2

I read the answers from the following two posts describing why daily leveraged ETFs are a bad idea to hold in the long term:

Investing in a leveraged index ETF for retirement. Risky?

Why are daily rebalanced inverse/leveraged ETFs bad for long term investing?

I also get that, if you are investing in a 3x daily ETF tracking the S&P500, say, and that index drops by 33.3% in one day, you lose everything. As you get closer to that limit, the rebalancing costs on days of losses increase to the value of your entire portfolio at that extreme.

However, is it really realistic that such an index might drop by 33.3% in one day? (As an aside, what happens if it drops by more than that -- would the leveraged ETF's price go negative or just stay at 0?)

Furthermore, while the critics of long-term holding of leveraged ETFs are right to point out rebalancing costs, in practice the CAGR of a simulated 3x daily leveraged stock has been around 11.56% since 1993 (including both major recessions) whereas SPY has been around 7.31%. You don't get more long term than that.

https://repl.it/@mureytasroc/SPY

So I get why you wouldn't want to invest your retirement portfolio in a 3x leveraged ETF, especially if your time horizon is short. However, if you have a certain amount of money that you could easily live without, and have a long time horizon, is keeping that money in a 3x leveraged ETF really that bad (or is it good)?

As a random interesting fact, at least in the above specified time frame, my simulations show that a 3x leverage yields the best returns out of all integer leverage multipliers (slightly more than 2x).

Murey Tasroc
  • 123
  • 4

2 Answers2

4

I've read many of these articles such as those in your links and they are right to a point. But as you noted, many leveraged ETFs have indeed provided the leverage that they advertised and then some. So who's wrong?

A small factor in this is the higher Expense Ratio of leveraged funds. They have to rebalance daily. This takes its toll under all price scenarios.

Many cite the example of Beta Slippage where for example, an ETF moves up X percent one day and then drops the same amount the next day. The result is a loss, even though the underlying recovered the same percentage. If this were to occur over a long period of time, the leveraged loss would be significant. Volatility is the enemy of leveraged ETFs. IOW, the more volatile the ETF, the more it loses over time.

Add these two factors together (Expense Ratio and Beta Slippage) and you'll see a lot of leveraged products that underperform their expectation.

But here's the other side of the story. If the underlying is trending, Beta Slippage is less of a factor, possibly very little at all. In some extended periods, the leveraged ETF not only met the 2X or 3X benchmark but it outperformed the expected result as well.

Given the upward bias of the markets over the long term (7.31% since 1993), this explains your 3X return of 11.56%. That's about 1.5X which is far short of 3X. In many periods during that 26 year period, the 3x yielded better than 3X. So timing does matter but unfortunately, there's no way to time the timing :->)

Bob Baerker
  • 76,304
  • 15
  • 99
  • 175
3

A day-to-day drop of 33% is rare indeed -- but even much smaller drops can sap the value out of a leveraged fund.

How about a 3% day-to-day drop instead? The S&P 500 has done that twice this month.

To see the effect of that, consider a hypothetical fund with a capital of $1,000,000 on Monday morning. It's 3× leveraged with respect to an index that drops from 100 to 97 during Monday. This causes the fund to drop in value by 9% to to $910,000.

On Tuesday, yesterday's scare turns out to be due to false rumors. The index rises back from 97 to 100, a 3.093% increase. Our fund's value accordingly grows by 9.279% from $910,000 to $994,439.

Note well that the market has recovered fully, but the hiccup eradicated 0.5% of the fund's value. And it works out exactly the same way if the market had climbed from 97 to 100 one day and fallen back the next -- with day-to-day leveraging you lose money either way.

If jitters of this size happen just once or twice per month, the cumulative effect of such short-term fluctuations will eat up any legitimate return the market would give you.

hmakholm left over Monica
  • 3,370
  • 1
  • 22
  • 24
  • You could make the same argument to say that days of sharp increase have an extra return... and what about the fact that a 3x leveraged etf tracking spy has done better than spy according to simulations? – Murey Tasroc Aug 17 '19 at 03:54
  • 1
    @MureyTasroc: No -- days of sharp increases that fall back the next day lose money too. The product of (1-3×0.03) by (1+3×0.03093) is less than 1 no matter which order you multiply them in -- even though (1-0.03)×(1+0.03093)=1. – hmakholm left over Monica Aug 17 '19 at 04:00
  • I meant days of consecutive increase – Murey Tasroc Aug 17 '19 at 04:02
  • And the historical data? That is at the heart of my question – Murey Tasroc Aug 17 '19 at 04:03
  • 1
    "... with day-to-day leveraging you lose money either way.... If jitters of this size happen just once or twice per month, the cumulative effect of such short-term fluctuations will eat up any legitimate return the market would give you." If that was the case then how do you explain that while the SPY is up 16.65% this year, the SSO (2x) is up 1.87x and the SPXL (3x) is up 2.82x ? Your conclusion is invalid. – Bob Baerker Aug 17 '19 at 13:18
  • @BobBaerker: perhaps they are not following the exact "triple the percentage movement each day" goals that is being assumed here. – hmakholm left over Monica Aug 17 '19 at 13:22
  • I didn't ask you why these leveraged funds failed to meet the entire 2x or 3x expectation. I asked why (as expected) they generated multiples of the SPY's return (1.87x and 2.82x) when your conclusion is that when the market bounces up and down, "it will eat up any legitimate return the market would give you." Your conclusion is wrong. – Bob Baerker Aug 17 '19 at 13:36
  • @Murey Tasroc - Historical data? Every month you're going to have some number of pairings of back to back up and down or down and up percentage moves. If the trend is up for the month, a long leveraged fund's gains will exceed the beta slippage. For example, in January of 2019, while the SPY returned 8%, the SSO (2x) returned 15.73% and the SPXL (3x) returned 24.01%. And yet there were 4 pairs of days when the SPY bounced and then dropped a similar percentage. This idea that beta slippage eradicates all gains is incorrect. – Bob Baerker Aug 17 '19 at 13:50
  • @BobBaerker: The only mathematically possible way they could get as high as you say is by not following a strict "thrice the percentage difference each day" history. Since log(2.82) is larger than three times log(1.1665), there must have been at least one day (and more likely many) when the funds grew by more than 3x the percentage increase of the index. – hmakholm left over Monica Aug 17 '19 at 13:55
  • @Henning Makholm - I'm not going to try to understand how you conclude that by comparing log of two numbers means that there had to be "many days when the funds grew by more than 3x the percent increase of the index". Both funds (2x and 3x) under performed in the time period mentioned (1.87x and 2.82x). The beta slippage could have been very small. These have different management, different leverage (2x vs 3x), different expense ratios, and possibly different leverage methods. You're just making assumptions with no basis in fact. You can verify your 'theory' by actually examining the data. – Bob Baerker Aug 22 '19 at 23:16