Investment without Q

Abstract This paper proposes an alternative to standard investment-Q regressions. Policy functions summarize the key predictions of any dynamic investment model, are easy to estimate and, unlike Tobin’s Q, account for a large fraction of the variation in corporate investment. As such policy functions are much better suited to evaluate and estimate dynamic investment models. Using this superior characterization of firm investment behavior we use indirect inference methods to estimate deep parameters of a structural model of investment and show that investment adjustment cost parameters are generally better identified from estimated policy function coefficients.