FIRM SPECIFIC HUMAN CAPITAL AS AN EMPLOYER DISCIPLINE DEVICE

Abstract

The inability of employers to monitor perfectly the level of effort of their employees is a potentially serious impediment to labor market efficacy. Indeed, a number of recent studies have concluded that this may lead to involuntary unemployment (Shapiro and Stiglitz [1984], Sparks [1986]); an inefficient sectoral allocation of workers (Oi [1990], Strand [1986]); and discrimination against productively identical workers (Bulow and Summers [1986]). This paper shows that the lock-in effect of firm-specific human capital can help alleviate problems of worker moral hazard and thereby promote labor market performance. I. INTRODUCTION As the surveys of Akerlof & Yellen [1986, Ch. 1], Carmichael [1990], Lang and Kahn [1990], and Weiss [1990] amply illustrate, the last decade witnessed the emergence of a large body of literature which falls under the general rubric of wage models of the labor market. The hallmark of this approach is that some essential aspect of worker performance depends upon the wage paid by the firm. Much recent work--Bulow & Summers [1986], Oi [1990], Shapiro and Stiglitz [1984], Sparks [1986], and Strand [1986]-- has considered the case in which (due to monitoring difficulties) worker effort depends (positively) upon the wage. The popularity of the approach undoubtedly stems from its ability to apparently explain many stylized features of labor market behavior in an attractively simple and straightforward manner. In the papers cited above, Shapiro and Stiglitz [1984] and Sparks [1987] show that labor market equilibrium can be characterized by involuntary unemployment; Bulow and Summers [1986] demonstrate that, inter alia, efficiency wage models are consistent with equilibrium discrimination against productively identical agents; and finally, Oi [1990] and Strand [1986] show that efficiency wage payments can generate an inefficient sectoral allocation of workers. Yet, recent research now suggests that the results obtained from this class of model are highly sensitive to the specific assumptions made at the outset about the completeness of contractual arrangements. Indeed, as Carmichael [1985] astutely observed, if contracts can include bond payments (even if financed by borrowing from imperfect capital markets), the standard imperfect-monitoring variant of the efficiency wage hypothesis fails to predict the emergence of equilibrium involuntary unemployment (i.e., all workers receive their reservation utility). However, proponents of efficiency wage models have responded to this criticism by arguing that workers would be unwilling to post large bonds with firms due to a problem of employer moral hazard. Specifically, workers recognize that firms may hire them in order to steal their bond payment ex post. Once this limitation on admissible bond payments is recognized, it might then be tempting to conclude that the main results of the efficiency wage literature will re-emerge. In this paper I consider this issue, in an environment which also allows for investment in firm-specific human capital (the basic structure is similar to Hashimoto [1981]). Following Macleod and Malcomson [1987; 1989], I assume that: (1) worker performance measures are sufficiently nebulous to preclude verification by third parties and (2) as in Carmichael [1983], the courts are unable to ascertain whether, in the event of separation, a worker quit or was fired. This informational structure generates a potentially serious joint worker-employer moral hazard problem: workers may shirk and employers may simply hire labor in order to collect any bond payments or exit fees stipulated in the contract. As emphasized by both Carmichael [1989; 1990] and Macleod and Malcomson [1987; 1989], the solution to this contracting problem requires that both parties receive a non-negative return from the continuation of the match (relative to what can be obtained from its dissolution). …