Reputational Dynamics

Reputational dynamics expresses the view that reputation should be conceptualized as an intrinsically dynamic phenomenon. This intuition is reflected in informal accounts of corporate reputation that indicate changing value such as reputational capital or a moral bank account. These metaphors (“reputation is like a financial asset”) capture the fact that the value of a company’s reputation can wax and wane over time in response to company actions or external events. This entry discusses reputational dynamics from the perspective of modern game theory. It discusses the commonly used game-theoretic concept of reputation as “hidden information,” outlines some of the limitations of this view, and offers an alternative perspective.

Overview Using Modern Game Theory

In modern game theory, reputation is modeled as a belief structure in a dynamic game of incomplete information. Customers and the general public have only partial information about an important feature of a company. That feature can be “quality” or it can be related to responsible business practices, for example, environmental stewardship or safe labor practices. One particular germane, but mathematically complex, model can be summarized as follows: A firm can engage in two forms of business practices: (1) responsible or (2) irresponsible. For example, a company may engage in safe or unsafe labor practices. The company’s labor practices are not directly observable by consumers. Rather, at any period, the consumers form beliefs about the company’s labor practices represented by a probability distribution pt that the firm’s practices are safe. Consumer demand is an increasing function of this belief that we can refer to as the company’s “reputation.” That is, consumers are assumed to care about whether the company engages in safe labor practices and adjust their purchasing behavior accordingly. Such product or firm attributes are often referred to as “credence goods”—that is, attributes of a product or company that consumers care about but that cannot be directly verified by consumers. Hence, consumers must base their purchase decisions on beliefs about the company rather than on direct experience.

A company’s reputation can change stochastically over time, as a function of the firm’s actions, for example, by investing in costly safety features or an exogenous shock, for example, an accident. The shock is assumed to be necessary (but not sufficient) for the level of safety to change. This creates a tendency for the level of safety at time t to persist until time t + 1. At the end of the period, a signal may or may not be realized, such as a report about an injury to workers. The probability that this signal is observed by consumers depends on the true underlying level of safety in the plant, and consumers update their beliefs in a rational fashion.

Suppose now that an unsafe factory is more likely to generate a reported accident than a safe factory. If an accident is observed, consumers infer that the factory is more likely to be unsafe; if no accident is observed, consumers infer the opposite. Because firm actions affect the probability of safe and unsafe business practices, which in turn affect the likelihood of the signal, consumers need to make inferences about the equilibrium investment level of the firm, as a function of its current level of safety. This updating process, in turn, affects the firm’s investment incentives. For example, if an unsafe factory is certain to have a reported accident, while a safe factory is guaranteed not to have an accident, the firm’s type is fully revealed in one period: If the factory is safe, no accident is observed. In turn, consumers infer from the absence of an accident that the firm’s factory is safe. By contrast, if the factory is unsafe, a reported accident will occur, and consumers infer from the report that the firm’s factory is unsafe. In this example, reputation is fully informative and separates the firms’ types. This generates a strong incentive for a firm whose factory is safe to invest so that it continues to be safe in the next period, while there will be a strong incentive for a firm whose factory is unsafe to invest so that its factory becomes safe.

Limitations

The practical usefulness of modeling reputation as hidden information can be limited. First, in game-theoretic models, a company’s reputation can “collapse” in as little as one period, once consumers can correctly infer a company’s type as irresponsible. Empirical research suggests that reputational change is more gradual and can be repaired. Second, the models depend on the public’s ability to rationally update their beliefs, yet extensive empirical research on branding suggests that actual brand perception is not well accounted for by rational belief formation but is subject to a variety of biases, emotional responses, and so forth. Third, the actual processes that shape consumer perception are not easily summarized as a probabilistic signal but involve firms and other constituencies, for example, the news media, consumer advocacy groups, and organized interests that compete in the arena of public opinion. Understanding how these processes work is often at the core of reputation management.

Alternative Perspective

An alternative perspective considers reputation as a stochastic process where rt is the current reputation of a firm. A good reputation is valuable to the firm, here modeled through a reduced form profit function π(r), where π(r) is strictly increasing and strictly concave in r. Thus, the model captures the belief that a good reputation is an asset to the firm, perhaps because it translates to higher demand for the firm’s products by concerned consumers or gives the firm an edge in recruiting executive talent and so on. Similar to other assets, corporate reputation is assumed to exhibit decreasing marginal returns.

Corporate reputations can increase or decrease. Specifically,

rt=rt−1+ft−at,

where ft ∈ {0,1} is a positive shock to the firm’s reputation and at ∈ {0,1} is a negative shock.

A firm can invest in responsible business practices xt, for example, better labor standards. Such investments are valued by the activists and consumers. Activists, on the other hand, can start a corporate campaign yt that may damage a company’s reputation. More investment in responsible business practices x increases the probability of a positive reputational shock in that period, while greater campaign intensity y in that period increases the probability of a negative shock. If both, a positive and a negative shock occur in the same period, they offset each other, and the firm’s reputation remains unchanged. This structure yields a stochastic dynamic game between a firm and an activist group, which can be solved for the Markov perfect equilibria.

The model yields some interesting insights. For example, it can explain why activist groups target particularly well-known firms and why the same firms are targeted repeatedly even after they have already invested in responsible business practices. One reason Starbucks is thought to be targeted so much is that the decreasing marginal returns of investing in responsible business practices creates an incentive for companies to “coast on their reputation,” for example, invest less and less in responsible business practices as their reputation increases. To preclude such coasting, activists will launch corporate campaigns to destroy some of the firm’s reputational equity to keep the firm “hungry.” Activists, however, cannot be too extreme or too effective in causing reputational harm. In that case, firms simply give up investing in their reputation.

Conclusion

Game-theoretic models of corporate reputation provide important insights about reputational dynamics, but many questions remain. Modeling corporate reputation as hidden information depends on rational belief formation by customers and the public, an assumption that strikes many observers as unrealistic and ill founded in the psychology of public perception. Recently, alternative models were developed that capture accounts of corporate reputation common in the management and marketing literature. Such accounts capture general accounts of public perception. As a next step, they need to be connected with detailed empirical studies of how the public perceives companies and how such perceptions change in response to actions by companies and their critics. There is a small but growing literature that has started to investigate these issues. For example, in the context of a corporate crisis, companies are usually perceived as villains, sometimes as heroes, and almost never as victims. The reason is that consumers view companies as an unusual entity that can form intentions but not suffer pain. Establishing a deeper connection between the microprocesses of how the public perceives companies and theoretical accounts of reputational dynamics remains an important challenge for future research.