The formation of a core-periphery structure in heterogeneous financial networks

Balancing liquidity and returns through interbank markets: Endogenous interest rates and network structures

“…By deriving realistic properties of the interbank lending network in the absence of credit risk, we show that the observed properties are primarily the result of bilateral transactions between banks rather than any differences in their ex ante properties. As Cohen-Cole et al ( 2010), Haᴌaj and Kok (2015), Lux (2015), and Castiglionesi and Navarro (2020) show the emergence of core-periphery structures in the presence of counterparty risk and Farboodi (2021) and in 't Veld et al (2020) show core-periphery structures for heterogeneous banks, our results suggest this is a universal property of interbank lending markets.…”

“…By assuming banks are homogenous in most aspects apart from their initial liquidity buffers and the necessary adjustments to the amounts of loans and deposits to accommodate this shock while maintaining the total asset size, we are not assuming heterogeneity in other respects and the liquidity shock is decisive for the emergence of interbank network features such as the core-periphery structure. We agree that any differences between banks play key roles in determining whether a bank is core, as pointed out in the literature (see, e.g., in 't Veld et al, 2020;Lux, 2015;Näther, 2018), and in real markets, banks are indeed heterogenous in many such aspects. The advantage of making our restrictive assumption is that the model gives a different perspective by enabling us to explore the emergence of some interbank network features even without heterogeneity in size or investment opportunities.…”

“…For interbank networks, empirically an error score in the region of 0.3 is commonly found. Other algorithms as developed in Holme (2005) and Da Silva et al (2008) lead to similar results.…”

“…Specifically, Farboodi (2021) shows that banks with risky investment opportunities form the core, and those lacking such opportunities populate the periphery. in 't Veld et al (2020) require banks of different sizes to establish such a network structure. Differences in information about other banks' behavior in the interbank market or a more efficient processing of such information also lead to a core‐periphery structure as shown in Näther (2018).…”

“…Such models are highly unsatisfactory as they are based on statistical models rather than on the actual behavior of individuals-banks in our case-and thus do not provide a framework to assess the impact of incentives and regulations on the properties of interbank networks. Farboodi (2021) and in 't Veld et al (2020) employ heterogeneous banks to achieve a core-periphery structure. Specifically, Farboodi (2021) shows that banks with risky investment opportunities form the core, and those lacking such opportunities populate the periphery.…”