Managing what you measure: Evidence from federal infrastructure surveys

We study how federal information-collection requirements affect public infrastructure spending of local governments. Exploiting the introduction and revision of the federal survey on highway infrastructure, we find that county roads that are randomly sampled by the survey attract more local government spending. The increased spending is concentrated in capital outlays for road improvements and is somewhat pronounced for roads in poor condition. The spending seems successful in expanding the county road network and usage and reducing fatalities. Collectively, our evidence suggests that federal information-collection efforts shape the allocation of resources by providing information on road segments in need of resources but also by directing attention and resources to roads that happen to be sampled, irrespective of their need for resources.