| 12:30 - 13:45
Institute for Fiscal Studies -
7 Ridgmount Street, London WC1E 7AE
Different unemployment insurance schemes exist to support displaced formal workers across countries, and their disbursement policies differ substantially. State-contingent transfers, such unemployment insurance (UI) programs, are schemes that transfer benefits periodically to displaced workers who remain non-employed. In contrast, some schemes transfer a benefit amount in a lump-sum fashion at layoff, such as severance pay programs (SPP) and most unemployment insurance savings accounts (UISA) programs. Although most countries provide workers with some combination of periodic benefits and lump-sum schemes, little is known about how the different disbursement designs affect the insurance value they provide to displaced workers. Moreover, the relative importance of these schemes vary systematically across the development path: lump-sum transfers are more common than UI in developing countries, where the state has limited capacity to monitor workers and a large share of the workforce is informal. This paper contributes to fill this gap using evidence from Sao Paulo, Brazil. We analyse high frequency expenditure data matched to a longitudinal microdata of formal employment for over 400,000 workers. Using multiple sources of individual income variation from unemployment insurance schemes available to workers in Brazil, we document several empirical findings on how different unemployment insurance schemes affect consumption smoothing. Across our empirical analysis we find that consumption is highly sensitive to the timing of disbursement of benefits. First, we find a substantial increase in consumption at layoff, when a lump-sum becomes available, followed by a long-term loss in consumption if workers remain non-formally employed. This increase is larger for workers with higher lump-sum despite similar long-term losses. Second, we find that consumption drops after UI exhaustion, and we zoom in at daily data to find that the higher consumption levels workers experience before UI exhaustion is concentrated around the UI payday, and that the change in expenditure is smooth around exhaustion. These empirical patterns shed light on the relevance of disbursement policies for social insurance, and provide novel evidence on the insurance value of unemployment benefits in a developing country context.