Stress Testing

Stress testing and capital planning are increasingly linked too many risk management processes that require coordination across risk, treasury, and financial planning and analysis functions.

A successful stress testing program demands flawless integration and execution of numerous tasks, including program governance, process and results validation, documentation, data quality management, economic scenario development, expected loss modeling, forecasting, and reporting. 

Establishing an integrated approach to the use of stress testing results require infrastructure, analytics, data, clear governance, and active participation of all stakeholders across divisions, business units and geographies.

.

Stress testing involves running computer simulations to identify hidden vulnerabilities in institutions and investment portfolios to evaluate how well they might weather adverse events and market conditions.

However, a robust risk management framework demands the implementation of scenario simulation where the distribution is extremely skewed towards the tails (tail events); a the situation that happens in very rare cases. Such shocks could be caused by various macro-economic or idiosyncratic events, which can consequently spread widely to previously thought of as uncorrelated choices of assets (Systematic or else Undiversified risk). 

 Vivid examples of historical crises that resulted in large losses of invested capital within a certain period (varying from a few days to a few months) include the Black Monday of 1987, the Gulf War of 1990, the Asian Crisis of 1997, the Russia Devaluation of 1998 and the Financial Crisis of 2008.

Although a large portion of such losses was often due to excessive leverage and over-concentration of positions, one wonders if an an investor should use such extreme divergences from normality to stress test their seemingly well-diversified multi-asset portfolios.

Admittedly, we should not rely on the the assumption that history repeats itself since background conditions and driving factors often differ vastly between distant periods of economic and market activity. However, to assess and verify that adequate capital is preserved to cover any unexpected losses, investors could attempt to estimate the impact on the portfolio performance that the re-occurrence of such damaging historical events could have.

In this way, stress testing would give us an idea of how stretched the loss-tolerance levels of a risk management the strategy would turn out to be during a crisis of historical precedence.

To conduct such an analysis, one would need to select a the historical crisis of relevance (a subjective choice) and apply changes to the risk factors driving the price of various asset classes (e.g. equities, bonds, credit, commodities, foreign exchange) accordingly, to assess their impact on the current portfolio if an identical market condition occurs. Such market shocks might be global or local in nature; consequently, geographical disparities in valuation changes would need to be identified and incorporated.

The date ranges one could use for such stress testing, could largely match the length of the a referenced historical event, with the return change being the cumulative return over the entire testing period.

This requires the risk management system to give as much flexibility on stress testing modeling as needed since every portfolio or strategy will not be subject to the same risk factors. It should at least contain quantitative information about extreme market movements during historical crises, the application of which, to the current portfolio would have to be straightforward. 

It would also be advantageous to provide the functionality for custom stress modeling based on a combination of lessons taken from several past crises such as significant price movements, correlation dislocations, etc. 

The latter will give the manager the confidence required for estimating the portfolio performance on extreme market conditions with scenarios tailor-made for its unique nature.

Once the adverse scenario simulation has been chosen, regardless of it being based on historical events or custom (hypothetical) market conditions, it can be applied to the current portfolio for a detailed risk evaluation analysis. Methods such as Stressed VaR that is perceived to be common market practice will provide sufficient information, whilst the ability to decompose the risk exposure into its core sources will provide a more detailed view of the concentration risk the portfolio may face. 

Moreover, the ability to implement portfolio tail the analysis will enhance the manager’s confidence in regards to the capital adequacy to cover significant losses even in the most detrimental conditions.