Signaling

Signaling refers to market actors' use of visible attributes that convey information to or change the beliefs of other actors in the market. Because signals are under the control of the signaler, they are, at least partially, designed to communicate. Hence, signaling theory essentially captures an economic view of reputation in the market. Signals are observable proxies for unobservable attributes of individuals or organizations. For example, an individual's conspicuous consumption may be a signal of wealth and social status. Organizations may invest in reputation building to signal the greater quality of their products and services. Signals are necessary in uncertain economic environments where sellers know more than buyers (i.e., where information asymmetry exists). In this context, we must distinguish between signals and indices (such as race or gender), which generally are considered unalterable (especially in the short run). A good example of a signal is education, which may, in an uncertain world, communicate to an organization which applicants are the most productive and, thus, who will contribute the most value to an organization.

Michael Spence, whose pioneering work on signaling equilibria won him the Nobel Prize in economics in 2001, used this example of education as a signal in job markets in his path-breaking book Market Signaling. In the context of any labor market, the signaling equilibrium is defined as a set of conditional probabilistic beliefs for the employer that, when translated into offered wages, employee investment responses, and new market data, are confirmed by the new market data relating education levels to productivity. Hence, in signaling equilibrium, the employer's beliefs are self-confirming.

Spence showed that usually there are many signaling equilibria, not just one. The simplest case involves comparing two groups of employees where one group doubles the productivity of the other. If education is assumed to function as a signal of differential productivity, the low-productivity applicants will not invest in education at all, whereas rational high-productivity applicants will invest in that education level where the gap between the wage rate and the education cost (= cost of the signal) is maximized. However, Spence also showed that sometimes everyone would be better off if they presented themselves in an undifferentiated pool of applicants. The extent to which the market is able to find a so-called “pooling equilibrium” that still allows differentiation between the two employee groups will depend on the proportions of low- and high-productivity employees and the government's tax scheme on education. Spence showed that with the right tax scheme, such pooling can be made to fulfill the economic criterion of Pareto efficiency.

The standard explanation of signaling and its associated multiple equilibria works because the cost of that [Page 1927]signal is negatively correlated with the attribute valued in the market. But even when education costs rise with the valued attribute (i.e., signaling costs vary the wrong way with respect to productivity), a signaling equilibrium may be attained. This outcome will be possible when the attainment of education also raises productivity to a large extent. In other words, when education is productivity inducing enough to justify its costs, the absence of a negative cost correlation does not necessarily destroy a signaling equilibrium (as initially assumed by Spence). Rather, a signaling equilibrium can still be identified. However, in this equilibrium, the private return to education falls short of the social return and hence causes underinvestment in education.

...