Private versus Public Corporate Ownership: Implications for Future Profitability. Kristian D. Allee Michigan State University

Transcription

1 Private versus Public Corporate Ownership: Implications for Future Profitability Kristian D. Allee Michigan State University Brad A. Badertscher University of Notre Dame and Teri Lombardi Yohn Indiana University Draft: September 2011 Preliminary and Incomplete Do not cite without permission. Abstract: There is a long history of academic research that attempts to provide insight into the effect of public ownership on firm performance. We extend this line of research by evaluating the differences in future profitability between publicly traded and privately held firms. We posit that private firms are more profitable than public firms and this is mainly driven by differences in the profit margin of the firms. We empirically test these predictions using a matched sample of private and public firms. Our results indicate that private firms have greater future profitability than public firms. We also find differences in the future profit margin but not future asset turnover of the private firms relative to the public firms. These results provide useful evidence in understanding the overall effect that public and private ownership has on the future profitability. JEL classification: Keywords: private companies, profitability, future performance, and ownership structure All errors are our own.

2 I. INTRODUCTION The advantages and disadvantages of public corporate ownership have been debated for years (Berle and Means 1932). Public ownership allows for greater access to credit, enhanced stock-based management compensation packages, external monitoring of the business, and to greater publicity for the company. These advantages of public ownership have the potential to increase investment opportunities, attract the best employee talent, and lead to an enhanced reputation relative to what would be possible if the company were private. On the other hand, there are potential disadvantages of public ownership. The diffuse ownership and separation of ownership from control could potentially create agency problems such that managers do not make decisions that are in the best interest of shareholders. The cost of regulation, especially since the implementation of the Sarbanes-Oxley Act (SOX), can be daunting, and the disclosure requirements can diminish the competitiveness of the business. In addition, public monitoring can lead managers of public companies to focus on short-term metrics instead of long-run future profitability. In general, there are fundamental differences between private and public ownership that could potentially affect corporate growth and profitability. Public ownership could aid companies in generating higher future profitability or it could inhibit companies from meeting their potential relative to private ownership. Warren Buffett has noted the potential concerns with publicly traded companies and has recently demonstrated a preference for investments in private companies. Hough (2011), in a discussion of Warren Buffett s letter to Berkshire Hathaway shareholders, states that Wall Street s short-sighted focus on stock earnings hinders company performance, whereas private companies are free to prosper. Google (2004), prior to its initial public offering, noted that outside pressures too often tempt [public] companies to sacrifice long term opportunities to meet quarterly market expectations." Furthermore, Berman (2011) notes that a 1

3 benefit of private investment into Automated Data Systems is that all the margin information, so useful to competitors, stays hidden from public view. These arguments suggest that there are tradeoffs associated with being publicly owned that could hinder future profitability. Therefore, the benefits of public ownership associated with greater access to credit, publicity, and management talent could be completely offset by agency issues, a short-term focus from market pressures, and regulatory and disclosure requirements associated with public ownership. This could lead public companies to experience lower future profitability relative to their private counterparts. Despite the interest in both practice and academics on the effect of private versus public ownership structure on firm performance, little research has examined this issue. The lack of empirical evidence is primarily due to the lack of data on private companies. In this study, we are able to provide empirical evidence on the relative future profitability of private and public companies by using data from Sageworks Inc., which has recently made private firm data more available to academic researchers. The Sageworks database contains balance sheet and income statement data for more than 100,000 unique private firms over the period 2001 to 2010 and was designed to assist accounting firms and banks in performing analytical procedures and ratio analyses on private clients. These data allow our study to take a first step toward examining the relative differences in future profitability between public and private companies from a broad set of industries using actual financial statement data. We hypothesize, after matching on year, industry, size, and current profitability, that private companies experience higher future operating profitability than public companies. We also hypothesize that the greater profitability for private companies is driven by higher future profit margins rather than higher future asset turnovers. Finally, we recognize that public companies will 2

4 likely experience lower costs of debt due to the disclosure requirements of a public company, but also suggest that this will not offset the hypothesized lower operating profitability resulting from diminished profit margins. Therefore, we predict that private companies will have higher future return on equity. To test these hypotheses, we first match public and private companies based on year, industry, firm size (net operating assets), and current profitability (return on net operating assets, profit and loss firms). We then examine whether firm profitability, defined as return on net operating assets, one, three, and five years ahead, differs between public and private firms, holding constant the matched variables in univariate results. In multivariate tests, we control for current growth and profitability as well as the components of profitability. We also test for differences between private and public companies with respect to the components of return on net operating assets, asset turnover and profit margin, to improve our understanding of the differences in profitability between public and private firm ownership. Finally, we investigate differences in the cost of debt for public and private companies, and test for differences in future return on equity between public and private firm ownership to assess whether financing costs affect future profitability differentially. We find evidence consistent with our predictions. Specifically, we find that private companies have significantly higher return on net operating assets one, three, and five years ahead. We find that the relatively higher profitability for private versus public companies is driven by higher profit margins. We find no significant difference in future asset turnover between public and private companies. We also find that while private companies experience a significantly higher cost of debt, private companies also experience a higher return on equity after controlling for the relative disadvantage of debt in their capital structure. 3

5 These results suggest that, for private and public companies in the same industry, of similar size, and similar current profitability, private companies experience higher future profitability in the form of return on net operating assets, profit margins, and return on equity. These results provide evidence on the overall effect of public versus private ownership on corporate future profitability. These results provide empirical evidence helpful to managers, lenders, and equity fund managers, both public and private, in terms of the future costs and benefits associated with ownership differences. Our results are also consistent with differences in future profitability deriving from the costs associated with being a public firm and not necessarily from the financial reporting pressures of reporting consistent and increasing earnings as many researchers and practitioners have suggested. This study contributes to the literature on the effect of ownership structure on firm performance. Prior work provides evidence on the effect of specific corporate ownership characteristics on current profitability (Demsetz and Lehn 1985; Demsetz and Villalonga 2001). While these studies examine the relation between current profitability and specific ownership structures within public companies, they do not provide evidence regarding the overall effect of public versus private ownership on current or future profitability. This study also contributes to the literature that examines firm performance before and after an initial public offering (Pagano, et al. 1998; Jain and Kini 1994; Mikkelson, et al 1997) or a public offering via a reverse leveraged buyout (DeGeoge and Zeckhauser 1993). Our study contributes to this literature by examining a sample of private firms compared to a matched sample of public firms. By examining the future profitability of these two samples, as opposed to examining companies around public offerings, the observed future performance is not influenced by the effects of the public offering itself. Our sample of firms are, therefore, likely to be more 4

6 stable, which allows for a more complete analysis of the impact of private versus public ownership on firm performance and is able to provide additional insight into the association between private versus public ownership and long-run performance. The remainder of the paper is organized as follows. Section II provides some background and develops the hypotheses. Section III describes sample selection and matching procedures, and provides descriptive statistics. Section IV presents the results of the empirical analyses, and Section V offers concluding remarks. II. BACKGROUND, HYPOTHESES DEVELOPMENT, AND RESEARCH DESIGN Prior Literature There is a long history of academic research that attempts to provide insight into the effect of public ownership on firm performance. Much of the research has examined the association between specific ownership characteristics and firm performance. For example, Himmelberg et al. (1999) and Demsetz and Lehn (1985) find no relation between return on assets and managerial ownership. Demsetz and Villalong (2001) find no relation between various ownership structure characteristics and firm performance. Holderness and Sheehan (1988) and Denis and Denis (1994) find no association between firm performance and the diffusion of ownership. Other research has examined specific costs associated with ownership structure. For example, Ang et al. (2000) find that agency costs increase with non-managerial ownership. Research has also examined the performance of companies after initial public offerings to assess the association between public ownership and firm performance. For example, Pagano et al. (1998) find a reduction in the profitability of companies after an initial public offering for a sample of Italian companies. Jain and Kini (1994) find a reduction in operating profitability after an initial 5

7 public offering for a sample of U.S. firms. Mikkelson et al. (1997) find reduced profitability from prior to the initial public offering to the end of the first year after public offering for a sample of U.S. firms, but find no further decline in profitability for the ten years after the public offering. DeGeorge and Zeckhauser (1993) find a reduction in firm performance after companies go public through a reverse leveraged buyout. While these studies provide insight into whether specific ownership characteristics are associated with firm performance or whether having an initial public offering is associated with a change in profitability, it is difficult to assess whether public companies are associated with differential future profitability relative to private companies. Research on specific ownership characteristics and company performance cannot provide insight into the overall effect of public versus private ownership on firm performance. In addition, research on changes in company performance around public offerings is unable to provide insight into whether there are differences in firm performance between public and private companies in a stable environment when there is not a significant inflow of capital. A direct empirical comparison of future profitability between private versus public firms would provide insight into this issue. However, the difficulty in obtaining data on private firms has led to a lack of research in this area. Some studies have overcome the data problem by focusing on regulated industries, such as banking and insurance companies, or using data collected from surveys of private companies. For example, Ke et al. (1999) perform a univariate comparison of a sample of 45 privately-held property-liability insurers and 18 publicly-held property-liability insurers, and detect no significant difference in profitability between public and private firms. However, this analysis examines a small sample of firms within one industry. Using the Forbes survey of the 500 biggest private companies in the United States, Coles et al. (2003) find that private firms are less profitable than 6

8 similar public firms, when profitability is measured as operating margin and profit margin. However, Coles, et al (2003) uses estimated data and does not employ a matched pair research design. We contribute to this stream of literature by presenting empirical evidence on differences in future profitability between a matched sample of private and public companies. Hypotheses Development The public corporation is believed to have numerous advantages over its private counterpart (Renneboog et al. 2007). For example, public firms are likely to be able to invest in more profitable projects to due greater access to capital. Additionally, public firms have access to more media exposure, greater publicity and an increased reputation. Public firms have greater use of stock price-based compensation packages that can attract the best employee and management talent. All of these factors could result in public companies outperforming private companies in terms of long-run future profitability. On the other hand, recently there have been arguments that public companies may be less profitable in the long-run than private companies. The agency conflict in which managers may not act in the best interest of shareholders is the most debated potential disadvantage of public ownership (Berle and Means 1932). Detailed disclosure requirements for public companies may also hinder profitability (Pagano and Roell 1998). In fact, Brau and Fawcett (2006) find, in a survey of CFOs, that disclosing information to competitors and SEC reporting requirements are among the five most important reasons why private firms remain private in the U.S. Therefore, firms organized as private companies have distinctive long-run advantages over public companies due to increased exposure of public firm corporate strategies and trade secrets. In addition, the cost of regulation, especially since the implementation of SOX, can hinder profitability. CFOs surveyed by Brau and Fawcett (2006) suggest that expenses associated with listing requirements 7

9 imposed by securities exchanges, SEC rules and regulations, and accounting requirements for public companies, estimated at over a million dollars annually, can affect long-run profitability (Hartman 2006). Moreover, organizing as a private firm avoids some of the pressures of myopic investment decisions due to demands from short-term oriented investors (Stein 1988; 1989). 1 For example, Beatty et al. (2002) compare samples of publicly and privately held bank holding companies because they expect public banks diffuse shareholders to be more likely than private banks more concentrated shareholders to rely on simple earnings-based heuristics in evaluating firm performance. Beatty et al. (2002) expect public banks managers to face more pressure than private bank managers to report earnings in line with expectations and find evidence consistent with their expectations. In summary, while we recognize there are valid reasons to expect public ownership to enhance a firm s performance, we hypothesize that these will be overpowered by the costs associated with public ownership. Specifically, public ownership accentuates agency conflicts, requires disclosure of detailed information to competitors, is relatively more costly due to listing requirements imposed by securities exchanges and the SEC, and can lead managers of public companies to focus on short-term metrics instead of long-run future profitability in order to report earnings in line with expectations. This leads to our first hypothesis: H1: The future operating profitability is higher for private companies relative to the public companies. Operating profitability, defined as return on net operating assets, is a multiplicative function of the company s asset turnover and profit margin (Fairfield and Yohn 2001). Public 1 Charles Koch, chief executive of Koch Industries Inc., the United States second-largest private company, claims chief executives that obsess about delivering those ever-increasing and predictable quarterly earnings are going to sacrifice long-term value in the end (Shlaes 2007). 8

10 ownership could affect the two components of profitability differentially. Asset turnover is likely to differ between public and private companies if public companies obtain a capital infusion and overinvest in projects such that each incremental dollar of investment in net operating assets generates fewer sales. However, prior research suggests that companies tend to decrease investment and decrease leverage after initial public offerings. This is consistent with Myers and Majluf (1984) and suggests that overinvestment into less effective net operating assets is not likely to occur with public ownership. On the other hand, public ownership does require increased disclosures about the profitability of individual segments and costs. The increased disclosure requirements are likely to expose information on margins to competitors (Berman 2011). This suggests that public ownership might lead to greater competition in high margin products, which will eventually lead to lower margins for the publicly disclosing firms. This leads us to predict that public ownership is likely to lead to lower future profit margins relative to private ownership in our second hypothesis: H2: The future profit margin is higher for private companies relative to public companies. As noted above, public ownership allows for greater access to capital, leads to improved reputation, and requires improved disclosures about the company. These factors are likely to benefit public companies with respect to access to and cost of debt financing. Consistent with this notion, Pagano et al. (1998) document that Italian public companies experience lower borrowing costs than private companies subsequent to an initial public offering. A company s return on equity is a function of its operating profitability (return on net operating assets) as well as its leverage and the spread between operating profitability and the cost of debt. We hypothesize above that public companies will experience lower future operating profitability. We also expect public companies to have a lower cost of debt relative to private companies. However, we argue 9

11 that the effect of public ownership on return on net operating assets is likely to outweigh the effect of public ownership on the cost of debt with respect to the company s return on equity. We make this assertion based on the notion that public companies are likely to rely less on debt after an equity offering (Pagano et al. 1998). Therefore, the benefit of the lower cost of debt is likely to have a small effect on return on equity. This leads to our third hypothesis: H3: The future return on equity is higher for private companies relative to public companies. Research Design To test our hypotheses that private companies are likely to experience greater future profitability than public companies, primarily due to differences in profit margins, we estimate the regression model shown in equation (1). 2 (1) The left-hand side variable is a future profitability metric of interest. We evaluate the following future profitability metrics: (1) future return on net operating assets, measured as RNOA t+1, RNOA t+3, and RNOA t+5 ; (2) future profit margin, measured as PM t+1, PM t+3, and PM t+5 ; and (3) future asset turnover, measured as ATO t+1, ATO t+3, and ATO t+5. RNOA is calculated as net operating income (before any financing costs or investment income) in the numerator, and average net operating assets (operating assets net of operating liabilities) in the denominator 2 We examine all of our hypotheses using a matched-pair design which is described later in this section. 10

12 (Fairfield et al. 2009). 3 PM is the firm profit margin and equals operating income divided by sales and ATO is asset turnover defined as sales divided by net operating assets. RNOA is the change in RNOA as described above from t-1 to t. NOA is the change in average net operating assets from t-1 to t. ATO and PM are the changes in ATO and profit margin from t-1 to t, respectively. The indicator variable PRIVATE is coded one when the firm is a private firm and zero otherwise. The coefficient of interest for our first and second hypothesis is b 1 for which we expect a positive and statistically significant sign on the coefficient. We also run equation (1) using future RNOA as the dependent variable for private firms with audited financial statements, and their matched public firms to examine the future profitability differences between public and private firms holding constant the presence of audited financial statements. 4 Although this specification reduces our sample size, we want to be sure that our results are not driven by non-audited private firms. To test our last hypothesis, H3, we analyze the future profitability of public and private firms after including financing costs by examining future ROE, defined as net income divided by stockholders equity. The difference between RNOA and ROE is generally related to the capital structure of the company. All else equal, a more levered firm will have a higher ROE than a less levered firm and that will result in a larger gap between RNOA and ROE, but that difference also depends on the ability of the firm to earn a return on net operating assets greater than the cost of the debt incurred. Public firms likely have greater access to and lower costs of debt due to the disclosure, audit, and regulatory requirements of the SEC. Our univariate results confirm that 3 Operating income equals sales minus costs of goods sold, overhead costs, and depreciation and amortization. Net operating assets equals stockholders equity minus cash and short term investments plus interest plus debt in current liabilities plus long-term debt. 4 Note here that we recognize that we cannot hold constant the cost of the audited financial statements and that the cost of the audit for a public firm is likely higher than the cost of the audit for a private firm due to the increasing costs of litigation for the auditor (Badertscher et al. 2011). 11

13 public firms indeed have lower cost of debt relative to private firms. However, since public firms have access to public equity, they are unlikely to finance their operations through significant increases in debt. Thus, while there is a benefit associated with being public, on the margin, we hypothesize that it is not likely to reverse the effects we hypothesize for RNOA after incorporating the effects of debt. Therefore, ROE t+1, ROE t+3, and ROE t+5 are the dependent variables in equation (2) and H3 predicts a positive and statistically significant coefficient on c 1 below. (2) New variables in equation (2) are defined as follows: BVE is the change in book value of equity from t-1 to t; LEV is long-term debt divided by the book value of equity; and LEV is the change in that variable from t-1 to t. As stated earlier, leverage will only be a benefit to current and future ROE when the firm earns a return on net operating assets greater than the cost of the debt incurred from the liability. This difference is often referred to as the spread. The importance of spread for increasing ROE can be established by disaggregating ROE as follows: (3) According to equation (3), return on equity can be increased by i) increases in the rate of return on the firm s net operating assets, ii) increases in leverage, and iii) decreases in the cost of debt relative to the return on net operating assets (Wahlen et al. 2011). Thus, we also examine ROE t+1, ROE t+3, and ROE t+5 by breaking current ROE into its three components (i.e., RNOA, Leverage, and Spread). Specifically, we examine the d 1 coefficient in equation (4) to further test H3. 12

14 _ _ _ _ _ _ _ _ (4) Because of sample limitations we define SPREAD_POS (SPREAD_NEG) equal to one if RNOA is greater (less) than the cost of debt and zero if the cost of debt is missing because the firm has no debt and/or interest expense. All other variables are previously defined. Matching Procedure Our research question involves examining the performance of public and private firms relative to one another. As documented by prior research, public companies are substantially larger than private companies. Using a sample of private and public firms, Asker et al. (2010) provide evidence that the median public firm has total assets of $246.2 million, compared to $1.3 million for private firms. To ensure that our results are not driven by size we use a matched dataset designed to identify large private companies and small public companies. Specifically, we match public and private firm-years based on fiscal year, firm industry, net operating assets, return on net operating assets, and whether the firm reported a loss. Consistent with Asker et al. (2010), our matching procedure is a variant of nearest-neighbor matching used in the program evaluation literature (Imbens and Wooldridge 2009). Starting in 2001, for each private firm, we identify a public firm in the same four-digit NAICS industry (equivalent to three-digit SIC), same fiscal year, closest in terms of net operating assets (NOA), such that max(noapublic, NOAprivate) / min(noapublic, NOAprivate) < 2, closest in terms of return on net operating assets (RNOA), such that max(rnoapublic, RNOAprivate) / min(rnoapublic, RNOAprivate) < 2, and whether the private and public firms 13

15 are loss firms (pre-tax net income <0). If no match can be found in a given fiscal year, the private observation is discarded and a new match is attempted for that firm in the following year. Once a match is formed, it is kept intact for as long as both the public and private firms remain in our sample. For new firms added after 2001 we follow the same approach in each year they have data. III. DATA SOURCES, SAMPLE SELECTION, AND DESCRIPTIVE STATISTICS Data Sources and Sample Selection Our database combines data on private companies obtained from Sageworks Inc., a firm that collects private firm data and develops financial analysis tools, with public companies obtained from Compustat. The Sageworks database was designed to assist accounting firms and banks performing analytical procedures and ratio analyses on private clients for benchmarking purposes. In order to conduct such analyses, Sageworks users input their clients financial statement information into the Sageworks system which then becomes part of the collective database used in our study. As a result, Sageworks obtains financial statement information directly from the private companies auditors or banks and not from the private firms themselves. Sageworks is similar to Compustat in that it contains accounting data from the income statement and balance sheet. In addition to financial information, the private firm s four digit NAICS industry code, legal form (S-Corp, C-Corp, partnership, limited liability), fiscal year end, state, and type of audit report (e.g., compilation, review, tax return, or audit) are also available. The auditors that utilize Sageworks software include most national mid-market accounting firms as well as hundreds of regional audit firms. Unlike Compustat, Sageworks exclusively covers private firms and all data are held anonymously so that no individual firm can be identified. Firms leave the Sageworks database due to mortality or switching to an auditing firm that does not utilize the 14

16 Sageworks software. Sageworks has a dedicated staff of accounting and programming specialists who review, examine, and monitor the data on a continuous basis. To construct our sample of private companies, we follow a similar process as Minnis (2011) and exclude from the Sageworks database all non-u.s. based companies as well as observations with data quality issues. Specifically, we delete all firm-years that fail to satisfy basic accounting identities as well as when net income (NI), cash flow from operations (CFO), accruals (ACC), or property, plant and equipment (PPE) are greater than total assets at year-end. We also require firms have assets and sales greater than $100,000 and must be of legal form C-Corp to ensure comparability to public firms. To be part of the sample of public firms, a firm must have non-missing amounts of assets, sales, and net income from Compustat during our sample period, be incorporated in the U.S. and have equity that is publicly traded. Consistent with prior literature, we exclude from both the public and private samples financial firms (NAICS 52) and regulated utilities (NAICS 22). Both the public-firm and privatefirm samples cover the period from 2001 (the beginning of the Sageworks database) through 2010, giving us a ten-year panel of data. After requiring lagged data to calculate changes, current data for examination, future data of at least one year, and requiring a public match as described above we are left with 642 unique private firms in the sample. We match these private firms with their nearest neighbor public firm allowing public firms to serve as a neighbor for more than one private firm if necessary. The final sample contains 1,196 private firm-years and an equal number of public firm-years. Descriptive Statistics The descriptive statistics in Table 1 are reported so as to ensure that our matching procedure worked as planned. Our matching procedure appears to have effectively held constant 15

17 firms current profitability and size between public and private firms within the same industry, in the same year, and either reporting profit or loss in the matching year. Specifically, we observe no statistical differences in the mean or median RNOA, ASSETS, NOA, or LOSS variables between public and private firms for our sample. However, we do observe that these firms are different in other aspects of financial performance and financial structure in the matching year. Public firm mean profitability measured in terms of total assets (ROA) is lower than private firm profitability on total assets, but that difference is not significant in terms of the median ROA. Furthermore, we observe in Table 1 that the mean levels of PM and ROE (ATO) are statistically significantly higher (lower) for private companies in the year of the match. [Insert Table 1 Here] Also in Table 1 we observe that private companies have more sales and have less debt than public companies. Consistent with our expectations as well as the prior literature (e.g., Pagano et al. 1998), we also find an economically and statistically significant difference in the cost of debt (COD) between public and private firms. Specifically, we find that on average publicly traded firms are charged approximately 400 basis points less than private firms for their debt and appear to benefit from the greater access to capital, improved reputation, and required disclosures about the company as expected. However, as noted in our hypothesis development, Table 1 reports that the proportion of debt used in the capital structure of these public companies is relatively small with the mean (median) long-term debt to average total assets of 7.7 (0.3) percent. Panel A of Table 2 reports descriptive statistics for all firm years of our sample, as opposed to only the descriptive statistics of the initial matching sample year (Table 1). Specifically, Table 2 Panel A reports the descriptive statistics for the sample of private and public firms for all years they are in the sample. From Panel A of Table 2 we observe that private firms are more profitable 16

18 than public firms on average, but this does not address potential future profitability differences in the firms. This result is similar to Asker et al (2010) who report a statistically significant difference in ROA between their sample of private firms and matched public firms also using the Sageworks database. However, Asker et al. (2010) do not look at future profitability of these firms or match on profitability characteristics. Rather, for their tests involving investment behavior of public and private firms, they attempt to control for these profitability differences. Our research question addresses whether the differences between public and private firms observed in Panel A of Table 2 and Table 1 of Asker et al (2010) are persistent over time. In Table 2, Panel A we observe that future profitability, measured as RNOA t+1, RNOA t+3, and RNOA t+5 are all significantly greater for private firms than for the matched public firms in our sample. This leads to preliminary univariate support for our first hypothesis. Additionally, it appears that the magnitude of the difference between samples is monotonically increasing in both means and medians as time passes from the current year to five years out. [Insert Table 2 Here] Figure 1 plots the median levels of RNOA over time for our matched firms. It shows that the monotonic increases observed in the difference between public and private firms from Panel A of Table 2 are a function of increasing profitability for the private firms and/or decreasing profitability for the matched public firms in our sample. We observe the same pattern if we plot mean levels of RNOA over time for our matched firms. [Insert Figure 1 Here] Panel A of Table 2 also provides further evidence for the expected difference in cost of debt between public and private firms. We find that there are significant differences in the cost of debt between our matched sample of public and private firms, not just in the matching year. Public 17

19 firms have significantly lower cost of debt and it appears that private firms pay a 65 (91) percent higher mean (median) premium than that of public firms for the debt in their capital structure. Examining the descriptive statistics on net income (NI) reported in Panel A of Table 2 we note that for our sample of matched firms there is no statistically significant difference in reported net income or the standard deviation of net income between public and private firms. There is an increasingly large literature discussing the pressures of managers of public corporations to report smooth and predictable earnings (Bartov 1993; DeAngelo et al. 1996; Barth et al. 1999; Beatty et al. 2002; Graham et al. 2005; Petrovits 2006; Allee et al. 2011). In a survey of more than 400 executives, Graham et al. (2005) report, [p]redictability of earnings is an overarching concern among CFOs and the executives believe that less predictable earnings command a risk premium in the market. We cannot generalize beyond our sample, but it appears that any profitability differences we observe in this paper are not likely due to differential pressure for predictably increasing earnings between public and private firms. Panels B and C of Table 2 report the Pearson correlations (above the diagonal) and Spearman correlations (below the diagonal) for the private and public firms in our sample, respectively. We see strong correlations between RNOA and its components ATO and PM, as well as RNOA and other profitability metrics such as ROA and ROE. We also observe strong correlations in the changes of RNOA and the changes in its components. There are few differences in the correlations of the variables between public and private firms (i.e., comparisons of the magnitudes and directions of the correlations between panels B and C of Table 2), except for the relation between current RNOA and future RNOA and the change in NOA ( NOA) and RNOA t+3 and RNOA t+5. It is apparent that current RNOA is much more highly correlated with future RNOA for private firms than public firms, consistent with our first hypothesis. Also, it 18

20 appears that NOA has differential effects on future profitability for public and private firms, consistent with profitability differences between the types of firms deriving from operating assets. IV. EMPIRICAL RESULTS Tables 3, 4, and 5 report the main results in this study regarding the differences in future profitability between our two samples. Panel A of Table 3 reports the results of equation (1) to test our first hypothesis on a multivariate basis. The results are consistent with our first hypothesis, both in the full regression and in the simplified regression (including only our PRIVATE firm variable, current RNOA, and the interaction term between the two variables). In both cases PRIVATE is positive and statistically significant for RNOA t+1, RNOA t+3, and RNOA t+5. 5 The statistical significance decreases as we get further into the future, going from a one sided p-value of less than for PRIVATE on RNOA t+1 to 0.04 for PRIVATE on RNOA t+5. However, we note that the sample size decreases across the horizons from 2,392 firm years for RNOA t+1, to only 202 firm years for RNOA t+5. The decreased statistical power is likely the cause of the lower significance. Not surprisingly, current RNOA is an important predictor of future RNOA for all future years examined suggesting persistence in RNOA for both public and private firms. However, we do not find differential persistence of RNOA between the two types of firms as the coefficient on PRIVATE*RNOA is not significant in most specifications. Interestingly, no other variable in our model is consistent in predicting future profitability of public or private firms. This suggests that, for our sample of firms, using the components of RNOA or their changes appears to be uninformative in the prediction of future firm RNOA, dissimilar to the literature on predicting public firm profitability (Fairfield and Yohn 2001). However, given that the prior results were 5 We test for statistical significance of the parameter estimates by using heteroskedasticity robust standard errors, in regressions with errors clustered by private firm and year. As the public firms in our sample can be used multiple times in the analysis, we also ran the regressions clustering by public firm and year and observe statistically similar results. 19

21 observed on larger public companies with more established production practices, this result is not necessarily inconsistent with those results. [Insert Table 3 Here] In Panel B of Table 3 we run equation (1) on our sample of firms, but we limit the analysis to only private firms with audited financial statements and their matched public firms who naturally have audited financial statements due to regulatory requirements. In Panel B of Table 3 we find similar results to Panel A of Table 3, but it appears that the requirement to have audited financial statements does attenuate the statistical significance of the future profitability differences between public and private firms, though the coefficients remain similar in size or are somewhat larger. We find a one sided p-value of for PRIVATE on RNOA t+1 and 0.06 for PRIVATE on RNOA t+5 when the sample is limited to firms with audited financial statements. It is difficult to determine whether this attenuation in the statistical significance of our results is due to the decrease in sample size or to a decrease in the future profitability effect observed in Panel A of Table 3. A decrease in the future profitability effect is consistent with the observation by CFOs surveyed in Brau and Fawcett (2006) that expenses associated with securities exchange listing requirements, SEC rules and regulations, and accounting requirements, estimated at over a million dollars annually, can severely hinder future profitability of public companies (Hartman 2006). It is also consistent with the literature on firms going private. Specifically, Block (2004) surveys 110 managers from a sample of 236 firms that went private between 2001 and 2003 and finds that the cost of being public is the number one reason for going private by smaller firms. This relates directly to the passage of the Sarbanes-Oxley Act in Therefore, requiring the private firms to have an audit appears to level the playing field a bit, but there are still future profitability differences between public and private firms in our sample. 20

22 Examining Table 2 we can see mean differences in the profit margin and asset turnover components of RNOA for public and private firms. In H2 we argue that public ownership leads to greater competition in high margin products through increased exposure via public disclosure requirements, which will eventually lead to lower corporate margins for public companies. Although it is plausible that asset turnover could differ between public and private companies, we do not think that this is likely. Therefore, we do not hypothesize differences in asset turnover between public and private firms but do expect a difference in profit margin between our sample firms. Figure 2 plots the median levels of PM and ATO over time for our matched public and private firms. We can see from these graphs that there appears to be a difference in PM for private firms relative to public firms, but that there does not appear to be consistent differences in ATO between the two samples. [Insert Figure 2 Here] Panels A and B in Table 4 examine profit margin (PM) and asset turnover (ATO), respectively. The results from the simple model and full model correspond with our hypothesis that public firms have lower profit margins in the future relative to private firms. We observe that, much like future RNOA results, future profit margins are difficult to predict and only current profit margin and RNOA are helpful in consistently predicting future PM. Panel B reports the results of the regression on equation (1) with future asset turnover, ATO t+1, ATO t+3, and ATO t+5, as the profitability metric of interest. Consistent with our discussion, we find that private firms do not differ consistently relative to public firms in terms of future asset turnover. Current ATO appears to be the only variable that is consistently predictive of future ATO, suggesting the relative importance of persistence in operating efficiencies. [Insert Table 4 Here] 21

23 To more fully examine the relative future costs and benefits associated with organizational form we also look at future profitability for public and private firms in our sample in terms of ROE t+1, ROE t+3, and ROE t+5. This allows for a test of our last hypothesis on the differential future profitability of public and private firms including financing costs. Figure 3 plots the median levels of ROE over time for our matched public and private firms. The plots reveal that while both firms seem to be exhibiting decreasing returns to equity over time, the public sample is decreasing at a greater rate than the private firms in our sample. This suggests that private firms may be more profitable than public firms even after factoring the beneficial costs of financing for public firms relative to private firms. [Insert Figure 3 Here] We test H3 by estimating equations (2) and (4) on our sample of matched firms in Table 5. Table 5, Panel A reports the results from the regression on future ROE specified in equation (2). We observe that private firms appear to still have higher future profitability, when future profitability is measured as ROE t+1, ROE t+3, and ROE t+5, even after controlling for the effects of leverage on the firms. The result in the fifth year out is somewhat attenuated relative to the results on RNOA reported in Table 3; however, it is still significant with a one-sided p-value of on only 202 observations. This is relatively persuasive support for our third hypothesis that, for private and public companies matched yearly on industry, size and current profitability, the future return on equity is higher for private companies relative to public companies. As in prior regressions, predicting future ROE using current profitability and its components is difficult. None of the other variables in the regression appear to consistently aid in prediction, including current ROE. [Insert Table 5 Here] 22

24 Panel B of Table 5 reports results consistent with and somewhat stronger than the results observed in Panel A of Table 5. After decomposing ROE into its RNOA, LEV, and SPREAD components, we find that private firm future profitability is greater than public firm future profitability and for ROE t+5 this result is significant at a one sided p-value of We also find that positive spread is a significant predictor of future profitability and that this effect is even somewhat greater for private firms than for public firms. We know of no theory that would suggest that a positive spread is more beneficial to the future profitability of a private firm than a public firm, but find it motivating that a careful decomposition of ROE into its components consistent with theory can aid in this type of analysis. Note that for ROE t+5 the decreasing sample size left us with a sample of firms reporting no negative spreads and no private firms with positive spreads and leverage; hence, we cannot estimate all variables and their interactions in model (5) for ROE t+5. V. CONCLUSION In this paper we examine the relative future performance of a sample of private and public companies matched each year on industry, size and current profitability. We hypothesize and find that private firms are more profitable in the future than public firms. We also hypothesize that this difference is likely due to differences in the profit margin of the firms and not the asset turnover component of their return on net operating assets. We acknowledge that there are financing benefits for being a public company. We find that there is a significant difference in the cost of debt between public and private companies likely due to the disclosure, audit, and regulatory requirements of the SEC. However, since public firms have access to public equity, they are less likely to finance their operations through debt. Thus, while there is a benefit associated with being 23

25 public, on the margin, we hypothesize and find that it only slightly attenuates the effects we observed in examining return on net operating assets. There are several limitations to our study that we acknowledge. First, due to sample size and time-series data limitations our ability to test future profitability is potentially limited. However, existing studies (e.g., Sloan 1996; Fairfield et al. 2003) use windows as short as one year ahead when examining future profitability and therefore our horizon is not inconsistent with, and some cases longer than, that used in prior research. Additionally, our results are generalizable only to the extent that the private firms in our sample and their corresponding public matches are representative of private and public companies in general. The Sageworks database is populated by firms that come in contact with auditors and bankers that take the extra time to benchmark their clients ratios and financial data against what they must believe to be a useful and archetypical sample of firms. However, this process potentially leads to particular types of firms being included in the Sageworks sample. However, this concern should be alleviated due to our matching on the same industry. Finally, the variables available to us to examine ownership structure, instead of public versus private ownership in general, are restricted and limit our ability to identify the specific mechanisms resulting in the observed differences. Our results provide evidence on the overall effect of public versus private ownership on corporate future profitability and provide empirical evidence helpful to managers, lenders, and equity fund managers, both public and private, in terms of the future costs and benefits associated with ownership differences. Our results are also indicative of differences in future profitability between private and public firms deriving from the costs associated with being a public firm and not necessarily from the financial reporting pressures of reporting consistently increasing earnings as many researchers and practitioners have suggested. Future research may wish to examine a 24

26 specific sample of public firms with these incentives and try to match on private firms to observe differences in financial reporting behavior and future profitability. 25

27 References Allee, K.D., S.J.W. Hamm, and D.D. Wangerin Have Accounting Standards Affected M&A Deal Structure? Evidence from Earnouts Working paper Michigan State and Ohio State Universities. Ang, J., R. Cole and J. Lin Agency Costs and Ownership Structure. The Journal of Finance 55: Asker, J., J. Farre-Mensa and A. Ljungqvist Does the Stock Market Harm Investment Incentives? Working paper New York University. Badertscher, B., B. Jorgensen, S. Katz, and W. Kinney Litigation Risk and Audit Fees: The Role of Public Equity. Working paper. Barth, M. E., J. A. Elliott, and M. W. Finn Market rewards associated with patterns of increasing earnings. Journal of Accounting Research 37 (2): Bartov, E The Timing of Asset Sales and Earnings Manipulation. Accounting Review 68 (4): Beatty, A. L., B. Ke, and K. R. Petroni Earnings management to avoid earnings declines across publicly and privately held banks. The Accounting Review 77 (3): Berle, A. and G. Means The Modern Corporation and Private Property. Chicago: Commerce Clearing House. Berman, D Going on Safari with Warren Buffett. The Wall Street Journal. March 1, Block, S.B The Latest Movement to Going Private: An Empirical Study. Journal of Applied Finance Vol. 14, No. 1 Brau, J. C., and S. E. Fawcett Initial public offerings: An analysis of theory and practice. Journal of Finance 61 (1): Coles, J. L., M. Lemmon, and L. Naveen A Comparison of Profitability and CEO Turnover Sensitivity In Large Private and Public Firms. Working Paper - Arizona State University. DeAngelo, H., L. DeAngelo, and D. J. Skinner Reversal of fortune - Dividend signaling and the disappearance of sustained earnings growth. Journal of Financial Economics 40 (3): DeGeorge, F. and R. Zeckhauser The Reverse LBO Decision and Firm Performance: Theory and Evidence. The Journal of Finance 48: Demsetz, H. and K. Lehn The Structure of Corporate Ownership: Causes and Consequences. Journal of Political Economy 93: