Why are the Unemployed So Unhappy? Evidence from Panel Data

The growing concern about the extent of joblessness in advanced Western economies is fuelled by the perception that the social costs of unemployment substantially exceed the costs of an economy operating below its potential. Rather, it is suspected that unemployment imposes an additional burden on the individual, a burden that might be referred to as the non-pecuniary cost of unemployment. Those costs arise primarily since employment is not only a source of income but also a provider of social relationships, identity in society and individual self-esteem. Darity and Goldsmith (1996) provide a summary of the psychological literature on the link between loss of employment and reduced wellbeing. Substantial efforts have been made in the past to quantify these nonpecuniary costs of unemployment. (See Junankar 1987; Björklund and Eriksson 1995 and Darity and Goldsmith 1996 for surveys of previous empirical studies.) To begin with, one can think of costs directly in terms of decreased psychological wellbeing. Beyond that, decreased wellbeing may express itself through adverse individual outcomes such as increased mortality, suicide risk and crime rates, or decreased marital stability. These possibilities have been explored by previous research. The general finding is that unemployment is associated with substantial negative non-pecuniary effects (see e.g. Jensen and Smith 1990; Junankar 1991). The case seems particularly strong for the direct negative association between unemployment and psychological wellbeing. For instance, Clark and Oswald (1994), using the first wave of the British Household Panel Survey, report estimates from ordered probit models in which a mental distress score is regressed on a set of individual characteristics, unemployment being one of them. They find that the effect of unemployment is both statistically significant and large: being unemployed increases mental distress by more than does suffering impaired health. Other researchers have used different measures of psychological wellbeing and yet obtained the same basic result, a large negative effect of unemployment on well being. Björklund (1985) and Korpi (1997) construct wellbeing indicators from symptoms of sleeplessness, stomach pain, depression and the like, while Goldsmith et al. (1995, 1996) measure