The authors estimate the causal effects of the reform on employment trajectories using a difference-in-differences method, based on administrative data tracking patterns of employment, unemployment and non-employment (the MiDAS panel from Dares).
Their level of observation is the end of a contract. Each end of contract is associated with a series of indicators relating to the individual’s trajectory following that end of contract: recourse to unemployment benefit, any amount of benefits received, the number of days worked by the individual during the 6 and 12 months following the end of their contract, the type of jobs taken up, the number of contracts taken up, the number of days per contract, etc. The authors compare changes in these indicators according to the contract end date, across different groups defined by the length of the gap between contracts during the period used to calculate the SJR.
To take one example, the number of days worked is compared for contract end dates before and after 1 October 2021. It is concluded that the reform has a positive effect on employment if the increase is greater (or if the decrease is smaller) for individuals with long periods between contracts, compared with those with short periods. This allows us to account for general labour market trends, both before and after the reform, and for systematic differences between groups with different lengths of time between contracts, whilst using differences in exposure to the reform to measure its effects.
This method nevertheless assumes that the effects of seasonality on the labour market are similar across the different groups with varying lengths of time between contracts. Our results suggest that this assumption is plausible, provided that we compare indicators relating to cumulative employment over a period of at least six months following the end of the contract across these groups.