Well-being analytics for policy use : Modelling health and education outcomes in Italy

Francesco Savazzi (University of Evry Val d'Essonne)

 

Abstract:

The present paper presents methodologies to forecast and conduct policy analysis for three well-being indicators with the goal of informing the Italian government’s budget planning process. For each of the three indicators (healthy life expectancy, overweight and obesity, and early school leaving), a model is developed that allows projecting future trends under a status quo scenario and that allows estimating the impact of policy and budget levers on future outcomes. The micro-economic models for being in good health have a moderate explanatory power with an R2 ranging between 0.2 and 0.3. The strongest predictors of good health are by far the prevalence of chronic diseases, followed by low mental health, sport practice and diet. Overall, the combined changes in inputs yield an improvement in the share of people declaring being in good health by 2.7 ppt, from a baseline of 62% among people older than 18. The micro-economic model for being in excess weight has lower explanatory power (R2 between 0.05 and 0.15). As a result, the combined changes in inputs yield a relatively small decrease by 0.5 ppt starting from a baseline of 47.6% of the population. The most important predictors are those associated with a healthy diet. Finally, the cross-region macro-economic model of early school leaving has high explanatory power (R2 above 0.90) and highlights a wide range of ‘push and pull’ factors. The combination of benchmark inputs yields a decrease in the rate of early leavers by 1.8 ppt, starting from a baseline of 13.1%. Overall, these results highlight the large scope for policy intervention to improve well-being outcomes, as well as the multiplicity of policy levers.

Co-authored with Murtin F, Vincent Siegerink, and Julien Bonnet.