S&P 500 Leveraged ETFs ‘Til The New Year

“October: This is one of the peculiarly dangerous months to speculate in stocks. The others are July, January, September, April, November, May, March, June, December, August and February.”

― Mark Twain, 1894

With the famous October 1929 market crash, the October 1987 crash, and the financial crisis roughly starting in October 2008 (16.7% drop), this quote seemingly was quite prescient. However, the only month over the last 100 years that has an average negative return is September. Exhibit 1 shows each month’s average return, standard deviation, minimum, and maximum from Jan. 1926 through Dec. 2023.

Whether September 2024 follows the historical average remains to be seen as it has experienced a jumpy start. The histogram in Exhibit 2 shows there have been 47 positive and 51 negative returns in September with some negative skewness, meaning a few larger negative returns relative to positive.

For the extreme bearish investor, or bullish as November and December are two of the better months (only behind July), what are the options?

Options of course, although many may find that unpalatable, may not have an options approved account, or may not even be able to trade them in certain retirement accounts. Buying on margin is another possibility, but then one needs a margin account and may have to deal with margin calls. Perhaps the easiest way for the very bullish/bearish to make market calls is using leveraged exchange-traded funds, (LETFs) which have become increasingly popular since their introduction in 2006. There are now more than 170 of these funds with $100 billion in assets on U.S. exchanges.

For the S&P 500, there are daily +/- 2.0x (ProShares SSO and SDS), +/- 3.0x (Direxion SPXL & SPXS, ProShares UPRO & SPXS), and a 4.0x exchange traded note (BMO’s SPYU). If you happen to reside outside the U.S. or have a foreign brokerage account, the London Exchange has a 5.0x fund (Leverage Shares SP5Y). This September, Tradr added 2.0x calendar LETFs with a weekly (SPYB), monthly (SPYM), and a possible quarterly reset (SPYQ) being introduced in October. Note this is only a partial list and the same opportunities exist for the Nasdaq, SOX (semi-conductors) and a multitude of other underlying indexes and even a few individual stocks that are magnified. This analysis will limit itself to the S&P 500 LETFs listed in Exhibit 3.

Long-term LETF Simulated Returns

With all warnings aside about using historical results to make dubious timing bets using monthly averages with only September having a statistically significant difference (which may be spurious at that), Exhibit 4 shows the returns that theoretically could have been made after accounting for transaction costs, borrowing costs, and expense ratios for the various funds. For the interested reader, see Trainor, Chhachhi, Brown (2020) for the details on how these returns are calculated. Daily return data is from the Center for Research in Security Prices value weighted S&P 500 index from Jan. 1926 through Dec. 2023 which before 1957 was the S&P 90.

It should be noted daily LETFs are not meant to double their underlying index return for longer than a day due to “beta” drift which is generally negative, i.e. the index return multiple tends to fall over time although can increase during trending markets with low volatility. These realized leverage ratios can even have wrong way leverage, i.e. a bullish LETF is down when the underlying index is up. FINRA does warn: ” Because they reset each day, leveraged and inverse ETFs typically are inappropriate as an intermediate or long-term investment. They may be appropriate, however, if recommended as part of a sophisticated trading or hedging strategy that will be closely monitored by a financial professional. “

Tradr’s new calendar LETFs overcome this issue to some extent, at least through their reset periods, meaning their weekly and monthly 2.0x LETFs will multiply the underlying index weekly or monthly return by 2.0x respectively. In addition, research shows bullish LETFs can be held for longer periods of time and even reduce risk if actively managed, (Scott and Watson, 2013; Trainor, 2011b, George & Trainor, 2017; Trainor, Chhachhi Brown, 2020). The easiest way to imagine LEFTs reducing risk is consider that a typical 60/40 portfolio only needs 30% of a 2.0x to attain 60% exposure, and one theoretically could now have 70% in “safer” bonds, ignoring maturity and default risk, both of which can be managed.

In addition, the long-run performance of broad market bullish LETFs has been nothing short of outstanding. In fact, the top six performing ETFs over the last 10 years are all LETFs. While the SPY is up 237%, (10 years through 9/13/2024) the S&P 3.0x LETFs are up over 730% and the 2.0x LETFs are up over 500%. Note all of these LETFs can suffer large and painful losses rather quickly, but in the right context and actively managed, LETFs can be held for longer periods of time. For inverse funds, this is NOT the case.

Given the caveats above, Exhibit 4 shows results based on monthly holding periods. The weekly calendar LETF is not included as the weeks are mismatched relative to monthly returns and results generally lie between the daily and monthly calendar reset. The 2.0x funds have average and median returns approximately 1.6x the underlying index although this is just based on aggregated statistics. The 3.0x, 4.0x, and 5.0x show similarly fading return multipliers. With that said, for a correct market call, 3.0x to 5.0x have monthly maximums from 161% to 342% which occurred in April 1933.

Before you run out and buy a 4.0x or 5.0x, Exhibit 5 shows the ending values of $1 for each fund if you had bought-and-held since Jan. 1926, Jan. 1940, and March 1957 when the S&P 500 was introduced. The $0.03 ending value is not a typo for the 5.0x. It can do extraordinarily well in upward trending low volatility markets. However, it also can quickly suffer almost catastrophic losses, recall the market’s 20% loss on Oct. 19, 1987 and multiply! In fact, the 2.0x LETFs outperform both the 3.0x and 4.0x if one includes the volatile 1930s.

The 1940 and 1957 start dates show 2.0x is better than 1.0x, the monthly 2.0x is better than the daily 2.0x, and the 3.0x gives maximum returns, better than both the 4.0x which is better than the 5.0x. There is little to suggest that buying and holding a 4.0x or higher other than for shorter term market bets is a good long-term strategy (see Trainor 2017).

This is not mathematically unexpected as the relationship between geometric and arithmetic returns can be approximated as X_PT = X_RT – 0.5 s_t² where X_PT is the geometric return, X_RT is the arithmetic return, and s_t² is the variance of returns. The variance term outweighs the higher leveraged returns much past 3.0x for the typical volatility experienced by the S&P 500 resulting in declining geometric returns. Using annual returns, the annual average and standard deviation for the S&P 500 from 1926 through 2023 is 12.16% and 21.02%. This gives an approximate long-term geometric annual return of 0.1216 – 0.5*.2102² = 9.95%. For the 5.0x with a 44.76% average (28% median) and 102.05% standard deviation the expected geometric annual return is 0.4476 – 0.5x*1.0205² = -7.31%!

For the inverse funds, long-term holdings generally result in a 100% complete loss as markets generally trend up. Mathematically, they suffer “beta” drift even worse than bullish LETFs and when the market does decline, their expected positive results are mitigated as volatility generally spikes during severe market drops. During the financial crisis, there were a number of investors unhappy with the fact their inverse LETFs declined despite the market being down as well, (Maxey, 2009). Thus, one should use these inverse funds for short-term market bets only. Exhibit 6 shows their simulated return statistics.

September through December

Exhibit 7 shows the Sept., Oct., Nov., and Dec. average and median returns for the various LETFs. For September, the inverse funds have been positive with a -3.0x earning 142%, theoretically anyway in 1931. This is the only month they average positive returns. October does not show any large differences between any of the bullish funds with the 5.0x being best and the 2.0x calendar LETF better than the daily, especially based on median returns and outperforming its 2.0x daily counterpart 66% of the time.

November and December favor the 5x with 7.22% and 6.26% average returns respectively. The calendar monthly reset again does slightly better than the 2x daily LETF but this is generally expected as it does not suffer beta decay over its monthly reset period.

In summary, results are what would be expected on average but there is significant variation ranging from -98% (5.0x) to 142% (-3.0x). Although not shown, all the 2.0x or greater funds have losses at one time or another exceeding 25% with maximum gains equally substantial. As shown earlier in Exhibit 4, in any given month, LETFs could lose 50% or more.

Quarterly Results

Although there is currently no quarterly reset LETF (one has been registered with the SEC), the possibility merits a quick examination of the daily vs monthly vs quarterly LETF for Oct. through Dec. Exhibit 8 shows the results. On average, the last quarter is a very good quarter to leverage up (S&P 500 3.90%, average quarter is 2.91%) with a 5.0x median return of over 20%, unfortunately coupled with a -98.7% minimum. Over the last quarter, LETF minimums start at -50% offset by large possible gains as well. The probability the last quarter will be positive is over 75% compared to the average for all quarters of 68%. For just the 2.0x, the calendar monthly or quarterly reset do have moderately better average and median returns while outperforming the daily LETF 56% of the time.

For those who may be considering the weekly or monthly calendar resets, note the 2.0x leverage will only hold if one invests at the beginning of the week or month. If the market has increased after this date, the leverage will be less than 2 and if the market has decreased, the effective leverage will have increased. This is also the case for daily LETFs but the effect is muted as the market moves little within a day relative to a week, month, or quarter if such a fund appears. To determine the leverage within the reset period, the following equation from Trainor (2011a) can be used:

Leverage = (β – 1)/(1 + βr_t) + 1

where β is the initial leverage ratio and r_t is the return of the underlying index from when the fund is rebalanced to when the fund is purchased within the rebalancing period.

Conclusion

The Ides of September are upon us. September is the only month the S&P 500 has an average negative return. On the horizon, the final quarter of the year approaches, which has a greater market return than average (3.9% vs 2.9%) with a greater possibility (75% vs 68%) that the quarter’s return will be positive. Both of these factors are favorable for LETFs.

This study has examined the gamut of S&P 500 LETFs currently available, from -3.0x to 5.0x. At the 2.0x level, investors now have choices from daily, weekly, monthly, and possible quarterly resets, the last of which may be available as soon as this October. For monthly or quarterly holding periods, Tradr’s SPYM and potentially SPYQ have some advantages with their longer resets relative to the daily 2.0x LETF, but the returns are not that much different. However, there is a much greater degree of realized leverage certainty if held for the duration of the reset period.

For the extremely bullish investor, a 3.0x, 4.0x, or even 5.0x could do well over the last quarter of this year, based on history anyway. However, regardless of the bullish LETF chosen, maximum losses start at -50% with -98% a possibility. One should partition their portfolio carefully if using these funds over any extended time period. The 2.0x funds and 3.0x funds can effectively be used long-term as long as one understands these funds are exposed to sudden large and sometimes drastic losses. They need to be actively managed. If over time your LETF does well, rebalancing is key as your portfolio may become overexposed to the extreme volatility that LETFs can experience.

References

Maxey, D. 2009. “ProShares Draws Suit Over a Leveraged ETF,” Wall Street Journal, Aug. 7.

George, J. & Trainor, W. (2018). Portfolio insurance using leveraged ETFs, Financial Services Review, 26, 1-17.

Scott, J. & Watsun, J. (2013). The floor-leverage rule for retirement. Financial Analysts Journal,69, 45-60.

Trainor, W. 2011a. “Daily vs Monthly Rebalanced Leveraged Funds,” Journal of Finance and Accountancy 6: 1-14.

Trainor, W. 2011b. Solving the Leveraged ETF Compounding Problem. Journal of Index Investing, 1: 66–74.

Trainor, W. 2017. “Leveraged Exchange-Traded Funds: When Four is Not More,” Journal of Investment Consulting 18(1): pp. 31-40.

Trainor, W. Chhachhi, I. and C. Brown. 2020. “A Portfolio of Leveraged ETFs.” Financial Services Review, 28(1): 35-48.