ARTICLES
The Link Between Job Satisfaction and Firm Value,
With Implications for Corporate
Social Responsibility
by Alex Edmans
Executive Overview
How are job satisfaction and firm value linked? I tackle this long-standing management question using a
new methodology from finance. I study the effect on firm-level value, rather than employee-level
productivity, to take into account the cost of increasing job satisfaction. To address reverse causality, I
measure firm value by using future stock returns, controlling for risk, firm characteristics, industry
performance, and outliers. Companies listed in the “100 Best Companies to Work For in America”
generated 2.3% to 3.8% higher stock returns per year than their peers from 1984 through 2011. These
results have three main implications. First, consistent with human resource management theories, job
satisfaction is beneficial for firm value. Second, corporate social responsibility can improve stock returns.
Third, the stock market does not fully value intangible assets, and so it may be necessary to shield managers
from short-term stock prices to encourage long-run growth.
T
he link between job satisfaction and firm value
is of great interest to both academics and prac-
titioners. It has implications for how managers
approach employee recruitment, retention, and mo-
tivation, and more generally speaks to the impor-
tance of human resource management (HRM) for a
firm’s overall business strategy. Indeed, Landy
(1989) described this relationship as the “Holy
Grail” of organizational behavior. But there is still
much debate on whether these variables are actually
related in practice. This paper aims to tackle this old
question—a predominantly management topic—
with a new methodology from the finance literature.
The Link in Theory
The theoretical benefits of job satisfaction are
reasonably clear. HRM theories (e.g., Becker &
Gerhart, 1996; Likert, 1961; McGregor, 1960;
Pfeffer, 1994) suggest several channels through
which job satisfaction can improve firm value.
One is recruiting and retaining key employees.
The resource-based view (RBV) of the firm (Bar-
ney, 1986; Wernerfelt, 1984) argues that firms
develop sustainable competitive advantage by
building resources that are both valuable and hard
for competitors to poach. HRM theories assert
that employees are indeed valuable. They are the
key source of value creation in the modern firm,
particularly in knowledge-based industries such as
pharmaceuticals and software. However, satisfying
the latter requirement—being hard to poach—is
difficult because talented employees can walk out
of the door. This is where the role of job satisfac-
tion lies. A satisfying workplace can foster job
Alex Edmans (aedmans@wharton.upenn.edu) is an Assistant Professor of Finance at The Wharton School of the University of Pennsyl-
vania, a Faculty Research Fellow at the National Bureau of Economic Research, and a Research Associate at the European Corporate
Governance Institute.
2012 1Edmans
Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express written permission.
Users may print, download, or email articles for individual use only. http://dx.doi.org/10.5465/amp.2012.0046
Bio
rich5/zol-aome/zol-aome/zol00412/zol3034d12z
xppws S⫽1 12/12/12 13:24 Art:
embeddedness (Mitchell, Holtom, Lee, Sablynski,
& Erez, 2001) and ensure that talented employees
stay with the firm. Relatedly, job satisfaction can
provide a valuable recruitment tool.
A second channel through which job satisfac-
tion can improve firm value is through worker
motivation. Under Taylorism (Taylor, 1911), mo-
tivation could be achieved by extrinsic factors
such as piece pay or the threat of firing. This
approach worked because employees’ output was
easy to measure—for example, the number of wid-
gets produced. Nowadays, workers’ tasks, such as
building client relationships and mentoring sub-
ordinates, are much harder to quantify. Thus, pro-
viding incentives based on output may be ineffec-
tive or even destructive (Kohn, 1993). Since
extrinsic motivators are less effective, this in-
creases the role for intrinsic motivators such as
satisfaction. Organ’s (1977) social exchange
model argues that an employee views pleasant
working conditions as a “gift” from the firm, and
responds with a “gift” of increased effort— even if
effort is not directly rewarded by an incentive
scheme. Job satisfaction can promote organiza-
tional citizenship behavior (Organ, 1988), where
an employee goes above and beyond the formal
requirements of the job and internalizes the firm’s
objective function as his or her own.
The Link in Practice
While the benefits of job satisfaction are clear in
theory, empirically documenting them in prac-
tice is much harder. Most studies investigate the
correlation between job satisfaction and individ-
ual job performance (rather than firm value),
measuring satisfaction using a survey. An early
meta-analysis by Brayfield and Crockett (1955)
concluded that there was “minimal or no relation-
ship” between the two; Vroom (1964) and Locke
(1965) later showed similar findings. While these
conclusions may have been driven by the relative
lack of studies in the 1950s and 1960s, Iaffaldano
and Muchinsky (1985) meta-analyzed 217 corre-
lations from 74 studies and found a mean correla-
tion of only 0.17. They concluded that job satis-
faction and job performance are “only slightly
related to each other,” dubbing the relationship
an “illusory correlation” and a “management fad.”
This meta-analysis proved highly influential and
pervaded management thinking for the rest of the
millennium. In particular, it was widely cited as
evidence that there is no link between satisfaction
and performance.
Recent studies have uncovered more positive
results. An updated meta-analysis by Judge,
Thoresen, Bono, and Patton (2001) found a mean
correlation of 0.30. Studying the related concept
of happiness/positive affect, Barsade and Gibson
(2007), Fisher (2010), and Lyubomirsky, King,
and Diener (2005) found a positive link with
individual success. While these studies are at the
individual level, the study of Ostroff (1992) and
the meta-analyses of Harter, Schmidt, and Hayes
(2002) and Harter, Schmidt, Asplund, Killham,
and Agrawal (2010) argued that the organization
is the appropriate unit of analysis as it takes into
account interactions between workers. All three
found stronger positive correlations than individ-
ual-level analyses. While the above analyses in-
vestigated performance, Allen, Bryant, and Var-
daman (2010), Griffeth, Hom, and Gaertner
(2000), and Lee, Mitchell, Holtom, McDaniel,
and Hill (1999) showed a negative link between
satisfaction and employee turnover, providing ev-
idence for the retention benefits of job satisfaction
discussed earlier.
A separate but related literature studies the link
between firm performance and high-performance
work practices (Huselid, 1995) or high-perfor-
mance workplace systems (HPWS) (Becker &
Huselid, 1998). Job satisfaction is an output mea-
sure—an attitudinal variable that is not directly
controlled by management but indirectly affected
through changing HRM policies. In contrast, an
HPWS is an input measure—a set of practices that
strategic HRM theorists consider performance en-
hancing, such as incentive compensation and em-
ployee participation. These practices are con-
trolled by management and are an input into job
satisfaction; indeed, job satisfaction is a potential
channel through which HPWSs can improve firm
performance.
Huselid (1995) showed that HRM practices are
correlated with lower turnover and, on a per-
employee basis, higher sales, market value, and
profits. Perceived organizational performance is
2 NovemberAcademy of Management Perspectives
rich5/zol-aome/zol-aome/zol00412/zol3034d12z
xppws S⫽1 12/12/12 13:24 Art:
positively linked to HRM practices (Delaney &
Huselid, 1996) and work-family programs (Perry-
Smith & Blum, 2000); Konrad and Mangel
(2000) found a positive association between work-
life programs and firm productivity. More broadly,
the review paper by Pfeffer and Veiga (1999)
detailed the various theoretical benefits of
HPWSs and cited supporting evidence. Combs,
Liu, Hall, and Ketchen (2006) meta-analyzed 92
studies and found an overall correlation between
HPWSs and firm performance of 0.2.
Difficulties in Identifying an Effect
The positive correlations reported by recent re-
search cannot be automatically interpreted as ev-
idence that job satisfaction improves firm value.
First, positive correlation could result from reverse
causality from performance to satisfaction, since
most of the above studies are cross-sectional. It
could be that performance directly causes satisfac-
tion (employees are happier working for a success-
ful firm), or that strong performance allows man-
agers to increase wages and benefits, which in turn
raises satisfaction. It is tempting to try to address
reverse causality by undertaking longitudinal stud-
ies that lag the satisfaction measure (e.g., by re-
lating satisfaction in 2001 to performance in
2002) and arguing that satisfaction must have
caused performance because it preceded it. But it
may be that satisfaction in 2001 was a result of
strong performance in 2001, and this strong per-
formance in 2001 also led to strong performance
in 2002 because performance is persistent. In
short, the correlation between satisfaction in 2001
and performance in 2002 could arise because a
third omitted variable (performance in 2001)
drives both.
1
Second, the above studies typically focus on job
performance as the dependent variable, but this is
critically different from firm value for three rea-
sons. First, if job performance is measured at the
individual level, it is unclear how this aggregates
at the firm level. It could be that there is a positive
correlation between individual satisfaction and
individual performance on average, but this is
driven by low-level workers and the correlation is
negative for senior managers. Since the latter are
more important for firm value, the overall link to
firm performance could be negative. Second, even
considering organizational performance, there are
several possible dimensions—such as productivity,
absenteeism, and turnover—and it is unclear how
to weight them. For example, if satisfaction is
correlated with higher productivity but also
higher turnover, it is unclear whether satisfaction
is beneficial overall. Similarly, it is unlikely that a
study will be able to analyze every single dimen-
sion of performance that is relevant to firm value;
it may omit some important ones.
Third, some performance measures do not take
into account the costs of achieving higher job
satisfaction. Even if job satisfaction improves job
performance, it may still reduce firm value net of
costs. Cappelli and Neumark (2001) found that
“high-performance” practices, which delegate re-
sponsibility to employees, increase labor costs.
Thus, even though they also found weak evidence
of improved productivity, there is no effect on
overall labor efficiency (output per dollar cost).
Moreover, the cost of increasing job satisfaction
may include not only direct financial expenses,
but also increased risk. Firms with high job satis-
faction are often human capital–intensive, mak-
ing them vulnerable to market downturns because
human capital has little value in bankruptcy.
While a profit measure (such as earnings) takes
financial expenses into account, it does not con-
sider risk; indeed, there is no accepted way to
adjust accounting variables for risk.
While most papers study firm performance, a
few do analyze firm value. Abowd (1989) showed
that announcements of pay increases reduce mar-
ket valuations dollar for dollar. Diltz (1995) found
that stock returns are uncorrelated with the Coun-
cil on Economic Priorities (CEP) minority man-
agement and women in management variables,
and are negatively correlated with family benefits.
Dhrymes (1998) found no relationship with
KLD’s employee relations variable, and Gorton
and Schmid (2004) showed that greater employee
involvement reduces firm value. These results are
1
Studying the change in performance between 2001 and 2002 is not a
solution either, as changes in performance are usually correlated over time.
An improvement in performance between 2001 and 2002 could be caused
by an improvement in performance between 2000 and 2001 (which also
leads to satisfaction in 2001).
2012 3Edmans
rich5/zol-aome/zol-aome/zol00412/zol3034d12z
xppws S⫽1 12/12/12 13:24 Art:
consistent with the view of job satisfaction held by
the economics and finance literatures: Companies
maximize their profits by ensuring that employees
are paid no more than their outside option, and so
providing them with “excess” satisfaction is costly
to shareholders. Practitioners also emphasize the
costs of creating a satisfying workplace: Tsao
(2003) in BusinessWeek and Zimmerman (2004)
in the Wall Street Journal discussed investors’ con-
cerns that Costco offers excessively generous pay
and benefits to employees, compared to its frugal
competitor Wal-Mart. The title of Zimmerman’s
article, “Costco’s Dilemma: Be Kind to Its Work-
ers, or Wall Street,” captures the zero-sum view
that job satisfaction is necessarily at the expense
of firm value.
This Paper’s Approach
T
his paper’s methodology aims to address both
of the above concerns: reverse causality, and
the need to measure firm value rather than
firm performance. It uses a measure of firm value:
the stock market value. In an efficient market, the
market value eventually takes into account all
publicly available information that is relevant for
firm value. How important a variable is can be
measured by its dollar effect on the stock market
value (Fama, 1970). The market value takes into
account all the channels through which job satis-
faction affects firm value, weighting them accord-
ing to their relative importance, and also consid-
ers costs. For example, assume that job satisfaction
leads to high employee turnover, but higher turn-
over reduces firm value by only $5 million (since
the high turnover is only for low-level workers).
Assume also that job satisfaction leads to high
productivity among senior managers, which in-
creases firm value by $50 million, and that the
cost of the job satisfaction program is $20 million.
Eventually, the overall increase in market value
will be $25 million.
However, simply studying the firm’s current
market value runs into two issues. First, it is still
subject to reverse causality—a high current mar-
ket value could lead to high satisfaction. Thus, I
instead study future stock returns— by relating
satisfaction in December 2001 to stock returns in
2002, for example. The stock return in 2002 is the
change in the market value between December
2001 and December 2002. Notably, unlike other
performance measures, stock returns are not per-
sistent: The change in market value between 2001
and 2002 is unrelated to the change in market
value between 2000 and 2001 in an efficient mar-
ket.
2
If satisfaction in December 2001 was caused
by strong performance in 2001, the market value
would already be high in December 2001 and so
we should not expect high returns in 2002.
Second, the market may not incorporate the
full benefits of job satisfaction immediately. While
it should incorporate tangible information such as
performance immediately, it may take time to
incorporate intangible information such as job
satisfaction, as it is difficult to quantify its impact
on firm value. Thus, a firm’s satisfaction in De-
cember 2001 may not be fully incorporated into its
value by December 2001, but may affect the mar-
ket value only when it leads to future tangible
outcomes that are directly valued. If job satisfac-
tion in December 2001 leads to employees being
productive and the firm announcing high earnings
in June 2002, the stock price will increase in June
2002. The company’s stock return between De-
cember 2001 and December 2002 will capture this
earnings announcement. Importantly, the use of
stock returns captures all possible channels, not
just earnings, through which job satisfaction may
affect firm value. Other channels include a new
product launch, positive customer satisfaction rat-
ings announced in the press, positive Wall Street
analyst reports, the signing of a new contract, or
the filing of a patent.
3
In sum, studying future
stock returns allows for the possibility that the
market takes time to recognize the benefits of job
satisfaction.
A further benefit is that using stock returns lets
us control for risk. While stock returns are rarely
used in the management literature, they are fre-
2
If the market is inefficient, stock returns may be persistent due to
“momentum” effects. As described later, I control for momentum.
3
Of course, it is difficult for the stock market to estimate the impact of
some of these outcomes on overall value precisely. However, the use of
market values does not assume that the market estimates the value of every
outcome precisely—only that it forms an unbiased estimate that is equally
likely to be an overestimate as an underestimate. If a new product launch
leads to the market value rising by $100 million, this implies that the true
value of the product is just as likely to be $110 million as $90 million.
4 NovemberAcademy of Management Perspectives
rich5/zol-aome/zol-aome/zol00412/zol3034d12z
xppws S⫽1 12/12/12 13:24 Art:
quently studied by the finance literature. This
literature has developed well-established empiri-
cal techniques—in particular, to control for risk—
from which this paper draws. In contrast, there is
no accepted way to adjust other measures of firm
performance for risk.
The Best Companies List
I
measure job satisfaction using the list of the
“100 Best Companies to Work for in America.”
This list was first published in a book in March
1984 (Levering, Moskowitz, & Katz, 1984) and
updated in February 1993 (Levering & Moskowitz,
1993). Since 1998, it has been published in the
January issue of Fortune magazine each year. The
list has been overseen by Robert Levering and
Milt Moskowitz throughout its 28-year existence,
and is compiled by the Great Place to Work
®
Institute in San Francisco.
4
A firm’s ranking in the Best Companies list
comes from two sources. Two-thirds of the score
comes from employee responses to the Trust In-
dex
©
employee survey. The survey questions were
developed through an extensive process that in-
volved a review of academic literature; interviews
with managers, employees, and workplace experts;
focus group sessions; and discussions with manage-
ment consultants, survey design experts, and re-
searchers. This process created 120 statements,
which were narrowed down to 57 following ex-
tensive testing with groups of employees, post-
survey interviews, and cluster and factor analysis.
The survey was then beta-tested in a variety of
workplace settings to ensure that each survey
statement was measuring correctly.
The final 57 statements span five categories:
credibility, respect, fairness, pride, and camarade-
rie. The first three concern workers’ trust in man-
agement, the fourth concerns workers’ perceptions
of the job, and the fifth concerns workers’ rela-
tionships with other employees. Indeed, the Insti-
tute defines a “great place to work” as one where
“employees trust the people they work for, have
pride in what they do, and enjoy the people they
work with.” Employees provide ratings on a scale
of 1 to 5 and answer two open-ended questions.
The survey statements are proprietary and not
generally made public; the Institute kindly pro-
vided them for this paper (the questions, but not
individual responses). Table 1 provides a sub-
sample of survey questions. Across all levels of
employees, 250 are randomly selected in each
firm; respondents fill in the surveys anonymously
and return their responses directly to the Institute.
The response rate is around 60%.
The remaining one-third of the score comes
from the Culture Audit
©
. This is a questionnaire
completed by management—similar to the survey
instruments used by the HPWS literature to eval-
uate a firm’s HRM practices—on the firm’s demo-
graphic makeup, pay and benefits programs, and
culture. Again, the questions are proprietary, and
the Institute kindly provided them for this paper.
They include the following topics: diversity
(proportion of women and minorities in senior
positions), turnover (voluntary, involuntary, and
retirements), compensation (average cash com-
pensation, retirement benefits, employee stock
ownership plans, stock options, profit sharing),
benefits (health care, training, on-site perks), time
off (paid vacations, sabbaticals, community in-
volvement), and work-family issues (parental
leave, child care). The audit also includes numer-
ous open-ended questions, examples of which are
given in Table 2.
The Best Companies list has several advantages
as a measure of firm-level job satisfaction. First
and most important is that we have data on stock
returns of the Best Companies for a 28-year period
that includes both recessions and booms. This
time series is substantially longer than those used
in prior literature, which typically considers one
point or, at most, a couple of years. This length
helps ensure that the results are not driven by a
specific time period or market conditions (e.g., it
may be that job satisfaction is valuable only in
recessions). Second, some previous studies focus
on individual dimensions of job satisfaction, such
as pay satisfaction or coworker satisfaction. This
4
Although the Institute was not founded until 1990, Levering and
Moskowitz used the same criteria for the 1984 list, although they surveyed
employees directly rather than through a questionnaire. The interviews
used for the 1984 list were the primary basis for the subsequent question-
naire. Indeed, when the Trust Index was first used, it replicated the results
that Levering and Moskowitz had obtained when they conducted inter-
views for their 1984 list.
2012 5Edmans
T1
T2
rich5/zol-aome/zol-aome/zol00412/zol3034d12z
xppws S⫽1 12/12/12 13:24 Art:
