Appendix 6-B THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES

Appendix 6-B THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES As noted in the chapter, the LIFO to FIFO choice provides an ideal research topic as the choice has 1. conflicting income and cash flow (tax effect) implications, and 2. data availability allowing for adjustment from one method to the other permitting as-if comparisons in research design. Earlier research focused on market reaction to FIFO-to-LIFO switches and the motivation for using one method as compared to the other. This line of research was consistent with the market-based and positive accounting approaches 1 to research prevalent at that time. More recently, in line with the renewed interest in security valuation issues, researchers have examined the relationship between equity valuation and alternative methods of inventory reporting. Equity Valuation Issues Jennings, Simko, and Thompson (1996) examined the contention that 1. LIFO income statements were more useful than non-lifo statements, and 2. Non-LIFO balance sheets were more useful than LIFO balance sheets by comparing which set of statements better explained the distribution of equity values for a set of LIFO firms. The as if non-lifo statements were created by using the LIFO reserve disclosures and the methodology described in the chapter. Their results were mixed. Consistent with their expectation, they found that LIFO-based income statements explained more of the variation in equity valuations than non-lifo income statements. However, they found that LIFO balance sheets were more useful than their non-lifo counterparts a surprising result given that non-lifo balance sheets are closer to current (rather than outdated LIFO) costs. Jennings et al. explained these results by noting the negative empirical relationship (reported earlier by Guenther and Trombley (1994)) between a firm s value and the magnitude of the LIFO reserve. 2 They argue (and demonstrate using a theoretical model) that if firms cannot (fully) pass on input price increases to their customers, a larger LIFO reserve indicates lower future profitability. In such cases, a negative relationship is expected between firm value and the LIFO reserve. Thus, the poor performance of the non-lifo balance sheet may be explained as follows. When the LIFO reserve is added to LIFO inventory to create the non-lifo balance sheet inventory, the positive relationship between value and assets may be offset by the loss of infor- 1 See Chapter 5 for further discussion. 2 This result seems anomalous because a higher LIFO reserve is indicative of higher asset values. W8

THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES W9 mation (with respect to the effects of inflation) that is provided by the LIFO inventory and LIFO reserve individually. As the elasticity of output prices with respect to input price changes fall, the LIFO and LIFO reserve components of non-lifo inventory have increasingly different implications for future net resource inflows, and loss of information through aggregation increases. 3 An alternative deferred tax explanation for the negative relationship between firm value and the LIFO reserve is offered by Dhaliwal, Trezevant and Wilkins (2000). They argue that the LIFO reserve indicates a potential future tax liability if the inventory (or firm) is liquidated or sold. Whichever argument is correct in explaining the negative relationship between firm value and the LIFO reserve, these results and those with respect to the comparison of LIFO and non-lifo balance sheets point out the need for well-grounded economic analysis when preparing a research design for empirical testing. The LIFO/FIFO Choice As the chapter discussion indicates, there may be sound reasons for firms to stay on FIFO. In addition to those related to LIFO liquidations and declining prices, these reasons include burdensome record keeping requirements, the inability to write down obsolete inventory, and the desire to maximize taxable income when using up a tax loss carryforward. Another reason is the desire to avoid the negative effect of LIFO on a firm s reported earnings. This motivation depends on whether (as discussed in Chapter 5) a market-based or financial contracting argument is used. The market-based argument says that, whether or not the market is efficient and can see through the FIFO/LIFO choice to the real economics of the firm, managers who believe that the market can be fooled by lower reported earnings are reluctant to use LIFO. Alternatively, the financial contracting approach considers the impact of the FIFO/ LIFO choice on management compensation and debt covenant restrictions. The bonus plan hypothesis argues that when top management compensation is based on income, the firm is less likely to use the LIFO method if the resultant lower earnings reduce their compensation. The debt covenant hypothesis argues that the negative effect of LIFO on a firm s reported income and ratios increases the probability that a firm will violate debt covenants regarding such financial measures as working capital, net worth, income, and the dividend payout ratio. Highly leveraged firms may be especially reluctant to use LIFO for that reason, notwithstanding the tax benefits. Studies of the FIFO/LIFO choice generally examine the impact of the choice on firms financial performance in terms of both market reaction and management behavior, as well as the effect on firms financial statements. These studies and the hypotheses tested are affected by both the progression in academic accounting theory and economic factors (such as higher inflation) that caused a resurgence in the adoption of LIFO in the mid-1970s. Market-Based Research LIFO has been permitted in the United States since before World War II, and its rate of adoption understandably follows the rate of inflation. In the 1970s, when the rate of inflation reached double-digits, LIFO adoptions soared. Approximately 400 companies switched from FIFO to LIFO in 1974 alone. This period coincided with heavy academic emphasis on market-based empirical research and the efficient market hypothesis, and the effect of the FIFO/LIFO switch was viewed as an ideal area for research. Given these conditions, the functional fixation hypothesis was tested to see whether: The market accepts financial statements as presented and thus views the switch to LIFO unfavorably since income is depressed. 3 Ross Jennings, Paul J. Simko, and Robert B. Thompson III, Does LIFO Inventory Accounting Improve the Income Statement at the Expense of the Balance Sheet?, Journal of Accounting Research, (Spring 1996), p. 105.

W10 APPENDIX 6-B THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES The market is efficient in the sense that it sees through reported data and views the switch to LIFO positively since cash flow increases. Proponents of the efficient market hypothesis predicted that the market would see through the switch and react favorably to the cash flow effects. Surprisingly, the results were equivocal. Sunder (1973) examined a sample of firms that changed to LIFO in the period 1946 to 1966 and found that prior to the switch these firms experienced positive abnormal returns (Figure 6B-1a). At the time of the change itself, the reaction was slightly negative or nonexistent, as investors seemed to ignore the positive cash flow effect. Moreover, the risk (beta) of firms that switched to LIFO increased in the months surrounding the switch. This result was similar to that of Ball (1972), who examined the market reaction to several accounting changes, FIFO/LIFO included. The positive reaction in the year of the switch was interpreted by some as a sign that the market anticipated the switch and had reacted prior to the actual announcement. Others felt that firms that switched had been having good years and could thus afford the negative impact of the switch, and that these studies suffered from a self-selection bias. Subsequent studies such as Eggleton et al. (1976), Abdel-khalik and McKeown (1978), Brown (1980), and Ricks (1982) extended this research by controlling for earnings-related variables and focusing on the large number of firms that switched in the 1974 to 1975 period. Generally, their results confirmed a negative market reaction in the year of the switch. Ricks, for example, used a control sample of non-lifo adopters (matched on the basis of industry and earnings calculated as if the control company was also on LIFO) and computed the cumulative average return differences between the two groups. His results, presented in Figure 6B-1b, clearly indicate better market performance for firms that did not adopt LIFO. Although these lower market returns were reversed within a year, the initial prolonged negative reaction is difficult to understand. One explanation for this anomalous behavior is that firms that switched to LIFO were those most affected by inflation. Thus, the market may have reacted negatively to the added risk (higher inflation) of these firms, explaining the lower returns and higher risk measures. 4 The difficulty with this explanation is that the sample firms were matched by industry. Thus, we must assume that the sample firms were somehow more adversely affected by inflation than other firms in the same industry. Biddle and Ricks (1988), discussed shortly, also found evidence consistent with this explanation. Implicitly, these studies help explain why firms stayed on FIFO; they wanted to avoid the unfavorable market reaction resulting from the adoption of LIFO. Biddle and Lindhal (1982) attempted to resolve some of these issues by arguing that previous studies did not consider the amount of tax savings from the LIFO adoption. They found a positive association (see Figure 6B-1c) between the market reaction and the estimated tax savings: The results in this study are consistent with a cash-flow hypothesis, which suggests that investor reactions to LIFO adoptions depend on the present value of tax-related cash-flow savings. After controlling for abnormal earnings performance, larger LIFO tax savings were found to be (cross-sectionally) associated with larger cumulative excess returns over the year in which a LIFO adoption (extension) first applied. 5 Biddle and Lindhal studied 311 LIFO adopters from the period 1973 to 1980. The pattern of abnormal returns reported is similar to Sunder s findings (Figure 6B-1a). Neither study used a control group, 6 making these results not directly comparable to those of Ricks. 4 This argument is consistent with the Jennings et al. (1996) explanation (discussed earlier) that the negative association between equity values and the LIFO reserve was related to the inability of firms to pass on higher input prices. 5 Gary C. Biddle and Frederick W. Lindahl, Stock Price Reactions to LIFO Adoptions: The Association Between Excess Returns and LIFO Tax Savings, Journal of Accounting Research, Autumn 1982, Part II, pp. 551 588. 6 Biddle and Lindahl instead used the size of the tax saving as a within-group control.

THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES W11 FIGURE 6B-1 Abnormal returns: Inventory method studies. Sources: (a) 1946 1966 Adopters: Shyam Sunder, Relationship Between Accounting Changes and Stock Prices: Problems of Measurement and Some Empirical Evidence, Journal of Accounting Research, Supplement 1973, pp. 1 45, Fig. 2, p. 18. (b) 1974 1975 Adopters: William E. Ricks, The Market s Response to the 1974 LIFO Adoption, Journal of Accounting Research, Autumn 1982, pp. 367 387, Fig. 2, p. 378. (c) 1973 1982 Adopters: Gary C. Biddle and Fredrick W. Lindahl, Stock Price Reactions to LIFO Adoptions: The Association Between Excess Returns and LIFO Tax Savings, Journal of Accounting Research, Autumn 1982, pp. 551 588, Fig. 1, p. 569. Thus, it is possible that there was some systematic but unexplained factor affecting the 1974 to 1975 adoptions, and that the research results were sensitive to the research design and the time horizon examined. Biddle and Ricks (1988), using daily data, confirmed that there were negative excess market returns around the preliminary dates of firms adopting LIFO in 1974. There is little evidence of

W12 APPENDIX 6-B THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES significant excess returns (negative or positive) near the preliminary dates of firms adopting LIFO in other years. 7 To explain the negative returns, they examined analyst forecast errors for the 1974 LIFO adopters. They found that analysts significantly overestimated the earnings and did not fully appreciate the magnitude of the impact of inflation. 8 In other years, however, the error in analyst forecasts for LIFO adopters was not significant. Further, they found that the negative returns were positively correlated with the forecast error, indicating that the market (as well as analysts) was surprised by the actual reported earnings. Thus, the negative returns were due to the surprise when the market realized that it had underestimated the impact of inflation. As the firms that adopted LIFO were presumably those most affected by inflation, the negative surprise reaction hit them hardest. In later years, however, the market learned from experience and the impact of inflation was more readily factored into earnings estimates. Although these studies shed some light on the market reaction to LIFO adoption, they still do not explain why some firms remain on FIFO. On the contrary, Biddle (1980) found surprising the finding that many firms voluntarily paid tens of millions of dollars in additional income taxes by continuing to use FIFO rather than switching to LIFO. 9 Contracting Theory Approach The contracting theories of accounting choice focus on this issue. Abdel-khalik (1985) examined the bonus plan hypothesis and its implicit corollary that management-controlled firms, in which ownership is widely held, are more likely to use FIFO than owner-controlled firms. The rationale for this argument was that when management is more removed from ownership of the firm, then management compensation rather than the wealth of the firm becomes the primary motivator for manager actions. Thus, the LIFO-induced tax savings are less important to the management-controlled firm. Abdel-khalik found that manager-controlled FIFO firms had relatively higher incomebased bonuses. On the other hand, there was no evidence that differences in compensation plans were related to the FIFO/LIFO choice. In explaining this (non)finding, Abdel-khalik hypothesized that either 1. firms switching to LIFO modify their compensation arrangements, or 2. as some executives have indicated to me, the FIFO-based income continues to be used in determining annual bonus. 10 Hunt (1985) examined the bonus plan and debt convenant hypotheses. His results did not support the bonus plan hypothesis. Contrary to expectations, he found that LIFO firms tended to be less owner-controlled. Hunt, however, did find support for the debt covenant hypothesis, especially with respect to the leverage and interest coverage ratios. His evidence also indicates a threshold level of dividend payout ratios above which firms are reluctant to use LIFO. Dopuch and Pincus (1988) examined the bonus plan, debt covenant, and taxation hypotheses in one study and found that the taxation effect provided the best explanation for the LIFO/FIFO decision. They compared the holding gain that would have accrued to LIFO firms had they stayed on FIFO with the holding gain for firms that remained on FIFO. 7 Gary C. Biddle and William E. Ricks, Analyst Forecast Errors and Stock Price Behavior Near the Earnings Announcement Dates of LIFO Adopters, Journal of Accounting Research, Autumn 1988, pp. 169 194. 8 At that time, LIFO adoptions were unusual, and it took time for analysts to learn to estimate the impact. That they did learn is evidenced by the reduced earnings forecast errors for LIFO adopters in later years. 9 Gary C. Biddle, Accounting Methods and Management Decisions: The Case of Inventory Costing and Inventory Policy, Journal of Accounting Research, Supplement 1980, pp. 235 280. 10 A Rashad Abdel-khalik, The Effect of LIFO-Switching and Firm Ownership on Executive s Pay, Journal of Accounting Research, Autumn 1985, pp. 427 447.

THE FIFO/LIFO CHOICE: EMPIRICAL STUDIES W13 They found larger holding gains for LIFO firms, resulting in higher tax savings. In addition, the holding gain grew as they approached the switch date. Dopuch and Pincus argued that this indicated the long-term FIFO firms in our sample have not been forgoing significant tax savings, in which case remaining on that method is certainly consistent with FIFO being an optimal tax choice, given other considerations. In contrast, long-term LIFO firms would have forgone significant tax savings.... Finally, using the long-term FIFO sample s average holding gains as a base, our change-firms average holding gains became significantly larger than the FIFO average as they approached the year in which they switched, and this difference continued to grow subsequently. 11 Further, Dopuch and Pincus argued that financial analysts could have calculated the increased holding gains for the switch firms and thus anticipated the switch. Therefore, the inconclusive findings of the market reaction studies could be a result of ignoring the advance warning market agents had regarding the switch. More recently, Jennings et al. (1992) supported this advance warning contention. They constructed a model that predicted which firms in the 1974 to 1975 period were more likely to adopt LIFO. The model accurately forecast adopting/nonadopting firms approximately two-thirds of the time. Furthermore, the prior probability of adoption affected the market reaction. The less likely candidates for adoption (according to the model) had more positive market reactions when they adopted LIFO. Similarly, firms that were originally viewed as likely candidates for adoption, but did not adopt, suffered negative market reaction when they failed to adopt LIFO. However, in summing up the research in this area, the editor of The Accounting Review stated We continue to be relatively uninformed about these issues and know little about the real reasons that many firms do not switch to LIFO when it appears that they would benefit by positive tax savings. 12 11 Nicholas Dopuch and Morton Pincus, Evidence of the Choice of Inventory Accounting Methods: LIFO Versus FIFO, Journal of Accounting Research, Spring 1988, pp. 28 59. 12 Editor s Comments, The Accounting Review, Vol. 67, No. 2, April 1992, p. 319.