AN EXAMINATION OF THE DETERMINANTS OF EXECUTIVE COMPENSATION LEVELS IN IARGE MANUFACTURING FIRMS OVER TIME Dissertation for the Degree Of Ph. D. v MICHIGAN STATE UNIVERSITY ' CHARLES GLEN THARP [1977 ~ II III I LIBRAA Y 58 ' IIIIIII IIIIIIIIII III I II IIIIII III 61 Michigan State University This is to certify that the thesis entitled AN EXAMINATION OF THE DETERMINANTS OF EXECUTIVE COMPENSATION LEVELS IN LARGE MANUFACTURING FIRMS OVER TIME presented by Charles Glen Tharp has been accepted towards fulfillment of the requirements for Ph. D 0 degree in Labor and Industrial Relations Major professor 0-7639 I '3',“ I, my ‘3 39'223 _ AN EXAMINATION OF THE DETERMINANTS OF EXECUTIVE COMPENSATION LEVELS IN LARGE MANUFACTURING FIRMS OVER TIME BY Charles Glen Tharp The major thrust of this analysis is to examine the factors which influence the levels of pay received by our nation's highest paid corporate executives. The basic hypotheses tested are that the levels of compensation received by top corporate executives will vary according to the size of the firm they direct, cor- responding corporate profitability, and the amount of human capital possessed by the executive. The sample for which this analysis was conducted consisted of eighty of the nation's largest manufacturing firms. These firms were analyzed over the time period from 1961 to 1975, thus yielding a pooled sample of 1,200 observations. Multiple regression analysis was used to examine the relationship between vectors of variables, measuring corporate size, profitability, the executive's stock of human capital, and the level of compensation the executive received. Charles Glen Tharp As a result of this analysis it was discovered that the various components of the executive pay package are influenced by different corporate and individual characteristics. Specifically, the salary component of the pay package was found to be significantly correlated with the size of the firm managed and the individual manager's stock of human capital. When examining changes in the level of base salary it was found that the pro- fitability of the firm was the only vector which exerted a significant influence on the amountxxfthis change. Changes in the bonus component of the executive pay package were found to be positively and significantly correlated with changes in both the size of the firm and its corresponding level of profitability. The value of the stock options awarded to the executive were dis- covered to be contingent upon only the change in corpo- rate size hithe year of the stock award. The conclusions drawn from the above results are that firms utilize the various components of the execu- tive pay package to reward different types of executive performance. Changes in base salary levels are used to reward executive behavior which results in an increase in the level of corporate profitability. Changes in the amount of the stock option award are contingent upon the executive's ability to increase the scale of the firm's operations. Finally, executive bonuses are tied to Charles Glen Tharp the performance of the corporation in terms of both pro- fitability and growth in size. AN EXAMINATION OF THE DETERMINANTS OF EXECUTIVE COMPENSATION LEVELS IN LARGE MANUFACTURING FIRMS OVER TIME BY Charles Glen Tharp A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY School of Labor and Industrial Relations 1977 *4/C)¢?OCL;1 ACKNOWLEDGMENTS I would like to thank my committee--Thomas Patten, Daniel Hamermesh, Steven Director and Frederic Wickert for their guidance in this undertaking. Credit for all of the accomplishments in this study I share with them. The shortcomingscfifthis analysis, however, are mine alone. The person who deserves the most credit for the completion of this work is my wife, Ann. For without her patience and support I could never have completed my degree. ii TABLE OF CONTENTS LIST OF TABLES . . . . . . INTRODUCTION . . . . . . . Chapter I. II. III. IV. VI. DEVELOPMENT OF A MODEL OF THE PAY DETERMINATION PROCESS FOR TOP CORPORATE EXECUTIVES . A REVIEW OF THE LITERATURE CONCERNING THE RELATIONSHIP BETWEEN EXECUTIVE COMPENSATION AND CORPORATE CHARACTERISTICS . . . A REVIEW OF THE LITERATURE CONCERNING THE RELATIONSHIP BETWEEN EXECUTIVE COMPENSATION AND HUMAN CAPITAL STOCK . . . . . . A PRESENTATION OF THE SAMPLE, METHODOLOGY, MODEL, AND HYPOTHESES TO BE TESTED EXPLORED . . . . . Sample . . . . . Procedure . . . p Model . . . . . . Hypotheses . . . . STATISTICAL ANALYSES AND Analysis . . . . . SUMMARY AND CONCLUSIONS BIBLIOGRAPHY . . . . . . . APPENDICES . . . . . . . iii RESULTS AND Page iv 27 69 Table 1. 10. 11. 12. LIST OF TABLES Values of Average Partial Correlation Coefficients and Corresponding Com- puted Values Of t O O O O O O O O 0 Regression Results: Profits, Sales and Compensation . . . . . . . . . . Regression Results: Market Values, Sales, and Compensation . . . . . . . . . Executive Compensation--Senior Officer (Correlation Coefficients, r) . . . . . Executive Group Compensation (Correlation coeffiCientS' r) O O O O O I O O 0 Correlation Coefficients Between Paired Independent Variables . . . . . . . Correlation Table for Transformed Variables (Ratios to Assets) . . . .. . . . . An Analysis of the Determinants of the Base Salary Component of Executive Pay . . . An Analysis of the Determinants of the Bonus Component of Executive Pay . . . . An Analysis of the Determinants of the Stock Option Component of the Executive Pay Package . . . . . . . An Analysis of the Determinants of Changes in the Salary Component of the Executive Pay Package . . . . . . . . . . . An Analysis of the Determinants of Changes in the Bonus Component of the Executive Pay Package . . . . . . . . . . . iv Page 36 42 43 50 51 54 64 118 119 120 121 122 Table Page 13. An Analysis of the Determinants of Changes in the Stock Option Component of the Executive Pay Package . . . . . . . 123 14. An Analysis of the Significance of the Relationship Between the Salary Com- ponent of the Executive Pay Package and the Vectors of Corporate Profit- ability, Corporate Size, and the Executive's Stock of Human Capital . . . 124 15. An Analysis of the Significance of the Relationship Between the Bonus Com- ponent of the Executive Pay Package and the Vectors of Corporate Profit- ability, Corporate Size, and the Executive's Stock of Human Capital . . . 125 16. An Analysis of the Significance of the Relationship Between the Stock Option Component of the Executive Pay Package and the Vectors of Corporate Profit- ability, Corporate Size, and the Executive's Stock of Human Capital . . . 126 17. An Analysis of the Significance of the Relationship Between Changes in the Level of the Salary Component of the Executive Pay Package and the Vectors of Changes in Corporate Profitability, Changes in Corporate Size, and the Executive's Stock of Human Capital . . . 127 .18. An Analysis of the Significance of the Relationship Between Changes in the Level of the Bonus Component of the Executive Pay Package and the Vectors of Changes in Corporate Profitability, Changes in Corporate Size, and the Executive's Stock of Human Capital . . . 128 19. An Analysis of the Significance of the Relationship Between Changes in the Level of the Stock Option Component of ' the Executive Pay Package and the Vectors of Changes in Corporate Profit- ability, Changes in Corporate Size, and the Executive's Stock of Human Capital . . . . . . . . . . . . 129 V INTRODUCTION The major thrust of this study is to discover and to analyze the factors which influence the level of pay received by the top managers of the nation's largest industrial firms. Given that the top executives of mod- ern corporations are thought to have the power to guide the firm down the road of financial success or to blindly lead it to the brink of corporate ruin, it is of great interest to examine the extent to which the level of pay received by this group of managers is associated with the performance of the firm. It has become common practice for the managers of large corporations to command levels of pay which are in the six-figure range and to receive a comprehensive and imaginative array of fringe benefits covering every contingency of living from long-term disability to financing the college education of their children (Forbes, p. 3). In light of these tremendously large pay levels received by the top management group, one may be drawn to ask on what basis these compensation decisions are made. The analysis presented in this study will attempt to answer this question. There are basically two approaches in conducting a study of this nature. One approach would be to gather l information from a sample of firms via use of a survey questionnaire as to the factors which the board of dir- ectors considers in the decision-making process for executive compensation levels. Although representing a possible approach, the use of a survey questionnaire is plagued with several problems which render this method of data collection impractical for the compensation researcher. One of the major drawbacks of the survey questionnaire approach comes in gaining the cooperation of corporations in revealing information on the rather sensitive area of executive pay. Further, large corpora- tions are generally so inundated with survey question- naires from governmental agencies, trade associations, private research firms, and fund raising organizations that the likelihood of the corporation providing a prompt and thoughtful response to the inquiries of a graduate student would be greatly reduced. In the event that one were able to secure a sufficient amount of participation in a data gathering effort of this nature, the likely result of such an inquiry would provide the researcher with responses which reflect the formal statement of corporate pay policies immured in pages of some dusty policy book and which may be more representative of how pay decisions should be made rather than how they actually are made. An alternative and more viable approach to gain- ing information concerning the pay determination process for top corporate executives is to engage in a form of "policy capturing." In this approach, the researcher first deve10ps a model which is thought to be descriptive of the pay determination process and then goes about the task of collecting the relevant data to make an empirical testing of the model developed possible. Hence, the emphasis in policy capturing is not to examine what organizations say they do with respect to executive pay decisions but rather what they actually do. The methodol- ogy employed in this policy capturing effort is presented and discussed in Chapter IV of this study. At this juncture, it is appropriate to emphasize what the scope of this analysis encompasses and to iden- tify the areas of research on executive compensation which are not explored in this study. The intent of this study is to develop and test a model of the pay determina- tion process which identifies and measures the relation- ship between the levels of compensation received by top corporate executives and the relevant corporate and per- sonal characteristics specified in the model. The formal development and statement of the model utilized in this analysis is presented in Chapter I. The focus of this analysis is not to develop and test a theory of managerial motivation or to derive results which will tell organizations how to "correctly" develop and implement an executive pay structure. The importance of properly inte— grating the executive pay plan with overall corporate planning, tieing pay to individual performance toward pre- determined goals and objectives, and the importance of tax considerations in making executive pay decisions have all been prescribed by writers in the areas of organiza- tional behavior and executive compensation as constituting the proper approach to designing an executive pay system (Moore, 1968). The emphasis of the approach taken to the study of executive compensation by earlier authors is that of a normative nature, i.e., telling organizations what they should do with respect to pay determination. The approach taken in this current analysis is that of a positive nature, analyzing the pay determination process as it actually exists, rather than a prescriptive state- ment of how-executive pay decisions should be made. The research efforts to date which have examined the factors influencing executive compensation levels will be reviewed in Chapters II and III of this study. A careful and thorough critique of these earlier studies will demonstrate the efforts which have been made to model the behavior of organizations with respect to executive pay decisions as well as point out areas in which addi- tional research is needed on this topic. The value of any study rests not only with the results obtained, but also with the manner in which the analysis is conducted. Chapter IV contains a statement of the model upon which this study is based, the methodol- ogy employed to test this model, and the sample over which the analysis was conducted. It is hoped that by revealing the theoretical, methodological, and empirical bases of this study that the results reported herein will be given greater reliability and increase the ability to . generalize from the findings. The statistical analysis and the empirical results derived from this study are presented in Chapter V. The corresponding inferences and conclusions drawn from these results are presented in the final chapter, Chapter VI. The benefits to be accrued from a study of this nature are that new insights will be gained into the process utilized by large corporations in making pay dec- isions for top corporate officers. Further, the findings for this top group in management may be generalized and applied to successively lower levels of management in an attempt to identify the factors which influence the levels of pay received by the individuals in various levels of corporate management. CHAPTER I DEVELOPMENT OF A MODEL OF THE PAY DETERMINATION PROCESS FOR TOP CORPORATE EXECUTIVES In this section, I will present a discussion of the process whereby firms set the level of compensation received by the top corporate executives. This process will be examined in terms of the relevant economic and behavioral models of wage determination. The major focus of this study is not to develop a model of executive motivation with respect to financial rewards, but rather to provide a conceptual and methodological framework through which the policies of large corporations about executive pay decisions may be captured and analyzed. The end result of conducting a study of this nature will be to discover which characteristics of the corporation influence the level of pay received by top executives as well as to gain insights into the goals and objectives established and pursued by large publicly held corpora- tions. When attempting to develop and test a model of the pay determination process for top corporate officers, it is necessary to first have an understanding of the institutional aspects of this process within the modern corporation. Historically, when the typical firm was owner-managed, the task of determining which factors affected the level of compensation received by the owner- manager was quite simple. One merely looks at the level of net profits to gauge what the return to the owner- manager would be for his services. Under this type of control structure within the firm, the question of the determinants of the level of executive compensation was a moot issue. However, with the growth of the giant corporation came the phenomenon of the professionally managed firm. Typical of this structure is the use of the board of directors of the corporation to set corpor- ate policies and objectives which are to be pursued by the management of the firm. Along with this policy setting task, the directors of the corporation are man- dated the responsibility of insuring that the level of pay awarded to the executives of the corporation reflect the corporate objectives established by the board of directors. Hence, it is through the use of the board of directors, in concert with the input of the corporate compensation staff, that pay decisions for top executives are made. The aim of this study is to capture through empirical analysis which objectives the directors of the I‘ll-i" modern professionally managed corporation elect to pur— sue, and which characteristics of the firm most strongly influence the level of compensation received by the top corporate executives. With respect to this undertaking, David Belcher, a noted writer in the field of compensa- tion administration, has stated that compensation theorists face a "huge task" (Belcher, 1974, p. 17). Belcher defines this task as being "to specify the fac- tors that determine compensation, the manner in which they do so, and the relative and absolute importance of each factor" (Belcher, 1974, p. 17). Traditional economic theory of the firm states that the rational employer will follow a profit—maxindzing course of behavior. Given that the large corporation is not owner-managed, profit-maximization may not represent an accurate description of the actions of the directors of the professionally managed corporation. There is a possibility that the utility function of the directors of the corporation may contain goals and objectives other than strict profit-maximization which they may elect to pursue. Oliver Williamson in his writings on the theory of the firm points out that the modern large firm is characterized by a separation of ownership from manage- ‘ment, and further given the existence of a degree of monopoly power in the product market, traditional theories of the firm may fail to explain adequately the behavior of the individual firm and the behavior of the management within that firm. On this point Williamson notes: Where the range of behavior that is consistent with survival is narrowly bounded (because of a purely competitive market) the question of motivation is of small importance. However, some (most) firms appear to have access to advantages that bring substantial relief from extinction. Here an understanding of motiva- tion may be essential (Williamson, 1967, p. 129). In terms of Williamson's writings, motivation is expresmai in relation to the goals and objectives which the corpor- ate managers set for the firm to achieve. An understanding of the objectives of the corpor- ation is essential to the analysis of the determinants of executive compensation. Regardless of the model employed to study the determinants of executive remuneration (be it derived from the behavioral sciences or economics), the basic construct of the model is that there should be a relationship between the level of pay an individual receives and the performance of the individual as mea- sured against a specific goal or standard of performance. As was pointed out earlier in this section, the focus of this present study is to capture the policies of corpora- tions with respect to the setting of executive compensa- tion levels. Hence, the appropriate model of the pay determination process to be utilized is that of an econo- mic model. The specific model examined is that of the marginal productivity theory of wage determination. At 10 this time it is apprOpriate to point out that the deci- sion to employ the marginal productivity theory in no way denies the validity of the behavioral scientists' approach to the study of compensation. The major dif- ference between the economists' model of the pay deter- mination process and that of the behavioral scientists' is in terms of the focus of the analysis. The behavioral scientists direct their study of pay upon the motiva- tional aspects of pay and the individual worker's per- ceptions of the effort-reward linkage (Lawler, 1971, p. 119). Therefore, the appropriate variables for this level of analysis are the individual's feeling of equity, the likelihood of favorable outcomes, and the linkage between effort and performance. Economists on the other hand, direct their attention toward the economic measures of value and productivity. These variables represent the basic concepts upon which the employment transaction are built. The employer compensates the worker in accordance with the value of the individual's productive services to the output of the firm, assuming that the worker is equally productive across firms. Given that the emphasis of this present study is to examine the extent to which various economic characteristics of the firm influence the level of pay received by corporate executives, it appears more appropriate to utilize the marginal 11 productivity approach to studying pay than to employ a behavioral science approach in this policy capturing effort. The name most commonly associated with the marginal productivity theory is that of John Bates Clark (Clark, 1899). Simply stated, the marginal productivity theory of employment holds that the economically rational employer will employ labor at a wage rate which is com- mensurate with the value of that worker's contribution to the functioning of the firm.1 Expressed in the nota- tion of the theory, the rational employer will compensate the worker at a level of pay (W) which corresponds to the worker's value of marginal product (VMP). Hence, the employer will set the level of compensation so as to sat- isfy the identity: =VMP. The economic rationale for setting the level of pay such that it is equal to the value of the worker's marginal physical product is that this method of pricing inputs corresponds to the least cost means of production. To pay the worker a higher wage level would be to compensate the individual at a rate which was greater than he could command in the market, and thereby result in an unnecessary addition to the 1While the marginal productivity theory, as dev- eloped by Clark, is a theory of labor demand it may be utilized as a theory of wage determination if one ignores the problem of causality between employment and wage levels. . l2 costs of the firm. Likewise, to pay the worker a lower wage rate would result in an unnecessary amount of labor turnover as a result of other employers bidding away from the firm's employ these underpaid workers. In order to utilize the marginal productivity theory outlined above for the examination of the pay determination process for top executives, certain assump- tions about the behavior of the firm, and the board of directors who makes the pay decisions within the firm, must be made. The assumptions needed to make the marginal productivity theory applicable to the analysis of execu- tive pay are: (1) that the board of directors of the corporation, who have the responsibility for making pay decisions, behave in an economically rational manner, i.e., are utility maximizers, and (2) that the executives of the firm are rewarded in accordance with the objective; of the corporation. By making these assumptions, then the relationship between various measures of executive performance and the level of compensation received can be examined. Before one can measure the relationships between the level of executive pay received and the individual executive's contribution to achieving the goals and objectives of the firm, an identification and descrip- tion of the actual goals of the firm is necessary. Virtually all of the large firms in the United States conform to the control structure of the 13 professionally managed firm. Under this structure, the owners of the firm are the shareholders who voice their interests in the operations of the firm through the annual meeting of the corporation. It is the directors of the corporation who guide the operation of the firm's activities and thus decide which policies will be imple- mented and pursued. Therefore, given this separation of ownership from the day-to-day management of the firm, there exists the Opportunity that the goals and objec— tives which the directors of the corporation establish and actively pursue may not always represent the best interests of the shareholders. As was alluded to earlier, traditional theories of the firm assume that the objective of the firm is that of maximizing profitability. There would be no question- ing of this view of the firm if ours were a perfectly competitive economy. However, since our economy is not a purely competitive one, this relaxation of the competi- tive assumption allows a degree of latitude in that the professional managers of the corporation may elect to establish corporate policies aimed at goals other than that of strict profit-maximization. This situation has led William Baumol, for example, to conclude from his studies of the firm that the typical large corporation seeks not to maximize its profits, but rather to maxi- mize the growth of the scale of the firm's operations. 14 When referring to the growth of the firm Baumol speaks in terms of the sales revenues of the firm as being repre— sentative of its size. "The typical large corporation in the United States seeks to maximize not its profits but its total revenues which the businessman calls his sales (Baumol, 1958, p. 187). The rationale offered by Baumol as to why managers would depart from pursuing a course of behavior aimed at profit-maximization is that it is in their own best self-interest to do so: "executive sal- aries appear to be far more closely correlated with the scale of the operations of the firm than with its profit- ability" (Baumol, 1959, p. 46). Based upon the assumptnni that executives are rewarded in accordance with the objectives of the firm, one is afforded the opportunity not only to examine the determinants of the levels of compensation received by corporate executives, but also to test the growth-maximization hypothesis offered by Baumol as being characteristic of the objectives estab- lished by the large corporation. By utilizing the marginal productivity model of the wage determination process one can make specific pre- dictions as to the nature of the relationship between executive compensation levels and organizational charac- teristics. If one were to assume that the typical large corporation is a profit maximizer, as is asserted by the traditional theories of the firm, then one would be drawn 15 to predict that the level of pay received by the top corporate executives would be positively and signifi- cantly correlated with the level of profitability of the firm. Under this assumed profit maximizing course of behavior, one would expect the rational firm to set the level of pay such that W=VMP: where W represents the level of executive compensation, and VMP is measured in terms of the individual executive's contribution to the profitability of the firm. Likewise, if one were to adhere to the Baumol hypothesis that the firm is not a profit-maximizer but rather a growth-maximizer then one would expect that the rational firm would set the level of executive compensa- tion such that W=VMP, where VMP is measured in terms of the growth of the firm. Hence, given that one views the objectives of the firm to be the promotion of growth, the variables expected to bear a strong relationship to executive compensation levels would be measures of the corporation's scale rather than its profitability. In order to apply the marginal productivity theory to the determination of wage levels, one must take the assumption that the individual is equally productive across firms. Under this assumption the marginal pro- ductivity theory, which is most often utilized to determine the level of employment of an input, can be utilized to analyze the determination of wage levels. 16 Under the marginal productivity approach to the analysis of wages, one would ideally want to measure the value of the individual's direct contribution to the output of the firm. However, when examining the contributions of the individual executive, it becomes exceedingly diffi- cult, if not impossible, to identify and to quantify specific actions which constitute the performance of the executive. Therefore, in order to utilize the marginal productivity theory of wages to analyze executive com- pensation, one is forced to find proxies for the direct contributions of the executive to the functioning of the firm. Several writers in the areas of executive compen- sation and organizational behavior have suggested that one may use corporate performance to proxy the performance of the individual executive. On this point, Belcher notes that “the job contributions of top management are assumed to represent an identity between the individual and the organization"(Belcher, 1974, p. 524). This point is echoed by Kenneth E. Foster when in reference to executive pay decisions he states that "pay at this level cannot be equated to direct contributions to out- put; and the broad, diffuse nature of the tasks performed makes it difficult to evaluate and compare the require- ments of various managerial positions" (Foster, 1969, p. 80). This suggests that since one cannot clearly measure the direct contributions of the individual 17 executive to the performance of the corporation, one may utilize measures of corporate performance to proxy for the performance of the individual. Although utilizing corporate measures of performance to proxy for those of the individual executive may tend to overstate the impact of the individual executive's actions on the functioning of the firm, the use of such measures may still be acceptable given that the actions and decisions of the top executives so strongly and so ubiquitously affect the functioning of the organization. One must take great care in specifying the manner in which the characteristics of the organization relate to the performance of the executive and the demands placed upon the executive. It can be reasonably assumed that it is a more demanding task to direct the opera- tions of a large corporation than it is to manage a relatively smaller one. Thus, one would expect to find that larger firms would pay higher levels of executive compensation than would smaller firms because the execu- tive managing the larger firm would be perceived to have a greater value of marginal product, given his greater managerial responsibilities, than would his counterpart in the smaller organization. Likewise, it would be expected that the manager who can direct his organiza- tion to a higher level of profitability is performing at a higher level and contributing at a greater rate than 18 his counterpart in a less profitable corporation. In terms of the marginal productivity theory, the contribu- tion of the individual executive to the functioning of the firm may be viewed in terms of the scope of the job he holds (organizational size) and the contribution he makes to the realization of organizational goals and objectives (whether they be of a growth or profitability nature). It is expected that these different types of performance (job scope vs. the achieving of organiza- tional goals) may be reflected in the various components of the pay package. On a cross-sectional basis it is expected that the level of base salary received by the top executives of large corporations will be directly related to the size of the firm's operations, i.e., scale, and therefore the magnitude of the resources over which the executive exercises responsibility. The dir- ection of this relationship should be that the executives of larger firms will receive larger base salaries than those in relatively smaller firms. Corporate size also enters into the executive pay decisions of the firm by virtue of the manner in which organizations compare themselves in the labor market. Large corporations compete with other large corporations in the labor mar- ket to attract and retain high caliber managerial personnel. To insure their competitiveness, firms per- iodically sample the labor market via the use of 19 executive compensation surveys. The various executive compensation surveys (American Management Association, Forbes, Business Week, Hay, and Sibson) categorize the results of their analysis on the basis of sales level and the particular industry of which the firm is a mem- ber (Rock, 1972, pp. 3-30). Therefore, given that cor- porations make extensive use of compensation survey results, and also that level of firm sales is a key organizational evaluation benchmark in these surveys, one would expect executiveocompensation levels to bear a strong relationship to the scale of the firm's operations, of which sales revenues is one measure. The second type of performance which would be expected to influence the level of compensation received by top executives is the achieving of organizational goals and objectives. As was pointed out earlier, there are basically two major objectives which the directors of the corporation may elect to pursue, either maximiza- tion of profitability or growth-maximization. If the goals which the board of directors of the corporation chooses to pursue correspond to a maximization of the growth of the scale of the firm's operations, then it would follow that the level of pay received by the top corporate officers would change in accordance with changes in the scale of the firm's operations. This would indicate that the changes in the level of pay 20 received by the executive would relate to the changes in the value of his contributions to the performance of the firm, which conforms to the marginal productivity model of the wage determination process. Likewise, if the directors of the corporation elect to follow a profitability-maximization goal for the firm's operations, then one would expect that changes in the level of execu- tive compensation received would be positively and significantly related to changes in the level of profit- ability of the firm. If we assume that the total pay package received by the executive is built upon the level of base salary awarded, then there are two ways in which the amount of compensation received by the executive may vary. Firstly, if the performance of the individual is perceived to have increased he may be granted an increase in base salary. This would indicate that the determination of base salary levels is contingent upon the results of the executive as well as the magnitude of the resources which he man- ages. The second way in which the level of pay received by the executive may increase to reflect increased per- formance would be through the use of what is termed incentive compensation. For top corporate executives the most common forms of incentive compensation are cash bonuses and stock option awards. Under this type of compensation, performance toward organizational goals 21 may be rewarded by granting more cash, through the bonus or by awarding shares of stock. When conducting this type of analysis there are potential problems in using measures of changes in cor- porate growth and corporate profitability to proxy the performance of the individual executive in that the growth in the firm's scale of profitability may be signi- ficantly influenced by the state of the general economy or of the specific industry in which the firm operates. What is needed is a procedure by which to separate this broader economic influence on the performance of the firm so as to get a better measure of the contributions of the individual executive toward achieving corporate goals and objectives. In order to adjust for this gen- eral economic effect on the performance measures of the firm, I will specify the corporate performance variables as a ratio to the industry average for that variable. For example, if one were looking at the correlation between executive compensation and corporate performance in the automobile industry, the measures of the character- istics of the firm should be expressed as a ratio to the median value of those variables for the automobile industry. Such a specification of performance measures in their relative values would serve to more accurately measure the contribution of the individual by factoring out the influence of the external economy or the state 22 of the specific industry upon the performance of the individual firm. The preceding paragraphs presented an analysis of the wage determination process in terms of the mar- ginal productivity theory of wages. The emphasis of the model concerns the relationship between the performance of the individual and the level of compensation received. There is also a likelihood that the personal character- istics of the executive will influence the level of compensation received. The analysis of the relationship between personal characteristics of the worker and the level of pay received is termed the study of "human capital" in the employment transaction. Individuals invest in themselves through the accumulation of human capital based on the belief that they will reap a posi- tive return in the form of higher wages as a result of this investment. The traditional human capital variables of age, education, and experience will be examined in an attempt to discover the extent to which these personal characteristics affect the level of pay received by the top corporate executives in large U.S. firms. Within the context of the marginal productivity approach to wage determination, there are basically two ways in which the individual's stock of human capital can influence the level of pay received. Firstly, if the board of directors of the corporation feel that it 23 cannot accurately identify and measure the contribution of the individual executive to the functioning of the firm, then the amount of human capital possessed by the executive may be used as an indicator of the individual's potential to perform within the organization. In this situation, the firm would be rewarding the individual for his perceived value of marginal product rather than his actual contribution to the firm. Within a particu- lar firm there may be present enough indicators of the individual executive's performance so as to make the resorting to the use of perceived marginal product unlikely. The situation in which perceived marginal product will most likely be the basis for compensation decisions is by other firms who are competing for the executiveksemployment and corresponding productive services. These potential employers will not have access to measures of the individual's productivity by which to make employment and compensation decisions; therefore, they will be forced to rely on the use of human capital measures to gauge the potential contribu- tions which the executive may make to the functioning of their particular corporation. The second way in which personal characteristics of the executive may influence the level of pay received is if the firm is in some manner actually "consuming" the human capital of the individual and compensate him 24 accordingly. If the utility function of the board of directors is such that having highly educated execu- tives from the "best schools" is deemed to be desirable, then one would logically expect that there would be a positive correlation between level of formal education and executive pay. Similarly, one may also expect that school type (Ivy League, Big Ten, etc.) may also influ- ence the level of pay received. This would be the case if the directors felt that the firm would be viewed more favorably by stockholders or investors if the cor- poration were headed by a "Harvard man." In this example, the type of school which the executive attended would represent part of the individual‘s contribution to the functioning of the firm in that it would measure the value of the individual's marginal product toward the output of the firm termed status. If the individual executive can bring more prestige to the firm as a result of his educational background, it is only logical that the firm should set the wage level so as to reflect this contribution of the individual. Therefore, one would predict by use of the marginal productivity approach to this analysis that the level of executive pay would be positively correlated with educational level and type of school attended as a result of this "credentialism" effect. Chapter III of this study will present a review of the literature concerning the 25 relationship between human capital and executive compen- sation. The discussion presented above concerning the relationship between the scale of the firm, the profit- ability of the firm, the executives' stock of human capital, and the level of compensation received can be expressed in the form of specific hypotheses which lend themselves to empirical testing. Underlying each of these hypotheses is the assumption that the appropriate model to be employed when analyzing the determinants of executive compensation is that of the marginal produc- tivity theory of wages, i.e. that firms set the level of the executive's compensation in accordance with, and to correspond to the value of his contribution to the per— formance of the firm. H-l. The level of compensation received by top corporate executives in the form of base salary will be positively and significantly correlated with the scale of the firm's operations. This positive relationship between base salary and size will reflect a reward to the executive who has greater amount of responsibility than the manager of a smaller firm. Changes in the level of executive compensa- tion received will be positively and significantly correlated with changes in the level of performance of the firm. This relationship between pay changes and changes in corporate performance will be reflected in changes in the level of base salary received as well as the level of bonus and stock option grant awarded. 26 H-3. Changes in the level of executive compensa- tion received will be more strongly related to changes in the performance of the firm relative to the performance of other firms in the industry of which the individual firm is a member than just the absolute level of performance of the firm. H-4. There is a positive relationship between the amount of human capital possessed by the individual executive and the level of compensation he receives. This positive relationship between pay and human capital will reflect the effect of human capital on perceived marginal productivity and in addition may reflect a positive return to credentialism on the part of the individ— ual. The methodology to be employed in testing these hypotheses is that of multiple regression analysis. A discussion will be presented in Chapter IV of the specific sample to be examined and the measures of performance to be included in the analysis. The following chapter, Chapter II, contains a review of the studies which have been conducted to date in an attempt to uncover which factors influence the pay determination process relative to top corporate execu- tives. By reviewing these studies one may see areas where fruitful analysis has been made and also be alerted to areas which need to be more thoroughly explored and the analysis of which needs to be refined. CHAPTER II A REVIEW OF THE LITERATURE CONCERNING THE RELATIONSHIP BETWEEN EXECUTIVE COMPENSATION AND CORPORATE CHARACTERISTICS In this chapter, a critique and review of the existing body of literature concerning the relationship between corporate characteristics and executive compen- sation levels in large U.S. corporations will be pre- sented. By carefully analyzing the research efforts to date which have attempted to discern which characteris- tics of the modern corporation most strongly influence the level of pay received by top corporate executives, one can not only gain insights into the approaches taken by previous researchers, but also identify areas in which further research is needed. Hence, by gleaning the positive contributions from each of these earlier studies one can adopt the posture of "standing on the shoulders of giants" when attempting to advance the body of knowledge which currently exists as to the pay deter- mination process for top executive positions. 27 28 The question as to the determinants of the incomes received by top corporate executives was examined very early in the economic literature of the twentieth century by F. W. Taussig and W. S. Baker. In their 1925 Quarterly Journal of Economics article the authors posed the fundamental issue in the study of executive incomes as being, "Are the driving motives the same for the executives, as for the individual proprietors of older days?" (F. W. Taussig and W. S. Baker, 1925, p. 2). In an attempt to answer this ques- tion Taussig and Baker conducted a survey of a large number of firms representing twenty-four different industries in the American economy. Based upon the results of this survey the authors concluded that execu- tive salaries in the United States were generally rigid downward and further were not adjusted upward year by year on the basis of annual earnings of the firm (Taussig and Baker, 1925, p. 2). Contrary to what was witnessed in European and British firms during this time, the authors found that firms in the United States made relatively little use of incentives for executives based upon corporate earnings. The attitudes that pre- vailed during the early 19003 in American corporations was that executive incomes were fixed in advance and that resulting profits were to be divided among the shareholders of the firm (Taussig and Baker, 1925, p. 22. 29 The authors did, however, assert that over the long-run changes in executive compensation were associated with changes in the profit level of the firm, but no hard evidence was presented to support this claim. The issue of the determinants of executive com- pensation lay relatively dormant from the 19205 until the early 19505. During this period other pressing matters filled the business and economic journals as a result of the disruptions injected into the functioning of the economy as a result of the great depression of the 19305 and the war-related activities during the decade of the 19405. The issue of executive compensation resurfaced in the literature in the early 19505. In 1951 Arch Patton, a since noted writer in the field of executive compensation, published the results of a study he con- ducted on the trends in the administration of executive compensation during that period (Arch Patton, 1951, p. 50). Patton utilized data derived from the American Management Association's survey of executive compensa- tion levels within major U.S. firms. Patton combined this compensation data with Security and Exchange Com- mission data on the performance of SEC reporting firms in order to analyze the relationship between corporate performance and executive compensation levels. The matching of firms represented in both the AMA and SEC 30 data resulted in a sample of 411 companies spanning 22 major industrial categories. Patton found executive compensation levels to vary by industry, and to vary between companies within a given industry according to the profit levels of the firms. On an industry by industry basis, the author discovered that executives receiving above average levels of compensation were concentrated in very competi- tive industries (like retail sales, textile, and department stores). Industries characterized by slow rates of technological and product change (like public utilities, heavy machinery, and nonferrous metals) were found to pay their executives below average levels of compensation. From this finding Patton concluded that firms in industries requiring substantial innovative and creative thinking on the part of management had to offer higher levels of compensation to attract and retain such individuals. Between companies within the same industry, Patton found that executive compensation levels varied positively with profit levels. This discovery led the author to the conclusion that "Profit level of the individual company was by far the most important deter- minant of executive compensation" (Patton, 1951, p. 58). Patton went on to more positively assert that "Broadly speaking, this survey points to a basic principle of 31 compensation that is frequently overlooked: salary increases can only come from profit increases. Only after executives have increased the profits of their company can they anticipate higher salaries. It does not appear to be a question of which came first, the chicken or the egg. The profit apparently must come first" (Patton, 1951, p. 58). Patton's analysis of executive compensation aroused the interests of other writers during this period. Later research efforts, however, failed to support the conclusions observed by Patton in his path- breaking study. Specifically, David Roberts, using the same data sources as Patton did earlier, arrived at polar conclusions from Patton's. Roberts found executive compensation to be significantly related to only one factor, corporate size (David Roberts, 1956, pp. 270- 295). Roberts' study covered the years 1945, 1948, 1949 and 1950. For this period data were gathered on the size and profitability of 410 major companies. Size of the firms was measured by total sizes. Roberts basically tested two conclusions offered by Patton that (1) executive compensation levels are subject to an industry effect and (2) within industries executive com- pensation levels vary directly and positively with the profit level of the individual firm. 32 With respect to the hypothesized industry effect on executive compensation levels, Roberts found such an industry differential to exist. Roberts also pointed out, however, that industries differ in the size firms of which they are composed and that when this variance in firm size is adjusted for the industry effect on compensation disappears (Roberts, 1956, p. 274). Using correlation analysis, Roberts tested the relationship between executive compensation, profits, and sales. A cross-sectional analysis was made of a subsample of 77 or the 410 firms in the survey. The resulting coefficients revealed that the intercorrelation between sales and profits was so high (+ .91) as to make the separate effect of size and profits on executive compensation indistinguishable. Switching to a time- series mode of analysis for these same firms over the period 1935-1950 Roberts found the variability in com- pensation levels to more closely approximate that of sales than that of profits. Thus, the author concluded that executive compensation was more closely related to the size of a firm (its sales) than to its profit— ability (Roberts, 1956, p. 276). The conclusion as to the strong relationship between executive compensation and sales which Roberts derived from his research is echoed by W. J. Baumol in his writings on the activities of management in large 33 American corporations. Baumol advanced the hypothesis that, "the typical large corporation in the United States seeks to maximize not its profit but its total revenues which the businessman calls his sales" (Baumol, 1958, p. 187). The rationale offered by Baumol as to why managers would depart from pursuing a course of profit maximization is that it is to their best self interests to maximize sales because, "executive salaries appear to be far more closely correlated with the scale of opera- tions of the firm than with its profitability" (Baumol, 1959, p. 46). Although offering no formal testing of this hypothesis, Baumol's observed relationship between corporate sales and executive compensation is consistent with the findings of Roberts' earlier statistical analy- sis. Expressing the view that the question as to the determinants of the level of compensation received by corporate executives was not clearly and unambigiously answered in earlier studies, McGuire, Chiu, and Elbing conducted a study which represented a continuation of the previous works of Patton and Roberts (McGuire, Chiu, and Elbing, 1962, pp. 753-761). McGuire, Chin and Elbing conducted a time-series analysis of the correlations between executive incomes, sales and net profits for 45 of the largest 100 industrial firms in the United States. Their analysis covered the 34 7-year period from 1953 - 1959. The authors' data sources were Fortune and Business Week magazines. Fortune reports, on an annual basis, data concerning the performance of the nation's 500 largest industrial firms (as ranked by total dollar value of current year's sales). The organizational characteristics reported by Fortune are sales, assets, net income, stockholders' equity, number of employees, and total return to inves— tors. Business Week provides data concerning the total compensation of the top two executives of approximately 150 of the nation's largest business firms. These data are provided on an annual basis, usually reported in the May or June issues, and presents total compensation figures disaggregated into its salary and bonus compon- ents. By combining Fortune and the Business Week data, the authors were able to obtain performance measures as well as executive compensation figures for the 45 firms in their study. The authors examined seven sets of correlations among the variables: sales, profits, and executive compensation. These seven sets of correlations are: 1. The gross relationships between sales, executive incomes and profits for identical years from 1953 through 1959. 2. Executive compensation lagged one year behind sales and profits. 3. Executive income lagged two years behind sales and profits. 35 4. Sales and profits lagged one year behind execu— tive compensation. 5. Year-to-year incomes were correlated with year- to-year changes in sales and profits. 6. Year-to-year changes in executive incomes were lagged one year behind year-to-year changes in sales and profits. 7. Year-to—year changes in executive incomes were lagged two years behind changes in sales and profits (McGuire, et al., 1962, pp. 754-755). The results of these correlations are presented in Table 1. A5 is evident from the coefficients and t values shown in Table l, the correlations between executive income and sales (columns A and B) were con- sistently higher and more significant than those between executive compensation and profits (columns C and D). From these results, the authors were led to conclude that Baumol's hypothesis is supported, i.e., executive compensation does appear to be more closely correlated with sales than with profits (McGuire, et al., 1962, p. 758). I Baumol's suggested relationship between execu- tive compensation and sales, rather than with profits, has great significance for an advanced economy such as that of the United States in which ownership of corpora- tions is separated from management. If managers do in fact behave in a manner that does not maximize share- holders' well-being, then, many of the assumptions under- lying economic analyses of the firm may no longer be 36 .mmh .a .mmmH .maHnHm cam .sHao .muHsooz "momDOm em.H scam. Ne.~ 555m. NIH mumm5 .54 cam ma .eum mumm5 .54 Hey mm.H moom. mm.~ mamm. NIH mumm5 .a< can ma “mum mumm5 .54 Hoe Hm.H ommH. mm.~ mmHv. mumm5 because .ma .m< .5< Ame e¢.¢ mmmm. mm.e mHmm. H umm5 .m can m No umm5 .5 Hey NH.5 mmmm. 05.5 mmmm. o umm5 .m can m “N umm5 .5 Ame om.e mmmm. Hm.e ommm. o Hmm5 .m can m «H Hmm5 .5 Amy -.e Shem. mo.m mmHm. 5mm5 Hummuso AHV 55 m5 H m u m Hoe Hoe Ame lac .9 mo mmsHm> omusm IEOU mcflocommmuuou ocm mpcmfloflmmooo COHUMHOHHOU Hafiuumm mmmuw>¢ mo mosao>II.H mqmde 37 valid. Wilbur G. Lewellen and Baine Huntsman in their 1970, American Economic Review article, entitled, "Mana— gerial Pay and Corporate Performance," attempt to answer the question as to whether corporate managers' compensa- tion is more closely correlated with total corporate revenues or with measures of shareholders' welfare (the authors utilize profits and equity market value as two measures of shareholders' well-being). Another objec- tive of the Lewellen and Huntsman study is to correct and improve upon some of the statistical and measurement biases they felt were present in the McGuire, Chin and Elbing study discussed earlier in this paper. Lewellen and Huntsman examined a sample of 50 firms drawn from the top 100 firms listed in the Fortune survey of the nation's 500 largest industrial firms. The authors examined the cross-sectional relationships between executive compensation and the performance of the corporations at three—year intervals from 1942 to 1963. The authors constructed the following equation to examine the relationship between executive compensation (C), corporate profits (P), and corporate sales (S): (1) Cit = a0 + a1 Pit + a2 Sit + Uit Subscript i denotes the firm, and t represents the time period to which the measure corresponds. The random 38 disturbance term is denoted by U. The authors conclude that equation (1) may be used as "a basis for observing the magnitude of the coefficients, al and a2 and the levels of statistical significance attaching thereto, the above specification provides a natural vehicle for inferring the relative influence of the two independent variables upon compensation" (Lewellen and Huntsman, 1966, p. 712). In essence, this was the approach taken by McGuire, Chin and Elbing in their earlier time series analysis. However, Lewellen and Huntsman warn that the use of equation (1) may lead to several sources of stat- istical bias. The authors found that the error terms of equation (1) were not random but rather, were in propor- tion to the dependent variable (executive compensation). Further, the authors discovered that those firms that were large in scale also had high sales and profits levels, thus, posing the threat of collinearity resulting from the scale-associated linkage between the independent variables (Lewellen and Huntsman, 1966, p. 712). Lewel- len and Huntsman concluded that because "the error terms (of equation (1)) tended to vary directly with the dependent variable, an appropriate weighting procedure is to divide each variable in (l) by any one of the several scale-related deflators . . . Moreover, by creating ratios in which both numerator and denominator are associated with the firm's size, the weighted regression 39 approach eliminates the basic reasons for expecting high degree of correlation between the variables as a conse- quence of a common scale factor" (Lewellen and Huntsman, 1966, p. 713). The weighting factor selected was that established by Miller and Modigliani in an earlier study, viz., book value of assets (Miller and Modigliani, 1966). Aside from the Miller and Modigliani precedent of using assets as a weighting factor, the resulting deflated equation, equation (2), has a meaningful economic inter- pretation. Given that all the variables in equation (2) are expressed as a ratio of assets, this implicitly represents a process whereby management maximizes sales or profits subject to the available resource constraint, i.e., maximization of sales or profits per dollar of resources employed (Miller and Modigliani, 1966). The new deflated equation is as follows: U C' 1 (sit) it P. t (it) (2)—l=a(—)+a ——+a + Ait 0 Ait 1 Ait 2 Ait Ait In equation (2), A. is total book value of 1t th assets of the i firm in time period t. The authors found that by using the new deflated equation, the new (Uit) Ait This deflated equation was also found to reduce the col- error terms are roughly constant over the sample. linearity problem among the independent variables. 40 In order to improve the measure of executive com- pensation over that utilized in earlier studies, Lewellen and Huntsman added "current income equivalents" (denoted by c*) of various deferred and contingent components of the total executive compensation package. Further, to reduce measurement problems, the authors used two mea- sures of profitability--net profits (P) and equity market value (V)--in an attempt to adjust for any inconsistencies between short-run and long-run profit maximization con- cepts (Lewellen and Huntsman, 1966, p. 714). Upon running equation (2) with the data for the years 1942 to 1963, it was revealed that the coefficients of the profit measure are positive for each cross- section, and the coefficients were also highly signifi- cant, for all of the runs (at the .05 level of signifi- cance). The coefficients for the sales variables were not significant in any run, and further, the sign of the coefficient was not consistent over the runs. It was also discovered that adding deferred and contingent com- ponents of compensation to the measure of total compensa- tion did not improve the fit, rather, the multiple correlation coefficients were reduced when the expanded measure of total compensation were used. A similar result occurred when equity market value was substituted for profits in the equation, for it seemed to have no 41 effect on the findings. The results of this analysis are presented in Tables 2 and 3. Based upon the findings of their study, Lewellen and Huntsman concluded that executive compensation is more closely correlated with profits, rather than sales as was hypothesized by Roberts and Baumol, and later supported by the work of McGuire, Chiu, and Elbing. A possible explanation for these contradictory findings may be the refinements in methodology introduced by Lewellen and Huntsman, as well as their use of a dif- ferent time period over which the study took place. Robert T. Masson in a 1971 Journal of Political Economy article presented his analysis of the relation- ship between the financial returns received by top corporate executives and firm performance (Masson, 1971). The author examined a relatively small sample of 39 firms in the electronics, aerospace, and chemical industries for the years 1947 to 1966. Masson looked at the relationship between changes in executive compensa- tion levels and changes in corporate variables of sales, earnings per share, and rate of return on a share of stock over this 10-year period. The measure of executive compensation used by Masson was very comprehensive in that it included salary and bonus, as well as present value of stock options, pensions and other deferred 42 .mH5.m .mmmH .cmamucsm can cmHHmzmH "mumaom 505. 00. v.5H 55.0 0.050 00.N 0.H5 N00. 00.H 0.0HI H0.0 0.550 vm.0 0.00H 000H H05. 00. 0.H 00.N N.HOHH 0v.0 0.00H 0H0. 00. H.0H NH.0 5.000 00.0 0.HNH 000d 00¢. 50. 0.00 50. 0.H0¢ 00.H 0.0%H 0N0. ow. 0.0 00.0 0.000 00.5 0.NHH 500a H05. 0N. 5.0 I 0H.N 5.00HH 00.v 0.00 0H0. 50. H.0Nl 5N.0 0.NOHH 00.0 0.0NH @00H 000. 00. 0.vHI vH.0 0.NONH 00.0 H.00 500. Om. 0.0 I 00.0 0.500H 00.NH 0.00H HO0H 000. ON. 0.0 50.H 5.5¢v 00.0 0.05 000. 00. 0.NH 0v.N 0.0H¢H 00.0 5.00 0VOH 005. HN.H 0.5m! 00.N 0.0HOH N0.5 0.Hv vv5. H0. 0.00! 00.H N.5000 N0.0 0.HOH 0v0H NHO. 0v.H 0.0NI 50.0 H.0NOH 00.0H H.0v 000. 0N.H 0.00! 00.N 0.0000 00.0 00.00 Nv0H Hm fl NM 9 H0 u 00 NM H mm # Hm u 06 “00% II.II I.III . u u A He u .s u u A He u I Hmc I Has Ho 1 .m5 Hmc .o "cowvmavm scammohmmm ”cowumnvm coammoumom .coqumcmmEoo can .moamm .munoum «madamom cowmmmHmQMII.N Manda 43 .5H5.m .oomH .cmsmucsm one :wHHmamH “momoom mos. Hm. H.mH mn.m m.mv ~5.~ m.m5 emm. e5. H.0HI NH.¢ m.mm mo.m H.omH mmmH was. «H. e.v I mo.v 5.mm mm.m H.mMH mHm. mo.H 5.5H no.m m.mm mm.m 5.omH oomH owe. we. m.mm we. m.mm Ho.~ n.mvH «mm. om.H v.mH MH.m m.mm no.5 m.mHH 5mmH men. he. o.MH oH.~ o.mm mv.v «.mm mHm. mm. m.v I mm.m o.HOH m5.m o.m~H emmH 05m. mm. m.o 5m.~ H.m5 mm.m H.mo mmm. mm. H.vH m5.~ m.mOH Hm.~H H.HHH HmmH mmm. me. m.~H mm. m.Hm oo.m m.m5 mam. em. o.m mm.m m.mmm m5.m m.om mva mom. ow. o.mHI mm.~ m.m0H Ho.5 m.Hv «on. me. m.mmI m5.m m.mo~ mo.o m.mm memH mHm. 0H. m.H I mm.m m.m0H mo.mH am.Hv mHm. we. m.m~I ~m.~ «.mmm ~5.m o.mm mva mm H mm u Hm u on mm u «m u Hm u om umm5 .5 u u H0 u .s u u A He u A . H55 Ho . .mc l .55 Ho "cofium50m cowmwoumom “coaumowfi scammmummm .cofiummcmmaou new .moamm .mmadm> umxumz «muasmmm cowmmmumwwwwwwflfllfifln name A 44 compensation. The methodology utilized was that of multiple regression analysis. Masson found that executive compensation levels tend to be more closely correlated with rate of return on stock than with any other variable (Masson, 1971, p. 1285). Masson further failed to find any support for the Baumol hypothesis, discussed earlier, that the com- pensation of executives is more closely correlated with sales-maximization than with profit-maximization. Based upon his analysis the author states, "it appears that stock market performance may be the most important determinant of executive returns" (Masson, 1971, p. 1285). Although the analysis conducted by Masson attempted to improve upon the past studies to reveal the determinants of executive compensation the results of his analysis are suspect for three major reasons. First, the author selected to analyze a very small sample of firms. By utilizing only 39 firms in his sample the ability to generalize from his results is greatly reduced. Secondly, the generalizability of the results derived from this research are further reduced by the fact that the industries studied (electronics, aero- space, and chemicals) were all growing at a remarkably rapid pace during the period studied (1947-1966) and therefore may tend to distort the analysis. 45 Finally, the measure of executive compensation used included the stock option portion of the total compensation package, a very popular form of remunera- tion during this period. The problem comes in the way in which Masson computed the value of the stock option component. Rather than using value of stock when granted, the author calculated the present value, to the year granted, of the stock option when exercised. Util- izing the value of stock as exercised, rather than when granted, valuation of the stock component of the execu- tive pay package would to a great extent measure the executives' ability to play the stock market rather than the relationship between corporate variables and the level of executive compensation. Given this method of valuing the stock component of the compensation package it is not surprising that the level of executive compen- sation was most highly correlated with the firms' stock market performance. Steven R. Cox and Donald Shauger examine the relationship between executive compensation, firm sales, and profitability (Cox and Shauger, 1973). The authors attempt to test Baumol's sales-maximization hypothesis by conducting a cross-sectional analysis of the rela- tionship between the level of executive compensation, firm sales, and profitability in 1969 and 1970 for a 46 sample of 100 American manufacturing corporations in 19 different industries. The authors obtained data on executive compensa- tion levels for the sample firms from the Business Week annual survey of executive compensation. Profit and sales figures were drawn from Moody's 1971 Industrial Manual. The measure of profitability used in this study was the ratio of gross profit before taxes to stock- holders' equity. Three different measures of executive compensation were used in their analysis: (1) salary plus bonus, (2) salary plus bonus plus deferred and contingent compensation arrangements awarded in year t and (3) salary plus bonus plus deferred and contingent compensation plus stock options exercised in year t (Cox and Shauger, 1973, p. 31). Upon regressing the various measures of executive compensation with sales and profitability the authors found compensation to be significantly correlated with both sales and profitability. However, although both were significant, profitability was more strongly correlated with compensation than was sales, especially when the more comprehensive measures of compensation were used. In the recessionary year of 1970 the impor- tance of profitability in the explanation of variances in executive compensation levels is further increased, thus implying that the importance of the various 47 determinants of compensation are not constant over the business cycle. The Cox and Shauger study described above represents an advance in the analysis of the determinants of executive compensation in that the authors test vari- ous specifications of the compensation, sales and profits relationship over a large sample of firms representing numerous industries. However, the study is severely limited by the fact that it uses a cross-sectional mode of analysis in a very atypical time period. The period studied, 1969 to 1970, encompassed a time of economic recession. The relative influence of corporate charac- teristics on executive compensation levels may be dif- ferent during a recessionary period than during normal periods, especially given the fact that salaries are generally inflexible downward. Another problem with this study is the manner in which the authors computed their total compensation measure. The authors used as the valuation of the stock option component of the pay pack- age the value of the options exercised in the observation year. Use of the value of the options exercised would tend to misrepresent the level of compensation received during that year because the stock Options being exer- cised may have been granted up to five years in the past. Thus, what the authors were picking up in the total compensation measure by using options exercised was not 48 compensation received from the firms in that year, but rather the return to the executive's ability in the stock market. A more appropriate valuation of the stock option component of the total pay package would have been the value of stock options granted to the executive by the firm in the observation year. David Ciscel hihis 1974 article entitled "Determinants of Executive Compensation" reviews the earlier research efforts in this area and offers his own analysis on the topic (Ciscel, 1974). Ciscel selected as a sample 210 of the 250 largest industrial corpora- tions in the United States. Corporate characteristics and performance data for this sample of firms were obtained from the Fortune ranking of the nation's 500 industrial corporations (ranked by sales). Executive compensation data for the executives of these 210 firms were derived from the Forbes listing of executive com- pensation levels for the "Forbes 500" firms. Two measures of executive compensation were used by the author. First, salary plus bonus of the chief executive was used, following the precedent of McGuire et a1. Secondly, total executive group compensation was used. The rationale for examining this second measure of com- pensation is the hypothesis expressed by J. K. Galbraith that power and control in the modern large corporation has moved from a single executive to a whole executive 49 group (Galbraith, 1971, p. 171). The corporate variables examined were sales, assets, net income, and number of employees for the years 1969, 1970 and 1971. Ciscel defines the first two variables, sales revenue and assets, as representing managerial concern with corporate growth and size. After tax profits, net income is used by the author to proxy shareholders' interests. Follow- ing the Galbraithian hypothesis, employees are used to represent the corollary that total company employment is also a reward for successful management. Finally, executive compensation is examined as a function of the chief executive's tenure with the corporation. It was found, however, that length of service had no impact on the level of compensation received by the senior execu- tive (Ciscel, 1974, p. 616). Table 4 presents the simple correlation coef- ficients between executive compensation, sales, assets, net income and number of employees. As is evident from the table there seems to be no strong correlation between executive compensation and any of the variables. Further, the correlation coefficients demonstrate little variance across the independent variables. Table 5 presents the results of the same analysis using executive group compensation in place of chief executive compensation. In each year the compensation level of the executive group appears to be more highly 50 TABLE 4.--Executive Compensation--Senior Officer (Correlation Coefficients, r). 1971 1970 1971 Years with Company .161 Sales .329 .364 Assets .374 .399 Net Income .305 .280 Employees .358 .364 1970 Years with Company .110 Sales .350 .400 Assets .376 .411 Net Income .357 .364 Employees .363 .376 1969 Sales .304 .350 Assets .354 .391 Net Income .287 .287 Employees .323 .339 SOURCE: Ciscel, 1974, p. 614. 51 TABLE 5.--Executive Group Compensation (Correlation Coefficients, r). 1971 1970 1971 Sales .575 .602 Assets .573 .589 Net Income .485 .471 Employees .696 .714 1970 Sales .601 .642 Assets .568 .590 Net Income .473 .477 Employees .716 .737 1969 Sales .570 .607 Assets .555 .579 Net Income .482 .485 Employees .678 .701 SOURCE: Ciscel, 1974, p. 615. 52 correlated with sales and assets than with net income. Further, executive group compensation and the number of company employees were highly correlated. The results reported in Table 5 lead the author to conclude that, "in general, the data seemed to con- firm the Galbraithian hypothesis that the financial reward of management is closely tied to the growth and size of the mature corporation. Secondly, the number of employees-~the size of the technostructure--seemed to be associated with the compensation of top management" (Ciscel, 1974, p. 617). Upon reflection, however, I fail to see the significance of Ciscel's finding that executive group compensation is highly correlated with the number of employees of the firm. Given that an increase in the number of employees of a firm will usually result in an increase in the size of the management team needed to deal with this increased work force it is likely that the size of the executive group will increase given a significant increase in the size of the work force for which they are responsible. Since the measure of compen- sation employed in this analysis is executive group compensation, this measure is merely a product of the number of executives in this group times their average level of compensation. Hence, the total executive group compensation level may increase as a result of 53 either the number of executives increasing, the average level of compensation per executive increasing, or some combination of both of these effects. The form of Ciscel's analysis doesn't allow one to determine which of these effects--increasing number of executives, or increasing average compensation per executive-—has taken place as a result of an increasing number of employees in the firm. - The stronger relationship found to exist between executive group compensation and sales and assets than with net income is also suspect. Table 6 shows the independent variables in the analysis to be highly inter- correlated. Ciscel himself notes this problem, although not attempting to correct for it, by stating that "this conclusion must be sharply tempered by the occurance of strong collinearity-~collinearity that permanently obscures the identification of the hypothesized rela- tionship" (Ciscel, 1974, p. 617). Making note of the fact that earlier studies have failed to reach a consensus as to what the determi- nants of executive compensation levels are, John R. McKean and R. Joseph Monsen attempt to settle this controversy in their 1975 article entitled "Executive Compensation and the Theory of the Firm: an Empirical Study" (McKean and Monsen, 1975). The authors examined a sample of 37 of the "Fortune 500" industrial corporatunus 54 .mmH .d .mhmH .cwmeoz can cameo: “momsom 0mm. new. Nam. has. MHA. sew. mmw50Hasm «am. hem. 5H5. mom. «mm. mam. msoocH umz mam. 5mm. «as. mum. mam. mHm. mummmm com. «me. mew. amm. OHS. mam. mmHmm mmmH was. Ham. mam. mam. mHs. mmm. mmm5on5m «mm. 555. mom. 55m. mam. Ame. maoocH sz ooo.H mmm. «we. «55. mam. sea. mummma ooo.H cam. «mm. 5mm. mam. mmHmm osmH ooo.H Hon. 5H5. mom. mmm5oHEEm ooo.H mom. Ham. mEoOEH umz ooo.H mmm. mommma ooo.H mmHmm HhmH mummmd mmHmm moomonEm mEoocH uoz muwmma mmamm osmH HhmH .mmHQMAHm> unwocommocH owuflmm cmm3umm mucwwowmmmoo GOHUMHOHHOUII.0 mqmfia 55 over the period from 1954 to 1965. A total of 428 observations was obtained by pooling data across firms and over time (McKean and Monsen, 1975, p. 126). Executive compensation, as obtained from the Business Week annual survey of executive compensation, was measured by salary plus bonus plus stock bonus. (Although it is not clear how the value of the stock bonus was calculated.) This measure of compensation was regressed against the corporate variables of sales, assets, profits, stock prices, tenure, industry, and control of the firm, as well as change over time in these variables (McKean and Monsen, 1975, p. 126). Two models were tested to determine these relationships. First, executive compensation levels were regressed against current year characteristics of the firm. Secondly, executive compensation was related to the cumulative performance of the firm over the past 12 years. The authors found, by using the first specifica- tion of the model as applied to the sample that sales in the current year was the most significant variable studied, and that profits are not statistically signifi- cant. Other variables found to be of significance were tenure of the chief executive (+), owner control (+), and industry type (McKean and Monsen, 1975, p. 130). Vari- ables that failed to be significant were assets, profits, 56 trend, stock price, change in profits, or the company's previous year's levels of these measures (McKean and Monsen, 1975, p. 120). When executive compensation was related to cumulative performance of the firm over the past 12 years, sales and change in sales were found to be positively and significantly correlated with executive compensation levels. Type of control also had a signifi- cant effect on executive compensation over time, with owner controlled firms paying higher levels of compensa- tion than manager controlled firms. Based upon their analysis the authors conclude that their results are consistent with those of McGuire, Chiu and Elbing that executive compensation is more closely correlated with sales than with profitability. However, when McKean and Monsen make use of the weighted regression techniques introduced by Lewellen and Huntsman [1970 AER] to adjust for heteroscedasticity and multi- collinearity of the independent variables, their results are changed. By using a weighted least squares approach (weighting by firm assets) the authors find that in the current year compensation is significantly correlated with only the variables profits and stock prices. Although when in their analysis WLS is used for the examination of the effect of cumulative past performance on executive compensation neither profits, nor sales, nor any other measure of performance is significant 57 (McKean and Monsen, 1975, p. 131). Thus the initial results reported in this study, that executive compensa- tion was most closely correlated with sales, may have merely been a product of the methodology used which failed to consider statistical biases because of the variability of the error terms not being randomly distributed over the sample and as a result of the independent variables being highly intercorrelated. .'David J. Smyth, William J. Boyes, and Dennis E. Peseau examine the relationship between executive com- pensation, sales and profits in 557 large U.S. corpora- tions during the year 1971 (Smyth, Boyes, Peseau, 1975). The major objective of this study is not to identify the various corporate characteristics and performance vari- ables that influence the level of compensation received by top corporate executives, but rather to examine the relationship between levels of executive compensation and sales-maximizing behavior or profit-maximizing behavior of the executive. Specifically, the authors stated that their intent in this study is to test the sales vs. profit-maximization hypotheses as being characteristic of the behavior of the large modern corporation. The vehicle through which their sales vs. profits question will be examined is via an analysis of executive compensation. "If executive remuneration is a function of profits but not of sales, then we conclude 58 that the evidence supports the profit-maximization hypothesis: if executive remuneration is a function of sales but not of profits, then the evidence would support the sales-maximization hypothesis; and if remuneration is a function of both profits and sales then we conclude that the evidence supports a managerial model in which the firm has a utility function in which both profits and sales are arguments and we are able to estimate the trade-off between profits and sales" (Smyth, Boyes, and Peseau, 1975, p. 72). Smyth, Boyes, and Peseau draw their data from the Forbes Annual Directory Issue (1972) which provides data for the year 1971 concerning corporate profits, sales, total assets, and other characteristics of the nation's 500 largest firms, as ranked by dollar value of sales. Forbes also provides compensation information for the total remuneration of the top executive of the firm and for the top executive group in each firm. In the Forbes data total remuneration consists of salary, bonus and directors' fee, but excludes deferred compensation and the stock option component of the total pay package. The methodology employed by the authors is that intro- duced by Lewellen and Huntsman of weighted least squares, where the weighting factor is the total book value of corporate assets. The basic models utilized are: 122 A EG and T Where EC = EG = P: S: A: U: 59 _ a P S U — — + b — + c — + _ a P S U _ — + b — + c — + — executive compensation average executive group compensation net profits of the firm sales revenues total assets disturbance term (Smyth, Boyes, Peseau, 1975, p. 76). The corresponding regression coefficients and t-values (in parentheses) are given below for running the two models above for the year 1971. 23.9... A 100108 (1) + 1.0731 (g) (28.81) A (9.42) A + 0.0251 (g) [R2 = 0.651] (4.48) __ = 3.3578 (1) 0.2030 (g) A (36.61) X T (6.38) A 0.0105 5 2_ 1 (6 70) A [R —0.765J + (Smyth, Boyes, and Peseau, 1975, p. 78) In the above equations, both the profits and the sales variables are highly significant. The evidence strongly 60 supports the hypothesis that sales and profits both sig- nificantly influence the level of compensation received by the top executive, as well as the top executive group of the large corporations studied. From these results the authors conclude that "the firm has a utility func- tion that includes both sales and profits" (Smyth, Boyes, and Peseau, 1975, p. 79). Smyth, Boyes and Peseau in their research effort have improved upon the earlier studies in this area in that they selected a very large sample to analyze, 557 firms. By utilizing a weighted least squares approach the authors also reduced the problems of heterosedasticity and collinearity of the independent variables in the model. However, this analysis fell short on several points. First the authors failed to utilize a compre- hensive measure of executive compensation. Failure to include deferred components and the stock options would tend to ignore approximately one half of the total pay package (Lewellen, 1975, p. 168). Secondly, the authors looked at only two corporate variables in trying to explain variances in executive compensation levels, sales and profits. There is a wide array of corporate scale and profitability measures that may be correlated with compensation which were ignored in this analysis. Fin- ally, the mode of analysis employed, cross-sectional analysis, greatly limits one's ability to generalize 61 from the results of this study. What would have been more preferable would have been for the authors to con- duct a time-series analysis over a long enough period of time, say 15 years, so that the possibility of any one particular point of the business cycle affecting the relationship between the variables would have been reduced. In a quite recent study (1975) Foster, Garro, and Rosario conduct an analysis to determine which corporate variables are most significantly correlated with execu- tive compensation levels. The authors note the fact that many articles and surveys have shown that there is a strong correlation between executive compensation, sales and profits. They state, however, that they are interested in the analysis of other factors that contri- bute to the determination of executive compensation. Foster, et al., select the chief executive officer as the executive position to be analyzed. The basic issue explored in their study was: "does the total cash com— pensation of a chief executive officer vary significantly and positively with the business fortunes of his firm on any other measure than increased sales volume? . . . It is the intent of this article to explore such other determinants--or correlates-~of CEO compensation" (Fosuan Garro, and Rosario, 1975, p. 99). 62 Foster, et a1. hypothesize that CEO compensation would tend to vary with the following variables: 1. Annual business results. 2. Long-term business results. 3. Managerial accountability (size of the firm managed) (Foster, et al., 1975, p. 102). The authors describe the first two factors as represent- ing short-term and long-term performance variables respectively, while the third factor is a key measure utilized in establishing the base salary ranges for chief executive positions. To test this hypothesis, the authors employ a multiple regression analysis on a cross-sectional basis by examining the 100 largest companies in the top Fortune 500 listing of U.S. manu- facturing firms for the year 1972. As in earlier studies, Business Week's annual survey of executive com— pensation was used to obtain compensation data. Foster, Garro and Rosario first constructed a multiple regression equation which contained 14 indepen- dent variables thought to be potentially beneficial in attempting to explain the level of CEO compensation. The 14 independent variables included represent various performance as well as responsibility measures. Upon performing the multiple regression analysis of these variables with CEO compensation, it was discovered, as had been pointed out in earlier studies, that the 63 independent variables were highly intercorrelated because of the existence of a scale effect. Taking the Lewellen and Huntsman approach, the authors created deflated variables, derived by dividing all variables by corporate assets, to control for the influence of size upon the correlation coefficients. After regressing the "deflated" variables against CEO compensation, the authors found that the performance measures (net earnings, return on share- holders' equity, return on capital, and earnings per share growth) correlated in the high .70's and low .80's with CEO compensation. It was also found, however, that the size variables (sales, assets, and number of employees) were only correlated in the .50's with CEO compensation (Foster, et al., 1975, p. 104). An analysis of the individual correlation coefficients for the deflated variables is presented in Table 7. As is evident from this table, CEO compensation is most highly correlated with five-year returns on capital (.81), next was five-year returns on equity (.80), following in correlation strength were the twelve- month measures of these variables. Sales, assets, sale growth, and number of employees were all correlated with CEO compensation in the .50 to .60 range. Based on the results of the "deflated" multiple regression analysis presented in Table 7, Foster, et al., 64 .OHH .m .memH .oHuamom one .ouumo .umumom .momoom 00. v0. 05. 55. H0. 00. em. 5m. 00. mm. 50. moo mamfim 3050m NHUOmm Nflmom 0000 000M SEOHO mom muwmmd Chum mOHmm . uoz .Hmuomma op moHummc mmHanum> emauommcmue Hon mHnue coHumHmuuooII.h mamas 65 conclude that the hypothesized relationship between CEO compensation, short-term or annual business results, long-term business results, and managerial accountability is supported. This latest study represents a stride forward in the analysis of the determinants of executive compensa- tion because of the wide range of performance and scale variables examined. Further, the methodology used-- multiple regression techniques, deflated regressors, and stepwise variable inclusion--is much more appropriate than the single factor analysis utilized in earlier studies. Although the Foster, et al., study shows signs of refining the analysis of executive compensation, it stops short in its efforts. One may point to three flaws in this study that deserve attention. Firstly, the use of the position of chief execu- tive officer is somewhat misleading. An examination of corporate titles, as collected and published in the Egg and Bradstreet, Reference Book of Corporate Managers, reveals that the chief executive officer is almost invariably the chairman of the corporation. Because of the differing policies among corporations, there are many situations in which the title of chairman merely represents a figurehead position, or may merely serve as a place to maintain a senior level advisor on corporate policy. Thus, to assign the responsibility for corporate 66 performance to the chairman may many times overstate his role in determining the fate of the corporation. To overcome this problem of lack of comparability of posi- tions across firms and industries I would propose that the authors should have adopted the approach used by Roberts, namely, that of examining the highest paid executive regardless of position title (Roberts, 1959, p. 274). The second point upon which the authors failed in this analysis was in their measure and utilization of the executive pay package. One of the authors' hypotheses was that managerial accountability, i.e., the size of the firm managed, is "a key job evaluation vari- able used in establishing the base salary ranges for chief executive positions" (Foster, et al., 1975, p. 102). Although this hypothesis implies that salary may be contingent upon a scale variable, the level of analysis which the authors undertook examined total compensation (salary plus bonus) and no attempt was made to separately analyze the effects of different firm variables on salary and bonus independently. Further, the authors employed, as in earlier studies, an incom- Iplete measure of executive compensation. By only looking at salary plus bonus the authors ignored the significant portion of the total executive pay package composed of stock options and deferred compensation. 67 Finally, the authors' observations for the com- pensation and performance measures were for the year 1972. As we know, during this period, the nation was subject to the wage and price controls of the Nixon administration. Further, during this period, the econ- omy was on the upswing, rising from the mild recession the nation experienced during the period from 1969 to 1970. Thus, the time span observed by the authors was one characterized by gains in business performance . while at the same time, there were political pressures to hold down increases in compensation levels. The effect Of these two Opposite influences may be to buffer or to partially negate the upward pull of good corporate performance on the level of executive compen- sation. Upon review of the articles and corresponding research efforts to date concerning the determinants Of executive compensation one is left with a general feel- ing of inconclusiveness. There are articles which assert that sales of the corporation is the controlling factor influencing the level of compensation received by tOp executives. However, there also exists a body of studies having found support for the hypothesis that corporate profitability exerts the greatest influence upon the level of remuneration received by top corporate Officials. Finally, one can even find studies that 68 conclude that one cannot discern which factors, either of a scale or profitability nature, are the most important in the determination of executives' pay levels. Of the articles critiqued in this chapter there are none which do not suffer from at least one of the following weaknesses: - poor methodological foundation, especially with respect to collinearity of the explana- tory variables. - small or unrepresentative sample upon which the study was based. - improper specification of corporate variables, especially those dealing with corporate profit- ability. - failure to accurately measure the total execu- tive pay package. - selection of atypical years in which to conduct analysis. The intent of this study is to reduce some of the con— fusion and inconclusiveness which currently exists in this area of inquiry. Further, many of the variable measurement problems, methodological biases, and sampling problems will be improved upon or corrected in this research effort. CHAPTER III A REVIEW OF THE LITERATURE CONCERNING THE RELATIONSHIP BETWEEN EXECUTIVE COMPENSATION AND HUMAN CAPITAL STOCK In Chapter II, a review and critique of the existing body of literature concerning the relationship between structural variables, i.e., organizational characteristics, and the pay levels received by top corporate executives was presented. The emphasis in this section, however, is directed toward evaluating the relative impact of the executive's stock of human capital upon the level of pay received. Individuals invest in themselves through the process of accumulating education, experience, and training. The aim of this investment is to enhance their productivity and hence employability and earning power. There have been numerous studies conducted attempting to measure the effects of different levels of human capital stock upon one's opportunities and ability to compete in the labor market. The focus Of these analyses has primarily centered on lower level participants in 69 70 organizations--skilled, semi-skilled, and unskilled workers--while very few studies exist which have directed their analysis at higher level participants within the organization. The mission undertaken in this section is to present a review and critique of the existing body of literature concerning the application of human capital analysis to top corporate executives and to present a model of the manner in which one would expect executive pay levels to be influenced by human capital factors. Organizations claim that top managers are rewarded on the basis of the position they hold within the corporation and their individual performance within that position. Belcher notes that organizations do not formally recognize the individual's personal character- istics, except the executive's tax situation, when making executive compensation decisions. "Organizations imply that top management is paid primarily for performance and secondly for the job to which they are assigned. Personnel characteristics are not assumed to be recog- nized" (Belcher, 1974, p. 532). Basing pay decisions on the scope of the position held, and the executive's performance within that position conforms to the prescrip- tion forwarded by E. E. Lawler as to how organizations may better utilize pay systems to elicit desired worker behavior (Lawler, 1971, p. 119). 71 Although it is stated by corporate directors and behavioral science researchers alike that it is desirable to base rewards on position and performance, and not to base pay decisions on the personal characteristics of the executive, there exists inconsistent evidence as to how closely the preceding statements describe the actual pay determination process for top executives. Robert Sibson, based upon the results of his annual management compensation survey, presents results which tend to support the above statement that personal characteristics of the executive do not influence the level Of pay received by the individual. As a result of his research, Sibson concludes that: - Length of service does not usually affect salary in any measurable way. Companies tend to pay a new man on the job the same as the man he succeeds. - Age does not correlate with salary level very well. Younger chief executives receive about the same as their older peers (Sibson, 1971, p. 30). Thus, Sibson concludes that the two traditional human capital variables of age and tenure do not seem to exert any significant influence upon the pay decisions concerning top executives' compensation levels. However, other studies into the relationship between pay and personal characteristics of top managers have failed to support the findings of Sibson. A study of McKinsey and Company on executive compensation found 72 that new chief executives generally were given lower levels of compensation than were afforded the individuals they succeeded (McKinsey and Company, 1970). This find- ing would tend to indicate that there is a seniority element which positively influences the level of pay an executive receives. Patton provides evidence to support the findings of the McKinsey study. Patton studied the relationship between executive pay, performance ratings received by the executives, and the executive's age across 86 large firms for the year 1967. What Patton found was that the pay differential between high and medium performers tended to increase with age (Patton, 1968). Patton notes that there is a positive relation— ship between age and pay level, especially for the top performers, and that specifically, "the 90th percentile executive is paid approximately $1,000 additional for each year of age until he 'peaks out' at age 62" (Patton, 1968, pp. 36-37).. To date the most thorough analysis of the rela- tionship between executive compensation and human capital accumulation is that by Kenneth Foster in his 1969 article entitled "Accounting for Management Pay Dif- ferentials" (Foster, 1969). Foster's article was prompted by what he felt to be an absence of any signi- ficant work exploring the process of executive pay determination. In reference to the research previously 73 directed at this issue, Foster notes that, "there is virtually no empirical evidence available that provides any information as to which factors are important, or to what degree they are important over the broad spectrum of management pay practices" (Foster, 1969, p. 82). The aim of Foster's study is to correct this situation by examining the degree to which a number of variables, indicative of both the individual managers' and the corporations' characteristics, are related to managerial pay levels across organizations. The sample examined by Foster consisted Of 19 firms, with all of which the author had close business and professional ties. The time frame of this analysis consisted of studying these 19 firms on a cross-sectional basis for the year 1968. By use of a survey question- naire, Foster collected data from the sample organiza- tions concerning positions in their management hierarchy and the characteristics Of the individual managers occupying these positions. The focus of this analysis centered on three functional areas within management: (1) computer program- ming, (2) engineering/scientific, and (3) marketing management (Foster, 1969, p. 83). For these positions and their incumbents, information was Obtained concerning the level Of base salary and cash bonus, years of pro- fessional and supervisory experience, amount of payroll 74 and number of employees supervised, age and degree held by the manager, the total number of employees and dollar volume of sales for the organization. By use of multiple regression analysis Foster found that years of professional experience was the variable which had the greatest explanatory power rela- tive to the observed variance in the level of managers' base salary (Foster, 1969, p. 85). Further, Foster found that 85% of the variance in the base salary of managers could be accounted for by the four variables: (1) years of professional experience, (2) number of employees supervised, (3) dollar value of payroll super- vised, and (4) average salary of personnel supervised (Foster, 1969, p. 86). With respect to the two tradi- tional human capital variables included in his analysis, age and tenure, Foster found that managerial pay was correlated .64 with years of supervisory experience, .60 with years of professional experience, and .39 with age (Foster, 1969, p. 84). These findings have led Foster to conclude that seniority tends to be a highly influen- tial factor relative to the pay decisions for managerial personnel. Although Foster's analysis represents the most comprehensive study to date concerning managerial pay 'and personal characteristics, the results obtained are 75 suspect on the basis of the small sample studies and the methodology utilized. Foster states that his sample size is 300 because this is the number of managerial positions studied. How- ever, given that all of these managerial positions are within the 19 firms surveyed, the true sample examined is the 19 firms whose pay practices the author is monitor- ing. Further, given that Foster has restricted himself to studying technical managers (the vast majority of whom were computer programming and engineering/scientific managers) the results may be biased by the nature of this group. The common practice is to utilize "maturity curves" as a basis for the determination of pay levels for this group of employees and thus one would expect a strong correlation between experience and the pay level of these technical managers (Sibson, 1967). The independent variables contained in the multi- ple regression analysis of Foster may be subject to severe problems of collinearity. The variables sales and number of employees are correlated to such a high degree (+ .70) as to make their inclusion in a single multiple regression equation result in their independent contributions being indistinguishable. The same criti- cism holds for the variables of number of employees supervised and the total dollar value of payroll 76 supervised. The author made no efforts to correct for these problems in the specification of his model. Because of the methodological and sampling prob- lems discussed above, the results of the Foster study tend to shed very little new light upon the question posed by Belcher concerning which variables affect the pay levels received by top management, and the degree to which they affect this determination. What is needed to improve and update the research in this area is a clearly specified model incorporating both human capital and organizational characteristics to be tested over a broad range of firms for a longer period of time. The preceding review of the existing body of literature concerning the role of human capital in the pay determination process reveals the general lack of sound and thorough analytical analysis in this area. Rather, the major emphasis of the analyses which have been conducted concerning the personal characteristics of top corporate executives has been of a descriptive nature. Lewellen has analyzed the age distribution of top corporate Officials as part Of his study of the characteristics of management within the modern corpora- tion (Lewellen, 1975). Descriptive information of this type concerning the age, level of educational attainment, and business background of the top managers of large U.S. 77 firms is available through the annual survey efforts of both Standard and Poors as well as Dun and Bradstreet. Recently, Standard and Poors Corporation surveyed 74,000 executives for its 1977 Register of Corporations, Direc- tors and Executives. Based upon the survey, it was found that 30 percent of these executives attended just 12 schools (Lansing State Journal, 1976, C-ll). The most frequently attended school was Harvard University. The results of this survey suggest that type of school attended will have a bearing upon one's probability of achieving a top management position with the corporation. Hence, there appears to be a reward to type of school attended in the form of probability of acquiring a top management position. Likewise, one could question whether this is the only reward accrued by the individual as a result of schooling. It would also be of interest to examine the extent t0‘which type Of school attended influences the level of pay received by tOp managers. It is quite probable that type Of school attended not only influences the individual's chance Of entering the ranks of top management, but also may be reflected in higher level of pay either as a result Of the individual having a higher perceived marginal product because Of the type of school attended or as a result of the firm paying a higher return for "credentialism" within the firm. 78 In Chapter V, the level of formal educational attainment as well as the type of school attended will be examined in an attempt to determine the effect of education on the level of compensation received by the individual executive. In an effort to depart from merely reporting and presenting an array of descriptive information concerning the personal characteristics of top corporate executives, a model is needed which explains the manner in which the personal characteristics of the individual impact upon and influence the pay decisions of the board of directors of the corporation. Such a model is provided by Yoran Weiss. Weiss, in his analysis of the return to invest- ments in higher education, provides an excellent model depicting the role of human capital factors in the wage determination process (Weiss, 1971). The basic model developed represents an application of marginal productiv- ity theory to the determination of wage levels. Weiss states that the variables in the individual's stock of human capital are the ability endowment of the individual (A), his accumulated work experience (H), and the amount of schooling the individual has acquired (E). The manner in which these human capital factors influence the wage level received by the individual is through their effect on the perceptions of the employer as to the individual's ability to contribute to the functioning of the firm. 79 The basic elements of the wage function hypothesis seem self-evident. Employers are willing to Offer higher wages to the able, educated, and exper- ienced workers, who are presumably capable Of doing anything the less able, educated, and experienced worker can do, and a little more. In other words, able, experienced and educated employees are more "productive" and in relatively limited supply. We assume, therefore, that the first order derivatives of the wage function are all positive (Weiss, 1971, p. 833). It is acknowledged by the author that higher levels of educational attainment may yield the employee a higher level of income as a result Of "credentialism." In the case of credentialism, the firm actually consumes the human capital Of the individual in and Of itself because it is deemed to be to the advantage of the cor- poration to have in its employ individuals with certain levels of education and experience. In this instance, the productivitywmf the individual and his contribution to the firm is expressed in terms of the status and prestige he brings to the organization as a result Of his creden- tials. Although the population which Weiss elected to study was that of scientists holding advanced degrees, the basic model he utilizes is applicable to the study of the executive pay determination process. The use Of the marginal productivity theory of wages as postulated in Weiss' wage function hypothesis allows one to utilize the descriptive information on executives' age, education, and experience provided by the Dun and Bradstreet and the 80 Standard and Poors surveys to analyze the determination of executive compensation levels. The econometric state- ment of the marginal productivity approach as applied to the study of pay determination is presented in the fol- lowing chapter. The objective of this current study is to provide insights into the role which the executive's stock of human capital plays in the pay decisions of the corporate board of directors. The previous studies in this area have failed to construct a model of the executive pay determination process through which to analyze'and inter- pret the results of their analysis. Further, by not limiting their analysis to top executive groups, and given the lack Of sound methodological practices render the results Of these previous studies to be of little practical usefulness for examining the pay determination process for top corporate executives. It is hoped that many of the theoretical and methodological deficiencies ‘will be corrected in this current analysis. CHAPTER IV A PRESENTATION OF THE SAMPLE, METHODOLOGY, MODEL, AND HYPOTHESES TO BE TESTED AND EXPLORED The basic Objective of this study is to present a comprehensive and up-to-date analysis of the deter- minants of executive compensation. Earlier studies on this topic, discussed in the preceding chapters of this analysis, provide a conceptual and methodological launching pad from which to base further research and analysis. Although there will be similarities between the design of the study described herein and these earlier studies, several new considerations will be added in an attempt to refine the analysis. Contrary to the focus of earlier studies of executive compensation, the primary goal of this study ‘will not be to settle the question whether firms that are not owner-managed, as is the case of most large U.S. corporations, pursue a profit-maximizing or a sales revenue-maximizing course of action. The test of this sales-profit controversy as hypothesized by Baumol is beyond the scope of this analysis. Rather, as was 81 82 pointed out earlier, the intent of this study is to examine the relationship between executive compensation levels, corporate characteristics, and human capital variables in an attempt to discern which of these factors most strongly influence the executive pay determination process within large publicly held corporations. In this chapter, a presentation will be made of the sample to be studied and the methodology to be employed. Further, the model to be utilized in testing the relationship between various structural and personal characteristics will also be formalized and presented. Sample In order to examine the pay determination process for top corporate management it was necessary to select firms of sufficient size and with a wide public distribu- tion of their common stock so as to ensure that manager behavior may be viewed as a phenomenon relatively separ- ate from ownership. Because of the ease Of access, and the comparability of the data reported across firms, the sample of corporations for this study was drawn from the Fortune listing of the top 500 U.S. industrial firms. The Fortune Directory of the 500 Largest U.S. Industrial Corporations was used to obtain data on cor- porate characteristics and measures. Data concerning top executive compensation was drawn from Business week's 83 "Annual Survey of Executive Compensation." This survey provides data on executive compensation, disaggregated into salary, bonus and deferred compensation, and stock options (exercised as well as granted) for the two top executives, chairman and president usually, of approxi- mately 150 of the nation's largest corporations. A matching of the corporations represented in both the Business Week and the Fortune surveys yields a sample of 80 firms. These 80 firms are examined in this analysis over the lS-year time period from 1961 to 1975. It is felt that selecting such a relatively long period of time to study would reduce some of the cyclical varia- tions which may serve to confound the analysis. To overcome the problem of the lack of comparability of position titles across firms, alluded to in the Foster et a1. study, I will utilize the concept introduced by Roberts of looking at the highest paid executive, regardless of position title (Roberts, 1959, p. 276). Information concerning the personal character- istics of the top executives of the nation's largest firms is readily available from the Dun & Bradstreet Reference Book of Corporate Managements. This annual survey provides information on the age, education, experience, and tenure for approximately 75,000 top American executives. Combining the information on the personal characteristics of the executives represented 84 in the Dun and Bradstreet survey with the compensation information reported in the Business Week survey provides one with the necessary information in order to test the relationship between executive compensation levels and the stock of human capital possessed by these top execu- tives. The sample selected is very representative of the nation's top 500 industrial firms in that the firms selected range from number 1 in the Fortune rankings to as low as number 323. The 80 firms in the sample span approximately 20 different industries, therefore, increasing the representativeness of the sample. Further, given that the firms constituting the nation's largest industrial firms are generally regarded as the pace setters in personnel related activities, the results of studying this group may serve to provide insights into the compensation practices of smaller organizations as well. Procedure Three general groups Of variables will be exam— ined to test their relationship to the level of pay received by top management. These three groups of variables may be described as those (1) having to do with the demands placed upon the executive as a result of the size of the organization and hence the amount of 85 resources the executive must manage, (2) those factors measuring or representing the performance of the firm and of the executive, and (3) the personal characteris- tics Of the executive, i.e., his stock of human capital. The size Of the organization serves as a job evaluation type factor for top executive positions. The size of the firm will, to a large extent, define the scope of the responsibilities placed upon the executive as a result of the amount of resources he must manage. Using the Wachtel and Betsey terminology, these struc- tural characteristics represent "demand-side" variables in that they measure the demands placed upon the indiv- idual by the size of the firm (Wachtel and Betsey, 1972, p. 121). Characteristics of the firm which constitute size measures are total dollar value of sales, book value of assets, and total number of employees of the organization. A In terms of the marginal productivity theory developed and presented in Chapter I, this vector of scale variables will influence the pay decisions of the board of directors in that managing a large firm is perceived to represent a greater contribution on the part of the individual than managing a relatively smaller firm. Thus, the scope Of the job the individual holds, i.e., the amount of resources managed, will represent a measure of the individual's contribution to the firm and 86 he will be rewarded accordingly. Equation (3) depicts this relationship between firm size and executive pay level. (3) Executive Payjt = f(SEit, ASit, EEit) where: SE = total dollar value of sales revenues of the firm AS = total book value of the firm's assets EE = total number of employees of the firm j = the jth executive i = the ith firm t = time period t. Based upon the marginal productivity model of labor demand, one would expect that the coefficients for the relationship between executive pay and the scale characteristics of the firm will be positive. This would signal that those executives managing larger firms are performing at a higher level than their counterparts in smaller firms and are therefore rewarded at correspond- ingly higher levels. The second vector to be analyzed in the pay determination process is that Of the contribution Of the individual to the achieving Of organizational goals and objectives. As was pointed out earlier, there are 87 basically two goals which the organization may elect to pursue, growth-maximization or profitability-maximization. Performance of the firm, which is utilized to proxy the performance of the individual executive toward the achieving of desired goals and objectives, will be employed to analyze changes in executive pay. If the individual executive is rewarded in accordance with his contribution to the performance of the firm, one would expect pay levels to change in the same direction as I changes in the measures of corporate performance. If the goals established for the corporation by the board of directors are to pursue a course of growth-maximization, then one would expect executive pay to be positively correlated with changes in the scale of the firm's Operations, i.e., DPAY = f(Dscale), where DPAY measures year-to-year changes in the level of compensation received by the highest paid executive. More specifically, in terms of the individual measures in the vector of corporate size, the relationship may be 2 expressed as: DPAY = f(DSE, DAS, DEE). 2Implicit in this analysis is the fact that I am ignoring inverse causation between the dependent and the independent variables. 88 Likewise, when employing the marginal productiv- ity approach to the analysis of executive pay determina- tion, one would expect pay to vary with changes in corporate profitability if the goals established by the board relate to a profit-maximization rather than a growth-maximization objective. In this situation: DPAY = f(DNISE). NISE represents net income as a percent of shareholders' equity. Earlier studies in this area have utilized net income of the corporation to represent and measure the profitability of the firm. However, this measure is far from a pure measure of corporate profitability in that it represents a scale factor as well. A more meaningful and more easily comparable mea- sure of profitability would be one which expresses profitability as a percent of resources employed, i.e., return on investment. Dividing net income by the amount of capital utilized, one adjusts for the scale component of the net income measure. Therefore, net income as a percent of shareholders' equity will be used in this analysis when examining the relationship between execu- tive compensation and corporate profitability. The personal characteristics of the individual executive have already been acknowledged as having the potential to influence the level of pay received. The 89 extent to which these human capital factors influence the pay determination process is discernable only after one has identified and measured the relevant personal characteristics which the executive brings to the labor market. These "supply-side" variables of the wage determination process have been the subject of numerous labor market studies (Mincer, 1970). The most commonly studied human capital factors are age, education, and experience. An examination will be made of the extent to which these factors correlate with the levels of pay received by top corporate executives. The specific form of this relationship between the executive's accumulated stock of human capital and the level of pay received is expressed in equation (4). (4) PAY = f(age, education, tenure) The elements comprising this human capital vector are the age of the individual executive (AGE), the level of for- :mal educational attainment (ED), the type of school attended (SCH), years with the present employer (YRS), years in current position (YICP), and whether the execu- tive was recruited directly into his current position or worked his way up through the internal structure of the corporation (05). To allow for the possibility that the effects of these human capital factors may vary in their (degree of influence over the range of values they may 90 assume, the variables age, years with the present employer, and years in current position will be expressed in a quadratic form so that the rate of change in the dependent variable can be measured for changes in the independent variables. Thus, the vector of personal characteristics is given by equation (5). (5) PAY = f(age, level Of educational attainment, type of school attended, years with company, years in current position, AGESQ, YRSSQ, YICPSQ), where AGESQ is age squared, YRSSQ is years with current employer squared, and YICPSQ is years in current position squared. Based upon a knowledge of human capital theory, one would expect that the correlation coefficients for the variables AGE, ED, YRS, and YICP would be positive representing a return, in the form of higher pay, to the executive fOr his investment in human capital accumula- tion. The signs of the coefficient for the squared human capital variables indicate the change in pay as the level of human capital possessed increases. For example, if the coefficient for the relationship between age of the executive and level of pay received were posi- tive, this would indicate that the executive's level of earnings would increase as he got older. However, if the coefficient for the age squared term were negative, this 91 would indicate that the compensation level increases at a decreasing rate as the executive ages. The methodology employed to test the relationship between executive compensation, corporate size, corporate profitability, and personal characteristics is that of multiple regression analysis. The model presented below will be tested over the 80 sample firms for the lS-year period from 1961 to 1975. Various specifications Of the executive pay package were regressed against the corpor- ate and personal characteristics in an attempt to determine which factors influenced the individual com- ponents of the pay package (salary, bonus plus deferred, and stock Option grants). Previous studies on this topic tended to concentrate on the total level of pay received and no attempts were made to examine the pay package in a disaggregated form. Based upon an under- standing of the institutional aspects of the pay deter- mination process for top corporate executives, it is hypothesized that base salary will be more strongly influenced by the scale of the firm's operations and the individual executive's stock of human capital (Crystal, 1970). The more volatile components of the pay package, bonus and stock option grants, are thought to be more sensitive to the progression of the firm toward the achieving of corporate goals and objectives. Thus to merely focus the analysis on total compensation in its 92 aggregated form would tend to ignore the possibility that the different components of the executive pay package are differentially influenced by the vectors of corporate size, profitability, and human capital stock. Earlier studies examining the determinants of top executive compensation levels suffered additional methodological problems in their analysis. The scope analysis in the previous studies on this topic have been based on a cross—sectional mode of analysis. Specifically, the research efforts of Patton, Roberts, McGuire, et al., Lewellen and Huntsman, Cox and Shauger, and Ciscel, all concentrated on studying the firms in their samples on a cross-sectional basis over a number of years. Before one can draw any statistical inferences from a study based upon this mode of analysis, one must check to determine whether the regression coefficients estimated by assign- ing subsets (year-to-year Observations) of the total sample to two or more different subsets do in fact belong to the same population (Dutta, 1975, pp. 173-174). The apprOpriate test to determine if one is dealing with the same structure when analyzing subsets of a population is the Chow-test (Chow, 1960). None of the above studies :made mention of utilizing such a test to determine the validity of inferences made about the population drawn from the subsets analyzed. However, for the analysis presented in the following chapter, a Chow-test was 93 performed and the result of this test indicates that the subsets of the structure (year-by-year analysis from 1961 to 1975) do in fact belong to the same overall structure (supported by the finding that at the 95% level the F- critical of 1.75 exceeded the F-calculated of 1.48 obtained from the Chow-test). Finally, performance of the firm will also be expressed as a ratio to that of the median performance of firms in the industry to which it belongs. This mea- sure of relative firm performance will be tested to see if executives are rewarded on the basis of absolute per- formance or relative corporate performance. The latter measure would have the advantage of controlling for the effect of the economy on the performance of the industry of which the individual firm is a member. To adjust for this industry effect on the various corporate variables, I will specify the corporate per- formance variables not in their gross values to be compared across industries, but as a ratio to the industry average for that variable. An example would be if one were looking at the correlation between executive compensation and corporate performance in the automobile industry, that the sales, assets, net income, etc. mea- sures of the firm be divided by the average value of these variables for the auto industry. Such a specifica- tion of the model would also serve to test to see if 94 executives are rewarded for performance relative to that of other firms in their particular industry. 94283.1. The basic model developed and utilized in this analysis is presented in equation (6) below. This model integrates into a single multiple regression equation the three vectors discussed above--organizational size, cor- porate performance, and the individual's stock of human capital--which are hypothesized to influence the level of pay received by the nation's top corporate executives. (6) C = a0 + alSE + a2AS + a3EE + a4NISE + aSAGE + a AGESQ + a YRS + a YRSSQ + a YICP 6 7 8 9 + aloYICPSQ + allED + alZSCH + U The variables contained in equation (6) are: C = executive compensation level of the highest paid executive SE = total sales revenues of the corporation AS = book value of the corporation's assets EE = total number of employees of the corpora- tion NISE = net income expressed as a percent of \ shareholders' equity~ AGE = age of the highest paid executive (AGE) 2 AGESQ 95 YRS = the number of years which the executive has been with the firm prior to assuming his current position, i.e., other years with the firm while not in current posi- tion YRSSQ = (YRS)2 YICP = the number of years which the executive has held his current position, where total years with the corporation-YRS =YICP YICPSQ = (YICP)2 ED = level Of formal schooling beyond high school SCH = type of college or university attended (ivy league, big ten, etc.) U = the disturbance term of the equation. As was noted in earlier studies, the use of equation (6) in its present form poses serious problems of collinearity among the variables in each of the three vectors (scale, profitability, and human capital). The result of this collinearity problem among the independent variables is that the individual contribution of each variable in explaining the observed variances in the dependent variable is rendered indistinguishable (Dutta, 1975, p. 44). The manner in which this multicollinearity problem in the specification of the model was handled in earlier studies was by utilizing a weighted least squares approach. As a result of dividing the variables in equation (6) by a measure of firm size (assets), it was (discovered that the problem of multicollinearity was greatly reduced. Although utilizing a weighted least 96 squares approach may reduce the degree of collinearity present, this represents but one of the possible econo- metric techniques available to deal with the problem. An alternative method for dealing with the prob- lem of the independent variables in the regression equation being highly intercorrelated is to analyze the variables not in terms of their individual contributions to the explanatory power of the model, but rather to examine the significance of the vector which they repre- sent. This is a preferable method for the purposes of this present analysis, for the emphasis of this inquiry is not focused on the significance Of individual vari- ables, but rather the significance of the firm and the individual characteristics which are represented by a combination Of the individual variables. To clarify this point, an example may be drawn from equation (6). In equation (6) the variables Of sales, assets, and number of employees all represent different measures of the size of the firm, i.e., the scale of the firm's operations. And since these three variables do measure the same characteristics of the firm, they are highly intercor- related (.80+) so as to make their individual contribu- tions to the explanatory power of the model indistinguish— able. Therefore, it would seem logical to examine these variables as representing the vector called scale of the firm and to analyze the significance of this vector 97 rather than merely looking at the significance of the separate elements within this vector. This way of view- ing the model, in terms of vectors representing corporate and individual characteristics, suggests that equation (6) be conceptualized as: (6) Executive Pay Level = a0 + a1 [scale] + a2 [profitability] + a3 [human capital] +IU. Where the elements of the scale vector are SE, AS, and EE, the profitability vector contains NISE, and the human capital vector is composed of AGE, YRS, SCH, ED, YICP, AGESQ, YRSSQ, and YICPSQ. (See page 117 for the defini- tions of these variables.) An additional correction which will be made in this analysis as compared to earlier studies on this tepic is that base salary, which is the more stable component of the executive pay package, will be regressed against values of the independent variables for the pre- ceding year. The rationale for regressing salary in time period t against measures of corporate size, profitability, and human capital measures for time period t-l is that this lagged specification conforms more closely to the actual timing of executive pay decisions by the board of 98 directors of the corporation. Given that the pay deci- sions regarding base salary are made at the beginning of the year, the only available measures of performance and corporate size are those which have been realized for the preceding year, as the performance measures of the coming year are not yet determined. Implicit in the assumptions of the earlier studies that have regressed salary:h1yeart: on measures of corporate characteristics also in year t is that the pay decisions for the executive are postponed until the board has had a chance to determine what the level of performance of the firm has been during that year. This specification of the pay model may be appro- priate for the more volatile components of the pay pack- age, bonus and stock option grant, but it is clearly inappropriate for the larger and more stable component of the pay package represented by base salary. Therefore, to examine the factors which influence the level of base salary received by executives in this sample, equation (6) will be specified such that the independent vari- ables are lagged one year behind the measure of salary. The basic hypotheses to be tested in this study were developed and presented in Chapter I of this analy- sis, and are restated below for the reader's convenience. An accurate statistical examination of these hypotheses will serve to alleviate some of the confusion and incon- clusiveness one encounters when surveying the literature 99 and previous research efforts concerning the determinants Of executive compensation. Hypotheses The level of compensation received by top corporate executives in the form of base salary will be positively and significantly correlated with the scale of the firm's operations. This positive relationship between base salary and size will reflect a reward to the executive who has greater amount of responsibility than the manager of a smaller firm. Changes in the level of executive compensa- tion received will be positively and signi- ficantly correlated with changes in the level of performance of the firm. This relationship between pay changes and changes in corporate performance will be reflected in changes in the level of base salary received as well as the level of bonus and stock option grant awarded. Changes in the level of executive compensa- tion received will be more strongly related to changes in the performance Of the firm relative to the performance of other firms in the industry of which the individual firm is a member than just the absolute level of performance of the firm. There is a positive relationship between the amount of human capital possessed by the individual executive and the level of compensation he receives. This positive relationship between pay and human capital will reflect the effect of human capital on perceived marginal productivity or may mea- sure a positive return to credentialism on the part of the individual (although this distinction is not possible to empirically discern). CHAPTER V STATISTICAL ANALYSES AND RESULTS The analysis presented in this chapter is based upon an in depth study of the pay received by the highest paid executives in 80 of the nation's largest industrial firms. This inquiry was conducted on a cross-sectional basis for these 80 firms over the time period from 1961 to 1975. The components of the executive pay package which were analyzed are salary, bonus and deferred compensation, plus stock option grant.3 The first two components of the executive pay package are rather straightforward and quite easy to measure. Salary and bonus are reported in their dollar values for the year in which they were earned and received. Stock option grants, however, are somewhat more difficult to assign a dollar value. Stock Option grants only become of value to the individual when the price of the stock increases over the life Of the grant, therefore allowing the executive 3Due to the manner in which the data were reported, bonus level and amount Of deferred compensation can not be disaggregated into their individual components. 100 101 to realize a gain from the exercise of the Option (Cheeks, 1974, p. 112). The approach taken when dealing with executive stock option grants in earlier studies has been to valuate the stock option at date of exercise discounted back to date of grant. This method Of valua- tion tends to confuse the study of the determinants of executive compensation with that of the study of execu- tive wealth. By valuing the stock option component of the pay package as of date of exercise, one is measuring to a large extent the executive's ability to play the stock market. This may be an interesting concept if one were attempting to measure the total income received by the executive as a result of his non-work, as well as work related behavior. The interest in this study, though, is to examine the levels of compensation received by executives as a result of their work related activi- ties. The focus here is upon the financial rewards offered to the individual executive by the organization in relation to the demands placed upon the executive by the firm and as a result of the personal characteristics the executive brings to the organization. To achieve this goal better, the proper valuation of the stock option component of the pay package would be that which Twould measure the cost to the organization of providing this form of reward. 102 Foster in his analysis of the cost-effectiveness of stock options as a form of executive compensation has suggested that the proper manner in which to view the value of a stock Option grant is as an interest-free loan to the executive by the company (Foster, 1973, p. 13). Foster provides a good illustration of this point in the following example he constructs: If he [the executive] were to receive a grant of 1,000 shares with a fair market value of $100 per share, for all practical purposes he has been given a loan of $100,000 . . . But if the company elected to Offer a new public stock issue on the date of the grant, the entire payment from the sale of shares ($100,000) would have been received at date of sale . . . From this point of view, then, the only expense to the company is the cost of foregoing the use Of the $100,000 (Opportunity cost) over the exercise period (Foster, 1973, pp. 13-14). Foster further suggests that the proper interest rate at which to valuate this Opportunity cost is the borrowing rate which corporations must pay on funds borrowed. Since in this study I am examining only the largest of the nation's corporations, the interest rate at which these organizations borrow is the prime lending rate. Hence, valuation of the stock Option component of the pay package is computed by multiplying the number of shares received by the market value Of these shares at date of grant times the prime interest rate, in the year of the grant, compounded over the life of the grant. 103 Analysis The model developed in Chapter IV, and presented below as equation (7) was employed to analyze the rela- tionship between levels of executive salary and various corporate and personal characteristics. (7) Salaryt = a0 + a1 SE _1 + azNISEt_1 + a YRSSQ 3AGESQt_1 + a 4 t-l +a 5YICPSQt_ + a AGE 1 6 t-l + a YRS + a YICPt_ 7 t-l 8 1 + a9EDt_l + SCH + Ut-l The definitions of the variables contained in equation (7) are presented on page 117 of Chapter V. The reader will note that the independent variables in equation (7) are lagged one year behind the measure of base salary level. It is thought that this lagged specification of the salary equation more closely conforms to the actual process of compensation determination utilized by the boards of directors of large corporations. Table 8 presents the results obtained from utilizing equation (7) to analyze the base salary levels received by the executives in the sample. Examining the coefficients contained in Table 8 tells the reader little 104 about the individual contribution of each variable in explaining variances in base salary due to the high degree of intercorrelation of the elements within each vector. Hence it would be of greater utility to examine the significance Of each vector in explaining variances in the level of base salary received by the executives in the sample rather than concentrating on the individual elements. If one were to restate the general hypotheses presented in Chapter IV into specific testable proposi— tions one would have a natural vehicle for testing the significance of each of these vectors in the executive pay determination process. The first hypothesis to be tested states that the level of salary received is posi- tively and significantly correlated with the scale of the firm's operations. One may express this postulated link between salary level and scale in a statistically testable form as the null hypothesis Ho which states that the relationship between salary and scale is not significant: HO : B (SCALE) = 0 The value of the F-statistic which one obtains from this procedure is 38.19 (see Table 14). The value of the F-statistic as computed far exceeds the critical 105 value of F at the .05 level of significance, 2.60. Therefore, on the basis of the above test one must reject the null hypothesis that the scale Of the firm's opera- tion has no significant effect on the level of salary received by the highest paid corporate executive. This result would draw one to conclude that corporate size does significantly influence the level of base salary received. An examination of the coefficients of the individual variables contained in the scale vector reveals that the direction Of this relationship between scale and salary is positive. The basis for asserting this positive relationship is that the scale variables of sales and assets have positive coefficients and are significant at the .05 level (see Table 8). The hypothesized positive relationship between salary level and the individual executive's stock of human capital is also supported by the results reported in Table 14. Stating the null hypothesis as: H0 : B (HUMAN CAPITAL) = 0 one is provided the opportunity to test the contention that the level of salary received is not significantly influenced by the amount of human capital the executive has accumulated. 106 The F-statistic for the vector of human capital variables is 37.14. As in the case of the scale vector, this value Of F as computed far exceeds the value of F-critical at the .05 level. On the basis of this result one may reject the null hypothesis that the executive's stock Of human capital does not significantly influence the level of pay received. Within the human capital vector, the individual elements which achieve significance are age and type of school from which the executive graduated. Salary level appears to increase with the age of the executive. Additionally, those executives who graduated from the Big Ten (Dll) experience greater earnings than do execu- tives who graduated from other colleges or universities. This result indicates that those executives who graduated from the Big Ten are more likely to be employed by firms paying relatively higher base salaries than will their peers from other schools. Although not specifically hypothesized, it was discovered in this analysis that the level of profit- ability of the firm does not significantly influence the level of salary received by the top executive group. Based upon the results reported in Table 14, one cannot reject the null hypothesis that the relationship between 107 level of profitability and salary level is not signifi- Vcantly different from zero. H0 : B (PROFITABILITY) = 0 The computed F-statistic for the profitability vector is 1.84, whereas the critical value of F at the .05 level is 3.84, thus not allowing one to reject the null hypothesis. When one examines the more volatile incentive based components of the executive pay package, bonus and stock option grant, one Obtains results similar to those obtained for base salary. The level of bonus received is significantly correlated with the scale of the firm's operations. The relationship between the vector of scale variables and bonus level is significant at the .05 level. This find- ing indicates that the larger firms are more likely to supplement base salary with cash and deferred bonuses than their smaller counterparts (see Tables 9 and 15). Likewise, level of bonus received was significantly correlated with the level of profitability of the firm, indicating that more profitable firms pay higher levels of bonus than did relatively less profitable firms. Interestingly, Table 15 indicates that the level of bonus received is influenced by the amount of human capital possessed by the individual executive. Specifi- cally, there is a negative relationship between years 108 with company and bonus level. This result signals that top executives who are recruited into the firm at a very high level within management are more likely to receive larger bonuses than those individuals who have worked their way up through the organizational structure over a longer period of years. Another result which may be derived from Table 9 is that the level of bonus received increases with the age of the executive, but that it increases at a decreas- ing rate over the individual's working life. This rela- tionship is demonstrated from the negative and significant coefficient associated with the AGESQ variable. The coefficient for age squared measures the rate of change in bonus increases as age increases, whereas the coef- ficient AGE measures the magnitude and direction of change in bonus with changes in the age of the executive. The negative coefficients for the dummy variables D3 and D4, which represent bachelor's degree holders and advanced degree holders respectively, imply that the non-college-graduate in the ranks of tOp management receives higher levels of bonus than do their degreed counterparts. The negative coefficients corresponding to D12 and D14 indicate that executives holding degrees from the University of California and those from MIT are concentrated in firms that pay their top executives 109 lower bonus levels than those firms headed by graduates of other colleges and universities. Finally, an analysis of the F-statistics pre- sented in Table 16 reveals that the value of the stock option grant received by the executives in this sample is significantly influenced by both the scale and the profitability of the firm. The interpretation of this finding is that larger firms award their executives with higher stock Option grants than do the smaller sized firms. Likewise, the more profitable the firm's opera- tions, the greater the amount of the stock Option grant awarded to top management. The vector of human capital elements was not found to influence significantly the value of the stock awarded the executive indicating that stock option decisions are based solely on corporate characteristics.' Specifically, Table 10 reveals that the value of stock Option grants received is positively influenced by the level of sales of the firm and the return to stockholders' equity. The only human capital variable to achieve significance was that of school type from which the executives graduated. Those executives who graduated from Big Ten colleges and universities are more frequently found to be in the employ of firms awarding larger stock option grants. One of the hypotheses advanced in this study is that the level of compensation received by top corporate 110 executives will be positively and significantly correla— ted with the profitability of the firm relative to that Of other firms in the industry of which it is a member. However, when measures of relative profitability are utilized in equation (7), the profitability vector fails to achieve significance at the .05 level for any of the components of the pay package (salary, bonus, or stock options). This result suggests that firms do not base executive pay decisions on relative profitability, but on the absolute profitability level of the firm (see Tables 14, 15 and 16). The preceding analysis represents a "snap-shot" look at the sample firms at a point in time to see how pay level is related to the level of various organiza- tional and personal characteristics. An interesting analysis would be to see how changes in the level of pay received by top management varies with changes in the level of firm size, performance, and personal character- istics of the executive. This level of analysis would measure the responsiveness of executive compensation levels to changes in growth and profitability of the firm. Equation (8) presents the model employed in the o I I I 4 analy51s of changes in executive compensation levels. 4When equation (8) was expanded by including dummy variables for year of Observation, a positive time trend was discovered to exist for executive pay levels. 111 = . . + . (8) DECit a0 + alDSEl + aDEEl aDASl t t t + a DNISEi + a YRSi + a t 3 t 4YICPi 2 t 5 1t D4. + a AGE. + U 5 1 1 6 t 7 t it Equation (8) shows change in executive compensation levels (DEC) from year t-l to year t as a function of one-year changes in company sales (DSE), number of employees (DEE), value of assets (DAS), and company pro- fitability (DNISE). The human capital measures were not expressed in change terms, however, given that their values for each executive change by one every year that the executive is in the sample- By examining the relationship between changes in executive pay levels, changes in the scale of the firm's operations, and changes in profitability, one can test whether higher levels of performance by the individual executive, as proxied by corporate performance, are rewarded by increases in level of pay received as the marginal productivity theory would predict. Table 17 reveals that change in the level of salary received by this group of top executives is significantly correlated with only one of the vectors of variables, that of change in profitability. This result 5For the analysis of changes in salary levels, equation 8 was examined for changes from year t-2 to year t—l. 112 is further supported by the coefficients presented in Table 11 which show that change in the level of base salary received is significantly correlated with change in net income as a percent of shareholders' equity, but with no other characteristic of the firm or of the individual executive. This result indicates that those executives who can increase the profitability of their firm are rewarded in the form of increased salary levels. Changes in the level of bonus received by top corporate executives are also found to be positively correlated with changes in the level of the firm's profitability (see Table 18). However, unlike base salary, changes in the level Of bonus received are addi- tionally influenced by changes in the scale of the frnms operations. It is further demonstrated in Table 18 that the vector of human capital measures does not influence changes in bonus level. Turning to the examination of the individual variables within each of these vectors, one finds in Table 12 that all of the scale measures are positively and significantly correlated with changes in the bonus component of the executive pay package. Talbe 19 presents the analysis conducted in an effort to determine which factors influence observed changes in the level of stock Option grant received by top corporate executives. Only the vector of scale variables was found to significantly affect changes in 113 the value of stock Option grants received. This result indicates that large corporations utilize the stock option grant as a means of rewarding the executive's success in expanding the scale of the firm's Operations. Referring to Table 13, the stock option is used to reward executives for achieving growth in the firm's sales revenues and dollar value of assets. When change in relative corporate profitability is substituted into equation (8), changes in the profit- ability vector fail to be significantly related to changes in any of the components Of the executive pay package (see Table 7). This confirms the findings for pay levels where level of relative corporate profit- ability represented an insignificant influence on the level of executive compensation awarded to this top executive group. A summary of the results derived from the analy— sis presented in this chapter is outlined below: - level of base salary received by top corporate executives is positively and significantly correlated with the scale of the firm's opera- tions as well as with the individual executivefis stock of human capital. — level of bonus and deferred compensation received by the managers of large U.S. corpora- tions is found to be significantly and nega- tively correlated with the scale of the firm's operations, its profitability, and the execu- tive's accumulated stock of human capital. - level of stock option grant awarded to top mana- gers is positively related to firm size and profitability. 114 - relative levels of profitability have no signifi- cant influence On the level of executive pay. - changes in the level of base salary received by top corporate managers are positively related to changes in the level of profitability of the firm. - changes in the amount of bonus awarded to top executives is positively correlated with both changes in the size of the firm and changes in the level of profitability. - the only vector of variables found to be signifi- cantly related to changes in the value of stock option grants was the vector of scale character- istics Of the firm. Further, the nature of this relationship was discovered to be in a positive direction. - changes in the relative level of profitability of the firm were found not to be significantly correlated with changes in the various compon- ents of the executive pay package. A somewhat different level of analysis examines the distribution of the executive pay package across its various components (salary, bonus, and stock options). This approach represents more of a descriptive analysis of the executive pay package in that the focus is on the ratio of the bonus and stock option components to the total level of pay. However, from this descriptive analy- sis, one may gain insights into the manner in which the composition of the pay package changes as the level of compensation changes. For the sample studied in this analysis, the average level of the total executive pay package was $294,430. Base salary accounts for 73% of the total 115 ($214,785), bonus represents another 19% ($56,547), and stock option grant comprises the remaining 8% ($23,330) of the total package. The latter two components of the executive com- pensation package (bonus and stock option grant) repre- sent the incentive based portion of the executive's remuneration from the corporation. Given that these two forms of incentive pay are much more volatile than the more stable element of base salary, it is of interest to examine how this portion of the pay package changes as total pay changes, i.e., to examine how [(bonus and stock option grant)/TOTAL compensation] changes as total compensation changes. If we set bonus and stock option grant equal to I, and label total compensation T, the question becomes how does (I/T) change as T changes. This question may be posed in a regression format by the following equation: (9) (I/T) = a + bT However, to more accurately analyze changes in the depen- dent variable (I/T) as the independent variable (T) changes it is advantageous to express equation (9) as a log transformation: (10) log (I) = a + b log (T) 116 The resultant coefficient of this regression is 1.74 (which is significant at the .05 level). The interpreta- tion of this result is that for every 1% change in T (total compensation), bonuses and stock options change by 1.74%. Based upon this finding one can assert that bonus and stock option grant as a percent of total com- pensation increases as the level of pay rises, thus indicating that at higher levels of pay the importance of base salary in the executive pay package is reduced. SE AS EE NISE AGE AGESQ YRS YRSSQ YICP YICPSQ D3 D4 D5 D10 BILL D12 Dl3 D14 117 Definitions of Variables Presented in Tables 8 to 13 total sales revenues of the corporation book value of the corporation's assets total number of employees of the corporation net income as a percent of shareholders' equity age of the executive (age) 2 the number of years which the executive has been with the firm prior to assuming his cur- rent position (YRS) 2 total number of years the executive has held his current position (YICP)2 1 if the executive has a 4-year degree, 0 otherwise 1 if the executive has an advanced degree, 0 otherwise 1 if years with firm prior to current position equals zero, i.e. if YRS = O, 0 otherwise 1 if the executive graduated from an Ivy League university, 0 otherwise 1 if the executive graduated from a Big Ten university, 0 otherwise 1 if the executive graduated from the University of California, 0 otherwise 1 if the executive graduated from Stanford Uni- versity, 0 otherwise 1 if the executive graduated from MIT, Ocnherwise 118 TABLE 8.--An Analysis of the Determinants of the Base Salary Component of Executive Pay. Significant at Variable B t the .05 level 53 .26 2.78 AS .30 3.87 BE -.24 .65 NISE .50 1.36 YRS --25 -52 YRSSQ .89 1.67 YICP -.25 -51 YICPSQ —.90 .55 AGE 3.76 5.31 AGESQ --63 -63 D3 2.75 .39 D4 12.65 1.59 05 6.24 .54 010 8.47 1.46 Dll 17.55 2.07 012 48.27 3.75 0.13 -17.97 1.26 014 10.19 1.08 Adjusted R2 = .51397 119 TABLE 9.--An Analysis of the Determinants of the Bonus Component of Executive Pay. Significant at Variable B t the .05 level 53 .33 2.98 AS -.37 .40 BE .24 5.43 NISE 3-34 7-70 YRS -l.85 3.29 YRSSQ .82 1.32 YICP -.44 -93 YICPSQ .83 1.37 AGE 2.35 2-34 AGESQ -.31 2.69 D3 -27.58 3-28 04 -41.89 4.49 D5 -l4.96 1.10 D10 10.68 1.57 011 --49 -50 012 -33.02 2.19 013 19.21 1.16 01.4 -27.04 2.44 Adjusted R2 = .34138 120 TABLE lO.--An Analysis of the Determinants of the Stock Option Component of the Executive Pay Package. Significant at Variable B t the .05 level 33 .32 3.36 * AS -.66 .90 EB -.20 .57 NISE .88 2.26 * YRS --36 -79 YRSSQ .86 ~55 YICP —1.10 .91 YICPSQ -.56 ~35 AGE .28 .42 AGESQ ~42 '45 03 -7.15 1.06 D4 -5.89 .79 D5 7.54 .69 D10 -3.96 .73 011 23.05 2.91 * 912 -7.52 .62 013 -l6.35 1.23 014 5.73 .65 Adjusted R2 = .05427 121 TABLE ll.--An Analysis of the Determinants of Changes in the Salary Component of the Executive Pay Package. Variable B t $293335“ DSE -.12 1.14 DEE . .43 .62 DAS -.20 .51 DNISE .36 1.98 * YRS -.16 .37 AGE _.10 .54 YICP .15 .51 D3 1.06 .76 D4 -1.06 .57 D5 ‘ -6.08 1,01 D10 -.61 .55 D11 —6.06 1.06 D12 1.29 .51 D13 .53 .61 914 5.62 .97 * DSE, DEE, DAS, and DNISE denote first differ- enCes in these variables Adjusted R2 = .00780 122 TABLE 12.--An Analysis of the Determinants of Changes in- the Bonus Component of the Executive Pay Package. Variable B t 5295885.? DSE .12 7.37 * DEE .44 3.41 * DAS .82 11.35 * DNISE 1.03 2.92 * YRS -.21 .86 AGE .18 .33 YICP _.39 .67 D3 -7.72 .97 D4 -8.85 .95 D5 -10.81 .98 D10 2.85 .44 D11 -23.30 2.41 * D12 .67 .32 013 .22 .45 D14 -4.92 .47 *DSE, DEE, DAS, and DNISE denote first differ- ences in these variables Adjusted R2 = .24994 123 TABLE 13.--An Analysis of the Determinants of Changes in the Stock Option Component of the Executive Pay Package. Significant at Variable B t the .05 level 0SE .14 2.56 * DEE .34 .80 DAS .96 3.91 * DNISE -.51 .42 YRS -1.43 1.73 AGE --10 -54 YICP .91 .46 03 -20.55 .76 04 -33.61 1.10 05 17.66 .47 010 -28.56 1.30 011 —68.78 2.10 * 012 29.95 .61 013 24.62 .46 014 -58.25 1.63 *DSE, DEE, DAS, and DNISE denote first differ- ences in these variables Adjusted R2 = .03151 124 TABLE 14.--An Analysis of the Significance of the Rela— tionship Between the Salary Component of the Executive Pay Package and the Vectors of Cor- porate Profitability, Corporate Size, and the Executive's Stock of Human Capital. Ho: B(corporate profitability) = 0 F-calculated = 1.84 F-critical = 3.84 (at the .05 level) Therefore, cannot reject Ho Ho: B(human capital) = 0 F-calculated = 37.14 F-critical = 1.75 (at the .05 level) Therefore, reject Ho Ho: B(corporation size) = 0 F-calculated = 38.19 F-critical = 2.60 (at the .05 level) Therefore, reject Ho Ho: B(relative corporate profitability) = 0 F-calculated = .24 F-critical = 3.84 (at the .05 level) Therefore, cannot reject Ho 125 TABLE 15.--An Analysis of the Significance of the Rela- tionship Between the Bonus Component of the Executive Pay Package and the Vectors of Corporate Profitability, Corporate Size, and the Executive's Stock of Human Capital. Ho: B(corporate profitability) = 0 F-calculated = 59.31 F-critical = 3.84 (at the .05 level) Therefore, reject Ho Ho: B(human capital) = 0 F-calculated = 6.67 F-critical = 1.75 (at the .05 level) Therefore, reject Ho Ho: B(corporate size) = 0 F-calculated = 79.79 F-critical = 2.60 (at the .05 level) Therefore, reject Ho Ho: B(relative corporate profitability) = 0 F-calculated = .36 F-critical = 3.84 Therefore, cannot reject Ho 126 TABLE 16.--An Analysis of the Significance of the Rela- tionship Between the Stock Option Component of the Executive Pay Package and the Vectors of Corporate Profitability, Corporate Size, and the Executive's Stock of Human Capital. Ho: B(corporate profitability) = 0 F-calculated = 736.76 F-critical = 3.84 (at the .05 level) Therefore, reject Ho Ho: B(human capital) = 0 F-calculated = 1.82 F-critical = 1.75 (at the .05 level) Therefore, reject Ho Ho: B(corporate size) = 0 F-calculated = 9.67 F-critical = 2.60 Therefore, reject Ho Ho: B(relative corporate profitability) = 0 F-calculated = .30 F—critical = 3.84 (at the .05 level) Therefore, cannot reject Ho 127 TABLE l7.--An Analysis of the Significance of the Rela- tionship Between Changes in the Level of the Salary Component of the Executive Pay Package and the Vectors of Changes in Corporate Pro- fitability, Changes in Corporate Size, and the Executive's Stock of Human Capital. Ho: B(change in corporate profitability) = 0 F-calculated = 3.86 F-critical = 3.84 Therefore, reject Ho Ho: B(human capital) = 0 F-calculated = .31 F-critical = 1.75 Therefore, cannot reject Ho Ho: B(change in corporate size) = 0 F-calculated = .80 F-critical = 2.60 Therefore, cannot reject Ho Ho: B(change in relative corporate profitability) — 0 F-calculated = .12 F—critical = 3.84 Therefore, cannot reject Ho 128 TABLE l8.-—An Analysis of the Significance of the Rela- tionship Between Changes in the Level of the Bonus Component of the Executive Pay Package and the Vectors of Changes in Corporate Pro— fitability, Changes in Corporate Size, and the Executive's Stock of Human Capital. Ho: B(change in corporate profitability) = 0 F-calculated = 8.54 F-critical = 3.84 Therefore, reject Ho Ho: B(human capital) = 0 F-calculated = 1.06 F-critical = 1.75 Therefore, cannot reject Ho Ho: B(change in corporate size) = 0 F-calculated = 169 F-critical = 2.60 Therefore, reject Ho Ho: B(change in relative corporate profitability)==0 F-calculated = 1.70 F-critical = 3.84 Therefore, cannot reject Ho 129 TABLE 19.--An Analysis of the Significance of the Rela- tionship Between Changes in the Level of the Stock Option Component of the Executive Pay Package and the Vectors of Changes in Corpor- ate Profitability, Changes in Corporate Size, and the Executive's Stock of Human Capital. Ho: B(change in corporate profitability) = 0 F-calculated = .02 F-critical = 3.84 Therefore, cannot reject Ho Ho: B(human capital) = 0 F-calculated = 1.53 F-critical = 1.75 Therefore, cannot reject Ho Ho: B(change in corporate size) = 0 F-calculated = 6.34 F-critical = 2.60 Therefore, reject Ho Ho: B(change in relative corporate profitability)==0 F-calculated = .10 F-critical = 3.84 Therefore, cannot reject Ho CHAPTER VI SUMMARY AND CONCLUSIONS In this chapter, an analysis and summary of the results obtained from the statistical analysis conducted in Chapter V will be presented. The conclusions and inferences which may be drawn from these empirical find- ings will be discussed in terms of the model of the executive compensation process developed in Chapter I. The specific model employed in this analysis is that of the marginal productivity theory of wage determination. Specific hypotheses were derived on the basis of this model and formally stated in a manner which affords the researcher an opportunity to test empirically the rela- tionships postulated. In the text which follows, the findings of the empirical analysis of the preceding chapter will be examined in light of these specific hypotheses. The first hypothesis advanced was that there exists a positive and significant relationship between the size of the corporation and the level of base salary received by the highest paid executive of the firm. The basis for this asserted relationship between salary 130 ‘1” 131 level and firm size is that it is assumed that the management of a large corporation represents a relatively more difficult and demanding task than the management of a smaller firm. Thus, the firm must compensate the executive in the larger corporation at a higher level of pay than his counterpart in a smaller corporation or the executive may be bid away from the firm's employ. The results reported in Table 8 and Table 14 sup- port this hypothesized relationship between firm size and the level of base salary received by the top corporate executive. Within the vector of firm size, the variables sales revenues of the firm and total book value of the firm's assets were positively and significantly correla- ted with the level of salary received by this top execu- tive group. The multiple regression coefficients for the individual elements within the scale vector, however, do not measure the true magnitude of the relationship between the individual variable and the level of base salary received. Rather, due to the existence of a great amount of collinearity between the elements in the scale vector (see Appendix E) the explanatory power of each element in the vector will be understated. As a result of this collinearity problem one is barred from making exact estimates on the effect of a percentage change in the independent variables (i.e. those elements of the scale vector) on the magnitude of the dependent 132 variable, base salary. However, for the purposes of this analysis it is sufficient to note that the vector of scale variables is significantly correlated with the level of salary received, and the coefficients of the individual variables in this vector reveal that the dir- ection of this relationship is positive. The discovered relationship between firm size and executive salary level represents a logical finding based. upon the common practice of using corporate size as a benchmark in the various executive compensation surveys which are conducted to gain information concerning the market for top executive talent. In addition to reveal- ing that larger firms reward their executives with higher base salaries than do their smaller counterparts, this finding has great value to the executive when plan- ning his career. If the executive is following a course of action which is hoped to lead to a maximization of the level of base salary received, which is the most stable and more predictable component of the executive pay package, then the executive should attempt to struc- ture his career path so that position movements are made in the direction of moving always to a larger firm. The rationale for wanting to manage a larger firm is that the larger the scale of the firm's operations the more likely that the executive will receive a relatively 133 higher level of salary than is awarded his counterpart in a smaller corporation. There is an additional point which is implied in the analysis concerning the level of base salary received by the executives in this sample. The inference which may be derived is that firms do not employ an ability-to- pay approach to the determination of top executive sala- ries. The basis for doubting the existence of an ability-to-pay policy with respect to executive salaries is the lack of any significant relationship between the profitability vector and level of executive salary (see Table 11). The finding that profitability does not significantly influence decisions with respect to base salary is consistent with the existence of a competitive market for top executive talent. The firm desiring to attract and retain high caliber managers must be willing to pay at least the market rate for these individuals regardless of the profitability situation of the firm. The finding that firm size is a significant influence on the level of salary received is consistent with the findings of Roberts, McGuire, Elbing and Chiu, Ciscel, and McKean and Monsen. However, the interpreta- tion given to this link between firm size and executive salary is not consistent with that of earlier studies. The finding that executive salaries are significantly correlated with firm size in no way supports or rejects 134 the hypothesis forwarded by Baumol that the modern cor- poration pursues a sales-maximization rather than a profit-maximization course of behavior. Rather, a more appropriate manner in which to test the sales vs. profit-maximization question would be to examine how the pay levels of top executives change as the level of profitability and level of firm sales change. By exam- ining first differences in both pay and corporate characteristics one is afforded the opportunity to gauge the pay-off to executives for achieving either a growth in corporate size or profitability, and based upon the assumption that executives are rewarded in accordance with corporate goals and objectives, one can make infer- ences as to the goals which corporations actually choose to pursue. The second hypothesis proposed in this study is that changes in the level of executive compensation received by the highest paid executives in the sample corporations will be positively and significantly corre- lated with changes in the performance of the firm. The performance referred to in this hypothesis is the ability of the executive to guide the firm along a course of operation which related to growth-maximization of the firm's scale or which relates to profit-maximization (depending on which view of the behavior and objectives of the modern professionally managed corporation is more 135 accurate). The hypothesized link between the performance of the organization toward established goals and objec— tives and the observed changes in the level of compensa- tion received by the top executive would conform to the marginal productivity theory of wage determination in that the executive would receive increases in his level. of compensation in response to the changes in the level of his contribution to the performance of the firm. Fur- ther, by examining the relationship between changes in corporate performance and executive compensation one is provided with a natural vehicle to determine whether the objectives of the typical large professionally managed organization relate to profit-maximization, or whether they are more closely aligned with a growth-maximization posture. Table 17 shows the results of analyzing the rela- tionship between changes in the level of base salary received and changes in the scale and profitability of the firm. Based upon this analysis it was discovered that changes in base salary level are significantly influenced by changes in the profitability of the firm's operations. Further, it is shown in Table 17 that changes in base salary level are not influenced to a significant degree by changes in the scale of the firm. This finding indicates that the typical large corporation follows a policy of basing changes in executive salary 136 levels on the profitability of the firm. Remembering that base salary represents only one component of the executive pay package, it is not possible on the basis of this finding to assert that the typical corporation pursues a profit-maximizing course of behavior. Rather, one must reserve any statements concerning the goals of the professionally managed corporation until the analy- sis of the remainder of the executive pay package (bonus and stock option grant) has been completed. The relationship between changes in the level of bonus received by top corporate executives and the perfor- mance of the firm is presented in Table 18. It is revealed ianable 18 that changes in the level bonus awarded to this top executive group is significantly related to changes in both the scale of the firm and with its profitability. This finding may lead one to conclude that the decisions of the board of directors of the typi- cal large corporation for executive bonus levels are based on the firm's performance in terms of both growth in scale and changes in profitability. In terms of the marginal productivity model, the finding that both growth in scale and growth in profitability influence the bonus decisions of the board signals that the board of directors value the contribution of the individual executive toward the achieving of both of these outcomes by the organization. 137 The final component of the executive pay package, stock option grant, changes in relation to changes in the size of the firm's operations. Table 19 demonstrates that the vector of scale variables is significantly cor- related with changes in the value of the stock option grant awarded the executive. Unlike changes in the level of cash bonus received, policy decisions concerning stock option grants are not significantly influenced by the profitability of the firm. The results of the preceding analysis concerning changes in the level of compensation received by top corporate executives lead one to the conclusion that the board of directors in its decision-making activities bases decisions for the various components of the execu- tive pay package on different factors. Policy decisions concerning adjustments in salary level are based upon consideration of the performance of the firm in terms of profitability. Table 11 shows that the direction of this relationship is positive. Year-to-year adjustments in the level of cash bonus awarded to top executives is based upon a consideration of both the performance of the corporation in terms of growth and its performance in terms of profitability. Variation in the value of the stock Option grant given by the board to the executive is based upon a policy which considers the growth in the scale of the firm's operations, but gives an insignificant 138 amount of attention to the performance of the firm in terms of profitability measures. The policy stance of the typical large corpora- tion for executive compensation decisions as revealed in the above discussion is to utilize the different compon— ents of the executive pay package to reward various types of performance on the part of the executive. Changes in base salary are employed to reinforce executive behavior .which results in a growth in the profitability of the corporation. Similarly, year-to-year changes in the value of the stock Options awarded to top management reflect the success of the executive in directing the corporation on a course of action which results in an increase in the scale of the firm's operations. And finally, the amount of cash bonus awarded to the execu- tive reflects performance of the firm in terms of both increases in profitability and in growth in scale. The conclusion which one infers from the above findings is that the goals and objectives pursued by the large professionally managed corporation are aimed at both firm growth and profitability. Hence, the utility function of the board of directors does not contain only one goal but rather encompasses a desire for both growth of the firm's operations as well as valuing increases in the profitability of the corporation. This finding is consistent with the earlier analysis of Smyth, Boyes, and 139 Peseau in which the authors found that both size and pro- fitability are significant factors in the determination of executive compensation levels (Smyth, Boyes, Peseau, 1975, p. 79). Hence, to merely assert that large corpor- ations pursue either a strict profit-maximization or a strict growth-maximization course of behavior is not possible based upon the results of this analysis of the executive pay determination process. The finding that changes in executive pay levels are positively related to changes in corporate profit- ability and corporate growth is consistent with Lawler's views on how rational compensation systems should be administered (Lawler, 1971). The pay for performance link does appear to exist for this top managerial group. A possible reason why earlier studies in this area have failed to support the existence of this link has resulted from their focus on levels of pay as Opposed to changes in pay levels, and the use of net income to measure per- formance without removing the scale component of this measure. The third hypothesis, H-3, asserting that observed variances in executive compensation levels are better explained by variances in corporate performance relative to other firms in the same industry rather than relative to firms aggregated across industries is not supported by the findings of this analysis. The 140 substitution of relative profitability measures in place of net income as a percent of shareholders' equity in equation (8) resulted in the relative profitability vec- tor not being significantly correlated with any of the components of the executive pay package (see Tables 14, 15, and 16), hence, indicating that executive pay deci- sions are not influenced by the performance of the firm relative to other firms in its industry. When changes in executive compensation levels are regressed against changes in the relative profit- ability of the firm, the same results are obtained as were derived for the analysis of the relationship between pay levels and level of relative profitability. Tables 17, 18, and 19 show that the vector of changes in relative profitability fails to be significantly correla- ted with changes in any of the components of the execu- tive's compensation level. The finding that the relative profitability of the corporation, i.e. relative to the profitability of other firms in the same industry, is not significantly related to the level of pay received by top executives, or to changes in the level of pay received, would signal that the policies established and pursued by large corporations for executive compensation decisions do not contain relative profitability as an important factor in the pay determination process. 141 This finding of the inability of relative per- formance measures to add to the explanatory power of the model may be to a large extent a function Of the owner- ship process in large publicly held corporations. The shareholders of large publicly held corporations have a wide range of investment opportunities open to them which span many industries. Given that top management is ultimately responsible to the shareholders of the company, the relevant comparison when judging per- formance becomes how firms in general have performed, not merely the performance of a specific industry. The fourth hypothesis Offered in this study is that the level of human capital possessed by the indiv- idual executive will exert a positive and significant influence upon the level of compensation the executive receives. The rationale for postulating this positive link between the individual's stock of human capital and the level of compensation received is two-fold. Firstly, the board of directors may feel that they cannot ade- quately estimate the value of the individual executive's contribution to the functioning of the corporation and therefore seek to utilize indirect measures of the ‘executive's potential to contribute to the organization. Such measures of executive ability are education, experi- ence, and tenure in the organization. Secondly, the board may deem it to be desirable to directly "consume" 142 ‘the human capital of the executive because the personal characteristics of the individual may be thought to be ‘valuable to the organization in and of themselves. Rewarding individuals on the basis of their education and experience irrespective of any expected performance effects is termed "credentialism." Rewarding on the basis of credentialism represents a form of pay for the contribution of the executive to the output of the cor- poration in terms of status and prestige. Table 8 shows that the level of human capital possessed by the individual executive does exert a significant influence upon the level of base salary he receives. Specifically, the age of the executive and the type of school attended positively and significantly influences the level of base salary he is awarded by the board. Level of base salary received is directly correla- ted with the age of the executive. However, this rela- tionship may not be of the magnitude implied by the multiple regression coefficient (3.76) due to the fact that the human capital variables of years with the com- pany, and years in current position are highly inter- correlated with the age Of the executive. Despite this inability to distinguish the individual contributions of the age and tenure variables to explaining variations in the level of base salary paid across firms, one is still able to state with confidence that the level of human 143 capital possessed by the executive is positively and significantly correlated with the level of base salary earned by this sample of top executives. It is further demonstrated in Table 8 that the type of college or university attended by the executive influences the level of base salary he will receive in the top management position he holds. The coefficient for the variable Dll measures the relationship between the level of base salary received and whether the execu- tive graduated from a "Big Ten" university or not. What the coefficient for this variable says is that if the executive graduated from a Big Ten university he is more likely to receive a significantly higher level of base salary than those who graduate from other universities. This difference amounts to approximately $17,500 more in base salary than that commanded by executives with degrees from other universities. Thus it may be inferred from this result that graduates of the universities in the Big Ten who make it into the ranks of top management are more likely to elect a career path which results in their being a top executive in a relatively high paying firm. The same interpretation may be given to the coefficient for the variable D12 which measures the impact on base salary of the executive possessing a degree from a school within the California university system. 144 Although it was pointed out earlier that graduates from the Ivy League schools are more likely to be found in the ranks of top management, those who graduate from a Big Ten school or the California university system are more likely to earn higher salaries once they enter the top management group. A possible rationale for this finding is that during the observation years, 1961 to 1975, graduates from Big Ten and California universities were more likely to be employed by the aero-space, auto- mobile, and chemical industries which were experiencing tremendous growth during this period. Therefore, the growth of the industry in which executives from these schools were concentrated may have resulted in their receiving higher base salaries. The individual executive's stock of human capital was also found to influence the level of cash bonus received. Based upon the results reported in Table 9 one may conclude that the Older the executive the more likely that he will receive a larger bonus award, but that there is a diminishing effect of age on level of bonus received as the level of the executives' age increases (see the coefficients for the variables age and age- squared, respectively). School background also exerts an influence on the level of bonus awarded to the executive by the board of directors. Graduates of the University of California school system (represented by 012) and 145 those executives who graduated from M.I.T. (D14) are more likely to be employed by corporations who pay lower levels of bonus than are graduates of other schools. Unlike the two previous forms of compensation discussed (base salary and cash bonus) the level of stock option grant is not significantly correlated with the vector of human capital elements. Thus, one may conclude that the benefits gained from investment in one's self, i.e. accumulation of human capital by the executive, are realized in the form of higher salary and bonus levels but does not influence the value of the stock option grant award received. This would be consistent with the idea that corporate performance influences stock related pay-offs. With respect to the compensation policies adopted by large corporations for executive compensation decisions it is evident that from the preceding analysis that the level of human capital possessed by the executive does influence the pay decisions of the board. The analysis contained in this study provides the reader with several insights into the manner in which executive pay levels correlate with various corporate and personal characteristics. The focus of this inquiry, however, was limited to examining monetary rewards pro- vided to the executive by the corporation in return for his services. The monetary compensation awarded to the 146 executive represents only a portion of the total rewards the executive receives as a result of the employment transaction. Absent from this current analysis is a valuation and examination of the nonpecuniary benefits which the corporation provides to the executive. When discussing the failure of earlier analyses of the earnings function to consider the nonpecuniary rewards in their specification of the pay model, Greg J. Duncan points out that, "the addition of nonpecuniary factors may change these estimates" (Duncan, 1976, p. 463). Duncan states that the analysis of the deter- minants of pay needs to be expanded so as to focus not only on direct compensation but also to include fringe benefits, working conditions, and the amount of satis- faction derived from one's job (Duncan, 1976, pp. 467- 468). Duncan suggests how the analysis of the earnings function may be expanded and improved, contains special significance for the study of executive earnings as a result of the vast array of power and status which accompanies their positions. This author feels that additional research is needed into the area of executive motivation and the effects of various forms of both pecuniary and non- pecuniary rewards on the executive's motivation to achieve organizationally desirable goals and objectives. The motivational implications of executive reward systems 147 represent a key consideration which the board of direc- tors must take into account when determining the level and form of compensation to give the top management group. The emphasis of this current study, and the research efforts which preceded it, is that of an econo- mic rather than of a behavioral science orientation. By concentrating on the economic rather than the behavioral elements of the pay model, one is denied the ability to make prescriptive statements as to how corporations may more effectively structure executive pay systems. Hopefully, this will be corrected in future research efforts merging the economic and behavioral variables into a single model of the pay determination process. BIBLIOGRAPHY 148 BIBLIOGRAPHY Alexander, Clifford E. "Research in Executive Compensa- tion: Multi-Factor Analysis, Multiple Regres— sion." 1976 Regional Conference Proceedings. New York: American Compensation Association, 1975. "Back from the Dean." Forbes, March 15, 1976, pp. 22-23. Baumol, William J. Business Behavior, Value and Growth. New York: McGraw-Hill, 1959. . "On the Theory of Oligopoly." Economica, Vol. 25, August, 1958, pp. 187-198. Belcher, David W. Compensation Administration. Engle- wood Cliffs, New Jersey: Prentice-Hall, 1974. Berg, J. Gary. Managing Compensation. New York: Ameri- can Management Association, 1976. Blackett, O. W. Management Compensation: a Statistical Study of the Relationship of Executive Compensa- tion to Capital Employed and Earnings. U. of M. Press, April, 1953. Boulding, Kenneth E. "Implications for General Economics of More Realistic Theories of the Firm." Ameri- can Economic Review, Vol. 42, May, 1952, pp. 30- 44. Brown, Dilworth F. "The Comparative Performance Index: a New Concept in Bank Officer Compensation." Bank Administration, April, 1976, pp. 61-66. Cain, Glen. "The Challenge of Segmented Labor Market Theories to Orthodox Theory: a Survey." Journal of Economic Literature, December, 1976, Vol. XIV, No. 4, pp. 1215-1258. Campbell, J. P.; Dunnette, M. D.; Lawler, E. E.; and Wicke, K. E. ‘Managerial Behavior, Performance ' and Effectiveness. New York: McGraw-Hill, 1970. 149 150 Cheeks, James B. How to Compensate Executives. Homewood, Ill.: Dow Jones-Irwin, Inc., 1974. Chow, Gregory C. "Tests of Equality Between Sets of Coefficients in Two Linear Regressions." Econo— metrica, Vol. 28, NO. 3, July, 1960, pp. 591-605. Ciscel, David H. "Determinants of Executive Compensa- tion." Southern Economic Journal, 1974, Vol. 40, No. 4, pp. 613-617. Clark, John Bates. Distribution of Wealth. New York: The MacMillan Company, 1899. Cox, Steven R. and Shauger, Donald. "Executive Compensa— tion, Firm Sales and Profitability." Intermoun- I tain Economic Review, Vol. 4, Spring, 1973, 3 pp. 29-39. : Crystal, Graef S. Financial Motivation for Executives. New York: American Management Association, 1970. Cyert, R. M. and March, J. G. A Behavioral Theory of the Firm. Englewood Cliffs, N. J.: Prentice-Hall, Deardon, John. "How to Make Incentive Plans Work." Harvard Business Review, Vol. 50, July-August, 1972, pp. 117-124. Duncan, Greg J. "Earnings Functions and Nonpecuniary Benefits." Journal of Human Resources, Vol. XI, No. 4, Fall, 1976, pp. 462-483. Dunnette, Marvin D. "Compensation: Some Obvious Answers to Unasked Questions." Compensation Review, First Quarter, 1969, pp. 8-16. Dutta, M. Econometric Methods. Cincinnati: South- Western Publishers, 1975. Eckstein, O. and Wilsen, T. A. "The Determination of Money Wages in American Industry." Quarterly Journal of Economics, Vol. 76, August, 1962, pp. 379-414. Ellig, Bruce R. "A Current Look at Qualified Stock Options." Compensation Review, First Quarter, 1973. 151 Fisher, Franklin M. "Tests of Equality Between Sets of Coefficients in Two Linear Regressions: an Expository Note." Econometrica, Vol. 38, No. 2, March, 1970, p. 361. Forbes, Vol. 119, No. 7, April, 1977, p. 30. Foster, Kenneth B. "How Cost-Effective are Stock Options?" Compensation Review, Second Quarter, 1973, Vol. 5, No. 2, pp. 11-18. ; Garro, James; and Rosario, Lina. "Determinants Of CEO Pay." 1975 National Conference Proceed- ings. New York: American Compensation Associa- tion, 1975. "Accounting for Managerial Pay Differentials." Industrial Relations, Vol. 9, No. 1, October, 1969, pp. 80-87. Fox, Harold. Top Executive Compensation, Studies in Per- sonnel Policy 204. New York: National Industrial Conference Board, 1966. Galbraith, John K. The New Industrial State. New York: Houghton and Mifflin, 1971. Goodman, Susan. "Salaries: Industry by Industry." The MBA, August-September, 1973, pp. 9-10. Gordon, Michael. "An Evaluation of Jaques' Studies of Pay in Light of Current Compensation Research." Personnel Psychology: Vol. 22, 1969, pp. 369-89. Gordon, R. A. Business Leadership in the Large Corpora- tion. Berkeley, Calif.: University of Califor- nia Press, 1961. Glueck, William F. Personnel: a Diagnostic Approach. Dallas, Texas: Business PubliEations Incorpora- ted, 1974. Gruenfeld, L. W. "A Study of the Motivation of Indus- trial Supervisors." Personnel Psychology, Hanoch, Giora. "An Economic Analysis of Earnings and Schooling." Journal of Human Resources, No. 2, Summer, 1967, pp. 310-329. 152 Hause, John C. "Earnings Profile: Ability and Schoolingfl' Journal of Political Economy, Vol. 80, May-June, 1972, pp. 108-138. Hettenhouse, George. "Cost-Benefit Analysis of Executive Compensation." Harvard Business Review, Vol. 48, July-August, 1970, pp. 114-124. Howe, R. "Price Tags for Executives." Harvard Business Review, Vol. 34. Lawler, Edward E. Pay and Organizational Effectiveness: a PsychologiEal View. New York: McGraw-Hill, 1971. Lewellen, Wilbur. "Recent Evidence on Senior Executive Pay." 1975 National Conference Proceedings. New York: American Compensation Association, 1975. Lewellen, William G. and Huntsman, Blain. "Managerial Pay and Corporate Performance." American Econo- mic Review, September, 1970, pp. 710-720. McGuire, Joseph W.; Chiu, John S. Y.; and Elbing, Alver 0. "Executive Incomes, Sales and Profits." American Economic Review, Vol. 52, September, 1962, pp. 753-761. McKean, John R. and Monsen, R. Joseph. "Executive Com- pensation and the Theory of the Firm: an Empiri- cal Study." Industrial Marketing Management, Vol. 4, 1975, pp. 125-132. McKinsey and Company. "A Study of Executive Compensa- tion in 216 Large Publicly Held West Coast Com- panies." 1970 (mimeographed). Masson, Robert T. "Executive Motivations, Earnings, and Consequent Equity Performance." Journal of Political Economy, Vol. 79, November-December, 1971, pp. 1278-1292. March, J. G. and Simon, H. A. Organizations. New York: John Wiley and Sons, 1958. Miller, M. H. and Modigliani, F. "Some Estimates of the Cost of Capital to the Electric Utility IndustryJ' American Economic Review, Vol. 56, June, 1966, pp. 333-391. 153 Mincer, Jacob. "The Distribution of Labor Incomes: a Survey with Special Reference to the Human Capi- tal Approach." Journal of Economic Literature, Vol. VIII, March, 1970, pp. 1-26. Moore, Russell F., ed. Compensating Executive Worth. New York: American Management Association, 1968. Patten, Thomas H. "OD, MBO and the R/P System: a New Dimension in Personnel Administration." Person- nel Administration, March-April, 1972, pp. 14-23. Patton, Arch. Men, Money and Motivation. New York: McGraw-Hill, 1961. . "Current Practices in Executive Compensation." Harvard Business Review, Vol. 29, No. 1, January, 1951' pp. 56-640 . "Deterioration in Top Executive Pay." Harvard Business Review, Vol. 43, 1965, pp. 106-118. . "Executive Compensation: the Past, Present, and Future." Financial Executive, July, 1976. . "What is Executive Experience Worth?" Business Horizons, Vol. XI, No. 5, October, 1968, pp. 31- 40. Rees, Albert and Schultz, George P. Workers and Wages in an Urban Labor Market. Chicago: University of Chicago Press, 1970. Roberts, David R. Executive Compensation. New York: Free Press of Glencoe, 1959. . "A General Theory of Executive Compensation Based on Statistically Tested Propositions." Quarterly Journal of Economics, 1956, pp. 270-295. Rock, Milton L. Handbook of Wage and Salary Administra- tion. New York: McGraw-Hill, 1972. Shepherd, W. G. "On Sales-Maximizing and Oligopoly Behavior." Economica, Vol. 29, November, 1962, pp. 420-424. Sibson, Robert. Wages and Salaries. New York: American Management Association, 1967. 154 . "Executive Pay--the Long Term is Where the Action Is." Nation's Business. November, 1971, p. 30. Simon, H. A. "The Compensation of Executives." Socio- metry, Vol. 20, March, 1957, pp. 32-35. . "New Developments in the Theory of the Firm." American Economic Review, Vol. 52, May, 1962, pp. 1-150 Salter, Malcolm. "Tailor Executive Compensation to Strategy." Harvard Business Review, Vol. 51, March-April, 1973. Smyth, David J.: Boyes, William J.: and Peseau, Dennis E. Size, Growth, and Executive Compensation in the Large Corporation. New York: Holmes and Meier, 1975. Taussig, F. W. and Baker, W. S. "American Corporations and Their Executives: a Statistical Inquiry." The Quarterly Journal of Economics, November, 1925. Taylor, J. "Toad or Butterfly? A Constructive Critique of Executive Compensation Practices." Industrial and Labor Relations Review, July, 1968, pp. 494- 496. Torrance, George. Prevalence of Bonus Plans in Manufac- turing, Studies in Personnel Policies, No. 185, New York, Industrial Conference Board, 1962. Wachtel, Howard M. and Betsey, Charles. "Employment at Low Wages." Review of Economics and Statistics, 1972, pp. 121-129. Weiss, Yoram. "Investment in Graduate Education." American Economic Review, Vol. 61, No. 5, Decem- ber, 1971, pp. 833-852. Williamson, Oliver. The Economics of Discretionary Behavior: Managerial Objectives in a Theory of the Firm. Chicago: Markham Publishing Company, 1967. Winton, Donald G. and Sutherland, Charles R. "A Perfor- mance-Based Approach to Determining Executive Incentive Bonus Awards." Compensation Review, Vol. 8, No. 1, First Quarter, 1976, pp. 14-27. APPENDICES 155 APPENDIX A A LIST OF THE CORPORATIONS REPRESENTED IN THE SAMPLE UPON WHICH THE ANALYSIS IN THIS STUDY WAS BASED Boeing General Dynamics McDonnell Douglas Rockwell International United Technologies Avco Lockheed Martin Marletta American Motors Corporation Bendix Chrysler Ford Motor Company General Motors Corporation TRW Borg Warner Armstrong Cork Allied Chemical Dow Chemical E. I. du Pont de Nemours & Co. Eastman Kodak Monsanto Union Carbide American Cyanamid Americaanelephone and Telegraph Columbia Broadcasting Co. General Telephone American Can Co. Continental Can Co. Owens-Illinois Pittsburgh Plate-Glass CO. Johnson & Johnson Eli Lilly Inc. Merck C0. General Electric RCA Westinghouse Zenith General Foods Coca-Cola 156 Nabisco Pepsi Co. Alcoa Anaconda Kaiser Kennecott Reynolds Phelps Dodge International Business Machines NCR Xerox Exxon Gulf Oil Phillips Petroleum CO. Shell Oil Standard Oil of California Texaco Crown Zellerbach International Paper CO. Mead Paper Co. St. Regis Caterpillar Tractor Co. Deere & Co. International Harvester Allis Chalmers Colgate Palmolive Proctor & Gamble Armco Bethlehem Steel Co. Inland Steel CO. National Steel Republic Steel Co. United States Steel Firestone Tire and Rubber General Tire B. F. Goodrich Tire and Rubber Goodyear Tire and Rubber Swift & C0. General Mills APPENDIX B MEANS AND STANDARD DEVIATIONS OF THE VARIABLES ANALYZED IN THIS STUDY Variable Salary ($000) Bonus and Deferred ($000) Stock Option Grant ($000) Total Compensation ($000) Sales Revenues ($0000000) Assets (50000000). Net Income ($000000) Employees (000) Net Income as a Percent of Shareholders' Equity (%) Age of the Executive Years with Company Years in Current Position 14.9.9.9. 214.78 56.55 23.33 294.43 310.64 281.12 184.78 86.81 11.63 57.08 28.40 7.05 157 Standard Deviations 72.26 91.34 82.65 160.77 513.90 511.28 360.42 118.94 6.10 5.63 11.55 5.53 APPENDIX C AN ANALYSIS OF THE SENSITIVITY OF LEVELS OF EXECUTIVE PAY TO CHANGES IN THE LEVEL OF CORPORATE PERFORMANCE Presented below is an analysis of the responsive- ness of the level of executive compensation received by top corporate executives to changes in the level of cor- porate activity. To measure this responsiveness a simu- lation was conducted whereby the minimum and maximum values of the relevant corporate variables were substitu- ted into equation (1) below in an attempt to ascertain the resultant values of executive pay. As is evident from the pay levels reported below which correspond to the minimum and maximum values of the corporate variables, executive pay levels are very res- ponsive to changes in the level of corporate performance. (1) Exec Pay = a + a SE + a AS + a BB + a NISE Minimum Maximum Salary $68,800 $1,988,120 Bonus 4,080 428,740 Stock Option 3,120 869,440 158 APPENDIX D DISTRIBUTION OF EXECUTIVES IN THE SAMPLE ACROSS SCHOOL TYPE Reported below is the distribution of the execu- tives in this sample over the types of schools repre- sented in this analysis. The figures are for the year- by-year composition of the sample of 80 firms. School Frequency Ivy League (D10) 33 Big Ten (D11) 10 Other Schools 22 Univ. of California (D12) 2 Stanford (D13) 2 MIT (D14) 3 159 160 HNVBN.I ovmnm. omvvo. Hmmvo.l mhmao.l thHo.I H¢mmo.l mmmmo. ommmo. QUHW 00000.0 mnmmm. aware. hvmma. Hammo. mmnoo. mmmoo.l evoma. ommva.l wmw mmaho. 00000.0 mmano. ovoma. mmooa. mhmno.l mmmmo. vmoom. Nmomo.l m0< 000 NmH ooo mhh 00m mvm mam mam omm mm 00. 00000. 00000. 00000. 00000.- 00. 00000. 00000. 00000.- 00000. 00.0 00000. 00000. 00000. 00000. 00. 00000.0 00000. 00000. 00000. 00. 00000. 00000.0 00000. 00000. 00. 00000. 00000. 00000.0 00000. 00. 00000. 00000. 00000. 00000.0 00. 00000. 00000. 00000. 00000. 00. 00000. 00000. 00000. 00000. mm mm mmmm mmmmmm 00000 mHqu MIR OZOZ< mBZmHUHhhmOU ZOHEdqmmmOU mAQZHm m xHozmmmd vvoma. vmowm. mammm. mmmvm. vammm. wmmmo. mmhma. ooooo.H omnmo. Numamm ommva.l Nmomo.l ommmv. vamam. vuamv. onmmm. nvooa. omnmo. ooooo.a mucom mm» m0< mm m4 mm mmHz cofluao xooum xumHmm mscom