The significance of gender in explaining senior executive pay variations: an exploratory study.
Our research presents two models, one which analyzes the determinants of annual executive pay and a second which examines the determinants of total executive pay. The latter includes components of long-term compensation in addition to annual pay. Within each model, we provide a number of specifications of the earnings equations. Our results show that various corporate and executive attributes influence both executives' annual compensation and their total compensation. Gender is found to be a statistically significant independent variable in our estimation of total executive pay but not in our estimation of annual pay.
In the following section we present background material for this study. We discuss previous research concerning the extent and sources of the gender pay gap, the role of equal pay legislation on the pay gap, and the determinants of chief executive officer salaries. We conclude the background section by describing our model. Next we discuss variables used in the model and data included in our empirical analysis. We then proceed to a discussion of our empirical analysis. We end with a discussion of the limitations and conclusions of our study.
Extent and Sources of the Gender Pay Gap
Research efforts over more than three decades have attempted to disentangle factors that account for the gender pay gap and, more recently, its decline. Blau and Kahn (1997) presented Bureau of Labor Statistics data showing that the female-to-male ratio of median weekly earnings of full-time workers, which was 64.2 percent in both 1967 and 1979, rose to 74 percent by 1991. According to Blau and Kahn, two important factors in explaining the pay gap and its decline are occupational distributions of men and women and differences in their age-experience profiles. With regard to differences in the occupational choices of men and women, Groshen (1991) examined five specific industries and found that the proportion of the pay gap due to occupational differences within these industries ranged from 11 to 26 percent. In our study, occupation is essentially held constant since we are examining a single occupation, corporate executives.
The low representation of women among executives in large corporations raises the issue of the "glass ceiling." Blau et al. (1998) reported that in 1972 almost 12 percent of all men were classified in executive, administrative, and managerial occupations, while only 4.6 percent of women were in such occupations. By 1995 these percentages had increased to 14.6 percent of men and 12.8 percent of women. Despite these remarkable gains for women in management, Daily et al. (1999) noted that there were only two female CEOs among Fortune 500 companies in 1987 and still only two in 1996. By the end of 1999 there were three.
In addition to occupational distributions, differences in experience levels of men and women have figured prominently in explanations of the gender pay gap. In a longitudinal study of gender and pay, Blau and Kahn (1997) found that women had less full-time experience; that deficiency in experience explained about one third of the pay gap in 1988. Differences in union membership, occupation, and industry combined to account for 28 percent of the gap. Blau and Kahn found, additionally, that the narrowing of the experience gap (from 7.5 years in 1979 to 4.6 years in 1988) accounted for most of the decline in the pay gap over this period.
Equal Pay Legislation
Another source of the declining gender pay gap could be the effect of legislation passed in the 1960s. Nobile (1996) provides a thorough discussion of the relevant laws. The
Equal Pay Act was passed in 2963 as an amendment to the Fair Labor Standard, Act. Flynn (1999) emphasizes the narrowness of this law. The law prohibits employers from paying men and women differently for the same work strictly on the basis of gender. Pay differences can be justified, however, for reasons such as seniority and the quality and quantity of work performed.
The second important piece of legislation is Title VII of the Civil Rights Act of 1964 which prohibits discrimination on the basis of race, color, religion, national origin, or sex. Interestingly, as Blankenship (1993) noted, discrimination on the basis of sex was added late by those hoping that the addition would prevent the law from passing. This law applies not just to discrimination in pay, but also to discrimination in hiring, promoting, and firing workers. As Gerhart and Milkovich (1992) pointed out, a company's promotion system itself may be the source of unequal pay.
The concept of comparable worth--equal pay for jobs requiring equal effort, preparation, and responsibility--is another approach suggested for eliminating the gender pay gap in light of differences in the occupational choices of men and women. Flynn (1999) stated that comparable worth is recognized neither under the Equal Pay Act nor under the Civil Rights Act. Moreover, according to both Nobile (1996) and Pinzler and Ellis (1989), the courts have not reacted favorably to comparable worth claims. Researchers have also questioned whether the Equal Pay Amendment has been enforced adequately to effectively reduce the pay gap (Flynn, 1999).
Determinants of Chief Executive Officer Compensation
There have been a number of studies identifying the determinants of chief executive officer (CEO) compensation. Although we do not include CEOs in our sample of executives, this body of literature provides useful background for this study. The results of studies on determinants of CEO compensation have been mixed in showing which specific factors were influential in determining the pay of chief executive officers, but they generally have shown that corporate performance and corporate size are determinants of CEO pay. Consequently, it is important to consider the influence that company performance and size have in explaining pay levels among other top executives.
In research on corporate performance and CEO compensation, the performance measures studied include return on assets, return on equity, stockholders' return, market return, and profits. Antle and Smith (1986) found a positive association between executive compensation and return on assets, even when controlling for stock returns. Ely (1991) studied the relation between CEO compensation and research and development expenditures. She found that including long-term compensation in addition to cash compensation increased the association between CEO compensation and stockholder returns. In addition, she reported that return on equity was related to total CEO compensation. Mayer-Sommer and Bedingfield (1989) noted that average executive compensation was related to average sales and market return variables. Deckop (1988) found that CEO compensation was positively related to corporate profit but not sales. This effect varied by industry. Deckop concluded that the compensation of CEOs varied directly with their firm's profits as a proportion of sales. In addition, Hall and Liebman (1998) noted that stock options had increased the sensitivity of CEO pay to company performance.
Researchers have analyzed the impact of corporate-size factors on CEO pay, because these variables act as surrogates for the CEO's level of responsibility or job complexity. The larger the company, the more complex is the CEO's job and the greater is the CEO's responsibility. For example, Foster (1981), theorizing that firm size was related to job responsibility, found that a combined measure of size accounted for 77 percent of the variation in CEO cash compensation. Bowlin (1998) tested the relation between CEO compensation and firm size in the defense industry. He concluded that, relative to non-defense industry executives, defense industry CEOs were overpaid for size measured as market value. Veliyath et al. (1994) found that firm size, measured as sales, was among the most significant influences on executive compensation, even controlling for corporate strategy and risk.
Since there were only two female CEOs in Standard and Poors top 500 companies in 1997, we chose to expand our study to include the toppaid executives in a company. In a sense, the five top-paid officers make up the organization's top management team (TMT), though researchers define TMT in various ways. Finkeistein and Hambrick (1990) defined TMT as all corporate officers who are also board members. Their study showed an average TMT size of 3.5 officers. A 1997 study by Hoffman et al. included "inner" and "outer" TMTs. Inner TMTs are defined as in Finkeistein and Hambrick (1990), while outer TMTs include members of executive management with the title of vice president or above. Though Finkeistein and Hambrick (1990) and Hoffman et al. (1997) looked at corporate performance, neither study considers compensation of the TMT. In our study, however, we focus on compensation and extrapolate that the compensation of all members of the TMT will be in line with that of the CEO in that corporate performance and corpora te size will be determinants of TMT pay.
We use data on top executives from both companies with women among the five top-paid executives and companies without top-paid women executives to investigate whether gender itself, apart from other factors, explains variations in pay. We seek to determine whether gender is a significant determinant of pay among top executives.
Based on the above literature review and discussion, we set forth the following hypothesis and apply it to both the case of annual executive compensation and total (annual plus long-term components) executive compensation:
Differences in executive pay between women and men from both companies with and without top executive women can be explained by variations in company characteristics, executive attributes, and gender.
The model can be expressed as follows:
EC = [[alpha].sub.0] + [X.sub.1][B.sub.1] + [X.sub.2][B.sub.2] + [delta]F (1)
EC is executive compensation; [X.sub.1] is a matrix of company characteristics including size, performance, company pay philosophy, and industrial sector variables; [X.sub.2], is a matrix of executive attribute variables such as age and experience; and F is a dummy variable indicating whether the executive is a female or male. Variables for company characteristics and executive attributes are included in the model to control for their influence on executive pay. The specific variables included in the model are described in the next section.
VARIABLES AND DATA
Samples of Executives
We limited our research to the pay of the top-paid executives in a company. The Securities and Exchange Commission (SEC) requires that companies publish the compensation of the five top-paid executives in the proxy statement. We examined these proxy statements for the Standard and Poors 500 companies through the Security and Exchange Commission (SEC) EDGAR database for 1997 and found 54 women ranking among the top-paid executives of these large firms. We chose 49 of these women, excluding all CEOs (highest paid executives) as well as three other women with incomplete records.
In addition to the 49 women in our study, two groups of top-paid executive men were included in the data set. First, we added all top-paid men (excluding CEOs) in the companies with top-paid executive women. We added executive men from companies that did not have top-paid executive women. Companies without top-paid executive women were selected by matching them to companies with top-paid executive women based on industry (SIC code) and sales (a proxy for size). All top-paid executive men from these companies, except for CEOs, were included. We examined historical data to verify that these companies did not include women among the highest paid executives in 1992 or 1995. Since there are no top-paid women executives in these companies, we assume that the glass ceiling still exists. Including men from these companies thus allows us to test whether or not the presence of a glass ceiling influences executive pay. In the following discussions, these executive men and women are referred to as "top-paid" or "senior" executives.
Variables in the Model
Executive Pay. Annual executive pay is defined as annual compensation (base pay plus bonuses and other annual compensation) reported in the company's proxy statement, as required by the SEC. Many previous studies of executive pay determinants (including studies of CEO pay determinants) have used annual pay as the dependent variable in the model; we continue that practice for the first part of our study of the top-paid executives. We develop a model in which we estimate the determinants of annual pay so that we can compare our results to those of earlier studies such as Catalyst's.
Total executive pay is defined as annual pay plus the forecasted value of executive stock options awarded during the year and other long-term pay as reported in the company's proxy statement. Although the SEC requires that long-term compensation information be included in the proxy statements, there is inconsistency among companies in how it is estimated. The SEC permits companies to choose the method by which they forecast the value of stock option awards. Most companies use the Black-Scholes model and forecast the value of the options at a zero, five, or ten percent return. When multiple rates of return were reported, we chose the value using a five percent rate.
Company Performance. (1) As discussed above, previous studies have indicated that CEO pay is correlated with company performance. Thus, we include two measures of company performance in our study. In addition to operating income, we include a return on assets measure (operating income divided by total assets) as an indicator of performance. Since operating income could also reflect the size of the company (the larger the company, the higher the income), return on assets normalizes the income variable for size. These performance variables are tested in our study for their influence on senior executive pay. Data for these variables come from Compustat PCPlus.
Company Size. As indicated earlier, past studies have reported that company size is a determinant of CEO pay. We include a company-size variable in our model, since we assume that CEO pay and the pay of other top executives are similarly influenced. Although company size is clearly a company-level variable, it might alternatively be considered as an indicator of an executive's level of responsibility--the larger the company, the greater the executive's responsibility. Among the variables used in past studies to represent company size are dollar value of sales, year-end market value, and book value of total assets. All of these measures are used in our study. Again, data for the company-size variables were obtained from Compustat PCPlus.
Corporate Pay Philosophy. We introduce the CEO's pay as an independent variable to represent the company's pay-scale philosophy. An individual in one company may receive higher pay than a similar individual in another company because the company itself is a higher-paying company, independent of its size or performance. CEO pay is used to capture this effect. The CEO would be the highest paid executive in the company and his or her pay would set the scale of pay for the other members of the TMT.
Our approach is consistent with Mahoney (1979) who showed that compensation across two hierarchical levels is proportional. Both in a field study of a Canadian organization and in a lab experiment, he found that differences in compensation between hierarchical levels ranged from 30 to 40 percent. This suggests a relation among pay levels or, in other words, a corporate pay philosophy. We use annual CEO pay in our annual pay model and total CEO pay in our total pay model. CEO pay was obtained from the company's proxy statement found in the SEC's EDGAR database.
Industry. In order to account for the industry in which executives are employed, we identify each firm as being in the manufacturing sector or not. This designation is determined by the primary Standard Industrial Classification (SIC) code reported by each firm.
Presence of TMT Women. We have purposely included men from companies without top-paid executive women as well as men from the same companies as the executive women in our study. This approach allows us to determine whether variations in compensation can be explained by the presence of women in the TMT. In other words, we can see whether the absence of a glass ceiling affects compensation. A dummy variable is used to indicate whether the executive is from a company in which women are present among the top-paid executives.
Line Versus Staff. (2) A number of factors can be viewed as relating to an executive's responsibility and, hence, compensation. One such factor is whether the executive is in a line (e.g., operations) position or in a staff (e.g., legal or financial) position. A common belief is that if an individual wants to make it to top management, the individual needs to take on the responsibility inherent in operations management. A line or staff determination was made by reviewing job titles and descriptions contained in proxy statements. An alternative indicator of an executive's responsibility is whether or not the executive is on the board of directors of the company. Since we determined that virtually all directors were in line positions, we did not include this variable in our regression results.
Age and Experience. Some prior research on gender and executive pay indicates that pay is influenced by an individual's age, educational level, years of work experience, and tenure with the firm (Blau et al., 1998; Rajagopalan and Prescott, 1990; Reskin and Ross, 1992). On the other hand, Harris and Helfat (1997) looked at annual compensation of new CEOs and found that age and years on the board of directors did not explain variations in compensation. In this study, we consider both age and experience as alternative (since they are highly correlated) proxies for an executive's level of responsibility within the firm. Information on age and experience was drawn from Standard and Poor's Register of Corporations, Directors and Executives 2000, Who's Who in Finance and Industry 2000-2001, and company proxy statements found on the EDGAR website. We used age as a proxy for general human capital and years of experience with the company as a proxy for specific human capital. A notable limitation of the data sources i s that they provided age and experience information on only about half of the executives in our sample.
Gender. The final variable, and the one of primary interest for this research, is the gender of the executive. Given similar company characteristics and executive attributes for two individuals, we would expect to see no difference in their compensation. Gender of the executive was determined through reviewing the proxy statement for female names and pronouns. For those executives for whom the name could be either male or female, a telephone call was made to the company to determine the gender of the executive.
Based on our model (Equation 1) and the discussions of variables, we expect signs on variables to show that executive pay is positively influenced by the size and performance of the company. We further expect that top-paid executives in companies with high CEO pay will be paid higher than those in companies with low CEO pay. Executives with greater job responsibility are expected to be paid more than those with less responsibility. Finally, we set forth the following hypotheses regarding gender:
H1: Controlling for company characteristics and executive attributes, the annual compensation of female executives will be equal to that of male executives.
H2: Controlling for company characteristics and executive attributes, the total compensation of female executives will be equal to that of male executives.
ANALYSIS AND DISCUSSION
Table 1 reports mean values for the dependent variables used in our two models. Mean annual executive pay for 1997 was about $1.4 million, while the mean total pay (including long-term pay elements) was almost three times greater at just over $3.8 million per year. Values of company performance variables (operating income and return on assets) are also shown in Table 1. Company size variables are measures of assets, market value, and sales. CEO pay, our measure of company pay scale philosophy, is reported as $2.4 million for annual CEO pay and almost $12 million per year when long-term earnings are included.
The variable indicating manufacturing firm takes on a value of one for individuals in manufacturing firms and zero for others. Just over 52 percent of individuals in our sample are employed in the manufacturing sector. The variable for the presence of women takes on a value of one for executives from firms with women among the five top-paid executives and zero for those from firms without women among the five top-paid executives. Just under 52 percent of the individuals in the study are employed in firms that have a woman among the five top-paid executives.
The variable for line equals one if the executive is in a line position, which is the case for over 65 percent of executives in this study. The variable "Female" takes on a value of one if the executive is a woman and zero if a man. Just under 15 percent of the executives in our study are women. Human capital variables, specifically age and experience, are omitted from the regressions shown in Tables 2 and 3; results of regressions in which they are included are discussed later.
For variables expressed in dollars, the natural log of the variable is used in the regression analysis. This log-linear form is consistent with an earnings function that assumes a non-linear impact of a continuous independent variable on earnings. Variables which may take on negative values, such as operating income, are modified by adding the number that will make the observation for that variable positive. For example, the lowest operating income for a company was a loss of -$937.8 million; therefore, we scaled all operating income values up by $937.8 million.
Tables 2 and 3 present the results of five regressions run separately for annual executive compensation and then total executive compensation. Only one measure of size or performance could be used in a regression equation at a time because of multi-collinearity among size and performance variables. The purpose of each regression is to determine factors that account for variations in executive pay. In each regression analysis, the natural log of compensation is the dependent variable. Continuous independent variables include return on assets and the natural logs of assets, market value, operating income, sales and CEO pay. Dummy variables are included to represent whether an individual is employed by a manufacturing firm, whether women are represented among the five top-paid executives of the individual's firm, whether the executive is in a line position, and whether the executive is a woman.
Annual Compensation. Table 2 indicates the results of two-tailed tests of the statistical significance of the coefficients of all variables as determinants of annual compensation. Adjusted R-squared statistics indicate that between 45 and 57 percent of the variation in annual compensation is explained by variations in the set of independent variables. All company size and performance variables except for return on assets are positive and significantly related to executive pay at the .01 level. Being in a larger and higher-performing company increases annual pay, ceteris paribus. Being in a company with higher CEO pay is associated with higher pay among other top-paid executives.
When statistically significant, the sign on the variable representing manufacturing firm is negative; thus, those in manufacturing firms tend to be paid less, ceteris paribus. Whether or not an executive is in a company with a woman among the five top-paid executives has no bearing on annual executive compensation. Individuals with more responsibility, as identified by being in a line position, are paid more than those in a staff position. Finally, and most importantly, being a woman is not associated with lower annual compensation, once company performance, company size, and other control factors are taken into account. We cannot reject hypothesis H1, that, ceteris paribus, top-paid executive women receive the same annual pay as top-paid executive men.
Total Compensation. The results of regression equations in which total compensation (annual plus long-term) is used as the dependent variable appear in Table 3. The independent variables in the five equations are identical to those shown in Table 2 for annual pay. The variables in each equation explain between 57 and 61 percent of the variation in total compensation. Company size and performance measures, with the exception of return on assets (Equation 4 in Table 3), are positively related to total compensation. CEO pay is also an important determinant of the total pay of other top-paid executives in the company. Being in a manufacturing firm is associated with lower total compensation. Being in a firm where women are among the five top-paid executives has no role in explaining variations in executives' total earnings, whereas being in a line position suggests higher total pay.
All of these results are similar to the finding reported in Table 2 for annual pay. The important difference in including long-term pay categories in the dependent variable is that gender is now a statistically significant determinant of variations in pay among the sample of executives. In all five equations, the sign on the variable Female is negative, indicating that top-paid women executives receive lower total pay than men do, ceteris paribus. (3) Thus, we can reject hypothesis H2 that top-paid executive women receive the same total pay as top-paid executive men.
Alternative Analyses of Level of Responsibility
Age and Experience. We noted earlier that age and experience with the company could serve as proxies for general and specific human capital. We were able to obtain such information for two non-mutually exclusive subsets of the 336 cases in our study: 187 cases with information on age and 159 with information on experience with the company. We performed tests of the differences of means of four variables (annual compensation, total compensation, age, and experience) for the group for which we had age (or experience) data and the group for which we were missing this information. Although annual and total compensation, age, and experience are all higher for those in the subsets with human capital data, the difference is statistically significant only for age.
For these subsets of cases, women are younger than men (48.5 years and 52.2 years, respectively) and the difference is statistically significant beyond the .01 level. Women also have less experience with their firms than men (13.6 years versus 16.9 years), but this difference is not statistically significant. The equations presented in Tables 2 and 3 were run for the sub-samples of executives, first including the variables age and age squared and then including the variables experience (years with the company) and experience squared. For annual pay regressions, when the coefficient on age is statistically significant (at either the .05 or .10 level) it is negative, and the coefficient on age squared is positive. In other words, annual compensation decreases at an increasing rate with respect to age. This may not be surprising considering the mean age of the executives and the fact that younger top executives may have been brought into the company at higher pay. Harris and Helfat (1997), for example, found tha t new CEOs hired from outside the company earned a pay premium. The coefficient of the gender variable, "Female," is not statistically different from zero when age and age squared are included in the regression equation. For estimates of total pay, neither age, age squared, nor gender are statistically significant determinants of pay.
With regard to an executive's experience, neither experience nor experience squared is statistically significant when included in the regression equations for annual pay. With the exception of one equation, the variable "Female" is negative and statistically significant (at the .05 or .10 level) in the regression equations for annual pay. For equations estimating total pay, neither experience, experience squared, nor gender are statistically significant. We are reluctant to generalize from these results, since we had data for only about half of the executives in our sample. It would be valuable to have this human capital information for all cases; however, published sources were incomplete.
Pay Level Ranks. Since each company reports data for the five top-paid executives, we define five pay level ranks, first (highest) through fifth (lowest). Half of the women among the top-paid executives examined in this study are concentrated at the fifth pay level rank. Though women constitute only 14.6 percent of our sample of 336 executives, they account for 30.8 percent of those who are at the lowest (fifth) pay level rank. It could be argued that executives who are the fifth highest paid individuals in their respective firms have less responsibility than the fourth highest paid and above. For this reason, we have separately examined the group of executives at the lowest of the five top pay levels and have rerun regression equations in Tables 2 and 3 for this group only. Variations in annual pay among this group of men and women are explained by CEO pay and the specific measure of corporate size or performance, with pay being positively related to each of these variables. In two of the five regressions (r egression 1 and regression 4), the coefficient on the gender variable is significantly different from zero (at the .10 level and .05 level, respectively), and in both cases the coefficient is positive. Thus, for those who are the lowest paid of the top five executives, being female is associated with higher annual pay. It should be remembered that gender did not figure in as a determinant of annual pay variation in the regressions shown in Table 2 for the full sample of executives.
Estimates of total pay for the group of executives in the fifth pay level rank reverse the conclusions of Table 3 regression results with respect to gender. Gender is not a determinant of pay variations of total pay for the individuals who are the lowest paid of the top five in their companies. There is less variation in total compensation among the seventy-eight individuals at the fifth pay level rank than among the full set of 336 executives in the sample, and this might account for our results. In any event, we would conclude that for the group that includes the lowest paid of the TMT within each company, women earn as much in total compensation as do men.
LIMITATIONS AND CONCLUSIONS
This study provides a starting point for evaluating the role of gender as a determinant of executive pay variation. We have overcome two important limitations of Catalyst's results, first, by estimating variations in total compensation and not just annual compensation and, second, by including a number of control variables. Yet several data limitations remain. First, the companies included in the sample were larger companies; our findings might not apply to smaller companies. Second, the results are based on only the top-paid executives (excluding CEOs) as reported in company proxy statements. These findings may not hold true for mid-level managers. Third, men in companies without women among the five top-paid executives came from companies that were within the same SIC as the companies with executive women. Since these companies were not selected randomly with respect to industry, generalization of our results to other industries would not be appropriate. Heavy industry firms in steel and automobile manufact uring, for example, did not have a woman as one of their five top-paid executives and thus were not represented in our samples. Finally, we were able to obtain information on human capital variables, age and experience, for only about half of the executives in our sample. When these variables were used in the sub-sample, our results were somewhat different, perhaps because of the age, experience, and earnings characteristics of this group. We conclude that it would be fruitful to explore other alternative measures of executive responsibility.
With these limitations in mind, our results indicate that variations in annual executive pay among senior women and men in the same companies can be explained by company performance, company pay-scale philosophy, company size, industrial sector, and executive responsibility but not by an executive's gender. Women who have broken through the glass ceiling and are among the highest paid executives in a company receive annual pay comparable to similarly situated men, controlling for company size, performance, pay-scale philosophy, manufacturing sector, and executive responsibility. For total pay (annual plus long-term), gender does help to explain variations in pay, given values of the control variables listed above. Our results suggest that studies of gender determinants of executive pay need to make a distinction between annual and total compensation measures and need to incorporate indicators of executive responsibility and company characteristics.
Table 1 Descriptive Statistics Variable Mean Standard N Deviation Executive Pay--Annual $1,403,767 $ 2,242,691 336 Executive Pay--Total 3,842,221 8,190,112 336 Assets (in millions) 28,174 59,012 333 Market Value (in millions) 16,100 23,331 326 Operating Income (in millions) 1,565 2,936 330 Return on Assets 10.33% 7.66% 330 Sales (in millions) $8,581 $ 11,146 336 CEO Annual Pay 2,351,384 2,411,823 336 CEO Total Pay 11,796,886 25,549 921 332 Variable Percent of Cases Manufacturing Firm: Yes 52.4% (176) No 47.6% (160) Presence of Women in Top Five-Paid Positions: Yes 51.8% (174) No 48.2% (162) Line Position: Yes 65.2% (219) No 34.8% (117) Female: Yes 14.6% (49) No 85.4% (287) Table 2 Regression Results for Annual Compensation (Coefficient (standard error)) Equation 1 Equation 2 Equation 3 Equation 4 Intercept 5.573 *** 5.000 *** 6.035 *** 5.479 *** (.477) (.503) (.487) (.537) Ln Assets .242 *** (.026) Ln Market .220 *** Value (.030) Ln Operating .285 *** Income (.030) Return on .401 Assets (.497) Ln Sales CEO Annual .397 *** .461 *** .394 *** .560 *** Pay (.037) (.036) (.037) (.036) Manufacturing .096 -.157 ** -.067 -.165 ** Firm (.067) (.065) (.064) (.076) Presence of .034 .008 .011 .027 Women (.067) (.071) (.067) (.076) Line Position .230 *** .210 *** .209 *** .224 *** (.065) (.069) (.067) (.074) Female -.131 -.118 -.123 -.145 (.095) (.101) (.098) (.108) Adjusted [R.sup.2] .560 .524 .567 .447 F 71.467 *** 60.734 *** 70.055 *** 45.236 *** N 332 325 317 329 Equation 5 Intercept 5.004 *** (.492) Ln Assets Ln Market Value Ln Operating Income Return on Assets Ln Sales .271 *** (.036) CEO Annual .437 *** Pay (.037) Manufacturing -.194 *** Firm (.064) Presence of .027 Women (.069) Line Position .210 *** (.067) Female -.130 (.098) Adjusted [R.sup.2] .533 F 64.818 *** N 335 *** significant at the .01 level ** Significant at the .05 level Table 3 Regression Results for Total Compensation (Coefficient (standard error)) Equation 1 Equation 2 Equation 3 Equation 4 Intercept 4.319 *** 4.739 *** 5.305 *** 4.663 *** (.516) (.525) (.511) (.555) Ln Assets .208 *** (.030) Ln Market .241 *** Value (.038) Ln Operating .275 *** Income (.036) Return on .426 Assets (.557) Ln Sales Ln CEO Total .528 *** .492 *** .478 *** .633 *** Pay (.036) (.040) (.039) (.035) Manufacturing .039 -.215 *** -.143 ** -.225 *** Firm (.079) (.074) (.072) (.085) Presence of .007 -.028 -.057 -.024 Women (.078) (.080) (.078) (.084) Line Position .196 *** .194 ** .189 ** .192 ** (.076) (.078) (.076) (.082) Female -.234 ** -.220 -.200 * -.253 ** (.111) (.116) (.112) (.121) Adjusted [R.sup.2] .585 .575 .612 .523 F 78.121 *** 73.350 *** 83.195 *** 60.288 *** N 328 321 313 325 Equation 5 Intercept 3.956 *** (.535) Ln Assets Ln Market Value Ln Operating Income Return on Assets Ln Sales .236 *** (.040) Ln CEO Total .551 *** Pay (.036) Manufacturing -.208 *** Firm (.073) Presence of -0.14 Women (.079) Line Position .194 ** (.077) Female -.235 *** (.113) Adjusted [R.sup.2] .568 F 73.532 *** N 331 *** significant at the .01 level ** significant at the .05 level * significant at the .10 level
(1.) We examine five company characteristics, namely, company performance, company size, corporate pay philosophy, industry, and the presence of executive women.
(2.) We identify two attributes of executive in the model. The first is whether the executive is in line or a staff position; the second is the executive's age or, alternatively, work experience.
(3.) When the variable Female is omitted from the equations shown in Table 2., the adjusted R-squared value decreases by .004 or .005 points.
Antle, R. and A. Smith. 1986. "An Empirical Investigation of the Relative Performance Evaluation of Corporate Executives." Journal of Accounting Research 24 (1): 1-39.
Blankenship, Kim. 1993. "Bringing Gender and Race In: U.S. Employment Discrimination Policy." Gender and Society 7 (2): 1993.
Blau, F. D., M. A. Ferber and A. E. Winkler. 1998. The Economics of Women, Men, and Work, 3rd Ed. Upper Saddle River, NJ: Prentice-Hall, Inc.
_____ and L. M. Kahn. 1997. "Swimming Upstream: Trends in the Gender Wage Differential in the 1980s." Journal of Labor Economics 15(1): 1-42.
Bowlin, W. F. 1998. "Executive Compensation in the U.S. Defense Industry." Journal of Cost Analysis (Fall): 87-102.
Catalyst. 1998. 1998 Census of Women Corporate Officers and Top Earners. New York, NY: Catalyst.
Daily, C. M., S. T. Certo and D. R. Dalton. 1999. "A Decade of Corporate Women: Some Progress in the Boardroom, None in the Executive Suite." Strategic Management Journal 20: 93-99.
Deckop, J. R. 1988. "Determinants of Chief Executive Officer Compensation." Industrial and Labor Relations Review 41(2): 215-226.
Ely, K. M. 1991. "Interindustry Differences in the Relation Between Compensation and Firm Performance Variables." Journal of Accounting Research 29(1): 37-58.
Finkeistein, S. and D. C. Hambrick. 1990. "Top-Management-Team Tenure and Organizational Outcomes: The Moderating Role of Managerial Discretion." Administrative Science Quarterly 35(3): 484-504.
Flynn, Gillian. 1999. "Protect Yourself from an Equal-Pay Audit." Workforce 78 (6): 144-146.
Foster, K. E. 1981. "Does Executive Pay Make Sense?" Business Horizons 24(5): 47-51.
Gerhart, B. and G. T. Milkovich. 1992. "Employee Compensation: Research and Practice." In Handbook of Industrial and Organizational Psychology, 3d ed. Ed. M. D. Dunnettee and L. M. Hough. Palo Alto, CA: Consulting Psychologists Press.
Groshen, E. L. 1991. "The Structure of the Female/Male Wage Differential." Journal of Human Resources 26 (3): 457-472.
Hall, B. J. and J. B. Liebman. 1998. "Are CEOs Really Paid Like Bureaucrats?" Quarterly Journal of Economics 113(3): 653-692.
Harris, D. and C. Helfat. 1997. "Specificity of CEO Human Capital and Compensation." Strategic Management Journal 18: 895-920.
Hoffman, J. J., R. A. Lheureux and B. T. Lamont. 1997. "The Effect of 'Inner' and 'Outer' TMT Size on the Performance of International Firms." Journal of Managerial Issues 9(1): 121-135.
Mahoney, T. 1979. "Organizational Hierarchy and Position Worth." Academy of Management Journal 22 (4): 726-737.
Mayer-Sommer, A. P. and J. P. Bedingfield. 1989. "A Reexamination of the Relative Profitability of the US Defense Industry: 1968-1977." Journal of Accounting and Public Policy 8(2): 83-119.
Nobile, Robert J. 1996. "How Discrimination Laws Affect Compensation." Compensation and Benefits Review 28 (4): 38-42.
Pinzler, Isabelle Katz and Deborah Ellis. 1989. "Wage Discrimination and Comparable Worth: A Legal Perspective." Journal of Social Issues 45 (4): 51-65.
Rajagopalan, N. and J. E. Prescott. 1990. "Determinants of Top Management Compensation: Explaining the Impact of Economic, Behavioral, and Strategic Constructs and the Moderating Effects of Industry." Journal of Management 16(3): 515-538.
Reskin, B. F. and C. E. Ross. 1992. "Jobs, Authority, and Earnings Among Managers: The Continuing Significance of Sex." Work and Occupations 19(4): 342-365.
Standard and Poor's Register of Corporations, Directors, and Executives 2000. 2000. Charlottesville, VA: McGraw-Hill.
Veliyath, R., S. P. Ferris and K. Ramaswamy. 1994. "Business Strategy and Top Management Compensation: The Mediating Effects of Employment Risks, Firm Performance and Size." Journal of Business Research 30: 149-159.
Who's Who in Finance and Industry 2000-2001. 1999. New Providence, NJ: Reed-Elsevier.
|Printer friendly Cite/link Email Feedback|
|Author:||Renner, Celia; Rives, Janet M.; Bowlin, William F.|
|Publication:||Journal of Managerial Issues|
|Date:||Sep 22, 2002|
|Previous Article:||Reexamining the relationship between action preferences and managerial risk behaviors *.|
|Next Article:||Preparing students for success in team work environments: the importance of building confidence.|