Do Investment-Cash Flow Sensitivities Measure Financial Constraints? New Evidence from Unpacking the Black Box

Abstract

The relationship between investment and cash flow has been a subject of considerable empirical debate. In this paper, our aim is to shed a new light on this relationship. It is argued that a firm’s observed investment and financing decisions are endogenous, in an ex-post behavioral sense, and that, therefore, investment-cash flow sensitivity (when estimated in a reduced-form, linear investment equation) is likely to be an ambiguous, and possibly misleading, measure of (current and expected future) financial constraints. Empirical evidence is provided in support of our case, based on a two-stage estimation strategy. In the first stage, allowance is made for unobserved firm heterogeneity as related to cash flow; we generalize the Q-investment equation and estimate individual, firm-specific investment-cash flow sensitivities by applying an entropy-based fixed-effect estimator. In the second stage, we look inside the black box and investigate how and to what extent these sensitivities are “driven” by important underlying factors. A number of striking ceteris-paribus results emerge: a) investment-cash flow sensitivity is monotonically decreasing in the level of cash flow; b) cutting back on investment, hoarding more cash, taking less on debt, building up debt capacity, and paying low or zero dividends (actions that are typically associated with the presence of tighter financial constraints) tend to produce a smaller value of investment-cash flow sensitivity; c) investment-cash flow sensitivity is negatively related to cash-flow volatility and positively related to investment volatility; and d) firm size and asset tangibility do not contribute to explaining the variation in investment-cash flow sensitivities across firms. From these results, it follows that the relation between the investment-cash flow sensitivity’s magnitude and the degree of financial constraints is indeterminate a priori. While such results call into question conventional wisdom, they help to explain the many contradictory findings encountered in the literature.