Size does matter: an examination of the economic impact of Sarbanes-Oxley.

Abstract

We employ an event study analysis to examine the market reaction to congressional agreement on the passage of the Sarbanes-Oxley Act of 2002. We find that as firm size increases, the negative impact of Sarbanes-Oxley's passage decreases. The results show that the difference in abnormal returns between the smallest and largest firms is 3.91%.

The Corporate Context of Sarbanes-Oxley

In a volatile world, burdened by corporate scandals and a decline in investors' confidence, lawmakers crafted the Sarbanes-Oxley Act of 2002 (hereafter abbreviated as SOX) with an enthusiastic desire for corrective action. Passed by Congress as a reaction to financial scandals such as Enron, WorldCom, Adelphia, Global Crossing, and Tyco, SOX enhanced standards for corporate accountability and penalties for corporate wrongdoing. SOX contains 11 extensive titles, ranging from extra responsibilities for audit committees to tougher criminal penalties for white-collar crimes such as securities fraud. In plain English, the objective of the law was to make financial reporting more transparent and executives more accountable, changes that were ultimately planned to restore investors' confidence in financial markets and enhance corporate governance.

SOX requires executives, boards of directors, and independent auditors to take specific actions that are intended to produce more reliable, timely, and useful financial information to the public. Greater transparency and more reliability in corporate reporting means that companies function in a moral and ethical environment that enhances their credibility. This is a key benefit that should lower the firm's cost of capital, which may improve growth. Thus, at least theoretically, SOX could provide a win-win situation for both companies and investors.

Although many agreed that the changes to the corporate governance system in the U.S. were necessary, some now believe that SOX has imposed an unnecessary burden on companies because of its high compliance costs. A challenging issue is Section 404 of SOX, Management Assessment of Internal Controls, which requires publicly held firms to identify financial reporting risks, establish related controls, assess their effectiveness, fix any material control deficiencies, and then re-test and re-document all of the above. A March 2005 survey by Financial Executives International shows that the first year compliance costs on section 404 of SOX alone averaged $4.36 million per company, and large companies with more than $5 billion in revenues spent more than $10 million per company.

Critics argue that for many firms, the costs of complying with SOX outweigh the benefits. For example, these costs are unreasonably high for small firms. If the compliance costs are at least in part fixed, small firms may bear a disproportionate burden. The American Electronics Association (AeA) claims that Section 404 compliance costs serve as a "regressive tax on small business." The purpose of this study is to examine the market impact of the enactment of SOX, in an effort to determine the market's reaction to the passage of this regulation with respect to firm size.

While the true costs of SOX compliance are not easy to measure, the benefits are even more difficult to estimate. The promised benefits of SOX, which include more transparent disclosure, improved corporate governance, and enhanced investor confidence, are difficult to measure in the short run. Thus, to investigate the impact of SOX on firms of different sizes, we use an event study analysis. Using an event study methodology allows us to gauge, in a single framework, the market's perspective of the benefits versus the costs of SOX. In particular, we find that small firms experienced a much larger negative abnormal return on the day that Congress agreed on the passage of SOX than did large firms. We find an almost monotonic relationship between firm size and adverse impact of SOX. In the next section, we discuss the major provisions of SOX, and we detail some of the provisions that have resulted in the highest compliance costs. In addition, we discuss how these costs are disproportionately burdensome to small firms.

SOX Provisions and Compliance Costs

Exhibit 1 includes the major titles of the Sarbanes-Oxley Act (SOX) of 2002. SOX addresses financial reporting, CEO and CFO certifications, internal controls, risk management, the composition of audit committees, corporate codes of ethics, whistleblower protection and attorney conduct.

Some of the most costly provisions include:

* Section 302, which requires the CEO and CFO to certify each annual or quarterly report filed;

* Section 906, which provides criminal penalties for the CEOs or CFOs who certify the reports knowing that they do not comply with requirements of the Security Exchange Act of 1934;

* Section 401, which requires that annual and quarterly reports disclose all material off-balance-sheet transactions and arrangements, and that pro forma financial information be presented appropriately.

However, according to many studies, the most costly provision of SOX is Section 404, which requires management oversight of the internal control structure of the firm, and requires an annual assessment of this control structure. The costs related to the fact that the CEO and CFO must certify the accuracy of financial statements are expected to ripple through the entire corporation, leading to increased monitoring by the audit committees.

Other costs are those related to increased audit fees and insurance premiums, larger expenses on internal control software, and higher consulting-related fees. These costs are extremely difficult to measure directly because firms may not have a direct method to measure the monitoring costs.

By many estimates, however, Sarbanes-Oxley's compliance costs have been much higher than expected, especially for smaller firms. Compliance costs for Section 404 alone have mounted to more than $35 billion, 20 times higher than regulators originally estimated. In Exhibit 2, three SOX cost-compliance studies are summarized. An AeA report entitled "The Sarbanes-Oxley Section 404: The 'Section' of Unintended Consequences and its Impact on Small Business" shows that Section 404 compliance costs for firms with $100 million or less in revenues amount to 2.55% of total revenues, while larger firm costs are lower (0.27% for firms $500-999 million; 0.16% for firms $1-4.9 billion; and 0.06% over $5 billion). In addition, Carney [5] and a Financial Executives International survey [8] both present evidence of mounting SOX compliance costs.

In addition to the explicit costs presented in Exhibit 1, several academic studies have examined the implicit cost of SOX compliance. For example, Engel et al. [7] find that the increased financial reporting and internal control requirements introduced by SOX are principal motives for companies in choosing to "go dark" or "go private." In a study using a comprehensive sample of 400 companies, Leuz et al. [9] note that on average, there is a large negative market reaction of approximately -10% to firms' decisions to go dark, the abnormal return at the announcement date being particularly negative for smaller firms. Investors anticipate that by deregistering, companies have lower-than-expected growth opportunities, lower profitability, and potential internal control problems. These results are in line with our conclusion that smaller companies are damaged more than large companies by the passage of SOX.

Block [3] notes that for some firms, the cost of remaining public more than doubled after the passage of SOX. This increase in costs is due to higher audit costs, increased insurance premiums, and outside director fees. Ribstein [14] finds that due to increased costs associated with passage of SOX, particularly in personal liability, executives may become more risk-adverse in the post-SOX environment. As a result, companies will tackle less-risky projects, which may reduce the value of their firms. Thus, even though in the short run the cost-saving rationale makes sense, in the long run, this decision to avoid risky projects will not only suppress creativity, but also have a negative impact on overall economic growth. Zhu and Small [17] find that American Depository Receipts (ADR) listings from foreign issuers decreased, and ADR delistings increase, after the passage of SOX. The increased cost of regulation has, in effect, discouraged foreign firms from being listed on the U.S. security exchanges.

Zhang [16] investigates the private cost and benefits of SOX by examining changes of the market index around the most significant legislative events. He finds that the cumulative abnormal returns around the events leading to the passage of SOX are significantly negative. Although the total market value loss of NYSE, AMEX and NASDAQ around the most significant three events amounts to $1.4 trillion, this loss is not directly attributed only to SOX. Li et al. [11] consider market reactions to the events by estimating the deviation of the market returns from the average raw returns of nonevent days in 2002. They find evidence of a positive association between stock returns and the extent of earnings management, suggesting that investors viewed SOX as beneficial. Chhaochharia and Grinstein [6] and Rezaee and Jain [13], using the same event-study methodology, also find that SOX is value-increasing. In the next section we detail our event-study analysis, and we reconcile possible reasons for the contradictory results of previous event study analyses of the passage of SOX.

Data and Empirical Method

Because our primary objective is to examine the impact of SOX on different-sized firms, we create size-based quintiles and deciles by partitioning our sample in two ways. First, we create a market-capitalization-based sample. We take all firms listed in the Center for Research in Security Prices (CRSP) database and remove all ADRs, all warrants, and all non-ordinary shares. We calculate the market capitalization of each firm on the last trading day in July 2002. Using this market capitalization value, we partition the sample into five quintiles based on the market capitalization for each security. After applying the sample filters, there are 5,395 firms in the final market capitalization-based sample. Second, we create an asset-based sample by taking all firms listed in the annual COMPUSTAT database, again deleting all non-ordinary shares, ADRs and non-U.S. listed securities, and partition them based on their total assets. After filtering the data, 4,303 firms remain in the asset-based sample, which we partition into five-asset-based quintiles. In a subsequent section, to examine further the firm size relationship, we break the asset-based sample into ten deciles.

We calculate abnormal returns for the entire sample using the market-adjusted approach (Brown-Warner 1980). Abnormal returns are calculated as: A[R.sub.it] = [R.sub.it] - [^.[alpha].sub.i] - [^.[beta].sub.i][R.sub.mt] where [R.sub.it] is the return to firm i on day t, [^.[alpha].sub.i] is an intercept term estimated using the 225 trading days returns 90 days prior to July 24, 2002, [^.[beta].sub.i] is the market beta for the ith firm estimated using the using the 225 trading days returns prior to July 24, 2002, [R.sub.mt] and is the market return on day t. We use the CRSP value-weighted market index as a proxy for the market return. We focus our event study analysis around July 24, 2002, because this is the day that the House and Senate reached an agreement on the final regulation. We use this day as the event day because much of the reaction should occur on the day that the rule was agreed upon, not the day that it was signed into law by the President.

Cumulative abnormal returns (CARs) are the sum of the one-day abnormal return estimates, and we calculate CARs for a window of one day before and the day of the announcement (-1, 0) and for one day before, the day of, and one day after the announcement (-1, 0, 1). Using these specifications allows our event study to capture anticipation of agreement (day -1) and any post-announcement stock price movement (day 1). We obtain the stock price and returns date for the event study analysis from the CRSP database. After estimating the abnormal returns for each firm, we average them within their quintile/decile rank, which allows us to assess the economic impact the passage of SOX had on firms of different sizes. We discuss the results of this analysis in the next section.

Firm Size and the Impact of the Passage of SOX

Exhibit 3 contains the abnormal returns for the market capitalization and asset-based quintile samples. Firms in the smallest market capitalization quintile have market capitalizations less than $22.14 million, and firms in the largest quintile have market capitalizations over $2.45 billion. As a reference, firms in quintile 3 have market capitalizations between $70.9 million and $227 million. As can be seen in Exhibit 3, firms in the two smallest market capitalization quintiles experienced the largest negative abnormal returns (i.e., the largest adverse economic impact), while firms in the largest-volume quintile experienced slightly positive abnormal returns. This indicates that small firms lost 3% of their value, while large firms actually experienced a very small increase in value, while controlling for general market movements on the day that the House and Senate agreed on the final SOX legislation. These negative abnormal returns for small firms are both economically and statistically significant, and provide evidence of an uneven burden borne by small firms.

Because market-capitalization quintiles may not be robust enough to measure a size factor, we repeat the analysis using an asset-based sample. Again we break the sample into five quintiles based on total assets. Firms in the smallest asset-based quintile had assets worth less than $51.9 million, and firms in the largest quintile had assets worth more than $1.88 billion. As a reference, firms in quintile 3 had assets between $183 million and $525 million. As can be seen in Exhibit 3, the asset-based quintiles exhibited a similar monotonic increase in adverse impact as firm size decreases. Again, however, the second to the smallest quintile has the largest negative abnormal returns.

To further examine the relationship between firm size and SOX, we break the sample into ten deciles to see if negative abnormal returns are decreasing in size on a finer scale. In addition, we estimate multiple-day abnormal returns for the asset-based deciles to assure the robustness of the analysis. Exhibit 4 contains the estimates from these specifications and we graph the results from the one-day abnormal return model in Diagram 1.

Again, the almost monotonic increase in adverse impact to small firms is evident. However, we note that the third decile has the largest negative abnormal returns in all specifications, but the increasing adverse impact is undeniable. This pattern holds across specifications of different size-based proxies, and when we consider the pre-announcement expectation, announcement day, and post-announcement drift (-1,0,1); and the pre-announcement expectation, announcement day (-1,0) specifications. In each of these specifications, smaller firms faired poorly versus larger firms.

The true relationship between firm size and economic impact of SOX is not apparent if one only looks at the average one-day negative abnormal return of -1.79% (-1.93%) for the aggregated asset-based (market capitalization-based) sample. The difference in abnormal returns for firms in the largest asset-based decile and the smallest decile is around -4% when considering the one-day abnormal specification. Only after examining the abnormal returns in decile and quintile forms does the true disproportional economic impact on firms of different sizes become apparent.

Future Research

While our article details the differential impact associated with the passage of SOX, our results should be taken in context of the limited amount of time that has passed since the regulation has taken effect. Further research is needed to gauge the long term impact of SOX on publicly listed firms. It is possible that the costs associated with compliance will decrease as the auditing industry adjusts to the new methods and regulatory requirements. But further research is needed to examine how firms have coped with the increased costs. For example, have the costs been absorbed by the firms or passed on to consumers? In addition, have the costs decreased since the enactment of SOX? It is likely that the regulatory burden associated with compliance will decrease as firms adjust to the process. A larger question is: What has happened to the firms that have chosen to de-list from their primary exchanges (i.e., go dark)? Do they remain private or do they, at a later date, decide to re-list; or do they choose other financing alternatives such as private placements?

Prior to the enactment of SOX, many firms had internal control systems that were similar to those imposed by SOX. Of particular interest to corporate governance researchers is an analysis of the compliance cost differential between firms that had adequate compliance procedures in place before the passage of SOX, and those that did not. This would provide additional insight into the benefits of establishing and maintaining adequate governance oversight and control mechanisms. In addition, the reaction of the announcement of the passage of SOX would differ between the two sets of firms. Firms with inadequate control systems would experience lower abnormal returns than the firms that had adequate control systems in place. Instead of our research providing a definitive answer to the impact of SOX, we hope our research opens to door to future study in this area.

Conclusion

More reliable financial statements, effective internal control mechanisms, and improved ethical culture may lead to improved long-term performance. However, at what cost? The actual compliance costs of SOX are difficult to measure directly. Firms have an incentive to over-report the costs of SOX in hopes of having the law repealed or changed. In addition, the benefits of SOX are difficult to measure in the short run. Thus, we chose to employ an event study analysis to examine the abnormal market reaction to the passage of SOX, which allowed us to gauge non-linearity in the benefits-cost structure of SOX. We show that small firms experienced larger negative abnormal returns associated with the passage of SOX than did large firms. The relationship between firm size and negative abnormal returns was almost monotonic; that is to say, the negative returns associated with SOX decreased as firm size increased. The results of our study should lead to further consideration of the burdens associated with SOX compliance, and we hope more effort is placed on determining the impact that this regulation has imposed on small firms.

References

1. American Electronics Association (AeA). "Sarbanes-Oxley Section 404: The 'Section' of Unintended Consequences and its Impact on Small Business," 2005.

2. Asthana, S., Balsam, S. and Kim, S. "The Effect of Enron, Andersen, and Sarbanes-Oxley on the Market for Audit Services," SSRN Working Paper, 2004.

3. Block, S. B. "The Latest Movement to Going Private: An Empirical Study," Journal of Applied Finance, 2004, 36-37.

4. Brown, S. and Warner, J. "Measuring Security Price Performance," Journal of Financial Economics 8, 1980, 205-258.

5. Carney, W. J. "The Costs of Being Public After Sarbanes-Oxley: The Irony of 'Going Private'," Emory Law and Economics Research Paper, 2005.

6. Chhaochhaira, V. and Grinstein, Y. "Corporate Governance and Firm Value--the Impact of the 2002 Governance Rules," Cornell University Working Paper, 2004.

7. Engel, E., Hayes, R. M., and Wang, X. "The Sarbanes-Oxley Act and Firms' Going-Private Decisions," SSRN Working Paper, 2004.

8. Financial Executives International. "Sarbanes-Oxley Compliance Costs Exceed Estimates," Press Release, 2005.

9. Leuz, C., Triantis, A. J., and Wang, T.Y. "Why Do Firms Go Dark? Causes and Economic Consequences of Voluntary SEC Deregistrations," AFA 2006 Boston Meetings Paper, 2004.

10. Levitin, M. J. and Snider, S. "Looking Ahead: The Market for Buyouts of Public Companies, Part II, Buyouts," LexisNexis Academic, 2002.

11. Li, H., Pincus, M. P.K. and Rego, S.O. "Market Reaction to Events Surrounding the Sarbanes-Oxley Act of 2002," SSRN Working Paper, 2004.

12. Linck, J. S., Netter, J. M. and Yang, T. "Effects and Unintended Consequences of the Sarbanes-Oxley Act on Corporate Boards," AFA 2006 Boston Meetings Paper, 2006.

13. Rezaee, Z. and Jain, P.K. "The Sarbanes-Oxley Act of 2002 and Security Market Behavior: Early Evidence," SSRN Working Paper, 2005.

14. Ribstein, L. E. "Sarbanes-Oxley After Three Years," University of Illinois Law & Economics Research Paper, 2005.

15. Romano, R. "The Sarbanes-Oxley Act and the Making of Quack Corporate Governance," SSRN Working Paper, 2004.

16. Zhang, I.X. "Economic Consequences of the Sarbanes-Oxley Act of 2002," University of Rochester, Working Paper, 2005

17. Zhu, H. and Small, K. "Has Sarbanes-Oxley Led to a Chilling in the U.S. Cross-listing Market?" Loyola College, Forthcoming CFA Journal.

Ken Small, Coastal Carolina University

Octavian Ionici, Loyola College in Maryland

Hong Zhu, Loyola College in Maryland Exhibit 1: MAJOR PROVISIONS of SARBANES-OXLEY ACT OF 2002 Title I. Public Company Accounting Oversight Board (Sections 101-109) II. Auditor Independence (Sections 201-209)

Section 201 Prohibits auditor from providing certain non-audit-

related services

Section 203 Requires rotation of the senior audit officials

Section 204 Requires the firm's auditor to communicate with the

audit committee

Section 206 Addresses conflicts of interest that may arise if

management were employed by the auditor III. Corporate Responsibility (Sections 301-308)

Section 301 Requires audit committee members be independent

board members

Section 302 Requires CEO/CFO certification of the firm's

financial statements

Section 303 Addresses improper influence of audits IV. Enhanced Financial Disclosures (Sections 401-409)

Section 401 Requires periodic reports including off-balance

sheet transaction

Section 403 Requires disclosure of transactions involving

management and principal stockholders

Section 404 Addresses management's assessment of internal

controls

Section 407 Requires audit committee contain one financial

expert V. Analyst Conflicts of Interest (Section 501) VI. Commission Resources and Authority (Sections 601-604) VII. Studies and Reports (Sections 701-705) VIII. Corporate and Criminal Fraud Accountability (Sections 801-807)

Section 802 Imposes criminal penalties for altering documents

Section 803 Makes debts non-dischargeable if incurred in

violation of securities fraud laws.

Section 806 Provides protection for employees of publicly

traded companies who provide evidence of fraud

Section 807 Provides criminal penalties for defrauding

shareholders of publicly traded companies. IX. White-Collar Crime Penalty Enhancements (Sections 901-906)

Section 906 Imposes corporate responsibility for financial

reports X. Corporate Tax Returns (Section 1001) XI. Corporate Fraud and Accountability (Section 1101-1107)

Section 1102 Covers tampering with a record or otherwise

impeding an official proceeding

Section 1106 Increases criminal penalties under Securities

Exchange Act of 1934

Section 1107 Addresses retaliation against informants Exhibit 2: SUMMARY of STUDIES on SARBANES-OXLEY COMPLIANCE COSTS

Companies Time Frame and Studies Selected Focused Costs Results American Financial Section 404 * 0.06% of Revenue when

Electronics Executives Compliance Revenue > 5 Billion

Association International Costs * 0.16% of Revenue when

(AeA) 2005 study and 5 Billion > Revenue >

AeA's Section 1 Billion

404 working * 0.27% of Revenue when

group's 1 Billion > Revenue >

estimation 500 Million

* 0.53% of Revenue when

500 Million > Revenue

> 100 Million

* 2.55% of Revenue when

Revenue < 100 Million Financial 217 public Year 1 of * Internal Labor Costs:

Executives companies Section 404 $1,337,935

International with average Compliance * External Consulting

(FEI) March revenue of Costs and Technology Costs:

2005 Survey $5 billion $1,716,987

* Auditor Attestation

Fees: $1,301,050

* Total Section 404

Compliance Costs:

$4,355,972 Carney 2005 44 companies Year 2004 * Median Net Income:

filed to go Compliance $421,000

private that Cost of * Average costs of

mentioned the securities compliance to

costs of laws including securities laws:

compliance SOX $291,000

* Average reported SOX

costs: $174,000

Exhibit 3: ONE-DAY ABNORMAL RETURNS by QUINTILE

Exhibit 3 includes results of the event study analysis around the day that the U.S. House and Senate agreed on the final Sarbanes-Oxley bill (July 24, 2002). The standard market model is used to construct the abnormal returns. Abnormal returns are calculated as: A[R.sub.it] = [R.sub.it] - [[alpha].sub.[iota]] - [beta]([R.sub.mt]), where [R.sub.it] is the return to firm i on day t, [[alpha].sub.[iota]] is an intercept term, [beta] is the market beta for the ith firm estimated using the 225 trading days prior to 90 trading days returns before July 24, 2002, and [R.sub.m] is the market return on day t. The CRSP value-weighted market index is used as a proxy for the market returns. Quintile Assets Day (0) Market Capitalization Day (0) 1 (Smallest) -2.97*** -2.54***

(14.13) (12.46) 2 -3.17*** -3.66***

(16.82) (24.08) 3 -1.86*** -2.53***

(10.86) (14.57) 4 -1.20*** -.94***

(6.66) (3.32) 5 (Largest) .22*** .04***

(12.08) (10.68) All Firms -1.79*** -1.93***

(16.02) (19.57) *** indicates significance 1% level ** indicates significance 5% level * indicates significance 10% level Patell Z-statistic in parentheses

Exhibit 4: ABNORMAL RETURNS BY DECILE (Asset Deciles)

Exhibit 4 includes results of the event study analysis around the day that the U.S. House and Senate agreed on the final Sarbanes-Oxley bill (July 24, 2002). The standard market model is used to construct the abnormal returns. Abnormal returns are calculated as: A[R.sub.it] = [R.sub.it] - [[alpha].sub.[iota]] - [beta]([R.sub.mt]), where [R.sub.it] is the return to firm i on day t, [[alpha].sub.[iota]] is an intercept term, [beta] is the market beta for the ith firm estimated using the 225 trading days prior to 90 trading days returns before July 24, 2002, and [R.sub.m] is the market return on day t. The CRSP value-weighted market index is used as a proxy for the market returns. Cumulative abnormal returns (CAR) are the sum of the one-day abnormal return estimates.

AR CAR CAR Number Decile Day (0) Days (-1, 0) Days (-1, 0, 1) 1 (Smallest) -3.27*** -4.33*** -3.43*** 428

(10.14) (9.84) (6.32) 2 -2.26*** -4.04*** -4.27*** 433

(9.84) (10.89) (8.94) 3 -3.38*** -4.76*** -5.11*** 420

(12.16) (13.66) (11.74) 4 -2.98*** -4.29*** -4.82*** 437

(11.64) (14.41) (12.15) 5 -1.91*** -3.04*** -3.31*** 432

(8.09) (11.06) (8.83) 6 -1.81*** -2.75*** -2.41*** 423

(7.26) (10.23) (5.94) 7 -1.63*** -2.51*** -2.63*** 433

(7.71) (11.49) (7.67) 8 -.77** -2.00*** -2.07*** 429

(1.62) (9.88) (6.08) 9 -.21*** -1.48*** -1.35 429

(2.77) (6.54) (.713) 10 (Largest) 64*** -.83 -.28*** 439

(14.27) (.048) (7.199) All Firms -1.79*** -3.00*** -2.96*** 4,303

(16.20) (30.98) (19.30) *** indicates significance 1% level ** indicates significance 5% level * indicates significance 10% level Patell Z-statistic in parentheses

Diagram 1: ONE-DAY ABNORMAL RETURNS GRAPH (Asset Deciles)

Diagram 1 is a graph of one-day abnormal returns from the asset-based sample. The standard market model is used to construct the abnormal returns. Abnormal returns are calculated as: A[R.sub.it] = [R.sub.it] - [[alpha].sub.[iota]] - b(Rmt), where [R.sub.it] it the return to firm i on day t, [[alpha].sub.[iota]] is an intercept term, [beta] is the market beta for the ith firm estimated using the 225 trading days prior to 90 trading days returns before July 24, 2002, and [R.sub.m] is the market return on day t. The CRSP value-weighted market index is used as a proxy for the market returns

[GRAPHIC OMITTED]