Estimation and Inference by Stochastic Optimization: Three Examples

TL;DR

This paper demonstrates how stochastic optimization algorithms, specifically resampled Newton-Raphson and quasi-Newton methods, significantly reduce computation time for estimation and inference in structural models, with empirical and simulation evidence.

Contribution

It introduces and empirically validates two resampled algorithms that accelerate estimation and bootstrap inference in structural models, improving efficiency over standard methods.

Findings

Computation time reduced from 5 hours to 15 minutes using rqN.

The methods maintain accuracy in estimation and inference.

Efficiency gains are demonstrated in dynamic panel regression.

Abstract

This paper illustrates two algorithms designed in Forneron & Ng (2020): the resampled Newton-Raphson (rNR) and resampled quasi-Newton (rqN) algorithms which speed-up estimation and bootstrap inference for structural models. An empirical application to BLP shows that computation time decreases from nearly 5 hours with the standard bootstrap to just over 1 hour with rNR, and only 15 minutes using rqN. A first Monte-Carlo exercise illustrates the accuracy of the method for estimation and inference in a probit IV regression. A second exercise additionally illustrates statistical efficiency gains relative to standard estimation for simulation-based estimation using a dynamic panel regression example.