Åsa Rosén, SOFI, Stockholm University; Homepage: http://www.sofi.su.se/~aro

2007-2014

Numerous empirical studies document that women, immigrants, and black earn less, are less likely to be promoted, or to find a job than men, natives, and whites (Altonji and Blank, 1999). In the overwhelming majority of these studies, the differences cannot be fully explained by differences in education, training, and other attributes which are likely to affect workers' productivity. Existing theoretical work on discrimination tends to focus on explaining wage differentials within firms.

One of the project's main objectives is to theoretically examine the scope for discrimination during recruitment. Several empirical studies report that job applicants with foreign-sounding names have a 40-50 percent lower chance of being invited to an interview, even when other characteristics are identical (OECD, 2008, Bertrand and Mullainathan, 2004). The key theoretical issue is to what extent market forces (competition) eliminate discriminatory hiring practices, and if not, whether labour market policies are effective in mitigating discrimination.

Contrary to early competitive models of discrimination, recent theoretical work introduces search friction. This approach has several advantages. In particular, search models incorporate externalities, e.g., the impact which the treatment of one group, say natives, has on members of the other group (immigrants). Furthermore, unemployment is an equilibrium feature of search models and can therefore be analysed simultaneously with wage differentials and discriminatory recruitment strategies.

Besides exploring the possible persistence of discriminatory recruitment practices, the project also makes a contribution to search theory, the prevalent theoretical framework used in modern labour economics and its macroeconomic applications. The research in this project is also relevant for the evaluation of labour market policies, such as the effectiveness of education and employment subsidies or of employment protection legislation.

The project is divided into two parts. The first part is joint work with Steinar Holden (University of Oslo) and explores discriminatory recruitment practices within a (search) model of self-fulfilling beliefs and vicious circles. (For earlier such models see e.g., Arrow, 1973; Coates and Loury,1993). The second part is joint work with Espen Moen (BI, Oslo) and develops a novel search model.

Part I: Discriminatory Hiring, Discriminatory Firing, and Employment Protection

Several papers analyse the effects of employment protection on key economic variables, such as unemployment, employment fluctuations, and investments in human capital (e.g., Addison and Teixeira, 2003; Garibaldi and Violante, 2005). This literature largely neglects possible interactions between employment protection and labour market discrimination. This issue is the subject matter of this part of the project.

In the paper "Discrimination and Employment Protection", Holden and I provide a novel explanation for discriminatory outcomes for equally productive groups in labour markets where it is costly for firms to lay off workers. The intuition for the key argument is as follows: Workers with low job-finding rates are riskier to hire because they may not leave the firm even if the firm-worker match turns out to be poor. Discrimination becomes self-enforcing, as it is precisely because a group is discriminated against that this group has lower job-finding rates, and is thus less attractive to hire. While non-discriminatory outcomes also exist in our model, these equilibria vanish if a sufficiently large proportion of employers have a taste for discrimination. Thus, the existence of discriminatory employers makes discriminatory recruiting practices profitable also for profit-maximizing firms, which is in sharp contrast to Becker's model (Becker, 1971). In addition, if there is a sufficiently large proportion of employers who hire according to the population shares of each group - a hiring quota - the discriminatory equilibrium vanishes.

The paper's main contribution is the theoretical result that discriminatory recruitment practices can persist. It is worth mentioning that the predictions of our theory are consistent with salient features of the labour market situation of immigrants. For example, in almost all OECD countries, immigrants are much more likely to have temporary employment than native-borns. This is in line with the notion that employers are reluctant to hire immigrants directly on a permanent basis (with employment protection). The Economic Survey of Sweden 2007 argues that one of the key elements to combat exclusion is "to reduce the risk associated with hiring someone who turns out not to be the right person for the job" - precisely as suggested by our model.

The paper is in the second round at the Journal of the European Economic Association and has been presented at the University of Oslo, the Frisch Centre, ETH Zurich, Norwegian School of Management, Norwegian School of Economics and Business Administration, the Norwegian Meeting for Labour Economists, Stockholm University, Umeå University, the CEPR/IZA 2009 Summer Symposium in Labour Economics, the 2009 Meeting of the Society of Economic Dynamics, and the 2010 World Congress of Econometric Society.

While working on the paper "Discrimination and Employment Protection", the question has emerged to what extent labour market outcomes will be affected by heterogeneous groups, that is, when groups are ex ante different: Will the outcome be proportional to the productivity differences, or will they be larger? Furthermore, can there be discrimination at the firing stage? In the project "Hiring and Firing Discrimination with On-the-job Search and Employment Protection", which is work in progress, Holden and I explore these issues. Preliminary results indicate that firms apply stricter hiring standards to the group with lower average productivity. Thus, this group has higher unemployment rates because the group members have lower productivity but also because they are treated differently at the hiring stage. Furthermore, we find that workers from the disadvantaged group are more likely to be fired for a given productivity, in addition to the stricter hiring standards. In a setting where wages do not only depend productivity, but are also group dependent, numerical simulations demonstrate that firms may still use stricter hiring-standards for disadvantages workers and at the same time are more prone to fire these workers. This outcome is more likely when minimum wages and unemployment benefits are relatively high. This research has been presented at the Norwegian Economists' Meeting.

Part II: Frictions and Heterogeneous Workers

There are two standard approaches to modelling frictions in labour markets; random search and directed search. Both approaches incorporate relevant features of labour markets, but each has its shortcomings. The random search framework captures well that allocating the right job to the right worker is costly for both firms and workers. However, offered wages do not affect the number and composition of applicants in random search models. While this is precisely the strength of directed search models, this latter framework does not reflect well the costs of matching heterogeneous workers with heterogeneous firms. The reason is that workers are typically assumed to observe all vacancies but direct their search to only one vacancy.

In the paper "Invisible Markets", Moen and I develop a competitive search model with a given number of search markets, where workers only observe the wages offered in a limited number of submarkets. As a result, heterogeneous workers are not able to fully self-select into different submarkets, and workers end up searching in markets in which also workers of other types search for a job. We show that the equilibrium of the model is a "Maximum Segmentation Allocation" (MSA) outcome: Workers are segmented to the largest degree possible, given the information constraints. The expected income of a given worker depends positively on the fraction of workers of his type in the economy, which gives rise to feedback effects. For instance, the return on human capital investment is an increasing function of the fraction of workers who invest, and this may lead to multiple equilibrium outcomes. Furthermore, the model shows that outcomes in the labour market do not only depend on the productivity of the worker but also on whether a worker belongs to a majority or minority group in the economy. Accordingly, the theory has the potential to account for discriminatory outcomes. This paper has been presented at the Meeting of the European Economic Association and the Society of Economic Dynamics.
References
Addison,J.T. and P. Teixeira, (2003), "The Economics of Employment protection", Journal of Labor Research, 24, 85-129.
Altonji, J.G and R.M. Blank, (1999), "Race and Gender in the Labor Market" in Orley Ashenfelter and David Card (eds.), Handbook of Labor Economics, Volume 3C, New York: North-Holland, pp. 3143-3259.
Becker, G.S., (1971), The Economics of Discrimination, Chicago: University of Chicago Press. (original edition,1957).
Bertrand, M. and S. Mullainathan, (2004), "Are Emily and Greg more Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination," American Economic Review, 94, 991-1013.
Coates, S. and G.C. Loury, (1993), "Will Affirmative-Action Policies Eliminate Negative Stereotypes?" American Economic Review, 83, 1220-1240.
Garibaldi, P and G.L. Violante, (2005), "The Employment Effects of Severance Payments with Wage Rigidities," The Economic Journal, 115, 799-832.
OECD, (2008), Employment Outlook.

Publications

Discrimination and Employment Protection, 2012, with Steinar Holden, University of Oslo.

Invisible Markets, 2010, with Espen R. Moen, Norwegian School of Management, Oslo.

Hiring and Firing discrimination with On-the-Job Search and Employment Protection, 2014, mimeo, with Steinar Holden, University of Oslo.