Measuring the Dynamic Effects of Building Digital Infrastructure

The paper measures the impact of public investment in fiber broadband infrastructure on private investment incentives, hiring and firing, and local market structure using a structural dynamic framework together with machine learning methods for causal inference. We model intensive and extensive margins of public investment and its distributional effects and simulate the impact of alternative investment policies. Using detailed Swedish data from 2009-2020, we quantify local authorities’ and firms’ tradeoffs to invest by modeling their decisions, cost structure, and long-run benefits from technology improvements. Investment and hiring setup costs in telecom and tech services are higher in high-access markets than in low-access markets. Counterfactual simulations show that eliminating public investment in broadband infrastructure reduces fiber access by 25 percentage points in sparse areas and 14 percentage points in dense areas in 2020. Removing public investment reduces firms’ incentives to enter, lowers long-run profits, depresses employment, and weakens private investment in telecom and tech services. A network capacity fee on tech services to fund infrastructure and decrease public investment accelerates fiber deployment but redistributes costs and reduces private investment, potentially weakening long-term innovation incentives.