supply chain lending and accounting conservatism · supply-chain information and conservative...
TRANSCRIPT
Supply Chain Lending and Accounting Conservatism
Guojin Gong
Smeal College of Business
Pennsylvania State University
University Park, PA 16802
Shuqing Luo
Business School
National University of Singapore
1 Business Link, Singapore 117592
March 2014
Comments Welcome
We thank workshop participants at National University of Singapore and the 2013 American
Accounting Association Annual Meeting for helpful comments and suggestions.
Supply Chain Lending and Accounting Conservatism
ABSTRACT: Economic interdependence between suppliers and customers generates
proprietary information useful for credit risk assessment. Such proprietary information can
mitigate information asymmetry between borrowers and lenders, and substitutes for conservative
accounting numbers in debt contracting. We hypothesize that a pre-existing lending relationship
with a borrower’s customer(s) grants the lender access to proprietary supply-chain information
and reduces the lender’s demand for, and therefore the supply of, accounting conservatism by the
borrowing supplier. Consistently, we document a negative effect of the pre-existing lender-
customer lending relationship on the borrowing suppliers’ accounting conservatism at loan
origination. This negative effect is robust to instrumental variable estimation, propensity score
matching approach, and a placebo test. Further, this negative effect becomes stronger when the
suppliers and customers operate in an opaque information environment, when the customers
represent important business partners and have long-lasting relationship with their lenders, and
when the suppliers extend sizeable trade credit to their customers. In additional analysis, we find
less use of accounting-based covenants when lenders have a pre-existing lending relationship
with the borrower’s customer(s). Our findings demonstrate the informational value of proprietary
supply-chain information for lenders and its impact on borrowers’ reporting practices.
Keywords: Supply-chain relationship; Information asymmetry; Debt contracting; Conservatism.
Data Availability: Data are available from public sources indicated in the text.
1
I. INTRODUCTION
Business relationships between suppliers and customers create significant operational and
financial interdependence within supply chains. Such economic interdependence generates
valuable information about supply-chain partners’ financial health and business risks.1 Prior
literature documents considerable influences of supply-chain information over corporate
managers and equity market participants.2 Yet, the implication of supply-chain information for
lenders, one important stakeholder group, has received little attention. As supply-chain partners
typically engage in non-arm’s-length transactions and collaborative activities, supply-chain
information is often proprietary in nature and unavailable to the public.3 Lenders, as the
dominant source of corporate financing, can potentially access such proprietary information
when contracting with supply-chain partners. In this study, we examine the informational value
of supply-chain information in lenders’ decision making and its associated effect on borrowers’
financial reporting practices. Specifically, we examine whether a pre-existing lending
relationship with a borrower’s major customer(s), a situation where the lender is granted access
to proprietary supply-chain information, reduces the lender’s demand for, and therefore the
supply of, conservative accounting by the borrowing supplier at loan origination.
During a lending process, borrowers are better informed than lenders about their own credit
quality, and lenders are more sensitive to borrowers’ losses than they are to borrowers’ profits.
The information asymmetry between borrowers and lenders, coupled with lenders’ asymmetric
payoffs, leads to agency cost of debt (Jensen and Meckling 1976). An extensive literature
1 See Cooper, Ellram, Gardner and Hanks (1997), Blackwell and Blackwell (1999), Lambert and Cooper (2000),
Lee, So, and Tang (2000), Baiman and Rajan (2002a, 2002b), and Kulp, Lee, and Ofek (2004). 2 See, for example, Kulp, Lee, and Ofek (2004), Hertzel, Li, Officer, and Rodgers (2008), Cohen and Frazzni
(2008), and Murfin and Njoroge (2011). 3
Examples of proprietary supply-chain information include customers’ purchase plans, demand forecasts,
collaborative product design and development, and other relationship-specific investments.
2
proposes that lenders demand conservative accounting numbers to reduce deadweight agency
costs, because conservatism can generate verifiable lower bound measures of net assets that
assist lenders in screening potential borrowers and in monitoring borrower’s creditworthiness
after loan issuance (e.g., Leftwich 1983; Watts and Zimmerman 1986; Watts 1993, 2003a, 2003b;
Ahmed, Billings, Morton, and Stanford-Harris 2002; Zhang 2008).4 Consistently, prior evidence
suggests that higher lender-borrower information asymmetry leads to more conservative
accounting by the borrowers (e.g., Ball, Kothari, and Robin 2000; Ball and Shivakumar 2005;
Gormley, Kim, and Martin 2011; Erkens, Subramanyam, and Zhang 2013).
Through a pre-existing lending relationship with major customer(s), a lender can
communicate with the customer(s) privately and request timely financial disclosures, covenant
compliance information, financial projections, and even strategic plans that may or may not well
aligned with the interests of the suppliers (e.g., Standard and Poor's 2007; Sufi 2007). Proprietary
information obtained from the borrowing customer(s) has important implications for the
supplier’s business operations and financial health. The lending experience with major
customer(s), therefore, enhances the lender’s understanding of the supplier’s business risks and
creditworthiness. This tends to reduce the reliance on conservative accounting numbers when the
lender makes the loan-granting decision and deliberates potential monitoring costs (hereafter, we
term the lender who has prior lending experience with the customer when granting loans to the
supplier as a “supply-chain lender”). As conservatism in conjunction with debt contracting is
costly to implement (e.g., Beneish and Press 1993; Gigler, Kanodia, Sapra and Venugopalan
2009), suppliers that borrow from supply-chain lenders adopt less conservative accounting
4 Throughout the paper, we use conservatism to refer to the timely incorporation of economic losses into accounting
earnings (Basu 1997; Watts 2003a), which is also called asymmetric timeliness of loss recognition or conditional
conservatism. From a debt contracting perspective, conditional and unconditional conservatism are substantially
different concepts. While unconditional conservatism seems at best neutral and possibly inefficient, conditional
conservatism can increase the efficiency of debt contracting (Ball and Shivakumar 2005).
3
choices to save associated costs.5 In contrast, a non-supply-chain lender who does not have prior
lending experience with major customer(s) has less information to assess the supplier’s credit
quality prior to loan initiation. Therefore, to mitigate agency costs, the lender is likely to demand
more conservative financial reporting when granting loans to suppliers. Accordingly, we
hypothesize that suppliers borrowing from supply-chain lenders report less conservatively than
suppliers borrowing from non-supply-chain lenders.
We note that our hypothesis is not entirely straightforward. Proprietary supply-chain
information often involves forward-looking events (e.g., customers’ planned purchase orders and
product development plans), and the credibility of such “soft” information remains a concern.
Conservative accounting numbers presumably are more credible than “soft” information due to
extensive public scrutiny of financial statements. It is possible that lenders consider proprietary
supply-chain information and conservative accounting numbers as complementary sources of
information, and do not reduce the demand for conservatism despite the possession of customers’
confidential information. Furthermore, the economic interdependence between a supplier and its
major customer(s) leads to positively correlated business risks along the supply chain (e.g.,
Hertzel et al. 2008). By lending to members of the same supply chain, a lender faces more
concentrated credit risk and has incentive to more diligently monitor borrowers, heightening the
demand for accounting conservatism. We also note that demands for conservatism come from
multiple sources including other lenders,6 corporate boards, auditors’ and managers’ legal
5 The costs of conservative accounting include the adverse effect that covenant violations have on firms (which are
exacerbated by “false alarms”), the potential negative effect on managers’ investment decisions, and possible
inefficiencies for equity valuation (e.g., Gigler, Kanodia, Sapra, and Venugopalan 2009; Guay and Verrecchia 2006;
Roychowdhury 2010; Barth, Caprio, and Levine 2004). 6 It is possible that other lenders’ conservatism demand is muted due to the certification role of supply-chain lenders,
because the fact that a supply-chain lender is willing to initiate or continue the loan indicates that the borrowing
supplier does not face immediate and serious concern about credit quality. In untabulated tests, we compare loan
terms between non-supply-chain lenders and concurrent non-supply-chain lenders for suppliers also borrowing from
supply-chain lenders. We find that the use of accounting-based covenants is significantly less prevalent when the
4
liability concerns, and shareholders (e.g., Watts 2003a, 2003b). The effect of supply-chain
lending experience might be outweighed by conservatism demands from other sources. Whether
the supply-chain lending experience affects the lender’s demand for accounting conservatism
remains an empirical question.
We start by examining the association between the existence of lender-customer lending
relationship and suppliers’ accounting conservatism at loan origination. We measure accounting
conservatism using firm-specific C-Score developed in Khan and Watts (2009). Results show
that at loan origination, suppliers borrowing from supply-chain lenders report less conservatively
than those borrowing from non-supply-chain lenders, after controlling for credit quality and
financial health of both the suppliers and their customers. This finding is consistent with the
notion that proprietary supply-chain information, by mitigating lender-borrower information
asymmetry, substitutes for conservative accounting numbers in debt contracting.
Extending credit to members of the same supply chain is an endogenous decision, and
omitted variables that lead to this decision may correlate with lenders’ demand for, as well as
borrowers’ supply of, accounting conservatism. To alleviate the endogeneity concern, we employ
an instrumental variable approach, a propensity score matching approach, and a placebo test. We
continue to find a negative effect of lender-customer lending relationship on the borrowing
supplier’s conservatism at loan origination, suggesting that endogeneity is unlikely to be the
primary driver of our finding.
Admittedly, in addressing our research question we face the empirical challenge of being
unable to directly observe the flows of proprietary information from customers to their lenders.
To alleviate this concern and further explore the impact of proprietary supply-chain information,
borrowing supplier also has a supply-chain lender. This result suggests that, conditional on the presence of supply-
chain lenders, other (non-supply-chain) lenders recognize the reduced conservatism by the borrowing suppliers and
accordingly adjust the covenant structure to reduce the reliance on accounting numbers.
5
we conduct several cross-sectional analyses. First, the access to propriety supply-chain
information should be more valuable when the supplier’s credit quality is more difficult to assess.
Consistently, we find a more pronounced negative relation between the pre-existing lender-
customer relationship and borrowing suppliers’ conservatism when the supplier and its major
customer(s) operate in an opaque information environment, where informational opacity is
proxied by low quality analyst forecasts and large magnitude of discretionary accruals. Second,
we expect that strong economic tie between supply-chain members and long-lasting lender-
customer lending relationship would generate more valuable proprietary information relevant for
evaluating the supplier’s credit quality. As expected, results show that greater customer
importance, proxied by the proportion of major customers’ purchases in their suppliers’ total
sales, and longer lender-customer lending relationship, proxied by the number of years that a
lender has had lending experience with the customers, strengthens the negative relation between
lender-customer lending relationship and borrowing suppliers’ conservatism. Third, among
various types of information obtainable from major customers, credit-relevant information is the
most important to assess the creditworthiness of their suppliers. Supporting this conjecture, we
find that larger trade credit extended to major customer(s), an important source of credit-relevant
information, strengthens the negative relation between lender-customer lending relationship and
borrowing suppliers’ conservatism.
In additional analyses, we find that supply-chain lenders accept fewer accounting-based
covenants than non-supply-chain lenders. This finding echoes existing evidence of fewer debt
covenants due to acquisition of private information by lenders (Ciamarra 2012; Bharath, Dahiya,
6
Saunders, and Srinivasan 2011; Engelberg, Gao, and Parsons 2012),7 thus providing further
support for our hypothesized effect of proprietary supply-chain information on lender-borrower
information asymmetry. This result also speaks to the use of conservatism in debt contract design.
Beatty, Weber, and Yu (2008) point out that lenders “who in spite of having the authority to
prescribe the financial information that they want, typically include at most very crude
adjustment to their contracts.” Our evidence indicates that better informed lenders do make
modifications in the design of loan covenants in accordance with a lower demand for
conservatism.
We contribute to the literature examining economic consequences of supply-chain network.
Prior studies document that economic interdependence and information integration between
supply-chain partners affect operating profit (Kulp et al. 2004), financial distress (Hertzel et al.
2008), corporate investment (Murfin and Njoroge 2011), opportunistic financial reporting
(Ramam and Shahrur 2008), and cross-firm return predictability (Cohen and Frazzni 2008;
Pandit, Wasley, and Tzachi 2011). A few recent studies also examine how various features of the
supplier-customer relationship affect loan contract terms (Koh, Teoh, and Tham 2011; Li and
Yang 2011; Files and Gurun 2011). The literature overlooks the potential interplay between
lenders and supply-chain partners in the debt contracting process. Lenders, being the major
capital provider, can privately communicate with borrowing customers and influence borrowing
suppliers’ financial reporting practices. Our study employs a novel setting, supply-chain lending
relationship, to offer new evidence concerning the role of proprietary supply-chain information
(obtained over the course of a prior lending relationship) in affecting lenders’ deliberations and
borrowers’ financial reporting practices.
7 These studies find fewer restrictive debt covenants when lender-borrower information asymmetry is mitigated due
to lenders’ acquisition of private information through board representation (Ciamarra 2012), relationship lending
(Bharath et al. 2011), or personal connections between employees of banks and firms (Engelberg et al. 2012).
7
Our study is also closely related to the literature examining the impact of information
asymmetry on accounting conservatism. Several studies suggest that lower lender-borrower
information asymmetry, achieved through private communication, geographical connection,
board representation, or relationship lending, leads to lower accounting conservatism by the
borrowers (Ball et al. 2000; Ball and Shivarkumar 2005; Gormley et al. 2011; Erkens et al. 2013).
Relatedly, LaFond and Watts (2008) document a lower level of conservatism after a decrease in
information asymmetry between equity investors.8 Our evidence complements these studies by
identifying supply-chain lending as one important information channel through which lenders
obtain proprietary information from borrowers’ stakeholders (in particular, major customers).
Such information appears to substitute for borrowers’ financial reporting (in particular,
conservative accounting numbers) during the debt contracting process. We also document an
indirect effect of proprietary supply-chain information on debt contract design—such
information, by substituting for conservative accounting numbers in debt contracting, reduces the
reliance on accounting numbers in the design of debt covenants.
The rest of the paper proceeds as follows. We develop testable hypotheses in Section II.
Section III describes the sample selection process and presents the sample. Section IV reports
and discusses empirical findings. We conclude in Section V.
II. HYPOTHESIS DEVELOPMENT
Proprietary supply-chain information and credit risk assessment
Proprietary supply-chain information has significant implications for suppliers’ business
operations and financial health. A lender can use such information directly as a source of new
information, in addition to the information received from other sources (such as public SEC
8 Conversely, accounting conservatism can affect information asymmetry. Wittenberg-Moerman (2008) shows that
greater conservatism lessens the information asymmetry between credit investors, and improves the efficiency of the
secondary trading of debt securities.
8
filings, press releases, and even inside information provided by the supplier at loan origination),
to improve credit assessment. The lender can also use the proprietary supply-chain information
indirectly to verify information regarding the supplier’s business prospects and facilitate
information processing in credit analysis. Both direct and indirect use of proprietary supply-
chain information can enhance lenders’ understanding of the suppliers’ creditworthiness.
In a supply-chain network, suppliers often request a significant amount of proprietary
information from customers to facilitate collaborative activities and enhance investment
efficiency. In particular, suppliers require their customers’ production plans and demand
forecasts to effectively allocate resources and manage operations. From lenders’ perspective, it is
difficult to ascertain and verify customers’ demand information solely from publicly available
information, as customers may not fully reveal their product development and expected demand
to the public and even their suppliers due to concerns about the competitive harm of disclosures
(Baiman and Rajan 2002b).9 Proprietary information about customer demand, obtained over the
course of the lender-customer lending relationship, can enhance supply-chain lenders’
understanding of the suppliers’ business prospects and operational stability.10
Moreover, through prior lending experience with major customers, supply-chain lenders can
glean insights concerning the incentive alignment between customers and their suppliers.
Misaligned incentives along the supply chain can impede timely innovation and the development
of relationship-specific investments, reducing the suppliers’ potential cash flows (Baiman and
9 Although information sharing is common between customers and suppliers (Lee et al. 2000; Baiman and Rajan
2002a; Kulp et al. 2004), it does not preclude “partial disclosure” or “selective disclosure” along the supply chain
(Baiman and Rajan 2002b). For instance, the potential for hold-up in supply chains can impede information sharing
between business partners (Schloetzer 2012). 10
As one example, a customer’s new product development plan tends to bring in higher and more stable future
revenues for its supplier. The customer refrains from publicly disclosing this information in fear of intense
competition from peers, but may supply this information to its lender to obtain favorable loan terms. The lender thus
enjoys an information advantage over outsiders (including non-supply-chain lenders) when evaluating the supplier’s
debt capacity and debt repayment ability.
9
Rajan 2002a). Therefore, supply-chain lenders have greater ability to project the level and
persistence of cash flow streams by the suppliers, which facilitates the credit granting process.
Besides implications for business operations, proprietary supply-chain information also
reveals important aspects regarding suppliers’ financial health and liquidity risk. Suppliers often
extend a substantial amount of trade credit to customers (Petersen and Rajan 1997). Accounts
receivable (resulting from trade credit) serves as an important collateral in loan contracts (Mian
and Smith 1992), and failure to timely collect accounts receivable can lead to adverse
consequences such as default on loans and cutting investment (Murfin and Njoroge 2011).
Proprietary information regarding customers’ financial health can shed light on the liquidity risk
exposure of suppliers to their customers. The information can also help lenders evaluate the
reasonableness of sales forecasts and/or cash flow forecasts provided by the supplier’s
management team. This can help supply-chain lenders better evaluate the risk of extending credit
to the suppliers.
Hypotheses
As discussed above, through prior lending experience with customers, supply-chain lenders
gain proprietary information valuable for credit risk assessment of suppliers. Proprietary supply-
chain information can reduce lenders’ demand for conservatism by the borrowing supplier
through two channels. First, supply-chain lending experience improves lenders’ ability to screen
potential borrowers. Prior to loan origination, proprietary supply-chain information can facilitate
supply-chain lenders to identify high quality suppliers and price their loans accordingly; this
reduces the need for conservative accounting numbers to protect against potential downside risk.
In contrast, non-supply-chain lenders face greater information disadvantage in assessing
10
borrowing suppliers, strengthening their demand for reporting conservatism.11
Second, after loan
origination, supply-chain lending experience facilitates lenders’ monitoring of borrowers. Prior
literature suggests that conservative reporting allows lenders to more closely monitor borrowers’
creditworthiness on a timely basis, and thus limits managerial opportunism (e.g., Watts and
Zimmerman 1986; Watts 1993).12
Proprietary supply-chain information offers lenders valuable
and timely signals about the borrowing suppliers’ cash flow steams and credit risk, and thus can
reduce the reliance on conservative accounting numbers in the lender monitoring process.
Anticipating this monitoring benefit, supply-chain lenders can permit less conservative reporting
before granting loans to the borrowing suppliers.
While non-supply-chain lenders can also learn about the supply-chain network from
publicly available information and/or requesting proprietary information directly from the
borrowing suppliers at loan origination, without any private communication with the customers,
non-supply-chain lenders need to incur greater costs verifying and processing acquired
information. Hence, it is more cost effective for non-supply-chain lenders to rely on “hard”
information such as conservative accounting numbers when contracting with the supplier. Non-
supply-chain lenders, therefore, demand more conservative financial reporting to mitigate
potential downside risks. To the extent that practicing conservative reporting involves nontrivial
11
Our hypothesis is complicated by the fact that proprietary supply-chain information may reveal poor credit quality
about the borrowing suppliers and lead to more informed rejection of loan applications. To the extent that borrowing
suppliers selected by supply-chain lenders differ systematically from the suppliers borrowing from non-supply-chain
lenders, our evidence may be confounded by correlated omitted variables that affect borrowers’ credit quality and
lenders’ loan-approval decisions. To mitigate this endogeneity concern, we control credit quality for both the
supplier and its customer(s) in the empirical models. We also use a propensity score matching approach and an
instrumental variable approach to address the concern of correlated omitted variables. 12
After loan origination, lenders do not have direct control over borrowers’ accounting policy. Nevertheless, lenders
can influence borrowers to maintain or adopt more conservative accounting choices through lender-initiated loan
modification/renegotiation and covenant violation, and thus limit the scope for managerial opportunism.
11
costs,13
borrowing suppliers respond to the lower conservatism demand by supply-chain lenders
to save associated costs. This reasoning leads to the following hypothesis, stated in the
alternative form.
H1: A pre-existing lending relationship with the borrower’s customer is associated with less
accounting conservatism by the borrower.
We further postulate several factors that may affect the informational value of propriety
supply-chain information in debt contracting. First, when customers operate in an opaque
information environment, gaining access to customers’ inside information allows supply-chain
lenders to enjoy greater information advantage than non-supply-chain lenders. In a similar vein,
the suppliers’ information opacity can also increase the potential benefits of supply-chain
lending because information opacity makes it more costly for non-supply-chain lenders to
acquire and analyze credit-relevant information when assessing the suppliers’ creditworthiness.
Hence, we expect that information opacity for both customers and borrowing suppliers increases
the information advantage of supply-chain lenders and further lowers their demand for
accounting conservatism.
H2a: Information opacity of the customer strengthens the negative effect of lender-customer
lending relationship on the borrowing supplier’s accounting conservatism.
H2b: Information opacity of the supplier strengthens the negative effect of lender-customer
lending relationship on the borrowing supplier’s accounting conservatism.
Second, the benefits of gaining access to propriety supply-chain information should be
greater when a supplier’s business relies more heavily on its customers, as greater business
reliance produces more and higher quality proprietary information for credit risk assessment.
Relatedly, supply-chain lenders are able to acquire more high-quality proprietary information
13
While it is debatable whether conservatism improves debt-contracting efficiency (Gigler et al. 2009; Caskey and
Hughes 2012), our paper is agnostic regarding this theoretical debate. In particular, our arguments only require
nontrivial costs associated with conservative accounting, and do not imply lower debt-contracting efficiency for
supply-chain lenders.
12
when they have repeated lending experience with the customers and hence are more familiar
with the customers’ business risks. Endowed with more high-quality proprietary information,
supply-chain lenders are likely to further reduce their reliance on conservative accounting
numbers in the lending process.
H3a: The importance of the customer in the supplier’s business strengthens the negative
effect of lender-customer lending relationship on the borrowing supplier’s accounting
conservatism.
H3b: The duration of prior lending experience with the customer strengthens the negative
effect of lender-customer lending relationship on the borrowing supplier’s accounting
conservatism.
Finally, of all types of information accessible to supply-chain lenders, credit-relevant
information is most important to their decision making. Trade credit extended to customers
represents an important source of credit-relevant information because it directly reveals the
supplier’s credit exposure to customers. Supply-chain lenders can evaluate the collectability of a
supplier’s accounts receivable by assessing its customers’ ability to repay trade credit in a timely
manner. As accounts receivable often serves as important collateral in loan contracting (Mian
and Smith 1992), a large amount of trade credit extended to the customers (i.e., greater credit
exposure to the customers) indicates greater informational value of proprietary supply-chain
information. Consequently, the amount of trade credit extended to major customers tends to
strengthen the negative effect of supply-chain lending on the borrowing supplier’s accounting
conservatism.
H4: The amount of trade credit extended to the customer strengthens the negative effect of
lender-customer lending relationship on the borrowing supplier’s accounting
conservatism.
13
III. SAMPLE, VARIABLE MEASUREMENT, AND DESCRIPTIVE STATISTICS
Sample
To build the sample, we start by collecting information on public firms and their disclosed
major customers. Our data collection is facilitated by Regulation S-K, which requires public
issuers (with the exception of small businesses filing 10-KSB) to reveal the identity of their
major customers as well as SFAS 14, which requires public companies to disclose the amount of
revenues from each major customer.14
These disclosures are collected and compiled by
Compustat Segment files, from which we identify a sample of 27,041 supplier-customer-year
observations with valid firm identifier (GVKEY) for both the suppliers and their customers
between 1988 and 2010.15
We then exclude 6,464 observations if the suppliers and/or the
customers belong to utility industry (SIC code 4000-5000) or financial industry (SIC code 6000-
7000). These procedures yield a sample of 20,577 supplier-customer-year observations,
representing 15,339 unique supplier-years.
For each supplier-customer-year observation, we search whether the supplier initiated new
loan(s) during the year. We do so in the Loan Pricing Corporation’s DealScan database which
has a comprehensive coverage of the global loan markets, providing information on the identity
of lenders and borrowers, loan characteristics, and loan contract terms. For each supplier-lender
pair, we then search whether the same lead arranger has a lending relationship with the supplier’s
major customers prior to (and remain outstanding during) the year that the lender initiated loans
14
Accordingly to SFAS 14 (paragraph 39), a major customer is defined as any customer whose purchases represent
at least ten percent of an issuer’s consolidated revenue and if the loss of the customer would have a material adverse
effect on the issuer and its subsidiaries. While SFAS 14 has been superseded by SFAS 30 in 1979 and SFAS 131 in
1997, the disclosure requirement about major customers remains. 15
We start our sample with year 1988 as this is the year Dealscan starts to have a more continuous coverage of
global loan markets.
14
to the supplier (i.e., supply-chain lender).16
We find that 16.7% of the suppliers (2,560 of 15,339)
initiated a lending relationship during a supplier-customer-year when their major customers have
outstanding loans in the year. Among these suppliers 41.7% (1,068 of 2,560) share at least one
lead arranger with their major customers during a supplier-customer-year.
We further retrieve financial data from Compustat annual files, stock trading information
from CRSP daily and monthly files to construct accounting conservatism measures and control
variables included in the regression analyses. The final sample includes 1,135 supplier-customer-
year observations, 61.7% of which (700 of 1,135) share a common lead arranger with their major
customers in the same year.
Variable measurement
Supply-chain lending relationship
We create an indicator variable CHAIN_LENDING to denote the existence of a lender-
customer lending relationship. CHAIN_LENDING equals one if a supplier borrows from a lead
arranger who also serves as a lead arranger of an outstanding loan to the supplier’s major
customer(s) (i.e., supply-chain lender), and equals zero if a firm borrows from a lead arranger
who does not have outstanding lending relationship with the supplier’s major customer(s) (i.e.,
non-supply-chain lender).
Accounting conservatism
Prior research suggests that lenders demand accounting conservatism, in particular
asymmetric timeliness of loss recognition, in the loan contracting process (e.g., Watts 2003a,
16
Note that 74.7% (11,455 of 15,339) of the suppliers have reported only one single major customer. When supplier
firms have reported more than one major customers in a fiscal year, we define supply chain lenders based on
whether a lender acts as a lead arranger in an outstanding lending relationship with any one of the supplier’s major
customers prior to initiating loans to the suppliers. In subsequent analysis where we examine how customer
characteristics affect lenders’ demand for suppliers’ accounting conservatism, we retain the largest customer (who
makes the greatest amount of purchases from the supplier) for each supplier-year observation.
15
2003b). To measure asymmetric timeliness of loss recognition, we follow Khan and Watts (2009)
to construct a firm-year measure C-Score (CSCORE). CSCORE stems from Basu’s (1997)
asymmetric timeliness metric, but allows inter-temporal and cross-sectional variations in
accounting conservatism. To measure C-Score, Khan and Watts (2009) begin with Basu (1997)
piece-wise linear regression, described below as Equation (1.1), and propose an augmented
regression, described below as Equation (1.2).
Xi,t = β1 + β2Di,t + β3Ri,t + β4Di,t×Ri,t + εi,t (1.1)
Xi,t = β1 + β2Di,t + Ri,t (μ1+ μ2ln(MVE)i,t + μ3MBi,t + μ4LEVi,t)
+ Di,t×Ri,t (λ1+ λ2ln(MVE)i,t+ λ3MBi,t + λ4LEVi,t) + (δ1ln(MVE)i,t
+ δ2MBi,t + δ3LEVi,t + δ4Di,t×ln(MVE)i,t+ δ5Di,t×MBi,t+ δ6Di,t×LEVi,t) + εi,t (1.2)
In both equations, i indexes firm and t indexes year. X is income before extraordinary items
divided by lagged market value of equity. R is 12-month compounded return starting nine
months before the year end. D is an indicator variable that equals one when R is negative, and
zero otherwise. In Equation (1.2), ln(MVE) is the logarithm transformation of market
capitalization, MB is market-to-book ratio, and LEV is book leverage.
In Equation (1.1), the coefficient β3 reflects the sensitivity of earnings to good news, and the
coefficient β4 captures the incremental sensitivity of earnings to negative news. Following Khan
and Watts (2009), for each year we specify β3 and β4 as a linear function of firm characteristics
that have been shown to significantly affect the degree of conservatism (e.g., LaFond and
Roychowdhury 2008; LaFond and Watts 2008), including firm size, market-to-book ratio, and
book leverage. C-Score (CSCORE) is specified as CSCORE = β4 = λ1+ λ2ln(MVE)i,t+ λ3MBi,t +
λ4LEVi,t.
We estimate Equation (1.2) for each year during our sample period using data from the
intersection of CRSP and COMPUSTAT. The coefficient estimates are applied to the above
16
specification for C-Score to generate the firm-year measure of accounting conservatism. A
higher value of CSCORE represents more conservative financial reporting.
Control variables
We control for firm size (ln(MVE)) and firm age (ln(FIRMAGE)), because larger firms and
more matured firms are likely to have richer information environment, which diminish the
importance of conservatism in debt contracting (Khan and Watts 2009). We also include
investment cycle (INVCYCLE), market-to-book ratio (MB),17
return volatility (RETVOL), sales
volatility (SALEVOL), and bid-ask spreads (BID_ASK), as firms having longer investment cycle,
more growth opportunities, greater return and sales volatility, and larger bid-ask spreads tend to
suffer from greater information asymmetry, and hence are subject to higher demand for
conservative reporting. Since conservatism increases in credit risk, we include leverage (LEV)
and Altman’s Z-score (ALTMAN) in the regression analysis. We add auditor choice (BIG_FOUR)
since prior studies suggest that big-four auditors demand a higher level of accounting
conservatism due to greater bargaining power and reputation concern (Basu et al. 2001; Krishan
2005). We also identify industries that involve high litigation risk (LITIGATION), as firms
exposing to higher litigation risk tend to recognize bad news promptly to preempt lawsuits.
We further control for customer fundamentals, customer-supplier relationship, and customer
importance since these factors may affect the supplier’s accounting conservatism. Specifically,
we include the customers’ performance (CU_ROA), leverage (CU_LEV), and Altman’s Z-score
(CU_ALTMAN). In general, suppliers face lower business risks when their major customers
exhibit higher profitability, lower leverage, and lower bankruptcy risk, which reduce lenders’
potential losses and necessitate a lower level of accounting conservatism. We also include the
17
Although Khan and Watts (2009) predict a positive effect of MB on CSCORE, they also point out that “buffer
problem” (Roychowdhury and Watts 2007) may make the effect hard to observe.
17
number of years that the supplier-customer relationship has existed (ln(DURATION)) and the
supplier’s sales to the specific customer divided by the supplier’s total sales (CU_IMPT). A
closer customer-supplier relationship and a heavier reliance on major customers imply greater
bargaining power by major customers, leading to greater conservatism demand (Hui et al. 2012).
Detailed definitions for these control variables are provided in the Appendix.
Finally, we include industry and year fixed effects to account for cross-industry variations
and inter-temporal trends in accounting conservatism.
Descriptive statistics
Table 1, Panel A reports the number of suppliers that have lending relationship with supply-
chain lenders versus those with non-supply-chain lenders by loan origination year. On average,
about 62% (700 of 1,135) suppliers borrow from supply-chain lenders and 38% (435 of 1,135)
borrow from non-supply-chain lenders. The number of observations declines after 2007,
reflecting the adverse consequence of financial crisis on corporate debt financing activities.
Table 1, Panel B presents the sample by Fama-French industry classification (excluding utility
and financial industries). We find that suppliers borrowing from supply-chain lenders spread
broadly across all industries, with the fraction ranging from as low as 3.3% in the consumer
durables industry to as high as 25.9% in the manufacturing industry.
[Insert TABLE 1 Here]
Table 2 reports summary statistics for observations with supply-chain lenders and those with
non-supply-chain lenders. Consistent with H1, CSCORE for suppliers borrowing from supply-
chain lenders (mean = 0.11 and median = 0.12) is significantly lower (p-value < 0.01) than the
corresponding CSCORE for suppliers borrowing from non-supply-chain lenders (mean = 0.25
and median = 0.24). In addition, suppliers that borrow from supply-chain lenders are larger in
18
market capitalization, have longer company history, shorter investment cycle, higher market-to-
book ratio, lower stock return volatility and sales volatility, smaller bid-ask spreads, higher
leverage, and higher bankruptcy risk. These suppliers are also more likely to employ reputable
auditors and operate in litigious industries. Furthermore, customers from these suppliers tend to
have higher leverage, higher bankruptcy risks, longer business relationship with their suppliers,
and represent more important partners to their suppliers.
[Insert TABLE 2 Here]
The Pearson and Spearman correlations, presented in Table 3, confirm the univariate
evidence in Table 2. In particular, our variable of interest, CHAIN_LENDING, is negatively and
significantly correlated with CSCORE (Pearson and Spearman correlations are both -0.29, p-
value < 0.01). We next turn to multivariate analyses to further study the relation between these
two variables.
[Insert TABLE 3 Here]
IV. EMPIRICAL RESULTS
The effect of lender-customer lending relationship on borrowers’ accounting conservatism
To examine whether a pre-existing lending relationship with major customer(s) is associated
with lower accounting conservatism by the borrowing supplier, we estimate the following
regression model at the time that the lender originates loans to the supplier:
CSCOREi,t = β1+ β2CHAIN_LENDINGi,t + β3’CVi,t (2)
where i indexes firm and t indexes year. The dependent variable, CSCORE, measures the degree
of accounting conservatism by the borrowing suppliers. The variable of interest,
CHAIN_LENDING, intends to capture the effect of lender-customer lending experience with
major customer(s). CV is a vector of control variables on firm and industry characteristics, as
19
described in the previous section. Robust standard errors are clustered at the firm level and
adjusted for heteoskedasticity.
Table 4 reports the regression results. Consistent with H1, the coefficient estimate on
CHAIN_LENDING is negative and significant (coefficient = -0.041, p-value < 0.01). This result
suggests that relative to suppliers borrowing from non-supply-chain lenders, supplier firms that
borrow from supply-chain lenders would report less conservatively (measured by CSCORE) by
0.041, the magnitude of which accounts for 25.6% of the sample mean.
[Insert TABLE 4 Here]
Results on control variables are generally consistent with our predictions. Specifically,
supplier firms in our sample report more conservatively when they have smaller market
capitalization, shorter company history, higher leverage, lower growth, larger bid-ask spread,
higher leverage, higher bankruptcy risk, and higher likelihood of employing reputable auditors.
We also find that supplier firms report more conservatively when their customers have higher
financial leverage and are of greater importance in the supplier’s customer base.
Identification strategy
Although we include a battery of control variables in the estimation of Equation (2), we
cannot rule out the possibility that the negative relation between the lender-customer lending
relationship and suppliers’ conservatism is driven by omitted variables that correlate with both
the existence of lender-customer lending relationship and suppliers’ conservatism. We address
this endogeneity concern below.
Instrumental variable approach
We first adopt an instrumental variable (IV) approach that relies on an exogenous variation
in supply-chain lending relationship due to the geographic proximity between a supplier and its
20
major customer(s). Prior literature proposes that geographic proximity can magnify the supply-
chain contagion effect that supply chain partners tend to make similar business decisions
(McFarland, Bloodgood, and Payan 2008), as geographic proximity enhances firms’
opportunities to exchange ideas and be cognizant of other important information in their fields
(Feldman 1994; Audretsch and Feldman 1996). Hence, a supplier that is in geographic proximity
with its major customer(s) is likely to choose the same credit providers as its major customer(s).
On the other hand, geographic proximity should not directly affect the supplier’s choice of
accounting conservatism other than through affecting the lender-customer lending relationship.18
To implement the IV approach, we estimate the following two-stage model.
CHAIN_LENDINGi,,t = β1+ β2DISTANCEi,t-1 + β3’CVi,t-1 + εi,t (3)
CSCOREi,t = β1+ β2FITTED_ CHAIN_LENDINGi,t + β3’CVi,t + εi,t (4)
In the first-stage probit regression, we endogenize the supply-chain lending decision by
regressing CHAIN_LENDING on DISTANCE, the geographic distance between the supplier and
its major customer(s), and CV, the same set of control variables as those included in Equation (2).
The fitted value of CHAIN_LENDING is then used to explain supplier firms’ accounting
conservatism in the second-stage regression.
[Insert TABLE 5 Here]
Table 5 reports the estimation results. In the first-stage regression results, we find that
DISTANCE is negatively and significantly related with CHAIN_LENDING, consistent with
geographic proximity increasing the probability that the supplier borrows from the same lender
18
A stream of finance research argues that a close geographic distance between borrowers and lenders mitigates
lenders’ information acquisition and monitoring costs, thus reducing loan spread (Degryse and Ongena 2005) and
fostering relationship banking (Berger, Miller, Petersen, Rajan, and Stein 2005; Bharath et al. 2011). Note that, our
instrumental variable measures the distance between a supplier and its major customer(s), not the distance between a
supplier and its lender (which may affect the information asymmetry between the supplier and the lender). We are
unable to measure the supplier-lender distance, as Dealscan reports lenders’ headquarters, as opposed to the
branches that actually initiated the loans to the borrowing suppliers.
21
as its major customer(s).19
This finding supports that DISTANCE satisfies the relevance criterion
as an IV. We also confirm the orthogonality criterion, as regressing the second-stage residual on
DISTANCE yields a statistically insignificant adjusted R2 of -0.0003.
20 In terms of control
variables, supplier firms with longer company history, shorter investment cycle, higher market
liquidity, greater default risk, less profitable customers, higher leveraged customers, and long-
lasting supplier-customer relationship are more likely to borrow from the same lender as their
major customers.21
In the second-stage regression results, we find that the estimated coefficient on
FITTED_CHAIN_LENDING is negative and significant (coefficient = -0.028, p-value = 0.07).
The IV estimation results are consistent with a causal effect from supply-chain lending
relationship to suppliers’ accounting conservatism.
Propensity score matching approach
We also conduct a propensity score matching (PSM) analysis to further address the
endogeneity concern. PSM analysis allows us to identify a sample of suppliers that borrow from
non-supply-chain lenders but are otherwise comparable to the sample of suppliers borrowing
from supply-chain lenders. Specifically, we first estimate a probit model by regressing
19
Hui et al. (2012) show that customers have demand for the suppliers’ accounting conservatism. One particular
concern about our IV is that customers’ conservatism demand may vary with the geographic distance between
suppliers and customers. This concern is mitigated by the insignificant correlation between CSCORE and
DISTANCE. Moreover, Hui et al. (2012) show that customers’ conservatism demand increases with the bargaining
power the customer has over the supplier. To the extent that geographic proximity increases customers’ bargaining
power (e.g., losing a local customer is more costly than losing a customer located further away), the possible effect
of our IV on customer’s conservatism demand should work against us finding predicted results. 20
We also follow Frank (2000) and Larcker and Rusticus (2008) to compute the Impact Threshold for a
Confounding Variable (ITCV) to assess how closely an unobservable confounding variable would have to be
correlated with CHAIN_LENDING and CSCORE in order to overturn the OLS results in Table 4. The results suggest
we would need an omitted variable with an impact of at least 49% higher than that of any of our control variables to
change our inferences. We believe it is unlikely that a confounding variable will overturn the negative relation
between CSCORE and CHAIN_LENDING. 21
In untabulated results, we also control for customer-lender relationship, including the duration of lending
experience with the customers and the importance of customer loans in lenders’ loan portfolio. Adding these
controls does not qualitatively alter the results in Table 5, and these additional controls are statistically insignificant.
22
CHAIN_LENDING on firm and industry characteristics included in Equation (3). Next, for each
supplier with CHAIN_LENDING = 1, we select one supplier with CHAIN_LENDING = 0
without replacement that has the closest predicted probability (i.e., propensity score) of
borrowing from a supply-chain lender. This matching procedure yields 188 observations with
CHAIN_LENDING = 1 and 188 observations with CHAIN_LENDING = 0.
In Table 6, we report two diagnostic tests to verify the effectiveness of PSM procedures.
Table 6 Panel A shows that the differences in observable characteristics between supply-chain-
lending observations and matched non-supply-chain-lending observations are mostly statistically
insignificant. Table 6 Panel B reports that the difference in propensity scores between these two
samples is minimal. Together, these statistics suggest that PSM procedures have effectively
removed meaningful observable differences between the two samples and ensure that the
difference in accounting conservatism is likely to be only caused by the lender-customer lending
relationship.
[Insert TABLE 6 about here]
Table 6 Panel C presents the difference in CSCORE across the two samples. As shown, the
average CSCORE for suppliers borrowing from supply-chain lenders is 0.183, which is
significantly lower than the average CSCORE of 0.246 for suppliers borrowing from non-supply-
chain lenders (difference = -0.063, p-value < 0.01). Hence, the propensity score matching
analysis also support a negative effect of supply-chain lending relationship on suppliers’
accounting conservatism.
A placebo test
The endogeneity concern rests on the assumption that certain omitted variables are
correlated with both the lender-customer lending relationship and the supplier’s accounting
23
conservatism. To the extent that such omitted variables are stable over time, our baseline results
should hold even when the lender-customer lending relationship is established in the periods
after the lender originated loans to the supplier. In contrast, our hypothesis is built upon the
argument that supply-chain lenders’ information advantage at loan origination comes from the
proprietary information acquired via the pre-existing lending relationship with the borrowers’
customers. In other words, our argument is valid only when the lender-customer lending
relationship is established before the loan origination to the borrowing suppliers.
To implement the placebo test, we create an indicator variable,
PSEUDO_CHAIN_LENDING, that equals one if a supplier borrows from a lender before the
period that the same lender originates a loan to the supplier’s major customer(s), and zero if a
supplier borrows from a lender after a different lender originates a loan to its major customer(s)
(i.e., the same control sample with CHAIN_LENDING = 0). We identify 374 supplier-year
observations with PSEUDO_CHAIN_LENDING = 1 and 435 observations with
PSEUDO_CHAIN_LENDING = 0. In Table 7, we re-estimate Equation (3) after replacing
CHAIN_LENDING with PSEUDO_CHAIN_LENDING, and find an insignificant coefficient on
PSEUDO_CHAIN_LENDING. This insignificant finding helps rule out the possibility that
certain omitted time-invariant factors driving the supply-chain lending choice also affects the
borrowing suppliers’ conservatism.
[Insert TABLE 7 about here]
Cross-sectional analyses
We consider three factors that potentially affect the negative relation between supply-chain
lending choice and suppliers’ accounting conservatism. First, we examine whether information
opacity enhances the marginal effect of proprietary supply-chain information (obtained through a
24
pre-existing lender-customer relationship) on borrowers’ accounting conservatism. We use two
proxies for information opacity: information cost (INFCOST) following Duchin, Matsusaka, and
Ozbas (2010) and absolute value of discretionary accruals (|DACC|) based on Hutton, Marcus
and Tehranian (2009). INFCOST is an index created by ranking firms on different dimensions of
information acquisition costs including the number of analysts, analyst forecast dispersion, and
analyst forecast error. Discretionary accruals are estimated from Modified Jones’ Model. Higher
values of INFCOST and |DACC| indicate more opaque information environment.
We then estimate Equation (2) separately for firms (either customers or suppliers) with
above the median value of information opacity and those with below the median value. Table 8,
Panel A presents the results based on information opacity of the customers, and Table 8, Panel B
reports the results based on information opacity of the suppliers. Consistent with H2a and H2b,
in both panels we find that the negative relationship between CHAIN_LENDING and CSCORE is
statistically significant only in the subsamples with high information opacity. In both panels,
Chow tests show that the differences in the coefficient on CHAIN_LENDING across subsamples
are statistically significant.
[Insert TABLE 8 about here]
Next, we examine whether the quality of proprietary supply-chain information, proxied by
customer importance in suppliers’ sales (CU_IMPT) and the duration of lender-customer lending
relationship (CU_LYRS), strengthens the negative effect of supply-chain lending on the
borrowers’ accounting conservatism. We estimate Equation (2) separately for firms with above
median value of CU_IMPT or CU_LYRS and those with below median value. Table 9 presents
regression results. Consistent with H3a and H3b, the negative coefficient on CHAIN_LENDING
is only significant for the subsamples with greater customer importance or longer prior lending
25
experience with customers. Chow tests show that the differences in the coefficient on
CHAIN_LEANDING across subsamples are statistically significant.
[Insert TABLE 9 about here]
Finally, we examine the importance of credit-relevant information to our baseline results. To
measure credit-relevant information, we approximate the amount of trade credit extended to a
major customer (TRADE_CREDIT) as the supplier’s total accounts receivable multiplying the
percentage of total sales made to this customer, scaled by the supplier’s net worth. We estimate
Equation (2) separately for firms with above median value of TRADE_CREDIT and firms with
below median value. The results are reported in Table 10. Consistent with H4, the coefficient
estimate on CHAIN_LENDING is negative and significant only in the subsample with larger
amount of trade credit. Chow test shows that the difference in the CHAIN_LENDING coefficient
across subsamples is statistically significant.
[Insert TABLE 10 about here]
Robustness check using an alternative measure of conservatism
To lend further credence to our baseline results, we adopt an alternative measure of
accounting conservatism by augmenting Basu’s (1997) piece-wise linear regression, as specified
in Equation (1.1), to include CHAIN_LENDING and its interaction term with Di,t×Ri,t. H1
predicts that prior lending experience with major customers reduces lenders’ demand for, and the
supply of, borrowers’ accounting conservatism. Thus, we expect the coefficient on
CHAIN_LENDINGi,t×Di,t×Ri,t to be negative.
Table 11 reports the regression results. Confirming the main result, we find a negative and
significant coefficient on CHAIN_LENDINGi,t×Di,t×Ri,t (coefficient = -0.368, p-value < 0.10)
under column (1). In column (2), after adding control variables as in LaFond and Roychowdhury
26
(2008) and LaFond and Watts (2008), we find that the coefficient on
CHAIN_LENDINGi,t×Di,t×Ri,t remains negative and significant.22
[Insert TABLE 11 about here]
The effect of lender-customer lending relationship on borrowers’ covenant design
Our findings above support the notion that proprietary supply-chain information mitigates
the information gap between lenders and borrowers, thereby leading to less conservative
reporting by the borrowing suppliers. The access to proprietary supply-chain information,
coupled with borrowers’ less conservative accounting practices, can reduce the reliance on
accounting numbers in designing debt covenants. Accordingly, we may observe less frequent use
of accounting-based or financial covenants in the lending agreements between supply-chain
lenders and borrowing suppliers. Consistently, both univariate and multivariate results, reported
in Table 12 Panel A and Panel B, show that suppliers borrowing from supply-chain lenders are
subject to fewer financial covenants than suppliers borrowing from non-supply-chain lenders.
[Insert TABLE 12 about here]
Lenders often impose multiple monitoring mechanisms in the lending agreements, and
financial reporting practices can alter lenders’ tradeoff of alternative monitoring mechanisms
(Costello and Wittenberg-Moerman 2011). To further shed light on lenders’ decision making, we
examine the relation between supply-chain lending experience and other loan terms, such as
yield spread, collateral requirements, loan maturity, and general covenants. In untabulated results,
we find that a lender’s prior experience with the borrower’s major customer(s) is associated with
lower yield spread, less use of collateral requirements, and more general covenants. We find no
evidence that supply-chain lending experience affects loan maturity. These results generally
22
Our baseline results are also robust to the asymmetric accrual-cash flow measure proposed by Ball and
Shivakumar (2005).
27
support the beneficial role of proprietary supply-chain information in reducing lender-borrower
information asymmetry, which indirectly benefits borrowing suppliers through cheaper debt
financing costs and fewer restrictive collateral requirements.
The effect of supply-chain lending relationship on customers’ accountingn conservatism
While our hypotheses focus on the effect of supply-chain lending experience on suppliers’
accounting conservatism, our arguments imply a similar effect of supply-chain lending
experience on customers’ accounting conservatism. Testing the latter effect, however, is
empirically difficult due to the asymmetry disclosure requirements: firms are required to disclose
their major customers, but not their major suppliers. Prior evidence suggests that suppliers,
identified through the COMPUSTAT segment file, are not economically significant to their
major customers (Fee, Hadlock and Thomas 2006).23
Consequently, lenders’ information gains
from accessing the suppliers’ private information are limited, and may not have a discernible
effect on lenders’ demand for accounting conservatism by the borrowing customers.24
To shed light on this issue, we identify whether a major customer originated loans with a
lender when the same lender has an outstanding lending relationship with the major customer’s
supplier. Among 2,839 customer-originated loans, we find 60.1% of customers (1,706 of 2,839)
originated loans from the same lender as their suppliers with whom the lender has outstanding
loans. In contrast with our earlier results, Table 13 shows that the relationship between the pre-
existing lender-supplier lending relationship (CU_CHAIN_LENDING) and the customers’
conservatism is insignificant (coefficient on CU_CHAIN_LENDING = 0.002, p-value = 0.842).
This result, however, does not necessarily imply that a pre-existing lending relationship with
23
For our sample firms, total sales that suppliers made to their major customers comprise only 2.1% of the total
sales of those customers. 24
Our conjecture is in line with prior evidence. For example, Hertzel et al. (2008) find a stock market wealth effect
for suppliers resulting from their customers’ bankruptcy, but fail to find similar effect for customers resulting from
their suppliers’ bankruptcy.
28
suppliers has a minimal effect on lenders’ conservatism demand for borrowing customers, due to
substantial measurement errors in identifying major suppliers under the current asymmetry
disclosure requirements.
[Insert TABLE 13 about here]
V. CONCLUSION
Supply-chain network generates economic interdependence across business partners. By
lending to a supplier’s major customer(s), the lender gains access to proprietary supply-chain
information and enjoys a unique information advantage in assessing the supplier’s
creditworthiness. Given that accounting conservatism is costly to implement, we conjecture that
proprietary supply-chain information lowers lenders’ demand for, and therefore the supply of,
conservatism by substituting conservative accounting numbers in the loan contracting process.
Consistent with our conjecture, we find that when lenders have prior lending experience with the
suppliers’ major customers, borrowing suppliers practice less conservative reporting at loan
origination. This finding is more pronounced when the customers and suppliers operate in an
opaque information environment, and magnifies with customer importance, long-lasting lender-
customer relationship, and the significance of trade credit suppliers extend to major customers.
Our findings support the importance of proprietary supply-chain information in reducing
agency cost of debt and assist lenders’ decision making. The evidence also sheds light on the
interaction between lenders and supply-chain partners, and its implications for borrowers’
accounting practices during the lending process. Future research may extend our analyses to
examine the implications of proprietary supply-chain information for the secondary trading of
debt securities and the potential spillover effect to equity market participants.
29
REFERENCES
Ahmed, A. S., B. K. Billings, R. Morton, and M. Stanford-Harris. 2002. The role of accounting
conservatism in mitigating bondholder-shareholder conflicts over dividend policy and in
reducing debt costs. The Accounting Review 77(4): 867-890.
Audretsch, D. B., and M. P. Feldman. 1996. R&D spillovers and the geography of innovation
and production. American Economic Review 86(4): 253-273.
Baiman, S., and M. Rajan. 2002a. Incentive issues in inter-firm relationships. Accounting,
Organizations and Society 27: 213-238.
Baiman, S., and M. V. Rajan 2002b. The role of information and opportunism in the choice of
buyer-supplier relationships. Journal of Accounting Research 40(2): 247-278.
Ball, R., Kothari, S., Robin, A., 2000. The Effect of International Institutional Factors on
Properties of Accounting Earnings. Journal of Accounting and Economics 29: 1-51.
Ball, R., and L. Shivakumar. 2005. Earnings quality in U.K. private firms. Journal of Accounting
and Economics 39: 83-128.
Barth, J., G. Caprio, and R. Levine. 2004. Bank regulation and supervision: what works best?
Journal of Financial Intermediation 13(2): 205-248.
Basu, S. 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of
Accounting and Economics 24: 3-37.
Basu, S., Hwang, L., Jan, C., 2001. Differences in conservatism between Big Eight and Non-Big
Eight auditors. Working paper, Baruch College, City University of New York.
Beatty, A., J. Weber, and J. J. Yu. 2008. Conservatism and debt. Journal of Accounting and
Economics 45: 154-174.
Beneish, M.D., and E. Press. 1993. Costs of technical violation of accounting-based debt
covenants. The Accounting Review 68: 233-257.
Berger, A. N., N. Miller, M. Petersen, R. Rajan, and J. Stein. 2005. Does function follow
organizational form? Evidence from the lending practices of large and small banks.
Journal of Financial Economics 76(2): 237-269.
Bharath, S., S. Dahiya, A. Saunders, and A. Srinivasan. 2011. Lending Relationships and
loan contract terms. Review of Financial Studies 4: 1141–1203.
Blackwell, R. D., and K. Blackwell. 1999. The century of the consumer: Converting supply
chains into demand Chains' supply chain management review.
http://www.manufacturing.net/scm Accessed: 24/05/2004.
Caskey, J. and J. Hughes. 2012. Assessing the impact of alternative fair value measures on the
efficiency of project selection and continuation. The Accounting Review 87(2): 483-512.
Ciamarra, E. 2012. Monitoring by affiliated banker on boards of directors: Evidence from
corporate financing outcomes. Financial Management 41: 665-702.
Cohen, L. and A. Frazzini. 2008. Economic links and predictable returns. Journal of Finance 63
(4): 1977-2011.
Cooper, M., L. Ellram, J. Gardner, and A. Hanks. 1997. Meshing multiple alliances. Journal of
Business Logistics 18(1): 67-89.
Costello, A. and R. Wittenberg-Moerman. 2011. The impact of financial reporting quality on
debt contracting: Evidence from internal control weakness reports. Journal of Accounting
Research 49(1): 97-136.
Degryse, H. and S. Ongena. 2005. Distance, lending relationships and competition. Journal of
Finance 60(1): 231-266.
30
Duchin, R., J. G. Matsusaka, and O. Ozbas. 2010. When are outside directors effective? Journal
of Financial Economics 96: 195-214.
Engelberg, J., P. Gao, and C. A. Parsons. 2012. Friends with money. Journal of Financial
Economics 103: 169-188
Erkens, D., K. Subramanyam, and J. Zhang. 2013. Affiliated bankers on board and conservative
accounting. Working paper, University of Southern California.
Fee, C. E., C. Hadlock, and S. Thomas. 2006. Corporate equity ownership and the governance of
product market relationships. Journal of Finance 61(3): 1217-1251.
Feldman, M. P. 1994. The Geography of Innovation. Kluwer Academic Publishers, Boston.
Files, R. and U. Gurun. 2011. Lenders’ response to restatements along the supply chain. Working
paper, University Texas at Dallas.
Frank, K., 2000. Impact of a confounding variable on a regression coefficient. Sociological
Methods and Research 29(2): 147–194.
Gigler, F., C. Kanodia, H. Sapra, and R. Venugopalan. 2009. Accounting conservatism and the
efficiency of debt contracts. Journal of Accounting and Economics 47(3): 767-797.
Gormley, T., B. Kim, and X. Martin. 2012. Do firms adjust their timely loss recognition in
response to changes in the banking industry? Journal of Accounting Research 50(1): 159-
196.
Guay, W. and R. Verrecchia. 2006. Discussion of an economic framework for conservative
accounting and Bushman and Piotroski (2006). Journal of Accounting and Economics 42:
149-165.
Hertzel, M. G., Z. Li, M. S. Officer, and K. J. Rodgers. 2008. Inter-firm linkages and the wealth
effects of financial distress along the supply chain. Journal of Financial Economics 87:
374-387.
Hui, K. W., S. Klasa, and E. Yueng. 2012. Corporate suppliers and customers and accounting
conservatism. Journal of Accounting and Economics 53, 115-135.
Hutton, A., A. Marcus, and H. Tehranian. 2009. Opaque financial reports, R2 and crash risk.
Journal of Financial Economics 94: 67-86.
Jensen, M. and W. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs and
ownership structure. Journal of Financial Economics 3: 305-360.
Khan, M., and R. Watts. 2009. Estimation and empirical properties of a firm-year measure of
accounting conservatism. Journal of Accounting and Economics 48: 132-150.
Koh, W., S. Teoh, and M. Tham. 2011. How major customers affect supplier loan yield and
covenants. Working paper, University of California at Irvine and Nanyang Technological
University.
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.
Kulp, S., H. Lee, and E. Ofek. 2004. Manufacturing benefits from information integration with
retail customers. Management Science 50(4): 431-444.
Lafond, R., S. Roychowdhury. 2008. Managerial ownership and accounting conservatism.
Journal of Accounting Research 46: 101–35.
Lafond, R., and R. Watts. 2008. The information role of conservative financial statements. The
Accounting Review 83: 447-478.
Lambert, D., and M. C. Cooper 2000. Issues in supply chain management. Industrial Marketing
Management 29(1): 65-83.
31
Larcker, D., and T. Rusticus. 2008. On the use of instrumental variables in accounting research.
Journal of Accounting and Economics 49; 186-205.
Lee, H., K. So, and C. Tang. 2000. The value of information sharing in a two-level supply chain.
Management Science 46(5): 626-643.
Leftwich, R. 1983. Accounting information in private markets: Evidence from private lending
agreements. The Accounting Review 58(1): 23-42.
Li, N., and Z. Yang. 2011. Customer relationship and debt contracting. Working paper,
University of Texas at Dallas.
McFarland, R. G., J. M. Bloodgood, and J. M. Payan. 2008. Supply chain contagion. Journal of
Marketing 72(2): 63-79.
Mian, S., and C. Smith. 1992. Accounts receivable management policy: Theory and evidence.
Journal of Finance 48: 169-200.
Murfin, J., and K. Njoroge. 2011. Small lending big: the real effects of trade credit demands on
constrained suppliers. Working paper, Yale University.
Pandit, S., C. Wasley, and Z. Tzachi. 2011. Information externalities along the supply chain: the
economic determinants of suppliers’ stock price reaction to their customers’ earnings
announcements. Contemporary Accounting Research 28: 1304-1343.
Petersen, M., and R. G. Rajan. 1997. Trade credit: Theories and evidence. Review of Financial
Studies 10(3): 661-691.
Raman, K., and H. Shahrur. 2008. Relationship-specific investments and earnings management:
Evidence on corporate suppliers and customers. The Accounting Review, 83(4): 1041-1081.
Roychowdhury, S. 2010. Discussion of “Acquisition profitability and timely loss
recognition. Journal of Accounting and Economics 49(1): 179-183.
Roychowdhury, S., and R. L. Watts. 2007. Asymmetric timeliness of earnings, market-to-book
and conservatism in financial reporting. Journal of Accounting and Economics 44: 2-31.
Schloetzer, J. 2012. Process integration and information sharing in supply chains. The
Accounting Review 87(3): 1005-1032.
Standard and Poor's. 2007. A guide to the loan market. Standard & Poor's, New York.
Sufi, A. 2007. Information asymmetry and financing arrangements: evidence from syndicated
loans. The Journal of Finance 62: 629-668.
Titman, S. 1984. The effect of capital structure on a firm's liquidation decision. Journal of
Financial Economics 13: 137-151.
Watts, R. L. 1993. A proposal for research on conservatism. Unpublished, 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., and J. Zimmerman. 1983. Agency problems, auditing and the theory of the firm:
some evidence. Journal of Law and Economics 26(3): 613-633.
Watts, R. L., and J. Zimmerman. 1986. Positive accounting theory. Prentice-Hall, London.
Wittenberg-Moerman, R. 2008. The role of information asymmetry and financial reporting
quality in debt trading: Evidence from the secondary loan market. Journal of Accounting
and Economics 46: 240-260.
Zhang, J. 2008. The Contracting conservatism of accounting conservatism to lenders and
borrowers. Journal of Accounting and Economics 45: 27-54.
32
APPENDIX
Variable Definitions
Supply-chain lending relationship
CHAIN_LENDINGit = A dummy variable that equals one if a supplier firm borrows from a
supply-chain lender and zero if a supplier firm borrows from a non-
supply-chain lender, where a supplier chain lender refers to a lender who
acts as a lead arranger in an outstanding lending relationship with any one
of the supplier’s major customers prior to initiating loans to the suppliers.
CU_CHAIN_LENDINGit = A dummy variable that equals one if a major customer firm borrows from
a supply-chain lender and zero if a major customer firm borrows from a
non-supply-chain lender, where a supply chain lender refers to a lender
who acts as a lead arranger in an outstanding lending relationship with
any one of the major customers’ supplier firm prior to initiating loans to
the major customer firm.
PSEUDO_CHAIN_LENDINGit = A dummy variable that equals one if the lender of a supplier firm
establishes a lending relationship with the supplier’s major customer after
lending to the supplier firm, and zero if a supplier firm borrows from a
lead arranger who do not have a pre-existing lending relationship with the
customers.
Customer-supplier relationship
DISTANCEit-1 = The distances (in 1,000 miles) between the corporate headquarter of the
supplier firm i and that of the major customer in fiscal year t-1.
ln(DURATION)it = Natural logarithm of one plus DURATION, the number of years that a
supplier-major customer relationship has existed as of the end of fiscal
year t.
CU_IMPTit = The relative importance of the customer to the supplier firm i in fiscal
year t, measured as the total sales made to the major customer during the
year divided by the supplier’s total annual sales (SALE).
SUP_IMPTit = The relative importance of the supplier firm i to the customer firm in
fiscal year t, measured as the total sales the supplier firm made to the
major customer during the year divided by the total annual sales (SALE)
of the customer firm in the year.
Supplier characteristics
(In Table 13, variables for the supplier’s major customer(s) are prefixed with “CU”.) CSCOREit = The firm-year conservatism measure based on Khan and Watts (2009).
The following annual cross-sectional model is first estimated: Xi,t= β1 +
β2Di,t + Ri,t (μ1+ μ2ln(MVE)i,t + μ3MBi,t + μ4LEVi,t) + Di,t×Ri,t (λ1+ λ2
ln(MVE)i,t+ λ3MBi,t + λ4LEVi,t) + (δ1 ln(MVE)i,t + δ2MBi,t + δ3LEVi,t +
δ4Di,t× ln(MVE)i,t+ δ5Di,t×MBi,t+ δ6Di,t×LEVi,t) + εi,t. CSCORE for
supplier firm i in fiscal year t is then computed as λ1+ λ2 ln(MVE)i,t+
λ3MBi,t + λ4LEVi,t..
ln(MVE)it = Natural logarithm of market value of equity MVE (CSHO × PRCC_F)
measured at the end of fiscal year t for firm i.
33
ln(FIRM_AGE) = Natural logarithm of the age of supplier firm i at the end of year t,
measured as the number of years supplier firm i has been listed by the
Center for Research in Security Prices (CRSP).
INVCYCLEit = Depreciation (DP) in fiscal year t divided by total assets (AT) for supplier
firm i. MBit
= Market value of equity, calculated as (CSHO × PRCC_F), divided by
book value of equity (CEQ) at the end of fiscal year t for supplier firm i.
RETVOLit = The standard deviation of daily stock returns for supplier firm i in fiscal
year t.
SALEVOLit = The standard deviation of the natural log of revenues for supplier firm i
measured from year t-5 to year t-1.
BID_ASKit = The average daily closing bid-ask spread for firm i in fiscal year t.
LEVit = Total book value of debt (DLTT + DLC) divided by total assets (AT)
measured at the end of fiscal year t for supplier firm i.
ALTMAN = Altman’s bankruptcy score, calculated based on 1.2 ×(ACT-LCT)/AT +
1.4 × (RE/AT) + 3.3 × (OIADP/AT) + 0.6 * (PRCC_F × CSHO)/LT +
SALE/AT for supplier firm i in in year t.
BIG_FOURit = An indicator variable that equals 1 if the auditor is one of the big four (or
five) auditing firms, and zero otherwise.at the end of fiscal year t for
supplier firm i.
LITIGATIONit = A dummy variable that equals one if the supplier firm i’s main operations
in year t are in a highly litigious industry (biotechnology (2833-2836 and
8731-8734), computers (3570-3577 and 7370-7374), electronics (3600-
3674), and retail (5200-5961) in fiscal year t, and zero otherwise (based
on Rogers and Stocken 2005).
ROAit = Net income before extraordinary items (NI) in fiscal year t, scaled by the
total assets (AT) at the beginning of year t for supplier firm i.
Xit = Net income before extraordinary items (NI) in fiscal year t, scaled by the
market value of equity at the beginning of year t for supplier firm i.
Rit = 12-month compounded returns starting 9 months before the end of fiscal
year t for supplier firm i.
Dit = A dummy variable that equals one if Ri,t is negative and zero otherwise.
Additional variables used in partitioning analysis
(In Table 8, variables for the supplier’s major customer(s) are prefixed with “CU”.)
INFCOSTit = A measure of supplier firm i’s information cost in fiscal year t. Following
Duchin et al. (2010), we first create an information index by averaging
the firm’s reverse percentile rank of analyst coverage (i.e., the number of
analysts following the firm), percentile rank of analyst forecast
dispersion, and percentile rank of analyst forecast error in the sample
firms of fiscal year t. We then scale the index to range from zero to one.
|DACCit| = Prior three years sum of absolute value of the discretionary accruals for
supplier firm i in year t, where the discretionary accruals is computed
based on modified Jones model.
CU_LYRSi,t = The number of months (divided by 12) that a lead bank has had lending
experience with a supplier i’s customer(s) when a new loan was initiated
to the supplier i in year t.
34
TRADE_CREDITit
= The relative importance of the trade credit extended to the customer for
supplier firm i, calculated as (the total sales made to the customer firm/
total sales of the supplier firm) × (accounts receivable / net worth of the
supplier firm), where net worth of the supplier firm is the total assets (AT)
minus total liabilities (LT) minus the intangible assets (INTAN) of the
supplier firm i.
Loan characteristics
FINCOV it = The number of financial covenants divided by the total number of
covenants.
ln(LOAN_SIZE)it = Natural logarithm of the total loan amount for supplier i in year t from
the lender.
ln(MATURITY)it = Natural logarithm of a loan’s maturity in number of months for supplier
i in year t. SECURE_LOANit = A dummy variable that equals one if a loan agreement for supplier i in
year t has collateral and zero otherwise.
RESOLVERit = A dummy variable that equals one if a loan agreement for supplier i in
year t is on revolving basis and zero otherwise.
SYNDICATED_LOANit = A dummy variable that equals one if a loan for supplier i in year t is a
syndicated loan by multiple lenders and zero otherwise.
35
TABLE 1
Sample Distribution
Panel A: Sample distribution by loan initiation year
Year CHAIN_LENDING = 1 CHAIN_LENDING = 0
N % N %
1988 0 0.0% 6 1.4% 1989 1 0.1% 5 1.2% 1990 3 0.4% 6 1.4% 1991 10 1.4% 7 1.6% 1992 8 1.1% 12 2.8% 1993 5 0.7% 14 3.2% 1994 11 1.6% 10 2.3% 1995 13 1.9% 15 3.5% 1996 20 2.9% 25 5.8% 1997 17 2.4% 29 6.7% 1998 18 2.6% 27 6.2% 1999 28 4.0% 26 6.0% 2000 50 7.1% 31 7.1% 2001 45 6.4% 29 6.7% 2002 60 8.6% 46 10.6% 2003 59 8.4% 37 8.5% 2004 73 10.4% 37 8.5% 2005 68 9.7% 23 5.3% 2006 65 9.3% 18 4.1% 2007 73 10.4% 14 3.2% 2008 51 7.3% 11 2.5% 2009 20 2.9% 7 1.6% 2010 2 0.3% 0 0.0% Total 700 100% 435 100%
Panel B: Sample distribution by Fama-French 12 industry
Fama-French’s 12 industries CHAIN_LENDING = 1 CHAIN_LENDING = 0
N % N %
Consumer Non-Durables 167 23.9% 48 11.0%
Consumer Durables 23 3.3% 24 5.5%
Manufacturing 181 25.9% 103 23.7%
Energy and Coal Extraction 55 7.9% 36 8.3%
Chemicals and Allied products 48 6.9% 13 3.0%
Business Equipment 119 17.0% 136 31.3%
Wholesale, Retail, and Some Services 40 5.7% 34 7.8%
Healthcare, Medical Equipment, and Drugs 67 9.6% 41 9.4%
Total 700 100% 435 100%
This table reports sample distribution for 1,135 supplier-year observations by new loan issuance year from 1988
to 2010 in Panel A and by industry in Panel B, conditional on whether a supplier borrows from a bank that has
outstanding lending relationship with at least one of the supplier’s major customer(s) (CHAIN_LENDING).
36
TABLE 2
Summary Statistics
Variables CHAIN_LENDING = 1 CHAIN_LENDING = 0 t-test of mean
difference
z test of median
difference Mean Median Mean Median
CSCORE 0.11 0.12 0.25 0.24 -10.24 ***
-9.88 ***
MVE (billion$) 8.23 1.05 2.52 0.41 5.73 ***
17.53 ***
FIRM_AGE 23.32 15.50 12.39 8.75 11.77 ***
9.23 ***
INVCYCLE 0.05 0.04 0.06 0.05 -5.01 ***
-5.15 ***
MB 3.14 2.16 2.65 1.89 2.89 ***
3.19 ***
RETVOL 0.12 0.10 0.18 0.16 -12.11 ***
-13.41 ***
SALEVOL 0.04 0.03 0.09 0.06 -7.02 ***
-10.35 ***
BID_ASK 0.00 0.00 0.02 0.01 -13.85 ***
-15.31 ***
LEV 0.28 0.26 0.24 0.19 2.82 ***
3.75 ***
ALTMAN 3.52 2.68 5.23 3.47 -5.50 ***
-5.55 ***
BIG_FOUR 0.91 1.00 0.87 1.00 2.05 **
2.12 **
LITIGATION 0.25 0.00 0.31 0.00 -2.37 **
-2.41 **
CU_ROA 0.07 0.08 0.06 0.07 1.55
1.57
CU_LEV 0.24 0.26 0.22 0.21 2.89 ***
3.89 ***
CU_ALTMAN 3.29 2.87 3.78 3.59 -4.87 ***
-5.04 ***
DURATION 4.71 4.00 3.19 3.00 7.49 ***
6.67 ***
CU_IMPT 0.19 0.16 0.22 0.17 -3.11 ***
-2.75 ***
DISTANCE (1,000 mile) 0.90 0.73 1.02 0.89 -2.66 ***
-2.08 **
This table reports the descriptive statistics of the variables used in the regression analyses of the effect of supply chain lending relationship on accounting
conservatism of the borrowing suppliers. The sample includes 1,135 supplier-year observations that initiated a loan from 1988 to 2010. See the Appendix
for variable definitions. *, **, and *** denote significance levels at 10%, 5% and 1% respectively based on two-tailed t (z) test statistics.
37
TABLE 3
Spearman/Pearson Correlation Matrix
Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18)
(1) CSCORE 1 -0.29 -0.68 -0.24 0.08 -0.4 0.54 0.1 0.58 0.04 -0.04 0.19 -0.09 -0.02 0.16 0.02 -0.18 0.11 (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (0.15) (0.17) (<.01) (<.01) (0.49) (<.01) (0.43) (<.01) (<.01)
(2) CHAIN_LENDING -0.29 1 0.52 0.28 -0.15 0.11 -0.4 -0.16 -0.43 0.11 -0.16 -0.06 0.02 0.05 0.12 -0.02 0.2 -0.08 (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (0.03) (0.41) (0.12) (<.01) (0.42) (<.01) (0.01)
(3) ln(MVE) -0.3 0.19 1 0.32 -0.11 0.48 -0.57 -0.13 -0.62 -0.01 0.02 -0.23 0.08 0.05 -0.08 0.09 0.13 -0.14 (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (0.62) (0.56) (<.01) (0.01) (0.08) (0.01) (<.01) (<.01) (0.00)
(4) ln(FIRM_AGE) -0.26 0.28 0.32 1 -0.16 0.03 -0.36 -0.28 -0.39 0.04 -0.17 -0.06 0.03 0.02 -0.03 0.03 0.18 -0.14 (0.00) (0.00) (0.00) (0.00) (0.30) (0.00) (0.00) (<.01) (0.19) (<.01) (0.06) (0.39) (0.41) (0.26) (0.34) (<.01) (<.01)
(5) INVCYCLE 0.09 -0.15 -0.1 -0.21 1 0.07 0.2 0.06 0.2 0.01 -0.04 -0.03 0.04 -0.04 -0.05 -0.12 -0.11 0 (<.01) (<.01) (<.01) (<.01) (<.01) (0.02) (<.01) (0.03) (<.01) (0.67) (0.13) (0.30) (0.19) (0.15) (0.13) (<.01) (<.01) (0.94)
(6) MB -0.25 0.08 0.2 0.05 0.04 1 -0.22 -0.06 -0.22 0.05 0.37 -0.06 0.1 0.03 -0.09 0.12 0.01 0.09 (<.01) (0.01) (<.01) (0.13) (0.16) (<.01) (0.06) (<.01) (0.11) (<.01) (0.03) (<.01) (0.37) (<.01) (<.01) (0.69) (<.01)
(7) RETVOL 0.45 -0.36 -0.25 -0.33 0.19 -0.01 1 0.25 0.79 -0.06 0.12 0.11 0.05 -0.08 -0.06 0.02 -0.24 0.11 (<.01) (<.01) (<.01) (<.01) (<.01) (0.74) (<.01) (<.01) (0.03) (<.01) (<.01) (0.09) (0.01) (0.06) (0.47) (<.01) (<.01)
(8) SALVOL 0.08 -0.14 -0.11 -0.24 0.1 -0.11 0.17 1 0.25 -0.08 0.15 0.09 0.14 -0.1 -0.12 -0.03 -0.18 0.09 (0.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (0.01) (<.01) (<.01) (<.01) (<.01) (<.01) (0.27) (<.01) (<.01)
(9)
BID_ASK 0.54 -0.43 -0.25 -0.34 0.2 0.19 0.74 0.17 1 -0.09 0.19 0.15 0.08 -0.1 -0.04 0.05 -0.25 0.14 (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (<.01) (0.01) (<.01) (0.14) (0.13) (<.01) (<.01)
(10) LEV 0.06 0.08 -0.06 0.02 0.02 0.21 0.01 -0.02 -0.01 1 -0.6 -0.01 -0.09 0.05 0.2 -0.1 0.05 -0.06 (0.06) (<.01) (0.04) (0.57) (0.61) (<.01) (0.67) (0.45) (0.63) (<.01) (0.80) (<.01) (0.13) (<.01) (<.01) (0.08) (0.04)
(11)
ATLMAN -0.01 -0.18 0.03 -0.17 -0.02 -0.06 0.12 0.15 0.17 -0.38 1 0.07 0.21 -0.05 -0.17 0.18 -0.07 0.18 (0.65) (<.01) (0.38) (<.01) (0.55) (0.05) (<.01) (<.01) (<.01) (<.01) (0.02) (<.01) (0.08) (<.01) (<.01) (0.01) (<.01)
(12) BIG_FOUR 0.18 -0.06 -0.09 -0.06 -0.05 0.09 0.06 0.09 0.13 -0.01 0.05 1 0.02 -0.03 0.05 -0.05 -0.01 0.13 (<.01) (0.03) (<.01) (0.04) (0.08) (<.01) (0.05) (<.01) (<.01) (0.62) (0.12) (0.40) (0.39) (0.10) (0.11) (0.76) (<.01)
(13) LITIGATION -0.09 0.02 0.11 0.03 0.01 0.05 0.03 0.14 0.05 -0.09 0.14 0.02 1 -0.03 -0.17 0.16 -0.06 0.08 (<.01) (0.41) (<.01) (0.39) (0.84) (0.11) (0.39) (<.01) (0.09) (<.01) (<.01) (0.40) (0.36) (<.01) (<.01) (0.03) (<.01)
(14) CU_ROA -0.05 0.05 0.02 0.03 -0.04 -0.04 -0.1 -0.06 -0.13 0.03 -0.01 -0.03 0.01 1 -0.11 0.49 0 0.07 (0.11) (0.11) (0.55) (0.33) (0.15) (0.20) (<.01) (0.03) (<.01) (0.32) (0.71) (0.40) (0.99) (<.01) (<.01) (0.96) (0.02) (15) CU_LEV 0.16 0.09 -0.05 -0.06 0 0.12 -0.05 -0.08 -0.03 0.17 -0.16 0.07 -0.15 -0.22 1 -0.28 -0.07 -0.05 (<.01) (<.01) (0.10) (0.06) (0.98) (<.01) (0.11) (0.01) (0.25) (<.01) (<.01) (0.01) (<.01) (<.01) (<.01) (0.01) (0.10) (16) CU_ALTMAN
0.01 -0.06 0.05 -0.05 -0.08 -0.01 0.08 -0.01 0.09 -0.11 0.12 -0.05 0.12 0.42 -0.3 1 -0.12 0.1 (0.62) (0.06) (0.08) (0.09) (0.01) (0.76) (<.01) (0.75) (<.01) (<.01) (<.01) (0.11) (<.01) (<.01) (<.01) (<.01) (<.01) (17) ln(DURATION) -0.18 0.2 -0.01 0.18 -0.11 0.05 -0.19 -0.18 -0.21 0.05 -0.08 -0.01 -0.07 0 -0.07 -0.13 1 0.12 (<.01) (<.01) (0.83) (<.01) (<.01) (0.07) (<.01) (<.01) (<.01) (0.10) (0.01) (0.62) (0.03) (0.93) (0.01) (<.01) (<.01) (18) CU_IMPT 0.12 -0.1 -0.08 -0.15 0.05 0.05 0.1 0.17 0.13 -0.07 0.18 0.16 0.05 0.04 -0.04 0.03 0.06 1 (<.01) (<.01) (0.01) (<.01) (0.10) (0.10) (<.01) (<.01) (<.01) (0.03) (<.01) (<.01) (0.10) (0.18) (0.20) (0.37) (0.06)
This table shows Spearman and Pearson correlations among selected variables for 1,135 supplier-year observations that initiated a loan from 1988 to 2010. The
upper (lower) right triangle of the matrix shows Spearman (Pearson) correlations. See the Appendix for variable definitions. Two-tailed p-values are reported in
parentheses.
38
TABLE 4
Supply-Chain Lending and Accounting Conservatism
Independent variables Predicted sign Dependent variable = CSCORE
Intercept ? 0.248*** (5.24)
CHAIN_LENDING - -0.041*** (-3.44)
ln(MVE) - -0.001*** (-3.77)
ln(FIRMAGE) - -0.012* (-1.68)
INVCYCLE - -0.153 (-0.89)
MB ? -0.011*** (-5.57)
RETVOL + 0.057 (0.56)
SALEVOL + -0.002 (-0.05)
BID_ASK + 5.000*** (8.24)
LEV + 0.087** (2.47)
ALTMAN - -0.002* (-1.78)
BIG_FOUR + 0.075*** (4.25)
LITIGATION + -0.050 (-1.43)
CU_ROA - 0.010 (0.09)
CU_LEV + 0.109** (2.26)
CU_ALTMAN - -0.001 (-0.18)
ln(DURATION) - -0.005 (-0.64)
CU_IMPT ? 0.107***
(2.92) Year fixed effects Included Industry fixed effects Included Number of observations 1,135 Adjusted R
2 0.682
This table reports the OLS regression results on the effect of supply chain lending relationship on accounting
conservatism for the borrowing suppliers. The sample includes 1,135 supplier-year observations that initiated a
loan from 1988 to 2010. The dependent variable is CSCORE. See the Appendix for variable definitions. Year and
industry fixed effects are included in the regressions, but their coefficients are not reported for brevity. Standard
errors clustered by firm and adjusted for heteroskedasticity are reported in parentheses. *, **, and *** denote
significant at 10%, 5%, and 1% respectively based on two-tailed t test statistics. See Appendix 3 for variable
definitions.
39
TABLE 5
Supply-Chain Lending and Accounting Conservatism: Instrumental Variable Approach
Independent variables First-stage:
CHAIN_LENDING
Second-stage: CSCORE
Intercept 0.805*** -0.030 (6.90) (-0.29)
DISTANCE -0.046*** (-2.86)
FITTED_CHAIN_LENDING -0.028* (-1.82)
ln(MVE) 0.000 -0.001*** (0.07) (-2.95)
ln(FIRMAGE) 0.003*** -0.001* (2.69) (-1.93)
INVCYCLE -0.871** -0.088 (-2.01) (-0.44)
MB 0.005 -0.009*** (1.35) (-5.09)
RETVOL -0.291 0.137 (-1.33) (1.32)
SALEVOL -0.036 0.006 (-0.55) (0.24)
BID_ASK -8.243** 6.069*** (-6.09) (5.17)
LEV -0.038 0.084** (-0.45) (2.53)
ALTMAN -0.011*** -0.002 (-3.38) (-1.42)
BIG_FOUR 0.013 0.053*** (0.29) (3.43)
LITIGATION 0.094 -0.079** (0.98) (-2.34)
CU_ROA -0.607* -0.064 (-1.87) (-0.43)
CU_LEV 0.308** 0.174*** (2.37) (2.60)
CU_ALTMAN 0.006 0.004 (0.79) (1.04)
ln(DURATION) 0.011** -0.004* (2.47) (-1.71)
CU_IMPT -0.015 0.089** (-0.16) (2.49)
Year fixed effects Yes Yes Industry fixed effects Yes Yes Number of observations 1,135 1,135 Adjusted R
2 0.344 0.619
This table reports the two-stage regression results on the effect of supply chain lending relationship on accounting
conservatism for the borrowing suppliers. The sample includes 1,135 supplier-year observations that initiated a loan
from 1988 to 2010. The dependent variable in the first stage is CHAIN_LENDING. The dependent variable in the
second stage is CSCORE. The predicted sign of the coefficient estimates is presented for first-stage/second-stage model
respectively. See the Appendix for variable definitions. Year and industry fixed effects are included in the regressions,
but their coefficients are not reported for brevity. Standard errors adjusted for heteroskedasticity are reported in
parentheses. *, **, and *** denote significance levels at 10%, 5% and 1% respectively based on two-tailed t test
statistics.
40
TABLE 6
Supply-Chain Lending and Accounting Conservatism: Propensity Score Matching
Panel A: Summary statistics of the matched samples
Variables CHAIN_LENDING = 1
(N=188)
CHAIN_LENDING = 0
(N=188) T-test of
mean difference Mean Mean
ln(MVE) 6.053 5.652 1.56 ln(FIRMAGE) 2.575 2.563 0.21 INVCYCLE 0.053 0.057 -1.14 MB 2.629 2.794 -0.56 RETVOL 0.152 0.155 -0.36 SALEVOL 0.055 0.062 -1.26 BID_ASK 0.029 0.033 -2.39**
LEV 0.744 0.674 0.56 ALTMAN 3.927 3.775 0.36 BIG_FOUR 0.898 0.877 0.65 LITIGATION 0.305 0.294 0.23 CU_ROA 0.063 0.071 -1.59 CU_LEV 0.281 0.251 0.86 CU_ALTMAN 3.568 3.664 -0.91 ln(DURATION) 1.448 1.270 2.94*** CU_IMPT 0.196 0.210 -1.11 Panel B: Summary statistics of the estimated propensity score
Propensity Scores No. of Obs. Min. P5 P50 Mean S.D. P95 Max.
CHAIN_LENDING =1 188 0.024 0.078 0.661 0.603 0.240 0.892 0.962 CHAIN_LENDING =0 188 0.004 0.105 0.600 0.565 0.241 0.871 0.912 Pair Difference ––– 0.000 0.001 0.036 0.075 0.097 0.310 0.464
Panel C: Difference in CSCORE between the matched samples Two Groups
Variable CHAIN_LENDING
= 1
(N = 188)
CHAIN_LENDING = 0
(N = 188) t-test of
mean difference
Wilcoxon test of
median
difference Mean Median Mean Median
CSCORE 0.183 0.164 0.246 0.233 -2.83*** 2.82***
This table reports the propensity score matching results on the effect of supply chain lending relationship on
accounting conservatism for the borrowing suppliers. To construct the matched sample, we first estimate the first-stage
regression, reported under Column (1) of Table 5, to generate a propensity score for each firm in the sample. For each
of the 700 supplier firms having supply chain lending relationship (“treatment firm”), we match with a supplier firm
without supply chain lending relationship (“control firm”) that has the same two-digit SIC code in the same year and
has the closet propensity score as the treatment firm (matched without replacement). For 188 treatment firms, we are
able to identify a matching firm from the control firms. Panel A reports the univariate comparisons between treatment
and control firms, and their corresponding t-statistics for the mean differences. Panel B reports the distribution of
estimated propensity scores for the treatment and control firms. Panel C reports the univariate comparisons in C-Score
between treatment and control firms, and their corresponding t-statistics (Wilcoxon z statistics) for the mean (median)
differences. See the Appendix for variable definitions. *, **, and *** denote significance levels at 10%, 5% and 1%
respectively based on two-tailed t (z) test statistics.
41
TABLE 7
Supply-Chain Lending and Accounting Conservatism: A Placebo Test
Independent variables Predicted sign
Dependent variable = CSCORE
Intercept ? -0.001 -0.001 (-0.01) (-0.00)
PSUDO_CHAIN_LENDING ? -0.024 (-1.64)
ln(MVE)
- -0.001 -0.001 (-0.57) (-0.57)
ln(FIRMAGE) - -0.011 -0.011 (-1.03) (-1.05)
INVCYCLE - -0.357* -0.371** (-1.85) (-1.98)
MB ? -0.010*** -0.010*** (-2.79) (-2.79)
RETVOL
+ -0.038 -0.038 (-0.37) (-0.37)
SALEVOL
+ -0.026 -0.026 (-0.68) (-0.68)
BID_ASK
+ 5.130*** 5.004*** (6.99) (6.78)
LEV
+ -0.022 -0.019 (-0.44) (-0.37)
ALTMAN
- -0.006** -0.006** (-2.52) (-2.52)
BIG_FOUR
+ 0.055*** 0.054*** (2.59) (2.59)
LITIGATION
+ -0.014 -0.013 (-0.36) (-0.34)
CU_ROA - 0.099 0.082
(0.74) (0.62)
CU_LEV + 0.180*** 0.185***
(3.37) (3.50)
CU_ALTMAN - -0.004 -0.005
(-1.13) (-1.26)
ln(DURATION) - -0.001 -0.002 (-0.04) (-0.11)
CU_IMPT
+ 0.089** 0.088** (1.98) (1.98)
Year fixed effects Included Included Industry fixed effects Included Included Number of observations 809 809 Adjusted R
2 0.581 0.583
This table reports results on a placebo test of the effect of supply chain lending relationship on accounting
conservatism of the borrowing suppliers. The sample has 809 supplier-year observations from 1988 to 2010,
including 435 supplier-years that initiated a lending relationship with a non-supply chain lenders
(CHAIN_LENDING = 0 and PSUDO_CHAIN_LENDING = 0) and 378 supplier-years that have their major
customers borrow after the same lead arranger initiated a loan to the supplier firm (PSUDO_CHAIN_LENDING =
1). See the Appendix for variable definitions. Standard errors clustered by firm and adjusted for heteroskedasticity
are reported in parentheses. *, **, and *** denote significance levels at 10%, 5% and 1% respectively based on
two-tailed t test statistics.
42
TABLE 8
Supply-Chain Lending and Accounting Conservatism: The Role of Information Opacity
Panel A: Information Opacity for the Customers
Independent variables Predicted
sign
Dependent variable = CSCORE
HIGH
CU_INFCOST
LOW
CU_INFCOST
HIGH
|CU_DACC|
LOW
|CU_DACC|
Intercept ? -0.005 0.155 0.302*** 0.048 (-0.04) (1.60) (4.32) (0.56)
CHAIN_LENDING - -0.038** -0.001 -0.055*** -0.004 (-2.56) (-0.05) (-3.13) (-0.21)
ln(MVE) - -0.001*** -0.007 -0.001* -0.001*** (-3.43) (-1.41) (-1.76) (-3.63)
ln(FIRMAGE) - -0.005 -0.014 -0.019* -0.009 (-0.57) (-1.24) (-1.81) (-0.82)
INVCYCLE - -0.503** 0.288 -0.069 -0.377 (-2.19) (0.90) (-0.23) (-1.54)
MB - -0.014*** -0.000 -0.008** -0.013*** (-6.80) (-0.08) (-2.57) (-5.14)
RETVOL - 0.148 0.107 -0.009 0.266 (1.19) (0.54) (-0.06) (1.58)
SALEVOL + -0.012 0.057 -0.015 -0.008 (-0.34) (0.81) (-0.32) (-0.18)
BID_ASK + 4.599*** 5.860*** 5.481*** 3.621*** (6.54) (4.97) (6.05) (4.34)
LEV + 0.117** 0.045 0.047 0.148** (2.58) (0.61) (0.82) (2.58)
ALTMAN - -0.001 -0.003 -0.003* 0.001 (-0.85) (-0.97) (-1.94) (0.53)
BIG_FOUR + 0.091*** 0.017 0.092*** 0.054** (4.23) (0.44) (3.28) (2.39)
LITIGATION + 0.018 -0.051 -0.047 0.009 (0.52) (-1.10) (-1.25) (0.17)
CU_ROA - 0.166 -0.382 -0.213 0.142 (1.05) (-1.58) (-1.39) (0.64)
CU_LEV + 0.086 0.033 0.024 0.239*** (1.49) (0.34) (0.36) (3.06)
CU_ALTMAN - -0.000 -0.002 -0.001 0.007 (-0.08) (-0.18) (-0.34) (1.12)
ln(DURATION) - -0.008 0.011 0.002 -0.021 (-0.65) (0.48) (0.14) (-1.42)
CU_IMPT ? 0.114** 0.065 0.080 0.129***
(2.36) (0.87) (1.46) (2.67) Year fixed effects Included Included Included Included Industry fixed effects Included Included Included Included Number of observations 820 315 544 591 Adjusted R
2 0.631 0.679 0.584 0.687
43
Panel B: Information Opacity for the Suppliers
Independent variables
Predicted
sign
Dependent variable = CSCORE
HIGH
INFCOST
LOW
INFCOST
HIGH
|DACC|
LOW
|DACC|
Intercept ? 0.258*** -0.493*** 0.316*** -0.041 (4.02) (-3.77) (4.34) (-0.14)
CHAIN_LENDING - -0.051*** 0.034 -0.072*** -0.010 (-3.59) (1.37) (-4.01) (-0.50)
ln(MVE) - -0.001*** -0.002** -0.000 -0.002*** (-3.14) (-2.15) (-0.53) (-4.36)
ln(FIRMAGE) + -0.018** -0.021 -0.020 -0.010 (-1.97) (-1.45) (-1.45) (-1.17)
INVCYCLE - -0.241 -0.009 -0.277 -0.110 (-1.14) (-0.03) (-0.95) (-0.48)
MB - -0.010*** -0.005 -0.011*** -0.012*** (-5.64) (-1.10) (-3.58) (-5.14)
RETVOL - -0.009 0.213 0.201 -0.108 (-0.07) (1.23) (1.37) (-0.75)
SALEVOL - -0.022 0.087* -0.038 0.017 (-0.54) (1.77) (-0.89) (0.39)
BID_ASK + 5.237*** 4.200*** 3.490*** 6.299*** (6.62) (4.03) (4.05) (7.47)
LEV - 0.120*** -0.062 0.103* 0.102** (3.15) (-0.72) (1.83) (2.23)
ALTMAN - -0.001 -0.004* -0.002 -0.001 (-1.01) (-1.73) (-1.04) (-0.41)
BIG_FOUR + 0.109*** 0.019 0.075*** 0.066** (4.71) (0.53) (3.06) (2.46)
LITIGATION + -0.028 0.048 -0.048 -0.009 (-0.69) (1.04) (-1.24) (-0.17)
CU_ROA - -0.026 -0.150 -0.125 0.088 (-0.17) (-0.51) (-0.77) (0.47)
CU_LEV + 0.070 0.302*** 0.075 0.082 (1.21) (2.99) (0.94) (1.28)
CU_ALTMAN - 0.000 -0.004 0.000 -0.000 (0.07) (-0.96) (0.12) (-0.06)
ln(DURATION) + 0.003 -0.023 -0.000 -0.013 (0.22) (-0.97) (-0.02) (-1.03)
CU_IMPT - 0.119** 0.099 0.092 0.156*** (2.43) (1.45) (1.51) (3.11)
Year fixed effects Included Included Included Included Industry fixed effects Included Included Included Included Number of observations 821 314 522 613 Adjusted R
2 0.653 0.628 0.587 0.669
This table reports the OLS regression results on the effect of supply chain lending relationship on accounting
conservatism of the borrowing suppliers, conditional on informational opacity for the customers (Panel A) and their
suppliers (Panel B). Information opacity is measured using informational cost (INFCOST) and the absolute value of total
accruals (|DACC|). The sample includes 1,135 supplier-year observations from 1988 to 2010. See the Appendix for
variable definitions. Standard errors clustered by firm and adjusted for heteroskedasticity are in parentheses. *, **, and
*** denote significance levels at 10%, 5% and 1% respectively based on two-tailed t test statistics.
44
TABLE 9
Supply-Chain Lending and Accounting Conservatism: The Role of Information Amount/Quality
Independent variables Predicted
sign
Dependent variable = CSCORE
HIGH
CU_IMPT
LOW
CU_IMPT
HIGH
CU_LYRS
LOW
CU_LYRS
Intercept ? 0.258*** -0.050 -0.206 0.254***
(3.66) (-0.82) (-1.62) (3.98)
CHAIN_LENDING - -0.056*** -0.032 -0.065*** -0.025
(-3.39) (-1.50) (-3.36) (-1.21)
ln(MVE) + -0.002*** -0.001 -0.002*** -0.001*
(-3.47) (-1.62) (-4.11) (-1.81)
ln(FIRMAGE) + -0.007 -0.018 -0.013 -0.012
(-0.66) (-1.55) (-1.23) (-1.20)
INVCYCLE - -0.295 0.061 -0.397 0.008 (-1.26) (0.19) (-1.55) (0.03)
MB ? -0.009*** -0.013*** -0.013*** -0.011*** (-3.63) (-3.61) (-4.83) (-3.99)
RETVOL + -0.048 0.106 0.148 0.038
(-0.32) (0.69) (0.94) (0.30)
SALEVOL - -0.031 0.045 -0.002 -0.000 (-0.74) (0.80) (-0.03) (-0.01)
BID_ASK - 5.056*** 5.203*** 3.833*** 5.548***
(5.98) (5.36) (4.30) (6.40)
LEV + 0.060 0.135* 0.158*** 0.038 (1.30) (1.81) (2.71) (0.75)
ALTMAN + -0.002** -0.002 -0.002 -0.001 (-2.14) (-0.72) (-0.92) (-0.86)
BIG_FOUR - 0.072*** 0.072** 0.073*** 0.049*
(3.59) (2.02) (3.10) (1.66)
LITIGATION + -0.011 -0.079 -0.001 -0.053 (-0.35) (-1.34) (-0.01) (-1.35)
CU_ROA - 0.114 -0.127 0.081 -0.089 (0.73) (-0.57) (0.36) (-0.52)
CU_LEV + 0.108 0.044 0.182** 0.067 (1.58) (0.61) (2.53) (0.96)
CU_ALTMAN + 0.001 -0.001 0.006 -0.003 (0.23) (-0.40) (0.91) (-0.99)
ln(DURATION) - 0.005 -0.012 -0.008 -0.005 (0.31) (-0.78) (-0.54) (-0.35)
CU_IMPT + 0.079 0.535** 0.088 0.150*** (1.51) (2.38) (1.37) (2.93)
Year fixed effects Included Included Included Included
Industry fixed effects Included Included Included Included
Number of
observations
567 568 543 592
Adjusted R2 0.638 0.620 0.638 0.639
This table reports the OLS regression results on the effect of supply chain lending relationship on accounting
conservatism of the borrowing suppliers, conditional on customer importance (CU_IMPT) and the duration of customer-
lender relationship (CU_LYRS). The sample includes 1,135 supplier-year observations from 1988 to 2010. See the
Appendix for variable definitions. Standard errors clustered by firm and adjusted for heteroskedasticity are in parentheses.
*, **, and *** denote significance levels at 10%, 5% and 1% respectively based on two-tailed t test statistics.
45
TABLE 10
Supply-Chain Lending and Accounting Conservatism: The Role of Trade Credit
Independent variables
Predict
ed
sign
Dependent variable = CSCORE
HIGH
TRADE_CREDIT
LOW
TRADE_CREDIT
Intercept ? 0.325*** 0.005 (4.82) (0.07)
CHAIN_LENDING - -0.048*** -0.016 (-2.85) (-0.73)
ln(MVE) - -0.003* -0.001*** (-1.75) (-3.30)
ln(FIRMAGE) + -0.016 -0.004 (-1.44) (-0.46)
INVCYCLE ? 0.046 -0.288 (0.18) (-1.20)
MB + -0.009** -0.013*** (-2.08) (-5.71)
RETVOL - 0.155 -0.149 (1.13) (-0.90)
SALEVOL + -0.006 -0.009 (-0.14) (-0.21)
BID_ASK + 4.214*** 6.575*** (5.33) (6.53)
LEV - 0.018 0.113*** (0.29) (2.60)
ALTMAN - -0.003 -0.001 (-1.41) (-0.89)
BIG_FOUR + 0.091*** 0.039 (3.72) (1.63)
LITIGATION + -0.004 -0.068 (-0.12) (-0.94)
CU_ROA - -0.168 0.141 (-1.01) (0.72)
CU_LEV + 0.017 0.246*** (0.20) (3.69)
CU_ALTMAN - -0.004 0.009 (-1.35) (1.48)
ln(DURATION) ? -0.012 -0.003 (-0.74) (-0.22)
CU_IMPT - 0.105* 0.094* (1.74) (1.79)
INTERCEPT Included Included Industry fixed effects Included Included Number of observations 567 568 Adjusted R
2 0.624 0.673
This table reports the OLS regression results on the effect of supply chain lending relationship on accounting
conservatism of the borrowing suppliers, conditional on the importance of trade credit in suppliers’ net worth
(TRADE_CREDIT). The sample includes 1,135 supplier-year observations from 1988 to 2010. See the Appendix for
variable definitions. Standard errors clustered by firm and adjusted for heteroskedasticity are in parentheses. *, **, and
*** denote significance levels at 10%, 5% and 1% respectively based on two-tailed t test statistics.
46
TABLE 11
Supply-Chain Lending and Accounting Conservatism: Alternative Measure of Conservatism
Independent variables Predicted sign Dependent variable = X
(1) (2)
Intercept ? -0.287** -0.352*** (-2.04) (-2.84)
D - 0.368 -0.746*** (1.09) (-3.05)
R
+ 0.014 0.228*** (0.45) (3.77)
D× R
+ 0.285** 0.255*** (2.12) (3.49)
CHAIN_LENDING
- -0.139** -0.009 (-2.04) (-0.21)
CHAIN_LENDING × D - 0.874 -0.028 (1.31) (-0.24)
CHAIN_LENDING × R
+ -0.037 -0.002 (-0.68) (-0.06)
CHAIN_LENDING× D × R
- -0.368* -0.711* (-1.77) (-1.89)
ln(MVE)
+ 0.106*** (5.13)
ln(MVE) × D
+ -0.097*** (-3.26)
ln(MVE) × R
+ -0.067*** (-4.83)
ln(MVE) × D× R
+ -0.091 (-0.76)
LEV
+ -0.720*** (-6.04)
LEV× D
+ -0.061 (-0.21)
LEV× R
+ 0.492*** (4.68)
LEV× D × R
+ -0.824 (-0.68)
MB
? 0.004 (1.09)
MB × D ? -0.003* (-1.92)
MB × R
+ -0.011*** (-2.67)
MB× D × R
? -0.120*** (-3.09)
LITIGATION
+ -0.333*** (-3.42)
LITIGATION × D + -0.126 (-0.77)
LITIGATION × R
+ 0.044 (0.64)
LITIGATION × D × R
+ -0.767 (-1.48)
Number of observations 1,135 1,135 Adjusted R
2 0.061 0.296
This table reports the OLS regression results on the effect of supply chain lending relationship on accounting
conservatism of the borrowing suppliers, using Basu (1997) piece-wise linear regression. The sample includes 1,135
supplier-year observations that initiated a new loan from 1988 to 2010. See the Appendix for variable definitions.
47
Standard errors clustered by firm and adjusted for heteroskedasticity are in parentheses. *, **, and *** denote
significance levels at 10%, 5% and 1% respectively based on two-tailed t test statistics.
48
TABLE 12
Supply-Chain Lending and Covenant Design
Panel A: Summary statistics for loan contract terms
Variables
CHAIN_LENDING = 1
(N=515)
CHAIN_LENDING = 0
(N=343) t test of mean
difference
z test of median
difference Mean Median Mean Median
FINCOV 0.216 0.250 0.324 0.333 -7.416 ***
-6.812 ***
SPREAD 154.60 150.00 237.26 200.00 -10.46 ***
-11.13 ***
LOAN_SIZE (million$)
484.67 250 101.14 30.00 6.36 ***
16.45 ***
MATURITY
3.89 4.09 3.77 3.64 2.63 ***
4.67 ***
SECURE_LOAN
0.47 0.00 0.68 1.00 -6.08 ***
-5.95 ***
RESOLVER
0.75 1.00 0.78 1.00 -0.94
-0.94
SYNDICATED_LOAN
0.97 1.00 0.70 1.00 10.46 ***
11.25 ***
This table reports summary statistics of loan contract terms. The sample includes 858 supplier-year observations
that initiated a new loan from 1988 to 2010, after excluding observations with missing loan contract terms. See the
Appendix for variable definitions. *, **, and *** denote significance levels at 10%, 5% and 1% respectively based on
two-tailed t (z) test statistics.
49
Panel B: Supply-chain lending and the relative use of financial covenants versus general covenants
Independent variables Dependent variable = FINCOV
Intercept 1.313***
(4.70)
CHAIN_LENDING -0.072***
(-3.74)
SPREAD -0.028**
(-2.19)
ln(LOAN_SIZE) -0.027***
(-3.31)
ln(MATURITY) 0.010
(0.71)
SECURE_LOAN -0.015
(-0.77)
RESOLVER 0.014
(0.71)
SYNDICATED_LOAN -0.139***
(-3.67)
ln(MVE) -0.001***
(-2.82)
ROA 0.302***
(4.35)
LEV 0.104*
(1.74)
MB -0.004* (-1.89)
ALTMAN 0.004
(1.33) Industry fixed effects Included
Year fixed effects Included
Number of observations 822
Adjusted R2 0.222
This table reports the OLS regression results on the effect of supply chain lending relationship on the use of
financial covenants in loan contracts. The sample includes 822 supplier-year observations that initiated a new
loan from 1988 to 2010, after excluding observations with missing loan contract terms. The dependent variable
is FINCOV. See the Appendix for variable definitions. Standard errors clustered by firm and adjusted for
heteroskedasticity are in parentheses. *, **, and *** denote significance levels at 10%, 5% and 1% respectively
based on two-tailed t test statistics.
50
TABLE 13
Supply-Chain Lending and Customers’ Accounting Conservatism
Independent variables Predicted sign Dependent variable = CU_CSCORE
Intercept ? 7.183*** 7.238***
(3.88) (3.91)
CU_CHAIN_LENDING ? 0.002 (0.20)
CU_ln(MVE) + 0.001*** 0.001***
(9.69) (9.61)
CU_ln(FIRMAGE) + -0.037*** -0.037***
(-4.28) (-4.31)
CU_INVCYCLE - -0.666*** -0.661*** (-2.75) (-2.71)
CU_MB ? 0.001 0.001 (0.88) (0.88)
CU_RETVOL + 0.765*** 0.765***
(4.72) (4.72)
CU_SALEVOL - -0.103*** -0.104*** (-2.77) (-2.77)
CU_BID_ASK - 13.085*** 13.084***
(14.58) (14.57)
CU_LEV + 0.201*** 0.201***
(4.94) (4.94)
CU_ALTMAN + -0.002 -0.002 (-0.93) (-0.91)
CU_BIG_FOUR - -0.019 -0.019
(-0.27) (-0.27)
CU_LITIGATION - -0.068*** -0.068*** (-6.48) (-6.47)
ROA - 0.019 0.018 (0.65) (0.61)
LEV + 0.019 0.018 (0.80) (0.76)
ALTMAN + 0.001 0.001 (1.43) (1.43)
ln(DURATION) - -0.024*** -0.024*** (-3.68) (-3.67)
SUP_IMPT + 0.023 0.020 (0.19) (0.16)
Year fixed effects Included Included
Industry fixed effects Included Included
Number of observations 2,835 2,835
Adjusted R2 0.339 0.338
This table reports the OLS regression results on the effect of supply chain lending relationship on accounting
conservatism for the borrowing major customers. The sample includes 2,836 customer-year observations that
initiated a loan from 1988 to 2010. The dependent variable is CSCORE for the major customers. The regression
variables are measured for the customers with the prefix “CU”, and the variables ROA, LEV, and ALTMAN are
measured for the customer’s supplier firm. See the Appendix for variable definitions. Year and industry fixed
effects are included in the regressions, but their coefficients are not reported for brevity. Standard errors clustered
by firm and adjusted for heteroskedasticity are reported in parentheses. *, **, and *** denote significance levels at
10%, 5% and 1% respectively based on two-tailed t test statistics.