Governments' Home Bias and Efficiency Losses: Evidence from National and Subnational Governments

We use one million procurement contracts awarded in France and Spain to quantify the importance of home bias in explaining governments’ purchases. We propose two strategies that allow us to filter out other factors that may explain the high geographical concentration in government procurement. We use the fact that “home†may have a different meaning for subnational and national governments. We identify their relative bias by comparing how much of the same product local and non-local establishments sell to national and subnational agencies in the same location. We also exploit a quasi-natural experiment that reduced the number of French regions. We compare how much establishments sell in their “new local†regions before and after the change. We use a trade model to quantify what our reduced-form estimates imply and find that governments’ home bias increases local expenditures shares by up to 29% and reduces governments’ output by up to 8%.