Investment-cash Flow Sensitivities: Constrained versus Unconstrained Firms

From the existing literature, it is not clear what effect financing constraints have on the sensitivities of firms’ investment to their cash flow. I propose an explanation that reconciles the conflicting empirical evidence. I present two models: The unconstrained model, in which firms can raise external funds, and the constrained model, in which firms cannot do so. Using low dividends to identify financing constraints in my generated panel of data produces results consistent with those of Fazzari, Hubbard, and Petersen (1988); using the constrained model produces results consistent with those of Kaplan and Zingales (1997). The literature documenting the sensitivity of firms’ investments to fluctuations in their internal funds, initiated by Fazzari, Hubbard, and Petersen (1988), is large and growing. The sensitivity is measured by the coefficient obtained from regressing investment on cash flow, controlling for investment opportunities using Tobin’sQ. Fazzari, Hubbard, and Petersen view firms as constrained when external financing is too expensive. In that case, firms must use internal funds to finance their investments rather than to pay out dividends. Fazzari, Hubbard, and Petersen identify firms with low dividends as “Most constrained” and firms with high dividends as “Least constrained.” As reported in Table I, “Most constrained” firms have investments that are more sensitive to cash flows than “Least constrained” firms. Kaplan and Zingales (1997) disagree with the interpretation of the result. Their identification of financially constrained firms is based on the qualitative and quantitative information contained in the firms’ various reports. Kaplan and Zingales identify firms without access to more funds than needed to finance their investment as “Likely constrained” and firms with access to more funds than needed to finance their investment as “Never constrained.” In contrast to Fazzari, Hubbard, and Petersen, Kaplan and Zingales do not consider firms that choose to pay low dividends even though they could pay out more as constrained. As Table I indicates, the investments of “Likely constrained” firms are less sensitive to cash flows than the investments of “Never constrained” firms. [Table I goes about here.] The regression results discussed above depend crucially on the criterion used to identify whether a firm experiences financing constraints. To explain the empirical evidence, I construct two models: An unconstrained model, in which firms have perfect access to external financial markets, and a constrained model, in which firms have no access. Section I presents the two models, and Section II describes the calibration necessary to solve the models. Series are simulated from the two models and pooled to represent the theoretical sample. Using this laboratory, I investigate whether the empirical results can be replicated. I find that they can. Section III discusses two main findings. First, using low dividends to identify firms with financing constraints leads to Fazzari, Hubbard, and Petersen’s result that low-dividend firms’ investment is more sensitive to cash flow than high-