Research

We evaluate the impact of the rent-control regulation implemented by the City of Paris in July 2019 on the Parisian rental market. Using 559,300 observations from January 2018 to June 2023, we apply a difference-in-differences model in which the control units are located in eight major French cities with particularly tight but unregulated rental markets. We find that the policy decreased rents by 3.7% to 4.2% in Paris on average, with stronger effects for small apartments. If every dwelling complied with the regulation, rents would have decreased by 8.2% to 8.7%. Extending the analysis to five additional regulated cities confirms the effectiveness of the policy, and we find no evidence of a decline in rental supply attributable to rent control.

This paper examines the relationship between local labour-market size and firms’ returns to training. Labour pooling may raise returns through matching and learning, while the risk of trained workers being poached may reduce them in larger markets. Estimates based on a large sample of French industrial firms show that returns to training increase with labour-market size, suggesting that labour-pooling effects dominate poaching effects. Average returns lie between 6.7% and 7.7%.

Synthetic difference-in-differences provides an efficient way to estimate causal effects under a latent-factor model, but it normally relies on panel data. This paper adapts the method to repeated cross-sectional data with treatment at the group level. In addition to unit and time weights computed from aggregated data, it introduces weights that account for the varying number of observations in each cross-section. Simulations show that these additional weights improve estimator performance.

Mining deposits are not randomly distributed across space, yet many studies of mining and conflict compare active mining regions with all other areas. We instead use areas with discovered but never-active mines as a control group. Georeferenced evidence shows that these areas are more comparable to places with active mines and that discovery itself does not affect conflict. With this control group, the estimated effect of mining activity on conflict is close to zero and insignificant; estimated effects of mineral-price changes are also 20% to 40% smaller.

This paper studies how agglomeration economies affect firms’ training provision. Firms may train less in dense labour markets because trained workers can be poached, or because matching and learning reduce the need for firm-provided training. After controlling for poaching through firm turnover and local-labour-market turnover, employment density still has a negative and significant effect on training. This points to matching and learning effects rather than poaching as the more likely mechanism.

This paper studies how firms allocate training among workers, focusing on access by educational attainment and location. Firms train fewer workers in denser areas but train those workers for longer, so the two effects broadly offset one another in the total amount of training. Agglomeration therefore changes the allocation of firm-provided training without increasing inequalities in training by years of schooling.