About the Data
Three federal series, one continuous record.
The cycle-drawdown view tracks total annual U.S. home sales — new construction plus existing — and measures each year's reading as a percentage change from the prior cycle peak. When sales set a new all-time high, the drawdown resets to zero. When sales fall below the prior peak, the drawdown shows the cumulative percentage decline. The metric answers a question that raw sales volume cannot: where are we in the cycle, and how bad is the current downturn relative to history?
Computing the drawdown requires identifying cycle peaks, which the Almanac does mechanically: a peak year is any year in which total sales exceed all prior years. The series has produced four peaks since the combined record begins in 1968: 1978 (7.99M combined), 2005 (8.36M), 2021 (6.89M), and — notably — the 1986 level (4.22M existing only, as the new-construction series peaked separately) does not register as a combined peak because the Volcker-era trough had not fully recovered by then.
The drawdown framework isolates the duration and depth of each bear market with a clarity that volume charts cannot provide. In volume terms, the 2024 reading of 4.74M looks similar to the 2012 reading of 5.12M or the 2001 reading of 6.09M. In drawdown terms, the 2024 reading reflects a -31% drawdown from the 2021 peak, while 2012 was -45% from the 2005 peak and 2001 was actually at a local high (not in drawdown at all). The drawdown framework correctly identifies 2024 as a mid-cycle bear market, not a recovery.
One limitation of the drawdown view: it is sensitive to which years are identified as peaks. If the combined series is computed differently — using only the existing-home-sales series for the full period, or using the Census new-sales SAAR rather than the annual unit total — the peak years shift and the drawdown percentages change. The Almanac uses the most straightforward definition: annual combined units, and the highest year observed up to that point.
Notable Cycles
Four genuine peaks, four wholly different recoveries.
The 56-year drawdown record has produced three major bear markets exceeding -40% and one current moderate drawdown. Their causes and character are distinct.
The Volcker bear (1979–1982), peak-to-trough: -45%
The 1978 combined peak of 7.99M (817K new + 3.99M existing, with differences in unit matching) gave way to a trough of approximately 4.41M in 1982 — a 45% collapse driven by the Federal Reserve's disinflation policy. The speed of onset was remarkable: the drawdown went from 0% in 1978 to -39% in 1981 in three years. The recovery was equally rapid: once rates fell, pent-up demand cleared the trough in three years. By 1986, total sales had recovered to near the 1978 level, even with the Volcker-era decade still fresh. This was a rate-driven bear: there was no excess supply, no fraud, no institutional collapse — just arithmetic pricing buyers out of the market. Remove the rate, buyers return.
The subprime bear (2005–2011), peak-to-trough: -45%
The 2005 combined peak of 8.36M (1.28M new + 7.08M existing) produced the deepest and most prolonged bear market in the series. The drawdown reached -45% by 2011 (4.57M combined), matching the Volcker-era trough in severity but taking nearly twice as long to reach it (six years vs. three). The recovery was dramatically slower: it took until 2021 — a full 16 years — for total sales to approach 2005 levels. The delay reflects the fundamental difference between a rate-driven bear and a credit-driven bear with structural excess supply. Foreclosures created shadow inventory that took years to clear. Credit standards tightened. A generation of first-time buyers was delayed or priced out by student debt and tight lending. The Volcker bear resolved in 3 years; the subprime bear required 16.
The rate-lock bear (2021–present), drawdown so far: -31%
The 2021 combined peak of 6.89M (771K new + 6.12M existing) has been followed by a drawdown to 4.74M in 2024 — a 31% decline in three years. This bear market is structurally unlike either predecessor. It is not rate-driven in the Volcker sense (demand collapsed but supply was fine) and not credit-driven in the subprime sense (there is no excess inventory or foreclosure wave). Instead, the primary mechanism is rate lock-in: existing owners are immobilized by below-market mortgages, creating a supply shortage that keeps prices elevated even as volume falls. The resolution pathway — waiting for either rates to fall or owners to accept portfolio losses — is slower and more uncertain than the rate-driven bear of 1982.
Definitions
- Sales
units, K or M
- Price
median, current $
- Rate
30-yr fixed, % APR
- SAAR
Census
- EHS
NAR
- PMMS
Freddie Mac
- Recession
NBER monthly