How Milton Berg Called the 1987 Crash

Valuation you could actually measure

By September 1987, the relationship between stock and bond yields put equity valuations above every prior peak in the data — 1972, 1968, even 1929. Berg is the first to say valuation alone is never a timing tool; the greatest long-term winners can look expensive for decades. But as one input among several independent warnings, an all-time extreme is not something to wave away.

Breadth was quietly falling apart

While the averages made headlines, the market's internals were diverging. The Dow Jones Utility Average peaked on January 22, 1987 on record volume. The NYSE advance-decline line topped in March. Yet the S&P 500 rallied another +24.91% into late August — a narrowing advance in which fewer and fewer stocks carried the indexes higher. Divergences like these are a classic feature of major tops, visible to anyone tracking breadth rather than price alone.

Policy and sentiment crossed wires

The Federal Reserve had not raised rates in three years, and that long calm had bred excessive optimism. When the Fed finally hiked on September 4, 1987, sentiment indicators did something perverse: they turned more bullish. Investors celebrating into a tightening is a hallmark of complacency — the crowd was positioned for a market that no longer existed.

A machine built to sell into a falling market

In mid-September 1987, Berg attended a Goldman Sachs meeting at which a Nobel laureate enthusiastically promoted “dynamic hedging.” To Berg, the implication was immediate: portfolio insurance would mechanically force selling into declines, converting any ordinary downturn into a cascade. It was not a prediction about investor psychology — it was arithmetic about how the strategies were built.

That structural insight has a modern echo. To this day Berg refuses to recommend leveraged ETFs, which by construction must sell as markets fall and buy as they climb — the same mechanical pro-cyclicality, wearing a new wrapper.

The timing window

Finally, the cycle work of Paul Macrae Montgomery — the analyst Berg names as his greatest single influence — pointed to a volatility window running from August 24 to November 5, 1987. The major indexes peaked in late August, inside the window. The crash arrived on October 19, inside the window.

What he did — and why it still matters

Berg did not just write the warning down. He had already reduced equity exposure in the option-writing fund he oversaw to roughly 20%, and his October 1987 sell signal at Oppenheimer — where his three mutual funds each held top five-year Lipper rankings — helped earn him the Institute of Econometric Research and Sylvia Porter's Personal Finance Magazine's mutual fund manager of the year award.

The lesson of 1987 is not that any single indicator called the crash. It is that five unrelated kinds of evidence — an extreme valuation ratio, breadth divergences, a perverse sentiment response to policy, a structural selling mechanism, and an independent cycle window — all pointed the same direction at once. Taken together, they tipped the odds away from a routine correction and toward a serious break.

That is the template Berg has systematized ever since: never one signal, always the cluster. It is the same architecture that flagged the April 2025 low from the other direction.