Diversification: Myths and Mastery

I believe diversification is one of the most powerful but misused concepts in finance. Professional investors use it as an excuse to buy horribly priced assets, index funds use it as the prime rationale for  their existence, and some people go completely away from it trying to get rich quick. I am here to tell you that Diversification is a bit like fire; If used properly it can do wonderful things for you (cook your food, heat your home) but if misused and misunderstood it can burn your home down or worse.

First of all, what is diversification: Wikipedia defines it as follows: 

In finance, diversification is the process of allocating capital in a way that reduces the exposure to any one particular asset or risk.

Basically, don't put all of your eggs in one basket. For example, if you put all of your money into Microsoft you have a whole bunch of things that could go wrong and would cost you dearly. At the highest level, the overall markets could go down which would affect most companies but there is also risk that something happens only to Microsoft - A competitor comes in with a disrupting technology (something different but better) for their product. They could have a lawsuit that threatens the existence of the company. It could be discovered that their reported earnings had been fraudulent. A natural disaster could destroy key properties or kill key people. I could go on and on but you get the point. There are many things that can affect the price of Microsoft (or any company) that is independent (or idiosyncratic) of what happens to other companies. This is not a problem in itself but because there are many companies and the risk can effectively be removed from your portfolio (by buying a group of companies), you are not compensated (with greater returns) for this risk. It should also be noted that there is also a possibility of independent factors working in your favor and not always against you. 

Side Note: It is important to understand that current price and expected return are inversely related. In other words, the more you pay for something the lower the expected return and vice versa. The only exception to this is if a company goes bankrupt, in this case your return is -100% no matter what price you pay. This is a little understood concept so I will give you an example. If you could see the price of Microsoft 10 years in the future was going to be $80, it is clear that buying today at $40 is much better than buying at $50. 

To take it one step further, it does not matter what the end number is, it could be lower than current price even. The result is still the same, just not as appealing, you lose less if you pay less.  

Our problem is that we try to back into the correct price for the company today by estimating where the stock price will be in 10 years. Even worse, we feel confident in our estimation. We then base our current price on that estimation. This might be ok if that estimation remained static but it does not. It moves around (sometimes violently) based on current news and events. I have found it much more profitable to focus on what we know (the current price) and realize that the end price is out of our control and unknowable. This should help you focus on current price in a healthy way (lower price is better for a good company).

How to diversify: Research has found that it only takes 20 different investments to diversify a portfolio. It is most effective if the investments vary by many dynamics, such as sector, size, asset class, geography, etc. My experience tells me that diversification of individual asset exposures is crucial to creating a portfolio that will do well in many circumstances and you will notice that all of my investments are well diversified. I am using ETF's and mutual funds that are invested in many securities, in some cases many thousands of securities. So a single investment has huge diversification embedded in it. 

Where people go wrong with diversification is they expand the definition to include all risks. They believe that if you balance your portfolio among all the available investments then you will have a protected portfolio in case of a downturn. This does not work because of one simple fact. When we hit a big downturn, like the recession of 2008, all risky assets drop. In investment terms, correlations go to 1. Diversification does not help when all asset prices are moving the same direction. 

So, how do you protect yourself? One idea that many have is to get out of the markets when things are expensive. This has not worked well historically because the markets can (but do not always) move much farther both above and below fair value than expected. Also these moves can last a very long time and over time the fair value is going up. Imagine this scenario; you get out of equities because they are overvalued today, they remain overvalued for 10 years before they come back to fair value but by that time the fair value is higher (maybe much higher) than when you sold. You have made a big mistake being out of the market even though you may have been right in your evaluation at the time you decided to exit.

This has caused many to jump to the conclusion that a diversified, constant allocation to both risky and risk free assets is the best solution. If you cannot use current valuations relative to the "fair" valuation to decide when to take on more or less risk, then pick a risk level you are comfortable with and just hold it constant (this is the current institutional thinking). 

My model: I believe we can do better than that. I focus on letting the market tell me when to take on more or less risk. I have looked at many economic and investment indicators and plotted them against the subsequent returns of many different indices. I broke each indicator's values into 5 groups (quintiles) highest to lowest and looked at the average subsequent 1 year return of the indexes for each quintile's results.  Most did not give any meaningful information but some had very strong relationships with subsequent returns of various indices. 

For example two of the indicies average returns modeled out like this vs the same indicator:

A) 1st to 5th Quintile Average Returns: -11.7%,  2.5%,  1.7%, 2.3%, 12.3%

B) 1st to 5th Quintile Average Returns:  13.8%,  8.7%,  4.2%, 9.7%,  -1.3%

There is a very clear preference for "A" in quintile 5 and "B" in quintile 1 for this indicator. Now this was backward looking which does not always work going forward, however I have been using this for 10 years on a forward looking basis with similar results. It is also important to note that this is an average and that there is significant variability in individual results.  What that means is that you do not want to put all your bets on one 12 month period.

I structure my strategy to take advantage of this information (avoiding indices that were less likely to perform well and going into ones that were more likely to do well). I also structure it in a way that gets diversification over time with minimal transaction costs. I recommend a new allocation each month on the one asset class that is expected to do best based on current market indicator positioning. Every month a new position is taken and the one that is one year old is liquidated. This is called laddering. Over the course of a few years the portfolio will take many positions (one per month) and the benefits of this diversification should bring returns close to the averages over time. This all assumes the relationships I have found continue to work.

For More Details:

If you want more background on the system start with the Tactical Allocation (introduction) blog, for ideas on how to implement (trade) the strategy see the Technical Stuff blog and for a statistical analysis of the past results (some actual and some backtested) see the Into The Weeds blog.

Other than that I post what I am going to do at month end and the following business day I post what my trades actually were and give a performance update.

Past performance is not a guarantee of future performance.  This strategy is presented for informational purposes only and is not a solicitation to buy or sell any securities. October is one of the peculiarly dangerous months to speculate in stocks in.  The others are July, January, September, April, November, May , March, June, December, August and February. ~ Mark Twain