Case Study 2

Advanced Risk Management

Due: Feb 1st, 2022, 8am

1 Instructions

1. Read the case study �Credit General, S.A.�in your course reader.

2. The case states that on the pre-speci�ed maximum sterling position,
Credit General had a daily 99 percent VAR of DM 4.4 million and a
10-day 99 percent VAR of 13.8 million. Use the data in the study to
verify (approximately) that these VAR numbers are accurate.

3. At the end of the day, Credit General had a realized daily loss of DM
10.50 million and a total (including unrealized) loss of DM 34.51 mil-
lion. How can you reconcile the loss with the 4.4 million VAR number
computed in the previous question? The realized loss was about 2.5
times larger than the VAR and the total loss was a staggering 8 times
as large as the VAR. Was the VAR calculation grossly inaccurate or
did something else go wrong?

4. Writeamemoonbehalf ofMr. PierreMarsonto theBoardofDirectors
of Credit General. Explain why the realized losses are larger than the
VAR measures. In your memo, be sure to use the analysis from points
1-3 above.

5. In the memo you must also make a recommendation on how to decrease
the exposure to pound sterling.

(a) Delineate several possible ways of decreasing Credit General�s
exposure to the pound.

(b) Analyze the two possibilities that seem viable to you.

(c) Explain which of the two seems more appropriate to implement.

(d) Be speci�c on what trades you would take, the amounts of the
trades, and what would be the expected outcome of these trades.

(e) Compute the VAR of the portfolio with the decreased pound
exposure.

1

Harvard Business School 9-296-011
August 15, 1995

This case was prepared by Professor André F. Perold as the basis for class discussion rather than to illustrate either effective
or ineffective handling of an administrative situation. The organization and some of the facts in this case have been
disguised.

Copyright © 1995 by the President and Fellows of Harvard College. To order copies or request permission to
reproduce materials, call 1-800-545-7685 or write Harvard Business School Publishing, Boston, MA 02163. No
part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in
any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without the
permission of Harvard Business School.

1

Crédit Géneral, SA

Late one afternoon in June, 1995, Mr. Pierre Marson was faced with an important decision.
As head of the Asset & Liability Committee (ALCO) at Crédit Géneral, S.A., he was responsible for
approving overnight positions that exceeded established guidelines. Marson had just been informed
that the bank’s foreign exchange group was saddled with a significant pound sterling exposure. At
£700 million, this position vastly exceeded the £150 million daily maximum. Moreover, on a mark-to-
market basis, it was already showing a loss.

Marson had to decide whether to allow the position to remain as is overnight. To quickly and
substantially reduce the position would be very costly given current market conditions. It would also
lock in a significant loss. Another possibility was to try to hedge the exposure with the use of options.
This, too, could prove costly but would leave open the possibility of making a profit if the sterling
market were to rebound during the next several days.

Background

Crédit Géneral was a large French institution engaged in a broad array of financial businesses
in institutional and retail markets worldwide. These included lending, deposit-taking, investment
and merchant banking, brokerage, insurance, and asset management. In recent years, some of the
bank’s principal activities such as foreign exchange trading and derivatives issuance accounted for a
significant and growing portion of its revenues. Exhibit 1 shows the bank’s summary financial
statements for 1994.

The foreign exchange group operated principally out of Paris, aided in overnight trading by
the bank’s dealing rooms in New York and Singapore. The group traded most currencies in at least
medium size, and French francs, the Japanese yen, the U.S. dollar, and the Deutschemark in very
large size. In these latter currencies, the group focused on serving major customers, presenting itself
as being especially competent in trading in large volume.

Becoming successful in large volume currency trading was something of a chicken-and-egg
problem. It required doing large volume business with the major players in these markets—such as
corporations, hedge funds, and central banks—in order to learn enough about supply and demand to
make informed pricing and positioning decisions on large trades.

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

296-011 Crédit Géneral, SA

2

The sterling/mark trade

The sterling position arose out of a trade between Crédit Géneral and a German corporate
client. The client had recently collected on a £1 billion receivable, which it wished to convert into
Deutschemarks. It had indicated its desire to do the trade fairly quickly, given its anxiety about the
possible effects on the currency markets of new U.S. economic data to be released during the
upcoming week. The client would give the business to Crédit Géneral if the bank would do the trade
on a principal basis at a reasonable price.

In a principal trade, Crédit Géneral would quote and guarantee a price in advance. Its gross
profit would be derived from the difference between the guaranteed price and the price at which the
trade actually would be executed. If the trade were done on a so-called agency or best-efforts basis,
Crédit Géneral would be paid a commission and the client would bear the risk associated with
executing the trade.

Ordinarily, Crédit Géneral would not bid for sterling on such a scale. It had less expertise in
sterling than currencies like the Deutschemark and the U.S. dollar, and sterling in any event tended to
be considerably less liquid than the major currencies. A yard (billion pounds) of sterling was at least
20 times the size of the typical interbank trade. Nevertheless, the foreign exchange group saw this as
an opportunity both to satisfy the needs of an important client and to enhance its stature in the
sterling market.

At around 08:00 London time, when the European currency markets were beginning to open
and the Asian markets were about to close, Crédit Géneral traders were busy assessing the depth of
the sterling market. Their goal was to develop a strategy for executing the trade, and to assess its
feasibility in the first place. The sterling market was most liquid during London morning hours, and
least liquid overnight during Asian trading hours.

The traders had four basic channels of communication with other market makers. The first
was the Reuters screen-based dealing system which allowed direct two-way communication between
dealers. This was the most active currency trading mechanism, used principally for rapid
communication and smaller-sized trades. The second channel was the electronic broker market. This
screen-based system was used by dealers to communicate indirectly and anonymously with other
dealers through online broker-intermediaries. The third channel was likewise a broker-intermediated
market except that communication took place between dealer and broker by telephone. The fourth
channel involved dealers talking directly with one another by telephone.

The information gleaned by the traders suggested that the sterling market would continue the
relatively calm pattern it had exhibited during the last week. Expected market volatility, as implied
by the prices of one-month sterling put and call options, was slightly lower than in recent days. These
volatilities were around 10% for the sterling/mark, and 14% for the dollar/mark.1 (Exhibits 2-5 give
historical information relating to the levels and volatility of the sterling/mark, the dollar/mark, and
the sterling/dollar.) Moreover, Crédit Géneral’s traders had received indications of interest from a
large U.S. bank to buy as much as several hundred million pounds. They also had received
indications of interest from many other market makers to buy in smaller size. Finally, the group
believed that Prime Minister John Major’s political problems were overblown and that the pound
would strengthen in the near future.

1A key determinant of option prices is expected price volatility. In the widely-used Black-Scholes formula, price
volatility is expressed in terms of the standard deviation of fractional price changes. The Black-Scholes formula

for near-term, at-the-money currency put or call options is approximately equal to 0.4σ t P, where σ is the
annualized standard deviation of fractional price changes, t is the life of the option expressed as a fraction of a
year, and P is the price of the underlying asset.

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

Crédit Géneral, SA 296-011

3

The foreign exchange group concluded that the yard sterling could be laid off relatively easily
if all channels were hit as rapidly and as close to simultaneously as possible. Moreover, if the bank
had to hold some of the position for a few days, it likely would profit if the pound strengthened as
forecast.

Shortly after 09:00, Crédit Géneral approached the client and enquired whether the bank
would be bidding competitively against other institutions. The client responded “No, it’s just you.”
Crédit Géneral then offered the client 2.2356 DM/ £ on the full yard. This price was 50 pips outside
the current bid side of the market which was being quoted on Reuters screens at 2.2406/2.2413
(bid/offer) for small trades. The present 7 “pip” bid/ask spread was typical for trades in the range of
£10 to £20 million. (1 pip = .01 pfenning = .0001 marks.) Medium-to-large corporate deals would
sometimes go inside this spread, for example, when a dealer was eager to get the business for its
information value. The client accepted Crédit Géneral’s price, and the deal was consummated.

Trade execution

Crédit Géneral immediately moved to execute the trade as planned, attempting to sell the
yard sterling for marks. However, the sterling market suddenly seemed to have “evaporated.” The
price of sterling fell rapidly, as did liquidity. The U.S. bank still was willing to do several hundred
million pounds, but now at 3.5 pfenning (350 pips) below the price which Crédit Géneral had
guaranteed the client. Moreover, market makers who had previously expressed interest no longer
were willing to buy in anything other than small size. Then a rumor surfaced that a British
conglomerate, engaged in takeover discussions with a German company, would soon be entering the
market to buy marks for sterling in significant quantities. The rumor ultimately proved unfounded
but further unsettled the market nevertheless.

The Crédit Géneral traders quickly laid off what they could on the U.S. bank—£250 million—
at a 3.5 pfenning loss, and then spent the rest of the day attempting to get the remainder done. The
traders did their best to “talk up” the market, but to little avail. By late afternoon, they had managed
to do only an additional £50 million, leaving the bank with its current exposure of £700 million.
Sterling now was being quoted on Reuters screens at a 10 pip spread: 2.2013/2.2023.

At 17:00, the head of the bank’s risk management group was appraised of the situation. His
initial reaction was disbelief that this could have occurred, and he indicated in the strongest terms
that the position would have to be reduced. Not having the authority to approve an overnight
position in excess of 20% over established guidelines, he quickly involved the bank’s treasurer, with
whom major risk management decisions often were coordinated, and Mr. Pierre Marson, who, as
head of ALCO, did possess the requisite authority.

Assessment of the situation

Crédit Géneral used a “value-at-risk” methodology to quantify the risk exposures of
individual trading positions, as well as the combined positions of a particular trading desk or the
firm’s aggregate portfolio. It measured value-at-risk as the 99th percentile worst-case outcome that
the position would have experienced on a one-day basis over the past three years. Sometimes it also
utilized a 10-day value-at-risk measure, defined analogously.

The bank presently used value-at-risk to set guidelines on maximum position sizes. In the
near future, it also was planning use value-at-risk to allocate capital and capital costs to individual
businesses.

The approximately 3.5 pfenning movement in the sterling/mark price had resulted in a total
mark-to-market loss of DM 34.5 million, made up of a realized loss of DM 10.5 million, and an
unrealized loss of DM 24 million, derived as shown in Exhibit 6. These figures contrasted starkly

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

296-011 Crédit Géneral, SA

4

with the one- and ten-day value-at-risk estimates on the maximum £150 million position allowed in
the guidelines. Over the last three years, the worst daily move in sterling had been -1.3%, for a one-
day value-at-risk of DM 4.4 million, and the worst 10-day move in sterling had been -4.1%, for a ten-
day value-at-risk of DM 13.8 million. The realized loss alone was more than twice the bank’s value-
at-risk allowance for sterling. And the total mark-to-market loss was more than twice the ten-day
value-at-risk. Finally, the total loss represented almost a month’s worth of the bank’s typical
revenues from foreign exchange trading.

Marson feared that to close out the position in the present circumstances would result in
losses potentially much greater than the present DM 34.5 million. On the other hand, to keep the
position overnight would expose the bank to the risk of further adverse moves. A standard method
of protecting against severe down moves was to purchase out-of-the-money sterling/mark put
options. However, this market was in even worse shape than the spot market. The options were
being bid in small size at prices that implied expected market volatility of around 12%. (Exhibit 7
gives theoretical options prices for different strike prices and volatilities.)

Another alternative was to try to reduce the riskiness of the position by taking offsetting
positions in currencies that were correlated with the sterling/mark. The most liquid candidate was
the dollar/mark. Over the last three years, the correlation between the daily price changes of these
currencies had averaged .3 and had been as high as .75 earlier in the year. Most recently, the
correlation had been around .4 (see Exhibit 5). Since it was only mid-day in the U.S. markets, there
still was ample liquidity available both in the dollar/mark currency as well as dollar/mark options.
The options were being quoted at implied volatilities just under 14%. Exhibit 8 shows the risk
reduction that can be obtained by taking offsetting positions in a correlated currency, for varying
assumptions about the correlation between the currencies.

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

Crédit Géneral, SA 296-011

5

Exhibit 1
Crédit Géneral Financial Statements as of December 31, 1994

Consolidated Income Statement
(all numbers in millions)

FF DM US

Interest receivable and similar income 71,428 20,709 13,361
Interest payable and similar charges (59,938) (17,379) (11,212)
Net commissions 4,212 1,221 788
Net income from securities transactions and

other banking operations
6,206 1,799 1,161

Net banking revenue 21,908 6,351 4,098

Other net income 705 205 132
Net gains on sale of long-term shareholdings

and properties
2,583 749 483

Net change in provisions on long-term
shareholdings and properties

(431) (125) (81)

Share in net income of companies accounted for
by the equity method

625 181 117

Total revenue from operations 25,390 7,361 4,749

General and administrative expenses (14,814) (4,295) (2,771)
Depreciation of fixed assets (808) (234) (151)
Goodwill amortization (314) (92) (59)
Net income before provisions and tax 9,453 2,739 1,767

Net provision for loan losses and other risks (5,538) (1,606) (1,036)
Net income before tax 3,914 1,133 731

Income tax (1,413) (409) (264)
Net income 2,502 724 467

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

296-011 Crédit Géneral, SA

6

Exhibit 1 (continued)
Crédit Géneral Financial Statements as of December 31, 1994

Consolidated Balance Sheet
(all numbers in millions)

Assets
FF DM US

Treasury operations and interbank transactions 330,129 51,048 32,934
Customer loans and credits 315,862 91,580 59,084
Credits under finance and operating leases 33,380 9,678 6,244
Securities received under resale agreements 141,793 41,111 26,523
Trading securities 132,944 38,545 24,868
Securities held for sale 75,569 21,910 14,135
Investment securities 44,136 12,797 8,256
Other accounts relating to securities 13,494 3,912 2,524
Long-term shareholdings and related

receivables
13,247 3,840 2,478

Share in net equity of companies accounted for
under the equity method

25,330 7,345 4,738

Fixed assets 7,906 2,292 1,478
Goodwill 3,453 1,001 646
Other assets 52,740 15,292 9,866
Total assets 1,035,921 300,350 193,774

Liabilities
Treasury operations and interbank transactions 307,867 89,261 57,588
Customer deposits 170,697 49,491 31,930
Securities delivered under repurchase

agreements
157,207 45,580 29,406

Trading securities 66,510 19,283 12,441
Bonds and other negotiable debt securities 195,537 56,693 36,576
Other accounts relating to securities 24,035 6,969 4,496
Accruals, provisions and other liabilities 51,722 14,997 9,675
Subordinated and undated term borrowings 11,435 3,316 2,139
Net worth 48,409 14,036 9,055
Total liabilities and net worth 1,035,921 300,350 193,774

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

2
9
6
-0

1
1

-7
–

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

296-011 Crédit Géneral, SA

8

Exhibit 3
Distribution of Daily Currency Movements

6/1/92 – 5/31/95

Daily % Price Change

Percentile DM/£ DM/$ $/£

Minimum -4.00 -2.89 -4.19
1% -1.44 -1.90 -2.25
5% -0.83 -1.19 -1.26

10% -0.61 -0.88 -0.82
25% -0.30 -0.38 -0.37
50% 0.00 -0.01 0.00
75% 0.25 0.37 0.34
90% 0.51 0.83 0.82
95% 0.74 1.32 1.11
99% 1.29 1.99 1.99

Maximum 2.30 3.17 3.47

Standard Deviation
(daily)

0.52 0.74 0.74

Standard Deviation
(annualized)

8.22 11.7 11.7

Correlation of Daily Price Movements

DM/£ v DM/$ DM/£ v $/£ $/£ v DM/$

0.30 0.29 -0.74

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

2
9
6
-0

1
1

-9
–

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

2
9
6
-0

1
1

-1
0
–

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

Crédit Géneral, SA 296-011

11

Exhibit 6
Calculation of realized and unrealized losses on the sterling/mark trade

Traded thus far
Guarantee to client

Remaining exposure
Guarantee to client
Total

Price
(DM/£)

2.2006
2.2356

2.2013
2.2356

£ Quantity
(millions)

300
300

Realized loss

700
700

Unrealized loss

Total loss

DM Value
(millions)

660.18
670.68
(10.50)

1,540.91
1,564.92

(24.01)

(34.51)

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

296-011 Crédit Géneral, SA

12

Exhibit 7
Theoretical Prices of One-month Currency Put Optionsa

(Spot currency price = 100)

Strike price

Volatility At the money Out of the money
(Annualized Std. Dev.) 100 98 96

8% 0.92 0.24 0.04
10% 1.15 0.41 0.10
12% 1.38 0.59 0.20
14% 1.61 0.79 0.32
16% 1.84 1.00 0.47

a Based on the Black-Scholes formula

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.

Crédit Géneral, SA 296-011

13

This document is authorized for use by Nanxin Chen, from 12/16/2021 to 03/30/2022
in the course: WI22 MGTF403-71035: Advanced Financial Risk Management (Valkanov) –MF22, University of California, San Diego

Any unauthorized use or reproduction of this document is strictly prohibited.