The Best 3x Leveraged ETF

by: Greg Hudson

If you had to pick one 3x leveraged ETF to hold over 30 years, which would it be?

A 30+ year historical simulation of 10 3x leveraged ETFs reveals winners and losers.

You probably shouldn't pick just one: a balanced portfolio of leveraged ETFs would have outperformed on a risk-adjusted basis.

In my last article, I explained how I built simulations of the 30+ year performance (net of fees) of leveraged ETFs (UPRO) and (TMF). Today, I'll open the aperture to a selection of ten 3x leveraged ETFs and see who would have been the winner.

First, let's introduce the challengers. Below, you'll see the 3x leveraged ETFs I'm simulating, the indices whose daily returns these ETFs seek to triple, and the proxies I used to create these simulations:

ETF Index My Proxy
(CURE) Health Care Select Sector Index (FSPHX)
(DFEN) Dow Jones US Select Aerospace & Defense Index (FSDAX)
(FAS) Russell 1000 Financial Services Index (FIDSX)
(TECL) Technology Select Sector Index (FSPTX)
(TMF) ICE Treasury 20+ Year Bond Index (VUSTX)
(TNA) Russell 2000 Index (CSMIX)
(TPOR) Dow Jones Transportation Average (FSRFX)
(UPRO) S&P 500 (VFINX)
(UTSL) Utilities Select Sector Index (FKUTX)

This is by no means a comprehensive list of 3x leveraged ETFs on the market, but it's a reasonable cross-section that covers a few broader indices, defensive sector-focused indices whose lower volatility may stand up better to increased leverage, cyclical sectors that have been on a tear lately, and a leveraged bond ETF for comparison.

Most importantly, I could find proxies with daily performance data going back to 1986 for all 10 of them. None of these proxies are going to be perfect - and many are quite imperfect. In most cases, I had to base my simulation on actively managed sector-focused mutual funds, which will at best roughly mirror the daily index changes that these ETFs are intended to triple. That being said, unless the proxy significantly outperformed the actual index over 30 years, these simulations should generally be a bit on the conservative side due to 1) effectively tripling the proxy's expense ratio and then also adding the leveraged ETF's expense ratio, and 2) in the case of TQQQ (which has no fund proxies going back to 1986), not including dividends. See my last article for more explanation of how I built this simulation.

Keeping in mind the limited accuracy of this simulation, let's see who the winners are. Let's keep it clean! Now, come out boxing!

Ladies and gentlemen, we have a winner! Thanks to its proxy, FSPHX, largely avoiding both the dotcom bust and the financial crisis, our simulated CURE boasts an enviable 29.2% annual return over the last 30+ years and crushes the competition, besting second place DFEN by over 6% annually.

Now, let's look at their stats:

Rank Ticker CAGR Sharpe Sortino Max Drawdown Max DD Date
1 CURE 29.2% .95 1.44 79% 10/26/87
2 DFEN 23.1% .66 1.04 97% 3/9/09
3 TPOR 22.2% .64 1.02 97% 3/9/09
4 TMF 19.5% .72 1.23 49% 8/21/13
5 UPRO 18.5% .60 .92 97% 3/9/09
6 TNA 12.8% .51 .79 97% 3/9/09
7 UTSL 11.7% .65 1.02 52% 3/9/09
8 TQQQ 11.6% .50 .79 99.9% 3/9/09
9 FAS 8.0% .46 .74 99.7% 3/9/09
10 TECL 6.9% .52 .83 99.9% 11/20/08

Notably, our simulated CURE also has the highest Sharpe and Sortino, followed by TMF.

But just look at those drawdowns! This is why the experts tell us that leveraged ETFs are a terrible long-term investment. With the exception of TMF and UTSL, you'd have been wiped out during bear markets (most notably, the financial crisis). And the pundits are also right - most of these long-term annual returns are nowhere near 3x the long-term annual return of the unleveraged proxy, and some are even significantly worse. TECL's proxy, FSPTX, would have given you annual returns of 11.5% over the last 30+ years, whereas our simulated TECL would have gotten you only 6.9%, with three times the risk!

Who in their right mind would invest in these things over the long term? Well, me. Am I out of my mind? I don't think so, because as I mentioned in this article, leveraged ETFs would have performed admirably over this period if they were just part of a diversified, regularly rebalanced portfolio.

Just compare the above single 3x ETF performance stats to a simulated portfolio of 35% UPRO and 65% TMF, which I proposed in my last article. Over the same period (rebalanced 8 times annually with a $10 rebalancing fee like that charged by Motif), this simulated balanced portfolio would have produced the performance stats below:

CAGR Sharpe Sortino Max Drawdown Max DD Date
25.4% 1.04 1.74 46% 3/9/09

CURE may have been the winner of our single 3x ETF challenge, but this balanced portfolio of UPRO and TMF clobbers it on a risk-adjusted basis (Sharpe, Sortino, and max drawdown). Even though UPRO and TMF were only our fourth and fifth place finishers on their own, the magic of diversification and rebalancing makes the combination of the two a much better risk-adjusted long-term play than our champion.

I would not be happy if your conclusion from reading this article is that you should put everything in CURE. It performed quite well in our simulation, but:

  1. That simulation was based on an actively managed fund that may have significantly outperformed the actual index on which CURE is based.
  2. It suffered a nasty 79% drawdown.
  3. Past performance is not predictive of future performance. What might happen to healthcare stocks if a Democrat becomes President and moves us to single payer? I don't know, and I doubt you do either.

Ultimately, there isn't any one 3x leveraged ETF I'd recommend holding over the long term. You really probably shouldn't unless you love gambling or have an infinite time horizon. But, as I'll continue to explore in further posts, a well diversified, balanced portfolio of multiple leveraged ETFs might be a perfectly sane part of a long-term investment strategy.

Disclosure: I am/we are long UPRO, TMF, TQQQ, TECL, TPOR, UTSL, DFEN, FAS, CURE.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: Past performance is not predictive of future performance. My simulations are based on data from Yahoo! Finance, which may contain errors.