A 1990s case on liquidity cycles, rank rotation, and cross-market sleeves

A historical book, not a single setup

The archive describes an early live application of a purely quantitative basket. About 90 percent of profits came from about 5 percent of trades. That result forced a shift from averaging many patterns to studying what the thin slice shared.

In this workflow, rank rotation means ordering candidates by those shared characteristics and concentrating in the small slice that historically produced most of the results. Diversification means spreading risk across several equity, bond, and currency markets so one trade sits inside a broader book rather than standing alone.

Liquidity as the regime map

The operator treated the liquidity cycle as a core regime map. The liquidity cycle is the sequence in which easier or tighter money first moves through bonds, then stocks, then the real economy. An expansion in liquidity during disinflation washes through the most liquid assets first.

When the economy weakens enough, the described sequence is that short-term rates fall and the yield curve steepens. That path was used as a weeks-to-months read on how liquidity is flowing through the system.

Here the yield curve is the spread between longer-term and shorter-term rates, used as a weekly-to-monthly growth and liquidity gauge. The market regime is the prevailing mix of liquidity, inflation, and growth that tells the book whether to expand, rotate, or stand aside. Asset allocation means choosing how much weight each market sleeve receives as those conditions change.

Weights across markets

A preferred allocation sleeve when conditions were supportive was five or six equity markets, four or five bond markets if bonds were favored, and three or four currencies. Five asset classes were treated as a workable minimum rather than a hard ceiling.

Unlike data and older regimes

Combining unlike data, such as a momentum tool with a volume-based tool, was presented as more useful than stacking several momentum measures. The top trades mixed quantitative signals with other information rather than more of the same.

The case argued that methods should be stress-checked on older regimes such as 1982 or 1973-74, because techniques that look robust in one stretch can fail when the cycle changes.

By late 1993 the same framework still found many individual names cheap relative to their growth even when the broad US market was not cheap on a macro basis. Most new listings were described as more stretched than the market itself.