STRESS TESTING AND SCENARIO ANALYSIS

Stress Testing and Scenario Analysis of Barclay's Bank

Stress Testing and Scenario Analysis of Barclay's Bank

Introduction

Stress Testing

Stress testing is a risk management tool for firms to evaluate the potential impact of an event or movement on the firm's asset quality, profitability, capital and other financial variables. Stress testing aims to primarily identify latent exposures - those that are less obvious, perhaps hidden across a wide variety of instruments, credits and derivatives positions. It focuses on traded market portfolios, which are well suited to stress testing because they can be marked to market on a regular basis. Stress testing of funding liquidity and operational risk is common in financial institutions such as banks (the scope of this paper) although loan book stress testing is less frequent (Stress Testing, 2004).

Stress tests are a valuable aid in assessing the stability of banking systems. They permit a forward-looking analysis and adopting a uniform approach to identifying potential risks to the banking system as a whole. Unlike Value at Risk (VaR), which reflects price behavior in everyday markets, stress tests simulate portfolio performance during abnormal market periods. Thus, in general, VaR and stress test exposure estimates are not added to each other.

Stress testing is needed for the following reasons (Stress Testing, 2004):

Capturing the impact of exceptional but plausible large-loss events on a portfolio

Checking if the capital buffer is sufficient under stress conditions

Introducing forward-looking elements in the capital assessment process

Reducing reliance on model parameters (e.g., when historical correlation may no longer be valid)

Ascertaining changes in the business environment, e.g., in liquidity

Reviewing changed horizons and liquidity of instruments

Supporting portfolio allocation decisions beyond the range of normal business conditions

Identifying hidden correlations within portfolios

Assessing the tail events beyond the level of confidence assumed in a given statistical model because, under stress conditions, the following may occur (Stress Testing, 2004):

Less predictability in the behaviour of stress factors

Rapid price movements and contagion may affect other markets

Shocks may spread across multiple markets

Economic conditions in affected regions may suddenly deteriorate

A stress test measures the x% tail in a profit and loss distribution of an asset, while VaR measures the mark-to-market gain or loss that an asset would experience, should one or more of the underlying economic factors that determine the asset's value experience a specific change in value.

Stress tests capture exceptional but plausible events: they provide information about risks falling outside those typically captured by the VaR framework (Figure 1). These risks include those associated with extreme price movements and with forward-looking scenarios not reflected in the recent history of the price series used to compute VaR.

Scenario Analysis

Scenario analysis is a method that tries to describe logical and internally consistent sequences of events to explore how the future might, could or should evolve from the past and present. The future is inherently uncertain. Through scenario analysis different alternative futures can be explored and thus uncertainties addressed. As such, scenario analysis is also a tool to deal explicitly with different assumptions about the ...