Nowcasting Thailand Economic Activity Using the Google Mobility Data


Forecasting economic activity in the covid-19 pandemic era is challenging. At the same time, policymakers and business leaders require timely evaluation of the health of the economy. Even in the period before the pandemic, most of Thailand’s economic indicators lag in an announcement. This paper explores the benefit of adding publicly available data, i.e., the Google Mobility data, for assessing the economic situation. We compare the forecasting performance of the Autoregressive Moving Average (ARMA) model, ARMA with explanatory variable (ARIMAX), and Mixed-data sampling (MIDAS) model using the Google mobility index. We consider the monthly service production index and manufacture production index. We find that the Google Mobility data help to improve the forecasting performance of various service sector. While the models with the Google Mobility index perform worse than the pure time series model, i.e. ARIMA in finance, public administration and manufacturing sector.

Chaleampong Kongcharoen
Chaleampong Kongcharoen
Assistant Professor of Economics

I’m an assistant professor of economics at Thammasat University. My research interests are time series econometrics, and empirical macroeconomics.