A dynamic model of job networking and social influences on employment

https://doi.org/10.1016/S0165-1889(03)00079-4Get rights and content

Abstract

This paper explores an economy in which personal connections facilitate job search. In the model, a firm receives information on the productivity of those job applicants with social ties to its current employees. In addition to providing a theory of networking, the model endogenously generates two classic theories in economic sociology. First, there is a highly non-linear relationship between average human capital in a group of socially connected individuals and the group's employment rate. Small changes in group composition may cause large changes in employment, as suggested in Wilson's ‘social isolation’ explanation for high unemployment rates among poor African-Americans. The model also supports Granovetter's ‘strength of weak ties’ hypothesis, which holds that acquaintances are more valuable job contacts than close friends.

Introduction

A line of empirical research going back to the 1930s indicates that social networks play an important role in job search. Bewley (1999) lists 24 studies between 1932 and 1990 that estimate the fraction of jobs obtained through friends or relatives, with most estimates between 30 and 60 percent. Despite this, the theory of job networking remains relatively underdeveloped. This paper analyzes a model economy in which networking arises because firms have limited information on the skills of job applicants. A firm's current employees provide information on the job-specific skills of their friends, thus improving the likelihood of a productive match. As a result of these information flows, a job searcher's employment probability and expected productivity are increasing in the number of employed workers among his or her friends.

In addition to providing a simple theory of networking's role in the labor market, the model has a number of implications for the behavior of employment rates within social groups (groups of individuals who share social ties, potentially defined by neighborhood, ethnicity, or other characteristics). First, the long-run employment rate of an individual is weakly increasing in the initial employment status, number of social ties, and human capital of any other member of his or her social group, and strictly increasing in his or her own initial employment status, social ties, and human capital. A second set of results characterizes the group-level relationship between human capital and long-run employment rates. There exists a critical level of human capital such that no group member is employed in the long run if all members are below that level, but the long-run employment rate is strictly positive if all members are above that level. The relationship is highly non-linear; near the critical value, small changes in group composition lead to large changes in employment. The third set of results describes the relationship between the social network itself and the long-run behavior of the employment rate. First, small groups without ties to the rest of society are at risk of falling into a zero-employment trap. Second, the structure of the social ties within the group matter. All else being equal, employment is increasing in the proportion of social ties that are ‘weak’ ties.

In addition to being of independent interest, these results formalize in an economic model two classic sociological theories: Granovetter's (1973) ‘strength of weak ties’ hypothesis and Wilson 1987, Wilson 1996 ‘social isolation’ theory of inner-city unemployment. The strength of weak ties hypothesis holds that acquaintances are more valuable in job search than close friends and family, because they provide a more diversified source of information. There is a sizable literature on sociology on the strength of weak ties. Montgomery (1992) uses a formal model to analyze its empirical implications, but assumes rather than derives a benefit to weak ties. In contrast, an aggregate version of the strength of weak ties arises endogenously in the model presented here.

The social isolation theory holds that poor access to job networks plays a key role in explaining high unemployment rates among low income African-Americans. Like the strength of weak ties, it has a long history in social science. In an early contribution, Kain (1968) argues that racial segregation leads to poorer job prospects for African-Americans, in part through their isolation from the job network. Almost three decades later, Wilson (1996) argues that growing isolation from job networks explains the significant decline in employment in low income African-American communities from the late 1960s to mid-1990s, and that this decline explains much of their contemporaneous growth in social problems. Wilson suggests that this is an unintended consequence of desegregation: with the end of formal segregation, the black middle class left traditional ghettos, depriving the remaining residents of critical contacts with employers. This small change in neighborhood composition led to a large deterioration in employment conditions through a vicious cycle of increased unemployment and further weakening of job networks. These dynamics arise endogenously in the model developed in this paper. Finally, I provide evidence supporting the social isolation theory: the cross-sectional relationship between average human capital and employment rates across US Census tracts exhibits a non-linearity very similar to that predicted by the model and by Wilson. While this empirical result does not in itself prove the social isolation theory is correct, the match between model and data is striking.

Taken as a whole, the model provides new insight into a number of ideas raised in other social sciences but previously unaddressed in economic theory.

Section snippets

Related literature

The theoretical literature in economics on job networking owes much to the contribution of Montgomery (1991). In his two-period model, firms lack information on the productivity of job candidates, but positive correlation in productivity between friends leads firms to offer higher wages to the friends of productive period-one employees. In more recent work (Calvó-Armengol and Zenou, 2002; Calvó-Armengol and Jackson, 2002), individuals randomly discover job vacancies, and currently employed

Description of the model

The model features a population of workers and firms. Productivity varies across worker–firm matches, and firms have limited information on the productivity of a particular match. Exogenous social ties between workers facilitate the transmission of additional information to firms. A firm's current employees provide references, or reports on the quality of the worker–firm match, for each of their friends. Social ties are thus valuable in job search.

Implications

This section describes the dynamic behavior of the efficient allocation. To facilitate this discussion, I introduce some additional notation. Let ni(t) be the employment status of worker i in period t:ni(t)≡1ifnif(t)>0forsomef,0otherwiseand let n(t)≡(n1(t),n2(t),…,nN(t)). For a finite number of workers, let n̄(t) be the employment rate:n̄(t)≡1Ni=1Nni(t).For an infinite number of workers n̄(t) must be defined as a probability limit.

The state of the economy at time t can be described by the pair

Incorporating non-networked job search

The basic model presented in previous sections is highly stylized. In particular, the assumptions guarantee that all jobs are obtained through networking and none are obtained through formal search. However, empirical research indicates that about half of jobs are obtained through networking and the other half are obtained through more formal methods. This section adds a formal job search component to the model. Because there is already a vast literature on formal search, the modeling here is

Conclusion and further directions

This paper has developed a model in which personal connections help transmit important information about the quality of a worker-job match. The resulting dynamics of the model have several interesting features. Long run employment is affected by the number of ties in the social network and the proportion of ties that are weak, as well as individual characteristics such as human capital and reservation wages. The relationship between any of these factors and long-run employment is highly

References (22)

  • D.T. Mortensen et al.

    Job reallocation, employment fluctuations, and unemployment

  • T.F. Bewley

    Why Wages Don't Fall During a Recession

    (1999)
  • Calvó-Armengol, A., Jackson, M.O., 2002. Social networks in determining employment and wages: patterns, dynamics, and...
  • Calvó-Armengol, A., Zenou, Y., 2002. Job matching, social network, and word-of-mouth communication. Working Paper 2797,...
  • L. Datcher

    The impact of informal networks on quit behavior

    Review of Economics and Statistics

    (1983)
  • R.I.M. Dunbar

    Grooming, Gossip, and the Evolution of Language

    (1996)
  • R. Durrett

    Oriented percolation in two dimensions

    Annals of Probability

    (1984)
  • Glaeser, E.L., 2000. The future of urban research: nonmarket interactions. Brookings-Wharton Papers on Urban Affairs,...
  • M. Granovetter

    The strength of weak ties

    American Journal of Sociology

    (1973)
  • M. Granovetter

    Getting a Job: A Study of Contacts and Careers

    (1995)
  • G. Grimmett

    Percolation

    (1989)
  • Cited by (12)

    • Referrals and information flow in networks increase discrimination: A laboratory experiment

      2018, Social Networks
      Citation Excerpt :

      If one of the groups has a better access to informal job search, this is detrimental for the other group, as in the case of referrals from the “old boys” network in a wide range of contexts (Rogers, 2000; McBrier, 2003; McDonald, 2011; Bian et al., 2015). Research concerning referrals highlighted how the hiring mechanism could enhance inequality of employment and wages (Montgomery, 1991; Krauth, 2004; Fontaine, 2008). Given that contacts might be homophilous with regard to internal quality, the extensive use of referrals lead necessarily to growing inequality (Montgomery, 1991; Beaman and Magruder, 2012).

    • Optimal taxation and social networks

      2014, Macroeconomic Dynamics
    • Searching with Friends

      2023, Journal of Labor Economics
    View all citing articles on Scopus

    This paper benefits from comments by the editor and anonymous referees, William Brock, Lorne Carmichael, Kim-Sau Chung, Steven Durlauf, and Jonathan Parker, as well as workshop participants at University of Wisconsin, the Santa Fe Institute Graduate Economics Workshop, the 2001 CEA meetings, and the 2001 WEHIA conference. All errors are mine.

    View full text