By Vern Sumnicht, MBA, CFP® IInvesting

Improving Modern Theory xceptional volatility and the lack of of study from traditional MPT known as An asset-allocation algorithm should overall investment return during the post-. This field not use standard deviation to measure “lost decade” have investors and of study is giving legitimacy to obvious risk, but instead should only consider the theirE advisors touting the apparent death weaknesses in the way MPT is applied by possibility of losing money as a measure of modern portfolio theory (MPT). On the investors. What follows is an outline of the of risk. By doing so, one would determine contrary, to paraphrase Mark Twain, “The weaknesses of how MPT has traditionally that the investment represented by the rumors of MPT’s death have been greatly been applied, along with solutions to blue line was more risky because the blue exaggerated.” improve these weaknesses. investment line indicates a much higher A quick review of the basic tenets probability of losing $50 than does the of MPT begs the question: Which of the Problem #1: Standard deviation is a investment represented by the green line. principles derived from MPT are no longer poor measure of risk. This definition of risk is supported relevant? The use of standard deviation (SD) as by post-modern portfolio theory and • Are investors no longer risk averse? a measure of risk assumes all investment behavioral research. Perhaps • Are equity and markets no returns follow a bell-shaped, symmetrical most importantly, it is supported by the longer fairly priced or efficient? curve. Yet, seldom does any investment common sense of individual investors. An • Is the allocation of the portfolio, return distribution actually resemble allocation that uses standard deviation as as a whole, more important than a classic “bell curve.” Plotting most its measure of risk is an algorithm that can individual security selection or investment return distributions will result be improved. The optimal asset allocation ? in a graph that is skewed either positively model would consider “the possibility of • Should investors no longer invest for (like the green curve below) or negatively loss,” rather than standard deviation, as the long-term? (like the blue curve) in Figure 1. the measure of risk. • Is there no longer an Efficient By using standard deviation as the Frontier where every level of risk measure of risk, the blue and green curves Problem #2: Is your diversification truly has an optimal allocation of asset are considered equal (i.e., same expected reducing risk? classes that will maximize returns? return and SD). Yet, these are not equal The traditional equity • Would investors rather be investments: The green curve has volatility that most advisors and investors recognize concentrated in a few asset classes with more upside risk, and the blue curve as the norm for allocating equity portfolios than be diversified among a greater has volatility with more downside risk. An consist of large-, mid-, and small- number of asset classes with low unexpected gain (so-called upside “risk”) capitalization growth and value, along correlation to each other? is not what an ordinary investor considers with international and emerging markets. risk. Losing money is risk! Continued on page 24 Common sense instructs most Return On Investment Fig 1 investors that these basic principles, Figure 1: Return on Investment derived from MPT, remain as relevant 25 today as they were the day they were conceived. Most of the present confusion 20 seems to derive from mean variance 15 optimization (MVO); this is the asset- allocation formula used to determine 10 the of optimal asset-

allocation portfolios. But MVO was Probability 5 not intended to be used for managing portfolios, and it should not be considered 0 equivalent to MPT. -100 -50 0 Expected 50 100 With advances in research and -5 Return technology, academia has derived a field

22 Napfa advIsor OctOber 2010 Investing

Over the past few years, these Journal of Financial Planning in 2006 correlations, a more comprehensive and Iequity asset classes have become highly (www.fpanet.org/journal/articles/2006_ dynamic approach is needed in making correlated (see Table 1). When the Issues/jfp0206-art7.cfm), William J. asset allocation decisions.” correlation coefficients among asset Coaker demonstrates the instability Coaker’s findings confirm that the classes within investment portfolios begin of correlation variables. He concludes investment environment is constantly to converge on 1.0, they lose the value of that, “rather than rely on historical Continued on page 26 diversification for reducing risk (losses) within those portfolios. Table 1 To make matters worse, research Traditional Asset Class Correlation Coefficient shows that in down markets—when For three-year period ended June 30, 2010 Midsize Midsize Small Small Large Large Emerg Int'l diversification is needed the most to help Growth Value Value Growth Value Growth Markets Equities protect from losses—these asset classes Stocks Stocks Stocks Stocks Stocks become even more highly correlated and Midsize Growth 1 thus provide even less risk protection. This Stocks process was well-defined in 2008, when Midsize Value 0.95 1 an overall liquidity crisis drove most asset Stocks classes downward in unison as investors Small Value 0.87 0.96 1 of all types liquidated everything they Stocks could sell. Small Growth 0.96 0.96 0.94 1 The ideal asset allocation should Stocks be applied using asset classes with low Large correlation. By comparing the matrix Value 0.91 0.97 0.94 0.91 1 Stocks of correlation coefficients in Table 1 to Large the matrix in Table 2, it is apparent that Growth 0.98 0.94 0.87 0.95 0.93 1 advisors using Table 2 asset classes are Stocks Emerg allocating investment portfolios to asset 1 Markets 0.89 0.80 0.80 0.82 0.80 0.89 classes with lower correlations. Int'l Also, the correlations among asset 0.91 0.88 0.79 0.86 0.89 0.92 0.94 1 classes change over time. Therefore, Equities investors must monitor the correlations Domestic Indexes = Russell Indexes; Emerging Mkts = MSCI Emerging Mkts Index; Intl Equities = MSCI EAFE within their portfolios to ensure proper diversification. Table 2 Improved Asset Class Correlation Coefficient Problem #3: A more robust and For three-year period ended June 30, 2010 objective asset allocation algorithm is Gold Bonds Utilities Tech Real Est. Energy Health Finance Materials needed. Investors and advisors constantly Gold 1 struggle to make the mean variance optimization algorithm (MVO) work Bonds 0.29 1 to determine their optimal portfolio allocation. Step one is determining Utilities 0.13 -0.06 1 expected returns, standard deviations, and correlation coefficients for each asset Tech -0.03 -0.27 0.73 1 class. Then, each asset class’ minimum and maximum percentage allocation is Real Est. 0.07 -0.07 0.49 0.74 1 set. MVO has been the method of choice Energy 0.30 -0.15 0.76 0.67 0.48 1 for investors to determine optimal asset

allocations since it was developed in the Health 0.15 -0.03 0.67 0.71 0.67 0.49 1 1950s. Despite decades of use, this buy/ hold asset allocation algorithm has many Finance -0.01 -0.03 0.47 0.72 0.89 0.44 0.69 1 flaws that can be improved. In “The Volatility of Correlation Materials 0.23 -0.18 0.72 0.87 0.74 0.81 0.70 0.71 1

Important Implications for the Asset Indexes: Gold: Comex Spot Settlement, Bonds: Barclays Capital 20+ Yr. Treasury Bond, iShares DJ US Sector indexes, Allocation Decision,” published in the respectively.

24 Napfa Advisor October 2010 Investing

changing in a random fashion, and adviser’s fee structure “completely” a shareholder for 10 years or 10 days. investments are affected by many more or “fairly well.” Yet, in a low-return Furthermore,I many funds wait until late factors than expected return, standard environment, reducing costs becomes in December to make their capital gains deviation, and correlation. Some of these an increasingly critical way to improve distributions known, leaving little or no economic and capital market factors investment returns for any portfolio. time for investors to make other decisions Continued on page 26 include money supply, GDP growth, Typically, actively managed equity- to offset the tax liability. These costs can inflation, dividend yields, interest rates, based mutual funds without commissions add an additional 0.5 percent to 2 percent unemployment, etc. These variables and (no-load mutual funds) charge about 1 to the annual cost for investors. others need to be brought into the algorithm percent to 2 percent each year inside the Index ETFs can be a low-expense for determination of optimally allocated fund, to manage and operate the fund. alternative to actively managed mutual portfolios (Efficient Frontier). A more However, there are additional costs that funds that both avoid unnecessary robust and comprehensive model than are not typically considered in the cost expenses and enable an investor to balance MVO is needed to improve the investment equation. Transaction costs—the costs to his or her asset allocation according to performance results obtained from asset buy and sell stocks inside the fund— can MPT principles. allocation and re-balancing decisions. add another (undisclosed) 0.5 percent to 1.5 percent annually to an investor’s cost Apply the Principles Problem #4: Portfolio management to hold that investment. Understanding the weaknesses in expenses are significant. Additionally, when fund managers how MPT is applied enables investors to In general, the individual investor are faced with redemptions, they must improve the risk-adjusted returns that their has little understanding of the impact of sell securities from the fund’s holdings to portfolio achieves from asset allocation. fees on net return. A survey by Boston- provide for the redemptions. When • The asset allocation algorithm based State Street Corp. and Knowledge@ these securities are sold for a gain, the sale shouldn’t use standard deviation to Wharton (the online business journal triggers a taxable event for the remaining measure risk. for the Wharton School) found that just investors holding the fund in non-qualified • Asset allocation should be applied 43 percent of investors understand their accounts, whether the investor has been using asset classes that have truly low correlation to one another. • The asset-allocation algorithm needs if You’re Thinking of WaYS To to be more robust than MVO. That is, more than three basic factors are groW Your adviSorY BuSineSS, required to determine optimal asset allocation. Changes in capital market and economic factors, like money supply, inflation, unemployment, dividend yields, etc., need consideration in determining optimal asset allocation. • The asset-allocation approach should be applied using fee-sensitive investment vehicles to mitigate Think ScoTTrade adviSor ServiceS. management, transaction, and tax expenses as much as possible. Scottrade advisor Services delivers the tools and resources advisors like you need to run your business, your way: The tenets of modern portfolio • No matter the number • Scottrade Advisor • Whether it’s your theory still hold true. What is needed of assets under Services provides personalized team or is a rethinking of how to apply the management, you can you with a dedicated a branch office, all rely on a team that has relationship team that Scottrade services are principles. extensive knowledge gets to know you and focused on helping of Scottrade to deliver your business. your business grow, first-class service. without ever competing for your clients. Vern Sumnicht, MBA, CFP®, is the To find out more about how Scottrade advisor ® Services can support your independent business, CEO of iSectors , a NAPFA Resource visit advisor.scottrade.com or call (866) 306-7135. Partner. For more information, call Brokerage Products and Services offered by 800-iSectors (1-800-473-2867) or email Scottrade, Inc. - Member FINRA/SIPC. [email protected].

October 2010 Napfa Advisor 25