Key Messages 
  • Computational general equilibrium models can be used to consider the economy-wide implications of planned adaptation actions.
  • The insight presents the extension of a CGE model to account for adaptation through an increased demand for irrigation services to reduce the adverse effects of climate change in agriculture.
  • Results suggest that even though irrigation expansion can reduce the initial production losses induced by climate change, final effects on GDP are small and/or ambiguous. This because (a) agricultural value added tend to build anyway a minor part of country value added (b) international trade and competitiveness effects in the market of agricultural commodities triggered by the application of irrigation produce both winners and losers independently upon the initial productivity gains.


Some adaptation responses can be driven by self-regulatory mechanisms or autonomous reaction, other by planned policy intervention. Typical examples of autonomous adaptation at the national and international level are changes in agents’ behaviour adjusting/adapting to new price conditions. These autonomous responses can be strengthened by planned strategies, often promoted by governments.

Planned adaptation will require increased expenditures to protect private assets at risk as well as to cope with the adverse effects of climate change, in this case in agricultural activities. The economy-wide effects of planned adaptation can be modelled through a computable general equilibrium (CGE) framework. CGE models consider, by construction, the so called “market-driven adaptation” i.e. the functioning of autonomous mechanisms – primarily, demand and supply reactions to endogenous changes in relative prices – which characterize instantaneous resource allocation across two market equilibria in response to exogenous economic shocks (Bosello and Parrado, 2014).

The literature surrounding global CGE models explicitly considering water resources and irrigation as production factors in agriculture is limited, although a number of relevant models have been developed (e.g. GTAP-W1, GTAP-W2, ICES-W, and GTAP-BIO-W).

This insight presents the changes made to a CGE model (ICES) extended with an irrigation module to integrate farmers’ adaptation through an increased demand for irrigation services to reduce the adverse effects of climate change in agriculture. The new specification modifies the model’s land supply structure in order to consider different land rents and imperfect flexible land conversion between pasture and cropland (irrigable and rainfed land) and among different crop industries. Moreover, it takes into account the additional capital, operational and maintenance costs that farmers face when they decide to expand irrigation.

Policy and methodological developments 

The basic ICES model is a multi-region, multi-sector, recursive dynamic CGE model of the global economy, derived from the GTAP-E model (Burniaux and Truong, 2002), which in turn is the energy environmental extension of the standard GTAP model (Hertel, 1999). Features of ICES are similar to most CGE models: domestic production is determined by a series of nested constant elasticity of substitution (CES) functions, which specify substitutability between primary factors, energy and non-energy intermediates. The demand side of the economy is characterized by a representative household, who receives income from primary factors, and allocates it across private consumption, public consumption and saving so as to maximize per capita aggregate utility, according to a Cobb-Douglas function. A global bank collects global net savings and allocates them amongst regions according to relative rates of return to capital. Bilateral trade is specified by assuming imperfect substitutability to consider product heterogeneity by country origin (Armington, 1969).

The main innovation of the ICES-IRR model is related to the specification of the crop production function. Farmers decide first which combination of rainfed and irrigated land to use. The use of irrigable land requires capital and infrastructure. Hence, irrigated land is not only more productive, but also more costly. To account for this, a new intermediate factor, called “irrigation services”, is included. Irrigable land and irrigation services are combined to determine irrigated land, which in turn is an imperfect substitute of rainfed land.

Irrigation services are on their turn a new production sector that uses energy, water distribution services, capital and labour. To take into account limitations on the potential use of irrigation services imposed by water availability the supply of this sector is constrained by including a fixed factor. As usual it can be interpreted as the water (scarcity) rents.

Another significant improvement in the ICES-IRR model is the new land supply structure. In the standard ICES model, a fixed land supply at the country level is imperfectly substitutable across different crops depending on relative crop prices and land rents. Imperfect land transformability is represented by a one-level Constant Elasticity of Transformation (CET) function. This specification implies to consider land equally and easily transformable across different activities (e.g., wheat, rice, millet, livestock, etc.) and land types (e.g. cropland, pasture land etc.).This is not very realistic. In particular, irrigated cropland is usually more valuable and “less substitutable” because it should meet specific conditions in terms of slope, drainage, texture, soil depth, etc. (FAO, 1997).

Following Taheripour et al. (2013), a three-level CET function was developed. At the bottom level, land in each region is assumed to be transformable among pasture and cropland, while at the second level cropland is allocated between irrigable land and dryland. Finally at the top level, land supply among crops is treated as in the standard ICES model. Differently from Taheripour et al. (2013), where irrigable land is supplied only to irrigated crop production, in ICES-IRR both irrigable and rainfed land can be used by all crop-producing industries.

The modelling framework was applied to compare a no-adaptation scenario, where climate change impacts are imposed assuming fixed amount of irrigated and rainfed land and an adaptation scenario, in which farmers are allowed to expand irrigated land to contrast yield losses from climate change. Results suggest the following:

  • In the no-adaptation scenario, lower latitude countries are those most negatively affected either in terms of decreased crop production or lower GDP. GDP can reduce by -1.4% in Asian countries by mid-century. Some higher latitude countries, e.g. Northern EU and the Former Soviet Union could experience slight GDP gains as a consequence of higher crop yields.
  • When irrigation can be expanded in the Adaptation scenario, macroeconomic results do not differ much from the No Adaptation case, although tiny average GDP gains can be observed in many regions: Sub Saharan Africa, South Asia, China, East Asia, Mediterranean and East Europe, South Korea, Australia, USA and Canada, in particular in RCP 8.5.

International trade also matters, influencing demand patterns. In general, regions with lower increases in domestic prices compared with world prices would also export more and vice versa. Climate change will thus reallocate agricultural production from most to less affected sectors and countries.

Main implications and recommendations 

Results from running the ICES-IRR suggest some interesting insights into the economy-wide implications of adaptation in agriculture through irrigation. Model results demonstrate that irrigation expansion can be an effective adaptation option in particular for lower latitude countries enabling higher production and lower GDP losses. However, gains compared to the no adaptation case are tiny in percentage terms. Converting rainfed into irrigable land and expanding irrigation services is costly and in the end increases further agricultural prices which compresses demand expansion. The final effect of flexible irrigation is a reallocation of crop production from developed to developing countries which are advantaged in relative terms by a combination of lower irrigation costs with the initial climatic impacts.

The extended ICES-IRR model is a useful addition to existing modelling approaches for investigating the economy-wide effect of adaptation in agriculture. The modelling approach to include irrigation as a planned adaptation strategy includes a new specification that modifies the model land supply structure in order to consider different land rents and imperfect flexible land conversion between pasture and cropland, irrigable and rainfed land and among different crop industries. Moreover, it takes into account the additional capital, operational and maintenance costs that farmers face when they decide to expand irrigation. This is a novelty compared to the existing literature, in which few studies analyse the role of irrigation as an adaptation strategy (Berittella et al., 2006; Calzadilla et al., 2013), and, most importantly, treat irrigation as an exogenous variable rather than as an autonomous farmers' decision.


Armington, P. S. (1969), A theory of demand for products distinguished by place of production. IMF Staff Papers 16 (1), 159-178.

Berrittella, M., Rehdanz, K., Tol, R. (2006), The economic impact of the south-north water transfer project in China: A computable general equilibrium analysis. Working Papers FNU-117, Research Unit Sustainability and Global Change, Hamburg University and Centre for Marine and Atmospheric Science.

Bosello, F., Parrado, R. (2014), Climate change impacts and market driven adaptation: The costs of inaction including market rigidities. Working Papers 2014.64, Fondazione Eni Enrico Mattei.

Burniaux, J.-M., Truong, T. P. (2002), GTAP-E: An energy-environmental version of the GTAP model. GTAP Technical Paper 16, GTAP Technical Paper.

Calzadilla, A., Zhu, T., Rehdanz, K., Tol, R. S., Ringler, C. (2013), Economy-wide impacts of climate change on agriculture in Sub-Saharan Africa. Ecological Economics 93, 150-165.

Food and Agricultural Organization (FAO) (1997), Irrigation potential in Africa: A basin approach. Rome: Food and Agricultural Organization of the United Nations.

Hertel, T. W., Tsigas, M. (1999), Structure of GTAP. In T. W. Hertel (Ed.), Global trade analysis: Modeling and applications. Cambridge University Press.

Taheripour, F., Hertel, T., Liu, J. (2013), Introducing water by river basin into the GTAP-BIO model: GTAP-BIO-W. GTAP Working Papers 77, Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University.