Using the Wrong ETF for the Right Job

Leveraged ETFs (LETFs) are some of the most loved and hated securities – for quick tactical trades there’s nothing more convenient, but their multi-day price movements require unique attention. The limitations of equity LETFs are well known – they create a buy-high/sell-low position called “short gamma” (think of short gamma as fighting against the market tide, while a long gamma position benefits from the tide – in long gamma you’d be selling into gains and buying on dips). An LETF application underappreciated by researchers is the use of inverse LETFs over long-term treasury bonds (“going short” the 30yr U.S. Treasury or other positively-convex-assets) – contrary to popular belief, there are ranges of outcomes where the inverse LETF is beneficial to managing these positions.

LETFs

Most regular ETF users are familiar with the limitations of LETFs over multiple trading days.[i] In brief, inverse ETFs and LETFs perform a daily reset of their leverage (e.g. -1x, +2x) inclusive of each day’s gains or losses – they sacrifice a conventional uncompounded market position (or “delta”) for consistency of leverage for each day’s incremental investor. It is the changing of the “delta” based on each day’s outcome which typically limits the use of LETFs. Each day’s new buyers get [+2x] for the day, but all other holders have a stochastic “delta” position with path dependency.  

Many investors and traders use absolute and relative reference points for trade entry and exit. Absent corporate actions, using price reference points on AAPL, MSFT, or XOM can create discipline, and relative price levels can be useful to indicate new levels of richness or cheapness. I would caution traders not to use or confuse LETF price points as index indicators or index proxies – when using LETFs, use the basic uncompounded index levels and not the LETF prices as indicators for entry and exit.

Using LETF Daily Compounding Affirmatively

Fixed Income investing differs from most other markets – we focus on yields and spreads, but it’s the prices that matter. For U.S. Treasuries and non-callable bonds, prices and yields map one-for-one (i.e. if I know the yield, I know the price), but option-embedded sectors of the fixed income markets (i.e. mortgage-backed) require additional parameters and assumptions for price and yield to meet.

The following graph illustrates both a long position in a 30-year Treasury bond (curves upper left to lower right) and a short position in a 30-year Treasury bond (curves lower left to upper right); percentage returns are indicated on the y-axis and yields are indicated on the x-axis with percentage returns graphed from a baseline yield of a 3.0%.

Well known to bond investors, the price/yield relationship is curvilinear and a long position in U.S. Treasuries benefits from “positive convexity” (or gamma)[ii] – as yields decrease from 3% to 1%, gains accelerate for each incremental yield point decline, and as yields increase from 3% to 6%, the degree of loss decelerates with each incremental yield increase – a “win-win”. Turning to the short Treasury position (an increasingly popular position as investors hedge or attempt to time anticipated Fed policy), traders face adverse negative convexity – as yields rise, gains slow, but as yields decline, losses accelerate – a “lose-lose”; one of several costs in shorting Treasuries.  

The next graph looks at a leveraged short (or inverse) long-bond position held outright (“-2X”) and the same underlying bond held through a -2X LETF (“LEFT -2X”); percentage returns are on the Y-axis and change-in-yield is on X-axis. The outright position (“-2X”) is consistent with double (2 times) the previous graph, including the adverse returns curvature and a baseline yield of 3.0%.

In contrast to both an outright holding and conventional wisdom, the LETF has turned the “frown” (of being short a long-dated bond) upside down. As yields increase, the marginal gain improves – in short, the daily compounding of the LETF trumps the position’s negative convexity. Similarly, as yields decline, the marginal losses decelerate – again, the daily compounding (or in this case daily decompounding) trumps the underlying negative convexity. When LETFs are applied to positions with convexity, the compounding effects of the LETF can create a naturally beneficial position.

In the interest of brevity, this analysis does not cover the daily impact of volatility or high levels of daily return reversals (i.e. sequences of U, D, U, D markets), but this analysis is likely to reflect those opportunities where volatility is moderate and trends are strong.

While leveraged ETFs and inverse ETFs often get a bad rap, I believe they can fill an important role in fixed income investing for a range of investors – particularly as investors navigate the current inflection in global fixed income.

There’s a wide variety of other fixed income instruments with varying degrees of negative and positive convexity and some of these markets appear to be fertile ground for continued ETF development – both regular LETFs and inverse LETFs. Please feel free to contact me for any commercial or research interest in the topic.

###

[i] My September 2015 post entitled “What Block Chain and Ledger Technology Can Do For ETFs” provides some basic numerical examples relating to LETF multiday returns. I also recommend the Credit Suisse ETF Trade Strategy papers (authors Lin and Mackintosh) including “Triple Trouble” dated October 2011, and the paper “Leveraged ETFs – All You Wanted to Know but Were Afraid to Ask” – Avellaneda and Jian-Zhang, February 2010. While they require considerable attention, LETFs, and in particular inverse LETFs, can address the needs of sophisticated investors without the complications of other institutional alternatives.

[ii] The term “convexity” is typically used for bonds, and the term “gamma” is typically used in options, but they have the same practical meaning – “is the responsiveness to changes in the underlying (i.e. yields) changing as the underlying changes", and when return graphs are not straight lines, convexity/gamma comes into play – all else equal more positive convexity is good and more negative convexity is bad. Typically investors pay for positive convexity (in the form of reduced yields) and investors get paid for being short convexity, but ETF formats do not always follow these rules of nature.