Title: Forecasting public investment using daily stock returns
Authors: Hiroshi Morita - Hosei University (Japan) [presenting]
Abstract: Financial market variables contain a lot of information to forecast the variations of macroeconomy. By taking advantage of such a desirable property of financial data, the predictability of the Japanese public investment is investigated by using daily excess stock returns of construction industry to contribute to the recent discussion on fiscal foresight. To examine the relationship between monthly public investment and daily stock returns without any time aggregation, we employ the VAR model with MIDAS regression, in which the optimal weights for connecting high frequency data and low frequency data are estimated in addition to the VAR coefficients and variance-covariance structure. Our results reveal that the VAR model with MIDAS regression reduces the mean square prediction error (MSPE) in the out-of-sample forecast by as much as 14 percent in comparison with the no-change forecast. Moreover, based on the local projection method, we also find that fiscal news shock estimated in our VAR model has a delayed positive effect on output after significant negative effect for almost the first one year.