Network analysis of the EU insurance sector

This paper contains an analysis of the network of the 29 largest European insurance groups and their financial counterparties. Insurance companies have direct exposures to other insurers, banks and other financial institutions through the holdings of debt, equity and other financial instruments. These exposures can cause direct contagion and thereby the spread of systemic risks. This analysis focuses on direct linkages between EU insurers and banks. Sectoral data show that at least 20% of insurers’ assets are investments in banks. As a result insurers are an important source of funding for banks. This paper adds to the expanding research on financial market networks and on systemic risks in the insurance sector. This paper considers these 29 insurers’ top ten exposures for each instrument. They represent about 10% of their total assets, which indicates a low level of concentration. More than half of the reported exposures result from investments in bonds issued by banks. In addition, some insurers have a higher exposure to banks within their own financial conglomerate. Other exposures reported include securities lending transactions and repos (both mostly collateralised with cash), as well as interest rate swaps. The network of insurance groups, banks and other financial institutions displays low interconnectivity overall, compared for instance to the interactions of the largest EU banks alone. The density of the network is relatively low. The characteristics of this network illustrate that credit and funding events cannot be expected to spread easily through direct contagion. The network shows a core-periphery structure, which partly results from the scope of the data collection: only 29 insurance groups reported exposures, and thus they form the core, while institutions from which no information was collected are in the periphery. However, the systemic importance of a few insurance groups stands out. These groups show higher levels of connectivity, proximity to credit events within the network, and importance for financial flows. Network measures for each of these aspects refer to the criteria of interconnectedness and substitutability, which are well-known in the policy debate on the systemic relevance of financial institutions. While the particular form of institutional importance varies significantly across insurance groups, the central role of a few “champions” in this network may require supervisory attention. Size, in terms of total assets and issued capital is an important factor but not the only one, determiningan insurer’s centrality in the network. Measures of the centrality of the banking counterparties also show a positive, non-linear relationship with the size of the banks. Distress in the network, as expected, causes only limited direct contagion. Insurers’ solvency positions are sufficiently large and their concentration of exposures is sufficiently low as to avoid direct contagion from a default of one of their counterparties, ultimately leading to their own default. This is also true for simultaneous distress at their top ten banking counterparties, with the exception of two insurance groups. The analysis presented is partial, as it relies only on the information collected. In particular, the analysis would benefit from including the exposures of banks and other financial institutions to insurers. This could be a useful enhancement for consideration in the future. JEL Classification: L14, G18