Does CPI move the U.S. dollar?
Verdict
- Direction (release_day): Not Significant
- Volatility (release_day): Not Significant
- Volatility (5d): Not Significant
CPI moves Treasuries hard, but the dollar only mildly — about 1.3× a normal day (just short of significant). The dollar leaned slightly weaker on CPI days in 2023–2025, but that tilt is not a reliable rule.
Bottom line: For the dollar, CPI is a second-order event — far smaller than its effect on Treasury yields.
The Data
| Dimension | Horizon | Value | Baseline | Test stat | p-value | Verdict |
|---|---|---|---|---|---|---|
| Direction | release_day | -0.17% | 0.00% | -2.74 | 0.009 | Not Significant |
| Volatility | release_day | 0.31% | 0.25% | 1.26 | 0.059 | Not Significant |
| Volatility | 5d | 0.67% | 0.59% | 1.13 | 0.189 | Not Significant |
- Direction (release_day): Over 2023–2025 the dollar leaned down on CPI days (−0.17%, p=0.009 in-sample). But this is a single policy regime with no out-of-sample test — consistent with our standard, we do not treat the direction as a robust, repeatable rule.
- Volatility (release_day): ≈1.26× a normal day, just short of significant (p=0.06). CPI moves Treasuries far more (×1.75) than the dollar.
- Volatility (5d): No elevation over a week.
Methodology
- Events (N): 37
- Window: 2023-01-12 → 2025-12-18
- Baseline: Unconditional distribution of k-trading-day % returns
- Look-ahead protected: Yes
- Tests: ttest_1samp_vs_zero, bootstrap_vs_baseline
Caveats
- The in-sample downward tilt covers a single 2023–2025 regime; we do not certify it as robust.
- Compare with our CPI → 10-year Treasury study (a much larger, significant reaction).
- Surprise (vs. forecast) conditioning is not applied.
- Historical statistics for informational purposes only, not financial advice. Results may vary with sample, period, and baseline definition.
Source
- Nominal Broad U.S. Dollar Index, Federal Reserve Board via FRED (Tier A) — DTWEXBGS
- CPI release dates, U.S. Bureau of Labor Statistics (via FRED) (Tier A) — Consumer Price Index