QQQ daily returns vs 10-year yield changes — correlation and variance explained (≈3y window)
Contrary to the shorthand that rising Treasury yields automatically “crush” growth stocks, the day-to-day link between QQQ returns and 10‑year yield moves over the past ~3 years is barely there: Pearson r = -0.0569 and the OLS regression explains just 0.32% of QQQ’s variance. Put simply, yield moves and QQQ daily returns move independently most of the time.
The test compares close‑to‑close QQQ returns to daily changes in the 10‑year, with a 60‑day rolling correlation and an OLS slope estimate; the fitted beta implies only a tiny negative return per 10 bps and is not statistically robust. The charts, full statistics and sensitivity checks that justify this conclusion are presented in the analysis below.
For QQQ over the past ~3 years, does the Nasdaq actually trade inverse to the 10-year Treasury yield the way the 'higher rates crush growth' narrative insists — is there a reliable negative correlation between daily QQQ returns and daily 10-year yield changes, and how much of QQQ's day-to-day variance does the rate move even explain? Thesis: the daily correlation is only weakly negative and accounts for almost none of QQQ's variance, so the 'yields up, growth down' reflex is far flimsier than the macro consensus claims.
How this was measured
Resampled QQQ minute bars to daily closes and computed close-to-close daily returns. Reindexed the 10-year Treasury constant-maturity yield onto the QQQ trading-day calendar with forward-fill, then took first differences to get daily yield changes (in percentage points; +0.10 = +10 bps). Measured Pearson and Spearman correlations between QQQ returns and daily yield changes, and fit an OLS regression of return ~ alpha + beta * dY to obtain the slope and R² (variance explained). A 60-trading-day rolling correlation traces stability over time. Window limited to the last ~3 years within available coverage (2023-07-17 to 2026-06-30).
The key numbers
Reading the numbers
Over 737 trading days the Pearson r is only -0.0569 — essentially zero — and R² is 0.0032417, meaning changes in the 10y explain about 0.3% of QQQ's day-to-day variance; the slope's p-value 0.1225 is not statistically significant.
The charts
The scatter is a broad cloud centered near (0,0) rather than a tight downward line, so there is no clear day-to-day tradeoff visually. The regression shows a tiny negative slope: the OLS beta implies about -0.00137296 in QQQ return for a 10 bp rise in the 10y, and Pearson r = -0.0569. With R² = 0.0032417 the yield move explains almost none of the daily return dispersion, and the beta p-value 0.1225 means the slope isn't statistically distinguishable from zero.
The 60-day rolling Pearson r bounces all over: across 678 windows it ranges from -0.5385 to 0.3244 (mean -0.0398), and the median rolling r is essentially zero at -0.00584349. That tells you there are short stretches with stronger negative correlation (and occasional positive pockets), but on balance the relationship is unstable and hovers near zero rather than showing a consistent, reliable negative link.
Per-calendar-year Pearson correlation (ret vs Δ10y)
| year | pearson_r | n_days |
|---|---|---|
| 2,023 | -0.3667 | 116 |
| 2,024 | -0.0401 | 250 |
| 2,025 | 0.1489 | 248 |
| 2,026 | -0.272 | 123 |
The takeaway
Short answer: no — over the last ~3 years there is only a very weak negative daily link between QQQ returns and changes in the 10‑year yield, and rate moves explain almost none of QQQ's day-to-day variance. Concretely, across 737 trading days Pearson r = -0.0569 (Spearman -0.0873), the OLS slope implies about -0.137% for QQQ per +10 bps in the 10y, and R² is just 0.32%. The slope is not statistically robust (p = 0.1225, roughly a 12‑in‑100 chance this effect is noise), and short‑window measures swing: yearly correlations ran from -0.367 (2023) to +0.149 (2025) while the median 60‑day r is about -0.0058. Practical takeaway: for daily trading or risk attribution the simple “yields up → growth stocks down” rule is flimsy — daily yield moves account for virtually none of QQQ’s volatility, though that doesn’t preclude stronger intraday or regime‑specific reactions around macro events.
The fine print
- 10y yields were forward‑filled onto the QQQ trading calendar; holidays and posting gaps were handled by ffill.
- Close‑to‑close returns can miss sharper intraday co‑moves around CPI/Fed prints; high‑frequency event analysis could show different patterns.
- Window covers ~3 years and includes shifting regimes; correlations are regime‑sensitive and can flip sign.
- Outlier days (large macro shocks) can materially influence correlations despite the Spearman check.