21
Corporate suppliers and customers and accounting conservatism $ Kai Wai Hui a,1 , Sandy Klasa b,2 , P. Eric Yeung c,n a Department of Accounting, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong b Department of Finance, University of Arizona, Tucson, AZ 85721-0108, United States c J.M. Tull School of Accounting, University of Georgia, Athens, GA 30606-6252, United States article info Article history: Received 14 September 2009 Received in revised form 12 November 2011 Accepted 29 November 2011 JEL classification: M41 K12 D82 Keywords: Financial disclosures Conservatism Suppliers Customers abstract We argue that a firm’s suppliers and customers prefer it to account more conservatively due to information asymmetry and these stakeholders’ asymmetric payoffs with respect to the firm’s performance. We predict that a firm meets this demand for accounting conservatism when suppliers or customers have bargaining advantages over it that enable them to dictate terms of trade or whether trade occurs at all. We show that when a firm’s suppliers or customers have greater bargaining power, the firm recognizes losses more quickly. Our findings provide insights into how a firm’s powerful suppliers and customers are associated with its accounting practices and also support the contracting explanation for accounting conservatism. Published by Elsevier B.V. 1. Introduction A firm’s corporate suppliers and customers use its accounting performance to gauge its underlying economic performance. Suppliers and customers will consider the firm’s accounting performance when evaluating if it can meet its short-term trading obligations. Likewise, long-term suppliers and customers study the firm’s accounting performance to assess its long-term financial viability and the risk of making relationship-specific investments. For instance, a supplier involved in a long-term transaction with a firm might acquire specialized assets to manufacture specific products for the firm, while a customer involved in a long-term transaction may tailor its production technology to process products purchased from the firm. If the firm were to go out of business, the supplier and customer could face large switching costs. Suppliers and customers involved in long-term transactions with a firm will also consider its accounting performance when evaluating the risk that the firm might experience financial distress and no longer honor their implicit claims. For example, a supplier could be concerned with whether the firm can continue providing it with revenue from multi-year Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/jae Journal of Accounting and Economics 0165-4101/$ - see front matter Published by Elsevier B.V. doi:10.1016/j.jacceco.2011.11.007 $ We gratefully acknowledge the comments of S.P. Kothari (the editor), Sudipta Basu (the referee), Ashiq Ali, Steve Baginski, Linda Bamber, Dennis Beresford, Brian Cadman, Kevin Chen, Jenny Gaver, Gilles Hillary, John Jiang, Clive Lennox, Paul Zarowin, and seminar participants at the Chinese University of Hong Kong, the University of Georgia, and Singapore Nanyang Technological University. Sandy Klasa thanks the University of Arizona for the financial support provided by the McGuire Professorship. Eric Yeung gratefully acknowledges the financial support from the Terry-Sanford Research Award at the University of Georgia. n Corresponding author. Tel.: þ1 706 542 3615. E-mail addresses: [email protected] (K.W. Hui), [email protected] (S. Klasa), [email protected] (P.E. Yeung). 1 Tel.: þ852 2358 7563. 2 Tel.: þ1 520 621 8761. Journal of Accounting and Economics ] (]]]]) ]]]]]] Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal of Accounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Contents lists available at SciVerse ScienceDirect

Journal of Accounting and Economics

Journal of Accounting and Economics ] (]]]]) ]]]–]]]

0165-41

doi:10.1

$ We

Beresfo

Univers

financia

Awardn Corr

E-m1 Te2 Te

PleasAcco

journal homepage: www.elsevier.com/locate/jae

Corporate suppliers and customers and accounting conservatism$

Kai Wai Hui a,1, Sandy Klasa b,2, P. Eric Yeung c,n

a Department of Accounting, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kongb Department of Finance, University of Arizona, Tucson, AZ 85721-0108, United Statesc J.M. Tull School of Accounting, University of Georgia, Athens, GA 30606-6252, United States

a r t i c l e i n f o

Article history:

Received 14 September 2009

Received in revised form

12 November 2011

Accepted 29 November 2011

JEL classification:

M41

K12

D82

Keywords:

Financial disclosures

Conservatism

Suppliers

Customers

01/$ - see front matter Published by Elsevier

016/j.jacceco.2011.11.007

gratefully acknowledge the comments of S.

rd, Brian Cadman, Kevin Chen, Jenny Gaver,

ity of Hong Kong, the University of Georgia, an

l support provided by the McGuire Professo

at the University of Georgia.

esponding author. Tel.: þ1 706 542 3615.

ail addresses: [email protected] (K.W. Hui), sklasa

l.: þ852 2358 7563.

l.: þ1 520 621 8761.

e cite this article as: Hui, K.W., et alunting and Economics (2011), doi:1

a b s t r a c t

We argue that a firm’s suppliers and customers prefer it to account more conservatively

due to information asymmetry and these stakeholders’ asymmetric payoffs with respect

to the firm’s performance. We predict that a firm meets this demand for accounting

conservatism when suppliers or customers have bargaining advantages over it that enable

them to dictate terms of trade or whether trade occurs at all. We show that when a firm’s

suppliers or customers have greater bargaining power, the firm recognizes losses more

quickly. Our findings provide insights into how a firm’s powerful suppliers and customers

are associated with its accounting practices and also support the contracting explanation

for accounting conservatism.

Published by Elsevier B.V.

1. Introduction

A firm’s corporate suppliers and customers use its accounting performance to gauge its underlying economicperformance. Suppliers and customers will consider the firm’s accounting performance when evaluating if it can meetits short-term trading obligations. Likewise, long-term suppliers and customers study the firm’s accounting performanceto assess its long-term financial viability and the risk of making relationship-specific investments. For instance, a supplierinvolved in a long-term transaction with a firm might acquire specialized assets to manufacture specific products for thefirm, while a customer involved in a long-term transaction may tailor its production technology to process productspurchased from the firm. If the firm were to go out of business, the supplier and customer could face large switching costs.Suppliers and customers involved in long-term transactions with a firm will also consider its accounting performancewhen evaluating the risk that the firm might experience financial distress and no longer honor their implicit claims. Forexample, a supplier could be concerned with whether the firm can continue providing it with revenue from multi-year

B.V.

P. Kothari (the editor), Sudipta Basu (the referee), Ashiq Ali, Steve Baginski, Linda Bamber, Dennis

Gilles Hillary, John Jiang, Clive Lennox, Paul Zarowin, and seminar participants at the Chinese

d Singapore Nanyang Technological University. Sandy Klasa thanks the University of Arizona for the

rship. Eric Yeung gratefully acknowledges the financial support from the Terry-Sanford Research

@eller.arizona.edu (S. Klasa), [email protected] (P.E. Yeung).

., Corporate suppliers and customers and accounting conservatism. Journal of0.1016/j.jacceco.2011.11.007

Page 2: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]2

maintenance contracts on the products it sells to the firm, while a customer might be worried about whether the firm canprovide it with continued high quality spare parts and service for the products it buys from the firm (Titman and Wessels,1988; Bowen et al., 1995; Banerjee et al., 2008).3

We predict that the importance of a firm’s economic performance to its suppliers and customers leads to a demandfrom these stakeholders for the firm to report more conservatively, as measured by how quickly the firm recognizes badnews compared to good news (Basu, 1997). This demand is in part a consequence of managers having incentives to exploittheir asymmetrically informed position relative to other firm stakeholders by being less forthcoming about bad news thangood news. Like creditors, a firm’s suppliers and customers should prefer it to recognize bad news more quickly than goodnews as they could incur significant costs if the firm went out of business, but potentially gain little from the firmperforming above expected levels. Our argument that a firm’s suppliers and customers will prefer it to account moreconservatively leads to the study’s main hypothesis that a firm only meets the underlying demand for accountingconservatism from its suppliers or customers when these stakeholders have bargaining advantages over it that allow themto dictate terms of trade or whether trade occurs at all. This hypothesis is analogous to hypotheses made in the literatureexamining the debt-contracting process. For instance, Watts and Zimmerman (1986) and Watts (2003a, 2003b) predictthat because creditors usually have bargaining advantages over firms and thus dictate borrowing terms, one reason firmsadopt more conservative accounting policies is that accounting conservatism helps to limit creditors’ downside risk, whichin turn enables firms to obtain lower interest rates on their loans.

We use two firm-level measures and one industry-level measure of supplier or customer bargaining power to test ourpredictions. First, we expect that larger suppliers or customers have bargaining advantages over a firm and consequentlywe construct measures for the size of suppliers or customers relative to the firm. Second, we assume that a customer hasbargaining power if it accounts for a large fraction of a firm’s sales. Similarly, we expect that a particular supplier hasbargaining power if a firm buys a large fraction of its supplies from the supplier. Our industry-level measure for suppliers’or customers’ bargaining power is how concentrated suppliers’ or customers’ industries are.

We find support for our main prediction that if a firm’s suppliers or customers have relative bargaining advantages overit, there is greater conservatism in the firm’s financial reporting. Specifically, we document positive associations betweenthe bargaining power measures of suppliers and customers and the extent to which a firm reports bad news more quicklythan good news.

Subsequently, we examine whether the relations between the bargaining power measures of suppliers and customersand accounting conservatism for a firm are affected by the extent of a firm’s dependence on its suppliers or its customers’dependence on the firm. We assume that if a firm operates in an industry that uses a production technology requiringsignificantly more material inputs than labor inputs then the firm’s dependence on its suppliers is greater and that if thefirm’s customers operate in such industries then their dependence on the firm is larger. We predict that if the relationsinvolving the measures of suppliers’ bargaining power are driven by firms meeting suppliers’ demand for accountingconservatism when suppliers have bargaining power and the ability to set terms of trade, then these relations should bemore pronounced when a firm depends more on its suppliers. Likewise, we predict that if the relations involving themeasures for customers’ bargaining power are due to firms meeting customers’ demand for accounting conservatismwhen customers have bargaining power, then these relations should be less pronounced if customers’ dependence on afirm is greater and consequently their bargaining position relative to the firm is decreased. Consistent with our predictions,we find that the positive associations between the measures of the bargaining power of suppliers and a firm’s accountingconservatism are more pronounced when a firm operates in an industry with high material costs relative to labor costs.Also, the positive associations between the customer bargaining power measures and a firm’s accounting conservatism areless pronounced when a firm’s customers operate in such industries.

If a firm operates in a less fragmented industry with high barriers to entry, its resulting market power should decreasethe relative bargaining power of its suppliers and customers (Porter, 1980). Consequently, if the associations betweensuppliers’ and customers’ bargaining power and a firm’s accounting conservatism are caused by firms meeting thesestakeholders’ demand for accounting conservatism when they have greater bargaining power then we expect theseassociations to be less pronounced when a firm operates in an industry with high barriers to entry. We find that theseassociations are indeed less pronounced when a firm operates in an industry with high barriers to entry.

We also investigate if the associations between the measures of the bargaining power of suppliers and customers and afirm’s accounting conservatism are more pronounced when these stakeholders are likely to be involved in long-termtransactions with the firm and their demand for conservatism would consequently be greater. To evaluate whether a firmmay be involved in long-term transactions with its customers or suppliers, we consider whether a firm’s customers likelyrely on it for maintenance services and spare parts or its suppliers rely on it for future revenue via service contracts or thesale of spare parts. Consistent with expectations, we find that when suppliers or customers are more likely to be involved

3 Spekman (1988), Spekman et al. (2002) and Soosay et al. (2008) report that, in the United States over the last several decades, a large fraction of

firms have moved toward maintaining long-term relationships with their corporate suppliers and customers. Klein et al. (1978) argue that in many cases

such long-term relationships result in the possibility that one or more of the contracting parties will engage in post-contractual opportunistic behavior.

Coase (2000, 2006), however, is skeptical that Klein et al. (1978)’s use of the Fisher Body-General Motors case is an example of post-contractual

opportunistic behavior and argues that in some contexts such opportunism is not a major determinant of industrial organization.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 3: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 3

in long-term transactions with a firm, the positive associations between the measures of the bargaining power of suppliersand customers and a firm’s accounting conservatism are more pronounced.

Our study contributes to three streams of literature. First, our findings provide support to the contracting explanationfor accounting conservatism that the demand for conservatism in part occurs because it allows firms to structure moreefficient contracts with related parties (e.g., Watts, 2003a). Second, our paper increases the understanding of the impact ofthe characteristics of a firm’s suppliers and customers on its accounting practices. Specifically, our results suggest thatwhen a firm contracts with suppliers or customers that have bargaining power over it and likely dictate terms of trade orwhether trade occurs at all, the firm meets their demand for more conservatism in its financial reporting practices. Finally,our study broadly contributes to the literature that examines how product market competition influences accountingpolicy. Although the extent to which a firm competes for economic profits with suppliers and customers is an importantdeterminant of its survival (e.g., Porter, 1980; Perry, 1989), prior research on the impact of product market competition onaccounting policy focuses on the competitiveness of a firm’s own industry (i.e., Bamber and Cheon, 1998; Harris, 1998;Verrecchia and Weber, 2006; Ali et al., 2010). Our findings suggest that the bargaining power of a firm’s suppliers andcustomers is also an important determinant of its accounting practices.

The remainder of the paper is organized as follows. Section 2 develops hypotheses and discusses related empiricalpredictions. Section 3 discusses methodological issues. Section 4 presents our empirical results. Section 5 concludes.

2. Hypothesis development and empirical predictions

In this section, we provide arguments for why a firm’s suppliers and customers prefer/demand that it accounts moreconservatively. We hypothesize that firms are more likely to meet this demand when their suppliers or customers havebargaining advantages that allow them to dictate terms of trade or whether trade occurs at all. We also derive severalempirically testable predictions.

2.1. Contracting with suppliers and customers and accounting performance

An important facet of Coase’s (1937) transaction cost theory is that self-interested contracting parties have incentives tobehave opportunistically under uncertainty and information asymmetry. Also, as the nature of transactions gets more complexit becomes increasingly difficult to ex-ante write a complete contract to deter opportunism. Building on transaction cost theory,Williamson (1979) analyzes the governance structures of transactions, including transactions between a firm and its suppliersor customers. Williamson (1979, p. 234) argues that opportunism is a central concept in the study of transaction costs, and thatopportunism is especially relevant for economic activity involving transaction-specific investments in human and physicalcapital. He classifies transactions into three types. First, discrete transactions are well-defined agreements between contractingparties with easily verifiable performance and are short-term in nature. Second, long-term transactions are characterized ashaving incompletely specified contingent claims under uncertainty and are longer-term in nature. Finally, relational transactionsare also long-term in nature and involve sustainable on-going relationships between contracting parties that typically requirethe parties to make transaction-specific investments.

Accounting information can help alleviate the information problems associated with each type of supplier-customertransaction. First, suppliers and customers are concerned that a firm might not meet its short-term trading obligations,which are typically well defined in trading contracts (i.e., discrete transactions). For example, suppliers expect to receivecash payments after shipping products to the firm, and customers anticipate product deliveries after paying upfrontdeposits. The breach of such obligations can disrupt supply chains and cause losses for suppliers and customers. To theextent that a firm’s accounting performance signals its ability to meet future short-term contractual obligations, suppliersand customers can use a firm’s accounting information to appraise the risk of potential short-term trading losses.

Second, suppliers and customers are concerned with a firm’s ability to fulfill long-term implicit claims (i.e., state-dependentclaims in long-term contracts). For example, a customer of a durable goods producer expects to receive multi-year warrantyservices on the products it purchases from the producer and is therefore concerned with the producer’s ability to provide thislong-term support. Similarly, a durable goods producer expects to earn revenue from multi-year maintenance services on theproducts it sells to a particular firm and thus cares about the firm’s long-term prospects. Transaction cost theory suggests that itis very difficult to write explicit contracts that perfectly cover all contingencies associated with implicit claims in long-termcontracts. Ball (1989) proposes that accounting can be viewed as a specialist function for providing information that assists firmsin establishing their quasi-pricing system used for contracting purposes and that auditing is viewed as a specialist adjudicationfunction, used in completing contracts whose payoffs have been defined in terms of accounting numbers. In this functionalcompletion role of accounting, conservatism is desirable because it helps complete ex ante incomplete contracts by providingmore reliable quasi-prices.4 Furthermore, as stated earlier, transaction cost theory posits that, in many cases, there is a highlikelihood that the parties involved in long-term contracts engage in post-contractual opportunistic behavior. Klein et al. (1978)and Klein (2000) argue that long-term relation-specific investments create assets that have appropriable ‘quasi-rents’, which

4 We note that for private firms, accounting conservatism could be even more important given that these firms do not have publicly available signals

of firm performance, such as stock returns.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 4: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]4

encourage contracting parties to behave opportunistically and create hold-up problems. Prior studies (e.g., Telser, 1980; Klein andLeffler, 1981; Bull, 1987) propose that when deciding whether to transact with a particular firm or what terms to offer the firm,suppliers and customers place significant importance on the firm’s reputation for honoring stakeholders’ implicit claims. Bowenet al. (1995) and Burgstahler and Dichev (1997) argue that firms with higher reported earnings are perceived by their suppliersand customers as being more likely to remain financially viable in the long-term and to honor long-term implicit claims.

Third, a firm’s accounting performance is likely to be particularly important for suppliers and customers in the contextof relational transactions in which these firm stakeholders make significant, irreversible transaction-specific investments.For example, a supplier might sign long-term contracts with workers and acquire specialized assets to manufactureproduction inputs for a particular firm. Likewise, a customer that has signed a long-term contract with a firm mightacquire specialized machinery to process unique products made by the firm. Because such investments are oftentransaction-specific, the supplier and customer could face prohibitive switching costs if the firm were to go out of business.Thus, before entering into relational transactions and when considering whether to make additional investments intoexisting relationships with particular firms, suppliers and customers will likely carefully examine a firm’s accountingperformance to assess the prospects for the firm’s long-term financial viability.

2.2. Suppliers and customers and the demand for accounting conservatism

We argue that the importance of a firm’s economic performance for its suppliers and customers leads to a demand fromthese stakeholders for the firm to report more conservatively. This argument follows from the contracting rationale foraccounting conservatism (Watts, 2003a) that conservatism helps address moral hazard caused by parties to the firm havingasymmetric information and asymmetric payoffs. As such, our argument is in part analogous to the proposition that thedemand for accounting conservatism is partly due to it facilitating contracting between the firm and creditors. Similar tocontracting between firms and creditors, contracting between firms and their suppliers and customers is made difficult dueto agency problems resulting from information asymmetry. Further, like creditors, suppliers and customers have asymmetricpayoffs with respect to a firm’s performance. For instance, if the firm’s future performance is above expected levels,customers would gain little from this as long as at expected performance levels the firm would have fulfilled its contractualobligations to these stakeholders. Also, although a firm’s suppliers could benefit if it performs unusually well and this leadsto it buying more from suppliers, the expected upside to suppliers due to the firm’s strong performance could still be limited.In contrast, if a firm were to experience financial difficulty or even go out of business, suppliers or customers involved inshort-term discrete transactions, long-term transactions, or relational transactions with the firm could incur large costs if itreneged on its contractual obligations to them or did not honor these stakeholders’ implicit claims.

As pointed out by Basu (1997) and Watts (2003a), managers have incentives to exploit their asymmetrically informedposition relative to other stakeholders of the firm and be less forthcoming about bad news concerning firm performancethan good news. Because, like creditors, a firm’s suppliers and customers are likely more concerned with bad news about afirm’s performance than good news they should have a preference for the firm to report more conservatively in order tolearn about poor firm performance more quickly.5

We hypothesize that a firm is more likely to report conservatively when its suppliers and customers have bargainingadvantages over it and are consequently able to dictate terms of trade or whether trade with the firm occurs at all.6 Ourprediction regarding bargaining power is similar to predictions made in the literature examining the debt-contractingprocess. For example, this literature predicts and finds that because creditors typically have bargaining advantages overfirms and are able to set borrowing terms, firms that follow more conservative accounting policies and consequently limitcreditors’ downside risks receive loans with lower interest rates (Watts and Zimmerman, 1986; Ahmed et al., 2002; Watts,2003a, 2003b; Zhang, 2008). Our prediction is also supported by the evidence of more conservatism when the auditor hasgreater bargaining power. For example, Basu et al. (2001) find that firms audited by Big Eight auditors have greaterconditional conservatism, consistent with Big Eight auditors having bargaining advantages compared to non-Big Eightauditors and consequently being able to impose conservatism.7

Our central argument is that conservatism is one viable mechanism that suppliers and customers could use to protectthemselves when contracting with a firm. We note that there are other alternative contracting mechanisms that could

5 Suppliers and customers with sufficient bargaining power could impose universal stringent credit terms on trading partners and reduce their

reliance on a firm’s financial statements. However, uniform credit terms and a smaller reliance on firms’ financial statements would not necessarily help

suppliers and customers involved in long-term or relational transactions with a firm that need to assess the risk of the firm going out of business.

Likewise, uniform credit terms would not be ex-ante efficient in the case of short-term transactions in which suppliers and customers can mitigate a

higher risk of a firm not meeting its short-term trading obligations with worse terms of trade. This is analogous to the situation in which it would not be

efficient for creditors to impose stringent lending terms to all borrowers without regards to their individual financial positions.6 In making this prediction, we follow Ahmed et al. (2002) and Zhang (2008) and assume that a firm’s stakeholders can infer its commitment to

reporting conservatively by considering the firm’s reputation for having reported conservatively in the past. Similar to debt contracting, contracting with

suppliers and customers is a multi-period game and a firm would be penalized by its suppliers and customers if they discovered that although ex-ante

the firm appeared to be committed to following conservative accounting practices, ex-post the firm deviated from these practices.7 Francis and Krishnan (1999) show that Big Six auditors are more likely to practice reporting conservatism (i.e., more likely to issue modified audit

reports) than non-Big Six auditors. Ruddock et al. (2006), however, find little empirical support that higher than expected non-audit services compromise

auditors’ bargaining power and result in greater conservatism.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 5: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 5

alleviate concerns about short-term potential trading losses. For example, suppliers could protect themselves againstpotential short-term trading losses by using other contracting arrangements such as forfeiting deposits or contracting outcollection through discounting receivables with banks, factoring, or the securitization of receivables (Mian and Smith,1992). However, conservatism is viable because it also benefits other contracting parties (i.e., debt holders), whicheffectively reduces the cost of using it. In addition, the degree of conditional conservatism is largely under managementcontrol, while other kinds of mechanisms rely on costly contracting arrangements with external parties (i.e., involvingextensive negotiating, contracting, information gathering, and enforcement costs).8,9

2.3. Empirical predictions

As developed in the previous subsection, the principal hypothesis that we test is as follows:

Hypothesis. A firm meets the demand for conservative accounting practices from its suppliers and customers when thesestakeholders have relative bargaining advantages over the firm that enable them to dictate terms of trade or whether tradewith the firm occurs at all.

This hypothesis leads to empirical predictions that accounting conservatism, as measured by more timely recognitionof bad news than good news (Basu, 1997), is positively associated with measures for the bargaining power of suppliers andcustomers. In this paper, we use two firm-level measures and one industry-level measure of supplier or customerbargaining power. First, consistent with recent work (e.g., Schumacher, 1991; Snyder, 1996; Inderst and Wey, 2007), weexpect the size of suppliers and customers to be positively associated with their bargaining power. Therefore, we constructmeasures of the average market value of a firm’s suppliers or customers relative to the market value of the firm. Second,we follow Porter (1980) and assume that a particular customer has bargaining power if it accounts for a large fraction of afirm’s sales, while a particular supplier has bargaining power if it provides a firm with a significant fraction of its inputgoods. Our industry-level measure of suppliers’ and customers’ bargaining power is average industry concentration in theindustries of a firm’s suppliers or customers. Prior work predicts and finds that when a firm’s suppliers or customersoperate in more concentrated industries, the firm’s more limited choices with respect to its trading partners provides itssuppliers or customers with bargaining power relative to the firm (e.g., Galbraith, 1952; Lustgarten, 1975; Porter, 1980;Schumacher, 1991; Fee and Thomas, 2004).10 Using our three measures of the bargaining power of suppliers andcustomers, the main empirical prediction that we test in the paper is:

Prediction 1. There are positive associations between more timely recognition of bad news than good news for a firm andmeasures of the bargaining power of the firm’s suppliers or customers.

Porter (1980) argues that suppliers have bargaining power when their products are important inputs for a buyer’s business.As a result, we expect that if a firm operates in an industry that uses a production technology requiring significantly morematerial inputs than labor inputs, the firm’s dependence on its suppliers increases the suppliers’ bargaining power and thefirm’s sensitivity to suppliers’ preferences that it have greater accounting conservatism. On the other hand, if a firm’s customersoperate in industries that require more material than labor inputs, their dependence on the firm to supply material inputsshould decrease their bargaining position relative to the firm and make it less sensitive to their preferences for it to havegreater conservatism. Thus, as further tests of our principal hypothesis, we test the following two predictions:

Prediction 2a. The positive associations between more timely recognition of bad news than good news for a firm andmeasures of the bargaining power of its suppliers are more pronounced when the firm operates in an industry in whichmaterial costs are high relative to labor costs.

8 It could also be the case that suppliers and customers with bargaining power would ask a firm to provide them with additional internal or

operational data so that they could evaluate the risks of contracting with the firm. However, such requests might not necessarily result in a lower

demand from these stakeholders for a firm to have greater accounting conservatism. First, if such internal financial and operational data was provided to

suppliers or customers, it would not be audited on a timely basis by outside independent public accounting firms. Hence, this data would be susceptible

to manipulation by a firm. Second, internal operational data produced by different firms are likely to be idiosyncratic and are often not comparable across

peer firms, which could potentially make it difficult for suppliers or customers to evaluate the risk of a firm based on this data. Finally, for competitive

reasons, a firm might refuse to provide internal proprietary data to suppliers or customers that already have bargaining advantages over it, or if it

provides this data it may misreport it so that suppliers or customers are less likely to gain competitive advantages through their possession of this data

(Baiman and Rajan, 2002).9 Even in cases in which there is a high degree of cooperation between a supplier and customer, conditional conservatism could still be useful if at

least some degree of information asymmetry exists between the two parties. For instance, in cases in which the supplier and/or customer firm needs to

make additional relationship-specific investments, conditional conservatism could make it easier for each party to assess the downside risk of incurring

switching costs if the other party went out of business. To the extent that the supplier or customer firms might have concerns about making further

relationship-specific investments and that conditional conservatism could reduce these concerns, the use of conditional conservatism could increase the

probability that these investments would be made.10 We use an industry-level measure because even if we ex-post observe a firm contracts with just one supplier or customer, ex-ante the firm should

consider all firms in its suppliers’ or customers’ industries. Also, publicly traded U.S. manufacturing firms often transact with multiple suppliers and

customers within each of their suppliers’ and customers’ industries. Hence, the overall bargaining power of a firm’s suppliers or customers is likely a

function of multiple suppliers or customers in the same industry.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 6: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]6

Prediction 2b. The positive associations between more timely recognition of bad news than good news for a firm andmeasures of the bargaining power of its customers are less pronounced when the firm’s customers operate in industries inwhich material costs are high relative to labor costs.

As argued in Porter (1980), if a firm operates in a less fragmented industry this provides the firm with bargaining powerover suppliers and customers as they have limited trading choices. We therefore expect that a firm operating in a lessfragmented industry is not as sensitive to suppliers’ and customers’ preferences for it to have greater accountingconservatism. We use barriers to entry in a firm’s industry to proxy for the likelihood that the firm operates in an industrythat is less fragmented. An additional prediction that we test to examine the empirical validity of our main hypothesis is asfollows:

Prediction 3. The positive associations between more timely recognition of bad news than good news for a firm andmeasures of the bargaining power of its suppliers and customers are less pronounced when the firm operates in anindustry with higher barriers to entry.

Finally, we expect that when the demand for accounting conservatism from a firm’s suppliers or customers is greater,this leads to more pronounced associations between the measures of suppliers’ and customers’ bargaining power and afirm’s accounting conservatism. Firms in durable goods industries or firms with higher selling, general, and administrativeexpenses often manufacture unique products that require specialized servicing or spare parts (Titman and Wessels, 1988;Banerjee et al., 2008). Thus, customers of these firms rely on them to fulfill a large number of implicit claims such asservicing the products they have bought from a firm or having parts available for these products, while suppliersproducing more specialized products rely on their customers to honor implicit claims such as providing them with incomefrom maintenance services or the sale of spare parts. Hence, customers of firms that produce unique goods will likely havestrong preferences that these firms report conservatively so that they can more easily evaluate these firms’ prospects forlong-term financial viability. Likewise, suppliers of unique goods will likely have strong preferences that their customersreport more conservatively.11 Thus, as additional tests of our principal hypothesis, we test the two following predictions:

Prediction 4a. The positive associations between more timely recognition of bad news than good news for a firm andmeasures of the bargaining power of its suppliers are more pronounced if a firm’s suppliers produce unique goods.

Prediction 4b. The positive associations between more timely recognition of bad news than good news for a firm andmeasures of the bargaining power of its customers are more pronounced if the firm produces unique goods.

3. Data

3.1. Sample

We study Compustat manufacturing firms (i.e., firms with Compustat primary six-digit North American IndustryClassification System (NAICS) codes between 311111–338999) over the 1995–2007 period. The sample period of1995–2007 is chosen because our data on industry concentration is collected from Census of Manufactures publications forthe 1997 and 2002 U.S. Economic Censuses. We focus on firms in the manufacturing sector for four reasons. First, typicalmanufacturing firms have important relationships with their corporate suppliers and customers, while in other sectors this isoften not the case. For instance, firms in the retailing sectors do not typically sell products to corporate customers, while firmsin the agricultural and mining sectors do not rely on major materials suppliers. Second, in the manufacturing sector whether afirm reports conservatively is likely to be of relevance to its suppliers or customers, as firms in this sector often make largeinvestments to set up production facilities and are consequently concerned about long-term contractual relationships withother firms in the supply chain. In contrast, switching suppliers or customers is relatively easier for firms in more laborintensive sectors, such as service sectors, and thus firms in such sectors may be less concerned about contractual relationshipswith other firms. Third, data for our industry-level measure of bargaining power, the industry concentration ratio, is widelyavailable only for industries in the manufacturing sector. Finally, the industry materials to labor costs ratio that we construct totest empirical predictions 2a and 2b is only available for manufacturing firms.

11 We consider whether a firm’s suppliers are durable goods producers or the firm itself is such a producer because this proxies for whether there

may be a larger demand from a firm’s suppliers or customers for it to have greater accounting conservatism. However, we acknowledge that whether a

firm is a durable goods producer could also impact its market power and consequently its bargaining power. For instance, Coase (1972) argues that

durable goods monopolists are only able to enjoy market power for a limited period of time because once an initial amount of output is sold the

monopolist will need to cut its price to obtain further profits and this continues until price falls to marginal cost. However, a number of subsequent

papers make counter-arguments suggesting that durable goods monopolists are able to maintain a considerable amount of market power over time. For

instance, Stokey (1981), Bulow (1982), and Kahn (1986) report that in cases in which production capacity is limited or marginal costs are increasing, a

durable goods monopolist is able to keep at least some of its market power over time.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 7: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 7

3.2. Proxies for bargaining power

To construct our first firm-level bargaining power measure for suppliers and customers resulting from relative size(Relative Size), we calculate the average market value of the companies in a firm’s supplier or customer industries (definedas six-digit NAICS industries) and divide this by the firm’s own market value. In order to create our second firm-levelmeasure for whether a firm has a customer that accounts for a large fraction of its sales or it buys an important fraction ofits input goods from a given supplier (Major Stakeholder), we first verify in the Compustat Segments file whether a firmreports that it has a customer that accounts for a significant fraction of its sales. Next, we collect the names of thesecustomers and manually check whether they are included on the Compustat database, and if so, if they are manufacturingfirms. If we determine that, during a particular year, one of our sample firms has a significant customer that is a Compustatmanufacturing firm, we set an indicator variable equal to one for that firm having a significant manufacturing customerthat year, otherwise the indicator variable takes on a value of zero. To examine whether during a particular year a firm hasan important supplier, we first manually check whether during any year over our sample period the firm is listed as animportant customer of another Compustat manufacturing firm. If so, we next use data from the Compustat Segments file tocalculate the ratio of the value of the materials it buys from this supplier to the firm’s cost of goods sold that year. If thisratio is at least 5% we set an indicator variable equal to one for the firm having an important supplier that year, otherwisethe indicator variable takes a value of zero.

Our industry-level measure of bargaining power (Concentration Ratio) is suppliers’ and customers’ industry concentra-tion, measured as the Herfindahl-Hirschman index. We collect industry concentration ratios from the 1997 and 2002Census of Manufactures publications. The Census of Manufactures is published during years in which a U.S. Census takesplace. The industry concentration measures in these publications are calculated from data collected from U.S. CensusSurveys. All private and public firms in the U.S. are required by federal law to respond to these surveys (Title 13 of the U.S.code). Ali et al. (2009) show that industry concentration ratios calculated by the U.S. Census, which are calculated withdata from both public and private firms in an industry, more accurately reflect market structure in an industry than doratios constructed with data from only the Compustat database, which is comprised almost entirely of publicly-tradedfirms.12,13 Following prior work such as Kovenock and Phillips (1997), Mackay and Phillips (2005), Campello (2003, 2006),Haushalter et al. (2007), Klasa et al. (2009), and Ali et al. (2009, 2010), we assume that the U.S. Census industryconcentration data are valid for windows of years surrounding a U.S. Economic Census year. Specifically, we assume thatthe values of the 1997 and 2002 industry concentration measures are valid for the 1995–1999 and 2000–2007 periods.

3.3. Measure of accounting conservatism

To examine the relation between accounting conservatism and the bargaining power of suppliers and customers, weuse a measure for the conservatism of a firm’s accounting practices based on Basu’s (1997) asymmetric earnings timelinessmeasure (Timeliness). This measure considers the relation between earnings and stock returns and captures the extentto which news about economic losses are recognized on a more timely basis in accounting earnings than news abouteconomic gains. Following Basu (1997), we measure conservatism as the degree of asymmetric timeliness of accountinggains and losses as reflected in stock returns. Basu’s measure is particularly relevant in our context because a firm’ssuppliers or customers are likely more concerned with bad news about a firm’s performance than good news about thisperformance. Furthermore, the timely recognition of bad news counters managers’ incentives to engage in earningsmanagement (e.g., LaFond and Watts, 2008; LaFond and Roychowdhury, 2008).14

Empirically, we measure Basu’s (1997) asymmetric timeliness of earnings by running firm-specific time-seriesregressions using rolling windows of at least the preceding seven years for each firm-year:

Xit=Pit�1 ¼ a0þa1DRitþa2RETitþa3DR� RETitþeit ð1Þ

where X is the earnings per share, P is the price at the beginning of the year, RET is the annual stock return, and DR is adummy equal to one if RET is negative, and zero otherwise. The asymmetric timeliness measure is calculated as the ratio of(a2þa3)/a2, which captures the timeliness of earnings in recognizing bad news (as reflected in stock returns) relative to the

12 Ali et al. (2009) show that when using U.S. Census industry concentration measures, there is evidence consistent with theoretical predictions that

more-concentrated industries are populated by fewer and larger firms with higher price-cost margins. However, using industry concentration measures

calculated with only Compustat data, they are unable to find such results. Further, Ali et al. (2009) find that the correlation of Compustat-based industry

concentration measures and U.S. Census-based measures is only 13%.13 Census-based industry concentration ratios are calculated for six-digit NAICS industries within the manufacturing sector. We note that prior to

1997 the Census of Manufactures reported industry concentration measures for four-digit SIC industries rather than for six-digit NAICS industries. Given

that industry concentration measures calculated at the four-digit SIC and six-digit NAICS levels are not directly comparable, we do not use data from

Census of Manufactures publications that were published prior to 1997.14 We focus on conditional conservatism primarily because it captures new information received by management in each period that is important for

contracting purposes, while unconditional conservatism only utilizes information known at the inception of an asset’s life and incorporates little new

information (Basu, 2005). In addition, Ball and Shivakumar (2005) argue that unconditional conservatism adds noise to payoffs to contracting parties and

thus could reduce contracting efficiency. Conditional conservatism measured by the timely recognition of bad news is particularly powerful in testing

contracting-based explanations for conservatism. Consistent with these arguments, Qiang (2007) finds that contracting induces only conditional

conservatism and that unconditional conservatism is in part induced by taxation and regulation.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 8: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]8

timeliness of earnings in recognizing good news (Basu, 1997; Pope and Walker, 1999; Givoly and Hayn, 2000; Francis et al.,2004). We expect that Timeliness on average is greater than one if accounting is conservative on average for manufacturingfirms during our sample period.

3.4. Corporate supplier and customer data

We hand-collect data on a firm’s suppliers and customers from 2008 IBISWorld reports. IBISWorld is an independentpublisher of U.S. industry research, and its annual reports cover over 700 different market segments. To create annualreports, IBISWorld analysts use a variety of sources ranging from government, company, and industry association statistics,to company reports, surveys, and personal contact. Each report provides information about an industry’s product segments(e.g., products and geographic spread), market characteristics (e.g., overall size and competitors within the industry),industry conditions (e.g., barriers to entry, regulation, cost structure, and technology and systems), and other information.

For each industry group at the five-digit NAICS level, IBISWorld also reports supplier and customer industry linkages atthe five-digit NAICS industry group level. For example, the suppliers for hardware manufacturing (five-digit NAICS code:33251) include industry groups such as metal pipes (33121) and copper rolling drawing (33142), while the customersinclude industry groups such as automobile manufacturing (33611), aircraft manufacturing (33641), and furnituremanufacturing (33721). For each firm, we first identify its primary five-digit NAICS industry group based on Compustat,and then collect data on its supplier and customer industry groups at the five-digit NAICS level based on IBISWorld. Thebargaining power proxies for suppliers and customers, Size Ratio and Concentration Ratio, are measured at the six-digitNAICS industry level within the identified five-digit NAICS industry groups.15

After data requirements to calculate accounting conservatism measures, supplier and customer bargaining powerproxies, and control variables used in our multivariate regressions (see details in Section 4), we are left with a final sampleof 7,064 firm-years over the 1995–2007 period. Table 1 shows the number of observations within each three-digit NAICSindustry group. Similar to the Compustat population, our sample consists of large proportions of computer and electronicproducts producers (30%), chemicals producers (20%), and machinery producers (12%).

Table 1 also reports the average number of six-digit NAICS supplier and customer industries for each six-digit NAICSindustry within a three-digit NAICS industry group. Among these industry groups, transportation equipment producershave on average the greatest number of supplier industries (32.3), while computer and electronic products producers haveon average the greatest number of customer industries (244.7). On the other hand, primary metal producers have onaverage the fewest supplier industries (4.3), while petroleum and coal products producers have on average the fewestcustomer industries (1.8).16

4. Empirical results

4.1. Univariate analysis

Panel A of Table 2 presents for three-digit NAICS industry groups mean values of the conservatism measure (Timeliness)and the bargaining power proxies (Relative Size, Major Stakeholder, and Concentration Ratio).17 This panel shows that firmsin the leather and allied products, food, and electrical equipment industry groups have relatively high conservatism(Timeliness42), while firms in the apparel, textile product mills, and petroleum and coal industry groups have relativelylow conservatism (Timelinesso0). This panel also shows that there is no clear pattern of correlations among the three

15 Given that IBISWorld reports supplier and customer industry linkages at the five-digit NAICS industry level and that we know the component six-

digit NAICS industries within a five-digit NAICS industry, Relative Size can be constructed at the five- or six-digit industry level. On the other hand,

because Census of Manufactures publications define industry at the six-digit NAICS level and from the data in these publications it is not possible to

construct industry concentration measures for five-digit NAICS industries, we have to collect suppliers’ and customers’ industry concentration ratios (the

Herfindahl-Hirschman index) at the six-digit NAICS level. We then average concentration ratios for all supplier or customer industries for a given firm.

Our multivariate results using the industry concentration based measure are qualitatively the same when we use a measure calculated by first

approximating individual suppliers’ or customers’ industry concentration ratios at the five-digit NAICS level (i.e., the average value for all six-digit NAICS

industries within a five-digit NAICS industry) and then averaging these ratios across all of a firm’s five-digit NAICS supplier or customer industries.16 If we calculate the average number of five-digit NAICS supplier and customer industries for each five-digit NAICS industry within a three-digit

NAICS industry group we find that again transportation equipment producers have on average the greatest number of supplier industries (8.3) and

computer and electronic products producers have on average the greatest number of customer industries (90.8). Using the five-digit NAICS classification

we also document that plastics and rubber products have on average the fewest supplier industries (1.4) and that firms in the printing and related

support industries have on average the fewest customer industries (1.0).17 As documented in Panel A of Table 2, the average size of a firm’s suppliers and customers relative to that of the firm is typically much higher than

one. This is driven by the right-tail skewness of the distribution of market values for a firm’s suppliers and customers. We note that, on average, the

median market value of a firm’s suppliers and customers is typically close to the market value of the firm and that our multivariate results are robust to

calculating the relative size variable using the median market value of a firm’s suppliers or customers. However, we believe calculating the relative size

variable using the mean market value of a firm’s suppliers and customers makes the most sense, because the mean value captures the effect on a firm’s

accounting conservatism of the firm having a few suppliers or customers that are considerably larger than it and are likely to have substantial influence

over it.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 9: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 1Sample distribution and number of supplier and customer industries across three-digit NAICS industry groups.

Three-digit

NAICS code

Industry group name No. of

firm-years

% of

Sample

% of

Compustat

No. of supplier

industries

No. of customer

industries

311 Food 293 4.15 3.66 11.8 6.9

313 Textile Mills 64 0.91 0.77 5.5 24.8

314 Textile Product Mills 32 0.45 0.24 13.5 6.5

315 Apparel 88 1.25 1.54 13.9 14.0

316 Leather and Allied Products 72 1.02 0.80 6.7 7.9

321 Wood Products 76 1.08 0.87 13.1 12.6

322 Paper 208 2.94 2.23 10.7 28.2

323 Printing and Related Support 78 1.10 0.89 10.4 10.0

324 Petroleum and Coal Products 159 2.25 1.74 8.2 1.8

325 Chemicals 1,385 19.61 23.02 11.7 12.1

326 Plastics and Rubber Products 189 2.68 2.67 4.5 10.7

327 Nonmetallic Mineral Products 116 1.64 1.52 11.0 20.7

331 Primary Metal 262 3.71 3.19 4.3 14.3

332 Fabricated Metal Products 383 5.42 3.57 17.7 22.3

333 Machinery 837 11.85 10.08 30.8 26.0

334 Computer and Electronic Products 2,129 30.14 33.67 15.0 244.7

335 Electrical Equipment, Appliances,

and Components

359 5.08 3.91 18.6 17.8

336 Transportation Equipment 334 4.73 5.63 32.3 155.5

Total 7,064 100.00 100.00

This table reports the number of firm-year observations within each three-digit NAICS industry group in our final sample, as well as the fraction of our

sample and of all Compustat manufacturing firms in each particular three-digit NAICS industry group. This table also reports for each three-digit NAICS

industry group the mean values of the number of supplier and customer industries (defined at the six-digit NAICS level) for a particular six-digit NAICS

industry.

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 9

bargaining power proxies. This suggests that the three proxies for bargaining power of suppliers and customers capturedifferent aspects of the overall bargaining power of a firm’s suppliers and customers.18

Panel B of Table 2 presents univariate correlations between the conservatism measure (Timeliness) and the bargainingpower proxies. Pearson correlations are reported above the diagonal, while Spearman correlations are reported below thediagonal. Consistent with the main hypothesis of the paper and Prediction 1, we find significant positive correlationsbetween Timeliness and Relative Size of suppliers and between Timeliness and Concentration Ratio of supplier industries.Panel B of Table 2 also shows that the correlations of Timeliness with Relative Size, Major Stakeholder, and Concentration

Ratio for customers are significant and positive, which further supports our main hypothesis and Prediction 1.

4.2. Multivariate model and descriptive statistics of variables

We use the following empirical Ordinary Least Square (OLS) multiple regression model to test whether accountingconservatism is related to the proxies for suppliers’ or customers’ bargaining power. This model controls for otherdeterminants of accounting conservatism:

Timelinessit ¼ b0þb1Supplier Bargaining Poweritþb2Customer Bargaining Poweritþb3Sizeit

þb4Market-to-Bookitþb5Leverageitþb6CFOStditþb7Payoutitþb8Def ault Riskitþb9Spreadit

þb10Litigationitþb11Cash Flowitþb12Self HHIitþb13Import Penetrationitþb14Unionit

þb15BigNAudititþb16USGovernmentitþX

t

btYeartþeit ð2Þ

In this model Supplier Bargaining Power and Customer Bargaining Power represent the alternative bargaining power proxiesfor suppliers and customers (Relative Size, Major Stakeholder, and Concentration Ratio). We control for Size (measured as themarket value of the firm) because prior studies show that large firms report less conservatively (Giner and Rees, 2001;LaFond and Watts, 2008; LaFond and Roychowdhury, 2008). We include Market-to-Book (measured as market valuedivided by book value of equity) in the regression because firms with a high market-to-book ratio are expected to havelarger growth opportunities and therefore are likely to face higher contracting costs (Watts, 2003a). This implies a positiveassociation between the firm-specific timeliness measure of conservatism and the market-to-book ratio.

18 Recall that Table 1 presents the average number of supplier and customer industries, which does not correspond with the concentration of each

supplier and customer industry. Nevertheless, the number of supplier and customer industries could be related to the firm’s choices of choosing

contracting parties, which in turn affects conservatism. There is some evidence that for industries that have fewer supplier and/or customer industries

conditional conservatism is higher. For instance, food (311) and leather and allied products (316) industries have the most conservative earnings

(Timeliness42.5) and they have few customer industries (o8).

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 10: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 2Univariate analysis of conservatism and bargaining power proxies of suppliers and customers.

Panel A: Average values of conservatism measure and bargaining power proxies for each three-digit NAICS industry group

Three-digit

NAICS code

Industry group name Suppliers Customers

Timeliness Relative

Size

Major

Stakeholder

Concentration

Ratio

Relative

Size

Major

Stakeholder

Concentration

Ratio

311 Food 2.65 20.91 0.01 853.10 53.62 0.02 786.21

313 Textile Mills 0.02 16.17 0.00 875.06 20.37 0.06 554.41

314 Textile Product Mills �0.47 31.31 0.03 519.74 80.55 0.16 1443.36

315 Apparel �1.62 11.24 0.02 469.90 50.80 0.00 1174.64

316 Leather and Allied Products 2.96 14.74 0.01 418.16 71.79 0.01 1096.28

321 Wood products 1.12 7.75 0.05 335.83 4.54 0.00 327.12

322 Paper 0.33 15.88 0.01 986.47 35.15 0.03 880.68

323 Printing and Related Support 1.08 29.07 0.00 746.55 17.40 0.09 45.61

324 Petroleum and Coal Products �0.26 5.26 0.01 1031.17 2.05 0.01 1292.99

325 Chemicals 1.69 59.33 0.04 1298.97 59.59 0.10 1263.56

326 Plastics and Rubber Products 1.84 42.51 0.01 381.00 163.35 0.06 1440.68

327 Nonmetallic Mineral Products 0.02 11.90 0.03 602.06 9.79 0.05 577.51

331 Primary Metal 1.46 21.73 0.03 1331.41 110.57 0.05 804.40

332 Fabricated Metal Products 0.68 36.48 0.05 512.72 59.70 0.08 961.20

333 Machinery 1.54 23.29 0.04 530.66 28.39 0.09 838.18

334 Computer and Electronic Products 1.90 26.47 0.07 790.31 58.47 0.14 495.43

335 Electrical Equipment, Appliances, and

Components

2.14 47.04 0.02 524.43 34.45 0.04 899.28

336 Transportation Equipment 1.28 20.65 0.03 958.84 94.54 0.18 1005.11

Panel B: Univariate correlation

Pearson (above the diagonal) and

Spearman (below the diagonal) correlations

Suppliers Customers

Timeliness Relative

Size

Major

Stakeholder

Concentration

Ratio

Relative

Size

Major

Stakeholder

Concentration

Ratio

Timeliness 0.053 0.017 0.030 0.043 0.031 0.027Relative Size of (Suppliers) 0.061 �0.014 �0.002 0.297 �0.013 0.039Major Stakeholder (Suppliers) 0.006 0.012 0.009 �0.010 0.290 �0.025Concentration Ratio (of Supplier Industries) 0.043 0.011 0.019 0.035 0.040 �0.136Relative Size (of Customers) 0.069 0.794 0.013 0.032 0.001 0.218Major Stakeholder (Customers) 0.032 0.028 0.291 0.059 0.056 �0.004

Concentration Ratio (of Customer Industries) 0.020 0.054 0.001 0.228 0.117 0.022

Panel A of this table presents average values of the conservatism measure (Timeliness) and the three bargaining power proxies (Relative Size, Major

Stakeholder, and Concentration Ratio) across supplier and customer firms in each three-digit NAICS industry group. Panel B presents Pearson (above the

diagonal) and Spearman (below the diagonal) correlations between the conservatism measure and the bargaining power proxies. Bold figures indicate

significance at the two-tailed 10% level.

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]10

We expect that the demand from bondholders for conservatism arising from shareholder-bondholder conflicts is higher ifa firm has more financial leverage. Consequently, we partly control for this demand with Leverage, a firm’s long-term debtscaled by book assets. Because conflicts of interest between shareholders and bondholders are also expected to be greaterwhen the volatility of a firm’s cash flows is higher or payouts to shareholders are larger, we control for CFOStd, the standarddeviation of cash flows over the preceding six years (Basu, 1995; Givoly and Hayn, 2000), and Payout, the annual amount ofcommon dividends and stock repurchases scaled by book assets (Watts, 1993; Ahmed et al., 2002).19 While we expect apositive coefficient on each of these three variables, we note that Payout could be negatively associated with conservatismbecause firms that pay high dividends could have large free cash flows and a lower level of economic growth opportunities(Ahmed et al., 2002). We also control for DefaultRisk, which is a continuous variable estimated as in Vassalou and Xing (2004)that measures a firm’s default probability. Controlling for a firm’s default risk is important because shareholder-bondholderconflicts of interest and the demand from bondholders for a firm to report conservatively should be greater if the likelihoodof the firm defaulting on its debt is larger. Also, suppliers’ and customers’ preferences for a firm to report conservativelyshould be greater if the firm faces a larger default risk. We expect a positive coefficient on DefaultRisk.

We expect a positive coefficient on Spread, the average daily bid-ask spread of stock price over a year for a firm, because thebid-ask spread is expected to be wider for firms that suffer from more information asymmetry with investors and the demand

19 Another reason to control for Leverage and CFOStd is that the asymmetric timeliness regression overestimates conditional conservatism in the

presence of debt (Beaver and Ryan, 2009).

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 11: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 11

for conservatism increases with information asymmetry about a firm’s prospects (LaFond and Watts, 2008). We further controlfor litigation risk with the variable Litigation, a variable that is calculated as in Rogers and Stocken (2005). Litigation is expectedto be positively associated with conservatism (Beaver 1993; Watts, 1993, 2003a; Basu, 1997). We also include as a controlvariable CashFlow, cash flow from operations scaled by book assets, in our models. For firms that have lower cash flows, furtherreductions in reported cash flows due to conservative accounting practices could be costly and thus make firms with lowercash flows less likely to follow conservative accounting policies. We predict a positive coefficient on CashFlow.

We further control for the firm’s own industry concentration ratio, as captured by the Herfindahl-Hirschman index (SelfHHI),because firms in more concentrated industries could be more likely to hide bad news which would lead to lower accountingconservatism (Dhaliwal et al., 2009). In addition, it could also be the case that if firms in more concentrated industries facehigher regulatory and political costs due to their market power, then these firms might have greater unconditional conservatismin an attempt to understate their profits and reduce such costs (e.g., Basu, 2005; Qiang 2007). Because of the general negativeassociation between conditional and unconditional conservatism, this could also lead to a predicted negative coefficient onSelfHHI. Due to the fact that our industry concentration measures are calculated using only data for U.S. firms, we also include asa control the intensity of global competition in a firm’s industry. Specifically, we control for the import penetration ratio in afirm’s industry (ImportPenetration). Following the international trade literature (e.g., Trefler, 1996; Bernard et al., 2006), thisvariable is defined as the value of imports divided by domestic absorption (value of shipments minus value of exports plus valueof imports).20 We control for the proportion of the workers in a firm’s industry that are unionized (Union) because the higher isthis proportion, the more likely it is that a firm has to contract with a union that has significant bargaining power.21 Unions withbargaining power could pressure management to invest in fewer high-risk projects, which could reduce creditors’ demand forconservatism (Farber et al., 2010). Alternatively, firms might follow more conservative accounting practices in the presence of apowerful union in an attempt to bargain for lower wages (Liberty and Zimmerman, 1986; Leung et al., 2009).

We control for the impact of auditors’ bargaining power on conservatism (Basu et al., 2001; Krishnan, 2005) with anindicator variable for if a firm is audited by a Big N auditor (BigNAudit) and year dummies because general legal scrutinyvaries over time (Basu, 1997). Finally, we control for whether a firm has the U.S. government (USGovernment) as animportant customer. We construct this variable by manually searching through names of important customers of a firmprovided in the Compustat Segments file and verifying whether an important customer is the U.S. government.22 Due topolitical concerns, firms that have the U.S. government as a major customer may choose to report more conservatively.Thus, we predict a positive coefficient on USGovernment.

Panel A of Table 3 reports univariate statistics for the dependent and independent variables appearing in Eq. (2). Weobserve significant variation in the asymmetric timeliness measure. The mean value is 1.512, indicating that earnings onaverage recognize bad news (as reflected in stock returns) 1.5 time faster than good news. We also observe considerablevariation in the size of suppliers and customers relative to the size of the firm. Panel A also documents that 4.1% (9.5%) ofour sample firms have an important supplier (customer).23 Finally, this panel provides evidence suggesting that there isconsiderable heterogeneity across firms in terms of the average industry concentration ratios of supplier and customerindustries. Panel B in Table 3 provides variable definitions. Because we use variants of Eq. (2) throughout the study, forbrevity we do not repeat variable definitions in subsequent tables.

4.3. Multiple regression findings

4.3.1. The relation between accounting conservatism and the bargaining power of a firm’s suppliers and customers

Table 4 provides the results of OLS regressions that test the main empirical prediction of the study, that conservatism infirms’ accounting practices is positively associated with the measures for the bargaining power of their suppliers andcustomers. The first model is a baseline model that does not include the bargaining power measures as independentvariables. The results for the second model show a significant positive coefficient on Relative Size for suppliers (0.235,t¼2.13) and a significant positive coefficient on Relative Size for customers (0.103, t¼2.23). These findings suggest that theextent to which a firm has more timely recognition of bad news than good news is greater when the size of suppliersor customers relative to its own size is larger and these stakeholders consequently have bargaining power over the firm.

20 The import penetration variable is calculated at the three-digit NAICS industry level. Data on the value of shipments for three-digit NAICS

manufacturing industries are obtained from the U.S. Census Bureau. Data on imports and exports for three-digit NAICS industries are available online

from TradeStats Express, which is maintained by the International Trade Administration of the U.S. Department of Commerce.21 Data on annual industry unionization rates are obtained from the Union Membership and Coverage Database, which is publicly available at

www.unionstats.com. This database reports industry unionization rates for three-digit Census Industry Classification (CIC) industries. These rates

represent the percentage of total workers in a CIC industry that are covered by unions in collective bargaining agreements. We are able to determine

which four-digit SIC industry codes correspond to three-digit CIC codes and hence, are able to assign industry unionization rates to our sample firms. The

data we use are for the 1995–2005 period. We assume the data for 2005 are valid for the years 2006 and 2007.22 To create the indicator variable for whether a firm has the U.S. government as a major customer, we identify firms in the Compustat Segments

Customer file for whom the customer type (Variable CTYPE) equals ‘‘GOVDOM’’. We next manually check the customer name to ensure that the customer

is a department of the U.S. government.23 Although we do not tabulate this, we also find that for the sample firms that have an important customer the mean (median) percentage of a firm’s

sales that is accounted for by the large customer is 20.8% (14.0%). Likewise, for the sample firms that have an important supplier the mean (median)

percentage of a firm’s cost of goods sold that is accounted for by purchases from the important supplier is 29% (23.3%).

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 12: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 3Descriptive statistics of variables of interest.

Panel A: Descriptive statistics

Mean STDEV 25% 50% 75%

Measure of conservatism

Timeliness 1.512 6.226 �0.556 1.580 3.504

Supplier Bargaining Power Proxies

Relative Size 0.297 0.772 0.005 0.033 0.191

Major Stakeholder 0.041 0.198 0.000 0.000 0.000

Concentration Ratio 0.815 0.419 0.484 0.719 1.027

Customer Bargaining Power Proxies

Relative Size 0.528 1.630 0.005 0.040 0.257

Major Stakeholder 0.095 0.294 0.000 0.000 0.000

Concentration Ratio 0.899 0.660 0.438 0.788 1.197

Control variables

Size 5.861 2.487 4.017 5.757 7.533

Market-to-Book 2.589 2.096 1.313 1.978 3.018

Leverage 0.153 0.143 0.020 0.129 0.244

CFOStd 0.068 0.052 0.033 0.053 0.084

Payout 0.036 0.053 0.000 0.015 0.045

DefaultRisk 0.031 0.128 0.000 0.000 0.000

Spread 0.025 0.058 0.004 0.013 0.032

Litigation 0.017 0.030 0.002 0.006 0.017

CashFlow 0.088 0.084 0.043 0.088 0.134

SelfHHI 0.807 0.710 0.288 0.592 1.024

ImportPenetration 0.278 0.165 0.131 0.258 0.386

Union 12.884 9.464 6.400 10.500 16.700

BigNaudit 0.818 0.386 1.000 1.000 1.000

USGovernment 0.043 0.205 0.000 0.000 0.000

Panel B: Variable definitions

Variables Definition

Timeliness Basu’s (1997) asymmetric timeliness measure of conservatism: estimated by running firm-specific time-series regressions

using rolling windows (at minimum the preceding seven years) for each firm-year as: X/

P¼a0þa1DRþa2RETþa3DR�RETþe, where X is earnings per share, P is the price at the beginning of the year, RET is annual

stock return, DR is dummy equal to one if RET is negative. The timeliness measure is calculated as the ratio of (a2þa3)/a2

Relative Size (Suppliers) Ratio of average market value of individual firms in supplier industries to the market value of the firm, divided by 100

Relative Size (Customers) Ratio of average market value of individual firms in customer industries to the market value of the firm, divided by 100

Major Stakeholder

(Suppliers)

Equals one if a firm’s purchases from a particular manufacturing supplier account for at least five percent of its cost of

goods sold in a given year, and zero otherwise

Major Stakeholder

(Customers)

Equals one if a firm discloses that it has a major manufacturing customer, and zero otherwise

Concentration Ratio

(Suppliers)

For a six-digit NAICS industry, average of its supplier industries’ Herfindahl-Hirschman index values, deflated by 1,000

Concentration Ratio

(Customers)

For a six-digit NAICS industry, average of its customer industries’ Herfindahl-Hirschman index values, deflated by 1,000

Size Natural log of market value of equity (data25�data199)

Market-to-Book Market (data24�data199) to book (data60) ratio (observations with negative book equity are dropped)

Leverage Long-term debt (data9) divided by total assets (data6)

CFOStd The standard deviation of CashFlow defined as cash flow from operations (data308) divided by total assets (data6),

measured over prior six years

Payout Common dividends (data21) plus stock repurchases (collected from the Securities Data Company database) divided by

total assets (data6)

DefaultRisk Default risk is calculated following Vassalou and Xing (2004), Eq. (9)

Spread The average daily closing bid-ask spread during the year in the CRSP daily file

Litigation The litigation risk measure is calculated following the Rogers and Stocken (2005), Eq. (4), which predicts the probability of

a securities class action lawsuit during a period using a vector of variables including market value, daily share volume,

beta, stock returns, daily stock return volatility, skewness of daily returns, minimum of daily returns, and litigation

industry indicators

CashFlow The ratio of operating cash flow (data308) to total assets (data6)

SelfHHI The Herfindahl-Hirschman index value of the industry in which the firm operates, deflated by 1,000

ImportPenetration The import penetration ratio is defined as the value of imports divided by domestic absorption (value of shipments minus

value of exports plus value of imports) at the three-digit NAICS industry level

Union Percentage of unionized workers in a firm’s three-digit CIC (census industry classification) industry

BigNaudit Equals one if the auditor is one of the big four (or five) auditing firms, and zero otherwise

USGovernment Equals one if the firm has the US government as a major customer, and zero otherwise

Ratio variables calculated with Compustat data are winsorized at 1st and 99th percentiles of their sample distributions.

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]12

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 13: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 4The effect of supplier and customer bargaining power on a firm’s timely recognition of bad news.

Bargaining Power Measures

Pred. Sign Baseline Regression Relative Size Major Stakeholder Concentration Ratio

Main effects

Supplier Bargaining Power þ 0.235nn 2.117nnn 0.430nnn

(2.13) (2.71) (2.76)

Customer Bargaining Power þ 0.103nn 1.523nnn 0.409nnn

(2.23) (3.04) (4.17)

Control variables

Size – �0.202nnn�0.131nn

�0.201nnn�0.197nnn

(�3.80) (�2.23) (�3.79) (�3.68)

Market-to-Book þ 0.082n 0.076n 0.082n 0.081n

(1.84) (1.71) (1.85) (1.82)

Leverage þ 0.975n 0.872 0.965n 1.123nn

(1.73) (1.56) (1.72) (2.00)

CFOStd þ �2.698 �2.556 �2.681 �2.372

(�1.33) (�1.25) (�1.32) (�1.16)

Payout ? �1.277 �1.362 �1.209 �1.370

(�0.82) (�0.87) (�0.78) (�0.88)

DefaultRisk þ 0.534nn 0.409n 0.534nn 0.517nn

(2.38) (1.79) (2.38) (2.29)

Spread þ 1.354n 1.026 1.359n 1.490nn

(1.78) (1.05) (1.78) (2.04)

Litigation þ 2.235nnn 2.012nnn 2.293nnn 2.308nnn

(3.36) (3.00) (3.48) (3.47)

CashFlow þ 2.165n 2.272nn 2.212nn 2.305nn

(1.94) (2.04) (1.99) (2.06)

SelfHHI – �0.181n�0.182n

�0.201n�0.141

(�1.66) (�1.67) (�1.85) (�1.30)

ImportPenetration ? 0.290 0.443 0.273 0.572

(0.56) (0.84) (0.52) (1.11)

Union ? �0.010 �0.010 �0.010 �0.012

(�1.35) (�1.34) (�1.35) (�1.54)

BigNAudit þ 0.208 0.274 0.213 0.223

(0.98) (1.28) (1.01) (1.05)

USGovernment þ 0.667nn 0.671nn 0.724nn 0.659nn

(2.05) (2.06) (2.21) (2.02)

Year-fixed effects Yes Yes Yes Yes

F-statistic 11.08nnn 9.44nnn 19.08nnn

Adj. R2 (%) 11.99 12.46 12.55 12.56

N 7,064 7,064 7,064 7,064

All variables are defined in Panel B of Table 3. All t-statistics (in parentheses) are based on Huber/White/sandwich robust standard errors and adjusted for

clustering of observations by firm. nnn, nn, and n indicate two-tailed significance at 1%, 5%, and 10% significance levels, respectively. F-statistics are for the

joint significance of test variables.

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 13

The findings for the third model show that when the bargaining power measure is Major Stakeholder that the estimatedcoefficient on this measure is significantly positive if a firm has an important supplier (2.117, t¼2.71) and is alsosignificantly positive if a firm has an important customer (1.523, t¼3.04). This suggests that if a firm relies on a supplierfor a large amount of its input materials or a particular customer accounts for a large fraction of its sales then thebargaining power of these stakeholders leads the firm to report more conservatively. The last column reports the resultswhen the bargaining power proxy is Concentration Ratio. As predicted, we find significant and positive coefficients onConcentration Ratio for suppliers (0.430, t¼2.76) and for customers (0.409, t¼4.17). These results indicate that a firm’stimely recognition of bad news increases with the concentration of its suppliers’ and customers’ industries. Overall, theresults in Table 4 provide strong support for our hypothesis that a firm’s accounting conservatism is positively associatedwith the bargaining power of its suppliers and customers.

The coefficients on most of the control variables in Table 4 models are significant with expected signs. For instance, theconservatism measure is negatively associated with Size and positively associated with Market-to-Book.24 Also, consistent

24 A positive association between Basu’s (1997) timeliness measure and the market-to-book ratio is consistent with theory. However, some prior

empirical studies find a negative association between the market-to-book ratio and the timeliness measure because, over a short horizon, prior

unrecognized increases in asset values (i.e., greater market-to-book ratio) reduce the necessity to recognize asset value losses (Basu, 2001;

Roychowdhury and Watts, 2007). We do not have this ‘buffer problem’ and find evidence consistent with theory because we estimate timeliness of

earnings using a long rolling window prior to the event year (Roychowdhury and Watts, 2007; Ball et al., 2011). However, if we use the lagged market-to-

book ratio we find a negative association between the timeliness measure and the market-to-book ratio.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 14: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]14

with the potential for larger stockholder-bondholder conflicts of interest leading to greater accounting conservatism, wefind significant positive coefficients on Leverage and DefaultRisk. In addition, we find that conservatism is positivelyassociated with the proxies for information asymmetry (Spread) and litigation (Litigation). We also find a positivecoefficient on CashFlow, suggesting that following conservative accounting practices is less costly for firms with greatercash flows. Finally, consistent with political motives leading firms that have the U.S. government as an important customerto report more conservatively, we find a significant positive coefficient on USGovernment.

A comparison of the adjusted R2 values for the first model with those for the other three models show that the adjustedR2 values of Models 2–4 are approximately 3.9–4.8% greater than that of the base-line model that does not include thesupplier and customer bargaining power measures. These increases are roughly equivalent to those when we include size,leverage, and cash flow in the baseline model. Thus, including the supplier and customer bargaining power measures inour model that explains conditional conservatism increases the explanatory power of the model. Likewise, the results ofF-tests indicate that the supplier and customer bargaining power proxies are jointly significant in our models.

4.3.2. The effect of material purchases on the bargaining power of suppliers and customers

Table 5 reports evidence on whether industry production technology modifies the relations between a firm’s accountingconservatism and the bargaining power of its suppliers and customers in predictable ways (Predictions 2a and 2b). Prediction 2astates that if a firm operates in an industry that uses a production technology requiring significantly higher material inputs thanlabor inputs, the firm’s dependence on its suppliers makes it more sensitive to their preference that it have greater accountingconservatism, leading to a more pronounced association between timely recognition of bad news and measures for thebargaining power of suppliers. Prediction 2b, on the other hand, states that if material inputs relative to labor inputs are large incustomers’ industries then the dependence of a firm’s customers on it for materials reduces their bargaining power relative tothe firm. This would then lead to a less pronounced association between timely recognition of bad news and measures for thebargaining power of customers. To obtain empirical estimates for the importance of material and labor costs in an industry, wecollect from Annual Survey of Manufactures publications, which are published by the U.S. Census, necessary information tocalculate the ratio of materials costs to labor costs in six-digit NAICS industries in the manufacturing sector.

The models in Table 5 are similar to Eq. (2), except that we include as independent variables, (i) HiML, an indicatorvariable that equals one if a firm’s industry materials costs to labor costs ratio is greater than that of the sample median forall manufacturing industries in a given year, and zero otherwise, (ii) HiCustomerML, an indicator variable that equals one ifthe mean value of the materials costs to labor costs ratio in a firm’s customer industries is greater than that of the samplemedian for all manufacturing industries in a given year, and zero otherwise, (iii) the interaction of HiML with proxies for thebargaining power of suppliers, and (iv) the interaction of HiCustomerML with proxies for the bargaining power of customers.

The results in the first column in this table show a positive coefficient on the supplier interaction variable, Relative

Size�HiML (0.414, t¼2.40). This is consistent with the notion that when a firm relies more on its suppliers for materials,this increases its sensitivity to their preference that it have greater accounting conservatism. On the other hand, we find anegative coefficient on the customer interaction variable (�0.211, t¼�2.07), which suggests that when the bargainingpower of customers is reduced due to customers’ dependence on a firm for large amounts of material inputs, customers’diminished ability to dictate terms of trade reduces the firm’s incentive to follow more conservative accounting practices.

In the second and third models in Table 5, we replace Relative Size with Major Stakeholder and Concentration Ratio,respectively, and interact them with HiML for supplier industries and HiCustomerML for customer industries. Consistentwith the results in the first column, we find that HiML leads to significantly more positive coefficients on Major Stakeholder

and Concentration Ratio (3.852, t¼3.58; 0.348, t¼3.35). Also, HiCustomerML results in significantly more negativecoefficients on Major Stakeholder and Concentration Ratio (�1.369, t¼�1.74; �0.167, t¼�2.15). These results are furtherevidence consistent with Prediction 2a that the more a firm relies on its suppliers for materials, the greater is its sensitivityto their preferences for it to have more accounting conservatism. Likewise, these findings provide additional support forPrediction 2b that the more customers rely on a firm for materials, the lower is their bargaining power relative to the firm,which reduces the firm’s incentives to meet their demand for conservatism. Finally, a comparison of the adjusted R2 valuesof the Table 5 models with the value for the baseline model in Table 4 shows that the Table 5 values are approximately 6.2–8.9% larger than that for the Table 4 baseline model. Likewise, when comparing the adjusted R2 value for each of the threemodels in Table 5 with that for the respective model in Table 4 that uses the same supplier and customer bargaining powermeasures, we find that the adjusted R2 values of the Table 5 models are 2.2–4.0% greater than those of the Table 4 models.

4.3.3. The effect of barriers to entry on the bargaining power of suppliers and customers

We report in Table 6 the findings on whether high barriers to entry in a firm’s industry decrease the bargaining powerof suppliers and customers and result in a less pronounced positive association between the conservatism measure andproxies for suppliers’ and customers’ bargaining power (Prediction 3). Barriers-to-entry in an industry are calculated as inKaruna (2007), as the weighted average gross value of the cost of property, plant, and equipment for firms in an industryweighted by each firm’s sales market share in the industry. The models in Table 6 are similar to Eq. (2), except that weinclude as independent variables HiBTE, an indicator variable that equals one if barriers-to-entry in a firm’s industry aregreater than that of the sample median for all manufacturing industries in a given year, and zero otherwise, andinteractions of HiBTE with the supplier and customer bargaining power measures. We note that the Pearson correlationbetween HiBTE and Market-to-Book is low and only �0.058 for our sample.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 15: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 5Industry material purchases and the effect of supplier and customer bargaining power on a firm’s timely recognition of bad news.

Bargaining Power Measures

Pred. Sign Relative Size Major Stakeholder Concentration Ratio

Main effects

Supplier Bargaining Power þ 0.016 0.189 0.259nn

(0.02) (0.40) (2.01)

Customer Bargaining Power þ 0.202n 2.161nnn 0.481nn

(1.94) (2.79) (2.16)

Interaction variables

Supplier Bargaining Power�HiML þ 0.414nn 3.852nnn 0.348nnn

(2.40) (3.58) (3.35)

Customer Bargaining Power�HiCustomerML – �0.211nn�1.369n

�0.167nn

(�2.07) (�1.74) (�2.15)

Control variables

HiML ? 0.242 0.311n�0.813nn

(0.71) (1.67) (�2.12)

HiCustomerML ? 0.554 0.465nnn 0.601nnn

(1.57) (3.15) (2.94)

Size – �0.143nn�0.212nnn

�0.194nnn

(�2.40) (-3.94) (-3.62)

Market-to-Book þ 0.089nn 0.089nn 0.083n

(2.00) (2.03) (1.88)

Leverage þ 0.905 1.039n 1.143nn

(1.61) (1.85) (2.03)

CFOStd þ �2.817 �2.924 �2.073

(�1.37) (�1.43) (�1.02)

Payout ? �1.112 �0.813 �1.260

(�0.71) (�0.53) (�0.81)

DefaultRisk þ 0.375 0.472nn 0.475nn

(1.63) (2.08) (2.11)

Spread þ 0.978 1.144 1.335n

(1.02) (1.41) (1.69)

Litigation þ 1.973nnn 2.205nnn 2.138nnn

(2.94) (3.34) (3.22)

CashFlow þ 2.325nn 2.273nn 2.163nn

(2.09) (2.03) (1.96)

SelfHHI – �0.177 �0.210n�0.224nn

(�1.60) (�1.91) (�2.00)

ImportPenetration ? 0.774 0.595 1.057nn

(1.48) (1.14) (1.98)

Union ? �0.015nn�0.015nn

�0.013

(�1.97) (�1.99) (�1.61)

BigNAudit þ 0.240 0.225 0.230

(1.11) (1.06) (1.09)

USGovernment þ 0.686nn 0.732nn 0.505

(2.08) (2.21) (1.45)

Year-fixed effects Yes Yes Yes

F-statistic 8.82nnn 11.08nnn 15.51nnn

Adj. R2 (%) 12.73 13.03 13.06

N 7,064 7,064 7,064

HiML equals one if the value for the material costs to labor costs ratio for a firm’s six-digit NAICS industry is above the sample median for all industries in

a given year, and equals zero otherwise. Material costs to labor costs for firms is collected from US Census Annual Survey of Manufactures publications.

HiCustomerML equals one if the mean value of the material costs/labor costs ratio of a firm’s six-digit NAICS customer industries is above the sample

median for all firms in a given year, and equals zero otherwise. All other variables are defined in Panel B of Table 3. All t-statistics (in parentheses) are

based on Huber/White/sandwich robust standard errors and adjusted for clustering of observations by firm. ***, **, and * indicate two-tailed significance

at 1%, 5%, and 10% significance levels, respectively. F-statistics are for the joint significance of test variables.

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 15

The findings reported in Table 6 are consistent with Prediction 3. Specifically, the results in the first column showsignificant negative coefficients on the interaction terms Relative Size�HiBTE for suppliers (�0.398, t¼�2.10) and Relative

Size�HiBTE for customers (�0.199, t¼�2.41), suggesting that when a firm conducts its business in an industry withhigher barriers to entry, this reduces the bargaining advantages of suppliers and customers over the firm and the extent towhich they are able to dictate terms of trade. As a result, the firm will be less concerned with suppliers’ or customers’preferences for accounting conservatism. The findings in the second column in Table 6 document that the coefficients onthe interaction of HiBTE with Major Stakeholder for both suppliers and customers are significantly negative (�2.802,t¼�1.74; �1.684, t¼�1.73). Likewise, the results in the third column indicate that the coefficients on the interaction of

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 16: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 6Barriers-to-entry and the effect of supplier and customer bargaining power on a firm’s timely recognition of bad news.

Bargaining Power Measures

Pred. Sign Relative Size Major Stakeholder Concentration Ratio

Main effects

Supplier Bargaining Power þ 0.452nnn 4.011nnn 0.962nnn

(2.86) (3.07) (3.69)

Customer Bargaining Power þ 0.198nnn 2.631nnn 0.659nnn

(3.12) (3.36) (4.59)

Interaction variables

Supplier Bargaining Power�HiBTE – �0.398nn�2.802n

�0.980nnn

(�2.10) (�1.74) (�2.99)

Customer Bargaining Power�HiBTE – �0.199nn�1.684n

�0.475nn

(�2.41) (�1.73) (�2.47)

Control variables

HiBTE ? 0.176 3.427nn 1.533nnn

(0.82) (2.11) (4.75)

Size – �0.131nnn�0.231nnn

�0.224nnn

(�2.21) (-4.27) (-4.12)

Market-to-Book þ 0.077n 0.083n 0.084n

(1.73) (1.87) (1.88)

Leverage þ 0.895 0.926n 1.089n

(1.60) (1.65) (1.95)

CFOStd þ �2.701 -3.214 �2.974

(�1.31) (�1.58) (�1.44)

Payout ? �1.423 �0.834 �0.828

(�0.91) (�0.53) (�0.53)

DefaultRisk þ 0.394n 0.402n 0.427n

(1.72) (1.76) (1.86)

Spread þ 0.955 1.169 1.270n

(0.93) (1.44) (1.67)

Litigation þ 2.135nnn 2.085nnn 2.038nnn

(3.19) (3.15) (3.04)

CashFlow þ 2.139n 2.419nn 2.387nn

(1.91) (2.19) (2.13)

SelfHHI – �0.185 �0.282nn�0.234nn

(�1.16) (�2.53) (�2.09)

ImportPenetration ? 0.342 0.241 0.575

(0.65) (0.45) (1.11)

Union ? �0.010 �0.013n�0.016nn

(�1.30) (�1.73) (�2.06)

BigNAudit þ 0.312 0.245 0.214

(1.45) (1.15) (1.01)

USGovernment þ 0.657nn 0.774nn 0.660nn

(2.02) (2.38) (2.03)

Year-fixed effects Yes Yes Yes

F-statistic 9.80nnn 4.36nn 12.22nnn

Adj. R2 (%) 12.62 12.81 12.86

N 7,064 7,064 7,064

HiBTE equals one if the value for barriers-to-entry in a firm’s industry is above the sample median in a given year, and equals zero otherwise. Barriers-to-

entry are calculated as in Karuna (2007), and all other variables are defined in Panel B of Table 3. All t-statistics (in parentheses) are based on Huber/

White/sandwich robust standard errors and adjusted for clustering of observations by firm. ***, **, and * indicate two-tailed significance at 1%, 5%, and

10% significance levels, respectively. F-statistics are for the joint significance of test variables.

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]16

HiBTE with Concentration Ratio for both suppliers and customers are also significantly negative (�0.980, t¼�2.99;�0.475, t¼�2.47). These results provide further support for Prediction 3. Further, comparing the adjusted R2 values of theTable 6 models with the value for the baseline model in Table 4 shows that the Table 6 values are approximately 5.3–7.3%larger than that for the Table 4 baseline model. Similarly, when comparing the adjusted R2 values of each of the threeTable 6 models with that for the respective model in Table 4 that uses the same supplier and customer bargaining powermeasures, we find that the adjusted R2 values of the Table 6 models are 1.3–2.4% greater than those for the Table 4 models.

4.3.4. The effect of specialized products on suppliers’ and customers’ demand for accounting conservatism

In Table 7 we test Prediction 4a that associations between timely recognition of bad news and the supplier bargainingpower measures are more positive if a firm’s suppliers tend to be durable goods producers or they have higher selling,general, and administrative expenses and Prediction 4b that associations between timely recognition of bad news and

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 17: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

Table 7Durable/specialized goods and the effect of supplier and customer bargaining power on a firm’s timely recognition of bad news.

Bargaining Power Measures

Pred. Sign Relative Size Major Stakeholder Concentration Ratio

Main effects

Supplier Bargaining Power þ �0.154 �0.032 0.630 0.184 �0.527 0.306

(�1.30) (�0.19) (0.56) (0.27) (�1.61) (0.64)

Customer Bargaining Power þ 0.016 0.013 1.433nnn 0.782 0.194nn 0.171n

(0.63) (0.31) (2.65) (1.15) (2.41) (1.74)

Interaction variables

Supplier Bargaining Power�DDurSupplier þ 0.845nnn 3.223nnn 1.892nnn

(4.50) (2.73) (3.29)

Customer Bargaining Power�DDur þ 0.134nnn 0.273n 0.236nnn

(3.97) (1.70) (2.81)

Supplier Bargaining Power�HiSupplierSGA þ 0.519nn 4.084nnn 0.312nn

(2.15) (3.67) (2.22)

Customer Bargaining Power�HiSGA þ 0.195nn 1.396n 0.380nn

(2.24) (1.94) (2.44)

Control variables

DDurSupplier ? �0.806nnn�3.731nnn

�0.519nn

(�4.38) (�2.94) (�2.40)

DDur ? �0.779nnn�0.393n

�0.878nnn

(�3.50) (�1.89) (�3.15)

HiSupplierSGA ? �0.484nnn�4.553nnn

�1.089

(�3.08) (�4.13) (�1.59)

HiSGA ? �0.260 �0.080 �1.484nnn

(�0.98) (�0.54) (�3.18)

Size – �0.154nn�0.145nn

�0.212nnn�0.200nnn

�0.231nnn�0.222nnn

(�2.54) (�2.43) (�3.91) (�3.78) (�4.24) (�4.18)

Market-to-Book þ 0.071 0.087n 0.081n 0.105nn 0.081n 0.090nn

(1.59) (1.93) (1.82) (2.36) (1.81) (2.02)

Leverage þ 0.552 0.754 0.851 0.974n 0.775 1.045n

(0.99) (1.35) (1.51) (1.75) (1.38) (1.88)

CFOStd þ �3.438n�2.762 �3.053 �2.553 �2.951 �2.722

(�1.67) (�1.32) (�1.49) (�1.26) (�1.42) (�1.28)

Payout ? �0.948 �1.331 �0.978 �1.001 �0.710 �1.083

(�0.61) (�0.85) (�0.63) (�0.64) (�0.46) (�0.69)

DefaultRisk þ 0.435n 0.411n 0.576nn 0.520nn 0.585nnn 0.576nn

(1.89) (1.79) (2.55) (2.31) (2.59) (2.53)

Spread þ 0.735 1.007 1.338n 1.386n 1.296n 1.324n

(0.69) (1.03) (1.81) (1.85) (1.89) (1.79)

Litigation þ 1.685nn 2.048nnn 2.115nnn 2.178nnn 2.052nnn 2.195nnn

(2.49) (3.06) (3.17) (3.28) (3.06) (3.30)

CashFlow þ 2.436nnn 2.330nn 2.428nn 1.876n 2.797nn 2.699nn

(2.19) (2.04) (2.17) (1.66) (2.50) (2.36)

SelfHHI – �0.253nn�0.157 �0.246nn

�0.204n�0.160 �0.045

(�2.30) (�1.44) (�2.25) (�1.89) (�1.45) (�0.41)

ImportPenetration ? 0.112 0.538 0.007 0.586 �0.141 1.289

(0.18) (0.98) (0.01) (1.09) (�0.24) (2.20)

Union ? �0.008 �0.014n�0.008 �0.016nn

�0.013n�0.014n

(�1.12) (�1.86) (�1.04) (�2.09) (�1.70) (�1.85)

BigNAudit þ 0.313 0.253 0.234 0.207 0.265 0.188

(1.44) (1.18) (1.09) (0.98) (1.23) (0.90)

USGovernment þ 0.677nn 0.742nn 0.727nn 0.744nn 0.677nn 0.775nn

(2.09) (2.28) (2.23) (2.25) (2.08) (2.41)

Year-fixed effects Yes Yes Yes Yes Yes Yes

F-statistic 24.59nnn 7.48nnn 8.31nnn 10.38nnn 10.86nnn 8.90nnn

Adj. R2 (%) 12.94 12.61 12.83 13.06 13.27 13.42

N 7,064 7,064 7,064 7,064 7,064 7,064

DDurSupplier equals one if the fraction of a firm’s supplier industries that are durable goods producers is above the sample median for all manufacturing

industries in a given year, and equals zero otherwise. Durable goods producers operate in industries with SIC codes 245, 250–259, 283, 301, and 324–399.

DDur equals one if the firm operates in a durable goods industry, and equals zero otherwise. HiSupplierSGA equals one if the mean value for selling,

general and administrative expenses (data182) deflated by sales (data12) for a firm’s supplier industries is above the sample median in a given year, and

equals zero otherwise. HiSGA equals one if the value of selling, general, and administrative expenses deflated by sales for a firm is above the sample

median for a given year, and equals zero otherwise. All other variables are defined in Panel B of Table 3. All t-statistics (in parentheses) are based on

Huber/White/sandwich robust standard errors and adjusted for clustering of observations by firm. ***, **, and * indicate two-tailed significance at 1%, 5%,

and 10% significance levels, respectively. F-statistics are for the joint significance of test variables.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 17

Page 18: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]18

customer bargaining power measures are more positive if a firm operates in a durable goods industry or the firm hasgreater selling, general, and administrative expenses. The first, third, and fifth models in Table 7 are similar to Eq. (2)except that we include the independent variable, DDurSupplier, an indicator variable that equals one if the fraction of afirm’s supplier industries that are durable goods producers is greater than the sample median for all manufacturingindustries in a given year, and zero otherwise, and the independent variable DDur, an indicator variable that equals one if afirm is a durable goods producer, and zero otherwise. Also, we include interactions of DDurSupplier with the bargainingpower proxies for suppliers and interactions of DDur with the bargaining power proxies for customers. The second, fourth,and sixth models in Table 7 are similar to Eq. (2) except that we include the independent variable, HiSupplierSGA, anindicator variable that equals one if the mean value for selling, general, and administrative expenses scaled by sales for afirm’s supplier industries is greater than that of the sample median for all manufacturing industries in a given year, andzero otherwise, and the independent variable, HiSGA, an indicator variable that equals one if selling, general, andadministrative expenses scaled by sales for a firm is greater than the sample median value in a particular year, and zerootherwise. As well, we include interactions of HiSupplierSGA with the bargaining power proxies for suppliers andinteractions of HiSGA with the bargaining power proxies for customers.

Supporting Predictions 4a and 4b, we find significant positive coefficients on all the interaction variables in the Table 7models. These results are consistent with the proposition that when a firm’s suppliers produce specialized goods, they rely onthe firm to provide them with future revenue via service contracts and spare parts and consequently their demand for thefirm to have accounting conservatism is greater. Similarly, these results support the idea that when a firm producesspecialized goods its customers rely on it to provide them with maintenance services for the products they buy from the firmand parts for these products, and as a result, customers’ demand for conservatism is larger. Overall, the Table 7 results suggestthat when the demand from suppliers or customers with bargaining power that a firm follows more conservative accountingpractices is likely greater, this increases the firm’s propensity to follow such practices. Also, a comparison of the adjusted R2

values of the Table 7 models with the value for the baseline model in Table 4 shows that the Table 7 values are approximately5.2–11.9% larger than that for the Table 4 baseline model. Likewise, when comparing the adjusted R2 values for each of the sixmodels in Table 7 with those for the models in Table 4 that use the same supplier and customer bargaining power measures,we find that the adjusted R2 values of the Table 7 models are 1.2–6.8% greater than those for the Table 4 models.

4.3.5. Reconciliation of our findings with those in Bowen et al. (1995)

Bowen et al. (1995) predict that firms needing to fulfill greater implicit claims with suppliers and customers are morelikely to manage their earnings upwards by choosing aggressive accounting methods (i.e., less unconditional conservatism).One of their proxies for whether a firm has to fulfill significant implicit claims for its customers is whether the firmproduces durable goods. They find that if a firm is a durable goods producer, it is more likely to use income-increasingaccounting methods. It is useful to note that in the first, third, and fifth models in Table 7 we find that the coefficients onDDur are significantly negative, indicating less conditional conservatism when a firm produces durable goods.25 Given thatin Table 7 we also find that the positive associations between a firm’s conditional conservatism and its customers’bargaining power are more pronounced if the firm is a durable goods producer, we examine whether for durable goodsproducers if one-standard deviation increases in customer bargaining power would offset the effect of less conditionalconservatism from being a durable goods producer. We find that while such increases in customer bargaining powerwould lead to economically important increases to a durable goods producer’s conditional conservatism, these increaseswould not offset the main effect of lower conditional conservatism for durable goods producers.26 Overall, the Bowen et al.(1995) findings and our Tables 4–7 results could indicate that although those firms that need to fulfill many implicit claimswith suppliers and customers might have less unconditional conservatism, when a firm transacts with suppliers orcustomers that have relative bargaining advantages over it and likely dictate terms of trade, this positively impacts thefirm’s accounting conservatism.

4.3.6. Robustness checks

unc

the

con

et a

effe

PA

The conservatism measure Timeliness could be mis-specified when the good news coefficient (denominator) is negative.Thus, for robustness we rerun all of our analyses using an inverted timeliness measure. All of the inferences from thestudy’s results are qualitatively the same when we use this measure.

� As an additional firm-level measure of supplier or customer bargaining power, we consider whether a firm is in part

vertically integrated (i.e., the firm has upstream or downstream operations in at least one of its suppliers’ or customers’industries). If the firm is partially vertically integrated, this reduces suppliers’ or customers’ overall bargaining power as

25 The negative coefficients on DDur suggest that durable goods suppliers are less conditionally conservative, in addition to their being less

onditionally conservative (Bowen et al., 1995). These findings may suggest a positive relation between conditional and unconditional conservatism in

context of firms’ contracting with customers, which differs from the general negative association between conditional and unconditional

servatism. However, we acknowledge that differences with respect to the firms that are studied and sample period between our study and Bowen

l. (1995) makes our findings and those in Bowen et al. (1995) not directly comparable.26 For instance, from the first model in Table 7 we find that a one-standard deviation increase in customer bargaining power reduces the negative

ct on conditional conservatism of being a durable goods producer from �0.779 to �0.561 (¼�0.779þ0.134�1.63).

lease cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 19: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

PA

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 19

the need for contracting with some of its suppliers or customers is eliminated (Williamson, 1971, 1979; Porter, 1980).Thus, we expect that a firm’s upstream (downstream) vertical integration is inversely associated with suppliers’(customers’) overall bargaining power. For the Tables 4–7 analyses, we run additional models in which we use indicatorvariables for whether a firm vertically integrates into at least one of its supplier or customer industries as measuresfor supplier or customer bargaining power. The results of these models provide additional strong support for ourpredictions.

� We find that the study’s results are qualitatively the same if we calculate the relative size bargaining power measures

using sales instead of market value, or measure relative size as the size of a firm relative to the total size of its supplieror customer industries or the average size of its supplier or customer industries.

� The study’s results are robust to controlling for litigation risk using an alternative litigation risk measure (Francis et al.,

1994), which is based on the industry in which a firm operates.

� The study’s results are robust to dropping the use of year fixed effects and clustering errors by firm and year (Petersen,

2009).

� We do not include a control for bankruptcy risk in our regression models due to the high correlation this variable would

have with the firm’s total leverage. However, as a robustness test we include in our regression models a firm’s ex-antebankruptcy risk, estimated as in Campbell et al. (2008). The study’s findings are qualitatively unchanged when weinclude this control in our models. Also, our findings are robust to excluding the ex-ante default risk variable andincluding as control variables leverage and a firm’s ex-ante bankruptcy risk. Finally, our findings are robust to excludingleverage from our models and including as control variables the ex-ante default risk and bankruptcy risk measures.

� We test our hypothesis with an alternative conditional conservatism measure that relies on the asymmetric accrual-

cash flow relation (Ball and Shivakumar, 2005). Specifically, we focus on the three-way interaction term of cash flows,negative cash flows dummy, and the supplier/customer bargaining power measures in the pooled cross-sectionalregression of accruals. The results support our main hypothesis.

� As an alternative to our firm-level timeliness measure, we considered a pooled measure estimated at the firm’s six-digit

NAICS industry level. The empirical support for our hypothesis is robust to the use of this alternative measure.

� As in all studies that rely on industry classifications, we have measurement error in our data as many firms operate in

multiple segments. However, the empirical support for our hypothesis is robust to the use of the sub-sample of firmsthat operate in only a single segment.

� Prior research suggests that there is variation across time in antitrust enforcement (Lin et al., 2000) and that time-series

variation in litigation risk affects accounting conservatism (Basu, 1997; Holthausen and Watts, 2001). Changes insuppliers’ and customers’ bargaining power associated with time-series variation in antitrust enforcement could resultin predictable variation through time of the effect of a firm’s suppliers and customers on its accounting conservatism.We collect data on the number of antitrust cases per year over our sample period from sources such as theAdministrative Office of the Unites States Courts. Consistent with the prediction that during periods when antitrustenforcement and litigation exposure is higher that the effect of a firm’s suppliers and customers on its conditionalconservatism is reduced, we find that during years when the number of antitrust cases is greater than the sampleperiod median value that the associations between the measures for suppliers’ and customers’ bargaining power with afirm’s level of conditional conservatism are significantly less pronounced.

� After Statement of Financial Accounting Standards No. 131, firms were no longer required to report the names of

customers that were important for their business. However, firms still had to disclose the existence of such customersand how much of the firm’s sales were accounted for by various important customers. We verified that the proportionof firm-years during which we code a firm as having an important customer (supplier) does not differ between the1995–1997 and the 1998–2007 periods. Over these two periods the fraction of firm-years during which a firm has animportant supplier (customer) are 3.7% (8.4%) and 4.2% (9.8%). The differences in these fractions across sample periodsare not statistically significant.

5. Conclusion

Corporate suppliers and customers involved in short-term transactions with a firm will consider its accountingperformance to evaluate the firm’s ability to meet short-term trading obligations. Likewise, suppliers and customersengaged in long-term transactions with a firm will use its accounting performance to assess the prospects for the firm’s long-term financial viability and the risk of making relation-specific investments or that the firm might not honor their implicitclaims. A firm’s suppliers and customers are likely to be more concerned with bad news about a firm’s prospects than goodnews. However, a firm’s managers typically have incentives to exploit their asymmetrically informed position relative toother firm stakeholders by being less forthcoming about bad news concerning the firm than good news. We expect that thisleads to a demand from a firm’s suppliers and customers for it to recognize bad news in its earnings more quickly. Wehypothesize that a firm meets the demand for accounting conservatism from its suppliers or customers when thesestakeholders have bargaining advantages over it that allow them to dictate terms of trade or whether trade occurs at all.

Supporting our hypothesis, we document positive associations between measures for the bargaining power of suppliersand customers and the extent to which a firm has more timely recognition of bad news than good news. To strengthen ourconclusions, we run further tests that examine how these associations are modified by the reliance of a firm on its

lease cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 20: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]]20

suppliers, the reliance of a firm’s customers on it, the market power of a firm, and the importance for a firm’s suppliers orcustomers that it has greater accounting conservatism. The results of these tests suggest that the positive associations wedocument are indeed the result of firms meeting the demand for accounting conservatism from suppliers or customersthat have relative bargaining advantages and likely determine terms of trade.

Our study makes three main contributions. First, our evidence supports the contracting explanation for accountingconservatism that the existence of conditional conservatism is in part due to it allowing firms to structure more efficientcontracts with related parties. Second, our paper increases the understanding of how a firm’s suppliers and customersaffect its accounting practices by showing that when a firm contracts with suppliers or customers that have bargainingpower over it, the firm meets their demand for greater accounting conservatism. Finally, our paper contributes to theliterature that examines how product market competition impacts a firm’s accounting practices by showing that, inaddition to competition within its own industry affecting these practices, the extent to which a firm competes foreconomic profits with suppliers and customers also affects these practices.

References

Ahmed, A.S., Billings, B.K., Morton, R.M., Stanford-Harris, M., 2002. The role of accounting conservatism in mitigating bondholder–shareholder conflictsover dividend policy and in reducing debt costs. The Accounting Review 77, 867–890.

Ali, A., Klasa, S., Yeung, E., 2009. The limitations of industry concentration measures constructed with Compustat data: implications for finance research.Review of Financial Studies 22, 3839–3871.

Ali, A., Klasa, S., Yeung, E., 2010. Oligopolistic Industries and Corporate Disclosure Policy. Working Paper. University of Texas at Dallas.Baiman, S., Rajan, M., 2002. The role of information and opportunism in the choice of buyer–supplier relationships. Journal of Accounting Research

40, 247–278.Ball, R., 1989. The firm as a specialist contracting intermediary: application to accounting and auditing. Unpublished Manuscript.Ball, R., Kothari, S.P., Nikolaev, V., 2011. On Estimating Conditional Conservatism. Working Paper.Ball, R., Shivakumar, L., 2005. Earnings quality in UK private firms: comparative loss recognition timeliness. Journal of Accounting and Economics

39, 83–128.Bamber, L.S., Cheon, Y.S., 1998. Discretionary management earnings forecast disclosures: antecedents and outcomes associated with forecast venue and

forecast specificity choices. Journal of Accounting Research 36, 167–190.Banerjee, S., Dasgupta, S., Kim, Y., 2008. Buyer–supplier relationships and the stakeholder theory of capital structure. Journal of Finance 63, 2507–2552.Basu, S., 1995. Conservatism and the Asymmetric Timeliness of Earnings. Ph.D. Thesis. University of Rochester, Rochester, New York.Basu, S., 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics 24, 3–38.Basu, S., 2001. Discussion of ‘‘On the asymmetric recognition of good and bad news in France, Germany and the United Kindom’’. Journal of Business

Finance and Accounting 28, 1333–1349.Basu, S., 2005. Discussion of ‘‘Conditional and unconditional conservatism: concepts and modeling’’. Review of Accounting Studies 10, 311–321.Basu, S., Hwang, L., Jan, C., 2001. Differences in Conservatism Between Big Eight and Non-Big Eight Auditors. Working Paper. Baruch College, CUNY.Beaver, W.H., 1993. Conservatism. Working Paper. Stanford University.Beaver, W.H., Ryan, S.G., 2009. Risky Debt, Mixed-Attribute Accounting, and the Identification of Conditional Conservatism. Working Paper.Bernard, A., Jensen, J., Schott, P., 2006. Survival of the best fit: exposure to low-wage countries and the (uneven) growth of U.S. manufacturing plants.

Journal of International Economics 68, 219–237.Bowen, R., Ducharme, L., Shores., D., 1995. Stakeholders’ implicit claims and accounting method choice. Journal of Accounting and Economics 20, 255–295.Bull, C., 1987. The existence of self-enforcing contracts. Quarterly Journal of Economics 102, 147–159.Bulow, J., 1982. Durable-goods monopolists. Journal of Political Economy 90, 314–332.Burgstahler, D., Dichev, I., 1997. Earnings management to avoid earnings decreases and losses. Journal of Accounting and Economics 24, 99–126.Campbell, J., Hilscher, J., Szilagyi, J., 2008. In search of distress risk. Journal of Finance 63, 2899–2939.Campello, M., 2003. Capital structure and product market interactions: evidence from business cycles. Journal of Financial Economics 68, 353–378.Campello, M., 2006. Debt financing: does it hurt or boost firm performance in product markets? Journal of Financial Economics 82, 135–172.Coase, R.H., 1937. The nature of the firm. Economica 4, 386–405.Coase, R.H., 1972. Durability and monopoly. Journal of Law and Economics 15, 143–149.Coase, R.H., 2000. The acquisition of Fisher Body by General Motors. Journal of Law and Economics 43, 15–32.Coase, R.H., 2006. The conduct of economics: the example of Fisher Body and General Motors. Journal of Economics and Management Strategy

15, 255–278.Dhaliwal, D., Huang, S., Khurana, I., Pereira, R., 2009. Product Market Competition and Accounting Conservatism. Working Paper. University of Arizona.Farber, D., Hsieh, H., Jung, B., Yi, H., 2010. Labor Unions and Accounting Conservatism. Working Paper. University of Missouri.Fee, C.E., Thomas, S., 2004. Sources of gains in horizontal mergers: evidence from customer, supplier, and rival firms. Journal of Financial Economics

74, 423–460.Francis, J., Krishnan, J., 1999. Accounting accruals and auditor reporting conservatism. Contemporary Accounting Research 16, 135–165.Francis, J., LaFond, R., Olsson, P., Schipper, K., 2004. Costs of equity and earnings attributes. The Accounting Review 79, 967–1010.Francis, J., Philbrick, D., Schipper, K., 1994. Shareholder litigation and corporate disclosure. Journal of Accounting Research 32, 137–164.Galbraith, J.K., 1952. American Capitalism: The Concept of Countervailing Power. Houghton Mifflin Co., Boston.Giner, B., Rees, W., 2001. On the asymmetric recognition of good and bad news in France, Germany, and the United Kingdom. Journal of Business Finance

& Accounting 28, 1285–1331.Givoly, D., Hayn, C., 2000. The changing time-series properties of earnings, cash flows and accruals: has financial accounting become more conservative?

Journal of Accounting and Economics 29, 287–320.Harris, M.H., 1998. The association between competition and managers’ business segment reporting decisions. Journal of Accounting Research

36, 111–190.Haushalter, G.D., Klasa, S., Maxwell, W.F., 2007. The influence of product market dynamics on a firm’s cash holdings and hedging behavior. Journal of

Financial Economics 84, 797–825.Holthausen, R., Watts, R., 2001. The relevance of the value-relevance literature for finance accounting standard setting. Journal of Accounting and

Economics 31, 3–75.Inderst, R., Wey, C., 2007. Buyer power and supplier incentives. European Economic Review 51, 647–667.Kahn, C., 1986. The durable goods monopolist and consistency with increasing costs. Econometrica 54, 275–294.Karuna, C., 2007. Industry product market competition and managerial incentives. Journal of Accounting and Economics 43, 275–297.Klasa, S., Maxwell, W.F., Ortiz-Molina, H., 2009. The strategic use of corporate cash holdings in collective bargaining with labor unions. Journal of

Financial Economics 92, 421–442.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007

Page 21: Journal of Accounting and Economicsmedia.terry.uga.edu/socrates/publications/2011/12/... · 1988; Bowen et al., 1995; Banerjee et al., 2008).3 We predict that the importance of a

K.W. Hui et al. / Journal of Accounting and Economics ] (]]]]) ]]]–]]] 21

Klein, B., Crawford, R., Alchian, A., 1978. Vertical integration, appropriable rents, and the competitive contracting process. Journal of Law and Economics21, 297–326.

Klein, B., Leffler, K., 1981. The role of market forces in assuring contractual performance. Journal of Political Economy 89, 615–641.Klein, B., 2000. Fisher-General Motors and the nature of the firm. Journal of Law and Economics 43, 105–142.Kovenock, D., Phillips, G.M., 1997. Capital structure and product market behavior: an examination of plant exit and investment decisions. Review of

Financial Studies 10, 767–803.Krishnan, G.V., 2005. Did Houston clients of Arthur Andersen recognize publicly available bad news in a timely fashion? Contemporary Accounting

Research 22, 165–193.LaFond, R., Roychowdhury, S., 2008. Managerial ownership and accounting conservatism. Journal of Accounting Research 46, 101–135.LaFond, R., Watts, R.L., 2008. The information role of conservatism. The Accounting Review 83, 447–478.Leung, W., Li, O., Rui, O., 2009. Labor Unions and Accounting Conservatism. Working Paper. Chinese University of Hong Kong.Liberty, S., Zimmerman, J., 1986. Labor union contract negotiations and accounting choices. The Accounting Review 61, 692–713.Lin, P., Raj, B., Sandfort, M., Slottje, D., 2000. The US antitrust system and recent trends in antitrust enforcement. Journal of Economic Surveys 14, 255–306.Lustgarten, S., 1975. The impact of buyer concentration in manufacturing industries. Review of Economics and Statistics, 125–132.Mackay, P., Phillips, G.M., 2005. How does industry affect firm financial structure? Review of Financial Studies 18, 1433–1466.Mian, S.L., Smith, C.W., 1992. Accounts receivable management policy: theory and evidence. Journal of Finance 48, 169–200.Perry, M., 1989. Vertical integration: determinants and effects. In: Schmalensee, R., Willig, R. (Eds.), Handbook of Industrial Organization, North-Holland,

New York.Petersen, M.A., 2009. Estimating standard errors in finance panel data sets: comparing approaches. Review of Financial Studies 22, 435–480.Pope, P.F., Walker, M., 1999. International differences in the timeliness, conservatism, and classification of earnings. Journal of Accounting Research

37, 53–87.Porter, M., 1980. Competitive Strategy. Free Press, New York.Qiang, X., 2007. The effects of contracting, litigation, regulation, and tax costs on conditional and unconditional conservatism: cross-sectional evidence at

the firm level. The Accounting Review 82, 759–796.Rogers, J.L., Stocken, P.C., 2005. Credibility of management forecasts. The Accounting Review 80, 1233–1260.Roychowdhury, S., Watts, R.L., 2007. Asymmetric timeliness of earnings, market-to-book and conservatism in financial reporting. Journal of Accounting

and Economics 44, 2–31.Ruddock, C., Taylor, S.J., Taylor, S.L., 2006. Nonaudit services and earnings conservatism: is auditor independence impaired? Contemporary Accounting

Research 23, 701–746.Schumacher, U., 1991. Buyer structure and seller performance in U.S. manufacturing industries. Review of Economics and Statistics 73, 277–284.Snyder, C.M., 1996. A dynamic theory of countervailing power. Rand Journal of Economics 27, 747–769.Soosay, C.A., Hyland, P.W., Ferrer, M., 2008. Supply chain collaboration: capabilities for continuous innovation. Supply Chain Management 13, 160–169.Spekman, R.E., 1988. Strategic supplier selection: understanding long-term buyer relationships. Business Horizons 31, 75–81.Spekman, R.E., Spear, J., Kamauff, J., 2002. Supply chain competency: learning as a key component. Supply Chain Management 7, 41–55.Stokey, N., 1981. Rational expectations and durable goods pricing. The Bell Journal of Economics 12, 112–128.Telser, L.G., 1980. A theory of self-enforcing agreements. Journal of Business 53, 27–44.Titman, S., Wessels, R., 1988. The determinants of capital structure choice. Journal of Finance 43, 1–19.Trefler, D., 1996. Trade liberalization and the theory of endogenous protection: an econometric study of U.S. import policy. Journal of Political Economy

101, 138–160.Vassalou, M., Xing, Y., 2004. Default risk in equity returns. Journal of Finance 59, 831–868.Verrecchia, R.E., Weber, J., 2006. Redacted disclosure. Journal of Accounting Research 44, 791–814.Watts, R.L., 1993. A Proposal for Research on Conservatism. Working Paper. University of Rochester.Watts, R.L., 2003a. Conservatism in accounting, Part I: explanations and implications. Accounting Horizons 17, 207–221.Watts, R.L., 2003b. Conservatism in accounting, Part II: evidence and research opportunities. Accounting Horizons 17, 287–301.Watts, R.L., Zimmerman, J.L., 1986. Positive Accounting Theory. Prentice-Hall, Englewood Cliffs, NJ.Williamson, O.E., 1971. The vertical integration of production: market failure considerations. American Economic Review 71, 112–123.Williamson, O.E., 1979. Transaction-cost economics: the governance of contractual relations. Journal of Law and Economics 22, 233–261.Zhang, J., 2008. The contracting benefits of accounting conservatism to lenders and borrowers. Journal of Accounting Economics 45, 27–54.

Please cite this article as: Hui, K.W., et al., Corporate suppliers and customers and accounting conservatism. Journal ofAccounting and Economics (2011), doi:10.1016/j.jacceco.2011.11.007