DeepDyve requires Javascript to function. Please enable Javascript on your browser to continue.
Listing approach, political favours and earnings quality: Evidence from Chinese family firms Listing approach, political favours and earnings quality: Evidence from Chinese family firms
Huang, Qiongyu; Cheng, Minying; Li, Wenjing; Wei, Minghai
2014-01-02 00:00:00
China Journal of Accounting Studies, 2014 Vol. 2, No. 1, 13–36, http://dx.doi.org/10.1080/21697221.2014.880167 Listing approach, political favours and earnings quality: Evidence from Chinese family firms a b c b Qiongyu Huang , Minying Cheng *, Wenjing Li and Minghai Wei School of Economics and Statistics, Lingnan Statistics Science Research Center, Guangzhou University, People’s Republic of China; Business School, Sun Yat-Sen University, People’s Republic of China; School of Management, Jinan University, People’s Republic of China The listing approach taken by Chinese family firms relates to the political favours that a family firm can obtain from local government, which could have a significant impact on corporate financial reporting behaviour. Using a sample of Chinese family firms over the period from 2003 to 2008, we find that family firms going public through initial public offerings (IPOs) obtain more bank loans, more subsidies and lower tax rates than firms going public through a takeover. Further investigation of the relation between listing approach and the quality of financial reporting shows that the earnings quality is systematically worse for family firms going public through IPOs than those going public through a takeover. Subsample regressions reveal that, in IPO firms, the magnitude of political favours obtained is negatively related to the quality of financial reporting. Keywords: earnings quality; family firms; information effect; listing approach; political favours; resource effect 1. Introduction In the past decade, literature has highlighted the effects of family control on earnings quality. It has been shown that, because of the alignment effect of family control, family ownership is associated with better earnings quality (Ali, Chen, & Radhakrishnan, 2007; Wang, 2006). However, Chaney, Faccio, and David (2011) find a negative relation between political connection and earnings quality. Because of a lesser need to respond to market pressures to increase information quality, politically connected companies can afford lower earnings quality. According to these findings, family ownership and political connections have opposite effects on earnings quality. It is interesting to examine whether the effects of political connection will offset the alignment effect of family ownership in a setting where political connection is very important. In this paper, we use the listing approach of family firms in China as a natural setting to examine how political connections affect the earnings information quality in family firms. The listing approach chosen by family firms is of particular importance in understand- ing the role of political favours and disclosure strategies in family firms, because it indicates the origin and background of a family firm in China as well as the resource that a family firm can obtain from the government and the market, which will in turn affect the firm’s incentives to supply higher or lower earnings information quality. There are *Corresponding author. Email: mn03chmy@mail2.sysu.edu.cn Paper accepted by Liansheng Wu. © 2014 Accounting Society of China 14 Huang et al. two main ways for family firms to become listed on the Chinese stock market. One is through initial public offerings (IPOs). The other is taking over a public company listed in the stock market (takeover). Choosing between an IPO and a takeover as a route to public listing is a complicated trade-off among multiple benefits and costs (Brau, Francis, & Kohers, 2003). However, such benefits and costs have a significant impact on the incentives to improve earnings quality. One of the benefits in China is the local political favours a firm can obtain (Du & Xu, 2009). Because local government is willing to inspire more firms to obtain public listings through IPOs in order to enhance the particular local government’s political performance (Du & Xu, 2009), family firms that choose to go public through an IPO usually receive additional support from local government as compensation. As suggested in Chaney et al. (2011), politically connected firms provide earnings information of lower quality because they have a lesser need to respond to mar- ket pressures. We ask: if such were the case for the politically favoured firms in China, would family firms provide poorer or better earnings information when they could obtain more political favours from local government? To find out whether there is a systematic difference in earnings quality between family firms that went public through IPOs (IPO firms) and family firms that went pub- lic through takeovers (takeover firms), we first examine whether more political favours could make IPO firms more reluctant to increase earnings quality. We find that IPO firms enjoy significantly more bank loans, more government subsidies and lower effec- tive tax-rates than takeover firms, suggesting that IPO firms can obtain more political favours than do takeover firms. Second, we examine the relation between listing approach and earnings quality. We find that the earnings quality is systematically lower in IPO firms than in takeover firms. IPO firms have more absolute abnormal accruals, lower earnings informativeness and higher persistence of transitory loss components in earnings. Finally, in the subsample of IPO firms, we find that political favours are nega- tively associated with earnings quality. These findings are robust to alternative specifications even after we control for the potential endogeneity that drives a firm to go public via an IPO. It suggests that IPO firms obtain more political favours than takeover firms, which reduces the incentives of those IPO firms to provide high earn- ings quality as a family firm. Our findings make several contributions to the literature. First, in the research on financial reports in family firms, prior literature focuses on the alignment effect and the entrenchment effect on earnings quality, driven by family ownership and family management (Ali et al., 2007; Chen, Chen, & Cheng, 2008; Wang, 2006). However, the theory and evidence from US family firms may not apply to family firms in other countries because of institutional differences (Ali et al., 2007). We propose that the effect derived from political favours is even more important for shaping earnings quality in a country with strong political intervention. Such an effect might override the effects of ownership structure and agency problems on the financial reporting behaviour of family firms. Second, it enriches the literature on the consequence of political connections and the mechanism through which political connections affect earnings quality. Prior studies find that investors in civil law countries suffer lower information quality (Bushman & Piotroski, 2006; Bushman, Piotroski, & Smith, 2004), and top executive political ties are correlated to opacity (Chaney et al., 2011). We use the listing approach of family firms in a relationship-based economy as a special setting to examine why politically supported firms provide earnings information of lower quality. We provide evidence China Journal of Accounting Studies 15 that local political favours enjoyed by IPO firms lower their incentives to improve information quality. Third, our research has strong policy implications. The effects of the listing approach on earnings quality stem from favours by local government. It is the govern- mental officials who interfere in the allocation of economic resources. Such political interference affects the corporate financial reporting behaviour, and hence the informa- tion transparency in the stock market. Our evidence implies that policy makers should take this into account when deciding whether the government should intervene, and how to intervene, in the resource allocation in an emerging market. The rest of this paper is organised as follows: Section 2 reviews prior studies. Sec- tion 3 develops the hypotheses. Section 4 describes the research design. Sections 5 and 6 report the empirical results and robustness check. Section 7 concludes the paper. 2. Literature review Corporate information qualities are substantially influenced by family control. On the one hand, family control may decrease the information transparency in family firms (entrenchment effect). Because controlling shareholders have the incentives and abilities to exploit minority shareholders (Shleifer & Vishny, 1997), they also have incentives to cover expropriation from minority shareholders. Evidence shows that, in order to hide such exploitation from minority shareholders, owners of family firms are inclined to provide financial reports with lower earnings quality, such as lower value relevance (Fan & Wong, 2002) and less earnings conservatism (Bona-Sánchez, Pérez-Alemán, & Santana-Martín, 2011). On the other hand, the presence of controlling shareholders may have positive effects on information disclosure (the alignment effect). By appointing a founding fam- ily member as CEO of the firm, the combination of family control and family manage- ment aligns the interests of shareholders and managers (Anderson & Reeb, 2003). Also, long-term orientation and reputation protection encourage family firms to focus on long-term interests. Thus, family firms are likely to use high quality financial reports to communicate with outside investors so as to lower their costs of debt (Anderson, Mansi, & Reeb, 2003). Wang (2006) finds that founding family ownership is related to higher quality of earnings information, suggesting that the alignment effect plays a dominating role in the financial reporting practice in family firms (Ali et al., 2007). Apart from corporate ownership structures and corporate governance, there are other factors that influence family firms’ reporting behaviour. Among them, political connec- tion is one of the most important factors, especially in an environment characterised by strong political intervention, which has not received sufficient research attention until recent years. Fan, Wong, & Zhang (2012) report an increase in earnings quality of family firms subsequent to succession, which is attributed to the loss of the entrepre- neur’s reputation and political/social networks. Specifically, favours from politicians and bureaucrats in secret are important proprietary knowledge and specific capital among firms that engage in political rent-seeking activities. Such proprietary knowledge and specific human capital is associated with opacity and low informativeness of accounting earnings (Fan & Wong, 2002). These findings predict that politically con- nected firms tend to limit their information disclosure to the public, so that they can reduce the leakage of proprietary information to the public and potential competitors, which implies a negative relation between political connections and accounting infor- mation quality. 16 Huang et al. 3. Hypothesis development Family firms have the choice of going public via an IPO or a takeover. Taking over a publicly traded company is often an attractive opportunity for private firms and presents an alternative to the IPO route (Brau et al., 2003), because firms could encounter sig- nificant costs in an IPO, such as listing fees (Foucault & Parlour, 2004) and underpric- ing (Booth & Chua, 1996). This is especially true for Chinese family firms, because state-owned enterprises (SOEs) enjoy preferential access to the capital market in China. In the transition economy and developing capital market in China, the IPO quota is a scarce resource under the control of the government (Francis, Hasan, & Sun, 2009; Jiang, Liang, & Chen, 2009). The China Securities Regulation Commission (CSRC) has maintained consistent control over deciding which company is permitted to under- take an IPO (Jiang et al., 2009). The natural connections between SOEs and the government institutions facilitate bias towards SOEs when the CSRC distributes the IPO quotas (Chen, Lee, & Li, 2008). Moreover, the initial purpose of developing the Chinese stock market has been to solve the financial difficulties in SOEs since the early 1990s (Wang, Xu, & Zhu, 2004). Hence, SOEs enjoy preferential access to the capital market, while it is quite difficult for family firms to obtain the opportunity to conduct an IPO. As such, taking over a listed firm becomes an alternative for family firms to raise finance in the Chinese stock market. Given the difficulties for family firms to go public via IPOs, family firms that choose an IPO may receive favours from external institutions such as the local govern- ment, as support or compensation. Du and Xu (2009) document that various Chinese local governments devote themselves to promoting local companies to go public, so that substantial stock market investment funds can be channelled into potentially pro- ductive companies. Compared with family takeover firms that need to go public via taking over a poorly-performing firm, family IPO firms that can go through the strict IPO process are usually the market leaders in a city or in a province (Du & Xu, 2007). They are better at boosting local economic growth, which is very important for local governmental officials in their tournament competition and political career (Zhu, 2004). In an economy where it is so difficult to push a local firm to conduct IPOs, IPO firms not only bring about GDP and tax growth, but also enhance the reputation of that par- ticular local government (Aharony, Lee, & Wong, 2000). Thus, local officials are likely to provide additional support for their business. There may be concerns that takeover firms are firms rescued by the government, so that they are also likely to receive more support from the government. However, support such as a bailout is usually given before a firm is sold, so that the firm in financial distress could be free from a takeover (Faccio, Masulis, McConnell, & Offenberg, 2006). But takeover firms do not necessar- ily obtain support after they are listed, especial when they are family firms locating in another district, which is quite common in the takeovers (Sun & Luo, 2011). To exam- ine whether IPO firms have stronger political favours from local government, which drive them to lower their financial reporting quality, we propose Hypothesis 1 as follows. H1. Family firms that went public through IPOs obtained more political benefits than those that went public through takeovers. With the political favours offered by local government, the incentives of IPO firms to provide high quality of earnings information could be reduced, which can be caused by the following effects. China Journal of Accounting Studies 17 The first is the information effect. Family firms have the incentives to hide their reliance on the political favours received from local government by providing less transparent accounting information. Political favours include preferential access to loans (Claessens, Feijen, & Laeven, 2008; Khwaja & Mian, 2005), longer term credit (Charumilind, Kali, & Wiwattanakantang, 2006), and a higher likelihood of being bailed out (Faccio et al., 2006). However, firms that count on political favours to become competitive or preserve their current position may suggest a negative signal to the market: the firms lack core competitiveness and the political favours could disap- pear anytime (Fan, Wong, & Zhang, 2007). To avoid an uncertain impact on the stock price and the commercial market, IPO firms that obtain political favours have incen- tives to hide their dependence on these connections by providing less transparent finan- cial reports. In addition, IPO firms are willing to mask their reliance on political benefits even if the market considers it to be a good signal that family firms obtain additional favours from the local government. China is a country characterised as a relationship-based economy and governed by strong political intervention in economic activities (Allen, Qian, & Qian, 2005). Favours from politicians and bureaucrats are one of the important specific capitals in business (Faccio, 2006). To prevent the leakage of proprietary infor- mation to competitors (Fan & Wong, 2002), family-specialised resources such as the entrepreneur’s reputation and political/social networks are associated with lower earn- ings quality (Fan et al., 2012). Hence, IPO firms tend to invite opacity as a strategy that allows firms to avoid unwanted political or social scrutiny. The second effect is the resource effect. On the one hand, compared with family firms that receive fewer political favours, IPO firms have less incentive to provide high-quality financial reports, because politically favoured firms can obtain an external resource much more easily than other firms with the help of government (Leuz & Oberholzer-Gee, 2006), especially in a market that experiences extensive intervention by the government. For example, politically connected firms have a lower cost of debt (Sapienza, 2004) and can provide less collateral to borrow money (Charumilind et al., 2006). So they might have less need to compete for resources by providing high-quality financial reports. Chaney et al. (2011) provide evidence that firms with politically related large sharehold- ers, or top directors, provide poor quality of accruals, but they are not penalised by higher cost of debt. As such, IPO firms are less willing to improve earnings quality when they are looking for an external resource if they can get help from local government. On the other hand, takeover firms have the incentive to improve information trans- parency to reduce information asymmetry. Compared with IPO firms, takeover firms do not enjoy preferential access to political favours. But they have the demand to attract external investors or to obtain external capital at a lower cost. More importantly, take- over firms have great incentives to conduct the seasoned equity offerings (SEOs) for funding. Because, compared with IPO firms, takeover firms do not have the chance to finance via an IPO. They rely significantly on SEOs to obtain external capital. Opacity hampers firms from raising finance in the capital market. Leuz and Oberholzer-Gee (2006) find that non-connected firms are more inclined to seek global financing than well-connected firms, implying that firms lacking political connections may have to seek other costly options in order to relax capital constraints. Family firms that acquire a public company can skip over the procedures in the pre-IPO stage. Thus, takeover firms do not provide as much information as the IPO firms when they go public. In addition, the disclosure requirements for acquirers are much simpler than IPO firms. 18 Huang et al. To increase liquidity and acquire the opportunities to raise external finance, takeover firms have the incentive to provide higher quality of information. In sum, we propose Hypothesis 2 as follows. H2. Family firms that went public through IPOs provided lower quality of earnings than those that went public through takeovers. To examine whether the lower earnings quality of IPO firms stems from their stronger political favours, we propose Hypotheses 3 as follows. H3. Among family firms that went public through IPOs, those obtaining more political favours provided lower quality of earnings. 4. Research design 4.1. Sample and data Our initial sample consists of family firms listed on the Shanghai Stock Exchange and the Shenzhen Stock Exchange from 2003 to 2008. The sample period starts from 2003 because disclosure of the ultimate owners has been mandated by CSRC since 2003, and we manually collected the listing approach and family ownership data until 2008. Following Wei, Wu, Li, and Chen (2011), a firm is considered as a family firm if its ulti- mate controller can be traced as a person or a family. We classify family firms into two categories according to the way a firm went public, i.e. IPO firms and takeover firms. If an IPO firm completed its IPO in or after 2003, it is included in the sample from the year when it became listed. Similarly, if a takeover firm completed the takeover in or after 2003, it is included in the sample from the date when the family finished acquisi- tion of the prior listed firm. Data on family ownership and CEO attributes (whether the CEO is a family member or a professional manager hired from the labour market) are manually collected from the IPO prospectuses and annual reports. Information on the background of top executives is also collected by hand from annual reports. For cases with ambiguous information, we searched Google as a cross-check. Financial data are collected from the Wind database, and data on stock returns and prices are collected from the China Stock Market and Accounting Research (CSMAR) database. We elimi- nate an observation if (1) it is a financial institution and (2) it has insufficient data for estimating earnings quality measures. Eventually, we obtain a total of 2,492 observations taken from 486 unique firms for primary tests. To avoid the influence of outliers, we winsorize each continuous variable at the 1st and 99th percentile of their distribution. Table 1 presents the sample distribution. The sample includes 1121 IPO observations and 1371 takeover observations. Comparing the number of newly listed firms, we can see that the number of family firms going public through an IPO increased, while the number of firms going public through a takeover decreased over time. 4.2. Empirical models To test H1, we employ bank loans, subsidies and effective tax rates as explanatory variables to obtain evidence on political favours derived by IPO firms (Adhikari, Deras- hid, & Zhang, 2006; Charumilind et al., 2006; Claessens et al., 2008; Faccio et al., 2006; Khwaja and Mian, 2005; Wu, Wu, & Rui, 2010). To test H2 and H3, we employ abnormal accruals, earnings informativeness, and the persistence of transitory loss components in earnings as the measures of accounting information quality following China Journal of Accounting Studies 19 Table 1. Sample description. Number of family firms Number of newly listed firms Year Listway=1 (%) Listway=0 Total Listway=1 (%) Listway=0 Total 2003 97 (36.09) 169 266 20 (22.03) 54 74 2004 144 (41.08) 209 353 44 (47.31) 49 93 2005 149 (40.55) 216 365 7 (18.92) 30 37 2006 180 (41.49) 249 429 30 (38.46) 48 78 2007 251 (49.02) 261 512 74 (62.18) 45 119 2008 300 (52.99) 267 567 59 (69.41) 26 85 Total 1121 (44.90) 1371 2492 234 (48.15) 252 486 Note: Listway is the approach for firms to go public, which is equal to 1 if a firm goes public by IPOs, and equal to 0 if a firm goes to public through takeover. the majority of earnings quality literature (e.g. Ball & Shivakumar, 2005; Basu, 1997; Dechow & Dichev, 2002; Fan & Wong, 2002; Francis, LaFond, Olsson, & Schipper, 2005; Wang, 2006). Ordinary least squares (OLS) regressions are used in the following analysis. T-statistics in the pooled regressions are based on robust standard errors clus- tered at the firm level to diminish the potential heteroscedasticity and the firm-level correlation across years (Petersen, 2009). 4.2.1 Listing approach and political favours We estimate the following model to analyse the relation between the listing approach and access to political favours: Pol proxy ¼ v þ v Listway þ v Controls þ Year&Industry þ e (1) i;t i i;t 0 1 i;t i where Pol_proxy represents three proxies for political favours, LRG, Subsidy and ETR, and where LRG = (Bank Debt /Total Assets – Bank Debt /Total Assets ), which fol- t t t t–1 t–1 lows Claessens et al. (2008); Subsidy = the total amount of subsidies received from the government, deflated by firm revenue and multiplied by 100; and ETR = (Income Tax Expenses – Deferred Tax Expenses )/Profit before Interest and Tax , which follows t t t Adhikari et al. (2006). Listway is a dummy variable that equals one if a family firm goes public via IPOs, and zero otherwise. Results are robust if we use industry-adjusted Pol_proxy as the dependent variables. We control the following factors in model (1). Size is the firm size computed as the natural log of total assets. LEV is the firm leverage, measured as the ratio of total liabil- ities to total assets. ROA is operating profitability, measured as the ratio of earnings to total assets. Age is the age of a firm since the firm went public. MB is the ratio of mar- ket value of firm equity on book value. CAPEX is future investment opportunities, mea- sured as capital expenditure deflated by firm assets. Famown is the ownership that is held by the family. In addition, Tangibleasset, the proxy for asset tangibility measured as the ratio of fixed assets to total assets, is controlled when regressing LRG. Sales growth (Growth) is controlled when regressing Subsidy. Nominal tax ratio (Nomtax)is controlled when regressing ETR. We also controlled for the year and industry dummies. Definitions of variables are summarised in Appendix A. Because the asset size varies substantially in the year of IPOs, LRG would have anomalous changes in that case. Thus, when regressing LRG, we eliminate the observations in the year the firm listed. We expect that Listway is positively related to LRG and Subsidy, and negatively related to ETR. 20 Huang et al. 4.2.2. Listing approach and earnings quality Following Wang (2006), we use the following three models to analyse the relation between the listing approach and earnings quality. (i) Abnormal accruals Abnormal accruals are estimated using the piecewise nonlinear model: ACC ¼ a þ a CF þ a CF þ a CF þ a DCF þ a DCF CF þ Year&Industry þ e (2) t 0 1 t 2 t1 3 tþ1 4 t 5 t where ACC is the total accruals at t scaled by average total assets at t, where total accruals are earnings before extraordinary items minus operating cash flows. CF is the operating cash flows at t, scaled by average total assets at t. CF is the operating cash t-1 flows at t−1, scaled by average total assets at t. CF is the operating cash flows at t+1 t+1, scaled by average total assets at t. DCF equals one if the change in cash flows at t is less than zero (CF – CF < 0), and zero otherwise. t t −1 ABS ACC is the absolute value of the error term from equation (2), which is the proxy for earnings management. A higher value of ABS ACC indicates a greater level of earnings management and lower earnings quality. ABS ACC represents the depen- dent variable in the following equation: ABS ACC ¼ d þ d Listway þ d Controls þ Year&Industry þ u (3) t 0 1 t i i We control the following factors in model (3). Famceo is a dummy variable that equals 1 if a firm’s ultimate controlling shareholder is also the CEO/chairman of the firm, and 0 otherwise. Inst is the institutional ownership at the year end. Insider is the percentage of equity owned by managers and directors (family members excluded), and Loss is a dummy variable that equals one if net income is less than 0, and zero other- wise. VC is the divergence of cash flow rights and voting rights, which is adopted in previous research (Fan & Wong, 2002; Francis et al., 2005) to investigate the entrench- ment effect of family firms. PC is a dummy variable that equals 1 if the chairman or CEO is politically connected and 0 otherwise (Chaney et al., 2011; Fan et al., 2007), which is controlled to fully investigate whether the poorer quality of accounting information stems from the political background of top executives. Other controlling variables (i.e. Famown, Size, ROA, Lev, Growth and Age) are defined in model (1). We expect that the coefficient on δ to be positive if IPO firms report earnings of lower quality than takeover firms. (ii) Earnings informativeness Earnings informativeness, measured by earnings response coefficients (ERCs), is deemed to be the second proxy for earnings quality. RET ¼ b þ b NI þ b NI Listway þ b NI Controls þ Year&Industry þ t (4) t t t i i t 0 1 2 t i where RET is the 12-month cumulative raw return ending four months after the fiscal year-end at t. NI is the net income for year t, scaled by the market value of equity at t–1. If accounting earnings, NI , is positively associated with RET , Listway is expected t t t to have a negative effect on earnings informativeness, suggesting that IPO firms have a lower market return than takeover firms when they report the same level of accounting earnings. We expect β to be negative. 2 China Journal of Accounting Studies 21 (iii) Persistence of transitory loss components in earnings We estimate the persistence of transitory loss components in earnings as follows. DNI ¼ c þ c DDNI þ c DNI þ c DDNI DNI þ c Listway t t1 t1 t1 t1 t 0 1 2 3 4 þ c DDNI Listway þ c DNI Listway þ c DDNI DNI Listway t1 t1 t1 t1 5 t 6 t 7 t P P P þ c Controls þ c DDNI Controls þ c DNI Controls i t1 i t1 t i i i þ c DDNI DNI Controls þ Year&Industry þ t t1 t1 i t ð5Þ where ΔNI is the change in net income before extraordinary items at t, scaled by aver- age total assets at t–1. D△NI equals 1 if △NI <0, and 0 otherwise. Basu (1997) t −1 t −1 finds that negative earnings changes (transitory loss components in earnings) are less persistent than positive earnings changes. Thus, we expect that the sign of the coeffi- cient on DΔNI ∗ ΔNI (γ ) is negative. The estimate on DΔNI ∗ ΔNI * Listway t-1 t-1 3 t-1 t-1 t (γ ) presents the incremental persistence of transitory losses of IPO firms. A signifi- cantly positive estimate on γ indicates that the transitory losses are more persistent for IPO firms than for takeover firms. 4.2.3. Listing approach, political favours and earnings information quality To test H3, we focus on the subsample of IPO firms and use the measure of political favours to replace Listway in order to re-estimate models (3) to (5). ABS_ACC , RET t t and ΔNI remain the dependent variables to examine the level of abnormal accruals, earnings informativeness and persistence of transitory loss components in earnings. We use factor analysis to extract Factor 1 from the proxies for political favours (LRG, Sub- sidy and ETR), and use Factor 1 to substitute these proxies in models (3) to (5) to examine whether more political favours are related to poorer information quality in IPO firms. We also use alternative factor analysis methods, the principal-component factor analysis and iterated principal factor analysis, to estimate alternative factors (Factor 2 and Factor 3) as the extraction from bank loan, subsidy and effective tax-rate. We expect a negative relation between the extractions of political favours and earnings quality in the IPO subsample. 5. Results 5.1. Descriptive statistics Table 2 presents descriptive statistics for the two subsamples. Regarding the political favours (Panel A), the average (median) growth rate of bank loan size in IPO firms is 1.34% (0.42%) each year, while the mean (median) in takeover firms is 0.2% (–0.14%). The average (median) subsidy in every hundred RMB of sales is ￥0.913 (￥0.238) for IPO firms, which is higher than ￥0.533(￥0.016) for takeover firms. The IPO firms also enjoy a lower effective tax rate (with a mean of 19.86%) than that of the takeover firms (27.65%). Untabulated univariate analysis shows that the average (median) political favours of IPO firms are significantly higher than those of takeover firms at the 5% or 1% level. In general, IPO firms have significantly higher growth in loan size, more government subsidies and lower effective tax rate. Panel B shows the descriptive statistics of variables in models (3) to (5). The aver- age abnormal accrual (ABS_ACC) is 0.053 for IPO firms and 0.063 for takeover firms, while medians are 0.039 and 0.035, respectively. For the earnings informativeness 22 Huang et al. Table 2. Descriptive statistics. Listway=1 (N==1121) Listway=0 (N=1371) Mean SD Median P25 P75 Mean SD Median P25 P75 Panel A: Political favour variables 0.200 LRG(%) 1.340 9.500 0.420 –3.040 4.980 13.147 –0.140 –5.409 5.180 0.553 Subsidy 0.913 1.650 0.238 0.019 0.930 1.630 0.016 0.000 0.290 27.647 ETR(%) 19.863 22.207 13.925 8.362 23.188 33.307 14.989 2.881 33.008 Panel B: Earnings quality variables 0.063 ABS_ACC 0.053 0.051 0.039 0.018 0.071 0.081 0.035 0.015 0.076 0.387 RET 0.243 0.982 –0.062 –0.435 0.555 1.176 –0.122 0.396 0.847 0.002 NI 0.031 0.062 0.035 0.013 0.057 0.086 0.015 0.000 0.040 0.020 △NI 0.011 0.075 0.006 –0.011 0.026 0.150 0.003 –0.024 0.039 0.022 △NI 0.018 0.004 0.010 –0.005 0.032 0.139 0.004 –0.018 0.038 t-1 0.435 D△NI 0.306 0.461 0.000 0.000 1.000 0.000 0.000 0.000 1.000 t-1 Panel C: Control variables 20.687 Size(billion) 20.955 0.876 20.845 20.327 21.504 1.000 20.758 20.062 21.350 –0.002 ROA 0.049 0.078 0.052 0.024 0.086 0.119 0.019 0.000 0.050 0.681 LEV 0.454 0.238 0.437 0.312 0.567 0.431 0.601 0.457 0.741 –0.105 Growth 0.112 0.362 0.171 0.040 0.270 0.768 0.082 –0.146 0.265 1.644 MB 1.504 0.755 1.255 1.082 1.635 1.046 1.222 1.044 1.760 9.934 Age 4.620 3.469 4.000 2.000 7.000 0.339 8.000 10.000 12.000 0.247 Tangibleasset 0.258 0.150 0.239 0.142 0.354 0.171 0.227 0.113 0.350 0.186 Inventory 0.163 0.117 0.140 0.089 0.205 0.188 0.122 0.053 0.244 0.038 CAPEX 0.083 0.067 0.066 0.031 0.119 0.050 0.018 0.005 0.051 0.971 Beta 0.939 0.402 0.988 0.692 1.215 0.337 0.996 0.758 1.171 0.249 Loss 0.085 0.279 0.000 0.000 0.000 0.433 0.000 0.000 1.000 0.315 Famown 0.415 0.164 0.407 0.277 0.533 0.135 0.286 0.227 0.376 0.329 Famceo 0.723 0.460 1.000 0.000 1.000 0.472 0.000 0.000 1.000 0.095 Inst 0.176 0.193 0.106 0.017 0.275 0.155 0.022 0.001 0.118 0.001 Insider 0.049 0.098 0.001 0.000 0.041 0.009 0.000 0.000 0.000 Note: Listway is the approach for firms to go public, which is equals 1 if a firm goes public by IPOs, and equals 0 if a firm goes to public through takeover. China Journal of Accounting Studies 23 model, on average, the 12-month cumulative return (RET ) for IPO firms is lower than takeover firms (0.243<0.387), while IPO firms have much larger net income (NI ) than takeover firms (0.031>0.002). The combined descriptions of RET and NI suggest that the earnings informativeness of IPO firms is lower. △NI , △NI and D△NI are t t–1 t–1 important variables for the persistence of transitory losses analysis. The average change of net income before extraordinary items (△NI ) is 0.011 for IPO firms and 0.020 for takeover firms, while the average D△NI are 0.306 and 0.435, respectively, indicat- t–1 ing that takeover firms are more likely to report negative earnings changes. In terms of controlling variables (Panel C), compared with takeover firms, IPO firms have a greater return on assets (ROA), a lower leverage ratio (LEV ), and higher growth in sales (Growth). IPO firms present a higher ratio of tangible assets (Tangib- leasset) and a lower market-to-book ratio (MB) than takeover firms. IPO firms are more likely to have family CEOs in the position. Institutional investors (Inst) are more will- ing to invest in IPO firms, and non-family insider ownership (Insider) is higher for IPO firms. IPO firms are less likely to report a loss (Loss). 5.2. Correlation analysis Table 3 tabulates the Pearson and Spearman correlations of the variables. Listway is positively correlated with governmental subsidies (Subsidy) in both the Pearson correla- tion test and Spearman correlations test. Listway is negatively correlated with effective tax ratio (ETR) in the Pearson and Spearman correlation tests. But the correlation of bank loan growth and Listway is insignificant, suggesting the need to control for other variables in further examinations. Further, Listway is negatively correlated with the absolute abnormal accrual (ABS_ACC ) in the Pearson correlation test. The correlation coefficients between controlling variables are smaller than 0.4 and greater than –0.75, suggesting that the multicollinearity problem will not be serious. 5.3. Multiple regressions 5.3.1. Listing approach and political favours Table 4 reports regression results on the association between the listing approach dummy and political favours. Column (1) shows that IPO firms are positively related to the growth of bank loan, significant at the 5% level. The magnitude of the coefficient indicates that IPO firms are associated with a 1.2% higher growth of loan on asset each year, which is economically large. Column (2) shows that IPO firms obtain larger subsidies from the government, significant at the 1% level. The subsidies that IPO firms receive are RMB 0.408 more than those of other firms for every hundred RMB of sales. In our sample, the family firm with the median sales of 553 million RMB receives 2.256 million RMB more from the government if it goes public through an IPO. Column (3) shows that listing through IPOs results in a 4.5% decrease in the effective tax rate after controlling for a series of factors, significant at the 1% level. In general, multiple regressions results indicate that IPO firms gain more favourable resources than takeover firms, which is consistent with H1. 5.3.2. Listing approach and earnings information quality Table 5 reports results from the OLS regression using the absolute value of abnormal accruals as the dependent variable. The coefficient on Listway is 0.010 and is 24 Huang et al. Table 3. Correlation analysis. ABS_ACC LRG Subsidy ETR Listway Famown Famceo VC PC Size ROA LEV Growth Inst Insider MB Age Loss *** *** *** *** *** *** * *** *** *** ABS_ACC 0.059 0.003 –0.216 0.011 –0.013 0.009 –0.061 –0.024 –0.181 0.061 0.061 –0.039 –0.004 0.022 0.203 –0.080 0.256 *** ** *** *** ** * *** LRG 0.184 –0.002 –0.021 0.026 0.04R0 –0.000 0.005 –0.020 0.029 –0.173 0.187 –0.050 –0.041 0.021 –0.030 0.002 0.145 *** *** *** *** *** *** *** *** ** *** *** *** *** Subsidy 0.022 0.005 –0.077 0.284 0.121 0.191 –0.114 0.026 0.101 0.191 –0.197 0.044 0.178 0.229 0.070 –0.019 –0.139 *** * *** ** *** *** * * *** ETR –0.113 0.000 –0.040* –0.039 0.010 0.006 0.027 0.012 0.116 0.054 –0.006 0.092 0.089 –0.025 –0.041 0.037 –0.486 *** *** *** *** *** *** *** *** *** *** *** *** *** *** Listway –0.071 0.010 0.102 –0.060 0.314 0.388 –0.200 0.148 0.114 0.326 –0.359 0.157 0.258 0.385 0.017 0.051** –0.215 *** *** *** *** *** *** *** *** *** *** *** *** *** Famown –0.057 0.007 0.029 –0.031 0.316 0.161 –0.086 0.145 0.123 0.241 –0.161 0.162 0.116 –0.013 –0.103 –0.132 –0.175 *** *** *** *** *** *** *** *** *** *** *** *** ** *** *** Famceo –0.073 –0.025 0.060 –0.009 0.387 0.183 –0.250 0.188 0.081 0.227 –0.185 0.130 0.213 0.215 0.047 –0.067 –0.168 *** *** *** *** *** *** *** *** *** *** *** *** *** *** VC –0.014 0.034 –0.054 0.018 –0.150 –0.174 –0.241 –0.068 0.086 –0.195 0.177 –0.100 –0.093 –0.262 –0.094 0.200 0.103 ** *** *** *** *** *** *** *** *** *** *** * *** PC –0.043 –0.027 0.019 –0.020 0.148 0.134 0.185 –0.062 0.079 0.138 –0.072 0.078 0.123 0.087 0.004 –0.035 –0.090 *** *** *** *** *** *** *** *** *** *** *** *** *** *** Size –0.289 0.003 –0.114 0.005 0.144 0.151 0.106 0.046** 0.095 0.128 0.100 0.177 0.340 0.107 –0.370 0.237 –0.232 *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ROA –0.409 –0.334 0.076 0.053** 0.240 0.187 0.204 –0.113 0.109 0.227 –0.417 0.394 0.413 0.228 0.247 –0.191 –0.658 *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** LEV 0.458 0.251 –0.066 0.000 –0.322 –0.177 –0.190 0.104 –0.089 –0.207 –0.508 –0.081 –0.238 –0.153 –0.132 0.173 0.320 *** *** *** *** *** *** *** *** *** *** *** *** *** *** Growth –0.260 –0.102 –0.080 –0.002 0.172 0.146 0.125 –0.031 0.072 0.222 0.397 –0.261 0.180 0.101 0.031 –0.107 –0.328 ** *** *** *** *** *** *** *** *** *** *** *** *** *** Inst –0.028 –0.014 0.041 –0.013 0.187 0.125 0.176 –0.071 0.116 0.297 0.272 –0.153 0.122 0.149 0.239 0.055 –0.257 *** *** *** *** *** *** *** *** *** *** *** *** Insider –0.002 –0.014 0.123 –0.035* 0.339 –0.079 0.219 –0.193 0.025 –0.082 0.172 –0.177 0.094 0.065 0.025 –0.068 –0.130 *** *** *** *** *** *** *** * MB 0.279 0.004 0.106 –0.026 –0.075 –0.137 –0.016 –0.049** –0.004 –0.373 0.020 0.210 –0.100 0.187 –0.001 –0.035 –0.004 ** *** *** *** *** *** *** * *** *** *** *** *** *** Age –0.048 0.002 –0.070 0.063 0.093 –0.125 –0.053 0.121 –0.039 0.220 –0.093 0.099 –0.068 0.027 –0.198 0.004 0.069 *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ** Loss 0.372 0.172 –0.073 –0.227 –0.215 –0.173 –0.168 0.084 –0.090 –0.246 –0.747 0.391 –0.353 –0.193 –0.115 0.059 0.051 This table presents the Pearson and the Spearman correlation coefficients of the variables. The right-hand side of this table is the results of Pearson correlation and the left-hand side of this table is the results of Spearman correlation. *** ** * , and indicate significance at the 5% and 10%, respectively. China Journal of Accounting Studies 25 significant at the 1% level, suggesting that family firms listing through IPOs have a higher level of absolute abnormal accruals (lower earnings quality). The coefficient on Famown is 0.015 and is insignificant, and the coefficient on Famceo is –0.003 and insignificant. The alignment effect observed by Wang (2006) cannot be observed in our sample. Our results suggest that IPO firms report poorer earnings quality. Table 6 reports results of the OLS regression examining the difference in earnings informativeness between IPO and takeover firms. The coefficient on NI*Listway is – 1.298 and significant at the 5% level, suggesting that IPO firms have lower level earn- ings informativeness (lower earnings quality). However, the coefficients on NI*Famown and NI*Famceo are insignificant here, suggesting that neither the entrenchment effect argument nor the alignment effect argument can be proved here. One possible explana- tion is that the effect of political favours on earnings informativeness is so strong that both the entrenchment effect and alignment effect diminish. Table 7 reports the results of the OLS regression estimating the persistence of transitory loss components in earnings for IPO and takeover firms. A firm is consid- ered to provide higher financial reporting quality if the transitory loss components in earnings are less persistent than transitory gain components (Ball & Shivakumar, 2005; Basu, 1997). The coefficient on D△NI *△NI *Listway is 0.581 and signif- t–1 t–1 icant at the 5% level, suggesting that family firms listing via IPOs have a higher persistency in transitory loss than transitory gains (lower earnings quality). The coefficients on D△NI *△NI *Famown and D△NI *△NI *Famceo are not sig- t–1 t–1 t-1 t–1 nificant, suggesting that neither the entrenchment effect nor the alignment effect is considerable. In summary, the results of the analysis of abnormal accruals, earnings informative- ness and persistency of transitory losses consistently demonstrate that IPO firms are associated with lower earnings quality than takeover firms, which is consistent with H2. The entrenchment effect and the alignment effect, as discussed in Fan and Wong (2002) and Wang (2006), are not significant here, while the information effect and resource effect are consistently significant through Tables 5 to 7. The evidence implies that politically supported family firms (firms listing via IPOs) are trying to avoid the uncertainty of political favours and hide their reliance on special resources from com- petitors by providing a lower quality of financial reporting. In addition, they are less motivated to compete for external resources than takeover firms by enhancing earnings quality. 5.3.3. Listing approach, political favours and earnings information quality Table 8 reports the results from the subsample regressions examining whether political favours drive IPO firms to report lower earnings quality. In Panel A, the coefficients on Factor 1 to Factor 3 are positively significant at the 10% level for IPO firms, indicat- ing that higher political favour is associated with more earnings management in IPO firms. In addition, in Panel B, the coefficients on NI*Factor 1 to NI*Factor 3 are nega- tively significant at the 10% level suggesting that higher political favour is associated with lower earnings informativeness in IPO firms. In Panel C, the coefficients on D△NI *△NI *Factor1 to D△NI *△NI *Factor 3 are positively significant at the t–1 t–1 t–1 t–1 5% level. Results suggest that for IPO firms, the magnitude of political favour obtained is negatively correlated with the quality of financial reporting, which is consistent with our H3. The evidence confirms our arguments that IPO firms provide lower earnings quality because of the political favours they received. 26 Huang et al. Table 4. Approaches to going public and political ties: multiple regressions (1) LRG (2) Subsidy (3) ETR Variables Coefficient T-stat. Coefficient T-stat. Coefficient T-stat. *** Constant 0.070 (1.24) 5.125 (4.56) –0.089 (–0.59) ** *** *** Listway 0.012 (2.52) 0.408 (4.01) –0.045 (–2.92) *** ** Size –0.002 (–0.67) –0.213 (–3.96) 0.017 (2.35) *** LEV –0.019 (–1.51) –0.093 (–0.87) –0.047 (–2.91) *** *** ROA –0.149 (–2.99) –0.196 (–0.47) –0.345 (–4.12) * * Age –0.001 (–1.71) –0.025 (–1.66) 0.003 (1.35) *** ** MB –0.001 (–2.91) 0.008 (1.03) –0.002 (–2.01) ** *** CAPEX 0.092 (2.39) –0.035 (–0.05) –0.306 (–3.82) Famown 0.007 (0.39) –0.063 (–0.19) –0.018 (–0.43) ** Tangibleasset –0.047 (–2.29) *** Growth –0.098 (–3.19) *** Nomtax 0.234 (3.17) Year&Industry Yes Yes Yes Adj.R 0.032 0.034 0.056 Obs. 2,238 2,466 2,308 *** ** * , , indicate significance at 1, 5 and 10%, respectively. The variable in bold is our main focus. Table 5. Multivariate analysis of approach to go public and abnormal accruals. Dependent variables: ABS_ACC Coefficient T-stat. *** Cons 0.273 (6.93) *** Listway 0.010 (2.69) Famown 0.015 (1.50) Famceo –0.003 (–1.00) Insider 0.033 (1.76) *** Size –0.013 (–6.51) *** Roa –0.133 (–3.15) *** Lev 0.065 (8.99) Growth –0.005 (–1.41) *** Inst 0.045 (5.43) ** Age –0.001 (–2.53) *** Loss 0.017 (2.95) VC –0.002 (–1.49) PC –0.000 (–0.11) Year&Industry Yes Adj. R 0.345 Obs. 2391 *** ** * , , indicate significance at 1, 5 and 10%, respectively. The variable in bold is our main focus. 6. Sensitivity analysis 6.1. Endogenous tests If there were any unobserved characteristics affecting the listing approach and correlated with information quality or political favours, we would have found spurious correlations, because Listway would be correlated with the error term. To mitigate these potentially endogenous issues, we employ a two-stage regression analysis. In the first China Journal of Accounting Studies 27 Table 6. Multivariate analysis of approach to go public and earnings informativeness. Dependent variable: RET Coefficient T-stat. *** Cons –0.246 (–2.94) NI –0.576 (–0.08) ** NI*Listway –1.298 (–2.08) NI*Famown 1.190 (0.62) NI*Famceo 0.509 (0.98) ** NI_insider –8.240 (–2.17) NI*Size 0.369 (1.05) NI*MB –1.793 (–1.24) NI*Lev 0.254 (0.55) NI*Inst 3.605* (1.92) NI*Age –0.001 (–0.01) *** NI*Loss –5.878 (–5.19) NI*VC 0.080 (0.57) NI*PC –0.363 (–0.84) Year&Industry Yes Adj. R 0.345 Obs. 2391 *** ** * , , indicate significance at 1, 5 and 10%, respectively. The variable in bold is our main focus. stage, we employ variables in Brau et al. (2003) to estimate the probability of a family firm going public through IPOs with the following model: ProbitðListway ¼ 1Þ¼ a þ a Herfindahl þ a HighTec þ a Lev Ind 0 1 2 3 (6) þa MB Ind þ a IpoVol þ a MRT þ a RFR þ e 4 5 6 7 where Listway is the treatment variable that equals one if the family firm goes public through IPOs, and zero otherwise. Brau et al. (2003) examine factors that influence the public acquirers’ choice between IPOs and takeovers. They find four groups of factors that influence the probability of an IPO, and we follow their exogenous variables according to data availability. The first group includes industry-related factors. Herfin- dahl is a proxy for the level of concentration within an industry, which is the sum of squared market shares of all members in a particular industry, calculated by using sales. HighTec indicates whether a family firm operates in high-tech industries. Based on classifications made by the CSRC, HighTec equals 1 if a family firm operates in areas of biotechnology, computers, electronics, communications, medical, and pharmaceuticals, and 0 otherwise. Lev_Ind is the capital structure in industries and acts as a proxy for industrial financial risk (measured as debt/assets). MB_Ind is the market-to-book ratio of the industry where the family firm is operating (measured as total market value/total assets), to capture the potential influence of the industry valuation on the choice of firms to go public. The second group consists of market-timing factors that capture the level of investor sentiment. IpoVol is the proxy of ‘hot issue’ periods in the IPO market, cal- culated as the natural logarithm of the annual volume of IPOs in the listing year. MRT is the proxy of investor enthusiasm, measured as the accumulative return on the stock market in the listing year. The third group contains Funding Demand Factors. We use the 1-year bill rate (RFR) as a proxy for borrowing costs at the time of the transaction. In the second stage, we estimate models (1), (3), (4) and (5) with the Inverse Mills Ratio which is calculated with the predicted values of Listway that are estimated from the first-stage regressions. 28 Huang et al. Table 7. Multivariate analysis of approach to go public and the persistence of transitory loss components. Dependent variable: △NI Coefficient T-stat. Cons –0.062 (–0.64) D△NI –0.254 (–1.54) t–1 △NI –1.218 (–1.36) t–1 D△NI *△NI –0.226 (–0.11) t–1 t–1 Listway 0.015* (1.96) D△NI *Listway –0.018* (–1.78) t–1 △NI *Listway –0.165 (–1.23) t–1 D△NI *NI *Listway 0.581** (2.29) t–1 t–1 Famown 0.017 (0.76) D△NI *Famown 0.026 (0.83) t–1 △NI *Famown 0.291 (0.68) t–1 D△NI *△NI *Famown 0.960 (0.98) t–1 t–1 Famceo –0.000 (–0.03) D△NI *Famceo 0.015 (1.47) t–1 △NI *Famceo 0.014 (0.11) t–1 D△NI *△NI *Famceo 0.001 (0.00) t–1 t–1 Insider –0.012 (–0.31) D△NI *Insider –0.052 (–0.75) t–1 △NI *Insider 0.714 (0.82) t–1 D△NI *△NI *Insider –3.780* (–1.95) t–1 t–1 Size 0.003 (0.58) D△NI *Size 0.014 (1.59) t–1 △NI *Size 0.054 (1.27) t–1 D△NI *△NI *Size –0.025 (–0.24) t–1 t–1 Lev –0.037 (–1.12) D△NI *Lev –0.102* (–1.94) t–1 △NI *Lev 0.204 (1.63) t–1 D△NI *△NI *Lev –0.799*** (–3.56) t–1 t–1 Inst 0.065*** (3.39) D△NI *Inst –0.039 (–1.22) t–1 △NI *Inst –0.394 (–0.93) t–1 D△NI *△NI *Inst –1.743* (–1.79) t–1 t–1 Age –0.003** (–2.06) D△NI *Age 0.001 (0.82) t–1 △NI *Age 0.000 (0.00) t–1 D△NI *△NI *Age 0.012 (0.30) t–1 t–1 VC –0.002 (–1.06) D△NI *VC –0.003 (–0.90) t–1 △NI *VC 0.002 (0.05) t–1 D△NI *△NI *VC –0.075 (–1.07) t–1 t–1 PC 0.001 (0.14) D△NI *PC –0.017* (–1.80) t–1 △NI *PC 0.047 (0.35) t–1 D△NI *△NI *PC –0.308 (–1.45) t–1 t–1 Year&Industry Yes Adj. R 0.254 Obs. 2397 *** ** * , , indicate significance at 1, 5 and 10%, respectively. The variable in bold is our main focus. The results of the first- and second-stage regressions are reported in Table 9. Panel A reports the results from the first stage. Coefficients on HighTec, Ipo_Vol, and RFR China Journal of Accounting Studies 29 Table 8. Political favours and earnings quality in the sample of IPO firms. Panel A: Political favours and abnormal accruals (dependent variables: ABS_ACC) Factor1 0.006 (1.85) Factor2 0.023 (1.90) Factor3 0.013 (1.73) Controls Yes Yes Yes Adj. R 0.261 0.261 0.261 Obs. 1077 1077 1077 Panel B: Political favours and earnings informativeness (dependent variables: RET) NI*Factor1 -0.686 (-1.81) NI*Factor2 -2.718 (-1.81) NI*Factor3 -1.671 (-1.80) Controls Yes Yes Yes Adj. R 0.721 0.721 0.721 Obs. 851 851 851 Panel C: Political favours and Persistence of Transitory Loss (dependent variables: △NIt) * * ** D△NI △NI Factor1 0.233 t–1 t–1 (2.58) * * *** D△NI △NI Factor2 0.940 t–1 t–1 (2.60) * * ** D△NI △NI Factor3 0.519 t–1 t–1 (2.22) Controls Yes Yes Yes Adj. R 0.234 0.234 0.231 Obs. 1077 1077 1077 *** ** * , , indicate significance at 1, 5 and 10%, respectively. are significantly and positively related to Listway, suggesting that a family firm is more likely to go public through an IPO if the firm is operating in a high-tech industry, or the firm goes public in a hot market with a high cost of debt. Herfindahl is significantly and negatively related to Listway, suggesting that family firms are less likely to go public through IPOs when the industry is highly competitive. Panels B and C report the regression result from the second stage. Panel B displays the association between Listway and political favour. The coefficient on Listway is posi- tively related to bank loan growth, significant at the 10% level, and positively related to government subsidies, significant at the 1% level. The effects of Listway on effective tax rate, although not statistically significant, yield findings with expected directions. Panel C presents the association between Listway and earnings quality. The coefficient on Listway is positively significant at the 5% level in the ABS_ACC model. The coeffi- cient on NI*Listway is negatively significant at the 10% level in the ERC model. The coefficient on D△NI *△NI *Listway is positively significant at the 1% level in the t–1 t–1 30 Huang et al. Table 9. Endogenous tests: Two-stage regression. Panel A: First stage: The probability to go public via an IPO vs. a takeover Variables Cons HighTec IpoVol RFR MRT Lev_ind MB_ind Herfindahl Coeff. -3.823*** 0.331* 0.241*** 0.227*** –0.206*** 0.003 –0.001 –3.236*** T-stat (5.66) (1.75) (5.06) (5.27) (–10.81) (1.40) (–1.97) (–2.85) Panel B: Second stage: Listing approach and political favours LRG Subsidy ETR Listway 0.042* 1.241*** -0.015 (1.86) (4.46) (-0.31) Controls Yes Yes Yes Year&Ind Yes Yes Yes Obs. 2,391 2,147 2,397 Panel C: Second stage: Listing approach and earnings information quality ABS_ACC ERC △NI Listway 0.021** NI*Listway –0.834* D△NI *△NI *Listway 0.540*** t-1 t-1 (2.22) (–1.84) (3.27) Controls Yes Controls Yes Controls Yes Year&Ind Yes Year&Ind Yes Year&Ind Yes Obs. 2391 Obs. 2147 Obs. 2397 *** ** * , , indicate significance at 1, 5 and 10%, respectively. China Journal of Accounting Studies 31 conservatism model. These results confirm our findings that IPO firms are associated with more political favours as well as poorer accounting information quality. 6.2. Earnings management through initial public offerings Previous research finds evidence that earnings go upwards before IPOs and decrease in years after IPOs, suggesting a reverse V shape in the pre- and post-IPO period (Aharony et al., 2000; Loughran & Ritter, 1997). The decrease of earnings after IPOs is mainly caused by earnings reversals resulting from earnings manipulation before or during the IPO year. In order to control the effect of earnings manipulation around IPOs, we replicate the multivariate analysis in models (3) to (5) excluding observations of IPO firms in the IPO year and the next two years after IPOs. Untabulated results show the coefficient on Listway is positively significant at the 10% level. The coeffi- cient on NI*Listway is negatively significant at the 10% level. The coefficient on D△NI *△NI *Listway is positively significant at the 10% level. Taken together, we t–1 t–1 obtain similar results from multiple regressions after eliminating observations in the IPO year and two years after IPOs for IPO firms, which indicates that our results are not driven by earnings manipulation around IPOs documented in the prior literature (Aharony et al., 2000; Loughran & Ritter, 1997). 6.3 Firm-year observations in financial distress In the regressions of Tables 4 to 8, we include the sample firms trapped in financial distress. These firms have difficulties in operation or finance, and some of them are constrained in stock transfer by the CSRC, such as limit-up, limit-down, and particular transfer dates. Therefore, we remove the observations in financial distress. In the regres- sion of LRG, the coefficients on Listway are 0.009, statistically significant at the 10% level. In the regression of Subsidy and ETR, the coefficients on Listway are 0.413 and –0.056 respectively, statistically significant at the 1% level. In the ABS_ACC model, the coefficients of Listway are significantly positive at the 1% level. In the REC model, the coefficients on NI*Listway are also significant at the 10% level. The coefficient on D△NI *△NI *Listway in the conservatism model is insignificant but positive at a t–1 t–1 level close to 10%. In general, we obtain similar results after eliminating observations in financial distress. 6.4. An alternative measure of earnings management In model (2), we estimate abnormal accruals using the Dechow and Dichev (2002) model modified by Ball and Shivakumar (2005) and applied by Wang (2006) as the dependent variable. Alternatively, we estimate abnormal accruals using a modified Jones’ model (Dechow. Sloan, & Sweeney, 1995) and re-estimate model (3). Untabulat- ed results show that the coefficient on Listway is 0.007, significant at the 10% level. When we control for the dummy indicator capturing whether an IPO firm is in its IPO year or the next two years after IPOs (IPO ), the coefficient on Listway remains positive although insignificant. Further, after removing the observations in financial distress, the result remains the same. These results confirm our findings that IPO firms have a higher level of earnings management than the takeover firms. 32 Huang et al. 6.5. Alternative measures of bank loan growth In model (1), we use the growth in bank loan scaled by total assets as a proxy of growth in bank loan. Alternatively, we apply the absolute value of growth in bank financing as the proxy (calculated as the natural log of absolute value of loan growth and keep the plus and minus). The re-estimated model shows that IPO firms do gain more bank loans than takeover firms. The coefficient on Listway is significant at the 1% level. In addition, Claessens et al. (2008) find that both short-term and long-term bank debts increase following corporate contributions to deputies. We also test whether the results differ between short-term and long-term bank debts. Using the growth of loan on assets, growth of absolute loan size, and adjusting both measures with the industry- median as the proxies, we find that short-term bank credit increases significantly in IPO firms at the 1% level, and the growth of long-term loan size also increases signifi- cantly in IPO firms at the 10% level. 6.6. Lagged political favours and earnings information quality In Section 5, we examined the effects of political favours in year t on the earnings information quality in year t. One concern is whether the political favours in year t have an impact on firms’ earnings in year t or year t+1. Family firms can use the resources to boost earnings in year t or year t+1 or even later. Thus, they can mask earnings information in year t or t+1 if they obtain the favours in year t. To investigate this lagged effect, we examine the impact of lagged political favours on earnings information quality in t. We use factor analysis as discussed in Section 5.3.3 to form lagged political favour factors. Unreported results show that lagged Factor 1 to Factor 3 is positively correlated with ABS_ACC in IPO firms, which is consistent with the notion that higher lagged political favours are associated with more earnings man- agement in IPO firms. The interaction of NI and lagged factors is negatively significant at the 1% level for IPO firms, suggesting that lagged political favours lower informa- tiveness in the current year for IPO firms. The interaction terms of lagged factors and D△NI *NI become insignificant. t–1 t–1 7. Conclusions The approach taken by family firms in going public reflects the political favours they can obtain from local government. This study examines the listing approach of family firms to shed light on the role of external political favours in shaping earnings quality. Using 2492 Chinese firm-year observations from 2003 to 2008 as the sample, we find that IPO firms derive significantly more bank loans, government subsidies and tax rate discount than takeover firms, suggesting that IPO firms receive more political favours from local government. Further, we document that the quality of reported accounting information is systematically worse for IPO firms than takeover ones. Compared with takeover firms, IPO firms experience larger abnormal accruals, lower earnings informa- tiveness and higher persistence of transitory loss components in earnings. Among IPO firms, the political favours they obtain have a negative impact on their earnings quality. Results are robust to alternative specifications. Our results suggest that IPO firms have incentives to cover the uncertainty and proprietary advantages from political favours, and do not have incentives to compete for external resources by improving earnings quality. China Journal of Accounting Studies 33 Our results extend recent studies on the determinants of financial reporting practice in family firms. We provide evidence that political influence plays an important role in shaping the accounting information quality of family firms in the context of emerging markets with heavy political interference. Our study also contributes to the literature on the impact of government intervention on earnings quality. Findings indicate that firms obtaining greater government supports have incentives to prevent leakage of proprietary information, i.e. the political favours, to competitors and the public. Reducing political interferences seems to be a key to a greater transparency cultivating the efficiency of the capital market. Acknowledgements We are grateful for useful comments from Qingyuan Li, Qiliang Liu, Hongbo Pan, Minggui Yu, workshop participant at Wuhan University, and we thank Jigao Zhu for his comments at the China Journal of Accounting Studies conference in Chengdu. We are extremely grateful to the anonymous referees, executive associate editor Liansheng Wu, language editor Pauline Weetman and joint editor Jason Xiao. We thank the National Natural Science Foundation of China (Approval Nos. 71272202 and 71372167) and Guangdong Natural Science Foundation (Approval No. S2013010013051) for financial support. Any remaining errors are our own. Appendix A. Definition of variables. Dependent variables The growth of loan, = (bank loan size at the end of year t/asset size at the end of LRG year t) - (bank loan size at the end of year t–1/asset size at the end of year t–1). Subsidy Government subsidies, = subsidy in year t/sales in year t*100. ETR Effective tax rate, = (income tax expenses in year t-deferred tax expenses in year t)/Profit before interest and tax in year t. ABS_ACC Absolute value of abnormal accruals of year t. RET 12-month cumulative raw return ending four months after the fiscal year-end of t. △NI Change in net income before extraordinary items at t, scaled by average assets at t-1. Independent variables Listway The approach for firms to go public, which is equal to 1 if a firm went public through initial public offerings, and equal to 0 if a firm went public through a takeover. NI The net income of year t, scaled by the market value of equity at the end of t−1. D△NI Equal to one if △NI <0, and zero otherwise. t −1 t −1 Control variables Famown The ownership that is held by the family. Famceo A binary variable which is equal to 1 if a firm’s ultimate controlling shareholder is also the CEO/chairman of the firm, and zero otherwise. VC Voting rights divided by cash flow rights, where voting right is the weakest link in the chain of control rights, and the cash flow right is the product results of the ownership stakes in each level along the chain. PC A dummy variable which is equal to 1 if the CEO or Chairman is a current or former officer of the central or local government or the military, and zero otherwise. (Continued) 34 Huang et al. Appendix A. (Continued) Control variables Size The natural log of total assets. Lev The firm leverage at year t, measure by total liabilities divided by total assets. Roa Return on assets, the ratio of net income divided by average total assets at year t. Growth The ratio of growth on sales divided by average total assets at year t. MB The ratio of market value of firm equity to book value. Inst Institutional ownership at the end of year t. Insider Nonfamily insider ownership, measured by the percentage of equity owned by managers and directors (family members excluded). Age The period of time in years since a firm became public. Loss A dummy variable which is equal to one if net income < 0, and zero otherwise. Tangibleasset A proxy for asset tangibility, computed as the ratio of fixed assets to total assets. CAPEX The future investment opportunities, measured as capital expenditure deflated by total assets. Beta Market risk, estimated from the market model using corporate monthly returns as the regressor and value-weighted average market monthly returns as the predictor. Inventory The ratio of inventory to total assets. Notes 1. Fan and Wong (2002) argue that firms with proprietary knowledge and specific human capital tend to concentrate their ownership and decision rights in the individuals who possess the specific knowledge (Christie, Joye, & Watts, 2002), because ownership concentration pre- vents leakage of proprietary information about the firms’ rent-seeking activities. 2. Both WIND and CSMAR databases are widely regarded as the most comprehensive and authoritative data sources of listed firms in China. 3. The sample size varies in different tests due to the insufficient data for corresponding vari- ables in the tests. 4. Kwahja and Mian (2005) use logarithms of loan size as the measure of credit access to derive the benefit of political connections. However, takeover firms usually need to burden heavy liabilities for the acquired firms. Hence, total loan size is not a good proxy for credit access in our sample. Instead, growth in loan size after a family firm goes public would bet- ter capture the preferential access to bank credit. 5. Concerning firms reporting either negative income (negative denominator) or tax refunds (negative numerator), their ETRs are distorted in certain situations (Adhikari et al., 2006). One example is a firm with a book loss (negative denominator) and tax refund (negative numerator) because ETR for this firm would be positive even though it pays no taxes. Another example is that some profitable subsidiaries in a group pay taxes (positive denomi- nator) but the group reports a book loss as a whole (negative denominator) because ETR for this firm is negative even though it pays taxes. To address this problem, we use the recoding scheme proposed by Gupta and Newberry (1997) and Adhikari et al. (2006), setting ETR: (1) to zero for firms with tax refunds; and (2) to one for firms with positive taxes and nega- tive/zero income or cash flow. 6. Listway is positively correlated with the absolute abnormal accrual (ABS_ACC ) in the Spear- man correlations test. This can be caused by the skewness of sample distribution from nor- mality or there are outliers in the sample. To address the potential problem, we also estimate model (3) using median regression. The coefficients on Listway are positively significant at the 1% level, which confirms that IPO firms report more abnormal accruals. 7. The results stays similar if the dependent variable is adjusted with the industry median (unta- bulated). 8. Other than the information effect and resource effect, IPO firms are different from takeover firms with respect to the shares and funds they raise in the IPO process. IPO firms need to China Journal of Accounting Studies 35 use these funds, therefore they will invest more in accruals. This could lead to high accruals for IPO firms. Using both the DD model and modified Jones’ model allows us to control for the normal accruals from new investments in working capital and fixed assets. To further control for the funding and investment effects for IPO firms, we control for capital expenses and use sales alternatively as the proxy for size in model (3). Results do not change in the alternative specifications. References Adhikari, A., Derashid, C., & Zhang, H. (2006). Public policy, political connections, and effec- tive tax rates: Longitudinal evidence from Malaysia. Journal of Accounting and Public Policy, 25, 574–595. Aharony, J., Lee, C. J., & Wong, T. J. (2000). Financial packaging of IPO firms in China. Journal of Accounting Research, 38(1), 103–126. Aharony, J., Wang, J., & Yuan, H. (2010). Tunneling as an incentive for earnings management during the IPO process in China. Journal of Accounting and Public Policy, 29(1), 1–26. Ali, A., Chen, T., & Radhakrishnan, S. (2007). Corporate disclosure by family firms. Journal of Accounting and Economics, 44(1–2), 238–286. Allen, F., Qian, J., & Qian, M. (2005). Law, finance, and economic growth in China. Journal of Financial Economics, 77(1), 57–116. Anderson, R., & Reeb, D. (2003). Founding-family ownership and firm performance: Evidence from the S&P 500. Journal of Finance, 58, 1301–1328. Anderson, R., Mansi, S., & Reeb, D. (2003). Founding family ownership and the agency cost of debt. Journal of Financial Economics, 68, 263–285. Ball, R., & Shivakumar, L. (2005). Earnings quality in U.K. private firms: Comparative loss recognition. Journal of Accounting & Economics, 38(1), 83–128. Basu, S. (1997). The conservatism principle and asymmetric timeliness of earnings. Journal of Accounting and Economics, 24(1), 3–27. Bona-Sánchez, C., Pérez-Alemán, J., & Santana-Martín, D. J. (2011). Ultimate ownership and earnings conservatism. European Accounting Review, 20(1), 57–80. Booth, J. R., & Chua, L. (1996). Ownership dispersion, costly information and IPO underpricing. Journal of Financial Economics, 41, 291–310. Brau, J. C., Francis, B., & Kohers, N. (2003). The choice of IPO versus takeover: Empirical evi- dence. The Journal of Business, 76, 583–612. Bushman, R., & Piotroski, J. (2006). Financial reporting incentives for conservative accounting: The influence of legal and political institutions. Journal of Accounting and Economics, 42 (1–2), 107–148. Bushman, R., Piotroski, J., & Smith, A. (2004). What determines corporate transparency? Journal of Accounting Research, 42, 207–252. Chaney, P. K., Faccio, M., & David, P. (2011). The quality of accounting information in politi- cally connected firms. Journal of Accounting and Economics, 51,58–76. Charumilind, C., Kali, R., & Wiwattanakantang, Y. (2006). Connected lending: Thailand before the financial crisis. Journal of Business, 79(1), 181–218. Chen, X., Lee, C. J., & Li, J. (2008). Government assisted earnings management in China. Jour- nal of Accounting and Public Policy, 27, 262–274. Chen, S., Chen, X., & Cheng, Q. (2008). Do family firms provide more or less voluntary disclo- sure? Journal of Accounting Research, 46, 499–536. Christie, A., Joye, M., & Watts, R. (2002). Decentralization of the firm: Theory and evidence. Journal of Corporate Finance, 9(1), 3–36. Claessens, S., Feijen, E., & Laeven, L. (2008). Political connections and preferential access to finance: The role of campaign contributions. Journal of Financial Economics, 88, 554–580. Dechow, P., & Dichev, I. (2002). The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review, 77(s-1), 35–59. Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earnings management. Accounting Review, 70, 193–225. Du, J., & Xu, C. (2007). Regional competition and regulatory decentralization: Case of China. Word Economy and Finance Research Programme: Economic & Social Research Council. 36 Huang et al. Du, J., & Xu, C. (2009). Which firms went public in China? A study of financial market regula- tion. World Development, 37, 812–824. Faccio, M., Masulis, R. W., McConnell, J. J., & Offenberg, M. S. (2006). Political connections and corporate bailouts. Journal of Finance, 61, 2597–2635. Fan, J., & Wong, T. J. (2002). Corporate ownership structure and the informativeness of account- ing earnings in East Asia. Journal of Accounting and Economics, 33, 401–425. Fan, J., Wong, T. J., & Zhang, T. (2007). Politically-connected CEOs, corporate governance and post-IPO performance of China’s partially privatized firms. Journal of Financial Economics, 84, 330–357. Fan, J., Wong, T. J., & Zhang, T. (2012). Founder succession and accounting properties. Contem- porary Accounting Research, 29(1), 283–311. Foucault, T., & Parlour, C. A. (2004). Competition for listing. The RAND Journal of Economics, 35, 329–355. Francis, B. B., Hasan, I., & Sun, X. (2009). Political connections and the process of going pub- lic: Evidence from China. Journal of International Money and Finance, 28, 696–719. Francis, J., LaFond, R., Olsson, P., & Schipper, K. (2005). The market pricing of accruals quality. Journal of Accounting and Economics, 39, 295–327. Gupta, S., & Newberry, K. (1997). Determinants of the variability in corporate effective tax rates: Evidence from longitudinal data. Journal of Accounting and Public Policy, 16(1), 1–34. Jiang, D., Liang, S., & Chen, D. (2009). Government regulation, enforcement, and economic con- sequences in a transition economy: Empirical evidence from Chinese listed companies imple- menting the split share structure reform. China Journal of Accounting Research, 2,71–100. Khwaja, A., & Mian, A. (2005). Do lenders favor politically connected firms? Rent provision in an emerging financial market. Quarterly Journal of Economics, 120, 1371–1411. Leuz, C., & Oberholzer-Gee, F. (2006). Political relationship, global financing, and corporate transparency: Evidence from Indonesia. Journal of Financial Economics, 81,411–439. Loughran, T., & Ritter, J. (1997). The operating performance of firm’s conducting seasoned equity offerings. Journal of Finance, 52, 1823–1850. Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: Comparing approaches. The Review of Financial Studies, 22(1), 436–450. Sapienza, P. (2004). The effects of government ownership on bank lending. Journal of Financial Economics, 72, 357–384. Shleifer, A., & Vishny, R. W. (1997). A survey of corporate governance. Journal of Finance, 52, 737–783. Sun, Y., & Luo, D. (2011). Government competition, resource allocation and ‘Shell Resource’ transfer. Journal of Management Science (in Chinese), 24(1), 11–20. Wang, X., Xu, L. C., & Zhu, T. (2004). State-owner enterprises going public: The case of China. Economics of Transition, 12, 467–487. Wang, D. (2006). Founding family ownership and earnings quality. Journal of Accounting and Economics, 44, 619–656. Wei, Z., Wu, S., Li, C., & Chen, W. (2011). Family control, institutional environment and cash dividend policy: Evidence from China. China Journal of Accounting Research, 4(1–2), 29–46. Wu, W., Wu, C., & Rui, O. M. (2010). Ownership and the value of political connections. Work- ing Paper. Shanghai Jiao Tong University and Chinese University of Hong Kong. Zhu, H. (2004). Decentralization, fiscal incentive and privatization of SOEs in China. World Economy (in Chinese), 12,14–24.
http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png
China Journal of Accounting Studies
Taylor & Francis
http://www.deepdyve.com/lp/taylor-francis/listing-approach-political-favours-and-earnings-quality-evidence-from-wy2F9500Z0
