Seeking Similarity: How Immigrants and Natives Manage in the Labor Market World Bank October 16, 2012 Olof Åslund Lena Hensvik Oskar Nordström Skans IFAU and Uppsala University 1
Immigrants labor market integration Migrants (in particular refugees) work less, are more often unemployed, and earn less than natives. Immigrants are sorted non-randomly across firms, even conditional on industry, region and human capital Sorting appear strongly linked to economic outcomes 2
Åslund and Skans (2010, ILRR) 3
Strong association between the degree of separation from natives and economic outcomes 4
Weaker link between pure ethnic segregation and economic outcomes 5
6
Motivation Ethnic workplace segregation is well documented in the US, Sweden and Germany Hellerstein and Neumark (2008), Åslund and Skans (2009, 2010), Dustmann et al (2011) The underlying process is not well-understood In this paper we focus on the role of managers 7
Motivation Managers are key players, yet we know very little about their effects. Immigrants are underrepresented among hiring managers (1.5 percent) relative to hires (7.2 percent) We use economy wide data to study the impact of manager ethnicity on hires, wages and quits at the establishment level. 8
Motivation (continued) Adds to a set of recent studying the role of manager race or origin using single firm data. Manager race and the race of hires/quits in the US Giuliano, Levine and Leonard (2009, 2011), Giuliano and Ransom (2011) Managers favor workers of their own ethnicity within firms Bandiera, Barankay and Rasul (2009) We are the first to study the role of manager origin for recruiting patterns using economy wide data. 9
Findings Estimate the impact of manager origin on the origin of new hires, on quits, and on wages. Strong impact of manager origin on the composition of new hires, Product market competition larger effects, Similarity predict higher wages and lower quit rates among workers, but only if not accounting for unobserved individual heterogeneity. Consistent with models of employee selection/statistical discrimination Managers may be better at attracting or selecting high aptitude workers of a similar background. 10
Why should similarity between managers and workers matter? Think of a search and matching world with (potential) preference discrimination (e.g. Rosén, 2003, JOLE) 1. Preference discrimination (utility from similarity)? Lower productivity threshold 2. Endogenously higher productivity (Lazear, 1999)? Sorting may be efficient 3. Informational advantages about match or agent quality? Stronger signals, or segregated networks 11
Data Linked longitudinal employer-employee dataset Main sample cover managers and new hires 1997-2005 Focus on plants < 50 employees All public sector plants and a large sample of private plants. Around 30 percent of all small to medium plants in the private sector Our main used dataset consists of over 700,000 new hires during 9 years, at around 100,000 different establishments. 12
Data Origin determined by country of birth. Natives : Including Western immigrants Immigrants: Eastern Europe and Rest of the world Managers : (1) Top manager (ISCO-88) (2) Middle manager (ISCO-88) (3) Highest wage at plant. 13
14
15
16
Lower Immigrant share among recruiting managers than among hires in all sectors 17
A straightforward model 18
Baseline estimates 19
Baseline estimates 20
With self-employed owners 21
22
23
24
25
26
27
28
29
30
Stronger effects with more competition High competition Low competition Private sector Public sector Baseline 0.072 0.044 0.074 0.038 (0.011) (0.013) (0.038) (0.007) Establishment 0.048 0.035 0.091 0.024 Fixed effects (0.018) (0.020) (0.026) (0.015) 31
Other heterogeneity 1) Industry differences, but positive estimates across the board 2) No clear patterns related to size of establishment 3) Small differences depending on worker characteristics 4) Small differences depending on manager characteristics 32
The impact on wages and separations 33
34
35
36
Additional findings 1) Managers with a higher fraction of immigrants among their former coworkers recruit more immigrants This impact is not large enough to explain recruitment differences between immigrant and native managers 2) Similarity also predict recruitent patterns when managers recruit former coworkers. 37
Why is similarity important? Smaller estimates for separations, where workers are relatively more important Suggest that it is more about managers, less about workers Stronger effects when product markets are competitive Not in line with preference discrimination Higer wages within establishment, no wage impact when individual heterogeneity is accounted for Not (quite) in line with match specific productivity Suggest that it is more about selecting high quality workers 38
Conclusions Manager origin is an important predictor of recruited workers origin, even within firms and establishments. Underrepresentation of immigrants at the demand side may contribute to matching difficulties for immigrant workers Suggest positive externalities from promoting immigrant career progression 39