This paper shows that altruism – the preference for transferring resources to future generations – is heterogeneous and persistent within dynasties, and an important driver of intergenerational inequality, given its influence on both physical and human capital transmission. Using Italian microdata, I document two facts. First, youths expecting an inheritance are substantially more likely to attend university, conditional on parental wealth, income, education, and transfers. Second, individuals who received inheritances are substantially more likely to intend to leave bequests, conditional on current resources. I develop a quantitative life-cycle model in which dynastic altruism follows a Markov process across generations, calibrated to match both empirical facts. Removing heterogeneity in bequest motives shrinks the enrollment gap between expected heirs and non-heirs by 63%; removing heterogeneity in coresidence and inter vivos transfers shrinks it by 37%; the anticipated wealth transfer itself has instead a large negative effect, as in its absence the gap widens by 53%. Persistence in altruism is critical: with altruism uncorrelated across generations, expected heirs become less likely to study, as the disincentivizing wealth effect dominates. Inheritance taxation funding student grants substantially raises enrollment and weakens the link between inheritance expectations and educational attainment.

UPF Applied Lunch Seminar, BSE Jamboree, Helsinki PhD Workshop on Economics of Education, EALE 2023 (Prague), UPF-CREI Macro Lunch, 1st Imperial College PhD Conference in Economics and Finance, EDP Jamboree 2024 (Barcelona), 3rd PhD and Post-Doctoral Workshop in Economics and Finance (Naples), 10th CEPR European Workshop on Household Finance (Copenhagen), Workshop on the Origins of Socioeconomic Inequality (Luxembourg), 28th Theories and Methods in Macro - T2M (CREST), 28th Workshop on Dynamic Macroeconomics (Vigo), 8th Doctoral Workshop on Quantitative Dynamic Economics (Konstanz), 5th Banca d'Italia Workshop on Human Capital (poster)

We document a robust positive relationship between income inequality and current account balances. In advanced economies, a one percentage point increase in the share of disposable income held by the top 1 percent is associated with a 0.8 percentage point increase in the current account balance. This relation is driven by inequality in the permanent component of income and operates primarily through higher private saving rather than changes in investment. We rationalize these findings using a tractable two-country heterogeneous agent model with non-homothetic preferences, where saving rates increase with permanent income. The model predicts capital flows from unequal to equal countries and helps explain the observed co-movement between inequality and global imbalances since the 1980s.

(* if by coauthor) 12th PhD Student Conference on International Macroeconomics (Paris Nanterre), European Central Bank IPA Economic Meeting*, Bank for International Settlements Research Seminar*, 2023 Annual Meeting of the Central Bank Research Association - CEBRA (New York), 23rd RIEF Doctoral Meeting*, 27th Theories and Methods in Macro - T2M (Amsterdam), PSE-CEPR Policy Forum 2024, UCL Stone Center PhD Conference on Income and Wealth Inequality, EEA-ESEM 2024 (Rotterdam)*, BdF-EUI Conference on International Macro, Banque de France–Pierre Werner Chair Conference on International Macro

EuHEA Conference 2024 (Vienna), Age-IT Meeting (Naples)

In surveys, it is seldom possible to obtain the most accurate available measure of any variable of interest, given the high costs (monetary and not) associated to the necessary procedures. In the context of Understanding Society, we consider the setting of blood pressure measurements as gathered by nurses and compare them with self-taken and interviewer-taken ones. Assuming classical measurement error for nurse-taken measures and non-classical error for the other two sources, we estimate the parameters of interest through Generalised Method of Moments (GMM) in order to correct for any possible bias in measures exhibiting non-classical error.