Research Framework
The DLI Liquidity Score synthesizes structural data from the Federal Reserve system and capital markets through a four-stage quantitative pipeline:
Direct integration with FRED, US Treasury Fiscal Data, and NY Fed Markets APIs — automated ingestion with cross-frequency alignment
Each indicator is direction-aligned so higher means tighter; most DLI inputs use rolling 5-year tightness percentiles, while indicator pages still show 10-year median/MAD z-scores for magnitude context
The 10 scoring indicators are grouped into 4 transmission tiers: Policy/Reserves (65%), Funding/Plumbing (10%), Credit/Intermediation (5%), Risk/Price (20%). The policy tier scores Fed balance sheet, TGA, and a conditional ON RRP depletion-transition signal; ON RRP is no longer forced into a linear Net Liquidity sign and does not keep penalizing the score after a low balance has stabilized.
The four sub-indices are aggregated via a CISS-style quadratic form √(s′·(W∘Σ)·s) — W the outer product of tier weights, Σ a rolling 60-day correlation matrix: the composite amplifies when channels co-move and dampens when they offset. Regime is classified via rolling 5-year P20 / P80 thresholds; the former Group-A momentum kicker is retired. Important: DLI remains one headline score measuring the dollar-liquidity stance over weeks to months; it is not a market-stress or crisis detector.
The site tracks 12 indicators; 10 core ones form 4 tiers with group weights Policy/Reserves 65%, Funding/Plumbing 10%, Credit/Intermediation 5%, Risk/Price 20%. DLI remains one single headline score for the policy-and-plumbing stance over weeks to months. The policy tier scores Fed balance sheet, TGA, and a conditional ON RRP depletion-transition signal; ON RRP participates only when the balance has recently fallen sharply toward a low-buffer state, and exits the daily group average once a depleted balance has stabilized.
Net Liquidity and M2 are tracked for context. The DLI scores the underlying policy levers directly: Fed balance sheet, TGA, and a conditional ON RRP depletion-transition signal.
| Tier | Weight | Indicator | Tightening Signal |
|---|---|---|---|
| Policy / Reserves | 65% | Fed BS Size | ↓ Falling = Tighter liquidity |
| TGA Balance | ↑ Rising = Tighter liquidity | ||
| ON RRP | Conditional depletion signal: active only on recent sharp drawdowns | ||
| Funding / Plumbing | 10% | SOFR-IORB | ↑ Rising = Tighter liquidity |
| SRF Usage | ↑ Rising = Tighter liquidity | ||
| Credit / Intermediation | 5% | Cash Buffer | ↓ Falling = Tighter liquidity |
| HY Spread | ↑ Rising = Tighter liquidity | ||
| Risk / Price | 20% | VIX | ↑ Rising = Tighter liquidity |
| Dollar Index | ↑ Rising = Tighter liquidity | ||
| 10Y Real Yield | ↑ Rising = Tighter liquidity |
The DLI Score uses adaptive rolling 5-year percentile thresholds to classify liquidity into three states — loose, neutral, and tight — keeping signals stable and comparable across macro cycles.
| Percentile | Liquidity Condition | Risk Bias (Secondary) | Interpretation |
|---|---|---|---|
| ≤ P20 | Loose | Risk-seeking tilt | Loose liquidity — generally favorable for risk assets |
| P20 – P80 | Neutral | Balanced | Conditions in the mid-range — watch marginal shifts and drivers |
| ≥ P80 | Tight | Defensive tilt | Significantly tight liquidity — risk assets face strong headwinds |
Liquidity indicators carry 1-7 days of release lag, and the indicators themselves move daily. A regime label at t therefore does not constrain asset performance at t+20 — by then the DLI itself has moved into a different state.
Current State
DLI tells you whether dollar liquidity right now is loose, neutral, or tight. That is what it is built for.
Coincident Correlation (≤10d)
If DLI tracks real liquidity, it shows up as coincident correlation within ~10 days. Beyond that horizon, the indicators themselves have drifted.
No Forward Prediction
We deliberately do not show "average 20 / 60-day forward returns under each regime" — those numbers mix unrelated states and produce spurious signals.
Central-bank financial-conditions / liquidity indices (Chicago Fed NFCI, ECB CISS, etc.) are validated contemporaneously, not by forward returns. The DLI is the same kind of object — a policy-and-plumbing-driven coincident *stance* index — so the right test is coincident correlation (the asset view shows the 1 / 5 / 10 / 20-day decay), not forward backtests. To be clear: the DLI is not equivalent to a market-stress or crisis index — it measures how loose or tight the liquidity stance is, not the acute-stress events themselves.
This site is for informational and educational purposes only. It does not constitute financial advice. All data comes from public sources; we do not guarantee completeness or timeliness. Investment decisions should be based on personal research and professional consultation.