FIRST DRAFT – INCOMPLETE – PLEASE DO NOT CITE

How should a policy-maker prioritize interventions to improve the public infrastructure with which firms operate? We propose an approach to this question that uses managers ’ rankings of how important public infrastructure constraints are for the operation and growth of their firm. We provide a conceptual framework within which such data can be interpreted. This centres on the public input character of infrastructure and uses an O-ring production function to model the impact of poor quality infrastructure on output. Using survey data from over 72,000 firms in 95 countries we verify the consistency of the within-country firm level variation and the cross-country variation in average reported constraints with predictions of the model and with well-established facts about economic development. We then show how the policy-maker can use the country-level data and bench-marks estimated from the cross-country data in prioritizing public infrastructure investment