A Derivative Life

The Way of an Entrepreneur

by J. Orlin Grabbe

Packing Heat

When I was in grad school, and finally found a few thesis topics sufficiently interesting that I could get inspired about and could also get approval to write on, I came across a partial differential equation that intrigued and mystified me, in a context (utility theory) I wasn't interested in, but which seemed to hold some insight or revelation about something I was writing about (Levy stable random variables). And I kept looking for references that would maybe have details that I could get a clue about what this equation meant. Now, in economics departments in those days almost no one understood option pricing theory—in fact when I first taught in the Finance Department of the Wharton School of the University of Pennsylvania, I was the only one in the department that knew anything about it. (This changed in a few years).

But I was in Harvard graduate school, working on my thesis. One day I was reading through a volume of Paul Samuelson's collected papers that I had purchased—reading the easy ones; perusing the difficult ones I found interesting; and scanning the difficult ones that didn't grab me to see if there might be anything cute to steal (mark for future reference). I got to the option-pricing section, pre-and-post-Black-Scholes approaches. There I saw the equation that mystified me, or at least the basis of it. It was the heat equation from physics (though not raised to a fractional power, which was my mystery equation). So I read all the option articles, and was amazed that I understood a good bit, yet felt stupid that this material relating to my seemingly futile search to understand my mystery equation had been there all along, and I hadn't known anything about it, and that no one in economics courses I had taken had bothered to mention it. That was because none of the professors understood any of this stuff—stochastic calculus, and the context of option-pricing theory—so they ignored it. Just a few top people had done work, and no one else understood what was supposed to be going on, and so they didn't embarrass themselves by trying to talk about it.

I was talking to my friend Chris Piros at the National Bureau of Economic Research, and he said Robert Merton (one of the few option guys) was doing his class in Advanced Capital Markets at MIT which he only does every 2 years. Yeah? I said, not really caring that much. Harvard and MIT had an agreement between the economics departments so graduate students could take classes in either place. Only gradually did it become clear that stochastic calculus, and option pricing theory, and solution to partial differential equations (i.e. theoretical option equations) were some of the main components of the course. So I went down to sit in with Chris. Merton mostly bullshitted in class, but it was still interesting to see what his attitude and approach were, but there was also a set of class notes to study (which went through a lot of technical material and substituted for the lectures), and a xeroxed bulk pack of every option-pricing article written in the 1970s. The latter was pure gold; in grad school you quickly learned that you would spend twice as much time trying to get hold of an assigned article as you would have to either read or study it. So that last semester while finishing my thesis I read through all the articles and went from being totally ignorant of the area to some non- casual acquaintance with everything that had been done to that point from the point of view of academic theory.


Flash forward. I am at Wharton doing my International Finance class for the first time. I think the entire field should be redone, so started creating my own reorganization and outline of material, and since I knew options in connection with forward contracts had begun to appear in the foreign exchange market, I used this as an excuse to put options on the course outline. I had no tolerance for professors who ignored what was happening in the markets, or for practitioners who pretended theory was irrelevant since I had already myself (to some distress from my colleagues—"who does this arrogant Assistant Professor think he is?") thrown all the irrelevant stuff out. And I would read Euromoney magazine, and go up to New York, say to Morgan Guaranty, to chat with FX traders about what was happening in the market and have lunch in their oak-paneled dining room, and try to figure out (first for myself) how to translate between the different languages of discourse, and bridge the gap between the academic sphere and the "real world", and communicate in both directions so I could sell the academic material better to MBA students and also forge bridges to market environments which I found much more entertaining than academic seminars. I quickly discovered that none of the foreign exchange people knew anything about options; meanwhile, none of the stock options people had a clue what was different about the foreign exchange from the stock market. And they both thought the other camp was full of shit. And so I'm trying to figure out how to exploit the fact that I can bridge the gap, and maybe also get a good paper published in the process, and perhaps attract some attention in the markets in case I decide someday to try to make the big money.

So I give my first option lecture, and this student comes up and says did I know that the Philadelphia Stock Exchange was going to start trading FX options, and gave me two legal brochures that explained the details of the proposed contract specifications, which he had but had decided he had no use for. Boom. I'm trying to create the rationale to do what I want to do, and reality quickly adjusts itself to meet my requirements.

Meanwhile in the Finance Dept. someone is trying to get a big chunk of grant money, several million dollars, and there is money (a $1000 grant) available to fellow professors who write individual sections of the overall proposal. My topic isn't terribly sexy, but $1000 dollars is $1000 dollars, and I try to figure out something that will pass muster and will let me spend time educating myself in something I want to spend time on. I forget the research question but I managed to contort out an acceptable (to me) answer, and proposed creating a foreign currency option pricing model using assumptions more appropriate to the foreign exchange market, and using fractal probability models. Tall order, but it was just a proposal and I got paid. Then Wharton, MIT, Brookings Institute, and Princeton emerged as the finalists, and so now I actually had to write a paper, but I'm going to get paid $2000 for this. Damn, now I had to do what I proposed, but maybe I can pull it off, and I definitely wanted that $2000. I spent almost the entire allocated time going nowhere with my probability assumptions, until it became obvious I would have nothing at all to turn in. So I changed to the usual Brownian motion (Gaussian) assumptions, and tried to deal with two stochastic interest rates and some other things, and saw this wasn't going to be a piece of cake either, but at least I made some progress, but to pull it all off, I had to solve my own partial differential equation.

The Watergate Burglars

I knew approximately what the answer had to be and tried the time-honored method of guessing the solution, and then differentiating it to see if it met the conditions of the equation and also the boundary conditions. Somehow it would never quite work out. And so one weekend I went down to Washington D.C. with my girlfriend (I think to see an ex-classmate of mine, who was in the Japanese Ministry of Finance, and who was assigned as the assistant to the Japanese Director of the IMF). And we stayed in the Howard Johnson's there, where I liked to stay, across the street from the Watergate, and the Kennedy Center behind the Watergate; and a short walk in the other direction put you in Georgetown with all the bars and restaurants, and a good bookstore; and it was the Howard Johnson's from where the Watergate burglars (G. Gordon Liddy, Howard Hunt, James McCord and the others) had set up surveillance; so I liked to stay there and it was very pleasant but cheap; and this time as the clerk was assigning us a room I asked for a citation of which rooms were available and there was one on the end on the 7th floor which I immediately requested, because if memory served me this was one of the Watergate burglar rooms (but if not next door). And I had taken my equations with me, because I had almost run out of time, and I kept thinking the elusive solution was just around the corner. So we had made a tour of the Kennedy Center and the Washington Mall in the afternoon, and had gone back to the room to rest and nap before walking over to Georgetown to eat and drink that evening, and I was sitting up in bed with the curtains closed and the bed light on, and my girlfriend asleep, and going over my equations again and trying to see a different form that would work and suddenly something new struck me, what I needed to change, and a shock of adrenaline went through me, and I started to tremble as I wrote on my notepad, and I was afraid to differentiate it and have disappointment slap me in the face again, but I had to do it. And it worked. And I went over it all again, and did it all again, and it still worked. And we went out and had a fine time that evening in Georgetown.

And in the morning, I did it all again, and yes, I still had the solution, and I was praising Jesus for the Watergate burglars. And things started to happen quickly. I was at an MBA party and met a friend of a Wharton MBA student who had just been assigned to the FX option floor of the PHLX stock exchange (he traded all of 2 contracts the first month), and we talked about FX options and I mentioned my paper. And he (Mark) wanted a copy, but after I gave it to him, he confessed he had no idea how any of this was useful to him. And I was pissed. I was supposed to be the professor, the theoretician, and he was supposed to be the practical MBA type who took the theory and went out and created the practical application, and he wasn't doing his damned job! So I decided I wasn't going to put up with this nonsense any longer, and went over to Radio Shack and bought a TRS80 Model 3, with two 128-k disk drives, with a big 48k ram, and a Fortran compiler, and took it home, and started reading the manual and found out how to turn it on, and for the first time ever set at the keyboard of a personal computer, and read the manual, then figured out how to install the Fortran compiler, learned about the mysteries of compiling and linking (before that I had always just turned in punch cards, and the anointed priests in the computer center did something mysterious and later the computer god would barf up a printout). I had no interest in computer science because to me a session of punching data into cards, followed by a trip to the computer center, was about as exciting as spending a day digging postholes on the ranch, followed by a day of going to church. So after my first and only computer course in which I spent 10 weeks learning some nonsense, and a little Basic and Fortran, and how IBM expected 20 lines of code a day, I knew I had no career interest in this area, and moreover had no desire to (and never did) take another computer course in my life. So once I actually figured out how to compile and link a Fortran program, I took my option paper, and started writing a pricing equation, and had to look for a good approximation routine for the normal distribution, and included that in the program. And then when I finally got sensible numbers I expanded the program to do a little print out of option prices across strikes and maturities—one that looked pretty much like the screen on the PHLX FX option floor.

I purchased my Trash 80 on the morning of the first day, conquered the mysteries of the personal computer and finished my program by the end of the second day. The next day I went to see Mark (the derelict FX option guy), and stuck the printout in his face, and said "Here, this is what my paper says." And he was stunned. And was speechless and then said maybe he should get a PC and do this too and I later gave him the source code. And he totally forgot, because he didn't realize it, to ask me again how he could use the numbers, which was lucky because I still wasn't quite sure, but our whole level of communication had taken a new turn.

Publish in the Parish

And now that I had numbers, I added sample numbers to my theoretical paper, and submitted it to a journal and it was accepted in the same issue as another FX option- pricing paper — and these two articles became the basic academic articles on the subject. Meanwhile, someone mentioned the paper in some context, and I got several calls from banks wanting a copy, and was thrilled for about two hours until I quickly realized I wasn't going to get anything back except maybe a little prestige; meanwhile the morons in the banking world were being paid big bucks to cannibalize other people's stuff. And I became a little surly, and then got a call from someone in Drexel, Burnham, Lambert's Philadelphia office who said he had a problem with FX options, and I agreed to meet him next day at his office, but went there determined that whatever this jerk wanted, he wasn't going to get a damned thing without paying for it. And he told me he was trying to explain to corporate treasurers how to use FX options, and educate them so he could sell them broking services, and was looking for something that would explain how to use FX options—and in particular he was trying to sell something to Sun Oil company who had a $600 million Canadian dollar exposure, and how FX options could be part of their plan.

And I was cautious, trying to figure out what this guy (John) wanted me to do, and it gradually became clear he had no clue what should be done—that's why he had called me. And since I had no idea what should be done either, I asked more questions trying to figure what the corporate treasurer needed, and what the problem was. And he showed me a five-page paper that he said some other firm had done, but it had made no sense to, nor had communicated with, the treasurer. I sat there and eagerly started reading this sample, and realized instantly I could do better than this, that this paper was a piece of shit, and that what the treasurer needed was something with reference to real numbers, and real contracts, and real dollar profit-loss calculations, and half the paper I was reading read like it had been copied out of a textbook somewhere. In fact, in my mind I concluded that John had written this, and that he wanted to buy something a lot better. (Only later did I find out it was Goldman Sachs that had written the piece of shit that I was expected to replace.) Clearly John thought I was an expert—he didn't know I had never done this before—so I told John I charged $100 an hour (I had read in the Wall Street Journal that that was the going rate), intending to figure out what to do, and then to only charge him the number of hours that I would have spent if I had immediately sat down and done what I already knew how to do. And I went home and tried to find a better prototype of what needed to be done, and remembered I had read something possibly relevant in my option packet back in my collection of 70s options articles, and sure enough found a paper that I decided to imitate in part. I had my trusty Trash 80 to crunch theoretical FX options prices, and knew all the relevant contract data and brokerage fees, all of which I included in the calculations, and looked up current Canadian dollar prices and created an initial section on currency risk exposure based on Canadian Dollar volatility, and then churned out minimum, maximum, and expected dollar values, using specific real option contracts and strikes that were available on the PHLX, and included transactions costs. Real dollars to make a treasurer ecstatic. It was the weekend and I spent 10 to 12 hours the first day. So by the time I got started I realized there could have been no way I could have done it in a morning (4 hours), so I decided to charge 6 hours. By the end of the 10 to 12 hours I spent that day, I had decided to maybe charge 8 hours, a whole day. But then I spent a good bit of the next day, finishing the numbers and checking them, and writing up the paper—written and printed out also on my trusty Trash 80 computer and printer—and I realized there never are any 8 hour days, they are always 10 hours, and moreover what was going to make me feel satisfied was $1000 for the work. So I decided to charge John 10 hours.

I took the finished paper over to his office, and he was out, so I left it on his desk in an envelope with his name on it. He called me a few hours later. He was ecstatic. But he quickly asked "how many hours . . ." He sounded worried. "Ten" I said promptly. "Okay," he said, still sounding worried. Maybe he was hoping for fewer, but I didn't care at that point. I had assumed Drexel was paying, but John wrote me a personal check. And then he ran out to Sun Oil, and the treasurer told him this was the first time he had understood any of this FX option stuff. And John ran around everywhere waving his paper (the one I had written), trying to drum up business, and thereby also spreading my name throughout the small FX option corporate, banking, and investment banking world.

FX Systems

I sent one of my students, Farid, to Merrill Lynch International Bank to work on a project. They wanted a handy FX option calculator, which he did for them in Visual Basic, using my Fortran and Basic pricing code as a starting point. This created the thought that this could be a starting point to make money on the basis of my FX option expertise. It could be done on a PC, and it could be started small and pyramided into something bigger. Farid and I ended up living in the same building in Society Hill Towers, down near Penn's Landing in Philadelphia. He was working as a computer consultant for a British company. We would occasionally run into each other, and say we should start a company. I got the form from the Pennsylvania Secretary of State, filled it out with my apartment as the business address, and sent in the $75, and we formed FX Systems. It was kind of fun, a game to tell everyone about. We decided we would exploit the new PC wave, and thats how we would do it—we wouldn't have to be heavily capitalized, and we could start it all from home. We each put in $500 and opened a corporate FX Systems account at a local bank. I still had my Trash 80, but Farid had just purchased one of the new IBM ATs. We had technology. We looked around as to what language to use. It had to be C—that was the only one going to make it in the business world. I didn't know any C. Farid had only used some Quick C, but not much. Never mind. Business dictated C, and anyway that part was more Farid's problem at this point. I added an American FX option pricing routine to the European one—no one seemed to have this. So a simple calculator would be interesting for that reason.

We had to have a math coprocessor to run the calculator—otherwise it was too slow. But even though banks were buying the IBM ATs, and most of them had ATs, none of them had math coprocessors, so we had to take our own computer to the demo. We got a large trunk with casters on one end which would hold the computer and monitor and keyboard, etc., with form-fitting foam cut out for just that purpose. (It was a company that built such things for the rock music business and Farid had them build an AT container.) Our first marketing run we scheduled 7 to 10 interviews in Manhattan. The first one was a breakfast meeting in a hotel on the Upper East Side. We just discussed our company, and what we hoped to do, with some people who were friends of the aforementioned John. Then we took our computer trunk downtown to Wall Street. Most cabs would not take us—they had to have a large car trunk (boot) so that our computer trunk would fit inside. Finally we got to Wall Street; it's getting warm and lifting and pushing the trunk around in a suit is uncomfortable, the air is hazy, and start-stop traffic is bad. We go into the building where the first bank is supposed to be housed (I had never been there and had to looked it up—it was some foreign bank). Well, the bank had moved 6 months earlier— up to Park Avenue. So we went back on the crowded sidewalk pushing our trunk, and spend forever finding another cab, and then took a nauseating Manhattan cab ride back uptown. We come rolling our trunk into the lobby, saying we have an appointment. Well, we are not allowed to use the front elevator, and we have to go around to the back and use the freight elevator. (And that was always true: we always had to use the freight elevator.) And when we get to our floor, we have to take a circuitous route to the reception. The guy we talk to is a jerk. He wants to see what we are selling, and has no interest in the calculator. We smoothly try to say it is just a demo, that what we are about is creating tailor-made software. We want to hear about what he needs done, what problems he has. He shows the system he has, and what is wrong with a key page, and what he wants done, and says the company selling the system won't change it for him.

Well, we can't change it either, since it's not our system, and we don't have an alternative system to offer (we suggest he contract us to build one to his specification— but we have a problem: no software references, no customers, and everyone knows beginning computer companies disappear after an average lifetime of 3 to 6 months, and who wants to supervise a long and expensive software project just to get a single page fixed?). We push our trunk back to the freight elevator and hit the sidewalk. We have to cross the street to go back downtown. It has warmed up to 90 degrees, and the casters on the trunk stick in the soft tar as we cross the intersection. We'll have to be more careful. We finally get a cab and go back down to Wall Street. Farid goes somewhere to give a demo by himself, and I meet John at Brown Brothers Harriman where he is going to introduce us, supposedly. I gradually realize that the guy we talk to is not in the market for anything, and that John is using myself and Farid as props to try to pyramid an image for himself to generate some business for himself. Then we meet Farid and go over to Morgan Guaranty. This is more pleasant. They like looking at the calculator. But they don't want the calculator. They want to know how much we want for the option models. Farid and I look at each other. We don't want to sell the models. There's no money in that: you are giving away your advantage. What you want to do is put the models in software and sell the same software over and over to different people. We say we don't see any money in selling models. But we agree to think about it. Nevertheless, the meeting is pleasant, except I am getting extremely sick as a result of the smog and exhaust and motion of the cab trips up and down Manhattan. I say I'm not feeling well and need to go outside for some air. I hurry out because I think I'm about to throw up in the bank. Out on the street I realize that I have to lay down. I look around and see Trinity Church at the end of Wall Street. I go there, walk through the church and out back which gives entrance to the tiny park and there are two park benches there. I stretch out on one of the benches, a street person in a suit. Slowly, slowly over time I start to feel better. After about an hour John and Farid find me. We go one more place and call it a day. I catch the train back to Philadelphia.

We do the computer trunk, freight elevator, routine numerous times again. Each time as we finish and push the trunk back toward the freight elevator, I can feel the people snickering at us behind our backs. But we have to sell something to somebody, convince them to buy something from us, despite the fact we have no customers, no track record, and no credibility. Yet slowly we are learning the market, learning it better than anyone else, learning what people think they wanted, say they wanted, think their problems are, translated into what we knew they needed. Meanwhile I am still at Wharton, with a reputation in FX option pricing area, and Farid is still a computer consultant. So we get in the door. But we don't sell anything.

A Sop to Cerberus

The lawyers at the Philadelphia stock exchange talk to me about writing CFTC filings for foreign currency futures contracts that the Philadelphia Board of Trade is planning to offer (an adjunct to the FX cash options). I look at the CFTC list of requirements, and sample copies of submitted filings of other contracts that were approved (acquired through the freedom of information act) and we go to lunch. They asked how much I charged. I say $1000 per day. That doesn't bother them. They asked for an estimate of the number of days; clearly they were interested in total cost. I said I wasn't sure but I would agree to cap it at 10 days, since it would take a minimum of 10 days, or two days per filing (among other things I would have to write up a description of each market, along with 5 years worth of data analysis for that market). More likely it would actually take 3 ½ to 4 weeks, but I would only charge them for 10 days, or $10,000 total. The lawyers ducked into their plates with that statement—clearly they were hoping to get it cheaper. But they sold it to the board; the board thought the filings were just bureaucratic paper crap, and didn't want to pay more than a couple of thousand, but the lawyers were worried saying it was much more important than that (and they didn't want to try to do it themselves, or to begin the search all over again to find someone to do it). Finally the board agreed. I started the project, and then they called to ask if I could deliver the one on the ECU (now the euro) first. I had been leaving that one to last, since it was different in many ways from the others, but bumped that one to priority. It turned out the CME was suing the Philadelphia Board of Trade saying they had no right to trade the ECU. But no one in Philadelphia or Chicago really knew what an ECU was (was it an index or a currency?) or had a clue as to how the European Monetary System operated (except for some superficial details). So suddenly my ECU paper was in heavy demand as the basis for whatever was going to be argued in the lawsuit. The PBOT won the lawsuit against the mighty CME, and my name was made in Philadelphia, and no one thought the $10,000 was a big deal any more. I took $6,000 of the $10,000 and bought my own IBM AT, with a 20 meg hard drive and 640k ram. Now FX Systems had two real computers. We should have been hot stuff. But we spent the next year without selling any software. We met various diverse and strange characters, and tried a hundred different approaches.

I gave a speech at the PHLX option conference and created an uproar by saying that of the 3 vendors of software in the next room, I was familiar with 2 of them—and of those two, one was getting the numbers right and one wasn't. (I didn't explain which was which.)

The Good Book

The first edition of my book International Financial Markets came out; it quickly became the best-selling MBA textbook , and appeared in bank training programs everywhere, and in central banks around the world. Bill Lipschutz at Salomon Brothers, and who my ex-student Andy Krieger traded with (Krieger became famous a few years later when he made $300 million one year trading FX options at Banker's Trust), was on the cover of my book. Lipschutz was the largest off-floor trader at the PHLX, and he and Krieger were the largest FX options traders in the over-the-counter bank market. Lipschutz's wife bought 50 copies of my book to give to friends.

Finally, over a year after FX Systems had started, Drexel, Burnham, Lambert in Philadelphia became interested in talking to us (they were specialists in 2 or 3 of the five currencies on the floor). They ran their risk analysis of their FX option book on a Lotus spread sheet, and put in everything by hand, and ran pricing sheets from a service they hated, and would have to spend 3 hours after the close of the market day doing the paperwork. They wanted to replace their Lotus spreadsheet with a real program, and have the ability to run their own pricing sheets, and to be able to go home at a decent hour. They got an okay from New York, but they didn't have much of a budget to spend. And so we got our first software project. We agreed on $10,000, cheap—but we would own the software which we could sell to others, and we could use Drexel as a reference. Farid and I rented offices across the street from the PHLX. Wharton was pissed that I was never seen except in class and scheduled office hours.

We finished the Drexel project and I decided to resign from Wharton. Drexel was the FX specialist in several currencies, and suddenly everyone could see our screens on Drexel's computers over on the specialist counter. They wanted to have what Drexel had, which meant they had to buy from us. Susquehanna was the largest trader on the PHLX floor overall, in stock and currency, and Jeff Yass and the gang came over for a demonstration, and wanted what we had, but wanted some additional features which we agreed to make. (Naturally we would keep the enhanced software.) We sold them $37,000 worth of stuff. Then Bank of America-Barry Teague, the other main specialist came over. They didn't want the general risk analysis—they wanted to keep their own system—but they wanted to buy our pricing sheet module: they needed it to keep from being ripped off by Drexel and Susquehanna. And so on down the line; we had broken the barrier and had things to sell. We sold to a dozen more floor traders and floor trading operations.

Merrill Lynch in New York wanted to talk about offering me a position for $250,000 a year or so to write their international bond letter (their interest was a consequence of my textbook). I would have jumped at it a year earlier, but now I saw it as another academic type job, which I was trying to break away from. I turned them down.

Then the PHLX wanted us to build a specialist quotation system for sending out indicative price quotes to Reuters. The specialists didn't want to reveal their pricing sheet, but the quotes could be fudged to within a reasonable ballpark. We built the system, which included a communication module and they were happy. We did a haircut module for Wagner Stott. We agreed to build several futures exchange-based systems for a trading company in gold, silver, eurodollars, etc., but screwed it up, and eventually ceased any more work after two systems (which they did use) , and gave part of the money back.

All the time I would program any pricing functions in C and Farid would take care of the rest, and I would check the program. Our customers would come into the office and ask imponderable things like, "I was trading today, and by my pricing sheet I was only willing to bid 37, but CRT picked it up for 38. Why?" Or, "why are your long-dated British option prices different from TOG's?"

The GAF Gaff

But the exchange sales slowed, and we tried to take up the slack with GAF corporation, with whom we had gotten a contract to build a corporate treasurer's risk management system. GAF was much in the news those days, the chairman Samuel J. Heyman having attempted a couple of corporate takeovers, both of which failed, but GAF walked away with millions of dollars profit each time. I wasn't happy with our performance on getting the system built, but on the other hand we kept getting jerked around by them. We were doing it for the corporate treasurer, but had to respond to all sorts of idiots, such as the head of the MIS department. He would say, "Well, here is a boxcar program, and I understand what it does, but I have no idea of what your system is supposed to do." And I would say, "well on this page you have this set of money-market instruments marked to market, and the amount of accumulated interest, and a hedge ratio calculated as the number of eurodollar futures contracts." And he would look at me blankly, since he had no idea what I was talking about, and then say, "Well, I still don't know what your program does. Can I look at your flow charts?" And I would look at Farid, and think mentally to myself, "It's up to you to tell him there are no f--king flow charts."

And then my ex-student Andy Krieger asked me to come trade FX with him at Banker's Trust, and I seriously thought about it, but dropped that mostly because I didn't like Banker's Trust's attitude.

So to take up some slack (we had pretty much saturated the PHLX) I did a consulting job on DM, S&P, and gold trading strategies with a New York firm. They didn't just want a report with data analysis; they wanted me to meet with them every two weeks, and discuss what I had done, and answer their questions. What they were buying was not a final report, but a prototype of how to do this themselves.

No Longer Faking It

At FX Systems, it seemed clear that we would have to change our focus from exchange systems and more to consulting, or else get into over-the-counter banking FX risk management systems, which were dominated by a company called Devon. Or we could get our corporate treasurer system built, and that wasn't going well because of all the interference and wasted energy. Then Drexel, Burnham New York decided that the PHLX floor option system should be abandoned, so that the PHLX books could be integrated with the over-the-counter books, and it would all be run through the Devon system, and pricing sheets would be generated by Devon also. The PHLX floor operation revolted, saying they had to have our pricing sheets, though they didn't care about the rest. So Drexel New York wanted us to make our pricing modules available through Devon, which would be available only to Drexel. So we agreed, charging them $90,000 for about 3 weeks work. We ended the year cash rich, and took out a 5-year lease on an office space catty-corner across the street. We hired two full time programmers in addition to our two part-time programmers, and a receptionist, and I hired an assistant to help check prices. We leased our own duplicating machine, and had our own kitchen and conference room. Suddenly, for the first time, we seemed like a company. It wasn't a game or a joke anymore.

For "security" I was now supplying to our programmers all the pricing functions in compiled form. There was an FX math module, a bond math module, a money market math module, a stock math module; later there would be a commodity math module, a barrier option module, etc. To assist the programmers and for my own checking purposes, I would also program some sort of self-standing calculator with screen inputs to generate the relevant outputs, and supply the compiled math module along with the calculator executable and calculator source to the programmer for his own reference in the system he was working on. Then I would also have to check the numbers coming out of the system in progress.

I have checked a lot of such systems-in-progress where I have programmed the math functions, explained to the programmer exactly which variables to pass, provided all the necessary interpolation functions on interests, volatilities, and what not. The programmer is very intelligent (say a Chinese graduate engineering student who got disillusioned with academia and decided to make some money instead), is well intentioned and works hard. In that context, if I got a system-produced report with 50 pricing numbers on it, say involving different types of options—on spot, forwards, futures, physicals, LME contracts, forward start options, barrier options, lookback options, etc.—and also spot, forward, future, LME, swap and other contracts, involving different currencies with different settlement conventions and different year bases; and all converted into a base currency for valuation purposes . . . well, what would I expect to get? Once upon a time I might have expected to earn my pay after having already provided all the calculation functions myself in C code (having previously done the math and the finance) by now finding the 2 numbers in the implementation that were wrong, and feeling I had done a good job. That was years ago. My realistic expectation now is I would find in the neighborhood of all 50 separate output numbers wrong. Getting things right is a tiresome, laborious, uphill job.

And always, always, ALWAYS, I would have to justify the numbers to the customers, who might bitch because the numbers didn't look like some other system he saw, or said it conflicted with market prices. These were all people in the market: we were dealing with RAW REALITY. Or, on a more theoretical level, I would get hired to check someone else's prices. I remember having to tell the MIT mathematicians and quants at Shearson Lehman that their option prices were wrong and, moreover (since this was natural part of the consulting job) figure out what was probably the programming or math error from the output numbers alone. So I gave my results to them in front of their boss who had hired me, and they excitedly denied it, and pulled out their equations, and source code and pointed excitedly at it, and so I had to take their output numbers which they couldn't deny and set there and calculate financial relationships in front of them and their boss and show there was a mistake, and then take those same numbers and back-solve to what I said the error was (and, yes, I admit it, what I claimed the error was seemed totally weird and otherwise implausible), until they had to shut up and look at each other, and their boss concluded the meeting. Still it took them 2 days to find their mistake, and to acknowledge I was exactly right.

And in the meantime no one could any longer pull this shit on us like "how do we know you can deliver", because we had some satisfied customers (assuming there is such a thing), and they couldn't say "we are in the market and your programs are just theoretical and blah blah" because all our customers were in the market and all our systems were used in the market, and they couldn't say "you are just programmers and our quants have checked your numbers and they are wrong" because they had to deal with me in that case and quants quickly learned to be scared of me, and everyone knew about my textbook, and so they never ever (and never had) argued that we didn't know finance. So we were getting somewhere. But we were still a small software computer who had to figure out how to keep getting new customers.


I offered the GAF person we were working with (Gerry) a job as Sales Manager. The idea being we could finish the corporate treasurer risk-management system without dickhead interference, and he could sell the system to other corporations (he was confident he could do so) for a percentage of the sale. (Neither he nor Farid nor myself was ever paid a salary. Farid and I split the profits and commission people got paid only if they performed. That way is risky, but always pays better if you succeed, and pays nothing if you fail.)

Well, Gerry got increasingly unhappy with GAF bureaucracy, then saw his annual bonus after working 12 hours days, and quit and came to work for us. And, guess what, the completion of our GAF contract got put in the hands of the MIS director (let's call him Doobie), who was now positive that there was a conspiracy between Gerry and us to exploit GAF. And things deteriorated, and then Doobie circulated a memo among several top officials at GAF, and right after that threatened legal action to my partner over the phone. So I had to insert myself for self-protection, and wrote a response to his memo, going over the whole history of the ridiculous time-wasting things we had to do, and finally ending with a point-by-point response to Doobie's memo. I started it with "Dear Mr. Doobie, I am writing to let you know we will no longer tolerate your insults, your verbal inaccuracies, and your threats of legal action." (I knew the latter two words would push the Chairman's button.) And I sent it to all those on Doobie's list, and also the Chairman, and also to the head FX trader/interest rate trader (since I knew this was good for at least a half-dozen people reading the memo—as my risk was it would be buried). There were several days of silence. Then I got a call from an investment bank in New York, and the guy laughing on the other end said he had heard about the "war" between GAF and FX Systems, and I said there was no war, but I had to write a memo to protect myself. Well, the chairman of mighty GAF ordered Doobie to go down to Philadelphia and patch things up. So Doobie had to take his humiliating journey, down to the little software firm of only 8 people, and we decided the only patch-up would be a contract detailing any complaints, and what we would do to fix them, and I expected Doobie to put up a big fight after his humiliation, but when we sat down at the table, the first thing he did was look at me and say "I don't want to get into a memo-writing war with you". He was already beaten. So he signed the contract, and went back to GAF, and in a few months was gone after 15 years, without having ever reached retirement. The word got around and people learned not to pick a fight with us.

First, Kill All the Lawyers

So, in no particular order, we sold corporate treasurer risk management systems to companies like Shearson, Sara Lee, McDonald's, Ontario Hydro, Dow Chemical (to several different divisions), and a number of other places (pieces of it to Bear Stearns, Drexel, etc.). I remember McDonald's sent down a 20-page contract. The longest we had ever signed before had been 6 pages. Well, it 1) contained things we hadn't agreed to (but some lawyer thought should be in any contract), along with 2) gibberish, ambiguous things about software written by someone who had obviously never turned on a computer (what exactly did it say?), and 3) also a section with references to things outside the contract, had non-sequiters and bad grammar (this section obviously a cut and paste job). So I crossed out 8 pages of stuff, and rewrote some, and we signed it and sent it back. A call came in wanting to speak to our "legal department". "Who is it?" I asked Christie, our receptionist and secretary. It was McDonald's corporate attorney, and when I spoke to him, he started off with the lawyerly bullshit that this was the type of contract McDonald's had to have, and they couldn't tolerate changes, etc. So first I pointed out some things we hadn't agreed to, then I went to one gibberish excised section and asked him what it meant (and, yes, I knew he was out of his element, and so wouldn't try to say, so I gave him one possible interpretation and why we therefore would not sign), and then I went to the cut and paste job and pointed out how it referred to things not even in the contract. Fortunately for him there were no witnesses, but all the changes stayed, because the attorney otherwise had to take the risk I would point out his incompetence to top management.

Meanwhile we were working on an over-the-counter FX system, which is hard to do when all resources are otherwise employed and you are always behind schedule. But we finally got something together, a start, which would be a basis to expand on. We realized our Sales Manager, while good at selling to corporations, wasn't good at selling to banks—he didn't have the right personality to get along with traders, so Farid had to do most of the bank selling and I supplemented as needed. We finally sold our first OTC FX System to Banker's Trust, Singapore, and since this was our first OTC system sale, and really our first foreign sale (we had previously sold a calculator to some obscure bank in Zurich, but that didn't count), I went half way around the world with Farid, so we could take a side trip to Bangkok, and see if the Thai food there was as good there as it was in Philadelphia. Then we started picking up the odd customer for the OTC FX System, but no major bank. We sold a system to a division of General Motors, and one to Continental Grain. Then to a regional bank or two, and to Andy Krieger who had started his own trading operation. Then we picked up Banker's Trust, London, and after that Banker's Trust, New York, and a couple of odd London banks. And we started penetrating other top tier banks: Bank of Montreal, Toronto-Dominion, and so on.

We spent a year going to Paris and trying to work out an agreement with a French software company that had, as clients, two hundred European financial institutions. We finally got the agreement signed—just in time for them to declare bankruptcy. It turned out to be very costly, but there had been some consolation, however. We would stay at the Hotel Maurice by the Louvre, with a view of the Eiffel Tower on the other side of the Seine, and if you stepped out on the small balcony on the junior suite on the top floor, you could look back up the Seine to the left and see the towers of Notre Dame. We signed the agreement in December, and following a dinner at a Thai-French restaurant, one of our clients drove us through the Bois de Bologne, and joined the caravan of cars inching along a half-mile section of road that was lined with prostitutes. Most were wearing fur coats, and as cars crawled by they would display their wares—usually some skimpy form of underwear (or perhaps not) underneath the coats. My girlfriend Sandy, who had come along on this trip, was less amazed by the whole somewhat amusing spectacle than by the simple fact that the women were standing in snow in their high heels. Iin between such diversions, we managed to sell systems to concerns like Societe General, Bank Paribas, Credit Agricole, and so on.

About this time I had begun pursuing another venture, the creation of what John and I called a "world commodity bank" (a private venture), which Farid didn't object to because if it succeeded there would be significant benefits to FX Systems and Farid. But after a while, I had to sell my interest in FX Systems to finance the new venture, the only contentious point being I sold my half to our Sales Manager, and Farid and Gerry didn't get along. (They in fact split up one and one-half years later, part of the agreement being neither could use the name FX Systems. Farid's half became FNX, Ltd., and I in fact supervised the building of another system by them, which I will get to.)

Slaying the Dragon

But shortly after I sold my half of FX Systems, the OTC FX System, for which I had created every numerical calculation of any type in, begin to be picked up by all those banks who had been locked-in to Devon for years: Citibank, Deutsche, Credit Suisse, Union Bank of Switzerland, Bank of Tokyo, Barclays, etc.

Credit Suisse, New York was the breakthrough that led to the others, but FX Systems had a problem with Credit Suisse at first in that Zurich sent two PhD's down, and they had to approve the theoretical part of the system, and no one at FX Systems could handle them. So I let FX Systems hire me as a consultant and made the two PhD's happy (the head FX trader at Credit Suisse, New York was already sold). The PhD's had seen some non-standard things in the pricing models, and like most academics who hadn't spent enough time with the math and the markets, assumed the problem was mine, when it was in fact their own lack of understanding both of option pricing and market realities.

Then Farid was thrown out of the New York office of Union Bank of Switzerland—the largest bank in Switzerland (it's now the United Bank of Switzerland), and one of the largest banks in the world based on real capital. (Only morons—meaning most of the financial media—base "size" on assets or deposits, as if the quicker a bank makes non-performing loans makes it "bigger" rather than smaller. But this is the basis of the myth of "large" Japanese banks.) What had happened was that I had become an expert witness for the law firm of a trader in New York who was suing UBS. I had taken the job as a favor to a friend, but I hadn't really wanted to do it. So I went looking for the smoking gun to prove the trader guilty, so I could tell his lawyers, so they could pay me and kiss me goodbye. But I didn't find it. I found a lot of things wrong, but it could be shown that this was all approved by the bank. Now Farid had gone to UBS New York by invitation to demo the OTC FX system, and UBS had just found I was representing the trader, and Farid comes in and spots one of our old clients from the PHLX, who is now working for UBS, and says hi Lenny, and Lenny looks at him unfriendly and says, "We're taking a corporate attitude on this." No one could quite buy the fact that I wasn't really still at FX Systems, and so Farid wasn't allowed to see the person who had invited him, and Farid was immediately escorted out of the bank.

The case had been in the New York papers, and there was considerable prestige involved, the UBS attorney having declared "we're not going to settle this case out of court"—and had apparently assured Zurich everything was a piece of cake, and if they went to court they would crush the trader, and then counter-sue him. (And part of the reason for this was if the trader really had a case, then UBS New York management had been negligent in several ways.) Now Swiss banks are different from American banks. You can be in an accounting firm involved in a lawsuit against Citibank, and yet go down to Citibank to take out a business loan, and no one would blink an eye, because one part of the bank doesn't care what is going on in another part. Not so UBS: they were taking a corporate attitude.

When I finally got all the evidence I needed and submitted my expert witness report, it hit Zurich like a bombshell. And they pulled their corporate lawyer in New York off the case, and got someone with more trial experience and I spend an entire day giving a deposition. And they got their own expert witness, a smooth old-timer from the Chicago option markets who sounded good (but whom I could eat for lunch—yet what would the jury think?) and he assured UBS they were right. So, now, what happened on this piece-of-cake case in which now they had their own expert witness, and which they had put all their prestige on the line in the statement they wouldn't settle out of court? Well, they promptly settled out of court.

The only thing they had against them was me, and I made one of the world's largest banks back down and eat crow. Yet six months later UBS Zurich purchased the OTC FX system. Why? Well, despite their semi-public humiliation, they had to acknowledge it was the best system in the world, even though every number in it was programmed by their "enemy," me. They told Farid in Zurich: "We realize we've had our differences with Dr. Grabbe in the past, but let him know that he is welcome here in Zurich any time." And, truly, there never were any hard feelings on my part. I knew Zurich had been snookered by New York.


John and I never quite sold the idea of the "world commodity bank", but we sold a piece of it to Barclays and John took a position there, but I only agreed to become a consultant four days a week. But part of the plan involved building a commodity system, which of course would handle foreign exchange, and exchange-traded commodities, but in particular also handle those strange metals traded at the London Metals Exchange and associated OTC contracts that had special characteristics. The plan had been that FX Systems (just now turned FNX) would build it for us, and I would see that it was done properly. But then Barclays New York lawyer (lawyers are always the leading criminals in any organization) said I had a conflict of interest, and so John and Barclays kept me off the project for 6 months.

Now, here is the situation. John knows exactly what he has in mind. FNX has all the software built to date, along with all my pricing modules, and in fact have a new pricing module I supplied for the commodity system. But nothing happens. They spent six months yelling at each other. You see, it isn't that easy. There wasn't anyone around with sufficient understanding of all parts of the process to show them how to put such a system together. Well, Barclays finally caved in to the inevitable, and I began to check the system and explain to FNX what they had to do, and so supervised another entirely new system every number of which I created and checked. (When it was finally built, Barclays to my disgust promptly abandoned it; but then FNX sold copies to two of the biggest traders at the LME, Barclay rivals, so it turned out to be another creation that endured.)

Toward the end of the commodity system project, Barclays (I was actually working for their investment bank, BZW, which conducted most of the trading operations, and not for Barclays, the commercial bank) began to get static from the Securities and Futures Authority (SFA), which is sort of a British equivalent of a combined Securities Exchange Commission and Commodity Futures Trading Commission. There was British legislation that required large traders to produce a particular risk report and submit it regularly to the SFA.

Naturally the task was given to me, so that the report would become one of the many reports the commodity system could produce. I read through the legislation, and when I finally decided I had figured out what was requested, I realized there were all sorts of undefined things in the law. So I wrote a long letter to the SFA explaining these undefined things and asking for clarification. I got no response. So, I thought, the file has been properly papered, and if they haven't responded, then I am free to make my own reasonable assumptions (which I knew would be better than any coming from the SFA, anyway). So I implemented the report with respect to those contracts that the system was ready to handle.

So this was presented to the SFA in London. And guess what? The reason they hadn't responded to my questions was that no one had ever before implemented the report. I was the first one to do so. The SFA needed someone to show them how to do it—and there aren't many of us around.

You see, to be an entrepreneur, you have to be an artist, and you have to be a creator. And you if combine that with a knowledge of reality—if you really know what you are doing—then there is no stopping you.

J. Orlin Grabbe is the author of International Financial Markets, and is an internationally recognized derivatives expert. He has recently branched out into cryptology, banking security, and digital cash. His home page is located at http://orlingrabbe.org/ .


from The Laissez Faire City Times, Vol 5, No 15, April 9, 2001