Tuesday, July 21, 2009

Stressed-Out VaR

Posted on Financial Times Alphaville by Tracy Alloway:

One of the centrepieces of Lord Turner’s review of banking regulation is the potential use of so-called stressed VaR — an attempt to rectify the failings of the ‘normal’ Value at Risk model, one of the pillars of upcoming banking regulations on market risk.

From the report:

The predominant assumption [in the financial industry] was that increased complexity had been matched by the evolution of mathematically sophisticated and effective techniques for measuring and managing the resulting risks. Central to many of the techniques was the concept of Value-at-Risk (VAR), enabling inferences about forward-looking risk to be drawn from the observation of past patterns of price movement. This technique, developed in the early 1990s, was not only accepted as standard across the industry, but adopted by regulators as the basis for calculating trading risk and required capital, (being incorporated for instance within the European Capital Adequacy Directive).

This implies that any use of VAR models needs to be buttressed by the application of stress test techniques which consider the impact of extreme movements beyond those which the model suggests are at all probable. Deciding just how stressed the stress test should be, is however inherently difficult, and not clearly susceptible to any mathematical determination.
Put simply, VaR is a way for banks to estimate how much of their portfolio they may expect to lose in a given timeframe at a certain amount of confidence. As an example, here’s a chart from JP Morgan, which has just published a whopping note on UK banks, risk and regulation, showing the distribution of HSBC’s 2007 trading revenue:

JPM chart of HSBC 2007 trading

If you look at the chart, the 99 per cent VaR would be about $70m as only 1 per cent of the losses over HSBC’s 259 trading days are greater than this amount (three days). Now, contrast that with HSBC’s 2008 trading pattern:

JPM chart of HSBC 2008 trading

That year there were 19 days when losses were greater than $70m, or 7.3 per cent of the time. Furthermore, in 2008 there were two days when the loss was greater than $170m. This is a good illustration of how the VaR does not give an indication of the size of loss, only the frequency, and is highly dependent on the assumption of the distribution — the inputs. So far, so old (and for an excellent overview of the shortcomings of VaR do read the full section in the Turner Review) and hence the stressed VaR proposal - which is aimed at holding VaR calculations to significantly more stringent market conditions. That means using inputs taken from times of financial stress relevant to the bank’s portfolio — so 2007/2008 as opposed to say, 2005/2006.

Only, as JPM notes, even stressed VaR has its limitations:

Given that VaR (or even stressed VaR) does not measure the maximum potential loss, and does not capture liquidity and systemic risks, we believe that it will be supplemented with additional position constraints, such as gross leverage ratios and balance sheet liquidity requirements.

That would have massive capital implications for the banks.

In fact, JP Morgan are estimating a 30 - 100 per cent increase in the UK banking sector’s capital requirements as a result of stressed VaR, additional incremental risk charges (aimed at capturing some of the liquidity risk not encapsulated by VaR) and the requirement for ‘higher quality’ of capital.

Stressed out VaR? Or stressed-out banks?

The full 140-page JP Morgan banking capital experience in the Long Room.

No comments: