1、option compensation for outside directors affects corporate bond yields in the secondarymarket. Our results show that the greater the ratio of outside directors stock and optionpioneerfromtheiruisitthe aggrpolicies. A fundamental solution to these conicts (other than equity-based compensation for ma
2、nagers) is to create a board ofJournal of Corporate Finance 14 (2008) 512531Contents lists available at ScienceDirectJournal of Corporate Financejournal homepage: to monitor managers on behalf of stockholders. However, like managers even directors may lack incentives to act in thebest interest of s
3、hareholders. To reduce the directorshareholder conicts, a growing number of rms have awarded equity-basedcompensation to outside directors over the last decade.1Recent evidence suggests that equity-based compensation helps alignincentives of outside directors and shareholders (Fich and Shivdasani, 2
4、005; Bryan et al., 2000).However, there exists little evidence on the effects of equity-based compensation for directors on the directorbondholderconicts. In this empirical paper, we examine two major hypotheses about the impact of the structure of outside directorcompensation on the agency costs of
5、 debt. Specically, our monitoring hypothesis posits that equity-based compensation forform of managerial work-shirking, perqrm.Sinceshareholdersstandto beardirectorswouldimprove board oversight andproblems such as managerial shirking, perqbondholders. On the ip side, equity-based Corresponding autho
6、r. Tel.: +1 419 530 2437; fax:E-mail addresses: mine.ertugrulutoledo.edu (M.1For example, Perry (1999) nds the percentage oincreased from 48% to 70% from 1992 to 1995.0929-1199/$ see front matter 2008 Elsevier B.V.doi:10.1016/j.jcorpn.2008.09.004e consumption, and overinvestment emerge when managers
7、 own but a fraction of theegate agencycosts,they haveanincentive toseekoptimalinvestmentandnancing1. IntroductionThe literature on agency problems,have incentives to expropriate wealthbondholders require higher returns oncompensation to total compensation, the lower the average yield spreads on the
8、rmsoutstanding bonds, withstockcompensation having alargerimpactthanoptioncompensation.Further, the effect of equity-based compensation on yield spreads is stronger for rms withlower-rated debt. 2008 Elsevier B.V. All rights reserved.ed by Jensen and Meckling (1976), points out that shareholders of
9、leveraged rmsbondholders through investing in riskier projects. Anticipating these incentives,capital resulting in higher cost of debt. Further, shareholdermanager conicts in theG34G32J33Keywords:Director incentivesCost of debtCorporate governanceAgency costsBoard compensation practices and agency c
10、osts of debtMine Ertugrula, Shantaram HegdebaUniversity of Toledo, College of Business Administration, Department of Finance, Mail Stop 103, Toledo, OH 43606-3390, United StatesbUniversity of Connecticut, School of Business, Department of Finance, 2100 Hillside Road, Unit 1041, Storrs, CT 06269-1041
11、, United Statesarticle info abstractArticle history:Received 14 September 2007Revised 10 September 2008Accepted 11 September 2008Available online 18 September 2008Extant theory and empirical evidence indicate that equity-based compensation can align theinterests of managers with those of shareholder
12、s, but it has a side effect of aggravatingbondholder-shareholder conicts by increasing managers risk-shifting incentives. Recentevidence conrms that extending equity-based compensation to outside directors also iseffective in aligning their interests with those of shareholders, but its adverse effec
13、ts on thedebt-related agency problems are unknown. In this paper, we examine how stock and stockJEL classication:reduce corporate bond yield spreads. Effective monitoring will help decrease agencyuisite consumption, and overinvestment and benet all stakeholders includingcompensation schemes (that in
14、crease the convexity of payoffs to directors) could+1 419 530 2873.Ertugrul), Shantaram.Hegdebusiness.uconn.edu (S. Hegde).f S Anderson et al., 2003) suggest that bond covenants rarely succeed in completelyeliminatingstockholderbondholder conicts, and bondholders will demand a premium for bearing th
15、e agency cost. This implies that theresidual bond yield to maturity (after controlling for other determinants such as duration, rating, liquidity, and issuercharacteristics)willimpoundapremiumforthecostofdebtarisingfromagencycosts.FollowingDuffee(1998)andAndersonetal.(2003), we measure the cost of s
16、easoned (outstanding) debt by the yield spread, which is the difference between the weightedaverage yield on the rms outstanding year-end debt obligations (excluding convertible bonds and oating rate bonds) and the516 M. Ertugrul, S. Hegde / Journal of Corporate Finance 14 (2008) 512531yield on a co
17、nstant maturity Treasury security with similar duration (published by Federal Reserve Bank of New York).3.2.2. Compensation and ownership variablesWe use measures of stock and option compensation for outside directors and their ownership as test variables, whilecontrolling for the structure of compe
18、nsation and ownership of managers. All our measures of compensation are annual. The ratioof equity-based compensation tototal compensation fordirectors (Dir. Stock and Option Compensation) is computed as the value ofstock grants plus the BlackScholes value of options, divided by total compensation.
19、Total compensation is equal to the sum ofdirector meeting fees multiplied by number of meetings, annual cash retainer, value of stocks, and BlackScholes value of optiongrants. To compute the BlackScholes value of option grants, we need information about the exercise price and time to expiration.Foll
20、owing Yermack (2004), we assume that exercise price is equal to the average stock price during the grant year and time tomaturity is 10 years. We use standard deviation of stock prices over 60 months (BS_VOLATILITY in Execucomp) and the companysaverage dividend yield for 3 years (BS_YIELD in Execuco
21、mp) in our computations. The risk free rate is set equal to the 10-yearconstant maturity Treasury bond yield as of year end. Due to these assumptions, the BlackScholes value of options for directors issubjecttoerror.Thesemeasuresaresupplementedbythepercentageoftotalcompensationgrantedintheformofstoc
22、ks(Dir.StockCompensation) and options (Dir. Option Compensation). We also calculate the percentage of stock ownership by outside directors(Dir. Stock Ownership), but we do not have data on director option ownership. The denitions of these and other variables arepresented in Appendix A.10Applyingthes
23、ameprocedures,weconstructthefollowingvariablestocontrolforthepercentageofequity-basedcompensationandequityownershipfortopveexecutivesfromproxystatements:Mng.StockandOptionCompensation,Mng.StockCompensation,Mng. Option Compensation and Mng. Stock Ownership.11To calculate option ownership, we divide t
24、he sum of the numberof optionsgranted that year and the number of options unexercised at the end of the year by the number of shares outstanding (Mng. OptionOwnership). We add the last two to create a measure of stock and option ownership (Mng. Stock and Option Ownership).Following Core and Guay (20
25、02) and Coles et al. (2006), we also include the following sensitivity variables. Dir. Sensitivity to Pricerepresents the change indirectorscompensation forone percentage pointchange in stock price.12Dir. Sensitivity to Volatility denotesthe change in directors wealth for a 0.01 change in the annual
26、ized standard deviation of stock returns.13Similarly, we calculate thesensitivities of top executives wealth to stock price (Mng. Sensitivity to Price) and volatility (Mng. Sensitivity to Volatility).3.2.3. Issue and rm characteristicsFollowingReebetal. (2001), Andersonetal.(2003), andOrtiz-Molina(2
27、006),wecontrolforissuecharacteristicsaswellasrmcharacteristicsinourmultivariate regressions.Ourcontrolsforbondissuecharacteristicscoverbondduration(measuredinyears),9Loss of observations of this magnitude is not uncommon in bond yield studies, see Anderson et al. (2003, 2004) who use a sample of 252
28、 rms. Further, sinceour analysis includes only rms with outstanding, public debt, we also run a Heckman selection model where the selection equation is whether the rm is in oursample or not. Our main results remain qualitatively the same.10Directors can be outside or inside directors. In our dataset
29、, 94% of the inside directors are among top ve executives and their compensation is included inmanagerial compensation variables.11We use the BlackScholes value of the options provided by Execucomp, which denes volatility as the standard deviation of stock price for 60 months(BS_VOLATILITY) and divi
30、dend yield as the companys average dividend yield for three years (BS_YIELD). These options values contain measurement error as theydo not incorporate unique characteristics of incentive stock options (i.e., stock options are not traded, are subject to vesting requirements, and directors andmanagers
31、 have limited ability to hedge their wealth).12Overall director wealth can also be affected by events that ruin the reputation of a director such as a corporate scandal due to ineffective monitoring. Loss ofreputation can result in loss of directorships which might signicantly affect the wealth of t
32、he director. However, it is very difcult to construct a measure ofdirector wealth including director reputation. We control for average number of directorships held by outside directors in our determinants of compensationregressions. In untabulated results, the number of directorships was not signic
33、antly correlated with yield spreads.13Since we do not have data on directors option holdings, we could only calculate the delta and vega of the options granted during the year. Therefore, Dir.Sensitivity to Price and Dir. Sensitivity to Volatility estimate the sensitivities of directors pay to chang
34、e in stock price and volatility with error. In contrast, Mng.Sensitivity to Price and Mng. Sensitivity to Volatility account for cumulative stock and option holdings of managers and hence are more reliable measures ofsensitivities of managerial wealth.Table 1Summary statisticsVariable Min Mean Media
35、n Max Standard deviationYield spread (%) 0.81 3.58 3.15 12.74 2.00Dir. stock and option compensation (%) 0.00 58.47 60.09 100.00 26.46517M. Ertugrul, S. Hegde / Journal of Corporate Finance 14 (2008) 512531credit ratings, and bond liquidity. Duration controls for differences in maturity and coupon r
36、ates, and credit ratings proxy fordefault risk. A potential problem with using bond ratings as an independent variable is that the effect of outside director equity-based compensation may have already been incorporated in the ratings. Recent evidence by Bhojraj and Sengupta (2003) andAshbaugh-Skaife
37、 et al. (2006) suggests that bond ratings are affected by corporate governance mechanisms. Moreover, Standardand Poors and Moodys report that they incorporate the strength of governance mechanisms into their ratings. To purge out theeffects of compensation, ownership, and governance mechanisms, we r
38、egress bond ratings on compensation and ownershipvariables for both managers and directors and the governance variables, and use the error term from this regression as anindependentvariable in our tests. We measure bond liquidityas the length of time that the bond has been outstanding (bond age).Pas
39、t research has documented that bond prices reect liquidity premium and more recently issued bonds are more liquid thanpreviously issued bonds (Green and Odegaard, 1997).Wealsocontrolforrmcharacteristicsinourmultivariateregressions.Largerrmsareconsideredsaferinvestments,sincetheyDir. stock compensati
40、on (%) 0.00 18.17 7.97 90.67 22.87Dir. option compensation (%) 0.00 40.25 39.74 100.00 31.81Dir. stock ownership (%) 0.00 0.94 0.11 27.08 3.53Dir. stock compensation ($000) 0.00 22.26 8.42 714.89 43.40Dir. option compensation ($000) 0.00 98.29 32.34 3,713.99 253.20Dir. sensitivity to price ($000) 1.
41、10 696.51 42.98 23,824.83 3,157.38Dir. sensitivity to volatility ($000) 0.00 8.00 3.98 74.60 12.28Mng. stock and option compensation (%) 0.00 53.23 54.36 96.93 21.84Mng. stock compensation (%) 0.00 7.99 0.00 62.26 13.40Mng. option compensation (%) 0.00 45.20 44.04 93.51 21.81Mng. stock ownership (%)
42、 0.00 2.71 0.50 37.33 6.10Mng. option ownership (%) 0.10 4.07 2.82 23.31 4.14Mng. stock and option ownership (%) 0.25 6.83 3.70 49.67 8.42Mng. sensitivity to price ($000) 1.12 2,381.20 852.27 48,241.62 5,635.96Mng. sensitivity to volatility ($000) 0.00 686.41 360.20 5,860.69 920.46Board independence
43、 0.11 0.69 0.71 1.00 0.17G-Index 3.00 10.09 10.00 16.00 2.58Board size 5.00 10.51 10.00 20.00 2.39Director tenure 0.00 8.68 8.20 24.17 3.52This table presents summary statistics for yield spreads, ownership, compensation, and governance variables. Our sample consists of 783 rm-year observationsdrawn
44、 from years 20002002. All variables are dened in Appendix A.tend to have more diversied assetsand largerasset bases.We measurermsize as the logof total assets.To control forcreditrisk,we use leverage ratio dened as long-term debt divided by total assets. Since more protable, better performing rms ar
45、e lesslikely to default, they will be characterized by lower yields. We measure protability as operating income scaled by total assets(ROA)andrmperformancewithsalesgrowth. Tocontrolfortheriskof therm,weusethestandarddeviationofROAforthepast5years.Sincehighgrowthrmshavemorerisk-shiftingincentives,weu
46、seTobinsQtocontrolforthermsgrowthopportunities.As Bhojraj and Sengupta (2003) nd a negative relation between institutional ownership and the presence of outside directorsand bond yields, we also include institutional ownership and percentage of independent directors as control variables in ourregres
47、sions. Consistent with Klock et al. (2005), we use Gompers et al. (2003) Index (G-Index) to control for shareholder rights.Following Anderson et al. (2004), we include board size and director tenure to account for board characteristics in our regressions.It is possible that outside directors who are
48、 afliated with nancial institutions might have different inuence on the cost of debt.To explore this possibility, we identify institutional investors, venturecapitalists, and investment bankers in ourlistof directorsbymatching the names of the companies that they work for to a list of institutional
49、investors from 13F lings and to a list of venturecapital and investment banks from SDC. We include the percentage of institutional investors, venture capitalists, and investmentbankers serving on the board as control variables in our regressions. Finally, we include year dummies and FamaFrench (1997)industry dummies to control for industry differences.143.2.4. Summary statisticsTable 1 presents descriptive statistics for yield spreads, compensation, ownership, an
