Target Financial Reporting Quality and M&a Deals That Go Bust

Target Financial Reporting Quality and M&A Deals that Go Bust*
HOLLIS A. SKAIFE, University of Wisconsin–Madison DANIEL D. WANGERIN, Michigan State University

1. Introduction This study investigates whether target ﬁrms’ ﬁnancial reporting quality affects the likelihood that merger and acquisition (M&A) deals will ultimately be terminated. Managers looking to increase their market share, enter new markets, or diversify their operations will consider acquiring another company based on the company’s performance, geographic locations, and lines of business, respectively. If the potential target is a U.S. publicly traded company, an acquirer’s initial assessment of the expected beneﬁts associated with the acquisition of the company is based on publicly available information. Generally, the acquirer obtains limited private information from the target prior to the signing of the acquisition agreement. Although an acquisition agreement creates a binding contractual obligation for both entities to go forward with the deal, it does not guarantee completion of the deal. The acquisition agreement typically contains a warranty by the target that its ﬁnancial statements are prepared in accordance with generally accepted accounting principles (GAAP). If this warranty is breached, the deal can be terminated. We hypothesize that low-quality ﬁnancial reporting by target ﬁrms prior to the announcement of a deal increases the likelihood that a target ﬁrm’s U.S. GAAP warranties stated in the acquisition agreement are breached. Therefore, we predict that deals involving targets with low-quality ﬁnancial reporting are more likely to be terminated (i.e., go bust). Based on prior research, we identify ﬁve measures of low-quality ﬁnancial reporting: the magnitude of discretionary accruals (Dechow, Sloan, and Sweeney 1995); the likelihood of a weakness in internal control (Doyle, Ge, and McVay 2007; Ashbaugh-Skaife, Collins, and Kinney 2007), off-balance-sheet liabilities (Barth 1991); analysts’ forecast error (Lang and Lundholm 1996); and analysts’ forecast dispersion (Barron, Kim, Lim, and Stevens 1998). The ﬁrst three measures are intended to capture noise and ⁄ or bias in ﬁnancial reporting that affects the relevance and representational faithfulness of the target’s ﬁnancial statements. Greater discretionary accruals, greater likelihood of internal control problems, and more off-balance-sheet liabilities are indicators of less reliable, less relevant, low-quality ﬁnancial reporting. The last two measures are used to assess the precision of the target’s ﬁnancial information where greater analysts’ forecast error and greater analysts’ forecast dispersion signal less precise, less certain, low-quality ﬁnancial reporting related to targets’ performance and operations. We combine these measures to construct a low-quality ﬁnancial reporting (LQFR) score and use this score in our empirical tests examining the M&A market consequences of targets’ ﬁnancial reporting quality.
* Accepted by Dan Segal. We appreciate the comments and suggestions of Dan Segal, an anonymous reviewer, Xia Chen, Jeremiah Green, Antonio Macias, David Veenman, Terry Warﬁeld, participants at the 2010 AAA annual meeting, and workshop participants at the University of Amsterdam, University of California–Irvine, University of Edinburgh, Santa Clara University, and the University of Wisconsin– Madison.

Contemporary Accounting Research Vol. 30 No. 2 (Summer 2013) pp. 719–749 Ó CAAA doi:10.1111/j.1911-3846.2012.01172.x

720

Contemporary Accounting Research

Acquirers rely extensively on targets’ ﬁnancial reporting when negotiating the purchase price with the target. LQFR contributes to greater uncertainty about a ﬁrm’s expected future earnings, creating greater uncertainty about the expected net future cash ﬂows related to the acquisition as well as potentially greater operating risk for the acquirer that increases the acquirer’s cost of capital. Both of these consequences of targets’ LQFR suggest acquirers will discount transactions involving LQFR targets resulting in lower, if any, deal premiums. On the other hand, LQFR target ﬁrms’ shares could trade at a discount because their ﬁnancial reporting is less reliable (see, e.g., Ashbaugh-Skaife, Collins, Kinney, and LaFond 2009), resulting in higher deal premiums. An acquirer offering a premium to a LQFR target will face stronger incentives to terminate the deal once it obtains access to the target’s private information and, as a result, gains a better understanding of the target’s value. Therefore, we begin our analysis by examining the relation between takeover premiums and targets’ ﬁnancial reporting quality. Corroborating prior research, we ﬁnd takeover premiums are greater when a deal involves multiple bidders, is structured as a tender offer, uses all cash consideration, the acquirer and target operate in the same industries, and the acquirer is a public company. Consistent with prior research, we also ﬁnd that premiums are lower when the acquirer uses stock-based consideration. After controlling for these deal characteristics, we ﬁnd that premiums are greater for deals involving targets with LQFR, and this ﬁnding holds after controlling for target ﬁrms’ operating and ﬁnancial characteristics linked in prior research to M&A takeovers, as well as ﬁrms’ ﬁnancial reporting quality. This ﬁnding suggests that based on targets’ publicly available ﬁnancial information, acquirers perceive additional value beyond that priced by investors in LQFR targets at the time the acquisition agreement is signed. However, the ofﬁcial offer via the signing of the acquisition agreement does not guarantee the deal will go through. Once the acquisition agreement is signed, the acquirer begins transactional due diligence. Transactional due diligence allows the acquirer more extensive access to private information about a target’s operating performance and ﬁnancial position by gaining rights to the target’s ﬁnancial records (e.g., accounting estimates, revenue recognition policies, signiﬁcant valuation accruals) and contracts (e.g., lease agreements). A crucial aspect of the acquirer’s due diligence process entails veriﬁcation of the target’s representations and warranties made in the acquisition agreement. As stated above, a key warranty made by publicly traded targets is that their ﬁnancial statements are prepared in accordance with GAAP. If the acquirer discovers during transactional due diligence a breach in the target’s warranty that its ﬁnancial statements are prepared in accordance with GAAP, then the deal can be terminated. We ﬁnd hostile deals, deals with multiple bidders, and deals where the acquiring ﬁrm has a minority interest (i.e., a toehold) prior to the announcement of a majority share acquisition are more likely to be terminated. In contrast, deals where the offer is made directly to target shareholders (i.e., tender offers) and where the target and acquirer operate in the same industries are more likely to be completed. After controlling for these deal characteristics, we ﬁnd deals involving LQFR targets are more likely to be terminated, with the marginal effect of LQFR increasing the likelihood of termination more than 9 percent. The result is robust to excluding deals involving multiple bidders, controlling for targets’ operating and ﬁnancing characteristics, and overbidding by the acquiring ﬁrm. We initially model the outcome of the M&A as binary; the deal is either completed or terminated. However, if by gaining access to private information via transactional due diligence the acquirer realizes that there is greater risk to the takeover of a LQFR target, then the acquirer might renegotiate the offer in an attempt to reduce the purchase price rather than terminating the deal altogether. Consistent with this conjecture, we ﬁnd a signiﬁcantly positive relation between targets’ LQFR and deal renegotiation as well as termination,
CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

721

highlighting another important consequence of targets’ ﬁnancial reporting quality in M&A. Targets’ LQFR can not only contribute to M&A deals being renegotiated or ultimately terminated, but also to publicly traded targets issuing restated ﬁnancial statements. Upon further analysis, we ﬁnd that targets involved in failed deals are more likely to restate their ﬁnancial statements soon after the deal goes bust relative to (1) other ﬁrms in the market and (2) ﬁrms with similar levels of ﬁnancial reporting quality as measured by our low-quality ﬁnancial reporting score. The number one reason failed targets restate their ﬁnancial statements relates to the accounting for liabilities associated with leases, contingencies, and expense recognition. Improper asset valuation is the second most common reason for failed targets’ restatements. The results of our restatement analysis provide insights into the speciﬁc violations of GAAP that are reﬂected in our LQFR score and that contribute to M&A deals going bust.1 Our study makes several contributions to the literature. The work of Erickson and Wang 1999 suggests that part of the wealth losses of acquiring ﬁrms is due to preacquisition earnings management activities by acquiring ﬁrm managers in stock-for-stock mergers.2 Contemporaneous research provides evidence that the quality of target ﬁrms’ ﬁnancial information is also associated with the wealth losses of acquiring ﬁrms in completed deals (Anilowski, Macias, and Sanchez 2009; McNichols and Stubben 2012). Our study complements these studies by examining the role of target ﬁrms’ ﬁnancial reporting in all M&A deals — completed and terminated — and provides evidence that targets’ LQFR directly relates to acquirers’ deal premiums, deal renegotiation, and the probability that the M&A deal falls through. Thus, our research provides new insights into the capital market consequences of ﬁnancial reporting quality. Our study also contributes to the literature documenting the determinants and consequences of ﬁnancial reporting restatements, as our study is the ﬁrst to explore ﬁnancial reporting restatements in the M&A setting. We provide evidence that suggests that the scrutiny of transactional due diligence intended to uncover unrecorded liabilities or overvalued assets results in failed targets more often restating their ﬁnancial statements. To date, academic literature has given little attention to transactional due diligence, although its importance is widely acknowledged by the investment community. Moreover, we construct a comprehensive measure of LQFR that encompasses the relevance, reliability, and precision of publicly available ﬁnancial information, and demonstrate that this measure is useful in understanding the market for corporate control and in predicting ﬁnancial statement restatements. The rest of the paper is organized as follows. Section 2 provides an overview of the role of ﬁnancial reporting and due diligence in M&A. Section 3 develops our proxy of LQFR. Section 4 describes our sample and data sources, and provides descriptive statistics. Section 5 presents the results of our empirical tests. Section 6 concludes and offers suggestions for future research.

1.

2.

Detection of an accounting misstatement as a result of transactional due diligence does not necessarily result in the target having to restate its ﬁnancial statements. This is because the information uncovered via transactional due diligence may not be considered material enough to warrant a restatement, but is material enough for the deal to be terminated. It is the judgment of the target’s management, board of directors, and auditor whether amended ﬁnancial statements are ﬁled with the SEC. This fact biases against ﬁnding any difference in ﬁnancial statement restatements between failed targets and other ﬁrms. Erickson and Wang (1999) fail to ﬁnd evidence suggesting that target ﬁrms, in contrast to acquiring ﬁrms, engage in signiﬁcant preacquisition earnings management behavior. Although targets have similar incentives to manage earnings upwards prior to an acquisition, targets have difﬁculty anticipating if and when a bid from an acquiring ﬁrm will be received.

CAR Vol. 30 No. 2 (Summer 2013)

722

Contemporary Accounting Research

2. Overview of due diligence in M&A The timeline presented in Figure 1 illustrates the due diligence process in M&A. In the initial stage of due diligence, referred to as preliminary due diligence, an acquirer relies on publicly available ﬁnancial statements about potential targets when targets are U.S. Securities and Exchange Commission (SEC) registrants. Public information about a potential target also comes from other SEC ﬁlings (e.g., Form 4 that reports changes in ownership by top management and directors), analyst reports, articles in the business press, industry summaries, and nonﬁnancial information disseminated through other means (e.g., material on corporate websites, product brochures, etc.). Managers of acquiring ﬁrms become more informed about potential targets during preliminary due diligence, but are no more informed than any other market participant who relies solely on ﬁrms’ publicly available information. Once preliminary due diligence is complete, the acquirer and target commit to negotiating a deal by signing a conﬁdentiality agreement. In a conﬁdentiality agreement, the target promises to provide the acquirer with limited access to private information in exchange for the acquirer’s promise to keep conﬁdential any information obtained. The signing of a conﬁdentiality agreement, which is not publicly disclosed, allows the acquiring ﬁrm to begin its due diligence review. The due diligence review typically includes an examination of management reports and projections, as well as investigation of any research and development projects. The acquirer’s due diligence review team also conducts interviews of key target employees and engages in site visits to gather additional private information. The acquirer incorporates the limited private information obtained in the due diligence review to update its initial valuation of the target formed on the basis of public information, decides whether to make a formal offer for the target, and, if so, determines the amount of the bid. The signing of the acquisition agreement signals that the target has accepted the acquiring ﬁrm’s bid. If the target ﬁrm is a U.S. SEC registrant, the deal is required to be disclosed via Form 8-K. More importantly for our study, once the acquisition agreement is signed, the acquirer begins its transactional due diligence. The acquisition agreement

Figure 1

The M&A due diligence process. Note. Adapted from Wangerin 2012. before acquisition agreement after acquisition agreement

Transaction timeline

confidentiality agreement signed

acquisition agreement signed

deal completed or terminated

Due diligence timeline

preliminary due diligence

due diligence review

transactional due diligence

acquirer evaulates M&A target candidates based on public information and negotiations begin

acquirer obtains limited private information from target firm and negotiates terms of the acquisition agreement

acquirer gains access to target firm's financial records tests for accuracy of target's warranties and representations that its financial statements are prepared in accordance with GAAP

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

723

includes covenants, representations, warranties, and conditions to closing that are critical to the completion of the deal. Transactional due diligence is intended to verify whether the representations and warranties contained in the acquisition agreement are true. In other words, transactional due diligence gives the acquirer a ﬁnal opportunity to verify exactly what it is buying before assuming the risks of ownership. In this ﬁnal stage of the due diligence process, the acquiring ﬁrm veriﬁes the accuracy of the information provided by the target including whether the target’s ﬁnancial statements are prepared in accordance with GAAP. Breaches of representations and warranties by the target give the acquirer the right to terminate a deal after the acquisition agreement is signed.3 We posit that targets with LQFR are more likely to breach the representations and warranties made in the acquisition agreement that their ﬁnancial statements are in compliance with GAAP and therefore are more likely to be involved in deals that are terminated. 3. Low-quality ﬁnancial reporting (LQFR) score There is a rich literature examining the characteristics of publicly traded ﬁrms’ ﬁnancial reporting. Much of this literature focuses on managers’ incentives to introduce a bias in ﬁnancial reporting to meet ﬁnancial reporting objectives (e.g., see Healy and Wahlen 1999 for an early review of the earnings management literature). Recently, more attention is given to the unintentional errors that are introduced into ﬁnancial reporting and the consequences this has for ﬁrms’ information environments (e.g., see Lambert, Leuz, and Verrecchia 2007 for theoretical arguments on how unintentional errors in ﬁnancial reporting affect ﬁrms’ information environment). We develop a comprehensive measure of LQFR that captures both types of ﬁnancial reporting problems. Speciﬁcally, the ﬁrst three measures of LQFR consist of discretionary accruals, likelihood of internal control problems, and off-balance-sheet liabilities. These three measures are intended to capture noise and ⁄ or bias in ﬁnancial reporting that affect the relevance and representational faithfulness of the target’s ﬁnancial statements. The last two measures included in LQFR are used to assess the precision of the target’s ﬁnancial information where greater analysts’ forecast error and greater analysts’ forecast dispersion signal less precise, less certain, low-quality ﬁnancial reporting related to targets’ performance and operations. By using an LQFR measure consisting of multiple properties to test our hypothesis, we diminish the concern that results are driven by a measurement bias inherent in one particular ﬁnancial reporting quality proxy (McNichols 2003). Appendix 1 provides details on our LQFR score. The magnitude of discretionary total accruals is the ﬁrst measure included in the LQFR score. We use total accruals rather than current accruals because total accruals have a more meaningful economic impact on the assessment of short- and long-term target performance. By using the magnitude of discretionary accruals, we allow target’s ﬁnancial reporting to be of low quality due to unintentional errors or earnings management. Unintentional errors result in more noise in the ﬁnancial statements, whereas earnings management results in biased ﬁnancial statements. For our study, both types of errors result in ﬁnancial statements being less representative of the target’s underlying performance. We employ performance-matched discretionary accruals because this procedure allows us to simultaneously control for the effect of the relation between a ﬁrm’s accruals and its
3. In an attempt to provide insights into whether it is the acquirer or the target that walks away from the deal, we search for target and acquiring ﬁrms’ 8-Ks ﬁled on the date that the deal is known to be terminated for a sample of terminated deals (n = 25). The language of the 8-K, as well as the corresponding press release accompanying the 8-K, provides no clear statements as to which party canceled the deal or reason for termination. The lack of transparency of the 8-K ﬁling comes as no surprise, as acquirers are not able to disclose private information obtained during due diligence under the terms of conﬁdentiality agreements with the target. The opacity of these disclosures adds to the contribution of our restatement analysis discussed below.

CAR Vol. 30 No. 2 (Summer 2013)

724

Contemporary Accounting Research

operating performance (Kothari, Leone, and Wasley 2005). The speciﬁc steps we take in calculating the magnitude of discretionary total accruals (|AAC|) are reported in panel A of Appendix 1. LQFR is increasing in the magnitude of discretionary accruals. Our second indicator of LQFR is the probability of having a material weakness in internal control. Recent research suggests ﬁrms with ineffective internal control over ﬁnancial reporting have less reliable ﬁnancial information. Ashbaugh-Skaife, Collins, and Kinney (2007) and Doyle et al. (2007) model the likelihood of having an internal control problem as a function of operational complexity, organizational change, accounting application risk, and ﬁrm resources that can be dedicated to internal control. Speciﬁcally, the likelihood of weaknesses in internal control is increasing in the number of segments, existence of foreign sales, sales growth, inventory intensity, incidence of net losses, bankruptcy risk, resignations of the independent auditor, and whether the ﬁrm engaged in an M&A or restructuring activities. The likelihood of weaknesses in internal control is decreasing in ﬁrm size. The probability of having a material weakness in internal control, denoted as WIC, is computed for each target ﬁrm using the parameters reported in Ashbaugh-Skaife et al. 2007. We use an estimate of internal control problems because the public reporting of internal control weaknesses is not available for all years of our analysis and only ﬁrms qualifying as accelerated ﬁlers were required to report on their internal control under SOX 404 during the latter years of our analysis. LQFR is increasing in the likelihood of having a material weakness in internal control.4 Off-balance-sheet assets and liabilities diminish the valuation usefulness of ﬁrms’ ﬁnancial statements because the ﬁnancial statements do not capture the resources and claims to resources that affect ﬁrms’ source and use of cash. We compute an estimate of a ﬁrm’s off-balance-sheet liabilities (OFFBS_LIAB) because unrecorded or undervalued liabilities pose a greater cost to the deal from the acquirers’ perspective and include this measure as our third attribute of LQFR. Based on Barth 1991, OFFBS_LIAB is calculated as follows. First, the residual from cross-sectional industry-year regressions of stock prices on assets and liabilities estimated at the 3-digit, 2-digit, and 1-digit SIC code levels requiring a minimum of 20 ﬁrms is estimated. Negative residuals represent off-balance-sheet net liabilities and are transformed by taking their absolute value so that larger values are interpreted as greater off-balance-sheet liabilities. Positive residuals are set equal to zero. LQFR is increasing in OFFBS_LIAB.5 We incorporate two analyst-based proxies for ﬁnancial reporting quality into our LQFR score. Prior research indicates that greater analyst forecast error and greater analyst forecast dispersion reﬂect less precise ﬁnancial information available to capital market participants who incorporate ﬁnancial information into their resource allocation decisions. |AFE| denotes the absolute value of analyst forecast error in the year immediately prior to the acquisition, where analyst forecast error is the difference between the actual reported earnings per share and the median analyst estimate eight months prior to ﬁscal year-end. DISP denotes analyst dispersion, which is computed as the standard deviation
4. We conduct two robustness checks to see if our results are sensitive to this research design choice. First, we use only ﬁrm-year observations post 2003 and replace WIC with a dummy variable equal to one when the target ﬁrm reports a material weakness in internal control under SOX 404 and zero otherwise. This test yields quantitatively similar results. Second, we reestimate the parameters from the Ashbaugh-Skaife et al. 2007 model using ﬁrm-year observations from 2004–2008 and substitute these values to calculate WIC. Again, the results are quantitatively similar. We also compute |AAC| and OFFBS_LIAB using the Fama-French 48-industry classiﬁcation rather than 3digit, 2-digit, and 1-digit SIC industry membership. When using the Fama-French 48-industry classiﬁcation to construct our LQFR score, we ﬁnd the results of our deal premium, termination, and renegotiation analyses continue to hold.

5.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

725

of analyst forecasts eight months prior to the ﬁscal year-end. Both measures are deﬂated by previous year end-of-period stock price and require a minimum of three analyst forecasts. LQFR is increasing in analyst forecast error (|AFE|) and analyst forecast dispersion (DISP). LQFR_SCORE is calculated as the average of the decile ranks of the ﬁve ﬁnancial reporting quality variables. The average is computed as the sum of the ﬁrm’s decile ranks of |AAC|, WIC, OFFBS_LIAB, |AFE|, and DISP divided by the number of ﬁnancial reporting quality variables with nonmissing data. We do not require targets to have all ﬁve proxies of ﬁnancial reporting quality available to calculate LQFR_SCORE in order to obtain the most comprehensive sample possible. Consequently, employing LQFR_SCORE as a proxy for ﬁnancial reporting quality allows our analysis to be the most representative of the public target M&A market, enhancing the external validity of our ﬁndings. In addition, because the denominator of LQFR_SCORE is a function of the number of dimensions of a target’s ﬁnancial reporting quality that is publicly available, LQFR_SCORE captures the differences in target ﬁrms’ information environments. Thus, using LQFR_SCORE rather than a single dimension of ﬁnancial reporting quality increases the internal validity of our study. Panel B of Appendix 1 reports the descriptive statistics on the ﬁve measures of ﬁnancial reporting quality as well as our LQFR score. It is clear from the descriptive statistics that sample sizes vary, conditional on the ﬁnancial reporting quality measure employed. More importantly, the univariate tests provide initial evidence that there are signiﬁcant differences in ﬁnancial reporting quality between targets of terminated deals versus targets of completed deals. The descriptive statistics indicate that the mean value of |AAC| is higher for targets of failed bids relative to targets that are taken over, although the difference is not statistically signiﬁcant.6 The propensity for having internal control problems is greater for targets of deals that go bust (mean = 0.5119; median = 0.5362) than targets of completed deals (mean = 0.4482; median = 0.3381). Median OFFBS_LIAB is also signiﬁcantly larger for failed bids (0.0692) relative to completed bids (0.0120). The dispersion of analyst forecasts is also greater when the takeover bid is withdrawn as the mean DISP for targets in the terminated sample is 0.0087 and 0.0051 for completed bids. Overall, the signiﬁcant differences in the mean and median LQFR_SCORE between failed targets (mean = 5.6689; median = 5.5500) and targets of completed deals (mean = 5.1449; median = 5.000) provides initial evidence that low-quality ﬁnancial reporting plays a role in M&A deals that go bust. Panel C of Appendix 1 displays the Pearson correlation coefﬁcients for the ﬁnancial reporting quality measures. All of the correlations between the ﬁnancial reporting quality proxies are signiﬁcant at the 0.10 level or better. The magnitudes of the correlations are well below 1.00, indicating that the alternative measures of ﬁnancial reporting quality pick up distinct attributes of ﬁrms’ ﬁnancial reporting quality. By construction, correlations between LQFR_SCORE and the other ﬁve information risk proxies are relatively high, ranging between 0.548 for WIC to 0.329 for |AFE|. However, the narrow range of the correlations between LQFR_SCORE and the individual proxies suggests LQFR_SCORE is a comprehensive measure of ﬁnancial reporting quality that is not driven by any one measure.

6.

While the mean values of |AAC| are not statistically different, untabulated statistics indicate that targets in terminated deals report signiﬁcantly more positive abnormal accruals than targets of completed deals. To the extent that all target ﬁrms have similar incentives to manage earnings upward in order to extract a higher bid from the acquirer, we conduct a sensitivity test using only positive abnormal accruals as an indicator of low ﬁnancial reporting quality in the calculation of LQFR_SCORE. We draw similar conclusions.

CAR Vol. 30 No. 2 (Summer 2013)

726

Contemporary Accounting Research

4. Research design Sample The sample construction is summarized in panel A of Table 1. Our sample originates from the SDC Platinum Mergers & Acquisitions Database and includes all deals settled between 2002 and 2008. The sample includes only takeover bids for U.S. publicly traded ﬁrms (i.e., public targets) because we require ﬁnancial reports to construct the LQFR score.7 We also require stock acquisitions to be greater than 50 percent of targets’ outstanding shares to be consistent with prior research (Hsieh and Walkling 2005). We start our analysis in year 2002 because prior to that time M&A deals could be accounted for using either the pooling of interest or purchase method and ﬁrms were not subject to the additional reporting and trading requirements instilled by the Sarbanes-Oxley Act of 2002. Both of these changes potentially have an effect on the frequency and terms of M&A deals.8 We end our analysis in year 2008 so that we can identify the settlement date for the M&A deals comprising our sample. These restrictions result in an initial sample of 3,239 deals, of which 2,772 are completed and 467 are terminated. We use COMPUSTAT as the primary source of ﬁnancial statement data, and this requirement reduces the sample by 1,601 observations lacking the data to conduct our empirical tests. This results in a sample of 1,638 observations with complete data for our deal termination analysis of which 234 are terminated. Finally, SDC does not report data necessary to compute premiums for 170 observations, so 1,468 observations are used in the deal premium analysis of which 1,263 transactions are completed and 205 are terminated. Panel B of Table 1 presents the number of deals settled each year partitioned by whether a deal is terminated or completed. The composition of the sample by settlement year is shown based on the reduced sample for our deal premium analysis. The distribution of the sample across years is relatively constant with the exception of 2008, which has the lowest number of settled deals at 66. This is because the 2008 subsample represents deals announced prior to but settled during 2008. Methodology Prior literature demonstrates that deal premiums are associated with various transactionspeciﬁc and target-speciﬁc characteristics, so we investigate the relation between LQFR_SCORE and deal premiums using the following OLS regression: PREMIUM ¼ b0 þ b1 LQFR SCORE þ b2 HOSTILE þ b3 MULTIBID þ b4 TOEHOLD þ b5 TENDER þ b6 INDUSTRY þ b7 STOCK þ b8 CASH þ b9 PUBLICBIDDER þ b10 SIZE þ b11 rCFO þ b12 GROWTH þ b13 ROE þ b14 MTB þ b15 LEVERAGE þ cn Year dummies þ e ð1Þ;

where PREMIUM is computed as the ratio of the offer price to the target’s share price
7. We do not require acquirers to be publicly traded ﬁrms because private acquirers account for more than half of the M&A transactions using U.S. public targets (Boone and Mulherin 2008; Bargeron, Schlingemann, Stulz, and Zutter 2008). As a result of this research design choice, not all acquiring ﬁrms have publicly available ﬁnancial information. For those transactions for which the acquirer is a U.S. publicly traded ﬁrm having the necessary data on COMPUSTAT or I ⁄ ⁄ B ⁄ E ⁄ S to compute a LQFR_SCORE (n = 468), we calculate the Pearson and Spearman pairwise correlations between acquirers’ and targets’ LQFR_SCORE and ﬁnd them to be not statistically signiﬁcant at 0.088 and 0.093, respectively. Prior research suggests that acquiring ﬁrms overpaid for targets to obtain the accounting beneﬁts under the pooling method, as pooling allowed the deal to be recorded using the book value of the target’s net assets as opposed to the fair value of net assets (Robinson and Shane 1990; Weber 2004). We leave open for future research the investigation of whether the option of pooling accounting is associated with fewer M&A deals being terminated.

8.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust
TABLE 1 Sample Panel A: Sample construction Completed deals Number of M&A deal announcements in 2001–2007 as reported in SDC Firms missing data necessary for our empirical tests Number of observations with required data for deal termination analysis Firms missing deal premium data Number of observations with required data for deal premium analysis Panel B: Composition of sample by settlement year Year 2002 2003 2004 2005 2006 2007 2008 Notes: Completed deals 165 180 201 184 206 287 40 Terminated deals 28 37 27 31 29 27 26 2,772 )1,368 1,404 )141 1,263 Terminated deals 467 )336 234 )29 205

727

Total 3,239 )1,601 1,638 )170 1,468

Total 193 217 228 215 235 314 66

In panel B, sample sizes are reported based on the number of observations with all required data for the deal premium analysis (n = 1,468). Deals settled in 2008 represent those that were announced in 2007 or prior.

four weeks prior to the announcement date as reported by SDC, minus one (Boone and Mulherin 2007; Bargeron et al. 2008), and LQFR_SCORE is as deﬁned in Appendix 1. As stated earlier, it is not clear what role the quality of targets’ ﬁnancial reporting plays in determining deal premiums. On one hand, the acquiring ﬁrm might offer more for targets that report more aggressive operating performance (e.g., higher operating margins by accelerating revenues or deferring the recognition of expenses) or stronger ﬁnancial positions (e.g., greater off-balance-sheet liabilities) via LQFR. Alternatively, acquirers may see through the LQFR and discount targets that present greater information uncertainty related to future expected net cash ﬂows. Consequently, we make no prediction on the relation between LQFR_SCORE and PREMIUM. Schwert (2000) demonstrates bid premiums are higher in hostile takeover attempts (HOSTILE) and Walkling and Edmister (1985) document higher bid premiums when there is competition among multiple bidders to obtain control of the target (MULTIBID). Deal premiums are lower when the acquiring ﬁrm already holds an ownership interest in the target ﬁrm at the time of the takeover bid (Walkling and Edmister 1985; Betton and Eckbo 2000). To control for the acquiring ﬁrm’s prebid ownership of the target we include TOEHOLD and expect it to be negatively associated with PREMIUM. We also control for whether the acquirer’s bid is negotiated or made directly to the target ﬁrm’s
CAR Vol. 30 No. 2 (Summer 2013)

728

Contemporary Accounting Research

shareholders by including the indicator variable TENDER and whether the target and acquirer operate within the same industry (2-digit SIC) with the indicator variable INDUSTRY. Stock-for-stock acquisitions qualifying as tax-free exchanges reduce target ﬁrm shareholders’ tax liability and are associated with lower acquisition premiums (Ayers, Lefanowicz, and Robinson 2003). In contrast, acquirers using cash to gain control of target ﬁrms are hypothesized to pay higher deal premiums (Jensen 1986; Schwert 2000). To control for the effects of stock-based (cash) consideration, we include the indicator variable STOCK (CASH) coded one when the consideration consists solely of stock (cash). Based on prior literature, we expect a negative coefﬁcient for STOCK and a positive coefﬁcient for CASH. Bargeron et al. (2008) ﬁnd public bidders pay higher premiums to gain control of target ﬁrms. The indicator variable PUBLICBIDDER is included in the model and is coded one when the acquiring ﬁrm is a publicly traded company, zero otherwise. Based on prior literature, we expect PUBLICBIDDER to be positively associated with PREMIUM. All deal characteristics are taken from the SDC. Prior research reports that ﬁrm-speciﬁc operating characteristics affect the quality of ﬁrms’ ﬁnancial information (Dechow et al. 1995; Becker, DeFond, Jiambalvo, and Subramanyam 1998; Francis, Maydew, and Sparks 1999; Dechow and Dichev 2002; Francis, LaFond, Olsson, and Schipper 2004, 2005) and documents that targets’ operating characteristics also affect deal premiums. Therefore, it is important to control for innate ﬁrm characteristics correlated with our LQFR score that potentially are also related to deal premiums in order to reduce the threat of a correlated omitted-variables problem.9 (1) includes six additional control variables that capture target ﬁrms’ operating and ﬁnancial risks. The ﬁrst, SIZE, is the log of the value of the transaction which captures the size of the target ﬁrm involved in the deal. Second, rCFO, is the standard deviation of net cash ﬂows from operations of the target ﬁrm over the ﬁve years prior to the acquisition, requiring a minimum of three years. Third, GROWTH is measured as the average percentage growth in sales over the previous three years. The fourth control is ROE, measured as income before extraordinary items divided by book value of common equity. MTB is the ﬁfth control, measured as the ratio of the market value to book value of common equity. The sixth control is LEVERAGE, measured as the ratio of long-term debt to book value of common equity. ROE, MTB, and LEVERAGE are measured at the ﬁscal year end prior to the year in which the deal is announced. Prior literature ﬁnds target ﬁrm characteristics including growth, performance, and leverage are associated with various acquisition outcomes (Walkling and Edmister 1985; Palepu 1986; Dong, Hirshleifer, Richardson, and Teoh 2006), so we make no prediction on the relation between GROWTH, ROE, rCFO, MTB and LEVERAGE and deal premiums. Schwert (2000) ﬁnds shareholders in smaller target ﬁrms receive higher premiums. Accordingly, we expect a negative coefﬁcient for SIZE. Finally, the model includes annual ﬁxed effects to control for changing economic conditions over time and their effect on deal premiums. Variable deﬁnitions are provided in Appendix 2. We test our prediction that LQFR increases the likelihood that M&A deals are terminated by estimating the following logistic regression model:

9.

WIC is a function of many of the same ﬁrm-speciﬁc operating characteristics and |AAC| incorporates relative performance, so ﬁrm-speciﬁc characteristics are to some extent inherently controlled for in the construction of LQFR_SCORE.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust ProbðTERMINATED ¼ 1Þ ¼ Fðb0 þ b1 LQFR SCORE þ b2 HOSTILE þ b3 MULTIBID þ b4 TOEHOLD þ b5 TENDER þ b6 INDUSTRY þ b7 STOCK þ b8 COLLAR þ b9 SIZE þ b10 rCFO þ b11 GROWTH þ b12 ROE þ b13 MTB þ b14 LEVERAGE þ cn Year dummies þ eÞ

729

ð2Þ;

where TERMINATED equals one if the bid is withdrawn (SDC), and zero otherwise. All other variables are as previously deﬁned. The deal characteristic variables represent attributes that prior research documents are associated with the likelihood that the M&A deal is completed or terminated (Hsieh and Walkling 2005; Betton and Eckbo 2000; Walkling 1985). Deals where acquiring and target ﬁrms are involved in friendly discussions regarding the combination are less likely to be terminated, so we predict a positive relation between HOSTILE and the likelihood the deal goes bust. If multiple ﬁrms are competing to take over a particular target, the competition will result in an increase in the likelihood that the deal will go bust for each bidder. Consequently we predict a positive coefﬁcient on MULTIBID. There is some evidence to suggest that the probability of the deal being completed increases when the acquirer has already established an ownership relationship (Hsieh and Walkling 2005); however, we also acknowledge the possibility that the cost of obtaining control may outweigh its beneﬁts for certain targets (e.g., because of integration problems and the accounting effects of consolidation). Therefore, we also leave the coefﬁcient on TOEHOLD unsigned. Prior research provides mixed results on whether deals involving acquiring ﬁrms’ tender offers are more or less likely to be completed, so we make no prediction for the association between TENDER and the deal being terminated. We also include the indicator variable INDUSTRY to control for differences in the likelihood of deal termination across within- and cross-industry deals. Prior research shows that deals are less likely to be completed when acquirers use their stock as the deal consideration because oftentimes the use of acquirers’ shares as method of payment proxies for overvaluation of the acquirers’ shares (Shleifer and Vishny 2003). The overvaluation of shares at the time of the deal announcement can ultimately result in the deal going bust if there is a signiﬁcant drop in share price prior to the completion of the deal. We predict a positive relation between STOCK and the likelihood the deal goes bust. M&A deals involving collars are shown to have lower probability of completion (Hsieh and Walkling 2005), so we expect a positive relation between COLLAR and the likelihood of the deal being terminated. We also include the six additional control variables identiﬁed in (1) capturing target ﬁrms’ operating and ﬁnancial risks. We have no priors on the relations between target ﬁrms’ operating and ﬁnancing characteristics and the probability of termination; therefore, we leave the related coefﬁcients unsigned. Descriptive statistics Table 2 displays the descriptive statistics for the variables used in the empirical tests partitioned by whether the deal is completed or terminated. With the exception of deal premiums, the descriptive statistics suggest that the characteristics of completed deals are distinctly different from the characteristics of deals that go bust. A greater percentage of terminated deals relate to hostile takeovers (34.63 percent) relative to completed deals (2.22 percent). When there is competition for the target in terms of multiple bidders, a greater percentage of deals go bust – 23.90 percent versus 2.77 percent for completed deals. Acquiring ﬁrms also own, on average, more of the targets’ shares at the time of the

CAR Vol. 30 No. 2 (Summer 2013)

730

Contemporary Accounting Research

TABLE 2 Descriptive statistics Panel A: Deal characteristics Completed (n = 1,263) PREMIUM Mean Median Std. dev. HOSTILE Mean MULTIBID Mean TOEHOLD Mean Median Std. dev. TENDER Mean INDUSTRY Mean STOCK Mean CASH Mean PUBLICBIDDER Mean COLLAR Mean Terminated (n = 205) LQFR_SCORE Mean Median Std. dev. SIZE Mean Median Std. dev. rCFO Mean Median Std. dev. GROWTH Mean Median Std. dev. ROE Mean Median Std. dev. MTB Mean Median Std. dev. LEVERAGE Mean Median Std. dev. Panel B: Target characteristics Completed (n = 1,263) Terminated (n = 205)

0.3707 0.2844 0.4471 0.0222 0.0277 0.0473 0.0000 0.1661 0.1417 0.5194 0.1599 0.5455 0.6683 0.0475

0.4134 0.2863 0.5645 0.3463*** 0.2390*** 0.0921*** 0.0000*** 0.2073 0.0829** 0.3122*** 0.1122* 0.5854 0.4000*** 0.0243

5.2591 5.0000 1.8311 5.6757 5.6495 1.9549 0.5162 0.1171 1.3686 0.1119 0.0659 0.3357 -0.0317 0.0683 1.0629 2.3694 1.8158 3.1487 0.6902 0.2164 1.9106

5.6601*** 5.5000*** 1.9371 5.3892* 5.2589* 2.0581 0.5465 0.1215 1.3301 0.0985 0.0502 0.3290 0.0779* 0.0559 1.1652 2.1793 1.6303** 3.3348 0.8467 0.2918 2.3879

Notes: Descriptive statistics are reported based on the reduced sample for the deal premium analysis where n = 1,468. All continuous variables are winsorized at the 1st and 99th percentiles. For continuous variables, signiﬁcant differences in means (medians) are based on the appropriate onetailed or two-tailed t-test (Wilcoxon rank-sum test). For indicator variables, signiﬁcant differences are based on a chi-squared frequency test. Variable deﬁnitions appear in Appendix 2. *, **, and *** represent signiﬁcance levels of 0.10, 0.05, and 0.01, respectively.

bid announcement for terminated deals (9.21 percent) relative to deals that are ultimately completed (4.73 percent). The median values of TOEHOLD are also signiﬁcantly different as 28 percent of acquiring ﬁrms in failed bids have non-zero values of TOEHOLD compared to only 10 percent in completed deals. Terminated deals are also less likely to involve tender offers (8.29 percent), or have acquirers and targets operating in the same
CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

731

industry (31.22 percent) relative to completed deals (14.17 percent and 51.94 percent, respectively). The descriptive statistics indicate that terminated deals are less likely to use the stock of the acquiring ﬁrm as consideration in the deal (11.22 percent) relative to completed deals (15.99 percent). Public bidders are also involved in terminated deals (40.00 percent) less frequently than in completed deals (66.83 percent). The descriptive statistics in Table 2 indicate that target characteristics differ on several dimensions between completed versus terminated deals. Failed targets are smaller and have lower market-to-book ratios (mean SIZE = 5.3892; median MTB = 1.6303) than ﬁrms that are taken over (mean SIZE = 5.6757; median MTB = 1.8158). Failed targets, on average, are more proﬁtable (mean ROE = 7.79 percent) than ﬁrms that are taken over (mean ROE = )3.17 percent); however, the differences in means is attributable to completed acquisitions involving loss ﬁrms. The majority of targets involved in completed and failed deals are proﬁtable and the difference in medians is not statistically signiﬁcant. Consistent with our predictions and the descriptive statistics reported in Panel B of Appendix 1, Table 2 also shows that failed targets have lower quality ﬁnancial reporting. The mean LQFR_SCORE for failed targets is 5.6601, which is greater than the mean LQFR_SCORE (5.2591) for targets involved in completed deals. The median LQFR_SCORE for failed targets is 5.500, greater than the median of 5.0000 for targets in completed deals. All differences are statistically signiﬁcant at the 0.01 level. Table 3 provides correlations among the deal and target characteristics. The upper right-hand portion of the table presents Pearson product-moment correlations, while the lower left-hand portion presents the Spearman rank-order correlations. To facilitate discussion, we focus on the Pearson product-moment correlations. LQFR_SCORE is signiﬁcantly positively correlated with PREMIUM, TOEHOLD and TENDER, and negatively related to INDUSTRY, PUBLICBIDDER, and COLLAR. In addition, LQFR_SCORE is negatively correlated with the target characteristics included in our prediction model. PREMIUM exhibits positive and signiﬁcant correlations with HOSTILE, MULTIBID, TENDER, and CASH while negative and signiﬁcant for STOCK, SIZE, rCFO, GROWTH, ROE, and MTB. These statistics indicate that it is important to control for deal and target characteristics when examining the relation between ﬁnancial reporting quality, deal premiums, and the likelihood that M&A bids will fail. 5. Results Target ﬁnancial reporting quality and deal premiums Table 4 presents the results of estimating (1) that examines the relation between deal premiums and the quality of target ﬁrms’ ﬁnancial reporting. We begin by estimating a benchmark model using deal characteristics that prior research documents are related to deal premiums (‘‘Benchmark model’’). As expected, we ﬁnd premiums offered to targets are larger for deals involving multiple bidders. Our ﬁndings indicate that during our period of analysis, deals involving tender offers result in greater premiums. Our results are consistent with prior research that ﬁnds premiums offered to targets are less when stock is the only form of payment and greater when cash is the only consideration. The signiﬁcantly positive coefﬁcient on PUBLICBIDDER indicates that public acquirers’ offers include a larger premium, on average, than private acquirers’ acquisition bids. In the second set of columns of Table 4, we report the results when adding LQFR_SCORE to the premium model used in prior research. In general, the signs and signiﬁcance of the coefﬁcients on MULTIBID, TENDER, STOCK, CASH, and PUBLICBIDDER are similar to the ‘‘Benchmark model’’ results reported in Table 4. We also ﬁnd a marginally signiﬁcant positive coefﬁcient on INDUSTRY after adding LQFR_SCORE to the ‘‘Benchmark model’’. More important for our study, we ﬁnd LQFR_SCORE is

CAR Vol. 30 No. 2 (Summer 2013)

732

TABLE 3 Pairwise correlations 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

1

2

CAR Vol. 30 No. 2 (Summer 2013) 0.08 0.26 0.03 0.00 )0.04 0.09 0.02 )0.06 0.05 )0.08 )0.05 0.04 )0.03 )0.07 0.15 )0.06 0.04 0.02 )0.09 0.01 0.13 )0.10 0.13 0.18 )0.15 )0.43 0.04 0.36 0.31 )0.07 0.05 0.14 0.05 0.03 0.02 )0.04 )0.03 )0.02 )0.04

Contemporary Accounting Research

0.04

0.23 0.14 0.02 0.04 0.00 0.14 )0.05 )0.01 )0.10 )0.03 )0.06 )0.02 0.01 0.04 0.17 )0.10 )0.10 )0.18 )0.02 0.02 )0.06

0.09 0.02 0.00 0.17 )0.20 )0.09 )0.07 )0.12 )0.08 0.03 0.03 )0.11 )0.02 )0.02 0.08 0.10 0.03 0.02 0.02 0.02 0.10 )0.11 0.01 0.03 0.05 0.01 )0.04 0.01 )0.03 )0.04 0.12 )0.07 )0.03 0.06 )0.25 0.01 )0.01 0.02 )0.01 0.01 0.21 0.06 )0.05 0.07 )0.03 )0.03 0.01 )0.01 0.01 )0.04 )0.12 0.37 0.05 )0.09 0.07 0.02 0.04 0.01 0.06 0.02 )0.39 0.33 0.15 0.01 )0.01 0.08 0.11 )0.06 0.07 0.02 )0.25 )0.19 0.00 )0.12 )0.11 )0.07 0.02 )0.03 )0.01 )0.32 0.17 )0.10 0.18 0.04 0.12 0.01 0.16 )0.02 )0.21 0.17 )0.04 0.04 0.01 0.04 0.01 0.00 )0.03

Deal characteristics 1 PREMIUM 2 HOSTILE 0.04 0.09 3 MULTIBID 4 TOEHOLD 0.03 5 TENDER 0.12 6 INDUSTRY 0.04 7 STOCK )0.12 8 CASH 0.13 9 PUBLICBIDDER 0.03 10 COLLAR 0.03 Target characteristics 11 LQFR_SCORE 0.17 12 SIZE )0.15 13 rCFO )0.14 14 GROWTH )0.03 15 ROE )0.15 16 MTB )0.08 17 LEVERAGE )0.10

)0.02 0.08 0.09 0.02 )0.02 0.00 0.00

0.05 0.07 )0.09 0.00 0.04 )0.11 )0.04 )0.48 )0.10 )0.10 )0.14 )0.12 )0.10 )0.21 )0.04 0.06 )0.01 )0.13 0.17 0.04 )0.51 0.43 0.20 0.14 0.24 0.05 0.00 0.01 )0.02 0.03 )0.11 0.02 0.01 )0.20 0.69 0.00 0.01 0.04 0.11 )0.04 0.03 0.03 0.06 )0.04 0.16 0.06 )0.21 0.30 0.07 0.01 0.16 0.01 )0.05 )0.04 0.01 )0.07 )0.06 0.02 0.02 )0.28 0.33 0.17 0.22 0.03 )0.02 )0.13 0.01 0.11 0.06 )0.04 0.27 0.03 )0.32 0.46 0.14 0.35 0.20 0.31 )0.01 )0.05 0.07 0.04 )0.12 0.02 0.02 )0.21 0.20 0.13 0.03 0.02 0.18

Notes:

Pearson (Spearman) correlation coefﬁcients are reported in the upper right (lower left) portion of the table. Bold text indicates correlations are statistically signiﬁcant at p-value < 0.10.

Target Financial Reporting Quality and M&A Deals that Go Bust
TABLE 4 Regression analysis of deal premiums LQFR and controls coeff.

733

Benchmark model Pred. sign Intercept LQFR_SCORE HOSTILE MULTIBID TOEHOLD TENDER INDUSTRY STOCK CASH PUBLICBIDDER SIZE rCFO GROWTH ROE MTB LEVERAGE Year dummies Sample size Adjusted R2 Notes: ± ± + + – ± + ) + + ) ± ± ± ± ± coeff. 0.371 0.047 0.281 )0.026 0.089 0.027 )0.071 0.083 0.065 t-stat 6.65*** 0.77 3.10*** )0.27 2.23** 1.08 )1.83** 2.71*** 2.29** coeff.

LQFR t-stat 3.33*** 4.57*** 0.93 2.96*** )0.35 1.95* 1.31* )2.10** 2.26*** 2.47***

t-stat

0.191 0.031 0.058 0.265 )0.035 0.077 0.033 )0.081 0.070 0.069

Included 1,468 0.085

Included 1,468 0.096

0.381 4.14*** 0.019 2.46** 0.079 1.28 0.272 3.10*** )0.078 )0.80 0.085 2.20** 0.029 1.17 )0.087 )2.26** 0.054 1.72* 0.087 3.11*** )0.022 )2.21** 0.001 )1.16 )0.021 )0.49 )0.023 )0.82 )0.007 )1.72* 0.010 1.21 Included 1,468 0.107

This table reports OLS regression results where the dependent variable, PREMIUM, is deﬁned as the ratio of the acquirer’s initial offer price to the target’s share price four weeks prior to the announcement date as reported by SDC, minus one. Standard errors of the parameter estimates are clustered at the ﬁrm level. See Appendix 2 for variable deﬁnitions. *, **, and *** represent signiﬁcance levels of 0.10, 0.05, and 0.01, respectively.

positively and signiﬁcantly associated with PREMIUM, indicating that acquirers offer more to targets with low-quality ﬁnancial reporting. This suggests that acquirers perceive additional value beyond that priced by investors in LQFR targets at the time the acquisition agreement is signed. This ﬁnding holds after controlling for targets’ operating and ﬁnancing characteristics (‘‘LQFR and controls’’ model), although the signiﬁcance of the coefﬁcient on LQFR_SCORE is weaker (p-value < 0.05). One possible explanation for ﬁnding a positive relation between PREMIUM and LQFR_SCORE is that target ﬁrms’ shares trade at a discount because of a higher cost of capital due to LQFR (e.g., Ashbaugh-Skaife et al. 2009). Consistent with this explanation, we ﬁnd in untabulated tests that targets with above average LQFR_SCORE have signiﬁcantly lower market-to-book and industry-adjusted PE ratios (p-value < 0.01).10 Acquirers potentially bid more aggressively for LQFR targets knowing that there is an option to

10.

Because LQFR potentially affects both the acquirer’s offer price and target’s preacquisition share price, we reestimate (1) using the target’s total assets per share as the deﬂator in PREMIUM. We continue to ﬁnd a positive and signiﬁcant coefﬁcient on LQFR_SCORE.

CAR Vol. 30 No. 2 (Summer 2013)

734

Contemporary Accounting Research

terminate the transaction when a breach in GAAP is discovered during transactional due diligence. 11 The likelihood of the deal being terminated Having documented that targets’ LQFR is positively associated with deal premiums, we turn to testing whether LQFR increases the likelihood of deals being terminated. Table 5 reports the results of estimating (2). We estimate a regression that excludes LQFR_SCORE (‘‘Benchmark model’’), a regression adding LQFR_SCORE (‘‘LQFR’’) and then a regression that includes deal and target characteristics (‘‘LQFR and controls’’). All estimations of (2) result in signiﬁcant v2 values, indicating that the model is signiﬁcant in explaining the likelihood that deals will be terminated. The pseudo R2 is around 25 percent for each model, and each speciﬁcation correctly predicts TERMINATED for over 88 percent of the 1,638 deals included in the sample. Given that the signs and signiﬁcance levels are similar for all three estimations, we limit our discussion to the ‘‘LQFR and controls’’ model.12 We ﬁnd the coefﬁcients on HOSTILE and MULTIBID to be positive and signiﬁcant, indicating that hostile M&A bids and deals involving more competition (i.e., more than one bidder) are more likely to be terminated. The signiﬁcantly negative coefﬁcient on TENDER indicates that offers made directly to target shareholders are more likely to be completed. This ﬁnding is potentially due to the fact that tender offers are often completed more quickly than other takeovers because they do not require approval through a special meeting of shareholders following an SEC proxy review process.13 We also ﬁnd a negative and signiﬁcant relation between INDUSTRY and the likelihood of termination, suggesting acquirers and targets operating in the same industry encounter fewer frictions in their efforts to complete the deal. The results indicate a signiﬁcantly positive relation between TOEHOLD and the likelihood that the deal will fail. This ﬁnding is consistent with the notion that acquirers fail to gain control via a hostile takeover attempt of a ﬁrm in which it had partial interest.14 The ﬁnding is also consistent with the conjecture that acquirers with partial ownership begin negotiations to take control of the target, but forego control by withdrawing the offer once they access targets’ private information related to targets’ ﬁnancial position and performance. The signiﬁcant positive coefﬁcient on LQFR_SCORE supports our hypothesis that deals involving targets with low-quality ﬁnancial reporting are more likely to be terminated. The marginal effect of LQFR_SCORE indicates that a small increase in
11. As a sensitivity test, we replace LQFR_SCORE with the ﬁve distinct ﬁnancial reporting quality measures of |AAC|, WIC, OFFBS_LIAB, |AFE|, DISP, and reestimate (1). We ﬁnd a positive and highly signiﬁcant coefﬁcient on |AAC|, indicating that acquiring ﬁrms offer larger premiums for targets with greater magnitudes of discretionary accruals. While acquiring ﬁrms offer larger premiums for such targets, it could also be the case that acquirers know that targets with larger discretionary accruals are more likely to violate GAAP thereby giving acquirers a chance to terminate the deal once they can access the policies, procedures, and contracts of target ﬁrms. We explore this issue by examining failed targets’ reasons for ﬁnancial statement restatements in section 5 below. Approximately one-third of terminated deals in the full sample involve multiple bidders. To ensure that our results are not driven by these observations, we also reestimate (2) using only single-bidder deals and ﬁnd that our inferences are unchanged. In untabulated tests, we interact HOSTILE and MULTIBID with LQFR_SCORE, given the potential differences in information ﬂows between the target and acquiring ﬁrm during the due diligence process. We continue to ﬁnd a signiﬁcant positive association between LQFR_SCORE and TERMINATED, but ﬁnd no signiﬁcant association for the interaction terms. The positive and signiﬁcant correlation between HOSTILE and TOEHOLD (0.218) lends some support for this explanation.

12.

13.

14.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust
TABLE 5 Logistic regression of the likelihood the M&A deal is terminated Benchmark model Pred. sign coeff. Intercept LQFR_SCORE HOSTILE MULTIBID TOEHOLD TENDER INDUSTRY STOCK COLLAR SIZE rCFO GROWTH ROE MTB LEVERAGE Year dummies Sample size Number terminated Number completed Pseudo R2 Wald v2 statistic % correctly predicted Type II error % Marginal effect of LQFR_SCORE Notes: ± + + + ± ± ± + + ± ± ± ± ± ± 3.476 3.035 2.390 0.814 )1.061 )0.669 )0.250 )0.739 z-stat LQFR and controls coeff. z-stat

735

LQFR coeff. z-stat )2.47** 2.61*** 10.83*** 8.10*** 2.92*** )3.15*** )3.62*** )0.70 )1.15

Ordered logit coeff. z-stat

6.32*** )0.884 0.104 10.73*** 2.933 8.31*** 2.346 1.78* 1.159 )3.31*** )1.082 )3.85*** )0.641 )0.98 )0.175 )1.07 )0.820

Included 1,638 234 1,404 0.254 211.05*** 89.13 9.88

Included 1,638 234 1,404 0.253 220.59*** 88.89 9.8 0.096***

2.656 3.80*** 0.094 2.09** 3.070 10.86*** 2.401 8.16*** 0.899 1.90* )1.059 )3.30*** )0.621 )3.54*** )0.214 )0.82 )0.687 )0.98 )0.068 )1.07 0.000 )0.63 0.008 0.03 0.091 1.06 )0.011 )0.39 0.095 2.54** Included 1,638 234 1,404 0.261 223.1*** 89.07 9.84 0.094***

0.107 2.896 2.161 0.682

2.58*** 9.61*** 7.10*** 1.46

)0.413 )2.52** 0.232 1.12 0.134 0.33 )0.066 )1.19 0.000 0.31 )0.116 )0.54 0.061 0.84 )0.010 )0.38 0.091 2.31** Included 1,441 217 1,174 0.18 202.04***

The ﬁrst three speciﬁcations of (2) report logistic regression results where the dependent variable, TERMINATED, is an indicator variable coded equal to one if the deal is terminated, zero otherwise. The last speciﬁcation of (2) reports results of an ordered logistic regression where the dependent variable is coded equal to zero if the deal is completed without renegotiation, one if the deal is completed and the ﬁnal offer price is less than the initial offer price, and two if the deal is terminated. Standard errors of the parameter estimates are clustered at the ﬁrm level. See Appendix 2 for variable deﬁnitions. The marginal effect for LQFR_SCORE is estimated at the median of its distribution, holding all control variables constant at the sample means. To determine the signiﬁcance of the marginal probabilities, we obtain standard errors using the delta method (Greene 2003, 674). *, **, and *** represent signiﬁcance levels of 0.10, 0.05, and 0.01, respectively.

CAR Vol. 30 No. 2 (Summer 2013)

736

Contemporary Accounting Research

LQFR_SCORE increases the likelihood of the deal going bust by over 9 percent.15 The ﬁndings reported in Table 5 suggest that LQFR is an important determinant of failed takeover bids.16 Interestingly, the only other target characteristic signiﬁcantly related to deals being terminated is LEVERAGE, where the results indicate that targets with greater leverage are more likely to be involved in deals that go bust. One explanation is that acquirers closely examine targets’ contracts via transactional due diligence, and realize there are more off-balance-sheet liabilities that need to be booked as a result of the M&A deal, creating too much postacquisition leverage for the acquirer. Our analysis of target ﬁrms’ ﬁnancial statement restatements discussed below identiﬁes the improper accounting for leases, contingencies, and off-balance-sheet liabilities as the primary reason for the accounting restatements that occur soon after deal termination. Having to book these additional liabilities as a result of completing the deal could trigger debt covenant violations for the acquiring ﬁrms. Thus far, we have modeled the outcome of the M&A deal as binary: the transaction is completed or terminated. However, after discovering ﬁnancial reporting problems during transactional due diligence, an acquirer might still attempt to complete the deal if the terms of the acquisition agreement can be renegotiated resulting in a lower purchase price, as a lower purchase price can offset the increased risk and uncertainty inherent with the target’s LQFR. To investigate more fully the consequences of targets’ ﬁnancial reporting quality in M&A, we explore whether LQFR affects the likelihood of completion, renegotiation, or termination of the deal. Speciﬁcally, we reestimate (2) using an ordered logistic regression that includes renegotiation by the acquirer as another potential outcome for each transaction. The dependent variable YJ is set equal to one of the three outcomes ranked from least to most severe, where: ( YC ¼ 0 COMPLETED Yj ¼ YR ¼ 1 RENEGOTIATED YT ¼ 2 TERMINATED To identify transactions in which the acquirer renegotiates the terms of the acquisition agreement to pay a lower price, we compare the amounts of the acquirer’s initial and ﬁnal bids reported by SDC for completed deals. When the ﬁnal bid is less than the initial bid, this indicates the acquirer paid a lower price and YJ is coded equal to one for renegotiation. COMPLETED is coded equal to zero if the transaction is completed with no renegotiation and TERMINATED is as previously deﬁned. We predict the relation between LQFR_SCORE and RENEGOTIATED to be positive because acquirers are likely to attempt to renegotiate the offer once they determine that LQFR targets’ ﬁnancial information is not representative of the targets’ underlying economic value after gaining access to targets’ private ﬁnancial information during transactional due diligence.

15.

16.

Holding all other variables constant at their mean, we evaluate marginal effects for LQFR_SCORE at the median of its distribution. Results are quantitatively and qualitatively similar when marginal effects are evaluated at the mean, 25th percentile, and 75th percentile. By construction, LQFR_SCORE reﬂects differences in target ﬁrms’ information environments because the denominator is a function of the number of dimensions of ﬁnancial reporting quality that are publicly available (e.g., not all targets are followed by analysts, so their LQFR_SCORE are based on three dimensions of ﬁnancial reporting quality). To investigate whether there are differences between the ﬁve dimensions of ﬁnancial reporting quality in explaining deals going bust, we replace LQFR_SCORE with |AAC|, WIC, OFFBS_LIAB, |AFE|, DISP, in (2). We ﬁnd marginally signiﬁcant positive coefﬁcients on |AAC|, WIC, and OFFBS_LIAB, suggesting that these unique attributes of ﬁnancial reporting play an important role in the likelihood of deals being terminated. We ﬁnd further evidence of these ﬁndings as we examine the reasons for ﬁnancial statement restatements presented in section 5.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

737

When estimating the ordered logistic regression, we use a subsample in which renegotiation is most likely to occur by excluding deals involving tender offers as these transactions involve negotiating directly with target shareholders rather than target directors.17 Within the subsample excluding tender offers (n = 1,441), we ﬁnd 50 transactions where the acquirer renegotiates a lower ﬁnal bid and 217 terminated deals. This indicates that renegotiation is not pervasive, but occurs at a rate of roughly 23 percent (or lower) relative to the frequency of termination. Results of the ordered logit analysis are reported in the last two columns of Table 5. The key ﬁnding is the parameter estimate for LQFR_SCORE is positive and highly statistically signiﬁcant. Thus, it appears that low-quality ﬁnancial reporting by the target increases the likelihood that bids will be renegotiated and in more severe circumstances terminated. Robustness tests of overbidding and the likelihood of deal termination We acknowledge the possibility that the positive association between LQFR_SCORE and the likelihood of termination might also be explained by overbidding at the time the acquisition agreement is signed as we document earlier.18 For example, consistent with prior research we ﬁnd the likelihood of deal termination is higher when there are multiple bidders competing to gain control of the target. The winner’s curse hypothesis suggests that the winning bidder is more likely to overpay (Thaler 1988). Our results are robust to excluding transactions involving multiple bidders; however, we conduct several sensitivity tests to assess whether overbidding by the acquirer affects the likelihood of termination. First, greater deal premiums can reﬂect acquirer overbidding. If acquirers bid aggressively and later discover adverse information through transactional due diligence, they could seek justiﬁcations for terminating the deal. Second, Bargeron et al. (2008) ﬁnd public bidders pay relatively higher deal premiums relative to private bidders. They attribute this phenomenon to higher potential synergy gains due to the combined entity and a lower likelihood for public bidders to terminate the deal. Third, acquiring ﬁrm managers may reassess whether to complete the acquisition depending on shareholders’ perceptions of the net beneﬁts associated with the deal. Luo (2005) ﬁnds evidence consistent with the hypothesis that acquiring ﬁrm managers learn from negative market reactions to M&A deal announcements as reﬂected by a positive association between announcement returns and the likelihood of deal completion. To the extent the positive association we document between LQFR_SCORE and the likelihood of deal termination reﬂects initial overbidding followed by a greater incentive to terminate the deal, we expect the proxies for overbidding to be positively associated with the likelihood of termination. In Table 6, we report the results of robustness tests where we reestimate (2), adding controls for overbidding by acquirers. In the ﬁrst set of columns (‘‘Controlling for deal premiums’’), we add the variables PREMIUM and BID_INCREASE, where BID_INCREASE (taken from SDC) is coded one when the acquirer increases its ﬁnal offer to gain control of the target ﬁrm, zero otherwise. The coefﬁcients on both variables are insigniﬁcant. The coefﬁcient on LQFR_SCORE, however, remains positive and signiﬁcant, and its marginal effect is similar to that reported in Table 5.
17. In this analysis, we focus on renegotiation where the acquirer ultimately pays a lower price compared to the price initially offered. In tender offers the acquirer can offer a higher price to induce target shareholders to tender a sufﬁcient number of shares to transfer control. As expected, we ﬁnd no instances where the acquirer’s initial bid exceeds its ﬁnal bid for tender offer deals. Results are robust to estimating the model using all deals in the sample and while including the TENDER indicator variable. We thank an anonymous referee for pointing out this potential explanation for why deals involving targets with low-quality ﬁnancial reporting are terminated.

18.

CAR Vol. 30 No. 2 (Summer 2013)

738

Contemporary Accounting Research

TABLE 6 Robustness tests of deal termination controlling for overbidding Controlling for deal premiums Pred. sign LQFR_SCORE Controls for overbidding PREMIUM BID_INCREASE PUBLICBIDDER CAR Controls for deal characteristics Controls for target ﬁrm characteristics Year dummies Sample size Number terminated Number completed Pseudo R2 Wald v2 statistic % correctly predicted Type II error % Marginal effect of LQFR_SCORE Notes: This table reports logistic regression results where the dependent variable, TERMINATED, is an indicator variable coded equal to one if the deal is terminated, zero otherwise. Standard errors of the parameter estimates are clustered at the ﬁrm level. BID_INCREASE is coded one if the acquirer increased its initial takeover bid, zero otherwise. CAR is the cumulative abnormal announcement return using the CRSP value-weighted index for the acquirer’s shares over the three-day window surrounding the announcement of the deal. See Appendix 2 for other variable deﬁnitions. The marginal effect for LQFR_SCORE is estimated at the median of its distribution, holding all control variables constant at the sample means. To determine the signiﬁcance of the marginal probabilities, we obtain standard errors using the delta method (Greene 2003: 674). *, **, and *** represent signiﬁcance levels of 0.10, 0.05, and 0.01, respectively. + ± ± ± ± coeff. 0.114 )0.178 )0.353 z-stat 2.13** )0.81 )1.01 )0.591 )3.05*** )5.022 )2.27** Included Included Included 1,468 205 1,263 0.271 219.31*** 89.07 9.73 0.092*** Controlling for market reaction coeff. 0.878 z-stat 1.92**

All controls coeff. 0.111 )0.151 )0.385 )0.594 )5.831 Included Included Included 1,468 205 1,263 0.322 218.67*** 89.85 8.98 0.082*** z-stat 2.06** )0.65 )1.10 )2.79*** )2.38**

Included Included Included 1,468 205 1,263 0.311 226.27*** 90.26 8.70 0.085***

In the columns of Table 6 labeled ‘‘Controlling for market reaction’’, we add the variables PUBLICBIDDER and CAR as additional controls for overbidding. We deﬁne CAR as the cumulative abnormal announcement return using the CRSP value-weighted index for the acquirer’s shares over the 3-day window surrounding the announcement of the deal. PUBLICBIDDER is as previously deﬁned. We ﬁnd coefﬁcients for both variables to be negative and statistically signiﬁcant. The negative coefﬁcient on PUBLICBIDDER suggests public acquirers are more reluctant to terminate acquisitions due
CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

739

to higher perceived synergies.19 The negative coefﬁcient on CAR is consistent with Luo 2005, suggesting acquirers are more likely to terminate acquisitions perceived negatively by the market. Importantly, the coefﬁcient for LQFR_SCORE remains positive and signiﬁcant. Similar results are obtained when all four controls for overbidding are included in the termination model (as reported in the last columns of Table 6) and the marginal effect of LQFR_SCORE remains statistically and economically signiﬁcant. Taken together, these results suggest that LQFR rather than overbidding is the primary determinant of a deal being terminated. In the next section, we provide additional evidence that supports our claim that acquirers uncover targets’ ﬁnancial reporting problems in transactional due diligence, ultimately increasing the probability of deals going bust. Financial restatements of failed targets Recall that after signing the acquisition agreement, the acquirer begins its transactional due diligence. If the acquirer discovers a breach in the target’s GAAP warranty, the acquirer cannot publicly disclose this information because of separate conﬁdentiality agreements signed prior to the acquisition agreement. Therefore, it is not known at the time the deal terminates what information the acquirer obtained in the transactional due diligence process that leads the deal to go bust. However, if, as a result of transactional due diligence, misstatements in the target’s ﬁnancial reporting are discovered and the target’s management, board of directors, and ⁄ or auditor assesses the misstatements as material, the publicly traded target is required to ﬁle amended (i.e., restated) ﬁnancial statements with the SEC. Therefore, to provide further evidence that targets’ LQFR contributes to the likelihood that deals go bust, we examine the incidence of restatements made by failed targets over the period beginning with the announcement date of the transaction and ending in the second ﬁscal year after the deal goes bust. A restatement by a failed target over this time period conﬁrms that the target’s previously ﬁled GAAP ﬁnancial statements are of low quality. It also provides insights as to the nature of the breached representations and warranties made by targets discovered by the acquirer during transactional due diligence. Panel A of Table 7 reports the frequency of restatements of failed targets relative to two control samples. The ﬁrst control sample is the frequency of restatements among ﬁrms covered by COMPUSTAT and Audit Analytics (the data source of ﬁnancial statement restatements). We ﬁnd that 39 failed targets (15.06 percent) have restatements within two years of the takeover announcement. This is nearly double the percentage of restatements for ﬁrms covered on both COMPUSTAT and Audit Analytics over the period from 2002 to 2008 (8.17 percent).20
19. As stated earlier, our sample does not require acquirers to be publicly traded. As a sensitivity test, we reestimate the premium and termination models using only public-for-public deals. Speciﬁcally, we drop observations related to private bidders and add acquirer size, deﬁned as the log of one plus the acquirer’s market value of equity at the end of the ﬁscal year prior to the deal, to the premium and termination models. We continue to ﬁnd the coefﬁcient on LQFR_SCORE to be positive and signiﬁcant in the deal premium analysis but lose signiﬁcance on the INDUSTRY, CASH, and PUBLICBIDDER coefﬁcients. In the termination model, the LQFR_SCORE, TOEHOLD, TENDER, and INDUSTRY coefﬁcients become insigniﬁcant. It is important to point out that dropping private bidders from the analysis contributes to a substantial decrease in power, because the sample sizes drop 56 percent (1,468 to 635 observations) in the premium model and 58 percent (1,638 to 682 observations) for the deal termination model. It is unclear whether the change in sample composition or lack of statistical power contributes to the insigniﬁcant coefﬁcients. Our analysis includes all restatements reported by Audit Analytics. We make no distinction between the underlying causes of restatements (e.g., error or fraud) because any unreported material misstatement by a target results in a breach of its GAAP representation. Over the period 2002–2008, we ﬁnd 4,428 ﬁrm-year observations with a restatement reported by Audit Analytics, consistent with Plumlee and Yohn 2010 who ﬁnd 3,744 restatements over the period 2003–2006.

20.

CAR Vol. 30 No. 2 (Summer 2013)

740

Contemporary Accounting Research

TABLE 7 Analysis of failed targets’ ﬁnancial statement restatements Panel A: Ex post M&A bid ﬁnancial statement restatements

All restatements Number of ﬁrms covered by Audit Analytics 2002–2008 Number of ﬁrms with restatements Percentage of ﬁrms restating

Failed targets 259 39 15.06%*** Failed targets 153 20 13.07%**

COMPUSTAT 49,749 4,428 8.17% Matched sample 153 11 7.19%

Restatements with available data for multivariate analysis Number of ﬁrms covered by Audit Analytics 2002–2008 Number of ﬁrms with restatements Percentage of ﬁrms restating Panel B: Logistic regression of the likelihood of restatement

Benchmark model Pred. Sign Intercept TERMINATED LQFR_SCORE lnMVE ROE LEVERAGE AUDITOR LEASE rCFO GROWTH MTB DEMPLOYEE LAGRET Year dummies Sample size Number restated Pseudo R2 Wald v2 statistic Pct. Correctly predicted Type II error rate Marginal effect of TERMINATED Marginal effect of LQFR_SCORE ± + + ± ± – – + ± ± ± + + coeff. )2.611 z-stat )31.36***

LQFR and terminated deals coeff. )3.064 0.473 0.065 )0.035 )0.014 0.033 )0.099 0.248 0.000 0.025 )0.010 z-stat )27.70*** 2.07** 6.04*** )3.18*** )1.11 3.21*** )2.13** 4.36*** 1.27 1.28 )4.28***

Additional controls coeff. z-stat

)0.061 )0.016 0.033 )0.110 0.288 0.000 0.036 )0.010

)6.12*** )1.21 3.12*** )2.36** 5.10*** 1.62 1.81* )4.13***

Included 45,001 3,748 0.026 659.47*** 91.67 8.33

Included 45,001 3,748 0.028 708.55*** 91.67 8.33 0.041** 0.076***

)3.458 )21.71*** 0.488 1.73** 0.083 5.53*** )0.013 )0.83 )0.043 )1.69* 0.049 3.61*** 0.009 0.14 0.195 2.46*** 0.000 0.16 )0.027 )0.63 )0.011 )2.75*** 0.004 0.81 )0.082 )2.06** Included 29,788 2,327 0.029 455.84*** 92.19 7.81 0.04* 0.071***

(The table is continued on the next page.)

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust
TABLE 7 (Continued)
Panel C: Reason for failed target restatement Number of failed targets 14 10 8 7 39

741

Reason for restatement Leasing, contingencies, liabilities and expense recognition Asset valuation Consolidation, foreign subsidiaries, and off-balance-sheet issues Revenue recognition

Percent restating 35.90 25.64 20.51 17.95 100.00

Panel D: Restatement periods subject to due diligence Mean Total number of quarters restated Number of quarters restated prior to deal termination Percentage of restatement period subject to due diligence Terminated transactions with overlapping due diligence and restatement periods 8.90 7.38 83.00 Number of failed targets 30 Median 7.00 4.00 57.14 Percentrestating 76.92

Notes: Panel A reports differences in the frequency of restatements by failed targets relative to two control samples. Failed target restatements occur after the M&A announcement date up to two years following the termination of the deal. The ﬁrst control sample is the universe of COMPUSTAT ﬁrms covered by Audit Analytics during 2002–2008. The second control sample is a matched sample where failed targets are matched to ﬁrms covered by COMPUSTAT and Audit Analytics based on industry (2-digit SIC), year, LQFR_SCORE, and size. Signiﬁcant differences in the frequency of restatements between failed targets and the control sample are based on a chi-squared test. Panel B reports a logistic regression of the probability of restatement. Standard errors of the parameter estimates are clustered at the ﬁrm level. lnMVE is the natural log of one plus the market value of equity; DEMPLOYEE is the percentage change in employees, scaled by the percentage change in assets; LAGRET is the lagged one-year cumulative abnormal return using the CRSP value-weighted index. See Appendix 2 for other variable deﬁnitions. The marginal effect for LQFR_SCORE is estimated at the median of its distribution, and TERMINATED by taking the difference in the predicted probability of restatement when TERMINATED = 1 and TERMINATED = 0, holding all control variables constant at the sample means. To determine the signiﬁcance of the marginal probabilities, we obtain standard errors using the delta method (Greene 2003: 674). The reason for a failed target restatement is summarized in panel C and based on the Accounting Category Key ﬁeld provided in Audit Analytics: Leasing, contingencies, liabilities and expense recognition = Key 7, 12, 21, and 42; Asset valuation = Key 1, 3, and 14; Consolidation, foreign subsidiaries, and off-balance-sheet issues = Key 11, 13, and 24; Revenue recognition = Key 6. Panel D reports summary statistics on the length of the restatement period (number of quarters) subject to due diligence. We determine the number of periods restated using the RES_BEGIN_DATE and RES_END_DATE ﬁelds in Audit Analytics. We then determine the number of quarters restated prior to deal termination by comparing the restatement period to the termination date for the deal reported by SDC. Terminated transactions with overlapping due diligence and restatement periods are those transactions where the termination date falls between the beginning and ending dates of the restatement period. *, **, and *** represent signiﬁcance levels of 0.10, 0.05, and 0.01, respectively.

CAR Vol. 30 No. 2 (Summer 2013)

742

Contemporary Accounting Research

The second control sample is constructed by matching the 153 failed M&A target ﬁrms that have nonmissing LQFR_SCORE values to nontarget ﬁrms based on industry (2digit SIC), year, LQFR_SCORE, and size (as measured by total assets at ﬁscal year-end). We ﬁnd that 20 of the 153 failed targets with nonmissing LQFR_SCORE values (13.07 percent) have a restatement within two years of the takeover announcement. This is in contrast to the matched sample where only 11 of the 153 matched ﬁrms (7.19 percent) restate their ﬁnancial statements during the same period. The differences in the frequency of restatements between the failed targets and both control samples are statistically signiﬁcant (at p-value < 0.05 or better). In order to control for other factors related to the likelihood of ﬁnancial statement restatements, we conduct a multivariate analysis by estimating the following logistic regression: ProbðRESTATEMENT ¼ 1Þ ¼ Fðb0 þ b1 TERMINATED þ b2 LQFR SCORE þ b3 SIZE þ b4 ROE þ b5 LEVERAGE þ b6 AUDITOR þ b7 LEASE þ b8 rCFO þ b9 GROWTH þ b10 MTB þ b11 DEMPLOYEE þ b12 LAGRET þ cn Year dummies þ eÞ ð3Þ; where RESTATEMENT is equal to one when the ﬁrm restates its ﬁnancial statements within one year from its ﬁscal year-end date and zero otherwise. We add TERMINATED to the model in order to test our prediction that failed target ﬁrms are more likely to issue restated ﬁnancial statements soon after the deal goes bust. Dechow, Ge, Larson, and Sloan (2011) models accounting misstatements as a function of ﬁnancial reporting quality, operating leases, change in employees, and stock returns. We include LQFR_SCORE (as previously deﬁned) as the proxy for ﬁnancial reporting quality, which we expect to be positively associated with the likelihood of a restatement. The variable LEASE is an indicator variable coded one when the ﬁrm has operating leases, zero otherwise. To control for changes in employees, we include the variable DEMPLOYEE, computed as the percentage change in employees scaled by the percentage change in assets. To control for stock returns, we include the variable LAGRET, computed as the lagged one-year cumulative abnormal return using the CRSP value-weighted index. Based on prior literature, we also control for size, accounting performance, leverage, and auditor quality (Beneish 1997; Jones, Krishnan, and Melendrez 2008). We use the natural log of market value of equity (lnMVE) as our measure of size. AUDITOR is equal to one if the ﬁrm contracts with a Big 4 auditor and zero otherwise, and LEVERAGE is as deﬁned previously. In addition, we include the variables rCFO, GROWTH, and MTB to control for ﬁrms’ operating and ﬁnancial risks. Annual ﬁxed effects are included in the model to control for variation in the number of restatements over time. The ﬁrst column in panel B of Table 7, labeled ‘‘Basic model’’, reports the results of estimating (3) excluding LQFR_SCORE and TERMINATED. We also exclude the variables DEMPLOYEE and LAGRET to report results for the largest possible sample. In line with prior research, we ﬁnd large ﬁrms using high-quality auditors are less likely to restate their ﬁnancial statements. Firms with higher leverage, historical sales growth, and lower market-to-book ratios, and ﬁrms using operating leases are more likely to issue restated ﬁnancial statements. The model is highly signiﬁcant and correctly predicts over 91 percent of the outcomes. The ‘‘LQFR and terminated deals’’ column reports the results of estimating (3) that includes LQFR_SCORE and an indicator variable coded one if the ﬁrm is a failed target (TERMINATED). We ﬁnd the coefﬁcient on LQFR_SCORE to be positive and highly

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

743

signiﬁcant, indicating that ﬁrms with low-quality ﬁnancial reporting are more likely to restate their ﬁnancial statements. We also ﬁnd a positive and signiﬁcant coefﬁcient on TERMINATED. This ﬁnding is consistent with the notion that the transactional due diligence efforts of the acquiring ﬁrm uncovers ﬁnancial statement irregularities and errors of the target that culminate in the M&A deal going bust and the failed target having to restate its ﬁnancial statements. As shown in the ‘‘Additional controls’’ column, these results are robust to controlling for DEMPLOYEE and LAGRET, reducing the sample size to 29,788 observations. Panel C of Table 7 summarizes the reasons for failed targets’ ﬁnancial statement restatements and offers descriptive evidence about the nature of the failed targets’ ﬁnancial reporting issues potentially discovered as a result of transactional due diligence. Over onethird (35.90 percent) of the restatements relate to improper accounting for leases, contingencies, liabilities, and recognition of expenses. This ﬁnding supports the conjecture that acquiring ﬁrms are more likely to terminate the deal after discovering off-balance-sheet liabilities. These liabilities have the potential to add too much ﬁnancial risk to the transaction once the proper identiﬁcation and valuation of liabilities is determined. The next most common reason for restatements is related to asset valuation issues (25.64 percent), followed by restatements due to improper consolidation and accounting for foreign subsidiaries and off-balance-sheet transactions (20.51 percent). We also ﬁnd that 7.95 percent of the restatements relate to instances of improper revenue recognition, which prior research notes is one of the more common reasons, in general, for restating ﬁrms’ ﬁnancial statements. In panel D of Table 7 we provide summary statistics on the number of failed targets’ restated quarters that were subject to due diligence by the acquiring ﬁrm. Failed targets that ﬁled restatements, on average, restated 8.90 ﬁscal quarters (median = 7.00). On average, 7.38 quarters (82.92 percent) are restated for the periods ending prior to the deal termination date. Overall, 30 of the 39 failed targets (76.92 percent) that ﬁled restated ﬁnancial statements did so for quarters that overlap with the transactional due diligence period. Collectively our analysis of restatements corroborates our claim that targets with LQFR are more likely to be involved in failed M&A deals. 6. Conclusion Our study provides evidence that target ﬁrms’ low-quality ﬁnancial reporting has multiple consequences for M&A deals. First, we document that acquirers pay higher premiums for targets that have low-quality ﬁnancial reporting. Second, we provide evidence that lowquality ﬁnancial reporting by targets increases the likelihood that deals are renegotiated and, more importantly, terminated. Speciﬁcally, we ﬁnd that almost 14 percent of all deals go bust, and low-quality ﬁnancial reporting by targets increases the chances of termination by over 9 percent. Third, our research identiﬁes a new determinant of ﬁnancial statement restatements as we document failed targets are more likely than other ﬁrms to ﬁle restated ﬁnancial statements. Finding failed targets are more likely than other publicly traded ﬁrms to restate their ﬁnancial statements supports our claim that low-quality ﬁnancial reporting contributes to M&A deals being terminated. Prior literature examines the consequences of low-quality ﬁnancial reporting, linking it to higher costs of debt and equity capital. Our paper offers new evidence on the consequences of low-quality ﬁnancial reporting, speciﬁcally as it relates to the market for corporate control, as we show that low-quality ﬁnancial reporting constrains the exchange of ownership and control via M&A. Future research can explore how low-quality ﬁnancial reporting by targets in completed M&A deals affect the quality of ﬁnancial reporting of the combined entity.

CAR Vol. 30 No. 2 (Summer 2013)

744

Contemporary Accounting Research

Appendix 1 Low-quality ﬁnancial reporting (LQFR) score
Panel A: Financial reporting quality measures |AAC| Magnitude of the target’s performance-adjusted discretionary total accruals at the end of the ﬁscal year immediately prior to the acquisition announcement. Performance-adjusted discretionary accruals are computed following Kothari et al. 2005 using the following steps. First, expected accruals for nonﬁnancial ﬁrms are calculated by estimating the following OLS regression by 3-digit, 2-digit, and 1-digit SIC code such that each industry-year group has at least 20 ﬁrms: TAit ¼ b0 þ b1 ð1=ASSETSitÀ1 Þ þ b2 DSALESit þ b3 PPEit þ eit ; where TA represents total accruals, ASSETS is total assets, DSALES is the change in sales minus the change in accounts receivable, and PPE is net property, plant, and equipment. The terms TA, DSALES, and PPE are scaled by lagged total assets. The residual from the regression above represents the estimated discretionary accrual. To performance-adjust, we form ten portfolios for each industry group based on the decile rankings of prior year return on assets (ROA). We subtract the median discretionary accrual for each ROA-decile portfolio from the estimated discretionary accrual, label this measure AAC, and take its absolute value. WIC The probability of a target having a material weakness in internal control over ﬁnancial reporting estimated using the parameters of the logistic regression model reported in Ashbaugh-Skaife et al. 2007 noted as: X11 Xbj ¼ À3:996 þ 0:087 Ã Segments þ 0:361 Ã Foreign þ 0:402 Ã M&A j¼1 þ 0:417 Ã Restruct þ 0:059 Ã Growth þ 1:163 Ã Inventory þ À0:036 Ã Size þ 0:475 Ã Loss þ À0:015 Ã Rzscore þ 2:008 Ã Auditor Resign WIC ¼ exp P11 j¼1 Xbj Ã ð1=1 þ

P11

j¼1

Xbj Þ;

where Segments is the number of business segments of the target; Foreign is an indicator variable coded one if the target reports foreign sales, zero otherwise; M&A is an indicator variable coded one if the target engaged in a merger or acquisition, zero otherwise; Restruct is an indicator variable coded one if the target underwent restructuring in any of the prior three years, zero otherwise; Growth is the average percentage growth in sales; Inventory is the average inventory level; Size is the average market value of equity; Loss is the percentage of the prior three years the target reported a net loss; Rzscore is the target’s Altman Z-Score decile ranking among all COMPUSTAT ﬁrms; Auditor_Resign is an indicator variable coded one if the target’s auditor resigned during the 15-month period prior to the target’s ﬁscal year-end, zero otherwise. All averages for continuous variables are taken over the previous three ﬁscal years. OFFBS_LIAB Off-balance-sheet liabilities are computed using the residual from the following regression: PRC ¼ b0 þ b1 ASSETS þ b2 LIAB þ e; where PRC = stock price at ﬁscal year-end (COMPUSTAT); ASSETS = total assets per common share outstanding (COMPUSTAT); and LIAB = total liabilities per common share outstanding (COMPUSTAT). The regression is estimated cross-sectionally at the 3-digit, 2-digit, and 1-digit SIC code levels such that each industry-year has a minimum of 20 ﬁrms. Negative residuals represent off-balance-sheet net liabilities and are transformed by taking their absolute value. Positive residuals are set equal to zero.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

745

|AFE| The absolute value of analyst forecast error for the target ﬁrm in the year immediately prior to the announcement of the acquisition, requiring a minimum of three analyst forecasts. Analyst forecast error is computed as the difference between the actual reported earnings per share and the median analyst estimate 8 months prior to the ﬁscal year-end (I ⁄ B ⁄ E ⁄ S), scaled by end of period stock price from the previous year (COMPUSTAT). DISP The dispersion of analyst forecasts for the target ﬁrm in the year immediately prior to the announcement of the acquisition, requiring a minimum of three analyst forecasts. Analyst forecast dispersion is computed as the standard deviation of analyst forecasts 8 months prior to the ﬁscal year-end (I ⁄ B ⁄ E ⁄ S), scaled by end of period stock price from the previous year (COMPUSTAT). LQFR_SCORE Low-quality ﬁnancial reporting score measured as the average of the target’s ﬁnancial reporting quality variables. The average is computed for each target ﬁrm as the sum of its decile rank of |AAC|, WIC, OFFBS_LIAB, |AFE|, and DISP divided by the number of ﬁnancial reporting variables with nonmissing data. Panel B: Descriptive statistics on ﬁnancial reporting quality measures Financial reporting quality measure |AAC| Completed (n = 1,130) Terminated (n = 215) WIC Completed (n = 896) Terminated (n = 143) OFFBS_LIAB Completed (n = 1,168) Terminated (n = 192) |AFE| Completed (n = 771) Terminated (n = 113) DISP Completed (n = 771) Terminated (n = 113) LQFR_SCORE Completed (n = 1,404) Terminated (n = 234) Mean Median Std. dev.

0.1156 0.1170 0.4482 0.5119 * 0.9244 1.1806 0.0276 0.0357 0.0051 0.0087** 5.1449 5.6689***

0.0648 0.0636 0.3381 0.5362* 0.0120 0.0692*** 0.0045 0.0051 0.0011 0.0014* 5.0000 5.5500***

0.1593 0.1562 0.4347 0.4528 3.1314 3.5186 0.1090 0.1278 0.0159 0.0023 1.9849 1.9840

*, **, and *** represent signiﬁcance levels of 0.10, 0.05, and 0.01, respectively.

CAR Vol. 30 No. 2 (Summer 2013)

746

Contemporary Accounting Research

Panel C: Pearson correlations between ﬁnancial reporting quality measures 1 1 2 3 4 5 6 |AAC| WIC OFFBS_LIAB |AFE| DISP LQFR_SCORE 2 0.084 3 0.162 0.108 4 0.093 0.163 0.434 5 0.072 0.094 0.342 0.562 6 0.445 0.548 0.401 0.329 0.408

Bold text indicates that correlations are statistically signiﬁcant at p-values < 0.10.

Appendix 2 Variable deﬁnitions
Deal characteristics PREMIUM RENEGOTIATED TERMINATED HOSTILE MULTIBID TOEHOLD TENDER INDUSTRY STOCK CASH PUBLICBIDDER COLLAR The ratio of the acquirer’s initial offer price to the target’s share price four weeks prior to the announcement date (SDC), minus one. An indicator variable equal to one if the deal is completed and the ﬁnal offer price is less than the initial offer price, zero otherwise (SDC). An indicator variable equal to one if the takeover bid is terminated, zero otherwise (SDC). An indicator variable equal to one if the deal is reported by SDC as unsolicited or hostile, zero otherwise (SDC). An indicator variable equal to one if there are multiple bidders for the target, zero otherwise (SDC). The percentage of the target’s common shares held by the acquirer on the acquisition announcement date (SDC). An indicator variable equal to one if the acquiring ﬁrm structures its bid in the form of a tender offer, zero otherwise (SDC). An indicator variable set equal to one if the target and acquirer share the same primary 2-digit SIC. An indicator variable equal to one if the consideration for the acquisition consists of the acquiring ﬁrm’s stock, zero otherwise (SDC). An indicator variable equal to one if the consideration for the acquisition consists of only cash, zero otherwise (SDC). An indicator variable set equal to one when the acquiring ﬁrm is a publicly traded company, zero otherwise (SDC). An indicator variable set equal to one if the deal includes a collar provision that restricts the value and ⁄ or amount of equity shares to be exchanged in the bid, zero otherwise (SDC).

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

747

Target characteristics LQFR_SCORE SIZE rCFO Low-quality ﬁnancial reporting score (see Appendix 1 for details). The natural log of the value of the consideration offered by the acquirer as reported in the deal value ﬁeld reported by SDC. The standard deviation of net cash ﬂows from operations (COMPUSTAT) of the target ﬁrm over the ﬁve years prior to the acquisition, requiring a minimum of three years. Average percentage growth in target sales (COMPUSTAT) over the previous three years. Target income before extraordinary items divided by book value of common equity (COMPUSTAT). Ratio of the target’s market value to book value of common equity (COMPUSTAT). Ratio of target’s long-term debt to book value of common equity (COMPUSTAT).

GROWTH ROE MTB LEVERAGE

References
Anilowski, C., A. J. Macias, and J. M. Sanchez. 2009. Target ﬁrm earnings management and the method of sale: Evidence from auctions and negotiations. Working paper, SSRN eLibrary. Ashbaugh-Skaife, H., D. W. Collins, and W. R. Kinney, Jr. 2007. The discovery and reporting of internal control deﬁciencies prior to SOX-mandated audits. Journal of Accounting and Economics 44 (1-2): 166–92. Ashbaugh-Skaife, H., D. W. Collins, W. R. Kinney, Jr., and R. LaFond. 2008. The effect of SOX internal control deﬁciencies and their remediation on accrual quality. The Accounting Review 83 (1): 217–50. Ashbaugh-Skaife, H., D. W. Collins, W. Kinney, Jr., and R. LaFond. 2009. The effect of SOX internal control deﬁciencies on ﬁrm risk and cost of equity. Journal of Accounting Research 47 (1): 1–43. Ayers, B. C., C. E. Lefanowicz, and J. R. Robinson. 2003. Shareholder taxes in acquisition premiums: The effect of capital gains taxation. The Journal of Finance 58 (6): 2783–801. Bargeron, L. L., F. P. Schlingemann, R. M. Stulz, and C. J. Zutter. 2008. Why do private acquirers pay so little compared to public acquirers? Journal of Financial Economics 89 (3): 375–90. Barron, O. E., O. Kim, S. C. Lim, and D. E. Stevens. 1998. Using analysts’ forecasts to measure properties of analysts’ information environment. The Accounting Review 73 (4): 421–33. Barth, M. E. 1991. Relative measurement errors among alternative pension asset and liability measures. The Accounting Review 66 (3): 433–63. Becker, C. L., M. L. DeFond, J. Jiambalvo, and K. R. Subramanyam. 1998. The effect of audit quality on earnings management. Contemporary Accounting Research 15 (1): 1–24. Beneish, M. D. 1997. Detecting GAAP violation: implications for assessing earnings management among ﬁrms with extreme ﬁnancial performance. Journal of Accounting and Public Policy 16 (3): 271–309. Betton, S., and B. Eckbo. 2000. Toeholds, bid jumps, and expected payoffs in takeovers. Review of Financial Studies 13 (4): 841–82. Boone, A. L., and J. H. Mulherin. 2007. How are ﬁrms sold? The Journal of Finance 62 (2): 847–75.

CAR Vol. 30 No. 2 (Summer 2013)

748

Contemporary Accounting Research

Boone, A. L., and J. H. Mulherin. 2008. Do auctions induce a winner’s curse? New evidence from the corporate takeover market. Journal of Financial Economics 89 (1): 1–19. Dechow, P. M., and I. D. Dichev. 2002. The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review 77 (4): 35–68. Dechow, P. M., W. Ge, C. R. Larson, and R. G. Sloan. 2011. Predicting material accounting misstatements. Contemporary Accounting Research 28 (1): 17–82. Dechow, P. M., R. G. Sloan, and A. P. Sweeney. 1995. Detecting earnings management. The Accounting Review 70 (2): 193–225. Dong, M., D. Hirshleifer, S. Richardson, and S. Teoh. 2006. Does investor misvaluation drive the takeover market? The Journal of Finance 61 (2): 725–62. Doyle, J., W. Ge, and S. McVay. 2007. Determinants of weaknesses in internal control over ﬁnancial reporting. Journal of Accounting and Economics 44 (1-2): 193–223. Erickson, M., and S. Wang. 1999. Earnings management by acquiring ﬁrms in stock for stock mergers. Journal of Accounting and Economics 27 (2): 149–76. Francis, J., R. LaFond, P. M. Olsson, and K. Schipper. 2004. Costs of equity and earnings attributes. The Accounting Review 79 (4): 967–1010. Francis, J., R. LaFond, P. Olsson, and K. Schipper. 2005. The market pricing of accruals quality. Journal of Accounting and Economics 39 (2): 295–327. Francis, J. R., E. L. Maydew, and H. C. Sparks. 1999. The role of Big 6 auditors in the credible reporting of accruals. Auditing 18 (2): 17–34. Greene, W. 2003. Econometric analysis (5th ed.). Upper Saddle River, NJ: Pearson Education. Healy, P. M., and J. M. Wahlen. 1999. A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13 (4): 365–83. Hsieh, J., and R. A. Walkling. 2005. Determinants and implications of arbitrage holdings in acquisitions. Journal of Financial Economics 77 (3): 605–48. Jensen, M. C. 1986. Agency costs of free cash ﬂow, corporate ﬁnance, and takeovers. The American Economic Review 76 (2): 323–29. Jones, K. L., G. V. Krishnan, and K. D. Melendrez. 2008. Do models of discretionary accruals detect actual cases of fraudulent and restated earnings? An empirical analysis. Contemporary Accounting Research 25 (2): 499–531. Kothari, S. P., A. J. Leone, and C. E. Wasley. 2005. Performance matched discretionary accrual measures. Journal of Accounting and Economics 39 (1): 163–97. Lambert, R., C. Leuz, and R. E. Verrecchia. 2007. Accounting information, disclosure, and the cost of capital. Journal of Accounting Research 45 (2): 385–420. Lang, M., and R. Lundholm. 1996. Corporate disclosure policy and analyst behavior. The Accounting Review 71 (4): 467–92. Luo, Y. 2005. Do insiders learn from outsiders? Evidence from mergers and acquisitions. The Journal of Finance 60 (4): 1951–982. McNichols, M. F. 2003. Discussion of ‘‘Why are earnings kinky? An examination of the earnings management explanation.’’ Review of Accounting Studies 8 (2-3): 385–91. McNichols, M. F., and S. Stubben. 2012. The effect of target-ﬁrm accounting quality on valuation in acquisitions. Working paper, SSRN eLibrary. Palepu, K. G. 1986. Predicting takeover targets: A methodological and empirical analysis. Journal of Accounting and Economics 8 (1): 3–35. Plumlee, M., and T. L. Yohn. 2010. An analysis of the underlying causes attributed to restatements. Accounting Horizons 24 (1): 41–64. Robinson, J. R., and P. B. Shane. 1990. Acquisition accounting method and bid premia for target ﬁrms. The Accounting Review 65 (1): 25–48. Schwert, G. W. 2000. Hostility in takeovers: In the eyes of the beholder? The Journal of Finance 55 (6): 2599–640.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&A Deals that Go Bust

749

Shleifer, A., and R. W. Vishny. 2003. Stock market driven acquisitions. Journal of Financial Economics 70 (3): 295–311. Thaler, R. H. 1988. Anomalies: The winner’s curse. The Journal of Economic Perspectives 2 (1): 191– 202. Walkling, R. A. 1985. Predicting tender offer success: A logistic analysis. The Journal of Financial and Quantitative Analysis 20 (4): 461–78. Walkling, R. A., and R. O. Edmister. 1985. Determinants of tender offer premiums. Financial Analysts Journal 41 (1): 27–37. Wangerin, D. D. 2012. The consequences of M&A due diligence for post-acquisition performance and ﬁnancial reporting. Working paper, SSRN eLibrary. Weber, J. P. 2004. Shareholder wealth effects of pooling-of-interests accounting: Evidence from the SEC’s restriction on share repurchases following pooling transactions. Journal of Accounting and Economics 37 (1): 39–57.

CAR Vol. 30 No. 2 (Summer 2013)

Target Financial Reporting Quality and M&a Deals That Go Bust

Similar Documents

Investing

University of Rochester Roundtable on Corporate M&a and Shareholder Value

P1 Acca Lsbf

Earning Quality

Cisco Systems Ethical Strategy

Mr Ahsan Aslam

The Relevance/Irrelevance Issue

Risk Management

After Bailout

Consumer Behvior

Juvenile Justice and Delinquency Prevention Act

Fddddddd

Middle Market Investment Banking

Gaurav

Jamaica Water Properties

Popular Essays