Top Executive Compensation in Less Developed Capital Markets

Publisher's copyright statement: This is the peer reviewed version of the following article: Pereira, I. & Esperança, J. (2015). Top executive compensation in less developed capital markets. Corporate Governance. 15 (1), 122-133, which has been published in final form at https://dx.doi.org/10.1108/CG-11-2014-0128. This article may be used for non-commercial purposes in accordance with the Publisher's Terms and Conditions for self-archiving.


I. Introduction
Executive compensation is a popular topic in both the popular press and the scholarly literature (Boyd et al 2012).Executive compensation is of considerable interest to the business and academic communities, as well as policymakers (Graham et al, 2012).
The compensation paid to CEOs of large publicly traded corporations rose dramatically during the 1980s and 1990s, stimulating much debate on the determinants of managerial pay (Murphy 1999).
The expansion of executive compensation modes has led to a rising concern about their adequacy, shared by stakeholders, academics and practitioners (Bruce et al, 2007).A vast empirical literature has followed the earlier agency theory explanations about the potential and draw-backs of performance related pay mechanisms (Barkema and Gómez-Mejia,1998;Mishra et al, 2000;Bebchuk and Fried, 2003).
The practice of a single individual serving as both CEO and board chair (CEO duality)has been the subject of academic interest for more than 20 years (Krause et al (2014).
Chief executive officers' compensation has always been in the media spotlight, and particularly so since the financial crisis of 2008.(Datta, 2014).Moreover, the 2008 economic downturn has trigged much discussion about the total amount and the nature of executive compensation, regarded as possible contribution to excessive risk taking.Kaplan (2008), Bogle (2008) and Walsh (2008) provide an interesting debate on this issue.
The high level of pay in recent years has been attributed to the need to compensate executives for the risk generated by a greater use of incentive pay (Frydman and Saks, 2010).Academic research on executive compensation mirrors the prominence of this topic within the business community (Boyd et al, 2012) Although a clear link between pay mechanisms and performance has been hard to find, there is evidence that performance based pay is more often used in the context of higher potential for goal divergence between owners and managers ( Bebchuk and Fried, 2004;Duffhues and Kabir, 2008;Bruce et al, 2007).However, firms that implement executive compensation plans based on performance generally adopt more ambitious and difficult strategies than firms that rely on fixed pay (Dow and Raposo, 2005).The evidence thus calls for public attention for reexamining the effectiveness of current pay system (Lin et al (2013).
We argue that the dominance of principal-agent theory as an approach to investigating executive pay has led to an overly narrow focus which may be unhelpful when considering cross-country differences and probably also hinders within-country analysis.(Bruce et al.,2005).
This result was mostly obtained in the context of well developed capital markets, with a higher separation of ownership and control.Conyon et al. (2011) find that higher levels of pay for U.S.CEOs relative to their U.K .andE.U. counterparts can be explained (at least in part) by their higher stock and option incentives ,their findings leave open the question of why incentives are so much higher for U.S.CEOs.Top executives are significantly more highly paid and hold more equity incentives in countries with stronger insider trading restrictions (Denis and Xu (2013).
Although some recent studies have moved away from the Anglo-Saxon world (Minhat and Abdullah, 2014-Malaysia, Duffhues and Kabir, 2008-Netherlands, Elston and Goldberg, 2003-Germany, Brunello et al, 2001-Italy, Alcouffe and Alcouffe, 2000-France and, Ortín-Ángel and Salas-Fumás, 1998-Spain), the state of the art in executive compensation research still largely ignores the situation of less developed capital markets, where the owner managed firm is predominant.
It is therefore important to find out if these countries tend just to "follow the lead" of the Anglo-Saxon practices, regardless of the levels of separation of ownership and control, or if the predictions of the agency theory also apply within this context.By shedding light on this reality we also contribute to generating a more dynamic analysis of executive compensation practices and trends.Countries with less developed capital markets, which are typical in continental Europe, may be regarded as being at an early stage of separation of ownership and control which may be enhanced as the widespread development of local capital markets contributes to the rising role of public corporations.
In section II we present the main hypotheses relating to the choice of compensation mechanisms employed by firms.In section III we present the data and the methodology used to test the hypotheses.In section IV we present the empirical findings and in section V we summarize the main conclusions of the study.

Characteristics of principals
The concept of agency has been widely used to analyse relations between owners (principals) and managers (agents) within organisations.Even CEO compensation is a negotiation between CEO and a principal (Yao and Applelbaum, 2009).Agency theory in particular addresses issues of opportunism between principals and agents.
Agency theory has been developed along two closely related routes (Jensen, 1983).One route focuses on identifying situations in which the principal and the agent have conflicting goals and then describing the governance mechanism that limit the agent's self-serving behavior.Jensen and Meckling (1976) analyzed the ownership structure of the firm and how equity ownership by managers can align the interests of principals and agents.Fama (1980) analyzed the role of capital markets to discipline agent's self-serving behavior.Fama and Jensen (1983) analyzed the role of board of directors as an information gathering mechanism that the stockholders of large firms could use to monitor the agent's self-serving behavior.Jensen (1984) and Jensen and Roeback (1983) extended this latter idea to analyze controversial practices such as golden parachutes and corporate raiding.This strand of literature has been mainly concerned with describing governance mechanisms that solve the agency problem.
In many firms, managers are closely involved with key aspects of daily operations and so benefit from an information asymmetry in relations with more disengaged owners.This enables the agents (managers) to act opportunistically against the owners (principals) in the form of hidden information (adverse selection) associated with the fact that executives sometimes have hidden information that can be omitted when the company makes compensation contract to get personal advantages in the future.Another problem is hidden action (moral hazard), which is described by Katz and Rosen (1998): the principal can not observe the agent's actions and also the principal and agent agree as to what action the executive must develop.
In agency theory financial contracts and institutions can be usefully explained as efficient mechanisms for dealing with, and possibly overcome moral hazard.Moral hazard and the necessary expenditures to overcome it constitute a form of agency cost, arising from the separation of principal and agent.A principal-agent problem exists within any firm since its activities are a collection of contracts between principals and agents.This problem arises if the principal delegates some authority to the agent to act on its behalf.But if the agent has more information than the principal the latter may not get what they want because the task has been delegated to the former.Principals delegate some control over their affairs to agents who may lack incentives to act in the customer's best interests and can plead adverse selection when the outcome is poor.This situation clearly creates incentive problems because the principal cannot observe the agent's actions, or because the principal has inferior information compared to the agent.
A related problem of informational asymmetry arises from a situation in which managers seeking finance might not being able to convince the owners about the profitability of the project as claimed.Since managers with low-quality investments can gain by asserting that their intended project is of good quality, the initial claim that the investment will be profitable cannot be taken at face value.This creates the classic adverse selection problem.The principal-agent literature concentrates on the relationship between two parties (stockholders and managers) who possess different levels of information and skills with regard to the firm's operations.In the literature, agency costs are often associated with the control that a principal can exert over the agent.Control, in this case, is related to the right to monitor and impose a given set of results (Grossman and Hart, 1986).In firms where there is a clear separation between ownership and control, such as stateowned or publicly traded, principals will have lower levels of control and hence be more prone to adopt variable pay schemes.
Hypothesis 1: Executives in listed public firms will receive a higher proportion of their salaries in the form of variable pay.

Characteristics of agents
Agency theory posits that the fundamental goal of firms is to maximize efficiency.The theory suggests that firms will choose pay strategies that reduce monitoring costs.This is efficient because the principal is buying the agent's work, which is best observed on the outcome achieved.In agency theory a routine task is one in which the action can be defined more or less precisely.The theory posits that routine tasks will be positively related to behavior based pay contracts (fixed or equity pay) and negatively related to the use of outcome based contracts (variable or stock pay).This is because routine tasks allow principals to specify the behavior of agents in the contract.Agents with a higher educational background are more likely to undertake non-routine tasks and hence enter into outcome based contracts.Agents with more experience in labour markets are also more likely to be exposed to non-routine tasks and hence enter into outcome based contracts.Productivity of managers at top executive levels should be higher and also influence the productivity of subordinates leading these agents to enter into outcome based contracts.
Hypothesis 2: Executives with higher educational levels will receive a higher proportion of their salaries in the form of variable pay.
Hypothesis 3: Top executives with longer experience in labour markets will receive a higher proportion of their salaries in the form of variable pay.

Relationship between principal and agent
In agency theory principals monitor agents (though agents may also monitor principals).
In this context, principals can reduce monitoring costs by engineering trust between themselves and their agents.One way in which trust can be engineered between transacting parties is by increasing the number of transactions.Principals and agents involved in long-term transactions should generally be better informed in relation to each other's behavior and therefore more easily agree to compensation schemes based on a fixed pay rather than variable pay.Firms that do not expect a long-term relationship to develop, and hence do not expend resources in socializing their managers, put less emphasis on behavior and more on actual outcomes and would be more prone to adopt outcome based compensation schemes (Stroh et al, 1996).One way to identify a firm's expectation regarding the length of its relationship with its manager's is through its human resource policy.Employment security, clear promotion ladders, and investments in training and development are all signals to managers that the firm expects to maintain a long-term relationship.Because the principals are better able to observe their agent's behavior in long-term relationships, managers in firms with human resource policies that encourage such relationships should receive a smaller proportion of their compensation in the form of variable pay than managers in firms that do not have these policies.
Hypothesis 4: Top executives engaged in long term relationships with principal will receive a lower proportion of their salaries in the form of variable pay.
Another implication of agency theory for organizational behavior stems from risk sharing that arises when cooperating parties have different attitudes toward risk.The key issue here is that the principal and the agent may prefer different actions because of their different risk preferences (Eisenhardt, 1989).The focus of the principal-agent relationship is to determine the optimal contract, behavior versus outcome, between the two.The very simple model assumes goal conflict between the principal and the agent, and an agent who is more risk averse than the principal.The argument is that the agents are unable to diversify their employment and hence are risk averse as opposed to the principals who are able to diversify their investments and hence are risk neutral.Under complete information -when the principal can observe the agent's behavior, then a behavior based contract is more efficient.Under incomplete information -when the principal can not observe the agent's behavior, due to moral hazard or adverse selection an outcome based contract is more efficient.Mature firms will tend to control moral hazard and adverse selection problems more easily than their emerging counterparts.
Hypothesis 5: Executives in more mature firms will receive a lower proportion of their salaries in the form of variable pay.

Characteristics of firms
Agency theory has two important implications for organizational behavior.The first is the treatment of information (the second is the treatment of risk, which is addressed in the subsequent paragraph).In particular, agency theory considers information as a commodity which can be acquired at a cost.A direct implication of this is that firms can invest in information systems to control agent opportunism.One very common information system used by firms to monitor executives is the board of directors, who are able to monitor and control the compensation level offered to CEO (Petra and Dorata, 2008).From an agency perspective, boards can be used as a monitoring device for shareholder interests (Fama and Jensen, 1983).When boards provide richer information, top executives are more likely to engage in behavior that is consistent with stockholders' interests.When boards provide rich information, compensation is more likely to be behavior based.Because the behavior of executives is better known, compensation based on knowledge of executive behavior is more likely.Executives, in this context, are rewarded for taking well-conceived actions whose outcomes may be unsuccessful.
Behavior such as using greenmail and golden parachutes, which tends to benefit agents more than stockholders, is less likely when boards monitor stockholder's interests.The richness of board information can be observed from the frequency of board meetings, number of board members or board members representing a particular ownership interest..There is a large body of evidence that connects the firm size to compensation: Jensen andMurphy (1990), andGregorič et al. (2010), among others.Large firms will tend to display these characteristics more often than small firms and rely more on behavior based compensation.
Hypothesis 6: Executives in large firms will receive a lower proportion of their salaries in the form of variable pay.
Information on the agent's behavior can be acquired at a cost.This issue was addressed earlier.Volatility introduces two problems: one is the inability to plan for the future; and the other is the risk shifting.In agency theory, volatility coupled with the willingness to accept risk influences the nature of contracts between the principal and the agent.When volatility is high, the costs of shifting risk to the agent are also high, and hence behavior based contracts will be more prevalent.This is a standard assumption in agency theory.
This idea can be extended to growth.When a firm experiences high growth and high return on sales agents will be more prone to share risks with principals.
Hypothesis 7: Executives in firms experiencing high growth levels will receive a higher proportion of their salaries in the form of variable pay.
Hypothesis 8: Executives in firms experiencing high return on equity will receive a higher proportion of their salaries in the form of variable pay.
Hypothesis 9: Executives in firms experiencing high productivity levels will receive a higher proportion of their salaries in the form of variable pay.
In Lisbon there are firms with more administrative complexity, so it is to believe that firms in Lisbon tend to use more variable pay.
Hypothesis 10: Top executives in firms located in the centre of economic activity will receive a higher proportion of their salaries in the form of variable pay.

Characteristics of industry
Industry characteristics may also influence the nature of the principal-agent contract.For example, manufacturing is more capital-intensive than services.Manufacturing is also well structured into standardized routines and procedures when compared to services.A direct implication of this is that monitoring costs in manufacturing may be less acute when compared to services.The quantity and quality of the service output provided by one firm will depend more on the personal attention and diligence of agents than will the output provided by manufacturing with a mechanized production process.Incentives will be more effective in increasing output in services than manufacturing.
Hypothesis 11: Executives in service firms will receive a higher proportion of their salaries in the form of variable pay.
Inovation and R&D are being positive related to variable pay (Balkin and Gómez-Mejia, 1987;Henderson and Fredrickson, 1996).Manufacturing and services are highly heterogeneous.For example, high-tech firms will perform more research and development activities, generally more difficult to supervise, and will more frequently be subject to innovation.The need to closely integrate different functional areas will lead to a broader definition of routine and procedure programmability.In these situations, variations in the effort of managers and workers are likely to induce larger effects on performance than in non-high tech firms generally associated with more structured processes.
Hypothesis 12: Executives in high-tech firms will receive a higher proportion of their salaries in the form of variable pay.

Other
The principal-agent literature concentrates on the relationship between two parties who possess different levels of information and skills with regard to the firm's operations.In the literature, agency costs are often associated with the control that a principal can exert over the agent.Control, in this case, is related to the right to monitor and impose a given set of results.In the context of the relationship between two offices of the same firm, as is the case of a headquarters and subsidiary, or even one subsidiary and another subsidiary of a multinational, such problems may however be less acute because each of the subsidiaries belongs to the same ownership unit (Arrow, 1975).The management of foreign subsidiaries is usually ascribed to an agent.This agent is largely responsible for maintaining relationships with the parent while running the business.His origin is thus largely employed to measure centralization or delegation of control.Where he is native, it is associated with the centralisation of control; and where he is a foreigner, it is associated with the delegation of control.The rationale is that in the case of an expatriate, the multinational benefits from a larger cross-border consistency of foreign office behavior but may suffer from lower information on the foreign market as agency costs may rise exponentially with the number of autonomous subsidiaries and the number of managerial functions.
Hypothesis 13: Top executives in multinational firms will receive a lower proportion of their salaries in the form of variable pay.

III -Data and methodology
In this study, the data was collected from primary and secondary sources in contrast with a large number of studies that relied on secondary sources such as those supplied by compensation consultants.
Secondary data was extracted from 500 Largest and Best Portuguese Firms of Exame, a business newspaper.It regularly publishes data on the largest firms operating in Portugal.
It is considered to be a highly reliable source of business information but it does not disclose in-depth details as to how firms are selected.This was a source for information about the firms including location, nationality, sector, number of employees, productivity, sales growth, return on equity and firm age.
Primary data was obtained through a survey e-mailed to the human resource director of each 500 firm, in order to obtain information about how executives in Portugal are being paid: the repartition of total remuneration in fix and variable remuneration and fringe benefits.In this survey we asked information on CEO, CFO and Commercial Director.
The questionnaire was simple in order to maximize the response rate and consistency.In fact we tried that human resource directors took no longer than five minutes to answer the survey.In the survey, was asked collected executive specific data, including age, experience in the firm and education.
We first called the 500 firms to ask for the email address of the Human Resource Director.Then the questionnaire was administered by e-mail with no attachments.We made a phone follow up to enhance the number of answers.Of the 500 firms approached, by email, 104 provided answers.However, 3 firms were discarded for providing incomplete questionnaires.The sample consists of 101 firms.It represents about 20% of the overall population.
The data set consists of one dependent variable -the percentage of variable compensation paid by the firm and 13 explanatory variables (for each of the hypotheses stated in the previous section).
The dependent variable used in the study is the percentage of variable compensation on the total compensation which is similar to Stroh et al (1996) as it appeared less sensitive to respondents than alternative measures used elsewhere: log wage (Ewing, 1996) or bonus payment (Ortín-Ángel and Salas-Fumás,1998;Bruce et al, 2007).
The characteristic of the principal was measured through a dummy (X1) variable that takes a value of 1 if the firm is listed or publicly owned and 0 otherwise.The characteristics of the agents were captured through two variables: a dummy (X2) that takes a value of 1 if the manager has a university degree and 0 otherwise; and variable (X3) that measures the age of the executive.The relationship between the principal and the agent is captured through a variable that measures the number of years of the existing contract between the two (X4) and the maturity of the firm (X5) in number of years since its foundation.
Firm-specific factors were captured through four variables: a variable that measures the size of the firm in number of employees (X6); a variable that measures the growth in sales in relation to the previous year (X7); a variable that measures the return the on equity ratio (X8); a variable that is measured as the ratio of value added over the number of employees (X9) as a proxy for productivity; and a dummy (X10) that takes a value of 1 if the firm is located in the centre of economic activity, in this case Lisbon, and 0 otherwise.Industry-specific factors were captured through two variables: a dummy (X11) that takes the value of 1 if the firm belongs to the service sector and 0 to the manufacturing sector; and another dummy (X12) that takes a value of 1 if the firm is high-tech and 0 otherwise.Because the data set consists of both domestic and multinational firms an additional dummy (X13) was created to capture differences between these two groups of firms: it takes a value of 0 if the firm is multinational and 1 if it is domestic.
The descriptive data relating to these variables are summarized in Table 1.The table shows that firms in Portugal compensate their managers on average with 13.5% in the form of variable pay.This figure is much lower than the average observed in markets with a larger presence of listed firms.Jensen, Murphy and Wruck (2004) shows that since the mid nineties variable compensation of the S&P 500 CEOs has largely exceeded fixed compensation.
Table 2 displays the bivariate correlations between the variables in the data set.The table shows that the variables in the data set are not linearly dependent.The highest correlation (-0.54) was observed between nationality (X13) and ownership (X1), showing that multinational firms have a larger propensity than domestic firms to be of the listed type.
The basic model used to analyse the responses to the questionnaire is of the linear form where i Y is the ith firm ( i=1…101), ji X is the jth characteristic ( j=1…13) of the ith firm and α , j β are fixed coefficients.The stochastic terms i ε are assumed to be independently and identically normally distributed, in contrast to the ji X , that are fixed.
Since the dependent variable is bounded to the left we used a Tobit procedure to estimate the basic equation.

IV -Empirical findings
The main empirical findings of the study are displayed in Table 3.The estimation shows that public corporations are more intensive users of variable pay.
This was the strongest discriminator for variable versus fixed compensation found in this study, consistent with many theoretical and empirical studies on this topic.Separation of ownership and control enhances agency problems, requiring co-alignment of interest as also observed recently in Ang et al (2000) that finds a significant association between separation of ownership and control and agency costs in a sample of 1708 small American corporations.Our result is also consistent with Bebchuk and Fried's (2003, p. 21) prediction that "managerial power substantially affects the design of executive compensation in companies marked by a separation of ownership and control".Indeed, as variable pay has led to a significant rise of the compensation package, it is in the executives' interest to enforce its implementation.
We also find that the higher the education level the higher the propensity to receive a larger amount of salary in the form of variable compensation.Stroh et al (1996), find no significant influence of education in the design of compensation mechanism in 29 of the Fortune 500 firms.Contrary to the theory based prediction, firms with lower productivity levels were found to pay higher levels of salary in the form of variable compensation suggesting that firms use the compensation mechanism to boost productivity levels.
Whether in practise this is achieved or not is a matter for further empirical inquiry.
As predicted, firms located in Lisbon, pay higher levels of salary in the form of variable compensation.By contrast, age of CEO, duration of contract and age of firm show no statistically significant relationship with the level of variable pay as also observed in Stroh et al (1996).Firm size, sales growth and return on equity also showed no significant impact on level of variable pay unlike observed in Jensen and Murphy (1990) and Leonard (1990).On the contrary, Abdullah (2006) found no association between directors remuneration and firm profitability.
Surprisingly, factors relating to the characteristics of the industry show mixed signs: a negative relationship between the level of variable and the service industry and positive relationship between the level of variable and the high-tech industry.Whilst the relationship with the service sector is not significant at a statistically meaningful level, the relationship with the high-tech sector is found to be significant at the 22% level.The findings suggest that it is certainly desirable to pursue the analysis of the influence of industry on the design of compensation mechanisms.Our findings also reveal the level of variable pay to be positively related to domestic firms suggesting that agency costs may be acute in these rather than multinational firms.
Overall, the findings of our study confirm that although the average weight of the variable pay for the largest Portuguese firms is relatively low compared to US firms, ownership is a strong determinant in the design and level of compensation.

V -Summary and conclusions
This paper shows that agency theory is particularly attuned to the analysis of situations in which contracting problems are difficult.These include situations in which there is a substantial goal conflict between principals and agents such that agent opportunism is likely.By emphasizing these issues, the paper uses agency theory to deduct testable hypotheses and generates empirical findings consistent with this theory.Other issues such as compensation in high tech and service firms, where monitoring is particularly difficult, are also addressed.
The study was carried out in the context of a small economy with a developing capital market, shedding new light on an issue hitherto predominantly studied in the context of Anglo-Saxon economies, namely the US.Lack of off the shelf data has been the main deterrent to research in the context of less developed capital markets.
In this study we collected data through a questionnaire emailed to the 500 largest non financial Portuguese firms.Despite the limitations inherent to the small size of the sample, we could confirm that the findings of previous empirical studies are partially confirmed in the Portuguese context.
We found that listed and publicly owned firms pay higher levels of variable compensation which is consistent with Ang et al (2000) and Bebchuk and Fried (2003).
We also found that the higher the education level the higher the propensity to receive a larger amount of salary in the form of variable compensation.This contrasts with Stroh et al (1996), who found no significant influence of education in the design of compensation mechanism.Firms located in Lisbon, pay also higher amounts in the form of variable compensation.Contrary to the prediction, firms with lower productivity levels were found to pay higher levels of salary in the form of variable compensation.
There are limitations as to the extrapolation of our results as we can not measure the level of potential idiosyncrasy.Ideally, the study should be replicated in different contexts to control for country specific influences.Nevertheless, the main finding that performance related pay mechanisms are less used in countries where public corporations and potential agency problems are less pervasive should hold.

Firms
generally operate in volatile environments.The future can reserve either success or failure, or an intermediate outcome.Firms are said to operate under volatile conditions because the economic environment, government policies, competitors, technological change, and so on, may cause uncontrollable variations in the firm's profit function.

Table 1
Descriptive statistics of variables in the data set Std

Table 2
Bivariate correlations of variables in the data set

Table 3
Tobit estimation for the size of variable compensation Y *** Significant at the 1% level; * Significant at the 10% level