Can hedge funds predict takeover offers and outcomes? The influence of hedge fund ownership on takeover likelihood and offer success

Uhlenkamp, Lisa M.; Schwetzler, Bernhard; Althammer, Wilhelm Article Can hedge funds predict takeover offers and outcomes? The influence of hedge fund ownership on takeover likelihood and offer success Schmalenbach Journal of Business Research (SBUR) Provided in Cooperation with: Schmalenbach-Gesellschaft für Betriebswirtschaft e.V. Suggested Citation: Uhlenkamp, Lisa M.; Schwetzler, Bernhard; Althammer, Wilhelm (2025) : Can hedge funds predict takeover offers and outcomes? The influence of hedge fund ownership on takeover likelihood and offer success, Schmalenbach Journal of Business Research (SBUR), ISSN 2366-6153, Springer, Heidelberg, Vol. 77, Iss. 2, pp. 309-355, https://doi.org/10.1007/s41471-025-00211-y This Version is available at: https://hdl.handle.net/10419/323729 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ ORIGINAL ARTICLE https://doi.org/10.1007/s41471-025-00211-y Schmalenbach Journal of Business Research (2025) 77:309–355 Can Hedge Funds Predict Takeover Offers and Outcomes?—The Influence of Hedge Fund Ownership on Takeover Likelihood and Offer Success Lisa M. Uhlenkamp · Bernhard Schwetzler · Wilhelm Althammer Received: 31 January 2024 / Accepted: 26 February 2025 / Published online: 8 May 2025 © The Author(s) 2025 Abstract This study investigates the impact of a company’s shareholder structure, particularly hedge fund ownership, on its likelihood to become target of a takeover offer. Additionally, it explores the consequences of such an offer within the context of the German takeover market. In line with prior research, we find that the presence of a hedge fund stake increases the likelihood of a firm receiving a takeover offer. However, the underlying reasons for this correlation remain a subject of debate in the academic literature, specifically whether it is due to hedge funds’ superior investment strategies, their role as activists, or their access to private information. We find that ownership structure is the primary predictor of takeover activity, while publicly available data offers limited predictive power. These findings, supported by supplementary tests, lend credence to the hypothesis that hedge funds profit from private information. In relation to the outcome of a takeover offer, we observe that unique characteristics of German corporate law, particularly the high level of minority protection, empower hedge funds to employ strategies beyond the conventional speculation on a takeover offer. A prevalent strategy among hedge funds in German takeovers entails the speculation on higher compensation through subsequent structural measures, such as the signing of a domination agreement or a squeezeout following a successful offer. This creates a particular prisoner’s dilemma for Lisa M. Uhlenkamp HHL Leipzig Graduate School of Management, Jahnallee 59, 04109 Leipzig, Germany E-Mail: [email protected] Bernhard Schwetzler Chair of Financial Management, HHL Leipzig Graduate School of Management, Jahnallee 59, 04109 Leipzig, Germany E-Mail: [email protected] Wilhelm Althammer Chair of Macroeconomics, HHL Leipzig Graduate School of Management, Jahnallee 59, 04109 Leipzig, Germany E-Mail: Wilhelm.Althamm[email protected]e K 310 Schmalenbach Journal of Business Research (2025) 77:309–355 hedge funds as target shareholders: to maximize gains, they should refrain from tendering their shares, while other shareholders must tender theirs for the offer to succeed. Our results on the impact of pre-offer hedge fund ownership reflect this tension: we find some evidence of a positive impact on offer success, though this effect becomes insignificant when focusing on conditional offers with a minimum acceptance threshold. Furthermore, our findings imply a negative relationship between pre-offer hedge fund ownership and the offer premium, thereby supporting the view that bidders anticipate hedge funds’ low sensitivity to premium increases when deciding whether to tender their shares. In conclusion, our results underscore the growing influence of hedge funds in German takeover offers. Keywords M&A · Takeover likelihood · Hedge fund · Ownership structure · Offer premium JEL Classification G14 · G18 · G23 · G32 · G34 1 Introduction In July 2017, two private equity firms, Cinven and Bain, announced that their conditional takeover offer for STADA AG, a listed pharmaceutical company, with an offer price of C66.25 had successfully passed the required minimum acceptance rate. Subsequent events included the acquisition of a substantial stake in the target stock by the US hedge fund Elliott, which then threatened to block the vote required for the proposed domination agreement. Elliott successfully demanded compensation of C74.4 C per share for all shares. In the subsequent judicial review of the offer, the compensation for all minority shareholders was further increased to C81.73. In June 2020, the majority shareholders initiated a squeeze-out of the remaining minority at a compensation offer C98.51. This sequence of events has given rise to a public debate about the consequences of increasing hedge fund activities in Germany, shedding light on the engagement of hedge funds in takeover offers under German takeover and corporate law. The present study aims to contribute to this discourse by analyzing the impact of hedge fund ownership on the probability of a company being targeted by a takeover offer and, after an offer has been made, on the outcome of the offer. The consequences of hedge funds engaging in capital markets are generally controversial. Critics and regulators express concerns that hedge funds may contribute to financial market volatility and amplify systemic risk (Chan et al. 2006). Hedge fund activism might harm shareholder value by being overly focused on short-term outcomes, neglecting longer-term profitable investments. Proponents, however, contend that hedge funds possess a unique capacity to influence corporate governance decisions and enhance shareholder value in comparison to other market participants, such as mutual and pension funds, thereby contributing to the enhancement of financial market quality. Lower regulatory requirements enable concentrated investment in select firms, amplified by higher leverage levels and augmented capital flexibility, thereby building upon locked-in capital. Moreover, hedge fund managers possess K Schmalenbach Journal of Business Research (2025) 77:309–355 311 strong incentives and are less susceptible to conflicting interests arising from their role as external investors (Brav et al. 2008; Agarwal et al. 2015; Bebchuk et al. 2015). An important aspect of the discourse concerns the question of whether hedge funds are better able to predict takeover offers. Greenwood and Schor (2009), Boyson et al. (2017), and Dai et al. (2017) have found that the pre-offer hedge fund stake in a firm is associated with a higher likelihood of that firm becoming a takeover target. However, the underlying mechanisms remain a subject of debate. Specifically, the question of whether hedge funds possess superior trading skills, have access to confidential information, or stimulate takeovers by behaving actively remains unresolved. The illicit nature of insider trading makes it challenging to directly observe the leakage of private information regarding impending takeover offer announcements that is subsequently exploited by traders. Conversely, trading in rumored targets offers an attractive investment opportunity. However, when information is not public but only available to a specific group of market participants, it provides an unfair advantage, thus distorting the capital allocation of the takeover market. Hedge funds have increased their investment in the German takeover market and have become relevant players in the takeover offer process (Dobmeier et al. 2020). Consequently, it is important to closely observe their actions to analyze their impact on the German takeover market. This is particularly relevant given that the German market for corporate control has historically been oriented towards stakeholders’ interests (employees, banks, but also customers and suppliers), and has converged in recent years to a stronger focus on shareholders’ interest (Mietzner et al. 2011). The strong minority protection of German corporate law enables a distinctive investment strategy, used by hedge funds in relation to a takeover offer, known as “backend” or “endgame” speculation. As in the above example of STADA, it may be advantageous for hedge funds, as pre offer shareholders, to refrain from accepting the takeover offer and tender their shares. This allows them to speculate on the success of the offer and the potential for higher compensation in a later domination agreement or squeezeout. This paper aims to contribute to a better understanding of the consequences of such a strategy for the success of a takeover offer. The present study hypothesizes that hedge funds actively trade based on private information regarding upcoming takeover offers. An extensive panel data set of German publicly listed firms from 2006–2020 is constructed, consisting of 5068 firmyears, thereof 260 deals that were announced to 225 distinct firms. In order to draw conclusions on a firm’s takeover likelihood, we apply different matching procedures. This approach enables us to investigate the hedge fund stake in treatment firms (i.e., takeover targets) compared against matched control firms (i.e., similar firms and hence potential targets). We also analyze the impact of hedge fund ownership on the outcome of a takeover offer. The findings of this study lend support to the main hypotheses, suggesting that the ownership structure of a firm serves as an important predictor of takeover deals. Specifically, the results indicate that a greater hedge fund stake in a firm is associated with a significantly higher likelihood of takeover, while a greater index fund stake significantly decreases the probability of facing a takeover offer. Publicly observable K 312 Schmalenbach Journal of Business Research (2025) 77:309–355 financial, balance sheet, and market information, such as the return on equity or the market-to-book ratio, appear to offer no substantial information on which firm is becoming a takeover target and, consequently, is an investment opportunity. This result provides first evidence that hedge funds might possess private information on impending takeover offers that guides their trading decisions, but do not possess superior skills in interpreting public information (Dai et al. 2017). To support these findings, we conduct a series of tests, the results of which indicate that hedge funds may benefit from private information by maintaining close ties with advisors to the target, such as insiders involved in deal financing. Furthermore, we find that targets with a high level of hedge fund engagement are associated with better prediction benchmarks, i.e., a higher level of correct and a lower level of false target predictions. This systematic trading pattern is again indicative of the exploitation of private information. We also examine the impact of hedge fund ownership on offers that have been made and investigate whether hedge fund trades systematically impact the outcome of a takeover offer. This occurs against the background of the potential “backend” speculation described above. This strategy creates a particular prisoners’ dilemma for hedge funds: to maximize gains, they should refrain from tendering their shares into the offer, as long as sufficiently many other shareholders tender theirs for the offer to succeed. The findings on the impact of pre-offer hedge fund ownership reflect this tension. A higher hedge fund stake prior to the offer is associated with a higher takeover success rate; however, this effect becomes insignificant when focusing on conditional offers with a minimum acceptance threshold. Furthermore, bidders offer lower premiums in takeover offers where the target has a higher hedge fund stake as owners. This phenomenon could be interpreted as bidders anticipating the hedge funds’ reduced sensitivity to premium increases due to the “backend” speculation, leading them to offer a lower premium. Given the weaker statistical significance of our findings on offer outcomes, we interpret them with caution, attributing them to the speculation of hedge funds regarding higher compensation in the “backend” following the successful completion of an offer. Our results suggest that hedge funds possess extensive networks and exploit private information regarding impending takeover offers. While German takeover legislation does not oblige investors to disclose their motivations, we cannot entirely exclude the possibility that this outcome may be partly attributable to hedge funds stimulating takeover offers through shareholder activism. Our study contributes to the existing literature on takeover likelihood prediction and hedge funds’ engagement in takeover offers. To the best of our knowledge, we are the first to investigate the takeover likelihood of publicly listed firms in Germany. Studies on the European, US, and UK markets include Brar et al. (2009), Dai et al. (2017), Danbolt et al. (2016), respectively. Furthermore, we contribute to the “channel” discourse on whether hedge funds have superior skills in identifying prospective takeover targets, trade on private information (e.g., Dai et al. 2017; Fich et al. 2019) or exhibit activist behavior, thereby stimulating takeover offers (e.g. Greenwood and Schor 2009,Boysonetal.2017). Finally, we contribute to the literature by exploring how hedge funds’ engagement in takeover offers affects the properties, characteristics and outcome of the deal, in terms of takeover success K Schmalenbach Journal of Business Research (2025) 77:309–355 313 rates and offer premiums (e.g., Dai et al. 2017; Corum and Levit 2019;Wuand Chung 2022). The remainder of the paper is organized as follows. Section 2 provides an overview of the German corporate governance system and legal framework, as well as the regulation on insider trading. Section 3 offers a review of the literature on takeover likelihood prediction and hedge fund engagement in takeover offers. Section 4 presents the data sample, variable definitions and summary statistics. Section 5 proceeds with the results of our empirical analysis and robustness. Section 6 concludes the paper. 2 German Corporate Governance System and Legal Framework The German corporate governance system was characterized by the prominent role of banks in financing and monitoring publicly listed corporations, a highly concentrated ownership structure, a pronounced orientation towards stakeholders’ interests with co-determination, and the representation of various stakeholder groups on the supervisory board (Franks and Mayer 2001; Mietzner and Schweizer 2014). Regulatory changes, including the enactment of the German Securities Acquisition and Takeover Act (Takeover Act or WpÜG), and modifications of the corporate tax law in 2002, prompted a convergence of the German corporate governance system towards an Anglo-Saxon shareholder-oriented model. The enactment of the Takeover Act resulted in an increase in takeover activity within the German market. The modifications in corporate tax legislation prompted banks to reduce their engagement in German firms, thereby creating opportunities for foreign capital investments and a diversification of ownership structures (Schmidt 2004; Goergen et al. 2008;Weber 2009). However, with universal banks engaging less in the corporate control of German publicly listed firms, and the more dispersed ownership structure leading to lower levels of shareholder participation, the German corporate governance market experienced a control vacuum, which attracted shareholder activists, specifically hedge funds, to step in and obtain stakes in firms with low levels of governance and profitability (Bessler et al. 2015). The process of acquiring a German publicly listed firm entails specific regulations on disclosure requirements to uphold market transparency and enable investors to acquire all relevant information on a specific stock. The bidder must publish her decision to pursue a takeover via an ad hoc news portal as soon as possible (§ 10 (1) and (3) WpÜG). Material market rumors regarding a potential takeover offer may also obligate the potential target company to disseminate an ad hoc message taking a position. All parties involved in the takeover offer process who possess private information that may considerably influence stock prices are prohibited from trading in the target shares in question until all relevant information is published to capital market participants (The Mayer Brown Practices 2022;BaFin2022). The German Takeover Act differentiates mandatory and voluntary offers. A mandatory offers is initiated by a shareholder (the bidder) who surpasses the 30% ownership threshold, thereby acquiring control of the target company according to the German Takeover Act. These offers may not be subject to conditions for becoming valid. In contrast, K 314 Schmalenbach Journal of Business Research (2025) 77:309–355 a voluntary takeover offer provides more flexibility, and can be subject to conditions to become valid, such as minimum acceptance rate (§ 29 (2), and § 35 WpÜG). These offers are often used to expand control over the target. With regard to shareholder activism, German legislation demands that shareholders disclose only limited information regarding their intentions, in contrast to the requirements stipulated under US regulations. According to the Risk Limitation Act of 2007, shareholders are obligated to report their holdings and the investment purpose thereof once they surpass the 10% threshold (§ 43 Securities Trading Act WpHG). This implies a convergence of German regulations towards those in the US. However, unlike US regulations under Schedule 13D and 13G, German investors are not required to disclose their intentions to be active or passive (see Mietzner et al. 2011; Weber and Zimmermann 2011; Bessler et al. 2015). Finally, German corporate law plays a significant role in takeovers and related hedge fund strategies. The German stock corporation act AktG offers a high level of legal protection to minority shareholders in a corporate takeover: firstly, achieving a majority of shares in a takeover offer does not allow the majority shareholder to squeeze out the remaining minority at a compensation equal to the offer price. Secondly, owning the majority of shares does not grant the investor full control over the target. In order to exercise full control, a domination agreement must be signed between the corporation and the majority shareholder, which itself requires a 75% vote in a shareholder meeting (§ 293 AktG). In such cases, minority shareholders are entitled to receive an adequate compensation, if they decide to relinquish their shares to the majority shareholder (§ 305 AktG). Alternatively, if they opt to maintain their investment in the company, they are entitled to receive a guaranteed dividend on a recurrent basis (§ 304 AktG). Complete minority squeeze-outs necessitate a 90% or 95% ownership stake of the majority shareholder (§ 62 UmwG, § 327 AktG). In all cases, the compensation of minority shareholders who depart from the corporation is subject to a court verification in an appraisal proceeding, as outlined in the German Arbitration Act (§ 2 Spruchverfahrensgesetz, SpruchG). As a result, attaining 100% ownership of the target is a long and in many cases costly process. The true cost of acquiring 100% of the shares in a German takeover offer is in most cases unknown, as this would require the tracking all structural measures and the corresponding compensation offers, including the potential increase in the court procedure. Aders et al. (2021) conducted a study on a sample of German control taking takeover offers; they found the pure offer premium to be on average 39%. However, the combined cost of offer premium and eventually increased compensation for a domination agreement and a squeeze-out for the shares acquired after the offer during the “endgame” is more than 70%. Consequently, the exploitation of the strong minority protection and speculation regarding higher compensation after a successful offer has become a prevalent investment strategy among hedge funds. A substantial number of hedge funds enter the takeover process even after the offer announcement has been made (Dobmeier et al. 2020). K Schmalenbach Journal of Business Research (2025) 77:309–355 315 3 Literature Review and Hypotheses Development 3.1 Takeover Likelihood Prediction 3.1.1 Hedge Fund Engagement Palepu (1986) provided the seminal work for the field of takeover likelihood prediction. He identified management inefficiency, undervaluation of the firm, size as well as a mismatch of firm growth and available resources as key factors in predicting takeover likelihood. Greenwood and Schor (2009), Boyson et al. (2017), and Dai et al. (2017) have identified a correlation between a pre-offer hedge fund stake in a firm and an increased likelihood of that firm becoming a takeover target. However, the channel of effects remain the subject of a debate. Specifically, the question of whether hedge funds possess superior trading skills, access to private information, or stimulate takeovers by behaving actively remains unresolved. Hedge funds that have invested in potential takeover targets can adopt a passive or activist investor role, and in both cases they benefit from the takeover-induced increase in target stock prices.1As passive investors, hedge funds identify and invest in firms that are likely to become targets, either based on superior trading skills or private information. In contrast, activist hedge funds aim to influence corporate governance decisions and stimulate a takeover offer by attracting a potential bidder. The presence of activist hedge funds has been shown to improve takeover decisions by reducing poorly performing takeovers, resulting in a smaller number of takeovers and smaller deals (Wu and Chung 2022). Superior trading skills, built on public information such as financial statements, industry dynamics, and rumors/news in the media may enable hedge funds to generate trading profits by investing in eventual takeover targets. While several studies find that takeover targets possess similar characteristics to each other as opposed to firms not subject to takeover attempts, e.g., undervaluation captured by the market-to-book value, or management inefficiency captured by a low return-on-equity, a broad strand of the literature argues that target prediction models suffer from limited precision and are unable to generate trading profits based on their predictions (Palepu 1986;Powell2001;Braretal.2009; Danbolt et al. 2016). Daietal.(2017) present evidence of trading on substantial insider information.2 Weber and Zimmermann (2011) analyze the disclosure process of hedge funds ac1Schwetzler and Uhlenkamp (2020) identified takeover announcement returns of 10.0% and abnormal returns including a 42 day pre-bid run-up period of 23.6%. These results underscore the potential for trading profits in a German sample. Betton et al. (2014) and Bessler and Schneck (2015) report similar returns for samples from the US and European, respectively. 2Bodnaruk et al. (2009) investigated the information leakage channel and found that financial advisors associated with investment banks process insider information on upcoming takeover offers. Similar findings were reported by Lowry et al. (2019) for the option market. Hedge funds trading on private information is also suggested by Massoud et al. (2011) in the lending market, by Agarwal et al. (2013) in the context of confidential holdings that are announced belatedly through additional regulatory filings, and by Qian and Zhong (2018) who provide evidence on information leakage on IPO stocks. K 316 Schmalenbach Journal of Business Research (2025) 77:309–355 quiring stakes of 5% or more in German publicly listed companies and come to a similar result. The literature outlines several channels through which activist hedge funds might attract a potential bidder and hence stimulate a takeover offer. Activist hedge funds may assume a marketing role by “putting firms into play” (Greenwood and Schor 2009,Boysonetal.2017), or they may serve to reduce information asymmetries in takeover offers due to their role as informed monitors. They could lower managerial resistance by threatening a proxy contest, thereby increasing the likelihood of a takeover offer (Corum and Levit 2019). Alternatively, they could counteract potential empire building and overpayment by bidders (Wu and Chung 2022). Our primary hypothesis is that hedge fund ownership has a positive impact on the probability of a company receiving a takeover offer (H1). We utilize information regarding the bidder and the informational environment of the offer to differentiate between potential explanations of hypothesis H1. First, we split our data sample into two subsample, based on whether the bidder is a financial investor or not. This approach is motivated by the findings of Bodnaruk et al. (2009)andDai et al. (2017), who provide evidence that hedge funds exhibit a higher degree of investment in deals with a heightened probability of insider information leakage. They approximate the information leakage probability by the number of advisors to the target, and the external funding source, such as investment banks, which involves a larger network of individuals with insider information. In this study, we approximate information leakage by the type of bidder. As private equity firms, compared to strategists, do not have a special law and M&A department, the number of advisors and thus the probability of information leakage is higher compared to strategic bidders. Significant differences in the impact of hedge fund ownership between the two subsamples are interpreted as support for the hypothesis of private information (H2). In a second subsample, we exclude offers that involved rumors and/or ad hoc announcements to further analyze the role of information leakage for hedge fund pre-offer ownership (H3). The following hypotheses are postulated: H1 A higher hedge fund stake prior to the offer is associated with a higher takeover likelihood. H2 A higher hedge fund stake prior to the offer is associated with a higher takeover likelihood if the bidder is a financial investor. H3 A higher hedge fund stake prior to the offer is associated with a higher takeover likelihood when excluding deals that are related to market rumors/news. H4 Hedge fund trading patterns are superior in identifying targets. The success rate of takeover predictions increases with the in the hedge fund’s pre-offer-stake in the target. K Schmalenbach Journal of Business Research (2025) 77:309–355 323 Table 1 Summary Statistics Variable NMean S.D. Q1 Median Q3 Ownership characteristics Hedge fund stake 5068 0.011 0.032 0.000 0.000 0.007 Institutional stake 5068 0.178 0.177 0.030 0.125 0.272 Strategic stake 5068 0.214 0.292 0.000 0.045 0.350 Individual stake 5068 0.175 0.236 0.000 0.030 0.327 Foreign stake 5068 0.194 0.229 0.004 0.111 0.296 Index funds stake 5068 0.013 0.028 0.000 0.000 0.011 Firm-specific characteristics Firm size: Total assets 5068 12.56 2.42 10.90 12.23 14.10 Inefficient mgmt: Return on Equity 5068 0.042 0.231 0.003 0.088 0.159 Growth resource imbalance dummy 5068 0.249 0.433 0 0 0 Undervaluation: Market to book ratio 5068 0.001 0.001 0.000 0.001 0.001 Leverage 5068 0.189 0.166 0.035 0.163 0.297 Liquidity 5068 0.317 0.228 0.137 0.262 0.460 Momentum: Stock price momentum 5068 0.113 0.470 –0.175 0.048 0.323 Momentum: Trading volume 5068 1.193 4.694 0.013 0.089 0.559 Industry-specific characteristics Industry disturbance dummy 5068 0.194 0.395 0 0 0 Industry concentration 5068 1,746 2.164 365 767 2.252 Macro-specific characteristics Market sentiment 5068 0.790 0.399 1 1 1 Bidder characteristics Toehold 260 0.462 0.289 0.272 0.455 0.736 Toehold acquisition 260 0.212 0.409 0 0 0 Largest shareholder 260 0.604 0.490 0 1 1 Financial investor 260 0.492 0.501 0 0 1 Foreign 260 0.465 0.500 0 0 1 Same industry 260 0.435 0.497 0 0 1 Deal characteristics Success rate 260 0.377 0.313 0.086 0.308 0.652 Offer premium 260 0.195 0.285 0.000 0.105 0.296 Rumor 260 0.085 0.279 0 0 0 Ad hoc 260 0.077 0.267 0 0 0 Cash offer 260 0.923 0.267 1 1 1 Hostile 260 0.196 0.398 0 0 0 Competing offer 260 0.019 0.138 0 0 0 Multiple rounds 260 0.154 0.361 0 0 0 Mandatory offer 260 0.215 0.412 0 0 0 Minimum accept. rate 260 0.188 0.300 0 0 1 Crisis 260 0.254 0.436 0 0 1 Variables are defined in Table 12 K 324 Schmalenbach Journal of Business Research (2025) 77:309–355 period.11 In our sample, the average success rate is 37.7%. Furthermore, bidders offer a premium of on average 19.5% on the three-months weighted average stock price (VWAP), constituting the minimum offer price according to German regulation (§ 5 (1) WpÜGAV). 5Results 5.1 Takeover Likelihood Prediction 5.1.1 Methodology In the first part of our analyses, we examine the relation between a company’s shareholder structure and its likelihood to become target for a takeover offer. Our focus is on the hedge fund stake. In a first step, we examine differences between target firms and control firms within the data sample, based on mean difference testing, Pearson and Spearman correlations and variance inflation factors. We apply the following binominal logit regression model, as proposed in the extant literature (Palepu 1986; Ambrose and Megginson 1992;Braretal.2009; Danbolt et al. 2016; Dai et al. 2017; Tunyi et al. 2019): Prob .targetit D1/ D˛Cˆownership i;t1ˇCˆcontrols i;t1C"it (1) The dependent variable targetit takes the value of one if a company (i) receives a takeover offer in a period (t), and equals zero otherwise. The model computes each company’s likelihood of being subject to a takeover offer in period (t) conditional upon its ownership characteristics and other observed characteristics in the preceding period. Control variables comprise firm-, industryand macro-specific characteristics; their impact on takeover likelihood has been documented in the extant literature. These control variables are computed on a year-end basis prior to the offer announcement. In order to control for a potential look-ahead bias, we lag all independent variables by one period. Furthermore, the June approach to accounting data is adopted to maintain appropriate lags in the model (Soares and Stark 2009). This approach accounts for the fact that most companies publish their financial and balance sheet information with a delay of up to half a year for the December year-end. Consequently, if a takeover offer is announced in the first half of the year (e.g., June 2020), only their financial data from the previous financial year-end are publicly available (from December 2018), which serves bidders as a base for their acquisition decision. This approach also addresses reverse causality, a potential endogeneity issue. Our analyses might be prone to omitted variable bias, a further potential source of endogeneity. To address this concern, we include a comprehensive set of control 11 The discussion of the legal German peculiarities have the consequence that “success” cannot be adequately measured by the bidder successfully acquiring enough shares to have the majority of votes. Actually, some offers in our sample are made by a bidder already owning the majority of the shares. K Schmalenbach Journal of Business Research (2025) 77:309–355 325 variables derived from prior research, and include industry and year fixed effects to control for time-invariant unobserved heterogeneity. To assess the model’s capability to predict takeover targets, the sample is split into a training and a hold-out sample. The last two years of the panel data set, comprising of firm-year observations from 2019 and 2020 (16% of all firm-year observations), are allocated to the hold-out sample (e.g., Espahbodi and Espahbodi 2003). The objective is to approximate a real-life setting of investigating future targets as closely as possible. We run our multivariate analyses on the training sample to mitigate bias and possible overfitting, although at the potential cost of estimation accuracy.12 The extant literature indicates that takeover targets differ substantially from nontarget firms, e.g., in terms of firm size and undervaluation. In order to draw conclusions on a firm’s takeover likelihood, we apply different matching procedures and perform our regression analyses based on matched firm-year observations. Matching allows us to investigate the hedge fund stake in treatment firms (takeover target), compared to the hedge fund stake in the matched control firms (similar firms and hence potential targets). First, we employ multivariate distance matching on the ten nearest neighbors on the year of offer announcement and firm size (and later also on the degree of undervaluation) (see for example Brar et al. 2009). Second, propensity score matching is employed on all covariates and exact matching is performed on year and industry on the ten nearest neighbors (and later applying radius caliper). Propensity score matching is implemented by first executing a logit regression model on the takeover probability (refer to Eq. 1), which offers insights on the fundamental firm characteristics associated with an increased threat of a takeover offer within our sample. Second, each firm’s probability (propensity score) of receiving a takeover offer in the next period is computed. Third, target and control firms are matched exactly on year and industry (based on the two-digit standard industry classification). We calculate mean differences in control variables between actual and potential target firms to control for the quality of our matching procedure, and test whether it leads to sufficiently balanced samples. Finally, we select observations that are successfully matched to be included in our regression analyses (see for example Dai et al. 2017; Tunyi 2019). 5.1.2 Univariate Results The study aims to examine whether actual targets and potential targets (control firms) differ structurally by testing mean differences, analyzing Pearson and Spearman correlations and variance inflation factors. Table 2presents mean values of firm characteristics considered to be associated with a firm’s takeover likelihood. The data sample comprises 260 target firm-years and 4808 control firm-years (complete unmatched sample). The table also reports the p-values of t-tests of the mean differences. Firm characteristics are computed year-end prior to the offer announcement. 12 Another technique to construct the hold-out sample is based on natural proportions of takeover activity per year (Palepu, 1986;Braretal.,2009). K 326 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 2 Univariate Analyses—Mean Difference Testing Variable Mean targets Mean non-targets Mean difference Ownership characteristics Hedge fund stake 0.021 0.011 0.010*** Institutional stake 0.175 0.178 –0.002 Strategic stake 0.178 0.216 –0.038** Individual stake 0.087 0.180 –0.093** Foreign stake 0.181 0.194 –0.014 Index funds stake 0.009 0.014 –0.005*** Firm specific characteristics Firm size: Total assets 12.449 12.569 –0.120 Inefficient mgmt: Return on Equity 0.032 0.042 –0.010*** Growth resource imbalance dummy 0.281 0.248 0.033 Undervaluation: Market to book ratio 0.904 1.063 –0.159*** Leverage 0.223 0.187 0.035*** Liquidity 0.285 0.318 –0.033** Momentum: Stock price momentum 0.103 0.114 –0.011 Momentum: Trading volume 1.204 1.192 0.012 Industry-specific characteristics Industry disturbance dummy 0.242 0.191 0.051 Industry concentration 1.919 1.736 182 Macro-specific characteristics Market sentiment 0.783 0.790 –0.007 Variable definitions are in Table 12 Mean differences for targets and non-targets are displayed in the last column ***, **, and * represent statistical significance at the 1, 5, and 10% level, respectively We find that targets exhibit a significantly divergent ownership structure compared to control firms. They exhibit a higher hedge fund stake (mean difference of +1.0%) and a lower index fund stake (mean difference of –0.05%) at year-end before the offer announcement (both significant at the 1% level). Furthermore, target firms exhibit a significantly lower share of strategic and institutional shareholders (significant at the 5% and 1% level, respectively), and a lower level of institutional shareholders relative to control firms (though statistically insignificant, aligning with Ambrose and Megginson 1992). Concerning firm-specific characteristics, target firms are rather inefficiently managed, as illustrated by a lower return on equity. Furthermore, target firms exhibit a lower valuation level, as measured by the market-to-book ratio, and a higher leverage ratio (both significant at the 1%). These outcomes align with the observations reported by Brar et al. (2009) and Danbolt et al. (2016). In summary, based on our p-values, target and control firms exhibit distinct structural differences in terms of ownership characteristics and selected firm-specific variables. However, we do not identify significant differences in industryand macrospecific characteristics between target and non-target firm-years in the data sample as measured by univariate analysis. Pearson correlation coefficients and variance inflation factors were estimated, revealing a modest correlation between the variables as indicated by the Pearson K Schmalenbach Journal of Business Research (2025) 77:309–355 327 correlation matrix (below the 0.7 threshold suggested by Dormann et al. 2013)and low variance inflation factors (below the3 threshold according to Hair et al. 2019). We conclude that multicollinearity concerns within out dataset are unlikely. The results of this analysis are available on request. 5.1.3 Takeover Likelihood Estimation In order to test our primary hypotheses concerning the relationship between a company’s shareholder structure and its takeover likelihood, we estimate a variety of logistic regression models, as specified in Eq. 1.Table3presents the coefficients and p-values of the logistic regressions for five distinct models, based on the underlying matching procedure. Model 1 represents the baseline regression without matching, meaning that all target and control firms in our sample are included in the regression model. Model 2 includes only those firms that are matched via multivariate distance matching on the ten nearest neighbors on the year of offer announcement and firm size. In model 3, matchings are based additionally on the level of undervaluation approximated by the market-to-book ratio. Variables that enter the matching procedure are excluded as repressors in the respective models. Model 4 builds on target and control firms that are propensity score matched on year and industry (two-digit standard industry classification code) on the ten nearest neighbors of 0.25 caliper. In model 5, matchings are based on radius caliper of 0.25. The quality of the matching procedure is controlled for by computing the differences in our variables of interest and control variables that are matched and hence enter the regression analysis. The results indicate that the matched sample exhibits reduced mean differences compared to the unmatched sample. Overall, our matching procedures, including multivariate distance matching and propensity score matching, result in a significant reduction in the mean bias from the unmatched to the matched sample. These results are available upon request. The estimation results reported in Table 3support our hypotheses H1, which posits a higher hedge fund stake is associated with a higher take-over likelihood, and H5, which posits that a higher stake of index funds implies a lower threat of a takeover offer (with the exception of model 3). Furthermore, a higher strategic stake reduces the takeover likelihood. The sign of this coefficient is consistent with the expectation that this investor group, characterized by its long-term interest in the target company, may act as a deterrent to potential bidders. The impact of institutional investors is not significant. This findings align with the inconclusive empirical results regarding their role, which suggests that institutional investors may either function as a delegated monitor in the capital market due to their superior information acquisition and analysis capabilities, or as a passive investor, a role that can be detrimental to corporate governance (see e.g., Jensen 1993; Bebchuk and Hist 2019; Heath et al. 2022).13 Furthermore, individual stakes are shown to exert a negative impact on takeover likelihood, which may be attributed to their capacity 13 Kim et al. (2024) examine the role of engagement costs as a cause of inconclusive outcomes. They demonstrate that for a specific category of offers (stock-for-stock bids), the proportion of institutional ownership increases the likelihood of receiving a bid. K 328 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 3 Takeover Likelihood Estimation Dependent variable: Probability of becoming a target Baseline, unmatched Multivariate distance matching Propensity score matching (1) (2) (3) (4) (5) Hedge fund stake 0.176 (0.197) 1.067* (0.075) 1.867** (0.018) 1.176*** (0.007) 0.469** (0.015) Institutional stake –0.056* (0.076) –0.247 (0.112) –0.166 (0.284) –0.131 (0.153) –0.077 (0.172) Strategic stake –0.084*** (0.000) –0.527*** (0.000) –0.342*** (0.000) –0.228*** (0.000) –0.129*** (0.000) Individual stake –0.173*** (0.000) –0.857*** (0.000) –0.612*** (0.000) –0.376*** (0.000) –0.275*** (0.000) Foreign stake 0.016 (0.466) –0.054 (0.631) 0.037 (0.732) –0.002 (0.978) 0.000 (0.991) Index funds stake –0.743*** (0.004) –2.348** (0.024) –1.903 (0.197) –2.313*** (0.004) –1.148** (0.013) Firm size: Total assets 0.001 (0.735) ––0.009 (0.166) 0.003 (0.454) Inefficient mgmt: Return on Equity 0.001 (0.977) –0.066 (0.502) 0.152 (0.16) –0.035 (0.501) –0.017 (0.594) Growth resource imbalance dummy 0.010 (0.205) –0.034 (0.453) –0.033 (0.543) 0.038 (0.137) 0.014 (0.367) Undervaluation: Market-to-book Ratio –5.901 (0.288) –40.877 (0.135) – –15.091 (0.365) –9.263 (0.359) Leverage 0.002 (0.942) –0.118 (0.509) –0.104 (0.513) –0.031 (0.701) –0.018 (0.717) Liquidity –0.041** (0.028) –0.214* (0.051) –0.271** (0.047) –0.139** (0.017) –0.074** (0.029) Momentum: Stock price momentum 0.003 (0.753) 0.034 (0.564) 0.005 (0.938) 0.002 (0.935) 0.005 (0.765) Momentum: Trading volume –0.002* (0.084) –0.005 (0.308) –0.013 (0.138) –0.004 (0.184) –0.003 (0.145) Industry disturbance dummy –0.003 (0.784) 0.0420 (0.399) 0.000 (0.995) 0.014 (0.634) 0.009 (0.603) Industry concentration 0.000 (0.226) 0.000 (0.131) 0.000 (0.966) 0.000 (0.558) 0.000 (0.565) Market sentiment 0.028 (0.902) –2.145** (0.038) –3.452** (0.016) –0.176 (0.82) 0.240 (0.607) Industry FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Constant Yes Yes Yes Yes Yes Pseudo R-square 0.136 0.221 0.187 0.095 0.158 N 4014 1688 924 1203 2016 The sample size depends on the matching procedure applied All variables are defined in Table 12 Standard errors are clustered by firm and year. P-vales are reported in parentheses ***, **, and * represent statistical significance at the 1, 5, and 10% level, respectively K Schmalenbach Journal of Business Research (2025) 77:309–355 329 to act as a deterrent for potential bidders. The impact of foreign stakes, however, is not significant, as the extant theoretical frameworks do not permit any predictions regarding the sign. Overall, we find mainly the ownership structure as a predictor of takeover deals. Conventional publicly observable financial, balance sheet and market information, such as the return on equity or the market to book ratio, appear to not provide substantial information which firm is becoming a takeover target and hence an interesting investment opportunity. This observation suggests that hedge funds possess private information regarding impending takeover offers that guides their trading decisions (in line with Dai et al. 2017). In all models, the only control variable with a significant and negative impact on the likelihood of a takeover offer is the target’s holdings of liquid assets. One possible explanation for this finding lies in the stringent requirements imposed by the German Stock Corporation Act (AktG) on using the target’s cash holdings to finance the purchase price. For instance, a domination agreement, which permits limited access to these funds, requires a shareholder vote with a 75% approval threshold. Consequently, substantial cash holdings in the target company are often perceived as a potential deterrent to takeover offers, as they could be utilized for defensive measures, such as share buybacks. While the findings provide support for hypothesis H1, the current analysis is not yet capable of distinguishing between two competing explanations. These explanations include the possibility that hedge funds are exploiting private information, or that they are stimulating takeover offers due to shareholder activism. To further refine our analysis, we split our sample and conduct an additional analysis, this time focusing on bids which either include or exclude financial investors. Following the discussion in Sect. 3.1.1., we hypothesize that these offers are subject to a higher or lower information leakage and, as a result, should be expected to exert a stronger or weaker impact by hedge funds on the takeover probability. Table 4presents the results for the probability of becoming a takeover target, based on firm-year observations selected through propensity score matching. Model 1 and 2 are based on a subsample of deals where the bidder is a financial investor. Conversely, in model 3 and 4 the bidder is not a financial investor. The results align with the reasoning outlined above and support our second hypothesis. The findings indicate a more pronounced association of the hedge fund stake prior to the offer and the threat of a takeover offer for model 1 and 2, when the bidder is a financial investor, in terms of both relevance and significance. Conversely, when the bidder is not a financial investor, this relation becomes insignificant. This finding lends further credence to the argument that hedge funds benefit from private information by maintaining close connections with advisors to the target or insiders involved in the financing of the deal. Second, we investigate whether we would obtain qualitatively similar results by excluding deals that are associated with rumors in the news/media or deals with ad hocs disseminated by the target or bidder firm on a possible upcoming takeover prior to the offer announcement (as motivated by Dai et al. 2017). These data points are hand collected from Bafin offer documents and via comprehensive news/media research. The results are presented in model 5 (excluding deals with rumors) and 6 K 330 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 4 Takeover Likelihood Estimation—Subsample Analyses Dependent variable: Probability of becoming a target Subsample: Deals with bidder being a financial investor (excl. in model 3 and 4) Subsample: Excl. deals with rumors/ad hocs (1) (2) (3) (4) (5) (6) Hedge fund stake 1.499*** (0.000) 0.513*** (0.005) 0.334 (0.255) –0.018 (0.906) 0.976*** (0.005) 0.432** (0.013) Institutional stake –0.230** (0.047) –0.091 (0.151) –0.125 (0.255) –0.069 (0.287) –0.128 (0.186) –0.060 (0.272) Strategic stake –0.168** (0.019) –0.068* (0.051) –0.313*** (0.000) –0.180*** (0.000) –0.243*** (0.000) –0.130*** (0.000) Individual stake –0.314*** (0.002) –0.176*** (0.000) –0.443*** (0.000) –0.286*** (0.000) –0.383*** (0.000) –0.240*** (0.000) Foreign stake 0.118 (0.119) 0.048 (0.21) –0.069 (0.449) –0.037 (0.495) –0.016 (0.824) –0.011 (0.791) Index funds stake –1.969* (0.093) –0.839 (0.154) –2.664*** (0.005) –1.398*** (0.006) –4.079*** (0.004) –2.152*** (0.002) Firm size: Total assets 0.002 (0.812) 0.000 (0.937) 0.018** (0.025) 0.007 (0.105) 0.005 (0.453) 0.000 (0.999) Inefficient mgmt: Return on Equity –0.074 (0.257) –0.042 (0.212) 0.017 (0.8) 0.023 (0.565) 0.009 (0.859) 0.007 (0.82) Growth resource imbalance dummy 0.059* (0.065) 0.025 (0.154) 0.007 (0.839) 0.003 (0.849) 0.031 (0.215) 0.014 (0.339) Undervaluation: Market-to-book Ratio –22.626 (0.293) –12.063 (0.299) –1.390 (0.946) –3.837 (0.741) –6.227 (0.718) –5.132 (0.608) Leverage –0.128 (0.262) –0.077 (0.182) 0.044 (0.662) 0.032 (0.561) –0.011 (0.899) –0.009 (0.864) Liquidity –0.179** (0.015) –0.078** (0.036) –0.079 (0.333) –0.052 (0.236) –0.149** (0.018) –0.076** (0.023) Momentum: Stock price momentum 0.010 (0.776) 0.001 (0.951) 0.014 (0.632) 0.009 (0.597) –0.009 (0.74) –0.005 (0.729) Momentum: Trading volume –0.006 (0.243) –0.003 (0.204) –0.002 (0.559) –0.001 (0.521) –0.004 (0.147) –0.002 (0.152) Industry disturbance dummy 0.012 (0.763) 0.010 (0.644) 0.048 (0.192) 0.023 (0.233) 0.032 (0.313) 0.015 (0.406) Industry concentration 0.000 (0.135) 0.000 (0.147) 0.000 (0.231) 0.000 (0.31) 0.000 (0.438) 0.000 (0.429) Market sentiment 0.643 (0.607) 0.384 (0.558) –0.770 (0.395) –0.218 (0.665) –0.171 (0.855) 0.327 (0.551) Industry FE Yes Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Yes Constant Yes Yes Yes Yes Yes Yes Pseudo R-square 0.120 0.184 0.120 0.201 0.118 0.193 N 661 1282 711 1290 1072 1887 The sample size depends on the matching procedure applied. All variables are defined in Table 12 P-values are reported in parentheses. ***, **, and * represent statistical significance at the 1, 5, and 10% level, respectively K Schmalenbach Journal of Business Research (2025) 77:309–355 331 (excluding deals with ad hocs) in Table 4; they support our third hypothesis. The findings indicate that a higher hedge fund stake significantly increases the probability of a subsequent takeover offer, even when excluding deals where the overall market anticipates a takeover offer. The assertion that rumors and ad hocs should also capture hedge fund activism to some extent reinforces the notion that private information is highly relevant to hedge funds trading in potential target companies. 5.1.4 Model Prediction Ability In this section, we seek to ascertain whether our likelihood models, which are based on different matching procedures, yield more precise and accurate future takeover target predictions than the baseline model and a random selection of potential targets from the population. Our primary interest lies in determining whether hedge funds’ trading patterns are profitable in the sense that hedge funds achieve a higher level of correct takeover target predictions and a lower level of false predictions in firms in which they are heavily invested. First, we estimate the cut-off probability in the training sample to classify firms as likely takeover targets. Subsequently, we apply this cut-off probability to the hold-out sample to assess the predictive capability of our models. Finally, we investigate whether targets with a high stake of hedge fund investment (e.g., based on the upper tercile of hedge fund investment) are associated with higher prediction benchmarks. To derive the optimal probability threshold, a trade-off between type I error (not identifying a target firm) and type II error (classifying a control firm as a target firm) must be balanced. First, we align with Palepu (1986) in suggesting the minimization of total misclassifications. We derive the cutoff by identifying the intersection of the probability density curves of target firms and non-target firms (a method also employed by Espahbodi and Espahbodi 2003). Subsequently, we implement the methodology suggested by R. Powell (2004), which involves maximizing the proportion of target firms in the portfolio to formulate a profitable trading strategy. It is noteworthy that type I errors are associated with higher opportunity cost compared to type II errors. The probability threshold is derived by dividing the predicted targets into ten equal-sized portfolios based on their takeover probability and identifying the portfolio with the highest concentration ratio (share of actual targets to total firms). Danbolt et al. (2016) recommend selecting the upper two deciles, i.e., the 20% of firms with the highest takeover likelihood, to enter the investment portfolio. The out of sample prediction results are presented in Table 5, while the acquisition probability distribution is illustrated in Table 15 in the appendix. The hold-out sample consists of all firm-year observations from 2019 and 2020, which approximate a reallife setting of investing in future targets. This results in a total of 795 observations in the hold-out sample, including 34 targets and 761 non-target firms. The cutoff probabilities range from a low of 6.1% in model 1 to a high of 53.0% in model 3. TP (TN) stands for true positive (true negative) classifications, i.e., the correct prediction of targets (TP) and non-targets (TN). Similarly, FP (FN) denotes false positive (false negative) classifications, representing type-I error (FN) and typeII-error (FP), respectively. Our study examines three distinct prediction benchmarks: precision is the number of correctly predicted targets relative to the total number of K 332 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 5 Out-of-sample Prediction Test Results Predictive power of takeover likelihood models Baseline, unmatched (in %) Multivariate distance matching (in %) Propensity score matching (in %) (1) (2) (3) (4) (5) Cut-off probability 6.1 34.7 53.0 19.9 11.8 Precision 5.7 4.8 5.1 4.4 4.1 Recall 19.4 71.9 59.4 52.9 20.6 Accuracy 83.2 38.2 47.7 49.3 76.0 This table presents the results from the prediction tests of the hold out sample The cut-off probability is determined as the intersection of the target and non-target probability density curves The reported benchmarks are precision TP / (TP+ FP), recall TP / (TP+ FN) and accuracy (TP+ TN) / (TP+TF+FP+FN) Table 6 Out of Sample Prediction Test Results—Investing into largest two Deciles of Takeover Probability Predictive power of takeover likelihood models Baseline, unmatched (in %) Multivariate distance matching (in %) Propensity score matching (in %) (1) (2) (3) (4) (5) Cut-off probability 7.7 67.6 71.6 25.5 15.4 Precision 6.8 4.8 5.4 4.1 2.5 Recall 16.1 59.4 59.4 38.2 8.8 Accuracy 87.2 47.7 50.3 59.5 81.4 predicted targets, reflecting type-II error. Recall is the number of correctly predicted targets relative to the number of actual targets, reflecting type-1 error. Accuracy is the total number of correct predictions relative to the total number of firms. Higher recall values denote lower opportunity costs. For instance, model 5 correctly predicts seven target firms (true positive), but fails to identify 27 true targets to invest in (false negative), resulting in a recall of only 20.6%. The simple baseline model 1 demonstrates superior precision and accuracy compared to models 2–5, while the more advanced models 2–5 exhibit higher recall performance. The recall analysis indicates that model 2 minimizes the opportunity cost of not identifying a target firm among the models presented, while model 1 exhibits the highest opportunity cost (71.9% vs. 19.4%, respectively). Notably, model 5 does not outperform a random selection of firms for investment from the hold out sample, with a precision of 4.1% compared to a share of 4.3% of targets in the holdout sample. In accordance with Danbolt et al. (2016), Table 6presents the prediction results for investment in the 20% of firms with the highest predicted takeover likelihoods. The distribution of actual targets and non-targets for models 5 and model 1 is illustrated in Tables 13,14 and 15 in the appendix. The highest concentration ratios are achieved in the final decile, as previously reported by Brar et al. (2009), though not K Schmalenbach Journal of Business Research (2025) 77:309–355 339 takeover offer premiums, suggesting the potential for trading on private information. Conversely, Wu and Chung (2022) posit that activist hedge funds’ a beneficial role in takeovers. Their analysis indicates that activist hedge funds can enhance takeover decisions, leading to a reduction in the frequency of takeovers and smaller deals. In light of these findings, we have conducted an OLS (ordinary least squares) regression analysis in accordance with following econometric specification to test our hypothesis: Offer Premiumij D˛CˆOwnership in t ij ˇCˆControls ij Cı1.crisis/Cı2.rumor/ Cı3.adhoc/ CjC"ij (3) Individual target companies/takeover offers are denoted by i, while jcaptures target industries and εij is a residual error term. The dependent variable, offer premium, is defined as the offer price per share on the three month volume-weighted average target stock price. The regression model is constructed analogously to Eq. 2in the previous section. The results are presented in Table 10. For model 1, ownership characteristics are computed one month prior to the offer announcement (three and six months for models 2 and 3, respectively). The findings do not support hypothesis H9, which posits a positive relationship between hedge fund stake and premium offered. Instead, a negative relationship is observed, though it is only weakly significant in models 1 and 2 and insignificant in model 3. This finding is interpreted in light of the potential “backend” strategy of hedge funds: being aware of this strategy, bidder might offer reduced premiums at elevated hedge fund stakes, anticipating (partial) support for the offer from hedge funds speculating on an increase of the compensation after the completion of the takeover. However, given the only weak significance of this finding, this interpretation should be treated with some care. It is noteworthy that only the hedge fund stake one month and three months prior to the offer announcement are significantly associated with the offer premium, aligning with the findings of Bodnaruk et al. (2009), who identified only the quarter prior to the offer as being relevant. Furthermore, while a significant relationship for the offer premium is found based on the three months volume-weighted average price (VWAP), the results turn insignificant for the market premium or the markup, which reflect the premium on the current market conditions (results are unreleased and can be provided upon request).16 The controls demonstrate the anticipated outcomes with respect to their impact on the offer premium. Conditional offers exhibit significantly higher premiums, while premiums for mandatory offers are significantly lower than those for voluntary offers. 16 The market premium is reflecting the current stock price environment. It captures the premium of the offer price per share on the stock price two days before offer announcement. K 340 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 10 OLS Regression on the Takeover Offer Premium Dependent variable: Offer Premium t= one month prior to the offer t= three months prior to the offer t= six months prior to the offer (1) (2) (3) Hedge fund stake –0.440* (0.083) –0.530* (0.093) –0.357 (0.314) Institutional stake 0.041 (0.481) 0.060 (0.487) 0.131 (0.294) Strategic stake –0.092 (0.107) –0.161** (0.028) –0.088 (0.222) Individual stake –0.035 (0.526) –0.079 (0.235) –0.012 (0.843) Foreign stake 0.045 (0.466) 0.082 (0.36) 0.102 (0.16) Index funds stake –0.168 (0.83) –0.340 (0.63) –0.614 (0.415) Firm size 0.001 (0.946) 0.008 (0.274) 0.005 (0.604) Toehold 0.001 (0.99) 0.107 (0.383) 0.129 (0.289) Toehold acquisition 0.017 (0.568) 0.036 (0.245) 0.018 (0.552) Bidder largest shareholder –0.044 (0.159) –0.058 (0.19) –0.068 (0.175) Bidder financial investor –0.120** (0.024) –0.085 (0.197) –0.083 (0.212) Bidder foreign 0.033 (0.184) 0.048 (0.121) 0.034 (0.296) Same industry –0.070 (0.182) –0.069 (0.253) –0.061 (0.329) Rumor –0.069 (0.113) –0.086* (0.055) –0.089** (0.046) Ad hoc 0.012 (0.727) –0.024 (0.495) –0.011 (0.775) Cash offer 0.047 (0.348) 0.029 (0.597) 0.001 (0.991) Hostile –0.020 (0.525) –0.039 (0.263) –0.029 (0.494) Competing offer –0.020 (0.774) –0.011 (0.873) –0.013 (0.858) Multiple rounds –0.020 (0.624) –0.032 (0.445) –0.034 (0.451) Mandatory offer –0.090*** (0.001) –0.116*** (0.005) –0.103*** (0.006) Minimum acceptance rate 0.107*** (0.004) 0.107** (0.010) 0.106** (0.014) Crisis 0.073** (0.024) 0.123*** (0.005) 0.093** (0.020) K Schmalenbach Journal of Business Research (2025) 77:309–355 341 Table 10 (Continued) Dependent variable: Offer Premium t= one month prior to the offer t= three months prior to the offer t= six months prior to the offer (1) (2) (3) Industry FE Yes Yes Yes Constant Yes Yes Yes R-square 0.304 0.293 0.285 Adj. R-square 0.181 0.168 0.160 N221 221 222 Standard errors are clustered by year of offer announcement P-vales are reported in parentheses ***, **, and * represent statistical significance at the 1, 5, and 10% level, respectively 5.2.3 Robustness The findings of our analyses might be susceptible to endogeneity. To address a potential omitted variable bias, we include a broad range of control variables suggested by the extant literature. For instance, we incorporate rumor/ad hoc dummies, which are designed to capture market expectations of an impending takeover offer. Additionally, our regression models in the previous two sections incorporate industry fixed effects accounting for unobserved industry trends that could potentially influence our results. Standard errors are clustered on the year of offer announcement to account for potential residual correlation across time. The inclusion of year fixed effects and the clustering of standard errors by time and year yield corroborating results. (Results are unreported and available upon request). We address potential measurement error of the independent variables by analyzing and presenting various models in terms of the month or quarter that ownership characteristics are computed on. This extension lends further support to our interpretation that hedge funds might use private information of upcoming takeover offers. 6Conclusion In our study, we investigate the relation between a company’s shareholder structure and its likelihood of becoming a target for a takeover offer in the German takeover market. We focus on the ownership stake of hedge funds. Furthermore, the study aims to shed light on the channel of effects, that is whether hedge funds possess superior trading skills, exploit private information, or stimulate takeovers. Additionally, we analyze the impact of pre-offer ownership of hedge funds on the outcome of given takeover offers. In accordance with our hypotheses, we find that a higher hedge fund stake is associated with a significantly higher takeover likelihood. Conversely, a higher index fund stake significantly decreases the probability of facing a takeover offer. Our K 342 Schmalenbach Journal of Business Research (2025) 77:309–355 findings reveal that publicly observable financial, balance sheet and market information offer limited insight into whether a company is becoming a takeover target, and consequently an investment opportunity for hedge funds. This observation provides initial evidence that hedge funds may have access to private information regarding impending takeover offers. To further substantiate this conclusion, we conduct several tests. We differentiate between bidder types and find that hedge funds benefit from private information when they have access to advisors to the target, financial investors, and/or insiders involved in the financing of the deal. We also find that firms with a high level of hedge fund engagement have higher prediction success rates, i.e., a higher level of correct target predictions and a lower level of false predictions. This systematic trading pattern is suggestive for the exploitation of private information. We also investigate whether hedge funds systematically impact the takeover success rate, and observe the high level of minority protection by German corporate law to offer an additional strategy to hedge funds by speculating on a higher compensation connected to the signing of a domination agreement or a squeeze out after a successful takeover offer. This “backend” or “endgame” speculation in German takeover offers meanwhile is popular among hedge funds (Dobmeier et al. 2020). To reap these benefits, hedge funds pursuing such a strategy as pre-offer target shareholders may weigh the economic effects of the takeover offer itself against the potential benefits of a successful offer allowing this “endgame” speculation. This additional strategy complicates the observation of straightforward effects of hedge fund ownership on the direct outcome of takeover offers. We find a higher hedge fund stake prior to the offer to be associated with a higher takeover success rate in cases without a minimum acceptance condition. However, when considering conditional offers, no statistically significant impact is observed. The results imply that hedge funds carefully evaluate the potential benefits and drawbacks of accepting or rejecting an offer, and speculate on a higher compensation in case of an offer success. Additionally, bidders offer lower premiums in offers with a higher preoffer hedge fund stake, suggesting that they take the potential speculation on higher compensation in the later endgame into account. Conducting a range of robustness tests, we find our results to be stable: we alter the model specifications, in terms of the variables applied to capture firm characteristics prior to the bid; additionally, we estimate our regression models based on multivariate distance matching and propensity score matching in various forms. The final interpretation of our results points towards hedge funds exploiting private information on impending takeover offers. While we find statistically significant support for this hypothesis, we cannot completely rule out that the reverse channel that hedge funds stimulate takeover offers by activism might be effective. We recognize these limitations to our analyses which are mainly based on lacking data. A notable distinction in regulatory frameworks is evident in the German market, where investors are not obligated to disclose their intentions to be active or passive, as is mandated in the US with the schedule 13D and 13G, respectively. K Schmalenbach Journal of Business Research (2025) 77:309–355 343 7 Appendix Table 11 Sample Distribution by Year and Balance of Targets to Non-targets Panel A: Distribution by Year Year Target firmyears Non-target firmsyears Total Balance: Targets to total, in % 2006 12 190 202 5.9 2007 35 242 277 12.6 2008 27 276 303 8.9 2009 17 283 300 5.7 2010 15 289 304 4.9 2011 21 315 336 6.3 2012 18 315 333 5.4 2013 15 310 325 4.6 2014 19 330 349 5.4 2015 13 351 364 3.6 2016 10 388 398 2.5 2017 15 378 393 3.8 2018 9 380 389 2.3 2019 20 376 396 5.1 2020 14 385 399 3.5 Total 260 4808 5068 5.1 Panel B: Distribution by Industry Industry Target firmsyears Non-target firmsyears Total Basic Materials 1 68 69 Chemicals 5 377 382 Consumer Goods 35 569 604 Consumer Services 32 472 504 Financial Services 51 253 304 Health Care 27 465 492 Industrials 47 1414 1461 Oil & Gas 7 74 81 Technology 47 834 881 Telecommunications 1 120 121 Utilities 7 162 169 Others 0 0 0 Total 260 4808 5068 Panel A shows the distribution of firms that became targets by year of offer announcement and German publicly listed firms that did not receive a takeover offer. Since we analyze takeover likelihood based on firm characteristics of the previous year end, the number of controls is based on data at hand of the prior year. The balance calculates the share of targets to non-target firms in the sample. All variable definitions are in Table 12 Panel B shows the distribution by industry, based on the industry classification by Thomson Reuters and by Cleary Gottlieb and Hamilton (2017). All variable definitions are in Table 12 K 344 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 12 Variable Definitions. (Variable definitions are partly based on Dobmeier et al. 2019) Variable Definition Source Target Dummy set to one if company is becoming a takeover target in respective year Bafin takeover offer documents and Thomson ONE’s screening function Ownership characteristics Hedge fund stake Shareholders with investor sub-type: Investment Advisor/Hedge Fund and hedge fund specific investment style: Hedge Fund, Momentum, Aggressive Growth, Global Macro, Specialty, Long/Short. Or with investor sub-type Hedge Fund Thomson Reuters Eikon, based on Dobmeier et al. (2020) Institutional stake Number of shares institutional shareholders (excluding index funds and ETFs) own in target company divided by number of shares not under bidder’s control at the offer Thomson Reuters Eikon Index fund stake Number of shares index funds and ETFs own in target company divided by number of shares not under bidder’s control at the offer Thomson Reuters Eikon Individual stake Number of shares individual shareholders (i.e., individual persons and families) own in target company divided by number of shares not under bidder’s control at the offer Thomson Reuters Eikon Strategic stake Number of shares strategic shareholders own in target company divided by number of shares not under bidder’s control at the offer Thomson Reuters Eikon Foreign stake Number of shares foreign shareholders (i.e., cross-border shareholder) own in target company divided by number of shares not under bidder’s control at the offer Thomson Reuters Eikon Firm specific characteristics Firm size: Total assets Natural logarithm of total assets (book value based on income statement) – Inefficient mgmt: Return on Equity Net income relative to last and current year’s average common equity Thomson Reuters Eikon Growth resource imbalance dummy Dummy set to one in case of above sample average sales growth and leverage, but below average liquidity or below sample average sales & leverage but above average liquidity. Based on average of 2-year sales growth, leverage and liquidity Thomson Reuters Eikon Undervaluation: Market to book ratio Market capitalization relative to book value of total assets Thomson Reuters Eikon Undervaluation: Dividend yield Dividend per share relative to share price Thomson Reuters Eikon K Schmalenbach Journal of Business Research (2025) 77:309–355 345 Table 12 (Continued) Variable Definition Source Target Dummy set to one if company is becoming a takeover target in respective year Bafin takeover offer documents and Thomson ONE’s screening function Leverage Total debt (short term and current portion of long term debt) relative to total assets Thomson Reuters Eikon Liquidity Ratio of cash & equivalents (cash and marketable securities based on book value) relative to the firm’s total assets Thomson Reuters Eikon Momentum: Stock price momentum One year stock return: Monthly average closing price relative to monthly average closing price in prior year Thomson Reuters Eikon Momentum: Trading volume Number of shares traded monthly relative to total free-float market capitalization Thomson Reuters Eikon Industry-specific characteristics Industry identifier 4-digit SIC code (standard industry classification) Thomson Reuters Eikon Industry disturbance dummy Dummy variable set to one if a takeover took place in the prior year in the same industry (4-digit SIC code) Thomson Reuters Eikon Industry concentration Market shares proxied by summing squared total sales of all firms in the respective industry (4-digit SIC code) Thomson Reuters Eikon Macro-specific characteristics Market sentiment Dummy variable set to one if the Prime All Share index has a positive total return in the previous year Thomson Reuters Eikon Bidder characteristics Toehold Percentage of shares in the target company owned by the bidder prior to the offer (incl. pre-negotiated shares/secured through irrevocable undertakings) Offer document on BaFin website, Thomson Reuters Eikon Toehold acquisition Dummy set to one if the bidder acquired a stake in the target during six months prior to the offer Offer document on BaFin website Largest shareholder Dummy set to one if the bidder is the largest shareholder prior to the offer announcement Thomson Reuters Eikon Financial investor Dummy set to one if the bidder is a financial investor, including private equity Thomson Reuters Eikon Foreign Dummy set to one if the bidders operates from a non-German speaking country Thomson Reuters Eikon Same industry Dummy set to one if the bidder operates in the same industry as the target company Thomson Reuters Eikon, Cleary Gottlieb and Hamilton (2017) K 346 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 12 (Continued) Variable Definition Source Target Dummy set to one if company is becoming a takeover target in respective year Bafin takeover offer documents and Thomson ONE’s screening function Deal characteristics Success rate Number of shares bidder acquired during the acceptance period relative to shares the bidder did not have under control at the offer Constructed based on offer document on BaFin website Offer premium Premium of the offer price per share on the three months volume weighted average target stock price, (OP/VWAP-1) Offer document on BaFin website Rumor Dummy set to one if the takeover offer is preceded by a rumor in the media/news speculations one year before offer announcement Offer document on Bafin website, news/ad hoc online research Ad hoc Dummy set to one if target or bidder disseminated an ad hoc notification on a possible upcoming offer one year before offer announcement Offer document on BaFin website, news/ad hoc online research Cash offer Dummy set to one if the compensation offered by the bidder is cash only Offer document on BaFin website Hostile Dummy set to one if the target management recommends their shareholders to reject the offer Statement of board of directors, Bundesanzeiger Competing offer Dummy set to one if the takeover offer is accompanied by a competing bid Offer document on BaFin website Multiple rounds Dummy set to one if the acceptance period is extended due to changes to the offer conditions by the bidder Offer document on BaFin website Mandatory offer Dummy set to one if the takeover offer constitutes a mandatory offer by regulation Offer document on BaFin website Minimum acceptance rate Dummy set to one if the takeover offer is subject to a minimum acceptance rate of shares for becoming valid Offer document on BaFin website Crisis Dummy set to one if the takeover offer was published in a crisis year, i.e. financial crisis in 2008 and 2009 and Euro crisis 2012 Offer document on BaFin website This table provides definitions of the variables used in the analyses K Schmalenbach Journal of Business Research (2025) 77:309–355 347 Table 13 Distribution of Estimated Acquisition Probability in Training Sample—Model 5 based on Propensity Score Matching Estimated acquisition probability Total of firms Actual targets Matched nontargets Intervals, in % Mid value Number Number f1(p) (in %) Number f2(p) (in %) f1(p)/ f2(p) 0.000 0.069 0.035 32 3 1.4 29 1.6 0.89 0.070 0.139 0.104 1550 83 39.7 1467 81.2 0.49 0.140 0.209 0.174 174 36 17.2 138 7.6 2.26 0.210 0.279 0.244 96 22 10.5 74 4.1 2.57 0.280 0.349 0.314 55 14 6.7 41 2.3 2.95 0.350 0.419 0.384 53 22 10.5 31 1.7 6.14 0.420 0.489 0.454 24 9 4.3 15 0.8 5.19 0.490 0.559 0.525 18 10 4.8 8 0.4 10.81 0.560 0.629 0.595 7 4 1.9 3 0.2 11.53 0.630 0.699 0.665 5 4 1.9 1 0.1 34.58 >0.7 – – 2 2 1.0 0 0.0 – Total – – 2016 209 100 1807 100 – The ten intervals (upper and lower bound displayed) range from the minimum to the maximum likelihood of 74.12% The table shows the number of target and control firms within each interval as well as the percentage (see f1(p)andf2(p)) The cut-off probability is the value where the ratio of f1(p)/f2(p) equals 1 (11.8% solved via interpolation) Table 14 Distribution of Estimated Acquisition Probability in Training Sample—Model 5 based on Propensity Score Matching—Equal-sized Portfolios Estimated acquisition probability Actual targets Non-targets Cutoff prob., in % Benchmarks Range, in % (min–max) Mid value Nf1(p) (in %) Nf2(p) (in %) Precision (in %) Recall (in %) Accuracy (in %) 0.00 0.02 0.01 2 1 200 11 0.00 10.4 100.0 10.4 0.02 0.03 0.02 5 2 197 11 0.02 11.4 99.0 20.2 0.03 0.04 0.03 8 4 193 11 0.03 12.5 96.7 29.8 0.04 0.05 0.04 9 4 193 11 0.04 13.8 92.8 38.9 0.05 0.06 0.06 13 6 188 10 0.05 15.2 88.0 48.0 0.06 0.08 0.07 20 10 182 10 0.06 17.1 82.3 56.7 0.08 0.11 0.09 11 5 191 11 0.08 18.9 72.7 64.7 0.11 0.15 0.13 22 11 179 10 0.11 23.4 67.5 73.7 0.15 0.25 0.20 43 21 159 9 0.15 29.6 56.9 81.5 0.25 0.74 0.49 76 36 125 7 0.25 37.8 36.4 87.2 Total – – 209 100 1807 100 – – – – The target and non-target firms are divided into ten equal-sized portfolios and presented in descending takeover likelihood The table shows the number of target and control firms within each interval as well as the percentage (see f1(p)andf2(p)) The cut-off probability is the lower threshold of each decile K 348 Schmalenbach Journal of Business Research (2025) 77:309–355 Table 15 Distribution of Estimated Acquisition Probability in Training Sample—Model 1—Equal-sized Portfolios Estimated acquisition probability Actual targets Non-targets Cutoff prob Benchmarks Range (min–max) Mid value Nf1(p) (in %) Nf2(p) (in %) Precision (in %) Recall (in %) Accuracy (in %) 0.00 0.01 0.01 5 2 397 10 0.00 5.6 100.0 5.6 0.01 0.02 0.01 10 4 391 10 0.01 6.1 97.8 15.4 0.02 0.02 0.02 7 3 395 10 0.02 6.5 93.3 24.9 0.02 0.03 0.02 12 5 389 10 0.02 7.2 90.2 34.5 0.03 0.03 0.03 8 4 393 10 0.03 7.9 84.8 43.9 0.03 0.04 0.04 14 6 388 10 0.03 9.1 81.3 53.5 0.04 0.06 0.05 14 6 387 10 0.04 10.5 75.0 62.8 0.06 0.08 0.07 26 12 376 10 0.06 12.8 68.8 72.1 0.08 0.12 0.10 34 15 367 10 0.08 16.0 57.1 80.8 0.12 0.90 0.51 94 42 307 8 0.12 23.5 42.0 89.1 Total – – 224 100 3790 100 – – – – Table 16 Hedge Fund Investment Success—Actual and Potential Target Firms split into Quintiles based on Hedge Fund Investment Hedge fund investment Prediction benchmarks Classifications Tercile NMean (in %) Precision (in %) Recall (in %) Accuracy (in %) TP TN FP FN Panel A: Model 5 (cut-off probability based on intersection of density curves) 1 999 0.0 27.7 73.4 76.2 80 681 209 29 3 211 0.1 19.6 52.6 86.1 10 151 41 9 4 403 0.5 20.0 48.6 76.9 17 300 68 18 5 403 5.1 24.8 63.0 93.2 29 269 88 17 Sum/ Mean 2016 1.1 – – – 136 1401 406 73 Panel B: Model 1 (cut-off probability based on intersection of density curves) 1 1904 0.0 14.9 74.6 73.0 85 1305 485 29 3 505 0.1 9.9 47.4 81.8 9 404 82 10 4 803 0.5 10.5 48.6 79.7 17 623 145 18 5 802 4.9 14.3 67.9 69.5 38 519 227 18 Sum/ Mean 4014 1.1 – – – 149 2851 939 75 Panel C: Model 5 (investing into largest two deciles of takeover probability) 1 999 0.0 32.1 65.1 81.2 71 740 150 38 3 211 0.1 29.4 52.6 84.4 10 168 24 9 4 403 0.5 23.3 40.0 83.4 14 322 46 21 5 403 5.1 27.6 52.2 78.9 24 294 63 22 Sum/ Mean 2016 1.1 – – – 119 1524 283 90 K Schmalenbach Journal of Business Research (2025) 77:309–355 355 Wang, J., and B. Branch. 2009. Takeover success prediction and performance of risk arbitrage. Journal of Business & Economic Studies 15(2):10–25. Weber, A. 2009. An empirical analysis of the 2000 corporate tax reform in Germany: effects on ownership and control in listed companies. International Review of Law and Economics 2(9):57–66. Weber, P., and H. Zimmermann. 2011. Hedge fund activism and information disclosure: the case of Germany. European Financial Management 19(5):1017–1050. Wooldridge, J.M. 2003. Cluster-sample methods in applied econometrics. American Economic Review 93(2):133–138. Wu, S.Y., and K.H. Chung. 2022. Hedge fund activism and corporate M&A decisions. Management Science 68(2):1378–1403. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. K