Printer Friendly

To be or not to be public: the impact of SOX.

Introduction

In the midst of the greatest economic/market meltdown since the Great Depression, the U.S. government has enacted a number of new laws/regulations to address the problems that played a role in the current crisis. Some commentators and others say that we should avoid rushing to put new rules in place as was done with the passage of the Sarbanes-Oxley Act (SOX) of 2002--which was passed on the heels of several major financial frauds which came to light in 2001 and early 2002. SOX is the most dramatic piece of legislation relating to shareholders and corporations since the 1930s.

The Sarbanes-Oxley Act, signed by President Bush on July 30, 2002, contains provisions that are intended to ensure more accurate disclosure of financial information to investors. The Act:

1. Empowers a public company accounting oversight board to inspect accounting firms. The accounting firms are charged an annual fee, and are assessed by the board every one to three years. (Section 1)

2. Allows public accounting firms to offer non-audit consulting services to an audit client only if the client's audit committee pre-approves the non-audit services to be rendered before the audit begins. (Subsection 201)

3. Prevents a public accounting firm from auditing a client firm whose CEO, CFO, or other employees with similar job descriptions were employed by the audit firm within one year prior to the audit. (Subsection 206)

4. Requires that the firm uses an independent audit committee, which consists of only outside board members, and one of those members must be a financial expert. (Subsection 301, Subsection 407)

5. Requires that firms with at least $75 million in assets that file 10-Ks improve their internal control systems (this provision was implemented as of November 15, 2004). (1, 2) (Subsection 404)

6. Requires that the CEO and CFO of firms that are of at least a specified size level certify that the audited financial statements are accurate, and holds them accountable for their verification. (Subsection 302)

7. Requires more timely and enhanced disclosure of the financial statements, (especially for off-balance sheet items). (Subsection 401)

8. Specifies major fines or imprisonment for employees who mislead investors or hide evidence. (Sections 8, 9 and 11)

9. Provides for the forfeiture of bonuses if financial statements are restated. (Subsection 304)

10. Eliminates personal loans. (Subsection 402)

All provisions listed above are intended to prevent fraudulent financial reporting, and increase the transparency of publicly-traded firms. However, some critics argue that SOX imposes a large penalty on publicly-traded firms by increasing the cost of disclosure. As a result of SOX provisions, publicly-traded firms are subject to larger audit fees, and higher costs of maintaining the required internal financial controls. Consequently, its provisions may have motivated firms to go private rather than incur the explicit and implicit costs of complying with the law's requirements. We investigate this issue by examining whether SOX affected the propensity of firms to go private, and the effect of SOX on valuation of firms going private.

Our study contributes to the literature in the following ways. First, we use a more focused definition of going private than previous related studies (discussed shortly), in which we examine firms whose shares no longer trade in any market and which no longer provide publicly available financial information. Therefore, we exclude firms that deregister ("go dark"). We think that our definition of going private provides the cleanest way to assess whether SOX encourages public firms to go private. Second, our sample is extended through 2005 so that we can assess a longer time period following SOX. This is relevant because of shift in the possible increased incentive to go private since SOX and the possible extra reward (valuation effect) for firms that have gone private since SOX. In particular, there was a resurgence in private equity deals, particularly after 2003 until the market collapse in the latter part of 2007. Third, we examine the reasons that are cited for going private, and test for those that are different after SOX compared to before SOX. We incorporate the most important reasons in our multivariate tests. Fourth, we test how the propensity to go private is affected by additional firm characteristics such as CEO ownership, institutional ownership, stock price performance, and four interaction terms that capture the incremental sensitivity of the decision to go private in response to firm characteristics since SOX. Fifth, we test how the valuation effects of going private are affected by firm characteristics, so that we can more closely assess the impact of a firm's ownership structure, governance, and the subsequent market response.

We identify a sample of 262 listed firms that went private over the years 1999-2005, a period which surrounds the Act. Our results indicate that SOX increased the propensity of public firms to go private, but that impact has been mitigated over time. We also find that firms are more likely to go private when they have high levels of CEO ownership, even more so post-SOX. Firms are also more likely to go private when they have poor stock market performance and liquidity, are out of favor with the market and are smaller in terms of assets since SOX became law.

In the period soon after the enactment of SOX, the valuation effect of going private is significantly greater than in the latter period of our study. Smaller firms experience larger valuation effects after SOX as do firms that cited cost of being public as a reason, and were in an industry that was out-of-favor with the market. Overall, the characteristics of firms going private and the market's perceptions of going private have changed since the passage of the Sarbanes-Oxley Act.

Related Literature

Lehn and Poulsen (1989) assess going private transactions that occurred between 1980 and 1987, and determine that the propensity to go private is positively related to the firm's cash flow and inversely related to the firm's growth rate. Slovin, Sushka, and Bendeck (1991) suggest that going private transactions are triggered by private information that a bidder possesses about the target. They assessed a sample of going private transactions in the 1980-1988 time period, and estimated the 2-day abnormal return to be 17.35 percent, and 6.49 percent over the 14 days prior to the two-day announcement window (statistically significant).

Leuz, Triantis and Wang (LTW, 2008) look at firms that either deregister or go private before and after the passage of SOX. The main focus of their paper is on firms that go dark. They find a large increase in going dark transactions following SOX, and significant negative cumulative abnormal returns both before and after SOX. These negative returns are attributed to poor future prospects, as well as increased costs associated with SOX. However, post-SOX, these returns are also associated with weak corporate governance or when there is weak protection for outside shareholders.

LTW also conduct an event study on 194 firms going private before SOX and 117 firms going private after SOX for the years 1998 through 2004. Their sample includes all listed companies as well as OTCBB and Pink Sheet companies. They find that cumulative abnormal returns are positive and significant for both the pre-SOX and post-SOX periods. Based on probit regression, they find that the characteristics of going dark and going private firms are different and therefore represent different corporate decisions.

Marosi and Massoud (2007) assess a sample of 261 going-dark firms. They find that firms with fewer valuable growth opportunities, greater insider ownership, lower institutional ownership, higher leverage and lower market momentum are more likely to go dark. They also find that investors experience significant valuation declines after the going dark announcement, along with significantly reduced liquidity. They attribute going dark transactions primarily to increased costs of auditing associated with SOX.

Engel, Hayes and Wang (EHW, 2007) assess going private transactions from the first quarter of 1998 through May 2005, but also include firms that go dark. Twenty-one percent of the firms in their total sample continued to trade in the Pink Sheets after deregistration, and 36 percent of the firms in their post-SOX sample continue to trade after deregistering with the SEC. They find that there has been a modest increase in going private transactions after SOX became law. They also find that firms that are smaller and have high inside ownership experienced more favorable stock price effects from going private announcements in the post-SOX period than in the pre-SOX period. Leuz (2007) discusses EHW's paper and questions whether their findings are supportive of SOX resulting in a significant increase in going private activity. He reiterates that going dark and going private transactions "are economically very different." (3) He argues that if going dark firms are excluded from EHW's sample, the increase in firms going private after SOX is no longer significant. He also points out that some of the positive returns that they observe may be due to the resolution of agency problems between controlling shareholders and the public shareholders. In this sense, SOX could be viewed as being beneficial to minority shareholders.

Finally, Kamar, Karaca-Mandic and Talley (KKT, 2009) conduct a cross-country analysis of firms going private between 2000 and 2004. Their U.S. sample includes companies from NYSE, AMEX, NASDAQ, as well as the Over-the-Counter Bulletin Board and Philadelphia Stock Exchange. However, they do not indicate how many going private transactions involved OTCBB firms. Moreover, in looking at acquisitions by private buyers, they do not distinguish between management led buyouts compared to those undertaken by private equity firms. KKT use going private transactions of firms in Western Europe and Canada as a control, since firms in these areas of the world were not subject to the provisions of SOX. Their results indicate that SOX resulted in small U.S. firms going private in the year after SOX was enacted. However, their results do not show a similar effect for larger firms. Finally, they conceded that their study may only be specific to the period they looked at.

Going Private

The Securities and Exchange Commission defines a going private transaction in Rule 13E-3 as a publicly traded firm that reduces its number of shareholders of record to less than 300 (500 for firms with assets of $10 million or less) or is no longer listed on a national exchange or quoted in an inter-dealer registered national quotation system. These firms can deregister with the SEC, and are no longer required to file with the SEC. (4) Some of these firms continue to trade in the Pink Sheets. Leutz et al. (LTW, 2008) have referred to this process as "going dark." However, some going dark firms continue to provide financial information to their shareholders. On the other hand, other firms go private by buying out the public shareholders (their shares no longer trade in any market), and no longer provide publicly available financial information about the company. Thus, the SEC definition of going private does not distinguish between firms that go dark and those which have fully withdrawn from our equity markets.

Going private involves a bid by management and/or an acquirer group for all outstanding shares not already owned by management or the bidder group. The bid price is negotiated between the acquirer and the board of directors. In many cases, the board appoints a special committee to evaluate the bid. The special committee typically hires its own lawyer as well as a financial adviser to represent the shareholders, and often solicits other offers for the firm. The board and the special committee have a legal obligation to review all bids, to reject any unfair offers, to maximize the price obtained for the shareholders, and to ensure that all bidders are treated fairly, even if the management bidding group is put at a disadvantage.

Marginal Benefits From Going Private

In theory, public firms should go private if the marginal benefits of doing so exceed the marginal cost. A firm that goes private avoids the costs cited in the following section. In addition, firms going private can save on the cost of annual stock exchange listing fees, the cost of making periodic filings with the SEC, implicit costs of providing potentially useful information to competitors, and the expenditure of management time and company resources in dealing with public shareholders (holding quarterly conference calls, holding the annual meeting, etc.). In addition, firms may be able to take a long-term focus on running the business by not having to meet analysts' and investors' quarterly earnings expectations.

Marginal Costs of Going Private

The decision to go private results in additional administrative expenses. There are direct costs of going private. The firm has to obtain legal advice and a fairness opinion from investment banks. Often, a special committee of the board is established (when required by state law) to evaluate the current bid and seek competing ones.

There are also indirect costs, including loss of the capital markets as a monitoring mechanism, lower liquidity, the loss of a transparent valuation, and a more limited ability to raise capital. There may be higher risk premiums on debt from using more financial leverage. In addition, going private may lead to potential lawsuits against managers for manipulating financial numbers to facilitate a buyout. When shareholders protest through litigation or a proxy fight, or when management seeks to undermine an outside bidder, the costs of legal fees and advisory services are even higher. These types of costs were incurred by firms going private before and after SOX.

However, there are some additional costs caused by SOX. The costs of an internal control system may have been most pronounced in the first year (especially for firms that did not have a system in place), but the system is supposed to be reevaluated each year, forcing firms to continue incurring additional costs. The Financial Executives Institutes (5) conducted a survey of Section 404 (internal control) implementation costs in March 2006 covering calendar 2005. Of the 274 companies which replied, 238 had market capitalizations above $75 million and average revenues of $5.7 billion. For these firms, the average total costs for Section 404 compliance were $3.8 million vs. $4.44 million in 2004 (the first year of implementation). In 2010, a survey by Financial Executives International found that the average audit costs for a public company were $3.3 million up 2% from 2009; those fees increased 2.2% in 2009 compared to the previous year (6).

Foley and Lardner, LLP (7) have conducted several annual surveys, and found that the average cost of being public for a company with less than $1 billion in revenues has increased by 36% in 2003, 33% in 2004, decreased by 16% in 2005, and increased by 1% in 2006. By comparison, the cost of complying with SOX for firms with revenues over $1 billion increased by 51% in 2004, decreased by 6% in 2005, and increased by 8% in 2006. Audit fees for companies with revenues of less than $1 billion increased by 16% in 2003, by 96% in 2004, by 16% in 2005, and by 4% in 2006. For firms with revenues greater than $1 billion, audit fees increased by 61% in 2004, by 2% in 2005, and by 5% in 2006. A 2004 Korn/Ferry International survey of Fortune 1000 firms estimated the cost of complying with SOX for the average firm was $4.1 million.

In addition to the explicit costs due to SOX, there are implicit costs of managers using more of their time for compliance rather than to manage operations. Furthermore, since the Act significantly increases penalties for noncompliance, some managers may make decisions that are focused on avoiding liability rather than on serving shareholders. Firms can also avoid the costs of reorganizing their boards in order to comply with Sarbanes-Oxley if they go private.

Hypothesized Impact of SOX on the Tendency to Go Private--Given the substantial increase in explicit and implicit costs of being public, we hypothesize that provisions of SOX had a more pronounced impact on the benefits than the costs of going private. Therefore, we expect that SOX should increase the tendency to go private. SOX could also increase the rewards to the shareholders of firms that go private, especially firms that were not obtaining the benefits of being public.

Hypothesized Impact of SOX on Valuation Effects of Going Private--Due to the increase in potential benefits from going private since SOX, we expect that the valuation effects of going private should be more favorable since SOX. A direct comparison of valuation effects from going private transactions after SOX versus before SOX period is conducted in order to test this hypothesis.

Methodology

We apply cross-sectional models to test our hypotheses regarding the impact of SOX on the likelihood of going private, and the valuation effects of going private. Our models include additional variables so that we can properly control for other factors while testing the impact of SOX on the tendency and valuation effects of going private.

Testing the Impact of SOX on the Tendency to Go Private

In order to test the impact of SOX on the firm's going private decision, we apply a cross-sectional model to firms that were publicly traded at the beginning of each year, in which the dependent variable is whether the firm went private during that year. Our main independent variable is a dummy variable SOX that is set equal to 1 for firms in years since the inception of SOX, and zero for firms in years prior to SOX. Our tests controls for other variables that could also influence the tendency to go private, which are described next.

CEO Ownership

CEOs who view their firm's shares as undervalued by the market may have a stronger incentive to go private if they own a relatively large proportion of the firm's shares. Since the CEO is likely to have some influence on the going private decision, firms may more easily go private without resistance if the CEO holds a relatively large portion of the shares. Halpern, Kieschnick, and Rotenberg (1999) find that firms with a high level of management ownership are more likely to go private. Goh et al (2009) discuss how management uses the going private transaction to "cash out". We use the percent of shares owned by the CEO (CEOWN) to measure CEO ownership

We also expect that the propensity for firms with high CEO ownership to go private would be more pronounced after SOX, since CEOs of these firms should be especially motivated to go private when the costs of remaining public increase. To test for this, we include an interaction variable CEO*SOX, where SOX is a dummy variable set equal to 1 for going private transactions that occur after SOX, and zero otherwise.

Management and Director Ownership

High-level (non-CEO) management and directors with share ownership may also have more to gain by cashing out in a going private transaction. Furthermore, they are better able to exploit their large ownership to achieve a majority of the votes to take the firm private. We use the percent of the shares held by directors and high-level executives (DEOWN) to measure this effect.

As described above for the CEO, we also expect that firms with high non-CEO executive/director ownership would have even more of an incentive to go private post-SOX because of the higher costs of remaining public after SOX. To capture this, we include an interaction variable DEOWN*SOX (dummy equal to one post-SOX).

Institutional Ownership

Many firms that have been surveyed cite the lack of institutional ownership as a reason for going private. We use the percent ownership by institutions (Instown) to measure this. We also test an interaction term Instown*SOX (dummy equal to 1 post-SOX) to determine whether institutional ownership after SOX has affected the likelihood to go private.

Recent Stock Price Performance

Firms that have experienced weak stock price performance recently may be more likely to go private, because their shares can be purchased by management or another buyer at a low cost. Halpern et al (1999) confirm that firms that go private tend to be underperformers, and might be viewed as being out-of-favor with investors. We use Runup to measure the percentage change in the stock price in the period from 200 days to 20 days before the announcement date.

Cash

Firms whose stock price is relatively low compared to cash per share may have reached a level of undervaluation that does not justify public listing. In addition, excess cash reduces the cost of taking the firm private. We use the ratio of equity market capitalization to cash (called Mktcash) to measure this variable.

Free Cash Flow

A firm that has a higher level of tree cash flow should have a higher debt capacity and thus be in a better position to use more debt with which to finance a going private transaction. Thus, a firm's free cash flow level should be positively related to the probability of going private. Free cash flow is defined as [Operating income--Taxes paid--Interest expense--Preferred dividends--Common dividends] / Average total assets (FCFAT). This is similar to the definition used by EHW (2007) and Lehn and Poulsen (1989).

Liquidity

If a firm has low stock market liquidity, it may be unable to attract investors or engage in a secondary offering of stock. Lack of liquidity is frequently cited by firms as one of the most important reasons for going private. We measure liquidity as trading volume in the year prior to the transaction divided by the average number of shares outstanding.

Profitability

Firms that are less profitable may be more willing to go private in order to eliminate the costs of being public. We use net income divided by assets (ROA) to measure profitability. We also include an interaction variable ROA*SOX to measure the effect of the firm's profitability on the likelihood of going private since SOX.

Price/Book Ratio

A firm should be more likely to go private if it believes its stock is undervalued by investors. In addition, a low multiple may indicate a lack of growth opportunities, which can also influence a firm's decision to go private (see Lehn and Poulsen, 1989). We hypothesize that the likelihood of a firm going private is inversely related to its price/book multiple (PBK).

A firm with a low price/book multiple might have more of an incentive to go private after SOX given their limited growth opportunities. We include an interaction variable PBK*SOX to measure the relationship between a firm's price/book ratio and the likelihood of going private since SOX.

Firm Size

Larger firms should exhibit economies of scale in their costs of reporting and compliance, while small firms would be hit hardest by the incremental costs of being public. Therefore, we expect that the likelihood of going private is inversely related to size. We use log of book value of assets (Logat) in the year prior to the going private transaction as a size variable.

Given the increased costs associated with complying with SOX in addition to the costs of already being public, it is possible that size may provide the most important incentive to go private after SOX. In order to see if this is the case, we interact assets with a SOX indicator variable equal to 1 if the firm goes private post-SOX (AT*SOX).

Testing the Impact of SOX on Valuation Effects of Going Private

In order to test the impact of SOX on valuation effects of going private, we also conduct a cross-sectional analysis that incorporates the possible impact of the Sarbanes-Oxley Act along with other control variables on the valuation effects. Some of the characteristics that can affect the incentive to go private may also affect the market's valuation in response to the announced going private transaction. We hypothesize that the valuation effects are associated with the following factors:

SOX Variable

The potential benefits and therefore valuation effects of going private post-SOX may have changed over time, as both market participants and managers adapt to SOX. We apply two dummy variables so that we can isolate the separate effects for the period following the passage of SOX until the end of 2003 compared to 2004-2005. While the cutoff dates are admittedly arbitrary, we expect that the firms that went private relatively soon after SOX was passed may have been able to extract the most benefits from going private, and therefore should experience larger valuation effects. Conversely, firms that did not go private until after 2003 may not have been as convinced that going private was the best solution in response to SOX. In addition, the market may have perceived that their intentions to go private are not directly attributed to SOX, since they waited for more than one year after the act to go private. A dummy variable SOX1 is set equal to 1 for going private transactions that occurred after SOX and through the end of 2003. A second dummy variable, SOX2 is set equal to 1 for going private transactions that occurred in the 2004-2005 period.

Cost Motive

Firms that go private only because of the cost savings may have less to benefit from going private than other firms which are perceived to be undervalued by the market. Therefore, firms that explicitly state their motive as cost savings may experience less pronounced valuation effects. A dummy variable, COBP, with a value equal to one is used if the cost of being public was cited as a motive.

Two interaction terms, (COBP*SOXI) and (COBP*SOX2) are also used to test whether the impact of SOX on valuation effects of going private is conditioned on the cost motive. While the SOX dummy variable already captures the potential increase in the valuation effects due to the cost of remaining public after SOX, the interaction term measures the incremental valuation effects when the market may perceive that the motive is cost savings versus possible undervaluation. The cost savings argument may be more valid shortly in the first period after SOX than later on.

Recent Stock Price Performance

Firms that have experienced weak stock price performance can go private at a low cost, and therefore may elicit more favorable valuation effects. As before, we use the variable Runup to measure recent stock price performance.

Management Buyout

A going private transaction can be initiated by the management of the firm or by a buyout specialist (Slovin, Sushka, and Bendeck, 1991). Buyout firms may have to pay a higher premium to take a firm private because they may need to offer a price that is acceptable to the firm's management in order to avoid resistance. We use a dummy variable (Mgmt) which is set equal to 1 for buyouts by management, and zero for buyouts by private equity firms or raiders.

CEO Ownership

In general, investors in firms with high CEO ownership may have less ability to affect the size of the buyout premium than firms with low CEO ownership. Thus, there should be an inverse relation between CEO ownership and valuation unless independent directors are successful in attracting higher bids. As before, we use the percent of shares owned by the CEO (CEOWN) to measure this variable.

Two interaction terms, (CEOWN*SOX1) and (CEOWN*SOX2), are also used to test whether the relation between CEO ownership and valuation effects of going private is affected by SOX in two periods since its passage.

Probability of Going Private

Since the market may anticipate (at least in part) a firm going private, we need to control for this. We do this by estimating the probability of this happening (called GP-Prob) for each firm using the results from conditional logit model 1. (8)

Institutional Ownership

Low institutional/diffuse retail ownership might result in a lower valuation effect than in the case of firms with high institutional ownership. We use the percentage of shares outstanding held by institutions (Instown) to measure institutional ownership.

Inside Directors

If managers make up a relatively high proportion of the board, they may have more influence, which could enable them to extract more value from a going private transaction. This would result in a lower value in the going private transaction. The variable, Pctinsid, is used to represent the percent of the board made up of insiders.

Conflicts of Interest

We expect that the valuation effects will be lower for firms that are subject to conflicts of interest, since these firms are more likely to try to minimize the price paid to shareholders. We use a dummy variable, Conflict, which is set equal to 1 if the firm reports that there are related party transactions or evidence of connected lending in the year prior to the going private transaction.

Special Committee

As indicated earlier, in many cases, a special committee is set up to evaluate the bid. We expect boards with special committees made up of independent directors to result in positive valuation effects. We use a dummy variable, Speccom, which is set equal to 1 if the special committee is independent.

Antitakeover

If a firm recently amended its shareholder rights agreement ("poison pill"), the board may have less bargaining power, and we would expect the valuation effects of going private may be lower. We use a dummy variable, Rights, which is set equal to 1 if a firm amended its shareholder rights agreement within two years of the going private bid to represent antitakeover conflicts.

Price/Book Ratio

Firms that have a high share price relative to book value may attract a lower premium in a going private transaction. Since the stock price already reflects growth opportunities, bidders would have little incentive to offer a large premium. We use the ratio of stock price to book value of equity (PBK) to measure this effect. Two interaction terms, (PBK*SOX1) and (PBK*SOX2), are also used to determine if there is a differential impact of SOX on PBK in the period up through 2003 and from 2004 through 2005.

Profitability

Firms that are less profitable may have more to gain from going private, because they can avoid the costs of remaining public which would make them even less profitable. Profitability is measured by the firm's industry-adjusted ROA.

Firm Size

The costs incurred by firms due to SOX can be more easily absorbed by large firms than by small firms. In addition, small firms are less likely to benefit from being public. Therefore, small firms may have an incentive to offer a higher premium in order to make it more likely that outside shareholders will accept the going private offer. Size is measured as the log of book value of assets in the year prior to going public (Logat).

Sample

We obtain a sample of firms flagged as going private in Securities Data Corporation's Mergers and Acquisitions Database. We only include U.S. firms which have traded on a national stock exchange--New York Stock Exchange, American Stock Exchange, or NASDAQ. We exclude firms trading on the OTC Bulletin Board (OTCBB) or in the Pink Sheets which went private. Since domestic firms trading in these markets account for a very small fraction of the market capitalization of all domestic equities trading in the U.S., there are no public policy implications from these firms going private as a result of SOX. We next conduct a Lexis Nexis search to identify the announcement dates for the going private transactions before conducting any empirical tests. The announcement date is the first date that information appears in the media that there has been a bid on the firm, not the filing of the SC13E-3 which can occur much later. We restrict our analysis to going private transactions that are completed and confirm the sample by examining 13E-3 and SC TO-T filings in EDGAR. We exclude firms that SDC has incorrectly identified as going private as a result of being acquired but which did not go private. Finally, to qualify for the sample, a going private firm must no longer trade anywhere in the U.S., nor provide publicly available financial information about itself.

To determine whether a firm continues to trade, we first investigate the delisting dates from CRSP for all sample firms. We then check Yahoo Finance and the Pink Sheets to verify that it no longer was trading in any market. Firms that are forced to go private because of being delisted by an exchange are excluded. If the stock continued to trade on the Pink Sheets, it was eliminated from the going private sample and classified as a "going dark" firm. To determine whether a firm provides financial information to the public, we search the internet to see if the firm had a website where such information was provided. (9) We found that the majority had websites, but these only described product lines. We exclude banks, other regulated financial institutions, and utilities, since these firms file publicly available financial statements with their regulators. Finally, we check EDGAR to see if any of the going private firms continue to provide financial information as a result of having publicly traded debt. These firms are excluded from the sample.

A detailed description of the sample selection process is provided in Appendix A.

The going private transactions must have occurred in the period from 1999 through 2005, must have consistent stock price data available in CRSP over a 200-day estimation period before the announcement, and have data available from Compustat. Furthermore, we obtain information on the size of the board of directors, ownership by directors, CEO and other executives. Information on financial advisors, anti-takeover devices, and motives for going private transactions is also obtained from SEC filings.

We create a control sample from SDC's Global New Issues database of firms that are in the same industry (based on 3-digit SIC code), have a similar size (as measured by assets) and book-to-market ratio (as per Barber and Lyon, 1997), and went public within a year of the sample firms that went private. Our matching process ensures that the age of the sample firm since it has gone public is similar to that of the matched control firm. We do this since many firms may have some "life cycle" propensity to go private. Furthermore, many firms that went private in our sample went public during the tech/dotcom bubble. The characteristics of firms that went public during that bubble differ from those that occurred earlier and later on. Finally, we verified that the control firms had not gone private as of October 2006.

Classification of Firms Going Private

The primary sample consists of 262 firms that went private over the period from 1999 through 2005. A breakdown of transactions by year is shown in Table 1, Panel A. About 46 percent of the going private transactions in the sample occur before SOX, versus 54 percent since SOX. Interestingly, in the first half of 2003, more firms went private than went public (Thomson Financial Corporation Global New Issue and Mergers and Acquisitions database).

The industry SIC classifications for the firms going private are shown in Table 1, Panel B. There are going private transactions in nearly every industry, although more of the transactions have occurred in the chemicals and plastics industry (SIC code 3000) and technology (SIC code 7000) sectors.

In Table 1, Panel C, we identify the type of bidder for all the going private transactions as well as those that occurred before and after SOX. About 43.5% of all going private transactions were initiated by management while 56.5% of the transactions were initiated by buyout specialists. The proportion of going private transactions initiated by management was 58.3% before SOX versus 33.8% since SOX. Buyout specialists became relatively more active post-SOX.

Change in Motives for Going Private

We obtain information about the motives for going private by looking at the proxy statements and press releases of firms that went private. Table 2 provides the motives which are cited most often by these firms or which have been mentioned in the business media as reasons to go private. In the table the sample is split into three periods: (1) pre-SOX, (2) from SOX until the end of 2003 (referred to as post-SOX1), and (3) 2004-2005 (referred to as post-SOX2). We test for significant differences between pre-SOX and either post-SOX1 or post-SOX2 using two tests--Cochran's and Mantel-Haenszel. Notice that the reasons most often cited for going private are lack of liquidity, the cost of being public, the premium that is being offered and lack of analyst coverage.

Interestingly, the percentage of firms going private which cite the cost of being public as a motive increased from the pre-SOX period to the post-SOX1 period, but then declined in the post-SOX2 period. Lack of liquidity is important pre-SOX and is dramatically larger post-SOX2 compared to post-SOX1. The premium being offered is significant in both post-SOX periods and is cited by a substantially larger number of firms in post-SOX2 compared to post-SOX1. Finally, institutional ownership is cited significantly more often in post-SOX2, while this is not the case in post-SOX1. Given the importance of these reasons, we incorporate these factors in our multivariate tests.

Among the other reasons that are often mentioned in the financial media and by business people are reduction in the expenditure of management time, the lack of benefit from being public to raise capital, the need to avoid disclosure and the benefit of being able to take a long term focus. These reasons accounted for less than a fifth of the sample of firms that went private before and after SOX.

Analysis

Conditional Logistic Regression Analysis

We run 3 conditional logistic regression models of the form:

Pj = exp([beta]'Xj) / [summation]j exp([beta]'Xj), for j = 1,0

where p is the probability of a firm going private and Xj is the vector of explanatory variables.

The results are presented in Table 3.

Since some of the independent variables are highly collinear, various models that contain different variables are used. The CEOWN variable is positive and significant, supporting the hypothesis that firms with higher levels of CEO ownership are more likely to go private. The CEOWN*SOX interaction term is positive and significant, supporting the hypothesis that the firms with higher levels of CEO ownership are even more likely to go private since SOX. The DEOWN variable and DEOWN*SOX variables are not significant.

The EHW study (2007) also applied a logit analysis (1) and found that a variable combining director and executive management variable was positive and significant. Our use of two variables to separately test the impact of CEO ownership and other executives and directors suggests that the CEO ownership may have more influence on the propensity to go private. Moreover, our results also show that since SOX, the likelihood of going private is positively and significantly related to CEO share ownership.

While the institutional ownership variable is not significant, the interaction term, Instown*SOX, is positive and significant. The Runup variable is negative and significant as is the liquidity variable. ROA is not significant, but ROA*SOX is negative and highly significant. This also occurs for the price to book ratio and its interaction term. These results indicate that firms with a higher level of institutional ownership are more likely to go private since SOX, as are firms with a weak stock performance before going private. Profitability and price-to-book do not affect the likelihood of going private. However, after SOX, firms with weak profitability and which are out of favor are more likely to go private. Size (assets) does not affect the likelihood of going private before and after SOX. However, when we interact SOX with assets, we find that size matters significantly. Small firms are much more likely to go private after SOX. Overall, there is very strong evidence that post-SOX additional characteristics affected the propensity to go private.

Measurement of Valuation Effects

To explore the valuation effects of going private transactions, we use standard event study methodology to estimate the market's reaction to the announcement to go private. We estimate the parameters of the market model over a 100-day estimation period chosen from t = -110 to t = -11, where t=0 is the event date. Scholes-Williams betas are used to account for infrequent trading and significance tests are based on the standardized cross-sectional method (Boehmer, Musumeci and Poulsen, 1991). Our significance tests adjust for cross-sectional dependence in the regression residuals. We adjust for serial dependence as well as for changes in event induced variance. CARS are based on the equally weighted market portfolio. We attempt to explain the source of value gains or losses with cross-sectional regressions on CARs. (2)

We examine mean (median) CARs for four different windows: (-10,+10), (-1,+1), (-1,0), and (0,0), and test for significance using t-tests and generalized sign tests. Results are shown in Table 4.

We also compare the percent of positive CARs to negative CARs before and after SOX. For the entire sample, mean and median CARs are significantly greater than zero (at the 1% level) for all event windows. Means range between 15.32% on day 0 to 24.71% for days -10 to +10. About 86% of the CARs is positive and highly significant (at the 1% level). We also split the sample of going private transactions based on whether they occurred before or after SOX was enacted. In both periods, the CARs are significantly positive (all at the 1% level). Interestingly, mean CARs of the going private transactions after SOX was enacted are not significantly different from before SOX. However, median CARS are significantly less after SOX in two of the four windows we look at. We then partition the post-SOX results into two periods. The mean CAR from going private in the post-SOX1 period (SOX through 2003) is significantly greater than that of the post-SOX2 period (2004-2005). (3)

Cross-Sectional Analysis of Valuation Effects

We ran 4 cross-sectional models with the (0,0) CARs as the dependent variable. (4) The results are shown in Table 5. (5) We investigate the characteristics of the firms and the transactions that lead the market to view the deals favorably or unfavorably, including whether the market anticipated the transaction for a given firm.

The SOX1 variable is positive and significant (at the 5% level) in all model specifications, consistent with a higher valuation effect after the passage of the Sarbanes-Oxley Act described above. However, SOX2 is insignificant. The cost of being public (COBP) variable is negative and significant (at the 10% level). The interaction term between COBP and SOX1 is positive and significant, which suggests that the cost motive leads to more favorable valuation effects shortly after SOX went into effect. Yet, the significance vanishes in the SOX2 period. Thus, the market responded more favorably to the cost motive for firms that went private relatively soon after the inception of SOX, but was not sensitive to this motive in the 2004-2005 period.

The Runup variable is negative and significant in all the models in which it appears, which reflects a more favorable effect for going private firms that recently had relatively poor stock price performance. The management variable is not significant. The CEOWN variable is generally negative and significant indicating that greater CEO ownership results in lower valuation from going private. The CEOSOX1 variable is negative all five models, but only modestly significant in one. This would suggest that CEO ownership did not affect value in the period just after SOX. However, the CEOSOX2 variable is positive, and modestly significant only in one of the five models. The DEOWN variable is not significant, but is modestly, positively significant in the two years post-SOX. These results offer modest evidence of greater alignment between CEOs and other officers/directors in the two years after SOX was enacted

The GP-Prob variable (6) is positive and modestly significant in four of the models. Institutional ownership and the percent of directors who are insiders does not affect valuation. With regard to governance, conflicts of interest generally reduce value, but the having an independent special committee or amended poison pills has no affect. The PBK variable is negative and modestly significant for both periods, implying that firms that were more out-of-favor experience better valuation effects. This relationship is somewhat more pronounced only in the post-SOX1 period. The profitability variable (ROA) is negative but is not significant. Size as measured by assets is negatively related to value, but is not significant. However, in the post-SOX period, size (as given by the interaction variable) is significantly negatively related to value. Thus, smaller firms are able to increase value by going private post-SOX, at least in the period that we examine.

Additional Tests

A useful extension of our prior tests would be to investigate whether SOX discouraged privately held firms from going public post-SOX, and whether private firms which were discouraged from going public share characteristics similar to our going private sample. (7) Using the Securities Data Corporation's Global New Issues database, we identify 36 non-financial firms that withdrew IPOs during the time period of the study for which data on total assets, net income, and cash and equivalents were available in IPO filings. In this group of 36 firms, 10 withdrew offers pre-SOX and 26 withdrew offers post-SOX. These firms were private prior to the initiation of their IPOs (i.e., they were not carved-out or spun-off, or firms that had been taken private or went bankrupt and subsequently re-emerged). We are able to obtain some basic descriptive statistics from SDC on these firms, because they are required to disclose some financial information in their prospectus and registration statements. It is difficult to generalize convincingly from our sample firms to all private firms that did not go public, since the former chose to initiate the IPO process in the first place.

In untabulated tests, we find that there is no difference in the characteristics of firms that withdraw offers before and after SOX, using nonparametric tests appropriate for small samples. Prior to SOX, there were no significant differences in terms of size (as proxied by total assets) between the firms that withdrew IPOs and going private firms. Additionally, there is no difference between firms that withdrew IPOs and firms that remained public, although firms that withdrew IPOs prior to SOX were less profitable (using ROA) than those that remained public or those that went private. Firms that withdrew IPOs prior to SOX were significantly less liquid (as proxied by the ratio of cash to assets) than those that remained public, but not those that went private.

However, using a sub-sample of our going private and staying public firms matched on industry, we find that post-SOX the firms that withdrew IPOs tend to be significantly smaller (based on total assets), less profitable (in terms of ROA), and less liquid (as proxied by the ratio of cash to assets) than firms that remained public during the sample period. Size and profitability are insignificantly different between the going private firms and firms that withdraw IPOs (i.e., which stay private) after SOX. Additionally, firms that withdraw IPOs tend to have less liquidity than those that go private post-SOX. While our sample is small, our results tentatively support our overall findings, namely, that SOX was an impetus for firms to either go private (or stay private), particularly for weak firms.

Summary

The Sarbanes-Oxley Act of 2002 represents the most significant piece of legislation dealing with the governance and accounting of pubic firms since the 1930s. Some critics of the new law have suggested that there are unintended consequences to public companies resulting from increased costs due to its implementation. We assess going private transactions before and after the Sarbanes-Oxley Act became law in an attempt to determine whether the causes or effects of going private have changed since SOX. We conduct a comprehensive analysis of firms going private transactions that took place before and after SOX, a period covering the years 1999 through 2005. We only examine firms whose shares are no longer traded and which no longer provide financial information to investors. We do not include going dark firms that deregister with the SEC. In order to better understand why firms go private, we consider factors cited by firms in going private. Lack of liquidity, the cost of being public, and the premium offered to go private are mentioned by at least 25% of the firms going private.

We find that firms are more likely to go private when they have higher CEO ownership, and this relationship is even more pronounced since the passage of SOX. The likelihood of going private is also inversely related to firm liquidity both before and after SOX. We find that firms which are smaller and exhibit weak profitability post-SOX are more likely to go private (when the cost of being public increased.) In addition, the propensity to go private is positively related to institutional ownership, but negatively related to the ratio of price to book value in the period after SOX.

The valuation effects of firms going private are mixed before and after SOX. Small firms that go private experience larger valuation effects post-SOX. Those that cited the cost of being public were negatively affected in both periods, yet in the period just after SOX, the effects are positive. Firms with weak stock market performance have better valuation effects pre and post-SOX. Valuation is generally negatively related to firms that were out of favor (price to book value) in both periods. However, the impact of the latter factor was greater in the period just after SOX. Overall, the characteristics of firms going private and the market's perceptions of going private have changed since the passage of the Sarbanes-Oxley Act.

Finally, we would like to include the caveat that our results may not be due entirely to Sarbanes-Oxley Act. Following SOX, many changes took place on the New York Stock Exchange and NASDAQ that strengthened listing and disclosure requirements. Furthermore, there may be additional oversight by all of the traditional monitors of the firm (shareholders, analysts, boards, and the media) that may have led firms to practice better governance and exhibit greater transparency.

Appendix A Sample selection

Search in SDC: All going private transactions in the U.S. of firms traded on the NYSE, AMEX, or NASDAQ between 1999 and 2005--698

Completed Transactions--435

Deals where data could be verified as having correct information from Lexis Nexis--402

Deals where it was verified that the firm did not voluntarily deregister or conduct a reverse stock split to go private (all of this information was verified in Lexis Nexis)--375

Deals where it was verified that the firm no longer provides financial information from Yahoo Finance and web search--346

Firms with data in CRSP for the lull estimation period--282

Firms whose going private documentation and most recent proxy statement is available in 10K-Wizard or Edgar--26

References

Atiase, R., "Predisclosure Information, Firm Capitalization, and Security Price Behavior Around Earnings Announcements," Journal of Accounting Research, 23 (Spring 1985), pp. 21-36.

Barber, B.M., & J.D. Lyon, "Detecting Long-run Abnormal Stock Returns: The Empirical Power and Specification of Test Statistics," Journal of Financial Economics, 43 (1997), pp. 341-372.

Barth, M.E., & R. Kasznik, R., 1999, "Share Repurchases and Intangible Assets," Journal of Accounting and Economics, 28 (December 1999), 211-241.

Bebchuk, L., & M. Kahan., "'Adverse Selection and Gains to Controllers in Corporate Freezeouts," in R. Morck (ed), Concentrated Corporate Ownership (University of Chicago Press, Chicago, IL, 2000).

Boehmer, E.J. Musumeci & A. Poulsen, "Event Study Methodology Under Conditions of Event-Induced Variance," Journal of Financial Economies, 30 (December 1991), pp. 253-272.

Brown, B., B. Davis & E. Wade, "Think Sarbanes-Oxley Doesn't Apply to Private Firms? Think Again," Houston Business Journal, December 6, 2002 C:3.

Engel, E., R. M. Hayes & X. Wang, "The Sarbanes-Oxley Act and Firms' Going Private Decisions," Journal of Accounting and Economics, 44 (September 2007), pp. 116-145.

Firth, M., "Dividend Changes, Abnormal Returns, and Intra-Industry Firm Valuations," Journal of Financial and Quantitative Analysis, 31 (June 1996), pp. 189-210.

"Fleeing From Public Ownership: Going Private Helps Many Companies Escape Sarbanes-Oxley and Stock Price Pressures," Dealmaker's Journal, November 1, 2003.

Goh, J., M. F. Gombola, F.Y. Liu & D.W. Chou, "Going Private Restructuring and Earnings Expectations: A Test of the Release of Favorable Information for Target Firms and Industry Rivals," Working paper, Drexel University (2009).

Halpern, P., R. L. Kieschnick & W. Rotenberg, "On the Heterogeneity of Leveraged Going Private Transactions," Review of Financial Studies, 12 (Summer 1999), pp. 281-309.

Hand, J.R.M., "Resolving LIFO Uncertainty: A Theoretical and Empirical Reexamination of 1974-75 LIFO Adoptions and Nonadoptions," Journal of Accounting Research 31 (Spring 1993), 21-49.

Kamar, E., P. Karaca-Mandic & E. Talley," Going Private Decisions and the Sarbanes-Oxley Act of 2002: A Cross-Country Analysis," Journal of Law, Economics and Organization, 25 (May 2009), pp. 107-133.

Lehn, K., & A. Poulsen, "Free Cash Flow and Stockholder Gains in Going Private Transactions," Journal of Finance, 44 (July 1989), pp. 771-787.

Leuz, C., (2007), "Was the Sarbanes-Oxley Actually This Costly? A Discussion from Event Returns and Going Private Decisions," Journal of Accounting and Economics, 44 (September 2007), pp. 148-165.

Leuz, C., A. Triantis & Y. Wang (2008), "Why Do Firms Go Dark? Causes and Economic Consequences of Voluntary SEC Deregistrations," Journal of Accounting and Economics, 45 (August 2008), pp. 181-208.

Li, H., M. Pincus & S. Rego, "Market Reaction to Events Surrounding the Sarbanes-Oxley Act of 2002," Working paper, University of Iowa, 2004.

Marosi, A., & N. Massoud, "Why Do Firms Go Dark?' Journal of Financial and Quantitative Analysis, 42 (June 2007), pp. 421-442.

Slovin, M.B., M.E. Sushka & Y. M. Bendeck, "The Intra-Industry Effects of Going Private Transactions," Journal of Finance, 46 (September 1991), pp. 1537-1549.

White, H., "A Heteroscedasticity-Consistent Covariance Matrix Estimator and A Direct Test for Heteroscedasticity," Econometrica, 48 (May 1980), pp. 817-838.

Ivy Xiying Zhang, "Economic Consequences of the Sarbanes-Oxley Act of 2002," Journal of Accounting and Economics, 44 (No. 1, 2007) pp. 74-115.

Leonard Rosenthal *

Bentley University

Kimberly C. Gleason

University of Pittsburgh

Jeff Madura

Florida Atlantic University

(1) For smaller firms and firms that are domiciled outside the U.S., Section 404 regulations have been modified several times. Pursuant to the Dodd-Frank Act (2010), in September 2010, the SEC suspended Section 404(b) for firms with a market cap of shares held by 'non-affiliated' holders of less than $75 million. As a result, these firms do not have to file an auditor's attestation on internal controls as part of their 10-K.

(2) Firms may not be able to completely evade SOX by going private. Some private firms may abide by specific SOX requirements, such as the establishment of internal controls, as a signal to vendors, customers, acquisition targets, potential bidders and creditors of increased transparency that may improve their business prospects. Firms that are currently private, but considering going public in the future, will be expected to have maintained SOX compliance for a period of years prior to going public. In addition, document retention and whistleblower aspects of SOX apply to both public and private firms (Brown et al, 2002).]

(3) See Leuz (2007), p 159.

(4) SEC Rule 12g55-1, adopted in 1965 determines the number of record holders below which firms can deregister. The number of record holders is almost always less than the number of beneficial holders since many investors have their shares in 'Street name'. Firms can engage in various transactions to bring the count to fewer than 300.

(5) See "FEI Survey on Sarbanes-Oxley Section 404 Implementation," March 2006, May 2007.

(6) Financial Executives Research Foundation annual "Audit Survey Fee" for 2010, 2009 and 2008.

(7) "The Cost of Being Public in the Era of Sarbanes-Oxley," Foley & Lardner, LLP. May 19, 2004; "The Cost of Being Public in the Era of Sarbanes-Oxley, "Foley & Lardner, LLP, June 16, 2005, June 15, 2006, August 2, 2007.

(8) See Hand (1993), Barth and Kasznik (1999).

(9) Some of the firms we excluded may be required to furnish financial statements to their shareholders either mandatorily or upon written request by shareholders. Since this information is limited to a small group of people, firms in this category do not fit a reasonable definition of providing publicly available information.

(1) In untabulated results, we run logistic regressions on a sample of going-dark firms over the same period as our going private transactions. We find that going-dark firms exhibit several different characteristics compared to the going private firms, consistent with LTW (2008).

(2) Our method of deriving valuation effects does not capture the initial impact that resulted from the regulations that arose from the SOX Act. The market could have repriced shares of firms at the time the regulations were announced, whether the firms ultimately went private or not. Our focus is isolated to confirmation that the firm intends to go private.

(3) We also conducted extensive sensitivity tests based on the dates used in the EHW (2007), Zhang (2007) and Li et al. (2004) papers, varying the "post SOX" period to 7/30/2002, and secondly, by eliminating all transactions that took place between 2/12/2002 and 7/30/2002. The results from applying these tests are quantitatively similar to those reported in the paper.

(4) Cross-sectional regressions using (-1,+1) CARs yielded qualitatively similar results.

(5) We use White's (1980) correction for heteroscedasticity in all regression models.

(6) The EHW (2007) study found that a variable that combines the CEO and other executives and director ownership was positively and significantly related to valuation effects of going private since SOX. This result differs from ours and is likely due to EHW using a different sample.

(7) We thank an anonymous reviewer and the editor for this suggestion.

We are grateful to Lucien Bebchuk, Victor Kalafa, Christian Leuz, and especially Gordon Karels and Husayn Shahrur for their helpful comments. Any remaining errors are the responsibility of the authors.

* Corresponding author. Tel.: 1-781-891-2516; Fax: 1-781-891-2896

E-mail address: lrosethal@bentley.edu
Table 1--Sample Descriptive Statistics

Panel A gives a breakdown of firms going private by year. Panel B
shows the distribution of firms in numbers and percent by SIC code.
Panel C shows the bidder structure in percent.

Panel A: Sample by Year

Year                                 Number         Percent

1999                                 41             15.65
2000                                 44             16.79
2001                                 32             12.21
2002                                 47             17.94
2003                                 41             15.65
2004                                 27             10.31
2005                                 30             11.45
Total                                262            100.00
Pre-SOX (1/1/1991-2/12-2002)         121            46.18
Post-SOX 1 (2/13/2002-12/31/2005)    84             --
Post-SOX2 (1/1/2004-12/31/2005)      57             --
Post SOX All                         141            58.32

Panel B: Sample by SIC

SIC                                  Firms Going    Percent
                                     Private        of Total

1000                                 27             10.31
2000                                 28             10.69
3000                                 56             21.37
4000                                 28             10.69
5000                                 40             15.27
6000                                 2              0.76
7000                                 56             21.37
8000                                 25             9.54
Total                                262            100.00

Panel C: Sample by Bidder Structure (in percent)

                       All        Pre-SOX    Post-SOX

Management Involved    43.45      58.33      33.75
Buyout Specialist      56.45      41.67      66.25

Table 2--Motives for Going Private

Reasons                        n      % of     Pre-    %Pre-
                                      total    SOX     SOX1

Liquidity                      113    43.13    43      35.54
Cost of being public           109    41.60    41      33.88
Premium                        87     33.21    16      13.22
Lack of analyst coverage       75     28.63    26      21.49
Institutional ownership        63     24.05    20      16.53
Industry conditions            63     24.05    10      8.26
Takes long-term focus          48     18.32    21      17.36
No benefit from raising        33     12.60    8       6.61
  capital in public markets
Avoid disclosure               33     12.60    9       7.44
Reduction of managerial        30     11.45    10      8.26
  time expenditure
Industry out of favor          24     9.16     21      17.36
  with market

Reasons                        Post-    %Post-   Post-    %Post-
                               SOX1     SOX2     XOX2     SOX2

Liquidity                      28       33.33    42       73.68
Cost of being public           43       51.19    25       43.86
Premium                        27       32.14    44       77.19
Lack of analyst coverage       21       25.00    28       49.12
Institutional ownership        14       16.67    29       50.88
Industry conditions            15       17.86    38       66.67
Takes long-term focus          16       19.05    11       19.30
No benefit from raising        13       15.48    12       21.05
  capital in public markets
Avoid disclosure               5        5.95     19       33.33
Reduction of managerial        14       16.67    6        10.53
  time expenditure
Industry out of favor          1        1.19     2        3.51
  with market

Reasons                        Post 1 vs. Pre    Post 2 vs. Pre
                               Odds Tests        Odds Tests

Liquidity                      0.93              15.32 ***
                               (0.64)            (15.32) ***
Cost of being public           10.66 ***         5.22 ***
                               (9.60) ***        (5.63) ***
Premium                        19.91 ***         14.52 ***
                               (18.24) ***       (13.44) ***
Lack of analyst coverage       2.00              7.84 ***
                               (1.35)            (6.88) ***
Institutional ownership        0.28              24.31 ***
                               (0.11)            (20.39) ***
Industry conditions            8.60 ***          25.65 ***
                               (7.29) ***        (34.85) ***
Takes long-term focus          1.48              1.46
                               (1.05)            (121)
No benefit from raising        7.67 ***          10.33 ***
  capital in public markets    (6.37) ***        (9.21) ***
Avoid disclosure               0.05              14.25 ***
                               (0.05)            (9.22) ***
Reduction of managerial        6.74 ***          2.36
  time expenditure             (5.58) ***        (3.45)
Industry out of favor          9.82 ***          7.84 ***
  with market                  (8.34) ***        (6.88) ***

Odds are done using Cochocran's and Mantel-Haenszel tests, are used,
respectively, for significant differences pre and post-SOX.

*** Significant at the .01 level

Table 3--Conditional Logistic Analysis

Dependent variable is Going Private Firm--1, Control Firm = 0. CEOWN
is percent of shares owned by the chief executive officer at the time
of going private transaction. CEO*SOX is equal to CEO ownership *SOX.
DEOWN is the percent of shares owned by non-CEO officers and
directors. DEOWN*SOX is an interaction variable equal to director and
executive ownership * SOX. Instown is the percent of shares held by
institutional investors at the time of going private transaction.
Instown*SOX is an interaction variable is equal to institutional
ownership*SOX. Runup is the daily average percent change in the stock
price from 200 days prior to the announcement of the going private
transaction to 20 days before the going private transaction. Logat is
log of book value of assets in millions of dollars, in the year prior
to the going private transaction. Mktcash is the ratio of equity
market capitalization to cash. FCFAT Free cash flow is defined similar
to EHW (2007) and Lehn and Poulsen (1989), as follows: Free cash flow
= [operating income--Taxes paid--Interest expense--preferred
dividends--common dividends]--average total assets. Liquidity is
average daily trading volume over the three months prior to the
transaction divided by the average number of shares outstanding. ROA
is defined as net income to total assets. ROA*SOX is an interaction
variable equal to the product of ROA and SOX. PBK is the price-to-
book ratio. PKBK* SOX is an interaction variable equal to the product
of price-to-book*SOX. AT* SOX is an interaction term equal to the
product of total assets and SOX. Wald statistics are shown beneath
parameter estimates.

               Model 1       Model 2       Model 3       Model 4
               (Wald-        (Wald-        (Wald-        (Wald-
               statistic)    statistic)    statistic)    statistic)

Intercept      -0.234 **     -0.147 ***    -0.237 ***    -0.347 ***
               (4.98)        (5.28)        (5.33)        (6.28)
CEOWN           0.102 ***                                 0.004 **
               (5.22)                                    (3.40)
CEOWN*SOX       0.243 **                                  0.113 ***
               (3.55)                                    (3.78)
DEOWN          -0.226
               (1.45)
DEOWN*SOX      -0.732
               (1.11)
Instown         0.025
               (1.07)
Instown*SOX     0.174 ***
               (3.66)
Runup          -0.014 **     -0.008 ***    -0.029 ***    -0.005 *
               (2.99)        (3.85)        (5.88)        (2.45)
Mktcash        -0.001        -0.001        -0.001        0.088
               (0.72)        (0.47)        (0.52)        (1.54)
FCFAT           0.132         0.175         0.112         0.074
               (1.77)        (1.26)        (1.29)        (1.82)
Liquidity      -0.012 ***    -0.058 ***    -0.035 ***    -0.009 ***
               (3.88)        (4.98)        (5.26)        (4.85)
ROA            -0.235        -0.117        -0.108*       -0.107
               (2.09)        (1.80)        (2.43)        (1.22)
ROA*SOX        -0.112 ***    -0.108 ***    -0.145 ***    -0.003
               (4.23)        (5.71)        (5.87)        (1.08)
PBK            -0.221        -0.074        -0.832        -0.006
               (1.60)        (1.47)        (1.64)        (1.53)
PBK*SOX        -0.236 ***    -0.285 ***    -0.183 ***    -0.210 *
               (5.21)        (6.37)        (4.73)        (2.54)
Logat           0.114                                     0.088
               (1.35)                                    (1.54)
AT*SOX                                                   -0.066 ***
                                                         (3.22)
N              388           388           388           388
Chi-sq         31.01 ***     27.84 ***     22.13 ***     31.01 ***

*** Significant at the .01 level

** Significant at the .05 level

* Significant at the .10 level

Table 4--Cumulative Abnormal Returns of Firms that Go Private

                         (-10,10)       (-1,1)         (-1,0)
                         Mean           Mean           Mean
                         Median         Median         Median

All                      24.71          15.32          16.68
N = 262                  20.00          19.33          10.89
                         (12.74) ***    (13.34) ***    (10.86) ***
Pre-SOX                  25.92          23.77          17.15
N = 121                  23.13          18.79          13.24
                         (10.39) ***    (11.26) ***    (8.84) ***
Post-SOX All             23.40          22.47          16.18
N = 141                  15.56          16.86          6.95
                         (8.87) ***     (8.40) ***     (6.79) ***
SOXI             30.03   27.41          20.67          19.97
N = 84                   26.55          19.49          11.09
                         (6.81) ***     (6.76) ***     (5.82) ***
SOX2             13.63   15.19          9.56           9.41
N = 57                   10.08          12.08          2.41
                         (7.26) ***     (5.88) ***     (4.99) ***
T (Post SOX -            -1.25          -1.38          -0.84
  Pre-SOX)
Z (Post SOX -            (4.06) **      (1.86)         (5.82) **
  Pre-SOX)
T(SOX]-                  2.87 ***       3.08 ***       2.79 ***
  SOX 2)
Z(SOX I-                 (3.72) **      (5.09) ***     (6.22) ***
  SOX2

                 (0.0)          (-1,1)+/-
                 Mean           Pct pos/neg
                 Median         Gen Sig Z

All              15.32          226/36
N = 262          9.33           96.26/13.74
                 (10.24) ***    (12.57) ***
Pre-SOX          14.96          101/20
N = 121          11.08          83.47/16.53
                 (8.11) ***     (6.92) ***
Post-SOX All     15.70          125/16
N = 141          13.72          88.65/11.35
                 (6.64) ***     (15.76) ***
SOXI             73/11
N = 84           11.44          86.91/13.09
                 (5.62) ***     (5.88) ***
SOX2             52/5
N = 57           9.88           91.22/8.77
                 (3.89) ***     (7.98) ***
T (Post SOX -    1.07
  Pre-SOX)
Z (Post SOX -    (2.68)
  Pre-SOX)
T(SOX]-          3.02 ***
  SOX 2)
Z(SOX I-         (2.85) *
  SOX2

Mean and median CARS in response to going private transactions. CARS
are for various windows around the announcement date. T-and
standardized cross-sectional model (SCS) generated Z-statistics are
used to test for differences in means (T) and medians (Z).

*** Significant at the .01 level

** Significant at the .05 level

* Significant at the .10 level

Table 5--Cross-Sectional Analysis (0,0) CARS

Model            1            2            3

Intercept         0.086        0.140 **     0.148 **
                 (1.58)       (2.25)       (2.37
SOX1              0.198        0.285 **     0.104 **
                 (2.05)       (2.36)       (2.18)
SOX2             -0.010       -0.042       -0.085
                (-1.07)       (0.58)      (-1.05)
COBP             -0.012 *     -0.009 *     -0.037 *
                (-1.65)      (-1.74)      (-1.90)
COBP*SOX1         0.025 **
                 (2.07)
COBP*SOX2        -0.011
                (-1.22)
Runup            -0.084 **    -0.104 **    -0.048 **
                (-2.58)      (-2.47)      (-2.15)
Mgmt             -0.112       -0.074        0.022
                (-1.41)      (-1.25)       (0.52)
CEOWN            -0.001       -0.003 **    -0.001 *
                (-1.25)      (-2.07)      (-1.75)
CEOWN*SOX1       -0.001       -0.001       -0.002 *
                (-0.52)      (-1.47)      (-1.73)
CEOWN*SOX2       -0.002        0.002 *      0.003 **
                (-1.30)       (1.85)       (1.99)
DEOWN             0.008
                 (1.47)
DEOWN*SOX1        0.017
                 (1.59)
DEOWN*SOX2        0.043 *
                 (1.74)
GP-Prob           0.017 *      0.024 *
                 (1.68)       (1.74)
Instown           0.002
                 (1.24)
Pctinsid         -0.202       -0.105 *     -0.185
                (-1.58)      (-1.68)      (-1.20)
Conflict         -0.028 *     -0.182 **    -0.085 **
                (-1.88)      (-2.05)      (-2.02)
Speccom           0.022        0.028        0.017
                 (1.27)       (0.85)       (0.86)
Rights           -0.016       -0.004       -0.009
                (-1.08)      (-1.45)      (-1.55)
PBK              -0.028 *     -0.015 *     -0.047 *
                (-1.69)      (-1.78)      (-1.74)
PBK*SOX1
PBK*SOX2
Profitability    -0.017       -0.007       -0.011
                (-0.66)      (-0.85)      (-1.22)
Logat            -0.002       -0.004       -0.001
                (-1.04)      (-1.58)      (-1.37)
AT*SOX
N                244          251          248
ADJ. R-SQ        0.0521       0.0682       0.0640
F                3.07 **      4.85 ***     4.30 **

Model            4            5

Intercept         0.026 **     0.110 **
                 (2.41)       (2.39)
SOX1              0.112 **     0.095 **
                 (2.36)       (2.04)
SOX2             -0.015       -0.030
                (-1.22)      (-1.05)
COBP                          -0.003
                             (-1.16)
COBP*SOX1                      0.008 *
                              (1.69)
COBP*SOX2
Runup            -0.027 **    -0.038 **
                (-1.99)      (-2.04)
Mgmt              0.001        0.010
                 (1.28)       (0.80)
CEOWN            -0.001       -0.001
                (-1.38)      (-1.43)
CEOWN*SOX1       -0.001       -0.001
                (-1.55)      (-0.82)
CEOWN*SOX2        0.002        0.001
                 (1.64)       (1.02)
DEOWN
DEOWN*SOX1
DEOWN*SOX2
GP-Prob           0.022 *      0.007
                 (1.78)       (0.86)
Instown
Pctinsid         -0.172       -0.012
                (-1.43)      (-0.22)
Conflict         -0.014 *     -0.022
                (-1.74)      (-0.44)
Speccom           0.008        0.015
                 (0.47)       (1.37)
Rights           -0.004       -0.013
                (-1.28)      (-0.75)
PBK              -0.016 *     -0.027 *
                (-1.69)      (-1.84)
PBK*SOX1         -0.021
                (-1.85)
PBK*SOX2         -0.003
                (-1.30)
Profitability    -0.001       -0.002
                (-1.38)      (-0.46)
Logat            -0.001       -0.001
                (-1.40)      (-1.22)
AT*SOX                        -0.350 **
                             (-2.02)
N                250          250
ADJ. R-SQ        0.0544       0.0448
F                4.10 ***     2.75 ***

SOX = 1 if the announcement took place after the introduction of
Sarbanes-Oxley by Representative Oakley (2-12-2002),and before the end
of 2003, and 0 otherwise. SOX2=1 if the announcement took place during
2004-2005, and 0 otherwise. COBP=1 if the cost of being public is
mentioned as a motive for going private, and 0 otherwise. COBP*SOX1
(COBP*SOX2) is an interaction variable = 1 if the reason for going
private is the cost of being public and if the transaction occurred
during SOX period 1 (2). Runup is the average daily percentage change
in the stock price from 200 days prior to the announcement of the
going private transaction to 20 days before going private. Mgmt=1 if
the buyout was organized by management, and zero otherwise. CEO is the
percent of shares owned by the chief executive officer. CEO*SOX1
(CEOSOX2) is an interaction term for the two periods following SOX.
DEOWN is the percentage of shares owned by non-CEO management and
directors. DEOWN*SOX1 (DEOWN*SOX2) is an interaction that captures the
additional sensitivity of the CAR to the two periods following SOX.
GP-Prob is obtained from conditional logit model (1) in Table 3 and
captures the anticipation by the market of the firm going private.
Instown is the percent of shares held by institutional investors.
Conflict=1 if the reports that there were conflicts in the year before
going private, and zero otherwise. Speccom=1 if going private special
committee was independent, and zero otherwise. Rights=1 if the firm
implemented or amended a poison pill in the two years preceding the
going private transaction. PBK is the price to book ratio. PBK*SOX1
(PBK*SOX2) is an interaction variable for transactions following SOX.
Profitability is industry adjusted return on assets. Logat is the log
of book value of assets in millions of dollars in the prior to going
private. AT* SOX is an interaction term equal to the product of total
assets and SOX.

*** Significant at the .01 level

** Significant at the .05 level

* Significant at the .10 level
COPYRIGHT 2011 Creighton University
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Rosenthal, Leonard; Gleason, Kimberly C.; Madura, Jeff
Publication:Quarterly Journal of Finance and Accounting
Date:Mar 22, 2011
Words:11926
Previous Article:When do firms issue exchangeable debt?
Next Article:The impact of independent and overlapping board structures on CEO compensation, pay-performance sensitivity and accruals management.
Topics:

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |