Identity, Gender, and Subjective Well-Being

Abstract

Using the self-reported level of happiness as a measure of subjective well-being, this study examines the relationship between gender identity and subjective well-being with data from Taiwan. The findings suggest that an individual's perceptions about the ideals of women's gender roles in the labor market, the family, and politics are strongly related to his or her assigned social category, the prescriptions and characteristics associated with the social category, and the actions taken to match the ideals of gender identity. Consistent with Akerlof and Kranton's (2000) identity model, it is also found that an individual's gains or losses in gender identity lead to increases or decreases in the level of happiness.

Keywords:

• identity
• gender
• well-being
• happiness

Notes

1 An excellent survey on economic analysis and happiness can be seen in Frey and Stutzer (2002).

2 Davis (2006, 2007) argues that the identity defined in Akerlof and Kranton (2000) is not reflexive and is determined in terms of third-party observable relationships between the individual and social characteristics. The concept of identity, or the sense of self employed in Akerlof and Kranton, is drawn on a generalized account of psychodynamic personality theory to treat the individual as an identity separate from all its different social identity. By doing so, the concept of identity can be simply defined as an argument in the standard utility-maximizing framework. Thus, Davis (2006, 2007) suggests that the identity defined in Akerlof and Kranton's model might better be labeled a social image of the self, which the individual adopts rather than a self-image. Other recent studies, using the social-psychology sociological identity approach, on issues of household financial organization and women's demand for entrepreneurial capital, include Sonnenberg (2008) and Fletschner and Carter (2008).

3 Before constructing the composite index Identity, the correlations are calculated between the 23 variables from the responses of the questions representing the elements related to women and work, gender roles in home and family life, and women in politics. The correlations between these 23 variables show that the work and family and political elements are not highly correlated. Thus, one variable, Identity, is constructed by summing the responses to these questions (as shown in Table 1). Using a principal-components analysis to construct the identity variable and undertaking the estimation on the relationship between gender identity and happiness also show similar results. However, principal-components analysis is more suitable for constructing indices from the variables of factors that are highly correlated. In this study, using the composite index, Identity, appears to have robust results from the estimations.

4 See Greene (2000) for details of an ordered probit model.