Liquidity Risk In Corporate FIXED INCOME Bond Markets George Chacko Harvard Business School & IFL 1
Roadmap � Introduction � Liquidity Risk Research � Motivation � Liquidity Measurement � Liquidity Factor Construction � Empirical Results for Liquidity Risk � Practical Implications of Liquidity Risk 2
3 Capital Structure Arbitrage
Capital Structure Arbitrage Worldcom 6.95 30Y Issuance Date: Aug-1998 Amount: $1.75 BB Callable Spread over benchmark Treasury Strip (%) 16 Forecast Spread Caa Actual Traded Spread Ba2 14 12 10 Baa2 8 6 4 2 0 Oct-00 Apr-01 Oct-01 Apr-02 Jul-00 Jan-01 Jul-01 Jan-02 4
Corp Bond Market Liquidity Issue Trading Frequency - Median bond trades less than once a quarter 16000 100.00% 100.00% 90.00% 14000 80.00% Cumulative Percent Issues 12000 70.00% Number of Issues (Total: 24170) 10000 60.00% 8000 50.00% 39.23% 40.00% 6000 24.33% 30.00% 4000 13.40% 20.00% 2000 3.58% 10.00% 0 0.00% 1 Trade/Week 1 Trade/M 1 Trade/Qtr > 1 Trade/Qtr No Trades Trading Frequency 5 Source: State Street Global Markets
Liquidity Trend in Bond Mkt Average Trade Size Percentiles (millions of US dollars): YR94 YR95 YR96 YR97 YR98 YR99 YR00 YR01 YR02 YR03 YR04 MIN 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 10% 0.36 0.44 0.43 0.48 0.50 0.43 0.40 0.42 0.37 0.35 0.28 20% 0.75 0.83 0.84 0.94 0.97 0.82 0.72 0.73 0.67 0.66 0.55 30% 1.06 1.11 1.18 1.23 1.32 1.12 1.01 1.03 0.94 0.91 0.78 40% 1.43 1.50 1.63 1.68 1.78 1.54 1.38 1.43 1.22 1.16 1.03 50% 1.84 2.02 2.09 2.16 2.34 2.08 1.93 1.98 1.66 1.52 1.30 60% 2.30 2.63 2.71 2.85 3.10 2.88 2.56 2.65 2.21 1.97 1.65 70% 3.02 3.59 3.61 3.72 4.15 3.89 3.45 3.59 2.99 2.50 2.17 80% 4.10 4.99 4.97 5.06 5.56 5.31 5.02 5.12 4.30 3.46 2.88 90% 6.20 7.22 7.33 8.00 9.16 8.93 8.23 8.42 7.06 5.75 4.55 MAX 100.31 99.92 100.67 111.99 224.98 249.93 152.53 199.98 271.99 199.98 100.28 6 Source: State Street Global Markets
TRACE Comparison CUSIP 172967BC4 (CITIGROUP), 4/14/2004 -- 10/4/2002 TRACE High (via Bloomberg) 115 TRACE Low (via Bloomberg) TRACE 1MM+ High 113 TRACE 1MM+ Low 111 109 107 105 103 101 99 4/14/2004 4/21/2004 4/28/2004 5/5/2004 5/12/2004 5/19/2004 5/26/2004 6/2/2004 6/9/2004 6/16/2004 6/23/2004 6/30/2004 7/7/2004 7/14/2004 7/21/2004 7/28/2004 8/4/2004 8/11/2004 8/18/2004 8/25/2004 9/1/2004 9/8/2004 9/15/2004 9/22/2004 9/29/2004 7 Source: State Street Global Markets
Limitations of Liquidity Measures � Conventional Measures of Liquidity: � Trading Volume � Bid-Ask Spread � However, if securities are extremely illiquid, conventional measures don’t work well � Rather than looking at actual trading, one solution is to look at a security’s propensity to trade. 8
Latent Liquidity � Latent liquidity: a quantitative measure of propensity to trade for individual securities � Rationale: � For a bond dealer, it is easier to access a bond issue if it is held in high-turnover portfolios � If a bond issue is held by high-turnover funds, it is likely that security has a higher propensity to trade. � So, a security’s propensity to trade can be constructed by looking at the aggregate trading characteristics of owners of that security 9
10 Latent Liquidity Properties Higher Liquidity Lower Liquidity
11 Latent Liquidity Properties Higher Liquidity Lower Liquidity
12 Latent Liquidity Properties Higher Liquidity Lower Liquidity
Liquidity Risk Factor Construction � We sort the US corp bond universe into 3x3x3 = 27 buckets � Duration � Credit Risk � Latent Liquidity � We then form three portfolios: � HML Duration � LMH Credit Risk � LMH Latent Liquidity � These portfolios represent interest rate, credit, and liquidity risk factors 13
Liquidity Risk Factor Time Series 140 130 Liquidity Index 120 110 100 90 80 11/27/1993 4/11/1995 8/23/1996 1/5/1998 5/20/1999 10/1/2000 2/13/2002 6/28/2003 Date 14
Factor Regressions � With these factors, we can now do factor regressions to compute individual security betas. � We first compute credit, duration, and liquidity betas for the US corp bond universe. � We then do a 5x3x3 sort of these securities based on these betas – 5 liquidity portfolios, 3 credit portfolios, and 3 duration portfolios � Using these 45 portfolios, we then conduct a series of tests to check the importance of the liquidity risk factor. 15
Empirical Results Liquidity Risk Alpha Alphas of Portfolios Sorted on Liquidity Betas L M/L M H/M H H - L CAPM -0.54% 0.71% 1.25% 1.94% 2.36% 2.90% Duration -0.36% 0.69% 1.31% 2.13% 2.78% 3.14% Duration, Credit -0.56% 0.63% 1.09% 1.68% 2.15% 2.71% 16
Empirical Results Contribution of Liquidity: 1 Incremental R 2 of Liquidity Factor Liquidity Portfolios H H/M M M/L L Credit H 5% 12% 18% 23% 30% Portfolios M 5% 13% 21% 25% 32% L 4% 13% 22% 26% 34% 17
Empirical Results Contribution of Liquidity: 2 Incremental R 2 of Liquidity Factor Liquidity Portfolios H H/M M M/L L Duration L 4% 14% 21% 27% 36% Portfolios M 3% 16% 20% 28% 37% H 6% 17% 23% 30% 39% 18
Practical Implications Convertible Arbitrage Benchmark Regressions Alpha DEF TERM Rm-Rf SMB HML UMD Liq. Adj.R2 0.0029 -0.66 -0.33 0.27 0.3859 1.39 -1.43 -1.21 3.65 0.0011 -0.02 0.09 -0.19 0.07 0.08 -0.02 0.24 0.4897 0.59 -0.13 1.1 -2.45 2.45 1.28 -0.09 2.93 0.0012 -0.19 0.06 0.1 0.01 0.26 0.4565 0.67 -2.58 1.82 1.54 0.24 3.47 0.0004 -0.66 -0.33 0.055 0.58 -1.43 -1.21 0.0026 -0.02 0.08 -0.15 0.07 0.08 -0.03 0.1598 3.51 -0.15 1.08 -2.74 2.44 1.26 -0.09 0.0035 -0.17 0.06 0.09 0.01 0.1566 3.32 -2.07 1.8 1.51 0.25 19
Practical Implications Treasury Yield Curve Average Contribution of Factors to Bond Yields (RMSE) Maturity Curvature Term Liquidity 0.5 2 3 5 1 3 7 10 2 7 9 16 3 13 16 27 5 29 37 56 7 38 46 73 10 21 64 97 20
Practical Implications Back to WorldCom Worldcom 6.95 30Y Issuance Date: Aug-1998 Amount: $1.75 BB Callable Spread over benchmark Treasury Strip (%) 16 Forecast Spread Caa Actual Traded Spread Ba2 14 12 10 Baa2 8 6 4 2 0 Oct-00 Apr-01 Oct-01 Apr-02 Jul-00 Jan-01 Jul-01 Jan-02 21
Practical Implications Credit vs. Liquidity Spread 1/1/01 -1/1/02: Change in credit spread is minimal 22
Practical Implications Credit vs. Liquidity Spread Baa Index Ba Index 23 Source: State Street Global Markets
Practical Implications Liquidity-Driven Asset Allocation � Problem: � Allocate portfolio across a set of Moody’s Baa1 or higher rated long duration securities. � Set: BLS, CAT, BA, CCE, IBM, D,ALL, WFC, PFE, SBC � Scenarios � Scenario 1 (Optimizing on Total Risk) � Scenario 2 (Optimizing on Liquidity risk) � Scenario 3 (Optimizing on Credit risk) 24
Practical Implications Optimizing on Liquidity Risk Sub-Optimal Sharpe: 1.05 Sharpe 1: 1.69 Sharpe 2: 1.96 25 Source: State Street Global Markets
Practical Implications Optimizing on Credit Risk Sub-Optimal Sharpe: 0.19 Sharpe 1: 0.72 Sharpe 2: 0.84 26 Source: State Street Global Markets
Recommend
More recommend
