Heterogeneity in Subjective Wellbeing : An Application to Occupational Allocation in Africa

Falco, Paolo ; Maloney, William F. ; Rijkers, Bob ; Sarrias, Mauricio

World Bank, Washington, DC

关键词: ACCOUNTING;    ATTRITION;    AVERAGE WAGES;    CLERICAL WORKERS;    DEVELOPMENT ECONOMICS;   

DOI  :  10.1596/1813-9450-6244
RP-ID  :  WPS6244

学科分类：社会科学、人文和艺术（综合）

来源: World Bank Open Knowledge Repository

【 摘 要 】

Using an extraordinarily rich paneldataset from Ghana, this paper explores the nature ofself-employment and informality in developing countriesthrough the analysis of self-reported happiness with workand life. Subjective job satisfaction measures allowassessment of the relative desirability of different jobs inways that, conditional wage comparisons cannot. Byexploiting recent advances in mixed (random parameter)ordered probit models, the distribution of subjectivewell-being across sectors of employment is quantified. Thereis little evidence for the overall inferiority of the smallfirm informal sector: there is not a robust averagesatisfaction premium for formal work vs. self-employment orinformal salaried work, and owners of informal firms thatemploy others are on average significantly happier thanworkers in the formal private sector. Moreover, theestimated distribution of parameters predicting satisfactionreveal substantial heterogeneity in subjective well-beingwithin sectors that conventional fixed parameter models,such as standard ordered probit models, cannot detect:Whatever the average satisfaction premium in a sector, alljob categories contain both relatively happy and disgruntledworkers. Specifically, roughly 67, 50, 40 and 59 percentprefer being a small-firm employer, sole proprietor,informal salaried, civic worker respectively, than formalwork. Hence, there is a high degree of overlap in thedistribution of satisfaction across sectors. The results arerobust to the inclusion of fixed effects and alternatemeasures of satisfaction. Job characteristics,self-perceived autonomy and experimentally elicited measuresof attitudes toward risk do not appear to explain thesedistributional patterns.