thetrading.tools
Market QuestionsEconomy · 9 gauges · daily since February 1994 · as of October 7, 2026

Is a recession coming?

No single indicator calls recessions reliably. This page reads nine of the most-watched, each dated by the day it was published so the history only uses what was known at the time: the yield curve, the Sahm rule, jobless and continuing claims, the leading index, building permits, real retail sales, industrial production and corporate credit spreads.

Is a recession coming?2026-10-07
NO SIGN
On October 7, 2026, 9 of the 9 gauges read clear and 0 read warning.
clear
9
of 9 gauges
watch
0
of 9 gauges
warning
0
of 9 gauges

No sign when clear gauges outnumber warning ones by 3 or more; warning when warning ones lead by 3 or more; watch in between.

Latest read

No sign. On October 7, 2026, 9 of the 9 gauges read clear and 0 read warning. All 9 gauges have read clear on 577 sessions since February 1994, in 10 separate stretches. From days with no recession in the prior year, no sign verdicts were followed by a recession within 12 months 8% of the time, against 11% for all such days.

Sources, methodology & freshnessThe site's tools: Yield curve, 10-year minus 3-month, Sahm rule, Jobless claims, Continuing claims, change over a year and more · Daily after the US market close.Data as of 2026-10-07 · Open ↓
Source
The site's tools: Yield curve, 10-year minus 3-month, Sahm rule, Jobless claims, Continuing claims, change over a year and more
Methodology
9 gauges, each clear, watch or warning by a fixed rule; verdict by count (net of 3 either way); replayed daily since February 15, 1994
Updates
Daily after the US market close.Data as of 2026-10-07
Maintained & reviewed by Yuriy Matso — methodology shown on the page.
01

The 9 gauges

Reading clear: Yield curve, 10-year minus 3-month; Sahm rule; Jobless claims; Continuing claims, change over a year; Leading economic index, 6-month change; Building permits, change over a year; Real retail sales, change over a year; Industrial production, change over a year; Corporate credit spread. Each rule was fixed on round numbers near the gauge’s quartiles before any forward result was examined.

GaugeReadingRank since 1994Clear ifWarning ifState
Yield curve, 10-year minus 3-month+1.06 pts43rd pctile≥ +0.5 ptsinverted (below 0)clear
Sahm rule0.0050th pctilebelow 0.30.5 or moreclear
Jobless claims (4-week) above their 52-week low+0.5%20th pctile≤ +10%≥ +20%clear
Continuing claims, change over a year−10.4%28th pctile≤ +5%≥ +15%clear
Leading economic index, 6-month change+0.5173rd pctilerisingdown 1 point or moreclear
Building permits, change over a year+4.2%54th pctile≥ 0≤ −15%clear
Real retail sales, change over a year+2.5%57th pctile≥ +2%≤ 0clear
Industrial production, change over a year+1.4%40th pctile≥ +1%≤ −1%clear
Corporate credit spread (Baa minus 10-year)1.48 pts4th pctile≤ 2.0 pts≥ 3.0 ptsclear
Rank is today’s reading against every session since February 15, 1994 (0 lowest, 100 highest). Economic series are dated by the day they were published, so each past day only uses what was known then; values are today’s revised data.
02

The verdict every day since 1994

The line is clear gauges minus warning ones, from +9 to −9; the shaded bands are the no sign and warning verdicts. Since February 1994 the verdict has been no sign on 76% of sessions, watch on 17% and warning on 8%.

Range:
NO SIGNWARNING-9-30391995200020052010201520202025SPY+9
Net count of the 9 gauges, every session from February 15, 1994 to October 7, 2026. The grey line is SPY on its own scale. The same scoring produces today's verdict, so the last point is the answer card.
03

How unusual is today's reading?

All 9 gauges have read clear on 577 sessions since February 1994, in 10 separate stretches. The earliest began in March 1994 and the most recent in August 2026.

04

What followed

The test is whether a verdict was followed by a new recession, counted from days with no recession in the prior year. Warning signs have raised the odds well above the base rate, and they also gave false alarms; the record covers only three recessions, so treat the rates as rough. From days with no recession in the prior year, no sign verdicts were followed by a recession within 12 months 8% of the time, against 11% for all such days.

ConditionSessionsRecession within 12 monthsSPY next year
Every day8,21511.3%n=6,681+13.1%83% up
No sign verdict6,2108.4%n=5,814+12.7%83% up
Watch verdict1,37631.3%n=828+15.3%83% up
Warning verdict62920.5%n=39+24.8%87% up
Share of sessions followed by an NBER recession month within the next 12 months, counted only from sessions with no recession month in the prior 12, and the median SPY return over the next 252 sessions. Overlapping daily windows, so neighbouring days are not independent.
First warning verdict, first dayCountNext monthNext quarterNext year
January 17, 20014 of 9−0.1%−7.1%−15.1%
April 10, 20085 of 9+2.1%−7.9%−36.9%
April 15, 20206 of 9+2.6%+15.9%+49.7%
May 18, 20234 of 9+4.3%+4.1%+26.4%
SPY price return from the first qualifying session after 63 sessions without one. Open windows have not finished yet.

How the Economy Verdict Works

  1. 1
    Read 9 gauges
    Each row comes from one of the site's tools and links to it: Yield curve, 10-year minus 3-month; Sahm rule; Jobless claims (4-week) above their 52-week low; Continuing claims, change over a year; Leading economic index, 6-month change; Building permits, change over a year; Real retail sales, change over a year; Industrial production, change over a year; Corporate credit spread (Baa minus 10-year).
  2. 2
    Give each one a state from a fixed rule
    Every gauge reads clear, watch or warning by a rule printed beside it. The rules were set on round numbers near each gauge's quartiles before any forward result was examined.
  3. 3
    Count, do not blend
    The verdict is a count of clear and warning gauges, never a weighted score, so every part of it can be checked against its own page.
  4. 4
    Replay it and test it
    The same scoring runs on every session since February 1994, so the chart and today's verdict cannot disagree, and what followed each verdict is compared with every day in the window.

Who Uses the Economy Verdict

Anyone asking the question
One dated answer, with every gauge behind it and a link to each.
Traders
A count that changes before the label does: watch how many gauges move, not only the verdict.
Anyone checking a claim
The rules and the history are on the page, including what followed past readings.

Pro Tips

01
Watch the count
A move of two or three gauges says more than the verdict label, which needs a lead of the threshold to change.
02
Open the gauges
Each row links to its own tool, with its full history and method.

Common Issues & Solutions

A gauge here differs from its own page▾
Some tools show several versions of a series; this page uses the one named in the row, measured on the date shown.

Frequently Asked Questions

What are the best recession indicators?▾
The yield curve, the Sahm rule, jobless claims and leading indexes are the most cited. Each has given false alarms, which is why this page counts nine of them.
How is a recession officially defined?▾
In the US, the National Bureau of Economic Research dates recessions after the fact, as a significant decline in activity spread across the economy and lasting more than a few months. Its dates are what the study on this page is graded against.
How is the verdict on this page decided?▾
Each of 9 gauges reads clear, watch or warning by a fixed rule shown beside it. The verdict is no sign when clear gauges outnumber warning ones by at least 3, warning when warning ones lead by at least 3, and watch in between.
What is the answer today?▾
No sign. On October 7, 2026, 9 of the 9 gauges read clear and 0 read warning.

Explore Other Tools

Last updated: 2026-10-07