The Leibniz Institute SAFE organizes a brown bag seminar on
Clustering-based Sector Investing
Matteo Bagnara (Leibniz Institute SAFE)
Abstract: Industry classification groups firms into finer partitions to help investments and empirical analysis. No solution has been proposed yet for the well-documented problems of standard industrial definitions, like their stale nature and coarse categories for firms operating in multiple areas. We allocate firms to homogeneous groups maximizing the within-group explained variation through a clustering approach that employs 69 firm characteristics. The resulting novel economic sectors represent a better investment set compared to existent classification schemes for portfolio optimization and for trading strategies based on within-industry mean-reversion. We provide a new metric to quantify feature importance for clustering methods, finding that size drive differences across classical industries while book-to-market and financial liquidity variables matter for clustering-based sectors.
