Appendix 1- Program Trading and Portfolio Insurance One subject that has received widespread publicity in recent years is program trading. Perhaps never in the
history of financial markets has there been more criticism about a trading approach that was less understood. I would
venture a guess that less than one out of ten people opposed to program trading even know the definition of the
term. One source of confusion is that program
trading is used interchangeably to describe both the original activity
and as a more general term encompassing various computer-supported trading strategies (for example, portfolio
insurance).
Program trading represents a classic arbitrage activity in which one market is bought against an equal short
sale in a closely related market in order to realize small, near
risk-free profits, resulting from short-lived distortions in
the price relationship between such markets. Program traders buy or sell an actual basket of stocks against an equal
dollar value position in stock index futures when they perceive the actual stocks to be underpriced or overpriced
relative to futures. In effect, program trading tends to keep actual stock and stock index futures prices in line. Insofar
as every program-related sale of actual stocks is offset by a purchase at another time and most program trades are
first initiated as long stock/short futures positions (because of the uptick requirement in shorting actual stocks),
arguments that program trading is responsible for stock market declines are highly tenuous. Moreover, since the bulk
of economic evidence indicates that arbitrage between related markets tends to reduce volatility, the relationship
between increased volatility and program trading is questionable at best.
Portfolio insurance refers to the systematic sale of stock index futures as the value of a stock portfolio
declines in order to reduce risk exposure. Once reduced, the net long exposure is increased back toward a full position
as the representative stock index price increases. The theory underlying portfolio insurance presumes that market
prices move smoothly. When prices witness an abrupt, huge move, the results of the strategy may differ substantially
from the theory. This occurred on October 19, 1987, when prices gapped beyond threshold portfolio insurance sell
levels, triggering an avalanche of sell orders which were executed far below the theoretical levels. Although portfolio
insurance may have accelerated the decline on October 19, it could reasonably be argued that the underlying forces
would have resulted in a similar price decline over a greater span of days in the absence of portfolio insurance. This is
a question that can never be answered. (It is doubtful that program trading, as defined above, played much of a role
in the crash of the week of October 19, since the severely delayed openings of individual stocks, tremendous
confusion related to prevailing price levels, and exchange restrictions regarding the use of the automated order entry
systems severely impeded this activity.)
163