Methodology

How the economic-cycle reading is built.

The model is a structured interpretation of several imperfect signals. This page exposes its inputs, weights, phase bands, adjustments, and failure modes so readers can judge the result instead of treating it as a black box.

Model overview

VisionBoard exposes two related outputs. The cycle position combines six economic signals and maps the result to a descriptive phase. The health score applies a separate bounded scoring rubric, including jobs and the alternative-reality check. Keeping the two visible helps readers inspect both the cycle label and the underlying economic health assessment.

Interpretation

A phase label is a model output, not an official recession declaration, probability guarantee, market-timing instruction, or promise about future returns.

Cycle-position weights

The phase shown on the dashboard begins with this cycle-position model. Consumer sentiment, industrial production, and the policy rate contribute normalized historical signals. GDP growth, the yield curve, and the policy-rate/inflation relationship use disclosed threshold rubrics before the six components are combined.

Components of the dashboard cycle-position model
SignalWeightWhat it represents
Real GDP growth rubric25%Reported pace of aggregate economic growth.
10-year / 2-year yield-curve rubric25%Curve shape as a signal of financial conditions and recession risk.
Policy-rate / inflation relationship20%The interaction between effective federal-funds rates and consumer inflation.
Industrial-production normalized signal15%Industrial output relative to its recent history.
Consumer-sentiment normalized signal10%Household confidence relative to its recent history.
Policy-rate normalized signal5%An additional financial-conditions input.

Health-score weights

The separate health score uses the following nominal weights: real GDP growth 23%, yield curve 23%, the rate/inflation relationship 18%, payroll employment 14%, consumer sentiment 9%, industrial production 3%, and the alternative-reality check 10%. The current core fetch does not supply a payroll observation, so that slot is omitted unless another valid observation is provided. When any observation is unavailable or invalid, the health calculation normalizes over the valid weighted inputs that remain. That can reduce comparability between runs with different source coverage.

Displayed phase bands

The standard cycle coordinate is bounded from 0 to 100. Lower values indicate healthier expansion conditions; higher values indicate greater contraction stress.

0–<25

Expansion

Broadly constructive growth conditions.

25–<50

Boom

Strong activity with greater overheating potential.

50–<75

Recession

Meaningful contraction signals in the composite.

75–100

Depression

Extreme stress in this model—not an official historical classification.

The reality adjustment

The model includes a bounded stress measure for cases where broader readings are worse than selected headline readings. Only an adverse gap adds stress; a broader measure that is more optimistic does not create a bonus.

  • Labor: the modeled broad-unemployment estimate minus the official U-3 rate, above a 4.5 percentage-point baseline and scaled over the next 3 points. The broad estimate is not a retrieved U-6 observation.
  • Credit: the positive gap between the selected reality and official credit-stress readings, scaled over 2 points.
  • Inflation: the positive gap between the selected reality and official inflation readings, scaled over 2 points.

Available components are bounded from zero to one and averaged. At maximum stress, the adjustment can move the standard cycle coordinate 25 points in the adverse direction—one complete phase band—and reduce the health score by 25 points. For the separate growth check, it can reduce reported GDP growth by as much as 2.5 percentage points.

A stagflation label additionally requires both elevated inflation and stagnation: either an inverted curve or stress-adjusted GDP growth below 1%. Hot inflation alone is not treated as stagflation.

Freshness, revisions, and fallbacks

Economic releases arrive at different frequencies and are often revised after publication. Market observations can be delayed, unavailable, or sourced from a fallback provider. The dashboard should be read with its timestamps and source-state messages, especially around release windows.

If required inputs are missing, some views use reduced input coverage or a neutral fallback. A fallback is operational continuity—not new evidence. We do not treat it as equivalent to a fully refreshed reading.

Known limitations

  • The weights and thresholds are judgment-based, not a statistically proven forecast of returns.
  • Historical relationships can break during policy changes, shocks, and structural shifts.
  • Initial releases can differ materially from revised data; the model may change as the source changes.
  • A national composite can hide regional, sector, household, or company-level differences.
  • The model does not incorporate a reader’s goals, liabilities, time horizon, taxes, or tolerance for loss.

Questions or reproducible challenges are welcome through the corrections process.