U.K. executive compensation practices: new economy versus old economy.
Keywords: managerial compensation; remuneration committee; executive stock options; new economy.
JEL Classification: G30, G34, J33.
In comparison with the huge body of literature on compensation practices of U.S. companies, the corresponding literature in the U.K. is still relatively sparse. This is particularly true for studies using a wider definition of remuneration to include long-term incentive plans and executive stock-option schemes.
The aim of this paper is to examine the remuneration practices of U.K.-listed companies, focusing mainly on directors' pay. Specifically, we investigate whether there are systematic differences between companies in the so-called New Economy (consisting of Internet, e-commerce, or dot-com firms) and those in the traditional Old Economy. In order to minimize the distorting impacts of comparing companies across disparate industries, we restrict our analysis to a sample of New and Old Economy companies involved in the same ultimate "business," namely retailing. Specifically, we compare and contrast the remuneration practices of traditional Old Economy retailers with those of their New Economy (e-commerce and dot-com) competitors.
The emergence of the New Economy is a recent phenomenon, and the majority of the companies in this sector were only floated in the second half of the 1990s, particularly during the latter years of the Internet bubble that burst in Spring 2000 (see, e.g., Ljungqvist and Willhelm  for statistics on Internet IPOs in the U.S.). Since remuneration practices are likely to differ systematically between recently listed firms and their more seasoned counterparts (e.g., Beatty and Zajac 1994), we compare New Economy companies not only with seasoned, established Old Economy firms, but also with a sample of traditional Old Economy retailers that were only recently listed on the stock exchange; we term the latter "New-Old Economy" companies. Specifically, New-Old Economy companies are defined as companies that use traditional methods of commerce (traditional retailing), and were floated on the London Stock Exchange during the three years prior to the remuneration report examined in our study (i.e., between 1996-99).
Unlike most previous U.K. studies, we extend the analysis beyond the level of the CEO, by including data on other executives, and examine whether stock option plans are available to lower-level employees besides top executives. Existing U.K. and U.S. evidence suggests that the use of option plans is more broad-based in New Economy companies (Anderson et al. 2000; Ittner et al. 2003; Murphy 2003) and in companies with flatter organizational hierarchies (Conyon et al. 2001).
We document the structure of remuneration packages in U.K. companies, including the relative prevalence and values of short-term cash compensation and of long-term, equity-based compensation. In contrast with most previous studies our analysis includes both CEOs and also other directors. We also examine the determinants of the level and structure of executive pay in terms of the characteristics of the firm, including membership of our three subsamples (or economy sectors), size, growth/ growth opportunities, financial policy, ownership structure, and governance arrangements.
Recent results for the U.S., presented by Ittner et al. (2003) and Murphy (2003), suggest that equity-based pay is more prevalent among New Economy companies. It is interesting to examine whether this finding holds not only for the U.S., but also in the U.K. There are many similarities between the U.K. and U.S. in terms of their institutional framework and the functioning of their capital and managerial labor markets. As argued by Conyon and Murphy (2000), one might as a result reasonably expect similarities in executive-pay practices between the two countries. However, Conyon and Murphy (2000) find substantial differences in pay practices, particularly in terms of a significantly greater use of stock options in the U.S. Their study provides a breakdown of pay by size and industry, but they do not give comparative figures for New and Old Economy companies, or for newly floated versus established companies. Thus, the question arises whether New Economy firms in the U.K. behave more like their established counterparts in the U.K. Old Economy, or more like other New Economy firms based in the U.S. (or elsewhere). Our study is the first to address this important issue.
We attempt to model the variation across executives in the characteristics of their executive stock options, specifically the moneyness of their options, (1) time to maturity, pay-performance sensitivity, the ratio of long-term to short-term pay, and the performance sensitivity of long-term pay relative to short-term pay, expressed as the "leverage" of the pay package.
Existing research in the U.S. has found that stock options are almost invariably granted at-the-money, that is, the exercise price is set (roughly) equal to the stock price at the grant date. For instance, Hall and Liebman (1998) report that 94 percent of the executive share options in the U.S. were granted at-the-money, and Murphy (1999) estimates the proportion at over 95 percent. By contrast, we find that in our U.K. retailing sample a surprisingly large number of companies grant executive stock options either out-of-the-money or even, contrary to U.K. corporate-governance best practice guidelines, in-the-money. To our knowledge, this is the first U.K. study to document a significant difference between the exercise price of executive stock options and the share price at the grant date. We seek to explain the grant-date moneyness of executive stock options in terms of a range of firm, director, and governance characteristics, including the economic sector of the firms, board and ownership structure, and the composition of the remuneration committee.
For each compensation feature that we model, we consider the influence of the fact that the firm is a New Economy firm and/or that it has recently experienced an IPO. Thus, we use the control sample of New-Old Economy firms to identify the special features of the New Economy group that cannot be explained by the fact that they are relatively new to the stock market. The present study contributes to the small, existing body of work on U.K. executive pay including stock options and long-term incentive plans (LTIPs). We do so by comparing the compensation practices of relatively established Old Economy companies, which typically make up the samples of previous studies (e.g., Main et al. 1996; Conyon et al. 2000; Conyon and Sadler 2001), with those of a new breed of company in the so-called New Economy.
The next section provides an overview of the U.K. institutional framework. In the third section, we develop and outline our research questions and describe the data and methodology used in this study. The fourth section presents the results of our analysis and the final section concludes the paper with a brief summary and discussion.
In recent years, a number of external factors have led to greater interest in compensation issues. Globalization has been one of these factors as in a supposedly global market for executives, firms find it necessary to benchmark their compensation against international companies. As Conyon and Murphy (2000) indicate, the generally higher compensation of U.S. CEOs has put pressure on CEO compensation in the U.K., Europe, and elsewhere. A further external factor is the growing influence of the shareholder value movement, which has led to a greater focus on performance-related compensation mechanisms, such as executive stock options. More recently, pressure on institutional investors such as unit trusts and insurance companies has increased the focus on financial performance and on executive compensation packages. In the U.K., the focus on corporate governance, including executive remuneration, over the past decade has given rise to a number of reports and codes of best practice. These have resulted in a dramatic increase in the level of disclosure of executive pay data. (2)
The recommendations of the Cadbury Committee (1992) were the first attempt to produce guidelines on "good practice" on important corporate governance issues. These include the responsibilities of executive and nonexecutive directors, the clarity and form of the information companies provide, the responsibilities of the auditors, and the links between shareholders, boards, and auditors. It did not, however, provide detailed guidance on the structure of the compensation packages. Instead it suggested this subject be put on the agenda of a subsequent committee, which eventually became known as the Greenbury Committee.
After the publication of the report of the Greenbury Committee (1995) (hereafter the Greenbury report), the U.K. government announced the withdrawal of the permitted tax advantages for new issues under the approved executive share option schemes. In addition, the November 1995 Budget introduced a new type of option scheme with a very low upper limit of only 20,000 [pounds sterling] on individual option holdings. Also in response to the recommendations of the Greenbury report, the previously acceptable practice of setting the exercise price of executive share options at a 15 percent discount to the current share price at the grant date was ruled out. Overall, the Greenbury report did not favor traditional executive share options schemes, calling not only for the repeal of any remaining tax advantages on certain schemes, but also encouraging companies to replace option schemes with other forms of LTIPs. In the U.K., LTIPs typically involve the award of shares, not cash (unlike in the U.S., where LTIPs can involve either shares or cash bonuses). Also, in the U.K., the award and vesting of shares in LTIPs is typically contingent on the executive achieving certain performance criteria, and the shares awarded must be held for a prescribed minimum period.
The recommendations of the Cadbury and Greenbury Committees were incorporated in the Combined Code of the London Stock Exchange (LSE), which effectively regulates the remuneration practices of LSE-listed companies (Combined Code 2000). All companies that are (or intend to become) listed on the LSE must include in their annual reports a statement about their compliance with the Code, detailing and giving reasons for any noncompliance. The Code has become such a prominent focal point that even companies listed on the Alternative Investment Market (AIM), that do not formally have to comply with the Code, still follow its recommendations in order to increase the attractiveness of their shares to investors.
The Code stipulates a high level of disclosure of information related to executive compensation and obliges all companies to clearly state their remuneration policy in a separate section of their annual reports. In addition it requires a complete description of the remuneration packages of each director, executive and nonexecutive.
The Code requires LSE-listed companies to have a remuneration committee that sets top executive remuneration, and recommends that this committee should consist wholly or mainly of independent nonexecutive directors. In practice, nonexecutives are typically directors of other companies' and members of the legal, accountancy, or consultancy professions. Whether nonexecutives are in fact independent in practice is often a moot point (e.g., Main 1993). Prior to 1995, remuneration committees in British companies frequently included the most senior executive director as a member (Conyon 1994). Nonexecutive directors often owe their positions to the executive directors, and outside members, made up of other business people, have similar (personal) objectives as the company's executives. Ezzamel and Watson (1998) find evidence of a "cozy" relationship between directors and nonexecutives, resulting in a ratcheting up of executive pay.
The Combined Code encourages remuneration committees to seek the advice of independent pay consultants to ensure objectivity in the remuneration policy. In practice, however, these consultants are often appointed on the recommendation of the company directors, which may impair their "independence." In addition, consultants tend to use cross-sectional comparisons in making pay recommendations (Tosi and Gomez-Mejia 1989; Mangel and Singh 1993), which can give rise to a ratcheting-up of pay over time. Comparisons are generally made with companies of a similar size and often in the same industry. This benchmarking tends to result in similar executive pay in companies of similar size, and the link between pay and company performance receives relatively less attention.
The accounting and tax treatment of executive compensation in the U.K. differs from that in the U.S. Of particular relevance to our study is the treatment of executive stock options, particularly as a determinant of the spread between the exercise price and the underlying share price on the grant date (the moneyness of the options). In the U.S., Murphy (1999) notes that neither the company nor the executive are taxed upon the granting of an executive stock option, and it is only at the exercise date that both are taxed, with the precise tax treatment depending on whether the stock options are "qualified" or "nonqualified." However, Bebchuk et al. (2002) point out that the U.S. Internal Revenue Code does not regard in-the-money options as "performance-based compensation," and as Conyon and Murphy (2000) note, the amount of non-performance-based compensation that is deductible from company profits as a business expense is limited to $1 million.
According to Main et al. (1996), an important influence on U.K. companies' use of executive share option schemes in Britain has been the U.K. tax authorities, the Inland Revenue. In the U.K., there are approved and unapproved share option schemes (corresponding to the qualified and unqualified schemes in the U.S.). For both approved and unapproved share option schemes, no tax is normally payable on the grant date assuming the exercise period does not exceed ten years (e.g., RM2 Partnership 2003). For unapproved options, gains are taxed as personal income upon exercise of the options, regardless of whether the shares acquired are subsequently sold. Approved options are not taxed on exercise, but the gains from a later sale of the shares are taxed at the capital gains rate (although there may be some tax exemptions and reliefs). Approved options are normally granted at the money (RM2 Partnership 2003), but because of institutional restrictions, they are rarely granted to top executives (Conyon and Murphy 2000, F665). Currently under approved schemes, each employee can only be granted options on shares up to a total value of 30,000 [pounds sterling] at the date of grant. (3) For the period of this study there was normally no corporation tax deduction for firms issuing shares under a stock option scheme.
In the U.S., the accounting treatment of regular stock options (with a fixed exercise price and expiry date) discourages the granting of options in-the-money as companies incur an accounting charge equal to the spread between the underlying share price and the exercise price on the grant day amortized over the life of the option (Murphy 1999). The accounting treatment of options with unspecified exercise prices and/or variable terms to maturity is less favorable than that of regular options (Murphy 1999; Bebchuk et al. 2002). By contrast, in the U.K., companies, were not until recently required to make a charge in their profit and loss accounts for the cost of options granted to employees. However, this was changed as of January 2004 and the compulsory adoption of the International Financial Reporting Standard, which requires that the fair value of employee stock options granted is recognized as an expense (International Accounting Standards Board 2003). (4)
THEORETICAL FRAMEWORK AND RESEARCH ISSUES
The conventional theoretical framework for understanding reward contracts is agency theory. (See Lambert  for a recent review.) Agency theory focuses on the risk characteristics of the enterprise (including any related information asymmetry issues) as the prime determinants of the shape and nature of reward contracts. The New Economy stocks are the most risky firms in our sample, and the Old Economy stocks are the least risky, and so we expect to see differences between the reward contracts of the New Economy stocks and the other firms in our sample.
While we do attempt to interpret our evidence by reference to agency theory, we view this theory only as a starting point for a more nuanced view that reflects a number of practical factors that are currently beyond our modeling capabilities.
Some practically important economic factors are difficult to model analytically. In particular the liquidity position of the firm, bankruptcy risk, and the risk of the firm being taken over are all factors that figure in the calculus of executives and their employers. We suspect that all of these factors are more likely to play a greater role in the case of our New Economy subsample.
The U.K. market for senior managerial talent attracts the sustained attention of politicians and other corporate stakeholders such as customers and trade unions. The high levels of compensation attached to the senior executive positions of very large corporate enterprises tend, from time to time, to attract attention and controversy. As a result, executive compensation committees may feel constrained in what they can pay executives, and this may spill over into the design of the compensation scheme itself. For example, schemes that have the potential to result in very large bonuses (or capital payouts on stock option realizations) may be limited by fear of public rebuke. Bebchuk et al. (2002) see this "outcry" factor as the most important constraint on management's ability to pay themselves excessive rewards. We hypothesize that such "outcry" factors will carry less force in highly innovative situations where there is an opportunity to argue that the special circumstances of a new kind of enterprise require a more aggressive form of compensation arrangement (the period just before the collapse of Enron seems to be a case in point). Thus we expect the New Economy (and perhaps even the New-Old economy) subsample to be relatively unaffected by this factor.
As explained in the previous section, U.K. executive compensation practices have been heavily influenced by the listing regulations of the London Stock Exchange and by various codes of practice. It is possible that our New Economy and New-Old Economy subsamples may have felt a need to comply closely with these regulations in order to achieve the status of an LSE listing.
Critics of agency theory have focused on the unrealistic nature of the assumptions underlying such models. However, these criticisms have lacked force because of their failure to advance an alternative theory capable of generating empirically interesting and testable hypotheses beyond those yielded by the standard agency model. It is only recently that more forceful critiques have been published that challenge agency theory both on the basis of its assumptions and on the basis of its empirical validity. This new work has established two empirically informed lines of criticism of the agency model.
Bebchuk et al. (2002) attempt to establish a more general view of the determinants of executive compensation that introduces a role for managerial power as an additional factor conditioning compensation arrangements. This new perspective provides a number of insights on how reward contracts are actually set. For example, the authors argue that it is difficult to reconcile the reluctance of firms to link rewards to relative company performance with an agency theoretic view of the world. This reluctance is easier to reconcile by a theory that gives greater recognition to rent-seeking behavior by company executives. We believe this bold hypothesis, with potentially far-reaching implications for public policy, is worthy of further investigation. Further work is needed both on the theoretical refinement of the hypothesis and on the construction of an empirical research design that is capable of identifying the incremental explanatory value of measures of managerial power for observable variations in reward structures. In the context of the present work, the Bebchuk et al. (2002) paper points to the importance of capturing the power relations between company executives and boards of directors. Our empirical work tests for the specific influence of the composition of the compensation committee on the form of the executive reward contract.
Murphy's (2002) Perceived Cost Perspective
In his commentary on Bebchuk et al. (2002), Murphy (2002) postulates a "perceived-cost view" of executive compensation, and presents new empirical evidence in favor of this hypothesis. Murphy (2002) argues that, because of cash flow and accounting considerations, companies routinely (and inappropriately) treat stock options as an inexpensive way to convey compensation; i.e., compensation committees routinely underestimate the true cost of the options they issue to executives. As a recognized authority on empirical research on executive compensation, Murphy's (2002) acknowledgement of evidence of a systematic misperception of the costs of executive stock options deserves special attention. (5)
The aim of this study is to investigate the practices of U.K.-listed companies of granting stock options and LTIPs, and to compare and contrast practices (1) between traditional and New Economy firms and (2) between old or "seasoned" (listed) companies and their more recently listed counterparts. In order to control for the importance of industry factors, we focus on retailers in the New Economy (dot-coms and e-commerce retailers) and Old Economy (traditional retailers). We examine the composition of compensation in terms of short- and long-term compensation and cash and equity-based pay. We model five specific features of executive contracts: the moneyness of the options, the term to maturity of the options, the pay/performance leverage of the contracts, long-term to short-term pay ratio, and the pay/performance sensitivity.
We expect the nature of compensation packages to vary systematically across the three subsamples. However, to some extent, these differences will simply be the effect of other firm characteristics, particularly of firm size, which are known to have a significant positive relation with compensation levels (Murphy 1999). An important research question is whether there is a pure New Economy effect over and above the impact of firm size and sales and other control factors. In addition to measures of firm size, we introduce controls for risk (volatility of stock returns), growth and growth opportunities, and the financial policies of the firm. In the light of the recent governance literature, we test for the influence of ownership structure and the composition of the remuneration committee on the levels of executive pay and the five specific contractual features of interest.
Agency theory leads one to expect the sensitivity of pay to performance to decrease as risk increases (for example, Lambert 2001). We therefore expect factors related to risk to be negatively associated with measures of pay/performance sensitivity. If one also allows for the costs of contracting in a multiperiod context, then agency theory would lead one to expect higher levels of long-term relative to short-term pay in firms that are developing rapidly. For example, the cumulative firm-specific knowledge of the management team may be essential to the firm's ability to exploit and generate future growth opportunities. Thus we expect to see higher relative long-term pay, and longer terms to maturity of stock options, in firms with high-growth rates, high levels of R&D, and low levels of book-to-market value (an inverse proxy of growth opportunities).
We expect indicators of strong governance to lead to higher levels of leverage, pay-performance sensitivity, and pay ratios, and lower base salaries, lower moneyness, and lower maturities. (6) The theories of Bebchuk et al. (2002) and Murphy (2002) point to a search for less obvious forms of wealth transfer. Executives in firms that are weakly governed might seek to exploit the situation by pushing up the moneyness or time to maturities of their stock options. Thus, after controlling for the growth prospects of the firm, we expect to find the clearest evidence of governance effects for these two variables. (7)
DATA AND METHODOLOGY
Until the second half of the 1990s, executive pay research in the U.K. was hampered by poor data quality due to the lack of detailed company disclosure of executive remuneration details. However, since then the level of disclosure of compensation details (and other governance information) in companies' annual report has increased dramatically. According to Conyon and Sadler (2001), the information on executive remuneration disclosed in most U.K. annual reports is now more detailed than under U.S. disclosure rules. Conyon and Murphy (2000) point out that data on prior share option grants are now more detailed in the U.K. than in the U.S., and Conyon and Sadler (2001) use this fact to quantify the biases induced in U.S. executive pay studies due to incomplete information disclosure.
The data used in our analysis is extracted from the Remuneration Committee (RC) Reports in the 1999 Annual Reports and Accounts of our sample of U.K. listed companies. (8) The RC Report typically discloses the salaries and bonuses paid to individual directors, and gives details of any stock options and long-term incentive plans (LTIPs) held by and granted to the directors. Our sample consists of executive compensation data for 549 directors (68 CEOs, 253 other executives, and 228 nonexecutives) of 72 U.K. retailing companies in the traditional and the New Economy sectors.
We identified the industry classification of companies using the 1999 FTSE classification (based on data from Datastream). The firms in our New Economy sample were chosen from the "Computer and Software Services" and "e-commerce Retailers" sectors of the London Stock Exchange (LSE). Using data from Datastream on FTSE industry classifications, we identified 24 companies in these two sectors. All of them are constituents of the FTSE techMARK Index. (9) Correspondingly, we randomly selected 24 firms for the Old Economy sample and 24 for the New-Old Economy sample, both from the "General Retailers" and "Food and Drug Retailers" sectors. We selected New-Old Economy firms that went public between 1996 and 1999, using the "U.K. New Issues" official lists for 1996 to 1999 from the LSE website.
All data relating to current and previous share options and LTIP grants, on numbers of and participants in schemes, and on the composition of the remuneration committee were taken from companies' annual reports. Data on other company characteristics, such as market capitalization, sales, number of employees, intangible and total assets, and return on equity, and on share prices, returns on stocks, and on treasury bills were collected from Datastream.
Following common practice in the recent literature (e.g., Conyon and Murphy 2000; Conyon and Sadler 2001; Murphy 1999), total compensation is defined as the sum of base salary and annual bonus, plus the value of share options and LTIPs. Where an executive has been employed for less than one complete financial year, we annualize the salary and bonus by multiplying the reported figure by the ratio of 12 over the number of months in post. We follow the approach of Conyon and Murphy (2000) who value LTIP share grants at the face value of the shares on the grant date and deduct 20 percent for U.K. LTIPs because the award of the shares is contingent on the director's attainment of performance criteria.
Following Conyon and Sadler (2001) and Conyon and Murphy (2000), we value stock options using the modified Black and Scholes (1973) (hereafter Black-Scholes) option-pricing model (see the Appendix for details). As in Conyon and Murphy (2000), no discount for performance contingency is applied to share-option schemes because the performance criteria on share-option schemes are rarely binding. Further, where they are linked to share price, the intrinsic value (the larger of the share price less the exercise price or zero) will be low in the event that the criteria are actually binding.
Murphy (1999) notes that the Black-Scholes formula measures the value of a standard, tradable European option to a well-diversified investor. By contrast, executive stock options are not tradable and are usually forfeited if the director leaves the company. In addition, risk-averse directors, who are prohibited from short-selling the shares of their companies and have all their human capital tied up with the company, are less able than ordinary, outside investors to hedge away the risk of the options (Murphy 1999). Further, the Black-Scholes formula ignores the possibility of early exercise, which may have either a positive or negative effect on the value of the options, and which, in turn, may depend on directors' private information. For a more detailed discussion of these difficulties and of the likely discrepancy between the cost of options to the company and their value to the director, see Murphy (1999). Finally, as pointed out by Conyon and Murphy (2000), the value of options to both the company and its directors will also be affected by the presence of performance criteria that determine whether the share options will vest or not. Conyon and Murphy (2000, F645) observe that "share options granted in the United Kingdom typically vest only upon attainment of some performance criteria, often based on earnings-per-share growth."
Conyon and Sadler (2001) examine the likely biases induced by the need to estimate parameters in the Black-Scholes formula. Contrasting U.S. disclosure with that in the U.K., they conclude that, based on the level of disclosure of compensation information in U.K. annual reports since the financial year 1997/98, there "exists sufficient information in U.K. company annual reports to analyze the design of British CEO compensation contracts" (Conyon and Sadler 2001,252).
Consistent with Conyon and Sadler's (2001) assessment, we find that the data in U.K. annual reports allows us to observe or infer information on the Black-Scholes parameters for CEOs and all directors. Data for the expiry date and the exercise price of the options is from the annual reports of the companies. The time to maturity (or expiration) was measured as the difference in calendar years between the expiry date of the option and December 31, 1999. The spot price is the closing share price on the day of the issuance of the option unadjusted for any subsequent capitalization changes and is taken from Datastream (datatype UP). The risk-free interest rate was measured as the 1999 average annual yield of the seven-year U.K. government bonds, and the dividend yield was computed as the average of the prior 48 monthly observations on cash dividend yield. (l0) Data for both the government bond and dividend yield are from Datastream. Volatility was estimated using the standard deviation of the monthly continuous compounded returns over the prior 48 months, multiplied by the square root of 12 to yield an annualized figure.
To assess the relative importance of equity (or long-term) incentives relative to cash (or short-term) incentives, we estimate two measures. First, the "pay ratio" is the ratio of the value of long-term pay (LTIPs and options) over short-term pay (salary and bonus). (11) Second, "leverage" measures the relative sensitivity of executives' wealth (or rather, their income from the employment relation) as a result of changes in the firm's equity value, as a fraction of their cash compensation. Leverage L is estimated as the pay-performance sensitivity of the executive's share options and LTIPs (similar to the measure employed by Core and Guay , but ignoring the executive's shareholding in the firm) divided by the executive's cash compensation (base salary + annual cash bonus):
(1) L = [n.sub.1] x [d.sub.O] x 0.01p + [n.sub.2] x [d.sub.LTIP] x 0.01p/s + b
where [n.sub.1] is the number of options held by the director, [n.sub.2] is the number of LTIP shares held by the director, [d.sub.O] is the delta (12) of the stock option, [d.sub.LTIP] is the delta of the long-term incentive plan, p is the company's year-end stock price, s is the base salary of the director, and b is the annual cash bonus of the director. The numerator indicates the total change in the values of the options and LTIP shares that the director holds, for a 1 percent change in the stock price (Core and Guay 1999, 154-155). If the director holds a portfolio of options or LTIP shares, then we take the sum of the deltas of each component of the portfolio. For simplicity we assume, as in Conyon and Murphy (2000, 657), that [d.sub.LTIP] is equal to 1. (13) Leverage of 1 (L = 1) indicates that a 1 percent change in the stock price (0.01*p) would increase the director's equity-based pay (stock options and LTIPs) by an amount equal to his cash compensation (base salary plus annual bonus).
To examine whether there are systematic differences in remuneration practices among Old, New-Old, and New Economy firms, after controlling for known determinants of remuneration such as firm size and sales, we run OLS regressions with remuneration as dependent variable and two sector dummies, firm assets and sales, and measures of other firm characteristics (including governance proxies) as the independent variables. The regressions are estimated separately by director type (CEOs, executive, and nonexecutive directors) and by type of compensation (salary plus bonus, equity-based "long-term" compensation, and total pay; all expressed in natural logarithms):
(2) ln(salary + bonus) = [a.sub.0] + [a.sub.1]ln(MV) + [a.sub.2]ln(Sales) + [a.sub.3][New.sub.dummy] + [a.sub.4] [New-Old.sub.dummy] + a'X + [epsilon];
(3) ln(long-term) = [b.sub.0] + [b.sub.1]ln(MV) + [b.sub.2]ln(Sales) + [b.sub.3] [New.sub.dummy] + [b.sub.4] [New-Old.sub.dummy] + b'X + [epsilon];
(4) ln(total pay) = [c.sub.0] + [c.sub.1]ln(MV) + [c.sub.2]ln(Sales) + [c.sub.3] [New.sub.dummy] + [c.sub.4] [New-Old.sub.dummy] + c'X + [epsilon];
where MV is the average market capitalization for the examined year, Sales is the total sales for the year, New and New-Old are sector dummies, and X is a vector of further explanatory and control variables comprising indicators of firm, director, ownership, and corporate-governance characteristics.
The dependent variable long-term is the value of the long-term pay (i.e., the sum of the values of stock options and LTIPs) and total pay is the sum of base salary, annual bonus, stock options, and LTIPs. The estimated coefficients of ln(MV) and In(Sales) can be interpreted as the elasticities of total pay (or of its separate components) with respect to size and revenues.
Since nonexecutive directors are not normally awarded stock options or LTIPs (see Table 2), the regressions for long-term and total pay are not estimated for nonexecutives. In addition to the separate regressions by director type, we estimate three further regression models (for short- and long-term compensation and for total compensation) using pooled data for all directors and interactive sector/CEO dummies.
Finally, we construct regression models for the determinants of moneyness (measured as the ratio of the share price on the grant day of the options over the stated exercise price), time to maturity (measured as the difference between the expiry date of the option and December 31, 1999; all figures are expressed in years), pay-to-performance leverage, the ratio of long-term to short-term pay, and pay-performance sensitivity (PPS).
In the regression analysis of moneyness as the dependent variable, we first estimate and report a model that includes firm-size measures (sales and assets) and other firm characteristics, but excludes all governance or ownership variables. Given that sales are a major driver of firm value in the retailing sector, we believe it is important to control for differences in sales relative to total assets. This is achieved by including both sales and assets as control variables in all the regressions. In this and all other specifications, the size measures are both interacted with the New and New-Old Economy dummy variables in order to estimate the marginal effects of firm size in the two sectors. We then introduce our governance and ownership variables without sector interaction. Finally, we allow the regression coefficients to vary across the three sectors for each (set of) governance/ownership variable(s) in a separate regression. Instead of including interaction terms for all variables at the same time, we adopted this sequential method to avoid potential multicollinearity among the interaction terms of different variables.
We employ the same modeling strategy in analyzing the other four contract features. We first report the results of the basic model excluding ownership and governance characteristics. We then report the results of adding in the governance and ownership characteristics without sector interaction. Finally, we report, in a separate panel, the regression parameters and significance levels associated with each individual governance and ownership variable when, one by one, the parameters associated with the variable are allowed to vary across sectors. (14)
Table 1 shows descriptive statistics for our sample of firms by economic sector at the end of 1999. Firms in the Old Economy sample are on average substantially (and statistically significantly) larger in terms of book value of assets, sales, and numbers of employees. Also, the average market capitalization of the Old Economy sample is larger than that of the New-Old sample. However, shortly before the bursting of the Internet and dot-com bubble in Spring 2000, the differences in the average market capitalizations the Old Economy sample and the New Economy companies were not statistically significant.
The New Economy sample is characterized by significantly higher rates of employee and sales growth, a higher ratio of R&D expenditure to sales, a lower ratio of book-to-market value, lower dividends per share, and a higher level of volatility (measured by Datastream as the standard deviation of the weekly share price during 1999). These findings correspond closely to the differences between New and Old Economy firms in the U.S. reported by Ittner et al. (2003).
Compensation Practices in the New and Old Economy
Share option schemes are widely in use in our sample companies. Table 2 gives detailed figures on the four main types of compensation (salaries, bonuses, share options, and LTIPs) for executive and nonexecutive directors. Panel A shows the numbers and proportion of directors in our sample receiving a particular compensation scheme, broken down by both economic sector (i.e., subsample) and by director type (CEO, other executives, and nonexecutives).
Focusing first on CEO compensation, Table 2 shows that options schemes for the CEO are most common in the Old Economy sample. This is perhaps surprising, but it should be noted that the firms are, on average, largest in the Old Economy, and previous results document that share option schemes are more common in large companies (e.g., Conyon and Murphy 2000, Table 2). This pattern is even more pronounced in the case of LTIPs, which are awarded to only one and two CEOs in the New and New-Old samples, respectively, but to six of the Old Economy CEOs.
These results, however, do not extend to the other executives below the CEO. Instead, for other executives, the New Economy directors are more likely to receive share options. The fact that the number (as opposed to the percentage) of other executives in New Economy firms receiving share options is lower than the corresponding number in the Old Economy sample is due to the far lower overall number of executives in the smaller New and New-Old Economy sample companies (see Column 1 in Panel A of Table 3). Examining the evidence on LTIPs, there is relatively little difference between the subsamples in terms of the proportions of other executives receiving such compensation.
An interesting observation is the existence of share options granted to nonexecutive directors in the New Economy firms (and to a very small extent in recently listed traditional, New-Old Economy firms). By contrast, nonexecutives are paid purely in the form of a fee (shown in the base salary column) in the Old Economy companies. In all sample companies, nonexecutives receive neither bonuses nor LTIPs.
Panel B of Table 2 shows the average amounts (in thousands of pound sterling) paid to the directors of the sample companies in the form of salary, bonus, share options, and LTIPs. The share option value is measured as the option value on the date of grant calculated using the Black-Scholes formula (see the Appendix). The LTIPs were valued at 80 percent of face value for performance-contingent awards. The figures reveal that while the proportion of CEOs receiving stock options is lower in the New than in the Old Economy sample (Table 3), the average value of the options received by each New Economy CEO eligible to receive stock options is double that of their Old Economy counterparts (although the difference in means for the CEOs is not statistically significant). (15) By comparison, the average option value for the CEOs of the New-Old Economy sample is much smaller than that for either of the other subsamples. The ranking of the subsamples is the same in terms of the medians of CEOs' share option values, although the difference between the New and Old Economy sample is not as pronounced. For the other executives (besides the CEO), both the proportion of directors that are awarded stock options and the average and median values of these options are largest for the New Economy retailers. The high average and median values of the option portfolios of CEOs and other executives of New Economy companies reflect the comparatively high market valuations of New Economy companies stocks during the sample period. The difference in the values of the option portfolios of CEOs versus other executives is least pronounced for the New-Old Economy firms and most pronounced for their seasoned Old Economy counterparts.
We next counted the number of stock-option schemes in operation in our sample firms. The firms in the Old Economy sample have the greatest average number of schemes, 3.39, as compared to 1.63 and 1.24 for the New-Old and New Economy firms, respectively. Moreover, all the Old Economy firms have at least one scheme, while some of the other firms have no schemes. This suggests that the complexity of the remuneration practices, in terms of the existence of staggered cohorts of award schemes, is a function of history, in terms of the age of the company or the period of its listing on the stock exchange. In this case, the New and New-Old Economy firms may have had too little time to reach the "steady state" number of overlapping option cohorts.
We also examined whether our sample firms extend their share option schemes from CEOs and executives to other directors and other company employees. Our findings correspond to those reported by Conyon et al. (2001) for the U.K., who find that firms with fewer hierarchical layers have more broadly based share option schemes. Specifically, we find that the younger New Economy companies, which typically have flatter organizational hierarchies, are most likely to extend their schemes to all employees--almost 64 percent of these firms do so--while none of the comparatively young but traditional New-Old Economy firms and less than 10 percent of Old Economy firms provide schemes to all their employees. The majority of Old Economy firms (67 percent) restrict their schemes to CEOs and other executive directors, and the lowest level that is reached by New-Old Economy firms (that disclose information on the participants of schemes) is the entire Board including nonexecutives. In interpreting these findings, we note that over half of the New-Old and New Economy firms provided no details about the participants of their schemes in their annual report for the sample period.
Composition of the Remuneration Committee and the Voting Power of External Shareholders
Bebchuk et al. (2002) point to power relations within the firm as a potential explanatory factor for executive compensation levels and compensation structures. Therefore, we focus on the compensation of the remuneration committee and the presence of significant external shareholders as factors that potentially influence the level of managerial compensation and the design of compensation contracts in the subsequent analysis.
Conyon and Peck (1998) document the adoption of remuneration committees by U.K. firms over the period 1991-94. They find, consistent with the Cadbury Committee compliance report published in 1995, and other studies (e.g., Conyon and Mallin 1997) that after 1992 remuneration committees were virtually universally adopted with 99 percent of FTSE 100 (largest 100 companies in terms of market capitalization) companies having a remuneration committee in 1994. For the Mid 250 (the next largest 250 companies, after FTSE 100, in terms of market capitalization) companies, Conyon and Mallin (1997) report an adoption rate of 87 percent for their 1994-95 sample. Conyon and Peck (1998) also report that the recommendation that the remuneration committee should be largely or wholly composed of nonexecutives had been widely implemented with an average proportion of nonexecutives on the RC of over 90 percent.
Panel A of Table 3 provides descriptive statistics about the composition of the compensation committees of the firms in our sample. Two Old Economy companies (slightly over 8 percent of companies in that subsample) disclose that they have no RC. Also six New Economy companies and three New-Old Economy firms give no information about their RC in their 1999 annual reports. Only 67 percent of the Old Economy firms comply with the recommendation that an RC should be composed entirely of nonexecutives. In six Old Economy firms the CEO or other executives are members of the RC. Compliance with this recommendation is slightly higher (lower) in the New (New-Old) Economy samples. Five New Economy companies and eight New-Old Economy firms have executives on the RCs of their companies. Independent pay consultants serve on RCs in between 23-45 percent of firms in our three subsamples, with the lowest participation among the New Economy firms.
Table 3 also reports, in Panels B and C, two indicators of the voting power of external shareholders on the RCs. The first shows the number of external shareholders holding at least 3 percent of the firm's equity. We note that the median number is roughly the same in all three sectors (3.5 to 4). The second set of ownership statistics shows the proportion of equity owned by shareholders who have at least 3 percent of the equity. The distribution is again roughly similar across the three sectors.
Our regression models often include total executive shareholdings as one ownership characteristic. Panel D reports the descriptive statistics for this variable, for the firms that are included in our regression models. We note that the New-Old Economy subsample exhibits significantly higher mean and median values for this variable compared to the Old Economy. The New Economy subsample is not significantly different in mean or median from the New-Old subsample, but it is significantly larger than the Old subsample in median.
Finally, Panel E of Table 3 gives figures on the shareholdings of directors in the three subsamples. As may be expected, overall, the fractional holdings of directors tend to be higher in the younger, New and New-Old, companies than in the established, Old Economy, retailers. However, the differences are not as clear-cut and pronounced as one may expect. In fact, for the CEO, there is no significant difference between their average holdings in Old and New Economy firms. CEOs in the New-Old Economy firms hold significantly larger stakes than in the other two subsamples. As for the holdings of other executives, there are significant differences in their average holdings in the younger (New and New-Old Economy) firms as opposed to their older counterparts, but the magnitudes of the differences in the corresponding medians are small (although statistically different from zero). The findings reflect the higher ownership concentration and equity retention of initial founder-owners of recently floated companies (in the New-Old sample).
The Determinants of Pay Levels
Table 4 presents the results of regressions of short-term compensation (salary and bonus), long-term compensation (stock-options and LTIPs), and total compensation, on measures of company size, sector dummies, a dividend policy dummy (1 if the firm pays dividends), and variables describing ownership and corporate governance, including the composition of the remuneration committee (RC). Panel A presents separate results for each type of director. Panel B presents pooled results for all executive directors and introduces dummy variables for the CEOs. The coefficients on the firm size and revenue, or sales, variables (expressed in natural logarithms) can be interpreted as the size and revenue elasticities of total pay and of the different pay components.
In both panels the explanatory value of the model is much greater for the salaries and bonuses of executives than it is for the long-term pay of executives. For example, in Panel B the explanatory value of the salary and bonus regression is 57 percent compared with only 13 percent for long-term pay.
Panel A (Panel B) indicates a significant positive relation between salary and bonuses and firm size for other executives and nonexecutives (all executives). However, we find no significant relation between long-term pay and firm size. The New-Old Dummy is significant in Panel B and for other executives in Panel A. The New Economy dummy is not significant in either panel. The results of the pooled regression model (for CEOs and other executives) indicate that CEOs have significantly higher total pay than other executives. The coefficients on the CEO dummy interacted with the economic sector dummies are not significant. (16)
Table 4 also considers the influence of the compensation committee on the level of managerial pay. Consistent with the theory advocated by Bebchuk et al. (2002), we find that the presence of other executives on the RC is positively related to total executive pay, while larger boards tend to lead to lower levels of pay, and the CEO on the remuneration committee leads to higher levels of total pay. Total external shareholdings exert a negative influence on pay, though this is only significant for salaries and bonuses. As in previous studies, we find that large internal director shareholdings exert a negative influence on executive pay. However, three other findings run contrary to the Bebchuk et al. (2002) theory. First, we find that the presence of an independent consultant on the RC is positively related to managerial pay; hence, there is no evidence here of a moderating influence on pay. Second, including only nonexecutives on the remuneration committee does not exert a negative influence on pay. Third, the ratio of outside to total number of directors leads to higher levels of pay.
Overall, this evidence suggests that attempts to control executive pay through independent consultants and external directors are largely ineffectual. With respect to current debates about how to control excessive executive pay, our results suggest that stimulating external institutional owners to take a more critical approach toward executive pay awards may be more productive than relying on nonexecutives and independent consultants.
Before examining the determinants of various features of the executive reward contract, we briefly examine the performance hurdles attached to the award of stock options and LTIPs in our sample.
The Greenbury Committee recommended that the award of executive stock-options and LTIPs should be conditional on meeting "challenging" performance criteria. It advocates performance measures such as shareholder return, and also suggests the use of relative performance evaluation so that directors are not rewarded for general increases in stock-market valuations and economic performance (Conyon and Mallin 1997). For the Old Economy sample 20 firms disclosed information about the performance criteria they employed. Of these, 19 referred to earnings per share growth as either the only, or one of several, performance criteria. The firms in this sample also referred to peer group performance (twelve mentions), individual targets (three mentions), profit targets (three mentions), and total shareholder return (five mentions). A substantial proportion (half or more) of New and New-Old Economy companies do not disclose information about the performance criteria employed. Assuming that firms that did use performance criteria have an incentive to disclose these criteria, we believe it is reasonable to assume that firms making no disclosures did not use performance criteria. The performance criteria most regularly disclosed by New-Old and New Economy firms was earnings per share (in total 19 out of 48 firms mention this measure). Very few of the New-Old or the New Economy firms referred to peer group performance as a relevant performance target.
The Moneyness of Executive Stock-Option Schemes
We measure "moneyness" as the share price on the grant day, SP, relative to the fixed exercise price of the options, EP. To our knowledge, there has been no systematic, in-depth analysis of the (determinants of the) moneyness of U.K. executive stock options at the grant date, and only one previous study, by Conyon and Mallin (1997), reports descriptive statistics on the exercise price relative to the grant-date share price. Conyon and Mallin (1997) document the disclosure of remuneration data by the companies comprising the FTSE 100 and Mid 250 indices of the U.K. stock market, primarily during 1994-95. They report an average discount, defined as (SP-EP)/SP, on executive share options of 0.3 percent for the Mid 250 companies and of 0.06 percent for FTSE 100 companies. They conclude that there appears to be an almost universal absence of discounts, but they also note (without providing evidence or references) that this was not always the case, and that "until recently many of the options that were issued subject to performance criteria were issued at discount to the prevailing market price" (Conyon and Mallin 1997, 47). Conyon and Mallin (1997) cite the Greenbury report, which called for executive share options never to be issued at a discount. The report also notes that for approved share option schemes, the U.K. tax rules (at the time of the report) allowed gains from the exercise of executive options and the subsequent sale of the shares to be taxed as capital gains instead of as income immediately upon option exercise, provided that the discount of the exercise price relative to the share price on the grant date did not exceed 15 percent. The 15 percent rule for approved schemes was abolished in the Finance Act 1996, but the requirements, previously stated in the Taxes Act 1988, that for approved schemes, the exercise price must be stated when the option is granted, and that it must not be "manifestly" below the market price of equivalent shares, were retained. It should be noted that the Finance Act refers only to approved schemes, while the Greenbury recommendations cover all schemes.
Contrary to these recommendations, the moneyness (SP/EP) of the executive stock options in our sample, shown in Table 5, reveal that a substantial proportion of these options are granted either at a considerable discount or a premium to the grant-day share price. The discounts are in stark contrast to both the Greenbury recommendations and current U.S. practice, where over 95 percent of options are granted at the money (Murphy 1999).
We find that the incidence of exercise-price discounts is particularly prevalent within our New Economy sample where the median moneyness, (SP/EP), is around 1.15, a 13 percent discount. The New Economy average moneyness is, however, largely the result of a small number of outliers as large as a staggering 37.5. Within the New Economy sample (and for the full three-sector sample as a whole), the distribution of moneyness is highly positively skewed (and leptokurtic) with a 75th percentile of the New Economy sample at 1.9, the 90th percentile at 19.5, and the 95th and 99th both at 37.5 (while the 1st, 5th, 10th, and 25th percentiles are 0.66, 0.71, 0.95, and 1.04, respectively).
By contrast to the large discounts granted to New Economy executives, the median moneyness for the other two samples is quite close to 1 for executives below the CEO, and also for Old-Economy CEOs. The New-Old sample shares some of the characteristics of the New Economy, in that there are some large outliers and the median CEO moneyness is closer to that in the New Economy than in the Old Economy. Thus, part of the New-Economy result may be due to the young age of the companies and the "IPO effect."
The Determinants of Moneyness
Table 6 presents the results of our attempt to model the determinants of moneyness. Our first model regresses moneyness (SP/EP) on the sector dummies and measures of firm size and other firm characteristics. As outlined in the "Data and Methodology" section above, the slopes of (the logarithms of) assets and sales are allowed to vary for the New and New-Old Economy in order to assess the marginal sector effects.
These results confirm that New Economy firms, and to a lesser (and statistically insignificant) extent, New-Old firms, have significantly higher levels of moneyness. Our further regression analysis, summarized in the remainder of the table, attempts to identify the underlying causes of this finding. The magnitude of the coefficient of the New Economy dummy is due to the impact of outliers in the dependent variable. (17)
The marginal effect of assets on moneyness in Model (1) is only significant for the New Economy, with a negative coefficient that compounds the negative coefficient of the uninteracted asset variable. By contrast, the marginal effects of assets and sales are insignificant for the New-Old sample. (18)
One possible explanation for deviations of the exercise price and the grant-date share price are unexpected share-price fluctuations between the date the exercise price is set and the actual grant date. Such fluctuations seem more likely for riskier stocks. A further reason to expect a possible link between moneyness and a measure of stock-return volatility, as a proxy of the New-Economy effect, is that this sector has both the highest average volatility and the highest average moneyness. However, contrary to expectations, volatility has no statistically significant impact on moneyness in any of our regression models.
Several other measures of firm characteristics designed to pick up the salient differences between the three sectors are included in all our models of moneyness. These are employee growth, the book-to-market ratio, the ratio of R&D expenditure to sales, a dummy variable indicating the payment of a dividend in the sample year, and the capital leverage of the company. In all our regression models, we find these variables to be statistically significant. The positive coefficients on employee growth, and R&D to sales, and the negative coefficient on leverage, are all consistent with the need to retain key employees in a high-growth situation. The significant positive value on the dividend dummy possibly reflects a need to protect the value of executive stock options from erosion by the payment of dividends. In this context, the positive coefficient on book-to-market is surprising. This could be because book-to-market proxies for bankruptcy risk as well as growth opportunities. Since we have several other measures in the regression that capture growth opportunities we may be left with a risk effect in book-to-market. Further inspection of the data suggests that the association between high book-to-market and high moneyness occurred almost entirely among traditional retailers. By contrast the high observations in the New Economy are associated with low book-to-market.
Model (2) introduces governance and ownership variables. The [R.sup.2] of the model increases from 33 percent to 39 percent, and almost all the newly added variables are statistically significant while the results for the other variables remain qualitatively unaltered. The composition of the remuneration committee (RC) appears to have a significant effect with both the presence of the CEO and that of independent consultants resulting in higher moneyness. The impact of the CEO is consistent with the Bebchuk et al. (2002) power argument. Further, the fact that independent consultants have no moderating impact, and instead appear to collude with CEOs, suggests that they cannot be relied upon to be effective monitors of remuneration policy. The influence of external shareholders on moneyness is mixed. Greater external shareholdings significantly reduce moneyness, but increasing the number of external shareholders with stakes of 3 percent or above (holding their combined shareholdings constant) increases moneyness. The latter result may reflect free-rider problems in shareholders' monitoring of management.
The magnitude of total inside shareholdings (defined as the sum of the share stakes of the CEO and all other executive directors of the firm) has no significant effect on moneyness. This could reflect two opposing effects: a greater ability to influence remuneration policy versus a greater alignment of executives' interests with those of other shareholders.
Model (3) retains all the variables of Model (2) and also allows the slopes of the interactive dummy variables indicating the CEO's presence on the RC of companies in the New and the New-Old Economy sample to vary in order to evaluate the marginal New and New-Old Economy effects. This increases the [R.sup.2] of the regression considerably from 39 percent to 57 percent. The results suggest that the marginal effect of CEO membership on the RC is positive and significant in both these sectors, but the magnitude of the effect is substantially (and statistically significantly) larger in the New Economy. Thus, there is evidence of an IPO effect, but also of an additional New Economy effect, with regard to the ability of CEOs to influence executive pay policy on the RC. However, this New Economy effect is obviously limited to those companies in the sector whose RC includes the CEO. As the descriptive statistics in Table 3 indicate, CEO membership on the RC occurs in only 24 percent of New-Economy firms (that disclose RC membership details), and is thus less common in the New-Economy sector than in the other two. Interestingly, the estimated coefficient of the (uninteracted) "CEO on RC" variable becomes negative, more highly significant, and in absolute terms, three times larger than in Model (3). Thus, surprisingly, controlling for the marginal New and New-Old Economy effects, the RC membership of the CEO significantly reduces moneyness.
In Model (3), the New-Economy dummy variable is no longer significant, and its coefficient sign changes to negative, when the slope of the "CEO on RC" variable is allowed to vary. This suggests that the differential effects of CEO membership on the RC among younger (New and New-Old Economy) companies can explain the marginal New Economy effect on the intercept in Models (1) and (2). (19)
Model (4) estimates the marginal New and New-Old Economy effects of including independent consultants on the RC. Unlike CEO membership, these effects appear not to be significant. However, controlling for the marginal effects almost doubles the (positive) coefficient estimate of the uninteracted "Consultants on RC" variable, and increases its statistical significance. This suggests that the positive consultant effect on moneyness, which contradicts the view that independent pay consultants act as moderating force, is mainly a feature of Old Economy firms.
Model (5) examines the interactive marginal effects of external stakeholders. Models (2), (3), (4), and (6) find a significant negative influence of uninteracted total external shareholdings on moneyness. Model (5) suggests that this influence is solely due to the marginal effect of (the shareholdings of) external stakeholders in New Economy firms. These shareholders have a substantial and significant limiting effect on moneyness consistent with the view that they act as efficient monitors of remuneration policy.
Including interacted variables for the number of external stakeholders renders all three variables, the uninteracted variable and the two interactive terms, (both jointly and individually) insignificant. Model (6) examines the marginal effects of inside shareholdings (the sum of the stakes of the CEO and other executives). Controlling for the marginal sector effects increases both the magnitude and the statistical significance level of the uninteracted inside-shareholdings variable. This suggests that it is mainly in the Old Economy where larger inside shareholdings reduce moneyness. This is potentially because, given the low level of executive ownership in the Old Economy sample (Table 3), an increase in executive shareholdings is likely to align the objectives of executives and shareholders more closely, and reduce executives' incentives to extract wealth from the company. The marginal effects of inside holdings in the New and New-Old Economy are positive but insignificant, possibly because with a relatively higher level of inside holdings, a further increase in executives' stakes will have only a relatively minor incremental incentive-alignment effect, and will instead increase executives' influence on pay policies.
Time to Maturity of Executive Stock-Option Schemes
Table 7 shows descriptive statistics on the difference (in calendar years) between the expiry dates of the options held by executive and nonexecutive directors and the year-end 1999. Since a given director may hold options from several cohorts of awards and with differing maturity dates, the figures are calculated on the basis of the average time to maturity of all the options held in the portfolio of a given executive.
The table reveals that the stock options of CEOs tend to have shorter maturities than those of the other executive directors in all three subsamples. This CEO effect is significant for the sample as a whole (in terms of the differences in both the means and medians) and for the Old Economy sector (for the means only), but not for the two other subsamples.
Examining the differences across the sectors, the CEOs and executives of New Economy companies have the longest terms to maturity. However, the differences in the means between the New Economy and the other two samples are only statistically significant for the CEOs. Also the corresponding differences in the medians are statistically not significant. Since a longer time to maturity normally increases option value, this partly explains the higher values of the share options received by New-Economy directors seen in Table 2 above. Comparing the two groups of traditional retailers (the Old and New-Old samples), there is only a slight, and statistically insignificant, difference.
The pattern for the other executives below the CEO is similar, but the differences between New and Old Economy are much less pronounced and not statistically significant. By contrast, the differences between New and Old Economy, on one hand, and the New-Old sample, on the other, are greater, and the difference in the medians between the New and New-Old Economy executives is statistically significant.
Pay-Performance Sensitivity of Executive Stock-Option Schemes
The pay-performance sensitivity (PPS) of executive stock options is estimated using the measure employed by Conyon and Murphy (2000) as the sum of (1) the percentage shareholdings of the director, (2) the number of options as a percentage of the total number of outstanding shares multiplied by the option "delta," and (3) the number of LTIP shares as a percentage of the total number of outstanding shares multiplied by the LTIP delta (which is assumed to equal 1). Descriptive statistics for the PPS of the directors in our sample are shown in Table 8.
The median PPS of CEOs is higher than that of other executives for all three samples, and a similar ranking is seen in the averages, except for the New Economy sample where the other executives' average PPS is higher than the CEOs'. The greater sensitivity of CEOs' option values to their stock prices makes sense in terms of incentivizing CEOs whose decisions are likely to have more impact on their companies' share prices than the decisions of other executives.
With regard to sector comparisons, the average PPS of New-Old CEOs is greater than in the other two sectors. For the other executives below the CEO, the lowest average and median PPS are observed in the Old sector. Both these results are driven by the differences in the directors' shareholdings illustrated in Table 3.
Thus, the principal difference in the CEOs' PPS is between the relatively high PPS of the New-Old CEOs and the significantly lower PPS of CEOs in the two other sectors. By contrast, among the executives below the CEO, it is the New Economy directors who have the highest average and median PPS, followed by the New-Old sample. Several of these sector differences in other executives' PPS are statistically significant (though not reported in the table, the difference between the New and Old Economy means was actually significant at the 6 percent level). Again, this is primarily due to the larger executive shareholdings in the New and New-Old sectors (Table 3), but also partly due to the greater prevalence of share options (and LTIPs) in the New Economy (Table 2).
Finally, in the New Economy sample, we observe considerably higher cross-sectional variation in the PPS for executives below the CEO (in terms of the standard deviation of PPS) than in the other two samples. By contrast, for CEOs, the PPS is least variable in the New Economy and most variable among New-Old companies.
Leverage of Compensation Packages and the Ratio of Long-Term to Short-Term Pay
As explained above, our leverage measure focuses solely on the pay-performance sensitivity and incentive effects induced by executive pay, while excluding the effects of directors' shareholdings. Our examination of leverage complements the analysis of the "regular" PPS measure (Table 8), and helps to disentangle remuneration effects from ownership effects.
Table 9 shows that the leverages of CEOs tend to be higher than those of other executives. The differences between CEOs and other executives appear largest in the New-Old and smallest in the Old Economy sector. However, the means are not statistically different in any of the sectors (nor overall). In addition, the only statistical differences in the medians (significant at 10 percent) are those between CEOs and other executives in the Old Economy, and for the sample overall. Thus, there is no clear evidence of a CEO effect in pay leverage.
We find no significant differences in the leverage of the CEOs of different sectors. Similarly, there is no significant difference between the leverage of New and Old Economy executives below the CEO, but the (average and median) leverage of New-Old Economy executives is significantly lower than in the two other sectors. Cross-sectional variation in the leverage of other executives is lowest in the New-Old Economy. By contrast, the variation in CEOs' leverage is the highest in that sector.
Panel E of Table 9 reports the (cross-sectional) correlation coefficients indicating the extent to which the leverage of the executives of the firms in each sample is related to the leverage of their CEOs. It appears that the correlation between CEO and other executives of the same firm is particularly high (99 percent) for New Economy firms, and comparatively smallest (although still rather high in absolute terms at 82 percent) for Old Economy firms.
Table 9 also presents descriptive statistics for the ratio of long-term pay to short-term pay. This measure assesses the relative importance of equity-based pay in the directors' pay package, but unlike the leverage measure, it is not designed to capture the sensitivity of pay to performance. The mean and median pay ratios of CEOs are higher than those of other executives in all three sectors. To some extent, this corresponds to the finding of higher leverage for CEOs (except for Old Economy CEOs).
New Economy CEOs appear to have higher pay ratios than CEOs of traditional retailers (Old and New-Old Economy). The same is true for other executives, although the differences are less pronounced. These findings again are roughly similar to the leverage results. Thus, both measures indicate the comparatively high prevalence and values of long-term pay in the New Economy sector.
Determinants of Time to Maturity, Leverage, Pay Ratio, and PPS
Table 10 reports our models of the determinants of time to maturity, leverage, pay ratio, and PPS in terms of firm, director, and governance characteristics. For each dependent variable, we start with a basic model comprising only measures of firm characteristics. In the second model, we add director characteristics (specifically, a dummy variable indicating whether the director is the CEO), governance variables describing the composition of the remuneration committee (RC), and the corporate ownership structure in terms of the numbers and holdings of external stakeholders and total inside shareholdings (by the CEO and other executives). In both (and all subsequent) models, we control for the marginal effects of firm size (in terms of sales and assets) on the dependent variables by interacting each size variable with both the New and the New-Old Economy dummy variables. (20)
For each dependent variable, six further regression models were estimated to assess in turn the impact of controlling for the marginal sector effects of the New and New-Old Economy on the coefficient of the uninteracted explanatory variable. Thus, we adopt the same modeling approach as in our analysis of moneyness reported in Table 6. However, instead of reporting the full regression results for the four dependent variables in Table 10, we report only the coefficient estimates of the specific explanatory variable whose slope we allow to vary across sectors. (21)
Time to Maturity
Model (1) of time to maturity explains around 29 percent of the variation in the dependent variable. The model shows a marked difference between the Old sector and the other two sectors as illustrated by the large and highly significant dummy variables for the New and New-Old Economy companies. This result is consistent with younger (IPO) firms with high current and expected future growth acting to retain the executives that took the firm up to and through the IPO-stage.
There is little evidence of size effects with respect to time to maturity. The only significant size coefficient in Model (1) is on the interactive term for New Economy directors, whose option portfolios seem to have lower average time to maturity the higher the sales of their firms. This is consistent with the retention or lock-in motive being less important for the New Economy companies at a relatively advanced stage of development. The significant negative coefficient on the dividend dummy could similarly indicate less need for executive lock-in for companies that are sufficiently developed to pay dividends or it could be linked to the need to protect the value of stock options from erosion by dividend payments. The significant negative coefficient of the ratio of R&D expenditures to sales is somewhat surprising. A possible interpretation is that it partially offsets the positive IPO effect captured by the two large, and highly significant sector dummy variables. The significant positive effect of capital leverage of the company on time to maturity suggests that the executives of riskier firms are locked in for longer periods.
Contrary to expectations, none of the other firm characteristics included in Model (1), notably employee growth, book-to-market ratio, and volatility, has a significant effect on the time to maturity. (22) Perhaps this is because in this regression the New and New-Old Economy dummy variables serve as superior proxies for these characteristics.
Model (2) introduces governance variables and a CEO dummy variable. With regard to governance factors, we find a number of significant effects consistent with arguments based on Bebchuk et al. (2002) suggesting that rent-seeking executives use their influence to obtain longer terms to maturity and increase the values of their option awards. Thus, CEO membership on the RC has a significantly positive effect on term to maturity, while independent consultants on the RC have a significant negative effect. The membership of other executives on the RC has no significant effect. Further, higher inside ownership results in longer maturities, while external stakeholders have no significant impact. Finally, CEOs appear to have significantly shorter terms to maturity than other executives.
Examining the results on marginal sector effects in Panel B of Table 10, we find that the positive effect of CEO membership of the RC is confined to the New-Old Economy. By contrast, the negative effect of independent consultants on the RC occurs primarily in the Old Economy while the marginal effect in the New-Old Economy just offsets the coefficient of the uninteracted variable.
The positive effect of larger inside shareholdings and the negative effect of the CEO dummy variable on maturity terms observed in Model (2) are also both driven by the Old Economy sample.
Our attempts to model leverage produced comparatively weak results with an [R.sup.2] of Model (2) of only 9 percent. The only firm characteristics with a significant effect are R&D to sales and the dividend dummy indicating that the pay packages of directors of more research-oriented companies (primarily those in the New Economy sector; see Table 1) have higher leverage, while those of dividend-paying companies (likely those in the mature Old Economy sector) have lower leverage.
Both independent consultants and other executives tend to reduce leverage if they are members of the RC. Both these effects are driven by the Old Economy sample, as illustrated by the results in Panel B. By contrast, in the New Economy, the membership of CEOs on the RC has a positive impact on leverage. In the Old Economy, leverage is further reduced by a larger number of external stakeholders, but this is offset by the marginal sector effects in the other subsamples. Again in the Old Economy, larger inside shareholdings result in lower leverage, but this is more than offset by the positive effect of inside shareholdings in the New Economy sample.
The New-Old sector dummy and firm characteristics are significant determinants of the pay ratio, although several of the effects appear contradictory. Thus, sales have a positive effect on pay ratio in the New Economy but a negative effect in the New-Old sample, and yet, assets have a positive effect for New-Old companies. At the same time, the significantly negative New-Old dummy indicates a lower intercept for younger traditional retailers. The positive effects of employee growth and R&D to sales are consistent with a New Economy effect, but the equally positive effect of the book-to-market ratio suggests the opposite. As argued in the discussion of the moneyness results in Table 6, high book-to-market ratios may in this context indicate recent dramatic share-price drops (relative to book values) and act as proxies of financial distress risk and cash constraints. Finally, CEOs have significantly higher pay ratios than other executives after controlling for other factors.
None of the governance variables are significant in Model (2), although when we allow for the slopes of the governance variables to vary across sectors (Panel B of Table 10), we find significant positive marginal effects of other executives' RC membership in the New Economy and of inside shareholdings in the New-Old Economy.
Pay-Performance Sensitivity (PPS)
None of the explanatory variables are significant in Model (1) of PPS, and only two variables are significant in Model (2), inside shareholdings and the CEO dummy, both with a positive effect on PPS. Nevertheless, the regression as a whole is clearly significant and the model can explain over 30 percent of the variation in PPS (indicating joint significance of the explanatory variables). Letting the slopes of the variables vary across sectors, we find that the first effect (of inside shareholding) is driven by the Old Economy and the second (the CEO effect) by the New-Old sample. These positive effects, particularly among companies operating in the traditional retailing sectors, appear to contradict the executive rent-seeking argument of Bebchuk et al. (2002). In fact, both effects have little to do with the impact of remuneration practices on PPS, but are due to the fact that individual directors' shareholdings enter additively in both the dependent and independent variables. Excluding the specific executive's holdings (which enters additively into the PPS measure) from the inside shareholdings measure results in a coefficient of similar (absolute) magnitude and statistical significance as the coefficient on inside shareholdings in Model (2), but with the opposite (negative) sign. (23) Similarly, the CEO effect in the New-Old sample seems to be due to the particularly large shareholdings of CEOs in that sector (see Table 3).
Our study examines the compensation practices of a sample of U.K.-listed retailing companies consisting of three subsamples: (1) established Old Economy retailers employing traditional retailing methods, (2) e-commerce or New Economy retailers, and (3) younger, recently listed, but traditional retailers, which we term "New-Old Economy" firms. We examine the compensation arrangements of other executives as well as CEOs. The features of compensation that we study are, principally, total salary and bonus, total long-term pay, moneyness, time to maturity, leverage, pay ratio, and PPS.
With regard to total compensation, we find that our model explains a much greater proportion of the variation in salary and bonus than the variation in long-term pay. After controlling for other factors, there is no remaining evidence of a pure New Economy effect. However, we did find that the New and New-Old sectors had higher total pay than the Old after controlling for size and the other factors. With regard to the Bebchuk et al. (2002) managerial rent-seeking theory, we found some supportive results and some results that ran contrary to their standpoint.
Probably the most novel results of this study are those relating to moneyness. While U.S. companies typically grant executive stock options at the money, this is often not the case in the U.K. Our basic regression model explains 39 percent of the cross-sectional variation in this variable. Our results on size effects, growth effects, and financial policy effects are broadly consistent with economic intuition. However, the effects of governance and ownership variables are somewhat mixed, yielding no conclusive evidence either for or against the Bebchuk et al. (2002) rent-extraction hypothesis. On balance, however, we would argue that the results favor (weakly) the Bebchuk et al. (2002) theory. Especially interesting are the significantly negative influence on moneyness of large external share ownership, and the positive influence of CEO membership of the RC. Somewhat surprising was the apparent failure of independent consultants to protect the interest of shareholders. Perhaps this indicates that they are effectively not independent. After all, consultants are hired by the (executives of the) company, and not by the shareholders. With regard to subsample differences, we found a strong differential sector effect for CEO membership of the RC. There was a very strong incremental positive effect in the New Economy, and a significantly positive, but smaller, incremental effect in the New-Old subsample.
Our time-to-maturity regression model explains 40 percent of the cross-sectional variation in this variable. This is an aspect of compensation that is relatively opaque, and so it is here that one might expect to find evidence in favor of the Bebchuk et al. (2002) hypothesis. At the same time, the choice of maturity terms may also be influenced by other considerations related to retaining and incentivizing executives. Growth companies, in particular, may choose longer terms to maturity to focus their executives' attention on future payoffs. Consistent with this effect, we find a strong positive IPO effect associated with significantly longer maturity terms in the New and New-Old Economy sectors. Consistent with the Bebchuk et al. (2002) theory are the significant negative coefficient on independent consultants, and the significant positive effects of CEO membership of the RC and inside (directors') shareholdings. An additional finding of interest is that CEOs tend to be awarded shorter time to maturity options. Overall, none of the main effects are inconsistent with the Bebchuk et al. (2002) theory, but there is evidence of other factors, such as the IPO effect, influencing the lengths of maturity terms.
The results for leverage, pay ratio, and PPS were much weaker, in terms of explanatory value, and yielded little additional insights into the governance and ownership issues. With regard to the agency model, we found no significant marginal effect of volatility on the three variables. However, we did find a significantly positive influence of employee growth and the ratio of R&D expenditure to sales on the pay ratio, which is consistent with an agency perspective in which there are costs of contracting ahead over several periods (one expects to see greater use of long-term pay where the value of the finn is largely driven by its growth opportunities). In addition, we found that directors' shareholdings exert a significant positive influence on PPS (through the direct, additive impact of individual directors' shareholdings on the calculation of the PPS measure). With regard to subsample differences, we found no pure New Economy effect in leverage, pay ratio, or PPS.
With regard to comparisons between CEOs and other directors, we discovered a number of significant effects. We found that CEOs have significantly higher salaries and bonuses and higher levels of long-term pay. The moneyness of options does not vary significantly between CEOs and other executives. However, CEOs have shorter times to maturity and higher pay ratios (particularly in the Old Economy), and they have higher sensitivity of pay to performance (particularly in the New-Old Economy). They also have higher leverage levels in all three sectors, though not significantly so.
In several of our models, we found evidence of significant size effects. In particular, we were surprised to find a tendency for the size effects to vary across sectors, and for the need to incorporate both sales and total assets as two different measures of size in the regression models. It is possible that these findings are unique to the retailing sector, where sales are an important value driver.
The research of this paper can be generalized in a number of ways. First, we were only able to use data for a single financial year. It would be interesting to look at the consequences of the bursting of the dot-com bubble in the U.K., as was done by Murphy (2003) who confirmed that most of the results reported in Ittner et al. (2003) for U.S. companies in 1999-2000 continued beyond 2000. Further U.K. work in this area could explore a greater range of industries to test some of the governance issues that we have studied for retailing. However, no other industrial sector readily lends itself to the three-way subsample comparisons of traditional and New Economy companies conducted in this study.
Black and Scholes Model
We calculate the value of the stock options, granted to the directors as part of their remuneration packages, at grant date through the Black and Scholes (1973) formula adjusted for continuously paid dividends:
Option Value = [Pe.sup.-ln(1+d)T] N(Z) - [Xe.sup.-ln(1+r)T] N(z - [sigma] [square root of T])
where P is the market price of the stock at grant date; d is the dividend yield, which is computed as the average of the prior 48 monthly observations on cash dividend per share; r is the risk-free rate (average yield on seven-year government bonds); X is the exercise price; T is the time-to-maturity of the option (in years); [sigma] is the stock volatility, calculated as the standard deviation of monthly continuously compounded returns over the prior 48 months multiplied by the square root of 12; and N(z) the cumulative normal distribution of:
z = ln(P/X) + ln(1 + r) - ln((1 + d) + [[sima].sup.2]/2)T/[sigma] [square root of T]
Share Option's Delta
Delta is a measure of the sensitivity of the option value to small changes in the price of the underlying asset (stock). We calculate delta as the partial derivative of the Black-Scholes value with respect to stock price or:
[partial derivative](Option Value)/[partial derivative](Price) = [e.sup.-dT] N(z)
In our leverage measure we use the sensitivity of the stock option value with respect to a 1 percent change in stock price or else the pound change. This is defined as:
Sensitivity of option value to stock price = [e.sup.-dT] N(z) x ([price/100])
When a director holds a portfolio of options, then the delta of the portfolio is calculated as the sum of the deltas of each component of the portfolio.
TABLE 1 Descriptive Statistics of Sample Companies Panel A: Old Economy Variable Obs. Mean Median Market capitalization (million 24 2245.10 (d) 581.44 (d) [pounds sterling]) Sales (million [pounds sterling]) 24 3504.10 (b,d) 1201.30 (b,d) Sales Growth (year t-1) 24 0.098 (b,d) 0.093 (b,d) Net Earnings per Share (EPS) 24 2.04 15.44 (b,d) Dividends per Share 24 7.94 (b,d) 5.95 (b,d) Total No. of Employees (OOOs) 24 39.05 (b,d) 15.88 (b,d) Employee Growth (year t-1) 24 0.059 (b) 0.061 (b) R&D/Sales 24 0.00 (b) 0.00 (b) Total Assets (million [pounds 24 2306.23 (b,d) 802.85 (b,d) sterling]) Volatility 24 0.32 (b,d) 0.32 (b,d) Book to Market Value 24 0.58 (b) 0.38 (b) Company Leverage 24 0.120 (b,d) 0.120 (b,d) Panel B: New Economy (a) Variable Obs. Mean Median Market capitalization (million 22 1567.00 (c) 473.26 (c) [pounds sterling]) Sales (million [pounds sterling]) 22 353.22 (b) 144.87 (b) Sales Growth (year t-1) 22 0.728 (c) 0.339 (b,c) Net Earnings per Share (EPS) 22 0.62 4.15 (b) Dividends per Share 22 1.30 (b,c) 0.68 (b) Total No. of Employees (OOOs) 22 2.44 (b) 1.18 (b) Employee Growth (year t-1) 22 0.461 (b,c) 0.225 (b,c) R&D/Sales 22 0.22 (b,c) 0.08 (b,c) Total Assets (million [pounds 22 389.48 (b) 154.55 (b,c) sterling]) Volatility 22 0.74 (b,c) 0.61 (b,c) Book to Market Value 22 0.054 (b,c) 0.040 (b,c) Company Leverage 22 0.049 (1) 0.002 (1) Panel C: New-Old Economy Variable Obs. Mean Median Market capitalization (million 24 137.77 (c,d) 33.83 (c,d) [pounds sterling]) Sales (million [pounds sterling]) 24 329.74 (d) 56.12 (d) Sales Growth (year t-1) 24 0.389 (d) 0.160 (c,d) Net Earnings per Share (EPS) 24 5.94 6.06 (d) Dividends per Share 24 3.20 (c,d) 2.98 (d) Total No. of Employees (OOOs) 24 4.12 (d) 0.57 (d) Employee Growth (year t-1) 24 0.087 (c) 0.000 (c) R&D/Sales 24 0.00 (c) 0.00 (c) Total Assets (million [pounds 24 127.08 (d) 33.14 (c,d) sterling]) Volatility 24 0.50 (c,d) 0.44 (c,d) Book to Market Value 24 0.48 (c) 0.39 (c) Company Leverage 24 0.048 (d) 0.011 (d) Panel A: Old Economy Variable Max. Min. Std. Market capitalization (million 12802.88 71.38 3487.59 [pounds sterling]) Sales (million [pounds sterling]) 18796.00 100.43 5218.95 Sales Growth (year t-1) 0.556 -0.267 0.138 Net Earnings per Share (EPS) 47.45 -260.47 60.92 Dividends per Share 21.00 0.00 6.90 Total No. of Employees (OOOs) 205.22 1.42 57.07 Employee Growth (year t-1) 0.411 -0.396 0.189 R&D/Sales 0.02 0.00 0.00 Total Assets (million [pounds 10552.00 81.43 3331.96 sterling]) Volatility 0.48 0.17 0.09 Book to Market Value 3.73 -0.18 0.76 Company Leverage 0.425 0.000 0.108 Panel B: New Economy (a) Variable Max. Min. Std. Market capitalization (million 7261.21 144.23 1935.11 [pounds sterling]) Sales (million [pounds sterling]) 1760.63 3.74 446.85 Sales Growth (year t-1) 5.039 -0.227 1.247 Net Earnings per Share (EPS) 30.64 -76.77 22.46 Dividends per Share 4.40 0.00 1.50 Total No. of Employees (OOOs) 10.27 0.11 2.98 Employee Growth (year t-1) 1.679 -0.233 0.586 R&D/Sales 2.77 0.00 0.58 Total Assets (million [pounds 3928.18 34.94 818.20 sterling]) Volatility 2.10 0.26 0.42 Book to Market Value 0.340 -0.110 0.124 Company Leverage 0.279 0.000 0.082 Panel C: New-Old Economy Variable Max. Min. Std. Market capitalization (million 1425.03 3.07 299.12 [pounds sterling]) Sales (million [pounds sterling]) 5465.70 4.54 1101.28 Sales Growth (year t-1) 3.128 -0.005 0.752 Net Earnings per Share (EPS) 25.00 -33.55 12.15 Dividends per Share 7.40 0.00 2.74 Total No. of Employees (OOOs) 66.14 0.04 13.36 Employee Growth (year t-1) 0.474 -0.059 0.133 R&D/Sales 0.00 0.00 0.00 Total Assets (million [pounds 1724.90 0.85 352.86 sterling]) Volatility 1.01 0.23 0.21 Book to Market Value 2.50 -0.7 0.59 Company Leverage 0.203 0.000 0.067 The sample consists of the population of 24 New Economy (dot-com/e-commerce) retailers, and 48 traditional, Old Economy retailers. The sample of 48 traditional retailers comprises 24 firms randomly selected from the population of traditional retailers that were listed on the London Stock Exchange between January 1996 and January 1999, and a further 24 firms randomly selected from the population of established or "seasoned" Old Economy retailers (listed before 1996). All fractions are presented in decimals. (a) There were no full accounting details in Datastream for two New Economy firms. (b) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between Old and New Economy. (c) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between New and New-Old. (d) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between Old and New-Old. TABLE 2 Director Cash and Equity-Based Compensation by Type of Director and Economic Sector Panel A: Number of Directors Receiving Cash and Equity-Based Compensation Number of Directors Receiving Annual Number Base Salary Bonus of Directors Obs. % Obs. % CEO 68 68 100 39 57.3 Executives 253 247 97.6 115 45.4 Nonexecutives (a) 228 225 98.7 0 0 Panel Al: Old Economy CEO 22 (b) 22 100 13 59.1 Executives 119 119 100 54 45.4 Nonexecutives (a) 100 100 100 0 0 Panel A2: New Economy CEO 22 (b) 22 100 17 77.3 Executives 56 55 98.2 33 60.0 Nonexecutives (a) 62 59 95.2 0 0 Panel A3: New-Old Economy CEO 24 24 100 9 37.5 Executives 78 73 93.6 28 35.9 Nonexecutives (a) 66 66 100 0 0 Share Options LTIP Shares Obs. % Obs. % CEO 51 75.0 9 13.2 Executives 177 69.1 33 12.9 Nonexecutives (a) 9 3.9 0 0 Panel Al: Old Economy CEO 19 86.4 6 27.3 Executives 83 68.0 18 14.7 Nonexecutives (a) 0 0 0 0 Panel A2: New Economy CEO 16 72.7 1 4.4 Executives 43 76.8 6 11.0 Nonexecutives (a) 8 12.9 0 0 Panel A3: New-Old Economy CEO 16 66.6 2 8.3 Executives 51 65.4 9 11.5 Nonexecutives (a) 1 1.5 0 0 Panel B: Amount (Pounds) of Cash and Equity-Based Compensation Base Salary Annual Bonus ([pounds sterling] ([pounds sterling] Number 000s) 000s) of Directors Mean Median Mean Median CEO 68 276 227 131 28 Executives 253 168 178 62 0 Nonexecutives (a) 228 31 22 0 0 Panel B1: Old Economy CEO 22 349 (c) 321 (c,e) 106 40 Executives 119 218 (c,e) 247 (c,e) 55 (e) 0 (c) Nonexecutives (a) 100 39 (c,e) 25 0 0 Panel B2: New Economy CEO 22 197 (c) 189 (c) 246 93 (d) Executives 56 141 (c,d) 169 (c,d) 125 (d) 22 (c,d) Nonexecutives (a) 62 24 (c) 20 0 0 Panel B3: New-Old Economy CEO 24 286 143 (e) 46 0 (2) Executives 78 112 (d,e) 95 (d,e) 24 (d,e) 0 (2) Nonexecutives (a) 66 24 (e) 20 0 0 Share Options LTIP Shares ([pounds sterling] ([pounds 000s) sterling] 000s) Mean Median Mean Median CEO 606 81 54 0 Executives 351 37 34 0 Nonexecutives (a) 41 32 NA NA Panel B1: Old Economy CEO 523 121 (c) 76 0 Executives 155 (c) 56 (e) 35 0 Nonexecutives (a) NA NA NA NA Panel B2: New Economy CEO 1,139 (d) 162 (c) 25 0 Executives 458 (c,d) 63 49 0 Nonexecutives (a) 43 32 NA NA Panel B3: New-Old Economy CEO 217 (d) 45 56 0 Executives 172 (d) 16 (e) 21 0 Nonexecutives (a) 34 34 NA NA Panel A: If a director holds options from more than one scheme, then the option holdings are treated as a portfolio; see the Appendix. Panel B: The mean, median, and number of observations (directors) were calculated based on the data of all directors eligible to get an award/grant, not only for those directors that actually received such an award (that is, zero values are included). NA means that no director in that (sub-) sample was eligible to receive the given type of award/grant. We have annualized the salary and bonus for 35 directors, who held their posts for less than a year. (a) Nonexecutive directors receive fees not base salary. (b) Two of the Old and New Economy firms have no CEO but instead two joint managing directors (MD); we classify the MDs as other executives. (c) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between Old and New Economy. (d) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between New and New-Old. (e) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between Old and New-Old. TABLE 3 Descriptive Statistics on the Composition of the Remuneration Committee (RC) and Ownership Variables Panel A: Composition of the Remuneration Committee CEO is Other Member of Executives Sector Obs. RC on RC Old Economy 24 5 (21%) 5 (21%) New Economy 17 * 4 (a) (23%) 3 (c) (18%) New-Old Economy 20 * 8 (b) (40%) 3 (d) (15%) Panel B: Number of External Stakeholders Obs. Mean Median Old Economy 24 3.625 3.500 New Economy 17 ** 4.588 4.000 New-Old Economy 20 ** 3.750 4.000 Panel C: Total Shareholdings of External Stakeholders Obs. Mean Median Old Economy 24 0.276 0.285 New Economy 17 ** 0.345 0.340 New-Old Economy 20 ** 0.296 0.275 Panel D: Total Shareholdings of Executive Directors (CEO and Other Executives) Obs. Mean Median Old Economy 24 0.078 (3) 0.009 (1,3) New Economy 17 ** 0.155 0.072 (1) New-Old Economy 20 ** 0.271 (3) 0.268 (3) Panel E: Directors' Shareholdings (All Economic Sectors) Obs. Mean Median CEO 57 0.066 0.004 Executives 234 0.026 0.001 Panel E1: Directors' Shareholdings (Old Economy) CEO 22 0.033 (3) 0.001 (3) Executives 119 0.009 (1,3) 0.0001 (1,3) Panel E2: Directors' Shareholdings (New Economy) CEO 16 0.028 (2) 0.009 (2) Executives 48 0.046 (1) 0.002 (1) Panel E3: Directors' Shareholdings (New-Old Economy) CEO 19 0.135 (2,3) 0.096 (2,3) Executives 67 0.043 (3) 0.000 (3) Only Non- Independent Executives Consultants Sector on RC on RC No RC Old Economy 16 (67%) 10 (42%) 2 (8%) New Economy 12 (71%) 4 (23%) 0 (0%) New-Old Economy 12 (60%) 9 (45%) 0 (0%) Panel B: Number of External Stakeholders Max. Min. Std. Old Economy 7.000 0.000 1.813 New Economy 11.000 1.000 2.874 New-Old Economy 8.000 0.000 2.049 Panel C: Total Shareholdings of External Stakeholders Max. Min. Std. Old Economy 0.730 0.000 0.180 New Economy 0.800 0.040 0.221 New-Old Economy 0.620 0.000 0.180 Panel D: Total Shareholdings of Executive Directors (CEO and Other Executives) Max. Min. Std. Old Economy 0.682 0.000 0.166 New Economy 0.769 0.000 0.216 New-Old Economy 0.758 0.000 0.234 Panel E: Directors' Shareholdings (All Economic Sectors) Max. Min. Std. CEO 0.641 0.000 0.129 Executives 0.724 0.000 0.094 Panel E1: Directors' Shareholdings (Old Economy) CEO 0.641 0.000 0.136 Executives 0.360 0.000 0.043 Panel E2: Directors' Shareholdings (New Economy) CEO 0.160 0.000 0.041 Executives 0.721 0.000 0.136 Panel E3: Directors' Shareholdings (New-Old Economy) CEO 0.450 0.000 0.148 Executives 0.724 0.000 0.118 Panel A: Figures shown indicate the number of sample firms in each category in the three economic sectors (subsamples). In parentheses we provide the corresponding percentages (calculated over the total number of companies in the sector for which we have details on the RC composition). (a) In two companies in the sector the RC consults the CEO but the CEO is not a member of the RC. (b) In a further two companies the CEO is consulted but is not a member of the RC. (c) In one further company the executives are consulted but are not members of the RC. (d) In a further two companies executives are consulted but are not members of the RC. Panel B: Descriptive statistics on the number of external shareholders that possess 3 percent or more of the company's outstanding shares. Panel C: Descriptive statistics on the cumulative fraction of outstanding shares in the possession of external stakeholders. Panel D: Descriptive statistics (by company) on the total executive directors' shareholdings of the sample companies as a fraction (decimal) of outstanding shares. Panels E and E1-3: Descriptive statistics (by director) on the individual fractional shareholdings of the executive directors (CEO and other executives). In Panels B-D, the number of observations indicates the number of sample firms in a given category. In Panels E and E1-3, the number of observations indicates executive directors. * Data on RC composition was unavailable from the annual report for 7 companies in the New Economy and 4 in the New-Old Economy sample. ** Observations on sample firms and executives that are not used in the regressions of Tables 9 and 10, due to unavailability of data on the RC composition of the sample company, are excluded. This reduces the number of companies in the New and New-Old samples. (1) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between Old and New Economy. (2) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between New and New-Old. (3) Difference in the means (medians) for corresponding director and compensation types between subsamples is significantly different from 0 at the 10 percent (5 percent) level (two-tailed) between Old and New-Old. TABLE 4 Regression Results for the 1999 Directors' Compensation Panel A: Regression Models Estimated Separately for CEOs, Other Executives and Nonexecutives Dependent Variable In(salary + bonus) Other Non Independent Variable CEO Execs Execs Intercept 7.83 *** 4.65 *** 6.41 *** ln(MV) 0.13 0.21 *** -0.05 ln(Sales) 0.08 0.13 *** 0.24 *** New Economy 0.04 0.12 0.24 New-Old Economy 0.10 0.29 * -0.01 Board Size -0.04 ** -0.08 *** -0.02 Out/Total Dirs. on Board 1.10 0.57 0.09 No of Ext. Stakeholders 0.03 0.05 -0.03 Total Ext. Shareholdings -1.13 ** -1.09 ** 0.19 Total Inside Shareholdings -0.53 -0.70 ** 0.06 Dividend Dummy 0.52 0.23 * 0.27 CEO on RC 0.09 0.55 *** -0.07 Only Nonexecs on RC 0.28 0.83 *** -0.28 Ind. Consultants(s) on RC 0.38 0.25 *** 0.19 Other Executive(s) on RC 0.30 0.58 *** -0.41 * Sample Size 57 234 214 [R.sup.2] 0.57 0.55 0.24 F-statistic 13.58 18.67 6.54 Dependent Variable In(long-term) Other Non Independent Variable CEO Execs Execs Intercept 3.08 8.24 *** NA ln(MV) -0.13 -0.05 NA ln(Sales) 0.44 0.19 NA New Economy 0.21 0.21 NA New-Old Economy -0.42 0.20 NA Board Size 0.06 -0.01 NA Out/Total Dirs. on Board 4.82 * 1.77 NA No of Ext. Stakeholders 0.24 0.04 NA Total Ext. Shareholdings -1.37 1.13 NA Total Inside Shareholdings 2.97 0.35 NA Dividend Dummy -0.38 -0.48 NA CEO on RC 0.04 -1.17 * NA Only Nonexecs on RC -0.07 -0.69 NA Ind. Consultants(s) on RC -0.22 -0.24 NA Other Executive(s) on RC -0.44 0.75 NA Sample Size 48 171 NA [R.sup.2] 0.19 0.10 NA F-statistic 1.22 2.49 NA In(total pay) Other Non Independent Variable CEO Execs Execs Intercept 4.48 * 4.10 *** NA ln(MV) 0.12 0.29 *** NA ln(Sales) 0.21 0.09 NA New Economy 0.45 0.28 NA New-Old Economy 0.19 0.52 ** NA Board Size -0.03 -0.09 *** NA Out/Total Dirs. on Board 2.83 *** 1.16 ** NA No of Ext. Stakeholders 0.03 0.03 NA Total Ext. Shareholdings -1.14 -0.41 NA Total Inside Shareholdings -0.17 -0.90 ** NA Dividend Dummy 0.18 0.06 NA CEO on RC 1.00 * 0.50 * NA Only Nonexecs on RC 1.12 * 0.94 ** NA Ind. Consultants(s) on RC 0.34 0.28 ** NA Other Executive(s) on RC 0.39 0.92 *** NA Sample Size 57 234 NA [R.sup.2] 0.48 0.39 NA F-statistic 7.02 10.48 NA Panel B: Regression Models Using Pooled Data for CEOs and Other Executives Dependent Variable In(salary + In(total Independent Variable bonus) In(long-term) pay) Intercept 5.13 *** 7.01 *** 4.04 *** ln(MV) 0.19 *** -0.07 0.24 *** ln(Sales) 0.12 *** 0.25 0.13 New Economy 0.12 0.34 0.40 New-Old Economy 0.23 * 0.20 0.47 ** CEO Dummy 0.62 *** 0.99 ** 0.89 *** CEO x New Economy -0.13 -0.28 -0.20 CEO x New-Old Economy 0.10 -0.49 -0.10 Board Size -0.07 *** 0.01 -0.08 *** Out/Total Dirs. on Board 0.71 ** 2.55 ** 1.56 *** No of Ext. Stakeholders 0.05 0.08 0.03 Total Ext. Shareholdings -1.07 *** 0.54 -0.57 Total Inside Shareholdings -0.68 *** 0.75 -0.77 *** Dividend Dummy 0.30 ** -0.52 0.06 CEO on RC 0.49 *** -0.98 0.58 ** Only Nonexecs on RC 0.75 *** -0.68 0.92 *** Ind. Consultant(s) on RC 0.27 *** -0.25 0.29 ** Other Executive(s) on RC 0.54 *** 0.52 0.81 *** Sample Size 291 219 291 [R.sup.2] 0.57 0.13 0.44 F-statistic 24.39 2.81 12.64 *, **, *** Significantly different from zero at 10 percent, 5 percent, and 1 percent, respectively (White-corrected Standard Errors). MV is the average market value of the firm's equity during 1999, and Sales is total sales in 1999. The coefficients on ln(MV) and In(Sales) can be interpreted as the elasticities of directors' pay with respect to size and revenues. We estimate and report separate regressions for short-term compensation (salary and bonus), long-term (share options and LTIPs), and total pay (short- and long-term). Similarly, we estimate and report separate regressions for CEOs, other executives and nonexecutives; however, we could not estimate separate regression for nonexecutives' long-term and total pay due to the small number of observations of long-term pay to nonexecutive directors. Two dummy variables to denote companies in the New Economy and New-Old Economy sectors (or samples) and variables on firm size, board size, proportion of outside directors on the board, external (stakeholders of 3 percent or more) and inside (executive directors) ownership; dividend dummy and the composition of the Remuneration Committee are included as additional explanatory variables. TABLE 5 Descriptive Statistics for the Moneyness (SP/EP) of Executive Stock Option Portfolios Panel A: Descriptive Statistics of Moneyness (SP/EP) for Total Sample (All Economic Sectors) Obs. Mean Std. Min. CEO 47 2.711 6.437 0.284 Other executives 165 2.174 5.181 0.461 Discount 0-15% 43 1.058 0.043 1.003 Discount greater 92 4.067 7.986 1.154 than 15% Panel B: Old Economy CEO 19 1.175 0.616 0.284 Other Executives 83 1.105 (a,c) 0.295 0.632 Panel C: New Economy CEO 14 5.737 11.332 0.676 Other Executives 38 4.979 (a,b) 10.206 0.662 Panel D: New-Old Economy CEO 14 1.770 1.850 0.461 Other executives 44 1.769 (b,c) 1.804 0.461 25th 75th Percentile Median Percentile Max. CEO 0.997 1.081 1.645 37.500 Other executives 0.983 1.060 1.378 37.500 Discount 0-15% 1.021 1.040 1.105 1.149 Discount greater 1.224 1.545 2.106 37.500 than 15% Panel B: Old Economy CEO 0.918 1.021 1.202 2.991 Other Executives 0.979 1.024a 1.195 2.576 Panel C: New Economy CEO 1.000 1.115 1.923 37.500 Other Executives 1.040 1.163a 1.895 37.500 Panel D: New-Old Economy CEO 1.000 1.105 1.666 7.810 Other executives 0.994 1.039 1.616 7.810 The table shows figures for the moneyness of executive stock options at the grant date measured as the weighted average ratio of the share price on the grant day (SP) over the exercise price of the option (EP). The second column of the table (Obs.) reports the total number of directors' portfolios of executive stock options. Since a given director may hold options awarded at different times and under more than one scheme, the figures below refer to averages of the portfolios held by directors who were awarded stock options. Observations on (portfolio) SP/EP that were excluded from the regressions reported in Table 6, due to unavailability of data on the RC composition of the sample company, were also excluded here. The table also reports, for each category, the mean (calculated over the number of observations in the second column), standard deviation, minimum and maximum and quartile cut-off points for the distribution of the portfolio SP/EP. (a) Difference in the means (medians) of corresponding director types between the Old and New Economy subsamples is significantly different from zero at the 1 percent (5 percent) level (two-tailed). (b) Difference in the means of corresponding director types between New and New-Old Economy subsamples is significantly different from zero at the 5 percent level (two-tailed). (c) Difference in the means of corresponding director types between the Old and New-Old Economy subsamples is significantly different from zero at the 5 percent level (two-tailed). TABLE 6 Determinants of Moneyness (share price at grant date over exercise price SP/EP) (1) (2) (3) Intercept 1.54 6.38 17.82 *** New Economy 66.15 ** 57.48 ** -15.56 New-Old Economy 7.18 7.37 -19.35 ** ln(Sales) 2.54 *** 3.33 *** 6.46 *** ln(Sales) x New Economy 1.33 1.35 -0.80 ln(Sales) x New-Old Economy -1.07 -0.56 -2.78 ** ln(Total Assets) -2.83 *** -3.91 *** -7.55 *** ln(Total Assets) x -4.63 ** -4.20 ** 1.56 New Economy ln(TotalAssets) x 0.64 0.06 3.66 ** New-Old Economy Volatility -0.11 0.40 0.81 Employee Growth (t-1) 3.63 ** 4.51 *** 2.51 * Book/Market Value 1.35 *** 1.70 *** 0.96 ** R&D/Sales 5.04 ** 6.13 *** 8.06 *** Dividend Dummy 5.07 *** 5.85 *** 6.49 *** Company Leverage -10.52 *** -17.81 *** -29.60 *** CEO on RC 1.47 * -5.42 *** Ind. Consultant(s) on RC 1.73 ** 1.39 ** No of External Stakeholders 0.72 ** 1.19 *** Total External Shareholdings -10.76 *** -14.38 *** Total Inside Shareholdings -3.35 -2.21 CEO on RC x New Economy 18.27 *** CEO on RC x New-Old 4.36 *** Economy Ind. Consultant(s) on RC x New Economy Ind. Consultant(s) on RC x New-Old Economy No of External Stakeholders x New Economy No of External Stakeholders x New-Old Economy Total External Shareholdings x New Economy Total External Shareholdings x New-Old Economy Total Inside Shareholdings x New Economy Total Inside Shareholdings x New-Old Economy Sample Size 212 212 212 [R.sup.2] 0.33 0.39 0.57 F-statistic 1.66 1.35 2.35 (4) (5) (6) Intercept 3.68 0.85 8.55 New Economy 56.40 ** 68.63 ** 50.07 New-Old Economy 12.06 11.78 4.75 ln(Sales) 2.81 *** 3.33 *** 3.31 *** ln(Sales) x New Economy 1.95 2.69 1.46 ln(Sales) x New-Old Economy -0.06 -1.53 -0.22 ln(Total Assets) -3.30 *** -3.69 *** -3.98 *** ln(Total Assets) x -4.70 ** -6.10 ** -3.97 * New Economy ln(TotalAssets) x -0.67 0.77 -0.16 New-Old Economy Volatility 0.24 1.57 0.61 Employee Growth (t-1) 4.28 *** 5.75 *** 4.07 ** Book/Market Value 1.67 *** 1.81 *** 1.55 ** R&D/Sales 6.07 *** 7.70 *** 6.08 *** Dividend Dummy 5.64 *** 6.03 *** 5.98 *** Company Leverage -15.49 *** -15.86 *** -19.17 *** CEO on RC 1.90 * 1.51 1.50 * Ind. Consultant(s) on RC 2.58 *** 1.61 ** 1.67 ** No of External Stakeholders 0.75 ** 0.23 0.89 ** Total External Shareholdings -9.32 ** -3.94 -12.93 ** Total Inside Shareholdings -2.47 -0.68 -0.39 ** CEO on RC x New Economy CEO on RC x New-Old Economy Ind. Consultant(s) on RC x -3.10 New Economy Ind. Consultant(s) on RC x -1.66 New-Old Economy No of External Stakeholders x 0.90 New Economy No of External Stakeholders x -0.45 New-Old Economy Total External Shareholdings x -17.02 ** New Economy Total External Shareholdings x 7.54 New-Old Economy Total Inside Shareholdings x 6.75 New Economy Total Inside Shareholdings x 2.26 New-Old Economy Sample Size 212 212 212 [R.sup.2] 0.40 0.41 0.39 F-statistic 1.20 1.23 1.27 *, **, *** Statistically significantly different from zero at 10 percent, 5 percent, and 1 percent, respectively (White-corrected Standard Errors). The dependent variable is moneyness, calculated as in Table 5. The moneyness regressions were estimated using observations on directors' portfolios of stock options. The regression models include sector dummies for the New and New-Old Economy samples; firm size variables uninteracted and interacted with the sector dummies; volatility, employee growth for the previous year, the ratios of book to market and of R&D expenditure to sales, a dividend dummy and company (capital) leverage; variables of external stakeholders (holding 3 percent or more) and inside (executive directors') shareholdings, and dummy variables for aspects of the composition of the remuneration committee (RC), specifically the RC membership of the CEO and independent pay consultants (both un- and interacted). TABLE 7 Time to Maturity of Sample Option Portfolios Panel A: TtM for Total Sample (All Economic Sectors) Obs. Mean Median Skewness CEOs 47 6.45 6.32 -0.01 Other Executives 165 7.57 8.37 -0.67 Nonexecutives 9 8.46 9.70 -1.56 Panel B: Old Economy CEOs 19 6.19 (a) 6.15 0.36 Other Executives 83 7.76 8.69 -0.65 Nonexecutives 0 NA NA NA Panel C: New Economy CEOs 14 7.45 (a,b) 8.68 -0.82 Other Executives 38 7.82 8.99 (b) -1.16 Nonexecutives 8 8.45 9.70 -1.56 Panel D: New-Old Economy CEOs 14 5.76 (b) 5.48 0.28 Other Executives 44 6.87 6.81 (b) -0.27 Nonexecutives 1 8.52 -- -- Kurtosis Max. Min. Std. CEOs 1.88 10.05 2.02 2.36 Other Executives 2.40 10.46 1.66 2.23 Nonexecutives -- 10.59 5.06 2.97 Panel B: Old Economy CEOs 2.50 10.01 2.78 2.15 Other Executives 2.32 10.42 2.54 2.12 Nonexecutives NA NA NA NA Panel C: New Economy CEOs 2.39 10.05 2.25 2.46 Other Executives 3.13 10.46 1.66 2.48 Nonexecutives -- 10.59 5.06 2.97 Panel D: New-Old Economy CEOs 2.00 9.36 2.02 2.35 Other Executives 2.02 10.10 2.02 2.08 Nonexecutives -- -- -- -- Time to maturity (TtM) is measured as the average of the differences between the expiry dates of the options in a portfolio and the 31 December 1999. The number of observations in column two refers to portfolios of options awarded to directors (as explained in the notes to Table 5). CEO and Other Executive observations on TtM that are not used in the regressions of Table 10, due to unavailability of data on the RC composition, are excluded. All figures are expressed in calendar years. NA indicates that no options were granted to the nonexecutive directors of this sector. (a) Difference in the means between Old and New Economy subsamples is significantly different from zero at the 1 percent level (two-tailed). (b) Difference in the means (medians) between New Economy and New-Old Economy subsamples is significantly different from zero at the 1 percent level (two-tailed). TABLE 8 Descriptive Statistics for the Pay-Performance Sensitivity of Directors' Pay by Economic Sector and Director Type Panel A: PPS for Total Sample (All Economic Sectors) Obs. Mean Median CEO 57 0.069 0.011 Executives 234 0.028 0.000 Nonexecutives 228 0.009 0.000 Panel B: Old Economy CEO 22 0.037 (c) 0.002 (a,c) Executives 119 0.010 (c) 0.000 (a,c) Nonexecutives 100 0.005 0.000 Panel C: New Economy CEO 16 0.032 (b) 0.015 (a,b) Executives 48 0.048 0.004 (a) Nonexecutive 62 0.007 0.000 Panel D: New-Old Economy CEO 19 0.137 (b,c) 0.104 (b,c) Executives 67 0.044 (c) 0.002 (c) Nonexecutives 66 0.015 0.000 Std. Std. Min. Max. Dev. Error CEO 0.000 0.641 0.129 0.017 Executives 0.000 0.724 0.095 0.006 Nonexecutives 0.000 0.473 0.042 0.003 Panel B: Old Economy CEO 0.000 0.641 0.136 0.029 Executives 0.000 0.360 0.043 0.004 Nonexecutives 0.000 0.195 0.026 0.003 Panel C: New Economy CEO 0.000 0.161 0.041 0.010 Executives 0.000 0.721 0.137 0.019 Nonexecutive 0.000 0.242 0.036 0.005 Panel D: New-Old Economy CEO 0.000 0.464 0.147 0.034 Executives 0.000 0.724 0.118 0.014 Nonexecutives 0.000 0.473 0.062 0.008 The table shows descriptive statistics of the pay-performance sensitivity (PPS) of director's pay. The PPS is calculated according to the Conyon and Murphy (2000) definition: PPS = (Shares held as a fraction of firm shares) + (Options held as a fraction of firm shares) x (Option Delta) + (LTIP shares as a fraction of firm shares) x (LTIP Delta) Option delta is a measure of the sensitivity of the option value to small changes in the share price (see the Appendix). If a director holds options from more than one scheme, then the director's options are treated as a portfolio, and the director's delta is calculated as the (unweighted) sum of the deltas of the options from the different schemes. LTIP delta equals 1. The numbers of observations shown in column two refer to director and (unlike Tables 5 and 7) include directors who received no equity-based compensation. CEO and Other Executive observations on PPS that are not used in the regressions of Table 10, due to unavailability of data on the RC composition, are excluded. (a) Difference in the means (medians) between the Old and New Economy subsamples is significantly different from zero at the 1 percent level (two-tailed). (b) Difference in the means (medians) between the New and New-Old Economy subsamples is significantly different from zero at the 1 percent (10 percent) level (two-tailed). (c) Difference in the means (medians) between the Old and New-Old Economy subsamples is significantly different from zero at the 1 percent (5 percent) level (two-tailed). TABLE 9 Summary Statistics for the Leverage and Pay-Ratio Measures by Sector and Director Type Panel A: Descriptive Statistics of Leverage and Pay Ratio Measures for Total Sample (All Economic Sectors) Leverage Obs. Mean Median Std. CEO 57 0.021 0.003 0.064 Executives 234 0.013 0.001 0.054 Panel B: Old Economy CEO 22 0.011 0.004 0.017 Executives 119 0.014 (c) 0.002 (c) 0.068 Panel C: New Economy CEO 16 0.027 0.003 0.065 Executives 48 0.018 (b) 0.002 (b) 0.053 Panel D: New-Old Economy CEO 19 0.028 0.002 0.093 Executives 67 0.006 (b,c) 0.001 (b,c) 0.015 Pay Ratio Mean Median Std. CEO 2.204 0.564 4.577 Executives 1.027 0.226 2.240 Panel B: Old Economy CEO 1.681 0.456 3.059 Executives 0.629 (a) 0.230 1.506 Panel C: New Economy CEO 4.217 1.002 7.455 Executives 1.967 (a) 0.257 3.391 Panel D: New-Old Economy CEO 1.114 0.439 1.664 Executives 1.052 0.162 2.129 Panel E: Cross-Sectional Correlation of the Leverage of CEOs' and Other Executives' Compensation Leverage of CEO Compensation New-Old Old Economy New Economy Economy Leverage of other 0.82 ** 0.99 ** 0.87 ** executives' compensation Panel F: Cross-Sectional Correlation of CEOs' and Other Executives' Pay Ratio CEOs' Pay Ratio New-Old Old Economy New Economy Economy Other executives' pay ratio 0.99 ** 0.95 ** 0.74 ** For the total sample and the three subsamples separately, Panels A, B, C, D report the mean, median, and standard deviation (Std.) of the pay ratio and the leverage measure L. The pay ratio is defined as long-term pay (the monetary values of the director's executive stock options and LTIPs) over short-term pay (the monetary value of salary and bonus). The remuneration leverage is defined as: L = [n.sub.1] x [d.sub.o] x 0.01p + [n.sub.2] x [d.sub.LTIP] x 0.01p / s + b where L is the leverage of the compensation scheme, [n.sub.1] is the number of options held by the director, [n.sub.2] is the number of LTIP shares held by the director, [d.sub.o] is the delta of the stock option, [d.sub.LTIP] is the delta of the long-term incentive plan, p is the company's year-end stock price, s is the base salary of the director, and b is the annual cash bonus of the director. Panel E shows the cross-sectional correlation (across firms in a given subsample) between the leverage of the CEO's compensation and the leverage of the other executives' compensation of the same firm. Panel F shows the cross-sectional correlation (across firms in a given subsample) between the pay ratio of the CEO's compensation and the pay ratio of the other executives' compensation of the same firm. In this table, the number of observations refers to executives, and includes directors who received no equity-based compensation. Observations on the directors of sample companies that were excluded from the regressions reported in Table 10, due to unavailability of data on the RC composition, are also excluded here. ** Correlation significantly different from zero at the 1 percent level (two-tailed). (a) Difference in the means of corresponding director types between Old and New Economy subsamples significantly different from zero at the 10 percent level (two-tailed). (b) Difference in the means (medians) of corresponding director types between New Economy and New-Old subsamples significantly different from zero at the 10 percent (5 percent) level (two-tailed). (c) Difference in the means (medians) of corresponding director types between Old and New-Old Economy subsamples significantly different from zero at the 10 percent level (two-tailed). TABLE 10 Determinants of Time to Maturity, Leverage, Pay Ratio, and PPS Panel A: Basic Model Time to Maturity (1) (2) Intercept -9.51 *** -14.60 *** New Economy 21.20 *** 21.58 *** New-Old Economy 22.12 *** 26.51 *** ln(Sales) 0.56 0.06 ln(Sales) x New Economy -1.99 ** -1.11 ln(Sales) x New-Old Economy -0.28 -0.14 ln(TotalAssets) 0.26 1.05 * ln(TotalAssets) 1.02 0.11 x New Economy ln(Total Assets) -0.84 -1.23 x New-Old Economy Volatility -0.37 0.86 Employee Growth (t-1) -0.49 -1.44 ** Book/Market Value 0.01 -0.03 R&D/Sales -1.79 ** -1.37 Dividend Dummy -1.17 * -1.97 *** Company Leverage 6.36 *** 7.87 *** CEO on RC 0.95 *** Ind. Consultant(s) on RC -1.08 *** Other Execs on RC 0.15 No. of External Stakeholders -0.08 Total External Shareholdings 0.39 Total Inside Shareholdings 2.59 *** CEO Dummy -0.91 *** Sample Size 212 212 [R.sup.2] 0.29 0.40 F-statistic 8.05 12.12 Leverage (1) (2) Intercept -0.06 -0.04 New Economy -0.09 -0.04 New-Old Economy 0.08 -0.05 ln(Sales) 0.01 0.01 ln(Sales) x New Economy 0.01 0.02 ln(Sales) x New-Old Economy -0-01 -0.01 ln(TotalAssets) -0.01 0.01 ln(TotalAssets) -0.01 -0.01 x New Economy ln(Total Assets) 0.01 0.01 x New-Old Economy Volatility -0.02 -0.01 Employee Growth (t-1) 0.01 0.01 Book/Market Value -0.01 -0.01 R&D/Sales 0.03 ** 0.03 ** Dividend Dummy -0.02 -0.04 ** Company Leverage 0.03 0.05 CEO on RC -0.01 Ind. Consultant(s) on RC -0.02 ** Other Execs on RC -0.02 ** No. of External Stakeholders -0.01 Total External Shareholdings -0.02 Total Inside Shareholdings 0.02 CEO Dummy 0.01 Sample Size 291 291 [R.sup.2] 0.04 0.09 F-statistic 14.38 3.20 Pay Ratio (1) (2) Intercept 1.58 1.15 New Economy -3.61 -4.08 New-Old Economy -9.57 ** -11.00 * ln(Sales) -0.42 -0.10 ln(Sales) x New Economy 1.83 ** 1.42 * ln(Sales) x New-Old Economy -1.57 *** -2.05 *** ln(TotalAssets) 0.46 0.13 ln(TotalAssets) -1.59 * -1.15 x New Economy ln(Total Assets) 2.20 *** 2.77 * x New-Old Economy Volatility -0.85 -1.02 Employee Growth (t-1) 1.37 * 1.44 * Book/Market Value 0.46 0.57 * R&D/Sales 4.46 *** 4.23 *** Dividend Dummy -1.45 -1.35 Company Leverage -1.48 -1.08 CEO on RC -0.42 Ind. Consultant(s) on RC 0.37 Other Execs on RC 0.17 No. of External Stakeholders -0.01 Total External Shareholdings -0.06 Total Inside Shareholdings 1.19 CEO Dummy 0.96 * Sample Size 291 291 [R.sup.2] 0.21 0.23 F-statistic 22.98 9.61 PPS (1) (2) Intercept 0.32 ** 0.04 New Economy 0.23 0.12 New-Old Economy -0.04 -0.11 ln(Sales) -0.02 0.01 ln(Sales) x New Economy 0.02 -0.01 ln(Sales) x New-Old Economy 0.03 -0.01 ln(TotalAssets) 0.01 -0.01 ln(TotalAssets) -0.03 0.01 x New Economy ln(Total Assets) -0.03 0.02 x New-Old Economy Volatility -0.04 -0.03 Employee Growth (t-1) 0.05 0.02 Book/Market Value -0.01 0.01 R&D/Sales 0.03 0.01 Dividend Dummy 0.02 -0.01 Company Leverage -0.10 -0.01 CEO on RC -0.01 Ind. Consultant(s) on RC 0.01 Other Execs on RC -0.01 No. of External Stakeholders -0.01 Total External Shareholdings 0.01 Total Inside Shareholdings 0.29 *** CEO Dummy 0.04 ** Sample Size 291 291 [R.sup.2] 0.12 0.31 F-statistic 5.10 14.50 Panel B: (Partial) Regression Results Controlling for Marginal Sector Effects Time to Maturity Leverage (2) (2) Model CEO on RC 0.50 -0.01 (3) CEO on RC x -1.38 0.03 ** New Economy CEO on RC x 2.76 *** -0.01 New-Old Economy Model Ind. Consultant(s) -1.48 *** -0.03 * (4) on RC Ind. Consultant(s) -0.90 -0.01 on RC x New Economy Ind. Consultant(s) on RC x 1.47 ** 0.04 * New-Old Economy Model Other Execs on RC -0.66 -0.01 ** (5) Other Execs on RC x 1.98 ** -0.01 New Economy Other Execs on RC x 1.55 -0.02 New-Old Economy Model No. of External Stakeholders -0.02 -0.01 ** (6) No. of External Stakeholders x -0.20 0.01 ** New Economy No. of External Stakeholders x -0.15 0.01 * New-Old Economy Total External Shareholdings 0.82 0.04 Total External Shareholdings x -0.15 -0.10 New Economy Total External Shareholdings x -6.13 0.01 New-Old Economy Model Total Internal Shareholdings 4.64 ** -0.04 ** (7) Total Inside Shareholdings x -4.44 0.21 ** New Economy Total Inside Shareholdings x -1.62 0.02 New-Old Economy Model CEO Dummy -1.18 *** 0.01 (8) CEO Dummy x 0.92 0.01 New Economy CEO Dummy x 0.01 0.02 New-Old Economy Pay Ratio PPS (2) (2) Model CEO on RC -0.86 -0.01 (3) CEO on RC x 1.85 0.01 New Economy CEO on RC x -0.25 -0.02 New-Old Economy Model Ind. Consultant(s) 0.31 0.01 (4) on RC Ind. Consultant(s) 1.30 0.01 on RC x New Economy Ind. Consultant(s) on RC x -0.18 -0.01 New-Old Economy Model Other Execs on RC -0.49 -0.01 (5) Other Execs on RC x 2.36 ** 0.01 New Economy Other Execs on RC x 0.03 -0.02 New-Old Economy Model No. of External Stakeholders -0.20 0.01 (6) No. of External Stakeholders x 0.24 -0.01 New Economy No. of External Stakeholders x 0.35 -0.01 New-Old Economy Total External Shareholdings 4.52 -0.01 Total External Shareholdings x -10.29 0.01 New Economy Total External Shareholdings x -5.72 0.05 New-Old Economy Model Total Internal Shareholdings -1.42 0.24 * (7) Total Inside Shareholdings x 7.65 0.15 New Economy Total Inside Shareholdings x 1.76 * 0.04 New-Old Economy Model CEO Dummy 1.10 0.02 (8) CEO Dummy x 0.73 -0.04 New Economy CEO Dummy x -1.02 0.08 ** New-Old Economy The dependent variables are time to maturity, as in Table 7, leverage, pay ratio, and pay-performance sensitivity, as in Tables 8 and 9. The time-to-maturity regressions were estimated using observations on directors' portfolios of stock options. The leverage and pay ratio regressions were estimated using observations for the directors awarded stock options and/or LTIPs, with a single observation of leverage/pay ratio for each director. The regressions include sector dummies for the New and New-Old Economy samples; firm size variables uninteracted and interacted with the sector dummies: volatility, employee growth for the previous year, Book to Market, R&D over Sales, dividend dummy and company leverage variables; variables of external (stakeholders of 3 percent or more) and inside (executive directors) ownership; binary dummy variables for the CEO (coded 1 if the observation is for the CEO, and 0 otherwise), and dummy variables for aspects of the composition of the remuneration committee (RC), such as the RC membership of the CEO, other executives and independent pay consultants. Panel B presents the within-sector sensitivity (variables are interacted with the New and New-Old sector dummies). Each group of interacted variables is calculated on separate regressions, as in Table 6. *, **, *** Statistically significantly different from zero at 10 percent, 5 percent, and 1 percent, respectively (White-corrected Standard Errors).
The authors thank Winnie Chong for excellent research assistance. Stephen Young kindly helped with data collection. This paper has benefited from the comments of the participants at the EFMA 2003 Conference, the BAACGSIG 2nd International Conference on Corporate Governance, Cardiff Business School, and the First Performance and Reward Conference, Manchester Metropolitan Business School. This draft of the paper has benefited greatly from the helpful comments of the Forum Associate Editor, Christopher Ittner, and of an anonymous referee. Professor Stathopoulos gratefully acknowledges the financial support provided to him by ESRC (Award No. R42200034399) through a Ph.D. studentship.
(1) "Moneyness" is defined here as the ratio of the share price at issue to the exercise price of the option.
(2) The main developments are documented in Inland Revenue (1988), The Cadbury Report (1992), The Greenbury Report (1995), Hampel Report (1998), The Combined Code (Part of the London Stock Exchange Listing Requirements) (1998), Hermes Statement on Corporate Governance and Voting Policy (1998), Internal Control: Guidance for Directors on the Combined Code (Turnbull Report) (1999), Company Law Reform (Green Paper) (1999), and Financial Services Market Act (2000).
(3) A further scheme, the Enterprise Management Incentive (EMI), was introduced after the end of our sample period for smaller quoted and unquoted companies in the U.K. Finance Act 2000. The tax treatment of options under the EMI discourages the granting of options in-the-money as income tax is payable at exercise equal to the spread between the fair market value (the share price) on the grant date and the exercise price (RM2 Partnership 2003).
(4) The Financial Accounting Standards Board (FASB) is considering amending U.S. accounting rules along International Accounting Standards Board (IASB) lines. We are grateful to the anonymous referee for pointing this out.
(5) Murphy's (2002) insight may be especially relevant to the New Economy sample because most of these firms were generating negative free cash flows during the period in question.
(6) We suspect that executives seek to hide rent transfers by either granting themselves or by persuading others to grant them options of greater maturity.
(7) We control for growth prospects because one would expect firms with high-growth opportunities to offer longer times to maturity.
(8) Focusing on the calendar year 1999, we selected the annual report for which most of the reporting period occurred during the calendar year 1999; e.g., for companies with a financial year-end in September, we selected the annual report for the financial year 1998/1999, and for those with an April year-end, we chose financial year 1999/2000.
(9) techMARK is a market within the official market of the London Stock Exchange. It went live on November 4, 1999 with over 190 U.K. and international companies from a wide range of FTSE industrial sectors, whose success depends on technological innovation. The market is open to innovative technology companies with a primary or dual primary listing in London, irrespective of their size, industry or location. The purpose of techMARK is to create a new way for growing technology companies to access capital to finance expansion.
(l0) The New-Old Economy companies (companies newly issued in the London Stock Exchange) have fewer than 48 observations; for these companies, we use all available monthly observations in Datastream
(11) We are grateful to the anonymous referee for suggesting this measure.
(12) Delta is a measure of the sensitivity of the option value to small changes in the share price (see the Appendix).
(13) The LTIP delta is a measure of change in value of a LTIP share for a small change in the share price, lf we assume that the directors will remain employed in the company for a long period (so that all time-restrictions on LTIP shares can lapse), then the LTIP delta for LTIPs that are not contingent on performance is 1. If the LTIP is performance contingent, then the delta is close to 0 if there is little chance of vesting, close to 1 if it is more likely to become vested, and greater than 1 if a slight change in share price has a large impact on the likelihood of vesting.
(14) We do not report the full regression results of these separate models; the results are available from the authors on request.
(15) The table shows the average values of the options received for all eligible directors. Note that some directors eligible to receive stock options may not have received any.
(16) We excluded observations with zero long-term pay from the long-term pay regressions. For the sake of completeness, we reestimated the regressions as Tobit models with the zeros included; the results were not materially different.
(17) Winsorizing the dependent variable at the 95th percentile reduces the coefficient estimate of the New Economy dummy to 8.8 (while increasing the [R.sup.2] to 35 percent). It also reduces the coefficient of the New-Old Economy dummy. The results of the winsorized regression are not reported here but are available from the authors on request.
(18) However, winsorizing the dependent variable at 95 percent renders all the uninteracted and interacted sales and asset variables statistically significant.
(19) Earlier specifications included further governance variables and characteristics of directors, including an individual directors' shareholdings, and a dummy variable discriminating between the CEO and other executives. However, none of these variables were statistically significant.
(20) Other measures of directors' characteristics (age, shareholdings, and years in post and with the company) and governance (the numbers of directors and the proportion of outside, nonexecutive directors on the board) were included in earlier specifications; they were subsequently excluded due to their lack of statistical significance.
(21) We included observations with zeros for leverage, pay ratio, and PPS. When we reestimated the regressions as Tobit models, the results were not materially different.
(22) Earlier specifications also included other measures of firm characteristics, shown in Table 1, but these were found to be insignificant and subsequently excluded.
(23) The results of this regression are not shown in Table 10, but are available from the authors on request.
Anderson, M., R. Banker, and S. Ravindran. 2000. Executive compensation in the information technology industry. Management Science 46: 530-547.
Beatty, R. P., and E. J. Zajac. 1994. Managerial incentives, monitoring, and risk bearing: A study of executive compensation, ownership, and board structure in initial public offerings. Administrative Science Quarterly 39 (2): 313-335.
Bebchuk, L. A., J. M. Fried, and D. I. Walker. 2002. Managerial power and rent extraction in the design of executive compensation. The University of Chicago Law Review 69:751-846.
Black, E, and M. Scholes. 1973. The pricing of options and corporate liabilities. Journal of Political Economy 27: 637-654.
Cadbury Committee. 1992. Report of the Committee on the Financial Aspects of Corporate Governance. London, U.K.: Gee & Co. Publishing Ltd.
Combined Code. 2000. The Combined Code: Principles of Good Governance and Code of Best Practice. London, U.K.: The Committee on Corporate Governance and Gee Publishing Ltd.
Company Law Reform. 1999. Companies Act 1999. London, U.K.: The Stationery Office Limited.
Conyon, M. J. 1994. Corporate governance changes in U.K. companies between 1988 and 1993. Corporate Governance: An International Review 2 (2): 87-99.
--, and C. Mallin. 1997. Director's Share Options, Performance Criteria and Disclosure: Compliance with the Greenbury Report. London, U.K.: Research Board, The Institute of Chartered Accountants in England & Wales.
--, and S. I. Peck. 1998. Board control, remuneration committees and top management compensation. Academy of Management Journal 41 : 146-157.
--, and K. J. Murphy. 2000. The prince and the pauper? CEO pay in the United States and the United Kingdom. The Economic Journal 110 (November): F640-F671.
--, S. I. Peck, L. Read, and G. V. Sadler. 2000. The structure of executive compensation contracts: U.K. evidence. Long Range Planning 33: 478-503.
--, --, and --. 2001. Performance pay and corporate executive compensation contracts: U.K. evidence. European Management Journal 19 (1): 73-82.
--, and G. V. Sadler. 2001. CEO compensation, option incentives, and information disclosure. Review of Financial Economics 10:251-277.
Core, J., and W. Guay. 1999. The use of equity grants to manage optimal equity incentive levels. Journal of Accounting and Economics 28: 151-184.
Ezzamel, M., and R. Watson. 1998. Market comparison, earnings and the bidding-up of executive cash compensation: Evidence from the United Kingdom. Academy of Management Journal 41 (2): 221-233.
Financial Services Authority (FSA). 2000. Financial Services and Markets Act 2000. London, U.K.: The Stationery Office Limited.
Greenbury Committee. 1995. Report on Directors' Pay. London, U.K.: Gee & Co. Publishing Ltd.
Hall, B. J., and J. B. Liebman. 1998. Are CEOs really paid like bureaucrats? Quarterly Journal of Economics 113 (3): 653-691.
Hampel Report. 1998. Committee on Corporate Governance: Final Report. London, U.K.: Gee & Co. Publishing Ltd.
Hermes Statement. 1988. Hermes Statement on Corporate Governance and Voting Policy. Hermes Pension Management, Ltd.
International Accounting Standards Board (IASB). 2003. Standards. Available at: http://www.iasb.org.uk/ standards/index.asp.
Inland Revenue. 1988. Taxes Act 1988. London, U.K.: The Stationery Office Limited.
Ittner, C. D., R. A. Lambert, and D. F. Larcker. 2003. The structure and performance consequences of equity grants to employees of new economy firms. Journal of Accounting and Economics 34: 89-127.
Lambert, R. A. 2001. Contracting theory and accounting. Journal of Accounting and Economics 32: 1-83.
Ljungqvist, A., and W. J. Willhelm. 2003. IPO pricing in the dot-com bubble. Journal of Finance 58 (2): 723-752. Main, B. G. M. 1993. Pay in the boardroom: Practices and procedures. Personnel Review 22 (7): 3-14.
--, and J. Johnston. 1993. Remuneration committees and corporate governance. Accounting and Business Research 23: 351-362.
--, A. Bruce, and T. Buck. 1996. Total board remuneration and company performance. The Economic Journal 106: 1627-1644.
Mangel, R., and H. Singh. 1993. Ownership structure, board relationships and CEO compensation in large USA corporations. Accounting and Business Research 23 (91A): 339-350.
Murphy, K. J. 1999. Executive compensation. In Handbook of Labor Economics, 3, edited by O. Ashenfelter, and D. Card. Amsterdam, The Netherlands: North-Holland Publishing Co.
--. 2002. Explaining executive compensation: Managerial power versus the perceived cost of stock options. The University of Chicago Law Review 69: 846-869.
--. 2003. Stock-based pay in new economy firms. Journal of Accounting and Economics 34: 129-147.
RM2 Partnership. 2003. Share options. Available at: http://www.rm2.co.uk/share%20options.htm.
Tosi, H., and L. R. Gomez-Mejia. 1989. The decoupling of CEO pay and performance: An agency theory perspective. Administrative Science Quarterly 34 (2): 169-189.
Turnbull Report. 1999. Internal Control: Guidance for Directors on the Combined Code. Institute of Chartered Accountants in England and Wales (ICAEW).
Konstantinos Stathopoulos, Susanne Espenlaub, and Martin Walker
University of Manchester
|Printer friendly Cite/link Email Feedback|
|Author:||Stathopoulos, Konstantinos; Espenlaub, Susanne; Walker, Martin|
|Publication:||Journal of Management Accounting Research|
|Date:||Jan 1, 2004|
|Previous Article:||Compensation committees and CEO compensation incentives in U.S. entrepreneurial firms.|
|Next Article:||Aggregation and measurement errors in performance evaluation.|