“Earnings management and seasoned equity offerings: evidence from Taiwan started Go Incubation Board for Startup and Acceleration firms”

Although a large body of empirical research focuses on listed companies, less is do- ne regarding small and medium enterprises. Under the authorities’ support, Taipei Exchange (TPEx) started Go Incubation Board for Startup and Acceleration Firms (GISA) in January 2014. This research yields insight into earnings management activities around seasoned equity offerings (SEO) based on GISA firms in Taiwan and the effectiveness of external corporate governance. Data for the study come from the GISA Market Observation Post System of TPEx and Taiwan Economic Journal. The results reveal that GISA firms with the incentives of raising funds are prone to upward accru-al-based earnings management during SEO to avoid long-term negative consequences. Especially, firms with paid-in capital more than TWD (NT$) 30 million, higher fundraising amounts, or smaller-sized firms, tend to increase discretionary accruals. Finally, Certified Public Accountants (CPAs) and Big 4 accounting firms effectively serve as external corporate governance on mitigating earnings management. This study makes some contributions to GISA literature. First, expands the prior research, the different earnings management level before and listed on GISA, to the firms listed on GISA. Second, link up the relationship between the SEO and earnings management of GISA in Taiwan. Finally, it provides several contributions to regulators, for instance, the ef- fectiveness of the counseling system provided by CPAs or Big 4 accounting firms. Also, the CPAs and Big 4 accounting firms serve as supervisors on corporate governance.


INTRODUCTION
According to the White Paper on Small and Medium Enterprises in 2016 published by the Ministry of Economic Affairs, enterprises' paidin capital less than NT$ 50 million takes up nearly 98%, which equals 1,383,981 firms of total enterprises in Taiwan. Only the remaining 1.4% and 0.6% of companies meet the criteria of paid-in capital in emerging listing and over-the-counter listing. Therefore, it is hard for small and medium enterprises (SMEs) to raise funds from the capital market.
To simplify the securities issuance procedures of the start-up company and improve the information transparency, the US set up Jumpstart Our Business Startups Act (JOBS Act) in 2012. Two years later, in 2014, the Taipei Exchange (TPEx) have modeled Go Incubation Board for Startup and Acceleration Firms (GISA) after US JOBS Act Title III Crowdfunding, providing the SMEs with innovation, creativity, and paid-in capital under NT$ 50 million, options to raise funds on the capital market. In other words, the GISA is an equity-based crowdfunding portal without the function of equity transactions. The GISA portal can provide a bridge between companies and investors to increase information transparency and financial report reliability. Until the end of May 2017, 77 enterprises were listed on the GISA and effectively raised nearly NT$ 245 million.
Prior literature suggests the relationship between listed initial public offering (IPO) or seasoned equity offerings (SEO) companies and earnings management on the US stock market (Cohen & Zarowin, 2010;Fan, 2007). However, the earnings management activities around SEO and external corporate governance based on GISA firms in Taiwan have received little attention. Also, there is little difference in set-up purposes between the GISA board and the stock market. Following the obligations, the GISA companies are required to disclose the information on the GISA Market Observation Post System. Thus, GISA firms that aim to raise funds from the capital market or to list on the mainboard in the future may have the incentives to conduct earnings management. However, most of SMEs in Taiwan do not meet the criteria to list on TPEx or TWSE. If these SMEs listed on GISA and have the incentives to SEO, they may make efforts to attract more outsiders' capital investment. One of the methods to raise more funds or be listed on the mainboard is to conduct earnings management. The research mainly focuses on the companies listed on the GISA board. This study aims to discuss whether or not SEO firms have more earnings manipulation for those GISA firms than non-SEO firms. Furthermore, it compares the pre-and post-earnings performance during the next SEO period. Finally, it is examined whether the certified public accountants (CPAs) and Big 4 accounting firms can effectively lighten the earnings manipulation behavior.
Overall, this study makes several contributions to the GISA and earnings management literature. First, GISA firms conducting SEO have higher incentives to conduct discretionary accruals than non-SEO firms. Next, outsiders cannot effectively lessen or heighten earnings management behaviors from management. Besides, this paper expands the prior research of earnings management that firms have different earnings management level before and after listing on GISA in an emerging economy. Finally, this paper contributes to regulators that the CPAs and Big 4 accounting firms serve as supervisors on corporate governance. The results imply the effectiveness of the counseling system provided by CPAs or accounting firms.
The remaining part of this paper is organized as follows. Section 1 reviews the relevant prior research and introduces the hypotheses. Furthermore, this paper presents the research model and discusses the research methodology in section 2, and the experimental results are subsequently analyzed in section 3. Last section presents the conclusion of the study and provides crucial suggestions and future implications.

LITERATURE REVIEW
This paper will review the literature concerning the GISA, earnings management, and external corporate governance. The first sub-section gives an overview of GISA in Taiwan, including the establishment purpose, the motivations on GISA firms, and its current situation. The second sub-section reviews the literature on earnings management, focusing on IPO and SEO. The last sub-section analyzes prior research related to the effect of external corporate governance on earnings management.

Go Incubation Board for Startup
and Acceleration Firms (GISA) According to the White Paper on Small and Medium Enterprises in Taiwan published by the Ministry of Economic Affairs in 2016, the number of SMEs with paid-in capital below NT$ 50 million is apart from the existing emerging established stock market and mainboard (previously called as OTC) listing for public offerings. Table 1 shows that the government hoping the multi-level capital markets will pave the way for economic and business development in Taiwan. Based on the agency theory, Chung, Firth, and Kim (2002) develop executives reporting incentives to manage earnings in the book titled "Positive Accounting Theory". Companies may offer different contracts to insiders or outsiders, such as compensation plan or debt covenant, to alleviate agency conflicts and reduce agency problems. However, the measurement criteria of these contracts are based on the accounting information, which will easily be altered once the managements are under the compensation of debt pressure. Moreover, the bigger the corporation, the higher the political pressure they face. In order not to raise politically sensitive, the management may have a higher propensity to choose the income-decreasing accounting policy. The earnings management motivations are derived from the following hypotheses.
The bigger operating size or better profitability, the more political pressure companies should bear. The pressure on the company makes the managers lower the earnings to get less attention from the government and supervisors. The prior literature also indicates that the government or overseers keep eyes on the large companies more than the small ones. The big corporations will downward earnings to lessen unnecessary political costs (Kothari, Leone, & Wasley, 2005). Sheng (1997) also proves that large companies with higher earnings performance will cause attention from society. Gunny (2005)  degree of accrual earnings manipulation will be larger for debt repayment than plant expansion. Several kinds of literature investigate the joint decision of discretionary accruals and asset dispose of SEO. The authors find that in SEO and previous SEO period, the companies have higher discretionary accruals. Besides, the results show that discretionary accruals and assets disposal is a complementary correlation in SEO and pre-SEO period and prove that managements consider two methods jointly.

Accrual earnings management
According to the prior research, the accruals can be divided into discretionary and non-discretionary. Discretionary accruals mean that the managers adjust cash flow from an opportunity set of GAAP. Non-discretionary shows the manager adjust cash flows authorized by accounting standard-setting bodies. Owing to the information asymmetry between management and stakeholders, the manipulate behaviors are hard to detect. Enterprises often use discretionary accruals. To sum up, discretionary accrual is a good index of earnings management. Chin et al. (2000) put forward the method using total accruals, including both discretionary and non-discretionary components, to measure earnings management. Cohen and Zarowin (2010) estimated by change in total accruals. Healy (1985) indicates that the residual from a regression of total accruals on change in sales and property, plant, and equipment. Further, Healy (1985) proposes the Modified Jones Model, which is the residual from a regression of total accruals on change in sales and property, plant, and equipment, where revenue is adjusted for changes in receivable in the event during the period. DeAngelo (1986) calculates residual from the regression of non-cash current assets as fewer liabilities on lagged levels of these balances and adjusted for the increase in revenues, expenses and property, plant, and equipment.

External corporate governance in mitigating earnings management
Corporate governance is a system used to direct and supervise an organization's performance. Dechow, Sloan, and Sweeney (1996) find that through corporate governance, internal control mechanisms not only can improve the reliability of the financial statements but also mitigate the earnings manipulation.
GISA is different from Main Board without setting corporate governance regulation. However, during the public integrative counselling step, the counselor will provide a standard internal control system for the GISA firms as an example. With sound corporate governance mechanisms, companies can work more efficiently to improve information transparency and reduce the possibility of accounting fraud.
Prior literature proves that financial statements audited by CPAs can lessen the information asymmetry and protect stockholders' equity (Fama & Jensen, 1983). CPA with high auditing quality can inhibit earnings management level due to the management's fear of disclosure (Dechow et al., 1996). Financial reports audited by CPAs can enhance credibility. However, according to GISA regulation, CPAs auditing is not a mandatory item if firms' paid-in capital is less than NT$ 30 million.
The evidence shows that the US companies with Big 4 auditors have higher earnings quality than the firms with non-Big 4 auditors (Teoh & Wong, 1993). Kim, Chung, and Firth (2003) find that Big 6 auditors are more effective than non-Big 6 auditors in detecting opportunistic earnings management. Francis and Wang (2008) indicate that only for firms with the well-known international Big 4 auditors, the earnings quality is better, and investor protection becomes stronger. Thus, Big 4 auditors enforce higher reporting quality. These findings suggest that better quality of audit quality is related to effective monitoring, limiting accrual-based earnings management.
Brickley, Coles, and Terry (1994) prove that when the percentage of outsiders increases, the discretionary accruals would decrease. Later, Peasnell, Pope, and Young (2005) also indicate the negative relationship between outsiders' proportion and managers' income-increasing abnormal accruals. On the other hand, Shivakumar (2000) finds that to influence investors' decisions and have a better offering price, and the management tends to manipulate earnings before SEO. That is to say, have a better offering price or receive sufficient funds from outsiders' investors (higher outsiders' subscription rate), the SEO firms tend to manage earnings around SEOs; and the better the earnings performance, the higher possibility to receive more investment amount. Most of the prior literature verifies that the external corporate governors, such as certified public accountants and Big 4, alleviate the agency problems among management, stockholders, and stakeholders (OECD). As for outsiders' subscription rates, there are mixed results in the literature. In general, the better external corporate governance, the better effectiveness of mitigating earnings management behavior.

Hypotheses development
In the past, almost 98% of the companies in Taiwan  According to registrations on GISA, one of the missions that the registered GISA firms should complete is to reach the capital raising target. However, the GISA firms listed not only for the channel of fundraising but also looking forward to listing on the mainboard and getting sufficient resources from the capital market in the future. Thus, this paper expects that GISA firms with the financial needs share the same characteristic as the mainboard listed companies. If the GISA firms have the incentives to raise funds or list on the mainboard in the future, the SEO firms will have a higher level of earnings management behaviors than non-SEO firms. Therefore, SEO firms with fundraising needs are likely to have earnings management behavior. Thus, the first hypothesis is formulated as follows: H1: After the GISA has been listed, the company conducting SEO has a higher level of earnings management than non-SEO.  Shivakumar (2000) finds that to have a better offering price or receive sufficient funds from outsiders' investors, and the management tends to conduct earnings management around SEOs. According to these empirical results, the second hypothesis is set as follows: H2a: For SEO firms listed on GISA, the level of earnings management is lower when audited by the CPAs.
H2b: For SEO firms listed on GISA, earnings management level is lower when audited by Big 4 accounting firms.
H2c: For SEO firms listed on GISA, there is an association between the earnings management level and outsiders' subscription rate.  Roychowdhury (2006). Previous studies (Healy, 1985; DeAngelo, 1986) use total accruals as a proxy for earnings management while Jones (1991), Dechow, Sloan, and Sweeney (1995) use discretionary accruals as a measure of earnings management. The reason to exclude non-discretionary accruals is that they are used to reflect business conditions that the managers cannot control. This paper considers a modified version of the Jones Model in the empirical analysis. In the modified model, non-discretionary accruals are estimated during the event period. The advantage of Modified Jones Model is to eliminate the conjectured tendency in measuring discretionary accruals with an error when managers use the discretion right for revenues: , 01 , 01 , 01

Sample selection
The

Descriptive statistics
For  Table 5 reports the Pearson correlation coefficient and Spearman rank correlation coefficient between the various proxies from the data set. To examine the multi-collinearity, a case of multiple regression is used in which the predictor variables are themselves highly correlated. If the absolute value of the correlation coefficient is larger than 0.7, two variables are highly correlated.

Correlation analysis
The correlation between DA and AMT (Total fundraising amount) are significantly correlated. By examining Variance Inflation Factor in regression equations (7), (8), and (9), the result for AMT (Total fundraising amount) is .17, 1.19, and 1.26, respectively, which is less than 5. This will not cause a huge impact. Also, there is a correlation between CPA and Big 4, but these two are the separate independent variables in different regressions. Thus, there is no multi-collinearity problem among the variables. Tables 6 and 7 show the empirical results of the earnings management activities around SEO of GISA firms, examine the impact of CPAs, Big 4 accounting firms, and outsiders' subscription rate of SEO firms on earnings management. Note: DA, the difference between total discretionary accruals and non-discretionary accruals. SEO, seasoned equity offerings. Company, conduct SEO equals 1; otherwise 0. CPA, firm audit by certified public accountants equals 1; otherwise 0. BIG4, firm audit by Big 4 accounting firms equals 1; otherwise 0. PIC, paid-in capital larger than NT$ 30 million equals 1; otherwise 0. FY, company foundation year. Natural logarithm of company i in year t. LOSS, an indicator variable that equals 1 if the firm is reporting a loss; otherwise 0. GROW, growth, measured by the difference between sales t and t − 1 divided by sales t − 1. LEV, leverage. Debt to assets ratio for company i in year t.

Main results for hypotheses
To start with, Table 6    funds from the market or listed on the mainboard in the future, the smaller GISA firms will have upward discretionary accruals. Roychowdhury (2006) finds that the bigger GISA firms will downward discretionary accruals to minimize political costs. For column (2), the CPAs' capability in mitigating earnings management among GISA firms, including SEO and non-SEO firms, is examined. The coefficient on CPA is −0.335 (p-value = 0.061), negatively related to earnings management. For column (3), the Big 4 accounting firms' capability in alleviating earnings management behaviors among GISA firms is examined. The coefficient on BIG4 (firm audit by Big 4 accounting firms) is −0.350 (p-value = 0.092), negatively related to earnings management at the 0.1 level.
The previous empirical result proves that firms with SEO purposes tend to have earnings management behavior than non-SEO firms. Table 7 reports the effect of the supervision from CPAs on accrual-based earnings management. Column (4) shows that the CPAs can effectively reduce the discretionary accruals; the coefficient is −0.421 (p-value = 0.072), significantly negative at 0.1 level. These results also prove that public integrative counseling from CPAs had paid off and effectively reduced earnings manage-ment level. Specifically, the estimation coefficient on AMT (Total fundraising amount) is 0.001 (p-value = 0.000), significantly positive at the 0.01 level. To raise sufficient funds from the capital market, the firms would upward manipulate discretionary accruals. The coefficient on CINV (SEO purposes for capital investment) is −0.680 (p-value = 0.038), negatively related to earnings management. Thus, SEO purposes for capital investment are not the main reason for conducting earnings management.
Column (5) shows the same results as column (4) that Big 4 accounting firms provide better earnings quality in mitigating discretionary accruals. The coefficient of BIG4 is −0.505 (p-value = 0.020) significantly negative at 0.05 level. This result also proves that public integrative counseling from international accounting firms worked and effectively reduce the level of earnings management. The estimation coefficient on AMT (Tot a l f u ndraising amount) is 0.001 (p-value = 0.000), significantly positive at the 0.01 level. The firms would upward manipulate discretionary accruals to raise sufficient funds from the capital market. The coefficient on CINV (SEO purposes for capital investment) is −0.856 (p-value = 0.011), negatively related to accrual-based earnings management. Column (6) shows that there is no relationship between outsiders' subscription rate and earnings management, especially in SEO firms in the GISA board, the coefficient is 0.238 (p-value = 0.360), not significant. This result reveals that either viewpoint of outsiders and earnings management has been adopted, especially for SEO firms in GISA in Taiwan. The estimation coefficient on AMT (Total fundraising amount) is 0.001 (p-value = 0.000), which is significantly positive at the 0.01 level, which remains the same as prior regression. The coefficient on CINV (SEO purposes for capital investment) is −0.640 (p-value = 0.059), negatively related to earnings management.
This paper predicts the company, which is listed on GISA. Conducting SEO has a higher level of earnings management than non-SEO in H1. The empirical results support H1 in several ways. First, compared with non-SEO firms, the SEO firms with further SEO purposes will conduct earnings management, especially using discretionary accruals. Besides, there is a positive relationship between paid-in capital and earnings management. Firms with paid-in capital more than NT$ 30 million tend to conduct income-increasing abnormal accruals. Moreover, due to the political costs, the larger-sized SEO firms will have downward earnings management behavior during fundraising compared with non-SEO. Last but not least, both CPAs and Big 4 serve as gatekeepers on earnings manipulation.
This paper further predicts that in SEO firms listed on GISA, the level of earnings management is lower when audited by the CPAs/ Big 4 accounting firms, and there is an association between the level of earnings management and outsiders' subscription rate in H2a, H2b, H2c. The empirical results also support H2a, H2b, except for H2c. Both CPAs and Big 4 accounting firms can effectively monitor the SEO firms to reduce the level of discretionary accruals. The result supports previous studies (Teoh & Wong, 1993;Worret, 2016). This paper also finds that SEO firms with higher fundraising amounts are prone to upward accrual-based earnings management. Healy and Wahlen (1999) illustrated that capital market motivations are the most significant incentives for managers to manipulate earnings. Thus, the governance should pay more attention to the earnings management behaviors of high fundraising amounts companies.
In summary, it is found that GISA firms with the incentives of raising funds may have a counteraction of upward post-SEO performance by accrual-based earnings management. Moreover, large-sized firms are prone to decrease discretionary accruals due to political costs. Also, firms with higher fundraising amounts and paid-in capital more than NT$ 30 million tend to have income-increasing abnormal accruals. Finally, consistent with prior literature, CPAs and Big 4 accounting firms play an important role in alleviating earnings management.

CONCLUSION
In 2014, TPEx set up a new stage, GISA, for SMEs to raise funds from the capital market, enlarge business scale and publicity, and even list on the stock market in the future. This paper examines the earnings management activities around seasoned equity offerings of GISA firms and verifies the impact of CPAs, Big 4, and outsiders on earnings management. Firms listed on the GISA not only find channels of fundraising but also look forward to listing on the mainboard and getting resources from the capital market in the future. Therefore, most of the GISA firms with SEO purposes may conduct upward earnings management, especially for accrual-based. The accruals reverse the same as previous research once the company gets sufficient funds from the capital market. Further, this paper reaches the consensus as to prior literature on earnings management incentives. To decrease political costs, large-sized companies are prone to decrease discretionary accruals. Based on capital market incentives, higher fundraising amounts will lead to an increase in accrual-based earnings management. Also, firms with paid-in capital more than NT$ 30 million tend to have income-increasing abnormal accruals. Finally, this paper shares similar empirical results on the role of external corporate governors in mitigating earnings management and maintaining audit quality. For GISA firms in Taiwan, CPAs and Big 4 accounting firms serve as gatekeepers in alleviating earnings manipulation behaviors. As for SEO purposes, the higher the funding amounts, the higher the possibility that management may intervene in the earnings process. CPAs and Big 4 accounting firms still stand firm and remain at their posts.

DA
The difference of total discretionary accruals and non-discretionary accruals for company i in year t