does the stock market fully value intangibles? employee...
TRANSCRIPT
Electronic copy available at: http://ssrn.com/abstract=985735
Does the Stock Market Fully Value Intangibles?
Employee Satisfaction and Equity Prices�
Alex Edmans
Wharton School, University of Pennsylvania
June 23, 2008
Abstract
This paper analyzes the relationship between employee satisfaction and long-run stock
performance. A portfolio of the �100 Best Companies to Work For in America�earned an
annual four-factor alpha of 4% from 1984-2005. The portfolio also outperformed industry-
and characteristics-matched benchmarks, and the results are robust to the removal of
outliers and other methodological changes. Returns are even more signi�cant in the 1998-
2005 sub-period, even though the list was widely publicized by Fortune magazine. These
�ndings have three main implications. First, employee satisfaction is positively correlated
with shareholder returns and need not represent excessive non-pecuniary compensation.
Second, the stock market does not fully value intangibles, even when independently veri�ed
by a publicly available survey. This suggests that intangible investment generally may not
be incorporated into short-term prices, underpinning managerial myopia theories. Third,
certain socially responsible investing (�SRI�) screens may improve investment returns.
Keywords: Employee satisfaction, market e¢ ciency, short-termism, managerial myopia,
human capital, human resource management, socially responsible investing
JEL Classification: G14, J28, M14
�2428 Steinberg Hall-Dietrich Hall, 3620 Locust Walk, Philadelphia, PA 19104-6367. (215) 746 0498. I amgrateful to Henrik Cronqvist, Xavier Gabaix, David Hirshleifer, Tim Johnson, Moza¤ar Khan, Lloyd Kurtz,Andrew Metrick, Milt Moskowitz, Stew Myers, Mahesh Pritamani, and seminar participants at the 2008 Ox-ford Finance Symposium, 2007 Conference in Financial Economics and Accounting at NYU, the 2007 SociallyResponsible Investing annual conference, MIT Sloan, Rockefeller, UBS, Virginia Tech, and Wharton for valuedinput. Special thanks to Amy Lyman of the Great Place To Work Institute for answering numerous questionsabout the Fortune survey, and to Patrick Sim for research assistance.
1
Electronic copy available at: http://ssrn.com/abstract=985735
�[Costco�s] management is focused on ... employees to the detriment of shareholders. To
me, why would I want to buy a stock like that?��Equity analyst, quoted in BusinessWeek
1 Introduction
This paper analyzes the relationship between employee satisfaction and long-run stock price per-
formance. An portfolio of the �100 Best Companies to Work For in America�earned a Carhart
(1997) four-factor alpha of 0.34% per month from 1984-20051, or 4% per year. These �gures
exclude any event-study reaction to list inclusion and only capture long-run drift. Returns re-
main signi�cant when calculated over industry- and characteristics-matched benchmarks, when
equal- or value-weighting, when adjusting for outliers, and when controlling for a large number
of other characteristics known to a¤ect returns. The outperformance is even stronger from 1998,
even though the list was published in Fortune magazine and thus highly visible to investors.
These �ndings contribute to three strands of research. The �rst is the increasing importance
of human capital in the modern corporation. The second is the equity market�s failure to fully
incorporate the value of intangible assets, which underpins managerial myopia theories. The
third is the e¤ect of socially responsible investing (�SRI�) screens on investment performance.
The traditional �rm that pervaded throughout much of the 20th century was predominantly
capital intensive and focused on generating cost e¢ ciency through scarce physical assets. Em-
ployees were seen as merely a cost to be minimized, rather than a source of positive value
creation (see, e.g., Taylor (1911)). Management practices therefore centered around extracting
maximum e¤ort from workers, while minimizing their compensation.
Management philosophies have dramatically changed over the past century. The current
competitive environment places a signi�cantly greater emphasis on quality and innovation,
for which human, rather than physical capital, is particularly important (Zingales (2000)).
Accordingly, the human relations movement (e.g. Maslow (1943), Hertzberg (1959), McGregor
(1960)) has rapidly grown in in�uence. It recognizes employees as key organizational assets,
rather than expendable commodities, and focuses on achieving competitive advantage through
recruiting, developing and retaining a superior workforce.
Despite the intuitive logic of the human relations movement, there is little existing evidence
that �rms with satis�ed employees perform more strongly. Supporting the traditional view
that improvements in worker welfare must come at the expense of shareholder returns, Abowd
(1989) �nds that higher pay reduces equity values dollar-for-dollar. This void provides the
�rst motivation for this paper. To my knowledge, it constitutes the �rst study showing that
employee satisfaction is positively associated with shareholder returns, rather than representing
ine¢ ciently excessive non-pecuniary compensation.
The second goal of the paper is to study the market�s valuation of employee satisfaction.
Even if CEOs believe that human capital investment improves long-run value, they may still
1Throughout this paper, year t refers to the returns for the Best Companies list published in that year.Since the list is published part-way through each year, the return period ends the following year. For example,1984-2005 returns are calculated from April 1984 through January 2006.
2
underinvest. This problem has been formalized by a number of managerial myopia models, such
as Narayanan (1985) and Stein (1988, 1989). The fundamental problem is that such investment
is intangible, and so its only immediately observable e¤ect is reduced earnings. Since low pro�ts
may stem from poor �rm quality, the stock price rationally falls. Fearing such a decline, an
equity-aligned manager may ine¢ ciently forgo investment in the �rst place.
Despite the widespread belief that managerial myopia is a signi�cant issue (see, e.g., Porter
(1992) and the survey by Graham, Harvey and Rajgopal (2005)), there is limited empirical
evidence either documenting or contradicting its existence in reality. This paper sheds light
on the real-life importance of short-termism by studying the key assumption underlying my-
opia theories: that intangible investment cannot be credibly communicated to outsiders. This
explains my focus on long-horizon returns, and my choice of the �Best Companies�list as the
measure of employee satisfaction. This list was published in Fortune from 1998 and thus be-
came highly observable and widely known, yet the results are even stronger for this subperiod.
Finding positive event-study reactions to Fortune inclusion would not imply that the market
underestimates the bene�ts of employee satisfaction (since it immediately incorporates the news
upon release), nor would �nding superior returns based on a proprietary measure unavailable
to most investors.
By delaying portfolio formation until the month after the publication of the list, I give
the market ample opportunity to react to its content. However, I still �nd signi�cant outper-
formance of the Best Companies portfolio. That even highly visible, independently veri�ed
measures of intangibles are not fully incorporated into stock prices implies that intangibles in
general are undervalued by the stock market �the vast majority of which have no equivalent
method of public certi�cation such as the Best Companies list. This in turn supports the key
assumption that underlies managerial myopia concerns, and suggests that managers evaluated
according to the short-term stock price may indeed have signi�cant disincentives to invest for
long-run growth. Note, however, that the underreaction to the list does not imply market irra-
tionality. Given that theory provides no clear predictions on the e¤ect of employee satisfaction,
the market may have simply been unaware of the bene�ts to shareholders.
The third implication of the study relates to the pro�tability of SRI strategies, whereby
investors only select companies that act in a manner that they de�ne as socially responsible.
Traditional portfolio theory (e.g. Markowitz (1959)) suggests that SRI reduces returns, since
it restricts an investor�s choice set. Indeed, many existing studies �nd that SRI screens have a
negative, or at best zero, e¤ect on returns. This paper suggest that SRI screens may improve
investment performance �at least when the screen focuses on employee welfare. Rather than
leading to the undesirable exclusion of potentially good investments, certain SRI screens may
focus the choice set on particularly attractive securities. This is because a �rm�s concern for
other stakeholders may be ultimately bene�cial to shareholders, yet not be priced by the market
as �stakeholder capital�is di¢ cult to value.
There are several potential explanations of the positive correlation found by this paper. One
is that high employee satisfaction causes superior future stock returns. Even though employee
satisfaction is intangible and not directly valued by the market, it subsequently manifests
3
in positive tangible outcomes that are incorporated into prices. I indeed �nd that the Best
Companies exhibit superior future accounting performance. However, other potential channels
may exist, such as employee satisfaction leading to favorable non-�nancial news releases by the
�rm (e.g. announcing a new contract or �ling a patent), favorable news articles by the press
(e.g. positive customer reception to a product), positive equity analyst reports (e.g. on growth
potential), the absence of unfavorable news (e.g. reports of worker unrest, studied by Mas
(2007)), or the launch of new products. If this explanation accounts for a signi�cant portion of
the positive correlation, it suggests that �rms may be able to improve shareholder returns by
introducing employee-friendly programs.
However, as with other long-run event studies (e.g. Gompers, Ishii and Metrick (2003),
Yermack (2006), Liu and Yermack (2007)), we do not have a natural experiment with random
assignment of the variable of interest to �rms, and so alternative explanations also exist. First,
there may be reverse causality from �rm performance to satisfaction. This paper uses stock
returns (rather than accounting performance) as the primary dependent variable to partially
address this interpretation: successful �rms should already have high stock prices today and
thus not earn superior future returns. However, reverse causality can still occur in the pres-
ence of private information: employees with inside knowledge that their �rm has strong future
prospects will report higher satisfaction today. While existing studies on employee trading be-
havior suggest that workers have no superior information on their �rm�s future stock returns
(e.g. Bernartzi (2001), Bergman and Jenter (2007)), it is not possible to rule out this expla-
nation directly. A second alternative explanation is that employee satisfaction is irrelevant
for shareholder value and simply proxies for other variables that are positively linked to stock
returns. While I control for an extensive set of observable characteristics, by their very nature
unobservables (such as good management practices) cannot be directly controlled for. If the
above explanations account for the bulk of the link between employee satisfaction and stock
returns, improving employee welfare will not necessarily cause a �rm�s stock returns to improve.
However, the two other conclusions of the paper still remain �the existence of a pro�table SRI
trading strategy, and the market�s failure to incorporate the contents of the Best Companies
list (even if list inclusion is a proxy for good management, another intangible).
While some other papers in the �nance and management literatures have investigated the
link between alternative measures of employee welfare and corporate outcomes, this paper
features a number of di¤erences. Most notably, it uses the Best Companies list, which has
a number of advantages. First, it is available for a particularly long time series (22 years),
which is particularly important to ensure that the results are not driven by outliers. The
vast majority of alternative measures o¤er few data points. Second, the list is arguably the
most comprehensive measure of a �rm�s working conditions, receiving signi�cant attention
from shareholders, company management, employees and human resource departments. While
most other measures only consider stated company policies (which can be manipulated), this
list conducts an extensive �grass-roots� analysis of the employees themselves. If this list is
indeed particularly accurate, I might �nd a link with corporate performance even though many
previous studies found no relationship. Third, the list has been highly publicized by Fortune
4
magazine, thus allowing me also to investigate the market�s valuation of intangibles and draw
implications for managerial myopia. In addition to the contrasting measure of satisfaction,
this paper�s methodology is also di¤erent, by focusing on long-run stock returns. Compared
to accounting returns, this outcome variable is more directly linked to shareholder welfare,
allows for risk controls, and captures all potential channels through which satisfaction may
improve investor returns. Compared to event-study returns, it captures the full magnitude of
the relationship if the market does not respond immediately to list publication.
This paper is organized as follows. Section 2 discusses the a priori reasons for hypothesizing
a link between employee satisfaction and shareholder value. Section 3 discusses the data and
methodology, Section 4 presents the results, Section 5 discusses the possible explanations for
the �ndings, and Section 6 concludes.
2 Why Might Employee Satisfaction Matter?
It may seem highly intuitive that �rms should perform more strongly if their employees are
happier, perhaps even removing the need to document such a relationship empirically. How-
ever, existing theoretical and empirical research is far from unambiguous, thus providing the
motivation for this study.
First, employee satisfaction may represent an ine¢ ciently high level of compensation. Hotelling�s
(1932) lemma for the �rm�s pro�t function (that changes in labor costs leads to a change in
pro�ts of equal and opposite magnitude) underpinned �zero-sum� beliefs that shareholders
maximize their return by minimizing employee wages. For example, Taylor�s (1911) �scienti�c
management�theory viewed employees as no di¤erent from any other input, and thus sought
to extract maximum e¤ort while minimizing total pecuniary and non-pecuniary compensation.
High satisfaction may indicate that workers are being allowed to slack or are given super�uously
pleasant working conditions, to the detriment of shareholder value. Indeed, agency problems
may lead to managers tolerating insu¢ cient e¤ort and/or excessive pay, at shareholders�ex-
pense. The manager may derive private bene�ts from improving his colleagues�compensation,
such as more pleasant working relationships (Jensen and Meckling (1976)), or seek to enjoy the
�quiet life�and avoid the e¤ort and potential confrontatinos required to bargain employees all
the way down to their reservation wage (Bertrand and Mullainathan (2003)). Alternatively,
high wages may constitute a takeover defense and lead to managerial entrenchment (Pagano
and Volpin (2005)). Cronqvist et al. (2008) �nd that salaries are higher when managers are
more entrenched, which supports the view that high employee compensation is ine¢ cient.
Second, employee satisfaction may represent an ine¢ cient form of compensation compared
with cash, for the same reason that CEO perquisites are often viewed as ine¢ cient compensation
(e.g. Yermack (2006)). The CEO is forced to consume the perk even if his marginal valuation
is low, whereas cash is fully fungible and could be used to buy the perk if the CEO desires
it. Indeed, in the early 20th century, cash was viewed as the most e¤ective motivator: given
relatively harsh economic conditions, workers were mainly concerned with meeting their physical
5
needs (such as food and shelter), which could be addressed with money.
Existing empirical evidence indeed fails to document a positive association between employee
satisfaction and stock returns. Consistent with the �zero-sum�philosophy, Abowd (1989) �nds
that announcements of pay increases reduce stock market valuations dollar-for-dollar. Diltz
(1995) demonstrates no link between shareholder returns and the Council on Economic Priori-
ties employee relations variable, and Dhrymes (1998) �nds the same result for KLD Research
& Analytics�employee relations measure. This view is also common among practitioners: Tsao
(2003) in BusinessWeek and Zimmerman (2004) in the Wall Street Journal discuss investors�
concerns that Costco o¤ers excessively generous pay and bene�ts to employees. The title of
Zimmerman�s article, �Costco�s Dilemma: Be Kind To Its Workers, or Wall Street?�, captures
the views that employee satisfaction is necessarily at the expense of investor returns.
On the one hand, the above research renders the relationship between employee satisfaction
and shareholder returns non-obvious, and thus potentially interesting to study. On the other
hand, to justify empirical investigation, it is also necessary to have convincing reasons, grounded
in existing theories, for why a positive link might exist. A plausible a priori hypothesis is
important to mitigate �data-mining� concerns and reduce the risk that erroneous inferences
are drawn from accidental patterns in the data.
Human relations theories argue that satisfaction is an e¢ cient form of compensation in
the modern �rm. Maslow (1943) and Hertzberg (1959) stress that money is only an e¤ective
motivator up to a point: once workers�basic physical requirements are met (which is increasingly
true in the more a uent current economic environment), they are increasingly motivated by
non-pecuniary factors such as recognition and self-esteem. While perks (such as country club
membership) can be bought with cash from third parties, on-the-job satisfaction cannot be
externally purchased and can only be provided directly by the �rm.
Moreover, the e¢ ciency wage hypothesis (Akerlof and Yellen (1986)) argues that employee
satisfaction may represent an e¢ cient level of compensation, since shareholders may bene�t
from paying workers in excess of their reservation wage. While the e¢ ciency wage hypothesis
was initially stated in terms of salary, �excess�job satisfaction may be even more powerful if
employees value it more highly than cash, for the reasons stated above.
One potential channel is motivation, and is supported by theories in both neoclassical eco-
nomics and sociology. In the traditional �rm, employees were primarily required to follow
prescribed processes, the output of which was easily measurable. This made motivation simple:
managers could simply measure output, and reward or punish the workers accordingly through
monetary �piece rates�or the threat of dismissal (Taylor (1911)). However, in the modern �rm,
workers are now increasingly called upon to perform skilled tasks, the output of which can be
very di¢ cult to measure accurately (for example, building client relationships). Output-based
incentives may thus be ine¤ective or even destructive (Kohn (1993)).2 If the employee�s output
is not quanti�able, but still observable (e.g. her cooperativeness with colleagues), the manager
can �re her if she shirks. In this case, providing �excess�job satisfaction may be an e¤ective
2Manso (2007) and Ederer and Manso (2007) show that motivating employees to innovate may requiretolerance of failure.
6
motivator since the employee risks being �red, and losing such satisfaction, if she exerts low
e¤ort (Shapiro and Stiglitz (1984)).
If output is neither veri�able nor observable, extrinsic motivators such as piece pay or
the threat of severance are ine¤ective. However, simply removing extrinsic motivators may
encourage workers to shirk. This is where intrinsic motivators become important. An employee
may view pleasant working conditions a �gift� from the �rm, and respond with a �gift� of
increased e¤ort (Akerlof (1982)).3 Sociological theories argue that job satisfaction can lead to
employees identifying with the �rm and internalizing its objectives in their own utility functions,
thus inducing e¤ort even if its outcomes are not directly rewarded in the employment contract
(Maslow (1943), Hertzberg (1959), McGregor (1960)). Supporting this hypothesis, Mas (2007)
�nds that labor unrest in Caterpillar led to reduced product quality. Unlike quantities, quality
is a non-contractible measure of e¤ort that is di¢ cult to control extrinsically and was only
discovered by customers after a sustained period of use.
A second channel is retention. In the early 20th century �rm, retention was a minor issue
as employees predominantly performed unskilled tasks and were easily substitutable; quits
were nonchalantly met by new recruitment. By contrast, employees are the key source of
value creation in many modern �rms (Zingales (2000)), such as those in the service sector
(e.g. �nancial �rms) or in high-quality manufacturing (pharmaceuticals, electronics). However,
unlike physical capital, human capital is inalienable and owned by the workers themselves,
not managers or shareholders, and can thus leave the �rm at little cost. E¤ective retention is
therefore crucial for building competitive advantage through a superior workforce.4 Similarly,
the resource-based view of the �rm (e.g. Wernerfelt (1984)) argues that sustainable competitive
advantage is attained through nurturing and retaining inimitable assets, such as human capital.
Recent employee-centric theories of the �rm such as Rajan and Zingales (1998, 2001), Berk,
Stanton and Zechner (2006), and Lustig, Syverson and van Nieuwerburgh (2007) show that it
can be addressed by paying workers in excess of their market wage or granting them a pecuniary
share of surplus. However, if employees value job satisfaction more highly than salary, the former
is an even more e¤ective retention tool. For the same reason, superior working conditions can
be a powerful recruitment tool.
In sum, while the relationship between employee satisfaction and corporate performance is
not su¢ ciently unambiguous to remove the need for empirical documentation, there appears
to be an adequately strong a priori hypothesis for a positive link to motivate such a study
and address data mining concerns. Indeed, some recent studies document a positive relation-
ship between employee-friendly policies and productivity and/or perceptions of organizational
performance (Delaney and Huselid (1996), Konrad and Mangel (2000), Perry-Smith and Blum
(2000), Bloom, Kretschmer and Van Reenen (2006)). However, these papers do not analyze the
3See Bénabou and Tirole (2003) and Carlin and Gervais (2007) for additional economic models of intrinsicmotivation and work ethic. Falk and Kosfeld (2006) show experimentally a positive relation between trust andproductivity.
4If the labor market is perfectly competitive, and human capital is fully transferable, superior workers canonly be retained by paying them in full for their higher productivity. Retention is still desirable if there arecosts of replacing employees (e.g. recruitment and training costs).
7
e¤ect on shareholder returns, which has a number of advantages described below. In addition,
their measures of employee welfare are only available for a short time series.
Closest to this study is a contemporaneous working paper by Faleye and Trahan (2006),
who show that the Best Companies exhibit superior contemporaneous accounting performance
than benchmark �rms. They focus only on the 1998-2004 period when the list was published
by Fortune, and do not examine long-run returns. Lau and May (1998) �nd a similar link to
accounting pro�ts using the 1993 list. However, the causality of this relationship is unclear,
since strong performance could cause higher satisfaction.
This paper uses stock returns as the primary dependent variable for number of reasons.
First, it su¤ers from fewer reverse causality issues: a pro�table company should not exhibit
superior future returns as its quality should already be incorporated in the current stock price.
Unlike employee satisfaction, pro�ts are tangible and thus likely to be valued by the market.
In contrast to accounting performance, stock returns should not be persistent (after controlling
for momentum e¤ects). However, the use of stock returns does not entirely rule out reverse
causality explanations, as discussed in Section 5. Second, the stock price is most directly
linked to shareholder value, capturing all of the possible mechanisms through which employee
satisfaction may have an e¤ect. While accounting performance may indeed be one channel, it
is unlikely to be the only (or even the most important) one, in particular since the bene�ts of
intangible investment may not manifest in accounting variables for several years. As previously
noted, employee satisfaction may lead to many other tangible outcomes that are directly valued
by the market and thus a¤ect the stock price, such as favorable non-�nancial news releases.
Indeed, prior studies have shown that the Best Companies perform strongly on non-accounting
performance measures, such as customer satisfaction (Simon and DeVaro (2006)) and employee
loyalty (Fulmer, Gerhart and Scott (2003)).5 Third, examining stock returns allows study of
not only the link between satisfaction and performance, but also whether this link is impounded
into prices, and the pro�tability of applying an SRI screen for investment returns.67 Finally,
stock return analysis allows for controls for risk, in particular to ensure that high returns are
not due to compensation for covariances with macroeconomic risk factors.
Having chosen to study stock returns, a second decision point is the horizon. I study long-
run returns as the event-study reaction will substantially understate the importance of employee
5Fulmer et al. (2003) investigate employees�desire to remain with the �rm in a year�s time. This measurewas collected by the Great Place to Work Institute but not used to form the Best Companies lists, and thus avalid outcome variable.
6Filbeck and Preece (2003) examine the relationship between inclusion in the 1998 Fortune list and stockreturns from 1987-1999. Interpretations may therefore be a¤ected by reverse causality: employee satisfactionmay be caused by strong past stock returns. They also �nd that Best Companies do not outperform size- andindustry-matched benchmarks. Fuller et al. (2003) �nd that returns over 1995-2000 to the Best Companies inthe 1998 list did not signi�cantly outperform matching �rms. At a conference, Kurtz and Luck (2002) presentedresults of the Best Companies�performance using the BARRA and North�eld attribution models. Goenner(2007) similarly controls for the market beta.
7Tobin�s Q is another potential dependent variable. However, if employee satisfaction is not valued by themarket, it will not show up in Q. By contrast, if employee satisfaction is fully valued, there may still be norelationship between Q and satisfaction. If all �rms choose satisfaction optimally given their circumstances (i.e.�rms with unskilled labor invest little in employee welfare), there should be no correlation. Demsetz and Lehn(1985) made this point in relation to Tobin�s Q and managerial ownership.
8
satisfaction if the market does not fully incorporate intangibles. Indeed, Hannon and Milkovich
(1996) �nd no event-study reaction to the publication of the 1984 list; Faleye and Trahan
(2006) �nd event-study returns of around 0.5%, signi�cantly lower than the long-run returns in
this paper. Conversely, considering only short-horizon returns might lead to overstated results.
Even if employee satisfaction is irrelevant for performance, the market might erroneously believe
that it is important and irrationally react to Fortune list inclusion. Gilbert et al. (2006) �nd
that the market reacts to a meaningless macroeconomic variable that investors erroneously pay
attention to, and Huberman and Regev (2001) document a �rm-level case of reaction to non-
information. In addition, analyzing long-run returns allows inference of whether intangibles
are impounded into prices. Event-study abnormal returns (with no drift) would imply that the
market responds to intangible information as soon as it is made public. By contrast, positive
drift indicates that the market does not fully value intangibles, even when made visible by
a study as widely disseminated as the Fortune one. It would also suggest a pro�table and
actionable trading strategy.
This paper is also related to a growing literature on socially responsible investing (SRI).
Existing evidence on the pro�tability of SRI strategies is mixed at best. Moskowitz (1972),
Luck and Pilotte (1993) and Derwall et al. (2005) �nd some evidence that SRI screening
improves returns, although based on small samples. Hamilton, Jo and Statman (1993), Kurtz
and DiBartolomeo (1996), Guerard (1997), Teoh, Welch and Wazzan (1999), Bauer, Koedijk
and Otten (2005) and Schröder (2007) report that SRI portfolios have similar returns to their
benchmarks. Hong and Kacperczyk (2008) document superior returns to �sin� stocks, such
as tobacco and gambling, that would be screened out by a SRI strategy. Geczy, Stambaugh
and Levin (2005) show that investors can experience signi�cant losses by restricting themselves
to SRI mutual funds; the magnitude of such losses depends on their prior beliefs about asset
pricing models and fund managers�skills. Brammer, Brooks and Pavelin (2006) �nd a negative
e¤ect of environmental and community screens. This study provides support for an SRI strategy
that involves investment in �rms with superior employee relations.
Finally, the �long-run event study�methodology of this paper is shared by a number of
other papers that link �rm characteristics to future stock returns, such as Gompers, Ishii and
Metrick (2003) on corporate governance, Yermack (2006) on corporate jet usage, and Liu and
Yermack (2007) on CEO home purchases. The variable studied in this paper is particularly
widely publicized. The paper therefore analyzes not only the importance of a particular variable
for shareholder value (as with other event studies), but also whether it is fully incorporated by
the market.
3 Data and Summary Statistics
My main data source is the list of the �100 Best Companies to Work for in America�. This list
was �rst published in a book in March 1984 by Levering, Moskowitz and Katz, and updated
9
in February 1993 by Levering and Moskowitz.8 From 1998, it has been featured in Fortune
magazine each January.
The list has used consistent criteria throughout its 22-year existence, and has been headed
by Robert Levering and Milt Moskowitz throughout. It is compiled from two principal sources.
Two-thirds of the total score comes from employee responses to a 57-question survey created
by the Great Place to Work R Institute in San Francisco. This survey covers topics such as
attitudes toward management, job satisfaction, fairness in the workplace, and camaraderie. 250
employees are randomly selected in each �rm, �ll in the surveys anonymously, and return their
responses directly to the Institute. The response rate is around 60%.9 The remaining one-third
of the score comes from the Institute�s own evaluation of factors such as a company�s demo-
graphic makeup, pay and bene�ts programs, and the �rm�s response to a series of open-ended
questions about its culture. The companies are scored in four areas: credibility (communica-
tion to employees), respect (opportunities and bene�ts), fairness (compensation, diversity), and
pride/camaraderie (teamwork, philanthropy, celebrations). After evaluations are completed, if
signi�cant negative news about a �rm�s employee relations comes to light, the Institute may
exclude that company from the list. It is important to note that Fortune has no involvement in
the company evaluation process, else it may have incentives to bias the list towards advertisers
(Reuter and Zitzewitz (2006)). To be considered for the list, a �rm must have been in existence
for at least seven years and have at least 1,000 employees.
Previously used measures of employee welfare (e.g. the KLD employee relations variable)
are based on observable human resource practices and outcomes, such as minority composi-
tion of the workforce, or the existence of a �extime work policy. These measures are likely
noisy as they are easy to manipulate �a �rm that placed little weight on employee welfare
might hire a minority as a nonexecutive director to �check the box�. This issue is similar to
how short-term earnings may be a misleading measure of a �rm�s fundamental value. The
Best Companies survey is particularly thorough as, in addition to considering such practices
and outcomes, it involves an in-depth �grass-roots�analysis of employee satisfaction through
extensively surveying the workers. Consequently it is arguably the most respected and presti-
gious measure of a �rm�s working conditions, receiving signi�cant attention from shareholders,
company management, employees and human resource departments.
Since 1998, the Best Companies list has been published in the �rst issue of Fortune mag-
azine each year. The publication date is typically in mid-January, and the issue reaches the
newsstands one week before the publication date. If the stock market perceives a positive link
between satisfaction and shareholder value and fully incorporates it into prices, the contents of
the list should be impounded into prices by at least the start of February. Therefore, February
1 is the date for formation and rebalancing of the portfolios from 1998 onwards. Similarly, the
1984 portfolio is formed on April 1, and the 1993 portfolio is formed on March 1.
8These dates are for the hardback edition. The paperback editions were published approximately a year later,but it is the hardback publication date that is relevant as investors would have been able to trade on the listafter publication of the hardback issue.
9While the Institute was not founded until 1990, Levering and Moskowitz used the same criteria for the 1984list, although they surveyed employees directly rather than through a questionnaire.
10
Table 1 details the number of companies in the Best Companies list in year t that had stock
returns available on CRSP in at least one month from the portfolio formation date until the
next portfolio formation date. The table also gives the number of �rms added to and dropped
from the list. Over the entire 1984-2005 time period, 224 separate publicly traded companies
were included in a Best Companies list.
On April 1, 1984, I form a portfolio containing the 74 publicly traded Best Companies in
that year, and measure the returns to this portfolio from April 1984 to February 1993. (The
results are conducted for both equal-weighted and value-weighted portfolios). The portfolio is
reformed on March 1, 1993 to contain the 65 �rms included in the new Best Companies list, and
returns are calculated from March 1993 through January 1998. This process is repeated until
January 2006 and I call this �Portfolio I�. If a �rm de-lists (e.g. goes bankrupt, or is acquired),
then its delisting return is used in its �nal month. At the start of the next month, the proceeds
are reinvested in all of the other stocks in the portfolio, based on their relative weights in the
portfolio at that point in time.10 If a Best Company is not traded in the �rst month after
list publication but goes public before the next list, I add it to the portfolio from the �rst full
month after it starts trading. 78 �rms therefore feature in Portfolio I from 1984-1993, since
four �rms in the initial list became public over that period. I include Best Companies with
only ADRs in the U.S., since an investor constrained to hold U.S. shares would have been able
to invest in such �rms. The results are unchanged when excluding �rms with ADRs, or �rms
that go public mid-way through the year.
Table 2 presents summary statistics on the original 74 Best Companies in March 1984, and
the 69 Best Companies in the �rst Fortune list in January 1998. The mean market capitalization
was $4 billion in 1984, rising to $22 billion in 1998. One notable statistic is that the dividend
yields are low: 2.7% (1.2%) in 1984 (1998), which is consistent with signi�cant reinvestment
of earnings in human capital. The average market-book ratio is a high 2.3 (4.9) and the mean
proportion of total assets accounted for by intangibles is only 0.9% (4.5%). Together, these
results suggest that these companies have little human capital on the balance sheet. This may
result from accounting standards hindering capitalization of this asset.
The most common industries in 1984 were consumer goods (7 companies), hardware (7),
measuring and control equipment (5), retail (5), and �nancial services (5). In 1998 they were
consumer goods (7), �nancial services (6), software (5), pharmaceuticals (5), hardware (4),
and electronic equipment (4). Human capital is plausibly an important input in all of these
industries, with the link perhaps less obvious for consumer goods.
4 Analysis and Results
To ensure that any outperformance of the Best Companies does not result simply from their
high exposure to risk factors, I run monthly regressions of portfolio returns on the four Carhart
(1997) factors, as speci�ed by equation (1) below:
10Results are unchanged if I instead reinvest any takeover proceeds in the new parent, under the rationalethat at least part of the merged entity exhibits superior employee satisfaction.
11
Rit = �+ �MKTMKTt + �HMLHMLt + �SMBSMBt + �MOMMOMt + "it (1)
where:
Rit is the return on Portfolio i in month t, in excess of a benchmark. Three di¤erent
benchmarks are used, described below.
� is an intercept that captures the abnormal risk-adjusted return, and is the key variable
of interest.
MKTt is the return on the CRSP value-weighted index in excess of the risk-free rate. This
represents a market factor.
HMLt is the return on a zero-investment portfolio which is long (short) high (low) book-
market stocks. This represents a value factor.
SMBt is the return on a zero-investment portfolio which is long (short) small (large) stocks.
This represents a size factor.
MOMt is the return on a zero-investment portfolio which is long (short) past winners
(losers). This represents a momentum factor.
"it is a generic error term which is uncorrelated with the independent variables.
All the regressors are taken fromKen French�s website. There remains considerable academic
debate as to whether the four factors proxy for economic risk or mispricing. I do not take a
stance on this issue as either explanation is a potential source of omitted variables bias: employee
satisfaction may be itself irrelevant and simply correlated with a �rm attribute positively related
to stock returns (either risk or mispricing). The alpha in equation (1) re�ects the excess return
compared to passive investment in a portfolio of the factors. Standard errors are calculated
using Newey-West (1987), which allows for "it to be heteroskedastic and serially correlated;
results are very similar if spherical standard errors are assumed.
The returns Rit are calculated over three di¤erent benchmarks. The �rst is the risk-free
rate, taken from Ibbotson Associates. The second is an industry-matched portfolio using the
49-industry classi�cation of Fama and French (1997). This is to ensure that outperformance is
not simply because the Best Companies operated in industries that enjoyed strong returns. It
also controls for any industry-speci�c risks not captured in the systematic risk factors of the
Carhart (1997) model. The third is the characteristics-adjusted benchmark used by Daniel et
al. (1997) and Wermers (2004)11, which matches each stock to a portfolio of stocks with similar
size, book-market ratio and momentum. This is to ensure that the outperformance is not simply
because the Best Companies are exploiting the size, value and/or momentum anomalies. It is
conservative, but not necessarily super�uous, to subtract the returns on the Daniel et al. (1997)
benchmarks before running the four-factor regression, as characteristics can have explanatory
power even when controlling for covariances (Daniel and Titman (1997)).
4.1 Core Results11The benchmarks are available via http://www.smith.umd.edu/faculty/rwermers/ftpsite/Dgtw/coverpage.htm.
12
My hypothesis is that Portfolio I generates signi�cant alphas over its benchmarks and risk
factors. This is a joint test of two sub-hypotheses: employee satisfaction is positively associated
with corporate performance, and the market fails to fully incorporate this link even when the
list is made publicly available.
Table 3 presents the core results of the paper, for the entire 1984-2005 period. Portfolio
I indeed generates signi�cant returns over all benchmarks, regardless of whether it is equal-
or value- weighted. For example, the monthly alpha over the risk-free rate is 0.34% for both
equal- and value-weighted returns. This equates to 4% annually. Note that the coe¢ cient on
the momentum factor is always negative, and sometimes signi�cant. This is inconsistent with
the idea that good stock performance leads employees to respond positively to the survey, and
that the Best Companies simply capture a momentum e¤ect. Unreported annual results show
that the outperformance is consistent over time, with Portfolio I beating the market in 18 out
of the 22 years from 1984-2005, including every year in the 2000-2002 period when the market
declined sharply.
The outperformance in Table 3 may result from the market simply not being aware of the
Best Companies list until 1998, since it was only published in book form. Even though the
list was publicly available and therefore potentially tradable by any investor, time constraints
prevent investors from analyzing all potentially value-relevant information. Thus, non-salient
information may not be noticed by market participants (e.g. Hou and Moskowitz (2005)).
Therefore, while the full-sample results are consistent with two of the paper�s three main im-
plications (the positive association between employee satisfaction and stock returns, and the
pro�tability of an SRI strategy), they need not imply that the market ignores highly visible
measures of intangibles.
Table 4 therefore repeats the analysis for the 1998-2005 subperiod when the list was featured
in Fortune magazine. If the market does fully value intangibles when they are widely publicized,
the alphas should be insigni�cant in this subperiod. I �nd that the opposite is the case: the
returns to the portfolio are even higher, with an equal-weighted (value-weighted) Portfolio I
earning a 0.64% (0.47%) monthly alpha. Recall that the portfolios are not formed until the
month after Fortune publication, so the results are not driven by irrational event-study reaction
to the publicity resulting from list inclusion.12
One explanation for the higher returns in the 1998-2005 subperiod, despite the greater pub-
licity, is that the list became updated annually, allowing it to contain more recent information
about a �rm�s employee satisfaction. This contrasts with the 1984-1997 period in which the
list was updated only once in fourteen years. Thus, Portfolio I may have contained companies
which su¤ered sharp declines in employee satisfaction since the last list. Section 4.3 provides
evidence that the information in list updates is valuable for enhancing investment returns.
12An alternative hypothesis is that the market erroneously reacts negatively to list inclusion under the beliefthat employee satisfaction is wasteful expenditure, and the long-run outperformance is correction of this tem-porary underpricing. This hypothesis is contradicted by the slightly positive event-study returns documentedby Faleye and Trahan (2006), which I also con�rm in unreported results.
13
4.2 Further Robustness Tests
The above subsection showed that the Best Companies�outperformance was not due to covari-
ance with the Carhart (1997) factors nor to selecting industries or characteristics associated
with abnormal returns. This subsection conducts further robustness tests.
To test whether the results are driven by outliers, I winsorize the top 10% and bottom
10% of returns. The winsorization is conducted by portfolio and by month: for example, the
returns of the top decile of �rms in Portfolio I in June 2000 are replaced by the 90th percentile
return among all �rms in Portfolio I in June 2000, and similarly for the bottom decile. Table
5 illustrates the four-factor alphas for the winsorized portfolios, for both 1984-2005 and the
Fortune subperiod. Again, three di¤erent benchmarks are used and the returns are both equal-
and value-weighted. The alphas remain signi�cant in the vast majority of speci�cations; in
some speci�cations the statistical signi�cance increases as winsorization reduces the portfolio
standard errors. The results in other tables are also robust to winsorization.
An additional concern is that the explanatory power of Fortune list inclusion stems only
from its correlation with �rm characteristics associated with superior returns other than the
size, book-to-market or momentum variables already studied in Tables 3 and 4. Calculating the
returns on a benchmark portfolio with similar characteristics is only feasible when the number
of characteristics is small, else it is di¢ cult to form a benchmark. I therefore use a regression
approach to control for a wider range of characteristics than the three studied by Daniel et al.
(1997). Speci�cally, I run a Fama-MacBeth (1973) estimation of equation (2) below:
Rit = at + btXit + ctZit + "it (2)
where:
Rit is the return on stock i in month t, either unadjusted or in excess of the return on the
industry-matched portfolio.
Xit is a dummy variable that equals 1 if �rm i was included in the most recent Fortune
survey.
Zit is a vector of �rm characteristics.
"it is a generic error term which is uncorrelated with the independent variables.
The �rm characteristics included in Zit are taken from Brennan, Chordia and Subrah-
manyam (1998). These are as follows:
SIZE is the natural logarithm of i�s market capitalization at the end of month t� 2.BM is the natural logarithm of i�s book-to-market ratio. This variable is recalculated each
July and held constant through the following June.
Y LD is the ratio of dividends in the previous �scal year to market capitalization measured
at calendar year-end. This variable is recalculated each July and held constant through the
following June.
RET2-3 is the natural logarithm of the cumulative return over months t� 3 through t� 2.RET4-6 is the natural logarithm of the cumulative return over months t� 6 through t� 4.
14
RET7-12 is the natural logarithm of the cumulative return over months t�12 through t�7.DV OL is the natural logarithm of the dollar volume of trading in security i in month t� 2.PRC is the natural logarithm of i�s price at the end of month t� 2.
The results are presented in Table 6. For both the unadjusted and industry-adjusted speci�-
cations, Best Companies inclusion is associated with an abnormal return of over 40 basis points
for the full sample, and an even higher return in the Fortune subperiod. This suggests that
the Best Companies�outperformance does not result from their correlation with the observable
characteristics studied by Brennan et al. (1998).
4.3 Alternative Portfolio De�nitions
This subsection analyzes the returns to three alternative portfolios. This allows me to investi-
gate whether updates of the Best Companies list provide value-relevant information to investors,
or instead whether the results are principally driven by the list published in one particular year.
Portfolio II is not reformed or reweighted each year: it simply calculates the returns to the
original 74 Best Companies from April 1984 to January 2006. This portfolio represents the
simplest trading strategy, as no rebalancing is required and no transactions costs incurred: an
investor could have simply held the same portfolio for 22 years. For the Fortune subsample,
this portfolio calculates the returns of the 69 Best Companies in the 1998 list from February
1998 to January 2006.
Portfolio III adds to the original 1984 portfolio (1998 portfolio for the Fortune subsample)
any new companies which appear on subsequent lists, but does not drop any �rm that is later
removed. The motivation is that some companies may have dropped out of the Top 100, but
still exhibited superior employee satisfaction than the average �rm (e.g. now be in the Top
150).
Conversely, Portfolio IV includes only companies dropped from the list. Speci�cally, it is
created on March 1, 1993 and includes any companies that were in the 1984 list but not in the
1993 list. On February 1, 1998, any companies that were in the 1993 list but not in the 1998
list are added, and so on. If a �rm is later added back to the list, it is removed from Portfolio
IV. (For the Fortune subsample, it is created on February 1, 1999.) Like Portfolio I, Portfolios
III and IV include �rms that go public after list formation.
Portfolios II and III should outperform their benchmarks, since they contain �rms with high
employee satisfaction for at least part of the period. I can also form a tentative hypothesis on
the relative performance of Portfolios I-III. Portfolio I should perform the most strongly, since
it represents the most up-to-date list. On the other hand, if Portfolio II performs similarly to
Portfolio I, this would imply that the previous results were driven by a single portfolio: the
1984 (or 1998) list, and thus only around 70 �rms. It would also cast doubt on the list as a good
measure of employee satisfaction, since the subsequent updates do not provide value-relevant
information.
15
While both Portfolios II and III fail to drop companies that have fallen out of the latest
Fortune list, the di¤erence is that Portfolio III contains any companies newly added to the
list. Therefore, it should outperform Portfolio II if recent lists provide useful information.
The hypothesis is tentative as it is di¢ cult to evaluate rigorously: since the three portfolios
contain many common stocks, their returns will very similar and will be likely statistically
indistinguishable. However, we can still verify whether the di¤erences are of the hypothesized
sign.
I also predict that Portfolio IV performs worse than Portfolios I-III, since the former con-
tains companies outside the Top 100 for employee satisfaction. Whether it also underperforms
its benchmarks depends on market incorporation of intangibles. If the market at all times cap-
italizes the value of employee satisfaction, the removal of a company from the list signals that
this variable has declined from previous market expectations. Therefore, under the assumption
that satisfaction improves performance, Portfolio IV should earn negative returns.
However, if employee satisfaction is important but not incorporated by the market, such a
prediction is not generated. In the extreme, if the Best Companies list is completely ignored,
employee satisfaction only feeds through to returns when its bene�ts manifest in future news
releases and earnings announcements. Hence the abnormal return of �rm i depends on its
level of employee welfare compared to the average �rm, rather than compared to the market�s
previous assessment of �rm i�s level of welfare. If �rm i is outside the Top 100, it may still
exhibit above-average satisfaction (e.g. be in the Top 150) and thus generate superior abnormal
returns.
Table 7 illustrates the results. The returns to Portfolio I-III are positive over all time-periods,
benchmarks and weighting methodologies, and statistically signi�cant in most speci�cations.
Portfolio III indeed underperforms Portfolio I and outperforms Portfolio II, suggesting that
the list updates contain useful information and that the earlier results are attributable on the
224 �rms included in the Best Companies list across the entire time period, rather than only
the 70 �rms that featured in the �rst list. These results o¤er a potential explanation for why
outperformance is particular strong over 1998-2005. In the Fortune subperiod, the list was
more updated every year, whereas for 1984-1997 it was updated only once in a fourteen year
period. Indeed, the marginally insigni�cant results for the 1984 Portfolio II arise because it
contained �rms such as Polaroid, Delta Airlines, Dana and Armstrong that featured only in
the 1984 list and su¤ered very weak performance from 1993 onwards.
Also as predicted, Portfolio IV underperforms Portfolios I-III in all speci�cations except
for the equal-weighted speci�cation from 1984-2005. This strong performance disappears when
winsorizing or value-weighting. However, Portfolio IV only underperforms its benchmarks in
one speci�cation (value-weighted from 1998-2005), and outperforms signi�cantly in some spec-
i�cations. This result further suggests that the market did not fully react when the companies
in Portfolio IV were initially added to the list.
If the bulk of Portfolio II�s outperformance occurred in 1998, this would suggest that market
reacts to the Fortune survey within one year, which is not unusually slow compared to other
news. Bernard and Thomas (1989) �nd that the incorporation of earnings announcements may
16
take up to 180 days; since this is signi�cantly greater than the 2-4 week delay between Fortune
publication and portfolio formation, the portfolio will capture part of this reaction. In unre-
ported results I �nd that Portfolio II slightly underperforms in 1998 and outperforms in every
year from 1999-2005, suggesting that non-incorporation of the Fortune survey is signi�cantly
slower than for other information. (The same time-series result is true for Portfolios I and III.)
Similarly, starting Portfolio II in 1999 (i.e. buying and holding the 1999 list) also leads to
economically and statistically signi�cant returns.
5 Discussion
Section 4 has documented a signi�cant correlation between employee satisfaction and future
stock returns that is robust to controls for risk, industries and �rm characteristics. There are
a number of potential explanations for this association:
Hypothesis A: Employee satisfaction causes superior future stock returns.
Hypothesis B: Employee satisfaction is irrelevant for shareholder value, and the higher re-
turns stemmed from irrational market reactions or demand from SRI funds.
Hypothesis C: Expectations of superior future stock returns cause high employee satisfaction
today.
Hypothesis D: There is no causal relationship in either direction between employee satisfac-
tion and stock returns, but a third variable causes both.
Hypothesis A argues that employee satisfaction causes superior �rm performance, through
improving recruitment, retention and motivation as discussed in Section 2. While workforce
quality and motivation are also di¢ cult for the market to observe, they may subsequently
manifest in a number of desirable tangible outcomes that are visible to the market and lead
to improved returns. Examples include improved accounting performance and positive non-
�nancial news releases by the �rm, the press or equity analysts. If this hypothesis accounts
for a meaningful portion of the overall correlation between satisfaction and stock returns, it
suggests that employee-friendly programs can improve corporate performance.
Since a causal explanation is agnostic as to the channel through which employee satisfac-
tion has an e¤ect, the principal analysis used the stock price as the dependent variable, as it
captures all the potential channels. However, the stock price has some limitations. While it
should incorporate all mechanisms that a¤ect shareholder value, it may also be in�uenced by
factors unrelated to shareholder value, such as irrational speculation. Hypothesis B is a second
interpretation of the results: that the Best Companies did not experience a true increase in
fundamental value, and instead their superior stock returns resulted from overvaluation. This
may have occurred if employee satisfaction is irrelevant for shareholder value, but the market
erroneously believed that a relationship exists and reacted irrationally when the list was �rst
featured in Fortune, or investor attention was drawn to the featured companies as a result of the
associated publicity. However, irrational reactions are typically concentrated immediately after
the announcement of irrelevant news, as shown by Huberman and Regev (2001) and Gilbert et
17
al. (2006).13 A similar explanation is that list inclusion leads to buying by SRI funds merely
because it allows the stocks to pass SRI screens, rather than because employee satisfaction is
bene�cial for shareholder value. Such purchases may take time to be executed and need not
occur within the month of list announcement. However, these trades are likely to occur within
one year. Portfolio II underperforms in 1998 and outperforms in every year from 1999-2005,
suggesting the results are not driven by in�ows from SRI funds or the publicity resulting from
list inclusion.
The absence of a coherent story supporting an alternative explanation, however, does not
constitute direct evidence in favor of the proposed explanation. I therefore investigate the ac-
counting performance of the Best Companies, which is not a¤ected by market overvaluation.
Note that this is not the only channel through which employee satisfaction may improve share-
holder value, and may not even be the most important one: particularly in the modern �rm, the
main bene�t of superior human capital may be di¢ cult-to-measure outputs such as the quality
of new products or processes invented, or the strengthening of client relationships, which only
manifest in accounting performance in the long-term. Indeed, Simon and DeVaro (2006) �nd
that the Best Companies exhibit superior customer satisfaction and Fulmer, Gerhart and Scott
(2003) �nd the same for employee loyalty. The strength or weakness of an accounting channel
does not preclude a link through other mechanisms, but I study accounting performance as it
is the most measurable.
Table 8 regresses various accounting performance measures on an indicator variable for
whether the �rm was a Best Company in the previous year. All values are industry-adjusted by
subtracting the median value for the Fama-French (1997) industry. As in Gompers, Ishii and
Metrick (2003), I use median (least-absolute-deviation) regressions owing to the presence of large
outliers, and the log book-to-market ratio as an additional regressor to control for �expected�
di¤erences in pro�tability. I �nd that the Best Companies are associated with higher next-
year pro�t margins, return on equity, and pro�t per employee, for both measures of pro�t
(operating income and net income); all di¤erences are signi�cant at the 1% level.14 The Best
Companies are also associated with higher one-year EPS growth. The only inconsistent result
is that the Best Companies exhibit lower one-year sales growth, although sales per employee
are signi�cantly higher.
While Table 8 looks at next-year performance measures, Table 9 analyzes long-term growth.
It compares the Best Companies in the 1984 list with all other �rms in Compustat according
to their 22-year (1983-2005) growth in sales, operating pro�t and net income. I calculate both
the percentage annualized growth rate and the growth rate in dollars per employee. Again, all
variables are industry adjusted; in addition, they are winsorized at the 1st and 99th percentiles
to remove outliers. While we have too few Best Companies to test the di¤erences statistically,
13Anginer, Fisher and Statman (2007) investigate the returns to another Fortune list, �America�s MostAdmired Companies,�and �nd negative long-horizon to list inclusion, suggesting that inclusion in a Fortune listdoes not lead to irrational price increases.14Stock returns may rationally respond to levels of pro�tability, rather than just changes. Even if pro�tability
was high the previous year, the market will attach a certain probability to this being temporary. High prof-itability in the current year (even if it is at a similar level to the previous year) may increase stock returns asthe market increases its posterior belief that pro�tability is caused by underlying �rm quality, rather than luck.
18
we can draw tentative conclusions based on economic signi�cance. The Best Companies exhibit
a mean industry-adjusted EPS growth of 1.1% per year, compared to the 0.3% exhibited by
other companies, and represents an economically important annualized di¤erence of 0.8%. Per
employee, the growth in net income is over $12,000 higher. The Best Companies also exhibit
higher growth in Tobin�s q, operating pro�t, operating pro�t per employee and sales per em-
ployee, with only growth in sales being slower. Taken together, the results in Tables 8 and
9 provides suggestive evidence that the abnormal returns to the Best Companies was at least
partially due to their superior accounting performance over the period.
Since the setting is not a natural experiment with random assignment of employee satis-
faction to �rms, we cannot make strong claims about causality from employee satisfaction to
shareholder returns. Hypothesis C is that superior performance leads to satisfaction. The use
of stock returns (rather than accounting pro�ts) as a dependent variable addresses concerns
of reverse causation in the absence of private information �past, current and expected future
pro�tability should all be impounded in the current stock price, and so pro�table �rms should
not outperform going forwards. However, if employees have private information about their
�rm�s expected future stock price performance, those with positive information will plausibly
report higher satisfaction today. (Note this direction is not unambiguous, since employees that
predict higher future returns will perceive the stock as undervalued today, potentially reducing
satisfaction). Since any thorough measure of satisfaction (rather than just an observation of
policies or outcomes) must come from workers, it is di¢ cult to think of other measures that
would be immune to this interpretation. However, this hypothesis can be evaluated indirectly
by using prior research on employee trading behavior. Benartzi (2001) shows that employees
make incorrect decisions when allocating their 401(k) accounts to company stock, and Bergman
and Jenter (2007) �nd that �rms are able to lower total compensation by granting their workers
overvalued options in lieu of salary. Both of these studies are inconsistent with the notion that
employees have superior information about future stock returns.
Hypothesis D is that the link between satisfaction and returns arises because a third unob-
servable variable causes both, such as good management practices (Bloom, Kretschmer and Van
Reenen (2006)). In other words, the explanatory power of Fortune inclusion only arises because
it is correlated with an omitted variable. While the analysis in Table 8 ruled out correlation
with observable determinants of returns, by their very nature unobservables cannot be used as
regressors. The standard solution is to introduce �rm �xed e¤ects to absorb the unobservables
and identify purely on within-�rm changes in the variable in question. Unfortunately, this ap-
proach is not appropriate here because �xed e¤ects require the unobservables to be constant
over time, but a change in employee satisfaction could be caused by changes in management
practices. In addition, there is limited within-�rm variation in Fortune inclusion: many �rms
remain in the list for several years, and a �rm removed from the list may still exhibit signi�-
cantly above-average satisfaction (e.g. be in the Top 150). Thus, such an approach would be
biased towards �nding no relationship (Zhou (2001)). 15
15An alternative approach would be to use random variation in some �rm-speci�c characteristic that wascausal for employee satisfaction but has no direct e¤ect on stock returns. Unfortunately, I have been unable
19
If the results were entirely driven by a combination of Hypotheses C and D, then satisfaction
has no causal e¤ect on returns and the introduction of employee-friendly programs (without
altering other management practices) would have no impact. However, other conclusions from
this paper would be una¤ected. It still remains that the market does not incorporate intangibles
(be they satisfaction or good management) even when made publicly available, and that an
investor could have earned signi�cant risk-adjusted returns by trading on the Best Companies
list.
Another important caveat, shared by many other long-run event studies, is that the sample
size is small. The Best Companies survey only contains 100 �rms per year (of which ap-
proximately 2/3 are publicly traded). Since these �rms are all in the right tail of employee
satisfaction, this small sample may not re�ect the relationship between shareholder returns and
the whole range of levels of satisfaction, as it may be that a positive link only exists at very
high levels. In addition, while the paper documents superior returns to an SRI screen based on
employee relations, its results may not extend to other SRI screens (e.g. environmental policy).
Finally, note that there are also factors that may lead to the results being understated.
Under Hypothesis A, the portfolio returns only capture the bene�ts of employee satisfaction
that have manifested in tangible outcomes within the time period studied. However, the e¤ects
of employee satisfaction (such as developing a new patent) may not become visible for several
years, and thus not be captured by the current analysis. In addition, some �rms may choose not
to be considered for the Best Companies list, perhaps because their reputations for employee
welfare are already strong and they do not need independent certi�cation. Thus, there may
be many companies with high employee satisfaction and stronger returns than the mean Best
Company not considered by this analysis. Finally, since employee satisfaction is particularly
di¢ cult to measure accurately, measurement errors will bias the results towards zero.
6 Conclusion
This paper �nds that �rms with high levels of employee satisfaction generate superior long-
horizon returns, even when controlling for industries, factor risk or a broad set of observable
characteristics. These �ndings imply that the market fails to incorporate intangible assets fully
into stock valuations � even if the existence of such assets is veri�ed by a widely respected
survey. This suggests that the market may have even greater di¢ culty in valuing other forms
of intangible investment, and provides empirical support for theoretical models of managerial
myopia. In addition, the results imply that an SRI screen based on employee welfare may
improve investment performance, although they may not extend to other SRI screens.
The results are consistent with the view that employee satisfaction causes stronger corporate
performance, potentially through improved recruitment, retention and motivation. However,
there are alternative interpretations of this association which the data cannot entirely rule out.
The economic magnitudes documented by the paper suggest that future research that success-
to identify such an appropriate instrument. For example, �natural experiments� such as exploiting labor lawregulatory change are not �rm-speci�c.
20
fully identi�es the underlying causes of superior performance may have important implications.
If superior employee satisfaction caused even a portion of the 34 basis point monthly abnormal
return, then employee-friendly programs can substantially improve shareholder value.
21
Table 1: Summary Statistics
The second column details the number of Best Companies that had returns available on
CRSP for at least one month between publication of the list of that year, and the subsequent
list. The third column gives the number of new companies added to the Best Companies
list of that year. The fourth column contains the number of companies on the previous Best
Companies list which no longer feature in the current list.
Year of List Best Companies Added Dropped
1984 78
1993 69 30 39
1998 70 34 33
1999 68 26 28
2000 60 20 28
2001 55 15 20
2002 55 14 14
2003 61 14 8
2004 57 11 15
2005 58 11 10
22
Table 2: Summary Characteristics
This table illustrates summary statistics for the 74 companies in the 1984 �100 Best Com-
panies to Work For in America�list that were public on April 1, 1984, and the 69 companies
in the 1998 list published in Fortune that were public on February 1, 1998. All data are of the
end of March 1984 (January 1998, respectively) and taken from CRSP and Compustat. The
last three items are based on Compustat data for 1996 (1982, respectively). They are missing
for companies that were not traded in 1996 (1982) and excluded for companies for which only
the ADRs are traded.16
# obs Mean Std. Dev. Min Max
1984 list
Market Cap ($ bn) 74 3.99 9.48 0 69.47
Price ($) 74 37.43 19.64 5.91 113.75
Volume (m) 74 3.59 4.71 0 20.43
Dividend yield (%) 70 2.74 2.29 0 10.11
Market/book 70 2.33 1.85 0.01 9.78
Intangibles as a % of total assets (%) 70 0.92 2.13 0 8.38
1998 list
Market Cap ($ bn) 69 21.51 39.78 0.03 204.59
Price ($) 69 50.99 25.48 5.38 127.56
Volume (m) 69 34.27 71.67 0 406.38
Dividend yield (%) 62 1.19 1.19 0 5.97
Market/book 61 4.86 4.86 -3.14 29.10
Intangibles as a % of total assets (%) 62 5.01 7.50 0 28.88
16To calculate book equity for the Market/Book ratio, I start with stockholders�equity (Compustat item 216)if it is not missing. If it is missing, I use total common equity (item 60) plus preferred stock par value (item130) if both of these are present. Otherwise, I use total assets (item 6) minus total liabilities (item 181), if bothare present. To obtain book equity, I subtract from shareholders�equity the preferred stock value, where we useredemption value (item 56), liquidating value (item 10), or carrying value (item 130), in that order, as available.Finally, if not missing, I add in balance sheet deferred taxes (item 35) to this book-equity value, and subtractthe FASB106 adjustment (item 330).
23
Table 3: Risk-Adjusted Returns
This table documents the results of monthly regressions of portfolio returns on the four
Carhart (1997) factors,MKT ,HML, SMB, andMOM . The regression is speci�ed in equation
(1). The dependent variable is the portfolio return less either the risk-free rate, the industry-
matched portfolio return, or the characteristics-matched portfolio return. Panel A contains
equal-weighted returns and Panel B contains value-weighted returns. The regressors are the
returns to zero-investment portfolios designed to capture market, value, size, and momentum
e¤ects. The alpha is the excess risk-adjusted return. The sample period is April 1984-January
2006.
Excess returns over
Risk-free Industry Characteristics
Panel A (equal-weighted)
� 0.34 0.22 0.25
(3.49***) (2.97***) (2.97***)
�MKT 1.11 0.07 0.12
(38.08***) (3.41***) (4.74***)
�HML 0.03 0.04 0.02
(0.64) (1.23) (0.61)
�SMB 0.15 0.14 0.03
(3.08***) (4.45***) (0.80)
�MOM -0.13 -0.04 -0.08
(4.76***) (2.20**) (3.97***)
Panel B (value-weighted returns)
� 0.34 0.20 0.19
(3.03***) (2.70***) (2.63***)
�MKT 0.95 -0.03 -0.01
(30.29***) (1.30) (0.27)
�HML -0.46 -0.09 -0.15
(8.13***) (2.39**) (3.77***)
�SMB -0.24 -0.25 -0.04
(4.77***) (7.23***) (1.27)
�MOM -0.04 -0.00 -0.02
(0.95) (0.06) (1.03)
# obs 262 262 262*: Signi�cant at the 10% level; **: Signi�cant at the 5% level; ***: Signi�cant at the 1% level
24
Table 4: Risk-Adjusted Returns from 1998
This table documents the results of monthly regressions of portfolio returns on the four
Carhart (1997) factors,MKT ,HML, SMB, andMOM . The regression is speci�ed in equation
(1). The dependent variable is the portfolio return less either the risk-free rate, the industry-
matched portfolio return, or the characteristics-matched portfolio return. Panel A contains
equal-weighted returns and Panel B contains value-weighted returns. The regressors are the
returns to zero-investment portfolios designed to capture market, value, size, and momentum
e¤ects. The alpha is the excess risk-adjusted return. The sample period is February 1998-
January 2006.
Excess returns over
Risk-free Industry Characteristics
Panel A (equal-weighted)
� 0.64 0.46 0.57
(3.70***) (3.28***) (4.08***)
Panel B (value-weighted returns)
� 0.47 0.30 0.32
(2.06**) (2.05**) (2.11**)
# obs 96 96 96
*: Signi�cant at the 10% level; **: Signi�cant at the 5% level; ***: Signi�cant at the 1% level
25
Table 5: Risk-Adjusted Returns of Winsorized Portfolios
This table documents the results of monthly regressions of portfolio returns on the four
Carhart (1997) factors,MKT ,HML, SMB, andMOM . The regression is speci�ed in equation
(1). For each portfolio and for each month, the returns of the constituent stocks are winsorized
at the 10% and 90% levels. The dependent variable is the winsorized portfolio return less either
the risk-free rate, the industry-matched portfolio return, or the characteristics-matched portfolio
return. Panel A contains equal-weighted returns and Panel B contains value-weighted returns.
The regressors are the returns to zero-investment portfolios designed to capture market, value,
size, and momentum e¤ects. The alpha is the excess risk-adjusted return. The sample period
is April 1984-January 2006 for the left-hand column, and February 1998-January 2006 for the
right-hand column.
1984-2005: excess returns over 1998-2005: excess returns over
Risk-free Industry Characteristics Risk-free Industry Characteristics
Panel A (equal-weighted)
� 0.19 0.07 0.13 0.51 0.33 0.48
(1.88*) (1.01) (1.52) (2.93***) (2.43**) (3.44***)
Panel B (value-weighted)
� 0.30 0.14 0.14 0.47 0.30 0.33
(3.49***) (2.02**) (2.02**) (2.25**) (2.23**) (2.36**)*: Signi�cant at the 10% level; **: Signi�cant at the 5% level; ***: Signi�cant at the 1% level
26
Table 6: Characteristics Regressions
This table documents the results of monthly regressions of individual stock returns on a
Fortune list inclusion dummy and the characteristics used in Brennan, Chordia and Subrah-
manyam (1998). SIZE is the natural logarithm of the �rm�s market capitalization (in billions)
in month t � 2. BM is the natural logarithm of the �rm�s book-to-market ratio as of the
calendar year-end before the most recent June. YIELD is the �rm�s dividend yield as of the
calendar year-end before the most recent June. RET2-3, RET4-6 and RET7-12 are the natural
logarithm of the compounded returns in, respectively, month t� 3 to month t� 2, month t� 6to month t � 4, and month t � 12 to month t � 7. DVOL is the dollar trading volume (inmillions) in month t � 2. PRC is the price at the end of month t � 2. The sample period isFebruary 1998-January 2006.
1984-2005 1998-2005
Raw Industry-Adjusted Raw Industry-Adjusted
Best Company 0.46 0.44 0.55 0.52
(4.05***) (4.28***) (2.38**) (2.60***)
SIZE 0.14 0.13 -0.02 -0.05
(1.57) (1.89*) (0.11) (0.32)
BM 0.25 0.24 0.12 0.10
(4.57***) (5.64***) (1.05) (1.21)
YIELD -0.05 -0.04 -0.03 -0.02
(4.21***) (4.39***) (2.26**) (2.31**)
RET2-3 0.80 4.61 1.19 6.02
(2.61**) (0.12) (1.76*) (0.19)
RET4-6 0.91 3.71 1.51 4.63
(3.61***) (0.77) (2.78***) (0.26)
RET7-12 1.03 2.53 0.92 3.07
(5.98***) (0.22) (2.62**) (0.23)
DVOL 1.26 1.06 1.65 1.40
(1.44) (1.53) (0.04) (0.10)
PRC -0.33 -0.31 -0.58 -0.47
(2.72***) (2.65***) (2.31**) (1.92*)
Constant 2.58 1.47 2.76 1.96
(7.16***) (3.25***) (3.89***) (2.19**)*: Signi�cant at the 10% level; **: Signi�cant at the 5% level; ***: Signi�cant at the 1% level
27
Table 7: Alternative Portfolio De�nitions
This table documents the results of monthly regressions of the returns of Portfolios II, III
and IV on the four Carhart (1997) factors, MKT , HML, SMB, andMOM . The regression is
speci�ed in equation (1). The dependent variable is the portfolio return less either the risk-free
rate, the industry-matched portfolio return, or the characteristics-matched portfolio return.
Panel A contains equal-weighted returns and Panel B contains value-weighted returns. The
regressors are the returns to zero-investment portfolios designed to capture market, value, size,
and momentum e¤ects. The alpha is the excess risk-adjusted return. The sample period is
April 1984-January 2006 for the left-hand column, and February 1998-January 2006 for the
right-hand column.
1984-2005: excess returns over 1998-2005: excess returns over
Risk-free Industry Characteristics Risk-free Industry Characteristics
Panel A (equal-weighted)
�, II 0.18 0.13 0.11 0.60 0.44 0.56
(1.61) (1.33) (1.08) (3.25***) (3.56***) (3.85***)
�, III 0.29 0.20 0.21 0.61 0.46 0.55
(3.24***) (3.02***) (2.58***) (3.72***) (3.80***) (4.11***)
�, IV 0.36 0.26 0.29 0.50 0.38 0.43
(2.64***) (1.96*) (2.34**) (1.68*) (1.55) (1.94*)
Panel B (value-weighted)
�, II 0.22 0.17 0.15 0.46 0.24 0.40
(2.28**) (2.09**) (2.11**) (1.96*) (1.62) (2.40**)
�, III 0.23 0.13 0.12 0.32 0.13 0.21
(2.84***) (2.25**) (2.43**) (1.65) (1.11) (1.75*)
�, IV 0.18 0.09 0.05 -0.33 -0.39 -0.22
(1.54) (0.79) (0.60) (1.12) (1.87*) (1.18)
*: Signi�cant at the 10% level; **: Signi�cant at the 5% level; ***: Signi�cant at the 1% level
28
Table 8: Accounting Performance
This table reports the results of median (least-absolute-deviation) regressions of accounting
performance measures on an indicator variable for whether the �rm was a Best Company in
the previous year. The log book-to-market ratio is used as a control variable. The sample
is the universe of Compustat �rms for 1984. All variables are �rst winsorized at the 1st and
99th percentiles, and then industry-adjusted by subtracting the median value from the relevant
Fama-French (1997) industry.
BC dummy ln(B/M) Constant
Operating Income/Equity 0.0517*** -0.0289*** 0.0766***
(0.0052) (0.0003) (0.0004)
Net Income/Equity 0.0936*** -0.0402*** 0.1133***
(0.0089) (0.0005) (0.0006)
Operating Income/Sales 0.0315*** -0.0069*** 0.0366***
(0.0037) (0.0002) (0.0003)
Net Income/Sales 0.0500*** -0.0096*** 0.0566***
(0.0057) (0.0003) (0.0004)
Operating Income/Employees 10.4822*** -0.7440*** 4.0881***
(0.0538) (0.0320) (0.0401)
Net Income/Employees 17.0935*** -1.0246*** 6.4052***
(0.8484) (0.0505) (0.0633)
Sales/Employees 97.2878*** 3.8142*** 153.3986***
(6.5789) (0.3913) (0.4902)
1-year Sales Growth -0.0331*** -0.0416*** 0.0799***
(0.0094) (0.0005) (0.0007)
1-year EPS Growth 0.0759*** -0.0376*** -0.01113***
(0.0377) (0.0021) (0.0027)
*: Signi�cant at the 10% level; **: Signi�cant at the 5% level; ***: Signi�cant at the 1% level
29
Table 9: 22-Year Accounting Growth
This table compares the growth in accounting performance measures between the Best
Companies and all other Compustat �rms, over 1983-2005. For a percentage growth rate to be
calculated, the �rm have positive levels of the accounting variable in both 1983 and 2005. The
growth rates are �rst winsorized at the 1st and 99th percentiles, and then industry-adjusted by
subtracting the median value from the relevant Fama-French (1997) industry.
Best Companies Other Di¤erence
Q Growth 0.51 -0.02 0.53
Sales Growth (% annualized) -0.04% 0.53% -0.57%
Operating Pro�t Growth (% annualized) 1.22% 0.37% 0.85%
Net Income Growth (% annualized) 0.73% 0.59% 0.14%
EPS growth (% annualized) 1.08% 0.28% 0.80%
Growth in Sales per Employee ($k) 142.76 61.78 80.99
Growth in Operating Pro�t Per Employee ($k) 13.95 5.51 8.45
Growth in Net Income Per Employee ($k) 23.96 11.55 12.41
30
References
[1] Abowd, John (1989): �The E¤ect of Wage Bargains on the Stock Market Value of the
Firm�. American Economic Review 79, 774-800
[2] Akerlof, George and Janet Yellen (1986): �E¢ ciency Wage Models of the Labor Market.�
Cambridge University Press, Cambridge UK
[3] Akerlof, George (1982): �Labor Contracts as Partial Gift Exchange.�Quarterly Journal
of Economics 97, 543-69
[4] Anginer, Deniz, Kenneth Fisher and Meir Statman (2007): �Stocks of Admired Companies
and Despised Ones.�Working Paper, University of Michigan
[5] Bauer, Rob, Kees Koedijk and Roger Otten (2005): �International Evidence on Ethical
Mutual Fund Performance and Investment Style.� Journal of Banking and Finance 29,
1751-1767
[6] Bénabou, Roland and Jean Tirole (2003): �Intrinsic and Extrinsic Motivation.�Review of
Economic Studies 70, 489-520
[7] Benartzi, Shlomo (2001): �Excessive Extrapolation and the Allocation of 401(k) Accounts
to Company Stock.�Journal of Finance 56, 1747-1764
[8] Bergman, Nittai and Dirk Jenter (2007): �Employee Sentiment and Stock Option Com-
pensation.�Journal of Financial Economics, forthcoming
[9] Berk, Jonathan, Richard Stanton and Josef Zechner (2006): �Human Capital, Bankruptcy,
and Capital Structure.�Working Paper, University of California at Berkeley
[10] Bernard, Victor and Jacob Thomas (1989): �Post-Earnings-Announcement-Drift: Delayed
Price Response or Risk Premium?�Journal of Accounting Research 27, 1-36
[11] Bertrand, Marianne and Sendhil Mullainathan (2003): �Enjoying the Quiet Life? Corpo-
rate Governance and Managerial Preferences.� Journal of Political Economy 111, 1043-
1075
[12] Bloom, Nick, Tobias Kretschmer and John Van Reenen (2006): �Work-Life Balance, Man-
agement Practices and Productivity.�Working Paper, Stanford University
[13] Brammer, Stephen, Chris Brooks and Stephen Pavelin (2006): �Corporate Social Perfor-
mance and Stock Returns: UK Evidence From Disaggregate Measures.�Financial Man-
agement 35, 97-116
[14] Brennan, Michael, Tarun Chordia and Avanidhar Subrahmanyam (1998): �Alternative
Factor Speci�cations, Security Characteristics, and the Cross-Section of Expected Stock
Returns.�Journal of Financial Economics 49, 345-373
31
[15] Carhart, Mark (1997): �On Persistence in Mutual Fund Performance.�Journal of Finance
52, 57-82
[16] Carlin, Bruce and Simon Gervais (2007): �Work Ethic, Employment Contracts, and Firm
Value.�Journal of Finance, forthcoming
[17] Cronqvist, Henrik, Fredrik Heyman, Mattias Nilsson, Helena Svaleryd and Jonas Vlachos
(2008): �Do Entrenched Managers Pay Their Workers More?�Journal of Finance, forth-
coming
[18] Daniel, Kent, Mark Grinblatt, Sheridan Titman and Russ Wermers (1997): �Measuring
Mutual Fund Performance with Characteristic-Based Benchmarks.� Journal of Finance
52, 1035-1058
[19] Daniel, Kent and Sheridan Titman (1997): �Evidence on the Characteristics of Cross-
Sectional Variation in Common Stock Returns.�Journal of Finance 52, 1-33
[20] Delaney, John and Mark Huselid (1996): �The Impact of Human Resource Management
Practices on Perceptions of Organizational Performance.�Academy of Management Jour-
nal 39, 949-969
[21] Demsetz, Harold and Kenneth Lehn (1985): �The Structure of Corporate Ownership:
Causes and Consequences.�Journal of Political Economy 93, 1155-1177
[22] Derwall, Jeroen, Nadja Guenster, Rob Bauer and Kees Koedijk (2005): �The Eco-
E¢ ciency Premium Puzzle.�Financial Analysts Journal 61, 51-63
[23] Dhrymes, Phoebus J. (1998): �Socially Responsible Investing: Is It Pro�table?�The In-
vestment Research Guide to Socially Responsible Investing, The Colloquium on Socially
Responsible Investing
[24] Diltz, J. David (1995): �Does Social Screening A¤ect Portfolio Performance?�Journal of
Investing Spring, 64-69
[25] Ederer, Florian and Gustavo Manso (2007): �Incentives for Innovation: Evidence from a
Laboratory Experiment.�Working Paper, Massachusetts Institute of Technology
[26] Faleye, Olubunmi and Emery Trahan (2006): �Is What�s Best For Employees Best For
Shareholders?�Working Paper, Northeastern University
[27] Falk, Armin and Michael Kosfeld (2006): �The Hidden Cost of Control.�American Eco-
nomic Review 96, 1611-1630
[28] Fama, Eugene and Kenneth French (1997): �Industry Costs of Equity.�Journal of Finan-
cial Economics 43, 153-193
32
[29] Fama, Eugene and James MacBeth (1973): �Risk, Return, and Equilibrium: Empirical
Tests.�Journal of Political Economy 81, 607-636
[30] Filbeck, Greg and Dianna Preece (2003): �Fortune�s Best 100 Companies to Work For in
America: Do They Work For Shareholders?�Journal of Business, Finance and Accounting
30, 771�797.
[31] Fulmer, Ingrid Smithey, Barry Gerhart and Kimberly Scott (2003): �Are the 100 Best
Better? An Empirical Investigation of the Relationship Between Being a �Great Place to
Work�and Firm Performance.�Personnel Psychology 56, 965-993
[32] Geczy, Christopher, Robert Stambaugh and David Levin (2005): �Investing in Socially
Responsible Mutual Funds.�Working Paper, University of Pennsylvania
[33] Gilbert, Thomas, Lars Lochstoer, Shimon Kogan and Ataman Ozyildirim (2006): �Re-
leases of Previously Published Information Move Aggregate Stock Prices.�Working Paper,
University of California at Berkeley
[34] Goenner, Cullen (2007): �Investing in Fortune�s 100 Best Companies to Work for in Amer-
ica.�Journal of Economics, forthcoming
[35] Gompers, Paul, Joy Ishii and Andrew Metrick (2003): �Corporate Governance and Equity
Prices.�Quarterly Journal of Economics 118, 107-155
[36] Graham, John, Campbell Harvey and Shivaram Rajgopal (2005): �The Economic Im-
plications of Corporate Financial Reporting.� Journal of Accounting and Economics 40,
3-73
[37] Guerard, John (1997): �Is There a Cost to Being Socially Responsible in Investing?�Jour-
nal of Investing 6, 11-18
[38] Hamilton, Sally, Hoje Jo and Meir Statman (1993). �Doing Well While Doing Good:
The Investment Performance of Socially Responsible Mutual Funds.�Financial Analysts
Journal November/December: 62-66.
[39] Hannon, John and George Milkovich (1996): � The E¤ect of Human Resource Reputation
Signals on Share Prices: An Event Study.�Human Resource Management 35, 405-424
[40] Hertzberg, Frederick (1959): �The Motivation to Work.�John Wiley & Sons, New York
[41] Hong, Harrison and Marcin Kacperczyk (2008): �The Price of Sin: The E¤ect of Social
Norms on Markets.�Working Paper, University of British Columbia
[42] Hotelling, Harold (1932): �Edgeworth�s Taxation Paradox and the Nature of Demand and
Supply Functions.�Journal of Political Economy 40, 577-616.
33
[43] Hou, Kewei and Tobias Moskowitz (2005): �Market Frictions, Price Delay, and the Cross-
Section of Expected Returns.�Review of Financial Studies 18(3), 981-1020
[44] Huberman, Gur and Tomer Regev (2001): �Contagious Speculation and a Cure for Cancer:
A Nonevent that Made Stock Prices Soar.�Journal of Finance 56, 387-396
[45] Jensen, Michael and William Meckling (1976): �Theory of the Firm: Managerial Behavior,
Agency Costs, and Capital Structure.�Journal of Financial Economics 3, 305-360
[46] Kohn, Al�e (1993): �Why Incentive Plans Cannot Work.� Harvard Business Review,
September-October 1993
[47] Konrad, Alison and Robert Mangel (2000): �The Impact of Work-Life Programs on Firm
Productivity.�Strategic Management Journal 21, 1227-1237
[48] Kurtz, Lloyd and Dan DiBartolomeo (1996): �Socially Screened Portfolios: An Attribution
Analysis of Relative Performance.�Journal of Investing 5, 35-41
[49] Kurtz, Lloyd and Christopher Luck (2002): �An Attribution Analysis of the 100 Best
Companies to Work for in America.�Presentation to North�eld Investment Conference,
Fish Camp, California, May 5-7
[50] Lau, R.S.M. and Bruce May (1998): �A Win-Win Paradigm for Quality of Work Life and
Business Performance.�Human Resource Development Quarterly 9, 211-226
[51] Levering, Robert and Milton Moskowitz (1993): �The 100 Best Companies to Work For
in America.�Plume, New York NY
[52] Levering, Robert, Milton Moskowitz and Michael Katz (1984): �The 100 Best Companies
to Work For in America.�Addison-Wesley, Reading MA
[53] Liu, Crocker and David Yermack (2007): �Where are the Shareholders�Mansions? CEOs�
Home Purchases, Stock Sales, and Subsequent Company Performance.�Working Paper,
Arizona State University
[54] Luck, Christopher and Nancy Pilotte (1993): �Domini Social Index Performance.�Journal
of Investing 2, 60-62
[55] Lustig, Hanno, Chad Syverson and Stijn van Nieuwerburgh (2007): �IT, Corporate Pay-
outs, and Growing Inequality in Managerial Compensation.�Working Paper, University
of California, Los Angeles
[56] Manso, Gustavo (2007): �Motivating Innovation.�Working Paper, Massachusetts Institute
of Technology
[57] Markowitz, Harry (1959): �Portfolio Selection: E¢ cient Diversi�cation of Investments.�
Wiley, New York
34
[58] Mas, Alexandre (2007): �Labor Unrest and the Quality of Production: Evidence from the
Construction Equipment Resale Market.�Review of Economic Studies, forthcoming
[59] Maslow, Abraham (1943): �A Theory of Human Motivation.�Psychological Review 50,
370-396
[60] McGregor, Douglas (1960): �The Human Side of Enterprise.�McGraw-Hill, New York
[61] Moskowitz, Milt (1972): �Choosing Socially Responsible Stocks.�Business and Society 1,
71-75
[62] Narayanan, M. P. (1985): �Managerial Incentives for Short-Term Results.� Journal of
Finance 40, 1469-1484
[63] Newey, Whitney and Kenneth West (1987): �A Simple Positive-De�nite Heteroskedasticity
and Autocorrelation Consistent Covariance Matrix.�Econometrica 55, 703�708
[64] Pagano, Marco and Paolo Volpin (2005): �Managers, Workers, and Corporate Control.�
Journal of Finance 60, 843-870
[65] Perry-Smith, Jill and Terry Blum (2000): �Work-Family Human Resource Bundles and
Perceived Organizational Performance.�Academy of Management Journal 43, 1107-1117
[66] Porter, Michael (1992): �Capital Disadvantage: America�s Failing Capital Investment
System.�Harvard Business Review 70, 65-82
[67] Rajan, Raghuram and Luigi Zingales (1998): �Power in a Theory of the Firm.�Quarterly
Journal of Economics 113, 387-432
[68] Rajan, Raghuram and Luigi Zingales (2001): �The Firm as a Dedicated Hierarchy: A
Theory of the Origins and Growth of Firms.�Quarterly Journal of Economics 116, 805-
851
[69] Reuter, Jonathan and Eric Zitzewitz (2006): �Do Ads In�uence Editors? Advertising and
Bias in the Financial Media.�Quarterly Journal of Economics 121, 197-227
[70] Schröder, Michael (2007): �Is There a Di¤erence? The Performance Characteristics of SRI
Equity Indices.�Journal of Business Finance and Accounting 34, 331-348
[71] Shapiro, Carl and Joseph Stiglitz (1984): �Equilibrium Unemployment as a Worker Disci-
pline Device.�American Economic Review 74, 433-444
[72] Simon, Daniel and Jed DeVaro (2006): �Do the Best Companies to Work for Provide
Better Customer Satisfaction?�Managerial and Decision Economics 27, 667-683
[73] Stein, Jeremy (1988): �Takeover Threats and Managerial Myopia.� Journal of Political
Economy 46, 61-80
35
[74] Stein, Jeremy (1989): �E¢ cient Capital Markets, Ine¢ cient Firms: A Model of Myopic
Corporate Behavior.�Quarterly Journal of Economics 104, 655-669
[75] Taylor, Frederick (1911): �The Principles of Scienti�c Management.� Harper Brothers,
New York
[76] Teoh, Siew Hong, Ivo Welch and Paul Wazzan (1999): �The E¤ect of Socially Activist
Investment Policies on the Financial Markets: Evidence from the South African Boycott.�
Journal of Business 72, 35-89
[77] Tsao, Amy (2003): �A Showdown at the Checkout for Costco.�BusinessWeek, August 28,
2003
[78] Wermers, Russ (2004): �Is Money Really �Smart�? New Evidence on the Relation Between
Mutual Fund Flows, Manager Behavior, and Performance Persistence.�Working Paper,
University of Maryland
[79] Wernerfelt, Birger (1984): �The Resource-Based View of the Firm.�Strategic Management
Journal 5, 171-180.
[80] Yermack, David (2006): �Flights of Fancy: Corporate Jets, CEO Perquisites, and Inferior
Shareholder Returns.�Journal of Financial Economics 80, 211-242
[81] Zimmerman, Ann (2004): �Costco�s Dilemma: Be Kind To Its Workers, or Wall Street?�
Wall Street Journal, March 26, 2004
[82] Zhou, Xianming (2001): �Understanding the Determinants of Managerial Ownership and
the Link between Ownership and Performance: Comment.� Journal of Financial Eco-
nomics 62, 559-571
[83] Zingales, Luigi (2000): �In Search of New Foundations.�Journal of Finance 55, 1623-1653
36