The Brady Commission report grave a coherent, factual account of the fateful week from Oct. 14 to Oct. 20, but it stopped short of drawing the appropriate theoretical conclusions. Events have demonstrated that financial markets are inherently unstable. Moreover, the instability is cumulative: The longer markets function without supervision explicitly aimed at maintaining stability, the greater the danger of an accident like Oct. 19, 1987.

These conclusions contradict the generally accepted view that market participants are supposed to act on the basis of rational expectations and markets are supposed to produce a permanent state of equilibrium—provided they are allowed to operate unhindered by regulation.

Unfortunately, this theory isn’t consistent with reality. The concept of rational expectations doesn’t apply when those expectations relate to events that are themselves contingent on the participants’ decisions. The events cease to be uniquely determined and participants are confronted with genuine uncertainty. They are forced to take a view and the bias they bring to bear plays a role in the outcome.

There are many ways participants’ biases can influence events. When people lose confidence in a currency, its decline tends to reinforce domestic inflation, thereby validating the decline. When investors have confidence in a company’s management, the rise in share price makes it easier for management to fulfill investors’ expectations. When bankers lend to developing countries, the growth of those economies can go a long way to validate the bankers’ confidence until an eventual reversal sets off a self-reinforcing process in the opposite direction. I call such initially self-reinforcing but eventually self-defeating connections “reflexive.”

Reflexive connections do not operate with equal force in all markets at all times. Nevertheless, the patterns often show similarities. For instance, the resemblance between the crashes of 1987 and 1929 is uncanny. The tendency for the dollar to overshoot, both on the up and the down side, is equally noteworthy.

The reason reflexive processes follow a dialectic pattern can be explained in general terms: The greater the uncertainty, the more people are influenced by the market trends; and the greater the influence of trend-following speculation, the more uncertain the situation becomes.

In currency markets there has been a mutually reinforcing connection between the relative importance of international capital movements, which have become progressively more trend- following, and excessive exchange-rate fluctuations. The monetary authorities recognized this fact in the 1985 Plaza Agreement, the 1987 Louvre Accord and the massive central-bank intervention since Jan. 4.

In the stock market, however, the growth of a trend-following bias has largely escaped attention. Yet the market has come to be dominated by professional investors whose ability to attract clients depends on their performance. When one is judged in comparison with market averages, it is difficult to keep one’s own judgment independent of the market trend; so difficult, indeed, that many people give up: hence the growing popularity of indexing, An index fund follows the market trend automatically. To protect against a reversal of the trend, portfolio insurance programs were invented. They are also trend-following devices, because they automatically sell when the market declines and buy when the market recovers. The same is true of option-writing programs. All these devices allow the individual participant to limit his risk, but only at the cost of increasing the instability of the market.

Eventually, the reliance on trend-following devices became greater than the capacity of the market to accommodate them. When the market started to fall, it continued to accelerate until it became disorganized, and some of the supposedly automatic programs could not be executed.

Computerized arbitrage trading between index futures and actual stocks cannot be blamed for what happened. By providing a connection between the two markets, it merely facilitated the execution of orders—although the fact that sell orders could be executed kept on generating new sell orders until the connection broke down. The ensuing panic disparity between index futures and actual stock prices then generated additional selling pressure.

Much of the discussion about liquidity or its lack is misplaced; what matters is the balance between buyers and sellers. Trend-following speculation (such as indexing, performance measurement and technical analysis and trend-following devices (such as portfolio insurance and option writing) disrupt the balance. Financial markets need a measure of liquidity to permit execution of buy and sell orders without excessive transaction costs; but beyond a certain point, liquidity, or its illusion, can be harmful because it encourages trend-following behavior.

To prove the point, the stock market dropped 6.7% last Friday. The size of the drop was not only unexpected but also inexplicable by anything other than instability due to trend-following behavior.

It is easier to diagnose the disease than to prescribe a remedy. The Brady Commission is right to recommend a single supervisory agency for all financial markets. The agency should be expressly charged with maintaining stability. How it will discharge its duty is more problematic, because supervisory agencies develop their own biases. In particular, they tend to fight yesterday’s excesses instead of focusing on tomorrow’s. If the illusion of excessive liquidity contributed to the crash of 1987, the current danger is excessive illiquidity: Any measures specifically designed to prevent a recurrence of the October experience are likely to be counterproductive. What is needed is the recognition that financial markets are inherently unstable and stability ought to be made the objective of public policy.