Measures for gender rebalancing in local elective assemblies. An impact assessment on Italian reforms

The general objective of the paper is to present the results of an evaluation of the effects produced by the regulatory measures and institutional practices introduced in recent years for the purposes of gender rebalancing in the elective offices. Many countries have introduced rules to accelerate the rebalancing process. The general feeling is that their effectiveness may vary a lot, because they need to be well designed and effectively implemented to achieve good results. A statistical recognition of gender shares among candidates and elected representatives for all institutions in Italy, in the period 2000-2016shows a growing tendency. Reforms intervene on this slow trend giving further propulsion to the rebalancing process. So the counterfactual approach is fundamental to verify if there has been a significant impact, detangling it from the deadweight effect. The impact evaluation is limited to the case of municipalities. In the other cases the lack of a credible counterfactual forces us to limit to statistical analysis of observed change. The paper presents and compares the results of three evaluation approaches: Panel models, natural experiment and RDD. Panel regressions including individual fixed effects can be applied to isolate the elements that influence the share of elected representatives , so addressing the problem of endogeneity. The approach makes it possible to verify the existence of trends in the presence of women in politics, in place independently of the legislative reform introduced in 2012. At the same time it allows to test the significance of the difference between the shares of women and men before and after the reform, providing a measure of the net impact of the policy. The second evaluation design exploits the condition of natural experiment that arose in the years immediately after the reform. This condition derives from the fact that, even if the law with the new gender measures came into force for all the municipalities subject to the legislation at the same time, the latter have electoral different dates for the election of the municipal council. Finally we apply a regression-discontinuity-design, comparing units that are immediately below or immediately above a policy application threshold represented by the size of the municipality in terms of inhabitants. Parties in municipalities smaller than 5.000 people are just asked to present lists with candidates of both genders, while municipalities over the threshold have to apply various mechanisms. The approaches are compared and discussed.