Abstract: Restoring the underpinnings of John Roemer’s conceptualization of inequality of opportunity, we introduce an innovative empirical approach to measure unfair inequalities through Bayesian networks. This methodology enhances the comprehension of the income formation process through structural learning algorithms, both generating an informative index and revealing the underlying data generating process. We demonstrate how this proposal relates to traditional measurement methods through simulated data and we provide an application to Germany, Italy, and Sweden. With the application, we illustrate the potential of Bayesian networks in this context by looking into the learned networks for the three countries.

Abstract: Measures of income polarization are often motivated as quantifying conflict potential in a society. The key premise is that individuals share a feeling of “group identity” with people with similar income and feel “alienated” from people distant in income. The group identity element makes the concept of polarization distinct from the concept of inequality. Existing measures of polarization however assume static societies. We argue that this is unsatisfactory and propose a measure that incorporates the dynamics of income over time in its elaboration. This inter-temporal income polarization measure introduces memory parameters that allow past income differences to determine the degree of current alienation and identification in a society. This leads to measures of income polarization that are sensitive to the history of interpersonal proximity and distances in income trajectories. We illustrate the empirical relevance of this longitudinal perspective with an application based on administrative data on labour income of cohorts of Italians.

Abstract: We study how local labor demand shocks affect internal migration using the universe of labor market flows for Italy. First, we document two novel facts: i) large and systematic differences between gross and net job and internal migration flows arise both across space and over time; ii) each gross flow is an important driver of the net growth rates. We estimate the causal impact of different-sign labor demand shocks on internal migration flows using as instrumental variable plausibly exogenous large mass hire and layoffs events. Our estimates reveal that job creation has a strong effect on the in-migration rate, whereas job destruction has a much milder effect on the out-migration rate. Crucially, we document that the large responsiveness of in-migration does not work through an increase in the number of relocating  workers, but rather through changes in their chosen destination alternatives. We also find that the effects of job creation on in-migration flows have a much larger geographical reach than those of job destruction, as out-migration flows are locally concentrated.

Abstract: The GC Wealth Project, a central project of the Graduate Center’s Stone Center on Socio-Economic Inequality, is a multi-year effort aimed at expanding and consolidating access to the most up-to-date research and information on wealth, wealth inequalities, and wealth transfers and related tax policies, across countries and over time.

The GC Wealth Project website — first launched in June 2023 — is organized around two main components: a data warehouse of gathered and novel data that can be visualized in a variety of ways through the interactive dashboard, and a Digital Library of Research on Wealth Inequality. Both are designed to provide researchers, policymakers, journalists, and others interested in wealth and wealth taxation with open, unlimited access to an array of high-quality information and resources.

All of the data, including the tailored visualizations that users can create using the interactive dashboard, can be exported.

Abstract: Relying on a matched survey-administrative dataset for the private sector covering years 1975-2018, we investigate the trend of intragenerational earnings mobility in Italy over the past 40 years. We focus on cohorts of workers born from 1940 to 1973 and compare their earnings patterns from age 35 to age 45 refining the method proposed by Nichols (2008) to distinguish ‘good’ and ‘bad’ earnings mobility. We document that the long-run tendency of increasing cross-sectional earnings inequality in Italy has been coupled with widening persistent differences within the same generation. We also uncover a more recent trend of deterioration of the nature of mobility – less career progression in favour of more instability. Moreover, depicting novel intragenerational Great Gatsby curves, we find that cohorts exposed to more earnings inequality also experience more permanent inequality and fewer chances of growth, especially because of the impact of the Great Recession.

Abstract: The work compares across cohorts and different levels of education the early-stage evolution of several labour market outcomes, with the aim of studying whether and to what extent education matters for the level, growth and stability of earnings. By using a rich longitudinal dataset developed from merging survey and administrative data, this article describes the evolution of earnings in the five years following education completion in Italy comparing differently educated workers born between 1970 and 1984. We find evidence of an 'education premium' in terms of faster school-to-work transition, higher employability and higher earnings. Moreover, education is associated with positive, faster and more volatile earnings growth. However, no clear-cut changes across cohorts in the association between the various outcomes and the level of education emerge from our analysis.

Abstract: The economic literature provides evidence that standard demographic characteristics and human capital variables explain at most one third of wage inequality in Mincerian earnings equations. This work explores the unexplained inequality by using Italian linked employer-employee data from 1998 to 2016. I provide evidence that, from the end of the 1990s to 2016, the type of contract became increasingly important in explaining wage inequality, especially annual and weekly, but around 70% of annual earnings inequality, and 40% of weekly and hourly wages remain unexplained by observable characteristics. My results suggest that part of the unexplained variance is due to differences between the firms in which the workers are employed, but the largest contribution remains workers’ unobserved heterogeneity. 

Keywords: earnings inequality, linked employer-employee data, AKM model, Italy.

This thesis stems from research work I carried out in various universities and institutions to provide new perspectives on the analysis of mobility in different research contexts. The first two chapters deal with earnings mobility and its association with inequality (Chapter 1) and polarization (Chapter 2), while the third chapter is concerned with geographical mobility in response to labour market shocks (Chapter 3). In all three cases, the focus of the analysis is on the Italian labour market. However, while the first and last works are applied in nature, the second one is a theoretical paper whose application is instrumental to understanding the theory and demonstrating its empirical relevance.

Abstract: The report investigates the role played by the middle class in Italy over the last decades, in comparison with other European countries, using various proxies of economic well-being (i.e. individual labour income, equivalised disposable income, wealth), and exploiting various cross-sectional and longitudinal sources (i.e. AD-SILC, EU-SILC, SHIW). Adopting an economic definition of the middle class, based on individual positions with respect to median positions in the income scale, we first assess how the relative weight of the middle class – in terms of population share and share of total income – has evolved over time, and then use longitudinal data to explore mobility in and out of the middle class.