Swensen Portfolio: the Risk X-Ray
Yale's late endowment chief wrote one book for individual investors, and this is its allocation: six asset classes, every one chosen for a distinct economic role, with a 20% real-estate slice that no other portfolio in this catalog matches. The analysis below tests Swensen's core claim, that six genuinely different return streams diversify better than a larger pile of similar ones, against the measured correlations.
David Swensen, Unconventional Success (2005)
Analysis as of Friday 2026-08-28 · recomputed weekly from daily market data · methodology
What the Swensen Portfolio holds
The published allocation: 30% US stocks, 15% developed international, 5% emerging markets, 20% REITs, 15% long Treasuries, 15% TIPS.
| Fund | Measured role | Weight | Expense ratio |
|---|---|---|---|
| VTI Vanguard Total Stock Market ETF | correlated core | 30% | 0.03% |
| VEA Vanguard FTSE Developed Markets ETF | correlated core | 15% | 0.03% |
| VWO Vanguard FTSE Emerging Markets ETF | correlated core | 5% | 0.06% |
| VNQ Vanguard Real Estate ETF | risk driver | 20% | 0.13% |
| TLT iShares 20+ Year Treasury Bond ETF | diversifier | 15% | 0.15% |
| TIP iShares TIPS Bond ETF | correlated core | 15% | 0.18% |
Where the risk actually sits
Measured through the covariance matrix, VTI dominates the risk budget at 36.3% of total risk for 30.0% of capital; at the other end, TIP accounts for only 3.5%.
| Fund | Share of risk | Risk | Capital | Beta | Role |
|---|---|---|---|---|---|
| VTI | 36.3% | 30.0% | 1.01 | correlated core | |
| VNQ | 26.7% | 20.0% | 0.59 | risk driver | |
| VEA | 18.1% | 15.0% | 0.79 | correlated core | |
| TLT | 9.9% | 15.0% | 0.11 | diversifier | |
| VWO | 5.3% | 5.0% | 0.75 | correlated core | |
| TIP | 3.5% | 15.0% | 0.05 | correlated core |
How the pieces move together
VTI + VEA + VWO move as one block (average correlation 0.77): together they are 50.0% of the capital and 59.8% of the risk.
TLT + TIP move as one block (average correlation 0.86): together they are 30.0% of the capital and 13.5% of the risk.
| Holding | VTI | VEA | VWO | VNQ | TLT | TIP |
|---|---|---|---|---|---|---|
| VTI 30% | 1.00 | |||||
| VEA 15% | 0.77 | 1.00 | ||||
| VWO 5% | 0.71 | 0.83 | 1.00 | |||
| VNQ 20% | 0.55 | 0.56 | 0.42 | 1.00 | ||
| TLT 15% | 0.13 | 0.16 | 0.14 | 0.49 | 1.00 | |
| TIP 15% | 0.20 | 0.23 | 0.15 | 0.56 | 0.86 | 1.00 |
Pearson correlation of weekly returns · 156-week common window ending 2026-08-28 · pairbook.io
The worst it would have been
Over the analysis window, this mix would have fallen at worst -8.9% (from the week of 2023-09-15 to 2023-11-03), shallower than the S&P 500's -16.9% over the same weeks, and it has since recovered that peak.
Worst weeks of the mix: -5.0% (week ending 2025-04-11), -2.9% (week ending 2023-10-27), -2.9% (week ending 2026-03-13).
FAQ
Why such a large REIT allocation?
Swensen treated real estate as a third pillar between stocks and bonds, equity-like returns with an inflation link. Whether REITs still behave that way is measurable: their correlation with the stock sleeve, above, is the test.
Why TIPS and long Treasuries at the same time?
They protect against opposite scenarios: long nominals against deflationary shocks, TIPS against inflationary ones. Holding both is deliberate scenario coverage, and their near-zero mutual correlation is the design working.
Run this analysis on your own portfolio
The same X-ray, risk contributions, correlation blocks, ETF overlap, runs on any mix of 4,700+ US stocks and ETFs: paste your holdings into the free portfolio X-ray, or ask an AI assistant with the PairBook MCP server.
