US Government Shutdown Liquidity Tracker

Optimal timing: Wednesday evenings or Thursday mornings (after weekly TGA/reserves data release)

Quick Start

Basic Usage (Current Shutdown Analysis)

python scripts/analyze_shutdown.py --output results.json
python scripts/visualize.py results.json --output chart.png

This analyzes the 2025 shutdown (Oct 1 - present) with default settings.

Custom Date Range

python scripts/analyze_shutdown.py \
  --start-date 2018-12-22 \
  --baseline-date 2018-12-15 \
  --end-date 2019-01-25 \
  --output results_2018.json

Output Format

The analysis produces:

1. JSON data file containing:

   • Raw daily data (EFFR, SOFR)
   • Weekly data (TGA, reserves)
   • Key time points (baseline, shutdown start, TGA peak, latest)
   • Liquidity status assessment (EASING/TIGHTENING/STABLE/MIXED)

2. Visualization chart (PNG) with three panels:

   • TGA vs Bank Reserves (dual-axis weekly data)
   • EFFR vs SOFR (daily rates)
   • SOFR Premium over EFFR (liquidity stress indicator)

3. Structured conclusion:

   • Current status (e.g., "EASING")
   • Explanation (e.g., "TGA releasing, reserves recovering")
   • Key metrics vs baseline and peak

Core Analysis Logic

The Transmission Mechanism

Government Shutdown
    ↓
Federal spending stops (but revenues continue)
    ↓
TGA accumulates at Federal Reserve
    ↓
Bank reserves drain (mechanical Fed balance sheet effect)
    ↓
Liquidity scarcity → SOFR premium expands
    ↓
"Stealth tightening" (市场实际融资成本 > Fed政策意图)

Status Determination

The script classifies liquidity conditions into four states:

EASING (压力缓解):

• TGA falling >$10B from peak
• Reserves rising >$10B from trough
• Indicates: Shutdown ending or fiscal spending resumed

TIGHTENING (压力加剧):

• TGA rising >5% from baseline
• Reserves falling >2% from baseline
• Indicates: Shutdown's stealth tightening effect persists

STABLE (相对稳定):

• TGA/reserves changing <$20B from peak
• Indicates: Liquidity conditions steady

MIXED (复杂信号):

• Conflicting signals require continued monitoring

Key Metrics

SOFR Premium = SOFR - EFFR (in basis points)

Interpretation guide:

• 0-5 bps: Normal conditions
• 5-15 bps: Moderate stress
• 15-30 bps: Significant stealth tightening
• >30 bps: Acute crisis (may trigger Fed intervention)

Historical Context

For detailed historical analysis, see references/historical_cases.md.

Summary:

Critical insight: The transmission efficiency depends on reserve abundance. In QE environments with ample reserves, shutdowns don't affect markets. In QT or high-rate environments with scarce reserves, shutdowns create measurable tightening.

Data Sources

All data sourced from Federal Reserve Economic Data (FRED) API:

• TGA (WTREGEN): Treasury General Account balance, weekly
• Bank Reserves (WRESBAL): Total reserves, weekly
• EFFR (EFFR): Effective Federal Funds Rate, daily
• SOFR (SOFR): Secured Overnight Financing Rate, daily

For technical details on data series, update schedules, and interpretation, see references/data_sources.md.

Important: TGA and reserves update weekly on Wednesdays. For most current analysis, run this skill on Wednesday evenings or Thursday mornings.

Workflow for User Requests

Scenario 1: "What's the latest on the shutdown liquidity situation?"

1. Run analyze_shutdown.py with defaults (2025-10-01 start)
2. Generate visualization
3. Present:
   • Current status (EASING/TIGHTENING/etc.)
   • Latest metrics (TGA, reserves, SOFR premium)
   • Brief comparison to peak stress point
   • Conclusion statement

Scenario 2: "Compare this to the 2018 shutdown"

1. Run analysis for both periods:
   • 2025: Oct 1 - present
   • 2018-19: Dec 22, 2018 - Jan 25, 2019
2. Generate both charts
3. Present side-by-side comparison:
   • TGA accumulation magnitude
   • Peak SOFR premium
   • Fed intervention (if any)
   • Monetary environment context
4. Reference historical_cases.md for detailed context

Scenario 3: "Is the situation getting better or worse?"

1. Run analysis
2. Focus on:
   • Trend from TGA peak to latest (is TGA releasing?)
   • Reserves recovery from trough
   • SOFR premium vs baseline
3. Present trend assessment with clear directional language
4. Optionally show week-over-week changes

Output Presentation Best Practices

1. Lead with conclusion: State status (EASING/TIGHTENING) upfront
2. Show key metrics concisely:

   TGA: $941B (-$17B from peak)
   Reserves: $2,863B (+$15B from trough)
   SOFR Premium: 4 bps (vs 19 bps peak)
   

3. Visualize: Always include chart for complex cases
4. Contextualize: Reference historical episodes when relevant
5. Avoid jargon overload: Explain "stealth tightening" simply if user seems unfamiliar

Advanced Usage

Custom Baseline

When analyzing a specific episode, set an appropriate pre-shutdown baseline:

python scripts/analyze_shutdown.py \
  --start-date 2025-10-01 \
  --baseline-date 2025-09-24 \
  --end-date 2025-11-07

The baseline should be ~1 week before shutdown starts (to capture "normal" conditions).

Monitoring Routine

For ongoing tracking:

1. Weekly check (Wednesdays/Thursdays):

   • Run analysis
   • Note status changes
   • Update user if significant shift

2. Event-triggered checks:

   • Shutdown announcement → Start tracking
   • SOFR premium spikes (>15 bps) → Generate alert
   • Fed intervention (SRF usage) → Document
   • Shutdown resolution → Final analysis

Limitations and Caveats

1. Weekly data frequency: TGA/reserves only update weekly, limiting real-time precision
2. Month/quarter-end effects: SOFR naturally spikes at period-ends (unrelated to shutdowns)
3. Other liquidity factors: QT, regulatory changes, seasonal patterns also affect reserves
4. Attribution challenge: Hard to isolate shutdown effect from concurrent events
5. No predictive power: This skill describes current conditions, doesn't forecast

Troubleshooting

No recent data?

• Check if today is before next Wednesday data release
• Most recent weekly data is typically ~1 week lagged

SOFR premium calculation fails?

• Verify both EFFR and SOFR have data for the date range
• SOFR introduced April 2018; unavailable before

Chart rendering issues?

• Ensure matplotlib is installed
• Check date range has sufficient data points (need >2 weekly observations)

References

See bundled documentation:

• references/historical_cases.md - Detailed analysis of 2013, 2018-19, 2025 shutdowns
• references/data_sources.md - FRED API technical reference

External resources:

• Original PDF report (user-provided) for full theoretical framework
• NY Fed SOFR page: https://www.newyorkfed.org/markets/reference-rates/sofr
• FRED data: https://fred.stlouisfed.org/