Virgin CEO Richard Branson famously quipped, "If you want to be a millionaire, start with a billion dollars and launch a new airline." I would like to humbly suggest a quicker way: Buy a levered ETF.
Literature on the decay of levered ETFs (LETFs) is burgeoning. I won't repeat it, but here are a few recent examples:
- Wall Street Journal: Beware 'Leveraged' ETFs
- Motley Fool: Another Sign to Avoid These Investments
- ZeroHedge: Leveraged ETFs - Why Do We Have Them?
So far, the warnings have exerted surprisingly little drag on LETF growth:
Source: Securities Litigation & Consulting Group, 2010
According to IndexUniverse, as of May 2012 the US alone had 275 leveraged and inverse ETFs with nearly $33 billion in assets. The Motley Fool estimated LETF assets at $45 bilion in January 2012. LETFs have also taken off in many international markets, most recently in Japan.
Who are all of these LETF investors? Is it just a mass of unsophisticated individuals? Are they holding these instruments for a day or less, as the prospectuses solemnly recommend? No to both. Institutions own a significant percentage of shares outstanding in these four popular LETFs (other LETFs are similar):
Shares held by Institutions
Russell 1000 Financials (RGUSFL Index)
Source: Bloomberg, compiled from 13F and Schedule-D filings, 9/4/2012
And these are not just any institutions. They are a who's who of the finance world: Goldman, UBS, Morgan Stanley, JP Morgan, Credit Suisse, Bank of America, and Knight Capital. A second group of institutional holders looks more like "investors" than "market makers," though of course the line is blurry: Stevens Capital, Susquehanna, Wellington Management, Universal Property & Casualty, etc. Most of these institutions (in both groups) have reported long LETF positions for at least the last four quarters. Daily turnover data also implies an average holding period considerably longer than one day [see table 4 p.12 of this paper]
This begs the question: Why would some of our best-paid investment professionals buy and hold instruments which - as most academics and practitioners agree - are highly likely to lose value over time? Is it possible that the "market makers," including Goldman, are creating these toxic instruments just so that some of their clients could short them, at the expense of other clients? Why, that would be outrageous and unprecedented… On the other hand, perhaps they know something we don't? If that latter explanation sounds as tempting to you as it does to me, let me share a Warren Buffett anecdote from Berkshire's 1985 annual report:
"Let me tell you the story of the oil prospector who met St. Peter at the Pearly Gates. When told his occupation, St. Peter said, "Oh, I'm really sorry. You seem to meet all the tests to get into heaven. But we've got a terrible problem. See that pen over there? That's where we keep the oil prospectors waiting to get into heaven. And it's filled - we haven't got room for even one more." The oil prospector thought for a minute and said, "Would you mind if I just said four words to those folks?" "I can't see any harm in that," said St. Pete. So the old-timer cupped his hands and yelled out, "Oil discovered in hell!" Immediately, the oil prospectors wrenched the lock off the door of the pen and out they flew, flapping their wings as hard as they could for the lower regions. "You know, that's a pretty good trick," St. Pete said. "Move in. The place is yours. You've got plenty of room." The old fellow scratched his head and said, "No. If you don't mind, I think I'll go along with the rest of 'em. There may be some truth to that rumor after all."
How good do you have to be at market timing to overcome the decay inherent in LETFs? To answer that, we must first measure the baseline, without any timing: Suppose you bought a single LETF share twelve years ago, at the beginning of January 2000 and held it through today. How much money did you make or lose? Did you do better than the underlying (unlevered) index?
This question is difficult to answer because the first batch of LETFs was introduced circa 2006, and most LETFs are considerably younger. Fortunately, many of the indices underlying LETFs have longer histories which we can use. For example, UDOW, which started trading on 2010-02-11, is supposed to deliver 3x the daily return of the Dow Industrials (INDU). We have INDU's daily percentage returns prior to 2010-02-11. We can multiply them by 3 to get UDOW's expected daily percentage returns. We can then daisy-chain UDOW's expected returns from 2000 through 2010-02-11 to arrive at UDOW's first closing price, $26.26 as adjusted for subsequent share splits. We can then work back to get UDOW's price on January 1st, 2000. In fact, we can be even more accurate: Rather than multiplying the INDU's return by 3, which is UDOW's goal, we can measure how closely UDOW came to achieving its goal by regressing its daily percentage returns against those of INDU from 2010-02-09 through the present. Here is that regression:
So the best-fit formula is:
UDOW percentage daily return ~= 2.97 * INDU percentage daily return + 0.000293
The coefficients make sense: Note the realized beta (2.97x) is slightly below the promised beta (3x) due primarily to management and trading costs, and the alpha is almost zero. In addition, the formula is highly accurate at describing UDOW's daily returns since inception, with an R^2 of over 99.6%. This result is typical of most LETFs on our shopping list.
Now that we've gotten the methodology out of the way, we can come back to the original question: What was your single LETF share worth on January, 2000?
Reading the first line of this table: Our "biggest loser" is the LETF pair SOXL/SOXS, which promises 3x and -3x daily returns on the Philadelphia Semiconductor Index (SOX Index on Bloomberg). From January 2000 through today, the underlying index lost 45% and both LETFs lost almost 100%. In fact, if you bought a single share each of SOXL and SOXS in January, 2000 you would paid just over $123k and $38 million (!) respectively. These two shares would now be worth just over $30 and $34. Generalizing this, if you bought a single share of each LETF in the table in January 2000, you would have paid a total of $53 million for a portfolio that is now worth $1,225.
Two other takeaways:
- You can make money buying and holding bullish LETFs, but generally not as much as you would buying the underlying index. As such, bullish LETFs seem to be the perfect "return-free risk" instruments: Twice the drawdown for less return.
- Long term losses for bearish LETFs are absolutely staggering, even in our "lost decade," one of the worst periods for equities in the history of US capital markets. Bearish LETFs are surely one of the most expensive ways to insure your portfolio against losses. The cure is much worse than the disease.
Clearly, no one would hold a bearish LETF for such a long time period. Everyone who buys it believes they are getting an effective hedge against (or upside from) the coming market drop, and they expect to get out before "too much" decay happens. Consider, however, the first share of SDOW ever created (-3x INDU). Since February 2010, that first SDOW share passed through an uninterrupted chain of shareholders, each one thinking they can time the market. Yet as a group, those shareholders lost 23.6% and, if SDOW had been introduced in January 2000, they would have lost over 99%. The odds are just not very good. So if you are one of those amazing few investors skilled enough to extract a positive return from such an awful instrument, please give me a call: I would like to hire you. Until then, I would prefer the other side of the trade.
Disclosure: Stepwise Capital and its affiliates have short positions in several levered ETFs mentioned in this article. 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.