VOO, SPY, QQQ And SOXX Vs Leveraged ETFs : May 2022 Decay Showdown (NASDAQ:QQQ)

Key points carried forward from February 2022

• Leveraged ETF decay costs have trebled in the last 5 months but now appear to have stabilized and plateaued
• Leveraged ETF decay costs tend to be directional: in the absence of any further market shocks, I expect stable (not increasing) LETF decay costs.
• If markets stabilize or recover, I would expect LETF decay costs to start decreasing from current levels.
• 2X ProShares Ultra S&P 500 (ARCX:SSO) will leak only 4% per year (up from 1.5% in February 2022) relative to holding 1X SPDR S&P 500 ETF (NYSEARCA:SPY) or Vanguard S&P 500 ETF (VOO). SSO still represents a very inexpensive form of gearing/leverage.
• However, 3X Direxion Daily Semiconductor Bull (ARCX:SOXL) LETFs at current levels, all things being held equal, is set to lose 47.3% (annualized loss relative to the semiconductor index iShares Semiconductor ETF (SOXX)) of its investment value to decay costs every year (up from 41% in February 2022) .
• Similarly, the short LETFs ProShares UltraPro Short QQQ (SQQQ) and ProShares UltraPro Short S&P500 (SPXU) are very expensive at the moment. They lose nearly half – nearly 50% (annualized cost) of their investment value per year to costs in current market conditions.
• Investors should migrate from 3x to 2x LETFS, avoid LETFs on less liquid indices other than on the S&P500 and Nasdaq100, and avoid short LETFs in current market conditions.

The charts in this article build on the (somewhat different) leveraged ETF decay pricing methodology which uses bond / fixed income bootstrapping techniques to “bootstrap” the LETFs returns relative to their underlying index, thereby isolate and eliminate price movements and compounding effects from the equation and what remains is the decay curves per Figure 1 above. I have tried to explain the methodology logic and modelling in my previous articles (see here, here, here and here) as well as annexures thereto. (comments welcome). Please note that I am not expressing any views whatsoever on the likely price movements or weightings in the underlying indices (including SPY, VOO, Vanguard Small Cap Index Fund (VB), and semiconductor indices). I am specifically only looking at the investment decision relating to the holding costs of LETFs relative to other lower-cost LETF (less expensive) alternatives that provide an equivalent index exposure. The objective is only to reduce LETF holding or decay costs in times of a market correction, not to time the market or suggest sector weightings (other than to highlight that certain illiquid, difficult to hedge sectors incur significantly higher LETF decay costs than SPY, VOO and Invesco QQQ Trust (QQQ)).

3X LETF data analysis:

• ProShares UltraPro QQQ 3X (NASDAQ:TQQQ)
• Direxion Daily Semiconductor Bull 3x SOXL
• Direxion Daily S&P 500 Bull 3X (SPXL)
• ProShares UltraPro S&P500 (UPRO)
• ProShares UltraPro Short QQQ SQQQ
• Direxion Daily Small Cap Bull 3X (TNA)
• ProShares UltraPro Short S&P500 SPXU

2X LETF data analysis:

• ProShares Ultra QQQ (QLD)
• ProShares Ultra S&P 500 (SSO)

Volatility (VIX) refer pink arrow in Figure 2 is trending down. More on volatility later.

The above tabulation (Figure 3) and the bar chart below (Figure 4) rank the leading LETFs with the lowest cost S&P and Nasdaq 2x LETFs on the left and higher cost 3X, and Short, esoteric LETFs on difficult, costly to hedge indices (including gold, minerals, semiconductors, and financial LETFs) on the right.

The following example will hopefully illustrate how an appreciation of LETF decay can be used to understand risks and potentially save costs.

LETF and CFD REBALANCING SCENARIO: Assume that an Investor has a $1m portfolio invested into Vanguard S&P500 VOO. Investor has also looked to increase (and profited from) his index exposure during the recent bull run by investing $100k into TQQQ Contracts For Difference CFDs as well as $50k SOXL CFDs as well as some other AAPL and Nvidia CFDs. All CFD investments are “on margin”, requiring no further capital outlay further than the $1m VOO’s. Investor’s CFD margining/finance costs are 7% per year. In current market conditions please calculate all likely annualized costs associated with this holding and please advise Investor of alternatives to optimize this holding. Please assume that Investor would like to keep his total index exposure constant i.e. the same index exposure as before the restructure. (Assume that brokerage costs are trivial in relation to LETF decay and CFD finance costs: (which they are)).

SHORT ANSWER: Investor can save annualized cost of ~$32,500 per year (3.25% of his investment value) in decay/leakage/finance costs by moving from 3X LETFs to 2X LETFs, and by moving out of CFDs (whilst maintaining the same segment index exposure as previously (Investor intends to stay in the market to catch the “bounce”))

LONG ANSWER: The costs can be summarized as follows: CFD finance and interest costs are 7% on USD150k CFDs or $10,500 per year. Likely TQQQ “leakage/decay” costs (relative to QQQ) are currently ~28% (refer Figures 1-4 and in particular the Tabulation in Figure 3 above TQQQ for 28 April 2022) per year = $28000 on a $100k TQQQ investment. Likely SOXL “leakage/decay” costs (relative to the underlying Semi-conductor index is currently ~48% per year[refer Figure 3 SOXL 28 April 2022 decay) = $24000 on $50k SOXL investment. So total Leveraged ETF plus CFD finance costs are $62500 per year on the $150 000 Leveraged ETF investments. This $62500 additional cost relative to a direct investment in the underlying index is significant in a sideways or downward market.

INDEX EXPOSURE: Investor believes in staying long through the dips and doesn’t want to miss out on the bounce (if/when that bounce ever materializes). Although only $100k has been invested into Nasdaq TQQQs (on margin) these $100k TQQQ’s produce 3X $100k = ~$300k of Nasdaq QQQ exposure. Likewise the semiconductor SOXLs have 3 x $50 = $150k index exposure on a $50k SOXL investment. Investor also has $1m S&P500 exposure by virtue of his Vanguard VOO investment. So Investor has $1.45m exposure on a $1m investment. (Roughly speaking he has an effective “Beta” multiple of 1.45 times the market on his total portfolio.)

PROPOSED STRATEGY: Refer Figures 1-4: 3X LETFs are very expensive (disproportionately expensive) relative to 2X LETFs. Obviously 2X LETFs require 50% more investment to achieve the same index exposure as 3X LETFs, but even so 2X LETFs are less expensive in multiples. (3X LETFs sail closer to the wind, can risk under-collateralization, forced closure/restructure in adverse market conditions, and their leakage costs and tracking errors are considerably higher).

Moreover all LETF leakage/decay costs have increased dramatically in recent months (increased in multiples).

Investor should consider swapping his 3X LETFs for 2X LETFs and unwind his CFD positions, but the question is how to do this while maintaining his current overall index exposure.

PROPOSED SOLUTION:

1. Sell $100k 3X Nasdaq TQQQ (costing 28% decay + 7% CFD interest per year: total cost $35k per year) and buy $150k 2X Nasdaq QLD (costing 10% per year decay: total cost = $15k per year). $20k per year net saving per year for same $300k QQQ Nasdaq exposure (3X TQQQ $100k = 2X QLD $150k = $300k QQQ Nasdaq 100 exposure) :Take the kids on a $20k ski holiday / weekend with the saving.
2. Sell $50k 3X SOXL Semiconductor LETF (costing 48%decay + 7%CFD interest per year: total cost $35k per year) and buy $150k 1X SOXX Semiconductor index (cost free) thereby saving a further $27.5k per year for the ski holiday (both SOXL and SOXX strategies yielding the same $150k Semiconductor index exposure)
3. BUT Strategies 1. And 2. require cash, so this will require selling Vanguard S&P 1x VOO of $150k and $150k respectively = $300k VOO sale to fund 1 and 2. (We will get to VOO dividend yield later). We are now left with $700k VOO S&O exposure.
4. BUT 3. Above has diluted our $1m VOO S&P exposure to $700k. We need to further sell VOO and buy 2X S&P SSO (SSO is the 2X LETF for VOO) sufficient to re-gain $1m S&P index exposure. Because SSO has a 2X multiplier, we need to sell a further $300k VOO and buy $300k SSO leaving $400k VOO and $300k SSO (with index exposure of $400k VOO S&P exposure and 2X $300k SSO exposure = $400+ $600 = $1m S&P exposure: Back to square one)
5. The cost of selling 3. and 4. above is foregone dividends on VOO of roughly 1.8% per year on the sale of $600k VOO in 3. and 4. (less dividends tax if any (assume 30% dividend withholding tax) Plus the additional decay costs on SSO of 2.5% on the $300k. So reduce the kids Ski fund by $7.56k and $7.5k respectively = $15k for 3. and 4 respectively.
6. So the total calculated saving of 1 to 5 above is $20k + $27.5k less $15k = $32,500 saving per year for the same index exposure as previously. The outcome of 1-6 above is a $32 500 saving per year on a $1m investment or 3.25% saving per year for the kids ski fund.

7. Check the issuer concentration / diversification / liquidation specific risks of the above strategy:

• RISK ANALYSIS: BEFORE SCENARIO: $1m in Proshares VOO (v low liquidation risk) plus $100k Proshares TQQQ (may experience disconnect in Mar2020 COVID type turmoil conditions making TQQQ medium risk), plus $50K in Direxion 3X SOXL (significant Mar2020 disconnect, tracking errors): Direxion 3X SOXL is considered medium to high hedging difficulty and medium to high risk of restructure risk in March 2020 type conditions. (Direxion had to restructure a number of their LETFs immediately post March 2020 and convert 3X to 2X. (3X LETFs sail considerably closer to the wind in market turmoil conditions than 2X LETFs and are subject to considerably higher tracking errors and decay / leakage in Mar2020 conditions than 2X (refer Figure 5 above).
• RISK ANALYSIS: AFTER SCENARIO: The move from 3X to 2X LETFs has considerably reduced LETF liquidation risks.
• PS: The above scenario is based on a real-life case study: The above strategy has reduced ~83% of the above fund portfolios’ holding costs by virtue of the above restructure. Some single stock CFDs remain including e.g. APPL CFDs remain.

8. Consider the remaining CFD Close-out risk: (e.g. consider CFD collateralization/coverage/margining cover ratios before and after (e.g. are remaining CFDs sufficiently collateralized compared to previously?)) A forced close-out of remaining CFD and equity positions due to insufficient margining typically triggered by a temporary downward market spike (inverse spike) (accompanied by Investor ‘missing-the-bus” on any subsequent bounce) would be a bleak and sub-optimal outcome to say the least and to be avoided at all costs. Investor should check marginingcollateralization ratios through the worst March 2020 COVID crash scenarios. Although I consider the failure risk relating to the individual Leveraged ETFs to have reduced considerably (by virtue of migrating LETFs away from more risky (semiconductor) LETFS to less risky S&P and Nasdaq LETFs and away from more risky 3X to less risky 2X LETFs, In the portfolio above under consideration, there are still some CFD’s remaining (remain including APPL CFDs etc). However the following risk has increased: Investor has substituted 1X S&P ETFs for 2X and has doubled the downside “beta”. So CFD collateralization ratios through the worst possible scenarios need to be re-evaluated to ensure that margining and collateral levels are capable of being well covered even through the worst possible crashes. (Forced foreclosure / closeout of the remaining CFDs would be a disaster (as the markets tend to plummet and then quickly bounce back, and a closeout would miss this bounce). Conclusion: Because we have traded out of the majority of CFDs, the remaining CFDs will remain very well covered through the worst possible crash scenarios even though we have substituted $600k VOOs for the more volatile SSO 2X LETFs.

Counter-intuitively and seemingly implausibly: 2X LETFs hold more downside risk for Investor than 3X LETFs (for the same $300 index exposure)

9. One further very final factor to consider is that we have substituted 3X LETFs for 2X LETFs. So we have had to effectively invest 50% more into LETFs to achieve the same index exposure as previously. 3X LETFs have capped downside (limited to the $100 investment) whereas by swapping to a lower multiple LETF, we have had to invest 50% more to achieve the same index exposure and that downside cap has now (theoretically) been expanded by 50%, with more possible downside risk. But the counterargument is that 2X LETFs decline at a slower rate than 3X LETFs and moreover, we are not going to close out for the foreseeable future by which time hopefully the markets have recovered. Furthermore we have drastically reduced the LETF holding costs. So on the balance, I am in favor of the move from 3X LETFs to 2X LETFs in current market conditions (“de-risking”).

QUESTION: Thinking a little more about 9. above, I questioned whether an investor should be indifferent between the following 3 strategies (before considering decay costs, costs of capital, dividends) :

• $300 investment in 1X QQQ (with a Nasdaq exposure of $300)
• $150 invested in 2X QLD (with a Nasdaq exposure of $300 [2X $150 = $300])
• $100 invested in 3X TQQQ (with Nasdaq exposure of $300 [3X $100 = $300])

SHORT ANSWER: Yes (but only in the short term), Investor should be indifferent between the 3 strategies because each produces the same $300 Nasdaq exposure.

LONG ANSWER: Again counter-intuitively, in the long term, or in rapidly moving directional markets, the above 3 strategies differ considerably. Out of curiosity I ran the above scenario for 10 years with $300k invested in QQQ, $150k in 2X QLD, and $100k in 3X TQQQ, all 3 scenarios initially starting with $300k of Nasdaq exposure. The divergence over 10 years is surprising and remarkable:

CONCLUSION: Starting with the same $300k NASDAQ index exposure in 2010 the 3X LETF 3X TQQQ handicapped to an $100k TQQQ investment outperformed the $300k QQQ by 4.5 times (measured in gross USD P&L terms). 2X LETF QLD with $150k initial investment outperformed $300k QQQ by 2.1 times. The TQQQ Turbo Rabbit finished miles ahead of the QQQ Tortoise despite Rabbit’s 3 times starting handicap.

These conclusions initially appear counter-intuitive: Surely $300k initial Nasdaq investment should yield the same returns over time?

The same conclusion holds for S&P500 LETFs: (albeit with lower S&P growth than Nasdaq): The Rabbit finished miles ahead of the tortoise despite Rabbits’ 3-fold starting handicap.

Intriguingly LETFs also perform better than the underlying index through the crash:

A further oddity can be observed through the 2020 crash: One would expect LETFs (the Turbo Rabbits) to fare worse than the index through March 2020 (particularly if they outperform when the index is rising), but the opposite holds true. $300k of LETFs S&P SPY exposure lost $102K through the crash whereas the 3X LETFs lost only $77k. So LETFs appear (peculiarly) to outperform the index both in bull and bear phases. If the TQQQ Turbo Rabbit moves forward more than 3X faster than the QQQ Tortoise, then one would expect the same Turbo effect in reverse gear, but Rabbits turbo is considerably suppressed, de-tuned in reverse gear. Please note that the following Chart 11 (as well as charts 9 and 10 above are referenced to actual prices (after taking into account LETF decay, leakage, synthetic gearing costs, volatility costs, LETF tracking errors etc), without which costs the LETFs would have outperformed the index by an even greater margin than that represented in the charts in both rising and falling market scenarios)

The SPY and QQQ indexes (Grey lines above) underperformed the LETFs in all 3 charts above, in both bull and bear phases (even after taking into account considerable LETF decay costs).

So if LETFs outperform the index in both bull and bear phases, then why not only invest into 3X LETFs, unlock 2/3rds of your capital and spend the $200k on lifestyle, houses, expensive cars and other vices?

ANSWER: The answer lies in the math. Caveat: The following section may well serve as a good substitute for a strong sleeping pill should anyone really be having difficulty in falling asleep at night:

As further corroboration of TQQQ LETF outperforming the 3X objective on the downside please observe TQQQ’s drop last Friday 29 April 2022 (refer Fig 12 above). On the face of it TQQQ should have dropped at least 3 X QQQ i.e. 3 X 4.50% = 13.50% drop before considering TQQQ decay / leakage relative to QQQ, and TQQQ should have dropped more than 13.5% if we factor in LETF decay. However TQQQ only dropped 13.15%. This intra-day difference is massive. From Annexure B minute-by-minute tracking error analysis and also from the LETF tracking error variances Figure 13 below, we see that TQQQ tracks QQQ extremely accurately and we can be 95% confident and certain that the actual TQQQ intra day return is accurate within the bands of 99.97133% and 100.003% of -13.15% (i.e. 95% confidence of tracking error between – 13.1462% and -13.1504%). i.e. the -13.15% TQQQ price drop is for all intents and purposes the correct answer (which is light years away from 3 X QQQ i.e. 3 X 4.50% = 13.50%). Why does TQQQ perform better than its 3X stated objective in sharp intra-day downturns? If this always holds true then why not only invest $100k i.e. one third of your capital into 3X LETFs, unlock 2/3rds of your capital and spend the remaining $200k on lifestyle, houses, expensive cars and other vices? Again the answer lies in the math.

[I’m going to leave the math for next month’s riveting LETF instalment. Comments welcome and Kudos to correct responders. Please refer my previous articles and annexures. But in thinking about the math please also bear in mind that in the short term there are no free lunches/windfall gains/slights of hand/miraculous profits the mysteriously materialize in the financial markets arising from mathematical anomalies. There is always a catch. The markets are a zero sum gain: Someone wins at somebody ease’s expense, and the banks/issuers/platforms are usually not the losers. Gaps/pricing anomalies are usually quickly arbitraged out.]

Daily LETF tracking errors

Let’s have a look at daily variances between achieved LETF multiples and how these do or don’t correlate to volatility:

Key observations:

Volatility has decreased (green circles)

So too has the number of dark red cells in the last 10 days. Dark red cells indicate a negative variance: i.e. an underachievement of daily instances of LETF 2X and 3X LETF objectives. Dark red cells indicate days of LETF decay. Dark red cells (days of most pronounced LETF decay/leakage/slippage/tracking error) usually occur in the most turbulent marked conditions. Extreme market frothiness makes hedging LETFs more difficult resulting in greater variances (both positive and negative), and market turbulence is generally associated with greater decay.

The onset of increased decay generally has lagged volatility spikes (for reasons suggested in previous articles)

Green cells in the last 10 days indicate positive LETF multiplier variances on certain days (which would reduce LETF decay).

A number of academic articles suggest that volatility is a driver of LETF decay. The term “Volatility decay” is widely used. Volatility (VIX) is derived (back-solved) from option prices. Option prices reflect the cost of inter alia insuring against market downturns. In frothy, uncertain markets the volatility “fear index” increases (option premiums become more expensive).

Also in next months article: The statistical concepts of Correlation vs Causation: Does volatility cause or drive LETF decay? (or do declining market prices independently drive both Volatility and LETF Decay factors separately and unrelated from one another?): (Comments also welcome)

Finally a quick litmus test check on SOXL Price Charts vs LETF Decay Charts:

LETF Decay charts are otherwise heavily distorted by exponential price movements, and compounding which distorts and masks the effects and calculation of the decay/leakage/ tracking error costs imbedded in all LETFs.

However current sideways markets provide us with a rare glimpse and hitherto unseen opportunity to isolate and eliminate the effects of index price movements and their compounding effects on LETFs. Simply put: If index prices increase and decrease over a period and then end up at the same level as they started, then the product of all daily price movement gain and loss factors over the period will equate back to exactly ONE. We will thereby have isolated and eliminated all of the distorting effects of price changes and compounding from our analysis for that particular period under review. This can be manually achieved by zooming in and out of price charts in order to identify a period of flat index return (blue line), and the comparing LETFs (red chart below) to their flatline underlying indices (blue chart below and flat blue line). The resulting differential (with price movements and compounding eliminated) can only be attributed to LETF Decay. In the case of Direxion Daily Semiconductor Bull 3X Shares SOXL we observe approximately 46% loss (refer red arrow below) in the LETF over the period 1 Feb 2021 to 30 April 2022 (15 months), or roughly 38% annualized loss of the LETF relative to its underlying index. This is a big loss of $38,000 per year on a $100,000 SOXL investment.

The following decay curves can be mathematically derived over any period by bootstrapping the LETF relative to its underlying index. (refer previous articles) The decay model yields a SOXL decay of 38.4% for the period (which for the propeller-heads out there is one minus the product of each daily decay factor (compounded) over the period) This SOXL ~38% loss can also be roughly visually corroborated by the red highlights below (where Decay ranged between 15% and 50% over that period and appears to average around 38% for the period). [The correct math is a little more involved than simple arithmetic averaging, but eye-balling, averaging, visual corroboration are fine, close enough for the litmus test purposes]

Conclusions:

• Investors are advised to steer well clear of high decay LETFs in current market conditions, both from a safety and cost perspective.
• The same market and index exposure can oftentimes be achieved by moving from 3X LETFs to 2X LETFs.
• No free lunches in the financial markets.
• I intend to move the small “at risk” LETF weightings back from 2X to 3X LETFs when leakage factors return to normal November 2021 levels (without betting the farm). We should see a return to normal levels in the next few months (refer “table tops” or “wave pattern” in chart below)