How Firms Avoid Losses-Use of Net DTA Essay

How Firms Avoid Losses: Evidence of Use of the Net Deferred Tax Asset Account David Burgstahler* University of Washington Gerhard G. Mueller Endowed Professor in Accounting W. Brooke Elliott University of Washington Michelle Hanlon University of Michigan Business School November 26, 2002 _________________________________________________________________________________ ABSTRACT: This paper investigates whether firms use discretion in accounting for deferred taxes to increase earnings and avoid reporting a loss.

We find that firm-years with small scaled profits reduce (relative to the prior year) the proportion of the gross deferred tax asset reserved by the valuation allowance more than firm-years with small scaled losses. We find no evidence that the firm-years that have seemingly moved from having a small scaled loss to a small scaled profit using changes in the net deferred tax asset have greater expected future taxable income to support this change under SFAS 109.

We will write a custom essay sample on
How Firms Avoid Losses-Use of Net DTA Essay
or any similar topic only for you
Order now

Our results also suggest that firms that increase earnings through the net deferred tax asset have relatively lower costs to managing earnings to avoid a loss, that is, these firms have a smaller pre-managed loss. ______________________________________________________________________________ *Corresponding author: David Burgstahler, (206) 543-6316, dburg@u. washington. edu. We are grateful for helpful comments from Bob Bowen, Frank Hodge, Kathryn Kadous, Shiva Rajgopal and Terry Shevlin and workshop participants at the University of Washington and the 2002 UBCOW conference.

We would also like to thank Corey Murata for his assistance in the data collection process. 1. Introduction Since its introduction in 1992, Statement of Financial Accounting Standard No. 109 Accounting for Income Taxes (SFAS 109) has been controversial. Unlike previous statements, e. g. , SFAS No. 96 and Accounting Principles Board (APB) No. 11, SFAS 109 requires managers to value and record deferred tax assets to the extent that they are “more likely than not” to be realized.

Any portion that does not meet the “more likely than not” criteria is reserved using a valuation allowance against the gross deferred tax asset. Petree et al. (1995) claim that recognizing the net deferred tax asset is the most complex and subjective area of SFAS 109. 1 Because the “more likely than not” criteria is subjective, managers have substantial discretion in determining the size of the valuation allowance and therefore also reported earnings. 2 This paper investigates whether firms manage the net deferred tax asset to avoid reporting a loss.

An example of possible earnings management is found in Network Appliance’s 1996 annual report. During a year when the gross deferred tax asset declined slightly from $3. 2 million to $2. 8 million, Network Appliance reduced its valuation allowance from 100% of the $3. 2 million gross asset to approximately 25% of the $2. 8 million gross asset, or $0. 67 million. If Network Appliances had maintained the allowance at 100%, they would have reported a loss instead of showing a small positive earnings number for 1996. 1 For example, see Khalaf (1993), Peavey and Nurnberg (1993), and Petree (1995). Also see Miller and Skinner (1998) which outlines the following reasons to expect earnings management: (1) no well-established guidelines for determining the appropriate level of the allowance; (2) appropriate level of the allowance depends on managers’ expectations about future earnings; and (3) the provision is large enough to allow managers to make material adjustments to accounting earnings. Adjustments to the valuation allowance in many cases flow directly to after-tax income from continuing operations through the provision for deferred taxes (tax expense). As discussed in Bauman, Bauman and Halsey (2000) and Hanlon and Shevlin (2001), adjustments to the valuation allowance do not always affect reported net income. This issue is discussed in greater detail below. 3 Network Associates fiscal 1996 is a firm-year in our sample.

In general, firms can manage the net deferred tax asset account to increase reported income in two primary ways 1) by reducing the valuation allowance while the deferred tax asset remains constant and 2) by not recording the appropriate valuation allowance when the deferred tax asset changes. To investigate both types of earnings management behavior, we focus on changes in the percentage of the gross deferred tax asset reserved by the valuation allowance (hereafter the relative valuation allowance or RVA) as opposed to simply changes in the valuation allowance.

Thus, our measure incorporates changes in the deferred tax asset as well as changes in the valuation allowance. We use the Burgstahler and Dichev (BD, 1997) methodology by comparing firm-years with small scaled profits to firm-years with small scaled losses under the assumption that ex post there is a higher probability that firms with small scaled profits engaged in earnings management. However, firms that were successful in managing earnings to achieve the small scaled profit could have used a wide array of other methods, and we do not know the ordering of the methods that firms use to manage earnings.

As a result, our tests comparing firm-years with small scaled profits to firm-years with small scaled losses may not detect earnings management through the use of the net deferred tax asset even if it exists for a smaller subset of these firms. Thus, we also investigate a sub-group of small scaled profit and loss firms based on their apparent use of the net deferred tax asset to move from a small scaled loss to a small scaled profit. We first provide evidence consistent with BD that frequencies of small losses are abnormally low while frequencies of small profits are abnormally high in our sample.

We initially split our sample into two groups, firm-years with reported small scaled profits (denoted by EM for earnings manager) versus firm-years with reported small scaled losses (denoted by 2 NEM for non-earnings manager). 4 Although these firms arguably have a wide array of reserves at their disposal to use to manage earnings, we find evidence that EM firm-years reduce the relative valuation allowance more than NEM firm-years. By not reserving as much of the gross deferred tax asset, these firms recognize more deferred tax benefit, which reduces income tax expense and increases earnings.

In order to control for reasons for changes in the relative valuation allowance other than opportunistic earnings management, we investigate whether EM firm-years have higher expectations of future taxable income which would justify a greater reduction in the amount of the deferred tax asset offset by the valuation allowance (i. e. , more of the asset would likely be realizable in the future). However, we find no evidence of significant differences between the expected future taxable incomes for firm-years in the two groups.

We next further partition both the EM and NEM groups into sub-groups and focus on the firm-years that would have had losses if there had not been a change in the net deferred tax asset account. 5 We classify firms drawn from the EM group that had losses excluding changes in the net deferred tax asset account and small profits including changes in the net deferred tax asset account as EMVA. 6 Similarly, we define the NEMVA sub-group as firm-years that had losses both excluding changes in the net deferred tax asset account and small losses including changes in the net deferred tax asset account. Thus tests on these sub-groups focus on observations where the net deferred tax asset changed the sign of earnings from a loss to a profit (EMVA) or failed to change earnings from a loss to a profit (NEMVA). When we compare the change in the relative 4 Similar to Burgstahler and Dichev (1997), small losses and small profits are defined as firms with net income in year t scaled by beginning-of-the-year market value of common equity for year t in the range -0. 015 to +0. 015 (i. e. those earnings observations which are close to zero). 5

The net deferred tax asset is defined as the gross deferred tax asset less the valuation allowance. 6 The EMVA sub-group makes up one third of the EM group. 7 Twenty of the NEM firm-years with had small profits prior to the change in the net deferred tax asset and are not included in the NEMVA sub-group. 3 valuation allowance for the EMVA and NEMVA sub-groups we find that the firm-years in the EMVA sub-group reduce the proportion of the gross deferred tax asset reserved by the valuation allowance by a significantly larger amount even after controlling for any differences in expected future taxable incomes.

In addition, we find that firm-years that increased earnings through the net deferred tax asset have lower costs to managing (i. e. , a smaller pre-managed loss). The remainder of the paper proceeds as follows. The next section summarizes the accounting for income taxes and prior research. Section three describes the research design and states the hypotheses. Section four describes the sample selection and the data requirements. Section five reports the results and section six concludes. 2. The Accounting for Income Taxes and Prior Research . 1 Accounting for Income Taxes SFAS 109 provides the current guidelines for the financial accounting for corporate income taxes. SFAS 109 permits the recognition of the tax benefits for items that provide future tax deductions (deferred tax assets). However, SFAS 109 requires managers with deferred tax assets to forecast the portion of future tax benefits that do not meet the criteria that they are “more likely than not” to be realized and to establish a valuation allowance for this portion.

It is because of the “more likely than not” criteria that the valuation allowance is considered subject to discretion. Reported income can be increased by changing the net deferred tax asset account in two primary ways. First, when a company records a deferred tax asset related to a change in their non-tax accruals (e. g. , warranty and bad debt reserves) or the creation of a net operating loss (NOL), managers can choose whether or not to establish a valuation allowance against the 4 deferred tax asset.

If a valuation allowance is established, the deferred tax benefit related to the valuation allowance is not recognized and income tax expense is not reduced. If the valuation allowance established is smaller than required given the future expected profitability of the firm, the income tax expense is reduced, and income is increased by an amount larger than justified by the expected utilization of the deferred tax benefit. Second, a firm can reduce a previously established valuation allowance even when there is no change in the deferred tax asset.

Again, in this case, if expected future taxable income is not sufficient to justify this change, the income tax expense is reduced (or the income tax benefit is increased) and reported net income is increased relative to what these amounts would have been absent the change in the valuation allowance . More subtle versions of these two cases occur when the deferred tax asset and the valuation allowance change in non-proportional ways. For example, if a firm ecords a decrease in deferred tax assets but simultaneously records a larger decrease in the valuation allowance, there will again be a decrease in income tax expense and an increase in reported net income. It is important to simultaneously consider the changes in each account in order to identify both firms that manage by 1) not recording the appropriate level of valuation allowance when deferred tax assets change or 2) by reducing the valuation allowance without a corresponding decrease in the gross deferred tax asset. . 2 Prior Research Prior research has investigated changes in the valuation allowance as a mechanism to manage earnings. For example, Miller and Skinner (1998) investigate whether managers choose the level of the deferred tax asset valuation allowance opportunistically, by testing two wellknown hypotheses: the leverage hypothesis and the smoothing hypothesis (described by Watts 5 and Zimmerman, 1986).

Using regression analysis over a cross-section of firms, Miller and Skinner find no relation between changes in the allowance and changes in leverage or changes in pretax operating earnings (the proxy for the incentive to smooth earnings) and conclude that managers do not opportunistically manage the valuation allowance, but establish the allowance consistent with the guidance of SFAS 109. 8

Visvanathan (1998) tests whether changes in the valuation allowance and changes in current income vary systematically with earnings management motives such as avoidance of violating debt covenants and achieving incentives in bonus for firms in the S 500. He finds no evidence of income-smoothing and no direct associations between changes in the valuation allowance and debt-to-equity ratios or the existence of bonus plan-based incentives. Another example is Schrand and Wong (SW, 2000) who investigate whether bank managers decrease the valuation allowance when earnings before the adjustment to the valuation allowance are below the consensus analyst forecast. SW report evidence consistent with incomeincreasing earnings management when pre-managed earnings are below the forecast. SW investigate a specific set of firms with the incentive to manage earnings and is the first study to report evidence consistent with the use of the valuation allowance as an earnings management tool.

We examine earnings management activities of a broader set of firms (beyond the banking 8 Under the leverage hypothesis, firms close to violating debt covenants are expected to avoid income-reducing accounting choices. To test for smoothing, Miller and Skinner include the change in pretax operating earnings as an independent variable. Assuming the target level for current operating earnings is the previous year’s earnings, smoothing behavior is evident if managers increase (decrease) the valuation allowance when earnings increase (decrease).

The reported coefficient estimate for change in pretax operating earnings was not significantly different from zero. 9 Visvanathan focuses on changes in the valuation allowance after controlling for changes in deferred tax assets and deferred tax liabilities. Although he simultaneously considers changes in the valuation allowance and changes in the deferred tax asset and liability accounts, he does not examine how changing the proportion of the deferred tax asset reserved by the valuation allowance can increase reported earnings.

Also, he does not examine earnings management to avoid missing a target (i. e. , avoid a reported loss). 6 industry) around a different benchmark (avoid a reported loss) using a broader measure (changes in the proportion of the deferred tax asset reserved by the valuation allowance). Bauman, Bauman and Halsey (BBH, 2001) examine the use of the valuation allowance as an earnings management tool for a sample of Fortune 500 firms using a contextual approach that identifies specific earnings targets (e. . , to avoid a loss, avoid reporting an earnings decrease, avoid missing an analyst forecast, and taking a “bath”) around which earnings management may occur and find little evidence of earnings management. Similar to our study, BBH examine earnings management to avoid a reported loss. However, BBH do not select a sample where firms have strong incentives or even the ability to avoid a reported loss by making opportunistic changes in the valuation allowance.

From their original sample, only 29 of their 122 firm-year observations have a loss excluding changes in the valuation allowance (hand collected from the rate reconciliation in the income tax footnote). Thus, the sample of firms that BBH use to examine whether firms use the valuation allowance to manage earnings to avoid a loss consists of only 29 firm-years and these firms do not necessarily have reported losses close to zero. Of these 29 firm-years, only 10 firm-years change the valuation allowance in a manner that results in a positive adjustment to net income. 0 Our study differs from BBH by examining changes in the valuation allowance together with changes in the deferred tax asset and by focusing on a setting where firms have strong incentives to avoid a reported loss. 11 10 Although BBH find virtually no cross-sectional evidence and little contextual evidence consistent with the use of the deferred tax asset valuation allowance to manage earnings around specific earnings targets, an important contribution of this paper is the discussion of the disclosures regarding the income tax provision.

BBH argue that researchers should use the amount representing the change in the valuation allowance listed in the income tax rate reconciliation (i. e. , a list of items that cause a firm’s effective tax rate to be different from the statutory tax rate) as opposed to the net change in the valuation allowance account when examining earnings management using the valuation allowance. In Appendix 1, we discuss the implications for our research of using each measure; the change in the valuation allowance and the line item listed in the rate reconciliation. 1 We note that Phillips, Pincus and Rego (2002) and Hanlon (2002) are studies which, while based on examining tax disclosures, are different from our study because the focus is not on the management of the tax expense itself, but rather the focus is on the information in the deferred tax expense account about non-tax accruals. In contrast, we examine after-tax earnings management of firms that manage earnings by reducing the firm’s tax expense. 7 3. Research Design and Hypotheses

Recent evidence suggests that managers have strong incentives to avoid reporting losses. BD (1997) estimate that 30-44% of firms with small pre-managed losses manage earnings to create positive earnings. Hayn (1995) reports that firms whose earnings are expected to fall just below zero earnings engage in earnings manipulation to help them cross the ‘red line’ for the year. Degeorge, Patel and Zeckhauser (1999) observe discontinuities in earnings distributions that indicate threshold-based earnings management.

They conclude that earnings management is driven by three hierarchically ordered thresholds: report positive earnings, sustain recent performance, and meet analysts’ expectations. We focus on the highest threshold in the hierarchy: to report positive earnings. Following BD, we assume that firm-years with small profits are more likely to have managed earnings upward than firm-years with small losses.

Using reported earnings, we classify firm-years as probable earnings managers (EM), those that reported a small scaled profit, and probable non-earnings managers (NEM), those that reported a small scaled loss. We further partition the sample using a calculated variable, pre-managed earnings. As discussed above, a reported loss can potentially be avoided by reducing the valuation allowance while the deferred tax asset remains constant or by recording less than the appropriate valuation allowance when the deferred tax asset changes.

Thus, to further partition the sample we assume that the entire change in the net deferred tax asset reflects earnings management and calculate earnings as they would have been reported in the absence of this earnings management. 12 Among the EM, we classify firms according to whether the effect of the change in the net deferred tax asset was to 12 We note that it is possible to have a change in the net deferred tax asset that results in a positive adjustment to reported income while the relative valuation allowance, RVA, remains constant.

Thus, our sub-group selection variable and dependent variable in the following test are not the same. 8 move the firm-year from a pre-managed loss to a reported profit (EMVA) or whether the premanaged earnings figure was a profit even without the change in the net deferred tax asset (EMOTHER), where “other” denotes the fact that the sign of reported earnings was not attributable to the net deferred tax asset alone but may have been attributable to other, unknown earnings management (or may not be attributable to earnings management at all).

Among the NEM, we classify firm-years according to whether the pre-managed earnings figure was a loss even without the effect of the change in the net deferred tax asset (NEMVA) or whether the change in the net deferred tax asset caused a pre-managed profit to become a reported loss (NEMOTHER). The sample partitions are illustrated in Figure 1, Panel A. In summary, we have the strongest reason to believe that the change in the net deferred tax asset was used to avoid a loss for EMVA.

For the EMOTHER group, earnings may have been managed to avoid a loss but the change in the net deferred tax asset was not sufficient by itself to create a profit. For the NEMVA group, earnings were not successfully managed to avoid a loss. Finally, for the NEMOTHER group, earnings were not successfully managed to avoid a loss and the effect of the change in the net deferred tax asset was sufficient by itself to change a premanaged profit to a reported loss. Our maintained hypothesis is that in the absence of earnings management, the aluation allowance would have been maintained at the same percentage of the gross deferred tax asset as in the previous year. We calculate the ratio of the valuation allowance to the gross deferred tax asset for each firm-year and investigate how this ratio changes from the prior to the current year. A decrease in the relative valuation allowance is consistent with earnings management, though it might also be explained by an increase in the prospects for future taxable income. 9 The change in the relative valuation allowance is calculated as: ? RVA = VA t GDTA ? t VA t ? 1

GDTA t ? 1 (1) where VAt and VAt-1 are the valuation allowance in year t and t-1 respectively and GDTAt and GDTAt-1, are the gross deferred tax assets in year t and t-1 respectively. 13 Because firms with post-managed earnings levels slightly above zero are more likely to reflect earnings management than firms with post-managed earnings levels slightly below zero, we compare the change in the relative valuation allowance for EM firm-years to the change in the relative valuation allowance for NEM firm-years and predict that EM firm-years will have a larger decrease in the ratio.

We also predict that firm-years in the EMVA group will have a greater decrease in this ratio than firm-years in the NEMVA groups or firm-years in the NEMVA, NEMOTHER and EMOTHER groups combined, ALLOTHER. Formally, our hypotheses are as follows: Hypothesis 1: EM firm-years decrease the relative valuation allowance more than NEM firm-years. Hypothesis 2: EMVA firm-years decrease the relative valuation allowance more than NEMVA firm-years. Hypothesis 3: EMVA firm-years decrease the relative valuation allowance more than ALLOTHER firm-years.

Because firms that expect to have sufficient future taxable income are justified in reducing the relative valuation allowance, we also include regression tests that control for expectations of future taxable income. Our regression is as follows (firm subscripts are omitted): 13 All of these variables were hand collected from the notes to the financial statements from Compact Disclosure. 10 ?RVA = ? + ? 1 EM t + ? 2 ? DTL (t ? ( t ? 1)) + ? 3? DTL (t ? ( t ? 1)) * EM + ? 4 ? ROA (t ? ( t ? 1)) + ? ? ROA (t ? ( t ? 1)) * EM + ? 6 ? MTB t + ? 7 ? MTB t * EM + ? (2) where ? RVA is the change in the firm’s ratio of the deferred tax valuation allowance to the gross deferred tax asset from year t to year t-1, EM is an indicator variable equal to one if the firm-year is a small-scaled profit firm-year and zero if the firm-year is a small-scaled loss firm-year, ? DTL is the change in the firm’s deferred tax liabilities as a proportion of gross deferred tax assets from year t to year t-1 , ?

ROA is the change in the firm’s return on assets from year t to year t-1, and ? MTB is the change in the firm’s market-to-book ratio from year t to year t-1. The indicator variable, EM, is used to estimate the difference between the EM firm-years and the NEM firms-years after controlling for other factors that proxy for expected future taxable income. If firm-years in the EM group reduce the relative valuation allowance more than firmyears in the NEM group after controlling for expected future taxable income, ? 1 will be significantly negative.

This would provide evidence consistent with firms using the net deferred tax asset to manage earnings to avoid a loss. In subsequent regressions, we compare the EMVA sub-group to the NEMVA sub-group by replacing EM with an indicator variable which equals one for firm-years in the EMVA sub-group and 0 otherwise. We estimate this regression first by using only firm-years in EMVA and NEMVA sub-groups and then again by including all firm-years in the sample. We include several variables as proxies for expected future taxable income.

Deferred tax liabilities, upon reversal, constitute taxable income. If management decides these deferred tax liabilities will generate sufficient taxable income against which they can use their deferred tax assets, then a reserve through the valuation allowance is not required. Thus, if firms have an 11 increase in deferred tax liabilities a reduction of the portion of the gross deferred tax asset reserved by a valuation allowance may be in response to this increase. If this is the case in our sample, we expect ? 2 in equation (2) to be significantly less than zero.

Additionally, if EM firmyears have changes in the relative valuation allowance which are more associated with changes in the deferred tax liabilities than NEM firm-years, then we expect ? 3, the coefficient on the interaction term EM*? DTL to be significantly greater than zero. We include the change in the return on assets as a proxy for earnings growth and the market-to-book ratio as a proxy for growth expectations. If a firm experiences current earnings growth or has a high market-tobook, the firm may have prospects for future taxable income which may justify a lower portion of the deferred tax assets to be reserved.

If this is the case, we expect ? 4 and ? 6 to be significantly less than zero. Again, if EM (or EMVA) firm-years are different than the appropriate comparison sample with respect to these variables, thus potentially justifying a change in the relative valuation allowance, ? 5 and ? 7 will be significantly different than zero in our test of equation (2). 4. Sample Selection and Data Requirements We initially consider all observations on the Compustat Annual Industrial File for the years after SFAS 109 became effective, 1993-1998, that meet minimal data requirements.

The earnings variable we focus on is after-tax net income (NIt, Compustat data item #172). In the results reported, the earnings variable is scaled by beginning-of-the-year market value of common equity for year t (MVEt-1, Compustat data item #25 x Compustat data item #199), consistent with BD (1997). Banks, financial institutions, and firms in regulated industries (e. g. , 12 utilities) are deleted. 14 We retain only those firm-years with net income in year t scaled by beginning-of-the-year market value of common equity for year t in the range -0. 015 to + 0. 015 (i. e. those earnings observations which are close to zero). In addition, we limit our sample to those firm-observations that have deferred tax footnote data available on the Compact Disclosure Database for both year t and year t-1. We exclude firm-years that have a zero valuation allowance in both year t and year t-1. Finally, we eliminate those firm-years with negative book value. The resulting sample consists of 482 firm-years. Table 1 summarizes the sample selection process. Figure 1, Panel B, reveals the number of firm-years in our sample and partitioned as we discuss above.

We classify 304 of our sample firm-years as EM, those with a small scaled profit (scaled earnings in the range of 0 to +1. 015). The 178 firm-years in our sample that report a small-scaled loss (scaled earnings in the range of -0. 015 to 0) are classified as NEM. Among the EM, we classify 100 firm-years as EMVA and 204 firm-years as EMOTHER. Among the NEM, we classify 158 firm-years as NEMVA and 20 firm-years as NEMOTHER. 15 We classify all 382 firmsyears in the NEMVA, NEMOTHER and EMOTHER groups combined as ALLOTH. 14

For regulated firms, conflicting incentives to report lower earnings arise whenever there are economic benefits from reporting lower earnings to regulators. For financial institutions, incentives to avoid losses may be linked to regulatory oversight. To focus on cases without these complications, firms with SIC code 6700 and SIC codes between 4400 and 5000 and 6000 and 6500 were deleted from the sample. 15 Firm-years in the EMOTHER group reported small scaled losses, but had small scaled profits prior to the change in the net deferred tax asset.

It is interesting descriptively to note that approximately 39% of firm-years with a ‘premanaged’ loss, as we define it above, reported a profit using a change in the net deferred tax asset while only approximately 9% of the firms with a ‘pre-managed’ profit reported a loss because of a change in the net deferred tax asset. Absent using the net deferred tax asset for earnings management to avoid reporting a loss, one would not expect different rates of movement between these two groups. 13 5. Results 5. 1 Descriptive Statistics

Figure 1 shows the distribution of reported earnings for the entire sample (range is from – 0. 015 to + 0. 015 scaled earnings) in increments of 0. 0025. The sharply higher number of observations in the scaled earnings change intervals to the right of zero relative to the intervals to the left of zero is consistent with BD’s (1997) finding of a much higher frequency of small positive earnings relative to small negative earnings. 16 To conserve space, we report descriptive statistics on only our sub-groups of firm-years. Table 2 reports these descriptive statistics for the EMVA, NEMVA and ALLOTH sub-groups.

The firm-years in the EMVA sub-group are comparable in terms of assets and market value to firmyears in the other sub-groups. The level of deferred tax assets increases by approximately the same amount, on average, for firm-years in both EMVA and NEMVA. The valuation allowance decreases by a significantly greater amount for the firm-years in the EMVA sub-group compared to firm-years in both the NEMVA and ALLOTH sub-groups. The level of the deferred tax liabilities decreased, on average, for all sub-groups from year t-1 to year t.

Thus, while the level of deferred tax liabilities is higher for the EMVA sub-group in both years, the changes between years are not significantly different between the sub-groups. Furthermore, the decrease in deferred tax liabilities for the EMVA sub-group does not appear to justify the decrease in the percentage of the deferred tax assets reserved by the valuation allowance for this sub-group. The change in return on assets (ROA) is not significantly different between the three sub-groups. Finally, similar to the deferred tax liabilities, the market-to-book ratio for each of these groups of 6 The significance of the irregularity near zero is confirmed by the statistical test used in Burgstahler and Dichev (1997) as the standardized difference for the intervals just to the left and just to the right of zero are –2. 4 and 2. 0 in our sample, respectively. These differences are significant at the 5% level. 14 firm-years decreases. The change in the market-to-book ratio is not significantly different in the EMVA sub-group as compared to the NEMVA sub-group. Table 3 reveals that EMVA firm-years are distributed across many industries, as are firmyears in all other sub-groups.

No single industry represents more than 16% of either group. Thus, any differences between the sub-groups are not likely attributable to industry makeup of the groups. Table 4 reports Pearson and Spearman correlation coefficients for our variables. Both correlation analyses reveal that ? RVA is significantly negatively correlated with the ? DTL. This is consistent with an increase in the deferred tax liabilities of a firm leading to a reduction in the amount of gross deferred tax asset reserved because expected future taxable income is higher.

The correlations of ROA and the market-to-book with the change in the percentage of deferred tax assets reserved by a valuation allowance provide little evidence that these measures are used when changing the proportion of the deferred tax asset. We continue to include these variables as controls, however, to be consistent with prior research (e. g. , Miller and Skinner 1998 and Schrand and Wong 2000). 5. 2 Univariate results Table 5 presents the results of t-tests for the change in the relative valuation allowance between select sub-groups.

Panel A provides evidence consistent with EM firm-years reducing the proportion of the gross deferred tax asset reserved by the valuation allowance from year t-1 to year t (i. e. , relative valuation allowance) by more than NEM firm-years. EM firm-years reduce the ratio, on average, by 0. 042 while NEM firm-years reduce the ratio by 0. 010. The difference of the mean change of this ratio is significant (t-stat = 2. 02, p-val = 0. 044. ). 15 Table 5, Panels B and C reveal similar relations for the EMVA sub-group relative to both the NEMVA and ALLOTH sub-groups.

Firm-years in the EMVA sub-group decreased the relative valuation allowance by a larger amount, 0. 126 on average, than firm-years in the NEMVA subgroup, 0. 028 on average (t-stat = 4. 211, p-val = 0. 000). Firm-years in the EMVA sample also decreased the relative valuation allowance by a larger amount than the firm-years in the ALLOTH sub-group, 0. 005 on average (t-stat = 5. 704, p-val = 0. 000). Table 6 presents the results of chi-square tests comparing the frequency of increases and decreases in the relative valuation allowance between firm-years in the sub-groups.

Although the t-tests presented in Table 5 confirm that firm-years in groups with a higher likelihood of managing earnings decrease the ratio by a larger amount than firm-years in groups less likely to have managed earnings, it is possible that firm-years in the former groups decrease the ratio no more often than firm-years in the latter groups. Panel A compares the EM group to the NEM group and provides little evidence that EM firm-years decrease the relative valuation allowance from year t-1 to year t more often than NEM firm-years.

However, Table 6 Panels B and C provide evidence consistent with firm-years in the EMVA sub-group decreasing the relative valuation allowance from year t-1 to year t more often than firm-years in the NEMVA or ALLOTH sub-groups (chi-square stat = 13. 606, p-val = 0. 001 and chi-square stat = 35. 401, p-val = 0. 001, respectively). This is consistent with firms using changes in the net deferred tax asset account, specifically changes in the proportion of the gross deferred tax asset reserved by the valuation allowance, to increase earnings and avoid a reported loss.

Figures 3a through 5c provide graphical representations of the above findings. Figures labeled with an “a” present a histogram of the relative valuation allowance in year t-1, VAt-1 / 16 GDTAt-1 (from 0 to 100%), for the specified sub-group, figures labeled with a “b” present a histogram of the relative valuation allowance for year t, VAt / GDTAt, and figures labeled with “c” present a summary of the difference between “a” and “b”, i. . , the change in the frequency of observations for each ratio interval. A shift in the frequency of observations from the rightmost intervals (a large portion of the deferred tax asset reserved by the valuation allowance) to the leftmost intervals (a small portion of the deferred tax asset reserved by the valuation allowance) is indicative of changing the valuation allowance and the deferred tax asset in order to increase earnings. 7 A comparison of Figure 3c (EMVA) to Figure 4c (NEMVA) and Figure 5c (ALLOTH) provides evidence consistent with firm-years in the EMVA sub-group managing the valuation allowance as a percentage of gross deferred tax assets in order to increase earnings more than other sub-groups of firm-years. 5. 3 Multivariate results 5. 3. 1 EM versus NEM Table 7, Panel A presents the results from estimating equation (2) including all firm-years in the sample. 8 The coefficient on the indicator variable, EM, is significantly negative (? 1 = -0. 121, p-val = 0. 000) consistent with EM firm-years reducing the relative valuation allowance by a larger amount than NEM firm-years after controlling for any linear effects of differing expectations of future taxable income realizations. The coefficient on ? DTL is significantly negative consistent with increases in deferred tax liabilities (i. e. , higher expected future taxable 17

Examining the level of this ratio in each year allows us to observe the frequency of observations from the EMVA sample that moved from the leftmost intervals to the rightmost intervals indicating a substantial change in the valuation allowance as a percentage of the gross deferred tax assets. Examining the change in this ratio would not provide the same information. 18 We check for outliers using Cook’s D statistic in estimating equation (2). Three observations with a Cook’s D value greater than 2 were deleted from the sample. 17 income) leading to a reduction in the proportion of the deferred tax asset reserved.

However, EM is still significant after controlling for this effect. Further, ? DTL does not explain more of the change in the dependent variable for EM firm-years than NEM firm-years; ? 3 is not significantly different than zero. No other coefficients are significant at the 5% significance level. This suggests that the control variables do not explain a significant portion of the change in the ratio for either group, and that the amount of the change in the ratio explained by factors associated with future taxable income is not greater for the EM firm-years than the NEM firmyears.

Thus, an expectation of higher future taxable income for EM firm-years does not seem to explain the larger decrease in the relative valuation allowance. This evidence is consistent with the change in the net deferred tax asset being motivated by earnings management to avoid a loss, rather than by other economic factors included in equation (2). 19,20 5. 3. 2 EMVA versus NEMVA Next, we re-estimate equation (2) defining the indicator variable, EMVA, as equal to one for firms-years in the EMVA sub-group and equal to zero for firm-years in the NEMVA sub-group.

Table 7, Panel B presents the results for this test. The evidence is consistent with firm-years in the EMVA sample reducing the valuation allowance as a proportion of the gross deferred tax asset by a larger amount than firm-years in the NEMVA sample (? 1 = -0. 093, p-val = 0. 001). The only other coefficients that are significant at the 5% confidence level are the coefficients on ? DTL (? 2 19 As a sensitivity test, we re-estimate equation (2) with return on assets in year t+1 as an additional independent variable and find the results are quantitatively similar.

Thus, using future realizations of income as a proxy for expectations of future taxable income does not alter our inferences. 20 We also estimate the following equation where the dependent variable is simply the change in the valuation allowance from year t to year t-1 and find that, similar to prior research, changes in the deferred tax assets and deferred tax liabilities explain a significant portion of changes in the valuation allowance. ? VA = ? + ? 1? GDTA + ? 2 ? GDTA * D + ? 3 ? DTL + ? 4 ? DTL * D + ? 5 ? ROA + ? 6 ? ROA * D + ? 7 ? MTB + ? 8 ? MTB * D + ? 18 = -0. 040, p-val = 0. 20) and on ? DTL * EMVA (? 3 = -0. 065, p-val 0. 006). 21 Thus, ? DTL explains more of ? (VA / GDTA) (t – (t –1)) for firm-years in the EMVA sub-group than for firm- years in the NEMVA sub-group. However, EMVA remains significantly different than zero suggesting that the change in deferred tax liabilities does not explain the entire difference in the change of the relative valuation allowance between firm-years in the EMVA and NEMVA subgroups. This evidence is consistent with firms changing the relative valuation allowance in order to increase income and avoid reporting a loss. 5. 3. EMVA versus ALLOTH Finally, we re-estimate the regression equation (2) defining the indicator variable, EMVA, as equal to one for firm-years in the EMVA sample and zero for firm-years in the ALLOTH sample. Table 7, Panel B presents the results. The evidence is consistent with firm-years in the EMVA sub-group reducing the relative valuation allowance by a larger amount than the firmyears in the ALLOTH sub-group (? 1 = -0. 119, p-val = 0. 000). Again, the only other coefficients that are significant at the 5% significance level are the coefficient on ? DTL (? 2 = -0. 035, p-val = 0. 00) and the coefficient on ? DTL * EMVA (? 3 = -0. 070, p-val 0. 000, respectively). The coefficient on the interaction term, ? 3, is also significantly negative. Thus, ? DTL explains more of ? (VA / GDTA) (t – (t –1)) for firm-years in the EMVA sub-group than for firm-years in the ALLOTH sub-group. However, EMVA remains significantly different than zero suggesting that the change in deferred tax liabilities does not explain the entire difference in the change of the relative valuation allowance between firm-years in the EMVA and ALLOTH samples. Again, 21 A negative coefficient on ?

DTL is expected because deferred tax liabilities represent future taxable income and thus an increase in deferred tax liabilities supports a decrease in the valuation allowance without a corresponding reduction in the deferred tax asset under SFAS 109. 19 this evidence is consistent with the change in the relative valuation allowance being motivated by earnings management to avoid a loss, rather than entirely by other economic factors included in equation (2). 5. 4 Additional Analysis Because we find evidence consistent with some firms adjusting the proportion of the ross deferred tax asset reserved by the valuation to increase earnings and meet an earnings target, we investigate why these firms use this tool for earnings management while other firms do not appear to do so. We examine whether EMVA firm-years have lower costs of eliminating a negative pre-managed loss than NEMVA firm-years, as reflected in the magnitude of the premanaged loss. We contend that if EMVA firm-years have a smaller pre-managed loss; it is less costly to manage the earnings to avoid a loss because a smaller adjustment to the net deferred tax asset is required.

Figures 6a and 6b provide graphical evidence that firm-years in the EMVA sub-group have, on average, smaller magnitude pre-managed losses than firm-years in the NEMVA subgroup and, thus, may explain why the EMVA sub-group was able to report a ‘post-managed’ profit. Figure 6a presents the distribution of pre-managed earnings for EMVA firm-years and Figure 6b presents the distribution of pre-managed earnings for NEMVA firm-years. A comparison of the frequency of observations in each of the graph intervals (0. 0025 in width from 0 to -0. 75 earnings scaled by market value of equity) reveals that for the EMVA sub-group 48% of the observations are concentrated in the first three intervals to the left of zero whereas only 25% of the observations for the NEMVA sub-group are concentrated in these same intervals. This finding indicates that, on average, the EMVA firm-years had a smaller magnitude loss to 20 eliminate than did the NEMVA firm-years, and the smaller magnitude loss would presumably have been less costly to manage away. 6. Conclusions By comparing the changes in the valuation allowance as a percentage of gross deferred tax assets (i. . , relative valuation allowance) for firm-years with small scaled profits (EM) and small scaled losses (NEM), as well as for the EMVA versus both the NEMVA and ALLOTH subgroups, we investigate a specific form of earnings management where incentives to manage earnings are present. We find evidence consistent with managers manipulating the net deferred tax asset account to increase earnings and avoid a loss. We find that EM and EMVA firm-years decrease the relative valuation allowance by a larger amount than NEM, NEMVA and ALLOTH firm-years.

Additional analysis reveals that this result holds even after controlling for other factors that may indicate higher expected future taxable income to support a greater a reduction in the proportion of the deferred tax asset reserved. We also find that the firm-years in the EMVA sub-group were more likely to have smaller “pre-managed” earnings than firm-years in the NEMVA sub-group. We interpret this as evidence that firm-years in the EMVA sub-group have lower costs to manage the relative valuation allowance to avoid a loss because they have a smaller loss to eliminate to achieve a profit.

Overall, in contrast to prior research (with the exception of SW), we find evidence consistent with the suspicions of many that the subjectivity allowed under SFAS 109 in establishing the valuation allowance is used to opportunistically manage earnings. 21 References Accounting Principles Board (APB), 1967, Accounting for Income Taxes, Accounting Principles Board Opinion No. 11, New York, NY: American Institute of Certified Public Accountants. Bauman, C. , M. Bauman and R. Hasley, 2001, Do firms use the deferred tax asset valuation allowance to manage earnings? , JATA Conference Supplement, Vol. 23, 27-48. Behn, B.

K. , T. V. Eaton, J. R. Williams, 1998, The determinants of the deferred tax allowance account under SFAS No. 109, Accounting Horizons, Vol. 12, No. 1, 63-78. Brown, P. R. , 1999, Earnings Management: A subtle (and troublesome) twist to earnings quality, The Journal of Financial Statement, Vol. 4, 61-63. Burgstahler, D. and I. Dichev, 1997, Earnings management to avoid earnings decreases and losses, Journal of Accounting and Economics, Vol. 24, 99-126. Dechow, P. M. and D. J. Skinner, 2000, Earnings management: Reconciling views of accounting academics, practitioners, and regulators, Accounting Horizons, Vol. 4, No. 2, 235-250. Degeorge, F. , Patel, J. and R. Zeckhauser, 1999, Earnings management to exceed thresholds, Journal of Business, Vol. 72, No. 1, 1-33. Financial Accounting Standards Board (FASB), 1987, Accounting for Income Taxes, Statement of Financial Accounting Standards No. 96, Norwalk, CT: FASB. Financial Accounting Standards Board (FASB), 1992, Accounting for Income Taxes, Statement of Financial Accounting Standards No. 109, Norwalk, CT: FASB. Gleason, C. A. and L. Mills, 2002, Materiality and contingent tax liability reporting, The Accounting Review, Vol. 77, No. 2, 317-342. Hanlon, M. 2002, The Persistence and Pricing of Earnings, Accruals and Cash Flows When Firms Have Large Book-Tax Differences. Working paper, University of Michigan. Hanlon, M. and T. Shevlin, 2002, Accounting for the Tax Benefits of Employee Stock Options and Implications for Research, Accounting Horizons, Vol. 16, No. 1, 1-16. Hayn, C. , 1995. The information content of losses, Journal of Accounting and Economics, Vol. 20, 125-53. 22 Healy, P. M. and J. M. Wahlen, 1999, A review of the earnings management literature and its implications for standard setting, Accounting Horizons, Vol. 13, No. 4, 365-383. Khalaf, R. Read Those Footnotes! ”, Forbes, February 1993. Miller, G. S. and D. J. Skinner, 1998, Determinants of the valuation allowance for deferred tax assets under SFAS No. 109, The Accounting Review, Vol. 73, No. 2, 213-233. Peavy, D. E. , and H. Nurnberg, 1993, FASB 109: Auditing considerations of deferred tax assets, Journal of Accountancy, May, 77-81. Petree, T. R. , G. J. Gregory, and R. J. Vitray, 1995, Evaluating deferred tax assets, Journal of Accountancy, March, 71-77. Phillips, J. , M. Pincus and S. Rego, 2001, Earnings management, tax planning and book-tax differences, Working paper, University of Iowa.

Schrand, C. and M. H. F. Wong, 2000, Earnings management and its pricing implications: evidence from banks’ adjustments to the valuation allowance for deferred tax assets under SFAS 109, Working paper, University of Pennsylvania. U. S. Securities and Exchange Commission (SEC), 2000b, Proposed Rule: Supplementary Financial Information, File no. S7-03-00, Washington, D. C. : Government Printing Office. Visvanathan, G. , 1998, Deferred tax valuation allowances and earnings management, Journal of Financial Statement Analysis, Vol. 3, No. 4, 6-15. 23

APPENDIX 1 Discussion of the rate reconciliation versus changes in the net deferred tax asset Bauman et al. (2001) discuss the implications for research on the valuation allowance of using the actual change in the valuation allowance when not all changes in the valuation allowance directly affect income. 22 In our primary analysis, we do not use the information in the rate reconciliation but instead use the gross deferred tax asset and valuation allowance amounts as reported in the tax footnotes. There are several reasons why we chose to do this.

First, we did collect the line item amount from the rate reconciliation for the change identified as a change in the valuation allowance. However, many firms-years from each of the samples did not have an amount listed even when there was a change in the valuation allowance, no merger or acquisition in the year, no line item in additional paid in capital for the tax benefits of stock options nor a discussion of a net operating loss consisting of the tax benefits of stock options, or other items that may normally cause the valuation allowance to be different from the amount in the rate reconciliation.

Similar to the conclusion in Bauman et al. , we suspect that this is due to the change in the valuation allowance in the rate reconciliation being netted with other items in the rate reconciliation. 23 Another reason we did not use the rate reconciliation number is that we investigate the use of the deferred tax asset as well as the change in the valuation allowance to manage earnings. 22 There are two main exceptions as outlined in Bauman et al.

First, under business combinations accounted for under the purchase method if a valuation allowance is recognized for deferred tax assets related to an acquired entity at the acquisition date, the tax benefits for those items that are first recognized (by elimination of the valuation allowance) after the acquisition date are applied to reduce, in order 1) goodwill related to the acquisition, 2) other non-current intangible assets related to the acquisition and 3) income tax expense. Second, are tax benefits charged or credited to shareholders’ equity.

These items include 1) increases or decreases in contributed capital, 2) expenses for employee stock options recognized differently for financial reporting and tax purposes, 3) dividends paid on unallocated shares held by an employee stock ownership plan, and 4) certain deductible temporary differences and carryforwards existing at the date of quasi-reorganization. 23 Firms are only required to separately disclose times in the rate reconciliation if they are 5% or greater of the income tax expense calculated at the statutory rate. See Gleeson and Mills (2000) for a further discussion of disclosure requirements in the rate reconciliation. 4 Thus, the number from the rate reconciliation is not the appropriate variable. For example, assume a firm incurs a net operating loss (NOL) due to stock option compensation deductions upon the exercise of stock options by its employees. At this time, a deferred tax asset is created for this NOL. If a valuation allowance is established on a NOL related to stock options, it is not listed in the rate reconciliation because option deductions nor the valuation allowance on them affect the tax expense, which is exactly the problem Baumen et al. ite. In tests such as ours, if we use the rate reconciliation instead of the actual valuation allowance, we would pick up the increase in the deferred tax asset and a zero from the rate reconciliation which would infer an increasing effect on income. However, there would be no affect on income as the change in the asset is offset by a change in the valuation allowance. As a result, we use the deferred tax asset and the valuation allowance as disclosed in the income tax footnote. We note, as in Bauman et al. 2001) and Hanlon and Shevlin (2002), that some of the changes in the valuation allowance may not affect income and may not be offset by a change in a deferred tax asset. Thus, we may inadvertently include this as a change in the valuation allowance that we assume affects income. For example, in a situation where a firm reduces the valuation allowance on a deferred tax asset of a net operating loss consisting of the tax benefits of stock option deductions without reducing the related deferred tax asset.

However, we do not expect this case to be more prevalent in our samples of firms suspected of earnings management as compared to firm-years with small scaled losses. The industry distribution is similar for all sets of firms and thus we have no ex ante reason to believe that one group uses stock options or undergoes other transactions that have the effect of changing the valuation allowance without affecting income more than other groups. Because we investigate a specific set of firm-years and compare these firm-years to a specific set of control firm-years as opposed to investigating a 25 road cross-section of firm-years, we argue that changes in the valuation allowance that do not affect income are less of a problem for our tests. However, to the extent that EM and EMVA firm-years do use more stock options or undergo more mergers and acquisitions our tests are biased. 26 Table 1 Summary of the Sample Selection Process Total Firms Available on 1998 Compustat for years 1993-1998 with NIt / MVEt-1 between -0. 015 and + 0. 015. 2,234 Less: Firm year observation with no match between Compustat and Compact Disclosure Firm year footnote data not available on Compact Disclosure

Firms with no mention of deferred tax asset(s) in footnotes Firms with incomplete deferred tax footnotes 252 757 191 157 Total Firms with Compact Disclosure Information 877 Less: Firms with a zero valuation allowance in year t and year t-1 Firms with no rate reconciliation in tax footnote Firms with negative book value Firms without deferred tax data available for year t and year t-1 Firms with Cook’s D > 2 in regression tests. 341 8 28 15 3 Entire Sample Total firms with complete data 482 EM Group Firms with Earnings t / MVE t-1 > 0 and < + 0. 015 304 NEM Group

Firms with earnings t / MVE t-1 < 0 and > -0. 015 178 EMVA Sub-group EM firms with negative pre-managed earnings 100 EMOTHER Sub-group EM firms with positive pre-managed earnings 204 NEMVA Sub-group NEM firms with negative pre-managed earnings 158 NEMOTHER Sub-group NEM firms with positive pre-managed earnings 20 ALLOTH Sub-group EMOTHER, NEMVA and NEMOTHER sub-groups combined 382 Note: Earnings (post-managed) are defined as Compustat item #172. MVE is calculated as Compustat items (#25 x #199). Pre-managed earnings are defined as net income in year t (Compustat data item #172 i. e. ost-managed earnings) minus the change in the net deferred tax asset (net deferred tax assets in year t minus net deferred tax assets in year t-1, where net deferred tax assets are defined as gross deferred tax assets in year t minus the valuation allowance in year t). All deferred tax data is hand collected from the firms’ deferred tax footnote 27 0. 31 0. 43 -0. 13 0. 72 0. 76 -0. 04 -0. 01 3. 16 3. 52 -0. 36 Level of Valuation Allowance Year t Year t-1 Year t – Year t-1 Level of Deferred Tax Liability Year t Year t-1 Year t – Year t-1 Return on Assets Year t – Year t-1 MTB Year t Year t-1 Year t – Year t-1 1. 77 2. 50 -0. 26 -0. 2 0. 41 0. 35 0. 00 0. 24 0. 37 -0. 07 10. 92 8. 34 1. 94 1. 14 1. 45 -1. 08 -0. 08 0. 15 0. 08 -0. 15 0. 09 0. 15 -0. 21 4. 05 2. 65 0. 23 4. 05 4. 50 0. 33 0. 02 0. 84 1. 00 0. 10 0. 46 0. 74 -0. 01 54. 28 49. 10 7. 97 3. 54 4. 48 -0. 94 0. 00 0. 44 0. 52 -0. 08 0. 48 0. 51 -0. 03 83. 91 66. 15 17. 76 3. 04 2. 44 0. 09 0. 03 0. 03 -0. 07 1. 97 1. 22 2. 73 1. 82 -0. 49 -1. 52 -0. 03 -0. 09 0. 17 0. 19 0. 00 0. 41 0. 13 0. 41 0. 15 -0. 01 -0. 07 10. 39 9. 21 1. 23 3. 40 5. 01 0. 04 0. 04 0. 53 0. 65 0. 02 0. 87 0. 91 0. 02 35. 58 35. 18 6. 43 944. 93 119. 29 33. 19 629. 93 851. 39 155. 51 44. 50 763. 54 NEMVA (N = 158) Mean Median Q1

Q3 926. 35 75. 41 28. 75 572. 04 3. 68 4. 74 -1. 06 0. 00 0. 54 0. 57 -0. 03 0. 52 0. 53 -0. 01 79. 59 87. 46 -7. 87 2. 10 2. 83 -0. 48 0. 00 0. 16 0. 17 0. 00 0. 47 0. 50 0. 00 9. 07 8. 29 0. 25 1117. 22 109. 96 952. 97 130. 59 1. 19 1. 68 -1. 45 -0. 01 0. 02 0. 01 -0. 04 0. 16 0. 17 -0. 04 2. 66 2. 32 -0. 21 0. 03 0. 08 3. 71 4. 94 0. 18 0. 01 0. 55 0. 59 0. 03 0. 89 0. 92 0. 03 32. 69 31. 12 2. 54 26. 09 509. 81 32. 23 600. 81 ALLOTH (N = 382) Mean Median Q1 Q3 891. 34 72. 19 23. 63 445. 65 EMVA (n = 100) ALLOTH (n = 204) Return on Assets*** Year t 0. 02 0. 01 0. 00 0. 00 0. 01 0. 01 0. 01 Year t-1 0. 02 0. 03 -0. 02 0. 09 -0. 01 . 02 -0. 08 Notes: Valuation allowance and deferred tax liability amounts are scaled by the level of gross deferred tax assets from their respective years. Values in bold are significantly different from comparable values for the EMVA sub-group at the 0. 10 significance level. *** The level of ROA for EMVA firm-years is compared to the level of ROA for ALLOTH firm-years with positive net income for year t. Compustat data item numbers: Total Assets (#6), MTB (#25 x #199) / #60, MVE (#25 x #199), ROA (#48 + #18) / lag (#6) Valuation allowance, deferred tax asset and deferred tax liability data are hand collected from the tax footnote. 07. 84 87. 89 19. 95 MVE Year t Year t-1 Level of Deferred Tax Assets Year t Year t-1 Year t – Year t-1 1271. 41 137. 43 1133. 36 147. 82 Total Assets 43. 84 595. 31 47. 81 535. 31 EMVA (N = 100) Q1 Q3 Mean Median 996. 25 141. 87 46. 26 633. 59 Table 2 Descriptive Statistics for the EMVA, NEMVA and ALLOTH Sub-groups 28 Table 3 Industry Classification and Distribution for the EMVA and ALLOTH Samples EMVA (N = 100) ALLOTH (N = 382) 2-digit SIC Name Frequency % of 100 Frequency % of 382 01 AGRICULTURE PRODUCTION-CROPS 0 0. 00% 1 0. 26% 02 AGRIC PROD-LVSTK,ANIMAL SPEC 0 0. 00% 1 0. 26% 10 METAL MINING 1 1. 00% 9 2. 6% 13 OIL AND GAS EXTRACTION 8 8. 00% 19 4. 97% 14 MNG, QUARRY NONMTL MINERALS 0 0. 00% 1 0. 26% 15 BLDG CNSTR-GEN CONTR,OP BLDR 2 2. 00% 1 0. 26% 16 HEAVY CONSTR-NOT BLDG CONSTR 0 0. 00% 1 0. 26% 20 FOOD AND KINDRED PRODUCTS 1 1. 00% 10 2. 62% 23 APPAREL & OTHER FINISHED PDS 0 0. 00% 2 0. 52% 24 LUMBER AND WOOD PDS, EX FURN 1 1. 00% 0 0. 00% 25 FURNITURE AND FIXTURES 0 0. 00% 1 0. 26% 26 PAPER AND ALLIED PRODUCTS 2 2. 00% 4 1. 05% 27 PRINTING,PUBLISHING & ALLIED 4 4. 00% 3 0. 79% 28 CHEMICALS & ALLIED PRODS 10 10. 00% 32 8. 38% 29 PETE REFINING & RELATED INDS 0 0. 00% 3 0. 79% 30 RUBBER & MISC PLASTICS PRODS 2. 00% 4 1. 05% 32 STONE,CLAY,GLASS,CONCRETE PD 2 2. 00% 3 0. 79% 33 PRIMARY METAL INDUSTRIES 1 1. 00% 6 1. 57% 34 FABR METAL,EX MACHY,TRANS EQ 1 1. 00% 10 2. 62% 35 INDL,COMML MACHY,COMPUTER EQ 10 10. 00% 29 7. 59% 36 ELECTR, OTH ELEC EQ, EX CMP 8 8. 00% 42 10. 99% 37 TRANSPORTATION EQUIPMENT 3 3. 00% 6 1. 57% 38 MEAS INSTR;PHOTO GDS;WATCHES 7 7. 00% 50 13. 09% 39 MISC MANUFACTURNG INDUSTRIES 4 4. 00% 3 0. 79% 42 MOTOR FREIGHT TRANS,WAREHOUS 0 0. 00% 3 0. 79% 50 DURABLE GOODS-WHOLESALE 1 1. 00% 9 2. 36% 51 NONDURABLE GOODS-WHOLESALE 4 4. 00% 6 1. 57% 53 GENERAL MERCHANDISE STORES 0 0. 00% 1 0. 26% 54

FOOD STORES 0 0. 00% 2 0. 52% 55 AUTO DEALERS, GAS STATIONS 0 0. 00% 2 0. 52% 57 HOME FURNITURE & EQUIP STORE 0 0. 00% 1 0. 26% 58 EATING AND DRINKING PLACES 1 1. 00% 11 2. 88% 59 MISCELLANEOUS RETAIL 1 1. 00% 7 1. 83% 65 REAL ESTATE 0 0. 00% 5 1. 31% 70 HOTELS, OTHER LODGING PLACES 0 0. 00% 1 0. 26% 72 PERSONAL SERVICES 1 1. 00% 0 0. 00% 73 BUSINESS SERVICES 16 16. 00% 60 15. 71% 78 MOTION PICTURES 1 1. 00% 5 1. 31% 79 AMUSEMENT & RECREATION SVCS 0 0. 00% 5 1. 31% 80 HEALTH SERVICES 3 3. 00% 8 2. 09% 82 EDUCATIONAL SERVICES 2 2. 00% 4 1. 05% 83 SOCIAL SERVICES 1 1. 00% 1 0. 26% 87 ENGR,ACC,RESH,MGMT,REL SVCS 1. 00% 10 2. 62% 99 NONCLASSIFIABLE ESTABLISHMNT 1 1. 00% 0 0. 00% Notes: No one industry makes up more than 16% of the EMVA or the ALLOTH sample. EMVA sub-group is distributed across many industries in a manner fairly consistent with the ALLOTH sub-group. 29 0. 160 ROA (t – (t-1)) -0. 301 0. 049 0. 113 -0. 027 ( DTL / GDTA) (t – (t-1)) 0. 110 -0. 146 -0. 217 0. 190 RVA MTB (t – (t-1)) -0. 122 -0. 124 Notes: Correlations with two-tailed p-values < 0. 10 in bold. 0. 053 -0. 072 0. 273 0. 652 -0. 041 -0. 616 -0. 629 DTLt / GDTAt 0. 059 -0. 055 -0. 551 0. 927 0. 040 -0. 648 -0. 615 DTLt-1 / GDTAt-1

×

Hi there, would you like to get such a paper? How about receiving a customized one? Check it out