5 September: Markus GEBAUER thesis

• Patrick FEVE, TSE researcher, University Toulouse 1 Capitole
• Franck PORTIER, TSE researcher, University Toulouse 1 Capitole
• Marti MESTIERI, University of  Northwestern
• Claudio MICHELACCI, Einaudi Institute for Economics and Finance

This Thesis contains three chapters that study firms with more than one worker per production unit in the context of search frictions and different employment contracts. 

The first chapter reexamines the canonical Mortensen-Pissarides-Diamond search model and introduces complementarities in labour into an otherwise standard framework. Specif- ically, I assume that two workers are needed to perform the task imposing an extreme form of complementarity and examine whether it is possible to implement optimality. I do this in a random search framework and then in a directed search framework. In both cases I maintain the assumption of Nash-bargaining. In a continuous framework the necessity to keep one worker employed while still searching for the second worker arises. This gives rise to additional externalities that make it impossible to correct for by Hosios-style conditions in a random search framework. In particular, not only do firms exert a congestion external- ity upon other firms when they post a vacancy, they also exert an externality by hiring the first worker of the team which then leads to the firm needing to post more vacancies looking for the second worker. Equally, the first worker arriving imposes a congestion externality upon other workers through the same channel. Introducing directed search reveals that it is only possible in the rarest of cases to implement optimality. 

The second chapter presents a search and matching model of dual labour markets. It exhibits higher matching rates and lower surplus for temporary workers, without assuming unobserved ex ante and ex post heterogeneity to motivate the use of different contracts. The model predicts that increases in firing costs lead to a decrease in permanent workers and an increase in wages for permanent workers, whereas hazard rates remain stable. I then empirically support these predictions using a regional French policy experiment, the “Contrat de Transition Professionelle”, which increased firing costs of permanent workers in firms below 1000 workers in 2006. It was implemented in 7 employment zones (“Zone d’Emploi”) which were already struggling economically thus not representative of France, so I employ Propensity Score Matching to achieve identification. I use matched Employer- Employee data from France to conduct this analysis which enables me to use individual level and firm level data. I find that indeed fewer permanent workers are hired in the treatment areas as opposed to the control areas, permanent workers to whom the increased firing costs applied were able to negotiate a higher wage and hazard rates for existing permanent employees did not change significantly. 

The third chapter takes a broader view and ask what (beyond labour market regulation) might have contributed to the growth of temporary job arrangements in the past decades all over the world. It theoretically investigates the decision of firms to hire temporary and permanent labour when workers and firms meet through a frictional directed search technology. Temporary workers differ from permanent workers in that they have a lower bargaining weight and look for a permanent job while on the temporary job. The findings are that permanent arrangements are more prevalent the more productive the aggregate production function is because firms want to retain more workers even in downturns. More efficient matching first increases the prevalence of temporary arrangements because workers are easier to replace and then decreases it as temporary workers become more likely to leave. In this sense a more efficient matching function is a double edged sword. Increasing the outside option means more temporary arrangements as it makes them less risky for the workers and hence more attractive. Bargaining weights for any type of worker have an ambiguous impact on the number of the affected workers since for relatively low starting values higher wages attract more workers at still relatively low wages which firms are willing to pay but ultimately for higher starting values of the bargaining weight wages are too high to justify more workers.