Most analyses of the principal-agent problem assume that the principal chooses an incentive scheme to maximize expected utility subject to the agent's utility being at a stationary point.An important paper of Mirrlees has shown that this approach is generally invalid.We present an alternative procedure.If the agent's preferences over income lotteries are independent of action, we show that the optimal way of implementing an action by the agent can be found by solving a convex programming problem.We use this to characterize the optimal incentive scheme and to analyze the determinants of the seriousness of an incentive problem.'Support from the U.K.

A LTHOUGH public policy analysts are revaluating income maintenance and income support programs, economists have devoted little attention to empirical research on one such program. Workers' Compensation (WC), a program in search of quantitative researchers, is about the same size as the Unemployment Insurance (UI) and Social Security Disability Insurance (SSDI) programs, and WC may have stronger supply effects than the UI program.' The state and federal no-fault insurance programs which constitute America's WC insurance system cost over $25 billion in 1980, and they covered, approximately, 90% of all wage and salary workers.2 During the period 1972 to 1978, the cost of WC as a percentage of covered payroll doubled and was probably equal to 2% of covered payroll.3 The WC program has enjoyed the support of both labor and industry. Employers favor this form of no-fault insurance because it guarantees a limit on the liabilities that they will incur due to the work-related injuries and diseases of their employees, and employees value the guaranteed medical expenses and payments that they receive under the program.4 Labor can view Workers' Compensation as a vast improvement over either the common law, which seemed to be designed to provide employees with strong safety incentives rather than to replace their lost income, or the employer liability laws that prevailed in most states until the early part of this century. Perhaps another reason for the position that the Workers' Compensation program has held in American social insurance has been that it is specialized in nature, and has constituted a relatively small share of the employers' overall cost. However, in recent years as the claim frequency under Workers' Compensation has risen dramatically and as policymakers and practitioners alike have consistently underestimated the cost consequences of liberalized Workers' Compensation benefits, analysts are beginning to reevaluate this very important form of social insurance. In this paper we analyze the two classes of Workers' Compensation injuries which account for most of the Workers' Compensation costs in the United States: temporary total and permanent partial injuries. In the next section we briefly describe some of the rudiments of the program after which we sketch an economic model of injury rates and suggest how they interact with wages and hours of work as levels of benefit change. In the fourth section of the paper we present empirical results which indicate that recent changes in the Workers' Compensation laws have had subReceived for publication April 1, 1982. Revision accepted for publication December 1, 1982. * Brigham Young University and Rutgers University, respectively. We wish to thank Steve Zrebiec for competent research assistance, and Monroe Berkowitz, Tom Brown, John F. Burton, Jr., Jennifer Field, and Fred Siskind for comments on an earlier draft. The views expressed herein are our own, and do not necessarily reflect those of Brigham Young University. ' Danziger, Haveman, and Plotnick (1981) guesstimate the reduction of work hours by transfer recipients as a percentage of total work hours of all workers as 1.2%, 0.7%, and 0.3% for SSDI, WC, and UI, respectively. 2 Dan Price's estimate (1981) that the WC program cost $20 billion in 1979 is a conservative one. He correctly attributes the full premium paid to private insurers, $14.3 billion, and to state funds to that year's cost, but he attributes only the benefits paid in 1979 by federal programs and firms that were selfinsured plus a 5%-to-10% markup for administrative cost to 1979 costs. This is equivalent to assuming that the federal programs and the firms which self-insure incur all of their WC losses during a calendar year. Actually, 1979 losses may be paid over many years, and injury or illness claims may arise many years after the end of calendar year 1979. These incurred losses and future claims should be fully reflected in current costs, but are so only to the extent that the actuarial price (premium) paid to private carriers and state funds is fully reflected in premiums collected. In addition, Price did not include the federal black lung benefits program funded by general revenues. 3Elson and Burton (1981) have examined the increasing trend in Workers' Compensation insurance. They present evidence which indicates that costs have doubled, for homogeneous classes of employers, in most states over the 1972 to 1978 period. 4There has been erosion of the certainty aspect of benefit payments due to litigation of claims. Vroman (1978) pointed out that certain permanent partial disability claims would be litigated with probability one. The high incidence of controversion has played a prominent role in calls for reform of the WC system and certainly was a factor in the state of Florida's decision to institute a wage loss system on August 1, 1979.

Commercial banks have been the subjects of a large body of empirical research employing regression and econometric models and discriminant analysis. The purpose of this paper is to empirically identify and describe relationships, including hedging behavior, between the asset side and the liability/capital side of the balance sheets of a cross-section of large U.S. banks. Canonical correlation analysis is the statistical technique that is employed. Unlike regression analysis which explains the behavior of a single dependent variable as a function of a set of independent variables, canonical correlation analysis relates two sets of variables. In the present case, one set of variables is the composition of the lefthand side of the balance sheet and the other set is the right-hand side. The variables used in this study are asset and liability/capital categories expressed as a proportion of total bank assets (i.e., a percentage breakdown of the balance sheet or a common size statement). These proportions are used in lieu of the more usual financial ratios and no information exogenous to the bank is employed.

Periods of high mortgage interest rates that characterized the early 1980s have escalated the use of assumption financing in home buying. In these periods, there is an increased demand for alternative sources of financing and a movement away from conventional sources. The purpose of this study is to provide a theoretical and empirical analysis of the effect of the increased use of assumption financing on selling prices of single-family residential housing. Aside from physical characteristics, the financing arranged by the homebuyer is considered a major influence in determining house prices. This is especially true in real estate appraisal practice where differences in types of financing are one of the major categories of adjustment (see [1], [4], and [10]).

This paper compares the optimal lending decisions of financial intermediaries that differ in their risk exposure. All intermediaries are assumed to face a loan demand described by a random applicant arrival process with each applicant offering a unique risk‐adjusted rate of return; loan demand is therefore uncertain in both quantity and quality. The intermediaries differ in terms of their risk exposure because of disparate funding practices. Intermediaries functioning as brokers minimize their exposure by borrowing funds only as demand is realized, whereas those behaving as asset‐transformers borrow in advance of realizing loan demand, thereby maintaining a loanable funds inventory and sustaining the related exposure. The optimal sequential lending policy is shown to involve setting a credit standard that becomes stricter with the length of the intermediary's planning horizon and the volume of loans outstanding. Most importantly, it is shown that brokers adopt stricter credit standards than asset‐transformers and therby reduce their volume of lending.

The financial intermediary's choice of operating as a broker with minimal risk exposure or as an asset‐transformer with interest rate risk is modeled as a funds inventory decision made prior to the resolution of uncertainty regarding the borrowing or lending interest rates. It is shown that an increase in the interest rate uncertainty leads the intermediary to reduce its exposure, thereby offering decreased asset‐transformation and more brokerage services. However, a stochastic increase in the interest rates leads to greater asset‐transformation and less brokerage services.