The previous article explained Portfolio Diversification and made an attempt at showing the benefits using only two different mutual funds. This article will explore some additional ways to improve portfolio diversification.

Use More Assets

One way to improve portfolio performance is to use more assets or to try to get assets that don’t correlate highly with each other. Three different portfolios are tested below from 1995 through January 2019. These portfolios are as follows:

1. Portfolio 1 is the 60% US stock40% US bonds looked at in the previous article.
2. Portfolio 2 contains additional assets, including those not expected to correlate highly with one another.
3. Portfolio 3 is the Harry Browne Permanent Portfolio discussed here.

The content of each portfolio and its weighting is shown below. Each portfolio was rebalanced each year.

Three Portfolios

(Click on image to enlarge)

Portfolio 2 contained 11 assets. The first three represent classes of US stocks (30%). The next three represent International stocks (20%). Then there are two categories of traditional US bonds (20%) and some “Junk” bonds (5%). The last two are REITs (5%) and Gold (10%). Both these should have low correlation to stocks.

Portfolio 3 is the Permanent Portfolio of Harry Browne.

The latter two portfolios are different and were expected to produce better diversification results. Here are the results:

(Click on image to enlarge)

Results

Comments on Results

Portfolio 2, with more assets and lower correlated ones, underperformed Portfolio 1 (the two asset 60 – 40 split) on both return and drawdown. It provided protection for the first market debacle (2001) but not the second (2008). Portfolio 3 performed as designed showing slow, low volatile growth over the entire period. Its drawdown was only 13% versus the other two which were in the thirties. The market benchmark of all stocks suffered a drawdown of 50%.

Depending on your degree of risk aversion, these results might not look promising. If you want low risk from these alternatives, you settle for a CAGR of only 6.72 (your money will double approximately every 13 years). If you want a higher return (double your money say every 9 years), prepare for drawdowns of 30% plus. Or, you can just buy the stock market index, double your money in 7 or 8 years while suffering through at least one drawdown of 50%.

The Uniqueness of the Period

The period looked at is somewhat unusual, at least with respect to history. It is highly volatile and contains large market swings. The chart below shows annual returns coded by decade. 

(Click on image to enlarge)

Only the decade of the 1930s and the first decade of this century had four losing years. The thirties had one year of 50% losses and three years of greater than 10% losses. The first decade of this century is the only one to have registered four years with greater than 10% losses. Unfortunately, the chart captures losses by calendar year. For investors, calendar years are artificial constructs (other than for income tax purposes). Time periods defined by cycles would be more useful because they would define run-ups and drawdowns.

If we were determined to improve portfolio diversification, there are other things we might do. However, continuing to search for better results easily slips into “curve-fitting” (searching for oddball combinations until we came up with one or several that looked good). This is backtesting done badly!

Terrific results using historical data are possible if you try enough different combinations of inputs. But this type of data-fitting is a recipe for disaster. Just because something looked good in the past does not mean it will perform well in the future. That is almost guaranteed for any curve-fitting session. To spot curve-fitting, ask yourself two simple questions:

1. Reality Check: Do the relationships being tested have any basis in economic reality?
2. Robustness: Do the relationships holdup when nearby values are used to run the backtests?

I am a fan of diversification and you should be also, especially if you are a novice investor. But even serious investors should be because all eggs should never be put into one basket. Regardless of how confident you may be about some future outcome, you are likely to be wrong. (Even the impossible has a 20% probability!)

Rather than searching for better diversification portfolios at this point, we shall move on to other methods to improve return and reduce risk. We will return to diversification once we have some new tools in our toolbox.

The next article will present one such technique. It is simple. Anyone can apply it. And, it produces amazing improvements in return and risk for recent markets.