Meeting and Beating Analysts Forecasts and Takeover Likelihood

Meeting and Beating Analysts Forecasts and Takeover Likelihood Abstract Prior research suggests that meeting or beating analysts earnings expectations has implications for both equity and debt markets: lower debt costs, higher valuation, and a higher probability of a rating upgrade. This paper examines whether meet-beat behavior also serves as a takeover deterrent. We find that meeting or beating analyst earnings forecasts reduces the probability of a firm being the object of a takeover attempt as well as the probability of the ultimate success of the attempt. However, we also find that this deterrent effect does not exist when firms are more likely to have met or exceeded expectations via earnings management. 29

Meeting and Beating Analysts Forecasts and Takeover Likelihood Introduction Anecdotal evidence and academic research have long acknowledged the importance that managers and financial markets place on meeting earnings benchmarks set by analysts forecasts. The literature on earnings surprises is extensive, incorporating important questions such as the incentives for and consequences of meeting or beating analysts forecasts and the relevance of earnings surprises for future performance. Kasznik and McNichols (2002) and Lopez and Rees (2002) provide evidence that markets rewards firms for meeting and beating analysts expectations in both short-term and long-term valuations. Jiang (2008) provides evidence that firms beating earnings benchmarks have a higher probability of rating upgrades, have lower cost of debt. These papers suggest that meeting or beating analysts expectations (hereafter, MBE) has implications for both equity and debt markets. In addition to market related incentives, Matsunaga and Park (2001) found a significant incremental effect on CEO s bonus where managers failed to meet quarterly earnings forecasts. There is another incentive for managers to pay attention to whether the firm meets or exceeds expectations, one which is largely unexplored in the literature: its potential effect as a takeover deterrent. The literature exploring the characteristics of takeover targets and takeover likelihood prediction models is extensive (e.g. Bebchuk, Cohen, & Wang, 2014; Cremers, Nair, & John, 2009). We contribute to both the MBE and takeover literatures by investigating the connection between a firm s MBE behavior and its chances of becoming a takeover target. Generally, acquirers choose targets when they expect to generate additional value either

(1) through realizing potential synergies 1 or (2) by eliminating inefficiencies in the target s management (Stein, 1988). Both of these reasons suggest that current performance is lacking compared to perceived value after the merger. Consistent with this idea, Healy, Palepu and Ruback (1992) provide evidence that merged firms show significant improvements in asset productivity related to their industries, leading to higher operating cash flow returns. The authors also find that there is a strong positive relation between post-merger increases in cash flows and abnormal stock returns at merger announcements, indicating that expectations of economic improvements underlie the equity revaluations of the merging firms. These studies provide evidence that takeover attempts are primarily driven by poor performance of the targeted firms. It follows that to the extent MBE behavior signals good performance, it could deter the likelihood of a takeover threat as well as the success of a takeover attempt. However, a considerable amount of research showing that some firms achieve earnings benchmarks through the use of earnings management (e.g. Burgstahler and Eames, 2006; Brown, 2001; Degeorge, Patel, and Zeckhauser, 1999) or other myopic behaviors (e.g. Ayers, Jiang & Yeung 2006; Bhojraj, Hribar, Picconi, and Mcinnis, 2009; Harris, Shi, and Xie, 2010). Thus, if firms MBE behavior is achieved through earnings management or other myopic behavior, it may not effectively deter takeover threats. The ultimate effect of MBE on takeovers is therefore an empirical question. Our results are consistent with MBE behavior serving as a takeover deterrent and lowering the likelihood of a takeover attempt. However, we also show that when MBE is likely to be the result of earnings manipulation, the likelihood of a takeover increases. Our results add to the literature on the benefits of meeting and beating analysts forecasts by identifying 1 Synergies created through a merger will either reduce costs (economies of scale in research and development, manufacturing, sales and marketing, distribution, administration etc.) or enhance revenues (cross-selling of products, expanded market share, or higher prices arising from reduced competition).

consequences of MBE for the market of corporate control, including a potential consequence of achieving MBE through earnings manipulation. The remainder of the paper is organized as follows. In the next section, we discuss the related literature and present our hypotheses. The third section describes our sample, data, and research design. In the fourth section, we present our main empirical results. We discuss our conclusions and potential extensions of the research in the final section. Background and Hypotheses Development Academics and practitioners alike have long been interested in what motivates acquirers to pick certain targets. Numerous studies have examined the determinants of takeover targets and explored takeover likelihood prediction models, including Ambrose and Megginson (1992); Shleifer and Vishny (2003); Rossi and Volpin (2004); Powell and Yawson (2005); Brar, Giamouridis, and Liodakis, (2009); and Palepu (1986). Several hypotheses that explain significant determinants of a firm s likelihood of acquisition have been put forth in the literature. For example, Palepu (1986) and Brar et al. (2009) discuss inefficient management, mismatch between recent growth and available financial resources, industry disturbance, undervaluation, small size, low market-to-book or price/earnings ratios, and leverage characteristics that may affect the probability of a firm becoming a takeover target. Thus, consistent with the findings in Healy et al. (1992), a firm that currently appears weak but has the potential for significantly improved performance may find itself a takeover target. MBE behavior, however, can signal good current performance, potentially decreasing the likelihood of a takeover attempt. Kasznik and McNichols (2002) find that firms meeting expectations experience a sequence of future earnings that is significantly greater than that of

firms that do not. This suggests that MBE behavior signals higher future earnings, increasing the current value of the firm; several papers (e.g. Kazsnik and McNichols, 2002; Lopez and Rees, 2002) find evidence consistent with this prediction. Since how well a firm performs is inversely related with the firm s need for external disciplining, we would expect a negative relation between MBE and the firms chances of becoming a takeover target. This leads to our first hypothesis: H1: Meeting or beating analysts expectations is negatively related to the likelihood of a takeover threat. On the other hand, if this performance is the result of managers manipulation of earnings, it might actually increase the likelihood of a takeover attempt. In a survey of over 400 managers, Graham, Harvey & Rajgopal (2005) found that 78% of managers use earnings management techniques that lower shareholder value to benefit from MBE. Bhojraj et al. (2009) provide evidence that firms that barely meet or beat analysts expectations engage in myopic behavior. Other studies (e.g., Ayers et al., 2006; Burgstahler and Eames, 2006; Degeorge et al., 1999) indicate that manipulators usually meet or narrowly beat analysts forecasts. Harris et al. (2010) provide evidence that firms that barely meet or beat earnings benchmarks are more likely to commit fraud. Keung, Lin, and Shih (2010) show that earnings response coefficients (ERC) for small earnings surprises are significantly lower than the ERCs for earnings surprises in adjacent ranges. Since acquirers are sophisticated investors, they may be able to see through such earnings management; MBE behavior in the form of a small earnings surprise may thus increase the likelihood of a takeover attempt consistent with the disciplining mechanism of takeovers. It is possible, however, that a small earnings surprise could increase the negative relation between MBE behavior and the probability of a takeover threat. Managers may be smoothing

earnings in order to signal positive future performance; this would naturally decrease the likelihood of the firm becoming a takeover target. Thus, we do not make a directional prediction for our second hypothesis: H2: Posting a small, positive earnings surprise affects the negative relation between meeting or beating analysts expectations on the likelihood of a takeover threat. If MBE behavior affects the likelihood of a takeover threat, then it may also affect the outcome (success or failure) of a takeover attempt. To the extent that MBE behavior predicts future earnings increases, it may be more difficult to acquire the firms based on the present value of future cash flows, consistent with the predicted effect on takeover threat in the first place. This leads to our third hypothesis: H3: Meeting or beating analysts expectations is negatively related to the likelihood of takeover success. However, MBE behavior achieved with a small earnings surprise may similarly affect this relation. If a small surprise is perceived as a sign of inefficient management or myopic behavior, the firm may be a good candidate for a takeover, increasing the likelihood of success of any takeover attempt. On the other hand, Skaife and Wangerin (2013) show that deals involving targets with low-quality financial reporting are more likely to be terminated. Thus, we once again do not make a directional prediction for our fourth hypothesis, as follows: H4: Posting a small, positive earnings surprise affects the negative relation between meeting or beating analysts expectations on the likelihood of takeover success. Data and Methodology We draw our data from the Compustat, IBES, and SDC Platinum databases. We start with all firms with available data in Compustat for the period 1989-2012, excluding utilities and

financials (177,969 firm-year observations). We merge this sample with IBES for analyst forecast information, and we keep only those observations that have available data to compute the earnings surprise. 2 Our IBES/Compustat sample has 74,711 firm-year observations for 9,717 firms. We winsorize all control variables to minimize the effect of outliers. We then create a dummy variable, TARGET, to identify firms included in SDC as firms that experienced a takeover threat. Figure 1 illustrates our variable measurement timeline, and Table 1 contains a summary of our sample selection procedure. [Insert Figure 1 and Table 1 About Here] We draw our control variables from Palepu (1986). Panel A of Table 2 presents descriptive statistics for the overall sample and for the subsample of firms that experienced a takeover threat (the takeover sample). In most cases, these variables are comparable in magnitude with previous literature (e.g. Ambrose and Megginson, 1992). The firms in our sample meet or beat analysts expectations about 58% of the time. The average return on equity (ROE) is negative (-4.2%) but the median ROE is positive (8.4%), indicating that the data is skewed to the right. The average ratio of cash to total assets (LIQUID) is 0.205 and the average leverage of the firm in our sample is 0.484. The average growth of sales (GROWTH) for our sample is 0.102 and the dummy GRDUMMY, which captures the growth resource mismatch as described in Palepu (1986), has a value of 0.209. The characteristics of firms in the takeover subsample indicate that some of these variables may be significantly different for targets vs nontargets; we conduct a more specific test of these differences in Table 3 below. Panel B of Table 2 presents the correlation matrix. Consistent with the hypotheses put forth in Palepu (1986), we observe a positive and significant correlation between the growth 2 We compute earnings surprise using the Summary and Actual files in IBES, where earnings surprise is defined as the difference between the actual earnings per share (EPS) and the latest mean analyst forecast.

dummy variable and the target dummy (indicating that firms with a larger growth-resource mismatch are more likely to be taken over) and a negative relation between performance, size, PE ratios and the target dummy (i.e. larger firms with higher ROE and high PE ration are less likely to be taken over). Although we show a positive correlation between the magnitude of unexpected earnings and the likelihood of being a takeover target, this result is based on the full sample without differentiating MBE firms and does not control for other determinants of takeover attempts. [Insert Table 2 About Here] Table 3 presents the results of mean difference tests for the characteristics of target and non-target firms. The results are consistent with the hypotheses put forth in Palepu (1986). Specifically, firms that are subject to takeover threats are significantly smaller, and have significantly lower ROE, lower profit and lower returns (i.e., have more inefficient management as reflected by performance) compared to their non-target counterparts. Furthermore, firms with lower PE ratios and higher growth dummies are likely acquisition targets, supporting the PE and growth-resource imbalance hypotheses proposed by Palepu (1986), respectively. [Insert Table 3 About Here] Results MBE and Subsequent Takeover Threats Our first hypothesis focuses on the association between meeting and beating analysts forecasts and the likelihood of a firm becoming a takeover target in the following year. Our models are based on the intersection of the Palepu (1986) paper identifying the determinants of takeover targets and the Lopez and Rees (2002) study investigating the effect of beating and missing analysts forecasts on the information content of unexpected earnings. We first replicate

Palepu (1986) and then add the MBE variables: unexpected earnings (UE); and BEAT, which takes the value of 1 if the earnings surprise is greater than or equal to zero. Prob (TARGET i,t+1 =1) = α + β 1 GRDUMMY i,t + β 2 GROWTH i,t + β 3 INDUSTRY i.t + β 4 LEV i,t + β 5 LIQUID i,t + β 6 MTB i,t + β 7 PE i,t + β 8 ROE i,t + β 9 SIZE i,t + (1) Prob (TARGET i,t+1 =1) = α + β 1 UE i,t + β 2 BEAT i.t + β 3 UE i,t *BEAT i,t + β 4 GRDUMMY i,t + β 5 GROWTH i,t + β 6 INDUSTRY i,t + β 7 LEV i,t + β 8 LIQUID i,t + β 9 MTB i,t + β 10 PE i,t + β 11 ROE i,t + β 12 SIZE i,t + (2) The dependent variable, TARGET, is a dummy variable that takes a value of 1 if a firm is subject to a takeover threat within the year subsequent to the earnings announcement at time t, and 0 otherwise. Following Palepu (1986), we control for several takeover determinants, including growth (GROWTH), resource mismatch (GRDUMMY), industry effects (INDUSTRY), leverage (LEV), liquidity (LIQUID), valuation ratios (MTB, PE), firm performance (ROE), and firm size (SIZE). 3 Our main variable of interest is the interaction variable between unexpected earnings and the dummy that signifies weather the firm achieved MBE and is denoted by (UE*BEAT) and all other independent variables are measured before the threat occurred (see Figure 1). We present the test of our first hypothesis in Table 4. In model 1 we replicate the results of Palepu (1986). Consistent with his study, GRDUMMY (0.317***) is positive and significant, which means that that a mismatch between growth opportunities and resources available increases the likelihood of a takeover attempt. We also find that SIZE significantly lowers the likelihood of a takeover attempt. While overall our results are consistent with Palepu (1986), 3 Detailed definitions of each variable are included in the appendix.

some differences exist. These differences are likely due to the different samples used in the studies. Palepu (1986) covers the years 1971-1979, while our time frame spans 1990-2012. Additionally, with the introduction of the SDC data base, the number of takeover attempt observations is also considerably larger. The Model 2 in Table 4 is our main contribution to the literature and the focus of our hypothesis. We find that the likelihood of a takeover attempt is significantly lowered when the firms meet or beat analysts expectations; UE*BEAT is significantly negative. Since our models utilize logit regressions, we separately calculate the marginal effect of the interaction variable (Ai and Norton, 2003). 4 The marginal effect of UE*BEAT (-0.024***) is also negative and significant. Our control variables are consistent with our first model. For example, the GRDUMMY is still significant and positive, and LEV and LIQUID still increase the likelihood of a takeover-threat. Furthermore, goodness of fit as measured by R 2 increases from 29% to 36% when we include our variable of interest. Table 4 Panel B tests our second hypothesis; because MBE can also be achieved by earnings management and myopic behavior, we test if the means of achieving MBE makes a difference in takeover likelihood. Consistent with prior research (Bhojraj et al, 2009), we investigate small earnings surprises where the surprise is larger than 0.00 but is smaller or equal to 0.01. Our results indicate that, in contrast to the previous results, when MBE is achieved by a small earnings surprise, our main variable of interest (UE*BEAT) and its marginal effect are both positive and significant. These results suggest that acquirers may differentiate firms based on how MBE behavior is achieved. These results extend and add to the findings of Skaife and 4 Ai and Norton (2003) explain why examining marginal effects, rather than focusing on coefficient estimates of the interaction effect, is more appropriate for nonlinear models and provide formulas for their computation. We refer the reader to their paper for more detailed information.

Wangerin (2013); we focus on the signal of meeting and beating analysts forecasts while Skaife and Wangerin (2013) primarily focus on earnings quality. MBE and Takeover Success In hypothesis one and two, we provide evidence that while MBE based on performance decreases the likelihood of receiving a takeover threat, MBE that may stem from myopic behavior increases it. Hypothesis 3 and 4 investigate whether MBE behavior affects the success of a takeover attempt. Consistent with our first two models, we estimate two more models where the dependent variable (SUCCESS) takes the value of 1 if the deal is completed successfully and 0 otherwise. Prob (SUCESS=1) =Logit( α + β 1 GRDUMMY i,t + β 2 GROWTH i,t + β 3 INDUSTRY i.t + β 4 LEV i,t + β 5 LIQUID i,t + β 6 MTB i,t + β 7 PE i,t + β 8 ROE i,t + β 9 SIZE i,t + (3) Prob(SUCESS=1)=logit( α + β 1 UE i,t + β 2 BEAT i.t + β 3 UE i,t *BEAT i,t + β 4 GRDUMMY i,t + β 5 GROWTH i,t + β 6 INDUSTRY i,t + β 7 LEV i,t + β 8 LIQUID i,t + β 9 MTB i,t + β 10 PE i,t + β 11 ROE i,t + β 12 SIZE i,t + (4) We present the test of our third hypothesis in Table 5. Model 3 again follows the Palepu (1986) study and tests whether the determinants of takeover also affect the likelihood of success. The results indicate that GRDUMMY increases the likelihood of success while SIZE decreases the likelihood of success. ROE also increases the likelihood of success. Model 4 provides the main test of interest. Consistent with our previous results, the marginal effect of UE*BEAT is significantly negative, indicating that MBE also decreases the probability of a successful takeover attempt. This finding provides evidence of an additional

incentive to meet or beat analysts forecasts. 5 Table 5 Panel B investigates the effect of MBE when achieved by a small earnings surprise. The results indicate that UE*BEAT and its marginal effect increase the likelihood of takeover success, consistent with the disciplining role of takeovers. In a recent paper, Skaife and Wangerin (2013) argue 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 to be breached therefore these deals are more likely to be terminated. 6 We take a different approach and argue that myopic behavior of the managers in achieving MBE increases the incentives to successfully complete the deal consistent with the disciplining role of takeovers. Conclusions We investigate the connection between meeting and beating analysts forecasts and firms chances of becoming a takeover target. To our knowledge this is the first paper that connects two important streams of research, on takeovers and on MBE behavior. While several papers provide evidence that MBE signals good performance and thus may serve as a takeover deterrent, other papers suggest MBE can be achieved myopically or by earnings management and may encourage the disciplining effect of takeovers. Our results provide evidence that this is indeed the case: when stemming from performance, MBE is a deterrent but when from potentially myopic behavior MBE encourages the disciplining of management. We find similar 5 Prior research (e.g., Doyle, Lundholm & Soliman 2006; Freeman & Tse, 1992; Zeibart, 1990) provide evidence that firms that post large earnings surprises have larger subsequent earnings surprises and returns that persist for a much longer duration. These firms also have stronger growth in cash flows and earnings when compared to firms with smaller surprises. 6 Although Skaife and Wangerin (2013) show that low earnings quality increases the likelihood of a failure, their table 4 results provide evidence that earnings quality increases the deal premium. The authors state that one explanation for this finding is that target firms share trade at a discount because of a higher cost of capital. In this paper, we argue that this is consistent with the disciplining role of takeovers. If the acquirers believe that they can better run the firm and lower the cost of capital, this might provide incentives to complete the deal.

results for the success of takeover attempts: while performance-based MBE lowers the success rate of the deals, myopic MBE increases it. Our study has one important limitation. We rely on the findings of prior research and argue that meeting and beating earnings by a small margin represents myopic behavior. This finding has been shown in so many influential papers that we have no reason to believe that it would not hold in our sample.

References Ai, C., and E. C. Norton. (2003). Interaction Terms in Logit and Probit Models. Economics Letters 80 (1): 123-129. Ambrose, B. W., and W. L. Megginson. (1992). The Role of Asset Structure, Ownership Structure, and Takeover Defenses in Determining Acquisition Likelihood. The Journal of Financial and Quantitative Analysis 27 (4): 575. Ayers, B. C., J. (Xuefeng) Jiang, and P. E. Yeung. (2006). Discretionary Accruals and Earnings Management: An Analysis of Pseudo Earnings Targets. The Accounting Review 81 (3): 617 652. Bebchuk, L., A. Cohen, and C. C. Y. Wang. (2014). Golden Parachutes and the Wealth of Shareholders. Journal of Corporate Finance 25: 140 154. Bhojraj, S., P. Hribar, M. Picconi, and J. McInnis. (2009). Making Sense of Cents: An Examination of Firms That Marginally Miss or Beat Analyst Forecasts. The Journal of Finance 64 (5): 2361 2388. Brar, G., D. Giamouridis, and M. Liodakis. (2009). Predicting European Takeover Targets. European Financial Management 15 (2): 4 450. Brown, L. (2001). A temporal analysis of earnings surprises: Profits versus losses. Journal of Accounting Research 39 (2): 221 242. Burgstahler, D., and M. Eames. (2006). Management of Earnings and Analysts Forecasts to Achieve Zero and Small Positive Earnings Surprises. Journal of Business Finance and Accounting 33 (5-6): 633 652. Cremers, K. J. M., V. B. Nair, and K. John. (2009). Takeovers and the Cross-Section of Returns. Review of Financial Studies 22 (4) (4): 1409 1445. Degeorge, F., J. Patel, and R. Zeckhauser. (1999). Earnings Management to Exceed Thresholds. The Journal of Business 72 (1): 1 33. Doyle, J. T., R. J. Lundholm, and M. T. Soliman. (2006). The Extreme Future Stock Returns Following I/B/E/S Earnings Surprises. Journal of Accounting Research 44 (5): 849 887. Freeman, R. N., and S. Y. Tse. (1992). A Nonlinear Model of Security Price Responses to Unexpected Earnings. Journal of Accounting Research (2): 185. Graham, J. R., C. R. Harvey, and S. Rajgopal. (2005). The economic implications of corporate financial reporting. Journal of Accounting and Economics 40 (1-3): 3 73.

Harris, D., L. Shi, and H. Xie. (2010). Is meeting and beating earnings benchmarks associated with accounting fraud. Unpublished paper, Syracuse University. Healy, P., K. Palepu and R. Ruback. (1992). Does corporate performance improve after mergers? Journal of Financial Economics 31:135-175. Jiang, J., (2008). Beating Earnings Benchmarks and the Cost of Debt. The Accounting Review 83: 377-416. Kasznik, R., and M. F. McNichols. (2002). Does Meeting Earnings Expectations Matter? Evidence from Analyst Forecast Revisions and Share Prices. Journal of Accounting Research 40 (3): 727 759. Keung, E., Z.-X. Lin, and M. Shih. (2010). Does the Stock Market See a Zero or Small Positive Earnings Surprise as a Red Flag? Journal of Accounting Research 48 (1): 91 121. Lopez, T., and L. Rees. (2002). The effect of beating and missing analysts forecasts on the information content of unexpected earnings. Journal of Accounting, Auditing & Finance 17 (2): 155 184. Matsunaga, S., and C. Park. (2001). The effect of missing a quarterly earnings benchmark on the CEO s annual bonus. The Accounting Review 76 (3): 313 332. Palepu, K. G. (1986). Predicting takeover targets. Journal of Accounting and Economics 8 (1): 3 35. Powell, R., and A. Yawson. (2005). Industry aspects of takeovers and divestitures: Evidence from the UK. Journal of Banking & Finance 29 (12): 15 40. Rossi, S., and P. F. Volpin. (2004). Cross-country determinants of mergers and acquisitions. Journal of Financial Economics 74 (2): 277 4. Shleifer, A., and R. W. Vishny. (2003). Stock market driven acquisitions. Journal of Financial Economics 70 (3): 295 311. Skaife, H. A., and D. D. Wangerin. (2013). Target Financial Reporting Quality and M&A Deals that Go Bust. Contemporary Accounting Research (2): 719 749. Stein. J. (1988). Takeover Threats and Managerial Myopia. Journal of Political Economy 96 (1): 61-80. Ziebart, D. A. (1990). The association between consensus of beliefs and trading activity surrounding earnings announcements. The Accounting Review 65 (2): 477 488.

APPENDIX Variables BEAT GRDUMMY GROWTH INDUSTRY LEV LIQUID MTB PE ROE SIZE SUCCESS TARGET t+1 UE Definition Takes the value of 1 if the surprise calculated as the difference between actual earnings minus the mean analyst forecast for the year is bigger or equal to 0.00 or 0 otherwise. Takes the value of 1 if the firm has a combination of either low growthhigh liquidity-low leverage or high growth-low liquidity-high leverage. The dummy is set to zero for all other combinations. Change in sales scaled by assets. The industry dummy is a 0/1 variable. It is assigned a value of 1 if at least 1 acquisition occurred in firms two digit SIC industry code during the sample period. Total Liabilities divided by shareholders equity. Cash scaled by total assets. Market to book ratio defined as the market value of equity divided by the book value of equity. Price at the end of the year divided by the earnings per share. Return on Equity calculated by dividing income before extra-ordinary items by shareholders equity. Total Assets. A dummy variable that takes the value of 1 if a deal has been reached and 0 otherwise. This is the STATUS1 variable in SDC data base. Dummy variable that takes the value of 1 if a deal was announced in SDC within the year following the earnings announcement date at time t (see Figure 1). Unexpected earnings calculated as the difference between actual earnings for the year minus the mean analyst forecast for the year.

FIGURE 1 Matching Timeline for data from Compustat/IBES/SDC Earnings Ann. Date t-1 Earnings Ann. Date t Earnings Ann. Date t+1 Deal Announcement Date Target t-1 =1 Deal Announcement Date Target t+1 =1 FYE t-1 ES t-1 FYE t ES t FYE t+1 ES t+1 This figure describes the procedure followed to match the data from the three databases (Compustat/IBES/SDC). For a given year t, accounting information from Compustat as of Fiscal Year End (FYE t ) is matched with the corresponding Earnings Surprise (ES t ), computed as of the earnings announcement date. For any firm that has a deal announced in SDC between the earnings announcement date at time t and the earnings announcement date at t+1, we make the dummy variable Target t+1 =1, to signify that the threat is following the earnings surprise (equivalently, we identify Target t-1 =1 for any deal announced in the period preceding the earnings surprise at time t).

TABLE 1.Sample Observations with non-missing data from Compustat (excluding Financial and Utility firms), over the period from 1990-2012. Number of firms that found a match in IBES and have earnings surprise information (defined as the difference between actual and last consensus analyst forecasted EPS) Merge IBES Sample with takeover information from SDC Platinum (no loss of observations in this step because the observations take the value of 0 if there is no matching threat). This is the main sample we use in our regressions. Number of Observations in the sample that have a credible threat following the earnings surprise at time t (TARGET t+1 =1). Firm-Year Unique Firms Observations 177,969 19,725 74,711 9,717 74,711 9,717 2,427 2,288

TABLE 2. Descriptive analysis of variables related to Takeover Threats This table provides information about the distribution of the variables that are used in our regressions. Variable definitions are presented in the data appendix. The sample covers the period from 1990 to 2012 and the analysis is done at firm-year level. Panel A presents the mean, standard deviation and the median values for all of our variables, for the main sample as well as the takeover subsample (we define the takeover subsample as the subset of firms that received a threat subsequent to a given earnings surprise, i.e. firm years where TARGET t+1 =1). Panel B presents the correlation matrix for our variables of interest Pearson correlations are documented in the top part of the table, and Spearman correlation in the bottom (correlations in bold are significant at 1%, 5% and 10% levels). PANEL A: Summary Statistics 1 Main Sample Take-Over Sample Mean Standard Median Mean Standard Median Deviation Deviation BEAT 0.578 0.493 1.000 0.594 0.491 1.000 GRDUMMY 0.209 0.406 0.000 0.263 0.440 0.000 GROWTH 0.102 0.255 0.102 0.119 0.268 0.093 LEV 0.484 0.255 0.475 0.477 0.266 0.454 LIQUID 0.205 0.231 0.108 0.235 0.246 0.134 MTB 3.278 60.932 2.064 3.336 23.282 3.336 PE 13.709 45.461 13.273 11.376 48.8 10.974 ROE -0.042 0.671 0.085-0.097 0.748 0.065 SIZE 3620.027 17980.919 319.416 1376.868 6215.641 199.700 UE -0.113 0.718 0.001-0.077 0.537 0.005 Number of Firm Year Observations 74,711 2,427

TABLE 2 (continued) PANEL B: Correlation Matrix (Main Sample n=74,711) BEAT GROWTH GRDUMMY LEV LIQUID MTB PE ROE SIZE TARGET UE BEAT 0.109-0.016-0.066 0.050 0.006 0.063 0.111 0.017 0.006 0.359 GROWTH 0.127-0.005-0.071-0.046 0.008 0.091 0.149-0.034 0.012 0.102 GRDUMMY -0.016-0.004 0.081-0.092 0.005 0.006 0.005-0.014 0.024 0.013 LEV -0.059-0.078 0.082-0.431 0.001-0.045 0.013 0.099-0.004-0.100 LIQUID 0.061-0.018-0.107-0.498 0.012-0.059-0.181-0.094 0.024 0.020 MTB 0.139 0.249 0.002-0.093 0.245 0.001-0.049-0.001 0.000 0.004 PE 0.150 0.221 0.007-0.073-0.096 0.227 0.111 0.014-0.009 0.052 ROE 0.204 0.287 0.024 0.111-0.165 0.239 0.406 0.049-0.015 0.144 SIZE 0.119-0.099-0.004 0.388-0.323-0.023 0.187 0.279-0.022 0.021 TARGET 0.006 0.016 0.024-0.007 0.018 0.003-0.016-0.026-0.041 0.009 UE 0.855 0.102-0.007-0.047 0.069 0.104 0.111 0.191 0.132 0.003

TABLE 3: Summary Statistics and tests of differences between Takeover threat firms versus and Non-Takeover threat firms. This table documents the differences in means for our variables of interest between the takeover and nontakeover samples (i.e. TARGETt+1=1 and TARGETt+1=0, respectively). In the last column, we present the magnitude of the differences, and we test whether these differences are statistically different from each other. Differences that are followed by *, **, and *** are significant at the 10%, 5% and 1% levels, respectively. Non-Takeover Observations (n=72,284) Mean Standard Takeover Observations (n=2,427) Mean Standard Deviation Tests of Differences Deviation BEAT 0.578 0.493 0.594 0.491-0.017 GRDUMMY 0.207 0.405 0.263 0.440-0.056*** GROWTH 0.102 0.254 0.119 0.268-0.018*** LEV 0.484 0.255 0.477 0.266 0.0063 LIQUID 0.204 0.231 0.235 0.246-0.031*** MTB 3.277 61.800 3.336 23.282-0.059 PE 13.787 45.342 11.376 48.8 2.411*** ROE -0.041 0.668-0.097 0.749 0.056*** SIZE 3695.0 18240.000 1376.900 6215.600 23118.500*** UE -0.114 0.723-0.004 0.718-0.038***

TABLE 4. Meeting and Beating Expectations and the Probability of a Takeover Threat This table reports the estimated coefficients (Chi square-values in parentheses) from the logit model investigating the impact of meeting or beating earnings -expectations on the probability of receiving a takeover threat in the subsequent period. The dependent variable TARGET t+1 equals 1 if there is a threat in the subsequent period and 0 otherwise. All variables are defined as described in the data appendix and the timing match is described in Fig. 1. The standard errors are clustered by firm and year. Coefficients followed by *, **, and *** are significant at the 10%, 5% and 1% levels, respectively. PANEL A: BEAT takes the value of 1 if the surprise is equal to or larger than 0.00. Model 1:TARGET i,t+1 = α + β 1 GRDUMMY i,t + β 2 GROWTH i,t + β 3 INDUSTRY i,t + β 4 LEV i,t + β 5 LIQUID i,t + β 6 MTB i,t + β 7 PE i,t +β 8 ROE i,t +β 9 SIZE i,t + Model 2:TARGET i,t+1 = α + β 1 UE i,t + β 2 BEAT i,t + β 3 UE i,t *BEAT i,t +β 4 GRDUMMY i,t + β 5 GROWTH i,t + β 6 INDUSTRY i,t + β 7 LEV i,t + β 8 LIQUID i,t + β 9 MTB i,t + β 10 PE i,t + β 11 ROE i,t + β 12 SIZE i,t + Model 1 Model 2 Intercept -14.054 (0.013) -14.026 (103.44)*** Theory Variables UE 0.181 (17.594)*** BEAT 0.1 (7.814)*** UE*BEAT -0.763 (32.667)*** Control Variables GRDUMMY 0.317 0.7 (44.520)*** (39.291)*** GROWTH 0.262 (11.10)*** 0.192 (5.576)*** INDUSTRY 11.006 (0.008) 10.986 (97.829)*** LEV 0.272 (9.867)*** 0.376 (19.020)*** LIQUID 0.372 (14.919)*** 0.377 (15.336)*** MTB -0.001 (0.031) -0.001 (0.346) PE -0.001-0.001

(2.385) (3.272)* ROE -0.0 (1.038) -0.056 (3.839)** SIZE -0.114 (86.876)*** -0.126 (120.03)*** Pseudo-R 2 0.29% 0.36% Number of observations 74,711 74,711 TARGET=1 2,427 2,427 TARGET=0 72,284 72,284 Standard errors clustered by firm, year and No Yes SIC Marginal Effect of UE*BEAT -0.024 (-3.139)***

PANEL B: BEAT takes the value of 1 if the surprise is bigger than 0.00 but smaller or equal to 0.01. TARGET i,t+1 = α + β 1 UE i,t + β 2 BEAT i,t + β 3 UE i,t *BEAT i,t +β 4 GRDUMMY i,t + β 5 GROWTH i,t + β 6 INDUSTRY i,t + β 7 LEV i,t + β 8 LIQUID i,t + β 9 MTB i,t + β 10 PE i,t + β 11 ROE i,t + β 12 SIZE i,t + Coefficient (Chi-Square) Intercept -14.024 (74.840)*** Theory Variables UE 0.119 (17.147)*** BEAT -2.854 (35.336)*** UE*BEAT 0.7 (.721)*** Control Variables GRDUMMY 0.319 (42.445)*** GROWTH 0.273 (11.708)*** INDUSTRY 11.038 (60.213)*** LEV 0.275 (10.463)*** LIQUID 0.337 (12.497)*** MTB -0.006 (0.175) PE -0.006 (1.643) ROE -0.033 (1.380) SIZE -0.118 (11.253)*** Pseudo-R 2 0.41% Number of Observations 74,711 TARGET=1 2,427

TARGET=0 72,284 Standard errors clustered by firm, year and SIC Yes Marginal Effect of UE*BEAT 9.624 (3.015)***

TABLE 5. Meeting and Beating Expectations and the Success of a Takeover Threat This table reports the estimated coefficients (Chi square-values in parentheses) from the logit model investigating the impact of meeting or beating earnings -expectations on the probability of a successful take-over threat. The dependent variable SUCESS equals 1 if the merger is completed successfully (a deal has been reached). All variables are defined as described in the data appendix and the timing match is described in Fig. 1. The standard errors are clustered by firm and year. Coefficients followed by *, **, and *** are significant at the 10%, 5% and 1% levels, respectively. Model 3:SUCESS i = α + β 1 GRDUMMY i,t + β 2 GROWTH i,t + β 3 LEV i,t + β 4 LIQUID i,t + β 5 MTB i,t + β 6 PE i,t +β 7 ROE i,t +β 8 SIZE i,t + Model 4:SUCESS i = α + β 1 UE i,t + β 2 BEAT i,t + β 3 UE i,t *BEAT i,t +β 4 GRDUMMY i,t + β 5 GROWTH i,t + β 6 LEV i,t + β 7 LIQUID i,t + β 8 MTB i,t + β 9 PE i,t + β 10 ROE i,t + β 11 SIZE i,t + Model 3 Model 4 Intercept 2.299 (87.581)*** 2.215 (65.122)*** Theory Variables UE 0.089 (0.878) BEAT 0.3101 (7.027)*** UE*BEAT -0.201 (0.244) Control Variables GRDUMMY 0.418 (10.299)*** 0.421 (10.05)*** GROWTH -0.174 (0.720) -0.239 (1.468) LEV 0.2 (0.935) 0.3102 (1.681) LIQUID 0.628 (5.332)** 0.548 (3.727)** MTB 0.001 0.001 (0.265) PE 0.002 (2.714)* (0.519) 0.002 (2.149)

ROE 0.201 (7.533)*** 0.178 (6.010)*** SIZE -0.212 (39.955)*** -0.227 (37.539)*** Pseudo-R 2 3.17% 3.57% Number of observations 2,427 2,427 SUCESS=1 1,957 1,957 SUCESS=0 470 470 Standard errors clustered by firm, year and No Yes SIC Marginal Effect of UE*BEAT -0.0 (-3.543)***

PANEL B: BEAT takes the value of 1 if the surprise is bigger than 0.00 but smaller or equal to 0.01. SUCESS i = α + β 1 UE i,t + β 2 BEAT i,t + β 3 UE i,t *BEAT i,t +β 4 GRDUMMY i,t + β 5 GROWTH i,t + β 6 LEV i,t + β 7 LIQUID i,t + β 8 MTB i,t + β 9 PE i,t + β 10 ROE i,t + β 11 SIZE i,t + Coefficient (Chi-Square) Intercept 2.331 (72.818)*** Theory Variables UE 0.159 (3.479)** BEAT -2.029 (5.6)*** UE*BEAT 0.182 (3.285)** Control Variables GRDUMMY 0.422 (10.073)*** GROWTH -0.188 (0.340) LEV 0.291 (1.471) LIQUID 0.616 (4.736)** MTB 0.001 (0.545) PE 0.002 (3.023)* ROE 0.198 (7.486)*** SIZE -0.216 (33.648)*** Pseudo-R 2 3.56% Number of Observations 2,427 TARGET=1 1,957 TARGET=0 470 Standard Errors clustered by firm, year and SIC Yes 29

Marginal Effect of UE*BEAT 2.745 (3.629)***