This paper evaluates how two major broadband subsidy programs from the 2021 Infrastructure Investment and Jobs Act, the Affordable Connectivity Program (ACP) and Broadband Equity, Access, and Deployment Program (BEAD), affect U.S. welfare and market structure. Using a rich dataset that merges FCC deployment data, the Urban Rate Survey, the American Community Survey, and hand-collected federal data, I begin with an event study that shows reduced prices per megabit and improved product quality post-intervention. To explore counterfactual scenarios and address a broader range of questions, I estimate a structural model that combines discrete choice demand with firms’ product-offering decisions. Counterfactuals reveal that BEAD increases firm entry (13.9-19.5%) and product variety (12.6-16.7%). However, it also raises marginal costs and prices (up to 18%), largely due to expanded offerings and growing input market pressures. By contrast, ACP increases consumer surplus (7.5-8.7%) and lowers costs and prices (2.7-3.8%) with a limited impact on firm entry and product variety. These findings underscore a key trade-off: BEAD promotes long-term competition at a rising cost, whereas ACP delivers short-term efficiency gains. Welfare improvements under BEAD diminish at higher subsidy levels, indicating that a subsidy rate of 25-50% may offer the most efficient balance between welfare gains and cost containment.

This paper introduces a new method that integrates machine learning techniques with traditional econometric tools to correct bias in two-step estimation for dynamic discrete choice models under incomplete information. Unlike most existing approaches that primarily address bias from conditional choice probabilities while taking the transition matrix as given, this method simultaneously addresses both sources of bias. I propose a two-step estimator that jointly includes the conditional choice probabilities and transition matrix as first-step estimates, providing a robust, non-iterative solution. The estimator achieves square root N consistency and asymptotic efficiency. In addition, I establish a simple connection between Neyman orthogonality and the Zero-Jacobian property, a key result that guarantees consistent and efficient estimators in single-agent models. Monte Carlo simulations using a common benchmark data-generating process demonstrate the good performance of the proposed method.