The effect of depreciation method choice on asset selling prices

Accounting, Organizations and Society 35 (2010) 757–774

Contents lists available at ScienceDirect

Accounting, Organizations and Society journal homepage: www.elsevier.com/locate/aos

The effect of depreciation method choice on asset selling prices q Scott B. Jackson a,⇑, Theodore C. Rodgers b, Brad Tuttle a a b

University of South Carolina, Moore School of Business, Columbia, SC 29208, USA Emory University, Goizueta School of Business, Atlanta, GA 30322, USA

a r t i c l e

i n f o

a b s t r a c t In this study, we examine whether an accounting choice that firms make for external financial reporting purposes influences the selling prices that managers seek to obtain when they dispose of used capital assets. From a normative perspective, managers should obtain the highest possible selling prices for used capital assets regardless of the firm’s depreciation method choice. However, theory predicts that depreciation method-induced differences in accounting book values will cause managers to systematically deviate from this normative prescription. Using multiple contexts, methodologies, and participant groups, we consistently find that managers sell used capital assets that have been depreciated using accelerated depreciation for lower prices than identical used capital assets that have been depreciated using straight-line depreciation. This effect even endures in the presence of fair value information about the asset being sold. We also provide theory-consistent evidence that depreciation method-induced differences in accounting book values influence managers’ asset selling price decisions because of the mental accounting process that they employ. Our study has economic efficiency implications for an array of situations in which depreciable assets are commonly sold. Ó 2010 Elsevier Ltd. All rights reserved.

Introduction This study examines whether an accounting choice that firms make for external financial reporting purposes influences the selling prices that managers seek to obtain when they dispose of used capital assets. Firms sell used capital assets in connection with divestitures, spinoffs, restructurings, reorganizations, downsizings, liquidations, fire sales,

q We are grateful to two anonymous reviewers, Arthur Allen, Wendy Bailey, Tim Doupnik, Louis Fayard, Bud Fennema, Vicky Glackin, Erin Hamilton, Noah Jackson, Kelvin Liu, Tom Lopez, Elaine Mauldin, Molly Mercer, Terence Pitre, Tammie Rech, Leigh Salzsieder, Bryan Stewart, Carlton Tartar, Scott Vandervelde, Rich White, Jennifer Winchel, Yi-Jing Wu, and workshop participants at Brock University, Florida State University, the University of Nebraska-Lincoln, the University of South Carolina, the 2008 American Accounting Association Annual Meeting, and the 2009 New England Behavioral Accounting Research Series for helpful comments. ⇑ Corresponding author. Address: School of Accounting, Moore School of Business, University of South Carolina, Columbia, SC 29208, United States. Tel.: +1 803 777 3100; fax: +1 803 777 0712. E-mail address: [email protected] (S.B. Jackson).

0361-3682/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.aos.2010.09.004

declining efficiency, plant and subsidiary disposals, productivity shifts, refocusing efforts, resource shortfalls, routine asset replacements, and earnings management.1 In peak expansion years during the period 1974–1992, approximately 7% of plant assets changed ownership through mergers, acquisitions, and outright asset sales (Maksimovic & Phillips, 2001). When selling used capital assets, managers 1 Studies that examine contexts in which used capital assets are sold include Ahn and Denis (2004), Atanassov and Kim (2009), Bartov (1993), Bates (2005), Colak and Whited (2007), Dittmar and Shivdasani (2003), Gaumnitz and Emery (1980), Herrmann, Inoue, and Thomas (2003), Hillier, McColgan, and Werema (2009), Hite, Owers, and Rogers (1987), Howe and McCabe (1983), John, Lang, and Netter (1992), John and Ofek (1995), Kang and Shivdasani (1997), Lang, Poulsen, and Stulz (1995), LoPucki and Doherty (2007), Maksimovic and Phillips (1998), Maksimovic and Phillips (2001), Mauer and Ott (1995), Pulvino (1999), Schlingemann, Stulz, and Walkling (2002), Warusawitharana (2008), and Yang (2008). These studies generally adopt an economics-based theoretical framework, which does not address whether and why psychological factors influence managerial decision making with respect to asset selling prices. Indeed, the aforementioned studies do not focus on the possible connection between accounting book values and managers’ asset selling price decisions. Our psychologybased theoretical framework suggests that such a connection may exist.

758

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

should attempt to obtain the highest possible selling prices. This objective is simple and intuitively appealing, but our results suggest that it may be difficult to follow. There are at least two factors that managers should ignore when making asset selling price decisions—historical costs and accounting depreciation (Blocher, Stout, Cokins, & Chen, 2008; Garrison & Noreen, 2003; Hilton, 2002; Horngren, Sundem, & Stratton, 2005; Titard, 1993). Historical costs are irrelevant because they relate to the past and cannot be changed regardless of any current decision. Likewise, accounting depreciation is irrelevant because it represents a method of cost allocation, not asset valuation (Kieso, Weygandt, & Warfield, 2007; Libby, Libby, & Short, 2007). Indeed, reductions in an asset’s book value need not (and usually do not) mirror the deterioration profile of the asset or changes in its fair value.2 As a result, depreciation method-induced differences in accounting book values are irrelevant when making asset selling price decisions, and these differences are even more clearly irrelevant when fair value information about the asset being sold is simultaneously available. This study draws upon research on mental accounting (Prelec & Loewenstein, 1998; Thaler, 1985, 1999) and mental depreciation (Gourville & Soman, 1998; Heath & Fennema, 1996) to predict that depreciation method-induced differences in an asset’s accounting book value will influence managers’ asset selling price decisions.3 Specifically, we contend that structural and relational similarities between mental depreciation and accounting depreciation will cause managers to apply mental accounting processes from the consumer domain of their lives to seemingly parallel situations involving depreciable business assets. If so, managers may be psychologically predisposed to sell used capital assets that have been depreciated using accelerated depreciation for lower prices than identical capital assets that have been depreciated using straight-line depreciation. Further, because mental accounting is spontaneous and may occur with minimal effort (Kahneman & Tversky, 1984), we expect that depreciation method choice will have a robust effect on managers’ asset selling price decisions even when managers simultaneously receive fair value information about the asset being sold. To test our prediction, we conduct five separate experiments using: (i) multiple decision contexts (sales through trade publication, trade-in, auction, and consignment), (ii) multiple methodologies (experiments with and without financial incentives), and (iii) multiple participant groups (practicing managers, experienced MBA students, and accounting students who have in-depth accounting knowledge). In each experiment, we manipulate the firm’s depreciation method choice, which causes the asset’s accounting

2 It is common to see significant gains and losses on asset disposals in firms’ financial statements. Such occurrences are evidence that the accounting book values of assets frequently diverge from their fair values. However, what cannot be determined from financial statement data is whether accounting book values influence managers’ asset selling price decisions. 3 Bowen, DuCharme, and Shores (1995) find that 68.7% of firms use the straight-line method only, 25.1% use a combination of the straight-line method and an accelerated method, and 4.9% use an accelerated method only. The remaining 1.3% of firms use an unclassified method or their method is missing.

book value to vary. At the same time, we hold the asset’s historical cost and physical attributes constant. We also manipulate the asset’s fair value in order to examine the possibility that participants use accounting book value as a proxy for fair value. If participants use book value in this manner then depreciation method-induced differences in accounting book values should have no effect on managers’ asset selling price decisions in the presence of fair value information. Our experiments reveal a robust depreciation method choice effect even when there are compelling reasons to expect that it will be absent. In each experiment, participants sell used capital assets that have been depreciated using accelerated depreciation for significantly lower prices than used capital assets that have been depreciated using straight-line depreciation. Even when participants simultaneously receive fair value information, our experiments reveal that depreciation method-induced differences in the asset’s accounting book value continue to exert a strong influence on participants’ asset selling price decisions. These findings arise even though participants hold equivalent beliefs about the asset’s physical condition. In addition, our experiments provide theory-consistent evidence that mental accounting processes at least partly underlie the depreciation method choice effect that we document. Finally, we provide an array of tests that help rule out alternative explanations for our experimental findings. This study makes two main contributions to the accounting literature. First, our study provides initial evidence, which is consistent across contexts, methodologies, and participant groups, that the price for which managers sell used capital assets is influenced by an accounting choice that firms make for external financial reporting purposes. A number of studies argue that decision errors and biases arise because individuals fixate on summary accounting numbers without paying adequate attention to the accounting methods that were used to generate those numbers (Arunachalam & Beck, 2002; Hand, 1990; Luft & Shields, 2001). In contrast to these studies, we provide evidence consistent with the view that depreciation method-induced differences in accounting book values influence managers’ asset selling price decisions not because managers fail to pay attention to accounting methods, but because of the particular meaning that managers ascribe to accounting information.4 Indeed, the close

4 In addition to this literature, researchers have also explored whether firms’ depreciation method choice has market-related consequences (Archibald, 1972; Beaver & Dukes, 1973; Kaplan & Roll, 1972), contracting consequences (Holthausen, 1981; Holthausen & Leftwich, 1983; Leftwich, 1981; Ricks, 1982; Watts & Zimmerman, 1986), and capital investment consequences (Jackson, 2008; Jackson, Liu, & Cecchini, 2009). Studies on market-related consequences suggest that investors correctly decipher the valuation implications of different depreciation methods (Ricks, 1982). Studies on contracting consequences indicate that a firm’s depreciation method choice may be value relevant even in the absence of market-related consequences because earnings influence how the cash flows of the firm are divided among contracting parties (Watts & Zimmerman, 1986). Studies on capital investment consequences suggest that accelerated depreciation is associated with larger capital investments than straight-line depreciation (Jackson, 2008; Jackson et al., 2009). However, none of these studies examine whether a firm’s depreciation method choice influences managers’ asset selling price decisions.

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

correspondence between mental depreciation and accounting depreciation appears to cause managers to attribute meaning to depreciation method-induced differences in accounting book values that are economically unwarranted. No prior study has documented or asserted that depreciation method-induced differences in accounting book values influence managers’ asset selling price decisions, and the results of our study cannot be inferred from prior research. Second, this study provides theory-consistent evidence that depreciation method-induced differences in accounting book values influence the way in which managers perceive and respond to sunk costs. Some studies manipulate the magnitude of sunk costs or progress towards completion (e.g., Conlon & Garland, 1993; Garland, 1990; Heath, 1995; Moon, 2001; Tan & Yates, 1995, 2002), while other studies hold sunk costs constant and manipulate project performance, personal responsibility, and other contextual variables (e.g., Brockner et al., 1986; Davis & Bobko, 1986; Simonson & Staw, 1992; Staw, 1976). Our study differs from sunk cost studies in that our stimulus is not sunk costs per se (we hold sunk costs constant in all of our experiments). Instead, our stimulus is depreciation method-induced differences in accounting book values (i.e., the portion of sunk cost that remains after deducting depreciation). Although accounting book values are irrelevant just as sunk costs are irrelevant, the differential transformation of sunk costs by straight-line depreciation versus accelerated depreciation appears to cognitively obscure the irrelevance of book values, causing managers to sell assets depreciated using accelerated depreciation for lower prices than assets depreciated using straight-line depreciation.5 Thus, our study extends sunk cost research by incorporating the modifying effect of accounting depreciation into our research design. This extension is important because it suggests that the sunk cost effect is modified by one of accounting’s most fundamental and enduring institutions—accounting depreciation. The remainder of this paper is organized as follows. The next section develops our hypothesis. The section after that describes five experiments and discusses their results. We then address a variety of alternative explanations for our experimental findings. The final section provides a summary and discusses certain limitations. Theory and hypothesis Research suggests that individuals maintain mental accounts that capture the costs and benefits of activities and events (Gourville & Soman, 1998; Hirst, Joyce, & Schadewald, 1994; Kahneman & Tversky, 1984; Prelec & Loewenstein, 1998; Thaler, 1985, 1999). According to Thaler (1999, 5 Of somewhat lesser importance, our study is also distinguished from sunk cost studies in that we focus on a decision made at the end of an asset’s useful life, after the firm has already committed itself to disposing of the asset. In many sunk cost studies, the focus is on decisions made at early or intermediate points in an asset’s life. The objective in many of these studies is to determine whether an initial investment decision causes individuals to escalate their commitment to a failing course of action. In contrast, the decision in our study (i.e., the selling price of a used capital asset) comes at the end of an asset’s useful life when escalation of commitment is irrelevant and cannot account for participants’ behavior.

759

p. 183), mental accounting is ‘‘the set of cognitive operations used by individuals and households to organize, evaluate, and keep track of financial activities.” Mental accounting processes appear to be spontaneous and may occur with minimal effort (Kahneman & Tversky, 1984; Thaler, 1999). An important determinant of how individuals mentally account for financial activities is the temporal order of costs and benefits. When costs and benefits occur in close time proximity (e.g., a dining experience in which the enjoyment of a meal occurs at about the same time as the payment), evaluation of a mental account involves the comparison of total costs to the associated hedonic benefits (Prelec & Loewenstein, 1998). However, when the cost of a financial activity occurs before consumers capture benefits from it (e.g., the purchase of a long-lived durable good), evidence suggests that consumers engage in a cognitive process known as ‘‘mental depreciation” (Heath & Fennema, 1996; Okada, 2001; Thaler, 1999). The process of mental depreciation involves allocating costs over consumption events in order to mentally align costs and benefits. Doing so enables consumers to avoid feeling as though they have experienced a loss at the time a durable good is purchased. The results of Heath and Fennema (1996) and Okada (2001) suggest that when consumers purchase a long-lived durable good, they create a mental account for the asset and then periodically allocate portions of cost to the durable good’s mental account as the asset is consumed. At the same time, consumers post the mental benefits of consumption to the durable good’s mental account, which temporally align the costs and benefits of acquiring the durable good. At any point during the durable good’s life, the portion of cost that has not been allocated to the durable good’s mental account comprises its ‘‘mental book value” (Okada, 2001). An asset’s mental book value is the consumer equivalent of a business asset’s accounting book value. When consumers dispose of a durable good before its mental book value has been depreciated to zero, they may experience cognitive pain because there are no consumption-related benefits to match with the remaining mental book value. However, the results of Okada (2001) and Prelec and Loewenstein (1998) suggest that consumers can avoid that pain by selling the durable good for an amount equal to or greater than its mental book value, thereby matching the asset’s cost with fully offsetting mental benefits. Thus, when a durable good has a high mental book value, consumers may desire to obtain a relatively high selling price to obtain sufficient mental benefits to fully offset the asset’s cost. On the other hand, when a durable good has a low mental book value, consumers may be willing to accept a relatively low selling price because the asset’s mental account has already accumulated sufficient mental benefits to fully offset the asset’s cost. This perspective is embodied in the sentiment frequently expressed by individuals that they have ‘‘gotten their money’s worth” out of an asset, which reflects the notion that consumers may be willing to accept a discounted price for a durable good that has a low mental book value. A number of researchers have observed that the physical recording process for accounting depreciation in

760

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

business settings is structurally and relationally similar to the cognitive recording process for mental depreciation in consumer settings (Heath & Fennema, 1996; Okada, 2001; Prelec & Loewenstein, 1998; Thaler, 1999). Specifically, accounting depreciation and mental depreciation are similar in that: (i) neither results in an immediate write-off of cost at the time of an asset’s acquisition, (ii) both involve periodic cost allocations, and (iii) both result in the write-off of book value at the time of an asset’s disposal. These structural and relational similarities may prompt managers to subconsciously apply the decision processes that they have developed for consumer durable goods to seemingly parallel situations involving business assets. This view is consistent with research on analogical reasoning (Gick & Holyoak, 1983; Novick, 1988; Sternberg, 1977) and with more general cognition-based theories in psychology (Anderson, 1998) indicating that when individuals encounter a new problem, they use problem solving techniques from previously encountered parallel problems. Despite the noteworthy similarities between mental book values and accounting book values, a fundamental difference between these concepts is that reductions in mental book values are triggered by physical usage while reductions in accounting book values are triggered by the mere passage of time. If managers do not fully internalize this difference, they may associate lower accounting book values with: (i) higher levels of past asset consumption and (ii) greater levels of asset physical degradation. For this reason, we expect that managers are predisposed to sell used capital assets that have been depreciated using accelerated depreciation for relatively low prices. Likewise, we expect that managers are predisposed to sell used capital assets that have been depreciated using straight-line depreciation for relatively high prices. Our primary research hypothesis, stated in alternative form, is as follows: Hypothesis: Ceteris paribus, managers sell used capital assets that have been depreciated using accelerated depreciation for lower prices than used capital assets that have been depreciated using straight-line depreciation. Experiments and results Experimental materials, dependent variables, and manipulations The experimental materials involve a single sales transaction for an asset which has an uncertain fair value. Used capital assets are often sold on a non-routine basis under conditions of less than perfect competition in decentralized and fragmented markets. Buyers and sellers in such markets face uncertainty about fair values because used capital assets are often unique and thinly traded. In these circumstances, definitive information about asset fair values is not known, so managers are likely to have considerable influence and discretion over selling prices. The decision contexts in this study require participants to assume the role of the general manager of a manufacturing company that is either: (i) selling one of the company’s machines through a trade publication on

a ‘‘best offer” basis (see Experiment 1), (ii) trading-in one of the company’s machines in connection with the acquisition of a new machine (see Experiment 2), (iii) selling one of the company’s machines in an auction (see Experiment 3), or (iv) selling one of the company’s machines on a consignment basis (see Experiments 4 and 5). In each experiment, we attempt to equalize participants’ beliefs about the machine’s physical condition by providing participants with identical information about the machine’s physical attributes. Specifically, we inform all participants that the machine: (i) is 4 years old, (ii) had an original expected life of 7 years, (iii) shows normal signs of wear and tear, (iv) had an original cost of $100000 and (v) had an estimated salvage value at the end of its 7-year life of $26000. There are two manipulated variables in each experiment (the dependent variables are discussed in connection with each individual experiment). The first manipulated variable is the company’s depreciation method (referred to as DEP in each of the experiments), which is manipulated between subjects at two levels. In the accelerated depreciation method condition, accumulated depreciation is $73968, resulting in an accounting book value of $26032. In the straight-line depreciation method condition, accumulated depreciation is $42284, resulting in an accounting book value of $57716. This manipulation is identical in all five experiments. The second manipulated variable is fair value information. We incorporate fair value information into our design to address whether participants use the asset’s accounting book value as a proxy for its fair value. If so, the effect of accounting book value on managers’ asset selling price decisions should disappear in the presence of fair value information. We manipulate fair value information in two ways. In Experiments 1, 2, and 3, we manipulate fair value between subjects as the presence or absence of fair value information (referred to as FV in the experiments). When fair value information is present, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $42000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” When fair value information is absent, participants are informed that the machine was recently inspected and ‘‘the inspector did not provide an estimate of fair market value.” In Experiments 4 and 5, we manipulate fair value information between subjects as low ($35000) or high ($49000) (referred to as FVLH in the experiments). When fair value is low, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $35000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” When fair value is high, participants are provided with the same information shown above except that fair value is $49000 rather than $35000. We explain why we have two fair value manipulations when we discuss Experiment 4.

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

Manipulation checks and salience checks The experimental materials for Experiments 1–5 contain three manipulation checks, each of which takes the form of a multiple-choice question. One manipulation check is for the company’s depreciation method choice, one manipulation check is for the machine’s book value, and one manipulation check is for the machine’s fair value. The percentage of correct responses for each of these manipulation checks is generally in excess of 95% in Experiments 1–5, but in no case do any fall below 90%. Our results include all participants, but excluding those participants who failed one or more of the manipulation checks does not alter any of our inferences or conclusions. The materials for Experiments 1–5 also contain four attention checks, each of which takes the form of a multiple-choice question (these are discussed in the section titled ‘‘Alternative explanations”). One attention check is for the machine’s physical condition, one attention check is for the machine’s original useful life, one attention check is for the period of time over which the machine has been used, and one attention check is for the earnings-related consequences of disposing of the machine. The percentage of correct responses for each of the four attention checks is generally in excess of 95% in Experiments 1–5, but in no case do any fall below 90%. Experiment 1 Purpose Our first experiment tests our research hypothesis using highly experienced business managers as participants. If we find no evidence of a depreciation method-induced book value effect with this group of participants, we believe that it makes little sense to proceed further. However, to the extent that we find evidence of a depreciation method-induced book value effect with experienced managers, we believe that it makes sense to consider additional decision contexts and issues. The context of this experiment corresponds to selling a used machine through a trade publication on a ‘‘best offer” basis. Participants Participants in Experiment 1 consist of practicing managers who voluntarily participated in an experiment administered through the mail. The names and addresses of 1750 managers were randomly selected from Standard and Poor’s NetAdvantage database of registered executives.6 To help match the task with participants who are likely to be familiar with selling used capital assets, we

6 Initially, we mailed out 1500 instruments, but one of the four conditions had only 12 respondents while each of the other three conditions had at least 23 respondents. We elected to mail out an additional 250 instruments in the condition that had only 12 respondents to help balance our sample sizes between cells (the managers in this subsequent mailing were also randomly selected from Standard and Poor’s NetAdvantage database of registered executives). Our conclusions are the same regardless of whether we use the participants from our initial mailing only or the participants from both the initial mailing and the subsequent mailing combined.

761

selected executives who were two or three positions from the top echelon of executives in their firms. Of the 1750 instruments mailed out, 172 instruments were returned as undeliverable. Of the 1578 instruments that presumably reached the intended recipients, 100 instruments were completed and returned, which results in a response rate of approximately 6% (100/1578).7,8 The mean age of participants is approximately 55 years and the mean amount of experience is approximately 34 years. Approximately 88% of the participants are male.9 In the post-experimental questionnaire, we ask participants how familiar they are with situations similar to the one in our study. Participants provide their response on a 10-point scale, with a response of ‘‘1” indicating that they are ‘‘unfamiliar” with this type of task and a response of ‘‘10” indicating that they are ‘‘familiar” with this type of task. The mean response to this question is 7.12, which suggests that our participants are reasonably well suited for the task that we ask them to perform. Results The dependent variables in this experiment are participants’ likelihoods of accepting seven different offer prices from prospective buyers of the used machine, which range from $20000 to $50000 in increments of $5000. Participants provide their likelihood ratings on scales ranging from 0 (very unlikely to accept the offer) to 100 (very likely to accept the offer). Our hypothesis predicts that participants are more likely to accept offer prices when the machine has been depreciated using accelerated depreciation than when the machine that has been depreciated using straight-line depreciation. Panel A of Table 1 provides descriptive statistics for the offer prices partitioned by the two manipulated variables (formal statistical tests are discussed next). When no fair value information is provided to participants, the mean likelihoods of accepting offer prices from prospective buyers of the machine in the accelerated depreciation condition are 33.70, 49.63, 67.04, 75.93, 84.07, 90.74, and

7 Initially, there were 101 participants, but one participant was eliminated because he/she indicated that he/she would be more likely to accept low offer prices than high offer prices for the asset. 8 We performed various analyses to assess the likelihood of response bias. The gender profile of the 1750 managers who were asked to participate in the experiment does not significantly differ from the gender profile of the 100 managers who actually participated in the experiment (p > 0.10). Likewise, based on a partition of the US into the western half and the eastern half, the geographic profile of the 1750 managers who were asked to participate in the experiment does not significantly differ from the geographic profile of the 100 managers who actually participated in the experiment (p > 0.10). We also partitioned the 100 responding managers into early versus late responders based on a median split of the postmark date. We find that: (i) participants’ responses to the dependent variables, (ii) participants’ age, gender, and experience profiles, (iii) the accuracy of participants’ responses to the manipulation checks, and (iv) the accuracy of participants’ responses to the attention checks do not differ based on whether the participants were early or late responders (p > 0.10 for all variables). Thus, we find no indication that: (i) managers who respond differ from managers who do not respond or (ii) managers who respond late differ from managers who respond early. 9 Demographic variables in this and subsequent experiments do not differ between the experimental conditions (p > 0.10), which suggests that participants were successfully randomized.

762

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

95.18. Likewise, the mean likelihoods of accepting offer prices from prospective buyers of the machine in the straight-line depreciation condition are 21.03, 30.38, 44.58, 54.34, 64.85, 77.65, and 91.11. Panel A of Fig. 1 plots the mean likelihoods for the seven offer prices. Visual inspection of Panel A provides initial evidence of a depreciation method-induced book value effect as indicated by the fact that mean likelihoods of accepting offer prices are always greater when the firm uses accelerated depreciation than when the firm uses straight-line depreciation. When fair value information is provided to participants, the mean likelihoods of accepting offer prices from prospective buyers of the machine in the accelerated depreciation condition are 36.25, 45.00, 59.58, 71.25, 85.42, 92.92, and 96.67. Likewise, the mean likelihoods of accepting offer prices from prospective buyers of the machine in the straight-line depreciation condition are 16.95, 23.91, 36.96, 50.00, 66.96, 83.04, and 93.47. Panel B of Fig. 1 plots the mean likelihoods for the seven offer prices. Visual inspection of Panel B provides initial evidence of a depreciation method-induced book value effect in the presence of fair value information as indicated by the fact that mean likelihoods of accepting offer prices are always greater when the firm uses accelerated depreciation than when the firm uses straight-line depreciation. To formally test our hypothesis, we estimate a 2  2 ANOVA for each of the offer prices.10 Recall that the manipulated variables and their abbreviations are defined in the section titled ‘‘Experimental materials, dependent variables, and manipulations”. As shown in Panel B of Table 1, DEP is highly significant for offer prices of $20000 (F-statistic = 9.02; p = 0.003), $25000 (F-statistic = 14.02; p < 0.001), $30000 (F-statistic = 17.38; p < 0.001), $35000 (F-statistic = 19.45; p < 0.001), $40000 (F-statistic = 19.07; p < 0.001), and $45000 (F-statistic = 13.24; p < 0.001), and insignificant for the offer price of $50000 (F-statistic = 2.46; p = 0.120). Also, FV and the interaction between DEP and FV are insignificant in each of the ANOVAs (p > 0.10). The significant main effect for DEP provides support for our hypothesis, and the insignificant interaction between DEP and FV suggests that fair value information does not eliminate the effect of depreciation method-induced differences in accounting book values.11 The lack of an interaction between DEP and FV is noteworthy. In the absence of fair value information, decision makers may use the machine’s accounting book value as a proxy for its fair value. If this is the only reason that the machine’s accounting book value influences participants’ selling price decisions, then the book value effect will disappear when fair value information is available. Contrary to this perspective, we find that the presence of fair value information does not eliminate the effect of accounting book values on participants’ likelihoods of accepting various offer prices. This finding suggests that

10 Instead of estimating a series of individual ANOVAs, we could take the average of participants’ seven individual responses and use the average in a single ANOVA. When we do this, our inferences and conclusions do not change. 11 In Experiments 4 and 5, we manipulate fair value as high or low and find a significant main effect.

participants are not merely using the machine’s accounting book value as a proxy for its fair value. Experiment 2 Purpose The first experiment involved selling a machine on a ‘‘best offer” basis, while the present experiment involves trading-in a machine in connection with the acquisition of a new machine. The results of Okada (2001) suggest that trade-ins may mitigate the effect of mental book values on consumer decisions because the undepreciated balance in an existing asset’s mental account can be transferred to the new asset’s mental account.12 Additionally, research in psychology suggests that seemingly inconsequential variations in tasks may cause individuals to make different decisions (Kahneman & Tversky, 1979; Kuhberger, 1998; Samuelson & Zeckhauser, 1988). Thus, it seems desirable to test our hypothesis in this alternative decision context. Participants Participants in Experiment 2 consist of 78 MBA students from a large public university. The mean age of the participants is approximately 30 years and the mean amount of experience is approximately 7 years.13 Approximately 71% of the participants are male. Of the 78 participants, 29 participants come from a part-time MBA program in which they had previously completed financial accounting and were a few weeks away from completing managerial accounting. These participants had previously received instruction on the meaning of accounting book value and the irrelevance of accounting book values for decision making. Participants were asked to voluntarily complete the experimental materials while in class and there was 100% participation. The remaining 49 participants came from a full-time MBA program in which students were about 4 weeks away from earning their MBA degree. To obtain their participation, we sent an email to 95 students seeking their voluntary participation in an experiment. The email contained a link to the experimental materials hosted on http:// www.SurveyMonkey.com. Participants were randomly assigned to one of the four experimental conditions. In the academic year preceding the administration of this experiment, participants received instruction on the meaning of accounting book value and the irrelevance of accounting book values for decision making.14 Results The dependent variable in this experiment is the tradein value that participants require for the used machine.

12 A similar mental accounting line of reasoning has been advanced in the context of transferring an asset from one investment account, which has declined in value, to a new investment account (Shefrin & Statmen, 1985). Transferring the asset rather than selling it outright allows the individual to avoid closing the mental account for the previous investment in the red. 13 In Experiments 2, 3, and 4, approximately 40% of participants’ total work experience is in a supervisory or managerial capacity. 14 We find no difference in mean responses to the dependent variable between the 29 part-time MBA students and the 49 full-time MBA students (p > 0.10).

763

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774 Table 1 Cell means and ANOVA results for Experiment 1. Offer amounts

No fair value information

Fair value information

Straight-line depreciation

Accelerated depreciation

Straight-line depreciation

Accelerated depreciation

(n = 26)

(n = 27)

(n = 23)

(n = 24)

16.95 15.50

36.25 29.01

23.91 15.88

45.00 30.50

36.96 16.90

59.58 27.58

50.00 21.53

71.25 22.52

66.96 21.83

85.42 17.93

83.04 14.60

92.92 13.67

93.47 8.85

96.67 7.61

Panel A: Mean likelihoods of accepting offer prices and standard deviations $20000 offer Mean 21.03 33.70 Std. dev. 25.28 32.24 $25000 offer Mean 30.38 49.63 Std. dev. 27.64 29.93 $30000 offer Mean 44.58 67.04 Std. dev. 27.97 31.96 $35000 offer Mean 54.34 75.93 Std. dev. 26.58 25.46 $40000 offer Mean 64.85 84.07 Std. dev. 25.91 19.47 $45000 offer Mean 77.65 90.74 Std. dev. 19.82 13.85 $50000 offer Mean 91.11 95.18 Std. dev. 16.06 11.22 Source of variation

DF

Panel B: Analysis of variance results Dependent variable is likelihood of accepting offer price of $20000 DEP 1 FV 1 DEP  FV 1 Error 96 2 R (%) = 8.77 Model F-statistic = 3.08 (p = 0.031) Dependent variable is likelihood of accepting offer price of $25000 DEP 1 FV 1 DEP  FV 1 Error 96 R2 (%) = 13.57 Model F-statistic = 5.02 (p = 0.003) Dependent variable is likelihood of accepting offer price of $30000 DEP 1 FV 1 DEP  FV 1 Error 96 R2 (%) = 16.78 Model F-statistic = 6.45 (p < 0.001) Dependent variable is likelihood of accepting offer price of $35000 DEP 1 FV 1 DEP  FV 1 Error 96 R2 (%) = 17.52 Model F-statistic = 6.80 (p < 0.001) Dependent variable is likelihood of accepting offer price of $40000 DEP 1 FV 1 DEP  FV 1 Error 96 R2 (%) = 16.78 Model F-statistic = 6.45 (p < 0.001) Dependent variable is likelihood of accepting offer price of $45000 DEP 1 FV 1 DEP  FV 1 Error 96 R2 (%) = 13.67 Model F-statistic = 5.07 (p = 0.003)

SS

F-statistic

p-Value

6357.93 14.68 273.48 67660.04

9.02 0.02 0.39

0.003 0.886 0.535

10125.95 767.14 21.12 69340.28

14.02 1.06 0.03

<0.001 0.305 0.865

12654.30 1414.49 0.17 69902.10

17.38 1.94 0.00

<0.001 0.166 0.987

11419.01 506.70 0.68 56372.23

19.45 0.86 0.00

<0.001 0.355 0.973

8841.88 74.22 3.67 44520.03

19.07 0.16 0.01

<0.001 0.690 0.929

3281.59 356.30 64.29 23785.86

13.24 1.44 0.26

<0.001 0.233 0.612

(continued on next page)

764

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

Table 1 (continued) Source of variation

DF

SS

Dependent variable is likelihood of accepting offer price of $50000 DEP 1 FV 1 DEP  FV 1 Error 96 R2 (%) = 3.25 Model F-statistic = 1.08 (p = 0.362)

F-statistic

327.94 91.99 4.84 12777.80

2.46 0.69 0.04

p-Value

0.120 0.408 0.849

The dependent variable is the likelihood of participants accepting offer prices for a used machine. Participants provide their responses on scales ranging from 0 (very unlikely) to 100 (very likely). The first manipulated variable is the company’s depreciation method choice, which is referred to as DEP. In the accelerated depreciation method condition, accumulated depreciation is $73968, resulting in an accounting book value of $26032. In the straight-line depreciation method condition, accumulated depreciation is $42284, resulting in an accounting book value of $57716. The second manipulated variable is the presence or absence of fair value information, which is referred to as FV. When fair value information is present, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $42000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” When fair value information is absent, participants are informed that the machine was recently inspected and ‘‘the inspector did not provide an estimate of fair market value.”

Panel A: No fair value information Mean likelihoods of accepting offer prices

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% $20000

$25000

$30000

$35000

$40000

$45000

$50000

Offer prices Straight-line depreciation Accelerated depreciation

Panel B: Fair value information Mean likelihoods of accepting offer prices

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% $20000

$25000

$30000

$35000

$40000

$45000

$50000

Offer prices Straight-line depreciation Accelerated depreciation Fig. 1. Mean likelihoods of accepting offer prices (Experiment 1).

765

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774 Table 2 Cell means and ANOVA results for Experiment 2. Levels of DEP

Levels of FV

Rows

No fair value information

Fair value information

Panel A: Mean trade-in values and standard deviations Straight-line depreciation Means $41604 Std. dev. $20206 n 18

$38497 $12307 21

$39931 $16274 39

Accelerated depreciation Means Std. dev. n

$24589 $12627 17

$25740 $10259 22

$25239 $11209 39

Columns Means Std. dev. n

$33339 $18805 35

$31970 $12899 43

$32585 $15728 78

Source of variation Panel B: Analysis of variance results DEP FV DEP  FV Error R2 (%) = 22.65 Model F-statistic = 7.23 (p < 0.001)

DF

SS

F-statistic

p-Value

1 1 1 74

4272.23 18.44 87.39 14732.11

21.14 0.09 0.44

<0.001 0.762 0.510

The dependent variable is the trade-in value that participants require for a used machine. The first manipulated variable is the company’s depreciation method choice, which is referred to as DEP. In the accelerated depreciation method condition, accumulated depreciation is $73968, resulting in an accounting book value of $26032. In the straight-line depreciation method condition, accumulated depreciation is $42284, resulting in an accounting book value of $57716. The second manipulated variable is the presence or absence of fair value information, which is referred to as FV. When fair value information is present, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $42000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” When fair value information is absent, participants are informed that the machine was recently inspected and ‘‘the inspector did not provide an estimate of fair market value.” For purposes of tabulating the sum of squares in Panel B, we divide the dependent variable by $1000.

Our hypothesis predicts that participants require a lower trade-in value when the machine has been depreciated using accelerated depreciation than when the machine has been depreciated using straight-line depreciation. Panel A of Table 2 provides cell means for the two levels of DEP partitioned by the two levels of FV, and Fig. 2 graphs the mean trade-in values. When no fair value information is provided to participants, their mean trade-in value in the straight-line depreciation condition is $41604 and their mean trade-in value in the accelerated depreciation condition is $24589. When fair value information is provided to participants, their mean trade-in value in the straight-line depreciation condition is $38497 and their mean trade-in value in the accelerated depreciation condition is $25740. Visual inspection of Fig. 2 provides initial evidence of a depreciation method-induced book value effect as indicated by the fact that mean trade-in values are larger when the firm uses straight-line depreciation than when the firm uses accelerated depreciation. To formally test our hypothesis, we estimate a 2  2 ANOVA, which is reported in Panel B of Table 2. The ANOVA reveals that there is: (i) a significant main effect of DEP (F-statistic = 21.14, p < 0.001), (ii) no significant main effect of FV (F-statistic = 0.09, p = 0.762), and (iii) no significant interactive effect of DEP and FV (F-statistic = 0.44, p = 0.510). The significant main effect of DEP provides sup-

port for our hypothesis. Again, the absence of an interaction between DEP and FV is noteworthy because it indicates that the presence of fair value information does not eliminate the effect of accounting book value on participants’ required trade-in values. In turn, this finding suggests that participants are not merely using the machine’s accounting book value as a proxy for its fair value. Experiment 3 Purpose The first two experiments do not include a financial incentive. Tuttle and Burton (1999) find that even a modest financial incentive can greatly increase information usage. This finding is noteworthy because a financial incentive could motivate participants to: (i) focus on fair value information and (ii) ignore accounting book value when making asset selling price decisions. Given that financial incentives are often a feature of actual decision contexts facing managers, it is desirable to test our hypothesis in a decision context involving real financial incentives. We rely on induced-value theory and offer participants compensation that is directly linked to their submitted selling price for the asset (Smith, 1976). We use a single round, sealed-offer auction.

766

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

$44000 $42000 $40000

Mean trade-in values

$38000 $36000 $34000 $32000 $30000 $28000 $26000 $24000 $22000 $20000

No fair value information

Fair value information

Straight-line depreciation

Accelerated depreciation

Fig. 2. Mean trade-in values (Experiment 2).

Participants Participants in Experiment 3 consist of 60 MBA students from a large public university. The mean age of the participants is approximately 31 years and the mean amount of experience is approximately 9 years. Approximately 72% of the participants are male. The experiment was administered near the end of students’ managerial accounting course, which is the second of two required accounting courses in the MBA curriculum. These participants had previously received instruction on the meaning of accounting book value and the irrelevance of accounting book values for decision making. Participants were asked to voluntarily complete the experimental materials while in class and there was approximately 95% participation.15 Results The dependent variable in this experiment is the auction price that participants require to sell the used machine. Our hypothesis predicts that participants will require a lower auction price when the machine has been depreciated using accelerated depreciation than when the machine has been depreciated using straight-line depreciation. The auction compensates participants $1 for every $5000 of selling price that they obtain, but only those auction prices below the median auction price are accepted and compensated. This situation mimics a market for used equipment in which there are twice as many sellers as buyers, and it creates an incentive for participants to submit low auction prices to avoid not selling their machine.16 The average payout for winning bids was approximately $4.25 for 10 minutes of time. Panel A of Table 3 provides cell means for the two levels of DEP partitioned by the two levels of FV, and Fig. 3 15 Initially, there were 61 participants, but one participant was eliminated from our analyses because he/she indicated an auction price of $250000. 16 Because communication between participants is not allowed, the most likely achievable equilibrium auction price is $5000, in which case participants would be paid $1.

graphs the mean auction prices. When no fair value information is provided to participants, their mean auction price in the straight-line depreciation condition is $37739 and their mean auction price in the accelerated depreciation condition is $19201. When fair value information is provided to participants, their mean auction price in the straight-line depreciation condition is $35407 and their mean auction price in the accelerated depreciation condition is $26300. Visual inspection of Fig. 3 provides initial evidence of a depreciation method-induced book value effect as indicated by the fact that mean auction prices are larger when the firm uses straight-line depreciation than when the firm uses accelerated depreciation. To formally test our hypothesis, we estimate a 2  2 ANOVA, which is reported in Panel B of Table 3. The ANOVA reveals that there is: (i) a significant main effect of DEP (F-statistic = 27.95, p < 0.001), (ii) no significant main effect of FV (F-statistic = 0.83, p = 0.366), and (iii) a marginally significant interactive effect of DEP and FV (F-statistic = 3.25, p = 0.077). The significant main effect of DEP provides support for our hypothesis. However, the presence of a marginally significant interaction between DEP and FV suggests that we should examine the effect of DEP on auction prices at each level of FV. When no fair value information is provided to participants, we find that the mean auction price of $19201 in the accelerated depreciation condition is significantly less than the mean auction price of $37739 in the straight-line depreciation condition (t-statistic = 4.72, p < 0.001). Similarly, when fair value information is provided to participants, we find that the mean auction price of $26300 in the accelerated depreciation condition is significantly less than the mean auction price of $35407 in the straight-line depreciation condition (t-statistic = 2.63, p = 0.014). Thus, our experimental results indicate that depreciation method-induced differences in the machine’s accounting book have an effect on auction prices even in the presence of fair value information and economic incentives to ignore

767

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774 Table 3 Cell means and ANOVA results for Experiment 3. Levels of DEP

Levels of FV

Rows

No fair value information

Fair value information

Panel A: Mean auction prices and standard deviations Straight-line depreciation Means $37739 Std. dev. $12071 n 16

$35407 $11168 14

$36651 $11519 30

Accelerated depreciation Means Std. dev. n

$19201 $9251 15

$26300 $7191 15

$22751 $8906 30

Columns Means Std. dev. n

$28769 $14195 31

$30697 $10257 29

$29701 $12383 60

Source of variation Panel B: Analysis of variance results DEP FV DEP  FV Error R2 (%) = 36.67 Model F-statistic = 10.81 (p < 0.001)

DF

SS

F-statistic

p-value

1 1 1 56

2859.62 85.01 332.81 5729.38

27.95 0.83 3.25

<0.001 0.366 0.077

The dependent variable is the auction price that participants require for a used machine. The first manipulated variable is the company’s depreciation method choice, which is referred to as DEP. In the accelerated depreciation method condition, accumulated depreciation is $73968, resulting in an accounting book value of $26032. In the straight-line depreciation method condition, accumulated depreciation is $42284, resulting in an accounting book value of $57716. The second manipulated variable is the presence or absence of fair value information, which is referred to as FV. When fair value information is present, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $42,000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” When fair value information is absent, participants are informed that the machine was recently inspected and ‘‘the inspector did not provide an estimate of fair market value.” For purposes of tabulating the sum of squares in Panel B, we divide the dependent variable by $1000.

$39000 $37000 $35000

Mean auction prices

$33000 $31000 $29000 $27000 $25000 $23000 $21000 $19000 $17000 $15000

No fair value information Straight-line depreciation

Fair value information Accelerated depreciation

Fig. 3. Mean auction prices (Experiment 3).

accounting book value. However, the effect of book value varies somewhat depending on whether fair value information is presented. Nonetheless, our findings again suggest that participants are not merely using the machine’s accounting book value as a proxy for its fair value.

Experiment 4 Purpose Experiment 4 tests our research hypothesis in an additional context (i.e., sale of a used machine on a consignment

768

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

basis), but more importantly it captures a perception measure to explore why depreciation method-induced differences in accounting book values influence managers’ asset selling prices. If the structural and relational similarities between mental depreciation and accounting depreciation cause managers to view these two types of depreciation similarly, as we contend in ‘‘our hypothesis development”, the perceived benefits from the machine should vary in response to the firm’s depreciation method choice. Thus, we expect that managers will perceive that an asset depreciated using accelerated depreciation has provided the company with relatively large past benefits, while managers will perceive that an asset depreciated using straight-line depreciation has provided the company with relatively small past benefits. If, however, firms’ depreciation method choice does not influence participants’ perceptions of past benefits, the claimed theoretical reason for our findings may be thrown into question. In addition to capturing a perception measure, Experiment 4 also implements an alternative fair value manipulation. In the three previous experiments, we manipulated fair value as either present or absent. We decided to revise the fair value manipulation as either low ($35000) or high ($49000) to control for the possibility that participants impute a fair value when that information is absent. Participants Participants in Experiment 4 consist of 76 MBA students from a large public university. The mean age of the participants is approximately 26 years and the mean amount of experience is approximately 4 years. Approximately 65% of the participants are male. The experiment was administered near the end of students’ managerial accounting course, which is the second of two required accounting courses in the MBA curriculum. The participants had previously received instruction on the meaning of accounting book value and the irrelevance of accounting book values for decision making. Participants were asked to voluntarily complete the experimental materials while in class and there was 100% participation.17 Perceived benefits To elicit participants’ perceptions about the benefits that the used machine has provided to the company, we hold the selling price of the machine constant at $32000 and ask participants to respond to five items on scales ranging from 1 (definitely disagree) to 7 (definitely agree). We hold the selling price constant at $32000 to equalize this source of benefit across experimental conditions. By doing so, we more accurately measure the benefits that participants perceive that the machine has provided to the company through its use in operations. If the selling price was not held constant, then there would be two sources of potential benefits (i.e., the benefits from using the machine and the cash benefits from selling the 17 Initially, there were 80 participants, but four participants were eliminated because they provided a minimum selling price of $0. We speculate that these participants may have indicated such a selling price under the belief that the unneeded used machine should be promptly disposed.

machine). Our interest lies in isolating and measuring depreciation method-induced differences in perceived benefits from using the machine. The five perceived benefit items are as follows:18 1. Assuming that the machine has been sold for $32,000, the company got its money’s worth out of the machine. 2. Selling the machine for $32,000 enabled the company to obtain positive overall benefits from the machine. 3. Assuming that the machine has been sold for $32,000, the company got sufficient total benefits from the machine. 4. By selling the machine for $32,000, the machine was economically beneficial on an overall basis. 5. Assuming that the machine has been sold for $32,000, the machine was good for the company on an overall basis. To assess whether these five items measure one unidimensional construct, we perform two analyses. First, to assess scale reliability we compute coefficient alpha (Cronbach, 1951), which provides an indication of how reliably a set of items reflects a single latent construct (Kerlinger & Lee, 2000; Nunnally & Bernstein, 1994). For the five perceived benefit items, we find that coefficient alpha is 0.90, which far exceeds the commonly accepted minimum reliability threshold of 0.70. Second, we factor analyze the five items and find that there is a single eigenvalue in excess of one (3.62). This factor explains approximately 72% of the total variance in the five items. Thus, we conclude that the five items listed above measure one unidimensional construct. In our statistical analyses, we use participants’ average response to the five items to measure perceived benefits.

Results The dependent variable is the minimum price that participants require for a used machine being sold on a consignment basis.19 Our hypothesis predicts that participants will require a lower consignment price when the machine has been depreciated using accelerated depreciation than when the machine has been depreciated using straight-line depreciation. Panel A of Table 4 provides cell means for the two levels of DEP partitioned by the two levels of FVLH, and Panel A of Fig. 4 graphs the mean prices. When fair value is low, the mean consignment price in the straight-line depreciation condition is $34857 and the mean consignment price in the accelerated depreciation condition is $26918. When fair value is high, the mean consignment price in the straight-line depreciation condition is $41721 and the mean consignment price in the accelerated depreciation condition is $29290. Visual inspection of Panel A of Fig. 4 provides initial evidence of a depreciation method-induced book value effect as indicated by the fact that mean consignment prices are larger when the firm 18 We formulated an initial set of 10 items which were evaluated by several colleagues for clarity. Based on their feedback, these items were rephrased. The revised items were then culled down to the five items that we deemed to be most closely aligned with the construct being measured. 19 We do not use the word ‘‘consignment” in the instrument, but the description of the circumstance that we provide to participants is a simple consignment arrangement.

769

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774 Table 4 Cell means and ANOVA results for Experiment 4. Levels of DEP

Levels of FVLH

Rows

Low fair value

High fair value

Panel A: Mean consignment prices and standard deviations Straight-line depreciation Means $34857 Std. dev. $10637 n 18 Accelerated depreciation Means $26918 Std. dev. $3221 n 20 Columns Means $30679 Std. dev. $8571 n 38 Panel B: Mean benefit perceptions and standard deviations Straight-line depreciation Means 4.50 Std. dev. 1.21 n 18 Accelerated depreciation Means 5.45 Std. dev. 1.06 n 20 Columns Means 5.01 Std. dev. 1.22 n 38

Source of variation Panel C: Analysis of variance results for consignment price DEP FVLH DEP  FVLH Error R2 (%) = 35.40 Model F-statistic = 13.15 (p < 0.001) Source of variation

$41721 $9745 20

$38470 $10622 38

$29290 $6678 18

$28042 $5221 38

$35833 $10432 38

$33256 $9831 76

4.24 1.41 20

4.36 1.31 38

4.94 1.24 18

5.21 1.16 38

4.57 1.36 38

4.79 1.30 76

DF

SS

F-statistic

p-Value

1 1 1 72

1965.67 404.13 95.56 4683.24

30.22 6.21 1.47

<0.001 0.015 0.229

DF

SS

F-statistic

p-Value

12.97 2.78 0.29 110.52

8.45 1.81 0.19

0.005 0.183 0.668

Panel D: Analysis of variance results for benefit perception questions DEP 1 FVLH 1 DEP  FVLH 1 Error 72 R2 (%) = 13.13 Model F-statistic = 3.63 (p = 0.017)

In Panel C the dependent variable is the consignment price that participants require for a used machine. In Panel D the dependent variable is the mean of the five perceived benefit questions (see ‘‘Perceived benefits”). The first manipulated variable is the company’s depreciation method choice, which is referred to as DEP. In the accelerated depreciation method condition, accumulated depreciation is $73968, resulting in an accounting book value of $26032. In the straight-line depreciation method condition, accumulated depreciation is $42284, resulting in an accounting book value of $57716. The second manipulated variable is the fair value of the machine, which is referred to as FVLH. In the low fair value condition, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $35000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” In the high fair value condition, participants are informed that the machine was recently inspected and ‘‘the inspector indicated that you might be able to sell the machine for about $49000, while reminding you that sales values for used machines are difficult to determine because they depend on a variety of situational factors such as local market conditions.” For purposes of tabulating the sum of squares in Panel C, we divide the dependent variable by $1000.

uses straight-line depreciation than when the firm uses accelerated depreciation. To formally test our hypothesis, we estimate a 2  2 ANOVA, which is reported in Panel C of Table 4. The ANOVA reveals that there is: (i) a significant main effect of DEP (F-statistic = 30.22, p < 0.001), (ii) a significant main effect of FVLH (F-statistic = 6.21, p = 0.015), and (iii) no significant interactive effect of DEP and FVLH (F-statistic = 1.47,

p = 0.229). Thus, our experimental results indicate that depreciation method-induced differences in the machine’s accounting book have an effect on participants’ consignment price, which provides support for our hypothesis. Further, we find that the significant main effect of depreciation method persists even in the presence of a significant main effect of fair value, which leads us to conclude that accounting book value is not being used as a surrogate

770

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

Panel A: Mean consignment prices $44000 $42000

Mean consignment prices

$40000 $38000 $36000 $34000 $32000 $30000 $28000 $26000 $24000 $22000

Low fair value

High fair value

Straight-line depreciation

Accelerated depreciation

Panel B: Mean benefit perceptions 6.0

Mean benefit perceptions

5.5

5.0

4.5

4.0

3.5

3.0

Low fair value Straight-line depreciation

High fair value Accelerated depreciation

Fig. 4. Mean consignment prices and mean benefit perceptions (Experiment 4).

for fair value (i.e., book value and fair value have their own independent effects). The dependent variable in our second analysis is participants’ perceptions about the benefits that the used machine has provided to the company. We expect that participants will perceive that the past benefits the machine has provided to the company are greater when the machine has been depreciated using accelerated depreciation than when the machine has been depreciated using straight-line depreciation. Panel B of Table 4 provides cell means for the two levels of DEP partitioned by the two levels of FVLH, and Panel B of Fig. 4 graphs the mean perceived benefits. When fair value is low, the mean benefit perception in the straight-line depreciation condition is 4.50 and the mean benefit perception in the accelerated depreciation condition is 5.45. When fair value is high,

the mean benefit perception in the straight-line depreciation condition is 4.24 and the mean benefit perception in the accelerated depreciation condition is 4.94. Visual inspection of Panel B of Fig. 4 provides initial evidence of a depreciation method-induced book value effect as indicated by the fact that mean benefit perceptions are larger when the firm uses accelerated depreciation than when the firm uses straight-line depreciation. To formally test our perception-based explanation, we estimate a 2  2 ANOVA, which is reported in Panel D of Table 4 (note that the perceptions rather than consignment price is the dependent variable). The ANOVA reveals that there is: (i) a significant main effect of DEP (F-statistic = 8.45, p = 0.005), (ii) an insignificant main effect of FVLH (F-statistic = 1.81, p = 0.183), and (iii) an insignificant interaction between DEP and FVLH (F-statistic = 0.19,

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

p = 0.668).20 Thus, our experimental results indicate that depreciation method-induced differences in the machine’s accounting book value have a significant effect on participants’ perceptions about the benefits that the used machine has provided to the company. This finding provides support for our cognition-based explanation for why depreciation method choice influences managers’ asset selling price decisions. Experiment 5 Purpose Experiment 4 provides support for our main research hypothesis and for our cognition-based explanation for the book value effect. However, it is possible that this support arises because the MBA participants lack an adequate understanding of accounting depreciation. Indeed, the judgments and perceptions of individuals who are welleducated in accounting may be entirely unaffected by depreciation method-induced differences in accounting book values. Thus, we rerun Experiment 4 using the same experimental materials, but with accounting students who arguably have in-depth accounting knowledge. By doing so, we are able to assess whether our results and conclusions are sensitive to the level of accounting knowledge of our participants. Participants Participants in Experiment 5 consist of 39 senior-level undergraduate accounting students and 26 master of accountancy students from a large public university. The mean age of the participants is approximately 22 years. Approximately 40% of the participants are male. The participants had previously received instruction on the meaning of accounting book value and the irrelevance of accounting book values for decision making. Participants were asked to voluntarily complete the experimental materials while in class and there was 100% participation. Perceived benefits Our measurement of perceived benefit is discussed in connection with Experiment 4. For the five items, we find that coefficient alpha is 0.94, which far exceeds the minimum reliability threshold of 0.70. We also find that there is a single eigenvalue in excess of one (4.02). This factor explains approximately 80% of the total variance in the five items. Thus, we conclude that the five perceived benefit items measure one unidimensional construct. In our statistical analyses, we use participants’ average response to the five items to measure perceived benefits. Results We discuss but do not tabulate the results for Experiment 5 because the results are very similar to the results for Experiment 4. To summarize, the experimental results 20 At the same time, we find that participants’ perceptions about the machine’s physical condition are uncorrelated with participants’ perceptions about the benefits that the machine has provided to the company (q = 0.048, p = 0.680). This finding suggests that these variables measure non-overlapping constructs.

771

indicate that depreciation method-induced differences in the machine’s accounting book value have a significant effect on participants’ consignment prices (p = 0.002), which provides support for our main hypothesis. Further, we find that the significant main effect of depreciation method persists even in the presence of a significant main effect of fair value (p = 0.009). There is no significant interaction. Our experimental results also indicate that depreciation method-induced differences in the machine’s accounting book value have a significant effect on participants’ perceptions about the benefits that the used machine has provided to the company (p < 0.001), which helps to explain why depreciation method choice has its documented effect. Thus, based on the results of this experiment, we conclude that in-depth accounting knowledge does not alter any of the inferences or conclusions drawn from Experiment 4.

Alternative explanations This section identifies and evaluates four alternative explanations for our experimental findings. The first alternative explanation is that participants believe that the firm’s depreciation method choice (or alternatively stated, the machine’s accounting book value) conveys information about the machine’s current physical condition. To evaluate this alternative explanation, we had participants in each experiment indicate their beliefs about the machine’s physical condition on a 10-point scale after reading information about the machine and being exposed to the manipulated variables. A response of ‘‘1” indicates that participants believe that the machine is ‘‘worn out” and a response of ‘‘10” indicates that participants believe that the machine is ‘‘in top condition.” In each of the experiments, we find that participants’ beliefs about the machine’s physical condition do not vary in response to the manipulated variables (p > 0.10), which indicates that we have successfully equalized participants’ beliefs about the machine’s physical condition, while at the same time manipulating book value and fair value. This finding was expected because we provide all participants with identical information about the machine’s physical attributes and condition (see discussion in the section titled ‘‘Experimental materials, dependent variables, and manipulations”). Thus, we do not attribute our results to depreciation method-induced differences in participants’ beliefs about the machine’s physical condition. The second alternative explanation is that participants believe that the firm’s depreciation method choice conveys information about the machine’s current market value (we address this point further in our discussion of the third alternative explanation). To evaluate this alternative explanation, we had participants respond to a depreciation knowledge question. This multiple-choice question asks participants to indicate which of the following three statements is true: (i) accounting depreciation is a method of allocating costs to the periods that benefit from an asset’s use, (ii) accounting depreciation is a method of determining the actual physical deterioration of an asset, and (iii) accounting depreciation is a method of determining how much an asset is worth if it were to be sold. Participants

772

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

in our experiments respond to this question correctly a minimum of 77% of the time by selecting the first option. Responses to this question suggest that most participants adequately understand what accounting depreciation means. Our inferences and conclusions are unaltered when we add an indicator variable for whether participants answered the depreciation knowledge question incorrectly. The third alternative explanation is that the firm’s depreciation method choice conveys information about the deterioration profile of the machine. The experimental materials inform participants that the company ‘‘depreciates machines over their expected useful lives using the same depreciation method for all machines.” This statement communicates to participants that the company’s depreciation method choice is a firm-wide policy that is not specifically linked to the deterioration profile of any specific asset. To evaluate whether participants interpret this statement in the manner intended, they respond to a multiple-choice question in the post-experimental questionnaire that contains the following option: ‘‘The Company depreciates machines over their expected useful lives using multiple depreciation methods, depending on the physical deterioration profile of each individual machine.” Only a very few participants select this option in Experiments 1–5, and excluding these participants from our analyses does not alter any of our conclusions or inferences. Thus, participants’ selling price decisions are not likely to be caused by an incorrect belief that the firm’s depreciation method choice is informative about the deterioration profile of the machine. The fourth alternative explanation is that participants attempted to avoid a financial statement loss on disposal, which motivated participants in the straight-line depreciation condition to seek higher selling prices. To control for the earnings-related consequences that disposing of the machine may involve, we inform participants that the machine’s selling price (Experiment 1), trade-in value (Experiment 2), auction price (Experiment 3), and consignment price (Experiments 4 and 5) will not significantly affect the company’s financial statements regardless of the amount. As a result, participants know that their selling price decision will not adversely affect the company or any of its stakeholders. In addition, Experiment 3 involved real compensation which was an increasing function of the machine’s selling price. This feature of Experiment 3 should motivate participants in all experimental conditions to act in their self-interest. Even in this experiment, our results strongly support our main research hypothesis. Two additional features of our experiments help to further address whether loss aversion is responsible for our results. First, in Experiment 1, participants receive offer prices from buyers ranging from $20000 to $50000 in $5000 increments. When the offer price is $20000 or $25000, there is a loss on disposal in both experimental conditions. When the offer price is $30000 and above, there is a loss in the straight-line depreciation condition and a gain in the accelerated depreciation condition. If loss aversion is driving the results of Experiment 1, we should observe no experimental effects at the first two offer prices because a loss arises in both experimental conditions. However, as discussed in Experiment 1, we observe the

hypothesized experimental effects at the two initial offer prices, not just those at or above $30000. Second, recall that Experiments 4 and 5 capture a perception measure that we use to understand why depreciation method-induced differences in accounting book values influence managers’ asset selling price decisions. If our theoretical explanation for the book value effect is correct, the perceived benefits from using the machine should vary in response to the firm’s depreciation method choice. On the other hand, if the only cognitive factor in operation is loss aversion, there should be no difference in perceptions across experimental conditions. As discussed in Experiments 4 and 5, we find that depreciation method-induced differences in accounting book values influence participants’ perceptions about the benefits that the machine has provided to the company. This finding affirms our theoretical explanation for our results.21 Summary and limitations When selling a used capital asset, there are at least two factors that managers should always ignore—an asset’s historical cost and accounting depreciation. However, depreciation method-induced differences accounting book values may be difficult for managers to ignore because the physical recording process for accounting depreciation in business settings is structurally and relationally similar to the cognitive recording process for mental depreciation in consumer settings. We contend that these similarities may cause managers to subconsciously apply the decision processes that they use for consumer durable goods to seemingly parallel situations involving business assets. To examine the effect of depreciation method choice on managers’ asset selling price decisions, we manipulate the firm’s depreciation method choice thereby varying an asset’s accounting book value. At the same time, we hold the asset’s historical cost and physical attributes constant. In each decision context that we examine, our evidence indicates that participants sell used capital assets that have been depreciated using accelerated depreciation for lower prices than used capital assets that have been depreciated using straight-line depreciation. This effect even endures when fair value information about the asset being sold is simultaneously provided to participants. Our results also support the intuition that individuals subconsciously apply the decision processes that they use for consumer durable goods to seemingly parallel situations involving business assets. Our study is subject to certain limitations, and we wish to mention two limitations here. First, the instruments used in our experiments may omit information that managers encounter in practice, and participants are not permitted to obtain additional information that they deem to be relevant. For example, managers cannot physically inspect the asset being sold and they cannot consult with other individuals before making their asset selling price decision. Second, although we consistently observe a 21 While we do not believe that loss aversion is the driving psychological force behind our results, we nonetheless acknowledge that loss aversion could play a second-order role.

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

significant book value effect, there are real-world factors that may dampen it. For example, managers may face justification and accountability pressures in their firms. These factors (and others) could negate or significantly reduce the depreciation method-induced book value effect that we document. As a result, we believe that some caution should be exercised when attempting to generalize our results to real-world settings. However, because of the robustness of our findings across contexts and participant groups, we believe that our study provides credible initial evidence that depreciation method-induced differences in accounting book values may influence managers’ asset selling price decisions. References Ahn, S., & Denis, D. (2004). Internal capital markets and investment policy: Evidence from corporate spinoffs. Journal of Financial Economics, 71(3), 489–516. Anderson, J. (1998). Production systems and the ACT-R theory. In Paul Thagard (Ed.), Mind readings: Introductory selections on cognitive sciences. Cambridge, MA: The MIT Press. Archibald, T. (1972). Stock market reaction to the depreciation switchback. The Accounting Review, 47(1), 22–30. Arunachalam, V., & Beck, G. (2002). Functional fixation revisited: The effects of feedback and a repeated measures design on information processing changes in response to an accounting change. Accounting, Organizations and Society, 27(1–2), 1–25. Atanassov, J., & Kim, H. (2009). Labor and corporate governance: International evidence from restructuring decisions. Journal of Finance, 64(1), 341–374. Bartov, E. (1993). The timing of asset sales and earnings manipulation. The Accounting Review, 68(4), 840–855. Bates, T. (2005). Asset sales, investment opportunities, and the use of proceeds. Journal of Finance, 60(1), 105–135. Beaver, W., & Dukes, R. (1973). Interperiod tax allocation and ddepreciation methods: Some empirical results. The Accounting Review, 48(3), 549–559. Bowen, R., DuCharme, L., & Shores, D. (1995). Stakeholders’ implicit claims and accounting method choice. Journal of Accounting and Economics, 20(3), 255–295. Blocher, E., Stout, D., Cokins, G., & Chen, K. (2008). Cost management: A strategic emphasis (4th ed.). New York, NY: McGraw-Hill Irwin. Brockner, J., Houser, R., Birnbaum, G., Lloyd, K., Deitcher, J., Nathanson, S., et al. (1986). Escalation of commitment to an ineffective course of action: The effect of feedback having negative implications for selfidentity. Administrative Science Quarterly, 31(1), 109–126. Colak, G., & Whited, T. (2007). Spin-offs, divestitures, and conglomerate investment. Review of Financial Studies, 20(3), 557–595. Conlon, D., & Garland, H. (1993). The role of project completion information in resource allocation decisions. Academy of Management Journal, 36(2), 402–413. Cronbach, L. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. Davis, M., & Bobko, P. (1986). Contextual effects on escalation processes in public sector decision making. Organizational Behavior and Human Decision Processes, 37(1), 121–138. Dittmar, A., & Shivdasani, A. (2003). Divestitures and divisional investment policies. Journal of Finance, 58(6), 2711–2744. Garland, H. (1990). Throwing good money after bad: The effect of sunk costs on the decision to escalate commitment to an ongoing project. Journal of Applied Psychology, 75(6), 728–731. Garrison, R., & Noreen, E. (2003). Managerial accounting (10th ed.). New York, NY: McGraw-Hill Irwin. Gaumnitz, J., & Emery, D. (1980). Asset growth, abandonment value and the replacement decision of like-for-like capital assets. Journal of Financial and Quantitative Analysis, 15(2), 407–419. Gick, M., & Holyoak, K. (1983). Schema induction and analogical transfer. Cognitive Psychology, 15(1), 1–38. Gourville, J., & Soman, D. (1998). Payment depreciation: The behavioral effects of temporally separating payments from consumption. Journal of Consumer Research, 25(2), 160–174. Hand, J. (1990). A test of the extended functional fixation hypothesis. The Accounting Review, 65(4), 740–763.

773

Heath, C. (1995). Escalation and de-escalation of commitment in response to sunk costs: The role of budgeting in mental accounting. Organizational Behavior and Human Decision Processes, 68(1), 38–54. Heath, C., & Fennema, M. (1996). Mental depreciation and marginal decision making. Organizational Behavior and Human Decision Processes, 68(2), 95–108. Herrmann, D., Inoue, T., & Thomas, W. (2003). The sale of assets to manage earnings in Japan. Journal of Accounting Research, 41(1), 89–108. Hillier, D., McColgan, P., & Werema, S. (2009). Asset sales and firm strategy: An analysis of divestitures by UK companies. European Journal of Finance, 15(1), 71–87. Hilton, R. (2002). Managerial accounting: Creating value in a dynamic business environment (5th edition). New York, NY: McGraw-Hill Irwin. Hirst, D., Joyce, E., & Schadewald, M. (1994). Mental accounting and outcome contiguity in consumer-borrowing decisions. Organizational Behavior and Human Decision Processes, 58(1), 136–152. Hite, G., Owers, J., & Rogers, R. (1987). The market for interfirm asset sales: Partial sell-offs and total liquidations. Journal of Financial Economics, 18(2), 229–252. Holthausen, R. (1981). Evidence on the effect of bond covenants and management compensation contracts on the choice of accounting techniques: The case of the depreciation switch-back. Journal of Accounting and Economics, 3(1), 73–109. Holthausen, R., & Leftwich, R. (1983). The economic consequences of accounting choice: Implications of costly contracting and monitoring. Journal of Accounting and Economics, 5, 77–117. Horngren, C., Sundem, G., & Stratton, W. (2005). Introduction to management accounting (13th ed.). Upper Saddle River, NJ: Pearson Prentice Hall. Howe, K., & McCabe, G. (1983). On optimal asset abandonment and replacement. Journal of Financial and Quantitative Analysis, 18(3), 295–305. Jackson, S. (2008). The effect of firms’ depreciation method choice on managers’ capital investment decisions. The Accounting Review, 83(2), 351–376. Jackson, S., Liu, X., & Cecchini, M. (2009). Economic consequences of firms’ depreciation method choice: Evidence from capital investments. Journal of Accounting and Economics, 48(1), 54–68. John, K., Lang, L., & Netter, J. (1992). The voluntary restructuring of large firms in response to performance decline. Journal of Finance, 47(3), 891–917. John, K., & Ofek, E. (1995). Asset sales and increase in focus. Journal of Financial Economics, 37(1), 105–126. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291. Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39(4), 341–350. Kang, J., & Shivdasani, A. (1997). Corporate restructuring during performance declines in Japan. Journal of Financial Economics, 46(1), 29–65. Kaplan, R., & Roll, R. (1972). Investor evaluation of accounting information: Some empirical evidence. Journal of Business, 45(2), 225–257. Kerlinger, F., & Lee, H. (2000). Foundations of behavioral research (4th ed.). New York, NY: Wadsworth Thomson Learning. Kieso, D., Weygandt, J., & Warfield, T. (2007). Intermediate accounting (12th ed.). New York, NY: John Wiley and Sons. Kuhberger, A. (1998). The influence of framing on risky decisions: A metaanalysis. Organizational Behavior and Human Decision Processes, 1(75), 23–55. Lang, L., Poulsen, A., & Stulz, R. (1995). Asset sales, firm performance, and the agency costs of managerial discretion. Journal of Financial Economics, 37(1), 3–37. Leftwich, R. (1981). Evidence on the impact of mandatory changes in accounting principles on corporate loan agreements. Journal of Accounting and Economics, 3(1), 3–36. Libby, R., Libby, P., & Short, D. (2007). Financial accounting (5th ed.). New York, NY: McGraw-Hill Irwin. LoPucki, L., & Doherty, J. (2007). Bankruptcy fire sales. Michigan Law Review, 106(1), 1–59. Luft, J., & Shields, M. (2001). Why does fixation persist? Experimental evidence on the judgment performance effects of expensing intangibles. The Accounting Review, 76(4), 561–587. Maksimovic, V., & Phillips, G. (1998). Asset efficiency and reallocation decisions of bankrupt firms. Journal of Finance, 53(5), 1495–1532. Maksimovic, V., & Phillips, G. (2001). The market for corporate assets: Who engages in mergers and asset sales and are there efficiency gains? Journal of Finance, 56(6), : 2019–2065.

774

S.B. Jackson et al. / Accounting, Organizations and Society 35 (2010) 757–774

Mauer, D., & Ott, S. (1995). Investment under uncertainty: The case of replacement investment decisions. Journal of Financial and Quantitative Analysis, 30(4), 581–605. Moon, H. (2001). Looking forward and looking back: Integrating completion and sunk cost effects within an escalation-ofcommitment progress decision. Journal of Applied Psychology, 86(1), 104–113. Novick, L. (1988). Analogical transfer, problem similarity, and expertise. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14(3), 510–520. Nunnally, J., & Bernstein, I. (1994). Psychometric theory (3rd ed.). New York, NY: McGraw-Hill, Inc. Okada, E. (2001). Trade-ins, mental accounting, and product replacement decisions. Journal of Consumer Research, 27(4), 433–446. Prelec, D., & Loewenstein, G. (1998). The red and the black: Mental accounting and savings and debt. Marketing Science, 17(1), 4–28. Pulvino, T. (1999). Effects of bankruptcy court protection on asset sales. Journal of Financial Economics, 52(2), 151–186. Ricks, W. (1982). Market assessment of alternative accounting methods: A review of the empirical evidence. Journal of Accounting Literature, 1, 59–102. Samuelson, W., & Zeckhauser, R. (1988). Status quo bias in decision making. Journal of Risk and Uncertainty, 7(1), 7–59. Schlingemann, F., Stulz, R., & Walkling, R. (2002). Divestitures and the liquidity of the market for corporate assets. Journal of Financial Economics, 64(1), 117–144. Shefrin, H., & Statmen, M. (1985). The disposition to sell winners too early and ride losers too long: Theory and evidence. The Journal of Finance, 40(3), 777–790.

Simonson, I., & Staw, B. (1992). Deescalation strategies: A comparison of techniques for reducing commitment to losing course of action. Journal of Applied Psychology, 77(4), 419–426. Smith, V. (1976). Experimental economics: Induced value theory. American Economic Review, 66(2), 274–279. Staw, B. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Organizational Behavior and Human Performance, 16(1), 27–44. Sternberg, R. (1977). Component processes in analogical reasoning. Psychological Review, 84(4), 353–378. Tan, H., & Yates, J. (1995). Sunk cost effects: The influences of instruction and future return estimates. Organizational Behavior and Human Performance, 63(3), 311–319. Tan, H., & Yates, J. (2002). Financial budgets and escalation effects. Organizational Behavior and Human Performance, 87(2), 300–322. Thaler, R. (1985). Mental accounting and consumer choice. Marketing Science, 4(3), 199–214. Thaler, R. (1999). Mental accounting matters. Journal of Behavioral Decision Making, 12(3), 183–206. Titard, P. (1993). Managerial accounting. Houston, TX: Dame Publications. Tuttle, B., & Burton, F. (1999). The effects of a modest incentive on information overload in an investment analysis task. Accounting, Organizations and Society, 24(8), 673–687. Warusawitharana, M. (2008). Corporate asset purchases and sales: Theory and evidence. Journal of Financial Economics, 87(2), 471–497. Watts, R., & Zimmerman, J. (1986). Positive accounting theory. Englewood Cliffs, NJ: Prentice-Hall. Yang, L. (2008). The real determinants of asset sales. Journal of Finance, 63(5), 2231–2262.