Discrimination in dynamic procurement design with learning-by-doing

Discriminatory programs that favor local and small firms in government procurement are common in many countries. This paper studies the long-run impact of procurement discrimination on market structure and future competition in industries where learning-by-doing makes incumbent firms more efficient over time. We consider a sequential procurement design problem in which local and global firms compete for public good provision. Both firms benefit from learning-by-doing if they provide the public good in the previous period but global firms only may be able to transfer learning-by-doing from different markets. We find that the optimal procurement has to be biased in favor of the local firm even when all firms are symmetric with respect to their initial cost distribution. This bias fosters future competition and reduces intertemporal expected transfers to providers.