Serving the Quantitative Finance Community

 
User avatar
Collector
Topic Author
Posts: 2572
Joined: August 21st, 2001, 12:37 pm
Contact:

How to fall

September 6th, 2001, 3:51 am

In the back wash of the collapse of Barings Banks, LTCM among others there have never been more focus on stress testing and worst case scenarios. Working for a bank or hedge fund you probably spend most of your time and a lot of the companys money to make sure "you" never will end up in a collapse situation.What about worst-cases in your personal life. I am not speaking about your bank account but:How to survive a fall onto subway tracksHow to escape from a car hanging over the edge of a clipHow to survive a airplane crashHow to stop a runaway horseHow to stop a car with no breaksYes very unlikely, just as unlikely as a financial institution collapse, but still it happens all the time. What you need is worst-case scenario skills. Today I just bought a book giving an introduction to such skills: "The Worst-Case Scenario Survival Handbook, by Piven and Borgenicht. The book is strongly recommended (I am not getting any commissions.).For example personally I am well prepared for falling (falling down a stair, falling when I am out rollerblading, falling when I am skiing down a big mountain.). I used to train Judo for many years(many years ago). One of the most important things in judo is to learn how to fall softly. Even now many years later I still "instinctively" know how to fall at any time. It has saved me from breaking my arms and legs several times (only broke my arm once.)In finance I think there have been to much focusing of how to avoid falling, and not to much on how to fall? If you fall you will always fall when you least expect it. That is when you need to know how to fall, you will still get some scratches, but in one-two-three you will be up at your feet again, ready to fight.(Sorry if this is to much, I think I may be had a drink too much this evening)You can never remove the tail probabilities of a worst-case scenario happening to you. (I loved Rudi Bogni’s story: "The Riskless Society"), but with more worst-case training you will increase the probability of getting away with only a few scratches. If you ever should visit Greenwich, CT, US I would be happy to show you how to fall on asphalt (not joking).
 
User avatar
Paul
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

How to fall

September 6th, 2001, 7:16 am

We bought that book for my son's birthday, he loved it!How to fall?Plan A: Panic. This is the Metalgesellschaft approach. Throw away all positions, even if they have positive value. Can be costly. How about the opposite?Plan B: The Ostrich. Hide your head in the sand and hope the problem goes away. Tried by Barings. Lose double the amount you should have lost. Not good. But...Plan C: Get a friend to throw themselves under you, preferably someone big and soft. This is the best approach, tested effectively by LTCM. You live to fight (or trade) another day having "learned you lesson." P
 
User avatar
DB

How to fall

September 6th, 2001, 8:45 am

Has anyone heard of the traders', gamblers' or possibly beginners' rule"When you're in trouble, just double"Any stories?
 
User avatar
Paul
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

How to fall

September 6th, 2001, 8:58 am

It wins every time at roulette...doesn't it?!P
 
User avatar
Aaron
Posts: 4
Joined: July 23rd, 2001, 3:46 pm

How to fall

September 6th, 2001, 11:48 am

Let me quote back: "your first loss is your least loss." In almost every financial disaster there is a series of doubling transactions. Procter & Gamble issued fixed rate debt and rates went down. Banker's Trust said, we'll give you the floating rate you wished you'd chosen, in return for a you agreeing to pay us a large amount of money with small probability. P&G won the first bet (also typical of disaster stories) but got addicted to that kind of deal. Once they started losing, they had to make them bigger and bigger until they were out $200 million betting on cross-currencies and interest rates unrelated to their business.The problem is clearly employees. Once they've done enough to get fired, there is no cost to trying to dig themselves out by doubling the bets until they get discovered. Assuming a reasonably clever employee working for a typically obtuse institution can get 10 doubles in, there are 1,023 people still employed and trusted for every Joseph Jett, Nick Leeson, et al. With demonstrated luck and penchants for rule-breaking and risk-taking, these people are no doubt highly placed. Of course, this group will insist that only idiots be hired for risk management so that no one will ask embarassing questions about past transaction. That explains a lot.
 
User avatar
DB

How to fall

September 6th, 2001, 1:27 pm

A related paradoxon is the stop-win strategy after suffering losses.Imagine a punter gets an allocation in an IPO to make a quick buck. Unfortunately it heads southwards, so he keeps adding the stock on its way down.Let's say the stock starts recovering and one monitors the average price paid against the spot.I bet once it is barely in the money the position will be liquidated, with maybe a tiny gain and a huge sigh of relief.So overall desperate risk-taking when it comes to losses, but risk-aversion when it comes to gains.I'm sure Henriette could provide an analysis.
 
User avatar
Paul
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

How to fall

September 6th, 2001, 1:48 pm

Start with $1MM. a) Lose it all..."Jeez, you lost 100%! You're fired!"b) Double it..."You only made how much, $1MM. You're fired!"On the downside it's relative amounts that matter, when you win it's the absolute amount.P
 
User avatar
Hamilton
Posts: 1
Joined: July 23rd, 2001, 6:25 pm

How to fall

September 7th, 2001, 3:20 am

Plan B: The Ostrich. Hide your head in the sand and hope the problem goes awayAlternatively, put a bag over your head and hope the problem goes away.On this thread, I notice 1 other guy with a bag over his head.
 
User avatar
Aaron
Posts: 4
Joined: July 23rd, 2001, 3:46 pm

How to fall

September 7th, 2001, 12:18 pm

Hmmm. . .So say an employee is entrusted with n different tasks (either simultaneously or consecutively) with capital investments C1, C2, . . .Cn. The outcome of each is O1, O2,. . .On. You suggest that the evaluation function is the sum of Oi for Oi>0 and a*Oi/Ci for Oi<0.To the extent money can be moved among outcomes, the employee should move money from the tasks with Oi<0 and Ci>a, first to tasks with Oi<0 and Ci<a, then to tasks with Oi>0. If there are still tasks with Oi<0 and Ci<a, the employee should move money from tasks with Oi>0.One example of this is found in mutual funds. Managers would overpay $1,000,000 in brokerage commissions for their biggest fund in order to get $500,000 in deliberately underpriced IPO for their smallest fund. The $10 billion fund lost 0.01% in annual performance, the $10 million fund gained 5%, the broker made $500,000. It was particularly attractive if the big fund were stuck with the kind of mediocre returns that would not attract new money, but did not drive away existing money.This passed for honesty among fund managers, by the way, the dishonest ones took the $500,000 personally. An older, and less efficient, variant of this is to buy 10,000 shares of stock X for $25 for the small fund; then buy 1,000,000 shares in the last ten minutes of trading for the quarter in the big fund, driving the price up to $35. The small fund adds $100,000 (1%) to its quarterly return, the big fund spends $10,000,000 more than its stock is worth, but this is only 0.1% and causes no performance penalty until the end of the next quarter.From a personal standpoint, it suggests writing off big unsuccessful tasks (become a movie star, fly to Mars, cure cancer) and devote the energy to minor unsuccessful tasks (get haircut, clean desk, check phone bill) until they are done; then work on things you're actually succeeding at.