Economic Breadth

How widespread is the economy's movement? Every member indicator casts one monthly vote — improving, abstaining, or deteriorating, judged against its own history. This page reports the count.

basis: revised live census · updated Sep 13

Net breadth
+33.3%
broad improvement
8 net of 24 members
Improving
10
moved beyond their own dead band
Abstaining
12
inside their own dead band
Deteriorating
2
moved beyond their own dead band

Each member counted at its latest monthly vote; the month record settles on the chart.

Growth is the default state: net breadth above +25 in 61.5% of months since 1963, below −25 in only 4.5% — improvement is diffuse and constant; broad deterioration is rare and concentrated.

Widest recorded: +100.0% (May 1978). Narrowest: -77.3% (Dec 2008).

Monthly census · 1963–2026

Data through: August 2026 · updated Sep 13

198719901992199519982001200320062009201220152018202020232026-90-4504590+25 · broad improvement (61.5% of months)−25 · broad deterioration (4.5% of months)
▬ net breadth▬ NBER recessions−− +25−− −25

Net breadth = improving minus deteriorating, as a share of the full electorate. Pink bands: NBER recessions. The last 40 years shown; expand the chart to reach the full record back to 1963.

Of the 34 months below −25, 94.1% were NBER recession months.

Not every broad deterioration ran into a recession: Feb 2003, Oct 2015, Feb 2025 deteriorated broadly with no recession following. Counting describes; it does not predict.

Breadth by regime

Regime record window: Aug 1981–Aug 2026 · 541 months

regimemonthsmedian net breadthrange
Healthy Expansion213+42.9%-5.9% to +86.7%
Soft Patch85+21.1%-22.7% to +57.1%
Contraction44-28.9%-77.3% to +18.2%
Overheating41+25.0%-12.5% to +80.0%
Early Recovery40+24.9%-28.6% to +86.7%
Cooling35+33.3%+4.2% to +53.3%
Disinflation22+20.8%0.0% to +60.0%
Transitional20+9.5%-22.7% to +45.8%
Below-Target Drift19+39.1%+16.7% to +50.0%
Unclassified14+10.1%-9.5% to +33.3%
Snapback8+62.5%+4.5% to +66.7%

Ranges overlap across regimes — a month's label does not fix its participation, and a participation reading does not name a regime. Label and participation are different facts.

Where the count and the regime label sit on the same side

Regime record window: Aug 1981–Aug 2026 · 527 months compared

Same side in 499 of 527 months (94.7%) · current run 16 months.

Both lean expansionary most months — net breadth is at or above zero in 88.6% of compared months and the regime label sits on the expansion side in 91.7% — so high agreement is partly expected; the disagreements are the informative months.

The 28 months they diverged

Mar 1982 · Jul–Aug 1982 · Oct 1982 · Sep 1990 · Apr–May 1991 · Mar 1993 · Nov 2000 · Jan 2001 · Jul–Aug 2001 · Dec 2001 – Jan 2002 · Feb 2003 · Jul–Dec 2007 · Feb 2008 · Jul–Aug 2009 · Nov–Dec 2019 · Jun 2022 · Feb 2025

Where the count and the label diverged — shown whether flattering or not.

24 voters, one per economic fact · deferred: existing home sales (2030), ISM mfg + services (~2031) · cell of record: 3-month window, p50 band · descriptive only — counts, never probabilities · methodology →