This study assesses the interaction between climate change and agricultural trade policies. We distinguish between two dimensions of agricultural trade policy: market insulation and subsidy levels. Building on the previous work of Tsigas, Frisvold and Kuhn (1997) we find that, in the presence of current levels of agricultural subsidies, increased price transmission --as called for under the Uruguay Round Agreement on Agriculture-- reduces global welfare in the wake of climate change. This is due to the positive correlation between productivity changes and current levels of agricultural support. Increases in subsidized output under climate change tend to exacerbate inefficiencies in the global agricultural economy in the absence of market insulation. However, once agricultural subsidies have also been eliminated, price transmission via the global trading system contributes positively to economic adaptation under climate change. This may partially explain the relatively slow growth of world grain import demand in recent years. In addition, bilateral agreements with East Asia, NAFTA, and the evolution of the CAP, have all had important impacts on the structure of world food and agricultural trade. The objective of this paper is to assess the relative role of each of the major forces-- consumer demand, factor accumulation, transport costs, and policy change--in driving changes in the composition of world food trade in 1980-1995. To do so, we employ a modified version of the Global Trade Analysis Project (GTAP) model of world trade which permits us to isolate the contribution of each of these related factors to the changing composition of world food trade in a general equilibrium context. We evaluate the relative role of each of these factors by simulating the model backwards in time, from 1995 to 1980 under different assumptions. This general approach, termed "backcasting" (i.e. backwards forecasting), takes as exogenous the basic drivers of change and attempts to explain the resulting change in food trade composition. The model-produced changes in the composition of agricultural and food trade are compared with historical trade data, to determine the relative importance of each factor on the changing composition of food trade. Given limited space, our focus will be on explaining the changes in the global composition of food and agriculture trade. A natural follow-on effort would target specific markets in more detail. This type of backcasting approach was first employed by Gehlhar (1997) who sought to explain the shift in exports of primary commodities to manufactures in East Asia in the 1980's. He calibrated the GTAP model to 1992 data, then implemented shocks to factor endowments and economy-wide total factor productivity (TFP) in order to force each economy back to its 1982 levels of population, land, labor, human capital, physical capital and technology. By comparing actual and predicted changes in export shares in this period, he found human capital accumulation played a key role in explaining the change in the aggregate composition of East Asian exports. Gehlhar, Hertel and Martin (1994) built on this work in an effort to predict future changes in the pattern of agricultural trade from 1992-2002. They also emphasized the importance of supply-side determinants of agricultural trade. In this paper, we go beyond this earlier work in a number of ways. First, we focus on the composition of agricultural exports, rather than simply looking at the share of agriculture in total trade. Secondly, we incorporate the Cranfield et al. estimates of Rimmer and Powell's recently developed, implicitly directly additive demand system (nicknamed AIDADS) into the GTAP model. This permits us to better capture the impact of demand-side changes on the pattern of global trade.