Interconnectedness as a measure of systemic risk : return spillovers between banking, insurance, and technology

Abstract

This thesis examines the return spillover effects between three sectors along 14 years framework. Our date extends from the beginning of January 2000 till December 2013. In this paper, we use three sectors to test for returns spillovers. Insurance, banking, and technology are the sector adopted in our model. Using the Diebold and Yilmaz 2012 model, we monitor the spillovers in returns on both company and index levels in two different modes: crisis and none crisis. Furthermore, extending the Diebold and Yilmaz 2012 spillover index, we use the frequencyConnectedness package developed by Barunik and Krehlik (2018) to monitor the frequency of connectedness between both the companies and indexes. This package is an extension for the DY spillover index and helps us monitor the connectedness closer than the original model. First, results show that the highest spillover exist between the insurance and banking sectors on both company and index levels. Second, high frequencies of connectedness available in the first bounds mean that markets seem to be processing quietly and quickly. Finally, 1 to 10 days is the bound with the highest spillover and the most accurate measure for spillovers. • Diebold & Yilmaz: We investigate the interconnectedness between three relatively important sectors: insurance, banking, and technology in the United States throughout three major crises that extended from 2000 till 2014 most notably the DotComm bubble, the financial crisis of 2007/8, and the European debt crisis that first sparked end of 2010. We measure systemic risk between these sectors by adopting the Diebold and Yilmaz models of 2009 and 2012 specifically the generalized variance decomposition approach on both levels: the index and the company. Results show a relatively high level of spillover effect among those sectors that of 55%. Furthermore, results underline that a strong correlation exists between banking and insurance rather than those two sectors with technology. Our findings present a contribution for regulators to further concentrate on limiting the interconnectedness between the banking and insurance sectors for the future. v • Frequency Dynamics (Barunik & Krehlik) We extend our study to monitor how the spillover among the banking, insurance, and technology sectors evolve throughout different time frequencies. Building up on the Diebold & Yilmaz (2012) model presented in our previous chapter we use Barunik & Krehlik (2018) package in R "frequencyConnectedness" to further dissect the interconnectedness between the different three sectors. Results show that markets have processed shocks imminently and little effect remained in the system in the long term. Moreover, the first ten days have the highest levels of spillovers. Our paper contributes in offering regulators the chance to monitor the first 10 days furthermore whether by adopting a higher frequency model that would give them the chance to find the inflection point, the point where markets are processing the data available. Doing this will further improve the ability of the regulators to limit the spread of losses among sectors. The study also has been conducted on both the index and company level.