Macro Compass

board as of 2026-08-03 · updated monthly · last refreshed 2026-08-17

No thresholds crossed.2 near-signal2 approaching1 inverted9 quiet9 context

read past issues

watchlist state board

business cyclenear-signal

Curve re-steepening after inversionnear-signal
The historical recession-imminent signal (valid with corroboration only, under the fiscal overlay)
present — 3m10y last inverted Oct-2025, now +0.79%
as of 2026-08-03 · US Treasury yield curve · 1982-2026
Sahm rule (real-time)not crossed
Trigger = recession likely already begun (confirmation only — never initiates)
0.07, falling
as of 2026-06-30 · BLS unemployment data (Sahm-rule computation) · 1970-2026
Yield curve 2s10s + 3m10ycontext
Graded through fiscal overlay, never alone
3m10y +0.79% / 2s10s +0.45%
as of 2026-08-03 · US Treasury yield curve · 1982-2026

debt & fiscal pressurenear-signal

Deficit as % of GDP vs. unemploymentnear-signal
>6% deficit with <5% unemployment = fiscal dominance regime
5.6-5.8% deficit / 4.2% unemployment
as of 2026-06-30 · US Treasury / BLS
Interest expense as % of federal tax receipts (net, trailing 12m)not crossed
Sustained >25% = pressure zone (net-interest basis)
19.5% net basis (gross would read 25.1%; NIPA-fallback 33.4%), Jun-26 MTS
as of 2026-06-30 · US Treasury, Monthly Treasury Statement
10-yr real yield (TIPS)not crossed
Persistent negative = repression operating
+2.47%
as of 2026-07-31 · US Treasury real yield curve · 2003-2026
Treasury auction tails / bid-to-cover / dealer take-downnot crossed
Severely tailed = tail >=3bp on a 10/30-yr with below-median bid-to-cover, twice within a quarter (operational definition, indicators.md)
Jul 10-yr and 30-yr both stopped through; dealers 7-9%
as of 2026-07-31 · TreasuryDirect auction results
Term premium (ACM 10-yr)context
Rising = fiscal risk priced; capped despite fundamentals = repression
0.84%, +33bp in July
as of 2026-07-31 · Federal Reserve Bank of New York, ACM term premium
Net interest / revenues, actual (context)context
none (context row)
18.5% (FY2025)
as of 2025-09-30 · Congressional Budget Office
Net interest / revenues, CBO baseline path (context)context
none — projection row, split from the actual row so it can never render as observed
25.8% (FY2036, Feb-2026 baseline)
as of 2026-02-28 · Congressional Budget Office

inflation & real economyapproaching

Labor productivity, 4-qtr trend (BLS)approaching
>=2.5% sustained 2 yrs on the 4-qtr average = escape path confirming (clock starts when the average crosses)
Q1-26 +2.80% (first quarter above the bar); 4-qtr avg 2.46%
as of 2026-03-31 · BLS Productivity and Costs · 1996-2026
Core goods CPI vs. pre-2020 trendnot crossed
Return to persistent goods deflation weakens T07
+0.79% y/y (Jun-26)
as of 2026-06-30 · BLS Consumer Price Index · 1996-2026
Wage growth (Atlanta Fed tracker) vs. unemploymentnot crossed
Phillips curve steepening
3.6% wage growth / 4.2% unemployment
as of 2026-06-30 · Federal Reserve Bank of Atlanta Wage Growth Tracker
Government + gov-adjacent share of job growthcontext
Bifurcation gauge (no numeric threshold)
56.2% of Apr-Jun net gains
as of 2026-06-30 · BLS Current Employment Statistics

dollar system~stirring

estimated distance to a reserve-and-settlement regime break, read across the practice's full dollar-system watch

Foreign official share of Treasury market (TIC)approaching
Sustained decline = recycling reversing (context only)
~12.5% (May-26) vs ~13.6% a year ago; level flat y/y
as of 2026-05-31 · US Treasury TIC
China UST holdings (incl. Belgium/Euroclear custody)not crossed
Accelerating decline
$1,131B combined, -1.5% y/y
as of 2026-05-31 · US Treasury TIC
TIC foreign net purchases of US equitiesnot crossed
4+ quarters of selling = reallocation real
+$903.8B over 12m to May-26; one negative month
as of 2026-05-31 · US Treasury TIC
Fed swap-line balancesnot crossed
Any material usage = dollar-shortage phase live (weekly cadence in stress; polarity qualifier v7 — a qualifying operation supplies dollars)
$132M — de minimis
as of 2026-08-03 · Federal Reserve H.4.1

trade & reshoring~nearing

policy leg: estimated distance to a durable trade-regime restructuring; the buildout leg reads cooler and is tracked separately

Manufacturing construction spendinginverted
The reshoring fact on the ground
$172.7B SAAR, -21.4% y/y, five straight declines
as of 2026-06-30 · Census Bureau construction spending
World trade % GDP vs 2008 peak (context)context
none — context gauge (plateau is fact, collapse is not)
56.7% (2024) vs 60.3% (2008)
as of 2024-12-31 · World Bank World Development Indicators

liquiditywatching

Fed net liquidity (balance sheet - TGA - RRP)context
Direction governs risk-asset tide
$5,825B; +$108B/3mo but -$162B in late July (TGA rebuild, RRP exhausted)
as of 2026-08-03 · Federal Reserve H.4.1

demographics & generational turnover~nearing

estimated distance to the demographic regime turn becoming operative: labor scarcity and the generational handoff, read together

Cohort asset shares (Fed Distributional Financial Accounts)context
Boomer share drawdown pace
Boomer net-worth share 51.6% (sprint extraction)
as of 2026-08-03 · Federal Reserve Distributional Financial Accounts
Next RMD-schedule waypoint (static statutory table, context)context
none — static table pointer
one-cohort-per-year through 2032; statutory lull 2033-34; last onset 2039
as of 2026-08-13 · Congressional Research Service

cycle base rates

Expansion age
NBER Business Cycle Dating Committee chronology (trough Apr-2020)
76 months (Aug-2026), from the Apr-2020 trough — longer than 8 of 12 postwar expansions, ~1.5x the 51.5-month postwar median
reference Aug-2026 · NBER business-cycle chronology
Term-spread probability
NY Fed, "The Yield Curve as a Leading Indicator" (published model output)
~20% at one significant figure for Jul-2027 (published model value 15.2%, from the Jul-2026 monthly-average spread +0.78); Estrella-Mishkin table cross-check ~10% at 4q; spot 3m10y +0.81 (2026-08-12)
reference Jul-2026 spread month (probability target Jul-2027) · Federal Reserve Bank of New York recession-probability model

Priors inform context, never triggers.

release calendar

2026-09-03Productivity & Costs, Q2 revised (BLS)BLS
2026-09-04Employment Situation, August (BLS)BLS
2026-09-11CPI, August (BLS)BLS
2026-09-30Personal Income & Outlays / PCE, August (BEA)BEA
2026-10-02Employment Situation, September (BLS)BLS
2026-10-14CPI, September (BLS)BLS
2026-11-05Productivity & Costs, Q3 prelim (BLS)BLS

4 occurred since the last snapshot · 4 more scheduled through 2026-12-04

2026-08-06Productivity & Costs, Q2 prelim (BLS)BLS
2026-08-07Employment Situation, July (BLS)BLS
2026-08-12CPI, July (BLS)BLS
2026-08-26GDP Q2 second estimate + Personal Income & Outlays / PCE, July (BEA)BEA
2026-11-06Employment Situation, October (BLS)BLS
2026-11-10CPI, October (BLS)BLS
2026-11-25Personal Income & Outlays / PCE, October (BEA)BEA
2026-12-04Employment Situation, November (BLS)BLS

about this board

This board is a monthly snapshot of a personal macro monitoring practice: a small set of indicators tracked against explicit thresholds, a handful of cycle base rates for context, and a calendar of the data releases that can move them.

Chip legend:

confirmed: threshold crossed and confirmed
near-signal: live but unconfirmed
approaching: moving toward its threshold
not crossed: quiet
inverted: moving against expectation
context: gauge, no crossing semantics
stale / not recorded: no current reading

Each trend group carries a proximity gauge. Colored gauges are measured: the most advanced state among the group's rows, stepping from quiet through approaching and near-signal to confirmed. It reads like a storm scale, not a forecast: it says how close the nearest tripwire is, not whether it will trip.

Gray gauges marked with a tilde (~distant, ~stirring, ~nearing, ~at hand) are estimated: an audited judgment of how close the group's trend is to turning, formed across the whole practice, including tripwires whose sources have not cleared the bar for republication here, and revisited quarterly. The line under the group name says which dimension of the trend the estimate reads. An estimate is a considered guess with a paper trail, not a measurement; when the practice cannot form one honestly, the gauge says "watching" instead.

Sparklines show each indicator's underlying public series at monthly or quarterly resolution; the row's reading is a dated observation and can differ from the line's last point in vintage and basis. Both curve rows chart the 3m10y spread. Series shown are the producers' published data, retrieved via FRED; the Sahm rule series is computed by the Federal Reserve Bank of St. Louis.

Each row names its public primary source; readings trace back to that source's own published release. New York Fed–sourced readings are © 2026 Federal Reserve Bank of New York, content from the New York Fed subject to the Terms of Use at newyorkfed.org; Atlanta Fed–sourced readings are reproduced with attribution to the Federal Reserve Bank of Atlanta. Neither institution endorses this board. Readings and thresholds are research bookkeeping, not signals or advice — nothing on this board should be read as a buy, sell, or hold recommendation.

Everything here is general information and personal research, not investment advice.