between gross job and workerflows. 3 A situation that arises naturally whenever agents can continue to search while matched, is one in which a matched agent contacts a new potential partner (who may also be matched) and each must decide whether to form a new match with the new partner or to stay with the old one. In a labor-market context, the employer who is trying to recruit an employed worker may have to face competition from the worker’s current employer, and in addition, the recruiting employer’s current employee may attempt to discour- age this employer from replacing him with the new worker (e.g., by accepting a smaller share of the matching surplus). Natural as they may seem, these generic situations have not been systematically analyzed in the literature. 4 One of the building blocks of our theory–and one of the contributions of this paper–is a simple and flexible noncooperative bargaining procedure that allows for competition among all parties taking part in such meetings. The equilibrium of the bargaining game we propose delivers the division of the gains from matching as well as privately eﬃcient creation and destruction of matches.
There is an increasing attention in the economic literature to the process of job creation and destruction (commonly referred to as gross jobflows) as well of hirings and separations (commonly referred to as gross workerflows). There is also an increasing interest for hiring and separations due to firm churning workers, that is in excess of job creation and destruction (e.g. Burgess et al., 2000). These processes appear to be at the heart of the functioning of market-based economies which entail a continuous process of creative destruction of jobs and job-matches: Each day, new firms start up; existing firms expand, contract and eventually shut down; individuals are hired to fill new positions or to replace previous employees on existing jobs; others quit or are dismissed. Within this context, many studies point to the importance of idiosyncratic firm-level characteristics, which appear to have similar effects in all countries, to explain both job and workerflows (e.g. Davis and Haltiwanger, 1999). Some studies argue that job-flow differences across countries are small while worker-flow differences are larger, and devote a lot of theoretical effort in explaining the disparity of patterns between job and workerflows (see for example Bertola and Rogerson, 1997, Pries and Rogerson, 2005). However, these statements are usually based on the comparison of national estimates, typically collected on the basis of different definitions and collection protocols. Cross-country comparable data on job and workerflows have been rarely used to describe cross- country differences (Haltiwanger et al., 2006, is one of the few exceptions) 1 and, as far as we know, no study have simultaneously used comparable data on both job and workerflows.
Like many transition economies, Slovenia is undergoing profound changes in the workings of the labor market with potentially greater flexibility in terms of both wage and employment adjustment. We investigate the impact of the changing labor market for Slovenia using unique longitudinal matched employer-employee data that permits measurement of employment transitions and wages for workers and links of the workers to the firms with whom they are employed. We can thus measure workerflows and jobflows in a comprehensive and integrated manner. We find a high pace of jobflows in Slovenia especially for young, small, private and foreign owned firms and for young, less educated workers. While jobflows have approached the rates observed in developed market economies, the excess of workerflows above jobflows is lower than that observed in market economies. A key factor in the patterns of the worker and jobflows is the determination of wages in Slovenia. A base wage schedule provides strict guidelines for minimum wages for different skill categories. However, firms are permitted to offer higher wages to an individual based upon the success of the worker and/or the firm. Our analysis shows that firms deviate from the base wage schedule significantly and that the
A new finding of the model is the negative correlation between the number of workers replaced and the number of jobs created in the second period. However, the limit of only one vacancy per period imposes this negative correlation, since firms are forced to choose only one of these actions. Our goal in the next section is to build a theoretical framework where firms may create new job positions and replace workers simultaneously. In addition, we show that the negative correlation between these variables still holds when we allow firms to choose freely the number of vacancies posted in each period. Moreover, this less restrictive environment also will provide richer implications concerning firm dynamics. Before moving to the next section we dedicate the remainder of this section to discussing two issues that were previously unaddressed: wage determination and model closure.
6 pools lead to a greater likelihood that individuals will seek opportunities to risk resources for increased resource gains (“gain spiral”) (Hobfoll and Shirom, 2000). In conclusion, Hobfoll promotes the opportunity of bringing in new resources, but also mentions that the gain of resources in itself has a modest effect, but acquires its saliency in the context of resource loss (Hobfoll, 2002). In the context of these economical harsh times, it is probable that most employees are confronted with loss of resources and it is especially important to bring in new resources or change the way people experience the loss of their resources, to prevent individuals from falling in to a loss spiral. In order to bring in new resources or change the way employees experience their loss of resources, it is fundamental to explore the resources available. To do this, I used a resource model with an extensive focus on resources in the context of working conditions, namely the Job Demands-Resources (JD-R) model, developed by Demerouti et al. (2001).
Economic growth provides only a partial explanation for regional differences in job and workerflows. According to the results, regional productivity, which also reflects the profitability of firms situated in a region, helps to explain a large part of differences in regional job and workerflows. The effect of productivity is found to be more pronounced in the case of jobflows than workerflows. The long-run difference between the lowest and the highest value of productivity between the regions and over time is estimated to be as large as 0.70 in job creation and 1.22 in the net rate of employment change. These figures are large but not totally out of line. The job creation rate may vary between 0 and 2, whereas the range of the net rate of employment change varies between –2 to 2. However, since the actual difference between the highest and the lowest value of the net rate of employment change is some 0.7 points, the magnitude of estimates has to be considered with some caution. 6 The estimates may pick up some unobservable factors that are not included in estimations.
A bstract: W hat can one infer about labor market ° ows from matched employer-employeepaneldata? T hepurposeofthis paperis tosketch possible answers tothis question.A generalbutsimplelabormarketequilibrium model ofhireandseparation° ows is developedinthepaper.T hemodelembodies the hypothesis thatworkerproductivity di®ers across employers and thatworker andemployer° ows re° ectresponses tothesedi®erences inalabormarketchar- acterized by friction.In themodeled market,each agentacts optimallytaking as given thewageo®erdistributionand markettightness and thesein turn are determined by theircollectiveaction.T heexistenceofalabormarketequilib- rium is establishedundertwodi®erentwagedeterminationmodels:rentsharing andwageposting.A demonstrationthatmarket° owparameters,searchandre- cruitinge®ortfunctions,andtheequilibrium wagedistributioncanbeestimated with matched job-work° owdatais theprincipalcontribution ofthepaper.
Seminal publications by Davis and Haltiwanger (1990, 1992) on gross jobflows statistics led the way to several analyses in the US and other countries. A sizeable literature has developed in the fields of labor and macro economics to measure and explain the various statistical and economic relations among the levels of gross job creation and destruction and workerflows. Better access to detailed microeconomic data has played a crucial role in this movement. One of the main results is the importance of idiosyncratic firm-level characteristics in explaining both job and workerflows (Davis and Haltiwanger, 1999). Studying the relative magnitudes of job and workerflows, Burgess, Lane and Stevens (2000, 2001) propose the useful notion of “churning flow” as the worker turnover in excess of jobflows. One main result identified by these authors is that churning is not randomly distributed across employers but is highly persistent in particular employers. They consider churning profiles as an equilibrium phenomenon that is associated with a particular set of optimal personnel policies. The literature yields a series of results regarding the correlation between wages and mobility rates.
To attempt to capture the differentially strong effect of the housing price bubble on the top group MSAs, we report the results of the MSA-level estimates of the respon- siveness of gross worker and jobflows to the local housing price index. By controlling for the national level of the labor market flow variable, national housing price move- ments, local and national labor market con- ditions, we can isolate the marginal contri- bution of the local HPI on the predicted flows. By allowing the effect to be hetero- geneous across MSAs, we allow for the pos- sibility that high-HPI MSAs had differen- tial responses to all of the control variables. The results are partially summarized in Ta- ble 1. For all four MSA-specific gross flow rates, the coefficient on the equivalent na- tional gross flow is essentially unity on av- erage, but with a substantial standard de- viation for the MSA-specific random com- ponent. In the case of gross workerflows, the random component has a standard de- viation of about 14 percentage points while for gross jobflows the standard deviation of the random component is about 25 per- centage points. Both of these estimates im- ply very substantial MSA-specific deviation in the gross flows. Online Appendix fig- ures C2 and C3 show that for all four gross flows, the estimated variation in the MSA- specific deviation from the national average is greatest for the top HPI group. That is, the most volatile local labor markets were those in which the housing price bubble was greatest.
different equilibria for different aggregate productivity levels. They also show that the amplification mechanism embedded in our model shows up clearly when considering the response of the unemployment and the excess reallocation flows to changes in aggregate productivity, a rather important character of actual labour markets, neglected by what has become the standard theory of equilibrium unemployment, i.e., the Mortensen-Pissarides model (Mortensen and Pissarides, 1994; Pissarides, 2000). Moreover our estimates show that excess reallocation varies negatively and the unemployment rate varies positively with job destruction shocks. Shimer (2005) argues that a destruction shock induces a positive correlation between excess reallocation and unemployment in the standard model, in stark contrast with the data. In our model, on the other hand, a higher destruction rate discourages excess turnover, because it shifts the composition of searchers towards the unemployed or people with lower expectations.
Two issues motivate the models. First, data coverage over the 1991-2000 period. Tariff and trade flow variables are available only up to 1998. This covers a large period but leaves the sharp 1999 devaluation out of the analysis, which is very important for manufacturing. So a model that does not require trade data would be required to provide some information on the 1999 devaluation on job and workerflows. Second, there is a concern when specifying a model with trade variables, such as the degree of openness (either measured by import penetration and/or export shares), tariffs and the aggregate exchange rate, is that these variables may be endogenous. For example, an increase in sector job destruction may induce an increase in tariffs in order to protect jobs (Muendler, 2002). At the same time, aggregate unmeasured shocks may affect both aggregate exchange rate movements and sector demand.
of b, the average replacement ratio over Þlled jobs in the base case is 75%. The elasticity of the unemployment stock with respect to replacement ratio predicted by our model is 6.0. Costain and Reiter (2003) argue that standard job matching models cannot be calibrated so as to match both the long-run response of unem- ployment to unemployment beneÞt and the business-cycle frequency volatility of unemployment. They show that introducing embodied technological change or sticky wages improve the model’s ability to match both types of stylised facts. An increase in the parameter a, which represents an increase in the extent of ranking between employed and unemployed job seekers, leads to lower un- employment and a virtually identical unemployment turnover rate. This seems counter-intuitive and is worth spending some time on. A higher value of a encourages more employed job search, and so leads to a higher quit rate and hence more vacancies. It also leads to a lower job destruction threshold and hence lower unemployment inßow rate. In this sense it functions like a fall in the cost of employed job search, which as noted above, reduces unemployment. This eﬀect cancels out the obvious direct impact on the chances of unemployed searchers from a greater disproportionate chance for employed searchers, so that the outßow rate is unchanged. The lower unemployment inßow rate reduces un- employment.
is reduced roughly 20%, relative to the benchmark model. While the reduction is substantial, readers may find it somewhat surprising that the model without the endogenous separation decision can generate such non trivial fluctuations in the job destruction rate. However, there are caveats in interpreting this result. First, this number is an upper-bound as mentioned above. In the model, firms are restricted from actively shedding workers. In other words, the result is based on the employment policy that is optimal only when endogenous job destruction imposes infinite costs on the firms. Second, shutting down endogenous separation shifts the balance of relative volatilities toward the job creation side, thereby making the model’s quantitative performance worse. Observe that the job creation rate is a lot more volatile than the job destruction rate in the model with exogenous separation only, implying that employment fluctuations are largely accounted for by the job creation rate. This is at odds with the empirical evidence. Furthermore, as can be seen in the last row of each panel, the model generates virtually no variations in the separation rate. 35 This is simply inconsistent with the facts about the separation rate.
JPOSRATIO Establishment’s ratio of quarterly temp to perm job creation flows 5.9067 23.6145 JNEGRATIO Establishment’s ratio of quarterly temp to perm job destruction flows 2.5828 13.4375 WPOSRATIO Establishment’s ratio of quarterly temp to perm hires 24.0556 112.7021 WNEGRATIO Establishment’s ratio of quarterly temp to perm separations 17.7020 150.2000 WNETRATIO Establishment’s ratio of quarterly temp to perm net employment flows 0.3078 26.3056 Establishment’s quarterly temp to perm contract conversion flows 1.4889 9.4991 Expectations for Next Quarter
Having presented our results we are now in a better position to compare our findings with those of other recent papers. In particular, the key papers that we note are Gal´ı, Smets, and Wouters (2011), Veracierto (2008), Christiano, Trabandt, and Walentin (2010), Shimer (2011), and Haefke and Reiter (2011). We think it is important to compare results both at a qualitative as well as quantitative level, since we think there are some robust qualitative findings in the literature. One key difference between our analysis and these other paper is that none of these other papers focuses on gross workerflows, either in steady state or over the cycle. We argue that this difference is likely to be critical. As we noted previ- ously, any model of participation is going to implicitly have a boundary that determines the participation and non-participation regions. In response to an aggregate shock, the size of the movements in participation and nonparticipation must surely be heavily influenced by the mass of individuals that are near the boundary. To us it seems unclear how one could be confident in having an empirically reasonable mass of people near the boundary without modeling the gross flows, since this is surely the single most relevant piece of information.
We ﬁnd that our benchmark model with shocks to frictions alone does a good job of accounting for the key features of ﬂuctuations in gross worker ﬂows between the three labor market states. We argue that the simulation results reﬂect some basic and intuitive economic forces present in a model of labor supply in the presence of frictions. These mechanisms actively involve the labor supply channel; even though the labor market participation rate displays limited and only weakly procyclical movements, the gross ﬂows in and out of not in the labor force (N ) into both employment (E) and unemployment (U ) are large, volatile, and show clear cyclical patterns. Although our benchmark model only has aggregate shocks to frictions, the presence of on-the-job search implicitly incorporates an endogenous, procyclical wage movement, as workers move up the job ladder more rapidly in good times. These endogenous procyclical movements in wages give rise to important labor supply eﬀects, so the labor supply channel is important in allowing the model to match the behavior of gross worker ﬂows.
Indeed, the above-used measure of the business cycle is quite naively chosen by the quarterly GDP growth rate. This procedure can be criticized because different measures could be more appropriate than GDP growth. Furthermore, it is often stated and argued that the labor market reacts with some delay to the business cycle. Therefore, it could be necessary to implement different alternative measures and a lag structure. However, this procedure still assumes that the influence of upswings and downswings on the transition rate is symmetric (the decrease/increase in an upswing is of the same size as the increase/decrease in a comparable downswing). Furthermore, it is assumed that the correlation stays the same over the whole period. Therefore, we implement an even more flexible structure to explore the issue of cyclical sensitivity more deeply. Year dummies for the period 1975–1992 for West Germany, and 1993–2004 for both parts of the country are included in the model. We treat the aggregate cycle as an unobservable, and estimate its impact to be τ t (t = 1, . . . , 30) for the typical worker.
As far as jobflows are concerned, Gomez-Salvador et al (2004) make use of the Amedeus database produced by Bureau van Dijk, containing comparable data at the firm level for European countries (and not at the establishment level) and covering all sectors except for the financial one. Amedeus covers firms that meet three basic criteria – they must have: an operating revenue greater or equal to 1 million euro; total assets greater or equal to 2 mil- lion euro; at least 10 employees (although for the UK, Germany, France and Italy these thresholds become 1.5 for revenues, 3 for assets and 15 for employees). This means that Amedeus underestimates the share of small firms.
The studies covering the U.S. have in common that they find job-to-job transitions to be procyclical (e.g. Petrongolo & Pissarides, 2001). By con- trast, the results of Burda and Wyplosz (1994) show that job-to-jobflows in France, Germany, and the UK are countercyclical. To address hetero- geneity in the cyclical sensitivity across different demographic groups, Kluve et al. (2009) employ a model that allows for heterogeneity in the cyclical dependence of labor market dynamics by means of cyclical loading factors. In their empirical implementation, they use the retrospective information of the German Socioeconomic Panel (SOEP) and the IAB employment subsam- ple. Their findings suggest that the re-employment rate, i.e. the transition from unemployment to employment, is the most decisive rate for differences in unemployment. Young workers experience more pronounced swings while women experience less pronounced swings in their re-employment rate. Addi- tionally to Kluve et al. (2009), this study also examines job-to-job transitions with respect to heterogeneous cyclical sensitivity.
Until 1999 the social security data contained information (predominantly) on permanent workers subject to social security. From 1999 onwards the data include records for other more marginal types of worker. Figure C1 shows that the hiring and separation rate is about 3.5 percentage points higher if we include all workers in the calculation as opposed to just permanent workers covered by social security. Figure C1 additionally shows that gaps in individuals’ social security record also increase the measure of hires and separations. A “gap” occurs if an individual works for establishment j in period t and at j in period t + k(k > 1) without an intervening spell of employment. If temporary layoffs are an important feature of the data, the inclusion of these gaps could make a difference. Including these gaps increases the measured hiring and separation rate by about 2.7 percentage points. In order to achieve a consistent series over the whole time period our measure of workerflows reported in the paper is based only on permanent workers covered by the social security system, and does not count a gap as a separation and hire. Note that none of these decisions changes our key conclusion as to the relationship between workerflows and jobflows.