Financial Constraints, Investment, and the Value of Cash Holdings*

DAVID J. DENIS Krannert Graduate School of Management Purdue University West Lafayette, IN 47907 [email protected]

VALERIY SIBILKOV Sheldon B. Lubar School of Business University of Wisconsin - Milwaukee Milwaukee, WI 53211 [email protected]

November, 2007

Abstract We provide robust evidence that cash holdings are more valuable for financially constrained firms than for unconstrained firms and investigate why this is so. Our results indicate that greater cash holdings are associated with higher levels of investment for both constrained and unconstrained firms, but that the marginal value of investment is greater for constrained firms. These findings suggest that higher cash holdings allow constrained firms to undertake valueincreasing projects that might otherwise be bypassed. As such, the evidence is consistent with the hypothesis that greater cash holdings of constrained firms are a value-increasing response to costly external financing.

*

We appreciate helpful comments received from Murillo Campello, Diane Denis, Mara Faccio, Mike Faulkender, Stu Gillan, John McConnell, Raghu Rau, Jin Xu, and finance seminar participants at the University of Iowa, Ohio University, and Texas Tech University.

1.

Introduction Modigliani and Miller (1958) argue that in a frictionless environment, companies can

fund all value-increasing investment opportunities.

That is, investment and growth do not

depend on the availability of internal capital. Once capital market imperfections are introduced, however, firms are not necessarily able to pursue all value-increasing investment opportunities. For example, in the models of Myers and Majluf (1984) and Greenwald, Stiglitz, and Weiss (1984), capital market frictions increase the cost of outside capital relative to internally generated funds. Consequently, some firms that have attractive growth opportunities invest less than the first-best optimum, leading to lower future growth and reduced operating performance and firm value. One way to mitigate these adverse effects is for firms with high costs of external finance (i.e. financially constrained firms) to rely more on internal financial resources: cash flow and cash holdings. Cash holdings can be valuable when other sources of funds, including cash flows, are insufficient to satisfy firms’ demand for capital.1

That is, firms facing external financing

constraints can use available cash holdings to fund the necessary expenditures. Consistent with this view, several studies report that firms with greater difficulties in obtaining external capital accumulate more cash.2 Similarly, Almeida, Campello, and Weisbach (2004) provide evidence that firms with greater frictions in raising outside financing save a greater portion of their cash flow as cash than do those with fewer frictions.

1

Several prior studies have analyzed cash flows, as opposed to cash holdings, as an internal source of capital. We provide a brief review of these studies in Section 2. 2

See Kim, Mauer, and Sherman (1998), Harford (1999), and Opler, Pinkowitz, Stulz, and Williamson (1999).

1

The above studies are consistent with two competing views of cash holdings. Under one view, higher cash holdings increase the value of constrained firms because they allow the firms to undertake valuable projects that might otherwise be bypassed. Alternatively, if financial constraints are a byproduct of potential moral hazard problems, high cash holdings might increase the likelihood of agency problems and empire-building by managers of constrained firms.3 We contribute to this debate by (i) providing robust evidence on the value of cash holdings for constrained and unconstrained firms, (ii) investigating the source of value differentials by analyzing the impact of cash holdings on investment for constrained and unconstrained firms, (iii) analyzing the marginal value of investment for constrained and unconstrained firms, and (iv) analyzing the reasons for low cash holdings for some constrained firms. Our study complements and extends recent studies by Faulkender and Wang (2006) and Pinkowitz and Williamson (2004) who study the value of cash holdings. Consistent with cash holdings alleviating financial constraints, Faulkender and Wang (2006) find that the relation between excess returns and changes in cash holdings is stronger for firms that are more likely to be financially constrained. Using a different methodology, Pinkowitz and Williamson’s (2004) findings are inconclusive with respect to the impact of financial constraints on the value of cash holdings. We note, however, that even if the value of cash holdings is greater for firms with greater financial constraints, the interpretation of this result is ambiguous. Greater cash holdings

3

Harford (1999) and Dittmar, Mahrt-Smith, and Servaes (2003) provide support for the hypothesis that cash hoarding by firms is value-reducing and can be a result of agency problems inside corporations, while Mikkelson and Partch (2003) argue that a policy of high cash holdings is not necessarily value-reducing and may in fact be an operating necessity.

2

might be more valuable to constrained firms because they allow the firm to invest when other sources of funds are costly, limited, or unavailable. In other words, greater cash holdings might allow firms that experience external financial constraints to avoid underinvestment and reduced growth. Alternatively, however, it might be the case that the higher value associated with greater cash holdings is a reflection of the market rewarding the firm for holding cash rather than overinvesting that cash in unprofitable projects.

Our evidence on the impact of cash on

investment, as well as on the marginal value of that investment for constrained and unconstrained firms sheds further light on this issue. Using a sample of 60,056 firm-year observations between 1985 and 2002, we report the following main findings. First, in regressions of firm value on cash holdings and a set of control variables, we find that the coefficient on cash holdings is significantly greater for constrained firms than for unconstrained firms. These findings are robust to alternative methodologies for measuring firm value, to measures of abnormal cash holdings and changes in cash holdings, and alternative methods of identifying financially constrained and unconstrained firms. Second, we find that cash holdings are positively associated with net investment (capital expenditures net of depreciation) for all firms. Moreover, the difference between the effects of cash holdings on net investment for constrained and unconstrained firms is statistically insignificant. In other words, firms with greater cash holdings invest more, regardless of whether they are constrained or unconstrained. Third, we find that the association between investment and firm value is significantly stronger for constrained firms than for unconstrained firms. Taken together, these findings are consistent with the view that (i) cash holdings are more valuable to constrained firms because cash allows constrained firms to increase investment, and (ii) the marginal investment of

3

constrained firms is more valuable than that of unconstrained firms. Finally, we investigate why some constrained firms have lower cash holdings despite the apparent benefits of holding more cash in these firms. On balance, our evidence supports the view that low internal resources and more costly external finance limit the available financial resources of low cash firms.

Specifically, we find that low cash, constrained firms have

significantly lower Altman’s Z-scores, interest coverage ratios, and cash flow margins than high cash, constrained firms. Moreover, the change in the cash flow margin over the prior three years is significantly lower for low cash, constrained firms than for high cash, constrained firms. Lowcash firms also issue equity less frequently and in lower quantities than the high cash firms. We find little evidence that the low cash holdings of some constrained firms are a byproduct of high agency cost firms wasting their cash reserves. Our study contributes to several other strands of the literature. First, it complements the evidence of Almeida, Campello, and Weisbach (2004) on the value of cash for constrained firms, and the sources of value increase. We demonstrate that, although more cash is associated with greater investment for both constrained and unconstrained firms, the marginal investment of constrained firms is associated with greater value increase than that of unconstrained firms. Second, we contribute to the debate in the cash holdings literature on whether high cash holdings are value-increasing.

Our results suggest that, for financially constrained firms, high cash

holdings are a value-increasing response to financing frictions. Third, our findings support the prediction of Myers and Majluf (1984) that firms facing financial constraints should retain cash and spend it later to avoid possible underinvestment.

4

The paper proceeds as follows. The next section summarizes the related literature and describes the potential contribution of our study.

Section 3 describes the data and the

methodology. Section 4 reports the main results and section 5 concludes.

2.

Related Literature Fazzari, Hubbard, and Petersen (1988) argue that when external financing is more costly

than internal financing, changes in cash flow will be an important determinant of marginal capital spending for constrained companies and the sensitivity of investment to cash flow will be increasing in the degree of financial constraints.

Although Fazzari et al. report evidence

consistent with their hypothesis, the interpretation of their findings has been challenged on both theoretical and empirical grounds.

Kaplan and Zingales (1997) question the validity of

investment-cash flow sensitivities as a measure of financial constraints. Using information from financial statements, they rank firms on the level of financial constraints and find that firms classified as less financially constrained actually exhibit greater investment-cash flow sensitivity. One explanation of the findings that they offer, favored also by Erickson and Whited (2000) and Alti (2003), is that Tobin’s Q is a noisy proxy for marginal Q. If cash flow contains information about investment opportunities and the profitability of assets in place, less constrained firms are more likely to adjust investment in response to shocks to investment opportunities. Thus, they have higher investment-cash flow sensitivity. Almeida, Campello, and Weisbach (2004) adopt an alternative approach to the question of whether costly external finance affects financial policies.

Rather than focusing on the

sensitivity of investment to cash flow, they focus on the cash flow sensitivity of cash. The intuition is that financially constrained firms should have a systematic propensity to save cash,

5

whereas unconstrained firms should not exhibit this propensity. Using several criteria for sorting firms into financially constrained and unconstrained, Almeida, Campello, and Weisbach (2004) find that the cash flow sensitivity of cash is positive for financially constrained firms, but statistically insignificant for financially unconstrained firms. Yet another strand of the literature studies the cross-sectional determinants of cash holdings. Opler, et al. (1999) find that cash holdings are negatively related to the level and the availability of a bond rating. That is, companies with a bond rating below investment grade and those having no bond rating available hold more cash than firms that have investment-grade bond rating. Similarly, Kim, Mauer, and Sherman (1998) and Harford (1999) report that cash holdings are positively associated with industry cash flow volatility. To the extent that firms with lower or no bond ratings and those that operate in industries with greater cash flow volatility face greater costs of external finance, the results support the view that financially constrained firms hold more cash than unconstrained firms. Although the above findings are consistent with higher cash holdings of constrained firms being a value-increasing response to costly external financing, another possibility is that constrained firms hold high cash reserves due to value-reducing agency problems and empirebuilding behavior by the managers. As noted in the introduction, several recent papers attempt to address the agency cost hypothesis and report mixed evidence. While studies by Harford (1999) and Harford, Mansi, and Maxwell (2006), Dittmar and Mahrt-Smith (2005), and Pinkowitz, Stulz, and Williamson (2006) report evidence consistent with the agency cost view, Mikkelson and Partch (2003) find no evidence that cash-rich firms perform any worse than firms with low cash holdings.

6

Our basic empirical approaches to measuring the value of cash are similar to those used in Faulkender and Wang (2006) and Pinkowitz and Williamson (2004). In addition to measuring the market valuation of cash holdings, however, we attempt to understand why cash is valued differently in constrained firms than in unconstrained firms.

Consequently, we investigate

whether the differential value of cash is related to investment policy, we analyze the marginal value of investment for constrained and unconstrained firms, and we investigate why some constrained firms have low cash holdings despite the apparent benefits of maintaining larger cash reserves.

3.

Sample Selection and Data Description The sample includes U.S. public companies with financial data available on Compustat’s

Industrial Annual P-S-T, Research, and Full Coverage files at any time over the period 1985 to 2002. The sample period is limited by the availability of Compustat data on bond and short-term debt ratings that are used to construct measures of financial constraints. We require firms to have at least $25 million in total book assets in 1994 dollars. Also, in order to eliminate the possible effects of regulations, we exclude companies in the financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. Finally we exclude firm-years with non-positive values for total book assets or cash holdings, or negative values for capital expenditures. These sample selection criteria result in 60,056 firm-year observations. The literature has proposed a variety of ways of identifying the level of financial constraints facing firms. However, there is no general agreement on which measure is the best proxy for financial constraints. As described below, we perform our analysis using several

7

alternative approaches for sorting firms into financially constrained and financially unconstrained.

1. Annual payout ratio. Fazzari, Hubbard, and Petersen (1988) argue that unconstrained firms are more likely to have higher payout ratios, while constrained firms are likely to have lower payout ratios. Therefore, each year, we assign those firms in the bottom (top) three deciles of the annual cash payout ratio distribution to the financially constrained (unconstrained) group. Following Almeida, Campello, and Weisbach (2004), payout ratio is defined as the ratio of dividends and common stock repurchases to operating income. Observations with a positive payout and zero or negative cash flow are assigned the highest payout ratio.

2. Average annual payout ratio. Alternatively, under the assumption that a firm’s level of financial constraints is constant through time, we assign firms in the bottom (top) three deciles of the average annual payout distribution to the financially constrained (unconstrained) group. Again, the payout ratio is equal to the ratio of dividends and common stock repurchases to operating income.

3. Dividend dummy variable. Under the assumption that constrained firms are unlikely to distribute cash in the form of dividends, we classify firms that pay dividends on common stock as financially unconstrained during the year. Firms that do not pay dividends are classified as financially constrained.

4. Debt rating. Following an approach similar to that in Whited (1992), Gilchrist and Himmelberg (1995), and Almeida, Campello, and Weisbach (2004), firms are classified as financially unconstrained if they have had their long-term debt rated by Standard & Poor’s (S&P Long-Term Senior Debt Rating is available on Compustat) and their debt is not in default (rating of “D” or “SD”). Firms are classified as constrained if they have debt outstanding that year, but have never had their public debt rated before (or long-term

8

debt rating is unavailable).

Firms with no debt outstanding are classified as

unconstrained.

5. Paper rating.

We construct a similar criteria using S&P Short-Term Debt Rating

available on Compustat. Firms are classified as financially unconstrained if they have had their short-term debt rated by Standard & Poor’s and their debt is not in default. Firms are classified as constrained if they have debt outstanding that year, but have never had their short-term debt rated before (or the rating is unavailable). Firms with no shortterm debt outstanding are classified as unconstrained.

6. KZ Index. Following Lamont, Polk, and Saa-Requejo (2001), the top 33% of firms ranked on the KZ index are classified as financially constrained, while the bottom 33% of firms ranked on the KZ index are classified as financially unconstrained. The KZ index is based on Kaplan and Zingales (1997) ordered logit regression and is calculated according to the following equation. KZ = -1.002*(Cash Flow/NetPPE)+0.283*MB+3.139*(Debt/Total Capital)-39.368*(Total Dividend/NetPPE)-1.315*(Cash holding/NetPPE)

Table 1 reports Pearson correlation coefficients for the alternative measures of financial constraints. All correlation coefficient are statistically significant at the 0.01 level and range from 0.07 to 0.80. The exception is the correlation between the bond rating measure and the KZ index, which has a correlation coefficient of -0.17. Although the high correlations imply that the measures are generally picking up similar information, it nonetheless appears that each measure picks up some unique information as well. We thus conduct our subsequent analysis using each of the alternative methods for classifying firms as financially constrained and unconstrained. Table 2 presents univariate comparisons of firm characteristics for sub-samples based on financial constraints. Consistent with the hypothesis that cash provides important benefits to

9

financially constrained firms, firms classified as financially constrained tend to hold more cash. For all classification criteria except the KZ index, the median and mean ratios of cash to total book assets are significantly higher for financially constrained firms. For the five measures excluding the KZ index, median cash holdings range from 3.4% of assets to 6.4% of assets for unconstrained firms and from 6.9% to 8.4% of assets for constrained firms. The pattern is reversed for the KZ index, where median cash holdings are 17.1% for unconstrained firms and 3.0% for constrained firms. The pattern in our sample is similar to that reported by Almeida, Campello, and Weisbach (2004). For the four main financial constraints criteria that they use, median cash holdings for firms in their sample ranges from 10.5% to 21.9% for the constrained firms and from 4.1% to 5.1% for the unconstrained. The slight difference in values might be due to the fact that Almeida, et al. limit their sample to manufacturing firms. The variation of cash holdings is higher for constrained firms than for unconstrained firms. The mean and the median standard deviation of the level of cash holdings and of the change in cash holdings are greater for financially constrained firms for all the criteria excluding the KZ-index. This is in line with the prediction of Myers and Majluf (1984) that firms facing higher costs of external financing should accumulate internal resources and use them later to finance investment. Constrained firms also have lower cash flow than do unconstrained firms. The average of the median ratios of cash flow to total assets is 8.6% for unconstrained firms and 6.9% for constrained firms. Firms that are classified as financially constrained are smaller and hold fewer tangible assets than unconstrained firms. Constrained and unconstrained firms have similar levels of leverage and net investment. The market-to-book ratio is slightly lower for the constrained firms, with the average of the median market-to-book of 2.02 for the unconstrained firms and 1.79 for the constrained firms.

10

Overall, our sample exhibits characteristics similar to those in prior studies. Specifically, we find that firms identified as likely to face financial constraints have lower cash flow, higher cash holdings, are smaller, and have more intangible assets than firms identified as unconstrained.

4.

Empirical Results In this section, we report our primary empirical results.

First, we investigate the

association between cash holdings and firm value for financially constrained and unconstrained firms using a variety of methodologies. Second, we investigate the effect of abnormal cash holdings on net investment and test whether the sensitivity of investment to cash holdings is different between financially constrained and financially unconstrained companies. Third, we compare the value of marginal investment for financially constrained and unconstrained firms. Finally, we examine alternative explanations for low cash holdings of some financially constrained firms.

4.1

The association between cash holdings and firm value If costly external finance constrains the firm from undertaking valuable investment

opportunities and cash holdings help alleviate this underinvestment, we expect a stronger positive association between cash holdings and firm value in firms that are more financially constrained. Alternatively, the positive association will be weaker or even negative for firms where cash holdings are primarily serving as a vehicle through which managers can overinvest.

11

4.1.1 Regressions of market-to-book ratio on cash holdings To test these hypotheses, we first estimate cross-sectional regressions of Tobin’s Q on cash holdings and a series of control variables. Tobin’s Q is calculated as total book assets plus market equity minus book equity over total book assets. The explanatory variables include cash holdings, firm size (measured as the log of book value of total assets), leverage, the ratio of research and development (R&D) to sales, and earnings before interest and taxes (EBIT). To test whether the association between firm value and cash holdings is stronger for financially constrained firms, we include the interaction of cash holdings with an indicator variable equal to one for firms with greater financial constraints. Henceforth, we refer to this indicator variable as the financial constraints dummy. The interaction term thus captures the difference in the effect of cash holdings on the value of the constrained and the unconstrained firms. We also include the financial constraints indicator as a separate explanatory variable to control for any differences in the valuation of constrained and unconstrained firms. The regressions are estimated using industry fixed effects to capture industry-specific unobservable factors that affect firm value and calendar-year dummy variables to capture any time-specific common variation in Q. Statistical significance is calculated using standard errors clustered at the firm level, which are robust to heteroskedasticity and autocorrelation. We obtain similar results if we follow the approach in Fama and French (1998) and compute statistical significance based on the time-series of coefficient estimates from annual firm value regressions. The results, reported in panel A of table 3, are presented for each of the six methods of identifying financially constrained and unconstrained firms. The coefficient on cash holdings is positive and statistically significant for each measure of financial constraints. The interaction term of cash and financial constraints dummy is also positive and statistically significant for each

12

measure except those based on debt ratings, indicating that the positive relation between cash holdings is significantly greater for firms classified as financially constrained. The difference between the value of cash for the constrained and the unconstrained firms is also economically important. Across the classification schemes with a greater positive relation for constrained firms, a dollar of cash is associated with 29 cents to $2 more value for constrained firms than for unconstrained firms. These estimates are in line with those of Pinkowitz and Williamson (2006). 4.1.2. Regressions of market-to-book on excess cash holdings One objection to using raw cash holdings in firm value regressions is that, because firm value is related to growth, cash and value can be endogenously determined by investment opportunities. Assume, for example, that the costs of financial constraints vary across firms and that internal cash holdings can effectively substitute for more costly external financing. If so, firms with the greatest costs of financial constraints will accumulate more cash. Because the costs of financial constraints are likely to be greatest for firms with high growth opportunities, high-cash firms will normally have higher values of Tobin’s Q than low-cash firms. Consequently, the positive relation between cash holdings and value may be spurious. Although the endogeneity bias described above seems unlikely to have a differential impact on constrained and unconstrained firms, we nonetheless control for possible endogeneity by estimating value regressions using a two-stage procedure similar to that in Harford (1999) and Mikkelson and Partch (2003). In the first stage, we follow Kim, Mauer, and Sherman (1998) and estimate the firm’s normal ratio of cash-to-total assets as a function of firm size, the market-tobook ratio, leverage, industry cash flow volatility (median of firm-level standard deviation of first differences in earnings before interest, depreciation and taxes over the prior twenty years), the duration of a firm’s cash cycle (average inventory age + average collection period – average

13

payment period), operating cash flow, the likelihood of financial distress, the return spread between a firm’s investments and the return on Treasury bills, and future economic conditions as measured by the log of the growth rate of industrial production.4 The model includes industry fixed effects, where industry is defined at the 2-Digit SIC level. The predicted value from the first-stage model thus represents a level of cash that is normal for a firm’s operations. Abnormal cash holdings are then computed as the difference between actual and predicted cash holdings, and are used in the second-stage regressions. Panel B of Table 3 reports the results from value regressions using the abnormal cash holdings variable.

As expected, the coefficients on abnormal cash holdings are lower in

magnitude than the coefficients on raw cash holdings from the previous set of regressions. Nonetheless, the coefficients on the interaction term of abnormal cash holdings and the financial constraints dummy are significantly positive for all measures of financial constraints except the KZ-index and the bond ratings criteria. In other words, controlling for potential endogeneity between Tobin’s Q and cash holdings, we still find a stronger positive association between value and cash for financially constrained firms for four of the six measures of financial constraints.5 Moreover, once we control for potential endogeneity, the economic magnitude of the difference between the sensitivity of value to cash for constrained and unconstrained firms increases. For 4

The use of the market-to-book ratio as a determinant of normal cash holdings presents a potential problem in that the market-to-book ratio is also a dependent variable in our second-stage regressions. (See also Dittmar and MahrtSmith (2007) for a discussion of this issue.) We note, however, that the first-stage uses the lagged market-to-book ratio whereas the second-stage uses the current ratio. Nonetheless, in robustness tests, we confirm that our conclusions are unaffected if we measure investment opportunities in the first stage using two-year lagged sales growth or the industry median market-to-book ratio. The only exception to this is that the coefficient on the interaction between abnormal cash holdings and the financial constraints dummy becomes statistically insignificant when the existence of a commercial paper rating is used as the proxy for financial constraints. All other results are qualitatively identical.

14

the classification schemes with a greater positive relation for constrained firms, a dollar of abnormal cash is associated with values that are 53 cents to $1.61 higher for constrained firms than for unconstrained firms. We conclude, therefore, that our finding that firm value is more strongly associated with cash holdings in financially constrained firms is not driven by the endogeneity of growth opportunities and cash holdings. 4.1.3. Regressions of excess stock returns on changes in cash holdings Faulkender and Wang (2006) develop a methodology that estimates the additional equity value resulting from changes in a firm’s cash position over the fiscal year.

Specifically,

Faulkender and Wang (2006) estimate a regression of annual abnormal stock return on the change in cash over the same year and several control variables. They interpret the coefficient on the change in cash as a measure of the value that investors put on a marginal dollar of cash. The control variables in the regression include change in book assets net of cash, change in earnings before interest and extraordinary items, change in R&D expenses, change in interest expenses, change in dividends, lagged cash holdings, leverage, and net financing during the fiscal year. All explanatory variables except leverage are standardized by lagged market equity. The dependent variable is excess stock return over the fiscal year, computed as the stock return over the fiscal year minus the return on a benchmark portfolio. The benchmark portfolios are 25 Fama-French value-weighted portfolios, constructed by independently sorting stocks on size and book-tomarket characteristics. To estimate the difference in the value of cash for constrained and unconstrained firms, we augment the Faulkender and Wang (2006) regressions with the financial constraints dummy 5

Almeida, Campello, and Weisbach (2004) also document that the KZ-index classification produces results that are at odds with those obtained using alternative measures of constraints. For this reason, Faulkender and Wang (2006) do not use the KZ-index as a measure of financial constraints.

15

and an interaction of change in cash with the dummy. The interaction measures the difference in the value of cash between constrained firms and unconstrained firms. The results, reported in Table 4, indicate that cash is more valuable for constrained firms. The interaction of the change in cash with the financial constraints dummy is positive for five of the six criteria (excluding the KZ-index), and statistically significant for four of the five positive cases (excluding the payout ratio). In those cases for which the interaction term is statistically significant, the coefficient estimates imply that the marginal value of cash is between 21 cents and 51 cents higher in constrained firms than in unconstrained firms. These estimates are consistent with the economic magnitudes reported in Faulkender and Wang (2006).

Overall, therefore, the evidence is fairly robust that cash is positively associated with firm value and that this association is stronger in firms that are likely to be financially constrained. In this sense, our findings support the view that a policy of greater cash retention by constrained firms, as documented by Almeida, Campello, and Weisbach (2004) and Opler, Pinkowitz, Stulz, and Williamson (1999), is more likely to be a value-increasing activity.

4.2

The effect of cash holdings on investment As noted earlier, the above findings are consistent with two interpretations. Greater cash

holdings might be more valuable to constrained firms because they allow the firm to invest when other sources of funds are costly, limited, or unavailable. In other words, greater cash holdings allow firms that experience external financial constraints to avoid underinvestment and reduced growth. Alternatively, however, it might be the case that the higher value associated with greater

16

cash holdings is a reflection of the market rewarding the firm for holding cash rather than overinvesting that cash in unprofitable projects.

To distinguish between these alternative

interpretations, we examine whether greater cash holdings are associated with greater investment and whether this association is stronger for constrained than for unconstrained firms. Table 5 reports the results of regressions of the determinants of investment. Following Lang, Ofek, and Stulz (1996), we measure investment expenditures as net investment, defined as capital expenditures net of depreciation. Net investment is deflated by total book assets. The regressions are estimated for next year’s net investment. Our primary independent variables of interest are abnormal cash holdings and the interaction between abnormal cash and a financial constraints indicator variable. The interaction thus measures whether the association between investment and cash is different between firms that are classified as financially constrained and those that are classified as unconstrained. To control for other possible determinants of investment, the regression model includes leverage, the market-to-book ratio, the R&D-to-sales ratio, the log of book value of total assets, fixed assets intensity, cash flow deflated by total book assets, the log of growth in sales for the previous two years, the sale of PPE deflated by total book assets, and dummy variables for equity and debt issues. Consistent with the predictions of Myers (1977), Lang, Ofek, and Stulz (1996) find that leverage is negatively related to future growth in low Q firms. The market-to-book ratio is used to control for growth opportunities. Larger and more established companies tend to be older, with fewer growth opportunities. Additionally, firm size may be related to potential agency problems, analyst coverage, and monitoring by the market for corporate control. We use the ratio of property, plant, and equipment to total book assets to control for asset tangibility, since firms with greater tangibility should rely more on investment in fixed assets. To control for

17

the availability of alternative internal funds, regressions include cash flow net of interest expenses divided by total assets.6 More financially constrained firms should exhibit a greater reliance on internal financing (e.g., Fazzari, Habbard, and Petersen (1988)). We allow for different coefficients on cash flow for constrained and unconstrained firms by including an interaction term of cash flow and financial constraints dummy. We include sales growth over the past two years into the regression model to allow for the multiplier effect of growth.7 Companies that exhibited higher growth in the past periods are likely to invest more and have higher subsequent growth. Indicator variables for equity and debt issues during year t and year t+1 are used to control for any effects that a debt or an equity issue, and the resources they provide to the company, can have on investment. We also include the value of PPE sales to control for cash inflow from equipment liquidations. Similar to the value regressions, investment regressions are estimated using industry fixed effects and include calendar-year dummies.

Statistical

significance is measured using robust standard errors that are clustered at the firm level. The results reported in table 5 indicate that the following year’s investment is positively related to current cash holdings.8 However, the coefficient on the interaction of abnormal cash and a financial constraints dummy is statistically insignificant. This evidence implies that firms with higher cash holdings invest more regardless of whether they are constrained or unconstrained. Although these findings are consistent with the idea that cash holdings allow constrained firms to increase investment, the results do not explain the higher value of cash for 6

Lang, Ofek, and Stulz (1996) argue that cash flow net of interest expense might proxy for a firm’s capital structure, as firms with higher leverage have higher interest expenses.

7

It is possible that earlier period’s growth on the right hand side is endogenously related to the left hand side investment variable. We use sales growth as it is least likely to be affected by endogeneity. We perform additional tests with alternative measures of prior growth, and this does not change the results.

8

Results are similar when the current year’s net investment is used as the dependent variable.

18

constrained firms, since the effect of cash on investment is insignificantly different between constrained and unconstrained firms. Therefore, in the next section, we analyze whether the profitability of investment is different for constrained and unconstrained firms.

4.3

Do greater cash holdings lead to more profitable investment? To examine the marginal profitability of investment for constrained and unconstrained

firms, we follow the valuation regression procedure of Fama and French (1998). The FamaFrench model relates firm value to a set of firm characteristics (earnings, assets, R&D, interest expense, dividends) and changes in these characteristics. Changes in variables are calculated over years -1 to 0 and 0 to +1 relative to the current year.9

Like Pinkowitz, Stulz, and

Williamson (2006), we modify the Fama-French specification by splitting the total change in assets into its cash and non-cash components. All variables are deflated by contemporaneous total book assets to avoid potential heteroskedasticity due to differences in firm size. The regressions include calendar-year dummies and are estimated using industry fixed effects. We obtain similar results if the regressions are estimated using the Fama and MacBeth (1973) methodology. Defining firm value as the total market value of the firm minus the book value of its assets, all divided by the book value of the firm’s assets (effectively, Tobin’s Q minus one), we estimate two sets of regressions. In the first set, firm value is the dependent variable, whereas the change in firm value is the dependent variable in the second set. The first set, which we refer to as ‘level’ regressions is intended to capture the expected effects of the explanatory variables on

9

Results are similar when changes are calculated over year -2 to 0 and 0 to +2 relative to the current year.

19

firm value. The second set, which we refer to as ‘change’ regressions is intended to capture unexpected effects. Our primary variables of interest are the coefficients on the non-cash asset variables. Fama and French (1998) interpret these coefficients as capturing the impact of investment on firm value.10 Thus, to test whether the marginal value of investment differs for constrained and unconstrained firms, we allow the coefficient on non-cash assets to differ for constrained and unconstrained firms. We also include the financial constraints indicator as a separate explanatory variable to control for any differences in the valuation of constrained and unconstrained firms. For these tests, we report only the results using average payout and bond ratings criteria for identifying constrained and unconstrained firms. However, with the exception of the results using the KZ-index noted below, our results are qualitatively the same using other measures of financial constraints. The results reported in Table 6 indicate that both prior and future changes in non-cash assets are positively related to firm value – i.e. the marginal value of investment is positive. More importantly, we find that the interaction between the financial constraints indicator variable and the change in non-cash assets is also significantly positive. The lone exception is the coefficient on the interaction of the prior change in non-cash assets and the financial constraints variable in the change regression using average payout as the measure of financial constraints.11

10

Note that the Fama-French measure of investment includes the combined effect of net capital expenditures and divestitures whereas in Table 5, we treat net capital expenditures and asset sales separately. 11

The only financial constraints criterion that produces negative coefficients on the interaction terms is the KZindex. This is in line with our earlier findings and with the findings of Almeida, Campello, and Weisbach (2004) who also find that the KZ-index produces results that are at odds with those produced by other financial constraints criteria.

20

These findings are consistent with the view that that the marginal value of investment is greater for financially constrained than for unconstrained firms. The significantly greater impact of investment on value for constrained firms implies that, although greater cash holdings are associated with greater investment for both constrained and unconstrained firms, the financially constrained firms appear to have more valuable marginal investment opportunities.12 Thus, collectively our findings suggest that cash is more valuable to constrained firms because (i) it allows them to increase investment, and (ii) this marginal investment is more valuable than that of unconstrained firms.

4.4

Why do some constrained firms have low cash holdings? Although our findings suggest that cash holdings are more valuable to constrained firms

because they allow them to increase investment in value-increasing projects, many constrained firms have surprisingly low cash reserves. For example, the median ratio of cash holdings to total assets for constrained firms in the first quartile of abnormal cash holdings is 0.021, as opposed to the sample median of 0.058. Constrained firms at the 25th percentile of abnormal cash holdings have average abnormal cash holdings scaled by total assets equal to -0.076. It is puzzling why some constrained firms hold so little cash when there appears to be such a large benefit from greater cash holdings. We consider three (not mutually exclusive) possible explanations for this puzzling behavior. One possibility is that low cash holdings are a result of potential or actual agency problems. If some managers tend to invest inefficiently, their boards of directors might limit the 12

Note that even after accounting for contemporaneous non-cash asset growth, cash holdings continue to have a positive impact on value. A plausible explanation is that higher cash holdings are associated greater current and future investment.

21

amount of cash at the disposal of managers in order to avoid the inefficient spending. Alternatively, entrenched managers might routinely waste cash reserves, leaving the firm with low cash holdings. A second possibility is that low cash, constrained firms exhibit weaker financial health such that they are either unable to accumulate cash reserves or are forced to draw down on previous balances of cash. Finally, it is possible that the costs of external finance are lower for some constrained firms than for others. If so, and if we measure ‘normal’ cash holdings with error, the optimal cash holdings for firms with lower costs of external finance will simply be lower. We examine these possibilities by comparing the characteristics of low cash, constrained firms with high cash, constrained firms. Specifically, we partition the sample of constrained firms according to their abnormal cash holdings, where firms with the ratio of abnormal cash holdings to total assets below (above) the median are classified as low (high) cash firms. To test the agency explanation, we measure managerial incentives as the percentage of equity (common stock + options) held by the top executive officer (CEO ownership) and by top executive officers (insider ownership) as reported in S&P’s Execucomp database.13 We also test whether the low cash firms invest more than their high cash counterparts. As reported in Table 7, CEO ownership and insider ownership are of similar magnitude for low and high cash constrained firms across all financial constraints criteria.14 A Wilcoxon rank-sum test fails to reject the equality of the distributions of CEO ownership for the low and

13

Because S&P Execucomp data starts in 1992 and includes a limited number of firms, we check whether the main results of the paper hold for the firms with ownership data available from Execucomp. The results are similar to those obtained for the full sample. 14

We report average values of all characteristics in Table 7. Because we find qualitatively identical results using median values, we do not report medians in the table, but instead conservatively report statistical significance by using the non-parametric Wilcoxon test.

22

high cash sub-samples for four out of the six financial constraints criteria. It thus appears that managers of low and high cash constrained firms have similar incentives. Furthermore, the average net investment is lower for low cash firms than for high cash constrained firms and the Wilcoxon test rejects the equality of the distributions for all constraints criteria. Thus, there is little support for the view that low cash constrained firms exhibit greater incentive problems or agency problems of overinvestment than do high cash constrained firms. There is evidence, however, that the low cash, constrained firms exhibit weaker financial health than the high cash constrained firms. Average operating cash flow is lower for the low cash firms, as is the average growth in the cash flow margin. Moreover, low cash, constrained firms exhibit significantly lower Altman’s Z-scores and interest coverage ratios.

Thus, a

plausible explanation for the low cash holdings of some constrained firms is that poor financial performance has drained prior cash reserves and/or prevented the firm from maintaining the optimal level of cash. Finally, the evidence in Table 7 indicates that the volume of external debt and equity issues is lower in low cash, constrained firms. We measure the value of debt issues as long-term debt issuance net of long-term debt reduction over total book assets. The value of equity issues is calculated as the one-year change in the number of shares outstanding (adjusted for stock splits) multiplied by the fiscal year close price.15 Using these measures, low cash firms issue less debt and less equity as a fraction of total book assets than high cash firms. Moreover, using the number of annual analyst recommendations for a firm to measure analyst coverage,16 we find that

15

Results are similar when the value of equity issues is calculated as the change in Treasury stock multiplied by the fiscal year close price. 16

We obtained the data on analyst recommendations from I/B/E/S database. Analyst coverage information is available for 24,840 firm-year observations.

23

low cash firms receive less analyst coverage than high cash firms. Under the assumption that firms with lower analyst coverage face greater information asymmetries between managers and outside investors, these findings imply that low cash firms face higher costs of external financing. 17 To summarize, the abnormally low cash holdings of some constrained firms cannot be explained by incentive problems, agency costs of overinvestment, or lower costs of external finance. Rather, the low cash balances appear to be due primarily to these firms being in weaker financial condition than high cash, constrained firms. This appears to inhibit their ability to maintain higher cash balances.

5.

Conclusions Our analysis yields three primary results. First, higher cash holdings are associated with

higher firm value and this positive association is significantly stronger for financially constrained firms than for unconstrained firms. Second, higher cash holdings are associated with higher levels of investment for both constrained and unconstrained firms. Third, there is a stronger positive association between investment and value for constrained than for unconstrained firms. We interpret these findings as consistent with the view that higher cash holdings of financially constrained firms are a value-increasing response to costly external financing. That is, higher cash holdings allow the firm to undertake positive net present value projects that would otherwise have been bypassed. Our findings are inconsistent with the view that constrained firms maintain higher cash balances to facilitate empire-building overinvestment.

17

Similarly, our

Brennan and Subrahmanyan (1995) and Chung and Jo (1996), among others, report evidence consistent with the hypothesis that analysts disseminate firm-specific information to investors.

24

findings do not support the hypothesis that the higher value of cash for constrained firms is a reflection of the market rewarding the firm for not overinvesting the funds. Our findings complement and extend those in Almeida, Campello, and Weisbach (2004) and Faulkender and Wang (2006). While their findings indicate that constrained firms are more likely to save cash out of their cash flow, our findings show that this behavior is value-increasing because it allows the firms to take on valuable investment opportunities. Together, the results of our study and that of Almeida, Campello, and Weisbach (2004) support the view that firms can mitigate the adverse effects of financial constraints by adopting a policy of greater cash retention. In this sense, our findings also have implications for the literature on corporate payout policy. In the presence of costly external finance, payout policy is relevant to stockholders because it impacts the ability of firms to undertake all future positive net present value projects. A payout that is too high constrains investment while a payout that is too low potentially creates agency problems of free cash flow. Our findings thus provide indirect support for life-cycle based theories of optimal payout policy in which younger, higher-growth firms limit dividends in order to conserve cash, while more mature, lower-growth firms increase payouts to mitigate overinvestment problems (see, for example, DeAngelo and DeAngelo (2006)).

25

References Almeida, Heitor, Murillo Campello, and Michael S. Weisbach, 2004, The Cash Flow Sensitivity of Cash, Journal of Finance 59, 1777-1804. Alti, Aydogan, 2003, How Sensitive is Investment to Cash Flow When Financing is Frictionless?, Journal of Finance 58, 707-722. Brennan, Michael J., and Avanidhar Subrahmanyam, 1995, Investment Analysis and Price Formation in Securities Markets, Journal of Financial Economics 38, 361-381. Chung, Kee H., and Hoje Jo, 1996, The Impact of Security Analysts' Monitoring and Marketing Functions on the Market Value of Firms, Journal of Financial and Quantitative Analysis 31, 493-512. DeAngelo, Harry and Linda DeAngelo, 2006, The irrelevance of the MM dividend irrelevance theorem, Journal of Financial Economics 79, 293-316. Dittmar, Amy, and Jan Mahrt-Smith, 2007, Corporate Governance and the Value of Cash Holdings, Journal of Financial Economics 83, 599-634. Dittmar, Amy, Jan Mahrt-Smith, and Henri Servaes, 2003, International Corporate Governance and Corporate Cash Holdings, Journal of Financial and Quantitative Analysis 38. 111133. Erickson, Timothy, and Toni M. Whited, 2000, Measurement Error and the Relationship between Investment and Q, Journal of Political Economy 108, 1027-1057. Fama, Eugene F.; and Kenneth R. French, 1998, Taxes, Financing Decisions, and Firm Value, Journal of Finance 53, 819-843. Faulkender, Michael, and Rong Wang, 2006, Corporate Financial Policy and the Value of Cash, Journal of Finance 61, 1957-1990. Fazzari, Steven, R. Glenn Hubbard, and Bruce Petersen, 1988, Financing Constraints and Corporate Investment, Brooking Papers on Economic Activity 1, 141-195. Gilchrist, Simon and Charles P. Himmelberg, 1995, Evidence on the Role of Cash Flow for Investment, Journal of Monetary Economics 36, 541-572. Greenwald, Bruce, Joseph E. Stiglitz, and Andrew Weiss, 1984, Informational Imperfections in the Capital Market and Macroeconomic Fluctuations, The American Economic Review 74, 194-199. Harford, Jarrad, 1999, Corporate Cash Reserves and Acquisitions, Journal of Finance 54, 19691997.

26

Harford, Jarrad, Sattar A. Mansi, and William F. Maxwell, 2006, Shareholder Rights and Corporate Cash Holdings, AFA 2006 Boston Meetings Paper. Hubbard, R. Glenn, 1998, Capital-Market Imperfections and Investment, Journal of Economic Literature 36, 193-225. Kaplan, Steven, and Luigi Zingales, 1997, Do Financing Constraints Explain Why Investment is Correlated with Cash Flow? Quarterly Journal of Economics 112, 169-215. Kim, Chang-Soo, David Mauer, and Ann Sherman, 1998, The Determinants of Corporate Liquidity: Theory and Evidence, Journal of Financial and Quantitative Analysis 33, 335359. Lamont, Owen, Christopher Polk, and Jesus Saá-Requejo, 2001, Financial Constraints and Stock Returns, Review of Financial Studies 14, 529-554. Lang, Larry, Eli Ofek, and René Stulz, Leverage, 1996, Investment and Firm Growth, Journal of Financial Economics 40, 3-29. McConnell, John J. and Henri Servaes, 1990, Additional Evidence on Equity Ownership and Corporate Value, Journal of Financial Economics 27, 595-612. Mikkelson, Wayne, and Megan Partch, 2003, Do Persistent Large Cash Reserves Hinder Performance?, Journal of Financial and Quantitative Analysis 38, 275-294. Modigliani, Franco and Merton Miller, 1958, The Cost of Capital, Corporation Finance and The Theory of Investment, The American Economic Review 53, 261-97. Morck, Randall, Andrei Shleifer, and Robert W. Vishny, 1988, Management Ownership and Market Valuation : An Empirical Analysis, Journal of Financial Economics 20, 293-315. Myers, Stewart C., 1977, Determinants of Corporate Borrowing, Journal of Financial Economics 5, 147-175. Myers, Stewart C., 1984, The Capital Structure Puzzle, Journal of Finance 39, 575-592. Myers, Stewart and Nicholas Majluf, 1984, Corporate Financing and Investment Decisions when Firms Have Information that Investors Do Not Have, Journal of Financial Economics 13, 187-221. Opler, Tim, Lee Pinkowitz, René Stulz, and Rohan Williamson, 1999, The Determinants and Implications of Corporate Cash Holdings, Journal of Financial Economics 52, 3-46. Pinkowitz, Lee, Rene M. Stulz, and Rohan Williamson, 2006, Does the Contribution of Corporate Cash Holdings and Dividends to Firm Value Depend on Governance? A CrossCountry Analysis, Journal of Finance, 61, 2725-2752.

27

Pinkowitz, Lee, and Rohan Williamson, 2004, What is a Dollar Worth? The Market Value of Cash Holdings, working paper series, Georgetown University. Stein, Jeremy C., 1996, Rational Capital Budgeting in an Irrational World, Journal of Business, 69, 429-455. Stein, Jeremy C., 2001, Agency, Information and Corporate Investment, in Handbook of the Economics of Finance, edited by George Constantinides, Milt Harris and René Stulz. Amsterdam: North-Holland. Whited, Toni M., 1992, Debt, Liquidity Constraints, and Corporate Investment: Evidence from Panel Data, Journal of Finance 47, 1425-1460.

28

Table 1 Correlation of Various Classifications of Financial Constraints Data from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with non-positive assets or cash holdings, or negative capital expenditures. See text for definitions of criteria used to categorize firm-years as financially constrained or unconstrained. Correlation coefficient estimates reported with p-values in parentheses below.

Payout Ratio

Average Payout

Financial Constraints Criteria Dividends Bond Ratings

Paper Ratings

Payout ratio (p-value)

1

Average payout (p-value)

0.7748 (0.00)

1

Dividends

0.8061 (0.00)

0.6542 (0.00)

1

(p-value) Bond ratings (p-value)

0.1882 (0.00)

0.2537 (0.00)

0.2106 (0.00)

1

Paper ratings (p-value)

0.3054 (0.00)

0.2846 (0.00)

0.2895 (0.00)

0.4566 (0.00)

1

KZ index

0.3880 (0.00)

0.3748 (0.00)

0.3026 (0.00)

-0.1694 (0.00)

0.0792 (0.00)

(p-value)

29

KZ Index

1

Table 2 Univariate Comparison by Financial Constraints Data from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with non-positive assets or cash holdings, or negative capital expenditures. See text for definitions of financial constraints criteria. Size is total book assets measured in 1994 dollars (adjusted for the CPI). Cash holdings, property, plant, and equipment, total debt, net investment, and cash flow are deflated by total book assets. Financial Constraints Criteria

Payout Uncon. Constr. (A) (B)

Avg. Payout Uncon. Constr. (A) (B)

Dividends Uncon. Constr. (A) (B)

Bond Ratings Uncon. Constr. (A) (B)

Paper Ratings Uncon. Constr. (A) (B)

KZ Index Uncon. Constr. (A) (B)

Cash holdings

Mean Median

0.114 0.052

0.167 0.079

0.142 0.064

0.186 0.079

0.099 0.048

0.176 0.084

0.083 0.040

0.171 0.081

0.063 0.034

0.155 0.069

0.240 0.171

0.067 0.030

St dev cash

Mean Median

0.066 0.051

0.087 0.075

0.076 0.057

0.083 0.069

0.059 0.046

0.088 0.075

0.058 0.044

0.081 0.066

0.043 0.037

0.078 0.061

0.099 0.088

0.059 0.042

St dev cash chg.

Mean Median

0.099 0.051

0.166 0.087

0.130 0.060

0.168 0.081

0.079 0.047

0.172 0.090

0.089 0.047

0.146 0.073

0.043 0.034

0.139 0.069

0.189 0.095

0.089 0.049

Cash flow

Mean Median

0.062 0.084

0.021 0.064

0.042 0.076

0.019 0.066

0.077 0.090

0.021 0.064

0.062 0.075

0.038 0.078

0.101 0.103

0.040 0.074

0.054 0.092

0.033 0.065

Net investment

Mean Median

0.016 0.004

0.024 0.002

0.021 0.004

0.038 0.008

0.030 0.011

0.029 0.003

0.024 0.007

0.032 0.008

0.020 0.012

0.030 0.007

0.015 0.005

0.032 0.005

Size

Mean Median

3811 483

739 153

2757 330

581 103

3926 504

726 137

5226 1179

926 120

12150 4736

1143 169

1963 223

2020 243

Total debt

Mean Median

0.292 0.245

0.298 0.237

0.290 0.227

0.278 0.217

0.285 0.243

0.288 0.226

0.416 0.354

0.231 0.173

0.294 0.268

0.281 0.223

0.145 0.083

0.406 0.382

PPE

Mean Median

0.364 0.300

0.309 0.232

0.319 0.251

0.300 0.225

0.369 0.311

0.300 0.221

0.395 0.341

0.297 0.230

0.414 0.360

0.315 0.246

0.193 0.158

0.445 0.405

Market-to-book

Mean Median

3.968 1.915

3.812 1.761

4.060 1.994

3.256 1.908

3.089 1.821

3.993 1.809

4.861 1.972

3.341 1.805

3.972 2.305

3.714 1.803

3.228 2.132

3.966 1.639

30

Table 3 Determinants of Firm Value Data from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with non-positive assets or cash holdings, or negative capital expenditures. See text for definitions of financial constraints criteria. The dependent variable is Tobin’s Q, which is equal to total book assets minus book equity plus market equity over total book assets. Size is the log of total book assets measured in 1994 dollars (adjusted for the CPI). Total debt, property, plant, and equipment, and EBIT are deflated by total book assets. Cash is cash and short-term investments deflated by total book assets. Abnormal cash is residual cash from Kim, Mauer, and Sherman’s (1998) model. The FC dummy takes a value of one if a company is identified as financially constrained and zero if a company is identified as financially unconstrained by the respective criterion. Regressions are estimated using industry fixed effects and include calendar-year dummies. Industry is defined at the 2-Digit SIC level. Presented R-squared excludes fixed effects. Statistical significance is computed using heteroskedasticity and autocorrelation robust standard errors that are clustered at the firm level. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Coefficient estimates are reported with t-statistics in parentheses below. Panel A

Cash Cash X FC dummy Total debt PPE Size RD-to-sales EBIT

Payout Ratio

1.945*** (12.15)

1.665*** (13.13) 0.390*** (2.91) -0.035 (0.71) -0.405*** (-7.80) 0.043*** (5.77) 0.671*** (11.40) 3.002*** (25.17) 0.008 (0.42)

1.571*** (13.21) 0.528*** (3.72) -0.117* (-1.92) -0.375*** (-6.08) 0.025*** (2.80) 0.552*** (9.51) 2.858*** (21.96) -0.037 (-1.33)

39685 0.210

34468 0.192

-0.139** (-2.65) -0.443*** (-7.53) 0.015 (1.64) 0.678*** (7.36) 3.158*** (15.89)

FC dummy

Observations Adjusted R-sq

Financial Constraints Criteria Average Payout Dividends

All Firms

60056 0.206

31

Bond Ratings

Paper Ratings

KZ Index

1.736*** (15.03) 0.292** (2.38) -0.142*** (-3.12) -0.441*** (-9.49) 0.018*** (2.88) 0.665*** (13.30) 3.163*** (28.72) 0.001 (0.03)

1.973*** (11.06) -0.034 (-0.18) -0.119*** (-2.63) -0.446*** (-9.59) 0.022*** (3.31) 0.679*** (13.56) 3.156*** (28.77) 0.048** (2.18)

1.939*** (4.53) 0.016 (0.04) -0.131*** (-2.87) -0.443*** (-9.55) 0.006 (0.98) 0.677*** (13.62) 3.163*** (28.85) -0.086** (-2.55)

1.666*** (16.71) 2.012*** (8.63) -0.566*** (-10.83) -0.328*** (-5.70) 0.047*** (6.06) 0.705*** (11.59) 3.481*** (25.75) 0.029 (1.09)

60056 0.206

60056 0.206

60056 0.206

33440 0.245

Table 3 – continued Panel B

Abnormal cash Abnormal cash X FC dummy Total debt PPEA Size RD-to-sales EBIT

All Firms

Payout Ratio

-0.083 (-0.68)

-0.358** (-2.38) 0.536*** (3.13) -0.293*** (-5.33) -0.623*** (-11.21) 0.032*** (4.38) 1.086*** (17.12) 2.567*** (20.21) 0.074*** (3.80)

-0.497*** (-3.42) 0.746*** (3.63) -0.326*** (-4.77) -0.566*** (-8.43) 0.027*** (3.01) 1.127*** (14.56) 2.643*** (17.36) 0.033 (1.14)

34022 0.145

25336 0.151

-0.340*** (-3.90) -0.650*** (-6.11) 0.016** (2.01) 1.248*** (14.16) 2.934*** (14.66)

FC dummy

Observations Adjusted R-sq

Financial Constraints Criteria Average Payout Dividends

46740 0.159

32

Bond Ratings

Paper Ratings

KZ Index

-0.511*** (-3.63) 0.638*** (3.91) -0.344*** (-6.93) -0.641*** (-12.89) 0.025*** (3.94) 1.226*** (19.08) 2.935*** (23.34) 0.080*** (4.42)

-0.080 (-0.52) -0.003 (-0.02) -0.336*** (-6.72) -0.650*** (-12.99) 0.018*** (2.65) 1.248*** (19.52) 2.934*** (23.30) 0.008 (0.41)

-1.611*** (-3.89) 1.617*** (3.88) -0.328*** (-6.60) -0.653*** (-13.07) 0.012* (1.78) 1.242*** (19.50) 2.926*** (23.28) -0.002 (-0.08)

0.155 (1.29) -1.869*** (-7.57) -0.635*** (-11.31) -0.438*** (-7.44) 0.029*** (4.04) 1.065*** (16.10) 2.971*** (21.03) -0.142*** (-5.42)

46740 0.161

46740 0.159

46740 0.160

29022 0.187

Table 4 Excess Stock Returns and Changes in Cash Holdings Data are from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with non-positive assets or cash holdings, or negative capital expenditures. See text for definitions of financial constraints criteria. The dependent variable is stock return over fiscal year minus the return on a benchmark portfolio. The benchmark portfolios are 25 Fama-French value-weighted portfolios. The control variables include change in book assets net of cash, change in earnings before interest and extraordinary items, change in R&D expenses, change in interest expenses, change in dividends, lagged cash holdings, leverage, and net financing during fiscal year. All explanatory variables except leverage are standardized by lagged market equity. The FC dummy takes a value of one if a company is identified as financially constrained and zero if a company is identified as financially unconstrained by the respective criterion. Regressions are estimated using OLS. Statistical significance is computed using heteroskedasticity and autocorrelation robust standard errors. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Coefficient estimates are reported with t-statistics in parentheses below.

Change in cash Constrained*change in cash Change in non-cash assets Change in earnings Change in R&D Change in interest expense Change in dividends Net financing Leverage Lagged cash Change in cash * lagged cash Change in cash * leverage FC dummy

Observations Adjusted R-sq

Financial Constraints Criteria Bond Dividends Ratings

Payout Ratio

Average Payout

1.114*** (7.67) 0.076 (1.15) 0.189*** (12.70) 0.331*** (13.67) 0.655*** (2.91) -1.643*** (9.83) 3.304*** (11.64) -0.052* (1.75) -0.007 (1.39) 0.287*** (10.16) -0.284** (2.58) -0.046*** (2.86) -0.002 (0.28)

1.156*** (10.27) 0.212** (2.35) 0.196*** (12.22) 0.353*** (11.84) 0.891*** (4.97) -1.650*** (7.47) 3.124*** (8.11) -0.051 (1.56) -0.011** (2.27) 0.266*** (9.13) -0.366*** (3.90) -0.067*** (3.55) 0.013 (1.41)

0.946*** (9.10) 0.249*** (3.74) 0.185*** (14.34) 0.361*** (15.20) 0.756*** (4.72) -1.604*** (10.22) 3.657*** (14.17) -0.053** (2.17) -0.004 (0.77) 0.259*** (10.28) -0.288*** (2.82) -0.048*** (3.35) -0.004 (0.53)

21947 0.156

17339 0.165

30749 0.153

33

Paper Ratings

KZ Index

1.020*** (10.34) 0.133** (2.14) 0.187*** (13.39) 0.362*** (14.85) 0.842*** (5.23) -1.672*** (10.70) 3.618*** (14.06) -0.039 (1.66) -0.005 (1.20) 0.261*** (10.32) -0.288*** (2.98) -0.048*** (3.69) -0.013** (2.06)

0.629*** (4.30) 0.511*** (3.86) 0.187*** (13.41) 0.361*** (14.79) 0.841*** (5.34) -1.682*** (10.73) 3.583*** (14.01) -0.039 (1.67) -0.005 (1.06) 0.268*** (10.39) -0.281*** (2.93) -0.053*** (3.91) -0.037*** (5.04)

1.181*** (8.58) -0.083 (0.93) 0.178*** (13.72) 0.344*** (16.88) 0.461** (2.25) -1.589*** (9.62) 4.015*** (10.65) -0.048* (1.83) 0.002 (0.36) 0.236*** (8.50) -0.274** (2.26) -0.049** (2.60) -0.030*** (4.00)

31365 0.155

31365 0.156

19250 0.155

Table 5 Determinants of Net Investment Data from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with nonpositive assets or cash holdings, or negative capital expenditures. See text for definitions of financial constraints criteria. The dependent variable is net investment, defined as capital expenditures net of depreciation in year (t+1). Size is the log of total book assets measured in 1994 dollars (adjusted for the CPI). Total debt, property, plant, and equipment, and cash flow are deflated by total book assets. Prior growth is measured as growth in sales for the past two years. Cash is residual cash from Kim, Mauer, and Sherman (1998) model. The FC dummy takes a value of one if a company is identified as financially constrained and zero if a company is identified as financially unconstrained by the respective criterion. PPE sales is the sale of PPE over total book assets. Regressions are estimated using industry fixed effects and include calendar-year dummies. Industry is defined at the 2-Digit SIC level. Statistical significance is computed using heteroskedasticity and autocorrelation robust standard errors that are clustered at the firm level. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Coefficient estimates are reported with t-statistics in parentheses below.

Payout Ratio Cash Cash X FC dummy Total debt Mrkt-to-book RD-to-sales Size PPEA Cash flow Cash flow X FC dummy Prior sales growth Equity issue Debt issue PPE sales FC dummy

Observations. Adj. R-sq

Financial Constraints Criteria Average Bond Payout Dividends Ratings

Paper Ratings

KZ Index

0.046*** (6.63) -0.010 (-1.22) -0.041*** (-14.33) 0.009*** (15.33) 0.000 (0.19) 0.000 (1.31) 0.062*** (14.26) 0.036*** (4.77) 0.041*** (4.61) 0.006*** (8.05) 0.010*** (11.06) 0.030*** (25.75) 0.433*** (8.30) 0.002 (1.45)

0.053*** (7.98) -0.015 (-1.59) -0.039*** (-10.89) 0.009*** (12.25) 0.002 (0.66) 0.000 (0.40) 0.067*** (11.61) 0.031*** (3.86) 0.074*** (6.54) 0.007*** (7.18) 0.011*** (9.41) 0.031*** (21.74) 0.511*** (7.70) 0.001 (0.48)

0.042*** (6.98) -0.003 (-0.41) -0.041*** (-15.29) 0.010*** (16.97) 0.000 (0.15) -0.001*** (-4.00) 0.068*** (16.88) 0.058*** (6.65) 0.016* (1.65) 0.007*** (9.46) 0.011*** (13.12) 0.030*** (29.87) 0.458*** (9.37) 0.000 (0.20)

0.043*** (5.67) -0.005 (-0.54) -0.041*** (-14.57) 0.010*** (16.98) 0.001 (0.59) -0.001*** (-3.90) 0.067*** (16.88) 0.034*** (3.25) 0.045*** (4.02) 0.007*** (9.60) 0.010*** (13.00) 0.030*** (29.88) 0.455*** (9.36) -0.005*** (-2.65)

0.049*** (3.54) -0.010 (-0.71) -0.041*** (-15.24) 0.010*** (17.08) 0.000 (0.12) -0.001*** (-3.32) 0.067*** (16.84) 0.016 (1.14) 0.055*** (3.90) 0.007*** (9.56) 0.010*** (12.92) 0.030*** (29.86) 0.458*** (9.42) -0.005** (-1.96)

0.042*** (9.67) -0.011 (-0.83) -0.040*** (-10.49) 0.009*** (13.88) -0.004* (-1.80) -0.001*** (-3.26) 0.068*** (13.61) 0.033*** (4.90) 0.073*** (6.93) 0.006*** (7.51) 0.009*** (9.29) 0.030*** (24.18) 0.551*** (9.21) -0.006*** (-3.05)

20913 0.202

15355 0.217

28996 0.208

28996 0.208

28996 0.208

18102 0.216

34

Table 6 Determinants of Value Data from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with nonpositive assets or cash holdings, or negative capital expenditures. See text for definitions of financial constraints criteria. The dependent variable is the level of the spread of value over cost and the change in the spread of value over cost from (t-1) to t. The spread of value over cost is equal to the total market value of a firm net of book value of its assets divided by the book value of its assets. The independent variables include prior and future changes in total book non-cash assets, in cash holdings, in earnings before interest and extraordinary items and after depreciation and taxes, in R&D expenditures, interest expense, and in total dividends paid, as well as the future change in the market firm value. Changes in variables are calculated over one-year periods. The level regressions also include current levels of earnings, R&D expenditures, interest expense, and total dividends paid. Prior change in assets x FC and Future change in assets x FC are interactions of prior and future change in total book assets with financial constraints dummies. All variables are deflated by contemporaneous total book assets. The financial constraints dummy takes a value of one if a company is identified as financially constrained and zero if a company is identified as financially unconstrained by the respective criterion. Regressions are estimated using industry fixed effects and include calendar-year dummies. Industry is defined at the 2-Digit SIC level. Statistical significance is computed using heteroskedasticity and autocorrelation robust standard errors that are clustered at the firm level. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Coefficient estimates are reported with t-statistics in parentheses below.

35

Table 6 - continued Dependent Variable Financial Constraints Criteria Prior change in non-cash assets Future change in non-cash assets Pr. ch. in non-cash assets x FC Ft. ch. in non-cash assets x FC Prior change in cash Future change in cash Pr. ch. in cash x FC Ft. ch. in cash x FC Earnings Prior change in earnings Future change in earnings R&D expenditures Prior change in R&D Future change in R&D Interest expense Prior change in interest Future change in interest Dividends Prior change in dividends Future change in dividends Future change in firm value Constraints dummy

Observations Adjusted R-squared

Level of Spread of Value over Cost Average Payout Bond Ratings 0.465*** (8.84) 0.739*** (15.74) 0.304*** (4.79) 0.109** (1.96) 1.038*** (8.86) 1.614*** (13.99) 0.563*** (3.31) 0.207 (1.33) 1.062*** (5.89) 0.220*** (2.81) 1.262*** (11.86) 6.084*** (19.70) 2.534*** (5.24) 8.168*** (16.11) 0.021 (0.04) -4.168*** (6.96) -7.145*** (10.52) 11.752*** (12.37) -1.085 (1.25) 10.776*** (8.12) -0.210*** (13.87) 0.052** (2.00)

0.328*** (7.97) 0.678*** (18.20) 0.480*** (10.30) 0.164*** (4.21) 1.213*** (7.99) 1.526*** (9.11) 0.211 (1.25) 0.120 (0.65) 1.214*** (8.52) 0.202*** (3.39) 1.390*** (16.32) 6.403*** (25.94) 2.378*** (6.11) 8.399*** (20.62) -0.449 (1.02) -3.768*** (8.52) -7.414*** (14.28) 11.231*** (14.65) -0.749 (1.02) 10.621*** (9.27) -0.208*** (17.83) -0.017 (1.07)

29964 0.310

54317 0.307

36

Change in Spread of Value over Cost Average Payout Bond Ratings 0.128*** (3.40) 0.402*** (12.54) 0.045 (0.85) 0.097** (2.28) 0.860*** (8.89) 0.850*** (9.10) 0.519*** (3.67) 0.387*** (3.07)

0.129*** (4.45) 0.330*** (13.16) 0.076** (2.15) 0.147*** (5.10) 0.852*** (7.70) 0.881*** (7.98) 0.374*** (2.91) 0.126 (1.02)

0.786*** (11.76) 1.010*** (14.50)

0.823*** (16.16) 1.048*** (19.84)

0.888** (2.07) 5.060*** (12.34)

0.400 (1.17) 5.210*** (15.48)

-2.163*** (4.60) -3.339*** (6.82)

-2.267*** (6.60) -3.578*** (9.93)

2.589*** (4.59) 2.544*** (4.37) -0.112*** (10.35) -0.033*** (4.30)

2.852*** (5.89) 2.419*** (4.73) -0.106*** (12.20) -0.018*** (3.93)

28367 0.190

51804 0.184

Table 7 Univariate Comparison of Financially Constrained Firms Data from Compustat for the period of 1985-2002. We exclude companies that belong to financial (SIC 6000-6999) and utility (SIC 4910-4939) industries. We exclude firms with total book assets less than 25 million in constant 1994 dollars and firm-years with non-positive assets or cash holdings, or negative capital expenditures. See text for definitions of financial constraints criteria. The sample includes financially constrained firms only. Firms with the ratio of abnormal cash holdings to total assets below (above) the median are classified as low (high) cash firms. Cash flow and net investment are deflated by total book assets. CEO ownership and insider ownership are the percentage of equity (common stock + options) held by the top executive officer and by top executive officers, respectively, as reported by S&P’s Execucomp database. Analyst coverage is the number of annual analyst recommendations for a firm. The value of debt issues is long-term debt issuance net of reduction over total book assets. The value of equity issues is one-year change in the number of stocks outstanding adjusted for stock splits multiplied by the fiscal year close price. 3-yr change in margin is the change in cash flow margin over the prior three-year period. Cash flow margin is income before extraordinary items plus depreciation and amortization over sales. Z-score is Altman’s (1968) Z-score. Interest coverage is earnings before interest and tax divided by interest expense. Mean values of each variable are reported. * denotes that the value for high cash firms is significantly different from that of low cash firms at the 0.05 level using the Wilcoxon Signed-Ranks Test.

Cash

Payout Low High (A) (B)

Avg. Payout Low High (A) (B)

Financial Constraints Criteria Dividends Bond Ratings Low High Low High (A) (B) (A) (B)

Paper Ratings Low High (A) (B)

KZ Index Low High (A) (B)

CEO ownership

0.057

0.053

0.066

0.060*

0.058

0.057

0.060

0.065*

0.053

0.056

0.046

0.049

Insider ownership

0.101

0.089

0.113

0.099*

0.100

0.096

0.103

0.108

0.091

0.094

0.078

0.083

Net Investment

0.027

0.033*

0.037

0.050*

0.030

0.040*

0.032

0.039*

0.031

0.038*

0.028

0.040*

Cash flow

0.009

0.029*

0.011

0.026*

0.010

0.028*

0.029

0.049*

0.032

0.046*

0.020

0.044*

Z-score

23.70

67.46*

26.95

80.70*

24.26

78.36*

23.73

72.63*

20.36

57.63*

8.95*

8.27

Interest Coverage

10.61

22.16*

20.37*

10.21

22.96*

13.43

27.79*

11.99

22.32*

5.05

4.32

3-yr change in CF margin

0.018

0.0157*

0.026

0.496*

0.012

0.143*

0.007

-0.051*

-0.007

0.083*

0.028

0.337*

Value of debt issued

0.058

0.076*

0.069

0.089*

0.066

0.087*

0.054

0.062*

0.061

0.078*

0.051

0.075*

Value of equity issued

0.006

0.009*

0.005

0.009*

0.010

0.012*

0.010

0.011*

0.010

0.013*

0.009

0.010*

Analyst coverage

56.0

62.7*

50.4

56.9*

53.8

58.4*

42.9*

45.6

52.7*

56.2

68.8

65.5

10.32

37