You are on page 1of 22

This article was downloaded by: [carlos lopez] On: 30 November 2011, At: 17:11 Publisher: Routledge Informa

Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Economic Systems Research

Publication details, including instructions for authors and subscription information:

Carlos Lpez-Morales & Faye Duchin
a a a

Rensselaer Polytechnic Institute, Department of Economics, Troy, NY, USA Available online: 30 Nov 2011

To cite this article: Carlos Lpez-Morales & Faye Duchin (2011): POLICIES AND TECHNOLOGIES FOR A SUSTAINABLE USE OF WATER IN MEXICO: A SCENARIO ANALYSIS, Economic Systems Research, 23:4, 387-407 To link to this article:

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Economic Systems Research, 2011, Vol. 23(4), December, pp. 387 407

Rensselaer Polytechnic Institute, Department of Economics, Troy, NY, USA
(Received 6 July 2011; In nal form 27 September 2011)

Downloaded by [carlos lopez] at 17:11 30 November 2011

Water stress in Mexico is intimately linked to agriculture, as irrigation claims 75% of national water withdrawals. The Mexican mix of irrigation technologies is dominated by ood techniques, utilized on 93% of irrigated land, while drip and sprinkler systems, both with higher application efciencies, are utilized on only 7% of irrigated land. This paper examines the extent to which government policies can induce the adoption of alternative irrigation technologies to promote a sustainable pattern of water withdrawals. The framework is an inter-regional inputoutput model formulated as a linear program that solves for cost-minimizing allocations of output that are constrained by regional factor availability. The model features endogenous choice among alternative agricultural technologies and determines commodity prices based on factor costs and on scarcity rents for limiting factors of production. The study denes and quanties sustainable endowments of water at the regional level and analyzes scenarios that combine fees or caps on water withdrawals with the availability of alternative irrigation technologies. We nd that water policies can induce technology adoption to achieve water sustainability, although the national price of agricultural output rises 5% to 8% relative to baseline levels. Furthermore, pricing water for irrigation can generate enough public revenue for the government to cover the full costs of technology adoption. Keywords: Irrigation technologies; programming; Mexico Water sustainability; Interregional input output model; Linear



The diverse development challenges that Mexico faces will grow in complexity in the twenty-rst century due to sharpened environmental problems. Trends of increased population, afuence, and climate variability will only intensify pressure on the countrys water resources. The water stress indicator, or the ratio of yearly withdrawals to average annual runoff, stands at 0.17, just below the 0.20 threshold of moderate stress when computed for the country as a single spatial unit (CNA, 2010; Falkenmark and Lindh, 1976). Because of a locational mismatch between economic activity and water availability, however, water scarcity manifests itself as a regional phenomenon. In fact, the national average of stress indicators for 13 Mexican regions is 0.47, greater than the high stress threshold. Activities responsible for 80% of Mexicos gross domestic product (GDP) are concentrated in the eight regions in the Center and in the North that exhibit stress indicators higher than the 0.40 scarcity threshold. These regions are home to 75% of the countrys population. Irrigated agriculture in Mexico claims the largest share of national withdrawals, with 71% of the total, while industrial sectors account for 20% and households for the rest

Corresponding author. E-mail:

ISSN 0953-5314 print; ISSN 1469-5758 online # 2011 The International InputOutput Association



(CNA, 2010). Mexicos irrigation infrastructure is distributed differently from regional water availability: 37% of irrigated land is located in water-scarce regions in the Center and in the North; water-abundant regions in the South, by contrast, contain only 3% of the countrys irrigated land. Flood by gravity, the most frequent irrigation technology in Mexico, is utilized on 93% of active irrigated land, while sprinkler or drip systems are present on only the remaining 7% (FAO-AQUASTAT, 2009). These technologies have application efciencies of 60%, 75% and 95%, respectively (Brouwer et al., 1989, Postel, 1999). It follows that a different mix could reduce signicantly the volume of water withdrawn for agricultural purposes, reducing stress in water-scarce regions. This paper examines the potential for water policies to change the regional distribution of agriculture and the mix of irrigation technologies to achieve a sustainable pattern of water withdrawals in Mexico. Contemplated policies include region-specic fees for irrigation water, which currently receives a full subsidy, and region-specic withdrawal restrictions via the issuance of a limited number of permits. To evaluate the impact of these measures on withdrawal patterns and economic costs, we implement an interregional, multifactor, multisector model of the Mexican economy that is based on the World Trade Model (Duchin, 2005), but applied to regions within a single country representing hydro-economic units that reconcile their economic and hydrologic characteristics. The model distinguishes irrigated and non-irrigated agriculture, dependent on blue and green water, respectively, and denes and quanties region-specic sustainable endowments of blue water to assess the environmental sustainability of patterns of blue water withdrawals. Regional endowments of green water are not quantied due to lack of data. However, by making explicit the dependence of non-irrigated agriculture on nonirrigated land, whose location is ultimately correlated with the location of green water endowments, we claim to capture the main geographical features of the constraints imposed by green water availability. The model is used to analyze two policy scenarios that offer irrigation options, and we compare the results to those of a baseline in the absence of water policies. A fourth scenario enforces water sustainability but in the absence of irrigation alternatives. We compare costs of adoption of water-saving technology, using a wide range of assumptions about upfront investment costs, time-horizons, and discount rates, with revenue generated by water pricing policies. The rest of the paper is organized as follows. Section 2 describes the Mexican water situation, and Section 3 reviews the relevant literature. The sustainable endowment of blue water is described in Section 4, along with the main characteristics of the model, including the representation of alternative irrigation technologies (i.e., a mix of sprinkler and drip irrigation systems), and the four scenarios. The results of the scenario analysis are presented in Section 5, and the nal section concludes.

Downloaded by [carlos lopez] at 17:11 30 November 2011

2. 2.1.

WATER IN MEXICO Water Withdrawals and Availability in Mexico

Mexicos Water Commission divides the national territory into 13 hydro-economic regions on the basis of economic and hydrologic criteria (see Figure 1). The present study further classies them in three groups of high, medium, and low average yearly


FIGURE 1. Hydro-economic regions of Mexico.


Downloaded by [carlos lopez] at 17:11 30 November 2011

Source: CNA (2010).

runoff for exposition purposes. Table 1 and Figure 1 distinguish the low-availability group, with 9% of national runoff, located in the North and the Center-North (regions I, II, VI, VII, and XIII); the medium-availability group, with 38% of runoff, in the Yucatan Peninsula and the Center (regions III, IV, V, VIII, IX, and XII); and the highavailability group, with 56% of total runoff, in the southern regions of the country (regions X and XI). Table 1 shows that agriculture is the largest user of water in all but regions XIII and X, the Mexico Valley and the Central Gulf, respectively. Domestic uses are the largest in region XIII, while industrial and thermoelectric uses are important in region X. Overall, food and energy exhibit high dependence on water, with thermoelectric plants representing 76% of the national generation capacity (SENER, 2010), and irrigated agriculture, which occupies only one quarter of active agricultural land, accounting for half the total value of agricultural output.


Water and Agricultural Land: A Critical Link

Regions with low and medium availability, where less than half of national runoff is generated, contain 97% of agricultural land in Mexico that is equipped for irrigation (37% and 60%, respectively; see Table 1). By contrast, the high-availability regions, with 55% of national runoff, contain only 3% of irrigated land. Overall, one can distinguish three classes of regions, where:



Mexican states, regions, availability groups, runoff, withdrawals, environmental water requirements, and agricultural land.
Runoff Withdrawals Environmental water requirements Agricultural land (103 ha) % 4

Downloaded by [carlos lopez] at 17:11 30 November 2011

Availability Group Low

States Baja California Baja California Norte Sonora Coahuila, Chihuahua Nuevo Leon Durango, Zacatecas San Luis Potos Estado de Mexico Distrito Federal Sinaloa Guerrero, Michoacan Puebla, Tlaxcala, Morelos Aguascalientes, Colima Guanajuato, Jalisco, Nayarit Hidalgo, Queretaro Tamaulipas Oaxaca Campeche, Yucatan Quintana Roo Veracruz Chiapas, Tabasco National

Name Baja Peninsula Northwest Bravo Basin Central-North Mexico-Valley North Pacic Balsas Basin Lerma-Santiago North-Gulf South Pacic Yucatan Peninsula Center-Gulf South Border

Region km3/yr % km3/yr % Agr. Km3/yr % of runoff Non-irrigated % Irrigated I II VI VII XIII III IV VIII IX V XII X XI 5 8 12 8 3 27 22 34 25 33 30 95 158 458 1 2 3 2 1 6 5 7 6 7 7 21 35 100 4 6 9 4 5 10 10 13 4 1 2 5 2 75 81 85 82 87 47 91 57 80 77 78 62 47 73 71 1 3 4 2 1 8 7 15 9 13 13 38 63 176 20 30 30 30 20 30 30 45 35 40 45 40 40 38 48 42 1083 2306 778 552 2473 2370 1754 1,016 1087 1,346 1,868 16,723 ,1 ,1 6 14 5 3 15 14 10 6 7 8 11 100 224 498 598 410 162 721 684 968 466 84 62 92 62 5,032


10 12 8 3 14 14 19 9 2 1 2 1 100



Sources: Runoff and withdrawals, CNA (2010). Environmental water requirements, own computations based on Smakhtin and Doll (2004). Land, SAGARPA (2009).

SUSTAINABLE USE OF WATER IN MEXICO (1) non-irrigated land is minimal (I and II, in the North); (2) both types of land are in use (mainly in the Center); (3) irrigated land is minimal (in the South).


A pattern of substitution and complementarity between rain-fed and irrigated production can be seen at this level of geographic aggregation. In the northern regions, crop requirements are fullled almost entirely by irrigation water; the Center utilizes a mix of both sources of water, with irrigation water smoothing rain variability and boosting land productivity; and the South, where rain is abundant, is not dependent on irrigation water. A decrease of agricultural output in the North, due to constraints on irrigation, may require an expansion in the use of irrigated or rain-fed land in other regions. Given the dominance of irrigated agriculture in water withdrawals, even small efciency increases could possibly translate into water savings of considerable magnitude.
Downloaded by [carlos lopez] at 17:11 30 November 2011


Water Policy

According to Mexican Federal Law, water resources are owned by the State. The government issues a limited amount of withdrawal permits, charging a fee per cubic meter withdrawn that depends on the geographic zone and the economic sector (Diario Ocial de la Federacion, 2009). National average fees among geographic zones are shown in Table 2. Positive fees are charged for industrial and domestic uses, but withdrawals for irrigation within permitted amounts have a zero price tag. The Mexican government plans to modify the pricing scheme, especially for irrigation water (CNA, 2008). Another policy option is to improve the monitoring and enforcement of the quantities withdrawn under the system of permits. The government is about to complete decentralization of the once heavily centralized monitoring system so as to better quantify volumes withdrawn and prevent clandestine withdrawals of water.



Relatively few studies in the economic literature on water have addressed the impact of water policy or of improvements in irrigation technologies on the environmental sustainability of withdrawal patterns. In this section, we describe rst the policy or technology experiments that have been carried out in water studies and then the analytical methods utilized in more general economic studies about water. 3.1. On Water Policy and Irrigation Technologies

Computable general equilibrium (CGE) models are used to analyze the global impact of water prices (Berrittella et al., 2005) and of withdrawal restrictions (Berrittella et al., 2007) in several regions of the world. These models assume prot maximizing rms and perfectly competitive markets, represent international trade using the Armington assumption (i.e., treat the same product produced in different countries as distinct products), and assess alternative scenarios using a monetary measure of welfare based on utility functions that are assumed to represent global consumer preferences. The rst



TABLE 2. Water exploitation fees by economic use in 2009. Use General Domestic (1) Domestic (2) Agriculture (1) Agriculture (2) Recreation Power Generation Aquaculture Pesos/m3 8.773 0.546 0.272 0.000 0.130 0.008 0.004 0.002

Downloaded by [carlos lopez] at 17:11 30 November 2011

Source: CNA (2010). Note: Fees are national averages among municipalities. General uses include industrial and other uses not included in the remaining categories. Domestic (1): 300 liters per capita per day. Domestic (2): .300 liters per capita per day. Agriculture (1): Use is smaller than permitted volume. Agriculture (2): Penalty per m3 exceeding the permit.

study performs several experiments in which water prices are specied exogenously for all sectors and all regions, or for water scarce regions, or for only the irrigated agriculture sector. The authors conclude that regional water prices have substantial impacts on the patterns of agricultural production, on international trade, and on measures of economic welfare. The second study imposes withdrawal restrictions on overdrawn aquifers and shows how international patterns of trade are modied by reduced water supply in selected regions. Velazquez et al. (2007) also use a CGE model, in this case to study the effect of priced irrigation water in Spains Andalusia region. They nd that agricultural water demand responds only very slightly to higher water prices; however, a sectoral reallocation of water results that does reduce water use. Calzadilla et al. (2008) develop a CGE model to study the global impacts of improved efciency in irrigation systems. In their experiment, exogenous improvements in regional irrigation efciencies are assumed, but no detail is provided as to the specic technologies. They nd that welfare benets are concentrated in water-scarce regions, while in water-abundant regions changes in welfare are mostly negative. The global effect on welfare, however, is positive. All the CGE studies reviewed above impose constraints on availability of some factors (capital, land, labor and, in some cases, water), but no consistent conclusions about water savings are examined by the authors. Llop (2008) combines water-pricing policies for Spain with exogenous efciency improvements in intermediate demand for water and studies the response of commodity prices using an input output (IO) price model. Her experiments assume that the water supply can be expanded as required by the policy interventions. Similar to CGE modelers, Llop is interested in changes in a money value for consumer welfare, which in her model is measured by changes in expenditures resulting from modied prices for a given nal demand. Unlike CGE models, however, Llops model does not involve an optimization process. In her results, water prices lead to substantial reductions in water use but increases in commodity prices and therefore reductions in welfare. Across-the-board exogenous technological improvement in water use, by contrast, exhibits the so-called Jevons Paradox: reductions in unit water requirements of all sectors lower costs but lead to



higher aggregate water consumption, although the latter seems to be driven by the assumed exogenous increase in deliveries from the water sector. The combination of pricing policies and improved technology is able to achieve water savings without the inationary effect on commodity prices.


Input Output Studies of Water

Several studies exploit the well-known ability of the quantity IO model to provide estimates for total requirements (direct and indirect) of primary factors of production, in this case water, at the sectoral level; these include Duarte et al. (2002) for Spain, Velazquez (2006, 2007) for Andalusia, Dietzenbacher and Velazquez (2007) also for Andalusia, and Lenzen and Foran (2001) for Australia. The main objective of this literature is to decompose the past demand for water to identify the sectors that were the most waterintensive both directly and indirectly. Guan and Hubacek (2008) integrate an IO model with a hydrological model to dene extended water demand and quantify it for the single region of northern China. Extended water demand adds to the total demand for water the volume of water that the hydrological system would require to dilute water pollution. Adding information about runoff from an external hydrological model, Hubacek and Sun (2005) analyze scenarios to examine how changes in Chinese lifestyles could affect the relation between Chinas water use and water supply. Unlike the studies discussed above, IO-based linear programming models offer production options, specify a choice criterion for selecting among them, and treat factor endowments as explicit constraints on economic activity. In two early contributions, Henry and Bowen (1981) and Harris and Rea (1984) apply similar linear programming models to identify when water endowments limit economic activities for two regional economies in the US. Both models maximize production subject to availability of labor and water, and the authors analyze scenarios that reduce water endowments until economic activity becomes physically infeasible. The binding water constraints generate positive shadow prices for water. These characteristics of linear programming are also exploited by Liu et al. (2009) to estimate region-specic shadow prices for water in nine major Chinese basins. The studies reported above operate at different spatial scales, from global multinational regions to sub-national provinces. The most frequent geographical units are based on political divisions: countries (treated separately or aggregated into multinational regions) or sub-national states and provinces (such as Andalusia in Spain, or Chinas northern region). A problem that arises with administrative units has been called the aggregation fallacy (Rosegrant et al. 2002), the assumption, whether explicit or not, of equal access to water from anywhere within the given region, when in fact component subregions may have very different endowments. One solution is to dene hydro-economic regions that are relatively homogeneous in both economic and hydrological attributes. Some examples of this strategy are Rosegrant et al. (2002) and Guan and Hubacek (2008). We follow this practice by dening 13 hydro-economic regions, making it possible to describe Mexicos geographical distribution of water and economic activity at a level of detail we nd acceptable and preferable to working at the national level or with very large sub-national regions. Nonetheless, some of the hydro-economic regions would clearly benet from further spatial disaggregation.

Downloaded by [carlos lopez] at 17:11 30 November 2011

394 4. 4.1. METHODS


Choice of Geographical Units

Downloaded by [carlos lopez] at 17:11 30 November 2011

This study uses the 13 hydro-economic regions dened by the Mexican National Water Commission (see CNA, 2010 and Figure 1), which provides data on regional runoff and withdrawals for economic use. Information about land use, employment of labor, capital, and production were compiled at the State level from various sources (see Table 3) and aggregated to match the hydro-economic boundaries (shown in Table 1). IO tables for Mexican regions are available in the literature (e.g., Chapa-Cantu et al., 2009), but they do not cover the entire country nor do they follow boundaries compatible with our hydro-economic regions. For lack of region-level information, we assumed that all regions share the average intermediate input structure that appears in the national input output table (INEGI, 2011a). The availability of water and other factors of production, however, is provided at the State level of detail from various sources (see Table 3). Thus, both our factor use per unit of output and factor endowments are region-specic (see Section 4.3).


Regional Sustainable Endowments of Water

The renewable supply of water, or average yearly runoff, for Mexicos hydro-economic regions is shown in Table 1. Regional runoff captures many of the key features of water availability, but it systematically overestimates regional endowments of water. Both environmental water requirements and the capacity of existing withdrawal infrastructure also need to be taken into account. The latter is called the effective supply, ES; it is measured in km3/yr and could be either larger or smaller than runoff. The fraction of
TABLE 3. Data sources for parameters and variables (for region m). Description Parameters Am Fm Source

Input-output matrix INEGI (2011a) Matrix of factor inputs Land: SAGARPA (2009) Labor and Capital: INEGI (2011b) Water: CNA (2010) Vector of factor endowments National nal demand Vector of factor prices Land: SAGARPA (2009), SEMARNAT (2010) Labor and Capital: INEGI (2011b) Water: CNA (2010), this study. INEGI (2011b) Land: INEGI (2011c), Lee Harris (2002) Labor and Capital: INEGI (2011b), Blancas (2006), Lee Harris (2002) Water: CNA (2010) This study

Exogenous variables

Fm y pm

Endogenous variables

xm p rm

Output National price vector Regional scarcity rents

Source: Own elaboration



mean annual runoff that remains potentially available after environmental water requirements are subtracted we call the environmentally sustainable supply, ESS. Finally, the effective environmentally sustainable supply of blue water, EESS, is calculated as the lower of the two quantities (shown as Equation 1). These measures of water supply are described in more detail by Duchin and Lopez (2011). When a regions effective supply is larger than the environmentally sustainable supply, sustainable withdrawals should be restricted to the latter, even if a fraction of the capacity of the withdrawal infrastructure must be left idle. Otherwise, the sustainable endowment is limited by the effective supply of water. EESS = min{ES, ESS} (1)

Downloaded by [carlos lopez] at 17:11 30 November 2011

Environmental water requirements (EWR) for each of the 13 Mexican regions, as esti mated using data from Smakhtin and Doll (2004), are shown in Table 1. Nationally, environmental water requirements amount to 38% of the national renewable supply. However, regions of low availability of water have lower environmental requirements as they have poorer aquatic ecosystems, while regions of higher availability of water tend to have higher requirements. In the absence of regional data on the effective water supply, we estimate these quantities for this study (see Section 4.5).


The Inter-regional Model

For this study we adapt Duchins World Trade Model (WTM) (Duchin, 2005) for an interregional analysis within a single country. Unlike many IO model applications focused on water, the WTM is an optimization model that chooses among production options; unlike other optimization models, including CGE models and other IO-based LP models, the WTM minimizes factor costs for given nal demand rather than maximizing production, prots, or a consumption-based measure of welfare. Earlier extensions and applications of the WTM include Strmman and Duchin (2006), Julia and Duchin (2007), He and Duchin (2009), and Strmman et al. (2009). Similar to He and Duchin (2009), this study analyzes the regional division of labor within a national economy. Like the model of Julia and Duchin (2007), ours allows for the choice among alternative agricultural technologies based on region-specic technologies and, within each region, distinguishing two classes of land, irrigated and non-irrigated. The representation of the choice of alternative technologies is based on Duchin and Lange (1992); but see a more parsimonious treatment of the choice among alternative technologies in Duchin and Levine (2011). Our database includes 15 economic sectors (they are shown later in Table 9) and ve factors of production (labor, capital, non-irrigated land, irrigated land, and blue water) for each of the thirteen regions. Output produced using technology l in region i is described by a sector-specic and region-specic set of coefcients arranged in matrices Ali for intermediate inputs and Fli for factor requirements. Each region has three alternatives to generate agricultural output (i.e., l 1, 2, 3): non-irrigated agriculture, ood irrigated agriculture, and a mix of drip and sprinkler irrigation technologies (see Section 4.4). The optimization is constrained by two conditions: total national output must satisfy total nal and intermediate demand, and the use of factors in a region may not exceed regional endowments. The primal linear program species the quantity relationships



and the dual model relates costs and prices:

m L

Min Z =
i=1 l=1

pi Fli xli

subject to
m L m

I Ali xli =
i=1 l=1 L i=1


Fli xli f i ; i = 1, . . . , m
l=1 m m

Downloaded by [carlos lopez] at 17:11 30 November 2011

Max W = p subject to

i=1 i=1

ri f i

I Al p Fl ri Fl pi ; i = 1, . . . , m; l = 1, . . . , L i i i where the matrices A and F have dimensions (15 15) and (5 15) respectively, while the vectors x, y and p all have dimensions (15 1). The vectors p, f and r have dimensions (5 1). m For the optimal solution, Z W, assuring that total consumption demand, p yi , is equal
m L

to factor earnings across all regions and technologies,

i=1 l=1

p Fli xli + i

m i=1

ri f i . Table 3

shows the exogenous and endogenous variables, parameters, and data sources. While the model solves for inter-regional ows within Mexico, it should be pointed out that the international trade of Mexico is included as part of the nal demand vector and, like domestic nal demand, is assumed to be the same under all scenarios.


Representing the Mix of Alternative Irrigation Technologies

Water requirements per unit of agricultural output, based on application efciencies from Brouwer et al. (1989) and Postel (1999), are shown in Table 4. Given the lack of more detailed (or more recent) data, these are assumed to be representative for Mexico. The table shows the withdrawal requirements for each irrigation technology per 100 m3 of water required. The application efciencies require that 166.7 m3 be withdrawn to supply this amount with ood irrigation, 133.3 m3 with sprinklers, and 105.2 m3 with drip systems. A 50 50 mix of sprinkler and drip systems can reduce by 30% the water withdrawals relative to the requirements of ood irrigation. The mix of alternative technologies achieves an efciency of 86%, which is shown in Table 4 to translate to withdrawal of 117 m3 to satisfy 100 m3 of crop requirements. Alternative irrigation technologies also have different requirements for factors other than water. Few studies comparing the input requirements among irrigation techniques, however, are found in the literature. IWMI (2006) suggests that alternative irrigation


TABLE 4. Water withdrawal requirements for irrigation technologies. Water Requirements (m3/unit) A 100 100 100 100 Application Efciency B 0.60 0.75 0.95 0.86 Withdrawal required (m3) C A/B 166.7 133.3 105.3 116.7


Gain relative to ood irrigation (%) 20 37 30

Technique Flood Sprinkler Drip Mix of technologies

Source: Own computations based on efciency data from Brouwer et al (1989) and Postel (1999).

Downloaded by [carlos lopez] at 17:11 30 November 2011

technologies produce both water savings and higher yields per hectare of land, the latter saving of lower magnitude than the former. According to Neibling (1997) labor and capital inputs in sprinkler and drip irrigation systems can be higher and require more skills than traditional ood techniques. Dalton et al. (2002) likewise conclude that labor and capital inputs vary signicantly among irrigation techniques, with sprinkler and drip systems at the higher end. Lacking data for the Mexican case, we make the assumption that the mix of irrigation technologies increases labor and capital requirements by 30%, raising water costs relative to ood irrigation, as shown in Figure 2. While these assumptions result in a somewhat stylized representation of irrigation technologies due to lack of data, we believe that they succeed in highlighting the important differences among them and they enable the endogenous choice of technologies that our modeling framework can provide. The cost differences among the technologies respond to regional differences in factor costs and by regional endowments.
FIGURE 2. Regional no-trade prices of agricultural output for three technologies.

Source: Own computations.

398 4.5.

C. LOPEZ-MORALES and F. DUCHIN Design of Scenarios

Downloaded by [carlos lopez] at 17:11 30 November 2011

The baseline scenario, S0, imposes neither constraints on withdrawals nor water fees. Since it is obvious that built infrastructure is adequate to make current withdrawals feasible, we take calculated baseline withdrawals as a lower bound for the capacity of the existing withdrawal infrastructure to be used in remaining scenarios. Data for the actual volume of such capacity is not available for Mexico from either internal or external sources (such as FAOs AQUASTAT), so the relation between observed withdrawals and the capacity of built infrastructure cannot be determined. We make the assumption that for each region the latter is greater by 30% than baseline withdrawals, which is known to be possible with the existing infrastructure. The gure of 30% was arrived at by studying the relation between actual withdrawals and the capacity of built infrastructure for some countries in FAOs AQUASTAT database (FAO, 2011). Constraints for land, labor, and capital in each region are based on documented data and remain the same for all scenarios (see Table 3 for data sources). We are interested in two possible courses of action for public policy to incentivize the adoption of alternative irrigation technologies as a means of achieving sustainable regional water withdrawals. The rst policy scenario, S1, establishes a region-specic price for water as a component of the vector pi of ex ante factor prices. (We analyzed several water fee structures but only one is reported in the following sections.) Positive water fees are imposed in this scenario for regions whose withdrawals are unsustainable in the baseline scenario. The other policy scenario, S2, establishes quantity restrictions, or caps, setting each regions water endowment at its effective environmentally sustainable supply. An additional scenario, S3, is formulated to assess the cost of water sustainability in the absence of water-saving technology. Water endowments are set equal to the environmentally sustainable effective supply but the choice of agricultural technologies is restricted to non-irrigated and ood irrigated. If there is a feasible solution, costs and agricultural prices are compared with the policy scenarios in which water-saving technology adoption is allowed. Table 5 shows the assumptions under all four scenarios.

5. 5.1.

RESULTS OF THE SCENARIO ANALYSIS Sustainability of Water Withdrawals

We assess regional water sustainability by comparing water withdrawals under each scenario to the regions effective environmentally sustainable supply (see Table 6). Under the baseline S0, four regions of the low availability group (I, II, VI, and XIII) have baseline withdrawals that are greater than the effective environmentally sustainable supply of water, meaning that withdrawal patterns are unsustainable and reductions of water use are required. In these regions, the effective supply is greater than the environmentally sustainable supply, such that withdrawals must be restricted for environmental purposes. Other regions have baseline withdrawals that are smaller than the effective environmentally sustainable supply. Expansions of the former are allowed up to full utilization of the effective supply. The pricing scenario, S1, sets positive water fees in regions with unsustainable water use under S0 (see Table 7 for a description of fees), while the caps scenario,


TABLE 5. Description of scenarios.
Endowments Scenario Baseline (S0) Water Unconstrained Land Labor Capital Agricultural technologies


Water fees

Pricing (S1) Caps (S2) No-adoption (S3)

Effective supply Effective environmentally sustainable supply Effective environmentally sustainable supply

Observed Observed Observed Non-irrigated, No use + use use ood irrigation, Expansion mix of drip and sprinkler Yes No Non-irrigated, No ood irrigation

Source: Own elaboration

Downloaded by [carlos lopez] at 17:11 30 November 2011

S2, restricts water withdrawals to each regions effective environmentally sustainable supply. Under both policy scenarios, withdrawals are sustainable for all regions. Scenario S1 reduces baseline withdrawals by 10%, while S2 reduces them by 7% (see Table 6). In both cases, reductions are localized in the four regions of the low availability group that exhibited unsustainable baseline withdrawals. With water fees, withdrawals in other regions are unchanged from S0, while caps allow for an expansion of withdrawals in region IX, a region with medium water availability.


Technology Adoption

The distribution of land use among agricultural technologies under the baseline and the policy scenarios is shown in Table 8. Under baseline conditions, the alternative irrigation technologies incur in-unit production costs higher than ood irrigation and non-irrigated agriculture (see Figure 2), and therefore they are not adopted in any region. Water policies under S1 and S2 promote adoption of alternative technologies among the four regions with unsustainable baseline withdrawals (see Table 8), with the result that sustainability of water use is achieved in all regions (See Table 6). In no region from the remaining groups are alternative technologies adopted. With water fees in S1, ood irrigation becomes more costly than the water-saving irrigation technology in regions of the low availability group (see Figure 3). In the remaining regions, ood irrigation is selected as it remains less costly than the alternatives. Adopting regions in the low availability group have adequate endowments of the other factors to satisfy the higher factor requirements of the water-saving irrigation technology, such that they use this option exclusively. This is not the case of region IX, in the medium availability group, in which a positive water price was also set. Inspection of the solution shows that the endowment of irrigated land is not fully utilized in this region because there is not enough labor to satisfy the higher labor requirements of the alternative irrigation technology. Overall, alternative irrigation technologies are adopted on 31% of irrigated land in use. With caps on water withdrawals (but unpriced water) under S2, ood irrigation remains cheaper than alternative techniques. However, a region may run out of water for irrigation.



Blue water supply and withdrawals under alternative scenarios (km3/yr). Water Supply Water Withdrawals D 3.7 5.7 8.4 5.4 2.8 15.6 15.2 18.7 3.7 0.7 2.8 4.3 1.0 87.9 Observed 3.8 6.4 8.5 3.7 4.7 10.5 10.4 13.2 4.5 1.3 1.7 4.6 2.0 75.3 Baseline S0 4.5 7.8 10 4.5 5 12.0 11.8 16.3 2.9 0.5 2.1 3.3 0.8 81.6 Pricing S1 3.3 5.8 7.4 4.5 2.8 12.0 11.8 16.3 2.9 0.5 2.1 3.3 0.8 73.5 Caps S2 3.7 5.7 8.4 4.5 2.8 12.0 11.8 16.3 3.7 0.5 2.1 3.3 0.8 75.6 No adoption S3 3.7 5.7 8.4 4.5 2.8 12.0 11.7 16.3 3.7 0.5 2.1 3.3 0.8 75.5

Downloaded by [carlos lopez] at 17:11 30 November 2011

Group Low


A 4.6 8.2 12 7.8 3.5 25.6 21.7 34.0 25.5 32.8 29.6 95.5 157.8 458.6

B 3.7 5.7 8.4 5.4 2.8 17.9 15.2 18.7 16.6 19.7 16.3 57.3 94.7 282.4

C 5.8 10.2 13.1 5.9 6.5 15.6 15.3 21.2 3.7 0.7 2.8 4.3 1.0 106



High National

Note: Column A is regional runoff as reported by CNA (2010). Column B is the environmentally sustainable supply that results from subtracting regional EWRs of Table 1 from runoff. Column C is the effective supply, which equals 1.3 times baseline withdrawals (S0). Column D shows the effective environmentally sustainable supply, which is the smaller of B and C.


TABLE 7. Water fees for the pricing scenario (S1) (pesos/m3). Region I II VI VII XIII III IV VIII IX V XII X XI S1 2.22 1.74 1.32 1.08 0.42 Yes S2 0.02 3.23 0.41 1.2 No


Water availability group Low

S3 3.7 2.9 2.2 1.8 0.7 Yes


Downloaded by [carlos lopez] at 17:11 30 November 2011

High Sustainable withdrawals

Note: Column S1 shows the exogenous, region-specic water fees assumed for the pricing scenario. They are derived as follows. First, they were set equal to the endogenous scarcity rents for water obtained under S2, which resulted in unsustainable withdrawals. Second, they were set equal to the higher scarcity rents for water under the no-adoption scenario, S3, which resulted in sustainable withdrawals. These numbers were decreased until withdrawals rst became unsustainable, which occurred at 0.6 times the scarcity rents under S3. The latter are the fees shown in the table for S1.

Depending on production costs and factor constraints, a given region may use the watersaving technology exclusively, not at all, or simultaneously with another irrigation technology or with non-irrigated agriculture. All regions that adopt the water-saving technology run into water supply constraints under S2. For regions that use it exclusively (II and XIII), land endowments are not constraining, but they are constraining for regions with several irrigation technologies in operation (II and XIII). For the country as a whole, the quantity restrictions of scenario S2 promote adoption of the water-saving irrigation technology on 24% of used irrigated land, with adopting regions localized in the group of regions with low water availability.


Impact on Commodity Prices

Commodity prices are affected by both the exogenous fees for water (S1) and endogenous scarcity rents on fully utilized water supply (especially under S2). See Duchin and Levine (2011) for a discussion of the two-part factor prices in the World Trade Model. Resulting prices and production costs under the different scenarios are shown in Table 9. The pricing scenario (S1) and caps scenario (S2) increase the price for the agricultural commodity by 5 8% relative to the baseline while barely increasing region-wide production costs. Increased agricultural prices raise the cost of food and reduce competitiveness in international markets. However, the price increases are kept relatively small under S1 and S2 by the adoption of water-saving irrigation techniques. By contrast, the results under the no-adoption scenario, S3, show that the agricultural price would increase by 36%


Downloaded by [carlos lopez] at 17:11 30 November 2011


Land use by irrigation technology: Baseline and policy scenarios (103 ha). Irrigated land Non-irrigated land S1 S2 63 55 1,408 2,998 837 718 2,764 2,794 1,821 360 1,087 812 1,941 17,658 S0 Flood 258 573 688 472 186 830 787 1,113 118 72 5,095 Flood 472 830 787 1,113 118 72 3,391 Alternative 258 573 688 24 1,542 Flood 80 258 472 830 787 1,113 234 72 3,844 S2


Group Low


S0 63 55 1,408 2,998 837 718 2,113 2,794 2,280 360 1,087 812 1,941 17,465

S1 63 55 1,408 2,998 837 718 2,764 2,794 2,280 360 1,087 812 1,941 18,116

Alternative 178 564 430 30 1,202


High National

Source: Own computations.



FIGURE 3. Regional no-trade prices of agricultural output among three technologies after water fees are imposed.

Downloaded by [carlos lopez] at 17:11 30 November 2011

Source: Own computations.

relative to the baseline without the option for water-saving irrigation. This large increase results from an enforced shift to non-irrigated agriculture due to a shortage of irrigated land associated with sustainable water constraints.


Adoption Costs and Government Revenue

Both water pricing and restrictions on withdrawals can promote sustainable use of water but elevate economic costs, reected in changes in the value of the objective function shown in Table 10. In addition to these costs, there are upfront expenditures for equipping additional land for irrigation, which are not accounted for in the model. FAO reports investment costs for drip and sprinkler irrigation (FAO, 2011): we utilize their low and high estimates (see Table 11), interpreting them as present values, and compute corresponding annuities at different time-horizons (5, 10, and 20 years) and discount rates (5% and 10%). These are reported and compared with government revenues, as calculated by the World Trade Model, in Table 12. The two policy intervention scenarios yield government income via the collection of water fees under S1 or the appropriation of scarcity rents on fully utilized water sources in S2. If the governments water revenue exceeds total adoption costs, then in principle agricultural producers can be subsidized to cover their additional costs. Table 12 compares economic costs of adoption to government revenue. Under the cap scenario, S2, water revenue does not cover adoption costs. In the pricing scenario, S1, however, costs can be fully covered by the governments water revenue. Provided that upfront investment costs do not lie at the higher end of the estimated range, and time horizons of irrigation investments are longer than ve years, farmers can be incentivized to achieve sustainable



TABLE 9. Commodity prices under the policy scenarios (price index relative to baseline price). Scenarios Sector Agriculture Livestock, Forestry & Fishing Oil and gas Non-Oil Mining Power generation Construction Processed Food Light manufacture Chemistry, Plastics Heavy Manufacture Commerce Transport Non-Financial Services Financial Services Social Services S1 1.05 1.01 1.00 1.00 1.00 0.98 1.01 1.00 1.00 1.00 1.01 1.01 1.01 1.02 1.01 S2 1.08 1.01 1.00 1.00 1.00 1.01 1.01 1.00 1.00 1.00 1.01 1.00 1.00 0.99 1.01 S3 1.36 1.05 1.00 1.00 1.00 1.01 1.05 1.01 1.00 1.00 1.04 1.03 1.02 1.01 1.04

Downloaded by [carlos lopez] at 17:11 30 November 2011

Source: Own computations. For non-traded commodities the average national price is shown.

TABLE 10. Increase in factor costs under policy scenarios (increase relative to baseline). Scenario Pricing (S1) Caps (S2)
Source: Own computations.

Factor costs (106 pesos) 4,749 48,422

TABLE 11. Estimated investment cost for irrigation technologies (pesos/ha). Technology Range Low Medium High Drip 1,429 2,075 5,934 Sprinkler 524 2,677 8,430 Mix 976 2,376 7,182

Source: Own computations based on FAO (2011).

use of water. Of course, one cannot conclude on the basis of these results that this compensation should be prioritized, as the opportunity costs of spending governments resources on other projects would need to be assessed. However, our results suggest that levying water fees is preferable to the alternative of limiting quantities of water



TABLE 12. Costs of technology adoption and government water revenues under policy scenarios (106 pesos). Pricing scenario (S1) 5 years 5% Low Medium High Government revenue 8,311 13,298 29,682 20,936 Caps scenario (S2) 5 years 10 years 5% 49,979 52,158 59,319 10% 50,378 53,117 62,116 20 years 5% 49,387 50,737 55,174 10% 49,834 51,811 58,305 10% 8,817 14,512 33,226 10 years 5% 6,746 9,542 18,729 10% 7,259 10,772 22,317 20 years 5% 5,986 7,719 13,411 10% 6,560 9,096 17,429

Downloaded by [carlos lopez] at 17:11 30 November 2011

5% Low Medium High Government revenue 51,199 55,085 67,856 16,650

10% 51,593 56,032 70,618

Source: Own computations. Note: Indicates that government water revenue exceeds total cost of adoption.

withdrawn: government revenue is likely to compensate for higher costs of assuring compliance in the former case, while in the latter case additional institutional frameworks are required to enforce the permit system and collect the scarcity rents generated.



This paper analyzes the potential for government policies involving pricing and caps on water withdrawals to induce adoption of new water-saving irrigation technologies capable of achieving sustainable water withdrawals in Mexico. We formulated a set of scenarios and customized the World Trade Model for their analysis. The study required designating hydro-economic regions, dening region-specic sustainable endowments of water resources, and compiling the required data from a variety of sources. We found that these policies could keep water withdrawals within each regions effective sustainable endowment of blue water, and government revenues were large enough to fully subsidize costs to agricultural producers. We envisage various extensions to the research reported in this paper. First, the agricultural sector needs to be disaggregated to distinguish the requirements of different kinds of crops. Second, surface and underground sources of blue water need to be distinguished to isolate instances of the overdraft of aquifers. Third, the pricing algorithm should be rened to determine the water fees that can promote sustainability at minimal economic costs. Finally, this modeling framework can be applied to scenarios about future growth in population and afuence, changes in diets, and changes in the availability of water as a result of



climate change. Another challenge is to examine the quality of wastewater and policies to avoid the contamination of water sources.

Acknowledgements We thank Nathaniel Springer for discussions throughout the design and implementation of the model. We also appreciate the thoughtful and thorough comments of two anonymous referees, most of which we have incorporated.

Berrittella, M., K. Rehdanz, R. Roson, and R. Tol (2005) The economic impact of water pricing: A computable general equilibrium analysis. Working Paper FNU-96, Hamburg University and Centre for Marine and Atmospheric Science, Hamburg.. Berrittella, M., A.Y. Hoekstra, K. Rehdanz, R. Roson and R.S. Tol (2007) The Economic Impact of Restricted Water Supply: A Computable General Equilibrium Analysis. Water Research, 41, 17991813. Blancas, A. (2006) Interinstitutional Linkage Analysis: A Social Accounting Matrix Multiplier Approach for the Mexican Economy. Economic Systems Research, 18, 29 59. Brouwer, C., K. Prins, and M. Heibloem (1989) Irrigation Water Management: Irrigation Scheduling. Training Manual No. 4. Food and Agriculture Organization, Rome, Italy. Calzadilla, A., K. Rehdanz, and R.S.J. Tol. (2008) Water Scarcity and the Impact of Improved Irrigation Management: A CGE Analysis. Kiel Working Papers No. 1436, Kiel Institute for the World Economy, Germany.. Chapa-Cantu, J., E. Ayala-Gaytan and I. Hernandez-Gonzalez (2009) Modelo de insumo producto para el Noreste de Mexico. Ciencia UANL, 12, 409 416. CNA. (2008) Programa nacional hdrico 2007-2012. Secretara de Medio Ambiente y Recursos Naturales, Mexico D.F. CNA. (2010) Estadsticas del agua en Mexico 2010. Secretara del Medio Ambiente y Recursos Naturales, Mexico D.F. Dalton, T., A. Files and D. Yarborough (2002) Investment, Ownership and Operating Costs of Supplemental Irrigation Systems for Maine Wild Blueberries. Technical Bulletin 83, Maine Agricultural and Forest Experiment Station, The University of Maine.. Diario Ocial de la Federacion. (2009) Ley Federal de Derecho. , Diario Ocial de la Federacion, November 27. (Last accessed, 23 September 2011). Dietzenbacher, E. and E. Velazquez (2007) Analysing Andalusian Virtual Water Trade in an Input-Output Framework. Regional Studies, 41, 185 196. Duarte, R., J. Sanchez-Choliz and J. Bielsa (2002) Water Use in the Spanish Economy: An Input-Output Approach. Ecological Economics, 43, 71 85. Duchin, F. (2005) A World Trade Model Based on Comparative Advantage with m Regions, n Goods, and k Factors. Economic Systems Research, 17, 141162. Duchin, F. and G. Lange (1992) Technological Choices and Prices, and their Implications for the US Economy, 1963-2000. Economic Systems Research, 4, 5376. Duchin, F. and S.H. Levine (2011) Sectors May Use Multiple Technologies Simultaneously: The Rectangular Choice-of-Technology Model with Binding Factor Constraints. Economic Systems Research, 23, 281 302. Duchin, F. and C. Lopez (2011) World Trade in Food as an Adjustment Mechanism to Water Scarcity. Submitted for publication.. Falkenmark, M. and G. Lindh (1976) Water for a Starving World. Boulder, CO, USA: Westview Press. FAO. (2011) Database on Investment Costs in Irrigation. Electronic database (Last accessed 23 June 2011). FAO-AQUASTAT. (2009) Mexico. Summary Fact Sheet. Country Fact Sheets. Electronic database (Last accessed 23 June 2011). Guan, D. and K. Hubacek (2008) A New and Integrated Hydro-economic Accounting and Analytical Framework for Water Resources: A Case Study for North China. Journal of Environmental Management, 88, 1300 1313.

Downloaded by [carlos lopez] at 17:11 30 November 2011



Harris, T. and M. Rea (1984) Estimating the Value of Water among Regional Economic Sectors Using the 1972 National Interindustry Format. American Water Resources Association, 20, 193 201. He, L. and F. Duchin (2009) Regional Development in China: Interregional Transportation Infrastructure and Regional Comparative Advantage. Economic Systems Research, 21, 3 22. Henry, M.S. and E. Bowen (1981) A Method for Estimating the Value of Water among Sectors of a Regional Economy. Southern Journal of Agricultural Economics, 13, 125 132. Hubacek, K. and L. Sun (2005) Economic and Societal Changes in China and their Effects on Water Use. Journal of Industrial Ecology, 1, 187200. INEGI. (2011a) Sistema de cuentas nacionales: Matrz de insumo producto 2003. Electronic database (Last accessed 23 June 2011). INEGI. (2011b) Sistema de consulta de los censos econonicos 2004. Electronic database (Last accessed 23 June 2011). INEGI. (2011c) VIII Censo agrcola, ganadero y forestal 2007. Electronic database. (Last accessed 29 June 2011). IWMI. (2006) Promoting Micro-irrigation Technologies that Reduce Poverty. Water Policy Brieng 23. International Water Management Institute, Sri Lanka. Julia, R. and F. Duchin (2007) World Trade as the Adjustment Mechanism of Agriculture to Climate Change. Climatic Change, 82, 393 409. Lee Harris, R. (2002) Estimation of a Regionalized Mexican Social Accounting Matrix using Entropy Techniques to Reconcile Disparate Data Sources. Trade and Macroeconomic Division Discussion Papers No. 97, International Food Policy Research Institute. Washington, DC. Lenzen, M. and B. Foran (2001) An Input Output Analysis of Australian Water Usage. Water Policy, 3, 321 340. Lopez-Morales, C. (2010) Policies and Technologies for a Sustainable Use of Water in Mexico. A Scenario Analysis. PhD Dissertation, Department of Economics, Rensselaer Polytechnic Institute, Troy, NY. Liu, X., X. Chen and S. Wang (2009) Evaluating and Predicting Shadow Prices of Water Resources in China and its Nine Major River Basins. Water Resources Management, 23, 1467 1478. Llop, M. (2008) Economic Impact of Alternative Water Policy Scenarios in the Spanish Production System: An Input-Output Analysis. Ecological Economics, 68, 288 294. Neibling, H. (1997) Irrigation Systems for Idaho Agriculture. Agricultural Experiment Station CIS1055, Cooperative Extension System, College of Agriculture, University of Idaho. Postel, S. (1999) Pillar of Sand: Can the Irrigation Miracle Last? New York, WW Norton & Company. Rosegrant, M., X. Cai and S.A. Cline (2002) World Water and Food to 2025: Dealing with Scarcity. Washington DC, USA: International Food Policy Research Institute. SAGARPA. (2009) Sistema de informacion agropecuaria y pesquera. Anuarios agropecuarios 1980 2008, Electronic database (Last accessed 29 June 2011). SEMARNAT. (2010) Sistema nacional de informacion ambiental y recursos naturales. Electronic database (Last accessed 29 June 2011). SENER. (2010) Sistema de informacion de energa. Electronic database. Secretara de Energa, Mexico (Last accessed 29 June 2011). Smakhtin, V.C.R. and P. Doll (2004) Taking into Account Environmental Water Requirements in Global Scale Water Resources Assessments. Comprehensive Assessment Research Report 2, International Water Management Institute, Colombo, Sri Lanka. Strmman, A. and F. Duchin (2006) A World Trade Model with Bilateral Trade Based on Comparative Advantage. Economic Systems Research, 18, 281 297. Strmman, A., E. Hertwich and F. Duchin (2009) Shifting Trade Patterns as a Means of Reducing Global Carbon Dioxide Emissions: A Multi-objective Analysis. Journal of Industrial Ecology, 13, 38 57. Velazquez, E. (2006) An Input-Output Model of Water Consumption: Analysing Intersectoral Water Relationships in Andalusia. Ecological Economics, 56, 226240. Velazquez, E., M. Cardenete and G. Hewings (2007) Water Price and Water Reallocation in Andalusia. A Computable General Equilibrium Approach. Working Paper No. 07.04. Department of Economics, Universidad Pablo de Olavide, Seville, Spain.

Downloaded by [carlos lopez] at 17:11 30 November 2011