Signal Theory

Signal theory (or signaling theory) describes how actors selecting among a choice set can successfully make determinations about the unobservable quality of their selection via observable signals. The theory grew out of labor economics, wherein job applicants signaled their unobservable quality via observable educational credentials, and has since been applied to diverse selection scenarios that run the gamut from anthropology to zoology. This entry explains the fundamentals of signal theory, describes four basic types of signals, and considers these in view of reputation.

Fundamental Principles

The key to understanding signal theory lies in knowing the difference between what is a signal in the vernacular and what constitutes a signal within signal theory. To do so requires that one should be able to clearly elucidate why high-quality actors can and do send a signal whereas low-quality actors do not. This is called the signaling equilibrium, wherein there is a separation between high- and low-quality actors based on their willingness and ability to signal. Thus, for instance, waving goodbye when leaving the room is not a signal in the signal theory sense because anyone can wave goodbye. Saying goodbye in the mother tongue of three individuals who all speak different languages, on the other hand, is consistent with signal theory because only high-quality individuals could do so. Quality refers to the individual’s underlying abilities and resource endowments and, in this example, would likely point to the individual’s learnedness or preparedness. The two key characteristics, therefore, of a valid signal are that (1) it should be observable and (2) low-quality actors would not or could not choose to signal.

Michael Spence originally developed signal theory in his seminal work on labor economics, for which he received the Nobel Prize in Economics. Spence described a hiring scenario wherein an employer would be willing to pay a high-quality worker $200 to complete a project but would pay a low-quality worker only $100. The employer does not know who is of high or low quality until after the person is hired. However, workers may be certified through a $60 training course. High-quality workers would pass the first time, but low-quality workers would probably have to take the training course twice, thus costing them $120 each. The high-quality worker would become certified and earn a net of $140, but the low-quality worker earns more if he or she is not certified ($100) than if he or she is certified ($80). Thus, certification is a valid signal that allows accurate selection of high- versus low-quality workers.

Signal Types

There are a variety of ways in which signals can differentiate between high- and low-quality actors. The most common is when there are differential signaling costs based on actor quality. For example, it is more economically feasible for high-quality firms to attain ISO (International Organization for Standardization) 14001 certification for environmental management standards than it is for low-quality firms because the high-quality firms would not have many changes to make to their existing processes. Similarly, a prestigious degree is less costly for a high-quality actor because a low-quality actor would presumably be unable to endure the rigors of either obtaining the degree or undertaking the activities required for admission. These might be called Spencian signals because they are most akin to Spence’s original formulation of the theory, though this label does not imply that other signal types do not conform to Spence’s ideas.

A second type of signal is costly in a different way. An actor could take on a credible commitment that operates as a pledge, or bond, to suffer negative consequences in the future if he or she does not deliver high quality. For instance, companies might offer warranties or money-back guarantees to consumers as a bonding signal of quality. Another good example would be company insiders who buy stock in their own companies, thus taking on personal risk that insiders at low-quality firms would presumably be unwilling to take. By making a financial commitment, insiders communicate their private knowledge to outsiders so that their willingness to take on personal risk serves as a bond to suffer the same loss outsiders would suffer if the company does not deliver high quality. For bonding signals, the cost of signaling is the same for all actors, but one can distinguish between high- and low-quality actors because the likelihood of incurring penalty costs is different.

A third type of signal shifts costs away from the actor and onto some third party. In many selection scenarios, there may be third parties that have their own information that could help distinguish between high- and low-quality actors. Therefore, third parties may provide an observable signal in the form of an endorsement that attests to the underlying quality of a particular actor in a choice set. Common endorsement signals include prestigious affiliates, discerning intermediaries, and certifying institutions. For instance, having a highly respected director on a board serves as an endorsement signal because such people would be unlikely to sit on the board of what they perceived to be a low-quality firm. In new ventures, notable investors and esteemed underwriters operate as endorsement signals for young companies seeking external financing or an initial public offering.

A fourth type of signal looks more to the future than to the past. Companies can signal their intent or plans for the future. This can be a valid signal within the signal theory framework if high-quality actors would be more likely than low-quality actors to engage in such a signal. For instance, a company might establish a foothold in a competitor’s market as a signal of its intent to retaliate should the competitor become aggressive. This could be a valid signal if only high-quality actors (i.e., those with the capability of retaliation) are willing to establish and maintain such a foothold over time. Stated differently, high-quality actors send the signal because they can bear the future costs of the intent that is signaled, but low-quality actors cannot bear those future costs.

In sum, signals are valid when they are observable, and signaling costs create a separation between high- and low-quality signalers. Spencian signals entail differential signaling costs, bonding signals impose differential penalty costs, endorsement signals are borne by a willing third party, and intent signals introduce possible costs in the future.

Signal Theory and Reputation

Each of the signal types described above could have important implications for reputation. Companies or individuals might use Spencian signals for reputation building. For instance, companies can say a lot of things in their press releases and marketing about their commitment to the environment, but these are not a good signal about the quality of their environmentalism because anyone can say those things. A better signal is Ford’s construction of a 10-acre garden atop the roof of one of its main manufacturing plants. Environmental commitment was the centerpiece of then CEO Bill Ford’s vision, and only a firm that was serious about the cause (a high-quality firm) would undertake the costs of such a project. At the individual level, an enterprising executive seeking to get on a company board might undergo some expensive and time-consuming training, such as that offered by the National Association of Corporate Directors. Doing so serves as a costly signal that only those who are committed to corporate governance would be willing to undertake.

Bonding signals would be useful for proving one’s reputation to outsiders. In these scenarios, insiders have private information about a firm, and they are willing to demonstrate that their information is accurate. For instance, a new venture seeking external capital beyond the founding team’s friends and family for the first time might establish a physical presence in the form of a storefront. Doing so imposes risks on the founding team because there are economic, and potentially social, costs associated with planting their flag in the ground if the company does not deliver high quality to its constituents. Individuals can also use bonding signals to prove their reputation. For example, a person providing professional services might offer a money-back guarantee if customers are not satisfied. Only high-quality service providers would be willing to assume such a risk, knowing that customers would be unlikely to be dissatisfied, so that the penalty costs associated with providing poor-quality service would be a nonfactor.

Endorsement signals allow individuals and firms to borrow the reputation of others. This is particularly useful for organizations and individuals who lack an established reputation of their own. Consider, for example, a firm expanding into a new country for the first time; it suffers from a liability of foreignness, wherein locals are not yet familiar with the company’s reputation. In such cases, the multinational firm might seek out a well-respected partner in the host country who could attest to the quality of the multinational’s products and services. Similarly, at the individual level, a job applicant might seek out a recommendation from a prestigious third party. Such a signal is increasingly valuable when the third party is closer to the one doing the hiring because the third party is putting his or her reputation on the line for this individual, which he or she would be unwilling to do for a low-quality actor.

Last, signals of intent provide a means for actors to validate promises about their future reputation. These are similar to bonding signals insofar as the actor does not necessarily incur the full cost of signaling, but signals of intent are more forward looking. A promise itself is not a useful signal because promises can go unfulfilled. However, costly actions that demonstrate intent can be useful for making determinations about a firm’s or an individual’s expected future reputation. For example, establishing a program to equip and train minorities for top management positions is a useful indicator of a firm’s commitment to future diversity and inclusion among its top management. Low-quality firms (i.e., those that have no intent to incorporate minorities in the top ranks) would be unlikely to institute such a program. Similarly, at the individual level, an enterprising young entrepreneur might join a sustainability support group to demonstrate that he or she intends to have an expected future reputation for his or her firm for caring about the environment.

In conclusion, a journalist once asked Spence, widely recognized as the originator of the theory, how one could receive a Nobel Prize in Economics for simply noticing that sometimes people do not know certain things about an actor they might select that the actor may wish them to know. The beauty of the theory, however, is in its simplicity. Spencian, bonding, endorsement, and intent signals are useful for separating high-quality selections from their low-quality counterparts. As a result, individuals and organizations can use signals to build their reputation, prove their reputation, borrow someone else’s reputation, or make promises about their expected future reputation.