Investment dynamics and forecast: Mind the frequency

Abstract

We analyze the in-sample fit and the out-of-sample forecast performance of the investment equation using different proxies for Tobin’s Q, controlling for cash flow, and using their frequency-decomposed components. We show that different frequencies of bond Q and cash flow significantly improve the empirical performance (both in-sample and out-of-sample) of the traditional investment equation. The key step is to filter out the noisy frequencies of the predictors and only retain those that have the greatest individual predictive power.