Nice to be back to posting. San Diego has been so full of fun, friends, family, and amazing local beer, that my blogging definitely suffered…
Is it even worth picking individual stocks? Even in quant portfolios? The more I delve into tactical asset allocation and it’s ease and benefits the more I’ve been asking myself this question. In this post I’m going to present some historical data that really brings this question to light. Then I’ll point you to some pretty amazing, yet complex research on this topic. The results may just lead you to give up any stock picking whatsoever even of it’s just the computer doing it for you.
OK. On to the data first. In the table below I show the all of the key portfolios I’ve presented on the blog in the past along with many of their historical portfolio statistics, e.g. return, standard deviation, sharpe ratio, etc…For the purposes of this post I’ve classified the portfolios into equity only, buy and hold, and tactical asset allocation portfolios. I’ve then sorted each category by compounded annual return, the most often used comparison metric.
I could talk about this table for days on end. I’ve highlighted all the TAA portfolios and then any other portfolio that beats the worst TAA portfolio (GTAA5). There are only 5 portfolios that beat the worst TAA portfolio in terms of annual return – two quant portfolios which are 100% stocks, those two quant portfolios mixed with the optimum portion of bonds to minimize risk, and only one buy and hold portfolio (IVY B&H 13). Let this sink in for a minute. The worst TAA portfolio beats the 100% SP500 portfolio and a 100% foreign stock portfolio. And the worst TAA portfolio beats all the highly touted buy and hold portfolios as well.
But of course it’s not only about compounded annual returns. Most investors care about risk adjusted returns. Sharpe ratio is the most common metric to compare risk adjusted returns. On this metric only two portfolios (both quant portfolios) beat the worst TAA portfolio (GEM). Using the Sortino ratio (which does not penalize positive returns) only the two combo quant portfolios I created beat even the worst TAA portfolio (GBM). When you consider the ease of implementation and what I call “stick to it-ness factor” it is really hard to beat any of the TAA portfolios. Impressive and eye opening to day the least! And wait to I show you in a subsequent post how SWRs compare across all these portfolios…
Is this something new and earth shattering? Not really, but it surely isn’t talked a lot about a lot especially in the finance industry. Most of them are too buys chasing fees. There is a lot of research comparing asset allocation vs individual security selection. A recent paper (free registration required) from GestaltU does a great job summarizing this research and adds to it as well. In a recent post they summarize the findings quite nicely. I quote the relevant portion of the post here.
In the meantime, it’s no secret that we are big fans of active asset allocation, which is sometimes called ‘tactical alpha’. Our enthusiasm stems from the following observations from our own research, and from other published sources:
1) Asset allocation dominates portfolio outcomes. From an empirical standpoint, Kaplan demonstrated that policy asset allocation explained, on average, 104% of institutional portfolio performance. In other words, portfolios would be better off, by about 4% (note: not 4 percentage points) by sticking to passive exposures for each reference portfolio sleeve than by allocating to active managers. From a theoretical standpoint, Staub and Singerdemonstrated that asset allocation explains 65% of orthogonal portfolio breadth, while security selection accounts for just 35%.
2) Asset risk premia are extremely time-varying, such that asset classes can underperform vs. long-term means for extended periods of time, sometimes decades.
3) Asset returns and covariances exhibit persistence in the short-term which enables economically significant forecasting of one or both parameters over short and intermediate horizons. This means it’s possible to systematically alter allocations to asset classes through time to take advantage of time-varying premia. (See AAA whitepaper here)
Filtering through the techie jibber jabber, basically asset allocation explains most portfolio performance. And this data is before applying factors such as momentum and value. And before applying risk management strategies the protect to the downside. Once factors and risk management is added to the mix you get the tactical asset allocation results similar to those presented in the table. Go back and look at the table I presented above. Ask yourself two questions. Is it worth picking individual stocks, even in quant portfolios, in the face of such strong TAA results? If you’re in buy and hold portfolios, isn’t it worth considering moving to TAA portfolios at least as part of you’re over all asset allocation? Food for thought.
22 Comments
Niklas · March 22, 2015 at 12:29 pm
Hi Paul,
thanks for an interesting post!
Have you made any testing or reseach how quant-portfolios like trending value behave if a stricter (than original -50%) stop-loss strategy is applied? One interesting srategy is trailing stop-loss where the loss is calculated from the highest value and not from the value where the stock is bought. Tim du Toit wrote in his blog about stop-loss strategies and he came to a conclusion that 20% trailing stop-loss strategy improves the performance compared to buy and hold strategy. It is also much easier to stick in quant strategies if max draw-down is smaller than in the original versions. You can find the blog here:
http://www.quant-value.com/truths-about-stop-losses-a-stop-loss-strategy-for-the-newsletter/
When I have calculated the allocations of various strategies in my portfolio, I have noticed that I have to put most of my assets to tactical asset strategies and only smaller amount to the quant strategies due to the large draw-down of quant-strategies. It would be interesting to improve these strategies by using i.e. trailing stop-loss. This would allow us to allocate more to these strategies and it would change these results a bit if it worked.
It would be nice to here your thoughts about this modification to the quant strategies. What do you think?
paul.novell@gmail.com · March 22, 2015 at 1:34 pm
Hey Niklas, very astute observation on THE biggest issue with quant portfolios, living with large drawdowns. I have been researching adding risk management rules to the quant portfolios quite a lot. And like Tim, I have found that the only one that really improves results over the normal quant portfolios is a trailing stop loss rule. I plan to post on this in the future but so far the rule that works best overall is a 30% trailing stop loss applied on a monthly basis. The trailing stop rule only improves drawdowns enough to match the index but is still better than none. But it doesn’t come close to the low drawdowns of the TAA strategies. More to come on this in later posts.
Paul
Ben O · March 22, 2015 at 2:03 pm
Hi Paul and Niklas,
Curious to read your post on the stop loss when you have it.
One method I am looking at for my slightly modified version of TV is basically the buy/sell signal of the taa approach. In other words: be 100% long TV when s&p > x-months moving average, be long 100% TV and short 100% SPY when under the moving average.
This is effective as the large drawdowns between TV and S&P are correlated. You end up with even better CAGRs and lower drawdowns… Just like TAA!
Let me know what you think.
paul.novell@gmail.com · March 22, 2015 at 8:24 pm
Hi Ben, in my backtesting using SMAs, like in the IVY TAA portfolios, in the TV portfolio makes returns worse. It does not improve returns, sharpe ratios, etc.. at all. The only item I have found to improve risk adjusted returns is using trailing stop losses.
Paul
Ben O · March 22, 2015 at 8:29 pm
Weird, it does in my adaptation of it. I use the S&P ma for deciding when to hedge though, not the ma for the TV portfolio. At least it for sure improves cagr for me and drawdown, the two things I am most concerned about.
Looking forward to your post on that topic!
Ben
paul.novell@gmail.com · March 23, 2015 at 10:13 am
Ben, interesting. If you want to send me a private email we can compare notes. Would be interesting to see differences in our approach.
Paul
Sam · March 22, 2015 at 12:30 pm
Paul,
I’m glad you have a Master’s in Finance, so I don’t have to read all those studies 🙂
The older I get, the more I’m convinced that the simplest rule based approach is best. I like the work you are doing, especially laying out many alternatives to compare.
Have you read Jason Kelly’s most recent book, THE 3% SIGNAL? An easy read where he builds on some of his previous books. I like the approach he takes using fictional characters over an extended period of time while weaving tons of news stories that ‘seemed’ important at the time. Has the book come up on your radar, and if so, care to offer an opinion? It seems like a most reasonable approach, with enough ‘to do’ to keep the most ardent tinkerer satisfied. Without be a spoiler, basically the approach is a value averaging with quarterly signals…..based on Michael Edleson’s book from ’06.
Really enjoy reading your thoughts on finance and the stories your wife writes about your adventures. My wife and I want to do the same thing you guys are!!
Thanks,
Sam
paul.novell@gmail.com · March 22, 2015 at 1:35 pm
Sam, I haven’t read Kelly’s most recent book. I’ll add it to my list. Thanks.
Paul
B · March 23, 2015 at 4:17 pm
Although the Trending Value screen from WWoWS does not beat all the TAA screens, it has pretty good numbers for a long-only screen. It looks like it does a good job of beating all the other buy and holds you list. It kills the buy and holds in performance and even the Sharpe is not that bad.
Thus, your idea of combining with bonds (30-70 TV Bond) and/or possibly combining with timing or trailing stops seems good. The TV screen seems like a winner to start with so any reduction in draw-down would be great. I look forward to the results.
Again – great job Paul! I wish I was tasting all those nice craft beers like you!
Tx – B
B · March 24, 2015 at 9:21 am
Paul,
I bought and read the “Dual Momentum” book based on reading about the GEM and GBM strategies you are tracking on your blog.
What etf’s are you using for tracking GBM?
Thank you!
paul.novell@gmail.com · March 24, 2015 at 11:40 am
VTI and VEU.
B · March 26, 2015 at 4:38 pm
Paul,
Thanks. I should have been more clear, I was wondering what you are using for the ETF’s as part of the BOND portion of GBM. Thanks again!
paul.novell@gmail.com · March 27, 2015 at 11:07 am
VGLT, JNK, SHY, IGOV. Buy top 1 based on 12 month total return. Would be invested in VGLT right now.
Paul
John · March 25, 2015 at 2:36 am
Is there a reason you prefer ETF’s over mutual funds?
paul.novell@gmail.com · March 26, 2015 at 8:29 am
John, in general ETFs are cheaper than mutual funds but for the big asset classes you can find index mutual funds that are as cheap as the ETFs. A lot of people also like in the increased liquidity of ETFs, can buy/sell anytime not just at the close of the day. For me, the biggest reason I like ETFs is that they are for more tax efficient than mutual funds so in taxable accounts they can make a huge difference in net returns overt time.
Paul
John · March 28, 2015 at 8:09 am
Yes. I was just thinking about it because you mentioned that MTUM was not that liquid and it seems if you put in a sizable limit order(especially in a retirement account) you might have trouble filling the order.
paul.novell@gmail.com · March 28, 2015 at 8:30 am
John, you bring up a good point. In certain instances using a mutual fund instead of an ETF may be the best approach, e.g. in a retirement account. I have not had any major issues using MTUM in several accounts tax and taxable. The ETF keeps growing in size so that is helping. There are some viable alternatives out there as well. But there is no reason not to use mutual funds in a non taxable account as long as the fees are the same or lower as the ETFs. All else being equal the saving on the bid/ask spread would be a plus. Something to add to my to do list – would be interesting to see what/if any gaps exist in replicating the asset classes with mutual funds. Have you looked at this by any chance?
Paul
John · March 28, 2015 at 2:43 pm
I charted VIGRX against MTUM on Morningstar and they seem to track each other closely so it seems like a viable alternative.
Jacq · March 28, 2015 at 5:46 pm
Hi Paul (long time lurker here) – any thoughts on portfolios made up of value stocks a la the University of Michigan “Value 40”?
http://webuser.bus.umich.edu/tradingfloor/earningstorpedo/value40/value40.htm
paul.novell@gmail.com · March 29, 2015 at 3:06 pm
Looks good on paper. Since they don’t divulge details of the strategy, except that it combines value/quality/momentum, I can’t really comment. And If I can’t replicate it myself I would never invest in it.
Paul
Mark · March 29, 2015 at 12:54 pm
Hi Paul,
Missed your posts while you were in San Diego, but glad to have you back posting. Couple comments/questions.
1) For people just jumping into the GTAA or any Asset Allocation investment for the first time, I think it is important to do a stop loss on initial orders. Unlike someone who has been in the market for awhile with gains under their belt that can absorb a loss, jumping in initially without some protection is unadvised in my opinion. Also, curious what you think about setting up trailing stop losses once the market is moving up in your GTAA3 or GTAA6 portfolio. I know GTAA automatically protects you with a monthly rebalance, but a trailing loss (% to be determined by each investors risk level) would protect gains if the market were to tank in any class before the monthly rebalance.
2) Many experts and signals like the 10 year CAPE and others, indicate the market may be running out of steam and we may be heading into bear markets in certain asset classes. What is your opinion of using Inverse ETFs in addition to global asset allocation? For those unfamiliar with inverse ETFs, they are basically a way to short the market using ETFs for defined indexes. They have several advantages to traditional short selling and can be used in most IRAs and 401ks. Have you considered this as a vehicle to add to your portfolio in down markets? Interested in your thoughts.
As always, thanks for the work you continue to do and your thoughtful responses.
Mark
paul.novell@gmail.com · March 30, 2015 at 9:50 am
Hi Mark. Thanks for the great questions.
1) I don’t think stop losses are a good idea when establishing these portfolios. Most investors this is meant to ‘protect’ couldn’t handle it IMO. Stop losses are easy on the way out, where they kill performance is on the way back in. I’d rather see investors average in to a portfolio over several months, even though the data shows that jumping in right away leads to better performance over the long term. As to trailing stop losses as an addition to AGG6 or AGG3, they don’t work. Backtest results show they hurt performance in these portfolios. Momentum needs to be given a chance to work.
2) I hate inverse ETFs. They are a trading vehicle, not an investment vehicle. And the fees are exorbitant. If I would do anything to protect the downside any more than the AGG systems do, I would consider hedging with index futures. They are cheap and efficient. Problem is that strategy doesn’t work. I’ve tested it over and over. It hurts performance over the long term.
These TAA portfolios are designed with risk reduction as an integral part of them. But you can’t mess with them too much. GTAA13 for example has had a max daily drawdown of 10%, and the AGG ports about 20%. This is wonderful performance vs any equity heavy buy and hold portfolio.
The other thing is that one of the main reasons I like these types of the portfolios is the ease of implementation and the requirement of only checking in once a month. I don’t want a TAA portfolio that requires as much work as a quant portfolio. For that I’ll go with a quant portfolio with it’s better returns.
My 2 cents…
Paul
Comments are closed.