Let's say someone has a few thousand saved in an IRA. Right now I have it sitting in Treasurys earning around 3%.

Over the past 50 years the S&P has averaged around 11% per year in total returns.

Let's say they wanted to invest the IRA balance in a 3x leveraged S&P ETF, specifically UPRO. If the S&P averages 11%, napkin math suggests an average ~33%, minus fees of about 1%. But it's not clear how the built-in leverage works for or against this method.

What happens when investing in these instruments long-term?

  • 17
    No. Not risky. Just one of the stupidest ideas ever. You are guaranteed to lose money over the long term with daily-leveraged ETFs. If you really want to invest with borrowed money, bite the bullet and borrow it yourself.
    – user296
    Commented Oct 20, 2011 at 4:13
  • 8
    I just wanted that to be painfully obvious to anyone who found this question. :)
    – user296
    Commented Oct 21, 2011 at 1:53
  • 4
    Read this: zerohedge.com/news/2016-01-02/…. Summary there are people betting that the entire leveraged ETF industry will go under, like the Subprime crisis. Most people (including at this site) do not understand how these work or why they carry the risk they do. I wouldn't put this as an answer as it would be incredibly unpopular to write (people don't like being called ignorant).
    – YaReally
    Commented Jan 4, 2016 at 12:43
  • 2
    Another good article for anyone who thinks leveraged funds are sprinkled with magic pixie dust: news.morningstar.com/articlenet/article.aspx?id=271892
    – timday
    Commented Jan 4, 2016 at 16:11
  • 1
    Sorry if my flagging was incorrect. There didn't seem to be a lot of consensus when I had asked on meta about how to handle all the questions we have on leveraged/inverse ETFs. I figured I'd try flagging a couple which were clearly duplicates, and chose the one that had slightly more total votes and views (and I'd answered) as the "standard" version. No hard feelings were intended. Perhaps we should just embrace the duplication.
    – user42405
    Commented Aug 23, 2017 at 16:31

9 Answers 9


Keep in mind, if the S&P were up 10% one day and down 10% next day, I am down 1%. But triple this, and 1.3 x .7 is .91, down 9%. This phenomenon is enough to make these 3x type ETFs not recommended for the long term.

Since you are considering this strategy with your hard-earned money, I respectfully suggest this exercise. Take an X day period, 10 days, 100, you decide. I'd go high. Create a spreadsheet with a column of the S&P index and next column its return for that day. In next column, triple it, and then calculate total return for the period. I may do this very thing for an article I plan to write, a follow on to my "ETF'd" which focused on the inverse ETFs out there.

My simple example was to exemplify one simple point, returns are cumulative. If the S&P can be down 1% over a 2 day period, but your triple ETF isn't down 3%, but a full 9%, what do you think you'll learn from the spreadsheet exercise? Be sure your time period includes the week the S&P moved over 50 points each day. Over a long period, those volatile days will happen with some regularity.

@chris thanks for the comments, it seems my original response was too simplistic, I added further thoughts with this edit.

  • 7
    Supporting reference. Refer to proshares.com/3xetfs/index.html: "UltraPro ETFs seek to provide a return of 300%, or 3x, the return of an index for a single day [...]" Commented Oct 19, 2011 at 11:54
  • 16
    @Travis, Joe is trying to tell you that these products are not designed for the long term. By design, they are more likely to fail to deliver you returns over the long term even if the market itself has performed admirably. Go here, read this, admire the comparison chart and allow your jaw to drop: seekingalpha.com/article/… (fixed the link) Commented Oct 19, 2011 at 11:57
  • Note that leveraged ETNs exhibit significant decay over time and have higher-than-average fees (compared to non-leveraged ETFs)
    – 0xFEE1DEAD
    Commented Aug 23, 2017 at 15:58

Leveraged ETF's don't work that way. They only provide the expected leverage over the short term - maybe a few days or weeks. On time frames longer than that, they go down and do not track the index.

You can see for yourself by comparing UPRO and S&P 500 on the same chart. The longer the time frame, the greater the difference.

By the way, one solution is to buy the unleveraged S&P 500 index on a regular basis (like monthly). Look for the one with the lowest fees (like Vanguard).

Another solution is to buy when the market is down. Since you will not be using the money for 40 years, this strategy would (in theory) provide much better returns.


I just compared UPRO and SPY from the time this question was originally posted on October 19th, 2011 through today (September 7th, 2016). See the chart here.

The return for SPY was 78.7% and the return for UPRO was 427.57%. I looked at nearly every 12 month segment over those 5-ish years and about 25% of the time SPY outperformed UPRO, and the other 75% of the time, UPRO outperformed SPY. UPRO was extremely volatile, and I'm not sure I could have held onto it for that entire time period, but over the ~5 year period since this question was posted UPRO performed 5x SPY.

I hope our friend didn't listen to the early answers to his question. I'm not sure we can count on market conditions being the same as the last 5 years, but if they are - this strategy might pay off.

Disclosure: I don't own any UPRO shares, but I'm now considering buying some.

  • 2
    Welcome to Personal Finance & Money. Please see the Help Center for tips on writing great answers. For example, you've offered helpful data and citations in this answer, but your math is way off. $5,000 * 5.2757 = $26378.50, not "multi million"s.
    – dg99
    Commented Sep 8, 2016 at 5:30
  • 2
    The price adjusted between each trading day to match the market open, so if you hold overnight then you lose money each day. This is by design.
    – user4946
    Commented Sep 9, 2016 at 18:41
  • I hope you did buy some back then! If so, congratulations on the gains! Commented Oct 27, 2021 at 4:53

It is actually a good idea, if you plan on investing for 25+ years. If you look at (hypothetical, zero-error tracking) historical data, you will see that you almost always do MUCH better investing in UPRO than SPY if you have a long time horizon. Since 1950, the annualized growth rate is just over 16%. Most people who say it is a bad idea haven't actually looked at what the hypothetical returns would be from real SPY data. There is decay during periods where the underlying index fund is flat, however this effect isn't nearly as bad as it is commonly exaggerated to be. There are other costs besides the decaying effect, for example the expense ratio is usually 0.95-0.98% on leveraged etf's. There are also other hidden costs incurred while operating these funds such as the commissions and bid/ask spread, which could potentially have a greater effect than the decay that others talk about. The extent is in the fund performance.

Another thing to think about, however, is that the investment will be more volatile. So, you would see the value of your portfolio vary widely. For example, the maximum drawdown (again, using hypothetical data) was 97.7% for the UPRO in the early 2000's.

In any case, I'm not trying to convince you to buy UPRO. However, you should know that you aren't crazy if you buy it and plan to hold long term. You are not guaranteed to lose money over the long run.

Disclosure: I hold shares in UPRO.


enter image description here

Let's suppose that you used a 3x during the 1980s with the Nikkei. How would you be looking right now?

Now, a few voters here may be angered that I don't write out a long answer, but the image I show is exactly what's wrong with these leveraged funds. They aren't too bad if you're timing is perfect, but if our timing was perfect, we wouldn't be using index funds anyway.

  • 4
    ...but what if you hold onto the asset until 2013? Wouldn't you have made a (huge) profit if you'd just waited it out a few more years? Like the OP I don't really get what's wrong with this. Commented Oct 5, 2018 at 0:06
  • Okay, I bought a leveraged ETF in April, during the pandemic.Should I sell it is soon as the pandemic is over and the index is on normal values (already there)? Or I still would be in plus anyway long-term?
    – Anixx
    Commented Aug 14, 2020 at 9:49

Bottom Line: If the S&P is at 1600 now, then one year from now it is at 1600 still, you will be way down. Leveraged ETF's suffer from decay, so you have two weights pulling it down and only 1 pushing it up. Also, if you are looking at an ETF that uses futures, the you Must understand futures, contango, and backwardation. You're swimming with sharks don't forget.

  • 2
    Can you elaborate as to why you have still suffered a loss, outside of inflation, if you have not exited your long position and the level of the security remains unchanged (or has re-attained the level it held a year ago)? Commented May 20, 2013 at 23:56
  • 2
    Just overlay the SPY ticker with the UPRO ticker on a chart and its clear that over long periods when the price of SPY returns to its past previous price, there is significant decay in the UPRO ETF. Commented May 21, 2013 at 10:32

If your strategy is buy and hold, you will undoubtedly do worse in comparison to 1x over the long run with a 3x etf, because normally 3x leverage is too much leverage. BUT, if your plan is to adjust leverage so as to achieve your target volatility, now you have a strategy that traditionally beats straight spy buy and hold. So let's say you decide you want to allocate 3000 towards spy. Let's say that you decide you want this part of your portfolio to contribute 10% overall volatility. Unlike returns, if you look at say a 3 month or 6 month period, you will see that the volatility of the past does a pretty good job of predicting the volatility of the future.

So strategy A is buy 3000 worth of spy and hold. Not bad. In strategy B, you would adjust your exposure to reflect the volatility. So if I see that the volatility is right at 10%, I will be 1000 upro and 2000 cash. When volatility increases I will reduce my exposure, so that I will actually have reduced my risk in comparison with the buy and hold. So if vol went to 15% I further reduce my exposure to keep the volatility where I want it, which would be 1000 * (10/15) or about 666 dollars. Most of the time, I will be in more than 1000 but usually less than 2000 and almost never will i be in the full 3000.

It's basically a poor man's leverage play. Not as efficient as other instruments, but, if you are not allowed leverage and can't afford futures, it is a good way to go. The talk about overnight rebalancing assumes that the rebalancing would always go against the 3x ETF. It seems to me that that can go both ways where sometimes the rebalance works out in your favor. You definitely should NOT buy and hold a 3x ETF, but if you implement this "risk parity" strategy, you should do quite well. Back test over the past 20 years yields a sharpe ratio 2-3 times better than buy and hold of SPY. If you implemented my strategy with the Nikei, you would have lost less than straight 1x Nikei. You have to consider not returns but risk adjusted returns. So the important thing is to adjust for that risk!


It's risky to buy UPPRO because the market is not always going to be a bull market and will from time to time go down or sideways. Leveraged ETFs will then lose value relative to the index, as pointed out in the other answers.

A better method is to short the inverse leveraged ETF and extract a profit out of the leveraged ETF losing value. A problem with this approach is then that a short position will usually incur fees while buying the ETF is free of charge. Shorting an inverse leveraged ETF for a more volatile index such as the NASDAQ is then more likely to work well on the long term.

If you compare shorting SQQQ with buying TQQQ, then what you see is that you get a larger profit when the index does not perform well. You can then get a profit even when the index decreases. But the price paid for this advantage is then that you'll have worse performance if the index rises steadily.

It is, however, possible to improve the results of shorting SQQQ by shorting this with the highest available leverage. If the leverage is L and the spread is u times the value of the SQQQ and we assume that u is kept constant, then the optimal strategy for shorting is to close and reopen the position every time the sell value of SQQQ increases by a factor of 1 - v, where v is given by a series expansion in powers of sqrt(u) as:

v = sqrt[2 u/(L+1)] + (4 L -1)u/[3 (L + 1)] + terms of order u^(3/2) and higher

Note that v is positive so the increase by a factor of 1 - v corresponds to a decrease in the value of SQQQ.

So, there are then two compoundings going on here. One is the daily compounding due to the rebalancing of SQQQ which over time leads to a loss as it's not optimized for maximum gain. We turn that into a profit by shorting it. And we do our own compounding by closing and reopening the short position in a way that's optimized for maximum profit.


YOU ARE NOT CRAZY. This is the RIGHT thing to do over the long term. Will it track the index exactly 3x? NO! Yes, we all get the daily rebalance and the 95bps in annual fees. However, if the trajectory of the S&P 500 over decades is up, you will come out greater than if you owned a vanilla S&P ETF. Yes, no question - we hit a recession, depression, etc....your investment will become crumbs, but again....the S&P will come out higher over time and you will rebound. The volatility will be very high, but this is the right thing to do. Cycle peak to cycle peak you will outgain massively, cycle peak to cycle trough will be extremely painful....you need to just keep contributions consistent every month and tuck it in the drawer. In 30 years....no brainer. If you believe the market will be flat over 30 years, you will get smoked, but we all know that's not true. Again, you may not get 33% CAGR....but it will be greater than the market.

  • 4
    Do you have any simulation data that demonstrates this would have been the case historically, i.e. that the intrinsic losses are outweighed by the likely upside? Commented Jul 2, 2017 at 8:56

You must log in to answer this question.