
The untapped potential of stress tests
Quants propose technique to include stress testing in portfolio allocation
Most financial institutions have stepped up their stress-testing efforts post-crisis. Driven by regulatory requirements such as the Comprehensive Capital Analysis and Review (CCAR) in the US, the European Banking Authority (EBA) stress test in Europe and international capital rules that use stressed risk measures, comprehensive stress tests are being conducted regularly by banks to ensure they have the necessary capital to withstand extreme market moves.
Both banks and buy-side firms also run specific scenarios internally – for instance, before the Brexit vote – for risk management purposes.
However, despite its widespread use, the application of stress-testing results still remains quite limited.
Stress testing often functions as a supplementary tool, added as an extra layer on top of usual risk management to work out how a portfolio might react given a certain type of market event. This is for hedging purposes, contingency planning or to help inform how much capital to hold against potential losses. However, when stressed situations arise, this doesn’t necessarily mean portfolios are automatically positioned to withstand them.
“While many organisations run portfolios through a number of historical and hypothetical scenarios, it is typically done to build awareness of possible outcomes. Using scenario output to plan a reaction function, as a way to rebalance in light of a future stress event, is relatively rare”, says a quant at a risk solutions vendor.
In this month’s first technical, Stress hedging in portfolio construction, Alex Ulitsky and Maurizio Ferconi, both managing directors at investment manager BlackRock in San Francisco, and Mehmet Bilgili, a vice-president at the same firm, develop a technique to use stress-testing results as part of the initial selection of the portfolio itself. In other words, the portfolio already reflects views on possible stressed outcomes.
Typically, businesses manage extreme risks by buying insurance or options rather than embedding stress testing directly in the optimisation process.
The quants, on the other hand, incorporate stress testing in portfolio construction by adding a constraint to the popular mean-variance optimisation method, which helps control losses during extreme events.
Mean-variance optimisation, widely used for portfolio construction, works by maximising returns relative to risk. This involves minimising an objective function based on the variance-covariance characteristics of the portfolio, which indicate the riskiness of the portfolio.
One novel aspect of the paper is being able to account for situations where stress events cannot be fully specified, as is often the case when assessing rare scenarios
The quants take this optimisation a step further by incorporating both normal market conditions and stressed conditions within the objective function so the portfolio can be positioned to sufficiently withstand stressed events right from the start.
What allows for the blending of normal and stressed market conditions in the technique presented by the quants are coefficients within the optimisation that can be set based on an investor’s level of risk aversion to a likely scenario. This allows for the two market states to be mixed and included in the allocation decision depending on the level of risk aversion, as opposed to dealing with the two conditions separately, which is the norm.
This translates into a cost-benefit trade-off between holding less of a risky asset and giving up on alpha – the return in excess of a benchmark – and being able to better withstand extreme events.
One novel aspect of the paper is being able to account for situations where stress events cannot be fully specified, as is often the case when assessing rare scenarios or ones that haven’t occurred yet. Because of the uncertainty around the values of the parameters, only a select range of possible values can be assigned to model parameters.
One risk manager at a European bank says the uncertainty in scenario parameters is typically addressed in practice by running multiple scenarios. The paper on the other hand, solves this by formulating a worst-case scenario optimisation, a technique which can account for the uncertainty.
It was only after the financial crisis that the industry started paying more attention to stress tests. Although stress testing has increasingly improved in terms of sophistication in recent years, it still functions as a separate tool from risk management, designed to understand what to expect during a potential stressed scenario.
In addition, building a realistic scenario is tricky because historical and hypothetical scenarios need not reflect future market conditions. This has encouraged firms to carry out reverse stress tests, which is a way of working backwards to see what type of scenario would cause a large loss or liquidity problem for the firm.
The authors of the paper further widen the scope of stress testing by using it to allocate portfolios that are already somewhat resilient to stressed events.
One thing that is clear from all of this is that stress testing is still an evolving discipline, and there is much more that can be done with it.
Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.
To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe
You are currently unable to print this content. Please contact info@risk.net to find out more.
You are currently unable to copy this content. Please contact info@risk.net to find out more.
Copyright Infopro Digital Limited. All rights reserved.
As outlined in our terms and conditions, https://www.infopro-digital.com/terms-and-conditions/subscriptions/ (point 2.4), printing is limited to a single copy.
If you would like to purchase additional rights please email info@risk.net
Copyright Infopro Digital Limited. All rights reserved.
You may share this content using our article tools. As outlined in our terms and conditions, https://www.infopro-digital.com/terms-and-conditions/subscriptions/ (clause 2.4), an Authorised User may only make one copy of the materials for their own personal use. You must also comply with the restrictions in clause 2.5.
If you would like to purchase additional rights please email info@risk.net
More on Our take
Getting a handle on model parameters
Mean reversion in rate parameters opens the door to dimensionality reduction
The case for believing in a Bessent put
Money market funds could prove critical in efforts to control 10-year yields
FRTB may bite harder for Europe’s CVA modellers
Farther reach of advanced approach and lighter load on total requirements mean limited takeaways from Canada and Japan’s implementation
Japan, Basel III and the pitfalls of being on time
Capital floor phase-in delay may be least-worst option for JFSA as US and Europe waver
FX traders revel in March Madness
Chaotic Trump policies finally bring diversity to flows – to the delight of market-makers
Market knee-jerks keep VAR models on their toes
With a return to volatility, increased backtesting exceptions show banks’ algos are stretched
A market-making model for an options portfolio
Vladimir Lucic and Alex Tse fill a glaring gap in European-style derivatives modelling
How AI agents could become investing’s crash test dummies
Firms mull the use of chatbot simulations to test organisational set-ups