Correlation measures how often two assets move together, on a scale from -1 (always opposite) through 0 (unrelated) to +1 (always together) — and diversification's entire benefit lives in that number being meaningfully below 1. For two decades the negative stock-bond correlation made the classic 60/40 portfolio work; in 2022 both fell together as inflation shocked both, per index records, and the correlation's sign flipped for the first sustained period since the late 1990s. NewsJay publishes information and education, not investment advice, and past relationships never guarantee future ones.
How does the number translate into portfolio behavior?
When assets are imperfectly correlated, their swings partially cancel: portfolio volatility lands below the weighted average of the parts, which is the only free lunch finance offers. Two assets with +0.9 correlation barely diversify each other; two at 0.0 cut each other's noise substantially; two at -0.5 offset so reliably that the pair can hold more total risk per unit of outcome than either alone. This is why combining a total-stock fund with a total-bond fund — correlation historically near zero to negative across long periods — achieved more risk reduction than combining five equity funds, whose internal correlations typically run above 0.8.
Why did the stock-bond correlation change?
The sign depends on which shock dominates. When growth shocks dominate — recessions, financial crises — bonds rally as rate-cut expectations take hold, holding the negative correlation of 2000 through 2021. When inflation shocks dominate — as in the 1970s and again in 2022 — rising expected rates hit both assets at once, and the correlation turns positive. The regime is not random noise; it tracks which macro variable markets fear most, which means an allocator can understand the regime even without predicting it.
| Regime | Dominant shock | Stock-bond correlation | Portfolio experience |
|---|---|---|---|
| 2000-2021 typical | Growth | Negative | Bonds cushion equity falls |
| 1970s, 2022 | Inflation | Positive | Both fall together |
| Within equities, always | Any crisis | Rises toward 1 | Stock diversification weakens exactly when needed |
Does equity diversification fail in crises?
Partially, and predictably. Correlations among individual stocks and sectors rise sharply in selloffs — in 2008 nearly everything equity fell together, whether it sold cement or software. This is why equity diversification protects against company-specific risk but only dampens, never removes, market risk. The assets that historically kept offsetting behavior in equity crises — high-quality Treasuries most reliably — do so through the rate channel, which is exactly the channel inflation shocks disable.
What about assets sold as uncorrelated?
Skepticism is the correct prior. Correlation estimates on short samples are unstable, and assets marketed as diversifiers — gold, managed futures, various alternatives — have records that vary by episode: gold's stock correlation has swung widely across decades; cash-like instruments anchor but earn little. The durable discipline is holding a few genuinely different risk exposures, understanding the channel through which each offsets equities, and testing the story against at least one historical regime where it failed.
How should an investor use correlation practically?
Three working rules. Assume equity correlations converge in crashes when sizing any portfolio of stocks — overlap between equity funds is nearly all shared market risk. Treat the bond sleeve's diversification as regime-dependent, stronger in growth shocks, and size it accordingly rather than assuming it always cushions. And distrust any asset whose diversification claim rests on one decade of data — the 2010s taught investors to expect bonds to always hedge, and 2022 repriced that lesson in a single year.
How should investors monitor correlation over time?
Correlation is estimated, not observed — rolling three-year windows on asset-class returns are the standard lens, and the estimates move enough that a quarterly glance at the stock-bond spread is plenty for allocation purposes. The monitoring question worth asking annually is directional: is the bond sleeve still behaving as ballast in equity selloffs, a question the last two or three drawdowns answer directly from the investor's own statements. The person whose bonds fell alongside stocks in 2022 and cushioned 2020 has seen both regimes inside five years — a better education than any coefficient table, and the practical basis for sizing each sleeve.
What is the diversification argument for international equity revisited through correlation?
International stocks diversify US equity imperfectly — correlations with the US market have run high in crises — but imperfect correlation is not absent diversification, and the currency channel moves foreign holdings when domestic markets struggle most. The honest framing treats international equity as equity risk with a different packaging, whose diversification benefit concentrates in currency and policy cycles rather than in escaping global drawdowns. The allocation that follows is modest rather than heroic: enough international exposure to matter, sized so its tracking-error regret during US-led rallies never triggers the sell decision that converts imperfect diversification into a loss.
How did the 2022-2026 sequence test portfolio construction?
The inflation shock and its long tail gave investors a full seminar: 2022's simultaneous stock-bond decline, the recovery years' restored ballast as growth fears returned, and the 2026 energy-spike episodes testing both regimes again. Portfolios built assuming the 2010s' negative correlation were repriced in 2022; portfolios that sized bonds for both shock types — and held inflation protection at the margin — traveled the whole sequence without a thesis change. The period's lesson is the article's in miniature: correlations are regime statements, regimes rotate, and the construction that survives is the one that never required a single regime to hold forever.
What is the one-sentence summary?
Correlation tells you whether two assets fail together, and diversification only exists between assets that can fail separately — so the allocator's real job is knowing which shocks each holding answers to, rather than memorizing coefficients that will have changed by the time they matter.
FAQ
What correlation number is good for diversification?
Lower is better, with big gaps mattering most: below roughly 0.3 provides meaningful risk reduction, near zero is excellent, negative is the gold standard — regime permitting. Between two equity funds, expect 0.8 or higher, which is why fund count alone diversifies little.
Can I hedge inflation risk in a portfolio?
Partially: Treasury Inflation-Protected Securities adjust principal with CPI by formula, making them the direct instrument; commodities and real estate carry episodic inflation sensitivity with their own volatility. Stocks hedge inflation over very long horizons but poorly in inflation shocks themselves.
Why did my diversified portfolio fall in 2022 anyway?
Because the inflation regime pushed the stock-bond correlation positive — the cushioning channel was disabled, not your construction broken. Portfolios built for growth shocks need a separate answer for inflation shocks, even a modest one.
Is zero correlation the same as safety?
No. An asset uncorrelated to stocks can still lose money on its own; correlation describes co-movement, not standalone risk. Cash is near-perfectly safe and near-zero correlated; plenty of volatile assets share the correlation with far worse outcomes.
For more context, read The 60/40 Portfolio: What It Holds and Why It Endures.
For more context, read three fund portfolio.
For more context, read How Often Should You Rebalance a Long-Term Portfolio?.
