Wage Dispersion, Job Vacancies and Job Search in Equilibrium

Abstract

In his survey, Models of Market Organization with Imperfect Information, Michael Rothschild paraphrases a colleague's views on the importance of imperfect information. friction caused by disequilibrium and lack of information accounts for variations in the numbers we observe at the fifth or sixth decimal place. Your stories are interesting but have no conceivable bearing on any question of practical economic interest (Rothschild, 1973, p. 1283). This same conviction led to the present paper. We simply did not believe that plausible imperfections in workers' market information would result in significant search activity, wage dispersion and job vacancies. Clearly, search would result if wage dispersion were great enough. But it seemed to us that, if firms responded rationally to workers' search behaviour, the resulting wage dispersion would not be great enough to induce significant search. We have constructed a model of the labour market in which the only imperfection is that workers do not have complete knowledge of the distribution of job offers. The equilibrium of the model is, in some cases, characterized by substantial search, wage dispersion and job vacancies. The equilibrium of the model diverges markedly from the corresponding equilibrium when information is perfect. One class of models which incorporates imperfect information deals with the search activity of workers who face a given distribution of wage offers and non-zero costs of search. Stigler's (1961, 1962) germinal articles fall into this class, as do most models of frictional unemployment (Gronau, 1971; McCall, 1970; Salop, 1973a). But individual searchers' imperfect information implies that employers will have some monopsony power. This insight gives rise to the second class of models, of which Mortensen's (1970) is the seminal one (see also Salop 1973b). Because individual firms are faced with a flow, supply of labour that is dependent upon their own wage rate, they must decide upon a wage policy: firms are not wage-takers. In such models workers' imperfect information may result in wage dispersion over firms. (The monopsony power and resulting wage dispersion are only transitory in Mortensen's model.) Our model is most closely related to Mortensen's. It differs from Mortensen's in that we systematically incorporate the stochastic nature of job search. The primary focus of this paper is on the question-could plausible imperfections in the information available to workers in an otherwise perfect world give rise to significant amounts of job search and wage dispersion in equilibrium?