By Eric Dutram, Assistant Vice President, DWS.

Eric Dutram, Assistant Vice President, DWS.

ETFs are often hailed as a major financial innovation. Seen by some as next-generation mutual funds, these securities are generally considered to be more tax efficient than their fund cousins while providing investors with intraday liquidity as well.

The key to this process is arguably the creation and redemption mechanism. This allows ETF issuers to create new shares when demand outstrips supply or redeem ETF shares when supply overtakes demand.

This system is generally between what is known as an Authorized Participant and the ETF issuer. When shares of an ETF are created, the Authorized Participant “AP” will deliver shares in-kind of the underlying constituents in exchange for units of the ETF. For redemptions, the opposite holds true; the AP would deliver the issuer the ETF in-kind for shares of the component securities.

While this may seem like a technical nuance, it is actually the heart of the ETF process. Due to the creation/redemption mechanism, APs can arbitrage away any difference between the components and the ETF’s price thereby helping to keep the trading price of an ETF in line with the value of its underlying securities.

In practice

This is easy to visualize for a well-known benchmark such as the S&P 500. An AP following this index would simply buy up all the component securities and deliver this as a basket to an ETF issuer. Such a process isn’t too difficult because all of the S&P 500 stocks are extremely liquid and there isn’t an overwhelming number of securities to purchase in order to create the basket.

While that is pretty straightforward, what about for the world of fixed income? Bonds do not trade like stocks do, while the Bloomberg USD High-Yield Corporate Bond Index – a major benchmark of sub-investment grade-rated debt – has over 2,000 securities.

And this doesn’t even begin to touch upon the issue of different maturity dates, again, something that is irrelevant for stocks. Clearly, the world of fixed income can be more complicated when it comes to building and maintaining an ETF. So how do fund managers solve these issues?

Replication vs. sampling

With a benchmark such as the S&P 500, a fund manager would use what is known as ‘full replication’. This means that the manager would buy each and every security that makes up a given benchmark. However, for a massive benchmark such as the high yield one described above – but also hard-to-trade equity markets and other illiquid investments – a ‘sampling’ process is used instead.

This means that the fund managers do not need to hold all of the securities in an index but instead obtain a representative sample. Sometimes, this can be but a small fraction of the original index amount, though the goal is to deliver similar returns and risks as the original. For the high yield bond world, our process looks like this:

Source: DWS.

We start off with the overall universe and then dial it down to the most liquid securities in the benchmark. From there, we create an optimized ETF which is a still smaller number of bonds which can replicate the overall performance characteristics of the index at large.

It is important to note that the final step consists of building a good creation basket so that APs can easily participate in the creation and redemption process. In other words, APs may only need to provide a subset of the ETF’s holdings to obtain new shares or redeem old ones, though it will be representative of the whole and generally more easily traded as well. The idea behind this is to promote overall liquidity and to make an AP’s job of gathering and dispersing securities all that much easier, so long as it remains representative of the overall ETF.

While this might seem like a complicated process, all of this is done to balance two competing and easy to understand factors that play into the construction process: trading costs and tracking error. The closer a portfolio gets to full replication, the greater the cost. However, the flipside is generally true: the lower the cost, the greater the tracking error.

Source: DWS.

The job of the capital markets team is to find the ideal point that balances both of these important items without sacrificing too much in terms of tracking or costs. This is why many fixed income ETFs utilize a sampling methodology. Without it, full replication would result in far higher costs while just buying a tiny subset would likely give too high a tracking error. The equilibrium is needed to find the optimal basket for a given investing situation.

Key takeaway

When it comes to ETFs, the creation and redemption mechanism is a key aspect of the process. However, this is only part of the story when it comes to fixed income investing thanks to the massive size of the bond market and its opaqueness.

Here, having a vigilant capital markets group on your side is arguably very important, as you need a team that is going to keep watch for any changes in the market and will stay committed to keeping both tracking error and trading costs low. This is perhaps the only way to build a lean portfolio that can give investors the type of exposure they are seeking in the world of fixed income while also taking advantage of the ETF structure.

(The views expressed here are those of the author and do not necessarily reflect those of ETF Strategy.)