Resumen de Essays on the labour market
In large labour market there are a lot of frictions, that arises limited mobility or informational asymmetries and imperfections, among many other things. But what intrigued me is how the labor market outcomes of workers many times depends also on characteristics that are not only related with workers' skill level or with the productivity of the match between a worker and a job. For example, the social group a worker belongs to, his social relationships, what people similar to him are doing, etc.
In this thesis, I propose three different examples of how the (possibly strategic) interaction between workers (and firms) contributes to explain how they are matched in a frictional economy. While in my first two papers I will describe two different reasons why assortativity between workers and jobs might fail, in the last one I will propose on analysis of the incentives of workers to use friends and relatives to collect information about job openings.
In the first article, Segmentation in a Labor Market with Two-Sided Heterogeneity: Directed versus Undirected Search, assortativity might fail to result ex post when the technology that matches worker and firm is completely random. Here, I propose a model of a frictional labor market with a continuum of workers types where segmentation depends on the ability of the workers to direct their search strategies to firms adopting different technologies. If the technology that arranges the meeting between job openings and applicants is totally random, a shortcut used in the literature to summarize search frictions, not surprisingly, some of the matches that form are not perfectly assortative. Waiting for a better match is costly, and hence agents might accept a partner that is not ideal. This is captured by the definition of the acceptance sets, that describe the reservation type that a firm might accept. This reservation type might be the lowest acceptable for a good firm, or the highest affordable for a not so productive firm. When this sets overlap, the randomness of the matching technology makes the profile of matching observed ex post conditional on the actual realizations of the random process. So, some degree of mismatch can be observed. Quite nicely though, the fact that not all meeting are acceptable induces some assortativity.
Perfect assortativity on the other hand requires a segmentation of the labour market. The novelty of my paper is that it relates this result with the workers ability to direct their application to one of the types of firms. This forces them to make a decision that eliminates one of the sources of frictions that where in the completely random scenario: the informational friction. When every worker could meet with any type of firm in the application stage, workers types exert a sorting externality among themselves since a lot of meetings are among unacceptable partners. When workers can direct their application, this sort of externality dies away, and every meeting becomes acceptable. Obviously where each worker is going to look for a job depends not only on where he is more productive, but also on the labour market conditions in each of the submarket. for the different types of technologies. Since in search models comparative statics results depend on the matching profiles, the model is well suited for this kind of exercises. Indeed, I use a calibrated version of the model model to shade light on the ambiguous relation between skill biased technological change and unemployment rate.
While unemployment increases for unskilled workers, it decreases for skilled workers.
The second paper, Discrimination, Mismatch and Unemployment, analyzes the mismatch between job and workers characteristics that can emerge due the interaction between search frictions, coordination externalities between firms and workers strategies and discriminatory hiring rules. I propose a wage posting model where firms have to decide which technology to adopt and use a non neutral tie breaking rule when receiving many equally productive applicants. The game has two symmetric equilibria, one where the profile of matches is perfectly assortative and one where workers which are penalized by the discriminatory hiring rule are mismatched in technologies where they are less productive. In both cases, discriminated workers receive a lower wage and the labor market is segmented. But while in the assortative equilibrium discriminated workers face lower unemployment spells, in the equilibrium with mismatch they face higher unemployment spells. This is not only a theoretical curiosity. Indeed, mismatch allows to reconcile frictional models where firms use discriminatory hiring rules with observed unemployment and wage gaps. For example, it is able to account for a stylized fact about Black Americans in the US: they earn lower wages conditioning on skill level (the so called wage gap) while facing higher unemployment spells.
While the wage gap has been successfully explained by introducing some sort of discrimination in the labour market. But usually a discriminatory labour market has the counterfactual implication that the discriminated workers face lower unemployment spells. The intuition lies in the equal profit conditions for firms. Since all of them have to gain the same profits (if zero or not depends on free entry), the one paying higher wages need to find easily workers, meaning that there workers that apply to the high wage face an higher unemployment rate. This happens because they are paid a lower wage and all firms must obtain the same level of profits. But when firms can decide on the adoption of technology, discriminated workers might decide to apply for jobs where they are less productive and face high unemployment spells to avoid discrimination in their ideal jobs, while firms open these kind of vacancies because they are cheaper.
Since in this equilibrium discrimination operates through the matching profile, it would not be detected by a researcher trying to estimate the wage gap taking the assignment of jobs to workers as exogenous or random.
Discrimination takes the weak form that firms use race as a selection criteria only when two applicants are totally equivalent in term of acquired skills. Since discrimination lowers the expected benefit of a discriminated worker to look for a job for which he is qualified, he might decide to apply to positions for which he is over-skilled. The resulting possibility of mismatch makes it possible in certain equilibria to observe observationally equivalent workers receiving lower wage and facing higher unemployment spells than workers belonging to different groups not discriminated against.
While in the other two chapters, I have focused on the way in frictional markets firms and workers strategies interact in determining assortative matching profiles, in the last chapter the focus is on the way workers look for a job. One reason why labor market are frictional is because workers and firms are not perfectly informed about each other location. This prevent the market to be a Walrasian spot market where positions are cleared. Workers get to know the existence of some vacancies in a random way, so that some unemployed workers find a match and some do not. But of course they may decide to share the information they are exposed to with unemployed friends. If these information is shared through word of mouth communication among workers that know each other, the information flow can be characterized studying the properties of the network of friendship that connects the workers. In most of the literature on the use of friends and relatives for job seeking purposes, these social networks are taken as a given, mostly for reasons of tractability. In Endogenous Job Contact Networks, a joint work with Andrea Galeotti, we develop a model where social networks are used to collect job offers, but workers decide strategically how much to invest in their network. This socialization game has a unique interior equilibrium, which depends on labor market conditions. Equilibrium job contact networks are dense and more productive in transmitting information when labor market turnover is moderate, while they are less productive and segmented into clusters for either high or low turnover.
As I mentioned, existing models of social networks and labor market generally assume that social networks are exogenous. This assumption implies a negative correlation between unemployment rate and proportion of job offers are filled through job seekers' social network. In contrast, in our model, the equilibrium response of job contact networks to changes in labor market conditions is sufficient to generate a positive correlation, which is consistent with the empirical patterns which we document using the quarterly UK labor force survey. Using the same data, we also construct direct measures of the use of social network, and show how they change with the change in labor market conditions.
This suggests that indeed it is important to consider the endogeneity of job contact networks to labor market conditions, and, in doing so, it question what is the role of social networks in explaining durable inequality among different group of workers. For this analysis, it is necessary to enrich the framework proposed by adding heterogeneity of workers and jobs, in order to study how social networks transmit information about the existence of a job, but also its characteristics.
