An article in the 12/2/06 Wall Street Journal – and here at MarketWatch.com – describes a simple three-fund portfolio that has provided higher returns and lower risk than the S&P 500 over one-, three-, and five-year spans.

Author Jonathan Burton – investment editor for MarketWatch – has written an article that all investors should read.  The thesis is simple: rather than building a complex portfolio in an attempt to eke out every last vestige of profit – with the attendant management and transaction fees such a strategy entails – you can create a well-balanced portfolio that performs well by selecting three good mutual funds: a U.S. stock index fund, and international stock index fund, and a U.S. bond index fund.

He quotes Meir Statman, a Santa Clara (California) University finance professor, who describes the strategy as a “‘cold-shower’ portfolio: you’ll do fine, but you’ll not have the biggest house in the fanciest neighborhood.”

The key to the strategy – as in all well-diversified portfolios – is to have “your money spread among investments that tend not to rise and fall in step with each other.”  Mr. Burton mentions a specific portfolio: 65% in the Vanguard Total Stock Market Index Fund (VTSMX), 20% in the Vanguard Total International Stock Index Fund (VGTSX), and 15% in the Vanguard Total Bond Market Index Fund (VBMFX).  Over the period from October, 2001 to October, 2006, this portfolio would have grown by more than 58%, compared to just under 21% for the S&P 500; furthermore, the proposed portfolio would have had less risk – lower volatility of returns – than the S&P 500.

While we at Bottom Line Gurus agree with the overall strategy that Mr. Burton suggests, we couldn’t help but analyze his recommended portfolio; it’s what we do.  Using our Portfolio Optimizer Pro software, we created a portfolio comprising the three suggested funds, and ran two simple analyses: one using five years of historical data from 10/1/01 to 10/1/06, and one using five years of historical data from 10/1/96 to 10/1/01.  The former covers the period during which the recommended portfolio outperformed the S&P 500 Index, and the latter covers the period immediately before that: data that would have been available to an investor on 10/1/01, and which could form the basis for his investment decision.

First: the holding period, 10/1/01 to 10/1/06:

From the efficient frontier graph, two observations are immediately clear:

The proposed portfolio has both a higher return and lower risk than the S&P 500 over the same period, as the author correctly stated.

The proposed portfolio is not even close to being efficient: at the same level of risk there was a portfolio that would have produced an additional 3.5% of return.

From the monthly return graph, one other observation is immediately clear:

The two stock index funds – VTSMX (yellow) and VGTSX (green) – are very much rising and falling in step with each other.  In fact, their correlation of returns over that period is +85%; that’s about as rising-and-falling-in-step-with-each-other as you can get: exactly what the author correctly says you want to avoid.

The efficient portfolio over that same period having the same risk as the proposed portfolio is seen here:

Note that this portfolio includes none of the US stock index fund: the international fund produced much higher returns and the high correlation of returns means that the US fund provided no useful additional diversification.

Note, too, that the benchmark US 10-Year Treasury Bond lies to the left of the efficient frontier created from the three index funds.  The frontier could be improved – greater return for the same risk or lower risk for the same return – by including the 10-Year Treasury (or any similar, interest-bearing cash equivalent).

Next, the five-year historical period immediately preceding the holding period, 10/1/96 – 10/1/01:

The range of the graph has been expanded to include all of the individual securities and the two benchmarks.  Two items are particularly noteworthy:

Over this five-year period the proposed portfolio, while still less volatile than the S&P 500, had lower returns than the S&P 500: 5.5% compared to 8.5%.

The average annual (compounded) return on the recommended international stock index fund, VGTSX, was -1.84% over the five-year period, with a correlation of returns to the US stock fund of +82%. 

While in hindsight it is well to suggest that VGTSX would have been a good investment between 10/01 and 10/06, in October, 2001 it is difficult to believe that a typical investor would have imagined that the average annual return would jump from -1.84% over the previous 5-year period to +15.4% over the next 5-year period.  A more reasonable three-fund portfolio would be one that includes funds that would have looked attractive on 10/1/01.

In summary: the advice given in this article is extremely worthwhile, but the implementation could have been improved, both in selecting funds that offer much better diversification to the portfolio and in selecting funds that would have appeared to be good prospects at the start of the holding period, rather than funds that only proved to be good prospects ex post facto.