Accounting Conservatism And

Relationship-specificity and accounting conservatism:
the effect of customer and supplier [1]

Charles J.P. Chen

CityUniversity of Hong Kong

Zengquan Li

Shanghai University of Finance&Economics

Xijia Su

CityUniversity of Hong Kong

YiweiYao[2]

CityUniversity of Hong Kong

February, 2008

Abstract:

Previousliterature providesevidence that conservatism could be driven by debt/compensation contract, litigation, regulation, and taxation reasons. This paper extends the above boundary by including customers and suppliers to explain the role of conservatism. Specifically, we test whether highercustomer/supplier specificity is associated with firm’smore accounting conservatism. Since customer/supplier specificity can create ex-post opportunisticincentive to the firm we expect conservatism can service a commitment role to induce the customer/supplier to undertake assets specific investment under this situation. Using asymmetric timeliness andcustomer/supplier R&D intensity as our major proxy for conservatism and specificity, we find, on both industry and firm level, firm’s accounting conservatism is positively associated with customers/suppliersspecificity. And this relation is weaker when an alternative mechanism, vertical integration,existsbetween the contracting parties or customer/supplier market concentration is low.

Key Words:

Specificity, Customer/Supplier,Accounting Conservatism

1. Introduction

Assets specificity is one of the three dimensions for characterizing transactions, also is held to be particular important (Williamson, 1985), while transaction costs are central to the study of economics (Williamson, 1979).But how corporate governance structure will be shaped by this contracting characteristic is quite unexplored, especially in terms of informal arrangements. Under this important topic, this paper tests whether accounting hasa role under this situations to mitigate the high transaction costs associated with assets specificityand how this role will be reflected in one of the most importantattributes of accounting, conservatism. Particularlywe examine whether firms use more conservative accounting as a commitment mechanism to induce suppliers/customers to undertake assets specificinvestments.

Williamson (1985) defines assets specific investment as ‘durable investment that are undertaken in support of particular transactions, the opportunity cost of which investment is much lower in best alternative uses or by alternative uses should the original transaction be prematurely terminated’ . This value differenceof the assets specific investment within and out of the specific relationis called quasi-rent (Alchian, 1984). We argue,for two reasons, the specific quasi-rent, which is created by customer/supplier’s assets specific investment, could be expropriated by the firm. The first reason is related to the information asymmetry between the firm and its customer/supplier. The second reason is due to the firm’s asymmetric loss function between customer/supplier’s overinvestment and underinvestment. The customer/supplier’s ex-ante anticipation of this ex-post expropriation will lead to overall contract efficiency loss (Kiein et al., 1978). And this deadweight loss will bebared by both parties. Hence, in equilibrium, both parties have incentive to employ some governance mechanisms to reduce the net loss (Alchian, 1984).This governance mechanism could be a variety of arrangements (Alchian, 1984,Williamson, 1979, 1985).In this paper, we predict, as one of important governance mechanisms, accounting should have a role in this context to mitigate agency cost and then enhance contact efficiency. Specifically, conservatism, as one of the most important attributes of accounting, through the different recognition requirement for gains versus losses, can mitigate the agency costs that resulting from information asymmetrybetween insiders and outsiders and asymmetric loss function of insiders (Lafond and Watts, 2008). So we raise our first hypothesisthat more customer/supplier specificity is associated with more conservative accounting of the firm.

Since previousliterature document that vertical integration is one of the important mechanisms to deal with mal-incentives created by assets specific investment (Williamson, 1975, 1979, 1985; Klein, Crawford and Alchian, 1978), we propose the positive association between customer/supplier specificity and firm’s conservatism could become weaker in the present of vertical integration due to the agency cost was largely reduced by fully integration. We further argue the effect of customer/supplier specificity can be more pronounced if the customer/supplier industry is highlyconcentrated. As documented in literature, industry concentration is a proxy for bargaining power. Therefore, it is easier or more effective for the customer/supplier, whose industry concentration is hightorequirea more favorable contracting system or asa protection to compensatetheir risk associated with assert specific investment.

We utilize three customer/supplier specificity measures: (ⅰ) Levy (1985) argues that research-intensive industries tend to involvespecialized inputs that require transaction-specific investment by suppliers. Allen and Phillips (2000) suggest that R&D-intensive industries are more likely to create relationship-specific investment. And R&D intensity also was used pervasivelyin the empirical literature on transaction cost economics to assess assets specificity (Boerner and Macher, 2001). Hence, we use the intensity of R&D expenditure of customer/supplier as our major proxy for customer/supplier specificity. (ⅱ) Following Nunn (2007),we construct an alternative supplier specificity measurebased on the type of market(whether or not the input of the firm is sold on an organized exchange and whether or not it is reference priced in a trade publication) in which the supplier’s products are traded in. (ⅲ) Following Kale and Shahrur（2007）, we use the dummy variable meaning whether a Strategic Alliance or Joint Venture(SA,JV) was formed between contact parties as the third proxy for customer/supplier specificity. Our tests rely on two measures to assess conservatism. First, trying toavoid the measurement error resulting from rents, we use firm level and three-year window estimated asymmetric timeliness as our major conservatism measure(Khan and Watts 2007). Following Ahmed and Duellman (2007), we also use an accrual-based conservatism measure, which is income before extraordinary items less cash flows from operations plus depreciation expenses deflated by average total assets, and average over a 3-year period concerted on year t, multiplied by negative one.

Our tests control for growth opportunity, measured with the sum of R&D and advertising expenses; for sales growth, Cash flow from operation as a proxy for profitability, uncertainty and investment cycle(measuredby stock return volatility and depreciation over lagged assets respectively), size, and for other conservatism drivers: equity contract, debt contract, litigation, regulation and taxation.

We find, on both industry and firm level, strong evidence that (ⅰ) the two conservatism measures are positively associated with customer/supplier specificity; (ⅱ) the above effect is weaker when vertical integration present or customer/supplier marketconcentration is low. These findings are robust to a number of sensitivity tests.

Previously, economics, marketing and finance literatures pervasively document the effect of customer and supplier on firm’s operation strategy, governance and financial structure. However, to date, little attempt have been made to address this issue in accounting with exceptions of Bowen et al. (1995) and a recent work by Raman and Husayan (2007). The first paperargues and finds that ongoing implicit claims between a firm and its customers, suppliers, employees, and short-term creditors create incentives for management to choose long-run income-increasing accounting policy, such as FIFO and straight-line depreciation method. The second one documents a more earnings management evidence when supplier/customer undertakes more relationship-specific investments, where earnings management is measured with discretionary accruals from modified Jones model and the frequency of large earnings increase.

However, for the following reasons these two studies fail to provide a complete and convincing story about how the ongoing relations between firm and its customer/supplier can influence the firm’s accounting choice. The first one is related to the underlying theory. As prior literature documents, since investors, if not in one game at least eventually over time, can see through or undo this earnings management or aggressive accounting policy especially when they are sophisticated or have a long term relationship with the firm, just like customer/suppliers. It is hard to understand why this short-term opportunistic behavior is assumed in equilibrium over long time period. While we expect in this ongoing relationship, to some extent, firm and its customer/supplier have common expectation that any opportunistic behavior will be realized eventuallyby the other party and then lead to some adverse effect, which will causea loss of overall contract efficiency. For example, this loss could be the real resourcespent to undo or adjust this manipulation or the inflated portion of earnings or book value by aggressive accounting policy, underinvestment or inefficient shorter duration. For the last one, Raman and Husayan (2007) do find empirical evidencethat the duration of the relation between the firm and its customer/supplier is negatively associated with earnings management under assets specificity investment context. Considering the long-term ongoing characteristic of the relation between firm and its customer/supplier, we propose a conservatism accounting rather that a liberal or opportunistic when assets specific investment is involved.

The second reason lies in the way they measure overall accounting method choice and earnings management and the way they choose sample. Bowen et al. (1995) use inventory recording and depreciation method to capture the overall aggressiveness of firms’ financial reporting. But as we know management use a portfolio rather that just one or two method to produce earnings and book value numbers. Given that their empirical evidence fully relying on this incomplete measure of overall accounting policy, it is difficult to generalize their conclusion to a complete story. Also, Bowen et al. (1995) arbitrarily select 1981, 1984, 1987, 1990 and 1993 as their sample years. With most of the relationship between customer/supplier lasting less than ten years (for our all available relation observations from 1980 to 2004, about 17% is longer than 10 years), they way they select their sample tends to create a selection bias towards shorter relations. Together with the reason we mentioned before, it is hard to argue their finding of income-increasing accounting is the true story in the on-going context. Regarding the earnings management paper, other than a lot of criticisms with their measure of earnings management, modified Jones model (Dechow et al., 1995), another problem is the possibility of alternative explanation. Usually customer/supplier experiences the same life cycle or ups and downs in growth as firm do. And accrual, both discretional and total, is positively related to growth opportunities. Therefore we can observe more discretionary accruals and positive earnings jump in the firm and higher R&D intensities,which exactly is their measure of relationship-specific investments of customer/supplier, in their customers/suppliers. If this is the case, it is not clear whether it is a synchronous growth story or earnings management story.

We try to mitigate the above limitations in the following ways. First, different from Bowen et al. (1995) and Raman and Husayan (2007), we favor anefficient contacting perspective on accounting choice instead of an opportunistic behavior one regarding the ongoing and endogenous nature of the claims between the firm and its customer and supplier. Consistently, we predict and find that the potentialconflict,which is caused by from assets specific investment, between the firm and its customer/supplierwill lead to a conservative accounting choice rather than an income-increasing one or earnings manager behavior. Second, we utilize multiple measures of both dependent and independent variables, namely, conservatism and customer/supplier specificity to mitigate measurement errors. Lastly, we use long time period sample window to avoid the potential sample selection bias in Bowen et al. (1995).

Conservatism is one of the most prominent characteristics of financial reportingwith long and significant influence on accounting practice (Basu, 1997). Based on previous researches, Watts (2003a) summarize four alternative explanations for accounting conservatism, contacting, litigation, taxation and regulation. So far, the extant literature provides ample evidence of the four conservatism explanations, contracting, litigation, taxation and regulation(Ball etal., 2000; Ball and Shivakumar 2005; Ahmed et al., 2002; Ahmed and Duellman, 2007; Ball et al., 2003; Bushman and Piotroski, 2006; Lafond and Watts, 2008; Qiang, 2007.).However,they are limited in the sense that these studies are constrained within the boundary consist of shareholders, creditors, managers, regulators and government. To understand accounting, the firm should be seen as a set of contracts among rational agents. Agents include managers, shareholders, creditors, employees, customers and suppliers, auditors and government. Contracts can be explicit or implicit, short-term or long-term (Sunder, 1987).Accounting choice therefore should be shaped by these forces all together. Without considering thoseimplicit contracts, the picture of roles of accounting we have so far is incomplete and could be biased. In this sense,by including customer/supplier into consideration, this study could contribute to conservatism literature by enriching our understanding on the reasons for conservatism.

The rest of the paper is organized as follows: we developed the main hypothesis in section 2, section 3presents theresearch design; in section 4 we describe our data and the sample; section 5reports empirical results; and section 6 gives major conclusions.

2. Hypothesis development

2.1 customer/supplier specificity and firm’s accounting conservatism

As one of important governance mechanisms, we expect accounting should also have a role in these assets specific investment situations to mitigate transaction cost andenhance contract efficiency. We examine this question in the context of the implicit claimsbetween the firm and its customers/suppliers with the presence of relationship-specific (RS) investment. Titman (1984) suggests that a firm with unique product may try to induce its customers/suppliers to undertake some RS investment. Except when these investments are transferable to alternative customers at low cost, which is rare, the set-up cost will lose value when the firm goes into bankruptcy or experience losses.

For two reasons, the specific quasi-rent couldbe expropriated by the firm under this situation. The first reason is related to the information asymmetry between the firm and its customer and supplier.Generally, insiders always hold private informationwhich is impossible can be accessed by outsiders or possible but with huge cost. And under this investment setting, the outsider’s (the customer/supplier who has undertaken assets specificinvestment) primary concern is the firm’s future prospect since the assets specific investment will lose value if the firm goes out business or experience sales decrease. Also customer/supplier could suffer from the firm’s reneging on existing contracts, which is more likely when firm’s future prospect worsen. But future prospect mainly relate to future growth options which areunverifiable (Smith and Watts, 1992). Therefore, we expect asevere information asymmetrybetween the firm and its customer/supplier under this assets specific investmentcontext. The second reason is that the firm’s loss function is asymmetry between customer/supplier’s overinvestment and underinvestment. We can image, in the trading relationship, the overinvestment of supplier / custom can not lead to losses to the firm, even worse, will improve the firms barging power due to the opposite excessive capacity which is non-marketable. However, the firm will suffer if its supplier/custom under-invest due to the under-supply of input or insufficient need for output. Therefore, the firm is provided with both incentive and ability to bias its book value and earnings upward.

However, with fully anticipation of this ex-post opportunistic behavior, customer/supplier will protect themselvesex-ante by systematically but inefficiently under-investing or even not investing at all or requiring better contracting terms to compensate the high risk they bear. Other costs created by this post opportunisticbehavior could be kinds of real resources devoted to the attempt to improve post-transaction bargaining position (Kiein et al., 1978).These costs are not just wealth transferring from one party to the other, but are dead loss will be beard by both parties. So in equilibrium, both parties will have incentive to employ some governance mechanism to reduce the deadweight loss (Alchian, 1984). There could be many arrangements that can be used for this purpose, such as a vertical integration and some enforceablelong-term contract (Klein et al, 1978).

In this paper we would like to look at an unexplored area, accounting. Lafond and Watts (2008) point out for two reasons conservatism can reduce agency problem caused by information asymmetry and symmetric reporting incentives. First, conservative accounting can provide the best ‘hard’ information on current performance for uninformed outsides. The higher verification standard for gains can eliminated the overstated unverifiable gains. On the other hand, more loss information will be provided under lower verification recognition standard. Therefore, on average, conservatism can well offset insiders’ asymmetric reporting bias as a persistent governance mechanism. The second reason is that this hard information provides a benchmark that makes it possible for alternative‘soft’ information sources to generate more credible information.