November 9, 2022
Great new paper by my friend and colleague, Louis Rouanet.
Louis has been doing yeoman’s work exploring the economic history and institutions of France. This paper brings his analysis to the modern era, as he shows the devastating impact of Francois Mitterand’s presidency.
Here’s the payoff from the abstract: By 1996 France would have been around 26% richer than if Mitterrand had not been elected.
And here’s the rest of the abstract:
“The attempt to implement democratic socialism in France following Francois Mitterrand’s election in 1981 is directly responsible for France’s economic underperformance since. This paper uses the synthetic control method to assess the causal effects of Mitterrand’s election. By 1996 France would have been around 26% richer than if Mitterrand had not been elected. Investment and the labor supply fell substantially while government consumption increased relative to our synthetic counterfactual. These effects are persistent. On the other hand, there is little evidence that Mitterrand’s election caused an increase in the average tax rate or inflation. We speculate that the persistence of the negative effects of democratic socialist policies is due to the industrial organization of rent-seeking in France. Rent-seeking in France is centralized and relatively anti competitive. This implies that the interests of the administrative State take precedence over that of private actors, that incumbents rent-seekers face less competition, and that bad policies tend to persist for longer.”
A few diagrams tell the story. Here’s GDP vs. the synthetic counterfactual:
Examining hours worked is particularly creative:
Investment:
In a world where a premium is placed on causal inference at the expense of anything else, I’m especially appreciative of how theoretically-grounded this paper is. According to Rouanet, Mitterand’s policies were particularly sticky due to the broader institutional context within which they were enacted. From the paper:
“As long as the political system does not generate large rent-seeking costs, competition between interest groups will have the tendency to weed out the most inefficient policies (Becker, 1985; Albrecht et al., 2022). On the other hand, the semi-presidential system established with the fifth republic in 1958, combined with the French tradition of government centralization created barriers to entry for interest groups in political markets. These institutional changes had some benefits as they, in theory, reduce rent-seeking. It also had costs as it hampered the tendency of competition among interest groups to weed out socially harmful policies.”
Read his conclusion for an insightful discussion of the interaction effects of policy and institutions (which are not the same thing).
Of course, synthetically exploring the overall institutional environment, as glacially as it moves, is likely not feasible. Exploring the interaction of policy with environment, however, is something I’d love to see more papers tackle. Doing so would bring theory back into the empirical discussion and reverse part of the trend the so-called credibility revolution set in motion.