Abstract
We introduce a novel reverse restricted MIDAS (RR-MIDAS) model, which allows us to forecast high frequency data using low frequency information. The RR-MIDAS model is applicable to more general mixed frequency data including the cases with larger differences in sampling frequencies, which are ineffectively handled by the reverse unrestricted MIDAS (RU-MIDAS) model. In Monte Carlo experiments, the RR-MIDAS model outperforms the other models such as RU-MIDAS and HF, in terms of predictive ability. The decent performance of RR-MIDAS model is demonstrated in a real-world application on forecasting US interest rate as well.
Acknowledgements
The authors are grateful to the Editor-in-Chief and three anonymous referees for their helpful comments and constructive guidance. This work was supported by the National Natural Science Foundation of China under Grant 71671056 and 71490725; the National Social Science Foundation of China under Grant 15BJY008.