Thread by @mbwong

My JMP is now posted: “Labor Market Consequences of Domestic Outsourcing: Evidence from Legalization in Brazil” economics.mit.edu/files/22108

Coauthored with my excellent classmate Mayara Felix. #EconTwitter #EconJobMarket 🧵

Our contribution: While prior work uses *firm-level* variation to measure the effect of outsourcing on wages, we use *market-level* variation from Brazil's 1993 outsourcing legalization to assess the aggregate effects of outsourcing on wages, employment, and welfare.

Our focus: Security guards, who experienced the largest rise in outsourcing following legalization

Data: RAIS matched employer-employee data with industry and occupation codes to identify outsourcing

Part 1: We show that outsourcing legalization led to a wave of occupational layoffs.

These layoffs moved incumbent workers to lower-wage firms and persistently reduced their wages. The loss of firm wage premia explains 40-50% of the wage reductions.

Part 2: We estimate the market-level effects of outsourcing.

We use a triple diff specification that leverages pre-legalization variation in the permissiveness of local courts and compares guards to less affected occupations.

Outsourcing legalization led to significant job reallocation. Employment of older workers fell, while employment of younger workers increased. Mean age in the occupation fell by 2 years.

Outsourcing legalization also raised employment by 8-15% and increased composition-adjusted wage by 2-4%. These effects are robust to controls for time-varying effects of exposure to concurrent trade liberalization.

Part 3: We use a parametric economic model to interpret these results and perform welfare analysis. In the model, efficiency gains raise wages and employment, while a reduction in worker bargaining power reduces wages.

Combining model + estimates, we infer that outsourcing legalization generated large efficiency gains, but had an insignificant effect on worker bargaining power in the occupation.

Welfare analysis of outsourcing legalization:

Cost = PDV of loss to laid-off incumbent workers = 8.4% of total wagebill

Benefit = Per annum increase in total surplus = 1.6-6.7% of total wagebill

Breakeven = 1-5 years

Conclusion: Banning outsourcing protects incumbent workers from layoffs. But it prevents efficiency gains from labor flexibility and specialization, so it is bad for overall welfare as well as potential entrant workers.

Footnote: Instead of banning outsourcing, policymakers can strengthen legal protections for outsourced workers to mitigate abusive practices like wage theft and unsafe conditions.

P.S. My excellent coauthor and classmate Mayara Felix is also on the job market. Hire her too!

Her website here: economics.mit.edu/grad/mfelix

Many thanks to @davidautor @johnvanreenen @townsend_rm & Bob Gibbons for guidance!! Thanks esp to Breno, Eduardo, and Magda for help with fieldwork! Thanks to @TobiasSalz @DrDaronAcemoglu @simon_jaeger @annastansbury @JADHazell and many others (see paper) for helpful advice!