When Banks Meet the Official Sector
Paper Session
Friday, Jan. 7, 2022 12:15 PM - 2:15 PM (EST)
- Chair: Uday Rajan, University of Michigan
Systemic Risk and Monetary Policy: The Haircut Gap Channel of the Lender of Last Resort
Abstract
We show that lender of last resort (LOLR) policy contributes to higher interconnectedness and the buildup of systemic risk in the banking sector. Our analysis uses a novel micro-level dataset that links the securities pledged by banks to obtain LOLR funding with the haircuts applied by the LOLR and by private repo markets. We exploit the variation across securities in the haircut gap, i.e. the difference in valuation haircuts between the private market and the central bank. We find that LOLR policy provides incentives for banks to increase their holdings of bonds with higher haircut gaps, especially those issued by other, interconnected, banks. This is consistent with theories of interbank monitoring rather than risk sharing. Stronger interconnectedness arises from an increase in home bias, especially for banks in distressed economies. Among domestic banks, higher haircut gaps increase the pledging of bonds issued by systemically important banks and stimulate the cross-pledging of bank bonds, in line with theories of bailout expectations in the event of a systemic crisis. Consistent with an increase in the demand for bank bonds with higher haircut gaps, we also document that LOLR policy stimulates their issuance by banks in distressed economies.Joint Determination of Counterparty and Liquidity Risk in Payment Systems
Abstract
We propose a methodology to assess how banks jointly manage their funding liquidity and counterparty risk in the context of an interbank payments system that operates with deferred net settlement (DNS). Our model allows us to evaluate critical features of the financial system that are usually analysed in isolation, such as the issuance of secured and unsecured credit obligations and the use of collateral and capital requirements. Throughout the day, banks issue payment orders, which represent claims on central bank balances that must be settled at the end of the day. A payments operator processes all payment orders and acts as the central counterparty (CCP). To remain risk neutral, the operator collects collateral from either the issuer or the recipient of a payment order using a defaulter-pay or survivor-pay arrangement. These arrangements closely resemble collateral and capital requirements in the wider banking system. Using intra-day data from the Canadian wholesale payments system, known as the Large Value Transfer System (LVTS), our results show that banks coordinate the issuance of payment orders to jointly manage their liquidity and counterparty exposures. Coordination leads to netting of credit exposures and unencumbering of collateral assets, which increases liquidity. In addition, we show that other things being equal, banks prefer to issue unsecured payments and do not see secured and unsecured payments as substitutes, although they coordinate the issuance of both types of payments. For unsecured payments, banks rely on both bilateral and multilateral coordination, whereas for secured payments, they rely almost exclusively on multilateral coordination. The differences in coordination arise because unsecured payments are contingent on the performance of a given counterparty, whereas secured payments only depend on the performance of the collateral supporting them. Thus, to the extend that collateral is homogenous, secured payments are fungible regardless of the issuer. Coordination and netting incentives increase with risk exposures and the cost of funding. We conclude that coordination disruptions may increase risk exposures and lead to collateral shortages and funding constraints, particularly among small participants, who tend to net less and require relatively more collateral to issue and to receive payments. In an extreme scenario, coordination disruptions could lead to gridlock and systemic risk. Therefore, coordination is an important risk management tool that should be considered when designing market infrastructures and regulations aimed at enhancing financial stability.Time Inconsistency in Stress Test Design
Abstract
We show that central banks face a time inconsistency problem when publishing bank stress test results. Before a stress test, they want to appear tough as the threat of letting banks fail the stress test incentivizes prudent behaviour. After the stress test, they want to act soft by releasing only partial information in order to reassure financial markets about bank health. We characterize an institutional design solution to this commitment problem: a social planner sets the framework within which the central bank communicates. We find that a hurdle rate framework, where all banks are judged to pass or fail relative to a common threshold, is optimal in many settings as it generates intermediate levels of both incentives and reassurance. With a hurdle rate framework, stress tests become an informational contagion channel, as changes in the health of one bank affect beliefs about the health of other banks. Thus, informational contagion can be a feature of a socially optimal institutional design in the presence of a time inconsistency problem.Discussant(s)
Matthew Plosser
,
Federal Reserve Bank of New York
Angelo Ranaldo
,
University of St. Gallen
Rodney Garratt
,
University of California-Santa Barbara
Nicolas Inostroza
,
University of Toronto
JEL Classifications
- G2 - Financial Institutions and Services