Court Intervention as a Governance Mechanism over CEO Pay: Evidence from the Citigroup Derivative Lawsuit

Abstract

We use an unanticipated court ruling in a lawsuit against Citigroup claiming corporate waste related to CEO pay to analyse court intervention as an alternative governance mechanism in cases of excess pay. We find a negative relation between announcement returns and excess pay, consistent with shareholders of these firms perceiving court intervention as net costly. However, we find a positive relation between announcement returns and excess pay accompanied by poor performance, suggesting that intervention is welcome when pay is more egregious. Finally, we find that firms with excess pay and whose shareholders welcome intervention reduce future pay relative to other firms, suggesting that the threat of court intervention is a potential mechanism to control excess pay.

Acknowledgements

We gratefully acknowledge the financial support of Boston College, Boston University, and the University of Virginia Darden School Foundation. We appreciate the research assistance of Craig Tessiatore and Jon Ashley. We thank an anonymous reviewer, David Erkens, Yaniv Grinstein, Justin Hopkins, Krish Menon, Susan Shu, Ewa Sletten, Anup Srivastava, and Jerry Zimmerman. We also appreciate the comments from seminar participants at Concordia University, Georgetown University, University of California at Berkeley, University of North Carolina at Charlotte, Portuguese Catholic University 2012 Accounting Conference, the American Accounting Association 2012 annual meeting, the 2013 American Accounting Association Management Accounting and Financial Reporting Section, the 2013 Ackerman Conference on Corporate Governance at Bar Ilan University.

Notes

¹ http://www.perkinscoie.com/news/pubs_detail.aspx?publication=2075&op=updates.

²Cooch (2011) defines the business judgement rule as ‘the presumption enjoyed by corporate boards that their actions are presumed to be made on an informed basis, in good faith, and in the honest belief that the action taken was in the best interests of the company’.

³Examples of the standard against which claims of corporate waste are judged include ‘an exchange that is so one sided that no business person of ordinary, sound judgment could conclude that the corporation has received adequate consideration’ (Brehm v. Eisner 746 A.2d (Del. 2000)); ‘ … the plaintiff must overcome the general presumption of good faith by showing that the board's decision was so egregious or irrational that it could not have been based on a valid assessment of the corporation's best interests’ (White v. Panic, 783 A.2d (Del. 2001)); and

The standard under which the Court evaluates a waste claim is whether there was ‘an exchange of corporate assets for consideration so disproportionately small as to lie beyond the range at which any reasonable person might be willing to trade.’ (Opinion of the Court of Chancery of the State of Delaware In Re Citigroup Inc. Shareholder Derivative Litigation, Civil Action No. 3338-CC, 24 February 2009.)

⁴See Thomas and Martin (2001) for a detailed discussion of the demand requirement and plaintiffs' success rates.

⁵Thomas and Martin (2001) document some success in executive pay cases, particularly in litigation against closely held corporations and in litigation in non-Delaware courts.

⁶Thomas and Wells (2011) offer an explanation for the hesitancy of courts to overturn board decisions related to pay, advocate the use of the courts in resolving pay issues, and propose an alternative doctrine under which courts could pursue involvement in the process.

⁷See In re Citigroup Inc. Shareholder Derivative Litigation, 964 A.2d 106 (Del. Ch. 2009) (No. 3338-CC) 2007 WL 5208459.

⁸A shareholder derivative lawsuit is a special case of shareholder suit and differs from class action lawsuits. Derivative lawsuits are brought on behalf of the corporation, as a whole, therefore representing the welfare of all shareholders. In these suits, the corporation is the claimant against a company's directors or officers. In contrast, class action lawsuits are filed by a subset of shareholders, when alleged negligence or fraud by a company's directors or officers leads to a loss of shareholder value, and only these shareholders receive the monetary rewards that may be obtained from the lawsuit.

⁹The plaintiffs were Montgomery County Employees' Retirement Fund, City of New Orleans Employees' Retirement System, Sheldon M. Pekin Irrevocable Descendants Trust Dated 1 October 2001, and Carole Kops.

¹⁰See the following for summaries of the claims and the outcome of the case: Cooch (2011), http://www.delawarelitigation.com/2009/02/articles/chancery-court-updates/chancery-court-dismisses-shareholder-claims-against-citigroup-for-failure-to-monitor-subprime-risks-but-allows-waste-claim-for-ceo-pay/.

¹¹The case was ultimately dismissed on 5 March 2012 (see court filings under docket number 3338-VCG (Del. Ch.) (Glasscock, V.C.)). We do not examine the market reaction to the dismissal because the significant, unexpected event was allowing the claim to proceed in court. Furthermore, we were unable to find any announcements, new wires, or other media disclosures related to it. Legal analysts clearly did not anticipate that eventual outcome, given their reaction to the initial ruling.

¹²Over 50% of all publicly traded firms in the USA choose Delaware as their legal home, with the state of second choice representing fewer than 5% of public firms (Bebchuk & Cohen, 2003; Daines, 2001).

¹³ http://www.duanemorris.com/alerts/alert3190.html (Retrieved May 7, 2012).

¹⁴While our main interest is in the failure to dismiss the corporate waste claim, the court also announced that it would dismiss a claim that Citigroup's directors breached their fiduciary duty by failing to adequately monitor and disclose exposure to subprime mortgages. It is possible that shareholders' reactions reflect concerns regarding their inability to obtain recourse through the courts regarding breach of fiduciary duty based on this ruling. While this concurrent ruling is not likely to have come as a surprise to shareholders (three subprime mortgage and credit–crisis shareholder derivative lawsuits had been granted motions to dismiss within the prior five months), we address this possibility in robustness tests.

¹⁵‘Delaware Chancery Court Revisits Oversight Liability and Corporate Waste’, Duane Morris LLP, 25 March 2009 (http://www.duanemorris.com/alerts/alert3190.html).

¹⁶We obtain this information from the Treasury Department (www.financialstabiliy.gov) as of 4 November 2009.

¹⁷See http://www.delawarelitigation.com/2009/02/articles/chancery-court-updates/chancery-court-dismisses-shareholder-claims-against-citigroup-for-failure-to-monitor-subprime-risks-but-allows-waste-claim-for-ceo-pay/ and http://www.willkie.com/files/tbl_s29Publications%5CFileUpload5686%5C2906%5CDelaware_Chancery_Court_Refuses_to_Hold_Citigroup_Directors.pdf.

¹⁸See Chase (2009) or http://www.bloomberg.com/apps/news?pid=newsarchive&sid=aClCw7s25FDk.

¹⁹The results from estimating Equation (1) are robust to using value-weighted returns with dividends, equal-weighted returns with and without dividends, three-day raw returns, and three-day size-adjusted returns, as well as to estimating the market model in the prior year (for 255 days ending on Day −255) to avoid the volatility associated with the financial crisis of 2008.

²⁰We obtain the state of incorporation (DEL) from Compustat as of the beginning of 2012 and assume that this designation is consistent with that in February 2009. However, to ensure that this assumption does not affect our conclusions, we estimate our regression excluding the indicator variable DEL, and our conclusions are unchanged.

²¹We first search for litigation cases filed with the district of Delaware using the following keywords: ‘derivative action’ or ‘derivative litigation’ or ‘derivative suit’ and ‘waste claim’. We then read all the claims and dropped the ones that were not related to excess compensation claims. We also excluded ‘say-on-pay’ lawsuits, mostly filed in the post-Citigroup period, but included stock options backdating lawsuits that were filled in the pre-Citigroup period to avoid overstating the impact of the Citigroup case.

²²Not surprisingly, %EXCESSPAY and EXCESSPAY are highly correlated. If we exclude %EXCESSPAY and allow EXCESSPAY to capture mean reversion, the results are stronger and our conclusions from Table 4 are unchanged.

²³Because of potential concerns with event date clustering, we also examine the market reaction using a portfolio approach where firm and market returns are aggregated by date (Schipper & Thompson, 1983). We test whether returns during the 3-day event window are different from the market over the period (379 trading days) 25 February 2008–24 August 2009. We find (in untabulated results) that the returns in the 3-day event window are negative, although not significant at conventional levels. These portfolio tests are inherently low powered, yet that the returns are negative provides some support that econometric properties of event clustering are not driving our results.

²⁴Using actual February 17 returns as a control, instead of the expected February 24 return, yields similar results.

²⁵Our conclusions are unchanged if MOMENTUM is excluded from these analyses.

²⁶This is obtained by evaluating all variables at their sample median, except POSCAR and POSCAR*EXCESSPAY which are evaluated at the median and 75th percentile of firms with positive announcement returns.

²⁷The Sefcik and Thompson (1986) methodology involves several steps. First, we create portfolios for each firm characteristic following the weights structure specified in Sefcik and Thompson for each firm. Then, we compute daily returns (excluding dividends) from 25 February 2008 to 24 August 2009 for each firm-characteristic portfolio. Next, we estimate a regression of the daily portfolio returns on the value-weighted market return (excluding dividends) and an event indicator variable for the period 24 to 26 February 2009. In this specification, the coefficient estimate on the event variable is unbiased, and the standard errors fully account for the cross-sectional disturbance heteroskedasticity and interdependence.

²⁸Specifically, we calculate the per cent analyst forecast error as the actual earnings per share announced minus the most recent mean analyst earnings per share forecast for the fiscal year obtained from I/B/E/S, deflated by the absolute value of actual earnings per share.

²⁹Data on EINDEX are obtained from Lucian Bebchuk's website, Retrieved July 18, 2011: http://www.law.harvard.edu/faculty/bebchuk/data.shtml. Bebchuk et al. (2009) define the EINDEX as an entrenchment index that is calculated based on six governance provisions that matter the most out from the provisions included in GINDEX.

³⁰Our conclusions are unchanged when we use either 2007 or 2008 excess pay for the 2008 event dates.