Alternatives to Buy and Hold (Part II)

In part I, I took an article by Matt Hougan of Index Universe to task.

Before I move on to more important topics, I'd like to briefly revisit my thoughts on what Matt Hougan had to say. Matt stated on February 5th, 2009 that the time for holding the market is now. He made no distinction between active 'traders' who supposedly have such high turnover they can never hope to succeed, and those who buy, close their eyes, and hope, er, hold for the long term. The common investor is becoming more sophisticated than Matt is giving them credit for and we realize there are myriad alternatives to either a) being a daytrader, or b) feeling like a sucker for buying and holding stocks on either their broker's advice or the advice of an online financial columnist who sees investing only in terms of diversification and expenses.

For someone like Matt Hougan, the key to any solid portfolio is diversification and low expenses. I couldn't agree more; however, as Mohamed El-Erian recently stated, “Diversification alone is no longer sufficient to temper risk. In the past year, we saw virtually every asset class hammered. You need something more to manage risk well.” (Kiplinger’s 2009).

With regards to expenses, a Seeking Alpha columnist recently wrote an article (coincidentally with Index Universe the subject as well) titled 'Asset Allocation ETFs: Low Expenses Are Nice, Losing Less Money Is Nicer'.

I couldn't have stated either point better – diversification and low expenses are essential elements of a portfolio, however if the last year has taught us anything it is that something more is needed to manage risk.

One such system which can be combined with a diversified, low expense portfolio is a moving average system. I detailed the basics in my previous article, but it essentially can be summarized as buying a security when it trades above a long term moving average and selling when it trades below the moving average. It is by no means a wonder drug, but the risk adjusted returns of moving average systems are superior to buy and hold (even when 2008 is excluded) and the turnover is low (unlike the implication Matt Hougan made in his article). Mebane Faber gives the best summary of the benefits of a moving average system in his recently updated paper. The system has produced 35 years of positive returns (and counting).

I would urge anyone reading this article to stop reading, download and read his article to gain a more in depth perspective on moving average systems. In addition, for a slightly different twist, check out Tom Lydon's ETF Trends website for another moving average system.

So how can we implement and—hopefully—profit from a well diversified, low expense, trend following system? Faber uses 5 indices in his paper, the S&P 500, EAFE, 10 Year Treasury, GSCI, and NAREIT. As I previously wrote, the common investor could mirror this portfolio by investing equally in ETFs that track the indices when the corresponding ETF is trading above your target moving average (for Faber the 10 month SMA):

• S&P 500 Index – SPY (or substitute VTI)

• EAFE – EFA (or substitute VEU)

• 10 Year Treasury – IEF

• NAREIT – VNQ, IYR, or ICF

• GSCI – GSP or DBC

When the ETF is below the 300 day SMA, sell and move to cash. Check the ETF's closing price once per month at the end of the month to determine your buy/sell position. This will reduce turnover (as opposed to checking prices daily/weekly, etc.) and in the long term still leads to much better risk adjusted returns.

However, for those investors who have a larger portfolio and wish to seek more diversification, they may want to expand beyond just the 5 indices/ETFs listed above. Below is an example of a portfolio with 22 positions partially inspired by Paul Merriman of FundAdvice.com. I have tweaked the portfolio to have a representation of commodities and alternative assets which were excluded by Merriman. All positions are ETFs with the exception of two (GSP and LSC, which are ETNs) thereby reducing credit risk. For those who wish to avoid ETNs altogether, for LSC I provide an alternative mutual fund RYMFX which has a high correlation to LSC.

Keep in mind that the portfolio would only be invested in SPY, for example, when SPY was trading above the predefined moving average target (such as the 300 day SMA used by Faber or the 200 day EMA used by Lydon). When SPY is trading below the moving average, the 5% of your portfolio that would otherwise be in SPY would be in cash.

US Stocks (20% of entire portfolio)

• 5% S&P 500 – SPY

• 5% Large Value - VTV

• 5% Small Value - VBR

• 5% Microcap - IWC

International (20%)

• 4% EAFE – EFA

• 3% EAFE Value – EFV

• 3% EAFE Small Cap – SCZ

• 3% Small Cap Value/Dividend – DLS

• 4% Emerging Markets – VWO

• 3% Frontier Markets – FRN

REITS (20%)

• 10% US REIT – VNQ

• 10% World ex-US Property – WPS

Bonds (20%)

• 4% Barclays 7-10 Year – IEF

• 4% Barclays TIPS – TIP

• 4% Investment Grade Corporate Bond – LQD

• 4% Barclays International Treasury Bond – BWX (or IGOV)

• 4% International Government Inflation Protected Bond – WIP

Commodities and Alternative Assets (20%)

• 4% GSCI – GSP (or substitute DBC as an ETF)

• 4% Element S&P CTI – LSC (or substitute RYMFX)

• 4% Precious Metals(gold and silver) – DBP

• 4% Current Harvest Fund – DBV

• 4% Agribusiness – MOO

This portfolio is not intended for everyone, in fact it may not be for anyone. For example, the real estate portion is equally weighted among only two holdings due to the limited choices among real estate ETFs. Some may wish to underweight the real estate sector to less than 20% of the overall portfolio in order to allocate in other areas. Also, MOO was added as an 'alternative' investment even though it holds equities—one could remove it and allocate more to the other alternative assets or replace it with another commodity ETF. One could also remove EFA and subdivide the EAFE allocation between a European ETF (such as VGK) and a Pacific ETF (such as VPL).

In addtion, those with a lower risk tolerance or close to retirement would probably be better served using a more conservative asset allocation (more on that in an upcoming article). Taxes are also a key consideration, any portfolio with a degree of turnover will have tax implications. Ideally, one would trade the system in an IRA. Finally, I am not intending to imply that the 300 day SMA is the optimal moving average. The 200 day EMA or SMA tend to be the most popular for identifying long term trends (Tom Lydon uses the 200 day EMA for his moving average ETF system).

For those seeking more than 5 positions but for whom 22 is too much, a 'compromise' portfolio of 12 positions may look like:

10% SPY

10% VBR

-------------

10% EFA

10% VWO

-------------

10% VNQ

10% WPS

-------------

6% IEF

7% TIP

7% BWX (or IGOV)

-------------

7% DBC

7% LSC

6% DBP

Even though a moving average system like the one Faber details has a low turnover (Faber's research shows an average annual round trip for various asset classes around .70) portfolio size plays a key role to how many funds are selected. For example, if you have $15,000 to invest then you would be better served sticking with fewer selections (such as the original five) in order to limit commissions as a percentage of the overall expenses on the portfolio. The larger the portfolio, the more positions can be justified due to the lower toll the trading commission takes as a percentage of the overall portfolio. Even at $7/trade, that still is around $10 per year per asset class on average in commissions (assuming .7 roundtrips). One should carefully consider the appropriate number of holdings based on the overall size of your portfolio. At $5000/position a $10 per year average commission is a .2% extra in fees. Then again, when faced with the alternatives, the risky strategy of buying and holding with no endgame in sight and facing 50%+ drawdowns, or paying a financial 'advisor' 1-2% for the same buy and hold portfolio and to tell you 'this time it really is the bottom', I think I know what I would do.

Hopefully this article can serve as a starting point for building a well diversified portfolio combined with a moving average trend system to capture the upside of the various sectors while avoiding much of the downside.

For more information on Faber, click here.

For more information on Tom Lydon's strategy, click here.

For an excellent chart of current ETF prices in relation to their moving average, visit Lydon's ETF analyzer.

Comments

[Chris B:]

No trading strategy works forever. People who continually chase the system they should have used last time end up behind the curve.

How many buy-and-hold investors bailed out during the tech bust of 2001-2002 and underperformed the indexes pursuing various trading and hedging strategies for the following 5 years when they should have bought and held? How many of those people then realized they weren't getting anywhere flipping assets and switched to buy-and-hold in 2006-2007?

Bottom line: if you flip flop trading strategies, expect to join the vast majority of investors who underperform the indexes over long periods of time. If you stick with one strategy, you have to accept down years.

As for me, if I devote 4-5 hours of research into identifying a great investment, I'm not going to sell it unless the market price exceeds my estimate of fair value. The 300d MA does not account for the actual value of the company, it accounts for the ebbs and flows and emotions of leveraged daily traders. If you own a company worth $50, and the market action drags today's price down to $25, why would you ever sell?

Mar 05 01:40 PM

[bigmoney:]

I like this approach, as it captures whatever sector is trending up for whatever reason, whether that be inflation, economic growth, or just irrational exuberance. I think these approaches have potential because many people will buy & hold, or move to cash and stay in cash.

I know people that are nearly 100% stocks, buy & hold fashion, and they stubbornly refuse to consider alternatives. They "dollar cost average". The current market sag is nerve-wracking, because many of the same macro-economic issues persist, with zombie banks, quasi-nationalization, etc. I'd say its quite possible to trend down from todays 6590 another -30%, and then trade sideways, while inflation does it thing. And in the not too distant future, large numbers of baby-boomers will be hoping to retire, or else "right-sized", and unable to secure quality employment.

Mar 05 11:57 PM

[Scott's Inv...:]

Chris b - I would encourage you to read the empirical results in Faber's paper as it doesn't sound like you've read the article. This is not a system I would encourage anyone to 'chase' but can serve as a long term portfolio strategy (combined with diversification,etc.). Buy and hold will outperform a moving average system in certain years (like you said, If you stick with one strategy, you have to accept down years), however the data on the s&p since 1900 is compelling (I think we would agree a fairly large sampling size) as is the date on the other major indices since 1973. In addition, there is a body of work outside of Faber which supports the superior risk-adjusted returns for a moving average system (see my first article for a few examples). This is not a system that was developed after the crash of 2008, it has produced 35+ years of positive returns (and counting).

Good luck in your research, if the market disagrees with your assessment of 'fair value' on a company (see 2008 or today?), I hope you have stop losses in place...

Mar 06 08:42 AM

[Scott's Inv...:]

Correction: Please note that Faber's system uses the 10 month SMA, not 300 day, I made the update but apparently too late for Seeking Alpha to make the correction.

Mar 06 08:51 AM

[Schneidehouse:]

Quick question . . . why not go short on the index instead of going to cash?

Mar 06 09:23 AM

[Scott's Inv...:]

Good question Schneidehouse - According to Faber going short historically does not improve the returns on the portfolio I think partly because historically the yield earned on cash helps contribute to positive returns (despite today's abysmal rates) and the margin costs of shorting an index would reduce overall returns. For a brief discussion, visit www.mebanefaber.com/20.../

Mar 06 09:34 AM

[exbury2:]

Hi there!
Since there are only 20 days of trading in a month,The 10 months moving average should not be far from the 200 days moving average.
Am i wrong?
Alain

Mar 06 11:32 AM

[Kinabalu:]

A 20% alternative asset allocation (equal to the U.S. stocks allocation) is far too high, particularly with a concentrated base fund like GSCI. There isn't enough history, and related portfolio research, to support anything more than 5% in that area.

Mar 06 01:04 PM

[Scott's Inv...:]

exbury2: Yes, I made the correction but apparently the article posted by Seeking alpha is still showing 300 day in a few areas - it should read '10 month sma'. Kinabalu: Read Faber's article, using the GSCI as a 20% allocation was the benchmark used which helped contribute to the risk adjusted returns. However, you could certainly adjust the percentage as you see fit. The goal is to get diversification across as many non-correlated assets as possible and there is an abundance of evidence that direct commodity exposure helps lower volatility and increase risk adjusted returns.

Mar 06 01:27 PM

[bcncv:]

Conceptually, I like the system, but Chris is right that no system works forever. Particularly, they always seem to work best just when you think they shouldn't.

Personally, I'm starting to develop a philosophy/strategy that you should apply diversification to trading strategies as well as asset classes. Theoretically (I have yet to test it), having a percentage of your portfolio as "buy and hold" and a percentage executing a strategy like this should significantly smooth out returns over time.

Mar 06 06:33 PM

[Kinabalu:]

On Mar 06 01:27 PM Scott's Investments wrote:

< Kinabalu: Read Faber's article, using the GSCI
> as a 20% allocation was the benchmark used which helped contribute
> to the risk adjusted returns. However, you could certainly adjust
> the percentage as you see fit. The goal is to get diversification
> across as many non-correlated assets as possible and there is an
> abundance of evidence that direct commodity exposure helps lower
> volatility and increase risk adjusted returns.

Scott

I read the article. My only comment on the moving average system would be to note that it's a model. Models have a distressing habit of exhibiting entirely different risk-return matrices when they are actually implemented in investment portfolios with the resulting market impact..

My comment above was specifically directed at the asset allocation choices selected to display the model characteristics. Again, 20% is far too high for alternative investments, and including the GSCI index as a so-called diversified basket of commodities, as Faber does in his article, is misleading. That index is, in fact, 74% energy weighted, a fact which may have caused the non-correlation achieved during the period studied. The abundance of evidence you cite certainly exists for a relatively short historical period but I am not convinced it will be there in the future.

Mar 06 07:34 PM

[Scott's Inv...:]

We can certainly debate the percentage which should be allocated to commodities, but the data goes back to 1974 so it is a fairly large sampling period. However, I would agree that it would be a bit risky to consolidate everything in a GSCI tracking ETF, hence my proposal to allocate the 'alternative' portion (also open for debate) in a larger portfolio among

4% GSCI – GSP (or substitute DBC as an ETF)

4% Element S&P CTI – LSC (or substitute RYMFX)

4% Precious Metals(gold and silver) – DBP

4% Current Harvest Fund – DBV

4% Agribusiness – MOO

One could also even substitute an absolute return fund or some other alternative asset.

Mar 06 08:10 PM

[Marc Michel:]

This system would have us selling virtually everything today as over 90% of all stocks are trading below their long term average. God help us if everyone adopted it.

Mar 07 06:03 PM

[richjoy403:]

Scott--I want to thank you for a well-constructed and most-timely article, with reference to other useful works. I have been searching for just such a strategy for the core of my portfolio (I expect for example that I will continue to take 'flyers' on a few interesting individual micro caps). I will most likely adapt a version with a slightly shorter SMA (maybe 120 days), as I do not consider the $8 trade costs important to the size of my portfolio. However, I certainly do need a strategy that will allow me to not spend hours every day obsessing over small market moves. There has to be more to life than staring at CNBC every day while making buy/sell decisions!

Mar 21 10:02 AM

[richjoy403:]

Scott--I am very much interested in your article, and have been thinking more about it through the weekend.

I especially appreciate that you have expressed some flexibility in the application of your model, and thus I am asking the following:

(1) Our economy is cyclical, having peaks and troughs about every 4 years (most recently about 5 years). Does the model work equally well when applied dogmatically (or uniformly if you like) at both near-peak and near-tough points of the economic cycle?...or for example, might one apply a tighter 6% or 7% stoploss to sell toward the late stages of an up-cycle so as to cautiously preserve profits, and likewise to cautiously limit losses on buys in what are expected to be the early stages of the next up-cycle? Likewise, in what is expected to be a near-trough bottom, could one expect to successfully modify the buy rule to 100 (or even 65) days EMA?

(2) Does this timing system work as well for individual stocks as for ETF baskets?..for example,would one have the same buy/sell rules for a high beta stock as for a basket of utility ETFs?

Any assistance you might lend will be much appreciated.

Mar 22 04:27 PM

[Scott's Inv...:]

Hi - You could apply the system to individual baskets of stocks, I think the primary thesis holds true across the equity markets as a whole; however, be mindful that a diverse asset allocation across all sectors is key in order to capitalize on trends - a simple trend system like this only helps in maximizing alpha, diversification and low expenses do the bulk of the work. Regarding stop losses, you can play around with different models to see how it affects returns, but gauging the peak/trough points in any cycle is not always easy to do. However, you raise a valid point that Tom Lydon addresses on ETFtrends - I could be mistaken but I believe his original model was exclusively based on the 200 day MA as the buy/sell signal. However, given how far we have fallen in most markets and the distance below the 200 day MA, he now proposes entering a position in the following manner for people who are deep in the red and make look to capitalize a little quicker on trend reversals (I believe this switch is simply for today's market conditions and overall he still adheres to the 200 day ma strategy): * When a fund crosses above its 50-day moving average, put 25% of the value of your portfolio.
* When the fund goes up 5%, put another 25% in.

Ultimately, the goal is to catch a trend, stay disciplined to a system, and cut your losses to preserve capital.

As a final note, the guy who intially got me started on trend systems is Mebane Faber - his new book is now available on amazon, I haven't read it (it's enroute) but I am sure there is plenty in there for both of us to soak up.

Mar 25 11:30 PM

[Scott's Inv...:]

Also, regarding your previous comment, I couldn't agree more! Putting a system together and having the discipline to follow it is the hard work, the buy/sell decisions come automatically which will allow you to take the remote and change the channel from CNBC. I've been playing around with some moving average ideas on the Forex market but have yet to come up with anything I feel comfortable with (mainly because I'm a novice at FX and need to work on position sizing/stop loss triggers). Eventually I'd like to clean up my blog and convert it to primarily tracking MA model portfolios like the ones I've written about, but I do have quite a few ETF charts on the right hand side if you need a quick reference for moving averages for a particular ETF.

Mar 25 11:36 PM

[hayek's right:]

With all the discussions on contra-ETFs, would it be possible to enhance your returns by using the 200-day moving averages to switch between long and short ETFs? Would using a 3%-5% guard band on switching decisions reduce the number of toggles (and transaction costs) to a reasonable level?

Apr 04 06:55 PM

[Scott's Inv...:]

To answer your first question, Faber studied that and showed that there was no significant edge gained by shorting the market below the average. The returns earned on cash contributed to the lower risk/return ratio for the strategy. Regarding the band, yes, that would decrease toggles and transaction costs (and potentially taxes), and could also slightly decrease returns on the underlying investment.

Apr 05 11:33 PM

[richjoy403:]

Thanks Scott, much appreciated!
...and I hope (as you suggested) you do in fact convert your blog to tracking MA model portfolios, as I think it would offer something much needed.

Apr 06 10:45 AM

[Scott's Inv...:]

Tom Lydon'ts ETF Trends website has a great (free) tool for tracking ETFs and the percentage they are above/below various moving averages. I have caught a few mistakes in their spreadsheet, so I would double check anything on the list.

Apr 06 01:17 PM