FormalPara Definition

Risk-taking refers to decision-making under risk or the willingness to accept the risk involved with a decision. Managerial risk-taking focuses on how managers in organizations perceive risk and make risky decisions. Manager’s willingness to take risk depends on her firm’s wealth position relative to certain reference points – aspiration point, success point, and survival points – as well as her characteristic as an individual decision-maker and the incentive structure in which she is in.

Body of Text

Risk-taking refers to making decision under risk. In this article, we discuss how strategic decisions under risk are made. We will specifically address how managers make decisions involving risk and how they perceive and interpret risk. We will then review some of strategy research on risk-taking, mainly focusing on how CEOs and top management teams make risk-taking decisions. Strategy research employs multiple methodologies, which range from interviews to formal modeling, in addressing such research question.

Managerial Risk-Taking

Manager’s perception or definition of risk is different from the assumptions made in normative models of risk-taking. Managers are familiar with the expected utility rule, but they do not use it (MacCrimmon and Wehrung 1986). More specifically, managers do not have a mean and variance approach to risk in that they only consider the downside risk of potential outcomes of their choices (March and Shapira 1987; Shapira 1995). Moreover, Shapira showed that managers care more about the magnitude of possible outcomes than the probabilities associated with the outcomes. Managers are most concerned about large magnitude of potential losses, and their definition of risk is whether there is a possibility of incurring huge losses.

In addition, taking risk when investing in financial markets is done in a passive manner since investors cannot influence what goes on in such market. However, Shapira also found that managers do not passively accept risk (in settings other than pure financial markets) in that they do not view risk to be inherent in the situation that they find themselves in. Instead, managers believe that risk can be managed. Managing risk, managers argue, is what differentiates risk-taking from gambling. In fact, the many successful executives are the biggest risk takers, but the more mature executives are more risk averse (MacCrimmon and Wehrung 1990). Managers also tend to be overly optimistic in making forecasts but, at the same time, overly cautious in making choices (Kahneman and Lovallo 1993).

Performance Feedback and Subsequent Risk Taking

Decision-maker’s risk-taking propensity, however, is not constant over time. In fact, decision-makers learn through making decisions and receiving feedback based on the decisions, and this feedback affects their risk-taking propensity. Investors suffer from “myopic loss aversion,” which refers to the combination of loss aversion and their tendency to evaluate equity portfolio frequently (Benartzi and Thaler 1995, p. 75). Myopically loss averse investors, however, take more risk when they evaluate their equity portfolio less frequently (Thaler et al. 1997). As the frequency of performance evaluation increases, decision-makers become more risk averse (Gneezy and Potters 1997). Decision makers are more risk averse for gains than for losses, thus, if they conform to standard learning models, they learn to favor less risky alternatives in the positive domain (March 1996) and to favor a certain outcome over an uncertain alternative (Denrell 2007). When such adaptive learning process is slow and imprecise, however, the likelihood of engaging in risky activities increases (Denrell and March 2001).

Performance feedback has a strong impact on firms’ risk-taking. In analyzing a large and longitudinal data set, Ref and Shapira (2017) show how performance feedback, when a firm is near its aspiration point as well as far away from it, can lead firms to change their risk-taking in a significant manner.

Decision Targets and Risk Preference

Although managers appear not to follow the mean and variance approach in risk-taking, researchers have investigated the empirical relationship between mean and variance of firm performance. Bowman’s (1980, 1982) paradox refers to the negative relationship between mean and variance of firm performance. More specifically, Bowman found that firms with higher average profit have lower variability in their profits as well. This means that low-performing firms are more risk-taking than high-performing firms. Low performance and lack of slack increase risk-taking, but the risks that are taken may also lead to poor returns (Bromiley 1991). Fiegenbaum and Thomas (1988) offered a prospect theory (Kahneman and Tversky 1979) type explanation to the paradox. They showed that a firm’s target returns play an important role in its risk-taking behavior. They found that firms that are below their return target are risk seeking, whereas firms that are above their return target are risk averse.

Similarly, risk preference depends on the decision-maker’s position relative to her standard (Fishburn 1977). Laughhunn et al. (1980) reported similar findings to what Fiegenbaum and Thomas (1988) did. More specifically, Laughhunn, Payne, and Crum found that managers of firms below their target returns were more risk seeking than those above their target returns. What is more interesting is that this relationship is present only when the potential for what they term as ruinous losses was not present. In other words, managers of firms below their target returns are risk seeking only if there is no possibility of incurring an unacceptable magnitude of loss. When the potential for ruinous loss is present, managers do not become risk seeking even if the firms are positioned below their target returns. This finding highlights that managers not only care about their firms’ position relative to their target returns but that they also care about other factors, such as the possibility of incurring an unacceptable magnitude of loss, that affect their risk preferences. These risk tendencies are explained by the model to be presented next.

The Variable Risk Preference Model

Another reference point that needs to be considered is the survival point (March 1988; March and Shapira 1992). March and Shapira argued that decision-makers have two reference points – aspiration point and survival point – to which they attend to one at a time. According to the behavioral theory of the firm, as performance level falls below the aspiration point, firms engage in more search and more risky behaviors (Cyert and March 1963). Building on this point, managers are thought to attend to one of the aforementioned reference points and which reference point they pay attention to affects their risk preferences. Survival point is the point at which all cumulative resources are exhausted, and the aspired level of return is the target performance level, which is thought to adapt and change through experience (Lewin et al. 1944; Shapira 1995). When focusing on survival, a manager’s risk preference has a monotonically increasing relationship with the total cumulated resources. In other words, as the magnitude of the total cumulated resources increase, managers take more risk. Under aspiration focus, managers are extremely risk seeking at very low levels of the total cumulated resources, but they become less risk seeking as they approach their aspiration level. Once they position above their aspiration level, they become risk averse, but then they start to become more risk seeking again as they move further above their aspiration level (Fig. 1).

Risk-Taking, Fig. 1
figure 183figure 183

Risk taken as a function of cumulated resources for fixed focus models of variable risk (March and Shapira 1992, p. 175)

Full size image

The model therefore has two main drivers, position and focus, which may lead firms to take different risks form the same asset position if they focus on different reference points. For example, firms that find themselves way below their aspiration level but still focus on it will take much higher risk than firms who find themselves in the same asset position but focus on the survival point.

Shapira interviewed practicing managers and showed that managerial risk taking, in fact, closely adhere to the descriptive model, and through analyzing organizational data, Miller and Chen (2004) found further evidence for this model. The variable risk preference model has been utilized to complement the transaction cost economics theory’s inadequate behavioral assumption of risk neutrality (Chiles and McMackin 1996). It was used to explain betting behavior of contestants in the Jeopardy! television game (Boyle and Shapira 2012) and has been further extended to account for three reference points: survival point, aspiration point, and success point (Hu et al. 2011).

Signal Detection Theory and the Project Selection Model

Signal detection theory examines how decision-makers handle signal and noise. Shapira (1995) used this framework to examine the values and utilities of outcomes that were prone to erroneous judgments and analyzed such decisions in terms of the type I and type II errors. The variable risk preference model depicts how manager’s risk preference changes as her attention shifts between the two reference points. While the variable risk preference model describes manager’s risk preference prior to making any strategic decision, the project selection model reflects how a manager’s different risk attitude affects project selection and its outcome evaluation. The project selection model describes how manager’s ex ante evaluation of a project is linked with the project’s ex post outcome evaluation and when strategic surprises occur (Lampel and Shapira 2001; Shapira 1995).

Managers accept any project with the ex ante evaluation that is above a certain threshold xc, and they reject projects if the ex ante evaluation is below that critical value. Similarly, projects with outcomes above a certain threshold yc are considered as successes, while projects with outcomes below that critical value are considered as failures. The values of xc and yc determine the likelihood of project success, project failure, and strategic surprise. The project selection model has its roots in the signal detection theory (Green and Swets 1966). The signal detection theory, however, focuses on the probability side of signal detection, while the project selection model highlights the utility aspect of project selection and evaluation. The project selection model has been used to study project selection, technological foresights and oversights (Garud et al. 1997), risk-sharing incentive contracts (Shapira 1993), and hurricane evacuation decision (Dye et al. 2014).

Risk-Taking and the Upper Echelon Perspective

So far we have referred to several strategy research studies that address the topic of risk-taking in organizations as we explicated the details of the two models: the variable risk preference model and the project selection model. Other strategy research that studies risk-taking behavior often adopt the upper echelons perspective (Hambrick and Mason 1984), which posits that strategic decisions and performance levels of organizations depend on the characteristics of top managers. Since this article focused mainly on managerial risk-taking, we briefly review some of the studies that adopt the upper echelons perspective in studying managerial risk taking. These studies can be broadly classified as either linking CEO traits to her risk-taking decisions or investigating the role of agency (Eisenhardt 1989) and incentive in CEO risk-taking. Researchers that study CEO traits generally assume risk attitude to be a stable personality characteristic, whereas researchers that look at the role of agency and incentive on CEO risk-taking view risk attitude to be context dependent. The proxy behaviors that are interpreted as high level of risk-taking are often R&D expenditure, mergers and divestitures, and the adoption of a new technology or innovation.

Trait

As mentioned above, research focusing on the link between manager trait and risk-taking propensity assumes risk attitude to be a stable characteristic of an individual. The two constructs that received the most attention are hubris (i.e., overconfidence) and narcissism. Manager hubris is related with risk-seeking behavior but only when managerial discretion is high (Li and Tang 2010). Narcissistic CEOs’ risk attitude, on the other hand, is contingent on the social feedback that they receive. Narcissistic CEOs become more risk seeking after receiving social praise (Chatterjee and Hambrick 2007) and more risk seeking in the domain in which there is higher audience engagement (Gerstner et al. 2013).

Agency and Incentives

Another stream of research looks at the effect of agency and incentives on managerial risk taking. These researchers assume that manager’s risk attitude is context dependent,and, more importantly, assume that managers serve their interest by varying their risk propensity in making strategic decisions. As CEOs wealth is more closely linked to stock volatility, CEOs become more risk-taking (Coles et al. 2006). Similarly, CEOs that receive stock options are more risk-taking than those that do not (Deutsch et al. 2011; Sanders and Hambrick 2007). Fixed incentive scheme, on the other hand, decreases risk-taking (Wright et al. 2007). More specifically, CEO restricted stock value relative to her reference point is linked to R&D intensity; negative deviation increases R&D, while positive deviation decreases it (Lim 2015). Relatedly, the risk preference of CEOs that are given stock options depend on the prior performance of the firm (Lim and McCann 2014). Wright et al. (1996) found that institutional ownership has a positive relationship with corporate risk-taking.

Conclusion

Managerial risk attitudes and decision-making under risk deviate from what is assumed in the expected utility theory. Managers consider downside risk and are particularly sensitive to the possibility of incurring a large magnitude losses. They pay less attention to probability estimates but, instead, pay more attention to the magnitude of possible outcomes. Furthermore, managers believe they can control the degree of risk involved in their decisions. The variable risk preference model and the project selection model depict some of these tendencies of managerial decision-making under risk. Other streams of strategy research report that CEO traits, agency, and incentive structure affect manager’s risk-taking propensity.