Jack Fosten: Autoregressive Difference-in-Differences
Event details
Author: Jack Fosten (City, London)
Abstract: Many studies in economics and finance use difference-in-differences (DID) to analyse the impact of policy or regulatory changes. Policy changes often manifest gradually in outcome variables, causing conventional DID estimates to potentially underestimate the medium- and long-run impacts. To address this problem, we propose an autoregressive DID model that explicitly incorporates this serial dependence and enables the estimation of treatment effects many periods into the future. We propose a bias correction to attenuate the OLS bias of the model. We apply the method to the effects of tax reform on financial assets of large firms, finding that DID underestimates long-run impacts.
Joint with: Ryan Greenaway-McGrevy (Auckland)
Host: Laura Coroneo (York)