Introduction Previous literature Directory UMM :Data Elmu:jurnal:L:Labour Economics:Vol7.Issue2.Mar2000:

Ž . Labour Economics 7 2000 181–201 www.elsevier.nlrlocatereconbase What effect does uncertainty have on the length of labor contracts? Kevin J. Murphy Department of Economics, Oakland UniÕersity, Rochester, MI 48309, USA Received 10 June 1998; accepted 24 June 1999 Abstract The literature on duration of explicit labor contracts has suggested that increased uncertainty should be associated with shorter labor contracts. More recently, it has been argued that the effect of uncertainty on contract duration depends on the type of uncertainty involved. Specifically, if the uncertainty pertains to aggregate real shocks, then contract durations should increase as workers seek to insure themselves against the repercussions of such shocks. Using a sample of 1876 labor contracts signed during the period 1977–1988, Ž this paper provides an empirical test of the foregoing hypothesis known as the efficient risk . sharing hypothesis . The paper presents results from estimation of a generalized-probit, simultaneous equation model, in which the dependent variables are contract length, indexation of the contract through a cost-of-living allowance, and the rate of wage change specified in the contract. The empirical findings confirm the efficient risk sharing hypothe- sis. q 2000 Elsevier Science B.V. All rights reserved. JEL classification: J50; C30 Keywords: Contract duration; Uncertainty; Efficient risk sharing

1. Introduction

The early literature on duration of explicit labor contracts suggests that increased economic uncertainty should be associated with shorter labor contracts Ž . Ž . Gray, 1978 . More recent work Danziger, 1988 , however, advances the notion Tel.: q1-248-370-3294; fax: q1-248-370-4275; e-mail: murphyoakland.edu 0927-5371r00r - see front matter q 2000 Elsevier Science B.V. All rights reserved. Ž . PII: S 0 9 2 7 - 5 3 7 1 9 9 0 0 0 3 8 - X that the effect of uncertainty on contract duration depends on the type of uncertainty involved. Specifically, nominal uncertainty should be associated with contracts of shorter length, but real uncertainty should be associated with contracts of greater duration. Using a sample of 1876 labor contracts signed during the period 1977–1988, this paper tests the Danziger hypothesis. The paper specifies several measures of economic uncertainty and incorporates these in a three-equation model in which the dependent variables are contract length, the rate of wage change specified in the contract, and the propensity to index against inflation via a cost-of-living Ž . clause in the contract a latent variable . The model is estimated using two Ž . techniques described by Amemiya 1978 for estimation of a simultaneous equa- tion generalized probit model. The remainder of the paper is structured as follows. Section 2 reviews previous theoretical and empirical work on the subject of contract duration as it relates to economic uncertainty. Section 3 presents the model of contract duration estimated in the empirical work. Section 4 relates the econometric details of estimating a simultaneous equation generalized probit model applied to the problem at hand. Section 5 discusses the empirical results. Section 6 closes the paper with a brief summary and conclusions.

2. Previous literature

The relationship of contract length and economic uncertainty was first ad- Ž . 1 dressed in the theoretical literature in a seminal paper by Gray 1978 . Her proposition is straightforward: contract negotiation is costly for both parties; therefore, the more costly the negotiations, the longer will the length of the contract tend to be as both sides wish to spread the costs of negotiation out over time. The reason that contracts do not last forever, however, is because uncertainty vis-a-vis variables of importance to the parties to the agreement limits the length of the labor contract. In other words, contingencies such as price inflation, changes in the structure of the firm’s product market, recessions, and so forth, can occur during the life of a contract. If such events were unforeseeable when the contract was negotiated, then the real wage will deviate from that which would equate demand and supply, and losses will accumulate due to inefficient production. If an increase in the degree of uncertainty present in the economic environment occurs, then the probability that unforeseen contingencies will arise during the course of the contract increases and the negotiating parties involved will wish to sign contracts of shorter length. 1 Although the issue was not at all new to the industrial relations literature at this point in time. See, Ž . Ž . for example, Stieber 1959 and Garbarino 1962 . Ž . Danziger 1988 develops the idea that labor contracts can insure workers against real shocks. In this light, workers are risk averse while firms are risk Ž . neutral or, at least, less risk averse . Over the life of a labor contract, real shocks to worker productivity can occur. Danziger argues that risk averse workers will wish to be insured against contemporaneous real shocks and will be willing to pay for such insurance by accepting lower wage growth than they would in the absence of such insurance. When the contract expires, workers are then exposed to the repercussions of the real shocks that transpired during the last contract. Danziger argues, therefore, that if a high degree of uncertainty exists regarding real shocks to the economy when a contract is negotiated, then the new contract will be commensurately longer to protect workers from the effects of such shocks. Danziger refers to this phenomenon as ‘‘efficient risk sharing’’ and he distin- guishes this from Gray’s hypothesis which he refers to as the ‘‘efficient produc- tion’’ hypothesis. Empirical work on the general topic of contract duration is not extensive. Studies that have examined the relationship between contract duration and uncer- tainty have focussed on the nominal uncertainty distinguished by Gray rather than on the real type of uncertainty distinguished by Danziger. Most studies of the relationship between contract length and nominal uncertainty, however, do con- firm the hypothesis that this type of uncertainty and duration are inversely related. Ž . The first article to examine this relationship was Christofides and Wilton 1983 . Using data on Canadian contracts spanning the years 1966–1975, they found, Ž consistent with Gray’s hypothesis, that nominal uncertainty as measured by the . variance of residuals from a distributed lag estimate of inflation was associated with reduced contract length, regardless of whether wages in the contract were indexed to the rate of inflation. Employing an alternative measure of nominal Ž . uncertainty based on the Livingston Index of Inflation Expectations , Vroman Ž . 1989 used a sample of contracts signed between 1958 and 1984 in the US manufacturing sector and also found that greater nominal uncertainty is associated with shorter contract length. Using a sample of Canadian contracts signed between Ž . 1978 and 1984, Christofides 1990 estimated a three-equation model allowing for interrelationships among contract duration, rate of wage change, and the elasticity of wage indexation. Though the primary focus of his study was not on the relationship between uncertainty and contract length, he did nevertheless find some evidence to support the hypothesis of a direct link between nominal Ž . uncertainty and contract duration. The study of Wallace and Blanco 1991 is the one that contradicts the theoretical prediction and the empirical findings regarding uncertainty and contract length. Their data set consisted of labor contracts signed in the US manufacturing sector, dating from 1968 to 1980. They found that nominal uncertainty has a significantly negative effect on contract length in the non-durable goods sector, but that it has a positive, though not significant, effect on duration in durable goods manufacturing. When they aggregated sectors they found that nominal uncertainty does not have a significant effect on contract length. Though all of the aforementioned studies test for a relationship between nominal uncertainty and contract length, none of the studies address the point made by Danziger concerning uncertainty regarding aggregate real shocks to the economy.

3. Theoretic underpinnings