Internal capital is a major source of funds for financing corporate investments. For example, Lamont (1997) estimates that for the 1981-1991 period, internal funds accounted for more than three quarters of capital expenditure outlays for U.S. non-financial corporations. This reliance on self-generated cash as a source of investment funds has prompted researchers to investigate the relationship between firms‘ investment decisions and internal resources. In a world where capital markets are perfect and all firms have free access to external sources of financing, investment decisions would be based solely on expected future profitability and, thus, not be affected by the availability of internally generated funds. In the real world, however, capital market imperfections exist, making internal funds less costly and therefore more attractive than external funds. This reliance on internal funds should be stronger for firms facing greater capital market imperfections, i.e., firms in countries where external capital markets are less developed. The interesting empirical question that arises therefore is whether corporate investments are indeed more sensitive to the availability of internal funds for firms in less developed economies. This is the question this paper addresses.
In recent years, there have been several studies examining cross-country differences in the development of financial markets and its impact on economic growth (La Porta, Lopez de Silanes, Shleifer, and Vishny (1997, 1998), Levine and Zervos (1998), Rajan and Zingales (1998), Beck, Levine, and Loayza (2000)). Starting with the seminal work of Fazzari, Hubbard,and Petersen (1988), there is also a large literature on the impact of external financing constraints on the sensitivity of investments to internal cash in a variety of domestic contexts (see, e.g., Fazzari, Hubbard, and Petersen (1988, 2000), Blanchard, Lopez-de-Silanes, and Shleifer (1994), Calem and Rizzo (1995), Gilchrist and Himmelberg (1995), Kaplan and Zingales (1997, 2000), Cleary (1999), Erickson and Whited (2000), and Houston and James (2001)).1 In view of the fact that current cash flow is likely to be positively correlated with future profitability/growth opportunities, these studies have typically used Tobin‘s q as an independent variable to control for it. While the domestic evidence on the impact of internal cash on investments is extensive, the international evidence is surprisingly sparse. Even the limited evidence that exists is mixed in nature. On the one hand, Hoshi, Kashyap, and Scharfstein (1991) and Shin and Park (1998) find that the sensitivity of investments to internal cash flow is less for Japanese and Korean firms that belong to corporate groups and thus have easier access to external capital. Similarly, Schaller (1993) finds the sensitivity to be lower for Canadian firms that are older and have more dispersed ownership. On the other hand, Kadapakkam, Kumar, and Riddick (1998) and Cleary (2001) find that in several developed countries, the sensitivity is lower for smaller and financially more constrained firms, i.e., firms for which external capital market frictions are likely to be greater. None of these studies examines the variation of investment-cash flow sensitivity across countries, however. Using cash stock as a measure of internal resources, and sales as a proxy for the marginal productivity of capital, Love (2000) finds the sensitivity to be greater for less developed countries. In a somewhat different context, Wurgler (2000) examines industry-level data and finds that investment is more sensitive to value added in countries with better developed financial markets, but is unable to distinguish between the effects of future profitability and current cash flow due to data limitations in the non-U.S. portions of his sample, which preclude measuring Tobin‘s q at the industry level.
In this paper, we present evidence from a large data set spanning 31 countries and 11 years (1987-1997). We find that after taking certain methodological considerations into account, and controlling for growth opportunities, the evidence is clear and unequivocal: corporate investments are more sensitive to internal cash flow for firms in countries with less developed financial markets. For example, the estimated investment-cash flow sensitivity is significantly higher for the non-OECD countries (0.230) than the OECD countries (0.141). Direct tests of the impact of financial development on investment-cash flow sensitivity also yield a strongly negative relationship. The result is robust to different estimation procedures, to six different measures of financial development, to three different measures of cash flow, and to the use of bootstrapped standard errors. We also find evidence of a strong negative relationship between investment cash-flow sensitivity and size (as measured by log of total assets) across countries, after controlling for the effects of financial development and growth opportunities, which is consistent with the original Fazzari, Hubbard, and Petersen (1988) hypothesis and the notion that smaller firms face greater information costs and are therefore more dependent on internally generated capital for their investment outlays.
Our contribution is also in terms of methodology. One of the concerns with earlier studies that relate the effect of financial development on the variable of interest (e.g., value added in Wurgler (2000)) is methodological. It has been customary to estimate this effect in two steps. In the first step, the sensitivity of the variable of interest is estimated for each country from separate country-by-country regressions. In the second step, the estimated sensitivities are regressed on the measure of financial development. In this study, we estimate the impact of financial development on investment-cash flow sensitivity in a single step, specifically by interacting our measure of financial development with cash flow and estimating its impact on investment using a pooled sample of firms from all countries. This approach addresses the errors-in-variables problem of two-step estimation and yields more efficient estimates.
Finally, Allayannis and Mozumdar (2001) have recently shown that negative cash flow observations may have a serious distortionary impact on estimated investment-cash flow sensitivities. The intuition is that when firms are in sufficiently bad shape (incurring cash losses), investments are down to their lowest possible levels and cannot be cut any further. In such situations, therefore, investment-cash flow sensitivity is extremely low.2 Including such negative cash flow observations in the sample reduces the estimated sensitivity for the entire sample. Allayannis and Mozumdar (2001) show that such observations play an important role in explaining the startling results reported by Kaplan and Zingales (1997) and Cleary (1999). We find that a similar problem exists in the international data, which possibly explains some of the ambiguous results reported in the literature earlier.
The rest of the paper is organized as follows. Section I describes our sample and estimation methodology. Section II presents the evidence on the impact of financial market development on investment-cash flow sensitivity in different countries. Section III examines the robustness of the results to alternative definitions of cash flow and financial development. Section IV explores the additional impact of firm size on investment-cash flow sensitivity. Section V concludes.
