A Simulation-Based Method to Estimating Economic Models with Privacy-Protected Data

Differential privacy algorithms typically distort data in ways that bias estimates from standard econometric methods. We describe a simulation-based econometric method that addresses this issue. The approach adapts the Method of Simulated Moments (MSM) for large datasets and models with high-dimensional fixed effects, when traditional MSM is computationally infeasible. We discuss the approach's application to estimating a gravity model of consumer visits using privacy-protected mobile device data. The methodological framework is flexible and applicable to a wide range of settings where economic models are estimated using privacy-preserved data.