Study: Do Bond ETF Creations & Redemptions Work?

November 10, 2008

 

For the study of discounts and premiums, Citi broke the research up by specific weeks during the market crisis. The first two weeks studied were September 15 through September 18 when AIG and Lehman dominated the news, and the week of October 6 through October 10, when the S&P fell 18%, when there was a treasury sell-off and muni market debacle.

For the entire taxable bond ETF group studied, the bid/ask spread differential and discount during these weeks was:

 

Date

B/A Midpoint (%)

Discount (%)

9/15

(1.00)

(1.27)

9/16

(0.71)

(1.09)

9/17

(1.76)

(2.12)

9/18

(2.26)

(2.55)

 

 

 

10/06

(1.60)

(1.89)

10/07

(1.18)

(1.43)

10/08

(2.64)

(2.92)

10/09

(1.86)

(2.17)

10/10

(6.46)

(6.93)

 

 

To make clear the extent to which exposure to corporates had a direct relationship with discount/premium level, Citi broke out four bond ETFs with varying degrees of corporates exposure: Vanguard Intermediate-Term Bond ETF (NYSE: BIV), which has a mix of intermediate government and corporate bonds; iShares iBoxx Investment Grade Corporate Bond Fund (NYSE: LQD), which only invests in high-quality, liquid corporate issues; and iShares iBoxx High Yield Corporate Bond Fund (NYSE: HYG) and the SPDR Lehman High Yield Bond ETF (NYSE: JNK), which invests in below-investment-grade issues. BIV trades closest to its indicative value, as it should.

 

Date

BIV B/A Midpoint (%)

BIV Price (%)

LQD B/A Midpoint (%)

LQD Price (%)

HYG B/A Midpoint (%)

HYG Price (%)

JNK B/A Midpoint (%)

JNK Price (%)

10/01

(0.72)

(0.92)

(0.58)

(0.76)

0.09

(0.62)

2.10

1.67

10/02

0.08

(0.05)

(2.30)

(2.43)

(0.84)

(1.31)

2.59

2.18

10/03

(0.25)

(0.56)

(1.14)

(1.28)

1.66

1.12

4.62

4.05

 

 

 

 

 

 

 

 

 

10/06

(1.44)

(1.60)

(5.56)

(5.73)

(2.06)

(2.63)

1.33

0.75

10/07

(0.56)

(0.76)

(1.15)

(1.26)

2.21

1.50

1.54

0.91

10/08

(0.99)

(1.13)

(1.07)

(1.21)

(0.09)

(0.39)

(1.85)

(2.39)

10/09

(2.12)

(2.29)

(1.18)

(1.45)

(0.01)

(0.43)

(1.20)

(1.50)

10/10

(5.49)

(5.56)

(5.69)

(6.26)

(9.62)

(10.15)

(7.79)

(8.43)

 

 

 

 

 

 

 

 

 

10/14

(0.23)

(0.38)

1.86

1.72

10.57

10.36

8.38

7.87

10/15

(0.99)

(1.12)

(1.08)

(1.23)

2.25

2.01

5.00

4.62

10/16

(1.14)

(1.33)

(0.89)

(1.08)

0.40

0.00

1.97

1.64

10/17

(1.02)

(1.12)

0.11

0.01

1.10

0.91

2.64

2.24

 

 

Citi's analysis suggests that it was thin markets, and many anomalies related to illiquidity, to which we can attribute the outsized discounts and premiums, as well as lack of quick arbitrage opportunities. Market makers' typical arbitrage plays that would have taken care of premiums as large as those seen on JNK and HYG (Oct. 14, for example) could be caused by a lack of liquidity in the market for market makers to short the ETF shares in a simultaneous transaction.

Citi writes: "Shorting such a quantity could move the market and thereby ... making the trade less profitable or even viable. ... In an in-kind redemption the market maker may have a difficult time purchasing the bonds, particularly in size without again moving the market."

Furthermore, the market makers did not trust the intraday NAVs as the basis for executing short trades, even in what normally would be a risk-free short trade with a discounted ETF. The net result is that trades that seem to be obvious may be impossible to execute when there is such short-term illiquidity in the markets.

Another notable finding was that HYG traded closer to its indicative value on many days than did LQD, the more liquid, higher-quality ETF. Citi analysts suggest that this may be due to creation basket size, with HYG holding 50 issues versus 100 bond issues in LQD.

Ultimately, market makers were either unwilling or unable to make tight markets in ETFs, Citi concludes. Furthermore, Citi suggests that the equity-based background of many of the arbitrage players in the ETF space may have left them ill-equipped to navigate the less-liquid and less-price-transparent bond markets. Citi analysts believe that the key to fixing these problems in future market upheavals could be changes to the creation and redemption process: fewer securities, more liquid securities, and a higher cash component in creation baskets.

In addition, more flexibility for cash as well as in-kind redemptions, and reducing minimum share size to create and redeem, could be beneficial changes. The minimum size now averages between 100,000-200,000 shares. While changing the basket components could make the risk of tracking error greater for the ETF, it could also allow market makers to tighten up markets when an ETF is drifting away from its indicative value.

 

Find your next ETF

Reset All