Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Earnings management, business cycle, and product market competition

Earnings management, business cycle, and product market competition China Journal of Accounting Studies, 2015 Vol. 3, No. 2, 136–157, http://dx.doi.org/10.1080/21697213.2015.1023694 Earnings management, business cycle, and product market competition a b b Hongjian Wang *, Qingyuan Li and Yana Chen a b Department of Accounting, Jinan University, Guangzhou, China; Economics and Management School, Wuhan University, Wuhan, China Using annual data from non-financial listed companies in China between 2001 and 2011, this paper empirically examines the degree and key mechanism of upwards earnings management by companies under cyclical fluctuations of the macro- economy. We find that: firstly, upwards earnings management is significantly and positively correlated with the growth rate of the macro-economy, indicating a rela- tively strong pro-cyclical effect; secondly, the pro-cyclical effect of upwards earn- ings management will be more significant if the management expects corporate profit to be lower than the average profit for the industry; thirdly, the pro-cyclical effect of upwards earnings management will be more significant if the company faces greater pressure from product market competition. Further tests reveal that the greater the incentive for management, the greater the financing restrictions, and the more sensitive the industry is to the economic cyclical fluctuations, then the more significant is the pro-cyclical effect of upwards earnings management. The results demonstrate that the key mechanism of upwards earnings management during the economic boom is the expected profit based on the difference between the previous year’s expected firm’s profit and that year’s average profit for the industry, as well as the pressure of product market competition. Thus, the combination of the macro- economic state and the degree of product market competition will play an important role in understanding earnings management behaviour. Keywords: business cycle; earnings management; pro-cyclical effect; product mar- ket competition 1. Introduction According to the survey by the China Europe International Business School of 1,214 executives of Chinese and foreign companies in 2013, 46% of local companies and 37% of foreign companies refer to macro-economic adjustments as a major considera- tion (Fan, Xu, & Zhou, 2013). In particular, the outbreak and rapid spread of the global financial crisis in 2008 resulted in violent fluctuations in the Chinese economy, causing the economic growth rate to fall from 14.16% (adjusted) in 2007 to 9.63% (adjusted) in 2008. Many companies thus underwent operating difficulties. During macro- economic cyclical fluctuations, the researchers have consistently focused on macro- economic fundamental indices, such as growth rate of gross domestic product (GDP), unemployment rate, and consumer price index (CPI). However, theoretical economists *Corresponding author. Email: whj.hust@163.com Paper accepted by Donghua Chen. © 2015 Accounting Society of China China Journal of Accounting Studies 137 have largely overlooked how macro-economic cyclical changes affect changes in micro-corporate behaviour. Figure 1 describes the relationship between the growth rate of China’s GDP and the average profit margin of listed companies from 2000 to 2012. As shown in Figure 1, the average profit margin of listed companies is relatively consistent with the GDP growth rate, especially after 2005, when the reform of non-tradable shares was further implemented. The consistency between profit margins and macro-economic cyclical fluctuations demonstrates that macro-economic cyclical fluctuations have a relatively important impact on micro-enterprise behaviour. As a core topic in financial accounting, earnings management has elicited consider- able attention from scholars, and extensive research has been conducted on this topic. A conclusion that can be made is that research concentrates on the motivation of earn- ings management (Barua, Lin, & Sbaraglia, 2010; Cohen & Zarowin, 2010), the meth- ods and governance of earnings management (Bo & Wu, 2009; Klein, 2002; Xie, Davidson, & Dasalt, 2003), and the economic consequences of earnings management (Chen & Yuan, 2004; Coles, Hertzel, & Kalpathy, 2006; Haw, Qi, Wu, & Wu, 2005; Xu & Chen, 2009). Most of these studies are based on a company-level perspective and ignore the possible effects that the macro-economic state might have on companies’ earnings management. Existing research on business cycles is primarily based on the macro- economic perspective, with some studies concentrating on its effects on micro-corporate behaviour. Previous studies have covered topics including business cycle and capital structure (Erel, Julio, Kim, & Weisbach, 2012; Korajczyk & Levy, 2003); business cycle and corporate investment (Li & Luo, 2010; McLean & Zhao, 2014; Ying, 2008; Yuan, Zhou, Liu, & Liu, 2012); and business cycle and return on capital (Klein & Marquardt, 2006; Li, 2009). However, few scholars focus on how the business cycle affects earnings management of companies. As an important strategic decision made by the management of a company, earnings management is conducted under certain speci- fic macro-economic circumstances. Therefore, studying earnings management based only on the company-level perspective will lead to limitations on the general applicabil- ity of the research results. Figure 1. The relationship between GDP growth rate and the average margins of listed compa- nies in China from 2000 to 2012. Notes: The data of GDP is from the National Bureau of Statistics and the average rate of profit from the annual reports of non-financial listed companies in the sample. 138 Wang et al. To date, scholars such as Jin (2005), Cohen and Zarowin (2007), Strobl (2013), and Chen (2013b) have started to pay attention in their research to the effect of the macro- economy on earnings management. Nevertheless, their studies take developed market economic countries as their context. Using the annual data from non-financial listed companies in China between 2001 and 2011, this paper examines empirically the degree, and key mechanism of companies’ earnings management under cyclical fluctua- tions of the macro-economy. We report the following findings. First, upwards earnings management is significantly and positively correlated with the growth rate of the macro-economy, indicating a relatively strong pro-cyclical effect; Second, the pro- cyclical effect of upwards earnings management will be more significant if the manage- ment expects corporate profit to be lower than the average profit for the industry; Finally, the pro-cyclical effect of upwards earnings management will be more signifi- cant if the company faces greater product market competition pressure. Further tests reveal that financing restrictions have no significant impact on the pro-cyclical effect. The results demonstrate that the key mechanism of upwards earnings management dur- ing the economic boom is the expected profit (based on the difference between the previous year’s profit of the company and the same year’s average industrial profit), as well as the pressure from product market competition. Thus, the combination of the macro-economic state and the degree of product market competition will be important in understanding earnings management behaviour. The major contributions of this paper are as follows. First, based on the studies of Jin (2005), Cohen and Zarowin (2007), Strobl (2013), and Chen (2013b), this paper uses annual data from non-financial listed companies in China as a sample, examining the different effects of cyclical fluctuations on the degree of firms’ upwards earnings management. Second, combining the nature of property rights with Chinese characteris- tics and the institutional background of industry regulations, this paper reveals the pres- sure from product market competition to be a key transmission mechanism of companies’ earnings management under the flourishing period of the economy. This finding provides evidence for users of accounting information to judge the changes in the quality of companies’ financial statements with macro-economic cyclical fluctua- tions. Third, this paper also deepens understanding of the potential effects of interac- tions between the macro-economy and micro-corporate behaviour on research in accounting and finance; and enriches the literature on the effects of macro-economic policies on micro-corporate financial and accounting behaviour. 2. Theoretical analysis and testable hypotheses In recent years, the relationship between the macro-economy and micro-corporate beha- viour has elicited increasing interest from researchers. Specifically, the global financial crisis stemming from the subprime crisis of the United States in 2008 spread rapidly to all economies of the world, triggering scholarly consideration of the mechanism by which the macro-economy could affect micro-corporate behaviour. Studies by Jin (2005), Cohen and Zarowin (2007), and Strobl (2013) were based on the background of Western countries, whereas Chen’s research (2013) explored the degree of compa- nies’ earnings management in different phases of the business cycle from the perspec- tive of industrial cyclicality. Thus, their studies do not to consider the particular aspects of the Chinese economy and how the economic cycle affects the motivation for earn- ings management. China Journal of Accounting Studies 139 Given that China is in a transitional period, earnings management behaviour in Chinese listed companies is more severe because companies have weak property protec- tion systems, undesirable market mechanisms, weak auditor independence and serious relationship-based transactions, and the role of political connections remains important (Piotroski & Wong, 2012). Therefore, exploring how the business cycle affects the mechanism of earnings management will be greatly significant to our understanding of the effect of macro-economic cyclical fluctuations on micro-corporate behaviour. 2.1. Business cycle and companies’ earnings management According to Prospect Theory, people’s levels of sensitivity to ‘loss’ and ‘gain’ are asymmetrical, i.e., the pain from loss is much greater than the happiness from gain, with the decision-maker’s ‘loss aversion’ behaviour as a concrete manifestation (Kahneman & Tversky, 1979). Extending to the capital market, if companies announce bad earnings news when the market trend provides positive expectations to investors, market investors will be surprised and may in turn respond intensely. Thus, the stock price of companies will suffer severely in the stock market (Conrad, Cornell, Landsman, 2002) because the economy is in an upswing, and investors tend to expect positive future trends. Under this circumstance, investors will not respond as strongly if companies announce good earnings news, but will respond dramatically if companies publish bad earnings news (Conrad et al., 2002). From the perspective of executive compensation contracts, executive compensation is not only influenced by the operating performance of the year, but also affected by the stock price. Thus, as rational economic individuals, executives can anticipate the severity of consequences (Chen, 2013a). We suggest that management, as rational eco- nomic persons, will report earnings selectively to maximise their own benefit. When making selective reporting of operating performance at the end of an accounting period, executives will consider the general market environment and the profit status of indus- trial competitors and make corresponding adjustments to the operating performance (He, 2013; Tomy, 2012). Specifically, when the whole market and industrial competi- tors perform well but a particular company in the industry is the first to publish bad news, its stockholders will interpret this as inability or insufficiency of efforts by the management, and stockholders will probably assume that the lower-than-average perfor- mance of the company is a result of bad management. In the face of this threat, rational executives will be motivated to manipulate profits upwards in order to protect their own compensation and to avoid reporting profits that are lower than those of their com- petitors (Conrad et al., 2002; Tomy, 2012). Business cycle refers to the phenomena of alternation and circulation between eco- nomic expansion and economic contraction that periodically occur in an economic operation. A business cycle can be divided into four stages (recession, floor, expansion, and peak) or two phases, namely the ascending (boom) stage and descending (reces- sion) stage. However, the divisions of a business cycle mainly depend on GDP growth rate, as can be found in the studies of Liu (2009), Li (2009), and Jiang and Liu (2011). When the macro-economy is in an ascending (boom) stage, the GDP growth rate tends to be high and the economic prospect is expected to be positive. When the macro- economy is in a descending (recession) stage, the GDP growth rate tends to be low and the economic prospect is expected to be negative. In the boom stage of the business cycle, the general performance of listed compa- nies in the capital market tends be good, and stockholders have high expectations of 140 Wang et al. the operating performance of such companies. A particular company in the industry posting a loss or releasing earnings lower than the industrial average level is bad news that shocks market investors. When the number of companies in the industry is high and the product market competition is stronger, investors have more choices of invest- ment opportunity and will choose companies with better performance. Meanwhile, badly performing companies are identified more easily, which brings a greater shock to the stock prices of companies posting losses (Strobl, 2013). From the perspective of an executive compensation contract, executive compensa- tion not only depends on the increase in earnings of this year over last, but also consid- ers the operating performance of industrial competitors. Even though the company’s performance is higher than that of the previous year, if it is lower than the industrial average level, the company will be regarded as failing to maximise the interests of its shareholders. This result influences shareholders’ evaluation of the effort and ability of the management, which eventually affects executive compensation and even the posi- tion of the managers (Albuquerque, 2011; Dye, 1992; Gibbons & Murphy, 1990). Therefore, forced by the pressure of market competition, the management has stronger motivation for earnings manipulation. In the ascending stage of the macro-economy, positive earnings manipulation appears pro-cyclical. However, in the descending stage of the macro-economy, the general performance of listed companies shows a downward trend and, at this time, investors have pes- simistic expectations of the future trend. Meanwhile, from the perspective of an execu- tive compensation contract, economic recession can serve as an excuse for the bad performance of the company, persuading shareholders into believing that the bad per- formance cannot be attributed to the laziness or inability of the management. In this way, the management has a motivation to take a ‘big bath’ in accounting terms, so that they can hide profit for the boom stage and engage in earnings competition during the next boom. Besides, hiding profits at this time helps gain more government compensa- tion (Wang, Li, & Xing, 2014). But if the loss is not severe during the economic reces- sion, the company will have a stronger motivation to exaggerate profit and turn around (Chen, 2013a, 2013b; Strobl, 2013). Therefore, in the descending stage of the macro- economy, companies not only have a motivation for upwards earnings manipulation in order to turn around, but also have a motivation for downwards earnings manipulation, hiding earnings to engage in earnings competition in the next boom stage. In the ascending stage of the macro-economy, the upwards earnings management of companies will be affected by management’s expectation and the difference between a firm’s profit and the industrial average profit. Specifically, in the ascending stage of the macro-economy, if the management expects their company’s profit to be lower than the industrial average level, they will have a strong motivation for upwards earnings man- agement to avoid investors’ interpretation of bad performance as bad management, which will pose a threat to their compensation or could lead to dismissal. On the con- trary, if the management expects a firm’s profit to be higher than the industrial average, they will have weak motivation for manipulating the profit. Therefore, when the man- agement expects a firm’s profit to be lower than the industrial average level, the motivation for the pro-cyclicality of upwards earnings management will be stronger in the ascending stage of the economic cycle. H1: Upwards earnings management is positively associated with the ascending stage of the business cycle (and hence is more pro-cyclical in the ascending than in the descending stage of the cycle) when management expects lower profits than the industry average for the current year. China Journal of Accounting Studies 141 The business cycle is proxied by the rate of change of GDP. The management expecta- tion of lower profits for the current year is proxied by the observation that the previous year’s profit is lower than the previous year’s industrial average. 2.2. Product market competition, the boom stage of business cycle, and upwards earnings management of the companies Following on from H1, product market pressure is a factor that leads to management’s profit manipulation during different stages of the business cycle. Thus, the degree of upwards earnings management for companies facing different product market pressures will be different. The impact of product market pressure is based on information and constraint mechanisms. The information mechanism of the product market means that competing companies are correlated in their production and operation information, which weakens the degree of information asymmetry between managers and shareholders, thus constraining earn- ings management (Chen & Xu, 2011; Holmstrom, 1982). The constraint mechanism of the product market means that market competition increases the likelihood of losses or bankruptcy of companies, forcing managers to exert more effort to improve the operat- ing performance to avoid the consequences of bad performance, such as reduced com- pensation or dismissal (Chen & Xu, 2011). Markrian and Santalo (2010) demonstrate theoretically that only when stockholders are able to observe fully the operational state of the company (i.e., the market is com- pletely efficient), could product market competition constrain earnings management through an information mechanism; otherwise, it will constrain competition. Markrian and Santalo’s(2010) empirical test using a large sample has supported this conclusion. Further, the empirical test by Chen and Xu (2011), using a sample of Chinese compa- nies, found that the stronger the product market competition, the stronger the motiva- tion for earnings management by companies. Thus, the greater the product market pressure, the more significant the pro-cyclicality of upwards earnings management. If product market competition is a key mechanism affecting upwards earnings management, then the degree of upwards earnings manage- ment between state-owned and non-state-owned enterprises in China will differ in the boom stage. On the one hand, compared with state-owned enterprises, non-state-owned enterprises tend to be younger and relatively small in size, and are usually in the competitive product markets (He, 2013). On the other hand, state-owned enterprises enjoy more preferential policies and enter the regulated and monopoly industries more easily than do non state-owned enterprises (Fang, 2007; World Bank, 2006). This discussion motivates Hypothesis 2a, which is as follows: H2a: Compared with state-owned enterprises, non-state-owned enterprises tend to face greater market competition pressure and show more significant pro-cyclicality of upwards earnings management. The evidence for H2a is that, for non-state-owned enterprises, there is a stronger posi- tive association between upwards earnings management and the change of GDP in the ascending stage of the business cycle. If product market pressure is a key mechanism affecting upwards earnings manage- ment of companies, some differences in the significance of pro-cyclicality for upwards earnings management between companies with different degrees of regulation may 142 Wang et al. exist. The reason is that China is an emerging market in a transformational period and imposes more restrictions on ‘the industries of national security significance, natural monopoly industries, industries providing important public services and goods, pillar industries, and industries of high and new technology’ (Luo & Liu, 2009; Xia & Chen, 2007) as regulated industries, thus limiting the access of private capital to these industries, including mining (B), oils/chemicals/plastic cement/plastics (C4), metal and non-metal (C6), electric power/gas/hydraulic production and supply (D), transportation and warehousing (F), information technology (G), and media and cultural industry (L) (Luo & Liu, 2009; Xia & Chen, 2007). Regulated industries build barriers to access by new enterprises, thus limiting the degree of product market competition in these industries. Hence, industrial regulation divides Chinese listed companies into two categories of strong market competition and weak market competition. This distinction alleviates the problem of endogeneity, and helps to better capture product market competition as a key mechanism affecting upwards earnings management of companies in the ascending stage of the macro-economic cycle. According to the above discussion, Hypothesis 2b is as follows: H2b: Compared with companies in regulated industries, companies in non-regulated indus- tries face greater market competition and show more significant pro-cyclicality of upwards earnings management. The evidence for H2b is that, for companies in non-regulated industries, there is a stronger positive association between upwards earnings management and the change of GDP in the ascending stage of the business cycle. 3. Research design 3.1. Method of calculating earnings management This paper uses the modified Jones Model to calculate the discretionary accruals of Chinese listed companies. It is considered to produce a better outcome according to the research conclusions of Xia (2003) and Huang and Xia (2009). Moreover, based on the studies by Lei and Liu (2006), Wang, Wang, and Gong (2009), and Zhang (2010), this paper uses the KLW Model proposed by Kothari, Leone, and Wasley (2005) to conduct a robustness test. The results show that the conclusions of this paper have not materi- ally changed. The calculation process using the modified Jones Model is as follows. According to the modified Jones Model (Dechow, Sloan, & Sweeney, 1995), the OLS method is used to estimate α , α , α in Model (1) by sub-year and sub- 1 2 3 industry samples. The estimated regression coefficients are then substituted into Model (2) to calculate the non-discretionary accruals. The non-discretionary accruals estimated through Model (2) are then substituted into Model (3) to obtain the discretionary accruals. To eliminate the effect of outliers, all regression variables are winsorized at the first and 99th percentiles. Meanwhile, to guarantee the reliability of the regression equation, the samples with fewer than 15 firm-year observations have been deleted. TA = A ¼ a ð1 = A Þþ a ðDREV = A Þþ a ðPPE = A Þþ e (1) t t1 1 t1 2 t t1 3 t t1 t China Journal of Accounting Studies 143 NDA ¼ a ð1 = A Þþ a ðDREV = A  DREC = A Þþ a ðPPE = A Þ (2) t 1 t1 2 t t1 t t1 3 t t1 DA ¼ TA =A  NDA (3) t t t1 t where TA is the total accruals, which equals the operating profit(NT ) in year t minus t t the net operating cash flow in year t (CFO ). A is the total assets at the end of year t t–1 t–1, while ΔREV is the variation between prime operating revenues of year t and t–1 year t–1. PPE is the total original value of fixed assets at the end of year t. NDA is t t the non-discretionary accruals in year t after being adjusted by the total assets at the end of year t–1. ΔREC is the variation between account receivables of year t and t–1 year t–1. DA is the discretionary accruals of year t. 3.2. Regression model Given the research hypotheses, we consider the indicators that will affect earnings man- agement of companies and put forward the regression equation (4), as follows: þDiscretionary accruals ¼ b þ b Cycle GDP þ b  Control þ e (4) i;t i;t 0 1 i +Discretionary_accruals is the positive discretionary accruals calculated using the modified Jones Model. There are 8,252 observations of positive discretionary accruals, and the remainder are the negative discretionary accruals. Cycle_GDP is the proxy vari- able for a business cycle. We define the proxy variable for a business cycle with one year lagged data because we suppose that the decision on earnings management this year tends to depend on the information from the previous business cycle (i.e., the transmission from macro-economic cyclical fluctuation to micro-corporate behaviour has a certain lag). This process will also ease any possible endogeneity problem in the effect of macro-economic cyclical fluctuations on micro-corporate behaviour. As shown in Figure 2, compared with macro-economic cyclical fluctuations, earnings management of companies shows an obvious lag, which indirectly indicates that defining the proxy variable for a business cycle with one year lagged data is reasonable. For the definition of a business cycle, this paper draws on the studies by Li (2009), Jiang and Liu (2011), Strobl (2013), and Wu (2013) to use the year-on-year growth rate of GDP in China as the proxy variable for a business cycle (Cycle_GDP). This defini- tion is based on the following two considerations. (1) Using a dummy variable to Figure 2. Annual GDP growth rate and earnings management of firms from 2000 to 2012. Note: The data for annual GDP growth are from the National Bureau of Statistics. Earnings man- agement is calculated for the companies in the sample using the modified Jones Model and data from the CSMAR database. 144 Wang et al. represent boom and recession periods will probably produce errors because of subjec- tive judgment. Thus, we use the continuous variable of year-on-year growth rate of GDP to avoid this kind of error. (2) Using this continuous variable can effectively avoid the systematic impact of implementation of new accounting standards and reform of the shareholder structure resulting from using a dummy variable. Because the GDP growth rate in this paper is one year lagged, the non-lag data total 11 years (from 2000 to 2010). Control in regression equation (7) represents the control variables based on previous studies. The control variables include market-to-book ratio (Market_to_book_ratio), equity incentive (Equity_incentive), rights offering (Rights_offering, Prior_rights_offer- ing), public offering (Public_offering, Prior_public_offering), debt paying ability (Leverage), size of the company (Firm_size), net cash flow (Cash_flow), profitability (ROA), auditing opinion (Audit_opinion), and the competence of independent auditors (BIG4_auditor). We have three moderator variables. First, to explore the specific mechanism of the pro-cyclicality of upwards earnings management, we use the difference between the adjusted profit (adjusted by discretionary accruals) of the company in the previous year and the previous year’s average industrial adjusted profit (adjusted by discretionary accruals) of that year as an indicator of management’s expectation for the profitofthe current year. We then have two moderator variables to represent the pressure of product market competition. They are state ownership and regulated industry status. Finally, we set the industry dummy variables and control for the industry’s possible effect on the regression model to guarantee the reliability of fitting the model. Specific definitions of variables are shown in Table 1. 3.3. Data This paper chooses the annual data between 2001 and 2011 from the China Stock Mar- ket and Accounting Research (CSMAR) database as the research sample. Given that estimated discretionary accruals must be one year lagged, the GDP growth rates are from 2000 to 2010. We obtain all our data from the CSMAR database and the National Bureau of Statistics. Sample companies with unavailable data and operating in the finance and insurance industries are excluded from the study. We observed a total of 14,676 discretionary accruals, from which we eliminated the negative discretionary accruals, leaving us with 8,252 firm-year observations for positive discretionary accru- als in the regression model. To minimise the influence of outliers, we winsorize all con- tinuous variables at the first and 99th percentiles, except dummy variables, such as proxies for a macro-economic cycle. Table 2 provides a distribution of the sample by industry. Of the 19 industries, most observations come from more cyclical industries, such as petrochemicals, metals and non-metals, machinery, real estate, wholesale and retail, and information technology, accounting for over 50% of the total sample. The operating performances of the companies in these cyclical industries tend to be more sensitive to the fluctuations of the macro-economy, thereby offering greater confidence in the results. Table 3 is the annual distribution of the sample, which shows a rising trend. Table 4 presents the main variables of the correlation coefficient in this paper. The results demonstrate that the economic cycle is significantly positively correlated with upwards earnings management (+Discretionary_accruals), which means the economic China Journal of Accounting Studies 145 Table 1. Definitions of the variables. Types Variable Meaning Definition Dependent variable +Discretionary_accruals The value of positive The positive discretionary accruals discretionary accruals calculated using the modified Jones Model. Independent test variable Cycle_GDP Business cycle Drawing on Li (2009) and Jiang and Liu (2011), we use annual GDP growth rate as the proxy for the business cycle with one year lagged data. Independent moderator variables Earnings_difference Expectation of An indicator variable that takes on the management value of 1 if the difference between the company’s profit of the previous year and the previous year’s industrial average level is negative and 0 otherwise. State_ownership Nature of property right If the actual controller is the government, it is defined as state-owned enterprise and takes the value of 0, and 1 otherwise. Regulated_industy Industrial regulation An indicator variable that takes on the value of 0 if the company belongs to regulated industry and 1 otherwise. Control variables Market_to_book_ratio Market-to-book ratio Measured as the market value of the company divided by the book value. Loss Losses An indicator variable that takes on the value of 1 if the profit of this year is negative and 0 otherwise. Equity_incentive Equity incentive An indicator variable that takes on the value of 1 in the year before issuing an equity incentive plan, and 0 otherwise. Rights_offering Annual dummy variable An indicator variable that takes on the (rationed shares) value of 1 in the year of rights offering, and 0 otherwise. Prior_rights_offering Annual dummy variable An indicator variable that takes on the (prior to rights offering) value of 1 in the previous three years of conducting rights offering, and 0 otherwise. Public_offering Annual dummy variable An indicator variable that takes on the (issuing additional value of 1 in the year of public offering, shares) and 0 otherwise. Prio_additional_shares Annual dummy variable An indicator variable that takes on the (prior to issuing value of 1 in the previous three years of additional shares) public offering, and 0 otherwise. Leverage Financial leverage Measured as debt divided by total assets. Firm_size Firm size Measured as the natural log of total assets. Cash_flow Net cash flow Measured as net cash flow from operating activities divided by total assets. (Continued) 146 Wang et al. Table 1. (Continued). Types Variable Meaning Definition ROA Return on assets Measured as net income divided by total assets. Audit_opinion Auditing opinion An indicator variable that takes on the value of 1 for the standard auditing opinion, and 0 otherwise. BIG4_auditor Big four accounting A dummy variable that equals 1 if the firms listed company is audited by Big Four and 0 otherwise. Industry_dum Industry dummy Based on standard classification of variables industries in China, the manufacturing industry is classified to secondary level. We set one dummy variable for each industry to control for industry fixed effects. Table 2. Sample distribution of positive discretionary accruals by industry. Industry N Percentage Industry N Percentage Agriculture, forestry, livestock 131 1.59 Other manufacturing 31 0.38 farming, fishery Mining 188 2.28 Utilities 383 4.64 Food & beverage 337 4.08 Construction 136 1.65 Textiles & apparel 284 3.44 Transportation 336 4.07 Paper & printing 131 1.59 Information 7,499 6.05 technology Petrochemicals 802 9.72 Wholesale & retail 639 7.74 trade Electronics 372 4.51 Real estate 927 11.23 Metals & non-metals 715 8.66 Social services 301 3.65 Machinery 1, 402 16.99 Communication & 109 1.32 cultural Pharmaceuticals 529 6.41 Total 8, 252 100.00 Table 3. Annual distribution of sample of positive discretionary accruals. Year N Percentage Year N Percentage 2001 475 5.76 2007 764 9.26 2002 550 6.67 2008 872 10.57 2003 661 8.01 2009 881 10.68 2004 649 7.86 2010 929 11.26 2005 664 8.05 2011 1,041 12.62 2006 766 9.28 Total 8,252 100.00 cycle does have an impact on the earnings management behaviour of the firm. The rest of the correlation coefficients of control variables are below 0.5, showing that there is no serious effect of multicollinearity on the regression results. China Journal of Accounting Studies 147 Table 4. Correlation matrix for the main variables. 12 34 567 8910 11 12 13 14 1. +Discretionary_accruals 1.000 *** 2. Cycle_GDP 0.081 1.000 *** *** 3. Market_to_book_ratio 0.120 −0.052 1.000 4. Loss −0.032 0.002 0.006 1.000 *** *** *** *** 5. Equity_incentive 0.123 0.047 0.093 −0.065 1.000 *** *** 6. Rights_offering −0.011 −0.065 −0.012 −0.034 −0.004 1.000 *** *** *** 7. Prior_rights_offering −0.029 −0.030 0.013 −0.050 0.012 0.014 1.000 *** *** *** *** *** ** *** 8. Prior_public_offering 0.052 0.134 0.060 −0.057 0.052 −0.028 −0.035 1.000 *** *** *** *** *** *** 9. Public_offering 0.114 0.080 0.025 −0.069 0.093 −0.030 0.007 0.018 1.000 *** *** *** *** *** ** *** *** 10. Leverage 0.069 0.045 −0.067 0.272 −0.077 −0.026 −0.009 0.079 −0.034 1.000 *** *** *** *** ** *** *** *** 11. Firm_size 0.016 0.059 −0.371 −0.177 0.051 0.022 0.013 0.040 0.152 0.085 1.000 *** ** *** *** *** *** *** *** *** 12. Cash_flow −0.350 0.024 0.119 −0.249 0.041 0.000 0.043 0.054 −0.018 −0.268 0.115 1.000 *** *** *** *** *** *** *** *** *** *** *** 13.ROA 0.105 0.034 0.148 −0.685 0.109 0.013 0.052 0.082 0.073 −0.434 0.194 0.408 1.000 *** *** *** *** ** *** *** *** *** *** *** *** 14. Audit_opinion −0.006 −0.031 0.110 0.393 −0.046 −0.023 −0.035 −0.044 −0.063 0.355 −0.201 −0.171 −0.417 1.000 *** *** ** *** *** *** *** *** 15. BIG4_auditor 0.006 −0.006 −0.060 −0.054 0.027 −0.013 −0.005 −0.001 0.010 −0.042 0.271 0.086 0.079 −0.035 Note: The table above presents the regression coefficient for each independent variable. Refer to Table 1 for definition of variables. T-statistics are presented in the parentheses below the coefficients and are corrected using a two-dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. 148 Wang et al. 4. Empirical results 4.1. Descriptive statistics Table 5 provides descriptive statistics for our observations. The mean value (0.098) of positive discretionary accruals (+Discretionary_accruals) is slightly greater than the median value (0.066), which is slightly right skewed. The mean and the median value of the dummy proxy variable for the business cycle (Cycle_GDP) are approximately the same, and Cycle_GDP presents a normal distribu- tion. Among the control variables, the mean and the median values of market-to-book ratio (Market_to_book_ratio) are 2.189 and 1.726, respectively, which are slightly right skewed. Moreover, the mean value of Loss (Loss) is 0.110, indicating that approxi- mately 11.0% of the sample companies are making losses. The mean value of the equity incentive (Equity_incentive) is 0.045, which indicates that 4.5% of the sample companies offer a stock option incentive. The mean value of Rights_offering (Rights_offering) is 0.013, indicating that approximately 1.3% of the sample companies conduct rights offering during the year. The mean value of issuing additional stock in a public offering (Public_offering) is 0.063, which shows that 6.3% of the sample companies issue additional stock. The mean and the median values of the leverage ratio, size of company (Firm_size), and net cash flow (Cash_flow) are quite similar, demonstrating that these variables present normal distributions. Furthermore, the mean value (0.031) of return on assets (ROA) is less than the median value (0.033), which is slightly left skewed. The mean value of auditing opinion (Audit_opinion) is 0.070, indi- cating that 7.0% of the sample companies have non-standard audit opinions. The mean value of Big4 auditors (BIG4_auditor) is 0.064, which shows that 6.4% of the sample firms are audited by Big4 auditors. Table 5. Descriptive statistics of the main variables. Variables N Mean STD MIN 25% Median 75% MAX +Discretionary_accruals 8,252 0.098 0.100 0.001 0.029 0.066 0.132 0.492 Cycle_GDP 8,252 0.104 0.017 0.083 0.092 0.100 0.113 0.142 Earnings_difference 8,252 0.578 0.494 0.000 0.000 1.000 1.000 1.000 Sate_ownership 8,252 0.489 0.500 0.000 0.000 0.000 1.000 1.000 Regulated_ownership 8,252 0.633 0.482 0.000 0.000 1.000 1.000 1.000 Market_to_book_ratio 8,252 2.189 1.607 0.536 1.224 1.726 2.589 10.629 Loss 8,252 0.110 0.313 0.000 0.000 0.000 0.000 1.000 Equity_incentive 8,252 0.045 0.207 0.000 0.000 0.000 0.000 1.000 Rights_offering 8,252 0.013 0.114 0.000 0.000 0.000 0.000 1.000 Prior_rights_offering 8,252 0.020 0.139 0.000 0.000 0.000 0.000 1.000 Public_offering 8,252 0.063 0.243 0.000 0.000 0.000 0.000 1.000 Prior_public_offering 8,252 0.151 0.358 0.000 0.000 0.000 0.000 1.000 Leverage 8,252 0.519 0.237 0.056 0.369 0.511 0.645 2.026 Firm_size 8,252 21.555 1.104 18.823 20.803 21.446 22.165 25.102 Cash_flow 8,252 0.020 0.080 −0.211 −0.021 0.024 0.066 0.277 ROA 8,252 0.031 0.077 −0.419 0.012 0.033 0.061 0.206 Audit_opinion 8,252 0.070 0.255 0.000 0.000 0.000 0.000 1.000 BIG4_auditor 8,252 0.064 0.245 0.000 0.000 0.000 0.000 1.000 Note: Refer to Table 1 for definition of variables. China Journal of Accounting Studies 149 4.2. Main results Table 6 reports the regression results for the test of Hypothesis 1. The dependent vari- able is positive discretionary accruals. Table 6. Regression of upwards earnings management on business cycle and on the difference between the company’s profit of the previous year and the industrial average of that year. (1) (2) (3) Larger than or Less than zero equal to zero Interaction ** Cycle_GDP 0.399 0.103 0.005 (2.09) (0.67) (0.03) Earnings_difference −0.029 (–1.31) *** Earnings_difference *Cycle_GDP 0.562 (2.89) *** *** *** Market_to_book_ratio 0.008 0.007 0.007 (3.32) (6.26) (5.37) *** *** *** Loss 0.019 0.030 0.024 (2.64) (4.55) (4.18) *** *** *** Equity_incentive 0.077 0.017 0.041 (5.10) (3.38) (5.12) Rights_offering −0.002 0.008 −0.000 (–0.28) (1.20) (–0.03) Prior_rights_offering −0.006 0.000 −0.005 (–0.66) (0.03) (–0.77) *** ** *** Public_offering 0.015 0.006 0.009 (2.63) (2.04) (3.20) *** *** *** Prior_ public_offering 0.032 0.028 0.030 (5.15) (3.01) (7.51) ** ** Leverage 0.015 0.021 0.022 (1.05) (2.39) (2.24) *** *** Firm_size 0.006 0.003 0.006 (2.37) (1.45) (2.94) *** *** *** Cash_flow −0.432 −0.658 −0.546 (–8.27) (–17.19) (–12.37) *** *** *** ROA 0.354 0.505 0.421 (4.96) (8.08) (8.30) Audit_opinion −0.007 0.010 −0.001 (–0.95) (1.01) (–0.09) BIG4_auditor 0.017 0.000 0.008 (1.12) (0.04) (0.84) ** *** Constant −0.099 −0.030 −0.105 (–1.97) (–0.85) (–3.05) *** Coefficient difference test of 6.66 Cycle_GDP (0.01) N 3,712 4,540 8,252 Industry_dum Yes Yes Yes *** *** *** F value 40.14 47.45 80.97 R 0.290 0.383 0.328 Note: The dependent variable is positive discretionary accruals. The table presents the regression coefficient for each independent variable. Refer to Table 1 for the definition of variables. T-statistics are presented in parentheses below the coefficients and are corrected using a two dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. 150 Wang et al. Columns (1) and (2) test the results of (1) those cases where management’s expected profit (the difference between the previous year’s profit of the company and the previous year’s industrial average profit) is less than zero, and (2) those cases where management’s expected profit is greater than or equal to zero. Column (3) shows the regression results of defining this relationship as a dummy variable (Earnings_difference) and interacting Earnings_difference with the proxy of business cycle (Earnings_difference*Cycle_GDP). The dummy variable Earnings_difference takes on the value of 1 if the difference between the company’s profit of the previous year and the previous year’s industrial average level is negative and the value 0 other- wise. According to the regression results in Table 6, the coefficient of the proxy for business cycle (Cycle_GDP) is significantly positive in Column (1), but non-significant in Column (2). The difference test (SUE test) results of the coefficients of Cycle_GDP in Columns (1) and (2) are statistically significant at the 1% level. We use this differ- ence test to show that one set of coefficients is significantly larger than the other set of coefficients. Furthermore, in Column (3), the coefficient of the interaction term (Earn- ings_difference*Cycle_GDP) is significantly positive. These results support Hypothesis 1, that upwards earnings management is positively associated with the growth rate of GDP when the previous year’s profit of the company is lower than the previous year’s industrial average level (i.e. management’s expectations are for lower profits). Hence we can say that upwards earnings management is more pro-cyclical in the ascending stage of the business cycle, Tables 7 and 8 report the regression results for the test of Hypotheses 2a and 2b. When the previous year’s profit of a company is lower than the previous year’s indus- trial average level, the coefficients of the proxy of business cycle (Cycle_GDP) in the subgroups of non-state-owned enterprises and non-regulated industry are significantly positive, and the coefficient difference test is significant. The results of the difference test (SUE test) show that the coefficients of the proxy of business cycle (Cycle_GDP) in the subgroups of non-state-owned enterprises and non-regulated industry are significantly larger than those in the subgroups of state- owned enterprises and regulated industry. When the profit of the companies is higher than or equal to the industrial average level, the coefficients of the proxy of business cycle (Cycle_GDP) are non-significant in either group and the coefficient difference test (SUE test) is significant. In addition, the interaction terms: State ownership  Earnings difference  Cycle GDP and Regulated industry  Earnings difference  Cycle GDP are both significantly positive. Taken together, these results indicate that when compa- nies produce profits lower than the industrial average level and face greater market competition pressure, they tend to show more significant pro-cyclicality of upwards earnings management in the ascending stage of the business cycle, which further sup- ports Hypotheses 2a and 2b. 4.3. Further tests According to the studies of Cheng and WarFeId (2004) and Xie and Lv (2011), management incentive is the important driving force for management to conduct earn- ings management. Thus, executive incentive might be an important factor of upwards earnings management. This paper uses executive stockholding and management China Journal of Accounting Studies 151 Table 7. Nature of property rights, difference in profit, business cycle, and upwards earnings management. (1) (2) (3) Company’s profit< Company’s profit ≥ Industrial profit Industrial profit Total sample Non-state- Non-state- owned State-owned owned State-owned ** * Cycle_GDP 0.646 0.262 0.224 −0.042 −0.094 (2.30) (1.66) (1.21) (–0.25) (–0.63) ** State_ownership −0.022 (–2.33) Earnings_difference −0.013 (–1.15) ** State_ownership *Cycle_GDP 0.199 (2.18) *** Earnings_difference *Cycle_GDP 0.390 (3.85) State_ownership* −0.034 Earnings_difference (–1.42) State_ownership* 0.395 Earnings_difference*Cycle_GDP (1.89) *** *** *** *** *** Market_to_book_ratio 0.009 0.009 0.006 0.008 0.007 (2.65) (3.65) (5.52) (4.89) (5.53) * *** ** *** Loss 0.024 0.015 0.033 0.023 0.023 (1.87) (1.94) (3.36) (2.51) (4.18) *** *** *** *** Equity_incentive 0.056 0.106 0.025 −0.010 0.039 (5.04) (3.56) (3.83) (–1.03) (4.85) Rights_offering 0.006 0.004 0.004 0.018 0.001 (0.15) (0.63) (0.56) (1.57) (0.19) Prior_rights_offering −0.027 0.011 −0.006 0.013 −0.004 (–1.47) (0.73) (–1.07) (1.35) (–0.62) * ** ** *** Public_offering 0.018 0.014 0.012 0.003 0.010 (1.88) (1.97) (2.20) (0.87) (3.28) *** *** ** *** *** Prior_ public_offering 0.039 0.029 0.031 0.024 0.030 (3.02) (3.12) (2.53) (2.74) (7.42) ** ** Leverage 0.002 0.042 0.021 0.028 0.021 (0.08) (2.00) (1.42) (1.55) (2.14) * ** *** Firm_size 0.008 0.005 0.004 0.001 0.006 (1.87) (1.14) (2.31) (0.58) (3.10) *** *** *** *** *** Cash_flow −0.392 −0.482 −0.670 −0.650 −0.546 (–5.47) (–7.33) (–18.24) (–11.28) (–12.37) *** *** *** *** *** ROA 0.334 0.390 0.487 0.521 0.415 (4.07) (3.26) (6.29) (6.84) (8.17) ** Audit_opinion −0.024 0.008 0.008 0.013 −0.000 (–2.04) (1.23) (0.70) (0.98) (–0.08) BIG4_auditor 0.034 0.014 −0.004 0.006 0.008 (1.29) (0.76) (–0.32) (0.82) (0.84) Constant −0.151 −0.058 −0.048 0.016 −0.102*** (–0.99) (–0.77) (–1.51) (0.34) (–2.72) Industry_dum Yes Yes Yes Yes Yes ** * Coefficient difference test of 4.31 3.25 Cycle_GDP (0.04) (0.07) N 1,617 2,095 2,419 2,121 8,252 *** *** *** *** *** F value 16.85 30.35 26.70 22.30 73.08 R 0.283 0.320 0.367 0.416 0.331 Note: The table above presents the regression coefficient for each independent variable. Refer to Table 1 for definition of variables. T-statistics are presented in the parentheses below the coefficients and are corrected using a two dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. 152 Wang et al. Table 8. Regression of upwards earnings management on business cycle, regulated and non- regulated industry and higher and lower profit related to industry average. (1) (2) (3) Company’s profit Company’s profit < Industrial profit ≥ Industrial profit Non- Non- regulated Regulated regulated Regulated ** Cycle_GDP (1) 0.585 0.064 0.048 0.225 0.055 (2.32) (0.29) (0.28) (1.44) (0.26) ** Regulated_industry −0.026 (–2.02) Earnings_difference 0.001 (0.05) Regulated_industry*Cycle_GDP −0.080 (–0.59) Earnings_difference*Cycle_GDP 0.200 (0.72) Regulated_industry*Earnings_difference −0.046 (–1.87) ** Regulated_industry* 0.556 Earnings_difference*Cycle_GDP (2.39) *** * *** *** *** Market_to_book_ratio 0.009 0.006 0.008 0.004 0.007 (3.07) (1.66) (6.58) (2.57) (5.33) ** ** *** *** Loss 0.027 0.011 0.025 0.037 0.024 (2.34) (1.35) (2.53) (6.01) (4.06) *** *** ** ** *** Equity_incentive 0.053 0.107 0.017 0.014 0.041 (3.81) (4.13) (2.54) (2.34) (5.09) Rights_offering 0.002 −0.009 0.002 0.019 −0.000 (0.16) (–1.02) (0.33) (1.73) (–0.07) Prior_rights_offering −0.023 0.014 −0.006 0.008 −0.005 (–1.76) (0.78) (–0.57) (1.53) (–0.81) *** *** *** Public_offering 0.027 −0.002 0.002 0.012 0.010 (3.52) (–0.26) (0.50) (3.00) (3.20) *** * *** ** *** Prior_ public_offering 0.040 0.021 0.037 0.015 0.030 (3.86) (1.66) (3.22) (2.12) (7.54) *** ** Leverage 0.018 0.011 0.040 −0.012 0.022 (1.02) (0.47) (4.11) (–0.98) (2.28) *** ** *** Firm_size 0.012 −0.004 0.006 −0.003 0.005 (4.00) (–0.67) (2.32) (–1.44) (2.87) *** *** *** *** *** Cash_flow −0.440 −0.410 −0.688 −0.616 −0.546 (–7.75) (–4.54) (–12.87) (–13.73) (–12.39) *** *** *** *** *** ROA 0.372 0.321 0.453 0.577 0.420 (4.56) (4.23) (6.22) (9.30) (8.30) *** Audit_opinion −0.005 −0.014 0.008 0.013 −0.001 (–0.38) (–2.85) (0.61) (1.52) (–0.10) ** BIG4_auditor 0.003 0.047 −0.005 0.006 0.008 (0.17) (2.07) (–0.63) (0.89) (0.85) *** ** *** ** Constant −0.323 0.166 −0.114 0.120 −0.075 (–2.97) (1.51) (–2.00) (2.99) (–2.27) *** Coefficient difference test of 9.26 1.71 Cycle_GDP (0.002) (0.191) Industry_dum Yes Yes Yes Yes Yes N 2,388 1,324 2,832 1,708 8,252 F value 28.15 22.61 41.12 27.26 75.48 R 0.277 0.300 0.390 0.373 0.330 Note: The table above presents the regression coefficient for each independent variable. Refer to Table 1 for definition of variables. T-statistics are presented in the parentheses below the coefficients and are corrected using a two-dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. China Journal of Accounting Studies 153 shareholding to represent the degree of management incentive, in order to explore macro-economic cyclical fluctuations’ impact on upwards earnings management under different degrees of management incentive. The untabulated regression results show that the coefficient of the interaction term (Earnings_difference*Cycle_GDP) in the sub- group of executive shareholding (coefficient 0.693, and T value is 2.60) and manage- ment shareholding (coefficient 0.642, and T value is 2.50) is greater than that in the subgroup without executive shareholding (coefficient 0.450, and T value is 2.97) and management shareholding (coefficient 0.424, and T value is 4.24). However the SUE test is not significant. We interpret these results as weakly indicating that management incentive is an important factor for upwards earnings management. A number of studies have demonstrated that financing need is an important motiva- tion for companies to conduct earnings management. During macro-economic cyclical fluctuations, the financing environment and financing needs of companies change. Thus, the degree of financial constraints of companies might become an important factor for the pro-cyclicality of upwards earnings management. From the perspective of financial constraint, this paper uses leverage ratio and size of the company as the proxies for financial constraints, and groups the data by the median value. Companies with high leverage ratio and small size are defined as companies with strong financial constraints; otherwise, they will be considered companies with weak financial constraints. The unre- ported regression results of the subgroups reveal that the coefficient of the interaction terms (Earnings_difference*Cycle_GDP) in the subgroup of high leverage (coefficient 0.635, and T value is 3.14) and small size (coefficient 0.777, and T value is 3.28) is greater than that in the subgroup of low leverage (coefficient 0.495, and T value is 1.58) and small size (coefficient 0.363, and T value is 2.39). However the difference test is not significant (SUE test). We interpret these findings as weakly indicating that financial constraint also is an important factor for upwards earnings management. According to the studies of Chen (2013b) and Li, Wang, and Rong (2013), the impact of degrees of macro-economic cyclical fluctuation on different industries is dif- ferent. This paper follows Kong (2010) and Li et al. (2013) to divide the listed compa- nies sample into two subgroups of industries with strong cyclicality and industries with weak cyclicality. Industries with weak cyclicality include agriculture, food and bever- age, medicine and biological products, medium and culture, and comprehensive indus- try. Industries with strong cyclicality are more sensitive to macro-economic fluctuations. The untabulated results of subgroup regression by industries with strong cyclicality and industries with weak cyclicality show that the coefficient of the interac- tion terms (Earnings_difference*Cycle_GDP) in the subgroup of strong cyclicality (co- efficient 0.508, and T value is 2.66) is more significant than that in the subgroup of weak cyclicality (coefficient 0.207, and T value is 2.80). We also find that the differ- ence test (SUE test) is significant. These results indicate that the pro-cyclicality of upwards earnings management under macro-economic cyclical fluctuations is more sig- nificant in industries with strong cyclicality, which is consistent with the findings of Chen (2013b) and Li et al. (2013). 4.4. Robustness tests We conduct additional sets of tests to assess the sensitivity of our results. We follow the studies of Jiang, Qu, Lu, and Li (2008) to use the Herfindhal Diversified Index (HHI index) based on operating income to measure the degree of market competition of the companies, and then obtain the median value by year. If the company’s HHI 154 Wang et al. index is larger than the median value of this year, then the company is facing lower market competition pressure; otherwise, the company is facing greater market competi- tion pressure. We conduct a new regression by using this variable to take the place of the nature of property right and industrial regulation. The unreported results show that the degree of sensitivity of macro-economic cyclical fluctuations has a different impact on upwards earnings management, which is manifested by the result that companies facing stronger market competition are more significant (compared with weak market competition). The table shows that the coefficient of the interaction term (Earnings_ difference*Cycle_GDP) is greater for strong market competition (0.569) than it is for weak market competition (0.429). However the difference test is not significant (SUE test). We interpret these findings as weakly indicating the pro-cyclicality of upwards earnings management. This result points towards the ‘market competition hypothesis’ of upwards earnings management under macro-economic cyclical fluctuations. 5. Conclusion As an important domain of financial and accounting research, earnings management has captured a great deal of attention from scholars. Existing studies are mainly based on the perspective of the micro-corporate level and are seldom based on the macro- economic level to study the effects on earnings management. However, decisions of companies are made under a given macro environment. Thus, focusing only on the micro level and ignoring the impact of the macro-economic environment limits the generalizability of the research results. Considering the reasons discussed above, this paper takes the annual data from non-financial listed companies in China between 2001 and 2011 as its sample to exam- ine empirically the degree, and key mechanism of companies’ earnings management under cyclical fluctuations of the macro-economy. We find the following: first, upwards earnings management is significantly and positively correlated with the growth rate of the macro-economy, indicating a relatively strong pro-cyclical effect. Secondly, the pro- cyclical effect of upwards earnings management will be more significant if the manage- ment expects corporate profit to be lower than the industrial profit. Thirdly, the pro-cyclical effect of upwards earnings management will be more significant if the company faces greater product market competition pressure. Further tests found that the greater the incentive of the management, the more sensitive the industry is to the eco- nomic cyclical fluctuations, and the more significant the pro-cyclical effect of upwards earnings management becomes. However, financing restrictions do not significantly affect the pro-cyclical effects. The results demonstrate that the key mechanism of upwards earnings management during the economic boom is the difference between the firm’s expected profit and industrial profit, as well as the product market competition pressure. Thus, the combination of the macro-economic state and the companies’ pro- duct market competition degree will play an important role in understanding earnings management behaviour. Admittedly, this paper is subject to some caveats. First, in terms of the definition of business cycle, although this paper uses annual GDP growth rate and divides the boom stage and recession stage by annual GDP growth rate, determining the stages of busi- ness cycles of China remains difficult. Second, because downwards earnings manage- ment has complex motivations and mechanisms, this paper focuses on the upwards earnings management under macro-economic cyclical fluctuations, which does not explain fully the comprehensive impacts of macro-economic cyclical fluctuations on China Journal of Accounting Studies 155 earnings management behaviour, thus requiring further studies to explore this question in the future. Acknowledgements We are most grateful to two anonymous referees and Liansheng Wu from Guanghua School of Management at Peking University for their valuable and constructive suggestions. We acknowl- edge the financial support from Key Projects of Philosophy and Social Sciences Research of Min- istry of Education (10JZD0019), National Natural Science Foundation of China (71072103; 71272228), Program for New Century Excellent Talents in University (NECT-120432) and Pro- gram for Young Scholars in Luojia. We are responsible for the consequences of this article. Notes 1. The reform of non-tradable shares is a process that allowed previously non-tradable shares held by controlling shareholders to be freely tradable on the exchanges. 2. Seemingly Unrelated Estimation, statistical package: http://www.stata.com/manuals13/rsuest. pdf. References Albuquerque, Ana. M. (2011). Do growth- option firms use less relative performance evaluation? Working Paper. Boston University. Barua, A., Lin, S., & Sbaraglia, A. M. (2010). Earnings management using discontinued opera- tions. The Accounting Review, 85, 1485–1509. Bo, X., & Wu, L. (2009). The governance roles of state-owned controlling and institutional investors: A perspective of earnings management. Economic Research Journal, (2), 81–90. (In Chinese). Chen, W. (2013a). Business cycle, industry prosperity and earnings management-empirical evidence from China’s listed companies. Journal of Auditing Research, (5), 96–105. (In Chinese). Chen, W. (2013b). Business cycles, cyclical industries and magnitude of earnings management: Empirical evidence from China’s listed companies. Nankai Business Review, (3), 26–35. (In Chinese). Chen, J., & Xu, Y. (2011). Product market competition, competitive situation and earnings man- agement of listed companies. Public Finance Research, (4), 58–61. (In Chinese). Chen, K., & Yuan, H. (2004). Earnings management and capital resource allocation: Evidence from China’s accounting-based regulation of rights issue. The Accounting Review, 79, 645–665. Cheng, Q., & WarFeId, T. D. (2004). Equity incentives and earnings management. The Account- ing Review, 80, 441–476. Cohen, D., & Zarowin, P. (2007). Earning management over the business cycle. Working paper. Stern School of Business of New York University. Cohen, D., & Zarowin, P. (2010). Accrual-based and real earnings management activities around seasoned equity offerings. Journal of Accounting and Economics, 50,2–19. Coles, J., Hertzel, M., & Kalpathy, S. (2006). Earning management around employee stock option reissues. Journal of Accounting and Economics, 41, 173–200. Conrad, J., Cornell, B., & Landsman, W. (2002). When is bad news really bad news? Journal of Finance, 57, 2507–2532. Dechow, P., Sloan, R., & Sweeney, A. (1995). Detecting earnings management. The Accounting Review, 70, 193–225. Dye, R. (1992). Relative performance evaluation and project selection. Journal of Accounting Research, 30,27–52. Erel, I., Julio, B., Kim, W., & Weisbach, M. (2012). Macroeconomic condition and capital rising. Review of Financial Studies, 25, 341–376. 156 Wang et al. Fan, Y., Xu, B., & Zhou, D. (2013). China business report of 2013. China Europe International Business School (in Chinese). Fang, J. (2007). Ownership, institutional environment and capital allocation. Economic Research Journal, (12), 82–92. (In Chinese). Gibbons, R.., & Murphy, K. J. (1990). Relative performance evaluation for Chief Executive Offi- cers. Industrial and Labour Relations Review, 43, 30S–51S. Hao, S., Jin, Q., & Zhang, G. (2011). Relative firm profitability and stock price sensitivity to aggregate news. The Accounting Review, 86, 1321–1347. Haw, I., Qi, D., Wu, D., & Wu, W. (2005). Market consequences of earnings management in response to security regulations in China. Contemporary Accounting Research, 22,95–143. He, W. (2013). Promotion incentives and earnings management journal of China soft science (in Chinese). Holmstrom, B. (1982). Moral hazard in teams. The Bell Journal of Economics, 13, 324–340. Huang, M., & Xia, X. (2009). An evaluation on specification and power of discretionary accruals models in the Chinese context. Nankai Business Review, (5), 136–147. (In Chinese). Jiang, L., & Liu, X. (2011). Research on economic cyclical fluctuations and cash holding behav- iors of listed companies. Accounting Research, (9), 40–46. (In Chinese). Jiang, F., Qu, Y., Lu, Z., & Li, Y. (2008). Product market competition and dynamic capital struc- ture adjustments. Economic Research Journal, (4), 99–110. (In Chinese). Jin, Q. (2005). Business cycle, accounting behavior and earnings management. Doctoral Thesis. Hong Kong University of Science and Technology. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–292. Klein, A. (2002). Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics, 33, 375–400. Klein, A., & Marquardt, C. A. (2006). Fundamentals of accounting losses. The Accounting Review, 81, 179–206. Kong, Q. (2010). Macroeconomic fluctuations, cyclical industry and capital structure choice. Journal of Beijing Institute of Technology, (6), 31–35. (In Chinese). Korajczyk, R. A., & Levy, A. (2003). Capital structure choice: Macroeconomic conditions and financial constrained. Journal of Financial Economics, 68,75–109. Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionary accruals. Journal of Accounting and Economics, 39, 163–197. Lei, G., & Liu, H. (2006). Between big shareholders’ control, financing scale and the degree of manipulating surplus. Management World, (1), 129–136. (In Chinese). Li, Y. (2009). Why does the operating performance of the Chinese listed companies deviate from the business cycle? Economic Research Journal, (3), 99–109. (In Chinese). Li, W. & Luo, M. (2010). An empirical study of pro-cyclicality of the capital adequacy ratio of commercial banks in China. Journal of Financial Research, (2), 147–157. (In Chinese). Li, Q., Wang, H., & Rong, X. (2013). Firm earnings persistence over the business cycle: Evi- dence from listed companies in china. China Accounting and Finance Review, (4), 129–191. (In Chinese). Liu, S. (2009). Retrospect and prospect of the 60 years of economic growth of China, and the discussion of the new round of business cycle. Economic Perspectives, 10,3–10. (In Chinese). Luo, D., & Liu, X. (2009). Political connection, entry barrier and corporate performance: Evi- dence from private listed companies. Management World, (5), 97–106. (In Chinese). Markrian, G., & Santalo, J. (2010). Product market competition, information and earnings man- agement. Working paper. IE Business School. McLean, D., & Zhao, M. (2014). The business cycle, investor sentiment, and costly external finance. Journal of Finance, 69, 1377–1409. Piotroski, J. D., & Wong, T. J. (2012). Institutions and information environment of Chinese listed firms. Working Paper. Stanford University, the Chinese University of Hong Kong. Strobl, G. (2013). Earnings manipulation and the cost of capital. Journal of Accounting Research, 51, 449–573. Tomy, R. E. (2012). Earnings persistence over the business cycle. Working paper. Stanford University. China Journal of Accounting Studies 157 Wang, H., Li, Q., & Xing, F. (2014). Financial crisis, government subsidies, and earnings manip- ulation-The empirical evidence from listed companies. Management World, (7), 157–167. (In Chinese). Wang, Y., Wang, L., & Gong, C. (2009). Reform of enterprise income tax, earnings management and its economic consequences. Economic Research Journal, (3), 86–98. (In Chinese). World Bank. (2006). China governance, investment climate, and harmonious society: Competi- tiveness enhancements for 120 cities in China (World Bank Report, No. 37759-CN). Wu, N. (2013). Business cycle, financial constraints and the dynamic collaborative choice of operating capital. Accounting Research, (8), 54–61. (In Chinese). Xia, L. (2003). Application of earnings management measuring models in the Chinese stock mar- ket. China Accounting and Finance Review, (6), 94–25. (In Chinese). Xia, L., & Chen, X. (2007). Marketization, SOE reform strategy, and endogenously determined corporate governance structure. Economic Research Journal, (8), 82–95. (In Chinese). Xie, B., Davidson, W. N., & Dasalt, P. J. (2003). Earnings management and corporate gover- nance: The role of the board and the audit committee. Journal of Corporate Finance, 9, 295–316. Xie, Z., & Lv, C. (2011). Research on the management equity incentive’s impact on earnings management. Public Finance Research, (6), 58–61. (In Chinese). Xu, H., & Chen, C. (2009). Accounting earnings quality, pricing of new shares and long term sustainment: Evidence from Chinese IPO market after institutional reform. Management World, (8), 25–38. (In Chinese). Ying, W. (2008). Empirical study on the impact of business cycle on corporate investment: Based on the perspective of sensitivity of investing cash flow. Public Finance Research, (5), 30–34. (In Chinese). Yuan, C., Zhou, H., Liu, W., & Liu, C. (2012). Empirical study on the pro-cyclicality of net capi- tal proportion in Chinese securities companies. Journal of Financial Research, (5), 70–79. (In Chinese). Zhang, W. (2010). Private placement of new shares and earnings management-empirical evidence from China’s Stock market. Management World, (1), 54–63. (In Chinese). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png China Journal of Accounting Studies Taylor & Francis

Earnings management, business cycle, and product market competition

Download PDF
22 pages

Loading next page...
 
/lp/taylor-francis/earnings-management-business-cycle-and-product-market-competition-Cf7FXj03WA

References (52)

Publisher
Taylor & Francis
Copyright
© 2015 Accounting Society of China
ISSN
2169-7221
eISSN
2169-7213
DOI
10.1080/21697213.2015.1023694
Publisher site
See Article on Publisher Site

Abstract

China Journal of Accounting Studies, 2015 Vol. 3, No. 2, 136–157, http://dx.doi.org/10.1080/21697213.2015.1023694 Earnings management, business cycle, and product market competition a b b Hongjian Wang *, Qingyuan Li and Yana Chen a b Department of Accounting, Jinan University, Guangzhou, China; Economics and Management School, Wuhan University, Wuhan, China Using annual data from non-financial listed companies in China between 2001 and 2011, this paper empirically examines the degree and key mechanism of upwards earnings management by companies under cyclical fluctuations of the macro- economy. We find that: firstly, upwards earnings management is significantly and positively correlated with the growth rate of the macro-economy, indicating a rela- tively strong pro-cyclical effect; secondly, the pro-cyclical effect of upwards earn- ings management will be more significant if the management expects corporate profit to be lower than the average profit for the industry; thirdly, the pro-cyclical effect of upwards earnings management will be more significant if the company faces greater pressure from product market competition. Further tests reveal that the greater the incentive for management, the greater the financing restrictions, and the more sensitive the industry is to the economic cyclical fluctuations, then the more significant is the pro-cyclical effect of upwards earnings management. The results demonstrate that the key mechanism of upwards earnings management during the economic boom is the expected profit based on the difference between the previous year’s expected firm’s profit and that year’s average profit for the industry, as well as the pressure of product market competition. Thus, the combination of the macro- economic state and the degree of product market competition will play an important role in understanding earnings management behaviour. Keywords: business cycle; earnings management; pro-cyclical effect; product mar- ket competition 1. Introduction According to the survey by the China Europe International Business School of 1,214 executives of Chinese and foreign companies in 2013, 46% of local companies and 37% of foreign companies refer to macro-economic adjustments as a major considera- tion (Fan, Xu, & Zhou, 2013). In particular, the outbreak and rapid spread of the global financial crisis in 2008 resulted in violent fluctuations in the Chinese economy, causing the economic growth rate to fall from 14.16% (adjusted) in 2007 to 9.63% (adjusted) in 2008. Many companies thus underwent operating difficulties. During macro- economic cyclical fluctuations, the researchers have consistently focused on macro- economic fundamental indices, such as growth rate of gross domestic product (GDP), unemployment rate, and consumer price index (CPI). However, theoretical economists *Corresponding author. Email: whj.hust@163.com Paper accepted by Donghua Chen. © 2015 Accounting Society of China China Journal of Accounting Studies 137 have largely overlooked how macro-economic cyclical changes affect changes in micro-corporate behaviour. Figure 1 describes the relationship between the growth rate of China’s GDP and the average profit margin of listed companies from 2000 to 2012. As shown in Figure 1, the average profit margin of listed companies is relatively consistent with the GDP growth rate, especially after 2005, when the reform of non-tradable shares was further implemented. The consistency between profit margins and macro-economic cyclical fluctuations demonstrates that macro-economic cyclical fluctuations have a relatively important impact on micro-enterprise behaviour. As a core topic in financial accounting, earnings management has elicited consider- able attention from scholars, and extensive research has been conducted on this topic. A conclusion that can be made is that research concentrates on the motivation of earn- ings management (Barua, Lin, & Sbaraglia, 2010; Cohen & Zarowin, 2010), the meth- ods and governance of earnings management (Bo & Wu, 2009; Klein, 2002; Xie, Davidson, & Dasalt, 2003), and the economic consequences of earnings management (Chen & Yuan, 2004; Coles, Hertzel, & Kalpathy, 2006; Haw, Qi, Wu, & Wu, 2005; Xu & Chen, 2009). Most of these studies are based on a company-level perspective and ignore the possible effects that the macro-economic state might have on companies’ earnings management. Existing research on business cycles is primarily based on the macro- economic perspective, with some studies concentrating on its effects on micro-corporate behaviour. Previous studies have covered topics including business cycle and capital structure (Erel, Julio, Kim, & Weisbach, 2012; Korajczyk & Levy, 2003); business cycle and corporate investment (Li & Luo, 2010; McLean & Zhao, 2014; Ying, 2008; Yuan, Zhou, Liu, & Liu, 2012); and business cycle and return on capital (Klein & Marquardt, 2006; Li, 2009). However, few scholars focus on how the business cycle affects earnings management of companies. As an important strategic decision made by the management of a company, earnings management is conducted under certain speci- fic macro-economic circumstances. Therefore, studying earnings management based only on the company-level perspective will lead to limitations on the general applicabil- ity of the research results. Figure 1. The relationship between GDP growth rate and the average margins of listed compa- nies in China from 2000 to 2012. Notes: The data of GDP is from the National Bureau of Statistics and the average rate of profit from the annual reports of non-financial listed companies in the sample. 138 Wang et al. To date, scholars such as Jin (2005), Cohen and Zarowin (2007), Strobl (2013), and Chen (2013b) have started to pay attention in their research to the effect of the macro- economy on earnings management. Nevertheless, their studies take developed market economic countries as their context. Using the annual data from non-financial listed companies in China between 2001 and 2011, this paper examines empirically the degree, and key mechanism of companies’ earnings management under cyclical fluctua- tions of the macro-economy. We report the following findings. First, upwards earnings management is significantly and positively correlated with the growth rate of the macro-economy, indicating a relatively strong pro-cyclical effect; Second, the pro- cyclical effect of upwards earnings management will be more significant if the manage- ment expects corporate profit to be lower than the average profit for the industry; Finally, the pro-cyclical effect of upwards earnings management will be more signifi- cant if the company faces greater product market competition pressure. Further tests reveal that financing restrictions have no significant impact on the pro-cyclical effect. The results demonstrate that the key mechanism of upwards earnings management dur- ing the economic boom is the expected profit (based on the difference between the previous year’s profit of the company and the same year’s average industrial profit), as well as the pressure from product market competition. Thus, the combination of the macro-economic state and the degree of product market competition will be important in understanding earnings management behaviour. The major contributions of this paper are as follows. First, based on the studies of Jin (2005), Cohen and Zarowin (2007), Strobl (2013), and Chen (2013b), this paper uses annual data from non-financial listed companies in China as a sample, examining the different effects of cyclical fluctuations on the degree of firms’ upwards earnings management. Second, combining the nature of property rights with Chinese characteris- tics and the institutional background of industry regulations, this paper reveals the pres- sure from product market competition to be a key transmission mechanism of companies’ earnings management under the flourishing period of the economy. This finding provides evidence for users of accounting information to judge the changes in the quality of companies’ financial statements with macro-economic cyclical fluctua- tions. Third, this paper also deepens understanding of the potential effects of interac- tions between the macro-economy and micro-corporate behaviour on research in accounting and finance; and enriches the literature on the effects of macro-economic policies on micro-corporate financial and accounting behaviour. 2. Theoretical analysis and testable hypotheses In recent years, the relationship between the macro-economy and micro-corporate beha- viour has elicited increasing interest from researchers. Specifically, the global financial crisis stemming from the subprime crisis of the United States in 2008 spread rapidly to all economies of the world, triggering scholarly consideration of the mechanism by which the macro-economy could affect micro-corporate behaviour. Studies by Jin (2005), Cohen and Zarowin (2007), and Strobl (2013) were based on the background of Western countries, whereas Chen’s research (2013) explored the degree of compa- nies’ earnings management in different phases of the business cycle from the perspec- tive of industrial cyclicality. Thus, their studies do not to consider the particular aspects of the Chinese economy and how the economic cycle affects the motivation for earn- ings management. China Journal of Accounting Studies 139 Given that China is in a transitional period, earnings management behaviour in Chinese listed companies is more severe because companies have weak property protec- tion systems, undesirable market mechanisms, weak auditor independence and serious relationship-based transactions, and the role of political connections remains important (Piotroski & Wong, 2012). Therefore, exploring how the business cycle affects the mechanism of earnings management will be greatly significant to our understanding of the effect of macro-economic cyclical fluctuations on micro-corporate behaviour. 2.1. Business cycle and companies’ earnings management According to Prospect Theory, people’s levels of sensitivity to ‘loss’ and ‘gain’ are asymmetrical, i.e., the pain from loss is much greater than the happiness from gain, with the decision-maker’s ‘loss aversion’ behaviour as a concrete manifestation (Kahneman & Tversky, 1979). Extending to the capital market, if companies announce bad earnings news when the market trend provides positive expectations to investors, market investors will be surprised and may in turn respond intensely. Thus, the stock price of companies will suffer severely in the stock market (Conrad, Cornell, Landsman, 2002) because the economy is in an upswing, and investors tend to expect positive future trends. Under this circumstance, investors will not respond as strongly if companies announce good earnings news, but will respond dramatically if companies publish bad earnings news (Conrad et al., 2002). From the perspective of executive compensation contracts, executive compensation is not only influenced by the operating performance of the year, but also affected by the stock price. Thus, as rational economic individuals, executives can anticipate the severity of consequences (Chen, 2013a). We suggest that management, as rational eco- nomic persons, will report earnings selectively to maximise their own benefit. When making selective reporting of operating performance at the end of an accounting period, executives will consider the general market environment and the profit status of indus- trial competitors and make corresponding adjustments to the operating performance (He, 2013; Tomy, 2012). Specifically, when the whole market and industrial competi- tors perform well but a particular company in the industry is the first to publish bad news, its stockholders will interpret this as inability or insufficiency of efforts by the management, and stockholders will probably assume that the lower-than-average perfor- mance of the company is a result of bad management. In the face of this threat, rational executives will be motivated to manipulate profits upwards in order to protect their own compensation and to avoid reporting profits that are lower than those of their com- petitors (Conrad et al., 2002; Tomy, 2012). Business cycle refers to the phenomena of alternation and circulation between eco- nomic expansion and economic contraction that periodically occur in an economic operation. A business cycle can be divided into four stages (recession, floor, expansion, and peak) or two phases, namely the ascending (boom) stage and descending (reces- sion) stage. However, the divisions of a business cycle mainly depend on GDP growth rate, as can be found in the studies of Liu (2009), Li (2009), and Jiang and Liu (2011). When the macro-economy is in an ascending (boom) stage, the GDP growth rate tends to be high and the economic prospect is expected to be positive. When the macro- economy is in a descending (recession) stage, the GDP growth rate tends to be low and the economic prospect is expected to be negative. In the boom stage of the business cycle, the general performance of listed compa- nies in the capital market tends be good, and stockholders have high expectations of 140 Wang et al. the operating performance of such companies. A particular company in the industry posting a loss or releasing earnings lower than the industrial average level is bad news that shocks market investors. When the number of companies in the industry is high and the product market competition is stronger, investors have more choices of invest- ment opportunity and will choose companies with better performance. Meanwhile, badly performing companies are identified more easily, which brings a greater shock to the stock prices of companies posting losses (Strobl, 2013). From the perspective of an executive compensation contract, executive compensa- tion not only depends on the increase in earnings of this year over last, but also consid- ers the operating performance of industrial competitors. Even though the company’s performance is higher than that of the previous year, if it is lower than the industrial average level, the company will be regarded as failing to maximise the interests of its shareholders. This result influences shareholders’ evaluation of the effort and ability of the management, which eventually affects executive compensation and even the posi- tion of the managers (Albuquerque, 2011; Dye, 1992; Gibbons & Murphy, 1990). Therefore, forced by the pressure of market competition, the management has stronger motivation for earnings manipulation. In the ascending stage of the macro-economy, positive earnings manipulation appears pro-cyclical. However, in the descending stage of the macro-economy, the general performance of listed companies shows a downward trend and, at this time, investors have pes- simistic expectations of the future trend. Meanwhile, from the perspective of an execu- tive compensation contract, economic recession can serve as an excuse for the bad performance of the company, persuading shareholders into believing that the bad per- formance cannot be attributed to the laziness or inability of the management. In this way, the management has a motivation to take a ‘big bath’ in accounting terms, so that they can hide profit for the boom stage and engage in earnings competition during the next boom. Besides, hiding profits at this time helps gain more government compensa- tion (Wang, Li, & Xing, 2014). But if the loss is not severe during the economic reces- sion, the company will have a stronger motivation to exaggerate profit and turn around (Chen, 2013a, 2013b; Strobl, 2013). Therefore, in the descending stage of the macro- economy, companies not only have a motivation for upwards earnings manipulation in order to turn around, but also have a motivation for downwards earnings manipulation, hiding earnings to engage in earnings competition in the next boom stage. In the ascending stage of the macro-economy, the upwards earnings management of companies will be affected by management’s expectation and the difference between a firm’s profit and the industrial average profit. Specifically, in the ascending stage of the macro-economy, if the management expects their company’s profit to be lower than the industrial average level, they will have a strong motivation for upwards earnings man- agement to avoid investors’ interpretation of bad performance as bad management, which will pose a threat to their compensation or could lead to dismissal. On the con- trary, if the management expects a firm’s profit to be higher than the industrial average, they will have weak motivation for manipulating the profit. Therefore, when the man- agement expects a firm’s profit to be lower than the industrial average level, the motivation for the pro-cyclicality of upwards earnings management will be stronger in the ascending stage of the economic cycle. H1: Upwards earnings management is positively associated with the ascending stage of the business cycle (and hence is more pro-cyclical in the ascending than in the descending stage of the cycle) when management expects lower profits than the industry average for the current year. China Journal of Accounting Studies 141 The business cycle is proxied by the rate of change of GDP. The management expecta- tion of lower profits for the current year is proxied by the observation that the previous year’s profit is lower than the previous year’s industrial average. 2.2. Product market competition, the boom stage of business cycle, and upwards earnings management of the companies Following on from H1, product market pressure is a factor that leads to management’s profit manipulation during different stages of the business cycle. Thus, the degree of upwards earnings management for companies facing different product market pressures will be different. The impact of product market pressure is based on information and constraint mechanisms. The information mechanism of the product market means that competing companies are correlated in their production and operation information, which weakens the degree of information asymmetry between managers and shareholders, thus constraining earn- ings management (Chen & Xu, 2011; Holmstrom, 1982). The constraint mechanism of the product market means that market competition increases the likelihood of losses or bankruptcy of companies, forcing managers to exert more effort to improve the operat- ing performance to avoid the consequences of bad performance, such as reduced com- pensation or dismissal (Chen & Xu, 2011). Markrian and Santalo (2010) demonstrate theoretically that only when stockholders are able to observe fully the operational state of the company (i.e., the market is com- pletely efficient), could product market competition constrain earnings management through an information mechanism; otherwise, it will constrain competition. Markrian and Santalo’s(2010) empirical test using a large sample has supported this conclusion. Further, the empirical test by Chen and Xu (2011), using a sample of Chinese compa- nies, found that the stronger the product market competition, the stronger the motiva- tion for earnings management by companies. Thus, the greater the product market pressure, the more significant the pro-cyclicality of upwards earnings management. If product market competition is a key mechanism affecting upwards earnings management, then the degree of upwards earnings manage- ment between state-owned and non-state-owned enterprises in China will differ in the boom stage. On the one hand, compared with state-owned enterprises, non-state-owned enterprises tend to be younger and relatively small in size, and are usually in the competitive product markets (He, 2013). On the other hand, state-owned enterprises enjoy more preferential policies and enter the regulated and monopoly industries more easily than do non state-owned enterprises (Fang, 2007; World Bank, 2006). This discussion motivates Hypothesis 2a, which is as follows: H2a: Compared with state-owned enterprises, non-state-owned enterprises tend to face greater market competition pressure and show more significant pro-cyclicality of upwards earnings management. The evidence for H2a is that, for non-state-owned enterprises, there is a stronger posi- tive association between upwards earnings management and the change of GDP in the ascending stage of the business cycle. If product market pressure is a key mechanism affecting upwards earnings manage- ment of companies, some differences in the significance of pro-cyclicality for upwards earnings management between companies with different degrees of regulation may 142 Wang et al. exist. The reason is that China is an emerging market in a transformational period and imposes more restrictions on ‘the industries of national security significance, natural monopoly industries, industries providing important public services and goods, pillar industries, and industries of high and new technology’ (Luo & Liu, 2009; Xia & Chen, 2007) as regulated industries, thus limiting the access of private capital to these industries, including mining (B), oils/chemicals/plastic cement/plastics (C4), metal and non-metal (C6), electric power/gas/hydraulic production and supply (D), transportation and warehousing (F), information technology (G), and media and cultural industry (L) (Luo & Liu, 2009; Xia & Chen, 2007). Regulated industries build barriers to access by new enterprises, thus limiting the degree of product market competition in these industries. Hence, industrial regulation divides Chinese listed companies into two categories of strong market competition and weak market competition. This distinction alleviates the problem of endogeneity, and helps to better capture product market competition as a key mechanism affecting upwards earnings management of companies in the ascending stage of the macro-economic cycle. According to the above discussion, Hypothesis 2b is as follows: H2b: Compared with companies in regulated industries, companies in non-regulated indus- tries face greater market competition and show more significant pro-cyclicality of upwards earnings management. The evidence for H2b is that, for companies in non-regulated industries, there is a stronger positive association between upwards earnings management and the change of GDP in the ascending stage of the business cycle. 3. Research design 3.1. Method of calculating earnings management This paper uses the modified Jones Model to calculate the discretionary accruals of Chinese listed companies. It is considered to produce a better outcome according to the research conclusions of Xia (2003) and Huang and Xia (2009). Moreover, based on the studies by Lei and Liu (2006), Wang, Wang, and Gong (2009), and Zhang (2010), this paper uses the KLW Model proposed by Kothari, Leone, and Wasley (2005) to conduct a robustness test. The results show that the conclusions of this paper have not materi- ally changed. The calculation process using the modified Jones Model is as follows. According to the modified Jones Model (Dechow, Sloan, & Sweeney, 1995), the OLS method is used to estimate α , α , α in Model (1) by sub-year and sub- 1 2 3 industry samples. The estimated regression coefficients are then substituted into Model (2) to calculate the non-discretionary accruals. The non-discretionary accruals estimated through Model (2) are then substituted into Model (3) to obtain the discretionary accruals. To eliminate the effect of outliers, all regression variables are winsorized at the first and 99th percentiles. Meanwhile, to guarantee the reliability of the regression equation, the samples with fewer than 15 firm-year observations have been deleted. TA = A ¼ a ð1 = A Þþ a ðDREV = A Þþ a ðPPE = A Þþ e (1) t t1 1 t1 2 t t1 3 t t1 t China Journal of Accounting Studies 143 NDA ¼ a ð1 = A Þþ a ðDREV = A  DREC = A Þþ a ðPPE = A Þ (2) t 1 t1 2 t t1 t t1 3 t t1 DA ¼ TA =A  NDA (3) t t t1 t where TA is the total accruals, which equals the operating profit(NT ) in year t minus t t the net operating cash flow in year t (CFO ). A is the total assets at the end of year t t–1 t–1, while ΔREV is the variation between prime operating revenues of year t and t–1 year t–1. PPE is the total original value of fixed assets at the end of year t. NDA is t t the non-discretionary accruals in year t after being adjusted by the total assets at the end of year t–1. ΔREC is the variation between account receivables of year t and t–1 year t–1. DA is the discretionary accruals of year t. 3.2. Regression model Given the research hypotheses, we consider the indicators that will affect earnings man- agement of companies and put forward the regression equation (4), as follows: þDiscretionary accruals ¼ b þ b Cycle GDP þ b  Control þ e (4) i;t i;t 0 1 i +Discretionary_accruals is the positive discretionary accruals calculated using the modified Jones Model. There are 8,252 observations of positive discretionary accruals, and the remainder are the negative discretionary accruals. Cycle_GDP is the proxy vari- able for a business cycle. We define the proxy variable for a business cycle with one year lagged data because we suppose that the decision on earnings management this year tends to depend on the information from the previous business cycle (i.e., the transmission from macro-economic cyclical fluctuation to micro-corporate behaviour has a certain lag). This process will also ease any possible endogeneity problem in the effect of macro-economic cyclical fluctuations on micro-corporate behaviour. As shown in Figure 2, compared with macro-economic cyclical fluctuations, earnings management of companies shows an obvious lag, which indirectly indicates that defining the proxy variable for a business cycle with one year lagged data is reasonable. For the definition of a business cycle, this paper draws on the studies by Li (2009), Jiang and Liu (2011), Strobl (2013), and Wu (2013) to use the year-on-year growth rate of GDP in China as the proxy variable for a business cycle (Cycle_GDP). This defini- tion is based on the following two considerations. (1) Using a dummy variable to Figure 2. Annual GDP growth rate and earnings management of firms from 2000 to 2012. Note: The data for annual GDP growth are from the National Bureau of Statistics. Earnings man- agement is calculated for the companies in the sample using the modified Jones Model and data from the CSMAR database. 144 Wang et al. represent boom and recession periods will probably produce errors because of subjec- tive judgment. Thus, we use the continuous variable of year-on-year growth rate of GDP to avoid this kind of error. (2) Using this continuous variable can effectively avoid the systematic impact of implementation of new accounting standards and reform of the shareholder structure resulting from using a dummy variable. Because the GDP growth rate in this paper is one year lagged, the non-lag data total 11 years (from 2000 to 2010). Control in regression equation (7) represents the control variables based on previous studies. The control variables include market-to-book ratio (Market_to_book_ratio), equity incentive (Equity_incentive), rights offering (Rights_offering, Prior_rights_offer- ing), public offering (Public_offering, Prior_public_offering), debt paying ability (Leverage), size of the company (Firm_size), net cash flow (Cash_flow), profitability (ROA), auditing opinion (Audit_opinion), and the competence of independent auditors (BIG4_auditor). We have three moderator variables. First, to explore the specific mechanism of the pro-cyclicality of upwards earnings management, we use the difference between the adjusted profit (adjusted by discretionary accruals) of the company in the previous year and the previous year’s average industrial adjusted profit (adjusted by discretionary accruals) of that year as an indicator of management’s expectation for the profitofthe current year. We then have two moderator variables to represent the pressure of product market competition. They are state ownership and regulated industry status. Finally, we set the industry dummy variables and control for the industry’s possible effect on the regression model to guarantee the reliability of fitting the model. Specific definitions of variables are shown in Table 1. 3.3. Data This paper chooses the annual data between 2001 and 2011 from the China Stock Mar- ket and Accounting Research (CSMAR) database as the research sample. Given that estimated discretionary accruals must be one year lagged, the GDP growth rates are from 2000 to 2010. We obtain all our data from the CSMAR database and the National Bureau of Statistics. Sample companies with unavailable data and operating in the finance and insurance industries are excluded from the study. We observed a total of 14,676 discretionary accruals, from which we eliminated the negative discretionary accruals, leaving us with 8,252 firm-year observations for positive discretionary accru- als in the regression model. To minimise the influence of outliers, we winsorize all con- tinuous variables at the first and 99th percentiles, except dummy variables, such as proxies for a macro-economic cycle. Table 2 provides a distribution of the sample by industry. Of the 19 industries, most observations come from more cyclical industries, such as petrochemicals, metals and non-metals, machinery, real estate, wholesale and retail, and information technology, accounting for over 50% of the total sample. The operating performances of the companies in these cyclical industries tend to be more sensitive to the fluctuations of the macro-economy, thereby offering greater confidence in the results. Table 3 is the annual distribution of the sample, which shows a rising trend. Table 4 presents the main variables of the correlation coefficient in this paper. The results demonstrate that the economic cycle is significantly positively correlated with upwards earnings management (+Discretionary_accruals), which means the economic China Journal of Accounting Studies 145 Table 1. Definitions of the variables. Types Variable Meaning Definition Dependent variable +Discretionary_accruals The value of positive The positive discretionary accruals discretionary accruals calculated using the modified Jones Model. Independent test variable Cycle_GDP Business cycle Drawing on Li (2009) and Jiang and Liu (2011), we use annual GDP growth rate as the proxy for the business cycle with one year lagged data. Independent moderator variables Earnings_difference Expectation of An indicator variable that takes on the management value of 1 if the difference between the company’s profit of the previous year and the previous year’s industrial average level is negative and 0 otherwise. State_ownership Nature of property right If the actual controller is the government, it is defined as state-owned enterprise and takes the value of 0, and 1 otherwise. Regulated_industy Industrial regulation An indicator variable that takes on the value of 0 if the company belongs to regulated industry and 1 otherwise. Control variables Market_to_book_ratio Market-to-book ratio Measured as the market value of the company divided by the book value. Loss Losses An indicator variable that takes on the value of 1 if the profit of this year is negative and 0 otherwise. Equity_incentive Equity incentive An indicator variable that takes on the value of 1 in the year before issuing an equity incentive plan, and 0 otherwise. Rights_offering Annual dummy variable An indicator variable that takes on the (rationed shares) value of 1 in the year of rights offering, and 0 otherwise. Prior_rights_offering Annual dummy variable An indicator variable that takes on the (prior to rights offering) value of 1 in the previous three years of conducting rights offering, and 0 otherwise. Public_offering Annual dummy variable An indicator variable that takes on the (issuing additional value of 1 in the year of public offering, shares) and 0 otherwise. Prio_additional_shares Annual dummy variable An indicator variable that takes on the (prior to issuing value of 1 in the previous three years of additional shares) public offering, and 0 otherwise. Leverage Financial leverage Measured as debt divided by total assets. Firm_size Firm size Measured as the natural log of total assets. Cash_flow Net cash flow Measured as net cash flow from operating activities divided by total assets. (Continued) 146 Wang et al. Table 1. (Continued). Types Variable Meaning Definition ROA Return on assets Measured as net income divided by total assets. Audit_opinion Auditing opinion An indicator variable that takes on the value of 1 for the standard auditing opinion, and 0 otherwise. BIG4_auditor Big four accounting A dummy variable that equals 1 if the firms listed company is audited by Big Four and 0 otherwise. Industry_dum Industry dummy Based on standard classification of variables industries in China, the manufacturing industry is classified to secondary level. We set one dummy variable for each industry to control for industry fixed effects. Table 2. Sample distribution of positive discretionary accruals by industry. Industry N Percentage Industry N Percentage Agriculture, forestry, livestock 131 1.59 Other manufacturing 31 0.38 farming, fishery Mining 188 2.28 Utilities 383 4.64 Food & beverage 337 4.08 Construction 136 1.65 Textiles & apparel 284 3.44 Transportation 336 4.07 Paper & printing 131 1.59 Information 7,499 6.05 technology Petrochemicals 802 9.72 Wholesale & retail 639 7.74 trade Electronics 372 4.51 Real estate 927 11.23 Metals & non-metals 715 8.66 Social services 301 3.65 Machinery 1, 402 16.99 Communication & 109 1.32 cultural Pharmaceuticals 529 6.41 Total 8, 252 100.00 Table 3. Annual distribution of sample of positive discretionary accruals. Year N Percentage Year N Percentage 2001 475 5.76 2007 764 9.26 2002 550 6.67 2008 872 10.57 2003 661 8.01 2009 881 10.68 2004 649 7.86 2010 929 11.26 2005 664 8.05 2011 1,041 12.62 2006 766 9.28 Total 8,252 100.00 cycle does have an impact on the earnings management behaviour of the firm. The rest of the correlation coefficients of control variables are below 0.5, showing that there is no serious effect of multicollinearity on the regression results. China Journal of Accounting Studies 147 Table 4. Correlation matrix for the main variables. 12 34 567 8910 11 12 13 14 1. +Discretionary_accruals 1.000 *** 2. Cycle_GDP 0.081 1.000 *** *** 3. Market_to_book_ratio 0.120 −0.052 1.000 4. Loss −0.032 0.002 0.006 1.000 *** *** *** *** 5. Equity_incentive 0.123 0.047 0.093 −0.065 1.000 *** *** 6. Rights_offering −0.011 −0.065 −0.012 −0.034 −0.004 1.000 *** *** *** 7. Prior_rights_offering −0.029 −0.030 0.013 −0.050 0.012 0.014 1.000 *** *** *** *** *** ** *** 8. Prior_public_offering 0.052 0.134 0.060 −0.057 0.052 −0.028 −0.035 1.000 *** *** *** *** *** *** 9. Public_offering 0.114 0.080 0.025 −0.069 0.093 −0.030 0.007 0.018 1.000 *** *** *** *** *** ** *** *** 10. Leverage 0.069 0.045 −0.067 0.272 −0.077 −0.026 −0.009 0.079 −0.034 1.000 *** *** *** *** ** *** *** *** 11. Firm_size 0.016 0.059 −0.371 −0.177 0.051 0.022 0.013 0.040 0.152 0.085 1.000 *** ** *** *** *** *** *** *** *** 12. Cash_flow −0.350 0.024 0.119 −0.249 0.041 0.000 0.043 0.054 −0.018 −0.268 0.115 1.000 *** *** *** *** *** *** *** *** *** *** *** 13.ROA 0.105 0.034 0.148 −0.685 0.109 0.013 0.052 0.082 0.073 −0.434 0.194 0.408 1.000 *** *** *** *** ** *** *** *** *** *** *** *** 14. Audit_opinion −0.006 −0.031 0.110 0.393 −0.046 −0.023 −0.035 −0.044 −0.063 0.355 −0.201 −0.171 −0.417 1.000 *** *** ** *** *** *** *** *** 15. BIG4_auditor 0.006 −0.006 −0.060 −0.054 0.027 −0.013 −0.005 −0.001 0.010 −0.042 0.271 0.086 0.079 −0.035 Note: The table above presents the regression coefficient for each independent variable. Refer to Table 1 for definition of variables. T-statistics are presented in the parentheses below the coefficients and are corrected using a two-dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. 148 Wang et al. 4. Empirical results 4.1. Descriptive statistics Table 5 provides descriptive statistics for our observations. The mean value (0.098) of positive discretionary accruals (+Discretionary_accruals) is slightly greater than the median value (0.066), which is slightly right skewed. The mean and the median value of the dummy proxy variable for the business cycle (Cycle_GDP) are approximately the same, and Cycle_GDP presents a normal distribu- tion. Among the control variables, the mean and the median values of market-to-book ratio (Market_to_book_ratio) are 2.189 and 1.726, respectively, which are slightly right skewed. Moreover, the mean value of Loss (Loss) is 0.110, indicating that approxi- mately 11.0% of the sample companies are making losses. The mean value of the equity incentive (Equity_incentive) is 0.045, which indicates that 4.5% of the sample companies offer a stock option incentive. The mean value of Rights_offering (Rights_offering) is 0.013, indicating that approximately 1.3% of the sample companies conduct rights offering during the year. The mean value of issuing additional stock in a public offering (Public_offering) is 0.063, which shows that 6.3% of the sample companies issue additional stock. The mean and the median values of the leverage ratio, size of company (Firm_size), and net cash flow (Cash_flow) are quite similar, demonstrating that these variables present normal distributions. Furthermore, the mean value (0.031) of return on assets (ROA) is less than the median value (0.033), which is slightly left skewed. The mean value of auditing opinion (Audit_opinion) is 0.070, indi- cating that 7.0% of the sample companies have non-standard audit opinions. The mean value of Big4 auditors (BIG4_auditor) is 0.064, which shows that 6.4% of the sample firms are audited by Big4 auditors. Table 5. Descriptive statistics of the main variables. Variables N Mean STD MIN 25% Median 75% MAX +Discretionary_accruals 8,252 0.098 0.100 0.001 0.029 0.066 0.132 0.492 Cycle_GDP 8,252 0.104 0.017 0.083 0.092 0.100 0.113 0.142 Earnings_difference 8,252 0.578 0.494 0.000 0.000 1.000 1.000 1.000 Sate_ownership 8,252 0.489 0.500 0.000 0.000 0.000 1.000 1.000 Regulated_ownership 8,252 0.633 0.482 0.000 0.000 1.000 1.000 1.000 Market_to_book_ratio 8,252 2.189 1.607 0.536 1.224 1.726 2.589 10.629 Loss 8,252 0.110 0.313 0.000 0.000 0.000 0.000 1.000 Equity_incentive 8,252 0.045 0.207 0.000 0.000 0.000 0.000 1.000 Rights_offering 8,252 0.013 0.114 0.000 0.000 0.000 0.000 1.000 Prior_rights_offering 8,252 0.020 0.139 0.000 0.000 0.000 0.000 1.000 Public_offering 8,252 0.063 0.243 0.000 0.000 0.000 0.000 1.000 Prior_public_offering 8,252 0.151 0.358 0.000 0.000 0.000 0.000 1.000 Leverage 8,252 0.519 0.237 0.056 0.369 0.511 0.645 2.026 Firm_size 8,252 21.555 1.104 18.823 20.803 21.446 22.165 25.102 Cash_flow 8,252 0.020 0.080 −0.211 −0.021 0.024 0.066 0.277 ROA 8,252 0.031 0.077 −0.419 0.012 0.033 0.061 0.206 Audit_opinion 8,252 0.070 0.255 0.000 0.000 0.000 0.000 1.000 BIG4_auditor 8,252 0.064 0.245 0.000 0.000 0.000 0.000 1.000 Note: Refer to Table 1 for definition of variables. China Journal of Accounting Studies 149 4.2. Main results Table 6 reports the regression results for the test of Hypothesis 1. The dependent vari- able is positive discretionary accruals. Table 6. Regression of upwards earnings management on business cycle and on the difference between the company’s profit of the previous year and the industrial average of that year. (1) (2) (3) Larger than or Less than zero equal to zero Interaction ** Cycle_GDP 0.399 0.103 0.005 (2.09) (0.67) (0.03) Earnings_difference −0.029 (–1.31) *** Earnings_difference *Cycle_GDP 0.562 (2.89) *** *** *** Market_to_book_ratio 0.008 0.007 0.007 (3.32) (6.26) (5.37) *** *** *** Loss 0.019 0.030 0.024 (2.64) (4.55) (4.18) *** *** *** Equity_incentive 0.077 0.017 0.041 (5.10) (3.38) (5.12) Rights_offering −0.002 0.008 −0.000 (–0.28) (1.20) (–0.03) Prior_rights_offering −0.006 0.000 −0.005 (–0.66) (0.03) (–0.77) *** ** *** Public_offering 0.015 0.006 0.009 (2.63) (2.04) (3.20) *** *** *** Prior_ public_offering 0.032 0.028 0.030 (5.15) (3.01) (7.51) ** ** Leverage 0.015 0.021 0.022 (1.05) (2.39) (2.24) *** *** Firm_size 0.006 0.003 0.006 (2.37) (1.45) (2.94) *** *** *** Cash_flow −0.432 −0.658 −0.546 (–8.27) (–17.19) (–12.37) *** *** *** ROA 0.354 0.505 0.421 (4.96) (8.08) (8.30) Audit_opinion −0.007 0.010 −0.001 (–0.95) (1.01) (–0.09) BIG4_auditor 0.017 0.000 0.008 (1.12) (0.04) (0.84) ** *** Constant −0.099 −0.030 −0.105 (–1.97) (–0.85) (–3.05) *** Coefficient difference test of 6.66 Cycle_GDP (0.01) N 3,712 4,540 8,252 Industry_dum Yes Yes Yes *** *** *** F value 40.14 47.45 80.97 R 0.290 0.383 0.328 Note: The dependent variable is positive discretionary accruals. The table presents the regression coefficient for each independent variable. Refer to Table 1 for the definition of variables. T-statistics are presented in parentheses below the coefficients and are corrected using a two dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. 150 Wang et al. Columns (1) and (2) test the results of (1) those cases where management’s expected profit (the difference between the previous year’s profit of the company and the previous year’s industrial average profit) is less than zero, and (2) those cases where management’s expected profit is greater than or equal to zero. Column (3) shows the regression results of defining this relationship as a dummy variable (Earnings_difference) and interacting Earnings_difference with the proxy of business cycle (Earnings_difference*Cycle_GDP). The dummy variable Earnings_difference takes on the value of 1 if the difference between the company’s profit of the previous year and the previous year’s industrial average level is negative and the value 0 other- wise. According to the regression results in Table 6, the coefficient of the proxy for business cycle (Cycle_GDP) is significantly positive in Column (1), but non-significant in Column (2). The difference test (SUE test) results of the coefficients of Cycle_GDP in Columns (1) and (2) are statistically significant at the 1% level. We use this differ- ence test to show that one set of coefficients is significantly larger than the other set of coefficients. Furthermore, in Column (3), the coefficient of the interaction term (Earn- ings_difference*Cycle_GDP) is significantly positive. These results support Hypothesis 1, that upwards earnings management is positively associated with the growth rate of GDP when the previous year’s profit of the company is lower than the previous year’s industrial average level (i.e. management’s expectations are for lower profits). Hence we can say that upwards earnings management is more pro-cyclical in the ascending stage of the business cycle, Tables 7 and 8 report the regression results for the test of Hypotheses 2a and 2b. When the previous year’s profit of a company is lower than the previous year’s indus- trial average level, the coefficients of the proxy of business cycle (Cycle_GDP) in the subgroups of non-state-owned enterprises and non-regulated industry are significantly positive, and the coefficient difference test is significant. The results of the difference test (SUE test) show that the coefficients of the proxy of business cycle (Cycle_GDP) in the subgroups of non-state-owned enterprises and non-regulated industry are significantly larger than those in the subgroups of state- owned enterprises and regulated industry. When the profit of the companies is higher than or equal to the industrial average level, the coefficients of the proxy of business cycle (Cycle_GDP) are non-significant in either group and the coefficient difference test (SUE test) is significant. In addition, the interaction terms: State ownership  Earnings difference  Cycle GDP and Regulated industry  Earnings difference  Cycle GDP are both significantly positive. Taken together, these results indicate that when compa- nies produce profits lower than the industrial average level and face greater market competition pressure, they tend to show more significant pro-cyclicality of upwards earnings management in the ascending stage of the business cycle, which further sup- ports Hypotheses 2a and 2b. 4.3. Further tests According to the studies of Cheng and WarFeId (2004) and Xie and Lv (2011), management incentive is the important driving force for management to conduct earn- ings management. Thus, executive incentive might be an important factor of upwards earnings management. This paper uses executive stockholding and management China Journal of Accounting Studies 151 Table 7. Nature of property rights, difference in profit, business cycle, and upwards earnings management. (1) (2) (3) Company’s profit< Company’s profit ≥ Industrial profit Industrial profit Total sample Non-state- Non-state- owned State-owned owned State-owned ** * Cycle_GDP 0.646 0.262 0.224 −0.042 −0.094 (2.30) (1.66) (1.21) (–0.25) (–0.63) ** State_ownership −0.022 (–2.33) Earnings_difference −0.013 (–1.15) ** State_ownership *Cycle_GDP 0.199 (2.18) *** Earnings_difference *Cycle_GDP 0.390 (3.85) State_ownership* −0.034 Earnings_difference (–1.42) State_ownership* 0.395 Earnings_difference*Cycle_GDP (1.89) *** *** *** *** *** Market_to_book_ratio 0.009 0.009 0.006 0.008 0.007 (2.65) (3.65) (5.52) (4.89) (5.53) * *** ** *** Loss 0.024 0.015 0.033 0.023 0.023 (1.87) (1.94) (3.36) (2.51) (4.18) *** *** *** *** Equity_incentive 0.056 0.106 0.025 −0.010 0.039 (5.04) (3.56) (3.83) (–1.03) (4.85) Rights_offering 0.006 0.004 0.004 0.018 0.001 (0.15) (0.63) (0.56) (1.57) (0.19) Prior_rights_offering −0.027 0.011 −0.006 0.013 −0.004 (–1.47) (0.73) (–1.07) (1.35) (–0.62) * ** ** *** Public_offering 0.018 0.014 0.012 0.003 0.010 (1.88) (1.97) (2.20) (0.87) (3.28) *** *** ** *** *** Prior_ public_offering 0.039 0.029 0.031 0.024 0.030 (3.02) (3.12) (2.53) (2.74) (7.42) ** ** Leverage 0.002 0.042 0.021 0.028 0.021 (0.08) (2.00) (1.42) (1.55) (2.14) * ** *** Firm_size 0.008 0.005 0.004 0.001 0.006 (1.87) (1.14) (2.31) (0.58) (3.10) *** *** *** *** *** Cash_flow −0.392 −0.482 −0.670 −0.650 −0.546 (–5.47) (–7.33) (–18.24) (–11.28) (–12.37) *** *** *** *** *** ROA 0.334 0.390 0.487 0.521 0.415 (4.07) (3.26) (6.29) (6.84) (8.17) ** Audit_opinion −0.024 0.008 0.008 0.013 −0.000 (–2.04) (1.23) (0.70) (0.98) (–0.08) BIG4_auditor 0.034 0.014 −0.004 0.006 0.008 (1.29) (0.76) (–0.32) (0.82) (0.84) Constant −0.151 −0.058 −0.048 0.016 −0.102*** (–0.99) (–0.77) (–1.51) (0.34) (–2.72) Industry_dum Yes Yes Yes Yes Yes ** * Coefficient difference test of 4.31 3.25 Cycle_GDP (0.04) (0.07) N 1,617 2,095 2,419 2,121 8,252 *** *** *** *** *** F value 16.85 30.35 26.70 22.30 73.08 R 0.283 0.320 0.367 0.416 0.331 Note: The table above presents the regression coefficient for each independent variable. Refer to Table 1 for definition of variables. T-statistics are presented in the parentheses below the coefficients and are corrected using a two dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. 152 Wang et al. Table 8. Regression of upwards earnings management on business cycle, regulated and non- regulated industry and higher and lower profit related to industry average. (1) (2) (3) Company’s profit Company’s profit < Industrial profit ≥ Industrial profit Non- Non- regulated Regulated regulated Regulated ** Cycle_GDP (1) 0.585 0.064 0.048 0.225 0.055 (2.32) (0.29) (0.28) (1.44) (0.26) ** Regulated_industry −0.026 (–2.02) Earnings_difference 0.001 (0.05) Regulated_industry*Cycle_GDP −0.080 (–0.59) Earnings_difference*Cycle_GDP 0.200 (0.72) Regulated_industry*Earnings_difference −0.046 (–1.87) ** Regulated_industry* 0.556 Earnings_difference*Cycle_GDP (2.39) *** * *** *** *** Market_to_book_ratio 0.009 0.006 0.008 0.004 0.007 (3.07) (1.66) (6.58) (2.57) (5.33) ** ** *** *** Loss 0.027 0.011 0.025 0.037 0.024 (2.34) (1.35) (2.53) (6.01) (4.06) *** *** ** ** *** Equity_incentive 0.053 0.107 0.017 0.014 0.041 (3.81) (4.13) (2.54) (2.34) (5.09) Rights_offering 0.002 −0.009 0.002 0.019 −0.000 (0.16) (–1.02) (0.33) (1.73) (–0.07) Prior_rights_offering −0.023 0.014 −0.006 0.008 −0.005 (–1.76) (0.78) (–0.57) (1.53) (–0.81) *** *** *** Public_offering 0.027 −0.002 0.002 0.012 0.010 (3.52) (–0.26) (0.50) (3.00) (3.20) *** * *** ** *** Prior_ public_offering 0.040 0.021 0.037 0.015 0.030 (3.86) (1.66) (3.22) (2.12) (7.54) *** ** Leverage 0.018 0.011 0.040 −0.012 0.022 (1.02) (0.47) (4.11) (–0.98) (2.28) *** ** *** Firm_size 0.012 −0.004 0.006 −0.003 0.005 (4.00) (–0.67) (2.32) (–1.44) (2.87) *** *** *** *** *** Cash_flow −0.440 −0.410 −0.688 −0.616 −0.546 (–7.75) (–4.54) (–12.87) (–13.73) (–12.39) *** *** *** *** *** ROA 0.372 0.321 0.453 0.577 0.420 (4.56) (4.23) (6.22) (9.30) (8.30) *** Audit_opinion −0.005 −0.014 0.008 0.013 −0.001 (–0.38) (–2.85) (0.61) (1.52) (–0.10) ** BIG4_auditor 0.003 0.047 −0.005 0.006 0.008 (0.17) (2.07) (–0.63) (0.89) (0.85) *** ** *** ** Constant −0.323 0.166 −0.114 0.120 −0.075 (–2.97) (1.51) (–2.00) (2.99) (–2.27) *** Coefficient difference test of 9.26 1.71 Cycle_GDP (0.002) (0.191) Industry_dum Yes Yes Yes Yes Yes N 2,388 1,324 2,832 1,708 8,252 F value 28.15 22.61 41.12 27.26 75.48 R 0.277 0.300 0.390 0.373 0.330 Note: The table above presents the regression coefficient for each independent variable. Refer to Table 1 for definition of variables. T-statistics are presented in the parentheses below the coefficients and are corrected using a two-dimensional cluster at firm and year level. ***, **, *represent significance at the 1%, 5%, and 10% level, respectively. China Journal of Accounting Studies 153 shareholding to represent the degree of management incentive, in order to explore macro-economic cyclical fluctuations’ impact on upwards earnings management under different degrees of management incentive. The untabulated regression results show that the coefficient of the interaction term (Earnings_difference*Cycle_GDP) in the sub- group of executive shareholding (coefficient 0.693, and T value is 2.60) and manage- ment shareholding (coefficient 0.642, and T value is 2.50) is greater than that in the subgroup without executive shareholding (coefficient 0.450, and T value is 2.97) and management shareholding (coefficient 0.424, and T value is 4.24). However the SUE test is not significant. We interpret these results as weakly indicating that management incentive is an important factor for upwards earnings management. A number of studies have demonstrated that financing need is an important motiva- tion for companies to conduct earnings management. During macro-economic cyclical fluctuations, the financing environment and financing needs of companies change. Thus, the degree of financial constraints of companies might become an important factor for the pro-cyclicality of upwards earnings management. From the perspective of financial constraint, this paper uses leverage ratio and size of the company as the proxies for financial constraints, and groups the data by the median value. Companies with high leverage ratio and small size are defined as companies with strong financial constraints; otherwise, they will be considered companies with weak financial constraints. The unre- ported regression results of the subgroups reveal that the coefficient of the interaction terms (Earnings_difference*Cycle_GDP) in the subgroup of high leverage (coefficient 0.635, and T value is 3.14) and small size (coefficient 0.777, and T value is 3.28) is greater than that in the subgroup of low leverage (coefficient 0.495, and T value is 1.58) and small size (coefficient 0.363, and T value is 2.39). However the difference test is not significant (SUE test). We interpret these findings as weakly indicating that financial constraint also is an important factor for upwards earnings management. According to the studies of Chen (2013b) and Li, Wang, and Rong (2013), the impact of degrees of macro-economic cyclical fluctuation on different industries is dif- ferent. This paper follows Kong (2010) and Li et al. (2013) to divide the listed compa- nies sample into two subgroups of industries with strong cyclicality and industries with weak cyclicality. Industries with weak cyclicality include agriculture, food and bever- age, medicine and biological products, medium and culture, and comprehensive indus- try. Industries with strong cyclicality are more sensitive to macro-economic fluctuations. The untabulated results of subgroup regression by industries with strong cyclicality and industries with weak cyclicality show that the coefficient of the interac- tion terms (Earnings_difference*Cycle_GDP) in the subgroup of strong cyclicality (co- efficient 0.508, and T value is 2.66) is more significant than that in the subgroup of weak cyclicality (coefficient 0.207, and T value is 2.80). We also find that the differ- ence test (SUE test) is significant. These results indicate that the pro-cyclicality of upwards earnings management under macro-economic cyclical fluctuations is more sig- nificant in industries with strong cyclicality, which is consistent with the findings of Chen (2013b) and Li et al. (2013). 4.4. Robustness tests We conduct additional sets of tests to assess the sensitivity of our results. We follow the studies of Jiang, Qu, Lu, and Li (2008) to use the Herfindhal Diversified Index (HHI index) based on operating income to measure the degree of market competition of the companies, and then obtain the median value by year. If the company’s HHI 154 Wang et al. index is larger than the median value of this year, then the company is facing lower market competition pressure; otherwise, the company is facing greater market competi- tion pressure. We conduct a new regression by using this variable to take the place of the nature of property right and industrial regulation. The unreported results show that the degree of sensitivity of macro-economic cyclical fluctuations has a different impact on upwards earnings management, which is manifested by the result that companies facing stronger market competition are more significant (compared with weak market competition). The table shows that the coefficient of the interaction term (Earnings_ difference*Cycle_GDP) is greater for strong market competition (0.569) than it is for weak market competition (0.429). However the difference test is not significant (SUE test). We interpret these findings as weakly indicating the pro-cyclicality of upwards earnings management. This result points towards the ‘market competition hypothesis’ of upwards earnings management under macro-economic cyclical fluctuations. 5. Conclusion As an important domain of financial and accounting research, earnings management has captured a great deal of attention from scholars. Existing studies are mainly based on the perspective of the micro-corporate level and are seldom based on the macro- economic level to study the effects on earnings management. However, decisions of companies are made under a given macro environment. Thus, focusing only on the micro level and ignoring the impact of the macro-economic environment limits the generalizability of the research results. Considering the reasons discussed above, this paper takes the annual data from non-financial listed companies in China between 2001 and 2011 as its sample to exam- ine empirically the degree, and key mechanism of companies’ earnings management under cyclical fluctuations of the macro-economy. We find the following: first, upwards earnings management is significantly and positively correlated with the growth rate of the macro-economy, indicating a relatively strong pro-cyclical effect. Secondly, the pro- cyclical effect of upwards earnings management will be more significant if the manage- ment expects corporate profit to be lower than the industrial profit. Thirdly, the pro-cyclical effect of upwards earnings management will be more significant if the company faces greater product market competition pressure. Further tests found that the greater the incentive of the management, the more sensitive the industry is to the eco- nomic cyclical fluctuations, and the more significant the pro-cyclical effect of upwards earnings management becomes. However, financing restrictions do not significantly affect the pro-cyclical effects. The results demonstrate that the key mechanism of upwards earnings management during the economic boom is the difference between the firm’s expected profit and industrial profit, as well as the product market competition pressure. Thus, the combination of the macro-economic state and the companies’ pro- duct market competition degree will play an important role in understanding earnings management behaviour. Admittedly, this paper is subject to some caveats. First, in terms of the definition of business cycle, although this paper uses annual GDP growth rate and divides the boom stage and recession stage by annual GDP growth rate, determining the stages of busi- ness cycles of China remains difficult. Second, because downwards earnings manage- ment has complex motivations and mechanisms, this paper focuses on the upwards earnings management under macro-economic cyclical fluctuations, which does not explain fully the comprehensive impacts of macro-economic cyclical fluctuations on China Journal of Accounting Studies 155 earnings management behaviour, thus requiring further studies to explore this question in the future. Acknowledgements We are most grateful to two anonymous referees and Liansheng Wu from Guanghua School of Management at Peking University for their valuable and constructive suggestions. We acknowl- edge the financial support from Key Projects of Philosophy and Social Sciences Research of Min- istry of Education (10JZD0019), National Natural Science Foundation of China (71072103; 71272228), Program for New Century Excellent Talents in University (NECT-120432) and Pro- gram for Young Scholars in Luojia. We are responsible for the consequences of this article. Notes 1. The reform of non-tradable shares is a process that allowed previously non-tradable shares held by controlling shareholders to be freely tradable on the exchanges. 2. Seemingly Unrelated Estimation, statistical package: http://www.stata.com/manuals13/rsuest. pdf. References Albuquerque, Ana. M. (2011). Do growth- option firms use less relative performance evaluation? Working Paper. Boston University. Barua, A., Lin, S., & Sbaraglia, A. M. (2010). Earnings management using discontinued opera- tions. The Accounting Review, 85, 1485–1509. Bo, X., & Wu, L. (2009). The governance roles of state-owned controlling and institutional investors: A perspective of earnings management. Economic Research Journal, (2), 81–90. (In Chinese). Chen, W. (2013a). Business cycle, industry prosperity and earnings management-empirical evidence from China’s listed companies. Journal of Auditing Research, (5), 96–105. (In Chinese). Chen, W. (2013b). Business cycles, cyclical industries and magnitude of earnings management: Empirical evidence from China’s listed companies. Nankai Business Review, (3), 26–35. (In Chinese). Chen, J., & Xu, Y. (2011). Product market competition, competitive situation and earnings man- agement of listed companies. Public Finance Research, (4), 58–61. (In Chinese). Chen, K., & Yuan, H. (2004). Earnings management and capital resource allocation: Evidence from China’s accounting-based regulation of rights issue. The Accounting Review, 79, 645–665. Cheng, Q., & WarFeId, T. D. (2004). Equity incentives and earnings management. The Account- ing Review, 80, 441–476. Cohen, D., & Zarowin, P. (2007). Earning management over the business cycle. Working paper. Stern School of Business of New York University. Cohen, D., & Zarowin, P. (2010). Accrual-based and real earnings management activities around seasoned equity offerings. Journal of Accounting and Economics, 50,2–19. Coles, J., Hertzel, M., & Kalpathy, S. (2006). Earning management around employee stock option reissues. Journal of Accounting and Economics, 41, 173–200. Conrad, J., Cornell, B., & Landsman, W. (2002). When is bad news really bad news? Journal of Finance, 57, 2507–2532. Dechow, P., Sloan, R., & Sweeney, A. (1995). Detecting earnings management. The Accounting Review, 70, 193–225. Dye, R. (1992). Relative performance evaluation and project selection. Journal of Accounting Research, 30,27–52. Erel, I., Julio, B., Kim, W., & Weisbach, M. (2012). Macroeconomic condition and capital rising. Review of Financial Studies, 25, 341–376. 156 Wang et al. Fan, Y., Xu, B., & Zhou, D. (2013). China business report of 2013. China Europe International Business School (in Chinese). Fang, J. (2007). Ownership, institutional environment and capital allocation. Economic Research Journal, (12), 82–92. (In Chinese). Gibbons, R.., & Murphy, K. J. (1990). Relative performance evaluation for Chief Executive Offi- cers. Industrial and Labour Relations Review, 43, 30S–51S. Hao, S., Jin, Q., & Zhang, G. (2011). Relative firm profitability and stock price sensitivity to aggregate news. The Accounting Review, 86, 1321–1347. Haw, I., Qi, D., Wu, D., & Wu, W. (2005). Market consequences of earnings management in response to security regulations in China. Contemporary Accounting Research, 22,95–143. He, W. (2013). Promotion incentives and earnings management journal of China soft science (in Chinese). Holmstrom, B. (1982). Moral hazard in teams. The Bell Journal of Economics, 13, 324–340. Huang, M., & Xia, X. (2009). An evaluation on specification and power of discretionary accruals models in the Chinese context. Nankai Business Review, (5), 136–147. (In Chinese). Jiang, L., & Liu, X. (2011). Research on economic cyclical fluctuations and cash holding behav- iors of listed companies. Accounting Research, (9), 40–46. (In Chinese). Jiang, F., Qu, Y., Lu, Z., & Li, Y. (2008). Product market competition and dynamic capital struc- ture adjustments. Economic Research Journal, (4), 99–110. (In Chinese). Jin, Q. (2005). Business cycle, accounting behavior and earnings management. Doctoral Thesis. Hong Kong University of Science and Technology. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–292. Klein, A. (2002). Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics, 33, 375–400. Klein, A., & Marquardt, C. A. (2006). Fundamentals of accounting losses. The Accounting Review, 81, 179–206. Kong, Q. (2010). Macroeconomic fluctuations, cyclical industry and capital structure choice. Journal of Beijing Institute of Technology, (6), 31–35. (In Chinese). Korajczyk, R. A., & Levy, A. (2003). Capital structure choice: Macroeconomic conditions and financial constrained. Journal of Financial Economics, 68,75–109. Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionary accruals. Journal of Accounting and Economics, 39, 163–197. Lei, G., & Liu, H. (2006). Between big shareholders’ control, financing scale and the degree of manipulating surplus. Management World, (1), 129–136. (In Chinese). Li, Y. (2009). Why does the operating performance of the Chinese listed companies deviate from the business cycle? Economic Research Journal, (3), 99–109. (In Chinese). Li, W. & Luo, M. (2010). An empirical study of pro-cyclicality of the capital adequacy ratio of commercial banks in China. Journal of Financial Research, (2), 147–157. (In Chinese). Li, Q., Wang, H., & Rong, X. (2013). Firm earnings persistence over the business cycle: Evi- dence from listed companies in china. China Accounting and Finance Review, (4), 129–191. (In Chinese). Liu, S. (2009). Retrospect and prospect of the 60 years of economic growth of China, and the discussion of the new round of business cycle. Economic Perspectives, 10,3–10. (In Chinese). Luo, D., & Liu, X. (2009). Political connection, entry barrier and corporate performance: Evi- dence from private listed companies. Management World, (5), 97–106. (In Chinese). Markrian, G., & Santalo, J. (2010). Product market competition, information and earnings man- agement. Working paper. IE Business School. McLean, D., & Zhao, M. (2014). The business cycle, investor sentiment, and costly external finance. Journal of Finance, 69, 1377–1409. Piotroski, J. D., & Wong, T. J. (2012). Institutions and information environment of Chinese listed firms. Working Paper. Stanford University, the Chinese University of Hong Kong. Strobl, G. (2013). Earnings manipulation and the cost of capital. Journal of Accounting Research, 51, 449–573. Tomy, R. E. (2012). Earnings persistence over the business cycle. Working paper. Stanford University. China Journal of Accounting Studies 157 Wang, H., Li, Q., & Xing, F. (2014). Financial crisis, government subsidies, and earnings manip- ulation-The empirical evidence from listed companies. Management World, (7), 157–167. (In Chinese). Wang, Y., Wang, L., & Gong, C. (2009). Reform of enterprise income tax, earnings management and its economic consequences. Economic Research Journal, (3), 86–98. (In Chinese). World Bank. (2006). China governance, investment climate, and harmonious society: Competi- tiveness enhancements for 120 cities in China (World Bank Report, No. 37759-CN). Wu, N. (2013). Business cycle, financial constraints and the dynamic collaborative choice of operating capital. Accounting Research, (8), 54–61. (In Chinese). Xia, L. (2003). Application of earnings management measuring models in the Chinese stock mar- ket. China Accounting and Finance Review, (6), 94–25. (In Chinese). Xia, L., & Chen, X. (2007). Marketization, SOE reform strategy, and endogenously determined corporate governance structure. Economic Research Journal, (8), 82–95. (In Chinese). Xie, B., Davidson, W. N., & Dasalt, P. J. (2003). Earnings management and corporate gover- nance: The role of the board and the audit committee. Journal of Corporate Finance, 9, 295–316. Xie, Z., & Lv, C. (2011). Research on the management equity incentive’s impact on earnings management. Public Finance Research, (6), 58–61. (In Chinese). Xu, H., & Chen, C. (2009). Accounting earnings quality, pricing of new shares and long term sustainment: Evidence from Chinese IPO market after institutional reform. Management World, (8), 25–38. (In Chinese). Ying, W. (2008). Empirical study on the impact of business cycle on corporate investment: Based on the perspective of sensitivity of investing cash flow. Public Finance Research, (5), 30–34. (In Chinese). Yuan, C., Zhou, H., Liu, W., & Liu, C. (2012). Empirical study on the pro-cyclicality of net capi- tal proportion in Chinese securities companies. Journal of Financial Research, (5), 70–79. (In Chinese). Zhang, W. (2010). Private placement of new shares and earnings management-empirical evidence from China’s Stock market. Management World, (1), 54–63. (In Chinese).

Journal

China Journal of Accounting Studies – Taylor & Francis

Published: Apr 3, 2015

Keywords: business cycle; earnings management; pro-cyclical effect; product market competition

There are no references for this article.