SEARCH

SEARCH BY CITATION

Keywords:

  • Corporate governance;
  • earnings management;
  • investor protection;
  • emerging markets

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

This paper studies the impacts of corporate governance on earnings management. We use firm-level governance data, taken from Credit Lyonnais Security Asia (CLSA), of nine Asian countries, in addition to the country-level governance data used in past studies. Our conclusion is as follows. First, firms with good corporate governance tend to conduct less earnings management. Second, there is a size effect for earnings smoothing, that is, large size firms are prone to conduct earnings smoothing, but good corporate governance can mitigate the effect on average. Third, there is a turning point for leverage effect, i.e. when the governance index is large, leverage effect exists, otherwise reverse leverage effect exists. It shows that a highly leveraged firm with poor governance is prone to be scrutinised closely and thus finds it harder to fool the market by manipulating earnings. Fourth, firms with higher growth (lower earnings yield) are prone to engage in earnings smoothing and earnings aggressiveness, but good corporate governance can mitigate the effect. Finally, firms in stronger anti-director rights countries tend to exhibit stronger earnings smoothing. This counter-intuitive result is different from Leuz et al. (2003).


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

Studies of earnings management (hereafter, EM) have recently attracted a lot of attention because of the accounting fraud at Enron, WorldCom, Xerox, Royal Ahold, HealthSouth, and so on. The increasing attention to the quality of reported earnings makes the study of earnings management important again (Levitt, 2000). Schipper (1989) and Healy and Wahlen (1999) state that earnings management is the alternation of firms' reported economic performance by insiders to either “mislead some stakeholders” or to “influence contractual outcomes”. It is premised that insiders engage in EM to dilute their rent-seeking activities from outsiders. That is, EM is used to reduce outsider interference and protect insiders' private control benefits. For instance, insiders can use their discretion in financial reporting to inaccurately reflect firm performance and consequently weaken outsiders' ability to govern the firm.

Furthermore, investors who are concerned with the stock returns examine the earnings frequently. This heightens capital market pressure and creates an additional incentive for firms to engage in earnings manipulation. The regular scrutiny of firms' financial performances by financial analysts in the past decades puts further pressure on firms to maintain earnings momentum to fulfil the expectation of the market (e.g. Barth et al., 1999; Myers and Skinner, 2002). Thus, knowing the factors that affect earnings management is helpful for investors to foresee the variation of earnings. In particular, will a company with good governance have less incentive to conduct earnings management?

Past studies regarding governance factors that affected the earnings management were typically based on the country-level to account for the variations (e.g. Leuz et al., 2003). These studies began with the finding that the laws that protect investors differ significantly across countries, in part because of differences in legal origins (La Porta et al., 1998, hereafter, LLSV). Leuz et al. (2003), for example, examined the relation between outside investor protection and earnings management across 31 countries using non-financial industries data. They found that strong investor protection at a country level reduces firms' EM activities. In contrast to the above studies using non-financial industries, Shen and Chih (2005) employed data of financial industries to calculate EM of 48 countries based on the methodologies of Degeorge et al. (1999) and Burgstahler and Dichev (1997). They then followed the anti-director right, legal enforcement, accounting disclosure and inside trading analysis technique of LLSV to account for the variations of EM across the countries. Their results show that accounting disclosure is the more effective method that explains variation of EMs across countries.

Many provisions of investor protection in a country may not be binding for all firms since firms have the flexibility in their corporate charters and bylaws to either choose to “opt-out” and decline specific provisions or adopt additional provisions not listed in their legal code. Klapper and Love (2004) claimed that firms could improve investor protection rights by increasing disclosure, selecting well-functioning and independent boards, imposing disciplinary mechanism to prevent management and controlling shareholders from engaging in expropriation of minority shareholders, and so forth. Hence, two firms from the same country may perform varying degrees of protection to their investors. Namely, the relationship between the country-level legal infrastructure and the firm-level corporate governance mechanisms is not a one-to-one correspondence. Though Richardson et al. (2002) have used the firm-level financial ratios to predict earnings management, they did not consider the role of corporate governance.

The aim of this paper is to study how corporate governance affects EM of Asian countries by using firm-level data. In a recent report, Credit Lyonnais Securities Asia (hereafter CLSA) calculated an index with corporate governance rankings for 495 firms across 25 emerging markets and 18 sectors. The descriptive statistics presented in the CLSA report show that companies that ranked high on the governance index have better operating performance and higher stock returns. Klapper and Love (2004) applied these governance variables and found that better corporate governance is highly correlated with better operating performance and market valuation. Also applying CLSA governance measures, Chen et al. (2003) found that both disclosure and non-disclosure corporate governance mechanisms have a significantly negative effect on the cost of equity capital. Doidge et al. (2004) also used CLSA data and found that firms in less developed countries have less incentive to improve firm-level governance because outside finance is expensive and the adoption of better governance mechanisms is expensive. Except for these three papers, to the best knowledge of authors, none have applied the corporate governance data directly to investigate the relevant issues. This paper uses the same governance ranking produced by CLSA to further investigate the relationship between firm-level governance and firm-level earnings management. Then, we studied how the corporate governance affects the relationship between firms' financial ratios and earnings management.

This paper is organised as follows. The next section describes the construction of the earnings management measures. We describe the data of CLSA, followed by a section describing the econometric models. Empirical tests results, and the robustness check are presented. The final section presents the conclusions.

Earnings management measures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

Because insiders can exercise discretion along a number of different dimensions, and the preferred earnings management method can differ across countries, a single earnings management measure may not yield a complete picture. Accordingly, we analysed four different measures of EMs. Our earnings measures are the same as Leuz et al. (2003), who classified the earnings management into two kinds of manipulations: earnings smoothing and earnings discretion. Bhattacharya et al. (2003) further classified the earnings management into three categories: loss avoidance, earnings smoothing and earnings aggressiveness. In this paper, we followed Leuz et al.'s (2003) approach in classifying earnings management into earnings smoothing (EM1 and EM2 in our notation below) and earnings aggressiveness (EM3 below).1 The fourth EM measure is simply the average of their rankings (AEM).

Because accounting earnings are the sum of accruals and cash flow from operation, we first introduce accruals and cash flow. The operational definition of accruals is:

  • image
  • ΔCAit = change in total current asset;

  • ΔCashit = change in cash/cash equivalents;

  • ΔCLit = change in total current liabilities;

  • ΔSTDit = change in short-term debt included in current liabilities;

  • ΔTPit = change in income taxes payable;

  • Depit = depreciation and amortisation expense.

Once the accruals are obtained, we calculate cash flow from operation.

  • image

The first smoothing measure captures the degree to which insiders use their discretion to alter the accounting component of reported earnings, that is, accruals, to reduce the variability of operating earnings:

  • image

where SD represents standard deviation, and cash flow from operation is the cash flow defined above.

A low value of this measure is indicative of insiders using their discretion to smooth reported earnings. That is, the higher EM1 implies firms are less prone to manage earnings.

The second measure of earnings smoothing is based on the contemporaneous correlation between the change in accounting and the change in operating cash flow.

  • image

where Spearman is the Spearman correlation coefficient. The concept of this measure is that the insiders may also use their discretion to report accounting accruals that offset economic shocks to the firm's operating cash flow that would otherwise affect reported earnings. A negative correlation implies the use of discretionary accounting accruals to offset undesirable cash flow shocks, hence, a large earnings management. Therefore, the higher the EM2, the less is the tendency to manage earnings.

The third measure is related to earnings aggressiveness, which represents the insiders that use their reporting discretion to misstate the firm's actual economic performance. It is noted that accrual itself involves judgement of managers; hence, the third measure simply uses the magnitude of accruals as a proxy for the amount of discretion insiders use to influence earnings. Thus, the insiders may manipulate the earnings by adjusting accruals to avoid loss. For firm i, EM3 is equal to the average value of the magnitude of accruals scaled by the absolute value of the firm's cash flow from operations to control for firm performance:

  • image

The larger EM3 are indicative of large-scale use of discretion to manipulate reported accounting earnings.

Our fourth EM measure, AEM, is the average ranking of the above three earning management measures. Following the concept of Leuz et al. (2003), we ranked firms' EM1, EM2 and EM3 and calculated their averages, that is, for firm i,

  • image

Note that Rank (EM1) and Rank (EM2) are reversed to be consistent with Rank (EM3). Hence, the higher the AEM, the more is the tendency to earnings management.

Data of CLSA

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

The CLSA report includes corporate governance (CG) rankings on 495 companies in 25 emerging countries in April 2001 and February 2002.2 In these reports, firms' corporate governance was assessed based on seven key criteria. The reports show that good corporate governance is associated with strong performance in several dimensions, including share price level, stock returns and accounting profitability.

The questions in the CLSA report cover seven broad categories: management discipline (DISC), transparency (TRAN), independence (INDP), accountability (ACCT), responsibility (RESP), fairness (FAIR) and social awareness (SOCI). The meanings of these categories are as follows. “Transparency” refers to the ability of outsiders to assess the true position of a company. “Discipline” refers to management's commitment to emphasise shareholder value and financial discipline. “Independence” refers to the board of director's independence of controlling shareholders and senior management. “Accountability” refers to the accountability of management to the board of directors. “Responsibility” refers to the effectiveness of the board to take necessary measures in case of mismanagement. “Fairness” refers to the treatment minority shareholders receive from majority shareholders and management. The last category, “social awareness”, refers to the company's emphasis on ethical and socially responsible behaviour.

Furthermore, CLSA included CG6, the simple average of the first six items and CG7, the weighted average of the seven items, with the weight being equal to 15 per cent for the first six items and 10 per cent for the last one.

In our case, only nine Asian countries were adopted, including Hong Kong, India, Indonesia, Korea, Malaysia, the Philippines, Singapore, Taiwan and Thailand. We skipped the remaining countries since the identification of the firms in CLSA could not be positively confirmed.3

Earning management and governance

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

Econometric model

The determinants in our earnings management equation consider the firm-level corporate governance variables (CG6 or CG7), as well as the financial ratios of firms. In addition to these two types of variables, the interactions of both variables are also attempted to investigate the robustness of the two types of variables. Following the suggestions of Richardson et al. (2002), our model is

  • image(1)
  • image(2)

where subscript i is the ith firm, EM is proxied by EM1, EM2, EM3 and AEM, respectively; Size is the log of firm's sales, Leverage is the debt ratio (total debt/total asset), External is the amount of external financing, EPS_Grow is the number of growth in EPS in the recent three years and Earnings_Yield is the earnings to stock price. We used the Worldscope database (2000), which contains up to 10 years (1991–2000) of historical financial data from home-country annual reports of publicly-traded companies around the world, to calculate all EM measures and financial variables above. To compute these variables we require that each firm have income statement and balance sheet information for at least three consecutive years. PGDP is a 6-year average (1994–1999) of the GDP per capita. Antidirect is the “anti-director rights” index from La Porta et al. (1998), and an aggregate measure of (minority) shareholder rights, ranging from zero to six. LAW is measured as the mean score across three legal variables used in La Porta et al. (1998): (a) the efficiency of the judicial system, (b) an assessment of rule of law and (c) the corruption index. All three variables range from zero to 10. Disclosure measures the inclusion or omission of 90 items in the 1990 annual reports (La Porta et al., 1998). Insider is the 5-year average (1997–2001) of the insider trading index taken from the World Competitiveness Yearbook (2002). Insider ranges from zero to 10, and the higher the variable, the less the extent of inside trading.

The relation between firm size (Size) and earnings management is controversial. One view is that capital market pressures are greater for larger firms because they are subject to closer scrutiny by the investment banks and analyst community, leading them to adopt aggressive accounting policies. Therefore, larger firms have higher incentives to manage earnings (Richardson et al., 2002), which is referred to as the size effect in this paper. The opposite view is that large firms are often requested to disclose their information and hence have less probability to manage earnings. Insiders of small firms, alternatively, are able to retain their private information more successfully than their counterparts of large companies, suggesting a reverse size effect (Lee and Choi, 2002). Hence, the coefficient on Size is uncertain.

Leverage is used to capture the impact of debt contracting on earnings management. Two opposite empirical evidences are found between the leverage and earnings management. One strand of evidence is that high leverage firms tend to manage earnings aggressively, which is dubbed as the leverage effect. This is suggested by Sweeney (1994) and Press and Weintrop (1990), who found that firms closer to violating debt covenants manage earnings more aggressively and that highly leveraged firms tend to violate debt covenants. Becker et al. (1998) support this view and found that managers respond to debt contracting by strategically reporting discretionary accruals. Richardson et al. (2002) presented the evidence that debt covenants (as proxied by leverage) are a motivation for aggressive accounting policies of restatement firms. Opposite to the above view, high leverage may also imply less earnings management, which is dubbed as the reverse leverage effect. Dechow and Skinner (2000) reported that firms with high leverage are less likely to report small increases in earnings. Ke (2001) found that the probability of reporting a small increase in earnings rather than a small decrease in earnings is higher for firms with low financial leverage. Chung and Kallapur (2003) did not find evidence of a statistically significant association between abnormal accruals and leverage. As a result, the relation between Leverage and EM is uncertain.

External is the average ratio of the amount of external financing relative to total assets. Firms, which raise external funds from capital market, tend to portray a rosy picture of future potential. Hence, firms with high external funds will actively engage in earnings management (Richardson et al., 2002), which is referred to as the external financing effect. On the contrary, firms with higher external funds will engage in less earnings management, since they are under closer scrutiny by the market and thus find it harder to fool the market by manipulating earnings, which is referred to as reverse external financing effect.

EPS_Grow is the number of years of consecutive growth of earnings to price (EPS) (Richardson et al., 2002). In particular, it is equal to 1, 2 or 3 if it has consecutive positive growth for 1, 2 or 3 years. A higher number of EPS_Grow has two implications. First, firms indeed perform well, making its earnings grow every year. Second, the strong performance is the result of earnings management because firms face pressure from capital markets to report growing earnings. Barth et al. (1999) found that the market reacts negatively to firms that break their string of consecutive earnings increases. To avoid this negative response, a firm with a decline in EPS will have more incentive for earnings management, suggesting coefficient on EPS_Grow is positively related to earnings management.

Earnings_Yield, which denotes the earning yield, or earnings-to-price ratio, is used to examine the market's perceptions of future growth. Prior research suggests that growth stock is particularly sensitive to stock price, especially around earnings announcements (Skinner and Sloan, 2002). Richardson et al. (2002) also expect that firms trading at low earnings yield will be under greater pressure to adopt aggressive accounting policies to deliver the anticipated growth in earnings. Hence, growth stock seems more likely to adopt earnings management. This paper's intent is not to distinguish growth from value stocks but rather simply to argue that growth firms, proxied by low Earnings_Yield, are more likely to engage in earnings management. Thus, Earnings_Yield is expected to be negatively related to the extent to which firms manage earnings.

The country-level governance variables are also taken into account. Antidirect is the measure of protection of small shareholders, LAW is the efficiency and integrity of the legal environment, which also includes corruption and enforcement, Disclosure is the transparency of the financial reports and accounting standard, Insider discusses the severity of insider trading, and the higher Insider is indicative of less severity of insider trading.

Equation (2) specifies the coefficient βj as the function of corporate governance. The model suggests that good governance may lessen the incentive of earnings management triggered by the above financial ratios.

Empirical evidence

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

Data description and basic statistics

Table 1 reports the basic statistics of EM1, EM2, EM3 and AEM. In addition to the country column, the first column is the number of firms used in each country. India has the highest number of firms of 48, and Philippines has the lowest number of firms of 8. The total number of firms used is 203, but this number varies across the years because of missing data in some firms. The highest mean of EM1 falls on Taiwan (0.467), followed by India (0.457) and Hong Kong (0.420). Recall that the higher the EM1, the lesser is the earnings smoothing. Hence, on average, these three countries show the lowest tendency to adopt earnings smoothing. In addition, while Taiwan has the lowest tendency to smooth earnings, its variation is also the largest. In contrast to the above three countries, Thailand has the lowest mean of 0.218 with the least variation of 0.153.

Table 1. Descriptive statistics of earnings measures (EM1, EM2, EM3, AEM) and CG of firms across countries
CountryEM1EM2EM3AEMCG6CG7
NMeanSDMaxMinMeanSDMaxMinMeanSDMaxMinMeanSDMaxMinMeanMean
  1. EM1, EM2, EM3 and AEM are the earning management measures defined in this paper. Following Leuz et al. (2003), EM1 and EM2 are measure earnings smoothing, and EM3 measure earnings aggressiveness. The higher EM1 (and EM2) implies firms are less prone to conduct earnings smoothing, and the higher EM3 implies firms are more prone to conduct earnings aggressiveness. AEM is simply the average of the rankings of EM1, EM2 and EM3 for each firm. The higher AEM implies firms have higher extent of earnings management. CG6 and CG7 are the average of corporate governance index in CLSA (2002). The higher CG6 (and CG7) represents better corporate governance.

Hong Kong210.4200.2681.1480.073−0.7430.4471.000−1.0001.0580.3071.5460.24786.52438.231150.3327.66762.8163.10
India480.4570.3972.0270.042−0.7960.4631.000−1.0000.6550.6014.3430.183100.79945.237177.0015.00061.3662.39
Indonesia140.2580.1730.5870.037−0.8560.202−0.429−1.0000.6500.2140.8100.213109.71454.398185.0031.33336.8436.84
Korea150.3330.2921.1630.065−0.9040.078−0.771−1.0000.8492.81611.3810.514120.60033.374188.0078.00057.4858.87
Malaysia330.3330.2601.0260.009−0.8410.2680.371−1.0000.6970.9355.4220.10297.10144.505185.6729.00061.3461.82
Philippines80.3740.2430.8030.107−0.9110.080−0.800−1.0000.6000.3021.0340.17597.45842.458157.3325.00051.2552.88
Singapore280.3530.2601.0850.083−0.8460.193−0.400−1.0000.7680.6002.4000.126104.26247.901191.6724.00063.2963.12
Taiwan210.4670.4742.1110.075−0.7860.3270.200−1.0002.2918.04837.3700.22387.46038.391175.3322.00059.9160.69
Thailand150.2180.1530.6170.049−0.9610.060−0.800−1.0001.5493.11912.5770.066131.04435.378189.3379.33362.3762.63
Total2040.3760.3212.1110.009−0.8330.3221.000−1.0000.9982.87237.3700.066102.00044.150191.66715.00059.9259.32

EM2 shows similar patterns as those of EM1 in that the above three countries (Taiwan, Hong Kong and India) still exhibit the highest numbers. Among these three countries, however, Hong Kong, rather than Taiwan, has the highest mean with the highest standard deviation. Because both EM1 and EM2 are designated to detect earnings smoothing, the earnings smoothing is thus less severe. This is probably because there are many tax deduction mechanisms in these countries to encourage investment and free trade.4

EM3, which measures earnings aggressiveness, shows a different result. The highest number still falls on Hong Kong (2.291), but the meaning of earnings management is reversed. Recall that EM3 is employed to detect the proportion of accruals to cash flow, and hence a higher EM3 implies a stronger tendency to conceal the true earnings by using the accruals. This evidence, together with the above results of EM1 and EM2, suggest that the earnings management in Taiwan seems to focus on adjusting income level (based on EM3), but this adjustment is not related to earnings smoothing (based on EM1 and EM2). This evidence is in fact consistent with the finding of Bhattacharya et al. (2003), who found that firms in Taiwan are actively engaged in earnings aggressiveness and loss avoidance but passively in earnings smoothing. Thailand and Hong Kong have the second and the third highest numbers of EM3. By contrast, the Philippines and Indonesia have the lowest number of 0.600 and 0.650. Thailand, Taiwan and Hong Kong have the highest market capitalisation ratios in the region, and this simple ranking appears to suggest that the EM3 is positively related to the development of capital market.

The AEM, adopted from Leuz et al. (2003), is the ranking of earnings management. Hong Kong and Taiwan have the lowest number of 86.54 and 87.46, respectively, whereas Thailand and Korea have the highest number of 131.04 and 120.60, respectively. Hence, in general, the former two countries manipulate less of the earnings than the latter two countries.

The corporate governance measures of CG6 and CG7 provided by CLSA are also interesting. Indonesia has the lowest CG6 (36.84), followed by the Philippines (51.25) and Korea (57.48), whereas Singapore has the highest CG6 (63.29), followed by Hong Kong (62.81) and Thailand (62.37). Thus, the ranking of CG6 seems to not be completely related to the above earnings measurement measures. The ranking of CG7 is similar to that of CG6 and hence is skipped here.

Table 2 reports the descriptive statistics of explanatory variables used in our sample. It is surprising to find that the mean of Size and External are similar across nine countries. Consistent with our expectation, the highest number of debt is in Korea (53.12) since Korean firms borrowed substantially during the Asian financial crisis; in contrast, the lowest number is in Hong Kong (18.95). Earnings yield (Earnings_Yield) also differs across countries. Hong Kong is 9.29 per cent but Korea is only −5.01 per cent.

Table 2. Descriptive statistics of firm characteristics across countries
CountryNSizeLeverageExternalEPS_GrowEarnings_Yield
  1. Size is the log of firm's sales, Leverage is the debt ratio (total debt/total asset), External is the amount of external financing, EPS_Grow is the number of growth in EPS in the recent three years and Earning_Yield is the earnings to stock price. We used the Worldscope database (2000), which contains up to ten years (1991–2000) of historical financial data from home-country annual reports of publicly-traded companies around the world, to calculate the financial variables above. To compute these variables, we require that each firm have income statement and balance sheet information for at least three consecutive years.

Hong Kong2113.6418.950.030.389.29
India4812.7322.160.040.425.94
Indonesia1413.1939.670.060.00−2.47
Korea1515.1153.120.120.13−5.01
Malaysia3313.1921.560.030.034.46
Philippines813.3731.090.070.133.02
Singapore2813.1722.130.040.112.72
Taiwan2113.2723.260.080.294.39
Thailand1512.4739.020.100.132.28
Mean 13.3530.110.060.182.74
Median 13.1923.260.060.133.02

Table 3 presents the correlation coefficients of all variables used in this paper. Except for the high correlation of up to 0.747 between LAW and Disclosure, the correlation coefficients between other pairwise explanatory variables are less than 0.3. Interesting to note is that Insider has a negative relation with CG6 and CG7. That is, the less insider trading of a country, the lower is the corporate governance. Recall that the higher the Insider, the less the extent of inside trading. Therefore, the negative association between Insider and CG seems to represent that the demand for sound corporate governance declines when insider trading of a country is not severe.

Table 3. EM, corporate governance and investor protection
 EM2EM3DISCTRANINDPACCTRESPFAIRSOCIAntidirectLAWInsiderDisclosureCG7CG6
  1. EM1, EM2 and EM3 are the earning management measures defined in this paper. DISC, TRAN, INDP, ACCT, RESP, FAIR and SOCI are the average of corporate governance index in CLSA (2002); CG6 is the simple average of the first six items, and CG7 is the weighted average of the seven items, with the weight = 15% for the first six items and 25% for the last one. Antidirect is the “anti-director rights” index from La Porta et al. (1998), is an aggregate measure of (minority) shareholder rights, ranging from zero to six. LAW is measured as the mean score across three legal variables used in La Porta et al. (1998): (a) the efficiency of the judicial system, (b) an assessment of rule of law and (c) the corruption index. All three variables range from zero to 10. Disclosure measures the inclusion or omission of 90 items in the 1990 annual reports (La Porta et al., 1998). Insider is the 5-year average (1997–2001) of the inside trading index taken from World Competitiveness Yearbook. Insider ranges from zero to 10, and the higher the variable, the less the extent of inside trading.

EM10.205−0.0850.120−0.0040.1230.024−0.0660.0710.1020.1500.071−0.0920.0570.0970.102
EM2 −0.0570.0360.029−0.034−0.018−0.033−0.0150.0660.0860.075−0.0220.011−0.0010.004
EM3  0.0650.097−0.0140.042−0.042−0.0070.039−0.0940.004−0.0530.0240.0380.040
DISC   0.1010.2830.1670.2310.3110.3630.1870.166−0.2610.1960.6200.626
TRAN    0.0410.1390.1120.0460.148−0.2340.100−0.1380.1750.3630.361
INDP     0.1690.3920.4030.1300.3080.349−0.0930.3290.6850.668
ACCT      0.1410.0680.3460.0540.025−0.3410.1840.5000.509
RESP       0.3540.0160.1440.258−0.0810.3050.6050.582
FAIR        0.0370.3030.317−0.1290.2860.6280.606
SOCI         0.073−0.029−0.5260.2750.4430.509
Antidirect          0.439−0.1630.3210.2470.243
LAW           0.0300.7470.3370.321
Insider            −0.392−0.363−0.393
Disclosure             0.4510.456
CG7              0.997

Regression results

Figure 1 plots the scatter plot between governance factors (CG6 and CG7) against four earnings measures. The pattern is unclear for these two variables because of strong heteroscedasticity. We adopted the weighted least square (WLS) method by employing dependent variables as the weights to remove the heteroscedasticity.5

image

Figure 1. Scatter plot of corporate governance (CG) and EM

Download figure to PowerPoint

Tables 4 to 7 report the estimated results of models (1) and (2) using EM1, EM2, EM3 and AEM as dependent variables, respectively. There are four different specifications in each CG, and the specifications depend on whether interaction terms are taken into account or not and hence eight equations are considered in each table.

Table 4. EM1 prediction function: weighted least square
 CG6CG7
(A)(B)(C)(D)(E)(F)(G)(H)
EM1EM1EM1EM1EM1EM1EM1EM1
  • ***, ** and *

    represent the level of significance at 0.01, 0.05 and 0.10, respectively.

Constant0.515 (1.592)0.089 (0.157)0.829*** (2.999)0.548 (1.009)1.566* (1.674)0.095 (0.166)0.855*** (3.075)0.562 (1.031)
CG0.007*** (2.622)0.008*** (2.951)  0.006*** (2.585)0.008*** (3.030)  
Size−0.044** (−2.050)−0.046* (−1.953)−0.074** (−2.051)−0.041* (−1.816)−0.045** (−2.091)−0.047** (−1.990)−0.076** (−2.132)−0.042* (−1.866)
Leverage−0.001 (−0.640)0.0004 (0.285)0.026*** (2.450)0.001 (0.658)−0.001 (−0.654)0.0004 (0.277)0.025*** (2.421)0.001 (0.704)
External0.665* (1.688)0.958** (2.247)−3.929 (−1.512)0.619 (1.484)0.671* (1.695)0.939** (2.200)−3.260 (−1.294)0.573 (1.371)
EPS_Growth0.095* (1.657)−0.002 (−0.032)−0.643 (−1.577)−0.076 (−1.192)0.107* (1.882)0.006 (0.096)−0.729* (−1.811)−0.072 (−1.121)
Earnings_Yield−0.003 (−1.218)−0.001 (−0.432)0.025 (1.230)0.001 (0.259)−0.004 (−1.322)−0.001 (−0.482)0.024 (1.223)0.001 (0.260)
PGDP 0.014 (1.460) 0.017* (1.903) 0.014 (1.462) 0.017* (1.906)
Antidirect 0.155*** (3.087) −0.466*** (−3.040) 0.158*** (3.087) −0.460*** (−3.037)
LAW −0.111 (−1.517) 0.224 (1.078) −0.119 (−1.560) 0.200 (0.960)
Disclosure 0.005 (0.469) 0.031 (1.277) 0.006 (0.494) 0.036 (1.457)
Insider 0.002 (0.035) −0.466** (−2.295) 0.003 (0.056) −0.510*** (−2.557)
CG × Size  0.001 (1.553)   0.001 (1.589) 
CG × Leverage  −0.0004*** (−2.550)   −0.0004*** (−2.518) 
CG × External  0.064* (1.685)   0.054 (1.459) 
CG × EPS_Growth  0.010* (1.730)   0.011** (1.995) 
CG × Earnings_Yield  −0.0004 (−1.437)   −0.0004 (−1.441) 
CG × Antidirect   0.009*** (4.220)   0.009*** (4.265)
CG × LAW   −0.005* (−1.738)   −0.005 (−1.639)
CG × Disclosure   −0.0003 (−1.107)   −0.0004 (−1.290)
CG × Insider   0.006** (2.259)   0.007*** (2.537)
R20.3980.4420.4560.4990.4060.4510.4640.510
Adjusted R20.3790.4070.4270.4580.3880.4170.4360.471
N202188202188202188202188
Table 5. EM2 prediction function: weighted least square
 CG6CG7
(A)(B)(C)(D)(E)(F)(G)(H)
EM2EM2EM2EM2EM2EM2EM2EM2
  • ***, ** and *

    represent the level of significance at 0.01, 0.05 and 0.10, respectively.

Constant−0.955*** (−2.719)−1.084* (−1.720)−0.594*** (−2.528)−0.495 (−0.894)−0.945*** (−2.684)−1.040* (−1.649)−0.568*** (−2.392)−0.401 (−0.727)
CG0.011*** (4.011)0.012*** (4.294)  0.011*** (4.115)0.013*** (4.502)  
Size−0.049** (−2.120)−0.033 (−1.284)−0.039 (−1.255)−0.021 (−0.911)−0.050** (−2.164)−0.034 (−1.333)−0.046 (−1.512)−0.024 (−1.040)
Leverage−0.003* (−1.826)−0.002 (−1.190)0.038*** (4.124)−0.001 (−0.815)−0.003* (−1.829)−0.002 (−1.183)0.038*** (4.263)−0.001 (−0.737)
External1.405*** (3.280)1.781*** (3.786)−12.509*** (−5.664)1.173*** (2.758)1.459*** (3.397)1.800*** (3.837)−11.634*** (−5.413)1.116*** (2.636)
EPS_Growth0.250*** (4.035)0.164*** (2.317)−1.558*** (−4.493)0.033 (0.502)0.264*** (4.281)0.167*** (2.365)−1.579*** (−4.597)0.029 (0.450)
Earnings_Yield−0.005 (−1.518)−0.003 (−0.977)0.047*** (2.765)−0.001 (−0.329)−0.005 (−1.594)−0.003 (−1.015)0.044*** (2.636)−0.001 (−0.354)
PGDP −0.008 (−0.750) −0.002 (−0.233) −0.007 (−0.702) −0.002 (−0.169)
Antidirect 0.071 (1.278) −0.986*** (−6.317) 0.075 (1.328) −0.987*** (−6.434)
LAW 0.007 (0.084) 0.616*** (2.912) 0.002 (0.022) 0.608*** (2.885)
Disclosure −0.012 (−0.935) −0.012 (−0.500) −0.012 (−0.956) −0.012 (−0.470)
Insider 0.076 (1.456) −0.092 (−0.446) 0.077 (1.466) −0.111 (−0.552)
CG × Size  0.0003 (0.910)   0.0004 (1.165) 
CG × Leverage  −0.001*** (−4.450)   −0.001*** (−4.589) 
CG × External  0.199*** (6.104)   0.187*** (5.864) 
CG × EPS_Growth  0.024*** (4.991)   0.025*** (5.122) 
CG × Earnings_Yield  −0.001*** (−3.152)   −0.001*** (−3.028) 
CG × Antidirect   0.016*** (7.109)   0.016*** (7.292)
CG × LAW   −0.009*** (−3.129)   −0.009*** (−3.119)
CG × Disclosure   0.0001 (0.229)   0.0001 (0.169)
CG × Insider   0.002 (0.647)   0.002 (0.771)
R20.5130.5110.7300.6250.5290.5320.7370.646
Adjusted R20.4980.4800.7160.5950.5140.5030.7230.617
N202188202188202188202188
Table 6. EM3 prediction function: weighted least square
 CG6CG7
(A)(B)(C)(D)(E)(F)(G)(H)
EM3EM3EM3EM3EM3EM3EM3EM3
  • ***, ** and *

    represent the level of significance at 0.01, 0.05 and 0.10, respectively.

Constant0.937*** (2.688)1.754*** (2.741)0.493 (1.634)1.320** (2.050)0.993*** (2.795)1.800*** (2.757)0.515* (1.673)1.320** (2.008)
CG−0.006** (−2.119)−0.006* (−1.945)  −0.006*** (−2.333)−0.006** (−2.199)  
Size0.002 (0.088)−0.005 (−0.200)0.057 (1.426)−0.006 (−0.209)0.001 (0.038)−0.005 (−0.190)0.054 (1.364)−0.005 (−0.201)
Leverage0.003* (1.951)0.003* (1.822)−0.015 (−1.301)0.003* (1.866)0.003* (1.842)0.003* (1.709)−0.014 (−1.192)0.003* (1.739)
External0.122 (0.288)0.242 (0.516)−1.507 (−0.531)0.191 (0.395)0.152 (0.349)0.290 (0.608)−1.623 (−0.582)0.233 (0.473)
EPS_Growth0.058 (0.947)0.067 (0.951)−0.195 (−0.437)0.060 (0.813)0.058 (0.928)0.071 (0.990)−0.114 (−0.254)0.064 (0.853)
Earnings_Yield−0.010*** (−3.183)−0.008*** (−2.467)0.053*** (2.411)−0.008** (−2.295)−0.010*** (−3.276)−0.008*** (−2.548)0.054*** (2.496)−0.008*** (−2.410)
PGDP −0.020** (−1.995) −0.021** (−1.988) −0.021** (−2.007) −0.022** (−2.020)
Antidirect −0.094* (−1.692) −0.152 (−0.847) −0.098* (−1.710) −0.146 (−0.807)
LAW 0.128 (1.590) 0.018 (0.073) 0.137 (1.609) 0.033 (0.133)
Disclosure −0.027** (−2.151) −0.015 (−0.514) −0.028** (−2.089) −0.018 (−0.613)
Insider 0.164*** (3.108) 0.295 (1.254) 0.164*** (3.047) 0.330 (1.400)
CG × Size  −0.001* (−1.696)   −0.001* (−1.690) 
CG × Leverage  0.0003 (1.531)   0.0002 (1.405) 
CG × External  0.028 (0.662)   0.030 (0.734) 
CG × EPS_Growth  0.003 (0.551)   0.002 (0.356) 
CG × Earnings_Yield  −0.001*** (−2.893)   −0.001*** (−3.001) 
CG × Antidirect   0.001 (0.355)   0.001 (0.290)
CG × LAW   0.002 (0.483)   0.002 (0.446)
CG × Disclosure   −0.0001 (−0.490)   −0.0002 (−0.371)
CG × Insider   −0.002 (−0.582)   −0.002 (−0.730)
R20.3570.3530.4120.3570.3580.3560.4140.361
Adjusted R20.3370.3110.3800.3030.3380.3140.3830.307
N196182196182196182196182
Table 7. AEM prediction function: weighted least square
 CG6CG7
(A)(B)(C)(D)(E)(F)(G)(H)
AEMAEMAEMAEMAEMAEMAEMAEM
  • ***, ** and *

    represent the level of significance at 0.01, 0.05 and 0.10, respectively.

Constant160.249*** (3.851)344.404*** (4.624)88.259*** (2.416)274.296*** (3.763)163.465*** (3.885)352.474*** (4.675)84.785** (2.292)277.629*** (3.765)
CG−1.073*** (−3.324)−1.048*** (−3.126)  −1.130*** (−3.544)−1.121*** (−3.367)  
Size0.362 (0.132)−0.012 (−0.004)11.190*** (2.343)−0.588 (−0.192)0.420 (0.152)0.003 (0.001)11.759*** (2.497)−0.518 (−0.169)
Leverage0.291 (1.588)0.139 (0.700)−3.470*** (−2.452)0.101 (0.514)0.272 (1.492)0.125 (0.632)−3.416*** (−2.454)0.076 (0.388)
External10.325 (0.204)−35.203 (−0.633)−74.108 (−0.216)−8.522 (−0.152)14.915 (0.291)−28.736 (−0.512)−167.983 (−0.501)1.228 (0.022)
EPS_Growth−8.393 (−1.143)−6.672 (−0.798)53.561 (0.994)−0.319 (−0.037)−9.522 (−1.291)−7.362 (−0.871)65.326 (1.220)−0.576 (−0.067)
Earnings_Yield0.243 (0.677)0.135 (0.357)−2.682 (−1.013)0.011 (0.029)0.231 (0.639)0.115 (0.300)−2.315 (−0.882)−0.033 (−0.086)
PGDP −2.820*** (−2.332) −3.126*** (−2.622) −2.943*** (−2.355) −3.278*** (−2.665)
Antidirect −23.392*** (−3.582) 26.574 (1.294) −24.242*** (−3.610) 27.375 (1.335)
LAW 22.044** (2.302) −22.780 (−0.818) 23.478*** (2.343) −17.794 (−0.631)
Disclosure −3.930*** (−2.617) −4.172 (−1.275) −4.081*** (−2.613) −4.919 (−1.474)
Insider 12.366** (2.004) 58.975** (2.164) 12.679** (2.024) 62.425** (2.312)
CG × Size  −0.158*** (−3.108)   −0.162*** (−3.296) 
CG × Leverage  0.054*** (2.667)   0.053*** (2.650) 
CG × External  1.629 (0.322)   3.182 (0.641) 
CG × EPS_Growth  −0.843 (−1.124)   −1.033 (−1.379) 
CG × Earnings_Yield  0.043 (1.108)   0.037 (0.964) 
CG × Antidirect   −0.725*** (−2.520)   −0.757*** (−2.624)
CG × LAW   0.655* (1.724)   0.609 (1.566)
CG × Disclosure   0.002 (0.039)   0.010 (0.232)
CG × Insider   −0.633* (−1.694)   −0.676* (−1.831)
R20.4630.4550.4880.4830.4700.4660.4970.495
Adjusted R20.4460.4210.4610.4410.4540.4320.4700.454
N202188202188202188202188

Table 4 presents the estimated results using EM1 as the proxy of earnings management. Coefficients of CG6 and CG7 are found to be positively significant, indicating that good corporate governance increases EM1. Recall that a higher EM1 denotes less earnings management, and hence good governance decreases the earnings smoothness, which is consistent with our intuition.

We next discuss the influences of the financial variables and how CG6 affects the impacts of these financial variables. Coefficients of Size are overwhelmingly negative, suggesting that larger firms tend to do more earnings smoothing. Hence, the size effect is supported, and as noted above, this is consistent with Richardson et al. (2002).6 Coefficients of the interaction between Size and CG6, however, are insignificant, suggesting that the good governance does not mitigate the earnings smoothing when the size of firm is large. Coefficients of Leverage are mostly insignificant except for specification (C). In specification (C), coefficient of Leverage is significantly positive (which is 0.026), but its interaction term with CG6 is significantly negative (which is −0.0004). Thus, the turning point of Leverage is decided by the formula 0.026 − 0.0004 × CG6, which is equal to 65. When CG6 is higher than this turning point, leverage effect exists, i.e. highly leveraged firms conduct more earnings smoothing. When CG6 is lower than this turning point, reverse leverage effect exists, i.e. highly leveraged firms conduct less earnings smoothing. It represents that a highly leveraged firm with poor governance is prone to be scrutinised closely and thus harder to fool the market by manipulating earnings.

External's coefficients are significantly positive in the first two specifications, supporting the reverse external financing effect, i.e. firms with higher external financing engage in less earnings smoothing. In specification (C), however, this effect disappears, as the coefficient of the interaction, CG × External, is significantly positive. EPS_Grow and Earnings_Yield seem not to have a strong effect on EM1, meaning that a firm with high earnings growth and a high Earning level is not prone to adopt earnings smoothing. Furthermore, coefficients of the interaction terms with CG, CG × Earning_Yield, are insignificant. In sum, CG variables do have an effect on earnings management. Moreover, it affects the manner that financial ratios influence earnings management.

Coefficients of macro and country-level governance variables are also interesting. GDP per capita has a positive effect on EM1 regardless of specifications. Accordingly, wealthy countries have less intention to adopt earnings smoothing, representing the effect of better accounting/reporting standards in more developed countries (LLSV, 1998). With regard to the country governance, coefficients of Antidirect are significantly positive in specification (B), suggesting that a firm in a country with good anti-director rights does less earnings smoothing, which is consistent with Leuz et al. (2003). Contrary to those found in Leuz et al. (2003), however, stronger anti-director rights show counter-intuitive result as it causes stronger earnings smoothing. This is because the same coefficient changes the sign to negative (−0.466) in specification (D), but this effect appears in low firm-level governance countries only, and not in high firm-level governance countries, because the coefficient of the interaction term, CG × Antidirect, is positive (0.009). Law and Disclosure do not have strong impact on EM. Insider has a negative effect (−0.466) on EM1 in specification (D), while the coefficient of the interaction term, CG × Insider, is positive (0.006), representing that good firm-level governance more effectively decreases earnings smoothing in countries with less inside trading.

Similar results are obtained when CG7 is used and hence are not discussed here.

Table 5 presents the estimated results using EM2 as a dependent variable. Recall that a higher EM2 implies it is less likely that accruals to offset economic shocks will be employed, and hence less earnings smoothing. Coefficients of CG6 and CG7 are found to be overwhelmingly positive regardless of specifications, suggesting that firms with weak governance are prone to use accruals to conduct earnings smoothing. These results are the same as those found in Table 4.

Coefficients of corporate controllable variables in Table 5 also demonstrate similar patterns as those reported in Table 4. Namely, Size, which also has a negative sign, implying size effect, but this effect is elusive as it is only significant in the first specification. The results of Leverage are similar to those reported in Table 4. There is a turning point for the external financing effect on EM2. That is, in the specification (C), the coefficient of External is −12.509, and the coefficient of interaction, CG × External, is 0.199, making the turning point of CG6 to be around 63.

The coefficients of EPS_Grow are positive in specification (A) and (B) in Table 5, representing firms with higher EPS_Grow, to avoid negative response by the market, have higher incentives to manage earnings to maintain their string of consecutive earnings increases, which is consistent with Barth et al. (1999). However, high EPS_Grow with high CG tends to encourage the earnings smoothing as the coefficient of the interaction term, CG × EPS_Grow, is significantly positive in specification (C). The coefficient of Earnings is positive (0.047) in specification (C), while that of the interaction term, CG × Earning_Yield, is negative (−0.001), suggesting that firms with higher growth (lower earnings yield) is prone to manage earnings (Skinner and Sloan 2002), especially when they have worse governance.

The results of Antidirect are similar to those shown in Table 4. It is seen that stronger enforcement of laws (LAW) can result in less earnings smoothing, since the coefficient is positive (0.616) in specification (D). Interestingly, this effect is stronger in countries with worse corporate governance, since the coefficient of the interaction term, CG × LAW, is negative (−0.009).

Coefficients of determinants in Table 5 are higher than those in Table 4, suggesting a better fitting. This is also evidenced by finding more significant coefficients in Table 5. Especially, when the interaction terms are considered, the adjusted-R2 increases from 0.5 to 0.7 roughly.

Table 6 reports the estimation results using EM3 as a dependent variable. Recall that a large number of EM3 refers to strong earnings aggressiveness. Hence, negative coefficients on CG6 and CG7 imply that better governance lessens the effect of earnings management. Three differences are found in this table from the previous two tables. First, there is a leverage effect in EM3, i.e. highly leveraged firms tend to conduct more earnings aggressiveness, but there is no reverse leverage effect, which is found in EM1 and EM2. Second, Size and External coefficients are insignificant for all specifications, suggesting that neither large firms, nor firms that rely on external finances are factors that affect the decision of earnings aggressiveness. Hence, there is a size effect for earnings smoothing but there is no size effect for earnings aggressiveness. Similarly, there is an external financing effect for earnings smoothing but no such effect for earnings aggressiveness. Third, consistent with our intuition, the results of Earnings_Yield suggest that a firm with higher growth (lower earnings yield) is more active in managing earnings, but good governance can mitigate the effect. This finding is similar to that in earnings smoothing (EM2).

Third, most of the country-level indices, including PGDP, Antidirect and Disclosure, are significant. Hence, firms in countries with good economic development and country governance engage in less earnings aggressiveness, which is consistent with Leuz et al. (2003).

Coefficients of determinants are smaller than those of the previous two tables, being consistent with the finding that most coefficients are insignificant.

Table 7 shows the estimated results of using AEM. Recall that AEM is the average of the rankings of EM1, EM2 and EM3 for each firm. The higher AEM implies firms have higher extent of earnings management. The coefficients of CG6 and CG7 are both significantly negative, implying that better governance is consistent with lower earnings management, on average. It is also worth to noting that coefficients of Size are positive in specification (C) and (G), suggesting that larger firms tend to do more earnings management. Coefficients of the interaction between Size and CG6, however, are significantly negative, suggesting that the good governance can mitigate the earnings management, on average. As other results do not change qualitatively, we skip the discussion.

Robustness check

We examine whether a firm belonging to a business group affects our results or not. This is because a firm belonging to a business group could have more channels to hide or increase its costs or revenues.7 Thus, it gets more chance to manage earnings. We obtain the information of a firm belonging (or not belonging to) business groups by first visiting the websites of our sample firms. It is easy if the websites explicitly and clearly record which group(s) a firm belongs to. When the statement is ambiguous, however, we key in the names of firms in Google and Yahoo! to examine any relevant news to obtain the information of belonging (or not belonging).8Table A2 shows our classification of belonging to a business group.

Tables 8 and 9 report the descriptive statistics for earnings management measures and corporate governance indices of firms belonging to (and not belonging to) a business group. It is shown that all paired mean differences in Tables 8 and 9 are statistically insignificant, showing that whether a firm belongs to a business group does not greatly affect earnings management or corporate governance. Table 10 shows that firms have higher earnings growth and higher earnings yield if they do not belong to business groups.

Table 8. Descriptive statistics of earnings measures (EM1, EM2, EM3, and AEM) of firms not belonging to business group (group = 0) and firms belonging to business group (group = 1)
 NEM1EM2EM3AEM
MeanSDMinMaxMeanSDMinMaxMeanSDMinMaxMeanSDMinMax
  1. Group = 0 means firms do not belong to certain business group; Group = 1 means firms belong to certain business group. EM1, EM2, EM3 and AEM are the earning management measures defined in this paper. Following Leuz et al. (2003), EM1 and EM2 are to measure earnings smoothing, and EM3 is to measure earnings aggressiveness. The higher EM1 (and EM2) implies firms are less prone to conduct earnings smoothing, and the higher EM3 implies firms are more prone to conduct earnings aggressiveness. AEM is simply the average of the rankings of EM1, EM2 and EM3 for each firm. The higher AEM implies firms have higher extent of earnings management.

Group = 0490.3590.2620.0491.376−0.8810.199−1.0000.0001.0301.8000.12612.577108.13640.08022.667188.000
Group = 11360.3820.3100.0092.027−0.8260.323−1.0001.0000.6870.6820.0665.42298.54944.63415.000191.667
t–test of the diff. Group (0 minus 1)−0.498−1.3771.3001.392
Table 9. Descriptive Statistics of corporate governance measures (CG6 and CG7) of firms not belonging to business group (group = 0) and firms belonging to business group (group = 1)
 NCG6CG7
MeanSDMinMaxMeanSDMinMax
  1. Group = 0 means firms do not belong to certain business group; Group = 1 means firms belong to certain business group. CG6 and CG7 are the average of corporate governance index in CLSA (2002). CG6 and CG7 are the average of corporate governance index in CLSA (2002). The higher CG6 (and CG7) represents better corporate governance.

Group = 04960.7019.18832.42979.74360.0249.36732.40079.600
Group = 113659.09312.88117.77185.82958.56113.00816.99586.800
t–test of the diff. Group (0 minus 1)0.9380.840
Table 10. Descriptive statistics of firm characteristics of firms not belonging to business group (group = 0) and firms belonging to business group (group = 1)
 NSizeLeverageExternalEPS_GrowEarnings_Yield
MeanSDMinMaxMeanSDMinMaxMeanSDMinMaxMeanSDMinMaxMeanSDMinMax
  1. Group = 0 means firms do not belong to certain business group; Group = 1 means firms belong to certain business group. Size is the log of firm's sales, Leverage is the debt ratio (total debt/total asset), External is the amount of external financing, EPS_Grow is the number of growth in EPS in the recent three years and Earnings_Yield is the earnings to stock price. We used the Worldscope database (2000), which contains up to ten years (1991–2000) of historical financial data from home-country annual reports of publicly-traded companies around the world, to calculate the financial variables above. To compute these variables, we require that each firm have income statement and balance sheet information for at least three consecutive years.

Group = 04913.0541.4549.26316.26925.61119.3180.00070.8010.0420.075−0.1370.2621.9391.0880.0003.0006.2117.968−32.88732.200
Group = 113613.3091.26410.88516.44727.23020.3770.00073.3500.0520.070−0.0590.5711.3161.1910.0003.0002.63611.810−62.61821.852
t–test of the diff. Group (0 minus 1)−1.089−0.496−0.8653.348***2.333***

Tables 11 reports the estimated results by adding a group dummy variable, which is equal to 1 if a firm belongs to a business group and 0 otherwise, when the earnings management is proxied by AEM. Table 12 addresses the issue of endogenous earnings yield (Earnings_Yield) where the instruments are all other explanatory variables (except for earnings yield), including group dummy variable and country-level governance indices as instruments. Again, the earnings management is proxied by AEM.9 Results of Tables 11 and 12 are similar as those obtained from Table 7. Our findings are, therefore, robust to the consideration of business groups and the endogeneity of the earnings yield. More importantly, the coefficients of group dummy variables are overwhelmingly insignificant, indicating that whether firms belong to a business group does not affect the extent of earnings management, at least in our sample firms.

Table 11. AEM prediction function: business group effect
 CG6CG7
(A)(B)(C)(D)(E)(F)(G)(H)
AEMAEMAEMAEMAEMAEMAEMAEM
  • ***, ** and *

    represent the level of significance at 0.01, 0.05 and 0.10, respectively.

Constant187.231*** (4.228)321.683*** (4.202)133.556*** (3.453)278.885*** (3.526)192.962*** (4.306)332.732*** (4.289)122.276*** (3.032)282.274*** (3.513)
CG−0.773** (−1.981)−0.788* (−1.848)  −0.864** (−2.246)−0.912** (−2.139)  
Group−9.231 (−1.178)−7.277 (−0.863)−11.998 (−1.536)−6.678 (−0.790)−9.136 (−1.155)−7.315 (−0.859)−67.914 (−1.060)−6.724 (−0.788)
Size−2.814 (−0.996)−1.573 (−0.500)9.959** (2.070)−1.811 (−0.571)−2.778 (−0.971)−1.571 (−0.498)14.662** (2.133)−1.729 (−0.544)
Leverage0.239 (1.294)0.123 (0.608)−2.424 (−1.529)0.117 (0.578)0.225 (1.217)0.110 (0.546)−2.494 (−1.595)0.098 (0.485)
External38.111 (0.644)14.526 (0.230)−685.482 (−1.510)3.342 (0.052)43.184 (0.717)21.914 (0.343)−756.742* (−1.728)11.570 (0.179)
EPS_Growth−2.651 (−0.341)−1.059 (−0.120)−68.191 (−1.051)−1.091 (−0.123)−3.661 (−0.464)−1.765 (−0.198)−68.345 (−1.034)−1.528 (−0.170)
Earnings_Yield0.153 (0.429)0.119 (0.314)−2.325 (−0.864)0.032 (0.081)0.140 (0.390)0.109 (0.286)−2.089 (−0.782)0.002 (0.005)
PGDP −0.004*** (−2.828) −0.003*** (−2.613) −0.004*** (−2.854) −0.003*** (−2.678)
Antidirect −19.055*** (−2.833) −0.657 (−0.021) −19.742*** (−2.851) 0.208 (0.007)
LAW 20.844** (2.205) −18.040 (−0.511) 22.115** (2.231) −10.861 (−0.302)
Disclosure −4.003*** (−2.636) −4.222 (−1.272) −4.156*** (−2.627) −4.862 (−1.439)
Insider 18.988** (2.542) 72.549** (2.171) 19.637** (2.597) 72.004** (2.182)
CG × Size  −0.180*** (−3.469)   −0.237*** (−2.952) 
CG × Leverage  0.038* (1.655)   0.038* (1.692) 
CG × External  11.048 (1.594)   12.397* (1.846) 
CG × EPS_Growth  0.948 (1.028)   0.947 (1.001) 
CG × Earnings_Yield  0.034 (0.878)   0.030 (0.777) 
CG × Antidirect   −0.278 (−0.608)   0.306 (0.509)
CG × LAW   0.568 (1.142)   0.487 (0.950)
CG × Disclosure   0.002 (0.054)   0.009 (0.200)
CG × Insider   −0.789 (−1.642)   −0.769 (−1.624)
R20.4750.4620.5010.4740.4800.4710.5110.484
Adjusted R20.4540.4200.4690.4230.4600.4310.4770.433
N185170185170185170185170
Table 12. AEM prediction function: instrumental variable methods
 CG6CG7
(A)(B)(C)(D)(E)(F)(G)(H)
AEMAEMAEMAEMAEMAEMAEMAEM
  • ***, ** and *

    represent the level of significance at 0.01, 0.05 and 0.10, respectively.

Constant161.592*** (3.890)363.269*** (3.751)87.691** (2.402)273.359*** (3.580)185.888*** (4.213)377.709*** (3.670)112.967*** (2.991)307.679*** (3.696)
CG−1.291*** (−2.651)−1.209** (−2.210)  −1.297** (−2.529)−1.272* (−1.818)  
Size0.355 (0.129)−0.120 (−0.039)11.551** (2.420)−0.470 (−0.147)−2.332 (−0.828)−1.661 (−0.513)10.873** (2.288)−2.502 (−0.728)
Leverage0.406 (1.605)0.137 (0.652)−3.501** (−2.488)0.093 (0.464)0.385 (1.551)0.081 (0.362)−2.184 (−1.409)0.112 (0.548)
External77.555 (0.605)−12.066 (−0.082)−102.477 (−0.297)−23.690 (−0.207)164.825 (1.162)118.605 (0.771)−758.361* (−1.727)103.922 (0.682)
EPS_Growth−10.069 (−1.367)−6.257 (−0.692)50.848 (0.943)−0.846 (−0.101)−5.181 (−0.667)−0.942 (−0.089)46.177 (0.717)−2.084 (−0.238)
Earnings_Yield1.782 (0.645)1.204 (0.401)−3.068 (−1.161)−0.315 (−0.137)2.923 (1.037)2.354 (0.768)−1.738 (−0.643)2.142 (−0.719)
PGDP −0.003* (−1.838) −0.003** (−2.187) −0.005*** (−2.584) −0.005** (−2.382)
Antidirect −26.266** (−2.449) −22.681 (−0.416) −25.756** (−2.304) −48.980 (−0.656)
LAW 23.417** (1.987) −22.681 (−0.814) 25.096** (2.318) 2.664 (0.065)
Disclosure −4.271** (−2.112) −4.727 (−1.027) −5.039*** (−2.592) −1.873 (−0.345)
Insider 15.458 (1.351) 62.042* (1.801) 25.871** (2.083) 48.652 (1.043)
CG × Size  −0.162*** (−3.197)   −0.181*** (−3.600) 
CG × Leverage  0.054*** (2.705)   0.034 (1.504) 
CG × External  1.998 (0.393)   12.324* (1.832) 
CG × EPS_Growth  −0.800 (−1.065)   0.645 (0.699) 
CG × Earnings_Yield  0.047 (1.261)   0.026 (0.669) 
CG × Antidirect   −0.805 (−1.229)   0.409 (0.371)
CG × LAW   0.651* (1.722)   0.326 (0.576)
CG × Disclosure   0.012 (0.151)   −0.056 (−0.575)
CG × Insider   −0.695 (−1.188)   −0.265 (−0.312)
R20.4610.4550.4890.4820.4770.4700.5030.483
Adjusted R20.4440.4210.4630.4410.4590.4330.4740.436
N185170185170185170185170

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References

This paper studies the impacts of corporate governance on earnings management. The index of corporate governance was taken from Credit Lyonnais Security Asia, who study firms in nine Asian countries. The measures of earnings management include earnings smoothing and earnings aggressiveness defined by Leuz et al. (2003) and Bhattacharya et al. (2003). Our conclusion is as follows.

First, corporate governance indices are found to be negatively related to the extent to which firms manage earnings. That is, firms with good corporate governance tend to conduct less earnings management. Second, there is a size effect for earnings smoothing, that is, large firms are prone to conduct earnings smoothing. It is worth noting that this effect or other types of size effect do not exist for the management of earnings aggressiveness, but good corporate governance indices seem to mitigate the size effect, on average.

Third, there is a turning point for leverage effect, i.e. when corporate governance index is higher than 60–70, leverage effect exists, otherwise, reverse leverage effect exists. Namely, higher leveraged firms conduct earnings smoothing more in a good governance regime but less in weak governance regime. It represents that a highly leveraged firm with poor governance is scrutinised closely by the market and thus finds it harder to fool the market by manipulating earnings.

Fourth, firms with higher growth (lower earnings yield) tend more actively to engage in earnings smoothing and earnings aggressiveness, but good corporate governance can mitigate the effect.

Finally, being consistent with the findings of Leuz et al. (2003), a firm in a country with good anti-director rights engages in less earnings smoothing and earnings aggressiveness. Contrary to those found in Leuz et al. (2003), stronger anti-director rights may result in stronger earnings smoothing. This effect, however, appears in low firm-level governance countries only, and not in high firm-level governance countries. Stronger enforcement of laws can result in less earnings smoothing, which is also consistent with Leuz et al. (2003), but this effect is stronger in countries with worse corporate governance. Therefore, on predicting earnings management across countries, in addition to country-level governance (such as Leuz et al., 2003), the role of corporate governance should have to be taken into account.

Notes
  • 1

    Following Leuz et al. (2003), we did not consider the effect of discretionary and non-discretionary earnings management.

  • 2

    The report in April 2001 entitled “Saints and Sinners: Who's Got Religion?” and “Make me Holy … but not yet!” in February 2002.

  • 3

    The names of the firms in CLSA were based on a literal translation, which may not have matched the actual names of the firms.

  • 4

    Taiwan provides tax deductions for high technology firms and other deemed crucial industries. Hong Kong is a free trade harbour. India also provides many tax deductions to encourage investments. Hence, this may be the reason of low smoothing management in the three countries.

  • 5

    More specifically, we attempted different weights, i.e. CG1, CG1.5 and CG2 in WLS. Reported results are based on CG1.

  • 6

    If firm size can be used as a proxy for information asymmetry and earnings management is measured by earnings smoothing, managers may choose to smooth earnings to lower earnings volatility and, in so doing, convey more valuable, more relevant information to uninformed investors (Fukui, 2000; Goel and Thakor, 2003). Larger firms have, therefore, higher incentives to engage in earnings smoothing.

  • 7

    For example, Claessens and Fan (2002) interpret that, in Asia, agency problems have been exacerbated by low corporate transparency, associated with rent-seeking and relationship-based transactions, extensive group structures and diversifications, and risky financial structures.

  • 8

    We key in the names of firms as well as the key words of subsidiary, consolidation and group in Google and Yahoo! to examine any relevant news.

  • 9

    Since the results are qualitatively the same if the earnings management is proxied by EM1, EM2 or EM3, we present the results of AEM only.

Appendix

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References
Table A1. CLSA corporate governance attributes
Governance attributeCLSA explanation
Disclosure attribute
Transparency (TRAN)The ability of outsiders to assess the true position of a company
(1) Disclosure of financial targets (e.g. three- and five-year ROA/ROE)
(2) Timely release of Annual Report
(3) Timely release of semi–annual financial announcements
(4) Timely release of quarterly results
(5) Prompt disclosure of results with no leakage ahead of announcement
(6) Clear and informative results disclosure
(7) Accounts presented according to IGAAP
(8) Prompt disclosure of market-sensitive information
(9) Accessibility of investors to senior management
(10) Website where announcements are updated promptly
Non-disclosure attribute
Discipline (DISC)Public commitment to CG and financial discipline
(1) Explicit public statement placing priority on CG
(2) Management incentivised toward a higher share price
(3) Sticking to clearly defined core business
(4) Having an appropriate estimate of cost of equity
(5) Having an appropriate estimate of cost of capital
(6) Conservation in issuance of equity or dilative instruments
(7) Ensuring debt is manageable, used only for projects with adequate returns
(8) Returning excess cash to shareholders
(9) Discussion in Annual Report on CG
Independence (INDP)Board is independent of controlling shareholders and is separate from senior management
(1) Board and senior management treatment of shareholders
(2) Chairman who is independent from management
(3) Executive management committee comprised differently from the board
(4) Audit committee chaired by independent director
(5) Remuneration committee chaired by independent director
(6) Nominating committee chaired by independent director
(7) External auditors unrelated to the company
(8) No representatives of banks or other large creditors on the board
Accountability (ACCT)Proper accountability of management to the Board
(1) Board plays a supervisory rather than executive role
(2) Non-executive directors demonstrably independent
(3) Independent, non-executive directors constitute at least half of the board
(4) Foreign nationals presence on the board
(5) Full board meetings at least every quarter
(6) Board members able to exercise effective scrutiny
(7) Audit committee that nominates and reviews work of external auditors
(8) Audit committee that supervises internal audit and accounting procedures
Responsibility (RESP)Record of taking measures in case of mismanagement:
(1) Acting effectively against individuals who have transgressed
(2) Record of taking measures in cases of mismanagement
(3) Measures to protect minority interests
(4) Mechanisms to allow punishment of executive/management committee
(5) Share trading by board members fair and fully transparent
(6) Board small enough to be efficient and effective
Fairness (FAIR)Treatment of minorities:
(1) Majority shareholders treatment of minority shareholders
(2) All equity holders having right to call general meeting
(3) Voting methods easily accessible (e.g. through proxy voting Quality of information provided for general meetings)
(4) Guiding market expectations on fundamentals
(5) Issuance of ADRs or placement of shares fair to all shareholders
(6) Controlling shareholder group owning less than 40% of company
(7) Portfolio investors owning at least 20% of voting shares
(8) Priority given to investor relations
(9) Total board remuneration rising no faster than net profits
Other attribute
Social Awareness (SOCL)Labour and environmental issues:
(1) Explicit policy emphasising strict ethical behaviour
(2) Not employing the under-aged
(3) Explicit equal employment policy
(4) Adherence to specified industry guidelines on sourcing of materials
(5) Explicit policy on environmental responsibility
(6) Abstaining from countries where leaders lack legitimacy (e.g. Myanmar)
Table A2. Classifications of belonging to a business group
No.CountryFirm nameGroupNo.CountryFirm nameGroup
 1Hong KongLi & Fung1 39IndiaPfizer India1
 2Hong KongCathay Pacific1 40IndonesiaUnilever1
 3Hong KongSwire1 41IndonesiaRamayana1
 4Hong KongHK Electric1 42IndonesiaTelkom1
 5Hong KongKingboard Chemical1 43IndonesiaSari Husada1
 6Hong KongEsprit Holdings1 44IndonesiaBimantara Citra1
 7Hong KongSmarTone1 45IndonesiaAstra Int'l1
 8Hong KongNew World Dev1 46IndonesiaSemen Gresik1
 9Hong KongHutchison Whamoa1 47IndonesiaGudang Garam1
 10Hong KongTexwinca Holdings1 48IndonesiaHM Sampoerna1
 11Hong KongASM Pacific1 49IndonesiaTempo Scan1
 12Hong KongYue Yuen1 50IndonesiaIndofood1
 13Hong KongSouth China Morning Post1 51IndonesiaMatahari1
 14IndiaDr1 52IndonesiaIndah Kiat1
 15IndiaCastrol1 53IndonesiaIndocement1
 16IndiaCipla1 54KoreaDaelim Ind1
 17IndiaHero1 55KoreaLG Chenical1
 18IndiaBPCL1 56KoreaHyundai Motor1
 19IndiaAsian1 57KoreaSK Telecom1
 20IndiaHindalco1 58KoreaKia Motors1
 21IndiaITC1 59KoreaSK Corp1
 22IndiaColgate-Palmolive1 60KoreaLG Electronics1
 23IndiaABB1 61KoreaSamsung Heavy Industry1
 24IndiaNestle1 62MalaysiaBAT1
 25IndiaTata1 63MalaysiaTanjong1
 26IndiaCummins1 64MalaysiaRoad Builder1
 27IndiaRanbaxy1 65MalaysiaCarlsberg1
 28IndiaNIIT1 66MalaysiaUMW1
 29IndiaSiemens India1 67MalaysiaIOI Corp1
 30IndiaGrasim Industries1 68MalaysiaEON1
 31IndiaVSNL1 69MalaysiaGamuda1
 32IndiaReliance Indystries1 70MalaysiaMalakoff1
 33IndiaBajaj Auto1 71MalaysiaIJM1
 34IndiaTata Tea1 72MalaysiaMISC1
 35IndiaBritannia Industries1 73MalaysiaTan Chong1
 36IndiaIOC1 74MalaysiaSime Darby1
 37IndiaONGC1 75MalaysiaPgas1
 38IndiaTata Infotech1 76MalaysiaJTI1
 77MalaysiaGuinness1132ThailandSiam Makro1
 78MalaysiaResorts1133ThailandElect'y Geneat'g1
 79MalaysiaTelekom Malaysia1134ThailandPTTEP1
 80MalaysiaGenting1135ThailandUnited Broadcast1
 81IndonesiaMatahari1136ThailandTelecom Asia1
 82IndonesiaIndah Kiat1137Hong KongCLP0
 83IndonesiaIndocement1138Hong KongMTRC0
 84MalaysiaYTL Power1139Hong KongGiordano0
 85MalaysiaMPI1140Hong KongHK&China Gas0
 86MalaysiaGolden Hope1141Hong KongTechtronic Inds0
 87MalaysiaK. Guthrie1142Hong KongJohnson Electric0
 88MalaysiaTenaga1143Hong KongFirst Pacific Company0
 89MalaysiaTRI1144Hong KongTelevision Broadcasts0
 90MalaysiaMAS1145IndiaHindustan0
 91MalaysiaMagnum1146IndiaBSES0
 92MalaysiaB Toto1147IndiaCadbury0
 93PhilippinesPLDT1148IndiaPunjab0
 94PhilippinesGlobe1149IndiaGlaxo0
 95PhilippinesABS-CBN1150IndiaTISCO0
 96PhilippinesManila Electric Co1151IndiaHPCL0
 97PhilippinesSan Miguel Corp1152IndiaL & T0
 98SingaporeSPH1153IndiaBHEL0
 99SingaporeST Egg1154IndiaIPCL0
100SingaporeSC Logistics1155IndiaNalco0
101SingaporeSC Marine1156IndiaZee Telefilms0
102SingaporeNOL1157IndiaSatyam Computers0
103SingaporeKeppel Land1158IndiaMTNL0
104SingaporeCity Dev1159IndiaNovartis India0
105SingaporeCerebos1160IndiaWockhardt0
106SingaporeVenture1161IndiaM & M0
107SingaporeNatsteel Ltd1162KoreaKepco0
108SingaporeKeppel Corp1163KoreaHite Brewery0
109SingaporeCreative1164KoreaCheil Jedang0
110SingaporeA-P Breweries1165KoreaPosco0
111SingaporeSingTel1166KoreaShinsegae Dept Store0
112SingaporeDelgro1167MalaysiaNestle0
113SingaporeHaw Par Corp1168MalaysiaStar0
114SingaporeWing Tai1169PhilippinesJollibee Foods Corp0
115SingaporeCycle & Carriage1170PhilippinesPetron Corp0
116SingaporeRobinson1171PhilippinesAyala Corp0
117SingaporeSingland1172SingaporeSing Airlines0
118SingaporeParkway1173SingaporeMarco Polo0
119SingaporeDatacraft1174SingaporeFCC0
120SingaporeF&N1175SingaporeWant Wanr0
121SingaporeUOL1176TaiwanSPIL0
122TaiwanChina Steel1177TaiwanCompeq0
123TaiwanDelta1178TaiwanGiant0
124TaiwanChina Airlines1179TaiwanNien Hsing0
125TaiwanNan Ya Plastics1180ThailandAdv Info Service0
126TaiwanFormosa Plastics1181ThailandTot'l Access Com0
127TaiwanHon Hai1182ThailandDelta Elec Thai0
128ThailandThai Union Frozen1183ThailandBangkok Bank0
129ThailandSiam City Cement1184ThailandLand&Houses (F)0
130ThailandSiam Cement1185ThailandBangkok Exp'way0
131ThailandBEC World1    

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Earnings management measures
  5. Data of CLSA
  6. Earning management and governance
  7. Empirical evidence
  8. Conclusion
  9. Appendix
  10. References
  • Barth, M. E., Elliot, J. A. and Finn, M. W. (1999) Market Rewards Associated with Patterns of Increasing Earnings, Journal of Accounting Research, 37, 387413.
  • Becker, C. L., DeFond, M. L., Jiambalvo, J. J. and Subramanyam, K. R. (1998) The Effect of Audit Quality on Earnings Management, Contemporary Accounting Research, 15, 124.
  • Bhattacharya, U., Daouk, H. and Welker, M. (2003) The World of Earnings Opacity, The Accounting Review, 78(3), 641678.
  • Burgstahler, D. and Dichev, I. (1997) Earnings Management to Avoid Earnings Decreases and Losses, Journal of Accounting and Economics, 24, 99126.
  • Chen, K. C. W., Chen, Z. and Wei, K. C. J. (2003) Disclosure, Corporate Governance, and the Cost of Equity Capital: Evidence from Asia's Emerging Markets, Working Paper, Department of Finance, Hong Kong University of Science and Technology.
  • Chung, H. and Kallapur, S. (2003) Client Importance, Non-audit Fees, and Abnormal Accruals, Accounting Review, 78(4), 931955.
  • Claessens, S. and Fan, J. P. H. (2002) Corporate Governance in Asia: A Survey, International Review of Finance, 3(2), 71103.
  • CLSA (2002) Make me Holy … but not yet! CG Watch: Corporate Governance in Emerging Markets.
  • Dechow, P. and Skinner, D. (2000) Earnings Management: Reconciling the Views of Accounting Academics, Practitioners, and Regulators, Accounting Horizon, 14, 235250.
  • Degeorge, F., Patel, J. and Zeckhauser, R. (1999) Earnings Management to Exceed Thresholds, Journal of Business, 72(1), 133.
  • Doidge, C. A., Karolyi, G. A. and Stulz, R. M. (2004) Why do Countries Matter so Much for Corporate Governance? ECGI – Finance Working Paper No. 50/2004. Available at SSRN: http://ssrn.com/abstract=580883 or DOI: 10.2139/ssrn.580883.
  • Fukui, Y. (2000) Earnings Management with the Help of Historical Cost Accounting: Not for Managers but for Investors, Working Paper, Tohoku University.
  • Goel, A. M. and Thakor, A. V. (2003) Why Do Firms Smooth Earnings? Journal of Business, 76(1), 151192.
  • Healy, P. and Wahlen, J. (1999) A Review of the Earnings Management Literature and Its Implications for Standard Setting, Accounting Horizons, 13, 365383.
  • Ke, Bin (2001) Why Do CEOs of Publicly Traded Firms Prefer Reporting Small Increases in Earnings and Long Strings of Consecutive Earnings Increases? Working Paper, Department of Accounting, The Smeal College of Business Administration, The Pennsylvania State University.
  • Klapper, L. F. and Love, I. (2004) Corporate Governance, Investor Protection, and Performance in Emerging Markets, Journal of Corporate Finance, 10, 703728.
  • La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R. (1998) Law and Finance, Journal of Political Economy, 106, 11131155.
  • Lee, B. B. and Choi, B. (2002) Company Size, Auditor Type, and Earnings Management, Journal of Forensic Accounting, 3, 2750.
  • Leuz, C., Nanda, D. and Wysocki, P. D. (2003) Investor Protection and Earnings Management: an international comparison, Journal of Financial Economics, 69, 506527.
  • Levitt, A. (2000) Testimony Concerning Commission's Auditor Independence Proposal Before the Senate Subcommittee on Securities Committee on Banking, Housing, and Urban Affairs, 28 September, http://www.sec.gov/news/testimony/ts152000.htm.
  • Myers, L. and Skinner, D. (2002) Earnings Momentum and Earnings Management, Working Paper, University of Michigan.
  • Press, E. G. and Weintrop, J. B. (1990) Accounting-Based Constraints in Public and Private Debt Agreements: Their Association with Leverage and Impact on Accounting Choice, Journal of Accounting and Economics, 12(January), 6595.
  • Richardson, R., Tuna, I. and Wu, M. (2002) Predicting Earnings Management: The case of earnings restatements, Working Paper, University of Michigan Business School.
  • Schipper, K. (1989) Earnings Management, Accounting Horizons, 3, 91102.
  • Shen, Chung-Hua and Chih, Hsiang-Lin (2005) Investor Protection, Prospect Theory, and Earnings Management: An International Comparison of the Banking Industry, Journal of Banking and Finance, 29, 26752697.
  • Skinner, D. and Sloan, R. (2002) Earnings Surprises, Growth Expectations and Stock Returns, Review of Accounting Studies, 7(2), 289312.
  • Sweeney, A. P. (1994) Debt-covenant Violations and Managers' Accounting Responses, Journal of Accounting and Economics, 17, 281308.
  • World Competitiveness Yearbook (2002) IMD International.
  • Worldscope database (2000) Online Computer Library Center (OCLC).

Chung-Hua Shen is currently a professor in the Department of Money and Banking in National ChengChi University, Taiwan. He received his PhD from Washington University in the USA. He teaches financial market and financial institutions, global financial systems and corporate governance. He was the Fulbright scholar in 2000 and Eisenhower Fellower in 2006. He has published papers in Journal of Banking and Finance, Journal of Money, Banking and Credit, Journal of Econometrics, Journal of International Money and Finance, International Journal of Economics and Finance, Review of Quantitative Finance and Accounting, Southern Economic Journal, Journal Policy Modeling, Eastern Economic Journal, Journal of Macroeconomics, Pacific Basin Finance Journal, Journal of Business and Economics, International Economic Journal, Applied Economics, Applied Financial Economics and International Journal for Forecasting.

Hsiang-Lin Chih is currently an associate professor and the director in Department of Cooperative Economics at the College of Business, National Taipei University, Taiwan. He teaches corporate finance, financial institutions man agement, cooperative economics, nonprofit financial management, and corporate social responsibility. He has published papers in Academia Economic Papers, Journal of Business Ethics and Journal of Banking and Finance.