University of Edinburgh Business School

Heads I win, Tails you lose? A career analysis of

executive pay and corporate performance∗

Ian Gregory-Smith†

Brian Main‡

January 2012

Abstract

The paper utilises a novel career perspective to examine managerial theories of organisa-tional control in the context of executive reward. Principal-agent theory, managerial powerand neo-institutionalism are evaluated. Detailed career histories of boardroom executivesin all FTSE350 companies between 1996 and 2008 are utilised. Using both a fixed effectspanel and a career based approach, rival hypotheses are tested. In addition to the pay-performance relationship, the probability of job loss as a function of performance is alsoestimated. The evidence presented points to the importance of institutional considerationsin the standard agency theory and managerial power perspectives, and lends support toa social theory of agency. The analysis is empirical in nature and an effort is made todraw on qualitative studies to further flesh out the discussion. The practical implicationsof the findings point to the inadequacy of existing arrangements to deliver efficient payoutcomes where performance is poor whilst highlighting dangers of relying on naive policyremedies that ignore the role played by institutional forces in their implementation. Alsohighlighted is the distinct advantage of boards adopting a cumulative or career-orientedapproach when evaluating executive performance. From a policy perspective, the case ismade for truly long term incentives, in the form of ‘Career Shares’.

Key words: Career Shares; Executive Pay; Pay-for-Performance; Social Agency Theory

∗Brian Main is grateful for research support under ESRC Grant: RES-062-23-0904.†University of Edinburgh‡Address for correspondence: University of Edinburgh Business School, 29 Buccleuch Place, Edinburgh,

EH8 9JS. Tel +44 (0) 131 6511375. E-mail: brian.main@ed.ac.uk

1

INTRODUCTION

Currently there is a disconnect between pay and performance in UK boardrooms. While the

public, in general, and business analysts, in particular, express disquiet regarding the level

of executive remuneration, it is this apparent disconnect between company performance de-

livered and remuneration received that provokes most concern regarding the effectiveness of

boards in securing value for money from executive directors (BIS, 2011; High Pay Commission,

2011; Hutton, 2010). The paper moves away from the conventional year-on-year analysis of

pay and performance to suggest that a richer insight can be obtained by using a cumulative

or career-based perspective. This allows the role of corporate governance to be evaluated in

a new light. In so doing, it provides support for the context-specific interpretation of agency

theory as suggested by Bruce et al. (2005); Gomez-Mejia et al. (2005); Wiseman et al. (2012).

Table I illustrates just why some concern regarding a disconnect between pay and performance

might well be justified. Using data introduced in detail below, the Table describes all executive

director careers in the FTSE350 that start and end between 1996 and 2008. The top panel

reports the distribution of total remuneration realised over these careers, both through direct

cash payments and through gains realised from equity-linked long term incentives (options,

performance share plans, etc.), all expressed in £2008m. Directors are then grouped according

to whether their shareholders were better off (value creators) or worse off (value destroyers)

at the end of the career period in question. It can be seen that the upper quartile of value

destroyers receive a reward at least as great as the typical (median) value creating executive

(£2.4m versus £2.0m). 1% of the value-destroying directors in our sample (21 directors) left

the shareholders of their companies in a worse state than when they started, and yet took

home in excess of £14.6m each. The data that underpin this Table are analysed in greater

detail below.

Over last 10 years there has also been a clear disconnect at the market level. This is apparent,

for example, through the lack of connection between the mean total CEO pay in the FTSE

2

100 and the annual variation in the FTSE 100 performance index (BIS, 2011, Fig 3, p11).

This disconnect at market level is observed even after a concerted effort to reform corporate

governance in the UK, as promulgated in a series of high profile reports on the topic (Cadbury,

1992; Greenbury, 1995; Hampel, 1998; Higgs, 2003; Walker, 2009) and codified in successive

versions of the UK’s Corporate Governance Code (FRC, 2010) and by the UK’s financial

regulatory authority (FSA, 2010). Furthermore, there is ample evidence that executive con-

tracts contain a high proportion of at-risk or performance-dependent pay (BIS, 2011, Fig 2,

p9). Indeed, qualitative studies of the remuneration decision process within boards generally

paint a picture of well intentioned independent directors striving to craft remuneration ar-

rangements that are both competitive in the executive market place and stretching in terms

of performance linkages, e.g., Bender and Moir (2006); Clarke and Conyon (1998); Lincoln

et al. (2006); Main et al. (2008).

So where does the relationship break down? It could be that there is indeed a strong linkage

between performance and remuneration, but that commentators observe and compare the out-

comes in a way that masks this relationship. For example, pay may be observed in a particular

year but the total remuneration realised pertains to performance over not only that particular

year but also (thanks to long-term incentives) to performance over the previous three years.

Similarly, total remuneration as awarded is in part contingent on performance over the coming

three years. However, the use of realised pay in a career framework, as presented in Table I

and discussed in detail below, avoids such timing pitfalls and, yet, continues to reveal a major

disconnect between pay and performance.

Such a disconnect not only offends the public’s sense of fair play (High Pay Commission, 2011),

but also runs contrary to a considerable body of management theory which predicts that at the

top of large companies the linkage of pay to performance should emerge as an important tool of

control over management behaviour (BIS, 2011; Core et al., 2003; Hutton, 2010). The tension

between the theoretically informed expectation and the observed empirical outcome invites

a more careful consideration of the theoretical underpinnings of the link between corporate

3

governance and directors’ remuneration. In particular, there is scope to assess the balance of

rival explanations of boardroom pay. The most direct of these is the principal-agent (Conyon

et al., 2011; Core and Larcker, 2002; Fama and Jensen, 1983; Gomez-Mejia et al., 2005;

Stathopoulous et al., 2005) or ‘contractarian’ (Davis, 2005) approach where non-executive

directors are seen as the representatives of shareholders, acting, among other things, to ensure

that remuneration arrangements are in the shareholders’ interests. In policy circles this is

usually alluded to under the ‘pay for performance’ label. Managerial power qualifies this

perspective by highlighting the asymmetry of information and bounded rationality issues that

raise the possibility of the process being captured by the incumbent management, who can

then extract generous pay awards for themselves (Bebchuk et al., 2002). From this perspective,

the presence of outside directors emerges as a key consideration in ensuring that shareholders

get a reasonable deal.

The situation is further complicated when consideration is given to the institutional context.

This weighs heavily when non-executive directors, in seeking legitimacy for their decisions re-

garding executive pay, fall back on being seen to do the right thing. They do this by following

the practice or guidance of others, while placing relatively little weight on designing pay to

serve as an incentive in its own right (DiMaggio and Powell, 1983; DiPrete and Eirich, 2010;

Eisenhardt, 1988). The field of directors’ remuneration provides abundant examples that are

consistent with such an explanation. When the ABI (ABI, 1987) suggested performance condi-

tions be applied to the vesting of executive share options and provided as an example earnings

per share (EPS) growth of inflation plus three percentage points, then suddenly the accepted

market practice became targeting EPS growth of exactly ‘RPI +3%’. Similarly, following

Greenbury (1995), the hitherto widespread use of three-year rolling contracts for boardroom

executives quickly disappeared to be replaced by a near universal one-year contract. While

the campaign for this change had been promoted for some time by the National Association

of Pension Funds (NAPF) and spearheaded by Alastair Ross Goobey, a managerial power

perspective might have predicted considerably more resistance than occurred once the Green-

bury Report (Greenbury, 1995) had been published. In a similar fashion: the adoption of

long term incentive plans in preference to share options that followed Greenbury (1995); the

4

abandonment of re-testing of performance conditions on long term incentives (ABI, 1999);

the move to having no award for less than median performance (ABI, 1999), and so on, are

innovations that were all quickly adopted with remarkably little push-back from boards or

executive directors.

The two important contributions of the paper are that, first, by utilising a career perspective

we offer a new lens through which to evaluate pay and performance relationships, thereby

providing fresh insights into the working of theories of corporate governance. Second, by

measuring remuneration in the form of realised pay, as opposed to pay awards, it avoids the

ambiguities of timing and uncertainty that beset such complex remuneration arrangements,

while offering a measure of reward that includes both cash or short term pay as well as the

longer term equity-based pay that has come to play a major part in executive reward (BIS,

2011). What results is a clear view of the extent to which non-executive representation on the

board can be viewed as a mechanism of control over executive reward and, by extension, exec-

utive decision making in general. Overall, our results point to the importance of incorporating

institutional considerations into the agency and managerial power explanations of board room

activity. In this, our conclusions regarding executive director remuneration enforce those of

Capezio et al. (2011) while using a wider and more robust measure of remuneration. In terms

of theory development and policy implications, our findings resonate with DiPrete and Eirich

(2010) and Wiseman et al. (2012).

THE GOVERNANCE OF DIRECTOR REMUNERATION

Theory

As explained above, much of the popular discussion of directors’ remuneration centres around

whether executives are being overpaid and whether shareholders are getting value for money

from the top executive team. These are far from trivial concerns and have received consid-

5

erable attention from scholars of management theory (Conyon et al., 2010; Core et al., 2003;

Frydman and Jenter, 2010; Murphy, 1999). There are, however, wider issues at stake. Pay is

seen as a major control device in providing managers with incentives to take decisions that are

in the shareholders interests (FSA, 2011, para592). The obverse of this being that observed

excess pay to executives can be taken as an indication of organisational slack (Wade et al.,

2006) and rent extraction by the executives (Bebchuk and Spamann, 2010).

In modern form, the identification of the control problem created by the separation of owner-

ship and control, and the accompanying development of a professional management class, is

generally attributed to Berle and Means (1932), although the issue was highlighted earlier by

Smith (1776). The concept of using remuneration design to address the problem by linking

pay with performance, and so aligning the interest of professional managers with those of

the shareholders (Murphy, 1985), emerges from the analysis of the widely-held corporation

as outlined by Jensen and Meckling (1976). This allows a pay mechanism to substitute for

imperfections in the direct supervision of the top management team (Murphy, 1999) - im-

perfections that arise owing to bounded rationality (Simon, 1947), asymmetric information

(Akerlof, 1970) and information impactedness (Williamson et al., 1975). The approach has

been characterised by Roberts (2001) as viewing the employment relationship as comprising

implicit and explicit contracts, and fits with the Alchian and Demsetz (1972) view of the firm

as a nexus of contracts. This is a world of optimal contracting (Demsetz and Lehn, 1985;

Hermalin and Weisbach, 2003) where enterprises ‘set optimal equity incentive levels’ (Core

and Guay, 2002, p.151)

This principal-agent or agency perspective on the management of enterprises with dispersed

ownership quickly came to dominate policy discussions in the field of directors’ remuneration,

and is commonly expressed in the much used triplet: ‘attract, retain and motivate’ (Green-

bury, 1995, para. 1.10). In academic circles, modeling behaviour in this manner was found to

be highly tractable (Grossman and Hart, 1983), although early empirical testing was less than

supportive, with the observed linkage between directors’ remuneration and their respective

6

company’s performance being found to be empirically so modest as to challenge the notion

that it could significantly influence decision making (Jensen and Murphy, 1990). However, as

time progressed, more aggressive use of long term share-based incentives in the boardroom

(Murphy, 2002) meant that later empirical estimates of the incentive effect rose to levels that

were more supportive of this perspective, and hence more able to explain directors’ remuner-

ation (Hall and Liebman, 1998). By this time, the theoretical perspective had deepened to

allow for the risk aversion of individual executives (Garen, 1994; Haubrich, 1994), and em-

pirical studies quickly incorporated this consideration to claim further support for the agency

interpretation (Hall and Murphy, 2002; Lambert et al., 1991). Growth in company size and

scarce managerial talent has also been used to explain the recent marked rise in director re-

muneration (Edmans and Gabaix, 2009; Gabaix and Landier, 2008).

The principal-agent framework soon came to be central to discussions of corporate gover-

nance: ‘Executives and shareholders can have divergent interests, ...... Remuneration struc-

tures should seek to address this.’ (ABI, 2011, para.v(c)). The potential vulnerability of the

process to being undermined by managerial or insider power (Bebchuk et al., 2002) imparts a

heightened importance to the role of non-executive directors. As indicated above, the recent

wave of reform of corporate governance in the UK is generally taken to date from the Cad-

bury Report (Cadbury, 1992) and, from the outset, the emphasis was on the inclusion and

role of non-executive or independent directors (Cadbury, 1992, para. 4.6). They are taken as

the embodiment of shareholder interests. As such, it is on their shoulders that responsibility

rests in terms of crafting reward arrangements such as to ensure executive directors see their

own interests as aligned with those of the shareholders. Failure in this task due to abuse of

managerial power is seen as leading to excess in executive director remuneration (‘rewards

for failure’ (BIS, 2011, p.22) being one manifestation). None of this, of course, rules out the

role of non-executives in directly monitoring performance (as would, in any case, be a neces-

sary part of any incentive pay scheme), or in providing general advice and counsel to the top

management team (Westphal, 1999). The non-executive directors themselves are assumed to

strive for such successful outcomes out of their own career or reputational concerns (Yermack,

7

2004) or from a professional and altruistic drive to pass on the advantage of their experience

to the next generation (Roberts et al., 2005).

Although it has come to dominate policy discussions of directors’ remuneration, the prin-

cipal agent perspective of directors’ remuneration as a designed incentive mechanism is not

without its critics. Surveys of the large numbers of empirical studies in the area (Dalton

et al., 2003; Devers et al., 2007; Tosi et al., 2000) find that estimates of the magnitude of

the pay-performance relationship remain empirically modest. These findings, in the face of

apparently continuously rising levels of reward (BIS, 2011; Hutton, 2010), lend strength to

alternative views of what determines executive remuneration - views that qualify the simple

agency theory picture presented above. The two main qualifications arise in the form of man-

agerial power (Pfeffer, 1981; Salancik and Pfeffer, 1977) and neo-institutionalism (Capezio

et al., 2011; DiMaggio and Powell, 1983; Scott, 1991). As indicated above, managerial power

sees the incumbent management (the executive directors) as exploiting their privileged po-

sition to extract generous levels of reward for themselves (Bebchuk and Fried, 2004). The

efforts of non-executive directors are needed to impose some discipline on this process which

is otherwise undermined by the power or influence of the incumbent top management team

(Bebchuk et al., 2002; Main et al., 1995). Restraints, if any, arise from a desire to stay

‘under the radar’ (Bebchuk et al., 2002, p.16), to avoid provoking ‘outrage’ (Bebchuk and

Fried, 2004, p.64) among shareholders or commentators. In such a situation, the prevalence

of non-executive directors is viewed as particularly important (Capezio et al., 2011; Gregory-

Smith, 2011). The Bebchuk critique of optimal contacting has been met by a spirited defence

(Conyon, 2006; Core et al., 2004; Hall and Murphy, 2003), but a meta analysis of 219 studies

in the area of managerial power and executive remuneration (van Essen et al., 2012) finds

that the prevalence of non-executive directors both lowers the level of CEO pay and increases

the pay-performance sensitivity.

The neo-institutional perspective (Main et al., 2008; Tarbert et al., 2008), places less faith in

the prevalence of non-executive directors. It generally accepts that they are well-intentioned

8

but, finding the design of executive director remuneration arrangements to be beset with

ambiguity and asymmetry of information (Akerlof, 1970; Williamson et al., 1975), and con-

fronted by their own bounded rationality (Simon, 1947), they are seen as reaching for certain

shorthand solutions to the problem. In such situations, they may find legitimacy: through a

‘mimetic’ process of following the lead of what other boards are doing (Devers et al., 2007;

DiPrete and Eirich, 2010; Porac et al., 1999; Zajac and Westphal, 1995); through a ‘coer-

cive’ process of conforming to existing regulatory codes (Barreto and Baden-Fuller, 2006); or

through a ‘normative’ process by drawing on their own experience-based standards as devel-

oped through service on other boards (DiMaggio and Powell, 1983; Meyer and Rowan, 1977;

O’Reilly et al., 1988; Perkins and Hendry, 2005; Scott, 2001). In each case, what emerges is

an isomorphism of practice whereby each board seems to follow what other boards are doing.

A driving consideration is legitimacy. This is a long way from the regular and periodic ad-

justment of the pay-performance relationship as envisioned in agency theory (Core and Guay,

1999; Core et al., 2003; Core and Larcker, 2002). In fact, rather than exerting a restraining

effect on the generosity or laxness of remuneration that abuse of managerial power threatens,

a higher presence of non executives may impart a sense of legitimacy and lead to more gener-

ous not less generous pay awards (Eisenhardt, 1988; Main and Johnston, 1993; Westphal and

Zajac, 1995).

In addition to the question of remuneration, there is always present the accompanying ques-

tion of continued employment. This distinction between the intensive margin and extensive

margin has long been recognised (Blundell et al., 2011; Heckman, 1993), and is no less im-

portant in the context of the demand for executive director services. Rather than more pay

for better performance (the intensive margin), directors can face the prospect of job loss for

poor performance (the extensive margin). Such binary outcomes (loss of employment versus

continuing employment) have been studied at boardroom level (Conyon, 1998; Conyon and

Florou, 2002; Gregory-Smith et al., 2009; Murphy, 1999; Weisbach, 1988) and a significant

performance sensitivity is observed. Clearly, if managerial power is a serious consideration

then it would be expected that this sensitivity would be enhanced where managerial power is

9

reduced. Everything expressed above regarding the strength of the pay-performance sensitiv-

ity can be re-expressed in terms of the performance sensitivity of job terminations. Empirical

results are more qualified here. Thus, Dahya et al. (2002) find that following the introduction

of the Cadbury (Cadbury, 1992) reforms in corporate governance the likelihood of job ter-

mination is more sensitive to poor performance, while Dedman (2003) find no change in the

relationship following these reforms. Reporting on the USA, Huson et al. (2001) also note a

failure of the sensitivity to change over a similar period of improved governance. The preva-

lence of non-executive directors can also be expected to make job tenure more performance

sensitive (Hermalin and Weisbach, 1998) and several papers find a significant relationship

between the performance sensitivity of CEO exit and the balance of non-executives on the

board (Gregory-Smith et al., 2009; Weisbach, 1988; Yermack, 1996).

Hypotheses

From the principal-agent perspective on remuneration as a control mechanism, the observed

connection between pay and performance should be empirically and statistically robust -

whether companies succeed or fail. Our career perspective offers a new framework in which

to test this well known hypothesis. It also allows us to generate some novel hypotheses on the

topic. Starting with the conventional approach, we can hypothesise:

Hypothesis 1: The observed pay-performance sensitivity is empirically and statis-

tically significant for successful and for unsuccessful companies.

Such a result should be even clearer in a career context where all the various performance

related pay-outs are taken into account and the timing ambiguities are minimised (Gong,

2010):

Hypothesis 2: The observed pay-performance sensitivity will be empirically larger

when measured over a director’s career when the effects of all long-term incentives

10

can be attributed as compared to estimates derived from annual observations of

remuneration and performance.

Whereas the principal-agent story regards the optimal board structure to be in place (Core

et al., 2003; Demsetz and Lehn, 1985), the managerial power perspective suggests that those

boards with a greater degree of outsider scrutiny will be subject to less abuse of managerial

power. Hence a greater preponderance of non-executives can be expected to lead to more

effective pay-performance contracting with the executive directors, as the more prevalent are

non-executive directors then the freer they will be from the influence of managerial power

when they negotiate the reward arrangements. It, therefore, follows that we can hypothesise

that, under the managerial power view, greater prevalence of non-executive directors should

imply greater value for money for shareholders:

Hypothesis 3: At any point a director’s career, for a given level of performance, the

level of executive reward will be lower the greater is the proportion of non-executives

on the board.

Continuing to adopt the managerial power view, the greater the prevalence of non-executives

on the board then the more emphasis will be paid to pay for performance. Thus in terms of

rate at which remuneration varies for a change in company performance:

Hypothesis 4: At any point a director’s career, the correlation between remunera-

tion earned and performance delivered will be higher the greater is the proportion

of non-executives on the board.

In a similar sense when it comes to terminating the careers of executive directors, the man-

agerial power perspective suggests that this should be more sensitive to performance (more

likely to happen when things go wrong) the greater the prevalence of non-executives on the

board. We can therefore hypothesise:

11

Hypothesis 5: The probability of seeing an executive director’s career terminate in

any year will be higher and will be more sensitive to performance in the presence

of a greater proportion of non-executives on the board.

It is difficult to hone a specific hypothesis to test the neo-institutional perspective, but were

reward to be much higher in the presence of a greater prevalence of non-executives then

this could be taken as an outcome of the effective legitimacy that a greater prevalence of

non-executives brings (Bender, 2004). The same could be said for an observation of lower

pay-performance sensitivity and performance-related career termination propensity.

The next section introduces the data used to test the hypotheses introduced above.

RESEARCH METHODS

Data

The source of the executive remuneration data used in this study is a commercial provider,

Manifest Information Services Ltd. This company has collected annual boardroom data on

all FTSE350 companies since 1995. These data are particularly rich in detail. Not only

are the cash payments of salary and annual bonus available, but so too are the details of all

realised gains through longer term incentive schemes such as share options, performance share

plans and so on. Personal details are available for each member of the board, in addition

to their precise start and end dates of service, thereby permitting the board composition

to be described at all times. Once a company is included in the sample frame, Manifest

continues to collect data on the company, even if it leaves the FTSE350, until it is wound up

or taken private. The Manifest data set used here extends through to 2008. For financial and

accounting data, DataStream is utilised.

12

Measures

The focal measure of remuneration used aims to capture realised remuneration. This includes

both cash payments in the form of salary and bonuses and also those gains realised from the

equity-linked long-term incentives such as executive share options or performance share plans.

The only source of financial reward that is not captured arises from pension benefits, which

are too imperfectly measured over the period to allow a consistent or credible treatment, here

or in any other previous UK study in the area. The measure of reward used is labelled ‘TDC’

to represent total direct compensation. This starts with total cash compensation in the form

of the director’s salary plus other cash payments such as any annual bonus received during

each year. To this is added the realised value of options and other equity based incentives

exercised during that year. By focusing on realised remuneration rather than remuneration

as awarded at grant date, it is possible to avoid the difficulty of calculating the expected

value of share options or performance share plans, both of which are generally subject to

quite complex performance conditions in the UK (Conyon and Murphy, 2000; Main, 2006).

This is not about what the board intended or wished should happen, but about what actually

transpired - realised reward. Focusing on realised remuneration also goes to the heart of the

debate, which concerns the effective link achieved between pay and performance. Adopting a

career perspective ensures any relationship between pay and performance is not obscured by

issues of timing wherein payments in one year actually refer, at least in part, to performance

delivered over an earlier time (Gong, 2010). All remuneration data are expressed in £2008.

The Manifest data base also reports on board membership and from this it is possible to com-

pute the size of board (‘Board’) as reported in each company’s Annual Report and Accounts,

and the percentage of these members who are non-executive (‘% NEDs’). For each executive

the date of birth permits age in years (‘Age’) to be computed. The start and end dates of

boardroom service are also recorded by Manifest and this is used to compute each executive’s

length of service in years as of the end of each financial year (‘Tenure’).

13

Additional company descriptive data are obtained from DataStream. As a control for com-

pany size, the logarithm of total sales is used (‘Sales’). Firm performance is primarily captured

by total shareholder return over the period in question (this reflects the return to holding the

share that arises both from dividend payments and changes in the price of the share). This is

available through the ‘RI’ index available in DataStream, where the start and end value of the

index is defined by the start and end of the relevant financial year or the start or end of the

executive’s career, when this occurs part way through a financial year. In order to express the

shareholder return as the total change in shareholder wealth, it is possible to use the average

total market capitalisation as reported in the ‘MV’ measure in DataStream. All financial data

are expressed in £2008. Summary statistics are presented in Table I.

Insert Table I: Career and Annual Summary Statistics

The upper panel in Table I presents the descriptive statistics for the 3,157 FTSE350 director

careers observed in their entirety between 1996 and 2008. The career controls are reported in

terms of their average value over the respective career. These 3,157 careers constitute 16,356

annual observations and descriptive statistics on those annual observations are provided in

the lower part of Table I. In the middle part of Table I, the career data is utilised to provide

an insight into the extent of the disconnect between pay and performance by separating out

‘value-creators’ (those directors where the shareholder of the respective company is better off

at the end of the career than at the start) from ‘value destroyers’ (those whose shareholders

are worse off at the end of the career than at the start). It has already been pointed out that

the median wealth-creating director realises a career reward of £2.1m that is lower than the

£2.4m received by the upper quartile value destroyer. In effect, 25% of the value-destroying

executives earn more than the median of the value-creating executives.

The histogram in Figure 1 makes use of these data on FTSE350 executive careers to contrast

the distribution of total remuneration enjoyed over each of these careers (in 2008£m) with the

performance of each executive’s respective company over the same period (measured as total

14

shareholder return). While the highest paid director in the sample (‘HPD’) can be seen to

both be richly rewarded and to deliver a high level of return to shareholders, the same cannot

be said for the other two illustrative examples in Figure 1 (labelled ‘Banker 1’ and ‘Banker

2’). There is an unambiguous “heads-I-win” and “tails-you-lose” aspect (Sanders, 2001) about

these results. Pay is right skewed - at worst, the career is brief and only moderately rewarding,

but at best it can be long and richly rewarding. On the other hand, shareholder returns over

these same careers are markedly more symmetric - while shareholders can gain much, they

also stand to lose it all, with past gains being easily swept away.

Insert Figure 1: Distribution of Pay and Performance

Hitherto, most research in this area has been structured around the connection between the

expected value of the observed annual award of executive pay and the performance of the

company in that year or in the previous year. While expected value has the advantage of

being forward looking, hence relevant for decision making, its calculation on an annual basis

can obscure the career pay-performance sensitivity (or lack thereof) - so-called long term in-

centives notwithstanding. Just one terrible year can destroy the pay-performance sensitivity

link on a career basis, as the“Banker 1” and “Banker 2” examples in Figure 1 illustrate. But

even absent cataclysmic career endings, our data reveal that keeping an eye on cumulative

pay and performance at each point in a director’s career allows the divergence between pay

and performance to become clear. Figure 2 plots for each career the cumulative position of

the shareholders and the cumulative position of the executive directors - at the end of their

first, second, third and so on year in office, for as long as they serve. While value-destroyers

render quite a distinct time path of value in shareholders’ hands (lower frame), they enjoy

a lower, but not radically different, path of personal prosperity over their career (upper frame).

Insert Figure 2: Value for Money

15

RESULTS

Pay performance elasticity

In terms of explaining the level of reward, Table II reports on regressions of the logarithm of

realised reward on a set of company and person specific descriptors. The top part of the Ta-

ble pertains to career-based regressions, while the lower part utilises annual data (controlling

for fixed effects). The initial estimate of the pay-performance sensitivity seems quite modest

(at 0.060). A test in column (2) demonstrates a significant difference in elasticity between

value creators and value destroyers. Separate estimation results in an elasticity of 0.17 for

value creators and 0.061 for value destroyers. A similar conclusion regarding the asymmetric

treatment of value creators and value destroyer is obtained using annual data (lower part of

Table II) but, importantly, the magnitude of the annual estimates is much higher (0.28 and

0.069, respectively) indicating that such a perspective underestimates the damage done to

shareholders during periods of under-performance. The annual estimates are similar to those

of 0.17 and 0.15 obtained by Bell and van Reenen (2011) and 0.32 obtained by Conyon et al.

(2011) using alternative UK data sources. Focusing on a narrower measure of cash pay, Guest

(2010) estimates the post-Cadbury elasticity at around 0.14 and Gregg et al. (2012) estimate

a pay-performance elasticity of 0.29 for 2006. Earlier analysis by Conyon and Murphy (2000)

produced an estimate of 0.12 (as opposed to 0.27 for the USA).

Separating the data into value-creators and value-destroyers, demonstrates that there is an

asymmetry in that the pay-performance sensitivity is significantly greater for value-creators,

with the connection for value destroyers being empirically more modest. This asymmetry of

treatment has been noted by Ezzamel and Watson (2002) and by Guest (2010). For value

creators, the career estimate of 0.170 indicates that by achieving a doubling of the increase

in shareholder wealth over their career the executives will be 17% better off. At the median,

producing an extra £304m in shareholder wealth results in an extra £368k on top of £2.1m

career earnings. On the other hand, for value destroyers the downside is less bleak. If instead

16

of merely destroying half of shareholder value (median return of value destroyers is -70.8%) a

full three quarters of value destroyed, the executives see their career reward fall from £1.31m

to £1.23m. While the former is consistent with prior estimates and empirically may be high

enough to sustain an agency perspective on arrangements, the downside picture points to-

wards rejection of Hypothesis 1.

The contrast between the upper and lower parts of Table II also reveals that the Career per-

spective presents a more modest pay-performance connection for all concerned. Using annual

data allows previous poor years to be discounted, and presents a more sanguine picture of the

pay-performance sensitivity than is justified from a career perspective. The advantage of the

cumulative or career approach is that the ambiguities of timing are avoided (Gong, 2010) but

the empirical estimates lead to a rejection of Hypothesis 2 and raise questions regarding the

‘optimal contracting’ perspective.

In all formulations presented in Table II, a larger size of board increases reward, and a higher

percentage of non-executives also increases the reward. These are statistically significant re-

sults and run contrary to the predictions of managerial power, while remaining consistent with

neo-institutional theory (as the larger and more non-executive dominated board feels greater

legitimacy in awarding generous pay awards). These results reject Hypothesis 3 and suggest

that, consistent with neo-institutional theory, there may be a legitimacy effect at work here as

the greater the conformance with governance codes (in terms of external representation on the

board) the more secure the board feels in designing highly remunerating reward arrangements.

Insert Table II: Pay-performance Elasticities. Career and Annual

17

Cumulative pay performance elasticities

The role of cumulative pay is further investigated in the pay regressions of Table III. Here,

rather than the career total or annual observation of pay (as in Table II), the cumulative record

of each director is examined as at the end of each year of that director’s career. This is the pic-

ture that would emerge if the remuneration committee incorporated all of the director’s record

to date when judging pay and performance. Once again, a marked and statistically significant

difference in outcomes is observed between the value-creators and the value-destroyers. For

value creators an elasticity with respect to cumulative performance of 0.19 is observed. For

the value-destroyers the effect fails to reach statistical significance - enforcing the impression

gathered from the descriptive statistics discussed above - that of a world of ‘heads I win and

tails you lose’.

As before, and consistent with the conclusions already drawn regarding Hypothesis 3, the

larger the board and the higher the proportion of non-executives on that board the greater

the level of realised pay. In addition, the interaction term between the % NEDs and the cumu-

lative TSR is significant in all cases, suggesting that the pay performance sensitivity is higher

for boards with a higher proportion of non-executives. Thus, not only does a greater presence

non-executives increase the level of remuneration but it also heightens the pay-performance

sensitivity - for both value creators and value destroyers. Such a result is consistent with

agency theory whereby the executive is offered a higher expected level of reward in return

for accepting a greater degree of pay at risk (Aggarwal and Samwick, 1999). This could also

be consistent with the isomorphism in practice that neo-institutional theory predicts. Under

the neo-institutional view, the higher the prevalence of non-executives then the greater will

be the pressure to act as other boards, and pressure to tie reward to performance will be

more forcefully put into practice - possibly aided by remuneration consultants who spread

‘best practice’ (Conyon et al., 2010). These interactions are plotted in Figure 3, where it

can be seen that the pay-performance elasticity is increasing in %NEDs but always higher for

value-creators than for value-destroyers.

18

The evidence here makes it impossible to reject Hypothesis 4. The managerial power inter-

pretation would predict that a greater preponderance of non-executives would leave the board

better able to craft a reward arrangement that leaves the executive directors facing a high

pay-performance sensitivity.

Insert Table III: Pay-performance Elasticities. Career and Annual

Insert Figure 3: Pay-performance Elasticity and percentage of non-executive

directors on the board

Probability of career exit

Of course, the principal-agent control effect of higher representation of non-executives could

be manifesting itself in the extensive form of the employment relationship, that is, through

career termination. To pursue this possibility, a probit estimation is reported in Table IV. In

addition to person and company-specific control variables, the level of cumulative reward paid

to date is included. As before, separating the sample into value-creators and value-destroyers

is statistically justified, but fails to reveal any significant role for non-executives - either in

their number, through the size of the board, or through their proportional representation

on the board. There is a significant performance relationship, with those directors currently

performing better being less likely to lose their post. Interestingly, for value-destroyers the

cumulative record of performance also weighs significantly - so that current poor performance

can be offset by a previous good record. This suggests an accumulated reservoir of goodwill

or credibility is an important feature in determining termination decisions.

However, there is no significant role observed for board size or the proportion on non-

executives. This remains true when an interaction term in introduced between %NEDs and

TSR. This evidence indicates that Hypothesis 5 can be rejected. The expected role of non-

executives in enhancing the extensive aspect of the pay performance relation (having one’s

19

career terminated for poor performance) is not observed. Even when the interacted with the

(significant) company performance variable (‘TSR’), the proportion of non-executives on the

board fails to demonstrate any significance.

Insert Table IV: Probit Estimates: Value Creators versus Value Destroyers

DISCUSSION

Summary of Findings

Using a large sample of executive director careers that are observed between 1996 and 2008,

it has been shown that in a substantial proportion of cases there is a clear disconnect between

pay and performance. Analysis reveals that the pay performance elasticity that lies at the

heart of the principal agent perspective is empirically more modest than conventional analysis

suggests. By adopting a career perspective, it is shown that the pay to performance sensitivity

is significantly lower than estimates based on annual observations suggest. This is particularly

the case for those directors whom we label value-destroyers (because they leave shareholders

worse off at the end of their career than at the beginning). These results are obtained us-

ing a wide measure of remuneration that includes the effects of long term incentives such as

executive share options and performance shares. To avoid ambiguities regarding expected

values, realised evaluations of remuneration are utilised. More importantly, the use of realised

values speaks directly to the public debate in this area which concerns the wide perception of

a disconnect between performance achieved and remuneration actually received (BIS, 2011;

High Pay Commission, 2011; Hutton, 2010). The debate centres around realised remuneration.

Far from exercising a restraining influence on pay, the presence of non-executive directors is

seen to enhance the level of reward - whether measured on a career basis or annually. Ex-

ploring the promise of a more integrative perspective on pay and performance, analysis of

20

cumulative remuneration as a function of cumulative performance is shown to be successful

in revealing the difference in pay-performance alignment between the value-creators and the

value-destroyers. The level and proportion of non-executives are, once again, sources of higher

not lower levels of remuneration. It is demonstrated in this context, however, that the pres-

ence of non-executives does significantly enhance the pay-performance elasticity for executive

directors.

Analysis of the length of executive director careers and the probability of dismissal finds that,

consistent with much of the literature, poor performance does lead to dismissal. While the

exit probabilities of value-creating directors depend only on contemporaneous performance,

for value-destroyers the cumulative record performance also plays a role, suggesting that there

is some institutional memory that mitigates current performance in the light of past success.

But the presence of non-executive directors does not sharpen the performance sensitivity of

exit - not even when allowing for an interaction with performance.

Contribution of the Study

The contribution of the study is to offer a new and more robust framework in which to scru-

tinise the observed pay and performance of executive directors from the perspective of rival

theories of management control. The novelty arises because rather than focusing on an an-

nual, year-by-year, set of observations of director remuneration, a whole-career approach is

adopted. This avoids the problem of which elements of remuneration to include (particularly

acute when dealing with long-term incentives). The resulting observations summarize the job

performed by the board in overseeing the reward aspects of the director’s career. A second

source of novelty is that the study utilises realised pay. This is an outcomes-based mea-

sure that avoids the problem of estimating expected outcomes - a problem that arises when

awarded remuneration is utilised and which has become increasingly acute with the advent in

the UK of markedly more idiosyncratic relative performance criteria, where the exact compo-

21

sition of the comparator peer group can vary from company to company and, indeed, from

time period to time period. Realised pay avoids such measurement complexities by simply

recording the outcomes enjoyed (or not) by the executive. As a measure, it speaks directly

to current policy concerns in this area (BIS, 2011; High Pay Commission, 2011; Hutton, 2010).

By minimising any potential mismatch between observed remuneration and performance, and

by avoiding the complexities of valuation that normally beset work in this area, this study

provides a more revealing focus on the key issue in management theory of how good a job

the board does in shaping the remuneration arrangements of the executive directors. The re-

sulting observations suggest that the pay-performance connection is weaker than alternative

perspectives suggest. The observed impact of non-executive directors seems to accord more

with an neo-institutional explanation of their role than that put forward by principal agent

theory or managerial power. The implications of these findings are discussed in some detail

below.

Implications for Theory and Practice

One approach to reducing the potential for a disconnect between pay and performance of the

type documented above, would be to prevent executives from cashing-out long-term incen-

tives and thereby enjoying the rewards of early success before shareholders can be assured

that any improvement in performance is not transient. To this end, we argue that all vested

long-term incentive rewards should be held in company shares (‘Career Shares’, see Bebchuk

and Fried (2010); Bhagat and Romano (2009); Main (2011); Main et al. (2011)) until the

executive leaves the board - and possibly for a year or so longer, so ensuring the succession

process is successful. This would engineer an automatic cumulative perspective to reward.

Those achieving high performance early in their career would have their feet held to the fire

throughout the remainder of their time in office as the later share price impacts on the value

of their reward. Those who underperform in that later period would experience an automatic

22

settling-up or claw-back as a faltering share price reduces their reward. On the other hand,

initially hesitant or slow beginnings are quickly forgiven as a later high performing share price

lifts the value of all rewards. All of this occurs automatically.

The results presented above make clear the extent of the ‘heads I win, tails you lose’ (Sanders,

2001) type of outcome that has arisen in the area of director’s remuneration. By adopting

a career perspective we demonstrate that the achieved pay-performance sensitivity is more

modest than it might appear from simply examining annual outcomes. For the class of value-

destroyers it is particularly weak, and yet their rewards remain significant - something that is

distinctly at odds with principal-agent theory. The presence of non-executive directors seems

to add a sense of legitimacy (in terms of high levels of remuneration) and, although there is ev-

idence of a higher pay-performance sensitivity in terms of cumulative reward, non-executives

seem to have no significant impact on the probability of career termination. This combination

of findings points to a board better described by the mimetic behaviour of neo-institutional

theory than in terms of the mechanism-design portrait offered by agency theory. In the words

of (Perkins and Hendry, 2005, p.1464), ‘what matters is how rewards appear, not whether per-

formance is being objectively overvalued’. If this is true, then it may be more effective to devise

policy aimed at remuneration design rather than simply attacking the remuneration quantum.

This focus on the design of remuneration has a two-fold advantage in policy terms. First,

although many commentators (Hutton, 2010) see the reforms in corporate governance over

the last two decades as having brought no relief from a seemingly remorseless rise in executive

remuneration, the UK has enjoyed one marked success in this field. The composition of ex-

ecutive reward has been radically re-defined and re-shaped under pressure from institutional

investors. This influence of institutional investors is most visibly apparent in the Guidelines

of the Association of British Insurers (ABI, 1987, 2011), which have marked some quite ma-

jor transformations in the way executive remuneration arrangements are shaped. Examples

include: the increased ceiling on option grants; the switch to performance share plans (‘Long

Term Incentive Plans’) from options; the abandonment of the ‘re-testing’ of performance con-

23

ditions; the introduction of peer groups and relative performance; the move away from ‘cliff

vesting’; the imposition of zero payout for below median performance; and so on. It is pos-

sible to be critical of some of these developments (Main, 2006), but impossible to deny their

impact. Effective strategy in any field should always build on what is achievable - proximate

objectives as opposed to mistaking goals for strategy (Rumelt, 2011). While impacting on the

level of remuneration may have proved elusive, there is no denying the effectiveness of UK

institutions in shaping the structure of reward. In this sense, a move to ‘Career Shares’ is

entirely achievable, as has been explained above and discussed in detail elsewhere (Bebchuk

and Fried, 2010; Bhagat and Romano, 2009; Main et al., 2011).

The second advantage to focusing on design is that such an approach is robust against the

prisoner’s dilemma problem that besets decisions on director pay (Main, 2011; Pepper, 2006).

When deciding on a quantum of pay, the board all too often finds the dominant strategy is to

err on the generous side. Being generous when other companies are being generous is rational,

if only to pay the ‘going rate’. Being generous in a world where no one else follows represents

a modest (in a large company) extra expense, but results in a top management team who

feel valued and pyschologically more disposed to reciprocate with a consummate level of co-

operation to shareholder benefit (Dabos and Rousseau, 2004; Fehr et al., 1997). On the other

hand, restraining pay brings only negative consequences. At best, the company matches the

market, but there is a chance of underestimating the market. Underestimating the market

(underpaying), in a world of increased transparency (DTI, 2002), results in the prospect that

the top management team will experience disruptive unplanned exits as people move to better

paying jobs or, at the very least, become disaffected and demoralised, regarding themselves

as being undervalued (Gregory-Smith and Main, 2011).

The results presented in Tables I through IV suggest that theorising around the operations of

the company board might be more fruitful if there were more attention paid to institutional

influences and perhaps less concentration on unadorned agency theory modeling. As has been

demonstrated here, policy prescriptions aimed at repairing the relationship between company

24

performance and director remuneration promise to gain more traction if an institutionally

richer perspective is developed.

Limitations and Areas for Further Research

As with any study that utilises panel data, there is always a sense that more could have been

gained with a longer time series of data. The time series here ran out in 2008 and an obvi-

ous pointer for further research is to ask what happens in the subsequent recession. There

must also be an important role to be played by further qualitative work in this field. The

institutional perspective portrayed above relies, perforce, on the interpretation of certain key

variables - essentially the size of the board and the proportion who are non-executive. There

remains considerable scope for qualitative studies centred around the workings of the board

and of the remuneration committee in particular. There is an existing kernel of work in this

area, e.g., Bender (2003); Lincoln et al. (2006); Main et al. (2011, 2008); Ogden and Watson

(2008), but more remains to be done.

Summary

Taking its inspiration from the Sanders (2001) observation that executive remuneration has

about it an aspect of ‘heads I win, tails you lose’, this paper documents the particular failure

of boards to tie the fortunes of their executive directors to the experience endured by their

shareholders. It does so while using a robust measure of remuneration (realised remuneration)

and in a setting (the director career) that minimises timing issues. The observed association of

larger boards and greater proportions of non-executives with higher levels of realised executive

remuneration raises a question over the relevance of managerial power approach in this area.

While it is necessary to guard against falling back on an ‘over socialized’ (Granovetter, 1985)

view of boardroom activity, it does seem that policy in the area might be better designed if

took account of the mimetic behaviour of non-executive directors. In so doing it could go with

25

the grain of observed action by introducing ‘career shares’ in much the same way as institu-

tional pressure led to the widespread adoption of long term incentive plans (Ltips) following

the Greenbury Report (Greenbury, 1995; Main et al., 2008).

This does not imply an outright rejection of agency theory. As Wiseman et al. (2012) makes

clear, one can accept the basic premise of this approach (individuals with an interest in an

enterprise may have divergent aspirations) while allowing for the particular social setting

in which the action is taking place. In discussing institutional influences on executive pay,

DiPrete and Eirich (2010) invoke the Coleman (1986) macro-to-micro and micro-to-macro

distinction to argue that good governance can produce higher not lower levels of executive

reward. In the present context, the argument would be that encouraging boards to take on

significant numbers of non-executives and emphasising that these non-executives should follow

clear and transparent processes when they determine executive remuneration (macro-to-micro)

may result in mimetic decision making that ends up ratcheting executive pay upwards, as each

company strives to keep up with the market. If, for the prisoners dilemma reasoning explained

above, each enterprise is keen not to be caught on the less generous side then the result is an

upward ratcheting of rewards (micro-to-macro). The argument is made above that in such

a situation it may be more effective to make use of the mimetic process to encourage the

adoption of genuinely long terms reward mechanisms such as ‘career shares’, which will at

least do something to ensure that value destroyers are not as generously remunerated as is

currently the case.

26

Table I: Summary Statistics

N p1 p25 p50 p75 p99 Mean St. Dev skew

Career Pay and Performance

TDC Realised (£M) 3157 .1335 .8704 1.72 3.325 21.76 2.983 4.384 5.676TSR 3157 -4.194 -.4348 .2165 .7052 2.238 .004589 1.208 -1.396Δ SW (£M) 3157 -7848 -84.41 50.18 438.2 27358 869.5 7181 2.967

Career Controls

Sales (£M) 3157 2.347 102.7 349.2 1407 34961 2552 9733 11.06Board 3157 5.143 8.5 10.29 12.5 20.3 10.74 3.149 .8205% NEDs 3157 .2388 .4255 .5104 .5889 .7744 .5069 .1186 -.07026Age 3157 33.83 43.5 48.14 53.49 64 48.42 6.843 .1553Tenure 3157 2.034 3.001 4.381 6.434 11.74 5.009 2.425 .9075

Value Creators

TDC Realised (£M) 1915 .1581 1.026 2.104 3.984 22.6 3.55 4.914 5.311TSR 1915 .003667 .2917 .5845 .9974 2.354 .7112 .5479 1.218Δ SW (£M) 1915 .4545 81.5 304 1061 40670 2243 7504 7.382

Value Destroyers

TDC Realised (£M) 1242 .1118 .731 1.31 2.377 18.43 2.11 3.221 6.243TSR 1242 -5.272 -1.523 -.7084 -.263 -.007749 -1.085 1.136 -1.763Δ SW (£M) 1242 -25150 -454.8 -144.7 -52.49 -1.235 -1248 6073 -9.532

Annual Pay and Performance

TDC Realised (£M) 16356 .01785 .1938 .3342 .6129 4.294 .5776 .9636 9.791TSR 16356 -1.87 -.1732 .06154 .2802 1.247 .01121 .537 -1.199Δ SW (£M) 16356 -5259 -43.36 11.83 116.9 6109 131.8 2034 1.719

Annual Controls

Sales (£M) 16356 1.551 93.54 340.1 1418 35974 2624 10168 10.88Board 16356 5 8 10 13 21 10.77 3.563 .7917% NEDs 16356 .2222 .4167 .5 .6 .8182 .5113 .1355 -.07248

1. Sample comprises FTSE350 executive directors serving between 1996 and 2008. The sample excludes careers less than 2 years and those careerscommencing prior to 1st January 1996.2. V alue Creators are directors who’s total shareholder return (TSR) is positive over their career. Career TSR is measured as the difference in thedirector’s company’s logged Datastream return index taken at the start of their career and at the end of their career. The annual calculation of TSRis simply the annual difference in the log of the return index. TSR is multiplied by the average Market Capitalisation over the director’s career to giveΔ Sℎareℎolder Wealtℎ (SW ) (or the market cap at the year end for the annual statistic).3. TDC realised is total compensation realised over the whole career, in Dec 2008 £M. This includes salary, bonuses, perks and the realised values fromshare options, deferred bonuses and vested equity incentives. This is our preferred measure of pay when analysing the efficiency of the remuneration contractover the director’s career.4. The Controls comprise Sales(turnover in Dec 2008 £M), Board (no. of directors at the year end), %NEDs (the percentage of the board comprisingnon-executive directors at the year end), Age (the age of the director at the year end) and Tenure (the directors’ tenure to date measured in years). Forthe career panel, the average value over the directors’ tenure was taken.

27

Table II: Pay-performance Elasticities. Career & AnnualCareer estimates

Full sample Value Creators Value Destroyers Asymmetric Test

TSRc 0.060*** 0.17*** 0.061*** 0.022(6.01) (6.02) (3.88) (1.63)

TSRc+ 0.12***(3.97)

Sales 0.21*** 0.23*** 0.19*** 0.21***(25.1) (19.1) (15.3) (25.3)

Board 0.032*** 0.033*** 0.027*** 0.033***(6.34) (5.33) (3.23) (6.53)

% NEDs 1.03*** 1.15*** 0.90*** 1.05***(9.32) (8.08) (5.35) (9.52)

Age -0.00029 0.00068 -0.0025 -0.00027(-0.16) (0.29) (-0.92) (-0.15)

Tenure 0.046*** 0.039** 0.061*** 0.045***(3.66) (2.33) (3.64) (3.65)

CEO 0.58*** 0.58*** 0.56*** 0.58***(18.5) (13.9) (11.9) (18.5)

Observations 3,157 1,915 1,242 3,157R-squared 0.666 0.681 0.618 0.668F-stat 212 145 68.3 214

Annual Estimates with individual & firm fixed effects

Full sample Value Creators Value Destroyers Asymmetric Test

TSR 0.11*** 0.28*** 0.069*** 0.038**(10.9) (8.18) (2.93) (2.15)

TSR+ 0.17***(4.86)

Sales 0.16*** 0.21*** 0.12*** 0.17***(8.63) (6.24) (5.47) (8.77)

Board -0.0072** -0.0038 -0.0056 -0.0072**(-2.08) (-0.78) (-1.02) (-2.10)

%NEDs 0.41*** 0.40*** 0.51*** 0.40***(5.36) (3.52) (4.34) (5.26)

CEO 0.20*** 0.18*** 0.24*** 0.20***(6.11) (4.05) (4.43) (6.15)

Observations 16,356 9,701 6,655 16,356R-squared 0.428 0.466 0.399 0.427F-stat 217 134 82.4 209

Robust t-statistics in parentheses*** p<0.01, ** p<0.05, * p<0.1

1. The first panel estimates the pay-performance elasticity on a career basis. The dependent variable is the loggedvalue of career TDC realised and TSRc is measured as the difference in the director’s company’s logged return indextaken at the start of their career and at the end of their career. The second panel estimates the pay-performanceelasticity on an annual basis, again using a realised measure of TDC (for comparability the realised measure ispresented here, but qualitatively similar results were obtained using a grant date based measure, results available onrequest).2. The benefit of including a career perspective can be seen by contrasting the career-based elasticities with thosewhich result from a year-on-year approach. Under the latter approach, the estimated pay-performance sensitivity onthe full sample appears as 0.11, versus 0.06 in the career estimates.3. The asymmetric test column tests whether pay is more sensitive to performance when the performance is positive.TSR+ is TSR when TSR>0. This variable is positive, suggesting the relationship between pay and performance isstronger when TSR is positive. Further, when this variable is included with the career estimates, the coefficient onTSRc returns as insignificantly different from zero. This suggests that on a career basis, a director who destroysmore value than her counterpart experiences no financial penalty relative to her counterpart.4. In addition to the controls reported above, industry and time dummies were included in the estimating equations.The industry dummies (and the tenure variable) were omitted in the annual panel as these are eliminated in the fixedeffects regression. Likewise, the age variable merely captures the within director time trend and hence was omittedfrom the annual panel.

28

Table III: Cumulative Pay-performance ElasticitiesFull Sample Value Creators Value Destroyers

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

TSRcum 0.057*** -0.15*** 0.19*** 0.037 0.026 -0.13**(4.83) (-3.51) (8.97) (0.46) (1.46) (-2.26)

%NEDs 1.20*** 1.16*** 1.37*** 1.26*** 0.87*** 1.01***(13.4) (13.1) (12.1) (10.6) (6.22) (6.66)

TSRcum‡ 0.43*** 0.30** 0.33***(4.96) (2.00) (2.93)

Controls

Sales 0.21*** 0.21*** 0.23*** 0.23*** 0.20*** 0.20***(29.1) (29.3) (23.7) (23.8) (17.7) (17.7)

Board 0.017*** 0.016*** 0.015*** 0.015*** 0.012* 0.012*(4.08) (4.00) (3.01) (2.96) (1.81) (1.90)

Age 0.0040** 0.0039** 0.0033 0.0033 0.0030 0.0027(2.38) (2.27) (1.56) (1.58) (1.10) (1.00)

CEO 0.52*** 0.53*** 0.54*** 0.54*** 0.51*** 0.51***(17.3) (17.5) (13.7) (13.7) (11.0) (11.1)

Tenure 0.13*** 0.13*** 0.11*** 0.11*** 0.15*** 0.15***(11.3) (11.3) (7.64) (7.69) (8.66) (8.82)

Observations 16,356 16,356 10,532 10,532 5,824 5,824R-squared 0.685 0.687 0.707 0.707 0.651 0.652F-stat 532 517 411 399 154 151

Robust t-statistics in parentheses*** p<0.01, ** p<0.05, * p<0.1

1 The above table reports OLS estimates using the company-specific pay accumulated to date by theexecutive director as the dependent variable. TSRcum is cumulative TSR, which measures the log differencein the return index starting at the director’s appointment date and ending at each year end until theirexit. TSRcum‡ interacts TSRcum with the average %NEDs to the year end. This allows us to capturethe potential increase in pay-performance sensitivity imposed by boards with a greater proportion of boarddirectors. Best practice introduced during the period has encouraged a steady increase in both non-executivemembership on the board and pay-performance sensitivity (PPS). As such there is a time component tothe relationship between %NEDs and PPS. Nevertheless, the table above explicitly controls for this timecomponent with year dummies (output omitted) in addition to the director tenure variable.

29

Table IV: Exit likelihood and cumulative payFull Sample Value Creators Value Destroyers

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

Pay 0.14*** 0.14*** 0.12*** 0.12*** 0.16*** 0.15***(8.25) (8.21) (5.43) (5.43) (5.53) (5.51)

TSR -0.19*** -0.27*** -0.16*** -0.34** -0.17*** -0.29**(-6.29) (-2.95) (-3.06) (-1.97) (-4.45) (-2.57)

TSRcum -0.052*** -0.053*** 0.052 0.051 -0.085*** -0.086***(-3.02) (-3.05) (1.60) (1.55) (-3.29) (-3.29)

TSR‡ 0.16 0.36 0.25(0.90) (1.13) (1.10)

%NEDs 0.018 0.025 -0.12 -0.16 0.21 0.28(0.17) (0.24) (-0.90) (-1.18) (1.31) (1.61)

Controls

Sales -0.018** -0.018** -0.024** -0.023** 0.010 0.0099(-2.00) (-1.99) (-1.98) (-1.96) (0.75) (0.74)

Board -0.0083* -0.0083* -0.0053 -0.0053 -0.0064 -0.0066(-1.91) (-1.92) (-0.96) (-0.95) (-0.89) (-0.92)

Age -0.021 -0.021 -0.017 -0.017 -0.023 -0.023(-1.24) (-1.25) (-0.83) (-0.83) (-0.79) (-0.80)

Age2 0.00038** 0.00038** 0.00038* 0.00038* 0.00034 0.00035(2.28) (2.28) (1.84) (1.84) (1.20) (1.21)

CEO -0.19*** -0.19*** -0.23*** -0.23*** -0.17*** -0.17***(-6.38) (-6.36) (-5.49) (-5.52) (-3.77) (-3.75)

Tenure 0.027*** 0.027*** 0.024*** 0.024*** 0.035*** 0.035***(3.77) (3.78) (2.58) (2.59) (2.90) (2.90)

Observations 16,197 16,197 10,414 10,414 5,783 5,783Pseudo Log-likelihood -6261 -6261 -3792 -3792 -2452 -2452Wald 719 718 529 529 402 402

Robust z-statistics in parentheses*** p<0.01, ** p<0.05, * p<0.1

1 The above tables reports probit coefficients (not marginal effects) where Pay is the log of cumulativeTDC realised. This captures the emerging connection between pay and the likelihood of exit. Directorswho have accumulated more during their tenure to date, are more likely to exit. The difference is greaterfor the value destroyers but the difference is not statistically significant (Wald test post biprobit in Stata11 (�2=0.67). The equivalent coefficient for column 1 if using log annual TDC realised is .066***2 TSRcum is cumulative TSR. The expected inverse relationship between exit likelihood and performance isfound and this is stronger for the value destroyers. Here the difference is statistically significant (�2=52.92).The TSR‡ interacts TSR with %NEDs and is not significant suggesting more non-executives do not tightenexit-performance relationship, rejecting H5.

30

Figure 1: Distribution of Career Pay and Performance

Banke

r 2

Banke

r 1

HPD

020

4060

80P

erce

nt

0 10 20 30 40 50 60 70 80

£M

Career Realised Pay

Banke

r 1HPD

Banke

r 205

1015

20P

erce

nt

−8 −4 0 4 8

Career TSR

Banker 1

Banker 2HPD

020

4060

80P

erce

nt

−50 −25 0 25£BN

Delta Shareholder Wealth

1. Distribution of CEO pay and shareholder returns from a career based perspective.2. Excludes careers commencing prior to 1st January 19963. Careers less than 2 years dropped4. HPD signifies the Highest Paid Director observed in our sample

31

Figure 2: Heads I win, tails you lose Value creators vs Value Destroyers

−2

−1

0

1

2

3

£M

1 3 5 7 9Tenure (years)

Value Creators Value Destroyers

Median Realised Pay

−2

−1

0

1

2

3

£100

M

1 3 5 7 9Tenure (years)

Value Creators Value Destroyers

Median delta Shareholder Wealth

1. Δ Sℎareℎolder Wealtℎ (SW ) is TSR is multiplied by the average Market Capitalisation over thedirector’s career2. Excludes careers commencing prior to 1st January 19963. Careers less than 2 years dropped

32

Figure 3: Pay-performance sensitivity and percentage of non-executive directors on the board.

−.1

0.1

.2.3

PP

S

.3 .4 .5 .6 .7% Non−executive directors

All directors Value CreatorsValue Destroyers

1. %NEDs is the percentage of non-executive directors that served on the board during the financial year,excluding the chairman. Best practice introduced during the period has encouraged a steady increase inboth non-executive membership on the board and pay-performance sensitivity (PPS). As such there is atime component to the relationship between %NEDs and PPS. Nevertheless, the graph above explicitlycontrols for this time component with year dummies and a variable capturing the length of the directorstenure (see table III )2. Excludes careers commencing prior to 1st January 19963. Careers less than 2 years dropped

33

References

ABI (1987). Share option and profit sharing incentive schemes. summary of revision to guide-lines to requirements of insurance offices as investors. Technical report, Association ofBritish Insurers.

ABI (1999). Responsible voting - a joint ABI - NAPF statement. Technical report, Associationof British Insurers and National Association of Pension Funds.

ABI (2011). Principles of remuneration. Technical report, Association of British Insurers.

Aggarwal, R. K. and A. A. Samwick (1999). The other side of the trade-off: the impact ofrisk on executive compensation. Journal of Political Economy 107 (1), 65–105.

Akerlof, G. A. (1970). The market for ‘lemons’: Quality uncertainty and the market mecha-nism. Quarterly Journal of Economics 84, 488–500.

Alchian, A. A. and H. Demsetz (1972). Production, information costs, and economic organi-zation. The American Economic Review 62, 777–795.

Barreto, I. and C. Baden-Fuller (2006). To conform or to perform? mimetic behaviour,legitimacy-based groups and performance consequences. Journal of Management Stud-ies 43 (7), 1559–1581.

Bebchuk, L. A. and J. Fried (2004). Pay without Performance: The Unfulfilled Promise ofExecutive Compensation. Harvard University Press.

Bebchuk, L. A., J. Fried, and D. Walker (2002). Managerial power and executive compensa-tion. University of Chicago Law Review 69, 751–785.

Bebchuk, L. A. and J. M. Fried (2010). Paying for long-term performance. University ofPennsylvania Law Review 158, 1915–1959.

Bebchuk, L. A. and H. Spamann (2010). Regulating Bankers’ Pay. Georgetown Law Jour-nal 98 (2), 247–287.

Bell, B. and J. van Reenen (2011). Firm performance and wages: Evidence from across thecorporate hierarchy. Centre for Economic Performance Discussion Paper No 1088 .

Bender, R. (2003). How executive directors’ remuneration is determined in two ftse 350utilities. Corporate Governance 11, 206–217.

Bender, R. (2004). Why do companies use performance-related pay for their executive direc-tors? Corporate Governance 12 (4), 521–533.

Bender, R. and L. Moir (2006). Best practice’ in setting executive pay in the UK encourage’good’ behaviour? Corporate Governance 67 (1), 75–91.

Berle, A. A. and G. C. Means (1932). The Modern Corporation and Private Property. WorldInc.

34

Bhagat, S. and R. Romano (2009). Reforming executive compensation: Simplicity, trans-parency and committing to the long-term. Yale Journal on Regulation 26 (39), 359–374.

BIS (2011). Executive remuneration - discussion paper. Technical report, Department forBusiness Innovation & Skills.

Blundell, R., A. Bozio, and G. Laroque (2011, October). Extensive and intensive margins oflabour supply: Working hours in the US, UK and France. Technical report, IZA Institutefor the Study of Labor.

Bruce, A., T. Buck, and B. G. M. Main (2005). Top executive remuneration: a view fromEurope. Journal of Management Studies 42 (7), 1493–1506.

Cadbury, A. (1992). The financial aspects of corporate governance, report on the committeeon the financial aspects of corporate governance. Technical report, GEE, London.

Capezio, A., J. Shields, and M. O’Donnell (2011). Too good to be true: Board structuralindependence as a moderator of CEO pay-for-firm-performance. Journal of ManagementStudies 48 (3), 487–513.

Clarke, R. N. and M. J. Conyon (1998). Corporate governance and director’s remuneration:Views from the top. Business Strategy Review 9 (4), 21–31.

Coleman, J. S. (1986). Are U.S. CEOs paid more than U.K. CEOs? inferences from risk-adjusted pay. American Journal of Sociology 91 (6), 1309–1335.

Conyon, M. and K. Murphy (2000). The prince and the pauper? CEO pay in the UnitedStates and United Kingdom. The Economic Journal 110, 640–671.

Conyon, M. J. (1998). Directors’ pay and turnover: An application to a sample of large UKfirms. Oxford Bulletin of Economics and Statistics 60 (4), 485–507.

Conyon, M. J. (2006). Executive compensation and incentives. Academy of ManagementPerspectives 20 (1), 25–44.

Conyon, M. J., J. E. Core, and R. G. Wayne (2011). Are U.S. CEOs paid more than U.K.CEOs? inferences from risk-adjusted pay. Review of Financial Studies 24, 402–438.

Conyon, M. J., N. Fernandez, M. A. Ferreira, P. Matos, and K. Murphy (2010, May). Theexecutive compensation controversy: A transatlantic analysis. Technical report, FondazioneRodolfo Benedetti, Italy.

Conyon, M. J. and A. Florou (2002). Top executive dismissal, ownership and corporateperformance. Accounting and Business Research 32 (4), 209–225.

Core, J., W. Guay, and R. Thomas (2004). Is us CEO compensation inefficient pay withoutperformance? a review of pay without performance: The unfulfilled promise of executivecompensation by Lucian Bebchuk & Jesse fried. Vanderbilt University & University ofPennsylvania Working Paper .

35

Core, J. E. and W. R. Guay (1999). The use of equity grants to manage optimal equityincentive levels. Journal of Financial Economics 28 (1), 151–184.

Core, J. E. and W. R. Guay (2002). The use of equity grants to manage optimal equityincentive levels. Journal of Financial Economics 28 (2), 151–184.

Core, J. E., W. R. Guay, and D. F. Larcker (2003). Executive equity compensation andincentives: a survey. Economic Policy Review 9 (1), 27–50.

Core, J. E. and D. F. Larcker (2002). Performance consequences of mandatory increases inexecutive stock ownership. Journal of Financial Economics 64, 317–340.

Dabos, G. E. and D. M. Rousseau (2004). Mutuality and reciprocity in psychological contractsof employees and employers. Journal of Applied Psychology 89 (1), 52–72.

Dahya, J., J. J. McConnell, and N. G. Travlos (2002, 02). The cadbury committee, corporateperformance, and top management turnover. Journal of Finance 57 (1), 461–483.

Dalton, D. R., C. M. Daily, S. T. Certo, and R. Roengiptya (2003). Meta-analysis of financialperformance and equity: Fusion or confusion? Academy of Management Annals 46 (1),13–26.

Davis, G. F. (2005). New directions in corporate governance. Annual Review of Sociology 31,143–162.

Dedman, E. (2003). Executive turnover in UK firms: the impact of Cadbury. Accounting andBusiness Research 33 (1), 33–50.

Demsetz, H. and K. Lehn (1985). The structure of corporate ownership: Causes and conse-quences. The Journal of Political Economy 93 (6), 1155–1177.

Devers, C. E., A. A. Cannella, G. P. Reilly, and M. E. Yoder (2007). Executive compensation:A multidisciplinary review of recent developments. Current Science 33 (6), 1016–1072.

DiMaggio, P. J. and W. W. Powell (1983). The iron cage revisited: institutional isomorphismand collective rationality in organizational fields. American Sociological Review 48 (1), 147–160.

DiPrete, T. A. and G. M. Eirich (2010). Compensation benchmarking, leapfrogs, and thesurge in executive pay. American Journal of Sociology 1153.

DTI (2002). Statutory Instrument 2002 No. 1986: Directors’ Remuneration Report Regula-tions. Department of Trade and Industry.

Edmans, A. and X. Gabaix (2009). Is CEO pay really inefficient? a survey of new optimalcontracting theories. European Financial Management 15 (3), 486–496.

Eisenhardt, K. M. (1988). Agency- and institutional- theory explanations: The case of retailsales compensation. Academy of Management Journal 31, 488–511.

36

Ezzamel, M. and R. Watson (2002). Pay Comparability Across and Within UK Boards: AnEmpirical Analysis of the Cash Pay Awards to CEOs and Other Board Members, pp. 489–514. Edward Elgar Publishing Ltd.

Fama, E. F. and M. C. Jensen (1983). Separation of ownership and control. Journal of Law& Economics 26, 301–25.

Fehr, E., S. Gaechter, and G. Kirchsteiger (1997). Reciprocity as a contract enforcementdevice. Econometrica 65 (4), 833–860.

FRC (2010). The UK corporate governance code. Technical report, Financial ReportingCouncil, London.

Frydman, C. and D. Jenter (2010, December). CEO compensation. NBER Working Papers16585, National Bureau of Economic Research, Inc.

FSA (2010, December). Revising the remuneration code. feedback on cp09/10 and final rules.Technical report, Financial Services Authority.

FSA (2011, December). The failure of the Royal Bank of Scotland: Financial Services Au-thority Board report. Technical report, Financial Services Authority.

Gabaix, X. and A. Landier (2008). Why has CEO pay increased so much? Quarterly Journalof Economics 123 (1), 49–100.

Garen, J. E. (1994). Executive compensation and principal-agent theory. The Journal ofPolitical Economy 102 (6), 1175–1199.

Gomez-Mejia, L. R., R. M. Wiseman, and B. D. Johnson (2005). Agency problems in diversecontexts: A global perspective. Journal of Management Studies 42 (7), 1507–1517.

Gong, J. J. (2010, July). Examining shareholder value creation over CEO tenure: A newapproach to testing effectiveness of executive compensation. Technical report, College ofBusiness, University of Illinois at Urbana-Champaign, Illinois.

Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness.American Journal of Sociology 91 (3), 481–510.

Greenbury, R. (1995). Directors remuneration: Report of a study group chaired by sir richardgreenbury. Technical report, Confederation of British Industry, London.

Gregg, P., S. Jewell, and I. Tonks (2012). Executive pay and performance: Did bankers’bonuses cause the crisis? International Review of Finance 12 (1).

Gregory-Smith, I. (2011). Chief executive pay and remuneration committee independence.Oxford Bulletin of Economics and Statistics. Article published online: 24 Aug 2011.

Gregory-Smith, I. and B. G. M. Main (2011, September). The executive labour market atwork. Technical report, University of Edinburgh Business School, Edinburgh.

37

Gregory-Smith, I., S. Thompson, and P. Wright (2009). Fired or retired? a competing risksanalysis of chief executive turnover. The Economic Journal 119 (536), 463–481.

Grossman, S. and O. Hart (1983). An analysis of the principal agent problem. Economet-rica 51, 7–45.

Guest, P. M. (2010). Board structure and executive pay: evidence from the UK. CambridgeJournal of Economics 34 (6), 1075–1096.

Hall, B. and J. Liebman (1998). Are CEOs really paid like bureaucrats? The QuarterlyJournal of Economics 63 (3), 653–691.

Hall, B. and K. Murphy (2002). Stock options for undiversified executives. Journal of Ac-counting and Economics 33, 847–869.

Hall, B. J. and K. J. Murphy (2003). The trouble with stock options. Journal of EconomicPerspectives 17, 49–70.

Hampel, S. R. (1998). Committee on corporate governance. Technical report, GEE, London.

Haubrich, J. G. (1994). Risk aversion, performance pay, and the principal-agent problem.Journal of Political Economy 102 (2), 258–276.

Heckman, J. J. (1993). What has been learned about labor supply in the past twenty years?Federal Reserve Bank of New York Economic Policy Review 83 (2), 116–121.

Hermalin, B. E. and M. S. Weisbach (1998). Endogenously chosen boards of directors andtheir monitoring of the CEO. The American Economic Review 88 (1), 96–118.

Hermalin, B. E. and M. S. Weisbach (2003). Boards of directors as an engdogenously deter-mined institution: A survey of the economic literature. Federal Reserve Bank of New YorkEconomic Policy Review 9, 7–26.

Higgs, D. (2003). Review of the role and effectiveness of non-executive directors. Technicalreport, Department of Trade and Industry, London.

High Pay Commission (2011). Cheques with balances: Why tackling high pay is in thenational interest. final report of the high pay commission. Technical report, The High PayCommission, London.

Huson, M. R., R. Parrino, and L. T. Starks (2001). Internal monitoring mechanisms and CEOturnover: A long-term perspective. The Journal of Finance 56 (6), 2265–2297.

Hutton, W. (2010, Dec). Interim report of the fair pay review. Treasury report, HM Treasury.

Jensen, M. and K. Murphy (1990). Performance pay and top-management incentives. Journalof Political Economy 98 (2), 225–263.

Jensen, M. C. and W. H. Meckling (1976). Theory of the firm: Managerial behavior, agency

38

costs and ownership structure. Journal of Financial Economics 3 (4), 305–360.

Lambert, R., D. F. Larcker, and R. E. Verrecchia (1991). Portfolio considerations in valuingexecutive compensation. Journal of Accounting Research 29 (1), 129–149.

Lincoln, D., D. Young, T. Wilson, and P. Whiteley (2006). The role of the board remunerationcommittee. how remcos function in determining top executive pay. Research report, Parc:Performance and Reward Centre.

Main, B. G. M. (2006). The abi guidelines for share-option based incentive schemes. settingthe hurdle too high? Accounting and Business Research 36 (3), 191–205.

Main, B. G. M. (2011). Executive pay - a career perspective. Edinburgh: The David HumeInstitute.

Main, B. G. M., R. Belfield, and K. Turner (2011). Is there a negotiation process in UKremuneration committees? Technical report, University of Edinburgh Business School.

Main, B. G. M., C. Jackson, J. Pymn, and V. Wright (2008). The remuneration committee andstrategic human resource management. Corporate Governance: An International Review 16,225–238.

Main, B. G. M. and J. Johnston (1993). Remuneration committees and corporate governance.Accounting and Business Research 23 (91A), 351–362.

Main, B. G. M., C. A. O’Reilly, and J. Wade (1995). The CEO, the board of directors andexecutive compensation: Economic and psychological perspectives. Industrial and CorporateChange 4 (2), 293–332.

Main, B. G. M., R. Thiess, and V. Wright (2011). Career shares as long term incentives.Journal of General Management 36 (4), 37–51.

Meyer, J. and B. Rowan (1977). Institutionalized organizations: Formal structure as mythand ceremony. American Journal of Sociology 83 (2), 340–363.

Murphy, K. (1999). Handbook of Labor Economics, Volume 3, Chapter Executive Compensa-tion, pp. 1–89. North Holland.

Murphy, K. J. (1985). Corporate performance and managerial remuneration: an empiricalanalysis. Journal of Accounting and Economics 7, 11–42.

Murphy, K. J. (2002). Explaining executive compensation: Managerial power vs. the perceivedcost of stock options. University of Chicago Law Review 69, 847–869.

Ogden, S. and R. Watson (2008). Executive pay and the search for legitimacy: an investigationinto how UK remuneration committees use corporate performance comparisons in long termincentive pay decisions. Human Relations 61, 711–739.

O’Reilly, C. A., B. G. Main, and G. S. Crystal (1988). CEO compensation as tournament and

39

social comparison: A tale of two theories. Administrative Science Quarterly 33 (2), 257.

Pepper, S. (2006). Senior executive reward: key models and practices. Gower Publishing Ltd.

Perkins, S. and C. Hendry (2005). Ordering top pay: interpreting the signals. Journal ofManagement Studies 42, 1443–1468.

Pfeffer, J. (1981). Power in Organisations. Pitman.

Porac, J. F., J. B. Wade, and T. G. Pollock (1999). Industry categories and the politics of thecomparable firm in CEO compensation. Administrative Science Quarterly 44 (1), 112–144.

Roberts, J. (2001). Trust and control in anglo-american systems of corporate governance:The individualizing and socializing effects of processes of accountability. Human Rela-tions 54 (12), 1547–1572.

Roberts, J., T. McNulty, and P. Stiles (2005). Beyond agency conceptions of the work ofthe non-executive director: Creating accountability in the boardroom. British Journal ofManagement 16 (1), 5–26.

Rumelt, R. (2011). Good Strategy, Bad Strategy. The difference and why it matters. NewYork: Crown Business.

Salancik, G. R. and J. Pfeffer (1977). Who gets power – and how they hold on to it. Organi-zational Dynamics 5, 3–21.

Sanders, W. G. (2001). Incentive alignment, CEO pay level, and firm performance: A case ofheads I win, tails you lose? Human Resource Management 40 (2), 159–170.

Scott, R. W. (1991). Unpacking institutional arrangements, Chapter The New Institutionalismin Organizational Analysis, pp. 164–182. Chicago: University of Chicago Press.

Scott, R. W. (2001). Institutions and Organizations. Thousand Oaks, California: Sage Pub-lications.

Simon, H. A. (1947). Administrative Behavior. Macmillan.

Smith, A. (1776). An Inquiry into the Nature and Causes of Wealth of Nations. RandomHouse.

Stathopoulous, K., S. Espenlaub, and M. Walker (2005). The compensation of UK executivedirectors: Lots of carrots but are there any sticks. Competition & Change 9 (1), 89–105.

Tarbert, H., K. Tee, and R. Watson (2008). The legitimacy of pay and performance com-parisons: An analysis of UK university vice chancellors pay awards. British Journal ofIndustrial Relations 46, 771–805.

Tosi, H., S. Werner, J. P. Katz, and L. R. Gomez-Mejia (2000). How much does performancematter? a meta-analysis of CEO pay studies. Journal of Management 26 (2), 301–339.

40

van Essen, M., J. Otten, and E. J. Carberry (2012). Assessing managerial power theory: Ameta-analytic approach to understanding the determinants of CEO compensation. Journalof Management 38 (3), forthcomimg.

Wade, J., C. O’Reilly, and T. Pollock (2006). Overpaid CEOs and underpaid managers:fairness and executive compensation. Organization Science 17 (5), 527–544.

Walker, D. (2009). A review of corporate governance in UK banks and other financial industryentities. Technical report, HM Treasury.

Weisbach, M. S. (1988). Outside directors and CEO turnover. Journal of Financial Eco-nomics 20, 431–460.

Westphal, J. D. (1999). Collaboration in the boardroom: behavioural and performance con-sequences of CEO-board social ties. Academy of Management Journal 42, 7–24.

Westphal, J. D. and E. J. Zajac (1995). Accounting for the explanations of CEO compensation:Substance and symbolism. Administrative Science Quarterly 40, 283–308.

Williamson, O. E., M. L. Wachter, and J. E. Harris (1975). Understanding the employmentrelation: the analysis of idiosyncratic exchange. Bell Journal of Economics 6 (1), 250–278.

Wiseman, R. M., G. Cuevas-Rodriguez, and L. R. Gomez-Mejia (2012). Towards a socialtheory of agency. Journal of Management Studies 49 (1), 202–222.

Yermack, D. (1996). Higher market valuation of companies with a small board of directors.Journal of Financial Economics 40 (2), 185–211.

Yermack, D. (2004). Remuneration, retention, and reputation incentives for outside directors.Journal of Financial Economics 59 (5), 2281–2307.

Zajac, E. J. and J. D. Westphal (1995). Accounting for the explanations of CEO compensation:substance and symbolism. Administrative Science Quarterly 40 (2), 283–308.

41