Macroeconomic and financial mixed frequency factors in a big data environment

Abstract

Abstract In this paper, we evaluate the predictive content of 3 new business condition indexes and uncertainty measures that are estimated using high-frequency financial and low-frequency macroeconomic time series data. More specifically, our measures are defined as latent factors that are extracted from a state space model that includes multiple different frequencies of non-parametrically estimated components of quadratic variation, as well as mixed frequency macroeconomic variables. When forecasting growth rates of various monthly financial and macroeconomic variables, use of our new mixed frequency factors is shown to result in significant improvement in predictive performance, relative to a number of benchmark models. Additionally, when used to forecast corporate yields, predictive gains associated with the use of our measures are shown to be monotonically increasing, as one moves from predicting higher to lower rated bonds. This is consistent with the existence of a natural pricing channel wherein financial risk (as measured using our volatility factors) contains more predictive information for lower grade bonds. We also find that a variety of extant risk factors including the Aruoba et al. [(2009a). Real-time measurement of business conditions. Journal of Business & Economic Statistics, 27(4), 417427] business conditions index also contain marginal predictive content for the variables that we examine, although their inclusion does not reduce the usefulness of our measures.