Thèse de doctorat en Sciences économiques
Sous la direction de Wilfried Sand-Zantman.
Soutenue le 09-06-2015
Essais en contrats publics
Le résumé en français n'a pas été communiqué par l'auteur.
This thesis is composed by three chapters, each one addressing a policy question originated in the complex nature of public sector's objective function. In the first chapter, a policymaker intends to auction divisible goods among large and small bidders. On top of being concerned about maximizing surplus of auction participants, he also wants to avoid concentrated allocations. Motivated by actual practices on the field of spectrum auctions, this chapter examines standard auctioning policies, composed by auction formats (uniform price or Vickrey) and devices to favour groups of bidders (price discounts or quantity restrictions). The chapter relates the choice of the auction format and the favoring devices with the relative weight that the policymaker assigns to surplus maximization and outcome concentration. More generally, this work highlights the role of bidders’ market power in smoothing the effect of asymmetries. The second chapter compares the adoption of labor incentive policies in the public and the private sector. I build a general equilibrium model in which the private and the public sector need to attract workers to produce private and public goods respectively. In a moral hazard framework, standard profit maximizing leads firms to provide incentives to workers whose expected production is above a threshold. Even though information rents are not per-se costly for a welfare maximizing Government, providing incentives may attract too many workers to the public sector, creating a problem of talents allocation. This tradeoff justifies the choice of an output threshold different than the one in the private sector for giving incentives to workers. Finally, in the third chapter, a Government privately informed about its level of commitment intends to attract a foreign investor to undertake a costly oil project with price-driven risk. I show that a Government with high level of commitment does not have enough contracting elements to separate from a less committed type that is not credible enough to allow the investor recover his sunk cost. Still, a project can be feasible when Government types pool their offers at the cost of compensating the investor with an expropriation premium. I find that contingent taxation can be used to minimize the exposure to such cost. An optimal contract keeps investor's profits constant at the highest level that can be credibly sustained by a low commitment Government in most states of nature, and compensates the investor with higher profits when prices are high enough.