اعتبار کشاورزی و کشش تقاضای اعتبار در شاآنشی و گانسو

This paper empirically estimates individual household credit demand elasticities based on 897 farm households surveyed in Shaanxi and Gansu provinces in the People's Republic of China (PRC) in October 2009. We used survey-based experimental techniques to extract individual household credit demand functions from which we estimated point demand elasticities. From a theoretical point of view, we proposed that as interest rates fell the demand for credit increased in elasticity, and this appears to hold in our data. We find a range of elasticities with mean point estimates of about − 0.6. We find that nearly 20% of farm households have nearly perfectly inelastic demands for credit but we also find that nearly 20% have elasticities above − 0.75 including some 15% that have elasticities greater than − 1.0. Previous studies that have argued against credit policies because of the low inelasticity of demand do not generally hold. There is much heterogeneity in credit demand and we would argue that a full spectrum of targeted credit policies can be used to address differences across farms.

Perhaps the most challenging problem facing farmers in the People's Republic of China (PRC) is access to credit. There are three key issues. First is the role of public policy and intervention in rural credit markets, second is the willingness of rural lenders to make loans to farmers under current risk conditions, and third is the actual demand for credit. It is this latter issue which is of primary concern to this paper. Understanding the demand for credit is a prerequisite to setting either credit policies or setting a path for rural credit reform. The key objective of the CBRC as it pertains to agriculture is to promote the development of financial services and increase competition in rural areas. Ultimately the goal is to match supply to demand and policies instituted to increase supply would be highly effective if aggregate demand was highly elastic and less so if highly inelastic. Likewise, if credit supply is chronically low then knowledge of credit demand and credit demand elasticities would be of limited value since setting policies through demand driven policies would be ineffectual. China's rural credit goals are based on both supply and demand driven processes. To accomplish these goals, the CBRC adjusted and relaxed its market-entry policy for rural banking institutions in December 2006. New types of rural financial institutions including village or township banks, lending-only companies, and rural mutual credit cooperatives are now permitted. These new institutions are contributing to the diversification of rural finance by providing wider coverage in the central and western regions. The CBRC Annual Report 2008 identified a “catfish effect,” wherein weaker financial institutions were responding to the new competitive realities with efficiencies realized in pilot areas. The policy objective is to ensure all counties and towns have access to financial services by 2012 through the establishment of new-type rural financial institutions, mobilizing these and existing institutions to open new branches, extend services, or offer mobile services. Much has changed since reforms were instituted in 2003. A host of new financial institutions have been registered and allowed to lend. For example in 2008, 91 new village banks were opened, 6 new lending companies were started, and 10 rural mutual credit cooperatives were initiated for a total of 109 new institutions (China Banking Regulatory Commission (CBRC) Annual Report 2008). The growth in village and township banks has been substantial. As of late 2011, some 570 VTBs had been licensed, while a further 103 preliminary approvals had been given by CBRC (i.e. for preparation to become licensed). Approximately one-third of these approvals were given in 2011 alone. Geographically, 196 institutions were approved in the Western region, 207 in the Central region, and 270 in the Eastern region (the numbers do not include a separate count of branches). The great majority of VTBs (84%) were sponsored by commercial banks – mainly city commercial banks (50%) and rural commercial banks (19%), with state-owned, joint-stock, and cooperative institutions (Rural Credit Cooperatives and Rural Cooperative Banks – respectively, RCCs and RCBs) also represented. VTBs are set up generally at county level, with a smaller number at township level, in addition to branches. Major international banks including HSBC, Citi, and StanChar have invested in rural finance provision through VTBs and other means.1 China has many means to affect credit demand including monetary policy and interest rate subsidies. According to Notice of the Peoples Bank of China on Adjusting Financial Institutions Deposit and loan Interest Rates, for RCCs, there is a ceiling loan interest rate of 2.3 times the base rate, and a floor that is 0.9 times the base rate. At the time of writing (January 2012) the base interest rate is 6.56% so that the range of interest that can be charged farmers is between 5.904% and 15.088%. RCC's dominate agricultural finance in rural China, but for commercial banks, there is no ceiling. With such a wide range of interest rates chargeable to farmers through the RCCs , Postal Savings, village banks (VTB), and non-deposit micro-credit companies (MCC) it is important to understand how farmers will respond in their demand for credit. To motivate this paper we illustrate how our results can be applied in a policy context. First, the base interest rates between 2009 when we first collected the data and the time of writing has increased by 23.50% as a monetary response to reducing credit demand.2 How do such policies affect agricultural loans? Assuming that the increase is passed onto farmers we multiply the mid-point elasticities in Fig. 5 by their respective frequencies and find a weighted elasticity of 0.598 (assuming a maximum elasticity of 1.3750). Multiplying this by − 0.2350 implies a weighted reduction in loan demand of 14.05% or CNY 140,530 for every CNY 1,000,000 in initial loan demand. From Fig. 5, 19.8% of households would reduce loan demand by only 2.94% for the lowest elasticity and 32.31% for the highest. In other words for every 1,000,000 lent to low elasticity borrowers representing 19.80% of the population the net reduction would be in the neighborhood of only CNY 29,375, but for the 13.4% facing an elastic demand their reduction in demand would be in the region of CNY 323,100.3 The average reported interest rate charged by RCC to our sample of farmers was actually 10.8%.4 As a second example suppose that the PBC agreed to a 3% subsidy to boost demand in rural areas. The farmers in our sample reported a total of CNY 6,504,911 in RCC loans outstanding. A 3% subsidy on 10.8% implies a 27.78% reduction in the interest rate. The weighted average elasticity is 16.6%. Thus we would estimate that credit demand would increase by CNY 1,080,538. For simplicity assume that the loans are evenly distributed across farms so that every farm had an average of CNY 21,611. The lowest elasticity group , comprising 19.8% of farmers would increase demand by only 3.47% or CNY 750.38, while the most elastic group, representing 13.4% would increase loan demand by 38.19% or by CNY 8,254.20 for a final loan of CNY 29,865.2. As these two examples show, understanding how credit demand elasticities are distributed across farm populations a much clearer picture can be inferred with regards to farmers' response to credit policies. With the economic problems faced by the Government of the People's Republic of China, the ministries of finance and agriculture, and the CBRC in matching demand to supply, regulating the institutions serving agriculture, and considering how best to use credit policy to achieve related goals of food security, rural income growth, and closing the income gap between rural and urban households, the type of micro-analysis presented in this paper is critically important. Resolving the economic problem requires as a first step an understanding of the relationship between interest rates and loan quantity. In addressing this economic problem this paper provides pro forma estimates of the demand elasticities for formal farm credit in Shaanxi and Gansu. Shaanxi and Gansu are particularly well suited for investigation because they are relatively poor areas in western China, and less developed than the agricultural economies found in Eastern China. With limited opportunities for non-farm enterprise, western China allows us to investigate farm credit demand with less influence from non-farm investments that drive credit demand in other parts of China such as the Yangtze River and Pearl River Delta areas. Farmers in Gansu in particular are very poor, while the areas we investigated in Shaanxi, while generally poor, had more varied agriculture. Between the two provinces, we are able to investigate credit demand across a range of agricultural crops and income classes. In general the literature on credit demand elasticities is scant and with the exception of a single paper by Weersink, Vanden Gungen, and Turvey (1994), we are unaware of any other paper that attempts to estimate credit demand elasticities for agriculture. Furthermore, we are unaware of any other study that goes beyond aggregation of data to extract demand elasticities at the individual farm level. Based on a specially designed survey of 897 farm households in Shaanxi and Gansu provinces, we derive the credit demand elasticity for each household and report the frequency of these elasticities. The novelty of our approach is in the design of an experimental demand supplement contained within a broader survey of agricultural and household credit and risk. Our experiment takes into account two factors raised by Feder, Lau, Lin, and Luo (1990). First, because we calculate the demand elasticity for each household we make no assumption of homogenous preferences. In fact, we find a range of credit elasticities from perfectly inelastic to elastic. Second, we make no prior assumption about whether the farm households are actually in an equilibrium that equates supply to demand. In fact, a strength of our approach is that it allows for disequilibrium—we include farm households whether they have formal debt or not, or whether the amount of debt held satiates demand. We then run regressions with these elasticities as the dependent variable and include independent variables that capture risk. We find only tepid support for the risk effect, but cannot reject it outright. Factors such as demography, culture, and attitudes and attributes of the rural credit delivery system are much stronger influences. In the next section, we review the literature on credit demand elasticity measurement. We then describe the experimental protocol used to derive individual household demand elasticities. This is followed by a summary of the experiment results and regressions (Tweedie and MLE) designed to investigate how factors gathered in the main survey affect credit demand elasticities. The paper then concludes.

As identified in the Introduction, there has been a backlash against using credit policy to encourage agricultural productivity in developing countries. The two main arguments that support this position are that demand elasticities are highly inelastic so that interest rate policies would generally be ineffective in spurring demand, and the second is institutional incompetence. The current state of the development literature and indeed the general literature on financial economics provide little empirical support for the first argument. The historical support for the second argument is better documented but perhaps a broader view or a precise definition of “credit policy” is required. This paper primarily tackles the problem from the point of view of the borrower and the general inelasticity of demand. Based on 897 farm households surveyed in Shaanxi and Gansu provinces in the PRC in October 2009, we used experimental techniques to extract individual household credit demand functions from which we estimated point demand elasticities. From a theoretical point of view, we proposed that as interest rates fell, the demand for credit increased in elasticity. This appears to hold in our data. The importance of this is that policy cannot assume as a matter of course that the elasticity is constant along the demand curve. In general, as interest rates are lowered, farmers will become more responsive to using credit. We say this as a matter of fact and not as a statement necessarily in favor of interest rate subsidies because this paper did not provide any measures of the general welfare benefits to such policies. Our results imply that one cannot generally assume that the demand elasticity is everywhere highly inelastic for all farms. We find a range of elasticities with mean point estimates of about − 0.6. We find that nearly 20% of farm households have nearly perfectly inelastic demands for credit but we also find that nearly 20% have elasticities above − 0.75, including some 15% that have elasticities greater than − 1.0. Previous studies that have argued against credit policies because of the low inelasticity of demand do not generally hold. There is much heterogeneity in credit demand and we would argue that a full spectrum of targeted credit policies can be used to address differences across farms. We provide GLM regressions using Tweedie and linear linked MLE to identify factors affecting demand elasticities. The Tweedie regression was used to account for the large number of farmers with zero elasticity. The two regressions were consistent in the interpretation but the Tweedie regression in all cases showed larger marginal responses. We find that the type of agriculture has no real bearing on the distribution of elasticities. Thus, we would not recommend that credit policies be targeted to any particular form of husbandry. We do find that farm profits and risk have a statistical effect on elasticity. In general, farms of higher risk as measured by the standard deviation of principal crop revenue will have lower demand elasticity, while farms with mean revenues will have more elastic demand elasticities. However, we do not find any statistical indication of a relationship between suggested acceptance of risk and the demand elasticity. We find that farms with higher savings rates have more inelastic demand, confirming our proposition that high savers substitute savings for credit while low to moderate savers may view borrowing and savings as complementary activities to liquidity preferences. We do find that farmers willing to obtain more credit, even at higher interest rates, have a more elastic demand, and as expected, farmers who would be willing to borrow more if interest rates fell would also have a more elastic demand. We find that current informal or formal indebtedness has a positive influence on the elasticity. We also find some interesting cultural indicators of loan demand. Farmers who are older tend to have higher elasticities as do more educated farmers. But this is offset by farmers who have been farming for a long time in what appears to be a statistical contradiction. We find that farmers brought up to avoid debt or simply prefer not to borrow from a bank have lower elasticities. We also find that a favorable opinion of the bank including flexibility in products and services has a positive effect on elasticity. In this context, credit policy may best be established as a marketing effort. To encourage greater use of debt for the purchase of productive assets, rural credit cooperatives and banks may do well by simply marketing and refining existing services to better meet the farmers' needs. This would increase the responsiveness of many farmers along the demand curve without having to resort to interest rate policies at all. We included the control variables in the regressions. In all cases, the four controls (county and/or province, amount of debt, direction of interest rates, and duration) were not statistically different from zero. Of these, the duration effect is most interesting. Karlan and Zinman (2008) found a duration effect in their elasticity measures and argued that loans of longer duration were more elastic than those of shorter duration. To examine this effect, we included 1-, 2-, and 3-year loans in our field experiment but find no evidence of a duration elasticity. That is, a farmer offered a 1-year loan is just as likely to have a high or low elasticity as one who is offered a 3-year loan. Much follow up work can be done. Our finding of more elastic demand for consumption relative to production suggests that credit policies may not have the production impacts as perhaps stated in this paper. Credit is fungible, and there may be a pass through of production credit to consumption. Whether fungibility is of economic significance depends on whether savings would be routed to consumption in any case, in which situation credit policy would be effective. Also, if the findings of broad heterogeneity in this study are found to hold elsewhere it suggests that responses to a general interest rate policy in agriculture will be hard to predict and could produce unexpected results.7 Further research should evaluate the impact of credit and credit policies on agricultural production and consumption patterns.