Skip to main content

​
Overview

Beyond Treasury yields, Whether ingests macro indicators to refine regime classification. These signals adjust the tightness and risk appetite scores when markets show stress beyond what the yield curve captures.

​
Data Sources

​
Bureau of Labor Statistics (BLS)

​
CPI_YOY
object
Series ID: CUUR0000SA0Endpoint: https://api.bls.gov/publicAPI/v2/timeseries/data/Calculation: Year-over-year percent changeUsage: Inflation pressure signal
​
UNEMPLOYMENT_RATE
object
Series ID: LNS14000000Endpoint: https://api.bls.gov/publicAPI/v2/timeseries/data/Calculation: Latest monthly valueUsage: Labor market health signal

​
Federal Reserve Economic Data (FRED)

​
BBB_CREDIT_SPREAD
object
Series ID: BAA10YEndpoint: https://fred.stlouisfed.org/series/BAA10YDescription: Moody’s Baa Corporate Bond Yield relative to 10-Year TreasuryUsage: Investment-grade credit stress signal
​
HY_CREDIT_SPREAD
object
Series ID: BAMLH0A0HYM2Endpoint: https://fred.stlouisfed.org/series/BAMLH0A0HYM2Description: ICE BofA US High Yield Option-Adjusted SpreadUsage: Junk bond stress signal; triggers tightness adjustments when ≥4.5%
​
CHICAGO_FCI
object
Series ID: NFCIEndpoint: https://fred.stlouisfed.org/series/NFCIDescription: Chicago Fed National Financial Conditions IndexUsage: Broad financial tightening signal; positive values add to tightness
​
VIX_INDEX
object
Series ID: VIXCLSEndpoint: https://fred.stlouisfed.org/series/VIXCLSDescription: CBOE Volatility Index (closing value)Usage: Equity market fear gauge; reduces risk appetite when ≥20

​
How Indicators Adjust Scores

​
Tightness Adjustments

Macro indicators can increase the tightness score when stress is detected:

​
High-Yield Credit Spread

// From lib/regimeEngine.ts:356
const HY_SPREAD_STRESS_THRESHOLD = 4.5;
const HY_SPREAD_TIGHTNESS_MULTIPLIER = 6;
const HY_SPREAD_TIGHTNESS_CAP = 15;

if (hySpread >= HY_SPREAD_STRESS_THRESHOLD) {
  adjustedTightness += Math.min(
    HY_SPREAD_TIGHTNESS_CAP,
    Math.round((hySpread - HY_SPREAD_STRESS_THRESHOLD) * HY_SPREAD_TIGHTNESS_MULTIPLIER)
  );
  contributors.push('HY OAS stress');
}
Example: If high-yield spreads are 5.5%, this adds (5.5 - 4.5) * 6 = 6 points to tightness (capped at 15).

​
Chicago Financial Conditions Index

// From lib/regimeEngine.ts:371
const CHICAGO_FCI_TIGHTENING_THRESHOLD = 0;
const CHICAGO_FCI_TIGHTNESS_MULTIPLIER = 10;
const CHICAGO_FCI_TIGHTNESS_CAP = 10;

if (chicagoFci >= CHICAGO_FCI_TIGHTENING_THRESHOLD) {
  adjustedTightness += Math.min(
    CHICAGO_FCI_TIGHTNESS_CAP,
    Math.round(chicagoFci * CHICAGO_FCI_TIGHTNESS_MULTIPLIER)
  );
  contributors.push('Financial conditions tightening');
}
Example: If NFCI is 0.5, this adds 0.5 * 10 = 5 points to tightness (capped at 10).

​
Risk Appetite Adjustments

Macro indicators can reduce risk appetite when investor caution spikes:

​
VIX Index

// From lib/regimeEngine.ts:386
const VIX_SHOCK_THRESHOLD = 20;
const VIX_RISK_APPETITE_MULTIPLIER = 1.5;
const VIX_RISK_APPETITE_CAP = 20;

if (vix >= VIX_SHOCK_THRESHOLD) {
  adjustedRiskAppetite -= Math.min(
    VIX_RISK_APPETITE_CAP,
    Math.round((vix - VIX_SHOCK_THRESHOLD) * VIX_RISK_APPETITE_MULTIPLIER)
  );
  contributors.push('Equity volatility shock');
}
Example: If VIX is 30, this subtracts (30 - 20) * 1.5 = 15 points from risk appetite (capped at -20).

​
VC Funding Velocity

// From lib/regimeEngine.ts:401
const VC_VELOCITY_SLOWDOWN_THRESHOLD = -5;
const VC_VELOCITY_RISK_APPETITE_MULTIPLIER = 1.5;
const VC_VELOCITY_RISK_APPETITE_CAP = 10;

if (vcVelocity <= VC_VELOCITY_SLOWDOWN_THRESHOLD) {
  adjustedRiskAppetite -= Math.min(
    VC_VELOCITY_RISK_APPETITE_CAP,
    Math.round(
      Math.abs(vcVelocity - VC_VELOCITY_SLOWDOWN_THRESHOLD) *
        VC_VELOCITY_RISK_APPETITE_MULTIPLIER
    )
  );
  contributors.push('VC funding slowdown');
}
Example: If VC velocity is -8%, this subtracts abs(-8 - (-5)) * 1.5 = 4.5 points (rounded to 5, capped at 10).

​
Tech Layoff Trend

// From lib/regimeEngine.ts:419
const LAYOFF_PRESSURE_THRESHOLD = 65;
const LAYOFF_RISK_APPETITE_MULTIPLIER = 0.6;
const LAYOFF_RISK_APPETITE_CAP = 12;

if (layoffs >= LAYOFF_PRESSURE_THRESHOLD) {
  adjustedRiskAppetite -= Math.min(
    LAYOFF_RISK_APPETITE_CAP,
    Math.round((layoffs - LAYOFF_PRESSURE_THRESHOLD) * LAYOFF_RISK_APPETITE_MULTIPLIER)
  );
  contributors.push('Tech layoff pressure');
}
Example: If layoff index is 75, this subtracts (75 - 65) * 0.6 = 6 points from risk appetite (capped at -12).

​
Policy Assessment Scoring

Whether uses a subset of macro indicators for its Policy Spec v1 scoring system:

​
Composite Tightness Score (CTS)

Combines base rate and BBB credit spread: 
// From lib/regimeEngine.ts:508
const cts = computeCompositeScore(
  baseRateZ ?? 0,
  POLICY_SPEC_V1.weights.baseRate,
  bbbZ ?? 0,
  POLICY_SPEC_V1.weights.bbbSpread
);
Weights:
  • Base rate: 0.7
  • BBB credit spread: 0.3

​
Composite Risk Appetite Score (RAS)

Combines curve slope and BBB credit spread: 
// From lib/regimeEngine.ts:509
const ras = computeCompositeScore(
  slopeZ ?? 0,
  POLICY_SPEC_V1.weights.curveSlope,
  bbbZ ?? 0,
  POLICY_BBB_WEIGHT_FOR_RAS
);
Weights:
  • Curve slope: 0.6
  • BBB credit spread: 0.4 (inverted)

​
Z-Score Normalization

All signals are normalized to z-scores before weighting: 
// From lib/regimeEngine.ts:493
const stats = computeRollingStats(value, history, fallbackMean);
const z = computeZScore(value, stats);
const directionalZ = directionalNormalize(z, direction);
const clippedZ = clipZScore(directionalZ, POLICY_SPEC_V1.zClip); // ±3 clip
Direction logic:
  • "HIGHER_IS_TIGHTER": Positive z-score indicates tightening (e.g., base rate, BBB spread, CPI, unemployment)
  • "HIGHER_IS_LOOSER": Positive z-score indicates loosening (e.g., curve slope)

​
Fetching and Caching

​
BLS API

Whether fetches BLS data via POST request: 
// From lib/macroSnapshot.ts:68
const response = await fetch('https://api.bls.gov/publicAPI/v2/timeseries/data/', {
  method: 'POST',
  headers: { 'content-type': 'application/json' },
  next: { revalidate: 900 }, // 15 minutes
  body: JSON.stringify({ seriesid: [seriesId], latest: true }),
});
Revalidate interval: 900 seconds (15 minutes)

​
FRED CSV

Whether fetches FRED data via CSV endpoint: 
// From lib/macroSnapshot.ts:41
const response = await fetch(
  `https://fred.stlouisfed.org/graph/fredgraph.csv?id=${encodeURIComponent(seriesId)}`,
  { next: { revalidate: 900 } }
);
Revalidate interval: 900 seconds (15 minutes)

​
Fallback to Snapshot

If any live fetch fails, Whether falls back to data/macro_snapshot.json: 
// From lib/macroSnapshot.ts:214
export const loadMacroSeries = async (fetcher: typeof fetch = fetch): Promise<MacroSeriesReading[]> => {
  try {
    return await buildLiveSeries(fetcher);
  } catch {
    return snapshotSeries; // Pre-loaded from macro_snapshot.json
  }
};
Snapshot data includes isLive: false and displays an “OFFLINE” badge in the UI.

​
Response Structure

​
id
MacroSeriesId
Enum: CPI_YOY | UNEMPLOYMENT_RATE | BBB_CREDIT_SPREAD | HY_CREDIT_SPREAD | CHICAGO_FCI | VIX_INDEX | ...
​
label
string
Human-readable name (e.g., “CPI Year-over-Year”)
​
value
number | null
Latest value in the series
​
unit
string
Enum: % | bps | x | index
​
explanation
string
Plain-English description of what the indicator measures
​
sourceLabel
string
Source attribution (e.g., “BLS public API”, “FRED”)
​
sourceUrl
string
Link to the original data source
​
formulaUrl
string
Link to documentation explaining the calculation
​
record_date
string
ISO date of the data point
​
fetched_at
string
ISO timestamp when Whether fetched the data
​
isLive
boolean
true if fetched from live API; false if using snapshot
​
history
SeriesHistoryPoint[]
Array of { date: string, value: number | null } for time-series charts

​
Boundary Contributors

When macro adjustments push scores across regime boundaries, Whether tracks the contributors: 
// From lib/regimeEngine.ts:434
return {
  tightness: adjustedTightness,
  riskAppetite: adjustedRiskAppetite,
  boundaryContributors: contributors, // e.g., ['HY OAS stress', 'Equity volatility shock']
  weakReadCount,
};
This list appears in the regime diagnostics panel to explain why the classification may be more severe than the yield curve alone suggests.

​
Weak Read Warnings

Whether counts how many macro adjustments are active: 
// From lib/regimeEngine.ts:142
const TWO_WEAK_READS_WARNING_COUNT = 2;

// In diagnostics:
twoWeakReadsWarning: weakReadCount >= TWO_WEAK_READS_WARNING_COUNT,
When two or more macro stress signals fire simultaneously, Whether displays a warning that conditions are deteriorating beyond normal yield curve patterns.

Build docs developers (and LLMs) love

Get started for freeTalk to us