Should I Trust You? Earnings Quality as a Signal in Earnout Agreements

Transcription

1 Should I Trust You? Earnings Quality as a Signal in Earnout Agreements Authors: Annalisa Prencipe Department of Accounting Università Bocconi annalisa.prencipe@unibocconi.it Luca Viarengo* Department of Accounting Università Bocconi luca.viarengo@unibocconi.it This draft: March 2016 Preliminary, please do not quote Comments Welcome *Corresponding Author

2 Should I Trust You? Earnings Quality as a Signal in Earnout Agreements Abstract This paper aims to test whether past earnings quality of the acquirer serves as a signal of its trustworthiness in view of the inclusion of an earnout agreement in an acquisition contract. Earnouts are contractual agreements that link part of the payment of an acquisition to the future performance of the acquired company. In the US, above 12% of the corporate acquisitions over the period include earnout agreements, and this percentage raises to over 20% in some industries. Typically, earnouts are contractual solutions aimed at bridging valuation disagreement between the potential acquirer and the seller of the target firm. Prior studies have highlighted earnouts usefulness in reducing adverse selection problems and the acquirer s valuation risk in M&A transactions. However, from the seller s perspective, such agreements are subject to moral hazard issues as the acquirer may not exert effort to achieve pre-determined benchmarks, as originally contracted. Moreover, many earnouts are based on accounting performance measures, which may be affected by accounting measurement issues. Consequently, earnouts agreements are often settled in courts of law. To reduce the probability of a litigation process, the uncertainty and costs associated with it, we expect the sellers to explore and investigate the quality type of the acquirer, in particular of its trustworthiness, prior to entering into an earnout agreement. A possible indicator of trustworthiness is the acquirer s past earnings quality. Based on a sample of 9,178 acquisition deals completed in the US between 2002 and 2014, we find that the acquirer s past earnings management is negatively associated to the probability of inclusion of an earnout in acquisition contracts. Keywords: Earnouts, acquisitions, earnings quality, earnings management. 2

3 1. Introduction An earnout is a contractual agreement that links part of an acquisition price to the future performance of the acquired company. Earnouts are fairly common in acquisition deals. In the US, over 12% of the corporate acquisitions, during the period , include earnout agreements. Earnout deals are also significant in terms of nominal value. Over the same period, the annual average value of earnout deals is close to 3 billion dollars, compared to the annual average of nonearnout deals of 32.6 billion dollars. Earnouts aim at bridging a valuation gap between the acquirer and the sellers, thus making the deal possible. When the negotiating parties are unable to reach an agreement on the value of the target company, earnouts contracts facilitate the closing of the deal by linking part of the acquisition price to the certain milestones, such as hitting target levels of performance of the acquired company. Earnouts are advantageous for acquirers as they allow risk sharing and reduce information asymmetry that typically affects acquisitions. In other words, earnouts allow to share the valuation risk between acquirer and sellers, and act as a mechanism of partial verification of the quality of the target company, as the sellers are willing to accept a deferred payment contingent on the performance of their firm only if they believe that the agreed-upon benchmarks, which trigger the additional payment, will be met. Earnouts are valuable also to the sellers, as they reduce the adverse selection risk for the target firm, thus allowing to negotiate a higher consideration. However, this benefit is costly, because, after the closing of the deal, the sellers bear not only the risk that the acquired company will not meet the pre-specified earnout requirements, but also the risk that the acquirer will behave opportunistically in the attempt to reduce, or avoid, the contingent payment. Such opportunism can be implemented by either managing the (reported and/or real) accounting results, or by reducing effort made to reach the target s benchmarks. The sellers of the target firm may be aware of the risk of opportunistic behavior by the acquirer, thereby introducing stringent rules on how to measure performance benchmarks. Such measurement 3

4 rules are applied in the attempts to reduce possible costly litigations. Nonetheless, legal disputes on the actual realization of the performance levels stated by the earnout are frequent 1. Trevis Laster, a judge who had to decide on a dispute related to an earnout payment, once said "an earnout often converts today's disagreement over price into tomorrow's litigation over outcome" 2. A survey conducted by the law firm Morrison & Foerster LLP (2012) on a sample of acquisitions in the hightech sector an industry where earnouts are used the most 3 with over 300 respondents reports that almost three-quarters of those who have used earnouts claimed that such clauses led to subsequent disputes or litigation, and nearly one-fifth of respondents estimated there had been post-deal conflicts over earnouts up to half of the times. In fact, practitioners warn M&A dealers to carefully define the accounting measures or entries to be included or excluded from the computation of the earnout 4. Recent anecdotal evidence helps clarifying this issue 5. In M acquired Acolyte Biomedica, a pharmaceutical company owned by Porton Capital, for an upfront payment of 10.4m and an earnout of 41m, contingent on net sales targets. In 2011, 3M was sued in the UK Court of Law by Porton Capital. The former owners of Acolyte Biomedica requested the Courts to rule for a higher earnout payment than that offered by 3M, accusing 3M for applying a highly conservative recognition approach for recording net sales. 6 Apparently, once the earnout acquisition deal is agreed and ownership has passed to the acquirer, only the target s former owners face transaction risk. That risk is associated with the acquirer incentive to either shirk and/or employ tactics to reduce the earnout payment. The target firm is 1 Some examples of earnout disputes ended up in litigation are Comet Systems Inc Shareholders Agent v. MIVA Inc, 980 A.2d 1024 (Del. Ch. 2008); Chambers v. Genesee & Wyoming Inc, 2005 WL (Del. Ch. Aug. 11, 2005); William J. LaPoint v. AmerisourceBergen Corp, 2007 WL (Del. Ch. Sept. 4, 2007), aff d, 956 A.2d 642 (Del. 2008). 2 Airborne Health, Inc. and Weil, Gotshal & Manges LLP v. Squid Soap, LP, C.A. No VCL (Del. Ch. Nov. 23, 2009). 3 Datar, Frankel and Wolfson (2001) report that earnout in the high tech sector are more than one third of the total. 4 See, for example, Fox and Wolf (2010), Crimmins, Gray, Waller, Brown (2010) or Shannon and Reilly (2011). 5 Porton Capital Technology Funds & Ors v 3M UK Holdings Ltd & Anor [2011] EWHC 2895 (Comm), 07 November The trial, however, ended largely in favor of 3M. 4

5 likely to be aware of the asymmetric risk inherited with that transaction. 7 Therefore, we expect that, during negotiations, both sides to the negotiations attempt to study the counterparty s quality, including its reliability, honesty, and trustworthiness, part of which depends on that party s past behavior. One possible signal of reliability and honesty is the firms financial reporting practices. Put differently, the sellers may gain additional insights into the degree of honesty by observing the acquirer s financial reporting quality in terms of earnings management, serving as an indication of trustworthiness. Following Healy and Whalen (1999) definition, earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers. Target owners may assume that, if the acquirer has the tendency to act opportunistically through earnings management to affect contractual outcomes, it is likely that it will act opportunistically also after the acquisition to deceive the target s owners. Therefore, sellers decision whether to accept an earnout agreement may depend on their perceived probability that the acquirer will use opportunistic tactics to avoid the contingent payment. The extent to which the acquirer manages earnings in the past may be used as a signal of such probability, i.e. the greater the earnings management, the higher the probability of future opportunistic behavior by the acquirer. Following this line of reasoning, we hypothesize that the decision to include an earnout agreement in an acquisition contract is positively (negatively) related to the acquirer s past earnings quality (earnings management). We test this hypothesis on a sample of 9,178 deals completed in the US over the period In order to capture the quality of earnings of the bidder, we rely on an earnings management proxy, 7 In Appendix A we provide an example of an earnout agreement (Source: 5

6 computed using a version of the modified Jones model (Dechow, Sloan and Sweeney, 1995). A lower earnings management proxy indicates a higher level of earnings quality, and therefore a higher acquirer s trustworthiness. While controlling for earnouts determinants as reported by prior studies, we provide statistical evidence of a significantly negative association between the acquirer s past earnings management and the likelihood of inclusion of an earnout in an acquisition deal. Overall, our results validate the hypothesis that the acquirer s past earnings quality is used as a signal by the seller to evaluate the former s reliability. Our paper contributes to the literature on earnouts by showing that the acquirer s reporting quality is used as a signal of the acquirer s trustworthiness and is a significant determinant of the decision by the seller to accept such contractual agreements. Our paper contributes also to the literature on the relation between M&As and earnings management, that so far has mainly been focused on the managerial incentives to manage earnings around the acquisition time to enhance the results of the negotiation. As far as we know, our study is the first to show that past earnings quality of the counterparty is a relevant determinant of the structure of acquisition contracts. The remainder of the paper is organized as follows. The next section provides a review of the relevant literature on earnouts and earnings management. In Section 3 we develop our hypothesis, and in Section 4 we describe the sample and data used in the empirical analysis. In Section 5 we describe our model, while in Section 6 we report our empirical results. In Section 7 we discuss some robustness tests. Section 8 concludes. 6

7 2. Literature review 2.1 Earnouts Compared to other areas in the disciplines of finance and accounting, research on earnouts is fairly recent. The seminal papers on the topic focus on the role of earnout contracts as mechanisms to reduce information asymmetry. Kohers and Ang (2000) analyze acquisitions of US companies over the period They report that earnouts are mainly used in acquisitions of private companies and subsidiaries due to the absence of a market price which increases the uncertainties on the value of the target and in acquisitions of companies operating in the service or high-tech sectors, characterized by higher uncertainty due to issues related to market demand for the target company products, human capital and growth opportunities. They also suggest that the likelihood of using earnout contracts is increasing in the size of the target, while it is decreasing in the size of the acquirer, and that the likelihood increases when the target firm operates in a different industry than the acquirer s. A later study by Datar, Frankel and Wolfson (2001) reaffirm Kohers and Ang s results, using international acquisitions over the period More recently, Ragozzino and Reuer (2009) analyze acquisitions of private target firms, pointing out that earnouts are used more frequently when the target is young, and when it operates in an industry that requires different expertise than that of the acquirer. Barbopulos and Sudarsanam (2012) focus on UK acquisitions, suggesting that when earnouts are optimally applied to reduce information asymmetry the acquirer experiences higher market returns compared to non-earnout transactions. Similar results are found by Kohers and Ang (2000). Using a sample of US deals over the period , Cain, Denis and Denis (2011) show that the performance parameters used in earnouts agreement are chosen in order to discover the actual value of the target firm. Cain et al. also suggest that the time period over which the earnout-related performance benchmark is measured tends to increase with the relevance of R&D costs and to decrease with the return volatility in the target s industry. 7

8 However, research on accounting or reporting issues related to earnouts is fairly limited. Allee and Wangerin (2013) study the impact of changes in accounting regulation on the use of earnouts. More specifically, they analyze the effects of the revised US accounting standards on business combinations (i.e. FASB ASC 805, formerly SFAS 141(R)), which modified the accounting recognition and valuation of contingent payments. While there was no obligation to recognize earnouts in financial statements prior to the revision, as of the fiscal year beginning after December 15, 2008, it is mandatory to value contingent payments at their fair value, and to evaluate these contingent payments each year until their expiration. Consistent with a financial reporting cost hypothesis, Allee and Wangerin (2013) report less frequent use of earnouts after the adoption of the new standard. However, the presence of a high-quality auditor tends to decrease such an effect. Based on a sample of deals carried out over the period July 1, 2006 June 30, 2011, Cadman, Carrizosa, and Faurel (2014) arrive at different conclusions, showing that the percentage of deals including earnouts did not change significantly after the new standard adoption. 2.2 Earnings management in M&As Although literature on earnings quality and earnings management is vast, evidence on the relation between M&A transactions and earnings management is rather limited. Using a sample of stockfor-stock deals completed between 1985 and 1990, Erickson and Wang (1999) suggest that earnings are managed upward by acquirers in the quarters preceding the deal, and it is proportional to the economic benefit for the acquirer, as captured by the relative size of the deal compared to the value of the acquirer s capitalization. The rationale behind this type of earnings management is that acquirers attempt to increase their stock price, thus reducing the relative cost of acquiring the target. Studying UK acquisition deals, Botsari and Meeks (2008) report similar results. Based on a US sample of acquisitions during , Louis (2004) reports that the effect of pre-acquisition earnings management on the acquirer s stock prices is reversed after closing the deal. Indeed, the 8

9 study reports a negative correlation between the level of earnings management and both short and the long term post-acquisition returns to the acquirer. Jensen (2005) contends that overvalued firms tend to inflate earnings and to undertake stock acquisitions to create the illusion of growth to satisfy market expectations. However, Ball and Shivakumar (2008) claim that, because large corporate events are characterized by higher than usual litigation and regulatory risk from inflating earnings, managers should be dissuaded from manipulating their reported earnings around these events. Gong, Louis and Sun (2008) reports that pre-merger abnormal accruals are a strong determinant of post-merger lawsuits. The effect of abnormal accruals is significant even after controlling for the post-merger abnormal return, which suggests that pre-merger earnings management has a first-order effect on the likelihood of a lawsuit. To the best of our knowledge, there are no prior studies that examine the acquirer s earnings management as an ex-ante signal of its reliability and at its consequences on the structure of acquisition contracts. 3. Hypothesis development As mentioned above, if an acquisition contract includes an earnout, the payment of part of the consideration is postponed to the future, and that part of payment depends on the achievement of pre-established benchmarks by the target company. Such benchmarks are often based on performance measures. The presence of an earnout reverses the information asymmetry problem that usually characterizes acquisitions. While before the closing the sellers have perfect monitoring over their company, and the bidder faces an information asymmetry problem which can be only partially unveiled through a due diligence process, the situation is reversed afterwards. Indeed, at the deal closing the bidder acquires the ownership and the control of the target company. As a consequence, 9

10 it is now the acquirer who potentially possesses more information on the acquired company, while the sellers (of the target company) face the issues and the consequences of information asymmetry. By acquiring control, the buyer becomes in-charge of measuring and reporting the post-acquisition performance of the acquired company. However, there is a risk of opportunistic behavior aimed at reducing or avoiding the contractual payment to the sellers. Indeed, the acquirer may affect the reported performance of the acquired company either by reducing the exerted effort or through accounting policies aimed at reducing the reported performance. The risk that the acquirer will engage in opportunistic behavior is increased by the sellers limited monitoring ability and lack of ability to influence the decisions of the acquirer itself. In case of disagreement about the benchmark outcomes, due to an alleged opportunistic behavior by the acquirer, the sellers may revert to a judicial court. However, the judges themselves may be affected by an information asymmetry that an investigation process may not completely resolve. Also, the judges may be called to decide on matters like the computation of accounting numbers, which are subject to discretion. The limited ex-post monitoring and enforcement possibilities, along with the risk to bear deadweight litigation cost in case of unsuccessful legal trial, will induce the sellers to engage in a rigorous and accurate ex-ante screening of their counterparties before accepting an earnout agreement. We hypothesize that bidder s past earnings quality is one of the signals employed by sellers to evaluate the bidder s trustworthiness. Using earnings management as an inverse proxy for earnings quality, we expect that a higher level of earnings management in the bidder s past financial statements will be assumed by the sellers as an indicator of higher risk of opportunistic behavior after the closing of the deal. Thus, the sellers would be more inclined to accept an agreement including an earnout when their counterparty showed low level of past earnings management. Based on this line of reasoning, we formulate our hypothesis as follows: 10

11 The likelihood of inclusion of earnouts in acquisition deals is negatively associated with the level of earnings management in the acquirer s past financial statements. In the next section we provide a description of the sample and the data used to test our hypothesis. 4. Data and variable description 4.1. Sample composition We collect acquisitions data over the period January 1 st, 2002 December 31 st, 2014 from Thomson ONE Banker, provided by SDC. This initial sample is merged with Compustat to retrieve the relevant accounting information. The final sample includes deals completed during the selected period, and for which the acquirer is a non-financial public company. Information on the total consideration paid, market value of the acquirer prior to the acquisition, and the accounting numbers required for estimating the abnormal accrual measure (i.e. our earnings management proxy) must be available. Table 1 provides details on the composition of the sample. The sample comprises 9,178 deals, out of which 1,138 deals include earnout agreements, i.e. 12.4% of total deals, a significantly higher proportion than the 9% reported by Cadman, et. al. (2014). There is a substantial variation in the number of earnout deals over the sample years, from 64 in 2009 to 117 earnout deals in The sample exhibits also a variation of the earnout deals proportion over the sample period, from 10.1% in 2003 to 15% in [Insert Table 1 here] Panel A of Table 2 provides additional descriptives of the sample. Using Fama-French 12 industries classification, we note a high proportion of earnouts deals in several industries: 22.4% in healthcare, followed by consumer durables (15.3%) and Computers (14.6%). Panel B of Table 2 11

12 classifies the sample according to the target s industry. Also here we note similar industrial concentration: 22,1% in healthcare, and 14.6% in computers. This sample attributes are similar to those reported by Datar et al. (2001). [Insert Tables 2 here] 4.2 Earnings management proxy To estimate the level of the acquirers earnings management we apply the modified-jones model as in Dechow et. al. (1995), adding ROA to the model as suggested by Kothari, Leone and Wasley (2005). We use a cross-sectional specification of the model by time and industry (i.e. quarter or year, consistently with the definition of the time variable). We follow Peasnell, Pope and Young (2000) and Jeter and Shivakumar (1999), who suggest that cross-sectional models seem to be better specified and have higher power than time series estimation models. The model used to estimate accruals is the following: ACCR j,t /TA j,t-1 = α[1/ta j,t-1 ] + β 1 [(ΔREV j,t - ΔREC j,t )/ TA j,t-1 ] + β 2 [PPE j,t /TA j,t-1 ] + β 3 ROA j,t (1) + ε j,t Where j and t designate firm and time, respectively, ACCR is the total accruals, ΔREV is the change in revenues, ΔREC is the change in accounts receivable, PPE designates property, plants and equipment, ROA is return on assets, and TA stands for total assets. Quarterly data are used to estimate the model, defining industries according to the Fama-French 12 industry classification. Abnormal accruals are defined as the actual accruals not explained by the expectation as estimated by model (1), i.e., EM j,t = ACCR j,t /TA j,t-1 E(ACCR j,t /TA j,t-1 ) 12

13 We apply the absolute abnormal accruals of the acquirer averaged over the four quarters preceding the acquisition, which we label EM. We opt for an absolute measure because we are more interested in the reliability of earnings rather than in the direction of their effects. We compute the average of the four quarters in order to capture a behavior that is consistent in time, rather than occasional. However, for robustness tests, we resort to variations of EM measures to validate the reported results. 4.3 Explanatory variables To test our hypothesis, we perform both a univariate and a multivariate analyses. In the latter case, we use the model that includes control variables as reported in prior studies, i.e., Kohers and Ang (2000), Datar et al. (2001), Barbopulos and Sudarsanam (2012), and Cadman et al. (2014). They suggest that earnouts are more frequently used in deals involving private companies and subsidiaries, for which the problem of asymmetry of information is relevant, also because of the absence of a market price for the target. Similarly, earnouts are more likely used for the acquisitions of targets operating in the service or high-tech industries, due to the high growth opportunities and relevant uncertainties related to the role of human capital that characterizes these industries. The presence of a toehold in the target company reduces the probability of an earnout, because the acquirer possibly possesses information on the target, and thus the valuation risk is less relevant than when a toehold is absent. Moreover, the likelihood of observing an earnout in a deal is positively associated with the size of the deal, and negatively associated with the value of the acquirer. We control also for whether the bidder and the target operate in same industry.8 Our model also controls for known determinants of earnings quality which may also affect the likelihood of an earnout. In particular, we control for the audit firm size, typically reported to be 8 There is a lack of consensus on the association between diversifying acquisitions, that is, deals in which the acquirer and the target operate in different industries, and the use of earnouts. Focusing on the US market, Kohers and Ang (2000) and Datar et al. (2001), who focus on acquisitions during the 80 s and the 90 s, report a positive association between cross-industry deals and the use of earnouts. Instead, Cadman et al. (2014), who focus on a more recent sample, find opposite results. 13

14 positively (negatively) associated with earnings quality (earnings management) (e.g., DeAngelo, 1981), and for ROA (e.g., McNichols, 2000; Kothari et al., 2005). In additional tests we control also for the quality of corporate governance, measured as the proportion of non-executive directors of the total number of board directors. We add the corporate governance variable for two reasons: it has been reported to be positively (negatively) associated to earnings quality (earnings management) (e.g., Beasley, 1996; Klein, 2002), and may be argued that corporate governance quality is an indicator of the acquirer reliability. 5. The logit model Since the decision to include an earnout in a deal is a dichotomous variable, a logit model is employed to test our hypothesis. We define and estimate the following model: Logit{Prob(Earnout it =1)} = β 0 + β 1 EM it + Σ β i Controls it + ε it (2) Our dependent variable is Earnout, a dummy variable that takes value 1 if an earnout agreement is part of the deal, zero otherwise. The model includes the following controls: HighTech = a dummy variable that takes value 1 if the target operates in the high-tech sector, 0 otherwise; Service = a dummy variable that takes value 1 if the target operates in the service industry, 0 otherwise; DealValue = log of the transaction price of the deal, including the earnout; MVacquirer = the log of the market value of the acquirer prior to the deal announcement; Subsidiary = a dummy variable that takes value 1 if the target is a subsidiary, 0 otherwise; Private = a dummy variable that takes value 1 if the target is a private company, 0 otherwise; 14

15 SameIND = a dummy variable that takes value 1 if bidder and target have the same two-digits SIC code; Toehold = a dummy variable that takes value 1 if the bidder holds a stake in the target before the acquisition, 0 otherwise; Stock = a dummy variable that takes value 1 if the upfront payment is at least partly in stocks, 0 otherwise; ROA = the return on assets of the acquirer calculated as the ratio of net income to total assets in the quarter (year) preceding the acquisition ; Big4 = a dummy variable that takes value 1 if the bidder s financial statements are audited by one of the big 4 audit firms, 0 otherwise; CorpGov = Acquirer s ratio of non-executive directors over total number of directors at the acquisition time. We also include year fixed effects to control for possible trends or changes in the regulatory or economic environment over time. 6. Empirical analysis 6.1 Descriptive statistics Table 3 provides descriptive statistics for the variables used in our study. Panel A compares the variables statistics between earnout and non earnout deals. The proportion of deals in which the target is a high-tech or a service company is significantly higher in the earnout group (34.6% vs. 22.7%, and 40.0% vs. 31.9%, respectively), consistent with prior literature. The table also exhibits that earnout transactions tend to be characterized by smaller-size acquirers (7.27 vs. 6.46, respectively). Moreover, when the acquisition involves contingent payments, the frequency of private companies is much higher than for the non-earnout sample (75.9% vs. 46.9%, respectively). The opposite seems to hold if we look at the proportion of subsidiaries (22% vs. 37.1%). However, 15

16 filtering out the relative presence of private companies, we notice that among non-private companies the frequency of subsidiaries is higher in deals involving earnouts. Earnout deals seem also to be characterized by a higher frequency of transactions in the same industry (63.9% vs. 60.7%), and by a higher proportion of acquisitions that include common stocks in the upfront payment (26.4% vs. 18.4%, respectively). The frequency of bidders having a toehold in the target company, instead, is significantly lower for earnout deals. Finally, if we compare the average value of our earnings management variable, we observe that it is rather similar in earnout and non-earnout deals (0.043 vs , respectively). Panel B provides the Pearson correlation matrix. The first column of the table reports correlations in line with prior literature. Indeed, earnouts are positively correlated with high-tech and service industries, with the target being a private firm and (weakly) with the target being in the same industry as the acquirer. Earnouts are negatively correlated with the deal value and the market value of the acquirer, with the target being a subsidiary, and with the acquirer having a toehold in the target., Earnouts seem to be (weakly) negatively correlated with the acquirer s auditor size and its corporate governance measure. Such correlation, however, is most probably associated with the high correlations of these variables with the acquirer s size (MVacquirer), as shown in column 6. The table reveals that earnings management (EM) is significantly negatively correlated to the acquirer s size (-0.337), with auditor size (-0.296), with ROA (-0.274) and the corporate governance variable (-0.110). These correlations are in line with those reported in the earnings management literature. The panel also indicates that earnings management is positively correlated with cases where deals include a payment partly with stocks (0.115), consistent with the extant literature (e.g., Erickson and Wang, 1999; Botsari and Meeks, 2008). [Insert Table 3 here] 16

17 6.2 Results The results of our main logit regression models are reported in Table 4. The table provides coefficients values of four versions of Model (2), where different controls are applied as explanatory variables. These regressions are based on 9,178 acquisitions and all four regressions are significant with all Chi2 values above 500. Model 1 is similar to the models applied in prior studies, with the inclusion of the auditor size (Big4) as an additional control variable. The regression coefficient of Big4 is insignificant, possibly due to the lack of variation in the sample, once considered its high correlation with the acquirer size. The variable SameIND, which controls for deals done in the same industry, is found insignificantly different from zero. This is not surprising as prior literature report conflicting results regarding this variable. All other variables in the model are consistent with prior findings. In particular, the negative coefficient of the acquirer size (MVacquirer) (-0.191) suggests that large firms are less inclined to use earnouts in acquisition deals, perhaps because they can likely afford higher misvaluation risk. The probability of a deal to include earnouts is increasing with the target being in the service (0.638) or high-tech (1.046) industries, with deal size (0.102), with the target being either private firms (2.507) or subsidiaries of other firms (1.589). All these results are consistent with those reported by Kohers and Ang (2000), Datar et al. (2001), Barbopulos and Sudarsanam (2012), and Cadman et al. (2014). Model 2 includes the earnings management variable (EM). Consistently with our hypothesis, the coefficient is significantly negative (coefficient of , with s.e. of 0.549), suggesting that the higher the acquirer s earnings management in the four quarters preceding the deal, the less likely the sellers to accept an earnout agreement to be included in the transaction. In Model 3 we control also for the presence of a toehold. We find that it is negatively associated with the probability of inclusion of an earnout, significant at the 10% confidence level (coefficient of and s.e. of 0.200). In model 4 we include the variable Stock, which captures the fact that part of the upfront payment is made in stocks. The latter variable shows a significant positive coefficient (0.268, with 17

18 s.e. of ), suggesting that common stocks and earnouts tend to be used as complementary tools for risk sharing purposes. In both Models 3 and 4, the coefficient on the earnings management variable remains negative and significant, and very close in their magnitude to that in Model 2. This empirical evidence supports our expectation that the sellers do a reverse due diligence on the acquirers, to screen the acquirer trustworthiness. In Table 5 we include an additional control for the corporate governance quality, proxied by the ratio between non-executive directors and total number of board members. The idea behind this inclusion is that the quality of the firm corporate governance may be considered as a corporate quality sign. Therefore it may serve either as a substitute or as a complement to the earnings management signal. The inclusion of this control reduces the sample size to 7,831 deals. Nonetheless, the results are very similar to those reported in Table 4. The corporate governance coefficient is negative but it lacks statistical significance, thus it does not seem to affect the likelihood of the inclusion of an earnout in the acquisition contract. [Insert Table 4 and Table 5 here] In the next section, we test the robustness of our results to various definition of earnings management (applying variations of window period estimations), and to stricter requirements on the sample selection process. 7. Robustness tests 7.1 Changing the time horizon for the measurement of earnings management In our main analysis we measure earnings management over the four quarters preceding the acquisition. Prior literature suggests that acquirers have the incentive to manage earnings upwards in the quarters preceding the deal in order to increase the stock price of their company, and thus to 18

19 reduce the cost of acquiring the target (e.g., Erickson and Wang, 1999; Botsari and Meeks, 2008). Therefore, the quarters immediately preceding the acquisitions may be considered a good indicator of the tendency to carry out earnings management in the presence of a contractual incentive. However, one may question the sellers s ability to observe the last quarter s results before deciding on the inclusion of an earnout in the agreement. In order to address this concern, we first repeat our logit models excluding the last quarter, i.e. the quarter just prior to the completion of the deal. Our results (untabulated) remain qualitatively unchanged, confirming that earnings management is negatively related to the likelihood of inclusion of an earnout in the acquisition contract. We repeat these analyses after calculating our test variable as the average of the earnings management proxy over a longer time horizon, i.e. the six quarters preceding the deal (both including and excluding the last quarter). The results (untabulated) remain qualitatively unchanged. In further tests, we check whether using annual data to measure the attitude to manage earnings by the bidder alters our main findings. Table 6 reports the results of the logit regressions when earnings management is estimated using annual data over a 3-year period preceding the deal. Basically, these revised definitions of earnings management measures do not alter our reported results. The earnings management variable is found to be negatively and significantly related to the likelihood of inclusion of an earnout in the contract (-0.509, s.e ). Similar results are shown when we also control for the corporate governance quality (Table 7). [Insert Table 6 and Table 7 here] Our main findings (untabulated) hold also when we exclude the last year prior to the acquisition and, therefore, consider only the earnings management carried out in years t-2 and t-3. 19

20 7.2 Propensity Score Matching In order to further check the validity of our hypothesis, we compute a t-test for the differences in the mean level of earnings management for the group of deals in which earnouts are included and for those in which they are not, using a Propensity Score Matching procedure (PSM), as described in Becker and Ichino (2002). The idea behind this methodology is to compare the level of earnings management (our outcome variable) between a group of treated subjects (earnout users) and a group of non-treated subjects (non-earnout users) selected as similar as possible to the treated observations. We follow this procedure to reduce the possibility that the comparison of subjects that are inherently different could bias the results. The propensity score is computed taking into consideration the following deal characteristics: the acquirer size, defined as total assets, the deal value, that is the total value of the transaction, the target operating in the high tech or service industry, and a dummy variable for the period before the financial crisis of The model is parsimonious, yet it satisfies the balancing hypothesis 10. Moreover, in order to improve the quality of the match, only the observations belonging to the common support of the treated and non-treated are used in the comparison. [Insert Table 8 here] The results of the PSM analysis, performed using the nearest neighbor method, are shown in Table 8. The level of earnings management is estimated over the four quarters preceding the deal, as in our main analysis. Consistently with our hypothesis, the acquirers of deals in which earnouts are 9 The size of the acquirer and the size of the deal are defined in terms of terciles of their sample distribution. Each observation is assigned to one of three groups: high, medium or small bidder size, and high, medium or small deal size. This partitioning is used for the computation of the propensity score. 10 If the balancing hypothesis is satisfied, treated observations and non-treated observations with the same propensity score share the same distribution of observable and unobservable characteristics. 20

21 used show a lower level of earnings management compared to their matched peers (i.e. non-users). The difference is significant at the 1% level (t-values of to ). 7.3 Introducing additional filters In line with Datar et al. (2001) and Cadman et al. (2014), in our main analyses we did not impose any restriction on the percentage of shares acquired in the deal. By using more restrictive filters on the sample, however, our results remain qualitatively the same. [Insert Tables 9 and 10 here] Table 9 reports the results obtained considering only the subsample of deals in which at least 20% of the shares of the target is acquired. In a setting like the US, in which ownership is generally fragmented, such percentage is usually sufficient to gain control on a company. The results in terms of signs and levels of significance remain substantially the same. In table 10 we apply an even more stringent filter, as we consider only deals in which the toehold in the target before the acquisition is lower than 50%, and the ownership of the acquirer after the acquisition is at least 90%. Our main results confirm to be robust, suggesting that past earnings management of the acquirer is negatively associated with the probability of inclusion of an earnout in the deal. 8. Conclusions Earnouts are contractual agreements that condition part of the payment of an acquisition to the future performance of the acquired company. 21

22 Earnouts are beneficial for acquirers as they reduce the information asymmetry that typically affects acquisitions, and act as a selection mechanism on the quality of the target company. They are advantageous also to the sellers, as they reduce the risk of adverse selection for the target company, thus allowing to negotiate a higher consideration. However, this benefit comes at a price, because, after the closing of the deal, the sellers bear not only the risk that the acquired company will not meet the earnout requirements, but also the risk that the acquirer will act opportunistically to reduce, or even avoid, the accounting based benchmarks for determining the contingent payment. Such opportunism can be implemented by either managing the accounting numbers, or by reducing the effort in directing the target s activities. Legal disputes on the actual realization of the performance levels stated by the earnout are indeed frequent. However, anecdotal evidence indicates that the risk to lose the legal case and to bear the cost of the litigation for the sellers is not trivial. Given the above-mentioned risks related to the ex post enforcement of earnout contracts, sellers who intend to agree for earnout contracts are likely to engage in ex-ante screening of the trustworthiness of their counterparties. To this purpose, past earnings quality of the acquirer may be used by the sellers as a signal of the former s reliability. Therefore, we hypothesize that the decision to include an earnout agreement in an acquisition contract is positively related to the acquirer s past earnings quality. We test this hypothesis on a sample of 9,178 deals completed in US between 2002 and In order to capture the quality of earnings of the acquirer, we rely on the absolute value of the earnings management proxy, computed using a variation of the modified Jones model (Dechow et al., 1995). A lower absolute value of our earnings management proxy indicates a higher level of earnings quality, and therefore of higher level of trustworthiness. After controlling for the determinants that prior studies have associated with the use of earnouts, through a logit model we show that there is a negative and significant relation between earnings management and the likelihood of inclusion of an earnout in an acquisition deal. Overall, our results 22

23 validate the hypothesis that the acquirer s past earnings quality is used as a signal by the seller to evaluate the former s reliability. Our paper contributes to the literature on earnouts by showing that the bidder s trustworthiness is a significant determinant of the decision by the seller to accept such contractual agreements. Our paper contributes also to the literature on the relation between M&As and earnings management, that so far has mainly been focused on the managerial incentives to manage earnings around the acquisition time to improve the results of the negotiation. Our study is the first to show that past earnings quality of the counterparty is a relevant determinant of the structure of acquisition contracts. 23

24 References Allee, K., Wangerin, D. (2013). Auditors Role in Financial Contracting: Evidence from SFAS 141 (R). Working paper. Ball, R., Shivakumar, L. (2008). Earnings quality at initial public offerings: managerial opportunism or public-firm conservatism. Journal of Accounting and Economics, 45, Beasley, M. S. (1996). An Empirical Analysis of the Relation between the Board of Director Composition and Financial Statement Fraud. The Accounting Review, 71, Becker, S. O., Ichino, A. (2002). Estimation of average treatment effects based on propensity scores. The Stata Journal 2(4). Barbopoulos, L., Sudarsanam, S. (2012). Determinants of earnout as acquisition payment currency and bidder s value gains. Journal of Banking & Finance, 36(3). Botsari, A., Meeks, G. (2008). Do acquirers manage earnings prior to a share for share bid?. Journal of Business Finance & Accounting, 35(5-6). Cadman, B., Carrizosa, R., Faurel, L. (2014). Economic Determinants and Information Environment Effects of Earnouts: New Insights from SFAS 141 (R). Journal of Accounting Research, 52(1). Cain, M. D., Denis, D. J., Denis, D. K. (2011). Earnouts: A study of financial contracting in acquisition agreements. Journal of Accounting and Economics, 51(1). Crimmins, P. M., Gray, B, Waller, J., Brown, M. (2010). Earn-outs in M&A Transactions. The M&A Journal, 10(10). Datar, S., Frankel, R., Wolfson, M. (2001). Earnouts: The effects of adverse selection and agency costs on acquisition techniques. Journal of Law, Economics, and Organization, 17(1). DeAngelo, L.E. (1981). Auditor size and audit quality. Journal of Accounting and Economics, 3,

25 Dechow, P. M., Sloan, R. G., Sweeney, A. P. (1995). Detecting earnings management. Accounting Review, 70(2). DuCharme, L. L., Malatesta, P. H., Sefcik, S. E. (2004). Earnings management, stock issues, and shareholder lawsuits. Journal of financial economics, 71(1). Erickson, M., Wang, S. (1999). Earnings management by acquiring firms in stock for stock mergers. Journal of Accounting and Economics, 27(2). Fox, D., Wolf, D. E. (2010). An Earnout Is In the Eye of the Beholder. Investment Dealers' Digest, 76(4). Gong, G., Louis, H., Sun A. (2008). Earnings management, lawsuits, and stock-for-stock acquirers market performance. Journal of Accounting and Economics, 46, Healy, P. and J. Whalen, (1999). A Review of the Earnings Management Literature and its Implication for Standard Setting, Accounting Horizons 13, Jensen, M. (2005). Agency costs of overvalued equity. Financial Management, 34, Jeter, D. C., Shivakumar, L. (1999). Cross-sectional estimation of abnormal accruals using quarterly and annual data: Effectiveness in detecting event-specific earnings management. Accounting and Business Research, 29(4). Jones, J. J. (1991). Earnings management during import relief investigations. Journal of Accounting Research, 29(2). Klein, A. (2002). Audit Committee, Board of Director Characteristics and Earnings Management. Journal of Accounting and Economics, 33, Kohers, N., Ang, J. (2000). Earnouts in Mergers: Agreeing to Disagree and Agreeing to Stay. The Journal of Business, 73(3). Kothari, S., Leone, A., Wasley, C., (2005), Performance Matched Discretionary Accrual Measures, Journal of Accounting & Economics 39, Louis, H. (2004). Earnings management and the market performance of acquiring firms. Journal of Financial Economics, 74(1). 25

26 McNichols, M.F. (2000). Research design issues in earnings management studies. Journal of Accounting and Public Policy, 19, Morrison and Foerster LLP (2012). M&A Leaders Survey. Press Releases 5/15/2012. Available at: Peasnell, K. V., Pope, P. F., Young, S. (2000). Detecting earnings management using crosssectional abnormal accruals models. Accounting and Business Research, 30(4). Ragozzino, R., Reuer, J. J. (2009). Contingent earnouts in acquisitions of privately held targets. Journal of Management, 35(4). Shannon, K., Reilly, M. (2011). Post-Closing Earnouts in M&A Transactions: Avoiding Common Disputes. American Bar Association Newsletter. Subramanyam, K. R. (1996). The pricing of discretionary accruals. Journal of Accounting and Economics, 22(1). 26

27 Table 1: Sample description - Deals by year This table shows the distribution of the sample deals by year over the period Year Nr deals Nr deals including earnout % % % % % % % % % % % % % % Total % 27

28 Table 2: Sample description - Deals by industry This table provides descriptive statistics on the sample deals by industry. In panel A deals are classified according to the industry of the acquirer, while in panel B deals are classified according to the industry of the target. Deals in which the acquirer operates in the financial industry are excluded. Panel A: Acquirer's industry Industry Nr deals Nr deals including earnout % Consumer NonDurables % Consumer Durables % Manufacturing % Oil, Gas, and Coal Extraction and Products % Chemicals and Allied Products % Computers, Software, and Electronic Equipment % Telephone and Television Transmission % Utilities % Services and Retail % Healthcare, Medical Equipment, and Drugs % Other (Mines, Constr, Entertainment, Transp) % Total % Panel B: Target's industry Industry Nr deals Nr deals including earnout % Consumer NonDurables % Consumer Durables % Manufacturing % Oil, Gas, and Coal Extraction and Products % Chemicals and Allied Products % Computers, Software, and Electronic Equipment % Telephone and Television Transmission % Utilities % Services and Retail % Healthcare, Medical Equipment, and Drugs % Financial % Other (Mines, Constr, Entertainment, Transp) % Total % 28

29 Table 3: Descriptive statistics and correlations This table provides descriptive statistics on the main variables used in the analysis. HighTech and Service are dummy variables that take value 1 if the target operates in the high tech or the service sector, respectively. DealValue is the log of the transaction price of the deal. MVacquirer is the log of the market value of the acquirer prior to the deal announcement. Subsidiary and Private are dummy variables that take value 1 if the target is a private company or a subsidiary, respectively. SameIND is a dummy variable that takes value 1 if bidder and target have the same two-digits SIC code. Toehold is a dummy variable that takes value 1 if the bidder holds a stake in the target before the acquisition. Stock is a dummy that takes value 1 if the upfront payment is at least partly in stocks. EM is the mean of the abnormal accrual measure in the four quarters preceding the deal. Big4 is a dummy variable that takes value 1 if the audit firm is belongs to one of the big four. ROA is the ratio of net income over total assets. CorpGov is the ratio of non-executive directors over the total number of directors of the bidder. Panel A provides, for each variable, mean and standard deviation, detailed for the subsample of earnout users, non earnout users, and in the whole sample. Panel B provides correlations. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. Panel A: Means and standard deviations Non-earnout users Earnout users Total Mean SD Mean SD Sign. Mean SD HighTech 22.70% 41.90% 34.60% 47.60% *** 24.10% 42.80% Service 31.90% 46.60% 40.00% 49.00% *** 32.90% 47.00% DealValue *** MVacquirer *** Subsidiary 37.10% 48.30% 22.00% 41.40% *** 35.20% 47.80% Private 46.90% 49.90% 75.90% 42.80% *** 50.50% 50.00% SameIND 60.70% 48.80% 63.90% 48.10% ** 61.10% 48.70% Toehold 6.50% 24.60% 2.50% 15.80% *** 6.00% 23.70% Stock 18.40% 38.80% 26.40% 44.10% *** 19.40% 39.60% Big % 39.60% 73.10% 44.40% *** 79.60% 40.30% ROA -0.50% 47.30% 0.30% 6.80% -0.40% 44.30% CorpGov 38.20% 37.90% 39.20% 37.20% 38.30% 37.80% EM