Simulating the Impacts of Credit Policy and Fertilizer Subsidy on Central Luzon Rice Farms, the Philippines: Comment
评论Rosegrant和Herdt关于菲律宾水稻产业的模型，指出其使用的安全优先决策规则不能真实反映农民的风险规避行为，并举例说明相同均值和标准差的不同分布有截然不同的财务含义。
Distribution (A) will be preferred to distribution (B) only by farmers who are risk seekers, although both distributions have the same mean (u = 1,0(0) and the same standard deviation (a = 3,000). The Journal article by Rosegrant and Herdt pre-sents an interesting and comprehensive dynamic and stochastic model of the Philippine rice industry. There is, however, one area in which it does not represent the real world. That is the risk aversion of farmers. The authors write, on p. 663, "Assuming a risk-averse, safety-first decision rule resulted in poorer model validation than the alternative risk-neutral, profit-maximizing behavioral rule. " Similar results also were reported by Roumasset. The problem, in my opinion, is with the safety-first decision rule. The form of this decision rule used in the article is Max d = u- ko, where u is the expected net income, and a is the standard devia-tion of the net income. This rule is criticized by Raiffa as "being terribly ad hoc " (p. 55). It is easy to give examples of two distributions with the same mean and standard deviation which have entirely different financial implications. Table 1 shows two such distributions (Raiffa, p. 56).