Sticky Decisions: Anchoring and Equity Stakes in International Acquisitions

Equity Ownership in International Acquisitions

Considerable research describes why executives acquire firms, enter a joint venture, enter an alliance, or set up wholly owned subsidiaries in foreign markets (e.g., Barkema & Vermeulen, 1998; Brouthers & Hennart, 2007; Cho & Padmanabhan, 2005; Vermeulen & Barkema, 2001). Executives prefer international acquisitions over other entry modes when they need to procure advanced technologies or reputable brands, enter foreign markets quickly, improve capacity utilization, or consolidate market power (Anand & Delios, 2002; Chen, 2008; Hennart & Park, 1993). The few studies that examine managers’ equity ownership decisions in international acquisitions rely on an array of theories (Chari & Chang, 2009; Malhotra & Gaur, 2014).

The transaction cost perspective suggests that managers consider contractual hazards when choosing a high-control (i.e., majority acquisition) or low-control (i.e., licensing, alliance, or international joint venture) entry mode. Specifically, firms that are more vulnerable to acute contractual hazards (e.g., costly hold-up, opportunistic renegotiations) than political hazards (e.g., expropriation) often prefer a majority acquisition as a governance structure, because it can manage contractual hazards in a cost-effective way (Henisz, 2000b; Zhao, Luo, & Suh, 2004). The real options perspective centers on how value-maximizing firms optimally allocate their financial resources to international expansion. For example, Chari and Chang (2009) find that firms purchase partial ownership in foreign target firms to overcome their relative information disadvantages and to capitalize on real options, in the form of follow-on equity investments, when the uncertainty gets resolved. Based on an information asymmetry perspective, Chen and Hennart (2004) propose that acquirers may opt for minority acquisitions when concerned with a holdup problem by the target. In a minority acquisition, target firm managers will retain equity stake in the merged firm and thus will refrain from activities that will erode the wealth of the firm (Chen & Hennart, 2004; Chi, 1994). Other studies using geographic distance as a proxy for information asymmetry find that greater geographic distance leads to lower equity ownership in foreign target firms (Malhotra & Gaur, 2014; Ragozzino, 2009). Drawing on an institutional perspective (North, 1990; Scott, 1995), other scholars propose that the regulatory and cultural underpinnings of foreign countries define the observed pattern of entry-mode decisions. By accentuating the liability of foreignness (Zaheer, 1995), institutional differences may encourage firms to pursue legitimacy in foreign markets through imitative behaviors (DiMaggio & Powell, 1983), especially in international settings (Chan, Makino, & Isobe, 2006; Davis et al., 2000; Lu, 2002).

To extend this literature, we emphasize that the high levels of uncertainty and information asymmetry in international acquisitions impose substantial cognitive demands on top managers. Therefore, it is important to understand how heuristics and mental models shape international equity decisions.

Managerial Cognition and Heuristics

According to early behavioral economics, decision makers with incomplete knowledge use heuristics or simple rules in uncertain environments (Gigerenzer & Gaissmaier, 2011; Gigerenzer & Goldstein, 1996; Kahneman, 2011). This research draws heavily on work by Tversky and Kahneman (1974), who integrate psychological realism into standard optimization- or equilibrium-based models (Rabin, 2013). A key focus is how heuristics might lead decision makers away from accurate information processing or reasoning. Recent studies suggest that heuristics do not always cause biases (Gigerenzer & Gaissmaier, 2011). In particular, Gigerenzer and Goldstein (1996) find that heuristics economize information and support timely decisions.

Behavioral strategy scholars have extended this work by studying how heuristics in managerial decision making influence strategic decisions, organizational processes, and organizational outcomes (Hodgkinson & Healey, 2008; Kaplan, 2011; Narayanan, Zane, & Kemmerer, 2011; Priem, Walters, & Li, 2011). Managers cope with their cognitive limitations by relying on mental frameworks (Nadkarni & Barr, 2008) or “simplified knowledge structures or cognitive representations about how the business environment works” (Gary & Wood, 2011: 569). These mental models can simplify spatial, temporal, and causal relationships (Gavetti & Levinthal, 2000; Weick, 1979). They also help filter information from the external environment so that managers can focus on relevant information for strategic decisions (Huff, 1982; Narayanan et al., 2011). In particular, managers might limit their search to recent information and nearby locations when searching for cues in uncertain environments (Gavetti & Levinthal, 2000; Levinthal, 1997; Levinthal & March, 1993).

These mental models are likely influenced by executives’ demographics or personal characteristics (Narayanan et al., 2011). A central tenet of the upper echelons perspective is that top managers typically must interpret incomplete or ambiguous information when making strategic decisions (Hambrick & Mason, 1984). However, their mental models tend to alter their perception of reality and thereby affect their strategic choices (Hambrick & Mason, 1984; Narayanan et al., 2011). Using the observable demographic attributes of top managers as proxies for their mental models, previous research shows that various demographic characteristics (e.g., age, tenure, educational background, experience, political orientation) shape how they perceive their businesses and environments and thus affect their strategic decisions (e.g., Carpenter et al., 2004; Christensen et al., 2015).

Behavioral strategy research also recognizes a role for managerial cognition in the acquisition process. Hitt and Tyler (1991: 329) evoke an upper echelons perspective to describe the process for evaluating a target firm:

Executives of firms considering acquisitions should carefully analyze both their external environment and internal operations and use the resulting analysis to evaluate potential acquisitions … given the limits of human information processing capabilities, a top executive evaluating different potential acquisitions can be expected to simplify the decision process by limiting the criteria considered and by weighing some criteria more heavily than others.

Garbuio, King, and Lovallo (2011) find that managers are inclined to acquire resources with which they are most familiar, independent of their relative economic or strategic significance. Maitland and Sammartino (2015) examine how executives at an Australian mining company make judgments when undertaking an acquisition in a politically hazardous African country. They propose that the executives analyzed and adapted to this environment by constructing small-world representations based on their understanding of the complex relationships among different actors and events. Thus, different heuristics can be used to identify which information to look for (“discovery heuristics”) and how to interpret it (“evaluation heuristics”). These insights are consistent with previous research on the link between cognition and search heuristics (Gavetti & Levinthal, 2000; Haley & Stumpf, 1989). In the case of international acquisitions, we propose that executives may be inclined to decide on the level of international equity ownership by engaging in anchoring.

Anchoring Theory

Anchoring constitutes a particularly interesting heuristic that can be traced to Tversky and Kahneman’s (1974) landmark article on prospect theory. Decision makers become overly influenced by some initial value, which biases their final decision in that direction (Tversky & Kahneman, 1974). A growing body of research, predominantly in experimental psychology, has investigated anchoring effects across many decisions (for a review, see Furnham & Boo, 2011). The mechanism can be explained according to a selectivity accessibility model (Mussweiler, 2003; Mussweiler & Strack, 1999a, 2000). People begin their estimation decisions with some initial value in mind or some reference to what the response or decision could be, which is the anchor. In their evaluation process, they search selectively for information, in their memory or through an external search that is consistent with this anchor, which is the selectivity phase. When they evaluate the gathered information to make their final decision, the anchor-consistent information is more readily accessible; hence, their final decision is close to their initial anchor. This is the accessibility phase.

Prior work on anchoring was performed in controlled laboratory settings, which exposed participants to anchors that were implausible or irrelevant to the focal decision. However, recent studies indicate that anchoring also happens when anchors fall within an acceptable or plausible range (Mussweiler & Englich, 2005; Mussweiler & Strack, 1999b). Plausible anchors come more easily to mind, and it requires considerable effort to adjust away from them (Epley & Giolvich, 2006). For example, Northcraft and Neale (1987) find that real estate agents estimating the price of a house anchor their decision on its list price, even though those same agents previously had denied that list prices would influence their final estimation. Similarly, Shapira and Shaver (2014) find that managers undertaking project investment decisions anchor their decisions on average profit measures, which shifted their choices away from the optimal, profit-maximizing investments. Beggs and Graddy (2009) show that a previous sale price (anchor) of a painting, rather than its objectively estimated value, determines its subsequent sale price at auctions. As with heuristics in general, despite the prevalence of anchoring in many decision domains, a systematic investigation in real business settings is a recent phenomenon (Beggs & Graddy, 2009).

Previous and Focal International Acquisition Equity Levels

International acquisitions occur under high levels of information asymmetry, increasing the difficulty that executives face in collecting valuable and reliable information about the target firm (Ragozzino & Reuer, 2011; Reuer, Shenkar, & Ragozzino, 2004; Seth, Song, & Pettit, 2000). Executives may resort to simplified practices and a due diligence process that is not sufficiently informative or analytically rigorous (Haspeslagh & Jemison, 1991). Preliminary evidence suggests that anchoring occurs in M&As. For example, Baker, Pan, and Wurgler (2012) show that managers’ decisions about acquisition premiums depend significantly on target firms’ peak stock prices in the previous 12 months. Premiums may increase for firms whose stock price spiked even once in the previous 12 months. Similarly, Malhotra et al. (2015) show that acquiring firms engage in anchoring when making premium decisions. Consistent with these findings and with evidence that common business anchors include past events and decisions (Hammond et al., 1998), we argue that previous international acquisition equity decisions in the local market influence international acquisition equity ownership decisions.

The literature on managerial cognition argues that managers’ cognitive frameworks shape their attention to specific, salient information in the environment (Marcel, Barr, & Duhaime, 2011). In particular, Thomas, Sussman, and Henderson (2001) find that cognitive frameworks lead managers to focus predominantly on past information. According to Gavetti and Levinthal (2000), past practices often act as templates for managers’ future decisions. For example, managers set future sales expectations by using previous years’ sales figures (Hammond et al., 1998). Thus, a preceding international deal’s equity stake likely serves as an anchor point for decision makers.

A previous international acquisition equity ownership level offers a meaningful anchor for two reasons. First, it is likely that acquiring firm executives have been exposed to previous international ownership equity decisions, because they perform transaction analyses to help make strategic decisions (Rosenbaum & Pearl, 2009). As two directors of UBS investment banks who have handled many large acquisition deals note, preceding transactional analyses serve to “anchor valuations” (Rosenbaum & Pearl, 2009: 84). Precedent analyses involve evaluations of previous acquisitions in the industry to gauge firm quality and local market conditions. Because the local market’s previous international acquisition is the most recent international acquisition, it also should be more easily recalled and offer the most current information. Therefore, preceding international acquisition equity levels should be highly salient in executives’ cognitive frameworks, which further increases the likelihood that executives notice and attend to this number in their evaluation of the focal deal.

Second, foreign acquirers are more likely to anchor on previous international deals by other foreign acquirers, which is in line with the influence of reference groups on decision making. The managerial cognition literature emphasizes how reference groups shape managerial strategy (Fiegenbaum & Thomas, 1995; Narayanan et al., 2011). That is, managers use reference groups to access information, set aspiration levels, and select reference points (e.g., Epley, 2004; Massini, Lewin, & Greve, 2005). People are often drawn to information from reference groups and use it to structure their perceptual fields (Sherif, 1953) and set strategies (Fiegenbaum & Thomas, 1995). Moreover, managerial attention increasingly focuses on information from reference groups because of its relevance. For example, international acquirers face regulations, market entry barriers, and social contexts that differ from those of domestic acquirers. International decision makers thus associate with other international acquirers as reference groups in similar situations. These combined features should induce acquiring firm executives to use the previous equity ownership level of other international acquirers as an anchor when making their own international acquisition equity ownership-level decisions.

Uncertainty and Anchoring Behavior

Anchoring effects depend on the level of uncertainty that surrounds the decision in question (Mussweiler & Strack, 2000). We consider two settings in which acquiring firm executives likely confront nontrivial informational deficiencies and learning challenges. In the first setting, a foreign target firm might operate in a different industry than the acquiring firm. Industry relatedness between the acquiring and target firms is an important determinant of uncertainty and information asymmetry in international deals (Chari & Chang, 2009; Malhotra & Gaur, 2014; Malhotra, Sivakumar, & Zhu, 2011). When foreign target firms operate in unrelated industries, acquiring firm executives confront unfamiliar circumstances and informational barriers, such as limited practical knowledge about the unrelated industry’s shared norms, procedures, and standards (Huff, 1982). By pursuing foreign target firms in unrelated industries, acquiring firms widen the gap between the information and experience they need to evaluate and execute an international acquisition and the knowledge and experience they already have accumulated in their own industry. This enlarged gap could accentuate information asymmetry problems, such that executives have relatively little information about the foreign target firm’s true potential or the future prospects for its industry (Chari & Chang, 2009; Chen & Hennart, 2004; Malhotra & Gaur, 2014; Reuer & Koza, 2000). These information deficiencies thus suggest that acquiring firms may find it difficult and costly to conduct information-intensive analyses, in terms of executive time, financial resources, and cognitive resources. Anchoring is particularly likely to emerge in this setting.

In the second setting, executives might face informational deficiencies when they pursue foreign target firms in unstable or highly volatile countries. Multinational enterprises risk being exploited when they invest in countries in which governments are not legally or politically restrained (Henisz, 2000a). In response to such political hazards, executives might employ mitigating governance structures, such as a majority acquisition (Henisz, 2000b). The inherently unstable nature of a politically hazardous environment prevents long-range planning and forecasting exercises that are standard business practices in institutionally developed environments, thus creating a role for heuristics. As Maitland and Sammartino (2015) show, senior executives attempt to cope with uncertainty in politically unstable environments by forming perceptions about complex structural relationships among different actors and events and their potential influence on those relationships. We propose that the same informational challenges could lead acquiring firm executives to exhibit anchoring behavior. Considering both channels, we predict the following:

Overconfidence and Anchoring Behavior

In keeping with an upper echelons perspective (Hambrick & Mason, 1984), previous research has linked aggressive bidding behavior by acquirers to hubris or overconfidence in their senior executives (Hayward & Hambrick, 1997; Roll, 1986; Seth et al., 2000). Overconfidence in senior executives may be explained in terms of differences in their self-assessments. For example, based on the concept of core self-evaluation, overconfident CEOs are likely to exhibit higher levels of self-esteem, self-efficacy, internal locus of control, and/or emotional stability than other CEOs (Hiller & Hambrick, 2005). On a whole, this implies that overconfident CEOs might have higher levels of core self-evaluation than other CEOs. The potentially enhanced levels of core self-evaluation in overconfident CEOs may also be manifested in their tendencies to judge themselves to be better than average (Malmendier & Tate, 2005). In addition, they might overestimate their abilities and prospects for success on particularly difficult tasks (Cain, Moore, & Haran, 2015; Windschitl, Kruger, & Simms, 2003). When undertaking tasks as difficult as international acquisition deals, CEOs may exhibit overconfidence because of the tendency to overestimate their ability or chance of success. We also expect overconfident CEOs to favor the accuracy of their memories, simple rules, and judgments over others’ (cf. Dunning, Griffin, Milojkovic, & Ross, 1990; Talarico & Rubin, 2003; Wells & Olson, 2003). Thus, when faced with international equity ownership decisions, they are predisposed to activate and act on knowledge from their own memories, simple rules, and judgments; hence, the previous levels of equity purchased in foreign target firms by others is less likely to emerge as an anchor among overconfident CEOs. Therefore, we propose the following:

Sample

We collected information on completed international acquisition deals between 1990 and 2009 from the Securities Data Corporation Platinum database. In line with prior studies, we omitted transaction values of less than $1 million (e.g., Fuller, Netter, & Stegemoller, 2002; Moeller, Schlingemann, & Stulz, 2005). We then merged firm-level information from the Compustat Global database and country-level information from the World Development Indicator database. We collected institutional measures from the World Bank. For our empirical analysis, we also collected firm-level information for the most recent preceding international acquisitions in the host country and target industry (anchor). After we merged all variables and removed missing values, the final sample of observations with complete data included 4,491 deals. The acquirers’ sample includes observations from 50 countries and covers 55 industries (using two-digit standard industrial classification [SIC] codes). The median deal value was US$37 million.

Dependent and Independent Variables

We determined the ownership acquired in the focal international acquisition from the Securities Data Corporation. This continuous variable ranges from 0.1% to 100% and is similar to measures used in other studies (Chari & Chang, 2009; Delios & Henisz, 2000; Malhotra & Gaur, 2014). The key explanatory variable is ownership acquired in the international acquisition preceding the focal deal in the host country and target industry (by four-digit SIC code). We calculated the ownership variable for the previous international acquisition by using a method similar to that used to measure ownership of the focal international acquisition.

In line with prior studies (Malhotra & Gaur, 2014; Malhotra et al., 2011; Reuer et al., 2004), we assessed deals in unrelated industries with a dummy variable that equals 1 if the acquiring firm and the target firm did not share the same four-digit SIC code and zero otherwise. To measure political stability, we used the Worldwide Governance Indicators developed by the World Bank, which range from -2.5 (high political risk) to 2.5 (low political risk).

Following Malmendier and Tate (2008) and Campbell et al. (2011), we created a CEO overconfidence dummy variable equal to 1 if the acquiring firm’s CEO has more than 67% “in-the-money” stock options (i.e., the stock price exceeds the exercise price by more than 67%) and 0 otherwise. ¹ When stock options are significantly in the money, a CEO who holds on to the option contract is considered to be highly confident about future stock performance. This measure has been found to affect various managerial and financial decisions (e.g., Deshmukh, Goel, & Howe, 2013; Galasso & Simcoe, 2011; Hirshleifer, Low, & Teoh, 2012; Malmendier & Tate, 2005, 2008). We gathered the CEO option information from the ExecutiveComp database. Because CEO compensation information is available for U.S. firms only, we tested Hypothesis 3 on a subsample of U.S. international acquisitions.

Control Variables

We controlled for several plausible explanations for acquiring firms’ foreign equity ownership structure. First, the transaction cost perspective contends that firms acquire greater ownership levels to counter opportunistic partners (Brouthers & Hennart, 2007; Chang & Rosenzweig, 2001). Opportunism increases with high asset specificity and when investments occur in uncertain environments. In line with other studies (e.g., Delios & Beamish, 1999; Mani, Antia, & Rindfleisch, 2007), we used R&D intensity to measure asset specificity, which is the R&D expense divided by the total revenue of the acquiring firm in the latest fiscal year-end before the focal acquisition. To account for environmental uncertainty, researchers commonly use the cultural distance between the home and host country (Erramilli & Rao, 1993; Luo, 2001; Padmanabhan & Cho, 1996). We accordingly adopted Hofstede’s (1980) four cultural dimensions (individualism, uncertainty avoidance, power distance, and masculinity), but we followed Kogut and Singh’s (1988) method and combined them into one composite variable.

Second, parent firms’ international experience provides them with capabilities that should influence their foreign equity ownership decisions (Delios & Beamish, 1999; Delios & Henisz, 2000). We controlled for three types of international experience—number of international acquisitions, international joint ventures, and international alliances—entered into by the acquiring firms in the 5 years before the focal acquisition.

Third, researchers have used real options to explain acquiring firms’ foreign equity ownership structure (Chari & Chang, 2009; Folta, 1998). We included several variables to control for real options. We measured target industry uncertainty by differentiating high-tech from non–high-tech industries. Using Loughran and Ritter’s (2000) definition of high-tech industries, our high-tech variable equals 1 if the target industry falls in the high-tech industry group and 0 otherwise. We also followed Beckman, Haunschild, and Phillips (2004) and used the average stock price volatility as another measure of industry uncertainty. Interest rates and foreign exchange rates also affect real options, because they determine the present value of the exercise price. We calculated the country-level interest rate variable by collecting the interest rate of the target country for each sample year from the World Development Indicator database. To calculate the effect of the foreign exchange rate, we noted the exchange rate of both acquiring and target countries in each sample year, then calculated the percentage change in the relative exchange rate between them from the beginning to the end of that year. A positive value means that the acquiring country’s currency appreciated against the target country’s currency in that year. In the regression model, we included the prior year’s interest rate and exchange rate measures. ²

Fourth, scholars have adopted institutional theory to explain foreign ownership structure. We controlled for a bandwagon effect that could encourage a certain entry mode strategy (Henisz & Delios, 2001; Xia, Tan, & Tan, 2008). We used a measure similar to Xia et al.’s (2008) and calculated the average foreign equity acquired in the host country industry during the year before the year of the focal deal. In addition to signaling any bandwagon effect, this variable controls for a general institutional effect on foreign ownership levels, including different regulatory restrictions on foreign ownership structures across countries and industries. An M&A wave in the host country also could cause a bandwagon effect. To control for this market wave effect, we include elapsed time, which equals the number of days between the focal acquisition and the previous acquisition in the target country and industry. We also counted the number of international acquisitions in the target country and industry in the past 12 months. Many acquisitions in a short period signal a peak in an M&A wave, so we controlled for this market sentiment effect.

Fifth, specific to international acquisitions, we controlled for the difficulty of evaluating target firms, which might explain the foreign equity structure of acquiring firms (Chari & Chang, 2009; Chen & Hennart, 2004). Large transactions are more complex and involve more costly due diligence than smaller transactions; therefore, we used transaction size to proxy for valuation difficulty (Chari & Chang, 2009). We also controlled for private versus public target firms. Private firms are not required to disclose information, and fewer analysts follow them, which increases the difficulty of evaluating them.

In addition to these theoretical explanations, we controlled for important transaction characteristics, such as tender offers, friendly versus hostile acquisitions, and payment method (stock payment = 1, cash or mixed payment = 0). Finally, we noted the target-country GDP growth rate, because rapidly growing countries provide more growth opportunities to multinational firms and attract more foreign direct investment (Xia et al., 2008).

Analytical Approach

We used a conventional regression method to test our hypotheses. The dependent variable (foreign ownership level) is bounded between 0.1% and 100%, so we used a Tobit regression method that addresses censored data, similar to prior equity ownership studies (Chari & Chang, 2009; Delios & Henisz, 2000; Malhotra & Gaur, 2014). Our sample covered multiple countries and industries, so we also included acquirer-country and industry dummy variables in the models. We controlled for year-specific variations in foreign ownership levels by incorporating year dummies from 1990 to 2009 in the regression model. Finally, in line with similar studies (Chari & Chang, 2009; Folta & Miller, 2002; Malhotra & Gaur, 2014), we calculated clustered standard errors to account for multiple international acquisitions by the same firm.

Supplementary Analyses

Our main results using Tobit regressions support our hypotheses. However, the positive correlation between the previous and the focal international equity ownership levels may also be because previous deals vicariously provide foreign acquirers with relevant insights, such as the quality of target firms in a given industry. There is no direct measure for vicarious learning, but this omitted variable could bias our inferences (Reeb, Sakakibara, & Mahmood, 2012). Therefore, we used an identification strategy to separate this vicarious influence from anchoring behavior.

Accordingly, we adopted an accepted methodology from the economics literature, developed by Genesove and Mayer (2001) and Beggs and Graddy (2009) and based on a hedonic regression approach. Specifically, we identified the key determinants of the international equity ownership decision (i.e., the previously described control variables). We then regressed the international equity ownership level of both the focal and previous international acquisition deals on these control variables. (To perform this regression, we also collected data for the control variables that related to the previous deal.) With this information, we predicted values for both the focal and previous international equity ownership levels, using a first-step hedonic regression model,

where R t is the predicted focal international equity ownership level, a function of the time t observable characteristics X t (i.e., control variables to explain the focal deal ownership), and

where R t − 1 is the predicted previous international equity ownership level, a function of the preceding deal’s observable characteristics X t − 1 (i.e., control variables to explain the previous deal ownership).

Then, we entered these predicted values in a second-stage estimation:

where Y t is the actual focal international equity ownership level purchased by a foreign-based acquirer in a given target country and industry, Y t − 1 is the actual previous international equity ownership level purchased by another foreign-based acquirer in the same target country and industry, and ε t is the error term. In equation 3, η captures the effect of the predicted focal international equity ownership level ( R t ) on the actual focal international equity ownership level; Y t − 1 − R t − 1 captures the deviation of the previous international equity ownership level from the past prediction of what that level should be; θ captures the effect of the stable, but unobserved or hard-to-measure characteristics of M&A deals in a given foreign country and industry. ⁴

The variable Y t − 1 − R t is our main focus, and the coefficient δ captures the anchoring effect we intend to estimate. That is, Y t − 1 − R t captures the deviation of the previous international equity ownership level from the current prediction of what the actual focal international equity ownership level should be. If foreign-based acquiring firms engage in anchoring, we expect that their decision to purchase a certain level of equity ownership in target firms in a given foreign country and industry will be strongly influenced by the previous equity ownership level purchased by other foreign-based acquirers in the same target country and industry. This level should be systematically different from what we objectively expect the focal international equity ownership level to be. Our baseline “no anchoring effect” null hypothesis is consistent with δ = 0. Intuitively, this hypothesis implies that the previous international equity ownership level does not systematically deviate from our most informed guess of what the actual focal international equity ownership level should be. ⁵ Therefore, it has no significant effect on the focal international equity ownership level when compared with a level that is justified by a fundamental analysis (predictive model). Contrary to the baseline hypothesis, if δ is statistically significant and positive, acquirers engage in anchoring when making international equity ownership decisions. The previous international equity ownership levels not only systematically deviate from reasonable expectations of what the acquirer’s focal international equity ownership levels should be but also exert significant, potentially unwarranted influences on the acquirer’s actual international equity ownership decisions.

Models 1 and 2 in Table 3 estimate the hedonic regressions for the previous acquisition in the target industry (the anchor) and the focal acquisition in the target industry, respectively.

OPEN IN VIEWER

Table 3

Hedonic Regression Models

Variables	Previous Deal Hedonic Model	Focal Deal Hedonic Model
R&D intensity	.51***	.18**
	(.11)	(.07)
Cultural distance	−.02***	−.02***
	(.01)	(.01)
Intl. acq. exp.	−.00	−.01**
	(.01)	(.01)
Intl. JV exp.	−.09***	−.06***
	(.02)	(.01)
Intl. alliance exp.	−.06***	−.05***
	(.02)	(.01)
Industry relatedness	.02	.05***
	(.01)	(.01)
High-tech target industry	−.00	.01
	(.02)	(.01)
Target country interest rate	−.07	−.01
	(.07)	(.06)
Exchange rate	.07	.03
	(.05)	(.05)
Average foreign ownership	.00***	.00***
	(.00)	(.00)
Elapsed time	.01***	.02***
	(.00)	(.00)
Target industry intl. acq.	.00	−.00
	(.00)	(.00)
Institutional distance	−.01	−.01
	(.01)	(.01)
Transaction value	.00***	.00***
	(.00)	(.00)
Private target	.07***	.06***
	(.01)	(.01)
Tender offer	.20***	.24***
	(.02)	(.01)
Friendly acquisition	.19***	.20***
	(.02)	(.01)
Stock payment	.00	.00
	(.00)	(.00)
Target GDP growth	.08	−.15
	(.24)	(.21)
Intercept	.59	.46
	(.40)	(.35)
Year, industry, and country fixed effect	Yes	Yes
Observations	2,862	4,491
R-square	.29	.32

Note: Standard errors are clustered on acquiring firms. R&D = research and development; intl. = international; acq. = acquisitions; exp. = experience; JV = joint venture; GDP = gross domestic product.

**

p < .05.

***

p < .01.

Using the predicted values for both the focal and previous international equity ownership levels, we estimated the anchoring regression model in Table 4. To assess the comparable size of the regression coefficients, we standardized all the variables. As Table 4 shows, the coefficient of the predicted focal international equity level ( η in equation 3) has a significantly positive effect on the actual focal international equity level, so our control model is a good predictor of the international acquisition equity ownership level. The regression coefficient for the deviation of the previous international equity ownership level from its predicted level also is positive and significant ( θ in equation 3); hence, the relatively stable but unobservable or hard-to-measure characteristics of preceding international deals affect the focal international equity ownership decisions.

OPEN IN VIEWER

Table 4

Anchoring Regression Models

Step 2 Anchoring Regression (Tobit Model)	Model 1
Focal predicted equity level ( η )	.61***
	(.02)
Anchoring effect (preceding – predicted equity acquired) ( δ )	.13***
	(.02)
Vicarious learning ( θ )	.05***
	(.02)
Constant	.22***
	(.05)
Year, industry, and country fixed effect	Yes
Observations	2,862
Model χ2	1,239.85***

Note: Standard errors are clustered on acquiring firms. The following model was estimated in the above regression:

Y t = η ⋅ R t + θ ⋅ ( Y t − 1 − R t − 1 ) + δ ⋅ ( Y t − 1 − R t ) + ε t .

Other sensitivity analyses: The results are robust to various sampling variations. First, our results are robust after removing U.K. and U.S. acquirers, which accounted for 34% of the 2,862 international acquisitions in our sample. Second, our results are robust after removing international acquisitions whose transaction values were less than $50 million. Third, our results are robust after excluding acquisitions in regulated industries, such as the financial industry or utilities. Finally, our results are robust after we removed any entries for which the focal and preceding international acquisitions were too close (100 days or less)—this ensures that our results are not due to clustering pattern of M&A transactions.

***

p < .01.

The coefficient for the anchoring effect ( δ in equation 3) is significantly positive (p < .01), and the standardized regression coefficient for this anchoring effect (.13) is twice that for the relevant information (.05). In terms of the economic significance of the anchoring effect, a 1 SD difference between the market’s previous international equity ownership level and the predicted focal international equity ownership level causes the anchor (i.e., the previous international equity ownership level) to pull the actual focal international equity ownership level to itself by nearly 13.3% of the standard deviation of ownership acquired in the focal acquisition. The standard deviation of the focal international equity ownership level is approximately 37%, so the effect of anchoring is approximately 4.9%. In our sample, the average market capitalization of the target firms was approximately US$1.1 billion. Anchoring would affect the average transaction value of an international acquisition by US$54 million. Our results also remained robust to other variations, as summarized in the footnote to Table 4.

Tests for Alternative Explanation

One alternative explanation for our results could be institutional isomorphism (DiMaggio & Powell, 1983; Lieberman & Asaba, 2006), in which acquiring firms imitate the equity stakes previously purchased by others to enhance their performance or bolster their legitimacy in the local market. Deliberative and sociological processes underscore institutional isomorphism. We expect managers who engage in mimetic behavior to search for firms or practices worth imitating. Organizations can bolster their survivability in uncertain markets by mimicking policies or practices that are perceived as legitimate. To separate our anchoring perspective from this institutional isomorphism perspective, we investigated whether anchoring effects persist, even when acquisition behavior falls short of legitimacy, and are thus potentially disadvantageous to an imitator.

Specifically, we considered hostile takeovers. When bidding firms make hostile offers, they circumvent top management and directly approach the shareholders. Hostile takeovers usually are a last resort (Gaughan, 2011). They spark resistance from top management (Gaughan, 2011; Schneper & Guillén, 2004) and workers (Schneper & Guillén, 2004). In international acquisitions, top managers of foreign target firms might lobby the host government to intervene. For example, during Mittal Steel’s hostile bid for France’s Arcelor, public opposition spread to then-President Jacques Chirac. Thus, by virtue of their unwelcome nature, hostile offers are unlikely to attract imitators. If anchoring effects capture legitimacy-seeking mimetic behavior, they should not arise in the case of hostile offers.

To test this, we ran our analyses on a subsample of international acquisitions in which the previous international acquisition in the local market was a hostile offer. We considered two measures, unfriendly deals and tender offers, which we collected from the Securities Data Corporation Platinum database. Most hostile acquisitions occur through tender offers (Browne & Rosengren, 1987; Gaughan, 2011). According to Models 1 and 2 in Table 5, a significantly positive relationship exists between previous and focal international acquisition equity levels for prior tender offers (p < .01) and unfriendly deals (p < .05). Thus, an anchoring effect is present even when the previous acquisition was hostile.

OPEN IN VIEWER

Table 5

Tests for Alternative Explanations

Dependent Variable	1	2	3	4	5
Focal Acquisition Ownership Acquired	Tender Offers	Unfriendly Deals	Unsuccessful Deals	Focal and Previous Acquirers From Different Countries	Previous Equity Level Exceeds Average Industry Deviation
Previous equity level	0.09***	0.04**	0.17***	0.06***	0.09***
	(0.01)	(0.02)	(0.03)	(0.01)	(0.01)
Control variables
Industry unrelatedness	−0.14***	−0.21***	−0.09***	−0.15***	−0.16***
	(0.00)	(0.01)	(0.01)	(0.00)	(0.01)
Target country political stability	0.01**	0.01	0.01	0.01**	0.01*
	(0.00)	(0.01)	(0.01)	(0.00)	(0.01)
R&D intensity	0.14***	0.18	0.15*	0.14***	0.02
	(0.04)	(0.16)	(0.09)	(0.04)	(0.07)
Industry stock volatility	0.01	−0.01	0.03	0.01	0.02**
	(0.01)	(0.01)	(0.02)	(0.01)	(0.01)
Cultural distance	−0.02***	−0.03***	−0.02***	−0.02***	−0.02***
	(0.00)	(0.00)	(0.01)	(0.00)	(0.00)
Intl. acq. exp.	0.01***	0.02***	0.01	0.01***	0.02***
	(0.00)	(0.01)	(0.01)	(0.00)	(0.00)
Intl. JV exp.	−0.04***	−0.04***	−0.01	−0.03***	−0.04***
	(0.01)	(0.01)	(0.04)	(0.01)	(0.01)
Intl. alliance exp.	−0.02**	−0.04***	−0.03*	−0.02***	−0.02**
	(0.01)	(0.01)	(0.02)	(0.01)	(0.01)
High-tech target industry	0.01	−0.00	−0.00	−0.01**	0.02*
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Target country interest rate	−0.01	−0.23***	−0.23**	0.01	0.09***
	(0.02)	(0.03)	(0.10)	(0.03)	(0.03)
Exchange rate	−0.01	0.10***	−0.04	−0.01	−0.06***
	(0.02)	(0.03)	(0.04)	(0.02)	(0.02)
Average foreign ownership	0.00***	0.00***	0.00***	0.00***	0.00***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Elapsed time	0.01***	0.02***	0.03***	0.01***	0.02***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Target industry international acq.	0.00	0.00	0.00*	−0.00	0.00***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Transaction value	0.00***	0.00***	0.00***	0.00***	0.00***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Private target	0.01***	0.06***	0.07***	0.02***	0.02***
	(0.00)	(0.01)	(0.01)	(0.00)	(0.01)
Tender offer	0.45***	0.48***	0.43***	0.47***	0.48***
	(0.01)	(0.02)	(0.02)	(0.01)	(0.01)
Friendly acquisition	0.16***	0.13***	0.09***	0.17***	0.17***
	(0.01)	(0.01)	(0.02)	(0.01)	(0.01)
Stock payment	0.00***	0.00***	−0.00***	0.00***	0.00***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Target GDP growth	−0.17**	−0.21	0.82**	−0.32***	0.05
	(0.08)	(0.17)	(0.34)	(0.08)	(0.11)
Inverse mill’s ratio	−3.34***	−3.61***	−2.89***	−3.52***	−3.51***
	(0.05)	(0.08)	(0.16)	(0.04)	(0.06)
Intercept	0.53***	0.77***	0.78***	0.59***	0.54***
	(0.02)	(0.04)	(0.07)	(0.02)	(0.03)
Year, industry, and country fixed effect	Yes	Yes	Yes	Yes	Yes
Observations	3,920	586	472	3,123	2,124
Chi-square	2,424.68***	528.52***	369.71***	1,967.11***	1,383.44***

Note: Standard errors are clustered on acquiring firms. R&D = research and development; intl. = international; acq. = acquisitions; exp. = experience; JV = joint venture; GDP = gross domestic product.

*

p < .10.

**

p < .05.

***

p < .01.

Similarly, acquiring firms that seek to imitate legitimate behavior should be reluctant to imitate other firms that make unsuccessful international acquisition bids, because such failure might emanate from the bidder’s departure from business norms and values. Furthermore, considering the potential decline in social support for a failed foreign bidder in local markets (Ahlstrom & Bruton, 2001; Bianchi & Ostale, 2006), other aspiring bidders may avoid practices that mimic those of the failed party. In line with institutional isomorphic reasoning, if a prior international acquisition bid was not perceived as successful, the focal acquirer has little reason to mimic its expressed equity ownership interest. To assess whether the previous international acquisition was successful, we used a short-term event study methodology and collected the daily stock price of the prior acquiring firm around the deal announcement period from DataStream. We also collected the corresponding stock market index for the same period. Similar to Brown and Warner (1985), we used standard market-adjusted models to calculate daily abnormal stock returns around the acquisition announcement date for each acquiring firm. That is, we focused on the event window t − 1 to t + 1 (one day before and one day after the announcement date) and calculated the daily abnormal return as follows:

where A R j , t was the daily abnormal return for firm j on day t, R j , t was firm j’s daily stock return on day t, and R m , t represented the daily return of the local stock market index on day t. We then added daily abnormal returns to measure the cumulative abnormal return (CAR) for acquiring firm j in the 3-day period (–1, +1) surrounding the acquisition announcement:

An announced international acquisition deal is successful, from the perspective of stock market investors and shareholders, if the CAR measure is positive, and it is unsuccessful if the CAR measure is negative. Model 3 in Table 5 shows the results for our analyses of a sample of unsuccessful international acquisitions. We still found a significantly positive relationship between the previous and focal international acquisition equity levels (p < .01). That is, an anchoring effect persists even when the previous acquisition was perceived as unsuccessful.

In line with the sociological foundation of the institutional isomorphism perspective, social or business networks constitute potential channels through which firms may externally acquire knowledge (DiMaggio & Powell, 1983). Imitators with social or relational ties to leading firms could have an advantage in terms of a better understanding of why and how leaders achieve legitimacy or superior performance (Haunschild, 1993; Haunschild & Beckman, 1998; Kostova & Roth, 2002). For example, prior research suggests that mimetic acquisition behavior is especially prevalent among acquirers with network ties in the form of board interlocks (Haunschild, 1993; Haunschild & Beckman, 1998). We tested whether our results hold if prior and focal foreign acquirers are not part of a board interlock. As acquirers from different countries are unlikely to be in the same interlocked board, we selected a subsample where both the focal and prior foreign acquirers are from different home countries. The results (Model 4 in Table 5) show a significantly positive relationship between the previous and the focal international acquisition equity levels (p < .01).

Finally, it can be argued that legitimacy-seeking foreign firms would not mimic the preceding acquisition equity level if it diverted from the industry norm. We measured the industry norm as the average level of equity purchased in foreign target firms of a given country and industry. If industry players tolerate deviations in equity purchases up to the average deviation from the industry norm, then the institutional isomorphism perspective suggests that legitimacy-seeking foreign acquirers should not imitate other foreign acquirers whose prior level of ownership of foreign target firms in the same industry exceeds the average deviation from the industry norm. Model 5 in Table 5 shows that, for a subsample of deals where the previous equity level exceeded the average deviation, the relationship between prior and focal international equity levels is significantly positive (p < .01). Overall, our critical tests show that institutional isomorphism is not a likely explanation for our results.