Special charge announcement labeling and timing choices: opportunism or signaling?

ABSTRACT

Special charge announcements often confuse investors trying to evaluate firm performance. In this paper, we investigate whether managers' choices of labeling (restructuring or write-off) and timing (prior to or simultaneously with earnings announcement) are driven more by desires to signal current and future operating performance or to opportunistically mask performance to affect market reactions. Using a sample of large negative special charge announcements from 1986 to 1992, we find the results to be consistent with a signaling scenario. We also find the market generally reacts appropriately to the announcements but are often slow to value improvements in future operating performance.

Keywords: Financial accounting, special charges, restructuring charges, write-offs, earnings announcements.

I. INTRODUCTION

Announcements of special charges often confuse investors trying to evaluate how well the firms have performed and will perform in the future. Further, as suggested by anecdotal reports, managers' choice of labeling special charges as either restructuring charges or write-offs may confuse investors even more (Smith and Lipin, 1996). In addition, the choice to announce special charges prior to or simultaneously with the earnings announcement may also affect investors' reactions. In this paper, we investigate whether managers' choices of labeling (restructuring or write-off) and timing (prior to or simultaneously with earnings announcement) are driven more by desires to signal current and future firm performance or more by desires to opportunistically mask performance to affect market reactions. To investigate whether a signaling or an opportunism scenario prevails, we examine both operating and market performance of announcing firms revealed in the years preceding and following the announcements.

We collected information about large negative special charge announcements from press releases during the time period 1986-1992. This sample of special charge announcements provides a rich data set for investigating how investors' perceptions may be influenced by management's selection of announcement options. During this period, there was little guidance from accounting authorities on special charges, allowing managers to exercise substantial discretion in determining when and to what extent special charges should be reported and announced. After increasing pressure from the AICPA, SEC, and other organizations, FASB attempted to limit management's discretion relating to special charge reporting by adopting EITF 94-3 in 1994 and SFAS 121 in 1996. These accounting standards added somewhat more constraint to how special charges should be reported. However, they did not eliminate the confusion. In a recent article, the Wall Street Journal (January 22, 1999, A2) reported:

"The SEC said company executives concoct a rosy portrait of earnings growth through an assortment of illegal accounting maneuvers. Among other things, the SEC said, companies take excessive restructuring charges. The SEC believes many companies are going too far--taking excessive reserves or write-offs in order to manipulate their results and hide the real health of their business."

Regarding restructuring charges, prior studies have examined firm performance (Carter, 2000; Atiase et al., 2001), earnings management (Moehrle, 2002; Weiss, 1999), impact on stock price (Brickley and Van Drunen, 1990; Bunsis, 1997; Kross et al., 1998) and on analysts' expectations (Chaney et al., 1999), and the underlying management actions (Hogan and Jeter, 1997). Other studies have examined the impact of write-off announcements on stock price (Strong and Meyer, 1987; Elliott and Shaw, 1988; Francis et al., 1996). This study is the first to address whether management's labeling of special charges (restructuring or write-off) and timing (prior to or simultaneously with earnings announcement) are appropriate and whether they matter to the market. The results provide evidence that management chooses the labeling and timing of special charge announcements intentionally to signal the performance of the firm, and the market is efficient enough to understand the signal. This finding is contrary to the common anecdotal belief that corporate managers opportunistically use special charges and the market is not efficient enough to see through the announcements. Our study provides empirical evidence that management's choice of the labeling and timing of special charge announcements can serve as a good indicator of announcing firms' performance. However, we also find that investors are often slow to value improvements in future operating performance for these firms.

The rest of this paper will proceed as follows. Section II examines relevant prior literature supporting the labeling and timing hypotheses. Section III describes the sample selection method and Section IV presents the results and analysis. Finally, Section V summarizes the study.

II. PRIOR RESEARCH AND HYPOTHESIS DEVELOPMENT

Corporate managers' intentions regarding the use of special charge labeling and timing in terms of signaling or opportunism has received scant discussion in the literature. Our study examines current and future operating performance of special charge firms to investigate the association between firm performance and the type of announcement. In addition, our study examines the market's reactions to the different types of special charge announcements, thus providing a more complete picture regarding managers' communication to investors. By doing so, we intend to extend the literature regarding whether managers are generally trying to signal true performance or to opportunistically manipulate investors' perceptions of poor performance when they make special charge announcements.

Labeling of Announcements

Our first research question is whether managers choose an announcement type to signal true performance to investors or to opportunistically manipulate investors' perception of poor performance. Prior research has provided some empirical evidence regarding the performance of special charge firms. Atiase, et al. (2001) found that the performance of restructuring charge firms is worse than that of non-restructuring firms in the pre-restructuring period, but it is better than that of non-restructuring firms in the post-restructuring period. Similarly, Carter (2000) found that restructuring firms realize improvements in future operating performance, but these improvements are not fully realized until at least three years following the restructuring.

Prior research has also examined corporate managers' use of special charges for the purpose of earnings manipulation. Francis et al. (1996) investigated a sample of write-offs between 1989 and 1992, including some described as restructuring charges. They examined the extent to which proxies for managerial incentives to manipulate earnings and proxies for asset impairment explain write-off decisions. For the full sample of write-offs, they found that both proxies are important in making write-off decisions. Moehrle (2002) and Weiss (1999) provided additional evidence of earnings manipulation by managers who report special charges.

Even though restructuring charges and write-offs result in a similar reduction of current period earnings, the market may react differently to the type of label. The words or labels that a company uses can play a big part in how the financial press--and market--reacts to an announcement. In the early 1990's, for example, Hewlett-Packard's first-ever layoffs were reported as a "restructuring" and as "reassignments," etc., and caused little attention. Alternatively, IBM and Digital's "first ever layoffs" led to a flurry of negative articles (PR News, 1997). According to signaling theory, signals should be alterable and therefore potentially subject to manipulation by the signal sender (Spence, 1973; Akerlof, 1976). The combination of management's choice of labeling, the potential difference in market reactions, and the fact that operating performance will only be fully revealed in the future is consistent with a signaling scenario.

Our first hypothesis tests the signaling scenario and involves the association between operating performance and managers' choice of labeling special charges (as either restructuring charges or write-offs). A strong firm can separate itself from a weak firm by giving a costly signal to the stock market. The weaker firm will be less inclined to do so because the signal is more costly (Bhattacharya and Dittmar, 2001). By referring to special charges as restructuring charges, management takes a higher risk of receiving more regulatory attention from accounting regulatory agencies. Even before EITF 94-3 and SFAS 121 were passed, the SEC has been looking closely at the announcements of restructuring charges. Robert Bayless, Chief Accountant of the SEC's Corporation-Finance division, warned the accounting profession in 1994 that many explanations of restructuring charges were "rather sparse and rather vague" and that some companies claim losses as restructuring charges when they are not (Harlan, 1994). Consistent with accounting regulators' concerns, recent articles in the news media pointed out that corporate managers play the "words game" in addition to the "numbers game" with respect to announcing their earnings numbers to analysts and investors (Weil, 2001)

As another example of the potential cost of restructuring charge announcements, 300 companies that labeled their special charges as restructuring charges received letters from the FASB in 1994 requesting that the firms provide legitimate reasons (Harlan, 1994). Such incidents suggest a higher probability that income numbers will need to be restated, and thereby give firms' accounting reporting practices a negative reputation in the stock market. Potentially, this reputation could drive down their stock price. All of these points increase the signaling costs for lower performing firms that opportunistically use the term restructuring charge instead of write-off.

The signaling scenario and the opportunism scenario offer fundamentally different predictions of the announcing firms' operating performance. While the signaling scenario predicts that managers of better performing firms will choose restructuring charges to send out signals, the opportunism scenario predicts that restructuring charge firms are not necessarily better performing companies than write-off firms. The first hypothesis tests the association between labeling and firm performance. Based on the signaling scenario, we predict that restructuring charge firms have better firm performance than write-off firms.

H1: Restructuring charge firms have better current and future operating performance than write-off firms.

Timing of Announcements

In addition to the labeling of the announcement, managers also choose when to announce special charges. To facilitate the study of management's discretionary choices of special charge announcement types, we categorize them into four unique groups using two criteria: labeling and timing. First, we use the labeling of the announcement to partition the sample into two main groups: write-off announcements (Group 1) and restructuring charge announcements (Group 2). Next, we use the timing of the announcement to partition the two groups further. Some managers announce special charges simultaneously with their firms' earnings release (Group A), while others announce prior to earnings release dates (Group B). The result is four distinct sub-groups: (1A) Write-off/simultaneous, (1B) Write-off/prior to, (2A) Restructuring charge/simultaneous, and (2B) Restructuring charge/prior to.

Signaling theories predict that if managers believe their firm's true performance is better than any other special charge firms, they will choose restructuring charges over write-offs to tell investors about their superior performance. Moreover, they will choose to make early announcements so that these announcements look more prominent to investors. For example, when Scottish bank Barclays announced its restructuring before its earnings, observers saw it as a signal that management was pushing ahead with business despite the absence of a full-time chief executive, and its share price increased nearly four percent (Flanagan, 1999). Thus, we expect Restructuring charge/prior to (Group 2B) firms to have stronger operating performance than any other group of special charge announcement firms.

At the other extreme, the signaling scenario predicts that if managers believe their performance is weaker than any other special charge firms, they will choose write-offs and wait until later to bury the news in the earnings report. Thus, we would expect Write-off/simultaneous (Group 1A) firms to have weaker operating performance than any other group of special charge announcement firms. The following hypotheses will be tested:

H2a: Current and future operating performance will be greater for firms reporting restructuring charges prior to earnings (Group 2B) than for any other special charge group.

H2b: Current and future operating performance will be lower for firms reporting write-offs simultaneously with earnings (Group 1A) than for any other special charge group.

Market Reactions

Prior studies that examined stock market reactions to the announcement of restructuring charges and write-offs provide varying conclusions. For example, Strong and Meyer (1987), Elliott and Shaw (1988), and Francis et al. (1996) all find negative overall market reactions to write-off announcements. On the other hand, Brickley and Van Drunen (1990), Bunsis (1997), and Kross et al. (1998) find positive market reactions around the announcement date of restructuring charges. The samples in these studies, however, have often been restricted to either restructuring charges or write-offs, but not both. In addition, most of the studies used various definitions of these charges, as well as relatively short sample periods (one to four years).

In this study, we include both restructuring charge and write-off announcements and differentiate between them (based on management's labeling) to determine if the market reacts more favorably to the former than to the latter. Both the signaling scenario and the opportunism scenario allow for the market to respond more favorably to restructuring charge announcements than to write-off announcements. But if the firms are sending accurate signals with their announcement type, and the market is efficient enough to understand the signal, then investors will also respond more favorably to early restructuring charge announcements than to any other types of announcements. Also, the market will react most unfavorably to late write-off announcements because investors rate these companies as the lowest performing firms. Thus, the market will react in the same order as firm operating performance predicted in Hypotheses 2a and 2b. The following hypotheses are tested:

H3a: The market responds more favorably to restructuring charge announcements than to write-off announcements.

H3b: The market responds more favorably to restructuring charge announcements prior to earnings announcements (Group 2B) than to any other special charge type.

H3c: The market responds least favorably to write-off announcements made simultaneously with earnings announcements (Group 1A) than to any other special charge type.

Many believe that corporate managers opportunistically use special charges and the market is not efficient enough to see through the announcements (Smith and Lipin, 1996; Condon, 1998). Our study first tests whether management's choice of the labeling and timing of special charge announcements serve as a good indicator of announcing firm performance. Second, we test whether investors use the labeling and timing of special charge announcements in their evaluations.

III. SAMPLE SELECTION AND DESCRIPTIVE STATISTICS

We obtained a sample of announcements of large negative special charges made by corporate managers in the news media. Our sample period covers seven years (1986-1992) prior to the adoption of EITF 94-3 and SFAS 121. During this sample period, the lack of relevant accounting rules may have allowed corporate managers more discretion in choosing the labeling and timing of special charge announcements than exists today. It also allows for several subsequent years of operating and market performance to be analyzed. We first used the annual Compustat file to identify 2,409 companies that reported special charges exceeding one percent of their total assets for our sample period. We then used the Lexis/Nexis database to search for initial special charge announcements made by these companies in the news media and found 887 announcements (see Panel A in Table 1).

Panel B in Table 1 categorizes the announcements by type (labeling and timing) of announcements for the sample period (see Appendix for an example of each type). A few interesting points are found when analyzing the announcements by type. First, during the sample period, more companies chose restructuring charges (576, or 65 percent) than write-offs (311, or 35 percent). Second, more firms made announcements on the earnings report date and buried the news in earnings releases (652, or 74 percent) than made separate announcements prior to the earnings report date (235, or 26 percent). Finally, restructuring charge companies made pre-announcements more often than write-off companies. Thirty-six percent (208 observations) of the 576 total restructuring charge announcements were made prior to the earnings report. On the other hand, among the 311 write-off announcements, only 9 percent (27 observations) were pre-announcements.

Panel C in Table 1 reports the industry distribution of the announcement types. There appears to be no industry clustering by announcement type. Using a Chi-square test, the number of announcements in each of the four announcement types is not significantly different statistically by industry (p>. 10).

IV. RESULTS

To investigate whether a signaling or opportunism scenario prevails, we report the results of a series of analyses to examine the relationship of announcement type to firm performance and to market reaction.

Univariate Comparisons of Firm Performance by Announcement Type

We assess announcement firm operating performance using current operating performance (before special charges) along with improvement in future operating performance. Our measure for current operating performance is Operating Income of the announcement year divided by Market Value of the previous year [OIMV]. Operating income [Compustat #178] is disclosed above special charges in the income statement. Market value is the previous year (year -1) closing price [#24] multiplied by the corresponding number of shares [#25]. To provide consistent comparisons of firm performance, we avoid using net income, total assets, and book value reported for the announcement year in comparing current firm performance because these items are affected by the amount of the reported special charges. Similarly, we use the change in operating income (before special charges) as the measure for improvement in future performance. And, because we require firms to have positive values for total assets, book value, and price for both the year of the announcement and the year previous to the announcement (decrease of 136 firms), and because of data availability in CRSP (decrease of 13 firms), the number of firm-announcements used hereafter is decreased from 887 to 738 announcements.

Panel A in Table 2 reports that special charge announcement firms generally have significant problems in their current operating performance. The average OIMV is negative 1.2 percent. However, announcement firms appear to make strong improvement with regards to operating performance after the third year following the initial announcement.

Panel B in Table 2 provides univariate comparisons of the firm performance between write-off firms and restructuring charge firms. Restructuring charge firms appear to have significantly better current operating performance than write-off firms. The OIMV is a positive 3 percent for restructuring charge firms and is significantly better than the negative 5 percent for write-off firms (the difference is statistically significant at the 5 percent alpha level using a non-parametric Wilcoxon test). In addition, although the improvement of future operating performance by restructuring charge firms is inferior for the first two years after the announcement, restructuring charge firms improve at a much faster rate than the write-off firms following the third year. The results for years four and five are statistically significant at the five percent alpha level. Therefore, the results provide evidence that support Hypothesis 1.

Panel C in Table 2 reports that restructuring charge pre-announcement (Group 2B) firms appear to have better current operating performance than other announcement firms. Their OIMV is a positive 5.1 percent, which is significantly better statistically than the other groups (p < .05 using a parametric t-test). Group 2B firms also show greater improvement in future performance after the second year. The differences for years three through five are significant at the five percent alpha level using a nonparametric Wilcoxon test.

As shown in Panel D, simultaneous write-off announcement (Group 1A) firms appear to have the lowest operating performance in comparison with other announcement firms. Their OIMV is a negative 6.5 percent, which is significantly lower than the others (statistically significant at the five percent alpha level using both parametric and non-parametric Wilcoxon tests). Regarding the improvement of future operating income, Group 1A is better than the other announcement firms only in the first two years after the announcement year. However, after that, the other announcement firms consistently exceed Group 1A firms in operating performance improvement. The results for years four and five are significant at the one percent alpha level using both parametric and non-parametric Wilcoxon tests. Overall, the results in Panel C and D of Table 2 provide evidence that support Hypotheses 2a and 2b.

Logistic Analysis of Labeling Choice and Firm Performance

We also use a logit model to examine the relationship between labeling choices and firm performance. Table 3 reports the results of the logit model where the dependent variable, RESTRUCTURING, takes a value of 1 if a restructuring charge announcement was made and a value of 0 if a write-off announcement was made. The model also includes control variables and a set of independent variables that represent past, current, and future operating performance of special charge firms.

Among the control variables, only SIZE (measured by the log of previous year sales) is significant in explaining the choice between a restructuring charge announcement and a write-off announcement. Of particular interest are the independent variables representing operating performance for the past year, current year, and five future years. Table 3 reports that only two are significantly related to labeling choice. The first one is CurOIMV (p=.001) and represents the level of current operating performance of special charge firms. In addition, OperShock4 (p=.02) represents future operating performance improvement from year 3 to year 4. These findings are consistent with the findings provided by univariate tests (see Table 2, Panel B).

Market Reactions to the Announcement--Univariate Comparisons

To examine market reactions to special charge announcements, we first employ a univariate test to compare announcement period price reactions to special charge announcements. Market-adjusted excess return is computed as the raw return minus the market return, using the CRSP equal-weighted index for the day. Panel A in Table 4 reports a significantly negative market reaction to special charge (restructuring charge/write-off) announcements overall. The Cumulative Abnormal Return [CAR] for the two-day announcement period is a negative 1.2 percent, which is statistically significant as indicated by the t-statistic value (two-tailed) of negative 3.4 (p=.001). For pre-announcement firms (Groups 1B and 2B), we also examine the combined market-adjusted CAR around the next quarterly earnings announcement date. The combined market return was significant with a negative CAR of 1.2 percent (parametric t = -2.9, p = .004), suggesting that in general the market reacted to special charge announcements in a significantly negative way, regardless of whether the announcements came with or without an earnings release.

Finding negative overall market reactions adds to the signaling evidence provided by Ajinkya and Gift (1984). They find, contingent upon management signal that the market reacts to the news in an unbiased fashion: an upward price reaction to "good news" and a downward reaction to "bad news." Our finding of overall negative price reaction to special charge announcements suggests that investors consider special charge news to have negative implications in general. The opportunism scenario cannot explain unfavorable market reactions to overall special charge announcements. If managers were to behave opportunistically, they would avoid making announcements as the market would respond unfavorably to the news.

Panel B in Table 4 reports that the market responded more favorably to restructuring charge announcements (Group 2) than to write-off announcements (Group 1). The average market response to the announcement of restructuring charges is a negative CAR of 0.6 percent and for write-off firms is a negative CAR of 2.2 percent (the difference is statistically significant at the 5 percent level using a parametric t-test). These results provide evidence to support Hypothesis 3a.

Panel C in Table 4 illustrates that the average market response around the announcement date for restructuring charge pre-announcements (Group 2B) was a negative CAR of 0.6 percent, and for remaining groups was a negative CAR of 1.3 percent. Although the market appears to have responded slightly more favorably to early announcements of restructuring charges than to other types of announcements, the difference is not statistically significant. Regarding write-off announcements that were made simultaneously with earnings reports (Group 1A), Panel D reports that the market responded less favorably on the earnings release date (negative 2.1 percent) than to other types of announcements (negative 0.7 percent). The difference is statistically significant at the five percent level. Overall, the results provide evidence to support Hypothesis 3c but not 3b.

Market Reactions to the Announcement--Multivariate Regressions

Table 5 reports the results of a multivariate analysis using OLS regressions. Multivariate analysis allows the isolation of stock price reactions to special charge announcements from the effects of other non-special charge information. The regression includes the amount of the special charge [Compustat #17] to measure the surprise of the special charge announcement. We predict that the market will react more negatively to larger restructuring charges or write-offs; therefore, the amount of special charge is predicted to have positive correlation with the market's response.

Because restructuring charges and write-offs are often announced together with earnings releases, we use the variable OPERSURPRISE (operating income surprise) to capture the information contained in the simultaneous earnings reports. OPERSURPRISE is defined as operating income [Compustat # 178] for the announcement year minus operating income for the previous year. Other control variables indicating firms' operating or financial performance include book-to-market ratio (market condition), debt/equity ratio (financial condition), (liquidity), and size (announcement or publication biases).

To examine how the market reacted to different labeling and timing types of announcements, a dummy variable is added to each regression. For example, to test H3a, a dummy variable D1 examines whether market reactions differ based on labeling difference. D1 is equal to one if announcements are for restructuring charges and zero if for write-off announcements. We use the following model to test H3a:

[CAR.sub.i] = [[beta].sub.0] + [[beta].sub.1] [(OPERSURPRISE).sub.i] + [[beta].sub.2] [(SPECIAL).sub.i] + [[beta].sub.3][(BOOKMARK).sub.i] + [[beta].sub.4] [(SIZE).sub.i] + [[beta].sub.5] [(DEBT).sub.i] + [[beta].sub.6] (D1) + [[epsilon].sub.i] (1)

where [CAR.sub.i] = firm i's market-adjusted return compounded over days 0 and +1. The daily market-adjusted return is computed as the raw return minus the return on the CRSP equally-weighted index for the day. OPERSURPRISE = operating performance surprise, measured by the log of reported operating income [Compustat #178] for the announcement year minus reported operating income for the previous year divided by total assets [#6] for the previous year. SPECIAL = special charge amount [#17], measured by the log of reported special items for the announcement year in Compustat divided by total assets for the previous year. BOOKMARK = the log of book value [#216] divided by market value (Prices [#24] multiplied by Number of Shares [#25]) for the previous year. SIZE = the log of sales [#12] for the previous year divided by total assets [#6] for the previous year. DEBT = the log of debt equity ratio (total debt [#181] divided by book value [#216]) for the previous year. Other regressions are defined similar to regression (1) except that each regression has its own corresponding dummy variable that represents the labeling and timing of announcement. To test H3b (H3c), the model has a dummy variable D2B (D1A) that is equal to one if the announcements belong to Group 2B (1A) and equal to zero otherwise. [Insert Table 5 about here]

Model 1 in Table 5 indicates that SPECIAL, which proxies for the amount of special charges (restructuring charges or write-offs), is highly significant in explaining the reactions of the market (p-value=.006). Its positive sign provides evidence that the market reacts negatively to larger restructuring charges and write-offs. OPERSURPRISE also has a positive relationship with market response but is not statistically significant (p=.12) in explaining market response. Hypothesis 3a predicts that the coefficient for the dummy variable (D1) should be positive because of the favorable effect of restructuring charge labeling on the market. The coefficient for D1 has the correct sign and is statistically significant (p=0.05); therefore, it supports Hypothesis 3a. However, it should be noted that the model explains only a small portion of the variation in the market-adjusted return.

Model 2 in Table 5 reports the analysis results regarding market reactions to early announcement of restructuring charges. Dummy variable D2B (for Group 2B announcements) has a positive sign as predicted, but it is not statistically significant (p=.30). Model 3 in Table 5 shows that the dummy variable D1A (for Group 1A announcements) has a negative sign as predicted, and it is moderately significant at the p=.10 level, providing some support for Hypothesis 3c. Again, however, the explanatory power of the model is low. In general, the results from multivariate regression analysis are consistent with the evidence provided by the univariate analysis. The market appears to react to the choice of labeling and timing of special charge announcements consistent with the signaling predictions as presented in hypotheses 3a and 3c.

Future Market Performance of the Announcement Firms

We also examined the market performance of announcement firms following the initial announcement dates. We compare it with the improvement in future operating performance and investigate whether the market was able to predict future operating performance ahead of time. Table 6 presents the results of subsequent market performance of announcement firms for the five years following the announcement date. Future market performance is measured as the sum of future excess returns. For example, future market performance for year +1 is the sum of excess returns from the fifth day after the announcement date to the same date in the following year. The results in Panel A of Table 6 show that market reactions to announcement firms are significantly less favorable than the market average after the third year following the announcement date. Compared to the improvement of operating performance of announcement firms after the third year reported in Table 2, this poor market performance after the third year appears to be incompatible. However, considering that operating performance of the announcement firms is relatively poor until the third year and that only after the third year is there improvement in operating performance for the announcement firms, it appears that investors need time to correctly evaluate the improvement of the announcing firms' performance.

Next, we examine whether the future market performance differs according to announcement labeling. Panel B of Table 6 reports comparisons of future market performance between write-off and restructuring charge announcement firms. The sum of excess returns for write-off firms for the first year after the announcements is larger than that for restructuring charge announcement firms. Compared to Panel B in Table 2, the superior market performance of write-off firms for the first year after the announcements appears to be consistent with these firms' greater operating performance improvement during the same period. On the other hand, while restructuring charge announcement firms show greater operating performance improvement than write-off announcement firms after the third year, the market has not yet correctly evaluated the improvement of restructuring charge firms. Therefore, it appears that because restructuring charge announcement firms show inferior operating performance for the first two years following the announcements, the market needs considerable time to adjust its perception of the quality of these firms.

Panel C in Table 6 compares future market performance of announcement firms by the four group announcement types. Recall that Panels C and D in Table 2 report that Group 2B firms improve their operating performance more than other groups of firms after the third year while Group 1A firms' improvement is less than that of other groups after the third year. However, the pattern of operating performance of these groups of firms is opposite for the first two years after the announcements. Compared to operating performance improvement, Panel C in Table 6 reports that the market appears to correctly evaluate operating performance of these firms for the short-term future (up to two years after the announcements). In particular, the market performance of restructuring charge pre-announcement firms (Group 2B) is the weakest among all the groups during the second year following the announcements. The negative 7.7 percent sum of future excess returns for Group 2B during the second year (significantly smaller than those of other groups at p=.05) is consistent with the inferior operating performance of Group 2B firms during the second year. However, after the third year following the announcements, it appears the market needs substantial time to adjust its evaluation of the firms in a timely manner. In summary, the market appears to be efficient in the short-term but less efficient in the long-term regarding understanding the true operating performance of special charge firms.

Finally, using the CRSP database, we also analyzed the announcement types of firms that did not survive for five years after making a special charge announcement. As shown in Table 7, thirty-nine percent of all announcement firms (346 firms out of 887) were lost in the database for some reason (e.g., bankruptcies, mergers, acquisitions, etc.) during the five-year period. However, the numbers of lost firms do not appear to differ according to announcement type.

V. SUMMARY

Firms continue to report many special charges, including restructuring and write-off announcements. Regulators have tried to limit the discretion available regarding special charges; however, considerable discretion persists. Due to the complicated information content and timing of these announcements, investors often have difficulty appropriately evaluating the true performance of these special charge firms. Managers' discretion in the labeling and timing of announcements has led to debate over whether their choices send a signal to the market to reveal their firms' real performance (signaling scenario) or to mask their firms' poor performance to favorably affect investors' perception of their firms (opportunism scenario).

In this study, we investigate the relationship between the choice of special charge announcement labeling and timing, current and future operating performance, and market reaction to the announcement. In doing so, we examine whether managers' choice of labeling and timing of announcement shows an intent that is more consistent with a signaling scenario or an opportunism scenario. The results are generally consistent with the predictions made in the hypotheses, suggesting that the managers of firms with healthier operating performance tend to label their special charge as a "restructuring charge" and announce it before the earnings announcement. "Weaker" firms, on the other hand, tend to choose the "write-off" label and announce it with their earnings announcement. The market tends to interpret the signals appropriately, although it often needs considerable time to correctly evaluate improvement in future operating performance for these firms.

This study is limited in that it examines only the labeling and timing of special charge announcements, not the underlying actions of management. Also, the modest explanatory power of the OLS regression models suggests that reasons other than those tested in this research contribute to market reactions of special charge announcements. However, this study provides evidence that, unlike the popular media perception that management seeks to mask their firms' true performance using special charge announcements and thus manipulate investors' perception of the firms, managers intentionally choose to announce and select a specific type of labeling and timing of announcement to send a signal to investors. By doing so, managers reveal realistic current and expected future performance of their firms and further differentiate their firms from other special charge firms of lesser quality. These results suggest that investors can use the labeling and timing of the announcements to evaluate the current and future performance of special charge firms.

REFERENCES:

Ajinkya, B., and Gift, M., "Corporate managers' earnings forecasts and symmetrical adjustments of market expectations", Journal of Accounting Research, vol. 22, 1984, 425-444.

Akerlof, G., "The economics of caste, the rat race, and other woeful tales" Quarterly Journal of Economics, vol. 90, 1976, 599-617.

Atiase, R., D. Platt, and Tse, S., "Operational restructuring charges and post-restructuring performance: The potential effects of write-off amounts on measured performance", Working Paper, University of Texas at Austin, 2001.

Bhattacharya, U. and Dittmar, A., "Costless versus costly signaling: theory and evidence from share repurchases", Working paper, Indiana University, Bloomington, 2001.

Brickley, J. and Van Drunen, L., "Internal corporate restructuring: An empirical analysis", Journal of Accounting and Economics, vol. 12, 1990, 251-280.

Bunsis, H, "A description and market analysis of write-off announcements", Journal of Business Accounting and Finance, vol. 24 (9/10), 1997, 1385-1400.

Carter, M., "Operating and stock market performance following corporate restructurings", Working Paper, Columbia University, New York, NY, 2000.

Chaney, P. K., Hogan, C.E., and Jeter, D.C., "The effect of reporting restructuring charges on analysts' forecast revisions and errors", Journal of Accounting & Economics, vol. 27 (3), 1999, 261-284.

Condon, B., "Pick a number, any number", Forbes, vol. 161 (6), 1998, 124-128.

Elliott, J. and Shaw, W., "Write-offs as accounting procedures to manage perceptions", Journal of Accounting Research, vol. 26 (Supplement), 1988, 91-119.

Financial Accounting Standards Board (FASB), "Liability Recognition for Certain Employee Termination Benefits and Other Costs to Exit an Activity (including Certain Costs Incurred in a Restructuring",. Emerging Issues Task Force Issue 94-3. Stamford, CT: FASB, 1994.

Financial Accounting Standards Board (FASB), "Accounting for the Impairment of Long-Lived Assets to Be Disposed Of", Statement of Financial Accounting Standards No. 121. Stamford, CT: FASB, 1995.

Flanagan, M., "Barclays takes axe to 6,000 jobs", The Scotsman, Edinburgh, Scotland (May 21), 1999, 25.

Francis, J., Hanna J., and Vincent, L., "Causes and effects of discretionary asset write-offs", Journal of Accounting Research, vol. 34 (Supplement), 1996, 117-134.

Harlan, C., "SEC wants explanations from firms planning to post restructuring losses", Wall Street Journal, February 25, 1994, B6.

Hogan, C. E., and Jeter, D.C., "The information content of restructuring charges: A contextual analysis", Working Paper, Vanderbilt University, 1997.

Kross, W., Park, T., and Ro, B., "The impact of operational restructuring announcements on stock price, risk, and trading volume", Working Paper, Purdue University, Layfayette, IN, 1998.

Moehrle, S., "Do firms use restructuring charge reversals to meet earnings targets?", The Accounting Review, vol. 77 (2), 2002, 397-413.

PR News. "Kodak's layoffs: Not a good snapshot of communication", Vol. 53 (No. 49, December 15), 1997.

Smith, R., and Lipin, S., "Are companies using restructuring costs to fudge the figures?", Wall Street Journal, January 30, 1996, A1.

Spence, M., "Job market signaling", Quarterly Journal of Economics, 1973, 355-374.

Strong, J. and Meyer, J., "Asset writedowns: Managerial incentives and security returns", Journal of Finance vol. 42 (July), 1987, 643-663.

Weil, J., "What's the P/E ratio? Well, depends on what is meant by earnings", Wall Street Journal, August 21, 2001, A1.

Weiss, I., "Earnings management in response to an exogenous non-recurring item", Working Paper, University of Chicago, IL, 1999.

APPENDIX

EXAMPLES OF VARIOUS ANNOUNCEMENTS OF SPECIAL CHARGES REPORTED IN THE NEWS MEDIA

* Group 1A (write-off announcements made on earnings release dates): "IOMEGA Announces Second Quarter Results" (PR Newswire, July 29, 1987)

"Iomega Corporation announced today the results for its second quarter, ending June 28, 1987. The company reported revenues of $21.5 million and a net loss of $19.4 million, or $1.28 per share, for the period. During the quarter, the company recorded write-offs of approximately $3.1 million associated with capital equipment and inventories for its 8-inch products and engineering changes to its new 5-1/4 inch products."

* Group 1B (write-off announcements made prior to earnings release dates): "Sega to Take Write-off and Post Loss, Illustrating Its Missteps in Video Games" (Wall Street Journal, March 16, 1998)

"Sega Enterprises Ltd. Said it will take a large write-off, report a huge loss and likely scale back sales of its flagship video-game machine, in a humbling admission that the once highflying upstart has fumbled its shot at dominating the video-game market. Sega said it would take a special charge of roughly 47 billion yen ($363.7 million) in order to write off accumulated losses of 40 billion-yen at Sega of America Inc. and a seven billion-yen loss at an Australian unit. Separately, the company said it will also write down 10.5 billion yen in inventories of excess parts and material, as well as 6.5 billion yen in goodwill related to several acquisitions."

* Group 2A (restructuring charge announcements made on earnings release dates): "ELECTRO-NUCLEONICS Reports Fiscal Year 1987 Results" (PR Newswire, September 10, 1987)

"Electro-Nucleonics, Inc. today reported the financial results for the fiscal year ended June 30, 1987.... Net loss for the year was $7.91 million or ($1.81) per share. Vincent V. Abajian, chairman and CEO, commented that, "During 1987, we have restructured the company to expand its position in the immunodiagnostic markets and reduce its emphasis on certain clinical chemistry products. The effect of this decision was a charge of $5.2 million to cost of sales.

* Group 2B (restructuring charge announcements made prior to earnings release dates): "CROSS & TRECKER Announces Restructuring Moves" (PR Newswire, October 8, 1987)

"Cross & Trecker Corporation announced today actions to reduce excess plant capacity and inventory as part of a corporate-wide program to lower costs and consolidate and restructure manufacturing operations."

Gyung Paik, Brigham Young University, Provo, Utah, USA Kip R. Krumwiede, Boise State University, Boise, Idaho, USA

Dr. Gyung H. Paik earned his Ph.D. at the University of Illinois, Urbana-Champaign in 2001. Currently he is an assistant professor of accounting at Brigham Young University, Provo, Utah.

Dr. Kip R. Krumwiede earned his Ph.D. at the University of Tennessee in 1996. Currently he is an assistant professor of accounting at Boise State University, Boise, Idaho.

TABLE 1
SAMPLE SELECTION FOR SPECIAL CHARGE ANNOUNCEMENTS BY YEAR, TYPE, AND
INDUSTRY

Panel A: Sample Announcements by Year (both restructuring and
write-off announcements)

                           Number of
Year              Special Charge Announcements

1986                         132
1987                         114
1988                         107
1989                         139
1990                         162
1991                         114
1992                         119
Total                        887

Panel B: Sample Announcements by Type of Announcements
Labeling
                                            Labeling

                           Write-offs      Restructuring
Timing                      (Group 1     Charges (Group 2)     Total

Announcement on
Earnings Report Date        284 (32%)        368 (41%)       652 (74%)
(Group A)
Announcement prior to
Earnings Report Date         27 (3%)         208 (23%)       235 (26%)
(Group B)
Total                       311 (35%)        576 (65%)       887 (100%)

Panel C: Announcement Type by Industry

                       Announcement Type (1,2)

                         (1A)            (1B)

                       Write-off       Write-off
                     Simultaneous        Pre-
Industry             Announcement    Announcement

Food                       4               0
                         (25%)           (0%)
Paper & Printing           5               0
                         (21%)           (0%)
Chemicals & Allied        23               0
  Products               (33%)           (0%)
Metals                    12               3
                         (24%)           (6%)
Machinery &               27               4
  Computer               (27%)           (4%)
Electronics &             23               1
  Electrical             (29%)           (1%)
Transportation             9               2
                         (23%)           (5%)
Measurement                9               1
  Instruments            (21%)           (2%)
Business Services         24               4
                         (34%)           (6%)
All Others                148             12
                         (37%)           (3%)
    Total                 284             27
                         (32%)           (3%)

                        Announcement Type (1,2)

                         (2A)            (2B)
                     Restructuring   Restructuring
                        Charge          Charge
                     Simultaneous        Pre-
Industry             Announcement    Announcement        Total

Food                      10               2              16
                         (63%)           (13%)          (100%)
Paper & Printing          12               7              24
                         (50%)           (29%)          (100%)
Chemicals & Allied        26              21              70
  Products               (37%)           (30%)          (100%)
Metals                    24              12              51
                         (47%)           (24%)          (100%)
Machinery &               36              34              101
  Computer               (36%)           (34%)          (100%)
Electronics &             38              17              79
  Electrical             (48%)           (22%)          (100%)
Transportation            16              12              39
                         (41%)           (31%)          (100%)
Measurement               19              13              42
  Instruments            (45%)           (31%)          (100%)
Business Services         31              11              70
                         (44%)           (16%)          (100%)
All Others                156             79              395
                         (39%)           (20%)          (100%)
    Total                 368             208             887
                         (41%)           (23%)          (100%)

(1) The numbers represent the number of announcements made by the
corresponding industry. The numbers in parentheses are the percentage
of the number of announcements in the specific announcement type when
the number of announcements made by the corresponding industry is
considered as 100 percent.

(2) The number of announcements in each of the four-group announcement
types (Group 1A, 1B, 2A, and 2B) is not statistically significantly
different by industry using a Chi-square test ([chi square] = 35.6,
p > .10).

TABLE 2
UNIVARIATE COMPARISON OF FIRM PERFORMANCE (N = 738)

                                                 Improvement in
                                 Current        Future Operating
                                Operating        Performance (2)
                             Performance (1)
                                 Year 0        Year +1    Year +2

Panel A: All firms
  Overall Special Charge
    Announcement Firms           -0.012          -2.4      -38.0
    (n = 738)

Panel B: Hypothesis 1
  Write-off Announcement
    Firms (n = 258)              -0.050          12.6       1.5
  Restructuring Charge
    Announcement Firms            0.030          -9.0      -59.2
    (n = 480)
      t-stat                     -1.7 *         1.8 **      1.3
      Wilcoxon z-score           -1.9 **         0.9        0.5

Panel C: Hypothesis 2a
  Restructuring Charge
    Pre-Announcement Firms        0.051          4.0       -126.2
    (Group 213: n = 182)
  Other Types of
    Announcement Firms           -0.037          -4.6      -12.0
    (n = 556)
      t-stat                     2.0 **          0.7      -2.9 ***
      Wilcoxon z-score             0.6           0.9      -1.8 **

Panel D: Hypothesis 2b
  Write-off Simultaneous
    Announcement Firms           -0.065          8.1        2.3
    (Group 1A: n = 235)
  Other Types of
    Announcement Firms            0.001          -6.8      -56.7
    (n = 503)
      t-stat                     -1.9 **         1.3       2.4 **
      Wilcoxon z-score           -1.7 **         0.4        0.9

                                          Improvement in
                                         Future Operating
                                          Performance (2)

                                 Year +3       Year +4    Year +5

Panel A: All firms
  Overall Special Charge
    Announcement Firms            -1.8           57.7       66.2
    (n = 738)

Panel B: Hypothesis 1
  Write-off Announcement
    Firms (n = 258)               -5.8           23.0       32.4
  Restructuring Charge
    Announcement Firms             0.4           76.4       84.6
    (n = 480)
      t-stat                      -0.3         -2.5 **    -1.9 **
      Wilcoxon z-score            1.4 *        -1.9 **    -1.8 **

Panel C: Hypothesis 2a
  Restructuring Charge
    Pre-Announcement Firms        18.1           94.7      154.9
    (Group 213: n = 182)
  Other Types of
    Announcement Firms            -7.6           46.8       40.1
    (n = 556)
      t-stat                       1.2           1.5      3.0 ***
      Wilcoxon z-score           1.9 **         1.8 **     2.2 **

Panel D: Hypothesis 2b
  Write-off Simultaneous
    Announcement Firms            -11.6          19.6       20.3
    (Group 1A: n = 235)
  Other Types of
    Announcement Firms             2.8           75.4       87.6
    (n = 503)
      t-stat                      -0.8         -2.7 ***   -2.7 ***
      Wilcoxon z-score           -1.2 *        -2.4 ***   -2.2 ***

(1) Current operating performance is measured by operating income
[Compustat #178] of the announcement year (Year 0) divided by market
value (price [#24] multiplied by number of shares [#25]) of the
previous year (year -1).

(2) Improvement in future operating performance is measured by
operating income of the year minus operating income of the previous
year. Because some sample firms did not survive is subsequent years,
only 655 firm-year announcements are used for future performance
comparisons for year +5.

Note: * = Significantly different at the 10% alpha level;
** = Significantly different at the 5% alpha level; *** =
Significantly different at the 1 % alpha level.

TABLE 3
LOGIT MODEL RESULTS: LABELING CHOICE AND FIRM PERFORMANCE
(N = 738; 258 WRITE-OFF VERSUS 480 RESTRUCTURING CHARGE
ANNOUNCEMENT FIRMS)

[RESTRUCTURING.sub.it] = a + [b.sub.1] * [(SPECIAL).sub.it]
+ [b.sub.2] * [(BOOKMARK).sub.it] + [b.sub.3] * [(CURRATIO).sub.it]
+ [b.sub.4] * (DEBT) + [b.sub.5] * [(SIZE).sub.it] + [b.sub.6] *
[(PASTOIMV).sub.it] + [b.sub.7] * [(CUROIMV).sub.it] + [b.sub.8] *
[(OPERSHOCKO).sub.it] + [b.sub.9] * [(OPERSHOCK1).sub.it]
+ [b.sub.10] * [(OPERSHOCK2).sub.it] + [b.sub.11] *
[(OPERSHOCK3).sub.it] + [b.sub.12] * [(OPERSHOCK4).sub.it]
+ [b.sub.13] * [(OPERSHOCK5).sub.it] + [[epsilon].sub.it]

                                  Predicted     Coefficient
Model (1)          Coefficient       Sign        Estimate

Intercept          Intercept          ?            1.9000
Control            SPECIAL            +            0.1000
Control            BOOKMARK           -           -0.0005
Control            CURRATIO           ?            0.0010
Control            DEBT               -           -0.0040
Control            SIZE               +            0.3600
Past (level)       PASTOIMV           +            0.0080
Current (level)    CUROIMV            +            0.1400
Current (change)   OPERSHOCKO         -            0.0200
Future (change)    OPERSHOCK1         -           -0.0040
Future (change)    OPERSHOCK2         ?           -0.0040
Future (change)    OPERSHOCK3         ?           -0.0200
Future (change)    OPERSHOCK4         +            0.0500
Future (change)    OPERSHOCK5         +            0.0080

                                   p-value
Model (1)          Chi-Square    (pr > ChiSq)

Intercept             68.70         0.001
Control                2.20         0.140
Control                0.01         0.900
Control                0.40         0.500
Control                1.10         0.300
Control               81.10         0.001
Past (level)           0.03         0.800
Current (level)        9.70         0.001
Current (change)       0.50         0.500
Future (change)        0.04         0.800
Future (change)        0.30         0.600
Future (change)        1.80         0.180
Future (change)        4.80         0.020
Future (change)        0.30         0.600

(1) Likelihood Ratio = 129.10 (p < .001)

RESTRUCTURING; = 1 if a restructuring charge announcement made, zero
if a write-off announcement made; SPECIAL = special charge amount
[#17], measured by log of reported special items for the announcement
year in Compustat divided by total assets for the previous year;
BOOKMARK = the log of book value [#216] divided by market value
(prices [#24] multiplied by number of shares [#25] for the previous
year); CURRATIO = the log of current asset [#4] divided by current
liability [#5] for the previous year; DEB = the log of debt equity
ratio (total debt [#181] divided by book value [#216]) for the
previous year; SIZE = the log of sales [#12] for the previous year;
PASTOIMV = past operating performance measure (level), measured by
the log of reported operating income [Compustat #178] for the previous
year divided by market value (prices [#24] times number of shares
[#25]) for the previous year; CUROIMV = current operating performance
measure (level), measured by the log of reported operating income
[Compustat # 178] for the announcement year divided by market value
(prices [#24] times number of shares [#25]) for the previous year;
OPERSHOCKt =changes in operating performance, measured by the log of
reported operating income for the current year (t) minus operating
income for the previous year (t - 1).

TABLE 4
MARKET REACTIONS TO SPECIAL CHARGE ANNOUNCEMENTS (N = 738)

                                 Market Reaction

                                                     CAR around the
                                   CAR around      Following Earnings
                                 Special Charge        Release 2
                                 Announcement (1)  (Pre-announcement
                                   (All Firms)           Firms)
Panel A: All Firms
  Overall Special Charge             -0.012              -0.003
    Announcements
      t-value                      (-3.4) ***            (-0.3)

Panel B: Hypothesis 3a
  Write-off Announcements          -0.022 ***            0.036
    (n = 258)
  Restructuring Charge
    Announcements (n = 480)         -0.006 **            -0.008
      t-value                        -2.3 **             1.9 **

Panel C: Hypothesis 3b
  Restructuring Charge Pre-
    Announcement (Grp. 2B:           -0.006              -0.008
      n = 182)
  Other Types of Announcement:
   (Group 1A, 1B, 2A: n = 556)       -0.013              0.036
     t-value                           0.9              -1.9 **

Panel D: Hypothesis 3c
  Write-off Simultaneous
    Announcement (Grp. 1A:           -0.021              0.000
      n = 235)
  Other Types of Announcement:
    (Group 1B, 2A, 213:              -0.007              -0.003
      n = 503) t-value               -2.0 **              0.30

                                 Market Reaction

                                    Combined
                                     CAR (3)
                                   (All Firms)
Panel A: All Firms
  Overall Special Charge             -0.012
    Announcements
      t-value                      (-2.9) ***

Panel B: Hypothesis 3a
  Write-off Announcements            -0.019
    (n = 258)
  Restructuring Charge
    Announcements (n = 480)          -0.008
      t-value                         -1.3

Panel C: Hypothesis 3b
  Restructuring Charge Pre-
    Announcement (Grp. 2B:           -0.014
      n = 182)
  Other Types of Announcement:
   (Group 1A, 1B, 2A: n = 556)       -0.012
     t-value                          -0.3

Panel D: Hypothesis 3c
  Write-off Simultaneous
    Announcement (Grp. 1A:           -0.021
      n = 235)
  Other Types of Announcement:
    (Group 1B, 2A, 213:              -0.008
      n = 503) t-value               -1.8 **

(1) CAR around Special Charge Announcements: market reactions to
special charge announcements are measured by 2-day (day 0 and day +1)
Cumulative Abnormal Returns (CAR: individual returns minus equal
weighted market returns) around the announcement date. For Group 1A
and 2A (simultaneous announcement groups), announcement date coincides
with the earnings report date. For Group 1B and 2B (pre-announcement
groups), announcement date is different from the earnings report date.

(2) CAR around the Following Earnings Release: for special charge
pre-announcement groups (Group 1B and Group 2B), market reactions to
the following earnings report are measured by five-day (day -2 through
day +2) CAR around the earnings report date.

(3) Combined CAR: for Group 1A and 2A (simultaneous announcement
groups), it is the same as CAR around special charge announcement; for
Group 1B and 2B (pre-announcement groups), it is the sum of CAR around
special charge announcement date and CAR around the following earnings
report.

Note: * = Significantly different at the 10% alpha level; ** =
Significantly different at the 5% alpha level; ***  Significantly
different at the 1% alpha level.

TABLE 5
OLS REGRESSION ANALYSIS: MARKET REACTIONS TO SPECIAL CHARGE
ANNOUNCEMENTS (N = 738)

General Model for Short-Window Market Reaction to Special Charge
Announcements:

[CAR.sub.it] = a + [b.sub.1] * [(OPERSURPRISE).sub.it] + [b.sub.2]
* [(SPECIAL).sub.it] + [b.sub.3] * [(BOOKMARK).sub.it] + [b.sub.4]
* [(SIZE).sub.it] + [b.sub.5] * [(DEBT).sub.it] + [b.sub.6]
* Dummy Variable + [[epsilon].sub.it]

Model 1 (H3a): Comparison of Write-off Announcements with
Restructuring Charge Announcements (Dummy Variable D1 = 1 if
restructuring charge, 0 otherwise)

Model 2 (H3b): Comparison of Restructuring Charge Pre-Announcement
[Group 2B] with Other Types of Announcements (Dummy Variable D2B = 1
if pre-announced restructuring charge, 0 otherwise)

Model 3 (H3c): Comparison of Write-off Simultaneous Announcement
[Group 1A] with Other Types of Announcements (Dummy Variable D1A = 1
if simultaneously announced write-off, 0 otherwise)

Variable (predicted sign)       Model 1           Model 2

Intercept (?)                 -1.260 (a)         -1.140
  [t-statistic] (p-value)    [-5.30] (0.010)    [-4.90] (0.001)
OPERSURPRISE (+)               0.061              0.061
  [t-statistic] (p-value)     [1.60] (0.120)     [1.57] (0.120)
SPECIAL (+)                    0.240              0.230
  [t-statistic] (p-value)     [3.40] (0.006)     [3.21] (0.001)
BOOKMARK (?)                   0.003              0.002
  [t-statistic] (p-value)     [0.80] (0.440)     [0.71] (0.480)
SIZE (-)                    -0.00014           -0.00008
  [t-statistic] (p-value)    [-0.04] (0.970)    [-0.02] (0.980
DEBT (-)                      -0.064             -0.007
  [t-statistic] (p-value)    [-0.98] (0.330)    [-1.01] (0.310)
Dummy variable (+,+,-)         0.320              0.190
  [t-statistic] (p-value)     [1.93] (0.050)     [1.02] (0.300)

Adjusted [R.sup.2]             1.7%               1.3%

Variable (predicted sign)       Model 3

Intercept (?)                 -0.990
  [t-statistic] (p-value)    [-4.30] (0.001)
OPERSURPRISE (+)               0.060
  [t-statistic] (p-value)     [1.60] (0.120)
SPECIAL (+)                    0.230
  [t-statistic] (p-value)     [3.30] (0.001)
BOOKMARK (?)                   0.002
  [t-statistic] (p-value)     [0.70] (0.480)
SIZE (-)                    -0.00008
  [t-statistic] (p-value)    [-0.03] (0.980)
DEBT (-)                      -0.006
  [t-statistic] (p-value)    [-0.97] (0.330)
Dummy variable (+,+,-)        -0.270
  [t-statistic] (p-value)    [-1.63] (0.100)

Adjusted [R.sup.2]          1.5%

(a) The first number for each variable is its OLS Regression
parameter estimate.

[CAR.sub.i] = firm is market-adjusted return compounded over days 0
and +1. The daily market-adjusted return is computed as the raw return
minus the return on the CRSP equally-weighted index for the day;
OPERSURPRISE = operating performance surprise, measured by the log of
reported operating income [Compustat #178] for the announcement year
minus reported operating income for the previous year divided by total
assets [#6] for the previous year; SPECIAL = special charge amount
[#17], measured by log of reported special items for the announcement
year in Compustat divided by total assets for the previous year;
BOOKMARK = the log of book value [#216] divided by market value
(Prices [#24] multiplied by Number of Shares [#25]) for the previous
year; SIZE = the log of sales (#12) for the previous year divided by
total assets [#6] for the previous year; DEBT = the log of debt equity
ratio (total debt [#181] divided by book value [#216]) for the
previous year.

TABLE 6
SUBSEQUENT MARKET PERFORMANCE OF SPECIAL CHARGE ANNOUNCEMENT
FIRMS (N = 541) (1)

                                    Future Market Performance (2)

                                   Year +1     Year +2     Year +3
Panel A: All firms
  Overall Special Charge
    Announcements                  0.022       0.011       0.0026
      t-value                      0.93        0.44        0.11

Panel B: Group 1 vs. 2
  Write-off Announcements          0.075       0.045       0.039
  Restructuring Charge             -0.007      -0.008      -0.017
    Announcements
     t-value                       1.80 **     1.10        1.10

Panel C: Four Sub-groups
  Group 1A (Write-off              0.071       0.053       0.046
    Simultaneous Announcement)
  Group 1B (Write-off Pre-         0.120       -0.023      -0.029
    Announcement)
  Group 2A (Restructuring Charge   -0.041 **   0.028       -0.053 **
    Simultaneous Announcement)
  Group 2B (Restructuring Charge   0.054       -0.077 **   0.049
    Pre-Announcement)

                                      Future Market
                                     Performance (2)

                                   Year +4     Year +5
Panel A: All firms
  Overall Special Charge
    Announcements                  -0.052      -0.086
      t-value                      -2.25 **    -3.95 ***

Panel B: Group 1 vs. 2
  Write-off Announcements          -0.042      -0.090
  Restructuring Charge             -0.058      -0.084
    Announcements
     t-value                       0.30        -0.20

Panel C: Four Sub-groups
  Group 1A (Write-off              -0.029      -0.110
    Simultaneous Announcement)
  Group 1B (Write-off Pre-         -0.170      0.059
    Announcement)
  Group 2A (Restructuring Charge   -0.052      -0.107
    Simultaneous Announcement)
  Group 2B (Restructuring Charge   -0.069      -0.042
    Pre-Announcement)

(1) The number of sample firms used for future market analysis is
541. Some announcement firms did not survive until year +5.

(2) Future market performance is represented by the sum of future
excess returns. Excess returns are measured by market-adjusted excess
returns (individual returns minus equal weighted market returns). For
example, future market performance from announcement date to year +1
is the sum of excess returns from the fifth day after the announcement
date to that date in the next year.

Note: ** = Significantly different at the 5% alpha level;
*** = Significantly different at the 1% alpha level.

TABLE 7
Subsequent Survival of Special Charge Announcement Firms

                             Announcement Type

                          Write-off       Write-off
                        Simultaneous        Pre-
                        Announcement    Announcement
Survivorship             (Group 1A)      (Group 1 B)

Number of firms that
made special charge
announcements (A)            284             27

Survivors (1) (B)            174             18
Number of lost firms
during the five years
following the
announcements (2)            110              9
(C)=(A)-(B)

Percentage of lost
firms (C) / (A)             (39%)           (33%)

                               Announcement Type

                        Restructuring   Restructuring
                           Charge          Charge
                        Simultaneous        Pre-
                        Announcement    Announcement
Survivorship             (Group 2A)      (Group 2B)     Total

Number of firms that
made special charge
announcements (A)            368             208         887

Survivors (1) (B)            227             122         541
Number of lost firms
during the five years
following the
announcements (2)            141             86          346
(C)=(A)-(B)

Percentage of lost
firms (C) / (A)             (38%)           (41%)       (39%)

(1) Firms that survived during the next five years after making
special charge announcements.

(2) Firms that did not survive for some reason (e.g., bankruptcies,
mergers, acquisitions, etc.) during the subsequent five years
following the announcement.