PERSPECTIVE

PUBLISHED ONLINE: 5 APRIL 2017 | DOI: 10.1038/NCLIMATE3253

Climate change through a poverty lens

Stephane Hallegatte1 and Julie Rozenberg2*

Analysis of the economic impact of climate change typically considers regional or national economies and assesses its impact on
macroeconomic aggregates such as gross domestic product. These studies therefore do not investigate the distributional impacts
of climate change within countries or the impacts on poverty. This Perspective aims to close this gap and provide an assessment
of climate change impacts at the household level to investigate the consequences of climate change for poverty and for poor peo-
ple. It does so by combining assessments of the physical impacts of climate change in various sectors with household surveys. In
particular, it highlights how rapid and inclusive development can reduce the future impact of climate change on poverty.

V
ery few quantified estimates of the impact of climate change
on poverty have been proposed1,2. The economic impact of
climate change is almost always analysed at aggregated lev-
els3. Studies consider regional or national economies, with various
sectoral breakdowns, to assess the impact of climate change on
macroeconomic aggregates such as gross domestic product (GDP)
or aggregate consumption. Most studies therefore show that poor
countries are more vulnerable than rich coutriesnot that poor
people are more vulnerable than rich people412. A few studies have
investigated the implications of such aggregate impacts for house-
holds, and have looked at impacts on poverty and poor people13. But
these studies follow a top-down approach where aggregate impacts
are estimated first, and the micro-level consequences for households
are considered second.
There are good reasons to start with the impact of climate change
on economic growth when investigating the impact on poverty. We
know that aggregate economic growth is fundamental for poverty
reduction: in the past decades, most of the reduction in poverty was
achieved by growing the size of economies, not by redistributing
income14,15. But economic growth is not the only channel through
which climate change can affect poverty, and aggregated approaches
are insufficient.
First, climate impacts on aggregate economic metricssuch as
GDPand impacts on poor people may be only weakly correlated.
Poor people represent an extremely small share of national income
for instance the income of the poorest 20% of the households in
Panama represent only 3% of GDPso the impacts of climate change
on poor people may have almost no impact on national income.
One can imagine a scenario where only the poorest are affected by
climate change: in that case, GDP would barely change, but poverty
may increase nevertheless. Second, the sectors and occupations rep-
resented in models of national economies may not be the ones that
provide income to many poor people. For instance, non-commercial
extraction from forest and other ecosystems represents up to 30% of
consumption in some poor tropical communities, reducing poverty
by up to 14% (ref. 16). Such consumptions are not represented in
national accounts and macroeconomic models, which cannot capture
the poverty effects of climate change impacts on ecosystems.
The poverty impact of climate change
In a recent World Bank report (Shock Waves: Managing the Impacts
of Climate Change on Poverty17) and in 15 associated background
papers, we analysed the impacts of climate change on poverty using
a bottom-up approach starting from the impacts of climate change
at the household level, in the spirit of previous work on food prices
and poverty 18,19. This section summarizes the main findings of this
report on the impact of climate change on poverty and poor peo-
ple, but also reports on new results from analyses done after the
publication of the report.
Findings confirm that poor people may be heavily affected
by climate change even when impacts on the rest of the popula-
tion remain limited. Many household surveys with self-reported
shocks show that poor people are more often affected by environ-
mental shocks, that poor people are losing morerelative to their
wealthwhen they are affected by a shock, and that poor people
receive less post-shock support from friends and family, the finan-
cial system, and social safety nets. Using Demographic and Health
Survey (DHS) data and hazard maps, we find that poor people are
more often exposed to floods, droughts, and extreme heat20,21. In
Nigeria, for instance, the most poor 20% of people are 50% more
likely to be affected by a flood, 130% more likely to be affected by a
drought, and 80% more likely to be affected by a heat wave than the
average Nigerian. Case studies in Bangladesh, India, and Honduras
also suggest that poor people are losing two to three times more
than non-poor people when hit by a flood or storm. Climate-related
shocks can keep people in poverty by making it more difficult for
households to accumulate assets, regularly wiping out their stock
of assets, or even creating irreversible impacts on human capital
(through health or educational impacts)2228.
These findings support a bottom-up approach, based on individual
or household-level vulnerability, instead of a macro-level approach.
To follow such an approach, we use a global database with 92 house-
hold surveys that describe the current distribution of income and
occupations in 92 countriesthe International Income Distribution
Data Set, created at the World Bank. Then, we use micro-simulation
techniques2932 to project the evolution of these households until
2030, driven by demographics and socioeconomic changes.
In our model, the population of each country is represented by
a set of thousands of representative households, described by the
share of the countrys population they represent (which is referred
to as their weight) and their characteristics, namely the number
of people in the household and their age, education level, sector of
employment, employment status, and income. To model the rep-
resentative households of the future, we change the income and
weight of each household in the model to reflect macro-economic
changes. The households weights are adjusted so that the total

Global Facility for Disaster Reduction and Recovery, World Bank, 1818 H Street NW, MC 6-606, Washington DC 20433, USA. 2Office of the Chief Economist,
1

Sustainable Development Practice Group, World Bank, 1818 H Street NW, MC 6-606, Washington DC 20433, USA. *e-mail: jrozenberg@worldbank.org

250 NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3253
population matches different age and skills compositions of the
population, participation in the labour force, and the labour share
of each sector, based on demographic scenarios developed for the
Shared Socioeconomic Pathways33 and assumptions on structural
change. This reweighting process models demographic changes (for
example, older people or more skilled people) and structural trans-
formations (for example, less people in unskilled agriculture). For
instance, a scenario with rapid structural and demographic change
will reduce the weight of representative households with low skills,
with members working in the agricultural sector, and with many
children, and will increase the weight of households with skilled
workers working in services, so that the aggregate characteristics
of the population is consistent with the aggregate scenario. Incomes
of each individual then evolve over time, based on assumptions
regarding productivity growth and skill premiums for the sector the
individual belongs to, and income redistribution through pensions
and social systems34 (see details in the Supplementary Material).
We used this technique to generate 1,200 reference scenarios
(scenarios without climate change) per country, by randomly com-
bining values within plausible ranges for the model parameters. As
discussed in the Supplementary Material, the parameters are treated
as independent, meaning that we can run baseline scenarios that
lack international consistency. Those scenarios are however not used
to predict the future but to investigate the conditions that can lead to
more or less optimistic poverty outcomes by 2030.We then use these
sets of conditions to create two representative reference scenarios
based on two Shared Socioeconomic Pathways (SSPs), SSP4 and
SSP535,36. The prosperity scenario has the population change of SSP5
and assumes rapid growth, a rapid decrease in the number of peo-
ple living below the extreme poverty line (reaching less than 3% of
the global population in 2030), low inequality, and universal access
to basic infrastructure services by 2030.In the poverty scenario, by
contrast, demographic changes are based on SSP4, growth is slower,
extreme poverty decreases much less (remaining at around 11% of
the global population in 2030), and inequality is much greater across
and within countries. Our prosperity scenario can be considered as
a quantified pathway for poverty in the SSP5 while our poverty sce-
nario can be considered a quantification of poverty in SSP4but
going only until 2030due to the limits of the methodology.
We then introduced in each of these two reference scenarios the
impacts of climate change on the income and real consumption of
each of the 1.4 million households that are in the database. In our
work so far, we took into account five channels, based on background
work realized for the report and on recent papers on the impacts
of climate change. In each channel, we consider an optimistic and
a pessimistic climate impact scenario. The difference between the
optimistic and pessimistic scenarios is not related to greenhouse
gas emissionsby 2030, changes in emissions have limited impacts
on the magnitude of climate changeand we deliberately leave
out the impact of climate mitigation policies on poverty through
energy prices. Instead, the difference between the two scenarios
comes from the sensitivity of the climate system and how changes
in climate conditions translate into physical impacts, for instance
on flood likelihood or agricultural yields (details are available in the
Supplementary Material). The five channels are:
1. The impact of climate change on agricultural productivity and
pricesas estimated by a global agro-economic model37and
its consequences for the income of farmers. The final impact
on poverty depends on the fraction of the population working
in agriculture, their productivity, and income diversification
within households.
2. The impact of climate change on food prices (using the same
modelling source), and the consequences of this for consum-
ers. The impact on poverty depends on the fraction of each
households budget that is dedicated to food consumption. In all
regions this fraction is decreasing with the household income.
PERSPECTIVE
3. The impact of climate change on natural disasters, assuming that
by 2030 the fraction of the population that will be affected annu-
ally by a flood or a drought increases by 20% in the optimistic
climate change scenario, and by 40% in the pessimistic climate
change scenario, and that the fraction of the population affected
annually by a storm surge or a cyclone increases by 10% in the
optimistic climate change scenario and by 50% in the pessimistic
climate change scenario38. Impacts are modelled as an income
loss following a methodology published elsewhere39.
4. The impact of climate change on labour productivity using recent
evidence suggesting that 1C in warming could reduce labour
productivity by 13%, at least for people working outdoors or
without air conditioning 21,40,41.
5. The impact of climate change on child stunting, malaria, and diar-
rhoea, as estimated by recent analyses4244. Impacts are modelled
through fixed treatment costs and number of working days lost.
All of these channels are modelled in a simple way, more as a dem-
onstration of the methodology than as a definitive assessment. There
is no double counting of the impacts because they represent dif-
ferent channels and physical mechanisms, however, it is very likely
that there are strong interactions across these channels, which may
increase or decrease the total impact. For instance, under-nutrition
makes people more vulnerable to many diseases, and this is not
taken into account. Similarly, no model of agricultural production
takes into account the effect of higher temperature on the ability of
farmers to work.
In addition, some potentially important channels are missing,
such as the effect on energy access (for example, through availabil-
ity of hydropower, cooling water for thermal plants, or biomass),
ecosystems, and specific sectors like tourism and electricity genera-
tion. Nevertheless, the impact on the number of people in extreme
poverty assessed with this methodology suggests the potential for a
large impact of climate change on poverty.
In the poverty scenario, the number of people in extreme poverty
in 2030 because of climate change is 35 million and 122 million,
in the optimistic and pessimistic climate impacts scenarios, respec-
tively. These results also emphasize how rapid and inclusive develop
mentas represented by our prosperity scenariocan prevent
most of the impacts of climate change on poverty. In this reference
scenario, the number of people in extreme poverty only because of
climate change is between 3million and 16 million, depending on
the climate impact scenarioa much smaller impact. The analysis
suggests that the dominant channel is through agricultural impacts
and effects on food prices, as shown in Fig. 1. These results also
provide a first-cut assessment of countries vulnerabilities, and can
guide further research towards the channels that are most likely to
affect the poorest. They can therefore help design climate change
adaptation policies that are targeted toward the poor and contribute
to poverty reduction.
After the publication of the Shock Waves report 17, and to bet-
ter understand the uncertainty on our results, we simulated these
climate change impacts not only in the poverty and prosperity ref-
erence scenarios but also in the 1,200 reference scenarios in each
country. We then sampled the country results to create global distri-
butions of impacts: we created 10,000 global baselines by randomly
selecting baselines in each country, and added low climate change
impacts and high climate change impacts in each of them. This
additional study confirms that the two most important channels
through which climate change affects poverty are the effect of agri-
culture prices on consumers and health impacts through malaria,
diarrhoea, and stunting (Fig.2).
A new finding is that the largest uncertainty stems from the effect
of agricultural pricesand that the worst scenarios in terms of cli-
mate change impacts on poverty are those in which the impacts of
higher agricultural prices are the worst (Fig.2). This is because the
impact of agriculture prices on poverty strongly depends on baseline

NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange 251

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
PERSPECTIVE NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3253

Agriculture
Prosperity scenario
(max. impact) Health
Labour productivity
Disasters
Prosperity scenario
(min. impact)

Poverty scenario
(max. impact)

Poverty scenario
(min. impact)

0 20 40 60 80 100 120 140

Additional people below the extreme poverty threshold (million)

Figure 1 | Number of additional people in extreme poverty in 2030 because of climate change, in the four scenarios. We use two socioeconomic
scenarios poverty and prosperity with different assumptions on population and productivity growth and inequality and two climate impacts
scenarios, with low-impact and high-impact assumptions on climate change physical impacts.

120
Climate change impact scenario
People living in extreme poverty by 2030 due to climate change

Low
100
High

80
(million)

60

40

20

Agriculture revenues Labour productivity Food prices Disasters Health

Figure 2 | Climate change impacts through each of the five channels, sampled from 1,200 scenarios in 92 countries. Each cell shows the number of
scenarios with a given income of the poor in the baseline in 2030 and a given impact of climate change on the impact of the poor.

assumptions and in particular on poor peoples income (the poorer people are particularly bad (for instance, larger than 5% of their
the people, the more they lose as a fraction of total consumption income) are mostly scenarios in which poor people would be poor
when food prices increase). This finding has concrete implications: in 2030 even in the absence of climate change. It shows very clearly
in a risk-based framework where bad scenarios are particularly that climate changeat least by 2030is a poverty multiplier: it
important, it means that knowing more about possible bad out- makes poor people poorer, and it increase the poverty headcount.
comes for agriculture pricesand how to avoid themshould be In terms of interventions, it means that policies that reduce pov-
apriority. erty in the scenario without climate change also reduce the impact
The initial study showed that rapid development is a good way of climate change. In other terms, poverty reduction, if well done,
of reducing the impact of climate change, at least up to 2030. The is adaptation.
new scenarios also confirm this finding and the idea that lifting
people out of poverty as soon as possible would reduce the future Comparison with previous approaches
impacts of climate change on poverty. Figure3 shows the impact This bottom-up approach also allows us to assess aggregated
of climate change on the income of the poor (here, the bottom impacts by summing up the individual-household impacts
20% in terms of income or consumption in each country) versus resulting in what we call national private household income losses,
the income of the poor in the reference scenario (without climate as a proxy for the gross national income (GNI). Some high-income
change) in 2030. Each cell shows the number of scenarios with a households may be missing from the surveys so we do not capture
given income of the poor in the reference scenario in 2030 and a 100% of the national private households. This calculation only
given impact of climate change on the income of the poor. It also accounts for a subset of all economic impacts, since it focuses on
shows that the scenarios where climate change impacts on poor households, and provides an estimate that is largely independent of

252 NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3253 PERSPECTIVE

pearsonr = 0.38; P = 0 pearsonr = 0.33; P = 0

Income loss of the bottom 20% due to climate change

20
8

Average national private household loss in 2030 (%)

6
15

4
in 2030 (%)

10
2

0
5

0
4

1000 2000 3000 4000

Yearly income of the bottom 20% in 2030 in the 5000 10,000 15,000 20,000
baseline scenarios (US$, 2005 PPP) GNI per capita in 2015 (US$, 2011 PPP)

Figure 3 | Income loss caused by climate change for the poorest 20%, as Figure 4 | National private household income losses (%) due to climate
a function of income of the bottom 20% in the baseline in 1,200 baseline change in 2030, in 1,200 baseline scenarios in 92 countries. PPP,
scenarios in 92 countries. PPP, purchasing power parity. purchasing power parity.

existing assessments which are based on economy-wide data and
focused on macroeconomic impacts. Figure4 provides the relation-
ship between current (2015) GNI per capita (on the x-axis) and pri-
vate household income loss (as a percentage of private household
income). It shows that countries that are poor today are likely to lose
more from climate change in terms of private household income in
the futureas already flagged in the literature4,6,8,45.
Estimated aggregate losses tend to be higher than most existing
estimates of aggregated economic impactseven though it is diffi-
cult to provide a quantified comparison because we do not have esti-
mates for developed countries. However, our results are consistent
with previous research in that they are highly heterogeneous across
countries, with higher aggregate impacts in poor countries. Impacts
are highly dependent on reference assumptions, with significantly
higher impacts when the population is poorer in the reference sce-
nario, with lower productivity, higher share of the budget dedicated
to food, less access to basic infrastructure (which affects exposure
to diarrhoea) and more people working in low-productivity agri-
culture. There are positive impacts on some countries and some
scenarios, largely linked to reduced exposure to disease (especially
where malaria is expected to become less prevalent 46) and impacts
on agricultural revenues where they are positive.
These results are based on a very different approach compared
to previous estimates, with different strengths and weaknesses to
other methodologies. The main strength of our approach is a bet-
ter accounting of the situation and vulnerability of households, and
the use of household survey data (instead of national accounts) that
may better represent household sources of income, especially for
the poorest households who represent a very small share of aggre-
gate economic output and consumption. The main weakness is the
absence of an internally consistent macro-economic framework
with interactions across sectors (for instance, if agricultural output
is significantly reduced, it has no impact on the non-agricultural sec-
tor or ability to invest in other activities). Our bottom-up approach
also misses some of the role of investment and international trade,
as it is only taken into account in food prices scenarios, and there
are concerns on the quality of household survey data, especially for
the richest households who especially in poor countries sometimes
represent a large share of the economy.
These weaknesses mean that the bottom-up approach is not
meant to replace a top-down approach based on macroeconomic
models (computable general equilibrium models or others). But it
provides an important complementary view, based on independent
data sources and methodology, focussing on the impact at the indi-
vidual and household scale. This view is made particularly impor-
tant by the fact that climate change is expected to affect the poor
disproportionally and that the impact on poverty may be more wor-
risome than the impact on aggregate GDP (a point also raised in the
IPCC fifth assessment report1).
Here, we find that impacts on the poor are likely to be larger than
the average impact, even within countries. Figure5 plots the income
losses at the aggregate level (x-axis) and for the bottom 40%, that is,
the 40% of households with the lowest income in the country (this
metric builds on the shared prosperity metric used by the World
Bank47). The figure mixes results for all developing countries in our
database, and for all baseline scenarios. It is, however, incomplete,
as we are missing results for developed countries. Acknowledging
this limitation, we find that the impacts on the bottom 40% are 70%
larger than those on the average population. As stressed elsewhere,
the fact that climate change worsens existing inequalities creates
an additional rationale to reduce emissions and stabilize climate
change at a low level48.
Implications for future research
Our exploration of micro-level vulnerability remains very simple
and opens avenues for future research along multiple dimensions.
First, our analysis should be completed by an assessment of the
impacts that climate change mitigation actions can have on the poor
through energy and food prices and other impacts on livelihoods. In
the short to medium term, this impact is likely to be more detrimen-
tal to the poor than the impacts of climate change if not corrected by
redistribution policies and social protection.
Second, as we already mentioned, a better and more exhaustive
representation of the channels through which climate change affects

NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange 253

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
PERSPECTIVE NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3253

20 under-nutrition can be caused by diarrhoeal disease54; climate-

related disease can reduce labour productivity, and thus income
and ability to save and invest in education. Non-climate shocks
need to be accounted for too: poor housing conditions worsen
15
vulnerability to all disease and natural disasters50,55; risk of evic-
tion due to weak land tenure leads to under-investment in housing
quality, and therefore higher vulnerability to natural disasters56.
Income loss of the bottom 40% (%)

10 Therefore, looking only at climate change impacts without taking

5
0 Studys Integrated Survey on Agriculture (LSMS-ISA) in 8 African
y = 1.7x
5
10
10 5 0 5 10
Average income loss in the population (%)
Figure 5 | Income losses for the poorest 40% in each country and each
scenario, plotted against the aggregated income losses in 1,200 baseline
scenarios in 92 countries.
peoples income is needed. The effects of disease could be repre-
sented in a more sophisticated way, and additional channels such
as access to energy and ecosystem services should be introduced.
Beyond the climate change channels, two more general research
questions remain, one linked to the methodology and the other to
the data.
In methodological terms, we need better representation and
modelling of how households accumulate assetsincluding social
and human capitaland escape or fall into poverty. The dynam-
ics of asset accumulation, including intergenerational dynamics
through health and education, should be represented explicitly,
in the spirit of Carter and Jenzen48. It is important to model how
households can save and accumulate assets, from their invest-
ment in their dwelling to financial savings and education of chil-
dren26,49,50. It is then necessary to account for what can interrupt
or reverse this accumulation of asset, from forced eviction from
their homes to natural disasters and disease. Agent-based models
could be an option to go further in that direction, and to repre-
sent the effect of shocks that affect households physical or human
capital, and their short-term and long-term prospects. They could
model the effect of health insurance, disaster risk insurance, or
social safety nets, and help assess the benefits of such policies. In
particular, such modelling would help measure the benefits of risk
management tools such as insurance and social safety nets, going
beyond avoided losses and considering how these instruments can
also help households shift toward higher-risk but higher-produc-
tivity occupations and technologies and invest in strategies that
can bring them out of poverty 5153.
Agent-based models would also allow for better accounting of
the interactions across various impacts of climate change, and for
the interactions between climate change impacts and other shocks.
In our simulations, each channel through which climate change
affects income is independent. In reality, however, there are good
reasons to assume that the effects of the five channels together
are worse than the sum of each channel: for example, the health
literature shows that under-nutrition makes people more likely
to suffer from other diseases such as malaria. It also shows that
into account interactions with the other factors that keep people in
povertyincluding voice and governance issuescould lead to a
large underestimation of climate change vulnerability 57.
In terms of data, needs are also large. First, recent surveys have
started to collect some of the data required to answer questions on
disaster and climate risk, such as the Living Standard Measurement
countries. But not all aspects are included in existing surveys and
there is still a real need to ask more and better questions on vul-
nerabilities to climate change and how to design policies that build
resilience and benefit poor people. Furthermore, panel household
surveys help better understand the dynamics of poverty reduction
and the role of shocks. There are already such panels, for instance
14 countries have LSMS panel surveys (http://go.worldbank.
org/2Y5QMGJ5C0). But the long periods between surveys (often
4years) make it difficult to link the occurrence of a shock (such
as a flood or a disease) to a change in income or poverty status.
High-frequency household surveys and systematic post-disaster
surveyssuch as the one carried out by Mumba58,59are needed to
better assess the role of environmental shocks in development and
poverty reduction.
Second, data are needed to better distinguish between struc-
tural and chronic poverty (based on assets) and stochastic and
temporary poverty (based on income, which is more volatile).
Doing so would help us distinguish shocks that create a tempo-
rary decrease in incomebut without irreversible impactsfrom
the shocks that bring people into poverty for extended periods of
time, or have irreversible impacts on childrens development and
prospects. The latter shocks are those that need to be avoided at all
costs, for instance using social safety nets and insurance60. More
knowledge will be critical to better design safety nets and insur-
ance schemes, and support poverty reduction in a world where
temporary poverty may be more important than chronic poverty.
Third, there is a need for better data to understand what is hap-
pening within households, capturing gender and generational
issues. Post-disaster surveys show that children are the first vic-
tims of natural hazards and shocks. Following weather shocks in
sub-Saharan Africa, asset-poor households typically provide chil-
dren with lower-quality nutrition and delay the start of school for
children61. They are also less likely to take sick children for medi-
cal consultations62. These behaviours have short-term and long-
term impacts, especially for children under the age of twosuch
as stunted growth and a greater tendency to get sick63,64. Disasters
and climate impacts also affect the distribution of consumption
between men and women65,66, a factor with large implications that
is occulted by databases with only aggregated household con-
sumption and income.
Fourth, it is important to know more on the interplay between
poverty, exposure, and vulnerability to disasters and environmen-
tal shocks and stresses (including for instance land degradation,
air pollution, and water scarcity). Much of the existing data col-
lected on climate shocks focuses on rural areas, in spite of large
vulnerability hot-spots in cities. Besides, in cities where the popu-
lation density is high, it is difficult to precisely measure the impact
of natural disasters (for example, annual flooding) on poor peo-
ple. An analysis of Mumbai, India, shows that the processes that
make poor people live in risky places tend to occur at the local
level and can only be analysed using high-resolution data59. But

254 NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3253
existing datasets are limited in terms of their spatial resolution and
some surveys even do not include the localization of households.
Poverty mapsfor instance those produced by the World Bank
Group67are extremely useful for climate-related studies, but they
are valid at the provincial or regional level only, and often cannot
support an analysis of risk exposure at the scale that is required for
floods and other heterogeneous disasters. Beyond technical and
cost issues, privacy concerns also matter, since they explain why
household location is often randomized within a 210km buffer,
making it more complicated to assess exposure to various hazards.
In conclusion, approaching climate change impacts from the
bottom-up, looking at households first and then aggregating at the
macro level, offers a new view on the climate change challenge.
By assessing and measuring the impacts of climate change with
poverty-related metrics, we can ensure that economic analysis bet-
ter captures one of the main reasons for concern regarding climate
change, namely the impact on the poorest and the most vulner-
able68. For instance, we can improve integrated assessment model
estimates of the social cost of carbon by providing region-specific
impacts and distribution of impacts within the region. We can
also support the design and assessment of new tools and policies,
such as social safety nets or health insurance systems, which can
be considered as climate adaptation policies. To take the agenda
forward, we call upon the scientific communities working on cli-
mate change and on poverty to intensify the emerging collabora-
tion2,18,25,51,6971. These interdisciplinary research programmes and
data collection efforts need to explore various pathways out of
poverty and their characteristics, including the role of household
resilience and environmental capital; the obstacles households face
to escape poverty and the shocks that bring them back, including
environmental and land degradation, water scarcity, environment-
related diseases, and natural hazards; and the policies that can
protect poor households from these shocks and accelerate capital
accumulation, progress in skills, productivity, wealth, and well-
being. The results of this research would be critical to better sup-
port policy-makers and help them assess the threat from climate
change and design the right adaptation and resilience policies.
Received 14 April 2016; accepted 16 February 2017;
published online 5 April 2017
References
1. IPCC. Summary for Policymakers. In Climate Change 2014: Impacts, Adaptation,
and Vulnerability (eds Field, C. B. et al.) (Cambridge Univ. Press, 2014).
2. Olsson, L. etal. in Climate Change 2014: Impacts, Adaptation, and Vulnerability
(eds Field, C.B. etal.) 793832 (Cambridge Univ. Press, 2014).
3. Arent, D.J. etal. in Climate Change 2014: Impacts, Adaptation, and
Vulnerability (eds Field, C.B. etal.) 659708 (Cambridge Univ. Press, 2014).
4. Tol, R.S. Estimates of the damage costs of climate change. Part 1: benchmark
estimates. Environ. Resour. Econ. 21, 4773 (2002).
5. Tol, R.S. The economic effects of climate change. J.Econ. Perspect.
23, 2951 (2009).
6. Nordhaus, W. Estimates of the social cost of carbon: concepts and
results from the DICE-2013R model and alternative approaches.
J.Assoc. Environ. Resour. Econ. 1, 273312 (2014).
7. Mendelsohn, R., Morrison, W., Schlesinger, M.E. & Andronova, N.G.
Country-specific market impacts of climate change. Climatic Change
45, 553569 (2000).
8. Hope, C. The marginal impact of CO2 from PAGE2002: an integrated
assessment model incorporating the IPCCs five reasons for concern.
Integr.Assess. 6, 1956 (2006).
9. Stern, N.The Stern Review on the Economics of Climate Change (Cambridge
Univ. Press,2006).
10. Yohe, G. & Tol, R.S. J. Indicators for social and economic coping capacity
moving toward a working definition of adaptive capacity. Glob. Environ. Change
12, 2540 (2002).
11. Mendelsohn, R., Dinar, A. & Williams, L. The distributional impact of climate
change on rich and poor countries. Environ. Dev. Econ. 11, 159178 (2006).
12. Sterner, T. & Persson, U.M. An even sterner review: introducing relative prices
into the discounting debate. Rev. Environ. Econ. Policy 2, 6176 (2008).
PERSPECTIVE
13. Skoufias, E., Rabassa, M., Olivieri, S. & Brahmbhatt, M. The Poverty Impacts of
Climate Change (World Bank, 2011).
14. Dollar, D. & Kraay, A. Growth is good for the poor. J.Econ. Growth
7, 195225 (2002).
15. Dollar, D., Kleineberg, T. & Kraay, A. Growth Still is Good for the Poor. Policy
Research Working Paper 6568 (World Bank, 2013).
16. Noack, F., Wunder, S., Angelsen, A. & Boerner, Jan. Responses to Weather and
Climate: A Cross-Section Analysis of Rural Incomes Policy Research Working
Paper 7478 (World Bank, 2015).
17. Hallegatte, S. etal. Shock Waves: Managing the Impacts of Climate Change on
Poverty (World Bank, 2015).
18. Hertel, T.W., Burke, M.B. & Lobell, D.B. The poverty implications of climate-
induced crop yield changes by 2030. Glob. Environ. Change 20, 577585 (2010).
19. Ivanic, M. & Martin, W. Implications of Higher Global Food Prices for Poverty in
Low-Income Countries. Policy Research Working Paper 4594 (World Bank, 2008).
20. Winsemius, H.C. etal. Disaster Risk, Climate Change, and Poverty: Assessing
the Global Exposure of Poor People to Floods and Droughts. Policy Research
Working Paper 3225 (World Bank, 2015).
21. Park, J., Hallegatte, S., Bangalore, M. & Sandhoefner, E. Households and Heat
Stress: Estimating the Distributional Consequences of Climate Change. Policy
Research Working Paper 7479 (World Bank, 2015).
22. Heltberg, R., Oviedo, A.M. & Talukdar, F. What are the Sources of Risk
and How do People Cope? Insights from Households Surveys in 16 Countries
(World Bank, 2014).
23. Sen, B. Drivers of escape and descent: changing household fortunes in rural
Bangladesh. World Dev. 31, 513534 (2003).
24. Krishna, A. Pathways out of and into poverty in 36 villages of Andhra Pradesh,
India. World Dev. 34, 271288 (2006).
25. Krishna, A. in Reducing Global Poverty: The Case for Asset Accumulation
(ed.Moser, C.O. N.) 6279 (Brooking Institution Press, 2007).
26. Moser, C.O. N. (Ed.) Reducing Global Poverty: The Case for Asset
Accumulation (Brookings Institution Press, 2007).
27. Carter, M.R. & Barrett, C.B. The economics of poverty traps and persistent
poverty: an asset-based approach. J.Dev. Stud. 42, 178199 (2006).
28. Dercon, S. Growth and shocks: evidence from rural Ethiopia. J.Dev. Econ.
74, 309329 (2004).
29. Savard, L. Poverty and inequality analysis within a CGE framework:
acomparative analysis of the representative agent and microsimulation
approaches. Dev. Policy Rev. 23, 313331 (2005).
30. Bourguignon, F., Ferreira, F.H. & Lustig, N. The Microeconomics of Income
Distribution Dynamics in East Asia and Latin America (World Bank
Publications, 2005).
31. Bussolo, M., De Hoyos, R. & Medvedev, D. Economic growth and income
distribution: linking macroeconomic models with household survey data at
the global level. Int. J. Microsimulation 3, 92103 (2010).
32. Olivieri, S. etal. Simulating Distributional Impacts of Macro-dynamics: Theory
and Practical Applications (World Bank, 2014).
33. KC, S. & Lutz, W. Demographic scenarios by age, sex and education
corresponding to the SSP narratives. Popul. Environ. 35, 243260 (2014).
34. Rozenberg, J. & Hallegatte, S. The Impacts of Climate Change on Poverty
in 2030and the Potential from Rapid, Inclusive, and Climate-Informed
Development. Policy Research Working Paper 7483 (World Bank, 2015).
35. ONeill, B.C. etal. A new scenario framework for climate change
research: theconcept of shared socioeconomic pathways. Climatic Change
122, 387400 (2013).
36. ONeill, B.C. etal. The roads ahead: narratives for shared socioeconomic
pathways describing world futures in the 21st century. Glob. Environ. Change
42, 169180 (2015).
37. Havlk, P. etal. Climate Change Impacts and Mitigation in the Developing World:
an Integrated Assessment of Agriculture and Forestry Sectors (IIASA, 2015).
38. Bouwer, L.M. Projections of future extreme weather losses under changes in
climate and exposure. Risk Anal. 33, 915930 (2013).
39. Hallegatte, S., Vogt-Schilb, A., Bangalore, M. & Rozenberg, J. Unbreakable:
Building the Resilience of the Poor in the Face of Natural Disasters (World
Bank,2016).
40. Deryugina, T. & Hsiang, S.M. Does the Environment Still Matter? Daily
Temperature and Income in the United States (National Bureau of Economic
Research, 2014).
41. Heal, G. & Park, J. Feeling the Heat: Temperature, Physiology and the Wealth
of Nations NBER Working Paper No. 19725 (National Bureau of Economic
Research, 2013).
42. Hales, S., Kovats, S., Lloyd, S. & Campbell-Lendrum, D. (Eds) Quantitative
Risk Assessment of the Effects of Climate Change on Selected Causes of Death,
2030s and 2050s (World Health Organization, 2014).
43. Kolstad, E.W. & Johansson, K.A. Uncertainties associated with quantifying
climate change impacts on human health: a case study for diarrhea. Environ.
Health Perspect. 119, 299305 (2010).

NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange 255

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
PERSPECTIVE
44. Caminade, C. etal. Impact of climate change on global malaria distribution.
Proc. Natl Acad. Sci. USA 111, 32863291 (2014).
45. Hahn, R.W. & Ulph, A. Climate Change and Common Sense: Essays in Honour
of Tom Schelling (Oxford Univ. Press, 2012).
46. Caminade, C. & Jones, A.E. Epidemiology: Malaria in a warmer West Africa.
Nat. Clim. Change 6, 984985 (2016).
47. Global Monitoring Report 2014/2015: Ending Poverty and Sharing Prosperity
(World Bank, 2015).
48. Dennig, F., Budolfson, M.B., Fleurbaey, M., Siebert, A. & Socolow,
R.H. Inequality, climate impacts on the future poor, and carbon prices.
Proc. Natl Acad. Sci. USA 112, 1582715832 (2015).
49. Moser, C. & Felton, A. In Reducing Global Poverty: The Case for Asset
Accumulation (Ed. Moser, C.O. N.) 1550 (Brookings Institution Press, 2008).
50. Fay, M. The Urban Poor in Latin America (World Bank, 2005).
51. Bandyopadhyay, S. & Skoufias, E. Rainfall variability, occupational choice, and
welfare in rural Bangladesh. Rev. Econ. Household 13, 589634 (2013).
52. Elbers, C., Gunning, J.W. & Kinsey, B. Growth and risk: methodology and micro
evidence. World Bank Econ. Rev. http://dx.doi.org/10.1093/wber/lhl008 (2007).
53. Dercon, S. & Christiaensen, L. Consumption risk, technology adoption and
poverty traps: evidence from Ethiopia. J.Dev. Econ. 96, 159173 (2011).
54. Checkley, W. etal. Multi-country analysis of the effects of diarrhoea on
childhood stunting. Int. J.Epidemiol. 37, 816830 (2008).
55. Lall, S.V. & Deichmann, U. Density and disasters: economics of urban hazard
risk. World Bank Res. Obs. 27, 74105 (2012).
56. Field, E. Entitled to work: urban property rights and labor supply in Peru.
Q.J.Econ. 122, 15611602 (2007).
57. Tschakert, P. The Role of Inequality in Climate-Poverty Debates. Policy Research
Working Paper 7677 (World Bank, 2016).
58. Patankar, A. & Patwardhan, A. Estimating the uninsured losses due to extreme
weather events and implications for informal sector vulnerability: a case study
of Mumbai, India. Natural Hazards 80, 285310 (2014).
59. Patankar, A. The Exposure, Vulnerability and Adaptive Capacity of Households
to Floods in Mumbai. Policy Research Working Paper 7481 (World Bank, 2015).
60. Carter, M.R. & Janzen, S.A. Social Protection in the Face of Climate Change:
Targeting Principles and Financing Mechanisms. Policy Research Working Paper
7476 (World Bank, 2015).
61. Alderman, H., Hoddinott, J. & Kinsey, B. Long term consequences of early
childhood malnutrition. Oxford Econ. Pap. 58, 450474 (2006).
62. Jensen, R. Agricultural volatility and investments in children. Am. Econ. Rev.
90, 399404 (2000).
63. Yamano, T., Alderman, H. & Christiaensen, L. Child growth, shocks, and food
aid in rural Ethiopia. Am. J.Agr. Econ. 87, 273288 (2005).
64. Dercon, S. & Porter, C. Live Aid revisited: long-term impacts of the 1984
Ethiopian famine on children. J.Eur. Econ. Assoc. 12, 927948 (2014).
65. Del Ninno, C., Dorosh, P.A., Smith, L.C. & Roy, D.K. The 1998 Floods in
Bangladesh: Disaster Impacts, Household Coping Strategies, and Responses
(International Food Policy Research Institute, 2001).
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3253
66. Ajibade, I., McBean, G. & Bezner-Kerr, R. Urban flooding in Lagos, Nigeria:
patterns of vulnerability and resilience among women. Glob. Environ. Change
23, 17141725 (2013).
67. Hentschel, J., Lanjouw, J.O., Lanjouw, P. & Poggi, J. Combining census and
survey data to trace the spatial dimensions of poverty: a case study of Ecuador.
World Bank Econ. Rev. 14, 147165 (2000).
68. Smith, J.B. etal. Assessing dangerous climate change through an update of the
Intergovernmental Panel on Climate Change (IPCC) reasons for concern.
Proc. Natl Acad. Sci. USA 106, 41334137 (2009).
69. Del Ninno, C., Dorosh, P.A. & Smith, L.C. Public policy, markets and
household coping strategies in Bangladesh: avoiding a food security crisis
following the 1998 floods. World Dev. 31, 12211238 (2003).
70. Ahmed, S.A., Diffenbaugh, N.S. & Hertel, T.W. Climate volatility
deepens poverty vulnerability in developing countries. Environ. Res. Lett.
4, 034004 (2009).
71. OBrien, K.L. & Leichenko, R.M. Double exposure: assessing the
impacts of climate change within the context of economic globalization.
Glob. Environ. Change 10, 221232 (2000).
Acknowledgements
The authors acknowledge the contributions and feedback from the core author group
of the Shock Waves report, including Mook Bangalore, Laura Bonzanigo, Ulf Narloch,
TamaroKane, David Treguer, and Adrien Vogt-Schilb. The report was produced with
guidance and under the supervision of Marianne Fay and JohnRoome. Guidance
was provided by the reports peer reviewers: Carter Brandon, RichardDamania,
FranciscoH.G.Ferreira, Masami Kojima, Andrea Liverani, TamerRabie,
EmmanuelSkoufias, and Mike Toman. Purnamita Dasgupta, Stefan Dercon, KristieEbi,
Chris Field, Michael Oppenheimer, and Youba Sokona also contributed as external
advisors to the author team.
Author contributions
SH and JR developed the methodological approach and the model, and wrote the
manuscript. JR implemented and ran the model, performed the statistical analyses of the
results, and generated the figures.
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints. Publishers
note: Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional affiliations. Correspondence and requests for materials should be
addressed to J.R.
Competing financial interests
The authors declare no competing financial interests.

256 NATURE CLIMATE CHANGE | VOL 7 | APRIL 2017 | www.nature.com/natureclimatechange

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.