A deceptively simple idea that was meant to protect pension funds from loss ended up exacerbating the crash of '87.
In 1981, John O'Brien left A.G. Becker Funds Evaluation to join two academics in a new firm.
Leland O'Brien Rubinstein was to offer a revolutionary idea to pension funds and other institutional clients: portfolio insurance. The name was catchy, and the product was designed to protect portfolios against more than minimal losses.
The idea had been developed by Hayne Leland and Mark Rubinstein, finance professors at the University of California at Berkeley, and it was designed to protect portfolios against significant loss, no matter what the market did.
The concept was deceptively simple. Messrs. Leland and Rubinstein had developed a computer program that would tell a pension fund or the fund's money manager to sell stocks and increase cash in a carefully measured way as stock prices fell. By the time the stock prices had declined the maximum amount the pension fund could tolerate, the fund would be all cash.
For example, if the fund was willing to accept a 5% loss of capital on its equity portfolio, the program would begin to sell stocks as prices began to decline and the portfolio would be all cash by the time its value had declined 5%. As stock prices began to recover, the program would control the purchase of stocks until the portfolio was again fully invested.
In effect, the program was replicating a put option, but doing it through programmed, dynamic hedging.
The price of the option, in this case, was the slight underperformance the portfolio would suffer because its actions on both the upside and the downside would lag the market. That was in addition to the costs of selling and later buying the stocks.
The more volatile the market, the greater would be the underperformance and hence the cost of the insurance.
FUNDS ARE SKEPTICAL
Mr. Leland tested the program for a year with his own money and had great success. Then, after a year of trying to market it to a skeptical pension fund universe, he and Mr. Rubinstein joined forces with Mr. O'Brien, who was well-known and respected in pension fund circles. Soon after, Mr. O'Brien signed the first client of portfolio insurance -- the A. G. Becker profit-sharing plan. Six months later, three bigger clients came in on one day: the pension funds of Honeywell Inc., Gates Rubber Corp. and the Auto Club of South Carolina.
The Honeywell fund presented a significant challenge to LOR. It used seven equity managers, and changing the stock-cash mix for the fund meant working with each of those managers. At first LOR encountered some resistance, but Trust Company of the West volunteered to try the strategy with its Honeywell portfolio. When TCW found the program was successful and did not unduly interfere, the other managers agreed to try it.
Soon after, LOR was running a $200 million dynamic hedging program for Honeywell.
But the concept still was a hard sell. Managers resisted having parts of their portfolios taken away and replaced with cash.
"Fortunately,"said John O'Brien, "the S&P 500 futures market began to develop in 1982 and mature in 1983, so we could leave the assets in the plan and change the mix by selling S&P futures. It became a noninvasive procedure."
From mid-1984, all portfolio insurance strategies involved the use of futures. The first client for which the futures-based version was used was Manville Corp. "It was the guts of Jim Beasley (the treasurer) that caused it to happen," said Mr. O'Brien. "And it happened because they were in bankruptcy from asbestos claims. The finance committee wanted to skim off the pension surplus before they lost it. Mr. Beasley said: 'If you can sell it to the committee, we will do it.' " The committee bought it.
LOR got the bulk, but not all, of Manville's business. J.P. Morgan objected. Morgan had been the banker for Manville since the beginning of the firm, so it had clout. Morgan officials told the committee the bank could run a dynamic hedging program, and as a result, Morgan got 20% of the $350 million pension fund. Soon after, portfolio insurance became fully accepted as BEA Associates, Wells Fargo Investment Advisors and Aetna Capital Management licensed the technology.
The number of funds covered by portfolio insurance, and the amount of assets protected, began to surge.
Many pension executives signing up for portfolio insurance had become concerned about the level of the stock market as the Dow Jones industrial average climbed through 1500 and continued toward 2000.
One of those was Allen Reed, now president of General Motors Investment Management Co. In 1987, Mr. Reed was the president of Hughes Investment Management Co., overseeing the $3.2 billion Hughes Aircraft pension fund and $1.1 billion 401(k) plan.
"In 1987, like a lot of other pension plans, we got concerned about the market valuations and signed up for three portfolio insurance programs," he said. "The three managers were LOR, BEA and J.P. Morgan."
Meanwhile, pension executives who had portfolio insurance in place in 1986, when the market was more volatile than usual, found the program more expensive than expected. While portfolio insurance proponents had projected normal costs of 200 basis points to 300 basis points a year, the actual costs in 1986 were 50 to 100 points higher. The extra cost led Manville to modify its program -- which had aimed at a minimum return of 0% over a three-year time horizon -- to accept a loss of 5% as the minimum three-year return.
By the beginning of 1987, 18 of the top 200 pension funds and a large number of smaller funds used portfolio insurance to protect their assets. But there was skepticism that portfolio insurance would work as promised.
One who issued early warnings was Bruce Jacobs, a finance professor who later formed his own money management firm. Mr. Jacobs had warned the strategy was unstable and not equivalent to a true insurance policy.
One of the most prescient of the pension executives was John Steinbach, assistant treasurer of Kmart Corp. "I don't have any way to substantiate this," he said at the time, "but when the market starts to decline, I can't help but believe when these programs kick in, the selling of one program will make it so another program has to sell. Logically, it seems you're going to have more volatility." During the summer, Mr. Steinbach reduced the Kmart pension fund's equity exposure to 58% of assets from 70%.
Portfolio insurance received a test on Sept. 11, 1986, when the Dow dropped 86.6 points, its largest one-day point drop ever. The Standard & Poor's 500 stock index dropped 11.88 points. The product passed its test, but not without some difficulty, and critics said the cost of the insurance was increasing.
The pricing of stock index futures changed throughout that day, and when they were severely underpriced, Advanced Investment Management Inc., Pittsburgh, sold stocks in the cash market rather than sell the futures at too low a price.
"The liquidity wasn't there," said William M. Morris, senior investment consultant at William M. Mercer-Meidinger Asset Planning Inc., "and the futures contract was so undervalued that anybody who would have executed it was crazy."
LOR chose not to trade that day because of the market conditions.
Despite the questions raised by the test, interest in portfolio insurance increased as the stock market continued to climb after the mild correction. And some companies began to apply portfolio insurance to their defined contribution plans, among them Trans World Airlines Inc. and Burlington Industries Inc.
By the end of 1986, Wells Fargo reported gaining $7.5 billion in portfolio insurance business for the year, followed by Bankers Trust Co. with $2.1 billion. By October 1987, the total amount of assets so protected was $70 billion.
The ultimate test occurred Oct. 19, 1987, when the market dropped more than 500 points -- almost 23% -- in one day.
It failed the test.
In fact, portfolio insurance was blamed by many for deepening what otherwise would have been a normal market correction.
The market had fallen 250 points on the Dow the week before, with almost half of that occurring Friday, Oct. 16. Over the weekend more sell orders built up, and by noon on the 19th the Dow had dropped 100 points. It dropped another 200 points in the next two hours, and more than 200 points in the final hour.
Portfolio insurance firms struggled to keep their promises. They found they could not carry out their stock index futures sales at any reasonable price because there were few buyers.
Allen Reed reported the performance of Hughes Aircraft's three portfolio insurance managers was dramatically different. LOR could not trade and shut down its efforts. BEA Associates had moved the Hughes account 100% to listed put options before the crash, so the portfolio was protected. J.P. Morgan had moved to a mix of options and futures, and told Mr. Reed it was continuing to try to trade.
In the end, what Mr. Reed thought was the equivalent of a 5% deductible insurance policy turned out to be a 60% deductible. That is, the Hughes protected portfolios were down about 60%, as much as the market. "We were better off to have had it than not, but obviously things didn't turn out the way we thought," Mr. Reed said.
As Peter L. Bernstein elegantly explained in his book "Capital Ideas," the problem was the significant difference between an insurance policy and the synthetic insurance policy represented by portfolio insurance.
In a true insurance policy, the insurance company is contractually bound to make good in the event of a claim of loss, and that promise is backed by the insurance company's assets. But with portfolio insurance, there was no one bound to take the other side of the trade when portfolio insurance managers wanted to sell stock index futures.
Messrs. Leland and Rubinstein had assumed stock index arbitrageurs -- investors who bought stock index futures and sold stocks when the futures were cheaper than the stocks, and vice versa -- would always be ready to trade. At the opening on Oct. 19, according to John O'Brien, LOR and the other portfolio insurance managers were offering a lot of futures contracts for sale. At first the arbitrageurs bought the futures, as expected, but they had trouble selling the stocks they wanted to sell, as buyers of stock were scarce.
When the market started down dramatically, the arbitrageurs either got out of the way of the falling elevator or were overwhelmed by it, as those selling stocks kept stock prices falling as rapidly as the futures prices. The arbitrageurs were aware that there was at least $68 billion of portfolio insurance-programmed futures selling lined up to be executed if the market should continue to fall. They couldn't hope to cope with that, and they couldn't sell stocks at any reasonable price, so most stopped buying. With few buyers, the futures markets virtually ceased to operate at times on that Monday.
While many portfolio insurance managers did get some trades off, the prices were far lower than anticipated. Messrs. Leland and Rubinstein had overestimated the depth and liquidity of the futures market.
There were two problems with portfolio insurance.
First, it did not perform as advertised. Most pension executives who bought it expected either no losses or minimal losses. In fact, they suffered losses of 10% of the value of their equities or more, but less than the 21% the average equity portfolio suffered.
Second, portfolio insurance was blamed for deepening the crash. The critics blamed the futures selling for helping to trigger more sales of stock and setting off a vicious circle as stocks and futures followed each other downward.
Dynamic hedging did not disappear after the crash, but the name "portfolio insurance" almost did. Allen Reed acknowledged the Hughes Aircraft fund continued to use dynamic hedging with BEA Associates, although it modified the program to make it more "catastrophe insurance."
Portfolio insurance use continued to decline (at least in name) in 1988, as did the use of index arbitrage, the strategy that was supposed to be the other side of portfolio insurance.