Official data · Transparent rules

U.S. Business Cycle

Official indicators for growth, jobs, housing and financial conditions—organized with clear, transparent rules.

Loading official data Checking sources…

Independent-family breadth

Business-cycle pulse

Related indicators are averaged within families before each family receives one equal vote.

Loading—

Supportive—

Neutral—

Soft—

Weak—

Families—

Raw indicators loading…

Past-month comparison

Compare with one past month

Choose a month to compare current readings and states with observations through that month’s final day.

Month

Advanced data basis

Historical data

Known-at-the-time uses ALFRED vintages; unsupported series are shown unavailable rather than silently replaced.

Supportive Balanced Soft Weak Context only Unavailable

01Level

Where the latest transformed reading sits against up to ten preceding calendar years. The current observation is excluded from its own benchmark.

02Trend

How it changed over one month for daily/weekly series or three months for slower series.

03Speed

Whether that trend is improving, weakening, or staying steady.

04Overall

35% Level + 45% Trend + 20% Speed. All three must be available; lower-is-better indicators are reversed.

component = clamp((input − median) ÷ (1.4826 × median absolute deviation), −2, +2)

Labels: Supportive ≥ +0.35 · Balanced between −0.35 and +0.35 · Soft below −0.35 but above −1.05 · Weak ≤ −1.05. Defined-rule and context-only cards are not forced through this formula.

Read the machinery

Methodology & data rights

No black box

Three separate questions

Level asks where the indicator is now versus up to ten preceding calendar years. Trend asks which direction it moved, normally over three months for monthly data. Speed asks whether that trend is strengthening, weakening, or staying steady. The current observation is excluded from its own benchmark. Indicators with an official natural zero use that zero instead; for CFNAI-MA3, zero represents trend growth and the 0.70 scale reflects the Chicago Fed’s historically cited contraction threshold.

Broad family votes

The core pulse uses ten broad families—broad activity, employment, labor demand, income, production, consumption/sales, unemployment stress, housing, orders/business, and financial conditions. Related indicators are averaged within families before each family receives one equal vote. Raw indicator breadth is shown separately.

Honest historical comparisons

The month comparison uses observations through that month’s final day and shows current versus past readings directly. Revised history is the simple default. The advanced “Known at the time” basis uses reproducible ALFRED vintages only; unsupported series are left unavailable rather than silently replaced.

Published composites stay distinct

The Conference Board LEI and CEI use a manually maintained snapshot of the official monthly release. Their exact published readings are shown with attribution, kept separate from the Philadelphia Fed coincident index, and excluded from the OVCrypto breadth vote to avoid double counting.

Calendar dates, not array positions

YoY and multi-month changes match the actual target calendar period. If that period is missing, the transform is unavailable for that observation; it never silently substitutes the preceding month.

Defined thresholds stay defined

The real-time Sahm Rule is shown against its official 0.50 percentage-point trigger. It is a labor-market confirmation rule, not a forward recession probability or an NBER declaration.

Backdrop means context, not a vote

Inflation, money, policy rates, Treasury yields, inflation expectations, yield curves, the broad dollar, and oil are grouped in a separate Macro Backdrop. Their labels are mechanical descriptions of the published readings. They never receive a green or red score and never enter the business-cycle pulse.

FRED notice: This product uses the FRED® API but is not endorsed or certified by the Federal Reserve Bank of St. Louis. The current FRED Terms of Use apply to FRED-backed features; by using those features, visitors agree to those terms and the OVCrypto Terms of Use. Source-level copyright and usage restrictions still apply.

NBER shading: Recession bands are ex-post peak/trough context. They are not an input to the signal score and should not be read as a real-time declaration.

Conference Board snapshot: The LEI and CEI cards reproduce a small set of figures published in the official monthly release and link back to the release and technical notes. Snapshot history begins with the earliest month stored in this project and grows through the documented monthly maintenance workflow.

Status meaning: CURRENT means the latest provider request succeeded. STALE CACHE means a provider refresh failed and a labeled older cache is being shown. PROVIDER ERROR means no valid current response was available. UNAVAILABLE means the official source returned no usable observation.

Reload: The top Reload control requests the dashboard again. FRED/ALFRED observations are not stored in OVCrypto application-memory, Redis/database, browser, or CDN caches; other official providers may use short, labeled server caches to respect their request limits.