Does Ownership Structure Affect the Motivation of Compensation Contract of Earnings Management? —From China’s Data

doi:10.4236/jssm.2011.43041

Paper Menu >>

Journal Menu >>

Journal of Service Science and Management, 2011, 4, 351-356

doi:10.4236/jssm.2011.43041 Published Online September 2011 (http://www.SciRP.org/journal/jssm)

351

Does Ownership Structure Affect the Motivation of

Compensation Contract of Earnings Management?

—From China’s Data

Jing Long, Yanxi Li, Xiuwen Xu, Linlin Fu

School of Management & Economics, Dalian University of Technology, Dalian, China.

Email: [email protected], [email protected], {xxw133, franfulinlin}@hotmail.com

Received July 10th, 2011; revised August 16th, 2011; accepted August 25th, 2011.

ABSTRACT

This paper is to explore that if ownership structure exactly affects on earnings management, especially on how to con-

strain executive compensation motivation to earnings management in the emerging market of China. We use five vari-

ables to quantify various corporate governance mechanisms including ownership structure and executive compensation

for 1024 listed firms sample within two years. We find that earnings management is positively affected by executive

compensation motivation, while this influence is constrained mostly by ownership concentration. This finding reflects

the current situation for the weak corporate governance in the emerging market of China. Therefore, such a significant

relationship is useful to control level of earnings management and improve the China’s emerging market developing

healthily, stably and harmoniously.

Keywords: Ownership Concentration, Executive Compensation, Earnings Management, Corporate Governance

1. Introduction

Over three hundred years ago, Adam Smith raised the

issue of the separation of ownership and stewardship in

joint-stock corporations in his masterwork “The Wealth

of Nations”. It was therefore suggested that a set of ef-

fective mechanisms should be in place to resolve the

conflict of interests between firm owners and managers.

The early 2000s were the years in which a numbers of

high-profile corporate financial frauds took place, in-

cluding large companies such as Enron, WorldCom,

Tyco, Global Crossing and others [1]. To respond to

these corporate financial frauds, the US Congress en-

acted the Sarbanes-Oxley (SOX) Act in 2002 and intro-

duced a new era of corporate governance, including re-

quirements for auditor independence, independence of a

firm’s audit committee, the responsibility of a firm’s

CEO and CFO on financial reports, and the protections

on whistleblowers. Since the occurrence of financial

frauds and the enactment of SOX in early 2000s, corpo-

rate governance has become a crucial mechanism for

government regulations on both corporate and capital

market operations.

Berle and Means [2] believed that when cooperate

shares are widely spread over to a great number of small

shareholders, there is a good separation between the

ownership and the power of operational management, in

which the managers act as the agents for the shareholders.

However, Fama [3], and Jensen and Meckling [4] argued

that when the managers do not own a great number of

shares, they may pursue personal interest at the expense

of the interest of shareholders while making managerial

decisions. Therefore, the business in a form of corpora-

tion exists with interest conflicts between owners and

managers, which is referred as traditional agency theory

or equity agency problem. Moreover, this equity agency

problem would become a central agency problem, in

which new conflicts arise between controlling and non-

controlling shareholders when mangers also own signifi-

cant amount of shares through stock options, pyramidal

ownership structure, or crossing holdings [5,6].

Healy [7] has empirically researched on bonus plan

and earnings management. He found that the managers

would choose the correspondingly policies and process to

lower the report earnings if the bonus plan proposed the

minimum and maximum process to lower the report

earnings. While the managers would choose the opposite

policies to increase the report earnings, as there is no

limitation on bonus plan.

Does Ownership Structure Affect the Motivation of Compensation Contract of Earnings Management?

352

—From China’s Data

Watts and Zimmerman [8] found that corporate gov-

ernance attributes help investors by aligning the interests

of managers with the interests of shareholders and also

by enhancing the reliability of financial information and

the integrity of reporting process. It confirmed that the

mechanism of corporate governance helps restrict man-

gers in the behavior of earnings management.

The purpose of this paper is to explore the mechanisms

of corporate governance used by listed firms in the Chi-

nese emerging capital market, and the effect of govern-

ance and ownership on earnings management, especially

on how to constrain executive compensation’s motiva-

tion to earnings management. The results of this research

will help users better analyze and understand the finan-

cial statements prepared by the Chinese listed firms. In

addition, our results will also help regulators and policy

makers in policy making, enactment of regulations/laws

and their enforcement.

In addition of this section of Introduction, the remain-

der of this paper is organized into six more sections. Sec-

tion 2 presents research framework and hypothesis. Sec-

tion 3 discusses the methodology used in this research.

Section 4 presents data, descriptive statistics and the re-

search results. Section 5 provides the sensitivity test and

section 6 proposes conclusion and suggestion.

2. Research Framework and Hypothesis

Dennis and McConnell [9] defines corporate governance

as “… the set of mechanisms that maintain an appropri-

ate balance between the rights of shareholders … and the

needs of the board and management to direct and manage

the corporation’s affairs.” The corporate governance

mechanisms may help resolve the two sets of conflicts:

between owners and managers, and between controlling

shareholders and minority shareholders. It consist own-

ership structure, board of directors, executive compensa-

tions, financial disclosure, etc. In this paper, we consider

and assess ownership structure (e.g. concentration own-

ership), board of directors (e.g. board size, proportion of

independent directors, the duality of CEO), and the man-

agers (e.g. the top3 executive compensations), which

effect on earnings management.

2.1. Executive Compensation

We define the top managers as CEO, the chairperson of

the board of directors or supervisors, and the general

manager of a firm. Watts and Zimmerman [8] stated that

managers have incentives to advance the earnings from

the future to the current accounting period when a bonus

award plan exists. In their later studies, Dechow and

Sloan [10] consistently confirmed that managers would

use accounting judgments to increase earnings-based bo-

nus awards. We define sata-3 is the annual salaries of the

3 top managers. The following hypothesis flows from the

literature discussed as:

Hypotheses 1: The executive compensation (sala_3)

has a negative impact on earnings management.

2.2. Ownership Structure

It is believed that one of the most important ways through

which a firm maximizes its value is through well-de-

signed ownership structure of the firm’s shares. Concen-

trated equity ownership can be bad for the governance of

the firm since it gives the largest shareholders too much

discretionary powers of using firm resources in ways that

serve their own interest at the expense of other share-

holders. That is, too much concentrated ownership (the

largest shareholders) may accentuate the earnings man-

agement we mentioned earlier.

We investigate the related issue of how ownership

concentration affects earnings management, a topic which

has received a significant impact by the corporate gov-

ernance literature on the Chinese market. We define that

h10 is the ownership concentration variable represented

by the sum of the square of shares held by the ten largest

shareholders. Herfindahl is included to examine if the

power sharing among the largest shareholders was a bet-

ter mechanism of corporate governance rather than one

shareholder. Thus, hypothesis 2 is developed as follows:

Hypotheses 2: The ownership concentration (h10) has

a negative impact on earnings management.

Long, Li and Fu [11] suggested to use a cross item to

test the constrain of ownership structure. Then we use

sala_t3 × h10 as a multiplication cross item between ex-

ecutive compensation and ownership concentration to

test if compensation plan can be constrained by owner-

ship structure for the level of earnings management.

Based upon the aforementioned theory and studies, we

formulate hypothesis 3 as:

Hypotheses 3: The multiplication cross item between

executive compensation and ownership concentration

(sala_t3 × h10) has an impact on constraining motivation

of compensation plan to earnings management.

2.3. Board of Directors

2.3.1. Board Size

Beasley [12], Dechow [13] found that, the more numbers

of the board, the less ineffective supervision to the man-

agers, and the more possible to do earnings management.

Bai et al. [14] suggested board size have a positive effect

on earnings management. We define n_d as the number

of the board.

Hypotheses 4: When the number of the board (n_d) is

larger, there is a lower level of earnings management.

Does Ownership Structure Affect the Motivation of Compensation Contract of Earnings Management? 353

—From China’s Data

2.3.2. Propor t i on of Independent Directors

Dechow [13] corroborated that when there is a higher

proportion of independent directors, there is lower level

of earnings management. We define the proportion of

independent directors as id. Thus, we develop hypothesis

5 as:

Hypotheses 5: When the proportion of independent

directors (id) is higher in a firm, there is a lower level of

earnings management.

2.3.3. CEO Duality

Fama and Jensen (1983) assert that when the CEO dual-

ity exists I a firm, the monitoring function of the board

will be weaker, that would also likely lead to more earn-

ings management. We define CEO Duality as dua. Thus,

hypothesis 6 is developed as follows:

Hypotheses 6: The CEO duality (dua) has a positive

impact on earnings management.

3. Methodology of the Research

Earnings management has also been rampant in the

emerging market of China. China Securities Regulatory

Commission (CSRC) requires listed companies to meet

certain return on equity (ROE) criteria before they can

apply for permission to issue additional shares to existing

shareholders; the most important criterion for delisting a

listed company is a reported net loss for three consecu-

tive years. Whenever a contract or regulation is based on

accounting numbers, managers have an incentive to ma-

nipulate those numbers to serve their own or the firm’s

interests. We try to illuminate how controlling share-

holder and the board affect the level of earnings man-

agement.

In order to examine the relationships among ownership

concentration, the board of directors, and earnings man-

agement, we first establish the following regression equa-

tion:

DA =





1 sala_t3 +



2 h10 +



3 sala_t3 × h10



4 n_d +



5 id +



6 dua +



7 size_a +



8 leve +



(1)

where:

DA = discretionary accruals of the firm, as an indica-

tor for earnings management;

sala_t3 = natural logarithm of the total salary of the

top 3 executives

h10 = index of Herfindahl for ownership concentra-

tion;

sala_t3 × h10 = the multiplication cross item equaling

sala_t3 multiplied by h10;

n_d = number of the directors.

id = the proportion of independent directors;

dua = a dummy variable that equals 1 if the CEO is the

chairman or a vice chairman of the board of directors,

and 0 otherwise;

size_a = logarithm of the total assets;

leve = Debt ratio calculated by dividing liabilities by

assets

After the regression equation has been properly con-

structed, the next step is to find a way to measure the

discretionary accruals (DA). Jones[15] suggested that

earnings management can be achieved by various means

such as the use of accruals, changes in accounting meth-

ods, and changes in capital structure (e.g., debt defea-

sance, debt-equity swaps). This study focuses on total

accruals (TA) as the source of earnings management.

Total accruals (TA) include discretionary accruals (DA)

and non-discretionary accruals (NDA). Discretionary

accruals (DA) are used to measure the level of earnings

management of a firm, and are calculated as follows:

TAit =α1(1/Ait–1) + α2(ΔREVit/Ait-1) + α3(PPEit/Ait–1)+ εit

NDAit = α1(1/Ait-1)+α2(ΔREVit – ΔRECit )/Ait–1

+ α3(PPEit/Ait–1)

DAit = TAit – NDAit (2)

where:

DAit = dealing accruals in year t for firm i;

NDAit = non-dealing accruals in year t for firm i;

TAit = total accruals in year t for firm i;

ΔREVit =revenues in year t less revenues in year 1 – t

for firm i;

ΔRECit = accounts receivables in year t less accounts

receivables in year t – 1for firm i;

PPEit = gross property, plant, and equipment in year t

for firm i;

Ait = total assets in year t – 1for firm i;

εit = error term in year t for firm i

4. Data and Descriptive Statistics

4.1. Data Collection

We conduct our analysis with the full set of non-financial

companies which were continuously listed on the Shang-

hai Stock Exchange (SHSE) and the Shenzhen Stock

Exchange (SZSE) from 2006 to 2008 period. In these

three years, there are both top performance and bottom

influence in China’s stock market as a emerging market

of China. And this time period provides more recent data

and at the same time avoids duplication of time spans

with previous studies.

Based on above criteria, we extract 1024 SHSE and

SZSE listed firms each year and total 2048 firms for our

analyses. We use RESSET and CSMAR data sources to

collect the data required. These databases are used to

collect the corporate governance data and the accounting

data.

Does Ownership Structure Affect the Motivation of Compensation Contract of Earnings Management?

—From China’s Data

354

4.2. Summary Statistics

Table 1 provides some summary statistics for the above

variables. As revealed, these 1024 companies have an

average dealing accruals (DA) of –0.0259. The mean of

logarithm of total executives’ compensation is 13.68.

And the average of h10 is 0.1688.

The number of board is from 5 to 19, the average is

9.31, which is very fit for the regulation of the Chinese

Corporation Law. The mean of the proportion of inde-

pendent directors is 35.9%, which is satisfied not less

than “1/3” in the regulation of the Chinese Corporation

Law. And the mean of the duality of CEO is 0.1270,

which reflects CEO is mostly not the chairman of the

board in most companies.

Finally, in the controlling variables, the mean of scale

of assets (size_a) is 21.65, and debt ratio (leve) is

49.97%.

4.3. Related Coefficient Validity Test

Table 2 use Pearson related coefficient validity test and

Spearman related coefficient validity test to test the self-

relation among the independent variables. The Bottom

Left lists the coefficient of Pearson method and the upper

right lists the coefficient of Spearman method. Both two

methods have showed the acceptable validity test results

in Table 2.

4.4. Regression Models

In this section, we carry out our econometric analysis.

We investigate the effects of the corporate governance

mechanisms on earnings management. This study gives

us some ideas about the magnitude of the corporate gov-

ernance premiums which are used to measure the level of

corporate governance and earnings management.

Table 3 reports regression results of earnings man-

agement variables on the five variables used to measure

some of the governance mechanisms and earnings man-

agement for China’s listed companies. The regression

model is regressed by debt ratio and the size of assets as

two control variables. Three interesting findings are

listed in order.

First, the executive compensation has a positive rela-

tionship with earnings management and the coefficient is

statistically significant at 1% level in each model. This

result is on the opposite side of the hypothesis 1.

Second, the ownership concentration is positive sig-

nificantly in model 3 and 4 while not significantly in

model 2, which is different with the hypothesis 2. Thus,

as added in a multiplication cross item of executive

compensation and ownership concentration, the coeffi-

cient of the multiplication cross item is negative and ad-

verse to the sign of the coefficient of executive compen-

sation. And it is statistically significant at 1% level, same

as the hypothesis 3. This data result represents the con-

straint function of ownership concentration to executive

compensation on the level of earnings management.

Third, the board variables, such as number of the di-

rectors, proportion of independent directors, CEO duality,

Table 1. Descriptive statistics of dependent and independent

variables.

Variable Mean SD Min Max

DA –0.0259 0.1384 –0.7352 3.0349

sala_t3 13.6823 0.7865 11.1419 21.4206

h10 0.1688 0.1192 0.0023 0.7264

sala_t3 × h102.3068 1.6491 0.0310 10.5430

dua 0.1270 0.3331 0 1

n_d 9.3081 1.9482 1 18

id 0.3588 0.0500 0 0.6

size_a 21.6531 1.0932 18.8266 27.3463

leve 0.4997 0.1761 0.0182 0.9669

Table 2. Pearson and spearman correlation matrix.

DA sala_d3 h10 dua n_d id size_a leve

DA 1.0000 0.1934 0.0805 –0.0138 0.0281 0.0316 0.3336 0.2454

sala_t3 0.1559 1.0000 0.0449 –0.0143 0.1613 0.0147 0.4086 0.2892

h10 0.0446 0.0397 1.0000 –0.0525 –0.0068 –0.0182 0.1835 0.1621

dua –0.0070 –0.0211 –0.0615 1.0000 –0.1086 0.0628 –0.0903 0.0069

n_d 0.0149 0.1628 –0.0103 –0.1017 1.0000 –0.1819 0.2258 0.0182

id 0.0331 –0.0046 0.0041 0.0598 –0.1978 1.0000 0.0351 –0.0337

size_a 0.1981 0.3898 0.2520 –0.0972 0.2544 0.0186 1.0000 0.1317

leve 0.2842 0.1513 0.1021 –0.0002 –0.0223 –0.0001 0.0557 1.0000

Notes: the Pearson related coefficients are listed on the bottom left and the Spearman related coefficients are listed on the upper right.

Does Ownership Structure Affect the Motivation of Compensation Contract of Earnings Management? 355

—From China’s Data

Table 3. Regression results on earnings management for each model.

Model 1 Model 2 Model 3 Model 4

variable pred. sign coef. t_value coef. t_value coef. t_value coef. t_value

sala_t3 - 0.0150 3.62*** 0.0322 4.83*** 0.0315 4.59***

h10 - –0.0140 –0.54 1.4170 3.29*** 1.3327 2.98***

her_sa - –0.1035 –3.31*** –0.0979 –3.02***

id - 0.0401 0.64

n_d - –0.0024 –1.43

dua + 0.0012 0.13

size_a 0.0247 7.9*** 0.0294 9.9*** 0.0255 7.83*** 0.0267 7.85***

leve –0.0789 –4.43*** –0.0835 –4.69*** –0.0798 –4.47*** –0.0772 –4.20***

_cons –0.7278 –10.77*** –0.6180 –10.11*** –0.9779 –9.64*** –0.9866 –9.38***

F-value 40.05 35.28 26.33 16.10

R2 0.056 0.049 0.061 0.062

Adj-R2 0.055 0.048 0.059 0.059

Note: *, ** and *** indicate significance at 10%, 5% and 1% levels, respectively.

do not show statistically significant with earnings man-

agement in model 4. The hypothesis 4, 5 and 6 are not

passed the empirical test.

Among the two control variables, size_a and leve, are

positive and negative significant relationship with earn-

ings management at 1% level.

5. Sensitivity Test

From the above analysis, we find that the ownership

concentration has a significant constrain on compensa-

tion contract for earnings management. That is, the

stronger compensation motivation is, the higher level

earning management presents. With the constrain of the

ownership concentration, compensation contract shows a

indistinctively effect on the earning management.

For the sensitivity test, we use some alternative vari-

ables to see for the same sample data and same model if

we can get the same empirical results as the above. First,

for the executive compensation, we change the top 3

salary of chairman of board of directors or supervisors

or CEO to only top 3 managers, and only top3 board of

directors. Thus, we use these two variables to alternate

sala_t3, and we get the almost the same empirical re-

sults.

Second, we change the alternative variable for owner-

ship concentration from the top 10 stockholders (h10) to

the top 5 stockholders (h5). And the same significant

result and multiplication cross function show the con-

strain function of ownership concentration for executive

compensation.

Then, even for the further proving, we use the sensitiv-

ity test―robust test in Stata 10.0―to make if it still show

the same results as it illustrates as follows in Table 4.

6. Conclusions and Suggestion

This paper empirically studies the impacts of executive

compensation on earnings management and its constrain

by ownership. We find that some estimated effects of the

variables are consistent with theoretical predictions. In

particular, executive compensation has statistically sig-

nificant and positive effects on dealing accruals. This

emphasizes that the compensation contract is one of mo-

tivation of earnings management. The managers have

strong willing to make earnings management to improve

their compensation. While we still find that ownership

concentration could constrain the effects of compensa-

tion motivations on earnings management. It is owner-

ship concentration that constrains the compensation mo-

tivation on earnings management.

However, the corporate governance mechanisms, such

Table 4. Robust test for the regression models.

Model 3 Model 4

variable coef. t_value coef. t_value

sala_t3 0.0071582 1.73* 0.0076395 1.72*

h10 0.6354436 2.40*** 0.55208811.92**

sala_t3×h10 –0.0471493 –2.45*** –0.0417336–2.00**

dua 0.00161830.27

n_d –0.0027086–2.52***

id 0.02676860.66

size_a 0.023105 12.24*** 0.0251899 12.16***

leve 0.485133 21.04*** 0.4744188 19.36***

_cons –0.6800912–10.07

Does Ownership Structure Affect the Motivation of Compensation Contract of Earnings Management?

356

—From China’s Data

as the board, do not affect the compensation motivation

on earnings management. And the board size, board in-

dependence and the CEO duality do not present signifi-

cantly effect on earnings management, which shows the

weakness of corporate governance in China’s listed firms.

The reason why board’s function is so weak may be for

the current situation and background of China’s stock

market. Recently, in most listed firms, the board govern-

ance is still not so effectively and the independence of

the board is not so significant, which weaken the board’s

influence on earnings management. The board govern-

ance regulation and system need to improve, not just on

paper, but on real practice. Thus, as a emerging market,

the Shanghai Securities Exchange named year 2009 is

“the corporate governance year for China’s listed firms”,

which enhance more effective measures on corporate

governance, especially board governance.

Our findings have valuable implications for both the

security regulators and listed companies in the emerging

market of China. It is known that many security regula-

tors in the world, including both the developed and de-

veloping countries, have recognized the serious problems

of earnings management. They have proposed various

ways, known as the best practice codes, to reduce earn-

ings management and improve a firm’s overall govern-

ance standard. It is our belief that our study sheds light

on the relative importance of various corporate govern-

ance practices about earnings management, and should

provide useful information for Chinese regulatory au-

thorities to design the best practice codes tailored to the

Chinese institutional background and current level of

emerging market development. In addition, it also pro-

vide useful guide for firms to design their corporate gov-

ernance mechanisms so that they can decrease earnings

management, and enhance their corporate governance

and improve the Chinese emerging market to develop

healthily, stably and harmoniously.

7. Acknowledgements

We are grateful to many members of the 2nd Financial

Risk & Corporate Finance Conference in Dalian Univer-

sity of Technology and the 8th International Conference

on Supply Chain Management and Information Systems

in Hong Kong Polytechnic University. Both authors ac-

knowledge financial support from National Natural Sci-

ence Foundation of China (71172136) and the Funda-

mental Research Funds for the Central Universities

(DUT10ZD107, DC10040208) and New Century Excel-

lent Talents in University (NCET -10-0281). We are re-

sponsible for all remaining errors.

REFERENCES

[1] D. Hwang, B. Staley, Y. T. Chen, J.-S. Lan, “Confucian

Culture and Whistle-Blowing by Professional Account-

ants: An Exploratory Study,” Managerial Auditing Jour-

nal, Vol. 23, No. 5, 2008, pp.504-526.

[2] J. A. Berle and G. C. Means. “The Modern Corporation

and Private Property,” Macmillan, New York, 1932.

[3] E. F. Fama, “Agency Problems and the Theory of the

Fine,” Journal of Political Economy, Vol. 88, No. 2, 1980,

pp. 188-307. doi:10.1086/260866

[4] M. C. Jensen and W. H. Meckling, “Theory of the Firm:

Managerial Behavior, Agency Cost and Ownership Struc-

ture,” Journal of Financial Economics, Vol. 3, No. 4,

1976, pp. 305-360. doi:10.1016/0304-405X(76)90026-X

[5] R. La Porta, F. Lopez-de-Silanes and A. Shleifer, “Cor-

porate Ownership around the World,” Journal of Finance,

Vol. 54, No. 2, 1999, pp471-517.

[6] S. Claessens, S. Djankov and L. Lang, “The Separation of

Ownership and Control in East Asian Corporations,”

Journal of Financial Economics, Vol. 58, No. 1-2, 2000,

pp. 81-112. doi:10.1016/S0304-405X(00)00067-2

[7] P. M. Healy, “The Effect of Bonus Schemes on Account-

ing Decisions,” Journal of Accounting and Economics,

Vol. 7, No. 1-3, 1985, pp. 85-107.

doi:10.1016/0165-4101(85)90029-1

[8] R. Watts and J. L. Zimmerman, “Positive Accounting

Theory,” 1st Edition, Prentice-Hall, Upper Saddle River,

1986.

[9] D. K. Denis and J. J. McConnell, “International Corporate

Governance,” Journal of Financial and Analysis, Vol. 38,

No. 1, 2003, pp. 1-36. doi:10.2307/4126762

[10] P. M. Dechow and P. G. Sloan, “Executive Incentives and

the Horizon Problem: An Empirical Investigation,”

Journal of Accounting and Economics, Vol. 14, No. 1,

1991, pp. 51-89. doi:10.1016/0167-7187(91)90058-S

[11] J. Long, Y.-X. Li and L.-L. Fu, “Is It Governance or

Ownership That Affect the Motivation of Compensation

Contract on Earnings Management? An Empirical Study

on Chinese Listed Firms,” 8th International Conference

on Supply Chain Management and Information Systems,

Hong Kong, 6-9 October 2010, pp. 1-5.

[12] M. Beasley, “An Empirical Analysis of the Relationship

between Board of Director Composition and Financial

Statement Fraud,” The Accounting Review, Vol. 71, No. 4,

1996, pp. 443-465.

[13] P. M. Dechow, R. G. Sloan and A. P. Sweeney. “Detect-

ing Earnings Management,” The Accounting Review, Vol.

70, No. 2, 1995, pp. 193-225.

[14] C.-E. Bai, Q. Liu, J. Lu, F. M. Song and J.-X. Zhang,

“Corporate Governance and Market Valuation in China,”

Working Paper, 2004.

[15] J. Jones, “Earnigs Management during Import Relief

Investigations,” Journal of Accounting Research, Vol. 29,

No. 2, 1991, pp. 193-228.