The Effect of Gold in a Diversified Portfolio – Part 2

In Part 1 of this article (6/13/11) we looked at how including gold as a candidate asset class affected classic Modern Portfolio Theory (MPT) portfolios. We found that gold was only advantageous for the more conservative portfolios and even then only at modest percentage allocations. MPT is, of course, still very much relevant for those with defined financial objectives and long time horizons who wish to minimize the risks associated with achieving their financial goals. The concept of broad diversification among investment classes is a basic and universally accepted precept of sound investment practice.

Nevertheless, MPT does have its critics. Its chief assumption is that the future will be like the average of the past, a point that has been the subject of debate. While in the long haul that is very likely to be true, there will be significant periods of time when that does not hold. Extended bear markets or short term crashes such as those experienced in the last decade can wreak havoc with even properly diversified portfolios that can take many years to repair.

The question now becomes: Is there a way to avoid the bulk of losses experienced during downturns while still benefiting from the periods of recovery? The answer to this is yes, and we’ll see how this can be accomplished.

The application of market timing to MPT-constructed portfolios
What we will be looking here is the application of a cautious market timing modification to MPT. In particular we will examine the effect of including gold as a candidate asset class in the same way as we did in the first part of this article.

Briefly, the approach is to apply a timing oscillator optimized specifically for each asset class. When the oscillator turns positive, a long position is entered into with the portfolio percentage determined by MPT. The position is held (with monthly allocation adjustments as determined by MPT) until the oscillator goes negative, in which case the asset class is liquidated and moved into cash (T-bills, insured money markets, etc.). Utilization of this monthly rebalancing strategy has historically shown to achieve the same financial objectives as classic MPT but at much lower risk. More detailed information on this approach is available here.

Risk comparison between non-gold portfolios with and without the benefit of market timing
Let’s look at some results. Table 3a below shows the current allocations (through April 2011) for market timing modified MPT portfolios without gold as a candidate. Note that the risk associated with each targeted portfolio return (given in column 1) is defined by the standard deviation shown in column 2.

Table 3a. Current Market-Timing Modified MPT allocations without gold

The portfolios in the above table compare to those in Table 1 in Part 1 of this article. The table is reproduced here for your convenience:

Table 3b. Current MPT allocations without gold (no market timing)

An important takeaway from this article is to note that the application of an optimized market timing system considerably reduces risk in an MPT constructed portfolio. At higher portfolio percentages (8% and above), the risk is reduced by 50% or more which is amazing! Additionally, the probability of experiencing a losing year drops as well. In a 10% target return portfolio, one can expect a losing year every six years instead of every three.

Another important point to note is that higher returns can be achieved by using the market timing approach. Compare the top returns in both models: a 12.2% return is the maximum achievable return (at this point in time) in MPT-constructed portfolios with no market timing versus 13.8% in models with it.

Risk comparison between portfolios containing gold with and without the benefit of market timing
Table 4a below shows the current allocations (again through April 2011) for market timing modified MPT portfolios with gold as a candidate. The striking difference from Table 2 in Part 1 of this article (reproduced in Table 4b below) is the prevalence of gold throughout the range of alternative portfolios.

This result tends to indicate that gold is a worthwhile investment component in your portfolio but only if you watch it closely and are nimble and reactive in getting in and out of your positions. However, comparing the standard deviations with those in comparable Tables 3a and 3b above for those portfolios of equal required return, the advantage of including gold as a risk reduction tool is actually quite minimal.

Table 4a. Current Market Timing Modified MPT allocations including gold

Table 4b. Current MPT allocations including gold (no market timing)

A historical perspective
The above tables reflect current allocations and risks associated with MPT generated portfolios. These numbers, however, are not indicative of long-term performance. For that, we need to start at the beginning of our data (January of 1928) and calculate how each portfolio would have done since then.

To that end, let’s return to the conservative 6% compounded annual return portfolios we looked at in Part 1. Figure A below compares the classic MPT approach to the market timing modified MPT approach without gold as a candidate asset class. The lines show the growth of a portfolio following each of these two strategies with monthly rebalancing according to each model’s recommendations. The magenta line is the classic MPT portfolio (repeated magenta line from the graph in part 1) while the green line represents the performance of the market timing approach.

As you can see, both approaches met the target return and in fact exceeded it (7.1% for classic MPT, “long” and 7.3% for marketing timing modified MPT, “comb”). But that is not the important result as both were only trying to achieve a return of 6%. The significant result is that the market timing modified strategy achieved the target with a substantially lower standard deviation than the classic MPT portfolio, 5.4% versus 6.8% (a 20% reduction in risk).

Figure A. Historical comparison of MPT derived portfolios (without gold) for a 6% required return with and without the benefit of market timing


You can see from the above table that over the past 83 years both approaches not only attained but exceeded the required 6% return. The difference is that the portfolio that included market timing did so at approximately a 20% lower level of risk.

Now let’s compare the same portfolios but with the addition of gold. Figure B below shows that again both portfolios met and actually exceeded the targeted return, and again the market timing strategy (green line) did so at an even lower risk (25% lower) than the classic MPT portfolios!

Figure B. Historical comparison of MPT derived portfolios (with gold) for a 6% required return with and without the benefit of market timing

Adding gold increases risk
Comparing the risks from both plots for comparable strategies we see that including gold as an asset class actually caused a 7% increase in risk in the classic MPT portfolios and a 2% increase in risk for the market-timed MPT portfolios, although that was mitigated by the slightly higher returns achieved. Again, the application of market timing decreases the risk in both models.

From our analysis, we can make the following conclusions:
1. The application of market timing to MPT significantly decreases portfolio risk.
2. The addition of gold increases risk in both models (market-timing vs classic MPT models) but much less so in the market timing model.
3. Higher levels of portfolio returns may be achieved with the market timing model.

The conclusion reached in this article reinforces the conclusion stated in Part 1: There is little to justify gold as an asset class for long-term objective based investments.

FYI: Visit our website for further information on how we apply market timing and how your portfolio can benefit from it, too.

Acknowledgement: Pat Glenn contributed research and analysis to this article.

Leave a Reply

You must be logged in to post a comment.