## wpiea2025127-print-pdf - Section 6 concludes with the main results and policy implications

## Source details

**Canonical URL:** [wpiea2025127-print-pdf - Section 6 concludes with the main results and policy implications](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025127-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025127-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025127-print-pdf.pdf.json)

---

### Debt ceiling events and the reserve-draining channel through the TGA
- The debt ceiling is a legislative limit on the amount of national debt that can be incurred by the U.S. Treasury; first established in 1917 through the Second Liberty Bond Act.
- When the debt ceiling is reached:
  - The Treasury is prohibited from issuing net new debt, relies on incoming revenues and cash reserves in the Treasury General Account (TGA), and may initiate “extraordinary measures.”
  - The X-date is when extraordinary measures and the TGA are exhausted and the government cannot meet all obligations.
- Recent practice and institutional notes:
  - Congress sometimes suspends the ceiling for a fixed period rather than increasing it by a specific dollar amount.
  - Appendix Section B (not reproduced here) describes each debt ceiling episode since 2011.

### Transmission mechanism: TGA-induced reserve flows, repo market structure, and predicted channels
- Repo market structure:
  - Two main segments: tri-party repo (MMFs ↔ dealer banks) and bilateral repo (dealer banks ↔ leveraged funds).
  - Dealer banks intermediate between cash-rich investors (MMFs) and leveraged funds holding long-term Treasuries.
- Two types of TGA inflows and opposite effects on repo markets:
  - Fiscal inflows (tax payments net of transfer outlays):
    - Deposit transfers from banks and MMFs into the TGA reduce funds available for MMF lending → negative supply shock → lower RRMM positions → lower matched repo positions of dealers → lower repo borrowing by leveraged Treasury investors → higher repo funding price (repo spread) and higher Treasury yields.
  - Issuance inflows (proceeds from Treasury bond auctions to primary dealers or leveraged funds):
    - Increase demand for repo funding as dealers and leveraged funds need financing to absorb larger Treasury positions → higher repo funding and reverse repo holdings by MMFs → higher repo spreads.
- Empirical aim: use high-frequency TGA flows in normal times and around debt-ceiling episodes to measure implications for repo spreads and intermediated repo volumes.

### Stylized facts from TGA and repo spread dynamics
- TGA patterns around debt ceilings:
  - TGA balances are run down sharply when the debt ceiling is reached and extraordinary measures are declared (latest pronounced in January 2025).
  - When the debt limit is suspended or lifted, TGA balances surge as the Treasury restarts borrowing to replenish the TGA.
  - The subsequent 30-day pent-up borrowing by the Treasury amounted to 530 billion on average in the 2023 and 2021 episodes (compared with about 60 billion prior to COVID).
- Repo spread behavior:
  - The general collateral (GC) repo to IOR spread increases each time the debt ceiling was suspended or lifted, with a dramatic surge during September 2019.
  - Since mid-2020 the GC spread has become much more stable, consistent with the Fed’s commitment to the ample reserve regime.
- Measurement note: GC spread = spread of the general collateral repo rate over the overnight interest rate on reserves (IOR).

### Data, sample, and institutional evolution
- Sample and sources:
  - Most regressions use post-GFC data starting from the third quarter of 2009.
  - Daily GC repo rates from LSEG Datastream; daily TGA balances and Treasury operations from the U.S. Treasury; Federal Reserve balance sheet variables (reserves, ONRRP) from the Federal Reserve.
- MMF dataset:
  - Monthly MMF balance sheet data from SEC N-MFP filings; top 20 MMF families and 18 dealers account for more than 90% of repo transactions.
  - Haircut rate baskets: 2% (1.8%–2.1%), 3% (2.1%–3.1%), 4% (3.1%–4.1%), 5% (4.1%–5.1%), and greater than 5.1%.
  - Sample limited to repos with maturities between 1 and 5 days.
  - Resulting panel: MMF family–dealer–month observations from September 2010 to May 2024.

### Aggregate empirical findings: sensitivity of GC repo spreads to TGA changes
- Baseline regression specification:
  - ∆y_t = α + β ∆tga_t + Γ X_t + Θ S_t + ε_t; variables first-differenced; seasonal dummies for month start/mid/end; Newey West SEs with 10 lags.
  - TGA and ONRRP measured in trillions of USD; repo spreads in percentage points.
- Baseline estimation (Table 1 highlights):
  - Column (1): β on ∆tga = 0.2032*** (standard error 0.047).
    - Interpretation: a $100 billion increase in the TGA balance is associated with a 2 basis point rise in the GC repo spread (0.2032 percentage points per trillion USD → 0.2032% = 20.32 basis points per $1 trillion; equivalently 0.2 basis point per $100 billion).
  - Robustness:
    - Column (2) excluding 2019 spike: ∆tga = 0.2072*** (0.048).
    - Column (4) adding 8 lags of repo spread: estimate remains close to 0.2.
    - Column (5) adding 4 lags of ∆tga: contemporaneous ∆tga coefficient remains 0.2.
  - Broader market measures:
    - Column (6) (SOFR spread): ∆tga = 0.1312*** (0.035) → $100 billion TGA increase associated with just shy of 1 basis point rise in the SOFR spread.
    - Column (7) (10-year on-the-run Treasury yield to IOR spread): table header reports ∆tga = 0.1962*** (0.047); text reports 0.0905* (0.046) — preserve reported table values where applicable.
- Seasonality and sample:
  - Seasonality: month start, month mid, month end show consistently significant coefficients (examples: month start -0.0140** (0.006), month mid 0.0319*** (0.003), month end 0.0628*** (0.009)).
  - Observations example: 4,143 in column (1); sample 01jul2009–16may2025; R2 example 0.0275.

### Disentangling sources: net debt issuance vs net fiscal expenses
- Table 1, column (3):
  - ∆tga, net debt issuance = 0.2654*** (0.045).
  - ∆tga, fiscal = 0.1423** (0.072).
- Interpretation: a $100 billion TGA increase due to net debt issuance is associated with about 1 basis point larger widening in repo spread than a $100 billion increase due to net fiscal expenses.

### State-dependence, nonlinearities, and role of reserves/ONRRP/SRF
- Interaction results (Table 2):
  - ∆tga × reserves as percent of GDP, lag = -0.0880*** (0.027) in column (1).
  - ∆tga × ONRRP as percent of GDP, lag = -0.0233*** (0.008) in column (1).
  - ∆tga × post-SRF dummy = -0.3016** (0.131) in column (2).
- Quantitative interpretations:
  - As of the first quarter of 2025, banking sector reserves ≈ 11% of GDP; in the third quarter of 2019 reserves ≈ 7% of GDP.
  - For a $100 billion TGA increase:
    - Repo spread would rise by 4.3 basis points with current reserve levels (first quarter of 2025), which is about half of that in 2019.
  - ONRRP ≈ 1.8% of GDP in the first quarter of 2025; this mitigated about a quarter of the 1.4 basis point increase in repo spread caused by a $100 billion TGA increase, compared to a scenario with no ONRRP usage.
- Nonlinearity / large TGA changes:
  - Define “large change in TGA” as top percentile of ∆TGA measured as percent of previous day’s sum of reserves and ONRRP.
  - Interaction ∆tga × large∆tga dummy: not significant for GC repo spread (column (3): coefficient 0.1000, SE 0.166); positively significant for SOFR 75th percentile (column (4): 0.2156* (0.115)).
  - Implication: upper-tail (higher-rate) repo transactions are more sensitive to large TGA shocks than average repo transactions.
- Standing Repo Facility (SRF):
  - Post-SRF dummy interaction reduces sensitivity of repo spreads to TGA increases, indicating SRF provides insurance that lowers market sensitivity.

### Period comparisons, persistence, and propagation
- Period slopes (bin-scatter controlling for seasonal factors):
  - Post-GFC, pre-COVID (3Q2009–end-2019): associated with an 8 basis point rise in repo spread for a $100 billion TGA increase.
  - Post-COVID (2020–Dec 2024): associated with a 2 basis point rise for a $100 billion TGA increase.
  - TT&L active period (pre-GFC up to 3Q2007): changes in TGA largely muted by TT&L; figure annotation slope indicated as 0.8.
- Local projections (Figure 6):
  - Effects for a $100 billion TGA increase fluctuate between 1 to 3 basis points over the 30 days following the shock, fading after about 20 days.
  - Gray area in figure denotes 99% confidence interval; horizontal axis days 0–30.

### TGA behavior during debt ceiling standoffs and event-study facts
- During standoffs:
  - Treasury implements extraordinary measures leading to a draw down of the TGA; after lift/suspension the TGA is replenished quickly.
  - Event study pooling standoffs since 2012:
    - The TGA shrinks by more than half on average until the debt ceiling is lifted.
    - Net bill issuance is negative before the ceiling is lifted and positive after; net coupon issuance follows predictable quarterly cycles.
- Event-time metrics:
  - Mean net flows calculated using a 5-day moving window; shaded ranges represent ±1 standard deviation around the mean.

### Impact on repo spreads during debt ceiling standoffs
- Regression re-estimation and prediction exercise:
  - Re-estimate Equation 1 with ∆tga × reserves as a share of GDP interaction, lagged ∆tga terms, excluding observations within 30 days of a debt ceiling suspension or increase; apply estimates to predict spreads in the 30-day window before and after each suspension or increase.
- Findings:
  - Average GC Repo - IORB spread declines until the ceiling is lifted and increases afterward, consistent with TGA-driven predictions.
  - Magnitude of observed swing similar to predicted swings based on TGA-driven regression estimates.
  - Including the 6-month US sovereign CDS premium improves fit somewhat, but a prediction based only on the CDS premium does not fit well.
  - Pure TGA inflow effect around debt ceiling suspension drives the bulk of the swing in the repo spread.

### Predicted impact of the 2025 debt ceiling episode (scenario inputs and predictions)
- Context and inputs:
  - Debt limit reinstated at $36.1 trillion on January 2, 2025.
  - Bipartisan Policy Center projects X Date could arrive as soon as late-July.
  - Prediction assumptions:
    - TGA balance follows Bipartisan Policy Center trajectory until the X-date.
    - Figure 10 (descriptive): lower bound X-date late-July; upper bound X-date late-October; mid-point forecast end-August.
    - After the X-date, TGA replenished to pre-debt ceiling level at average historical pace.
    - ONRRP and reserve levels follow mid-point estimate from the Federal Reserve Bank of New York’s Survey of Market Expectations.
- Predicted repo spread effects (Figure 11 results):
  - If the debt ceiling is not lifted until the X-date:
    - The repo spread is predicted to decline by between 20 to 30 basis points depending on the speed of TGA drawdown, and then sharply rebound following the debt-ceiling lift-off.
  - Larger expected swing if the TGA is drawn down more slowly, because declining bank reserves and ONRRP balances from QT increase sensitivity.
  - Comparative magnitudes:
    - Smaller than prior episodes: ∼40 bps in 2017 and ∼70 bps in 2019.
    - Still an unusually large spike in the post-COVID era (since mid-2020 spreads have varied in a narrow band of less than 10 basis points around the federal funds target).
  - Current amplifiers: declining bank reserves due to QT and very low ONRRP levels.

### Main conclusions and policy implications
- Main conclusions:
  - Political brinkmanship around the federal debt-ceiling can have significant adverse consequences for the repo market and financial stability, with the Treasury market at its center.
  - Mechanical reserve injections and drainings resulting from TGA volatility have a significant effect on repo spreads and the volume of repo borrowing intermediated by the market.
  - The TGA-driven liquidity channel matters even more than sovereign risk in driving repo spread swings during debt ceiling episodes.
- Policy options and recommendations:
  - Re-establish a version of former Tax and Loan Note accounts to offset TGA fluctuations and insulate the financial system from associated fiscal flows (Huther, Pettit and Wilkinson (2019)).
  - In the absence of such arrangements and with recurring debt-ceiling impasses:
    - Maintain a large level of reserves to blunt the impact from TGA volatility.
    - Recognize trade-offs: maintaining an ample reserve regime offers benefits but involves costs related to a large Fed balance sheet.

*Source — wpiea2025127-print-pdf - Section 6 concludes with the main results and policy implications.*

### Section 6 concludes with the main results and policy implications.

### wpiea2025127-print-pdf - Section 6 concludes with the main results and policy implications.

### Debt ceiling events and the reserve-draining channel through the TGA
- The debt ceiling is a legislative limit on the amount of national debt that can be incurred by the U.S. Treasury; first established in 1917 through the Second Liberty Bond Act.
- Since the 1990s, debates over raising the debt limit have triggered fiscal standoffs, government shutdowns, and—in 2011—a historic downgrade of the U.S. credit rating.
- When the debt ceiling is reached:
  - The Treasury is prohibited from issuing net new debt to finance government operations, though it may still issue debt to roll over maturing obligations.
  - The Treasury relies on incoming revenues and cash reserves in the Treasury General Account (TGA).
  - The Treasury may initiate “extraordinary measures” (e.g., suspending or redeeming investments in certain government trust funds, transactions through the Federal Financing Bank) to buy time.
  - The X-date is the point when extraordinary measures and the TGA are exhausted and the government no longer has sufficient funds to meet all obligations.
- Recent practice: Congress has sometimes opted to suspend the ceiling for a fixed period rather than increase it by a specific dollar amount; suspension averts immediate crisis but does not resolve underlying structural and political tensions.
- Institutional note: Appendix Section B (not reproduced here) describes each debt ceiling episode since 2011.

### Transmission mechanism: TGA-induced reserve flows and the repo market
- Repo market structure summarized:
  - Two main segments: tri-party repo (MMFs ↔ dealer banks) and bilateral repo (dealer banks ↔ leveraged funds).
  - Dealer banks intermediate between cash-rich investors (MMFs) and leveraged funds that invest in long-term Treasuries.
- Two types of inflows into the TGA and their opposite effects on repo markets:
  - Fiscal inflows (tax payments net of transfer outlays):
    - Transfer of deposits from banks and MMFs into the TGA reduces the supply of funds available for MMF lending.
    - This manifests as a negative supply shock: lower reverse repo (RRMM) positions held by MMFs → lower matched repo positions of dealers → lower repo borrowing by leveraged Treasury investors.
    - Result: higher repo funding price (repo spread) and higher Treasury yields to compensate investors for increased funding costs.
  - Issuance inflows (proceeds from Treasury bond auctions to primary dealers or leveraged funds):
    - Increase demand for repo funding as dealers and leveraged funds need financing to absorb larger Treasury positions.
    - This demand shock increases repo funding and reverse repo holdings by MMFs, requiring higher repo spreads.
- Empirical aim: test supply and demand effects using high-frequency TGA flows in normal times and around debt-ceiling episodes; measure implications for repo spreads and intermediated repo volumes.

### Stylized facts from TGA and repo spread dynamics
- Historical patterns:
  - TGA balances are run down sharply soon after the debt ceiling is reached/triggered, typically when extraordinary measures are declared (latest pronounced in January 2025).
  - When the debt limit is suspended or lifted, TGA balances surge as the Treasury restarts borrowing to replenish the TGA.
  - The surge was particularly pronounced in the 2023 and 2021 debt ceiling episodes: the subsequent 30-day pent-up borrowing by the Treasury amounted to 530 billion on average (compared with about 60 billion prior to COVID).
- Repo spread behavior:
  - The general collateral (GC) repo to IOR spread increases each time the debt ceiling was suspended or lifted (grey-shaded zones in Figure 2), with a dramatic surge during the repo crisis in September 2019.
  - Since mid-2020 the GC spread has become much more stable, consistent with the Fed’s commitment to the ample reserve regime.
- Measurement note: GC spread is measured as the spread of the general collateral repo rate over the overnight interest rate on reserves (IOR); the spread captures additional compensation lenders require to lend in secured funding markets instead of depositing with the Fed.

### Data and the evolution of the Treasury General Account
- Sample and data sources:
  - Most regressions use post-GFC data starting from the third quarter of 2009.
  - Daily general collateral repo rates sourced from LSEG Datastream.
  - Daily TGA balances and Treasury issuance/redemption/fiscal operations from the U.S. Treasury.
  - Federal Reserve balance sheet variables (reserves, ONRRP) from the Federal Reserve.
- Treasury Tax and Loan (TT&L) accounts and institutional change:
  - Before the GFC, the Treasury managed cash via both the TGA at the Fed and TT&L accounts at private depository institutions; TGA targeted approximately $5 billion pre-GFC.
  - TT&L accounts received tax revenues and earned interest at a rate equal to the federal funds rate minus 25 basis points.
  - Post-GFC changes: as the federal funds rate fell below 25 basis points, the TT&L deposit rate dropped to zero; the Fed began paying interest on reserves, making it advantageous for the Treasury to maintain funds in the TGA to avoid reducing Fed remittances.
- MMF data:
  - Monthly MMF balance sheet data from SEC N-MFP filings; all U.S.-domiciled MMFs report detailed portfolio holdings as of the last business day of each month.
  - N-MFP provides borrower, collateral type, haircut rate, maturity, repo rate, and transaction amount.
  - Aggregation follows Huber (2023) at MMF family and dealer levels.
  - Focus sample: top 20 MMF families and 18 dealers, accounting for more than 90% of total repo transactions.
  - Haircut rate baskets: 2% (1.8%–2.1%, representing 70% of repo transactions), 3% (2.1%–3.1%), 4% (3.1%–4.1%), 5% (4.1%–5.1%), and greater than 5.1%.
  - Sample limited to repos with maturities between 1 and 5 days.
  - Resulting dataset: panel of MMF family–dealer–month observations from September 2010 to May 2024 (Form N-MFP dataset updates after May 2024 not available yet).

### Empirical identification and strategy
- Identification assumptions:
  - Aggregate-level: at a daily frequency, financial markets treat the TGA as exogenous—Treasury fiscal operations are assumed unrelated to market conditions day-to-day.
  - Disaggregated (MMF-level): individual MMFs and dealers similarly regard TGA changes as exogenous.
- Supporting evidence for exogeneity: prior literature indicates day-to-day or month-to-month Treasury operations are generally unrelated to market conditions, with fiscal policy minimally responsive to macroeconomic conditions at these frequencies (cited authors: Correa, Du and Liao (2020); Diamond, Jiang and Ma (2024)).

*Italic: Source — wpiea2025127-print-pdf - Section 6 concludes with the main results and policy implications.*

### 4.1  Aggregate Repo Market

### 4.1  Aggregate Repo Market

### Methodology
- Regression specification (Equation 1): dependent variable is the spread between the general collateral repo rate and the interest rate on excess reserves (IOR). All variables are first-differenced.
  - ∆y_t = α + β ∆tga_t + Γ X_t + Θ S_t + ε_t
- Seasonal dummies included for beginning, middle, and end of month (S_t).
- Controls (X_t) used in robustness checks; Newey West standard errors with 10 lags.
- TGA and ONRRP are measured in trillions of USD. Repo spreads are in percentage points.

### Baseline estimation results (Table 1)
- Column (1) baseline:
  - β on ∆tga = 0.2032*** (standard error 0.047)
  - Interpretation: a $100 billion increase in the TGA balance is associated with a 2 basis point rise in the general collateral repo spread (0.2032 percentage points per trillion USD → 0.2032% = 20.32 basis points per $1 trillion; equivalently 0.2 basis point per $100 billion).
- Robustness and alternative specifications:
  - Column (2) (excluding 2019 spike): ∆tga = 0.2072*** (0.048)
  - Column (4) (adds 8 lags of repo spread): estimate remains close to 0.2.
  - Column (5) (adds 4 lags of ∆tga): contemporaneous ∆tga coefficient remains 0.2.
- Broader market measures:
  - Column (6) (SOFR spread): ∆tga = 0.1312*** (0.035).
    - Interpretation: $100 billion increase in TGA associated with just shy of 1 basis point rise in the SOFR spread.
  - Column (7) (10-year on-the-run Treasury yield to IOR spread): ∆tga = 0.1962*** (0.047) in table header but reported coefficient in text: about 1 basis point rise in 10-year Treasury spread for $100 billion TGA increase (text reports 0.0905* in table column (7) with SE 0.046—preserve table values).
- Seasonality coefficients (representative):
  - month start: negative and significant in many specs (e.g., -0.0140** (0.006) in column (1))
  - month end: positive and significant (e.g., 0.0628*** (0.009) in column (1))
  - month mid: positive and significant (e.g., 0.0319*** (0.003) in column (1))
- Sample periods and observations (selected):
  - Observations: 4,143 in column (1)
  - Sample: 01jul2009–16may2025 for many columns
  - R2 example: 0.0275 for column (1)

### Net debt issuance versus net fiscal expenses (Table 1, column (3))
- Distinguishing TGA changes by source:
  - ∆tga, net debt issuance = 0.2654*** (0.045)
  - ∆tga, fiscal = 0.1423** (0.072)
- Interpretation: a $100 billion TGA increase due to net debt issuance is associated with about 1 basis point larger widening in repo spread than a $100 billion increase due to net fiscal expenses (net debt issuance appears to have a larger impact consistent with added repo demand from primary dealers warehousing Treasuries).

### State-dependence and nonlinearity (Table 2)
- Main findings:
  - Interaction of ∆tga with reserves as percent of GDP (lag): negative and significant.
    - ∆tga×reserves as percent of GDP, lag = -0.0880*** (0.027) in column (1)
  - Interaction of ∆tga with ONRRP as percent of GDP, lag = -0.0233*** (0.008) in column (1)
  - Interaction of ∆tga with post-SRF dummy (column (2)): ∆tga×post SRF dummy = -0.3016** (0.131)
- Quantitative interpretation provided in text (preserve original phrasing and numbers):
  - As of the first quarter of 2025, banking sector reserves ≈ 11% of GDP; in the third quarter of 2019 reserves ≈ 7% of GDP.
  - Estimates suggest that for a $100 billion TGA increase:
    - Repo spread would rise by 4.3 basis points with current reserve levels (first quarter of 2025), which is about half of that in 2019.
  - ONRRP ≈ 1.8% of GDP in the first quarter of 2025; this mitigated about a quarter of the 1.4 basis point increase in repo spread caused by a $100 billion TGA increase, compared to a scenario with no ONRRP usage.
- Nonlinearity / large TGA changes:
  - Define “large change in TGA” as top percentile of ∆TGA measured as percent of previous day’s sum of reserves and ONRRP.
  - Interaction ∆tga×large∆tga dummy: not significant for general collateral repo spread (column (3): coefficient 0.1000, SE 0.166), but positively significant for SOFR 75th percentile (column (4): 0.2156* (0.115)).
  - Implication: higher-rate repo transactions (upper tail) are more sensitive to large TGA shocks than average repo transactions.

### Time-varying sensitivity and facilities (Figure 4 and SRF)
- One-year rolling-window estimates show β (sensitivity of repo spread to ∆TGA) is negatively correlated with the sum of reserves and ONRRP.
- Standing Repo Facility (SRF) role:
  - Interaction with post-SRF dummy is negative (see column (2) in Table 2), suggesting SRF establishment reduces the sensitivity of repo spreads to TGA increases by providing market insurance of additional liquidity in stress.

### Period comparisons and role of Treasury Tax and Loan (Figure 5; textual findings)
- Bin-scatter evidence controlling for seasonal factors:
  - Post-GFC and pre-COVID period (third quarter 2009 to end-2019): associated with an 8 basis point rise in repo spread for a $100 billion TGA increase (red dots).
  - Post-COVID period (beginning of 2020 to Dec 2024): associated with a 2 basis point rise in repo spread for a $100 billion TGA increase (green dots).
  - TT&L active period (pre-GFC up to third quarter of 2007): changes in TGA largely muted by TT&L (blue dots); slope indicated as 0.8 in figure annotation.
- Interpretation: active TT&L operation and abundant Fed balance sheet post-COVID reduce spillovers from TGA changes to repo spreads.

### Persistence and propagation dynamics (local projections; Figure 6)
- Local projection estimates (extending column (1) of Table 1) up to 30 days after a TGA change:
  - Effects fluctuate between 1 to 3 basis points in response to a $100 billion TGA increase, before fading away after about 20 days.
  - Figure notes: gray area represents 99% confidence interval; horizontal axis days 0–30; vertical axis basis points response to $100bn increase in TGA.

*Italic: Source: IMF working paper chapter "4.1  Aggregate Repo Market" (extracted content).*

### 5.1  TGA and Debt Ceiling Standoffs

### 5.1  TGA and Debt Ceiling Standoffs

### TGA behavior during debt ceiling standoffs
- During debt ceiling standoffs, the Treasury implements “extraordinary measures” that lead to a draw down of the TGA.
- After the drawdown period, the Treasury replenishes the TGA immediately when the debt ceiling is suspended or lifted.
- Event study pooling debt ceiling standoffs since 2012 shows:
  - The TGA shrinks by more than half on average until the debt ceiling is lifted.
  - After the debt ceiling is lifted, the TGA quickly returns to near its previous level.
- Net bill issuance is the main driver of the TGA during debt ceiling standoffs:
  - Net bill issuance is negative before the debt ceiling is lifted.
  - Net bill issuance is positive after the debt ceiling is lifted.
- Net coupon issuance follows a periodic cycle aligned with the Treasury’s predictable quarterly refinancing auctions for bonds.

### Key statistics and figures (as presented)
- Event-time window: days relative to when the debt ceiling is suspended/lifted.
- Mean net flows calculated using a 5-day moving window; shaded ranges represent ±1 standard deviation around the mean.

### 5.2  Impact on Repo Spreads

### Regression approach and prediction exercise
- Re-estimate Equation 1 including:
  - An interaction term for changes in the TGA and reserves as a share of GDP (as in Column 1 of Table 2).
  - Lagged terms for changes in the TGA (as in Column 5 of Table 1).
  - Exclude observations within 30 days of a debt ceiling suspension or increase.
- Apply regression estimates to predict changes in repo spreads through the 30 day window before and after each suspension or increase in the debt ceiling.

### Findings on repo spreads during standoffs
- Comparison in Figure 9:
  - The average spread between the general collateral repo rate and the interest rate on excess reserves (GC Repo - IORB) declines until the debt ceiling is lifted and increases afterward, consistent with predictions driven by TGA changes.
  - The magnitude of the observed swing in the spread is similar to the predicted swings based on TGA-driven regression estimates.
- Including the 6-month premium of US sovereign credit default swaps (CDS) in the estimating regression:
  - Improves the fit of predictions somewhat (helps account for the large decline immediately after the debt ceiling increase).
  - A prediction based only on the CDS premium does not fit the data well.
  - The pure TGA inflow effect around debt ceiling suspension drives the bulk of the swing in the repo spread.

### 5.3  Expected Impact of 2025 Debt Ceiling Episode

### Context and assumptions
- The debt limit was reinstated at $36.1 trillion on January 2, 2025, following its suspension as part of the 2023 Fiscal Responsibility Act.
- The Bipartisan Policy Center projects that the X Date could arrive as soon as late-July.
- Prediction inputs and assumptions:
  - TGA balance follows the trajectory projected by the Bipartisan Policy Center until the X-date.
  - Figure 10 presents the mid-point and upper and lower bounds of the TGA balance forecast:
    - Lower bound X-date: late-July.
    - Upper bound X-date: late-October.
    - Mid-point forecast: end-August.
  - After the X-date, the TGA is replenished to its pre-debt ceiling level at the average historical pace.
  - ONRRP and reserve levels follow the mid-point estimate from the Federal Reserve Bank of New York’s Survey of Market Expectations.

### Predicted effects on repo spreads (Figure 11 results)
- If the debt ceiling is not lifted until the X-date:
  - The repo spread is predicted to decline by between 20 to 30 basis points depending on the speed at which the TGA is drawn down, and then sharply rebound following the debt-ceiling lift-off.
  - The expected swing is larger if the TGA is drawn down more slowly over time, because declining bank reserves and ONRRP balances from QT increase sensitivity of spreads to TGA changes.
- Comparative magnitudes:
  - Expected magnitude is somewhat smaller than associated swings in prior episodes: ∼40 bps in 2017 and ∼70 bps in 2019.
  - The predicted swing would still constitute an unusually large spike in the repo spread during the post-COVID era.
  - Since mid-2020, the spread has varied in a narrow band of less than 10 basis points around the federal funds target, with larger discrete jumps only at quarter ends and FOMC rate changes.
- Current conditions that amplify sensitivity:
  - Declining level of bank reserves due to ongoing QT by the Fed.
  - Very low level of ONRRP.
  - Both factors contribute to amplifying the swing in repo spreads in response to TGA flow volatility.

### Note on prediction model
- The prediction model includes:
  - An interaction term for changes in the TGA and reserves as a share of GDP (as in Column 1 of Table 2).
  - Lagged terms for changes in the TGA (as in Column 5 of Table 1).
  - Excludes observations within 30 days of a debt ceiling suspension or increase.

### Conclusion and policy implications

### Main conclusions
- Political brinkmanship around the federal debt-ceiling can have significant adverse consequences for the repo market and financial stability, with the Treasury market at its center.
- Mechanical reserve injection and draining resulting from TGA volatility have a significant effect on repo spreads and the volume of repo borrowing intermediated by the market.
- The TGA-driven liquidity channel appears to matter even more than sovereign risk in driving repo spread swings during debt ceiling episodes.

### Policy options and recommendations
- Options to mitigate spillovers to funding markets include re-establishing a version of former Tax and Loan Note accounts that could offset TGA fluctuations and insulate the financial system from associated fiscal flows (Huther, Pettit and Wilkinson (2019)).
- In the absence of such arrangements and with recurring debt-ceiling impasses:
  - Maintain a large level of reserves to blunt the impact from TGA volatility.
  - Note that maintaining an ample reserve regime offers additional benefits but also involves trade-offs related to the cost of a large Fed balance sheet.

*Source: 5.1–5.3, 6, and Appendix sections of the provided chapter content.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025127-print-pdf.pdf_