Decision brief · Regime risk
Why diversification can disappear when it is needed
Stock–bond correlation changed from negative to positive in the studied sample, reducing the protection expected from a balanced allocation.
Depth 1 · Answer
A balanced portfolio's protection depended on a correlation regime that changed sign.
In the frozen sample, 126-day SPY–TLT correlation averaged −0.35 during 2015–2019, reached +0.37 during 2022–2023, and remained positive at +0.31 at the sample end. In a separate static stress, pushing all asset correlations toward one raised the studied 60/40-plus-alternatives portfolio's annualized volatility from 11.6% to 15.8%.
Published 8 August 2026 · Underlying investigation last updated 3 August 2026
- SPY–TLT · 2015–19
- −0.35
- SPY–TLT · 2022–23
- +0.37 peak
- Portfolio vol · observed corr.
- 11.6%
- Portfolio vol · corr. → 1
- 15.8%
Why it matters
Multi-asset allocators and enterprise risk teams
The finding matters wherever historical correlations determine risk budgets, hedges, capital allocation, retirement glide paths, or the expected protection of a balanced portfolio.
Decision affected
Whether historical diversification deserves to be treated as insurance
Do not size protection from one estimated covariance matrix alone. Ask what happens if the macro driver changes and historically distinct sleeves begin sharing the same loss factor.
Depth 2 · Evidence
Two views expose the same hidden assumption
The first view is historical. A rolling 126-trading-day correlation between SPY and TLT captured the negative stock–bond relationship of 2015–2019 and its reversal during the 2022–2023 inflation shock. The values describe this sample; they do not identify the next regime.
The second view is counterfactual. The stress held individual asset volatilities fixed and blended the estimated correlation matrix toward perfect positive correlation, repairing the matrix to remain positive semidefinite. At the observed structure, the strategic portfolio had 11.6% annualized volatility. At the perfect-correlation limit, it reached the 15.8% weighted average volatility of its components—about four percentage points of diversification removed.
These are complementary lenses. History shows that a load-bearing relationship moved; correlation stress shows the portfolio consequence without claiming when or why the endpoint will occur.
Depth 3 · Application
Budget for the failure of relationships, not only assets
- Show portfolio risk across a range of correlation states instead of one point estimate.
- Identify which pairwise relationship contributes most to the diversification benefit.
- Pair covariance-based risk with named macro scenarios that explain why correlations might change.
- Monitor rolling relationships without treating the latest window as a forecast of the next regime.
- Describe “defensive” assets conditionally: protection against one shock may share exposure to another.
Depth 4 · Limits
A stress surface is not a timing model
- The SPY–TLT result comes from one historical sample and one 126-day rolling-window choice.
- The stress preserves individual volatilities while changing correlations; real crises can change both.
- Blending every correlation toward one is deliberately stylized, not a forecast of a specific crisis.
- The portfolio is one fixed strategic allocation with monthly rebalancing and liquid ETF proxies.
- The analysis does not establish that positive stock–bond correlation will persist.
Depth 5 · Method and code
Interrogate the portfolio through several non-overlapping lenses
Related research