Stress testing when Value-at-Risk (VaR) isn’t enough
Stress tests are an indispensable complement to Value-at-Risk (VaR) models.
While the risk factors, distributional assumptions, and pricing functions of VaR models vary, even the most sophisticated approaches should be combined with a stress testing framework in abnormal or crisis periods. In fact, Value-at-Risk models do not adequately capture volatility jumps or changing correlation structures and perform poorly when liquidity dries up, as seen by the Lehman crisis.
The derivatives rule (SEC 18f-4): Going beyond Value-at-Risk
As a result of the derivatives rule SEC 18f-4 passed on October 28, 2020, all SEC-registered mutual funds, ETFs and Business Development Companies (BDCs) with derivative notional exceeding certain threshold are required to appoint a derivatives risk manager in charge of implementing a regulatory framework for its fund’s derivatives use.
The necessary risk guidelines focus on reporting limits of fund leverage risk based on Value-at-Risk (VaR). However, VaR itself is not a leverage measure and other factors can cause a fund’s VaR to move away from its benchmark. For that reason, the derivatives and risk management framework also requires reports on portfolio stress testing and portfolio backtesting results on at least a weekly basis1.
That means you need a tool that gives you access to various factors to construct stress tests relevant to your portfolio and produce accurate results via a full repricing approach that doesn’t make any assumptions about distributions or first order approximations.
Why use portfolio stress tests in addition to Value-at-Risk measures
- Acts as a complement to fundamental factor risk models and analytics like VaR
- Captures risk from volatility jumps in times of crisis/abnormal periods
- More transparent and intuitive than VaR
- Helps to design better risk hedges
You may already have the tools to design a portfolio stress test, but as the old adage goes, ‘garbage in, garbage out.’ So here is what you need to do to ensure your stress tests yield the most useful result.
1) Be relevant to the portfolio
Are the relevant stress factors selected? For example, applying a constant -5% shift stress test to all equities in a long/short equity portfolio is not useful, whereas this same stress test applied to a portfolio of equity options (with nonlinear pricing) is more useful. Better yet would be a correlated or beta-adjusted shift in an equity index for both portfolios mentioned above. Furthermore, running ladder shocks, for e.g., -15%, -10%, -5%, etc. helps you understand the non-linear behavior of portfolios with derivatives under different market conditions.
2) Highlight hidden exposures to macro factors
It’s not sufficient to just track exposures to factors within a fundamental factor model. Stress tests enable you to calculate and track sensitivities to any factor including macro factors like oil, gold, rates, spreads. These serve as a complement to fundamental factor models. You can use this information to hedge against drastic macro events.
3) Use appropriate lookbacks, Exponentially Weighted Moving Average
(EWMA) weighting and sampling based on the factors being stressed. We recommend running rolling weekly returns for multi-asset portfolios in correlated or beta-adjusted stresses to handle any returns timings issues across markets and geographies. Choosing the lookback is also important, especially for macro stresses like oil or VIX as the current market regimes may not reflect correlations that would typically be exhibited under stress conditions. So, choosing a period in time when the correlations were stressed or better yet, utilizing an event-weighting approach to weight the days with the largest stress factor returns more are ways to utilize relevant correlations.
4) Combine market and credit stresses
For example, you can design an anticipated downgrade of an issuer by shifting spreads for bond and CDS holdings that reference the issuer and combine market and credit stresses.
Keeping these stress testing pointers in mind, you’ll be one step closer to tackling the regulatory requirements of Rule 18f-4. If you’re looking for insights on how to implement Value-at-Risk or portfolio backtesting within your program, need a turnkey reporting package or cloud-native risk system, learn more about Axioma Risk.
References:
Request a demo