![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
Exploiting the convergence of tax-based accounting standards to the judgement-based International Financial Reporting Standards (IFRS) as an information shock, this study examines whether the decrease in book-tax conformity improves earnings informativeness in China from the book-tax tradeoff perspective. Using A-share firms which experience this drastic regulatory change as the treatment firms and B-share firms which are not subject to such change as the benchmark firms, we find a significant decrease in the earnings informativeness for A-share firms but not B-share firms, and that the decrease is most pronounced for firms with strong financial reporting incentives. Additional analyses reveal that the decreases in earnings informativeness are due to financial reporting changes rather than changes in economic conditions. Our results shed light on the importance of considering underlying institutional factors in assessing the impact of changes in financial reporting on earnings quality.
Acknowledgements
We are grateful to the editors, associate editor (Martin Jacob), and two anonymous referees for their very constructive comments and suggestions. K. Hung Chan acknowledges partial financial support for this work by the Research Grant Council (RGC) of Hong Kong (project number UGC/FDS11/B03/18), Kenny Z. Lin acknowledges partial financial support from Lingnan University (grant number DR15A2) and Pauline W. Wong acknowledges partial financial support from Hang Seng University of Hong Kong (grant number URC-CRF-1516-06). All errors remain our own.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 All listed companies across the European Union, including Germany, France and Portugal (traditionally tax-dominated countries [Nobes Citation2011]), adopted IFRS during the fiscal year 2005. In addition to China, many transition economies, including Argentina, Brazil, Mexico, Russia, Vietnam, Romania and Poland, departed from tax-based accounting to converge with IFRS (Goncharov and Zimmermann Citation2006, Dobija and Klimczak Citation2010, Filip and Raffournier Citation2010, Phuong and Richard Citation2011).
2 We use roughly the same sample period as Chan et al. (Citation2010) for the consistent classification of different book-tax conformity periods. We do not extend the analysis beyond 2007 because of other events (e.g. the 2008 financial crisis and 2008 income tax law reform in China) and unobservable influences that could confound our results. Therefore, we adopt an approach that balances the tradeoff between using a short period reducing the effect of uncontrolled confounders and using a long time horizon providing a rich set of information for the examination.
3 Changes in accounting standards may confound the effects of book-tax conformity on earnings informativeness. However, prior studies in China found that moving towards international accounting standards generally improves accounting quality (e.g. Liu et al. Citation2011). Therefore, changes in accounting standards tend to be biased against finding results supporting our hypotheses. Thus, our results are conservative.
4 In addition to China, Hong Kong, Singapore and a growing number of countries, particularly in Europe (e.g. Croatia, Estonia, Lithuania, Russia and the Slovak Republic), adopted a flat tax system. Increasing calls for a flat tax system in the United States and United Kingdom are also observed (Bartlett Citation2005, Leigh Citation2011, Rabushka Citation2003). Although the US federal income tax is currently progressive, seven states, including Massachusetts and Utah, adopted a flat state income tax (Leigh Citation2011).
5 Under the Corporate Income Tax Law, tax losses can be carried forward for five years. Therefore, not much incentive exists to smooth losses for tax purposes.
6 For example, the majority of the firm years in Blaylock et al. (Citation2015) and Tang (Citation2015) is from G7 and other Western European countries and Australia.
7 For example, the CSRC investigated and sanctioned only 59 firm years between 2001 and 2007, with an average of 7.43 per year (Chen et al. Citation2010).
8 For example, the Supreme Court began accepting civil lawsuits against financial misstatements in the securities market only in January 2002. However, private investor plaintiffs remain uncommon, because Chinese law shoulders them with difficult burdens of proof.
9 In contrast to Ahmed et al. (Citation2013), we use a gradual convergence of IFRS as a specific source of exogenous variation in book-tax conformity to examine whether earnings quality declines when earnings management is less constrained by tax considerations.
10 However, the one-book system differs in China, where financial reporting conforms to tax regulations (during the pre-IFRS period), with the income tax amount being the same for financial and tax reporting purposes. In such a system, tax authorities have full command of financial reporting and can better prevent the understatement of earnings than the European one-book system.
11 For example, Yinguangxia (the Chinese equivalent of Enron) boasted RMB 745 million in profits by counterfeiting selling contracts and reporting nonexistent plants. Other large-scale scandals involved Zhenbaiwen (stock code 600898), Monkeyking (stock code 000535) and Yorkpoint (stock code 000008). These scandals also involved beyond-GAAP earnings inflation.
12 In addition, prior studies suggested that firms take advantage of the US progressive tax regime to smooth earnings, especially during high-conformity periods (Blaylock et al. Citation2015, Hanlon et al. Citation2008). To the extent that accounting information is distorted by income smoothing, investors place less weight on reported income, thereby reducing earnings informativeness for investment decisions during a high-conformity period (Tucker and Zarowin Citation2006). However, in China’s flat tax system, firms should have little incentive to engage in tax-motivated income smoothing.
13 Although some firms not applying for rights offerings may also manage earnings upward for purposes other than rights offerings (e.g. stock options, debt covenants and compensation schemes), other incentives to manipulate earnings found in mature markets are basically nonexistent in China (Liu and Lu Citation2007). For example, most firms in China feel little pressure in terms of debt covenant constraints, and incentive-based compensation schemes and stock options were rare during our sample period. Therefore, we focus on factors that strongly motivate earnings management behaviour in the Chinese setting.
14 We also considered testing the association between current earnings and future cash flows. However, we lack cash flow data for 1995–1997, because Chinese listed firms were not required to prepare cash flow (but fund flow) statements before 1998. Future research can use other earnings informativeness and earnings management measures to validate our results.
15 Atwood et al. (Citation2010) measure the book-tax conformity level by calculating the root mean squared error (RMSE) from a regression of current tax expenses on pretax book income, where a high (low) RMSE indicates low (high) book-tax conformity. We estimate the RMSE measures for our sample in each of the three periods to validate the decrease in the level of conformity in China. The results show that the RMSE level increased from 0.0076 in 1995–1997 to 0.0089 in 1998–2000 then to 0.0100 in 2001–2003, with the overall F-test rejecting the null hypothesis of equal means at the 5% level (F = 4.49, p = 0.011).
16 With book-tax conformity being influenced by accounting standards and tax laws simultaneously, isolating their individual effects on the properties of accounting information is difficult. Therefore, examining the relevance of accounting information in a setting with basically constant tax laws across firms is necessary. Our sample period satisfies the above requirement and covers relevant historical events in China’s corporate financial reporting. Evidence from this study has current implications for future practice and reforms in financial reporting.
17 In China, publicly listed firms experiencing financial distress are required by the CSRC to use the prefix ‘ST’ in their trading stock code. The ST system was initiated to detect poorly performing firms and release an early warning signal to the firm and investors.
18 In a sensitivity test, we also run a pooled regression by including an indicator variable (A-Share) and interaction terms with the time period variables in the model.
19 We also run a separate by-period regression for each of the three conformity periods to reconfirm the pooled regression results based on Model (1).
20 To measure the market development level, we use the National Economic Research Institute (NERI) index of marketisation of China’s provinces, with the first edition in 2001 and updates in 2003, 2004, 2007 and 2010 (Fan et al. Citation2010). The market development level in eastern and coastal areas is much more advanced than that in the central and western regions (Fan et al. Citation2010).
21 Instead of adding 1% to the prescribed ROE levels, we add 0.5% and 1.5% at a time to the prescribed ROE levels. The results are qualitatively similar to the main results. Details are discussed in the sensitivity tests.
22 The returns for both A-share and non-A-share markets appeared to be consistent with the Shanghai Stock Exchange Composite Index that peaked near 2,300 in August 2000 and fell to the 1,500 index level during 2001–2003. The negative (median) values of stock returns for both markets may indicate the generally unfavourable market conditions during 1995–1997.
23 As we have only three ERCs (i.e. for 1995, 1996 and 1997) before the reform, establishing a reliable trend for the A-share and non-A-share firms is not feasible. Nevertheless, we run separate models for the two groups of firms and plot the ΔE coefficients of each group. In an undisplayed figure, we note that the prereform trends of the treatment and control firms are similar.
24 In a similar vein, Tang and Firth (Citation2011, Citation2012) decompose book-tax differences into normal and abnormal components. However, the authors did not further decompose abnormal differences into book and tax effects.
25 The total accruals of a firm consist of DAs and nondiscretionary accruals (NDAs). Following Jones (Citation1991), we estimate the NDAs by regressing total accruals on the change in sales and fixed asset balance (all the variables are scaled by lagged total assets to ensure comparability across firms). The prediction error, or difference between the total accrual (defined as the change in the working capital accounts less depreciation) and NDA estimate, can be interpreted as the part of the accruals being ‘managed’.
26 Our controls are based on prior studies demonstrating that firm characteristics and financial variables are related to DAs. Specifically, we include firm size, debt ratio, firm ownership, listing time, profit margin, auditor size and the region and industry where the firm is operating.
27 The use of panel data will reduce the number of observations for the non-A-share firms by 39%, and the exclusion of loss firms will reduce the number of observations for such firms by approximately 12%. Given that we do not observe a clear trend for the ERCs of the non-A-share firms, our sensitivity tests focus only on A-share firms in all the panels in this table.