Working Papers

WORKING PAPERS

• “Untying the knot: Divorce Legalization and the Decline of Marriage”, with Sekyu Choi. Revise and Resubmit at Journal of Development Economics

• “Life Cycle Saving in a High Informality Setting – Evidence from Pakistan”, with Priyanka Kanth, Policy Research Working Papers 10121 

• “How Organized is the Informal Sector? The Role of Business Associations in Microenterprises  in West Africa”, with Kathleen Beegle,, Policy Research Working Paper 11101

•   “The Early Career Dynamics of Informality and Underemployment: Evidence from the Arab Republic of Egypt”, with Samia Ferhat, Policy Research Working Papers 10499

• “Compensation, diversity and inclusion at the World Bank Group” with Jishnu Das and Florian Tordoir, PRWP 8058

• Joubert et al. 2026. “Inclusive Pensions for an Aging World: Evidence and Strategies for Engaging Informal Workers” Working Paper. Washington, D.C.: CGAP

WORK IN PROGRESS

• "The determinants of pension benefits take-up, evidence from Chile", with Pablo Troncoso

Means-tested government transfers throughout the world exhibit high rates of nonparticipation, even in large well-established programs in rich countries, such as the United States’ SNAP or EITC programs. Non-take up issues are likely to be even higher in lower capacity settings but remain very poorly documented beyond aggregate take up rates, because data requirements are high. Characterizing non-take up at the individual level requires administrative data to identify program recipients, but also survey data to identify eligible non-recipients, who do not feature in admin records. We obtained administrative data on social pension applications and receipts which are linked to a longitudinal survey. The survey contains rich data on pension eligibility criteria but also on pension program knowledge and self-reported benefit receipt. We show that non-take up is most common among the most vulnerable, contrary to what ordeal theory would predict, implying that non-take up reduces the quality of program targeting. We also establish that using self-reported beneficiary status, as in much of the extant literature, can severely bias the characterization on non-takers, because individuals often misreport the benefits they receive. We now are in the process of examining how pension knowledge affects the probability of non-take up and plan to use a recent expansion of the means-testing threshold from the 60th to the 90th percentile as a natural experiment to evaluate the effect of means-testing on take up. 

• "In-house formalization and labor inspections", with Katharina Maria Fietz

This study considers how the enforcement of labor regulation affects “within-firm” or “post-hiring” formalization, a process where firms initially hire workers informally, possibly to assess their productivity, and later decide whether to offer costlier formal contracts. Rather than focusing solely on the direct impact of inspections on firms that are actually inspected, as in much of the literature, the research examines the more important deterrent effect of inspection risk on all firms. Using short panels collected for Brazil’s largest employment survey (PNADC) in which workers are interviewed in consecutive semesters and are asked about their tenure at their job, we identify formalization transitions that occur without a change in employer. We find that 29% of all formalization transitions occur under the same employer over the period and document how this phenomenon varies over time, region and type of firm. We use unique administrative data on the universe of labor inspections in Brazil since 1998, merged to employer-employee records for inspected and non-inspected firms to document variation in inspection intensity. We find a negative and statistically significant relationship between labor inspection intensity and the rate of within-firm formalization, suggesting that firms facing higher inspection risk are less likely to delay the formalization of their workers. 

•  "All at once: the impact of lumpsum pension payouts on retirement decisions”, with Thiago Scot and Tatiana Flores

This study uses a cohort-based regression discontinuity to examine how pension payout modalities impact labor supply decisions, wage reporting and the timing of retirement. It exploits a unique institutional feature: when the Dominican Republic privatized their pension system in 2003, individuals older than 45 at the time of first participation were granted the right to withdraw their accumulated pension balances as a lumpsum payment instead of annuities. Theoretically, lumpsum retirement payouts could incentivize individuals to retire earlier if they are biased towards immediate consumption at the expense of long-term planning. On the other hand, the prospect of lumpsum retirement benefits could increase incentives to participate, for example if individuals hope to start a small business in old age. Consistent with the first hypothesis, we find that pension claiming jumps by 15pp at the discontinuity, but we find no evidence that the lumpsum payouts option encouraged participation at younger ages. We are currently exploring complementary data sources, including survey and tax data (which is being exploited in a parallel engagement with the ministry of finance) to identify how receiving pension benefits as a lumpsum affects the subsequent consumption, employment and welfare of older individuals.

• Understanding Workers’ Valuation of Formal Employment in Central Asia, with Laurent Bossavie, Nathalie Gonzalez-Prieto, and Yashodhan Ghorpade

Due to lack of information, liquidity constraints or lack of trust in government, workers may not value the benefits associated with formality, particularly health and pension coverage. In that case, payroll contributions would be perceived purely as a tax and labor regulation would create large incentives to working informally. If, instead, workers know and value these benefits, it is likely that those who work informally do it by lack of opportunity rather than by choice. We run an online discrete choice experiment in Uzbekistan a la Maestas et al. (2023) in which individuals chose between hypothetical jobs with varying characteristics, including wage levels and formality. We also elicit the extent to which earnings are underreported. We also collect information on the financial and behavioral determinant of the willingness to contribute to social security.