Bond indexes are what they are. They represent the average dollar invested in the bond markets. Those that say that the indexes are flawed miss the point. Indexes represent the average return of an asset class, with all of its warts and wrinkles. That is the nature of an index; it earns what the asset class as a whole earns.
So what if big issuers dominate the index? The average dollar in bonds reflects that. Do you want to take a bet against the average? You probably do, and I do as well. But it is not the purpose of an index to make that bet, so much as to facilitate that bet for active managers.
I appreciated the book The Fundamental Index – Arnott did us a favor by writing it. The book shows how to do enhanced indexing off of fundamental factors. (A pity that the book went public at the point where most of those factors were overpriced.)
The trouble with enhanced indexing is scalability. Suppose Arnott’s fund and those like it grew large relative to the market as a whole. The components of his strategy that are smallest relative to their total market size will get bid up disproportionately. Eventually they will not be a favored investment of the strategy, and as they move to sell, they will find that they are large holders of something the market is not so ready to buy. As the price goes down, perhaps it becomes attractive again. Perhaps an equilibrium will be reached.
One thing is certain, though. The non-enhanced index can be held be everyone. The enhanced index will run into size limits.
What then for bond ETFs? Are they chained to inferior indexes? No. By their nature, bond indexes are almost impossible to replicate perfectly because of liquidity constraints. Many institutional bond investors buy and hold, particularly for unique issues. That’s why indexes are constructed out of liquid issues which will have adequate tradability. Who issues those bonds? The big issuers. It is not possible to create a scalable bond index in any other way, and even then, there will always be some bonds in the index that are impossible to find, and/or, because they are index bonds, they trade artificially rich to similar bonds that are not in the index.
Almost all bond indexers are enhanced indexers, because they don’t have enough liquidity to exactly replicate the index. Instead, bond indexers try to replicate the factors that drive the index, with better performance if they can manage it. That’s where choosing non-index bonds that are similar in characteristics, but have better yields comes in. That is the value of active bond management; it does not mean that the indexes are flawed, but that there are ways for clever investors to systematically do better, that is, until there are too many clever investors.
Morningstar prepared this piece on pricing difficulties with bond ETFs and open-ended bond funds. Yes, it is true that many bonds don’t trade regularly, and that matrix pricing gives estimates for prices on bonds that have not traded near the close, where an asset value must be calculated.
Remember the scandal over mutual fund front-running? In that case, stale pricing off of last trades enabled clever connected “investors” to place late trades where the calculated NAV was far away from the theoretically correct NAV (if assets traded continuously). In order to calculate the theoretically correct NAV (which the late traders did in order to make money), the mutual funds had to engage in a form of matrix pricing, adjusting the last trades to reflect changes in the market since each last trade until the close. Far from being inaccurate, matrix pricing is far superior to using the last trade.
I will take the opposite side of the trade from the Morningstar piece. Markets are not rational, especially bond ETF investors. I trust the NAV more than the current price; matrix pricing is complex, but it is pretty accurate. Yes, for some really illiquid, unique issues, it will get prices wrong, but that is a tiny fraction of the bond universe. We can ignore that.
Rationality comes to bond ETFs when sophisticated investors do the arbitrage, and create new ETF units when there is a premium to the NAV, or melt ETF units into their constituent parts when there is a discount to NAV. That pressure places bounds on how large premiums and discounts can become.
The more specific the bonds must be to create a new unit, the harder it is to do the arbitrage, and the higher the level of premium can become before an arbitrage can occur. If a less specific group of bonds can be delivered to create a new unit, i.e., the bonds must satisfy certain constraints on issuer percentages, issue sizes, duration [interest rate sensitivity], convexity [sensitivity to interest rate sensitivity], sector percentages, option-adjusted spread/yield, etc., then arbitrage can proceed more rapidly, and premiums over NAV should be smaller.
So, when there are large premiums to NAV, it is better to sell. Large discounts, better to buy. Of course, take into account that short bond funds should never get large premiums or discounts. If they do, something weird is going on. Long bond funds can get larger premiums and discounts because their prices vary more. It takes a wider price gap versus NAV before arbitrage can occur.
As for cash creations, those that run the ETF could publish a shadow ETF price, which would represent the price that they could create new units themselves, taking into account how they would like to change the ETF’s positions in order to better outperform while matching the underlying characteristics of the index. That shadow ETF price could not be a fixed percentage of the existing NAV. It would have to vary based on the cost of sourcing the needed bonds. This would run in reverse for cash-based redemptions, which would only likely be asked for when the ETF was at a discount. Better for the fund to do some modified “in-kind” distribution, agreed to in advance by the sophisticated unit liquidator.
Well, if there’s not enough liquidity in the bond market to accommodate our desired investment, why not create it synthetically through credit default swaps? That might work, but if the bonds are illiquid, often the derivatives are as well, or, the derivatives trade rich to where an identical bond would trade in the cash market. There is also credit risk from the party buying protection on the default swap; if he goes broke, your extra yield goes away, at least in part.
I don’t see derivatives as being a solution here, though they might be helpful in the short-run while waiting to source a bond that can’t be found. Derivatives aren’t magic; liquidity comes at a cost, and some of those costs aren’t obvious until a market event hits.
Also, I would argue that the rating agencies are better judges of creditworthiness on average than the prices of credit default swaps. Though rating agencies should be examined for their conduct in structured securities, their record with corporates is pretty good. The rating agencies do fundamental research; yields do reflect riskiness, but markets sometimes wander away from their fundamental moorings. Derivatives can trade rich or cheap to the cash market for their own unique reasons. Same for bond spreads – just because one bond has a higher spread than another similar bond, it does not mean that that bond is necessarily more risky.
When I was a corporate bond manager, I would occasionally find bonds that yielded considerably more than others of a given class. My job, and the job of my analyst was to find out why. Often the bond was not well known, or was a better quality name in a bad industry. On average, spreads reflect riskiness, but in individual situations, I would rather trust the judgments of fundamental analysts, including the rating agencies, though private analysts are better still.
So what should I do in the Current Environment?
I don’t think we are being paid to take credit risk at present, so stay conservative in bonds for now. Specifically:
- Underweight credit risk.
- With equities, stress high-quality balance sheets, and stable industries.
- Underweight financials, particularly banks and names that are related to commercial real estate.
- GSE-related residential mortgages look okay.
- TIPS don’t look good on the short end, but look okay on the long end.
- Be wary of paying premiums on bond ETFs… and maybe look at some closed-end funds that trade at discounts.
- The yield curve is steep, but that is ahead of a lot of long supply coming from the US Treasury. Stick to short-to-intermediate debt, and wait for supply to be digested. After that, maybe some long maturity positions can be taken as rentals, so long as inflation does not take off.
- Diversify into foreign bonds, but don’t go crazy here. The Dollar has run down hard, and opportunities are fewer. (I will have a deeper piece on this in time, I hope.)
This is a time to preserve capital, not reach for gains. Don’t grasp for yields that cannot be maintained.
PS — Thanks to the guys at Index Universe and Morningstar for the articles; they stimulated my thinking. I like both sites a lot, and recommend them to my readers. The articles that I cited had many good things in them, I just wanted to take issue with some of their points.