Abstract
We exploit administrative data describing monthly income trajectories of a sample of Italian Alternative Work Arrangements (the labor vouchers) users between 2008 and 2015 to address whether they replace standard labor income sources and welfare benefits. We estimate cross-income elasticities relying on a fixed effects specification and on a novel estimator controlling for individuals' sample selection in the labor market depending on observable and unobservable determinants. Estimates indicate that vouchers increase overall labor income and substitute standard contracts', but, when selection is taken into account, the positive effect on total labor income reduces while the substitution effects becomes larger. Vouchers effects are heterogeneous depending on sex, age, nationality, and on mafia widespread across provinces. We then set up an event study exploiting a threshold limiting workers' yearly voucher earnings in force between 2012 and 2014. We find that when the threshold is reached, voucher income drops and so does overall income, with an increasing effect over time. No effects are found on welfare transfers and labor income. AWAs effect in the Italian labor market thus depends on whether sample selection bias is addressed and on which types of workers use them and where.