Matthew Patterson is the Head of Investment Strategy at Accretive Asset Management LLC, a firm dedicated to developing products that help financial advisors better serve their clients. Accretive developed the indices underlying the BulletShares suite of target maturity date fixed income products offered by Guggenheim. Matt recently sat down with ETF Database to talk about some of the limitations of traditional fixed income ETFs, as well as some of the innovation in the space:
ETF Database (ETFdb): Explain the difference in “yield experience” when investing in bond ETFs as compared to investing in individual bonds.
Matthew Patterson (MP):: When you buy an individual bond, you have a pretty good idea of what your yield on the investment is going to be. You still have reinvestment risk and default risk of course, but the stream of cash flows you are expected to receive from an individual bond is known in advance, which allows you to calculate an expected yield to maturity at the time of purchase. Even if interest rates dramatically rise and reduce the fair market value of the bond, an investor always has the option of holding the bond to maturity and receiving the benefit of what he originally bargained for.
In contrast, investors have no way of predicting in advance what their expected yield to maturity will be when they purchase a traditional bond mutual fund/ETF. The most obvious reason for this is that traditional bond mutual fund/ETFs have no maturity date. They are designed to operate in perpetuity, so the managers of these funds are constantly buying and selling bonds in order to maintain a constant duration and keep the fund fully invested. This has the perverse effect of maintaining a relatively high sensitivity to interest rate changes even as an investors’ need to liquidate an investment draws nearer. Consequently, investors can actually experience a dramatic decline or increase on the yield of the investment if interest rate moves cause significant changes to the NAV at which investors can liquidate their investment.
ETFdb: Let’s talk about tracking error. We know that a lot of fixed income ETFs have shown considerable tracking error relative to the underlying index that they seek to replicate. What are some of the causes of the tracking error in fixed income ETFs?
MP: That is an interesting question because if you read a typical ETF prospectus, you might think that index sampling is the biggest source or tracking error. When I say index sampling, I am referring to the process of a fund investing in only a representative sample of securities that collectively are supposed to have an investment profile similar to the fund’s underlying index. But tracking error due to index sampling should, in theory, have a normal distribution with an average expected return of zero. In other words, the representative sample might outperform or underperform the index over any given period, but there is no reason to expect it to outperform or underperform the index on a regular basis, assuming that it is a truly representative sample.
I suspect that the tracking error that most people are concerned about is negative tracking error. I rarely hear people complain that their investments outperform their expectations. And one of the biggest causes of tracking error in fixed income ETFs is expenses. When I refer to expenses, most people assume I am talking about a fund’s stated expense ratio. While it is true that you can generally expect an index fund to underperform its index by at least as much as the expense ratio, it is important for investors to understand that the costs included in the expense ratio are not the only costs incurred by funds. Most significantly, funds incur transaction costs when they trade securities.
What we have found shocking is the extremely high level of turnover reported by the vast majority of fixed income ETFs. Transaction costs are driven primarily by how much trading an ETF engages in to track the characteristics of the underlying indices, and of course these trading costs will be higher with more illiquid securities, and particularly so with off-the-run corporate bonds that trade infrequently.
These costs are important because they decrease overall performance, but are not included in the expense ratio calculation of the fund, nor are they specifically disclosed in the fund shareholder reports. While they always result in underperformance relative to the index for the more illiquid asset classes, they are effectively invisible to investors.
There is another source of significant tracking error in fixed income ETFs and funds in general, and it is even less visible to fund investors than transaction costs. Since indices tracked by fixed income funds are generally rules-based transparent indices, anybody with access to market data can predict in advance what portfolio changes fixed income index managers will be making.
This creates a profitable opportunity for institutional investors to front run fixed income funds by selling short or providing extremely low bids on securities they know the index managers will have to sell, and buying securities and then asking for additional premium for securities that they know index managers will have to buy. And this front running creates tracking error for fixed income ETFs by temporarily driving down the prices of securities they must sell, and driving up the prices of securities they must buy.
Because pricing information in the fixed income market can be difficult to come by, the extent of this negative tracking error caused by front running is really difficult to quantify in the bond market. But I will say that a recent study in Journal of Empirical Finance estimates that such front running cost Russell 2000 index funds 38 basis points of performance on an annual basis. Based on the empirical work we have done, we suspect losses caused by index front running are even higher for fixed income funds.
ETFdb: So for example if there were a 1-5 year bond ETF that focused on securities with a time to maturity within one to five years, it would be relatively straightforward to know that the securities are about to go under a year to maturity are going to be sold, and those with just over five years are about to be bought as they come into that window?
MP: Precisely, and we have actually done work on looking at several indices that are out there and looking at what happens to the prices of deleted bonds in the month they are deleted from the index. You do actually see underperformance of those bonds relative to other bonds in the portfolio, but what is interesting is that the prices of those bonds subsequently rebound after the rebalancing of the index funds is done. So in the long run the index return really doesn’t reflect the temporary blip to the value of these funds, but it does affect the long run returns of the index funds that are forced to sell those bonds when the bonds are temporarily undervalued.
ETFdb: You have done some research on the minimum maturity rule. Explain what that is and how it will affect the risk return profile of bond ETFs.
MP: Traditionally most indices have employed a minimum maturity rule, which excludes bonds of a certain minimum maturity from a certain index. For example, the most popular bond index in the world, the Barclays Capital Aggregate Bond Index, employs a one year minimum maturity rule. That means that bonds are removed from the index when they reach one year to maturity. Minimum maturity rules make some sense when you consider that the Barclays Capital Aggregate Bond Index was originally intended to serve as a performance benchmark for bond managers rather than as the basis for investment products.
Credit risk and yield, and therefore the opportunities for bond managers to differentiate their performance, decline significantly as a bond approaches maturity. Excluding bonds of a certain minimum maturity from certain fixed income indices can have the effect of focusing in on the portion of the bond market where outperformance is most likely to be generated by talented bond managers.
We think the minimum maturity rule makes much less sense for bond indices intended to serve as the basis of investment products. ETFs seeking to track bond indices with minimum maturity rules are effectively forced to sell all of the bonds in their portfolios before they mature, incurring significant transaction costs that decrease the performance of the portfolio in the process. This makes no sense when you consider one of the primary benefits of a bond is that it returns your capital investment to you at maturity without any transaction costs.
By forcing ETFs that track them to sell bonds before maturity, fixed income indices that employ minimum maturity rules generate needless portfolio turnover, reducing returns because of transaction costs that could easily be avoided by simply holding bonds to maturity. It is an outcome that illustrates the power of path dependence. Years after fixed income indices became more valuable as investment strategies then as performance benchmarks, they continue to be constructed to serve the objectives of performance benchmarks rather than investment strategies.
ETFdb: Let’s talk about the BulletShares indices that underline some fixed income products from Guggenheim. Explain to us how these products work and how they are different from a lot of the traditional fixed income products out there.
MP: Based on the conversations we have had with financial advisors, our objective was to create an index methodology that can serve as the basis of investment products that combine the best attributes of individual bonds and bonds funds. So this meant developing an index methodology that would permit the creation of diversified investment portfolios while providing a stream of cash flows followed by the return of principal on maturity dates.
To accomplish this we developed the concept of the maturity targeted bond index, as opposed to the traditional approach to bond indexing which I would call a constant maturity bond index. By creating a separate index for each year of maturity–and then keeping every bond in the index until maturity, unless it otherwise fails to meet the index criteria–BulletShares indices allow for the creation of diversified bond portfolios that have cash flow profiles similar to those of an individual bond. Not only do investment products based on BulletShares indices help financial advisors manage their clients’ future cash flow needs by providing a known maturity date, they also allow financial advisors to dynamically hedge their future liability.
ETFdb: So why might that be appealing to a liability-driven investor—whether it be a pension fund that will owe billions of dollars or a family preparing to send a kid to college?
MP: Suppose an investor puts his money into what I would call a constant maturity intermediate term bond ETF–let’s say the investor needs his money in five years and the weighted average maturity of the bond ETF’s portfolio is five years. He is going to have a relatively decent match between the duration of his investment and his future liability at the time he makes his investment. But the problem is, as the liability draws nearer, the bond ETF continually rolls over shorter term bonds into longer term bonds in order to maintain a relatively constant duration. By the time the liability is only one year away the investor is exposed to an investment with an intermediate term duration, which means he is at risk of experiencing a significant capital loss if interest rates rise before he liquidates his investment.
In contrast, with an investment product based on the BulletShares methodology, duration, and therefore interest rate risk, declines as the liabilities draw nearer. It’s the retail equivalent of liability driven investing. That said, we don’t think you need to be a liability driven investor to take advantage of the benefits of investment products based on BulletShares indices. Those seeking to maintain a constant maturity exposure can simply establish a laddered portfolio that re-invests as holdings mature. And we believe that this methodology, when used as a part of a constant maturity ladder strategy, would result in better outcomes than the traditional high turnover, high tracking error fixed income ETFs.
ETFdb: Yet BulletShares products still bring the immediate diversification benefits that investors expect from ETFs, correct?
MP: Exactly. And one of the problems you will run into with individual bonds–in addition to the diversification challenges–is that the secondary market for those bonds tends to be relatively illiquid. If you have an exchange traded bond ETF, it actually trades on a screen-traded market where you can actually see the bid and the ask, you can see the depth on both sides of the market, and you can make intelligent trading decisions, and generally get better execution than you are going to get in the over-the-counter bond market. And in the over-the-counter bond market, which is a market largely transacted over the telephone, your ability to get decent execution on individual bonds is largely determined by the quantity and quality of relationships you have with bond dealers.
Based on conversations I’ve had with financial advisors, they are not pleased with the quality of execution they are getting on individual bonds.
Read more about the new high yield BulletShares products here.
[For more ETF insights, sign up for our free ETF newsletter]