Relationship-specific transactions, property right and accounting
conservatism
Name: Tao Ma
Student number: 11702192
Thesis supervisor: Dr. Ghazizadeh, Pouyan Date: June 21, 2018
Word count: 12,246
MSc Accountancy & Control, specialization Control
Statement of Originality
This document is written by student Tao Ma who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This paper focuses on the relationship between relationship-specific transactions and the company’s accounting conservatism. We argue that relationship-specific transactions with customers or suppliers would have a negative influence on the company’s accounting conservatism, especially for the non-state-owned enterprises.
This paper chooses 2010-2016 firms listed in A share as the sample and studies the above-mentioned relationship. We find that trading parties have established the relationship of trust through long-term cooperation, which can provide the private channel for communication stably. All of them are willing to share information in a low-cost and high-efficiency way for the motivation of cost minimization and interest maximization, which can help them get more information about the firms’ operation and finance. With suppliers and customers growing clout, it also can require the firm to share information through private channels, thereby reducing demands on the quality of public information, and accounting conservatism.
Furthermore, we consider the background of property rights and find that non-state-owned enterprises are more dependent on the relationship-specific transaction to maintain long-term and stable earnings performance for their weak position in the competitive market. In contrast, state-owned enterprises are in an advantageous with access to more resources and preferential policies, so suppliers and customers have far fewer effects on them. Therefore, non-state-owned enterprises with its suppliers and customers prefer to share information through private channels and have weak demand for stable information.
Contents
1 Introduction ... 5
2 Literature review and hypothesis ... 9
2.1 Relationship-specific transactions ... 9
2.1.1 Positive effect... 9
2.1.2 Negative effect ... 10
2.2 Accounting conservatism ... 12
2.3 Relationship-specific transactions and accounting conservatism ... 15
2.4 Hypothesis... 17
3 Methodology ... 20
3.1 Sample and data source ... 20
3.2 Variables and Model ... 21
3.2.1 Accounting Conservatism ... 21
3.2.2 Relationship-specific Transactions ... 22
3.2.3 Control Variables ... 24
3.2.4 Model ... 24
4 Results and Discussion ... 26
4.1 Descriptive statistics ... 26
4.2 Correlation analysis ... 29
4.3 Multiple linear regression ... 30
4.4 Robustness test ... 33
4.5 Discussion ... 36
1 Introduction
In the market, almost every company has business relationships with other companies, such as suppliers and customers. Thus, the quality of the company’s public information is important, because in general, a company’s suppliers and customers use the company’s financial performance to measure its underlying economic performance. A well-established market and reliable company information can help all participants quickly find the trading partner and conduct transactions fairly and efficiently. However, the choice of trading methods and trading partners in the market are constrained not only by the internal demand but also the external environment in which the company is located in. In some developing countries, such as China, the high transaction costs caused by the incompleteness of the external market have prompted companies to establish transactions based on a relational network which can greatly reduce the costs of transactions, which are defined as relationship-specific transaction.
Western scholars first focus on the specific transactions between Japanese automakers and suppliers or customers, and then begin to study the relationship between relationship-specific transactions and accounting conservatism. In the past few years, Chinese scholars began to focus on the relationship-specific transactions between the company and suppliers or customers. These studies have shown that mutual trust can be established between the trading parties through long term relationship-specific transactions. It can help to reduce the transaction costs because they can exchange information through private channels. However, at the same time, the transaction mentioned above can also induce the opportunism, thereby increasing the transaction costs and leading to inefficient cooperation.
China, as a country with a transition economy, has a strong relationship-oriented characteristics, coupled with the intensification of product market competition in recent years and immature market development, resulting in a generally high transaction cost among enterprises. And asymmetric market information hinders the fair trading, and increases the difficulty of enterprises seeking new suppliers or customers. So, the enterprises tend to build the long-term business relationship with its stakeholders to reduce the costs of transactions and solve the interest conflicts (Klein, Crawford, & Alchian, 1978; Williamson, 1979). Thereby, the demand for the disclosure and transparency of public information is reduced or even eliminated (Ball, Kothari, & Robin, 2000; Z. Li, Ye, & He, 2011). Moreover, through private channels, stakeholders can access the company’s information more directly (Burkart
& Ellingsen, 2004), and they prefer to use the private channels to exchange information as it is lower costs and more efficient. Accordingly, from the company’s perspective, the demand for public information and accounting conservatism is reduced as well. This paper focuses on whether relationship-specific transactions between Chinese listed companies and its stakeholders would lead to the decline of company’s accounting conservatism. The reasons for choosing China are as follows.
Firstly, after nearly 20 years of rapid development, China’s total economic output surpassed that of Japan and the European Union, ranking second in the world. As the world's largest emerging economy, China has become a major market growth point for many multinational companies. And according to the study conducted by Bespoke Investment Group, China’s stock market accounts for 16% of the global stock market’s value. China’s stock market has already been one of the largest stock markets in the world, so, there is a large sample which is accessible and can be studied.
Secondly, with the rapid growth of China's economy, China's capital market has also developed rapidly and has created a set of accounting systems that combine Chinese and Western characteristics. As a business language, accounting plays an important role in promoting global economic integration. However, the Chinese accounting academics' research methods for accounting problems in China have not been integrated into the international mainstream for a long time. Therefore, this paper wants to do an integration work, by using the international popular theories and methods to solve China's problems. Thirdly, at present, as China is undergoing economic transformation, the scope of labour division in society and market development are limited, effective legal systems, property rights systems, and sound financial market mechanisms have not yet been established, making it impossible for China to move from relational contracts to regular contracts in a short period of time. The incompleteness of the external market has increased the cost of contract implementation and supervision, induced opportunistic behaviour, and increased the degree of asset specificity(Klein et al., 1978), thereby hindering the normal trading activities. Rajan and Zingales (1998) compared the relationship-specific transaction system and the market trading system and found that because of the lack of formal legal protection, relationship-specific transactions make more use of the reputation mechanism of the business connection circle to ensure the implementation of the transaction. At this time, the transaction method based on the relationship can reduce the transaction cost more effectively and become the natural choice under the situation of the absence of the external legal mechanism
and the market effectiveness(Xia & Chen, 2007). Johnson et al. (2000) studied the enterprise relational contract in the transition economy and also found that in a country where the legal protection is not sound, the company prefers to make a contract with a company that has a relationship with itself. As a result, transactions continue to rely on relationships for a long period of time, so, it is more meaningful to study the relationship-specific transactions that take place in the Chinese market.
To do the research, Stata 15.0 and Excel were used, and this paper chooses 2010-2016 firms listed in A share as the sample and studies the relationship between relationship-specific transaction and accounting conservatism. This paper uses empirical Ordinary Least Square (OLS) multiple regression model to do the research, and we find that both sides of the transaction establish a trust relationship through long-term cooperation and can provide a stable channel for communication. And they are all willing to share information in a low-cost and efficient way to reduce the cost and maximize profit. With the increasing influence of suppliers and customers, it may also require companies to share information through private channels, thereby reducing the requirements for the quality of financial statements and the accounting conservatism. We also find that non-state-owned enterprises rely more on relationship-specific transactions because they are in the weak position in the competitive market. However, suppliers and customers have less influence on state-owned enterprises, because state-owned enterprises have strong bargain power in the transaction.
The contributions of this paper are: (1) The related research literature of relationship-specific transaction is combed and summarized in detail, which is helpful to understand the relationship-specific transaction behaviour more deeply. (2) The study of accounting conservatism is enriched from the perspective of stakeholders. Currently, the article is limited to both suppliers and customers. Therefore, this paper proposes different perspectives to study existing suppliers and customer's influence on accounting conservatism. (3) Most of the existing literature studies focus on the relationship-specific transactions that depend on the government-business network(S. Tang, Hu, & Sun, 2011). This article is concerned with the establishment of a relationship-specific transaction model based on the business network between the company and its major suppliers and customers.
This paper is divided into four sections. The first section is the introduction which mainly introduces the research issues and thesis framework, including the research background and research ideas, as well as the research contributions and overall structure of this paper. The second section is the theoretical basis, literature review, and hypothesis. It is mainly to review
the relevant theories involved in the research issues of this paper and to classify and review the relevant domestic and foreign research literature. Based on these theories and previous studies, we develop the hypothesis for this paper. The third section is about methodology. It mainly introduces the sample selection and data sources of the research design, the definition and regression model of major variables. The last section discusses results. It mainly summarizes the results of the previous empirical analysis and draws conclusions from the relevant research.
2 Literature review and hypothesis
2.1 Relationship-specific transactionsWilliamson (1985) argued that in order to reduce transaction costs, the company would maintain a stable business relationship with its suppliers or customers for a long time period by making the relational specific investment, and this kind of long-term cooperative relationship is the relationship-specific transaction. Due to cultural differences, there are also differences in relation-specific investment between eastern and western companies in establishing cooperative relationships. Western companies establish cooperative relationships in accordance with a specific business purpose. And in the process of cooperation, they focus less on the personal relations among individuals in both companies. But in China, in addition to commercial purposes, it is also concerned about the development of personal relations between the two parties (Xin & Pearce, 1996). Studies have revealed that this kind of transaction has both pros and cons. Moreover, the scholars study the governance mechanism of the relationship-specific transaction. Section 2.1 mainly reviews the previous research on effects of relationship-specific transaction and governance mechanism.
2.1.1 Positive effect
The establishment of relationship-specific transactions between companies and customers/suppliers can improve the efficiency of resource allocation, reduce transaction costs, and create a continuous relationship rent. This is a necessary condition for enhancing corporate performance and enhancing the core competitiveness of enterprises.
The study of Kalwani and Narayandas (1995) found that establishing a relationship-specific transaction with major corporate customers can not only improve the efficiency of the allocation of corporate resources, but also that its continuous procurement behaviour can provide long-term stability income guarantees for the company. Thereby, it can improve the company's inventory management and increase the recovery of accounts receivable. Kong (2011), based on the framework of the new transformation economy, examined the economic consequences of social relational networks from the perspective of major Chinese customers and suppliers. Research showed that social relations can bring benefits to enterprises. Especially in a weak institutional environment, relational networks can reduce transaction costs by reducing information asymmetry. That is, enterprises’ increased dependence on relational networks can reduce credit costs and advertising and sales cost. Patatoukas (2012)
found that the company's stable relationship sales can improve the company's operating efficiency, reduce sales costs, and can achieve the purpose of improving corporate financial performance. In related Chinese research, Z. Chen and Wang (2014) found that supply chain integration can reduce the company's costs, improve the efficiency of asset allocation, and enhance the enterprise’s financial performance through the benefits of relationship-specific transactions with upstream and downstream partners. Y. Li (2017) took Chinese companies in manufacturing industry from 2007 to 2015 as the sample to do the research, and the research revealed that relationship-specific transactions can significantly improve corporate performance as well as the inventory management efficiency.
The Chinese scholar Y. Wang (2006) extended the relational contract with a single market interaction to the connected market and discussed the interactive relationship between relational contract and formal contract governance mechanism, thus portraying the influence of social division of labour and market scope on micro-governance. Tao Wang and Hong Qin (2006) found that relationship rules, relationship persistence, specific investment, and explicit contracts can have an impact on the opportunism behaviour of suppliers, resulting in synergies and embezzlement effects. Y. Chen and Wang (2010) found that higher audit quality can help companies and suppliers first establish trust relationships, and then choose commercial credit models with lower transaction costs. Xiaoyang He (2011) found that in a relational trading environment, mutual trust between families can reduce transaction costs and increase the value of corporate relational assets. Zhijun Tang et al. (2012) found that in economic transformation, relational contracts can reduce information asymmetry and establish long-term relationships for mutual trust. Zhang, Ma, and Sheng (2012) conducted a study on the manufacturing industry and found that the company sent signals to suppliers or customers through high-quality auditors, indicating that they are in a good financial position to maintain relational transactions. Xu, Lin, and Wang (2015) found that relational trade fairs enabled companies to actively control operational risk and financial risk, thereby choosing a lower level of financial leverage. Bao and Zhao (2015) found that relationship-specific transactions enhanced the incentive of suppliers to hold cash. This motive was stronger in non-state-owned enterprises; at the same time, institutional environment, market development, and social integrity effectively relieved this motive.
2.1.2 Negative effect
Klein et al. (1978) and Williamson (1979) have found empirical evidence based on an opportunist perspective that relationship-specific transactions of enterprises may reduce the
quality of public information as well as the demand for public information, and lead to the occurrence of opportunistic behaviour. Khanna and Thomas (2009) took the Chilean listed company as the research sample and found that when companies cross-shared, jointly controlled, and established linkages through chained directors, there is a high probability that these companies' stock prices will rise or fall simultaneously. Z. Li et al. (2011) took private enterprises in China as the research sample and revealed that the existence of relationship-specific transactions made the demand for high-quality public information not high. Specifically, the stock price of the listed companies with a high level of relationship-specific transactions cannot reflect the real situation of the company. Because to a large extent, the company's stock price will rise and fall in sync with related parties. The research of Xu, Lee, and Lin (2014) revealed that the increase in the proportion of relationship-specific transactions with customers or suppliers would significantly reduce the investment efficiency of enterprises, that is, relationship-specific transactions can easily trigger the opportunistic behaviour of management while bringing relationship rents. In addition, when the degree of relationship-specific transactions of companies is increasing, the bargaining power of customers and suppliers will become stronger, and the possibility of using renegotiated opportunities to seek more private profits will increase, and a certain degree of “predatory” will occur. (Porter, 1992; Suutari, 2000)
The Chinese scholar Y. Tang (2009) analysed the relationship between relationship-specific transactions and company performance and found that the supplier's bargaining power will increase the company's dependence and transfer costs, which will lead to lower product quality, higher prices, and weaken the company's profitability. Based on the China’s economic transformation, S. Wang, Zhou, He, and Yu (2010) revealed that as the increase of relationship-specific investment, the company would more inclined to conduct long-term by local accounting firms. Z. Lin, Zheng, Peng, and Xu (2014) further argued that only when the firm obtains a standard audit opinion in the previous year does the relationship-specific transaction significantly improve the stability of the contract between the firm and the local auditing firm. Xu et al. (2015) took the Chinese companies in manufacturing industry as the research objects and showed that relationship-specific transactions motivated companies to do earnings management. In areas where the institutional environment is not well developed, companies are more willing to conduct relationship-specific transactions while reducing the level of earnings management to deliver useful private information. Z. Lin, Zheng, and Tang (2014) found that relationship-specific investment made it difficult for executives to work
through current performance to reduce the sensitivity of pay-performance from the perspective of executive compensation. And this effect is even more pronounced in regions with low levels of trust. Xu et al. (2015), based on the background of China's relationship-specific transactions, found that the higher the degree of customer concentration, the stronger the positive link between management power and equity incentives, resulting in a decline in company performance. Xu et al. (2014) revealed that relationship-specific transactions exacerbated excessive investment and underinvestment. However, state-owned companies can effectively control the relationship and slow down this negative relationship.
To sum up, we can see that on the one hand, this kind of transactions can establish long-term trust and cooperative relations, regulate mutual behaviours through private communication mechanisms, lower transaction costs, and have the effect of value creation; on the other hand, it induces opportunistic behaviours of both parties to transactions, and increases transactions. cost, resulting in inefficient cooperation, hinder the realization of value creation.
2.2 Accounting conservatism
The early definition of accounting conservatism was mainly the description of the normative category and lacked relevant theoretical support. Bliss (1924) believes that accounting conservatism is predicting all possible losses but not expecting any uncertain gains. Since then, the scholars have carried out similar statements. However, Ball and Shivakumar (2005) believe that there are certain defects in the above definition, and there are insufficient considerations for the specific conditions of timeliness. Basu (1997) gives a definition from an empirical point of view by constructing a surplus-equity model, using bad news and good news as specific conditions for timeliness, as opposed to bad news (loss), for good news (profit). Confirming that more evidence is needed, resulting in the accounting asset's asymmetry in recognizing the gains and losses, is the accounting conservatism.
Accounting conservatism stems mainly from accounting contracts and is influenced by laws and regulations (Watts, 1993). Watts (2003) classified the existence of conservatism as four factors: contract, litigation, regulation, and taxation. Since then, many scholars have done a lot of tests on this, and some scholars have also explained the reasons for the existence of conservatism from the perspective of corporate governance, audit quality, and so on. This article combines the existing literature and summarizes the factors that influence the accounting conservatism in the following four aspects. From the perspective of debt contract, Ahmed et al. (2002) found that due to the existence of debt covenants, the more serious the
conflicts of interest between shareholders and creditors, the higher the degree of accounting conservatism required by creditors. From the perspective of litigation, scholars believe that accounting standards alone cannot truly improve the quality of accounting information and must be complemented by sound laws, high-quality judicial and law enforcement standards. Therefore, the more stringent the legal penalties and regulatory mechanisms, the higher the degree of accounting conservatism. When the interests of shareholders and creditors are compromised, they are more likely to sue for compensation (Watts, 2003). As for regulation, accounting standards are necessary but not sufficient conditions for steady accounting information. The mere reliance on changes in accounting standards does not really improve the quality of accounting information. It requires strict legal penalties and strict administrative supervision mechanisms (C. Zhu & Li, 2008). Regarding taxation, in order to avoid duty reasonably, the management are likely to adjust accounting earnings and taxable income. (Shackelford, 2001).
The Chinese scholars' research on the influencing factors of accounting conservatism focuses on debt covenants, regulatory systems, and corporate governance. In terms of debt covenants, Z. Sun, Liu, and Wang (2005) suggested that the debt covenants of listed companies are positively correlated with accounting conservatism, and the accounting conservatism increases with the increase of debt contract conflicts. K. Zhu and Chen (2006) found that when corporate bank loan ratios were at a low level, non-state-owned companies showed better accounting conservatism than state-owned companies; but when the ratio was high, state-owned enterprises and non-state-owned enterprises did not show significant differences. J. Zhu (2011) conducted research on the listed companies in Shanghai and Shenzhen Stock Exchange market, which were sued by banks, and found that their accounting information is more robust. In terms of the regulatory system, Qu and Qiu (2007) also found that accounting standards alone cannot truly improve the quality of accounting information. It also required a strong legal system to protect it. Ce Chen and Lyu (2011) also found that stringent regulatory measures can enhance the accounting conservatism of the private enterprises which are listed in small and medium-sized boards and mainboards. Therefore, the more complete the construction of the legal system, the fewer the government's intervention in the market, the faster the marketization process and the better the economic development, the stronger the accounting conservatism(S. Zhu & Xia, 2009).
Moreover, state-owned shares can weaken the correlation between liability factors and accounting conservatism (Z. Sun et al., 2005). Cao, Li, and Sun (2005) argued that when the
remaining control power is greater than the residual income right, the major shareholder manipulates the accounting information by affecting the formation and issuance of accounting information, and delays the recognition of bad news to seek private interests. G. Sun and Zhao (2014) analysed the absence of state-owned enterprise owners and soft debt constraints from the perspective of behavioural finance, making management overconfidence more robust to robustness.
Both Chinese and foreign scholars have demonstrated that accounting conservatism has a governance function that can reduce information asymmetry and optimize investment decisions (Ball & Shivakumar, 2005; Holthausen & Watts, 2001). This paper mainly reviews financial costs and investment efficiency.
In terms of financing costs, accounting conservatism can effectively solve the interest conflicts, and have the effect of reducing debt costs and improving contract efficiency (Ahmed, Billings, Morton, & Stanford‐Harris, 2002). The study on the bid-ask spreads in the secondary loan market revealed that accounting conservatism can mitigate the information superiority of informed traders and improve bond trading’s efficiency (Wittenberg-Moerman, 2008). There is still controversy over equity financing. The study revealed that accounting conservatism did not play an important role in reducing the costs of equity capital (Francis, LaFond, Olsson, & Schipper, 2004). However, Lara et al. (2006) believe that the measurement method they used was improper, and then used a variety of methods to empirically test the negative correlation between robustness and ex-ante capital cost. Garcia et al. (2011) also found that higher accounting conservatism can lower the firm's costs of equity capital.
The Chinese scholars Mao and Dai (2009) proved that the robustness significantly reduced the company’s debt cost, but did not find any impact on the pre- and post-event costs of equity capital. Tao and Liu (2010) found that the robustness of debt maturity structure helps state-owned listed companies obtain long-term debt financing. Liu Yu and Yuan Lin (2010) found a positive correlation between accounting conservatism and the probability of refinancing of the company. Jinxin Zhang and Yi Wang (2013) found that unconditional robustness in advance compared with conditional stability afterward is more conducive to alleviating corporate financing constraints. Binyuan Luo (2014) also proceeded from two types of robustness, and conditional stability can also ease corporate financing constraints and reduce the problem of principal-agent conflicts.
In terms of investment efficiency, Watts (2003), Ball and Shivakumar (2005), and Francisa and Martin (2010) found that the long-term accounting conservatism of companies can restrict managers' unfavourable investment incentives and choose appropriate investment decisions. Venugopalan (2004) found that accounting conservatism can effectively constrain the behaviour of enterprise management and rationally allocate enterprise resources. Bushman et al. (2005) studied the relationship between corporate investment efficiency and accounting conservatism in different countries and revealed that in the country with higher accounting conservatism, companies can mitigate investment risk as quickly as possible and withdraw from loss projects. Bushman and Piotroski (2006) studied the internal restraint system and found that strong conservatism will reduce the recognition of investment opportunities. Biddle and Hilary (2006) found that the higher the soundness of public information, the more the information asymmetry between internal managers and external fund providers can be reduced, thus improving investment efficiency.
Chinese scholars S. Zhu and Xia (2009) used Bushman and others’ models to conduct empirical research and found that stable policies would encourage companies to underestimate their own financing capacity, making companies lack of funds, making the investment scale is limited, resulting in insufficient investment. Gang Sun (2010), Hongxia Liu and Lingling Suo (2011) both considered that the soundness of accounting performance reduces the investment misconduct of enterprises’ underinvestment and excessive investment.
2.3 Relationship-specific transactions and accounting conservatism
Klein et al. (1978) and Ball et al. (2001) found that stakeholders who rely on relationship-specific transactions to exchange information can share information through reputation and private communication mechanisms and regulate the behaviours of both parties. With regard to the need for disclosure of public information, firms in civil law countries are more
dependent on relationship-specific transactions than the company in the countries of the maritime law system. And civil law countries have lower levels of accounting conservatism. In addition, LaFond (2005) proposed that the relational infrastructure governance system uses private communication as the dominant information delivery mechanism, such as between suppliers and companies, thus reducing the need for accounting conservatism. Regarding the company's life cycle, Peek et al. (2006) found that more early adopters of relational financing will reduce the need for robustness of accounting.
The Chinese scholars X. Chen and Huang (2008) also found that the accounting conservatism generally shows a U-shaped distribution throughout the company's life cycle. This is due to the fact that in the early stages of growth, companies rely mainly on relational financing to enter the mature stage of development after financing. The approach tends to be market-oriented. And creditors’ requirements for the level of accounting conservatism have also increased. Z. Li et al. (2011)'s research on the relationship-specific transactions between privately-owned listed corporate groups in China proves that the replacement of private communication mechanisms has enabled companies to reduce the need for high-quality public information.
Almazan et al. (2006) researched on the effect of suppliers and customers on information disclosure and revealed that good news and information transparency can have an impact on this kind of transactions. Hui et al.(2012) studied the effect of the bargaining power of the transaction parties on the accounting conservatism and found that when the supplier/customer has a bargain power, the company will tend to maintain the accounting conservatism to show that the performance is good, as a commitment to the contractor, to maintain long-term relationship with each other. X. Wang (2014) examined the impact on the accounting conservatism from the perspective of the bargaining power of both parties. The study found that the strong bargaining power of customers will prompt suppliers to improve the degree of accounting conservatism, and the more specific investment in the relationship between companies, the stronger this effect will increase. And the more fierce competition in the customer's market, this effect will be weakened.
Recalling the current domestic and international research on relationship-specific transaction and accounting conservatism, they all stand on the contractual perspective between suppliers and customers and companies, and explore the influence of relational contracts on accounting conservatism. This paper has found that Chen et al. (2009), Yao et al. (2009), Hui et al. (2012), and X. Wang (2014) have all discovered that with the increasing number of relationship-specific transactions between suppliers and customers and businesses, contractual opportunistic behaviours increase, which requires accounting information as a governance mechanism to eliminate information asymmetry between the two parties, constrain opportunism, and avoid stuck risks, so the increase in relational transactions also has good influence on accounting conservatism.
Most of prior research focus on the negative influence which this kind of transaction has on accounting conservatism. However, the different conclusion might be drawn if the research
focus the positive influences. Nowadays, studies on how the relationship-specific transaction influence the accounting conservatism has not yet reached an agreement. From the perspective of stakeholders, China has only limited number of studies on accounting conservatism. Based on these points, this paper attempts to investigate whether a relationship-specific transaction between Chinese listed companies and suppliers or customers will lead to the decline of corporate accounting conservatism.
2.4 Hypothesis
According to Transaction Cost Theory, with the motivation to cut transaction costs, the company and suppliers or customers are motivated to do the relationship-specific investment, thereby maintaining the long-term business transaction relationship(Williamson, 1985). Because the assets of a relationship-specific transactions usually have a high asset specificity, it also limits the extent to which assets can be used for different purposes, thus, the companies will face huge losses if the relationship breaks down. Meanwhile, due to the high specific of asset, interest conflicts are difficult to mitigate through formal ways such as law system. Therefore, the two sides of the relationship-specific transactions usually use the private mechanism to exchange information and restrict the behaviour of both parties(Klein et al., 1978), and do not need to restrain the disclosure of public information (Z. Li et al., 2011). According to Stakeholder Theory, as the main stakeholders of the enterprise, suppliers and customers also do relationship-specific investment as what the shareholders do. Besides, they bear the important responsibility of the company's procurement and sales in the industrial chain, and share the business risk, so the enterprise is more regarded as an aggregation of various economic interest (Blair, 1995). Suppliers and customers can get the company's business information through more private channels (Burkart & Ellingsen, 2004). Meanwhile, with the increases of supplier and customer’s bargain power, stakeholders can require to exchange information through more private ways, because their influence over the company overtakes the company's management gradually (Zhang, Ma, & Zhang, 2012).
According to the Asymmetry Information Theory, there are various governance problems because of the differences in the symmetry of obtaining information among the company, suppliers and customers. As one of the governance mechanisms, accounting information can alleviate this information asymmetry. But combining the social background and theoretical analysis, it can be found that:
Firstly, under the background of economic transformation in China, the relational contract is the leading contractual governance mechanism in a long period of time, and the relationship-specific transactions can solve the conflicts and problems caused by the lack and imperfect of legal system and market mechanism more effectively in the transaction process. Secondly, in order to lower the transaction costs, both of suppliers/customers and the company want to maintain the long-term relationship-specific transactions, which also enables them to continuously communicate and deliver information through private channels. Compared to disclosure of public information, private communication is more cost-effective and efficient, thereby reducing the need for high-quality public information disclosure. Finally, for the purpose of maximizing the interests of stakeholders, the increase in relationship-specific transactions increases the specificity relationship specific assets. And as an aggregation of economic interests, the company, suppliers and customers can share the company's business information in a more timely manner through private communication channels, such as the bad news like financial losses. Thus, the need for accounting conservatism of public information is reduced. Accordingly, the first hypothesis is proposed:
Hypothesis 1. In relationship-specific transactions, there are negative correlations between
accounting conservatism of the company and degree of dependence between the company and suppliers or customers.
Furthermore, take the property right of the company into consideration. In China, based on property rights, enterprises can be divided into two types, one is state-owned, and the other is non-owned. Under the background of China’s economic transformation, the state-owned enterprises are supposed to take on social responsibilities such as solving employment problems and providing public goods (Y. Lin & Li, 2004). The connection with the government has made it easier for state-owned enterprises to acquire scarce resources, and to have access to financing guarantees and financial subsidies when it comes to operational risk or debt default risk (Johnson & Mitton, 2003).
In market competition, the government is more inclined to protect state-owned enterprises, so non-owned enterprises will face more restrictions in the market competition than state-owned enterprises. For instance, when applying for a loan, banks would require non-state-owned companies to provide more collateral and higher-quality financial statements to reflect their operations. Therefore, non-state-owned enterprises have higher accounting conservatism than state-owned enterprises and relationship-specific transactions enable non-state-owned enterprises to deliver corporate information through private channels of communication and
suppliers and customers, with a much higher effect on transaction costs than state-owned enterprises(K. Zhu & Chen, 2006).
All these deficiencies have determined non-state-owned enterprises are more dependent on suppliers or customers, so in relationship-specific transactions, suppliers and customers have relative strong bargain power, and they are more likely to share information through private channels. However, state-owned enterprises have a relatively strong bargain power in relationship-specific transactions, and have to take all kinds of stakeholders and social responsibilities into account, which make them tend to have better accounting conservatism. Thus, the second hypothesis about relationship-specific transactions and accounting conservatism is as follows:
Hypothesis 2. In relationship-specific transactions, the negative correlations between
accounting conservatism of the company and degree of dependence between the company and suppliers or customers are more pronounced when the company is non-state owned company.
3 Methodology
3.1 Sample and data source
In order to avoid the impact of changes in accounting standards in China during 2007-2008 and the impact on the financial crisis during 2008-2009, this paper selected companies listed in Shanghai and Shenzhen stock markets from 2010 to 2016 as the sample. All the data are downloaded from CSMAR Database and WIND Financial Information Database.
From the database, we can know the proportion of sales revenue from top five customers to total revenue (hereinafter “proportion of top 5 customers”) and the proportion of expenditure on top five suppliers to the total expenditure (hereinafter “proportion of top 5 suppliers”) of these companies. And in this paper, samples are divided into three groups, (1) if a company only has the data of the proportion of top 5 suppliers, this company would be placed in SUPPLIER, and its data of top 5 suppliers would be used to do the analysis; (2) if a company only has the data of the proportion of top 5 customers, this company would be placed in CUSTOMER, and its data of top 5 customers would be used to do the analysis; (3) if a company has the data of the proportion of both top 5 suppliers and top 5 customers, this company would be placed in SUPPLIER, CUSTOMER and SUPPLIER/CUSTOMER, and its data of top 5 suppliers and top 5 customers would be used to do the analysis.
For further selection, the following types of companies were excluded, (1) The companies did IPO that year, because accounting surplus of these companies fluctuate greatly in that year. (2) ST1 and *ST2 companies, because of abnormal financial situations of such companies. (3)
Listed companies in financial and insurance sectors. (4) The companies which issue A-shares3, B-shares4 and H-shares5 at the same time. (5) Insolvent companies. (6) The
companies whose data are missing. Finally, the sample size of SUPPLIER is 4715, the
1 ST: Special treatment for abnormal financial status or other abnormal conditions.
2 *ST: The stock faces delisting risk.
3 A-shares: Common RMB shares. It is issued by companies within China, and it can be subscribed and traded by domestic
institutions, organizations or individuals (excluding Hong Kong, Macao and Taiwan investors) in RMB.
4 B-shares: Special RMB shares, which are denominated in RMB, subscribed and traded in foreign currencies, and are listed
and traded on the domestic Shanghai Stock Exchange and the Shenzhen Stock Exchange.
5 H-shares: Overseas-listed foreign-invested shares, issued and traded in Hong Kong by companies which listed in Mainland
sample size of CUSTOMER is 9643, the sample size of SUPPLIER/CUSTOMER is 1757. To lower the influence of extreme value, 1% samples of continuous variables were Winsorize2 treated. The data is processed through Stata15.0 and Excel.
But it should be noted, as for the data of the top five suppliers and customers, due to the statistical lack of the CSMAR database, and some companies did not disclose the data of supplier and customer, the samples of supplier are missing a lot, which results in a smaller classification is based on the “Guidelines for Industry Classification of Listed Companies (2012 Revision)” which is issued by the China Securities Regulatory Commission in 2012.
3.2 Variables and Model
3.2.1 Accounting Conservatism
Currently, there are following methods used to measure accounting conservatism: Accumulated surplus - stock compensation method and C-SCORE which is based on Basu Model. C. Zhang et al. (2012), Zheng and Y. Liu (2014) verified the applicability of the C-SCORE indicator in China and found that the company's annual accounting conservatism is highly consistent among different measurement methods. Besides, some other Chinese scholars, such as Sun, Zhao et al. (2010), Shen et al. (2013), and Wang et al. (2014) have used the C-SCORE indicator method to determine the influencing factors and economic consequences of accounting conservatism as well. The problem was empirically studied and the conclusions consistent with other methods were obtained through robustness tests. So, in this paper, we will use Khan and Watts (2009) C-SCORE indicators which is based on Basu model as a measure of the level of accounting conservatism, and do the robustness test for C-SCORE indicators through the Basu model. The calculation steps are as follows.
EPSi,t/Pi,t-1 = β0 + β1Di,t+ β2RETi,t + β3Di,t RETi,t+ ε (4.1)
Among them, EPSi,t represents the earnings per share in the year t of company i, and Pi,t-1
represents the stock closing price at the end of April of the year t-1. Di,t is a dummy variable,
when RETi,t ≤ 0, Di,t = 1; when RETi,t > 0, Di,t = 0. RETi,t stands for stock annualized yield
changes from May, year t to April, year t+1, and it is calculated as:
RETi,t = ∏𝑡 + 1, 4 𝑡, 5 (1+ri,j)-1 (4.2)
In the traditional Basu model, the stock return rate is positive for good news, and the stock return rate is negative for bad news. Therefore, β2 measures the timeliness of confirming the
influence of good news, and (β2 +β3) measures the timeliness of confirming the influence of
bad news. The differences between β2 and (β2 + β3) is increment β3, it represents the degree
of accounting conservatism. When the company has an accounting conservatism, the timeliness of confirming the influence of bad news is higher than that of good news, then, β3
is expected to be greater than zero.
Based on the Basu model and on the basis of linear function of company size, debt level, and book value of market capitalization, Khan and Watts (2009) reconstructed the β2 and β3,
G − SCORE = β2 = μ0 + μ1 SIZEi,t + μ2 MTBi,t + μ3 LEVi,t (4.3)
C − SCORE = β3 = λ0 + λ1 SIZEi,t + λ2 MTBi,t + λ3 LEVi,t (4.4)
Among them, SIZEi,t , MTBi,t , LEVi,t respectively represent the natural logarithm, price to
book ratio and asset-liability ratio of the assets in the year t of the company i.
Next, substitute (4.2), (4.3) and (4.4) into (4.1), and control the influence of SIZEi,t, MTBi,t,
and LEVi,t on the regression of the equation, and obtain (4.5):
EPSi,t/Pi,t-1 = β0 + β1Di,t + RETi,t (μ0 + μ1SIZEi,t + μ2MTBi,t + μ3LEVi,t) + Di,tRETi,t (λ0 + λ1
SIZEi,t + λ2MTBi,t + λ3LEVi,t ) + δ1SIZEi,t + δ2MTBi,t + δ3LEVi,t + δ4Di,tSIZEi,t +
δ5Di,tMTBi,t + δ6Di,tLEVi,t + εi (4.5)
Finally, regress the equation (4.5), and substitute the estimated values μ0 ~μ3 and λ0 ~λ3 into (4.3) and (4.4) respectively to get β2 and β3. β2 and β3 are G-SCORE and C-SCORE for
company i in the year t. Furthermore, G-SCORE describes how sensitive accounting surplus is to good news, and as described above, C-SCORE is an indicator for company’s accounting conservatism. The higher value of C-SCORE, the better accounting conservatism of the company.
3.2.2 Relationship-specific Transactions
According to previous research, the measurement of relational transactions is based on the proportion of the company's purchase to the large suppliers and the proportion of sales to the major customers. This method reflects both the importance of suppliers and customers to the company and the degree to which the company relies on suppliers and customers. When the company's major procurement and sales focus more on the top five suppliers and customers,
the company and suppliers/customers are more dependent on each other, thus, there would be more specific assets in relationship-specific transactions.
Therefore, the purchasing and selling ratio indirectly measures the relationship between the enterprise and the main supplier or customer, which is theoretically persuasive, but limited by the detailed degree of the disclosure of suppliers and customers in Chinese company’s annual report, and in this paper, the ratios of the top five suppliers and customers to the total purchase and sales are used as the measurement indexes of the relationship-specific transactions. Specifically as follows:
(1) Supplier procurement ratio (SUPPLIER), the company's top five suppliers purchase aggregate value accounted for the annual purchase ratio, measuring the relationship-specific transactions between suppliers and company.
(2) Customer sales ratio (CUSTOMER), the company's top five customer sales totals accounted for the annual sales ratio, measuring the relationship-specific transactions between customers and company.
3.2.3 Control Variables
As for control variables, this paper referred to Khan and Watts (2009), Zheng and Liu (2014), Charles Chen, Li, Su, and Yao (2008) and Yao (2009), selected following variables as control variables:
Table 3.1 Control Variables
Control Variable Variable Code Variable Definition
Proportion of institutional
shareholding INSHOW
Number of institutional holdings divided by total shares
Net cash flow of operating
activities CFO
The amount of cash flow generated by the enterprise's operating activities divided by total assets at the beginning of the year
Investment cycle CYCLE Accumulative depreciation divided by initial total assets
Growth ability GROWTH (operating profit this year – operating profit last year)/operating profit last year
Profitability ROA Net profit divided by initial total assets
Volatility VOLATILITY The standard deviation of the rate of return for a stock in a fiscal year
The age of the company AGE Years after the enterprise IPO
Audit firm BIG Dummy variable, if the audit firm is ‘big four’, values are 1, otherwise 0
The size of the company SIZE Natural logarithm of corporate total assets
Financial leverage LEV Corporate debt ratio, total liabilities divided by total assets
Industry Ind
Set the industry dummy variable according to the Industry Classification Standard of the China Securities Regulatory Commission
Year Year Annual dummy variable used to control macroeconomic effects
3.2.4 Model
This paper draws on the model of Yao (2009) and constructs the following formula (4.6) to test Hypothesis 1:
C − SCORE = α0 + α1SUPPLIER + α2CUSTOMER + α3INSOWN + α4CFO + α5CYCLE +
α6GROW + α7ROA + α8VOLIATILITY + α9AGE + α10BIG + α11LEV +
α12SIZE + 𝜑1 𝑖.𝑌𝑒𝑎𝑟 + 𝜑2 𝑖.Ind + ε (4.6)
Among them, α0 is a constant term, ε is a residual term, and the rest are the coefficients of the
Then, we introduce SOE, a variable of property right, with a value of 1 if the ultimate controller of the company is a state-owned shareholder, and 0 otherwise. Hypothesis 2 is tested by regression of equation (4.7) to investigate the effect of the interaction terms SOE_SUPPLIER and SOE_CUSTOMER on the C-SCORE indicator of accounting conservatism.
C – SCORE = α0 + α1SUPPLIER + α2SOE_SUPPLIER + α3CUSTOMER +
α4SOE_CUSTOMER + α5SOE + α6INSOWN + α7CFO + α8CYCLE +
α9GROWTH + α10ROA + α11VOLIATILITY + α12AGE + α13BIG +
α14SIZE + α15LEV + 𝜑1𝑖.𝑌𝑒𝑎𝑟 + 𝜑2i.Ind + ε (4.7)
4 Results and Discussion
4.1 Descriptive statisticsAs it is mentioned in section 3.1, this paper selected companies listed in Shanghai and Shenzhen stock markets from 2010 to 2016 as the sample and divided the sample into three groups, SUPPLIER, CUSTOMER and SUPPLIER/CUSTOMER. This section presents a descriptive statistic for these sample groups. Table 4.1 is the annual distribution of the sample. As it was mentioned earlier, due to the statistical lack of the CSMAR database, and as some companies did not disclose the data of top 5 suppliers and top 5 customers, the samples are missing a lot. These factors make the supplier data sample missing in the previous three years more. Due to the large sample size and the relatively uniform data distribution in the past three years, the sample size of suppliers in the first three years is expected to have no adverse impact on the regression results.
Table 4.1 Sample Annual Distribution
SUPPLIER CUSTOMER SUPPLIER/CUSTOMER Year Samples Proportion Samples Proportion Samples Proportion
2010 81 1.72% 196 2.03% 0 0 2011 108 2.29% 892 9.25% 38 2.16% 2012 199 4.22% 1,221 12.66% 32 1.82% 2013 457 9.69% 1,549 16.06% 110 6.26% 2014 1,207 25.60% 2,108 21.86% 508 28.91% 2015 1,277 27.08% 2,185 22.66% 544 30.96% 2016 1,386 29.40% 1,492 15.47% 525 29.88% Total 4,715 100.00% 9,643 100.00% 1757 100%
Table 4.2 shows the industry distribution of the sample companies. This paper uses the “Guidelines for Industry Classification of Listed Companies (2012 Revision)” issued by the China Securities Regulatory Commission in 2012 to sort statistics， which include totally 18 industries. Because accounting methods of enterprises in financial industry are different from those of ordinary enterprise, for the sake of consistency, this article removes the financial and insurance industries from the industry sample selection process. On the other hand, in the process of screening samples, we also considered the issue of the consistency of the industry distribution of sample of SUPPLIER and CUSTOMER. From the perspective of industry distribution, the number of samples in manufacturing industry is the largest, and the sample size of both SUPPLIER and CUSTOMER accounts for more than 50% of the samples. In the
second place is the wholesale/retail industry, with a sample size about 4.5%. This distribution is mainly based on two reasons: First, manufacturing companies tend to have closer relationships with customers and suppliers, and reflect the arguments of this paper more clearly. Second, such distribution characteristics are also basically consistent with the distribution characteristics of Chinese listed companies.
Table 4.2 Sample Industry Distribution
SUPPLIER CUSTOMER SUPPLIER/CUSTOMER Industry Samples Proportion Samples Proportion Samples Proportion
C Manufacturing Industry 3396 72.03% 6641 68.87% 1106 62.95% F Wholesale/Retail 211 4.48% 383 3.97% 106 6.03% I IT 205 4.35% 531 5.51% 38 2.16% K Real Estate 167 3.54% 394 4.09% 132 7.51% E Architectural Services/Construction 115 2.44% 243 2.52% 42 2.39% B Mining/Metallurgy 112 2.38% 231 2.40% 85 4.84% D Utilities/Energy 109 2.31% 289 3.00% 93 5.29% A Agriculture/Fishing/Forestry 87 1.85% 187 1.94% 35 1.99% G Transportation/Logistic/Distribution 77 1.63% 203 2.11% 25 1.42% S Others 77 1.63% 199 2.06% 42 2.39% L Business Services 55 1.17% 102 1.06% 13 0.74% N Environment Related 37 0.78% 57 0.59% 16 0.91% H Hospitality/Tourism/Restaurant 23 0.49% 30 0.31% 11 0.63% M Science/Research 21 0.45% 55 0.57% 0 0
R Entertainment/Leisure/Sports & Fitness 19 0.40% 79 0.82% 12 0.68%
Q Public Health/Medical/Healthcare 3 0.06% 5 0.05% 0 0
O Personal Care & Services 1 0.02% 12 0.12% 1 0.06%
P Education 0 0.00% 2 0.02% 0 0
Total 4715 100.00% 9643 100.00% 1757 100%
Table 4.3 is the descriptive statistics for variables in this paper. Among them, the average value of SUPPLIER is 0.372, the maximum value is 0.968, the minimum values is 0.037, and the first quartile is 0.205. This indicates that the transactions between the company and large suppliers are more frequent and the company is more dependent on the relationship-specific transactions with the top five suppliers. The average value of CUSTOMER is 0.314, the maximum value is 0.988, the minimum values is 0.013, and the first quartile is 0.145, it indicates that the dependency between the company and top five customers is also high, that is, the customers of listed companies tend to be more concentrated, and the company relies more on the relationship with large customers. This is in line with the data collected by most domestic scholars at the moment (Z. Lin, Zheng, Peng, et al., 2014; S. Wang et al., 2010; Zhang, Ma, & Zhang, 2012). Thus, it is feasible to measure relationship-specific transactions by adopting the procurement and sales ratios of top five suppliers and customers. The remaining grouping variables and control variables are also consistent with previous studies.
Table 4.3 Descriptive Statistics for Variables
VARIABLE Mean Std.Dev. Smallest 25% 50% 75% Largest
SUPPLIER 0.372 0.218 0.037 0.205 0.324 0.502 0.968 CUSTOMER 0.314 0.225 0.013 0.145 0.253 0.431 0.988 INSOWN 0.344 0.249 0.000 0.117 0.330 0.539 0.874 CFO 0.041 0.131 -0.202 0.000 0.041 0.087 0.254 CYCLE 0.077 1.892 0.000 0.008 0.027 0.062 0.478 ROA 0.049 0.976 -0.192 0.013 0.036 0.065 0.204 VOLATILITY 0.132 0.124 0.054 0.095 0.119 0.152 0.307 AGE 11.880 6.135 2.000 6.000 13.000 17.000 22.000 BIG 0.057 0.233 0.000 0.000 0.000 0.000 1.000 SOE 0.447 0.497 0.000 0.000 0.000 1.000 1.000
4.2 Correlation analysis
Table 4.4 Correlation analysis
CSCORE SUPPLIER CUSTOMER SOE MARKET INSOWN CFO CYCLE GROWTH ROA VOLATI~Y AGE BIG SIZE LEV CSCORE 1 SUPPLIER -0.182*** 1 CUSTOMER -0.122*** 0.354*** 1 SOE 0.267*** 0.00200 0.112*** 1 MARKET -0.0340 -0.071*** -0.074*** -0.160*** 1 INSOWN 0.151*** -0.0240 -0.0160 0.127*** 0.0140 1 CFO -0.095*** -0.0110 0.0360 0.00600 0.0290 0.071*** 1 CYCLE -0.553*** 0.196*** 0.104*** -0.201*** 0.0210 -0.147*** -0.066*** 1 GROWTH 0 -0.0160 0.00100 0.00400 0.0270 -0.0320 -0.0260 -0.049** 1 ROA -0.274*** -0.0370 -0.0190 -0.094*** 0.074*** 0.160*** 0.404*** -0.00200 0.00200 1 VOLATILITY -0.174*** 0.057** 0.109*** -0.0290 0.00500 -0.0370 -0.059** 0.124*** 0.00100 -0.047** 1 AGE 0.058** 0.123*** 0.0100 0.187*** -0.158*** -0.089*** -0.119*** 0.0360 -0.00200 -0.101*** 0.00800 1 BIG 0.071*** 0.0270 -0.00300 0.044* 0.050** 0.056** 0.0350 -0.068*** 0.0100 0.00800 -0.061** -0.0100 1 SIZE 0.735*** -0.195*** -0.102*** 0.288*** -0.058** 0.222*** 0.00700 -0.567*** 0.00700 -0.00200 -0.159*** 0.077*** 0.122*** 1 LEV 0.609*** -0.087*** 0.00200 0.151*** -0.051** 0.059** -0.223*** -0.183*** -0.0210 -0.403*** 0.0200 0.128*** -0.00900 0.442*** 1 *** p<0.01, ** p<0.05, * p<0.1
From Table 4.4 we can see, the correlation coefficients among the three major variables of SUPPLIER, CUSTOMER and accounting conservatism (C-SCORE) are -0.182 and -0.122 respectively, and at the same time, they are all highly significant at the 1% level of significance. Since the correlation between suppliers, customers, and accounting conservatism is less than 0.2, it can be considered that the correlation between them is low, which preliminarily excludes the possibility of multicollinearity.
On the other hand, both suppliers and customers exhibited significant negative correlations with accounting conservatism, consistent with the main hypotheses in this article. Most of the other control variables also show the correlation with the explanatory variables, this verifies the effectiveness of the equation setting. At the meanwhile, the correlation coefficient between the control variables is very small (basically less than 0.2). It can be considered that there is no strong correlation between other control variables. This shows that the equation has no significant multicollinearity problem, indicating that the model does not need to compensate and adjust for multicollinearity problems.
4.3 Multiple linear regression
This paper uses empirical Ordinary Least Square (OLS) multiple regression model (4.6) to test whether accounting conservatism of the company is negative related to the degree of dependence between the company and suppliers or customers. The results of regression test are shown as Table 4.5.
The results prove that at the 1% level of significance, both supplier and customer show a significant negative correlation with degree of accounting conservatism. When considering the supplier and customer’s affection on degree of accounting conservatism, the negative relationship is still significant at the 10% and 5% level of significance. Besides, due to the different sample size among these three sample groups, the value of the adjusted R-squared is different. However, the adjusted R-squared of all groups are approximately equal to or greater than 0.7, indicating that the model is suitable for this research question.
Based on these findings, it can be considered that with the increase of the dependency between the company and suppliers or customers, they tend to share information through private communication channels, thereby reducing the degree of reliance on disclosure of public information, which in turn lower the accounting conservatism. This conclusion is also valid when considering the influence of suppliers and customers at the same time, thus proving the hypothesis 1.
Table 4.5 Regression Analysis Results
(1) (2) (3)
VARIABLES SUPPLIER CUSTOMER SUPPLIER/ CUSTOMER SUPPLIER -0.006*** -0.004** (-2.71) (-1.97) CUSTOMER -0.009*** -0.006** (-5.44) (-2.53) INSOWN -0.005*** -0.008*** -0.005*** (-3.18) (-6.25) (-2.95) CFO 0.008 0.003 0.007 (1.00) (0.55) (0.87) CYCLE -0.093*** -0.092*** -0.092*** (-5.05) (-7.64) (-5.02) GROWTH -0.000 -0.000 -0.000 (-0.32) (-0.76) (-0.38) ROA -0.094*** -0.074*** -0.095*** (-6.15) (-5.96) (-6.20) VOLATILITY -0.060*** -0.067*** -0.055*** (-4.64) (-7.10) (-4.10) AGE -0.004*** -0.006*** -0.005*** (-4.05) (-6.09) (-4.39) BIG -0.001 -0.004*** -0.000 (-0.67) (-3.04) (-0.30) SIZE 0.013*** 0.012*** 0.013*** (15.62) (24.20) (15.30) LEV 0.043*** 0.042*** 0.043*** (8.26) (12.13) (8.13) Constant -0.238*** -0.219*** -0.236*** (-14.81) (-21.77) (-14.38) Observations 4,715 9,643 1,757 Adjusted R-squared 0.732 0.667 0.732 Robust t-statistics in parentheses
*** p<0.01, ** p<0.05, * p<0.1
This paper further introduced dummy variables of the property rights, and the interaction terms SOE_SUPPLIER and SOE_CUSTOMER to do the regression of (4.7) to test the hypothesis 2. The regression results are shown in Table 4.6.
It can be seen from the table that the level of accounting conservatism of state-owned and non-state-owned enterprises is negatively affected by relationship-specific transactions, but the degree of response is different. Among them, at the 1% level of significance, the negative influence of supplier's relationship-specific transactions on the state-owned enterprises' accounting conservatism is less than on the non-state-owned enterprises. When considering the mutual influence of suppliers and customers, at the same significant level of 1%, the reduction in state-owned enterprises' accounting conservatism is significantly weaker than
that of non-state-owned enterprises. However, the relationship-specific transactions with customers did not produce any influences on the level of accounting conservatism of state-owned and non-state-state-owned enterprises. This shows that when companies with different property rights conduct relationship-specific transactions with suppliers or customers, non-state-owned enterprises are vulnerable to the influence of suppliers because they share information through private channels and they have weaker bargain power in market transactions. Thus, the accounting conservatism is significantly reduced. Compared with non-state-owned enterprises, non-state-owned enterprises' relationship-specific transactions have a weaker influence on accounting conservatism.
Table 4.6 Regression Analysis Results (Property Right)
(1) (2) (3)
VARIABLES SUPPLIER CUSTOMER SUPPLIER/CUSTO MER SUPPLIER -0.012*** -0.008*** (-3.17) (-2.68) SOE_SUPPLIER 0.011*** 0.009*** (2.88) (2.67) CUSTOMER -0.011*** -0.008* (-3.64) (-1.73) SOE_CUSTOMER 0.004 0.003 (1.00) (0.57) SOE -0.001 0.001 -0.001 (-0.47) (1.18) (-0.36) INSOWN -0.007*** -0.009*** -0.006*** (-3.88) (-6.76) (-3.69) CFO 0.008 0.003 0.007 (0.94) (0.50) (0.77) CYCLE -0.091*** -0.090*** -0.089*** (-4.96) (-7.61) (-4.93) GROWTH -0.000 -0.000 -0.000 (-0.30) (-0.66) (-0.31) ROA -0.089*** -0.072*** -0.090*** (-5.85) (-5.82) (-5.86) VOLATILITY -0.060*** -0.067*** -0.054*** (-4.64) (-7.15) (-4.10) AGE -0.005*** -0.006*** -0.006*** (-4.49) (-6.49) (-4.87) BIG -0.001 -0.004*** -0.000 (-0.61) (-3.08) (-0.19) SIZE 0.012*** 0.012*** 0.012*** (15.54) (24.24) (15.18) LEV 0.044*** 0.042*** 0.044*** (8.57) (12.14) (8.37) Constant -0.230*** -0.215*** -0.227*** (-14.56) (-21.39) (-13.93) Observations 4,715 9,643 1,757 Adjusted R-squared 0.736 0.668 0.735 Robust t-statistics in parentheses
4.4 Robustness test
In order to ensure the reliability of using the C-SCORE indicator to measure the level of accounting conservatism, this paper divided the sample of the supplier and the customer into three groups of CS1, CS2, and CS3 according to the C-SCORE from low to the high, and adopt the Basu model to perform the regression test to measure the level of accounting conservatism, results are in Table 4.7.
Table 4.7 Robustness of C-SCORE
SUPPLIER CUSTOMER VARIABLES CS1 CS2 CS3 CS1 CS2 CS3 DR -0.004 -0.007 -0.001 0.001 -0.004 -0.000 (-0.97) (-1.18) (-0.11) (0.22) (-1.16) (-0.01) RET1 -0.001 0.001 0.003 0.003 0.005 0.013** (-0.47) (0.12) (0.45) (1.12) (1.51) (2.25) DR_RET 0.015 0.049** 0.084*** 0.049*** 0.057*** 0.082*** (1.12) (2.31) (3.24) (3.43) (4.60) (4.81) Constant 0.036*** -0.009 -0.105*** 0.022*** 0.007 -0.002 (4.44) (-0.50) (-3.39) (5.24) (1.07) (-0.19) Observations 597 600 599 1,401 1,405 1,402 Adjusted R-squared 0.080 0.045 0.175 0.065 0.081 0.148 Robust t-statistics in parentheses
*** p<0.01, ** p<0.05, * p<0.1
We found that in the groups of SUPPLIER and CUSTOMER, the accounting conservatism level calculated by the C-SCORE indicator is consistent with the Basu model. That is, with the C-SCORE indicator from low to high, the level of accounting conservatism (the regression coefficient of DR_RET) which are based on the Basu model also increases, and all
