13 Questions on Risk Management
   What are the major risks resulting from ...
     Credit Risk
     Market Risk
     Liquidity Risks
     Legal Risk
     Operational Risk


















 

What are the major risks resulting from financial instruments?

Market Risk

Market risk is the risk to an institution’s financial condition resulting from adverse price or volatility moves of the assets contained in a firm’s portfolio. It is different from the firm’s mark-to-market, which is the current value of the firm’s financial instruments. Market risk represents what the firm could lose if prices or volatility changed. A firm must measure the market risks resulting from its portfolio of financial instruments and senior managers must decide the frequency of this measurement. Firms with active portfolios should calculate their exposures daily while those with small portfolios could do it less frequently.

Market risk can be measured as the potential gain or loss in a position or portfolio that is associated with a price movement of a given probability over a specified time horizon. This is the value-at-risk (VAR) approach. How it should be measured is a decision taken by the board of directors on the advice of senior managers; external consultants and auditors can be co-opted if senior managers feel that they have inadequate knowledge to deal with this very technical issue.

While the actual measurement of market risk is undertaken by the line managers, senior managers must decide on the key parameters which are used in any value-at -risk approach. These are the time horizon, i.e. what holding period is covered by the potential loss/gain number; and the confidence interval, i.e. what percentage of portfolio price changes is taken into account by the value-at-risk calculation.

Line managers must ensure that the market risk information sent to senior managers and the board is easily understood and accompanied with the appropriate caveats. The relevant time horizon and confidence interval used in the market risk calculations must be highlighted in any report because they are relevant to interpreting the numbers. For example, a daily VAR number of $10 million at 95% confidence interval means that the firm could lose/gain up to $10 million on 19 out of 20 trading days. A daily VAR of $10 million at 99% confidence interval means that the firm could lose/gain up to $10 million on 99 out of 100 trading days. Intuitively, the firm with a daily VAR of $10 million at 99 % confidence interval has a ‘less risky’ portfolio than the one with a daily VAR of $10 million at 95 % confidence. Converting this intuition into numbers requires that they be normalised to a common standard; Table 3 on page 24 shows VAR numbers which have been normalised to what a firm could probably lose/gain in one year (columns 3 and 4 in table 3).

A $10 million VAR based on 10-days and 99% confidence interval has been chosen as the benchmark because the Basle Committee for Banking Supervision has decided that these parameters form the basis for market risk capital for banks. The table shows clearly that giving your firm’s VAR without the accompanying time horizon and confidence interval is almost meaningless. For example, a firm with a daily VAR of $10 million (at 99% confidence interval) could lose up to $159 million a year compared with a firm with a 10-day VAR of $10 million (at 99% confidence interval), which could lose up to $51 million annually. Further, a firm with a daily VAR of $10 million at 95% confidence interval could lose up to $224 million annually. The table also drives home the point to shareholders that they must be told these key parameters in any information they receive about a company’s market risk exposure.

Table 3:

WHOSE PORTFOLIO IS MORE RISKY?

Parameters VAR Annual 1% Loss 1 (99% confidence) Annual 5% Loss 1 (95% confidence) Comparative Riskiness
10-days, 99% $10m $51m $36m 1.0 (benchmark)
10-days, 95% $10m $72m $51m 1.4
1-day, 95% $10m $224m $159m 4.4
1-day, 99% $10m $159m $112m 3.1
5-days, 95% $10m $102m $72m 2.0
5-days, 99% $10m $72m $51m 1.4
1(a) To convert a daily VAR into an annual figure, multiply $10m by the square root of 252 (15.87), a 5-day number by the square root of 52 (7.2), and the 10-day number by the square root of 26 (5.1). (Assume that there are 252 trading days, 52 weeks and 26 fortnightly periods a year and that there are five working days a week.)

For linear risk ,the fact that the square root of time rather than time itself is the correct method of scaling up or down (i.e. using square root of 52 as opposed to using 52 outright when scaling up a weekly number into an annual number) is rooted in statistical theory and can be mathematically proven. In any process where events are random and independent of each other (for e.g. stock prices), it can be shown mathematically that the variance of the process increases proportionally with the number of events. This distance from the average is usually represented by the standard deviation which is obtained by square rooting variance. Since variance is proportional to time, it then follows mathematically that standard deviation is proportional to the square root of time.

1(b) Since this column represents an annual 1% loss, VAR calculations based on 95% confidence have to be normalised upwards to 99% confidence. This is achieved by multiplying the annual 95% loss by 1.4. (1.4 is derived from dividing 2.33 by 1.65; 2.33 standard deviations is 99% confidence and 1.65 standard deviations is 95% confidence interval.)

(2)Repeat 1(a) to obtain the annual VAR losses. Since this column represents 95% confidence, all VAR calculations based on 99% confidence have to be normalised downwards to 95% confidence. This is achieved by dividing the annual 99% loss by 1.4 .

The table is extracted from: "Managing Derivative Risks" by Lillian Chew, John Wiley & Sons, 1996.

VAR is effective in measuring risk for normal market price changes. A prudent firm will supplement VAR with stress testing to examine the impact of various types of financial distress on the firm’s portfolio. The tests should identify events that could have an unfavourable effect on the institution and should assess the financial ability of the institution to withstand them. These analyses should consider not only the likelihood of adverse events but also worst-case scenarios. Such stress tests should not be limited to quantitative exercises that compute potential losses or gains; they should also include more qualitative analyses of the actions management might take under particular scenarios, for e.g. operating procedures and lines of communication. Senior managers should decide on the type of stress testing techniques to be used.


Box 8
Stressing the portfolio
Credit Suisse Financial Products conducts scenario analysis, i.e. it revalues its portfolio after determining an appropriate movement in the main underlying assets within a specified interval, to estimate the profit or loss of the portfolio under extreme market circumstances. Scenarios are based on the worst one-week market movements in the past five years and on any other major financial event outside the five-year period that is relevant to the bank’s business. Some of the scenarios the firm has examined include the bond market crash of February/March 1994, the Exchange Rate Mechanism crisis of 1992 and flight to quality currencies, the stock market plunge of 1987 and the movement of oil prices during the Gulf War. Scenario analysis is carried out weekly, and the results are reviewed by senior management. The results and movement are explained in detail on a monthly basis and are reviewed by the Board of Directors and the Market Risk Committee on a quarterly basis.

The results of Commerzbank’s stress simulations, together with any recommendations are presented each month to the full Board of Managing Directors.

In its 1995 annual report, the German bank disclosed the amount of capital the bank had to set aside to cover possible losses brought about by extreme market turbulence. If the bank wanted to set aside capital for a fivefold standard deviation move, it would need DM 370.6 million; to cover the greatest possible loss for a historical 250-day period, DM 205.9 million.

Sources: Credit Suisse Financial Products Annual Review, 1995; Commerzbank Annual Report, 1995

Continue

Case Studies * 13 Questions on Risk Management * What are the major risks resulting from financial instruments?