The value relevance of corporate restructuring charges

We also re-evaluated investors’ reaction to the restructuring announcements using the events study methodology. We evaluated investors’ reaction to 3-days, 5-days, 11-days and 21-days windows. Our results did not show any significant cumulative average returns (CAR) during these windows. The CARs for the 31-days window, however, became significantly positive. When the variable of performance enhancing restructuring charge is included in the short-window regression, the coefficient is not significantly positive. Instead it is insignificantly different from zero. These results suggest that investors are unable to correctly anticipate the full effect of restructurings during the announcement period.

Detailed information on the components of restructuring charges is usually provided in the annual reports. A press release of a restructuring plan typically provides reasons and scope, and sometimes an estimated amount of the restructuring charges. Information, such as precise amount and disaggregated components of the restructuring charges, is usually not available on the announcement date. Prior studies (e.g. Chaney et al. 2000; Kross et al. 2000) assume that component information is available at the restructuring announcement date.

We also use alternate metrics, such as the market adjusted returns and the CAPM based returns, as a robustness check. From a methodological standpoint, our metrics are an improvement over the simple market adjusted raw returns metric used by Khurana and Lippincott (2000).

The quantile regression was suggested by the editor.

Emerging Issues Task Force released EITF 94-3, “Liability Recognition for Certain Employee Termination Benefits and Other Costs to Exit an Activity (Including Certain Costs incurred in a Restructuring)”, in 1994, and set forth the financial reporting and disclosure requirements for a plan of restructuring. EITF94-3 provides a review and gives examples of individual costs typically included in restructurings as well as guidelines on the timing and measurement of the charges associated with the restructuring plan. EITF94-3 was superseded by SFAS No. 146 in June 2002.

Atiase et al. (2004) have recently examined the association between restructuring and the firm’s operating performance. They conclude that restructuring charges are associated with improved operating performance, but may not necessarily guarantee such improvement.

Details on identifying successful and unsuccessful restructurings are given in the research methodology section.

Investors’ reaction to the magnitude of restructuring charges has especially been pointed by the reviewer.

We also use industry-adjusted return on assets (ROA) as the performance metric and obtain qualitatively similar results.

We realize that some firms in the group of unsuccessful restructuring may improve their future operating performance beyond 2 years after restructuring. However, we do not study the period over 2 years because the operating performance over a longer period is likely to be affected by events other than restructuring, which may confound our results. Nevertheless, we acknowledge this limitation of the study.

As a robustness check, we use two other metrics, namely the market adjusted returns and the CAPM based metric. The results are similar to those based on the Fama–French based metric, which are presented in Table 6.

As mentioned earlier, we also conduct tests using two additional types of buy-and-hold abnormal returns. In the first case, we compute twelve-month buy-and-hold abnormal returns using monthly market-adjusted returns (MAR) by adjusting the raw returns with market returns, i.e. MAR _it   = R _it  − R _mt . In the second case, we obtain buy-and-hold abnormal returns by cumulating using monthly market-adjusted returns obtained from the CAPM model (CAPMAR), i.e. CAPMAR _it  = R _it  − R _ft  − β _i [R _mt  − R _ft ]. Here β _i was obtained by regressing (R _it  − R _ft ) on (R _mt  − R _ft ) for the 24-month estimation period prior to the restructuring fiscal year.

Since RCHG is reported on the pretax basis, to be consistent, we use pretax earnings and earnings change in the regression. Our findings do not change even if we use the after-tax RCHG by assuming a 35% corporate tax rate.

Several studies use this method to separate non-recurring charges from earnings. For example, Khurana and Lippincott (2000) decompose operating earnings in the same way to study the value relevance of restructuring charges in the earnings-returns regression. Elliott and Hanna (1996) decompose earnings from continuing operations into earnings before write-offs and write-offs to examine the earnings response coefficients of multiple write-offs. Bradshaw and Sloan (2002) find that informativeness of earnings improves after isolating special charges from earnings.

We add control variables of the firm size and growth, when CAPMAR is used to compute the cumulated abnormal returns. Size is the logarithmic transformation of the fiscal year-end market value of equity; and growth is the sum of market value of equity and the book value of debt scaled by the book value of total assets. In the model when MAR is used, we also include an additional variable, beta, to control for systematic risk.

The EITF96-9 issue discusses whether inventory markdowns associated with an exit plan or a restructuring activity should be classified in the income statement as a cost of goods sold or as an exit or a restructuring cost. The Task Force members agreed that, regardless of how the inventory markdowns are classified, it may be appropriate to disclose the amount of inventory markdowns associated with an exit plan or a restructuring activity. The SEC staff also believed that inventory markdowns should be classified in the income statement as a component of cost of goods sold.

The Task Force members also considered the following costs qualified as restructuring costs: costs for new systems development or acquisition, job retraining for terminated employees, moving expenses for terminated employees, the incremental cost of subcontracting future warranty work on products sold before and after the commitment date, and so on.

Prior to 2001, restructuring charge is not separately identified from special charges in Compustat and the collection of restructuring charge and its components requires manual collection. We restrict the initial sample to firms reporting negative special charges to obtain a smaller initial sample to start with. This restriction is reasonable because most restructuring companies have a negative special charge. However, under this method firms with positive special charges but has a restructuring charge were not identifiable. Lopez (2002) also imposed this restriction in his data selection process.

The extreme values are defined as stock prices less than $1.00 as of the beginning of the fiscal year and those whose earnings per share from continuing operations scaled by P _t−1 are among the bottom and top three percentile of the sample.

As discussed the section of Research Design, facilities closing and consolidations (FC) often involve with workforce reduction (WFR) as well as asset write-downs and write-offs (AWD), which reduces the number of firms that report a significant amount of facility closing costs.

The bottom and top one percent of dependent variable, BHR, and independent variables, E ^adj and RCHG, are winsorized.

We also included a dummy variable of PI in the regression analysis, and we find that the results on PI_RCHG and NPI_RCHG remain unchanged. The coefficient for the dummy PI variable, however, is negative, suggesting that the negative value of intercept increases, which means that investor reaction to restructuring is negative, but it becomes positive when restructuring is performance enhancing.

We obtain similar results when include the performance-enhancing indicator variable in the equations.

We also conducted an additional test on the total sample with a dummy variable of magnitude of restructuring charges. The results based on this test are consistent with separate analyses on the two groups, as reported in Table 6.

The results on the BHR (market adjusted), consistent with the results reported in the Panel A of Table 7, show that the coefficient for the performance-improving profit firms is insignificant.

The results from the quantile regression, while consistent with the results from the OLS regression, also show that the uppermost quantile and the lowermost quantile are somewhat different from the other quantiles.