The first leveraged ETF was launched in 2006, 13 years after the first ETF was rolled out back in 1993. Although there has been much debate over whether leveraged ETFs add to market volatility and risk, the market size of leveraged ETFs has topped $100 billion since their launch. Evidently, investors’ appetite for leveraged ETFs has been growing rapidly in the past decade. ETF providers upped the initial 2X leverage to 3X to satiate the market demand. But that growth coincided with the longest bull market in history. Now that we have hit a new bear market, should we still consider leveraged products? The purpose of this article to provide some information to help investors decide on questions like this.
Leveraged ETFs are fundamentally different from traditional leverage. First and foremost, investors in these ETFs are not subject to margin calls (unless the ETF is purchased on margin), a common counterparty risk enforced by a broker. (Leveraged ETFs are still subject to third-party risk because of the involvement of derivatives.) ETF providers use derivatives such as swap agreements and futures contracts to achieve double or triple leverage, and they rebalance daily so that targeted movement matches the tracking index. In place of borrowing cost in traditional leverage, leveraged ETFs charge ETF fees only. These fees are charged daily and are reflected in the end-of-day net asset value (NAV).
Secondly, leveraged ETFs rebalance daily. For this reason, the return for a period longer than a day will likely differ from each single day’s return compounded over the period. In addition, although some leveraged ETFs distribute dividends, they do not necessarily track the index dividend rate. This widens the long-term return discrepancy. Academic literature supporting this fact argues that the return of a k-times leveraged ETF over investment horizon [0,t] largely depends on the holding period volatility (see Jarrow 2010**). Theoretically, investors may gain/lose less than the k-times leveraged performance of the index over time. We will use the triple-leveraged ETF UPRO as an example below. (See http://www.proshares.com.) 
The darker-shaded areas represent those scenarios where returns are expected to be less than 3X the tracking index performance. Please note, that theoretically, one may lose all principal value if the index falls drastically but will not suffer from negative value unless the ETF was purchased on margin. The second column, “Three Times the One Year Index,” is just for 3X demonstration purposes. The scenario where the ETF loses its entire value is also highly unlikely, because it requires the market to fall either 33.33% on one day or continuously with high volatility. For example, Chart 1 suggests that if the index falls 30% with volatility at 50%, then the loss is 83.8%, less than the indicated 3X 90%. In any case, the purpose of this chart is to show the importance of volatility in the compounding of returns – the chart is not a representation of actual returns. Furthermore, the index’s annualized historical volatility rate for the five-year period ended on May 31, 2019, was 13.36, with the highest May-to-May volatility being 16.65%. Although the historical performance is not any indication of the future, we can reasonably shift expectation to the 10% and 25% columns in Chart 1.
 In order to demonstrate the impact of volatility on actual returns, we construct a theoretical 3X index and compare it with the first-launched 3X S&P 500 ETF, SPXL. Chart 2 above shows the cumulative return of the ETF since its inception in 2008 and the theoretical 3X index return on a daily compounding basis. There are some key takeaways. First, the fact that the actual ETF performance outpaced the theoretical index confirms the prior argument that volatility is an important factor. However, even though the volatility on average has been low in the past decade, the 230.52% difference between the ETF and the theoretical Index is still a significant number. In addition, the discrepancy cannot be explained by tracking error, which is about 0.10%. Second, neither the ETF nor the theoretical index lost its entire principal at the bottom of the bear market. The biggest losses occurred contemporarily on March 9, 2009. The ETF was down 71.32% while the index was down 72.05% at the low; and the ETF recouped its loss on September 15, 2009, a month ahead of the index. Moreover, the gap between the two lines closes each time volatility spikes (e.g. in 2018). But as volatility subsequently falls, the ensuing recovery in the ETF outpaces that of the index.
We have so far pointed out the main risks in leveraged ETFs: equity market risk, volatility risk, counterparty risk, compounding risk, and tracking error risk. There are some other risks inherited in these ETFs as well. For example, one may face intraday price risk, as these ETFs are often traded for daily purposes. But since the NAV is calculated at the end of the day, the intraday price performance may deviate from the stated multiple of the index. Another example is liquidity risk. Although the leveraged ETF industry has grown quickly, some ETF markets may be thin during volatile times, which exacerbates the price movement. Systematic risk applies to equity ETFs in general. We believe volatility risk and compounding risk are the deciding factors for investors to consider before they purchase these ETFs. Last but not least, given that leveraged ETFs are usually short-term trading vehicles, investors are subject to short-term capital gains taxation. *Data from Bloomberg.
**Please see Jarrow 2010 for k-times leveraged ETF evolution. Using Ito’s formula in continuous compounding, Jarrow shows that the evolution for the k-times leveraged ETF over the investment horizon [0,t] is: Note: Cumberland uses UPRO in its leveraged ETF strategy and may have a position in UPRO at the time of this publication. | 
No comments:
Post a Comment