Leveraged ETFs: The Data Says We're Missing Out
This post originally appeared on LinkedIn here.
Leveraged ETFs have been given a bad rap. With the right products and proper management, retail and professional investors alike can improve their returns.
Why it matters: to protect retail investors against volatility decay, the SEC, brokers, and ETF providers have warned against holding these products for any significant period. These warnings, while important, have failed to foster a productive conversation about the relative costs and benefits of these products.
- Long-term investors can expect to outperform broad market indices when using leveraged ETFs so long as the overall leverage level of their portfolio is modest and they rebalance daily
- Hedge funds and trading firms can exploit specific nuances of these products that may not be correctly priced by the market
The Intuition
Popular ETFs (exchange-traded funds) often attempt to track market indices. For example, the SPY ETF seeks to track the S&P 500.
Leveraged ETFs attempt to produce returns that are a multiple of the indices they seek to track. For example, the SSO ETF seeks to generate returns that are 2x of the S&P 500 on a daily basis.
Investors who believe that the S&P 500 will increase over the long run may be tempted to purchase SSO rather than SPY. They understand they will face a high degree of short-term risk. But, they reason, since they believe the S&P 500 will produce a positive return in the long run, they will be able to enjoy 2x the long-run return of the S&P 500.
This is false.
Two years ago, the SEC issued an investor bulletin stating:
Most leveraged ETFs “reset” daily, meaning that they are designed to achieve their investment objective on a daily basis. Their performance over longer periods of time may differ significantly from the performance of the underlying index or benchmark during the same period of time.
The intuition here can be a little tricky, so let's look at an extreme example.
Pretend a non-leveraged ETF goes down by 10%. The next day, it returns to its original level thanks to an 11.1% return. This means that a 3X leveraged ETF would decline by 30% initially. The next day, it would increase by 33.3%. However, since this 33.3% return is on a substantially lower price than that of our non-leveraged ETF, the non-leveraged ETF underperforms. This concept demonstrated here is called "volatility decay."
Another Example
Over a longer period, swings in a hypothetical non-leveraged ETF (or underlying index), translate to a substantial decline in the price of the leveraged ETF!
In general, several factors will increase the degree to which value is destroyed by leveraged ETFs:
- Greater degree of leverage employed
- Higher volatility of the underlying index
- Longer investment time horizon
The theory presented so far is why the SEC, brokers, and ETF providers warn of the potential danger of these products.
What's Missing from This Story
It is true that leveraged ETFs can destroy value and suppress returns.
But the longer-run performance of leveraged ETFs is more nuanced.
In fact, the longer-run performance might look more like this:
In the chart above, the gray line represents the return of a hypothetical ETF without any leverage. The yellow line represents 3X the return of the ETF without any leverage. In other words, the yellow line represents what we might expect a 3x leveraged ETF to return over a 3-year period if we did not concern ourselves with the nuances of daily resetting. The green line represents the returns of a 3X leveraged ETF that resets daily.
If correct, this chart shows there are times when leverage ETFs can perform worse than expectations (i.e. when the green line is below the yellow line). There are also times when they can perform better than their stated level of leverage would imply.
For example, when the ETF without leverage increases by 75%, the chart shows that the ETF with 3X leverage increases by about 300%. Our intuition might suggest that a 3X leveraged ETF should perform 3X the non-leveraged ETF (75% x 3 = 225%). However, it actually does better, despite the negative effects of volatility decay.
Additionally, when the ETF without leverage decreases by 50%, the leveraged ETF does not decrease by 150% [1]. Daily resetting can actually improve returns and limit risk under the right circumstances.
This is not a complete endorsement of replacing non-leveraged ETFs in your portfolio with leveraged ETFs. After all, this chart also shows double-digit losses when the non-leveraged ETF has a return of 0% over 3 years. Investors with shorter-term goals or a belief that a given index will not increase substantially over the investor's investment time horizon should avoid these products.
This chart is the product of simulating several hundred return scenarios for both a hypothetical non-leveraged ETF and a corresponding leveraged ETF, given a specified average return and volatility level. From there, I developed a model to predict the leveraged ETF return given the return of the non-leveraged ETF.
While this technique shows a far more nuanced view of expected returns, our key factors remain the same. Higher values of the factors below will cause the green and yellow lines to diverge:
- Degree of leverage employed
- Volatility of the underlying index
- Time horizon
For those familiar with options, the green line may appear similar to the payoff diagram for going long on a call option on the non-leveraged ETF. Hedge funds and traders may want to consider if there are opportunities to use leveraged ETFs as part of their options strategies.
Evidence
Does the model developed in the previous section actually describe how these products perform in the real world?
Largely yes.
Let's look at SSO (2X the S&P 500 on a daily basis) as an example. Examining it from its inception in June 2006 through late December 2022, we see a cumulative return of 404%. This is far higher than its non-leveraged counterpart, SPY, which returned 209%. [1]
Using the model discussed in the previous section, we predict a cumulative return of 392% over this 16-year period for SSO versus an observed cumulative return of 404%. Not a bad prediction!
The model performed similarly when predicting the long-term performance of another leveraged ETF: TQQQ (3X the daily performance of the NASDAQ).
How would investors who go long on these products likely fare over the long run?
Using a leverage level of 3X with the S&P 500 from 1950 to 2009 would have provided a compound annual growth rate (CAGR) of about 14%, according to a quant researcher at Double Digit Numerics [2]. A leverage ratio of 1 would have generated a CAGR of about 7% [3].
Even a modest amount of leverage can significantly improve performance. In the example above, if one used a 2X leverage level, they would earn a CAGR of nearly 12% [4]. This is only about 200 basis points less than using a leverage level of 3X. The reason for this is that when the leverage ratio increases, value is destroyed by the daily resetting nature of these products. For this reason, a leverage ratio of 4X in this example actually performs worse than a leverage ratio of 2X or 3X [5].
The Bottom Line
Leverage ETFs are risky. But, when managed well they can be a tool for long-term investors to increase their returns. One portfolio long-term investors could construct may consist of 80% SPY and 20% SSO. This would create a leverage ratio of 1.2 (0.8 x 1 + 0.2 x 2 = 1.2). Assuming the historical performance of the S&P 500 continues in terms of returns and volatility, this should modestly improve annual returns without introducing too much risk. However, the investor would need to rebalance their portfolio daily in order to maintain their target leverage level: something that would be a challenge to many retail investors. An ETF issuer like ProShares or Direxion ETFs could introduce products with leverage levels of 1.25X or 1.5X to solve this problem.
I look forward to your comments.
Appendix and Disclosure
A Common Criticism of this Approach
Isn't leverage borrowing money? Why would I borrow money to invest? That sounds risky!
This gut reaction is a good one. Using leverage does increase risk! It also, indirectly, often involves borrowing money. Leveraged ETFs use debt or derivatives to hit their performance targets. SSO, a 2X leveraged S&P 500 ETF, uses debt to hit its target performance.
However, the important distinction is this debt is held by the ETF and not by you, the investor, personally. Let's say the S&P 500 were to go down 75% in one day (and let's pretend that trading on the exchange is not halted). This means that SSO should decline by 150% that day. However, the price of ETFs cannot drop below 0. So, even though you're invested with borrowed money, you cannot lose more than your initial investment (before brokerage fees and trading costs, of course).
Additionally, most equity investors use leverage when they invest—whether they know it or not. After all, many public companies take on debt. Therefore, if you purchase shares in a company with debt, your investment performance is subject to the degree of leverage employed by the company.
What is most important is that you, the investor, control the level of leverage you employ so that you can reach your goals. Otherwise, you're leaving that up to the management of the firms in which you invest.
Notes
[1] This ignores the performance gained through reinvesting dividends. However, even if dividends are re-invested, the performance of SSO is far higher than SPY. End date of mid-December 2022 chosen as that is when I began my analysis.
[2] Cooper, Tony. "Alpha Generation and Risk Smoothing using Managed Volatility." 2010. Available at SSRN here. Numbers sourced from Figure 2 on page 8. Importantly, the figures here ignore expense ratios, transaction costs, etc. of using leverage directly or using leveraged products.
[3] Ibid.
[4] Ibid.
[5] Ibid.
Disclosure
This is not investment advice and is for informational purposes only. It should not be used to guide your investments. Speak with a financial advisor (preferably one that is a fiduciary) before making an investment in any financial product. There are risks associated with investing. The past performance of any security may not predict future results.