Seven-Year Mean-Reversion Forecast
Forecast returns over a horizon that spans the typical rise and fall of a bubble.
- Difficulty
- Expert
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 95%
GMO's forecast uses seven years because historical bubbles in its study took roughly six and a half years, on average, to rise and fall. The framework estimates an asset's return from valuation, growth, and income, with mean reversion supplying a long-term anchor. It can also compare a normal regime with alternatives, such as persistently depressed real cash rates, rather than assuming one future with certainty. The output is not a prediction of the market's path or turning point. It is a standardized expected-return estimate used to rank asset classes, expose weak portfolio assumptions, and allocate capital toward better-compensated risks. Scenario weighting and error analysis remain necessary because policy, margins, and required returns may differ from history.
Origin
GMO originally used ten-year forecasts, then adopted seven years after Ben Inker's historical bubble study found an average rise-and-fall period of six and a half years.
Core principles
- 01Forecast horizons should reflect the duration of market cycles.
- 02Starting valuation anchors long-term expected returns.
- 03Growth and income must complement valuation.
- 04Different economic regimes require explicit scenarios.
- 05Forecasts guide allocation rather than predict exact paths.
How to run it
- 1
Set the horizon from cycle evidence
Choose a period long enough for valuation and profitability to normalize. GMO rounds its historical six-and-a-half-year bubble cycle to seven years.
Pro tip Match the horizon to the mechanism being forecast rather than selecting a convenient calendar period.
Watch out A shorter horizon may be dominated by sentiment and policy narratives.
- 2
Estimate valuation return
Measure how a move from current valuation toward a defensible normal level would affect annualized returns.
Pro tip Use several proxies for profitability and valuation.
Watch out Do not assume every financial series has equally strong evidence of mean reversion.
- 3
Add growth and income
Include sustainable fundamental growth and the income received by the investor. Keep these components separate so assumptions can be inspected.
Pro tip Express results in real terms when the investor's goal is preserving purchasing power.
Watch out Nominal returns can obscure a failure to beat inflation.
- 4
Build alternative regimes
Recalculate returns under credible structural alternatives, such as permanently lower real cash rates or higher required risk premiums.
Pro tip Use scenarios to expose sensitivity rather than to manufacture a preferred answer.
Watch out Do not assign certainty to policy variables that cannot be forecast consistently.
- 5
Rank the opportunity set
Compare expected returns across regions and asset classes on a consistent basis. Note both absolute attractiveness and relative superiority.
Pro tip Highlight when one asset is far more attractive than the next-best alternative.
Watch out Relative cheapness does not necessarily mean attractive absolute returns.
- 6
Allocate and monitor
Use the forecasts as primary strategic inputs, then revisit them as prices and fundamentals change.
Pro tip Explain the process to stakeholders before inevitable periods of underperformance.
Watch out Do not abandon the horizon because a short-term narrative moves markets against the forecast.
In the wild
Near a market bottom, depressed prices produce a high seven-year expected return even while sentiment remains fearful. The forecast encourages investors to add risk because the starting valuation offers strong compensation, although the realized return path may remain volatile.
→ Capital is added when long-term expected returns are strongest rather than after confidence has already recovered.
A pension fund calculates returns under normal mean reversion and under a regime where real cash rates remain near zero. The alternative improves near-term equity forecasts because less valuation normalization is assumed, but it lowers sustainable long-term returns across stocks and bonds.
→ The fund sees that persistent low rates do not solve its long-term spending problem.
Common mistakes
Demanding precise timing
The forecast estimates long-term return, not the date or shape of the path toward fair value.
Using only one regime
Treating historical normalization as certain hides the consequences of structural changes in rates or risk premiums.
Confusing relative and absolute value
The cheapest available asset may still offer an unattractive absolute return when the entire opportunity set is expensive.
Is it for you?
Best for
It is best for institutions and patient investors making strategic asset-allocation decisions across a full market cycle.
Not ideal for
It is not ideal for investors who need precise one-year forecasts or cannot tolerate extended periods of benchmark underperformance.
From the transcript
“the average of those was actually six and a half years, and so we converted to a seven year forecast.”
“our for our seven year forecast, the framework is pretty simple. You have valuation and you have growth and income”
“our traditional forecasts assume normal mean reversion, normal mean reversion to cash rates, term premium on top of that, and then equity was premium on…”
From the episode
Matthew Kadnar Talks About His Transition From Law to Finance
Matthew Kadnar