Unemployment insurance and consumption in a context of high informality

A large share of the labour force in developing countries works informally. The lack of a verifiable employment registration not only limits the ability of governments to levy payroll or income taxes, it also constraints targeting of social insurance schemes such as unemployment insurance (UI). Along the development path, the share of formal employment systematically increases while the formal employment/layoff risk does not necessarily decrease and workers often end up switching between formal and informal employment over their careers. The question as to whether any UI system should be implemented to mitigate the welfare consequences of formal employment shocks and how it should be designed thus naturally arise along the development process.

In developing country contexts, the formal-informal margin of employment has been central to the policy debate on the design of social insurance schemes. In the case of UI, upon layoff, workers may find informal jobs that are not observed by the government. This additional margin of behavioural response could both increase the moral hazard problem in job searches and reduce the insurance value of UI benefits, which could help explain why UI is not as prevalent in developing countries. The existing empirical literature has by and large focused on the efficiency costs of UI under high informality. Comparatively, there is very limited evidence on the welfare consequences of transitions in and out of formal employment to inform the academic and policy debates.

In this project, we study how different measures of workers' consumption – an important component of their welfare – changes following a formal employment layoff in the state of Sao Paulo, Brazil. The setting of our project presents a unique opportunity to study the insurance value of UI in a context of high informality.

First, it is a labour market where UI systems exist alongside with a large informal sector (over 35% of private sector employment is informal in Sao Paulo). Second, one of the main reason why the existing evidence is limited for developing countries is that combining panel data on consumption and formal employment status is challenging. We will combine administrative data covering the universe of formal employment spells in Sao Paulo and administrative consumption data from receipts attached to individual identification number generated by a government's rewards program against Value-Added Tax (VAT) evasion.