Macro Mechanics #2: The Reverse Repo Facility
Pierce & Pierce Research | January 2026
Part of the Macro Mechanics Series
⏱️ 12-15 minutes • Beginner-friendly • No prior Fed knowledge required
Why This Matters
In Brief #1, we explained how the Treasury General Account works like a seesaw with bank reserves. When the TGA rises, reserves fall. When the TGA falls, reserves rise.
But for the past few years, that seesaw had a shock absorber. A massive cushion that could expand or contract to smooth out the impact of TGA swings on the banking system.
That cushion was the Overnight Reverse Repo facility. At its peak in late 2022, the ON RRP held over $2.4 trillion.
Current Status:
| Metric | Value | Status |
|---|---|---|
| Peak Balance | $2.4 trillion (Dec 2022) | - |
| Current Balance | ~$6 billion (Jan 2026) | 🔴 |
| Decline | -99.8% | 🔴 Buffer exhausted |
That cushion is now gone. This changes everything about how Fed balance sheet mechanics transmit to markets.
Section 1: What Is the Reverse Repo Facility?
The Overnight Reverse Repo facility is a program where the Fed borrows cash from eligible counterparties overnight. In exchange, the Fed provides Treasury securities as collateral and pays interest at the ON RRP rate.
Think of it like a savings account at the Fed, but for money market funds and other institutions that cannot hold regular bank reserves.
The facility serves two important functions: (1) it provides a floor under short-term interest rates by offering a risk-free return, and (2) it acts as a liquidity buffer that can absorb excess cash or release it back into the system. Both functions matter for understanding why its exhaustion is significant.
The Key Distinction:
| Institution Type | Where They Park Cash | Rate They Earn |
|---|---|---|
| Banks | Reserve balances at Fed | IORB (3.65%) |
| Money Market Funds | ON RRP facility | ON RRP rate (3.50%) |
Banks earn interest on reserve balances (IORB) by depositing cash at the Fed. But money market funds are not banks. They cannot hold reserves. Before the ON RRP existed, these funds had to find other places to park their cash overnight.
The Mechanical Effect:
| ON RRP Action | Effect on Markets |
|---|---|
| Funds deposit cash in ON RRP | 🔴 Liquidity removed from private markets |
| Funds withdraw from ON RRP | 🟢 Liquidity flows back into financial system |
This creates a dynamic buffer. The ON RRP can absorb excess liquidity when the system has too much, and release it when conditions tighten.
Section 2: How the Buffer Actually Works
To understand why the ON RRP mattered so much, you need to see its position in the Fed’s balance sheet plumbing.
Fed Balance Sheet Liabilities (Zero-Sum Relationship):
| Liability | What It Is | Who Controls It |
|---|---|---|
| Currency | Physical cash | Public demand |
| Bank Reserves | Cash banks hold at Fed | Fed policy + bank behaviour |
| TGA | Treasury’s checking account | Treasury operations |
| ON RRP | Money market fund deposits | MMF behaviour + rate incentives |
Here’s the constraint that governs everything: total assets must equal total liabilities. When one liability increases, something else must decrease.
Scenario A: No Buffer (What Happens Now)
| Step | What Happens |
|---|---|
| 1 | Treasury issues $100B in new debt |
| 2 | Investors buy the bonds |
| 3 | Payment comes from bank accounts |
| 4 | Reserves decrease by $100B |
| 5 | TGA increases by $100B |
| Result | 🔴 Reserves take full hit |
Scenario B: With Buffer (What Used to Happen)
| Step | What Happens |
|---|---|
| 1 | Treasury issues $100B in new bills |
| 2 | Money market funds withdraw from ON RRP to buy bills |
| 3 | ON RRP decreases by $100B |
| 4 | TGA increases by $100B |
| 5 | Reserves stay flat |
| Result | 🟢 Buffer absorbs impact |
One important nuance: The ON RRP only buffers flows to the extent that money market funds are willing to shift their allocations. If T-bill yields are higher than the ON RRP rate, funds prefer bills. If the ON RRP rate is higher, funds prefer parking cash at the Fed. The buffer is rate-sensitive, not automatic.
Section 3: Why It Grew to $2.4 Trillion
The ON RRP was designed as a monetary policy tool, not as a multi-trillion dollar liquidity absorber. But pandemic-era quantitative easing changed everything.
The QE Flood:
| Period | Fed Balance Sheet Addition | Primary Destination |
|---|---|---|
| March 2020 - April 2022 | +$4.8 trillion | Reserves + ON RRP |
This created an enormous flood of liquidity. Banks ended up with vastly more reserves than they needed. Money market funds ended up with vastly more cash than they could deploy at attractive rates.
Where Could This Excess Liquidity Go?
| Option | Yield Available | Problem |
|---|---|---|
| T-bills | Near zero / negative | Terrible returns |
| Private repo | Near zero | Flooded market |
| ON RRP at Fed | Small but positive | Risk-free ✓ |
Money market funds made a rational choice: park the excess at the Fed rather than accept near-zero or negative yields elsewhere.
ON RRP Growth Timeline:
| Date | ON RRP Balance | Status |
|---|---|---|
| Jan 2021 | Near zero | Normal |
| Dec 2021 | ~$1.5 trillion | 🟡 Absorbing QE excess |
| Dec 2022 | ~$2.4 trillion | 🔴 Peak buffer |
The facility absorbed QE’s excess liquidity. It became an unintentional pressure valve for a system drowning in cash.
Section 4: The Great Drainage
Starting in mid-2023, the ON RRP began draining rapidly. By late 2025, it had fallen to roughly $10-20 billion.
What Drove the Drainage? The Rate Spread Flipped:
| Period | T-Bill Yield | ON RRP Rate | Spread | Fund Preference |
|---|---|---|---|---|
| 2021-2022 | Near zero | 0.05-0.30% | Negative | ON RRP |
| Late 2023 | ~5.0% | ~5.30% | Slight positive | ON RRP |
| 2024-2025 | ~4.5-5.0% | ~4.25-5.05% | Variable | Bills |
| Dec 2025 | ~3.6% | 3.50% | +10 bps | Bills |
When T-bill yields exceeded the ON RRP rate, fund managers shifted to bills. Multiply this decision across thousands of funds managing trillions of dollars, and the ON RRP drained steadily.
Historical Drainage Timeline:
| Date | ON RRP Balance | Status |
|---|---|---|
| Dec 2022 | ~$2.4T | 🟢 Full buffer |
| June 2023 | ~$2.0T | 🟢 Beginning decline |
| Dec 2023 | ~$800B | 🟡 Rapid drainage |
| June 2024 | ~$400B | 🟡 Continuing decline |
| Dec 2024 | ~$100B | 🔴 Near exhaustion |
| Sept 2025 | ~$18B | 🔴 Effectively zero |
| Dec 2025 | ~$10B | 🔴 Buffer gone |
Fed officials acknowledged in the September 2025 FOMC minutes that ON RRP balances had declined substantially and could no longer serve as a meaningful buffer for reserve fluctuations.
Section 5: What Its Absence Means
With the ON RRP buffer exhausted, the Fed’s balance sheet dynamics have fundamentally changed.
Transmission Comparison:
| Condition | TGA Movement | Impact Path | Reserve Volatility |
|---|---|---|---|
| Before (with buffer) | TGA swings | ON RRP absorbs → Reserves stable | 🟢 Low |
| After (no buffer) | TGA swings | Direct hit to reserves | 🔴 High |
Concrete Implications:
| Implication | What It Means | Status |
|---|---|---|
| Direct transmission | Every TGA dollar change hits reserves | 🔴 Active now |
| Reserve scarcity risk | No cushion near the floor | 🔴 Elevated |
| Active Fed management required | Fed must step in directly | 🟡 Ongoing |
This is precisely why the Fed ended QT in December 2025 and began reserve management purchases of T-bills. The buffer used to do this job passively. Now the Fed must do it actively.
⚠️ Transmission Caveat:
As covered in Brief #1, reserve injections don’t mechanically flow to risk assets. Just because TGA swings hit reserves directly does not mean they automatically move asset prices.
| Factor | Why It Can Break Transmission |
|---|---|
| Bank hoarding | Banks may sit on reserves for regulatory compliance |
| Capital constraints | Basel III requirements (LCR, NSFR) trap some reserves |
| Credit demand | Reserves sitting idle don’t stimulate anything |
The relationship between reserves and risk assets is directional, not mechanical. But the absence of the buffer means the system has less slack to absorb shocks.
Section 6: Historical Precedents
Two episodes illustrate how the ON RRP buffered Treasury operations.
The 2023 Debt Ceiling Resolution
| Phase | TGA Balance | ON RRP Role |
|---|---|---|
| Jan 2023 | ~$490B | Buffer available |
| May 2023 | ~$49B | Ceiling forced drawdown |
| Post-ceiling rebuild | Back to $400B+ | ON RRP absorbed ~$500B of issuance |
| Reserve impact | Stable | 🟢 Buffer worked |
Money market funds shifted roughly $500 billion from ON RRP to purchase the new Treasury bills. The TGA rebuilt while reserves stayed relatively stable.
If this episode had occurred in 2025 with no ON RRP buffer, the reserve impact would have been severe.
The March 2020 COVID Shock
| Phase | ON RRP Balance | What Happened |
|---|---|---|
| Pre-pandemic | Near zero | Normal conditions |
| Post-QE flood | $2+ trillion | Absorbed excess liquidity |
This demonstrated the facility’s ability to absorb excess liquidity during periods of massive monetary expansion. Without it, short-term rates might have gone deeply negative.
The Lesson: The ON RRP served as both a floor under rates (during excess liquidity) and a buffer against Treasury operations. Its exhaustion removes both functions simultaneously.
Tail Risk Note: A liquidity crisis (similar to March 2020) could theoretically rebuild ON RRP usage if the Fed cuts aggressively or intervenes in ways that push ON RRP above T-bill yields. This is low probability but high impact—worth monitoring during stress events.
Section 7: Where We Are Now
Current ON RRP Dashboard:
| Metric | Value | Status |
|---|---|---|
| ON RRP Balance | ~$6B | 🔴 Exhausted |
| ON RRP Rate | 3.50% | Bottom of Fed range |
| T-bill yields | ~3.55-3.65% | Above ON RRP rate |
| Bank Reserves | ~$2.88T | 🔴 Below floor |
| Rate spread | +5-15 bps | 🔴 Bills preferred |
Why the ON RRP Stays Near Zero:
The Fed cut the ON RRP offering rate to 3.50% in December 2025 (the bottom of its policy range). This was intentional. With T-bill yields above this rate, money market funds have no incentive to park cash at the Fed.
Fed officials have been explicit that they want ON RRP usage near zero. This represents a return to “normal” conditions before QE flooded the system with excess liquidity.
What Would Rebuild the ON RRP?
| Scenario | Likelihood | Status |
|---|---|---|
| T-bill yields fall sharply below ON RRP rate | Low | 🔴 |
| Another round of QE floods system | Low | 🔴 |
| Treasury dramatically reduces bill issuance | Very low | 🔴 |
| Liquidity crisis forces Fed intervention | Low (tail risk) | 🟡 |
None of these scenarios appear imminent. The buffer should remain near zero through 2026 and likely beyond.
Section 8: Thesis Connection
For Pierce & Pierce readers, the ON RRP drainage is critical context for the current liquidity setup.
The Compound Effect:
Brief #1 explained how TGA drawdowns inject liquidity into reserves. Brief #2 shows why this matters more now than before. With no ON RRP buffer, TGA movements have unfiltered impact on the banking system.
Combined Reading:
| Factor | Current Status | Direction / Monthly Impact |
|---|---|---|
| TGA | Drawing down | 🟢 Injecting ~$40B |
| Fed T-bill purchases | Ongoing since Dec 12 | 🟢 Injecting ~$40B |
| ON RRP Buffer | Exhausted | 🔴 Cannot absorb $0 |
| Net effect | Direct to reserves | 🟢 ~$80B |
Section 9: What To Watch
Primary Data Sources:
| Source | What It Shows | Frequency |
|---|---|---|
| FRED: RRPONTSYD | ON RRP usage | Daily |
| NY Fed operations page | Daily participation | Daily |
| FRED: RRPONTSYAWARD | ON RRP offering rate | Daily |
ON RRP Level Interpretation:
| Level | Status | Interpretation |
|---|---|---|
| >$100B | 🟢 | Some buffer capacity remains |
| $20-50B | 🟡 | Minimal buffer, normal operations only |
| <$20B | 🔴 | Effectively exhausted, no shock absorption |
The Spread to Watch:
ON RRP Incentive = T-bill yield - ON RRP rate
Current: 3.60% - 3.50% = +10 bps → Bills preferred, ON RRP stays near zero
| Spread | Fund Preference | ON RRP Direction |
|---|---|---|
| Positive (T-bills pay more) | Bills | Stays near zero |
| Negative (ON RRP pays more) | ON RRP | Would rebuild |
Quick Reference
| Item | Detail |
|---|---|
| The Mechanic | ON RRP is where money market funds park cash at the Fed. High balances absorb excess liquidity. Near-zero balances mean TGA swings hit reserves directly. |
| The Nuance | The ON RRP is rate-sensitive, not automatic. Funds shift based on yield differentials. Also serves as a floor under short-term rates. |
| Peak Level | ~$2.4 trillion (December 2022) |
| Current Level | ~$6 billion (January 2026) |
| Buffer Status | 🔴 Effectively exhausted. Down 99.8% from peak. |
| Why It Matters Now | No buffer means TGA movements and Fed policy transmit directly to reserves. |
| What Would Change This | QE, sharp rate cuts, dramatically reduced Treasury issuance, or liquidity crisis. None appear imminent. |
| Monitor | FRED series RRPONTSYD (daily) and NY Fed operations page |
What’s Next
Brief #3: Bank Reserves & The “Ample” Floor
Why reserves are the liquidity that actually matters, what happened in September 2019, and why the Fed panicked in October 2025.
Brief #4: QT → QE (Balance Sheet Cycles)
The 42-month headwind that became a tailwind, and why “reserve management purchases” are not quite the same as QE.
Brief #5: Global M2 & The Liquidity Lag
Why Bitcoin correlates with money supply, and why it lags by months, not days.
Sources:
Federal Reserve Economic Data (FRED) | NY Fed Open Market Operations | Federal Reserve H.4.1 Release | Fed Staff Working Papers on Monetary Policy Implementation | FOMC Meeting Minutes (September, November, December 2025) | Office of Financial Research Repo Market Analysis
Methodology Note:
ON RRP data from FRED series RRPONTSYD through January 2, 2026. Historical debt ceiling analysis from Treasury Daily Statements and Congressional Research Service. All figures represent publicly available data from official sources.
This publication is for educational purposes only. Not financial advice. Cryptocurrency investments carry substantial risk of complete loss.
Pierce & Pierce Research
Democratising Institutional-Grade Crypto Research
