Abstract An attempt is made in this article to redefine underemployment and unemployment without making reference to an excess supply of labor or any causal mechanism of unemployment. Instead, underemployment and unemployment are defined in terms of equity which draws upon the individual's preferences. A specific proposal is that underemployment be defined by the presence of contribution inequity relative to at least half the persons employed in a field that the underemployed person might prefer to move into. Empirically, most recent survey data on preferences for contingent and other nontraditional employment are used to illustrate the application of the concept. The major finding is that nearly 10 million Americans in the nontraditional workforce are underemployed.
Keywords: underemployment, unemployment, contingent employment, consumption inequity, contribution inequity, superfairness
Since the Great Depression of the 1930s, unemployment has been recognized as a problem for economic policy. Nevertheless, defining unemployment remains a problem. A person is "involuntarily unemployed" if she or he has no job and would accept a suitable job. In this paper we derive a definition of "underemployment" and "unemployment" from considerations of equity. The concept of equity we will draw on, which is based on the preferences of the persons concerned, comes from neoclassical economics. This seems innocent enough: if a person has no job, but (in relevant circumstances) prefers to have no job, then she is not "involuntarily" employed. However, concepts related to Baumol's "superfairness" allow us to go considerably beyond that--without accepting the neoclassical idea that preferences are the only factors that matter. 
Changes in the workforce i.e. the rise of contingent work have also undermined the traditional concepts of unemployment. In the following sections we review some of those changes and then introduce "fairness" and "superfairness" as they appear in recent economic theory. We then define under employment in terms of equity, and proceed from the more general (and conceptually troubling) underemployment to the less general category of unemployment. We then draw on studies of contingent employment, in which we have data on preferences, to illustrate the application of the concept in a particularly crucial and important case.
DEFINITIONAL ISSUES: UNDEREMPLOYMENT, NONTRADITIONAL AND CONTINGENT EMPLOYMENT
The phrase "contingent employment arrangement" was coined by Audrey Freeman at a 1985 conference on employment security (Polivka and Nardone 1989). According to Freeman (1988), an important characteristic of "contingency" is a lack of attachment between the worker and employer, for these workers are hired only when there is an immediate need for their service. Some researchers thus define the contingent workforce as those workers whose ties with their companies are much looser than those held by conventional nine-to-five salaried employees (Lester 1996). For example, Christensen and Murphree (1988) characterized the contingent work (CW) as having four properties: (1) time: CW is no longer a nine-to-five regular job; (2) location of work: contingent workers are more likely to work at home; (3) permanency: there is no guarantee of future employment; and (4) social contract: CW lacks a traditional employer--employee contract promising reciprocal rights, protection and obligations between them.
However, the definition provided by Freeman has been attacked for inconsistency and possible misclassification because it allows jobs offering a high degree of employment stability, such as a self-employed professional, to be classified as contingent solely because they are not full-time and permanent positions. An alternative definition of contingent work was provided by Polivka and Nardone (1989) which is based on the terms of employment, consisting of such factors as job security, variability in hours of work, and access to benefits(Lester 1996: 71). …