Incentive alignment through performance-focused shareholder proposals on management compensation

Accepted Manuscript Incentive Alignment through Performance-Focused Shareholder Proposals on Management Compensation Steve Fortin, Chandra Subramaniam, Xu (Frank) Wang, Sanjian (Bill) Zhang PII: DOI: Reference:

S1815-5669(14)00014-9 http://dx.doi.org/10.1016/j.jcae.2014.06.001 JCAE 53

To appear in:

Journal of Contemporary Accounting & Economics

Received Date: Revised Date: Accepted Date:

22 July 2013 20 May 2014 29 May 2014

Please cite this article as: Fortin, S., Subramaniam, C., Wang, X., Zhang, S., Incentive Alignment through Performance-Focused Shareholder Proposals on Management Compensation, Journal of Contemporary Accounting & Economics (2014), doi: http://dx.doi.org/10.1016/j.jcae.2014.06.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Incentive Alignment through Performance-Focused Shareholder Proposals on Management Compensation Steve Fortin McGill University [email protected] Chandra Subramaniam University of Texas at Arlington [email protected] Xu (Frank) Wang* Saint Louis University [email protected] Sanjian (Bill) Zhang McGill University [email protected]

We thank the following for their helpful comments and suggestions: Mufaddal Baxamusa, Joy Begley, Francesco Bova, Sandra Chamberlain, Lilian Chan, Denis Cormier, David Farber, Ian Gow, Kiridaran (Giri) Kanagaretnam, Kin Lo, Russell Lundholm, Michel Magnan, Garen Markarian, Khalid Nainar, Matt Wieland, Jenny Zhang, workshop participants at the Chaire d’information financière et organisationnelle of UQAM, McMaster University, and University of British Columbia, and workshop participants at the Financial Management Association meeting, American Accounting Association annual meeting, the Canadian Accounting Academic Association meeting, and the Midwest Finance Association meeting. We also thank Andrew Metrick for generously making his data publicly available on his own website on governance index. Sanjian(Bill) Zhang also thanks the Social Sciences and Humanities Research Council of Canada for its financial support. *Corresponding author

Incentive Alignment through Performance-Focused Shareholder Proposals on Management Compensation ABSTRACT

We investigate an emerging pay-performance activism under a natural setting of performancefocused shareholder proposals rule (PSPs) (Rule 14a-8) established by the Securities and Exchange Commission (SEC) for top management compensation. We find that: (1) PSP sponsors successfully identify firms that suffer from a misalignment of managers and shareholders’ interests; (2) CEOs’ pay-for-performance sensitivity increases in the post-proposal period; and (3) shareholders benefit through positive stock returns as related to proposal filing dates; while (4) bondholders suffer significant negative returns and even more so for high-leverage firms. Our additional analyses suggest that perceived risk increase is the main driver of observed negative abnormal bond returns. However, we fail to find similar results for shareholder proposals not focused on performance (NPSPs). Collectively, our results indicate that shareholders benefit from this pay-performance activism through PSPs (but not NPSPs), but potentially at the expense of bondholders. Keywords: Bond Market Reaction, Shareholder Proposal, Incentive Compensation, Pay for Performance. JEL Classifications: M4, G30, M52. Data Availability: Data used are from public sources identified in this paper.

1

Incentive Alignment through Performance-Focused Shareholder Proposals on Management Compensation 1. INTRODUCTION

Corporate boards are conscious of the role that executive pay practices play in improving corporate governance and increasing shareholder wealth (Gammeltoft, 2010). Economic theory suggests that the key to aligning managerial compensation with shareholder interest is to increase the sensitivity of executive compensation to firm performance (Core et al., 2005; Jensen and Meckling, 1976). Firms finance their operations, however, with funds from both shareholders and creditors, e.g., bondholders. Thus, agency theory also concerns shareholder-bondholder agency conflict and the difficulty of concurrently aligning the interests of shareholders, bondholders, and managers (Ahmed et al., 2002; Jensen and Meckling, 1976; Ortiz-Molina, 2007). In the past decade, the business press has focused on excessive CEO pay, observed during the 2001 Enron/Worldcom scandals as well as the recent 2007–2008 credit crisis, e.g., AIG. Critics contend that contracting between CEOs and boards has been shadowed by pervasive managerial influence (Bebchuk and Fried, 2005; Crystal, 1992). Consistent with these concerns, shareholders have begun to use the “shareholder proposal rule” (Rule 14a-8) established by the Securities and Exchange Commission (SEC) to defend their interest and have submitted hundreds of proposals to many of the largest U.S. corporations. In this paper, we document a related concept, emerging pay-performance activism sponsored by shareholders, and examine how bondholders perceive such activism. We identify pay-performance activism as those shareholder proposals with a sharp focus on executive compensation issues tied to financial performance. Such performance-focused shareholder proposals (PSPs) are theoretically and practically different than those proposals that call only for

2 CEO pay constraints or that tie CEO pay to certain social and environmental actions (nonperformance-focused shareholder proposals [NPSPs]). The appendix provides two examples of both types. Our study centers on PSPs that demand that directors tie executive compensation more closely to firm performance, thereby realigning manager interests with those of their shareholders. Our first research question, a building block for our next and primary question, rests on the uniqueness of the increasing number of PSPs filed by investors. There has been some research into the effects of shareholder concerns or threats on stock price, financial performance, or executive compensation (Johnson and Shackell, 1997; Johnson et al., 1997; Karpoff et al., 1996). From a rational shareholder’s perspective, performance-focused proposals should be most beneficial to shareholder interests. Given that Section 953a of the Dodd-Frank Act of 2010 requires firms to disclose more details of their pay-for-performance practices, an examination of PSPs has the potential to provide timely insights to the SEC, the designated market regulator. Little is known, however, about the determinants and market impact of performance-focused shareholder proposals. 1 Thus, we extend prior research by providing new evidence on the differences between PSPs and NPSPs in terms of their economic determinants and consequences. Our primary research question is related to the probable negative side effects of PSPs on bondholders and their regulatory implications. Equity-linked compensation (especially CEO executive compensation, heavy with stock options) can increase risk-taking incentives for managers (Jensen and Meckling, 1976; Jensen and Murphy, 1990). Bebchuk and Spamann (2009), in reflecting on the crisis between 2007 and 2009, attribute bankers’ excessive risk1

In regard to determinants, most studies do not separate PSPs from NPSPs. The only exception is Ertimur et al. (2011), who found that 21.5% of their sample included proposals that linked pay to performance. While they found that all shareholder proposals, on average, appear to target large firms with high CEO pay, they do not differentiate between PSPs and NPSPs. Other prior research on compensation-related proposals has found mixed results for the usefulness of these proposals to change managerial behavior or to have any market effects.

3 taking behavior to the high equity component in executive compensation. They noted that, with the increase in executive pay sensitivity to stock price as well as to stock price volatility, management may serve the interests of shareholders through further risk-taking and at the expense of all other stakeholders, including bondholders. By constructing a sample of 136 S&P 500 companies that received at least one PSP between 1996 and 2006, we first test whether targeted firms that receive PSPs have, ex ante, pay practices that are suboptimal from an alignment perspective, as compared with control firms. The first of our two control groups consists of 262 S&P 500 companies that were not targeted by either PSPs or NPSPs between 1996 and 2006.2,3 Our second control group is comprised of 51 firms that received NPSPs during the same period. We start by studying the determinants of a firm that received a PSP relative to no proposals and/or received an NPSP. Next, we test for improvement in management incentive alignment following the proposal year by studying changes in pay-performance sensitivity for PSP firms, compared with control firms. Third, we examine stock returns as related to the proposal day to determine the effect of proposals on shareholders and the bond market reaction around the proposal day. We test our bond results at both the firm and bond levels and find that our results are robust under both settings. Our results show that firms with higher excess CEO compensation are more likely to receive PSPs, compared to firms that receive NPSPs or firms that receive no proposals. Second, firms that receive PSPs see their equity-based pay-for-performance sensitivity increase significantly following the proposal year, compared with control firms that received NPSPs or did not receive proposals. Third, as a result of PSPs, shareholders enjoy significantly positive abnormal stock returns, while bondholders suffer significant negative abnormal bond returns. 2

The S&P 500 firms are selected based on the S&P 500 list for the fiscal year 2005. We examine the robustness of our results by using two alternative methodologies to form two separate controlmatching portfolios, as seen in Section 5.7.

3

4 Further exploratory analyses suggest that high-leverage firms experience more negative abnormal bond returns than do low-leverage firms and that the volatility change after PSP proxy filing dates explains the negative bond reaction. This is consistent with the notion that pay design changes lead to more risk-taking behavior by target firms. We contribute to the literature as well as to the ongoing regulatory debate in several ways. First, to the best of our knowledge, we present the first evidence of bondholder reaction to shareholder proposals. Specifically, the realignment of manager and shareholder interests due to the PSP is associated with a decrease in bond returns. Because there is a trade-off between shareholder-manager interest alignment and shareholder-bondholder conflict (DeFusco et al., 1990; Klein and Zur, 2011; Ortiz-Molina, 2007), our results suggest that boards of directors and regulators should adopt a balanced approach in dealing with activist shareholder campaigns, particularly those concerning top management incentive compensation. The SEC was established in the 1930s with a mandate to protect investors in securities (both stocks and bonds). To fulfill its duty toward public bondholders, the second SEC chairman William Douglas lobbied Congress to pass the Trust Indenture Act of 1939 and established the bond trustee system in the United States. In response to the Dodd-Frank Act of 2010, the SEC released the new “Say-onPay” regulation in January 2011. From a bondholder’s perspective, the new SEC regulation might result in unintended consequences that have the potential to compromise its duty toward bondholders. Second, we provide the initial evidence that PSPs are very different from NPSPs 4 through our investigation of the determinants and consequences of the emerging payperformance activism through PSPs. Recent studies focus on the effect of overall shareholder

4

See also Ferri and Sandino (2009) and Ertimur et al. (2011), among others, for examples of studies of shareholder non-binding votes.

5 votes and related regulations on compensation issues (e.g., Carter and Zamora, 2009; Ferri and Maber, 2012). Prior research does not, however, differentiate PSPs from NPSPs. We extend the findings of such research by isolating PSPs from NPSPs based on incentive alignment and agency theory (Jensen and Meckling, 1976), on increased investor demand (Rappaport and Nodine, 1999), and on emerging trends in institutional practices (CalPERS, 2010). We find that PSPs and NPSPs are different with respect to the rationale for targeting a specific firm and their impact on firm pay-performance sensitivity. Shareholder proposals are often perceived by the business community and some popular business newspapers to be submitted by less-sophisticated investors and to have little effect on important governance matters, as compared with activist campaigns by large investors, such as hedge funds. Our results, however, suggest that PSP (but not NPSP) sponsors are more sophisticated investors, who understand the proper use of management compensation contracts in maximizing their own utility. 5 For instance, our results show that PSP sponsors target firm CEOs with excess compensation, while NPSP target CEOs without excess compensation. Firms targeted by PSP (but not NPSP) sponsors increase their CEO pay-for-performance sensitivity.6 Our evidence suggests that treating shareholder proposals based on whether they are performance-focused can provide additional insight into how different types of shareholders interact with their target firms. Third, our research also informs the ongoing debate about executive compensation regulation. The business press cites poor incentives as “one of the most fundamental causes” of the recent economic crisis (Blinder, 2009). According to Solomon and Paletta (2009), the Obama 5

This is also consistent with recent evidence of shareholders’ voting strategically on management proposals (Maug and Rydqvist, 2009). 6 We discuss, in the Background section, details of the rationale for separating PSPs and NPSPs and, in Sections 5.2 and 5.3, the empirical evidence of how PSPs and NPSPs are different in terms of target firms’ selection and in their effects on pay-for-performance sensitivity.

6 administration clearly believes that “more closely align[ing] pay with long term performance” is the lesson from the recent recession. Nevertheless, lack of pay-performance sensitivity does not seem to be the actual cause for the excessive risk taking between 2002 and 2007 (DeYoung et al., 2009). Further, the collapse of any high-leverage business, such as banking, e.g., Lehman Brothers, implies larger losses in absolute dollars for bondholders and average depositors than for shareholders. Interestingly, these anecdotal results are consistent with our untabulated test results that show that the bonds of high-leverage firms suffer more negative returns than do those of low-leverage firms. In summary, we are concerned that more shareholder activism, as further encouraged by the Dodd-Frank Act of 2010, could generate unintended negative effects for bondholders. The remainder of this study is organized as follows: Section 2 provides the background and literature review, and Section 3 presents the hypothesis development. We describe the research methodology in Section 4. Then Section 5 presents the results and, Section 6, the conclusion.

2. BACKGROUND OF PAY-PERFORMANCE ACTIVISM The 1942 Securities and Exchange Commission (SEC) Rule 14a-8 allows eligible shareholders7 to make their proposals in a company’s proxy statement for a vote at shareholder meetings. Popular shareholder proposals include board-related policies and executive compensation schemes, among other important corporate governance concerns (Georgeson, 2009). We focus on one emerging type of compensation-related proposals, PSPs, due to their importance to shareholder-manager interest alignment and their potential to affect bondholders.

7

With $2,000 in market value or 1% of voting shares.

7 This type of proposal has strong theoretical or public policy support (Jensen and Murphy, 1990) and is most likely to gain traction with boards and shareholders.8 The dichotomy between PSPs and NPSPs is based on economic theory and institutional practice. In regard to the new classification scheme, first, agency theory’s incentive alignment mechanism between the principal (shareholder) and the agent is essentially performance-focused (Jensen and Meckling, 1976; Jensen and Murphy, 1990); an agent should receive higher compensation if he or she enables the principal to gain larger wealth. Compared to politically- or socially-centered shareholder proposals, performance-focused ones can be tested more directly for their economic consequences. Second, shareholders are more receptive to the concept of payfor-performance now than in the early 1990s, when Jensen and Murphy first argued that it is not how much but how top management is paid. This shareholder receptiveness is supported by Rappaport and Nodine’s (1999) argument that shareholders want changes in “pay schemes that motivate companies to deliver more value.” In particular, they suggest that companies use indexed stock option, which is a primary goal of many of the PSPs. Third, the pay-forperformance principle is widely supported by the largest public pension funds, investment groups, and Institutional Shareholder Services in their voting guidelines (CalPERS, 2010; Institutional Shareholder Services [ISS], 2007; TIAA-CREF, 2010). Compared with other types of shareholder activism, such as hedge fund activism, the pay-performance activism as put forth in shareholder proposals is relatively less costly. First, as noted earlier, the eligibility requirement is low. Specifically, shareholders must own more than $2,000 in stock or 1% of the company to be eligible to send in proposals. Second, the 1992 SECreformed proxy solicitation rules and Regulations S-K and S-B pay disclosure rules allow

8

We discuss this point in more detail in the Hypothesis Development and Research Design sections.

8 shareholders to access compensation data easily and to coordinate “in proxy contests against management . . . at a significantly reduced cost” (Thomas and Martin, 1999, p.1039). Although the new rules on executive compensation disclosure provide ample means for shareholder activists to monitor company compensation practices, prior research finds mixed evidence on what kinds of firms receive shareholder executive-pay proposals. For instance, Karpoff et al. (1996) and Strickland et al. (1996) find that shareholder proposals are more common for large firms that have poor prior performance. Recent evidence (Cai and Walkling, 2011), however, indicates that target firms are simply large firms, not those with management incentive misalignment, poor governance, or poor performance. In addition, Institutional ownership has been found to be positively associated with the probability of attracting a shareholder proposal (Strickland et al., 1996), but this association is not evident in a more recent study (Ertimur et al., 2011). In fact, the only consistent result in the literature is that firm size is positively related to attracting a proposal. The impact of shareholder proposals on total executive pay level is even less clear. Thomas and Martin (1999) do not find any relationship between receiving a proposal and executive compensation or the sensitivity of CEO pay to performance. Similarly, Karpoff et al. (1996) find that there is no relationship between receiving a shareholder proposal and the level of executive compensation. In contrast, more recent studies with new samples in 1990s and 2000s reached the opposite conclusion. Ferri and Sandino (2009) and Ertimur et al. (2011) both find that shareholder proposals reduce executive pay.

9 3. HYPOTHESIS DEVELOPMENT 3.1. Lack of Alignment and Pay-Performance Activism There is a large body of agency literature on the alignment of shareholder and manager interest through compensation contracts (Carter et al., 2007; Jensen and Meckling, 1976; Holmstrom, 1979; Jensen and Murphy, 1990). It is unclear, however, whether shareholders target firms randomly or have enough sophistication to identify firms that suffer from a misalignment of shareholders and managers’ interests. We examine whether shareholders who submit PSPs are sophisticated investors. If true, they should be more likely to target firms that suffer from more severe agency problems. Core et al. (2008) propose a model to split CEO compensation between a “normal” portion justified by economic determinants and an “excess” portion, that is, the actual compensation minus the predicted compensation. They argue that the excess portion indicates that management incentives have exceeded what economic determinants warrant and that the existence of excess compensation is evidence of the misalignment of management incentives with shareholder interests. Similarly, we posit that the presence of excess compensation is an indication of misalignment. This leads to our first hypothesis, expressed in alternate form: H1: Firms with higher excess compensation are more likely to receive PSPs.

3.2 PSPs and Equity-Based Pay-for-Performance Sensitivity Next, we turn to firms’ reactions to shareholder executive-pay proposals. Although shareholder proposals are not legally enforceable upon a board of directors, as the board can choose to ignore them, shareholder proposals subject the board to social pressures. According to the literature, directors give more weight to shareholder proposals that receive over 10% of the

10 vote, viewed as a significant percentage of shareholder dissatisfaction (Thomas and Martin, 1999). Boards of directors who ignore repeated and economically sound proposals could face proxy contests, votes held against them in elections (Del Guercio et al., 2008), or litigation by some shareholders (Brill, 2004). In the past, boards also implemented a substantial portion of the requests made by shareholder “vote no” campaigns during director elections if withholding votes exceeded 20% (Del Guercio et al., 2008). Therefore, even if boards are not legally bound to implement shareholder proposals, they still could be forced to take action in line with economically sound proposals. Because PSPs are especially well grounded in a solid economic foundation, we posit that they should be supported by boards. This suggests that the stock-based pay would likely increase after firms receive PSPs. The increase in equity-based pay is also consistent with the notion that directors, when faced with shareholder pressure to improve pay-performance relations, may try to reach a certain level of managerial incentive by changing pay structures or shifting other pay components into stocks and stock options and, at the same time, by addressing concerns expressed in PSPs. Such structural shifts would increase the pay-for-performance sensitivity of stock-based compensation. This leads to the following hypothesis9: H2: The change in pay-for-performance sensitivity in equity-based compensation is greater for firms that receive PSPs than for control firms in the same period.

3.3. PSPs, Stockholder Returns, and Bondholder Returns Renneboog and Szilagyi (2010) document positive abnormal returns for all shareholder proposals, and Strickland et al. (1996) find positive returns for proposals from pension funds. 9

In H2, we do not differentiate between stocks and stock options. The current practice of increasing incentives is seen in stock options or restricted stock.

11 Bizjak and Marquette (1998), however, note negative returns for proposals that rescind poison pills. In another paper, Del Guercio and Hawkins (1999) show no evidence of significant market effects of all kinds of shareholder proposals, and we are not aware of any studies that specifically document the market reaction to PSPs. However, if PSPs have the effect of reducing excess compensation and realigning management and shareholders’ interests, then they should be associated with a positive market reaction from shareholders. Accordingly, we hypothesize a positive stock market reaction for PSP firms. This leads to our third hypothesis: H3: Stock markets react positively to shareholder proposal announcements for firms that receive PSPs. Merton’s (1974) bond pricing model predicts that bond prices are positively associated with total firm value and negatively associated with firm return volatility, usually proxied by firm stock return volatility in empirical studies, and leverage. If a firm reduces total CEO pay after receiving a PSP, then the firm value should increase due to the reduction of expected future cash outflows. Thus, bond prices would rise along with firm value. Further, the bond market may react positively if the bondholders expect the board and the CEO to re-contract to restrict valuedestroying actions, e.g., excessive perks for CEOs, when facing shareholder pressure in regard to pay for performance (Watts and Zimmerman, 1986). PSP proposals also could negatively affect bond prices by inducing more risk-seeking managerial actions, thus increasing firm return volatility. As we posit, firms that receive a PSP may choose to increase CEO equity-based compensation. Agrawal and Mandelker (1987) show that, with more stock and option holdings, managers tend to take on riskier investment projects with riskier financing. They find a positive association between managerial stock holdings and change in firm return volatility and leverage. DeFusco et al. (1990) show that board approval of

12 an executive stock option plan is associated with significantly negative bond returns and positive stock returns. Klein and Zur (2011) find that proposals by hedge funds result in the reduction of bondholder’s wealth by 3.9% as related to a 13D filing. In sum, the results from these studies are consistent with the notion that firms trade off between shareholder-manager alignment and shareholder-bondholder alignment (Ortiz-Molina, 2007). If the market expects increased firm return volatility due to more risk-seeking behavior by the top managers after they receive PSPs, then there should be a negative bond market price reaction associated with PSP filing days. Thus, our final hypothesis is: H4: Bond markets react negatively to shareholder proposal announcement for firms that receive PSPs.

4. RESEARCH DESIGN 4.1. Sample Selection We start with S&P 500 firms from the 2005 list. We collect shareholder executive-pay proposal data and proxy filing dates from the corporate proxy statements of S&P 500 companies for the filing years 1996 through 2006. We then obtain data from Execucomp, Compustat, CRSP, TRACE, Thomson Reuters, and Mergent Fixed Income Securities Database (FISD). After deleting 51 firms that were acquired during or immediately after the proposal year or whose proposals do not have proxy-filing dates or lacked required information, our primary test sample has 136 PSP proposal firms. Following Core and Larcker (2002) and Cheng and Farber (2008), we construct the first set of 262 control firms (1,752 firm-years) that receive no proposals during our sample period, between 1996 and 2006.10 Then, we identify the second set of 51 control

10

As a robustness check, we also select from the list of S&P 500 firms two no-proposal firms by industry, size, and book-to-market to match each PSP firm. We rerun the regressions, as seen in Panel A of Tables 4 and 5, based on

13 firms that received NPSPs during our sample period. Both sets of control firms must have all required information. For our bond return study, we use a reduced set of 86 firms that received PSPs due to insufficient bond trading data in the TRACE database during the sample period or around the proxy statement filing dates.11 We label the year “Proposal Year” (or “Year 0”) when a shareholder proposal is reported in the proxy. The sample of shareholder compensation proposals is classified as either PSP or NPSP. As discussed earlier, PSPs include those that specifically link pay to performance. NPSPs are primarily those that wield pressure for corporate support for certain social causes, e.g., environmental issues or human rights issue, or that call for capping or reducing compensation without consideration for performance. Figure 1 shows a substantial increase in the number of PSP proposals among S&P 500 firms over the 1996–2006 period. The number of PSP proposals ranges from a low of 3 in 1996 to a high of 65 in 2003. The figure also shows that the number of PSPs is lower than that of NPSPs in 1996–2000 but higher in 2001–2006. [Insert Figure 1 here]

4.2. The Probability of Receiving PSPs To test whether a firm with excessive CEO pay relative to performance is more likely to receive a PSP, we first find the deviations in executive pay from expected pay using a methodology similar to that of Core et al. (2008). We perform a pooled cross-sectional

this alternative control sample. The results are consistent with our main results. We discuss the details in Section 5.3. 11 Of the 136 PSP firms, 11 firms do not have any required bond trades in TRACE in our sample period, 34 firms do not have any required trades within 5 trading days around the proxy statement filing dates, and 5 firms do not have required trades for the same bond both before and after the proxy statement filing dates. Therefore we are left with 86 PSP firms in our bond return sample.

14 regression of all firms in the S&P 500 index with identified determinants for expected pay, as drawn from the literature, for the fiscal years between 1994 and 2005. The residuals from equation (1) provide us with the magnitude of the deviations of CEO pay from expected pay based on economic determinants: Log (TotalComp ) it −1 = β 0 + β 1 Log ( Sales ) it −1 + β 2 market _ to _ book it −1 + β 3 ROAit −1

(1)

+ β 4 stock _ returns it −1 + β 5 ROA _ volatility it −1 + β 6 stock _ volatility it −1 + β 7 Leverage it −1 + γ 1...k year _ indicators + γ 1...k industry _ indicators + u it −1

The dependent variable log(TotalComp) is the log-normalized total of CEO compensation. The log-transformation helps to explain the residual as a percentage deviation of the actual from the expected total compensation. The independent variables are firm size (Log Sales), growth opportunity (market_to_book), firm performance (ROA and stock_returns), firm risk (ROA_volatility and stock_volatility), and agency cost of debt (Leverage), which are known to be associated with executive compensation (Core et al., 1999; Mehran, 1995; Smith and Watts, 1992). After obtaining the total compensation residuals from the benchmark model, we combine our PSP sample and no proposal control sample and run a logistic regression using the following model:

PROPOSALit = β 0 + β1 ExcessCompit −2 + β 2 InternalGovit −1 + β 3 ExternalGovit −1 + β 4 Poor _ indicatorit −1 + β 5 FirmSizeit −1 + ε it

(2)

PROPOSAL is coded 1 for a firm with a PSP and 0 for a firm that did not receive any shareholder proposals on compensation. ExcessCompit-2 is the one-year lagged residual value prior to the proposal year obtained from the executive-pay benchmark model in Equation (1). We control for other governance factors that may affect the likelihood of a firm’s being targeted by a shareholder proposal. InternalGovit −1 is the percentage of shares held by the large institutional

15 blockholders, whose ownership is higher than 5% of a firm’s outstanding shares (Duchin, 2010; Yun, 2009). ExternalGovit −1 is a modified version of the governance index designed by Gompers et al. (2003). In Gompers et al.’s index, as interpreted in Cremers and Nair (2005), a higher value means lower anti-takeover vulnerability and, thus, weaker external governance (Dittmar and Mahrt-Smith, 2007; Klock et al., 2005). For convenience, we multiply the index value by -1 to derive a ExternalGovit −1 variable, which indicates stronger external governance with a higher index value. Poor _ indicatorit −1 is 1 if the five-year industry-adjusted stock return is negative and 0 otherwise. In addition to controlling for FirmSize, which is the log-transformed total assets, we augment this model with other control variables in sensitivity analyses, as seen in Table 4. Consistent with the prediction in H1, we expect a positive sign for β1 for PSP firms.

4.3. Impact of PSPs on Equity-Based Pay-Performance Sensitivity We measure CEO incentive alignment (or pay-performance sensitivity) as the change in the dollar value of annual stock options and restricted stock grants for a 1% change in the stock price (Baker and Hall, 2004; Core and Guay, 1999). Because we want to know whether boards of directors take action after receiving PSPs to improve pay-performance sensitivity, we use the measure from Core and Guay, in its logarithmic transformation, to proxy for CEO annual incentive alignment. We test the change in pay-performance sensitivity from T-1 year to T+1 year. Similar to Hartzell and Starks (2003), we use an ordinary least squares regression model in Equation (3) to control for factors that might affect compensation incentives: ΔLog ( AnnualIncentive) = β 0 + β 1 PROPOSAL + γ 1 Δ( ShareholderWealth) it + γ 2 Δ( ShareholderWealth) it −1 + ∑ γ k (OtherControlVariables) + ε it

(3)

16 where Δ Log(AnnualIncentive) is the change in pay-performance sensitivity of restricted stock and stock options from the pre-proposal period to the post-proposal period. We measure payperformance sensitivity similar to the method used by Core and Guay (1999). We use the difference in the sensitivity because the selection of control firms itself may be insufficient to eliminate completely any differences in pay structures between sample firms and control firms. PROPOSAL is 1 if a firm receives a PSP and 0 if a firm does not receive any executive pay proposals throughout the sample period. Δ(ShareholderWealth) is the change in shareholder’s wealth for the two-year period prior to the proposal year. This is the control for shareholder wealth change that is associated with CEO compensation (Hartzell and Starks, 2003). Other control variables include firm size (proxied by market capitalization) and growth opportunities (proxied by Tobin’s Q) because they potentially influence the equity compensation ratio (Mehran, 1995) measured at the beginning of the proposal year. Hartzell and Starks (2003) find that higher institutional ownership concentration is associated with subsequent increases in pay-performance sensitivity. Therefore, we include institutional blockholder ownership as an independent control variable.12 Firms may award CEOs large sums of equity compensation in some years but not in subsequent years. We control for this using unexpected prior year total compensation calculated from Equation (1). Finally, we use an indicator variable NEWCEO (1 if a new CEO takes office during the proposal year or following the proposal year and 0 otherwise) to control for CEO turnover (Hartzell and Starks, 2003).

12

We use our InternalGov as the control variable here because it represents institutional blockholder ownership. We also control for external governance (ExternalGov) because it may be related to CEO pay-for-performance sensitivity (Shaw and Zhang, 2010).

17 We hypothesize a positive sign for the coefficient β1 , which implies that we expect the equity-based pay-for-performance sensitivity to increase following the proposal year for firms that receive a PSP.13

4.4. Impact of PSPs on Stock Markets To test the stock market reactions as associated with the proposal filing dates, we compute the abnormal stock return around proxy filing dates based on Fama and French’s (1993, 1996) three-factor market model with a CRSP equal-weighted index. The event window is between the fifth trading day prior to day zero (proposal filing date) and the first trading day after day zero. Stated another way, the event window is [-5,+1]. We test the abnormal returns using the full proposal sample as well as the sample of firms with available bond trading data. We hypothesize positive abnormal stock returns around proxy filing dates.

4.5. Impact of PSPs on Bond Markets We measure bondholder returns around proxy filing dates following methodologies, similar to those of the literature, that use daily bond prices (e.g., Bessembinder et al., 2009; Hand et al., 1992). We use the total bond return approach (price change plus accrued interest) to calculate buy-and-hold bond return (BHR) as follows:

BHR =

[(Pt +1 − Pt −5 ) + AI ] Pt −5

(4)

where Pt +1 is the bond price on the first trading day after day zero, Pt −5 is the earliest available trading price for the same bond within the five trading days prior to day zero, and AI is the 13

Similar to what is seen in Hartzell and Starks (2003), none of the measures includes change in compensation due to an increase or a decrease in the value of stock and options that the CEO already holds.

18 accrued interest. Our event window design also allows us to capture the effect of any possible information leakage prior to the proxy statement filing with the SEC. We eliminate the observation if the bond did not trade within five trading days after day zero.14 The adjusted bond return is the holding period BHR of a sample firm bond less the same period BHR of a treasury bond (Hand et al., 1992) matched by maturity, as follows:

ABHRijt = BHRijt − TBHRijt

(5)

where ABHR ijt is the adjusted bond buy-and-hold return, BHR ijt is the corporate bond buy-andhold return between day t-5 and day t+1 for bond j at firm I, and TBHR ijt is the treasury buy-andhold return in the same period. 15 We hypothesize negative abnormal bond returns around proxy filing dates.

5. ANALYSIS OF RESULTS 5.1. Descriptive Statistics Table 1 presents a summary of the industry distribution for the PSP firms, the NPSP firms, and the no-proposal control firms. The PSP sample distribution is similar to that of the control firms in the S&P 500 as well as to the distribution of firms in the Compustat universe, with one exception: The sample has a larger percentage of energy and utility companies (twodigit SIC code 49). [Insert Table 1 here]

14

We calculate a weighted-average price if there is more than one trade during the trading day and eliminate noninstitutional trades to increase the power of the test (Bessembinder et al., 2009). We test the sensitivity of event windows of bond market reactions by using different windows, such as [-4, +1] or [-4, +4]. The results are invariant to these different choices. 15 As a sensitivity check, we test abnormal returns using a matching portfolio model. The results are qualitatively similar.

19 Table 2 presents descriptive statistics of the data sample based on firm-years, 136 for the PSP sample, 51 for the NPSP sample, and 1,752 for the no-proposal control sample. Compared to the control firms in the S&P 500 index, the PSP target firms are much larger in terms of size, based on market capitalization and shares outstanding, consistent with previous papers (e.g., Bizjak and Marquette, 1998; Cai and Walkling, 2011). In addition, total compensation for CEOs is also significantly higher for PSP and NPSP firms compared to control firms, similar to the results of Ertimur et al. (2011). For example, the mean (median) market capitalization for firms with at least one PSP is $29,366 ($14,395) million, compared with $11,191 ($6,358) million for control firms. The mean (median) CEO total compensation for firms with at least one performance-focused pay proposal is $12.0 ($9.2) million, compared to $7.2 ($5.2) million for control firms. The results are significantly different in each case, at least at the 0.05 level. There is no significant difference between PSP and NPSP firms for any of the variables identified here. Similar to PSP firms, NPSP firms are significantly different from control firms that receive no proposals. [Insert Table 2 here]

5.2. Probability of Receiving PSPs Table 3 presents the benchmark model, Equation (1), to determine the total compensation residuals as a proxy for excess pay used in the logistic regressions seen in Table 4. All coefficients, except for ROA volatility, in Table 3 are significant at conventional levels. [Insert Table 3 here]

20 In Table 4, we present the results of the regression for Equation (2). As predicted by H1, the probability of receiving PSPs increases when CEOs enjoy excess compensation, with a positive and significant coefficient (p-value < 0.01) for ExcessComp. Control variables show the expected signs and are significant at conventional levels. [Insert Table 4 about here]

To shed more light on whether PSPs are intrinsically different from NPSPs with respect to shareholder targeting and its impact, we study whether the probability of receiving a PSP (but not an NPSP) proposal is related to excess compensation. We re-estimate Equation (2) with the dependent variable as 1 if the firm receives an NPSP and 0 if it receives no proposal. The results presented in Panel B show that the coefficient on ExcessComp is -0.09 and is not statistically significant. However, similar to the situation with PSP firms, we find that investors target large and poorly performing firms with NPSPs. Finally, we reduce the sample to only those firms that receive proposals. We re-estimate Equation (2) with the dependent variable as 1 if the firm receives a PSP and 0 if it receives an NPSP. This test reduces the sample size from 1,888 to 187. Panel C in Table 4 shows that the coefficient of ExcessComp is positive and statistically significant (p-value = 0.053), even with the small sample size. This suggests that the PSP (but not the NPSP) firms are targeted due to higher excess compensation, which implies that PSPsponsoring shareholders are able to distinguish the target firms with excessive CEO pay to ameliorate agency cost.

21 5.3. PSPs and the Change in Equity-Based Pay-Performance Sensitivity In the next step, we test whether boards make changes to their compensation incentives following the proposal year and present the results in Table 5. In Panel A, the coefficient for the variable PROPOSAL is positive and significant at the 0.001 level. Our result suggests that payperformance sensitivity for PSP firms increases significantly more than for the control firms that receive no proposal during our study period, which provides empirical support for H2.16,17 In sum, our results are consistent with the hypothesis that boards of firms that receive PSPs change their compensation structure by increasing pay-performance sensitivity following the proposals more than those of control firms that did not receive any shareholder proposals. Next, we test whether PSPs differentially affect the pay-performance sensitivity from NPSPs. We do this by modifying Equation (3) and including an additional dummy variable for NPSP, as given in Equation (6) below: ΔLog ( AnnualInce ntive ) = β 0 + β 1 PSP + β 2 NPSP + γ 1 Δ ( Shareholde rWealth ) it + γ 2 Δ ( Shareholde rWealth ) it −1 + ∑ γ k (OtherContr olVariable s ) + ε it

( 6)

The variable PSP (NPSP) is equal to 1 if a firm receives a PSP (NPSP) proposal and 0 otherwise during the sample period. We expect the coefficient of PSP to be positive but not the coefficient of NPSP. Panel B of Table 5 shows that the coefficient of PSP continues to remain positive and significant (at less than the 0.001 level), which implies that pay-performance sensitivity increased following the proposal year. However, the NPSP coefficient is positive but not 16

To ensure that the results are not driven by firm size, we reduce the original sample into a size-truncated sample, following Pittman and Fortin (2004) and Blackwell et al. (1998), and rerun the regression in Table 5. The results are qualitatively similar. 17 In our test presented in Table 5, we use the sensitivity of equity-based pay to stock prices (delta) to provide incentives for managers to increase stock prices. We also test whether there is a change in managerial risk-taking behavior by computing the sensitivity of equity-based pay to stock volatility (vega) and rerunning the test for Table 5. The results are similar.

22 significant at any conventional level. This result implies that, unlike PSP proposals, not all compensation-related proposals have an effect on changing the CEO pay structure to increase the alignment between CEO pay and performance. [Insert Table 5 here]

We recognize that our sample could be biased by our matching of 136 PSP firm-years with 1,752 firm-years that cover all years during the 1996–2006 period. As an alternative, we select from the S&P 500 index two no-proposal firms by industry, size, and book-to-market to match each PSP firm. We rerun the regressions in Panel A of Tables 4 and 5 based on this control sample (272 no-proposal firm-years vs. 136 PSP firm-years). Similarly, we construct an alternative control sample for the NPSP firms (102 no-proposal firm-years vs. 51 NPSP firmyears) and rerun the regressions in Panel B of Tables 4 and 5. Untabulated results show that ExcessComp is positively associated with the probability of a firm’s receiving PSPs (coefficient = 0.407, p-value = 0.009). In contrast, ExcessComp is negatively associated with the probability of a firm’s receiving NPSPs (coefficient = -0.628, pvalue = 0.054), which indicates that NPSP sponsors target firms with lower excess CEO pay. We find the increase in pay-performance sensitivity of the PSP firms is significantly higher than that of the control firms (coefficient of the PROPOSAL variable = 1.378, p-value = 0.004), which is the same as the result under the 1,752 no-proposal sample. In contrast, there is no difference in the pay-performance sensitivity change between NPSP firms and their new control firms matched by industry, size, and book-to-market (coefficient of the PROPOSAL variable = -0.666,

23 p-value = 0.207). The results above are consistent with those reported in Tables 4 and 5, which indicates that PSP sponsors are sophisticated investors, whereas NPSP sponsors are not.18

5.4. Impact of PSPs on Stock Markets Given the confirmation of the increase in pay-performance sensitivity between T-1 year and T+1 year for the PSP sample, a rational stock market might anticipate the potential benefit derived from improved manager-shareholder alignment and react positively around the proxy filing dates. Table 6 presents the cumulative abnormal stock returns of firms as associated with the proxy filing date for the proposals. Panel A is for the full sample of 136 PSP targeted firms as well as the 86 targeted firms with available bond prices. The mean (median) cumulative abnormal stock return is 1.04% (0.94%) for the full sample.19 For the smaller, 86-firm subsample with available bond trading price information, the mean (median) cumulative abnormal stock return is 1.28% (1.15%). Our abnormal stock returns are statistically significant at the 0.01 level for the full sample and the restricted sample of firms with bond trading activity. Our results suggest that stock prices of firms react positively to PSP-type proposal filings. [Insert Table 6 here]

18

In our tests on the determinants of PSPs, our independent variable of interest is managers’ excessive pay (total pay minus predicted pay), as used in Core et al. (2008). In a robustness check, we follow Ertimur et al. (2011) by also testing the effect of total pay and predicted pay on PSP filing decisions. Our untabulated results show that the probability of a firm’s receiving PSPs is positively associated with CEO total pay (p-value < 0.01) but that the probability of a firm’s receiving NPSPs is not related to CEO total pay. When splitting CEO total pay into predicted total pay and excess total pay, we find the coefficient of predicted total pay is insignificant for PSPs but significantly positive for NPSPs (p-value < 0.05). In contrast, we find the coefficient of excess total pay is significantly positive for PSPs (p-value < 0.01) but negative for NPSPs. This finding is consistent with our main findings that PSP sponsors are sophisticated investors who target excess CEO pay while NPSP sponsors are not. 19 Given that the mean size of the PSP firm is large, at $29.4 billion, the cumulative abnormal return results in an increase in average firm value of $350 million, which is economically significant.

24 Panel B is for the full sample of 51 NPSP firms as well as the 27 NPSP firms with available bond prices. The mean (median) cumulative abnormal stock return is 0.58% (-0.14%) for the full sample. For the smaller, 27-firm subsample with available bond trading price information, the mean (median) cumulative abnormal stock return is 1.30% (1.12%). The abnormal returns are statistically insignificant for the full sample and marginally positive at 0.10 level for the restricted sample with bond trading activity. Next, we examine whether the abnormal stock returns are different between firms that receive PSPs versus NPSPs. We do not find a significant difference between stock returns between the two samples. Our (untabulated) results imply that either the market is unable to distinguish between firms that receive PSPs and NPSPs or that our weak result is due to our small sample size.

5.5. Impact of PSPs on Bond Markets The TRACE database provides treasury-adjusted bond returns on 86 firms that receive PSPs (689 individual bonds) and 27 firms that receive NPSPs (219 individual bonds).20 Because our stock return data are based on firm return, we test the robustness of treasury-adjusted bond returns using the firm-level (Bessembinder et al., 2009) and representative bond-level approaches (Dhillon and Johnson, 1994). Firm-level bond return is the average bond return weighted by the issuing value of each bond for the same firm. 21 A representative bond is the bond with the

20

For the 86 PSP firms in the bond sample, the average (median) abnormal bond return is -28 basis points (-25 basis points), significant at the 0.01 level for both the t-test and the z-test, based on individual bonds. The average amount of outstanding bonds is about $772 million per bond and $6.2 billion per firm (based on a separate calculation of the authors and on information from TRACE. Here, “outstanding bonds” are bonds with valid trading information from TRACE), and the average economic loss per firm for the 86 firms’ bondholders is about $15 million in the seven days surrounding the proxy filings of PSPs (with a total market value loss of $1.3 billion for the 86 PSP firms combined). Therefore, the negative impact on bondholder wealth is not only statistically significant but also economically significant. 21 Value-weighted averaging is better than equal-weighted averaging, according to the methodological study by

25 highest average number of daily trades within 300 days prior to day zero. If two bonds of the same firm have the same average daily trades, we pick the bond with the highest number of trading days. Firm-level and representative bond results are reported in Table 7. Panel A of Table 7 shows that the means and medians of abnormal returns are both significantly negative at firm-level (p-value < 0.01) and representative bond-level (p-value = 0.02). At the firm level, the mean (median) abnormal bond return is -37 (-22) basis points.22 The representative bond result is slightly weaker, and the mean (median) abnormal bond return is -35 (-19) basis points. Our bond results suggest that the bond market reacted negatively to shareholder PSP proposals filings, possibly due to anticipated future pay-performance sensitivity improvement between T-1 year and T+1 year for the firms that receive PSPs, as documented in section 5.3. 23 In untabulated robustness tests, we separate the 86 PSP firms into two groups based on the median leverage. The first group of firms is termed the “high-leverage group,” and the second, the “low-leverage group.” We find the high-leverage group had a mean abnormal firmlevel bond return of -54.3 bps, while the low-leverage group had a mean abnormal bond return of only -4.9 bps. The two means are significantly different (p-value < 0.02). We perform the same mean test with the representative bond approach (Bessembinder et al., 2009). The p-value is 0.04 under the representative bond approach. We also perform median tests under both bond approaches and reach the same conclusion. In sum, our results are largely consistent with the notion that higher leveraged firms’ bondholders will suffer more from the news with respect to a

Bessembinder et al. (2009). 22 We present both parametric and non-parametric test results, as suggested by Bessembinder et al. (2009) and Klein and Zur (2011). We place more emphasis on the results based on average abnormal bond return, as Bessembinder et al. (2009) specifically point out that “mean returns more accurately reflect the aggregate experience of investors.” 23 Our bond return losses are similar to a well-cited bond event study with the largest bond sample size ever in published papers. Billett et al. (2004) find that during the merger and acquisition announcement period, acquirer firm bonds experience negative mean and median abnormal bond returns of -17 and -3 basis points.

26 further increase in equity-based pay-performance sensitivity. This result could be informative to legislators as well as market regulators. Panel B in Table 7 illustrates the treasury-adjusted bond returns for 27 firms that receive NPSPs with available bond prices from TRACE. The results show that, at the firm (representative bond) level, the average abnormal bond return is positive 30 basis points (19 basis points), but the median bond return is negative for both cases. In all cases, however, bond returns are not significant at any conventional level. [Insert Table 7 here]

Finally, we compare the bond return between PSP and NPSP firms for the same seven days around the proposal date. Our untabulated results show that the average abnormal bond returns are significantly lower for PSP firms at firm-level, compared to firms that receive NPSPs, using mean and median tests at p-values of 0.04 and 0.10, respectively. We do not find, however, a significant difference in means or medians for the representative bond. In general, we find weak support that bondholders treat PSPs and NPSPs differently.

5.6. Determinant of the Shareholder-Bondholder Conflict In the second stage, we report exploratory tests of potential factors that would explain our observed evidence of positive shareholder reaction to PSPs and negative bondholder reaction. One of the consequences of an incentive alignment of shareholders and managers may be an expected increase in risk-taking behavior by managers. While shareholders benefit from risktaking, as they conceptually own an option on the firm’s assets, bondholders usually suffer negative effects. We examine stock return volatility associated with PSP filing dates to determine

27 whether the market provides evidence of a shift in firm risk during this period. Bondholders are not directly subject to equity volatility. In fact, bond value depends on total firm risk (asset), as opposed to the riskiness of the residual equity claim. Unfortunately, total firm risk is not observable, and, while stock return volatility is not a perfect proxy for total firm risk, it should be correlated to the theoretical construct and is widely used in the accounting and finance literature. In Table 8, we show the results of our computation of stock return volatility in the 30 days before and after the filings of PSPs. To refute the competing explanation that return volatilities for our PSP target firms are mainly driven by increasing overall market volatility, we present market-adjusted return volatility for two time periods (-30,-1) and (0, +29) and for all PSP targeted firms (136 firms) as well as the subset of 86 firms with available bond trading information. The statistically significant increase in market-adjusted return volatility after PSP proposal filings presented in Panel A further demonstrates that our result is not driven by the overall volatility change in the equity market. In Panel B, we show that market-adjusted stock return volatility for firms that receive NPSPs is unchanged between the pre- and post-proposal period. We also retest the market-adjusted stock return volatility for a 60-day window before and after the filing of PSPs. The (untabulated) results are qualitatively the same under the alternative window length. [Insert Table 8 here]

In conclusion, the volatility result is consistent with our prediction that PSPs could induce managers to become more risk seeking in future actions and that the stock market anticipates

28 their changes in behavior and signals it through rising return volatility.24 Finally, the increase in firm risk hurts bond prices.

5.7. Additional Robustness Tests As a robustness check for the measure of pay-performance sensitivity, we also use the change in the proportion of equity compensation as an alternative proxy of change in payperformance sensitivity. We rerun the analyses in Table 5 using change in equity compensation proportion as the new dependent variable and still get a positive and significant coefficient for PROPOSAL (coefficient = 0.086, p-value < 0.01). We also find that the proportion of equity compensation is higher for firms with greater growth opportunities, consistent with the literature (e.g., Mehran 1995). An alternative way to select control firms is through matching based on industry, size, and some key factors to rule out their impact. We repeat the tests in Table 5 by replacing the control firms with a one-to-one matched set of control firms based on the same industry and of similar size and prior differences in the pay-performance sensitivity in the pre-proposal period. The results are similar. We also find that the stock (bond) returns test and the stock return volatility test results are qualitatively similar with respect to one-to-one matching control firms. To the extent that the one-to-one matching method does not control for other potential factors than the three factors (industry, size, and equity compensation), we also construct a control group of firms matched by their propensity scores. First, we run a logistic regression using all observations and all variables in Table 4. Then, for each sample firm, we select two control firms with the least difference in propensity scores. Next, we compare the sample group 24

Based on our untabulated additional test, the negative bond return around PSP proposal filings cannot be attributed to changes in expected future cash flows (proxied by changes in analyst earnings forecast around PSP filing dates).

29 and control group to identify any factor with statistically significant differences between the two groups. Finally, when controlling for all remaining factors that are still different between the two groups, we compare the two groups on the pay-performance sensitivity, stock returns, bond returns, and the change in stock return volatility. The results remain qualitatively similar. Next, to ensure that our stock and bond return results are not driven by confounding effects, we rerun tests for the values in Tables 6 and 7 after eliminating proposals that have potentially confounding events (e.g., quarterly 10-Q filing, earnings announcements, merger/acquisition announcements) within three days around the proxy statement filing dates. The results are qualitatively similar. Finally, as seen in Sections 5.4 and 5.5, all the bond and stock tests are univariate by nature. To refute the competing explanation that changes in other firm characteristics, such as profitability, cash balance, discretionary spending, and leverage, are true drivers of our documented results, we follow Klein and Zur (2011) and conduct a multivariate regression to control for changes in other firm characteristics. The results are reported in Table 9. Our sample includes all the 86 firms that have available data to calculate the abnormal bond return. Similar to the results of Klein and Zur (2011), we also find that abnormal bond return is significantly and negatively associated with abnormal stock return (p-value = 0.019). In sum, the results are consistent with the notion that shareholders benefit from emerging pay-performance activism, but potentially at the expense of bondholders. [Insert Table 9 here]

30

6. SUMMARY AND CONCLUSIONS

Boards of directors are concerned about CEO compensation and firm performance. Our findings show that one special type of shareholder proposal, a performance-focused proposal (PSP), but not a non-performance-focused shareholder proposal (NPSP), has the potential to increase targeted firms’ equity-based (mostly stock option-based) incentive compensation, which aligns shareholder-manager interests potentially at the expense of bondholders. The results suggest that shareholders are sufficiently sophisticated to be able to identify firms that pay excess compensation to their managers when filing PSPs. Boards react to these economically supported proposals, which results in significant increases to CEO payperformance sensitivity following the receipt of a PSP. While shareholders react positively, more importantly, bondholders react negatively surrounding the proxy filing dates of these proposals. We find no significant change in expected future cash flows but document a significant increase in total firm risk following the proposals, which is consistent with the differential reaction of shareholders and bondholders. The implications of the findings are three-fold. First, we show that not all compensation related proposals are the same. We find that only PSP (but not NPSP) firms have excessive CEO pay. Unlike NPSP firms, PSP firms tend to increase the pay-performance sensitivity after being targeted. This evidence supports the notion that activist shareholders who focus on firm performance appear to understand compensation disclosure and the lack of incentive alignment between managers and shareholders. The results also corroborate the Wall Street Journal’s claim that “executive-pay activism” has been turned “into a potent mainstream force” (Lublin and Dvorak, 2007), and is consistent with research that demonstrates that shareholder concerns and

31 threats have a significant impact on corporate governance and performance (e.g. Del Guercio et al., 2008). Further, our study is the first that explores the impact of shareholder’s executive compensation proposals on the wealth of bondholders. We find significantly positive stock returns and negative bond returns for PSP firms around the period of the proxy filing that announces the proposal. We point out a probable conflict of mandates for the SEC when it passes regulation that favors only the interests of equity investors, possibly at the expense of bond investors. Finally, our approach of delineating PSPs and NPSPs sheds new light on shareholder proposal activism and compensation regulation. Collectively, the results of this study and prior studies (e.g. Cai and Walkling, 2011) suggest that the pay-performance activism has an influence on boards and top management, while non-performance-focused proposals may not. Our results may have implications for shareholder activism and compensation regulation. The study results are subject to several limitations. First, this paper is the first to show bond return to shareholder proposals, and as such, we suggest caution in making inferences beyond our specific setting. Second, our sample is small due to the lack of corporate bond trading data and the nature of our manual data collection of performance-focused shareholder proposals. Third, the performance-focused shareholder proposals are only in the context of management compensation. Future research on the determinants and the effects of these proposals is warranted in other contexts, such as shareholder proposals on director election or nomination.

32 REFERENCES Agrawal, A., Mandelker, G., 1987. Managerial incentives and corporate investment and financing decisions. J. Finance 42, 823–837. Ahmed, A., Billings, B., Morton, R., Harris, M., 2002. The role of accounting conservatism in mitigating bondholder-shareholder conflicts over dividend policy and in reducing debt costs. Account. Rev. 77, 867–890. Baker, G. P., Hall, B. J., 2004. CEO incentives and firm size. J. Labor Econ. 22, 767–798. Bebchuk, L. A., Fried, J. M., 2005. Pay without performance: overview of the issues. J. Applied Corp. Finance 17, 8–23. Bebchuk, L. A., Spamann, H., 2009. Regulating bankers’ pay. Working paper, Harvard Law School. Bessembinder, H., Kahle, K. M., Maxwell, W. F., Xu, D., 2009. Measuring abnormal bond performance. Rev. Financ. Stud. 22, 4219–4258. Billett, M., King, D., Mauer, D., 2004. Bondholders’ wealth effects in mergers and acquisitions: new evidence from the 1980s and 1990s. J. Finance 59, 107–135. Bizjak, J. M., Marquette, C. J., 1998. Are shareholder proposals all bark and no bite? Evidence from shareholder resolutions to rescind. J. Financ. Quant. Anal. 33, 499–521. Blackwell, D. W., Noland, T. R., Winters, D. B., 1998. The value of auditor assurance: evidence from loan pricing. J. Account. Res. 36, 57–70. Blinder, A., 2009. Crazy compensation and the crisis. The Wall Street Journal (May 28). Brill, J. M., 2004. The new compensation committee responsibilities: a roadmap for meeting the new standards and avoiding personal liability. The Corporate Counsel 29, 1-10.

33 Cai, J., Walkling, R. A., 2011. Shareholders’ say on pay: Does it create value? J. Financ. Quant. Anal. 46, 299–339. CalPERS. 2010. Global principles of accountable corporate governance. Carter, M. E., Lynch, L. J., Tuna, I., 2007. The role of accounting in the design of CEO equity compensation. Account. Rev. 82, 327–3357. Carter, M. E., Zamora, V., 2009. Shareholder remuneration votes and CEO compensation design. Working paper, Boston College. Cheng, Q., Farber, D., 2008. Earnings restatements, changes in CEO compensation, and firm performance. Account. Rev. 83, 1217–31250. Choi, S., 2000. Proxy issue proposals: impact of the 1992 SEC proxy reforms. J. Law Econ. Org. 16, 233–268. Core, J., Guay, W., 1999. The use of equity grants to manage optimal equity incentive levels. J. Account. Econ. 28, 151–3184. Core, J., Guay, W., Larcker, D. F., 2008. The power of the pen and executive compensation. J. Financ. Econ. 88, 1–25. Core, J., Guay, W., Thomas, R. S., 2005. Is U.S. CEO compensation broken? J. Applied Corp. Finance 17, 97–104. Core, J., Holthausen, R. W., Larcker, D. F., 1999. Corporate governance, chief executive officer compensation, and firm performance. J. Financ. Econ. 51, 371–406. Core, J., Larcker, D. F., 2002. Performance consequences of mandatory increases in executive stock ownership. J. Financ. Econ. 64, 317–340. Cremers, K. J. M., Nair, V. B., 2005. Governance mechanisms and equity prices. J. Finance 60, 2859–2894.

34 Crystal, G., 1992. In search of excess: the overcompensation of American executive. W. W. Norton & Company, New York. DeFusco, R. A., Johnson, R. R., Zorn, T. S., 1990. The effect of executive stock option plans on stockholders and bondholders. J. Finance 45, 617–627. Del Guercio, D., Hawkins, J., 1999. The motivation and impact of pension fund activism. J. Financ. Econ. 52, 293–340. Del Guercio, D., Seery, L., Woidtke, T., 2008. Do boards pay attention when institutional investor activists ‘‘just vote no’’? J. Financ. Econ. 90, 84–103. DeYoung, R., Peng, E. Yan, M., 2009. Executive compensation and the business policy choices at U.S. Commercial Banks. Working paper, University of Kansas. Dhillon, U. S., Johnson, H., 1994. The effect of dividend changes on stock and bond prices. J. Finance 49, 281–289. Dittmar, A., Mahrt-Smith, J., 2007. Corporate governance and the value of cash holdings. J. Financ. Econ. 83, 599–634. Duchin, R., 2010. Cash holdings and corporate diversification. J. Finance 65, 955–992. Ertimur, Y., Ferri, F., Muslu, V., 2011. Shareholder activism and CEO pay. Rev. Financ. Stud. 24, 535–592. Fama, E. F., French, K. R., 1993. Common risk factors in the returns on stocks and bonds. J. Financ. Econ. 33, 3–56. Fama, E. F., French, K. R., 1996. Multifactor explanations of asset pricing anomalies. J. Finance 51, 55–84. Ferri, F., Maber, D. A., 2012. Say on pay votes and CEO compensation: evidence from the UK. Rev. Finance 17, 527–563.

35 Ferri, F., Sandino, T., 2009. The impact of shareholder activism on financial reporting and compensation: the case of employee stock option expensing. Account. Rev. 84, 433–466. Gammeltoft, N., 2010. Blankfein emulating Buffett as No. 48 in Wall Street CEO pay. Bloomberg News (May 26). Georgeson Inc. 2009. Annual corporate governance review. Gompers, P., Ishii, J., Metrick, A., 2003. Corporate governance and equity prices. Quar. J. Econ. 118, 107–155. Hand, J. R., Holthausen, R.W., Leftwich, R. W., 1992. The effect of bond rating agency announcements on bond and stock prices. J. Finance 47 (2), 733–752. Hartzell, J. C., Starks, L. T., 2003. Institutional investors and executive compensation. J. Finance 58, 2351–2374. Holmstrom, B., 1979. Moral hazard and observability. Bell J. Econ. 10, 74–91. Institutional Shareholder Services (ISS), 2007. US proxy voting guidelines summary. Jensen, M., Meckling, W., 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. J. Financ. Econ. 3, 305–360. Jensen, M., Murphy, K. J., 1990. Performance pay and top-management incentives. J. Polit. Econ. 98, 225–264. John, K., Klein, A., 1995. Shareholder proposals and corporate governance. Working paper, New York University (FIN-98-046). Johnson, M. F., Porter, S., Shackell, M. B., 1997. Stakeholder pressure and the structure of executive compensation. Working paper, University of Michigan. Johnson, M. F., Shackell, M. B., 1997. Shareholder proposals on executive compensation. Working paper, University of Michigan.

36 Karpoff, J. M., Malatesta, P. H., Walkling, R. A., 1996. Corporate governance and shareholder initiatives: empirical evidence. J. Financ. Econ. 42, 365–395. Klein, A., Zur, E., 2011. The impact of hedge fund activism on the target firm’s existing bondholders. Rev. Financ. Stud. 24, 1735–1771. Klock, M., Maxwell, W., Mansi, S., 2005. Does corporate governance matter to bondholders? J. Financ. Quant. Anal. 40, 693–719. Lublin, J. S., Dvorak, P., 2007. The insiders: How five new players aid movement to limit CEO pay—mainstream figures mix with activists; a fix “from within.” The Wall Street Journal (March 13). Maug, E., Rydqvist, K., 2009. Do shareholders vote strategically? Voting behavior, proposal screening, and majority rules. Rev. Finance 13, 47–79. Mehran, H., 1995. Executive compensation structure, ownership, and firm performance. J. Financ. Econ. 38, 163–184. Merton, R. C., 1974. On the pricing of corporate debt: the risk structure of interest rates. J. Finance 29, 449–470. Ortiz-Molina, H., 2007. Executive compensation and capital structure: the effects of convertible debt and straight debt on CEO pay. J. Account. Econ. 43, 69–93. Pittman, J. A., Fortin, S., 2004. Auditor choice and the cost of debt capital for newly public firms. J. Account. Econ. 37, 113–136. Rappaport, A., Nodine, T. H., 1999. New thinking on how to link executive pay with performance. Harvard Business Review. 77, 91–101. Renneboog, L., Szilagyi, P. G., 2010. Shareholder activism through the proxy process. ECGIFinance Working Paper No. 275.

37 Shaw, K., Zhang, M., 2010. Is CEO cash compensation punished for poor firm performance? Account. Rev. 85, 1065–1093. Smith, C. W., Watts, R. L., 1992. The investment opportunity set and corporate financing, dividend, and compensation policies. J. Financ. Econ. 32, 263–292. Solomon, D., Paletta, D., 2009. U.S. eyes bank pay overhaul: administration in early talks on ways to curb compensation across finance. The Wall Street Journal (May 13). Strickland, D., Wiles, K. W., Zenner, M., 1996. A requiem for the USA: Is small shareholder monitoring effective? J. Financ. Econ. 40, 319–338. Thomas, R. S., Martin, K. J., 1999. The effect of shareholder proposals on executive compensation. University of Cincinnati Law Rev. 67, 1021–1081. TIAA-CREF. 2010. Policy statement on corporate governance. Watts, R. L., Zimmerman, J. L., 1986. Positive accounting theory. Prentice-Hall, Englewood Cliffs, NJ. Yun, H., 2009. The choice of corporate liquidity and corporate governance. Rev. Financ. Stud. 22, 1447–1475.

38

70 60 PSP

50

NPSP

40 30 20 10 0 1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Figure 1. Shareholder executive-pay proposals (1996–2006). The number of shareholder proposals for S&P 500 companies targeted by at least one performance-focused executive-pay shareholder proposal over the period of 1996–2006. NPSP = non-performance-focused proposals; PSP = performance-focused proposals.

39 Table 1 Industry distribution. Compustat n % 945 4.7

13

Oil and Gas Extraction

20

Food and Kindred Products

4

2.9

4

7.8

8

3.1

352

1.8

28

Chemicals & Allied Products

11

8.1

6

11.8

21

8.0

1,294

6.5

29

Petroleum Refining & Related Industries

3

2.2

1

2.0

4

1.5

102

0.5

35

Industrial, Commercial Machinery, Computer Equipment

9

6.6

3

5.9

16

6.1

883

4.4

36

Electronic And Other Electrical Equipment And Components, Except Computer Equipment

7

5.1

3

5.9

23

8.8

1,125

5.6

37

Transportation Equipment

7

5.1

1

2.0

5

1.9

302

1.5

38

Measuring, Analyzing, & Controlling Instruments; Photographic, Medical & Optical Goods; Watches & Clocks

4

2.9

2

3.9

14

5.3

923

4.6

40

Railroad Transportation

3

2.2

0

0.0

1

0.4

38

0.2

48

Communications

7

5.1

1

2.0

6

2.3

840

4.2

49

Electric, Gas, & Sanitary Services

21

15.4

2

3.9

13

5.0

654

3.3

60

Depository Institutions

7

5.1

4

7.8

18

6.9

1,797

9.0

63

Insurance Carriers

4

2.9

4

7.8

11

4.2

463

2.3

73

Business Services

10

7.4

2

3.9

20

7.6

2,389

12.0

All Others

35

25.7

18

25.4

89

29.0

7,863

39.4

100

51

100

262

100

19,970

SIC

Industry Description

Total

136

NPSP n % 0 0.0

NoProposal n % 13 5.0

PSP n % 4 2.9

100

This table shows the industry distribution of PSP firms, NPSP firms, no-proposal control firms, and all Compustat firms. The sample consists of 136 firms that received PSPs over the 1996–2006 period. There are 51 NPSP firms. The no-proposal control firms are all of the remaining S&P 500 firms (262 firms) that did not receive any shareholder executive-pay proposals during the same period. The other firms from the remaining S&P 500 list are deleted due to insufficient data.

40 Table 2 Summary characteristics.

Total Compensation ($K) Salary ($K) Bonus ($K) Stock options ($K) ROA Market Capitalization ($MM) Leverage Shares Outstanding (MM) CEO Stock ownership n

(1) PSP Firms 11,975 1,100 2,258 4,730 4.60%

Mean (2) NPSP Firms 11,532 1,142 2,727 4,716 5.2%

(3) Control Firms 7,211 898 1,234 3,460 5.30%

29,366

31,266

11,191

0.64 780 1.45% 136

0.64 826 2.04% 51

0.6 273 2.16% 1,752

1–2

2–3 1–3

(4) PSP Firms 9,212 1,080 1,632 3,004 3.50%

Median (5) NPSP Firms 7,279 1,033 1,305 2,717 3.1%

(6) Control Firms 5,206 902 898 1,966 5.00%

***

***

***

***

***

***

*

***

***

***

14,395

12,704

6,358

**

0.66 391 0.53% 136

0.62 335 0.83% 51

0.6 160 0.98% 1,752

***

*** **

4–5

5–6 4–6 ***

***

**

***

***

*** **

**

***

** ***

*** **

This table presents the descriptive statistics for the 136 sample firm-years with at least one PSP, 51 firms with only NPSPs, and 1,752 firmyears from the 262 control firms that did not receive any shareholder executive-pay proposals over the 1996–2006 time period. Total Compensation is the sum of the CEO’s salary, bonus, stock option grants, restricted stock, and other compensation. ROA is the net income before extraordinary items and discontinued operations divided by total assets. Market Capitalization is the closing price for the fiscal year times the number of common shares outstanding. Leverage is calculated as total liabilities divided by total assets. Shares Outstanding is the number of common shares outstanding as reported by the company. CEO Stock Ownership is the percentage of total shares outstanding, including stock options held by the CEO. Means and medians are provided. t-test (z-test) is for the differences in means (medians), with significance at the 0.01, 0.05, and 0.10 levels, given by ***, **, and *, respectively.

41 Table 3 OLS regression to determine excess CEO compensation. Independent

Estimate (p-value)

Intercept

4.486 (<.001)

Log(Sales)

0.380 (<.001)

Market-to-book

0.028 (<.001)

ROA

0.004 (0.095)

Stock returns

0.002 (<.001)

ROA volatility

0.005 (0.233)

Stock volatility

0.684 (<.001)

Leverage

-0.183 (0.035)

Adjusted R2

0.361

N

5,188

Control of year and industry effects

Yes

This table presents regression results from the estimation of Equation (1). The sample consists of 5,188 firm-year observations for all the S&P 500 firms in the period of study, 1996–2006. We compute all the explanatory variables at or for the period ending at year t. Log(TotalComp) is the dependent variable and is the natural logarithm of the CEO total compensation. Log(Sales) is the natural logarithm of Sales. Market-to-book is defined as the market value of equity divided by the book value of equity. ROA is the net income before extraordinary items and discontinued operations divided by total assets. Stock_returns is the total stock return in the last year less the median stock return in the same period and same industry (two-digit SIC code). ROA volatility is the standard deviation of ROA over five years. Stock volatility is the standard deviation of daily stock returns over 6 months. Leverage is the total liabilities divided by the total assets.

42 Table 4 PSPs and excess compensation. Panel A: PSP firms vs. no-proposal firms. (1)

(2)

Independent

Predicted sign

Estimate (p-value)

ΔPredicted. Prob.

Estimate (p-value)

ΔPredicted. Prob.

ExcessComp

+

0.552

73.68%

0.550

73.30%

(<.001) InternalGov

+/ –

0.005

(<.001) 0.55%

(0.624) ExternalGov

+/ –

0.141

+

0.542

15.19%

+

0.943 (<.001)

CEO Stock Ownership

+/ –

0.144

15.53%

(0.002) 71.91%

(0.012) FirmSize

0.48%

(0.666)

(0.002) Poor_Indicator

0.005

0.534

70.53%

(0.013) 156.80%

0.951

158.82%

(<.001) -0.032

-3.15%

(0.523) CEO Age

+/ –

-0.012

-1.17%

(0.483) Log Likelihood 2

Pseudo R N

-335.98

-335.46

31.26%

31.36%

1,888

1,888

This panel presents the logistic regression results of Equation (2), where the dependent variable is equal to 1 if a firm receives a PSP and 0 for all firm-years of the S&P 500 companies that did not receive any shareholder executive-pay proposals during our sample period. There are 136 sample firms, and 1752 control firm years. ExcessComp is the prior year residual value obtained from the executive-pay benchmark model in Equation (1). InternalGov is the percentage of shares held by large institutional blockholders, whose ownership is higher than 5% of a firm’s outstanding shares. ExternalGov is a Gompers et al.’s (2003) G index, multiplied by -1. Poor_Indicator is 1 if the fiveyear industry-adjusted stock return is negative and 0 otherwise. FirmSize is the logarithmic transformation of total assets. CEO Stock Ownership is the percentage of total shares outstanding held by the CEO, excluding options. CEO Age is the historical CEO age. ΔPredicted Prob. is the change in the predicted probability that occurs when the independent variable increases one unit in its value and is evaluated at the mean values of the remaining independent variables. The p-values (in parentheses) are based on maximum likelihood standard errors. The industry and year effects are controlled. Coefficients on the intercept, year and industry indicators are not shown, and p-values are based on one-tailed tests for hypothesized variables and two-tailed otherwise.

43 Table 4 (continued) Panel B: NPSP firms vs. no-proposal firms. Independent

Predicted sign

Estimate (p-value)

ΔPredicted. Prob.

ExcessComp

+

-0.090 (0.700)

-8.63%

InternalGov

+/ –

-0.023 (0.182)

-2.24%

ExternalGov

+/ –

0.075 (0.257)

7.81%

Poor_Indicator

+

1.034 (0.003)

181.20%

FirmSize

+

0.606 (<.001)

83.29%

CEO Stock Ownership

+/ –

-0.033 (0.560)

-3.23%

CEO Age

+/ –

-0.002 (0.935)

-0.21%

Log Likelihood

-172.34

Pseudo R2

25.75%

N

1,803

This panel presents the logistic regression results of Equation (2), except for the dependent variable, where the dependent variable here is equal to 1 if a firm receives an NPSP and 0 for all firm-years of the S&P 500 companies that did not receive any shareholder executive-pay proposals during our sample period. There are 51 NPSP firms and 1,752 control firm-years. ExcessComp is the prior year residual value obtained from the executive-pay benchmark model in Equation (1). InternalGov is the percentage of shares held by large institutional blockholders, whose ownership is higher than 5% of a firm’s outstanding shares. ExternalGov is a Gompers et al.’s (2003) G index, multiplied by -1. Poor_Indicator is 1 if the five-year industry-adjusted stock return is negative and 0 otherwise. FirmSize is the log-transformed total assets. CEO Stock Ownership is the percentage of total shares outstanding held by the CEO, excluding options. CEO Age is the historical CEO age. ΔPredicted Prob. is the change in the predicted probability that occurs when the independent variable increases one unit in its value and is evaluated at the mean values of the remaining independent variables. The p-values (in parentheses) are based on maximum likelihood standard errors. The industry and year effects are controlled. Coefficients on the intercept, year and industry indicators are not shown, and p-values are based on one-tailed tests for hypothesized variables and two-tailed otherwise.

44

Table 4 (continued) Panel C: PSP firms vs. NPSP firms. Independent

Predicted sign

Estimate (p-value)

ΔPredicted. Prob.

ExcessComp

+

0.484 (0.053)

62.18%

InternalGov

+/ –

0.026 (0.053)

2.60%

ExternalGov

+/ –

0.032 (0.766)

3.21%

Poor_Indicator

+

-0.157 (0.753)

-14.53%

FirmSize

+

0.510 (0.070)

66.58%

CEO Stock Ownership

+/ –

-0.017 (0.827)

-1.69%

CEO Age

+/ –

-0.037 (0.326)

-3.63%

Log Likelihood

-74.01

Pseudo R2

32.45%

N

187

This panel presents the logistic regression results of Equation (2), except for the dependent variable, where the dependent variable is equal to 1 (0) if a firm receives a PSP (NPSP) in the S&P 500 companies during our sample period. There are 136 PSP firms and 51 NPSP firms. ExcessComp is the prior year residual value obtained from the executive-pay benchmark model in Equation (1). InternalGov is percentage of shares held by large institutional blockholders, whose ownership is higher than 5% of a firm’s outstanding shares. ExternalGov is a Gompers et al.’s (2003) G index, multiplied by -1. Poor_Indicator is 1 if the five-year industry-adjusted stock return is negative and 0 otherwise. FirmSize is the log-transformed total assets. CEO Stock Ownership is the percentage of total shares outstanding held by the CEO, excluding options. CEO Age is the historical CEO age. ΔPredicted Prob. is the change in the predicted probability that occurs when the independent variable increases one unit in its value and is evaluated at the mean values of the remaining independent variables. The p-values (in parentheses) are based on maximum likelihood standard errors. The industry and year effects are controlled. Coefficients on the intercept, year and industry indicators are not shown, and p-values are based on one-tailed tests for hypothesized variables and two-tailed otherwise.

45

Table 5 Impact of PSP on CEO annual equity-based compensation incentives. Panel A: PSP firms vs. no-proposal firms. Independent

Predicted Sign

PROPOSAL

+

1.332 (<0.001)

Δ(ShareholderWealtht)

+/ –

22.377 (0.264)

Δ(ShareholderWealtht-1)

+/ –

-5.994 (0.731)

MarketCap

+/ –

-6.356 (0.345)

TobinQ

+/ –

0.236 (0.009)

NEWCEO

?

1.029 (<.001)

ExcessComp

–

-2.292 (<.001)

InternalGov

?

-0.001 (0.947)

ExternalGov

?

-0.006 (0.882)

2

Estimate (p-value)

Adjusted R

0.145

N

1,888

This table presents the OLS regression results of Equation (3). There are 136 sample firms and 1,752 control firm-years. The dependent variable is the change from Year T-1 to Year T+1 in the logarithmic transformation of the dollar change in annual stock options and restricted stock grants for a 1% change in the stock price. PROPOSAL is 1 if a firm received at least one PSP and 0 for control firms. Δ(ShareholderWealth) is the change in shareholder wealth for the two-year period prior to Year 0. MarketCap is the market value of equity. TobinQ is the market value of equity plus the book value of debt and divided by the book value of assets. We measure these three variables at the beginning of the proposal year. NEWCEO is 1 if the CEO changed either in the proposal year or the following year and 0 otherwise. ExcessComp is the prior year residual value obtained from the executive-pay benchmark model in Equation (1). InternalGov is the sum of the institutional blockholder ownership that is higher than 5% of the firm’s outstanding shares. ExternalGov is a Gompers, Ishii, and Metrick (2003)’s G index, multiplied by -1. The industry and year effects are controlled. Coefficients for intercepts, and year and industry indicators are not shown, and p-values are based on one-tailed tests for hypothesized variables and two-tailed otherwise.

46 Table 5 (continued) Panel B: PSP firms vs. NPSP firms Independent

Predicted Sign

PSP

+

Estimate (p-value) 1.409 (<.001)

NPSP

?

0.525 (0.391)

Δ(ShareholderWealtht)

+/ –

19.178 (0.190)

Δ(ShareholderWealtht-1)

+/ –

-11.746 (0.394)

MarketCap

+/ –

-5.212 (0.330)

TobinQ

+/ –

0.224 (0.010)

NEWCEO

?

1.049 (<.001)

ExcessComp

–

-2.301 (<.001)

InternalGov

?

-0.001 (0.901)

ExternalGov

?

-0.006 (0.892)

Adjusted R2

14.8%

N

1,939

This table presents the OLS regression results of Equation (6). There are 136 PSP firms, 51 NPSP firms, and 1,752 other firm-years. The dependent variable is the change from Year T-1 to Year T+1 in the logarithmic transformation of the dollar change in annual stock options and restricted stock grants for a 1% change in the stock price. The variable PSP (NPSP) is equal to 1 if a firm receives a PSP (NPSP) proposal and 0 for either NPSP (PSP) firms or other firms that did not receive any shareholder pay proposals during the sample period. Δ(ShareholderWealth) is the change in shareholder wealth for the two-year period prior to Year 0. MarketCap is the market value of equity. TobinQ is the market value of equity plus the book value of debt and divided by the book value of assets. We measure these three variables at the beginning of the proposal year. NEWCEO is 1 if the CEO changed either in the proposal year or the following year and 0 otherwise. ExcessComp is the prior year residual value obtained from the executive-pay benchmark model in Equation (1). InternalGov is the sum of the institutional blockholder ownership that is higher than 5% of the firm’s outstanding shares. ExternalGov is a Gompers et al.’s (2003) G index, multiplied by -1. The industry and year effects are controlled. Coefficients for intercepts, and year and industry indicators are not shown, and p-values are based on one-tailed tests for hypothesized variables and two-tailed otherwise.

47 Table 6 Stock market reactions. Panel A: PSP firms. Cumulated Abnormal Return (%) Sample

p-value

n

Mean

Median

t-tests

z-tests

Full PSP sample

136

1.04

0.94

<0.01

<0.01

Subsample with valid bond return

86

1.28

1.15

<0.01

<0.01

Panel B: NPSP firms.

Sample Full NPSP sample Subsample with valid bond return

n 51 27

Cumulated Abnormal Return (%) Mean Median 0.58 1.30

-0.14 1.12

p-value t-tests

z-tests

0.22 0.06

0.36 0.10

This table shows cumulative abnormal stock returns. The event window is between the 5th trading day prior to day zero, proposal filing date, and the first trading day after day zero. Therefore, the event window is [-5,+1]. We use the latest event date for each firm if the firm has multiple proposals in multiple years during the sample period. The cumulative abnormal stock returns are computed according to Fama and French’s (1993, 1996) three-factor model. Panel A is for the full PSP sample (136 firms) as well as its subsample of 86 firms with valid bond return information. Panel B is for the full NPSP sample (51 firms) as well as its subsample of 27 firms with valid bond return information; p-values are based on one-tailed tests.

48 Table 7 Bond market reactions. Panel A: PSP firms.

Sample

N

Basis Points (1 basis point = 0.01%) Mean Median

p-value t-tests

z-tests

Firm Level

86

-37

-22

<0.01

<0.01

Representative Bond

86

-35

-19

0.02

<0.01

Panel B: NPSP firms.

Sample

N

Basis Points (1 basis point = 0.01%) Mean Median

p-value t-tests

z-tests

Firm Level

27

30

-11

0.20

0.47

Representative Bond

27

19

-11

0.32

0.47

This table shows the treasury-adjusted abnormal bond returns. Panel A is for 689 bonds of 86 PSP receiving firms with available trading data in TRACE. Panel B is for 219 bonds of 27 NPSP receiving firms with available trading data in TRACE. The event window is between the 5th trading day prior to day zero, proposal filing date, and the first trading day after day zero. Therefore, the event window is [-5,+1]. We eliminate the observation if the bond did not trade within five trading days after day zero. We use the latest event date for each firm if the firm has multiple proposals in multiple years during the sample period. Firm level bond return is the weighted average of all bond returns for each firm weighted by the issuing amount of each bond for the same firm. A representative bond is the bond with the highest average daily trades within three hundred days prior to day zero. If two bonds of the same firm have the same average daily trades, we pick the bond with the highest number of trading days; p-values are based on one-tailed tests.

49 Table 8 Change in market-adjusted stock return volatility around PSPs. Panel A: PSP firms. Sample

N

(-30, -1)

(0, +29)

Test of Difference, p-value

Full sample

136

0.2171

0.2352

<0.01

Bond sample

86

0.1920

0.2143

<0.01

Panel B: NPSP firms. Sample

N

(-30, -1)

(0, +29)

Test of Difference, p-value

Full sample

51

0.2185

0.2226

0.38

Bond sample

27

0.1746

0.1756

0.47

The market-adjusted return is defined as the firm’s stock return minus the CRSP equal-weighted market return. Market-adjusted stock return volatility is measured during the 30 trading days before and after the proposal date. The full sample is the proposal sample. The bond sample includes proposal firms with available bond trading data in TRACE; p-values are based on one-tailed tests.

50 Table 9 PSP firms’ bond returns and stock return. Independent

Predicted Sign

Estimate

(p-value)

–

-0.098

(0.019)

+/ –

-0.011

(0.081)

–

-0.005

(0.236)

∆ROA

+/ –

0.148

(0.243)

∆OpMargin

0.006

(0.857)

∆Leverage

+/ – –

-0.048

(0.134)

∆CashBal

+

0.012

(0.615)

∆DPS

+/ –

0.005

(0.525)

∆Z-Score

+/ –

0.001

(0.920)

∆Size

+/ –

0.009

(0.594)

∆M2B

+/ –

+0.000

(0.860)

∆σ(ε)

?

0.123

(0.572)

∆σ (Unlevered Returns)

?

-0.227

(0.215)

EarningsAnn

?

0.008

(0.475)

∆Capex

?

-0.021 42.4%

(0.859)

Abnormal Stock Returns SubDebt SpecuGrade

Adjusted R2 N

86

This table presents the regression results of PSP bond returns on stock returns. There are 86 sample firms. The dependent variable is the [-5,+1] daily abnormal bond return. Abnormal Stock Returns is the [-5,+1] cumulative abnormal stock return. SubDebt is 1 if the target firm has subordinated debt and 0 otherwise. SpecuGrade is 1 if the firm-level bond rating prior to proposal date is BB+ or lower and 0 otherwise. ROA is EBITDA scaled by total assets. OpMargin is EBITDA scaled by sales. Leverage is total long- and short-term debt scaled by total assets. CashBal is cash and short-term investments scaled by total assets. DPS is the dividend per share. Z-Score is calculated based on the formulae in Altman (1968) to model bankruptcy probability. Size is the natural log of total assets. M2B is market value of equity divided by book equity. σ(ε) is the standard deviation of the market model residual. σ(Unlevered Returns) is the standard deviation of unlevered stock returns. EarningsAnn is 1 if the target firm announced earnings in the event window and 0 otherwise. Capex is capital expenditures less property sale divided by total assets. Prefix ∆ indicates a change variable. The industry and year effects are controlled. Coefficients for intercepts, and year and industry indicators are not shown; p-values are based on one-tailed tests for hypothesized variables and twotailed otherwise.

51 APPENDIX Samples of Shareholder Proposals on Executive Compensation 1. Performance-focused Proposals: Knight Ridder Inc., 1999 “RESOLVED: That the shareholders of Knight Ridder Inc. urge the Board of Directors to adopt an executive compensation policy that all future stock options grants shall be performance-based. For purposes of this resolution, performance-based stock options are defined as either: (1) stock options with the exercise price indexed to an appropriate S&P 500 peer group index (such as the index used in the Company's annual proxy statement); or (2) premium-priced stock options, which set the exercise price of the option above the current market value of the stock.”

2. Performance-focused Proposal: Raytheon, 2002 “RESOLVED, shareholders urge the Board of Directors to adopt a formal policy that a majority or all future stock option grants to senior executives be performance-based on core business operating results. Consistent with this topic, the amount of company pension income is to be subtracted from the financial results that are used to determine future stock option grants, and pension income is to be reported annually on the primary company web site for verification. Performance-based stock options are defined as: 1) Indexed options, whose exercise price is linked to the S&P Aerospace Index shown in the graphs on pages 26 and 27 in the 2002 proxy; 2) Premium-priced stock options, whose exercise price is above the market price of the grant date; or 3) Performance vesting options, which vest when the market price of the stock exceeds a specific target.”

3. Non-performance-focused Proposal: NCR Corporation, 2005 “Proposal: Management and Directors are requested to consider discontinuing all domestic partner benefits for highly paid executives making over $500,000 per year or, if not feasible, ask these executives to reimburse the company for these expenses.”

4. Non-performance-focused Proposal: General Electric, 1999 “Resolved: Shareholders request that a committee of outside directors of the board institute an Executive Compensation Review, and prepare a report available to shareholders four months after our annual meeting with the results of the Review and recommended changes. The review shall cover all pay, benefits, perks, stock options and special arrangements in the compensation packages for all company's top officers. Special attention should be made to consider GE's potential environmental liabilities with executive compensation.”