Systematic job search and unemployment

A great deal of attention has been paid recently to the micro-economic foundations of macro-economic theory and in particular to the derivation of the Phillips relationships from a study of the underlying labour markets. Following Stigler [12, 13], the methodology of some of these recent papers3 has been to study the behaviour of a worker searching for a new job in a labour market characterized by uncertain wage-differentials, vacancies and qualifications. Faced with this uncertain environment, the individual calculates a subjective wage-offer probability distribution for the entire market from his expectations of wage-rates and vacancies and then samples the firms randomly, stopping his search and accepting employment when he is offered a position at a wage at least as great as the " acceptance wage " which he had calculated from the wage-offer distribution. This " acceptance wage " has the property that the marginal cost of further search equals (or exceeds) the marginal benefit. An important deficiency of this approach is the assumption that workers are unable to distinguish between firms ex ante. Only after sampling the wage-offer at a firm is the job seeker able to characterize that firm as a highor low-wage employer. The job seeker travels around the labour market, sampling a firm here and a firm there and consequently, unless new information causes a change in his expectations, never changing his level of acceptance.4 Furthermore, the possibility that the job seeker will be given no offer is not analyzed although this aspect of the problem has a number of interesting implications. In fact, individuals are able to distinguish among firms ex ante, and they sample specific firms in a systematic fashion rather than just sampling the job market in general. Consequently, at every moment of time a rational job seeker must select a firm to sample as well as an acceptance wage. Allowing the job seeker this additional information on specific firms in the market and the extra degree of freedom in his search strategy obviously makes him better off, that is, he would be willing to pay a fee to be able to search systematically rather than sample the market randomly. Furthermore, his maximizing behaviour changes as a result of this new information. It will be shown that his optimal acceptance level declines with his duration of unemployment as he samples his best opportunities first and poorer ones later. It is also interesting that the criterion derived for rating firms is not merely an expected wage expression but one which also includes the probability of acceptance and the rate of time preference as separate arguments in addition to the expected wage.