The New Market Wizards: Conversations with America's Top Traders (45 page)

At that time, I went away for a vacation, which was fortunate, because the Hunt buying started to push silver prices sharply higher. I’m sure that if I were around, I would have gotten out at the first opportunity of breaking even. By the time I came back, the silver calls were in the money [i.e., the market price had risen above the strike price].

Although I thought the market was going to continue to go up, I couldn’t stand the volatility. One day, I decided to go into the silver pit just to get a feel for what was going on. I promised myself that I would keep my hands in my pocket. At the time, silver was trading at $7.25. I decided to sell twenty-five contracts against my calls just to lock in some profits. Before I knew it, I had liquidated my entire position. By the time the calls expired, silver prices had gone up to about $8.50.

The 1979–80 silver market was one of the great bull markets of all time. [Silver soared from $5 per ounce to $50 per ounce in a little over a year.] Did you have any inkling of how high prices might go?

None whatsoever. In fact, even $10 per ounce seemed extremely far-fetched. I don’t know anybody who bought silver at relatively low prices and got out at over $20. The traders who bought silver at $3, $4, $5, and $6 did one of two things. Either, by the time silver got up to $7, $8, or $9 they got out, or they rode the position all the way up and all the way down. I’m sure there are exceptions, but I’ve never met one. I did, however, know traders that went short silver at $9 and $10 because the price seemed so ridiculously high and ended up riding the position until they had lost their entire net worth. That happened to some of the best professionals I knew in the silver market.

Would Hunt have succeeded if the exchange didn’t step in and change the rules by allowing trading for liquidation only, thereby averting a delivery squeeze?

The exchanges didn’t have to change the rules to prevent Hunt from taking delivery. According to the rules, the exchange has the power to step in and say, “Ok, you want silver, you can have your silver, but you’re going to have to spread out the delivery periods.” Or they can allow trading for liquidation only. If the exchanges had just stood aside and allowed a noneconomically driven demand for delivery, they would have been abrogating their responsibilities.

At the time that the Hunts were standing for delivery of April silver, the forward contracts were trading at huge discounts. The Hunts had no immediate economic need for delivery. If all they really wanted was ownership of the silver, they could have switched their April contracts into the discounted forward months, locking in a huge net saving and also freeing up their capital for use in the interim. Or they could have purchased silver coins in the free market at $35 per ounce when the April contract was at $50. When instead of these economically sensible alternatives they insist, “No, no, no, we want to take delivery of the silver in April,” it indicates that they’re playing a game. That’s not what these markets are here for. So I feel that the Hunts got exactly what they had coming to them.

A lot of innocent parties were hurt by the Hunt activity. For example, take a mine down in Peru whose cost of production is under $5 per ounce. When the price gets up to $15, the mine decides to lock in a huge profit by hedging their next two years’ worth of production in the silver futures market. This makes all the economic sense in the world. However, when the price keeps on going up to $20, $25, $30, $35, they have to keep putting up more and more variation margin on their short futures position. Eventually, they run out of money and are forced to liquidate their position, going broke in the process.

I know that CRT’s basic emphasis is option arbitrage, but I’m curious, do you do any directional trading?

Yes, for my own account. I’ve always believed that technical trading would work. From time to time I dabbled in it, and in each case it worked very well. However, I didn’t like the way directional trading distracted me all the time. I turned my ideas over to CRT staff members who had both an interest in technical trading and the appropriate skills. They followed up by developing technical trading systems based on these concepts and assuming the responsibility for the daily trade executions. I rarely look at the system anymore, except for an occasional glance at the account statements.

How long has the system been operational?

Five years.

How has it done?

The system has been profitable in four of the five years it has traded, with an average annual gain of 40 percent for the period as a whole.

Did you do anything different in the losing year?

Ironically, it started out as a fantastic year. About halfway through the year, the system was really smoking, so we started increasing the position size very quickly. At one point, we must have nearly tripled it. That was the only time we increased trading size rapidly. As it turned out, if we had held the trading size constant, the system would have had a winning year.

Is there any human judgment involved in this system, or is it totally mechanical?

Early on, there was about a six-month period when human judgment was employed.

And it was usually detrimental?

Unbelievably detrimental.

It’s amazing how often that’s true.

Everyone says that.

What is your view of fundamental analysis versus technical analysis?

Back in the late I970s, I once gave a talk on technical analysis at a seminar. At lunch I ended up sitting at the same table as Richard Dennis. I asked him what percentage of his trading was technical and what percentage was fundamental. He answered with scorn in his voice, “I use zero percent fundamental information.” The way he answered, I was sorry I had asked the question. He continued, “I don’t know how you escape the argument that all fundamental information is already in the market.”

I asked, “How do you escape the argument that all the technical information is already in the market?”

He said, “I never thought of that.” I admired him for that. He had a humility about him that I think explains a lot of his success.

My basic argument was that there are a number of technicians trading with the same information and the distribution of success is a matter of who uses that information better. Why shouldn’t it be the same with fundamentals? Just because all the information is in the market doesn’t mean that one trader can’t use it better than the next guy.

To a major extent, CRT’s prominence is due to options. I assume that CRT is, in fact, the world’s largest trader of options. How did someone without any mathematical training—you were a philosophy major, as I recall—get involved in the highly quantitative world of options?

I have never had a course in math beyond high school algebra. In that sense, I am not a quant. However, I feel math in an intuitive way that many quants don’t seem to. When I think about pricing an option, I may not know calculus, but in my mind I can draw a picture of how you would price an option that looks exactly like the theoretical pricing models in the textbooks.

When did you first get involved in trading options?

I did a little dabbling with stock options back in 1975–76 on the Chicago Board of Options Exchange, but I didn’t stay with it. I first got involved with options in a serious way with the initiation of trading in futures options. By the way, in 1975 I crammed the Black-Scholes formula into a TI-52 hand-held calculator, which was capable of giving me one option price in about thirteen seconds, after I hand-inserted all the other variables. It was pretty crude, but in the land of the blind, I was the guy with one eye.

When the market was in its embryonic stage, were the options seriously mispriced, and was your basic strategy aimed at taking advantage of these mispricings?

Absolutely. I remember my first day in the T-bond futures options pit, when the market had been trading for only several months. Someone asked me to make a market in a back month. [The back months have considerably less liquidity than nearby months.] Since it was my first day, I felt really out of step. I was too embarrassed not to make a market. So I gave him a fifty-point bid/ask spread on a hundred-lot order. I said, “Look, this is my first day, and I don’t really trade the back months. I’m sorry, but this is the best I can do.”

His jaw dropped and he said, “You’re making a fifty-tick market on a hundred-lot!” He couldn’t believe anybody was making that tight of a market.

The bid/ask spreads were that wide back then?

Yes, there was far less volume than now. A fifty- or hundred-lot was considered a really large order.

Once you put on a position because the market provided you with a large edge for taking the other side, I assume that you tried to hedge the position to eliminate the risk. However, when you went to implement an offsetting position, didn’t you face the same problem of wide bid/ask spreads?

The first thing you would normally do is hedge the option position by taking an opposite position in the outright market, which had much broader volume. Then the job becomes one of whittling down positions that can bite you, and there are so many ways to do it. For example, if on the original trade I sold a call, I would now be looking to buy other calls and could afford to become the best bidder in the pit.

Obviously, in those early years, the option market was highly inefficient. It’s pretty easy to see how, in that type of situation, you could put on positions that were well out of line, hedge the risk, and make lots of money. However, I’m sure that with the dramatic growth in volume over the years, the market has become much more efficient, and those types of trades no longer exist. What kind of concepts can be used in today’s market to make money?

Yes, the market has become much more competitive, but so have we. As long as we stay a notch better than our competition, there will still be good profit opportunities.

So there are still mispricings in the market?

Absolutely. There will always be mispricings in the market. The notion that the market will trade at its precise theoretical fair value implies that someone will hold it there without getting paid. Why should anyone do that? The service of making a market, like any other labor, is one that people are not going to want to do for free, anymore than they would want to wait on tables for free. There’s work involved. There’s risk involved. The market has to pay someone to do it. It’s only a question of how much.

Is that payment a bid/ask spread?

Yes. If that edge did not exist, when someone walked in with a large buy or sell order, who would be there to take the other side?

The following section deals with theoretical questions related to options. Explanations for the layperson are provided within the bracketed portions of text.

What do you think are some of the conceptual flaws in standard option pricing models?

I don’t know how I can answer that question without disclosing information we don’t want to talk about.

Well, let me take the initiative. For example, is one of the flaws in the standard models that they don’t give enough probability weight to extreme price moves? In other words, actual price distributions have fatter tails than are implied by normal probability curves. Therefore, people using the standard models might then be inclined to sell out-of-the-money options at lower prices than would be war-ranted by the way markets really work.

Yes, that’s a flaw in the standard Black-Scholes model. When we first started, even our biggest competitors didn’t seem to have that figured out. and a lot of our profits came through that crack. By now, however, all the serious players have figured it out, and I assume that many commercially available models allow for it.

I would add, however, that it’s one thing to recognize that the tails are fatter than normal, and it’s another to know where to go from there. For example. do you simply fit your distribution to your empirical observations, and price options accordingly? That path has some serious problems. Do you take into account your hedging strategy? Are there other variables that none of the available models allow for? And, if there are, would their inclusion introduce so much complexity into the model as to make its application unweildly?

In other words, knowing that the distribution isn’t log-normal only opens a can of worms. Frankly, though, I still can’t understand why back then the competition hadn’t at least gotten the lid off the can.