Intervention analysis based on exponential smoothing methods: Applications to 9/11 and COVID-19 effects

摘要

This study extends intervention analysis beyond the ARIMA models, which are currently used by most scholars and practitioners, to exponential smoothing models. This allows us to obtain the benefits of exponential smoothing modeling in analyzing time series with interventions. Exponential smoothing modeling allows for easier seasonal adjustments, and complex seasonality can be more readily incorporated into the analysis. In this study, we propose a method of intervention analysis based on exponential smoothing models through an innovational state-space model, and we obtain maximum likelihood estimates by maximizing the likelihood function of the state-space model. We analyze two applications: the 9/11 effect on U.S. airlines and the COVID-19 effect on the current population of Seoul, Korea. From the proposed method, we estimate the intervention effects and seasonal components in each series. This results in seasonally-adjusted time series with both intervention and seasonality removed.