Selected Articles From Pensions & Investments

Who's Goal Are You Pursuing?  (Published 4th Quarter 1998)

The goal of DB plans is to fund their plan within their cost constraints. The goal of a 401K participant is to retire at a certain age. In both cases there is some return that must be earned at minimum to accomplish the investor's goal. If this Minimal Acceptable Return (MAR) is missing from your performance measure, the chances are you are helping someone else accomplish their goal, not yours.

An example of this problem can be seen in the "information ratio". The numerator consists of the active manager's return minus the return on a passive index. It provides information on whether the manager earned more than the passive index. Is that the goal of the investor, to beat an index? No, that's the goal of active portfolio managers, because that's how they get business from consultants. Your goal is not to get business for them. Beating an index may or may not accomplish your goal.

What is in the denominator of the information ratio? The deviations of the same thing that is in the numerator. It conveys information about whether the active manager is tracking the index. Suppose you are an active manager that consistently gets higher returns than the index. You will have a terrible tracking error. The more you outperform the index the worse you will look. That could result in firing someone like Warren Buffett or Tony Browne. Of course, there aren't many managers like these, so the real problem is, the denominator does not contain the MAR. Therefore, tracking error cannot measure the risk of not accomplishing the investor's goal. The information ratio conveys information that is relevant for most portfolio managers, but not individual investors who have to earn some MAR to accomplish their goal.

An alternative is the Omega Excess return, which is directly related to the investor's goal. It tells investors whether the manager beat a passive set of indexes after adjusting for the risk of not accomplishing their goal. It measures downside risk as deviations below the MAR. Furthermore, it factors in the investor's degree of risk aversion and the downside risk the manager took relative to the downside risk of a style benchmark. It seems to me, these are all things an investor should want to know. For example, the Janus 20 Fund had a five year unadjusted return of 29.6%. The downside beta indicates the fund took 10% less risk of falling below the MAR of 8.6% than a passive set of indexes that attempts to replicate the funds style. The Omega return says, after adjusting for all the things mentioned above, it earned 26.4%. The Omega Excess indicates the fund earned 2.6% more than a passive strategy. The R squared indicates the passive strategy accounts for 83% of the returns of the fund. While the calculations may be difficult, the numbers are easy to interpret.

Did it work better in the past than other performance measures? A study conducted by Bernardo Quan of DAL Investment Co. in San Francisco showed that Omega Excess returns had more predictive power than the Sharpe ratio, the Sortino ratio, or the information ratio for the 15 year period ending December 1997. (See Other Research on Home Page for details). Also, it has a low correlation with the Sharpe ratio and the Information Ratio, indicating there is information in the Omega Excess not contained in the others. But, as reported last quarter, the top omega funds did not do better than the S&P 500 in the horrible third quarter of '98. Although, they did beat the median managed equity account by 430 basis points (The median return reported on P2 of the December 14, 1998 was -14.3%).

Gambling vs. investing  (Published 4th Quarter 1999)

It’s No Joke

 “Doctor, it hurts when I do this”…’well, don’t do that’, goes the old Henny Youngman joke.   Many investors get hurt making decisions based on short term results.  The purpose of this article is to say, “don’t do that”, and offer a cure for the affliction.

 Most performance studies are based on manager data over a short time interval of five years or less.  Weekly publications highlight winners and losers for the week, monthly publications do the same for twice as long.  Academic studies look way out to five years, as if that were a long time.  Consequently, investors are forced to choose managers based on a short term interval.  The result is, they buy at the top of the current fad.  The WM Company in Scotland did an analysis of managers their clients fired versus the managers they hired.  You guessed it, most of the managers they fired, subsequently outperformed the ones they hired.

 It’s not enough to say, “don’t do that”, there needs to be a better basis for selecting managers in the first place.  I suggest the following:

 1.       Focus on the manager’s style, not the manager’s returns.

2.      Identify the upside potential of this style relative to its downside risk.

3.      Find managers who can beat their style benchmarks.

4.      Base judgments on what could have happened, not what did happen.

Over 80% of the returns for almost all managers can be explained by a combination of passive indexes (a style benchmark) one could purchase at 1/10^th  the cost of active management.  Data is available on these indexes for 20 years or more.  This provides much more stable estimates than the short term results of the manager.   So…don’t do that (use short term results of the manager), do this (use long term results of the manager’s style benchmark).

 Investors should focus on the highest average return for each manager’s style benchmark above the return they need to realize their goal.  Consistency of results that are relative to your goal are what matters, not who had the highest return last week relative to other managers.  I call this the upside potential.  The upside potential divided by the risk of not accomplishing your goal is called the Upside Potential Ratio (U-P ratio).

 The funds in Table 1 are ranked by their U-P ratio.  Washington Mutual has 93% more upside potential than downside risk, i.e., the upside potential is 193% of the downside risk.  The R squared for Washington Mutual indicates that 96% of the returns for this fund can be explained by a combination of style indexes (identified with a _US ending). Yet, it ranks higher than any of these individual indexes.  Who says diversification doesn’t matter.  It is interesting that most of the funds ranked higher than the individual style indexes but lower than the S&P 500.  

 Based solely on its returns for the last five years, Washington Mutual would not have made the first, or even the second quartile in this analysis.  But, based on twenty years of style data, Washington Mutual has more upside potential for the downside risk than any other mutual fund (see U-P ratio).  This ranking procedure presumes the investor is more interested in consistency of performance above 8.6%, than hiring the manager with the highest return last quarter, and more concerned about downside risk than deviations about the mean. 

 After identifying managers with the highest U-P ratio, one must determine if the manager can beat the style benchmark on a risk-adjusted basis.  This number is given by the Omega excess return.  Washington mutual earned 300 basis points (3%) more each year than its style benchmark after adjusting for downside risk.  The negative returns in the Omega excess column support the view that most funds cannot beat a passive set of indexes.

 The U-P ratio identifies which style could provide the most upside potential for the downside risk.  The Omega excess identifies who should beat the passive style indexes. Those funds that satisfy both criteria are shown in bold italics. What a manager could do, or should do, is determined by using the bootstrap procedure developed by Bradley Effron at Stanford University.  This procedure is explained in the U-P Ratio article on this web site. 

The Asset Allocation Question:

This performance analysis does not indicate how much should be allocated to each asset category, or how much should be allocated to active versus passive management.  The answer to these questions requires an asset allocation model. 

 I used a model called SAM to find the solution for an investor who wanted 30% fixed and 70% equity, with approximately the same mix of styles as the S&P 500.  The solution is shown in the Chart titled Asset Allocation.  SAM allocates more to Fidelity Dividend Growth than Washington Mutual mainly because of the constraint to maintain approximately the same mix of styles as the S&P 500. Fixed income totals less than 30% because some funds hold cash.

 The Cure:

Investing in mutual funds for your retirement should not be exciting.  Speculating is exciting.  Gambling is even more exciting.  So, if your palms are sweaty every time you check the financial page, this is the doctor saying, “don’t do that”.  Diversify until the sweating stops.

 The style indexes in this analysis were provided by Independence International Associates (IIA) in Boston.  Fund data was provided by LCG associates in Atlanta.  The statistics were generated by SAM, a proprietary product of the Pension Research Institute.

Home | Published Papers | Work in Progress | Past Conferences | Future Conferences
Contents Copyright© 1998, Pension Research Institute