“Plans based on average assumptions are wrong on average.” – Dr. Sam L. Savage

It is my pleasure to announce that the Portfolio Strategy & Research (PSR) team at Loring Ward has created an independent Advisory Council (our “Think Tank”) to guide us on investment-related themes and other aspects of financial services. Dr. Sam Savage is our first official member of this Council.

Sam has a very accomplished career in both academia and the private sector. After teaching at the University of Chicago Business School for 15 years, he came to Stanford University in 1990 as an Adjunct Professor in the Engineering School. Sam has also taught courses at Northwestern University, the Naval Postgraduate School, and the Cambridge University Judge Business School, where he is a Fellow. He received his Ph.D. in Computer Science from Yale University in 1973, followed by a year at the General Motors Research Laboratory.

Sam is also Executive Director of ProbabilityManagement.org, a 501(c)(3) nonprofit devoted to the improved communication of uncertainty and risk through the SIPmath™ Standard, which is based on his invention: the SIP data array (SIP stands for Stochastic Information Packet).

Sam is the son of Dr. Leonard J. Savage who, along with Dr. Milton Friedman, inspired Dr. Harry Markowitz to lay the statistical foundation for how we evaluate risky investments today. I have heard Dr. Markowitz mention several times to me how Friedman and the elder Savage are his idols in our industry.

The Flaw of Averages
One of Sam’s passions is the Flaw of Averages, a term he coined in a San Jose Mercury News article in 2000 on the dangers of making investment and other planning decisions in the face of uncertainty. In his subsequent book on this topic, Sam stresses the idea that “plugging average values of asset performance into a financial plan does not predict the average performance of the plan.” Since most economic and business processes fluctuate widely around these averages, rules based solely on averages often lead to bad decisions.

A simple example that Sam is fond of telling is of the man who decides to cross a river because he has been told that the average depth is three feet, but then he drowns when he reaches the part of the river that is 15 feet deep.

Let’s take another example that’s even closer to home. Let’s say that we have a client who has amassed a portfolio worth $1 million as of his retirement age. His advisor mentions that the expected return of his stock-heavy portfolio is 6% per year. The client acts upon that estimate and decides that he can afford to withdraw 6%, or $60,000 per year, for the rest of his life. Now, he has implicitly assumed his portfolio will earn a consistent 6% per year, with no deviation, and that his original $1 million will not diminish in value…ever.

As noted, this is a stock-heavy portfolio, and we know that stock returns can be volatile. What happens if in years three, four and five the stock market is down -10% each year? (We know that, historically, even worse peak-to-trough declines have occurred in the stock market.) If he kept withdrawing $60,000 per year from his now-diminished portfolio, he would likely exhaust his entire portfolio by year 20 of his retirement.

The Flaw of Averages error was choosing a withdrawal rate equal to the average or expected rate of return based solely on that average, and nothing else.

Sam has found that the Flaw of Averages permeates all industries. From a Behavioral Science perspective, people like to keep things simple, and decision-making using a simple average is easier to comprehend than a decision process that considers an average and the range of outcomes that could occur around that average. In fact, Sam’s invention of the SIPmath™ helps overcome such psychological barriers by representing uncertainties as sourced, auditable data.

The Bottom Line
Loring Ward believes strongly in the view that the sole use of averages (such as expected returns) must be tempered to account for the degree of risk — fluctuations around the average — that exists, regardless of whether the average turns out to unfold as expected.

Throughout our investment process at Loring Ward, from start to finish, we take this general approach to heart, including in the following ways:

Our Capital Market Assumptions (CMAs)
While our CMAs (key inputs into our portfolio construction and optimization process) focus on estimating the expected (average) returns of asset classes over time, we also explicitly estimate the expected fluctuations (volatilities or risks) of those returns around their averages, as well as the correlations of the returns across asset classes. Assessing which asset classes will have high fluctuations and/or high correlations with other asset classes is key in seeking to avoid the Flaw of Averages in our efforts to build portfolios for Loring Ward advisors and their clients.

Our Portfolio Construction
We believe that the best practice for optimal portfolio design is the well-known technique of Mean-Variance Optimization (MVO). Portfolio MVO is a valuable tool, but academic and professional research reveals that naïve application of MVO can amplify adverse consequences due to errors in estimating average asset class returns. So we apply several sophisticated techniques to build our model portfolios, including Resampled MVO, a form of financial simulation to model the tens of thousands of portfolio outcomes that might possibly occur. We also use behavioral constraints, such as caps on allocations to selected asset classes, to help minimize the tendency of a naïve MVO to “run away with” a single strong asset class.

Our Planning Tools
We make extensive use of Monte Carlo Simulation to help clients better understand the potential strength of their portfolio over time with respect to inflation, current savings rate, planned spending rate and capacity for financial risk. We are beginning to experiment with Dr. Savage’s SIP data arrays in this area of client portfolio analysis with the hopes of potentially designing simulation tools of the future.

The importance of understanding the strength of a portfolio over time cannot be overstated. If simulation suggests that a client’s portfolio has a high risk of exhausting itself in the future, it is important for the advisor and client to see the probability of this reality early, when steps can be taken to reduce the risk of financial failure. Good planning tools can help clients make better choices on variables like savings rate, spending rate and asset allocation to meet long-term financial goals. With Sam’s help, we are working on making meaningful improvements to our tools, and we look forward to sharing updates with you.

Our Ongoing Portfolio Oversight
We recommend a series of general threshold rebalancing guidelines on asset classes to maintain the appropriate risk profiles for client portfolios over time. These guidelines have been developed based on extensive simulations that incorporate many different market scenarios that could occur in each asset class within a model portfolio.

Our experience in the financial markets and in working with client portfolios has taught us that Sam’s warnings about the Flaw of Averages are real. At Loring Ward, we have built systems and procedures to design, build and maintain portfolios to help our clients meet their long-term financial goals.

We anticipate that Sam’s expertise may guide us further in many other areas. The PSR Think Tank works independently of our Investment Committee to complement our efforts with further diversification in expertise. Our collective goal is for continuous improvement and innovation as we design, build, and protect portfolios.

Model portfolio performance does not reflect the impact of all material economic and market factors might have had on an advisor’s decision making if clients’ money were actually managed at that time. Model performance is hypothetical and for illustrative purposes only.

17-226

 

Tweet about this on TwitterShare on LinkedInShare on FacebookEmail to someone