By A.V. Vedpuriswar
Acknowledgement
This presentation draws heavily from the work of Prof John C
Hull, an authority on Risk management.
Introduction
VAR tells us the maximum loss a portfolio may suffer at a
given confidence interval for a specified time horizon.
If we are 95% sure that the portfolio will not suffer more than $
10 million in a day, we say the 95% daily VAR is $ 10 million.
Problem
Average revenue = $5.1 million per day
Total no. of observations = 254. Std dev = $9.2 million
The point such that the no. of observations to the left = (254) (.05) = 12.7
(12.7 11) /( 15 11 )
.4
1.7 / 4
=
- $9.6 million
$ 15.2 million
Problem
The VAR on a portfolio using a one day time horizon is USD
100 million. What is the VAR using a 10 day horizon ?
Solution
Variance scales in proportion to time.
So variance gets multiplied by 10
316
(N2 = 12 + 22 .. = N2)
Note; This approach is valid only when the daily variances are
independent.
9
Problem
(12,500) (5)
27,951
Monthly VAR
( 12,500) (20)
55,902
(12,500) (125)
139,754
Annual VAR
(12,500) (250)
197,642
10
11
Problem
Consider a position consisting of a $100,000 investment in asset A and a $100,000
investment in asset B. Assume that the daily volatilities of both assets are 1% and
that the coefficient of correlation between their returns is 0.3. What is the 5-day
99% VaR for the portfolio?
Solution
The standard deviation of the daily change in the investment in each asset is
$1,000. The variance of the portfolios daily change is
1,0002 + 1,0002 + 2 x 0.3 x 1,000 x 1,000 = 2,600,000
The standard deviation of the portfolios daily change is $1,612.45.
14
Problem
Solution
= 2%
= (.02) (10,000,000)
= $200,000
Z (p = .01)
= Z (p =.99)
= 2.33
Daily VAR
= (2.33) (200,000)
= $ 466,000
10 day VAR
= 466,000 10
= $ 1,473,621
15
Problem
Consider a portfolio of $5 million in AT&T shares with a daily
volatility of 1%. Calculate the 99% VAR for 10 day horizon.
Solution
= 1%
(.01) (5,000,000)
= $ 50,000
Daily VAR
(2.33) (50,000)
= $ 116,500
$ 116,500 10
= $ 368,405
10 day VAR =
Problem
513,129
$1,622,657
220,277
= 219,369
17
Problem
Consider a portfolio made up of 5 year 5 % coupon government
bonds. The bonds are trading at $ 100. The historical annual
volatility is 1 %. Expected YTMs are normally distributed with zero
mean and volatility of 1%. Calculate the 95% one year VAR.
Solution
Worst YTM = actual YTM + 1.65 x Volatility
= 5 + 1.65 x 1 = 6.65%
18
Problem
Consider the following single bond of $10 million, a modified
duration of 3.6 yrs and annualized yield volatility of 2%.
Calculate the 10 day holding period VaR of the position with a 99%
confidence interval, assuming there are 252 days in a year.
Solution
VAR = $10,000,000* 0.02*3.6* 10/252 * 2.33 = $334,186
19
Problem
Assume that a risk manager wants to calculate VAR for an S&P
500 futures contract using the historical simulation approach. The
current price of the contract is 935 and the multiplier is 250. Given
the historical price data shown below for the previous 300 days,
what is the VAR of the position at 99% using the historical
simulation methodology?
Returns: -6.1%,-6%,-5.9%,-5.7%, -5.5%, -5.1%..........4.9%, 5%,
5.3%, 5.6%, 5.9%, 6%
Solution
The 99% return among 300 observations would be the 3rd worst
observation among the returns.
Among the returns given above -5.9% is the 3rd worst return, the
99% VAR for this position is therefore
(935)*250* (0.059) = $13,791.
20
Problem
Consider the portfolio of an American trader with two foreign
currencies, Canadian dollar and euro. These two currencies are
uncorrelated and have a volatility against the dollar of 5% and 12%
respectively. The portfolio has $2 million invested in CAD and $1
million in Euro. What is the portfolio VAR at 95% confidence level?
Solution
Variance of the portfolio return
$ .156205 million
$257,738
Undiversified VAR
$ 363,000
21
Problem
.0245
= .0245
$.1565 million
VAR
$ 258,225
Marginal VAR
(1.65) (156,500)
= 258,225 257,738
= $ 487
22
Problem
An American trader owns a portfolio of options on the US dollar-sterling exchange
rate. The delta of the portfolio is 56.0. The current exchange rate is $/ 1.5000.
Derive an approximate linear relationship between the change in the portfolio value
and the percentage change in the exchange rate. If the daily volatility of the
exchange rate is 0.7%, estimate the 10-day 99% VaR.
Solution
The approximate relationship between the daily change in the portfolio value, P,
and the daily change in the exchange rate, S, is P = 56 S
For a unit change in , $ will change by 1.5. It follows that
P = 56 x 1.5 x
Or
P = 84 x
The standard deviation of x equals the daily volatility of the exchange rate, or 0.7
percent.
The standard deviation of P is therefore 84 x 0.007 = $ 0.588.
So the 10-day 99 percent VaR for the portfolio is
0.588 x 2.33 x 10
23
Problem
Some time ago a company entered into a forward contract to buy 1 million for $1.5 million. The
contract now has 6 months to maturity. The daily volatility of a 6-month zero-coupon sterling
bond (when its price is translated to dollars) is 0.06% and the daily volatility of a 6-month zerocoupon dollar bond is 0.05%. The correlation between returns from the two bonds is 0.8. The
current exchange rate is $/ 1.53. Calculate the standard deviation of the change in the dollar
value of the forward contract in 1 day. What is the 10-day 99% VaR? Assume that the 6-month
interest rate in both sterling and dollar is 5% per annum with continuous compounding.
Solution
The contract is a long position in a sterling bond plus a short position in a dollar bond.
1.463 x 0.0005
24
Problem
Consider the contract on the dollar/euro exchange rate (EC)
traded on the CME. The notional amount is 125,000 Euros.
Assume that the annual volatility is 12% and the current price is
$1.05 per Euro. Find daily VAR at 99% confidence level.
Solution
VAR = (2.33) [(.12) / 252] (125,000 1.05)
= $ 2310
25
Problem
Consider a portfolio with a one day VAR of $1 million.
Assume that the market is trending with an auto correlation of
0.1. Under this scenario, what would you expect the two day
VAR to be?
Solution
V2
22 (1 + )
2 (1)2 (1 + .1)
2.2
2.2
1.4832
26
The first days movement will affect the third days movement
through the square of the correlation coefficient and so on.
Then the combined variance will be:
+ (2) () 2
Problem
Consider a portfolio with standard deviation of daily returns of 0.1
and autocorrelation of 0.3. Calculate the 5 day volatility.
Solution
= 0 .1; = .3
Variance
.080642
Volatility
.284
28
29
30
Generate Scenarios
The first step is to generate all the price and rate scenarios
necessary for valuing the assets in the relevant portfolio, as
well as the required correlations between these assets.
31
32
Probability Distribution
Monte Carlo simulations are based on random draws from a
variable with the required probability distribution, usually the
normal distribution.
The normal distribution is useful when modeling market risk in
many cases.
But it is the returns on asset prices that are normally
distributed, not the asset prices themselves.
33
34
Other approximations
There are also other approximations that use delta, gamma ()
and theta () in valuing the portfolio.
By using summary statistics, such as delta and gamma, the
computational difficulties associated with a full valuation can be
reduced.
37
38
39
Historical Simulation
40
Introduction
Unlike the Monte Carlo approach, it uses the actual historical
distribution of returns to simulate the VAR of a portfolio.
Steps
Collect data
Generate scenarios
Calculate portfolio returns
Arrange in order.
42
% Returns
- 16
- 14
- 10
-7
-5
-4
-3
-1
0
1
2
4
6
7
8
9
11
12
14
18
21
23
Frequency
1
1
1
2
1
3
1
2
3
1
2
1
1
1
1
1
1
1
2
1
1
1
Cumulative Frequency
1
2
3
5
6
9
10
12
15
16
18
19
20
21
22
23
24
26
27
28
29
30
So VAR = -7
Disadvantages
Simple
No normality assumption
Non parametric
Weighting of data
44
46
Model Validation
47
48
49
0-4 exceptions
Yellow zone :
5-9 exceptions
Red zone
10 or more exceptions
Problem
Based on a 90% confidence level, how many exceptions in
back testing a daily VAR model should be expected over a
250 day trading year?
Solution
10% of the time loss may exceed VAR
25
51
Problem
We are currently feeding a model with 600 days of data. The
VAR confidence level is 99%. Nine exceptions are observed.
Should we reject the model? Suppose it had been 12. Would
we reject the model?
Solution
.019
Problem
We back test a VAR model with 1000 days of data. The VAR
confidence level is 99%. 17 exceptions are observed. Should the
model be rejected at 5% significance level?
Solution
.01
2.64% < 5%
2.64%
Problem
A $10 million one year loan has a 1.25% probability of default. If
there is a default, the recovery can be anything from 0 to the full
loan value. Calculate the 99% VAR and conditional VAR.
Solution
Default zone = 1.25% starting from 98.75% going up to 100%.
Loss at 99% point =[ .25/1.25] X 10 = $ 2 million.
So the 99% VAR is $ 2million.
54
Stress Testing
55
Introduction
Stress testing involves analysing the effects of exceptional
events in the market on a portfolio's value.
56
57
Gaussian Copulas
Consider variables, V1 and V2 that are not normally
distributed.
Map the two variables on to a normal distribution.
Apply correlation
Create bivariate normal distribution.
59
Illustration
Consider 2 variables that have a uniform distribution. Using
Gaussian copula, and assuming a copula correlation of 0.3,
define a correlation structure.
0.25
0.50
0.75
0.25
0.50
0.75
60
0.25
0.50
0.75
0.25
-.675, -.675
-.675, 0
-.675, +.675
0.50
0,
0.75
+.675, -.675
+.675, 0
+.675, .+ 675
0.25
0.50
0.75
0.25
.095
.1633
.2157
0.50
.1633
.2985
.4134
0.75
.2157
.4134
.5953
-.675
0, 0
0,
+ .675
61
Illustration
x
percentile
percen
tile
0.1 5.00
-1.65
2.00
-2.06
0.2 20.00
-.84
8.00
-1.41
0.3 38.75
-.29
18.00
-.92
0.4 55.00
.13
32.00
-.47
0.5 68.75
.49
50.00
0.6 80.00
.84
68.00
.47
0.7 88.75
1.21
82.00
.92
0.8 95.00
1.65
92.00
1.41
0.9 98.75
2.24
98.00
2.06
62
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.1
.006
.017
.028
.037
.044
.048
.049
.050
.050
0.2
.013
.043
.081
.120
.156
.181
.193
.198
.200
0.3
.017
.061
.124
.197
.273
.331
.364
.381
.387
0.4
.019
.071
.149
.248
.358
.449
.505
.535
.548
0.5
.019
.076
.164
.281
.417
.537
.616
.663
.683
0.6
.020
.078
.173
.301
.456
.600
.701
.763
.793
0.7
.020
.079
.177
.312
.481
.642
.760
.837
.877
0.8
.020
.080
.179
.318
.494
.667
.798
.887
.936
0.9
.020
.080
.180
.320
.499
.678
.816
.913
.970
63
Maturity
0.8 years
Coupon
Annualised yield
Volatility
3 month
6 month
1 year
Solution
The current position involves the following:
Cash flow of $50,000 in .3 years
3.6 months.
5.50%
6.00%
Volatility
50,000
(1.056).3
.6
.06 (.04)
3
49,189
.068%.
Here = .068
12 + 22 + 2 12
= .06
2 = .10
= .90
$ (.2397) (49,189)
get: .06
1,050,000
(1.066).8
3.6
(.01)
6
=.066
= $997,662
= 0.16
LHS =.0256
.006096 .0256
(.6797) (997,662)
$50,000
$1,050,000
t = .3
t = .8
Total
--
319,589
331,380
12 month bond
678,074
678,074
--
37, 397
2,628,536
$1621.3
1621.3 x 10 x 2.33
$11,946