On-The-Job-Search And A Dynamic Analysis Of Labor Markets

Date of Award




Degree Name

Doctor of Philosophy (Ph.D.)




Most models of job search focus on developing optimal search rules when an individual is unemployed and searching for a new job. This study provides a theoretical and empirical framework for analyzing job search while either employed or not employed. The model considers On-The-Job-Search as a distinct labor market state and allows the occurrence of any random event, including the arrival of a job offer, even when an individual is not searching for a job. In order to analyze search decisions as well as employment decisions, the model permit transitions among four labor market states; not searching while not employed (traditionally called not in the labor force), searching while not employed (traditionally called unemployment), not searching while employed, and searching while employed.The econometric framework consists of a four state continuous-time hazard rate model, and the empirical analysis makes use of the baseline household survey of the Employment Opportunity Pilot Projects (EOPP) where continuous labor force history data are available for each earner in a total of 29,620 families that were interviewed.The study tests the hypothesis that the classifications employed and searching for a job and employed and not searching are behaviorally meaningless distinctions. The hypothesis is rejected suggesting that the task of building economic models that predict a distinction between employment with search and employment non-search is an empirically useful one. Another important finding is that employment with search typically follows a spell of unemployed search rather than a spell of employment without search. This finding casts some doubt on the traditional sequential job search models which assume search ceases when an acceptable job offer is found. When the transition rates are allowed to vary with personal characteristics and with time, the results indicate significant effects of many explanatory variables on the transition rates for each demographic group considered and negative duration dependence at least for the first 6 months of the spell duration.


Economics, Labor

Link to Full Text