Arbitrage, liquidity and exit: The repo and federal funds market before, during, and after the financial crisis

Transcription

1 Arbitrage, liquidity and exit: The repo and federal funds market before, during, and after the financial crisis Morten Bech (FRBNY), Elizabeth Klee (FRB), and Viktors Stebunovs (FRB) May 21, 2011 The views expressed in this paper are solely of the authors.

2 Overview Estimate the relationship between the Treasury GC repo rate and fed funds rate during three periods Normal : Aug 2007 Crisis : Aug Dec 2008 Aftermath : Dec Jun 2010 Investigate three issues in the fed funds and repo markets Policy transmission Pricing of risk Liquidity effects Simulate the effects of draining reserve balances on the fed funds and repo rates Study balance sheet mechanics of draining to clarify the simulation results

3 Background A repo is a sale of a security coupled with an agreement to repurchase it later For the cash borrower, it is a repo; for the cash lender, it is a reverse repo Treasury General Collateral is the most prevalent collateral Fed funds are (overnight) unsecured loans of reserve balances at the Fed that market participants make to one another Participation in the fed funds market is limited to commercial banks, GSEs Fed funds are used to satisfy requirements and settle payments The Fed sets a target for the fed funds rate, implemented through Open Market Operations These two markets are tightly connected (in theory) The Fed involved in both through Open Market Operations There is a significant overlap among market participants Both markets went through a few structural changes

4 Together no longer? Federal Funds Overnight GC Repo Spread Basis points Daily.

5 Federal Reserve s balance sheet Primary, secondary, and seasonal credit 3,000 Federal Reserve Bank Assets and Liabilities 2,500 Support for specific institutions (ML LLCs, Bear, AIG) 2,000 1,500 Assets Other credit facilities (PDCF, AMLF, CPFF, TALF) Central bank liquidity swaps 1,000 Term auction credit Agency debt and mortgage backed securities holdings 500 Repurchase agreements and all other assets Treasury securities held outright $ Billions Federal Reserve notes in circulation 1,000 Reverse RPs, capital, and all other liabilities U.S. Treasury accounts 1,500 2,000 Liabilities Deposits of depository institutions and other deposits 2,500 3,000 Jan 3, 2007 Jul 4, 2007 Jan 2, 2008 Jul 2, 2008 Dec 31, 2008 Jul 1, 2009 Dec 30, 2009 Jun 30, 2010 Dec 29, 2010 Jun 29, 2011 Wednesdays Last updated May 13, 2011.

6 Share of Cash assets in Total assets for U.S. commercial banks Share of Cash Assets in Total Assets of U.S. Commercial Banks, percent 12% 10% 8% 6% 4% 2% 0%

7 Econometric strategy VEC model ( Normal / Crisis ): x t = α(β x t 1 + c 0 ) + Φ(L) x t + ΞZ t + e t VAR model ( Aftermath ): x t = µ 0 + Φ(L)x t + ΞZ t + e t t - daily, sample from 2002 to 2010 x t - vector of fed funds and Treasury GC repo rates β - cointegrating slope c 0 - long-run risk premium proxy (captures default probability and repo recovery rate) α - vector of adjustment speed coefficients Z t - vector of exogenous supply shifters and controls e t - GARCH/TARCH error

8 Results highlights: Repo rate Estimation results Sample periods Jan Aug Aug Dec Dec Jun Econometric model VEC/GARCH VEC/GARCH VAR/GARCH Cointergrating vector Constant NA Slope NA Dynamic equation Repo equation Speed of adjustment NA Factor miss NS NA Net issuance Lag of net issuance NS LSAP NA NA NS TGA NA NA R^ DW Number of observations Total balances Median, $ billions Range, $ billions

9 Results highlights: Fed funds rate Estimation results Sample periods Jan Aug Aug Dec Dec Jun Econometric model VEC/GARCH VEC/GARCH VAR/GARCH Cointergrating vector Constant NA Slope NA Dynamic equation Fed funds equation Speed of adjustment NA Factor miss NS NA Net issuance NS NS NS Lag of net issuance NS NS LSAP NA NA TGA NA NA R^ DW Number of observations Total balances Median, $ billions Range, $ billions

10 Default probability-recovery rate isocurve c 0 pins down a default probability-recovery rate isocurve: p = c 0 /(1 + c 0 + δ) P delta isocurves 100% 90% 80% Aug 07 Dec 08 Jan 02 Aug 07 Probability of default 70% 60% 50% 40% 30% 20% 10% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Recovery rate

11 Adjustment speed: Repo rate Overnight Treasury GC Repo Rate Jan 02 Aug 07 Aug 07 Dec Basis points Days

12 Adjustment speed: Fed funds rate Federal Funds Rate Basis points Jan 02 Aug 07 Aug 07 Dec Days

13 The liquidity effect and its identification The liquidity effect is a negative relationship between a money measure and an interest rate There is a demand curve for some type of money The liquidity effect is related to the slope of this demand curve Methodologies and evidence are mixed Hamilton (1997) and others find a significant liquidity effect at a daily frequency, but not at a lower frequency Exogenous shifters help to identify the liquidity effect For fed funds: errors in supply of reserves, balances at the Treasury s checking account at the Fed, LSAP For repo: LSAP, net Treasury issuance

14 Simulating a $980 billion reverse repo: Set up Recognizing that demand for reserve balances is likely non-linear, what are the potential effects of draining nearly $1 trillion in liquidity? Starting point: $1 trillion in reserve balances, repo and fed funds rates at 18 and 15 basis points respectively (early 2010) Consider two scenarios, both drain $980 billion in reserve balances but over 10 business days in the first over 30 business days in the second

15 Simulating a $980 billion reverse repo: First $900 billion basis points Overnight Treasury GC Repo Rate Scenario 1 Scenario Federal Funds Rate 80 Scenario 1 70 Scenario basis points basis points Spread between Fed Funds and Repo Rates Scenario 1 Scenario $ billions Cumulative Reverse Repo Volume Scenario 1 Scenario

16 Simulating a $980 billion reverse repo: Next $80 billion Overnight Treasury GC repo rate Federal Funds rate basis points Scenario 1 Scenario basis points Scenario 1 Scenario basis points Spread between Fed Funds and Repo Rates Scenario 1 Scenario $ billions Cumulative Reverse Repo Volume Scenario 1 Scenario

17 Balance sheet mechanics of draining: First round Before a Large Scale Reverse Repo Federal Reserve, $ bill. Commercial Banks, $ bill. Dealers/Mutual Funds, $ bill. SOMA 2,500 Res. Bal. 1,500 Res. Bal. 1,500 D/MF Dep. 600 Bank Dep. 600 Repo Rev. Repo Loans 6,700 Oth. Dep. 7,000 Rev. Repo Debt 9,450 Currency 1,000 Equity 600 Sec. Hold. 9,500 Equity 650 2,500 2,500 8,200 8,200 10,100 10,100 After a Large Scale Reverse Repo: First Round Effects Federal Reserve, $ bill. CommercialBanks Banks, $ bill. Dealers/Mutual Funds, $ bill. SOMA 2,500 Res. Bal. 1,000 Res. Bal. 1,000 D/MF Dep. 100 Bank Dep. 100 Repo Rev. Repo 500 Loans 6,700 Oth. Dep. 7,000 Rev. Repo 500 Debt 9,450 Currency 1,000 Equity 600 Sec. Hold. 9,500 Equity 650 2,500 2,500 7,700 7,700 10,100 10,100

18 Balance sheet mechanics of draining: Second round After a Large Scale Reverse Repo: First Round Effects Federal Reserve, $ bill. Commercial Banks, $ bill. Dealers/Mutual Funds, $ bill. SOMA 2,500 Res. Bal. 1,000 Res. Bal. 1,000 D/MF Dep. 100 Bank Dep. 100 Repo Rev. Repo 500 Loans 6,700 Oth. Dep. 7,000 Rev. Repo 500 Debt 9,450 Currency 1,000 Equity 600 Sec. Hold. 9,500 Equity 650 2,500 2,500 7,700 7,700 10,100 10,100 After a Large Scale Reverse Repo: Second Round Effects Federal Reserve, $ bill. Commercial Banks, $ bill. Dealers/Mutual Funds, $ bill. SOMA 2,500 Res. Bal. 1,000 Res. Bal. 1,000 D/MF Dep. 500 Bank Dep. 500 Repo 300 Rev. Repo 500 Loans 6,700 Oth. Dep. 6,600 Rev. Repo 500 Debt 9,450 Currency 1,000 New Loans 500 New Dep. 500 Sec. Hold. 9,400 Equity 650 Equity 600 2,500 2,500 8,200 8,200 10,400 10,400

19 Conclusions Monetary policy transmission still functioned through the crisis, although Link changed as reserve balances ballooned, interest on reserves came into effect, lower zero bound reached No more common stochastic trend Probability of default appears to have jumped notably in the midst of the crisis Arbitrage appears to have slowed down

20 Conclusions (continued) Demand curves for repo and fed funds appear to be non-linear Simulations suggest that draining reserve balances should boost both repo and fed funds rate and eventually restore the pre-crisis spread between the two rates Result is likely robust to the explicit modeling of interest on excess reserves Large-scale reverse repo done in a short period strains both the repo and fed funds markets notably Studying balance sheet mechanics of draining illustrates How these market strains come about and dissipate How draining might be expansionary and boost M2