March 17, 2009
Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives
As usual, Bob Veres wrote a very insightful article last week questioning whether advisors who jumped1 out of the market sometime before November - and in some cases before September - of 2008 were skillful or lucky. Veres’ article dovetails with a recent article about research by Sebastein Page and his colleagues (Why Diversification is Failing), questioning claims by advisors espousing market timing strategies. Although it would be very helpful to see some actual verification of how these advisors used the economic and valuation data in 2008 to properly time the market, the reality is most at the moment look like geniuses.
As a mathematician and statistician, I have written about the shortcomings in the manner in which the financial services industry calculates risk and reward. In fact, since founding Carolopolis, our asset allocation strategies have rejected Modern Portfolio Theory (MPT) ideas, including mean-variance and correlation. Because we were aware events such as 2008 can and will occur, the option to ‘jump’ seems reckless and counterintuitive to our primary function of assisting clients in achieving their goals.
Using correct risk and return estimates can assist advisors in auditing their asset allocation strategies and demonstrating their skill and value to clients in today’s volatile environment.
Conceptualizing risk from an investor’s perspective
Since Harry Markowitz first wrote Portfolio Selection in 1959, the investment industry has used the standard deviation of a set of returns as a proxy for risk. But, from a mathematical standpoint, standard deviation is a measurement of volatility and the calculation assumes returns are always normally distributed. Applying such methodology to the monthly returns of the S&P 500 would indicate a loss greater than 12.8% in a single month would have nearly no chance of happening, yet such declines have occurred 11 times since 1926. Because MPT concludes that events as 2008 are so improbable they only occur once a century, is it any wonder that advisors claim this crisis was unique, unprecedented or unforeseeable?
Since standard deviation is not a measure of risk, we cannot identify risk with it. In fact, standard deviation flies in the face of how investors perceive risk – volatility on the upside is reward and volatility on the downside is risk. When posed the question, investors consistently rank risk as:
- Not being able to achieve a goal;
- Losing principal; and finally
- Not beating the market
In Portfolio Selection, Markowitz recognized the lack of ability of standard deviation to provide accurate insight. He alluded to the use of semi-variance as far superior to his mean variance approach, but the computational methods necessary for the semi-variance approach were nearly impossible in the 1950s. Semi-variance or downside risk only considers observations below the mean. While standard deviation provides measures of volatility, semi-variance looks only at the negative fluctuations of an asset. By neutralizing all values above the mean, semi-variance estimates the loss that a portfolio could incur.
Semi-variance would have indicated a loss of 37.68% for a given 12 month period was not unexpected before 2008. Today, a monthly pullback of 10.99%, as indicated by semi-variance, is well within the normal range for the S&P 500.
A third measure of risk - downside deviation - allows us to customize asset allocation and manager selection decisions to clients’ goals. While semi-variance calculates the downside risk based on the average returns of an asset, downside deviation considers returns below a fixed level – the investor’s Minimal Acceptable Return (MAR). The investor’s MAR is the absolute annualized return that must be reached in order to achieve a specific goal.
One great benefit of downside deviation is in verifying expected returns. Many financial planning and asset allocation programs use the average annual return of the S&P 500 to illustrate future returns and success probabilities. From 1970 to 2007 the average annual return of the S&P 500 was 8.78%; however, 45% of the time returns were less than that and the actual returns (annualized) were 7.72% - a full 106 basis points lower than the average return. Over 30 years, that could ruin a retirement income plan!1 Veres defines ‘jumped’ as reducing client equity exposure to 20% or less.
Display article as PDF for printing.
Would you like to send this article to a friend?
Remember, if you have a question or comment, send it to .
