Portfolio Optimization Isn’t Necessarily Worth the Effort

A few weeks ago, I offered a list of what assets an individual investor needs to hold in their portfolio. Second in order of importance on my list—behind stocks, bonds and cash—were REITs (real estate investment trusts). These real estate investments offer a dual benefit of diversification and favorable returns.

I bring this up because a study argued that REITs are not a separate asset class. The points made in the study are worth discussing to shed light on how some academics and practitioners use math to optimize portfolios. The usefulness of the findings for most individual investors is questionable. While one can certainly put a great deal of effort into optimizing a portfolio, the actual benefit often will not justify the effort. The vast majority of investors will benefit more from a simplistic strategy that can be adhered to over the long term than a complex strategy based on mathematical attempts to optimize a portfolio’s allocation.

Before delving into some of the specifics of the study, the authors of the aforementioned study (Jared Kizer and Sean Grover of Buckingham Asset Management), acknowledge the diversification characteristics of REITs. They included correlation data showing REITs marching somewhat to their own beat when compared to the S&P 500 and being nearly completely independent relative to five-year Treasury notes. So the argument wasn’t about whether including REITs can be beneficial to a portfolio, but whether they are necessary. It’s a discussion which, while thought-provoking, should not rank high on your list of considerations about how to allocate your portfolio.

Kizer’s and Grover’s argument starts with the correlations of other sectors. Both energy and utility stocks also have reduced correlations with the S&P 500 and five-year Treasuries. (REITs, energy and utilities are somewhat independent of each other well.) All three sectors have their own S&P Dow Jones and MSCI sector classifications, but utilities and energy stocks are not considered to be their own asset classes.

A different way of answering the question is to determine whether or not the returns of REITs are being driven by certain factors. The authors find that REITs have “positive and statistically significant exposure” to the size, value, term (the additional yield required for holding a longer-dated bond—in this specific case, the five-year Treasury) and the investment-grade yield premium (the extra yield required to hold corporate bonds instead of Treasury bonds). Given these characteristics, they believe owning small-cap value stocks and long-term bonds “should do a decent job of replicating the returns of REITs.” Since this is possible, Kizer and Grover say REITs are therefore not a distinct asset class.

For an individual trying to manage their own portfolio, attempting to adjust weights of small-cap value stocks and bonds beyond what they would have otherwise allocated to them solely to replicate REITs is a lot of work. It requires return data and analytical software most individual investors do not have, not to mention a strong background in mathematics and statistical analysis. Plus, by simply owning REITs or a REIT fund, an investor still realizes good diversification benefits.

I’m not picking on Kizer and Grover; actually, I found their study interesting. Where I raise the question is in its application to individual investors. Whereas some industries and processes require exactness, investing does not. Similarly, while higher level math is a prerequisite for some fields, it isn’t for investing. Instead, what successful investing requires is a disciplined process. A moderately good allocation strategy adhered to over the long term will always outperform precise allocations that are difficult to construct and maintain.