Fortune's Formula (22 page)

I
N THE EARLY
1970s,
Thorp got a lead that actor Paul Newman might be interested in investing. Newman had just done
The Sting
. (The plot concerns a con artist going by the name of Kelly who uses a delayed wire service scam to dupe a gangster into placing a ruinously large bet.) Thorp had a beer with Newman on the Twentieth Century Fox lot. Newman asked how much Thorp could make at blackjack if he did it full-time. Thorp answered $300,000 a year.

“Why aren’t you out there doing it?” Newman asked.

“Would you do it?” Thorp asked.

Thorp estimated that Newman made about $6 million that year. Thorp was making about the same.

Newman decided not to invest with Princeton-Newport. He expressed reservations about the way the firm made trades to minimize taxes. Newman explained that he was a highly visible liberal activist. He was number 19 on President Richard Nixon’s “enemies list.” Newman suspected the government gave his tax returns extra scrutiny. He did not want to do anything with his taxes that might give the slightest cause for suspicion.

Indeed, not all of the thinking at Princeton-Newport had to do with making money. Some had to do with keeping it. The tax implications of trades were carefully considered.

“I’ve estimated for myself that if I had to pay no taxes, state or federal, I’d have about thirty-two times as much wealth as I actually do,” Thorp told me recently. This statement shows how the power of compounding applies to expenses as well as profits.

Take Shannon’s pipe dream of turning a dollar into $2,048. You buy a stock for $1. It doubles every year for eleven years (100 percent annual return!) and then you sell it for $2,048. That triggers capital gains tax on the $2,047 profit. At a 20 percent tax rate, you’d owe the government $409. This leaves you $1,639. That is the same as getting a 96 percent return, tax-free, for eleven years. The tax knocks only 4 percentage points off the pretax compound return rate.

Suppose instead that you run the same dollar into $2,048 through a lot of trading. You realize profit each year, so you have to pay capital taxes each year. The first year, you go from $1 to $2 and owe tax on the $1 profit. For simplicity, pretend that the short-term tax rate is also 20 percent (it’s generally higher). Then you pay the government 20 cents and end the first year with $1.80 rather than $2.00.

This means that you are not doubling your money but increasing it by a factor of 1.8—after taxes. At the end of eleven years you will have not 2
¹¹
but 1.8
¹¹
. That comes to about $683. That’s less than half what the buy-and-hold investor is left with after taxes.

In the late 1970s, Jay Regan came up with a clever idea. At that time, a treasury bond was still a piece of fancy paper. Attached to the bond certificate were perforated coupons. Every six months, when an interest payment was due, the holder would detach a coupon and redeem it for the interest payment. After all the coupons were detached and the bond reached its maturity date, the bond certificate itself would be submitted for return of the principal.

Regan’s idea was to buy new treasury bonds, immediately detach the coupons, and sell the pieces of paper separately. People or companies that expected to need a lump sum down the road could buy a “stripped,” zero-coupon bond maturing at the time they needed the money. It would be cheaper than a whole bond because they wouldn’t be paying for income they didn’t need in the meantime. Other people might want the current income but not care about the future lump-sum payment. They would buy the coupons.

An even bigger selling point of Regan’s idea was a loophole in the tax law. Most of the pieces of paper from a dismembered bond would sell for a small fraction of their face value. This was as it should be. A zero-coupon $10,000 bond that matures in thirty years is not worth anywhere near $10,000
now
. Since there are no interest payments, the buyer can profit only by capital gains. That is possible only if the buyer pays much less than $10,000 for the bond now.

Fair enough. Buy a $10,000 bond, strip off the coupons, and resell the zero-coupon bond for, say, $1,000. This, it was theorized, ought to give you the right to claim a $9,000 capital loss on your current year’s taxes. At any rate, nothing in the tax code said how taxpayers were supposed to figure the cost basis of the various parts of the bond. The law said nothing because no one in Congress had thought of stripping treasury bonds at the time the laws were written.

Regan took the idea to Michael Milken. Milken thought it was great. A wealthy investor with a million dollars’ worth of capital gains had only to buy $1.1 million worth of new bonds, strip the coupons, and sell the stripped bonds for $100,000. Presto chango, the capital gains disappear. Despite the nominal tax loss, the seller was not really out anything. The $100,000 plus the coupons were still worth approximately the $1.1 million paid. It could even be argued (a few shades less convincingly) that the government wasn’t out anything. The taxes on the rest of the bond would be paid, later rather than sooner.

Milken set up a company called Dorchester Government Securities to market the idea to clients. Dorchester was based in Chicago and seems to have been little more than an address—One First National Plaza, Suite 2785. In 1981 Dorchester changed its name to Belvedere Securities. Regan and Thorp were made partners of Dorchester/Belvedere. The other partners included Michael and Lowell Milken and Saul Steinberg’s Reliance Group Holdings. Steinberg was one of Milken’s most successful junk bond raiders.

“Creative” tax shelters rarely last long. After a few tax seasons, the Treasury Department complained it was hemorrhaging revenues to a loophole that no one in Congress had ever intended. The 1982 Tax Equity and Fiscal Responsibility Act closed the loophole by requiring investors to account for the value of the coupons in claiming losses. At the same time, the new law confirmed the right to sell stripped treasuries (they are still being sold) and replaced paper bond certificates and coupons with electronic bookkeeping.

A
NOTHER GOVERNMENTAL DECISION
opened a new opportunity for Princeton-Newport. The U.S. government decided that AT&T was a monopoly after all. In 1981 the telecommunications giant was broken up into eight pieces. Each AT&T shareholder received shares of the seven “Baby Bells” (regional telephone companies) and of the “new” AT&T. It was possible for investors to get a head start on the breakup and buy shares of the Baby Bells and “new” AT&T before they were officially issued. Thorp’s computer alerted him to a weird disparity. The shares of old AT&T were slightly cheaper than the equivalent amounts of the new companies.

There were Wall Street analysts who spent careers analyzing AT&T, and they paid no attention to this. The price differential was so small that costs would eat up any profit…unless someone bought an awful lot of the stock.

Princeton-Newport’s capital was then about $60 million. Judging the deal to be as risk-free as these things get, Thorp borrowed massively to buy 5 million shares of old AT&T for Princeton-Newport and sold short a corresponding quantity of the eight new companies. The 5 million shares cost about a third of a billion dollars. That was a leverage ratio of about 6 on the hedge fund’s total capital.

The trade was the largest ever made in the history of the New York Stock Exchange. Thorp paid $800,000 in interest on the borrowed money. He still cleared a profit of $1.6 million on the dissolution of the former employer of Claude Shannon and John Kelly.

In April 1982 a new investment called S&P futures began trading. S&P futures allow people to place bets on the stock market itself, or more exactly on the Standard and Poor’s index of 500 large American companies.

A futures contract is an “option” that isn’t optional. In both types of contracts, two parties agree to a future transaction at a price they set now. With an option, one party—the option holder—has the right to back out of the deal. The option holder
does
back out unless he can make a profit by exercising the option.

With a futures contract, neither side can back out. The holder of a futures contract gets all the profits
or
losses that would accrue by buying the securities outright.

What’s the difference between buying S&P futures and investing in a plain old S&P 500 index mutual fund? The answer is that you put up a lot less money with futures. An S&P futures contract is a
cheap
ticket on a wheel of fortune that has big prizes and big penalties. Anyone who knows which way the S&P index is heading can make a huge profit.

Thorp did not know what the market was going to do. He did see a new way to turn a profit.

The two parties to an S&P futures contract are theoretically agreeing to the sale of a portfolio of all the S&P 500 stocks. No one actually buys the 500 different stocks to deliver. Instead, the parties figure who would owe whom and settle up in cash.

They settle not just on the transaction date but at the end of trading every day throughout the term of the contract. This is necessary because of the big losses possible. Daily settling ensures that no one gets too far behind, minimizing the chance of a big default.

What’s an S&P futures contract worth? Thorp suspected that people would be playing hunches. Brokerages had staffs of highly paid analysts who guesstimated where the S&P was going to be in so many months. Thorp believed this advice was virtually useless. When people invest based on useless advice, there may be an opportunity to profit.

Thorp used software to determine a fair price for S&P futures. He had to model the random walk of all 500 S&P stocks. Princeton-Newport’s minicomputers had a huge speed and storage advantage over what was available to most other traders. The computer model told Thorp that the S&P futures were, like many exciting new things, overvalued. That implied that Princeton-Newport could make money by selling S&P futures. But hedging the trade would mean buying all 500 S&P stocks, racking up a lot of trading costs.

Thorp did further calculations and concluded that buying selected sets of S&P stocks would provide sufficient protection. As Thorp computed a high probability of success, Princeton-Newport committed $25 million of its capital to the S&P futures, doing as many as 700 trades a day. There were days when the fund’s trades accounted for more than 1 percent of the total volume of the New York Stock Exchange.

The gravy train lasted about four months. The profit came to $6 million. Then the market got the message. The prices for S&P contracts dropped, and other traders started using computers. The price anomalies vanished.

In 1981, the year of the AT&T deal, Princeton-Newport achieved a return of 22.63 percent net of fees. For 1982, with the S&P futures trade, it was 21.80 percent. As fiscal 1982 ended, Thorp and Regan could boast that a dollar invested at the outset had grown to $6.61 in thirteen years.

By that time, Thorp and Regan had turned a conviction that the market can be beaten into one of the most successful investment partnerships of all time. It was rare enough to achieve greater-than-market return over thirteen consecutive years. Skeptical academics and some traders tended to judge such exceptional performance as a Faustian bargain. Successful arbitrageurs, it was held, were risk-takers. Sooner or later, they lose big.

Everything about Princeton-Newport refuted that view. The fund had never had a down year, or even a down quarter. With his talk of the Kelly formula to manage risk, Thorp gave every appearance of being “the first sure winner in history.”

The partnership was a marriage of opposites. Regan lived a continent away on a baronial 225-acre New Jersey ranch where he raised horses. In a 1986 profile of the partners in
Forbes
, it is Regan who supplies the sound bites. “Taking candy from a baby,” said Regan of one trade. “You back the truck up to the store and start loading it.”

Regan was “near the rumors, information and opportunities that are always rattling through the Wall Street network,” Thorp explained. “There’s a string of rumors coming down the pipeline. The further you are down the information chain, the less valuable the information is.”

Thorp was introspective, approaching the challenges of his work as a scientist. He measured his words like he measured everything else. Thorp was careful to characterize his fund’s performance as “getting rich slow,” as if more confident words might jinx things. Not until 1982 did he quit his “day job” teaching at UC Irvine.

Thorp was slow to display his now-considerable wealth. In the office, he dressed like a California professor on his day off, in mod shirts and sandals. When the Thorps finally decided it was time to buy a big house, they chose a hillside ten-bathroom home, said to be the largest in Newport Beach, with panoramic views from Catalina to the Santa Ana Mountains. It had a fallout shelter with 16-inch-thick concrete walls and steel doors. Ever mindful of the odds, Thorp computed that it could withstand a one-megaton hydrogen bomb blast as close as a mile away.

Neither Thorp nor Regan could have imagined how soon it would all end, or how.