Mixed-frequency models

Abstract

This chapter provides a comprehensive review of mixed-frequency models applied to tourism forecasting using data sampled at different frequencies. A series of mixed data sampling (MIDAS) models are discussed, including the basic and unrestricted MIDAS models, the MIDAS with autoregressive terms and other extended MIDAS models. Empirical examples are provided to test the forecasting performance of the MIDAS-type models. The results illustrate that the MIDAS-type models are effective tools to increase forecasting accuracy and decrease forecast failure risk as they fully exploit the available information sampled at different frequencies.