Unemployment Insurance in a Large Informal Economy

Liepmann and Pignatti (2024) estimates the MVPF of unemployment benefits in a country with a large informal economy. The paper studies Mauritius, a country with a population of 1.3 million people where 56.2 percent of employment is informal.

In Mauritius, individuals that lose a job for any reason other than voluntary resignation that has been held for at least 6 months without interruption are eligible to claim unemployment benefits (UBs) until employment is resumed (up to 1 year). The benefit amount is determined based on the number of months they are claimed and the beneficiaries’s former wage: beneficiaries are eligible for 90 percent of their former wage during months 1-3 post-separation, 60% during months 4-6, and 30% during months 7-12. When a beneficiary returns to the workforce in a formal position, contributions from their new position are observed by the Ministry of Social Security and UBs are automatically turned off. If a beneficiary finds a new informal position, they are expected to report their employment status to the labor office which would then manually de-register the beneficiary from UB receipt. However, there are no sanctions for failing to report a new job; individuals can claim UBs while working informally without detection or repercussions. The average length of UB receipt is 10.3 months and about 70 percent of participants stay in the program for its entire duration.

Liepmann and Pignatti (2024) use a difference-in-differences analysis to estimate the implications of losing a formal job. The paper finds that three years after the loss of a formal job, displaced workers are 43.7 percent less likely to hold a formal job and 38.2 percent more likely to be working informally. These displaced workers face an average reduction in monthly earnings of 50 percent for those that are enter informal employment and about 20 percent for those returning to formal employment. This difference holds for hourly rates and persists overtime, indicating that those re-entering the labor force via informal employment experience a significant earnings penalty. This fall in earnings is mirrored by a drop in consumption: consumption falls between 20 to 30 percent for all households in the  following a job loss.

The paper then conducts a regression-kink analysis to estimate the effect of UB generosity on benefit duration and time until formal re-employment. The analysis exploits the drop in benefit entitlements between months 1-3 to 4-7. The paper reports benefits below 90 percent of a beneficiary’s former wage causes them to exit the program faster and receive benefits for a shorter amount of time. Relatedly, the findings indicate benefit receipt increases by 0.324 months in response to a 1,000 Rupee (61 USD) increase in UB entitlement with a corresponding elasticity of 0.474. The elasticity of the duration before formal re-employment with respect to benefit levels is estimated to be 0.762 when formal re-employment is capped at 2 years after job loss.

MVPF SUMMARY

_________

The paper uses the methodology from Hendren and Sprung-Keyser (2020) to translate these estimates into an MVPF of unemployment benefits in a country with a large informal sector.

There is a large literature studying optimal UI policies, often focused on estimating the “Baily-Chetty” condition (Baily, 1978; Chetty, 2006). The Baily-Chetty condition asks whether individuals are willing to pay the net cost of additional unemployment insurance. The WTP for $1 of additional UI benefits is often measured as

so that individuals are willing to pay more than $1 to the extent to which they are risk averse, measured by the coefficient of relative risk aversion, gamma, and the extent to which they experience a consumption drop upon unemployment. Liepmann and Pignatti (2024) use \gamma = 2 and estimate that the change in consumption \frac{\Delta c}{c} = 0.282, yielding a WTP of 1+(2 \cdot 0.282) = 1.564.

The cost of $1 of UI can be written as 1+FE, where the fiscal externality from providing $1 of additional unemployment insurance is measured from the impact of additional UI on unemployment duration.

If individuals were required to pay the cost of their additional unemployment insurance benefits, then additional UI would increase (decreases) welfare if the consumption smoothing benefit is greater (less) than FE. But in practice, beneficiaries of additional UI benefits are not necessarily the ones who pay for the cost of those benefits. Therefore, Hendren and Sprung-Keyser (2020) use these same components to estimate the MVPF of UI:

The efficiency costs of the program are determined according to Equation (1) of Liepmann and Pignatti (2024). The equation is as follows:

where \eta_{B,b} is the elasticity with respect to length of UB receipt (estimated to be 0.474), \eta_{D,b} is the time until formal re-employment (estimated to be 0.762), \frac{D}{B} is the ratio of expected time outside of formal employment to expected duration of benefit receipt (estimated to be 0.664), and \frac{\tau}{b} is the ratio of tax rate and benefit generosity (estimated to be 0.057). Together, this yields a net efficiency cost of 1.502 to the government.

The estimated MVPF of unemployment benefits in a country with a large informal economy is therefore 1.564 / 1.502=1.01.

While the Baily-Chetty condition generally thinks of UI as solely a social insurance policy; the MVPF places UI in the broader context of policies with distributional incidence. This means that additional spending on UI benefits may be desirable not solely because of its correction of market failures, but also because it could be a more efficient method of redistribution.