WPS4758 Policy ReseaRch WoRking PaPeR 4758 How China's Farmers Adapt to Climate Change Jinxia Wang Robert Mendelsohn Ariel Dinar Jikun Huang The World Bank Development Research Group Sustainable Rural and Urban Development Team October 2008 Policy ReseaRch WoRking PaPeR 4758 Abstract This paper uses a cross sectional method to analyze farmers have adapted to current climate, provides insight irrigation choice and crop choice across 8,405 farmers in into how they will likely adapt when climate changes. 28 provinces in China. The findings show that Chinese Future climate scenarios will cause farmers in China farmers are more likely to irrigate when facing lower to want to reduce irrigation and shift toward oil crops, temperatures and less precipitation. Farmers in warmer wheat, and especially cotton. In turn, farmers will shift places are more likely to choose oil crops, maize, and away from potatoes, rice, vegetables, and soybeans. especially cotton and wheat, and are less likely to choose However, adaptation will likely vary greatly from region vegetables, potatoes, sugar, and especially rice and to region. Policy makers should anticipate that adaptation soybeans. In wetter locations, farmers are more likely to is important, that the magnitude of changes depends on choose soybeans, oil crops, sugar, vegetables, cotton, and the climate scenario, and that the desired changes depend especially rice, and they are less likely to choose potatoes, on the location of each farm. wheat, and especially maize. The analysis of how Chinese This paper--a product of the Sustainable Rural and Urban Development Team, Development Research Group--is part of a larger effort in the department to mainstream research on climate change. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at jxwang.ccap@igsnrr.ac.cn, robert. mendelsohn@yale.edu, adinar@worldbank.org (After 12/1/2008 adinar@ucr.edu), jkhuang.ccap@igsnrr.ac.cn. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team HOW CHINA'S FARMERS ADAPT TO CLIMATE CHANGE Jinxia Wang, Robert Mendelsohn, Ariel Dinar and Jikun Huang Jinxia Wang is an Associate Professor in the Center for Chinese Agricultural Policy (CCAP), Institute of Geographical Sciences and Natural Resource Research, Chinese Academy of Sciences, Beijing, China (jxwang.ccap@igsnrr.ac.cn). Robert Mendelsohn is Edwin Weyerhaeuser Davis Professor in the School of Forestry and Environmental Studies in the Yale University, USA, (robert.mendelsohn@yale.edu). Ariel Dinar is a Lead Economist in the Development Research Group of the World Bank, Washington DC, USA (adinar@worldbank.org). Jikun Huang is Director and Professor of CCAP (jkhuang.ccap@igsnrr.ac.cn) We would like to thank Lijuan Zhang, Jianmin Cao, Cheng Chen, Yumin Li, Xiangjun Xing and Hao Li for their assistance in data cleaning. We acknowledge financial suport from DECRG in the World Bank, State Office of Comprehensive Agricultural Development (SOCAD) in China, and Global Environmental Foundation and China's National Natural Sciences Foundation (70733004) and Chinese Academy of Sciences (KSCX-YW-09). We acknowledge the constructive comments by Carter Brandon. The views expressed in this paper are those of the authors and should not be attributed to any of the funding sources. SUMMARY This paper examines how farmers have adapted to the current range of climates across China. A cross sectional method is used to analyze irrigation choice and crop choice across 8,405 farmers in 28 provinces in China. A discrete choice logit model is used to capture the choice of irrigation and a multinomial logit model is used to capture crop choice. We find that both irrigation and crop choice decisions are climate sensitive. Chinese farmers are more likely to irrigate when facing lower temperatures and less precipitation. Farmers in warmer places are more likely to choose oil crops, maize, and especially cotton and wheat, and are less likely to choose vegetables, potatoes, sugar and especially rice and soybeans. In wetter locations, farmers are more likely to choose soybeans, oil crops, sugar, vegetables, cotton and especially rice, and they are less likely to choose potato, wheat and especially maize. The analysis of how Chinese farmers have adapted to current climate, provides insight into how they will likely adapt when climate changes. Future climate scenarios will cause farmers in China to want to reduce irrigation and shift crops towards oil crops, wheat, and especially cotton. In turn, farmers will shift away from potatoes, rice, vegetables, and soybeans. We find, however, that adaptation will likely vary from region to region. For example, irrigation is likely to fall in the eastern regions of China but increase in the western regions (provided there is sufficient water). One important weakness of this study is that it was not able to take into account water availability, which is a critical element of Chinese agriculture. The projections into the future take into account only climate change. The analysis does not take into account other background changes that may well occur including changes in prices, technology, and water availability. Analyses of climate impacts must take adaptation into account or they will overestimate damages. Policy makers must be aware that adaptation is an endogenous response to climate. They should anticipate that adaptation is important, that the magnitude of changes depends on the climate scenario, and that the desired changes depend on the location of each farm. 2 I. INTRODUCTION Although there is an extensive literature on the effects of climate on agriculture (Reilly et al 1996; McCarthy et al 2001), there are very few studies that have measured adaptation. Studies that compare the impacts of climate change that include adaptation, such as Ricardian studies (Mendelsohn, Nordhaus, and Shaw 1994, Mendelsohn and Dinar 1999; Mendelsohn et al. 2001; Mendelsohn and Dinar 2003; Seo et al. 2005; Kurukulasuriya et al. 2006; Fleischer et al 2007; Seo and Mendelsohn 2007, Wang et al. 2008), tend to find much lower damages than studies that do not include adaptation, such as agronomic analyses (Rosenzweig and Parry 2004; Reilly et al. 1996; McCarthy et al 2001; Parry et al. 2004). It is clear from this empirical evidence that it is very important to include adaptation in any impact analysis of long term climate change. Of course, adaptation is likely to be less important with respect to year to year weather fluctuations as farmers may have fewer options to adapt to sudden or abrupt changes. Adaptations are actions that people and firms take in response to climate change to reduce damages or increase benefits (IPCC 2007).1 What specifically do farmers do to adapt to climate? How have they adjusted to the climates that they live in today? A new series of studies have begun to examine this question. By comparing what farmers do in one climate zone versus another, the studies quantify how farmers have made long term adjustments to climate. For example, studies have examined how climate affects the choice of irrigation in Africa (Kurukulasuriya and Mendelsohn 2006a) and South America (Mendelsohn and Seo 2007). Studies have explored how climate affects livestock choice in Africa (Seo and Mendelsohn 2006) and South America (Seo and Mendelsohn 2007a). Previous studies have explored how climate alters crop choice in Africa (Kurukulasuriya and Mendelsohn 2006) and South America (Seo and Mendelsohn 2007b). All of the above mentioned adaptation studies find that farmers adjust irrigation practices, crop varieties, and livestock species to both temperature and precipitation levels. 1One should distinguish between adaptation to climate change, which spans over a long period, which differs from adaptation to climate variance, the changes in weather from year to year (Leary et al. 2006). Adapting to weather is important but it should not be confused with adapting to climate. 3 In the present analysis, we use the same cross sectional methods used in the above studies of Africa and South America to study farm adaptation in China. We expect that farmers in China have also adapted to the range of climates that they face in China. Analyzing a sample of 8,405 farms sampled across 28 provinces in one year, we estimate logit models of irrigation and multinomial logit regressions of crop choice to detect how these choices vary with long term temperature and precipitation. Matching the location of each farm to climate data and soils, it is possible to examine the effect of climate on these endogenous choices by farmers while controlling for several other factors. The available data allow us to measure the direct effects of temperature and precipitation on irrigation choice and crop choice. We specifically examine the choice of 9 major crops in China: wheat, rice, maize, soybean, potato, cotton, oil crops, sugar, and vegetables. Unfortunately, the amount of irrigation water a farmer uses is not available in the dataset. We do not know water availability or cost. If future climate scenarios reduce available water supplies, this is likely to affect these choices and the present study does not take this into account. This is an important omission for an agricultural system such as in China that relies heavily on irrigation. The paper is organized as follows. We briefly review the methodology of adaptation analysis in the next section. Section three discusses the available data and the construction of the variables in the data set. In the fourth section, we present the estimation results for current farmers. The fifth section then forecasts how future farmers would change their irrigation and crop choice for three different climate scenarios in 2050 and 2100. The paper concludes with a summary of the key results and a discussion of policy relevance. II. METHODOLOGY We assume that farmers make choices that maximize their income. We define income broadly to include both products they sell and consume. In this analysis, we are interested in modeling how they select from a number of discrete and mutually exclusive choices (McFadden 1981). In order to study irrigation choice, we rely on a dichotomous logit. We test how climate influences the probability of choosing whether to irrigate or 4 not, while controlling for a number of other independent variables such as soils, household characteristics and farm characteristics. In order to study crop choice, we rely on a multinomial logit regression. The multinomial logit examines the probability that a farmer chooses one of the 9 major crops grown in China. We do not examine minor crops. We assume that the choice among the 9 crops is independent of these other choices. The multinomial tests the influence of climate on the probability of choosing each crop controlling for a number of other independent variables such as soils, household characteristics, and farm characteristics. We model irrigation choice and crop choice separately. We assume that farmers choose the crop that yields the highest net profit. Hence, the probability that a crop is chosen depends on the profitability of that crop. We assume that farmer i's profit in choosing crop j (j=1, 2,..., J) is: ij =Vj(Ci,Ki,Si)+ (Ci,Ki,Si) (1) j where C is a vector of climate variables, K is a vector of exogenous characteristics of the farm, and S is a vector of characteristics of the farmer. The vector K includes soils, elevation and access variables; S includes variables such as the education of the farmer and land size. The profit function is composed of two components: the observable component V and an unobservable component that is in the error term . The farmer will choose the crop that yields the highest profit. Similarly, the farmer also chooses irrigation or rainfed farming based on which type of farming yields the highest income. The farmer will choose crop j over all other crops k (jk) if: (Zi) >k (Zi)fork j.[orif k(Zi)- (Zi)