Showing papers by "Francesco Bartolucci published in 2020"

Longitudinal networks of dyadic relationships using latent trajectories: evidence from the European interbank market

Abstract: Financial markets are ultimately seen as a collection of dyadic transactions. We study the temporal evolution of dyadic relationships in the European interbank market, as induced by monetary transactions registered in the electronic market for interbank deposits (e‐MID) during a period of 10 years (2006–2015). In particular, we keep track of how reciprocal exchange patterns have varied with macro events and exogenous shocks and with the emergence of the Global Financial Crisis in 2008. The approach adopted extends the model of Holland and Leinhardt to a longitudinal setting where individuals’ temporal trajectories for the tendency to connect and reciprocate transactions are explicitly modelled through splines or polynomials, and individual‐specific parameters. We estimate the model by an iterative algorithm that maximizes the log‐likelihood for every ordered pair of units. The empirical application shows that the methodology proposed may be applied to large networks and represents the process of exchange at a fine‐grained level. Further results are available in on‐line supplementary material.

Multilevel Model-Based Clustering: A New Proposal of Maximum-A-Posteriori Assignment

Abstract: We deal with the problem of latent variable prediction in the context of multilevel latent class models for categorical responses provided by individuals nested in groups. In particular, we propose a posterior assignment rule that jointly predicts the individual- and group-level latent variables. This proposal is alternative to the common maximum-a-posteriori rule, which is based on first predicting the latent variables at cluster level and, then, those at individual level. To illustrate the proposal, we show the results of two simulation studies and two applications on data related to the national and the international assessment of student skills.

Primary-school class composition and the development of social capital

Abstract: We study the development of social capital through adult civic engagement, in relation to social capital exposure having occurred during childhood based on experiences outside the family at primary school. We assume that the types of classmates in attendance at a child's school would have influenced her/his social capital. To identify the types of classmates, we take advantage of the heterogeneity in the ability levels of British primary-school classes during the 1960s. At that time, some schools were practicing a method of streaming, whereas others were not. Using British National Child Development data, we construct a single score of civic engagement and evaluate the effect on adult civic engagement of attending homogeneous-ability classes versus nonhomogeneous-ability classes and being in high-, average- or low-ability classes when enrolled in streamed schools. Our results show that children who were grouped in homogeneous-ability classes developed a lower interest in civic engagement than their peers who attended mixed-ability classes (nonstreamed schools). Moreover, among children who attended streamed schools, a lower attitude toward civic engagement was observed among low-ability students. Thus, streaming appears to be detrimental to social capital development, especially for low-ability individuals.