Network Connectivity, Systemic and Systematic Risk

Key Findings

• We show that (i) network exposures act as an inflating factor for systematic exposure to common factors; (ii) the power of diversification is reduced by the presence of network connections.
• Empirically, in the presence of network links, a misspecified traditional linear factor model presents residuals that are correlated and heteroskedastic. We support our claims with an extensive simulation experiment.
• This approach allows us to decompose the risk of a single asset (or a portfolio) into four components: (i) the systematic component, (ii) the idiosyncratic component, (iii) the impact of the asset interconnections on the systematic risk component, the contribution of network exposure to the systematic risk component, and (iv) the effect of interconnections on the effect of idiosyncratic risk on the systematic risk component (the amplification of idiosyncratic risks that generates systematic/non-diversifiable risk).
• Our approach also allows us to decompose the risk premium component of returns into three components: the risk premiums associated with (i) common factor exposures, (ii) the impact of asset connections on common factors and (iii) the amplification effects of idiosyncratic risk.
• The simulation analysis we perform shows that the new model we propose can be used to better understand the different components of systematic risk and volatilities and analyze the causes of systematic risk.
• Moreover, the simulation allows us to disentangle the error estimation of linear factor models that ignore the presence of network connections. In particular, we show that the residual correlations start drifting away from zero if network connectedness is ignored in the model estimation.
• Finally, we confirm the simulation analysis using real data based on equity returns and a network based on the input-output table. 

Policy Implications

This new model is relevant for policymakers and regulators since they need to be aware of the implications of the different possible policy choices on network connections and their effects on equilibrium stock returns and volatilities. For investors and other market participants, the model is relevant since they need to understand whether and to what degree network connectivity has an impact on risk premiums, volatilities, and spillovers between markets.