1993issue C121-12
Leading indices at bull-market peaks
The archive workflow scores how many leading-index series have already topped and compares that cluster with tape and monetary timing signals. The editorial use is as a regime-overlay, not as a standalone sell trigger.
- Treat a bull-market-peak as a confirmation problem: score how many leading-index series have already topped, then read that count as a regime-overlay.
- Weekly volume and weekly new highs were the only two of 14 tested series with a perfect lead-rate before every available bull-market-peak in the sample.
- Even weekly volume, the most tightly clustered leader, led by 2 to 59 weeks, so median-lead is a relative check rather than a calendar.
- New highs across most leaders argue for a bull-market correction; unusually long, widespread confirmation-and-divergence argues that a major top may already be in.
A confirmation problem
A bull-market-peak is a classified high after a large advance that is then followed by a classified bear market under explicit percentage and duration rules. The working problem is confirmation: how many historically leading series have already made their own highs.
A leading-index is a market or economic series that has historically reached its own peak before the broad equity benchmark peaked. Confirmation-and-divergence is the rule that a trend is more trustworthy when related averages or leading series make new highs together, and more suspect when they fail to confirm.
The historical sample
A long historical sample classified 31 bull markets and 30 bear markets in the Dow Jones Industrial Average from 1900 through the early 1990s, using explicit percentage-and-duration rules plus a limited Value Line exception for 1990.
In that sample the median bull market lasted 614 days and rose 69.1%, while the median bear market lasted 375 days and fell 26%.
How an index peak was marked
An index peak was defined as the highest point between the preceding bear-market low and the midpoint of the following bear market, with a few documented exceptions for lagging highs, a 1973 volume spike, the 1987 volume high, and a missing T-bill trough in 1978.
Lead-rate is the share of historical bull-market peaks at which a given series topped before the benchmark.
Breadth gauges and imperfect clocks
Weekly volume and weekly new highs were the only two of the 14 tested series with a perfect record of topping before every available bull-market-peak in the study. Each series is a breadth-gauge, a participation measure used to judge whether price strength is widely supported.
Even the most tightly clustered leader, weekly volume, still showed a wide lead-time range from 2 weeks to 59 weeks, so lead reliability is relative rather than clock-like.
Count the series that have already peaked
A practical check is to count how many of the 14 series have already peaked and to compare the current median-lead with the historical median-of-medians of about 26 weeks before a typical bull-market top. Median-lead is the typical number of weeks by which a series, or the median of several series, has peaked before the benchmark high.
Editorial note: the same cluster can be summarized as a composite-lead-index, a single constructed series that combines several leading indicators, often with weights based on each series' historical lead-rate. The archive workflow itself is the count of peaked series plus the median-lead comparison.
An archive-period snapshot
As of mid-September 1993, 9 of the 14 series (64%) already implied a peak, and the median series had peaked 29 weeks before the late-August 1993 high.
Editorial note: that snapshot belongs to the archive window. It is not a claim about later markets.
Sector leadership as a second layer
Brokerage stocks led market peaks in 4 of 5 cases in the illustrated comparison, showing that sector leadership can be used as an additional intermarket confirmation layer.
Confirm tape and monetary signals
The intended use of the leading-index set is to confirm or challenge tape and monetary timing signals. New highs across most leaders argue for a bull-market correction, while unusually long, widespread divergence argues that a major top may already be in.