Institutional Investors, Financial Health, and Equity … Investors, Financial Health, and Equity Valuation Dan S. Dhaliwala, Oliver Zhen Lia*, and Hong Xieb aUniversity of Arizona

Institutional Investors, Financial Health, and Equity Valuation

Dan S. Dhaliwala, Oliver Zhen Lia*, and Hong Xieb

aUniversity of ArizonabUniversity of Kentucky

Abstract

We investigate the relation between institutional ownership, financial health, and the market valuation weights on earnings and the book value of equity. We find that the valuation weight on earnings (book value) increases (decreases) with the level of institutional ownership for profit firms, while that on book value increases with the level of institutional ownership for loss firms. This valuation effect is not subsumed by incorporating current measures of financial health and is mainly driven by institutions with long investment horizons and monitoring incentives. We conclude that the institutional valuation effect is consistent with institutions playing a positive governance role.

JEL Classifications: M41, G12

Keywords: institutional ownership, equity valuation, financial health

1. Introduction

This paper investigates the relation between institutional ownership, financial health, and the market valuation weights on earnings and book value of equity. Our inquiry is motivated by two lines of research in the literature. The first line of research suggests that institutional investors play a positive role in corporate governance (e.g., Shleifer and Vishny, 1986 and 1997; Bushee 1998) and that they prefer to invest in financially healthy firms (e.g., Hessel and Norman, 1992; Del Guercio, 1996). The second line of research suggests that the market valuation weights on earnings and book value of equity are a function of firms’ financial health (e.g., Burgstahler and Dichev, 1997; Barth, Beaver and Landsman, 1998; and Collins, Pincus and Xie, 1999). Since

* Corresponding author: Oliver Zhen Li, Department of Accounting, University of Arizona, 1130 E Helen Street, Tucson, Arizona 85721. Email: [email protected].

We thank Brian Bushee, Gia Chevis, Dan Collins, Jevons Lee, James Myers, Mort Pincus, Joshua Rosett, Weimin Wang, David Ziebart, and workshop participants at Tulane University and the 2007 APJAE Symposium on Accounting at the National Taiwan University for helpful comments and suggestions. We especially thank Brian Bushee for generously providing us with his institution classification data.

Dan S. Dhaliwal, Oliver Zhen Li, and Hong Xie Asia-Pacific Journal of Accounting & Economics 17 (2010) 151–173

152

institutional investors prefer to invest in financially healthy firms and their presence likely dynamically improves firms’ financial health through their positive influence on corporate governance, we predict that the market valuation weights on earnings and book value of equity are related to the level of institutional ownership.

We find that firms with high levels of institutional ownership are financially healthier or less financially distressed than their low institutional ownership counterparts. For example, high institutional ownership firms enjoy higher ROE and ROA ratios in the current year and subsequent three years compared with low institutional ownership firms, regardless of whether these firms report profits or losses in the current year. In addition, high institutional ownership firms that report profits in the current year are more likely to continue to report profits and higher ROE and ROA ratios in the subsequent years than their low institutional ownership counterparts. On the other hand, high institutional ownership firms that report losses in the current year are more likely to turn around and report profits and higher ROE and ROA ratios in the subsequent years relative to low institutional ownership firms that report losses in the current year. In short, we find that firms with high levels of institutional ownership are financially healthier or less financially distressed, both in terms of current and future financial performance measures, than firms with low levels of institutional ownership.

We then examine the effect of institutional ownership on equity valuation weights on earnings and book value of equity using a valuation model following Barth et al. (1998) and Collins et al. (1999). As expected, we find that the market valuation weight on earnings (book value of equity) increases (decreases) with the level of institutional ownership for profit firms while the valuation weight on book value of equity for loss firms increases in the level of institutional ownership. Thus, we find that the valuation weights on earnings and book valuation of equity change systematically according to the level of institutional ownership.

Our findings on the valuation effect of institutional ownership are consistent with two potential roles of institutional investors documented in prior literature. First, institutional investors may simply self-select into financially healthy or growth firms (e.g., Hessel and Norman, 1992; Del Guercio, 1996). Second, institutional investors play a positive governance role that can dynamically influence firms’ current and future performance (Shleifer and Vishny, 1986 and 1997; Bushee, 1998). These two roles, however, are not mutually exclusive. We, therefore, can only provide evidence on whether our findings on the valuation effect of institutional ownership are more consistent with one role than the other role.

We conduct several tests for this inquiry. First, we incorporate current measures of financial health into our valuation model in addition to the level of institutional ownership. If institutional investors simply pick financially healthy firms, then the valuation effect of institutional ownership is likely to be subsumed once we control for current measures of financial health. If, on the other hand, institutional investors play a positive governance role, then the valuation effect of institutional ownership is unlikely to be subsumed by controlling for current measures of financial health. This is because the valuation weights on earnings and book value of equity reflect not only firms’ current but also their future financial health, which is affected by the governance role of institutional investors. We find that the valuation weights on earnings and book


153

value of equity remain systematically related to the level of institutional ownership after incorporating various measures of current financial health. This suggests that the level of institutional ownership goes beyond serving merely as a proxy for the current measures of financial health and the valuation effect of institutional ownership is consistent with institutional investors playing a positive governance role, which can dynamically affect firms’ future financial health.

Second, we develop a model linking the level of institutional ownership to its determinants, such as size, ROE and growth potential, following Badrinath, Gay and Kale (1989) and Gompers and Metrick (2001). We then examine the effect of the residual level of institutional ownership on equity valuation weights. We continue to observe that the residual level of institutional ownership, which is unrelated to its determinants such as size, ROE and growth potential, impacts equity valuation weights in the expected directions. This again suggests that the level of institutional ownership serves more than a proxy for current measures of financial health and that our findings on the valuation effect of institutional ownership is consistent with institutions playing a governance role.

Finally, we separate institutional investors into short-term investors (i.e., transient) and long-term investors (i.e., quasi-indexers and dedicated institutions) following Bushee’s (1998) classification scheme. We find that the valuation effect of institutional ownership is mainly driven by long-term institutional investors rather than short-term institutional investors. Since Bushee (1998) finds that long-term institutional investors, but not short-term institutional investors, play a positive governance role, the above finding further suggests that the valuation effect of institutional ownership documented in this paper is consistent with the governance role.

This paper contributes to the literature on equity valuation. First, we establish a link between the level of institutional ownership and the market valuation weights on earnings and book value of equity and provide evidence that such a link is consistent with institutional investors, especially long-term institutional investors, playing a positive governance role. Given the sheer size of and the rapid growth in institutional ownership in the U.S. capital market (Gompers and Metrick, 2001), we believe that it is important to explicitly examine whether and how the level of institutional ownership affects equity valuation weights. Our findings that institutional ownership affects equity valuation weights are consistent with many prior studies documenting that institutional ownership affects the market pricing of accounting information (e.g., Callen, Hope and Segal, 2005; Jiambalvo, Rajgopal and Venkatachalam, 2002; Bartov, Radhakrishnan and Krinsky, 2000).

Second, we extend Collins et al. (1999) and Barth et al. (1998). We confirm their finding that the valuation weights on earnings and book value of equity differ across profit and loss (financially healthy versus less financially healthy) firms. More importantly, we document that the valuation weights on earnings and book value of equity within profit or loss firms also differ according to the level of institutional ownership.

The remainder of the paper is organized as follows. The next section develops hypotheses. Section 3 describes the sample selection process and variables. We present


154

the descriptive statistics and empirical findings in Section 4. Section 5 summarizes and concludes the paper.

2. Hypothesis Development

(i) Level of Institutional Ownership and Financial Health

A large number of studies suggest that institutional investors play a positive role in corporate governance. Theoretically, Shleifer and Vishny (1986) and Admati, Pfleiderer and Zechner (1994) demonstrate that the presence of larger shareholders overcomes the free-rider problem associated with monitoring, leading to monitoring as an equilibrium outcome. The empirical literature appears to support this prediction. For example, Carleton, Nelson and Weisbach (1998) show that the Teachers Insurance Annuity Association – College Retirement Equities Fund (TIAA-CREF) is able to influence corporate governance issues successfully through private negotiations with firms’ management. Hartzell and Starks (2003) provide evidence that institutional concentration is positively related to the pay-for-performance sensitivity and negatively related to the level of executive compensation. They interpret their findings as being consistent with institutions serving a monitoring role in mitigating the agency problem between shareholders and managers.

In addition, Bushee (1998) finds that institutional investors, as a whole, reduce the likelihood that managers cut R&D to reverse earnings declines, suggesting that “institutional investors are sophisticated investors who typically fulfill a monitoring role in reducing incentives to manage earnings with cuts in R&D (p. 307).” Chan, Martin, and Kensinger (1990) show that the announcements of increased R&D spending are associated with a positive market reaction, especially for high-tech firms. Lev and Sougiannis (1996) provide evidence that R&D expenditure is positively associated with future earnings and that R&D capital is positively associated with future stock returns. These results, in conjunction with those in Bushee (1998), suggest that institutional investors, as a whole, can improve future firm performance through curtailing managers’ incentives to manage earnings with cuts in R&D. However, Bushee also finds that predominant ownership by transient (i.e., short-term) institutions significantly increases the likelihood that managers cut R&D to manage earnings (p. 307). This suggests that the positive governance role of institutional ownership is primarily due to the presence of long-term institutions (i.e., dedicated and quasi-indexer). Finally, Chung, Firth, and Kim (2002) find that the presence of large institutional investors inhibits managers from increasing or decreasing reported profits towards the managers’ desired level or range of profits, consistent with institutional investors monitoring and constraining the self-serving behavior of corporate managers.

Another line of research suggests that institutional investors self-select into financially healthy firms due to their fiduciary responsibility. Del Guercio (1996) argues that the prudent-man laws purport to protect beneficiaries by allowing them to seek damages from a fiduciary who fails to invest in their best interest. As a result, fiduciaries have an incentive to protect themselves from liabilities by tilting their portfolios toward


155

high-quality assets that are easy to defend in court. She finds that institutional investors tilt their investment portfolio towards high quality stocks and that more constrained banks tilt significantly more than other institutional investors. In examining the financial characteristics of institution-neglected and institution-held firms, Hessel and Norman (1992) find that firms strongly held by institutions (with at least 65 per cent of institutional ownership) are more profitable in terms of ROE and ROA, invest more heavily in R&D and are larger in size than firms neglected by institutions (with less than 15 per cent of institutional ownership). Thus, they also find that institutions prefer to invest in financially healthy firms.1 Regardless of whether institutional investors play a positive governance role or simply self-select into financially healthy firms, the extant literature suggests a positive association between institutional ownership and a firm’s financial health.

(ii) Level of Institutional Ownership and Equity Valuation

Prior valuation research suggests that the market valuation of accounting information depends on firms’ financial health. For example, Burgstahler and Dichev (1997) find that the valuation weight on earnings (book value of equity) is high (low) when the firm is financially healthy as measured by the ROE ratio. Furthermore, Barth et al. (1998) provide evidence that the pricing multiple on and the incremental explanatory power of earnings (book value of equity) increase (decrease) as financial health, measured either by the proximity to bankruptcy or by debt rating, increases. Finally, Collins et al. (1999) show that (1) the valuation weight on earnings (book value of equity) is higher (lower) for profit firms than that for loss firms and (2) the primary value attribute for loss firms is book value of equity, since losses are not indicative of expected future earnings and thus are of little value relevance.

To the extent that firms with high levels of institutional ownership are financially healthier than those with low levels of institutional ownership, we expect that the market valuation weight on earnings increases while that on book value of equity decreases as the levels of institutional ownership increase for firms that report profits. For loss firms, however, we expect that the valuation weight on book value of equity (the primary value attribute for loss firms) increases as the levels of institutional ownership increase because we expect loss firms with high institutional ownership to be less financially distressed and more likely to turn around than loss firms with low institutional ownership. We make no prediction regarding how the valuation weight on earnings changes in relation to the levels of institutional ownership for loss firms since losses are of little value relevance.

1 However, Smith (1996) finds that institutional activism increases shareholder wealth if it is successful in changing the organizational control structure of targeted firms and that performance related (as opposed to takeover related) activism increases shareholder wealth. He, however, finds no statistically significant changes in operating performance after vs. before activism for targeted firms. Similarly, Wahal (1996) finds no long-term improvement in operating performance for firms targeted by nine major pension funds in the post-targeting period. Note, both studies examine a small sample of firms that become targets of institutional investor activism and that they compare operating performance of targeted firms after vs. before activism, instead of comparing operating performance between firms with high vs. low level of institutional ownership.


156

Based on the above discussion, we test the following hypotheses regarding the valuation effect of institutional ownership:

Hypothesis 1: For profit firms, the valuation weight on earnings increases whereas that on book value of equity decreases as the level of institutional ownership increases.

Hypothesis 2: For loss firms, the valuation weight on book value of equity increases as the level of institutional ownership increases.

3. Sample and Variables

(i) Sample Selection

We obtain data from two sources. Accounting data are obtained from the 2000 annual Compustat files and data on institutional ownership are obtained from the SEC Compact Disclosure database for the period between 1989 and 1999. Institutional investors with more than US$100 million in equity ownership must report their holdings to SEC in their quarterly Section 13(f) filings. Institutional investors in the Compact Disclosure database include banks, insurance companies, mutual funds, college endowment funds, corporate investors, pension and retirement funds, broker dealers, and other investment advisors.

We obtain 30,727 observations from Compustat for the period between 1989 and 1999 with sufficient information to calculate all accounting variables used in this study (discussed in more detail below). We then merge the Compustat data with the institutional ownership data and obtain a total of 14,576 observations. To reduce the impact of extreme observations, we delete the top and bottom 0.5% of all variables defined below.2 Our final sample is composed of 13,612 firm-year observations, covering 11 years between 1989 and 1999.

(ii) Variable Measurement

Stock price (P) is the fiscal year-end stock price (Compustat data item #199). Earnings per share (NI) are the bottom-line net income (Compustat data item #172) divided by the number of shares outstanding (Compustat data item #25).3 Book value of equity per share (BV) is total book value of equity (Compustat data item #60) divided by the number of shares outstanding (Compustat data item #25). Book-to-market ratio (BTM) is the ratio between total book value of equity and market value of equity (Compustat data item #199×Compustat data item #25). Research and development intensity (RND) is total research and development expense (Compustat data item #46) scaled by the beginning of the period total assets (Compustat data item #6). Sales growth (GROWTH)

2 These variables are price per share, net incomes per share, book value of equity per share, total assets, book-to-market ratio, research and development intensity, sales growth and return-on-equity ratio.

3 Using income before extraordinary items (Compustat data item #18) yields qualitatively identical results as those reported in the paper.


157

is change in sales (Compustat data item #12) scaled by the beginning of the period total assets. The level of institutional ownership (INST) is measured as the ratio of the number of shares held by all institutional investors to the number of shares outstanding for each firm-year. Finally, we measure financial health using the ROE ratio, defined as net income (#172) divided by beginning of the period book value of equity (#60), and the ROA ratio, defined as net income scaled by beginning of the period total assets.

4. Results

(i) Descriptive Statistics

We first divide the full sample into low and high institutional ownership subsamples using the median level of institutional ownership (INST), obtaining 6,814 observations in the low institutional ownership portfolio and 6,798 observations in the high institutional ownership portfolio. We then separate the full sample into profit (NI ≥ 0) and loss (NI < 0) subsamples. There are 10,276 observations in the profit subsample with 4,701 (5,575) observations in the low (high) institutional ownership portfolio. On the other hand, the loss subsample contains 3,336 observations with 2,113 (1,223) observations in the low (high) institutional ownership portfolio.

Panel A, Table 1 reports descriptive statistics for the profit subsample. By construction, the mean and median levels of institutional ownership for the high institutional ownership portfolio (57.84% and 57.00%, respectively) are higher than those for the low institutional ownership portfolio (16.17% and 16.00%). The means and medians of price (P), earnings per share (NI), book value of equity per share (BV) and total assets (ASSET) for the high institutional ownership portfolio are all significantly higher than those for the low institutional ownership portfolio. In addition, high institutional ownership firms are more profitable in terms of the current ROE and ROA ratios. The mean and median ROE (ROA) for the high institutional ownership portfolio are 18.38% (9.37%) and 15.99% (7.85%), respectively, while those for the low institutional ownership portfolio are 16.62% (8.82%) and 13.22% (7.03%). The differences in mean and median ROE (ROA) between high and low institutional ownership portfolios are both highly significant, suggesting a positive relation between financial health and institutional ownership. On the other hand, the book-to-market ratio (BTM) is significantly higher for low institutional ownership firms than for high institutional ownership firms, suggesting that institutional investors are associated with growth opportunities. Furthermore, R&D intensity (RND) is also slightly higher for low institutional ownership firms than for high institutional ownership firms although there is no difference in sales growth (GROWTH) between these two groups.


158

Table 1 Descriptive Statistics

Panel A: Firms Reporting ProfitsLow Institutional Ownership High Institutional Ownership Difference in

Variable N Mean Median Std Dev N Mean Median Std Dev Mean Median

INST 4,701 0.1617 0.1600 0.1044 5,575 0.5784 0.5700 0.1485 –166.34*** –87.25***P 4,701 16.6025 11.7500 15.5393 5,575 30.5216 25.6880 20.0812 –39.58*** –43.67***NI 4,701 0.9095 0.6600 0.9030 5,575 1.6390 1.2992 1.2770 –33.79*** –36.36***BV 4,701 7.7120 5.8275 6.6629 5,575 12.0725 9.7892 8.0956 –29.85*** –36.43***ASSET 4,701 1450.74 77.90 5943.37 5,575 2489.90 495.47 5612.99 –9.06*** –46.99***BTM 4,701 0.6266 0.5266 0.4278 5,575 0.4986 0.4242 0.3333 16.69*** 15.68***RND 4,701 0.0626 0.0326 0.0803 5,575 0.0592 0.0295 0.0741 2.22** 0.89GROWTH 4,701 0.2068 0.1400 0.3096 5,575 0.1988 0.1318 0.2820 1.36 0.47ROEt 4,701 0.1662 0.1322 0.1486 5,575 0.1838 0.1599 0.1357 –6.25*** –11.41***ROAt 4,701 0.0882 0.0703 0.0728 5,575 0.0937 0.0785 0.0690 –3.91*** –6.60***ROEt+1 3,521 0.1131 0.1194 0.1850 4,352 0.1429 0.1449 0.1650 –7.54*** –9.05***ROEt+2 2,719 0.0981 0.1111 0.1839 3,433 0.1334 0.1406 0.1693 –7.82*** –9.14***ROEt+3 2,028 0.0927 0.1067 0.1864 2,648 0.1364 0.1444 0.1689 –8.37*** –10.27***ROAt+1 3,521 0.0632 0.0624 0.1024 4,352 0.0733 0.0705 0.0869 –4.72*** –4.43***ROAt+2 2,719 0.0565 0.0592 0.1028 3,433 0.0679 0.0675 0.0878 –4.71*** –4.43***ROAt+3 2,028 0.0539 0.0568 0.1010 2,648 0.0684 0.0682 0.0858 –5.30*** –5.51***Dt+1 3,521 0.1562 0.0000 0.3631 4,352 0.1206 0.0000 0.3257 4.58*** 4.57***Dt+2 2,719 0.1754 0.0000 0.3804 3,433 0.1398 0.0000 0.3469 3.83*** 3.83***Dt+3 2,028 0.1805 0.0000 0.3847 2,648 0.1326 0.0000 0.3392 4.52*** 4.51***

Panel B: Firms Reporting LossesLow Institutional Ownership High Institutional Ownership Difference in


INST 2,113 0.1574 0.1550 0.0959 1,223 0.5585 0.5414 0.1466 –85.65*** –48.10***P 2,113 7.9194 5.3750 8.7608 1,223 16.5714 11.6880 14.8915 –18.55*** –23.51***NI 2,113 –0.7997 –0.5671 0.8271 1,223 –1.2467 –0.8543 1.2177 11.40*** 10.96***BV 2,113 3.7118 2.6179 3.8574 1,223 7.8990 6.1827 6.5239 –20.47*** –25.15***ASSET 2,113 350.33 41.17 2478.06 1,223 1406.30 207.84 4068.15 –8.24*** –29.90***BTM 2,113 0.6688 0.5015 0.5515 1,223 0.6521 0.5229 0.4820 0.92 –1.34RND 2,113 0.1508 0.1034 0.1619 1,223 0.1242 0.0817 0.1355 5.08*** 3.64***GROWTH 2,113 0.0344 0.0112 0.2651 1,223 0.0406 0.0106 0.2770 –0.63 0.44ROEt 2,113 –0.2803 –0.2028 0.2605 1,223 –0.2010 –0.1257 0.2142 –9.51*** –9.61***ROAt 2,113 –0.1785 –0.1160 0.1759 1,223 –0.1148 –0.0655 0.1299 –11.94*** –10.86***ROEt+1 1,348 –0.1602 –0.1049 0.3173 860 –0.0479 0.0068 0.2877 –8.41*** –9.42***ROEt+2 899 –0.1043 –0.0367 0.3429 580 0.0023 0.0416 0.2783 –6.27*** –6.43***ROEt+3 617 –0.0588 0.0235 0.3457 404 0.0496 0.0850 0.2617 –5.37*** –5.01***ROAt+1 1,348 –0.1149 –0.0549 0.1967 860 –0.0399 0.0034 0.1527 –9.50*** –9.67***ROAt+2 899 –0.0812 –0.0157 0.2053 580 –0.0175 0.0196 0.1487 –6.46*** –5.83***ROAt+3 617 –0.0514 0.0125 0.2063 404 0.0090 0.0376 0.1379 –5.17*** –4.39***Dt+1 1,348 0.6506 1.0000 0.4770 860 0.4837 0.0000 0.5000 7.87*** 7.76***Dt+2 899 0.5417 1.0000 0.4985 580 0.3845 0.0000 0.4869 5.98*** 5.91***Dt+3 617 0.4554 0.0000 0.4984 404 0.3168 0.0000 0.4658 4.46*** 4.42***


159

Table 1 Descriptive Statistics (continued)

Panel C: Firm Performance and Changes in Institutional OwnershipFirms Decreasing

Institutional OwnershipFirms Increasing

Institutional Ownership Difference in


LINST 4,870 0.3931 0.3800 0.2452 5,211 0.3563 0.3385 0.2338 7.71*** 7.21***INST 4,870 0.3401 0.3100 0.2327 5,211 0.4403 0.4400 0.2434 –21.11*** –20.33***P 4,870 20.8713 14.6250 19.0630 5,211 22.9292 17.1250 19.3307 –5.38*** –7.51***NI 4,870 0.7093 0.6348 1.5283 5,211 0.8799 0.7987 1.4934 –5.66*** –6.74***BV 4,870 9.1660 7.0892 7.6918 5,211 9.4988 7.5158 7.5569 –2.19** –3.67***ASSET 4,870 1895.29 176.57 5611.56 5,211 1848.21 208.00 5332.13 0.43 –4.39***BTM 4,870 0.6103 0.4964 0.4346 5,211 0.5646 0.4626 0.4109 5.41*** 5.29***RND 4,870 0.0762 0.0386 0.0982 5,211 0.0764 0.0365 0.1035 –0.09 1.05GROWTH 4,870 0.1239 0.0837 0.2682 5,211 0.1622 0.1106 0.2842 –6.98*** –7.75***ROE 4,870 0.0620 0.0978 0.2379 5,211 0.0886 0.1208 0.2323 –5.68*** –7.29***ROA 4,870 0.0281 0.0474 0.1371 5,211 0.0413 0.0590 0.1345 –4.87*** –6.82***

The final sample contains 13,612 firm-year observations spanning 11 years from 1989 to 1999. The decrease in sample size in years t+1, t+2 and t+3 is due to the fact that some observations in the final sample have missing observations in one or more years after the sample year (year t). These observations are primarily in 1997, 1998 and 1999. Accounting data are taken from the 2000 annual Compustat files. Stock price (P) is the fiscal year-end stock price (Compustat data item #199). Earnings per share (NI) are the bottom-line net income (Compustat data item #172) divided by the number of shares outstanding (Compustat data item #25). Book value of equity per share (BV) is total book value of equity (Compustat data item #60) divided by the number of shares outstanding (Compustat data item #25). ASSET is total assets (Compustat data item #6). Book-to-market ratio (BTM) is the ratio between total book value of equity and market value of equity (Compustat data item #199×Compustat data item #25). Research and development intensity (RND) is total research and development expense (Compustat data item #46) scaled by beginning of the period total assets. Sales growth (GROWTH) is change in sales (Compustat data item #12) scaled by the beginning of the period total assets. Return on equity (ROE) is net income scaled by beginning of the period book value of equity. Return on asset (ROA) is net income scaled by beginning of the period total assets. The loss dummy (D) is equal to 1 if NI < 0, and 0 otherwise.Data on institutional ownership are obtained from the SEC Compact Disclosure database for the period between 1989 and 1999. Institutional investors with more than US$100 million in equity ownership must report their holdings to SEC in their quarterly Section 13(f) filings. Institutional investors in the Compact Disclosure database include banks, insurance companies, mutual funds, college endowment funds, corporate investors, pension and retirement funds, broker dealers, and other investment advisors. We calculate the level of institutional ownership (INST) as the ratio between the number of shares held by all institutional investors and the number of shares outstanding for each firm-year. LINST is the lagged level of institutional ownership. If INST > LINST, the firm has increased its level of institutional ownership; if INST ≤ LINST, the firm has decreased or maintained its level of institutional ownership.See Table 1 for definitions of other variables.*, **, and *** denote significance at the 0.10, 0.05 and 0.01 level, respectively, based on a two-tailed t-statistic.


160

We report descriptive statistics for the loss subsample in Panel B, Table 1. The mean and median levels of institutional ownership for the high institutional ownership portfolio are 55.85% and 54.14%, respectively, while those for the low institutional ownership portfolio are 15.74% and 15.50%. These percentages are very similar to their counterparts in the profit subsample. Also similar to the profit subsample, the means and medians of price (P), book value of equity per share (BV), and total assets (ASSET) for the high institutional ownership portfolio are significantly higher than those for the low institutional ownership portfolio. The mean and median earnings per share (NI) for the high institutional ownership portfolio, however, are more negative than their counterparts from the low institutional ownership portfolio. Despite larger losses per share, high institutional ownership firms are less financially distressed than low institutional ownership firms in terms of ROE and ROA ratios. The mean and median ROE (ROA) ratios for the high institutional ownership portfolio are –20.10% (–11.48%) and –12.57% (–6.55%), respectively, whereas those for the low institutional ownership portfolio are –28.03% (–17.85%) and –20.28% (–11.60%) in the current year. The differences between these means and the medians are all highly significant. Again, RND is higher for low institutional ownership firms than for high institutional ownership firms and there is no difference in GROWTH between the two groups. One aspect for the loss firm subsample that is different from the profit firm subsample is that BTM is no longer different between the low and high institutional ownership firms.

Panel C, Table 1 reports descriptive statistics for the subsample that exhibits decreases or no changes in the level of institutional ownership in year t relative to year t – 1 (hereafter, the decreasing subsample) and for the subsample that exhibits increases in the level of institutional ownership in year t relative to year t – 1 (hereafter, the increasing subsample). We find that the means and medians of stock price (P), earnings per share (NI) and book value per share (BV) for the decreasing subsample are significantly lower than those for the increasing subsample. Moreover, the ROE and ROA ratios for the decreasing subsample are also significantly lower than those for the increasing subsample, which is suggestive that the decreasing subsample is less financially healthy than the increasing subsample.

In Panels A and B, Table 1, we also compare the ROE and ROA ratios and the frequency of losses (D) in the subsequent three years between high and low institutional ownership firms. For both profit and loss firms, the mean and median ROE and ROA ratios are both significantly higher for high institutional ownership firms than for low institutional ownership firms. Also, loss firms with high institutional ownership are more likely to turn around and report profits in the subsequent years as compared to loss firms with low institutional ownership. For example, there are only 48.37%, 38.45% and 31.68% of loss firms with high institutional ownership continuing to report losses in year t + 1, t + 2 and t + 3, respectively, whereas there are still 65.06%, 54.17% and 45.54% of loss firms with low institutional ownership remaining in losses in year t + 1, t + 2 and t + 3. The above results show that loss firms with high institutional ownership are less financially distressed and are more likely to turn around in the subsequent years.

In summary, findings in Table 1 suggest that high institutional ownership firms are financially healthier, judged by current and future ROE and ROA ratios, and the propensity to report losses, than low institutional ownership firms. We expect the


161

differences in financial health between high and low institutional ownership firms to manifest themselves in equity valuation weights on earnings and book value of equity.

(ii) Valuation Effect of Institutional Ownership

We test Hypotheses 1 and 2 using the following regression model: Pit = β0 + β1NIit + β2BVit + β3Dit + β4NIit Dit + β5BVit Dit + β6INSTit + β7 NIit INSTit + β8BVit INSTit + β9NIit INSTit Dit + β10BVit INSTit Dit + uit, (1)

where D is a loss dummy that equals 1 if NI < 0, and 0 otherwise. Equation (1) allows the valuation weights on earnings and book value of equity to differ depending on (1) whether a firm reports a profit or a loss and (2) its level of institutional ownership. The coefficients β7 and β8 capture the interaction between INST and NI and the interaction between INST and BV, respectively, for profit firms. Findings that β7 > 0 and β8 < 0 support our Hypothesis 1. The coefficient β10, on the other hand, captures the interaction between INST and BV for loss firms. A finding that β10 > 0 supports our Hypothesis 2.

Panel A, Table 2, reports our findings. In this table and subsequent tables containing regression results, the t-statistics are computed using the Huber-White procedure that corrects for heteroskedasticity and auto-correlation. The coefficients on earnings (β1 = 7.8631, t = 10.73) and book value of equity (β2 = 0.6619, t = 6.06) for profit firms are both significantly positive as expected. In addition, the incremental coefficient on earnings for loss firms is significantly negative (β4 = –8.9250, t = –11.28) while the incremental coefficient on book value of equity for loss firms is significantly positive (β5 = 0.2414, t = 1.70). These findings are consistent with Collins et al. (1999).

Table 2 Institutional Ownership and Valuation Weights

Panel A: Institutional Ownership Panel B: Changes in Institutional Ownership

Variable Predicted Sign Estimate Variable Predicted Sign Estimate

Intercept +/– 2.3240 Intercept +/– 0.5561(4.95)*** (0.98)

NI + 7.8631 NI + 8.2776(10.73)*** (9.46)***

BV + 0.6619 BV + 0.7290(6.06)*** (6.23)***

D +/– –1.5597 D +/– –1.8337(–3.35)*** (–3.37)***

NI D – –8.9250 NI D – –10.0981(–11.28)*** (–10.04)***

BV D + 0.2414 BV D + 0.1505(1.70)* (0.98)

INST +/– 17.3320 LINST +/– 18.2641(13.79)*** (12.34)***

NI INST + 2.3833 NI LINST + 1.6946(1.70)* (1.09)


162

BV INST – –0.7327 BV LINST – –0.6372(–3.38)*** (–2.79)***

NI INST D +/– –0.3613 NI LINST D +/– 1.6089(–0.22) (0.87)

BV INST D + 0.7255 BV LINST D + 0.7093(2.69)*** (2.39)**

INC +/– 2.5832(5.87)***

NI INC + 0.6548(1.29)

BV INC – –0.2266(–3.10)***

NI INC D +/– –0.6123(–0.87)

BV INC D + 0.2267(1.90)*

Adj. R2 0.5439 Adj. R2 0.5590

# of Obs. 13,612 # of Obs. 10,081

D is a loss dummy that equals 1 if a firm reports losses during the year and 0 otherwise. LINST is the lagged level of institutional ownership, INC = 1 if INST > LINST and INC = 0 if INST ≤ LINST.See Table 1 for definitions of other variables.The t-values are computed using the Huber-White procedure, which corrects for heteroskedasticity and auto-correlation.*, **, and *** denote significance at the 0.10, 0.05 and 0.01 level, respectively, based on a two-tailed t-statistic.

The terms NI INST and BV INST capture the valuation effect of institutional ownership on earnings and book value of equity, respectively, for profit firms. The coefficient on NI INST is positive and significant (β7 = 2.3833, t = 1.70). The coefficient on BV INST is negative and significant (β8 = –0.7327, t = –3.38). These results support Hypothesis 1 that the valuation weight on earnings (book value of equity) increases (decreases) as the level of institutional ownership increases for profit firms. The term BV INST D captures the effect of institutional ownership on book value of equity for loss firms. The coefficient on BV INST D is significantly positive (β10 = 0.7255, t = 2.69), supporting Hypothesis 2 that the valuation weight on book value of equity increases in the level of institutional ownership for loss firms.

Next, we examine the valuation effect of changes in the level of institutional ownership. An increase in the level of institutional ownership can be viewed as a signal that the firm’s financial health will improve or it will turn around if in financial distress. If the market valuation weight on earnings (book value of equity) increases (decreases) in the level of institutional ownership for profit firms as we have hypothesized, then we should observe that the incremental valuation weight on earnings (book value of equity) is positive (negative) for profit firms whose levels of institutional ownership


163

increase in the current year as compared to the previous year. Since we hypothesize that the valuation weight on book value of equity increases with the level of institutional ownership for loss firms, we should observe that the incremental valuation weight on book value of equity is positive for loss firms whose levels of institutional ownership increase in the current year. We test the above implications of our Hypothesis 1 and 2 using the following equation:

Pit = β0 + β1NIit + β2BVit + β3Dit + β4NIit Dit + β5BVit Dit + β6LINSTit + β7NIit LINSTit + β8BVit LINSTit + β9NIit LINSTit Dit + β10BVit LINSTit Dit

+ β11INCit + β12NIit INCit + β13BVit INCit + β14NIit INCit Dit + β15BVit INCit Dit + uit, (2)

where LINST is the lagged level of institutional ownership. INC equals 1 if INST – LINST > 0 and equals 0 if INST – LINST ≤ 0. All other variables are as defined earlier. Note that we use LINST instead of INST in Equations (2) because INST = LINST + (INST – LINST). The dummy variable INC captures whether the current level of institutional ownership (INST) is increased or decreased as compared to its level in the previous year (LINST). The coefficients β12 and β13 capture the incremental valuation effect of increases in the level of institutional ownership for profit firms whereas the coefficient β15 measures the incremental valuation effect of increases in the level of institutional ownership on BV for loss firms. We expect that β12 > 0 and β13 < 0 and that β15 > 0.

Panel B, Table 2, presents the regression results for Equation (2). We find that the coefficient on NI INC is positive but insignificant (β12 = 0.6548, t = 1.29), not supporting Hypothesis 1; the coefficient on BV INC is significantly negative (β13 = –0.2266, t = –3.10), supporting Hypothesis 1; and that the coefficient on BV INC D is significantly positive (β15 = 0.2267, t = 1.90), supporting Hypothesis 2.

To summarize, the results in Table 2 generally support our hypotheses that the valuation weight on earnings (book value of equity) increases (decreases) for profit firms as the level of institutional ownership increases whereas the valuation weight on book value of equity increases for loss firms as the level of institutional ownership increases.

(iii) Roles of Institutional Investors in Equity Valuation

Our findings so far suggest that the level of institutional ownership is positively associated with firms’ financial health and that the valuation weight on earnings (book value of equity) changes systematically according to the level of institutional ownership for both profit and loss firms. These findings, however, are possibly consistent with two roles of institutional investors hypothesized in the literature: (1) institutional investors simply pick financially healthy firms into their portfolios, and (2) institutional investors play a positive governance role that can dynamically influence firms’ future financial health. In this section, we investigate whether our findings are more consistent with the former or latter.


164

(a) Valuation Effect of Institutional Ownership after Controlling for Current Measures of Financial Health

If institutional investors simply pick healthy firms, then the valuation effect of

institutional ownership on earnings and book value of equity is likely to be subsumed when measures of current financial health are incorporated in Equation (1). If, on the other hand, institutional investors play a positive governance role, then the level of institutional ownership likely proxies for both current and future financial health since institutional investors likely dynamically improve firms’ future financial health. Under this scenario, the valuation effect of institutional ownership on earnings and book value of equity is unlikely to be subsumed when measures of current financial health are incorporated in Equation (1). We conduct three tests along the above line of reasoning to examine whether the valuation effect of institutional ownership is more consistent with a stock picking role or a governance role.

In the first test, we incorporate ROE/ROA, two common measures of current financial health, into Equation (1):

Pit = β0 + β1NIit + β2BVit + β3Dit + β4NIit Dit + β5BVit Dit + β6INSTit + β7NIit INSTit + β8BVit INSTit + β9 NIit INSTit Dit + β10BVit INSTit Dit + β11ROEit (ROAit) + β12NIit ROEit (NIit ROAit) + β13BVit ROEit (BVit ROAit) + uit. (3)

If the valuation effect of institutional ownership on earnings and book value of equity (i.e., β7 > 0, β8 < 0 and β10 > 0) remains after incorporating ROE (ROA) in Equation (3), we would conclude that institutional investors are more likely to play a governance role than a stock picking role.

The results for estimating Equation (3) are provided in Table 3. Panel A shows the valuation effects of ROE and ROA (Model 1 and Model 2) alone. They are consistent with Barth et al. (1998) in that when the level of firm financial health increases, the equity valuation weight shifts from book value of equity to earnings. Panel B shows the valuation effect of institutional ownership after controlling for the effects of ROE or ROA. We continue to observe that β7 > 0 (2.6304, t = 1.89), β8 < 0 (–0.7016, t = –3.38) and β10 > 0 (0.8446, t = 3.15) in Model 1. These coefficients are all significant and in the right directions in Model 2 as well. These results suggest that the level of institutional ownership serves more than a mere proxy for the current measures of financial health such as ROE or ROA. The valuation effect of institutional ownership on earnings and book value of equity, therefore, is likely due to institutions playing a governance role.

In the second test, we use debt rating as an alternative measure of financial health following Barth et al. (1998). Only a small fraction of firms in Compustat reports debt rating. We estimate debt rating for firms without rated debt following the procedure in Barth et al. (1998). First, we estimate the following debt rating estimation equation in the cross-section, year-by-year, for firms with rated debt:

BRit = b0 + b1ASSETit + b2ROA it + b3DBTA it + b4DIVit + e it, (4)


165

where BR is a numerical value corresponding to a firm’s Standard and Poor’s subordinated debt rating (Compustat item # 320) or senior debt rating (Compustat item #280), ASSET is total assets, ROA is return-on-asset ratio, DBTA is debt (Compustat item #181) to total asset ratio and DIV is an indicator variable that is equal to one if the firm pays a dividend (Compustat item #21).

Next, we use Equation (4) and the estimated coefficients b0, b1, b2, b3 and b4 to predict debt rating for firms without rated debt. Thus, we have actual debt rating for firms with rated debt and predicted debt rating for firms without rated debt. We then create two indicator variables, SUB and SEN. SUB is set to one if a firm’s subordinated debt rating is below the sample median (i.e., financially healthy), and zero otherwise.4 Similarly, SEN is set to one if a firm’s senior debt rating is below the sample median, and zero otherwise. SUB and SEN are two alternative measures of current financial health.

Table 3 Institutional Ownership and Concurrent Measures of Firm Financial Health

Panel A: The Effect of Current Measures of Financial Health

Model 1: ROE Model 2: ROA Model 3: SUB Model 4: SEN

Variable Estimate Variable Estimate Variable Estimate Variable EstimateIntercept 6.4016 Intercept 5.6919 Intercept 6.1432 Intercept 6.0043

(20.58)*** (19.03)*** (16.03)*** (17.23)***NI 4.6196 NI 4.1902 NI 2.1370 NI 1.8303

(9.46)*** (12.15)*** (9.59)*** (8.71)***BV 0.9747 BV 0.9782 BV 1.0420 BV 0.9987

(24.56)*** (24.72)*** (15.40)*** (16.10)***ROE 7.5856 ROA 21.8736 SUB 5.9393 SEN 6.4780

(7.61)*** (12.37)*** (9.47)*** (10.27)***NI ROE 10.8034 NI ROA 25.3332 NI SUB 3.6415 NI SEN 3.9310

(14.89)*** (19.78)*** (9.22)*** (9.81)***BV ROE –1.3591 BV ROA –2.1492 BV SUB –0.3056 BV SEN –0.2689

(–2.87)*** (–3.26)*** (–3.63)*** (–3.27)***Adj. R2 0.5078 Adj. R2 0.5206 Adj. R2 0.5297 Adj. R2 0.5382# of Obs. 13,612 # of Obs. 13,612 # of Obs. 11,813 # of Obs. 11,813

Panel B: The Incremental Effect of Institutional Ownership

Model 1: ROE Model 2: ROA Model 3: SUB Model 4: SENVariable Estimate Variable Estimate Variable Estimate Variable Estimate

Intercept 1.3426 Intercept 0.5450 Intercept 1.7209 Intercept 1.6971(2.67)*** (1.12) (3.43)*** (3.44)***

NI 6.4814 NI 4.8132 NI 6.8370 NI 6.6466(6.74)*** (5.94)*** (8.76)*** (8.64)***

BV 0.8567 BV 0.9319 BV 0.5396 BV 0.5166(7.74)*** (8.74)*** (4.76)*** (4.56)***

D 0.9000 D 2.7735 D –1.1789 D –0.8725(1.33) (4.16)*** (–2.47)** (–1.79)*

4 Standard and Poor uses a lower numerical number to indicate higher credit standing.


166

NI D –5.5972 NI D –2.8976 NI D –7.7606 NI D –7.5539(–4.75)*** (–2.97)*** (–9.56)*** (–9.39)***

BV D –0.0256 BV D –0.0678 BV D 0.2279 BV D 0.2227(–0.16) (–0.43) (1.68)* (1.64)*

INST 16.5946 INST 15.3165 INST 16.2212 INST 15.5729(13.41)*** (12.48)*** (12.72)*** (12.38)***

NI INST 2.6304 NI INST 3.0266 NI INST 2.8824 NI INST 2.7620(1.89)* (2.20)** (2.03)** (1.94)*

BV INST –0.7016 BV INST –0.6755 BV INST –0.7122 BV INST –0.6775(–3.38)*** (–3.30)*** (–3.23)*** (–3.09)***

NI INST D –0.1173 NI INST D –1.0004 NI INST D –1.7022 NI INST D –1.7333(–0.07) (–0.62) (–1.05) (–1.07)

BV INST D 0.8446 BV INST D 0.8744 BV INST D 0.7217 BV INST D 0.7141(3.15)*** (3.22)*** (2.69)*** (2.68)***

ROE 6.0876 ROA 18.7061 SUB 2.7608 SEN 3.3526(4.65)*** (8.17)*** (4.89)*** (5.77)***

NI ROE 4.6448 NI ROA 16.3292 NI SUB 0.4819 NI SEN 0.6450(4.08)*** (10.11)*** (1.45) (1.92)*

BV ROE –1.0035 BV ROA –0.9083 BV SUB 0.1300 BV SEN 0.1428(–1.88)* (–1.47) (1.72)* (1.96)**

Adj. R2 0.5482 Adj. R2 0.5588 Adj. R2 0.6062 Adj. R2 0.6100# of Obs. 13,612 # of Obs. 13,612 # of Obs. 11,813 # of Obs. 11,813

SUB is a dummy that is equal to 1 if the firm’s subordinated debt rating is above its median, and 0 otherwise. SEN is a dummy that equals 1 if the firm’s senior debt rating is above its median, and 0 otherwise. See Table 1 for definitions of other variables.The t-values are computed using the Huber-White procedure, which corrects for heteroskedasticity and auto-correlation.*, **, and *** denote significance at the 0.10, 0.05 and 0.01 level, respectively, based on a two-tailed t-statistic.

We replace ROE and ROA in Equation (3) with SUB and SEN. We first report the valuation effects of SUB and SEN alone in Panel A, Table 3 (Model 3 and Model 4). We find that the valuation weight on earnings (book value of equity) is higher (lower) for firms with favorable debt rating. We thus replicate one of the major findings in Barth et al. (1998). More importantly, Panel B (Model 3 and Model 4), Table 3, shows that the valuation effect of institutional ownership remains unchanged after incorporating SUB or SEN. Specifically, the coefficient on NI INST (BV INST) is significantly positive (negative) and that on BV INST D is significantly positive, regardless of whether SUB or SEN is used as a measure of financial health. This suggests that the valuation effect of institutional investors is incremental to the effect of debt ratings documented in Barth et al. (1998) and provides further evidence that the valuation effect of institutional ownership is consistent with the governance role point of view. Interestingly, the coefficient on NI SUB becomes insignificant and those on BV SUB and BV SEN become significantly positive, instead of negative, in Panel B. This indicates that debt rating is perhaps a less reliable measure of financial health compared with the level of institutional ownership.


167

In the third test, we examine whether the residual level of institutional ownership, which is unrelated to the current measures of financial health, impacts equity valuation in a similar way as the level of institutional ownership. Specifically, we employ the following two regression models,

INSTit = b0 + b1log(ASSET)it + b2ROEit + b3BTMit + b4RNDit + b5GROWTHit + uINSTit, (5a)

Pit = β0 + β1NIit + β2BVit + β3Dit + β4NIit Dit + β5BVit Dit + β6uINSTit + β7NIit uINSTit

+ β8BVit uINSTit + β9NIit uINSTit Dit + β10BVit uINSTit Dit + eit, (5b)

where uINST is the residual level of institutional ownership from Equation (5a) and all other variables are as defined earlier.

In the first stage, we regress the level of institutional ownership on its determinants identified in prior studies in Equation (5a). Badrinath et al. (1989) and Gompers and Metrick (2001) suggest that institutional investors prefer to invest in large firms (i.e., b1 > 0). Hessel and Norman (1992) find that institutional investors prefer firms with high ROE ratios (i.e., b2 > 0). BTM, RND and GROWTH are all measures of a firm’s growth potential. We expect institutional investors to be attracted to firms with growth potential (i.e., b3 < 0, b4 > 0 and b5 > 0).

In the second stage, we estimate Equation (5b) to examine whether the residual level of institutional ownership still affects equity valuation weights as we hypothesize. In our setting, uINST is the portion of INST that is unrelated to log(ASSET), ROE, BTM, RND and GROWTH. Thus, any effect of uINST on valuation weights indicates that the level of institutional ownership proxies more than these determinants of institutional ownership (i.e., measures of current financial health and growth opportunities) and provide support for the governance role point of view.

The results of estimating these two equations are reported in Table 4. Panel A presents our findings from estimating Equation (5a). Consistent with prior studies, we find that institutional investors prefer to invest in large firms, firms with high ROE ratios, and firms with growth potential.5 Panel B, Table 4, reports our findings from estimating Equation (5b). The coefficient on NI uINST remains significantly positive (β7 = 2.3064, t = 1.75) and the coefficient on BV uINST remains significantly negative (β8 = –0.8652, t = –4.76). This suggests that the valuation weight on earnings (book value of equity) for profit firms increases (decreases) with the residual level of institutional ownership even after controlling for log(ASSET), ROE, BTM, RND and GROWTH. However, the coefficient on BV uINST D (β10 = 0.3762, t = 1.37) is positive but insignificant.6

5 Gompers and Metrick (2001), however, find that the level of institutional ownership is positively related to the book-to-market ratio, opposite to our finding.

6 We find similar results as those reported in Table 4 when we replace ROE in equation (5a) with ROA, SUB or SEN.


168

Table 4 Residual Level of Institutional Ownership and Valuation Weights

Panel A: Determinants of Institutional Ownership

Panel B: Valuation Effect of Residual Level of Institutional Ownership

Variable Predicted Sign Estimate Variable Predicted Sign Estimate

Intercept +/– 0.0485 Intercept +/– 7.9771(2.94)*** (21.99)***

Log(ASSET) + 0.0589 NI + 9.6416(20.46)*** (24.87)***

ROE + 0.0599 BV + 0.3614(5.08)*** (6.68)***

BTM – –0.0293 D +/– –3.3346(–4.11)*** (–7.48)***

RND + 0.1448 NI D – –10.3352(4.41)*** (–23.19)***

GROWTH + 0.0138 BV D + 0.7678(1.50) (9.81)***

uINST +/– 9.3181(6.35)***

NI uINST + 2.3064(1.75)*

BV uINST – –0.8652(–4.76)***

NI uINST D +/– –0.4787(–0.30)

BV uINST D + 0.3762(1.37)

Adj. R2 0.2347 Adj. R2 0.5205# of Obs. 13,612 # of Obs. 13,612

uINST is the residual level of institutional ownership estimated from a regression of the level of institutional ownership on its determinants reported in Panel A. See Table 1 for definitions of other variables.The t-values are computed using the Huber-White procedure, which corrects for heteroskedasticity and auto-correlation.*, **, and *** denote significance at the 0.10, 0.05 and 0.01 level, respectively, based on a two-tailed t-statistic.

To summarize, our tests in Tables 3 and 4 suggest that the level of institutional ownership captures value-relevant information above and beyond current measures of financial health. In addition, they suggest that the valuation effect of institutional ownership documented in this paper is incremental to those documented in the prior literature (e.g., Burgstahler and Dichev 1997, Barth et al. 1998). Coupled with findings in Table 1 that the level of institutional ownership is positively related to measures


169

of both current and future financial health, we conclude that the level of institutional ownership appears to be an ex ante parsimonious proxy for both current and future financial health, perhaps due to institutional investors positively influencing corporate governance and thus dynamically improving firms’ financial health. Thus, the valuation effect of institutional ownership is consistent with institutional investors playing a positive governance role.

(b) Short-Term Versus Long-Term Institutional Investors

So far, we have assumed that institutions are homogeneous. To provide further evidence on whether the valuation effect of institutional ownership documented in this paper is consistent with institutional investors playing a positive governance role, we examine whether different types of institutions impact the equity valuation weights on earnings and book value of equity differentially. Bushee (1998) finds that mangers are less likely to cut R&D to reverse an earnings decline when institutional ownership is high, implying that institutions, as a whole, are sophisticated investors who typically serve a monitoring role in reducing pressures for managerial myopic behavior. However, he also finds that a large proportion of ownership by transient (i.e., short-term) institutions significantly increase the probability that managers reduce R&D to reverse an earnings decline. Bushee’s results suggest that the large shareholdings and sophistication of institutional investors allow them to monitor and discipline firm managers, but this is not likely the case for short-term institutional investors. Callen et al. (2005) show that the difference between the variance contribution of domestic and foreign earnings decreases with the level of institutional ownership. Moreover, they show that, using Bushee’s (1998) institutional investor classification scheme, the above relation remains unchanged for long-term institutional investors (i.e., dedicated institutions and quasi-indexers) but does not exist for transient institutions.

Since transient institutions are unlikely to play a governance role, we expect that their presence will not systematically affect the valuation weights on earnings and book value of equity for both profit and loss firms as we observe for institutional investors as a whole. In contrast, dedicated institutions and quasi-indexers (i.e., long-term institutions) are known to play a positive governance role (Bushee, 1998). We, therefore, expect that their presence will systematically affect the valuation weights on earnings and book value of equity as we observe earlier. A finding that long-term, but not short-term, institutions systematically affect the valuation weights on earnings and book value of equity in the expected directions provides further evidence that the valuation effect of institutional ownership is consistent with institutions playing a positive governance role.

As in Bushee (1998), we divide institutional investors into transient, quasi-indexer and dedicated institutions. We define SHORT as the percentage ownership by transient institutions and LONG as the aggregate percentage ownership by quasi-indexers and dedicated institutions. We run the following regression:


170

Pit = β0 + β1NIit + β2BVit + β3Dit + β4NIit Dit + β5BVit Dit + β6SHORTit + β7NIit SHORTit + β8BVit SHORTit + β9NIit SHORTit Dit + β10BVit INSTit Dit + β11LONGit + β12NIit LONGit + β13BVit LONGit + β14NIit LONGit Dit + β15BVit LONGit Dit + uit. (6)

We expect that the valuation effect of institutional ownership is mainly driven by long-term institutional investors. We thus expect the valuation weights on earnings and book value of equity to change systematically with the level of long-term institutional ownership for profit and loss firms (i.e., β12 > 0, β13 < 0 and β15 > 0). We do not expect the valuation weights on earnings and book value of equity to change systematically with the level of short-term institutional ownership.

Table 5 reports our findings from estimating Equation (6). While the coefficients on NI SHORT, BV SHORT and BV SHORT D are either insignificant or in the wrong direction; the coefficient on NI LONG is positive and significant (β12 = 5.1573, t = 2.59), the coefficient on BV LONG is negative and significant (β13 = –0.8372, t = –2.80), and the coefficient on BV LONG D is positive and significant (β15 = 1.3057, t = 3.66). These results support Hypotheses 1 and 2 for long-term, but not short-term, institutional investors and provide further evidence that the valuation effect of institutional investors on earnings and book value of equity is consistent with institutions playing a positive governance role.

Table 5 Different Types of Institutional Ownership and Valuation Weights

Variable Predicted Sign Estimate

Intercept +/– 0.0327(0.07)

NI + 7.3242(9.28)***

BV + 0.7689(6.54)***

D +/– 0.1256(0.27)

NI D – –8.0712(–9.43)***

BV D + –0.1226(–0.89)

SHORT +/– 58.4349(14.78)***

NI SHORT +/– –6.4265(–1.69)*

BV SHORT +/– –0.8033(–1.35)

NI SHORT D +/– 10.2737(1.88)*


171

BV SHORT D +/– 0.7811(0.80)

LONG +/– 11.5868(7.02)***

NI LONG + 5.1573(2.59)***

BV LONG – –0.8372(–2.80)***

NI LONG D +/– –4.8075(–2.07)**

BV LONG D + 1.3057(3.66)***

Adj. R2 0.5939# of Obs. 13,120

We define percentage ownership by transient institutions as institutional investors with short investment horizons, SHORT. We define percentage ownership by quasi-indexer and dedicated institutions as institutional investors with long investment horizons, LONG.The t-values are computed using the Huber-White procedure, which corrects for heteroskedasticity and auto-correlation.*, **, and *** denote significance at the 0.10, 0.05 and 0.01 level, respectively, based on a two-tailed t-statistic.

5. Summary and Conclusion

This paper examines the effect of institutional ownership on the market valuation of earnings and book value of equity. Prior studies on institutional investors suggest that they play a positive role in corporate governance (the governance role) and that they prefer to invest in financially healthy firms (stock picking role). In addition, prior valuation research suggests that the market valuation weights on earnings and book value of equity are a function of firms’ financial health. We hypothesize that the valuation weight on earnings (book value of equity) increases (decreases) as the level of institutional ownership increases for profit firms whereas the valuation weight on book value of equity increases as the level of institutional ownership increases for loss firms. We find evidence supporting our hypotheses.

We conduct several tests to investigate whether our findings are more consistent with the governance role or stock picking role of institutional investors. We find that the level of institutional ownership continues to affect equity valuation weights on earnings and book value of equity in the expected directions after we incorporate measures of current financial health and that the residual level of institutional ownership, which is unrelated to measures of current financial health and growth opportunities, also affects equity valuation weights on earnings and book value of equity in the expected directions. These findings suggest that the level of institutional ownership captures value-relevant information that includes but is above and beyond current measures of financial health. Further, we find that the valuation effect of institutional ownership is mainly driven


172

by institutions with long investment horizons and monitoring incentives rather than institutions with short investment horizons and little monitoring incentives. Overall, our findings on the valuation effect of institutional ownership are more consistent with the governance role than with the stock picking role.

References

Admati, A. R., P. Pfleiderer, and J. Zechner (1994) “Large Shareholder Activism, Risk Sharing, and Financial Market Equilibrium”. Journal of Political Economy 102, 1097-1130.

Badrinath, S. G., G. D. Gay, and J. R. Kale (1989) “Patterns of Institutional Investment and the Managerial ‘Safety-Net’”. Journal of Risk and Insurance 56, 605-629.

Barth, M. E., W. H. Beaver, and W. R. Landsman (1998) “Relative Valuation Roles of Equity Book Value and Net Income as A Function of Financial Health”. Journal of Accounting and Economics 25, 1-34.

Bartov, E., S. Radhakrishnan, and I. Krinsky (2000) “Investor Sophistication and Patterns in Stock Returns after Earnings Announcements”. The Accounting Review 75, 43-63.

Beaver, W. (2002) “Perspectives on Recent Capital Market Research”. The Accounting Review 77, 453-474.

Bushee, B. J. (1998) “The Influence on Institutional Investors on Myopic R&D Investment Behavior”. The Accounting Review 73, 305-333.

Bushee, B. J. (2001) “Do Institutional Investors Prefer Near-Term Earnings Over Long-Run Value?”. Contemporary Accounting Research 18, 207-246.

Burgstahler, D., and I. Dichev (1997) “Earnings, Adaptation and Equity Value”. The Accounting Review 72, 187-215.

Burgstahler, D., J. Jiambalvo, and E. Noreen (1989) “Changes in the Probability of Bankruptcy and Equity Value”. Journal Accounting and Economics 11, 207-224.

Callen, J. L., O. Hope, and D. Segal (2005) “Domestic and Foreign Earnings, Stock Return Variability, and the Impact of Investor Sophistication”. Journal of Accounting Research 43, 377-412.

Carleton, W. T., J. M. Nelson and M. S. Weisbach (1998) “The Influence of Institutions on Corporate Governance Through Private Negotiations: Evidence from TIAA-CREF”. Journal of Finance 53, 1335-1362.

Chan, S. H., J. D. Martin, and J. W. Kensinger (1990) “Corporate Research and Development Expenditures and Share Value”. Journal of Financial Economics 26, 255-276.

Chung, R., M. Firth, and J. Kim (2002) “Institutional Monitoring and Opportunistic Earnings Management”. Journal of Corporate Finance 8, 29-48.

Collins, D. W., M. Pincus, and H. Xie (1999) “Equity Valuation and Negative Earnings: The Role of Book Value of Equity”. The Accounting Review 74, 29-61.

Del Guercio, D. (1996) “The Distorting Effect of the Prudent-Man Laws on Institutional Equity Investments”. Journal of Financial Economics 40, 31-62.

Feltham, G., and J. Ohlson (1995) “Valuation and Clean Surplus Accounting for Operating and Financial Activities”. Contemporary Accounting Research 11, 689-732.


173

Gillian, S. L., and L. T. Starks (2000) “Corporate Governance Proposals and Shareholder Activism: The Role of Institutional Investors”. Journal of Financial Economics 57, 275-305.

Gompers, P., and A. Metrick (2001) “Institutional Investors and Equity Prices”. Quarterly Journal of Economics 116, 229-259.

Hand, J. (1990) “A Test of the Extended Functional Fixation Hypothesis”. The Accounting Review 65, 764-780.

Hanlon, M., S. Rajgopal, and T. Shevlin (2003) “Are Executive Stock Options Associated with Future Earnings?”. Journal of Accounting and Economics 36, 3-43.

Hartzell, J. C., and L. T. Starks (2003) “Institutional Investors and Executive Compensation”. Journal of Finance 58, 2351-2374.

Hessel, C. A, and M. Norman (1992) “Financial Characteristics of Neglected and Institutionally Held Stocks”. Journal of Accounting, Auditing and Finance 7, 313-330.

Jiambalvo, J., S. Rajgopal, and M. Venkatachalam (2002) “Institutional Ownership and the Extent to Which Stock Prices Reflect Future Earnings”. Contemporary Accounting Research 19, 117-145.

Lev, B., T. Sougiannis (1996) “The Capitalization, Amortization, and Value-Relevance of R&D”. Journal of Accounting and Economics 21, 107-138.

Ohlson, J. (1995) “Earnings, Book Values, and Dividends in Security Valuation”. Contemporary Accounting Research 11, 661-687.

Shleifer, A., and R. W. Vishny (1986) “Large Shareholders and Corporate Control”. Journal of Political Economy 94, 461-488.

Shleifer, A., and R. W. Vishny (1997) “A Survey of Corporate Governance”. Journal of Finance 52, 737-783.

Smith, M. P. (1996) “Shareholder Activism by Institutional Investors: Evidence from CalPERS”. Journal of Finance 51: 227-252.

Wahal, S. (1996) “Pension Fund Activism and Firm Performance”. Journal of Financial and Quantitative Analysis 31, 1 – 23.

Walther, B. R. (1997) “Investor Sophistication and Market Earnings Expectations”. Journal of Accounting Research 35, 157-179.

Institutional Investors, Financial Health, and Equity … Investors, Financial Health, and Equity Valuation Dan S. Dhaliwala, Oliver Zhen Lia*, and Hong Xieb aUniversity of Arizona

Documents