Recovering Technologies That Account for Generalized Managerial Preferences: An Application to Non-Risks-Neutral Banks

The authors suggest that risk plays an important role in managerial production decisions. Managers make implicit and explicit decisions related to risk, return, and cost in setting target market, product, pricing and delivery decisions. Standard models of production and cost do not explicitly account for risk, assuming that managers are neutral toward risk. This simplification may undermine the model's usefulness when applied to an industry such as banking where risk plays an important economic role in the business. The standard model would label risk-averse banks at best, allocatively inefficient and at worst, technically inefficient. The authors attempt to maximize a managerial utility function, defined over profit, inputs, and outputs with respect to the mix of inputs and with respect to profit subject to the production constraint that the input mix must produce the given output vector. The solution to this utility maximization problem gives the manager's most preferred production plan. To the extent that managers have favored inputs whose employment they will increase at the expense of profit, the most preferred production plan will not be allocatively efficient. In fact, it may not even be technically efficient. The cost function that follows from the utility-maximizing production plan is sufficiently general to incorporate non-neutrality toward risk and to allow other managerial objectives in addition to profit maximization. Formulating the production plan from a model of constrained utility maximization suggests that the functional forms needed to implement the model can be derived by analogy to those of consumer theory. The Almost Ideal (AI) Demand System, adapted to accommodate generalized managerial preferences, yields input share equations and a profit (cost) function that, in the case of cost minimization, are identical to the translog cost function and input share equations. The model is estimated using 1989 and 1990 data from U.S. banks whose assets eq(This abstract was borrowed from another version of this item.)