What Does Climate Change Mean for Agriculture in Developing Countries? A Comment on Mendelsohn and Dinar

Mendelsohn and Dinar review much of the important work on the implications of climate change for agriculture, focusing particularly on developing countries. Their message is that efficient economic adaptation significantly reduces the estimated effects of climate change. Few dispute that some amount of adaptation is likely and that its potential contribution to reducing the negative impacts of global warming is large. One such study (Darwin and others 1995), which analyzed the global impacts using an ecozone (land class) methodology, found that without adaptation, average cereal production yields fell roughly 20 to 30 percent in four different climate scenarios. Through various channels of adaptation (modifying crops and techniques on existing farmland, shifting crops to new land, and responding to changing market prices), these losses were reversed, resulting in small increases in production worldwide (0 to 1 percent) even before considering the positive effects of carbon dioxide (CO 2) fertilization (table 1). Striking, however, are both the initial shock in cereal production in the study reported in table 1 and the range of impacts on yields (without adaptation) estimated by a variety of studies for different sites around the world (shown in table 2). The Ricardian method reported by Mendelsohn and Dinar and the ecozone (land class) method of Darwin and others (1995) are similar in that they use cross-sectional evidence to estimate the adaptation response to climate change that occurs over time. Darwin and others (1995) use this evidence to estimate productivity shocks that are introduced into a general equilibrium model. As Mendelsohn and Dinar note, the Ricardian method is limited because it does not account for market effects, that is, the fluctuation of prices reflecting market conditions. The result is thus strictly applicable only to a closed economy. Mendelsohn and Dinar note that this bias will be small if the global price effect is small, and they cite a study by Reilly, Hohmann, and Kane (1994) that shows small price effects in some scenarios. This single study is