Tous les séminaires
Jeudi 23 Février :
16H15 : Christian BROWNLEES (Universitat Pompeu Fabra)
"Community Detection in Partial Correlation Network Models"
17H30 : Patric GAGLIARDINI (University of Lugano)
"A Diagnostic Criterion for Approximate Factor Structure"
Salle S016, CREST-INSEE, Bâtiment Malakoff 2, 15 Bd Gabriel Péri, 92245 Malakoff
Christian BROWNLEES (Universitat Pompeu Fabra)
COMMUNITY DETECTION IN PARTIAL CORRELATION NETWORK MODELS
Many real-world networks exhibit a community structure: The vertices of the network are partitioned into groups such that the concentration of linkages is high among vertices in the same group and low otherwise. This motivates us to introduce a class of Gaussian graphical models with a community structure that replicates this empirical regularity. A natural question that arises in this framework is how to detect the communities from a random sample of observations. We introduce an algorithm called Blockbuster that recovers the communities using the eigenvectors of the sample covariance matrix. We study the properties of the procedure and establish consistency. The methodology is used to study real activity clustering in the U.S. and Europe.
Papier joint avecGuðmundur Stefán Guðmundsson, and Gábor Lugosi
Patrick GAGLIARDINI (University of Lugano)
A DIAGNOSTIC CRITERION FOR APPROXIMATE FACTOR STRUCTURE
We build a simple diagnostic criterion for approximate factor structure in large cross-sectional equity datasets. Given a model for asset returns with observable factors, the criterion checks whether the error terms are weakly cross-sectionally correlated or share at least one unobservable common factor. It only requires computing the largest eigenvalue of the empirical cross-sectional covariance matrix of the residuals of a large unbalanced panel. A general version of this criterion allows us to determine the number of omitted common factors. The panel data model accommodates both time-invariant and time-varying factor structures. The theory applies to random coefficient panel models with interactive fixed effects under large cross-section and time-series dimensions. The empirical analysis runs on monthly and quarterly returns for about ten thousand US stocks from January 1968 to December 2011 for several time-invariant and time-varying specifications. For monthly returns, we can choose either among time-invariant specifications with at least four financial factors, or a scaled three-factor specification. For quarterly returns, we cannot select macroeconomic models without the market factor.
Papier joint avecElisa Ossola, and Olivier Scaillet.
Jeudi 16 Mars :
- Hao Zhou (Tsinghua University) - - Ostap Okhrin (Dresden University of Technology)
Jeudi 23 Mars :
- David Veredas (Vlerick business School) – Sylvestre Frezal (Datastorm)