Generalists versus Specialists: Managerial Skills and CEO Pay*

Cláudia Custódio† Arizona State University – W. P. Carey School of Business

Miguel A. Ferreira‡ Universidade Nova de Lisboa – Faculdade de Economia

Pedro Matos§ University of Southern California – Marshall School of Business

This Version: October 2010

Abstract

We study whether the rise in CEO pay in the last decades is explained by an increase in the relative importance of general managerial skills. We construct an index of general managerial ability using detailed information on a CEO’s industry background, experience as top executive, and educational training. We find a positive relation between the general ability index and CEO pay using the sample of S&P 1,500 firms in the 1993-2007 period. This relation is driven by both industry mobility and experience as a top manager and is mostly pronounced among diversified firms. We find a 10% pay premium by comparing the actual CEO’s pay to the pay of a portfolio of CEOs who are specialists in the industries that match the CEO’s industry experience. Furthermore, we provide evidence that CEO pay increases the most when firms hire a new CEOs from outside the firm and switch from a specialist to a generalist executive. Our findings suggest that a rise in the importance of general relative to firm-specific managerial skills helps to explain the rise in executive pay.

JEL Classification: G34, J24, J33 Keywords: Executive compensation, Market for executives, Human capital, CEO

* We thank Ilona Babenko and Oguzhan Ozbas for helpful comments. This research is supported by a research grant from the Fundação para a Ciência e Tecnologia (FCT/POCI 2010). † E-mail: claudia.custodio@asu.edu ‡ E-mail: miguel.ferreira@fe.unl.pt § E-mail: pmatos@usc.edu

1

1. Introduction

The rapid rise in the level of U.S. executive pay in the last decades has sparked a heated debate

in the finance literature over the causes of this trend (Frydman and Jenter (2010)). Some authors

argue that the surge in CEO pay is due to “rent extraction” by powerful CEOs who use captive

boards of directors and a loosening of social norms to award themselves excessive pay packages

(Bebchuk, Fried, and Walker (2002)). Others counter that the U.S. corporate governance system

works reasonably well (Holmstrom and Kaplan (2001)) and point to market forces in the CEO

labor market to explain the evolution of CEO pay. These authors argue that the rise in CEO pay

has actually occurred simultaneously with an increase in external hiring and that externally hired

CEOs actually earn more, on average, than “entrenched” incumbents or CEOs promoted from

within firms (Murphy and Zabojnik (2004, 2008), and Frydman (2005)).

In the market-based explanation for the rise in executive compensation, CEO pay is determined

in a competitive labor market between firms and CEOs, and pay has increased because

conditions in this market have changed. Murphy and Zabojnik (2004) develop a model in which

CEOs with more transferable skills across firms and industries command higher market wages in

equilibrium. Murphy and Zabojnik (2008) and Frydman (2009) provide evidence that the nature

of CEO jobs in the last decades has changed and there is a greater emphasis on general skills

instead of firm-specific. This change could be a result of product market changes due to industry

deregulation (Hubbard and Palia (1995), and Cunat and Guadalupe (2009a)) and foreign

competition (Cunat and Guadalupe (2009b)). Other forces could be the rise of a knowledge-

based economy (Garicano and Rossi-Hansberg (2006)) where technology and changes in general

management practices increase the effect of CEO talent on firm value, or that the CEO job has

not changed considerably but that the increase in pay can be attributed to the substantial growth

in firm size and top talent being matched to larger firms (Gabaix and Landier (2006)). Ferreira

and Sah (2010) propose a model where more generalist managers tend to occupy the top of the

hierarchy as the complexity of the environment and the quality of communication in the

organization increases.

2

In this paper, we investigate whether the rise in CEO pay is explained by the rise in the relative

importance of general managerial skills relative to firm-specific skills. To test this hypothesis,

we use the panel data set of CEOs of S&P 1,500 firms in the 1993-2007 period obtained from

Execucomp. We collect full biographical data for the CEOs, namely age and educational

background but, most importantly, detailed information on all current and past positions in the

CEO’s professional career, including positions outside S&P 1,500 firms and non-board positions.

This detailed biographical data are drawn from the BoardEx database. Our final sample includes

more than 4,000 CEOs, whose resumes include more than 32,000 different past positions.

A major contribution of this paper is to measure the level of general managerial skills of each

CEO. We focus on industry mobility and top executive experience. For each CEO at the end of

each year, we measure the number of industries in which he worked (past number of industries),

whether he held a CEO position at a different company (past CEO dummy), and whether he has a

business degree (general education dummy). We construct an index of general managerial ability,

as given by the first factor of the principal component analysis of the three individual measures

of general ability. This index summarizes information on each manager’s general human capital

skills and allows us to classify a CEO as a “generalist” (high index scores) or a “specialist” (low

index scores). We find that the average general managerial ability index has increased over the

last 15 years for CEOs of S&P 1,500 firms. There has been also an upward trend in the average

level of CEO compensation in our sample.

We find that the general managerial ability index is positively and significantly associated with

total CEO pay. The effect is economically meaningful. A one-standard deviation in the index of

general managerial ability is associated with an additional 6% in total pay, which in monetary

terms means that a generalist CEO earns approximately a quarter of a million dollars of extra pay

per year than a specialist CEO. We find a positive association for both the cash and equity

components of pay. We control for other CEO characteristics (e.g., tenure and age) and firm

characteristics. Results are also robust to the inclusion of firm fixed effects, which control for

omitted time-invariant firm characteristics. Furthermore, we find that the past number of

industries variable (which proxies for industry mobility) and the past CEO dummy are,

individually, positively linked to CEO pay.

3

To measure the magnitude of the premium that generalist CEOs command over their non-

generalist or specialist peers, we develop a new measure of CEO excess compensation. This

measure captures the market wage premium of a CEO with multi-industry experience over the

prevalent pay of CEOs with single-industry experience that match the CEO past industry

experience. This excess compensation measure gives a more direct test of whether generalist

CEOs are paid at a premium when compared to a matched portfolio of equivalent specialists. For

each CEO-year, the excess compensation is the difference between the total compensation

awarded to the CEO and his imputed compensation, calculated as the average CEO

compensation of the portfolio of industries where the CEO worked up to a given year, where the

industry-level compensation is the median compensation of CEOs that worked only in a industry

up to a given year (single-industry CEOs). The measure of CEO compensation is inspired by the

measures of excess value used in the corporate diversification literature (Berger and Ofek

(1995)).

We find that CEO excess compensation is the highest in the telecom industry in our sample. At

the same time, CEOs in the telecom industry score the highest in the index of general managerial

ability. The rapid change in this industry in the nineties due to technological innovation,

deregulation and increase in competition could be behind the demand by firms for managers with

strong general skills. The average excess compensation of multi-industry CEOs has increased

over the last 15 years and is closely linked to the evolution of the total CEO pay. We find that

CEOs with a multi-industry background have higher excess compensation than single-industry

CEOs. We also find that the general ability index is significantly associated with higher excess

compensation, controlling for other CEO and firm characteristics.

We then examine whether the market wage premium of generalist CEOs is heterogeneous across

types of firms. We find a “general managerial ability” wage premium both in small and large

firms. Interestingly, however, the premium is more prevalent in diversified firms (i.e.,

conglomerates) than in stand-alone firms. We investigate this issues in more detail and find that

conglomerates pay more to executives that have moved across industries in the past. This result

is in line with Rose and Shepard (1997) and Anderson, Bates, Bizjak, and Lemmon (1998) who

4

also find that conglomerates pay more to CEOs in order to compensate for their managerial

ability.

Finally, we test more directly the implications of the Murphy and Zabojnik (2004) model that

CEOs with more transferable skills across firms and industries capture higher market wages

when they change jobs. Therefore, we examine in detail cases of CEO turnover and classify them

into an internal hire or an external hire when the CEO is hired from outside the firm. We further

classify an external hire into a switch from a specialist to a generalist or a switch from a

generalist to a specialist. We find that there is a stronger increase in CEO pay when a switch

from a specialist to a new generalist CEO via an external hire takes place. This is direct evidence

that general managerial human capital does command a premium in the CEO labor market. This

result is consistent with Harris and Helfat (1997) who find that CEOs that move across firms in

different industries receive a higher compensation premium than those that move across firms in

the same industry.

Our paper contributes to the literature on market-based explanations of the developments in CEO

pay (Murphy and Zabojnik (2004)). In a broader sense, our work shows that managerial

attributes are important to understanding CEO pay. Graham, Li, and Qiu (2009) find that

manager fixed effects explain a majority of the variation in executive pay. Our study uses

detailed biographical data on CEOs instead of executive fixed effects,. We complement the

evidence in Frydman (2005) who finds that the rise in executive pay is explained by the growing

importance of general managerial skills in the second half of the twentieth century using a

sample of the top 50 firms. She finds a positive association between CEO pay and the generality

of human capital, as measured by education and occupational mobility within a firm (i.e., the

number of organizational areas an executive worked such as production, sales, or human

resources). Our index focuses on mobility across industries and firms, rather than internal

mobility within a firm. We use a more comprehensive data set that covers the full sample of S&P

1,500 firms in recent decades. We directly test the hypothesis that generalist CEOs are paid more

by developing a measure of excess compensation, which shows that generalists earn a premium

when put against specialist CEOs working in the same industries. In addition, we show that the

5

pay premium is coming from the labor market because CEO pay mainly increases when a firm

hires a new generalist CEO from outside the firm.

Our work also extends Murphy and Zabojnik (2008) and Eisfeldt and Kuhnen (2010) by

providing additional evidence on the effect of firm-specific and general skills on CEO turnover.

These models predict a compensation premium not only for externally hired CEOs, but also for

hiring industry outsiders, particularly after industry shocks. Cremers and Grinstein (2010) show

that the external market for CEOs tends to respond more to industry shocks. Kaplan, Klebanov,

and Sorensen (2010) provide evidence that firms involved in leverage buyouts do trade-off

general versus firm-specific skills of CEOs while performing their hiring decisions. Overall, our

paper provides evidence that general managerial skills are linked to higher pay and a change in

the CEO job function toward emphasizing those skills and this could help to explain the rise in

CEO compensation over the last decades.

However, other authors have alternative views. Gabaix and Landier (2006) argue that the CEO

job has not changed but that the increase in pay can be mostly attributed to the growth in firm

size and to the assertive matching of top talent to large firms. Kaplan and Rauh (2010) argue that

the increasing return to generalist managerial skills does not explain the top end of the income

distribution. Instead, they argue that greater firm scale and superstars are more likely to explain

the increased skewness at the top of the income distribution.

The remainder of the paper is organized as follows. In the next section, we present the sample

and variables used in this study. In section 3 we present the empirical results on the relation

between CEO compensation and general managerial skills. In section 4 we investigate the

specific case of diversified firms or conglomerates. Section 5 concludes.

2. Sample and Data Description

Our initial sample consists of a panel of 25,562 CEO-firm-years in the 1993-2007 period drawn

from the Execucomp database.1,2 We manually match the executives in Execucomp that are

1 Although Execucomp contains data starting in 1992, we drop this first year due to the small number of observations (only 433 CEO-firms).

6

identified as CEOs in a specific year with their Boardex profiles in order to get data on their

characteristics including education and professional experience in terms of industries, firms, and

positions. The final sample contains 21,142 CEO-firm-years observations and 4,264 different

CEOs. The number of Execucomp CEO-firms in our sample per year is provided in the last

column of Table 1. We could not find a match in Boardex for 1,211 CEOs contained in our

initial sample. There is some survivorship bias in BoardEx, which affects primarily the match in

the start of the sample period. The percentage of CEO in Execucomp whose profile is in

BoardEx grows from 67% in 1993 to 80% in 1998 and then to more than 90% in the 2000-2007

period.3 The primary findings in the paper are robust when we use the 2000-2007 sample period.

We perform an additional data match between firms in Boardex and Compustat (or alternatively

Datastream for international firms) in order to obtain the industry classification of past positions

of the CEOs. Because Compustat and Datastream only include publicly-traded firms, our

analysis is restricted to past positions in publicly-listed firms. The sample of past positions

contains 32,500 observations.

2.1. Measuring CEO Compensation and General Managerial Ability

Figure 1 and Table 1 report the time series of the average CEO total compensation in the 1993-

2007 period. Total compensation consists of salary, bonus, value of restricted stock granted,

value of options granted, long term incentive payout and other compensation (Execucomp item

TDC1). In some tests we consider cash compensation and equity-based compensation separately.

We develop a new measure of excess compensation that aims to capture the compensation

premium (or discount) of a generalist or multi-industry CEO (i.e., an executive who worked in

several different industries) when matched to an equivalent portfolio of specialist or single-

industry CEOs. This measure is inspired by the excess value measure commonly used in the

corporate diversification literature (e.g., Berger and Ofek (1995)). It is computed as the

2 Cadman, Klasa, and Matsunaga (2009) and Gao, Li, and Lemmon (2010) argue that there are systematic differences in CEO compensation, both in terms of levels and structure, across private and public firms and for Execucomp and non-Execucomp firms. 3 The CEO profiles missing in BoardEx are mainly from executives that have retired or deceased before 2000.

7

difference between the CEO total compensation and its imputed compensation. The imputed

compensation is the average CEO compensation of the portfolio of industries where the CEO

worked based on his past work experience. The industry level compensation is given by the

median compensation of CEOs that worked only in that industry (single-industry CEOs). The

industry match is done at the four-digit SIC code level when there are five or more single-

industry CEOs or at the highest SIC level where there are at least five single-industry CEOs. In

the case of a single-industry CEO, the excess compensation measure is simply the difference

between the CEO total compensation and the median compensation of single-industry CEOs in

his industry.

Figure 1 and columns (3)-(4) of Table 1 show a significant increase in both average total

compensation and average excess compensation, as well as a positive correlation between the

two.4 Because the average excess compensation is always positive over time, we conclude that

multi-industry CEOs are paid at a premium. Moreover, this evidence suggests that the increase in

total CEO compensation is linked to an increase in the compensation premium of generalist

CEOs over specialists working in similar industries.

We investigate the importance of general human capital of CEOs is related to the increase in

CEO pay over the recent years. To do so, we create an index of the generality of the CEO’s

human capital (general managerial ability index) based on the CEO’s education and past work

experience (prior to the current position) in publicly-listed firms. We consider the following

proxies of general managerial ability:

Past number of industries: number of industries at the four-digit SIC level where a CEO has

worked in publicly-listed firms, prior to the current position.

Past CEO dummy: dummy variable that equals one if a CEO held a CEO position at another

publicly-listed firm, prior to the current position.

4 The analysis of the average excess compensation over time in Table 1 and Figure 1 is restricted the sample to multi-industry CEOs. This helps us to better interpret the generalist CEO excess compensation since the measure is simply capturing how much more (or less) a generalist is paid when compared to its equivalent portfolio of specialists.

8

General education dummy: dummy variable that equals one if the CEO has a business degree,

irrespective of its level (undergraduate, master and PhD).

In some tests and in the robustness section we use some variations of the prior three variables:

Past multi-industry experience: dummy variable that equals one if a CEO worked in more than

one industry (four-digit SIC), prior to the current position

Past number of firms: number of publicly-listed firms a CEO worked, prior to the current

position.

We use principal components to transform the proxy variables into a smaller number of factors

that contain the same information as the original proxy variables. The first principal component is

the linear combination of the three variables (past number of industries, past CEO dummy, and

general education dummy) used to proxy for general managerial ability. Following the usual

practice, we use all principal components whose eigenvalues exceed one (i.e., that have more

explanatory power than any one of the original proxies by itself). The results of this analysis are

presented in Table 2 and they give us a single factor for the general managerial ability index with

a eigenvalue of 1.14. As expected, the factor loadings are positive on all three proxies. The

factor gives about equal weight to the past number of industries and general education dummy,

and a much lower weight to the past CEO dummy. Thus, higher levels of general human capital

are reflected in a higher value of the index. We standardize the index to have zero mean and a

standard deviation of one.

Our measure is related to the index of generality of human capital defined in Frydman (2005).

Frydman’s measure takes into account the CEO's education and his occupational mobility within

the current firm (i.e., in different areas of the firm such as sales and production). However, it

does not look at the industry and top manager experience of a CEO. So we focus on the role of

external mobility rather that internal mobility in the CEO labor market.

We find an increase in the importance of the general managerial skills of CEOs that is consistent

with the findings in Frydman (2005). Figure 1 and Table 1 show that that there is a shift in the

relative importance of general versus firm-specific managerial skills with the index of general

9

managerial ability steadily increasing over time. This is consistent with executives running firms

in more recent years having more general skills, which are transferable across firms and

industries.5

Figure 2 shows the cross sectional distribution of the general index scores over the 1993-2000

and 2001-2007 periods. The figure shows a shift in the distribution of the index to the right,

which is consistent with an increase in the importance of general managerial skills during our

sample period.

2.2. Cross-Industry Variation

Table 3 shows the average total and excess compensation and the average general managerial

ability index by industry. As we have argued before, industry level phenomena, such as

technological shocks, changes in product market competition and regulation might determine the

need for top executives with more general skills. We find significant variation across industries

in terms of the index of general managerial ability as well as in terms of the differences in pay

between generalists and specialists CEOs.

The telecom industry has the highest average level of generality of CEO human capital (0.35)

and at the same time is also the industry where the CEOs get the highest average total pay ($6.9

million). Moreover, generalist CEOs (defined as the CEOs with a general managerial ability

index above the median) in this industry are the ones paid the highest amount. When looking at

our excess compensation measure we find that CEOs of telecom firms with resumes that include

positions in other industries receive on average $5.3 million more than CEOs that have spent all

their careers only in the telecom sector (single-industry CEOs). In fact, during the 1990s, the

telecom industry changed rapidly not only in terms of technological innovation (cell phones,

internet, etc), but also in terms of regulation as a result of the Telecommunications Act in 1996

(Laffont and Tirole (2000)). The rapid increase in competition associated with deregulation and

5The individual measures that are part of the index are also increasing over time, particularly during the first half of our sample period. The average past number of industries steadily increases from 1 in 1994 to 1.2 in 2002. In 1993, 40% of CEOs have general education, while after 2002 this figure increases to more than 50%. The average past CEO dummy increases from 34% in 1993 to 50% in 2002.

10

the need to rapidly adjust to technological innovation could increase the demand for managers

who can respond to this type of environment, namely managers with strong general skills.

The industry with the lowest general managerial skills index is Oil, Gas and Coal (-0.142). We

still find a premium in generalists' compensation in this industry. However, the difference in

average total compensation between generalists and specialists is less than 1 million dollars.

Except for CEOs in the Consumer Durables industry, we find a positive and statistically

significant generalist compensation premium across all industries, both when comparing the

average total compensation of generalist to specialists as well as when using the excess

compensation measure.

2.3. Summary Statistics of CEO and Firm Characteristics

Table 4 shows summary statistics for compensation, CEO characteristics and firm characteristics.

The average CEO in our sample receives $4 million of total compensation, which corresponds to

$1.2 million in cash and $2.1 million in equity incentives. The mean excess compensation is $1.4

million.

Besides the CEO characteristics that are included in the general ability index already described

above (past number of industries, past CEO dummy, general education dummy), we measure the

following additional CEO attributes:

CEO tenure: number of years a CEO has held the top executive position in the firm. We expect a

negative relation between the number of years in the CEO position and compensation, as more

recently hired CEOs are expected to have their compensation adjusted to recent market

conditions.

External hire dummy: dummy variable that equals one if the CEO is hired from outside the firm,

and zero otherwise (i.e., promoted internally). We expect a premium to exist for externally hired

CEO, since the company has to face labor market wages to hire him.

CEO age: the age of the CEO in years.

11

A CEO in our sample has worked in one other industry than his current industry and two other

firms, on average. 17% of the CEOs have previous experience as a top executive and 48% have a

business degree. They are 55 years old and spend 8 years in a given CEO position on average,

and about 22% of the CEOs are hired from outside the firm.

In our tests we control for the following firm characteristics:

Size: total sales in US$ millions. Gabaix and Landier (2005) argue that size is the main

determinant of CEO compensation, because CEO talent is matched to firm size.

Leverage: total debt divided by total assets.

Tobin’s Q: sum of total assets plus market value of equity minus book value of equity divided by

total assets. Tobin’s Q aims to capture the growth opportunities of the firm and tends to be

positively associated to CEO compensation.

ROA: earnings before interest, taxes, depreciation and amortization divided by total assets. This

is a measure of accounting performance which is expected to be positively correlated with

compensation. We include both contemporaneous and lagged ROA as control variables.

Volatility: annualized standard deviation of monthly stock returns.

Diversification dummy: dummy variable that equals one if a firm operated in more than one

business segment. Rose and Shepard (1997) and Anderson et.al (1998) find a positive premium

in compensation associated with diversified firms. In some tests we use the number of business

segments rather than the dummy.

Stock return: annual stock return. We expect a positive relation between stock market

performance and compensation. We include both contemporaneous and lagged stock return as

control variables.

Cash: sum of cash and short-term investments divided by total assets.

R&D: research and development expenses divided by total assets.

CAPEX: capital expenditures divided by total assets.

12

Firm age: number of years since a firm’s shares have been publicly listed.

The Appendix table provides definitions and data sources for all variables. All variables are

winsorized at the 5th and 95th percentile values.

2.4. Firm Characteristics and General Managerial Ability Index

We look at the match between firms and general managerial skills using univariate tests. Table 5

presents correlation coefficients between the general ability index and firm characteristics. The

table also presents mean firm characteristics for “generalist” and “specialist” CEOs and test for

the difference in firm characteristics between the two types of CEOs. A generalist is defined as a

CEO with general managerial ability index above the yearly median, while a specialist has an

index below the yearly median. As expected, we find that firms with generalist CEOs are bigger,

older and more diversified. We also find that firms with generalist CEOs have slightly higher

leverage and cash holdings, and lower stock return volatility. We do not find statistically

significant differences in terms of accounting and stock market performance (ROA and stock

returns) between generalists and specialists. The differences in Tobin’s Q, CAPEX and R&D are

not economically meaningful, although they are statistically significant.

The results in Table 5 suggest that firm size and diversification are relevant firm characteristics

in explaining the relation between the generality of CEO human capital and compensation. With

respect to firm size, for instance, at this stage we cannot reject the possibility that the positive

correlation between compensation and the generality of human capital is due to the fact that

bigger firms may tend to appoint generalist CEOs. A similar reasoning can be applied to

diversified firms. We further investigate these issues in multivariate tests in Sections 3 and 4.

With respect to CEO characteristics, we find that generalist CEOs tend to have shorter tenure and

to be more frequently hired from outside the firm. We expect this to be the case, since generalist

CEOs are more likely to have worked in different firms.

3. Do Generalist CEOs Get Paid More?

We investigate the relation between CEO compensation and the generality of managerial skills.

We first use panel regression tests where the dependent variables are CEO total, cash, or equity

13

compensation. We also use a measure of CEO excess compensation based on the CEO’s industry

experience. We then examine the relation between CEO compensation and the generality of

managerial skills in a sample of large versus small firms and diversified versus standalone firms.

We also investigate CEO compensation around CEO turnovers and switches of CEO type.

Finally, we conduct several robustness checks on our primary results.

3.1. CEO Compensation Tests

Table 6 presents our main test of whether CEOs with higher general managerial ability receive

higher compensation. We run both ordinary least squares (OLS) and firm fixed effects panel

regressions, where the dependent variable is the logarithm of CEO total compensation (columns

(1)-(4)), the logarithm of CEO cash (salary plus bonus) compensation (columns (5)-(6)) and the

logarithm of equity (restricted stock plus option awards) compensation (columns (7)-(8)). Our

explanatory variable of interest is the general managerial ability index. The OLS regressions

include industry (two-digit SIC) and year dummies. The firm fixed effects specification controls

for unobserved sources of firm heterogeneity and also include year dummies.6 Fixed-effects

methods solve “joint determination” problems in which an unobserved time-invariant variable

simultaneously determines both total compensation and the general managerial ability index. It is

equivalent to looking only at within-firm changes in the general managerial ability index.

Because only the effects of within-firm changes in total compensation are taken into account,

firm-specific omitted variables cannot explain the observed relationship between CEO

compensation and managerial attributes. All reported t-statistics are adjusted for

heteroskedasticity and within-firm correlation using firm-level clustered standard errors.

Columns (1)-(2) present the coefficients of a regression of total compensation on firm and CEO

attributes but not including the general managerial ability index. This is the base case regression.

We find that firm size is positively and significantly associated with total compensation, which is

consistent with the findings in Gabaix and Landier (2005) and others. 7 We consider other

commonly used firm-level factors affecting executive pay and find that total pay is positively and

significantly associated with growth opportunities (Tobin’s Q) and firm performance as

6 We obtain consistent findings using CEO fixed effects regressions. 7 Our results are robust when we use the market value of assets or equity as proxies for firm size.

14

measured by ROA and stock returns. The only CEO attribute that significantly and consistently

impacts total pay is the external hire dummy. We find that external CEO appointments carry a

positive and significant wage premium, which is consistent with Murphy and Zabojnik (2004,

2008).

Columns (3)-(4) present the coefficients of a regression of total compensation on the general

managerial ability index as well as firm and CEO attributes controls. There is strong evidence of

a positive and statistically significant relation between total pay and the index. The coefficient on

the general managerial ability index variable is positive and statistically significant. This is

consistent with CEOs with more general managerial skills carrying a wage premium. It is also

economically significant. Using the specification in column (3), a one-standard deviation

increase in the index is associated with an increase in total compensation of about 6%. This

corresponds to approximately a quarter of a million dollars of extra pay per year.

With respect to the other explanatory variables, the inclusion of the general managerial ability

index does not significantly affect the coefficients of the control variables in columns (1)-(2),

although the magnitude of the effects is slightly lower in some cases. We still see that firm size

and performance are positively associated with total pay and that external CEO appointments

carry a significant premium.

Columns (5)-(6) present estimates of similar regressions using cash compensation as the

dependent variable, while columns (7)-(8) use equity incentives as the dependent variable. In all

specifications we find a positive and significant relation between compensation and the general

managerial ability index. The magnitude of the effect, however, is stronger in the case of equity

compensation. Using the specifications in columns (5) and (7), a one-standard deviation increase

in the index is associated with an increase of 1.8% and 8.6% in cash and equity compensation,

respectively. In unreported regression, we find that general managerial ability index is positively

associated with equity compensation, if we run a Tobit specification for the ratio of equity

compensation to total compensation.

We have documented a positive relation between CEO total pay and an index that captures the

degree of generality of managerial skills. The index includes three components related with the

15

external mobility or mobility across industries of executives (past number of industries), past

experience as a CEO (past CEO dummy), and educational background (general education

dummy). Table 7 presents the estimates of the same regressions as those in Table 6 but now

using as main explanatory variables the individual components of the index rather than the

aggregate index. We find that the past number of industries and the past CEO dummy are

positively and significantly associated with total compensation, while the general education

dummy coefficient is insignificant. Thus, the evidence supports the hypothesis that both high

external mobility across industries and a past appointment as a CEO carry positive pay premiums.

A CEO that worked in different industries in the past receives an additional pay of 6% in terms

of total compensation, while a CEO that has a past CEO appointment receives 5% more. This

represents an increase in average annual total compensation of approximately $240,000 and

$200,000, respectively. The fixed effects specifications show consistent results. Having an

MBA-type education does not have a significant effect, perhaps reflecting the fact that MBA-

type education is nowadays widespread not only among executives, but also among non-

executive professionals. Indeed, if both executives and non-executives invest in a more general

education, this might explain the rise in executive compensation, but not the increasing gap

between executive and non-executive compensation. Our results challenge some of the findings

of Frydman (2005) and Murphy and Zabjonik (2008) on the role of business education, de per se,

explaining the rise in CEO pay over the last decades.

3.2. CEO Excess Compensation Tests

We have documented in the previous subsection that CEOs with more general skills earn a

significant pay premium. In this section, to examine whether general managerial skills is

associated with higher or lower CEO total pay, we measure the percentage difference between a

CEO’s total pay and the imputed pay given its past industry experience measured by the pay of

single-industry CEOs (i.e., the pay of CEOs that worked only in one industry over their career).

We measure excess compensation as the log of the ratio of a CEO’s total compensation to its

imputed compensation - i.e. the premium or discount in pay resulting from industry mobility.

This measures allows us to do a better match between the compensation of a CEO with more

general skills (generalists) and the CEOs with less general skills (specialists) in terms of their

16

industry experience. The measure of CEO compensation is inspired by the measures of excess

value used in the corporate diversification literature (e.g., Berger and Ofek (1995)).

In the regressions of Table 8 the dependent variable is the CEO excess compensation. We control

for other factors that could affect excess compensation. Firm size, performance, growth

opportunities and other firm and CEO characteristics are taken into account using the same set of

control variables used in Table 6. The main explanatory variables of interest are the past multi-

industry dummy and the general ability index. We use the past multi-industry dummy as

explanatory variable because it matches naturally the definition of the dependent variable in this

section. Columns (1)-(2) present the coefficients of the regression of CEO excess compensation

on the past multi-industry dummy as well as control variables. The coefficient on the past multi-

industry dummy captures the percentage difference in average excess compensation between

CEOs with a career path across more than one industry (multi-industry) and single-industry

CEOs (single-industry). The pay premium for multi-industry CEOs is more than 10% in the OLS

specification in column (1).

Columns (3)-(4) present the coefficients of a regression of CEO excess compensation on the

general managerial ability index as well as control variables. There is strong evidence of a

positive and statistically significant relation. A one-standard deviation increase in the index is

associated with 4.5% more CEO excess compensation using specification (3).

In summary, the findings using the excess compensation measure are consistent with those using

total compensation as the dependent variable. The magnitude of the effects is also similar. The

results so far support the notion that CEOs with more general managerial skills carry a

significant wage premium in the labor market. In particular, CEO compensation is higher for

executives that have not made their entire career in a single industry and therefore there is

evidence of an industry mobility wage premium.

3.3. Effect of Firm Size and Diversification

In the previous subsections, we find evidence of a positive relation between CEO compensation

and proxies of CEO general managerial human capital and industry mobility. In this subsection,

we investigate whether the relation between general managerial attributes and executive

17

compensation is heterogeneous across different types of firms. We first study firm size because

the size of a firm’s operations has been shown to be an important determinant of CEO

compensation (e.g., Gabaix and Landier (2006)). Second, we study corporate diversification as

measured by the diversification dummy (i.e., a firm operates in more than one business segment).

Both firm size and the number of business segments have been used as proxies of the scope and

complexity of the firm’s operations (e.g., Boone, Field, Karpoff, and Raheja (2007), and Coles,

Daniel, and Naveen (2008)).

Table 9 presents the estimates of regressions of CEO pay for groups of firms based on their size

and whether they are diversified. The specifications and the set of control variables are the same

as the ones used in Table 6. Our explanatory variable of interest is the general managerial ability

index.

Columns (1)-(4) of Table 9 present results by splitting the sample into small and large firms

based on whether firm size is below the yearly median (small firms in columns (1)-(2)) or above

the yearly median (large firms in columns (3)-(4)). We find a positive and statistically significant

relation between total compensation and the general managerial ability index in both small and

large firms. The magnitude of the effects is also similar in both groups of firms. We conclude

that our primary finding is not driven only by large firms. Smaller firms seem also to be willing

to pay a similar premium for general managerial skills.

Columns (5)-(8) of Table 9 present results by splitting the sample into with a single business

segment (standalone firms in columns (5)-(6)) and diversified firms (multi-business segments

firms in columns (7)-(8)). Interestingly, we find a positive and statistically significant coefficient

only for the diversified firms sample in the firm fixed effects specifications (column (8)). This

finding is consistent with the hypothesis that diversified firms are the ones that are in more need

and pay a higher premium for executives with general managerial skills. In other words, a

generalist manager is more valuable when matched to a generalist firm, such as conglomerates. 8

We explore this issue in more detail in Section 4 below.

8 Xuan (2009) finds that appointing specialist CEOs in multi-division firms might lead to inefficient capital allocation decisions, particularly if the CEOs was previously linked to a division of the firm.

18

3.4. Evidence from CEO Turnovers

Our evidence so far is based on firm-year panel regression estimates. In this subsection we

examine changes in compensation around CEO turnover, in particular when there is a switch of

CEO type. We expect to find an increase in CEO compensation when a firm switches from a

CEO with less general skills (specialist) to a CEO with more general skills (generalist). We

classify CEOs as generalists if their general managerial ability index is above the yearly median,

and we classify them as specialists if their index below the median. Furthermore, the effect is

likely to be more pronounced when the new CEO is hired from outside the firm as in this case

the firm has to pay to access the market (Murphy and Zabojnik (2004, 2008)).

We run regression tests where the dependent variable is total, cash, equity or excess

compensation. We Use the following explanatory variables to measure the effect on pay from a

CEO turnover: (1) no switch of CEO type, which is a dummy that equals one if there is a CEO

turnover at time t but there is no switch of CEO type; (2) switch to generalist – external hire,

which is a dummy that equals one if there is a CEO turnover at time t and a specialist CEO is

replaced by a generalist CEO hired from outside the firm; (3) switch to specialist – external hire,

which is a dummy that equals one if there is a CEO turnover at time t and a generalist CEO is

replaced by a specialist CEO hired from outside the firm; and (4) switch – internal hire, which is

a dummy that equals one if there is a CEO turnover at time t and the new CEO is internally

appointed. We include these four variables as regressors and therefore the intercept in the

regression captures the benchmark case of no CEO turnover at time t. Table 10 presents the

results.

Column (1) presents the coefficients of a regression of total compensation on the four

explanatory variables of interest and control variables (same set of variables used in Table 6).

The coefficient on switch to generalist – external hire is positive and significant, while the

coefficient on switch to specialist – external hire is insignificant at the 5% level. The effect on

CEO total compensation of switching from a specialist to a generalist CEO by external

appointment is substantial at about 26% extra pay.

19

Columns (2) and (3) present estimates of similar regressions but using the logarithm of CEO

cash and equity compensation as dependent variables. The results show that the increase in

compensation associated with a switch from a specialist to a generalist CEO is driven by equity

incentives, rather than by increases in salary or bonus. The results are also consistent with a

significant impact on CEO excess compensation of a switch from a specialist to a generalist CEO

(column (4)).

Overall, there is evidence of an increase in CEO pay when a firm switches from a specialist to a

new generalist CEO via an external hire. This is direct evidence that general managerial human

capital does command a premium in the CEO labor market.

3.5. Robustness

In this subsection we perform several robustness checks of our primary findings. Table 11

presents robustness checks with respect to the general managerial ability index. We alternatively

construct the general ability index using the past CEO dummy, the general education dummy,

but now using the past number of firms rather than the past number of industries (columns (1)-

(2)). The past number of firms measures the number of different firms where an executive

worked in the past. The intuition is that a generalist CEO should possess skills that are

transferable across firms (and not only industries) rather “firm-specific human capital” that it is

valuable only within a particular organization. The second variation of the general managerial

ability index is to count the past number of industries based on two-digit SIC codes rather than

four-digit SIC codes (columns (3)-(4)). The third variation of the index is to count the past

number of CEO positions rather than using the past CEO dummy variable (columns (5)-(6)). We

find strong evidence of a positive relation between CEO total pay and the alternative definitions

of the general managerial ability index in all specifications. Similar to the results in Table 6, the

effect of the index of general managerial attributes on CEO total pay is always economically and

statistically significant.

We also perform robustness checks on the measure of excess compensation. We calculate the

measure of excess compensation by performing the match between the multi-industry CEO and

the portfolio of single-industry CEOs using two-digit SIC codes instead of the four-digit SIC

20

codes. Using this coarser industry classification has two main implications: first, we ensure that

the industries are significantly different from each other, and second, that we have more single-

industry CEOs to use as a benchmark. In untabulated regressions (similar to those in Table 8),

we find the results to be robust to this alternative definition of the measure of excess

compensation. There is strong evidence of a positive and statistically significant relation between

excess compensation and the multi-industry dummy.

3.6. Discussion

We go one step further in our analysis and we look at whether there is evidence that higher pay

levels by generalist CEOs is consistent with an efficient functioning of the CEO labor market.

An alternative hypothesis is that generalist CEOs accumulate their multi-industry experience

because they are just “serial CEOs” that engage in “job hopping” (Giannetti (2010)).

Additionally, generalist CEOs may feature more prominently in databases of executive search

firms and have an easier time being recruited. Dasgupta and Ding (2010) emphasize the

enhanced role of executive search firms in the last decades. We investigate the relation between

generalist CEOs and firm performance to test this alternative explanation of our findings.

Specifically, the alternative hypothesis predicts that firms hiring generalist CEOs would see

performance and shareholder returns negatively affected. In untabulated regerssions, we find

evidence of a positive relation between shareholder value (Tobin’s Q) and the general ability

index using an OLS specification, although the relation is not statistically significant when we

use a firm fixed effects specification. We also look at whether generalist CEOs exhibit higher

turnover or “job hopping”. In unreported tests, we do not find evidence consistent with this

prediction. The relation between CEO turnover and the general ability index is statistically

insignificant and even negative in some specifications. Taken together, the findings do not

support the existence of market inefficiencies in the setting of CEO pay. Overall, the results are

consistent with an efficient market-based explanation of the significant wage premium awarded

to CEOs with general managerial skills.

21

4. A Detailed Look at Conglomerate Firms

In sub-section 3.3 we provide evidence that conglomerates pay more, on average, to CEOs with

multi-industry career backgrounds. This is not a mechanic effect that large and more diversified

firms pay their CEOs more. Interestingly, in our basic tests in Table 6 we do not find evidence

that conglomerates per se pay more to their CEOs irrespective of their level of general

managerial ability.

In order to perform a more direct test of the relation between firm diversification and CEO

compensation we define a new measure of excess compensation: conglomerates excess

compensation. While the excess compensation measure helps us to answer the question on

whether generalists are paid at a premium or a discount when compared to specialists, the

conglomerates excess compensation measure answers a different question: whether diversified

firms pay an additional premium to their CEOs. We use Compustat segment-level data to

construct the conglomerates excess compensation. We compare the actual CEO compensation of

a diversified firm to their imputed standalone compensation for individual business segments.

The imputed compensation is the average compensation of the portfolio of industries that

matches the business segments of the diversified firm where the industry-level compensation is

the median of across standalone firms in a industry. The industry match is done at the four-digit

SIC code level where there are five or more stand-alone firms or at the highest SIC level when

there are at least five stand-alone firms. In the case of standalone firms, the excess compensation

measure is simply the difference between the CEO total compensation and the median

compensation of standalone firms in its industry. The conglomerates excess compensation is the

log of the ratio of a CEO’s total compensation to its imputed compensation. This measure is

similar in spirit to the excess value measure used to estimate the value effect of diversification

(e.g., Berger and Ofek (1995)). The main difference is that we compare the compensation of a

conglomerate with the compensation of an equivalent portfolio of standalone firms in the same

industries instead of comparing firm valuation.

Rose and Shepard (1997) and Anderson et al. (1998) find a significant compensation premium in

diversified firms. However, while controlling for industry effects they take only into account the

core business of the conglomerate, ignoring the industries of the secondary business segments.

22

By using our measure of conglomerate excess compensation we make sure to match the

conglomerate’s compensation to a benchmark that includes all the industries where the

conglomerate operates in.

Table 12 presents the estimates of regressions where the dependent variable is the new measure

of conglomerates excess compensation. Our explanatory variables of interest are alternatively the

diversification dummy (equals one if the firm has more than one business segment) and the

firm’s number of segments.

The positive and significant coefficient of the diversification dummy in columns (1)-(2) is

consistent with CEOs of diversified firms earning higher wages relative to standalone firms. The

conglomerates CEO pay premium is substantial at 26% using the specification in column (1).

Columns (3)-(4) present the estimates using the number of segments as main explanatory

variable. We find a positive and significant relation between the number of business segments

and CEO total compensation.

Overall, using our measure of conglomerates excess compensation, we find evidence consistent

with Rose and Shepard (1997) who argue that diversified firms are more complex organizations,

and more difficult to manage, and therefore CEOs in diversified firms should be paid at a

premium relative to stand-alone firms. Anderson et al. (1998) argue that the compensation

premium in conglomerates is not necessarily associated with poorer corporate governance in

these firms. Our findings also suggest that conglomerates are only willing to pay a premium to

CEOs to compensate them for their general skills. In fact, Laux (2001) suggests that it is cheaper

to provide incentives in conglomerates that standalones.

An alternative view is that diversifications destroy value due to greater agency costs, which

would be associated with higher rent extraction by managers (Jensen (1986)). There is a wide

literature on the existence of a diversification discount (Berger and Ofek (1995), Lang and Stulz

(1994), Rajan, Servaes and Zingales (2000) among others) which could be consistent with higher

agency costs, however, there is also contradictory evidence that diversification de per se causes a

loss of firm value (Villalonga (2004)).

23

Finally, we find a positive and significant relation between the conglomerates excess

compensation and the index of general managerial ability in columns (5)-(6). This suggests that

not only conglomerates pay a premium to CEOs relative to standalone firms but that the

premium for general managerial skills is greater in conglomerates. In untabulated results, we also

find that that the coefficient of the interaction variable between the diversification dummy and

the general ability index is positive and significant.

We check for the robustness of the results using a finer measure of conglomerate excess

compensation, where the imputed compensation is now calculated imposing the requirement that

standalone firms should have single-industry CEOs (i.e., executives that have always worked in

the same industry). In untabulated results, we still find a compensation premium in diversified

firms when we use this finer measure of conglomerate excess compensation.

5. Conclusion

This paper investigates whether the increase in the importance of general managerial skills of top

executives explains the rise in CEO compensation. We construct a new measure of the generality

of human capital based on a CEO’s external mobility across industries, experience as top

executive, and education. We find a positive and significant relation between the index of

general managerial ability and CEO compensation using the sample of S&P 1500 firms in the

1993-2007 period.

We test more directly if there is a compensation premium for generalist CEOs. We construct a

new measure of excess compensation that compares the total compensation of a CEO that has

worked in many industries with the average compensation of a matched portfolio of single-

industry CEOs in the same industries where the CEO worked in the past. We find the excess

annual compensation premium is as high as 10%. Our results suggest that the effect of general

skills on compensation is particularly strong for diversified firms. We further investigate the case

of conglomerate firms and find that there is a compensation premium for generalists in

conglomerate firms as high as 25%.

24

Our paper provides direct evidence of the growing importance of general managerial skills

versus firm-specific skills. This trend is likely to increase the outside opportunities of CEOs with

general skills and therefore lead to higher levels of CEO pay. Our findings highlight the

importance of the inner workings of the CEO labor market to understand the evolution of CEO

pay in recent decades.

25

References

Anderson, Ronald, Thomas Bates, John Bizjak, and Michael Lemmon, 1998, Corporate

Governance and Firm Diversification, Working paper, University of Utah.

Bebchuk, Lucian, Jesse Fried, and David Walker, 2002, Managerial Power and Rent Extraction

in the Design of Executive Compensation, University of Chicago Law Review 69, 751-761.

Berger, Philip, and Eli Offek, 1995, Diversification’s Effect on Firm Value, Journal of Financial

Economics 37, 39-65.

Bertrand, Marianne and Antoinette Schoar, 2003, Managing with Style: The Effect of Managers

on Firm Policies, Quarterly Journal of Economics 68, 1169-1208.

Boone, Audra, Laura Field, Jonathan Karpoff, and Charu Raheja, 2007, The determinants of

corporate board size and composition: An empirical analysis, Journal of Financial

Economics 85, 66-101.

Cadman, Brian, Sandy Klasa, and Steven Matsunaga, 2009, Determinants of CEO Pay: A

Comparison of ExecuComp and Non-ExecuComp Firms, Accounting Review, forthcoming.

Coles, Jeffrey, Naveen Daniel, and Lalitha Naveen, 2008, Boards: Does one size fit all?, Journal

of Financial Economics 87, 329-356.

Cremers, Martijn, and Yaniv Grinstein, 2010, The Market for CEO Talent: Implications for CEO

Compensation, Working paper, Yale University.

Denis, David, Diane Denis, and Atulya Sarin, 1997, Agency Problems, Equity Ownership, and

Corporate Diversification, Journal of Finance 52, 135-160.

Cuñat, Vincente, and Maria Guadalupe, 2009a, Executive compensation and competition in the

banking and financial sectors, Journal of Banking and Finance 33, 495‐504

Cuñat, Vincente, and Maria Guadalupe, 2009b, Globalization and the provision of incentives

inside the firm: The effect of foreign competition, Journal of Labor Economics 27, 179‐212

26

Dasgupta, Sudipto, and Fei Ding, 2010, Search Intermediaries, Internal Labor Markets, and CEO

Pay, Working paper, Hong Kong University of Science & Technology

Eisfeldt, Andrea, and Camelia Kuhnen, 2010, CEO Turnover in a Competitive Assignment

Framework, Working paper, Northwestern University.

Ferreira, Daniel, and Raaj Sah, 2010, Who Gets to the Top? Generalists Versus Specialists in

Organizations, Working paper., London School of Economics.

Frydman, Carola, 2005, Rising Through the Ranks: The Evolution of the Market for Corporate

Executives, 1936-2003, Working paper, MIT Sloan School.

Frydman, Carola, and Dirk Jenter, 2010, CEO Compensation, Working paper, Rock Center for

Corporate Governance, Stanford University.

Gabaix, Xavier, and Augustin Landier, 2006, Why Has CEO Pay Increased So Much?, Quarterly

Journal of Economics 123, 49-100.

Gao, Huasheng, Kai Li, and Michael Lemmon, 2010, A Comparison of CEO Pay in Public and

Private US Firms, Working paper, University of Utah.

Garicano, Luis, and Esteban Rossi-Hansberg, 2006, Organization and inequality in a knowledge

economy, Quarterly Journal of Economics 121, 1383-1435

Giannetti, Mariassunta, 2010, Serial CEO Incentives and the Structure of Managerial Contracts,

working paper, Stockholm School of Economics.

Graham, John, Si Li, and Jiaping Qiu, 2009, Managerial Attributes and Executive Compensation,

Working paper, Duke University.

Holmstrom, Bengt, and Steven Kaplan, 2001, Corporate Governance and Merger Activity in the

United States: Making Sense of the 1980s and 1990s, Journal of Economic Perspective 15,

121-144.

Hubbard, R. Glenn, and Darius Palia, 1995, Executive pay and performance: Evidence from the

US banking industry, Journal of Financial Economics 39, 105-30.

27

Jensen, Michael, 1986, Agency costs of free cash flow, corporate finance, and takeovers,

American Economic Review 76, 323-329.

Kaplan, Steven, Mark Klebanov, and Morten Sorensen, 2010, Which CEO Characteristics Matter?

Journal of Finance, forthcoming.

Kaplan, Steven, Joshua Rauh, 2010, Wall Street and Main Street: What Contributes to the Rise

in the Highest Incomes? Review of Financial Studies, forthcoming.

Laux, Christian, 2001, Limited-Liability and Incentive Contracting with Multiple Projects,

RAND Journal of Economics 32, 514-526.

Laffont, Jean, and Jean Tirole, 2000, Competition in Telecommunications (MIT Press

Cambridge, MA.).

Lang, Larry, and Rene Stulz, 1994, Tobin’s q, Corporate Diversification and Firm Performance,

Journal of Political Economy 102, 1248-1280.

Murphy, Kevin J., and Ján Zábojník, 2004, CEO Pay and Appointments: A Market-Based

Explanation for Recent Trends, American Economic Review - Papers and Proceedings 94

192-196.

Murphy, Kevin J., and Ján Zábojník, 2008, Managerial Capital and the Market for CEOs,

Working paper, University of Southern California.

Rajan, Raghuram, Henri Servaes, and Luigi Zingales, 2000, The Cost of Diversity: The

Diversification Discount and Inefficient Investment, Journal of Finance 55, 35-80

Rose, Nancy, and Andrea Shepard, 1997, Firm Diversification and CEO Compensation:

Managerial Ability or Executive Entrenchment? RAND Journal of Economics 28, 489-514.

Villalonga, Belen, 2004, Diversification Discount or Premium? New Evidence from the Business

Information Tracking Series, Journal of Finance 59, 475-502

Xuan, Yuhai, 2009, Empire-Building or Bridge-Building? Evidence from New CEOs’ Internal

Capital Allocation Decisions, Review of Financial Studies 22, 4919-4948.

28

Table 1 CEO Compensation and General Managerial Ability Index by Year

This table presents average score of the general managerial ability index and average CEO total compensation per year from 1993 to 2007. The index of general managerial ability is the first factor of applying the principal components method to three proxies of general managerial ability (past number of industries, past CEO dummy, and general education dummy). The sample consists of Execucomp firms for which CEO profile data are available from BoardEx. Average CEO excess compensation for the sample of CEOs that worked in more than one four-digit SIC industry (multi-industry CEOs) are also shown. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix.

Year

Average general 

ability index

Average total 

compensation        

($ thousands)

Average excess 

compensation of 

multi‐industry CEOs 

($ thousands) Obs.

1993 ‐0.107 2,131                          744                               774            

1994 ‐0.155 2,270                          852                               1,011        

1995 ‐0.110 2,415                          1,149                          1,092        

1996 ‐0.104 2,994                          1,330                          1,184        

1997 ‐0.075 3,546                          1,728                          1,272        

1998 ‐0.045 3,734                          1,691                          1,392        

1999 ‐0.026 4,227                          2,088                          1,515        

2000 ‐0.004 4,577                          2,598                          1,595        

2001 0.022 4,783                          2,588                          1,548        

2002 0.058 4,400                          2,239                          1,572        

2003 0.051 4,191                          2,142                          1,642        

2004 0.073 4,672                          2,315                          1,653        

2005 0.076 4,769                          2,282                          1,670        

2006 0.075 4,952                          2,279                          1,652        

2007 0.067 5,054                          2,191                          1,570        

Total 0.000 4,092                          2,000                          21,142

29

Table 2 General Managerial Ability Index - Principal Component Analysis Results

This table presents the results of the principal component analysis of three proxies of general managerial ability (past number of industries, past CEO dummy, and general education dummy). Factor loadings on each proxy general managerial ability, eigenvalue, and proportion of variation explained by the first factor are shown. Variable definitions and data sources are provided in the Appendix.

Past nr. of industries Past CEO dummy

General education 

dummy

Loadings 0.756 0.013 0.755

Scores 0.662 0.011 0.662

Proportion explained

Eigenvalue

0.380

1.141

30

Table 3 CEO Compensation and General Managerial Ability Index by Industry

This table presents average score for the CEO general ability index and average CEO total compensation for each of the Fama-French 12 industries. Average CEO total compensation for the sub-samples of CEOs above (high index) and below (low index) the median yearly general ability index and the t-statistic of the difference, and average CEO excess compensation for the sample of CEOs that worked in more than one four-digit SIC industry (multi-industry CEOs) and its t-statistic are also shown. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix.

Industry

Average 

general 

ability index

Average total 

compensation ($ 

thousands)

Average total 

compensation ‐      

high general ability 

index ($ thousands)

Average total 

compensation ‐      

low general ability 

index ($ thousands)

Difference   

t‐stat

Average excess 

compensation of  

multi‐industry CEOs 

($ thousands)  t‐stat Obs.

Consumer nondurables 0.011 4,101                 4,992                         3,482                          6.94 2,110                         13.98     1,423

Consumer durables 0.077 3,587                 3,514                         3,650                          ‐0.48 883                            5.57       648

Manufacturing 0.144 3,354                 4,072                         2,689                          10.92 1,385                         16.77     2,815

Oil, gas, and coal ‐0.142 4,518                 5,125                         4,167                          2.99 2,189                         10.58     874

Chemicals 0.085 4,090                 4,411                         3,845                          2.18 1,925                         12.71     749

Business equipment ‐0.037 4,192                 4,873                         3,748                          7.29 1,975                         20.39     3,581

Telephone and television 0.350 6,907                 8,211                         5,536                          4.79 5,371                         15.20     435

Utilities 0.113 2,746                 3,146                         2,387                          4.56 832                            7.86       1,180

Wholesale and retail ‐0.129 3,585                 4,482                         3,033                          9.37 1,816                         16.13     2,495

Healthcare and drugs 0.015 4,379                 5,527                         3,456                          9.50 2,289                         15.65     1,627

Finance ‐0.066 5,182                 6,058                         4,648                          7.60 2,962                         22.66     2,911

Other ‐0.031 3,947                 4,683                         3,486                          6.82 1,952                         16.11     2,404

31

Table 4 Summary Statistics

This table presents the mean, median, standard deviation, minimum, maximum and number of observations for each variable. All variables are winsorized at the 5th and 95th percentile values. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix.

Variable Mean Median Std. Dev. Min Max Obs.

Total compensation ($ thousands) 4,092 2,407 4,274 458 16,490 21,142

Cash compensation ($ thousands) 1,217 946 858 306 3,500 21,142

Equity compensation ($ thousands) 2,124 865 2,973 0 11,013 18,199

Excess compensation ($ thousands) 1,430 0 3,917 ‐2,609 12,938 21,119

General ability index 0.00 0.16 1.00 ‐1.19 6.41 21,142

Past nr. industries 1.12 1.00 1.38 0.00 13.00 21,142

Past CEO dummy 0.17 0.00 0.37 0.00 1.00 21,142

General education dummy 0.48 0.00 0.50 0.00 1.00 21,142

Past multi‐industry dummy 0.57 1.00 0.49 0.00 1.00 21,142

Past nr. firms 1.67 1.00 1.72 0.00 18.00 21,142

CEO tenure 7.95 6.00 7.03 1.00 49.00 19,727

External hire dummy 0.22 0.00 0.41 0.00 1.00 19,396

CEO age 55.53 56.00 7.35 29.00 90.00 20,131

Sales ($ milions) 3,619.20 1,316.70 5,261.24 120.57 20,051.42 21,113

Leverage 0.22 0.22 0.17 0.00 0.55 21,038

Tobin's Q 1.91 1.52 1.04 0.97 4.83 21,078

ROA 0.13 0.13 0.08 0.00 0.29 20,740

Volatility 0.11 0.09 0.05 0.04 0.24 20,938

Diversification dummy 0.57 1.00 0.49 0.00 1.00 18,529

Stock return 0.13 0.10 0.37 ‐0.50 0.97 18,746

Cash 0.13 0.06 0.15 0.00 0.52 21,111

R&D 0.02 0.00 0.04 0.00 0.14 21,117

CAPEX 0.05 0.04 0.04 0.00 0.16 20,084

Firm age 23.15 18.00 19.02 0.00 82.00 20,982

Panel A: Compensation

Panel B: CEO characteristics

Panel C: Firm characteristics

32

Table 5 General Managerial Ability Index and CEO and Firm Characteristics

This table presents correlation coefficient of the general managerial ability index with firm characteristics, and averages of firm characteristics for the samples of CEOs above (high index) and below (low index) the median yearly general ability index and t-statistic of the difference. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix.

Correlation with 

general ability 

index

Average ‐ high 

general ability 

index

Average ‐ low 

general ability 

index

Difference   

t‐stat

CEO tenure ‐0.080 7.443 8.319 ‐8.680

External hire dummy 0.106 0.260 0.187 12.283

CEO age 0.062 55.74 55.38 3.4085

Sales ($ millions) 0.197 6,326 4,131 10.891

Leverage 0.054 0.233 0.217 7.128

Tobin's Q ‐0.020 1.889 1.925 ‐2.496

ROA ‐0.024 0.133 0.133 0.040

Volatility ‐0.093 0.101 0.109 ‐10.986

Diversification dummy 0.129 0.605 0.550 7.465

Stock return 0.002 0.134 0.129 0.887

Cash ‐0.071 0.117 0.133 ‐7.846

R&D 0.026 0.025 0.023 3.979

CAPEX ‐0.098 0.052 0.055 ‐4.515

Firm age 0.125 25.497 21.434 15.358

33

Table 6 CEO Compensation and General Managerial Ability Index

This table presents estimates of OLS and firm fixed effects panel regressions of the logarithm of CEO total, cash and equity compensation on the general managerial ability index and other CEO- and firm-level control variables. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4) (5) (6) (7) (8)

OLS

Fixed 

effects OLS

Fixed 

effects OLS

Fixed 

effects OLS

Fixed 

effects

General ability index 0.063*** 0.041*** 0.018** 0.019** 0.086*** 0.057**

[5.00] [3.14] [2.20] [2.01] [4.16] [2.21]

CEO tenure ‐0.007*** 0.000 ‐0.006** 0.000 0.001 0.006*** ‐0.001 ‐0.005

[‐2.76] [0.049] [‐2.41] [0.091] [0.51] [3.18] [‐0.24] [‐0.96]

External hire dummy 0.127*** 0.190*** 0.104*** 0.172*** 0.051** 0.071*** 0.286*** 0.352***

[4.01] [5.41] [3.26] [4.90] [2.47] [2.81] [5.80] [5.39]

CEO age ‐0.000 ‐0.002 ‐0.001 ‐0.003 0.005*** 0.001 ‐0.009*** ‐0.003

[‐0.11] [‐0.87] [‐0.48] [‐1.05] [3.49] [0.78] [‐2.82] [‐0.81]

Sales (log) 0.435*** 0.297*** 0.427*** 0.294*** 0.256*** 0.164*** 0.582*** 0.455***

[32.5] [11.0] [31.9] [11.0] [30.6] [9.18] [33.6] [9.09]

Leverage 0.378*** ‐0.323*** 0.378*** ‐0.330*** 0.190*** ‐0.185*** 0.456*** ‐0.342*

[3.97] [‐3.27] [3.97] [‐3.34] [2.94] [‐3.31] [3.20] [‐1.95]

Tobin's Q 0.132*** 0.092*** 0.128*** 0.091*** ‐0.003 ‐0.028*** 0.342*** 0.233***

[6.75] [5.09] [6.62] [5.08] [‐0.23] [‐2.79] [12.1] [7.76]

ROA ‐0.541** 0.384* ‐0.528** 0.386* 1.011*** 1.496*** ‐2.360*** ‐1.089***

[‐2.47] [1.77] [‐2.42] [1.78] [7.56] [11.3] [‐6.88] [‐2.92]

ROA (t‐1) ‐0.109 0.383** ‐0.098 0.393** ‐0.618*** ‐0.423*** 0.524* 0.796**

[‐0.55] [2.19] [‐0.49] [2.24] [‐4.92] [‐3.79] [1.67] [2.56]

Volatility 0.445* 0.252 0.511** 0.264 ‐0.749*** ‐0.857*** 2.030*** 0.756**

[1.77] [1.18] [2.04] [1.24] [‐4.77] [‐6.98] [5.11] [2.01]

Diversification dummy ‐0.015 ‐0.041 ‐0.019 ‐0.041 0.013 ‐0.016 ‐0.068 ‐0.189***

[‐0.53] [‐1.44] [‐0.69] [‐1.45] [0.75] [‐0.90] [‐1.56] [‐3.65]

Stock return 0.175*** 0.131*** 0.177*** 0.133*** 0.237*** 0.222*** ‐0.105** ‐0.117***

[6.93] [5.66] [7.07] [5.74] [14.7] [16.2] [‐2.51] [‐3.02]

Stock return (t‐1) 0.251*** 0.184*** 0.254*** 0.185*** 0.155*** 0.126*** 0.285*** 0.236***

[11.9] [9.62] [12.0] [9.72] [12.1] [11.4] [8.14] [7.19]

Cash 0.452*** 0.210 0.492*** 0.209 0.150* 0.160** 0.989*** 0.014

[3.19] [1.63] [3.50] [1.63] [1.75] [2.24] [5.55] [0.066]

R&D 2.087*** ‐1.911** 1.983*** ‐1.844** 0.496 ‐0.667 2.672*** ‐2.605*

[4.02] [‐2.57] [3.85] [‐2.47] [1.59] [‐1.64] [3.77] [‐1.87]

CAPEX ‐0.081 0.430 0.004 0.432 ‐0.742*** ‐0.246 0.068 0.024

[‐0.23] [1.32] [0.011] [1.33] [‐3.24] [‐1.26] [0.11] [0.042]

Firm age 0.000 ‐0.001 0.000 ‐0.001 0.002*** 0.002 ‐0.004*** ‐0.002

[0.19] [‐0.35] [0.15] [‐0.42] [4.03] [1.57] [‐3.44] [‐0.85]

Constant 3.260*** 3.390*** 4.142*** 1.566***

[14.0] [14.5] [16.0] [3.68]

Observations 11453 11453 11453 11453 11427 11427 8010 8010

R‐squared 0.500 0.262 0.503 0.263 0.555 0.326 0.470 0.225

Year dummies Yes Yes Yes Yes Yes Yes Yes Yes

Industry dummies Yes No Yes No Yes No Yes No

Equity compensation Cash compensation Total compensation (log)

34

Table 7 CEO Compensation and General Managerial Ability Index Components

This table presents estimates of OLS and firm fixed effects panel regressions of the logarithm of CEO total, cash and equity compensation on individual proxies of general managerial ability (past number of industries, past CEO dummy, general education dummy) and other CEO- and firm-level control variables. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4) (5) (6)

OLS Fixed effects OLS Fixed effects OLS Fixed effects

Past nr. industries 0.059*** 0.023** 0.021*** 0.013** 0.068*** 0.026*

[6.46] [2.42] [3.74] [2.00] [5.46] [1.66]

Past CEO dummy 0.052* 0.058*** 0.002 ‐0.000 0.128*** 0.068*

[1.93] [2.61] [0.12] [‐0.035] [3.18] [1.81]

General education dummy 0.007 0.041 ‐0.008 0.011 ‐0.014 0.018

[0.31] [1.51] [‐0.53] [0.59] [‐0.41] [0.41]

CEO tenure ‐0.006** 0.000 0.001 0.006*** 0.000 ‐0.003

[‐2.26] [0.18] [0.58] [3.18] [0.044] [‐0.82]

External hire dummy 0.087*** 0.162*** 0.045** 0.067** 0.237*** 0.325***

[2.70] [4.44] [2.18] [2.58] [5.43] [5.54]

CEO age ‐0.003 ‐0.003 0.004*** 0.001 ‐0.011*** ‐0.005

[‐1.26] [‐1.17] [2.98] [0.58] [‐3.78] [‐1.23]

Sales (log) 0.423*** 0.292*** 0.254*** 0.163*** 0.531*** 0.400***

[31.6] [10.9] [30.3] [9.14] [35.3] [8.82]

Leverage 0.365*** ‐0.331*** 0.187*** ‐0.184*** 0.435*** ‐0.253*

[3.85] [‐3.35] [2.89] [‐3.30] [3.37] [‐1.72]

Tobin's Q 0.126*** 0.092*** ‐0.004 ‐0.028*** 0.295*** 0.195***

[6.58] [5.15] [‐0.31] [‐2.78] [12.2] [7.29]

ROA ‐0.527** 0.398* 1.009*** 1.501*** ‐1.868*** ‐0.732**

[‐2.42] [1.83] [7.55] [11.3] [‐6.28] [‐2.31]

ROA (t‐1) ‐0.075 0.391** ‐0.608*** ‐0.421*** 0.430 0.641***

[‐0.38] [2.22] [‐4.85] [‐3.77] [1.64] [2.59]

Volatility 0.476* 0.270 ‐0.758*** ‐0.856*** 1.531*** 0.549*

[1.89] [1.27] [‐4.83] [‐6.97] [4.39] [1.75]

Diversification dummy ‐0.021 ‐0.040 0.012 ‐0.016 ‐0.056 ‐0.148***

[‐0.79] [‐1.40] [0.67] [‐0.89] [‐1.43] [‐3.39]

Stock return 0.184*** 0.132*** 0.240*** 0.222*** ‐0.090** ‐0.099***

[7.35] [5.71] [14.8] [16.3] [‐2.54] [‐2.97]

Stock return (t‐1) 0.257*** 0.185*** 0.156*** 0.126*** 0.229*** 0.192***

[12.2] [9.72] [12.2] [11.4] [7.75] [6.80]

Cash 0.458*** 0.207 0.134 0.157** 0.864*** 0.116

[3.27] [1.61] [1.57] [2.20] [5.50] [0.65]

R&D 1.995*** ‐1.872** 0.504 ‐0.663 2.949*** ‐1.450

[3.93] [‐2.51] [1.62] [‐1.63] [4.86] [‐1.33]

CAPEX 0.039 0.426 ‐0.730*** ‐0.246 ‐0.107 0.003

[0.11] [1.31] [‐3.20] [‐1.26] [‐0.20] [0.0058]

Firm age 0.000 ‐0.001 0.002*** 0.002 ‐0.004*** ‐0.003

[0.26] [‐0.43] [4.10] [1.58] [‐3.46] [‐1.33]

Constant 3.455*** 4.184*** 2.425***

[14.0] [15.8] [8.51]

Observations 11453 11453 11427 11427 8010 8010

R‐squared 0.506 0.264 0.556 0.326 0.493 0.246

Year dummies Yes Yes Yes Yes Yes Yes

Industry dummies Yes No Yes No Yes No

Equity compensation (log)Cash compensation (log)Total compensation (log)

35

Table 8 CEO Excess Compensation and General Managerial Ability

This table presents estimates of OLS and firm fixed effects panel regressions of the CEO excess compensation on a dummy that takes the value of one if a CEO worked in more than one industry (past multi-industry dummy) and the general ability index and other CEO- and firm-level control variables. The excess compensation is defined as the logarithm of the ratio of the CEO total compensation to its imputed compensation from single-industry CEOs that match the CEO’s past industry (four-digit SIC) experience. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4)OLS Fixed effects OLS Fixed effects

Past multi‐industry dummy 0.104*** 0.055**

[3.96] [1.99]

General ability index 0.045*** 0.027**

[3.29] [1.98]

CEO tenure ‐0.005** ‐0.001 ‐0.005* ‐0.000

[‐2.03] [‐0.25] [‐1.93] [‐0.12]

External hire dummy 0.074** 0.140*** 0.074** 0.139***

[2.28] [3.81] [2.28] [3.81]

CEO age ‐0.001 ‐0.002 ‐0.000 ‐0.001

[‐0.43] [‐0.65] [‐0.21] [‐0.54]

Sales (log) 0.334*** 0.233*** 0.335*** 0.233***

[27.4] [9.31] [27.6] [9.29]

Leverage 0.351*** ‐0.201** 0.355*** ‐0.204**

[3.90] [‐2.12] [3.94] [‐2.15]

Tobin's Q 0.098*** 0.066*** 0.098*** 0.066***

[5.12] [3.64] [5.09] [3.67]

ROA ‐0.085 0.340 ‐0.086 0.333

[‐0.40] [1.64] [‐0.41] [1.60]

ROA (t‐1) ‐0.054 0.234 ‐0.079 0.229

[‐0.28] [1.31] [‐0.40] [1.28]

Volatility 0.520** ‐0.064 0.544** ‐0.060

[2.01] [‐0.30] [2.11] [‐0.28]

Diversification dummy 0.010 ‐0.021 0.009 ‐0.022

[0.37] [‐0.73] [0.32] [‐0.76]

Stock return 0.158*** 0.123*** 0.158*** 0.123***

[6.09] [5.20] [6.12] [5.23]

Stock return (t‐1) 0.184*** 0.135*** 0.185*** 0.135***

[8.39] [6.87] [8.43] [6.92]

Cash 0.484*** 0.145 0.488*** 0.148

[3.61] [1.18] [3.66] [1.21]

R&D 2.470*** ‐1.484** 2.552*** ‐1.419*

[6.37] [‐2.06] [6.57] [‐1.96]

CAPEX 0.490 0.202 0.488 0.200

[1.46] [0.63] [1.45] [0.62]

Firm age ‐0.000 0.000 ‐0.000 0.000

[‐0.13] [0.17] [‐0.24] [0.16]

Constant ‐2.807*** ‐2.777***

[‐19.8] [‐19.7]

Observations 11442 11442 11442 11442

R‐squared 0.344 0.079 0.343 0.079

Year dummies Yes Yes Yes Yes

36

Table 9 CEO Compensation and General Managerial Ability Index: Effect of Firm Size and

Diversification

This table presents estimates of OLS and firm fixed effects panel regressions of the logarithm of CEO total compensation on the general managerial ability index and other CEO- and firm-level control variables. The small (large) firms sample consists of those firms whose sales are below (above) the yearly median. The stand-alone (diversified) firms sample consists of those firms with number of business segments equal to (above) one. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4) (5) (6) (7) (8)

OLS

Fixed 

effects OLS

Fixed 

effects OLS

Fixed 

effects OLS

Fixed 

effects

General ability index 0.057*** 0.053* 0.060*** 0.036** 0.061*** 0.008 0.064*** 0.047***

[2.78] [1.96] [4.00] [2.30] [2.98] [0.25] [4.45] [3.27]

CEO tenure ‐0.007** ‐0.007* ‐0.003 0.004 ‐0.007** ‐0.007 ‐0.006* 0.005*

[‐2.12] [‐1.69] [‐0.95] [1.39] [‐2.00] [‐1.55] [‐1.73] [1.67]

External hire dummy 0.088** 0.187*** 0.130*** 0.177*** 0.158*** 0.249*** 0.065* 0.149***

[2.12] [3.20] [3.13] [4.02] [3.19] [3.64] [1.71] [3.54]

CEO age ‐0.001 0.000 ‐0.001 ‐0.003 ‐0.002 ‐0.001 ‐0.001 ‐0.003

[‐0.47] [0.059] [‐0.41] [‐0.91] [‐0.59] [‐0.34] [‐0.32] [‐1.17]

Sales (log) 0.427*** 0.322*** 0.397*** 0.316*** 0.416*** 0.293*** 0.447*** 0.299***

[16.4] [7.36] [19.0] [7.63] [24.5] [6.41] [27.4] [8.74]

Leverage 0.327*** ‐0.423*** 0.315* ‐0.181 0.186 ‐0.390** 0.473*** ‐0.321***

[3.00] [‐3.11] [1.94] [‐1.28] [1.41] [‐2.53] [3.80] [‐2.64]

Tobin's Q 0.146*** 0.122*** 0.095*** 0.072*** 0.126*** 0.103*** 0.125*** 0.074***

[6.14] [4.52] [3.29] [2.94] [5.48] [3.61] [4.44] [3.07]

ROA ‐0.871*** 0.374 0.086 0.513 ‐0.514 0.506 ‐0.499* 0.305

[‐2.90] [1.28] [0.27] [1.55] [‐1.62] [1.40] [‐1.71] [1.12]

ROA (t‐1) ‐0.086 0.462* ‐0.045 0.209 ‐0.293 0.420 0.029 0.399*

[‐0.33] [1.92] [‐0.15] [0.83] [‐0.97] [1.42] [0.11] [1.75]

Volatility 0.596* 0.332 0.373 0.240 0.817** 0.274 0.246 0.298

[1.81] [1.23] [1.03] [0.72] [2.06] [0.75] [0.85] [1.18]

Diversification dummy ‐0.059 ‐0.026 0.031 ‐0.023

[‐1.59] [‐0.64] [0.88] [‐0.57]

Stock return 0.136*** 0.046 0.222*** 0.218*** 0.112*** 0.051 0.228*** 0.203***

[3.99] [1.39] [6.26] [6.38] [2.94] [1.36] [6.81] [6.75]

Stock return (t‐1) 0.236*** 0.146*** 0.264*** 0.219*** 0.234*** 0.167*** 0.263*** 0.201***

[8.05] [5.42] [9.13] [8.01] [7.13] [5.49] [9.27] [8.20]

Cash 0.817*** 0.314** ‐0.219 ‐0.088 0.739*** 0.222 0.254 0.145

[5.69] [2.16] [‐0.85] [‐0.38] [4.24] [1.16] [1.29] [0.82]

R&D 1.304** ‐1.853** 2.845*** ‐1.237 2.440*** ‐4.133*** 1.582** 0.011

[2.03] [‐2.02] [3.78] [‐1.02] [3.84] [‐3.47] [2.17] [0.010]

CAPEX 0.304 ‐0.382 ‐0.159 1.380*** 0.023 0.042 0.013 0.591

[0.68] [‐0.90] [‐0.32] [2.79] [0.046] [0.076] [0.030] [1.52]

Firm age ‐0.003* ‐0.002 0.000 ‐0.001 ‐0.000 ‐0.002 ‐0.000 ‐0.002

[‐1.72] [‐0.68] [0.48] [‐0.34] [‐0.088] [‐0.20] [‐0.42] [‐0.87]

Constant 3.469*** 5.447*** 3.560*** 4.926*** 3.490*** 5.837*** 3.305*** 5.037***

[14.2] [15.6] [11.1] [12.4] [11.6] [12.8] [13.5] [16.1]

Observations 5146 5146 6307 6307 4325 4325 7128 7128

R‐squared 0.327 0.215 0.442 0.283 0.469 0.220 0.535 0.264

Year dummies Yes Yes Yes Yes Yes Yes Yes Yes

Industry dummies Yes No Yes No Yes No Yes No

Small firms Large firms Standalone firms Diversified firms

37

Table 10 CEO Compensation and Turnover and Switch of CEO Type

This table presents estimates of OLS panel regressions of the logarithm of CEO total compensation on switch of CEO type dummy variables and other CEO- and firm-level control variables. A CEO is classified as a generalist (specialist) if he has general ability index above (below) the median. The switch of CEO type dummy variables are: (1) if there is a CEO turnover but no switch of CEO type (no switch of CEO type); (2) if a specialist CEO is replaced by a generalist CEO hired from outside the firm (switch to generalist – external hire); (3) if a generalist CEO is replaced by a specialist CEO hired from outside the firm (switch to specialist – external hire); and (4) if the new CEO is internally appointed (switch – internal hire). The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4)

Total 

compensation 

(log)

Cash 

compensation 

(log)

Equity 

compensation 

(log)

Excess  

compensation 

(log)

No switch of CEO type ‐0.036 ‐0.099*** 0.097** ‐0.052*

[‐1.20] [‐5.25] [2.15] [‐1.72]

Switch to generalist ‐ external hire 0.262*** ‐0.059 0.375*** 0.194**

[2.90] [‐1.18] [2.84] [2.14]

Switch to specialist ‐ external hire 0.199* 0.029 0.179 0.211*

[1.73] [0.47] [1.00] [1.72]

Switch ‐ internal hire ‐0.129*** ‐0.153*** 0.017 ‐0.117***

[‐3.25] [‐6.32] [0.30] [‐3.02]

CEO tenure ‐0.007*** ‐0.001 0.001 ‐0.006**

[‐2.74] [‐0.49] [0.31] [‐2.29]

External hire dummy 0.107*** 0.054*** 0.265*** 0.077**

[3.27] [2.61] [5.99] [2.35]

CEO age ‐0.000 0.005*** ‐0.008*** ‐0.000

[‐0.11] [3.53] [‐2.70] [‐0.099]

Sales (log) 0.434*** 0.258*** 0.541*** 0.340***

[32.4] [31.0] [36.0] [28.0]

Leverage 0.368*** 0.183*** 0.441*** 0.366***

[3.87] [2.84] [3.39] [4.14]

Tobin's Q 0.131*** ‐0.002 0.298*** 0.103***

[6.72] [‐0.18] [12.1] [5.36]

ROA ‐0.532** 1.021*** ‐1.866*** ‐0.063

[‐2.43] [7.62] [‐6.22] [‐0.30]

ROA (t‐1) ‐0.110 ‐0.631*** 0.422 ‐0.118

[‐0.55] [‐4.98] [1.61] [‐0.60]

Volatility 0.470* ‐0.752*** 1.554*** 0.529**

[1.86] [‐4.80] [4.45] [2.04]

Diversification dummy ‐0.014 0.015 ‐0.047 0.014

[‐0.51] [0.87] [‐1.19] [0.51]

Stock return 0.175*** 0.235*** ‐0.097*** 0.153***

[6.94] [14.6] [‐2.73] [5.94]

Stock return (t‐1) 0.255*** 0.150*** 0.234*** 0.177***

[12.0] [11.7] [7.79] [8.05]

Cash 0.452*** 0.145* 0.861*** 0.463***

[3.19] [1.70] [5.37] [3.43]

R&D 2.103*** 0.491 3.090*** 2.575***

[4.05] [1.57] [4.99] [6.70]

CAPEX ‐0.080 ‐0.771*** ‐0.251 0.382

[‐0.23] [‐3.36] [‐0.46] [1.14]

Firm age 0.000 0.002*** ‐0.004*** ‐0.000

[0.24] [4.12] [‐3.41] [‐0.36]

Constant ‐0.033 0.018 ‐0.511*** ‐0.003

[‐0.86] [0.79] [‐5.60] [‐0.077]

Observations 11469 11443 8020 11457

R‐squared 0.500 0.558 0.487 0.343

Industry dummies Yes Yes Yes No

38

Table 11 Robustness of CEO Compensation and General Managerial Ability Index

This table presents estimates of OLS and firm fixed effects panel regressions of the logarithm of CEO total compensation on the general managerial ability index and other CEO- and firm-level control variables. Alternative definitions of the general managerial ability index are used: index 1 uses past number of firms, past CEO dummy, and general education dummy; index 2 uses past number of industries (two-digit SIC), past CEO dummy, and general education dummy; and index 3 uses past number of industries (four-digit SIC), past number of CEO positions, and general education dummy. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4) (5) (6)

OLS Fixed effects OLS Fixed effects OLS Fixed effects

General ability index 0.076*** 0.037*** 0.061*** 0.040*** 0.076*** 0.037***

[5.93] [3.08] [4.78] [3.12] [5.93] [3.08]

CEO tenure ‐0.006** 0.000 ‐0.006** ‐0.000 ‐0.006** 0.000

[‐2.27] [0.13] [‐2.44] [‐0.021] [‐2.27] [0.13]

External hire dummy 0.089*** 0.168*** 0.106*** 0.176*** 0.089*** 0.168***

[2.74] [4.68] [3.33] [5.02] [2.74] [4.68]

CEO age ‐0.001 ‐0.003 ‐0.001 ‐0.002 ‐0.001 ‐0.003

[‐0.64] [‐1.04] [‐0.38] [‐0.97] [‐0.64] [‐1.04]

Sales (log) 0.426*** 0.292*** 0.427*** 0.294*** 0.426*** 0.292***

[31.8] [10.9] [32.0] [11.0] [31.8] [10.9]

Leverage 0.349*** ‐0.333*** 0.372*** ‐0.339*** 0.349*** ‐0.333***

[3.66] [‐3.41] [3.91] [‐3.46] [3.66] [‐3.41]

Tobin's Q 0.128*** 0.093*** 0.128*** 0.092*** 0.128*** 0.093***

[6.65] [5.18] [6.63] [5.14] [6.65] [5.18]

ROA ‐0.511** 0.380* ‐0.531** 0.369* ‐0.511** 0.380*

[‐2.34] [1.74] [‐2.43] [1.70] [‐2.34] [1.74]

ROA (t‐1) ‐0.083 0.374** ‐0.093 0.383** ‐0.083 0.374**

[‐0.42] [2.14] [‐0.47] [2.19] [‐0.42] [2.14]

Volatility 0.461* 0.289 0.533** 0.288 0.461* 0.289

[1.84] [1.36] [2.12] [1.35] [1.84] [1.36]

Diversification dummy ‐0.016 ‐0.042 ‐0.018 ‐0.042 ‐0.016 ‐0.042

[‐0.60] [‐1.47] [‐0.68] [‐1.48] [‐0.60] [‐1.47]

Stock return 0.180*** 0.132*** 0.178*** 0.133*** 0.180*** 0.132***

[7.19] [5.72] [7.11] [5.74] [7.19] [5.72]

Stock return (t‐1) 0.257*** 0.187*** 0.255*** 0.187*** 0.257*** 0.187***

[12.2] [9.80] [12.1] [9.79] [12.2] [9.80]

Cash 0.475*** 0.217* 0.489*** 0.219* 0.475*** 0.217*

[3.41] [1.69] [3.47] [1.71] [3.41] [1.69]

R&D 1.917*** ‐1.951*** 2.011*** ‐1.952*** 1.917*** ‐1.951***

[3.74] [‐2.60] [3.91] [‐2.60] [3.74] [‐2.60]

CAPEX 0.006 0.418 ‐0.011 0.435 0.006 0.418

[0.017] [1.28] [‐0.031] [1.33] [0.017] [1.28]

Firm age 0.000 ‐0.001 0.000 ‐0.001 0.000 ‐0.001

[0.36] [‐0.34] [0.16] [‐0.39] [0.36] [‐0.34]

Constant 3.477*** 3.416*** 3.477***

[14.5] [4.78] [14.5]

Observations 11469 11469 11469 11469 11469 11469

R‐squared 0.504 0.263 0.502 0.263 0.504 0.263

Year dummies Yes Yes Yes Yes Yes Yes

Industry dummies Yes No Yes No Yes No

Index 1 Index 2 Index 3

39

Table 12 Conglomerates Excess Compensation

This table presents estimates of OLS and firm fixed effects panel regressions of the conglomerates excess compensation on the firm’s diversification dummy and number of business segments and other CEO- and firm-level control variables. The excess compensation is defined as the logarithm of the ratio of the CEO total compensation to its imputed compensation from standalone firms that match the firm’s current business segments. The sample consists of Execucomp firms for which CEO profile data are available from BoardEx in the 1993-2007 period. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix. Robust t-statistics adjusted for firm-level clustering are reported in brackets. *, **, *** indicates significance at the 10%, 5% and 1% levels.

(1) (2) (3) (4) (5) (6)

OLS

Fixed 

effects OLS

Fixed 

effects OLS

Fixed 

effects

Diversification dummy 0.258*** 0.119*** 0.249*** 0.115***

[7.48] [3.29] [7.30] [3.17]

Nr. segments 0.042*** 0.017**

[3.88] [2.24]

General ability index 0.071*** 0.040**

[4.29] [2.31]

CEO tenure ‐0.009*** 0.002 ‐0.009*** 0.002 ‐0.009*** 0.002

[‐2.96] [0.83] [‐2.89] [0.79] [‐2.63] [0.85]

External hire dummy 0.106*** 0.198*** 0.104*** 0.199*** 0.083** 0.179***

[2.90] [5.06] [2.83] [5.10] [2.21] [4.53]

CEO age 0.002 ‐0.005* 0.003 ‐0.005* 0.002 ‐0.006**

[0.98] [‐1.84] [1.12] [‐1.78] [0.67] [‐2.00]

Sales (log) 0.343*** 0.229*** 0.340*** 0.233*** 0.336*** 0.227***

[23.1] [8.16] [23.5] [8.39] [23.0] [8.10]

Leverage 0.515*** ‐0.204* 0.511*** ‐0.196* 0.516*** ‐0.208*

[5.00] [‐1.78] [4.90] [‐1.72] [5.03] [‐1.82]

Tobin's Q 0.122*** 0.107*** 0.112*** 0.106*** 0.118*** 0.107***

[5.52] [5.01] [5.14] [4.94] [5.40] [4.99]

ROA ‐0.227 0.191 ‐0.165 0.197 ‐0.224 0.192

[‐0.88] [0.74] [‐0.64] [0.76] [‐0.87] [0.74]

ROA (t‐1) ‐0.422* ‐0.021 ‐0.453* ‐0.032 ‐0.405* 0.002

[‐1.82] [‐0.098] [‐1.96] [‐0.15] [‐1.74] [0.0077]

Volatility ‐0.169 ‐0.079 ‐0.205 ‐0.060 ‐0.080 ‐0.098

[‐0.54] [‐0.31] [‐0.66] [‐0.24] [‐0.26] [‐0.39]

Stock return 0.105*** 0.059** 0.118*** 0.063** 0.249*** 0.115***

[3.47] [2.13] [3.93] [2.27] [7.30] [3.17]

Stock return (t‐1) 0.208*** 0.142*** 0.215*** 0.144*** 0.106*** 0.062**

[7.95] [6.01] [8.25] [6.05] [3.51] [2.25]

Cash 0.366** 0.198 0.325** 0.193 0.207*** 0.144***

[2.54] [1.36] [2.25] [1.32] [7.93] [6.12]

R&D 0.964** ‐2.637*** 0.898** ‐2.615*** 0.411*** 0.184

[2.24] [‐3.14] [2.11] [‐3.14] [2.86] [1.26]

CAPEX ‐0.040 0.158 ‐0.242 0.140 0.766* ‐2.489***

[‐0.097] [0.39] [‐0.59] [0.35] [1.78] [‐2.97]

Firm age 0.002** ‐0.002 0.003*** ‐0.002 0.096 0.127

[2.40] [‐0.62] [2.60] [‐0.59] [0.23] [0.32]

Constant ‐2.758*** ‐1.315*** ‐2.698*** ‐1.329*** ‐2.677*** ‐1.272***

[‐15.6] [‐5.18] [‐15.1] [‐5.26] [‐15.4] [‐5.02]

Observations 9972 9972 9972 9972 9975 9975

R‐squared 0.367 0.112 0.361 0.111 0.373 0.113

Year dummies Yes Yes Yes Yes Yes Yes

Industry dummies Yes No Yes No Yes No

40

Figure 1 CEO Compensation and General Managerial Ability Index

This table presents average score of the general managerial ability index and average CEO total compensation per year from 1993 to 2007.The sample consists of Execucomp firms for which CEO profile data are available from BoardEx. Average CEO excess compensation for the sample of CEOs that worked in more than one four-digit SIC industry (multi-industry CEOs) are also shown. All variables are winsorized at the 5th and 95th percentile values. Variable definitions and data sources are provided in the Appendix.

‐0.2

‐0.15

‐0.1

‐0.05

0

0.05

0.1

0

1,000

2,000

3,000

4,000

5,000

6,000

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Average

 General Ability Index

Average

 CEO

  Compensation ($ thousands)

Average general ability  index Average total compensation Average excess compensation of multi‐industry CEOs

41

Figure 2 Distribution of General Managerial Ability Index

This figure presents the cross-sectional distribution of the scores of the CEOs on the general managerial ability index in 1993-2000 (Panel A) and 2001-2007 (Panel B). The index of general managerial ability is the first factor of applying the principal components method to three proxies of general managerial ability (past number of industries, past CEO dummy, and general education dummy). The sample consists of Execucomp firms for which CEO profile data are available from BoardEx. Variable definitions and data sources are provided in the Appendix.

Panel A: 1993-2000

Panel B: 2001-2007

0.1

.2.3

Fra

ctio

n

-2 0 2 4 6Scores for factor 1

0.1

.2.3

Fra

ctio

n

-2 0 2 4 6Scores for factor 1

42

Appendix A: Definition of Variables and Data Sources

Variable Description

Panel A: CEO Compensation

Total compensation Total CEO Compensation in thousand $, which consists of salary, bonus, value of restricted stock granted, value of options granted, long term incentive payout and other compensation (Execucomp TDC1).

Cash compensation Salary plus bonus in thousand $ (Execucomp TOTAL_CURR).

Equity compensation Value of restricted stock granted plus value of options granted in thousand $ (Execucomp RSTKGRNT + OPTION_AWARDS_BLK_VALUE).

Excess compensation Difference between the CEO total compensation and its imputed compensation from single-industry CEOs that match the CEO past industry experience. The imputed compensation is the average compensation of the portfolio of industries where the CEO worked where the industry-level compensation is the median compensation of CEOs that worked only in a industry up to given year (single-industry CEOs).

Conglomerate excess compensation

Difference between the CEO total compensation and its imputed compensation from standalone firms that match the individual business segments where the firm currently operates. The imputed compensation is the average compensation of the portfolio of industries based on firm’s current business segments where the industry-level compensation is given by the median compensation of CEOs in stand-alone firms.

Panel B: CEO Characteristics

General ability index First factor of applying the principal components method to three proxies of general managerial ability: past number of industries, past CEO dummy, and general education dummy.

Past number of industries Number of industries (four-digit SIC) a CEO worked based on his work experience in publicly-listed firms prior to the current position.

Past CEO dummy Dummy variable that takes the value of one if a CEO held a CEO position at another company based on his work experience in publicly-listed firms prior to the current position.

General education dummy Dummy variable that takes the value of one if a CEO has a business degree (undergraduate, master, MBA or PhD).

CEO tenure Number of years as CEO in the current position.

External hire dummy Dummy variable that takes the value of one if the CEO was hired from outside of the firm.

CEO age Age of CEO in years.

Past multi-industry dummy Dummy variable that takes the value of one if the number of industries (four-digit SIC) where a CEO worked based on his work experience in publicly-listed firms prior to the current position is greater than one.

Past number of firms Number of firms a CEO worked based on his past work experience (prior to the current position) in publicly-listed firms.

43

Appendix (continued)

Panel C: CEO Turnover

No switch of CEO type Dummy variable that takes the value of one if there is a CEO turnover and the firm does not switch of CEO type (Execucomp).

Switch to generalist - external hire

Dummy variable that takes the value of one if the firm switches of CEO type from a specialist (i.e., CEO with general ability index below the yearly median) to a generalist (i.e., CEO with general ability index above the yearly median) and the CEO is hired from outside the firm (Execucomp).

Switch to specialist - external hire

Dummy variable that takes the value of one if there the firm switches of CEO type from a generalist (i.e., CEO with general ability index above the yearly median) to a specialist (i.e., CEO with general ability index below the yearly median) and the CEO is hired from outside the firm (Execucomp).

Switch - internal hire Dummy variable that takes the value of one if the firm switches of CEO type by internal promotion (Execucomp).

Panel D: Firm Characteristics

Sales Log of sales in thousands of US$ (Compustat SALE).

Leverage Total debt divided by total assets (Compustat TD/AT) .

Tobin’s Q Sum of total assets plus market value of equity minus book value of equity divided by total assets (Compustat (AT+CSHOxPRCC_F-CEQ)/AT)).

ROA Earnings before interest, taxes, depreciation and amortization divided by total assets (Compustat EBITDA/AT)

Volatility Annualized standard deviation of monthly stock returns (CRSP).

Diversification dummy Dummy variable that takes the value of one if a firm has more than one business segment (Compustat).

Stock return Annual stock return (Compustat (PRCC_F(t)/AJEX(t)+DVPSX_F(t)/AJEX(t))/(PRCC_F(t-1)/AJEX_F(t-1))).

Cash Cash and short-term investments divided by total assets (Compustat CHE/AT).

R&D R&D expenses divided by total assets (Compustat XRD/AT).

CAPEX Capital expenditures divided by total assets (CAPX/AT).

Firm age Number of years since a firm has listed its shares (CRSP).

Number of segments Number of business segments (Compustat).