Peering Through the Fog of Uncertainty: Out-of-Sample Forecasts of Post-Pandemic Tourism

This paper uses an augmented gravity model framework to investigate the historical impact of infectious diseases on international tourism and develops an out-of-sample prediction model. Using bilateral tourism flows among 38,184 pairs of countries during the period 1995–2017, I compare the forecasting performance of alternative specifications and estimation methods. These computations confirm the statistical and economic significance of infectious-disease episodes in forecasting international tourism flows. Including infectious diseases in the model improves forecast accuracy by an average of 4.5 percent and as much as 7 percent relative to the standard gravity model. The magnitude of these effects, however, is likely to be much greater in the case of COVID-19, which is a highly contagious virus that has spread fast throughout populations across the world.