The Impact of CEO Long-term Equity-based
Compensation Incentives on Economic
Growth in Collectivist versus Individualist
Countries

Cynthia J. Campbell
Iowa State University
Gerdin Business Building
2167 Union Drive
Ames, IA 50011–2063, stati Uniti
campcj@iastate.edu

Rosita P. Chang
University of Hawaii at Manoa
2404 Maile Way
Honolulu, HI 96822, stati Uniti
rositac@hawaii.edu

Jack C. DeJong, Jr.
Nova Southeastern University
3301 College Avenue
Ft. Lauderdale, FL 33314, stati Uniti
jd1315@nova.edu

Robert Doktor
University of Hawaii at Manoa
2500 Campus Road
Honolulu, HI NA 96822, stati Uniti
bobdoktor@gmail.com

Lars Oxelheim, Corresponding Author
School of Business and Law
University of Agder
Servicebox 422, N-4604
Kristiansand, Norway
lars.c.oxelheim@uia.no
Research Institute of Industrial Economics (IFN)
P.O. Box 55665, SE-10215
Stockholm, Sweden
lars.oxelheim@ifn.se
E
Knut Wicksell Center for Financial Studies
Lund University
P.O. Box 7080, S-220 07 Lund, Sweden
lars.oxelheim@fek.lu.se

Trond Randøy
School of Business and Law
University of Agder
Servicebox 422, N-4604
Kristiansand, Norway
trond.randoy@uia.no

Astratto
This study examines the impact of the prevalence of long-term equity-based chief executive officer
(CEO) compensation incentives on GDP growth, and we address the moderating role of individualist

Asian Economic Papers 15:2

C(cid:2) 2016 by the Earth Institute at Columbia University and the Massachusetts

Institute of Technology

doi:10.1162/ASEP_a_00432

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

versus collectivist cultures on this relationship. We argue that long-term incentives given to CEOs
in some firms may convey to other CEOs that they too may be able to receive such incentives and
rewards if they emulate the incentivized and rewarded CEOs. In a longitudinal study across 22 na-
tions over a 5-year period, we find that the higher proportion of CEOs in a country are awarded
long-term equity-based incentive compensation, the greater future real GDP growth, particularly in
collectivist countries.

1. introduzione

The 2008–09 global financial crisis triggered a debate about the use of equity-based incen-
tives in chief executive officer (CEO) compensation. One central argument against the use
of such incentive-based CEO compensation is that it appears to have only a small or neg-
ligible impact on firm-level performance (Vedere, per esempio., Murphy 1999; Oxelheim and Randøy
2005; Frydman and Saks 2010; Minnick, Unal, and Yang 2011), particularly when firms
have weak corporate governance (Qin 2012). We elevate the issue to the country level—
asking whether CEO incentives are good for country-level “performance.” A motivation
for this study is Bowe, Filatotchev, and Marshall’s (2010) call for research on how “differ-
ent national business traditions and cultural traits can interact with the workings of finan-
cial institutions, thereby moderating the relationship between governance, finance and
business strategy.” Specifically, we look at how “different national business traditions”
interact with the financial “institution” of long-term equity-based incentives in CEO com-
pensation, and how it is potentially moderated by “cultural traits” of collectivist versus
individualist cultures.

We argue that the finding of a lack of a strong positive effect between executive long-term
equity-based compensation incentives (cioè., stock options, restricted stock and perfor-
mance shares) and firm performance does not imply there is not a positive effect from
these incentive schemes on the economic growth of a country. Infatti, we make the case
that one can expect significant national variation in the effectiveness of long-term com-
pensation incentives, as such incentives are institutionally embedded at the national level
(Bruce, Buck, and Main 2005).

We see two main reasons for the existence of a positive national growth effect. Primo, In
response to the empirical support of a weak link between CEO incentives and firm perfor-
mance and the implicit implication that it does not spur economic growth, we stress the
need to explicitly address this issue in a broader theoretical approach involving country-
based effects. We propose that a considerable part of the overall economic benefits from
long-term equity-based compensation incentives in firms is their spillover effects to the
national level, as it leads to wealth-creating decision-making beyond the focal firms. Noi
argue that when a CEO in one firm has long-term compensation incentives this leads to

110

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

changes in the behavior of CEOs of rival firms, and leads to a positive spillover effect
on these other CEOs and their firms’ performance. Over time, the effects of the CEO in-
centives cascade through the economy as CEOs and/or aspiring CEOs are motivated to
make wealth-creating economic decisions. The economic rationale is that the higher the
percentages of CEOs with long-term incentive remuneration, the better the economic per-
formance at the national level. Specifically, we add to the literature by extending the argu-
ments from tournament theory (Lazear and Rosen 1981; Blanes-i-Vidal and Nossel 2011),
addressing within-firm effects of executive compensation, to include between-firm effects
creating wider overall economic benefits. We claim that between-firm effects can be seen as
a learning process by CEOs and/or aspiring CEOs in other firms, and thus label our new
theory vicarious agency.

Secondo, existing research has only partly addressed the international business nature of
top-management compensation incentives (Strange et al. 2009; Oxelheim and Randøy
2013). Hence, there is a need to address how executive incentives represent a means to en-
hance firm productivity in one country more than in other countries. Fundamentally, we
argue that top-management long-term equity-based compensation incentives have the po-
tential to change the common corporate culture of a country in line with the “moderniza-
tion theory” of Inglehart (1997). We propose, Tuttavia, that this effect is influenced by the
relative degree of individualism/collectivism of the national culture. The cultural context
of collectivism would typically lead to less focus on individual top management incen-
tives, but the impact of such incentives within the cultural norm of collectivism might be
bigger. Per esempio, in the United States such incentives are expected, and thus produce
less impact, whereas in China offering long-term equity-based executive compensation
incentives differentiates the firm in terms of wealth-creating corporate goals as well as
enhances the recruitment of talented executives. Inoltre, when new practices are in-
troduced in collectivist cultures, such as China, this can be expected to produce stronger
social influence across firms, as demonstrated by Li and Tang (2010).

The paper is organized as follows. In Section 2 we review the general issue of CEO incen-
tives and national growth. In Section 3 we present the hypothetical construct of vicarious
agenzia, exploring the relation of CEO long-term equity-based incentive compensation to
national economic prosperity. Sezione 4 reports predictive analysis relating CEO incentive
pay to national GDP growth in general, and investigates the impact of the relative na-
tional societal level of individualism/collectivism upon the many facets of this dynamic.
Sezione 5 is a discussion of the policy implications of our findings. Sezione 6 concludes the
paper and suggests future research.

2. CEO long-term compensation incentives and national growth

Agency theory focuses on how incentives work within the firm to align interests between
owners and managers. Managerial power theory presents another view of the linkage

111

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

between owners and managers, suggesting executive incentives are not necessarily struc-
tured to alter executive behavior in a firm-level wealth-creating direction (Bebchuk and
Fried 2003).

Standard agency theory predicts that CEO long-term performance compensation incen-
tives affect firm performance. A tenet of normative agency theory (Jensen and Meckling
1976) is that individual firms should index executive compensation to remove market-
wide effects. Relative performance evaluation posits executive compensation reward only
the part of firm performance for which the executive can claim some control and responsi-
bility. Executive incentive awards should not reflect the benefit or loss arising from over-
all financial market performance over which the executive has no control (Bertrand and
Mullainathan 2001). Inoltre, managerial long-term compensation incentives are
more effective at inducing wealth-creating decisions when formulated in line with the
economic psychology of incentives—specifically, considering how executives prefer less
risky, less uncertain, and more immediate forms of rewards (Pepper and Gore 2014).

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

The literature addressing relative performance measurement suggests that CEOs per-
ceived to be highly talented are in demand by competing firms and as a consequence
accrue higher wages from their own firms than might be justified by firm performance
under their leadership (Rajgopal, Shevlin and Zamora 2006). Chang, Dasgupta and Hilary
(2010), Tuttavia, demonstrate that CEO compensation may reflect CEOs’ abilities and be
correlated with firm performance.

Few studies have analyzed the between-firm effects of incentive pay practices. One ex-
ception is the ratchet effect (Freixas, Guesnerie, and Tirole 1985) extended to the area of
executive pay, making the case that one executive’s pay is affected by other executives’
pay—typically within the same executive labor market—which we argue is mostly con-
fined within a country even if it is in the same industry. Similar to managerial power
theory, the ratchet effect is driven by “what the CEO can get” (managerial power at the
between-firm level) rather than how it affects motivation and wealth-creating incentive
alignment between managers and owners at the firm level (the agency theory argument in
general and the tournament theory in particular).

The ratchet effect represents a misanthropic view with a race to the top of compensation
without any related increase of efforts. Kuhnen and Tymula (2012) offer insight into the
ratchet effect in executive pay through a behavioral economic simulation. They demon-
strate that feedback may induce a ratcheting effect in productivity, mainly because of the
competition to be at the top of the rank hierarchy.

A natural experiment of the bidding-up or ratchet effect occurred in 2005. German author-
ities, following International Financial Reporting Standards, changed regulations, forcing

112

Asian Economic Papers

CEO Pay and National Economic Growth

Figura 1. Conceptual categorization of executive pay theories and proposed economic
growth effects

German-listed firms to disclose each executive member’s pay. An empirical analysis done
by Stadtmann and Wissmann (2008) confirms such pay increase effects from increased
pay transparency in Germany. These findings open the analysis to consideration of the
determination of the CEO’s compensation not only with respect to their own firm’s per-
formance, but also to the behavior of competing firms, or even to the overall economy
(Oxelheim, Wihlborg, and Zhang 2012).

We suggest that existing executive pay studies are too narrow in their theoretical scope
to capture effects at the national level. To visualize the contribution, we conceptualize
in Figure 1 executive pay theory in relation to two dimensions: level of analysis and the
extent that executive incentives contribute to national economic growth.

The underlying implication for country-level economic growth is that appropriate long-
term equity-based CEO incentives (with relative performance evaluation) should be good
for economic growth. Specific to managerial incentives, Bloom and Van Reenen (2010)
examine CEO incentive pay and firms’ productivity (a broader definition of firm perfor-
mance than equity price performance) and find a positive relationship. The results hold
only among some studies of small groups of firms, Tuttavia. Even though the positive re-
lationship between firm-level CEO incentives and firm performance is relatively weak, COME
long as a relationship exists, higher firm performance can have direct and indirect positive
effects on national prosperity.1 In contrast, incentive awards based on CEOs’ serendip-
itous good fortune, namely, luck (Oyer 2004), managerial power (Qin 2012), or ability
to skim corporate wealth through poor governance (Bertrand and Mullainathan 2001),
should be bad for overall economic growth.

The difference between the agency and managerial power theories including the between-
firm ratchet effect can be seen as philosophical. The agency theory reflects a philanthropic

1 This should hold even in the absence of non-competitive firm behavior.

113

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

view when it comes to wealth-creating efforts as a result of incentives, and the manage-
rial power theories are misanthropic, not leading to wealth-creating efforts as a result
of incentives.

Although the agency and tournament theories’ argument for motivation and incentives
through compensation is typically in a within-firm context, the vicarious agency theory
presented in the next section considers the between-firm motivational effects and the
spill-over impact of executive compensation incentives in other firms, particularly within
the same country.

3. Hypothetical construct of vicarious agency

The relation between CEO long-term equity-based incentives and the financial perfor-
mance of a CEO’s firm has been found by past research to have a weak link; nevertheless,
this study posits that aggregate firm performance within a country may be enhanced by
ubiquitous implementation of CEO pay incentives within that country.

The dynamics of the relation between the relative frequency of CEO equity-based incen-
tive pay and aggregate firm performance within the same country are complex. We argue
that there is a link between a CEO’s motivation to enhance the financial performance
of their firm and the carrot of future incentives and other rewards for having done so.
Accordingly, we hypothesize that when CEO “A” sees CEO “B” richly rewarded by the
pleased board of directors, then CEO “A,” believing that they may be similarly rewarded,
may be motivated to work harder learning from the rewarded CEO’s behavior, so as
to please their own board and thereby receive similar rewards as those achieved by
CEO “B.”

The philosophical view on human behavior here is similar to the one adopted in tourna-
ment theory. “Vicarious agency” is the term we choose for the between-firm context to de-
scribe the dynamics of CEO “A” being motivated as a consequence of observing CEO “B”
receiving a large compensation bonus. In so doing, we borrow the term “agency” from
classic economic agency theory (Jensen and Meckling 1976) and the term “vicarious” from
classic social learning theory in psychology (Bandura 1977). The notion is that CEOs learn
vicariously by observing the consequences of other CEOs’ behavior. Bandura’s concept
of social learning theory posits that individuals learn from observing the consequences
of other individuals’ behavior. We argue that CEOs, and/or potential CEOs, having ob-
served rewarded CEOs, engage in imitation and behavioral modeling to seek similar re-
wards. By combining the dynamics of agency theory and social learning theory, we arrive
at our hypothetical construct: vicarious agency.

Within a national network, incentive rewards to one CEO may have cascading effects
through vicarious agency, motivating numerous other CEOs to work with greater vigor

114

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

in pursuit of enhancing their firms’ performance. The overall enhanced motivation of
multiple CEOs will thus result in overall improvement of national prosperity. Così, our
hypothesis is somewhat akin to what happens in a chain reaction in physics. Two or three
CEOs see rewards given to another CEO. They then enhance their efforts in hopes of sim-
ilar rewards. Some succeed, E, in turn, their rewards are noted by other CEOs, who
then enhance their efforts, and the chain reaction grows. Not all CEOs succeed and not all
firms prosper, but more do prosper than might have if the incentives did not exist. IL
chain reaction reaches a critical mass, and the nation prospers overall. This is similar to
concepts of dissemination of board of director practices observed by Bouwman (2011).

To summarize, we posit that the existence of CEO long-term equity-based incentive com-
pensation not only influences firms’ executives who receive these incentives, but also mo-
tivates executives in competing firms, enhancing the competitive rivalry, and producing
economic prosperity as a whole. If true, one expects to find a positive relation between the
frequency of the use of such long-term equity-based compensation incentives in an econ-
omy and economic growth of that economy in the following time period.2 This is also
supported by past research that highlights the complex relation between CEO compen-
sation in one firm and the behavior of competing firms and the actors within those other
firms (per esempio., Rajgopal, Shevlin, and Zamora 2006). Based on this argument we formulate
Hypothesis 1.

Hypothesis 1: There is a positive relation between the prevalence of CEOs receiving long-term
equity-based compensation incentives in a nation and the real GDP growth of that nation.

Highlighting the importance of specific cultural dimensions to global business, Hofst-
ede (1980, 2001) provides the research community with an initial set of four dimensions
of cultural variation: power/distance, uncertainty/avoidance, masculinity/femininity,
and individualism/collectivism (I/C). Of these, the dimension sparking the greatest
interest among researchers in international business has been I/C (per esempio., Bochner 1994;
Triandis 1995; Li and Tang 2013; Gray, Kang, and Yoo 2013). Given the importance and
research legacy of the I/C cultural dimension in international business research, we inves-
tigate links this dimension may have on the theoretical construct of vicarious agency and
Hypothesis 1. From Hofstede’s (2001) catalogue of I/C research findings:

(1) The more individualist societies are, in general, the wealthier they are compared with

more collectivist societies.

2 Please note, we use frequency of stock options, restricted stock and performance shares incen-

tives, not the monetary amount of incentives. Excessive incentives are not argued to be better than
moderate incentives.

115

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

(2) Individualist societies have economies based on individual interests, whereas collec-

tivist societies’ economies are based on collective interests.

(3) Collectivist societies prefer reward allocation based on equality for in-group and eq-
uity for out-group, whereas individualist societies prefer reward allocation based on
equity for all.

(4) Treating friends better than others is normal and ethical in collectivist societies, al-

though such favoritism is considered unethical in individualist societies.

(5) In collectivist societies, personal relationships prevail over task and company in

business dealings, whereas company and task prevail over personal relationships in
business dealings in individualist societies.

(6) Collectivist societies exhibit large differences in the distribution of wealth,

whereas wealth is distributed more equally across sectors of the economy in
individualist societies.

These six I/C points can be used to understand the implications of this cultural dimension
for our theoretical model of vicarious agency. Items 1 E 2 imply that societies with a
strong individualist orientation are more likely to result in outcomes supporting Hypoth-
esis 1, but items 3 through 6 support the notion that collectivist societies are more likely to
have a greater effect from the dynamics of vicarious agency in Hypothesis 1.

Li and Tang (2013) report that executive hubris is stronger in the more collectivist Chinese
society than in the more individualist society of the United States. Executive hubris is de-
fined as an executive’s exaggerated self-confidence or pride (Hiller and Hambrick 2005).
The cultural context of collectivism may lead to situations where social influences are
more likely to condone the manifestation of executive hubris in business decisions involv-
ing diverse strategic corporate decisions, from excessive premiums paid for acquisitions
to greater acceptance of CEO long-term equity-based incentive compensation.

We consider the role of the in-group network of CEOs fundamental to the dynamics of
vicarious agency. The social influences surrounding in-group networking among CEOs,
be it through family ties, social settings, on the golf course, or as members of each other’s
board of directors, imply that CEO long-term incentives occur more easily, frequently,
and with less adverse societal reaction in collectivist cultures. This leads us to impart
greater significance to Hofstede’s catalogue of points 3 E 4. Hofstede’s point 3 ar-
gues for more restrictive utilization of CEO incentives in collectivist societies such that
they are limited to members of in-groups, such as top executives. Hence we formulate
Hypothesis 2.

Hypothesis 2: There is a stronger relation between the prevalence of CEO long-term
equity-based compensation incentives and real GDP growth in collectivist cultures than in
individualist cultures.

116

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

4. Empirical results

4.1 Variables and data

The dependent variable is the percentage of real GDP growth (GDPGR) reported by the
International Monetary Fund’s World Economic Outlook. GDPGR is in constant (inflation
adjusted) national currency.

We have two explanatory variables. First is the percentage of firms in a country provid-
ing long-term equity-based incentive remuneration in the form of stock options, restricted
stock, and performance shares with typical vesting periods of three to five years, to their
CEOs (LTIREM). Data are obtained from surveys by Towers Perrin between 2001 E
2005 for 22 countries. Towers Perrin surveyed their clients for the years 2001, 2004, E
2005. Their clients were mainly large firms; hence, our study has a potential bias toward
large firms. Because the Towers Perrin data on long-term equity-based incentive plans is
missing for the years 2002 E 2003, we observe the country-by-country trends and ap-
proximate the missing values by interpolation.3 Towers Perrin stopped providing the
survey data after 2005, restricting our study period to 2001–05, except that the dependent
variable GDPGR stops in 2006 (T + 1). Our second explanatory variable is the cultural di-
mension of I/C as measured by Hofstede’s individualism index value (INDIVCOLLECT).
The data source is Hofstede’s Cultures Consequences (1980, 2001).

Based on past research on economic growth, we identify four control variables to re-
duce the risk of making incorrect inferences. Primo, we include investment to real GDP per
capita in 2005 constant prices (INVRATIO). This is the only variable found in prior studies
of economic growth that survives specification bias in an extreme bound analysis setting
(Levine and Renelt 1992). These data are from Penn World Table Version 6.3. (Heston,
Summers, and Aten 2009). The second control variable is the Fraser Institute freedom
index (LEGAL). It measures the country’s legal system and security of property rights,
and is associated with economic growth (per esempio., Goldsmith 1995). The third control variable
is the Fraser Institute freedom index measuring the country’s labor market regulation
(LABOR). Both LEGAL and LABOR are measured on a zero-to-ten scale, with higher
values indicating a greater degree of freedom. These two variables are motivated by the

3 In analyzing the Towers Perrin data for 2001, 2004, E 2005 it is clear that the percentage of firms
providing long-term equity-based incentive pay to their CEOs does not vary wildly from year to
year, but changes slowly and gradually; clear trends are visible from 2001 A 2004 on a country
by country basis. Using interpolation assumes a linear trend from 2001 A 2004, which keeps our
errors of approximation small, since most countries experience only a 5–20 percent change from
2001 A 2004. Without interpolation, our estimators will be less efficient and the standard errors
larger as the information from 2002 E 2003 regarding the covariance between our independent
variables investment’s share of real GDP per capital in 2005 constant prices, legal structures, E
labor market regulations and the dependent variable, real growth in GDP, is lost.

117

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

research initiative of the Canadian Fraser Institute.4 The fourth control variable is the
population’s general education. Following Mork, Wolfenson, and Yeung (2004) in their
study of the relation between economic growth and billionaire innovators, we control for
each country’s general education level (EDUCATION). It is the average years of schooling
for the population age 15 years and older. We also control for other unobservable country
effects via the use of country random disturbance terms in our random effects models.

Tavolo 1 reports descriptive statistics. Descriptive statistics for LTIREM are for the years
data are provided by Towers Perrin: 2001, 2004, E 2005. We do not use interpolated
data for the years 2002 E 2003 in the descriptive statistics. The GDPGR statistics are
presented for years 2002 through 2006, because our hypotheses relate LTIREM in year
t affecting GDPGR in year t + 1. Lagging LTIREM reduces the problem of endogene-
ity to focus on the impact long-term equity-based executive incentive pay has on real
GDP growth, rather than focusing on whether high rates of economic growth cause more
firms to use such incentive pay. INDIVCOLLECT, INVRATIO, EDUCATION, and both
Fraser Institute freedom indices, LEGAL and LABOR, are measured for the years 2002
through 2006.

We examine the variables on a country-by-country basis, as well as subsamples of 11 indi-
vidualist countries and 11 collectivist countries. The average annual real GDP growth rate
(GDPGR) ranges from 10.1 percent in China to 0.7 percent in Italy. The mean percentage
of firms with long-term equity-based incentive remuneration ranges from 100 percent in
Canada and 96.7 percent in the United States, the United Kingdom (UK), and the Nether-
lands to 16.7 percent in India. The mean investment to real GDP per capita, INVRATIO,
ranges from 10.3 percent in South Africa to 43.0 percent in South Korea. For the zero-to-
ten scale of the Fraser Institute Freedom Indices, LEGAL ranges from an average score of
3.86 for Argentina to 8.95 for the Netherlands, and LABOR ranges from an average score
Di 3.39 for Germany to 8.51 for Hong Kong. The average years of total schooling ranges
from a low of 4.59 years for India to a high of 12.87 years for the United States. Hofstede’s
individualism index value (INDIVCOLLECT) ranges from a low of 18 for South Korea to
a high of 91 for the United States. Over the whole 22-country sample, the mean percent-
age of firms with long-term equity-based incentive remuneration averages 68.7 per cento,
and the mean annual real GDP growth rate averages 3.5 percent and the Hofstede indi-
vidualism index value averages 57.50. For the economic control variables, INVRATIO
medie 26.8 per cento, LEGAL averages 7.25, LABOR averages 5.65, and EDUCATION
medie 9.69 years.

4 We follow a common research tradition to control for indices of this kind, making explicit that

the validity is conditioned upon an assumption about equidistance. See Canadian Fraser Institute
(http://www.freetheworld.com/).

118

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

9

5

9

.

.

.

3
2

8

4
1

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

0
9

5
7

0
8

1
7

7
6

6
7

0
8

1
7

8
6

9
8

1
9

6
4

8
3

0
2

5
2

8
4

6
4

0
3

0
2

5
6

8
1

1
5

4
0

9
0

8
2

3
1

9
4

1
1

3
0

2
1

4
0

3
1

6
0

2
1

2
2

6
1

9
1

5
1

1
1

4
2

4
1

7
1

3
1

4
1

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

6
8

0
4

6
8

3
8

9
3

8
0

0
8

0
4

3
6

6
2

7
8

4
0

0
0

2
5

4
7

9
5

9
1

3
2

0
4

3
1

8
3

6
6

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

1
1

0
1

1
1

9

1
1

9

0
1

1
1

9

9

2
1

9

7

7

9

4

1
1

8

8

8

9

1
1

8
3

5
2

4
3

2
3

7
7

9
0

9
5

6
6

9
7

8
3

4
5

5
6

0
5

7
5

2
2

8
5

0
6

6
6

1
1

6
2

4
7

1
5

.

.

.

.

.

.

.

.

.

.

.

0

0

0

0

0

1

0

0

0

0

0

.

.

.

.

.

.

.

.

.

.

.

0

0

0

0

0

0

0

1

0

0

0

9

2

.

.

.

1

1

1

7

8

6

.

.

.

9

0
1

3
4

9
4

.

.

.

1

1

1

4
5

3
8

5
8

1
3

9
3

8
5

8
5

2
8

2
1

9
9

2
7

6
4

7
3

4
2

1
5

6
7

3
7

8
1

4
7

2
5

7
6

1
3

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

6

4

6

5

3

4

5

3

7

6

7

4

4

4

8

5

6

5

6

5

4

5

7

7

7

.

.

.

5

5

5

7
1

2
3

6
1

1
1

4
1

0
7

7
2

6
4

7
1

2
3

2
2

1
6

3
4

4
4

9
3

0
3

5
4

9
7

1
2

8
2

6
5

8
2

5

9

4

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

1

0

6
8

3
5

6
4

3
5

0
8

9
2

5
9

1
7

5
7

2
8

4
8

6
8

6
8

4
4

8
6

3
2

4
7

3
7

6
3

6
7

1
7

4
6

.

.

.

.

.

.

.

.

.

.

.

8

7

8

7

8

6

8

8

8

8

7

.

.

.

.

.

.

.

.

.

.

.

3

4

5

7

6

7

4

8

6

6

6

3

2

.

.

.

7

8

6

7

9

2

7

8

4

1

6

2

4

7

.

.

.

.

.

.

.

.

.

.

.

2

0

0

0

1

0

0

.

.

.

.

.

.

.

.

.

.

.

8

9

8

4

6

2

0

7
2

6
2

3
2

1
3

5
2

2
2

1
3

0

0

3

8

2

9

8

3

9

3

0

.

.

.

.

.

.

.

.

.

.

.

3

1

1

1

2

0

0

5

0

0

1

8

5

4

8

4

9

9

3

3

0

7

.

.

.

.

.

.

.

.

.

.

.

5
1

3
1

2
3

6
2

8
1

2
3

4
2

1
3

0
1

3
4

5
3

)
1
1
=
N

(

5

9

9

.

.

.

7

3

9

8

7

9

.

.

.

7
2

5
2

.

.

.

.

.

.

.

.

.

.

.

5

6

2

8

2

2

5

0

0

1

0

1

1

1

8

7

7

9

0

9

7

4

9

4

5

.

.

.

.

.

.

.

.

.

.

.

1

0

2

2

0

1

2

0

1

2

6

.

.

.

3

1

5

0

0

.

.

.

7
5

8
7

0
0
0
0

9
8
9

0

4
1
3

8
4
7

7
1
−

.

.

.

.

7
3

9

8

7
3

1

3

3

0

5

8

0
0
0
0

0

0
0
0
0

9
0
8
6

0

0
0
0
0

0

4
1
0
2

0

0
0
0
0

0

0
0
0
0

0

0
0
0
0

8
6
2

0
6
1

7
−

.

.

.

.

0

1
5

0

9
0
8
6

0
7
1

2
1
4

.

.

.

.

0

0
−

0

0
4
7

0
0
0
0

0

4
9
3

6
7

.

.

.

.

8
−

S
e
io
R
T
N
tu
o
C

4
1
0
2

2
5
6

5
8
2

.

.

.

.

0

1

1
−

0
0
0
0

8
3
6

5

0

7
8
7

1
3

T
S
io
v
io
T
C
e
l
l
o
C

D
N
UN

T
S
io
l
UN
tu
D
io
v
io
D
N
io

N
e
e
w
T
e
B

N
UN
e
M

F
o

sì
T
io
l
UN
tu
q
e

S
e
io
R
T
N
tu
o
C

)
1
1
=
N

(

T
S
io
v
io
T
C
e
l
l
o
C

D
N
UN

,
)
1
1
=
N

(

T
S
io
l
UN
tu
D
io
v
io
D
N
io

,
)
2
2
=
N

(

6
2

1

3

7

0

3

3

5

1
1

7

7

9

7

4

0

7

2

1

6

6

7

8

1

3

8

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

1

2

1

0

0

1

3

1

5

7

1

2

6

4

4

5

1

.

.

.

2

3

0
1

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

5

0

9

2

9

8

0

9

0

2

1
1

3
1

8
1

5

5

8
1

2

6

0

8

5

8

0

7

0

1
1

3
1

7

5
1

5

1
1

1
2

0
1

8

5

1

.

.

.

5
2

4
1

1
2

3

3

0

3

0

7

7

0

7

7

7

.

.

.

.

.

.

.

.

.

.

.

3
9

5
7

1
7

6
9

5
6

1
8

6
9

6
9

8
8

8
8

0
0
1

3

0

7

0

7

3

0

0

0

0

3

.

.

.

.

.

.

.

.

.

.

.

3
5

5
5

6
2

5
6

6
1

8
2

0
4

0
8

5
6

5
6

3
6

sì
R
T
N
tu
o
C

:

UN

l
e
N
UN
P

UN
io
l
UN
R
T
S
tu
UN

M
tu
io
G
l
e
B

UN
D
UN
N
UN
C

e
C
N
UN
R
F

sì
N
UN
M
R
e
G

sì
l
UN
T
IO

S
D
N
UN
l
R
e
H
T
e
N

D
N
UN
l
R
e
z
T
io

w
S

N
e
D
e
w
S

.

.

K
U

.

.

S
U

R
UN
S

,

G
N
o
K
G
N
o
H

UN
N
io
T
N
e
G
R
UN

l
io
z
UN
R
B

UN
N
io
H
C

UN
C
io
R
F
UN
H
T
tu
o
S

UN
e
R
o
K
H
T
tu
o
S

e
R
o
P
UN
G
N
io
S

o
C
io
X
e
M

N
io
UN
P
S

UN
io
D
N
IO

N
UN
P
UN
J

7

7

8

.

.

.

8
6

6
8

0
5

e
l
P
M
UN
S

e
l
o
H
W

:

C

l
e
N
UN
P

e
l
P
M
UN
S

e
l
o
H
W

T
S
io
l
UN
tu
D
io
v
io
D
N
IO

T
S
io
v
io
T
C
e
l
l
o
C

0
0
0
0

0

8
1
2

3
6

1
5
9

7
−

.

.

.

.

F
o

S
T
S
e
T

:

D

l
e
N
UN
P

T
S
e
T
–
F
UN
V
O
N
UN

e
tu
l
UN
v
–
P

e
tu
l
UN
v

T
S
e
T
–
T

e
tu
l
UN
v
–
P

0

0

2

0

3
2

7
2

6

8

0

0

6

7

2

2

9

9

2

2

8

9

4

5

5
6
≤
T
C
E
l
l
O
C
V
IO
D
N

IO

H
T
io

w
S
e
io
R
T
N
tu
o
C

T
S
io
v
io
T
C
e
l
l
o
C

:

B

l
e
N
UN
P

2

0

5

1

3
3

9
2

6

9

0

0

3

9

3

9

2

)
1
1
=
N

(

5
6
>
T
C
E
l
l
O
C
V
IO
D
N

IO

H
T
io

w
S
e
io
R
T
N
tu
o
C

T
S
io
l
UN
tu
D
io
v
io
D
N
IO

.

v
e
D

.

D
T
S

N
UN
e
M

.

v
e
D

.

D
T
S

N
UN
e
M

.

v
e
D

.

D
T
S

N
UN
e
M

.

v
e
D

.

D
T
S

N
UN
e
M

.

v
e
D

.

D
T
S

N
UN
e
M

.

v
e
D

.

D
T
S

N
UN
e

M

.

v
e
D

.

D
T
S

N
UN
e

M

T
C
E
l
l
O
C
V
IO
D
N

IO

N
O
IO
T
UN
C
U
D
E

R
O
B
UN
l

l
UN
G
E
l

)

%

(

O
IO
T
UN
R
V
N

IO

)

%

(

R
G
P
D
G

)

%

(

M
E
R
IO
T
l

S
C
io
T
S
io
T
UN
T
S

C
io

M
o
N
o
C
e
D
N
UN
N
o
io
T
UN
S
N
e
P
M
o
C

e
v
io
T
N
e
C
N
io
D
e
S
UN
B
–
sì
T
io
tu
q
e
M
R
e
T
–
G
N
o
l
N
o

S
C
io
T
S
io
T
UN
T
S

e
v
io
T
P
io
R
C
S
e
D

.
1

e
l
B
UN
T

l
l
e
w

S
UN

S
e
io
R
T
N
tu
o
C

2
2

e
H
T

F
o

H
C
UN
e

R
o
F

)
e
tu
l
UN
v

X
e
D
N
io

M

S
io
l
UN
tu
D
io
v
io
D
N
io

S
’
e
D
e
T
S
F
o
H

(

T
C
E
l
l
O
C
V
D
N

IO

IO

D
N
UN

,
)
P
tu

D
N
UN

sì

5
1

e
G
UN

N
o
io
T
UN
l
tu
P
o
P

e
H
T

R
o
F

G
N
io
l
o
o
H
C
S

l
UN
T
o
T

F
o

S
R
UN
e
sì

e
G
UN
R
e
v
UN
(

N
O
IO
T
UN
C
U
D
E

,
)
X
e
D
N

io

N
o
io
T
UN
l

tu
G
e
R

.
6
0
–
2
0
0
2
M
o
R
F

S
io

)
5
0
0
2

D
N
UN

4
0
0
2

,

1
0
0
2
(

M
E
R
IO
T
l

T
P
e
C
X
e

S
e
l
B
UN
io
R
UN
v

e
H
T

l
l
UN

R
o
F

D
o
io
R
e
P

e

M

io
T

e
H
T

.
S
e
io
R
T
N
tu
o
C

T
S
io
v
io
T
C
e
l
l
o
C

1
1

e
H
T

D
N
UN

,
S
e
io
R
T
N
tu
o
C

T
S
io
l
UN
tu
D
io
v
io
D
N
io

1
1

e
H
T

,
S
P
tu
o
R
G
B
tu
S

D
N
UN

e
l
P
M
UN
S

e
l
o
H
w

e
H
T

R
o
F

S
UN

T
e
k
R
UN
M
R
o
B
UN
l

e
T
tu
T
io
T
S
N
IO

R
e
S
UN
R
F
(

R
O
B
UN
l

,
)
X
e
D
N
io

e
R
tu
T
C
tu
R
T
S

l
UN
G
e
l

e
T
tu
T
io
T
S
N
IO

R
e
S
UN
R
F
(

l
UN
G
E
l

,
)
UN
T
io
P
UN
C

R
e
P

P
D
G

l
UN
e
R

F
o

e
R
UN
H
S

T
N
e
M

T
S
e
v
N
io
(

O
IO
T
UN
R
V
N

IO

,
)
H
T
w
o
R
G

P
D
G

l
UN
e
R

l
UN
tu
N
N
UN
(

R
G
P
D
G

,
)
S
e
R
UN
H
S

e
C
N
UN
M
R
o
F

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

–
R
e
P

D
N
UN

k
C
o
T
S

D
e
T
C
io
R
T
S
e
R

,
S
N
o
io
T
P
o

k
C
o
T
S

,
.
e
.
io

,

N
o
io
T
UN
R
e
N
tu
M

e
R

e
v
io
T
N
e
C
N
io

D
e
S
UN
B
–
sì
T
io
tu
q
e

M

R
e
T
–
G
N
o
l

H
T
io

w

S

M

R
fi

F
o

T
N
e
C
R
e
P
(

M
E
R
IO
T
l

S
e
l
B
UN
io
R
UN
v

F
o

N
o
io
T
UN
io
v
e
D

D
R
UN
D
N
UN
T
S

D
N
UN

N
UN
e
M
e
H
T

S
T
N
e
S
e
R
P

e
l
B
UN
T

S
io
H
T

:
S
e
T
o
N

.
S
N
o
io
T
UN
l
tu
C
l
UN
C

’
S
R
o
H
T
tu
UN

:
e
C
R
tu
o
S

119

Asian Economic Papers

CEO Pay and National Economic Growth

The descriptive statistics indicate that firms in the 11 individualist countries are statisti-
cally different from those in the 11 collectivist countries, as seen in panel D of Table 1. For
the individualist relative to the collectivist countries we find, at the 1 percent statistical
significance level, that the percentage of firms with long-term equity-based executive in-
centive pay is higher, the real GDP growth rate is lower, the degree of freedom afforded
by the legal structure and security of property rights is higher, the average years of to-
tal schooling is higher, and of course the Hofstede individualism index value is higher.
Not statistically significantly different between the collectivist and individualist coun-
tries are the per capita investment to GDP ratio (INVRATIO) and labor market regulation
index (LABOR).

4.2 Methodology

To test Hypothesis 1, we use panel data and perform a cross-sectional country-based time
series regression with random effects. We relate the impact of long-term equity-based
CEO incentive pay in one year on the real rate of growth in GDP in the following year
as well as the modifying effect of a country’s measure of I/C on the relation between
long-term equity-based incentive pay for CEOs and the real rate of economic growth.
Because Hofstede’s INDIVCOLLECT values do not change over time for our 22 countries,
we are precluded from using a fixed effects model as the INDIVCOLLECT values capture
some of the differences between the countries.5 Alternatively, we utilize a random effects
modello, which includes a random disturbance term, μ
io , representing a collection of factors
omitted in our model that are specific to each country. These random disturbance terms
are constant through time and enable us to pool the results from the various countries and
focus on the relation between lagged LTIREM, GDPGR, and INDIVCOLLECT while re-
ducing the bias that would otherwise result in our coefficients’ estimates being influenced
by unmeasured variables that are correlated with the GDP growth rate and would distort
the between country variability. To reduce the risk of miss-specifying our model by omit-
ting any economic variables, we include the four control variables: INVRATIO, LEGAL,
LABOR, and EDU (the natural log of EDUCATION). The model used is:

GDPGRit

= γ
0

+ γ

1LTIREMi(t−1)

+ γ

2INVRATIOit

+ γ

3LEGALit

+ γ

4LABORit

+ γ

5EDUit

+ γ

6INDIVCOLLECTit

+ μ
io

+ ε
Esso

(1)

where i represents the ith country, t denotes years 2002 through 2006, and the μ
the random disturbance terms for country i.

i terms are

To test Hypothesis 2, we use our model in equation (1) with the sample split into two
even groups by Hofstede’s individualism index value; the subsamples consist of 11

5 Including both fixed-effects dummy variables and Hofstede’s INDIVCOLLECT in the same regres-

sion results in perfect multicollinearity.

120

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

Tavolo 2. Support of Hypothesis 1
GDPGRit = γ0 + γ1LTIREMi(t−1)

+ γ2INVRATIOit + γ3LEGALit

+ γ4LABORit + γ5EDUit + γ

6INDIVCOLLECTit + μi + εit

Panel least squares regression results on GDP growth

All countries (N = 22)
Intercept

LTIREM(−1)

INVRATIO

LEGAL

LABOR

LN(EDUCATION)

INDIVCOLLECT

No. of panel observations
Adjusted R2
F-statistic
Durbin Watson statistic

0.1736
(3.8176)∗∗∗
0.0642
(3.2085)∗∗∗
0.1225
(2.0856)∗∗
−0.0007
(−0.2172)
0.0014
(0.5733)
−0.0794
(−3.1835)∗∗∗
−0.0007
(−3.0895)
110

∗∗∗

0.1575
4.3956∗∗∗
1.4221∗∗

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Notes: This table presents the panel least squares regression results using random effects for

Tutto 22 countries. The countries include Argentina, Australia, Belgium, Brasile, Canada, China,

France, Germany, Hong Kong, India, Italy, Japan, Mexico, Netherlands, Singapore, South

Africa, South Korea, Spain, Sweden, Svizzera, UK, and the United States. All regressions

are run with five years of data: 2001–05 if lagged, and 2002–06 if not lagged.

Numbers in parentheses are t-statistics.
∗∗

Statistically significant at α = 5 percent level and

∗∗∗

statistically significant at α = 1

percent level. The cross-section random effects are not reported as they sum to zero and are

not relevant to the discussions.

collectivist countries with INDIVCOLLECT scores less than or equal to 65 E 11 In-
dividualist countries with INDIVCOLLECT scores greater than 65. Because Hofstede’s
INDIVCOLLECT values do not change over time for our 22 countries, IL 11 collectivist
countries and 11 individualist countries are consistent throughout our five-year sample.
Bifurcating our sample into collectivist and individualist countries illuminates the im-
pact of Hofstede’s individualism index value on the direction and magnitude of the im-
pact of long-term equity-based incentive remuneration for CEOs on a country’s real GDP
growth rate.

4.3 Empirical results for Hypothesis 1

Tavolo 2 reports the results from equation (1). Our results indicate that long-term compen-
sation incentives for executives enhance national economic prosperity. In our 22-nation
sample we find a positive and statistically significant (at the 1 percent level) relation be-
tween the prevalence of long-term incentives for higher-level executives and national
economic prosperity.6 The regression coefficient of 0.0642 on lagged LTIREM also

6 Using clustered standard errors our results are qualitatively the same because of the small number

of years in our panel regressions.

121

Asian Economic Papers

CEO Pay and National Economic Growth

indicates economic significance; if the proportion of firms in a country using long-term
equity-based incentives for CEOs increases from 50 percent to 60 percent in one year, we
expect to see an average increase of 0.64 percent in the annual growth rate of real GDP
in the following year.7 The regression coefficient on INVRATIO is positive and large in
magnitude with significance at the 5 percent level. The regression coefficients on the two
freedom indices, LEGAL and LABOR, are statistically insignificant, but the regression
coefficients on both LN(EDUCATION) or EDU and INDIVCOLLECT are negative and
statistically significant at the 1 percent level. Our results support Hypothesis 1 in that
there is a significant positive relation between the percentage of firms offering CEO incen-
tive compensation in year t and GDP growth rate in the subsequent year across our global
sample of 22 nazioni.

4.4 Causality issues

The question remains whether our results support Hypothesis 1 or are simply a conse-
quence of the fact that enhanced GDP may cause greater frequency of CEO incentive
compensation. Such a reverse effect could be because in good economic times firms
have greater slack and are therefore more able and likely to offer compensation incen-
tives to their top executives. Ovviamente, our use of lagged LTIREM relative to GDPGR in
equation (1) helps to mediate against real GDP growth causing increased frequency of
CEO incentive remuneration. Nevertheless, to address this we alter our model in equa-
zione (1) to reverse the relation and regress the frequency of firms providing long-term
incentive remuneration to their CEOs on lagged real GDP growth including our other
economic control variables. Our model of equation (1) now becomes:

LTIREMit

= γ
0

+ γ

1GDPGRi(t−1)

+ γ

2INVRATIOit

+ γ

3LEGALit

+ γ

4LABORit

+ γ

5EDUit

+ γ

6INDIVCOLLECTit

+ μ
io

+ ε
Esso

(2)

The random effects model of Equation (2) enables us to pool the results from the various
countries and focus on the relation between lagged GDPGR and LTIREM while reducing
the bias that would otherwise result in our coefficients’ estimates from unmeasured vari-
ables that are correlated with the frequency of long-term CEO incentive pay and distort
the variability across countries. Our variable definitions are identical to our earlier model,
except that t now denotes years 2001 through 2005 because LTIREM is no longer lagged.

7 Our model is robust to the inclusion of alternative control variables, including the annual growth
rate of newly invested capital from Datastream, the Fraser Institute’s economic freedom index,
and Transparency International’s corruption perceptions index. Our unreported results find a
positive and statistically significant relation between lagged long-term equity-based incentive pay
for CEOs and the real GDP growth rate for all country samples and for all combinations including
alternative control variables.

122

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

Tavolo 3. Support of one-way causality in supporting Hypothesis 1.
Panel least squares regression results on long-term equity-based
incentive compensation

All countries (N = 22)

Intercept

GDPGR(−1)

INVRATIO

LEGAL

LABOR

LN(EDUCATION)

INDIVCOLLECT

No. of panel observations
Adjusted R2
F-statistic
Durbin Watson statistic

∗∗∗

−1.8174
(−5.6014)
0.0791
(0.3116)
−0.7071
(−2.1907)∗∗
0.0340
(2.2178)∗∗
0.0393
(3.7283)∗∗∗
0.9602
(5.5380)∗∗∗
0.0011
(0.6530)

110

0.3880
12.5169∗∗∗
0.5715∗∗∗

Notes: This table presents the panel least squares regression results using random effects for all

22 countries. All regressions are run with five years of data: 2000–04 if lagged, and 2001–05 if

not lagged. Numbers in parentheses are t-statistics.
∗∗

Statistically significant at α = 5 percent level and

∗∗∗

statistically significant at α = 1 per cento

level. The cross-section random effects are not reported as they sum to zero and are not relevant

to the discussions.

The results from Equation (2) are reported in Table 3. If increases in GDP cause greater
frequency of CEO incentive remuneration, we would expect to find a positive and signifi-
cant regression coefficient for lagged real GDP growth GDPGR(–1). What we find instead
is that the regression coefficient is insignificantly different from zero, suggesting that GDP
growth is not causing a greater frequency of firms providing CEO equity-based incentive
compensation. These results suggest that our results in Table 2 not only support Hypoth-
esis 1 regarding a positive relation between the frequency of firms providing CEO incen-
tive pay and the real GDP growth of the country, but also that the positive relation runs
only in one direction, with increased frequency of CEO incentive remuneration leading
increased real GDP growth of the country.

4.5 Empirical results for Hypothesis 2

To test Hypothesis 2, we use our model in equation (1), which includes Hofstede’s in-
dividualism index values, INDIVCOLLECT, for the 22 countries. Our results in Table 2
support Hypothesis 2 via the negative and statistically significant relation at the 1 per-
cent level between each country’s Hofstede individualism index value and national
economic prosperity. The regression coefficient on INDIVCOLLECT of −0.0007 indi-
cates that an increase in a country’s degree of individualism results in less GDP growth,
O, conversely, that an increase in a country’s degree of collectivism results in stronger
GDP growth.

123

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

Tavolo 4. Support of Hypothesis 2
GDPGRit = γ0 + γ1LTIREMi(t−1)

+ γ2INVRATIOit + γ3LEGALit

+ γ4LABORit + γ5EDUit + γ

6INDIVCOLLECTit + μi + εit

Panel least squares regression results on GDP growth

11 collectivist countries
with INDIVCOLLECT ≤ 65 with INDIVCOLLECT > 65

11 individualist countries

Intercept

LTIREM(−1)

INVRATIO

LEGAL

LABOR

LN(EDUCATION)

INDIVCOLLECT

No. of panel observations
Adjusted R2
F-statistic
Durbin Watson statistic

∗∗

0.2154
(2.5572)∗∗
0.1209
(3.1739)∗∗∗
0.2866
(2.5127)
−0.0028
(−0.4338)
0.0005
(0.0965)
−0.1369
(−3.1313)∗∗∗
0.0002
(0.2718)
55
0.1415
2.4829∗∗
1.4812

−0.0689
(−1.1672)
0.0143
(0.8672)
−0.0652
(−0.9867)
−0.0037
(−1.4463)
0.0059
(2.9761)∗∗∗
0.0485
(1.9720)∗
−0.0003
(−0.8032)
55
0.3068
4.9835∗∗∗
1.9179

Notes: This table presents the panel least squares regression results using random effects for all 22 coun-

tries split into two subsamples: 11 collectivist countries and 11 individualist countries. All regressions are

run with five years of data: 2001–05 if lagged, and 2002–06 if not lagged. Numbers in parentheses are

t-statistics.
∗
Statistically significant at α = 10 percent level;
∗∗∗
statistically significant at α = 1 percent level.

∗∗

statistically significant at α = 5 percent level;

Clearer support for Hypothesis 2, Tuttavia, is demonstrated by splitting our sample into
IL 11 collectivist countries and 11 individualist countries as demarcated by Hofstede’s in-
dividualism index values below or above, rispettivamente, our 22-country sample median
value of 65. Tavolo 4 uses our model of equation (1) and demonstrates strong support
for Hypothesis 2 by the much larger magnitude of the regression coefficient of lagged
LTIREM for the 11 collectivist countries than the 11 individualist countries. The collec-
tivist country regression coefficient of 0.1209 is over eight times the magnitude of the in-
dividualist country regression coefficient of 0.0143. Inoltre, the collectivist country
regression coefficient of lagged LTIREM is significant at the 1 percent level whereas the
individualist country regression coefficient is insignificant, also supporting Hypothesis
2. For the 11 individualist country subsample, both LABOR and LN(EDUCATION) Avere
positive and significant regression coefficients.

4.6 Robustness tests of Hypotheses 1 E 2

Having addressed the causality issue with respect to Hypothesis 1 we turn to the influ-
ence of outliers in terms of the percent of firms offering long-term equity-based incentives
in a country (cioè., stock options, restricted stock, and performance shares). Identifying
the outliers for the 11 individualistic countries from Table 1, Sweden is the lowest with
65.0 percent of firms offering long-term incentives and Canada is the highest with 100

124

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

Tavolo 5. Support of Hypothesis 1: Robustness tests
GDPGRit = γ0 + γ1LTIREMi(t−1)

+ γ2INVRATIOit + γ3LEGALit

+ γ4LABORit + γ5EDUit + γ

6INDIVCOLLECTit + μi + εit

Panel least squares regression results on GDP growth

Omit Japan
& Sweden
20 countries

Omit Singapore Omit four
& Canada
20 countries

outliers
18 countries

Intercept

LTIREM(−1)

INVRATIO

LEGAL

LABOR

LN(EDUCATION)

INDIVCOLLECT

No. of panel observations
Adjusted R2
F-statistic
Durbin Watson statistic

0.2038
(4.0200)∗∗∗
0.0753
(2.9510)∗∗∗
0.1361
(2.3276)∗∗
−0.0020
(−0.6279)
0.0018
(0.7077)
−0.0958
(−3.4665)∗∗∗
−0.0007
(−3.1741)
100

∗∗∗

0.176
4.534∗∗∗
∗∗∗
1.375

0.1901
(3.8373)∗∗∗
0.0744
(3.4190)∗∗∗
0.1275
(1.9903)∗∗
0.0014
(0.4090)
0.0006
(0.2440)
−0.0896
(−3.2931)∗∗∗
−0.0009
(−3.2590)
100

∗∗∗

0.147
3.844∗∗∗
∗∗
1.472

0.2244
(4.1036)∗∗∗
0.0873
(3.1862)∗∗∗
0.1365
(2.1679)∗∗
0.0002
(0.0688)
0.0009
(0.3356)
−0.1079
(−3.6089)∗∗∗
−0.0009
(−3.4002)
90
0.165
3.932∗∗∗
∗∗
1.417

∗∗∗

Notes: This table presents the panel least squares regression results using random effects for 18

O 20 of the original 22 countries as indicated. All regressions are run with five years of data:

2001–05 if lagged, and 2002–06 if not lagged. Numbers in parentheses are t-statistics.
∗∗

Statistically significant at α = 5 percent level;

statistically significant at α = 1 per cento

∗∗∗

level. The cross-section random effects are not reported as they sum to zero and are not relevant

to the discussions.

percent of firms offering such incentives. Outliers chosen for the 11 collectivist countries
are India8 and Japan averaging 16.7 percent and 28.3 per cento, rispettivamente, of firms with
long-term equity-based incentives and Singapore averages 80.0 per cento.

Tavolo 5 tests Hypothesis 1 using all 22 countries. We run three additional regressions:
(1) omitting the two lowest countries of Japan and Sweden in each subsample, (2) omit-
ting the two highest countries of Singapore and Canada in each subsample, E (3) omit-
ting all four countries identified as outliers. The results are qualitatively the same in all
three scenarios. Notably, there is even more support for Hypothesis 1 because in all three
regressions the coefficient on lagged LTIREM not only continues to be significant at the 1
percent level but is greater than the 6.42 percent in Table 2, ranging from 7.44 percent to
8.73 per cento. These three additional regressions also support Hypothesis 2 by the negative
and significant at the 1 percent level coefficients for Hofstede’s individualism index value.
The magnitudes of the coefficients are virtually identical to the coefficient in Table 2 using
Tutto 22 countries.

8 India for brevity reasons is excluded in the tables, but the result is qualitatively the same when
India is substituted for Japan as the lowest of the collectivist countries’ firms offering long-term
equity-based incentives.

125

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

Tavolo 6. Support of Hypothesis 2: Robustness tests for collectivist countries.
GDPGRit = γ0 + γ1LTIREMi(t−1)

+ γ2INVRATIOit + γ3LEGALit

+ γ4LABORit + γ5EDUit + γ

6INDIVCOLLECTit + μi + εit

Panel least squares regression results on GDP growth

Omit Japan
10 countries

Omit Singapore
10 countries

Omit Japan &
Singapore
9 countries

Intercept

LTIREM(−1)

INVRATIO

LEGAL

LABOR

LN(EDUCATION)

INDIVCOLLECT

No. of panel observations
Adjusted R2
F-statistic
Durbin Watson statistic

0.2898
(2.6658)∗∗
0.1415
(2.6632)∗∗
0.2614
(2.3774)∗∗
−0.0058
(−0.7895)
0.0002
(0.0396)
−0.1599
(−3.0149)∗∗∗
−0.0001
(−0.2392)
50
0.099
1.899
1.402

0.2456
(2.4869)∗∗
0.1375
(3.1740)∗∗∗
0.3028
(2.2603)∗∗
0.0016
(0.2125)
−0.0016
(−0.2750)
−0.1573
(−3.0495)∗∗∗
−0.0001
(−0.1308)
50
0.163
2.589∗∗
1.641

0.3371
(2.6919)∗∗
0.1683
(2.7520)∗∗∗
0.2594
(2.0406)∗∗
−0.0014
(−0.1643)
−0.0018
(−0.2874)
−0.1870
(−3.0155)∗∗∗
−0.0006
(−0.7286)
45
0.128
2.074∗
1.559

Notes: This table presents the panel least squares regression results using random effects for the 11 collectivist
countries with INDIVCOLLECT ≤ 65. All regressions are run with five years of data: 2001–05 if lagged, E

2002–06 if not lagged. Numbers in parentheses are t-statistics.
∗
Statistically significant at α = 10 percent level;
significant at α = 1 percent level.

∗∗

statistically significant at α = 5 percent level;

∗∗∗

statistically

Tests of Hypothesis 2 are reported in Tables 6 E 7 by running the same regression in
Tavolo 2 but splitting the sample into the 11 individualist countries with Hofstede’s indi-
vidualism index value greater than 65 and the 11 collectivist countries with Hofstede’s
individualism index value less than or equal to 65. The three regressions run for the sen-
sitivity analysis for the 11 collectivist countries are: (1) omitting Japan, (2) omitting Sin-
gapore, E (3) omitting both Japan and Singapore (Tavolo 6). The coefficient on lagged
LTIREM remains large in magnitude relative to the 1.43 percent coefficient for the indi-
vidualist countries reported in Table 4, ranging from 13.75 percent to 16.83 per cento, E
larger than 12.09 percent reported in Table 4 for collectivist countries. The coefficient is
significant at the 5 percent level when omitting Japan and significant at the 1 percent level
for the other two scenarios. With both outliers removed the coefficient is 16.83 per cento
and significant at the 1 percent level.

The three regressions run for the sensitivity analysis for the 11 individualist countries
are: (1) omitting Sweden, (2) omitting Canada, E (3) omitting both Sweden and Canada
(Tavolo 7). Although the coefficients are greater than the 1.43 percent reported in Table 4,
ranging from 1.63 percent to 3.87 per cento, they remain small relative to the collectivist
regression coefficients. In contrast to the Table 4 results with all 11 individualist coun-
tries where the coefficient of lagged LTIREM is statistically insignificant, two of the three

126

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

Tavolo 7. Support of Hypothesis 2: Robustness tests for individualist countries.
GDPGRit = γ0 + γ1LTIREMi(t−1)

+ γ2INVRATIOit + γ3LEGALit

+ γ4LABORit + γ5EDUit + γ

6INDIVCOLLECTit + μi + εit

Panel least squares regression results on GDP growth

Omit Sweden
10 countries

Omit Canada
10 countries

Omit Sweden
& Canada
9 countries

Intercept

LTIREM(−1)

INVRATIO

LEGAL

LABOR

LN(EDUCATION)

INDIVCOLLECT

No. of panel observations
Adjusted R2
F-statistic
Durbin Watson statistic

−0.0836
(−2.2987)∗∗
0.0378
(2.4899)∗∗
0.0532
(1.0923)
−0.0002
(−0.1190)
0.0032
(2.0206)∗∗
0.0146
(0.9021)
0.0000
(0.1864)
50
0.442
7.470∗∗∗
2.114

−0.0779
(−1.1119)
0.0163
(0.9263)
−0.0628
(−0.7978)
−0.0037
(−1.3029)
0.0060
(2.8074)∗∗∗
0.0522
(1.8344)∗
−0.0003
(−0.7924)
50
0.303
4.554∗∗∗
1.956

−0.0891
(−1.9240)∗
0.0387
(2.2855)∗∗
0.0540
(0.9026)
−0.0003
(−0.1235)
0.0036
(1.9435)∗
0.0176
(0.8691)
0.0000
(−0.0305)
45
0.407
6.023∗∗∗
2.128

Notes: This table presents the panel least squares regression results using random effects for the 11 individualist
countries with INDIVCOLLECT > 65. All regressions are run with five years of data: 2001–05 if lagged, E

2002–06 if not lagged. Numbers in parentheses are t-statistics.
∗
Statistically significant at α = 10 percent level;
significant at α = 1 percent level.

∗∗

statistically significant at α = 5 percent level;

∗∗∗

statistically

coefficients are now significant at the 5 percent level and the scenario omitting only
Canada, the high outlier, continues to be statistically insignificant for the coefficient of
lagged LTIREM.

The sensitivity analysis leads to the conclusion that leaving out the outlier countries does
not change support for both Hypotheses 1 E 2. The results are not being driven by the
outlier countries of Japan and Singapore for the collectivist countries or the outlier coun-
tries of Sweden and Canada for the individualist countries. Infatti, support for both hy-
potheses is strengthened with the removal of the outlier countries.

5. Discussion and implications

This study is motivated by the public reaction to large CEO pay packages, especially in
times of economic crisis and mediocre firm performance. This makes CEO pay incen-
tives a global public policy issue (Ertimur, Ferri, and Muslu 2011). Therefore, this paper
analyzes the relation between the prevalence of CEO long-term equity-based incentive
pay and its impact on national economic prosperity (GDP growth), and the moderat-
ing effect of individualist/collectivist culture on this relationship. We contribute to the

127

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D

/

l

F
/

/

/

/

1
5
2
1
0
9
1
6
8
4
6
2
3
UN
S
e
P
_
UN
_
0
0
4
3
2
P
D

.

/

F

B
sì
G
tu
e
S
T

T

o
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

CEO Pay and National Economic Growth

understanding of CEO compensation through the development of a new theory, a merger
between agency theory and social learning theory: vicarious agency.

After longitudinal data analysis for 22 countries we find support for the hypothesis that
the long-term equity-based incentive reward given to a CEO in one company may ac-
tually become an incentive to a host of CEOs in other networked firms (typically within
the same country). These CEOs are motivated to model their behavior after that of the
rewarded CEO. Although the CEO may well deserve the incentive reward, if this is
carefully overseen by the board, their behaviors may not have led to economic success
for the focal firm because success depends on good fortune as well as wealth-creating
decision-making by the CEO. Nevertheless, the board may be correct with regard to re-
warding the CEO, because as other CEOs mimic the behaviors of successful CEOs, greater
hustle spreads throughout the CEO and/or potential CEO community.

Daniel, Cieslewics, and Pourjalali (2012) indicate that national cultural practices in-
fluence the institutional environment of firms and this, in turn, influences numerous
corporate governance practices such as financial disclosure, board accountability, E
executive compensation practices. More specifically, recent research in international
business highlights how incentive systems are instrumental in bringing about economic
change (Yang and Stening 2013). Using China as a case, Yang and Stening write that “cul-
tural values evolve at the slowest pace while business-related values and political ide-
ologies are more dynamic and change at a considerable faster pace” (P. 438) and that
“particular changes to ownership and reward systems have triggered business related
value changes in China” (P. 438). Consistent with their statement, we argue that long-
term equity-based executive compensation incentives are particularly powerful mecha-
nisms of change and are of particular importance in collectivist countries such as China
(Adithipyangkul, Alon, and Zhang 2011, makes such arguments for China). These argu-
ments for changes through reward systems in collectivist countries are key motivating
factors behind our Hypothesis 2.

In Hypothesis 2, culture may influence executive compensation practices. Our findings
demonstrate the importance of comparative international research that addresses the dy-
namic impact of organizational behavior (CEO incentives) on country-level outcomes
(economic growth). We find that the vicarious agency dynamic appears more robust in
collectivist than individualist national societies. Inoltre, this may help us to more
fully understand the dynamic between CEO long-term equity-based incentive compensa-
tion and firm performance. Per esempio, it may be that CEO long-term equity-based in-
centive compensation is more effective in enhancing firm performance in societies where
institutional practices push for parsimonious use of such incentives. In cases where 85
percent or more of CEOs expect incentive-based compensation, as is the case in most

128

Asian Economic Papers

l

D
o
w
N
o
UN
D
e
D

F
R
o
M
H

T
T

P

:
/
/

D
io
R
e
C
T
.

M

io
T
.

/

e
D
tu
UN
S
e
P
UN
R
T
io
C
e
–
P
D