Horizon Kinetics December 2015: The Fixed Income Contrarian Compendium
H/T Chris Mayer
Horizon Kinetics – Murray’s Musings – The Unintended Consequences Of Bond Indices
Imagine a situation in which a nation has several central banks and each bank determines the appropriate interest rate in a given segment of the economy. One such segment might be non-investment-grade borrowers. Let us assume that, as the economy weakens, that segment’s bank elects to raise interest rates by several hundred basis points. Certainly most observers would agree that the interest rate increase would exacerbate an already precarious situation. Of course, no country would establish such a central bank. This scenario, however, is about to play out.
The two largest high yield bond ETFs in the U.S. are the iShares iBoxx High Yield Corporate Bond ETF (HYG) and the SPDR Barclays High Yield Bond ETF (JNK). HYG has about $14.4 billion in AUM; JNK has $9.7 billion. Year to date to December 11, 2015, HYG attracted $1.5 billion in new AUM; JNK has attracted roughly $1.1 billion in new AUM.
It might astonish some readers to learn that the 2015 year-to-date flows have occurred mainly in the fourth quarter. For the period September 30, 2015, to December 18, 2015, JNK attracted $1.26 billion and HYG attracted $2.2 billion, according to ETF.com.
Although high yield bonds as an asset class have been under pressure since the end of April 2015, the bond indices have been substantial net buyers of bonds. There has been no high yield bond index-related liquidity crisis because the indices have been buyers, as opposed to sellers, of bonds.
The outflow from high yield bond ETFs is a very recent phenomenon. In November 2015, JNK had $1 billion in redemptions while HYG still had a positive inflow of $172 million. In December 2015 (through December 18), however, JNK had outflow of $588 million and HYG had an outflow of $418 million. This is about $1 billion in 14 business days, or $71.4 million per business day. JNK owns 791 issues and HYG owns 1,015 issues, and there is obviously much overlap between them. Even if only 700 issues needed to be sold between these two funds on a daily basis, however, this comes to about $100,000 per day per issue. That is not a very large sum of money, and it is well to remember that the indices generally hold the most liquid bonds.
In theory, the bond market surely could accommodate $100,000 per day per issue. Unfortunately, these measures, as accurate as these unquestionably are, do not fully measure outflow. Consider that on December 11, 2015, 54,233,298 HYG shares traded. With a closing price for that day of $79.52, this amounted to $4.3 billion. JNK traded 34,999,231 shares on December 11, 2015. Using the closing price for that day of $33.69, this came to almost $1.2 billion of share trading volume. Although the ETFs trade as stocks, the potential buyers nevertheless are being asked to absorb $5.5 billion of high yield debt in a single day.
The system can function in this manner as long as the redemption requests are not actual outflow from the funds and are merely sales of the ETF. Therefore, it could be argued that, even in the event of illiquidity in the high yield bond market, there should be liquidity in the ETF market for high yield bonds in the form of ETF shares. The prices may be disappointing from the perspective of the seller but there would be liquidity.
Nevertheless, there would not necessarily be liquidity if the index composition is forced to alter by virtue of the rules of its own charter. For instance, it is certainly possible for a given bond to decline sufficiently in price so that its market value renders it too small for inclusion in a bond index. At that point, it might not be possible to sell the bond at a price that reflects genuine market value, since an index fund cannot scale out of a position. If the bond in question is no longer eligible for index inclusion, it must be sold at that point in time when exclusion has been determined. This would be known to the entire bond market trading community, and bond traders would be in a position to exert extraordinary market power.
Another exclusionary circumstance would be declaration of bankruptcy. High yield bond indices are not designed to hold securities in default. It might be a very good investment idea that the index holds bonds in default; however, defaulted bonds are not really part of the high yield asset class. In any case, defaulted bonds pay no interest, and a high yield fund is sold as an income-generating instrumentality. The only buyers of defaulted bonds are bankruptcy workout specialists. The bankruptcies not infrequently take years. The workout funds would be in an excellent trading position vis-à-vis the index, which would be a forced seller at a given point in time.
Consider, for example, the specific rules of the iBoxx U.S. Dollar High Yield Index as delineated in the Markit iBoxx U.S. Dollar Liquid High Yield Index Guide, a working paper of April 2012 (www.markit.com). According to these underlying fund rules, only fixed-rate bonds whose cash flow can be determined in advance are eligible for the index. Obviously, this specifically excludes bonds in default.
Issues rated D by Fitch, S&P, or Moody’s are specifically subject to index exclusion on the next index rebalancing date, according to Section 2.2, page 6, of those rules. All bond market participants would know this rule. Interestingly, an investment-grade bond that has been downgraded to junk status cannot enter the index for three months (Section 2.2.1). What happens when these bonds are excluded from the investment-grade indices but not yet included in the high yield indices?
According to Section 2.4.1 of the same working paper, the outstanding face amount of a bond must be greater or equal to $400 million to be eligible for index inclusion. Thus, if a company had sufficient cash flow to repurchase meaningful amounts of debt at a big discount to par value during a financial crisis, this bond would be excluded from the index. The resultant forced selling might easily help the issuer in building equity value on the balance sheet. However, the holder of the index would, by definition, not be able to participate in the likely appreciation of the debt of that issue. This security would be excluded.
According to Appendix 4, Page 10, of the working paper, an issuer must have at least $1 billion outstanding face amount of bonds to be eligible for inclusion. This requirement refers to all publicly traded debt of this issuer. Thus, in the next crisis, as some issuers improve the balance sheet by selling assets or new shares, and repay debt, these companies’ bonds might be excluded from the index as of the next rebalancing date subsequent to the action. Yet, issuers with increasing debt or deteriorating balance sheets would remain in the index until the default event.
If the high yield bond ETFs experience illiquidity events, one realistic scenario is not that the illiquidity is caused by the herd-like mentality of investors desirous of redemptions, although this certainly can happen; rather, one can readily envision that the rules of the index