مدیریت سود و عملکرد بازار دستیابی به شرکت ها

I examine the market's efficiency in processing manipulated accounting reports and provide an explanation for the post-merger underperformance anomaly. I find strong evidence suggesting that acquiring firms overstate their earnings in the quarter preceding a stock swap announcement. I also find evidence of a reversal of the stock price effects of the earnings management in the days leading to the merger announcement. However, the pre-merger reversal is only partial. I find evidence of a post-merger reversal of the stock price effects of the pre-merger earnings management. The results suggest that the extant evidence of post-merger underperformance by acquiring firms is partly attributable to the reversal of the price effects of earnings management. The study also suggests that the post-merger reversal is not fully anticipated by financial analysts in the month immediately following the merger announcement. However, consistent with suggestions in the financial press that managers guide analysts’ forecasts to “beatable” levels, the effect of the earnings management reversal seems to be reflected in the consensus analysts’ forecasts by the time of the subsequent quarterly earnings releases.

This study examines the market's efficiency in processing manipulated accounting reports and provides an explanation for the post-merger underperformance anomaly. Prior studies find that acquirers experience significant losses in the years subsequent to a merger announcement. Jensen and Ruback (1983, p. 20), who review this literature, comment that the “post-outcome negative abnormal returns are unsettling because they are inconsistent with market efficiency and suggest that changes in stock prices during takeovers overestimate the future efficiency gain from mergers.” Through an examination of the effects of earnings management on the performance of acquiring firms, I find that the reversal of the effects of pre-merger earnings management is a significant determinant of both the short-term and the long-term performance of stock-for-stock acquirers. Consistent with Erickson and Wang (1999), I find strong evidence suggesting that acquiring firms overstate their earnings reports in the quarter preceding a stock swap announcement. Erickson and Wang (1999) postulate that the market expects a firm to inflate its earnings prior to a stock swap and, consequently, discounts its stock price at the announcement of the stock swap whether the firm manages earnings or not. Anticipating this market behavior, an acquirer's best response is to manage earnings. Consistent with rational market expectations, Shivakumar (2000) finds a significant negative correlation between a firm's discretionary accruals and the market reaction to the announcement of a seasoned equity offering (SEO). In this study, I find no evidence that the market reaction over the three days around a merger announcement is negatively correlated with acquirers’ abnormal accruals. This apparent contradiction with Shivakumar (2000) likely stems from differences in the process leading to a merger as opposed to an SEO. A merger is the result of negotiations between an acquirer and a target, while an SEO involves only the issuing firm. Therefore, because of the negotiation process between the merging partners that is inherent in a merger, news leakage is more likely for mergers than for SEOs. This is consistent with results by Schwert (1996), for instance, who shows significant run-ups in targets’ prices as far back as one month (21 trading days) before a merger announcement.1 Hence, in a merger, the stock price adjustment is likely to start weeks before the merger announcement during the rumor phase of the merger. To accommodate this eventuality, I extend the event window one month prior to the merger announcement. Consistent with the rational expectations hypothesis, the correlation between the abnormal accruals and the abnormal returns of the acquiring firms that engage in stock swaps becomes significantly negative. Shivakumar (2000) finds no significant correlation between the discretionary accruals and issuers’ long-term performance when performance is measured using a matched-firm control sample. He concludes that the market completely undoes the effects of earnings management at the announcement of an SEO. However, Shivakumar's results may not be generalizable. To completely undo the effects of earnings management, it seems that investors would need to observe managers’ actions or, at least, fully understand managers’ opportunity sets and all the intricacies of earnings management. Because managers have so many ways to manipulate earnings, it is unlikely that the market would know exactly how much they have inflated their reports. Existing anecdotal evidence suggests that even the most sophisticated investors are fooled by earnings management, even when managers’ incentives to mislead the market are evident.2 Furthermore, if investors could completely undo earnings management, they would probably have enough incentives to do so for all firms, regardless of whether the firms are issuing stocks or not. Once investors determine the exact amount of earnings management, the signal about the direction of the earnings management provided by the stock offerings would be uninformative. In the case of a stock-for-stock merger, the managers of the target firms should have enough incentives and expertise to detect earnings management by acquirers. However, Shleifer and Vishny (2003) suggest that the managers of a target firm could agree to a stock merger even if they know that the acquirer's stock is overvalued. The managers of the target will agree to merge for reasons of retirement or illiquid stock ownership.3 The acquirer can also buy their agreement through the acceleration in the exercise of stock options, by granting them generous severance pay, or by keeping them in top positions. In addition, the target's managers can simply be duped into accepting the acquirer's inflated stocks. In a lawsuit against Walter A. Forbes and E. Kirk Shelton, the Securities and Exchange Commission accuses the top two former officers of CUC International Inc. (CUC) of undertaking “a program of mergers and acquisitions on behalf of CUC in order to generate inflated merger and purchase reserve.” The suit contends: “To entice HFS management into … [a] merger, Forbes and Shelton inflated CUC's earnings and earnings projections. Soon after CUC merged with HFS to create Cendant, Forbes and Shelton explicitly congratulated each other on being masterful ‘financial engineers’ … [for] duping HFS into agreeing to a merger with CUC.”4 Shivakumar (2000, p. 340) posits that, “since issuers cannot credibly signal the absence of earnings management, investors treat all firms announcing an offering as having overstated prior earnings, and consequently discount their stock prices.” This assumes that the market's estimation of earnings management is unbiased and implies that the announcement of a stock issuance will lead to a mean stock price decline as investors correct for the average earnings management. Errors in the market's assessment of earnings management will reverse in the subsequent years when the pre-issuance earnings management reverses or the managers deflate the market's expectations.5 Therefore, even if the market's assessment of earnings management is unbiased, the post-merger performance of the stock-for-stock acquirers will very likely have a negative correlation with the degree of the pre-merger earnings management. To illustrate my point, assume that, through earnings management, Firm A and Firm B inflate their values by 6% and 14%, respectively. Investors do not observe the managers’ actions. However, for the sake of argument, assume that the market forms unbiased estimates (on average) of the firms’ earnings management and discounts the value of each firm by 10% (the average percentage stock price inflation) at the merger announcement. Under this scenario, the abnormal return at the merger announcement (AR) and the pre-merger earnings management (EM) will be uncorrelated. There will only be a mean effect captured by the intercept of the regression of AR on EM. After the merger, when the pre-merger earnings management reverses or the managers deflate the market's expectations, everything else equal, the value of Firm A should increase by 4% and the value of Firm B should decrease by 4%, implying a negative correlation between earnings management and the post-merger performance. Hence, even under the strong assumption that the market's assessment of earnings management is correct on average (which seems unlikely given the anecdotal evidence), as long as the market does not know exactly the extent of pre-merger earnings management by each acquirer, a negative correlation should be observed between earnings management and post-merger performance. I find that stock-for-stock acquirers fare worse than cash acquirers over the three years subsequent to the merger announcement. More important, I find a significant negative correlation between the abnormal accruals and the long-term stock performance for the stock-for-stock acquirers. I also compare the acquirers’ reported earnings with the analysts’ earnings forecasts in the year subsequent to the merger. The results suggest that the reversal of the effects of the pre-merger earnings management by stock-for-stock acquirers is not fully anticipated by financial analysts in the month immediately following the merger announcement. However, I find that by the time of the subsequent quarterly earnings releases, the effect of the earnings management reversal is reflected in the consensus analysts’ forecasts. My findings are distinct from those reported by Sloan (1996), who finds a significant inverse relation between accruals and long-term returns. Sloan uses all firms that have necessary data on Compustat and the Center for Research in Security Prices (CRSP). However, consistent with Chen and Cheng (2004), I find evidence indicating that the accrual effect is not general; it is associated with incentives related to the accruals. Sloan (1996) estimates a univariate regression of raw return on accruals and a multivariate regression where he controls for beta, log market value, earnings-to-price, and book-to-market. Using similar models, I find that the coefficient on the accrual variable is negative for the stock-for-stock acquirers and positive for the cash acquirers.6 The balance of the paper is organized as follows. Section 2 describes the research design. The sample selection process is discussed in Section 3. The results are reported in Section 4. The study concludes in Section 5.

The results of the study are presented below. They include descriptive statistics, estimates of the acquirers’ abnormal returns around and after merger announcements, estimates of earnings management by the acquirers prior to merger announcements, the association between the acquirers’ abnormal returns and earnings management, estimates of analysts’ earnings forecast errors around merger announcements, and some sensitivity analyses.