A friend of mine has a knack of homing in on financial folly like a dowsing rod waved above the Atlantic finds water.
This week he proudly informed me that he’d bought a short ETF.
He hasn’t done so because he’s got too much money and thought the financial services industry deserved a donation, via yet another poorly understood product.
No, he believed that holding a short ETF would help protect his portfolio if we are hit by further bear market declines.
It might sound good in theory – although I’d argue it’s better to think about portfolio protection before the market has already slumped 20% and the crash is mainstream news, and that the best way to reduce your risk is simply to sell – but anyway, the maths of how a short ETF works is rather different to what my friend imagined.
If my friend holds the short ETF the way he planned to, he could lose money even if the market does fall.
Insurance that costs you money when you expect it to pay out is surely even worse than having no insurance!
In this post I’ll explain why my friend was wrong to buy a short ETF to try to insulate his portfolio from a prolonged downturn in the stock market. (Next week we’ll consider what he might have done instead).
What is a short ETF?
A short ETF (also called an inverse ETF) is an Exchange Traded Fund that delivers the opposite of the daily return from its underlying benchmark – often an existing, conventional ETF.
For example, the db X-Tracker FTSE 100 Short Daily ETF from Deutsche Bank (factsheet) delivers the opposite of the return you’d get from the bank’s normal FTSE 100 ETF tracker.
- If the FTSE 100 falls 1% in a day, the short ETF will rise 1%
- If the FTSE 100 rises 1% in a day, the short ETF will fall 1%
You can also get leveraged short ETFs, which usually have 2x or 3x in the title. As the name implies, these ETFs return two or three times the opposite of the daily return of their benchmark.
Short ETFs are synthetic ETFs with the usual risks you’d expect from such securities, such as counterparty risk.
But there’s a more fundamental problem with short ETFs – at least in the way many uninformed buyers aim to use them.
The snag: The only way is not up
The problem arises due to the way that the mathematics of compounding works.
Maths will probably not be on your side with a short ETF should you hold it for more than a day or two.
Over one day, short ETFs do what they say they’ll do on the tin (not that every buyer reads the label). They deliver the opposite of the benchmark’s return.
The trouble comes if you hold a short ETF for more than one day, let alone the weeks or months my friend had planned in order to try to offset any losses from a falling stock market.
Because of the impact of compounding, longer-term returns will be more or less than you’d expect from simply summing the inverse of the daily returns.
The impact is especially noticeable in volatile market conditions. Which is to say most market conditions – shares rarely go up or down in a straight line for long.
Examples of how short ETFs work in practice
This is all pretty counter-intuitive, so let’s illustrate it with a couple of examples.
Let’s say we’re bearish about the FTSE 100, because it’s just broken through the 10,000 level and we think that’s quite enough for now.
We decide to buy £10,000 worth of a short ETF that delivers the inverse of the FTSE 100.
In other words, if the FTSE 100 falls 1% in a day, we’ll gain 1% on our £10,000, and vice versa.
In addition, my friend – let’s call him Harry – decides to go one better, buying a 2x short ETF with his £10,000.
If the market falls 1% in a day, for example, Harry expects to gain 2%.
Example 1: A declining market
We buy our ETFs on Monday. Let’s say we’re even luckier than we deserve given we’re punting on a single day’s return from the FTSE. The market falls 2%.
Here are the returns1:
|
Index |
Index Change |
Short ETF |
2x Short ETF |
|
10000 |
|
£10,000 |
£10,000 |
|
9800 |
-2.00% |
£10,200 |
£10,400 |
So the index fell 2% and we made £200 on our £10,000 investment. Harry has made £400.
All as expected so far.
Now we get greedy and think the market will keep falling. (Why are we so confident? Perhaps we heard it on CNBC. Those guys are never wrong…)
Let’s say the market falls 2% a day every day for the rest of the week. That’s five days of 2% declines.
Think we’ll be up 10% at the end of the week? Think again!
Here’s how our returns stand by the end of play Friday:
|
Index |
Index Change |
Short ETF |
2x Short ETF |
|
10000 |
|
£10,000 |
£10,000 |
|
9800 |
-2.00% |
£10,200 |
£10,400 |
|
9604 |
-2.00% |
£10,404 |
£10,816 |
|
9412 |
-2.00% |
£10,612 |
£11,249 |
|
9224 |
-2.00% |
£10,824 |
£11,699 |
|
9039 |
-2.00% |
£11,041 |
£12,167 |
|
-9.61% |
10.41% |
21.67% |
As you can see, the way the compounding works, the FTSE 100 index has declined by a little less than the 10% you might have expected.
Similarly, the short ETFs are both more than 10% ahead. Already we can see the returns aren’t the simple opposite of the underlying index, due to compounding.
Incidentally if the market had actually risen by 2% a day – thwarting our purchase of the short ETF – then this relative performance would have been reversed.
Example 2: A volatile, declining market
Sometimes markets go in one direction for a week or so, but in practice they’re usually more volatile. Two steps forward and one step back is more common, whatever the overall direction of travel.
Here’s how our two short ETFs perform in a bumpier market:
|
Index |
Index Change |
Short ETF |
2x Short ETF |
|
10000 |
|
£10,000 |
£10,000 |
|
9500 |
-5.0% |
£10,500 |
£11,000 |
|
9975 |
5.0% |
£9,975 |
£9,900 |
|
9476 |
-5.0% |
£10,474 |
£10,890 |
|
9950 |
5.0% |
£9,950 |
£9,801 |
|
9453 |
-5.0% |
£10,448 |
£10,781 |
|
-5.47% |
4.48% |
7.81% |
In this example, the index has declined by 5.47%, but the short ETF has only climbed by 4.48%. Meanwhile the 2x short ETF has grown far less than the near-11% gain you might have expected from simply doubling the index’s decline.
All sorts of different outcomes are possible depending on exactly what numbers you plug in. Try creating a spreadsheet and having a play.
Example 3: A volatile yet flat market
Perhaps the most shocking example of how you can come unstuck with a short ETF is when the market is flat over the period, yet you still lose money.
Take a look at this sequence of returns:
|
Index |
Index Change |
Short ETF |
2x Short ETF |
|
10000 |
|
£10,000 |
£10,000 |
|
9000 |
-10.0% |
£11,000 |
£12,000 |
|
9450 |
5.0% |
£10,450 |
£10,800 |
|
9923 |
5.0% |
£9,928 |
£9,720 |
|
9625 |
-3.0% |
£10,225 |
£10,303 |
|
10000 |
3.9% |
£9,827 |
£9,500 |
|
0.00% |
-1.73% |
-5.00% |
Here we see that after five very volatile days, the index has ended up exactly where it began.
In contrast, we’ve lost money on the short ETF and Harry has pretty much taken a bath on his 2x leveraged one!
This isn’t because these ETF products don’t work. They are doing what they are meant to do, which is deliver the opposite of the daily return, with knobs on in the case of the 2x short ETF.
It’s all down to the mathematics of compounding.
Short ETFs deviate in the real world, too
This is not a case of odd examples chosen to prove a point. Such behaviour from short ETFs happens very often in the real world, too.
A striking example cited by the FT a while ago involved a double leveraged short ETF based on the FTSE Xinhua China 25 index.
After the Beijing Olympics, this index fell 34% in four months. Yet rather than making a 68% gain, the short ETF delivered a 56% loss!
More mundanely, you may simply find that your long-term holding of a short ETF does go in the right direction, but for the reasons I’ve cited it turns out to not be enough to offset the declines you’re trying to guard against.
For instance, between 1 April 2015 and now the FTSE 100 has declined by nearly 12%, but the DB X-Trackers DBX FTSE 100 Daily Short ETF (Ticker: XUKS) has only risen by 7%.
So much for protecting your portfolio.
Again, I stress that’s not to say these products are doing anything wrong – rather they are doing what they are meant to do, which is offering inverse correlation to the index on a daily basis, in this case to the FTSE 100.
But if you buy and hold them as longer-term portfolio protectors, then you are doing something wrong, though exactly what result you’ll get – good or ill – will depend on the lottery of volatility over the period.
Note too that there may be other reasons for deviations with short ETFs, involving costs, taxes, dividends, and also financing charges in the case of leveraged ETFs.
The compounding effect is the most important one though, and reason enough for everyone but day traders to avoid these ETFs entirely!
(And you already know that you should avoid being a day trader…)
Further reading on short ETFs:
- ETF Securities has a comprehensive PDF about short and leveraged ETFs.
- The US Securities and Exchange Commission has issued a warning about short ETFs.
- Here’s some better ways to protect your portfolio.
- For simplicity, I’m obviously ignoring the TER cost of the ETFs in all these examples. [↩]
Great example and lesson on compounding. Personally, I do like Inverse ETFs because they are easy to trade. But your important point should be a lesson that Inverse ETFs may have to rebalanced constantly if you hold them longer than 1 day.
Thanks for a very useful and informative article.
Are these leveraged and short ETFs actually designed for anything other than a quick punt for those too lazy to walk down to the bookmakers?
Very interesting.
I wonder if one approach for your nervous friend would be to allocate say 10% of his portfolio to such a short, but to ask his broker/platform to trigger the purchase when the FTSE has dropped to a certain price (almost like a spreadbet). At the time of writing the FTSE100 is 5827, so the trigger could be a price of 5700 or whatever number scares your friend. I think one problem is working out the relationship between the FTSE’s value and the short ETF’s price.
Is there an issue that many (most?) ETFs only have their price set once per day? If true, this will introduce a huge tracking error to the underlying index during periods of high volatility.
Even then what would be the benefit if the rest of the portfolio (if mainly UK equities?) continues to lose value in the short term (if that’s how your friend views it). He’d have a short ETF “making money” (perhaps) and the majority of his portfolio “losing money”.
Personally, I’d follow your normal advice and I’d keep the money allocated for the short ETF as cash so that when the FTSE dropped I’d have the funds to take advantage of the lower prices to buy.
p.s. I am a fan of ETFs: just the boring ones.
1. A friend, you say? I believe you…
2. Seriously now, good work, as ever.
3. These things are potentially lethal and so easy to misunderstand, as you say.
4. They do a great disservice to the ETF cause – and suit the purpose of certain opponents of passive investing, too. I’m referring to those people who deliberately misrepresent ETFs as if they all are as potentially dangerous as this particular type. Much better to get an ‘expert’ fund manager pick stocks for you, they say…
@Tony – Cheers for your thoughts. As I understand it, conventional indices are constructed to take into account the compounding of their underlying holdings, so that this sort of error does not emerge. Regarding my friend, yes, the theory was (and seems quite common, from a quick Google) that the short ETF would go up in value as his portfolio fell over a few weeks. The trouble is the daily compounding means a different kind of bet is being undertaken. I agree that a plunge protection fund makes far more sense if you’re a dabbler, but most people who are draw to active strategies, even semi-responsible semi-active ones like me, find it hard to sit in cash and wait, especially at today’s rates. (Doesn’t mean it’s not right to do that, just saying I think people feel the need to ‘do something’ and feel the Short ETF is something).
@Mattster — Thanks. They are basically designed for day traders, and to be fair to the provider the ProFunds PDF I link to at the bottom does make that clear to their customers, though I don’t suppose many of them read the literature. Most of us have no reason to go near them in my view.
@Alex – Hah, it wasn’t me guv. 🙂 When it comes to the active part of my portfolio (boo! hiss!) I just sell up if I’m uncomfortable. But that’s for another blog/day. Thanks for your comments, and I agree that the profusion of ETFs is bad for the passive cause.
@Ken — Oops, missed you — apologies! Thanks, and for commenting, too.
I wrote an article titled “Are Derivative-Based ETFs Sowing The Seeds Of The Next Financial Crisis?” for Seeking Alpha a few months ago. I concluded that ETF’s don’t do what it says on the tin (mimic the underlying asset) and that they are slowly mutating into more complex financial instruments like collateralized debt obligations, which I drew disturbing parallels with the subprime mortgage crisis. It would be therefore foolish for any retail investors to see them as a panacea to gaining exposure to virtually any asset.
Excellent post. I had toyed with buying some inverts or the VIX but certainly now I will steer clear of these ETFs. It’s entirely logical as an outcome but needs clear illustrations to see the workings. Thank you.
Short ETFs is something I have looked at, having had it recommended by a MoneyWeek contributor.
If you look at ETFs Double Short FTSE
http://www.etfsecurities.com/fund/lev/etfs_fund_ftse_100_le_en.asp
and click on the Factsheet. As I understand it, the fund tries to track the FTSE100 Leveraged Index. You will see Index Historical Performance which compares the (Short) Leveraged Index with the FTSE.
2011 2010 2009 2008
FTSE100 Levgd. Index 13.3% -28.4% -48.0% 32.2%
FTSE 100 -13.1% 12.6% 22.1% -31.3%
It appears that since 2008, if the FTSE falls the Index makes a profit it is approximately 100% of the FTSE drop, but
when the FTSE rises, the Index loss is approximately 200% of the FTSE rise.
That doesn’t seem very good odds to me.
Or have I missed something?
You haven’t missed anything, Richard. As demonstrated by my maths examples above, the performance of a short ETF held over a period of time is a reflection of volatility (and pretty random) not a predictable return. Leverage makes matters worse.
Mattster wrote “Are these leveraged and short ETFs actually designed for anything other than a quick punt”
Compare 50% LUK2 (2x FT100), 50% IGLT with 100% ISF and you’ll see that half in the 2x compares relatively closely with 100% in the 1x over time (the 50% in cash (which I’ve assumed were invested in IGLT here) generally covers the ‘cost of borrowing’/time-value that the leveraged fund incurs to maintain the 2x scaling).
50% loading into a 2x and 50% ‘cash’, can replicate the underlying (100% in 1x) reasonably closely – even over extended periods of time. Whilst the expense ratio is generally higher for 2x funds (0.5% for LUK2 for instance), holding half in that in effect reduces the cost down to 0.25%, which isn’t dissimilar to what 100% the 1x might cost.
Rebalancing back to 50-50 2x and cash daily will have provide the closest tracking, but the costs will kill. You can rebalance much less frequently, perhaps at 40% bands (when the weights have declined below 30%/risen above 70% for instance) and achieve similar general reward as the 1x with much less rebalance frequency (once every year or two perhaps), but likely with some tracking error – that has 50-50 probability of being better or worse.
This image shows an example of 50% in a US 2x SPY (SSO), 50% in SHY (short term treasury (cash)) compared to the SPY (S&P500) for year-to-date (no rebalancing as I don’t have a ETFReplay.com account and the free version limits you to ‘buy-and-hold’ only).
http://www.jfholdings.pwp.blueyonder.co.uk/ssoshy_vs_spy.png
Hi everyone — I’m hearing lots more about short ETFs again this year with the market declines, including a few comments on Monevator, so here’s an updated post from 2013 as a reminder that they don’t work the way most casual investors think they do.
Anyone overwhelmed by the detailed advice should just consider the fact that % changes mislead anyway, as a 10% fall followed by a 10% rise does not get you back where you started
90/100 * 110/100 = 0.99
So be very wary of adding a series of falls and rises as merely the sum of their percentages
Blimey. That makes spreadbetting against the index look almost honest 😉 None of that daily compounding and you can do it for three months hence. Costs are a bit ugly, but at least comprehensible in advance.
Or you could build up your portfolio while Mr Market is in a funk, and look back at the value you got in 10 years time!
Very timely, thanks. And trust Monevator to have warned of this ages ago. I too have a friend who buys these but as day trades (naughty, I know). But when we met up in the pub the other night after work he seemed very pleased with himself and his returns, though he sticks to bank stocks (I know..) Having said that, bank stocks for the next 6-12 months seem quite the trend amongst bankers now, at least in the States..
Thanks for the interesting article. Leaves me wondering how short ETFs compare with buying a put option on the index..
Either way “your friend” might be better to simply reduce his position if it’s so large that he can’t sit it out ..
@PC — I love that people keep putting “my friend” in ironic quotes… 😉 What is more incredible/unbelievable — that someone who has written an investment blog for eight years suddenly buys a short ETF without knowing how it works and then compounds his embarrassment by writing a blog about it and THEN republishes it a few years later to relive his embarrassment, or a person who writes an investment blog for 8 years actually *having* a friend? 🙂
Or that I have friends who do silly things with money? (Believe me, I have them…)
Hi, I think it would be helpful to also illustrate/explain for why this is not a problem for long etf index positions. Are there any exceptions? E.G 2x long?
“Short ETFs are synthetic ETFs with the usual risks you’d expect from such securities, such as counterparty risk.”
Ach, that reminds me of when AIG was in the mire, my three ETFs in commodities were all synthetic and all went to zero overnight. That was two weeks of nervousness I’d prefer not to repeat!
We’re a wee while on now from this article, and I would say that things have moved on a little now. It is now possible to buy an “Infinite Turbo” from SG which gives you the straightforward return of a short or long position on say, the FTSE 100, with the leverage of your choice. Each ETF has its own “knockout” where you lose all your capital if your trade goes significantly against you, but you can’t lose more than your capital stake. Could be a useful tool for those with strong hearts wanting to eliminate the random number generating mechanism of daily products.