Elsevier

Journal of Urban Economics

Volume 79, January 2014, Pages 39-58
Journal of Urban Economics

Do labor market networks have an important spatial dimension?

https://doi.org/10.1016/j.jue.2013.03.001Get rights and content

Abstract

We test for evidence of spatial, residence-based labor market networks. Turnover is lower for workers more connected to their neighbors generally and more connected to neighbors of the same race or ethnic group. Both results are consistent with networks producing better job matches, while the latter could also reflect preferences for working with neighbors of the same race or ethnicity. For earnings, we find a robust positive effect of the overall residence-based network measure, whereas we usually find a negative effect of the same-group measure, suggesting that the overall network measure reflects productivity-enhancing positive network effects, while the same-group measure may capture a non-wage amenity.

Introduction

Research in labor economics has explored the potential for network connections among workers along a number of dimensions – including common military service, attending the same school, or coming from the same family. There appears to be a relatively common finding that workers who are connected to each other in some way that could plausibly result in them sharing labor market information also seem to have similarities in labor market outcomes, consistent with a role for labor market networks.

In a series of papers, two of us have explored a particular dimension of labor market networks – spatial labor market networks that connect workers who live in the same neighborhood. Using matched employer–employee data, Hellerstein et al. (2011) show that neighbors are more likely to work in the same establishments than would be predicted simply by the fact that neighbors are likely to work near where they live,1 and for minorities and especially Hispanic immigrants the clustering of neighbors in the same workplace is dramatic. In addition, the study finds evidence suggesting that these residence-based networks are racially stratified. In particular, blacks are much more clustered at work with their black neighbors than with their neighbors overall (i.e., without regard to race). These findings suggest that labor market connections among neighbors may be an important source of network connections in the labor market.

In this paper, we turn to different types of evidence on labor market networks to explore further the role of residence-based labor market networks. In particular, we study evidence of the productivity of these networks in terms of turnover and earnings. We draw on theoretical work (Dustmann et al., 2011, Brown et al., 2012) that derives implications of labor market networks that arise because network connections lead to better labor market matches. What is new, however, is that we test these predictions for residence-based networks.

This inquiry is useful for two reasons. First, if the evidence we have assembled in our prior work is really identifying labor market networks, then these theoretical implications for job matches should carry over to workers potentially connected via residence-based labor market networks. Second, although there is every reason to expect that there are various types of labor market networks, and workers may be connected to others through more than one type of network, it is nonetheless useful to try to identify the most important sources of labor market connections. Nothing in our evidence contrasts the importance of residence-based networks with networks along fundamentally different connections, like common military service. However, our evidence can help distinguish between networks based on place of residence – which therefore have an important spatial dimension – and networks based simply on common race or ethnicity.

Given racial and ethnic segregation of neighborhoods, residence-based labor market networks could give rise to evidence that looks simply like networks based on common race or ethnicity. For example, Dustmann et al. (2011) find that workers who work with a larger share of workers from their ethnic group have lower turnover and higher wages – consistent with the predictions of their model. However, it may be that this arises because workers from the same neighborhood work together and are likely to be of the same ethnic group. Thus, looking explicitly at the role of residence-based labor market networks, and contrasting this with the role of networks that may be based solely on common race or ethnicity, can help us pin down the spatial nature of labor market networks.

Identifying an important spatial dimension of labor market networks is potentially significant for a number of reasons. First, if policy is to try to leverage labor market networks to get multiplier effects,2 then policymakers have to know which connections among workers are productive. Second, some evidence that is consistent with the existence of networks is alternatively consistent with the existence of discrimination, and distinguishing between the two is obviously important. For example, evidence like that in Dustmann et al. (2011) – showing lower turnover and higher wages when one works with more co-ethnics – could stem from labor market discrimination, if employers with a large share of an ethnic group treat workers from that ethnic group better. However, evidence of residence-based networks – which can give rise to the type of findings Dustmann et al. generate – is harder to explain as stemming from discrimination, and in that sense can give us more solid evidence on labor market networks.

And third, establishing that residence-based networks are important in determining labor market outcomes provides new perspectives on how to think about the interrelations between space – in particular, where people live – and the labor market. These issues are central to questions at the intersection of urban economics and labor economics. Residence-based labor networks can, for example, help explain how ethnic and racial residential segregation reinforces poorer labor market outcomes for minorities. But they can also potentially lead us to think about how to increase labor market connections among neighbors that might help offset some of these disadvantages – as may happen for Hispanic immigrants who often live in highly-segregated ethnic enclaves.

Section snippets

Research on the presence and nature of labor market networks

A growing body of evidence in labor economics points to the importance of labor market networks. Earlier evidence consists largely of survey findings indicating widespread reliance on friends, relatives, and acquaintances to find jobs. (See Ioannides and Datcher Loury, 2004, and more recent evidence for European countries in Pellizzari, 2010.) More recent research has studied and typically documented the similarity of employment outcomes for Veterans who served together in World War I (

Data

This project uses employer–employee matched data to construct measures of network ties and estimates of the impact of employment networks. The core data is a set of infrastructure files produced by the Census Bureau’s LEHD program (see Abowd et al., 2009). Previous research on residential employment networks in Hellerstein and Neumark (2003) used the 2000 DEED, based on matching 2000 Long-Form Census respondents from the “Sample Edited Detail File” (SEDF) to their establishment of employment.

Methods

To provide a point of reference between the LEHD data used in this paper, and the DEED data used in earlier research, we begin by re-computing the measures of the importance of labor market networks used in Hellerstein et al. (2011), but using the LEHD data instead.

For a population of workers, we first compute for each worker the share of co-workers (in the same establishment) who live in the same residential neighborhood as that worker.

Methods

Our main analysis concerns the effects of residence-based networks on labor market outcomes – specifically turnover and earnings. In the regression analysis of the effects of networks, we use measures related to those that go into our “effective network isolation index” defined in Eq. (3), but there are some differences. For our analysis of the impact of networks on turnover and earnings we measure the extent of labor market networks for worker i at employer e in year t as NIietO, with network

Conclusions and discussion

If labor market networks lead to better matches in the job market, they should reduce turnover and increase wages. We use matched employer–employee data with detailed information on where people live and where they work to test this hypothesis in the context of residence-based labor market networks. Recent research has suggested that these types of labor market networks may be present, but it has not assessed any evidence asking whether these networks are productive in terms of improving labor

References (35)

  • Julie A. Kmec et al.

    Does it pay to have a network contact? Social network ties, workplace racial context, and pay outcomes

    Social Science Research

    (2009)
  • Abowd, John M., Stephens, Bryce E., Vilhuber, Lars, Andersson, Fredrik, McKinney, Kevin L., Roemer, Marc, Woodcock,...
  • John Abowd et al.

    The New National Individual Characteristics File (ICFv4) and Quarterly Workforce Indicator (QWI) Tabulations by Worker Race, Ethnicity, and Education. Technical Memo

    (2011)
  • Andersson, Fredrik, Garcia-Perez, Monica, Haltiwanger, John, McCue, Kristin, Sanders, Seth, 2010. Workplace...
  • Åslund, Olof, Hensvik, Lena, Nordström Skans, Oskar, 2009. Seeking Similarity: How Immigrants and Natives Manage at the...
  • Patrick Bayer et al.

    Place of work and place of residence: informal hiring networks and labor market outcomes

    Journal of Political Economy

    (2008)
  • Dan A. Black

    Discrimination in an equilibrium search model

    Journal of Labor Economics

    (1995)
  • Brown, Meta, Setren, Elizabeth, Topa, Giorgio, 2012. Do Informal Referrals Lead to Better Matches? Evidence from a...
  • Thomas H. Cormen et al.

    Introduction to Algorithms

    (2001)
  • Federico Cingano et al.

    People I know: job search and social networks

    Journal of Labor Economics

    (2012)
  • Damm, Anna Piil, 2012. Neighborhood Quality and Labor Market Outcomes: Evidence from Quasi-Random Neighborhood...
  • Linda Datcher

    The impact of informal networks on quit behavior

    Review of Economics and Statistics

    (1983)
  • Dustmann, Christian, Glitz, Albrecht, Schönberg, Uta, 2011. Referral-Based Job Search Networks. IZA Discussion Paper...
  • Per-Anders Edin et al.

    Ethnic enclaves and the economic success of immigrants – evidence from a natural experiment

    Quarterly Journal of Economics

    (2003)
  • Cynthia Estlund

    Working Together: How Workplace Bonds Strengthen a Diverse Democracy

    (2003)
  • Laura Giuliano et al.

    Manager race and the race of new hires

    Journal of Labor Economics

    (2009)
  • Mark S. Granovetter

    Getting a Job: A Study of Contacts and Careers

    (1974)
  • Cited by (64)

    • The structure and growth of ethnic neighborhoods

      2023, Journal of Urban Economics
    • Gender differences in job flexibility: Commutes and working hours after job loss

      2022, Journal of Urban Economics
      Citation Excerpt :

      The literature demonstrates the importance of a spatial network of local neighbourhood-level interactions for job search and referrals (Hellerstein et al., 2011; Schmutte, 2015). Moreover, Hellerstein et al. (2014) provide evidence of spatial labour market networks, showing these networks affect turnover and earnings. In a different context, Rosenthal and Strange (2012) show that female entrepreneurs make firm location decisions such that they are more segregated and located farther away from dense areas and benefit less from agglomeration economies.

    • Geographic distribution of firms and expected stock returns

      2021, Journal of Economic Dynamics and Control
    View all citing articles on Scopus

    We are grateful to Gilles Duranton, Kristin McCue, Erika McEntarfer, Henry Overman, and Giorgio Topa for helpful comments. Any opinions and conclusions expressed herein are those of the author(s) and do not necessarily represent the views of the U.S. Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.

    View full text