Peasant Risk Aversion and Allocative Behavior: A Quadratic Programming Experiment

A quadratic risk programming model is used to examine the impact of yield uncertainty on peasant allocation of land among crops and use of hired factor services. The assumption of an exponential utility of income function permits sample estimation of the extent of risk aversion and interpretation of the dual solutions as shadow prices. Historical survey data on a Chinese village are used to show that optimization qualified by risk aversion proves superior to risk neutrality or credit constraints in explaining peasant allocative behavior. Key words: China, farm planning, quadratic programming, risk aversion. Quadratic risk programming is commonly considered a theoretically appealing technique for analyzing the impact of risk aversion on farm planning in terms of the trade-off be-tween the first and second moments of stochastic outcomes of farm plans (Hazell, Rae, Freund). It is often recognized that use of the expected income-variance (E-V) crite-rion of optimality is equivalent to maximiza-tion of the expected value of an exponential utility-of-income function if income is nor-mally distributed (Freund). However, it is rarely recognized that in this case the dual solution of the quadratic program has an in-terpretation analogous to the linear case, and a criterion emerges for measuring the extent of farmer risk aversion from readily accessible data. In reaction to the limited success of pro-grams in certain less developed countries in-tended to induce acceptance of technical changes among smallholders, a number of re-cent microstudies (Baker and Bhargava, Wol-gin, Schluter, Roumasset) have examined the relative importance of subjective factors such as risk aversion or liquidity requirements and objective factors such as credit or input avail-ability on peasant decision making, arriving at contradictory conclusions. This study utilize