Estimating the Stock-Flow Matching Model Using Micro Data

Abstract

We estimate the stock‐flow matching model using micro‐level data from a well‐defined labor market. Using a dataset of complete labor‐market histories for both sides of the market, we estimate hazard functions for job‐seekers and vacancies. We find that the stock of new vacancies has a significant positive impact on the job‐seeker hazard, over and above that of the total stock of vacancies. There is an even stronger robust result for vacancy hazards. Thus we find evidence in favor of stock‐flow matching, even when controlling for unobserved search heterogeneity and stratifying into submarkets defined by location and occupation.

Active labour market policies in Germany: do regional labour markets benefit?

“…The coefficient is positive and significant as well as significantly greater than the effect for the low-unemployment regions. As Andrews et al (2013) show, the stock of new vacancies has an additional positive effect on the job-finding rate of the job-seekers. The vacancies are on average "older" in low-unemployment regions compared to high-unemployment regions, which might explain the differences between low-and high-unemployment regions.…”

Dynamics and endogeneity of firms’ recruitment behaviour

“…Other studies like Coles and Smith (1998), Gregg and Petrongolo (2005), and Coles and Petrongolo (2008) show that the hazard rate of vacancies registered with UK Job Centres is highest in the first month after the start of the search process. Kuo and Smith (2009) and Andrews et al (2013) use weekly data from UK Job Centres and show that the hazard rate of vacancies is highest in the first two weeks. If we calculate the weekly vacancy-filling hazard from using the standard piecewise-constant (weekly) estimation approach, we obtain a fluctuating hazard with no clear pattern (see Figure A1a) in the Appendix).…”

“…The evidence on the shape of the hazard is mixed. Using the standard approach of calculating the vacancy-filling hazard Coles and Smith (1998), Coles and Petrongolo (2008), Kuo and Smith (2009), or Andrews et al (2013) provide support for stock-flow matching by showing that the hazard rate is highest in the first two weeks after registering the vacancy with the UK Job Centre and drops sharply thereafter. Davis et al (2017) apply the standard approach to the German Job Vacancy Survey (the same data-set as we use) and find that the hazard is rather stable, first increasing a bit and then decreasing slightly.…”

“…More recent evidence on the time spend on searching for a job is provided by Krueger and Mueller (2012) and on the use of different search channels by Kuhn and Skuterud (2004) and Kuhn and Mansour (2014). Kuo and Smith (2009), Andrews et al (2013), and Kettemann et al (2017) show that the longer a worker is unemployed the higher the probability that she matches with a newly posted vacancy. This rejects random search in favour of the stock-flow-matching.…”

Education and the Labour Market

“…More recent evidence on the time spent on searching for a job is provided by Krueger and Mueller (2012) and on the use of different search channels by Kuhn and Skuterud (2004) and Kuhn and Mansour (2014). Kuo and Smith (2009), Andrews et al (2013), and Kettemann et al (2017) show that the longer a worker is unemployed the higher the probability that she matches with a newly posted vacancy. This rejects random search in favour of the stock-flow-matching.…”