Thomas W. Hertel · Uris Lantz C.

 Baldos

Global Change and

the Challenges
of Sustainably
Feeding a Growing
Planet
Global Change and the Challenges
of Sustainably Feeding a Growing Planet
Thomas W. Hertel • Uris Lantz C. Baldos

Global Change
and the Challenges
of Sustainably Feeding
a Growing Planet
Thomas W. Hertel Uris Lantz C. Baldos
Department of Agricultural Economics Department of Agricultural Economics
Purdue University Purdue University
West Lafayette, IN, USA West Lafayette, IN, USA

ISBN 978-3-319-22661-3 ISBN 978-3-319-22662-0 (eBook)

DOI 10.1007/978-3-319-22662-0

Library of Congress Control Number: 2015947927

Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
Alexander and Sarah Hertel-Fernandez,
in hopes of a sustainable future

Nagisa, Danilo and Sergia Baldos,

in thanks for love and encouragement
Foreword

Imagine this setting. A workshop at the Massachusetts Institute of Technology in

September 2002. A man standing at the blackboard, surrounded by a group of
economists, holding forth on some new idea. Ideas are flying as much as the chalk
dust. It was difficult not to get drawn in. This was how I first met Tom Hertel. He
was the man at the chalkboard. I was maybe the only non-economist in the group,
and whatever Tom was saying went over my head, but it was clearly exciting to the
others. Tom saw me at some point and said something like, “Navin, this is where
your data would come in and be most useful”. This has been the story of our rela-
tionship since—Tom sees connections that few others do.
The origins of this book tell the same story, of Tom’s ability to synthesize and see
the big picture. It started with his Presidential Address to the Agricultural and
Applied Economics Association (AAEA) in 2010. My memory of that event is of
numerous phone calls and emails from Tom hunting me down for more and more
references on various global change issues; his thirst was insatiable. Tom has been
fascinated by the analytical work done by climate scientists and ecologists in the
fields of climate change, land use change, and global food security; his unique con-
tribution has been his ability to see how economists could contribute. He has often
been the only economist attending annual meetings of the American Geophysical
Union or the Open Science Conference of the Global Land Project. Tom developed
his ideas into a course during his sabbatical at Stanford, collaborating with the agri-
cultural ecologist David Lobell; that effort eventually became this book.1 Tom
brings a decidedly economic perspective on issues, but he is able to make connec-
tions to other disciplines, including my own in global land use and environmental
change, in new and insightful ways.
A more recent story of my interactions with Tom will highlight the joy of inter-
disciplinary collaborations. In 2012, I wrote a Guest Editorial for the top ecology
journal Frontiers in Ecology and the Environment,2 titled “Can intensive farming

1
Tom was ably assisted in his vision by his coauthor, the fabulous Uris Baldos, then a fresh PhD
student and now a postdoctoral fellow, who did all the empirical work behind the scenes.
2
This journal is a benefit of ESA membership and distributed to all members.

vii
viii Foreword

save nature?” In this editorial, my coauthor and I argued that intensive farming
could have environmental benefits by sparing natural ecosystems from the chain-
saw, popularly termed a “land sparing” strategy. But we concluded that a few impor-
tant questions remain open:
A major criticism of land sparing is that there is scant empirical evidence that agricultural
intensification actually leads to land sparing in practice. In fact, intensification is often
accompanied by further expansion, as neighboring farmers adopt the novel practices. This
implies either that the demand for the agricultural product has increased because of intensi-
fication or that the land sparing has occurred in another region of the world. Whether inten-
sification can create increased demand of agricultural products is an important question; it
depends on whether demand remains elastic as supply increases….Measuring land sparing
through empirical analysis is thus confounded by: (1) the lack of a “control” situation to
compare against—perhaps cropland would have expanded even faster had it not been for
intensification; and (2) the fact that national-level analysis may fail to account for “leakage”
of land sparing to other parts of the world.

A keen reader would have noticed some economic language in the above para-
graph—references to demand elasticity—concepts whose importance I appreciated
better through my interactions with Tom. I sent the editorial to Tom suggesting this
as an area where an economist could make an important contribution. Tom responded
that he had already conducted an analysis of this issue and submitted a paper to an
economic journal, but his paper had been rejected on the basis that these ideas were
not new to economists. But these ideas were new to land change scientists! Tom and
I thus collaborated on a broader analysis, to develop an analytical framework using
an economic perspective, to investigate under what conditions agricultural intensi-
fication can lead to land sparing. The findings of our study that historical intensifica-
tion in Latin America and Asia did indeed spare land for nature, but future
intensification in Africa is not guaranteed to do so, were considered sufficiently
important and interesting to a broad audience that our paper3 was published in the
top journal, Proceedings of the National Academy of Sciences.
Clearly, natural scientists have a lot to gain from interdisciplinary collaborations,
especially with economists interested in the same problem. Global environmental
change science (climate science, land use science, ecology) is capable of saying if
humans do X, the consequence will be Y. But these disciplines have a limited ability
of projecting what, indeed, humans will do, and more importantly, how they will
respond to the ongoing environmental changes (i.e., the endogenous response, in
economic language). In systems language, these are feedbacks to the system that
include human responses, i.e., humans are active participants in the system, rather
than passive respondents. How humans respond to price fluctuations is, certainly, a
limited perspective on the complexity of human behavioral responses, but it is an
important one. The logic is simple. When the price of food increases, humans will

3
Hertel, T. W., Ramankutty, N., & Baldos, U. L. C. (2014). Global market integration increases
likelihood that a future African Green Revolution could increase crop land use and CO2 emissions.
Proceedings of the National Academy of Sciences of the United States of America, 111(38),
13799–13804.
Foreword ix

generally buy less of it, depending on their base economic circumstances. All of us
who have stood at the grocery store comparing the prices and labels of various
options can relate to this. People like me, who grew up in India eating little meat,
not because of religious or cultural reasons, but because we could not afford it, can
relate well to the price sensitivity of food consumption.
So what does this book have to offer? First, every chapter has a nugget of “eco-
nomic wisdom.” And when the authors discuss economic concepts, they go back to
the basics, to the supply–demand curves, and explain them in a way that is easily
accessible to a nonspecialist audience. I am going to take examples from four differ-
ent chapters to illustrate what I mean. In Chap. 2, we learn how income determines
consumption, and in particular how consumer response to food price and income
changes depends on base income levels. Here we learn, for example, that if you are
rich, and devoting a tiny portion of your net income to food, changes in food price
will have less influence on your purchasing behavior, than if you were poor and
devoting a large portion of your income to food.
In Chap. 3, Hertel and Baldos introduce the important concept of Total Factor
Productivity (TFP), how it is different from yields, and why it is important to con-
sider. I must admit that I am a recent convert myself to appreciating the value of
TFP. Until Tom’s influence, I had always thought that yields (production per unit
area) are what matter, because it is what we derive from the land. To me TFP
sounded like an economic “construct” with no real-world significance, i.e., some-
thing only economists cared about. Chap. 3 makes the case for why non-economists
take TFP seriously. Increasing yields, derived simply from increasing inputs, say
little about the efficiency of the agricultural system. It is only TFP, through compar-
ing outputs to inputs, that evaluates efficiency of the system (i.e., getting more by
simply putting more in is not improving efficiency). Given our current environmen-
tal challenges related to food, growing more food by adding more water and fertil-
izers is not a solution.
In Chap. 4, the authors discuss the importance of price effects in forecasting
future demand and future land use. Typical biophysical analysis would project
demand based on population and income growth and then estimate the land supply
response required (with implications for land use, depending on the assumptions on
intensification). But bringing in the economics allows for capturing the influence of
higher price on consumer response (typically to reduce consumption), which
reduces the land supply response. On the land supply side, economics allows one to
better capture how a farmer would respond to increasing prices by either increasing
input use, or through expanding land. As Hertel and Baldos put it, the price effects
on demand and input use act as “shock absorbers” to reduce the land use response
from what one might naively predict without consideration of economics.
In the final Chap. 11, the authors pull all of their insights together to illustrate the
impact of ignoring economic logic. They exercise the SIMPLE model under various
common assumptions used by other studies to illustrate their shortcomings (or
“error signatures” as they call it). They start their exercise with a historical validation
x Foreword

of the SIMPLE model! The exclamation mark at the end of the previous sentence
because economic models are widely criticized by other disciplines for not validat-
ing their model results. Hertel and Baldos have taken this criticism to heart and
performed one of the rare model validations in that literature. They then use this as
a baseline to illustrate how various assumptions by other studies (e.g., ignoring
price effects on demand or on intensification) lead to under- or over-estimating crop
production, yields, crop prices, and land use. The authors end the book with the
grand experiment, of SIMPLE model projection to 2050, finding continuing, albeit
slowing, cropland expansion and flat-to-declining crop prices. They also predict
large reductions in malnutrition owing to strong income growth, which will be fur-
ther aided by global market integration.
Another major contribution of this book is that, in addition to presenting useful
economic thinking, the authors also present useful reviews across a broad range of
issues, crossing into other disciplines in doing so. Based on their all-encompassing
framework presented in Fig. 1.4, they cover a range of topics ranging from popula-
tion and income growth (Chap. 2) to water (Chap. 5), climate change (Chap. 6),
environmental services (Chap. 7), biofuels (Chap. 8), livestock (Chap. 9), and nutri-
tion (Chap. 10). In each chapter, the authors review what we know about the func-
tioning of the system and the impacts of exogenous or endogenous forces. To take
one example, in Chap. 7, I learnt about some unique REDD and PES schemes,
alongside a discussion of land tenure challenges to implementation. In this chapter,
they even delve into the ecological literature on the land sparing versus sharing
debate. Similarly, in Chap. 6 on climate change, they review plant physiological
responses to increasing carbon dioxide concentrations and temperature (discussing
photosynthesis and respiration processes, influence of vapor pressure deficit on sto-
matal conductance, etc., issues I learnt as a graduate student in global environmental
science). Rarely do economists cross so deeply into other disciplines. Because of
such comprehensive coverage, I think this book will be useful not only to those
interested in an economic perspective, but also to those interested in obtaining a
broad overview of various issues surrounding the food security and environmental
sustainability challenge.
In closing, let me recount an experience from attending a workshop in
November 2008, the Ernst Strüngmann Forum on “Linkages of Sustainability.”
I was invited as part of the “Land” group, to consider the land limitations to meeting
the future resource needs of human societies. We also had groups representing the
themes of Water, Energy, and Materials. We had in-group discussions in parallel,
but we also circulated among other groups. During one session, the Energy folks
were visiting us in Land, and tiring of hearing us talk in jargon about land telecon-
nections, land use transitions, etc., asked us, “Just tell us, is there enough land to
meet the needs of people in 2050?” We replied, “Well, it depends. What do you
mean by needs?” They said it would be a population of nine billion with a North
American lifestyle. We responded that we would certainly run out of water before
we ran out of land to produce that much food. So we went to the Water group and
asked them if there would be enough water. They responded that there is certainly
enough water in the ocean, but desalination is very energy intensive. So they flipped
Foreword xi

the question back to the Energy group, to ask if there was enough energy.
The Energy group responded that there could certainly be enough energy, but it
would be expensive! It depends on how much one is willing to pay for energy. We
looked around our groups for an economist but could not find one. I sorely missed
Tom Hertel!

Vancouver, BC, Canada Navin Ramankutty

Professor of Global Food Security
and Sustainability
Liu Institute for Global Issues
and Institute for Resources
Environment, and Sustainability
University of British Columbia
Vancouver BC, Canada
Preface

How did this book come about? This book has its origins in Thomas Hertel’s
Presidential Address to the Agricultural and Applied Economics Association
(AAEA) in 2010. Uris Lantz Baldos was a new doctorate student at the time, and his
first research assignment was to build a library of reference materials for the under-
lying paper. We started with an outline not dissimilar from the table of contents for
this book. And as we began to populate this library, we realized that there was an
immense amount of literature bearing on the question of long-run agricultural land
use at a global scale. In putting this reading list together, digesting it, and preparing
the AAEA address, we gained a far greater appreciation for the complexity of the
factors driving the long-run availability or resources for global food production, and
especially competing demands for services from the world’s land resources. Over
the last few years, we have benefited greatly from interactions with members of
diverse intellectual communities, all working on this important question, but typi-
cally from very different perspectives. This book seeks to bring together many of
these diverse views of the fundamental drivers behind changes in long-run global
land use.
From our perspective as economists, the most striking feature of this literature
was that many of the most influential papers on the topic of future land use, and the
implications for food security and the environment, abstracted from economic
behavior. Furthermore, we had the distinct impression that, in those places where
economics was being brought to bear, authors were being excessively influenced by
near-term developments in commodity markets—and therefore not taking a suffi-
ciently long-run view of the problem. This led Hertel to develop a very simple,
analytical framework capturing global supply and demand for agricultural land,
with which he was able to illustrate the aspects of this problem that were being
ignored by the physical scientists and those that were likely being mis-estimated by
many of the economists. This framework is provided in Appendix B of this book
and it is utilized extensively throughout the book as a synthesis device for highlight-
ing key economic features of the long-run supply and demand for global agricul-
tural land.

xiii
xiv Preface

In light of the usefulness of this simple analytical framework, and given our
backgrounds in global economic modeling, we decided to go one step further and
implement a numerical version of this framework, with the goal of “putting some
numbers” on these competing forces shaping future land use. In so doing, we did
not seek to replicate the many valuable economic models of global agriculture and
land use—a community which has been expanding rapidly as of late. Existing mod-
els are typically of high dimension, covering many time periods, commodities, and
regions—some are now even solved at the level of individual grid cells! Rather, our
goal was to keep the model as simple as possible, staying as close as we could to the
analytical framework. Indeed, under a few simplifying assumptions, the entire
numerical model can still be collapsed down to a single equation, through succes-
sive substitutions.
We have named the numerical model SIMPLE, which stands for “a Simplified
International Model of Prices, Land use and the Environment” (see Appendix C).
We have found SIMPLE to be an extremely useful framework for exploring a wide
range of issues related to global land use, food security, and the environment, allow-
ing us to explore issues as diverse as the interplay between climate adaptation and
mitigation, the nutritional consequences of climate change, food waste, the impacts
of technological change on land use, and the prospects for long-term commodity
prices. In each case, the simplicity of the model has allowed us to explore the issue
from a new point of view, offering additional analysis, as well as comprehensive
uncertainty quantification.
Perhaps our greatest success with SIMPLE has been in the classroom, where we
have used this as a teaching tool for an interdisciplinary course title which now
shares the same title as this book. (The syllabus is freely available online at https://
mygeohub.org/courses/global_change.) This course was first developed in collabo-
ration with David Lobell, who hosted Hertel’s sabbatical at Stanford University in
2011–2012. The course was targeted at Master’s level students, and we had partici-
pants from across the campus, including Environmental Earth System Sciences,
Engineering, Law, Business, and Public Policy. Each week covered a different
theme, with an introductory lecture presented by an expert in that field, followed by
student-led discussion of the readings, and finally, discussion of the weekly lab
exercise which consisted of a series of structured simulations using the SIMPLE
model. Uris Lantz Baldos and Thomas Hertel prepared the labs, with input from
David Lobell and the course teaching assistant: Robert Heilmayr. The success of
this format led Hertel to propose the course when he returned to Purdue University.
As of this writing, it has been offered three times at Purdue, to students in Agricultural
Economics, Earth and Atmospheric Sciences, Engineering, Hydrology, Agronomy,
and Forestry and Natural Resources. This book is built upon these course offerings,
the weekly guest lectures by domain experts, the associated lab assignments with
SIMPLE, as well as our journal papers. It is designed for use in similar courses on
other campuses, as well as for use as a stand-alone resource for those seeking an
in-depth exposure to this field.
Preface xv

How should you use this book? As noted, this book was developed in the context
of an interdisciplinary course and it is well suited to serving as a textbook. The core
analytical framework is taken from economics, and as such, the course lends itself
to be led by a faculty member from that field. However, the level of economics
employed is sufficiently straightforward so that the course could also by offered by
qualified faculty in departments such as geography, earth systems science, or related
fields of study. In the current course offering at Purdue University, we devote one
week to each topic, beginning the week with an overview lecture by a faculty mem-
ber with expertise in this area, followed by student-led discussion of the readings—
leading into discussion of the lab assignments and a review of the basic principles
of economics.
We have half a dozen lab assignments spaced out over the first 10 weeks of the
semester. These are based on the SIMPLE model, which is detailed in Appendix C
of this book. The labs are drawn from the empirical examples at the end of each
chapter, sometimes consolidating several themes into one assignment in order to
avoid having too many labs. (A sample syllabus and labs, along with software for
implementation, are available online for free at https://mygeohub.org/courses/
global_change). Each lab exposes the students to a new concept in economics as it
applies to global land use. By the time they have completed these assignments, stu-
dents are well placed to develop their own class project, which they present at the
end of class and subsequently write up in lieu of a final exam. (For a partial list of
past class projects, see Appendix A.)
This book can also be used as a reference for those embarking on research into
global land use, food security, and the environment. Each chapter offers a compre-
hensive and up-to-date survey of the literature in the main areas bearing on future
land use for agriculture and the environment. As such, this is a good place to begin
reading and research in these subfields. Those undertaking research in this field may
also take an interest in the SIMPLE model itself (see the link above). The download-
able labs offer a natural launching pad for the development and exploration of new
hypotheses about the interplay between global environmental and economic forces
operating on global agricultural land use, food, and environmental security.

West Lafayette, IN, USA Thomas W. Hertel

Uris Lantz C. Baldos
Abbreviations

AEZ Agro-ecological zones

AgMIP Agricultural Model Intercomparison and Improvement Project
APEC Asia-Pacific Economic Cooperation
BAU Business as usual
bbl Barrel
Bgy Billion gallons per year
B Billion
C Carbon
CERES Crop environment resource synthesis model
CGE Computable general equilibrium models
CO2 Carbon dioxide
CO2e Carbon dioxide equivalent
CP Current policies
DEC Dietary energy consumption
DSSAT Decision Support System for Agrotechnology Transfer
FAO United Nations Food and Agriculture Organization
FAPRI Food and Agricultural Policy Research Institute
g Grams
GCM Global circulation model
GDP Gross domestic product
GFAR Global Forum on Agricultural Research
GHG Greenhouse gasses
GL Gigaliters
GLy Gigaliters per year
GR Green revolution
ha Hectare
HADGEM Hadley Centre Climate Model
IEA International Energy Agency
IMPACT International Model for Policy Analysis of Agricultural Commodities
and Trade
IPCC Intergovernmental Panel on Climate Change

xvii
xviii Abbreviations

IWSR Irrigation water supply reliability

Kcal Kilocalorie
Kg Kilograms
LPJmL Lund-Potsdam-Jena managed Land Model
M Million
Mcal Megacalorie
Mts Metric tons
NP New policies
OECD Organisation for Economic Co-operation and Development
PES Payments for environmental services
PFP Partial factor productivity
R&D Research and development
REDD Reducing emissions from deforestation and forest degradation
RFS2 Renewable Fuel Standard Program
SIMPLE A Simplified International Model of agricultural Prices Land use and
the Environment
TEM Terrestrial ecosystem model
TFP Total factor productivity
UNDP United Nations Development Programme
UNICEF United Nations Children’s Fund
UNPD United Nations Population Division
USDA United States Department of Agriculture
VPD Vapor pressure deficit
WASDE World Agricultural Supply and Demand Estimates
WHO World Health Organization
Contents

1 Overview of Global Land Use, Food Security

and the Environment .............................................................................. 1
1.1 Motivation ........................................................................................ 1
1.2 Historical Perspective ....................................................................... 3
1.3 Economic Framework ...................................................................... 6
References ................................................................................................. 10
2 Population and Income as Drivers of Global Change ......................... 13
2.1 Determinants of Population Growth ................................................ 13
2.2 Geographic Differences in Food Consumption Patterns .................. 16
2.3 Income as a Determinant of Food Consumption Patterns ................ 20
2.4 Analyzing the Global Food and Land Needs
Due to Population and Income Growth ............................................ 23
References ................................................................................................. 26
3 Productivity Growth and Yields in the Global Crops Sector.............. 27
3.1 Historical Yield Growth ................................................................... 27
3.2 Prospects for Further Increases in Yields ......................................... 29
3.3 Measuring Productivity: Partial vs. Total Factor Productivity ......... 32
3.4 Differential Sources of Productivity Growth in SIMPLE ................ 36
References ................................................................................................. 38
4 Economic Responses to Scarcity ............................................................ 41
4.1 Overview .......................................................................................... 41
4.2 Consumer Response to Rising Prices: The Price Elasticity
of Demand for Food ......................................................................... 42
4.3 Supply Response at the Intensive Margin: Market Price
Elasticity of Crop Yields .................................................................. 44
4.4 Supply Response at the Extensive Margin: Price Elasticity
of Cropland Supply .......................................................................... 46

xix
xx Contents

4.5 The Three Economic Responses at Work: Cropland Use

from Global Biofuels........................................................................ 51
4.6 The Role of Economic Responses in the Global
Land-Sparing Debate ....................................................................... 51
References ................................................................................................. 54
5 Water, Food and Environmental Security............................................. 57
5.1 Water Use in Agriculture.................................................................. 57
5.2 Irrigated Agriculture......................................................................... 59
5.3 Adjusting to Water Scarcity ............................................................. 62
5.4 Interactions Between Irrigation and Land Use:
Application of SIMPLE ................................................................... 63
References ................................................................................................. 66
6 Climate Change Impacts in Agriculture ............................................... 69
6.1 Overview .......................................................................................... 69
6.2 Quantifying Climate Impacts on Agriculture................................... 73
6.3 Adaptation to Climate Change ......................................................... 75
6.4 Analysis of Climate Impacts and Adaptation Using SIMPLE ......... 79
References ................................................................................................. 81
7 Land-Based Environmental Services .................................................... 85
7.1 Externalities and Land-Based Environmental Services ................... 85
7.2 Carbon Sequestration ....................................................................... 86
7.3 Setting Aside Lands for Biodiversity
and Other Ecosystem Services ......................................................... 88
7.4 What Role for Wildlife Friendly Farming? ...................................... 90
7.5 Economic Analysis of the Impacts of REDD
on the Food System .......................................................................... 92
References ................................................................................................. 97
8 Biofuels as a Driver of Long Run Land Use Change ........................... 99
8.1 Historical Perspective and Policy Context ....................................... 99
8.2 Market-Mediated Impacts of Biofuels on Global Land Use ............ 102
8.3 Projecting Global Agricultural Land Use to 2050
with Biofuels .................................................................................... 108
References ................................................................................................. 113
9 Livestock and Processed Foods .............................................................. 115
9.1 Overview .......................................................................................... 115
9.2 Livestock Production and Consumption .......................................... 115
9.3 Environmental Impacts of Livestock Consumption ......................... 117
9.4 Economic Implications of Livestock Consumption ......................... 118
9.5 Incorporating Livestock and Processed Foods into SIMPLE .......... 120
9.6 Projecting Livestock and Processed Food Demand
Using SIMPLE ................................................................................. 122
References ................................................................................................. 124
Contents xxi

10 Food Security and Nutrition .................................................................. 125

10.1 Overview ........................................................................................ 125
10.2 Undernutrition: Current Trends
and Human Impacts ....................................................................... 126
10.3 Climate Change Risks and Food Security ..................................... 128
10.4 Agricultural Productivity and Future Food Security:
A SIMPLE Application.................................................................. 131
References ................................................................................................. 137
11 Global Change and the Food System in 2050 ....................................... 141
11.1 Looking Back to Understand the Future ........................................ 141
11.1.1 Global Validation of SIMPLE Over the 1961–2006
Period ............................................................................... 142
11.1.2 Identifying the Error Signatures of Other Models ........... 144
11.1.3 Regional Production and the Failure
of the Integrated Markets Hypothesis .............................. 148
11.2 Globalization, Market Integration and the Environmental
Impacts of Technology ................................................................... 150
11.3 Assessing Long Run Global Food and Environmental
Security .......................................................................................... 152
11.3.1 Implications for Future Cropland Conversion ................. 153
11.3.2 Implications for Food Prices ............................................ 154
11.3.3 Implications for Nutritional Attainment .......................... 156
References ................................................................................................. 158

Appendix A Class Projects .......................................................................... 161

Appendix B Analytical Framework: A Theoretical Model of Long

Run Demand and Supply for Agricultural Land ................ 165

Appendix C Description of the SIMPLE Model ....................................... 171

Index ................................................................................................................. 181

Chapter 1
Overview of Global Land Use, Food Security
and the Environment *

1.1  Motivation

Over the past decade, there has been a resurgence of interest in the global farm and
food system and its contributions to feeding the world’s population as well as ensur-
ing the environmental sustainability of the planet. The 2007/2008 commodity crisis
underscored the vulnerability of the global food system to shocks from extreme
weather events, energy and financial markets, as well as government interventions
in the marketplace. We have learned that a “perfect storm” in which all these factors
coincide can have a devastating impact on the world’s poor, as well as putting con-
siderable pressure on natural resources and the environment. As we look ahead to
the middle of this century, will the world’s agricultural resource base be up to the
task of meeting the diverse demands being placed on it? Will food security still be a
challenge for large numbers of the world’s population? We examine this issue, pay-
ing particular attention to the land and water resource base available in the world
today, factoring in potential changes in the quality and quantity of these resources
in the wake of climate change, and taking account of competing, non-agricultural
demands on these limited resources.
The number of people which the world must feed is expected to rise from 7.2
billion in 2104 to more than 9 billion by mid-century. When coupled with signifi-
cant nutritional improvements for the 2.1 billion people currently living on less than
$2/day (World Bank, 2008, p. 1), this translates into a very substantial rise in the
demand for agricultural production. The FAO estimated the increased demand at 60 %

* This chapter draws heavily on Hertel’s Presidential Address to the Agricultural and Applied
Economics Association (Hertel, 2011).

© Springer International Publishing Switzerland 2016 1

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_1
2 1  Overview of Global Land Use, Food Security and the Environment

of current production, with a figure nearer 100 % in the developing countries

(Alexandratos & Bruinsma, 2012, p. 64). Over the past century, global agriculture
has managed to offer a growing population an improved diet, primarily by increas-
ing productivity on existing cropland. However, a number of authors have docu-
mented signs of slowing yield growth for key staple crops (Deininger & Byerlee,
2010, Box 2.1). And public opposition to genetically modified crops has slowed
growth in the application of promising biotechnology developments to food produc-
tion in some parts of the world (Paarlberg, 2008). At the same time, the growing use
of biomass for energy generation has introduced an important new source of indus-
trial demand in agricultural markets (Energy Information Agency, 2010). To com-
pound matters, water, a key input into agricultural production, is rapidly diminishing
in availability in many parts of the world (McKinsey & Co, 2009), and many soils
are degrading (Lepers et al., 2005).
In addition, agriculture and forestry are increasingly envisioned as key sectors
for climate change mitigation policy. When combined, farming and land use
change—much of it induced by agriculture—currently account for about one-­
quarter of global greenhouse gas emissions (Vermeulen, Campbell, & Ingram,
2012), but, if incorporated into a global climate policy, these sectors could con-
tribute up to half of all mitigation in the near term, at modest carbon prices
(Golub, Hertel, Lee, Rose, & Sohngen, 2009). Any serious attempt to curtail these
emissions will involve changes in the way farming is conducted, as well as plac-
ing limits on the expansion of agriculture—particularly in the tropics, where
much of the crop land conversion has come at the expense of forests, either
directly (Gibbs et al., 2010) or indirectly, via a cascading of land use requirements
with crops moving into pasture and pasture into forest (Barona, Ramankutty,
Hyman, & Coomes, 2010). Limiting the conversion of forests to agricultural
lands is also critical to preserving the planet’s biodiversity (Green, Cornell,
Scharlemann, & Balmford, 2005). Finally, agriculture and forestry are likely to be
the economic sectors whose productivity is most sharply affected by climate
change. This will shift the pattern of global comparative advantage in agriculture
(Reilly et al., 2007), and may well reduce the productivity of farming in precisely
those regions of the world where malnutrition is most prevalent, while increasing
yield variability and the vulnerability of the world’s poor (Ahmed, Diffenbaugh,
& Hertel, 2009).
In light of these multiple challenges facing the global farm and food system,
this book brings together the relevant literature bearing on each of these topics. A
key feature of this book is that it is built around a unifying analytical framework
which is revisited in each chapter and which offers readers a modeling tool
(SIMPLE: a Simplified International Model of agricultural Prices, Land use and
the Environment) for quantifying these diverse factors within a single, economic-
environmental framework which readers can download and run on their own.
While greatly simplified, this modeling tool is also rich enough to capture a wide
variety of different forces operating on the global farm, food and environmental
systems, thereby giving readers the opportunity to examine issues well-beyond
those discussed in this book.
1.2 Historical Perspective 3

1.2  Historical Perspective

Before looking ahead several decades to mid-century, it is useful to look backwards

in time to gain an historical perspective on the issue of long term land use change
around the world. We focus on land use—as opposed to water and other natural
resources—since this is the most visible source of competition between agriculture
and non-agriculture uses. The history of global land use is also better documented
than that of other natural resources.
Ramankutty et al. (2006) point out that people have been inducing land cover
change since the beginning of human history. Indeed, as they point out, large scale
burning of the landscape in western Africa was documented as early as 500
B.C. Foley et al. (2005) observe that most societies follow a common sequence of
land use regimes, as they portray in Fig. 1.1, excerpted from their paper. Here, the
timeline begins on the left, with land cover largely under natural ecosystems, and
progresses to the right, with frontier clearings for subsistence agriculture and small-­
scale farms, which in turn gives way to intensive agriculture, the development of
urban areas, and the advent of land devoted to protected recreational activities and
biodiversity. The world’s present day land cover is extremely diverse, and encom-
passes points along this entire continuum. However, the portion of the globe devoted
to intensive agriculture, managed forestry, protected lands and urban areas has
clearly been growing with time. Today, about one-third of the world’s land cover is
devoted to agriculture, one-third to forests and one-fifth to savannas, grasslands and
shrub-lands; the remainder is either barren or low productivity land, with urban
areas comprising about 1 % of the world’s land cover (Ramankutty, 2010).

Fig. 1.1  Land use transitions. Source: Foley et al. (2005)

4 1  Overview of Global Land Use, Food Security and the Environment

In their historical analysis of crop land cover changes over the twentieth century,
Ramankutty, Foley, and Olejniczak (2002) document very diverse patterns of
change across major regions of the world. In Europe, cropland cover actually
declined over the past century, and cropland increases were relatively modest in the
U.S., East Asia and tropical Africa. However, cropland cover expanded dramati-
cally over the twentieth century in Latin America, Canada, Australia and Southeast
Asia. Figure 1.2, taken from their paper, plots population against hectares of crop-
land in 1900 and shows that areas with high population also had larger cropland
areas, with the global average cropland area equaling 0.76 ha/capita. Indeed, in
1900, most regions of the world fell quite close to this ray from the origin in their
graph. However, by 1990, under the influence of greatly improved agricultural pro-
ductivity, the slope of this line had declined by more than half, to just 0.35 ha/capita,
and, while many regions still fall along ray from the origin, some have begun to
deviate more sharply from this relationship. Indeed, Fig. 1.2 shows that both Russia
and the U.S. experienced stronger than average cropland area growth, relative to
population, and both China and South Asia experienced the reverse, with relatively
more rapid population growth. These divergences from the 0.35 ha cropland/capita
line, drawn for 1990, were enabled in part by falling costs of international transport
and declining trade barriers, both of which have facilitated increased international
trade in food products. They also reflect the inherent responsiveness of yield growth
over time to economic forces, including population pressure. With more pressure on
the land, and more abundant labor, there were strong incentives in East Asia to cul-
tivate the land more intensively, thereby boosting yield growth, relative to the land
abundant, population scarce, regions of the world (Hayami & Ruttan, 1985).
Over the final two decades of the twentieth century, land cover change acceler-
ated to unprecedented levels. Lepers et al. (2005) document these trends and high-
light deforestation “hotspots” (Fig. 1.3). Most of these hotspots were in the tropics,
with the Amazon leading the way, followed by Southeast Asia, Central Africa and
Central America. Russia, too, shows some deforestation hotspots in their analysis.
These authors also examine areas of wind and water erosion, chemical and physical
deterioration and decreased vegetative cover. When it comes to these forms of land
degradation, the authors find that the region with the most intensive losses is the
Middle East and near Asia.
This degradation of existing crop land, when combined with the seemingly inex-
orable growth in demand for food, fiber and fuel, has led many observers to suggest
that the world may run out of land. Malthus (1888) is perhaps the best known cham-
pion of this position. However, he is by no means alone. It seems that every decade
or two, the specter of the world running out of land is raised. As recently as 1985,
Buringh wrote in the Philosophical Transactions of the Royal Society of London
that “Recent studies show that on a global scale all land reserves will be lost within
one century and reserves of highly productive land will be lost in twenty-five years.”
Yet here, three decades later at the time of this writing, we are far from this outcome.
Nonetheless, fears are once again surfacing surrounding the availability land to
meet the demands of a growing population and expanding demands for biofuels,
with many analysts suggesting that the world is now experiencing a ‘new normal’
1.2 Historical Perspective 5

Fig. 1.2  Population density per cropland hectare: 1900 vs. 1990. Source: Ramankutty et al. (2002)

of increased resource pressures and high prices (Food & Agriculture Organization
of the UN Media Centre, 2013; OECD/FAO, 2013; World Bank, 2013). Assessing
these issues is clearly a challenging task, and one which is not amenable to simple
trend projections. What is needed is a more rigorous framework for analysis—
which is the topic of the next section.
6 1  Overview of Global Land Use, Food Security and the Environment

Fig. 1.3  Deforestation hotspots: 1980–2000. Source: Lepers et al. (2005)

1.3  Economic Framework

The processes behind global land use change are complex, and vary across scales
(Ramankutty et al., 2006). Some are dominant at local scale but less prominent
when one moves to global scale analyses (e.g., local zoning laws), while other pro-
cesses are front and center in global analyses, but are rarely present in local analyses
(e.g., global economic growth). In this chapter, we take, as our organizing frame-
work, the simplified, comparative static economic framework for analysis of global
agricultural land use change developed by Hertel (2011) and detailed in Appendix
B. A visual summary is offered in Fig. 1.4 where the global farm and food sector is
conceptualized as one in which land (along with water and other natural resources—
we lump them all into land for ease of exposition) is combined with non-land inputs
in order to produce agricultural output to satisfy the world’s food, fiber and fuel
demands. Both non-land inputs (such as fertilizer) and cropland conversion have
significant environmental impacts, which need to be weighed against the contribu-
tions to global food security engendered by greater agricultural output.
There are three broad classes of perturbations to this global food economy which
will play a role in the context of this book: the percentage growth in demand (D) for
agricultural output (A), denoted ΔAD, the percentage growth in productivity of land
(L), which affects the demand (D) for this natural resource, ΔLD, and shifts in the
supply (S) of land (L) to farming, ΔLS. The first of these shocks is driven by changes
in population, income, energy prices and biofuel mandates (Baldos & Hertel, 2014).
1.3  Economic Framework 7

Fig. 1.4  Analytical framework

Crop productivity is a function of the biophysical environment, and is importantly

affected by climate change, water resource availability, and soil degradation.
Productivity is also influenced by technologies, which are a lagged function of
research and development expenditures (Alston, Pardey, & Ruttan, 2008; Fuglie,
2012). The third driver of this global farm and food system involves shifts in the
supply of land to farming. This can be affected by many different forces (Lambin,
2012), including urbanization, climate change (Cassman, Grassini, & van Wart,
2010) as well as the growing demand for environmental services provided by land
and associated policies (Golub et al., 2013; Steinbuks & Hertel, 2014).
In addition to producing food, fiber and fuel, global agriculture has important
environmental impacts. GHG emissions from agriculture and deforestation (much
of which is driven by agriculture) account for more than a quarter of global GHG
emissions (Baumert, Herzog, & Pershing, 2005). This link is portrayed in Fig. 1.4
by the arrows stemming from land conversion and from the intensification of agri-
cultural production (e.g., application of nitrogen fertilizer, flooding of rice paddies).
The conversion of natural lands to agriculture is also the predominant cause of bio-
diversity loss (Balmford, Green, & Scharlemann, 2005), and the intensification of
agricultural production often has serious impacts on water quality (Vitousek et al.,
2009). In short, agriculture is very significant driver of environmental change (Foley
et al., 2005).
Thus far we have discussed the exogenous drivers of change in the global food
system. However, in order to accurately predict changes in global land use, we must
also factor in the endogenous responses of the system to outside pressures. In particu-
lar, there are three important ‘margins’ of economic response that have a direct bearing
on global land use (see italicized entries in Fig. 1.4). The first of these is the demand
margin, which is typically represented by the absolute value of the priceelasticity of
8 1  Overview of Global Land Use, Food Security and the Environment

demand (D) for agricultural (A) products, ηAD. When faced with higher prices,
consumers purchase less food. This has serious implications for food security as the
largest proportional reductions tend to come in the poorest countries who can least
afford the higher food prices (Muhammad, Seale Jr., Meade, & Regmi, 2011).
The second economic margin of interest is termed the ‘intensive margin’ of sup-
ply response, denoted by the price elasticity of supply with respect to commodity
prices, holding area constant: ηAS,I. This parameter reflects the fact that farmers,
when faced with higher prices, tend to intensify production in order to boost crop
yields, with the size of this response varying by crop and region of the world
(Keeney & Hertel, 2008). The third economic margin of response to scarcity in the
global food system is the extensive margin of crop supply, ηAS,E, which relates to the
expansion of crop production, in response to price changes, through increasing the
area cultivated (Lubowski, 2002). For a given level of global demand, there is a
direct tradeoff between the intensive and extensive margins of supply. If yields can
be increased at a sufficiently rapid pace, then additional area conversion is not nec-
essary. Indeed, the most remarkable fact about global agricultural land use over the
past 50 years is that the vast majority of increased crop production has come from
intensification. Higher yields and more intensive rotations have accounted for 86 %
of the increase in global crop production (Alexandratos & Bruinsma, 2012), thereby
lessening the need for cropland area expansion.
The equations underlying the analytical framework in Fig. 1.4 are formally
derived in Appendix B of this book. However, the main insights from this frame-
work—insights which will surface repeatedly and in many different contexts
throughout this book—can be gleaned from the following equation which repre-
sents the long run equilibrium percentage change in global agricultural land use, qL*,
as a function of the three exogenous perturbations, as moderated by the economic
margins of response (Hertel, 2011):

qL* = [(DAD + DLS - DLD ) / (1 + h AS, I / h AS, E + h AD / h AS, E )] - DLS (1.1)

Equation (1.1) offers several immediate insights which are pertinent to the debate
about long run land use at global scale. Firstly, if there is simultaneously no scope
for intensification of production ( h A = 0 ) and no price responsiveness in demand
S,I

( h A = 0 ), then (e.g.) a 5 % exogenous growth in demand, net of yield growth, gets

D

translated into a rise in equilibrium agricultural land use of (DA - DL ) = 5 % . As we

D D

will see in Chap. 2, this expression is a pretty accurate characterization of much of

the biophysical literature on long run land use. Of course, the global shifts in demand
and supply will often be built up from extremely detailed analyses at the sub-­
national level (Alexandratos & Bruinsma, 2012). The drawback of this approach is
that it abstracts from the fact that such shocks might in fact lead to an endogenous
response on the part of the farm and food system in the face of induced economic
scarcity. This is where the price elasticities of demand and supply (and hence the
contributions of economists) come into play.
Most studies of global-scale land use change to date have either greatly simplified
or ignored the potential for such endogenous responses of demand and yield to the
1.3  Economic Framework 9

exogenous drivers of this system. This includes the important and influential future
prospective studies of the UN Food and Agriculture Organization (Alexandratos,
1995; Alexandratos & Bruinsma, 2012; Bruinsma, 2009). When these economic
responses to scarcity are ignored, long run growth in land use is simply the outcome
of the footrace between demand growth and yield growth. Hertel (2011) identifies
two critical limitations of such studies. Firstly, they abstract from the impact of
induced land scarcity due to urbanization or land degradation. Secondly, such purely
biophysical projections overstate the long run expansion of agricultural area. By
including the price responsiveness of yields and consumer demand, economic analy-
ses of land use boost the size of the denominator in Eq. (1.1) and thereby lead to
smaller equilibrium land use changes. In effect the potential for intensification of
production and demand reductions in response to scarcity serve as “shock-absorb-
ers”, dampening the amount of land employed in agriculture over the long run in the
face of net demand growth. Any outward shift in net demand boosts food prices,
which, in turn tempers the realized demand for agricultural products. The ensuing
rise in land prices serves to encourage the intensification of production, with agricul-
tural producers (and indeed the long run activities of the agricultural research
establishment) substituting variable inputs for land to achieve higher yields.
A second important insight from Eq. (1.1) is that, what matters from the point of
view of equilibrium land use in agriculture is not the absolute size of the intensifica-
tion and final demand elasticities, but rather their size relative to the land supply
elasticity (i.e., the extensive margin of land use), ηAS,E. If, by some accident of fate,
all three elasticities were equal in magnitude, then we would have the result that
each of these three margins of economic response would contribute equally to
dampening the growth in net demand; only one-third of net demand growth would
be translated into agricultural land use change. In addition to this increase, we would
have to make an adjustment for farmland removals (a reduction in equilibrium land
use) from agriculture yielding the following net change in long run land use:

qL* = [(DAD - DLD ) / 3] - 2 DLS / 3. (1.2)

We will return to Eq. (1.1) time and again throughout the book, as we seek to under-
stand either the implications of particular assumptions/omissions made in the stud-
ies considered, or alternatively, as we seek to back out the implications of results
pertaining to the long run use of land in agriculture, as revealed by the simulation of
large complex models.
In summary, the analytical framework outlined in this section illustrates the need
for an interdisciplinary approach to studies of global land use change. To assess the
growth in global demand for agricultural output, one needs estimates from many
sources including: demographers who estimate the evolution of fertility and death
rates, energy experts forecasting the demand for bioenergy, and nutritionists and
economists estimating future consumption patterns. The changes in agricultural
yields over the next 40 years will depend on the potential for genetic advances as
predicted by crop scientists, expansion of irrigated areas based on hydrological
studies, climate impacts based on climate and crop sciences, as well as agronomists’
10 1  Overview of Global Land Use, Food Security and the Environment

estimates of potential yield gains from increased fertilizer use. The economic and
institutional barriers to expanding production are also important and depend on
access to markets and government policies as estimated by land use scientists and
economists. Shifts in the supply of agricultural land will depend on future patterns
of urbanization, water shortages, degradation of current farmlands, and demands for
environmental services including parks, carbon sequestration and biodiversity. Most
of these factors vary greatly by locality. Understanding the global picture in the
context of such heterogeneity requires globally consistent, geospatial data on land
use, productivity, soils, climate, water availability, institutions, and so on.
In the subsequent chapters of this book, we will explore these underlying deter-
minants of the stylized framework laid out in Fig. 1.4 in more detail. Specifically, in
the next chapter we will explore the key historical drivers of global food demand—
population and per capita income (ΔAD). This is followed by a chapter on productiv-
ity growth (ΔLD). In Chap. 4 we focus on the price responsiveness of supplies at the
extensive (ηAS,E) and intensive (ηAS,I) margins. We then bring in other constraints on
supply, including water (Chap. 5), climate change (Chap. 6) and competing demands
for land (Chap. 7), which can result in a shift the land supply curve facing agricul-
ture (ΔLS). Chapters 8 and 9 discuss two other important drivers of change in the
global cropping system and natural environment, namely biofuels and livestock. Up
to this point in the book, we largely use the terms ‘agriculture’ and ‘crops’ synony-
mously. However, as we will see in Chap. 9, the livestock sector is an important
driver of global land use and environmental change as well as nutritional outcomes.
This is followed by a chapter (Chap. 10) on nutritional attainment and a concluding
chapter (Chap. 11) in which we bring all of these factors together in order to make
projections of global food, nutrition and environmental outcomes in 2050.
Supporting materials are provided in the appendices for those who wish to obtain a
deeper understanding of the theory behind the economic framework and replicate
the results reported throughout this book.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Ahmed, S. A., Diffenbaugh, N. S., & Hertel, T. W. (2009). Climate volatility deepens poverty vul-
nerability in developing countries. Environmental Research Letters, 4(3), 034004. http://doi.
org/10.1088/1748-9326/4/3/034004.
Alexandratos, N. (Ed.). (1995). World agriculture: Towards 2010. Rome, Italy: Food and
Agriculture Organization of the United Nations and Wiley. Retrieved from ­http://www.fao.org/
docrep/V4200E/V4200E00.htm.
Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050: The 2012 revision
(Working Paper No. 12-03). Rome, Itally: Food and Agriculture Organisation of the United
Nations.
Alston, J. M., Pardey, P. G., & Ruttan, V. W. (2008). Research lags revisited: Concepts and evi-
dence from U.S. Agriculture (Staff Paper No. 50091). University of Minnesota, Department of
Applied Economics. Retrieved from http://ideas.repec.org/p/ags/umaesp/50091.html.
References 11

Baldos, U. L., & Hertel, T. (2014). Bursting the bubble: A long run perspective on crop com-
modity prices. Retrieved from http://www.gtap.agecon.purdue.edu/resources/res_display.
asp?RecordID=4574.
Balmford, A., Green, R. E., & Scharlemann, J. P. W. (2005). Sparing land for nature: Exploring the
potential impact of changes in agricultural yield on the area needed for crop production. Global
Change Biology, 11(10), 1594–1605. http://doi.org/10.1111/j.1365-2486.2005.001035.x.
Barona, E., Ramankutty, N., Hyman, G., & Coomes, O. T. (2010). The role of pasture and soybean
in deforestation of the Brazilian Amazon. Environmental Research Letters, 5(2), 024002.
http://doi.org/10.1088/1748-9326/5/2/024002.
Baumert, K. A., Herzog, T., & Pershing, J. (2005). Navigating the numbers: Greenhouse gas data
and international climate policy (p. 122). Washington, DC: World Resources Institute.
Retrieved from http://www.wri.org/publication/navigating-the-numbers.
Bruinsma, J. (2009). The resource outlook to 2050. By how much do land, water use and crop
yields need to increase by 2050? In FAO Expert meeting on How to Feed the World in 2050.
Rome: Food and Agriculture Organisation of the UN.
Buringh, P. (1985). The land resource for agriculture. Philosophical Transactions of the Royal
Society of London. Series B, Biological Sciences, 310(1144), 151–159.
Cassman, K. G., Grassini, P., & van Wart, J. (2010). Crop yield potential, yield trends, and global
food security in a changing climate. In D. Hillel & C. Rosenzweig (Eds.), Handbook of climate
change and agroecosystems: Impacts, adaptation, and mitigation (Vol. 1, pp. 37–51).
Hackensack, NJ: World Scientific.
Deininger, K. W., & Byerlee, D. (2010). Rising global interest in farmland: Can it yield sustain-
able and equitable benefits? Washington, DC: World Bank.
Energy Information Agency. (2010). Annual energy outlook 2010 (Annual Energy Outlook No.
DOE/EIA-0383(2010)). Washington, DC: US Department of Energy.
Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R., … Snyder, P. K.
(2005). Global consequences of land use. Science, 309(5734), 570–574. http://doi.org/10.1126/
science.1111772.
Food and Agriculture Organization of the UN Media Centre. (2013, April 12). Feeding nine billion
in 2050. Retrieved May 23, 2013, from http://www.fao.org/news/story/en/item/174172/icode/.
Fuglie, K. O. (2012). Productivity growth and technology capital in the global agricultural econ-
omy. In K. O. Fuglie, S. L. Wang, & V. E. Ball (Eds.), Productivity growth in agriculture: An
international perspective (pp. 335–368). Cambridge, MA: CAB.
Gibbs, H. K., Ruesch, A. S., Achard, F., Clayton, M. K., Holmgren, P., Ramankutty, N., & Foley,
J. A. (2010). Tropical forests were the primary sources of new agricultural land in the 1980s
and 1990s. Proceedings of the National Academy of Sciences of the United States of America,
107(38), 16732–16737. http://doi.org/10.1073/pnas.0910275107.
Golub, A. A., Henderson, B. B., Hertel, T. W., Gerber, P. J., Rose, S. K., & Sohngen, B. (2013).
Global climate policy impacts on livestock, land use, livelihoods, and food security. Proceedings
of the National Academy of Sciences of the United States of America, 110(52), 20894–20899.
http://doi.org/10.1073/pnas.1108772109.
Golub, A., Hertel, T. W., Lee, H.-L., Rose, S., & Sohngen, B. (2009). The opportunity cost of land
use and the global potential for greenhouse gas mitigation in agriculture and forestry. Resource
and Energy Economics, 31(4), 299–319. http://doi.org/10.1016/j.reseneeco.2009.04.007.
Green, R. E., Cornell, S. J., Scharlemann, J. P. W., & Balmford, A. (2005). Farming and the fate of
wild nature. Science, 307(5709), 550–555. http://doi.org/10.1126/science.1106049.
Hayami, Y., & Ruttan, V. W. (1985). Agricultural development: An international perspective.
Baltimore, MD: Johns Hopkins University Press.
Hertel, T. W. (2011). The global supply and demand for agricultural land in 2050: A perfect storm
in the making? American Journal of Agricultural Economics, 93(2), 259–275. http://doi.
org/10.1093/ajae/aaq189.
Keeney, R., & Hertel, T. W. (2008). Yield response to prices: Implications for policy modeling
(Working Paper No. 08-13). Purdue University.
Lambin, E. F. (2012). Global land availability: Malthus versus Ricardo. Global Food Security,
1(2), 83–87. http://doi.org/10.1016/j.gfs.2012.11.002.
12 1  Overview of Global Land Use, Food Security and the Environment

Lepers, E., Lambin, E. F., Janetos, A. C., DeFries, R. S., Achard, F., Ramankutty, N., & Scholes,
R. J. (2005). A synthesis of information on rapid land-cover change for the period 1981–2000.
BioScience, 55(2), 115–124. http://doi.org/10.1641/0006-3568(2005)055[0115:ASOIOR]
2.0.CO;2.
Lubowski, R. (2002). Determinants of land use transitions in the United States: Econometrics
analysis of changes among the major land-use categories. Cambridge, MA: Harvard University.
Malthus, T. R. (1888). An essay on the principle of population (9th ed.). London, UK: Ballantyne
Press.
McKinsey & Co. (2009). Charting our water future: Economic frameworks to inform decision-­
making. 2030 Water Resources Group: McKinsey & Co.
Muhammad, A., Seale, J. L., Jr., Meade, B., & Regmi, A. (2011). International evidence on food
consumption patterns: An update using 2005 International Comparison Program Data
(Technical Bulletin No. TB-1929) (p. 59). Washington, DC: Economic Research Service, US
Department of Agriculture. Retrieved from http://www.ers.usda.gov/Publications/TB1929/.
OECD/FAO. (2013). OECD-FAO agricultural outlook 2013-2022 (p. 326). OECD/FAO. Retrieved
from http://dx.doi.org/10.1787/agr_outlook-2013-en.
Paarlberg, R. L. (2008). Starved for science: How biotechnology is being kept out of Africa.
Cambridge, MA: Harvard University Press.
Ramankutty, N. (2010). Agriculture and forests: Recent trends, future prospects. In T. Graedel &
E. van der Voet (Eds.), Linkages of sustainability (Vol. 4, pp. 11–31). Cambridge, MA: MIT
Press.
Ramankutty, N., Foley, J. A., & Olejniczak, N. J. (2002). People on the land: Changes in global
population and croplands during the 20th century. AMBIO: A Journal of the Human
Environment, 31(3), 251–257.
Ramankutty, N., Graumlich, L., Achard, F., Alves, D., Chhabra, A., DeFries, R. S., … Turner,
B. L. (2006). Global land-cover change: Recent progress, remaining challenges. In Land use
and land cover change (pp. 9–39). The Netherlands: Springer. Retrieved from http://dx.doi.
org/10.1007/3-540-32202-7_2.
Reilly, J., Paltsev, S., Felzer, B., Wang, X., Kicklighter, D., Melillo, J., … Wang, C. (2007). Global
economic effects of changes in crops, pasture, and forests due to changing climate, carbon
dioxide, and ozone. Energy Policy, 35(11), 5370–5383.
Steinbuks, J., & Hertel, T. W. (2014). Confronting the food-energy-environment trilemma: Global
land use in the long run (No. WPS6928) (pp. 1–87). The World Bank. Retrieved from http://
documents.worldbank.org/curated/en/2014/06/19696268/
confronting-food-energy-environment-trilemma-global-land-use-long-run.
Vermeulen, S. J., Campbell, B. M., & Ingram, J. S. I. (2012). Climate change and food systems.
Annual Review of Environment and Resources, 37(1), 195–222. http://doi.org/10.1146/
annurev-environ-020411-130608.
Vitousek, P. M., Naylor, R., Crews, T., David, M. B., Drinkwater, L. E., Holland, E., … Zhang,
F. S. (2009). Nutrient imbalances in agricultural development. Science, 324(5934), 1519–1520.
http://doi.org/10.1126/science.1170261.
World Bank. (2008). Agriculture for development (pp. 1–386).
World Bank. (2013, April 15). Food crisis. Retrieved May 23, 2013, from ­http://www.worldbank.
org/foodcrisis/bankinitiatives.htm.
Chapter 2
Population and Income as Drivers
of Global Change

2.1  Determinants of Population Growth

For as long as there have been humans on the earth, demography has been a driver
of global environmental change. More people required more land—either for hunt-
ing and gathering, or eventually for the cultivation of crops. Demography, the sci-
ence of how human populations grow and evolve, offers a number of key insights
which are critical for understanding the likely changes in global food and environ-
mental pressures in the coming decades. For most of human history, the birth and
death rates were similar and both quite high (Fig. 2.1), resulting in very modest
population growth—at least compared to growth rates experienced over the past
century. However, with the advent of the industrial and agricultural revolutions as
well as modern medicine, improved sanitation, nutrition and health care, death rates
in the modernizing economies fell, while birth rates remained relatively unchanged—
at least initially. This opened a gap between the two, resulting in sharply increasing
population growth rates as shown in Fig. 2.1.
As long as families rely on child labor for farming and employment in the family
business, there is relatively little incentive to reduce the fertility rate—the average
number of children that an adult woman will give birth to. In many cultures, children
are also viewed as a form of old-age insurance—particularly before pensions and
social security are available to the population. When coupled with high child mortal-
ity rates (or at least the expectation of high rates), parents are loath to reduce the
number of children which they have. Even after parents decide to limit their number
of children, limited access to contraceptives, as well as traditional social and reli-
gious norms, can make it difficult to implement family planning in practice.
With mortality rates falling and fertility rates remaining high for a time, societies
in this stage of the demographic transition can experience explosive population
growth. This is the situation in much of Africa today. Fertility rates on the continent
remain as high as seven children/woman in the poorest countries of West Africa
(e.g., Niger), leaving the continent poised to account for 50 % of global population

© Springer International Publishing Switzerland 2016 13

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_2
14 2  Population and Income as Drivers of Global Change

Fig. 2.1  The stylized

demographic transition
model proposed by
Thompson (1929)

growth over the 2010–2050 period, even though it only accounted for 15 % of global
population at the start of this period (Bloom, 2011).
Of course, as countries become wealthier, a variety of factors contribute to the
decline in the birth rate. Improved social insurance reduces the need to rely on chil-
dren in retirement, the modernization of agriculture and the banning of child labor
in industry also reduce the economic incentive to have children, while urbanization
and higher educational standards raise the economic cost of raising children. In
addition, the education of women tends to result in later marriage, while improving
access to, and effectiveness of, contraception. Female education and career develop-
ment also raises the opportunity cost of temporarily leaving the labor force for preg-
nancy and childcare. All of these drivers of demographic change are encouraged by
rising incomes, and, in turn, they contribute to further income gains, thereby result-
ing in a virtuous circle through which fertility rates can fall sharply. Indeed, this is
already evident in some of the wealthier and more urbanized countries in Africa,
such as Tunisia, where fertility rates are close to replacement levels (around two
children/woman).
Better healthcare and improved sanitation have lengthened human life expec-
tancy; thus, it takes many decades for any significant demographic change to play
itself out. For example, the post-World War II baby boom in the United States cre-
ated echoes that continued to be felt as the baby-boomers had children, and then
these children have in turn had their own children. This is a very important outcome
of the demographic transition—commonly termed the demographic dividend. With
the decline of infant mortality, the share of population under 15 years of age grows
rapidly. If these young people are educated and can find jobs, then there are good
prospects for a boost in average per capita economic growth stemming from the fact
that the working age share of the population is temporarily elevated. Figure 2.2
illustrates this point, showing the number of individuals in each decadal age cohort
in the world’s developing vs. developed countries. Clearly the share of the popula-
tion in the developing countries that is of working age is much higher, and will grow
2.1 Determinants of Population Growth 15

Fig. 2.2  The demographic dividend. Source: Roberts (2011), as created from UNPD (2011)

in the coming decades, while the elderly represent just a small share of the total
population.
The demographic dividend has been contributed to the ‘East Asian economic
miracle’ experienced during the latter part of the twentieth century. Bloom (2011)
contrasts Indonesia and Nigeria—two of the world’s most populous countries—over
the period 1960–2010. In 1970, Indonesia’s ratio of working age to non-­working age
populations began to increase, rising from less than 1.2 in 1970 to more than 2.0 in
2010. This was accompanied by strong growth in the rate of increase in GDP/capita
in that country. In contrast, Nigeria’s ratio of working to non-working age popula-
tion fell from about 1.25 in 1960 to just above 1.0 in 1985, before rebounding.
However, it is not expected to reach 1.25 until 2030. Nigeria’s GDP/capita growth
rate has remained flat over most of this period. Of course there are many other fac-
tors at work in determining a country’s economic growth. However, there is no
doubt that the share of working age individuals in the population can play an impor-
tant role. And, of course, this entire process is reversible in the face of declining
birth rates, such as those being observed in Japan and parts of Western Europe. In
these regions, the increasing ratio of elderly to working age population can results in
a significant decline in national productivity and create a significant drag on eco-
nomic growth, leading to what has been dubbed a ‘demographic winter’.
For purposes of the analysis undertaken in this chapter, where we consider popu-
lation and income as exogenous drivers of change, we look at the growth rate of
population between the present and 2050 which, by region, as summarized in
Fig.  2.3. As anticipated from the forgoing discussion, virtually all the growth in
world population from 2010 to 2050 will take place in the developing countries,
with the most rapid rate of growth arising in Africa. Of course, there is considerable
uncertainty in these projections depending on how fertility rates evolve over the
coming decades. And the demographic uncertainty becomes much more pronounced
as we move out to the end of the century. Indeed, UNPD (2013) shows that, if fertil-
ity rates were to remain unchanged from 2010, world population in 2100 would
exceed 25 billion! On the other hand, if total fertility rates in 2100 were to fall
16 2  Population and Income as Drivers of Global Change

Fig. 2.3  World population growth, 1950–2050. Source: Roberts (2011) and UNPD (2011)
medium growth variant

below replacement levels, to 1.6 children/woman, as they have done in some devel-
oped countries, then global population could be as low as six billion in 2100.
Needless to say, the difference between feeding 6 billion people and feeding 25
billion in 2100 is staggering! Fortunately for us, we focus on 2050—a period for
which the demographic picture has largely already been cast, with expectations of
about 9 billion people on the planet at the mid-century mark.
Before leaving the topic of demographics, it is important to point out the implication
of a slowing global population growth rate for incremental global food and resource
requirements. Figure 2.4 (based on population projections made in 2000) illustrates
this point. Here, the declining growth rate translates into sharply declining annual
increments to global population. Indeed, according to this figure, we have already
passed the period (i.e. 1980s) during which the largest absolute number of people were
being added annually to the planet: more than 800 million people per year. This figure
is expected to drop sharply, falling to less than half that amount in the next few decades.
Of course, the future of global food demand is not just a question on how many more
mouths must be fed, but also what do these individuals actually consume and by how
much. This brings us to the determinants of per capita consumption around the world.

2.2  Geographic Differences in Food Consumption Patterns

Given the emphasis in this book on feeding the world sustainably, it is critical to
understand the stylized trends in food consumption patterns. One of the most com-
monly cited patterns of food demand change pertains to the differences which arise
when individuals or households move from the country-side to urban areas—particu-
larly for those households in developing countries that move from a relatively
2.2  Geographic Differences in Food Consumption Patterns 17

Fig. 2.4  Annual increments to global population (10-year average), 1750–2050. Source: UNPD
(2000, 2011)

traditional, peasant farm and to a modern city, as has been the case for millions of
migrant workers in Asia over the past several decades. Higher urban wages and lon-
ger work hours increase the opportunity cost of food preparation within the house-
hold; thus, it is expected that urban household dine outside more often and consume
more ready-to-eat and processed products—perhaps also a greater diversity of prod-
ucts. In light of the fact that the percentage of the world’s population residing in
urban areas is projected to rise from about 50 % today to 70 % in 2050 (UNPD,
2011), this could play an important role in global consumption patterns.
Timmer, Block and Dawe explored the differences in rural and urban consump-
tion patterns in the specific case of rice in Indonesia, India and Bangladesh. They
found that rice consumption—a traditional staple in these countries—is more pre-
dominant in rural areas. For example, rural households in the Jiangxi Province con-
sumed more than three times the amount of rice as do urban households, where
wheat, oils, livestock products and fruits and vegetables are more predominant.
Rural households often get 70 % of their daily calories from rice, whereas the urban
households seek greater dietary diversity. They also looked at rice consumption pat-
terns across income quintiles within the rural and urban areas. In all three countries
which they investigated, rice consumption continued to rise amongst the poorest 60
% of rural population. These impoverished households are seeking additional calo-
ries. However, in Indonesia, they found that nearly all of the urban population is
reducing its rice consumption. India is somewhat poorer, on average, and there the
poorest urban quintile is still increasing its rice consumption. In Bangladesh, the
poorest of the three countries which the authors surveyed, urban rice consumption
is only declining in the wealthiest quintile of the population. This evidence suggests
that income is a key underlying driver of food consumption patterns.
18 2  Population and Income as Drivers of Global Change

Table 2.1  Per capita food consumption by region (in grams/capita/day)

Food Commodities
Regions Crops Livestock Processed foods Total
Australia/New Zealand 1008 758 487 2253
Canada/U.S.A. 1154 786 537 2477
Central America & the Caribbean 992 413 277 1682
South America 952 505 316 1773
European Union+ 1151 622 443 2216
Eastern Europe 1262 605 351 2219
Middle East 1282 297 120 1698
North Africa 1468 220 112 1800
Sub-Saharan Africa 1069 89 155 1313
Southern Africa 816 288 292 1396
Central Asia 1149 455 78 1682
South Asia 809 214 89 1112
Southeast Asia 916 122 139 1176
China/Mongolia 1603 280 139 2022
Japan/Korea 1000 331 270 1601
Source: FAOSTAT (2013), and processed by the authors

Table 2.1 reports per capita consumption of crops, livestock and processed food
across the world based on the 15-region version of the SIMPLE model. Consumption
is broken into three categories: crops, livestock products and processed food and
these are reported as consumption by weight in grams/day/capita based from the
Food Security Statistics by FAO (2012). Focusing on the last column of Table 2.1,
we see that the total intake of food varies from 1112 grams/capita/day in South Asia
to more than twice that amount in Canada and the United States. Generally speak-
ing, regions with higher per capita incomes also have higher consumption levels—a
point to which we will return shortly. Equally important to highlight is the fact that
the share of composition of food consumption vary greatly across regions. In much
of Africa, the average consumer gets more than 80 % of their food (measured by
weight) directly from crops, whereas in Australia/New Zealand and Canada/U.S.A.,
this share is less than 50 %. Of course, this deficit in crop consumption is offset by
greater demand for livestock and processed foods. For example, in Australia/New
Zealand, a third of total food consumption, by weight, is comprised of livestock
products. Since livestock require feed and/or pasture, in the case of ruminants, a diet
rich in livestock products is much more demanding of the world’s land resources.
To understand the global food and land use implications of both the level and
composition of food consumption and size of population, we look at impacts of add-
ing more people across different parts of the world in the SIMPLE model (Table 2.2).
Specifically, we examine the changes in global crop demand and crop land use as a
consequence of adding ten million more people in each successive region while
abstaining from the dampening effect of rising prices on food demand. As can we
seen from this table, the ‘weight’ of consumers in different regions of the world on
global changes in crop demand and crop land use varies greatly. Feeding an addi-
2.2  Geographic Differences in Food Consumption Patterns 19

Table 2.2  Impact of adding ten million people in just one region
Variable
Crop output Crop land use
Region (in M Mts) (in 1000s ha)
Australia/New Zealand 32 985
Canada/U.S.A. 28 879
Central America & the Caribbean 13 405
South America 22 680
European Union+ 21 648
Eastern Europe 17 543
Middle East 17 536
North Africa 14 440
Sub-Saharan Africa 13 396
Southern Africa 10 308
Central Asia 11 344
South Asia 9 268
Southeast Asia 24 745
China/Mongolia 18 549
Japan/Korea 7 220
Note: Population is increased by ten million, one region at a time, and the
resulting global impacts are recorded here. Prices are kept constant in
these simulations
Source: Authors’ calculations using the SIMPLE model

tional ten million people in regions which currently have high levels of per capita
food consumption and have greater consumption of livestock and processed foods
(e.g. Australia/New Zealand and Canada/U.S.A., see Table 2.1), will require greater
expansion in global crop production and cropland use (32–28 M tonnes and 9.9–
8.8 M hectares, respectively). On the contrary, adding ten million more people in the
developing world wherein per capita food consumption is relatively low (e.g. Sub-­
Saharan Africa and South Asia) results in more modest impacts on global food
demand and land use. This consequence of regional consumption differences will
come into play later on as we consider the impact of future population growth on the
global food system and the environment. These global changes will also become
more pronounced once we consider the impacts of rising incomes on the level and
composition of food consumption as majority of the populace residing in the
­developing world will begin to consume more livestock and processed food, a phe-
nomenon which we currently observe in rapidly growing countries such as China.

2.3  Income as a Determinant of Food Consumption Patterns

In light of the fact that we seek to project global consumption patterns forward to
the year 2050, it is important to know how these consumption patterns are likely to
change over time. When Latin America becomes as wealthy as the United States
20 2  Population and Income as Drivers of Global Change

and Canada are today, will their consumption patterns looks like those currently
observed in these high income regions? Making such predictions relies on the idea
of a stable set of consumer preferences which results in changing consumption pat-
terns as per capita income rises. Southgate, Graham, and Tweeten (2010) summa-
rize the stylized relationships between living standards and consumption patterns.
At lower living standards, one finds the consumption patterns characterized by the
poorest countries in Sub-Saharan Africa and South Asia in which the bulk of the
food consumed comes in the form of cereals and other carbohydrates, with most of
the protein coming from beans, lentils and other plant-based protein sources. As
living standards rise, the total volume of food consumed also rises. But more dra-
matic than this aggregate volume increase is the change in dietary composition.
Livestock products assume a much more prominent place in the diet, as do sugars,
edible oils and fresh fruit and vegetables.
The common preference hypothesis embedded in the work of Southgate et al.
(2010) has been formally tested by econometricians studying global consumer
demands. For example, Clements and Chen (1996) conclude their paper on this
subject with the following observation: “Taken as a whole, the results reveal a sur-
prising degree of similarity in international consumption patterns and support the
idea that tastes are constant.… with respect to broad commodity groups.” They find
that per capita income is a very good predictor of the share of average households’
budgets devoted to food as a whole. Of course, income alone cannot fully explain
the mix of cereal, livestock and plant products which a household typically con-
sumes since other factors such as geography, lifestyle, culture and religion also
influence the composition of households’ food demand.
Capitalizing on this idea of common international preferences across aggregated
commodity categories, many authors has estimated how average per capita con-
sumption across countries is likely to evolve as a function of very large changes in
income—such as we are likely to see over the next several decades in much of the
developing world. Figure 2.5, excerpted from Reimer and Hertel (2003), reports the
results of one such exercise. Here, food products are broken up into two categories:
meat, dairy and fish, on the one hand, and crops, processed foods, beverages and
tobacco on the other. The third category of spending for which the total budget share
is plotted refers to a broad aggregate of non-food items, including services associ-
ated with housing, medical care, education and recreation. Each point in the figure
refers to a country in their sample and they are arranged along the horizontal axis
according to the logarithm of per capita private household spending. The vertical
axis displays the share of total spending devoted to these three categories.
There are several noteworthy points to be drawn from this figure. First of all, the
share of overall household expenditure devoted to crops at the lowest income levels
is very high indeed. However, this budget share falls rapidly as income rises. (Note
that a falling budget share does not necessarily translate into falling consumption,
since income is simultaneously rising.) On the other hand, livestock (including fish)
consumption actually increases its share of total consumers’ budgets as incomes
rise. This rising share is not enough to offset the decline in the crops aggregate, so
the overall food share falls, but it is indicative of the kind of dietary transition men-
2.3  Income as a Determinant of Food Consumption Patterns 21

Fig. 2.5  Predicted per capita total budget shares for selected products and services across the
world. Source: Reimer and Hertel (2003) based on estimated demand system using data from the
International Comparison Project (1996)

tioned by Southgate et al. (2010). At the highest income levels, the combined share
of these two food categories is very small indeed. It is also clear from Fig. 2.5 that,
while low income households spend relatively little on housing, education, recre-
ation and medical services, this rises steadily as a share of total expenditure, and it
rises even more sharply as an overall expenditure (due to the rising income levels).
In the SIMPLE model, and therefore in the projections used throughout this
book, we will rely on international cross-section relationships such as these in order
to predict how consumption behavior is likely to change as households become
richer. Given the particular requirements of the model, we draw on the estimates
presented in Muhammad, Seale Jr., Meade, and Regmi (2011) to determine how
consumption patterns evolve with income growth. Figure 2.6 plots the underlying
relationship between the per capita income and the share of consumers’ food budget
devoted to crops, livestock and processed foods given per capita incomes in 2006
and in 2050 (projected). In general, we see that food budget shares for food crops
are relatively smaller than shares for livestock and processed foods. Going forward
to 2050, additional income will be spent on the consumption of livestock and
processed food, resulting in increased food budget shares for these commodities.
On the contrary, the food expenditure share for food crops is expected to decline in
the future as household’s purchases of meats and dairy products continue to rise.
In order to further understand the relationship between income and consumption,
we report in Table 2.3 the impact on food consumption of a 10 % rise in per capita
income in each region. Note that the changes in per capita food demand are less than
22 2  Population and Income as Drivers of Global Change

Fig. 2.6  Predicted per capita food budget shares for crops, livestock and processed foods. Source:
Authors’ calculations, based on the estimated consumer demand elasticities from the Muhammad
et al. (2011) and income growth projections from Fouré et al. (2013)

Table 2.3  Impact of boosting income by 10 %, by region

Per capita food consumption (in %)
Region Crops Livestock Processed food
Australia/New Zealand −1.5 1.1 1.1
Canada/U.S.A. −1.7 0.9 0.9
Central America & the Caribbean  0.1 2.6 2.8
South America  0.4 2.8 3.1
European Union+ −1.4 1.2 1.2
Eastern Europe  0.4 2.8 3.1
Middle East  0.1 2.6 2.9
North Africa  1.1 3.5 3.9
Sub-Saharan Africa  2.4 4.7 5.3
Southern Africa  0.3 2.7 3.0
Central Asia  2.2 4.5 5.1
South Asia  2.1 4.4 5.0
Southeast Asia  1.4 3.8 4.2
China/Mongolia  1.2 3.6 4.0
Japan/Korea −1.5 1.2 1.2
Note: Per capita income is raised by 10 % one region at a time and the
resulting change in consumption in that region is recorded. Prices are
unchanging
Source: Authors’ calculations using the SIMPLE model

the rise in incomes as food demand is typically income inelastic. Changes in per
capita demand vary across regions, reflecting the differences in income levels,
responsiveness of food demand to income, and composition of consumption pat-
terns across the globe. We see that in Africa and South Asia, parts of the world
which have relatively low per capita food consumption, have stronger increases in
per capita food consumption for all food types compared to regions wherein food
2.4  Analyzing the Global Food and Land Needs Due to Population and Income Growth 23

consumption in high. This highlights the importance of income in shaping per cap-
ita demands in the developing world, particularly in regions which face food scar-
city. Looking at the changes across food types, the income responsiveness of food
demand for livestock and processed foods are generally higher than for crops. Per
capita consumption of staple crops is even declining in the developed regions (e.g.
Canada/U.S.A. and European Union+). The impact of rising incomes on both the
level and composition of food consumption will have important implications on
future global food needs as well as land use requirements and these impacts will be
further magnified with global population slated to exceed nine billion in 2050.

2.4  A
 nalyzing the Global Food and Land Needs
Due to Population and Income Growth

We conclude this chapter by taking a closer look at how the key drivers of food
demand influence the future global farm and food system. We start with our base
data in 2006 and establish two future scenarios with varying rates of population and
per capita income growth rates. In particular, we are interested in the effects of his-
torical (1961–2006) and future (2006–2051) trends of population and per capita
income on the global crop demand and crop land use. We rely on the 2012 World
Population Prospects (UNPD, 2013) in calculating both historical and future popu-
lation growth rates. The latter is calibrated to the medium variant growth rate—the
median among the set of available projections which assumes declining fertility
rates for regions with relatively large family size and slightly higher rates for coun-
tries wherein observed fertility is lower than replacement levels (i.e. roughly two
children per woman). We then combine population data with the historical as well
as future projections of real GDP to construct the growth rates of per capita income.
Historical income rates are taken from the World Development Indicators (2013)
while future projections are based on Fouré, Bénassy-Quéré, and Fontagné (2013).
In Fig. 2.7, we develop the economic intuition for how these key drivers of food
demand affect the global crop markets. Growth in population and income can be
illustrated as an outward shift in global crop demand (ΔAD). If we assume that the
equilibrium market price does not change ( P = P* ), then the shift in the global crop
demand curve needs to be accompanied by a shift in the supply curve of crops
(right, Fig. 2.7). With fixed market prices, the increase in global crop demand due
to population and income growth is fully transmitted resulting into an equivalent
increase in the market equilibrium crop quantities (D DA = q A ). Alternatively,
allowing for market prices to rise in response to rising global crop demand (left,
Fig. 2.7) will dampen the rise in the equilibrium market crop quantities (D DA > q A ).
It is important to note that the scenarios illustrated in this section cannot be con-
strued as projections yet, as we are just looking at one side of the picture—namely
demand. To see this clearly, let us return to Eq. (1.1) in Chap. 1 which portrays the
change in global land use as a function of various drivers as well as economic
responses to scarcity:
24 2  Population and Income as Drivers of Global Change

Fig. 2.7  Crop market equilibrium without (left) and with (right) fixed prices

qL* = [(D DA + D SL - D DL ) / (1 + h AS , I / h AS , E + h AD / h AS , E )] - D SL (1.1)

In the projections undertaken here, we assume that the supply curve for land is unaf-
fected, as are crop yield growth rates (D SL = D DL = 0) . Furthermore, we abstract from
any changes in yields or demand due to price changes (h AS , I = h AD = 0 ), so this equa-
tion collapses to qL* = D DA . In the subsequent chapters, we will introduce the supply-­
side of the story and additional drivers of global agriculture which will allow us to
make more robust assessment about future long-run changes in equilibrium price,
yields, output, consumption and land use.
Tables 2.4 and 2.5 report the changes in global crop demand and crop land use
owing to population and income growth based on historical and future rates, respec-
tively. Note that we are also interested in disaggregating the impacts across each
demand driver and region.1 Looking at the last row of Table 2.4, we can see that if
historical growth rates persist in the future and prices remained unchanged then
global crop demand and crop land use is expected to rise by 18.0 billion metric tons
and 365 million hectares with population significantly exceeding the contribution of
income. Our decomposition shows that at least 83 % of the total expansion in world
crop demand and crop land use would be due to population growth. Note that the
global annual historical growth rates of population and per capita incomes are com-
parable (1.7 % vs. 1.4 % per annum, respectively) which suggest that the regional
differences in population and income growth plays a crucial role in understanding
these global results. Indeed, we see that bulk of the changes at the global level is
driven by the developing world especially by South Asia, Southeast Asia and China/
Mongolia. Almost half of the increase in global crop demand and crop land use can

1
 We aggregated the results of the 15-region version of SIMPLE. Regions classfied as Developed
include Australia/New Zealand, Canada/U.S.A., European Union+ and Japan/Korea. Latin
America consists of Central America & the Caribbean and South America while Rest of World
comprise of Eastern Europe, Middle East, North Africa, Sub-Saharan Africa, Southern Africa and
Central Asia.
2.4  Analyzing the Global Food and Land Needs Due to Population and Income Growth 25

Table 2.4  Historical population and income growth rates: contribution to global crop demand (in
million metric tons) and global crop land use (in million hectares)
Crop demand (in M Mt) Crop land use (in M ha)
Regions Population Income Total Population Income Total
Developed 877 −127 750 17.6 −2.4 15.3
Developing 14,207 3059 17,266 281.1 68.6 349.7
Latin America 1823 295 2118 36.2 5.8 42
Sub-Saharan Africa 2006 −129 1877 40.3 −2.4 37.9
South Asia 2629 844 3472 51.6 16.9 68.5
Southeast Asia 2751 931 3682 53.9 19.2 73.1
China/Mongolia 3373 1111 4485 66.7 28.7 95.4
Rest of World 1625 7 1632 32.4 0.4 32.8
World 15,083 2932 18,016 298.7 66.3 365
Source: Authors’ calculations using the SIMPLE model

Table 2.5  Future population and income growth rates: contribution to global crop demand (in
million metric tons) and global crop land use (in million hectares)
Crop demand (in M Mt) Crop land use (in M ha)
Regions Population Income Total Population Income Total
Developed 397 −32 365 9.1 −0.7 8.5
Developing 4949 4401 9350 112.3 109.2 221.4
Latin America 469 334 803 10.7 8 18.7
Sub-Saharan Africa 2356 924 3279 53.3 21.7 75.1
South Asia 884 1152 2036 20 28.6 48.5
Southeast Asia 699 734 1433 15.9 17.9 33.7
China/Mongolia 132 937 1069 3 24.8 27.8
Rest of World 410 321 730 9.4 8.2 17.5
World 5345 4370 9715 121.4 108.5 229.9
Source: Authors’ calculations using the SIMPLE model

be attributed to population growth in these regions which ranges from 2.2 % to 1.6
% per annum based on historical rates.
Of course, assuming that historical growth rates will persist in the future is quite
naïve; hence, it is important to explore the global crop demand and crop land use
implications of expected future growth rates in population and income. Table 2.5
summarizes the global changes and the contribution of each driver and region. We
immediately observe that these results are quite different from the previous scenario
and this can be directly attributed to the expected slowdown in future population
growth. Globally, population will increase by 0.8 % per annum between 2006 and
2051 which is roughly half of the historical growth rate. Despite the slower growth,
we still observe significant increases in global crop demand (at 9.7 billion metric
tons) and crop land use (at 230 million hectares) due to the sharp rise in future per
capita income growth (around 2.1 % per annum, globally). Indeed, our decomposi-
tion reveals that the contribution of income growth to global changes in crop demand
and crop land use is close to that of population (45 % vs. 55 % respectively). This
26 2  Population and Income as Drivers of Global Change

suggests that, for the first time in human history, income growth will rival—and
eventually surpass—population growth as a driver of food demand. Looking at the
regional decomposition, we see again that population and income growth in the
developing world will have a strong impact on the global changes in crop demand
and crop land use. Indeed, around 68 % of the global changes in crop demand and
crop land use will be driven by regional demands in Sub-Saharan Africa, South Asia
Southeast Asia and China/Mongolia. In South Asia and China/Mongolia, income
growth will be the main driver of regional demands. Population growth in Sub-
Saharan Africa remains a key driver of demand in that region; thus, the addition of
people in this region will exert more influence on global crop output in the future.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Bloom, D. E. (2011). 7 Billion and counting. Science, 333(6042), 562–569. http://doi.org/10.1126/
science.1209290.
Clements, K. W., & Chen, D. (1996). Fundamental similarities in consumer behaviour. Applied
Economics, 28(6), 747–757. http://doi.org/10.1080/000368496328498.
FAO. (2012). FAO food security statistics. Retrieved from http://www.fao.org/economic/ess/en/.
FAOSTAT. (2013). FAO Statistical Database Accessed (Nov 2013). Retrieved from http://faostat.
fao.org/.
Fouré, J., Bénassy-Quéré, A., & Fontagné, L. (2013). Modelling the world economy at the 2050
horizon. Economics of Transition, 21(4), 617–654. http://doi.org/10.1111/ecot.12023.
International Comparison Program. (1996).
Muhammad, A., Seale J. L., Jr., Meade, B., & Regmi, A. (2011). International evidence on food
consumption patterns: An update using 2005 International Comparison Program Data
(Technical Bulletin No. TB-1929) (p. 59). Washington, DC: Economic Research Service,
U.S. Department of Agriculture. Retrieved from http://www.ers.usda.gov/Publications/TB1929/.
Reimer, J., & Hertel, T. (2003). International cross section estimates of demand for use in the
GTAP model. Center for Global Trade Analysis, Department of Agricultural Economics,
Purdue University. Retrieved from http://ideas.repec.org/p/gta/workpp/1190.html.
Roberts, L. (2011). 9 Billion? Science, 333(6042), 540–543. http://doi.org/10.1126/
science.333.6042.540.
Southgate, D., Graham, D., & Tweeten, L. (2010). The world food economy (2nd ed.). Oxford:
Blackwell.
Thompson, W. S. (1929). Population. American Journal of Sociology, 34(6), 959–975.
UNPD. (2000). World at Six Billion. New York, USA: United Nations.
UNPD. (2011). World population prospects: The 2010 revision. New York, NY: Population
Division of the Department of Economic and Social Affairs of the United Nations Secretariat.
Retrieved from http://esa.un.org/unpd/wpp/index.htm.
UNPD. (2013). World population prospects: The 2012 revision. New York, NY: Population
Division of the Department of Economic and Social Affairs of the United Nations Secretariat.
Retrieved from http://esa.un.org/unpd/wpp/index.htm.
World Bank. (2013). World development indicators. Washington, DC: World Bank. Retrieved from
http://data.worldbank.org/data-catalog/world-development-indicators.
Chapter 3
Productivity Growth and Yields
in the Global Crops Sector

3.1 Historical Yield Growth

The most important accomplishment of global agriculture over the past 50 years has
been the remarkable increase in crop productivity—more specifically crop yields.
As a consequence, despite a tripling of global crop output, area expansion has been
quite modest. According to the FAO (Alexandratos & Bruinsma, 2012), 86 % of the
expansion in output over this period was due to either higher yields (77 %) or
increased cropping intensity through multiple cropping (9 %). If this feat can be
repeated in the next 50 years, then prospects are favorable for improving food secu-
rity while limiting cropland incursion into environmentally sensitive areas. However,
some observers have argued that yield growth rates have slowed down in recent
years, thereby raising doubts about the potential to continue the historic increase in
crop yields by mid-century. Furthermore, some of this intensification has come at an
environmental cost—for example farm fertilizer and pesticide runoff resulting in
diminished water quality. This chapter will explore the debate over future produc-
tivity growth, introducing some key concepts along the way.
To begin, it is useful to examine long run yield trends. One of the crop-region
pairs with the longest historical yield data is corn produced in the United States.
Figure 3.1 plots corn yields in the U.S. over the past 150 years. From the figure, it
is quite clear that the remarkable growth in yields is a relatively recent phenomenon.
For most of recorded history, growth in agricultural output came at the extensive
margin—converting natural lands such as forests and grasslands for use in agricul-
ture. With the onset of crop breeding technology, the introduction of hybrids, along
with the intensive use of modern farm inputs, crop yields have grown strongly—
increasing fivefold since the 1930s. But how long can this growth continue?
Looking at the broad trends in productivity growth, Grassini, Eskridge, and
Cassman (2013) examined growth rates in yields for rice, wheat and maize across
many regions of the world from the 1960s to the present. They concluded that, in
general, a linear growth model offers the best fit to the pattern of historical yield

© Springer International Publishing Switzerland 2016 27

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_3
28 3 Productivity Growth and Yields in the Global Crops Sector

Fig. 3.1 Corn yield growth from 1866 to 2002, assembled by Sutch (2014) using updated data
from the USDA National Agricultural Statistics Service (2013)

growth for these three major staple grains (Fig. 3.2). Of course, linear yield growth
translates into a slowing rate of growth in yields, as each annual increment builds on
a larger base yield. Given this, Grassini et al. (2013) argue that yield projections
based on exponential growth—as is common in many economic models—are likely
to be grossly misleading. Furthermore, they note that such projections also ignore
the biophysical limits on yield growth captured by the yield plateaus shown in Fig.
3.2. Indeed, in the cases of rice and wheat, the authors find strong evidence of
emerging yield plateaus in the highest yielding regions, such as California and
Korea for rice and Northern Europe for wheat. What the factors lie behind the slow-
down in productivity? Is this due to policy reforms in these regions which have
resulted in lower prices and slowing investments in agriculture? Globally, research
and development expenditures in agriculture flattened out in the 1980s and 1990s
when world food prices declined. Or is it due to the fact that yields of existing crop
varieties are truly reaching some biophysical limit? What is the potential for increas-
ing crop yield growth in the coming decades?
When cast in terms of the footrace between growing demands and rising supplies
described in the previous chapter, the prospects of slower productivity growth in the
future, if confronted with a steady growth rate in demand (an unlikely assumption,
given the discussion in Chap. 2 about the slowing rate of population growth), implies
more cropland conversion in order to attain supply-demand equilibrium.
Furthermore, if global agriculture cannot keep pace with increased food demand in
the future, then food prices will surely rise in response to the ensuing scarcity. As
we will see, this is a complex topic, which requires an understanding of both the
biophysical and economic determinants of yields.
3.2 Prospects for Further Increases in Yields 29

Fig. 3.2 Trends in grain yield for rice, wheat and maize, by region (letters in parentheses describe the
mathematical function which best fits these data, e.g., L linear model). Source: Grassini et al. (2013)

3.2 Prospects for Further Increases in Yields

On a purely scientific level, crops can be viewed as ‘machines’ which convert sun-
light, CO2, nutrients and water into carbohydrate molecules via photosynthesis.
Efficient nutrient and water intake can be attained through proper farm manage-
ment; thus, from the point of view of boosting the fundamental yield potential of a
crop, the key question is whether net carbon uptake can be increased, and how long
such improvements are likely to take. Zhu, Long, and Ort (2010) discuss the
30 3 Productivity Growth and Yields in the Global Crops Sector

Fig. 3.3 Different types of yields and identifying the yield gap. Source: Van Ittersum et al. (2013)

prospects for improving net carbon uptake in plants. The channels for such improve-
ments range from near-term prospects (e.g. improving canopy structure, optimizing
canopy chlorophyll content) to innovations that will likely only occur in the long
run—if at all (e.g. conversion of C3 plants to C4, increasing mesophyll conduc-
tance). Many of these individual measures are expected to increase efficiency by up
to 30 %, suggesting that there is still substantial scope for improving yields by push-
ing the biophysical limits (Fischer, Byerlee, & Edmeades, 2014).
However, the greatest potential for increasing average yields globally likely rests
in closing the so-called ‘yield gap’. Figure 3.3, taken from Van Ittersum et al.
(2013), illustrates the key concepts underpinning any discussion of yield gaps. We
start with maximum potential yield (Yp), reflecting what is possible with current
cultivars, and assuming that yields are not limited by water (i.e., irrigation is pres-
ent) or nutrients. Potential yield in an irrigated environment is dictated by radiation,
temperature and CO2 concentrations as well as planting dates. In a rainfed system,
yields will be water-limited (Yw), and therefore dependent upon rainfall and soil
moisture so that: Yp > Yw. Due a variety of factors, including diminishing economic
returns to the application of fertilizer, labor and other inputs, as well as uncertainty
facing producers, the exploitable yield is less than potential and water limited yields.
Lobell, Cassman, and Field (2009) place this at about 80 % of Yp and Yw. Finally,
there is also the gap between the exploitable yield and what is achieved by the aver-
age farmer (Ya), and therefore the observed crop yield in agricultural census data.
Estimates of potential yields typically come from crop simulation models which
gauge the theoretical potential at a given point on the earth’s surface. Potential
yields can also be estimated from crop competitions (Lobell et al., 2009), frontier
analysis (Neumann, Verburg, Stehfest, & Müller, 2010), or by simply taking the
3.2 Prospects for Further Increases in Yields 31

maximum yields observed in farm surveys. The difference between exploitable

yield and average yield is termed the ‘yield gap’ and this is a central focus for those
arguing for boosting yields by ‘closing this gap’. Lobell et al. (2009) suggest that
yield gaps are already quite close to potential yields for much of the world’s irri-
gated agriculture, which in turn, accounts for 40 % of global production. So the
emphasis in closing these yield gaps will likely be on rainfed cropping systems
where yields are often less than 50 % of attainable yields. Ejeta (2013) suggests that
these gaps are smallest for wheat (40–95 %) and largest for maize (16–46 %) and
sorghum (8–33 %). Estimating yield gaps using a global cross section of gridded
data and controlling for agro-ecological conditions, Licker et al. (2010) find evi-
dence of very large yield gaps for maize, wheat, soybeans and rice, globally. By
their calculations, closing these gaps would result in an increase of 50 % in world
maize supply.
Of course, there are good reasons why such gaps exist and why closing these gaps
poses a significant challenge. In the Guinea Savannah Zone of Sub-Saharan Africa,
where the authors report the largest yield gaps, maize prices at the farm gate are very
low and the cost of purchased farm inputs such as fertilizers and pesticides are very
high. As a result, it is hardly surprising that average yields are very low. This point is
reinforced by Robert Herdt (1979) in his extensive study of yield gaps for rice in
Asia. He finds that: “what is technically possible is more modest than most observers
admit; the economics of substantially higher yields is not attractive.”
Based on the preceding discussion, it appears that there is great potential to
increase yields globally. And the constraints to such improvements are often driven
by economic factors. Of course, improving the economic circumstances of rural
producers in Africa is a question of broader economic development. Nonetheless,
there is significant optimism about this potential on some fronts. The World Bank
recently published a study entitled: “Africa: the Sleeping Giant” hypothesizing that,
with appropriate investments and policy reforms, the Guinea Savannah Zone of this
continent could become the next Brazilian Cerrado—supplying much of the
increased global agricultural output likely to be demanded over the coming decades
(World Bank, 2009). Jelle Bruinsma, a long-time observer of the global agricultural
scene also reflects this optimism in his expert projections for future yield growth,
suggesting that productivity improvements in Africa could contribute significantly
to global production increases between 2010 and 2050 (Bruinsma, 2009).
Ultimately, increases in yields at the farm level will depend on a host of socio-
economic factors including: institutional and economic reforms, public investments
in infrastructure, education, and agricultural public and private research and devel-
opment (R&D). Indeed, long term productivity growth in agriculture has been
shown to be closely linked to R&D expenditures (Fuglie, 2012). However, there are
long lags—sometimes two decades or more—before these investments can be trans-
mitted to increases in crop yields (Alston, Pardey, & Ruttan, 2008). So R&D invest-
ments made over the next 20 years will likely determine future productivity growth
rates by mid-century. The good news is that R&D expenditures have risen strongly
since 2000, partly in response to the recent spikes in global food prices. Between
2000 and 2008, global public spending on agricultural R&D rose by 22 % to nearly
32 3 Productivity Growth and Yields in the Global Crops Sector

$32 billion and global private spending rose at an even faster rate (by 26 %).
Domestic agricultural R&D spending is also rising strongly in developing regions
such as India, Brazil and China (Beintema, Stads, Fuglie, & Heisey, 2012). Whether
these rates will be sustained in the coming decades is an open question—particu-
larly if commodity prices resume their long-run historical decline and interest in
investing in the agricultural sector wanes. Whether it is acknowledged or not, R&D
appears to be strongly related to food prices (Fulginiti & Perrin, 1993)—particu-
larly in the context of today’s crisis-driven political economy.
In their comprehensive study of global crop yields, Fischer et al. (2014) offer the
following ‘pearls of wisdom’ on this topic:
In conclusion, no calamity is foreseen—but there is no room for complacency, especially of
the kind invoked by some advocates of biotechnology. Multidisciplinary agricultural sci-
ence remains the key to success. With complementary investments in infrastructure and
institutions, and relative freedom from civil unrest, the world should manage to sufficiently
feed its growing population. (p. 559)

3.3 Measuring Productivity: Partial vs. Total Factor

Productivity

Thus far we have focused solely on crop yields—a measure of partial factor produc-
tivity which reflects crop output per area of land input used (e.g., bushels per acre,
kilograms per hectare), while remaining silent on the use of other inputs. Basic
accounting dictates that crop yields are critical for global land use, since the required
cropland area is simply equal to total production, divided by yield. (Note that this is
a broad definition of crop yields since it includes cropping intensity effects due to
the cultivation of multiple crops on the same land in a given year.) Now we turn to
the question of how cost minimizing farmers might adjust yields endogenously in
response to changing market conditions.
Figure 3.4 illustrates how optimal input use—and implicitly crop yield—is deter-
mined within an economic framework wherein farmers are minimizing overall input
costs subject to a required level of crop output. Curved lines in the figure represent
isoquants faced by farmers—a mapping of all possible land (XL) and non-land (XNL)
input combinations which yields a given level of crop output (Q). The isoquant is
downward sloping which suggests that substitution is possible between land and
non-land inputs. In other words, a farmer can produce a pre-specified level of crops
by either using more land or applying more non-land inputs such as fertilizers,
human and/or animal labor and farm machinery. However, the capacity for input
substitution is imperfect due to diminishing marginal productivity for both inputs.
Based on Fig. 3.4, we see that farmers who are intensively producing crops will
have to apply more non-land inputs, per unit output, since the marginal productivity
of non-land inputs is diminishing (e.g. there are limits to increased crop production
from the application of additional fertilizer). The straight lines in Fig. 3.4 represent
3.3 Measuring Productivity: Partial vs. Total Factor Productivity 33

Fig. 3.4 Optimal land and non-land input use by farmers given different input prices. The left
panel shows input use under intensive production (i.e. less land use and high crop yields) while the
right hand panel shows the case of extensive production (i.e. more land use and low crop yields)

isocost lines which show the combination of land and non-land inputs that a farmer
can chose, given a fixed budget for input expenditures and prevailing input prices.
Note that the slope of this curve is determined by the ratio of input prices (PNL/PL).
When input prices are accounted for, the cost-minimizing combination of land and
non-land inputs that a farmer will chose will be at the point whereby the ratio of
marginal products for both inputs is equal to the ratio of input prices. Graphically,
this is represented by the tangency between the isoquant and isocost lines. Within
this framework, we can clearly see that input use, and consequently crop yield, will
depend on the prevailing input prices. If farmers face a high price of land relative to
non-land inputs (i.e. a flatter isocost line in the figure), more non-land inputs will be
used relative to land inputs, resulting in more intensive crop production (left panel
of Fig. 3.4). Under these conditions, crop yields will be higher since less land is
used. On the contrary, if non-land inputs are much more expensive than land inputs,
farmers will use more land inputs (i.e. more extensive crop production) and as a
result have lower crop yields (right panel of Fig. 3.4).
While land use is important for environmental purposes, when discussing food
prices and food security in this book, we will see that it is not yields per se, but rather
total factor productivity (TFP) that will matter most. As the term suggests, TFP mea-
sures the efficiency with which crop outputs are produced, factoring in all inputs—
not just land. So when the index of outputs rises faster than the index of all inputs
used in production, we say that there is TFP growth. To see why TFP is a key deter-
minant of crop prices, consider two examples of yield increases generating reduced
land use (Fig. 3.5). In the right panel of Fig. 3.5, a new crop variety is introduced in
the market which yields 10 % more output per hectare compared to existing variet-
ies, while keeping all inputs—including fertilizer, labor and capital, constant. This is
represented by the downward shift in the isoquant curve which suggests that techno-
logical improvements increase the productivity of inputs, thereby reducing the input
requirements needed to produce the same level of output.
34 3 Productivity Growth and Yields in the Global Crops Sector

Fig. 3.5 Rise in crop yields due yield increases owing to use due to changes in relative inputs
prices (left panel) and technological improvements (right panel)

Assuming costless technological adoption and unchanging input prices (i.e.

same slope of isocost line), the cost of production will be reduced in the right hand
panel, since input use is reduced (although outside this simplistic framework, the
gains are likely to be reflected in higher input prices—particularly for inputs, such
as improved seeds, associated with the new technology). This cost reduction is cap-
tured in the downward shift of the isocost line. If food demand remains unchanged,
lower cost production will be transmitted to consumers in the form of lower crop
prices. Cheaper food translates directly into increased food consumption and
improved food security, particularly for low income households. Of course, under
this scenario, crop yields will also be increasing due to the reduction in over-all
input use for a given level of crop output.
The left hand panel of Fig. 3.5 portrays a different set of circumstances in which
yields also rise, but for a different reason. Here, there is an exogenous rise in the
price of farmland—perhaps due to competition with suburban demands. With the
price of farmland rising, producers are likely to apply more fertilizer, labor and
other inputs in order to increase yields enough to cover the higher cost of land. This
is reflected by the steeper slope of the new (solid) isocost line. In this case, due to
the subsequent reduction in land inputs, we also observe rising yields. However,
unlike the previous case, this yield improvement is due to endogenous ‘intensifica-
tion’—moving away from land towards non-land inputs. Now, rather than down-
ward pressure on prices, the graph is presented in such a way that food prices remain
unchanged since total output is the same in the new equilibrium. In short, improved
yields by no means ensure lower crop prices and improved food security since
yields can be driven by changes in input use due to changing relative input prices.
On the other hand, TFP growth invariably leads to lower food prices, provided the
purveyor of improved technology is not able to extract all of the rents from farmers
in the form of higher input prices.
Despite its importance for food security, total factor productivity is rarely dis-
cussed in the literature on future changes in global agriculture. One of the main
3.3 Measuring Productivity: Partial vs. Total Factor Productivity 35

reasons for this lack of attention is the difficulty of measuring TFP growth. Accurate
assessments require measuring and accounting for ALL inputs used in agricultural
production, and this is a challenging (some would say impossible!) task. Provided
the omitted input varies randomly over time, or if the input is independent of the
total cost of agricultural production, this may not be a problem. For example, the
impact of good weather might temporarily boost production in some years or bad
weather might result in lower than average production in other years. This will add
additional noise when estimating TFP growth but should not affect the decadal
growth rate. On the other hand, better farm management practices which can be
achieved through improvements in farmer’s education and farm extension can lead
to steady increases in agricultural output. And if the unreported input is growing
over time and is a principle component of production costs, then its omission can
lead to an overstatement of TFP growth (Alston & Pardey, 2014).
More generally, total factor productivity can move in the opposite direction from
yields when there are incentives for changing the rate at which other inputs are
applied. Ludena, Hertel, Preckel, Foster, and Nin (2007) highlight the case of China,
where crop yields rose strongly starting in the early 1970s. However, total factor
productivity growth was flat until the implementation of agricultural reforms in
China in the 1980s. This implies that the growth in yields observed during the early
years was largely a consequence of applying more inputs to a fixed amount of land.
This phenomenon is not restricted to developing countries. In the wake of the reform
of the EU Common Agricultural Policy in the 1990s, both producer prices and sub-
sequently yields fell, yet TFP continued to grow relatively strongly. Understanding
agricultural productivity requires one to look beyond trends in yields.
Fuglie (2012) provides a comprehensive historical assessment of TFP growth for
global agriculture. Figure 3.6 summarizes his recent analysis of growth in global
agricultural output over the period 1961–2009 along with a decade-by-decade
breakdown. Importantly, in addition to reporting total decadal output growth, he
offers a decomposition of this growth into the following elements: (a) area expan-
sion, (b) growth in irrigated area, (c) growth in inputs per hectare of land, and (d)
total factor productivity growth. This decomposition offers several critical insights.
First of all, aggregate agricultural output growth was remarkably steady over this
period, varying from a high of about 2.7 %/year in the 1960s to a low of about 2.1
%/year in the 1980s. Annual output growth over the most recent decade has been
higher than in the preceding three decades and is only surpassed by the average
annual growth rate in the 1960s.
While the overall average growth rate in agricultural output was rather steady
over these five decades, the sources of growth varied greatly. In the wake of the
Green Revolution of the 1960s and 1970s, input intensification was the primary
source of output growth, worldwide. The new varieties of seeds boosted yields dra-
matically, but they also required far more inputs per hectare of land in order to be
effective (i.e. fertilizers, irrigation etc.). Therefore, the ensuing growth in yields was
not coincident with large increases in total factor productivity. This contrasts
sharply with productivity growth in the 1990s, when low commodity prices discour-
aged the application of additional inputs, yet output growth was maintained via the
36 3 Productivity Growth and Yields in the Global Crops Sector

%/year
3.0

2.5

2.0

1.5

1.0

0.5

0.0
1961-2009 1961-1970 1971-1980 1981-1990 1991-2000 2001-2009

TFP Irrigation
Inputs/Land Area expansion

Fig. 3.6 TFP growth in global agriculture: 1961–2009. Source: Fuglie (2012)

application of improved production techniques. Average annual agricultural growth

over this decade was nearly as high as in the 1970s, but it was largely TFP driven
and can be loosely attributed to institutional and market reform and increased inter-
national trade which allowed more efficient allocation of global production. And the
same is true of the most recent decade, although with the higher prices emerging in
the final few years of Fuglie’s sample, we do observe an uptick in the intensification
of production contributing to higher output growth. It is also notable that area
expansion has been less significant in the past two decades—coincident with the
higher TFP growth rates.

3.4 Differential Sources of Productivity Growth in SIMPLE

To conclude this chapter, we look at the implications of different measures of pro-

ductivity growth in global agriculture given a series of scenarios generated from the
SIMPLE model. In these simulations, we revisit the historical experience starting
3.4 Differential Sources of Productivity Growth in SIMPLE 37

from 1961 to 2006 and examine the changes in key agricultural variables namely
crop production, land use, yields and crop prices. By looking back at history, we can
also evaluate how well our model replicates observed data given our assumptions on
key economic margins and growth rates of exogenous drivers. Building on our his-
torical simulations in Chap. 2, we account for changes in global demand due to
growth in population and income. We use population data from the 2012 World
Population Prospects (UNPD, 2013) and real GDP data from the World Development
Indicators (2013). Total factor productivity growth for livestock and processed food
sectors are taken from Ludena et al. (2007) and Griffith, Redding, and Reenen
(2004), respectively.
We adopt differential measures and sources of crop productivity growth in each
scenario to illustrate the impact of these assumptions on the historical changes in
global agriculture. As illustrated above in Fig. 3.5, crop yield growth can driven by
either input substitution—the movement away from land towards non-land inputs due
to changes in relative input prices—or due to technological change—the overall reduc-
tion in input use due to technological improvements in crop production. Accordingly,
we will start with a scenario in which all of the yield change is due to changing relative
input prices. We will then consider a scenario in which the yield change is targeted
exogenously as ‘technological change’ and ignore the influence of relative input prices.
Finally, we will consider a scenario in which we shock TFP and allow yields to adjust
endogenously as a function of both technology and changing relative prices.
Historical crop TFP growth rates are taken from Fuglie (2012). We rely on
FAOSTAT (2014) in calculating historical changes in global crop production, global
crop land and global crop yields while changes in global crop prices are taken from
the Global Economic Monitors datasets by the World Bank (2013a, 2013b). These
observed changes at global scale are summarized in Fig. 3.7 using four different pan-
els. The dotted lines signify the observed changes in each global variable, while the
individual bars report model outcomes under the three different ‘scenarios’. As can be
seen in the upper left hand panel, global crop production rose by more than 200 %
over the 1961–2006 period. Yields rose by almost as much (dashed line in lower left
panel of Fig. 3.7). Therefore, only a modest increase global crop land (11 %) was
required, and global crop prices actually declined (29 %—lower right hand panel).
In the first, stylize scenario in Fig. 3.7, crop yield growth is postulated to come
entirely from input substitution. (We alter the relative price of land and nonland
inputs exogenously to target yields according to the observed change over this
period.) In this case, we grossly under-predict the increase in global crop production
and global crop land. Instead of declining, global crop price rises strongly, thereby
highlighting the importance of technical change in achieving the observed food
price path. On the other hand, if we incorporate crop yield growth via pure techno-
logical change and ignore input substitution (second set of bars in Fig. 3.7), we tend
to overstate global crop production growth and global crop land expansion as well
as global crop price reduction. Only when both input substitution and TFP growth
are incorporated do we see improvements in simultaneously replicating historical
changes in global crop production, crop land and crop prices. (Of course we no
longer hit yields exactly, since they are now endogenous.)
38 3 Productivity Growth and Yields in the Global Crops Sector

Fig. 3.7 Changes in key global variables from the historical simulations using the SIMPLE model

In summary, productivity growth is critical to understanding the past perfor-

mance of the global food system, as well as projecting the future. Within our eco-
nomic framework, yield growth will no longer be synonymous with productivity
growth. Rather, it will be a function of rates of total factor productivity growth, as
well as changing relative prices. The latter has the potential to induce endogenous
intensification, wherein additional nonland inputs are applied to the land, thereby
boosting output per hectare of cropland.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050: The 2012 revision
(ESA Working Paper No. 12-03). Rome, Italy: Food and Agriculture Organisation of the
United Nations.
Alston, J. M., & Pardey, P. G. (2014). Agriculture in the global economy. Journal of Economic
Perspectives, 28(1), 121–146. http://doi.org/10.1257/jep.28.1.121.
Alston, J. M., Pardey, P. G., & Ruttan, V. W. (2008). Research lags revisited: Concepts and evi-
dence from U.S. Agriculture (Staff Paper No. 50091). University of Minnesota, Department of
Applied Economics. Retrieved from http://ideas.repec.org/p/ags/umaesp/50091.html.
Beintema, N., Stads, G.-J., Fuglie, K., & Heisey, P. (2012). ASTI global assessment of agricultural
R&D spending: Developing countries accelerate investment (International Food Policy
References 39

Report). Washington, DC: International Food Policy Research Institute, Agricultural Science
and Technology Indicators, Global Forum on Agricultural Research.
Bruinsma, J. (2009). The resource outlook to 2050. By how much do land, water use and crop
yields need to increase by 2050? In FAO expert meeting on how to feed the world in 2050.
Rome, Italy: Food and Agriculture Organisation of the UN.
Ejeta, G. (2013). Global Agricultural Productivity. Guest Lecture in AGEC 596. Spring 2013.
Purdue University.
FAO. (2014, September 25). FAOSTAT. Retrieved September 25, 2014, from http://faostat.fao.org/.
Fischer, R. A., Byerlee, D., & Edmeades, G. (2014). Crop yield and food security: Will yield
increases continue to feed the world? Canberra: ACIAR.
Fuglie, K. O. (2012). Productivity growth and technology capital in the global agricultural econ-
omy. In K. O. Fuglie, S. L. Wang, & V. E. Ball (Eds.), Productivity growth in agriculture: An
international perspective (pp. 335–368). Cambridge, MA: CAB.
Fulginiti, L. E., & Perrin, R. K. (1993). Prices and productivity in agriculture. The Review of
Economics and Statistics, 75(3), 471–482. http://doi.org/10.2307/2109461.
Grassini, P., Eskridge, K. M., & Cassman, K. G. (2013). Distinguishing between yield advances
and yield plateaus in historical crop production trends. Nature Communications, 4, 2918. http://
doi.org/10.1038/ncomms3918.
Griffith, R., Redding, S., & Reenen, J. V. (2004). Mapping the two faces of R&D: Productivity
growth in a panel of OECD industries. Review of Economics and Statistics, 86(4), 883–895.
http://doi.org/10.1162/0034653043125194.
Herdt, R. W. (1979). An overview of the constraints project results. In Farm-level constraints to
high rice yields in Asia: 1974–1977 (pp. 395–421). Los Baños, Laguna, Philippines:
International Rice Research Institute.
Licker, R., Johnston, M., Barford, C., Foley, J. A., Kucharik, C. J., Monfreda, C., & Ramankutty,
N. (2010). Mind the Gap: How do climate and agricultural management explain the “yield gap”
of croplands around the world? Global Ecology and Biogeography, 19(6), 769–782.
Lobell, D. B., Cassman, K. G., & Field, C. B. (2009). Crop yield gaps: Their importance, magni-
tudes, and causes. Annual Review of Environment and Resources, 34(1), 179–204. http://doi.
org/10.1146/annurev.environ.041008.093740.
Ludena, C. E., Hertel, T. W., Preckel, P. V., Foster, K., & Nin, A. (2007). Productivity growth and
convergence in crop, ruminant, and nonruminant production: Measurement and forecasts.
Agricultural Economics, 37(1), 1–17. http://doi.org/10.1111/j.1574-0862.2007.00218.x.
Neumann, K., Verburg, P. H., Stehfest, E., & Müller, C. (2010). The yield gap of global grain pro-
duction: A spatial analysis. Agricultural Systems, 103(5), 316–326.
Sutch, R. (2014). America’s first biotech food: The strangely-political story of hybrid corn. In
Food+tech: From flint knives to cloned meat, New York, NY.
UN Population Division. (2013). World population prospects: The 2012 revision. New York, NY:
Population Division of the Department of Economic and Social Affairs of the United Nations
Secretariat. Retrieved from http://esa.un.org/unpd/wpp/index.htm.
USDA National Agricultural Statistics Service. (2013). Crop production historical track records
(No. 2157-8990, p. 240). Washington, DC: USDA National Agricultural Statistics Service.
Van Ittersum, M. K., Cassman, K. G., Grassini, P., Wolf, J., Tittonell, P., & Hochman, Z. (2013).
Yield gap analysis with local to global relevance—A review. Field Crops Research, 143, 4–17.
http://doi.org/10.1016/j.fcr.2012.09.009.
World Bank. (2009). Awakening Africa’s sleeping giant: Prospects for commercial agriculture in
the Guinea Savannah Zone and beyond (pp. 1–236).
World Bank. (2013a). Global economic monitor: Commodities. Washington, DC. Retrieved from
http://data.worldbank.org/data-catalog/commodity-price-data
World Bank. (2013b). World development indicators. Washington, DC. Retrieved from http://data.
worldbank.org/data-catalog/world-development-indicators.
Zhu, X.-G., Long, S. P., & Ort, D. R. (2010). Improving photosynthetic efficiency for greater yield.
Annual Review of Plant Biology, 61(1), 235–261. http://doi.org/10.1146/annurev-
arplant-042809-112206.
Chapter 4
Economic Responses to Scarcity*

4.1  Overview

The discussion in Chaps. 2 and 3 focused on exogenous drivers of the global food
system, including demographics, income per capita and agricultural productivity. A
more complete model of global sustainability would seek to explain how each of
these drivers evolves over time as a function of deeper factors, such as investments
in human and physical capital, and political and economic systems. However, in this
book we only focus on the endogenous responses of producers and consumers to
changes in the relative scarcity of food, land and related resources. This chapter
develops the theoretical framework on how these economic responses arise and how
these affect the long run change in cropland use, prices and food consumption. At
the same time, it also offers important insights into future patterns of cropland
expansion and intensification.
A good starting point for understanding the response of the global food system
to scarcity is to recall the analytical framework developed in Appendix B and dis-
cussed in Chap. 1, Eq. (1.1) which gives an expression for the percentage growth in
cropland use (qL*):

qL* = (D DA - D DL ) / (1 + h AS , I / h AS , E + h AD / h AS , E )

As before, the sign of the change in global land use depends on the footrace between
the percentage growth in demand (ΔAD) and supply—as measured by trend improve-
ments in cropland yield—(ΔLD), wherein (D DA - D DL ) > 0 Þ qL* > 0. That is, if
demand grows faster than trend yields, then cropland area will rise. However, the
magnitude of this change in global cropland use depends critically on the respon-
siveness of producers and consumers to scarcity in the food system, as captured by
the elasticity of crop supply at the intensive margin (ηAS,I), the extensive margin

* Parts of this chapter draw heavily on Hertel and Tyner (2013).

© Springer International Publishing Switzerland 2016 41

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_4
42 4  Economic Responses to Scarcity

(ηAS,E), and the price elasticity of demand for crops (-h AD ). If producers respond
aggressively to scarcity-induced higher prices, by intensifying their use of fertilizer
and other non-land inputs, i.e. h AS , I >> 0, then the amount of land conversion
required by rapid demand growth will be diminished. Similarly, if consumers
respond to higher prices by cutting back on food consumption, then -h AD < 0 and
this feature also moderates the land use expansion required in the face of strong
growth in population and/or income.
The relative impact of these two economic ‘shock-absorbers’ depends on the size
of the intensive-supply and demand elasticities relative to the extensive-supply
elasticity: h AS , E > 0 . If productive land is more available (i.e. high ηAS,E) then there is
greater potential for farmers to convert more natural lands into agricultural lands in
light of high food demand.
In this chapter, we explore the literature on all three of these economic responses.
We then conclude this chapter by looking at the regional responses to exogenous
shocks, in particular improved technology. Specifically, we revisit the historical
Green Revolution and shed additional light on the land use and GHG implications
of regional productivity growth. As it turns out, the impacts of regional technical
change depends importantly on the price elasticities of supply and demand across
regions of the world.

4.2  C
 onsumer Response to Rising Prices: The Price
Elasticity of Demand for Food

If food demand is relatively price-responsive, so that rising prices significantly cur-

tail food consumption, then this will moderate the need for food supply expansion
in the face of scarcity. This is clearly illustrated in Fig. 4.1 wherein we explore the
changes in quantity demanded by consumers with varying response to food prices.

Fig. 4.1  Changes in food demand given similar price increases for less price responsive (left
panel) and more price responsive (right panel) consumers, respectively
4.2 Consumer Response to Rising Prices: The Price Elasticity of Demand for Food 43

In the left panel, we see a relatively steep demand curve which captures the behavior
of consumers if they have lower responsiveness to market prices. Under these con-
ditions, there is little decline in crop demand (Q to Q*) by consumers given rising
food prices (P to P*). In the extreme, if consumers are wholly unresponsive to
market prices this is represented by a vertical demand curve. On the other hand, the
right panel of Fig. 4.1 illustrates the case of very price responsive consumer demand
in which case the demand curve is much flatter so that we see a much larger contrac-
tion in food demand given the same price rise as before.
However, diminished consumption of food is often not a socially desirable out-
come—particularly in parts of the world where malnourishment is most pervasive.
In 2006, more than 20 % of the population in Sub-Saharan Africa and South Asia
was malnourished (totaling 460 million people), with an average nutrition gap
reaching 252 kcal/day in South Asia (Baldos & Hertel, 2014). Regrettably, it is
precisely these low income households that are most responsive to price increases
(Muhammad, Seale Jr., Meade, & Regmi, 2011; Seale, Regmi, & Bernstein, 2003).
Upon reflection, the price-responsiveness of the poor is hardly surprising since a
household spending 70 % of its income on food will experience a 35 % decline in
real income when faced with a 50 % food price rise unless it can reduce food con-
sumption. Indeed, for low income households, with little savings and few assets, it
is impossible to accommodate such a price rise without reducing food consumption.
In contrast, for a high income household (e.g., in the U.S.) which might spend just
10 % of its income on food, a 50 % food price rise can be more readily accommo-
dated by adjusting the consumption of non-food items, as this shock represents a
more modest 5 % decrease in real income. An additional factor muting the impact
on these high income households is the high level of post-farm value-added in the
consumers’ purchases. Consumer prices in high income regions embody additional
costs from post-farm processing, transportation, marketing and advertising in addi-
tion to farm gate commodity prices. If these costs account for three-quarters of the
consumer prices and remain unchanged, then a 50 % increase in farm-level com-
modity prices would be expected to translate into just a 1/4 × 50 % = 12.5 % rise in
consumers’ food costs in the wealthy economies.
Muhammad et al. (2011) provide an international cross section analysis of con-
sumer demand elasticities for food, treating average per capita consumption in each
country as an ‘observation’ and assessing the impact of differing food prices across
countries on long run demand. They find wide variation in the price responsiveness
of food demand across countries. For example, in the case of the Democratic
Republic of Congo—one of the poorest countries in their sample, they estimate the
own-price elasticity of demand for breads and cereals to be −0.50. This suggests
that a 1 % increase in the market price of breads and cereals roughly translates to a
half a percent reduction in household purchases for these commodities, on average.
In contrast, the same elasticity in the U.S. is nearly zero. Overall, when income is
held constant, they find that the own-price elasticity of demand for food varies from
nearly −0.90 in the poorest countries to −0.30 in the United States.
In practice, the farm level price elasticity of demand—namely the price elasticity
facing producers [and hence the operational elasticity for use in Eq. (1.1) in
44 4  Economic Responses to Scarcity

Chap. 1]—depends not only on the behavior of consumer demand for food. It also
depends on the demand for agricultural intermediate inputs used by livestock pro-
duction and food processing, and on the price-responsiveness of biofuels demand.
In the case of livestock feed demand, there is significant potential for substitution
amongst feedstuffs (Beckman, Keeney, & Tyner, 2011). However, there is generally
less substitutability among raw agricultural products in the case of food processing
for human consumption. This has the effect of dampening the price elasticity of
demand at the farm gate, as any crop price rise becomes diluted by the addition of
other inputs in the supply chain. Finally, the growing biofuels component of food
demand is very different from the other sources of derived demand for crops in that
that it is often pre-determined by government mandate and therefore unresponsive
to scarcity. For example, in 2012 in the U.S., the corn ethanol mandate was 13.2
billion gallons, regardless of corn price, so the drought during that year that sent
corn prices over $315/Mt did not alter corn use for ethanol. This meant that other
corn users had to adjust proportionately much more in order to equilibrate demand
in response to diminished supplies.

4.3  S
 upply Response at the Intensive Margin: Market Price
Elasticity of Crop Yields

The intensive margin of supply response measures the percentage change in crop
output, holding cropland area constant, when there is a permanent price increase of
1 %. Therefore, this can also be referred to as the price elasticity of crop yields,
since yield measures the output per hectare of land. The response of yields to
increasing crop prices is dependent on the ease of substitution between land and
non-land inputs and this is illustrated in Fig. 4.2. Faced with higher output prices,

Fig. 4.2  Changes in the firm’s output and optimal input use in response to rising consumer and
input prices given imperfect input substitution (left panel) and no input substitution (right panel),
respectively
4.3 Supply Response at the Intensive Margin: Market Price Elasticity of Crop Yields 45

farms will increase their production (Q to Q*) and input use in both cases (XL and
XNL to XL* and XNL*). Changes in input demands will also alter relative input prices
(PL/PNL to PN*/PNL*). However, crop yields will not rise for farms with fixed propor-
tion production technology (i.e. input substitution is absent, right hand panel of Fig.
4.2). These firms will have to increase both inputs proportionally with the rise in
output; thus crop yields will remain unchanged. On the other hand, farms which can
alter their production technique in response to higher land returns will employ more
non-land inputs (e.g., fertilizer and labor), relative to land inputs, resulting in crop
yield growth.
Naturally, the elasticity of substitution between land and non-land inputs is
something that agricultural economists have long sought to measure. The larger this
response is, relative to the extensive margin, the smaller will be the required increase
in area to meet a given rise in demand. Keeney and Hertel (2008) review the econo-
metric yield elasticity estimates for corn in the U.S. in the post-World-War II period
and find that the estimates of the elasticity of corn yields with respect to corn price
have been declining as the industry has consolidated and become more homoge-
neous. After reviewing these studies, they settle on a yield elasticity of 0.25 for use
in their CGE model. This means that, if the price of corn rises, relative to the price
of variable inputs, by 10 %, yields will rise by 2.5 %. Berry and Schlenker (2011)
examine year to year variation in U.S. corn yields and argue that the short run elas-
ticity with respect to price is nearly zero, proposing 0.10 as their preferred value for
the yield elasticity. Huang and Khanna (2010) conduct a more extensive economet-
ric investigation across several U.S. crops and obtain an intermediate value for corn
(0.15), with a higher yield response to price for wheat (0.43) and lower for soybeans
(0.06). Goodwin, Marra, Piggott, and Mueller (2012) allow for intra-seasonal price
response and obtain an estimate of 0.25. All in all, while there remains some uncer-
tainty about the appropriate value for this parameter in the U.S., there is broad
agreement that the value is small (0.25 or less), and likely getting smaller over time.
Unfortunately, the evidence for other countries is generally quite is sparse. And
yet, as Golub and Hertel (2012) document, the uncertainty in yield response to price
in other major producing regions around the world is far more important for uncer-
tainty in global land use change in response (for example) to U.S. ethanol production
than is the uncertainty in the U.S. corn yield elasticity itself. This is a simple matter
of arithmetic: since the rest of the world encompasses a much larger area, the yield
elasticity in this the non-U.S. regions is more important in the global impact of a
U.S.-based shock to demand. There is evidence that the potential for supply response
at the intensive margin is much greater in developing countries than in the U.S. and
Europe. Fertilizer subsidies in Malawi have been shown to have a profound impact
on yields. (Recall that the yield elasticity measures the output response to a change
in crop price, relative to non-land input prices. So yields will rise if crop price rises,
or if fertilizer prices fall.) This result is hardly surprising, in light of the findings of
Potter, Ramankutty, Bennett, and Donner (2010) who report nitrogen fertilizer appli-
cation rates of less than 2.5 kg/ha on more than 50 % of global cropland. Clearly
46 4  Economic Responses to Scarcity

raising application rates from this low level on the other 90+ % of global croplands
in response to higher world prices is likely to have a significant impact on output.
More generally, high commodity prices may be expected to provide longer run
incentives to invest in boosting agricultural productivity—both on farms and in
research stations. Indeed, Fulginiti and Perrin (1993) find evidence for a large and
statistically significant influence of output prices on agricultural productivity itself
in a set of developing countries. Specifically, they find that a 10 % rise in past agri-
cultural prices boosts current agricultural productivity by 1.3 %. This comes in
addition to the tendency to intensify production via the use of nonland inputs.
While the intensification of production around the world moderates the need for
cropland expansion at the extensive margin, increased fertilizer use, in particular,
carries with it significant potential for environmental impacts (Foley et al., 2005;
Ramankutty, 2010; Tilman, Cassman, Matson, Naylor, & Polasky, 2002). Nitrogen
fertilizer applications are a major source of GHG emissions, accounting for a sig-
nificant share of total GHG emissions from agriculture as a whole (Baumert,
Herzog, & Pershing, 2005). In addition, agricultural runoff, coupled with excess
nitrogen applications has resulted in the eutrophication of waterways (Vitousek
et al., 2009). This kind of tradeoff is at the heart of the ongoing ‘land-sparing vs.
land-caring’ debate in the ecology and environmental biology literature (Balmford,
Green, & Scharlemann, 2005; Phalan, Onial, Balmford, & Green, 2011). Is it better
for the environment to cultivate a smaller amount of land intensively, or to farm a
greater area, but to do so more extensively (i.e. using less fertilizer and other non-­
land inputs, and perhaps interspersing trees with the crops)? This is a theme which
will resurface throughout this book.

4.4  S
 upply Response at the Extensive Margin: Price
Elasticity of Cropland Supply

The responsiveness of cropland to higher commodity prices has been a topic of

great interest over the past decade. As prices rose during the price spikes of 2007/08
and 2010/11, producers around the world responded by planting additional cropland
area. And this, in turn, led to increasing rates of deforestation in some parts of the
world, leading to concerns about the environmental impacts of this commodity
boom. We can easily see the implications of increased demand for crops in the land
markets in Fig. 4.3. Greater crop demand due to (e.g.) growing biofuel demand is
reflected as a rightward shift in the land demand curve with producers using more
inputs to meet demand. In areas wherein arable land is relatively more available
(leftmost panel in Fig. 4.3) we expect that the cropland supply is more responsive
leading to rising land rents and even faster cropland expansion. However, in regions
wherein additional cropland is relatively scarce (right hand panel in Fig. 4.3), we
observe a sharper rise in land rents, and less cropland area expansion. Cropland sup-
ply response is critical to any analysis of agricultural land use. Indeed, the supply
response at the extensive margin actually reversed analysts’ estimates of the impacts
4.4 Supply Response at the Extensive Margin: Price Elasticity of Cropland Supply 47

Fig. 4.3  Changes in cropland supply given similar increases in land demand when land is abun-
dant (left panel) and when land is scarce (right panel), respectively

of corn ethanol production on GHG emissions, worldwide (Searchinger et al.,

2008). As a direct result of increased corn-ethanol demand in the United States, U.S.
corn harvested area rose sharply, from about 70 million acres in 2001 to a peak of
87 million acres in 2007 (USDA, National Agricultural Statistics Services, 2012).
Of course, this new corn area had to come from somewhere, and much of it came
from soybeans, the harvested area of which fell from 74 to 65 million acres over this
same period (USDA, National Agricultural Statistics Services, 2012). Reduced area
in the U.S. contributed to higher soybean prices, which in turn encouraged addi-
tional production in many parts of the world—particularly South America, which is
quite responsive to developments in the U.S. market (Villoria & Hertel, 2011). In
this way, a price increase for corn in one part of the world is transmitted to other
countries and commodities and contributes to a general rise in the price of field
crops and cropland area worldwide (Naylor et al., 2007). With prices for crops rising
due to the biofuel boom, as well as rising incomes and richer diets—particularly in
Asia, there was strong incentive for expansion of harvested area, and this was indeed
the case over 2000–2013, when world harvested area for cereals, coarse grains, and
oil crops rose by nearly 151 M ha (Fig. 4.4). Furthermore, the area response in this
figure seems to be closely related to price (in this case, food grain prices).
Estimating the acreage response to commodity prices has a long tradition in agri-
cultural economics and this is a key parameter in virtually all of the agricultural
policy and forecasting models of the farm economy (e.g., FAPRI as used in Fabiosa
et al. (2010)). Most of these commodity models are focused on annual changes in
acreage allocated to particular crops. However, given our interest in very long run
changes, at a more aggregate level, it is instructive to draw on the study of Lubowski,
Plantinga, and Stavins (2006) who make use of site-specific USDA—National
Resources Inventory data over two decades to estimate a model of land cover transi-
tions involving aggregated crops as well as other, competing land uses. In their
model, the individual transition probabilities depend on the relative returns to land
use in the different types of cover. These estimates show that U.S. pasture is far more
48 4  Economic Responses to Scarcity

Fig. 4.4  Global harvested area for cereals, coarse grains and oil crops and food grain prices from
1985 to 2013. Source: FAOSTAT (2015) and World Bank GEM commodities database (2015)

responsive to changes in cropland returns than is forest cover—a finding that is

important from the standpoint of limiting carbon emissions from cropland expan-
sion, as conversion from pasture results in much smaller carbon fluxes (Plevin,
Gibbs, Duffy, Yui, & Yeh, 2011). Using this econometric model, Ahmed, Hertel, and
Lubowski (2008) elicit the elasticity of cropland supply in response to changes in the
rental rate on land allocated to cropping. Not surprisingly, this grows over time, start-
ing at 0.05 after 5 years and reaching 0.15 over the course of 20 years. From the
standpoint of the environmental impacts of cropland area expansion, it is important
to factor in not only how much land is converted, but also where this expansion
occurs. Does it arise in carbon rich environments? What are the implications for
biodiversity loss? If, in addition, the land which is converted has relatively low
yields, then more conversion is required to meet the global shortfall. West et al.
(2010) evaluate the ratio of carbon loss from cropland conversion to current agricul-
tural yields around the world, and find that it is particularly high in the tropics, sug-
gesting that induced land use change in that region is likely to carry with it a
particularly high environmental cost. This is of special concern, given the geographic
profile of harvested area change in recent years. Hertel and Tyner (2013) report that,
over the 2005/06–2011/12 period, the greatest cropland expansion has come in South
America, South Asia, East Asia and Sub-Saharan Africa, along with Eastern Europe.
Gibbs et al. (2010) seek to understand not only the broad region in which crop-
land expansion occurs, but more specifically, the land cover type from which the
added cropland is drawn. They focus on observations at specific sites throughout the
tropics sampled over the period from 1980 to 2000. A striking finding of their study
is that 55 % of the sampled sites which were converted to cropland over this period
had been covered by forest at the start of the period. Of course, all such historical
case studies are limited in that they do not control for other variables of interest
which might play a role in land conversion decisions.
4.4 Supply Response at the Extensive Margin: Price Elasticity of Cropland Supply 49

A critical factor in determining the amount of land which must be converted fol-
lowing an expansion in demand is the productivity of the newly converted land,
relative to existing cropland. From an economist’s point of view, the presumption is
typically that land in the same general region, but not currently under crops, must
be less productive—otherwise it would already be under cultivation by profit-seek-
ing producers. However, that need not be the case if there are other barriers to bring-
ing the land into crop production, or if additional investments are required which
may then bring that land up to the average productivity level of other land sharing
similar soil and climatic characteristics. The latter is the case in the Brazilian
Cerrado, where very significant investments in land improvements are required
before planting soybeans, but once these improvements have been made, soybean
yields on the newly converted lands are comparable to those on existing cultivated
hectares (Deininger & Byerlee, 2010).
Estimating the relative productivity of marginal land for each region of the world
is clearly a large task. The early work on estimating the global land use impacts of
biofuels expansion abstracted from this aspect of the problem, assuming instead that
any reduction in average yield due to area expansion would be offset by intensifica-
tion of production (Searchinger et al., 2008). Subsequent work by Hertel et al.
(2010) used an estimate 0.66 as the ratio of productivity of marginal lands relative
to average productivity, based on unpublished data on land idled under the
U.S. Conservation Reserve Program. Taheripour, Zhuang, Tyner, and Lu (2012)
refined the estimates of the productivity of marginal lands using the Terrestrial
Ecosystem Model (TEM) to estimate relative productivity of non-cropland to crop-
land at the grid cell level, thereupon aggregating to global Agro-Ecological Zones.
Using these more disaggregated estimates of the productivity of marginal land
instead of the previous globally uniform value, they find that the land requirement
for U.S. biofuel mandates is 25 % lower than previous estimates using the blanket
estimate of 0.66 for the productivity ratio..
In addition to the question of where land conversion is likely to arise, an important
question is: How much additional land could feasibly be brought into production in
the aggregate? The study by Lambin et al. (2013) highlights the fact that there are
important local constraints to land expansion in many regions, such as zoning ordi-
nances, nature preserves and insufficient infrastructure, which are not factored into
most, if not all, global analyses of land cover change. The presence of such biophysi-
cal and socio-economic constraints suggests that the elasticity of land supply in some
regions may be overstated—resulting in greater land scarcity than might otherwise
be predicted. In addition, the authors highlight the considerable uncertainty associ-
ated with current land use, at global scale. For example, they place the range of pos-
sible crop land cover in the year 2000 somewhere between 1510 and 1611 M ha. For
pastures, the range is even larger: 2500–3410 M ha. This uncertainty is compounded
when one seeks to project global land use in 2030. They suggest that the growth in
land requirements for agriculture, forestry, urbanization and protected areas could be
as low as 285 M ha and as high as 792 M ha. Depending on where on this spectrum
the true outcome resides, the pressure for additional deforestation could vary greatly.
50 4  Economic Responses to Scarcity

As pointed out by Lambin (2012), over the past two centuries, there have been
two competing views of global land use—each associated with a famous 19th phi-
losopher. On the one hand we have Thomas Malthus—famous for the Malthusian
scenario in which resources are finite and population grows to the point at which it
can no longer be sustained, with equilibration occurring through adjustment in liv-
ing standards and life expectancy. On the other hand we have David Ricardo, who
believed that, as population and income growth made land scarcer, the ensuing rise
in land rents would induce investments in land improvement, aimed at making the
existing land more productive.1
Ricardo’s hypothesis finds support in the history of the Brazilian Cerrado region
which remained largely uncultivated prior to the 1970s due to its acidic soils and
high market access costs. However, with the boom in soybean production in Brazil,
there were strong incentives to invest in new technologies which could modify the
soils and allow expansion of production into these previously unproductive areas
(World Bank, 2009). The Cerrado now accounts for more than half of Brazil’s total
soybean area and this experience has led analysts to suggest that a similar agro-
nomic and economic revolution could occur in the Guinea-Savannah Zone of
Africa—a region comprising 400 M ha of potential crop land, of which just 10 %
are currently cropped. Indeed, the high commodity prices during the period: 2006–
2012 resulted in a flood of private investment into Africa—the so-called “Land
Grab”—in the hopes of taking advantage of just such potential. However, to date,
the record of such investments is mixed, and due to poor documentation of land
tenure and weak land governance, such leases have sometimes been undertaken on
land which was communally held, thereby stripping the local population of their
livelihoods (Deininger & Byerlee, 2010). Of course such market-driven invest-
ments also tend to ignore the ecological costs of land expansion into environmen-
tally sensitive regions—particularly in the absence of strong local institutions with
a mandate to protect the environment. Lambin (2012) concludes his discussion of
the ‘debate’ between Malthus and Ricardo with the following remark (p. 86):
“Thus, while a Malthusian view of land as a finite resource is too narrow and
ignores many factors such as international trade and the great room for maneuver-
ing that derives from a changing geographic distribution of land use, a Ricardian
view of land is too optimistic as it ignores the social and ecological costs that
increase rapidly with the conversion of more natural ecosystems, indigenous lands,
ecological corridors, carbon-­rich ecosystems, or lands that provide other important
ecosystem services.”

1
 Geographers have added to the Ricardian adjustment mechanisms by documenting the way in
which land tends to be sorted over time to achieve a better match of land use with land potential
(Mather & Needle, 1998).
4.6 The Role of Economic Responses in the Global Land-­Sparing Debate 51

4.5  T
 he Three Economic Responses at Work: Cropland Use
from Global Biofuels

Hertel, Steinbuks, and Baldos (2012) examine the relative importance of the supply
and demand margins of economic response in the context of the International
Energy Agency’s projections of biofuel production growth over the 2006–2035
period, under baseline policies. Those authors find that, with the extensive margin
alone in place, global cropland would expand by a projected 8.7 % as a consequence
of this growth in biofuel demand. Bringing the intensive margin of supply response
into play drops the global land requirement to 3.8 %, and adding the demand margin
further reduces it to 3.1 %. This example offers a number of valuable insights. First
of all, analyses which ignore the economic responses to scarcity and simply require
more cropland as global demand grows, relative to supply, would overstate the
expected cropland conversion by more than 100 %, estimating an 8.7 % expansion
instead of a 3.1 % growth in cropland. This has major implications for global
­terrestrial carbon stocks and biodiversity, as we will see in subsequent chapters.
A second point pertains to the relative importance of the supply and demand mar-
gins. In this example, the bulk of the ‘work’ in mitigating the global scarcity induced
by biofuels comes from the intensive margin of supply—with demand response
adding relatively less. This makes sense, since, in the long run, consumer demand is
less responsive to price than is supply.

4.6  T
 he Role of Economic Responses in the Global
Land-­Sparing Debate

Bringing to bear the three margins of economic response to scarcity in the global
food system can shed important light on a variety of issues. One of these is the
debate over whether or not improvements in agricultural technology ‘spare land’
(Rudel et al., 2009; Stevenson, Villoria, Byerlee, Kelley, & Maredia, 2013). The
land-sparing hypothesis was first posed by Norman Borlaug (Borlaug, 1976), the
so-called ‘father of the Green Revolution’. The advances in crop-breeding which
grew out of his work, and that of his collaborators, resulted in a massive increase in
staple grain yields in Asia, Latin America and the Middle East. Borlaug argued that,
given a fixed demand for global food, being able to grow that food on fewer hectares
of land obviously spared additional land for nature. However, many localized case
studies of agricultural innovations apparently contradict this finding, concluding
instead that such innovations lead to increased deforestation and cropland conver-
sion (Angelsen & Kaimowitz, 2001). We can use our economic framework to shed
additional light on this question.
Figure 4.5 offers a simplified, supply-demand depiction of the impact of the
Green Revolution (GR) on global food markets. In the left hand portion of this dia-
gram, we see the supply curve for crops produced in the region experiencing the
52 4  Economic Responses to Scarcity

Fig. 4.5  Three panel depiction of the impact of the green revolution on global food markets.
Source: Hertel et al. (2014)

GR, which is modeled as an outward shift in this curve with producers supply more
output at the same price. This outward shift in regional supply has an impact on the
world markets where the global supply of crops is the horizontal summation of sup-
plies in the GR region and in the Rest of the World (RoW). With world supplies
shifting outwards, the (fixed) downward sloping world demand for food traces out
the ensuing decline in crop price, thereby signaling a relative abundance of crops on
the world market and triggering the price elasticities of demand and supply. In par-
ticular, consumers purchase more food (the demand margin) and producers in RoW
reduce supplies by reducing area (the extensive margin) and also lessening the
intensity of cultivation (the intensive margin) in response to lower world prices.
However, the impact on supplies in the GR region is ambiguous.
Those authors who argue against the land-sparing hypothesis typically focus on
land use change in region A—the region where the innovation occurs. Here, we can
see immediately that the quantity of crops produced expands from QA to QA*.
However, we cannot discern from the figure whether land use expands. This is
because yields also rise, permitting more output to be produced on current lands. It
turns out that the condition for expansion depends on the slope of the excess demand
schedule (EDA), which, in turn depends on the global price elasticity of demand for
food, as well as the price elasticity of supply in RoW (Hertel, Ramankutty, &
Baldos, 2014). If region A is small, relative to RoW, and if there is significant sup-
ply response in RoW, then the price elasticity of excess demand will be large (in
absolute value) and land use in region A will rise—regardless of the price elasticity
of food demand. And this land expansion in the wake of agricultural innovation is
what many of the case studies find, leading them to the apparent refutation of the
land-sparing hypothesis. However, they have ignored altogether the question of land
use in the Rest of the World which clearly declines.
Hertel et al. (2014) explore the land-sparing issue in considerable detail and find
that the answer depends critically on the relative sizes of the three margins of eco-
nomic response, as well as relative yields in the two regions at the outset of the
GR. Key results from their counterfactual analysis of the historical GR using the
SIMPLE model are shown in Fig. 4.6. The first set of bars in the figure report the
4.6 The Role of Economic Responses in the Global Land-­Sparing Debate 53

Fig. 4.6  Impact of the green revolution on global and regional output, area and yields: 1961–
2006. Green bars depict the historical baseline in the presence of the GR and red bars show the
counterfactual no-GR case. Source: Hertel et al. (2014)

historical (model baseline) evolution of output, cropland area and yields in the
world as well as in the GR and RoW regions. Note that the majority of the global
output expansion over this historical period came from increased yields, as opposed
to increased area. From the green bars (GR in place, as observed historically) we see
that global output roughly tripled, thereby driving prices down over this period.
However, when the authors rerun history (second set of bars in the figure), removing
the supply shift owing to the GR, output and yields increase much less, prices rise
and cropland area increases. Furthermore, we see from the bottom panel that crop-
land increases, relative to the historical baseline, in both the GR and RoW regions.
Clearly this historical GR was land-sparing.
In their paper, these authors go on to examine a prospective African GR begin-
ning in 2025 and extending through 2050. They find that such a GR may not be
land-sparing, due to the relatively low yields, and relatively strong cropland area
response, in Africa. When production shifts from the rest of the world to Africa,
more, rather than less land is converted globally. Thus it is not only the absolute size
of the economic responses to scarcity that matters in global sustainability analysis,
but also the relative size across regions of the world.
In summary, this chapter has introduced the key economic responses to scarcity
which serve to differentiate economic analysis of long run land use and sustainabil-
ity of the food system from purely biophysical approaches. Adding this dimension
to the analysis allows us to better understand the last 50 years of global agricultural
54 4  Economic Responses to Scarcity

development. We have seen that models which ignore these responses will overstate
future cropland area requirements, and are also likely to prove misleading when
evaluating contemporary issues such as the change in global GHG emissions in the
wake of technological change in agriculture. In the remainder of the book, these
economic margins of response to scarcity will play a central role in our analyses of
a range of global sustainability questions.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change

References

Ahmed, S. A., Hertel, T., & Lubowski, R. (2008). Calibration of a land cover supply function using
transition probabilities (GTAP Research Memorandum No. 14). West Lafayette, IN: Center for
Global Trade Analysis, Department of Agricultural Economics, Purdue University. Retrieved
from http://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2947.
Angelsen, A., & Kaimowitz, D. (Eds.). (2001). Agricultural technologies and tropical deforesta-
tion. New York, NY: CAB.
Baldos, U. L. C., & Hertel, T. W. (2014). Global food security in 2050: The role of agricultural
productivity and climate change. Australian Journal of Agricultural and Resource Economics,
58(4), 554–570. http://doi.org/10.1111/1467-8489.12048.
Balmford, A., Green, R. E., & Scharlemann, J. P. W. (2005). Sparing land for nature: Exploring the
potential impact of changes in agricultural yield on the area needed for crop production. Global
Change Biology, 11(10), 1594–1605. http://doi.org/10.1111/j.1365-2486.2005.001035.x.
Baumert, K. A., Herzog, T., & Pershing, J. (2005). Navigating the numbers: Greenhouse gas data
and international climate policy (p. 122). Washington, DC: World Resources Institute.
Retrieved from http://www.wri.org/publication/navigating-the-numbers.
Beckman, J., Keeney, R., & Tyner, W. (2011). Feed demands and coproduct substitution in the
biofuel era. Agribusiness, 27(1), 1–18. http://doi.org/10.1002/agr.20247.
Berry, S., & Schlenker, W. (2011). Empirical evidence on crop yield elasticities (Technical Report,
p. 25). Washington, DC: International Council on Clean Transportation. Retrieved from http://
www.theicct.org/empirical-evidence-crop-elasticities.
Borlaug, N. (1976). Forestry in a world of limited resources: Mobilizing world land resources to
4meet the growing needs for food, fibre, forest products, wildlife and recreation. In XVI world
congress, International Union of Forest Research Organizations, Oslo (pp. 171–245).
Deininger, K. W., & Byerlee, D. (2010). Rising global interest in farmland: Can it yield sustain-
able and equitable benefits? Washington, DC: The World Bank.
Fabiosa, J. F., Beghin, J. C., Dong, F., Elobeid, A., Tokgoz, S., & Yu, T.-H. (2010). Land allocation
effects of the global ethanol surge: Predictions from the international FAPRI model. Land
Economics, 86(4), 687–706.
Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R., … Snyder, P. K.
(2005). Global consequences of land use. Science, 309(5734), 570–574. http://doi.org/10.1126/
science.1111772.
Food and Agriculture Organisation of the UN. (2015). FAO Statistical Database. Rome, Italy.
Retrieved from http://faostat.fao.org/.
Fulginiti, L. E., & Perrin, R. K. (1993). Prices and productivity in agriculture. The Review of
Economics and Statistics, 75(3), 471–482. http://doi.org/10.2307/2109461.
Gibbs, H. K., Ruesch, A. S., Achard, F., Clayton, M. K., Holmgren, P., Ramankutty, N., & Foley,
J. A. (2010). Tropical forests were the primary sources of new agricultural land in the 1980s
References 55

and 1990s. Proceedings of the National Academy of Sciences of the United States of America,
107(38), 16732–16737. http://doi.org/10.1073/pnas.0910275107.
Golub, A., & Hertel, T. (2012). Modeling land use change impacts of biofuels in the GTAP-BIO
model. Climate Change Economics, 3(3) , 1--30.
Goodwin, B., Marra, M., Piggott, N., & Mueller, S. (2012). Is yield endogenous to price? An
empirical evaluation of inter- and intra-seasonal corn yield response. In AAEA annual meeting.
Seattle, WA: North Carolina State University.
Hertel, T. W., Golub, A. G., Jones, A., O’Hare, M., Plevin, R., & Kammen, D. (2010). Global land
use and greenhouse gas emissions impacts of U.S. maize ethanol: Estimating market-mediated
responses. BioScience, 3, 223–231.
Hertel, T. W., Ramankutty, N., & Baldos, U. L. C. (2014). Global market integration increases
likelihood that a future African Green Revolution could increase crop land use and CO2 emis-
sions. Proceedings of the National Academy of Sciences of the United States of America,
111(38), 13799–13804. http://doi.org/10.1073/pnas.1403543111.
Hertel, T., Steinbuks, J., & Baldos, U. L. (2012). Competition for land in the global bioeconomy.
Retrieved from http://www.gtap.agecon.purdue.edu/resources/res_display.
asp?RecordID=3982.
Hertel, T. W., & Tyner, W. E. (2013). Market-mediated environmental impacts of biofuels. Global
Food Security, 2(2), 131–137. http://doi.org/10.1016/j.gfs.2013.05.003.
Huang, H., & Khanna, M. (2010). An econometric analysis of U.S. crop yield and cropland acre-
age: Implications for the impact of climate change. In AAEA, CAES, & WAEA Joint Annual
Meeting. Denver, CO: Department of Agricultural and Consumer Economics, University of
Illinois-Urbana-Champaign.
Keeney, R., & Hertel, T. (2008). The indirect land use impacts of U.S. biofuel policies: The impor-
tance of acreage, yield, and bilateral trade responses. Center for Global Trade Analysis,
Department of Agricultural Economics, Purdue University. Retrieved from http://ideas.repec.
org/p/gta/workpp/2810.html.
Lambin, E. F. (2012). Global land availability: Malthus versus Ricardo. Global Food Security,
1(2), 83–87. http://doi.org/10.1016/j.gfs.2012.11.002.
Lambin, E. F., Gibbs, H. K., Ferreira, L., Grau, R., Mayaux, P., Meyfroidt, P., … Munger, J. (2013).
Estimating the world’s potentially available cropland using a bottomup approach. Global
Environmental Change, 23, 892–901.
Lubowski, R. N., Plantinga, A. J., & Stavins, R. N. (2006). Land-use change and carbon sinks:
Econometric estimation of the carbon sequestration supply function. Journal of Environmental
Economics and Management, 51(2), 135–152. http://doi.org/10.1016/j.jeem.2005.08.001.
Mather, A. S., & Needle, C. L. (1998). The forest transition: A theoretical basis. Area, 30(2),
117–124. http://doi.org/10.1111/j.1475-4762.1998.tb00055.x.
Muhammad, A., Seale, J. L., Jr., Meade, B., & Regmi, A. (2011). International evidence on food
consumption patterns: An update using 2005 International Comparison Program Data
(Technical Bulletin No. TB-1929, p. 59). Washington, DC: Economic Research Service, US
Department of Agriculture. Retrieved from http://www.ers.usda.gov/Publications/TB1929/.
Naylor, R. L., Liska, A. J., Burke, M. B., Falcon, W. P., Gaskell, J. C., Rozelle, S. D., & Cassman,
K. G. (2007). The ripple effect: Biofuels, food security, and the environment. Environment:
Science and Policy for Sustainable Development, 49(9), 30–43. http://doi.org/10.3200/
ENVT.49.9.30-43.
Phalan, B., Onial, M., Balmford, A., & Green, R. E. (2011). Reconciling food production and
biodiversity conservation: Land sharing and land sparing compared. Science, 333(6047),
1289–1291. http://doi.org/10.1126/Science.1208742.
Plevin, R. J., Gibbs, H. K., Duffy, J., Yui, S., & Yeh, S. (2011). Agro-ecological zone emission
factor model. Retrieved from http://plevin.berkeley.edu/docs/Plevin-AEZ-EF-Model-
Preliminary-­Sep-2001.pdf.
Potter, P., Ramankutty, N., Bennett, E. M., & Donner, S. D. (2010). Characterizing the spatial pat-
terns of global fertilizer application and manure production. Earth Interactions, 14(2), 1–22.
56 4  Economic Responses to Scarcity

Ramankutty, N. (2010). Agriculture and forests: Recent trends, future prospects. In T. Graedel &
E. van der Voet (Eds.), Linkages of sustainability (Vol. 4, pp. 11–31). Cambridge, MA: MIT
Press.
Rudel, T. K., Schneider, L., Uriarte, M., Turner, B. L., DeFries, R., Lawrence, D., … Grau, R.
(2009). Agricultural intensification and changes in cultivated areas, 1970–2005. Proceedings of
the National Academy of Sciences of the United States of America, 106(49), 20675–20680.
http://doi.org/10.1073/pnas.0812540106.
Seale, J., Regmi, A., & Bernstein, J. (2003). International evidence on food consumption patterns.
United States Department of Agriculture, Economic Research Service. Retrieved from http://
ideas.repec.org/p/ags/uerstb/33580.html.
Searchinger, T., Heimlich, R., Houghton, R. A., Dong, F., Elobeid, A., Fabiosa, J., … Yu, T. -H.
(2008). Use of U.S. croplands for biofuels increases greenhouse gases through emissions from
land-use change. Science, 319(5867), 1238–1240. http://doi.org/10.1126/science.1151861.
Stevenson, J. R., Villoria, N., Byerlee, D., Kelley, T., & Maredia, M. (2013). Green Revolution
research saved an estimated 18 to 27 million hectares from being brought into agricultural
production. Proceedings of the National Academy of Sciences of the United States of America,
110(21), 8363–8368. http://doi.org/10.1073/pnas.1208065110.
Taheripour, F., Zhuang, Q., Tyner, W. E., & Lu, X. (2012). Biofuels, cropland expansion, and the
extensive margin. Energy, Sustainability, and Society, 2, 25.
Tilman, D., Cassman, K. G., Matson, P. A., Naylor, R., & Polasky, S. (2002). Agricultural sustain-
ability and intensive production practices. Nature, 418(6898), 671–677. http://doi.org/10.1038/
nature01014.
US Department of Agriculture, National Agricultural Statistics Services. (2012). USDA Crop
Statistics. Retrieved from http://www.nass.usda.gov/Statistics_by_Subject/index.php?
sector=CROPS.
Villoria, N. B., & Hertel, T. W. (2011). Geography matters: International trade patterns and the
indirect land use effects of biofuels. American Journal of Agricultural Economics, 93(4), 919–
935. http://doi.org/10.1093/ajae/aar025.
Vitousek, P. M., Naylor, R., Crews, T., David, M. B., Drinkwater, L. E., Holland, E., … Zhang,
F. S. (2009). Nutrient imbalances in agricultural development. Science, 324(5934), 1519–1520.
http://doi.org/10.1126/science.1170261.
West, P. C., Gibbs, H. K., Monfreda, C., Wagner, J., Barford, C. C., Carpenter, S. R., & Foley, J. A.
(2010). Trading carbon for food: Global comparison of carbon stocks vs. crop yields on agri-
cultural land. Proceedings of the National Academy of Sciences of the United States of America,
107(46):19645–19648. http://doi.org/10.1073/pnas.1011078107.
World Bank. (2009). Awakening Africa’s sleeping giant: Prospects for commercial agriculture in
the Guinea Savannah Zone and beyond (pp. 1–236).
World Bank. (2015). Global economic monitor: Commodities. Washington, DC. Retrieved from
http://data.worldbank.org/data-catalog/world-development-indicators
.
Chapter 5
Water, Food and Environmental Security*

5.1 Water Use in Agriculture

How much water is used to produce the world’s food supply each year? Crop evapo-
transpiration alone (not including water for food processing and preparation) con-
sumes about one liter of water per calorie of food produced (Molden, 2007). Based
on this ratio, the total amount of water evaporated in a year to feed today’s 7.1 bil-
lion people would fill a canal 10 meters deep, 100 meters wide and 7.7 million
meters long so that it would encircle the globe 193 times! One-fifth of this total
water use comes from the application of irrigation water which, in turn, accounts for
70 % of total annual global freshwater withdrawals (Molden, 2007) and about 40 %
of crop production (Bruinsma, 2009). Since the latter is based on irrigated cropland
which amounts to less than 20 % of total cropland, it is clear that irrigated lands
exhibit above-average yields, and contribute importantly to dampening agricultural
land use expansion and strengthening global food security.
Water productivity in agriculture varies greatly by region, as does the irrigation
intensity of crop production. Productivity is highest in moist, temperate regions,
such as the UK, northern France, and Germany where Zwart, Bastiaanssen, de
Fraiture, and Molden (2010) estimate productivity rates of roughly 1.5 kg of wheat
per cubic meter of water. In contrast, in parts of the Middle East where conditions
are dry and hot, and evapotranspiration rates are very high, water productivity in
wheat production is just one-fifth as large. Keller and Seckler (2008) point out that,
once one controls for evaporative demand, there is a linear relationship between
crop biomass and transpiration. If you want to double crop output, then it would

*This chapter draws heavily on the OECD report written by Hertel and Liu (2014).

© Springer International Publishing Switzerland 2016 57

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_5
58 5 Water, Food and Environmental Security

appear you need to double water use—a fact which raises concerns in light of the
world’s need to boost crop production by as much as 70 % by mid-century
(Alexandratos & Bruinsma, 2012). However, Keller and Seckler (2008) recognize
that there are avenues for weakening the link between transpiration and crop output.
These include: boosting the harvest index (share of grain in total biomass), increas-
ing plant density which can reduce evaporative losses, and increasing the plant’s
transpiration efficiency.
Whereas Keller and Seckler (2008) focus on water’s role in the plant’s ability to
generate biomass, Rockström et al. (2010) take a broader view of hydrological effi-
ciency, considering the delivery of water to the plant as well. Focusing on rainfed
maize in Sub-Saharan Africa, they argue that reducing non-productive evaporative
losses and eliminating nutrient deficiencies so that all water accessible in the root
zone is used productively, farmers could theoretically boost yields from 1 to 3 t/ha,
on average. Limiting deep percolation of rainfall could result in yields rising from 3
to 5 t/ha And finally, they argue that technologies which allow farmers to capture all
the rainfall on a field could boost yields further, from 5 to 7 t/ha. In short, they do
not see hydrological limitations to a doubling of yields in this region of the world.
Rather, the key issue is one of managing existing water.
Focusing on what is currently achieved in practice, Rockström et al. (2010) plot
the inverse of water productivity (cubic meters/Mt) against yield for a wide range of
crops, as reproduced in Fig. 5.1. Depending on the type of crop, where it is pro-
duced, and under what conditions, water productivity varies widely. Nonetheless,
water productivity clearly increases with yields, but the rate of increase falls off
rapidly as yields rise, and it appears to plateau at 5 t/ha. Thus, low-yielding, small-
holder agriculture in the tropics can likely increase yields significantly without
requiring proportionate increases in water input. However, in this cross-section
analysis, modern high productivity systems appear to show the linear relationship
between water use and yields predicted by Keller and Seckler (2008).

Fig. 5.1 Nonlinear

relationship between
water productivity and
grain yields for cereal
crops. Source: Rockström
et al. (2010)
5.2 Irrigated Agriculture 59

Beyond improving the efficiency of on-farm water use, there is also great poten-
tial for improving the efficiency with which irrigation water is collected and deliv-
ered to the farm. This includes evaporation from water storage, which accounts for
a significant fraction of losses in some regions. For example, reservoir evaporation
in Texas amounted to about 61 % of total agricultural irrigation use during the year
2010 (Wurbs & Ayala, 2014). In Australia, Craig (2005) estimated this loss to be
about 40 % of the total storage volume. This evaporative loss could increase by
about 15 % by 2080, due to the effect of higher surface temperatures in the face of
climate change (Helfer, Lemckert, & Zhang, 2012). Increasing total usable water
storage by reducing this type of loss depends on the adoption of evaporation sup-
pression technology, which is driven by the marginal value product of the water to
be saved. Another avenue for water savings is the improved delivery of the water to
the field—and ultimately to the crop.
Water is also important for livestock production, which accounts for 20 % of total
water used by agriculture (De Fraiture et al., 2007). This includes both the direct use
like animal drinking and feed-mixing, as well as the indirect use for grazing and
growing feed crops. However, 98 % of the water consumption is attributed to the
latter—evapotranspiration of blue and green water in the production of feedstuffs.
Producing livestock is generally believed to be much more water intensive than pro-
ducing crops (Mekonnen & Hoekstra, 2012). For example, the global average water
footprint per Mt of beef is about 50 times that of vegetables and 10 times to that of
cereals. In the U.S., beef production requires 1.6 m3of irrigation water/Mcal pro-
duced, vs. less than 0.1 for wheat and 0.4 for rice (Eshel, Shepon, Makov, & Milo,
2014). However, water for livestock is relatively less costly if we compare beef and
crops in terms of protein content. The global water footprint per gram of beef protein
is five times of that for cereal protein, and water footprint per gram of beef fat is
three times of that for nuts fat (Mekonnen & Hoekstra, 2012).
Over time, the livestock industry worldwide has become more industrialized,
with greater reliance on intensive production techniques, and less reliance on graz-
ing and extensive practices (Taheripour, Hertel, et al., 2013). This has important
implications for water efficiency. Grazing systems are less efficient in feed conver-
sion and therefore require more land resources than the industrial systems, but graz-
ing systems require a lower fraction of feed concentrates, thereby placing less stress
on irrigation water, and generally result in less water pollution, thereby allowing for
greater reuse of water. Thus, the shift to intensive livestock production represents a
mixed bag when it comes to pressures on the world’s water resources.

5.2 Irrigated Agriculture

The FAO estimates that 16 % of the world’s cultivatedcropland is equipped for

irrigation (Alexandratos & Bruinsma, 2012). However, the intensity of irrigation
varies considerably across regions, as shown in Fig. 5.2. About 70 % of the area
equipped for irrigation is located in 15 Asian countries vs. 16 % in America, 8 % in
60 5 Water, Food and Environmental Security

Fig. 5.2 Area equipped for irrigation: historical and projected (M ha). Source: Alexandratos and
Bruinsma (2012)

Europe, 5 % in Africa and 1 % in Oceania (Siebert et al., 2010). Between the present
and 2050, the FAO expects that net global irrigated area will continue to expand, but
more slowly compared with the historical growth, as a result of overall adequacy of
current food supplies, increasing scarcity of suitable areas for irrigation, more
intense competition for water and rising importance of investment in other sectors
(Alexandratos & Bruinsma, 2012). The estimated 20 million ha net expansion by
mid-century is expected to occur almost exclusively in land scarce developing coun-
tries (Fig. 5.2) where increases in agricultural output will depend heavily on yield
increases (Alexandratos & Bruinsma, 2012; Bruinsma, 2009). Unfortunately, the
cropping areas most in need of irrigation are often located in regions where water
resources are naturally scarce and facing increasingly stiff competition from other
uses. This points to rising opportunity costs of water devoted to irrigation and
greater pressure to conserve irrigation water.
The potential for increasing water supply for irrigation is subject to infrastruc-
ture constraints (storage and withdrawal facilities) and water resource limits
(Rosegrant & Cai, 2002). The latter are determined both by hydrological flows and
by competing demands for water in residential, industrial and environmental uses.
Agriculture is typically the residual claimant of water within a given basin. Keeping
infrastructure capacity and water use efficiency unchanged, the water budget for
irrigation is expected to dwindle as the economy and populations grow. Strzepek
and Boehlert (2010) predict an 18 % reduction in worldwide water irrigation avail-
able for agriculture by 2050, mainly caused by increasing environmental flow
requirements and larger municipal and industrial demands. In hotspots like northern
Africa, India, China, parts of Europe, the western United States and eastern
Australia, the reduction in water for irrigation could be much more dramatic.
5.2 Irrigated Agriculture 61

Forty-two percent of the world’s irrigation water withdrawals come from ground-
water, and the rest from surface water (Döll et al., 2012). These two sources of water
offer fundamentally different attributes and also face differing hydrological con-
straints. Groundwater is generally less vulnerable to climate variability than is sur-
face water. It is available ‘on demand’. The temporal flexibility of groundwater
supply is critical for agriculture, since water stress is particularly problematic for crop
growth at specific points in the growing season. Therefore, we expect that the impor-
tance of ground water with regard to irrigation supply will intensify as more frequent
and severe extreme weather events increase variability in precipitation, soil moisture
and surface water (Taylor et al., 2012). During California’s 2007–2009 drought, for
example, the source of irrigation water shifted from surface water to mostly ground-
water. However, the slow rate of recovery of groundwater to a dynamic equilibrium
state means withdrawal can easily surpass replenishment and lead to groundwater
depletion. Over-drafting, especially in the regions that are becoming heavily depen-
dent on groundwater, weakens its buffering effect, potentially making agriculture
even more fragile in the face of longer drought duration. Pavelic et al. (2012) show
that the average residence time of shallow, accessible groundwater ranges from less
than 1 to 4 years, which explains why 2 or more years of continuous drought can pose
a serious problem to farmers relying on groundwater for irrigation.
Another advantage of groundwater has to do with accessibility to farmers. While
surface water rights are often predetermined so that access involves engagement
with other institutions, groundwater can generally be accessed by simply drilling a
well—something under direct control of the farmer. Both of these factors have con-
tributed to rapid growth in groundwater withdrawals in many regions. Burke and
Villholth (2007) show the recent evolution of ground water use in various regions
around the world. The growth in India since the inception of the Green Revolution
is staggering. Bangladesh, China, Mexico and Tunisia also show very strong growth.
In contrast to groundwater, global surface water withdrawals have been slowing
down recently, from a growth rate of 2 % in the 1980s to −1 % during 1990–2010
(Wada et al., 2013), largely due to the fact that surface water has already been heav-
ily exploited and that the construction of new reservoirs has been declining since the
1980s (Chao et al., 2008; Rosegrant et al., 2013). Nevertheless, surface water
remains the dominant source of irrigation in Europe (70 %), Southeast Asia (more
than 80 %) and South America (Wada et al., 2013). Surface water supplies are
highly climate-dependent. Watershed responses to reduced rainfall and higher tem-
peratures are typically amplified, due to vegetation interception and transmission
loss (Arnell, 2004). Thus, a 20 % reduction in rainfall might yield a 50 % reduction
in runoff (Turral et al., 2010). Although warming tends to increase total precipitation
and water discharge at the global level (Füssel et al., 2012; Milly et al., 2005), this
tendency may not translate directly into a more beneficial effect on surface water
irrigation in critical regions. This is due to the fact that climate models are predict-
ing an uneven spatial and temporal distribution of rainfall, in which the wet areas
get wetter and the dry areas drier. Besides, even where annual precipitation is not
expected to decline, seasonal shifts may cause substantial problems if the increased
water runoff in rainy season cannot be impounded due to limited storage capacity.
62 5 Water, Food and Environmental Security

5.3 Adjusting to Water Scarcity

As some regions in the world face increased water scarcity, the question arises: How
will this scarcity be resolved? And how will the method of resolution affect food
production, consumption, nutrition and environmental security? In a comprehensive
global study, undertaken at the level of individual river basins, and utilizing a suite
of economic and hydrologic models, Rosegrant, Ringler, Zhu, Tokgoz, and Bhandary
(2013) compare across two across two 2030 scenarios the level of Irrigation Water
Supply Reliability (IWSR: the ratio of irrigation water supplies to demand, 1.0 is
best). The first is the ‘Business As Usual’ (BAU) scenario under which agricultural
productivity and water use efficiency in the agricultural, industrial and domestic
uses sectors reflect current trends. The authors find that, under the BAU scenario,
the global IWSR falls from 0.77 in 2000 to 0.62 in 2050. The decline is particularly
sharp in the East and South Asia regions.
Liu, Hertel, Taheripour, Zhu, and Ringler (2014) examine the implications for
agricultural land use, production and trade of this BAU scenario using a global eco-
nomic model of international trade which embodies competition for water at the
river basin level. They find that projected shortages in South Asia and China are
likely to contribute to global cropland area expansion (7.6 M ha) as well as signifi-
cant shifts reductions in production in the affected regions. However, the impact on
global food prices as well as global economic welfare is quite modest as the
increased supply of food from more water abundant regions makes up for these
regional shortfalls. Indeed, they project that these future water shortages for irriga-
tion will result in increased ‘virtual’ water trade, with water intensive crops being
produced increasingly in the more water abundant regions, and shipped to the
regions experiencing scarcity—most notably Pakistan, India and China.
In sharp contrast to their BAU scenario, Rosegrant et al. (2013) also explore a
counterfactual, ‘Bioeconomy’ (BIO) scenario in which they allow for faster agricul-
tural productivity growth, due to increased R&D expenditures, as well as significant
improvements in water use efficiency—particularly for the non-agricultural sectors.
This, too, is important for irrigation, since water available for agriculture is often a
residual, based on availability after other water demands have been satisfied. In this
case, the combination of higher agricultural productivity and increased water use
efficiency allow for a global IWSR of 0.73 in 2030, with far smaller declines in the
Asian regions, as well as increases in some of the other regions. In short, making
more efficient use of water—both in irrigation and in non-farm uses—is critical for
ensuring global sustainability of agricultural production by mid-century.
Of course, improvements in the efficiency of water use hinge critically on the
management of river basins. Grafton et al. (2013) point out that there are significant
opportunities for improving the governance of water, and thereby its allocation
across uses—including in-stream flows providing environmental benefits. In their
review of four of the world’s major river basins, Grafton et al. (2013) highlight the
successes achieved in Australia (Murray-Darling Basin) and China (Yellow River)
where significant increases in environmental flows have been attained at only modest
5.4 Interactions Between Irrigation and Land Use: Application of SIMPLE 63

cost to existing users. In both cases, centralized planning and coordination was criti-
cal, although the methods of implementation were polar opposites: China imposed
reductions administratively (top-down), while in the Australian case, reallocations
were achieved via voluntary sales of water rights to the government (bottom-up).
While few countries have the administrative authority to manage water alloca-
tions in a top-down fashion, the Australian experience in dealing with water scarcity
is more likely to be generalizable and is therefore worthy of further discussion here.
Beginning in 1983, and continuing for the next decade, water trading was intro-
duced in Australia (Bjornlund & McKay, 2002). This has proven to be crucial instru-
ment for allowing adjustment to periodic droughts, such as those experienced in
2002/3 when water usage was cut by nearly 30 % in the Murray-Darling Basin.
Analysts expected only a modest drop in water required for irrigation, but producers
responded much more vigorously. Indeed, rice producers in the region went from
using all of their general water allocation in 2001/02 to using less than 10 % of this
allocation in 2002/03. And rice production in the Basin dropped from the 600,000–
800,000 ton range to nearly zero (National Water Commission, 2011). Instead of
growing and selling rice, the producers voluntarily sold their temporary, annual
water allocation to the government which offered a higher price than could be earned
from rice production in that year. This decline in rice production was offset by
increases elsewhere around the world. As a consequence, it was possible to supply
water to non-farm uses, as well as ensuring that valuable in-stream environmental
flows were maintained during the drought period. According to Will Fargher (2014),
General Manager of the Water Markets Group in the National Water Commission:
“While the drought has had a significant economic and social impact on the irrigated
agricultural sector, the ability to trade water means that remaining irrigation busi-
nesses are in much better shape than they otherwise would be. The existence of
water markets makes for a more resilient and responsive irrigation sector which is
better placed to adapt to seasonal conditions and other pressures for change.”

5.4 Interactions Between Irrigation and Land Use:

Application of SIMPLE

In addition to shaping future patterns of production, irrigation—and potential limita-

tions on the expansion of irrigated areas, will have important impacts on future pat-
terns of cropland expansion. Taheripour, Hertel et al. (2013) illustrate this point in
the context of their analysis of the land use impacts of biofuel expansion (there is
more on this topic in Chap. 8), however, this analysis is relevant for any source of
demand growth. They begin by estimating the pattern and magnitude of cropland
expansion resulting from the increased corn demand caused by boosting U.S. corn
ethanol production from observed levels in 2001 (very low) to 15 billion gallons/year
as mandated in the U.S. Renewable Fuel Standard. They find that, in the absence of
irrigation constraints, this requires an additional 3.75 M ha of cropland, worldwide.
64 5 Water, Food and Environmental Security

Fig. 5.3 Optimal input use in irrigated (left) and rainfed (right) crop production systems

However, when Taheripour, Hertel et al. (2013) factor in likely constraints on

future irrigation expansion, preventing growth in irrigated area in those regions
where water is deemed ‘physically scarce’ by the International Water Management
Institute, they find that 4.43 M ha of additional cropland is required. This is due to the
fact that yields are higher, on average, in irrigated areas, so curbing expansion there
requires greater expansion elsewhere. But this is not the end of the story. Since rain-
fed areas are, on average, richer in both above- and below-ground carbon, expansion
of rainfed cropland tends to result in greater CO2 emissions per hectare converted. As
a consequence, the authors find that the annualized indirect land use change emis-
sions jump by 28 %, from 35.6 to 45.4 g CO2e per mega-joule of ethanol capacity.
The SIMPLE model can be used to illustrate this same point. In this case, we
modify the model to allow for the separation of irrigated and rainfed areas using pro-
duction and crop land share information taken from Taheripour, Hertel et al. (2013).
Therefore, each of these cropping systems has a distinct production function in each
of the 15 model regions. The differences in optimal input use in irrigated and rainfed
crop production systems within SIMPLE is illustrated in Fig. 5.3. The right-hand
panel depicts input use under rainfed areas wherein the crop land input is high,
relative to non-land inputs, indicating extensive crop production and low yields.
While the isoquants and elasticity of substitution between land and non-land inputs is
the same in both cases, the initial level of yields, and therefore returns to land (as
reflected by the flatter slope in the isocost line), is higher in irrigated agriculture (left-hand
panel) due to the fact that we include returns to irrigation with the returns to land.
We evaluate the implications of hypothetical constraints on expansion of irri-
gated crop production systems in the context of the historical experience from 1961
to 2006. We take historical growth rates in population, incomes, and total factor
productivity as given—following the historical simulations presented in Chaps. 2
and 3. We consider three alternative irrigation scenarios: the historical baseline and
alternative, ‘counterfactual’ scenarios. The latter assume that irrigated area cannot
expand. In the first alternative scenario, irrigated production systems can still boost
yields through further intensification of production (e.g., introducing drip irrigation,
5.4 Interactions Between Irrigation and Land Use: Application of SIMPLE 65

Fig. 5.4 Composition of global crop production and cropland use under alternative irrigation sce-
narios: 1961–2006

applying additional labor and variable inputs), whereas in the final scenario, not
only is irrigated area fixed, but also farmers in the irrigated cropping systems cannot
boost yields by applying additional nonland inputs on the existing irrigated crop-
land. The results of the simulations are summarized in Fig. 5.4 wherein we report
the updated values of crop production and cropland use for year 2006.
In the historical baseline, irrigated crop production contributes roughly 40 % of
global crop production and 20 % of global land use in 2006. Constraining irrigated
area (first counterfactual simulation) results in surprisingly modest impacts on
global crop land use, with irrigated area falling by around 43 M ha while rainfed
area increases by just 8 M ha relative to the historical baseline. Limiting irrigated
areas has little effect on global crop production since producers under these systems
can instead intensify crop production in response to the ensuing crop price rise,
thereby increasing yields. However, if we also limit the ability of irrigated crop
producers to boost yields on their land, then we see increased pressure for low-yield
rainfed systems to expand production and land use. Rainfed cropland now expands
by 84 M ha, relative to the baseline. There are also significant impacts on crop price
in the latter scenario. When irrigated area alone is restricted, crop prices increase by
just 1.2 %, relative to baseline. However, when both area and yields are constrained,
prices in 2006 are 14.5 % above their baseline level. These results highlight the
importance of both area and yield response in irrigated production systems for mod-
erating the expansion of rainfed croplands and lowering food prices.
In summary, expansion of irrigated area has been a key element of the important
boost in yields and cropping intensity over the past five decades (Bruinsma, 2009).
If this option is limited in the future due to water scarcity, this will require greater
amounts of land conversion, and will push cropland expansion into more carbon-
rich, and likely ecologically sensitive, areas of the world. This highlights one of the
key sustainability challenges of the coming decades.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.
66 5 Water, Food and Environmental Security

References

Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050 (the 2012 revi-
sion) (ESA Working Paper No. 12 - 03). Rome, Italy: FAO. Retrieved from http://www.fao.org/
fileadmin/templates/esa/Global_persepctives/world_ag_2030_50_2012_rev.pdf.
Bjornlund, H., & McKay, J. (2002). Aspects of water markets for developing countries: experi-
ences from Australia, Chile, and the U.S.. Environment and Development Economics, 7(04),
769–795. http://doi.org/10.1017/S1355770X02000463.
Bruinsma, J. (2009). The resource outlook to 2050. By how much do land, water use and crop
yields need to increase by 2050? In FAO Expert meeting on how to feed the world in 2050.
Rome, Italy: Food and Agriculture Organisation of the UN.
Burke, J., Villholth, K. (2007). Groundwater: a global assessment of scale and significance, in:
Molden, D. (Ed.), Water for Food, Water for Life. Earthscan and International Water
Management Institute, London and Colombo, pp. 395–423.
Chao, B.F., Wu, Y.H., Li, Y.S., (2008). Impact of Artificial Reservoir Water Impoundment on
Global Sea Level. Science 320, 212–214. doi:10.1126/science.1154580.
Craig, Ian, Andrew Green, Michael Scobie, and Erik Schmidt. (2005). “Controlling Evaporation
Loss from Water Storages.” Report 1000580/1. Toowoomba, Australia: University of Southern
Queensland, National Centre for Engineering in Agriculture. http://www.ncea.org.au/
Evaporation%20Resources/index_files/Page1668.htm.
De Fraiture, C., Wichelns, D., Rockstrom, J., Kemp-Benedict, E., Eriyagama, N., Gordon, L. J., …
Karlberg, L. (2007). Looking ahead to 2050: scenarios of alternative investment approaches.
Retrieved from http://agris.fao.org/agris-search/search.do?recordID=QL2012001800.
Döll, P., Hoffmann-Dobrev, H., Portmann, F. T., Siebert, S., Eicker, A., Rodell, M., Strassberg, G.,
Scanlon, B. R. (2012). Impact of water withdrawals from groundwater and surface water on
continental water storage variations. J. Geodyn. 59–60, 143–156. doi:10.1016/j.jog.2011.05.001.
Eshel, G., Shepon, A., Makov, T., & Milo, R. (2014). Land, irrigation water, greenhouse gas, and
reactive nitrogen burdens of meat, eggs, and dairy production in the United States. Proceedings
of the National Academy of Sciences of the United States of America, 111(33), 11996–12001.
http://doi.org/10.1073/pnas.1402183111.
Fargher, W. (2014). Responding to scarcity: Lessons from Australian water markets in supporting
agricultural productivity during drought. OECD. Retrieved from http://www.oecd.org/tad/
sustainable-agriculture/49192129.pdf.
Füssel, H.- M., Heinke, J., Popp, A., Gerten, D. (2012). Climate Change and Water Supply, in:
Edenhofer, O., Wallacher, J., Lotze-Campen, H., Reder, M., Knopf, B., Müller, J. (Eds.),
Climate Change, Justice and Sustainability. Springer Netherlands, pp. 19–32.
Grafton, R. Q., Pittock, J., Davis, R., Williams, J., Fu, G., Warburton, M., … Quiggin, J. (2013).
Global insights into water resources, climate change and governance. Nature Climate Change,
3(4), 315–321. http://doi.org/10.1038/nclimate1746.
Hertel, T. W., & Liu, J. (2014). Implications of water scarcity for economic growth. Paris: OECD
Working Paper. ENV/EPOC 17.
Helfer, Fernanda, Charles Lemckert, and Hong Zhang. (2012). “Impacts of Climate Change on
Temperature and Evaporation from a Large Reservoir in Australia.” Journal of Hydrology 475
(December): 365–78. doi:10.1016/j.jhydrol.2012.10.008.
Keller, A., & Seckler, D. (2008, July). Transpiration: Constraints on increasing the productivity of
water in crop production. Winrock Water Forum.
Liu, J., Hertel, T., Taheripour, F., Zhu, T., & Ringler, C. (2014). International trade buffers the
impact of future irrigation shortfalls. Global Environmental Change, 29, 22–31.
Mekonnen, M. M., & Hoekstra, A. Y. (2012). A global assessment of the water footprint of farm
animal products. Ecosystems, 15(3), 401–415. http://doi.org/10.1007/s10021-011-9517-8.
Milly, P. C. D., Dunne, K. A., Vecchia, A. V. (2005). Global pattern of trends in streamflow and
water availability in a changing climate. Nature 438, 347–350. doi:10.1038/nature04312.
References 67

Molden, D. (2007). Water for food, water for life: A comprehensive assessment of water management
in agriculture. London, UK: Earthscan.
National Water Commission. (2011). Water markets in Australia: a short history. Canberra,
Australia: NWC.
Pavelic, P., Patankar, U., Acharya, S., Jella, K., Gumma, M. K. (2012). Role of groundwater in
buffering irrigation production against climate variability at the basin scale in South-West
India. Agric. Water Manag. 103, 78–87. doi:10.1016/j.agwat.2011.10.019.
Rockström, J., Karlberg, L., Wani, S. P., Barron, J., Hatibu, N., Oweis, T., … Qiang, Z. (2010).
Managing water in rainfed agriculture—The need for a paradigm shift. Agricultural Water
Management, 97(4), 543–550. http://doi.org/10.1016/j.agwat.2009.09.009.
Rosegrant, M. W., Cai, X., (2002). Global Water Demand and Supply Projections. Water Int. 27,
170–182. doi:10.1080/02508060208686990.
Rosegrant, M. W., Ringler, C., Zhu, T., Tokgoz, S., & Bhandary, P. (2013). Water and food in the
global bioeconomy: Challenges and opportunities for development. Agricultural Economics,
44(s1), 139–150.
Siebert, S., Burke, J., Faures, J. M., Frenken, K., Hoogeveen, J., Döll, P., Portmann, F. T., (2010).
Groundwater use for irrigation – a global inventory. Hydrol Earth Syst Sci Discuss 7, 3977–4021.
doi:10.5194/hessd-7-3977-2010.
Strzepek, K., Boehlert, B., (2010). Competition for water for the food system.
Taheripour, F., Hertel, T. W., & Liu, J. (2013). The role of irrigation in determining the global land
use impacts of biofuels. Energy, Sustainability and Society, 3(1), 1–18.
Taheripour, F., Hurt, C., & Tyner, W. E. (2013). Livestock industry in transition: Economic, demo-
graphic, and biofuel drivers. Animal Frontiers, 3(2), 38–46. http://doi.org/10.2527/af.2013-0013.
Taylor, R. G., Scanlon, B., Döll, P., Rodell, M., van Beek, R., Wada, Y., Longuevergne, L., Leblanc,
M., Famiglietti, J.S., Edmunds, M., Konikow, L., Green, T.R., Chen, J., Taniguchi, M.,
Bierkens, M. F. P., MacDonald, A., Fan, Y., Maxwell, R. M., Yechieli, Y., Gurdak, J. J., Allen,
D. M., Shamsudduha, M., Hiscock, K., Yeh, P. J. -F., Holman, I., Treidel, H., (2012). Ground
water and climate change. Nat. Clim. Change 3, 322–329. doi:10.1038/nclimate1744.
Wada, Y., Wisser, D., Eisner, S., Flörke, M., Gerten, D., Haddeland, I., Hanasaki, N., Masaki, Y.,
Portmann, F. T., Stacke, T., Tessler, Z., Schewe, J. (2013). Multimodel projections and uncer-
tainties of irrigation water demand under climate change. Geophys. Res. Lett. 40, 4626–4632.
doi:10.1002/grl.50686.
Wurbs, Ralph A., and Rolando A. Ayala. 2014. “Reservoir Evaporation in Texas, USA.” Journal of
Hydrology 510 (March): 1–9. doi:10.1016/j.jhydrol.2013.12.011.
Zwart, S. J., Bastiaanssen, W. G. M., de Fraiture, C., & Molden, D. J. (2010). A global benchmark
map of water productivity for rainfed and irrigated wheat. Agricultural Water Management,
97(10), 1617–1627. http://doi.org/10.1016/j.agwat.2010.05.018.
Chapter 6
Climate Change Impacts in Agriculture

6.1 Overview1

Up to this point, we have abstracted from the potential impacts of climate change on
land use, food security and the environment. In this chapter, we begin by reviewing
the evidence on climate change, its consequences for temperature and precipitation,
and the resulting mechanisms through which it affects crop production. These
include: the impact of temperature on crop development, how elevated CO2 levels
affect plant photosynthesis and respiration, increased water stress for crops due to
higher temperatures and uneven rainfall, extreme temperature damage, and addi-
tional crop damage from pests and disease. We then review the evidence quantifying
some of these impacts historically and use these as a basis for the global simulation
of climate change impacts using the SIMPLE model. In addition to analyzing the
basic impacts on land use and food prices, we also explore potential responses to
climate change such as adaptation through the development of heat and drought
tolerant crop varieties via increased funding for agricultural research. We also show
how successful adaptation to climate change can provide significant mitigation ben-
efits, as improved yields, relative to baseline, lessen the need for land conversion,
which, in turn, reduces further emissions from land use change.
The table has now been set for significant warming of the earth’s surface in the
coming decades. Climate change mitigation policies currently being debated will do
little to alter the expected rate of warming over the next 20–30 years due to the
momentum already in the energy and climate systems. The long-lived, carbon-
intensive energy systems currently in place in the rapidly growing developing econ-
omies of the world, along with continued reliance on expansion of commercial land
uses into carbon-rich natural environments, both serve to ensure that GHG concen-
trations in the atmosphere will rise in the near term. Current estimates suggest that

1
This section draws heavily on Hertel and Lobell (2014).

© Springer International Publishing Switzerland 2016 69

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_6
70 6 Climate Change Impacts in Agriculture

increased radiative forcings will result in temperature increases on the order of

0.3–0.4 °C per decade in most agricultural regions by 2050. Such temperature
increases are likely to threaten agricultural productivity growth—particularly in the
tropics where the bulk of the world’s poor currently reside and find their livelihoods
(IPCC, 2014).
Climate risks, in the form of intra- and inter-season variability, have always pre-
sented a challenge to farmers. Droughts, frosts, floods, heat waves, hail storms, and
other extremes are familiar worries. Indeed, natural climate variability causes so
many losses, and is so high on the list of current concerns to farmers, that many
scholars advocate focusing exclusively on coping with climate variability as a first
step towards dealing with longer-term trends (Cooper et al., 2008; Washington
et al., 2006). Whether or not such a focus is wise, it is clear that (1) current climate
risks are substantial and (2) climate trends will tend to amplify some risks and
reduce others, but is unlikely to create entirely new risks or reduce existing risks to
zero. It follows that any adaptation aimed at addressing future climate trends will
likely have some value in current climate, and conversely any effort to address cur-
rent climate risks will have some value in the context of future climate. A key ques-
tion is which risks are increasing fastest, and therefore which innovations are likely
undervalued if considered solely in the context of current climate variability. For
example, heat stress is widely acknowledged as a current constraint to wheat pro-
duction throughout the developing world (Kosina, Reynolds, Dixon, & Joshi, 2007),
but will become increasingly important as temperatures rise (Asseng, Foster, &
Turner, 2011; Ortiz et al., 2008). All biological processes are influenced by tem-
perature, and therefore the net effects of warming and elevated CO2 on crops are the
result of several separate yet interacting components. In this section, we review the
five most prominent of these components.
Crop development The speed with which crops proceed through different devel-
opment stages, and the resulting duration of total growth, are strongly dependent on
temperature (Parent & Tardieu, 2012). Development rates increase linearly with
temperature across a wide range, typically from 0 to 30 °C, with the exact range
depending on the crop. Because total accumulation of biomass and yield scales with
the duration of the season (as well as key stages like grain filling), the yield effect of
shortened duration from warming is frequently negative in both temperate and trop-
ical systems (Hatfield et al., 2011). The main exceptions occur when faster develop-
ment helps to avoid water stress at the end of season.
Photosynthesis and respiration Plants take up CO2 through the process of photo-
synthesis, and release CO2 during respiration, when the photosynthates are tapped
for energy needed for plant growth and maintenance. Between one-quarter and one-
half of the carbon uptake in photosynthesis is typically released as respiration
(Amthor, 1989). The difference between photosynthesis and respiration determines
the net carbon uptake by a crop. Both processes depend on temperature, with an
optimum temperature for net uptake in full sun varying according to plant type.
Warming can either increase or decrease net carbon uptake, depending on crop type,
the starting temperature relative to optimum, and whether the warming occurs at
6.1 Overview 71

day or night. C3 crops, which include rice, wheat barley and rye, are found in a wide
range of environments. Photorespiration in C3 plants increases with warming and
therefore results in lower optimal temperatures, ranging between 15 and 30 °C. C4
crops, which include maize, sorghum and millet, evolved more recently, converting
atmospheric carbon into a chemical compound with four (instead of three) carbon
atoms. Having originated in sub-tropical areas, C4 plants exhibit a higher optimum
temperature (30–40 °C) which has to do with their lack of photorespiration.
In addition to the direct effects of temperature on photosynthesis and respiration,
higher temperatures increase the saturation vapor pressure of air and, in the absence
of added moisture, will increase the vapor pressure deficit (VPD) between the leaf
and surrounding air (Lobell, Hammer, et al., 2013). Plants respond to higher VPD
by reducing stomatal conductance, leading to a decrease in CO2 flux into the leaf
and subsequent depression of photosynthesis rates. Thus, higher temperatures affect
photosynthesis both directly, via effects of warming on enzymes, and indirectly via
effects on leaf CO2 concentrations.
Elevated CO2 generally leads to an increase in leaf CO2 levels and a reduction in
stomatal conductance. The former leads directly to higher photosynthesis rates in
C3 plants, but typically not in C4 plants which are already saturated with CO2.
Elevated CO2 also increases the optimum temperature for C3 photosynthesis,
because it inhibits photorespiration (Long, 1991). The stomatal effect reduces tran-
spiration losses and leads to higher water use efficiencies in both C3 and C4 plants,
which can result in greater overall yields in dry conditions because plants are able
to reduce losses of soil moisture. However, the stomatal effect also increases canopy
temperature because of lower transpiration rates.
The overall strength of yield response to CO2can be constrained by a lack of nutri-
ents needed for grain biomass, so that regions with low fertilizer inputs—typically in
the tropics—are expected to show moderated responses to CO2 increase. At the same
time, tuber crops with much lower nutrient content in harvested plant parts are able
to respond more strongly to elevated CO2, with potatoes and cassava, for instance,
showing responses well above grain crops (Rosenthal et al., 2012). In general, tropi-
cal systems have a greater proportion of dry conditions and tuber crops, which will
favor CO2 responsiveness, but also have higher proportion of C4 crops, which dimin-
ishes CO2 responsiveness (Leakey, 2009). The net difference in CO2 responses for
aggregate agriculture in the tropical vs. temperate systems remains ambiguous.
Water stress As mentioned previously, higher temperatures increase saturation
pressure of water vapor in the atmosphere. Absolute humidity of the atmosphere is
also expected to increase, mainly due to increased evaporation over oceans, but only
enough to maintain constant relative humidity, with a corresponding increase in
overall Vapor Pressure Deficit (VPD) (Held & Soden, 2006). This higher VPD leads
to higher rates of soil evaporation and plant transpiration, both of which lead to
declines in soil moisture. Even in scenarios of increased rainfall, many regions still
exhibit a decline in soil moisture due to the evaporative changes (Meehl et al.,
2007). A decrease in moisture is significant for crop growth in both temperate and
tropical systems, but is likely more problematic in tropical areas where the length of
72 6 Climate Change Impacts in Agriculture

the viable growing period is determined by soil moisture. For example, projections
in Africa show consistent reductions in the growing period for most countries, with
a reduction of more than 20 % in Southern Africa and the Sahel by the end of the
twenty-first century (Thornton, Jones, Ericksen, & Challinor, 2011).
Because elevated CO2 improves the water use efficiency of plants, this leads to
increases in soil moisture for the same level of biomass production. The net effect
of elevated CO2 and warming on soil moisture and water stress is not yet well
known, and will likely depend on the particular combination of temperature, VPD,
and CO2 concentrations. A recent grassland study found that a CO2 concentration of
600 ppm was enough to completely counteract the moisture decline associated with
day/night warming of 1.5/3.0 °C (Morgan et al., 2011).
Extreme temperature damage Both cold and hot extremes can directly damage
plant cells, leading to severe injury or even death. Several reviews detail the specific
thresholds relevant for various crops (Luo, 2011; Porter & Gawith, 1999). Hot
extremes can be particularly damaging during the flowering period, where they can
irreparably damage reproductive organs and young seed embryos. For example, rice
spikelets exhibit dramatic increases in sterility when exposed to high air tempera-
tures during flowering, with this effect exacerbated under elevated CO2, presumably
because of decreased transpiration rates which contribute to further canopy warm-
ing (Matsui, Namuco, Ziska, & Horie, 1997).
Berry, Roberts, and Schlenker (2013) estimate a spline function which captures
the varying impact of extreme heat over the course of the (March–August) growing
season (maize in the U.S. Corn Belt). Their results show that the impact of extreme
temperatures on yields varies considerably over the course of the growing season,
with the addition of a daily degree day above the 29 °C threshold doing the greatest
damage during the month of July, which is a critical period in crop development. In
contrast, warmer temperatures at the beginning and end of the season can be
beneficial.
Global warming is expected to reduce the incidence of cold extremes and increase
the incidence of hot extremes, and indeed both trends are already clearly observed
in many regions (Alexander et al., 2006; Zwiers, Zhang, & Feng, 2011). Given that
cold extremes cause much more crop damage in temperate than tropical systems,
these temperature trends are more damaging to tropical systems. Indeed, the reduc-
tion of frost constraints in temperate systems presents a lot of adaptation opportuni-
ties that do not exist in tropical areas.
Pest and disease damage A final, but less understood, influence of warming and
increased CO2 concentrations is on major pests, weeds, and diseases. We refer inter-
ested readers to the recent review of Ziska, Blumenthal, Runion, Hunt, and Diaz-
Soltero (2011), while noting simply that both warming and CO2 are likely to affect
these biotic stresses in various ways. For instance, invasive weeds tend to be more
responsive than crops to changes in resource availability, such as elevated CO2.
Reduction in frost frequency will also likely expand the ranges of many important
pests and diseases. For example, Hannukkala et al. (2007) report a steady march
earlier in the calendar of the first observations of potato blight in Finland over the
6.2 Quantifying Climate Impacts on Agriculture 73

past decade. In the early 1990s, the first appearance of this blight was typically
between 60 and 100 days after planting. However, by the early 2000s, observations
as early as 20 days were common, thereby requiring considerably more effort on the
part of producers to deal with this pathogen. Ziska et al. (2011) document the north-
ward shift of the kudzu weed in the U.S. Corn Belt from 1971 to 2006. Overall,
there is little indication yet of whether changes will be more severe in temperate or
tropical systems.

6.2 Quantifying Climate Impacts on Agriculture2

Since this book strives to offer quantitative assessments of the major drivers of
change on global environmental and food security, it is necessary to translate the
preceding mechanisms through which climate affects crop yields into numerical
impact estimates. There are two broad approaches to this problem: crop simulation
modeling and statistical estimation of impacts.
Crop growth simulation models The predominant tool for assessing the impacts
of climate change on agricultural productivity is the biophysical crop growth simu-
lation model. The most widely used variant for climate change analysis is the Crop
Environment Resource Synthesis (CERES) model, as implemented through the
Decision Support System for Agrotechnology Transfer (DSSAT) (Jones et al.,
2003). This is a process-based model with extensive data requirements and explicit
spatial resolution, simulating crop growth as a function of soils, water availability,
temperature, and soil nitrogen dynamics at the level of an individual field. Crop
growth is broken into six phenological stages and leaf and stem growth rates are
calculated for each stage. Management factors are also considered, including choice
of variety, planting date, row spacing, irrigation and nitrogen fertilizer application
amounts and timing. By varying temperature and precipitation inputs, users can
simulate the impact of climate change on agricultural productivity. Results from
DSSAT and the broader family of crop simulation models underpin a number of
important studies of the impact of climate change on crop productivity (Rosenzweig,
Elliot, et al., 2013) and the subsequent impacts on global food security (Nelson
et al., 2010).
Important strengths of crop simulation modeling include: (a) the simulation of
growth by stage so that daily temperature data can be utilized and the impact of
extreme events on yields can be assessed, and (b) the ability of users to specify crop
varieties, fertilizer applications, as well as irrigation availability, all of which are
critical elements in any climate adaptation strategy. However, with few exceptions,
most crop models were initially developed with the explicit goal of aiding field
management decisions, such as cultivar choice, irrigation timing and fertilizer appli-
cation rates. Emphasis in crop modeling was therefore placed on factors like rates

2
This section draws heavily on Hertel and Rosch (2010).
74 6 Climate Change Impacts in Agriculture

of crop development, soil water dynamics, and nutrient supply and demand.
Developers of crop models have long cautioned against their use in climate change
studies, given the lack of development and testing in extreme climate conditions
(White, Hoogenboom, & Hunt, 2005; White, Hoogenboom, Kimball, & Wall,
2011). A review of 221 studies using crop models for climate change impacts, cov-
ering more than 70 different models, found that only six studies considered the
effects of elevated CO2 on canopy temperature, and similarly few studies considered
direct heat effects on seed set or leaf senescence (White et al., 2011).
However, existing crop models are continually being improved, and the AgMIP
project has provided an important vehicle for organizing the community of crop
modelers and structuring model comparison and improvement exercises
(Rosenzweig, Jones, et al., 2013). Four major caveats should be borne in mind, how-
ever, when drawing on crop models to assess the impact of climate change. (1) Only
a subset of relevant processes is included in any single model. For instance, most
models include treatment of crop development and photosynthesis responses to tem-
perature, but omit heat effects on grain set and pest damage. In general these omitted
processes are thought to become more damaging with climate change, so models
may provide estimates biased toward positive values. (2) The ability of models to
correctly predict effects of adaptation is inherently limited to the types of impacts
that are modeled in the first place. For example, effects of warming on crop duration
may be fairly easily addressed by switching to existing longer maturing varieties,
whereas effects on grain set could be more challenging and require development of
new varieties. (3) The types of processes omitted by models tend to be more impor-
tant in tropical than in temperate systems, including effects of high VPD on photo-
synthesis, heat stress on grain set and leaf senescence, and pest and disease pressures.
(4) Despite their use in global scale impact studies, most crop models are highly cali-
brated, field-based analyses which have rarely been validated at national or global
scale. This leads us naturally into the statistical approaches which are typically based
on more aggregate data and therefore more readily validated at aggregate scales.
Statistical approaches An alternative to the crop simulation approach is to esti-
mate statistical relationships between crop yields on the one hand, and temperature
and precipitation on the other. The advantages of this approach are that it requires
relatively less data and it can be readily implemented for large geographic areas
(e.g., nations or continents) which is the scale required for global assessments of
climate impacts on environmental and food security. However the approach is not
process-based, and so it relies on predicting future responses based on past relation-
ships. Thus changes in varieties grown, and other agro-ecological choices, such as
planting and harvesting dates, etc., are not taken into account, thereby limiting the
potential for examination of future adaptation. Path-breaking statistical studies
include the work of Schlenker and Roberts (2009) which identified thresholds for
extreme heat impacts on U.S. field crops and Lobell, Schlenker, and Costa-Roberts
(2011) who examine the impact of recent temperature trends on crop productivity at
global scale. More recent publications take advantage of improved data and
advanced statistical techniques to capture the effect (e.g.) of differential impacts of
extreme heat throughout the growing season (Berry et al., 2013).
6.3 Adaptation to Climate Change 75

All the yield-based studies of agricultural response to climate change assume

that the crops continue to be grown in the same locations, thereby failing to account
for potential adjustments in the mix of agricultural activities at any given location.
The hedonic approach, popularly known as the “Ricardian” approach, was initiated
by Mendelsohn, Nordhaus, and Shaw (1994). It recognizes that farmers will vary
the mix of activities to choose the one yielding the highest return on any given par-
cel of land. As a consequence, they focus on the impact of climate on land values,3
not yields. By associating climatic variation in a cross-section of data with variation
in land values, researchers aim to estimate the long run economic value of climate
and hence the impact of changes in climate, once adaptation to the new climate has
taken place.
The Ricardian approach has been applied in a wide range of circumstances,
including developed and developing countries. Its successful application relies on
two key assumptions. First, there must be an observable, long run equilibrium in
factor markets (especially land). Second, there should be no adjustment costs so that
land rents fully reflect the value of climate at any given location. Given these
assumptions, the hedonic approach typically utilizes cross-section data to estimate
long-run relationships, which are thus sensitive to omitted variable bias (Schlenker,
Hanemann, & Fisher, 2005) and therefore, may lead to false inferences about the
impact of climate change on agriculture.
In wrapping up his survey of a decade of Ricardian analysis in developing coun-
tries, Mendelsohn (2009, pp. 16–17) concludes that:
The studies generally confirm the hypothesis that tropical and subtropical agriculture in
developing countries is more climate sensitive than temperate agriculture. Even marginal
warming causes damages in Africa and Latin America to crops. Crops are also sensitive to
changes in precipitation. In semi-arid locations, increased rainfall is beneficial. However, in
very wet places, increased rainfall can be harmful. If climate scenarios turn out to be rela-
tively hot and dry, they will cause a lot of damage to farms in low latitude countries.
However, if climate scenarios turn out to be relatively mild and wet, there will be only
modest damages and maybe even beneficial effects. The magnitude of the damage depends
greatly on the climate scenario.

6.3 Adaptation to Climate Change4

It is useful to think about adaptation to climate change in two steps (Antle &
Capalbo, 2010). First comes adaptation based on existing technology, and, in the
second step comes adaptation development and use of new technologies. However,

3
The use of land values, as opposed to annual returns, has the additional advantage of embodying
the expectations of returns in a normal year, whereas annual net returns will be influenced to ran-
dom variations in annual production. However, in many developing countries land markets are
insufficiently developed to allow this approach, and so net returns are used.
4
This section draws heavily on Hertel and Lobell (2014).
76 6 Climate Change Impacts in Agriculture

it should be noted that there is a long lag time in development and dissemination of
new technologies (Alston, Pardey, & Ruttan, 2008). Of course the technologies that
are new to the farmer may not be new to the world, but rather simply a technique
which was not profitable in that region prior to climate change.
Adaptation using existing technologies It is informative to consider several
examples of adaptation based on existing technology. Consider, for example, the
challenges posed by increased weeds, pests and pathogens which may result from
climate change. Such an increase will likely require more intensive use of labor for
weeding as well as labor, machinery and chemicals for the application of herbicides
and pesticides. This would, in turn, be expected to reduce profitability, relative to
the pre-climate change scenario. The intensity of irrigation is another important
choice variable for farmers which may be affected by climate change and which
may be viewed as a near term management decision on those farms already equipped
with irrigation. At heightened temperatures, the rate of evapotranspiration rises and
the plant requires more water to maintain normal development. Again, the intensity
of variable input use rises, but and maximum expected value of the farm enterprise
falls once again.
Another example of how variable input use is likely to be autonomously altered
by climate change is motivated by crops’ response to elevated CO2 levels. In order
to translate higher CO2 concentrations into faster plant growth, nutrient availability
may need to be increased. This means a higher rate of fertilizer application is desir-
able. To the extent that their crops are already nutrient-constrained—as is the case
in many developing countries—producers may not be in a position to benefit from
such adaptation. Indeed, there is reason to believe that climate impacts are likely to
be larger, and adaptation more limited, in developing countries. In particular, many
temperate systems—which often correspond to wealthier countries—will likely see
increases in growing season length (defined as time between last and first frost),
whereas many tropical agricultural systems—upon which many of the world’s
poorest people rely—will likely see reductions in growing seasons (defined as the
period with sufficient soil moisture). Temperate farmers will have the option of an
earlier sowing date to escape hot conditions during critical periods such as flower-
ing, as well as adopting longer maturing varieties in order to compensate for faster
rates of crop development. Both of these can be effective at reducing simulated cli-
mate impacts in crop modeling studies, but such options are generally not as attrac-
tive under tropical systems.
Market integration can be an effective tool for asset-poor households to adapt to
climate-driven shocks to agriculture. However, poor households face many barriers
to participating in the very commercial markets which might facilitate adaptation to
weather-induced variability. Poor infrastructure, coupled with remote location,
leaves many farmers—as well as rural consumers—only weakly connected to prod-
uct markets. And the absence of credit markets can render the investment in new
technologies for climate adaptation prohibitive. Similarly, while off-farm work
would be an excellent income diversification option for many climate vulnerable,
rural households, access to the towns where such jobs are available is often costly
6.3 Adaptation to Climate Change 77

and may require temporary migration of the household member employed, thereby
removing their contribution to the farm.
In an interesting historical study of the rainfall and famine in colonial India,
Burgess and Donaldson (2010) find that the arrival of railroads—and hence ready
access to national markets—“dramatically constrained the ability of rainfall shocks
to cause famines in colonial India” (p. 450). To the extent that adverse climate
shocks are not uniformly distributed, there can be significant grains from trade
between grain surplus and grain deficit regions of the country/world (Ahmed,
Diffenbaugh, Hertel, & Martin, 2012). Government policies, too, can frustrate
access to markets. In the wake of the 2007–08 commodity crisis, many countries
imposed export bans on staple grains, thereby exacerbating the world price rise dur-
ing this period and likely throwing additional households around the world into
poverty (Anderson & Nelgen, 2011; Ivanic & Martin, 2008).
The role of new technologies in adaptation In agriculture, introduction of new
technologies has been the norm for at least the past two centuries, and these technolo-
gies have facilitated dramatic shifts in the location and environmental circumstances
under which crops are grown (Olmstead & Rhode, 2011). More recently, technology
improvements have come to be taken for granted, and models of future agricultural
production typically account for new technology by including some exogenous rate
of growth in yields. Of interest here, however, are those particular technologies that
would reduce the sensitivity of agriculture to weather, thereby helping to adapt to
climate change. While new crop seeds are the simplest example of an innovation that
could foster adaptation, we also include in this category agronomic innovations, such
as new methods of water harvesting or conservation agriculture.
Among the obvious candidates are innovations that improve production under
extreme heat and drought conditions. However, technologies that help improve cold
tolerance could also be beneficial, as they would allow faster transition of crops
northward into cooler locations. Technologies that facilitate earlier sowing, such as
those witnessed in the United States over the past few decades (Kucharik, 2006),
could also help to avoid harmful weather. Seeds with improved pest and disease
resistance could become more valuable if climate change exacerbates biotic stresses.
And cropping systems that are more capable of surviving inundations, such as the
new submergent tolerant rice varieties, will have added value as frequency of heavy
rainfall increases.
It is extremely difficult to predict the potential rate of improvement enabled by
any of these innovations. Some, for instance, argue that achieving drought tolerance
without incurring a significant (and potentially unacceptable, from the farmers’ per-
spective) yield penalty in good years is extremely unlikely (Sinclair, Purcell, &
Sneller, 2004). Indeed, evidence over the past few decades suggests that innovations
in maize production have made that crop more, not less, sensitive to drought (David
B. Lobell et al., 2014). Others remain optimistic about the potential for adaptation
through innovation, but there is widespread agreement that traits such as heat or
drought tolerance are unlikely to be improved very quickly.
Even for relatively straightforward improvements, the historical record under-
scores the substantial time lags associated with returns on agricultural investments,
78 6 Climate Change Impacts in Agriculture

with benefits peaking an average of 20 years after the initiation of the research (Alston
et al., 2008). Major innovations often take longer. Hybrid corn research started 59
years before release of the first variety, and Bt corn research started 96 years before
its release in 1997 (Alston, James, Andersen, & Pardey, 2010). New genetic tech-
niques are almost certain to speed things up, yet at the same time many of the traits
desirable for climate adaptation are complex and even modest gains are difficult. The
recent efforts toward releasing drought tolerant maize in the United States, for exam-
ple, have been characterized by companies trying hard to manage expectations.
Once technologies are developed, there are additional lags in their adoption. This
can be particularly true in the case of heat- or drought-tolerant seeds, which, unlike
herbicide or pest resistance, typically only exhibit clear benefits in years of moder-
ate stress. This feature can markedly slow adoption as farmers are not easily con-
vinced that the benefits outweigh the costs (Lybbert & Bell, 2010).
Institutionally, the capacity for research is generally higher in developed coun-
tries. Indeed, the large gap in research capacity of developed and middle-income
countries, on the one hand, and the poorest countries on the other, is large and grow-
ing, with the United States alone spending roughly five times the total for all of
Sub-Saharan Africa on public and private agricultural research and development
($10 B vs. $2 B in 2006 USD) per year (Pardey, 2006). Half of the 22 % jump in
global agricultural R&D spending from 2000 to 2008 was fueled by India and
China. In Latin America, 86 % of the region’s R&D spending growth is accounted
for by just three countries: Argentina, Brazil and Mexico (GFAR, 2011). Some
national research systems are simply too small and funding is too volatile to permit
them to retain high quality staff and reach the critical mass needed to effectively
address farmers’ scientific and technological needs. In Sub-Saharan Africa, research
funding is driven by large, donor-funded initiatives, which result in considerable
year-to-year volatility. Indeed, R&D spending volatility in this region is twice as
high as in Asia and Latin America, and Burkina Faso, Gabon, Mauritania and
Tanzania recorded spending volatility nearly four times as high as these other
regions (GFAR, 2011). All of these factors will result in considerable heterogeneity
of planned adaptation responses to climate change, with the poorest countries likely
experiencing the greatest challenges.
Perhaps the most important difference between rich and poor countries’ capabil-
ity to adapt to climate change lies in the ability of farmers to take on the risk associ-
ated with new technologies. Adaptation will be an investment decision, with
uncertainty associated with the costs and benefits of this investment. Many farm
level investments are irreversible, and bring with them a stiff penalty for being
wrong since the farmer cannot readily ‘undo’ their decision and recoup the costs
expended. Antle and Capalbo (2010) discuss the impact of such uncertainty on pro-
ducers’ investment decisions, emphasizing the fact that, under these circumstances,
it is often optimal to postpone making a decision until the uncertainty is further
resolved. In developed countries, farmers have many forms of insurance that they
can turn to if new technology fails. In developing countries, an investment that does
not pay off in the first year could be disastrous to family income and assets. This is
further exacerbated by a lack of access to credit. The inability to take such risks
6.4 Analysis of Climate Impacts and Adaptation Using SIMPLE 79

characterizes much of tropical cropping systems, and, for example, helps to explain
the relatively low use of fertilizer inputs.

6.4 Analysis of Climate Impacts and Adaptation

Using SIMPLE5

It is useful at this stage to turn to some analysis based on the SIMPLE model of global
crop production to better understand the linkages between climate change, agricul-
tural production, and food and environmental security. Lobell, Baldos and Hertel
(2013) begin with a baseline scenario which incorporates climate impacts. In keeping
with the recent IPCC report (IPCC, 2014), they treat these impacts as modifying the
trend rate of productivity growth. For example, that report suggested that, between
now and 2050, adverse climate impacts could result in the loss of about 1 year of
trend productivity per decade (0–2 % yield loss/decade). Lobell, Baldos and Hertel
(2013) base their impacts scenario on the annualized impacts of climate change esti-
mated by Müller et al. (2010) for the 2010 World Development Report (2010). Müller
et al. report impacts by mid-century both with and without effects of CO2 fertiliza-
tion. Since the focus of this analysis is on adaptation to increased temperatures and
changes in precipitation patterns, they used the climate impacts without CO2 effects.
Lobell, Baldos and Hertel (2013) abstract from these effects, since, in addition to
being highly uncertain, they will likely not be altered by adaptive research. On the
other hand, new varieties of crops and new management practices could be devel-
oped in an effort to neutralize the effect of changing temperature and precipitation.
Table 6.1 presents the shocks applied to the version of SIMPLE model used in
Lobell, Baldos and Hertel (2013) (annualized growth rates over the 2006–2050
period), including the impacts of projected 2050 temperature and precipitation
changes on TFP, which are negative in all regions except Europe—findings that are
broadly consistent with other global assessments in the literature, which anticipate
negative impacts in most developing countries (Nelson et al., 2010). In a second
scenario, Lobell et al. allow all regions to fully adapt to climate change by investing
more in Research and Development. This has the effect of restoring the baseline
TFP to the levels that would have prevailed without temperature and precipitation
changes. Europe, which is the lone region to benefit from climate changes to 2050,
is not adjusted under this scenario, as no planned adaptation is required. The authors
estimate the total amount of investment needed to fully adapt to climate change (i.e.,
restore TFP to levels without climate change) in each region by assuming an elastic-
ity of TFP with respect to investment of 0.3. That is, a 10 % increase in agricultural
investment is assumed to result in a 3 % increase in TFP. This value is taken from
Nelson (2009), where it is based on expert estimates on effects of R&D spending on
agricultural productivity growth. However, most econometric analyses in the litera-
ture give quite similar values for this key parameter (Thirtle, Lin, & Piesse, 2003).

5
This section draws heavily on Lobell, Baldos and Hertel (2013).
80 6 Climate Change Impacts in Agriculture

Table 6.1 Key growth rates from 2006 to 2050 (annual growth rate) used in Lobell, Baldos and
Hertel (2013)

Per capita TFP

Income regions Population income Livestock Processed food
Upper High 0.33 1.22 0.40 0.65
Lower High 1.02 2.17 0.40
Upper Middle 0.53 2.74 0.80
Lower Middle 0.22 5.03 2.20
Low 1.17 4.62 0.30

Built-up land Climate change TFP

Geographic regions expansion TFP shocks Biofuel use Crops
East Asia & Pacific 2.37 −0.41
Europe & Central Asia 1.30 0.04
Latin America & Caribbean 1.99 −0.17
Middle East & North Africa 2.55 −0.32
North America 1.91 −0.15
South Asia 3.32 −0.73
Sub-Saharan Africa 3.76 −0.16
Global 5.40 0.83

Key results for global land use change and associated GHG emissions are shown
in Fig. 6.1. In the baseline scenario, with no adaptation (and no CO2 effect) global
cropland area rises by more than 300 M ha between 2006 and 2051. Most of this
new cropland comes from forests or pasture land. When these lands are converted
to cropland, there is generally carbon released into the atmosphere in the form of
CO2, with the size of this release depending on how much land is converted and
where this conversion occurs. The right hand panel of this figure estimates that this
conversion results in the release of about 90 Gt of CO2 over this baseline.
When all regions adversely affected by climate change successfully implement
planned adaptation to eliminate the adverse impacts of temperature and precipita-
tion changes, productivity growth is higher, baseline yield grows more rapidly and
less land is required in 2050. With less additional land required, there is less crop-
land conversion and therefore less CO2 added to the atmosphere. This emissions
saving is an important potential source of GHG mitigation, and the authors divide
these reduced emissions into the estimated cost of the R&D required for adaptation
in order to come up with a median ‘price’ of abatement of $16.70/tonne CO2 abated.
This compares quite favorably to carbon ‘prices’ in the few regions that have intro-
duced carbon markets, leading the authors to conclude that ‘climate adaptation is an
effective form of mitigation’.
Climate change also has an important impact on food prices and hence consump-
tion and nutritional attainment. In another application of the SIMPLE model, Baldos
and Hertel (2014) focus on these impacts. We will defer discussion of these nutri-
tional impacts to Chap. 10. However, it is useful to examine the impacts on price at
this stage. The first panel of Fig. 6.1 reports the impact of climate change on global
References 81

Fig. 6.1 Simulated changes for 2050 relative to 2006 for (a) global price, (b) cropland area, and
(c) land use emissions change. Bars show median values across all simulations, and error bars
indicate 5th–95th percentile confidence interval. Source: Lobell, Baldos, et al. (2013)

crop prices. In the absence of adaptation (and abstracting from CO2 fertilization),
crop prices rise significantly over this baseline simulation. However, with success-
ful adaptation to higher temperatures and altered precipitation patterns, the price
rise is moderated, and the 95 % confidence interval comes close to including zero
price rise over this 44 year period.
In summary, assessing the potential impacts of climate change on agriculture,
land use, and crop prices, is a complex process. Currently available crop models are
incomplete in their treatment of all of the potential channels through which these
impacts can be felt. Nonetheless, it is informative to examine the potential impacts
of existing studies, using the economic framework employed in this book. In so
doing, we find that the temperature and precipitation components of climate change
will lower global crop yields, thereby requiring additional land conversion.
Successful adaptation to these climate shocks would restore yields to their baseline
levels, thereby reducing land conversion and associated GHG emissions and mod-
erating crop prices over the coming decades. In this sense, adaptation to climate
change can prove to be an effective form of climate change mitigation as well.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Ahmed, S. A., Diffenbaugh, N. S., Hertel, T. W., & Martin, W. J. (2012). Agriculture and trade
opportunities for Tanzania: Past volatility and future climate change. Review of Development
Economics, 16(3), 429–447. http://doi.org/10.1111/j.1467-9361.2012.00672.x.
Alexander, L. V., Zhang, X., Peterson, T. C., Caesar, J., Gleason, B., Klein Tank, A., … Rahimzadeh,
F. (2006). Global observed changes in daily climate extremes of temperature and precipitation.
Journal of Geophysical Research, 111, D05109. doi:10.1029/2005JD006290.
82 6 Climate Change Impacts in Agriculture

Alston, J. M., James, J. S., Andersen, M. A., & Pardey, P. G. (2010). Research lags and spillovers.
In J. M. Alston, J. S. James, M. A. Andersen, & P. G. Pardey (Eds.), Persistence pays (pp. 239–
269). New York, NY: Springer.
Alston, J. M., Pardey, P. G., & Ruttan, V. W. (2008). Research lags revisited: Concepts and evi-
dence from U.S. Agriculture (Staff Paper No. 50091). University of Minnesota, Department of
Applied Economics. Retrieved from http://ideas.repec.org/p/ags/umaesp/50091.html.
Amthor, J. S. (1989). Respiration and crop productivity. New York, NY: Springer.
Anderson, K., & Nelgen, S. (2011). Impact on world prices of many countries using trade policy
to stabilize domestic food prices. Presented at the IATRC theme day on agricultural price vola-
tility, Trade Policy and Food Security, St. Petersburg, FL.
Antle, J. M., & Capalbo, S. M. (2010). Adaptation of agricultural and food systems to climate
change: An economic and policy perspective. Applied Economic Perspectives and Policy,
32(3), 386–416. http://aepp.oxfordjournals.org.
Asseng, S., Foster, I., & Turner, N. C. (2011). The impact of temperature variability on wheat
yields. Global Change Biology, 17, 997–1012.
Baldos, U. L. C., and T. W. Hertel. 2014. Global Food Security in 2050: The Role of Agricultural
Productivity and Climate Change. Australian Journal of Agricultural and Resource Economics.
doi:10.1111/1467-8489.12048.
Berry, S. T., Roberts, M. J., & Schlenker, W. (2013). Corn production shocks in 2012 and beyond:
Implications for harvest volatility. NBER, pp. 59–81.
Burgess, R., & Donaldson, D. (2010). Can openness mitigate the effects of weather shocks?
Evidence from India’s famine era. American Economic Review, 100(2), 449–453. http://doi.
org/10.1257/aer.100.2.449.
Cooper, P. J. M., Dimes, J., Rao, K. P. C., Shapiro, B., Shiferaw, B., & Twomlow, S. (2008). Coping
better with current climatic variability in the rain-fed farming systems of sub-Saharan Africa:
An essential first step in adapting to future climate change? Agriculture, Ecosystems &
Environment, 126(1–2), 24–35.
Global Forum on Agricultural Research, G. F. on A. R. (2011, September 30). ASTI global assess-
ment of agricultural R Spending: Developing countries accelerate investment (IFPRI) | EGFAR
[Text.Homepage]. Retrieved May 19, 2013, from http://www.egfar.org/bookmarks/
asti-global-assessment-agricultural-rd-spending-developing-countries-accelerate-investment.
Hannukkala, A. O., T. Kaukoranta, A. Lehtinen, and A. Rahkonen. 2007. Late-Blight Epidemics
on Potato in Finland, 1933–2002; Increased and Earlier Occurrence of Epidemics Associated
with Climate Change and Lack of Rotation.” Plant Pathology 56 (1): 167–76. doi:10.1111/
j.1365-3059.2006.01451.x.
Hatfield, J. L., Boote, K. J., Kimball, B. A., Ziska, L. H., Izaurralde, R. C., Ort, D., … Wolfe, D.
(2011). Climate impacts on agriculture: Implications for crop production. Agronomy Journal,
103, 351–370. http://doi.org/doi:10.2134/agronj2010.0303.
Held, I. M., & Soden, B. J. (2006). Robust responses of the hydrological cycle to global warming.
Journal of Climate, 19(21), 5686–5699.
Hertel, T. W., & Lobell, D. B. (2014). Agricultural adaptation to climate change in rich and poor
countries: Current modeling practice and potential for empirical contributions. Energy
Economics, 46, 562–575. http://doi.org/10.1016/j.eneco.2014.04.014.
Hertel, T. W., & Rosch, S. D. (2010). Climate change, agriculture, and poverty. Applied Economic
Perspectives and Policy, 32(3), 355–385. http://doi.org/10.1093/aepp/ppq016.
IPCC. (2014). Climate change 2014: Impacts, adaptation, and vulnerability. In Contribution of
Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge, UK: Cambridge University Press.
Ivanic, M., & Martin, W. (2008). Implications of higher global food prices for poverty in low-
income countries. Agricultural Economics, 39, 405–416. http://www.blackwellpublishing.
com/journal.asp?ref=0169-5150&site=1.
Jones, J. W., Hoogenboom, G., Porter, C. H., Boote, K. J., Batchelor, W. D., Hunt, L. A., …
Ritchie, J. T. (2003). The DSSAT cropping system model. European Journal of Agronomy,
18(3–4), 235–265. http://doi.org/10.1016/S1161-0301(02)00107-7.
References 83

Kosina, P., Reynolds, M., Dixon, J., & Joshi, A. (2007). Stakeholder perception of wheat production
constraints, capacity building needs, and research partnerships in developing countries.
Euphytica, 157(3), 475–483.
Kucharik, C. J. (2006). A multidecadal trend of earlier corn planting in the Central U.S.A.
Agronomy Journal, 98(6), 1544–1550. http://doi.org/10.2134/agronj2006.0156.
Leakey, A. D. B. (2009). Rising atmospheric carbon dioxide concentration and the future of C4
crops for food and fuel. Proceedings of the Royal Society B: Biological Sciences. doi:10.1098/
rspb.2008.1517.
Lobell, D. B., Baldos, U. L. C., & Hertel, T. W. (2013). Climate adaptation as mitigation: The case
of agricultural investments. Environmental Research Letters, 8, 15–28. http://doi.
org/10.1088/1748-9326/8/1/015012.
Lobell, D. B., Hammer, G. L., McLean, G., Messina, C., Roberts, M. J., & Schlenker, W. (2013).
The critical role of extreme heat for maize production in the United States. Nature Climate
Change, 3(5), 497–501. http://doi.org/10.1038/nclimate1832.
Lobell, D. B., Roberts, M. J., Schlenker, W., Braun, N., Little, B. B., Rejesus, R. M., & Hammer,
G. L. (2014). Greater sensitivity to drought accompanies maize yield increase in the
U.S. Midwest. Science, 344(6183), 516–519. http://doi.org/10.1126/science.1251423.
Lobell, D. B., Schlenker, W. S., & Costa-Roberts, J. (2011). Climate trends and global crop pro-
duction since 1980. Science, 333(6042), 616–620.
Long, S. (1991). Modification of the response of photosynthetic productivity to rising temperature
by atmospheric CO2 concentrations: Has its importance been underestimated? Plant, Cell &
Environment, 14(8), 729–739.
Luo, Q. (2011). Temperature thresholds and crop production: A review. Climatic Change, 109(3),
583–598.
Lybbert, T. J., & Bell, A. (2010). Stochastic benefit streams, learning, and technology diffusion:
Why drought tolerance is not the new Bt. Agbioforum, 13(1), 13–24.
Matsui, T., Namuco, O. S., Ziska, L. H., & Horie, T. (1997). Effects of high temperature and CO2
concentration on spikelet sterility in indica rice. Field Crops Research, 51(3), 213–219.
Meehl, G. A., Stocker, T. F., Collins, W. D., Friedlingstein, P., Gaye, A. T., Gregory, J. M., …
Zhao, Z. -C. (2007). Global climate projections. In Climate Change 2007: The Physical Science
Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge, UK: Cambridge University Press.
Mendelsohn, R. (2009). The impact of climate change on agriculture in developing countries.
Journal of Natural Resources Policy Research, 1(1), 5–19. http://doi.
org/10.1080/19390450802495882.
Mendelsohn, R., Nordhaus, W. D., & Shaw, D. (1994). The impact of global warming on agricul-
ture: A Ricardian analysis. The American Economic Review, 84(4), 753–771.
Morgan, J. A., LeCain, D. R., Pendall, E., Blumenthal, D. M., Kimball, B. A., Carrillo, Y., … West,
M. (2011). C4 grasses prosper as carbon dioxide eliminates desiccation in warmed semi-arid
grassland. Nature, 476(7359), 202–205.
Müller, C., Bondeau, A., Popp, A., Waha, P., & Fader, M. (2010). Climate change impacts on
agricultural yields: Background note to the World Development Report 2010 (Background
Note). Potsdam, Germany: Potsdam Institute for Climate Impact Research.
Nelson, G. C., & International Food Policy Research Institute. (2009). Climate change: Impact on
agriculture and costs of adaptation. Washington, DC: International Food Policy Research
Institute.
Nelson, G. C., Rosegrant, M. W., Palazzo, A., Gray, I., Ingersoll, C., Robertson, R., … You, L.
(2010). Food security, farming, and climate change to 2050: Scenarios, results, policy options
(Research reports no. G. C. Nelson, et al.). International Food Policy Research Institute
(IFPRI). Retrieved from http://ideas.repec.org/p/fpr/resrep/geraldnelson.html.
Olmstead, A. L., & Rhode, P. W. (2011). Adapting North American wheat production to climatic
challenges, 1839–2009. Proceedings of the National Academy of Sciences of the United States
of America, 108(2), 480–485. http://doi.org/10.1073/pnas.1008279108.
84 6 Climate Change Impacts in Agriculture

Ortiz, R., Sayre, K. D., Govaerts, B., Gupta, R., Subbarao, G. V., Ban, T., … Reynolds, M. (2008).
Climate change: Can wheat beat the heat? Agriculture, Ecosystems & Environment, 126(1–2),
46–58.
Pardey, P. G. (2006). Agricultural research: A growing global divide? Washington, DC:
International Food Policy Research Institute.
Parent, B., & Tardieu, F. (2012). Temperature responses of developmental processes have not been
affected by breeding in different ecological areas for 17 crop species. New Phytologist, 194(3),
760–774.
Porter, J. R., & Gawith, M. (1999). Temperatures and the growth and development of wheat: A
review. European Journal of Agronomy, 10(1), 23–36.
Rosenthal, D. M., Slattery, R. A., Miller, R. E., Grennan, A. K., Cavagnaro, T. R., Fauquet, C. M.,
… Ort, D. R. (2012). Cassava about-FACE: Greater than expected yield stimulation of cassava
(Manihot esculenta) by future CO2 levels. Global Change Biology, 18(8), 2661–2675. http://
doi.org/10.1111/j.1365-2486.2012.02726.x.
Rosenzweig, C., Elliott, J., Deryng, D., Ruane, A., Arneth, A., Boote, K., … Jones, J. (2013).
Assessing agricultural risks of climate change in the 21st century in a global gridded crop
model intercomparison. Proceedings of the National Academy of Sciences of the United States
of America, 111(9), 3268–3273.
Rosenzweig, C., Jones, J. W., Hatfield, J. L., Ruane, A. C., Boote, K. J., Thorburn, P., … Winter,
J. M. (2013). The Agricultural Model Intercomparison and Improvement Project (AgMIP):
Protocols and pilot studies. Agricultural and Forest Meteorology, 170, 166–182. http://doi.
org/10.1016/j.agrformet.2012.09.011.
Schlenker, W., Hanemann, W. M., & Fisher, A. C. (2005). Will U.S. Agriculture really benefit from
global warming? Accounting for irrigation in the hedonic approach. The American Economic
Review, 95(1), 395–406.
Schlenker, W., & Roberts, M. J. (2009). Nonlinear temperature effects indicate severe damages to
U.S. crop yields under climate change. Proceedings of the National Academy of Sciences of the
United States of America, 106(37), 15594–15598. http://doi.org/10.1073/pnas.0906865106.
Sinclair, T. R., Purcell, L. C., & Sneller, C. H. (2004). Crop transformation and the challenge to
increase yield potential. Trends in Plant Science, 9(2), 70–75.
Thirtle, C., Lin, L., & Piesse, J. (2003). The impact of research-led agricultural productivity growth
on poverty reduction in Africa, Asia and Latin America. World Development, 31(12), 1959–
1975. http://doi.org/10.1016/j.worlddev.2003.07.001.
Thornton, P. K., Jones, P. G., Ericksen, P. J., & Challinor, A. J. (2011). Agriculture and food systems
in sub-Saharan Africa in a 4°C+ world. Philosophical Transactions of the Royal Society A:
Mathematical, Physical and Engineering Sciences, 369(1934), 117–136. http://doi.org/10.1098/
rsta.2010.0246.
Washington, R., Harrison, M., Conway, D., Black, E., Challinor, A., Grimes, D., … Todd, M.
(2006). African climate change: Taking the shorter route. Bulletin of the American
Meteorological Society, 87(10), 1355–1366.
White, J. W., Hoogenboom, G., & Hunt, L. A. (2005). A structured procedure for assessing how
crop models respond to temperature. Agronomy Journal, 97(2), 426–439.
White, J. W., Hoogenboom, G., Kimball, B. A., & Wall, G. W. (2011). Methodologies for simulat-
ing impacts of climate change on crop production. Field Crops Research, 124(3), 357–368.
http://doi.org/10.1016/j.fcr.2011.07.001.
World Bank. (2010). Development and climate change (World Development Report, pp. 1–399).
Washington, DC: World Bank.
Ziska, L. H., Blumenthal, D. M., Runion, G. B., Hunt, E. R., & Diaz-Soltero, H. (2011). Invasive
species and climate change: An agronomic perspective. Climatic Change, 105, 13–42.
Zwiers, F. W., Zhang, X., & Feng, Y. (2011). Anthropogenic influence on long return period daily
temperature extremes at regional scales. Journal of Climate, 24(3), 881–892. http://doi.org/10.
1175/2010JCLI3908.1.
Chapter 7
Land-Based Environmental Services

7.1  Externalities and Land-Based Environmental Services

The world’s land resources provide human and natural ecosystems with a wide
range of services, some of which are marketed, such as food, fiber and timber, while
most ecosystem services are not marketed. The economic framework employed
throughout this book focuses on food markets, and is able to shed light on the con-
sequences of changes in the supply and demand for food for regional and global
land use and environmental outcomes. However, there are many other land-based
services for which markets generally do not exist. Yet humans, as well as nature,
benefit from these ecosystem services which include, among others, carbon seques-
tration, flood prevention, water filtration, natural pollination, pest and disease con-
trol, local temperature moderation and biodiversity. This chapter focuses on these
‘environmental’, typically non-market services.
In many cases, benefits from these environmental services accrue to households
and firms external to those individuals making decisions bearing on the provision of
these services. For example, the forest owner making a decision to clear-cut her for-
est likely does not consider the impact on lowland flooding and siltation, availability
of natural pollinators, biodiversity, or global CO2 emissions. These impacts are all
felt primarily by households and firms who are not involved in the decision to cut
down the forest. Furthermore, if the beneficiaries of this flow of environmental ser-
vices are too numerous and diffuse, or if they lack the capability, knowledge and
expertise to organize themselves in time to prevent the deforestation [this is the
market solution to the presence of externalities (Coase, 1960)], then a strong case
can be made for government intervention. Indeed, this is why land is one of the most
regulated resources in most societies. This type of regulation is particularly impor-
tant in the context of projections for global food and environmental security. We
expect that households’ valuation of these ecosystem services will rise over time, as
per capita incomes rise (Jacobsen & Hanley, 2009), and, if these services are not
preserved, future generations may not have the option of restoring them. In this

© Springer International Publishing Switzerland 2016 85

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_7
86 7  Land-Based Environmental Services

chapter we will discuss the main externalities associated with global land use and
consider how their presence might alter our view of global food and environmental
security in 2050. In many cases, the proposed solution to these externalities involves
‘internalizing’ them by creating publicly facilitated markets offering payments to
local landowners who preserve or enhance the flow of desired ecosystem services.

7.2  Carbon Sequestration1

Perhaps the most prominent land-based externality under discussion today is carbon
sequestration. Appropriate management of the world’s soils and forests can contrib-
ute significantly to slowing the rate of CO2 accumulation in the atmosphere. Reilly
et al. (2012) simulate the case where there is perfect pricing of above- and below-­
ground carbon associated with land use. (Perfect pricing implies that those admin-
istering the program and the participants both have complete information about
carbon stocks.) This comes in addition to the pricing of carbon from energy com-
bustion. They estimate a net gain over the twenty-first century of 178 petagrams of
carbon from land-based carbon pricing.2 Golub, Hertel, Lee, Rose, and Sohngen
(2009) estimate that, in the near term, forest carbon sequestration could supply up
to 50 % of the annual flow of globally efficient greenhouse gas (GHG) abatement.
Sohngen (2010) estimates that inclusion of forest carbon sequestration within an
optimal climate policy could reduce the price of carbon associated with this optimal
climate policy by nearly half in 2100. Clearly there is much to be gained by provid-
ing incentives for individual decision makers to modify land cover and land use
practices to reduce deforestation rates and accommodate additional carbon stocks in
the soil, as well as in the form of above ground biomass. However, these studies also
show that the lowest cost forest carbon sequestration options are associated with
avoided deforestation in the tropics—largely in developing countries—where land
tenure is often uncertain and institutions governing land use are weak. What can
realistically be accomplished under such circumstances?
While this field is still in its scientific and policy implementation infancy, there
are a significant number of carbon sequestration initiatives underway presently, and
the number is growing each year. Peters-Stanley, Hamilton, and Yin (2012) offer an
annual snapshot of carbon sequestration projects, obtaining a total sample of 451
projects across multiple survey years. Voluntary projects dominate their survey,
with lesser contributions from the Clean Development Mechanism and other
compliance-­driven contracts. REDD+ projects dominated transacted volumes in
2010, but these were exceeded by afforestation/reforestation projects in 2011. Latin
America shows the highest volume of contracts, followed by North America, with
the volume of projects in Africa jumping by 150 % between 2010 and 2011. The
authors find that the majority of these contracts are with private land owners. Given

 This section and the next draw heavily on Hertel (2013).

1

 A petagram is equal to 1.0E12 kg.

2
7.2 Carbon Sequestration 87

the excess supply of carbon contracts on the market, relative to demand, purchasers
gravitate toward the easier contracts—which tend to be with the privately held
lands. The data on the price of individual contracts by land tenure type suggests that
sequestration on collectively held lands is roughly twice as costly as on privately
held lands (Peters-Stanley et al., 2012).
Unruh (2008) argues that, despite high biophysical potential, land tenure is a criti-
cal barrier to successful carbon sequestration policies—particularly in Africa. He
highlights the social, legal and economic disconnect between statutory land tenure
which applies in theory, and the customary systems which are predominant in practice,
but which are not recognized by formal laws. As a consequence, governments often
“ignore customary tenure systems and regard such areas as part of the public domain,
while at the same time lacking the capacity to enforce such a claim or resolve the
problems that such a claim produces”. He goes on to assert that “deriving functioning
legal and financial institutions and increasing tenure security for small-scale producers
is what international development has been attempting for decades, and the lesson is
that the challenges to achieving these are immense” (Unruh, 2008). This is hardly a
rosy forecast for the future of carbon payments on the world’s poorest continent.
Others are more optimistic about the potential for carbon sequestration contracts
on communal lands. And the empirical evidence suggests that progress has been
made in the establishment of carbon sequestration contracts on communally held
lands. Peters-Stanley et al. (2012) report a strong increase in the number of con-
tracts with collectively held lands from 2010 to 2011. Barbier and Tesfaw (2012)
point out that sequestration is an area where such contracting may work especially
well in Africa, precisely because tree planting shows a sustained commitment to the
land and is therefore one of the ways in which individuals can secure long-run use
rights in the context of communally held land. The TIST project in Kenya has suc-
cessfully exploited this feature of traditional tenure systems to implement contracts
even in the absence of legal property rights.3
Another type of institutional innovation for dealing with collectively managed land
is offered by the Juma Reserve in Brazil (The Economist, 2009), wherein each house-
hold in this indigenous community in the Amazon forest receives a debit card. Once a
month, the forest is monitored for evidence of disturbance. If the forest is not dis-
turbed, then each household gets an additional payment on their debit card. In addition
to lending an incentive for collective monitoring of the forest, this program offers the
potential for poverty alleviation, as the funds go directly to households in a region
where poverty is rife and the fixed payment will be most meaningful for the poorest
households. However, as with most such projects, the situation is more ­complex than
initially meets the eye and implementation has raised many practical problems which
must be overcome for this to be a long term success (“The Carbon Hunters,” 2010).
One of the most important frontiers for tropical forest protection is found in
Indonesia. There, deforestation for timber harvesting, as well as for establishment
of oil palm plantations has occurred in carbon-rich regions of tropical forest. This
resulted in Indonesia placing third behind the U.S. and China on the list of top GHG

 http://www.tist.org/i2/
3
88 7  Land-Based Environmental Services

emitters in the year 2000 (PEACE, 2007). In 2010, Norway pledged $1 billion USD
to support a pilot carbon payment program and implement a 2-year moratorium on
new oil palm and timber concessions in primary forests and peat lands. Busch et al.
(2015) evaluate this program against history, by asking the following question: ‘If
this moratorium had been put in place in 2000, what volume of GHG emissions
would have been avoided over the decade leading up to 2010?’. They place this
estimate at 58 Mt CO2e, or about 8 % of total emissions. This is a significant sav-
ings—particularly when compared to the other voluntary and compliance contracts
initiated over this period (Peters-Stanley et al., 2012).
The most important success story in reducing deforestation rates in recent decades
has been the experience in Brazil from 2004 to 2013. This is discussed at length in
Nepstad et al. (2014) who call attention to the 70 % decline in the Brazilian deforesta-
tion rate over this period. They attribute this decline to a variety of factors, including
improved satellite monitoring and strict enforcement, restriction of government credit
to districts with high deforestation rates, the emergence of environmentally sensitive
supply chains for soybeans and beef, expansion of protected areas, a slowdown in
road construction, and a rapid increase in the productivity of beef grazing land. With
a change in government in Brazil, as well as changing economic circumstances, it is
unclear whether this record will be sustained. However, the avoided deforestation
over the past decade has placed Brazil at the forefront of global climate mitigation
efforts, demonstrating that large scale land-based mitigation is indeed possible.
What effect would a global sequestration policy have on agricultural land use and
food prices? Golub et al. (2013) have simulated the impact of a global forest carbon
sequestration and find that this environmental policy has a particularly strong impact
on agricultural land use and land prices in the tropical, non-Annex I countries. This
feeds through to higher food prices which adversely affect the poor. Hussein, Hertel,
and Golub (2013) delve more deeply into the distributional impacts of the same
global forest carbon sequestration policy. They conclude that most of the benefits of
this policy flow to landowners (either private or collective owners of the land). Since
the poor generally control relatively little land (and when they do, it is often land of
lower value), they are unlikely to benefit greatly from appreciating land values.
Therefore, the predominant impact of global forest carbon sequestration policy on
the poor is likely to be through higher food prices. They find that this results in pov-
erty increases in 11 of their 14 sample countries, with most of the impact being
driven by the forest carbon sequestration in the tropics (Hussein et al., 2013).

7.3  S
 etting aside Lands for Biodiversity and Other
Ecosystem Services4

The global “market” for biodiversity conservation has grown strongly in recent
years, reaching nearly $3 billion USD annually (Madsen, Carroll, & Moore-Brands,
2010). Much like REDD, this market relies on payments to local landowners in

 This section draws heavily on Hertel (2013).

4
7.3 Setting Aside Lands for Biodiversity and Other Ecosystem Services 89

return for preserving biodiversity and other desirable characteristics associated with
natural lands. The encompassing term for these programs is ‘Payments for
Environmental Services’ (PES) and there is now a large literature on these programs
(Gong, Hegde, & Bull, 2013). One of the most successful documented examples is
the ‘Socio Bosque’ program in Ecuador. In addition to targeting biodiversity, it also
seeks to address local issues, including water management and poverty. And Socio
Bosque is also part of Ecuador’s REDD+ program (De Koning et al., 2011). In
short, it is motivated by a whole range of land-based externalities. Two years after
inception, this program encompassed more than 500,000 hectares of natural ecosys-
tems, with payments being made to more than 60,000 individuals (De Koning et al.,
2011). The targeted areas were those which: (a) were close to roads and rivers
(threat of deforestation); (b) were important in water provision to lower catchments;
(c) had a low percentage of protected natural areas (threat of loss of biodiversity);
and (d) had a high index of unmet basic needs (i.e. poverty).
While the theory of PES suggests that payments should vary over space and time
according to the opportunity cost of the land in competing uses, this is often difficult
to achieve in practice. In the case of Socio Bosque, data limitations and the desire for
transparency dictated instead a simple ‘progressive’ structure in which the first 50
hectares enrolled receives 30 USD/hectare/year for the 20-year duration of the con-
tract, the second 50 hectares (i.e. 51–100 hectares) receives 20 USD/hectare/year
and so on. As a result, the bulk of the community payments (80 %), when expressed
on a per household basis, are under 500 USD/year (De Koning et al., 2011). These
community payments are used for investments that address basic needs, as well as
for productive activities such as agriculture and community banking.
However, in order to enroll in Socio Bosque, a formal land title is needed, and
this precluded involvement by some poor households and communities which have
not yet formalized their land ownership (De Koning et al., 2011). In addition, poor
households are poor, in part due to their limited access to land, which in turn limits
their potential participation in this land-based contracting arrangement. The diffi-
culty in reaching the poorest households with PES programs is also evident in the
Mexican program of Payment for Hydrological Environmental Services of Forests
which was launched in 2003. Overall, that program has been reported to: “reach an
important part of the poorest population, but that for reaching the poorest of the
poor, special outreach is needed as this part of the population has less contact with
government institutions” (Muñoz-Piña, Guevara, Torres, & Braña, 2008).
Pagiola, Arcenas, and Platais (2005) evaluate the poverty impacts of payments
for environmental services (PES) in Latin America. They conclude: ‘PES programs
are not a magic bullet for poverty reduction, but there can be important synergies
when program design is well thought out and local conditions are favorable’
(p. 248). Because such payments are tied to land, their distributional impacts are
inherently tied to the distribution of land ownership in the target region. Since rural
land ownership is highly correlated with income, this immediately biases the pro-
grams towards the wealthier households. Also, transactions costs for the program
(e.g., contracting costs, management plans) are largely independent of farm size and
therefore most onerous for small farms. These fixed costs also create an incentive
90 7  Land-Based Environmental Services

for those administering the program to work with larger entities—a classic adverse
selection problem that reduces the poverty-reduction potential of PES. Of course, if
smallholder farmers can organize themselves into co-operatives, some of these
problems may be overcome. Pagiola et al. (2005) also highlight the importance of
land tenure and credit constraints. In the frontier areas where deforestation is most
active, land tenure is often insecure. When coupled with credit constraints, this
makes it very difficult for low-income households to participate in such programs.
As with carbon contracts, PES programs for biodiversity and other environmen-
tal services can have important indirect effects through the market place when
undertaken at large scale. Pelikan, Britz, and Hertel (2015) explore the global envi-
ronmental impacts of the E.U. biodiversity-targeted, agricultural set aside program
which is part of the recent proposal for a reformed Common Agricultural Policy.
They estimate that implementation of such a program would improve the environ-
mental status of the high-yielding regions of the E.U. by removing a significant
amount of land from production. However, the resulting price increases are expected
to trigger intensification of production in the more marginal areas of Europe, with
attendant environmental side-effects. With higher world prices, the authors expect
an additional 400,000 hectares of cropland conversion in the rest of the world,
resulting in GHG emissions of about 20 tonnes CO2e for every hectare of EU agri-
cultural land set aside for biodiversity (Pelikan et al., 2015).
In their reviews of the challenges of implementing national PES systems in devel-
oping countries, Angelsen (2014) and Gong et al. (2014) highlight several points.
The first is the challenge of defining and measuring the service provided. Given the
spatial heterogeneity of ecological systems throughout the tropics, accurate measure-
ment of environmental services can be extremely costly—where do you draw the
line? In this context it is interesting that the Socio Bosque program opted for an
extremely simple system (flat payments), as opposed to a more sophisticated
approach informed by geospatial data on land cover, carbon stocks and measures of
ecological diversity. Angelsen (2014) highlights the challenge of contract design in
light of asymmetric information which gives rise to both moral hazard and adverse
selection problems. More generally, Gong et al. (2014) emphasize the challenge
posed by high transactions costs of PES programs and the need for institutional inno-
vations which can lower these costs and encourage additional participation. Gong
et al. (2014) conclude their review of PES programs by noting that there is a: "desire
to simultaneously obtain a maximum level of environmental benefits, an increase in
economic efficiency and a reduction in inequality is a laudable goal, but project
developers must realize there are trade-offs, tough decisions have to be made.”

7.4  What Role for Wildlife Friendly Farming?

In addition to ecosystem services payments aimed at keeping certain lands out of

farming, there is considerable debate about the potential for altering the way in
which farming is undertaken, with some advocating a shift towards
7.4 What Role for Wildlife Friendly Farming? 91

‘wildlife-­friendly’ farming practices (Teja Tscharntke, 2012). Given need to feed a

growing, ever-richer population, there appears to be a trade-off between adopting
lower-yield, more environmentally friendly farming practices—dubbed ‘land-shar-
ing’, and more intensive, high-yielding approaches—dubbed ‘land-sparing’ (recall
Chap. 4). Furthermore, it is argued that appropriate agro-ecological intensification
of production could achieve both goals (Teja Tscharntke, 2012). Of course, it is an
empirical question whether the world could feed nine billion people with wildlife-
friendly farming practices and any such analysis is likely to be highly specific to
particular locations. The amount of data required to make a comprehensive ecologi-
cal assessment at global scale is truly daunting.
Phalan, Onial, Balmford, and Green (2011) have made a first attempt to quantify
the land-sharing/land-sparing trade-off by undertaking extensive field surveys in
southwest Ghana and northern India. They focus on birds and trees as measures of
ecosystem diversity and their field work allows them to characterize the relationship
between crop yields and bird or tree population density. Their observations lead
them to distinguish four different patterns of species density. In two of these cases,
density rises with crop yields, thereby suggesting these species will be ‘winners’
under future projections of expanding food production. Not surprisingly, tree spe-
cies are almost wholly absent from this category, with the winners being restricted
to some species of birds which flourish in the presence of crop cultivation. In the
case of these winning species, there are two categories: those for which density rises
sharply with initial yield growth, then plateaus—these species favor land sharing, as
they thrive on more cultivated area, and those species for which density rises
strongly only at high yields—thereby favoring land sparing. Their findings suggest
that land sharing is the preferred strategy for most bird species in Ghana. In India,
the picture is mixed, with a significant number of species being favored by the land
sparing approach. Of course nearly all tree species lose out as the intensity of culti-
vation increases. Amongst these ‘losers’, the land-sparing approach is favored due
to the fact that species densities fall sharply with the onset of cultivation, thereafter
flattening out. (As opposed to a gradual decline at low levels of yields which is more
characteristic of bird species.)
Of special interest to us are the projections which Phalan et al. (2011) undertake.
They acknowledge that future population and income growth will boost the demand
for food, and they take 2050 production targets as given, asking the question: Which
cultivation strategy will offer the greatest benefit to bird and tree densities in their
study regions, while meeting future food demand? They measure population size
for each species as a proportion of an ‘all forest’ baseline. Their findings are unam-
biguous: the land-sparing approach (high yields, less area) results in better out-
comes for the majority of bird and tree species in both Ghana and India. And, in the
case of Ghana, the land-sparing approach in 2050 appears to yield a better outcome
for more bird and tree species than is the case today under the current cultivation
strategy. They note that the latter is characterized by relatively low yields and exten-
sive area. We must bear in mind that these findings are site- and species-specific so
we cannot generalize from them to the entire country/continent or globe. However,
they do lend support to the idea that intensive cultivation on less land may be better
for overall biodiversity than extensive cultivation practices.
92 7  Land-Based Environmental Services

While the Phalan et al. (2011) study looks at impacts flowing from farming to the
environment, externalities also flow in the other direction. In many cases, agricul-
ture is the beneficiary of ecosystem services flowing from a healthy, diverse ecosys-
tem, including nutrient cycling, soil improvements, water provision, genetic
diversity, and reduced threat of pests and disease (Zhang, Ricketts, Kremen, Carney,
& Swinton, 2007). One such service which has been carefully evaluated is the avail-
ability of natural pollinators. Nearly 100 of the world’s important fruit, vegetable
and seed crops are dependent upon animal pollination (Klein et al., 2007). Loss of
these ecosystem services could result in a drop in global agricultural production of
between 3 and 8 % of all output, with much larger impacts on cultivated area which
would have to rise to compensate for the loss of pollination services (Aizen,
Garibaldi, Cunningham, & Klein, 2009).
Ricketts, Daily, Ehrlich, and Michener (2004) studied the value of forest-based
pollinators to coffee production in Costa Rica and find that, as a result of access to
natural pollinators, coffee grown within one kilometer of forests shows 20 % higher
yields, as well as significant increases in quality. The economic value of these forest
services is “commensurate with expected revenues from competing land uses and
far exceeds current conservation incentive payments” (Ricketts et al., 2004).
Garibaldi et al. (2013) confirm the value of wild pollinators in 41 cropping systems
across the world, arguing that pollination by managed honey bees supplemented,
rather than substituted for wild pollination services. The value of wild pollinators to
agriculture suggests a different pattern of agricultural cultivation than might other-
wise be the case. This highlights the desirability of interspersing forests amongst the
cropped areas. Overall, study of the interplay between agriculture and natural eco-
systems is still in its infancy, with the necessary data for continental and global scale
assessments still under development.

7.5  E
 conomic Analysis of the Impacts of REDD
on the Food System

It is useful to examine how one of the most important and wide-ranging policies
discussed in this chapter, REDD+, might affect economic scarcity in the global food
system, using the economic modeling tools developed in this book. A first step can
be made by recalling Eq. (1.1) from Chap. 1, which gives an expression for the
percentage growth in cropland use (qL*). This equation includes a term signifying a
backward shift in land supply owing to the emergence of a new source of competing
demand for land, where the percentage reduction in existing land for agriculture, at
current land rental rates, is represented by DLS > 0 :

qL* = [(DAD + DLS - DLD ) / (1 + h AS, I / h AS, E + h AD / h AS, E )] - DLS (7.1)

To recap key points from Chap. 1: the sign of the change in global land use
depends on the footrace between the percentage growth in demand (ΔAD) and sup-
7.5 Economic Analysis of the Impacts of REDD on the Food System 93

ply—as measured by trend improvements in cropland yield—(ΔLD), wherein hold-

ing the supply curve constant ( DL = 0 ), we have: (DA - DL ) > 0 Þ qL > 0 . The
S D D *

magnitude of this change in global cropland use depends critically on the respon-
siveness of producers and consumers to scarcity in the food system, as captured by
the elasticity of crop supply at the intensive margin (ηAS,I), the extensive margin
(ηAS,E), and the price elasticity of demand for crops ( -h A ). If producers respond
D

aggressively to scarcity-induced higher prices, by intensifying their use of fertilizer

and other non-land inputs, i.e. h A >> 0 , then the amount of land conversion
S,I

required by rapid demand growth will be diminished. Similarly, if consumers

respond to higher prices by cutting back on food consumption, then -h A < 0 and
D

this feature also moderates the land use expansion required in the face of strong
growth in population and/or income.
From Eq. (7.1) we can see that, for a non-zero reduction in cropland supply,
DLS > 0 , there will be both a direct and an indirect effect. The direct effect is cap-
tured by the last term in the equation, -DLS < 0 , which serves to reduce the amount
of land in agriculture. In other words, if 100 M ha of land which would have entered
agricultural production is permanently enrolled in a REDD+ program, then this
represents a direct reduction of 100 M ha of farm land. However, this is not the end
of the story, because withdrawal of these lands makes agricultural land, and food,
scarcer, thereby bringing into play the economic margins of response discussed in
detail in Chap. 4. This is the indirect effect and the reason why the supply shift also
appears in the first term on the right hand side of Eq. (7.1).
The most obvious response by producers is to find some other, unrestricted land
which can be farmed, and bring that into production. Indeed, if the net growth in
food demand, (DA - DL ) is unresponsive to price (zero price elasticities of pro-
D D

ducer yield and consumer demand: h A = h A = 0 ), then this is precisely how equi-
S,I D

librium will be restored in the global food economy. 100 M ha of new land (or its
equivalent in the case of varying land productivity) will be found, and the direct and
indirect effects will cancel out, leaving us with the following outcome:

qL* = [(DAD + DLS - DLD ) / 1] - DLS = (DAD - DLD ) (7.2)

In this case, the REDD+ policy will be ineffective at reducing global cropland use.
However, with rising food scarcity, food consumption is likely to decline some-
what, and producers will have an incentive to intensify production. These elements
boost the denominator of the first term in Eq. (7.1). If, for example, all three eco-
nomic margins of response are equal (regardless of their absolute value—it is only
relative values that matter here), then the equilibrium land use change outcome in
the face of a REDD+ policy will involve land use falling. However, this decline is
less than the full amount of the set-aside land:

qL* = [(DAD + DLS - DLD ) / 3] - DLS = [(DAD - DLD ) / 3] - 2 DLS / 3 (7.3)

In this case, the reduction in agricultural land area due to the REDD policy is equal
to two-thirds of the total REDD+ program, or 66 M ha.
94 7  Land-Based Environmental Services

We can implement a REDD+ policy in SIMPLE by shifting the land supply

schedules in each region of the model. For this application, we draw on the work of
Golub et al. (2013) who modeled the impact of a global carbon policy in which
developed countries implement a comprehensive climate mitigation program—tar-
geting fossil fuel emissions as well as non-CO2GHG emissions from industry and
agriculture, in addition to carbon sequestration incentives for forestry. The common
‘price’ of carbon across all mitigation alternatives is USD  27/tonnes CO2e. Results
were reported at the level of individual agro-ecological zones in each of their model
regions. For purposes of our study, we aggregate their cropland cover shifts, DL > 0
S

, to the level of the 15 regions of the SIMPLE model and focus on the equilibrium
change in global cropland area. We consider the cropland cover shifts5 generated
under scenarios A and B in Golub et al. (2013) (Table 7.1, right columns). In the
first scenario (A), GHG emissions taxes are implemented along with forest carbon
sequestration incentives for Annex I regions only (i.e., wealthier countries including
the U.S., E.U., Canada, Japan, etc...). The second scenario builds on the previous
scenario and introduces forest carbon sequestration incentives to Non-Annex I
regions which includes Brazil, India, China, and Sub Saharan Africa.
Aside from these land supply shocks, we also consider future growth in popula-
tion, incomes, biofuels and total factor productivity. Populationgrowth rates are
based on the medium-fertility variant projections from the UN World Population
Prospects (UNPD, 2013) while income growth rates are computed using the future
estimates of real gross domestic product from Fouré, Bénassy-Quéré, and Fontagné
(2013). We also calculate the growth in global biofuel consumption from the
“Current policies” scenario published in the World Energy Outlook (IEA, 2008,
2012). The latter forecasts are based on the results of a detailed world energy model
given exogenous growths in GDP and population as well as assumptions on future
energy prices and technology. With the “Current policies” scenario, all energy poli-
cies for the power and transportation sectors enacted as of mid-2012 are taken into
account in the projections. TFP growth rates for both the crop sector and livestock
sector is based on the global and regional projections from Ludena, Hertel, Preckel,
Foster, and Nin (2007) which are generated under the assumption of gradual conver-
gence in productivity across regions. Lacking global estimates, TFP growth in the
processed food sectors is based on the historical assessment of Griffith, Redding,
and Reenen (2004) for OECD countries.
Following Golub et al. (2013), we limit our projections to a 20-year period run-
ning from 2006 to 2026. We then report the figures relative to the baseline without
these policies. The results at the global level show that these policies have signifi-
cant impacts on global land use and crop production. Under scenario A, world crop
land use and crop production are roughly 38 M ha and 13 M Mt lower than the
baseline without these policies. We observe a larger reduction in cropland and pro-
duction when forest carbon sequestration incentives are applied in all regions

5
 We use Eq. (7.1) to calculate the effective cropland supply shift from the equilibrium changes in
croplands taken from Golub et al. (2013)). We use the estimates of global demand, intensive and
extensive margins in SIMPLE in order to come up with the global scalar needed to calculate the
effective cropland supply shift. The scalar is equal to 1.30 in this model.
Table 7.1  Future growth rates of exogenous variables (in % per annum rates)
Total factor productivity Land supply shifters
Regions Population Per capita income Biofuels Crop Livestock Processed foods Scenario A Scenario B
Eastern Europe −0.42 4.83 1.04 −0.07 −0.05
North Africa 1.06 3.58 −0.30 0.01 −0.06
Sub-Saharan Africa 2.45 3.78 0.43 0.05 −0.96
South America 0.64 2.51 2.64 0.01 −1.59
Australia/New Zealand 1.00 1.75 0.42 −0.19 −0.16
European Union+ 0.09 1.40 0.42 −0.04 −0.02
South Asia 0.83 4.84 1.71 0.00 −0.32
Central America 0.94 2.55 2.64 0.01 −0.32
Southern Africa 0.57 2.76 0.43 0.05 −0.96
Southeast Asia 0.79 3.62 2.38 0.02 −0.09
Canada/U.S. 0.66 1.06 0.42 −0.40 −0.35
China/Mongolia 0.09 5.55 2.38 0.01 −0.39
Middle East 1.02 2.38 −0.30 0.01 −0.06
7.5 Economic Analysis of the Impacts of REDD on the Food System

Japan/Korea −0.22 2.17 0.42 −0.09 −0.34

Central Asia 1.47 5.03 1.04 0.01 0.00
World 5.75 0.94 0.89
Sources: From left to right—UN World Population Prospects (2013), Fouré et al. (2013), IEA (2008, 2012). Ludena et al. (2007) for crops and livestock TFP
growths, Griffith et al. (2004) and Golub et al. (2013), respectively
95
96 7  Land-Based Environmental Services

Fig. 7.1  Regional changes in crop production and crop land use under REDD+ policies from 2006
to 2026 under Scenario A (left: Annex I only) and Scenario B (right: global mitigation)

(­scenario B) (around 221 M ha and 90 M Mt relative to the baseline). With the
REDD+ policies under scenario A, the decline in global crop price over this period
is quite close to the baseline (around −3.9 % vs. −4.3 % over the baseline). However,
the long run decrease in global crop price is almost halved (at −1.9 %) under sce-
nario B, highlighting the adverse impacts of stringent forest conservation policies
on food security.6
Regional changes in crop production and crop land use are summarized in
Fig. 7.1. In general, land use and production in Annex I regions both decline, given
the crop land supply shifts from the REDD+ policies in this region. Starting with
scenario A (left panel), we see that the crop sectors in the U.S. and Canada are
adversely affected with reductions in cropland use and crop output of 36 M ha and
66 M Mt, relative to the baseline. To offset this contraction, crop production in the
Non-Annex I regions will have to expand. We see strong expansion in crop produc-
tion (as well as crop land use) in Non-Annex I regions particularly in South East
Asia, China/Mongolia and Sub-Saharan Africa This, in turn, has an adverse impact
on tropical terrestrial carbon stocks.
If forest carbon sequestration incentives are also introduced in the developing
countries (non-Annex I region), we obtain the results reported under scenario B
(right-hand panel of Fig. 7.1). The non-Annex I regions generally have carbon-rich
forests and strong incentives to respond to REDD+ payments so forest cover expands
and crop land area and production contract in China/Mongolia, South Asia, Sub-
Saharan Africa and South America. In response, some Annex I regions actually
increase crop production, despite their own REDD+ programs, due to the global price
rise. In summary, initiatives aimed to reduce deforestation and forest degradation, as
well as preserve biodiversity, will have implications for global cropland availability
and could potentially influence future trends in crop production and food prices.

6
 Note that we are assuming food markets are perfectly integrated under these scenarios. The price
effects from such shocks will become more evident if we relax this assumption and introduce
distinction between international and domestic markets (see Chap. 11).
References 97

Supplemental learning resources on this chapter are available for free at https://mygeohub.
org/courses/global_change.

References

Aizen, M. A., Garibaldi, L. A., Cunningham, S. A., & Klein, A. M. (2009). How much does agri-
culture depend on pollinators? Lessons from long-term trends in crop production. Annals of
Botany, 103(9), 1579–1588. http://doi.org/10.1093/aob/mcp076.
Angelsen, A. (2013). The economics of REDD+. In Handbook in forest economics.
Angelsen, A. (2014). The Economics of REDD+. In Handbook in Forest Economics. S. Kant and
J. Alavalapati (eds.). Routledge. 290–306.
Barbier, E. B., & Tesfaw, A. T. (2012). Can REDD+ save the forest? The role of payments and
tenure. Forests, 3(4), 881–895. http://doi.org/10.3390/f3040881.
Busch, J., Lubowski, R. N., Ashkenazi, E., Austin, K., Baccinni, A., Boltz, F., … Yusuf, A. A.
(2013). Reducing emissions from deforestation in Indonesia: Carbon payments or a morato-
rium on concessions? In Preparation.
Busch, Jonah, Kalifi Ferretti-Gallon, Jens Engelmann, Max Wright, Kemen G. Austin, Fred Stolle,
Svetlana Turubanova, et al. (2015). “Reductions in Emissions from Deforestation from
Indonesia’s Moratorium on New Oil Palm, Timber, and Logging Concessions.” Proceedings of
the National Academy of Sciences of the United States of America 112 (5): 1328–33.
doi:10.1073/pnas.1412514112.
Coase, R. H. (1960). Problem of social cost. The Journal of Law & Economics, 3, 1.
De Koning, F., Aguiñaga, M., Bravo, M., Chiu, M., Lascano, M., Lozada, T., & Suarez, L. (2011).
Bridging the gap between forest conservation and poverty alleviation: The Ecuadorian Socio
Bosque program. Environmental Science & Policy, 14(5), 531–542. http://doi.org/10.1016/j.
envsci.2011.04.007.
Fouré, J., Bénassy-Quéré, A., & Fontagné, L. (2013). Modelling the world economy at the 2050
horizon. Economics of Transition, 21(4), 617–654. http://doi.org/10.1111/ecot.12023.
Garibaldi, L. A., Steffan-Dewenter, I., Winfree, R., Aizen, M. A., Bommarco, R., Cunningham,
S. A., … Klein, A. M. (2013). Wild pollinators enhance fruit set of crops regardless of honey
bee abundance. Science, 339(6127), 1608–1611. http://doi.org/10.1126/science.1230200.
Golub, A. A., Henderson, B. B., Hertel, T. W., Gerber, P. J., Rose, S. K., & Sohngen, B. (2013).
Global climate policy impacts on livestock, land use, livelihoods, and food security. Proceedings
of the National Academy of Sciences of the United States of America, 110(52), 20894–20899.
http://doi.org/10.1073/pnas.1108772109.
Gong, Y., R. Hegde, and G.Q. Bull. (2014). Payment for Ecosystems Services: Lessons from
Developing Countries. In Handbook in Forest Economics. S. Kant and J. Alavalapati (eds.).
Routledge. 210–224.
Golub, A., Hertel, T. W., Lee, H.-L., Rose, S., & Sohngen, B. (2009). The opportunity cost of land
use and the global potential for greenhouse gas mitigation in agriculture and forestry. Resource
and Energy Economics, 31(4), 299–319. http://doi.org/10.1016/j.reseneeco.2009.04.007.
Gong, Y., Hegde, R., & Bull, G. Q. (2013). Payment for ecosystems services: Lessons from devel-
oping countries. In Handbook in forest economics.
Griffith, R., Redding, S., & Reenen, J. V. (2004). Mapping the two faces of R&D: Productivity
growth in a panel of OECD industries. Review of Economics and Statistics, 86(4), 883–895.
http://doi.org/10.1162/0034653043125194.
Hertel, T. W. (2013). Land, environment and climate: Contributing to the global public good
(UNU-WIDER Working Paper 2013-107). Retrieved from http://www.wider.unu.edu/publica-
tions/working-papers/2013/en_GB/wp2013-107/.
Hussein, Z., Hertel, T., & Golub, A. (2013). Climate change mitigation policies and poverty in
developing countries. Environmental Research Letters, 8(3), 035009. http://doi.
org/10.1088/1748-9326/8/3/035009.
IEA. (2008). Energy technology perspectives 2008. Paris, France: IEA.
98 7  Land-Based Environmental Services

IEA. (2012). World energy outlook. OECD Publishing. Retrieved from http://www.oecd-ilibrary.
org/content/book/weo-2012-en.
Jacobsen, J. B., & Hanley, N. (2009). Are there income effects on global willingness to pay for
biodiversity conservation? Environmental and Resource Economics, 43(2), 137–160. http://
doi.org/10.1007/s10640-008-9226-8.
Klein, A.-M., Vaissiere, B. E., Cane, J. H., Steffan-Dewenter, I., Cunningham, S. A., Kremen, C., et al.
(2007). Importance of pollinators in changing landscapes for world crops. Proceedings of the
Royal Society B: Biological Sciences, 274(1608), 303–313. http://doi.org/10.1098/rspb.2006.3721.
Ludena, C. E., Hertel, T. W., Preckel, P. V., Foster, K., & Nin, A. (2007). Productivity growth and
convergence in crop, ruminant, and nonruminant production: Measurement and forecasts.
Agricultural Economics, 37(1), 1–17. http://doi.org/10.1111/j.1574-0862.2007.00218.x.
Madsen, B., Carroll, N., & Moore-Brands, K. (2010). State of biodiversity markets report: Offset
and compensation programs worldwide. Ecosystem Marketplace. Retrieved from http://www.
ecosystemmarketplace.com/documents/acrobat/sbdmr.pdf.
Muñoz-Piña, C., Guevara, A., Torres, J. M., & Braña, J. (2008). Paying for the hydrological ser-
vices of Mexico’s forests: Analysis, negotiations and results. Ecological Economics, 65(4),
725–736. http://doi.org/10.1016/j.ecolecon.2007.07.031.
Nepstad, Daniel, David McGrath, Claudia Stickler, Ane Alencar, Andrea Azevedo, Briana Swette,
Tathiana Bezerra, et al. (2014). Slowing Amazon Deforestation through Public Policy and
Interventions in Beef and Soy Supply Chains. Science 344 (6188): 1118–23. doi:10.1126/
science.1248525.
Pagiola, S., Arcenas, A., & Platais, G. (2005). Can payments for environmental services help
reduce poverty? An exploration of the issues and the evidence to date from Latin America.
World Development, 33(2), 237–253.
Paying local people to stop deforestation. (2009, September 24). The Economist.
PEACE. (2007). Indonesia and climate charge: Current status and policies. Jakarta, Indonesia:
DFID and the World Bank.
Pelikan, J., Britz, W., & Hertel, T. W. (2015). Green light for green agricultural policies? An analy-
sis at regional and global scales. Journal of Agricultural Economics, 66(1), 1–19. http://doi.
org/10.1111/1477-9552.12065.
Peters-Stanley, M., Hamilton, K., & Yin, D. (2012). Leveraging the landscape: State of the forest
carbon markets 2012. Washington, DC: Ecosystem Marketplace.
Phalan, B., Onial, M., Balmford, A., & Green, R. E. (2011). Reconciling food production and
biodiversity conservation: Land sharing and land sparing compared. Science, 333(6047),
1289–1291. http://doi.org/10.1126/Science.1208742.
Reilly, J., Melillo, J., Cai, Y., Kicklighter, D., Gurgel, A., Paltsev, S., … Schlosser, A. (2012).
Using land to mitigate climate change: Hitting the target, recognizing the trade-offs.
Environmental Science & Technology, 46(11), 5672–5679. http://doi.org/10.1021/es2034729.
Ricketts, T. H., Daily, G. C., Ehrlich, P. R., & Michener, C. D. (2004). Economic value of tropical
forest to coffee production. Proceedings of the National Academy of Sciences of the United
States of America, 101(34), 12579–12582. http://doi.org/10.1073/pnas.0405147101.
Sohngen, B. (2010). An analysis of forestry carbon sequestration as a response to climate change.
Copenhagen Consensus on Climate.
Teja Tscharntke, Y. C. (2012). Global food security, biodiversity conservation and the future of
agricultural intensification. Biological Conservation, 151(1), 53–59. http://doi.org/10.1016/j.
biocon.2012.01.068.
The Carbon Hunters. (2010, May 11). Frontline world. PBS. Retrieved from http://www.pbs.org/
frontlineworld/stories/carbonwatch/2010/05/the-carbon-hunters.html.
UN Population Division. (2013). World population prospects: The 2012 revision. New York, NY:
Population Division of the Department of Economic and Social Affairs of the United Nations
Secretariat. Retrieved from http://esa.un.org/unpd/wpp/index.htm.
Unruh, J. D. (2008). Carbon sequestration in Africa: The land tenure problem. Global Environmental
Change, 18(4), 700–707. http://doi.org/10.1016/j.gloenvcha.2008.07.008.
Zhang, W., Ricketts, T. H., Kremen, C., Carney, K., & Swinton, S. M. (2007). Ecosystem services
and dis-services to agriculture. Ecological Economics, 64(2), 253–260. ­http://doi.org/10.1016/j.
ecolecon.2007.02.024.
Chapter 8
Biofuels as a Driver of Long Run Land
Use Change

8.1  Historical Perspective and Policy Context1

Bioenergy has played an important role in the global economy ever since the inven-
tion of fire. Indeed, the use of wood for heating and cooking fuel remains important
today in many developing countries. However, bioenergy for transportation—com-
monly referred to as biofuels, is a more recent phenomenon. Ethanol has been pro-
duced as a fuel source since the 1970s. Beginning with the Energy Policy Act of
1978, the United States began providing explicit subsidies for ethanol, thereby kick-­
starting the industry that we see today. And the ethanol sector remains an industry
with a wide range of state and federal support in the U.S. [see Tyner (2008), Table 1,
for a detailed history]. The most important legislation today is the RFS2, which
mandates the annual percentage standards and required volumes for cellulosic and
advanced biofuels, biodiesel and total renewable fuel (US EPA, 2013).
The other major producer of ethanol for transportation is Brazil, where sugarcane-­
based ethanol has been produced for several decades. Indeed, until the recent bio-
fuel boom in the U.S. during the mid-2000s, Brazil was the world’s leading producer
of ethanol, accounting for at least half of the global supply of fuel ethanol (Brown,
2012). The ethanol industry in Brazil took off during the 1970s, when oil price
shocks and falling sugar prices led to the establishment of the National Alcohol
Program. Although the primary goal of this program is to save foreign exchange
from fuel importation by using ethanol as a substitute, it was also seen as a way to
boost incomes of sugarcane farmers. With the availability of low cost sugarcane
feedstock and credit support from the government, ethanol production targets were
quickly met. To further boost domestic demand for ethanol, both the Brazilian
­government and private industries pushed for the introduction of alcohol-based cars
and later flex-fuel vehicles (Hira & de Oliveira, 2009; Leite, Leal, Cortez, Griffin,

 This section draws heavily on Hertel and Tyner (2013).

1

© Springer International Publishing Switzerland 2016 99

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_8
100 8  Biofuels as a Driver of Long Run Land Use Change

& Scandiffio, 2009). Currently, around 88 % of ethanol production in Brazil is con-

sumed domestically and is used as transportation fuel (Barros, 2012).
Many of the same factors that led to the ethanol programs in the U.S. and
Brazil—high oil prices, a desire for energy security, cleaner energy sources, and an
interest in supporting agriculture, led to ambitious programs being proposed else-
where during the past decade. To reach its GHG targets in the Kyoto protocol, the
European Union mandated that around 5.75 % of road transportation fuel consumed
in the region should come from biofuels by the end of 2010. More recently, the
Renewable Energy Directive was adopted in the region which includes a binding
target for renewable energy used in road transportation and a stringent set of envi-
ronmental sustainability criteria for all biofuels consumed in the E.U. (Al-Riffai,
Dimaranan, & Laborde, 2010; European Council, 2009). Under the Medium and
Long-Term Development Plan for Renewable Energy, China planned to increase its
use of biodiesel and ethanol by 2 and 10 M Mt per year by 2020, respectively (GSI,
2008). But with food price spikes during the past years, government support for
grain-based ethanol in China has waned while the production of ethanol from non-­
food crops, such as sweet sorghum and cassava, and biodiesel from oil crops has
been prioritized (Scott & Junyang, 2012). In India, the government has recently
approved the National Policy on Biofuels which mandates that 20 % of diesel and
gasoline requirements should come from biodiesel and bioethanol by 2017 (Sorda,
Banse, & Kemfert, 2010). Unlike diesel, gasoline use in the Indian transportation
sector is limited; hence, it is expected that biodiesel production in India will expand
in coming decades (Arora et al., 2010).
With increased demand worldwide, some countries has focused on exporting
bioenergy in international markets. Argentina, a traditional producer and exporter of
oilseeds, is one of the top producers of biodiesel in the world. Although there are
existing domestic blending requirements in Argentina, most of the biodiesel pro-
duced in the country are exported. Around 60 % biodiesel production in Argentina
is sold outside of the country but more stringent environmental requirements
imposed in the destination markets—particularly the U.S. and the E.U.—could
potentially limit the volume of biodiesel exports in the coming years (Joseph, 2013).
Another world supplier of biodiesel is Indonesia. It exports around 70 % of its
domestic biodiesel production which is generally produced from palm oil (Slette &
Wiyono, 2013). To ramp-up domestic consumption of bioenergy, recent legislations
in Indonesia require that the shares of biodiesel in the transportation sector should
reach 20 % by 2025, while for ethanol, its share in gasoline fuel use should be at 15
% (Kumar, Shrestha, & Abdul Salam, 2013).
Despite their popularity, there are concerns about the impacts of bioenergy man-
dates across the world on the global food grains market. Fischer (2009) undertook a
study of these links for the 2009 FAO food summit. In their “TAR-V1” scenario the
mandatory, voluntary and indicative targets for biofuel use announced by both
developed and developing countries are implemented by 2020—boosting produc-
tion to twice the 2009 levels. Second generation technologies are assumed to
become available after 2015, and are only gradually deployed under this scenario.
This results in 2020 cereal prices which are 38 % above baseline, falling off to 27 %
above baseline by 2050, when second generation biofuels assume a larger share of
8.1 Historical Perspective and Policy Context 101

the total. The price rise is widespread and crops prices increase by roughly the same
amount as do cereals (35 % in 2030 and 27 % in 2050). This price rise boosts culti-
vated land globally by 38 M ha in 2020 and 48 M ha (or about 2.8 %) by 2050.
Since ethanol-based fuel competes with petroleum, the price of oil is a key deter-
minant in the commercial viability of producing these biofuels. Throughout the
period 1978–2003, oil prices remained in the $10–$30/bbl range, thereby limiting
serious expansion of biofuels, which required significant subsidies to be competi-
tive at these low oil prices (Tyner 2008, p. 647). The rise in oil prices after 2003,
reaching roughly $100/bbl by 2008, changed this economic landscape dramati-
cally—particularly since the earlier subsidies were left in place for a number of
years. This oil price boom, in combination with the RFS2 which ensured a market
for new capacity, led to a boom in ethanol production in the U.S., with a significant
impact on world demand for corn. Indeed over the marketing periods 2005/06–
2007/08, at least 43 % of the global increase in grain demand came from feedstock
demand by the U.S. ethanol industry (Westhoff, 2008). Higher demand has, in turn,
bid up the price of the feedstock, and high corn prices have led to an expansion of
planted area, thereby contributing to global land use change.
However, further expansion of the ethanol industry in the United States is con-
strained by the so-called “blend wall”—the limit on the amount of ethanol which
can be blended with petroleum to produce gasoline which is usable by all motor
vehicles (Tyner, 2008). With roughly 135 BGs per year of gasoline being consumed
in the U.S. in 2013, this places the total ethanol which can be absorbed at 13.5 BGs
per year (given 10 % blend rate), which is less than current capacity and also less
than the mandated 2015 production under the RFS2. While the US-EPA has ruled
that automobiles built after 2001 can use a 15 % blend, this is not yet operational, as
about one-third of the U.S. vehicle stock pre-dated 2001 as of 2012. Therefore, we
are unlikely to see much additional expansion of ethanol production in the United
States in the near term (Tyner, 2008). Indeed, the EPA has rolled back the RFS tar-
gets to slightly more than 15 bgy, of which about 13 bgy would come from corn-­
based ethanol. This marks a sharp reversal of momentum in this industry which had,
until recently, been booming.
The ethanol blend wall is also limiting the use of ethanol from second generation
biofuels, which appear to be economically infeasible in the medium term in any
case. A recent study on biofuels by the National Academy of Sciences estimates
that, at $111/bbl oil, the price gap between what processors can pay and suppliers'
likely costs, amounts to anywhere from $0.77 to $1.51/gallon for a variety of differ-
ent second generation feedstocks [see NRC (2011), Table 2]. For this reason they
conclude that the RFS2 cellulosic fuel mandate is unlikely to be met in 2022—a
factor which will come into play when we project the impact of future biofuels
developments on global land use.
One way of skirting the blend wall is to produce ‘drop-in’ fuels which substitute
readily for gasoline. This involves thermochemical processing which can take, as its
feedstock, cellulosic materials which are not destined for the food markets. Hertel,
Steinbuks, and Baldos (2013) explored the role of these second generation ‘drop-in’
biofuels in the context of global land use to the year 2100. They estimated that,
under the EIA’s reference scenario for oil prices, second generation cellulosic bio-
102 8  Biofuels as a Driver of Long Run Land Use Change

fuels produced from biomass such as switchgrass could become competitive by

2040, with the total land area under biofuel feedstock rising to 225 M ha by 2100.
They also explored the sensitivity of this source of global land use change to tech-
nological improvements as well as government GHG mitigation policies. They find
that an increase in the conversion efficiency by 50 % adds another 30 M ha by the
end of 2100 to land allocated for growing biofuel feedstocks. With the introduction
of GHG mitigation policies, unmanaged forest area expands, leading to increased
supply of cellulosic feedstocks and a scenario wherein second generation biofuels
become competitive 15 years earlier—in 2025. When it comes to climate mitigation
policies, there is scope for other sources of bioenergy to play a significant role.
Indeed, under some scenarios, scientists envision a ‘back to the future’ scenario in
which there is extensive burning of biomass for energy and power. Rose et al. (2012)
highlight the opportunities for such technologies in a future in which climate miti-
gation policy becomes a serious proposition. They estimate that land-based mitiga-
tion options could contribute 100–340 Gigatons of Carbon equivalent over the
century, which is 15–40 % of the total required for climate stabilization at 450 ppm,
with bioenergy providing up to 15 % of primary energy.

8.2  M
 arket-Mediated Impacts of Biofuels
on Global Land Use2

A central issue in the debate over the environmental effects of biofuels has to do with
the impact on global land use—a point which was first raised by Searchinger et al.
(2008) who concluded that, once the indirect land use impacts of corn ethanol were
factored in, this production of this renewable fuel actually raised global GHG emis-
sions. Subsequent studies have resulted in more moderate estimates of global land
use change from biofuels, but none have refuted this source of emissions altogether.
Starting from a relatively low level of 1.8 bgy of ethanol production in 2001, the
U.S. RFS2 envisioned an increase of 13.3 bgy by 2015. As of the writing of this
chapter, most of this increase has been achieved, with the remaining gap being
attributable to the blend wall discussed above. How much additional land has been
required to meet this new source of corn demand? One approach to estimating this
figure would be to simply look at the actual change in global cropland area since
2001. However, brief reflection suggests that this is fundamentally flawed, since
cropland use over this period has been influenced by many other factors—some of
which are likely far more important, including population, income and agricultural
productivity, as discussed in the preceding chapters. Therefore, we rely on the esti-
mates of indirect land use change from a global economic model. In this case we
draw on the GTAP-BIO-AEZ model, as documented in Hertel et al. (2010) which
was used by the California Air Resources Board to support the analysis underpin-

2
 This section is taken from Hertel et al. (2010). This analysis was originally undertaken
for the California Air Resources Board in conjunction with their setting of that state’s Low Carbon
Fuel Standard, for which these estimates were mandated.
8.2 Market-Mediated Impacts of Biofuels on Global Land Use 103

ning its proposed Low Carbon Fuel Standard. Those authors break their analysis
into four parts: (1) the ensuing reduction in food consumption, (2) the intensifica-
tion of crop production, (3) land use change in the U.S., and (4) land conversion in
the rest of the world.
Domestic market-mediated effects  The results reported in Table 8.1 begin with a
naïve estimate of the output change using the baseline ethanol conversion factor of
2.6 gallon per bu and baseline coarse grains yields of 335 bu per ha with the base-
line area of 36 M ha, which translate into a 42 % increase in baseline production
(Table 8.1, column 4). Of course, any rise in price which reduces consumption of

Table 8.1  Impact on U.S. land use of increasing U.S. corn ethanol from 6.6 to 56.7 GL/year
% change in % change in
individual coarse grains
Units Value variable hectares
Adjustments in coarse grains harvested area
Corn ethanol yield L/Mg 387
Change in ethanol production GL 50
Additional corn required Tg 129 42
2001 coarse grains yields Mg/ha 8.5
Additional equivalent area (using 2001 coarse M ha 15 42
grains yields)
Change in coarse grains output due to:
Change in exports −17 −4
Change in domestic sales 30 22
Decline in non-ethanol domestic sales −31 −17
Domestic sales to livestock −43 −17
Livestock feed demand: substitution of DDGS −37 −15
for the corn in livestock corn-based feed
Livestock feed demand: reduction of livestock −8 −3
corn-based feed
Livestock feed demand: reduction of all feed −1 −0.4
due to reduction in demand for livestock
Other domestic sales −0.3 −0.1
Change in sales to ethanol 757 47
Final change in corn output 17 17
Additional land once demand-side market M ha 6.1 17
forces considered (i.e. constant yields on land
with initial productivity)
Additional land needed when yield increase is M ha 5.0 14
taken into account on land with initial
productivity
Additional land needed when corn yield M ha 6.0 16
increase due to higher prices and corn yield
decline on other cropland converted into corn
are taken into account
Source: Hertel et al. (2010)
104 8  Biofuels as a Driver of Long Run Land Use Change

U.S. coarse grains and export demand is quite price-responsive. Based on the trade
elasticities in GTAP-BIO-AEZ, the 50.15 GL per year rise in ethanol production
reduces gross coarse grain exports from the U.S. by 17 % (Table 8.1, column 4). As
exports constitute 27.6 % of total sales in the base year, this reduces the coarse
grains area requirements in the U.S. by about 4 % (Table 8.1, line 8, column 5).
(Reduced exports will be made up in part by production somewhere else; see below.)
At this point, the 42 % increase in output is reduced to 36 % (Table 8.1, line 10).
Some domestic uses of coarse grains in the United States are also price responsive.
Livestock feed dominates domestic corn use and matters here because a complemen-
tary product of corn ethanol production is distillers’ dried grains with soluble (DDGS),
a product that can be fed to animals in place of grains and soymeal. In effect, convert-
ing a hectare’s worth of corn to fuel does not “use” all the feed value of corn.
Higher coarse grain-to-DDGS price ratios, encourage substitution of the latter for
the former (Taheripour, Hertel, & Tyner, 2008) and a large reduction (37 %) in domes-
tic corn used in livestock feed (Table 8.1, line 12). Other feedstuffs also displace corn-
based feed (Table 8.1, line 13). These two factors combine to provide a 42 % reduction
in the use of corn grain in feed. Higher livestock feed prices reduce consumption of
livestock products themselves. Other domestic uses of coarse grains (e.g., in the man-
ufactured foods and beverages sectors) are smaller, less responsive to price and are
therefore little affected. Taking all these factors into account, the domestic demand
(other than for ethanol) declines by 31 % (Table 8.1, column 4). Since non-ethanol
domestic sales are about two-thirds of baseline coarse grains production, these mar-
ket-mediated responses result in a further 17 % decline in total output requirements,
bringing the revised output requirement figure down from 36 % to about 17 % once
domestic livestock and other responses to changing prices are accounted for.
Switching from the demand to the supply-side, we must consider the response of
yields to higher market prices. If yields on existing coarse grains land increased by
17 % in response to higher corn prices, then there would be no land conversion
needed to meet the increased demand for corn due to ethanol. As discussed in the
preceding chapters, two competing forces are at play in the market-mediated
response of yields to biofuels production. First of all, higher corn prices induce
higher yields (the intensive margin). In response to ethanol expansion, Hertel et al.
(2010) obtain an average U.S. yield increase, owing to intensification, of 2.8 %
(Table 8.2). This means that, rather than rising by 17 %, the land employed by the
coarse grains sector only needs to rise by about 14 % (Table 8.1, final column).
Working in the opposite direction is the tendency for expansion of corn land to
reduce average yields as less productive land is brought into production. Here, there
are two factors at play. First, corn yields change as corn replaces other crops on
existing cropland. This extensive margin is based on observed land rents per hectare
in current use. In the United States, this expansion results in a decline in average
coarse grains yields as corn production expands into less corn-suited land. A second
extensive margin arises when cropland is expanded into pasture and forest lands.
The “extensive margin” row of the U.S. panel in Table 8.1 shows that that these two
factors combined tend to offset the intensification effect, resulting in a net yield
increase for coarse grains of just about 0.4 %.
8.2 Market-Mediated Impacts of Biofuels on Global Land Use 105

Table 8.2  Change in harvested area, by crop: U.S. and rest of the world
USA
Coarse Grains Oilseeds Sugarcane Other Grains Other Crops
Decomposition of output changes (%)
Output 17 −6.1 −1.7 −9.4 −1.7
Yield 0.41 −1.2 0.40 −0.43 −1.3
Area 16 −5.2 −2.1 −9.0 −0.59
Decomposition of yield changes (%)
Yield 0.41 −1.2 0.40 −0.43 −1.3
Intensive 2.8 1.3 1.8 0.86 0.47
Extensive −2.3 −2.5 −1.4 −1.3 −1.7
Harvested area (M ha)
6.0 −1.6 −0.02 −2.7 −0.01
ROW
Coarse Grains Oilseeds Sugarcane Other grains Other crops
Decomposition of output changes (%)
Output 1.0 1.4 −0.15 0.28 0.07
Yield 0.35 0.46 0.29 0.25 0.16
Area 0.69 0.98 −0.43 0.03 −0.11
Decomposition of yield changes (%)
Yield 0.35 0.46 0.29 0.25 0.16
Intensive 0.26 0.32 0.19 0.18 0.10
Extensive 0.09 0.13 0.09 0.07 0.06
Harvested area (M ha)
1.4 1.6 −0.10 0.20 −0.53
Source: Hertel et al. (2010)

Thus to obtain the U.S. coarse grains output of 17 % (Table 8.1) an area increase
of 16 % is required. This amounts to a rise of about 6 M ha of land over the baseline
harvested area (Table 8.1, column 3). How will this equilibrium increase in land
devoted to coarse grains be met? Table 8.2 reports adjustments in harvested area for
other U.S. crops, triggered by the expansion in land devoted to coarse grains. This
amounts to a 4.4 M ha reduction, with most of this coming out of area previously
devoted to oilseeds and other grains. In this analysis, the authors ignore the effects
of such crop-switching on GHG emissions—focusing only on emissions from con-
version of new cropland. However, crop-switching in the United States does leave
a significant gap in world supplies of these other products—some of which will be
produced elsewhere.
As expected, the reduction in total production of these other crops in the U.S. is
also influenced by yield changes which are also reported in Table 8.2. With the
exception of sugar crops, average yields fall—despite the presence of an intensifica-
tion effect. The reason for this decline is that the best soybean land, for example, is
converted to corn, thereby lowering average soybean yields (and similarly for
wheat, etc.). This extensification effect dominates the intensification effect and
106 8  Biofuels as a Driver of Long Run Land Use Change

therefore results in a larger decline in U.S. output and exports than would otherwise
be the case. Indeed, the estimated declines in exports of other grains (−15 %) and
oilseeds (−12 %) rival the percentage export reduction in coarse grains themselves
thereby contributing to increased cropland conversion in rest of the world (RoW).
The final piece of the land use puzzle in the United States is the conversion of
non-cropland to crops, the dominant source of indirect land use change GHG emis-
sions and thus a focus of the debate over ethanol as a renewable fuel. With 6.0 M ha
of increased coarse grains area and 4.4 M ha of reduced area for other crops, net
cropland conversion in the U.S. amounts to 1.6 M ha, which amounts to a roughly 1
% increase in total cropland. This analysis is silent on the precise nature of the land
transitions. A priori, one might expect that most of the cropland will come from
high quality pasture land, with increased demand for pasture infringing on forest
lands. The estimates here suggest that about 2/3 of the net reduction will occur in
pasture land and 1/3 of the net reduction (0.5 M ha) comes from forest cover. The
composition of these land cover changes vary greatly by in the U.S., with pasture
land declining in all areas, but forested lands declining only in the most productive
areas where corn is grown.
Market-mediated effects in the rest of the world  Not surprisingly, the reduction in
coarse grain exports from the U.S. to the RoW results in higher production overseas.
The aggregate increase in RoW coarse grains production is 1 % (Table 8.2, bottom
panel), with the largest contributions coming from Latin America, the E.U. and
China. The distribution of production increases depends not only on existing capac-
ity, but also on bilateral patterns of trade. Those regions that either import a signifi-
cant amount of corn from the United States, or compete with U.S. exports in third
markets, experience the largest increases in production (Villoria & Hertel, 2011).
In the case of non-coarse grain crops, the percentage changes in production in
RoW vary. For oilseeds (+1.4 %) the percentage increase in RoW production is even
higher than for coarse grains. This is a consequence of U.S. oilseeds being signifi-
cantly displaced (1.6 M ha). The rise in other grain production is smaller in percent-
age terms, and ‘other crops’ is smaller yet. RoW production of sugar crops actually
declines, as corn ethanol is substituted in the U.S. for imported cane ethanol.
As with the U.S., the increases in production in RoW are met by a combination
of yield and area increases (Table 8.2, bottom panel). In the case of coarse grains
and oilseeds, the area increase is twice as important as the yield increase, whereas
in the case of other grains the yield response is more important. In the cases of sugar
crops and other crops, area harvested actually falls, while yields increase modestly.
The bottom row of Table 8.2 reports the area changes in RoW. Overall, total crop-
land area rises in all regions excepting Southeast Asia with further conversion of
forest and pasture land to crops. The estimated total cropland conversion is 2.6 M ha
in RoW—the majority of which (2.4 M ha) is net conversion from pasture.
Market-mediated summary  Figure 8.1 summarizes market-mediated adjustments
on global cropland conversion following an increase in U.S. corn ethanol produc-
tion of 50.15 GL per year. This summary is obtained from a series of successively
less restrictive model solutions, each adding another element of the market-­mediated
8.2 Market-Mediated Impacts of Biofuels on Global Land Use 107

Fig. 8.1  Market-mediated reduction in global cropland conversion from 50.135 GL/year of corn
ethanol production (M ha based on 2001 yields). Source: Hertel et al. (2010)

effects. If the constraints were relaxed in a different order, this decomposition would
likely change. The first column reports the gross feedstock requirement (15.2 M ha)
for the 50.15 GL/year increment to U.S. ethanol production. This would apply if
resources (land, labor and capital) were in perfectly elastic supply—an assumption
typically used in life cycle analysis—so that there were no price responses whatso-
ever. The finite availability of suitable land induces a price response which in turn
engenders a reduction in nonfood demand as well as an intensification of livestock
and forestry activities (crop yields are still fixed at this point, as are food demands),
resulting in a reduction in cropland conversion to 11.3 M ha. Use of co-products
further reduces the demand for cropland conversion to 6.6 M ha. This is followed in
Fig. 8.1 by the impact of reduced food consumption, leaving about 4.4 M ha of
global cropland conversion. After that, we see that the competing effects on yields
of higher prices inducing more intensive crop production on the one hand, and crop-
land expansion lowering yields, on the other. These effects are largely offsetting at
the global level (−1.6 M ha vs. +1.4 M ha). This leaves a net cropland conversion
estimate of 4.2 M ha. Thus, market-mediated effects result in net land conversion of
just 0.28 ha for each gross hectare of corn production diverted to fuel use. When
adjusted for 2007 coarse grains yields, this figure is reduced to our final value of
3.8 M ha—just about two-fifths of previous estimates (Searchinger et al., 2008).
108 8  Biofuels as a Driver of Long Run Land Use Change

8.3  P
 rojecting Global Agricultural Land Use to 2050
with Biofuels

We now seek to examine the global land use implications of increased biofuel
demand worldwide between the present and 2050. For this, we again use the
SIMPLE model in which global biofuel demand acts as an exogenous shifter of
global crop demand. It is an indirect source of crop demand since first generation
biofuels generally require crop inputs as feedstock. As noted previously, second
generation biofuels, produced from cellulosic feedstocks, are potentially important
in the future but are not yet commercially viable and may not be for several more
decades. We start by examining how land use projections due to increased biofuel
demand are altered by the incorporation of each of the three key economic margins
of adjustment. On the consumption side, the demand margin governs the respon-
siveness of crop demand to changes in crop prices. On the production side, we have
the intensive and extensive margins of supply response. The intensive margin facili-
tates the substitution of non-land inputs for land, thereby boosting yields, while the
extensive allows for expansion in cropped area. This leads to four different experi-
ments, all with the same exogenous shock to biofuels feedstock demand, but allow-
ing for different market-mediated effects: (1) implementation of the extensive
margin only with short-run supply response, (2) extensive margin only (i.e. longer
run response), (3) extensive & intensive margins only, and (4) all margins. Aside
from these experiments, we also revisit our projections to 2050 introduced in the
previous chapters and include increased biofuel use as a driver of global agriculture
along with population, income and agricultural productivity in order to assess the
role of biofuels in the context of other drivers of global land use.
The biofuel growth rates for these experiments are based on the extended annual
growth rates of global biofuel demand for the years 2006–2035 taken from the
World Energy Outlook by the International Energy Agency (IEA) (2008, 2012b).
The IEA projections of energy demand are based on the simulations of the World
Energy Model—a dynamic partial equilibrium model of the energy markets (IEA,
2012a). It is a large-scale simulation model whose core components consist of final
energy consumption, energy transformation and energy supply from renewable and
non-renewable sources. Projections from the model are typically generated given
assumptions on economic growth, population, energy prices and technology. It can
also be used to assess the impacts of improved energy efficiencies, energy subsidies
and climate policies on long-run energy demand projections. To account for uncer-
tainty in the energy markets, projections of biofuel demand are generated under
three scenarios—Current Policies (CP), New Policies (NP), and 450 ppm climate
Policies (450P). The CP scenario takes into account all policies enacted as of mid-­
2012 while the NP scenario adapts CP’s assumptions plus the added implementa-
tion of modest GHG emissions and energy policies currently under discussion or
recently announced. Examples of these include the implementation of country-level
targets to reduce GHG emissions as discussed under the 2010 Cancun Agreements
and the initiatives to reduce fossil-fuel subsidies by G-20 and APEC member coun-
8.3 Projecting Global Agricultural Land Use to 2050 with Biofuels 109

tries. However, under the NP scenario not all countries meet their emission targets.
Finally, the 450P scenario is based on an ambitious energy pathway that can poten-
tially limit the rise in average global temperature relative to pre-industrial levels.
It adapts the policies under the NP scenario but with more bold targets aimed at
GHG stabilization at 450 part per million. Given the biofuel sector’s share of global
crop use in SIMPLE’s baseline data, projections in 2050 under the CP, NP, and 450P
scenarios translate into increases of about 16.2 %, 20.3 %, and 40.5 % in global crop
demand, respectively.
Before examining the results of these projections, it is useful to return to the
analytical framework laid out in Chap. 1, which offers a useful way of understand-
ing the land use impacts of biofuels and how these are shaped by the types of eco-
nomic responses discussed in the preceding chapters. For the moment, let us also
abstract from changes in demand due to population and income and exogenous
changes in yield due to improved technology. This leaves us with just an exogenous
perturbation to crop demand resulting from expansion of biofuels production (ΔAD).
In this case, Eq. (1.1) simplifies to the following expression for the percentage
change in global agricultural land use (qL*) in response to an outward shift in demand
for agriculture-based biofuel feedstocks:

qL* = [(DAD + DLS - DLD ) / (1 + h AS, I / h AS, E + h AD / h AS, E )] - DLS (8.1)

Recall from Chap. 1 that the denominator of Eq. (8.1) encapsulates three key mar-
gins of economic response to the biofuel expansion. These include the intensive
margin of supply response—represented by the elasticity of yields with respect to
commodity price, ηAS,I, the extensive margin of supply response, given by the price
elasticity of land supply with respect to commodity price, ηAS,E, and the absolute
value of the price elasticity of demand for agricultural output, h AS, I = 0 .
By ignoring the intensive and demand margins of adjustment, many biophysically-­
based analyses of indirect land use change from biofuels understate the denomina-
tor in Eq. (8.1) and thereby overstate the overall change in demand for cropland.
This was found to be the case with the initial Searchinger et al. study discussed
above, which set the intensive margin of supply response to zero. Another key point
from Eq. (8.1) is that it is the relative size of the elasticities that matters for land use
change from biofuels growth. A large value for the intensive margin of supply
response does not necessarily imply less land use change if the extensive margin is
also larger—and similarly for the price responsiveness of demand. We also note that
studies using ‘short run’ land supply elasticities will likely understate the extensive
margin of supply response (ηAS,E) such that the denominator of Eq. (8.1) is overly
large and the equilibrium land use change will be too small.
Table 8.3 summarizes the results for each experiment under the CP, NP and 450P
scenarios. As mentioned earlier, only the biofuel shock is implemented in the first
four experiments. Starting with the experiment in which the only margin of adjust-
ment is short-run extensive margin, we see an increase in global land use under the
CP, NP and 450P biofuel scenarios of 13.6 %, 17.1 % and 34.7 %, respectively.
When the supply of land is more responsive to changes in land rents, the estimated
110 8  Biofuels as a Driver of Long Run Land Use Change

Table 8.3  Global land use outcomes for each biofuel scenario: 2006–2050
Biofuel scenarios
Global land use change “Current policies” “New policies” “450 policies”
In % increase
Biofuels + Extensive margin (short run) 13.6 17.1 34.7
Biofuels + Extensive margin 14.9 18.8 38.6
Biofuels + Extensive & Intensive 6.7 8.4 16.7
margins
Biofuels + All margins 6.0 7.5 14.9
All Drivers + All margins 16.3 17.3 22.1
In million hectares
Biofuels + Extensive margin (short run) 192.3 241.7 490.6
Biofuels + Extensive margin 210.6 266.1 546.0
Biofuels + Extensive & Intensive 95.4 119.6 236.0
margins
Biofuels + All margins 84.3 105.8 210.3
All Drivers + All margins 231.1 244.9 312.3
Source: Authors’ calculations based on simulation of the SIMPLE model

Fig. 8.2  Land market equilibrium with long-run (left) and short-run (right) extensive margins

land expansion from increased biofuel use rises somewhat more. Under the exten-
sive margin case (i.e. longer run response), the corresponding expansion in global
land use under the CP, NP and 450P biofuel scenarios are around 14.9 %, 18.8 %
and 38.6 %, respectively.
The economic intuition behind these results is illustrated in Fig. 8.2. Note that
the extensive margin is linked to the land market and is determined by the supply
response for land—how much land is available given increases in land rents. Given
our baseline parameters, the supply of land is more responsive to changes in land
rents (i.e. flatter supply curve) when the long-run extensive margin is used (left,
Fig. 8.2). This implies that more land is available for use in crop production and, as
anticipated by Eq. (8.1), the equilibrium increase in land use due to greater biofuel
demand will be larger. However, if the extensive margin is calibrated to its short-run
8.3 Projecting Global Agricultural Land Use to 2050 with Biofuels 111

value (right, Fig. 8.2) then the supply of land is less responsive to changes in land
rents (i.e. steeper supply curve). With land becoming increasingly scarce, the equi-
librium land use change from increased biofuel demand is dampened.
Going back to Table 8.3, we observe significant reductions in global land use
expansion from biofuels when both intensive and extensive margins are active. As
predicted by Eq. (8.1), projections of land use expansion under these biofuel sce-
narios are sharply reduced compared to the case wherein only the extensive margin
is active (6.7 %, 8.4 % and 16.7 % in the CP, NP and 450P scenarios, respectively).
Recall that the intensive margin dictates how easily it is substitute between land and
non-land inputs during crop production (i.e. smoothly curved isoquants). Increased
biofuel demand encourages additional crop production which will then lead to
increased demand in both land and non-land inputs. However, growth in demand
will be different between these inputs. Remember that in our framework, the supply
of non-land inputs (e.g., fertilizer) is assumed to be more price responsive than for
land. This implies that the supply of non-land inputs is less scarce than land; hence,
crop output expands by using relatively more non-land inputs and thereby increase
crop yields endogenously (left, Fig. 8.3).
Without the intensive margin (right, Fig. 8.3), crops can only be produced using
fixed-ratios of non-land and land inputs (i.e. L-shaped isoquants). In the absence of
exogenous technological change, crop yields will be fixed under this scenario and
more land use expansion will occur for a given increase in crop demand.
When all margins are activated, (demand included) the corresponding land use is
further dampened, as predicted by Eq. (8.1). Note that with the presence of the
demand margin (left, Fig. 8.4), the rise in global crop demand will be smaller than
the increase in biofuel growth since some of the demand increase is dampened due
to the rise in crop prices (i.e. demand curve has a negative slope). In the absence of
this margin (right, Fig. 8.4), the crop demand is perfectly price inelastic (i.e. demand
curve is a vertical line) and global crop demand will increase in the same magnitude
as the growth in biofuel demand. With greater crop demand, land use from biofuel
expansion will be greater.

Fig. 8.3  Optimized input usage with (left) and without (right) the intensive margin
112 8  Biofuels as a Driver of Long Run Land Use Change

Fig. 8.4  Crop market equilibrium with (left) and without (right) the demand margin

Fig. 8.5  Contribution of biofuels to global cropland expansion: 2006–2050. Source: Authors’
calculations based on simulation of the SIMPLE model

Finally, when we combine biofuels with other drivers of agriculture, the resulting
land use change from the CP, NP and 450P scenarios are 231, 245 and 312 million
hectares, respectively. In comparison to other drivers, we see that the relative con-
tribution of biofuels is smaller than that of population and per capita income under
the CP and NP scenarios (Fig. 8.5). However, under the 450P scenario the rise in
biofuel demand is high enough such that its contribution to global land use change
may exceed that of population growth.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.
References 113

References

Al-Riffai, P., Dimaranan, B., & Laborde, D. (2010). Global trade and environmental impact study
of the EU biofuels mandate. Washington, DC: International Food Policy Research Institute.
Retrieved from http://environmentportal.in/files/biofuelsreportec.pdf.
Arora, D. S., Busche, S., Cowlin, S., Engelmeier, T., Jaritz, H., Milbrandt, A., & Wang, S. (2010).
Indian renewable energy status report: Background report for DIREC 2010 (No. NREL/
TP-6A20-48948) (p. 139). Springfield, VA: NREL; GTZ; REN21; IRADe.
Barros, S. (2012). Brazil—Biofuels annual (Annual GAIN report no. BR12013, p. 28). Sao Paulo,
Brazil: USDA Foreign Agricultural Service.
Brown, L. R. (2012). Full planet, empty plates: The new geopolitics of food scarcity. New York,
NY: W.W. Norton.
European Council. (2009). Directive 2009/30/Ec of the European Parliament and of the Council.
Retrieved from http://www.r-e-a.net/document-library/thirdparty/rea-and-fqddocuments/
FQD_090605_Directive_200930EC_OJ.pdf.
Fischer, G. (2009). How do climate change and bioenergy alter the long-term outlook for food,
agriculture and resource availability? In Session 2: The resource base to 2050: Will there be
enough land, water and genetic potential to meet future food and biofuel demands? Rome, Italy.
GSI. (2008). Biofuels at what cost? Government support for ethanol and biodiesel in China (p. 78).
Manitoba, Canada: International Institute for Sustainable Development; Global Subsidies
Initiative.
Hertel, T. W., Golub, A. G., Jones, A., O’Hare, M., Plevin, R., & Kammen, D. (2010). Effects of
US maize ethanol on global land use and greenhouse gas emissions impacts of US maize etha-
nol: Estimating market-mediated responses. BioScience, 60(3), 223–231.
Hertel, T., Steinbuks, J., & Baldos, U. (2013). Competition for land in the global bioeconomy.
Agricultural Economics, 44(s1), 129–138. http://doi.org/10.1111/agec.12057.
Hertel, T. W., & Tyner, W. E. (2013). Market-mediated environmental impacts of biofuels. Global
Food Security, 2(2), 131–137. http://doi.org/10.1016/j.gfs.2013.05.003.
Hira, A., & de Oliveira, L. G. (2009). No substitute for oil? How Brazil developed its ethanol
industry. Energy Policy, 37(6), 2450–2456. http://doi.org/10.1016/j.enpol.2009.02.037.
IEA. (2008). World energy outlook 2008. Organisation for Economic Co-operation and
Development. Retrieved from http://www.oecd-ilibrary.org/content/book/weo-2008-en.
IEA. (2012a). World energy model documentation. OECD/IEA.
IEA. (2012b). World energy outlook 2012. Organisation for Economic Co-operation and
Development. Retrieved from http://www.oecd-ilibrary.org/content/book/weo-2012-en.
Joseph, K. (2013). Argentina—Biofuels annual (Annual GAIN report, p. 14). Buenos Aires,
Argentina: USDA Foreign Agricultural Service.
Kumar, S., Shrestha, P., & Abdul Salam, P. (2013). A review of biofuel policies in the major biofuel
producing countries of ASEAN: Production, targets, policy drivers and impacts. Renewable
and Sustainable Energy Reviews, 26, 822–836. http://doi.org/10.1016/j.rser.2013.06.007.
Leite, R. C. de C., Leal, M. R. L. V., Cortez, L. A. B., Griffin, W. M., & Scandiffio, M. I. G. (2009).
Can Brazil replace 5% of the 2025 gasoline world demand with ethanol? Energy, 34(5), 655–
661. http://doi.org/10.1016/j.energy.2008.11.001.
NRC. (2011). (Report in brief). Washington, DC: National Research Council; National Academy
of Sciences.
Rose, S. K., Ahammad, H., Eickhout, B., Fisher, B., Kurosawa, A., Rao, S., … van Vuuren, D. P.
(2012). Land-based mitigation in climate stabilization. Energy Economics, 34(1), 365–380.
http://doi.org/10.1016/j.eneco.2011.06.004.
Scott, R., & Junyang, J. (2012). Peoples Republic of China—Biofuels annual (Annual GAIN
Report No. 12044, p. 28). Washington, DC: USDA Foreign Agricultural Service.
Searchinger, T., Heimlich, R., Houghton, R. A., Dong, F., Elobeid, A., Fabiosa, J., … Yu, T. -H.
(2008). Use of U.S. croplands for biofuels increases greenhouse gases through emissions from
land-use change. Science, 319(5867), 1238–1240. http://doi.org/10.1126/science.1151861.
114 8  Biofuels as a Driver of Long Run Land Use Change

Slette, J., & Wiyono, I. E. (2013). Indonesia—Biofuels annual (Annual GAIN Report No. ID1337,
p. 10). Washington, DC: USDA Foreign Agricultural Service.
Sorda, G., Banse, M., & Kemfert, C. (2010). An overview of biofuel policies across the world.
Energy Policy, 38(11), 6977–6988. http://doi.org/10.1016/j.enpol.2010.06.066.
Taheripour, F., Hertel, T., & Tyner, W. (2008). Biofuels and their by-products: Global economic
and environmental implications. Presented at the 11th GTAP Conference/2008 American
Agricultural Economics Association.
Tyner, W. E. (2008). The US ethanol and biofuels boom: Its origins, current status, and future
prospects. BioScience, 58(7), 646–653.
US EPA. (2013). EPA finalizes 2013 renewable fuel standards (No. EPA-420-F-13-042, p. 3).
Washington, DC: U.S. Environmental Protection Agency, Office of Transportation and Air
Quality.
Villoria, Nelson B., and Thomas W. Hertel. 2011. Geography Matters: International Trade Patterns
and the Indirect Land Use Effects of Biofuels. American Journal of Agricultural Economics 93
(4): 919–35. doi:10.1093/ajae/aar025.
Westhoff, P. (2008). Farm commodity prices: Why the boom and what happens now? Choices,
23(2), 6–10.
Chapter 9
Livestock and Processed Foods

9.1  Overview

Up to this point, most of the focus in this book has been on crop land expansion,
crop production and the use of crop outputs for food, feed and fuel. However, fresh
food sales account for less than one-quarter of global food sales and the fastest
growing categories of food consumption are processed foods, including breakfast
cereals, ready-to-eat meals, fats and oils and dairy products (Gehlhar & Regmi,
2005). Each of these processed food categories is, in turn, tied back to crop produc-
tion, but the food product being consumed has a great deal of ‘value-added’ which
comes in the form of manufacturing inputs, as well as wholesale/retail distribution
services. If we wish to understand the future demand for food crops, we need to
begin by understanding the evolving demands for livestock and manufactured food
products. Given the livestock sector’s relative importance in crop demand (40 % of
U.S. cropland is devoted to feed production), as well as its significant environmental
impacts (Eshel, Shepon, Makov, & Milo, 2014), we focus the bulk of our discussion
here on the livestock sector.

9.2  Livestock Production and Consumption

Figure 9.1, taken from Taheripour, Hurt, and Tyner (2013), shows the evolution of
production (which is equal to consumption, at global scale) for the world’s major
livestock products since 1970. (We abstract here from fisheries, given data limita-
tions.) Meat, milk and egg production has risen strongly, with the highest growth
rate for eggs where demand has quadrupled since 1970. And most of the growth in
global demand for eggs stems from China, where production grew tenfold over the
four decades: 1970–2010. China’s growth also fueled the largest share of the rise in
global meat demand over this period, with pork being the largest contributor,

© Springer International Publishing Switzerland 2016 115

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_9
116 9  Livestock and Processed Foods

Fig. 9.1  Trends in global livestock production, by region. Source: Taheripour et al. (2013)

showing about a tenfold rise in output since 1970. The growth in global milk pro-
duction was fueled to a great extent by India, where dairy products replace meat in
much of the local diets.
In order to support the strong increases in livestock consumption shown in
Fig.  9.1, a great deal of feedstuffs are required. While feed conversion factors vary
greatly by location and variety of animal, in the United States it takes roughly 7 kg
of dry feed to produce 1 kg of beef, and the average conversion factors for pigs and
poultry are roughly 3:1 and 2:1, respectively. This means that feeding people live-
stock products typically requires more land than simply feeding them staple grains,
for example. These conversion factors are even more striking when placed in caloric
terms. Eshel et al. (2014) compute the caloric feed requirements used to deliver a
calorie in the form of human-consumed livestock products in the United States. For
example, it takes roughly ten calories of feedstuff consumed by poultry to deliver a
9.3 Environmental Impacts of Livestock Consumption 117

single calorie of poultry meat for human consumption. This conversion factor is
slightly higher for pork and lower for eggs and dairy products. However, it is four
times as large for beef (Eshel et al., 2014).

9.3  Environmental Impacts of LivestockConsumption

The combination of high conversion rates from crops to livestock products and
extremely rapid growth in consumer demand for meat, milk and eggs in developing
countries has placed an increasing burden on the planet’s resources. Researchers
have drawn a link between rapid growth in meat demand in China, for example, and
deforestation in Latin America (Fearnside, 2001). This has led to increased interest
in the connection between livestock consumption and a variety of environmental
indicators. Eshel et al. (2014) have recently sought to quantify these linkages in the
context of the U.S. food economy. Figure 9.2 summarizes their findings with respect
to four environmental metrics: land use, irrigation water, greenhouse gas emissions
and applications of nitrogen fertilizer (reactive N).
Consider, for example, the first bar chart in Fig. 9.2, a, which reports annual land
use per Mega calorie (Mcal) produced for five categories of livestock products.
Eggs require just 3 m2/Mcal. Poultry meat ranks second with just under 5 m2/Mcal,
followed by pork and then dairy. In the case of dairy, the authors distinguish between
pasture land (cross-hatched area) and crop land. From this, it can be seen that about
two-thirds of the land requirements for dairy consist of pasture land, given current
farming practices in the United States. Beef production, on the other hand, requires
147 m2/Mcal—an amount so large that it cannot be shown on the same scale! While
the bulk of this land area is pasture, if one focuses solely on cropland used to ­support
U.S. beef production, the requirements are 3–5 times as large as for poultry, eggs
and pork production. This is in part due to the lower conversion efficiency for beef
mentioned above.

Fig. 9.2  Environmental performance of key livestock categories in the U.S. diet. Source: Eshel
et al. (2014)
118 9  Livestock and Processed Foods

Eshel et al. (2014) also report resource requirements for irrigated water, greenhouse
gases and reactive nitrogen. Most of the irrigation and nitrogen fertilizer require-
ments for livestock production derive from the production of feedstuffs. In the case
of irrigation water, most of the livestock products require 0.1 - 0.2m3/Mcal delivered
to human consumption. However, this figure for beef is much higher, totaling
16 m3/Mcal. In the case of nitrogen fertilizer, the story is similar, with most of the
livestock products showing overlapping confidence intervals. Beef however, stands
out as requiring about five times as much reactive Nitrogen, 176 gNr/Mcal.
The case of GHG emissions (Panel c in Fig. 9.2) is somewhat different from the
other environmental metrics since the livestock themselves (not just feedstuffs) are
a large source of methane emissions. Indeed, ruminant livestock are the largest
source of GHG emissions from agriculture, and a dominant source of non-CO2
emission worldwide (Golub et al., 2013). When compared to all sectors, and evalu-
ated in terms of CO2 equivalent GHG emissions per $ of output produced, the rumi-
nant meats sector has the highest emissions intensity of all sectors in the global
economy—higher even than the electric power sector (Avetisyan, Hertel, &
Sampson, 2014). This emissions intensity is considerably lower in the U.S. due to
higher animal yields and higher prices for animal products, but it still remains
among the handful of top sectors in terms of emissions intensity (Avetisyan, Golub,
Hertel, Rose, & Henderson, 2011). It is for this reason that beef stands out in panel
c of Fig. 9.2, with an emissions intensity of 10 kgCO2e/Mcal—once again placing it
nearly one order of magnitude above the other animal products in that figure.
For purposes of comparison with a plant-based diet, Eshel et al. (2014) also com-
pute resource requirements for several staple crops, including wheat, potatoes and
rice. These are reported in green text with accompanying arrows along the top of
panels a–d in Fig. 9.2. Potatoes (p) and rice (r) require land area of less than
1 m2/Mcal of plant product consumed. Wheat (w) land requirements are somewhat
higher, due to lower average yields, as this crop is typically produced on more mar-
ginal lands in the U.S. Rice production in the U.S. is heavily irrigated, placing this
crop above all of the livestock products according to that metric, with the exception
of beef. Potatoes are on par with the other livestock products, while wheat—typi-
cally a dry land crop in the U.S., is less demanding of irrigation water. Greenhouse
gas emissions per Mcal produced are lower for wheat and potatoes, while rice—a
large emitter of methane gas—has a GHG emissions intensity comparable to that of
dairy, poultry and pork. Finally, reactive nitrogen intensities for the crops are also
lower than for livestock products—largely due to their lesser land intensity.

9.4  Economic Implications of LivestockConsumption

Eshel et al. (2014) summarize current U.S. consumption patterns in the final panel
of Fig. 9.2. In particular, they report the share of average daily per capita caloric
intake represented by each of these product categories. Dairy is the most important
category of livestock consumption (measured in caloric terms), followed by beef,
9.4 Economic Implications of Livestock Consumption 119

poultry, pork and eggs. They argue that the heavy reliance on beef represents a
misallocation of resources, given the large environmental footprint of these beef-
derived calories. However, there are many consumer choices made in society which
have adverse environmental consequences. If the national goal was to minimize our
environmental footprint, we would ask everyone to walk or bicycle to work, cease
the use of air conditioners, and become vegetarians. Economists’ response to such
proposals is that it is not possible for scientists to dictate human behavior. Rather,
one must use market signals to encourage individuals to factor in the environmental
footprint of their individual consumption decisions.
Viewed through this economic lens, the problem with the current set of con-
sumer choices is that the market price for beef—and many other products—does
not reflect many of the environmental costs. For example, absent a carbon tax on
GHG emissions, consumers purchasing beef—or any other product for that mat-
ter—fail to factor into their decisions the social cost of increasing GHG concentra-
tions in the atmosphere. A similar problem arises in the case of irrigation water
when the rate of withdrawals exceeds recharge rates. Producers do not factor in the
cost to future residents of exhausting groundwater supplies or restricting in-stream
flows of water below environmentally sustainable levels. In the case of nitrogen
fertilizer applications, excess usage results in significant damage to ground water
quality, as well as creating ‘dead zones’ (hypoxia) in marine environments like the
Gulf of Mexico. Finally, in the case of industrial crop production, producers are
rarely asked to factor in the cost of reduced biodiversity from crop land expansion
and pesticide application. All of these are termed ‘externalities’ by economists.
These represent costs to society which are not internalized by producers and are
therefore not reflected in the prices of the food products purchased by consumers.
Therefore, all else equal, we expect consumers to buy more of “emissions inten-
sive” food products—for example beef—than would be the case if they were con-
fronted by its true social cost.
Avetisyan et al. (2011) focus on the GHG emissions externality and explore the
impact which a global tax on the emissions associated with ruminant meat produc-
tion would have on production, consumption and international trade in beef and
other ruminant meat products. Not surprisingly, they find that such a tax would
reduce consumption of beef, worldwide. However, the size of this reduction varies
greatly by region. In the U.S. and E.U., it is between 4 and 6 %, whereas in Brazil,
the reduction in consumption is more than 20 %. The reason for these variations can
be traced back to the stark differences in emissions intensities of ruminant meat
production across the globe. Whereas U.S. beef emits about 4.5 MkgCO2e/$ output,
the same figure is more than ten times as high in Brazil. As a consequence, an
externality-correcting tax on GHG emissions from beef would result in a sharp
reconfiguration of global beef production, favoring those regions with low emis-
sions intensities. Indeed Avetisyan et al. (2011) predict that such a tax would benefit
producers in the U.S. and E.U., where production would rise, with the surplus being
exported to high emissions intensity regions of the world.
Even without externality-correcting taxes, the global beef sector has been suffering
economically due to its lesser efficiency in the conversion of costly feedstuffs into
120 9  Livestock and Processed Foods

human consumed calories and protein. Between 2006 and 2013, in the wake of the crop
commodity boom, beef prices in the U.S. rose from $3.97/lb. to $5.29/lb. and has
resulted in a decline in total consumption of about 10 %, even as population and
incomes have risen. (The decline in per capita consumption over this period was even
larger—reaching nearly 20 % by the year 2014.) Many consumers have substituted
cheaper pork and poultry meat for beef (recall the production figures in Fig. 9.1).
Clearly households do respond to price signals, which is why economists argue for
taxes or regulations which will result in consumers factoring in the environmental costs
of livestock production when they make purchasing decisions in the grocery store.

9.5  I ncorporating Livestock and Processed Foods

into SIMPLE

As we have seen in Chap. 2, livestock products have fundamentally different demand

characteristics than crop commodities. Therefore, they cannot simply be viewed as
‘embodied crops’ when projecting future demand for global cropland. However, once
we have projected future demands for livestock and processed food products, we
need a vehicle for translating these additional food demands back to the farm-level—
obtaining what economists term the ‘derived demand’ for crops. Within the SIMPLE
framework used throughout this book, we handle the transformation of crops into
other food products through the addition of two, post-farm gate sectors: livestock and
other food processing. These sectors purchase crop outputs and combine these crops
with non-crop inputs in order to produce the consumer goods: livestock and other
processed foods. (Note that in the version of SIMPLE used in this book, we subsume
grazing within the non-cropland inputs into livestock production.)
The supply chains linking farm producers to consumption of processed food prod-
ucts are quite sophisticated and are continuously evolving (King & Venturini, 2005).
We do not attempt to model all of these factors in SIMPLE. Rather, in the spirit of
keepings the model simple, we allow for three separate factors to determine the
demand for crops by the (e.g.) livestock sector and these are illustrated in Fig. 9.3.

Fig. 9.3  Sources of derived demand of crops by the livestock sector in SIMPLE
9.5 Incorporating Livestock and Processed Foods into SIMPLE 121

In the left panel of Fig. 9.3, we see that increasing livestock demand will result
in greater production and an outward shift of the sector’s isoquant curve (Q to Q*).
Given that relative input prices (PF/PNF) do not change, a higher level of output
requires additional use of both feed (XF to XF*) and non-feed (XNF to XNF*) inputs. All
else constant, for example, we expect a 10 % rise in livestock demand to result in a
10 % rise in feed demand. We call this the ‘expansion effect’ of livestock production
on crop demand, and this is the first, and most intuitive driver of crop usage in the
livestock sector.
However, livestock technology has been evolving over the past decades and it is
now much more dependent on feed concentrates, which are replacing grazing and
dedicated feed crops (Taheripour et al., 2013). This leads us to the second source of
change in the livestock sector’s demand for crops, namely technological change
(middle panel of Fig. 9.3). This is represented by a new isoquant wherein the same
level of output is produced but technological change is biased towards more inten-
sive use of feedstuffs leading to a greater use of this input, relative to other inputs
(which include pasture). Therefore, even holding total livestock output constant as
well as relative input prices, we see the demand for crops in this sector to rise (XF to
XF*). We refer to this as the technological change effect.
The third, and final source of potential change in the livestock sector’s feed crop
demand in SIMPLE reflects the responsiveness of feed demand to changes in rela-
tive prices, which we term the “substitution effect” (right panel of Fig. 9.3). When
corn and soybean prices rose sharply after 2007, livestock producers shifted to lower
cost substitutes, including by-products from the food processing and biofuels sec-
tors. This attempt to minimize the cost of producing livestock in response to changes
in relative input prices is captured by the elasticity of substitution between crop and
non-crop inputs in SIMPLE. In the long sweep of post-WWII economic history,
crop prices have actually been falling, relative to non-crop input prices. This is por-
trayed in the right-most panel of Fig. 9.3 where we show how falling feed prices and
rising non-feed costs result in a steeper isocost curve (PF/PNF to PF*/PNF*). With the
capacity for input substitution, firms will react to changing input prices by using
more feed inputs and less of expensive non-feed inputs. In contrast to the livestock
sector, the food processing sector in SIMPLE only allows for the first two channels
for altering the demand for crop inputs since we do not allow for substitution of
between crop and non-crop inputs during the production of processed foods.
An important simplification in SIMPLE is the assumption that livestock and pro-
cessed foods are not traded. In fact, of the more than $3 trillion in global processed
food sales, only 10 % of the products were traded (Gehlhar & Regmi, 2005). Most
were produced locally, often using imported crop inputs. In fact, for most manufac-
tured food products, the preferred method of international engagement is foreign
direct investment. By placing the factory close to the consumer market, firms are
able to tailor their products to local tastes adopt local brands. This approach is less
dominant in the case of commodity-based food products such as fresh/frozen meat
and canned fruits and vegetables, where the simplification in SIMPLE is more prob-
lematic. However, since our focus is on global crop production and crop land use,
we find this to be a cost worth bearing. By considering only trade in crops, we are
122 9  Livestock and Processed Foods

able to reduce the long run equilibrium condition to just one market clearing condi-
tion—namely global supply is equal to global demand for crops.

9.6  P
 rojecting Livestock and Processed Food
Demand Using SIMPLE

We conclude this chapter with a series of projections designed to assess the impacts
of growth in the derived demand for crops by the livestock and processed food sec-
tors on global crop production, crop land use, yields and crop price. We build on the
projections developed in Chap. 7 and look at the future period 2006–2050 given
expected growth rates in population, income growth, biofuels as well as total factor
productivity in the crops, livestock and processed food sectors. Following Fig. 9.3,
we start by looking at changes in global agriculture due to the expansion effect in
the derived demand for crops by the livestock and processed food sectors.1
Specifically, the ‘Expansion: Pop only’ scenario considers the impact of population
growth alone on the derived demand for crops while the ‘Expansion: Pop + Inc’
scenario includes both population and income growth. We then build on the second
scenario and activate the input substitution effect (‘Expansion + Input Sub’) allow-
ing livestock producers to respond to changing feed prices by adjusting input use.
Lastly, we include a future baseline scenario which accounts for not only expansion
and substitution effects in the derived demand for crops but also input-biased tech-
nical change in both livestock and processed food sectors based on the total factor
productivity estimates from Ludena, Hertel, Preckel, Foster, and Nin (2007) and
Griffith, Redding, and Reenen (2004), respectively.
Figure 9.4 summarizes the results of our forward looking projections given the
three key sources of derived demand for crops. If we only consider the expansion
effect and just look at the contribution of population (‘Expansion: Pop only’) then
crop use in the livestock and processed food sector is projected to increase at a similar
rate to population growth. (Differences arise due to the price responsiveness of con-
sumer demand as well as interactions between regional population growth rates and
differences in regional consumption of livestock and processed food products.) Given
the slow growth in the world’s population (0.80 %/year), we see a modest increase
(0.69 %/year) in the combined global derived demand for crops by these sectors lead-
ing to a steady expansion in global crop production (by 59 %). Crop yield growth
rises nearly as fast as global crop production (at 55 %), resulting in a virtually flat
growth in cropland use (2 %) over the 2006–2050 period when population is the only
demand driver. With a slower increase in overall crop demand and rising productivity,
it is not surprising that crop prices decline sharply under this scenario (by −31 %).
When we include income in addition to population (‘Expansion: Pop + Inc’) the
global derived demand for crops by the livestock and processed food sectors grows
twice as fast (1.87 %/year) relative to the previous scenario. This highlights the

 Note that in all scenarios, consumer response to changing food prices is permitted.
1
9.6 Projecting Livestock and Processed Food Demand Using SIMPLE 123

Fig. 9.4  Decomposition of source of growth in the derived demand for crops by the livestock and
processed food sectors. First bar is based on population growth alone. The second bar adds income
growth to the drivers of change. The third and fourth bars add, successively, input substitution in
livestock and technological change in both the livestock and food processing sectors

importance of income growth in shaping not just global food consumption patterns,
as observed in Chap. 2, but also global crop production and land use. Under this
scenario, we see sharp expansion in global crop production (102 %) and a lesser
reduction in global crop price (−13 %). Despite strong yield growth (79 %), there is
a need to expand crop land use (13 %) in response to the remarkable increase in the
global derived demand for crops.
Including the substitution effect (‘Expansion + Input Sub’) results in more inten-
sive use of crop inputs due to declining long crop prices (now at −11 % with the
substitution effect present). The growth in the combined global derived demand for
crops increases slightly given intensification in crop input use in the livestock and
processed food sectors (1.94 %/year). We see a modest increase in global crop pro-
duction, cropland use and crop yields compared to the projections from the earlier
scenario (at 106 %, 14 % and 81 % respectively).
Finally, we incorporate input-biased technical change (based on historical trends)
into our projections in the four set of bars in Fig. 9.4. This increases the productivity
of non-crop inputs, relative to feedstuffs and results in further growth in the global
derived demand for crops (at 2.08 %/year). In the future baseline scenario, global
crop production increases by 113 %, with global crop land use and crop yields
expanding by around 15 % and 84 % respectively. Despite greater demand in the
future, global crop prices are still expected to decline (at −9 %). With growing
incomes, we expect strong demand for livestock products and processed foods in
124 9  Livestock and Processed Foods

the future and as we have shown this results in the expansion of the derived demand
for crops from these sectors. Furthermore, changes in relative input prices and tech-
nical improvements lead to intensification in crop input use by these sectors.
In conclusion, the world’s future food and land use needs will hinge greatly on the
future trends in the consumption and production of livestock and processed foods
sectors. Understanding the evolution of these consumption patterns, as well as the
technology of these industries is a critical piece of long run sustainability analyses.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Avetisyan, M., Golub, A., Hertel, T., Rose, S., & Henderson, B. (2011). Why a global carbon pol-
icy could have a dramatic impact on the pattern of the worldwide livestock production. Applied
Economic Perspectives and Policy, 33(4), 584–605. http://doi.org/10.1093/aepp/ppr026.
Avetisyan, M., Hertel, T., & Sampson, G. (2014). Is local food more environmentally friendly?
The GHG emissions impacts of consuming imported versus domestically produced food.
Environmental and Resource Economics, 58(3), 415–462. http://doi.org/10.1007/
s10640-013-9706-3.
Eshel, G., Shepon, A., Makov, T., & Milo, R. (2014). Land, irrigation water, greenhouse gas, and
reactive nitrogen burdens of meat, eggs, and dairy production in the United States. Proceedings
of the National Academy of Sciences of the United States of America, 111(33), 11996–12001.
http://doi.org/10.1073/pnas.1402183111.
Fearnside, P. M. (2001). Soybean cultivation as a threat to the environment in Brazil. Environmental
Conservation, 28(1), 23–38. http://doi.org/10.1017/S0376892901000030.
Gehlhar, M., & Regmi, A. (2005). Factors shaping global food markets (Agriculture Information
Bulletin No. 794) (pp. 5–17). United States Department of Agriculture, Economic Research
Service. Retrieved from http://ideas.repec.org/p/ags/uersab/33751.html.
Golub, A. A., Henderson, B. B., Hertel, T. W., Gerber, P. J., Rose, S. K., & Sohngen, B. (2013).
Global climate policy impacts on livestock, land use, livelihoods, and food security. Proceedings
of the National Academy of Sciences of the United States of America, 110(52), 20894–20899.
http://doi.org/10.1073/pnas.1108772109.
Griffith, R., Redding, S., & Reenen, J. V. (2004). Mapping the two faces of R&D: Productivity
growth in a panel of OECD industries. Review of Economics and Statistics, 86(4), 883–895.
http://doi.org/10.1162/0034653043125194.
King, R. P., & Venturini, L. (2005). Demand for quality drives changes in food supply chains. In
M. Gehlhar & A. Regmi (Eds.), Factors shaping global food markets (pp. 18–31, chapter 2).
United States Department of Agriculture, Economic Research Service. Retrieved from http://
ideas.repec.org/p/ags/uersab/33751.html.
Ludena, C. E., Hertel, T. W., Preckel, P. V., Foster, K., & Nin, A. (2007). Productivity growth and
convergence in crop, ruminant, and nonruminant production: Measurement and forecasts.
Agricultural Economics, 37(1), 1–17. http://doi.org/10.1111/j.1574-0862.2007.00218.x.
Taheripour, F., Hurt, C., & Tyner, W. E. (2013). Livestock industry in transition: Economic, demo-
graphic, and biofuel drivers. Animal Frontiers, 3(2), 38–46. ­http://doi.org/10.2527/
af.2013-0013.
Chapter 10
Food Security and Nutrition

10.1 Overview

So far, we have only looked into the aggregate regional and global change in food
production, prices and land use given the key drivers and economic responses gov-
erning global agriculture. However, these global changes in the farm and food sys-
tems have differential impacts on adequacy of food consumption and nutritional
intake by individual households. Prior to modern agriculture, large-scale famine and
food shortages due to crop failure and adverse weather conditions were quite com-
mon. Since then, the global economy has made significant gains in improving agri-
culture’s capacity to feed the world.
During the first World Food Conference (1974) the main concern was focused on
improving food availability by ensuring sufficient food supply and stocks world-
wide. However, between 1969/71 and 2005/07, per capita consumption of cereals
has increased by 10 % from 144 to 158 kg/year (Alexandratos & Bruinsma, 2012).
And the supply of dietary energy per capita—the most basic measure of nutrition—
has increased steadily in the developing world growing by around 0.7 % per annum
since 2000s (FAO, 2015).Yet amidst this time of plenty, hunger is still prevalent,
with roughly 800 million persons—mostly in the developing world—suffering from
caloric undernutrition. Therefore, the issue of food insecurity cannot be solely
attributed to problems on food production and supply.
It is now widely accepted that food accessibility, stability and utilization are also
key components of food security (FAO, 1996, 2009). Equally important to food
availability is access to food which depends on household income. Those individu-
als living in poverty typically experience chronic hunger, regardless of local food
availability. These households are also the most vulnerable to increasing food
prices. Indeed, the recent spikes in food prices have also renewed concerns over
stabilizing food supply since those at risk of hunger are often more susceptible to
short-term disruptions in the farm and food systems. The third leg of nutritional
well-being depends on food utilization which encompasses how individuals and

© Springer International Publishing Switzerland 2016 125

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_10
126 10 Food Security and Nutrition

households make use of the food available to them. By following best practices in
food storage, preparation and feeding, food can be better absorbed by the body lead-
ing to improved nutritional status. Health status is also an important determinant of
food utilization.
In this chapter, we assess the linkages between food production and food secu-
rity. We start by examining the current trends in undernutrition and how inadequate
consumption of food and nutrients adversely impact human development. This is
particularly important in the case of pregnant women and children since they face
greater health risks from chronic undernutrition. We then take a closer look at how
climate change affects future global agricultural production and food security. To
conclude this chapter, we use the SIMPLE model to project food security outcomes
given future trends in agricultural productivity.

10.2 Undernutrition: Current Trends and Human Impacts

Although the majority of the world’s population has sufficient food to eat, roughly
one in ten people still experience chronic hunger. Figure 10.1 reports the number of
undernourished persons worldwide, based on inadequate daily caloric intake—a
rudimentary measure of nutritional status (FAO, 2015). We see that most of those
who are undernourished reside in Sub-Saharan Africa, Eastern and Southern Asia.
As of 2014, it was estimated that around 800 million persons were suffering from
caloric undernutrition. The figure shows significant improvements in global food
security over the past two decades, with the number of persons suffering from hun-
ger declining by around 209 million during the period 1990/92–2012/14—even as
global population increased strongly. From this figure, we see that the undernutri-
tion headcount has been drastically reduced in Eastern Asia—particularly —where
roughly 140 million persons have been lifted out of chronic hunger. Observers have
noted that strong economic growth has boosted the purchasing power of house-
holds, while sustained growth in agricultural productivity has contributed to the
dramatic rise in China’s domestic food production (Zhang, 2011; Zhao, Luo, Deng,
& Yan, 2008). Indeed, while one of four persons suffered from chronic hunger in
China in 1990/92, the prevalence of caloric undernutrition in China has now nearly
reached the global average (FAO, 2015).
However, little has changed in the number of persons experiencing of chronic
undernourishment in Southern Asia, and this number increased in Sub-Saharan
Africa over the period displayed in Fig. 10.1. The persistence of hunger in Southern
Asia, specifically in India has continued to puzzle researchers. Unlike China, nutri-
tional status in India has stagnated despite strong economic growth over the past
decade. Studies have linked these dismal trends in nutritional outcomes in India to
ineffective government policies, persistence of social castes which are correlated to
undernutrition as well as absence of female empowerment—particularly for moth-
ers. This is important since they are most influential in determining children’s nutri-
tional status (Haddad, 2009; Headey, Chiu, & Kadiyala, 2012). The dismal trends in
10.2 Undernutrition: Current Trends and Human Impacts 127

Fig. 10.1 Undernourished persons by region. Source: FAO Food Security Statistics (2015)

food security in Sub-Saharan Africa have been linked to frequent occurrences of

conflicts, droughts and diseases. Food aid has helped alleviate the incidence of
extreme famine within the region, but there remain many disincentives for invest-
ment in domestic agricultural production—including the impact of food aid on local
agricultural prices (Sanchez & Swaminathan, 2005). Without significant invest-
ments in agriculture, chronic hunger will persist in this region. Going forward,
strengthening food security in Sub-Saharan Africa, Eastern and Southern Asia will
be critical if the world is going to eliminate the global hunger.
Relative to the rest of the population, undernutrition places an especially impor-
tant burden on the health and well-being of young children and pregnant women.
Black et al. (2013) review the evidence on the diseases and mortality linked to
undernutrition in mothers and children. Looking at recent statistics from UNICEF,
WHO, and World Bank (2012), and Olofin et al. (2013), Black et al. highlighted that
the incidence of stunting (low weight-for-height due to poor nutrition) and wasting
(low height-for-age from severe starvation and/or disease) are widespread in Africa
and Asia. In Africa, stunting affects 37 % of children under the age of 5 or roughly
56 million. Although the incidence of stunting is lower in Asia (at around 27 %) the
number of children affected is even larger than in Africa at around 96 million—
most of whom reside in South-central Asia. Roughly 19 million children are afflicted
by severe wasting worldwide. Stunting and wasting are significant contributors to
child morbidity in children under 5 years (around 1.1 and 0.9 million deaths glob-
ally, respectively).
Inadequate intake of micronutrients, which increases the risks of diseases such as
blindness and chronic diarrhea, is also linked to child morbidity and reduced lifes-
pan. Specifically, deficiencies in Vitamin A and Zinc are responsible for 273,000
deaths globally and 4 % of morbidity in children under 5 years. Suboptimal breast-
feeding practices by mothers, such as discarding colostrum, early weaning and
128 10 Food Security and Nutrition

using pre-lacteal supplements such as sugar water are also linked to increase risk of
mortality and disease in infants. In fact, improper breastfeeding alone contributes to
around 804,000 deaths and 12 % of morbidity in children under 5 years. Maternal
undernutrition, reflected in low body weight and short stature, is tightly coupled
with health loss in children and infants [Stevens et al. (2012) as cited in Black et al.
(2013)]. Across the world, chronic maternal undernutrition is most prevalent in
Southern Asia, where 1 out of 5 women between the ages of 20 and 49 years have a
low body mass index. In Africa and Southeastern Asia, roughly 10 % of women are
underweight. Maternal short stature and underweight is linked to fetal growth
restriction which results in 817,000 deaths in children below 5 years [Olofin et al.
(2013) as cited in Black et al. (2013)]. It also increases the risk of mortality for
mothers during labor and delivery.
The long-term health impacts of maternal and child undernutrition are quite
severe. A comprehensive review by Victora et al. (2008) on maternal and child
undernutrition in low to middle income countries highlighted that nutritional status
of pregnant mothers and children less than 2 years old have long term consequences
for adult health and well-being. The authors looked at case studies which examined
the relationship among maternal size and nutrition, size of newborn baby, and child
as well as child development on adult well-being. They found strong evidence link-
ing low birth weight and short stature to stunted adult height. Adequate nutrition
during childhood is important for adult height as height differences between stunted
and non-stunted children tend to persists to adulthood [Sachdev et al. (2005) and
Martorell, Khan, and Schroeder (1994) as cited in Victora et al. (2008)]. The authors
also looked at evidence relating child brain development to an adult’s capacity for
learning and cognitive performance. Indeed, the authors found several case studies
which linked poor cognitive skills as well as weak educational performance on early
childhood stunting. Although evidence is scarce, the authors mentioned that proper
lung function, bone mass, and immune system development hinge on proper nutri-
tion during pregnancy and infanthood. Due to the long term impacts of undernutri-
tion, especially when it occurs during fetal gestation and early childhood, it is not
surprising that adequate nutrition is strongly linked to human capital development
and economic growth.

10.3 Climate Change Risks and Food Security

It is widely accepted that climate change will introduce further uncertainty in the
global farm and food system. Chapter 6 developed this theme in detail, highlighting
how changes in the patterns of precipitation and temperature might affect crop
yields and crop production. Of course, such disruptions will also affect food mar-
kets and ultimately alter food consumption and nutritional status. In this section, we
explore this linkage, examining the risks to food availability, stability, accessibility
as well as utilization owing to climate change.
10.3 Climate Change Risks and Food Security 129

Disruptions in crop yields and food availability Crop yields are directly linked
tofood production and availability; thus adverse yield impacts of climate change can
have negative consequences for future food security. In an effort to link yields to
future nutritional outcomes, Lobell et al. (2008) examined climate change risks
across selected regions and crops using a statistical crop model—estimated from
historical data—and climate projections from global circulation models. To account
for the contribution of each staple crops to nutritional status, the authors considered
the share in daily caloric supply by crop. Based on their estimates, staple crops
which are key sources of dietary calories and are at greatest risk from climate change
include wheat in South Asia, rice in Southeast Asia and maize in Southern Africa.
There is now an emerging body of literature analyzing the global impacts of climate
change on food production using global economic models of agriculture (Nelson
et al., 2014; Von Lampe et al., 2014). Temperature and precipitation projections
from global circulation models are first applied to global, gridded biophysical crop
models in order to compute new yield trends given a changing climate. Once calcu-
lated, these productivity shocks are then incorporated in food demand and supply
projections in global agricultural models. There are several studies which use this
integrated approach and look at changes in price and per capita food consumption
due to climate change (Fischer, Shah, Tubiello, & van Velhuizen, 2005; Schneider
et al., 2011; Tubiello & Fischer, 2007).
A widely cited study on the global impacts of climate change is offered by Nelson
et al. (2010). Starting with temperature and precipitation projections from two global
climate models, the authors used the DSSAT global gridded crop model to quantify
the crop yield risks from climate change. These climate induced changes were then
fed into the IMPACT partial equilibrium model of agriculture, to assess the food
security outcomes from climate change productivity shocks at global scale. The
results show that climate risks on crop production are most severe in Sub-Saharan
Africa and in South Asia particularly for wheat production systems. With declining
yields and production, food prices rise strongly and this has detrimental conse-
quences for food availability and consumption. Indeed, the authors predict a decline
in average, per capita cereal and meat consumption in the developing world of 22 %
and 10 %, respectively. The authors also incorporated further measures of nutritional
and food security outcomes in their analysis. Changes in average consumption were
converted to daily caloric availability. On average, they predict that daily caloric
supply in the developing region would fall by 16 % due to climate change. The
authors also suggested that climate risks will lead to a rise in the head count of child
malnutrition (by 25 million) most of whom reside in South Asia and Sub-Saharan
Africa. Towards the end of this chapter, we conduct a similar analysis and apply the
SIMPLE model to examine the food security issues due to climate change.
Instability in food supply due to extreme weather events In addition to long-
term impacts on global agricultural yields and production, climate change poses
risks for the stability of food supply due to the expected greater frequency of extreme
weather events. Historically, adverse weather such as droughts and floods have long
been a source of volatility in agricultural markets particularly in food insecure
regions of the world. For example, the incidence of droughts in Sub-Saharan
130 10 Food Security and Nutrition

Africa is historically tied to harvest failures as food production in the region relies
heavily on rainfed farming. Selected estimates of mortality from droughts and sub-
sequent conflicts amount to 2.6 million persons during the 1900s [Devereux and
Maxwell (2001) as cited in Baro and Deubel (2006)]. Even in recent years, food
crises from droughts have affected at least 17.0 million persons in Africa (Sapir,
Vos, Below, & Ponserre, 2012).
South Asia is also quite vulnerable to extreme weather. The region regularly
faces heavy flooding during the monsoon months as 80 % of annual rainfall occurs
during this season. Excess rainfall is further exacerbated by overlapping flood peaks,
siltation of natural waterways, deforestation as well as poor flood control infrastruc-
tures (Mirza, 2010). Indeed, over 300 flood-related disasters have been recorded in
the region during the period 1976–2005. Owing to the size of population residing in
this region, the number of those affected by flooding is quite significant. During the
said period, floods have affected the lives of nearly one billion persons and have
resulted in more than 2000 deaths annually (Shrestha & Takara, 2008). Floods can
also result in food crises particularly when they result in significant production
losses in the agricultural sector (Del Ninno, Dorosh, & Smith, 2003; Douglas, 2009).
Moving forward, the expected rise in frequency of extreme weather events will
lead to greater risks in food insecure regions (Svirejeva-Hopkins et al., 2013).
In Sub-Saharan Africa, long-term rises in temperatures will likely lead to lower
precipitation and dampened ground water recharge rates, thereby resulting in
increased aridity. Regional projections show increased exposures to flooding and
drought in South Asia. At the moment, households residing in food insecure regions
are already facing significant risks from weather-related food shortages due to high
incidence of poverty, weak governance and poor institutions. Given increased vola-
tility in the coming decades, there is a widespread call-to-action regarding the
implementation of climate adaptation policies which will strengthen the resilience
of local economic and agricultural systems to adverse weather shocks (Collier,
Conway, & Venables, 2008; Stringer et al., 2009; Svirejeva-Hopkins et al., 2013).
Risks to food accessibility Adequate access to food hinges on affordable food
prices and sustained increases in household income. Those who live in poverty are
quite vulnerable to spikes in food prices since they allocate a significant share of
their income on food purchases and have limited physical and financial capacity to
cope with food crises. Crop failures and damages due to increased intensity and
frequency of extreme weather events have been linked to recent food price fluctua-
tions (Tadesse, Algieri, Kalkuhl, & von Braun, 2014). Climate change also poses
further risks to food accessibility by threatening the livelihoods of the poor and
intensifying the incidence of poverty. Indeed, poverty is expected to increase for the
urban poor, especially for wage laborers, who are most vulnerable to climate related
spikes in food prices (Ahmed, Diffenbaugh, & Hertel, 2009).
Barriers to food utilization Climate risks to human health from increased inci-
dence of water-, vector- and food-borne diseases threatens the proper utilization of
food and hinders overall improvements in nutritional attainment. Transmission of
diseases is tied to temperature and precipitation patterns. Indeed, it is well docu-
mented that increased precipitation and periodic flooding lead to contamination of
10.4 Agricultural Productivity and Future Food Security: A SIMPLE Application 131

water supplies and virulent spread of infection water- and vector-borne diseases
such as cholera, diarrhea, dysentery and dengue (Miller, Feachem, & Drasar, 1985;
Mirza, 2010). In addition to its direct impact on mortality, extreme heat and mois-
ture due to climate change also increases the incidence of food poisoning from sal-
monella and mycotoxins (McMichael & Lindgren, 2011; Paterson & Lima, 2010).
Food stockpiled by the household can also be contaminated by pests and blight.
Although temperature and precipitation is closely linked to human health outcomes,
definitively attributing these to long-term changes in climate remains a challenge
since observed disease outbreaks are often tied to natural short-run variation in
weather patterns (McMichael & Lindgren, 2011). Nonetheless, projections from
mathematical models of disease transmission—particularly for malaria and den-
gue—indicate greater spread of disease and increased risk of infection in the face of
warmer temperatures and increased moisture (Altizer, Ostfeld, Johnson, Kutz, &
Harvell, 2013). Poor socio-economic conditions will further exacerbate the risk of
exposure and infection to these diseases, particularly if adequate health infrastruc-
ture is absent.

10.4 Agricultural Productivity and Future Food Security:

A SIMPLE Application1

Food availability has been greatly enhanced by productivity gains in the agricultural
sector. Continuation of such trends will be critical in strengthening food security
between now and mid-century, as population, incomes and biofuel use continue to
grow. However, there is uncertainty about how agricultural productivity will evolve
in the coming years. As discussed in Chap. 3, the growth rate of total factor produc-
tivity—a measure of the growth in aggregate output relative to an index of all
inputs—in the global agricultural sector actually rose over the past two decades
(Ludena, Hertel, Preckel, Foster, & Nin, 2007). However, there are concerns on
some fronts that crop yields for key staple grains may be reaching their biophysical
limits in some regions (Alston, Beddow, & Pardey, 2009). This could have an
adverse effect on global food availability and prices. The future trajectory of crop
yields will also be affected by climate change, although the precise impacts are
uncertain and spatially heterogeneous (recall Chap. 6). Depending on location, the
temperature and precipitation impacts of climate change may enhance or dampen
crop yield growth. There is also the potential for crop yields to be enhanced via the
fertilization effect of rising CO2 concentrations in the atmosphere (Rosenzweig
et al., 2014).
In this section, we develop a set of projections of global food security and look
at how these are influenced by trends in agricultural productivity and the complexi-
ties introduced by climate change. We focus on the most rudimentary measure of
nutritional outcome; adequacy of dietary caloric consumption. Following FAO

1
This section draws heavily on Baldos and Hertel (2014) albeit with updated projections.
132 10 Food Security and Nutrition

(2010), we characterize the distribution of dietary energy consumption within each

region using information on the average caloric consumption, the minimum dietary
energy requirement and the standard deviation of caloric consumption. This frame-
work allows us to calculate three key measures of food security in addition to aver-
age per capita caloric consumption. These include undernutrition incidence,
headcount and the undernutrition gap. The incidence of undernutrition measures the
prevalence of caloric hunger by reporting the fraction of population whose daily
dietary energy intake is below the minimum requirement. Undernutrition count—
calculated from undernutrition incidence and population—shows the number of
persons who are suffering from caloric hunger. Finally, the undernutrition gap rep-
resents the intensity of caloric deprivation and reflects the average dietary energy
deficit that an undernourished person needs to close in order to satisfy the minimum
caloric requirement (FAO, 2012).2,3
Figure 10.2 illustrates how increases in the average dietary energy intake—
owing either to increased incomes or reduced prices—shifts the distribution of

Fig. 10.2 Distribution of caloric consumption for Sub-Saharan Africa in 2006 (solid black line)
and in 2050 (dashed blue line). The dotted line represent the minimum required daily caloric
intake. Source: Authors’ calculations

2
Mathematically, the undernutrition incidence and gap are equivalent to the poverty index and gap
measures as proposed by Foster, Greer, and Thorbecke (1984). Given this, we can use the concept
of poverty-growth elasticities in our model implementation to link these measures to the changes
in average per capita dietary energy intake. Widely used in the poverty literature, these growth
elasticities measure the percent changes in the indices of poverty and poverty gap given a one
percent change in average per capita income (Bourguignon, 2003; Lopez and Serven, 2006).
3
To account for dietary upgrading towards healthier diets (i.e. lower calories per gram of food), we
directly link changes in caloric content of food to per capita incomes. Higher incomes may result
in fewer calories per dollar spent on a given food type—as observed in crops and processed
foods—as well as consumers’ shift to a leaner and higher quality diet.
10.4 Agricultural Productivity and Future Food Security: A SIMPLE Application 133

caloric energy within a region. Specifically, it plots the predicted caloric distribution
for Sub-Saharan Africa in the years 2006 (solid black line) and 2050 (dashed blue
line). The vertical dotted line within the 2006 distribution represents the minimum
caloric requirement. The area to the left of this line is the fraction of the population
which is undernourished, having caloric consumption below the required amount.
Going forward to 2050, rising per capita incomes lead to increased food and caloric
consumption. Greater caloric intake shifts the distribution, resulting in a thinner tail
to the left of the unchanging minimum caloric requirement. The undernutrition inci-
dence in 2050 is the area bounded by the minimum dietary energy requirement and
the new caloric distribution curve. Once calculated, we can use this information to
derive the undernutrition headcount and gap.
We implement a series of carefully designed scenarios to assess how future
global food security will be affected by agricultural productivity and climate change.
Following our future simulations in Chap. 7, we start by establishing the baseline
scenario for 2050 (‘Future Baseline’) which is induced by growing population, per
capita incomes, biofuel production, as well as total factor productivity in the crops,
livestock and processed food sectors (Table 10.1). We then explore the food security
impacts of the stagnation of agricultural productivity (‘Demand only’ scenario).
Specifically, we only apply the demand shocks outlined in the preceding scenario,
omitting productivity growth and thereby highlighting its role in driving future
nutritional outcomes. We then assess how global food security will be affected by
climate change (two last columns of Table 10.1). In light of the great uncertainty
about the effects of elevated CO2 concentrations in the atmosphere on plant growth,
we consider crop yield effects from climate change under two alternative (bound-
ing) scenarios: (a) when there is no CO2 fertilization (‘No CO2 fert.’ scenario) and
(b) when there is the full CO2 fertilization effect (‘CO2 fert.’ Scenario), as predicted
by the crop model. Looking at the climate change yield shocks, we see that crop
production in developing regions, notably in South and South East Asia as well
as Sub-Saharan Africa, appears to be quite vulnerable to climate change.
Table 10.2 summarizes our projections of future food security. In reporting our
results, we focus on three key regions wherein historically chronic malnutrition is
prevalent. These include: Sub-Saharan Africa, China/Mongolia and South Asia. We
begin with our baseline scenario for 2050 wherein we report the future values of
selected food security outcomes when both demand and supply drivers are imple-
mented. In the future, our baseline suggests that population and agricultural produc-
tivity growth will be slower than in the 10-year historical period, whereas global
biofuel use and per capita incomes continue their steady rise. Our results show sig-
nificant improvements in nutritional outcomes relative to 2006. Globally, average
dietary energy intake increases by 24 % while the prevalence and average depth of
undernutrition further decrease by 83 % and 28 %, respectively. Sharp rises in aver-
age dietary energy consumption are observed South Asia and China/Mongolia—
regions with strong per capita income growth rates—while we see notable reductions
in the incidence of undernutrition in Sub-Saharan Africa. Given these improvements,
we see a significant reduction (around 77 %) in the global undernutrition count
which falls by 585 million between 2006 and 2050, despite increasing population.
 134

Table 10.1 Future growth rates of exogenous variables (in % per annum rates)
Total factor productivity Climate change yield impacts
Regions Population Per capita income Biofuel Crops Livestock Processed foods CO2 fert No CO2 fert
Eastern Europe −0.42 4.83 1.04 0.39 0.02
North Africa 1.06 3.58 −0.30 −0.04 −0.32
Sub-Saharan Africa 2.45 3.78 0.43 0.14 −0.16
South America 0.64 2.51 2.64 0.22 −0.17
Australia/New Zealand 1.00 1.75 0.42 0.07 −0.29
European Union+ 0.09 1.4 0.42 0.31 0.02
South Asia 0.83 4.84 1.71 0.36 −0.36
Central America 0.94 2.55 2.64 0.22 −0.17
Southern Africa 0.57 2.76 0.43 0.14 −0.16
Southeast Asia 0.79 3.62 2.38 0.4 −0.35
Canada/U.S. 0.66 1.06 0.42 0.23 −0.15
China/Mongolia 0.09 5.55 2.38 0.27 −0.07
Middle East 1.02 2.38 −0.30 −0.04 −0.32
10

Japan/Korea −0.22 2.17 0.42 0.07 −0.29

Central Asia 1.47 5.03 1.04 0.39 0.02
World 0.80 2.09 5.80 0.94 0.89
Sources: From left to right—UN World Population Prospects (2013), Fouré, Bénassy-Quéré, and Fontagné (2013), IEA (2008, 2012), Ludena et al. (2007), and
Müller et al. (2010)
Food Security and Nutrition
Table 10.2 Selected food security statistics
Future scenarios: 2050
Climate change
Demand only No CO2 fert. CO2 fert.
Regions Base data: 2006 Base scenario In Δ relative to ‘Base Scenario’
Average dietary energy consumption (in kcal/capita/day)
World 2761 3410 −586 −48 77
Sub-Saharan Africa 2110 2809 −478 −66 108
South Asia 2341 3502 −705 −67 108
China/Mongolia 2989 4140 −908 −44 71
Undernutrition index (in %)
World 12.0 2.0 6.0 0.2 −0.4
Sub-Saharan Africa 23.5 2.4 9.9 0.6 −0.8
South Asia 20.2 1.2 6.0 0.3 −0.3
China/Mongolia 9.6 0.9 5.0 0.1 −0.1
Undernutrition gap (in kcal/capita/day)
World 235 168 34 1 −1
Sub-Saharan Africa 207 137 40 4 −6
South Asia 252 163 42 3 −5
China/Mongolia 250 184 47 2 −3
Undernutrition count (in million)
World 764 179 554 25 −33
Sub-Saharan Africa 158 47 193 13 −16
South Asia 302 26 129 5 −7
10.4 Agricultural Productivity and Future Food Security: A SIMPLE Application

China/Mongolia 128 12 70 1 −2
Source: Authors’ calculations
135
136 10 Food Security and Nutrition

We now consider the ‘Demand only’ scenario wherein agricultural productivity

growth is ignored. Note that the subsequent columns of Table 10.2 report the differ-
ences in food security outcomes relative to our baseline scenario in 2050. From this
column of results, we see that improvements in nutritional outcomes are severely
dampened if agricultural productivity stagnates. As discussed in the previous section,
productivity growth does not only increase food availability but it also improves food
accessibility by keeping food prices low. Without productivity growth, food prices
will sharply rise and improvements in food security outcomes hinge on income
growth alone. The prevalence of undernutrition exceeds that in our future baseline by
more than four times (2.0 % vs. 8.0 % of the world’s population). The undernutrition
incidence and gap in Sub-Saharan Africa falls more slowly relative to the baseline,
highlighting the importance of agricultural productivity in strengthening food secu-
rity in this region. In this scenario, the decline in global malnutrition count between
2050 and 2006 is only around 210 million people. However, across regions the
increase in malnutrition count will be higher in the poorest countries, where the aver-
age caloric intake is low and the response to higher prices is most accentuated.
We now look at the changes in nutritional outcomes in light of potential crop yield
impacts of climate change. We consider both yield shocks generated in the presence
or absence of CO2 fertilization using the yield estimates from Müller et al. (2010).
Without CO2 fertilization (‘No CO2 fert.’), crop yields in most regions will be
adversely affected by the temperature and precipitation impacts from climate change.
With relatively lower crop yields, the reduction in crop prices from projected crop
TFP growth will be slightly dampened. At a glance, we see that the gains in food
security from 2006 to 2050 are reduced relative to our baseline scenario. At both
global and regional levels, the change in average dietary energy consumption (DEC)
increases and average depth of malnutrition are negligible. However, the relative
reduction in average DEC is greater (at least 38 % more than the global reduction) in
Sub-Saharan Africa and South Asia wherein consumers are more responsive to food
prices. The gravity of climate change impacts for food security is quite evident if we
look at the prevalence of malnutrition. At the global level, malnutrition incidence
increases by about 10 % relative to the baseline scenario. Across regions, the increase
in the prevalence of malnutrition is more than 25 % in Sub-Saharan Africa and South
Asia, relative to the baseline. Coupled with the steady growth in population, global
malnutrition count actually increases under this scenario (by about 25 million, rela-
tive to the baseline) and most reside in Sub-Saharan Africa and South Asia.
When the effects of CO2 fertilization are added in (‘CO2 fert.’), crop yields are
higher in most regions of the world, resulting in slightly lower crop prices and fur-
ther improvements in food security outcomes. However, similar to the previous case,
we do not observe these gains explicitly if we look directly at the average DECs but
relative to the increase in the global average, we see stronger gains (by more than 37
%) in average dietary energy intake in Sub-Saharan Africa, Central Asia and South
Asia, relative to the baseline. With CO2 fertilization in place, the global malnutrition
incidence further declines by 20 % relative to our baseline. Regions which benefit
most from reduced malnutrition headcount under this scenario are Sub-Saharan
Africa, South East Asia and South Asia. With CO2 fertilization effects, the number
References 137

of malnourished persons globally further declines by around 33 million relative to

the baseline.
These projections illustrate the uncertainty posed by climate change on global
food security as it may enhance or dampen improvements in nutritional outcomes in
the future depending on the strength of the yield impacts of CO2 fertilization. Future
food security will still hinge on socio-economic conditions as well as productivity
trends in the agricultural sector. However, climate change will become a relevant
driver of nutritional outcomes especially for those residing in tropical regions of the
world where chronic malnutrition is currently most prevalent.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Ahmed, S. A., Diffenbaugh, N. S., & Hertel, T. W. (2009). Climate volatility deepens poverty
vulnerability in developing countries. Environmental Research Letters, 4(3), 034004. http://
doi.org/10.1088/1748-9326/4/3/034004.
Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050: The 2012 revision
(Working Paper No. 12-03). Rome, Italy: Food and Agriculture Organisation of the United
Nations.
Alston, J. M., Beddow, J. M., & Pardey, P. G. (2009). Agricultural research, productivity, and food
prices in the long run. Science, 325(5945), 1209–1210. http://doi.org/10.1126/science.1170451.
Altizer, S., Ostfeld, R. S., Johnson, P. T. J., Kutz, S., & Harvell, C. D. (2013). Climate change and
infectious diseases: From evidence to a predictive framework. Science, 341(6145), 514–519.
http://doi.org/10.1126/science.1239401.
Baldos, U. L. C., & Hertel, T. W. (2014). Global food security in 2050: The role of agricultural
productivity and climate change. Australian Journal of Agricultural and Resource Economics,
58(4), 554–570. http://doi.org/10.1111/1467-8489.12048.
Baro, M., & Deubel, T. F. (2006). Persistent hunger: Perspectives on vulnerability, famine, and
food security in Sub-Saharan Africa. Annual Review of Anthropology, 35(1), 521–538. http://
doi.org/10.1146/annurev.anthro.35.081705.123224.
Black, R. E., Victora, C. G., Walker, S. P., Bhutta, Z. A., Christian, P., de Onis, M., … Uauy, R.
(2013). Maternal and child undernutrition and overweight in low-income and middle-income
countries. The Lancet, 382(9890), 427–451. http://doi.org/http://dx.doi.org/10.1016/
S0140-6736(13)60937-X.
Bourguignon, F. (2003). The growth elasticity of poverty reduction : explaining heterogeneity
across countries and time periods. In T. S. Eicher & S. J. Turnovsky (Eds.), Inequality and
Growth: Theory and Policy Implications (pp. 3–26). MIT Press. Retrieved from http://ideas.
repec.org/p/del/abcdef/2002-03.html.
Collier, P., Conway, G., & Venables, T. (2008). Climate change and Africa. Oxford Review of
Economic Policy, 24(2), 337–353. http://doi.org/10.1093/oxrep/grn019.
Del Ninno, C., Dorosh, P. A., & Smith, L. C. (2003). Public policy, markets and household coping
strategies in Bangladesh: Avoiding a food security crisis following the 1998 floods. World
Development, 31(7), 1221–1238. http://doi.org/10.1016/S0305-750X(03)00071-8.
Devereux, S., & Maxwell, S. (Eds.). (2001). Food security in Sub-Saharan Africa. London, UK:
Practical Action.
138 10 Food Security and Nutrition

Douglas, I. (2009). Climate change, flooding and food security in south Asia. Food Security, 1(2),
127–136. http://doi.org/10.1007/s12571-009-0015-1.
FAO. (2010). FAO statistical yearbook. Rome, Italy: Food and Agriculture Organization of the
United Nations.
FAO. (2012). FAO food security statistics. Retrieved from http://www.fao.org/economic/ess/en/.
Fischer, G., Shah, M., Tubiello, F. N., & van Velhuizen, H. (2005). Socio-economic and climate
change impacts on agriculture: An integrated assessment, 1990–2080. Philosophical
Transactions of the Royal Society B: Biological Sciences, 360(1463), 2067–2083. http://doi.
org/10.1098/rstb.2005.1744.
Food and Agriculture Organisation of the UN. (2015). FAO food security statistics. Rome, Italy.
Retrieved from http://www.fao.org/economic/ess/ess-fs/en/.
Food and Agriculture Organisation of the United Nations. (1996). Rome declaration on world food
security and world food summit plan of action. Presented at the world food summit, Rome,
Italy.
Food and Agriculture Organisation of the United Nations. (2009). Declaration of the world summit
on food security. Presented at the world summit on food security, Rome, Italy.
Foster, J., Greer, J., & Thorbecke, E. (1984). A class of decomposable poverty measures.
Econometrica, 52(3), 761–766. http://doi.org/10.2307/1913475.
Fouré, J., Bénassy-Quéré, A., & Fontagné, L. (2013). Modelling the world economy at the 2050
horizon. Economics of Transition, 21(4), 617–654. http://doi.org/10.1111/ecot.12023.
Haddad, L. (2009). Lifting the curse: Overcoming persistent undernutrition in India. IDS Bulletin,
40(4), 1–8. http://doi.org/10.1111/j.1759-5436.2009.00052.x.
Headey, D., Chiu, A., & Kadiyala, S. (2012). Agriculture’s role in the Indian enigma: Help or
hindrance to the crisis of undernutrition? Food Security, 4(1), 87–102. http://doi.org/10.1007/
s12571-011-0161-0.
IEA. (2008). World energy outlook. OECD. Retrieved from http://www.oecd-ilibrary.org/content/
book/weo-2008-en.
IEA. (2012). World energy outlook. OECD. Retrieved from http://www.oecd-ilibrary.org/content/
book/weo-2012-en.
Lobell, D. B., Burke, M. B., Tebaldi, C., Mastrandrea, M. D., Falcon, W. P., & Naylor, R. L.
(2008). Prioritizing Climate Change Adaptation Needs for Food Security in 2030. Science,
319(5863), 607–610. http://doi.org/10.1126/science.1152339.
Lopez, H., & Serven, L. (2006). A normal relationship ? Poverty, growth, and inequality (Policy
Research Working Paper Series No. 3814). The World Bank. Retrieved from http://ideas.repec.
org/p/wbk/wbrwps/3814.html.
Ludena, C. E., Hertel, T. W., Preckel, P. V., Foster, K., & Nin, A. (2007). Productivity growth and
convergence in crop, ruminant, and nonruminant production: Measurement and forecasts.
Agricultural Economics, 37(1), 1–17. http://doi.org/10.1111/j.1574-0862.2007.00218.x.
Martorell, R., Khan, L., & Schroeder, D. (1994). Reversibility of stunting: Epidemiological find-
ings in children from developing countries. European Journal of Clinical Nutrition, 48(Suppl
1), S45–S57.
McMichael, A. J., & Lindgren, E. (2011). Climate change: Present and future risks to health, and
necessary responses. Journal of Internal Medicine, 270(5), 401–413. http://doi.
org/10.1111/j.1365-2796.2011.02415.x.
Miller, C., Feachem, R., & Drasar, B. (1985). Cholera epidemiology in developed and developing
countries: New thoughts on transmission, seasonality, and control. The Lancet, 325(8423),
261–263. http://doi.org/10.1016/S0140-6736(85)91036-0.
Mirza, M. M. Q. (2010). Climate change, flooding in South Asia and implications. Regional
Environmental Change, 11(1), 95–107. http://doi.org/10.1007/s10113-010-0184-7.
Müller, C., Bondeau, A., Popp, A., Waha, K., & Fader, M. (2010). Climate change impacts on
agricultural yields: Background note to the World Development Report 2010 (Background
note). Germany: Potsdam Institute for Climate Impact Research.
References 139

Nelson, G., Rosegrant, M. W., Koo, J., Robertson, R., Sulser, T., Zhu, T., … Lee, D. (2010). The
costs of agricultural adaptation to climate change (Discussion paper no. 4) (p. 62). Washington,
DC: The World Bank.
Nelson, G. C., van der Mensbrugghe, D., Ahammad, H., Blanc, E., Calvin, K., Hasegawa, T., …
Willenbockel, D. (2014). Agriculture and climate change in global scenarios: Why don’t the
models agree. Agricultural Economics, 45(1), 85–101. http://doi.org/10.1111/agec.12091.
Olofin, I., McDonald, C. M., Ezzati, M., Flaxman, S., Black, R. E., Fawzi, W. W., … Nutrition
Impact Model Study (anthropometry cohort pooling). (2013). Associations of suboptimal
growth with all-cause and cause-specific mortality in children under five years: A pooled anal-
ysis of ten prospective studies. PLoS One, 8(5), e64636. http://doi.org/10.1371/journal.
pone.0064636.
Paterson, R. R. M., & Lima, N. (2010). How will climate change affect mycotoxins in food? Food
Research International, 43(7), 1902–1914. http://doi.org/10.1016/j.foodres.2009.07.010.
Rosenzweig, C., Elliott, J., Deryng, D., Ruane, A. C., Müller, C., Arneth, A., … Jones, J. W. (2014).
Assessing agricultural risks of climate change in the 21st century in a global gridded crop
model intercomparison. Proceedings of the National Academy of Sciences of the United States
of America, 111(9), 3268–3273. http://doi.org/10.1073/pnas.1222463110.
Sachdev, H. S., Fall, C. H., Osmond, C., Lakshmy, R., Dey Biswas, S. K., Leary, S. D., …
Bhargava, S. K. (2005). Anthropometric indicators of body composition in young adults:
Relation to size at birth and serial measurements of body mass index in childhood in the New
Delhi birth cohort. The American Journal of Clinical Nutrition, 82(2), 456–466.
Sanchez, P. A., & Swaminathan, M. S. (2005). Cutting world hunger in half. Science, 307(5708),
357–359. http://doi.org/10.1126/science.1109057.
Sapir, D. G., Vos, F., Below, R., & Ponserre, S. (2012). Annual disaster statistical review 2011.
Brussels, Belgium: CRED.
Schneider, U. A., Havlík, P., Schmid, E., Valin, H., Mosnier, A., Obersteiner, M., … Fritz, S.
(2011). Impacts of population growth, economic development, and technical change on global
food production and consumption. Agricultural Systems, 104(2), 204–215. http://doi.
org/10.1016/j.agsy.2010.11.003.
Shrestha, M. S., & Takara, K. (2008). Impacts of floods in south Asia. Journal of South Asia
Disaster Study, 1(1), 85–106.
Stevens, G., Singh, G., Lu, Y., Danaei, G., Lin, J., Finucane, M., … Global Burden of Metabolic
Risk Factors of Chronic Diseases Collaborating Group (Body Mass Index). (2012). National,
regional, and global trends in adult overweight and obesity prevalences. Population Health
Metrics, 10(1), 22. http://doi.org/10.1186/1478-7954-10-22.
Stringer, L. C., Dyer, J. C., Reed, M. S., Dougill, A. J., Twyman, C., & Mkwambisi, D. (2009).
Adaptations to climate change, drought and desertification: Local insights to enhance policy in
southern Africa. Environmental Science & Policy, 12(7), 748–765. http://doi.org/10.1016/j.
envsci.2009.04.002.
Svirejeva-Hopkins, A., Schwan, S., Schewe, J., Hare, B., Frieler, K., Rocha, M., … Menon, A.
(2013). Turn down the heat : Climate extremes, regional impacts, and the case for resilience—
Full report (No. 78424) (pp. 1–254). The World Bank. Retrieved from http://documents.
worldbank.org/curated/en/2013/06/17862361/turn-down-heat-climate-extremes-
regional-impacts-case-resilience-full-report.
Tadesse, G., Algieri, B., Kalkuhl, M., & von Braun, J. (2014). Drivers and triggers of international
food price spikes and volatility. Food Policy, 47(0), 117–128. http://doi.org/10.1016/j.
foodpol.2013.08.014.
Tubiello, F. N., & Fischer, G. (2007). Reducing climate change impacts on agriculture: Global and
regional effects of mitigation, 2000–2080. Technological Forecasting and Social Change,
74(7), 1030–1056. http://doi.org/10.1016/j.techfore.2006.05.027.
UN Population Division. (2013). World population prospects: The 2012 revision. New York, NY:
Population Division of the Department of Economic and Social Affairs of the United Nations
Secretariat. Retrieved from http://esa.un.org/unpd/wpp/index.htm.
140 10 Food Security and Nutrition

United Nations Children’s Fund, World Health Organization, The World Bank. (2012). UNICEF
WHO-World Bank Joint Child Malnutrition Estimates. New York: UNICEF; Geneva: WHO;
Washington, DC: The World Bank.
Victora, C. G., Adair, L., Fall, C., Hallal, P. C., Martorell, R., Richter, L., & Sachdev, H. S. (2008).
Maternal and child undernutrition: Consequences for adult health and human capital. The
Lancet, 371(9609), 340–357. http://doi.org/10.1016/S0140-6736(07)61692-4.
Von Lampe, M., Willenbockel, D., Ahammad, H., Blanc, E., Cai, Y., Calvin, K., … van Meijl, H.
(2014). Why do global long-term scenarios for agriculture differ? An overview of the AgMIP
Global Economic Model Intercomparison. Agricultural Economics, 45(1), 3–20. http://doi.
org/10.1111/agec.12086.
Zhang, J. (2011). China’s success in increasing per capita food production. Journal of Experimental
Botany, 62, 3707–3711.
Zhao, J., Luo, Q., Deng, H., & Yan, Y. (2008). Opportunities and challenges of sustainable agricul-
tural development in China. Philosophical Transactions of the Royal Society B: Biological
Sciences, 363(1492), 893–904. http://doi.org/10.1098/rstb.2007.2190.
Chapter 11
Global Change and the Food System in 2050

11.1 Looking Back to Understand the Future1

Estimates of future agricultural output and crop land requirements vary widely—in
part due to the many competing factors which influence the global food system
(Schmitz et al., 2014). To account for all these factors, many alternative, computa-
tional models of global agriculture have been developed. However, little has been
done to evaluate whether these models perform well over historical periods.
Therefore, before looking forward in time, it is important that we should first look
backwards and see if the SIMPLE model used throughout this book can be used to
explain the past. Perhaps most interesting will be an investigation of where the
SIMPLE model falls short. This information will be valuable when interpreting
results; we should not over-emphasize findings obtained from features of the model
which perform poorly in back-casting exercises. It is also a useful guide to future
model improvements. Finally, we will see that this historical period can provide an
excellent ‘laboratory’ for uncovering the ‘error signatures’ of commonly used models
of the global food and environmental system which omit fundamental mechanisms
required to replicate the past and are critical for future projections.

1
This section draws heavily on Baldos and Hertel (2013).

© Springer International Publishing Switzerland 2016 141

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0_11
142 11 Global Change and the Food System in 2050

11.1.1 Global Validation of SIMPLE over

the 1961–2006 Period

Since the SIMPLE model has been used in this book to make forward looking pro-
jections from 2006 to 2050, we choose to evaluate the model over a comparable
period of time looking backwards—in this case from 2006 to 1961.2 The most obvi-
ous metrics involve comparing model predictions to observed changes in the fol-
lowing global scale variables: (a) crop production, (b) crop price, (c) cropland area,
and (d) average crop yield. To derive these endogenous changes in SIMPLE, the
model is perturbed using the main exogenous drivers of global agriculture during
this historical period, including: population and per capita income (by demand
region) and total factor productivity (TFP) for crops (by supply region), livestock
and food processing (by demand region).3
The values for the exogenous drivers of the global food economy are reported in
Table 11.1. From the table, we see that population and per capita incomes grew
steadily during this historical period. Strong growth in population can be observed
in the lower high, upper middle (such as Brazil) and low income regions (such as
India).4 Likewise, steady growth in per capita income is observed with the lower
middle income region (including China) showing sharply higher per capita income
growth (4.3 % per annum). Crop supplies are mainly driven by the growth in TFP
which is the key measure of productivity improvement in the model. Crop sector
TFP grew by more than 1.15 % per annum, with the exception of Sub-Saharan
Africa where it grew by 0.9 % annually. With regard to the livestock sector, strong
TFP growth in the lower middle income region is observed. In contrast, livestock
TFP growth in the low income region grew by only 0.2 % per annum. Due to lack
of reliable regional estimates, a uniform rate of TFP growth in the processed food
sector is imposed across all regions.

2
Since it is confusing to report results going backwards in time, and since we do not have a full set
of data for parameterizing the model in 1961, we first simulate the model back to 1961 to establish
a benchmark economy for that year, and then we simulate the model forward again (reversing the
shocks) to reach 2006. The results from the latter simulation are what we report here.
3
We use in this section the original version of SIMPLE in which there a just five demand regions
and seven crop supply regions. Since we also assume perfectly integrated world markets (an
assumption which will be relaxed in the next section), it is not necessary to match demand with
supply regions. Later we will discuss a variant of the model which features segmented markets. In
this case, supply and demand are matched for each 15 geographic regions.
4
This classification of countries by per capita income follows World Bank conventions. Over time,
countries may graduate to higher income levels. For our purposes, we focus on their classification
as of 2001.
11.1 Looking Back to Understand the Future 143

Table 11.1 Growth rates for key exogenous variables: 1961–2006

TFP
Income Per capita Processed Geographic
regions Population income Livestock food Crops regions
Up Higher 0.79 2.62 0.92 1.89 East Asia & Pacific
Low Higher 2.64 2.69 0.92 1.78 Europe & Central
Asia
1.58 Latin America &
the Caribbean
Up Middle 2.07 1.71 0.75 0.89 2.19 Middle East &
North Africa
Low Middle 1.71 4.25 2.20 1.65 North America
1.15 South Asia
Low 2.26 2.35 0.16 0.91 Sub-Saharan
Africa
Sources: From left to right—UN World Population Prospects (2013), World Bank Development
Indicators (2013), Ludena et al. (2007), Griffith et al. (2004) and Fuglie et al. (2012)

Results from our global validation exercise are reported in Fig. 11.1.5 Based on
the figure, we observe that SIMPLE slightly overstates the global change in crop
production (top left panel, compare model total with actual data) over the 1961–
2006 period (204 % vs. 196 %). The model also slightly understates the historical
decline in crop price (25 % vs. 29 %) (bottom right panel). SIMPLE does a very
good job in partitioning supply growth between the intensive and extensive margins,
with changes in global cropland (top right panel) and global average crop yield (bot-
tom left panel) (16 % and 161 %, respectively) which are just slightly above the
observed values due to the higher level of global output predicted by the model.
Overall, these global results are encouraging since it demonstrates that SIMPLE
incorporates the most important drivers and economic responses governing long-
run changes in agriculture, at the global scale.
The third set of bars in Fig. 11.1 provides a decomposition of the historical
changes in global crop production, average yields, crop land and price, according to
the individual contributions of the historical drivers highlighted in Table 11.1. From

5
It should be noted that this version of SIMPLE was ‘tuned’ on two key dimensions of the global
food economy over the 5 year period from 2001 to 2006, prior to undertaking this 45 year historical
validation experiment. First, the unobserved intensification parameters in the livestock sector is
calibrated due to lack of robust estimates in the literature. We focus on the high income region,
which is deemed to be most representative of future developments in the livestock industry and we
then select the substitution parameter which best fits the data on feed input use for this region, over
the 2001–2006 period. This same value is subsequently assigned to all demand regions. Secondly,
the regression estimates of the income and price elasticities (see Appendix C) are adjusted by re-
estimating the linear regressions of the demand elasticities with per capita incomes using deflated
per capita incomes (divided by a factor of 4). In the initial calibration effort, the simulated change
in global crop demand for food (10.9 %) is nearly one-quarter greater than the historical change
(around 8.8 %). This adjustment closes this gap by dampening the magnitude of the regression
intercepts while maintaining the values of the regression slopes.
144 11 Global Change and the Food System in 2050

Fig. 11.1 Validating historical changes at the global level: 1961–2006

the decomposition of crop output, we can see that it was population (in red)—and
not income (in orange)—which was the main driver of historical growth in crop
demand. This is followed in significance by total factor productivity (in green). TFP
is an important source of yield growth and it helped dampen crop land use expansion
during this period. More importantly, the TFP growth has been the key driver behind
the historical reduction in food prices, which otherwise would have risen over this
period due to the pressures of increasing income and population worldwide.

11.1.2 Identifying the Error Signatures of Other Models

Existing global agricultural models produce widely divergent projections of global

crop production and land use in 2050 (Schmitz et al., 2014; Smith et al., 2010). This
is hardly surprising, given the widely varying assumptions imbedded in the models.
11.1 Looking Back to Understand the Future 145

Some of these differences may be inconsequential for simulating global food out-
comes, while others may be critically important. Absent a laboratory in which to
test these alternative assumptions, it is impossible to know which model results are
reliable. For this reason, it is valuable to have a standard set of validation experi-
ments against which to evaluate model performance, test new features, and set
future research priorities. Here, we do this, using the 1961–2006 period discussed
above. We consider, in turn, alternate assumptions underpinning different global
agricultural models which may be important for determining global production,
yield growth, land use and crop prices. We are particularly interested in whether or
not these assumptions result in certain ‘error signatures’ which might be used to
understand differences in long run projections from these different models.
The restrictive assumptions which we have chosen are designed to highlight
shortcomings in many of the existing global models and assess their relative signifi-
cance. They include: exogenous per capita food consumption (E1), fixed price and
income elasticities of demand for food (E2), short- to medium run supply elastici-
ties (E3), the absence of endogenous intensification of crop production (E4), and
historical, trend-based yield projections (E5). To illustrate the potential for interac-
tions amongst these restrictions, two experiments (E6.a and E6.b) are considered
which include multiple elements of the earlier experiments designed to reflect com-
binations of assumptions sometimes found in biophysical and in economic models
of global agricultural land use.
Figure 11.2 summarizes the results from these restricted experiments. In every
case, the key historical drivers of change: population, incomeand total factor pro-
ductivity growth, are identical to the historical baseline. The only difference lies in
the underlying economic behavior. A good starting point is the restriction in the way
crop consumption is modeled. Some integrated assessment models simply treat per
capita food consumption as being exogenously specified (Lotze-Campen et al.,
2008; Wise et al., 2009). As illustrated in Fig. 3.5, preserving historical per capita
food consumption (E1) leads to an understatement of the increase in global crop
demand and global crop production over this historical period. With less output
growth, but the same level of TFP growth, prices fall sharply, yields grow more
slowly, and global cropland use contracts.
A more common consumption specification in global agricultural models is to
have fixed (unchanging) price and income elasticities of food demand (Havlik et al.,
2013; Nelson et al., 2010). In this case, rather than becoming smaller in absolute
value as per capita incomes rise (Muhammad et al., 2011), the responsiveness of
demand to rising incomes is based on historical estimates of these values and is kept
constant (E2). Figure 3.2 clearly illustrates that under this scenario both global crop
demand and global crop production are overstated. This is due to the dominance of
the income effect over this projections period. With sharply rising incomes, a failure
to account for the diminishing impact of marginal increments to purchasing power
results in excessively high demand and a significant overstatement of historical pro-
duction, area and yield, while global crop price falls by only about half of its
observed value.
146 11 Global Change and the Food System in 2050

Fig. 11.2 Impact of restrictions on the historical changes at the global level: 1961–2006. Source:
Baldos and Hertel (2013)

Looking at the supply side of the global agricultural picture, it is useful to recall
from Chaps. 1 and 4 that there are two key margins of economic response at work:
the extensive margin (additional area) and the intensive margin (yield increases).
Starting with the parameters which influence the extensive margin, in scenario E3
the long-run supply elasticities for land and non-land inputs are replaced with their
corresponding short-run (5 years) values from the 2001–2006 tuning exercise.
Models which are based on econometric estimates of cropland area response are
likely to fall prey to this limitation (Golub et al., 2012; Nelson et al., 2010). This is
because such estimates are typically based on annual time series data from which it
is hard to extract long term supply response. This point is emphasized by Hertel
(2011) who offers indirect evidence that prominent global studies of biofuels
(Fischer, Hizsnyik, Prieler, Shah, & Velthuizen, 2009) and climate impacts (Nelson
et al., 2010) are not using long run supply elasticities in their models. With these
11.1 Looking Back to Understand the Future 147

short-run parameters in place, the results in E3 show that a smaller global supply
response leads to a rise in crop prices over this period, as cropland area is unable to
respond as vigorously to increased land demand for crop production. While yield
changes are comparable to their historical values over this period, production falls
short of its observed growth rate, despite rising crop prices.
The other critical component of supply is the response of yields to higher crop
prices and/or increased scarcity of land. While the size of this response is hotly
debated (Berry & Schlenker, 2011; Goodwin, Marra, Piggott, & Mueller, 2012;
Huang & Khanna, 2010; Keeney & Hertel, 2009), there is little doubt that signifi-
cantly higher prices do encourage farmers to respond with more intensive cultiva-
tion practices. Yet not all agricultural models incorporate this possibility (Calvin,
Wise, Page, & Chini, 2012), and it is often unclear how large this effect is in those
models that do allow for endogenous yield response (Havlik et al., 2013; Lotze-
Campen et al., 2008; Nelson et al., 2010). This issue is further explored in experi-
ment E4 wherein this intensive margin of supply response is eliminated. As a
consequence, yields grow more slowly than in the historical record—being driven
solely by TFP growth. Crop prices are essentially flat and cropland expansion is in
excess of 40 %—as opposed to the observe change of just 16 %. Clearly failure to
account for the intensive margin of supply response can be expected to lead to a
significant overstatement of future cropland requirements.
A slightly different approach involves explicitly targeting the rate of average
crop yield growth (as opposed to targeting TFP). This is relevant, since many
biophysically-based agricultural models treat productivity growth as arising solely
through crop yield improvements (Havlik et al., 2013; Lotze-Campen et al., 2008;
Nelson et al., 2010). Of course, if the growth rate of crop yields in the future yield
is known, one can expect that it would help greatly in making credible projections
of global land use change. But, as experiment E5 demonstrates, even knowing
yields with certainty does not result in accurate prediction of cropland change over
this historical period. Since land is only one of many agricultural inputs, accurately
projecting yields does not allow for an accurate prediction of the change in crop
prices over time, as can be seen from the bar for E5 in the lower right panel in Fig.
11.2. This in turn leads to the underestimation of the changes in crop production and
cropland use.
The last two experiments illustrate the potential impacts on historical projections
when some of the above restrictions are combined. A purely biophysical view of the
historical period would treat per capita food consumption as being exogenous and
would ignore the potential for endogenous crop yield response to higher crop prices
(i.e. no intensive margin), instead targeting crop yield growth (E6.a). Similar to the
first experiment, global crop production in this scenario is grossly understated
(upper left panel of Fig. 11.2). By targeting average yields and ignoring the eco-
nomic yield response, the changes in global cropland use and global crop price
move in the opposite direction of what was observed over this historical period.
Another interesting combination of restrictions is captured by E6.b, which seeks
to mimic the behavior of those global agricultural models which fail to account for
long run changes on both the demand and supply sides. Specifically, the price and
148 11 Global Change and the Food System in 2050

Fig. 11.3 Validating historical changes at the regional level: 1961–2006

income elasticities of demand for food do not evolve with per capita incomes in this
scenario and the short to medium run input supply elasticities are imposed. With an
overly responsive demand for food, the projections tend to capture the observed rise
in global crop production but erroneously predict the change in global crop price.
As the supply of land is less responsive to land rents, global crop demand can only
be met by increasing the use of non-land inputs; hence, global average crop yields
are overstated while global cropland expansion is understated under this scenario.

11.1.3 Regional Production and the Failureof the Integrated

Markets Hypothesis

Once we move from the global to the regional scale, the SIMPLE model’s predic-
tions of crop production are much less satisfactory (left panel, Fig. 11.3), with too
little output expansion in South Asia, and Sub-Saharan Africa, and too much expan-
sion in North America and Europe & Central Asia. Upon reflection, this makes
sense. Returning to Table 11.1, we see that the rate of growth in TFP in Africa is
very low over this period. If African consumers had actually had ready access to
world markets during this time, they would have purchased much more of the food
from the countries where TFP was rising—and therefore costs were falling.
However, this was not an option for many households in Africa, as they were largely
isolated from world markets through high trade barriers and prohibitive trade and
transport costs. Indeed, majority of the world’s rural populations still rely on subsis-
tence farming. In order to feed the rapidly growing African population over this
historical period (recall Table 11.1), production had to increase much more strongly
than would have been expected by the global economic model, given the slow TFP
growth over this period.
11.1 Looking Back to Understand the Future 149

Other agricultural models have also found it difficult to capture changes in the
regional production (McCalla & Revoredo, 2001). A central feature of many mod-
els of global crop production is the assumption of fully integrated global crop mar-
kets. Yet this was far from the truth throughout most of the historical period. This
state of affairs was highlighted by D. Gale Johnson who published a series of papers
and books on the topic of “World Agriculture in Disarray” (Johnson, 1973) over the
post WWII period. In this work, Johnson discusses the many distortions which
caused the global distribution of agricultural output to be inconsistent with eco-
nomic logic. The evolution of these distortions has subsequently been documented
in a path-breaking study by Kym Anderson (2009). Since the completion of the
Uruguay Round of talks, which resulted in establishment of the World Trade
Organization, agricultural support has been reformed in many parts of the world.
However, there remain significant barriers to free trade in agricultural products
(Anderson & Martin, 2005) and this suggests the need to incorporate such policies
into SIMPLE if it is to accurately reflect the regional evolution of future
production.
In addition to explicit government policies shaping the regional patterns of agri-
cultural production, there are other important barriers to international trade in agri-
cultural products, including poor quality domestic transport infrastructure,
burdensome customs procedures and poorly developed port facilities. These barri-
ers to trade loom particularly large in Sub-Saharan Africa (Wilson, Mann, & Otsuki,
2004), and have limited that regions’ engagement in the global trading system. As a
consequence of this insulation from world markets, Sub-Saharan Africa's output has
grown much more than would have been anticipated, given its relatively low rate of
productivity growth over the 1961–2006 period. And its increased output has largely
been directed to domestic consumption. This is reflected in the fact that its share in
global trade of agricultural products has declined by around 70 % during this his-
torical period (FAO, 2013).
In an effort to capture these effects, while retaining the relative simplicity of
SIMPLE, we have developed another version of the model—this time dropping the
integrated markets hypothesis. Instead we conceptualize the regional markets as
being made up of a continuum of producers and a continuum of consumers—some
of which have ready access to world markets and therefore can purchase imports,
and some of which do not. Under certain assumptions about the distribution of
household characteristics (Anderson, Palma, & Thisse, 1992), the market level
demand (and supply) which emerge from this type of economy can be summarized
by a constant elasticity of substitution (or a constant elasticity of transformation in
the case of producers) between domestic and foreign varieties. In this case, a high
elasticity of substitution reflects an economy which is tightly integrated into world
markets. On the other hand, a low value for this parameter reflects an economy in
which poor infrastructure and high trade costs result in a strong segmentation of
domestic and international markets such that international and domestic prices are
not so closely linked.
We have implemented a segmented markets version of SIMPLE in order to test
whether this can, in fact, aid in reproducing the historical changes in regional crop
150 11 Global Change and the Food System in 2050

production. (In this preliminary implementation, the elasticities of substitution in

consumption and transformation of supplies between domestic and foreign markets
were all set to 3.0. Nonetheless the extent of market integration still depends on the
observed share of purchases from, and sales to, international markets.) . The right
panel in Fig. 11.3 shows the projected crop output changes over the 1961–2006
period with the segmented markets specification. Consider once again the results for
Sub-Saharan Africa. With segmented markets, African crop production increases
much more rapidly since much of the population does not have access to world
markets and therefore the food must be grown locally. Overall, the correspondence
between simulated and observed production is much improved in the lower panel of
Fig. 11.3. Of course, this could be further improved by varying the degree of market
segmentation between producers and consumers within and across regions. And
further enhancements would come from introducing trade costs and border policies
explicitly. We leave it to the readers to explore such extensions of this model for
their own regions of interest.

11.2 Globalization, Market Integration

and the Environmental Impacts of Technology6

From the preceding discussion, it is clear that the assumption of fully integrated
world markets is overly strong for the post-WWII period. However, it is also unde-
niable that, as global transportation and communications infrastructure improves,
the world is becoming more tightly integrated. A model which faithfully reproduces
historical outcomes from the latter half of twentieth century may not be appropriate
for analysis of developments in the world food system in the middle of the twenty-
first century. In this section, we use the segmented and integrated markets versions
of SIMPLE in order to explore the interplay between globalization and the environ-
mental impacts of technological change.
Recall the discussion in Chap. 4 about the land use impacts of technological
improvement in agriculture. There, we found that the historical Green Revolution
(GR) potentially contributed to improved environmental outcomes by keeping the
conversion of cropland below the level which would otherwise have arisen in its
absence. While we did not belabor the point in that early chapter, the results reported
in Fig. 4.4 were based on the SIMPLE model with segmented markets. This imple-
mentation followed directly from our experience with model validation at regional
scale, as reported in Fig. 11.3. In this section, we look forward to 2050 and ask a
similar question—only this time with respect to the prospective African GR which
many scientists, investors, governmental and non-governmental organizations hope
to stimulate (World Bank, 2009).

6
This section draws heavily on Hertel et al. (2014).
11.2 Globalization, Market Integration and the Environmental Impacts of Technology 151

Fig. 11.4 Sensitivity analysis of the regional and global cropland change and their corresponding
carbon emissions given a future African Green Revolution under both segmented and integrated
markets: difference between with vs. without Green Revolution TFP growth. Error bars reflect 95
% confidence intervals obtained from Monte Carlo analysis with respect to parameter uncertainty.
Source: Hertel et al. (2014)

Figure 11.4 reports the impact on global land use and associated GHG emissions
attributable to an African GR which becomes effective in 2025 and which boosts
TFP in the Sub-Saharan Africa region by an amount comparable to that achieved in
Asia during the historical GR. Under the historically valid, segmented markets
hypothesis (left-hand panel in the figure), this GR, like the historical one, is shown
to be land- and emissions-sparing at global scale, i.e. the GR lowers cropland area
and GHG emissions associated with cropland conversion. (The error bars in Fig.
11.4 were generated by running the model many times, sampling from the full dis-
tribution of economic parameters governing SIMPLE.) However, if efforts to inte-
grate the African economies into the global marketplace are successful, and the
integrated markets hypothesis were to apply in a post-2025 period (results shown in
right-hand panel of Fig. 11.4), then the impact on global land use would be ambigu-
ous, since the error bar on global land use change in the upper right hand corner of
Fig. 11.4 brackets zero. More striking is the finding that GHG emissions from land
152 11 Global Change and the Food System in 2050

conversion (lower right hand chart) would rise globally as a result of the African GR
in the presence of integrated world markets.
Hertel, Ramankutty, and Baldos (2014) explore this apparent paradox in detail
and show that it can be traced back to the relative environmental efficiency of
African agriculture. They follow West et al. (2010) and define environmental
efficiency in terms of the amount of carbon released into the atmosphere when a
hectare of land is converted, relative to the crop yield on that same hectare of land.
We focus on the inverse of this ratio, which is small when yields are low, but carbon
stocks are high. Such is the case in Africa, where the emissions efficiency is less
than half of that in the rest of the world (Hertel et al., 2014). Therefore, when
African TFP rises, and agriculture in that part of the world expands, the amount of
land which must be converted in order to fill the global market deficit left by con-
traction in the rest of the world is relatively large (due to the low yields). Furthermore,
despite these low yields, the carbon released by land conversion is still high—per-
haps comparable to, or even greater than that in many places where conversion is
now avoided. So we have a situation in which GHG emissions from cropland con-
version are higher than under the baseline scenario in which this region’s technol-
ogy improves less rapidly. This situation is exacerbated by the tendency for
incremental output in Africa to come from the extensive margin—i.e., favoring
cropland expansion over intensification.
However, as Hertel et al. (2014) are at pains to emphasize, this finding does not
mean that the world should not invest in new technology for Africa. Indeed, as TFP
grows, and yields improve, eventually environmental efficiency in the region will
rise to the point where future TFP growth is no longer GHG increasing. In the
interim, the authors emphasize the importance of protecting the most carbon-rich
lands from conversion—a measure which will indirectly boost the average environ-
mental efficiency in the region. Furthermore, any environmental costs must be
weighed against the gains from improving nutrition (recall Chap. 10), as well as
boosting rural incomes and reducing poverty in Sub-Saharan Africa.

11.3 Assessing Long Run Global Food

and Environmental Security7

Each of the preceding chapters in this book has developed one major theme under-
pinning the future of global food and environmental sustainability. To the maximum
extent possible, we have sought to isolate the consequences of each individual driver
for global land use, crop output, prices, nutrition and environmental indicators. This
is how we typically advance scientific knowledge—by studying factors in isolation
from one another. However, in attempting to make statements about long run food
and environmental sustainability, we must bring all of these factors together and

7
This section draws heavily on Baldos and Hertel (2014, 2015).
11.3 Assessing Long Run Global Food and Environmental Security 153

Fig. 11.5 Global cropland change and associated drivers: 1961–2006 and 2006–2051

account for their interactions, including potential counter-acting effects as well as

synergies. This cannot be done without a formal model, and so we conclude this
book by undertaking a set of forward-looking projections for the global food econ-
omy over the period 2006–2050 using the SIMPLE model, thereby seeking to con-
tribute to the debate over long run sustainability.

11.3.1 Implications for Future Cropland Conversion

Let us first focus on global cropland use. Figure 11.5 reports historical and projected
changes, decomposed by the four major drivers of change: population, per capita
income, biofuels and technology. Focusing first on the historical change, we see
that, over the 1961–2006 period, global cropland rose by about 17 %, with popula-
tion (red bar) being by far the largest driver of change. Growth in per capita income
(orange bar) was much less important over this period, as much of the income
growth was in wealthier countries where consumers were already eating relatively
rich diets (recall Chap. 2). Finally, note from Fig. 11.5 that, without strong produc-
tivity growth (green bar) there would have been far more cropland conversion.
Now let us look ahead from 2006 to 2051, using baseline projections for popula-
tion from the UN World Population Prospects (2013) and real GDP growth from
CEPII (Fouré, Bénassy-Quéré, & Fontagné, 2013), along with biofuels from the
154 11 Global Change and the Food System in 2050

IEA ‘business as usual’ baseline (IEA, 2008, 2012). We can see from the next two
bars in Fig. 11.5 that income growth is now a more important driver of land use
change than population. Biofuels are also significant going forward, although as we
learned in Chap. 8, this source of growth in demand for cropland is highly uncertain
and very dependent on the future price of oil. The reason for reporting two sets of
cropland change projections in Fig. 11.5 is the great uncertainty in total factor pro-
ductivity growth going forward. If productivity growth in global agriculture contin-
ued at the rate achieved over past decades, there would be minimal need for (net)
cropland conversion at the global level. (Of course we expect gross conversion at
the local level, with some land moving into crops and other land transitioning out of
crops for a variety of reasons.) However, as we discussed in Chap. 3, there are seri-
ous concerns about the potential for the agricultural research and development
infrastructure to continue to deliver such large annual gains in TFP. And solely on
the basis of linear yield growth trends, we expect future rates of yield growth to
decline (Grassini, Eskridge, & Cassman, 2013). Add to this, the potential for cli-
mate change to slow productivity growth (IPCC, 2014), and it seems prudent to
project somewhat slower TFP growth in the coming decades. With this in mind, we
adopt the lower productivity growth projections of Ludena, Hertel, Preckel, Foster,
and Nin (2007) in the second set of projections in Fig. 11.5. With this rate of future
rate of productivity growth, future rates of global net cropland conversion return to
historical levels. However, even though the cumulative conversion rate is compara-
ble to the past, the drivers are quite different going forward, as illustrated by con-
trasting the colored components of the future growth projections with those from
the historical decomposition.

11.3.2 Implications for Food Prices

We turn next to the implications of this analysis for food prices—or more specifi-
cally, aggregated crop prices. Following the logic of the previous figure, Fig. 11.6
begins by breaking down the drivers of historical price trends. Since cropland con-
version is closely related to prices, it is not surprising that the relative ranking of
drivers of crop prices is the same as for cropland. While population was the most
important demand-side driver of crop prices historically, income growth will sur-
pass population as growth in the latter falls in the coming decades. Overall, the
strong historical TFP growth led to a trend of declining prices over the 1961–2006
period. However, as with the case of cropland conversion, whether prices rise or fall
in the future depends critically on future TFP growth. As before, we report two sets
of projections—one with historical rates of productivity growth and one with
(lower) projections for the 2006–2051 period. In the former case prices fall even
more sharply, while in the latter, the price decline is moderated.
Given the recent controversy over whether crop prices entered a ‘new normal’
period of permanently higher and possibly rising prices (Food and Agriculture
Organization of the UN Media Centre, 2013; OECD/FAO, 2013; World Bank,
11.3 Assessing Long Run Global Food and Environmental Security 155

Fig. 11.6 Historical and future price changes

2013), it is useful to think about future projections in a more careful manner. Even
if one agreed that SIMPLE offered a reasonable representation of global crop mar-
kets, there is significant scope for disagreement regarding the size of future income
growth, for example. In addition, there is uncertainty in the response of food
demands to rising incomes (recall Chap. 2). If our projections of income growth in
the developing world are too low, for example, and if our income elasticity of
demand for food in those regions is also too low, then one could imagine a future in
which crop demand growth was much stronger. If this happened to coincide with
baseline TFP projections which were overly optimistic, then one could end up with
future crop prices which were rising instead of falling. How likely is such an out-
come? Monte Carlo analysis permits us to investigate this question formally. By
specifying explicit probability distributions for all of the growth drivers as well as
for the economic response parameters in SIMPLE, we can then explore then entire
set of possible future outcomes.
Figure 11.7 reports results from this Monte Carlo analysis with the horizontal axis
reporting the projected percentage change in the SIMPLE model’s crops price index
between 2006 and 2050, and the vertical axis recording the frequency with which
each range of price change is observed. The most likely outcome (modal value) is a
price decline of about −23 %, with a mean change of −8 %. Roughly 72 % of the
outcomes foreshadow a price decline over this period. However, there remain 28 %
of total outcomes which correspond to a rise in prices over this period (with the high-
est increase at 56 %), so we cannot rule out the possibility of a ‘new normal’—at least
compared to the 1961–2006 period. However, these outcomes are less likely and tend
to correspond to more eccentric combinations of shocks and model parameters.
156 11 Global Change and the Food System in 2050

Fig. 11.7 Distribution of price changes: 2006–2050, given uncertainty in underlying drivers of
global change, as well as economic response parameters

Of course there are additional factors which bear on crop prices in the long run.
As discussed in Chap. 5, water scarcity can result in increased land conversion and
higher prices. However, based on the work of Liu, Hertel, Taheripour, Zhu, and
Ringler (2014), it appears that the global price impacts of future scarcity of irriga-
tion water will be minimal—as the global trading system is likely to simply reallo-
cate production to more water abundant regions. A more significant factor is climate
mitigation policy, which, as shown in Chap. 7, can have a very significant impact on
future cropland availability. An aggressive, global policy to encourage forest carbon
sequestration would sharply restrict future cropland conversion, resulting in higher
crop prices than otherwise projected. As we have shown, this issue can be explored
using SIMPLE, once estimates of the likely shift in regional cropland supply are
available.

11.3.3 Implications for Nutritional Attainment

In addition to discussing the implications for crop land conversion, GHG emissions
and prices, it is important to return to the primary objective of the global food sys-
tem, which is to ensure that the population is well-nourished. Towards this end, we
combine these forward-looking projections with the nutrition module developed in
Chap. 10 in order to say something about the long run prospects for caloric malnu-
trition. Figure 11.8 reports the change in global malnutrition headcount under two
alternative specifications of market structure: integrated and segmented regional
11.3 Assessing Long Run Global Food and Environmental Security 157

Fig. 11.8 Contributions of key drivers of agriculture to changes in global malnutrition count under
the LPJmL crop model and HADGEM GCM model for different trade regimes. White bars shows
the change in the number of malnourished persons from 2006 to 2050 while the colored bars high-
light the contribution of the key drivers of agriculture to the changes in global malnutrition count.
Source: Baldos and Hertel (2014, 2015)

markets. In addition, in order to factor in the range of potential climate impacts, we

draw on simulation results obtained from combining the HADGEM climate model
with the LPJmL crop model. As noted in Chap. 6, the most controversial aspect of
climate impacts in agriculture relates to the potential impacts for crop growth of
elevated CO2 concentrations in the atmosphere. Therefore, we consider results
wherein this is ignored (left-hand panel) and where it is full factored in (right hand
panel). Thus Fig. 11.8 offers four different scenarios for 2050.
The first point to note is that, regardless of market structure and climate impacts,
we expect significant reductions in caloric malnutrition between 2006 and 2050.
This is due to the strong increases in per capita income in the world’s poorest regions
(Africa and South Asia). This is very good news. However, there is a substantial
difference between the malnutrition headcount with and without the CO2 fertiliza-
tion effects. Absent the beneficial effects of elevated CO2 on crop growth—perhaps
due to the offsetting impact of elevated canopy temperatures on crop growth—
yields are lower and crop prices higher. This causes malnutrition to fall by just 427
million instead of 573 million under the segmented markets case. These results
underscore the importance of better understanding the impact of climate change on
crop yields.
Another point also becomes evident from Fig. 11.8, and that is the critical impor-
tance of more tightly integrated world markets for reducing malnutrition. In the case
on ‘no CO2 fertilization’, moving from segmented to integrated markets has as large
an impact on the malnutrition headcount as does the addition of the CO2 effects—
boosting this from −427 to −580 million persons. This is because the more tightly
integrated world markets allow consumers in the regions hardest hit by climate
158 11 Global Change and the Food System in 2050

change to readily access crop outputs from other regions that are less adversely
affected. These benefits are less pronounced, but still in evidence in the case of CO2
fertilization (right hand panel).
In summary, globalization—in the form of more tightly integrated world mar-
kets—appears to be a double-edged sword. On the one hand, it raises the prospect
of adverse environmental impacts from productivity growth in the world’s poorest
regions (e.g. Africa), where environmental efficiencies are relatively low and crop-
land area response is high. On the other hand, smoothly functioning global markets
for agricultural products are a boon to low income urban consumers—particularly
in the context of the most adverse climate change projections.

Supplemental learning resources on this chapter are available for free at https://mygeohub.org/
courses/global_change.

References

Anderson, S. P., Palma, A. D., & Thisse, J. F. (1992). Discrete choice theory of product differentia-
tion. Cambridge, MA: MIT Press.
Anderson, K., & Martin, W. (2005). Agricultural Trade Reform and the Doha Development
Agenda. World Economy, 28(9), 1301–1327. http://doi.org/10.1111/j.1467-9701.2005.00735.x.
Baldos, U. L. C., & Hertel, T. W. (2013). Looking back to move forward on model validation:
Insights from a global model of agricultural land use. Environmental Research Letters, 8(3),
034024. http://doi.org/10.1088/1748-9326/8/3/034024.
Baldos, U. L., & Hertel, T. (2014). Bursting the bubble: A long run perspective on crop commodity
prices. Retrieved from http://www.gtap.agecon.purdue.edu/resources/res_display.asp?
RecordID=4574.
Baldos, U. L. C., & Hertel, T. W. (2015). The role of international trade in managing food security
risks from climate change. Food Security, 7(2), 275–290. http://doi.org/10.1007/s12571-
015-0435-z
Berry, S., & Schlenker, W. (2011). Empirical Evidence on Crop Yield Elasticities (Technical
Report) (p. 25). Washington, D.C., USA: International Council on Clean Transportation.
Retrieved from http://www.theicct.org/empirical-evidence-crop-elasticities.
Calvin, K., Wise, M., Page, K., & Chini, L. (2012). Spatial Land Use in the GCAM Integrated
Assessment Model. In EMF Workshop on Climate Change Impacts and Integrated Assessment.
Snowmass, Colorado, USA: Pacific Northwest National Laboratory.
Food and Agriculture Organization of the UN Media Centre. (2013, April 12). Feeding nine billion
in 2050. Retrieved May 23, 2013, from http://www.fao.org/news/story/en/item/174172/icode/.
Fouré, J., Bénassy-Quéré, A., & Fontagné, L. (2013). Modelling the world economy at the 2050
horizon. Economics of Transition, 21(4), 617–654. http://doi.org/10.1111/ecot.12023.
Goodwin, B., Marra, M., Piggott, N., & Mueller, S. (2012). Is Yield Endogenous to Price? An
Empirical Evaluation of Inter- and Intra-Seasonal Corn Yield Response. In AAEA Annual
Meeting. Seattle, Washington, USA: North Carolina State University.
Golub, A. A., Henderson, B. B., Hertel, T. W., Gerber, P. J., Rose, S. K., & Sohngen, B. (2012).
Global climate policy impacts on livestock, land use, livelihoods, and food security. Proceedings
of the National Academy of Sciences, 1–6. http://doi.org/10.1073/pnas.1108772109.
Grassini, P., Eskridge, K. M., & Cassman, K. G. (2013). Distinguishing between yield advances
and yield plateaus in historical crop production trends. Nature Communications, 4, 2918. http://
doi.org/10.1038/ncomms3918.
References 159

Griffith, R., Redding, S., & Reenen, J. V. (2004). Mapping the Two Faces of R&D: Productivity
Growth in a Panel of OECD Industries. Review of Economics and Statistics, 86(4), 883–895.
http://doi.org/10.1162/0034653043125194.
Fischer, G., Hizsnyik, E., Prieler, S., Shah, M., & van Velthuizen, H. (2009). Biofuels and Food
Security: Implications of an accelerated biofuels production (No. Issue 38). Vienna, Austria:
Organization of the Petroleum Exporting Countries Fund for International Development.
Fuglie, K. O. (2012). Productivity growth and technology capital in the global agricultural econ-
omy. In K. O. Fuglie, S. L. Wang, & V. E. Ball (Eds.), Productivity Growth In Agriculture: An
International Perspective (pp. 335–368). Cambridge, MA, USA: CAB International.
Havlik, P., Valin, H., Mosnier, A., Obersteiner, M., Baker, J. S., Herrero, M., … Schmid, E. (2013).
Crop productivity and the global livestock sector: implications for land use change and green-
house gas emissions. American Journal of Agricultural Economics, 95(2), 442–448. http://doi.
org/10.1093/ajae/aas085.
Hertel, T. W. (2011). The Global Supply and Demand for Agricultural Land in 2050: A Perfect
Storm in the Making? American Journal of Agricultural Economics, 93(2), 259–275. http://doi.
org/10.1093/ajae/aaq189.
Hertel, T. W., Ramankutty, N., & Baldos, U. L. C. (2014). Global market integration increases
likelihood that a future African Green Revolution could increase crop land use and CO2 emis-
sions. Proceedings of the National Academy of Sciences of the United States of America,
111(38), 13799–13804. http://doi.org/10.1073/pnas.1403543111.
Huang, H., & Khanna, M. (2010). An Econometric Analysis of U.S. Crop Yield and Cropland
Acreage: Implications for the Impact of Climate Change. In AAEA, CAES, & WAEA Joint
Annual Meeting. Denver, Colorado, USA: Department of Agricultural and Consumer
Economics, University of Illinois-Urbana-Champaign.
IEA. (2008). World energy outlook. OECD. Retrieved from http://www.oecd-ilibrary.org/content/
book/weo-2008-en.
IEA. (2012). World energy outlook. OECD. Retrieved from http://www.oecd-ilibrary.org/content/
book/weo-2012-en.
IPCC. (2014). Climate change 2014: Impacts, adaptation, and vulnerability. In Contribution of
Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge, UK: Cambridge University Press.
Johnson, D. G. (1973). World Agriculture in Disarray. London, UK; New York, USA: Macmillan.
Anderson, K. (2009). Distortions to Agricultural Incentives: A Global Perspective, 1955-2007.
Washington, D.C., USA: World Bank Publications.
Keeney, R., & Hertel, T. W. (2009). The Indirect Land Use Impacts of United States Biofuel
Policies: The Importance of Acreage, Yield, and Bilateral Trade Responses. American Journal
of Agricultural Economics, 91(4), 895–909. http://doi.org/10.1111/j.1467-8276.2009.01308.x.
Liu, J., Hertel, T., Taheripour, F., Zhu, T., & Ringler, C. (2014). International trade buffers the
impact of future irrigation shortfalls. Global Environmental Change, 29, 22–31.
Lotze-Campen, H., Müller, C., Bondeau, A., Rost, S., Popp, A., & Lucht, W. (2008). Global food
demand, productivity growth, and the scarcity of land and water resources: A spatially explicit
mathematical programming approach. Agricultural Economics, 39(3), 325–338. http://doi.
org/10.1111/j.1574-0862.2008.00336.x.
Ludena, C. E., Hertel, T. W., Preckel, P. V., Foster, K., & Nin, A. (2007). Productivity growth and
convergence in crop, ruminant, and nonruminant production: Measurement and forecasts.
Agricultural Economics, 37(1), 1–17. http://doi.org/10.1111/j.1574-0862.2007.00218.x.
McCalla, A. F., & Revoredo, C. L. (2001). Prospects for Global Food Security: A Critical Appraisal
of Past Projections and Predictions (Food, Agriculture, and the Environment Discussion Paper
No. 35) (p. 81). Washington D.C., USA: International Food Policy Research Institute.
Muhammad, A., Seale Jr., J. L., Meade, B., & Regmi, A. (2011). International Evidence on Food
Consumption Patterns: An Update Using 2005 International Comparison Program Data
(Technical Bulletin No. TB-1929) (p.59). Washington, D.C., USA: Economic Research
Service, US Department of Agriculture. Retrieved from http://www.ers.usda.gov/Publications/
TB1929/.
160 11 Global Change and the Food System in 2050

Nelson, G., Rosegrant, M. W., Palazzo, A., Gray, I., Ingersoll, C., Robertson, R. D., … You, L.
(2010). Food security, farming, and climate change to 2050: Scenarios, results, policy options
(Research reports No. Gerald C. Nelson, et al.). International Food Policy Research Institute
(IFPRI).
OECD/FAO. (2013). OECD-FAO agricultural outlook 2013 (Annual outlook) (p. 326). OECD/
FAO. Retrieved from http://www.oecd.org/site/oecd-faoagriculturaloutlook/prices.htm.
Schmitz, C., van Meijl, H., Kyle, P., Nelson, G. C., Fujimori, S., Gurgel, A., … Valin, H. (2014).
Land-use change trajectories up to 2050: Insights from a global agro-economic model com-
parison. Agricultural Economics, 45(1), 69–84. http://doi.org/10.1111/agec.12090.
Smith, P., Gregory, P. J., van Vuuren, D., Obersteiner, M., Havlik, P., Rounsevell, M., … Bellarby,
J. (2010). Competition for land. Philosophical Transactions of the Royal Society B: Biological
Sciences, 365(1554), 2941–2957. http://doi.org/10.1098/rstb.2010.0127.
UN Population Division. (2013). World population prospects: The 2012 revision. New York, NY:
Population Division of the Department of Economic and Social Affairs of the United Nations
Secretariat. Retrieved from http://esa.un.org/unpd/wpp/index.htm.
West, P. C., Gibbs, H. K., Monfreda, C., Wagner, J., Barford, C. C., Carpenter, S. R., & Foley, J. A.
(2010). Trading carbon for food: Global comparison of carbon stocks vs. crop yields on agri-
cultural land. Proceedings of the National Academy of Sciences of the United States of America,
107(46), 19645–19648. http://doi.org/10.1073/pnas.1011078107
Wilson, J. S., Mann, C. L., & Otsuki, T. (2004). Assessing the potential benefit of trade facilitation:
A global perspective (Policy Research Working Paper Series No. 3224). Washington, D.C.,
USA: The World Bank. Retrieved from http://ideas.repec.org/p/wbk/wbrwps/3224.html.
Wise, M., Calvin, K., Thomson, A., Clarke, L., Bond-Lamberty, B., Sands, R., … Edmonds,
J. (2009). Implications of limiting CO2 concentrations for land use and energy. Science,
324(5931), 1183–1186. http://doi.org/10.1126/science.1168475.
World Bank. (2009). Awakening Africa’s sleeping giant: Prospects for commercial agriculture in
the Guinea Savannah Zone and beyond (pp. 1–236).
World Bank. (2013, April 15). Food crisis. Retrieved May 23, 2013, from http://www.worldbank.
org/foodcrisis/bankinitiatives.htm.
Appendix A
Class Projects

This appendix contains a sample of the class projects undertaken in the first 4 years
of this course offering. They illustrate the wide range of potential applications of
SIMPLE which creative students from a variety of different disciplines bring to the
table. Many of the SIMPLE version archives associated with these applications are
available on the web site: https://mygeohub.org/courses/global_change.
1. Food Waste and Post-harvest Losses: The UN-FAO estimates that one-third of
global food production is lost or wasted so that only two-thirds of production is
actually consumed. What are the implications of such losses for crop prices?
How would a reduction in post-harvest losses affect nutritional outcomes?
What measures could be implemented to encourage a reduction in food waste,
and what would be their impact?
2. Changing Nutrition Guidelines: The USDA is in the process of formulating a
new set of nutrition guidelines. For the first time they are considering adding
environmental impacts to these guidelines. How would such considerations
change the pattern of food consumption? How would changing consumption
patterns alter the pattern of global land use and GHG emissions?
3. Africa as the Sleeping Giant of Agriculture: In 2009, the World Bank published
a report suggesting that the Guinea Savannah Zone of Africa could become the
next breadbasket for the world. What would be the implications of such a devel-
opment for land use, food security and environmental quality, worldwide?
4. Empowerment of women: We know from course readings and discussion that
female education and empowerment has consequences for population growth,
agricultural productivity and nutrition. How significant are these effects? Since
they affect both supply and demand for food, how does their interaction affect
global food and environmental security?
5. Urbanization: What are urbanization’s impacts on demand for land? For histori-
cal reasons, urban development often occurs in ecologically and agrononmically
rich river deltas. What are the implications for global land use, food security and
the environment of further conversion, as the world becomes more urbanized?

© Springer International Publishing Switzerland 2016 161

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0
162 Appendix A  Class Projects

6. REDD: Similar to the urbanization experiment, students could explore the

impact of efforts to dedicate additional land to the production of environmental
services. This would be implemented through adjustments to the regional
conversion factors of land in environmental services to cropland.
7. Climate change: Chap. 6 explored the impacts of different types of climate
shocks on agricultural productivity and land use. A deeper dive might entail
converting existing studies of the impacts of climate change into shocks or
parameter adjustments within the SIMPLE own model and exploring the
results.
8. Water scarcity: Chap. 5 discusses emerging water scarcity in some parts of the
world. Since irrigated agriculture uses 70% of total freshwater withdrawals
globally, how will these shortages affect global food and environmental secu-
rity? To what extent will local shortages be offset by increased imports from
other regions? What are the consequences for virtual water trade?
9. Jevon’s paradox: Chap. 4 introduces Jevon’s paradox—a situation in which
yield growth might lead to greater land use. A deeper exploration of the condi-
tions under which we might expect such a result would be useful. What is the
demand elasticity required to generate this outcome? How do yield differentials
across regions affect the likelihood of this outcome? Can these results tell us
anything about the probability of experiencing Jevon’s paradox in the real
world?
10. Impacts of shifting population and income on global demand elasticity:
Regional differences in demographics and income growth will shift the balance
of global demand. How do these shifts change the aggregate global demand
elasticity? How does this compare to shifting demand elasticities within
regions?
11. Biophysical/economic interaction: Relative yields (local yields versus the
global maximum) might be indicative of how close a particular region is to the
biophysical limits of intensification given current technologies. To understand
how such biophysical limits interact with the economics governing the crop
market, one might want to consider a linear relationship between relative yields
and our intensification parameter. By plotting different regions on this linear
relationship, one could determine whether such a biophysical limit, through its
impact on the economic parameters, changes the outcomes of crop expansion
scenarios.
12. Impacts of different types of technology growth: In the SIMPLE model we can
simulate land augmenting, land dis-augmenting and technology neutral produc-
tivity growth. What are the realistic ranges of these types of productivity growth
going forward? If yield growth outpaces non-land technology growth, what
would be the impact on prices and extensification? Under different scenarios,
does the clear relationship between land prices and extensification begin to
break down?
13. Globalization: As discussed in Chap. 4, the effective elasticity of demand for a
local market depends on the rest of the world’s supply elasticity, the local pro-
duction’s share of the global market as well as the global demand elasticity.
Appendix A  Class Projects 163

Among other possibilities, an analysis of globalization’s impact might compare

the extensification impacts of productivity changes in small markets (share of
world supply is small) to large markets. A more sophisticated analysis would
compare results from a fully integrated world market with one in which regional
markets are segmented [see also Hertel, Ramankutty and Baldos (2014)]1,2.
14. Changing productivity of livestock and food processing: This book has empha-
sized the impacts of changing crop productivity. However, one could also assess
the implications of changes in the TFP of livestock production or food process-
ing. Contradictory effects of these downstream productivity changes (less crops
required to produce a good, but demand is now increasing) may lead to interest-
ing results depending on the assumed parameter values.
1 5. Market mediated responses: Hertel (2011)3 highlights the importance of con-
sidering economic factors when estimating the land use implications of changes
in biofuels demand. Chapter 8 pursues this topic in some detail as well. To what
extent are the biophysical estimates of the impacts of demand shocks (e.g. pop-
ulation growth, income expansion) moderated through economic processes
such as intensification and demand reduction?
1 6. Cost-benefit analysis of productivity growth: Numerous papers provide esti-
mates of the cost and scale of historical TFP growth. Using these estimates, one
could estimate the consumer and producer surplus generated by the yield
growth to evaluate the cost effectiveness. How does this compare with the origi-
nal investment in these technologies?

1
 Hertel, T. W., Ramankutty, N., & Baldos, U. L. C. (2014). Global market integration increases
likelihood that a future African Green Revolution could increase crop land use and CO2 emissions.
Proceedings of the National Academy of Sciences, 111(38), 13799–13804. http://doi.org/10.1073/
pnas.1403543111
2
 Baldos, U. L. C., & Hertel, T. W. (2013). Looking back to move forward on model validation:
insights from a global model of agricultural land use. Environmental Research Letters, 8(3),
034024. http://doi.org/10.1088/1748-9326/8/3/034024
3
 Hertel, T. W. (2011). The Global Supply and Demand for Agricultural Land in 2050: A Perfect
Storm in the Making? American Journal of Agricultural Economics, 93(2), 259–275. http://doi.
org/10.1093/ajae/aaq189
Appendix B
Analytical Framework: A Theoretical Model
of Long Run Demand and Supply
for Agricultural Land

The model developed in this appendix is the simplest possible analytical framework
for studying the six key factors shaping long run demand and supply for agricultural
land and characterized in Eq. (1.1) of Chapter 1. Begin with the demand for aggre-
gate, agricultural output (qDA). The slope of this demand schedule is governed by the
aggregate price elasticity of demand which (ηDA) captures the sales-weighted aver-
age responsiveness of all sources of demand to changes in the price of agricultural
output, including food, fiber and food. This demand curve is subject to outward
shifting over time due to exogenous growth in population, per capita income, and
other factors (ΔDA), so that the percentage change in long run demand may be
expressed as follows:

q AD = -h AD pA + D DA (B.1)

Now postulate one global production function for agricultural output, which
combines land (and water) with variable inputs (including labor, capital, fertilizer,
etc.), wherein the latter are deemed to be in perfectly elastic supply in the long run.
The potential for increasing yields in response to higher global prices for farm prod-
ucts—and hence higher returns to land—is captured by the elasticity of substitution
in production (σ). Assuming constant returns to scale (clearly inappropriate at the
farm level, but a good fit for industry behavior under entry and exit), zero pure eco-
nomic profits [Eq. (B.3)], and cost minimizing behavior, this gives the derived
demand for land in Eq. (B.2):

qLD = q AS - s ( pL - pA ) - D DL (B.2)

Here, qLD denotes the long run percentage change in global derived demand for
land which is a function of three factors. Firstly, all else constant, boosting the sup-
ply of agricultural output, qAS, by 50% will require a 50% increase in effective (pro-
ductivity adjusted) land. However, this land requirement may be diminished if there

© Springer International Publishing Switzerland 2016 165

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0
166 Appendix B  Analytical Framework…

is an economic incentive for intensification of production. Such an incentive will

arise when land becomes scarce, relative to other inputs.
Note that aggregate output price (pA) is just an index of input costs under the zero
profits/constant returns to scale, so that the percentage change in this variable may
be expressed as follows:

pA = åq j p j (B.3)
j

Where pj is the percentage change in the price of input j, and θj is the share of input
j in total costs of agricultural production. Therefore, a scarcity of land will be indi-
cated by a rise in the rental rate on land, relative to the composite price of inputs,
which is just the price of output, i.e. ( pL - pA ) > 0 . Under such circumstances,
producers will substitute variable inputs (e.g., labor, fertilizer) for land, thereby rais-
ing agricultural output per unit of land, subject to the limitations of technology, as
described by σ. Note that we also include the potential for exogenous “technological
change” in the form of yield growth (expressed as a percentage of global average
yields), ΔLD, which enters into this demand equation with a negative sign, such that
faster “trend” yield growth diminishes the global derived demand for land. This
exogenous element is included, as many long run projections reviewed include an
element of changing yields which is not determined by relative prices and therefore
exogenous to this simple partial equilibrium model.
To complete the model, we must add Eq. (B.4) describing the long run supply of
land to agriculture:

qLS = n LS pL - D SL (B.4)

As with aggregate demand, we include both a price-responsive element, indicating
that additional land will be converted to agricultural uses if the relative return to
land used in agriculture rises (based on the elasticity of land supply, νLS), and an
exogenous element, ΔLS, representing a shift in this land supply schedule owing to
the conversion of agricultural land to other uses due to forces outside this simple
model (e.g., urban sprawl).
The long run factor market closure for this model assumes that p j = 0, "j ¹ L .
That is, long run prices for the variable factors of production are not affected by
conditions within the agricultural sector. Applying this variable input price restric-
tion into Eq. (B.3), and rearranging to isolate the change in land rents on the left
hand side, yields a relationship whereby any long run change in agricultural prices
is capitalized into land rents:

pL = q L-1 pA (B.5)

Plugging Eq. (B.5) into the land supply Eq. (B.4), we obtain land supply in terms of
commodity price:
Appendix B  Analytical Framework… 167

qLS = n LSq L-1 pA - D SL = h AS, E pA - D SL (B.6)

So we can see that the elasticity of land supply with respect to commodity price is
likely to be considerably smaller than that the elasticity with respect to land rents.
We will call this the extensive margin of supply response to commodity price:
h AS, E = n LSq L-1 .
Now Rearrange Eq. (B.2) to isolate commodity supply on the left hand side. This
becomes our agricultural commodity supply equation:

q AS = qLD + s ( pL - pA ) + D DL (B.7)

And substitute in Eqs. (B.5) and (B.6) (recognizing that land supply must equal land
demand in long run equilibrium), yielding:

( )
q AS = n LSq L-1 pA - D SL + s q L-1 pA - pA + D DL

(B.8)

Which we can rearrange to become:

( )
q AS = éën LSq L-1 + s q L-1 - 1 ùû pA - D SL + D DL

(B.9)

The term in brackets [.] is the aggregate agricultural supply response to output price
and is comprised of the extensive margin discussed above, and the intensive margin,
( )
h AS, I = s q L-1 - 1 , which depends on the elasticity of substitution between variable
inputs and land (larger σ gives a larger supply response) and the cost share of land
(larger θL gives a smaller supply response). If land is the only input, then there is no
scope for intensification. Therefore, the total supply response of agriculture, when
measured in terms of responsiveness of output to output prices may be written as:
q AS / pA = h SA = h AS, I + h AS, E .
Equating commodity supply [Eq. (B.9)] to demand [Eq. (B.1)] yields:

(h S,I
A )
+ h AS, E pA - D SL + D DL = -h AD pA + D DA

(B.10)

Equation (B.10) may be solved for the long run, equilibrium commodity price
change as a function of the exogenous shocks:

( ) ( )
p*A = D DA + D SL - D DL / h AS, I + h AS, E + h AD = D / h

(B.11)

Equation (B.11) is extremely useful in understanding the commodity price

impacts of exogenous shocks to commodity demand (e.g., due to biofuel mandates),
to yields (e.g., due to climate change), or to land supply (e.g., due to urbanization).
This relationship is mediated by the three key elasticities in our model: the intensive
margin of supply response, the extensive margin, and the price elasticity of demand.
For small values of these combined elasticities, we can expect to see very large
168 Appendix B  Analytical Framework…

price changes. As any one of these margins becomes more price responsive, we will
see a rapid dampening of the commodity prices flowing from such shocks.
In order to deduce the implications for long run land use in agriculture, we plug
Eq. (B.11) into Eq. (B.6) to obtain:

qL* = h AS, E D / h - D SL (B.12)

Dividing top and bottom of the first term on the right hand side, we obtain the fol-
lowing expression for the long run equilibrium use of land in agriculture:

( ) ( )
qL* = éë D DA + D SL - D DL / 1 + h AS, I / h AS, E + h AD / h AS, E ùû - D SL

(B.13)

Equation (B.13) shows the role of each of the six key determinants of the long
run demand and supply for land in agriculture. These include three exogenous
shifters: (1) shifts in commodity demand (e.g., due to biofuels or population growth),
(2) shifts in land supply (e.g., due to urbanization), (3) exogenous growth in yields
(e.g., due to prior investments in agricultural R&D), and three elasticities of supply
and demand capturing the endogenous: (4) potential for intensification of agricul-
ture, (5) potential for agricultural land expansion, and (6) potential for demand
reduction.
From this expression, it can be seen that, if there is simultaneously no scope for
intensification of production ( h AS, I = 0 ) and no price responsiveness in demand
( h AD = 0 ), then the denominator of the term in brackets [.] collapses to one and the
land supply shocks cancel out. Therefore, any exogenous growth in net demand—
for example due to a 5% increase in population relative to trend yields—gets fully
translated into an equilibrium rise in land use in agriculture. As discussed in
Chap. 11, this is a pretty accurate characterization of much of the literature long run
land use. All of the effort in these studies goes into computing the net demand fac-
( )
tor, D DA - D DL , and, once that is done, the result is immediately translated into a
change in agricultural land use, abstracting from the fact that such shocks might in
fact lead to an endogenous response on the part of the farm and food system.
Equation (B.13) offers valuable insight into what is missed in such biophysical
studies of global agricultural land use. In particular, note the role which the price
responsiveness of yields and final demand play in this expression. By boosting the
size of the denominator in Eq. (B.13), they serve as “shock-absorbers”, dampening
the amount of land employed in agriculture over the long run in the face of positive
net demand growth. That is, any outward shift in net demand tends to boost food
prices, which, in turn tempers the realized demand for agricultural products. The
ensuing rise in land prices serves to encourage the intensification of production,
with agricultural producers (and indeed the long run activities of the agricultural
research establishment) substituting variable inputs for land to achieve higher
yields.
A second important insight from Eq. (B.13) is that what matters from the point
of view of equilibrium land use in agriculture is not the absolute size of the
Appendix B  Analytical Framework… 169

i­ ntensification and final demand elasticities, but rather their size relative to the land
supply elasticity (this is the extensive margin of land use, and is captured by ηAS,E).
Indeed, if all three elasticities are equal, then the denominator of the bracketed term
in Eq. (B.13) is just equal to three. In this case, each margin absorbs a third of the
adjustment, and the equilibrium land use change is equal to the sum of one-third of
the net demand growth, minus two-thirds of the agricultural land supply shock
( ) ( )
= 1 / 3 D DA - D DL - 2 / 3 D SL .
Appendix C
Description of the SIMPLE Model

As the reader will have noticed, there are many different variants of the SIMPLE
model, each one targeted to a different problem, including: GHG emissions, nutri-
tion, food waste, etc. And we anticipate that many more versions will emerge in the
wake of the publication of this book. Each one is documented in the context of
the first application of that particular variant of the model. (To obtain an up to date
list and download relevant software and files, visit https://mygeohub.org/courses/
global_change.) All of the variants of SIMPLE are built around a common core
model which was first documented in the supplementary online materials for
Baldos and Hertel (2013) which focused on the validation of this model over the
period: 1961–2006. The following material is drawn from this appendix.
1. Mathematical description of SIMPLE
Sets/indices
Commodities

i = ( Crops, Livestock , Proc _ Food, Non _ Food )

Income Regions

y = (Up _ Higher , Low _ Higher , Up _ middle, Low _ middle, Low )

Geographic Regions

æ E _ Asia _ Pac, Eur _ C _ Asia, L _ Amer _ Carr , ö

g =ç ÷
è M _ East _ N _ Afr , N _ America, S _ Asia, S _ S _ Africa ø

Parameters
εP(i,y) Price elasticity of commodity demand
εY(i,y) Income elasticity of commodity demand

© Springer International Publishing Switzerland 2016 171

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0
172  Appendix C Description of the SIMPLE Model Appendix C 

αP(i) Intercept of price elasticity regression with log of per capita

income
αY(i) Intercept of income elasticity regression with log of per capita
income
βP(i) Slope of price elasticity regression with log of per capita income
βY(i) Slope of income elasticity regression with log of per capita
income
σCROP(g) Substitution elasticity between cropland and non-land inputs
σPRFD Substitution elasticity between crops and non-crop inputs in the
processed food sectors
σLSTK Substitution elasticity between feed and non-crop inputs in the
livestock sectors
εLAND(g) Cropland supply response to cropland rents
εNLAND(g) Non-land supply response to non-land prices
θLAND(g) Cost share of croplands
θNLAND(g) Cost share of non-land inputs
θCRPFOOD(y) Cost share of crop inputs in the processed food sectors
θNCRPFOOD(y) Cost share of non-crop inputs in the processed food sectors
θCRPFEED(y) Cost share of feeds in the livestock sectors
θNCRPFEED(y) Cost share of non-crop inputs in the livestock sectors
Variables
Note: Lower case letters refer to the percentage change in the LEVELS vari-
ables (i.e. UPPER CASE letters). They are linked in the model code through
update equations and the non-linear model is solved as an initial value problem
using the GEMPACK program.4
Quantities
qPC(i,y) Per capita commodity demand
q(i,y) Regional commodity demand
xLAND(g) Cropland area
xNLAND(g) Non-land input quantity
xCRPFEED(y) Feeds used in the livestock sectors
xNCRPFEED(y) Non-crop inputs used in the livestock sectors
xCRPFOOD(y) Crop inputs used in the processed food sectors
xNCRPFOOD(y) Non-crop inputs used in the processed food sectors
xCRPBIOF Crop feedstock used in the global biofuel sector
Prices
pCROP Global crop price
p(i,y) Price of commodity
pLAND(g) Cropland rent
pNLAND(g) Non-land prices

 Harrison, W. J., & Pearson, K. R. (1994). Computing solutions for large general equilibrium
4

models using GEMPACK. Monash University, Centre of Policy Studies/IMPACT Centre.

Appendix C  Description of the SIMPLE Model 173

pNCRPFOOD(y) Price of non-crop inputs in the processed food sectors

pNCRPFEED(y) Price of non-crop inputs in the livestock sectors
Other variables
yPC(y) Per capita income
pop(y) Population
TFPCROP(g) Total factor productivity in the crop sector
TFPLIVESTOCK(y) Total factor productivity in the livestock sector
TFPPROC_FOOD(y) Total factor productivity in the processed food sector
AO(g) Input neutral productivity change in the crop sector
AF(i,y) Input-biased productivity change (livestock and processed
food sectors only, non-crop input-augmenting technical
change)
Equations
There are three broad types of equations in the model: consumer demands,
food and agricultural supplies, and commodity market clearing. Consumer
demands are simple log-linear relationships in which the own-price and income
elasticities vary as a function of per capita income level. The supply equations
for crops, livestock and processed foods are based on non-linear Constant
Elasticity of Substitution (CES) production functions. These are readily
expressed in linearized form (i.e. percentage change) as shown below. Note that
when this model is solved with the linearized-levels variable linkages, we obtain
the same solution as would be obtained by implementing the model in levels
form. There is only one commodity market clearing condition in this model, and
that is for crops at global scale. For more details on this mixed, linearized-levels
representation of an economic model, see Hertel, Horridge, and Pearson (1992).5
Consumer demand equations
e P (i ,y ) = a P (i ) + b P (i ) ln éYPC ( y ) ù Predicted price elasticities wrt. per capita
ë û
income level
e Y (i ,y ) = aY (i ) + bY (i ) ln éYPC ( y ) ù Predicted income elasticities wrt. per capita
ë û
income level
qPC (i ,y ) = e P (i ,y ) p(i ,y ) + e Y (i ,y ) yPC ( y ) Per capita commodity demand
q(i ,y ) = qPC (i ,y ) + pop( y ) Regional commodity demand
Crop supply/production equations
xLAND ( g ) = xCROP ( g ) - ao( g ) - s CROP ( g ) é pLAND ( g ) - pCROP - ao( g ) ù Derived demand
ë û
for cropland
x NLAND ( g ) = xCROP ( g ) - ao( g ) - s CROP ( g ) é pNLAND ( g ) - pCROP - ao( g ) ù Derived demand
ë û
for non-land
pCROP = ao( g ) + q LAND ( g ) pLAND ( g ) + q NLAND ( g ) pNLAND ( g ) Zero profit
condition

5
 Hertel, T. W., Horridge, J. M., & Pearson, K. R. (1992). Mending the family tree: A reconciliation
of the linearization of levels schools of applied general equilibrium modeling. Economic Modeling,
9, 385–407.
174  Appendix C Description of the SIMPLE Model Appendi

xLAND ( g ) = e LAND ( g ) pLAND ( g )

x NLAND ( g ) = e NLAND ( g ) pNLAND ( g ) Cropland supply
Non-land supply
Livestock supply/production equations
xCRPFEED ( y ) = q(“ Livestock ”, y ) - s CRPFEED é pCROP - p(“ Livestock ”, y ) ù
ë û Derived demand for
feeds
x NCRPFEED ( y ) = q(“ Livestock ”, y ) - af(“ Livestock ”, y ) - s CRPFEED
é pNCRPFEED y - p “ Livestock ”, y - af “ Livestock ”, y ù
ë ( ) ( ) ( )û
Derived demand for
non-crop inputs
p(“ Livestock ”, y ) = qCRPFEED ( y ) pCROP + q NCRPFEED ( y )
é pNCRPFEED y - af “ Livestock ”, y ù
ë ( ) ( )û
Zero profit condition
Processed food supply/production equations
xCRPFOOD ( y ) = q(“ Proc _ Food ”, y ) - s CRPFOOD
é pCROP - p “ Proc _ Food ”, y ù
ë ( )û
Derived demand for
crop inputs
x NCRPFOOD ( y ) = q(“ Proc _ Food ”, y ) - af(“ Proc _ Food ”, y ) - s CRPFOOD
é pNCRPFOOD y - p “ Proc _ Food ”, y - af “ Proc _ Food ”, y ù
ë ( ) ( ) ( )û
Derived demand for
non-crop inputs
p(“ Proc _ Food ”, y ) = qCRPFOOD ( y ) pCROP + q NCRPFOOD ( y )
é pNCRPFOOD y - af “ Proc _ Food ”, y ù
ë ( ) ( )û
Zero profit condition
Market clearing equations
pCROP = p(“ Crops ”, y )
Integrated world price
7 5 for crops
å
g =1
XCROP ( g ) = å éQ(“ Crops ”, y ) + XCRPFEED ( y ) + XCRPFOOD ( y ) ù
i =1
ë û
+ XCRPBIOF
Market clearing for crops
TFP equations
TFPCROP ( g ) = AO( g )

PNCRPFEED ( y ) X NCRPFEED ( y )
TFPLIVESTOCK ( y ) = AF(“ Livestock ”, y ) *
P(“ Livestock ”, y )Q(“ Livestock ”, y )

PNCRPFOOD ( y ) X NCRPFOOD ( y )
TFPPROC _ FOOD ( y ) = AF( Proc _ Food , y )
*
P(“ Proc _ Food ”, y )Q(“ Proc _ Food ”, y )

Appendix C  Description of the SIMPLE Model 175

2. Description of model tuning process

As with any computational model, some tuning is necessary in order to ensure
reasonable performance of the full equilibrium model. We refrain from tuning
SIMPLE over the same period for which the historical validation is undertaken
(i.e. 1961–2006). Instead, we focus on the period 2001–2006 for tuning pur-
poses. We focus the model tuning on three key dimensions of global
agriculture.
(a) We ensure that our economic yield response to crop price matches that from
the literature.6 Specifically, a 1% increase in global crop price should trans-
late to a 0.25% increase in crop yields
(b) We do not have robust estimates for the unobserved intensification parame-
ters in the livestock and food processing sectors. For the livestock sector, we
‘calibrate’ this parameter by focusing on the high income region, which we
deem most representative of future developments in the livestock industry,
and choose the parameter which best fits the data on crop input use for this
region, over this period. This value is subsequently assigned to all demand
regions. For the processed food sector, we set the elasticity of substitution
between crop inputs and non-crop inputs for the processed food sectors to
zero under the assumption that this relationship is fixed over time.
(c) We observed that in our initial calibration effort, the simulated change in
global crop demand for food (10.9%) is nearly one-quarter greater than the
historical change (around 8.8%). We adjusted the regression estimates of the
income and price elasticities by re-estimating the linear regressions of the
demand elasticities with per capita incomes using deflated per capita incomes
(divided by a factor of 4). This dampens the magnitude of the regression
intercepts while maintaining the values of the regression slopes.
3. Supplementary tables
Table C.1  Base data for year 2001
Crop Price of
Crop production data output Value of land Croplands Non-land non-land
East Asia & Pacific 1722 32,856 265,241 149,676 1.00
Europe & Central Asia 1251 34,478 350,493 98,128 1.00
Latin America & 689 6573 155,009 66,461 1.00
Caribbean
Middle East & North 200 3816 49,368 17,384 1.00
Africa
North America 717 19,761 230,211 56,241 1.00
South Asia 838 15,989 205,137 72,839 1.00
Sub-Saharan Africa 361 3444 144,979 34,822 1.00
(continued)

6
 Keeney, R., & Hertel, T. W. (2009). Indirect land use impacts of US biofuels policies: The impor-
tance of acreage, yield and bilateral trade responses. American Journal of Agricultural Economics,
91, 895–909.
176 Appendix C  Description of the SIMPLE Model

Table C.1 (continued)
Crop Price of
Crop production data output Value of land Croplands Non-land non-land
Units Million Mt. Million USD Thou. Ha. Million Index
USD (2001) = 1
Livestock production Feed Non-feed Price of
data non-feed
Upper high 457 1,320,324 1.00
Lower high 2 3595 1.00
Upper middle 102 143,402 1.00
Lower middle 379 212,636 1.00
Low 86 28,765 1.00
Units Million Mt. Million USD Index (2001) = 1
Processed food Crop inputs Non-crop Price of
production data non-crops
Upper high 579 654,547 1.00
Lower high 2 1625 1.00
Upper middle 465 231,879 1.00
Lower middle 620 196,061 1.00
Low 582 236,728 1.00
Units Million Mt. Million USD Index (2001) = 1
Value of consumption Crops Livestock Processed food Non food
Upper high 47,318 1,368,779 715,915 22,144,970
Lower high 355 3855 1794 102,908
Upper middle 15,210 154,221 281,178 1,768,666
Lower middle 117,193 252,802 261,788 1,322,221
Low 80,711 37,853 298,465 558,426
Units Million USD
Quantity of Crops Livestock Processed food Non food
consumption
Upper high 446 1,226,521 676,532 21,937,334
Lower high 3 3650 1586 101,371
Upper middle 143 139,037 243,646 1,692,069
Lower middle 1106 222,645 222,594 1,224,559
Low 761 32,403 227,299 507,690
Units Million Mt. Million USD
Other data Income Population Crop use in
biofuels
Upper high 28,705 856
Lower high 17,051 9
Upper middle 4933 494
Lower middle 1446 2090
Low 472 2142
Global 43
Units USD per Millions Million Mt.
capita
Appendix C  Description of the SIMPLE Model 177

Table C.2  Base data for year 2006

Price of
Crop production data Crop output Value of land Croplands Non-land non-land
East Asia & Pacific 1935 39,952 266,311 163,728 1.11
Europe & Central Asia 1372 40,545 352,415 105,610 1.09
Latin America & 780 7955 159,755 71,732 1.12
Caribbean
Middle East & North 220 4504 50,035 18,678 1.10
Africa
North America 790 23,372 231,357 60,706 1.10
South Asia 927 18,953 208,149 78,378 1.10
Sub-Saharan Africa 386 3906 147,877 36,623 1.08
Units Million Mt. Million USD Thou. Ha. Million USD Index
(2001) = 1
Livestock production Feed Non-feed Price of
data non-feed
Upper high 461 1,358,002 1.00
Lower high 2 4036 1.00
Upper middle 103 154,065 1.00
Lower middle 383 240,943 1.00
Low 108 33,780 1.00
Units Million Mt. Million USD Index
(2001) = 1
Processed food Crop inputs Non-crop Price of
production data non-crops
Upper high 621 660,268 1.00
Lower high 1 1816 1.00
Upper middle 605 236,636 1.00
Lower middle 726 230,162 1.00
Low 639 292,779 1.00
Units Million Mt. Million USD Index
(2001) = 1
Value of consumption Crops Livestock Processed Non food
food
Upper high 50,676 1,410,450 730,947 24,622,334
Lower high 409 4308 1985 125,389
Upper middle 16,825 165,765 305,490 2,162,496
Lower middle 132,568 284,551 312,830 2,068,221
Low 95,989 46,091 365,538 812,772
Units Million USD
Quantity of Crops Livestock Processed Non food
consumption food
Upper high 456 1,287,137 713,647 24,391,476
Lower high 4 4140 1813 123,516
Upper middle 151 154,964 272,304 2,068,843
Lower middle 1192 276,645 272,620 1,915,459
Low 863 39,461 285,925 738,928
Units Million Mt. Million USD
(continued)
178 Appendix C  Description of the SIMPLE Model

Table C.2 (continued)
Price of
Crop production data Crop output Value of land Croplands Non-land non-land
Other data Income Population Crop use
in biofuels
Upper high 30,853 885
Lower high 18,640 10
Upper middle 5623 523
Lower middle 2060 2169
Low 602 2316
Global 95
Units USD per Millions Million
capita Mt.

Table C.3  Parameters for 2001–06 tuning and 1961–06 validation experiments
Elasticities of substitution 2001 to 2006 1961 to 2006
Livestock 1.16
Processed Food 0
Crops 0.55
Non-land supply response 0.49 1.34
Land supply response
East Asia & Pacific 0.04 0.11
Europe & Central Asia 0.04 0.11
Latin America & Caribbean 0.20 0.55
Middle East & North Africa 0.11 0.29
North America 0.04 0.11
South Asia 0.10 0.28
Sub-Saharan Africa 0.20 0.55
Income elasticities
 Regression Intercept
  Crops 0.88
  Livestock 1.05
  Processed Foods 1.20
  Non-food 1.56
 Regression Slope
  Crops −0.10
  Livestock −0.09
  Processed Foods −0.10
  Non-food −0.05
Price elasticities
 Regression Intercept
  Crops −0.74
  Livestock −0.83
  Processed Foods −1.17
  Non-food −1.14
(continued)
Appendix C  Description of the SIMPLE Model  179

Table C.3 (continued)
Elasticities of substitution 2001 to 2006 1961 to 2006
 Regression Slope
  Crops 0.07
  Livestock 0.05
  Processed foods 0.08
  Non-food 0.04

Table C.4  Detailed results of the historical validation experiment (1961–2006)

Actual dataa
Geographic regions Crop productionb Crop landc Crop yields Crop priced
East Asia & Pacific 309.0 33.2 207.0
Europe & Central Asia 48.8 −13.5 72.1
Latin America & the Caribbean 295.5 63.7 141.7
Middle East & North Africa 354.7 27.1 257.9
North America 153.5 −7.2 173.1
South Asia 222.4 7.6 199.8
Sub-Saharan Africa 300.0 48.2 169.8
Global 196.4 15.7 156.2 −28.8

Simulation
Geographic regions Crop production Crop land Crop yields Crop price
East Asia & Pacific 291.6 14.9 240.9
Europe & Central Asia 229.8 12.1 194.2
Latin America & the Caribbean 243.4 46.6 134.2
Middle East & North Africa 445.7 46.9 271.5
North America 197.8 10.7 169.2
South Asia 113.4 13.6 87.9
Sub-Saharan Africa 77.1 12.6 57.3
Global 206.0 16.7 162.1 −24.0
a
Actual data based on 90 countries using 1961 data and average data from 2004 to 2008 to repre-
sent 2006
b
Production data includes 50 crops which are aggregated by converting quantities to corn equiva-
lent using 2004/08 relative prices
c
Refers to “Arable land and permanent croplands” as defined by FAO
d
Based on the world price index data for food grains from the Global Economic Monitor Data by
the WorldBank
Index

A ethanol, cellulosic, 101

Africa, 3, 4, 13–15, 18–20, 22–24, 31, 43, 48,
50, 53, 58, 60, 72, 75, 78, 80, 86,
87, 95, 127–130, 134, 136, 142,
143, 148, 150–152, 157, 158
Agricultural, 1, 2, 6–10, 35, 57
intensive agriculture, 3
investments, 28, 77
irrigated agriculture, 59–61
land, 109
land use, 57, 145 see also Cropland
models, 144, 145, 147
output, 6, 9, 27, 31, 35, 60, 109, 141, 149
production, 35, 62, 77, 79, 126, 127 see
also Crop production
productivity, 2, 4, 46, 49, 62, 70, 73, 79,
102, 108, 126, 131–137, see also
Crop productivity
research and development (R&D), 7, 9, 28,
31, 78, 79
yields see also Crop yields
Analytical framework, 2, 7–9, 41, 109 see also
Economic framework
B Climate Change Mitigation, 69, 81, 94
Biodiesel
Biofuel, 100
Biodiversity
payment for hydrological environmental
services of forests, 89
payments for environmental service, 89
Biofuel, 4, 10, 44, 49, 51, 95, 100, 102–112,
146, 153, 154
demand, 108, 110, 111
ethanol, 101
ethanol, corn, 102, 104, 106
ethanol, sugarcane, 99
feedstock, 101, 102, 108, 109
C
Caloric intake, 118, 126, 133, 136
Calories, 132
caloric see also Dietary Energy
Central Asia, 18, 19, 22, 24, 80, 95, 134, 136,
143, 148
China, 4, 18, 19, 22, 25, 26, 32, 35, 60–62, 78,
87, 94–96, 100, 106, 115, 117, 126,
133–135, 142
Climate change, 59, 69–81, 128–131
adaptation, 69, 75–81, 130
impacts, 9
mitigation policy, 2
pests and diseases, 72
precipitation, 69, 73–75, 79–81
rainfall, 69, 71, 75, 77
temperature, 69–74, 76, 79–81
variation, 75
yield impacts, 129
Climate Change Policy
carbon contracts, 87, 90
forest carbon sequestration, 94
REDD+, 92, 96
CO2
concentration, 71, 72, 76
effects, 79
emission, 64, 85
fertilization, 79, 81, 133, 136
levels, 69, 71, 76

© Springer International Publishing Switzerland 2016 181

T.W. Hertel, U.L.C. Baldos, Global Change and the Challenges of Sustainably
Feeding a Growing Planet, DOI 10.1007/978-3-319-22662-0
182
Consumption, 9, 16–23, 43, 44, 59, 62,
115–120
processed food consumption, 18, 19, 21,
23
Crop, 27, 37, 147
demand, 25, 43, 46, 108, 109, 111, 115,
121, 122
development, 69, 70, 72, 74, 76
input, 121, 123, 124
land, 37, 57, 59, 60, 62–65, 80, 102,
104–107, 115, 117, 119–123, 147,
150, 151, 153, 154, 156, 158
land area, 41, 44, 46–48, 53, 54
land conversion, 48, 51
land expansion, 41, 46, 48
land supply, 46–50
land use, 19, 23–25, 41, 51, 65
models, 73, 74, 76, 129, 133
output, 19, 27, 32, 34, 44, 57, 115, 120
prices, 33, 37, 44, 65, 81, 121–123, 147,
see also Food prices
production, 32, 37, 49, 57, 64, 65, 69, 79,
107, 110, 111, 115, 119, 121–123,
128, 129, 133, 143, 145
productivity, 7, 73, 74 see also Agricultural
productivity
yields, 27, 31–35, 37, 44–46, 73, 74, 81,
105, 107, 123, 128, 129, 131, 133,
136
D prices, 8, 9, 28, 31, 33, 34, 42, 43, 62, 80,
Deforestation, 4, 6, 7, 46, 49, 51, 130
Dietary energy See Calories;Caloric
Drought, 44, 61, 63, 69, 70, 77, 78, 127, 129
E supply, 125, 129
Economic framework, 6–10, 32, 38, 51, 81, 85
Economic margins
demand margin, 51, 52
extensive margin of supply, 42, 45–50
intensive margin of supply, 41, 44, 45, 51
Economic margins of response
extensive margin of supply, 93
intensive margin of supply, 93
Economic responses to scarcity, 9, 51, 53
Ecosystem services, 50
Elasticity
crop supply, 93
crop yields, 44–46
cropland supply, 46–50
land supply, 49
land supply elasticity, 9
Index
price elasticity of demand, 7–8, 42,
43, 52
price elasticity of demand for agricultural
output, 109
price elasticity of demand for crops, 93
price elasticity of excess demand, 52
price elasticity of land supply, 109
substitution, 45
yields with respect to commodity price,
109
Environment
demand for environmental services, 7, 10
environmental change, 7, 10
environmental impacts, 6, 7
environmental sustainability, 1
Environmental impacts, 46, 48, 117–118,
150–152
Externalities, 85–86
F
Feed, 1, 32, 57, 59, 121, 125
Feedstuff, 44, 59, 119, 121
Fertility rates, 13
Food, 28, 32
accessibility, 125, 130, 136
availability, 125, 128, 129, 131, 136
consumption, 34, 41–43, 125, 128, 129,
145, 147
demand, 4, 10, 28, 34, 42–44, 52, 129,
145, 155
125, 129, 130, 136, see also Crop
prices
production, 2, 62, 125, 126, 129, 130
security, 125–137
sector, 6, 37, 133, 142
system, 1, 2, 7, 38, 41, 51, 53, 125, 128,
141, 156
Forest carbon sequestration; land–based
carbon sequestration, 94
G
GHG emissions, 7, 46, 47, 54, 80, 81, 88, 90,
94, 102, 105, 106, 108, 118, 119,
151, 152, 156
Green revolution (GR)
African Green Revolution, 53
historical Green Revolution, 42, 52, 53
Greenhouse gas emissions, 2, 118 see also
GHG emissions
Greenhouse Gasses, 118
Index 183

H N
Heat effects, 74 Non-crop input, 120, 121, 123
Heat stress, 70, 74 Non-land inputs, 6, 32, 34, 37, 42, 44, 46, 64,
146, 148

I
Income, 20–26, 37, 42, 43, 50, 64, 76, 78, 80,
128, 130, 143–145, 148, 152, 155,
158
per capita income, 10
level of income, 20, 23
Income growth, 133, 142, 153–155
Input prices, 33, 34, 37, 44, 45
Input substitution, 32, 37, 44, 45,
121–123
Integrated markets, 148–151, 157
Intensification, 7–9, 27, 34, 35, 41, 46, 49,
64, 152
endogenous intensification, 38
intensification of production, 36
intensive production, 33
Intensive crop production, 33, 145
Irrigation
areas equipped for irrigation, 59, 60
irrigated agriculture, 59–61
irrigated areas, 63–65
irrigated crop production, 64
irrigated cropping systems, 65
Isocost, 33, 34
Isocost line, 33, 34, 64
Isoquants, 32, 33, 64
P
Per capita income, 22, 80, 95, 133, 134, 142,
143, 145, 148, 153, 157
Photosynthesis, 29, 69–71, 74
Population, 1, 2, 4–6, 10, 13–16, 19, 23–26,
28, 32, 37, 42, 43, 50, 60, 64, 80,
94, 95, 126, 127, 130–134, 136,
142–144, 150, 153, 156
Population growth, 4, 26, 94
Potential yield, 10, 30, 136
Price, 7, 8, 10, 34, 37, 42–52, 63, 65, 77, 80,
125, 127, 129, 131, 132, 136, 143,
145, 147, 154, 155
oil price, 100, 101
shocks, 99
spikes, 100
Processed Food, 120–124, 134
Projections, 5, 9, 10, 28, 31, 51, 72, 108,
129–131, 133, 137, 142, 144, 145,
147, 148, 153, 154, 156
R
Rainfed areas, 64
Rainfed cropland, 64, 65
Rainfed systems, 65

L
Land conversion, 42, 48, 49, 65, 69, 80, 81,
151–152, 156
Land use change, 2, 3, 6, 8, 9, 45, 48, 52, 64,
69, 80, 147, 151, 154
Land-sparing, 46, 51–54, 91
Livestock, 10, 18, 20, 22, 37, 44, 59, 80, 95,
103, 104, 115–124, 133, 134, 142,
143
Livestock feed, 44
M Sub Saharan Africa, 94
Malnutrition, 2, 156, 157
malnutrition count, 136
malnutrition incidence, 136
Margins of economic response
demand margin, 7, 9
extensive margin of supply, 8, 109
intensive margin of supply, 8, 109
Meat, 115, 129
Milk and eggs, 117
S
Segmented markets, 149–151, 157
SIMPLE model, 18, 19, 21, 22, 25, 36, 38, 52,
64, 69, 79, 80, 94, 108, 110, 112,
126, 129, 141, 142, 148, 150, 153,
155
South Asia, 4, 18–20, 22, 23, 25, 26, 43, 48,
62, 80, 95, 96, 129, 130, 133–136,
143, 148, 157
Southeast Asia, 4, 18, 19, 22, 25, 26, 61, 95,
106, 129, 134
Substitution effect, 122, 123
Supply of land, 7, 148
T
Technological change, 37, 54, 150
Technological improvements, 33, 34, 37
Total factor productivity (TFP), 32–38, 64,
131, 133, 134, 142, 144, 145, 154
184 Index

Trade, 4, 36, 50, 62, 63, 77, 148–150 ground water supply, 61
Tropical cropping systems, 79 irrigation, 73, 76
Tropical systems, 70–73, 76 irrigation supply, 61
irrigation water, 57, 59–62
soil moisture, 71, 72, 76
U water scarcity, 62–63, 65
Undernutrition See also Malnutrition water supply, 60
water use, 71, 72
water withdrawals, 57, 61
W
Water
efficiency of water use, 62 Y
ground water, 61 Yield gaps, 30, 31