State ownership and abnormal accruals in highly-valued firms: Evidence from China

Abstract

We examine how state ownership affects Chinese firms’ abnormal accruals during a period of high valuation. We find the magnitude of abnormal accruals first increases for up to three years of high valuation, and then reduces after the fourth year. We also find that managers turn to using abnormal real transactions after four consecutive years of high valuation. Next, we examine whether the degree of abnormal accruals in highly-valued firms differs between state-owned enterprises (SOEs) and non-NSOEs. Supporting the view that SOE managers have less incentive to sustain high stock prices, we find evidence that highly-valued SOEs have significantly lower levels of abnormal accruals than highly-valued NSOEs during the period of high valuation. Our findings contribute to the literature on the cross-sectional variation in the relation between managers’ pressure to sustain high stock prices and their accounting choices in firms with different ownership structures.

Introduction

The quality of reported earnings in financial statements is of considerable interest to practitioners, regulators, and academics. Jensen (2005) suggests that highly-valued equities induce managers to use abnormal accruals with the purpose of sustaining upward trends in earnings and stock price.¹ Using U.S. data, Badertscher (2011) finds that a firm’s level of abnormal accruals changes with the duration of high valuation. Specifically, in order to sustain their high stock price, managers tend to have high abnormal accruals in the early stages before moving to abnormal real transactions (often referred to as real earnings management). We examine how state ownership affects Chinese firms’ abnormal accruals during the period of high valuation. Our study on abnormal accruals in the Chinese market is informative to both investors and regulators, as Clinch and Wei (2011) find that that Chinese investors’ reaction to earnings is four times more than those in the U.S. We confirm Badertscher’s (2011) findings in the Chinese market and, more importantly, we test whether this relation differs in state-owned enterprises (SOEs) (where the government is the ultimate controlling shareholder) and non-state-owned enterprises (NSOEs) (where private entities or individuals are the controlling shareholder) in China. Agency costs are associated with abnormal accruals, and ownership structure is a primary determinant of agency costs; therefore, we link a company’s ownership structure with their abnormal accruals.

Aminadav and Papaioannou (2020) show that state ownership is prevalent and influencial around the globe. In many countries (e.g., Brazil, China, India, and Russia), the government holds controlling stakes in the largest and most important frms. SOEs and NSOEs in China differ in the nature of their ownership, agency problems, and governance mechanisms, which leads to differences in incentives and ability to manage earnings. The uniqueness of the Chinese setting allows us to compare the effects of high valuation on abnormal accruals across two groups of firms with different ownership structures. We propose two competing predictions on how state ownership affects the relation between sustained high valuation and abnormal accruals. On the one hand, managers of Chinese SOEs have double identities as both businessman and government agents. Their promotions are more based on fulfilling the various political and social objectives than by achieving superior operating and financial performance (Cao et al., 2019, Chen et al., 2018). Moreover, managers in Chinese SOEs usually hold fewer stocks and options than NSOE managers do. Thus, they have less monetary incentives to sustain high stock prices compared to NSOE managers. Finally, SOEs also have less incentive to improve financial performance in order to gain access to external finance because of the close links between SOEs and state-owned banks in China.

On the other hand, existing empirical evidence shows that monitoring over managers is weaker in SOEs in China, including appointing less-professional directors (Fan et al., 2007), hiring lower quality auditors (Wang et al., 2008), weaker monitoring from state-owned banks (Chen et al., 2010), and weaker monitoring from reputable underwriters during IPOs (Chen et al., 2013). In total, prior research suggests managers of SOEs in China have fewer incentives to manage earnings, while weaker monitoring may induce more ability for managers to manage earnings in SOEs than in NSOEs. Thus, it is an empirical question as to whether the abnormal accruals and abnormal real transactions of highly-valued SOEs behave differently from NSOEs.

Our analyses are based on a sample of 19,226 firm-year observations with sufficient data from the China Securities Markets and Accounting Research (CSMAR) database over the period from 2003 to 2017. We use the performance-adjusted discretionary accruals estimated from the modified Jones model to measure abnormal accruals (Kothari et al., 2005), and use the prior year’s P/E ratio or market-adjusted stock returns to identify highly-valued firms. We find that the level of income-increasing accruals is higher for highly-valued firms. Further, consistent with the U.S. findings in Badertscher (2011), we find the magnitude of abnormal accruals first increase for up to three years of high valuation, and the magnitude is reduced after the fourth year. This finding is consistent with the conjecture that the difficulty of consistently having positive abnormal accruals increases over time because of the reversing nature of accruals. We find that managers turn to engaging in abnormal real transactions after four consecutive years of high valuation. These results confirm the finding of Badertscher (2011) that managers’ choice of alternative abnormal accruals mechanisms depends on the duration of a firm’s high valuation.

These findings are related to the stream of literature on the effect of firm valuation on abnormal accruals. Efendi et al. (2007) find that firms that restate earnings exhibit signs of being overvalued in the years prior to the earnings restatement and engage in non-GAAP abnormal accruals. Chi and Gupta (2009) find that overvaluation intensifies accruals management. Houmes and Skantz (2010) document that highly-valued firms have significantly higher discretionary accruals. A recent paper by Yang and Abeysekera (2019) replicates Badertscher’s (2011) results in the Australian setting. However, it is unclear whether these results will hold for Chinese firms where it is common for public firms to be ultimately controlled by the government, i.e. they are SOEs. Managers in these firms, which are characterized by different capital structures and political embeddedness, may have different reporting behaviors. For example, managers of Chinese firms, especially those of SOEs with high political ranks, are more motivated to exercise caution (Chen et al., 2018). Thus, it is ex ante unclear if the relation between high valuation and abnormal accrual holds in countries other than the U.S. and Australia. We contribute to the literature by showing that, on average, Chinese firm’s abnormal accruals follow a similar pattern as the developed countries during periods of high valuation.

Next, we examine whether the degree of abnormal accruals in highly-valued firms differs between SOEs and NSOEs. Supporting the view that SOE managers have less incentive to sustain high stock prices, we find evidence that highly-valued SOEs present significantly lower levels of abnormal accruals than highly-valued NSOEs. Our findings contribute to the literature on the cross-sectional variation in the relation between managers’ pressure to sustain high stock prices and their abnormal accruals behavior in China, an environment where this relation is likely to vary substantially between SOEs and NSOEs. Studies using U.S. data largely assume that the effects of high valuation on abnormal accruals are uniform across firms (Badertscher, 2011, Chi and Gupta, 2009, Houmes and Skantz, 2010). Our findings suggest that state ownership affects this relation, i.e., managers of SOEs are less incentivized to have positive abnormal accruals even when their companies’ stocks are highly valued.

Our results contribute to our understanding of the impact of ownership structure on managers’ reporting behaviors. The importance of examining whether the behavior of managers in SOEs differs from the behavior of managers in NSOEs is elevated by the prevalence of government control of listed companies in many parts of the world, particularly in transitional economies. Choosing a Chinese setting is suitable for our research question due to the country’s prevalence of state ownership for both historical and political reasons (Huang et al., 2017). While it has long been established that state ownership is associated with inefficiency and underperformance, there is also evidence that ultimate ownership by government helps overcome problems such as underfinancing and underinvestment (Che and Qian, 1998) and the incidence of financial frauds (Chen et al., 2006). Given China’s rapid economic growth during the past few decades, economists resort to different theories to explain its success, such as federalism (e.g., Qian and Roland, 1998). Broadly speaking, our findings add firm-level evidence associated with the improving of firms’ earnings quality behind China’s economic growth.

Second, our findings of differential effects of high valuation on abnormal accruals between SOEs and NSOEs enhance our understanding of firms’ political background in general. Political connections are common in companies listed on the world’s stock markets (Faccio et al., 2006) and there is a growing interest in the consequence of firm political background from accounting and finance researchers. Prior studies suggest that political connections enhance firm value (e.g., Fisman, 2001, Johnson and Mitton, 2003). Our findings add to the role of political background in financial reporting quality, which ultimately adds to shareholder value. In this sense, our results also have implications for investors and financial analysts in transitional economies by explaining firms’ accounting choices. It also has policy implications for regulators regarding the heterogeneous reporting behaviors of managers with and without political background.

The remainder of our paper is organized as follows. Section 2 reviews related literature and develops our hypotheses. Section 3 describes our research design. Section 4 describes our sample selection procedures and presents the results of our main tests. Section 5 reports the results of our additional tests on real earnings management. Section 6 reports the results of our robustness tests and section 7 concludes our paper.