A0817
Title: Exploring financial networks using quantile regression and Granger causality
Authors: Samriddha Lahiry - Harvard University (United States) [presenting]
Sumanta Basu - Cornell University (United States)
Diganta Mukherjee - Indian Statistical Institute (India)
Kara Karpman - Cornell University (United States)
Abstract: Granger causality-based techniques to build networks among financial firms using time series of their stock returns have received significant attention recently. Existing Granger causality network methods model conditional means and do not distinguish between connectivity in the lower and upper tails of the return distribution - an aspect crucial for systemic risk analysis. Statistical methods are proposed that measure connectivity in the networks using tail-based analysis. They are able to distinguish between connectivity in the lower and upper tails of the return distribution. This is achieved using bivariate and multivariate Granger causality analysis based on regular, and Lasso penalized quantile regressions, an approach called quantile Granger causality. An asymptotic theory of quantile Granger causality estimators is provided under a quantile vector autoregressive model, showing its benefit over regular Granger causality analysis on simulated data. The proposed method is applied to the monthly stock returns of large U.S. firms and demonstrates that lower tail-based networks can detect systemically risky periods with higher accuracy than mean-based networks. In a similar analysis of large Indian banks, it is found that upper and lower tail networks convey different information about periods of high connectivity that are governed by positive vs negative news in the market.