The analysis of association between institutional ownership and listed corporate's positive earnings management in China

(1)

The analysis of association between

institutional ownership and listed

corporate’s positive earnings

management in China

Elaine Sun

10825533

22/06/2015

Word count: 10858

MSc Accountancy and Control, variant Accountancy

Amsterdam Business School

Faculty of Economics and Business

University of Amsterdam

Supervisor: Dr.

Sikalidis Alexandros

(2)

ii

Statement of Originality

This document is written by Elaine Sun who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Acknowledgments

This paper was substantially completed by the end of author’s master career. I would like to express my gratefulness to Sikalidis Alexandros, my supervisor. He gave a lot of help during the whole process, from initial topic selection and modeling to data analyses and final conclusion. This paper benefits a lot from his insightful comments and professional suggestions.

(4)

iv

Abstract

This paper studies the association between institutional ownership and aggressive earnings management, including both accrual based earnings management and real earnings management, of A-shared listed companies in China. There are two types of institutional investors – transient and long-term oriented institutional investors. They have different effects on firm earnings manipulation decision. Existence studies think that these two types of institutional investors are mutually exclusive, therefore they are linear related to firms’ earnings management. Different from prior literature, this study points out that transient and long-term orient institutional ownership can co-exist in listed firms. The association between institutional ownership and earnings management is expected to vary along with the increasing of percentage of institutional ownership. The regression results of this paper support the hypothesis that there is an inverted U relation instead of a linear relation. Turning points are found. Before the turning point, institutional investment is expected to be positively related to earnings management. Higher institutional ownership will lead to more upwards earnings management. Institutional investors cannot take their advantages to eliminate earnings manipulation. On the contrary, after the turning point, institutional ownership is negatively related to earnings management. Institutional investors can be positively involved in corporate governance.

Key words: Institutional investors; Accrual earnings management; Real earnings management; Inverted U relation; Turning point.

(6)

2

1. Introduction

In March 1998, Kai Yuan Securities Funds were listed on the capital market by China Southern Fund Management Ltd. It formally started the development of institutional investment in Chinese capital market. The rise of institutional investment is the significant change of global financial architecture during the past three decades. Statistics results show that institutional investment has rapid development in China as well. Institutional investors have become the one of the major investors in capital market. According to statistical data, until December 2012, the proportion of institutional ownership has already exceed 70% of tradable market cap of A-Share stocks in China. Half of institutional investors have become the top ten shareholders of listed companies in China. Useem (1996) points out that institutional investor is playing a more and more important role in corporate governance. It means that institutional ownership has negligible influence on company management.

Financial report is a principal approach for most investors to analyze the performance of listed firms they invest in. Based on financial reporting, investors are able to find out corporate financial position and operational result from time to time. Agency theory points out that information asymmetry between ownership and agency leads to agency problems. Agency problems between outside investors and managers result in opportunistic earnings management. According to motivation of IPO earnings management (Teoh, Welsh & Wong, 1998), equity incentives (Cheng, Warfield, 2005) and effects of bonus schemes (Healy, 1985), we can expect that listed firms have motivations to manage earnings. Through accounting flexibility, listed firm manipulate earnings to twist accounting authenticity. False financial information which affects investor’s judgment of stock price and market return of their investments. Wrong investment decision reduces the efficiency of resource distribution in capital market. In recent years, it appeared the crisis of honesty and credit in listed companies’ honesty and credit following a series of accounting fraud and swindle being exposed such as the

(7)

scandal of Enron, the event of Yinguangxia and the event of China source. How to restrain listed firms’ earnings management behavior to improve the quality of accounting information has become an important topic that cannot be avoided to upgrade the capital market.

Generally, institutional investors are regarded as sophisticated investors who have advantages, compared to individual investors, in collection and analysis of information about listed company (Hand, 1990; Kim et al. 1997; Bartov et al. 2000). They have variable and more convenient channels to achieve listed corporate’s in-time information, including both public and private information. Institutional investors should have benefits to eliminate the phenomenon of earnings management. A growing amount of researches have been carried out to explore the relationship between institutional investment and listed corporate’s earnings management. However, the debate about the resultant effects of institutional investor’s involvement in corporate governance has never stopped. Prior research points out that institutional investor can either encourage opportunistic earnings management for short term target or positively monitor corporate governance to eliminate earnings management (Bushee, 1998). There is not a common comment on its influence on earnings management so far. Many critics, Porter (1992) and Bhide (1993) both hold the view that the frequent and fragmented characteristic of institutional investment discourage investors from playing monitoring role in the company which they invested in and encourage earnings management. Further, some researchers even hold more absolute view that all kinds of earnings management behavior are fraud practice (Goel & Thakor, 2003). Empirical analyses find that earnings management does negatively affect investors’ benefits. Earnings management is a typical Zero-sum Game. Between companies and stakeholders, one party’s success comes at the expense of another (Shipper, 1989). For example, earnings management leads to investors’ incorrect estimation of stock price (Chaney & Lewis, 1995) and improper allocation of capital resources (Chen & Yuan, 2004). However, some researchers have opposite view that institutional investors are positively

(8)

4

(Chung et al. 2002). Further, some studies categorize different types of institutional investors and test different influence on corporate earnings management according to different investment characteristics (Gaspar et al. 2005; Kon, 2007).

Although the literatures carried out to investigate the relationship between institutional investors and earnings management are abundant, there has not been consensus till now. This study aims at testing competing views mentioned above by focusing on the relationship between institutional ownership and increasing earnings management.

RQ: What is the association between institutional ownership and positive earnings

management in listed company in China?

Recently, almost all researches carried out to study the role of institutional investors in Chinese capital market are dealing with the relationship between investor and accrual earnings management. This paper contributes to the literature by extending the research into the effects of institutional investment on corporate earnings management including both aggressive accrual based and real management. Secondly, this paper contributes to this field by pointing out the inverted U relationship between institutional investors and positive earnings management to settle the debate mentioned above. It proves that the influence varies with the increase of institutional investors’ share holdings. This paper is carried out based on the assumption that short term and long term investors co-exist in one listed firm. Most co-existing researches carried out in China hold the view that there is a linear relationship instead of curve relationship, because they study short-term and long-term institutional ownership separately. This paper’s regression results support the assumption. Conclusion of this paper can help to explain contrary outcomes of previous researches and to settle the debate mentioned above. Finally, as a leader in capital market, this paper calls institutional investor to pay more attention to positively play a role in corporate governance to monitor earnings manipulation and improve efficiency of capital market. This paper gives authorities advices to encourage development of institutional investors as well.

(9)

The next section based on literature review develops hypothesis by testing the research question of association between percentage of institutional ownership and listed firm’s aggressive earnings management behavior in China. Section three presents the research methodology used in this paper to carry out empirical analysis, including the measurement of depended variables, independent variables and control variables, research model and description of the samples. Section four reports the results of descriptive statistics, empirical analysis and robustness analysis and finally, section five illustrates a summary and conclusion.

2. Literature review and hypotheses development

2.1 Theoretical Framework

The concept of earnings management is first carried out by Schipper (1989). He points out that earnings management is a disclosure management for private interests of certain people. Corporate manager controls reporting process of financial statements on purpose to misleading stakeholders’ investment decisions or influence the consequences of covenant based on these financial reports. In 1999, Healy and Whalen extended Schipper’s research by expanding the definition of earnings management. Their study found that managers not only can control reporting process, but also can manage real operation actions, investment activities and financing activities to manipulate earnings reporting to misleading stakeholders. Real earnings management makes investors to believe that this company reaches earnings target based on normal operation while, actually, with sub-optimal management decisions (Roychowdhury, 2006). In China, there a large amounts of researches regarding the definition of earning management as well. Some researches point out that earnings management is a kind of behavior which aims at meeting stakeholders’ perspective through abusing accounting standards (Lu, 1999). Zhang and Liu in 2002 went further to identify earnings management as surplus controlling at particular points for particular outcomes. They points out that earning manipulation is the choice of the corporate’s administration

(10)

6

without breaking the accounting rules. Therefore, it is essentially different from accounting frauds. In order to better study the research question of this paper, definition of earnings management stated by Healy and Whalen is used in this research.

Capital market is an important component of financial market. Efficient capital market theory is first carried out by Fama in 1970. Based on judgment of listed company’s earnings information and stock price, institutional investor aims at making profit by investing in securities market. Prior research points out that there is positive correlation between corporate reporting earnings information and stock price. Bernard and Thomas (1990) present evidence that higher reporting earnings correspond to relatively higher stock prices and reporting earnings involve information positively related to firm value. Efficient capital market theory also indicates that stock price can correctly reflect all the information of listed company. Efficient market theory has three strength reduction assumptions. Firstly, all investors in capital market are rational investors who can make rational assessment of inter-value of stock. They are aware of all consequence of every investment decision. They are able to react rapidly to in-time information in the market so that they can make unbiased estimation of stock price to increase capital market efficiency. Secondly, efficient capital market theory admits that there are part of irrational investors. They randomly make investment decision. The influence of non-rational investment behavior cancels each other out. Therefore they cannot significantly influence stock price. Finally, although some of the irrational investments have homogeneity, rational investors’ arbitrage trading can remove irrational investors’ negative impact on stock price. In the end, market is efficient. According to efficient market theory, institutional investors are rational investors and rational arbitragers. They have the ability to influence market efficiency through gathering and analyzing accounting information of listed firms.

However, on the other hand, if institutional investors’ excessive dependence on firms’ reporting earnings information, administrations of listed firm may try to benefit from aggressive earnings management which does harm to investors value maximization.

(11)

Information asymmetry is the may cause of low capital market efficiency. In the capital market, there are mainly two kinds of asymmetries. One is between investosr and company managers. Investors make investment decisions based on operation information, financial position and development prospect of companies. Then, agency problem exists. Corporate managers can make profit by manipulating earnings. The other is between institutional investors and individual investors. Lots of researchers find out that institutional investors have more informational advantages than individual investors. Institutional investors are sophisticated investors. Superiorities are represented in two aspects. (1) Scale advantage. The cost of information collection is almost fixed. With more capital volume, the cost shared by per capital will be less. (Jun et al. 2011). Compared with individual investors, institutional investors have more convenient information collection methods to further reduce cost. Lastly, institutional investors are on behalf of the interests of smaller investors. In order to maximize investment income within specified time and acceptable bound of risks, institutional investors are more cautious with investment decision by asking for more private and detailed information about listed companies instead of only disclosure reporting. (2) Professional advantage. Institutional investor has professional financial consultants to analysis and process data with high-level skills, especially accounting information which is highly related to investment decisions. (Balsam et al. 2000; Collins et al. 2003). Under the circumstances, sophisticated investors are less likely to be fooled by manipulated earnings information.

Based on these theories, institutional investor should be rational investor who is able to detect company’s bias earnings information and take advantage to discover earnings management, play positive role in corporate governance and enhance capital market efficiency. But in fact, plenty of studies figured out thoroughly inconsistent consequences. Therefore, the following literatures will further discuss the relationship between institutional ownership and earnings management.

(12)

8

2.2 institutional investor’s ownership and earnings management.

Porter (1992), Mitra (2005) and many other researchers all hold the view that the percentage of institutional holdings is the key point which decide whether or not institutional investors can mitigate earnings management. Recent researches find that different investment characteristics, such as the holding period and transaction features, may have various influence on relationship between institutional investor and listed firm’s earnings management. Generally, there is a common conclusion that higher institutional ownership discourages myopic management behavior. Therefore, larger sharing holdings by institutional investor help to eliminate earnings manipulation (Rajgopal et al. 1999). Chen (2007) classifies institutional investors into short-term oriented institutional investors and long-term oriented institutional investors. The writer finds that short-term institutional investor cannot efficiently supervise corporate earnings management behavior, while long-term institutional investor is positively involved in corporate governance to mitigate earnings management.

2.2.1 Short-term oriented institutional investors and earnings management.

Some studies hold the view that institutional investors are inherently short-term investors. They have two ways of investments. (1) Myopic behavior. This kind of investors only buy stock when stock price goes up and sell it when price goes down. They only care about short-term income. They probably cannot detect earnings management, so he cannot play accurate role in corporate governance (Louis, 2004). (2) Sensitive strategy. Investors are active arbitragers in capital market. Investment decisions are more sensitive to corporate earnings management behavior. Investors are able to correctly analysis firms’ accounting information, detect mispricing and make right prediction about firms’ future performance. Decision making strategy is that if the mispricing can be remedied and return to normal, investors sell stock of corporate with high level earnings management and buy stocks of firm with lower earnings management. Investment decisions are rational and can benefit from arbitrage trading.

(13)

Researches find that most short-term institutional investors are rational arbitrageurs. Investors are classified into three groups: transient, dedicated and quasi-indexer based on their past investment patterns in terms of portfolio turnover, diversification and momentum trading (Bushee, 1998). Transient investments are diversified. Institutional investors have short-term investments and frequent transactions. Investors are willing to sacrifice long-term benefits for short-term profits. This kind of investment decision is sensitive to listed firms’ accounting information. Dedicated investors hold concentrated investments. Investments have lower turnover rate and are based on long-term perspective. Investors do not purchase or sale shares easily and frequently. Quasi-indexers set up investment portfolios based on market index. Investors have diversified investment portfolios and have comparatively long-term share holdings. Along with classification, in their study of institutional investor’s influence on abnormal anomaly, Collis, Gong & Hribar (2003) found out that transient institutional investors can play a more significant role in short-term arbitrage trading because arbitrage is a short-term strategy which requires high turnover rate. Lev & Nissim (2006), also based on Bushee’s (1998) research results, indicate that dedicated and quasi-indexers institutional investors’ change of share holdings are more sensitively to accrual accounting information.

Empirical tests also demonstrate that short-term oriented institutional investors care more about current return on investment. Investors’ incentives to monitor manager are weakened. Short-term investment strategy encourages opportunism earnings manipulation. Because of frequent transaction, short-term institutional investors even aggravate corporate manager’s motivations to manage income to meet market requirement (Li et al. 2011). In conclusion, all these arguments point out that short-term institutional ownership is likely to be positively associated with earnings management. To be specific, more short-term institutional investments are related to more earnings management behavior of listed firms.

(14)

10

2.4 Long-term oriented institutional investors and earnings management

On the other side, some studies hold the view that institutional investors are long-term. They pay more attention to long-term benefits. Investors hold more concentrated investment portfolios with lower turnover rate. It allows investors to have more time and motivation to gather related corporate information. Investors take advantages to communicate with corporate shareholder to get more private information. Decision making is more cautious towards listed firms’ earnings management behavior. Investors are probably to sell stock or avoid investment when discovering high level earnings management.

Generally, long-term oriented institutional investors can positively affect corporate governance and play a role of monitoring. Empirical tests indicate that frequent transaction of stock cannot obtain income higher than market average level. Long-term oriented investors are more likely to benefit from concentrated investment (Lakonishok et al. 1992). The higher institutional ownership, the longer investment horizon and the more motivation to monitoring corporates as well (Gaspar et al. 2005). Long-term investment means that institutional investors are more interested in listed firm’s long-term value. It helps to reduce agency problems. There is high possibility that institutional investors with large share holdings take part in monitoring corporate management. In this case, they will suffer less losses when firm is in trouble. To be specific, with the increase of share holdings, the cost of selling stocks is higher, because bulk sale of stocks leads to share price decrease (Black & Coffee 1994). If institutional investors are negatively involved in corporate governance, when company has a poor performance, investors with larger stock holdings losses more (Pound, 1992).

In additional, when a firm’s shares are concentrated in a small number of institutional investors, especially institutional investors who are sophisticated investors, the total monitoring cost is lower. It is more convenient for all investors collect information, share information and collectively supervise corporate management. The monitoring

(15)

process is improved. Meanwhile, with lower total monitoring cost, the average monitoring cost of every institutional investor is lower. It gives investors more incentives to monitor firm they invested in (Kon, 2003).

Further, long-term institutional ownership promotes communication between investors and corporations. It is easier for long-term investors to get more related information which enhance effectiveness of corporate governance and supervision.

To sum up, all the opinions above demonstrate that long-term institutional investors concentrate more on firm’s future performance. They care about firm’s long-term profitability. Their investment strategy can reduce earnings manipulation because derived from short-term institutional Investors, they reduce the current income growth pressure put on firms. In conclusion, it is possible that long-term institutional investors have intense motivation to monitor firm accounting procedure and helps to reduce earnings management. Therefore, long term oriented institutional ownership is expected to be negatively related to earning management. It means that more long-term shareholdings lead to less earning management behaviors.

2.5 Testable hypothesis in the Chinese settings

Based on the literature review listed above, this paper finds that researches related to association between institutional ownership and earnings management have two competing views. One is that short-term oriented institutional investors have passive portfolio investment. Investors are lack of incentives to take the role of monitoring because of monitory holdings. At this moment, institutional investment has negative liner relation with earnings management. That is to say, the more short-term oriented institutional ownership, the higher level of earnings management. On the other side, long-term oriented institutional investors have positive portfolio investment. Investors care more about future benefit instead of current income. Listed firm’s performance is effectively monitored by investors and earnings management level can be brought down.

(16)

12

All these studies assume that institutional investors of a listed company are either all short-term or all long-term investors. Most of these prior studies hold the view that these two kinds of institutional investors are mutually exclusively. However, short-term and long-term oriented institutional investors can co-exist. In China, although there are lack of studies directly test the association between institutional ownerships and earnings management of listed firms, prior literatures can prove that long-term and short-term institutional investors can co-exist instead of mutually exclusive. Liu and Xu (2012) according to Yan and Zhang’s (2009) classification of institutional investors analyze the influence of short-term and long-term institutional ownerships on future equity yield and market volatility. They find that short-term investors are likely to cause market fluctuation while long-term investors are more like to play to stabilize the market. Yao and Huang (2010) carry out the study about association between institutional investments and earnings management of A-share listed company from 2004 to 2007. They classify institutional investors into five categories: securities investment funds, QFII, social security funds, insurance funds and security companies. They find that securities investment funds and QFII positively reduce earnings management of listed company while the other three kinds of institutional investors’ investments are not significant related to earnings management level.

From the above, it can be concluded that with different classification of institutional investors, the outcomes of research questions can be different. It is reasonable to assume that short-term and long-term institutional investors co-exist in Chinese market. When both long-term and short-term investors exist, their influences on listed firm’s earnings manipulation behavior are opposite. For example, short-term oriented investors pursue maximization of current profit. It encourages listed firms to manipulate income upwards to increase stock price to meet investor’s requirement. On the contrary, long-term oriented investor seeks long-long-term value maximization. Investors have incentives to supervise to limit short-term earnings management behavior (Koh, 2003). Under this circumstance, the key point is to analyze which investor takes a dominant role. If long-term institutional investors are superior, there is expected to be a positive influence on

(17)

corporate governance. While if short-term institutional investors are superior, more earnings management is likely to occur (Luo, 2013).

Huang and Gao (2010) find out that the association between the influence of institutional ownership on insider control of listed company and earnings management. Through empirical analysis, researchers draw the conclusion that the association is an inverted-U relation in Chinese securities market. They are among the few researchers in China who point out the curve association instead of linear relationship between institutional ownership and earning management. The turning point of the concave relation is 26% ownership of institutional investors compared to total shares outstanding of listed company. That is to say that different share ratios of institutional investors have various influence on earnings manipulation. When share ratio is less than 26%, as outside investors, institutional investors are a passive shareholder. Insiders control on financial reporting is more aggressive (Klein, 2002). The level of listed firm’s earnings management is higher. When the ratio is above 26%, investors have stronger motivations to monitor corporate performance because of self-interest. It can help to relieve insiders’ earnings management.

All the analyses above lead to a hypothesis that association between institutional ownership and earnings management is not a linear relation. There should be a turning point before which institutional ownership does not eliminate earnings management behavior. On the contrary, it encourages more opportunistic earnings manipulation. Above the percentage, a negative association is expected. Therefore, this paper develop the following hypothesis.

H1: The association between institutional ownership and positive earnings management of listed companies in China follows an inverted-U relation.

(18)

14

3 Research Design

3.1 Measuring institutional ownership

There is not a universal definition of institutional investors. According to the traditional definition, institutional investor is different from individual investors. Brancato (1997) thinks that capital of institutional investor is managed by professional staffs or organizations, including private pension funds, retirement funds, mutual funds, closed-end trust funds and all kinds of donation funds. Securities Law of China also briefly and similarly defines institutional investor as investor except for individual investor. This paper will use this definition to carry out the research.

Derived from research of Collins, Gond and Hribar (2003), the percentage of institutional ownership is 𝐼𝑂_𝑖𝑡 which is calculated as the stocks owned by institutional investors divided by the total shares outstanding of listed company.

3.2 Measuring earnings management

Measurement of earnings management mainly focuses on companies’ aggressive accrual-based and real-based earnings management. This paper holds the opinion that accrual-based earnings management and real earnings management can co-exist in listed company. As explained in the theory section, managers has the incentive to up-ward manage earnings to meet institutional investors’ requirement. Institutional investors who are sophisticated investors can take information advantages to detect and help to reduce firms’ earnings management behavior.

3.2.1 Measurement of Accrual Earnings Management

One means of earnings management is manipulation of discretionary accruals through judgment in financial reporting without direct cash flow consequence. Derived from the research of Jones (1991), this paper uses cross-sectional analysis to calculate total

(19)

accruals of listed companies. Expectation model is listed below. 𝑇𝐴𝑖𝑡 𝐴𝑖𝑡−1= 𝛼1( 1 𝐴𝑖𝑡−1) + 𝛽1( ∆𝑅𝐸𝑉𝑖𝑡 𝐴𝑖𝑡−1) + 𝛽2(( 𝑃𝑃𝐸𝑖𝑡 𝐴𝑖𝑡−1) + 𝜀𝑡 （1） Where:

𝑇𝐴_𝑖𝑡: Total accruals in year t, for firm i which is measured by net income of firm less CFO.

∆𝑅𝐸𝑉_𝑖𝑡: Revenues in year t less revenues in year t-1 for firm i;

𝑃𝑃𝐸_𝑖𝑡: Gross property, plant, and equipment in year t for firm i;

𝐴𝑖𝑡: Total assets in year t for firm i; 𝜀𝑖𝑡: Error term in year t for firm i.

With this expectation model, three regression coefficients, 𝛼_𝑖, 𝛽_1𝑖 and 𝛽_2𝑖 can be obtained. Use these coefficients in the following formula to calculate the normal accruals. 𝑁𝐴_𝑖𝑡 = 𝛼₁( 1 𝐴𝑖𝑡−1) + 𝛽1( ∆𝑅𝐸𝑉𝑖𝑡−∆𝐴𝑅𝑖𝑡 𝐴𝑖𝑡−1 ) + 𝛽2(( 𝑃𝑃𝐸𝑖𝑡 𝐴𝑖𝑡−1) + 𝜀𝑡 (2) Where:

∆𝐴𝑅_𝑖𝑡: Accounts receivable in year t less accounts receivable in year t-1 for firm i.

Finally, the amount of accrual-based earnings management can be calculated by actual accrual based earnings management minus normal accrual based earnings management.

𝐴𝐵𝑆_𝐷𝐴_𝑖𝑡 = 𝑇𝐴𝑖𝑡

𝐴𝑖𝑡−1− 𝑁𝐴𝑖𝑡 (3)

(20)

16

𝐴𝐵𝑆_𝐷𝐴_𝑖𝑡 means downside accrual earnings management. Higher absolute value of ABS_𝐷𝐴_𝑖𝑡 illustrates higher extent of earnings management.

3.2.1 Measurement of Real Earnings Management

Companies can also manage earnings by manipulate their operational decisions. Derived from the idea of Roychowdhury (2006), there are three kinds of real earnings management behavior. First, sales manipulation. For example, companies accelerate sales through price discount or loose credit terms. It will lead to increasing operational income in short term to gain higher current profit. But cash flow cannot match income. Therefore, this kind of management can be measured by abnormal cash flow. Normal Cash flow can be estimated by the following formula:

𝐶𝐹𝑂𝑖𝑡 𝐴𝑖𝑡−1 = 𝛼1( 1 𝐴𝑖𝑡−1) + 𝛽1( 𝑆𝑖𝑡 𝐴𝑖𝑡−1) + 𝛽2( ∆𝑆𝑖𝑡 𝐴𝑖𝑡−1) + 𝜀𝑡 (4) Where:

𝑆𝑖𝑡: Sales in year t for firm i;

∆𝑆𝑖𝑡: Sales in year t less in year t-1 for firm i.

Every firm year, the abnormal cash flow AB_CFO equals actual CFO/𝐴_𝑖𝑡−1less normal (estimated) calculated by the model above. If there is sales manipulation, AB_CFO should be negative.

Secondly, reduction of discretionary expenditures which are usually R&D expenses. It means that managers have the right to decide whether and when to incur the expenditure. Managers can choose to reduce reported expenses to increase current profits. This is most likely to happen when such expenditure will not generate profits immediately and managers have the motivations to meet earnings targets by decreasing discretionary expenditures. However, this practice probably takes a risk of lower cash flow in the future. It can be measured by abnormal discretionary expenses. Normal discretionary

(21)

expense is measured by: 𝐸𝑋𝑃𝑖𝑡 𝐴𝑖𝑡−1 = 𝛼1( 1 𝐴𝑖𝑡−1) + 𝛽 ( 𝑆𝑖𝑡−1 𝐴𝑖𝑡−1) + 𝜀𝑡 (5)

Abnormal discretionary expenditure equals actual EXP/ 𝐴_𝑖𝑡−1 minus normal discretionary expense. When companies want to raise current earnings, AB_EXP should be negative. On the contrary, AB_EXP should be positive.

Thirdly, overproduction or increasing production to report lower COGS. With higher outputs, fixed costs are spread over a larger number of units. With lower COGS per unit, firm reports better operating margins. However, supplies are much more than demands. It implies that cash flow income from operation is lower than normal given sales level. This management can be measured by abnormal production costs. Normal production costs can be estimated by the formula below:

𝑃𝑅𝑂𝐷𝑖𝑡 𝐴𝑖𝑡−1 = 𝛼1( 1 𝐴𝑖𝑡−1) + 𝛽1( 𝑆𝑖𝑡 𝐴𝑖𝑡−1) + 𝛽2( ∆𝑆𝑖𝑡 𝐴𝑖𝑡−1) + 𝛽3( 𝑆𝑖𝑡−1 𝐴𝑖𝑡−1) + 𝜀𝑡 (6)

Abnormal production costs AB_PROD equals reported PROD/𝐴_𝑖𝑡−1 less estimated normal production costs. Positive AB_PROD illustrates the existence of production manipulation.

Based on existing study, this paper uses a combined proxy to represent all three kinds of real earnings management (Cohen et al. 2008). We calculate this combined proxy using the following formula.

COM_REM = AB_CFO − AB_PROD + AB_EXP (7)

Both accrual earnings management and real earnings management are lagged by prior year total assets. Lagged variables are used to eliminate the endogenous problem of

(22)

18

variables. This paper will mainly focus on positive earnings management because this paper holds the view that managers have more incentives to manage earnings upwards to meet institutional investors’ earnings target.

3.3 Model specification for hypothesis testing

In order to test the hypothesis which expects an inverted-U relation between institutional ownership and listed company’s earnings management, this paper uses the following regression models.

Firstly, the relationship between institutional investment and increasing accrual earnings management is tested by formula 8.

𝐴𝐵𝑆_𝐷𝐴𝑖 = 𝛽1𝐼𝑂𝑖+ 𝛽2𝐼𝑂𝑖2+ 𝛽3𝑆𝐼𝑍𝐸𝑖 + 𝛽4𝑃/𝐵𝑖 + 𝛽5𝐴𝐷𝐽_𝑅𝑂𝐴𝑖+ 𝛽6𝐵𝐼𝐺 𝑁𝑖+ 𝛽7𝐿𝐸𝑉𝑖+ 𝛽8𝐿𝑇𝐴𝐶𝐶𝑖+ 𝜖𝑖 (8)

According to hypothesis of this paper, institutional ownership should be positively related to abnormal accrual based earnings management while square of institutional ownership should be negatively related to abnormal accrual based earnings management. It means that 𝛽₁ should be positive and 𝛽₂ should be negative. Then the institutional investment and discretionary earnings management will have an inverted U relation as expected.

Secondly, the relationship between institutional investment and real earnings management is tested by the similar way.

𝐶𝑂𝑀_𝑅𝐸𝑀_𝑖 = 𝛽₁𝐼𝑂_𝑖 + 𝛽₂𝐼𝑂_𝑖2+ 𝛽₃𝑆𝐼𝑍𝐸_𝑖+ 𝛽₄𝑃/𝐵_𝑖 + 𝛽₅𝐴𝐷𝐽_𝑅𝑂𝐴_𝑖 + 𝛽₆𝐵𝐼𝐺 𝑁_𝑖+ 𝛽₇𝐿𝐸𝑉_𝑖+ 𝛽₈𝐿𝑇𝐴𝐶𝐶_𝑖+ 𝜖_𝑖 (9)

(23)

positive while coefficient of IO2_{should be negative as well. Under this assumption} there will be a turning point of percentage of institutional ownership, before which institutional ownership is positively related to real earnings management while after which institutional investment will negatively associated with earnings management. Considering that institutional ownership is not the only factor that influences listed company’s earnings management. Several control variables are set to filter out other considerations which may influence manager’s motivation to manipulate income. These control variables are firm size, market-to book equity ratio, industry-adjusted ROA, BIG4, Leverage, and lagged total accruals.

Size: Natural logarithm of market value of equity (MVE). It is difficult to estimate the

sign of coefficient of firm size. One the one hand, managers of bigger firms have more motivations to manage earing to attract more both shareholders and debtholders. One the other hand it is more difficult for managers to manipulate earnings because of better corporate governance in bigger firms and more monitoring from stakeholders.

P/B ratio: P/B ratio is measured as market value per share divided by book value per

share. It is hard to identify the sign of P/B ratio’s coefficient.

ADJ-ROA: Industry adjusted return on Asset. It is calculated by the following method.

ADJ _ROA =

𝑛𝑒𝑡 𝑖𝑛𝑐𝑜𝑚𝑒 + 𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑒𝑥𝑝𝑒𝑛𝑠𝑒 𝐴𝑖𝑡−1

It is expected that ADJ_ROA is positively related to earnings management. Firstly, when managers’ bonus is directly related to financial ratio such as ROA, they have more motivation to manage earnings. Secondly, when institutional investors mainly use financial ratio to evaluate firm performance, managers are more willing to manage earnings upwards to achieve goal.

(24)

20

BIG N：BIG N is a dummy variable which equals 1 when listed firm is audited by a

BIG N auditor, including KPMG, EY, DTT, PWC, Lixin, Zhong Rui Yue Hua, Tianjian, Xin Yong Zhong He, Daxin and Dahua. Auditor is expected to be negatively related to earnings management. BIG N represents better audit quality. With higher level of audit quality, listed firms are less likely to manipulate earnings.

LEV: Leverage is calculated as total long-term debt divided by total assets. Long-term

debt is related to firms’ long-term performance. It has conflicts with firms’ current performance. When institutional investors are focusing on short-term target, listed firms may sacrifice its long-term benefits by reducing long-term debt. Therefore, it is expected to be negatively related to earnings management.

LTACC: LATCC is lagged total accruals for listed firms which equals total accruals in

prior year. As sophisticated investors, as mentioned above, they will reduce investment or avoid investment when detecting earning management behaviors. Therefore, it is predicted that the relationship between last year earnings management and institutional ownership. However, it is hard to expect its direction of impact on current year earnings management.

3.4 Sample Selection

Listed firms’ financial information and institutional ownership data used in this research are obtained from the database of Tai’an. The selection criterial of sample firms used in this study includes time horizon which starts from 2010 to 2013 because that the global financial crisis of 07-08 influenced the economic market significantly. . There are, in total, 2133 (2010-2013) Listed A-shared companies on Chinese Stock markets. This paper excludes sample firms which are financial listed companies, because financial company’s business practice and accounting treatment are particularly different others industries. The measurement of financial listed company’s earnings management should be studied separately. Firms who miss necessary data used for

(25)

earnings management measurement are also excluded in this research. Firms with diversify resources or industrial are exclude as well. Firms with negative earnings management are no included in this paper because this paper mainly focuses on the influence of institutional investors’ investment decision on aggressive earnings management. This paper use random sampling method to choose observed value to compare with listed firm’s financial report published by CSRC (China Securities Regulatory Commission). No significant differences are detected. It can prove the accuracy of observations to a certain extent. In order to avoid the influence of extreme value on the outcomes of empirical analyses, this paper uses Winsorize method to exclude minimum (below 1% of total amount) and maximum (upper 1% of total amount) to achieve more reliable results.

Under the sample selection criteria, the total amount of sample firms used for testing is 244. There are 107 samples firms in 2010, 125 in 2011, 111 in 2012 and 61 in 2013. In total, there are 402 observations.

(26)

22

4. Results

4.1 Descriptive Results

Table 1 illustrates the descriptive statistics for dependent variables, independent variables and control variables of this paper. Firstly, the mean and medium of positive accrual earing management divided by last year total asset are 6.58% and 4.92% respectively. Regarding abnormal real earnings management, the mean is 9.89% and medium is 7.80%. Institutional ownership, on average, hold 33.82 percentage of total shares of sample observations. 25 % of samples’ institutional ownerships are lower than 6.78%. Three quarters of sample firms have less than 46.02% institutional holdings. This ratio is relatively low compared to other countries with mature capital market, such as American whose average ratio is much more than 50%. From descriptive statistics, it seems that listed firms with same institutional investment level have higher degree real earnings management than accrual earnings management. This needs further discussion in the following regression analyses.

Then, in terms of other variants, the long-term debt to total asset ratio of samples are relatively low in China. 50% of samples have a leverage ratio of less than 38.26% while only 25% of samples have a ratio of more than 58.26%. On the one hand, it is a good signal that most of listed firms in Chinese capital market do not have significant financial risks. On the other hand, lower leverage ratio may lead to worse liquidity of these listed firms. Long-term debt is normally related to long-term investment decision such as research and development. Lower leverage ratio may be a negative signal of scarifying future benefit to meet current financial goals. Regarding auditors, more than half of samples are using Big N auditors. Taking firm size and big N audit firm both into consideration, from the table we can tell that bigger firms are more willing to use Big N audits which is closely related to high quality financial information. The mean of price-to-book equity ratio which is measured by market value of equity divided by book value of equity is 3.332. It shows great difference between firm shares’ book value

(27)

and market value. The average and median of LTACC are 1.14% and 1.18% respectively.

Further, this paper carries out correlation analyses on all explanatory by using Pearson correlation coefficient test. When the correlation coefficients between each explanatory variants are less than 0.6, control variants will not significantly influence the empirical outcomes (Hossian, 1994). Table 2 below illustrates the outcomes of the test. Size is positively related to leverage level (LEV). Cotter’s finding in 1998 illustrates that bigger firms are more likely to hold long-term debt. Firm size is positively related to Big N as well. Bigger firms are more willing to use Big N auditors. Adjusted return on asset is also significantly positively associated with firm size. It means that bigger firm is more likely to obtain higher ROA. As expected, institutional ownership (IO) is negatively related to prior year lagged total accruals (LTACC). Institutional investors are likely to reduce investment when discovering earning management. That is to say, it is possible that institutional investors are able to detect abnormal earnings management or institutional investors can positively influence listed firms’ earnings manipulation. According to the table, it can be concluded that although dependent variants and control variants have influence on each other to a certain extent, they do not have the issue of multi-collinearity, therefore, the results of empirical analyses will not be significantly influenced.

(28)

24 Table 1 Descriptive Statistics N First Quartile Median Third Quartile Mean Std. Deviation IO 402 0.0678 0.2014 0.4602 0.3382 0.8755 IO^2 402 0.0045 0.0406 0.2119 0.2564 0.9206 SIZE 402 0.8513 1.055 1.239 1.047 0.2745 P/B 402 1.9170 2.847 4.411 3.332 1.818 ADJ-ROA 402 0.0352 0.0548 0.0786 0.0588 0.0336 BIG N 402 0 1 1 0.62 0.486 LEV 402 0.2206 0.3826 0.5682 0.3096 0.7051 LTACC 402 0.0267 0.0118 0.0469 0.0114 0.0609 ABS_ACC COM_REM 402 402 0.0202 0.0262 0.0492 0.0780 0.0892 0.1266 0.0658 0.0989 0.0619 0.0874

ABS_ACC, positive abnormal accrual based earnings management; COM_REM, positive abnormal real earnings management; IO, percentage of institutional ownership; IO2_{, square of percentage of} institutional ownership (IO); SIZE, natural logarithm of total assets; P/B ratio, the ratio of market value per share to book value per share; ADJ-ROA, lagged return on asset; BIG N, dummy variable which equals 1 when firm is audited by a BIG N auditor, otherwise 0; LEV long-term debt to asset ratio; LTACC, lagged last year total accruals.

(29)

Table 2

Correlations

IO IO^2 SIZE P/B

ADJ-ROA

BIG N LEV LTACC

IO Pearson Correlation 1

Sig. (2-tailed)

IO^2 Pearson Correlation .943

**

1

Sig. (2-tailed) .000

SIZE Pearson Correlation .487

** .424** 1 Sig. (2-tailed) .000 .000 P/B Pearson Correlation .278 ** .268** .622** 1 Sig. (2-tailed) .000 .000 .000 ADJ-ROA Pearson Correlation .417** .376** .396** .267** 1 Sig. (2-tailed) .000 .000 .000 .000

BIG N Pearson Correlation .017 .004 .017 -.019 -.005 1

Sig. (2-tailed) .732 .940 .734 .711 .922

LEV Pearson Correlation -.096 -.094 .314

**

-.225** -.211** .008 1

Sig. (2-tailed) .054 .060 .000 .000 .000 .875

LTACC Pearson Correlation -.032 -.031 .069 .040 .031 .035 .111

*

1

Sig. (2-tailed) .519 .533 .170 .423 .541 .485 .026

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

IO, percentage of institutional ownership; IO2_{, square of percentage of institutional ownership (IO); SIZE, natural logarithm of total assets;} P/B ratio, the ratio of market value per share to book value per share; ADJ-ROA, lagged return on asset; BIG N, dummy variable which equals 1 when firm is audited by a BIG N auditor, otherwise 0; LEV long-term debt to asset ratio; LTACC, lagged last year total accruals.

(30)

26

4.2 Regression results

4.2.1 Regression results of relationship between institutional investment and increasing accrual earnings management

Table 3 demonstrates the multiple regression results of model 1 which tests the relationship between institutional ownership and increasing accrual earnings management. As expected, there is a positive relationship between institutional investment (IO) and increasing accrual earnings management (ABS_DA) and a negative relationship between square of institutional investment (IO2) and abnormal discretionary earnings management. Results reported in Table 3 indicate that estimated coefficients are statistically significant (0.0440 and 0.0282 respectively). These support the inverted U relation instead of a linear relation between institutional ownership and accrual earnings management as estimated in the hypothesis part of this paper. Optimal regression model is used to study the concave relationship. There is a turning point in the curve which is drawn according to the regression model. At the turning point, the earnings management degree of listed firm is maximized with the certain amount of institutional holdings of firm shares. This paper assume that on the left side of this point, institutional investors turn out to be positively related to aggressive earnings management. While on the right side of this point, completely opposite relationship will be observed. According to the estimated coefficients for IO and IO2, at the turning point, the percentage of institutional ownership is calculated as the following formula:

(31)

Table 3: Regression Results of Model 1:

𝐴𝐵𝑆_𝐷𝐴_𝑖 = 𝛽₁𝐼𝑂_𝑖+ 𝛽₂𝐼𝑂_𝑖2+ 𝛽₃𝑆𝐼𝑍𝐸_𝑖 + 𝛽₄𝑃/𝐵_𝑖 + 𝛽₅𝐴𝐷𝐽_𝑅𝑂𝐴_𝑖+ 𝛽₆𝐵𝐼𝐺 𝑁_𝑖 + 𝛽₇𝐿𝐸𝑉_𝑖+ 𝛽₈𝐿𝑇𝐴𝐶𝐶_𝑖+ 𝜖_𝑖

Variable Expected sign Estimated Coefficients (p-value)

Constant ? .016 (.3441) IO + .333 (.0440) IO2 - -.362 (.0282) SIZE ? .045 (.0182) P/B ? .002 (.4343) ADJ-ROA + .143 (.039) BIG N - -.004 (.0450) LEV - .066 (.0710) LTACC ? .530 (.0545) Adjusted R2 .066 Sample Size (N) 402

ABS_ACC, positive abnormal accrual based earnings management; IO, percentage of institutional ownership; IO2_{, square of percentage of institutional ownership (IO); SIZE, natural logarithm of} total assets; P/B ratio, the ratio of market value per share to book value per share; ADJ-ROA, lagged return on asset; BIG N, dummy variable which equals 1 when firm is audited by a BIG N auditor, otherwise 0; LEV long-term debt to asset ratio; LTACC, lagged last year total accruals.

(32)

28

Figure 1

Fig. 1. The intercept in this figure is processed according to the computation procedure which is pointed out by Utama and Cready in 1997. The estimated coefficients of independent variables reported in table 3 are used in this optimal regression model. The estimated coefficients of control variables are multiplied against their average value. The sum of results is used as the constant in the model. Thus, this figure represents the association between institutional ownership and behavior of firms’ positive accrual earnings management under an ‘average’ firm change basis.

As assumption, turning point is the maximization point. That is to say, before which the turning point, sample firms engage in more increasing discretionary earnings management when institutional ownership increases. However, when institutional investment reaches 45.99% or is more than the turning point of institutional holdings, institutional investors perform to be play positive involved in reducing firms’ earnings management behavior. This is mainly because that at this point, institutional investors are more interested in term investment. They are more interested in their long-term benefits. Therefore, the more percentage of shares they are holding, the more incentives they have to positively involving in monitoring listed firms’ earnings

(33)

management behavior. It can also be interpreted as that transient institutional investors are not sensitive to accrual earnings management signals. Consistent with previous research, they cannot take their advantages and play their role effectively. They have tendency of opportunistic behavior. They will chase after go up and kill cheapen which is negatively affect efficient capital market.

Meanwhile, regarding control variables, regression results show that some variables are significantly related to earnings management. Firm size and adjusted return on assets are significant positively related to accrual earnings manipulation while auditor is negatively related to discretionary accrual earnings management with significant influence. These results suggest that expect independent variables, these variables may also partly affect firm’s earnings management. As explained in descriptive results, firm size is significantly and positively associated with firm’s adjusted return on asset. All the rest control variables are insignificant. LEV ratio and LTACC ratio have positive relationship with discretionary accrual earnings management decision, however they will not significantly influence manager’s earnings management decision.

4.2.2 Regression results of relationship between institutional investment and increasing real earnings management

Regression results also support the concave association between institutional ownership and positive real earnings management. Under the significant level of 0.05， the percentage of institutional investment significantly affect earnings management with a positive coefficients of 0.560 and the square of institutional ownership is significantly influence firm’s real earnings management decision with a negative coefficients of -0.584. The maximization point of this model calculated by the same formula should be

(34)

30

Table 4: Regression Results of Model 2:

𝐶𝑂𝑀_𝑅𝐸𝑀_𝑖 = 𝛽₁𝐼𝑂_𝑖 + 𝛽₂𝐼𝑂_𝑖2+ 𝛽₃𝑆𝐼𝑍𝐸_𝑖 + 𝛽₄𝑃/𝐵_𝑖 + 𝛽₅𝐴𝐷𝐽_𝑅𝑂𝐴_𝑖 + 𝛽₆𝐵𝐼𝐺 𝑁_𝑖 + 𝛽₇𝐿𝐸𝑉_𝑖 + 𝛽₈𝐿𝑇𝐴𝐶𝐶_𝑖 + 𝜖_𝑖

Variable Expected sign Estimated Coefficients (p-value)

Constant ? .090 (.1940) IO + .560 (.0452) IO2 - -.584 (.0491) SIZE ? .0142 (.0433) P/B ? .018 (.0600) ADJ-ROA + .651 (.0160) BIG N - -.098 (.0010) LEV - -.240 (.083) LTACC ? .0796 (.0641) Adjusted R2 .125 Sample Size (N) 402

ABS_ACC, positive abnormal accrual based earnings management; COM_REM, positive abnormal real earnings management; IO, percentage of institutional ownership; IO2_{, square of}

percentage of institutional ownership (IO); SIZE, natural logarithm of total assets; P/B ratio, the ratio of market value per share to book value per share; ADJ-ROA, lagged return on asset; BIG N, dummy variable which equals 1 when firm is audited by a BIG N auditor, otherwise 0; LEV long-term debt to asset ratio; LTACC, lagged last year total accruals.

(35)

Figure 2

Fig. 2. Similarly, the estimated coefficients of independent variables reported in table 4 are used in real earnings management optimal regression model. The estimated coefficients of control variables are multiplied against their average value. The sum of results is used as the constant in the model. Thus, this figure represents the association between institutional ownership and behavior of firms’ positive real earnings management under an ‘average’ firm change basis as well.

Turning point under model 2 is a little bit higher than that in model 1 (47.95% compared to 45.99%). It may be because that real earnings management is easier to be manipulated by manager while is harder to be detected by institutional investors. Managers have more choices to manipulate firm performance under real earnings management. Institutional investors cannot tell the evidence of real earing management behavior only based on financial report. Although, institutional investors are more sophisticate investors compared to others, they still need pay more effort to gather more information including both financial and non-financial information. It will be more costly for institutional investors to monitor firm performance. Therefore, the incentive of institutional investors to positively involve in monitoring firm management is

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0 % 1 0 % 2 0 % 3 0 % 4 0 % 5 0 % 6 0 % 7 0 % 8 0 % 9 0 % 1 0 0 % IN CR EA SIN G CO M -R EM INSTITUTIONAL OWNERSHIP

(36)

32

eliminated. As a result, the turning point is lagged.

The results of this paper consist with the hypothesis which holds the view that the association between institutional ownership and earnings management is a curvilinear relation. It is different from existing theories which only focus on liner relationship between institutional ownership and earnings management. Under the assumption that listed firms have co-existence of both short-term and long-term institutional investors, this turning point can help to explain the conflict opinions towards the association between institutional ownership and earnings management. As a result, institutional investors may negatively influence firms’ earnings management when the percentage of shares hold by institutional investors is relatively low. At this point, majority of institutional investors are transient oriented investors, they mainly focus on listed firms’ short-term return of investment. Therefore, managers have more motivation to engage in earnings management to meat investors’ requirement. On the contrary, when the percentage of institutional ownership is relative high, long-term oriented institutional investors are playing a major role. They care more about listed firms’ ability to making value for them in the long-term so they are more willing to take responsibilities to monitor firms’ earnings management which will negatively affect their benefit.

4.3 Sensitive analysis

4.3.1 Accrual earnings management

In the previous section, this paper uses Jones model to measure accrual earnings management. Later studies adjusted this model based on Jones’ theory. Dechow and Dichev (2002) takes the influence of operational cash flow on accruals in to consideration. Kothari, Leone and Wasley (2005) point out that performance evaluation should also be added in the formula to calculate abnormal accruals. In order to strengthen the persuasion of empirical results, this paper will do the robustness test by using the adjusted model carried out by Kothari, Leone and Wasley (2005). Their study

(37)

finds that including performance variance can reduce noisy and increase the reliability of calculated accrual earnings management. The formula used is listed below,

𝑇𝐴𝑖𝑡 𝐴𝑖𝑡−1 = 𝛼1( 1 𝐴𝑖𝑡−1) + 𝛽1( ∆𝑅𝐸𝑉𝑖𝑡 𝐴𝑖𝑡−1) + 𝛽2(( 𝑃𝑃𝐸𝑖𝑡 𝐴𝑖𝑡−1) + 𝛽3𝑅𝑂𝐴𝑖𝑡+ 𝜀𝑡 (10)

The only difference between formula 10 and 1 is 𝑅𝑂𝐴_𝑖𝑡 which is calculated by revenue divided by total asset at the beginning of the period. ROA is the representative of firm performance. The regression results of model 1 which using the modified abnormal accruals are similar to the original results. Institutional ownership is positively related to increasing abnormal accrual earnings (Beta=0.436). The square of institutional ownership is negatively related to increasing discretionary accruals (Beta=-0.508). And the turning point moves left to 42.91%. It means that when performance measurement is taken into consideration, accrual earnings management is more accurate estimated, institutional investors can better involved in corporate governance. They can better eliminate earnings management with lower ownership or shorter period of investment.

4.3.2 Endogenous Problem of Variables

Gao’s study which studies the relationship of institutional ownership and earnings manage in 2008 points out that endogenous problem of variables is the most significant limitation which affect the empirical results. This paper also holds the view that it is possible that the endogenous issue exists in this study. Therefore, this paper tries to carry out sensitive analyses to control endogenous problem.

Firstly, all variants used in this paper are lagged variants. Lagged variables are used to eliminate endogenous problem. Then this paper is going to employ two-stage least squares method to carry out robustness test. This paper chooses institutional ownership and PE ratio as tool to carry out two-stage least squares test. It can be found that the regression outcomes of robustness test is similar to those of original empirical analyses. Both estimated coefficients for IO and IO2_{are significant under the 5% significance}

(38)

34

level and have the expected sign. The negative coefficients for IO2_{proves the} inverted-U relation between institutional investment and aggressive earnings management. It is consistent with hypothesis.

(39)

5. Conclusion

5.1 Summary

This paper tests the association between institutional ownership and listed firm’s aggressive earnings management including both accrual earnings management and real earnings management in China. Existence literatures hold two contrary opinions towards the relationship between institutional ownership and earnings management decisions. Some of the studies reach the conclusion that institutional investment negatively affects listed firms’ earnings management. They hold the view that institutional investors focus on short-term benefits, therefore investments give managers more incentives to manage performance to meet investors’ requirement. Under this circumstance, institutional investors cannot properly play the role in monitoring firm management. However, the others hold the opposite view that institutional investors can positively monitor firm performance to decrease firms’ earnings management behavior. Because they think that compared to individual investors, institutional investors are sophisticated investors who have more advantages to detect firm earnings management. Institutional investors are investors who care more about their long-term benefit. They are more willing to take part in corporate governance. Therefore, with more institutional investment, the degree of firms’ earnings management will decrease.

However, this paper holds the view that both these two kinds of opinions have the same defects that they holds the assumption that each firm only has single type of institutional investors either transient investors or long-term investors. They are mutually exclusive. Therefore, the association between institutional ownership and aggressive earnings management is linear relation. This paper points out that this assumption has hypothetical error. In reality, short-term and long-term institutional investors co-exist in listed firms. The relationship depends on which type of investors are playing the

(40)

36 inverted-U relation instead of liner relation.

The results of this paper support the hypothesis of concave relation. As predicted, there is a positive relationship between institutional ownership and earnings management. The turning point for accrual earnings management is 45.99% institutional ownership and the turning point for real earnings management is 47.95%. After the turning point, a negative association is found. These findings can demonstrate that short term and long term investors can co-exist. The influence of total percentage of institutional ownership on firm’s earnings management depends on which type of investors are playing the dominant role in corporate governance. With more short-term oriented investors, managers are more motivated to manage earnings to meet short-term goals of investors. While with more long-term oriented investors, corporate performance is monitored by investors to focus on long-term goal, earnings management can be effectively eliminated.

5.2 Contribution

This research can benefit future research by extending the assumption that transient and long-term oriented investors can co-exist in listed firms. Taking this extended assumption into consideration, the debate towards the role of institutional ownership in eliminating earnings management can be explained and settled. Moreover, it can help to provide research evidence to calls for a more dominate role for institutional investors in corporate governance of listed firms they invested in, especially for firms with majority of short-term institutional investors. Because with more short-term investors, there will be higher possibility of earnings management. Institutional investors should be aware of the possibility and take actions to reduce earnings management behavior. It highlights the significance to use adjusted accounting-based performance evaluation to measure firm performance. In terms of the positive relationship, long-term oriented investors can help to eliminate earnings management. Therefore, for more efficient capital market, there should be less legal barriers to insure that institutional investors

The analysis of association between institutional ownership and listed corporate's positive earnings management in China