The authors show that the cash flow measure in past studies (profit plus depreciation), reflecting only a weak relationship with investment, does not represent actual cash flow available for investment. The cash flow measure used by the authors corrects for the effects of extraordinary items, deferred taxes, unremitted earnings in unconsolidated subsidiaries, losses from the sale of long-lived assets, and other operating cash flows. The authors also use broader measures for long-term investment and consider working capital investment and other uses of cash.

According to the authors, one dollar of cash flow from the current year results in a net working capital investment of $0.11 and a capital expenditure of $0.15, both statistically significant. The prior year’s cash flow, however, has a greater impact on investment. One dollar of lagged cash flow is used for total fixed investment ($0.68), working capital investment ($0.09), dividend payouts ($0.06), and decreases in equity ($0.20). At the same time, debt in the sample increases by $0.09.

The authors break the sample down into financially constrained and unconstrained firms. Recognizing that there is measurement bias with market/book value because book value is in the denominator, they use past returns to estimate q. After correcting for measurement error and controlling for q, they find that $1.00 of cash flow in the current and prior years is associated with an additional fixed investment of $0.63 for constrained firms and $0.32 for unconstrained firms. Working capital investment increases by $0.15 for constrained firms and decreases for unconstrained firms. Unconstrained firms also use the dollar of cash flow to increase dividends ($0.18), reduce debt ($0.24), and repurchase shares of stock ($0.29).

After controlling for expected cash flow, the authors find a negative correlation between investment and market/book value for unconstrained firms, which implies that profitable growth firms may have free-cash-flow problems. Furthermore, the larger investment related to cash flow for constrained firms indicates that financial constraints play a role in investment.

The authors study investment–cash flow sensitivities using data on 1,800 US nonfinancial firms each year over 1971–2009. They use three different measures for long-term investment; the most inclusive one captures the increase in all fixed assets, including acquisitions and patent purchases, no matter how they are financed.

Past studies have defined cash flow as income before extraordinary income plus depreciation, and results have suggested only a small impact on investment in recent years. The authors redefine the cash flow measure as one that approximates operating cash flow on the cash flow statement, excluding changes in working capital because they consider working capital a type of investment. The correlation between the traditional cash flow variable and the authors’ cash flow measure diminishes significantly over time, largely because of an increase in noncash special items beginning in the 1980s. The cash flow measure used in recent studies explains why a weak investment–cash flow association has been observed.

The authors find that cash flow is strongly linked to both short-term and long-term investment and that the association is strongest for firms with financial constraints.

The authors recognize and address problems of previous researchers who attempted to link cash flow to investment. They develop a more relevant cash flow measure that should be considered by future researchers. Their more inclusive model of how cash flow is used—for paying dividends, increasing working capital, repurchasing shares, reducing debt, and investing long term regardless of financing—may be useful for analysts.

The results indicate that financially constrained firms invest more when cash flow increases, demonstrating that financial constraints affect investment for this subgroup. The relevance of financial constraints may have implications for economists and policymakers trying to stimulate the economy.