A significant effort has been devoted by scientists from various disciplines to shed light on the causes and effects of climate change in recent years (Tol 2010). Although there is still controversy about the details (Idso and Singer 2009), it is widely accepted that climate change has already started to occur, and the impacts will increase through the 21st century (Agrawala and Fankhauser 2008; Parry et al. 2007; Stern 2006). There are a wide range of social and physical effects that are linked to climate change in the literature, and the most significant effects are expected to be increasing temperatures accompanied by declining precipitation plus increasing frequency of climatic extremes. Hence, agricultural production, which is among the most climate-dependent economic activities, is likely to be the most vulnerable sector (Fankhauser 2005; Rosegrant et al. 2008). Various methods are employed to translate the physical effects to economic shocks. The most popular way is introducing climate change shocks through the agricultural sector as yield or water requirement shocks. Although there is no consensus among the conclusions of these studies, some general results can be derived. The results suggest an average negative welfare effect between 1 to 2 percent of gross domestic (GDP) at the global level (Calzadilla et al. 2010; Tol 2012). Nonetheless, aggregate impacts are generally considered as weak due to the adjustment effects of economic agents and markets in response to climate-induced shocks (Bosello et al. 2010). However, effects are non-homogenous over space, time, sectors, and social groups. Country level analysis suggest more significant effects especially in the Middle East (Sowers and Weinthal 2010; World Bank 2010), Africa (Arndt et al. 2012a, 2012b; Pauw, Thurlow and van Seventer 2010; Thurlow, Zhu, and Diao 2012), and South Asia (Thurlow, Dorosh, and Yu 2012).
Among the growth levers in the economic landscape for developing countries, international trade is argued to offer a potential for adaptation in the face of climate change. This is achieved through the enabling channels of technological spill-overs and enhanced access to capital and infrastructure investments and production specialization. Trade has the potential to alleviate the climate-induced scarcity burden (especially in the agricultural sector) by bridging the differences between demand and supply conditions globally. Nonetheless, it can also increase climate-induced vulnerability in certain regions which specialize in the production of certain products in which they have a comparative advantage, while relying on imports to meet their demands for other commodities and services. Trade liberalization is reported to have welfare-improving effects (Calzadilla, Rehdanz, and Tol 2011; Chang, Chen, and McCarl 2012; Laborde 2011; Reilly and Hohmann 1993). However these effects are generally insufficient to compensate the adverse effects of climate change (Hertel and Randhir 2000; Reilly and Hohmann 1993). Welfare gains from trade liberalization depend primarily on the elimination of trade barriers such as tariffs and quotas, and subsidies. The effects are not uniform and depend on the geographic location (Calzadilla et al. 2010; Reilly and Hohmann 1993) and vulnerability of the region to climate change (Reilly and Hohmann 1993), and poor people are expected to be adversely affected more from the changes (Laborde 2011).
The objective of this paper is to analyse the climate change and trade linkages, and evaluate the potential of trade liberalization (i.e. tariff elimination) as a means of adaptation in the context of developing countries. Our focus is on Morocco and Turkey as case studies, where we use the GTAP model to investigate trade liberalization scenarios' welfare impacts under climate change. In Section 2, we present our methodological approach for developing the range of global yield forecasts and data sources. Section 3 discusses the results for the world and the regional patterns in welfare impacts.
Section 4 focuses on the Moroccan and Turkish cases, Section 5 summarizes our key findings and conclusions.
2 Methodological approach, scenarios, and data discussion
2.1 Modelling framework: GTAP Model
To estimate the impacts of climate-induced agricultural productivity shocks on the economy and the linkages with international trade, we use the Global Trade Analysis Project (GTAP) general equilibrium global trade model and its accompanying database. The GTAP model is a multicommodity, multi-region computable general equilibrium model (Hertel 1997). In the standard GTAP model, we assume markets are perfectly competitive and exhibit constant returns to scale. Consumers, as represented by the private household, maximize utility where consumption is modeled via a nonhomothetic constant difference of elasticity implicit expenditure function. Producers are assumed to maximize profits subject to a nested constant elasticity of substitution production function which bundles primary factors and intermediate inputs to produce final outputs. For the purpose of our analysis, a standard neo-classical closure is assumed where producers earn zero-profits, the regional household is on its budget constraint, and global investment equals global savings, with equilibrium imposed in all markets. World price of a given commodity is determined through the global trade balance.
We make use of the GTAP database version 7.0 which provides a disaggregation of agricultural production and harvested area by agro-ecological zones (1) (AEZ) by using the Food and Agriculture Organization (FAO) data on production, harvested area and price, available by country and 159 FAO crop categories. The GTAP database version 7.0 has been aggregated in order to accommodate the needs of the analysis. The new aggregation collapses the dimensions of the GTAP database into 16 regions, 15 sectors, and five factor endowments.
The objective of the analysis is to shed light on the potential impacts of trade policy as an adaptation tool in the face of climate change. Therefore, we will attempt to analyse the potential adaptive impacts of selected trade policy liberalization scenarios, especially for the agricultural sector. In particular, the analysis will focus on the macroeconomic and welfare linkages of trade liberalization and climate change impacts in terms of trade (TOT) flows and production impacts globally and more specifically in Morocco and Turkey. We conduct the analysis in a comparative static mode where projected yield shocks by 2050 are introduced into the model as productivity shocks to the technology parameters in the model. Table 1 summarizes the selected simulation scenarios. In order to investigate the effects of trade liberalization, we calculate the net effect of key variables by comparing the results from the climate change only scenario with the results from the climate change plus trade liberalization scenarios (Table 1 in the Appendix).
2.2 Discussion of data sources for climate-induced yield shocks
Given the objective of the study, we needed to develop a global dataset that takes into consideration the inherent uncertainty in terms of regional distribution of impacts and the heterogeneous nature of their magnitude across climate scenarios. We have identified two major sources for the yield impact data: IFPRI Food Security CASE Maps database (IFPRI 2010) and the Integrated Model to Assess the Global Environment (IMAGE) Version 2.2. Combining the two databases, we create a comprehensive set of projected yield change estimates that provides estimates of productivity shocks on the basis of the regional and sector aggregation adopted in the analysis.
The IFPRI food security CASE maps database
The IFPRI database provides projected yield impacts globally for six crops (rice, wheat, maize, cassava, groundnut, and soybean) under a wide range of scenarios based on simulations from the IMPACT model by 5-year increments until 2050 (Nelson et al. 2010).The results are provided for three overall scenarios (pessimistic, baseline, and optimistic) that capture the dynamics of economic growth based on assumptions about per capita GPD growth and population growth (Table 2).
The yield impacts are estimated via the Decision Support System for Agrotechnology Transfer (DSSAT) model version 4.5 using two Intergovernmental Panel on Climate Change (IPCC) climate scenarios that project future greenhouse gas emission (A1B and B1) and two global circulation models (GCM) for the climate (called CSIRO and MIROC), the combination of which represent two futures: a dry and relatively cool future under the combination CSIRO A1B and B1, and a wet and warmer future under the combination MIROC A1B and B1.
Hence, combining the scenarios capturing the projected per capita GDP growth and population growth, the SRES scenarios and the GCM climate models, results in 15 potential future pathways that encompass a wide range of plausible outcomes in terms of projected yields. Yield estimates from the IMPACT model are dynamically produced under each of the 15 pathways for irrigated and rainfed systems, and incorporate assumptions about exogenous yield growth (i.e. the intrinsic productivity growth rates, IPRs) and exogenous area changes (i.e. the intrinsic area growth rates, IARs). Figures 1a and 1b summarize a sample of the data on yield and IPR for wheat in the top five producing regions.
To isolate the effect of climate change on yield, we combine the yield and IPR data generated by the IMPACT model. As documented in Nelson et al. (2010), the IMPACT model captures the effects of climate change through the alteration of crop area and yield, represented by Equation (1) and Equation (2) as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
The key parameters of interest are [gy.sub.tni] and [ga.sub.tni]. They...