Jobs, inflation, and the Fed vs the bond market

Web-verified 2026-06-09. Data + sourcing behind docs/charts.html. All series annual, public (FRED/BLS/Treasury/BIS/BOJ), rounded; verify at the cited series. Overlay; not used in the proofs. The "Fed follows the 2Y, not the mandate" reading is a strong, labeled interpretation — the co-movement is fact, the intent is inference.

Two claims the charts test

1. A single "jobs number" and "inflation number" hide opposite trends — you must disambiguate to see the truth.
2. The Fed's policy rate tracks the bond market (the 2-year yield) far more tightly than its statutory dual mandate (2% inflation + maximum employment).

Disambiguating "employment"

• Headline payrolls overstated, revised late: the QCEW benchmarks cut ~1.5M jobs across 2024–25 (−818k, −911k), arriving 6–18 months after the prints that drove decisions.
• U-3 vs U-6: the headline ran 3.6–4.4%; U-6 (discouraged + involuntary part-time) ran ~3–5 pts higher (6.8–8.1%) — the slack "maximum employment" omits.
• Sectoral divergence: recent growth is concentrated in health care, leisure/hospitality, and state/local government, while manufacturing, information (tech), temp help, and federal government shrank. One aggregate masks opposite trends.
• Gig: full-time independents 13.6M → 27.7M; multiple-jobholders inflate the count (macro-gig-labor).

Disambiguating "inflation"

• CPI vs core PCE differ by weight/scope (PCE ~0.3–0.4 pt lower); neither sat at 2% for most of the decade (~0% in 2015, 4–8% in 2021–23, ~3% in 2024–26).
• CPI shelter (~⅓ of CPI) lags market rents ~1 year (ALNRI/Zillow/NTRI, macro-rent-cpi-divergence) — misstating turning points both ways.
• In gold, a US home is −81% since 1998 (macro-gold-silver-reprice) — "2% inflation" is an artifact relative to debasement.

The Fed follows the bond market — co-movement FACT; "responding to bonds" interpretation

• The funds target and the 2-year Treasury yield move together, and the 2Y turns first at every pivot: it fell below the funds rate ahead of the 2019 and 2024 cuts and rose ahead of/with the 2022 hikes. The Fed ratifies the market's expectation.
• If it steered to the mandate, the funds rate would track 2% + full employment — instead it sat at 0.25% through 3.6% core PCE (2021) and stayed ≥4.5–5.5% as core PCE fell to ~2.8% and U-6 rose (2023–25).
• The long end and the global bond market bind it too: US 10Y/30Y, Baa ~6%, and the 10Y JGB escaping YCC (0% → 2.0% Dec-2025 → ~2.66% 2026) — raising the cost of the yen carry trade that funds crowded longs (macro-carry-trades).
• Z3 (macro-fed-trap): there is no single feasible rate satisfying the divergent targets — so the mandate cannot be what sets the rate; financial conditions / the bond market do.

Rate differentials — sharp/fast vs smooth/delayed (measured)

Different spreads trade off noise against lead horizon (computed on the monthly cache, predicting the subsequent fed-funds change):

The horizon lengthens from the short end to the curve (1–8 months → 18+ months) — the fast-vs-delayed axis. Noise is not monotonic (the 3M−funds gap is the cleanest; the 2Y−funds and 10Y−3M are jumpiest). Use the right differential for the question: 3M−funds for the fastest clean read on an imminent move, 2Y−funds for the strongest read on the Fed's direction, the 2s10s curve for the smoothest (most delayed) cycle/recession read. Charts: docs/charts.html.

Bond-market breakdowns (charts page)

The differentials extend across the curve and the bond universe — sliced by geography (region → sub-region → country, GDP-weighted: US/CA, DE/GB/FR/IT, JP/AU → global) and type/quality (sovereign 10Y, corporate Baa/Aaa, household 30Y-mortgage, and corporate OAS by rating AAA/BBB/CCC). Plus 30Y term-premium differentials and credit spreads (Baa−10Y full history; HY/IG OAS recent). Where a licensed feed would be needed, accessible proxies are substituted (the licensed indices source from the same public tape anyway): corporate quality tiers are now charted from the real FINRA TRACE tape (the OAuth Query API corporateMarketBreadth/Sentiment — investment-grade vs high-yield vs convertibles advance/decline breadth and volume mix, monthly from the daily trading-day tape, via models/graph/fetch_tape.py); credit quality also via the free AAA→CCC rating ladder + EM-corporate OAS; state munis → per-state ETF distribution-yield time series (California CMF, New York NYF, national MUB, HY-muni HYD via the free Yahoo chart API — a real 10-yr series, not a snapshot), plus corporate-by-maturity (VCSH/VCIT/VCLT). The two remaining gaps both need a per-CUSIP feed: GICS-industry corporate OAS (TRACE file download mapped to SIC) and city/per-issuer munis (MSRB EMMA). The per-institution cut already exists in the repo via the FDIC BankFind API (models/graph/bank_exposure.py). Full provenance table on the charts page.

Bottom line

Disambiguation dissolves the official story, and the rate path maps onto the 2-year yield, not onto 2% inflation or full employment. The evidence supports the thesis: the Fed steers to the bond market / financial conditions, with the "dual mandate" as the public framing.

← Research index · structured data: macro-jobs-inflation-fed.json · macro-jobs-inflation-fed.md