The Informativeness of Non-GAAP Earnings: Empirical Evidence on China

Jinyu Wang¹, Georgios Georgakopoulos², Kanellos Toudas³, Paraskevi Boufounou⁴

¹Business School, University of Amsterdam, Amsterdam, Netherlands.
²Department of Agricultural Economics and Development, Agricultural University of Athens, Athens, Greece.
³Department of Agribusiness and Supply Chain Management, Agricultural University of Athens, Athens, Greece.
⁴Department of Economics, National and Kapodistrian University of Athens, Athens, Greece.

DOI: 10.4236/tel.[phone]                

Abstract

In 2019, China piloted the Science and Technology Innovation Board (STIB) to test the practicability of a registration-based system with information disclosure requirements different from those of the approval-based system widely used. Using empirical data, the relationship between the informativeness (measured by earnings persistence and value relevance) of non-GAAP performance measures (core earnings and non-recurring gains and losses), of companies listed in the STIB was explored. The main findings are both measures capture earnings persistence, and core earnings are value-relevant; the sign of the performance measure influences informativeness; the informativeness of non-recurring profits and losses improves after setting the board.

Keywords

Corporate Governance, Informativeness, Non-GAAP Performance Measures, Earnings Persistence, Value Relevance

1. Introduction

This paper aims to provide empirical evidence to investigate the informativeness of non-GAAP performance earnings and non-recurring gains and losses of the companies listed in the Science and Technology Innovation Board.

The informativeness of non-GAAP performance measures is an important research field since it has a significant influence on managerial disclosure decisions, therefore influencing the firm’s reputation, investors’ decisions, and the regulator’s supervision. Prior papers usually split non-GAAP earnings into core earnings and special items (Shiah-Hou & Teng, 2016; Burdekin & Harrison, 2021) and measure the informativeness of performance measures through two dimensions: earnings persistence and value relevance (Curtis et al., 2014; Henry et al., 2020) . Jennings & Marques (2011) examined the joint effects of corporate governance and regulation by the Securities and Exchange Commission (SEC) on the disclosure of manager-adjusted non-GAAP earnings in US and showed that prior to regulation, investors were misled by disclosures of non-GAAP earnings made by firms with weaker corporate governance, while there was no such evidence after the SEC intervention. Wang (2018) summarized the development of non-GAAP performance measures in the United States by combing the regulatory history of relevant departments based on the current regulatory framework in the United States, discovering that the non-GAAP earnings and the corresponding non-recurring capture the characteristic of informativeness in A-share companies. Non-GAAP earnings are more informative than non-recurring gains and losses (Wang, 2017) . Zhang et al. (2022) showed that the likelihood and frequency of non-GAAP earnings reporting are lower following a going-concern audit opinion (GCO). Young (2014) reviewed the academic and professional debate surrounding non-GAAP earnings reporting by management, arguing that the demand for customized performance reporting is a natural response to constraints imposed by a one-size-fits-all reporting system and that the non-GAAP phenomenon is part of a long-standing debate over the definition and presentation of periodic performance and his work was further discussed by Miller (2014) . Venter et al. (2014) measure the value relevance of headline earnings through the coefficient between stock price and headline earnings, stating that headline earnings are more value-relevant than GAAP earnings. Ramakrishnan & Thomas (1998) showed that the price-earnings link is described better by separating components of unexpected earnings and that systematic links develop among current earnings components, future earnings, and stock prices. Richardson et al. (2005) showed that less reliable accruals lead to lower earnings persistence (and that investors do not fully anticipate the lower earnings persistence) leading to significant security mispricing, thus suggesting that there are significant costs associated with incorporating less reliable accrual information in financial statements. McVay (2006) further studied core earnings, while Wieland et al. (2013) , using stock price and return data showed that core earnings are more value relevant than GAAP earnings. Cheng (2016) discussed IFRS non-GAAP earnings disclosures and fair value measurement.

This paper starts with the information disclosure system and unique characteristics of the Science and Technology Innovation Board, further explores the disclosure of non-GAAP performance measures under the Science and Technology Innovation Edition, and finally provides some suggestions for the development status of China’s capital market and the actual situation of listed companies, and to provide a practical reference for the application and development of non-GAAP financial measures in China’s A-share market in the future.

The CSMAR database is used to gather all the available figures of the companies listed on the Science and Technology Innovation Board. There are two models introduced in this paper. The final sample consists of 858 firm-year observations for the first regression from 2017 to 2020 and 183 firm-year observations for the second model from 2019 to 2020. Also, to test the moderating effect, the whole period is split into intervals before and after the policy.

The results of the empirical research show that both measures are persistent and core earnings are of value relevance. Second, the sign of the performance measure influences informativeness. Third, the informativeness of non-recurring profits and losses improved after the board was set, indicating the pilot policy strengthens information disclosure to some extent.

This study contributes not only to existing literature but also to economic reality. The results of the research build up the foundation for future literature. It provides a new perspective of qualitative research for the new board.

This article is organized as follows: First the main characteristics of China’s Stock Markets are briefly presented, placing special emphasis on the STAR market established in 2019 by the Science and Technology Innovation Board. Then a literature review follows, providing a critical analysis of the most representative studies on the informativeness of non-GAAP earnings and the research gaps that this study addresses. Following, the methodology used, and the main results obtained are introduced and discussed against previous studies. Finally, the main conclusions are summarized, the study’s practical and academic implications are marked and relevant suggestions for future study are offered.

2. Literature Review and Hypotheses

2.1. Background

China’s stock market is divided into Shanghai stock market and Shenzhen stock market (Zhen, 2013) . There are four boards in Shanghai Stock Exchange: Shanghai A-share (Main-Board Market), small and medium-sized board, the Growth Enterprises Market and the Science and Technology Innovation Board (SSE).

The official name of A-share is RMB common stock. It is an ordinary share issued by domestic companies for domestic institutions, organizations, or individuals (excluding investors from Taiwan, Hong Kong and Macao) to subscribe and trade in RMB. Small and medium sized enterprise board is a gathering board of small and medium-sized companies set up by Shenzhen Stock Exchange to encourage independent innovation. Companies in this board generally have the characteristics of fast growth of income, strong profitability, good stock liquidity and active trading. Different from the Main-Board Market, the Growth Enterprises Market provides financing channels and growth space for the entrepreneurial enterprises which cannot be listed on the main board market for the time being. In other words, the Growth Enterprises Market is an important supplement to the main board market and also occupies an important position in the capital market.

The above three boards all adopted the approval-based system in which the application materials for listing of enterprises are reviewed by the development and Examination Committee of CSRC. The registration-based system is put on trial by the Science and Technology Innovation Board. The Shanghai Stock Exchange is responsible for the examination of the issuance and listing, and the CSRC decides to approve the registration.

In 2019, the Science and Technology Innovation Board (STAR Market) was set up to deepen capital market reforms, optimize the capital market’s basic system and improve capital market functions. Unlike other boards, the STAR Market adopts a registration-based system.

Registration-based system is also called “declaration system”. Under this system, the government agencies do not conduct substantive audits on the securities companies to be issued and their value, but only review their public information. They explicitly refuse the unqualified securities issuance, otherwise they will automatically register and take effect after the legal audit time. Therefore, the registration-based system reflects the autonomy and freedom of the market economy.

The main assumption of the approval system is that the audit of the regulatory authorities can exclude the enterprises that do not meet the listing conditions, to ensure that the securities purchased by investors in the securities market are well qualified (Wang, 2020) .

The principles of information disclosure include 1) the principle of protecting investors’ rights and interests and 2) materiality. The principle of protecting the rights and interests of investors runs through the whole process of information disclosure (Wang, 2018) .

The STAR market adopts a combination of mandatory disclosure and voluntary disclosure. Since it is difficult to achieve the intended goal of disclosure by relying solely on a single approach of mandatory disclosure, regulators usually encourage companies to voluntarily disclose more information on their own initiative to fully reveal the company’s situation (Zhen, 2019) . Under the approval-based system, issuers should make public all the relevant materials for investors to choose and judge, as well as meet the requirements of securities issuance conditions (Chen, 2013) .

2.2. Non-GAAP Performance Measures

Non-GAAP performance measure is an adjustment item based on GAAP performance measures. Sometimes, listed companies want to provide detailed external users with the company’s operation and sustainable performance level (Baik et al., 2008) . By adjusting part of the data, they eliminate temporary, uncontrollable, unrelated, or non-cash items on the operating results. Therefore, Non-GAAP performance measures are introduced. Core earnings, EBITDA, and free cash flow are the most common proxies of non-GAAP performance measures (Wang, 2018) .

According to the theory of information asymmetry (Jensen & Meckling, 1976) , the information that each trader has in the market is different. For companies and investors, listed companies are often in a strong position because they have additional information (Wang, 2018) . Interest asymmetry may lead to agency problems, moral hazards, and adverse selection problems (Jensen & Meckling, 1992) .

2.3. The Informativeness of Financial Information

Earnings persistence refers to the likelihood that the earnings level will recur in future periods and thus can reflect the predictive value of the information (Nichols & Wahlen, 2004) . This predictive value shows the information can make a difference to the decision made by users and matters as a performance measure.

Non-recurring gains and losses consist of these expenses. Core earnings equals earnings minus non-recurring gains and losses. Thus, we focused on two main components to test the informativeness: 1) non-GAAP earnings and 2) non-recurring profits and losses. The first hypothesis is related to the earnings persistence of non-GAAP performance measures. The hypotheses tested are:

H1: The core earnings of the companies listed in the Science and Technology Innovation Board this year has positive relationship with the earning in the next year.

H2: Non-recurring profits and losses are persistent in the Science and Technology Innovation Board.

H3: The core earnings of the companies listed in the Science and Technology Innovation Board are value relevant.

H4: Non-recurring profits and losses of the companies listed Science and Technology Innovation Board have value relevance.

H5a: Positive core earnings are more persistent than negative core earnings.

H5b: Non-recurring profits are more persistent than non-recurring losses.

H6a: Positive core earnings are more value relevant than negative core earnings.

H6b: Non-recurring profits and losses are more value relevant than non-recurring losses.

H7a: The implementation of Science and Technology Innovation Board strengthens the relation between the core earnings in this year and the core earnings in the next year.

H7b: The implementation of Science and Technology Innovation Board strengthens the relation between the core earnings and the stock prices.

In summary, one of the goals of the Science and Technology Innovation Board is to preserve investors’ rights while also improving information quality. Furthermore, non-GAAP disclosure of some performance measures is once again mandatory. The information quality is expected to be influenced to some extent by the more complete, detailed, and strict disclosure rules. This study explores the Science and Technology Innovation Board of non-GAAP performance measurements of persistence and value relevance, based on earlier publications such as Wang (2017) and Shiah-Hou & Teng (2016) , to investigate the benefits and limitations of disclosure system, as well as to make recommendations for government enhancements. Therefore, the present research focuses on studying the informativeness of non-GAAP earnings of companies listed in the STAR market of China, as this is a newly established market, and no previous relevant studies exist.

Brosnan et al. (2023) reviewed the past 20 years literature on non-GAAP reporting and identified the important existing and emerging research areas concerning non-GAAP earnings disclosures. Table 1 below presents a review of the main relevant studies. Further, with respect to the topic’s practical implications Zhang et al. (2023) , showed that managers are less likely to report non-GAAP earnings with the increase of stock price informativeness, suggesting that information from stock prices discourages non-GAAP reporting from firm managers and showed that stock price informativeness influences managerial incentive of non-GAAP reporting through two channels: 1) direct monitoring and 2) the improvement in the information environment. Also, a number of studies focused on examining the Big 4 auditing influence to the quality of Non-GAAP reporting (Feng et al., 2022; Dak-Adzaklo et al., 2023) .

Table 1. Literature review.

Source: Own analysis.

3. Methodology

3.1. Sample Selection

The data comes from the CSMAR database. In this paper, the data of the database are sampled and checked with the financial reports of listed companies. The financial performance measures used in this paper are the standard data disclosed in the financial reports of listed companies. In this paper, Excel software is used to sort out, match, and calculate the data, and STATA is used to test the regression model (Table 2).

3.2. Earnings Persistence

The possibility that the earnings level would recur in future periods is referred to as earnings persistence. Following Freeman & Tse (1989) , the relation between current earnings performance and future earnings performance can be expressed as:

$Earning{s}_{t+1}={\alpha }_0+{\alpha }_{1}Earning{s}_t+{\epsilon }_t$ (1)

According to the equation earnings = cashflows + accruals, Sloan (1996) used a disaggregated model to verify his hypothesis. The regression can be replaced by:

$Earning{s}_{t+1}={\alpha }_0+{\alpha }_{1}Accrual{s}_t+{\alpha }_{2}Cashflow{s}_t+{\epsilon }_t$ (2)

Based on previous earning persistence regressions, we proposed the following regression to test the persistence of the non-GAAP earnings.

$\begin{array}{c}Earning{s}_{t+1}={\alpha }_0+{\alpha }_{1}C{E}_t+{\alpha }_{2}N{R}_t+{\alpha }_{3}D1C{E}_t+{\alpha }_4{D}_{2}N{R}_t+{\alpha }_5{D}_{2}N{R}_t\\ +{\alpha }_{6}LE{V}_t+{\alpha }_{7}OWN+{\alpha }_{8}RO{E}_t+{\alpha }_{9}COVID+{\alpha }_{10}AG{E}_t+{\epsilon }_t\end{array}$ (3)

The first empirical model is designed to test hypothesis 1, hypothesis 2, hypothesis 5a, hypothesis 5b and hypothesis 7a. The dependent variable and independent variables used are presented in Table 3 that follows.

Table 2. Sample characteristics.

Source: Own analysis.

Table 3. List of variables used to test earnings persistesnce.

Source: Own analysis.

3.3. Value Relevance

The model for testing the value relevance of earnings is first designed as followed:

$Returns=\alpha +\beta Earnings+\epsilon$ (4)

This regression model allowed us to answer the question of whether earnings are “value relevant”. Basu (1997) relied on the “reverse” regression model:

$Earnings=\alpha +\beta Returns+\epsilon$ (5)

This model helps us to answer the question of whether earnings summarize information, as reflected in stock price, in a timely manner.

Then Basu (1997) splits this reverse-regression model based on whether the sign of the returns’ variable is positive or negative.

$Earnings={\alpha }_0+{\alpha }_1+{\beta }_{0}Returns+{\beta }_{1}DReturns+\epsilon$ (6)

where D is a dummy variable when the stock returns are negative. Typically studies of this sort are based on variations of the following linear price-levels regression:

$P_t={\alpha }_0+{\alpha }_{1}BV{E}_t+{\alpha }_{2}EAR{N}_t+{\epsilon }_t$ (7)

where P is the market value of a share in the firm at time t, BVE is the book value of equity (per share) and EARN is the GAAP earnings (per share) of the firm. He constructed the following price model to study the impact of non-recurring profit and loss and net profit after deduction of non-recurring profit and loss on the stock price.

$\begin{array}{c}{P}_t={\beta }_0+{\beta }_{1}BVP{S}_t+{\beta }_{2}COEP{S}_t+{\beta }_{3}EIEP{S}_t+{\beta }_4{D}_{3}COEP{S}_t\\ +ControlVars+{\epsilon }_t\end{array}$ (8)

Based on previous models, we introduced variables that can measure their sign, proposing the following model for studying value relevance.

$\begin{array}{c}{P}_t={\beta }_0+{\beta }_{1}BVP{S}_t+{\beta }_{2}COEP{S}_t+{\beta }_{3}EIEP{S}_t+{\beta }_4{D}_{3}COEP{S}_t+{\beta }_5{D}_{4}EIEP{S}_t\\ +{\beta }_{6}LE{V}_t+{\beta }_{7}GRO{W}_t+{\beta }_{8}OWN+{\beta }_{9}COVID+{\beta }_{10}AG{E}_t+\epsilon \end{array}$ (9)

The Dependent variable and independent variables (main variables) used are presented in Table 4 that follows. The test of earnings persistence aims to support hypothesis 1, hypothesis 2, hypothesis 5a, hypothesis 5b and hypothesis 7a. The descriptive statistical results of earnings persistence model variables are presented in Table 5 below.

Table 4. List of variables used to test value relevance.

Source: Own analysis

Table 5. Descriptive statistics—earnings persistence.

Source: Own analysis.

4. Results

This paper uses OLS regression model to investigate the relationships.

4.1. Earnings Persistence

The test of earnings persistence aims to support hypothesis 1, hypothesis 2, hypothesis 5a, hypothesis 5b and hypothesis 7a. The descriptive statistical results of earnings per-sistence model variables are presented in Table 5 above.

Table 6 below, shows the correlation coefficient matrix of the earnings persistence model. The coefficient 0.777 of core earnings in year t and earnings in year t + 1 is bigger than 0.443, that of non-recurring gains and losses in year t and earnings in year t + 1. We can predict that the persistence of the core earnings is stronger than non-recurring gains and losses. Thus, it could be concluded that the implementation of STAR market is within expectation and the mandatory disclosure of non-GAPP mandatory disclosure of non-GAAP performance measures is necessary. 0.248 shows a negative correlation between the negative core earnings in year t and the earnings in year t + 1. Furthermore, the factor 0.096 can be used as predictive evidence for forecasting the relation between non-recurring losses in year t and earnings in year of t + 1. As for the control variables, there is no predictive relation between LEVERAGE and Earnings in year t + 1. OWN, ROE and AGE show significant positive relation with Earnings in year t + 1. However, the opposite pattern could be found in COVID, implying the macroeconomic influence is harmful to companies’ profitability.

Table 7 below presents the Multicollinearity tests for earning management by means of Variance Inflation Factor (VIF) values. All the values are below the threshold of 10. This evidence is consistent with no severe multicollinearity for the first model.

Table 6. Pearson correlation matrix—earnings persistence.

Source: Own analysis. ***p < 0.01, **p < 0.05, *p < 0.1.

Table 7. Multicollinearity tests—earnings persistence.

Source: Own analysis

Normality of the residuals is checked by Skewness/Kurtosis test in the residuals (to determine whether they are consistent with the normal distribution). Under the null hypothesis the residuals are normally distributed. The null hypothesis of normality is rejected when the p-value < 0.05. This indicates that the residuals of these models are normally distributed.

In order for OLS method to be efficient, the error terms of the model should be uncorrelated and should have a constant variance. When this requirement does not hold, heteroscedasticity exists. The Breusch-Pagan homoscedasticity test Chi2 for earnings persistence was estimated 206.238 (Heij et al., 2004) . For the models in all intervals, the p-value of the test is less than 0.05. Thus, the null hypothesis for homoscedasticity is rejected.

In this paper, Stata software was used to do OLS regression test on panel data. In the test of earnings persistence, the experimental group (STAR group) was divided into two-time intervals of 2017-2018 and 2019, and the control group was divided into the same intervals for comparison. The first interval is the period before China set up the STAR market while the second one is the period after the pilot policy. Table 8 below contains the results for testing H1, H2, H5a, H5b and H7a examining earnings persistence, from where there are evidence supporting H1, H2 and H5b while there is no evidence shown for H5a and H7a. The opposite conclusions emerge because of the outbreak COVID in 2019. Even though we introduced a control variable related to the pandemic, no significant result is shown in the table. Thus, further research could shed light to this issue.

Table 8. Results for hypotheses testing H1, H2, H5a, H5b and H7a—earnings persistence.

Source: Own analysis. ***p < 0.01, **p < 0.05, *p < 0.1.

Table 9 contains the results of the robustness regression for the hypotheses related to earnings persistence. The coefficients of the core earnings and non-recurring profits are significant at 1% level in all periods. The results of the OLS regression model indicate that both non-GAAP performance measures capture the characteristic of persistence, strongly supporting hypothesis 1 and hypothesis 2. Likewise, by comparing 2.104 and 0.341 (2.104 - 1.763), it is not hard to draw the conclusion that the non-recurring profits and losses are more persistent when it is positive than negative (H5b). There is no evidence for supporting hypothesis 5a since the coefficients of all the negative core earnings are not significant. Furthermore, 0.926 is smaller than 1.088 while 2.551 is bigger than 1.215. This contradictory phenomenon related to the persistence of different non-GAAP components cannot be treated as a convince of hypothesis 7a.

4.2. Value Relevance

The test of value relevance aims to support hypothesis 3, hypothesis 4, hypothesis 6a, hypothesis 6b and hypothesis 7b.

Table 10 below shows the descriptive statistical results of the earnings persistence model variables.

Table 11 below indicates that BVPS, COEPS, EIEPS and GROW positively and significantly correlate with P. The correlation coefficient between the book value of equity and the stock price P is 0.823, indicating a positive association, according to the Pearson test results. As a result, we should expect the stock price to favorably reflect the rise or decline in the company’s owner’s equity. The non-recurring profit and loss EIEPS and the net profit share price P have a correlation coefficient of 0.663, indicating a positive relationship. Therefore, it can be predicted that the non-recurring profit and loss have a low-value correlation

Table 9. Robustness test results—earnings persistence.

Source: Own analysis. ***p < 0.01, **p < 0.05, *p < 0.1.

Table 10. Descriptive statistics—value relevance.

Source: Own analysis.

Table 11. Pearson correlation matrix—value relevance.

Source: Own analysis. ***p < 0.01, **p < 0.05, *p < 0.1.

or no value correlation. The correlation coefficient between corps and stock price P is 0.691, showing a significant moderate correlation, and its correlation coefficient is slightly larger than that of non-recurring profit and loss. LEV has a negative relation with stock price at the significant level of 10%. The coefficient for LEV is −0.135. The coefficient for the variable GROW is 0.298 (***), indicating that GROW has a positive relation with stock price at the significant level of 1%.

The test for multicollinearity by means of Variance Inflation Factor (VIF) values is presented in Table 12 below. All of the values are below the threshold of 10. This evidence is consistent with no severe multicollinearity for the second model which tests value relevance.

The normality of the residuals is checked by Skewness/Kurtosis tests and the adj chi2 is 23.03 with prob 0.001. As the null hypothesis of normality is rejected when the p-value < 0.05, I it indicates that the residuals of these models are normally distributed.

The Breusch-Pagan was used as homoscedasticity test, and a chi2 value of 195.33 with prob 0.000 was estimated (p-value < 0.05). As the p-value of the test is less than 0.05, the null hypothesis for homoscedasticity is rejected.

Stata software was used to do OLS regression to test panel data. Firm and year. In the test of earnings persistence, the experimental group (STAR group) was divided into two-time intervals of 2019 and 2020, and the control group was divided into the same intervals for comparison. The first interval is the period before China set up the STAR market while the second one is the period after the pilot policy. Table 13 presents the results for testing H3, H4, H6a, H6b and H7b examining value relevance, and shows that there are evidence supporting H3, H6b and H7b while there is no statistically significant result shown for H3 and opposite results are shown to reject H7a. The coefficient of COVID in 2019 to 2020 is statistically significant at 10% level. This shows a negative impact of the COVID-19 on the stock price.

Table 14 below shows the results of the robustness regression for the value relevance hypotheses. In the OLS regression model, the coefficients (16.495, 10.406) in 2019 and total period of COEPS are statistically significant at 1%, which gives convincing support for hypothesis 3. Only the coefficient in the total period is significant at 10% level, standing for hypothesis 4. Hypothesis 6a and hypothesis 6b are rejected since there are no significant results for negative core earnings per share and non-recuring loss per share at 10% level. Similarly, there is no result supporting hypothesis 7.

4.3. Summary of Results

Table 15 below provides an overview of the results for all tested hypotheses. The main findings are: Hypothesis 1 and hypothesis 2 are supported by the regression results in this paper. Both non-GAAP earnings and the corresponding exclusions capture the characteristic of persistence. Consistent with the conclusions of Burgstahler et al. (2002) , this paper further confirmed that non-recurring gains and losses are not zero persistent. Venter et al. (2014) found that the non-GAAP earnings are of value relevance. Similarly, the non-GAAP earnings of those companies listed in the STAR market are value relevant, standing for hypothesis 3 (Li, Su, Dong, & Zhu, 2018) . There are significant coefficients for non-GAAP earnings. Hypothesis 5b is accepted in this paper. In consistency with Wang (2017) , the non-recurring gains are more persistent than non-recurring losses. We considered coronavirus may be one of the reasons for this unusual result. The economic depression caused by the epidemic makes the managers intentionally disclosure beneficial information. Hypothesis7b is

Table 12. Multicollinearity test—value relevance.

Source: Own analysis.

Table 13. Results for hypotesestesting H3, H4, H6a, H6b and H7b—value relevance.

Source: Own analysis. ***p < 0.01, **p < 0.05, *p < 0.1.

Table 14. Robustness test—value relevance.

Source: Own analysis. ***p < 0.01, **p < 0.05, *p < 0.1.

Table 15. Summary of main findings.

Source: Own analysis.

supported in the expected direction, showing that the value relevance of companies listed in the STAR market are stronger. Since there is nearly no quantitative research about the informativeness relevant to the STAR market. This finding serves as a reference for future literature. The result is in the opposite direction of hypothesis 5a. This finding is contradictory with previous literature (Basu, 1997) as they found the positive components of earnings are more persistent than negative ones. Likewise, the results for testing hypothesis 7a is statistically significant but in an opposite direction, suggesting the persistence of the non-GAAP earnings decreases after the pilot policy. We considered coronavirus might be a factor which influence the performance of capital market (Burdekin & Harrison, 2021) . There are no statistically significant results of the experimental group supporting H4, H6a and H6b. However, H6b can be proved by the results of the control group.

5. Conclusion

In 2019 the Science and Technology Innovation Board introduced optimizing the capital market system. This entirely new board is different from previous boards not only in terms of the listing conditions but also from information disclosure requirements. One of the most important purposes is to protect the investors rights tightly related to information disclosure. Investors are sometimes not competent enough or have limited time and energy to capture and understand all the information. Therefore, they will fixate the non-GAAP performance measures that better reflect a company’s economic reality than GAAP performance measures do. Therefore, this paper focuses on non-GAAP performances, especially core earnings and non-recurring profits and losses. Previous scholars always examined informativeness from two dimensions: persistence and value relevance. In that, we measure the informativeness of both core earnings and non-recurring profits and losses through these two dimensions separately. Meanwhile, A-share companies from industries are selected as the control group.

The earnings persistence is tested by examining the correlation between core earnings in year t and earnings in year t + 1. At the same time, the value relevance is measured by finding the relation between the stock price and the core earnings per share in the same year. The sensitivity of the magnitudes of the characteristics when the sign of the numbers is less than zero is also investigated.

The main findings are stated as follows. Firstly, the regression results show that the core earnings are persistent and value relevant and the non-recurring profits and losses are persistent, indicating that the non-GAAP earnings, and their exclusions, is informative in the STAR market. Secondly, the sign of the non-GAAP performance measure influences its informativeness. Thirdly, as predicted, the informativeness of non-recurring profits and losses is stronger after the pilot policy was implemented.

Therefore, this paper contributes to academic research on the topic in three dimensions. First, the regression results provide the statistical foundation for future literature. Secondly, this paper also provides a new perspective of the decomposition of GAAP performance measures. Thirdly, introducing the dummy variable remains an inspiration for future scholars to go deeper about this topic.

The main limitation of this research is as follows: First, a causal relationship between the variables was not evidenced, only associations. This implies that the results may be influenced by randomness. Secondly, the figures for testing value relevance before 2019 are hard to find, and therefore it was feasible to compare pre-STAR and post-STAR markets. To make up for this restriction, the results were compared with the control group after 2019. Future research could resolve the above issues. Moreover, future research could introduce industry variables and other additional factors as control variables in the regression model. Thirdly, the informativeness of the non-recurring profits and losses is even more significant than that of core earnings. This result is inconsistent with the control group. The company’s status might be the first reason for the unexpected result since companies in the STAR market are start-ups while the control group contains a loss of mature companies. In addition, it is expected that the coronavirus influences the development of these companies to some extent. Future research could consider these two influences.

Furthermore, as this is the first study on the informativeness of non-GAAP Earnings of companies listed on the STIB, its implications are particularly important to investors, auditors and the regulator of the STAR market. With China’s capital market development, more and more listed companies are expected to adopt non-GAAP financial measures. To maintain a good information disclosure environment and promote the healthy development of the securities market, adequate supervision and guidance is necessary.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.

References

[1]	Baik, B., Billings, B. K., & Morton, R. M. (2008). Reliability and Transparency of Non-GAAP Disclosures by Real Estate Investment Trusts (REITs). The Accounting Review, 83, 271-301. https://doi.org/10.2308/accr.2008.83.2.271
[2]	Basu, S. (1997). The Conservatism Principle and the Asymmetric Timeliness of Earnings. Journal of Accounting & Economics, 24, 3-37. https://doi.org/10.1016/S0165-4101(97)00014-1
[3]	Brosnan, M., Duncan, K., Hasso, T., & Hollindale, J. (2023). Happy 20-Year Anniversary Non-GAAP Earnings: A Systematic Review of the Literature. Journal of Accounting Literature. https://doi.org/10.1108/JAL-02-2023-0035
[4]	Burdekin, R. C., & Harrison, S. (2021). Relative Stock Market Performance during the Coronavirus Pandemic: Virus vs. Policy Effects in 80 Countries. Journal of Risk and Financial Management, 14, Article No. 177. https://doi.org/10.3390/jrfm14040177
[5]	Burgstahler, D., Jiambalvo, J., & Shevlin, T. (2002). Do Stock Prices Fully Reflect the Implications of Special Items for Future Earnings? Journal of Accounting Research, 40, 585-612. https://doi.org/10.1111/1475-679X.00063
[6]	Chen, Z. (2013). Research on the Legal System of Stock Issuance Supervision in Contemporary China. Ph.D. Thesis, Fudan University. (In Chinese)
[7]	Cheng, C. S. A. (2016). Discussion of ‘IFRS Non-GAAP Earnings Disclosures and Fair Value Measurement’. Accounting and Finance, 56, 99-112. https://doi.org/10.1111/acfi.12200
[8]	Curtis, A. B., McVay, S. E., & Whipple, B. C. (2014). The Disclosure of Non-GAAP Earnings Information in the Presence of Transitory Gains. The Accounting Review, 89, 933-958. https://doi.org/10.2308/accr-50683
[9]	Dak-Adzaklo, C. S. P., Nduakoh, D., Obiri-Yeboah, E., & Ofosu, E. (2023). Government Subsidies and Non-GAAP Reporting. SSRN. https://doi.org/10.2139/ssrn.4481851
[10]	Doyle, J. T., Lundholm, R. J., & Soliman, M. T. (2003). The Predictive Value of Expenses Excluded from Pro Forma Earnings. Review of Accounting Studies, 8, 145-174. https://doi.org/10.1023/A:1024472210359
[11]	Du, K., & Huddart, S. (2020). Economic Persistence, Earnings Informativeness, and Stock Return Regularities. Review of Accounting Studies, 25, 1263-1300. https://doi.org/10.1007/s11142-020-09531-2
[12]	Entwistle, G. M., Feltham, G. D., & Mbagwu, C. (2010). The Value Relevance of Alternative Earnings Measures: A Comparison of Pro Forma, GAAP, and I/B/E/S Earnings. Journal of Accounting, Auditing & Finance, 25, 261-288. https://doi.org/10.1177/0148558X1002500205
[13]	Feng, Z., Francis, J. R., Shan, Y., & Taylor, S. L. (2022). Do High-Quality Auditors Improve Non-GAAP Reporting? The Accounting Review, 98, 215-250. https://doi.org/10.2139/ssrn.4087878
[14]	Freeman, R. N., & Tse, S. (1989). The Multiperiod Information Content of Accounting Earnings: Confirmations and Contradictions of Previous Earnings Reports. Journal of Accounting Research, 27, 49-79. https://doi.org/10.2307/2491064
[15]	Heij, C., de Boer, P., Frances, P. H., Kloek, T., & van Dijk, H.K. (2004). Econometric Methods with Applications in Business and Economics. Oxford University Press.
[16]	Henry, T. F., Weitz, R. R., & Rosenthal, D. A. (2020). Non-GAAP Earnings Disclosure Post 2010 SEC Regulation Change. Journal of Corporate Accounting & Finance, 31, 114-134. https://doi.org/10.1002/jcaf.22435
[17]	Isidro, H., & Marques, A. (2021). Industry Competition and Non-GAAP Disclosures. Accounting and Business Research, 51, 156-184. https://doi.org/10.1080/00014788.2020.1798209
[18]	Jennings, R., & Marques, A. (2011). The Joint Effects of Corporate Governance and Regulation on the Disclosure of Manager-Adjusted Non-GAAP Earnings in the US. Journal of Business Finance & Accounting, 38, 364-394. https://doi.org/10.1111/j.1468-5957.2011.02238.x
[19]	Jensen, M. C., & Meckling, W. H. (1992). Specific and General Knowledge, and Organizational Structure. In L. Werin, & H. Wijkander (Eds.), Contract Economics (pp. 251-274). Basil Blackwell Ltd.
[20]	Jensen, M. C., & Meckling, W. H. (1976). Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, 3, 305-360. https://doi.org/10.1016/0304-405X(76)90026-X
[21]	Kolev, K., Marquardt, C. A., & McVay, S. E. (2008). SEC Scrutiny and the Evolution of Non-GAAP Reporting. The Accounting Review, 83, 157-184. https://doi.org/10.2308/accr.2008.83.1.157
[22]	Leung, E., & Veenman, D. (2018). Non-GAAP Earnings Disclosure in Loss Firms. Journal of Accounting Research, 56, 1083-1137. https://doi.org/10.1111/1475-679X.12216
[23]	Li, N., Su, H., Dong, W., & Zhu, K. (2018). The Effect of Non-Recurring Items on Analysts’ Earnings Forecasts. China Journal of Accounting Research, 11, 21-31. https://doi.org/10.1016/j.cjar.2017.10.001
[24]	Marques, A. (2006). SEC Interventions and the Frequency and Usefulness of Non-GAAP Financial Measures. Review of Accounting Studies, 11, 549-574. https://doi.org/10.1007/s11142-006-9016-x
[25]	Marques, A. (2010). Disclosure Strategies among S&P 500 Firms: Evidence on the Disclosure of Non-GAAP Financial Measures and Financial Statements in Earnings Press Releases. The British Accounting Review, 42, 119-131. https://doi.org/10.1016/j.bar.2010.02.004
[26]	McVay, S. E. (2006). Earnings Management Using Classification Shifting: An Examination of Core Earnings and Special Items. The Accounting Review, 81, 501-531. https://doi.org/10.2308/accr.2006.81.3.501
[27]	Miller, G. (2014). Discussion of “The Drivers, Consequences and Policy Implications of Non-GAAP Earnings Reporting” by Steven Young (2014). Accounting and Business Research, 44, 466-468. https://doi.org/10.1080/00014788.2014.900953
[28]	Nichols, D. C., & Wahlen, J. M. (2004). How Do Earnings Numbers Relate to Stock Returns? A Review of Classic Accounting Research with Updated Evidence. Accounting Horizons, 18, 263-286. https://doi.org/10.2308/acch.2004.18.4.263
[29]	Ramakrishnan, R. T. S., & Thomas, J. K. (1998). Valuation of Permanent, Transitory, and Price-Irrelevant Components of Reported Earnings. Journal of Accounting, Auditing & Finance, 13, 301-336. https://doi.org/10.1177/0148558X9801300309
[30]	Richardson, S. A., Sloan, R. G., Soliman, M. T., & Tuna, İ. (2005). Accrual Reliability, Earnings Persistence and Stock Prices. Journal of Accounting & Economics, 39, 437-485. https://doi.org/10.1016/j.jacceco.2005.04.005
[31]	Shiah-Hou, S. R., & Teng, Y. Y. (2016). The Informativeness of Non-GAAP Earnings after Regulation. Finance Research Letters, 18, 184-192. https://doi.org/10.1016/j.frl.2016.04.015
[32]	Sloan, R. G. (1996). Do Stock Prices Fully Reflect Information in Accruals and Cash Flows about Future Earnings? The Accounting Review, 71, 289-315. https://www.jstor.org/stable/248290
[33]	Venter, E. R., Emanuel, D., & Cahan, S. F. (2014). The Value Relevance of Mandatory Non-GAAP earnings. Abacus, 50, 1-24. https://doi.org/10.1111/abac.12020
[34]	Wang, K. (2020). Comparison between Approval System and Registration System and Policy Suggestions on Transition to Registration System in China. Business Information, 50, 291. (In Chinese)
[35]	Wang, R. (2018). Research on the Application of Non-GAAP Performance Measures in China. Ph.D. Thesis, Xiamen University. (In Chinese)
[36]	Wang, W. (2017). The Earnings Persistence, Value Relevance and Executive Compensation Effects of GAAP Earnings and Non-GAAP Earnings. MSc. Thesis, Chongqing University. (In Chinese)
[37]	Wieland, M. M., Dawkins, M. C., & Dugan, M. T. (2013). The Differential Value Relevance of S&P’s Core Earnings Versus GAAP Earnings: The Role of Stock Option Expense. Journal of Business Finance & Accounting, 40, 55-81. https://doi.org/10.1111/jbfa.12013
[38]	Young, S. (2014). The Drivers, Consequences and Policy Implications of Non-GAAP Earnings Reporting. Accounting and Business Research, 44, 444-465. https://doi.org/10.1080/00014788.2014.900952
[39]	Zhang, T., Kwack, S. Y., Si, Y., & Tian, G. (2022). Non-GAAP Earnings Reporting Following Going-Concern Opinions. Accounting & Finance. https://doi.org/10.1111/acfi.13028
[40]	Zhang, T., Tian, G., & Feng, H. (2023). Role of Stock Price Informativeness in Shaping Non-GAAP Earnings Disclosures. Asia-Pacific Journal of Accounting & Economics, 30, 266-283. https://doi.org/10.1080/16081625.2021.2020665
[41]	Zhen, A. (2019). Information Disclosure Is the Key to the Success of the Science and Technology Innovation Board. Financial Weekly, 8, 24. (In Chinese)
[42]	Zhen, Y. (2013). China’s Capital Markets. Chandos Publishing. https://doi.org/10.1533/9781780633602