I’ve long believed the most difficult part of being a Couch Potato investor is resisting temptation. Index investors are asked to be content with market returns, but they are bombarded daily by fund companies, advisors and market gurus who promise more.
Back in May 2012, I wrote about one of these enticing strategies, described in The Ivy Portfolio by Mebane Faber and Eric Richardson. The so-called Global Tactical Asset Allocation (GTAA) strategy grew out of Faber’s widely read research paper, A Quantitative Approach to Tactical Asset Allocation, first published in 2007. It begins with a diversified portfolio inspired by the Yale and Harvard endowment funds, combining traditional and alternative asset classes. The “tactical” part involves using market timing to move in and out of these asset classes based on 10-month moving averages.
Faber updated the paper in early 2013 and it now includes four full decades of data. From 1973 through 2012, the GTAA strategy shows exactly one negative year: a modest loss of –0.59% in 2008. And over those 40 years, the GTAA delivered an annualized return of 10.48% with a standard deviation of 6.99%, compared with a 9.92% return and higher volatility (10.28%) for a buy-and-hold strategy using the same five asset classes (US and foreign stocks, bonds, real estate and commodities).
The rest of the story
In my original post I described The Ivy Portfolio as “an excellent book that I would highly recommend to index investors,” and I stand by that. The authors do an admirable job of explaining the benefits of diversifying across uncorrelated asset classes, which is a core principle of the Couch Potato strategy. But in my follow-up post, I warned investors not to be tempted by the idea of avoiding drawdowns and getting higher returns: “At some point in the future we’ll have another bull market that lasts three or four years, maybe more. Timing strategies are guaranteed to underperform during prolonged bull markets (as they did throughout the 1990s and in the mid-2000s).”
We’re now well into a bull market that began in 2009, so let’s see how the GTAA has done since the book appeared. Faber’s updated paper includes data for both the GTAA and buy-and-hold portfolios from 2009 through 2012, and the timing strategy underperformed in all four years:
Buy & Hold | GTAA | |
2009 | 18.74 | 12.77 |
2010 | 14.04 | 3.65 |
2011 | 2.81 | 2.49 |
2012 | 11.57 | 1.26 |
Over this four-year period, the buy-and-hold benchmark had a four-year annualized return of 11.64%, during which $1 grew to $1.55. The GTAA had a four-year annualized return of 4.95%, during which $1 grew to just $1.21.
The real-world performance
All of these numbers are hypothetical, of course, so let’s consider how GTAA worked in the real world. In October 2010, AdvisorShares launched the Cambria Global Tactical ETF (GTAA) and hired Faber and Richardson to manage it. The ETF takes the simple five-fund strategy to a much more complex level: the fact sheet explains it’s based on “a diversified portfolio of approximately 50 to 100 ETFs.”
The fund’s performance makes it clear it bears little resemblance to the strategies laid out in Faber and Richardson’s writing. In 2011, when the five-fund GTAA portfolio returned 2.49%, the ETF lost 7.29%. The following year, when the hypothetical GTAA portfolio in the paper returned 1.26%, the ETF outperformed by five percentage points. Bottom line, if you were attracted to the strategy described in the research, start figuring out how to implement it yourself, because you won’t be able to access it through this ETF.
And how has the ETF performed overall? As of December 27, its one-year return was 1.93%, and its three-year return was 0.42% annualized. Let’s compare this to a pair of Faber and Richardson’s own buy-and-hold benchmarks. The ETF proxies for the five-fund portfolio are tracked by Faber on his website; those for the 10-fund ETF portfolio are suggested on page 191 of The Ivy Portfolio. (Returns are in USD. Three-year returns are annualized and assume no rebalancing. All data from Morningstar.)
Five-Fund Portfolio | ETF | 1 year | 3 years | ||
20% | US equities | VTI | 33.88 | 16.07 | |
20% | Foreign stocks | VEU | 14.18 | 5.49 | |
20% | 10-year government bonds | IEF | -6.30 | 4.42 | |
20% | US real estate | VNQ | 3.10 | 9.72 | |
20% | Commodities | DBC | -6.67 | -1.53 | |
Total | 7.64 | 6.83 | |||
10-Fund Portfolio | ETF | 1 year | 3 years | ||
10% | US large cap | VV | 32.10 | 15.94 | |
10% | US small cap | VB | 39.04 | 16.19 | |
10% | Foreign developed stocks | VEA | 21.39 | 8.33 | |
10% | Foreign emerging stocks | VWO | -4.60 | -2.28 | |
10% | US bonds | BND | -2.21 | 3.25 | |
10% | TIPS (inflation-protected bonds) | TIP | -9.19 | 3.47 | |
10% | U.S real estate | VNQ | 3.10 | 9.72 | |
10% | Foreign real estate | RWX | 9.06 | 10.70 | |
10% | Commodities | GSG | -0.70 | -1.10 | |
10% | Commodities | DBC | -6.67 | -1.53 | |
Total | 8.13 | 6.27 |
Over the full three years, both buy-and-hold portfolios enjoyed a cumulative gain of over 20%, while the ETF’s return was effectively zero. Every dollar invested in the fund would have grown to $1.01.
Bad timing
The point here is not to pick on one active strategy or one fund. My goal is to help investors avoid being seduced by similar promises. There will always be active strategies that show past outperformance or reduced risk based on hypothetical models, but as the Cambria Global Tactical ETF demonstrates, there are no guarantees they will work in the future. And it is extremely difficult or impossible to execute them in the real world, where there are management fees, transaction costs, taxes and highly competitive money managers ready to arbitrage away market inefficiencies.
You also need to ask yourself how patient you will be when the active strategy underperforms, as all of them will eventually. According to Faber and Richardson’s own data, the GTAA strategy unperformed the buy-and-hold model in 12 of the 17 years from 1975 through 1991. And including 2013 (using the Cambria ETF as a proxy) it has now fallen short in five straight years, and in eight of the last 11. Active investors may be attracted to low volatility, but they also want comparable performance. How many would continue with any active strategy after multiple years of dismal results like that?
Successful investing is about more than the funds in your portfolio: it also requires a change in thinking. To stick to an indexing strategy over a lifetime, you need to let go of the idea that you can improve it with some kind of active overlay. Before you think you’re adding value with market timing or picking individual securities, remember that the probability of success is low and the potential payoff is small: even the best active strategies cannot hope to achieve more than an incremental outperformance over the long term.
More important, the whole pursuit is an enormous distraction from what’s really important, which is saving regularly, diversifying widely, keeping costs low and tuning out the chatter of those who promise what they can’t deliver.
Great post!
I had actually emailed you about opinions on strategies early in the year because I was just at the point where I was going to be starting to invest.
I find the subject of market timing a little funny, because I’ve succeeded at it by accident. So far I have put money into my chosen funds twice, when I had the money and felt comfortable doing so. Both those times (April 15th and June 24th) just happened to be at local minimums for the surrounding months (especially for the TSX and international, but also the S&P500). So while the overall return on the TSX is 9.23% YTD at the moment, my personal return on my TSX fund is 14.74%. Something to note is that June 24th was an ex-dividend date, but the funds dropped by much more than the dividends that day.
Not that I’m advocating trying to time the market, just saying that it’s awesome when it happens to work out on its own. At only two data points, my own experience is purely a result of random coincidence (although maybe I should try and parley this into a successful career as an active advisor, I already have market-beating returns I can point to for the rest of my career!)
Your articles are a pleasure to read as always!
Cheers,
Dan
A great post – I really enjoy reading your blog!
One thing to be said for the timing method, is that is does allow the investor to avoid disasters.The loss it gives in 2008 is much less than a buy and hold strategy – and although the markets came back nicely since then – what if they had not??
The following chart shows how the timing method could help an investor avoid a disaster like the Great Depression. 2008 could have developed into this, if Central banks had not intervened.
http://www.advisorperspectives.com/dshort/charts/timing/SP500-MMA.html?Dow-monthly-MAs-1928-1940.gif
If markets keep going down, year after year, and you keep re-balancing into that decline, your losses keep compounding. Seeing a portfolio decline by such a huge amount would be very difficult to take. The timing method Faber endorses, avoids this problem.
So it under performs a lot of the time – the time it saves you from an 80% decline would make it worth it. – if that decline comes to pass. Don’t forget, it took 20 years to recover back to 1929 levels in the stock market – that could be a person’s entire retirement period. Perhaps those who worry about an 80% decline should not be in the stock market anyway. But it could happen again, and then people will wish they did worry about it.
Should we “buy and hold and re-balance” and assume the 80% decline can’t happen again, or should we have some exit strategy that might cause us to under perform the markets in good times? A scenario like The Great Depression would test the stomach lining of any “buy and hold” investor. We all need to determine how much risk we need to take, and be well aware of the maximum loss our portfolios could endure in a worst case situation, with the strategy we decide to use.
Typically great post Dan. Please keep up the superb work. I can’t remember the last time you posted something I didn’t immediately add to our “share” list.
@Jim: Glad you liked the post, but I fear you missed the main point, which is that market timing strategies sound tempting when discussed in the abstract (“It would have saved me from losses”) but in reality they are a recipe for disappointment.
Investors should always assume huge losses are possible. However, these losses can be mitigated by diversifying globally and by adding a healthy allocation to fixed income. No balanced portfolio is going to lose 80% unless we have global Armageddon, at which point your portfolio will be the least of your concerns.
And you must accept that equities are a very long-term investment: if you don’t have a long enough horizon to recover from expected losses, then you need to make your portfolio more conservative by reducing the equity component. Relying on market timing strategies to protect you from losses is unreliable, and likely to result in selling low and missing the recovery.
I like to remember the words of Ken Fisher as an antidote to magical thinking: “To my knowledge, no one has ever achieved market-like returns without some market-like downside. If you want to achieve something close to stocks’ long-term average, you must accept downside volatility. No way around that.”
Hi Dan:
I don’t think I did miss the point.
If you go back and look at Faber’s paper – particularly pages 32 – 34 , Faber disagrees with your post.
His paper says the strategy DOES work. And to emphasize it, his update focuses on the years 2006 – 2012, what he calls the “out of sample” data, the real world data that has occurred since the original study was completed.
He states of this “out of sample” time period (on page 34):
“The model performed exactly as one would expect it to from historical data. Namely, even though it only outperformed in three out of seven years, it beat buy and hold by over two percentage points per year, with much less volatility and most importantly to many investors, lower drawdowns.”
So Faber is saying it worked, and that it “beat buy and hold by over two percentage points per year” in the relevant time period.
The problem with your point, is that you have cherry picked your start year as 2009. Leaving out the year 2008 in your calculations, hugely distorts the findings and makes it appear that the strategy does not work, when in fact, Faber’s paper shows it does work. The whole point of the strategy is that it allows the investor to avoid massive drawdowns, thus improving overall results. When you left out the year 2008, the year with the massive drawdown, you left out the key year that shows the strategy DOES work. Avoiding the massive drawdown is the point of the strategy – and the point of my previous comment re: the period of 1929 – 1932.
If you start at 2006 – the year after his original paper was published, the strategy does outperform “buy and hold” – BECAUSE the strategy allows the investor to miss the massive drawdown of 2008 (the year you decided to exclude).
Which leads to my point (re: The Great Depression and other massive drawdowns) – being able to side step massive market meltdowns, allows you to preserve more capital than “buy and hold” does, and that is why Faber’s strategy DOES work.
From Faber’s paper:
2006 – 2012 the strategy out performed “buy and hold” by 2 percentage points per year
1973 – 2012 the strategy outperformed “buy and hold” by .56 percentage points per year
And it does this with much lower volatility and a much lower maximum drawdown on the portfolio, which allows the investor to more easily stay with the strategy.
If you cherry pick your start year, to a year after a massive drawdown ends, “buy and hold” may very well come out on top for the next few years. But massive drawdowns DO happen, and if you include them in the data (which you should), Faber’s strategy DOES work. Excluding them does not allow for a fair comparison.
Backtests are easy to deliver above market returns. All you do is tweak the parameters until your get the results you want. The issue is there is very smart people with expensive mainframes doing exactly the same thing. As soon as a winning strategy is found its exploited and then its not winning anymore rinse and repeat. The future is not predictable. I’m sticking with indexing.
@Jim: I “cherry-picked” 2009–12 because the previous version of Faber’s paper and The Ivy Portfolio include data up to and including 2008. I’m not sure why the paper starts with 2006. Possibly because starting with 2005 reduces the annual outperformance from 2% to 40 basis points.
In any event, I’m aware that the hypothetical backtests show outperformance over the long term: I explain it in the third paragraph of the post. No one is disputing the strategy looks wonderful on paper. But I can’t fund my retirement with hypothetical backtests. An investment strategy can only be said to “work” if one actually executes it. Faber and Richardson themselves tried to do that with their ETF and the results are nothing like they were expected to be. In its first full year (2011) it lost almost 8%, a bigger loss than any of the 40 years in the study.
My sense is that market timing is a strategy people love to read about, talk about and debate about, but hardly anyone actually does it in a systematic, disciplined way.
Dan, it’s cherry picking because they first published the paper in 2006. The data since then is thus out of sample and not vulnerable to data mining. Any timing strategy is likely to underperform when the market is rising. The test will be after the next big decline. I don’t think you can make any definitive conclusions over just part of a market cycle.
Loved the read but makes me cringe – I am moving my retirement fund into self-directed dicount brokerage in the new year and implementing the strategey (IF BMO ever gets its act together 1.5 mths so far – would not recommend them) I keep reading everybody laughing at bonds ie. Warren Buffet even. I am soo tempted to shorten up the recommended bond allocations – exactly what you are saying not to do here. Just to confirm you are not considering any movement in the allocations based on current enviroment (he asked knowingly).
@ Andrew F: – I could not have said it better myself.
@Jeremy: 2006-2012 is NOT a “backtest”. The original paper came out in 2005, so 2006 – 2012 (“out of sample”) is a verification that what the model predicted, actually worked for that period of time. 2 percentage points per year is significant.
@Canadian Couch Potato: I guess the concept of Indexing was a “hypothetical backtest” back in the 1970’s since “An investment strategy can only be said to “work” if one actually executes it”. Since it was not possible to execute Indexing prior to the mid-1970’s, the concept of Indexing was not an investment strategy, but was data-mined from a hypothetical backtest. All ideas and models have to start somewhere!
I always find it amusing when people talk about data-mining and back-testing as nasty terms, when that is exactly where the idea of Indexing originated. Looking at the US markets from 1926 to the mid-1970;’s and saying “Hey this works.”, was data-mining. After 1975, it was investing!
Indexing has not always worked in all countries. I still believe it is the basis of a sound investing strategy, but to a large extent, it depends on the country you look at. So is it a sound investing strategy, or is the success of Indexing in North American markets an anomaly? Maybe it is is a fluke that it has worked since it was data-mined in 1975. Time will tell. If you lived in Russia, or Germany, or Japan from 1900 – 1945, you might not think Indexing works. That is a 45 year period. We have only been “doing” Indexing for the past 37 years(since around 1975). Lots could still happen in the coming years to make Indexing look like a bad idea.
@Andrew F and Jim: There will be no resolving this debate, so I’ll step out of it. I look forward to hearing from investors who have actually used this market timing strategy and have real world results to share.
@Doug C: There is nothing wrong with using short-term bonds as a long-term holding if you want lower volatility, or with choosing a bond allocation appropriate to your own situation. That’s not market timing, it’s just risk management. The danger comes when you try to make repeated tactical shifts based on interest rate forecasts.
Great article. Im curious as to why the performance of the Cambria fund is SO bad. Management fees are high, and account for 1.59% according to the website. Okay, i get that. Trading fees would certainly be higher than a buy and hold, so theres something there for sure. But the total pefformance dif is almost 7-8%…..is that all just trading fees? Im jusy surprised to see such a cavernous difference between the theory and real world performance. Id have thought it would have been somewhat closer.
@Willy: The data in the papers and the book are largely based on a simple five-asset-class portfolio. The ETF appears to use a much more complicated strategy, presumably because they thought that would improve performance.
Ah, I see. So then, in fairness, this particular fund did not follow the prescribed “index”. Had it, we should have seen index returns, less the MER and trading fees. Instead, this fund tried to “actively manage” its way to better performance, which is not necessarily an indictment of this particular “index”.
Dan,
Thanks for all your insights with ETFs. It has made a remarkable difference with how I have set up my investment portfolio. We are fortunate for having this blog and your ongoing research and thoughtful comments
Have a happy new year and all the best for 2014!
Claire
@Willy: I think slightly more accurate would be to say that they designed a different, more complex index for the ETF to track. The initial portfolio was largely arbitrary, and designed largely for simplicity, so presumably the idea of the ETF was to optimize the same concept given that simplicity was no longer a limitation. Once the index rules were set, presumably there was no further active management. (Although I haven’t researched to see if they were tweaked after the fund was launched.)
I timed the market in my RIF account. Don’t fight the Fed. I selected a 25% Bond to 75% Stocks in 2013. here are YTD results of Dec. 31, 2013:
C$ Funds
TD Canadian Bond Index-e 8.4%
TD Canadian Index-e 12.6%
TD Corporate Bond Capital Yield-I -0.7%
TD Emerging Markets Inv Srs 3.9%
TD International Index Curr Neut-e 25.7%
TD NASDAQ Index-e 36.7%
TD U.S. Index Currency Neutral-e 32.1%
Total C$ Funds 16.4% return.
However, the Mawer Balanced Class A Fund in my LIF account had 20.2% return for 2013.
Happy 2014.
@gil: Not sure about your other numbers, but the TD Canadian Bond Index Fund did not return 8.4% in 2013. It lost close about 1.6%.
For the record, the ING Streetwise Balanced Growth Fund (also 75% equities and 25% bonds) returned 17.4%. Will have all the model portfolio performance numbers soon.
@Canadian Couch Potato The argument you made in this post doesn’t convince me for the following reasons:
1. You’ve compared a Buy and Hold portfolio against a GTAA one which will have on average 30% of it’s assets invested in cash at any time. Clearly this lowers risk and returns. A fair comparison would be the GTAA portfolio vs. a Buy and Hold portfolio holding 30% cash.
2. You compared absolute returns and not risk adjusted returns. (E.g. Sharpe ratio.) (Kind of the same point as above.)
3. You’ve ignored how well the Relative Strength Strategies (E.g. GTAA 13 AGG Top 6) have worked in that paper.
4. You’ve ignored draw downs when comparing the two. For a person growing their portfolio with regular contributions, I will agree that this doesn’t matter much but for someone in retirement, drawing from their portfolio it’s a hugely different story.
5. I am using a active strategy of my own device that is and has been working.
I personally use a variant of the Relative Strength Strategy and it is serving me well. Part of my variant ensures that I never have less than 65% in bonds (I am close to retirement and I don’t need to take any more risk than that) at all times so there is no more downside risk than a buy and hold portfolio with 65% bonds.
I’ve used a loosey goosey version of this strategy for ages but when it saved me from any losses in 2008 I woke up to embracing it more. Since then I’ve become more and more organized and rigorous with this strategy. At this point, I am fully committed to it for better or worse.
@Brian G: My own starting bias would be that agnostic Index Investing gives the maximum potential benefit for any given risk assumption, and that the transparency of the market rules out any hidden secrets; but I’m curious to hear any reasoning that might suggest you could lower your risk without jeopardizing your return potential by timing-driven asset allocation juggling. You say “Part of my variant ensures that I never have less than 65% in bonds”. That would mean you have a maximum of 35% in equities. This seems like a conservatively prudent percentage appropriate for your retirement proximity. So how much variation does your method allow? Would your equity allocation decrease markedly in times of increased perceived market risk? There doesn’t seem to be much room to move down from what seems like already quite a low risk position; or am I missing something?
@Oldie, I keep an open mind and I let my risk adjusted returns tell me if I am on track or not. I am well aware of my own hubris and that even a scheme that has worked in the past may one day stop working (this includes Couch Potato indexing, my scheme, etc.) Therefore I always invest using the idea that I could be wrong and I limit the downside to something I can accept if it goes pear shaped. My bias is that low cost indexed investing is good and maybe it is the best if you have a long horizon and are contributing regularly so you can take advantage of dips in prices. I would recommend it for any young person saving for retirement. Even for my scheme, it’s where I start regarding asset allocation and possible investment candidates.
Caveats aside…
As for my scheme, I won’t give specifics and I don’t recommend it because I don’t want to be responsible for anybody else but myself … but it is what I do. You are correct, at any time I can hold between 0% and 35% in “risky” assets (which includes longer duration bonds.) So how do I attempt to get excess returns if I can’t go above 35%? Mostly rotation with the occasional signal to go to all cash. At any point, I rank all risky investment options based on what my system thinks will do well and invest only in the top N. This system is an old one originally used on stocks and the parameters are numerous (E.g. what are the possible investments, what are the signals used, what ranking to use, etc.) Google “Relative Strength Robert Levy” to find the original paper on this idea.
@Brian G: Thanks for your prompt reply; I appreciate your candour and your consideration for others who might be lulled into following a scheme that might not work out. I must say I remain unconvinced. Part of the reason, with all due respect, may lie in the difference between your belief system and mine.
You basically believe in the underlying principle of the Index Investing system, but as I under stand it, want to tweak it to mitigate risks in the belief that there are subtle market signals that, if read accurately, can advise you to change course fast enough to mitigate the disaster that befalls all others who are too obtuse to decipher those signals, and yet not so prematurely that you miss some of the available profit left in the system. I, basically, don’t believe there are any consistent signs (that is, any mathematical tool that can massage the market data to produce a useful set of signs) that could tell us how to avert loss or, for that matter to capture profit, ahead of every one else. For this to exist would seem somewhat contradictory to me. For me, to believe in the first principle means to disallow belief in the usefulness of subsequent tweaking. The only decision to make is at the outset, and that is to assess one’s risk tolerance (in the face of ignorance regarding the future) and to fix one’s asset distribution accordingly. Therefore a statement to the effect that even the Couch Potato scheme might stop working in the future doesn’t make sense to me — how could it “stop working”? At very worst I could stop believing in it because of some market catastrophe that exceeded my risk tolerance, but that is hardly a failure of the scheme itself — it is merely delivering what it set out to do, to track the index (or indices).
@Oldie … I don’t expect to convince you, it’s just an explanation of what I do. However, the concerns 1 to 4 that I shared with Dan regarding this blog post not comparing apples to apples I think are valid ones (I hope).
This is what I know about markets: (1) past performance doesn’t predict future performance (2) pick any two investment instruments and they will be correlated rarely, sometimes, most of the time. You can’t use (1) but you can use (2). Index investing portfolios are built on (2) but they must assume static correlation to come up with a portfolio mix on the efficient frontier but static correlation is clearly wrong (e.g. that’s why most asset classed moved together n 2007-2008) … but it’s convenient. A rotation scheme uses (2) by assume mean reverting correlation rather than static correlation, this feels more right but the question is can you exploit it enough for it to be worth it.
Sorry, I pressed send to quickly. For what it’s worth, I was trying to answer your question of when a Couch Potato scheme “stops working.” My definition of “stops working” is when all the diversified assets you hold stop moving independently (usually in a down market) and become correlated, thus ruining the whole point of diversification reducing risk just when you need it the most.
As I’ve pointed out, this isn’t a big deal if you are in accumulation stage but it is in the income stage.
@Brian G: Thanks for the clarification, and for the sobering reminder of what might happen sometimes — and yes, this is one of the ugly scenarios that you have to include in your mental list of possibilities when investing passively, as well as heeding the ever-present footnote that you can never predict when anything is going to happen, market-wise. All you can possibly predict accurately is when you can expect to withdraw your investments and profits (if any); and so, well ahead of withdrawal time, one should have arranged to have that portion slated for withdrawal to have been systematically converted to GIC’s/ Cash etc. That isn’t attempting market timing, that’s just prudent investment cycle planning, allowing for unpredictable but possible disasters to occur without ruining your withdrawal plans.
I acknowledge that doing this will lessen the potential for profit during the period that the Equity portion was converted to Income. I suspect that even though our theoretical standpoints differ, from a practical viewpoint, our prudent asset switching actions prior to planned withdrawal would be similar. The difference might be if I were to have 15 years or more investing horizon room. I would then be inclined to feel that this risk had been taken into account already, and to keep on re-balancing asset classes to the pre-assigned levels regularly as usual.
@ Brian G: …but as you already pointed out “this isn’t a big deal if you are in accumulation stage…” so we are actually closer in behaviour than my rigid theoretical line in the sand would suggest, lol.
Great discussion guys. Thanks.
I first became aware of this strategy (selling when below the 10 month average) through this site, and I’ve been very interested in it ever since. I’ve done extensive back testing and played around with various time frames and variations on buy and sell rules.
While it’s not silver bullet that’s going to suddenly give you 20% returns forever, I don’t think its fair to completely discount its potential as a stabilizer in a long term portfolio.
The fact of the matter is no major correction has ever occurred without being preceded by some sort of breach of the 200 day/10 month average, often well in advance.
This also seems to only happen a couple times a year on average, maybe less. That’s pretty low maintenance for an “active” strategy.
In fact I saw one good point made by a few people.
Lets say this strategy does in fact under-perform the market, even by 2%. If a stock index returns 8% over time, and this method only returns 6%, but you don’t experience massive draw downs, don’t you think a lot of people would be happy to take the 6% and sleep better at night?
2% is the difference of a high MER that many people are already paying active funds for nothing. It’s a small price to pay for stability.
-I used 2% only because that seems to be roughly the difference that’s been thrown around in the arguments between pure passive and using this timing strategy and which over/under performs the other.
@Will: I find it hard to disagree with your logic and I have had similar thoughts about this approach. I can see it both ways, to be honest.
On one hand, this approach is simple, concise, “mechanical” in nature (ie just as with a passive portfolio, the decisions are automatic and made by calculation, not human judgement). Followed correctly, it is guaranteed to underperform in up times, but guaranteed to outperform in down times, which, for many psychological reasons discussed on this very blog, I think is important. I consider the fact that the fund cited in this article failed to perform as condemnation of the strategy – because the fund didn’t follow the prescribed strategy. I continue to believe that IF you have the fortitude to follow this strategy for 30+ years, inexpensively and without deviation, you’d be quite happy.
On the other hand, as CCP points out – find anyone who has. Owing to failings that are theoretically easy when discussed, but in reality very challenging, few are successful with this type of thing. It forces you to make a “decision” every month, on every asset class. That’s 60 opportunities/year for your “human weakness” or just plain error to get the better of you, and for you to deviate from the system. It does present higher trading costs as well, and could have tax implications.
Like I said, I see the merit to this strategy. I think, even for couch potatos, there is some merit to having some sort of “stop loss” mechanism that prevents you from having to watch a 40% drawdown (recognizing those are typically once-in-a-generation events). I think the psychological benefit of that alone makes it worth paying some small cost for. But there are real tradeoffs that may be just as bad or worse… so no slam dunk. Just my $0.02.
Sorry, meant to say I DO NOT consider the fact that the fund cited in this article failed to perform as condemnation of the strategy.
Will and others
At the beginning of ones saving and investing career and near and in retirement I find people are very sensitive to changes in portfolio value and fearful of losses. I understand what Oldie is saying about changing asset allocation in a life cycle.
For the young investor its because its hard to save and paper losses seem like money down the drain no matter how sophisticated ones knowledge of the benefits of passively investing and potential compounding time. Earning power is low, saving is hard as it competes with other needs such as saving for a house and car. I couldn’t stand to see my savings drop when I began because it would represent such a large proportion of income and it was very hard to set money aside.
When older I think the sensitivity comes from the relatively larger amounts involved and the fact that there is a higher risk of a bad outcome because of the need to live off of ones savings. Even with a relatively conservative allocation you may be only seeing 10% of a portfolio paper loss in a bad year but that 10% may be substantial dollars. It may be more than one earns in a year and it is very difficult to see a paper loss equivalent to or more than a years labour compensation.
For these reasons I get why any strategy that may reduce drawdown potential, even at the cost of long term returns, is appealing. I have spent time reading Fabers various strategies. In explaining the 10 month rules to someone else using active charts this person was amazed at what Will points out above: “The fact of the matter is no major correction has ever occurred without being preceded by some sort of breach of the 200 day/10 month average, often well in advance.” Does this have utility in a very practical sense?
I just wonder about the difficult psychology of dealing with major corrections or long term bears for those at the beginning and end of their saving and investing careers. Look how many people who have become completely risk averse since 2008 and 09.
You’re right, that fund they tried to make isn’t really following their strategy properly, they tried to make it way too complicated so no wonder it fails.
I prefer to evaluate it as a direct comparison to the couch potato portfolios. Keep it simple, only worry about the few index funds we use here.
I disagree that is forces you to make a decision “every month”. In fact it doesn’t really make you do much of anything other than a couple times a year in most cases. In fact the US market especially hasn’t come anywhere near touching the long term average in over 2 years, and until it does there is nothing to do but continue to hold or DCA more.
Yes you need to check it regularly but the majority of the time its just a glance and you know to go away and check next month. It’s a personal preference thing. For me it’s probably more hands off than I’d like and It would discipline me to not meddle more than I need to. For others it might be too much work, and that’s fine too
As for costs and tax implications, yes those are a concern. I would mainly employ this in a tax sheltered account , it’s a whole other animal if you are doing this in a regular account of course and likely not worth it.
Even trade fees can be mitigated or basically eliminated too. ETF’s at a broker like questrade would have no fees for buying and even for selling the fee is quite low.
I was actually thinking about this in terms of an e-series portfolio though. If you only have sell signals a couple times a year, and they are always the month apart, you shouldn’t ever hit the 2% early redemption fee, and buying back in is always free.
You can see on the chart at the following link(the chart is at the bottom of the web page) , that using the Ivy portfolio (which sells an asset when below the 200 day sma) does beat the 60/40 portfolio ( 60% stock / 40% bonds using buy and hold) over the long term.
So you don’t even have to concede anything like 2% for using this strategy. It works, and it works well. It stands up very nicely to all other strategies that use buy/hold.
http://mebfaber.com/2013/07/31/asset-allocation-strategies-2/
It would have saved you from the disaster of 1929 – 1932:
http://www.advisorperspectives.com/dshort/charts/timing/SP500-MMA.html?Dow-monthly-MAs-1928-1940.gif
In 2013, it would have got you out of long bonds in May and out of REITS in June – both of which have fallen a lot since then – you would still be out of them since they are still below their 200 day sma.
I am a big fan of this. It is easy to implement – takes the emotion out of the equation – it is easy to stick to because it gets you out during big market falls – and back in after the recovery starts – and saves you from the next depression. And it does all this without sacrificing returns over the long term.
@ andrew
You’re absolutely right about people’s perception of risk both early and later in their investment lives.
As I’m young I had a relatively small amount (around 5k) invested in a “balanced” fund with a 2% MER in my RSP back in 2007/2008. While the drawdown as somewhat concerning it also regained most of its value within a couple years and wasn’t too devastating to me. It wasn’t enough to make me really panic I guess.
I did after that devote myself to educating myself about the whole thing which led me to the couch potato as well as various other viewpoints (I’m currently in a basic e-series potato with a smaller bond weighting because I am young and feel I can accept the risk.)
I like the idea of employing a logical, informed strategy to a simple setup like the couch potato.
I’ve totally bought in to the idea of indexing rather than individual stocks or actively managed funds (though I’m still interested in more specific ETF’s to some extent, as long as they follow a consistent index type system and aren’t some funky active fund)
But I haven’t given up on the idea of being informed about what’s going on and keeping an eye on things, and even having a specific rule for managing the portfolio that keeps it on track and stable. Maybe Dan just hasn’t convinced me yet ;)
@Jim: I don’t think that’s apples to apples as the Ivy portfolio holds only 20% bonds, compared to 40%, and also suffers a max drawdown of -46% according to that same link you provided. So it doesn’t really solve the drawdown problem using those %s. Anyway, the point is that using a “stop loss” mechanism like this will NECESSARILY cost some return on the upside (it’s just mathematics), so it’s not free. But I still think that could be valuable. If you were offered a choice of (A) CCP’s Complete Couch Potato portfolio, or (B) CCP’s Complete Couch Potato portfolio plus a mechanism that would cost me 1% CAGR, but would virtually guarantee that I’d never see a drawdown higher than 10%, which would you take? That’s the fundamental promise here. I’d give serious consideration to the latter.
@ Willy
Indeed, I’ve done some backtesting but I still need to organize it all into a combined comparison to the 60/40 CCP. I’ve mostly been focusing on stabilizing the equity portion, as I’m not really as concerned about bond draw downs.
I think the beauty of this is that you can’t really deny the back testing other than from a psychological standpoint. Price vs Moving average is about the simplest technical you can have, and it’s a straight yes or no answer.
In fact I would argue that it’s much, much easier psychologically to follow this than Dan gives it credit for.
You have a solid signal telling you to sell, and following that signal potentially saves you massive losses. If you’ve been invested for a long time you are probably even selling at a profit in a book vs market value perspective as well. So yeah, I’ll sell no problem.
If its a fake out and you have to buy back in slightly higher, it’s probably still near the 200 day average and therefore somewhat “low” anyway. Maybe you lost a few %, big deal. A buy and holder might be re-investing more at this point anyway.
The hard part maybe is buying back in after the big drop. But I’d bet that isn’t too hard either if you’ve just successfully avoided a massive draw down and everybody around you is down 20-30% or more.
@Willy:
1. Using the 200 day sma does NOT cost you 1% in returns – if you look at the chart again, it does as well as any of the portfolios and BEATS 60/40 buy and hold – which is a couch potato portfolio. You seem to indicate that buy and hold beats selling on the 200 day sma by 1% – it does NOT.
2. I don’t know where your figure of 10% comes from re: the drawdown. Selling when an asset crosses the 200 day sma can be much more than a 10% drawdown. There is no “virtual guarantee” of 10% limit on the drawdown.
As the price of an asset drops, so does the moving average. Both can drop a lot more than 10% before the price crosses the moving average.
What doing this does do, is limit the drawdown to a much lower value than buy and hold allows – and saves you from a once in a life-time disaster like 1929-1932.
@will: Agreed. As I said, I share your views and can’t really refute any of it. I think if you followed this strategy, as prescribed, with discipline, then for all the reasons you mentioned (simple technical, binary decision), you should be very happy. I am confident that, overall, it will cause a drag on your returns vs. buy&hold (the size of which depends on market movements) – this much is certain, just pure mathematics. However I am also confident that I will not like to watch the value of my portfolio drop by 40% some day. :-) So take your pick, I guess.
(For the record I still follow a completely passive, Couch Potato style portfolio. I see the benefits/cannot refute the value in a mechanism like this, but I haven’t got up the gumption to implement it for myself yet, owing to some of the concerns I cited above – more “work” required, potential cost & tax issues, and a higher risk of deviation and loss if you do so at the wrong time).
@ Jim
I think the important thing people need to accept is that no, you won’t sell at the “top” or buy at the “bottom”, this method will still miss those. As you say the price has to start dropping well before the 200 day average is hit.
I tried a lot in vain of course to find a way to catch those tops and bottoms and it just isn’t possible.
For people inclined for a more advanced variation though, I do think there are a few ways to weed out the real drops from the temporary corrections, improving the accuracy and returns even more.
It’s much harder to back test though and of course introduces some arbitrary judgement calls. I’m nowhere near actually using it or recommending to others. More of a thought experiment for me at this point.
@Jim: To be clear, the numbers I used were “for example” and not meant to be strict. A strategy like this WILL cost you some premium in CAGR vs. buy and hold (its mathematically certain) – the magnitude of which is unknown and will depend on market movements. It will also reduce your maxmium drawdown – again the magnitude of which depends on market movements and is unknown ahead of time. So the point is, you pay something to get something here. The probability that you get something for free (ie reduced drawdown with no loss of CAGR) is unlikely.
@Willy:
Your “mathematically certain” outcome did not happen in the period 1970’s – 2012.
60/40 buy and hold CAGR: 9.48%
Ivy Portfolio (buy/sell at 200 day sma) CAGR: 9.92%
What DID happen, is that buy and hold was beaten by .44% per year on avg.
There was NO COST (as you say) to using this strategy – it BEAT buy and hold.
There was a “cost” to using buy/hold. That is CERTAIN from the data.
http://mebfaber.com/2013/07/31/asset-allocation-strategies-2/
@Jim and Willy
I think the important things to consider here are the conditions in which this strategy will truly lose vs buy and hold.
That being a case where the price drops below the 200 day average at the end of 1 month, triggering a sell, and then immediately goes on a massive bull run the next month, forcing you to buy back in at a much higher price.
Now considering the statistical fact that the market goes up over time, I can see why a passive adherent would be concerned about this potential lost profit. We all know the market goes up over time.
But how often does it do a 10/20/30% gain in 1 month?
Now what are the chances that hitting the 200 day moving average results in the following month being down 10/20/30 %? I’m pretty sure it’s happened more often than the former.
Yes, the strategy will have times when it loses you a few %. Maybe even within a certain year and certain conditions (like a whipsaw bull market) it will under perform for a year or so.
A lot of the time you are in the market getting the same gains as everybody else though, and I’m confident the losses you end up protecting yourself from far outweigh the short term gains you might miss sometimes.
@Jim: I am aware of the link you posted the first time. The problem is that the Ivy portfolio tested in that link did NOT employ the 200DMA buy/sell “stop/loss” that we are discussing. If it did, it didn’t work. :) Look down about 7 rows and you see the Max DrawDown for Ivy is 46%, very much higher than the 30ish% drawdown of the 60/40. In other words, you did not get “less drawdown and more CAGR”. You got more CAGR but “paid for it” with higher drawdown and volatility. Again, the math is simple. If you are using a moving average to trigger a decision, you’re going to “cut off” some ups and some downs – you can’t cross over an average without it. This is one of the lessons of passive investing (no free lunch, and those who promise it are usually leaving something out).
@Willy:
You are correct – the chart in the link I included does not use the 200 day sma part of the strategy – my mistake.
If you check the original paper that Dan mentions in his original post, the data we need is there – this is what I was thinking of when I sent the other link in error.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=962461
Look at page: 33
The chart for 1973 – 2012
This data does use the 200 day sma (12 month SMA as listed in the chart on page 33)
From that chart:
Buy and Hold return (CAGR) is: 9.92%
Max Drawdown is: -46%
The 200 day sma return (CAGR) is: 10.59%
Max Drawdown is: -14.09%
So from 1973 – 2012, the 200 day sma strategy beat buy and hold buy by an average of .67% per year (this is even more than the other link shows) and the max DD was only -14.09% .
So it DOES beat buy and hold and it DOES lower the max DD significantly.
There is no sacrifice in using this strategy over buy and hold.
Saying that you sacrifice return by using this strategy, contradicts the data from 1973 – 2012. It just isn’t true. It IMPROVES the return.
It increases the return AND it lowers the max DD significantly.
Maybe there is a free-lunch after all!
@ willy
I don’t really get the link he posted either, it’s just a bunch of portfolio’s with no specifics about how each worked, and yeah, it doesn’t make sense that the ivy had such a big drawdown.
I know how you feel on the issue of there being no free lunch.
But when I think of it theoretically, all extended upwards moves must by definition take place above the 200 day moving average and all extended downwards moves take place below it. It’s a simple fact of the math. If there is ever a sudden shift one way or the other from the wrong side it will cross that average. If it fails to then it’s not a true trend change.
I wonder as a thought experiment, if you were to set up auto trading system such that on a daily scale, you bought and sold any time the price closed above or below, effectively cutting the chart in half along the 200 day average and only being invested above it, what would happen then? Assuming no tax costs or comissions, I would think that overall it would perform positively, and never lose more than the long term average itself does. effectively you would be following that long term average perfectly forever.
In other words, what is the return of an index if you simply plot its 200 day average as a line?
Eventually it all reverts to the mean anyway
Here is another link looking at results from 1926 – 2012
Go down about two-thirds on the page to the heading: The Extraordinary Results.
There is a great chart there that shows visually how the timing model out-does buy and hold.
http://www.mutualfundobserver.com/2013/06/timing-method-performance-over-ten-decades/
I think you can also look at it in terms of what forces would cause this method to lag the index.
There are basically 2 “drags” on this method (not counting taxes and trading costs which would present a 3rd potentially, but can be mitigated)
a) the distance between the price you execute your trades and the moving average, both top and bottom. If the sell or buy trades have different deviations it effects the gains. You want both to be as close to the average as possible. The further either of them are the more of your gains get clipped (i.e. your signal was “late”)
b) whether the long term average has moved up or down between the signals. You of course want it to slope sharply downward after you sell, because that means your sell was correct and prices dropped.
If you sell and the average continues up, likely you got a buy signal the next month or two and the drop was negligible/temporary.
as for b) it will be hugely beneficial for you if the market tanks after you sell, and would be detrimental if the market goes on an upwards tear.
However, if the market tanks you will be staying out of it for quite a while as it drops. Meanwhile if the market springs back up you are likely back in immediately, probably even the next month.
Again as I said earlier, its impossible for you to get a sell signal, and then see the market shoot so sharply up that you get “losses” anywhere near the level you get if the market shoots down and you didn’t sell.
Look at any of the points where these indexes fall below the long term average. They either skip along the average sideways for a while and then pop back above, or they dip below a little bit or a lot before catching up to it again.
At no point does the market ever drop below the long term average and then immediately gap up 10% and proceed on a long bull run the next day
@Jim. I think we are both right, or at least saying the same thing differently. The GTAA does underperform in up periods. Faber acknowledges it directly in the paper, and you can see it in the chart on page 28. This is what im referencing. What you are seeing on page 33 is that, over the past 40 yrs, there have been 2-3 specific periods of intense downturns where the GTAA has vastly outperformed, as its designed to do. The outperformance in those periods has outweighed the underperformance during up periods, resulting in a higher average.
So the question really becomes, whats going to happen to you and i in the next 20 yrs? If there are any more large crashes as in 2008, then the GTAA willl certainly outperform and may likely weight out any underperformance. But those type of drawdowns are rare, only 3-4 times during the last 100yrs. So if they dont happen, you could pay a penalty. I find this very very analogous to insurance. I have insurance on my car. If i crash, I’ll be very glad i did and will come out ahead. If i never crash, i will have paid for all that insurance for nothing. GTAA is kind of like portfolio insurance to me. So i guess the real question is whether you think your stocks will crash substantially in your investing horizon. (And for the record i think this is at least very probable, ant some point, which is why part of me likes GTAA.) Cheers
Good point willy, the performance of a system like this really does depend on the type of market we have in the next few decades.
Looking at a long term monthly chart of the sp500, it’s hard not to feel like we’ve entered a period of repeated, giant whipsaws. Even though the crashes in the past like the 30’s depression were technically bigger, looking at a chart on a linear scale that only goes back about 20-25 years is just scary.
If you take the long term average from the 80’s and draw it smoothed out going froward, you pretty much get a line that leads right to where the 200 day average is today. We’ve basically gotten the average return for the last 20+ years, but we’ve had to endure two massive heart attacks to get there.
This also presents a tempting prospect… If you can even roughly “time” those two events, you get to grab a huge premium over the average return of the market by riding it from bottom to top twice.
So yes, it’s these two events that make a timing strategy truly profitable. Unfortunately that trumps up the real results somewhat if you are only backtesting to the mid 90’s (as I did with one method and got like 18% on the sp500).
You’re right though, it IS like insurance. If we get a calm rising market for the rest of our investing lives we are likely going to pay a premium for said insurance.
But something tells me the world isn’t going to exactly be the pinnacle of stability over the next 30-40 years…
The GTAA 13 AGG 6 strategy caught my attention in particular as a way to enhance returns over the more basic GTAA strategies. One measure of robustness is how well the strategy performs if the ranking / trading does not occur on the last trading day of the month, but some other day (say, the 2nd thursday of each month). It would seem that as long as a month was allowed to pass before re-ranking, the strategy shouldn’t care whether it is month-end or not.
I set up a spreadsheet and looked at 2011-today using yahoo’s adjusted price date. As described in the paper, the top 6 of 13 assets are identified each month by averaging the 1 yr, 6 mo, 3 mo, and 1 mo returns. If they are above the 10 mo sma, they are bought or held. If they fall out of the top 6 or below the 10 mo. sma, they are sold. The process is repeated every 21 trading days.
My numbers won’t match the published ones exactly, but since the beginning of 2011, trading at the end of the month yield a CAGR of 11.8%. So far, so good… the strategy is likely to underperform slightly in rising markets, and this does exactly that.
However, if the monthly trading date is moved from the end of the month to another day, the returns fall drastically. Perhaps the strategy is quietly benefiting from some kind of end of month anomaly, but since that is not explained in any part of its reasoning, it is hard to trust.
Days away from EOM : CAGR over last 3 yrs.
+14 : 4.1%
+13 : 2.9%
+12 : 4.7%
+11 : 3.3%
+10 : 4.9%
+9 : 5.6%
+8 : 6.0%
+7 : 6.2%
+6 : 4.5%
+5 : 6.7%
+4 : 5.4%
+3 : 8.0%
+2 : 7.1%
+1 : 10.3%
0 : 11.8%
-1 : 9.6%
-2 : 9.8%
-3 : 8.0%
-4 : 8.6%
-5 : 8.3%
-6 : 5.9%
I haven’t looked at the other GTAA strategies yet, but base on this, I have a hard time seeing how the GTAA 13 AGG 6 couldn’t be counted on to deliver anything like what is suggested by the paper.