This research project aims at evaluating the effects of covid-19 on economic inequality, with particular reference to European households before the lockdown, during the containment period and after the post-lockdown rebound. Attention is paid to the evaluation of the effectiveness of recovery funds and how successful are public policies in reducing poverty and inequality among European households. The term evaluation refers here to the normative assessment of public policy activities, from a ex-post viewpoint.
In order to pursue this goal longitudinal data are needed, as well as new tools to evaluate the effects of a public policy on the entire distribution rather than the effects on average incomes.
Income inequality estimation, usually measured through Lorenz curve and Gini concentration index, requires households microdata. The evaluation of the success of a public policy in reducing inequality requires, in addition, longitudinal data since same households should be followed over time. Longitudinal data allow to estimate household income distribution before and after the public policy (i.e. recovery funds).
The reference source for this evaluation exercise is the European Union (EU) Survey on Income and Living Conditions (SILC), in its longitudinal component. Starting from the pre-pandemic wave (2019), continuing analyzing the 2020 households data and finally looking at the successive waves (post-pandemic period), the project evaluates the effects of the policy. Data are confidential, but members of the research team are part of an approved project for using these data (2019 - 2024).
In terms of new statistical method, the project extends the traditional Average Treatment Effect approach. The novelty in the methodology lies in the possibility of comparing two entire distributions instead of two average values. Therefore, Gini concentration indx and Lorenz curve for treated (after the policy) and untreated (before the policy) households can be studied.
This research project aims at producing both theoretical advances and empirical analyses to evaluate the effects of the recovery funds on the European families after the COVID-19 pandemic.
In this respect, the present research is expected to produce highly innovative results from a methodological point of view, and also to increase the level of knowledge due to the impact of the public policy on the degree of inequality in the EU countries.
1. The project aims at developing innovative statistical methodologies to evaluate the effect of public policies, the recovery funds, (¿treatment¿) in terms of poverty and income inequality, when data come from complex sample surveys, possibly with different inclusion probabilities. The analysis of this kind of data requires to combine two different weights, namely:
a. design weights, necessary to account for the sampling design;
b. weights based on propensity scores, necessary to account for the unbalancing effect of relevant covariates in determining the probability of receiving treatment.
In particular, on the basis of weights a, b, new estimates of Gini concentration coefficient, as well as of the Lorenz curve, will be developed. Such estimates allow to overcome the most relevant consistency problems of ¿traditional¿ approaches not based on reweighting sample data. In fact, unweighting sample data could lead to estimators that are severely biased for any sample size and, as a consequence, inconsistent.
2. In many cases, the evaluation of the effect of a treatment is not a mere problem of estimation, but instead a problem of hypothesis testing. An important example is to test whether the Lorenz curve for treated individuals is above (i.e. closer to perfect equality) w.r.t. the Lorenz curve for untreated individuals. This requires the study of the (at least approximated) distribution of the adopted test-statistics. In this direction, the project aims at developing new resampling methods, based on pseudo-populations. Such resampling methodologies, in case of variable probability sampling designs, have been studied in Conti et al. (2019, 2020). In this project, these methodologies will be extended in order to include the two weighting systems a, b.
3. Empirical analysis:
the methodological advances we intend to develop in this project allow to evaluate the effects of the recovery funds as stated by the European Commission in May 2020 (The European Commission President Ursula von der Leyen will attend the European Parliament's extraordinary plenary on 27 May, where she will present the Commission's reconstruction plan) in reducing the degree of income inequality in the EU raised by the pandemic.
Income inequality in the EU counties can be estimated using EU-SILC data and the traditional measures like Gini index and/or the Lorenz curve. The statistical and econometric literature so far proposed methodology for evaluating the effect of a public policy only on average value.
This project intends to overcome this limit expanding the possibility of estimate the effect of a policy intervention on the traditional measures of inequality.
Once the longitudinal weighting scheme in EU-SILC is validated, the new framework will incorporate both the designs weights and the treatments weights.
Causal inference is of primary importance when it comes to choosing, designing and implementing public policies. The effects on inequality of the recovery funds cannot be evaluated using the traditional methods based on ATE.