BOARD REMUNERATION, PERFORMANCE, AND CORPORATE GOVERNANCE IN LARGE SPANISH COMPANIES. Rafel Crespí (UAB,UIB) Carles Gispert (UAB) Version December1998 Departament d'Economia de l'Empresa Universitat Autònoma de Barcelona Edifici B 08193 Bellaterra Phone: +34 3 581.14.51 Fax: +34 3 581.25.55 E-Mail: [email protected]E-Mail: [email protected]
29
Embed
board remuneration, performance, and corporate governance in ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
BOARD REMUNERATION, PERFORMANCE, ANDCORPORATE GOVERNANCE IN LARGE SPANISH
COMPANIES.
Rafel Crespí
(UAB,UIB)
Carles Gispert
(UAB)
VersionDecember1998
Departament d'Economia de l'EmpresaUniversitat Autònoma de BarcelonaEdifici B 08193 BellaterraPhone: +34 3 581.14.51Fax: +34 3 581.25.55E-Mail: [email protected]: [email protected]
2
BOARD REMUNERATION, PERFORMANCE, AND CORPORATE GOVERNANCE
IN LARGE SPANISH COMPANIES ..
Abstract:
In this paper we present empirical evidence on the relationshipbetween board remuneration of a sample of large Spanish companiesand a set of explanatory variables such as performance and size ofthe company. The objective is to contrast the existing evidence basedon the agency theory in an institutional environment with differencesfrom the literature.
We focus on the impact of company’s governance structure on therelation between pay and performance. Specifically we consider theownership concentration and the firm leverage as key determinat ofthe board-shareholders relationship.
Our results confirm the positive relationship between boardremuneration and company performance, which is stronger for bookvalues than for stock market measures. Industry performance alsoexplains the remuneration and provides useful information toevaluate board behaviour. Company size is also related to boardremuneration, and affects the pay-performance relationship, althoughis not relevant when we use an elasticity approach. Finally, thegovernance structure of companies is relevant when explaining thepower of the compensation-performance relationship, and differencesbetween the impact of ownership concentration and firm leverage onthis relationship are found.
JEL Classification: G3, J33
Keywords: Board Remuneration, Company Performance, Firm Size, Corporate
Governance.
The authors thank on Vicente Salas and Luc Renneboog for their comments toprevious versions of the paper. We also thank the participants, specially to MartinConyon, at Workshop on Corporate Governance, Contracts and Managerial Incentives(Berlin, July 1998) and Cambridge (1998) workshop on Corporate Governance. Theusual disclaimer applies. This work has received financial support from the DGCYT,Project Number PB-94-0708.
3
1. Introduction
What determines the company board’s remuneration? Are the board compensation
schemes related to the company performance figures or just to firm size? Is the
governance structure important when explaining the relationship between board
remuneration and company performance?
Some answers to these questions have been given in papers by Murphy (1985) or Gregg
et al (1993), who related executive compensation to corporate performance. Jensen and
Murphy (1990) estimate the pay-performance relationship for executives. Main et al.
(1996) relate boardroom pay and company performance, and in a recent study by
Conyon (1997) the impact of corporate governance innovations on top director
compensation is analysed. The mentioned studies link the performance of US and UK
companies with the remuneration of their top managers. The economics of these finding
are the existence of incentive contracts that reduce costs associated to the agency
relationship between shareholders and managers. The results are consistent with the
predictions of the agency model.
Only one of these papers (Main et al., 1996) considers the total board remuneration as a
magnitude of the pay side in its relationship with performance. This approach is
justified by the redefinition of the agency problem. On the principal side there are
shareholders and the agent is board of directors. The board has the attribution to
allocate company resources, acts on behalf of shareholders (the principal), and is only
under control of the stock market, the products market and shareholder meetings. Our
approach, is based on the legal rules that make company managers responsible in front
of the board, and the board responsible toward the shareholders.1 We have an additional
justification to use the total board compensation instead of top executives remuneration,
which is the availability of data that will be detailed in the Data and Methodology
section.
1 It’s important to underline here that in Spanish firms, the CEO and top executives are
usually board members. This category is known as insider board members, opposite to
outsider board members, who do not perform executive tasks in the company.
4
There is also empirical evidence on the impact of governance characteristics on the pay-
performance relationship. Conyon (1997) justifies the introduction of governance rules
and characteristics in the compensation-performance dependence in order to identify the
degree of potential agency problems and to evaluate how successful remuneration
schemes are in curbing this kind of problems. Ownership concentration and firm
leverage are two components of the company governance structure that we consider. An
extension of the model developed by Salas (1992) presents a theoretical framework in
which some corporate governance characteristics can be a substitute for compensation
schemes based on company performance. The aim is to detect, as the agency theory
predicts, differences in the determinants of board compensation depending on the level
of supervision that shareholders or creditors exert over the management.
Our empirical study is based on a data set of large and listed Spanish companies during
the period from 1990 to 1995. Peña (1978) and Ortín (1997) are examples of empirical
papers in Spain, and they focus on the company size as a key point to justify the
managers’ remuneration. The Spanish institutional environment as a thin, but growing,
stock market and the presence of significant shareholders with important ownership
participations allows us to compare the results of this "non market-oriented system"
with the "Anglo-Saxon system". Effectively, the conclusions of the literature come from
research applied to UK and USA data, where the level of ownership dispersion is higher
than in continental Europe and Japan. Therefore the control mechanisms of the agency
problem can rely more on market devices, like takeovers, than on direct contractual
incentives between shareholders and managers. In our context we expect board
remuneration to be less performance-dependent to the extent that governance structures
differ. It is also important to know to what extent these determinants of board
remuneration differ among corporate control systems.
Our findings support the hypothesis that performance explains board remuneration. We
also find a positive relationship between company size and board remuneration. This
positive relation is not significant when the link is between changes in compensation and
changes in company size. Finally, governance structure in terms of ownership
concentration is relevant to explain the remuneration performance relationship, but is
not relevant in terms of firm debt leverage.
5
The paper continues developing the theoretical framework. Section 3 is devoted to data
and methodology explanations, and the results are presented in Section 4. The paper
concludes in Section 5.
2. Determinants of Board Remuneration.
2.1 Remuneration and Company Performance.
From a theoretical perspective, the design of compensation contracts for the company
managers in an agency context has been an important subject in the microeconomic
literature2. The problem to solve is the determination of an optimal compensation
scheme that motivates managers to make a maximum effort, taking into account that
managers are risk-averse, and that contract is done in a context of asymmetric
information. The conceptual framework of the agency theory provides a set of useful
elements to evaluate the manager-shareholder relationship.
The assumption of non-observable costs for both parts in the relationship presents a key
difference from the neo-classical postulates (see Hart 1995). In a principal-agent model
the level of effort of the agent is private information, which is a basic condition for a
potential opportunistic behaviour. The model explains the outcome (y) of a company as
a function of the real effort carried out by mangers (e) and a set of variables which are
not under a manager’s control which are assumed randomly distributed (ε).
y = f(e,ε)
To motivate the agent’s best course of action, and punish him if fails to do so, the
principal should offer a contract where the manager’s compensation is based on
observable outcomes. The outcome has to be highly correlated with the non-observable
measure (e).
Tirole (1988), Rosen (1992) and Hart (1995) justify the use of a contract where
manager compensation (w) is based on an observable outcome of the company
w = g(y)
2 See, for example, Ross (1973), Mirrlees (1976) or Grossman and Hart (1983).
6
There is the assumption that effort (e) and results (y) move together, otherwise the
observability of y would be irrelevant in pushing managers toward higher effort levels
(e). In case of risk-neutral agents, the first best solution would be a contract with the
following form
w(y) = y - F
where shareholders sell the company to the agents, who would pay a fixed amount F to
them. With risk-averse agents, the second-best solution is a contract with a trade-off
between incentives and risk. If the aim of the contract is to achieve a higher level of
effort, the incentives should be very sensitive to results, which means that the risk borne
by the agent is very high. On the other hand, if the objective is to minimise the risk
borne by managers, the incentive contract would not be connected with performance
measures, and effort and motivation are values expected to be lower.
An expected consequence is that risk-averse agents would only accept expensive
contracts that compensate for the risk they bear. The amount of this premium could be
justified by the individual risk aversion coefficient and by the implicit risk of the
company performance measure.
The theory predicts a positive relationship between manager’s remuneration (measured
in different ways) and company results or performance. Several empirical papers
present evidence in this sense. Murphy (1985), Jensen and Murphy (1990) Gregg et al
(1993), Conyon and Peck (1996), Conyon (1997) and Main et al (1996), are relevant
references that support the previous reasoning.
Our approach considers board members as the first level of delegated authority from
shareholders in the agency relationship. They have to report the outcome of their
decisions to the shareholders and according to the incentive constrain, we expect that
their compensation will be linked to company performance measures.
Hypothesis 1: There exists a positive relationship between company performance and
board remuneration.
One of the factors which can affect compensation-performance based contracts is the
noise level of the performance measure. Holmstrom (1979) shows how the principal is
always better off with more information about agent behaviour than with less
information. The introduction of new observable variables (z) to the former contract
7
w = g(y), and a more accurate measure of the real effort applied by managers would
reduce the associated risk to the performance measure, decreasing the variance of
observable outcomes. Additional information can be introduced in the compensation
contract from the behaviour or performance of comparable managers in companies with
similar activities. Our contract would be then
w = g(y,z)
Gibbons and Murphy (1990) show evidence about the relative importance of the
evaluation of a manager’s behaviour. Poor company results can be caused by external
factors, which affect companies belonging to the same industry. In this case, a correct
evaluation of management effort would consider not only the company performance
figures, but also industry measures of performance (see also Lazear 1995).
According to this we expect a negative relationship between board compensation and
industry performance measures, balancing the positive relationship of the remuneration
and company performance. Empirical evidence with data sets of UK and US companies
is provided by Gibbons and Murphy (1990), Conyon and Leech (1994) and Conyon
(1997).
Hypothesis 2: There is an inverse relationship between industry performance and
board remuneration.
2.2 Remuneration and Firm Size.
Company size has usually been introduced in empirical research as a controlling
variable to explain levels of manager compensation.3 Empirical findings are supported
on the Rosen’s (1992) argument that the cost of wrong decisions and the benefits of
correct managerial behaviour are higher the larger is the company. A consequence is
that managers should be better paid to the extent that the company is larger. Larger
firms may employ better-qualified and better-paid managers. Anyway, as Murphy
(1998) exposes, recent data suggest that the positive relation between CEO pay and
company size has weakened over time, although it remains positive. For companies of
3 See Murphy (1985), Jensen and Murphy (1990), Gregg et al (1993) and Conyon
(1997). Also see Ortín (1997) for the Spanish case.
8
significantly different sizes, we expect that larger firms will pay more to their board
members than smaller ones.
Hypothesis 3: There is a positive relationship between company size and board
remuneration
A further step is to evaluate the impact of firm size on the board remuneration-firm
performance relationship. As Murphy (1998) and Gibbons and Murphy (1992) assert
empirically, the pay-performance relationship is smaller in large firms. They state that
the optimal pay-performance relation may decline with firm size for two reasons: (i) the
variance of changes in shareholder wealth increases with firm size, and (ii) the CEO’s
direct effect on firm value may decrease with firm size (Gibbons and Murphy, 1992, p.
489). The underlying proposition of the second point is the incentive principle of
Milgrom and Roberts (1992), where the level of performance incentives depends on the
marginal contribution of manager effort to company profitability. For larger firms is
more difficult that additional manager efforts will have the same positive impact on the
principal outcome as we could expect in smaller firms. For this reason, it will not be
worthy to introduce high-powered incentives that will only serve to increase the agent’s
exposure to income variations.
Hypothesis 4: The positive relationship between board compensation and company
performance is expected to be lower for larger companies than for
smaller ones.
Furthermore, Gibbons and Murphy (1992) also show that while the pay-performance
relationship varies significantly with firm size, the pay-performance elasticity is almost
invariant to firm size. The sensitivity approach in the pay for performance relationship
establishes the effect of increases of one monetary unit in the performance measure over
the shareholder's wealth amount. On the other side an elasticity approach in the pay
performance relationship does not require a correction for firm size (Murphy,1998). We
follow this our approach, as we comment in the next section.
2.3 Remuneration and Governance structures
A contract that establishes a compensation scheme depending on the company results
would be optimal in the absence of transaction costs. In a world of incomplete
contracts, where agency problems arise, the companies’ governance structures play an
9
important role. They are important as a decision-making mechanism for situations that
have not been specified in the initial contract (Hart 1995). These contracts contribute to
overcome the agency costs that appear from limited specifications in the initial
negotiation.
Governance peculiarities come from the institutional rules and markets where the
company operates and form the company governance characteristics. Institutional and
market characteristics such as the development of financial markets, the activity of the
market for the corporate control, the importance of banks as monitors, the legal rules
protecting the shareholder’s investments or the role of institutional investors are
important. The firm specific governance conditions refer, among others, to the structure
and composition of ownership, the creditor's control or the board composition. Both
kind of attributes are able to influence the board-shareholders relationship moderating
the problems of information asymmetries, increasing the supervisory activity or
avoiding free riding behaviour.These variables can influence the way in which control
and supervision of company managers is exerted. The higher the control and supervision
intensity, the higher will be the knowledge about the real effort made by managers in the
performance of their contracts.
This leads us to hypothesise that board remuneration will be influenced by the
governance structures. In companies where shareholders have less control over
managers (manager-controlled companies in the terminology of Berle and Means), the
discretionary power for managers will be higher, and a more powerful compensation
performance scheme is need. The impossibility to observe makes it more dependent upon
an indirect yardstick which is observable.
The justification is based on the idea that the impossibility to observe directly (and
supervise) the managers’ behaviour makes their remuneration more dependent on an
observable measure of their effort, which could be company performance. On the other
hand, in governance structures where shareholders can directly supervise4 the managers’
decision process, the compensation-performance relationship could be weaker. In this
4 In this context, there is the case where managers are board members and significant
shareholders. Boards of companies with highly concentrated ownership usually
represent an important proportion of the outstanding equity.
10
case shareholders would have superior information about the managers’ effort and the
remuneration could depend on the observation of the real variable (effort) in the
contract.
Salas (1992) presents a model of incentives and supervision where a risk-neutral
principal (in our case the shareholders) contracts a risk-averse agent (managers or
board members). The principal offers this efficient contract to the agent5
R y x y x( , ) = + +α β µ
where R is total compensation, βα, and µ are parameters to estimate when solving the
problem of maximising the principal expected return. Two related measures give
information about the real effort made by the agent: the company outcome (y) and the
manager's observed effort (x). Both measures provide imperfect information about real
effort because some noise randomly affects this. The following expression considers this
fact
2
1
εεδ
+=++=
exey
where e is the real effort, δ is a constant and ε ε1 2, are random variables with a mean of
0 and variances σ σ12
22, , respectively. These random variables include, among other
factors, the mentioned lack of precision in the measures.
The solution to the problem requires the following condition6
βµ
σσ
∗
∗ = 22
12
This necessary condition indicates that higher supervision, when reducing the variance
σ22 , also reduces the weight assigned to β , so compensation should be less
performance-dependent. On the other hand, a weak supervision would mean that
5 According to Wilson (1968), cited in Salas (1992).
6 For more details, see Salas (1992) , pp. 133.
11
compensation ought to be highly correlated with company performance, and risk is
transferred from the principal to the agent.
This theoretical approach provides us with a starting point to evaluate the expected
impact of governance mechanisms, such as supervision devices, on the compensation of
board members.
The level of ownership concentration can be considered as a proxy of intensity of
shareholders supervision. Lower levels of ownership concentration increase the
probability of shareholder’s free rider behaviour, decrease ownership control and allow
for higher discretionary power (less supervision) for board members. The expected
relationship is as follows:
Hypothesis 5: The intensity of the board pay for performance relation is inversely
related to the ownership concentration
The shareholder’s control of managers is complemented with the creditors ability to
monitor the debt contracts. In this sense the board members obligation to fulfil an
enforceable contract decreases the possibility of opportunistic behaviour. This also
reduces the amount of cash flow that they can freely allocate (Jensen, 1986). We
consider the firm leverage as a proxy of intensity of creditors supervision and free cash
flow availability, and lower levels of leverage can be interpreted in terms of higher
discretionary power for board members, and we expect the following relationship:
Hypothesis 6: The intensity of the company performance and board compensation
relationship is inversely related to the firm leverage
3. Data and Methodology
Connecting the six hypotheses introduced in a testable model lead us to consider the
explanatory variables of board remuneration as performance, size and governance
attributes. Governance characteristics and also company size are explanatory to the
board remuneration-performance relationship. Thus, the board remuneration equation
Collinearity statistics: all variables have a tolerance above 0.3Standard error in parentheses.*p<0.1, **p<0.05, ***p<0.01
24
Table 4
Board compensation and firm size
Qi= Quartiles of sales turnover (Q1 is the lower)
Q1 Q2 Q3 Q4
Total boardcompensation(average in MMpts.)
66.71 94.65 129.93 374.17
Board membercompensation(average in MMpts.)
8.83 9.35 12.98 23.69
Board members 7.55 10.12 10.01 14.65
N 124 125 124 125
25
Table 5
Pay-performance sensitivity to firm size
Qi= Quartiles of sales turnover (Q1 is the lower)
Q1 Q2 Q3 Q4
Total boardcompensation/Bt-1*
0.1022 0.0849 0.0473 0.0200
Board membercompensation/Bt-1
0.0138 0.0099 0.0043 0.0017
Total boardcompensation/Vt-1**
0.0147 0.0069 0.0044 0.0028
Board membercompensation/Vt-1
0.0028 0.0010 0.0006 0.0003
N 125 125 125 125
*Bt-1= Benefits before taxes one year lagged**Vt-1= Firm market value one year lagged
26
Table 6
Pay-performance sensitivity to ownership concentration
Qi= Quartiles of ownership concentration (Q1 is the lower)
Q1 Q2 Q3 Q4
Total boardcompensation/Bt-1
0.0837 0.0650 0.0667 0.0168
Board membercompensation/Bt-1
0.0115 0.0055 0.0095 0.0005
Total boardcompensation/Vt-1
0.0144 0.0059 0.0053 0.0057
Board membercompensation/Vt-1
0.0025 0.0008 0.0010 0.0010
N 141 141 141 142
Table 7
Pay-performance sensitivity to leverage
Qi= Quartiles of leverage (Q1is the lower leverage)
Q1 Q2 Q3 Q4
Total boardcompensation/Bt-1
0.0386 0.0693 0.0525 0.0764
Board membercompensation/Bt-1
0.0005 0.0087 0.0053 0.0096
Total boardcompensation/Vt-1
0.0066 0.0091 0.0040 0.0103
Board membercompensation/Vt-1
0.0012 0.0019 0.0005 0.0015
N* 81 82 81 82
* Banks are not included
27
Table8Impact of firm leverage on board compensation-firm performance relationship
Dependent variable: ∆lnREMit
LEVt<Median LEVt>Median
∆ln(1+ROAi,t-1) 0.029(0.105)
0.499***(0.083)
∆ln(1+AIROAi,t-1) -0.378*(0.226)
-0.187(0.415)
∆ln(1+ROAi,t) 0.208***(0.067)
0.068(0.083)
∆ln(1+AIROAi,t) -0.168(0.175)
0.306(0.288)
SHRi,t-2 0.051***(0.013)
0.052***(0.019)
AISHRi,t-2 -0.056**(0.027)
-0.061(0.044)
∆ln(St) -0.002(0.006)
0.048(0.029)
Year Dummies Yes Yes
Observations 118 106
F 3.966*** 6.104***
R2 0.270 0.391
Test of structuralchanges (Chow, F)
2.2787**
Standard error en parentheses*p<0.1, **p<0.05, ***p<0.01
28
6. REFERENCES:
CONYON, M.J. (1996): “Directors’ pay and turnover: an application to a sample of largeUK firms”, mimeo.
CONYON, M.J. (1997): “Corporate governance and executive compensation”.International Journal of Industrial Organization, 15, pp. 493-509.
CONYON, M.J. and D. LEECH (1994): “Top pay, company performance and corporategovernance”. Oxford bulletin of economics and statistics, vol.56 nº3, pp 229-247.
COUGHLAN, A. and R. SMITH (1985): “Executive Compensation, ManagementTurnover, and Firm Performance: An Empirical Investigation”. Journal ofAccounting and Economics, 7 (1-3), pp. 43-66.
GIBBONS, R, and K.J.MURPHY (1990): “Relative performance evaluation for chiefexecutive officers”. Industrial and Labor Relations Review, vol.43, Special Issue.pp. 30S-51S.
GIBBONS, R, and K.J.MURPHY (1992): “Optimal Incentive Contracts in the Presence ofCareer Concerns: Theory and Evidence”. Journal of Political Economy, 100 (3), pp.468-505.
GREGG, P., S.MACHIN, and S.SZYMANSKI (1993): “The Disappearing RelationshipBetween Directors’ Pay and Corporate Performance”. British Journal of IndustrialRelations, 31:1 March, pp.1-9.
GROSSMAN, S. and O. HART (1983): “An Analysis of the Principal-Agent Problem”.Econometrica, 51, pp. 7-45.
HART, O. (1995): “Corporate governance: some theory and implications”. The EconomicJournal, 105 (may), pp. 678-689.
HOLMSTROM, B. (1979): "Moral Hazard and Observability". Bell Journal of Economics,vol.10, nº1, pp. 74-91.
HOLMSTROM, B. and MILGROM, P. (1991): "Multitask Principal-Agent Analyses:Incentive Contracts, Asset Ownership, and Job Design". Journal of Law andEconomic Organization, vol. 7, supl., pp.24-52.
JENSEN, M.C. (1986): "Agency Costs of Free Cash Flow, Corporate Finance, andTakeovers". American Economic Review, vol.76, nº2, Maig, pp. 323-329.
JENSEN, M. and MURPHY, K.J. (1990): "Performance Pay and Top-ManagementIncentives". Journal of Political Economy, vol. 98, nº 2, pp. 225-264.
29
LAMBERT, R. AND D. LARCKER (1988): “An Analysis of the Use of Accounting andMarket Measures of Performance in Executive Compensation Contracts”. Journal ofAccounting Research, 25, pp. 85-129.
LAZEAR, E. (1995): Personnel Economics, MIT Press, Cambridge, MA.
MAIN, B., A. BRUCE and T. BUCK (1996): “Total board remuneration and companyperformance”. Economic Journal, 106, pp.1627-1644.
MILGROM, P. and ROBERTS, J (1992): Economics, Organization and Management.Prentice Hall.
MIRRLEES, J. (1976): “The Optimal Structure of Incentives and Authority within anOrganization”. Bell Journal of Economics, 7, pp. 105-131.
MURPHY, K.J. (1985): “Corporate performance pay and managerial remuneration: anempirical analysis”. Journal of Accounting and Economics, 7, pp.11-42.
MURPHY, K.J. (1998): “Executive Compensation”. Forthcoming in O. Ashenfelter and D.Card (eds.), Handbook of Labor Economics, Vol. 3, North Holland.
ORTÍN, P. (1997): “El tamaño de la empresa y la remuneración de los máximos directivos.Evidencia empírica para el caso español, años 1990-1992”. Revista de EconomíaAplicada, nº 13, vol.V.
PEÑA SÁNCHEZ DE RIVERA, D. (1978): “Modelos explicativos de las diferenciassalariales de los directivos españoles”. Economía Industrial. Marzo, pp.51-61.
ROSEN, S. (1992): “Contracts and the Market for Executives”, in Werin and Wijkander,(ed): Contracts Economics, Basil Blackwell, pp. 181-217.
ROSS, S. (1973): “The Economic Theory of Agency: The Principal’s Problem”. AmericanEconomic Review, 63, pp. 134-139.
SALAS FUMÁS, V. (1992): "Incentivos y supervisión en el control interno de la empresa:Implicaciones para la concentración de su accionariado". Cuadernos Económicos deICE, nº 52, 3, pp.127-145.
TIROLE, J. (1988): The theory of industrial organization. MIT Press.
WILSON (1968): “The Theory of Sindicate”. Econometrica.