Disclaimer: I am not a financial adviser and you need to do your own research.
I won't repeat this good article on investing for effective altruists: https://80000hours.org/2015/10/common-investing-mistakes-in-the-effective-altruism-community/ . But one of the key ideas is if you are investing for yourself and you are young, or if you are investing for charity, temporary falls in value should not worry you. Taken even further, if there is a big payoff, but some chance of losing all your money, this can still be a good idea. This means that you are risk neutral, and just care about the expected (probability weighted) outcome. If you are like this, then I have an investment opportunity you may be interested in.
It is three times leveraged Russia. The Russian stock market has fallen precipitously, and it has fallen even more so from the US perspective because of currency changes. If it recovers to fair market value (mean reversion) over seven years, the value of your investment would be about four times as much. However, with three times leverage, even taking into account that you need to pay interest on the implicit loan, one would expect more like 50 times the money. Now there are impacts of volatility that I do not fully understand that could reduce this. But I have seen a 3X leverage produce 15 times as much money as the initial investment in 5 years and it was not quite as good an opportunity as Russia.
The downside of Russia is not just the normal potential for temporary loss of money. It is also possible that there could be a forced sell out of foreign assets, freezing of assets, or even seizing of assets, depending on international relations. So this is definitely not for the faint of heart.
But if you are risk neutral like I am, you may want to give it a try. The ticker is RUSL. I just put a significant fraction of my portfolio into it today.
Also, one could even argue that helping out Russia when it is worst off could prevent Russia from doing desperate things, so there may be global catastrophic risk benefits.
Update: March 17, 2016:
If you decided to invest when I did, congratulations - you have made 80% on your money in less than two months! This now means that the expected long-term return is significantly lower. Then when you include the volatility drag (see discussion below), it is not clear that this leveraged investment is better than just unleveraged Russia. So it might make sense to start phasing out of leveraged Russia and buying ERUS, especially if it is a large percent of your portfolio. On the other hand, if you think that oil price will continue rapidly reverting towards the long-term marginal cost of production, since Russia is highly correlated with oil price, you may want to stay in leveraged Russia longer.
Your use of the phrase "fair market value" is a large red flag.
I've been speculating in stocks for 35 years. One of the hardest lessons I needed to learn was to not believe that last year's prices were fairer than today's prices.
Betting on mean reversion occasionally makes sense, but I've learned to only do it after careful analysis of the fundamentals (earnings, book value, etc).
By "fair market value," I just mean value at the long term mean valuation. Do you happen to have numbers on what percent of the time values did not revert to the mean?
TL;DR: Please please please don't do this. There are not many better ways to burn money.
There are three premises here:
I'm going to ignore 1 and focus on 2/3. RUSL and RUSS (the sister ticker; RUSS is x3 short) get their exposure by holding shares of RSX.
What would have happened historically if RSX had been overvalued and you called it right, and bought RUSS (this sister ticker to RUSL; it's triple leveraged short). What about if RSX was undervalued, and bought RUSL? If you look at raw prices, here's the last four years:
2012: RSX 27.62 -> 30.68, RUSS 34.4 -> 13.77, RUSL 34.27 -> 34.4
2013: RSX 30.68 -> 27.93, RUSS 13.77 -> 12.25, RUSL 39.5 -> 28.05
2014: RSX 27.93 -> 14.79, RUSS 12.25 -> 27.41, RUSL 28.05 -> 17
2015: RSX 14.79 -> 14.23, RUSS 27.41 -> 40.59, RUSL 17 -> 10.23
This doesn't necessarily look awful, but misses a critical factor; both RUSS and RUSL have had significant reverse splits* in the past few years. RUSS had a 1 for 4 split in 2015; that year's final prices should be thought of as 40.59 / 4 = 10.1475. RUSL had a 1 for 6 split in 2014; that year's final price should be thought of as roughly 17 / 6 = 2.8333.
It gets worse. RUSS has paid no dividends for the last few years, and RUSL paid one tiny dividend in 2014. Generally leveraged ETFs do not pay dividends. RSX, meanwhile, has paid dividends of 0.519, 0.638, 0.742 and 0.729 in 2015/14/13/12 respectively. A glance at RSX's price tells you these are not small numbers. Accounting for that the following becomes apparent:
In 2012 RSX went up. If you held it you made a tidy 13.7% return (accounting for dividends). Congratulations on your successful prediction! Unless you used that prediction to buy RUSL of course, in which case you made less than 1%.
In 2013 RSX went down. If you held it you lost 6.5%. If you saw that coming and invested in RUSS, you lost 11%.
In 2014 RSX fell massively. If you held it you lost nearly 45%. If you bought RUSS in anticipation, you do actually make money this time; 124%.
In 2015 RSX moved by less than 0.5% in either direction; down 0.28%. RUSS, however, lost 63%, more than wiping out the gains in the previous year.
Obviously if you guess wrong you do even worse. Reality says that this works really badly. If your theory and reality disagree, it's generally not reality that's wrong; this goes double if you aren't an expert in the industry.
*A reverse split of, e.g. 1 for 4 means that if you owned 400 shares of RUSL the day before the split, you own 100 shares the day after. In the US this is normally used to get prices back into the 10 - 100 range that US exchanges generally prefer their stocks to trade in.
Thanks for the good examples of volatility drag. This is very important for something as volatile as Russia. However, there is a tendency when valuations are very far from the mean to have large runs towards the mean. One example of a consistent run was from 1/30/2015 to 5/15/2015 in 3.5 months the base index, RSX, went up 39%. During the same time, RUSL went up 137%. Cubing would imply 167% increase, so not bad. Of course there is no guarantee that this will happen, but if it does, the investment will make a lot of money. But you are correct that if mean reversion takes a long time, leverage may be no better than the underlying index, and could even be negative returns.
Mean reversion plays generally have a timeline of years; I assumed that's what you were proposing. This should be obvious from the fact that you could have made the exact same mean reversion argument at the end of 2014 (2014's massive drops from the Ukraine crisis were when Russia arguably became undervalued according to your metrics), but actually RSX barely moved in 2015 as a whole. Sure if you cherry-pick the high in May 2015 you do well; hindsight-based cherry-picking always does well.
Over very short time frames (like <6 months), RUSS/RUSL are probably fine because there's limited opportunity for the transaction costs to build up. That is in fact what they are intended for; short-term bets. But if you're investing on the timeline of years, they're really bad, which is hopefully amply demonstrated by the (non-cherry-picked) data I gave for the last few calendar years.
Indeed, I did invest in non-leveraged Russia in 2014 because it was undervalued. But only now that it is extremely undervalued am I willing to make the bet on rapid upward movement where it makes sense to have leverage.
It looks like RUSL rebalances daily, which could end up eating a lot of your returns. I don't know much about this but I'd expect buying on margin with Interactive Brokers to be a better bet.
(See Brian Tomasik's analysis of leveraged investing.)
However, with three times leverage, even taking into account that you need to pay interest on the implicit loan, one would expect more like 50 times the money.
How did you come up with this figure? If a Russia ETF is currently at $10 and you expect it to increase to $40 within seven years, a 3x leveraged investment would increase to $120 (ignoring overhead costs and volatility drag), right? That's a 12x increase, not 50x.
Thanks for the link. It sounds like Brian is saying that margin works in a way like taking out a loan. If you had $100,000 to start with and took out a loan for $200,000 and invested all $300,000, and the value went to 50%, your net worth would be $-50,000. My understanding of these leveraged funds is that you cannot go negative. You can rebalance yourself to avoid this, but that is a lot of work.
Check out http://finance.yahoo.com/echarts?s=TECL+Interactive#symbol=TECL;range=my . This is the 3X investment that produced 15 times as much money, and we got that the underlying asset had a growth ratio of roughly the cube root of 15. So basically leverage is not multiplying the returns by three, but raising to the third power. You can see this with a simple case of underlying investment return of 10% per year over 20 years give 6.7 times as much money. However, if you double the annual return to 20%, after 20 years you get 38 times your money (and really doubling your daily return would mean 21% annual). It is does not quite achieve this I believe because of the impact of volatility (Brian says alternating days of gains and losses hurt) and of course the fees and interest on the implicit loan.
If that's true then it sounds like the way ETFs implement leverage is completely different from how margin works. I don't really know how ETFs implement leverage though.
If you use margin, you can rebalance at different intervals. ETFs are roughly the same as using margin, but rebalancing daily.
Rebalancing daily doesn't really "eat up your returns" in any meaningful way. It replaces a smaller probability of losing a larger amount of money with a larger probability of losing a smaller amount of money.
Doesn't it reduce your returns though? It increases transaction costs, and it increases variance drag. If a highly-leveraged stock is more volatile then it will perform worse under daily rebalancing than a stock with equal returns but less volatility.
Under the assumption that each day's expected return is a constant that doesn't depend on the previous days, the average growth rate of your investment is basically the same whether you rebalance daily or weekly or monthly. It decreases slightly as you go to longer time scales (e.g. at 2x leverage annually you will get wiped out pretty often, so your average growth rate is negative).
(Of course if days are correlated with each other, then different rebalancing periods will have different values. E.g. if there is short-term momentum then rebalancing frequently will be good and if there is short-term reversal then it will be bad.)
If you care about expected returns rather than growth rates, then rebalancing daily is a good idea iff you are using >1x leverage. Well, hard to say what it means for it to be a "good idea," since your optimal strategy is infinitely much leverage, which yields infinite expected returns. But having Nx leverage rebalanced daily will lead to higher expected returns than rebalancing monthly.
Yes, more volatile stocks with the same expected (i.e. arithmetic average) returns have lower growth rates. That's true whether or not you use leverage though. More volatile stocks with the same growth rate will have lower growth rates when leveraged, whether you rebalance daily or annually.
Transaction costs are not zero. If you assume they are you can make an easier model but you fail to capture the reality of how RUSS/RUSL act.
Thanks, Paul. To further address Michael's question, I think the reason why leverage gives returns raised to a power, rather than multiplying by a constant factor is the rebalancing. Let's say we have $100 and we take out a $100 loan and invest those $200. This is 2x leverage. However, if the fund increases 10%, we would have $220 and still a $100 loan, which is less than 2x leverage. In order to maintain the leverage, you should take out $20 more in loan and then have $120 in loan and $240 in the market. Then you can see why as the stock goes up, a given percent increase will give you a greater dollar increase if you rebalance. The converse is true if the stock goes down. And this is how you protect yourself from going to zero (a given percent decrease in the market means a smaller dollar value loss if you reduce your loan). The problem comes if gains and losses alternate each rebalancing period. One day it goes down and you sell some, and the next day goes up and you buy some. Since you don't want to sell low and buy high, I believe this is the volatility drag.
In the case of 3X Russia, it uses derivatives (typically swaps or futures) to implement much of the leverage. Also, it reserves the right to intervene on very large daily movements, which I interpret to mean that it would not let the fund go to zero if there were a 34% daily drop.
You note it, but I don't think your calculations truly take into account the counterparty / political risk of being unable to get liquid from Russian stocks. That risk factor is built into the current low price of Russian stocks - why do you think the market is wrong? Because of the longer time horizon of young investors?
Well, I guess I just see that markets nearly always recover to the long-term mean return (maybe wars are the exception?). So I would say that people are overly optimistic when the market is high relative to that mean, and people are overly pessimistic when the market is low relative to that mean. As Warren Buffett says, "Be greedy when others are fearful, and be fearful when others are greedy." Of course we don't know how long it will take to revert to the mean, so there is always short-term risk. But I've seen data indicating that seven years is a typical time, which is not too long. My calculations just say, "if it reverts to the mean." With all the issues with leverage, let's say mean reversion means only 10 times the money. And let's say that a random long-term investment is 8% per year-that's a factor of 1.7 over seven years. If you are risk neutral, you would have to believe there is a >83% chance of losing all the money for this to not be a good investment, which I highly doubt (17% chance of making 10 times the money is worth 1.7 times the investment).
How does such a strategy work on historical markets? Do you have a principled way of identifying these below-mean stocks? Low P-E ratio? Political conflict?
I am not an expert here, but I am getting advice from an expert. I believe it is a combination of things like low P/E ratio, low price-to-book, and also whether the investment is diversified (one individual company could easily not mean revert by going bankrupt. I think the next step up it is a mutual fund in one market sector, and then one country, and then diversified across multiple countries.) I believe that political conflict would result in low valuations, so you do not need to specifically look for political conflict.
Hmm I thought widely traded markets were pretty efficient with regard to publicly available technical information like that. The thing is that if you're using simple statistics like those, why not see how this would've performed on stocks traded 20 years ago? If they are successful on recent historical data, then why not quit one's day job to raise a fund of many millions of dollars to trade on that basis?
There are people who do this, like Jeremy Grantham: https://en.wikipedia.org/wiki/Jeremy_Grantham . We get tips from him in part. The difficulty with retaining investors with this strategy is you will often miss out on big returns when markets are overvalued, like the US in the last few years.
Where has the Russian stock market declined? The MICEX index look pretty flat.
I think much of the decline is due to falling oil prices. I don't necessarily expect them to mean revert much.
The iShares Russia ETF currently has a P/E of 6, which is extaordinarily low.
EDIT: P/E is 6 according to Yahoo! Finance but 16 according to iShares (h/t Ryan Carey for pointing this out). These values are not even close so I don't know what's going on here.
We looked into this fairly deeply and found that the lower value corresponds to the harmonic weighted average: https://en.wikipedia.org/wiki/Harmonic_mean. This is the one that is relevant for mean reversion. When the valuations of the companies vary widely, the different means give very different answers. The most robust is the Shiller P/E, and for Russia, it is 5 now.
I didn't downvote, but for what it's worth I think this may be a bit niche and of tangential relevance to be a top level post on the EA Forum. The 'EA Hangout' or investment Facebook groups would be better fits - see http://wiki.effectivealtruismhub.com/index.php?title=Discussion_groups . People would likely be interested there. I hope this is helpful feedback!
This really should go in the Crazy EA Investing Ideas group... I'd sent an invite but I don't know this guy's name.
Also, whatever happened to the self help group...?
Thanks for both of your feedback. I was not aware of those groups. See https://sites.google.com/site/daviddenkenberger/home for contact info. I realize it is kind of niche, but I believe it is one of the biggest mean reversion opportunities that has ever occurred (when you include the leverage done for you) so I thought a lot of charity investors would be interested. I am hoping to increase my positive impact on the world significantly with this one move.
Oh, uh, two things. 1: the term "crazy" in the name of the group is not derogatory, it is its actual name, and it is actually a serious group. 2: I was personally interested in the self help group, and just took the opportunity to ask Tom about it, and did not mean to imply that it had anything to do with you or your post.
Sent you an email invite.
Thanks for the clarification and the invite. "Radical" might have been a better term than "crazy." :-)
I would also like to be added to the crazy EA's investing group. Could you send an invite to me on here?
I left already, there wasn't much of interest.
Can you invite me to that group too? :)
It still exists! It's linked from that wiki page - I believe the admins Evan, Leo and Jacy have to approve new people.
I guess it's a secret group? Because the link tells me there is no group. I'll shoot Jacy a message, then. Thanks.
Evan/Leo/Jacy, could you add a description of how to join to http://wiki.effectivealtruismhub.com/index.php?title=Discussion_groups ?