## sr1146

## Source details

**Canonical URL:** [sr1146](https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr1146.pdf)

## Other formats

- [Markdown version](/medialibrary/media/research/staff_reports/sr1146.pdf.md)
- [Structured JSON version](/medialibrary/media/research/staff_reports/sr1146.pdf.json)

---

### 1. INTRODUCTION
- Role and importance of US Treasury securities:
  - Serve as a benchmark in global capital markets because of their "exceptional creditworthiness and liquidity."
  - Key uses: manage interest rate risk, price other securities, collateralize financing transactions, implement monetary policy, and act as a reserve asset.
  - Investors pay a premium for Treasuries because of their liquidity, leading to lower yields and hence lower US government borrowing costs.
  - Liquidity is a central attribute closely linked to Treasuries’ pricing and popularity.
- Historical episodes of market stress and implications for liquidity:
  - Major events: the 2007-09 global financial crisis (GFC); post-crisis regulatory changes and increasing role of principal trading firms (PTFs); the October 15, 2014 flash rally; Covid-19-related developments in March 2020 (the "dash-for-cash").
  - March 2020 disruptions:
    - Massive customer selling overwhelmed dealers’ capacity to intermediate trades.
    - Market liquidity deteriorated to its worst level since the GFC.
    - Pricing dislocations among Treasuries with similar cash flows and between cash Treasuries and Treasury futures emerged.
    - The Federal Reserve initiated massive asset purchases and other initiatives to restore market functioning.
- Policy responses and calls for greater resilience:
  - Academic and policy calls to make the Treasury market more resilient emerged after Covid-related disruptions.
  - Interagency efforts: Interagency Working Group on Treasury Market Surveillance (IAWG) annual reports (IAWG 2021, 2022, 2023, 2024).
  - Significant steps taken include: launch of a standing liquidity backstop; improved market transparency; approval of rules to promote increased central clearing of trades.
- Scope and structure of the paper:
  - Reviews market structure and prior work on Treasury liquidity.
  - Describes factors that have meaningfully affected liquidity provision since the GFC.
  - Discusses effects of these changes on market functioning in normal times and times of stress.
  - Concludes with policy implications, including recent policy initiatives intended to promote market resilience.

### 3.1 Postcrisis Regulatory Framework
- 3.1.1. Changes in the framework:
  - Post-GFC regulatory tightening focused on enhancing capital, liquidity, and transparency.
  - Major US dealers became bank holding companies (BHCs) or affiliated with BHCs and were drawn into BHC regulatory frameworks, including Basel III and the 2010 Dodd-Frank Act (DFA).
  - Basel enhancements:
    - Basel 2.5 (introduced in 2011) introduced a stressed value at risk (VaR) measure and an incremental risk charge.
    - Basel III mandated three changes crucial for Treasury market functioning:
      - (i) Increased the minimum risk-based capital requirements relative to Basel II.
      - (ii) Introduced the supplementary leverage ratio (SLR), making activities such as repo funding of Treasury collateral subject to prudential capital requirements for bank-affiliated dealers.
      - (iii) Introduced the liquidity coverage ratio (LCR) and the net stable funding ratio (NSFR).
    - Short-term funding via repos with Treasury collateral is treated favorably under both the LCR and the NSFR.
    - Phasing and timing: tightening of capital requirements and the SLR were phased in between 2013 and 2015; the LCR became effective in 2015 and the NSFR in 2018. Some final components (Basel III endgame) remain to be implemented as of the time of writing.
  - DFA reforms:
    - Required stress testing (DFAST) integrated into the Federal Reserve’s Comprehensive Capital Analysis and Review (CCAR).
    - Tightened reporting and oversight for derivatives markets.
    - The Volcker Rule constrained proprietary trading; it went into effect in April 2014 and was loosened to some extent in 2020.
  - SEC actions:
    - Rule 15c3-5 (implemented in 2011) required brokers-dealers to have robust risk management controls.
    - Adoption/amendment of trade transparency rules (Rule 605, Rule 606, Rule 17a-8) and rules governing dealer conduct with customers (Regulation Best Interest, Form CRS).
- 3.1.2. Impact of the changes:
  - BCBS (2022) evaluation:
    - Basel III reforms coincided with improved capital and liquidity positions, particularly at banks with the weakest capital and liquidity ratios.
    - Found no evidence that reforms impaired the aggregate supply of credit.
    - Banks complying with Basel III lowered their costs of equity and debt, with more pronounced declines for banks with lower initial capital ratios.
  - Market participant observations and academic findings:
    - Regulations broadly affected dealers’ business models and market-making practices.
    - Risk weights and credit risk charges increased trading costs for corporate bonds and credit derivatives; leverage ratio increases financing costs for relatively lower risk activities such as repo operations.
    - Research highlights:
      - Ryan & Toomey (2021): SLR can be a binding constraint for dealer intermediation in normal and stressed times.
      - Cochran et al. (2023): For banks with large dealer affiliates, the SLR has implied higher capital requirements than risk-based capital rules in recent years.
      - Adrian et al. (2017), Du, Hébert & Li (2023): Post-GFC regulatory tightening increased dealers’ balance sheet costs.
      - Gromb & Vayanos (2010); He & Krishnamurthy (2013); Breckenfelder & Ivashina (2023): Regulatory tightening attenuated liquidity provision.
      - Bräuning & Stein (2024): Relaxation of the SLR in 2020, when Treasuries and reserves were temporarily exempt, eased dealer constraints and improved Treasury market liquidity.
      - Boyarchenko et al. (2018), Du, Hébert & Li (2023), Wu & Jarrow (2024): Leverage ratios help explain negative swap spreads.
    - Financial Stability Board (2022) study: Evidence of post-GFC regulatory reforms adversely impacting dealers’ liquidity support is inconclusive.
  - Macroeconomic interaction:
    - Dealer willingness to provide liquidity is limited by macroeconomic conditions; leverage is procyclical.
    - Dealer risk-taking and balance sheet size evolve with market developments and macroeconomy.
    - Covid-19 demonstrated constrained dealer risk-taking capacity under extreme uncertainty.

### 3.2 Changing Auction Participation
- Question: Whether dealer intermediation of new Treasury issuance has changed materially amid balance sheet constraints and fast-growing public debt.
- Evidence on auction participation:
  - Dealer and broker purchases:
    - 65% of new note and bond issuance in January 2005 to July 2007.
    - 17% of new note and bond issuance in January 2022 to December 2024.
  - Increased participation from investment funds is associated with a well-intermediated auction; auction tail increases with volatility.
- Regression results (Table 1: Treasury auctions and dealer intermediation) — coefficients from regressing auction tails on primary dealer purchase shares and swaption implied volatilities (standard errors in parentheses):
  - 2-Year:
    - Dealer share: -0.009 (0.016)
    - Volatility: 0.021*** (0.004)
    - Constant: 1.902*** (0.322)
    - Observations: 216
    - Adjusted R2: 0.365
  - 3-Year:
    - Dealer share: 0.003 (0.016)
    - Volatility: 0.030*** (0.003)
    - Constant: 1.625*** (0.331)
    - Observations: 186
    - Adjusted R2: 0.485
  - 5-Year:
    - Dealer share: 0.038 (0.024)
    - Volatility: 0.023*** (0.006)
    - Constant: 2.893*** (0.465)
    - Observations: 216
    - Adjusted R2: 0.258
  - 7-Year:
    - Dealer share: 0.067*** (0.025)
    - Volatility: 0.027*** (0.005)
    - Constant: 2.932*** (0.440)
    - Observations: 185
    - Adjusted R2: 0.277
  - 10-Year:
    - Dealer share: 0.073*** (0.018)
    - Volatility: 0.028*** (0.005)
    - Constant: 2.617*** (0.412)
    - Observations: 209
    - Adjusted R2: 0.308
  - 30-Year:
    - Dealer share: 0.151*** (0.042)
    - Volatility: 0.028*** (0.006)
    - Constant: 3.573*** (0.511)
    - Observations: 194
    - Adjusted R2: 0.298
- Notes on Table 1:
  - Auction tail = highest yield accepted minus median yield of accepted bids (in basis points).
  - Dealer share = primary dealer purchase share less the average purchase share from the previous six auctions for the corresponding term (in percent).
  - Volatility = term-matched swaption implied volatility as of the previous day’s market close (in basis points).
  - Sample period generally runs from January 2007 to December 2024 with start-date exceptions for specific terms due to data limitations.
  - Newey-West standard errors reported. One, two, and three asterisks indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

### 3.3 Changing Ownership Patterns
- Global demand and reserve holdings:
  - U.S. Treasury securities are a vital reserve asset for central banks.
  - Central banks’ demand for U.S. dollars has remained broadly stable over the past 25 years.
  - Roughly 10 percentage point decline in the dollar's share of international reserves: from about 70 percent in 2000 to about 60 percent in 2024.
- Mutual funds, MMFs, hedge funds, and asset managers:
  - Shift in MMFs’ assets-under-management (AuM): significant decline in prime MMF AuM and commensurate increase in government- and Treasury-only MMF AuM.
  - During March 2020 dash for cash, MMFs generally received inflows; redemption pressures focused on prime MMFs.
  - Mutual funds’ redemption risks:
    - Bond mutual funds sold $236 billion in Treasuries in the first quarter of 2020, accounting for about one-third of total Treasury sales that quarter.
    - Funds sold the most liquid securities first to meet redemptions.
  - Asset managers’ futures positions:
    - Asset managers significantly increased long futures positions over the past decade.
    - Use of futures to take directional positions or extend duration at lower cost creates basis trades and attracts hedge funds.
  - Hedge funds and basis trade vulnerabilities:
    - Basis trade relies on leverage with financing in the repo market at low cost and overnight maturities, increasing demand for dealer balance sheet space.
    - Trades tend to be concentrated among a small number of institutions, adding concentration risk.
    - Hedge fund cash positioning largely offsets mutual fund futures positioning, increasing financial system interconnectedness.
- Implications for intermediation:
  - Mutual funds and hedge funds can support liquidity in normal times but may be forced to unwind positions in stress, increasing liquidity demand.
  - Dealers’ intermediation capacity may be constrained when liquidity is most needed.
  - Covid-19 episode:
    - Dealers were overwhelmed by extraordinary selling pressure from foreign official accounts, mutual funds, and hedge funds.
    - Dealers bought Treasuries and other fixed income securities in huge quantities but balance sheet constraints and internal risk limits prevented them from meeting demand on typical terms, amplifying volatility.
    - Fleming et al. (2022): increases in primary dealers’ net Treasury note and bond and agency MBS holdings were the highest on record for the four weeks ending March 18.
    - FSB (2022): extreme uncertainty drove dealer VaRs higher, aggregate risk limits did not change substantially, leading to reluctance to expand intermediation further.

### 3.4 Electronic Trading
- Adoption and effects:
  - Fully electronic interdealer broker (IDB) platforms launched in 1999 and 2000 and quickly gained market share.
  - Electronic trading improved liquidity: narrower bid-ask spreads, lower price impact, increased order book depth, and lower mean and variance of spreads.
  - Low latency trading associated with better day-to-day liquidity: narrower spreads, higher depth, lower price impact.
  - Electronic trading reduces trading costs, enhances competition, and promotes price discovery.
- Risks and episodes:
  - Concerns about operational risk, disruptive market practices, and short-term disruptions.
  - Knight Capital lost $440 million in 45 minutes in August 2012 due to an electronic trading glitch.
  - November 2023 cyberattack on a major clearing firm adversely affected market liquidity and illustrated clearing-channel risks.
- Role of principal trading firms (PTFs):
  - PTFs opened to electronic IDBs around 2004 and by the October 15, 2014 flash rally accounted for most IDB activity.
  - PTFs trade for their own account, use algorithmic strategies, carry little inventory overnight, and typically provide intraday but not inter-day liquidity.
  - October 15, 2014 flash rally: 10-year Treasury yield dropped 16 basis points and rebounded in a narrow 12-minute window; interagency staff report noted PTFs’ prominent role and mixed behavior during the event.
  - During the pandemic, PTFs increased activity sharply in late February and early March 2020 but then decreased level and share of activity in mid-March 2020 as volatility surged; PTFs reduced intermediation capacity relative to dealers, exacerbating dealers’ intermediation challenges.

### 4. BROKER-DEALER INTERMEDIATION AND TREASURY MARKET LIQUIDITY
- Dealer intermediation capacity — trends and drivers:
  - Indicative measures: Gross Treasury positions; Estimated market-making positions.
  - Both measures plunged during the GFC; some recovery followed, but since 2011 these measures have trended down as a percent of debt outstanding (dollar levels increased somewhat in recent years, especially since the pandemic).
  - Drivers of reduced capacity: dealers’ concerns about risk exposures, heightened volatility raising VaR, internal risk limits, regulatory constraints, and potential negative feedback loops between volatility and dealer behavior.
- Empirical relationship: liquidity, volatility (MOVE), and unit VaR — OLS results (Table 2)
  - Regression framework:
    - Dependent variable: year-to-year changes in the Treasury market liquidity index of Adrian, Fleming & Vogt (2023).
    - Independent variables: implied Treasury market volatility (MOVE) and unit VaR.
    - Sample: January 2000 through December 2024. Variables at quarter-end and standardized to mean 0 and standard deviation 1.
  - Reported coefficients and statistics:
    - Column (1): Volatility coefficient = 0.665*** (0.055); Observations = 96; Adjusted R2 = 0.533.
    - Column (2): Volatility coefficient = 0.595*** (0.074); Observations = 96; Adjusted R2 = 0.198.
    - Column (3): Value at risk coefficient = 0.587*** (0.127); Value at risk secondary coefficient = 0.255* (0.133); Observations = 96; Adjusted R2 = 0.562.
    - Notes: Newey-West standard errors in parentheses. One, two, and three asterisks indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
  - Interpretation:
    - Implied volatility (MOVE) is strongly positively associated with market illiquidity.
    - Dealer unit VaR also provides explanatory power for market liquidity when included.
- Quantile regressions — nonlinear effects and tail dynamics:
  - Method: quantile regressions at the 5th, 10th, 25th, 50th, 75th, 90th, and 95th percentiles using weekly data (January 2000 through December 2024), variables standardized.
  - Key findings:
    - Volatility (MOVE) exhibits strong nonlinear dependence on market liquidity: slope coefficients rise sharply at higher quantiles, indicating a larger impact of volatility when illiquidity is higher.
    - Dealer unit VaR shows a similar nonlinear pattern: slope coefficients rise with higher quantiles and turn positive at the upper quantiles.
    - Up to lower/mid quantiles, market illiquidity is negatively correlated with dealer VaR; in the upper quantiles, the relationship becomes strongly positive.
  - Broader interpretation:
    - Volatility and dealer risk taking provide additional explanatory power for extreme quantiles of market illiquidity.
    - These results are consistent with the mechanism that dealer risk constraints can amplify illiquidity, especially in stress episodes.
- Overall conclusions from the section:
  - Empirical results reveal a significant deterioration in market liquidity dynamics when dealers face intermediation challenges.
  - The US Treasury market underwent substantial transformations following the GFC that have left inherent vulnerabilities in crises.
  - Dealer risk constraints, measured by implied volatility (MOVE) and unit VaR, play an important role in explaining both average and extreme episodes of Treasury market illiquidity.

### 6. CONCLUSION
- Evolution of the Treasury market since the GFC:
  - Post-crisis regulatory changes bolstered financial resilience.
  - Ownership and buyer composition at auctions changed: investment funds at least partially displaced dealers in the primary market.
  - Electronic trading facilitated entry of PTFs into the interdealer market; PTFs engage in high-speed cross-market arbitrage.
- Effects on market functioning and liquidity:
  - Overall assessment: structural changes have on balance promoted market functioning and liquidity.
  - Primary market: dealers reduced their footprint; increased participation by investment funds ensured that auction metrics have not deteriorated.
  - IDB market: dealers account for a much smaller share of activity; entry of PTFs and increased activity of asset managers helped maintain high liquidity.
  - Caveats: concerns remain about non-dealer intermediation during stress because PTFs may pull back and asset managers face redemption pressures.
- Market stress in March 2020 and policy response:
  - During the dash-for-cash, mutual funds, hedge funds, and foreign central banks sold massive quantities of Treasuries.
  - Dealers stepped in but were overwhelmed, producing pricing dislocations and unusually poor illiquidity.
  - The Federal Reserve initiated asset purchases at an unprecedented speed and scale to restore market functioning.
- Policy and structural reforms pursued since 2020:
  - Notable changes include: launch of a standing repo facility; approval of rules to increase the share of trades that are centrally cleared; release of transactions data for the most actively traded Treasuries.
  - Additional efforts targeted promoting transparency of investment funds, ensuring consistent treatment of market participants and infrastructure, and understanding barriers to a market structure that relies less on dealers.
- Research outlook:
  - Increased data availability and recent reforms suggest Treasury market liquidity and resilience will remain a fruitful area of research.

*Source: sr1146 - New York Fed staff report content.*

### 1. INTRODUCTION

### 1. INTRODUCTION

### Role and importance of US Treasury securities
- US Treasury securities serve as a benchmark in global capital markets because of their "exceptional creditworthiness and liquidity."
- Key uses: manage interest rate risk, price other securities, collateralize financing transactions, implement monetary policy, and act as a reserve asset.
- Investors pay a premium for Treasuries because of their liquidity, leading to lower yields and hence lower US government borrowing costs (examples cited: Amihud & Mendelson 1991, Longstaff 2004, and Krishnamurthy & Vissing-Jorgensen 2012).
- Liquidity is a central attribute closely linked to Treasuries’ pricing and popularity.

### Historical episodes of market stress and implications for liquidity
- Major events calling Treasury market liquidity into question:
  - The 2007-09 global financial crisis (GFC).
  - Post-crisis regulatory changes and the increasing role of principal trading firms (PTFs).
  - The October 15, 2014 flash rally (Joint Staff Report 2015).
  - Covid-19-related developments in March 2020 (the "dash-for-cash").
- March 2020 disruptions:
  - Massive customer selling overwhelmed dealers’ capacity to intermediate trades (Duffie 2020).
  - Market liquidity deteriorated to its worst level since the GFC (Fleming & Ruela 2020).
  - Pricing dislocations arose among Treasuries with similar cash flows and between cash Treasuries and Treasury futures (Duffie 2020, Schrimpf, Shin & Sushko 2020).
  - The Federal Reserve initiated massive asset purchases and other initiatives to restore market functioning (Vissing-Jorgensen 2021).
- Research documenting liquidity events and market structure shifts:
  - Nguyen et al. (2020) characterize GFC disruptions for most actively traded Treasuries.
  - Musto, Nini & Schwarz (2018) examine pricing discrepancies in less-active securities.
  - Adrian et al. (2017) discuss post-crisis regulatory changes affecting dealer-intermediated markets.
  - The Joint Staff Report (2015) documents PTFs accounting for most interdealer activity by October 2014.

### Policy responses and calls for greater resilience
- Covid-related disruptions prompted academic and policy calls to make the Treasury market more resilient (examples cited: Duffie 2020, Liang & Parkinson 2020, Brookings 2021, Group of Thirty 2021).
- Interagency efforts and actions:
  - Interagency Working Group on Treasury Market Surveillance (IAWG) annual reports (IAWG 2021, 2022, 2023, 2024) have driven many policy efforts.
  - Significant steps taken include:
    - Launch of a standing liquidity backstop.
    - Improved market transparency.
    - Approval of rules to promote increased central clearing of trades.

### Scope and structure of the paper
- This paper assesses US Treasury market functioning given recent market developments and associated policy responses.
- Coverage includes:
  - A review of the market’s structure and prior work on Treasury liquidity.
  - A description of factors that have meaningfully affected liquidity provision since the GFC.
  - Discussion of the effects of these changes on market functioning in both normal times and times of stress.
  - A concluding discussion of policy implications, including recent policy initiatives intended to promote market resilience.

*Source: sr1146 - 1. INTRODUCTION (staff report content).*

### 3.1 Postcrisis Regulatory Framework

### 3.1 Postcrisis Regulatory Framework

### 3.1.1. Changes in the framework
- Post-GFC regulatory tightening focused on enhancing capital, liquidity, and transparency of financial institutions.
- Major US dealers became bank holding companies (BHCs) or affiliated with BHCs through mergers and were drawn into BHC regulatory frameworks, including Basel III and the 2010 Dodd-Frank Act (DFA).
- Key elements of Basel enhancements:
  - Basel 2.5 (introduced in 2011) upgraded measurement and management of market risk exposures, introducing a stressed value at risk (VaR) measure and an incremental risk charge.
  - Basel III mandated three changes crucial for Treasury market functioning:
    - (i) Increased the minimum risk-based capital requirements relative to Basel II, increasing the cost of capital to banks of risk taking.
    - (ii) Introduced the supplementary leverage ratio (SLR), making activities such as repo funding of Treasury collateral subject to prudential capital requirements for bank-affiliated dealers.
    - (iii) Introduced the liquidity coverage ratio (LCR) and the net stable funding ratio (NSFR), enhancing resilience to liquidity shocks and anchoring maturity transformation to retail deposits and other stable sources of debt financing.
  - Short-term funding via repos with Treasury collateral is treated favorably under both the LCR and the NSFR.
  - The tightening of capital requirements and the SLR were phased in between 2013 and 2015; the LCR became effective in 2015 and the NSFR in 2018. Some final components (Basel III endgame) remain to be implemented as of the time of writing.
- DFA reforms:
  - Required stress testing for large financial institutions and their affiliated dealers (DFAST), integrated into the Federal Reserve’s Comprehensive Capital Analysis and Review (CCAR).
  - Tightened reporting and oversight for derivatives markets.
  - The Volcker Rule constrained proprietary trading and limited investments in hedge funds and private equity; it went into effect in April 2014 and was loosened to some extent in 2020.
- SEC regulatory actions:
  - Rule 15c3-5 (implemented in 2011) required brokers-dealers to have robust risk management controls and supervisory procedures.
  - Post-GFC adoption and amendment of trade transparency rules (Rule 605, Rule 606, Rule 17a-8) and rules governing dealer conduct with customers (Regulation Best Interest, Form CRS).

### 3.1.2. Impact of the changes
- BCBS (2022) evaluation:
  - Basel III reforms coincided with improved capital and liquidity positions, particularly at banks with the weakest capital and liquidity ratios.
  - Found no evidence that reforms impaired the aggregate supply of credit.
  - Banks complying with Basel III lowered their costs of equity and debt, with more pronounced declines for banks with lower initial capital ratios.
- Market participant observations and academic findings:
  - Post-crisis regulations broadly affected dealers’ business models and market-making practices (CGFS 2014).
  - Risk weights and credit risk charges increased trading costs for corporate bonds and credit derivatives; leverage ratio increases financing costs for relatively lower risk activities such as repo operations.
  - Market participants identify the Basel III leverage ratio and higher risk-weighted capital as most significant for sovereign bonds; revisions to the market risk framework are primary drivers for corporate bond charges (CGFS 2016).
  - Research findings:
    - Ryan & Toomey (2021): SLR can be a binding constraint for dealer intermediation in normal and stressed times.
    - Cochran et al. (2023): For banks with large dealer affiliates, the SLR has implied higher capital requirements than risk-based capital rules in recent years, potentially affecting willingness and ability to intermediate in Treasury markets.
    - Adrian et al. (2017), Du, Hébert & Li (2023): Post-GFC regulatory tightening increased dealers’ balance sheet costs.
    - Gromb & Vayanos (2010); He & Krishnamurthy (2013); Breckenfelder & Ivashina (2023): Regulatory tightening attenuated liquidity provision.
    - Bräuning & Stein (2024): Relaxation of the SLR in 2020, when Treasuries and reserves were temporarily exempt, eased dealer constraints and improved Treasury market liquidity.
    - Boyarchenko et al. (2018), Du, Hébert & Li (2023), Wu & Jarrow (2024): Leverage ratios help explain negative swap spreads.
  - Financial Stability Board (2022) study: Evidence of post-GFC regulatory reforms adversely impacting dealers’ liquidity support is inconclusive.
- Macroeconomic environment interaction:
  - Dealer willingness to provide liquidity is also limited by macroeconomic conditions; leverage is procyclical (Adrian & Shin 2014).
  - Dealer risk-taking and balance sheet size evolve with market developments and macroeconomy (Adrian et al. 2015b, 2017, 2013; Chen et al. 2021).
  - Covid-19 demonstrated constrained dealer risk-taking capacity under extreme uncertainty (FSB 2022).

### 3.2. Changing Auction Participation
- Question: Whether dealer intermediation of new Treasury issuance has changed materially amid balance sheet constraints and fast-growing public debt.
- Evidence on auction participation:
  - Dealer and broker purchases:
    - 65% of new note and bond issuance in January 2005 to July 2007.
    - 17% of new note and bond issuance in January 2022 to December 2024.
  - Increased participation from investment funds is associated with a well-intermediated auction; auction tail increases with volatility.
- Regression results (Table 1: Treasury auctions and dealer intermediation) — coefficients from regressing auction tails on primary dealer purchase shares and swaption implied volatilities (standard errors in parentheses):
  - 2-Year:
    - Dealer share: -0.009 (0.016)
    - Volatility: 0.021*** (0.004)
    - Constant: 1.902*** (0.322)
    - Observations: 216
    - Adjusted R2: 0.365
  - 3-Year:
    - Dealer share: 0.003 (0.016)
    - Volatility: 0.030*** (0.003)
    - Constant: 1.625*** (0.331)
    - Observations: 186
    - Adjusted R2: 0.485
  - 5-Year:
    - Dealer share: 0.038 (0.024)
    - Volatility: 0.023*** (0.006)
    - Constant: 2.893*** (0.465)
    - Observations: 216
    - Adjusted R2: 0.258
  - 7-Year:
    - Dealer share: 0.067*** (0.025)
    - Volatility: 0.027*** (0.005)
    - Constant: 2.932*** (0.440)
    - Observations: 185
    - Adjusted R2: 0.277
  - 10-Year:
    - Dealer share: 0.073*** (0.018)
    - Volatility: 0.028*** (0.005)
    - Constant: 2.617*** (0.412)
    - Observations: 209
    - Adjusted R2: 0.308
  - 30-Year:
    - Dealer share: 0.151*** (0.042)
    - Volatility: 0.028*** (0.006)
    - Constant: 3.573*** (0.511)
    - Observations: 194
    - Adjusted R2: 0.298
- Notes on Table 1:
  - Auction tail = highest yield accepted minus median yield of accepted bids (in basis points).
  - Dealer share = primary dealer purchase share less the average purchase share from the previous six auctions for the corresponding term (in percent).
  - Volatility = term-matched swaption implied volatility as of the previous day’s market close (in basis points).
  - Sample period generally runs from January 2007 to December 2024 with start-date exceptions for specific terms due to data limitations.
  - Newey-West standard errors reported. One, two, and three asterisks indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

### 3.3. Changing Ownership Patterns
- Global demand and reserve holdings:
  - U.S. Treasury securities are a vital reserve asset for central banks.
  - Central banks’ demand for U.S. dollars has remained broadly stable over the past 25 years (Arslanalp, Eichengreen & Simpson-Bell 2022, 2024).
  - Roughly 10 percentage point decline in the dollar's share of international reserves: from about 70 percent in 2000 to about 60 percent in 2024.
- Mutual funds, MMFs, hedge funds, and asset managers:
  - Shift in MMFs’ assets-under-management (AuM): significant decline in prime MMF AuM and commensurate increase in government- and Treasury-only MMF AuM.
  - During March 2020 dash for cash, MMFs generally received inflows; redemption pressures focused on prime MMFs (PWG 2020).
  - Mutual funds’ redemption risks:
    - Bond mutual funds sold $236 billion in Treasuries in the first quarter of 2020, accounting for about one-third of total Treasury sales that quarter (Ma, Xiao & Zeng 2022).
    - Funds sold the most liquid securities first to meet redemptions; outflows contributed to selling pressures in fixed income markets.
  - Asset managers’ futures positions:
    - Asset managers significantly increased long futures positions over the past decade (Figure 5 referenced).
    - Use of futures to take directional positions or extend duration at lower cost creates basis trades; attracts hedge funds and bolsters their role.
  - Hedge funds and basis trade vulnerabilities:
    - Basis trade relies on leverage with financing in the repo market at low cost and overnight maturities, increasing demand for dealer balance sheet space.
    - Trades tend to be concentrated among a small number of institutions, adding concentration risk (Banegas & Monin 2023; IAWG 2023; IMF 2024a).
    - Hedge fund cash positioning largely offsets mutual fund futures positioning, increasing financial system interconnectedness (Barth et al. 2024).
- Implications for intermediation:
  - Mutual funds and hedge funds can support liquidity in normal times but may be forced to unwind positions in stress, increasing liquidity demand.
  - Dealers’ intermediation capacity may be constrained when liquidity is most needed.
  - Covid-19 episode:
    - Dealers were overwhelmed by extraordinary selling pressure from foreign official accounts, mutual funds, and hedge funds.
    - Dealers bought Treasuries and other fixed income securities in huge quantities but balance sheet constraints and internal risk limits prevented them from meeting demand on typical terms, amplifying volatility.
    - Policy response included the Federal Reserve’s market functioning purchases of Treasuries and agency MBS.
    - Fleming et al. (2022): increases in primary dealers’ net Treasury note and bond and agency MBS holdings were the highest on record for the four weeks ending March 18.
    - FSB (2022): extreme uncertainty drove dealer VaRs higher, aggregate risk limits did not change substantially, leading to reluctance to expand intermediation further.

### 3.4. Electronic Trading
- Adoption and effects:
  - Fully electronic interdealer broker (IDB) platforms launched in 1999 and 2000 and quickly gained market share.
  - Evidence indicates electronic trading improved liquidity: narrower bid-ask spreads, lower price impact, increased order book depth, and lower mean and variance of spreads (Mizrach & Neely 2006; Adrian, Fleming & Vogt 2023).
  - Low latency trading associated with better day-to-day liquidity: narrower spreads, higher depth, lower price impact.
  - Electronic trading reduces trading costs, enhances competition, and promotes price discovery (Domowitz 2002).
- Risks and episodes:
  - Concerns about operational risk, disruptive market practices, and short-term disruptions (Treasury Market Practices Group 2015).
  - Knight Capital lost $440 million in 45 minutes in August 2012 due to an electronic trading glitch (BBC Magazine, August 11, 2012).
  - November 2023 cyberattack on a major clearing firm adversely affected market liquidity and illustrated clearing-channel risks (Wall Street Journal, November 9, 2023).
- Role of principal trading firms (PTFs):
  - PTFs opened to electronic IDBs around 2004 and by October 15, 2014 flash rally accounted for most IDB activity (Joint Staff Report 2015).
  - PTFs trade for their own account, use algorithmic strategies, carry little inventory overnight, and typically provide intraday but not inter-day liquidity.
  - October 15, 2014 flash rally: 10-year Treasury yield dropped 16 basis points and rebounded in a narrow 12-minute window; interagency staff report noted PTFs’ prominent role and mixed behavior during the event.
  - During the pandemic, PTFs increased activity sharply in late February and early March 2020 but then decreased level and share of activity in mid-March 2020 as volatility surged; PTFs reduced intermediation capacity relative to dealers, exacerbating dealers’ intermediation challenges (Fleming et al. 2022).

*Source: sr1146 - 3.1 Postcrisis Regulatory Framework — https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr1146.pdf*

### 4. BROKER-DEALER INTERMEDIATION AND TREASURY MARKET LIQUIDITY

### 4. BROKER-DEALER INTERMEDIATION AND TREASURY MARKET LIQUIDITY

### Dealer intermediation capacity — trends and drivers
- Indicative measures of dealer intermediation capacity include:
  - Gross Treasury positions (sum of short and long positions).
  - Estimated market-making positions (sum across dealers of the minimum of long and short positions by security type and maturity).
- Historical evolution:
  - Both measures plunged during the GFC.
  - Some recovery occurred in the years immediately following the GFC.
  - Since 2011 these measures have trended down as a percent of debt outstanding (though dollar levels have increased somewhat in recent years, especially since the pandemic).
- Drivers of reduced intermediation capacity:
  - Limits on intermediation capacity reflect dealers’ concerns about risk exposures.
  - In stress periods, heightened volatility increases dealers’ risks and causes risk measures such as value at risk (VaR) to rise sharply.
  - Internal risk management processes and regulatory constraints limit dealers’ ability to respond as demand for intermediation intensifies.
  - A negative feedback loop can arise between volatility and dealer behavior, potentially producing markedly worse liquidity outcomes.

### Empirical relationship: liquidity, volatility (MOVE), and unit VaR — OLS results (Table 2)
- Regression framework:
  - Regressions use year-to-year changes in the Treasury market liquidity index of Adrian, Fleming & Vogt (2023).
  - Independent variables: implied Treasury market volatility (MOVE index) and unit VaR (based on Adrian & Shin (2014)).
  - Sample: January 2000 through December 2024.
  - All variables measured at quarter-end and standardized to mean 0 and standard deviation 1.
  - Volatility and VaR data are from Bloomberg.
- Reported coefficients and statistics (preserve exactly as in source):
  - Column (1): Volatility coefficient = 0.665*** (standard error (0.055)); Observations = 96; Adjusted R2 = 0.533.
  - Column (2): Volatility coefficient = 0.595*** (standard error (0.074)); Observations = 96; Adjusted R2 = 0.198.
  - Column (3): Value at risk coefficient = 0.587*** (standard error (0.127)); Value at risk secondary coefficient = 0.255* (standard error (0.133)); Observations = 96; Adjusted R2 = 0.562.
  - Notes: Newey-West standard errors in parentheses. One, two, and three asterisks indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
- Interpretation from OLS results:
  - Implied volatility (MOVE) is strongly positively associated with market illiquidity (higher MOVE → worse liquidity).
  - Dealer unit VaR also provides explanatory power for market liquidity when included.

### Quantile regressions — nonlinear effects and tail dynamics
- Method and purpose:
  - Quantile regressions estimate conditional slope coefficients for the 5th, 10th, 25th, 50th, 75th, 90th, and 95th percentiles of the liquidity distribution.
  - Objective is to understand how market liquidity depends on dealer risk constraints across the entire conditional liquidity distribution, not just on average.
  - Weekly data used: liquidity and volatility averages of daily values; VaR interpolated from quarter-end values. All variables standardized to mean 0 and standard deviation 1 before estimation. Sample: January 2000 through December 2024.
- Key findings (from Figure 8 description and text):
  - Volatility (MOVE) exhibits strong nonlinear dependence on market liquidity:
    - Slope coefficients rise sharply at higher quantiles (when illiquidity is higher), indicating a larger impact of volatility when liquidity is low.
    - Higher volatility is associated with more illiquid markets, and this relationship becomes stronger in the upper tail of illiquidity.
    - The difference in illiquidity responses between lower and upper quantiles is statistically significant.
  - Dealer unit VaR shows a similar nonlinear pattern:
    - Slope coefficients rise with higher quantiles and turn positive at the upper quantiles.
    - Up to lower/mid quantiles, market illiquidity is negatively correlated with dealer VaR (periods of adequate liquidity can correlate with lower dealer intermediation).
    - In the upper quantiles, the relationship becomes strongly positive, consistent with increased perceived dealer risk in stressed or illiquid conditions leading to reduced market-making and amplified illiquidity.
- Broader interpretation:
  - Volatility and dealer risk taking provide additional explanatory power for extreme quantiles of market illiquidity.
  - These results are broadly consistent with Duffie et al. (2023) and illustrate the economic mechanism highlighted by Adrian & Shin (2014): dealer risk constraints can amplify illiquidity, especially in stress episodes.

### Overall conclusions from the section
- Empirical results reveal a significant deterioration in market liquidity dynamics when dealers face intermediation challenges.
- The US Treasury market underwent substantial transformations following the GFC that have left inherent vulnerabilities, particularly during crises when demand for dealer intermediation can increase dramatically.
- Dealer risk constraints, measured by implied volatility (MOVE) and unit VaR, play an important role in explaining both average and extreme episodes of Treasury market illiquidity.

*Source: sr1146 - 4. BROKER-DEALER INTERMEDIATION AND TREASURY MARKET LIQUIDITY (staff report content).*

### 6. CONCLUSION

### 6. CONCLUSION

### Evolution of the Treasury market since the GFC
- Post-crisis regulatory changes bolstered the resilience of financial institutions and the broader financial system.
- Ownership and buyer composition at auctions changed: investment funds at least partially displaced dealers in the primary market.
- Electronic trading facilitated entry of PTFs into the interdealer market; PTFs engage in high-speed cross-market arbitrage.

### Effects on market functioning and liquidity
- Overall assessment: these structural changes have on balance promoted market functioning and liquidity.
- Primary market:
  - Dealers reduced their footprint in the primary market.
  - Increased participation by investment funds ensured that auction metrics have not deteriorated.
- IDB (interdealer brokerage) market:
  - Dealers account for a much smaller share of activity.
  - Entry of PTFs and increased activity of asset managers helped maintain high liquidity.
- Caveats:
  - Concerns remain about non-dealer intermediation during stress because PTFs may pull back and asset managers face redemption pressures.
  - Post-GFC regulatory changes contribute to dealer resilience.

### Market stress in March 2020 and policy response
- Market disruptions since the GFC culminated in notable strains in March 2020 with the onset of the Covid-19 pandemic.
- During the dash-for-cash, mutual funds, hedge funds, and foreign central banks sold massive quantities of Treasuries.
- Dealers stepped in as counterparts but were overwhelmed, producing pricing dislocations and unusually poor illiquidity.
- The Federal Reserve initiated asset purchases at an unprecedented speed and scale to restore market functioning.

### Policy and structural reforms pursued since 2020
- Notable policy and market-structure changes include:
  - Launch of a standing repo facility.
  - Approval of rules to increase the share of trades that are centrally cleared.
  - Release of transactions data for the most actively traded Treasuries.
- Additional efforts have targeted:
  - Promoting transparency of investment funds.
  - Ensuring consistent treatment of market participants and infrastructure.
  - Understanding barriers to a market structure that relies less on dealers.

### Research outlook
- Increased data availability and the recent reforms suggest that Treasury market liquidity and its resilience will remain a fruitful area of research in the years to come.

*Source: sr1146 - 6. CONCLUSION*

---


_Source: https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr1146.pdf_