Housing prices do not move randomly; they follow an 18-year rhythm that has mapped the last three U.S. housing peaks within a year of their actual top.
Most people treat housing cycles as reactions to interest rates or employment data. The Federal Reserve adjusts rates, economists release jobs reports, and headlines scream about affordability. Underneath that noise, land values in the United States have tracked a much older clock. In 1933, economist Homer Hoyt, then at the University of Chicago, analyzed rent and land values across major American cities and identified a recurring pattern. He argued that land is unique because supply cannot expand instantly to meet demand, creating a lag that repeats roughly every 18 years.
The mechanism is simple but slow. When prices rise, developers break ground. Zoning boards approve permits, construction crews pour concrete, and units are delivered. By the time those new homes hit the market, demand has often shifted. The Federal Reserve Bank of St. Louis tracks housing starts, which consistently lag price peaks by 18 to 24 months. This lag ensures that supply always arrives too late to stop the boom or too early to prevent the bust.
The four phases of the cycle
The cycle is not a sine wave. It is a sequence of four distinct phases, each with specific characteristics that repeat with high fidelity. The recovery phase begins with low prices and high vacancy. The expansion phase sees prices rise faster than rents as speculation enters. The hyper-supply phase occurs when construction peaks and vacancies turn negative. The recession phase follows, where prices correct and construction stops.
The National Association of Realtors tracks historical median sales prices, which align closely with Hoyt’s original data when adjusted for inflation. Comparing the last three full cycles reveals the stability of the 18-year interval.
| Cycle | Recovery Start | Peak Year | Years Elapsed | Phase Characteristics |
|---|---|---|---|---|
| 1 | 1971 | 1989 | 18 | Peak followed by 1990s recession |
| 2 | 1989 | 2006 | 17 | Peak followed by 2008 financial crisis |
| 3 | 2006 | 2024 | 18 | Peak followed by current correction |
The table shows the timing of the peaks. The 1989 peak was followed by a downturn in the early 1990s. The 2006 peak preceded the 2008 crash by exactly two years, accounting for the lag in construction delivery. The 18-year interval from 2006 implies a 2024–2026 inflection. The S&P CoreLogic Case-Shiller Index showed continued price growth through 2024 but a marked deceleration relative to the 2020–2022 trajectory — consistent with the late-expansion phase rather than confirmed peak.
The credit and construction lag
The 18-year cycle is driven by two factors: the inelasticity of land supply and the availability of credit. Land is fixed in location. You cannot build more of it in Manhattan or San Francisco. When demand grows, prices must rise to ration the existing supply. This price signal triggers new construction, but construction is slow.
Credit availability amplifies the lag. In the early expansion phase, lending standards are tight, which keeps prices stable. As prices rise, lenders loosen standards to capture market share. This influx of credit fuels the expansion phase, pushing prices further above fundamental value. By the time credit tightens again, usually near the peak, developers are already three years into building projects.
The math of the correction is brutal. During the recession phase, prices typically drop 20% to 30% from the peak. This drop is not linear. It accelerates when inventory turns. The Federal Reserve Bank of St. Louis data shows that in the 1990 and 2008 recessions, real median home prices fell by approximately 15% within 24 months of the peak. The 2006 peak saw a real price decline of nearly 30% over five years as the financial system unwound.
The current cycle mirrors the 2006 peak more than the 1989 peak. In 1989, interest rates were higher relative to income, which capped the bubble. In 2006 and 2024, low rates allowed leverage to push prices higher before the supply lag caught up. The difference is that the 2024 peak occurred with tighter lending standards than 2006, which may limit the severity of the crash. However, the price level itself remains the primary risk.
Where the cycle stands today
If the 18-year cycle holds, the 2006 peak to 2024 peak marks exactly 18 years. A buyer entering the market in 2024 is buying at the top of the expansion phase. The math says prices typically drop 20% to 30% during the recession phase. Waiting three years could save 15% on a $400,000 home, but borrowing costs might rise. The cycle does not care about the Federal Reserve’s current rate; it cares about the time it takes to build a house.
The pattern suggests that 2024 is the inflection point where supply begins to exceed demand. The 2026 to 2028 window is historically the period of highest vacancy. Buyers who purchase at the peak face the risk of negative equity for a decade. The 1989 buyers waited 10 years to break even in nominal terms. The 2006 buyers waited 14 years in nominal terms, longer in real terms.
The 18-year cycle is not a prophecy of doom, but a map of risk. It identifies when the risk of price correction outweighs the risk of missing out. The tradeoff is clear: buying at the peak offers immediate access to housing but carries a 20% downside risk over the next five years. Buying at the bottom offers a 15% price discount but carries the risk of higher rates. The cycle favors the buyer with cash who can wait for the 2026 correction. For most, the math says holding liquidity until the cycle turns is the safer bet. The 18-year clock is ticking, and the hands are at the same position they were in 1989 and 2006.