Trading Treasury Rates with ETFs - Part 2

In Trading Treasury Rates Part 1 I established a good reason for devising ways to trade U.S. Treasury interest rates without using the futures markets. In that article I offered two candidate ETFs, the iShares Barclays 20+ Year Treasury Bond Fund (TLT) and the Proshares Ultrashort 20+ Year Treasury Fund (TBT).

TLT is a long bond fund - it's value rises when the market value of long-term U.S. Treasury securities (USTS) rise (which happens when their market yields fall).

TBT in contrast is a short fund which does the opposite - it's value rises when the market value of long-term USTS falls. Further, unlike TLT, TBT is a leveraged fund and it is also a delta fund. A delta fund is designed to track the change in the value of an underlying security (rather than track the long-run value of the fund over time) day by day, percent by percent. If the fund is 2X short-leveraged like TBT, if on any given day TLT were to rise by X%, then TBT should fall by about 2X%.

The benchmark for both funds is the Barclays Capital 20+ Year U.S. Treasury Index, which is discussed in the next part of this series. TLT tries to follow the index exactly. TBT attempts to track two times (2X) the inverse of the daily change in this index each day.

Criteria for Choosing the ETFs

Now let me justify these two choices.

Of the many ETF candidates that we might choose, the following criteria should be satisfied

1. The ETF must be liquid, which implies that it should have high daily volume, especially when compared to peers.
2. If the strategy requires options, as our does, then a full range of put and call options must be available for the ETF, and the options that are near the money in near-term months must have both adequate open interest and daily volume.
3. The ETF should track well - meaning that if it promises to track an underlying index or other stated value, it should have a track record of doing that with a high degree of accuracy (the criteria would vary from one context to another, but in this context for a straight tracking ETF like TLT it should be above 95% and for a delta ETF it should be above 90%).

There are more than a dozen ETFs that track U.S. Treasury securities of various maturities. Candidates can be found in online searches at dedicated sites or in the Online Wall Street Journal Market Data section.

Daily Volume

Figure 2.1 U.S. Treasury Security Exchange Traded Funds (ETFs) shows a selection of mid- to long-term funds for both longs and shorts. A quick glance shows that only TLT and TBT have the necessary liquidity, as measured by average daily volume, to qualify for the first criterion above. The daily volume eclipses that of the other ETFs. The Proshares TBF may seem like a better candidate than TBT for short trades because it is not leveraged, but at 238,000 shares a day and no listed options it must be ruled out (N means no options, I means illiquid - inadequate open interest and/or daily volume).

Adequate Options

In Part 1 when describing general strategy I stated that if we were to use options for our strategy, we would use only puts and calls on TLT. If not using options then TBT provides a direct short opportunity but if options are required then puts on TLT are more suitable than calls on TBT.

Although the argument won't be supported with data here, because delta ETFs don't track with anywhere near the integrity of a straight tracking ETF like TLT, options on delta ETFs compound the error in tracking integrity. In a word, they are too volatile.

Even the TLT options barely qualify for serious trading. They tend to be liquid with adequate open interest and daily volume only right around the money. Also there is almost always a minimum spread between best bid and best ask of five cents and it is often higher. This is a large negative, because on many of the equity index ETFs this spread is seldom more than a penny.

Figure 2.1 Near-the-Money TLT Options shows the daily volume and open interest on TLT for May 20, 2010. As can be seen both volume and open interest is adequate for small trades (say 30 contracts or so) but not really for large trades. Although not shown here, it should be noted that volume and open interest on TBT options are at least as high as TLT and in volatile markets sometimes higher.

The Tracking Record

Traditional ETFs that track a known index over the long-run, like TLT, normally have a good tracking record (at above 95%) if they are popular and liquid.

Delta ETFs, which attempt to track only the change in the target day-by-day, sometimes with leverage like 2X for TBT, don't have such a good track record. If the Delta ETF has any sizeable error on the day-by-day target, that error can accumulate over time to where the long-term performance of the target has little correlation to the long-term performance of the ETF that tracks it.

This is partly because of how these ETFs are collateralized, which is discussed in the next part.

The criteria for measuring how well an ETF tracks its target, in this case the Barclay's Capital 20+ Year U.S. Treasury Index, compares the daily Net Asset Value (NAV) of the fund to that of the underlying index. ETF fact sheets and prospectuses can always be found on their parent web sites with a simple search and the fund performance relative to the target they track can be found in either or both.[1] If untrusting, a researcher with some spare time can use regression analysis to see how they track.

TLT tracks very closely, with less than 2% error because (a) iShares collateralizes TLT by simply buying the same securities as are used in the index in the same proportions, and (b) the stock is popular and liquid enough for the daily trading price to actually track the true Net Asset Value (NAV) of the holdings.

It turns out that even though TBT is a delta ETF and is collateralized with derivatives called Treasury Swaps, because it is also liquid, in the first quarter of 2010 only on a few occasions did it track with more than a 3% error. For short term trades this is completely adequate, although because this is a delta ETF, its usefulness as a long-term short hedge (or hedge against rising interest rates) is still questionable.

NEXT: Collateralization and pricing conventions for these two ETFs.

---

[1] In this case see the ProShares UltraShort 20+ Year Treasury Fact Sheet and Barclays 20+ Year Treasury Bond Fund (TLT) and once on that page, link from Related Links and Documents to the "Tracking Error Chart."