A geometrical analysis, 540. — A numerical analysis, 547. — A model with two types of labor, 558. — Monopoly elements, 559. — Concluding remarks, 560.

If labor supply curves are not perfectly inelastic, wage discrimination induces some minority workers to leave the labor force. Studies of discrimination that focus only on wage differentials overlook these disincentive effects on minority employment. This article introduces a method of estimating the employment effects of wage discrimination and applies it to data on men and women from the 1984 Survey of Income and Program Participation. The authors find that wage discrimination against women caused a net loss of over four million jobs, supporting Gary Becker's contention that discriminatory employers use labor inefficiently.

The authors use three duration models to estimate the effects of disability benefits on the hazard of returning to work and on the expected duration of work absences. The results show that disincentives exist even when disability benefits are not conditioned on the recipient remaining out of work. In addition, blacks and women are found to be absent longer than white men. Durations of work absences are also influenced by available wages, the type and severity of injury, the physical demands of the jobs for which the worker is qualified, and the willingness of employers to help the worker return to work.

Traditional theory has usually condemned trade restriction as an irrational or misguided interference with efficient resource allocation. The two notable exceptions have been the optimum tariff and the infant industry argument for protection. More recently, however, trade theorists have begun to investigate the possibility that tariffs are used to achieve certain objectives, that is, ones reflecting considerations not captured by individual private consumption.' Their work has evaluated the relative efficiency of tariffs, production subsidies, and consumption taxes as instruments for implementing such objectives as a given increase in domestic production or restraint of imports below the free trade level.2 This note seeks to extend the analysis by comparing a tariff (or equivalent quota) with a commonly used protective arrangement which we term a distributed as alternative means of pursuing various objectives of trade restriction. We do not intend to de-emphasize the cost, even when minimized, of achieving certain objectives through protection. Rather, to the extent that these objectives are the end result of a process of political bargaining, they may be difficult to alter merely through demonstration of their welfare cost to the economy as a whole. Of course, calculation and promulgation of this cost makes it harder for special interest groups to obtain economic privileges at the expense of the majority. However, we should not overlook the possibility that certain noneconomic objectives are truly public goods which provide sufficient collective utility to the majority to justify their cost in terms of foregone consumption. In either case, it is desirable to provide the least-cost method of achieving the objective. For purposes of this paper, noneconomic can best be understood as outside the model. It is well known that any protective goal can be achieved through some combination of a consumption tax and a production subsidy. However, tariffs and quotas are employed much more frequently than direct subsidies probably for reasons of political feasibility. In this paper we use a familiar model to show that a proportionally distributed quota implies a smaller tax and larger subsidy than a tariff or conventional quota, and thus provides a more economically efficient-but equally practical-means of achieving some objectives. As long as markets remain competitive, import quotas distributed in most of the usual ways' are analytically equivalent to tariffs, since they will acquire a premium value in the market reflecting the domestic scarcity of the importable good.4 The only difference is in the distribution of the proceeds. In fact, the import licenses can be auctioned to the public, in which case (assuming again that markets remain competitive and in particular that the supply of licenses is not monopolized) the distribution of the proceeds is also equivalent, since the auction should yield the same revenue as would be produced by a tariff. We designate this type of restriction below as a tariff/quota. In contrast, the proportionally distributed quota is a system in which import rights are assigned to producers or consumers in proportion to their individual production or consumption of the restricted good. We show below that which group receives the licenses is irrelevant to the resulting equilibrium, for a given objective of protection (again, markets remnaining competitive). This is analogous to the fact that the economic incidence