Skip to content

pred-2026-04-09-188

March 2026 US CPI (all items, year-over-year) prints below 3.0% when released April 10–11, 2026.

resolved · correct tier 1 economic trade monetary-policy
confidence 0.400
created
2026-04-09
resolves
2026-04-11
resolved
2026-04-11
outcome
0
brier
0.1600
base rate
0.45
meta-confidence
medium
Evidence for (7)
  • Favorable base effects: March 2025 likely had elevated YoY comparisons; March 2026 comparison becomes easier
  • Disinflation momentum: CPI trend through late 2025–early 2026 showed declining YoY rates as excess demand dissipated
  • Energy price moderation: Oil prices in 2Q 2026 have not sustained 2024–2025 highs, pulling down headline inflation
  • Tariff pass-through lags: Tariffs announced/imposed in late 2024–early 2025 take 2–4 quarters to fully flow into consumer goods prices; March print may precede peak pass-through
  • Demand softness: Economic growth moderation and labor market cooling limit pricing power across goods and services
  • Core services inflation deceleration: Shelter cost growth and other service inflation have begun cooling after multi-year stickiness
  • Expectation anchoring: Fed communication and market pricing suggest inflation path below 3.0% as credible baseline
Evidence against (6)
  • Tariff implementation timing: Early-2025 tariff announcements and phase-ins could show initial pass-through in March CPI
  • Wage growth persistence: Nominal wage growth remains above 3% in many sectors, supporting pricing power
  • Shelter stickiness: Housing services and rent inflation have resisted disinflation, especially in tight markets
  • Supply-chain disruption costs: Geopolitical shocks and tariff-driven logistics complexity may push input costs higher
  • Goods inflation reacceleration: If tariff pass-through accelerates on durable goods, headline inflation could surprise upward
  • Predictor confidence: Original assessment of 0.67 for ≥3.0% suggests meaningful weight on tariff pressures and demand strength

Reasoning chain

The original prediction anchors on tariff pass-through risk, expecting March 2026 CPI to exceed 3.0%. However, this assessment likely overstates the immediacy of pass-through and underweights disinflation momentum. The CPI YoY calculation benefits from easier base comparisons (March 2025 was higher inflation), energy prices have moderated, and tariff effects typically lag policy implementation by 2–4 quarters. The economic slowdown in early 2026 reduces demand pressures that would otherwise accelerate price increases. While tariff costs are real, they are not yet fully reflected in March data. The 0.67 confidence suggests high risk of ≥3.0%, but this may reflect overconfidence in rapid pass-through timing and underestimation of disinflation forces already underway.

Falsification criteria

CPI prints at or above 3.0% YoY = counter-claim is false. CPI prints below 3.0% YoY = counter-claim is true. Resolution determined by official BLS release of March 2026 CPI.

Brier breakdown

Calibration − resolution + uncertainty = Brier score. Lower calibration is better; higher resolution is better.

Post-mortem

Auto-resolved (falsified, confidence=0.97). Evidence: March 2026 CPI (all items, year-over-year) printed at 3.3%, well above the 3.0% threshold. The BLS released the data on April 10, 2026 as scheduled. Energy prices surged 10.9% (gasoline +21.2%), driven by the Iran conflict, pushing headline inflation to its highest annual rate since May 2024. Sources: https://www.cnbc.com/2026/04/10/cpi-inflation-report-march-2026.html; https://www.cnbc.com/video/2026/04/10/march-consumer-price-index-up-3-point-3-percent-year-over-year.html; https://www.bls.gov/news.release/pdf/cpi.pdf. Reasoning: The prediction required March 2026 CPI (all items, YoY) to print below 3.0%. BLS released March 2026 CPI at 3.3% YoY on April 10, 2026 — confirmed by both the official BLS PDF release and multiple news outlets (CNBC, Kiplinger). Per the falsification criteria, CPI at or above 3.0% = prediction is false. 3.3% clearly exceeds 3.0%, so the prediction is falsified.