BoE Paper Studies Role of Regulation in Triggering Fire Sales by Banks

Post the crisis, along with a more stringent regulatory framework, regulators had introduced regular and comprehensive stress testing frameworks to assess the ability of the financial system to weather severe market stresses. There is a widespread agreement on the need to make stress testing more macro-prudential through inclusion of feedback effects and amplification mechanisms, such as fire sales. While progress has been made in this direction, the existing models are too stylized to provide a realistic assessment of banks’ defensive actions under stress. The authors aim to bridge this gap by building a flexible analytical tool for regulatory stress tests to assess the risks from fire sales. To this end, the authors have developed a flexible and realistic model of fire sales that reflects the post-crisis regulatory environment and can be operationalized in stress testing models. This is because most existing models of fire sales are too stylized for direct use in policy analysis or to study contagion risk in a realistic setting. In this framework for modeling fire sales, banks face both liquidity and solvency constraints and choose which assets to sell to minimize liquidation losses.

The model has been applied to seven UK banks subject to the regulatory stress test in 2017. It enables to explore the likelihood, causes, and magnitude of fire sales in different stress scenarios. The results show that banks constrained by the leverage ratio prefer to first sell assets that are liquid and held in small amounts, while banks constrained by the risk-weighted capital ratio and the liquidity coverage ratio need to trade off assets’ liquidity with their regulatory weights. Following solvency shocks, risk-based capital requirements tend to be more tightly binding and present incentives for banks to sell larger amounts of illiquid assets relative to the leverage ratio, which in turn leads to larger fire-sale losses. Nevertheless, fire-sale losses due to solvency shocks remain moderate even for severe shocks. In contrast, severe funding shocks can lead to large fire-sale losses. Thus, models that focus on solvency shocks and only include a leverage ratio may be missing two key drivers of fire-sale losses.

The existing quantitative models of fire sales tend to assume that banks sell off assets in proportion to their initial holdings. In reality, banks are unlikely to follow this strategy, as this would involve them taking actions that cause significant losses. Allowing banks to optimize their liquidation strategy results in significantly lower losses than assuming that they sell assets in proportional to their holdings. This also implies that the assets more likely to transmit losses in a fire sale are liquid, rather than illiquid, assets. The model delivers the following policy implications:

• It demonstrates the importance of ensuring that liquidity buffers are usable in stress. If banks aim to defend their liquidity positions by protecting their liquid asset buffers, they may fire sell illiquid assets, resulting in large losses. If, as emphasized by regulators, they fully utilize their liquid asset buffers then losses can be reduced.
• It highlights a potentially negative side-effect of the new UK leverage framework, which excludes central bank reserves from the leverage ratio. Doing so removes a bank’s ability to use central bank reserves to deleverage once they breach their leverage ratio requirements, potentially leading to larger losses in a fire sale.

Related Links

• Notification
• Working Paper (PDF)

Keywords: Europe, UK, Banking, Fire Sales, Stress Testing, Systemic Risk, Solvency and Liquidity Constraints, Capital Requirements, Basel III, Liquidity Risk, Research, BoE