The great financial crisis in 2007-2008 demonstrated how excessive systemic risk can bring the financial system to the brink of a financial meltdown when it is hit by adverse shocks, such as by losses from the subprime mortgage market. It further demonstrated, once again, that a financial crisis tends to result in recessions, or even depressions, harming the real economy. Our focus is on developing methodologies to measure systemic risk and evaluating policies to design a more resilient financial system.
We have been working together with the Bank of England, the ECB and the Bank of South Africa to develop models that can be used to better understand systemic risk and to quantify its dangers. This includes a better understanding of the interaction between the channels of contagion through which risk is propagated and amplified. Our goal is to create tools that can be used to make quantitative forecasts of risk and that can be used to investigate alternative policies for reducing systemic risk.
Our work
This project focuses on:
- Systemic risk and stress testing
- Climate finance
- Market ecology
System-wide stress tests
A natural representation of the financial system is as a financial network. The nodes in the network represent the financial institutions/assets while the edges represent the various direct or indirect linkages between institutions. We model how financial shocks can propagate through the financial network via both direct (e.g. loans, bonds and derivatives) and indirect linkages (e.g. common asset holdings). Unlike microprudential stress tests which focus on assessing the direct impact of a crisis shock on the balance sheet of institutions, system-wide stress tests capture how initial shocks can propagate through the financial network and be amplified via interacting contagion mechanisms. System-wide stress tests can be used to measure systemic risk in the financial system and evaluate (macroprudential) policies to reduce such risk.
Ending “too big to fail”
The 2007-2008 financial crisis forced governments to choose between the unattractive alternatives of either bailing out a systemically important bank or allowing it to fail disruptively. Bail-in has been put forward as an alternative that potentially addresses the too-big-to-fail and contagion risk problems simultaneously. Does it indeed resolve the “too-big-to-fail" problem? An evaluation of the financial stability implications of the bail-in design suggests the answer is no (unless there are significant improvements in that design).
Basel II capital requirements & Basel III capital and liquidity buffers
We study the impacts of capital and liquidity regulation on financial stability. Our studies show that the Basel II capital regulation, which relies on setting minimum capital requirements based on the Value-at-Risk, results in destabilizing leverage cycles. Basel III capital regulation sought to avoid the pitfalls of its earlier counterpart by introducing regulatory capital buffers, which sit on top of capital requirements. The idea is that by making regulatory buffers “usable” – in the sense that these can be drawn down to absorb losses – destabilizing leverage cycles would be avoided. We show that financial stability critically depends on how usable capital and liquidity buffers are perceived to be by banks.
Large exposure limits
A key pillar of the post-crisis regulatory reforms is the introduction of large exposure limits to curb concentrated exposures to a single counterparty. In practice, adherence to these limits is measured based on an institution’s exposures to a counterparty arising via its loans, bonds, and other direct linkages. We show that only capturing “direct exposures” misses a significant part of exposures, defined as the loss an institution is exposed to upon the default of a counterparty. Instead, exposure measures should be revisited fundamentally to not only include direct, but also “indirect” and “higher-order” exposures. Indirect exposures arise from indirect linkages via common asset holdings. Higher-order exposures emerge from contagious spill-over losses, following direct and indirect losses.
Market ecology
Standard approaches to the theory of financial markets are based on equilibrium and efficiency. Using concepts and methods developed by biologists, we can view the wealth invested in a financial strategy as the abundance of a species. The wealth dynamics of the market ecosystem explain how market inefficiencies spontaneously occur. The composition of a market gives insight into the origins of excess price volatility and deviations of prices from fundamental values. While wealth dynamics act over longer timescales, the call from investors to deliver steady excess returns drives fund managers to innovate, leading to evolution of investment strategies and more rapid changes in market composition. Fine-grained reporting data allows regulators to track activities of market participants and their proportional sizes. Combined with modelling of the market activities, this gives regulators the tools to better predict long-term changes in financial markets, and the impact of policy changes such as the effect of quantitative easing. Classifying investment activities based on their expected impact on qualitative aspects on the market provide investors with better information to guide their choice of investment funds. Our market ecology research covers several areas Market ecology and financial stability; Risk management of investment strategies; and Opponent modelling in adaptive markets.
Key findings
- The financial system may be stable or unstable for a given microprudential stress test outcome. Microprudential stress tests should be complemented with macroprudential stress tests to account for systemic risk.
- Multiple interacting contagion mechanisms should be captured jointly to avoid underestimating systemic risk.
- Institutions might have large indirect or higher-order exposures to another institution, even though the direct exposures might be small or even zero.
Funders
Baillie Gifford, European Central Bank, Fidelity Investments, James S. McDonnell Foundation, JP Morgan Chase & Co.
Other outputs
- The Economic Simulation Library (ESL) provides an extensive collection of high-performance algorithms and data structures used to develop agent-based models for economic and financial simulation. https://github.com/INET-Complexity/ESL