The Distribution of Income: A Saddle Point Formulation
研究了收入分布的信息熵在给定经济约束下达到一阶平稳的条件，发现其是梯度为零且海森矩阵正定的最小值，与信息论中的最大值相反，并探讨了将其扩展为鞍点公式的可能性，对发展经济学有潜在应用。
An income distribution, for which Shannon's information entropy becomes first-order stationary under given economic side conditions, has been shown to represent a minimum with a vanishing gradient vector and positive definite Hessian matrix with respect to differentiation in the endogenous parameters of the distribution. This is contrary to the maximum encountered in information theory. It is analyzed whether the model could be extended to a full saddle point formulation by maximizing the information entropy with respect to differentiation in the exogenous side conditions. The distribution patterns generated suggest that the framework could find applications in development economics. Copyright 1991 by The editors of the Scandinavian Journal of Economics.