Company Stock in Defined Contribution Plans: Evidence from Proxy Voting

• Date: 01 March 2017
• DOI: http://doi.org/10.1111/fima.12147
• Published date: 01 March 2017
• Author: Heejin Park

Company Stock in Defined Contribution

Plans: Evidence from Proxy Voting

Heejin Park∗

I examine whether firms’ decisions to offer company stock in defined contribution (DC) plans

are explained by managers’ corporate control motives. Using a large sample of proxy voting

outcomes, I find that employee ownership in DC plans is significantly and positively associated

with the level of voting support for management-sponsoredproposals. This suggests that managers

encourage employee DC holdings in company stock in order to receive higher voting support in

favor of management. The effects of employee ownership on voting outcomes are significantly

greater in specific subsamples where employee vote is moreimportant to management.

From the perspectiveof employees, holding a large share of their defined contribution (DC) plan

invested in the stock of their employer is inefficient because it exposes them to an unnecessary

amount of diversifiable risk. Further, company stock in DC plans ties the value of the employee’s

human capital to the value of his/her retirement savings. This can lead to a situation where the

employee’s labor income and retirement savings fall simultaneously, exacerbating the personal

effects of negative economic shocks.

However, a nontrivial number of firms offer participants company stock as an investment option

in their DC plans. Moreover, they often offer employer matching with company stock even if

participants choose investment options other than the company’s stock. Holden and VanDerhei

(2003) show that among plans that offer employer stock in DC plans, on average, 38% of DC

assets are in their own company stock as of year-end 2002. Using 11-K filings during the period

from 1991 to 2000, Brown, Liang, and Weisbenner (2006) show that 39.3% of their sample firms

offering company stock in DC plans match with employer stock for employer contributions.

Given that employers’ decisions to include company stock in DC plans lead to unnecessary

risk for their employees, it is natural to question the rationale behind this decision. To date, only a

few studies have attempted to investigate the motivations behind this practice. Forexample, Rauh

(2006) argues that managers offer employees company stock in DC plans as a takeover defense

by showing that the changes in Delaware case law in the mid-1990s—which made takeovers

more difficult for outside shareholders to successfully complete—lowered employee ownership.

Rauh’s (2006) study suggests that employer stock is a substitute takeover defense. Brown et al.

(2006) investigate why companies decide to offer matches in company stock and find that fir ms

that have lower stock price volatility, lower bankruptcy risk, and a defined benefit (DB) plan are

more likely to provide company stock matches.

Motivated byRauh (2006), this paper explores whether f irms’ decisions to offer companystock

in their employees’ DC plans can be explained by managers’ corporate control motives. Since

the market for takeover activityhas not been ver y active since the 1990s, and instead shareholder

Special thanks go to my adviser,Diane Del Guercio, for reading countless drafts and providing invaluable support and

advice. I also thank an anonymous referee, Andrea Anthony, Julian Atanassov, Sudipto Dasgupta, Wei Jiang, Xiaoding

Liu, Keun Jae Park, Raghavendra Rau (Editor), Hai Tran, Jason Turkiela, and Wesley Wilson forhelpful comments and

suggestions.

∗Heejin Parkis with the Korea Workers’Compensation & Welfare Service in Seoul, South Korea.

Financial Management •Spring 2017 •pages 155 – 202

156 Financial Management rSpring 2017

activism through the proxy process has been more pervasive, I conduct tests in the setting of

annual proxy voting. In particular, I hypothesize that managers provide company stock in order

to receive a higher level of support for management in proxy voting. This hypothesis assumes

that employee ownersvote in favor of management. This assumption is supported by the fact that

voting rights on employee ownership in DC plans are largely delegated to plan trustees. Further,

plan trustees have voting authority for all unallocated shares as well as those allocated shares

when plan participants do not provide anydirection on voting. Because plan trustees are appointed

by management, the trustees are more likely to support management rather than act in the best

interests of their participants (Chaplinsky and Niehaus, 1990; Gordon and Pound, 1990; Chang

and Mayers, 1992). Even if employee owners participate in voting, because they earn wages that

are fixed claims on the firm’s cash flow (Jensen and Meckling, 1979; Faleye, Mehrotra, and

Morck, 2006), they may vote in favor of management and not in favor of shareholders. This is

because from the employee’s perspective, the present value of expected wages and benefits is

greater than the present value of an employee’s equity stake.

Using a panel of 10,093 US firms with 72,560 management-sponsored proposals and

4,436 shareholder-sponsored proposals over the 2003–2012 period, I find a positive associa-

tion between employeeownership in DC plans and voting support for management.1Specif ically,

firms with higher employee ownership receive higher levels of voting support for management-

sponsored proposals compared to firms with lower employee ownership. In management-

sponsored proposals, the ordinary least squares (OLS) results show that a 1-standard deviation

increase in percent of employee ownership in a firm’s equity market value raises “for” votes by

0.1 percentage points. Even though a small increase in votes for management may have little

impact on whether the proposal passes or fails, this is meaningful given the evidence that a

lower level of votes leads to changes in board, management, and other governance-related is-

sues (Cai, Garner, and Walkling, 2009; Fischer et al., 2009; Yermack, 2010; Ertimur, Ferri, and

Oesch, 2013). The results are robust after controlling for other factors affecting voting outcomes,

such as firms’ f inancial strength, managerial ownership,Institutional Shareholder Ser vices (ISS)

recommendation, voting mechanism, and proxy variables for corporate governance.

I conduct several additional subsample tests. First, I separately examine cases consisting of

management-sponsored proposals opposed by ISS. The previous literature shows that ISS is

the most influential proxy adviser, and votes can be swayed up to 20% due to the influence

of ISS recommendations (Bethel and Gillan, 2002; Cai et al., 2009; Choi, Fisch, and Kahan,

2010). Hence, this is likely to be a situation when the support of a block percentage of employee

ownership can be most useful to management. Simply put, employee voting rights can be more

usefully called for when the level of voting support for management is expected to be lower

due to the influence of a proxy adviser. The nextsubsample test consists of only close votes where

the levels of voting support are either just aboveor below 50%. The rationale for examining close

votes is that if management has information about the likely outcome of voting at a time when

it can do something to change the outcome, managers would encourage trustees and employees

to support their proposals for a successful vote.2Thus, given that the marginal effects of voting

support for management are greater for proposals that have around 50% of voting support, I

1In most cases, management tends to vote for management-sponsored proposals but vote against shareholder-sponsored

proposals. In my sample, more than 99.9% of management proposals are recommended by management, while 0.03%

of shareholder proposals are recommended by management. Therefore, voting support in favor of management can be

interpreted as having higher levels of voting support for management-sponsored proposal but lower levels of voting

support for shareholder-sponsored proposals.

2Listokin (2008) finds that there is a large difference between the frequency of management-sponsored proposals passing

with votes just above 50% and the frequency of proposals failing with votes just below 50%.In this study, close votes

Park rCompany Stock in Defined Contribution Plans 157

expect that the results will be stronger in close votes.3Last, I conduct subsample tests separately

for director election and say on pay frequency proposals. Specifically, compensation-related

proposals are closely related to managerial entrenchment because they directly influence the

economic welfare of management. The results show that the effect of employee ownership in

voting outcomes is significantly greater for director election with more than 20% withheld and

for say on pay frequency proposals.

Overall, I find that the effects of employee ownership are much larger for proposals in these

subsamples. For example, in say on pay frequency proposals, an increase in the percent of

employee ownership in firms’ equity market value by 1 standard deviation is associated with an

increase in the proportion of “for” votes of 0.97 percentage points. Further, once it is combined

with managerial ownership, the effect is even stronger. Specifically, increases in the percent of

employee ownership by 1 standard deviation and percent of managerial ownership by 1 standard

deviation raise the proportion of “for” votes by 3.86 percentage points. Considering that Cai

et al. (2009) find that a 1% decrease in the average votes for compensation committee members

associates with reductions in abnormal chief executive officer (CEO) compensation by $143,000

in the following year, the effects of employee ownership are economically meaningful.

My paper relates to several strands of literature. First, I complement Rauh’s (2006) f indings

by providing evidence that corporate control motives are still important in an era when hostile

takeovers are not significant. In particular, the findings of this paper indicate that managers offer

company stock in DC plans to their employees in order to receive higher voting support for

management in spite of unnecessary risk to their employees. As such, my findings contribute to

understanding the purpose of employee ownership in pension plans from the view of employers

(Mitchell and Utkus, 2002; Brown et al., 2006).

Second, my findings contribute to a growing literature on shareholder voting. A large body of

research has focused on the influence of proxy advisers’ recommendations on voting outcomes

(Bethel and Gillan, 2002; Cai et al., 2009; Alexander et al., 2010; Choi et al., 2010) or voting

behavior of institutional investors(Gillan and Starks, 2000; Parrino, Sias, and Starks, 2003; Davis

and Kim, 2007; Cotter, Palmiter, and Thomas, 2011; Morgan et al., 2011; Ashraf, Jayaraman,

and Ryan, 2012). In particular, this paper sheds light on a limited body of literature concerning

methods management may use to try to have influence on voting outcomes (Ferri and Oesch,

2014).

Finally, my paper also contributes to a growing area of research focused on DC plans. As the size

and importance of DC plans increase, it is evident that employees must become more responsible

for their ownretirement income security. However,an extensiveliterature on behavioral economics

suggests that plan participants struggle to make optimal investment decisions due to choice and

information overload (Benartzi and Thaler, 2001; Iyengar, Huberman, and Jiang, 2004; Iyengar

and Kamenica, 2006), undefined preferences and investment menu design (Benartzi and Thaler,

2001, 2002; Mitchell and Utkus, 2002; Elton, Gruber, and Blake, 2006), and procrastination and

inertia (Benartzi and Thaler, 2002; Choi, Madrian, and Metrick, 2002). Managers’ decisions to

offer employees company stock in DC plans in order to increase managerial power may lead

are defined as the levels of voting support between 30% and 70%. As shown in Figure A1 in the Appendix, there is a

discontinuity in the distribution of voting outcomes in management-sponsored votes at the 50% level.

3This is because with a small shift, management can win the proposal as the marginal effect in close votes of voting

shift from employees and trustees is large. However, management will withdraw or alter the proposal if managers predict

that the proposal will fail with far below 50% support. Furthermore, management will not act if managers predict that

the proposal will win with far above 50% support. Due to a limited sample size, here, close votes are defined somewhat

broadly as situations where the levels of voting support are between 30% and 70% and the proposals need greater than

50% support to pass.