What Leveraged ETFs Would Have Done in a Real Bear Market
Every backtest of TQQQ is measured over a period that begins in 2010, because that is when it launched. That happens to exclude the two worst declines the Nasdaq has ever experienced. Here is what the mechanics imply about the missing scenario.
#The gap in the record
TQQQ began trading in February 2010, roughly eleven months after the March 2009 market bottom, at the start of one of the longest bull markets in US history. Its entire track record sits inside an unusually favorable environment for a leveraged product.
The TQQQ versus QQQ comparison shows what that produced: a strong result over the last decade alongside a worst-case decline of about 82 percent during 2022. That 2022 episode is the closest thing to a stress test in the entire record, and it was a single year inside an otherwise rising decade.
#What the underlying index actually did
We do have the input. QQQ traded through the dot-com collapse, and the drawdown figure on the QQQ versus SPY comparison is stark: a decline of about 83 percent from March 27, 2000 to October 9, 2002.
That is the unleveraged index. Two and a half years, not a crash, with repeated failed rallies along the way. The path matters enormously for what follows.
#Why 3x does not mean three times the loss
A 3x daily fund does not deliver three times the multi-year return, in either direction. It delivers three times each single day’s return, compounded off a base that resets every day. Over an extended decline this produces something considerably worse than a simple multiplier, and it does so for a reason that is easy to state.
Consider a period where the index falls 10 percent and then rises 11.1 percent, returning exactly to where it started. A 3x daily fund falls 30 percent to 70 percent of its value, then rises 33.3 percent, reaching about 93 percent. The index is flat; the leveraged fund has lost 7 percent to nothing but the path.
This is volatility drag, and it accumulates with every oscillation. The dot-com decline was not a smooth slide. It was a long sequence of sharp drops and violent bear-market rallies, which is close to the worst possible environment for a daily-reset product.
#Where the arithmetic leads
Applying 3x daily leverage to a decline of that depth and duration produces losses that approach total. The precise figure depends on the exact daily sequence, but the shape of the answer does not depend on the details: an 83 percent underlying decline spread across two and a half volatile years leaves a 3x daily fund with a very small fraction of its starting value.
The recovery arithmetic is what makes this close to unrecoverable. Losing 95 percent requires a 1,900 percent gain to break even. Losing 99 percent requires 9,900 percent. For comparison, an 82 percent loss, which TQQQ did experience in 2022, already required a gain of about 456 percent to return to its prior peak.
- A fund does not need to reach zero to be effectively destroyed. Below about 90 percent, the required recovery exceeds anything the underlying index has historically produced in a reasonable timeframe.
- Leveraged funds can and do close after severe losses, or reverse split to keep the share price presentable. Neither restores the capital.
- Financing costs on the leveraged exposure rise with interest rates, and rates often rise during exactly the periods that hurt these funds most.
#The contribution plan complication
Someone contributing steadily through such a decline would be buying continuously as the fund collapsed, which sounds like it should help. Against an index that recovers, it does. Against a leveraged fund, the recovery arithmetic works differently.
Each contribution is subject to the same volatility drag from the moment it arrives. Money contributed near the bottom would have participated in the subsequent rebound, but money contributed during the two years of decline would largely have been consumed. And because the fund needs a multi-hundred-percent gain merely to return to its starting point, the shares purchased early in the decline contribute almost nothing to the eventual outcome.
This is the opposite of what happened to plain QQQ over the same decade, where a contribution plan finished ahead of SPY precisely because its cheap purchases near the bottom recovered fully.
#What these products are actually for
Leveraged ETF prospectuses state plainly that the funds seek their stated multiple of daily returns and are intended for short holding periods with daily monitoring. They are built as trading instruments. The buy-and-hold contribution plan this calculator models is not the use they were designed for, and the daily reset is the reason.
None of this makes them fraudulent or hidden. The mechanics are disclosed. It does mean that a backtest showing strong results over 2010 to today is measuring the product in the single most favorable environment it has ever encountered, and that reading it as evidence of long-term suitability is a mistake about what the sample contains.
#What to test
- Run TQQQ from November 2021 to December 2022 in isolation to see the 2022 decline without the surrounding recovery absorbing it.
- Run QQQ from 2000 to 2002 by itself. That is the input to every inference in this post, and it is real observed data.
- Run TQQQ over its full history alongside QQQ, then note out loud that the window starts in 2010 and why that matters.
- Compare the drawdown figures rather than the return figures. For leveraged products the drawdown is the number that determines whether the return was ever collectable.