به روز رسانی دوره ای از نیروی کار و در جستجوی کار

This paper examines patterns of worker reallocation in the search and matching model of the labor market. We show that on-the-job search is crucial for explaining the observed cyclical upgrading of workers to better employment opportunities in booms. This is due to the rising availability of employed searchers which facilitates recruitment for newly created high-wage jobs. The standard model fails to exhibit such behavior. At the same time, the model is consistent with salient features of labor market dynamics, such as the volatility of vacancies and unemployment, and a highly procyclical rate of job-to-job transitions. This suggests an important channel for the reallocation of workers across jobs as well as the propagation of aggregate shocks.

Search and matching theory, as outlined by Mortensen and Pissarides (1994) and Pissarides (1994) has become the dominant framework for the analysis of labor market dynamics over the business cycle.2 An initial focus on the cyclical behavior of job flows, such as job creation and destruction, has been met with reasonable success, as shown by Mortensen and Pissarides (1994), Cole and Rogerson (1999) and others. However, recent research has begun to uncover aspects of the data that the search and matching model is not able to reconcile without strong auxiliary assumptions. For instance, Hall (2005) and Shimer (2005) argue that the cyclicality of vacancies and unemployment can only be explained when assuming either implausibly large shocks to productivity or exogenously rigid wages. This paper introduces on-the-job search in a search and matching model to account for key patterns of labor reallocation over the business cycle. With workers in low-wage jobs searching while on the job, we find that high-wage sectors exhibit more cyclical employment growth.3 Apart from this new result, the model also generates strongly procyclical job-to-job worker flows, a dimension missing from many models. Consequently, while most new hires in a boom are in good jobs, the outflows from unemployment shift towards low-wage employment. Thus, there is a cyclical change in the composition of new employment opportunities for unemployed workers. The behavior of our model is in line with a number of empirical regularities on worker flows emphasized in the literature. It features a form of vacancy chain, since job-to-job quits induce creation of job vacancies (Akerlof et al., 1988 and Contini and Revelli, 1997). Hiring into new jobs and replacement hiring are strongly procyclical. As argued by Okun (1973), booms are associated with a relative larger supply of good jobs. Search on the job facilitates the reallocation of workers from bad to good jobs, and therefore the creation of good jobs in a boom. This point is also stressed by Mortensen (1994) and Mortensen and Pissarides (1999), and elaborated upon more recently by Barlevy (2002). Finally, workers that have been employed for a long time have lower quit rates, since they are more likely to have made the transition to a good job.4 In the model, job heterogeneity is introduced via two job types that differ in terms of profitability and thus wages. This offers a natural motivation for search on the job. Workers in low-wage (‘bad’) jobs search in order to gain employment in high-wage (‘good’) jobs. Good job vacancies can be matched with employed and unemployed job seekers, whereas firms in the bad job sector only hire unemployed workers. Wages are determined by surplus sharing for each matched job-worker unit and are continuously renegotiated. We calibrate the model to match salient long-run features of job and worker flows. The main reason for on-the-job search to play such an important role is that it accelerates the reallocation of workers across jobs. In a standard model, all job growth has to be fed from the pool of the unemployed, which is quickly exhausted in a boom. Instead, increasing search activity by the already employed opens a new channel of employment growth for new, higher value firms. The availability of employed searchers relaxes the economy's resource constraint: input into the production of new matches is rising since employed workers see an incentive for higher search intensity. In that sense, on-the-job search resembles variable factor utilization in the aggregate production function. The incentive to search more intensively is maintained throughout a boom, and the ability of the economy to allocate workers more effciently is thus enhanced. This paper also adds to a literature that explores the effciency of booms and recessions. Barlevy (2002) argues that on-the-job search is crucial for explaining mobility patterns in the labor market, and emphasizes its role for the effcient allocation of resources over the cycle. One view holds that recessions facilitate the reallocation from unproductive to more productive uses.5 However, this should imply that recessions are times when newly created jobs are the most productive. Barlevy cites evidence by Davis et al. (1996) that jobs created in recessions are of lower average quality and yield below average pay. He also argues that the fact that worker reallocation is procyclical indicates that the reallocation of resources to better uses may be impeded in recessions rather than in booms. Our results are consistent with Barlevy's, whose model features job heterogeneity and match quality heterogeneity. In our model, the main mechanism is the endogenous increase in good job supply induced by the rising availability of searchers. The key sectoral regularities that we refer to are discussed in McLaughlin and Bils (2001) and Aaronson and Christopher (2004), and others. The former show that there is cyclical upgrading of labor in U.S. manufacturing in that high wage jobs have more (pro-)cyclical employment. The latter report recent evidence on the cyclical changes in the composition of job quality, with most job growth during expansions occurring in high-wage industries, including the latest cycle up to 2004. Furthermore, Albaek and Sorensen (1998) report the reallocation of workers across employment opportunities in Denmark, which is largely from low-wage to high-wage jobs. They also document the importance of replacement hiring. The paper proceeds as follows. The next section develops the business cycle model with labor market frictions, and introduces on-the-job search. We discuss the calibration and solution of the model in Section 3, while Section 4 reports and interprets the results. Section 5 concludes.

We have developed a dynamic business cycle model of a frictional labor market with on-the-job search. Our analysis highlights the important role of on-the-job search for the cyclical reallocation of labor. Booms are times when relatively more high-quality employment opportunities become available for employed and unemployed workers. However, this is not necessarily enjoyed by unemployed workers, who face increasing competition for good jobs. Instead these workers shift their search activity to low-wage jobs, which in turn may offer them the chance of moving to high-wage jobs. A model without on-the-job search does not predict these patterns, but features parallel movements of employment in all sectors, and no change in the employment opportunities available to unemployed workers. In a companion paper (Krause and Lubik, 2004), we explore in detail the strong propagation of shocks to the aggregate economy induced by employed job search. The lack of amplification in the standard model is shown by Hall (2005) and Shimer (2005). In our model, the increased presence of searchers in the matching market reduces pressure to increase wages, and thus keeps the incentives for posting vacancies high. The real wage rigidity emphasized by Hall and Shimer as central to explain labor market dynamics thus arises endogenously with on-the-job search. While our model's key ingredient is objective heterogeneity in job productivities, the dynamics in a related paper by Nagypal (2004) are driven by workers' heterogenous subjective valuations of match quality. This induces increased job search in a boom as workers see improved opportunities for finding preferable jobs, amplifying shocks. Even though the model with on-the-job search explains important dimensions of the data, other aspects may have to be considered to give a full picture of the data. These are the endogeneity of the job destruction rate, labor force participation, and worker heterogeneity, which is likely to interact in interesting ways with job heterogeneity. From a business cycle perspective, inclusion of capital may have significant effects for the propagation of shocks. We plan to investigate these extensions in future work.