Jeremy Grantham's Mean Reversion Approach to Investing

Jeremy Grantham, co-founder of GMO (Grantham, Mayo, & van Otterloo), centers his entire investment philosophy on the observation that asset class valuations tend to revert toward their historical long-run averages — and that this tendency creates predictable, profitable entry and exit points. He treats mean reversion not as a bet on a single stock, but as a framework for identifying when entire asset classes have drifted so far above or below their typical price levels that a statistical correction becomes highly probable.

The Foundation: Mean Reversion as a Law of Markets

Grantham’s premise is straightforward: over long stretches, prices oscillate around a fair value anchored to fundamentals (earnings, cash flow, replacement cost). An asset trading at 25× earnings when the historical median is 15× is not intrinsically superior — it is merely expensive. Conversely, one trading at 8× is not a bargain trap; it is likely underprice. Given enough time, gravity pulls prices back to the mean.

This differs from the efficient-market view, which holds that current prices already reflect all available information. Grantham argues that behavioral forces — herding, euphoria, panic, career risk — push prices into predictable extremes. Once those extremes are large enough, the reversion is almost mechanical.

Measuring Valuation Extremes

Grantham’s framework relies on quantitative metrics tailored to each asset class:

• U.S. equities: price-to-earnings, price-to-book, dividend yield relative to bond yields
• Emerging market equities: the same ratios, often historically cheaper than developed markets
• Bonds: yields relative to inflation expectations, credit spreads
• Real estate: cap rates, price-to-replacement-cost
• Commodities: prices relative to cost of production, inventory cycles

At GMO, these ratios are tracked for decades or longer. When a metric strays 2+ standard deviations from the mean, it signals an extreme. Grantham’s strategy is to stage into the despised asset class gradually as extremes deepen and rotate out of the beloved one.

The Timing Challenge: Why Being Early Costs

Grantham’s candid admission is that mean reversion works, but timing is merciless. During the late 1990s, he correctly identified U.S. large-cap equities as absurdly overvalued — but the valuation stretched for another 2–3 years before the 2000 crash. Funds that followed his call to underweight U.S. stocks looked foolish for years, even though the eventual outcome validated the thesis.

This lag is the reason GMO manages money in long-dated funds (often 7–10 year mandates) rather than quarterly vehicles. Quarterly investors fire managers too early. Mean reversion is a multi-year bet, not a quick trade.

The 2008 Financial Crisis and Beyond

When the housing bubble and subsequent credit collapse unfolded, Grantham’s framework shone. He had flagged U.S. real estate as drastically overvalued (prices detached from rents, cap rates at historic lows), and the reversion was brutal. Conversely, emerging market equities had been despised and cheap; the reversion worked in reverse — eventually underperforming, but offering attractive entry points.

More recently, Grantham warned in 2021 that U.S. equities, especially mega-cap growth stocks, had entered a “bubble” by historical standards. His core position: wait for reversion, rotate to less-loved pockets (small cap, value, international), and exploit the eventual correction.

How Mean Reversion Shapes Portfolio Construction

In practice, Grantham’s approach:

1. Rank asset classes by their deviation from historical norms. If equities are 2.5 standard deviations above the mean and emerging markets are 1.8 standard deviations below, the contrast is stark.

2. Size portfolio tilts to the magnitude of the extreme. A 3-sigma event warrants a bigger tactical bet than a 1.5-sigma one. But he avoids picking a precise peak or trough; instead, he layers in as extremes deepen.

3. Set reversion targets. Rather than predicting the stock price, he forecasts where valuations should go. If a market trades at 18× earnings and the historical mean is 13×, a reasonable 5–7 year target is 13–15× — and earnings growth could push absolute prices higher even as multiples compress.

4. Rebalance ruthlessly. As assets rise back toward fair value, he trims them and redeploys to the next cheap asset class. This requires resisting the herd, which is psychologically taxing.

The Contrarian Edge: Why It Works (and Doesn’t Always)

Mean reversion is powerful because it exploits a fundamental human bias: extrapolation. When a stock or sector has risen for years, investors assume the trend continues. Grantham bets against that. When a market is hated, investors assume despair is permanent. He buys.

Yet the edge erodes over time as more capital follows the rule. Additionally, mean reversion assumes the long-run “mean” remains stable. In rare cases — a technological disruption, a permanent shift in industry structure, a change in regulatory regime — the historical average is no longer relevant. Grantham’s framework is most reliable for broad asset classes (equities vs. bonds, U.S. vs. emerging) and less reliable for individual securities or disrupted sectors.

Allocation Discipline and Humility

Despite his system’s track record, Grantham has become more humble about its limits. In recent years, he has acknowledged that mean reversion cycles are lengthening, that passive flows distort prices in new ways, and that climate and demographic shifts may alter long-run averages. He still anchors to mean reversion but now layers in scenario analysis and tail-risk hedging alongside the core thesis.

See also