|
Article Excerpt
I. INTRODUCTION
Firms often allow their divisions to negotiate transfer prices (Eccles 1985; Emmanuel and Mehafdi 1994). When divisions have the option to transact in outside markets if they cannot reach agreement internally, conventional economic reasoning holds that a selling division would hesitate to accept less and a buying division would hesitate to pay more than the external market price. However, if the outside market price generates widely different profits across divisions, then interdivisional concerns for fairness could lead the two divisions to agree on a price that distributes profits more evenly. Such fairness pressures could constitute an important source of the demand for accounting disclosures that reveal relative profit information (Luft 1997).
Luft and Libby (1997) use a case-based questionnaire to investigate preferences for fairness in negotiated transfer pricing. They find that executive M.B.A.s expect transfer prices to be significantly lower than the outside market price when the outside price favors sellers, but not when the outside price results in equal profits for the buying and selling divisions. However, Luft and Libby's (1997) participants did not actually negotiate transfer prices, nor did they face monetary consequences. The applicability of Luft and Libby's (1997) findings to practice hinges on whether fairness pressures influence not only the transfer prices people expect to negotiate, but also the transfer prices they actually negotiate. Our study addresses this question.
Comparing fairness effects in expected and actual transfer price negotiations requires us to consider how parties negotiate transfer prices. Borrowing from Bazerman et al. (2000), we contrast two ends of a social presence continuum. At one end, we use a computerized negotiation mechanism in which the only communications are bids, asks, and acceptances. This setting typifies the experimental economics literature (Friedman and Sunder 1994), and reflects the increasing use of depersonalized, technology-driven negotiation mechanisms in practice (Moore et al. 1999). Notwithstanding this trend, face-to-face interaction will always play an important role in organizations, limiting the practical insights from a strictly computerized setting restricted to bids, asks, and acceptances. Accordingly, to capture the other end of the social presence continuum, we also investigate a setting in which the parties negotiate face-to-face, with no restrictions on communication.
These two settings yield different conclusions regarding the linkage from initial expectations to actual transfer prices. In the restricted, computerized setting, participants indicate that they expect fairness-based price concessions similar to those documented by Luft and Libby (1997) when the outside market price favors sellers, but these expectations do not materialize in actual negotiations. By contrast, in the unrestricted, face-to-face setting, we find a pronounced fairness effect in both initial expectations and subsequent negotiations. Prior transfer pricing studies have employed computerized negotiations (e.g., Ghosh 1994, 2000), written negotiations (e.g., DeJong et al. 1989; Greenberg et al. 1994), and face-to-face negotiations (e.g., Chalos and Haka 1990; Ravenscroft et al. 1993), with little consideration of the possible behavioral influences of these different negotiation mechanisms. Our results suggest that at least in some settings, different negotiation processes can yield different conclusions.
In addition to offering a better understanding of negotiated transfer pricing, these findings underscore the importance of merging psychology-based behavioral accounting research with the tools of experimental economics (e.g., Kachelmeier 1996; Haynes and Kachelmeier 1998; Moser 1998; Waller 2001; Libby et al. 2002; Sprinkle 2002). Specifically, we find that the extent to which questionnaire-based judgments (typifying the cognitive-behavioral tradition) generalize to actual decisions in compensated, competitive settings (typifying the experimental economics tradition) depends on environmental features, such as the means of negotiation, that can either suppress or reinforce phenomena such as preferences for fairness.
Section II develops the hypotheses. Section III describes our experiment. Section IV presents tests of the primary hypotheses, and Section V presents supplemental analyses to corroborate the primary findings and to examine the incremental effects of negotiation experience and relative profit information. Section VI concludes and discusses the future research directions suggested by the study's limitations.
II. THEORY AND HYPOTHESES
Baseline Hypotheses for Expected Transfer Prices
Because our goal is to compare Luft and Libby's (1997) findings for questionnaire-based transfer pricing expectations to the results from actual cash negotiations, we must first establish a common baseline. Luft and Libby (1997) develop a case that highlights the friction between self-interest and organizational equity in negotiated transfer pricing. A selling division ("Parts") can supply a product at a cost of $20 to a buying division ("Assembly") that values the product at $80. An outside market also exists, in which the price of the same product is either $70 (the treatment condition, favoring Parts) or $50 (a control condition, equalizing profits between the two divisions). If the organization grants divisions the autonomy to go to the outside market as an alternative to transacting internally, then economic self-interest dictates that the parties should not settle for a transfer price less favorable than that available from the outside market, after adjusting for any transaction costs from transacting externally.
From a strict market perspective, one might question the need for negotiation, because the firm could simply impose the outside market price as the internal transfer price. But the applicability of market logic to the social organization of a firm is limited: "Market exchange tends predominantly to encourage calculative relations of a transaction-specific sort between the parties.... Internal organization, by contrast, is often better able to make allowance for quasimoral involvements among the parties" (Williamson 1975, 38). Negotiated transfer pricing captures these trade-offs, providing a mechanism whereby the firm's interdependent subunits can balance market factors against the broader social factors that affect long-term organizational effectiveness (Watson and Baumler 1975; Chalos and Haka 1990).
One such social factor is interdivisional equity. In the Luft and Libby (1997) scenario, a transfer price of $70 would yield a much larger profit for Parts ($70 - $20 cost = $50 profit) than for Assembly ($80 - $70 = $10). Negotiators who have access to relative profit information might resist the perceived inequity of this split, consistent with anecdotal evidence that firms consider fairness repercussions in establishing transfer pricing policies. (1) For example, Eccles (1985, 8) summarizes evidence from interviewing 54 managers at 13 firms, concluding:
The fundamental difficulty in managing transfer pricing involves establishing practices that will lead to decisions that enhance corporate performance, while at the same time measuring, evaluating, and rewarding performance in light of these practices in a way that managers perceive as fair.
Luft (1997) reviews the theoretical framework of models that have begun to incorporate preferences for fairness in bargaining (e.g., Kahneman et al. 1986; Bolton 1991; Rabin 1993), tying this framework to the role of management accounting as the means by which fairness pressures can arise. Drawing on this framework, Luft and Libby (1997) find that if it is reasonable to assume divisions have approximately equal entitlements in all other aspects, then negotiators expect greater price concessions below the outside market price when the outside price favors parts ($70 treatment condition) than when it splits profits evenly between Parts and Assembly ($50 control condition). To establish a common baseline, we hypothesize the same prediction:
H1: The transfer price negotiators expect will fall shorter of the outside market price when the outside price would result in significantly higher profits for Parts than when the outside price would result in equal profits for Parts and Assembly.
Evidence also suggests that expectations in negotiation are egocentric, meaning that each party overweights the view most beneficial to itself (Bazerman et al. 2000). Egocentrism is closely related to motivated reasoning (Kunda 1990), whereby an individual's preferences lead to biased expectations and judgments. In transfer pricing, even if the instrument that elicits expected transfer prices gives participants no reason to bias their responses, motivated reasoning and egocentrism suggest that participants will be unable to detach themselves from their assigned roles in forming such expectations. We posit this effect in our second hypothesis, based on the similar hypothesis predicted and detected by Luft and Libby (1997).
H2: Parts negotiators (sellers) will expect higher transfer prices than will Assembly negotiators (buyers).
Together, H1 and H2 imply that the largest departure of price expectations below the outside market price will be for the price Assembly expects when the outside price favors Parts. That is, when the outside price is equitable ($50), egocentrism simply means that buyers (sellers) are likely to expect a transfer price lower (higher) than $50. However, when the outside price favors sellers ($70), buyers' preference to be treated fairly (H1) compounds their egocentrism (H2), such that buyers are likely to expect to negotiate a price well below market. For sellers, any fairness effect (H1) may offset their natural egocentrism (H2), such that sellers expect a price close to the market price from which they would benefit. Equivalently, sellers may rationalize a different view of fairness, reasoning that they deserve the market price due to assumed cost efficiencies. In general, perceptions that one deserves a bigger slice of the pie tend to work against interpreting fairness as "equality" (Thompson and Loewenstein 1992; Hoffman et al. 1994; Kagel et al. 1996). Thus, buyers' expected price concessions might not materialize in actual negotiations. This is the question we consider next.
From Expectations to Outcomes
The first two hypotheses replicate Luft and Libby's (1997) tests of the transfer prices people expect to negotiate--but what would happen in actual negotiations with cash payment of obtained profits? One view is typified by Franciosi et al. (1995, 938), who characterize fairness as "expectations that are not sustainable." In a laboratory market study, they show that although fairness pressures influence initial expectations, incentives for wealth maximization eventually overpower fairness expectations in favor of classical economic equilibrium price predictions. Extended to transfer pricing, the Franciosi et al. (1995) argument suggests that when the outside price favors sellers, the negotiating parties (especially buyers) may initially expect price concessions that reduce the perceived inequity. But sellers may not agree to a more equitable price in actual negotiation, especially if they...
|
