The determinants of occupational choices and income generation in African labour markets: theory and evidence from field-experiments and survey data

Lead Research Organisation: University of Oxford
Department Name: Economics

Abstract

The path to economic development hinges crucially on workers' job opportunities and career choices. For instance, an economy that lacks sufficient wealth to create employment through investment in capital, education and infrastructure may stagnate in a state of perennially low economic growth (Banerjee and Newmann 1993). My research explores the opportunities available to African workers and the mechanisms that govern their career choices with a focus on the rapidly expanding informal economy. Who chooses to be an entrepreneur and why (or why not)? What are the returns to investment in informal enterprises and what is the value of education? More specifically, my goal is to understand how earnings uncertainty and the risks associated with informal jobs affect the well-being of workers and, hence, their career plans. Within policy circles the rapid 'informalisation' of developing economies remains a contentious subject, which my research helps clarify.
The main innovation in my doctoral thesis is the use of experimental data from behavioural field-experiments conducted in Ghana. My experiments consisted of several interactive tasks (games) aimed at measuring respondents' psychological attitudes in economic decision-making. In particular, I have constructed a measure of workers' willingness to take financial risks, and used it to test several hypotheses on real-life choices. Experiments of this kind, conducted in the challenging environment of developing countries, are currently at the research frontier in Economics.

Chapter 1 explores the challenges of precisely quantifying vulnerability to income losses. Using of a 'maximal heterogeneity' methodology I estimate the risk that a worker will lose his/her job on the basis of his/her prior job-history. The main strength of this approach is that it imposes minimal assumptions on the data and it hence produces results that are more general and robust than conventional methods. Identifying who is at high risk of job-losses is crucial for effective targeting of social protection policies.
In chapter 2 I test whether the decision to work in the formal rather than the informal economy reflects workers' preferences for higher income security. Using data from my field experiments, I find that individuals who are less averse to risk tend to prefer the informal sector, where income uncertainty is higher. This result shows that concerns of income instability play a significant role in workers' career choices.
Chapter 3 studies the direct impact of income instability on workers' happiness. Thanks to our new CSAE dataset on self-reported well-being in Ghana, I am able to perform a sophisticated panel analysis of variations in happiness as a function of changes in income uncertainty. My results indicate that uncertainty has a strong impact on well-being and can offset the benefits of income gains that come at the cost of less security. In line with chapter 2, this evidence confirms that psychological effects of job-insecurity are strong and it motivates the design of effective income safety nets in Africa (e.g. unemployment benefits).
Chapter 4 explores the determinants of productivity in Ghanaian informal enterprises. Using data on a sample of micro-entrepreneurs followed over several years, I estimate returns to their capital, education and work-experience. I find that capital and experience have the strongest impact on firm productivity, but the real income gains that result from micro-investment are too modest to constitute a viable engine of growth. The benefits from formal education are also limited, indicating that the skills acquired in school do not fit the needs of the growing informal economy.

During the PDF I will disseminate the results of my research through conferences and seminar presentations and seek their publication in academic journals. Further, I will use new CSAE data (including a new wave of experiments) to implement two extensions to my existing papers.

Publications

10 25 50