1/ 10/2000

 

Some Thoughts On 401K Plans

and the Active versus Passive Decision

by

Frank Sortino and Hal Forsey

The purpose of this paper is to present research ideas.  None of the information shown in this research paper should be considered an endorsement or recommendation of any mutual fund or investment strategy.  Many assumptions were made that could invalidate the results.   Anyone acting on the statistics shown here does so at their own risk.

 Many people are offering advice to 401K plan participants.   One of the best is Bill Sharpe's Financial Engines Company.  Those I respect least recommend a 100% allocation to the S&P 500.  Why stop there.  Why not just invest in the top ten stocks in the S&P 500, or better yet, the best stock.  The question is not, “should one allocate funds to fixed income or foreign securities?  The question is how much to allocate to each. That requires financial planning. There was a time in the mid to late 70’s when some pundits were questioning the logic of investing in stocks because they had underperformed bonds for a decade.  Quite a few pension funds actually decreased their holdings in equities relative to fixed income just when they should have been adding to them.  The opposite phenomenon is occuring now.   Building a strategy for the next decade based on what worked over the past decade is not prudent in our opinion.  

Another popular approach is called the "Life Style Allocation", which simply puts the same percentage in fixed income as your age, e.g., if you are 40 years old, put 40% in fixed income.  It is understandable that unsophisticated users want something that is simple to use.   But that doesn't mean the underlying methodology has to be simple minded.  The Life Style approach makes the ludicrous assumption that all 40 year olds have the same goals and the same amount of wealth to achieve the goal.   FinancialEngines is proof that it is possible to hide complex statistical methodology from the end user while letting them benefit from rigorous analytics.   While we profer a different methodology than Bill Sharpe, we respect the work he has done.  For our approach to asset allocation, see the article titled "The Dutch Triangle"in the Journal of Portfolio Management, Fall 1999.

 To repeat, 401K plan participants need more than a simplistic "do it yourself" model to assist them  in this extremely important decision.  Ideally each 401K investor would have the opportunity to pay a small fee for individual financial planning.  At minimum, they should have internet access to a software program that would ensure a prudent strategy for three or four different categories of investors.  PRI  has developed a program called SAM for the latter.  The input to SAM is controlled by sophisticated professional investors using a model called SetSAM.  The output requires the 401K participant to answer only two questions.  For more information on SAM or SetSAM  return to the home page and click on Questions.

This paper assumes the asset allocation decision has already been determined in SetSAM.  The question to be addressed is how to make the choice between active and passive equity managers.  The most popular tool for making this decision is the ratio shown in Exhibit 1.

 

Exhibit 1

 

Info Ratio.gif (15854 bytes)

 

 

 

At first glance, this seems to solve the problem nicely.  The numerator measures the excess return the active manager earned over a passive index. In CAPM terms this is the manager’s alpha.  The term in the denominator is the standard error of the alpha.  But what does that mean?  If the manager’s portfolio looks exactly like the index, there will be no tracking error.  But, there will also be no excess return in the numerator.  In this instance, the information ratio would identify a closet indexer who is not worth the active management fee.  This information appears valuable for reducing costs.    

 But, what is the manager being measured against?  The most popular index for this ratio is the S&P 500.  Most managers do not have the same composition as the stocks in the S&P 500.  To the extent the manager’s stocks behave differently than the stocks in the S&P 500, there will be a tracking error.  In fact, the only way there would not be a tracking error is if the manager consistently performed better (or worse) by the same exact percentage, e.g., always beat the index by 3%.  How likely is that?

 

 FLAWS

Suppose a manager was astute enough to avoid investing in Japanese banks in the ‘90’s.  That manager could have beat the index for Japan by an average of 800 basis points per year.  However, the tracking error would have been huge, which would have made the information ratio smaller than that of managers who held weak companies in a weak industry all the way to the bottom.  Could this happen? It did.  Could this cause a pension investment officer with an expensive consultant to reject such a manager?  It did.  The manager in question works with Invesco.  He beat the Japan index every year for eight years; sometimes by as much as 1300 basis points, and sometimes by as little as 300 basis points.   The result was a very large tracking error caused by excellent management.

 Let’s return to the S&P 500 index.  Surely this is not the only passive index available to pension funds.  Many firms offer a finer decomposition into large cap, small cap, value, and growth.  Why should every manager be compared to one index when a blend of passive indexes would better explain the source of their returns (and risk)?  Because, the information ratio derives from the CAPM, which says all investors want some combination of one risky asset (the S&P 500 is a surrogate for the market) and one riskless asset.  The goal of the CAPM is to solve the investment problem for all investors simultaneously.  Is that your goal?  The CAPM assumes equilibrium prevails in the world.  Is that your world?  You need to be clear on what it is you are trying to accomplish, and the environment in which you are operating.   

Would replacing the S&P 500 with a blend of indexes solve the problem?  There is a more fundamental problem with the information ratio.  It is not directly related to the goal of the investor.  If your goal is to retire at age 65, then investing in a market portfolio is merely one investment vehicle you might choose to accomplish your goal.  Beating the market is not your goal. Beating the market may or may not accomplish your goal.  If you must earn 8% to accomplish your goal and the market is down 10%, but your manager is only down 5%, the manager still incurred risk of not accomplishing your goal.  Given the assumptions of the CAPM, the information ratio appears to be relevant for everyone.  For the investor who can identify an MAR that will accomplish his or her goal, looks can be deceiving.

THE SOLUTION

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

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

III.               Find managers who can beat their style benchmarks.

IV.              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 1/10th  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. 

 

PERFORMANCE RELEVANT TO YOUR GOAL

As explained in "The Dutch Triangle" [1999], there is some return that must be earned at minimum in order to accomplish one’s goal.  If that minimal acceptable return (MAR) is not in the equation, the equation is not relevant to your goal.  One way to accomplish that is to use the Sortino ratio.

Sortino Ratio.gif (5534 bytes)

 

The R in the numerator is the realized return of the manager for some interval of time.  In this study the interval is five years.  The R in the denominator is only for returns lower than the MAR.  Since this ratio uses the manager's data, estimates are limited to relatively short term intervals.  We try to overcome this time sensitivity problem by using the bootstrap procedure developed by Bradley Efron at Stanford University.  If the Sortino Ratio is used, one could then use the Omega excess return to see if the manager beat the passive benchmark. 

Of course, one could substitute returns on the manager's style benchmark for the manager's data.   Instead we propose a new performance measure based on behavioral finance studies.  It is called the Upside Potential Ratio (U-P ratio).  For Details see the article on this Web site.

u-pratio.gif (11211 bytes)

 

The R in the numerator is not the manager's realized return.   It is calculated from the returns of the manager's style benchmark for 20 years or more.  The returns are bootstraped to generate a distribution of returns that could have happened.  Only those returns above the MAR are recognized in the numerator and only returns below the MAR are counted in the denominator.  The implicit assumption is that investors should focus on the average return above the MAR for the manager’s style benchmark, because it should be a more reliable predictor of future performance.   Consistency of results that are relative to your goal are what matters, not who had the highest return relative to other managers. 

 

Empirical Results

 The funds in the far left column of Table 1 are ranked by their U-P ratio.  The funds in red are ranked by their Sortino Ratio using a five year interval ending 11/99.  The rankings are quite different for the two performance measures.  In general the funds ranked by the U-P ratio are more conservative than those ranked by the Sortino ratio.  The style profile of Janus Investment Twenty is 80% aggressive growth, while Washington Mutual has 84%   evenly balanced across large value, small value, and aggressive value. Growth stocks have been in favor the past five years, but value stocks have outperformed growth over the last 25 years.  So what do you do with this information?  Those who believe in reversion to the mean might tend to favor the U-P ratio results.   Those who believe momentum is the best predictor of future performance might favor the Sortino Ratio.

 

                                                        Table 1

Omega 5 year Sortino
Fund U-P Ratio Excess R-squared Sortino Ratio Rankings Return Ratio
Washington Mutual Investr 189% 3.0% 96% Janus Inv Twenty 40.3% 3.82
Vanguard Windsor II 185% -0.8% 96% Fidelity Aggressive Growth 36.4% 3.68
Vanguard Windsor 184% -3.3% 87% Fidelity Dividend Growth 27.5% 3.56
Putnam Growth & Income/A 177% -2.4% 97% T Rowe Price Science&Tech 35.1% 3.53
Fidelity Adv Grth Opp/T 177% -2.9% 93% Vanguard Instl Index/Ist 27.5% 3.33
MFS Mass Inv Trust/A 177% 0.0% 96% Vanguard 500 Index 27.4% 3.3
T Rowe Price Growth & Inc 175% 0.3% 94% BGI Masterworks S&P 500 27.0% 3.26
Dreyfus Appreciation 175% -1.5% 95% Fidelity Spart US Eq Indx 27.2% 3.26
Amer Cnt Income & Gr/Inv 174% 1.1% 98% SSgA S&P 500 Index Fund 27.2% 3.26
BGI Masterworks S&P 500 174% 0.0% 100% Vanguard Growth & Income 27.5% 3.25
Fidelity Dividend Growth 173% 4.3% 89% Amer Cnt Income & Gr/Inv 26.8% 3.25
SSgA S&P 500 Index Fund 173% 0.2% 100% T Rowe Price Eq Index 500 27.2% 3.25
Investment Co of America 173% 0.3% 98% Dreyfus Basic S&P 27.1% 3.24
Vanguard Instl Index/Ist 173% 0.6% 100% Prudential Stock Index 26.8% 3.19
Vanguard Growth & Income 173% 1.6% 98% Dreyfus Appreciation 26.0% 3.15
Vanguard 500 Index 172% 0.4% 100% Putnam Investors/B 28.1% 3.14
Fidelity Spart US Eq Indx 172% 0.2% 100% MFS Mass Inv Growth Stock/A 32.1% 3.13
T Rowe Price Eq Index 500 172% 0.2% 100% Washington Mutual Investr 23.0% 3.13
Fundamental Investors 172% 2.5% 95% Fidelity 25.6% 3.12
Aggressive Value_US 172% 0.0% 100% T Rowe Price Blue Chip Gr 27.0% 3
Prudential Stock Index 172% -0.1% 100% MFS Mass Inv Trust/A 23.9% 2.97
Dreyfus Basic S&P 172% 0.1% 100% Janus Inv Janus 28.5% 2.96
Fidelity Equity Income 171% -1.8% 96% Fidelity Grth & Inc 23.9% 2.95
Federated Amer Ldrs/A 170% 1.0% 95% Fundamental Investors 23.2% 2.93
Federated Equity Inc/A 169% 0.9% 95% Fidelity OTC Portfolio 30.5% 2.91
Fidelity Equity Income II 168% -4.6% 94% Investment Co of America 22.6% 2.88
Merrill Basic Value/A 168% 1.1% 94% Fidelity Advisor Equity Growth 28.9% 2.86
Fidelity Grth & Inc 167% -0.9% 96% Galaxy Eqty Growth/Tr 25.8% 2.85
Fidelity 166% -0.5% 95% Vanguard U#S# Growth 28.9% 2.83
AXP Stock 165% -6.5% 95% AXP New Dimensions 26.5% 2.82
Galaxy Eqty Growth/Tr 165% -1.7% 95% Fidelity Growth Company 29.5% 2.78
Amer Cnt Select/Inv 163% -3.1% 92% Fidelity Contrafund 25.5% 2.77
Norwest Growth 163% -2.9% 93% Amer Cnt Select/Inv 25.0% 2.77
Neuberger Partners Fund 163% -1.1% 94% Federated Equity Inc/A 21.7% 2.76
Scudder Growth & Income 162% -0.2% 92% Aggressive Growth_US 35.1% 2.7
Putnam Investors/B 162% -1.5% 94% Putnam OTC Emerg Grth/A 29.2% 2.68
Dodge & Cox Stock 161% 2.9% 92% Vanguard PRIMECAP 29.4% 2.67
Franklin Mutual 160% 0.7% 84% Putnam New Opportunity/A 28.7% 2.65
Fidelity Blue Chip Growth 160% -3.7% 92% Large Growth_US 33.0% 2.61
AXP New Dimensions 160% -1.9% 96% INVESCO Dynamics 29.7% 2.6
T Rowe Price Blue Chip Gr 159% 1.4% 97% Large Value_US 27.0% 2.59
T Rowe Price Growth Stock 159% -0.8% 95% T Rowe Price Growth Stock 23.8% 2.58
American Mutual 158% 1.4% 95% MFS Capital Opport Fund/A 28.9% 2.56
Fidelity Magellan Fund 158% -1.4% 89% Founders Growth 26.6% 2.53
Fidelity Retirement Growth 156% -7.4% 87% Fidelity Magellan Fund 24.8% 2.52
Prudential Equity Inc/B 156% -3.6% 86% Fidelity Blue Chip Growth 24.1% 2.51
Neuberger Guardian Fund 155% -10.5% 85% Fidelity Adv Grth Opp/T 20.5% 2.49
Neuberger Genesis Fund 155% -1.7% 64% AXP Growth 27.1% 2.44
MFS Research Fund/A 154% -2.7% 92% AIM Value Fund/A 25.8% 2.43
T Rowe Price Spect Growth 154% -1.7% 91% Putnam Vista/A 27.0% 2.41
T Rowe Price Equity Inc 154% 3.3% 93% Putnam Voyager/A 27.1% 2.4
Fidelity Disciplined Equity 153% -2.9% 92% Norwest Growth 22.8% 2.37
Janus Inv Janus 153% 0.9% 83% Amer Cnt Growth/Inv 25.7% 2.37
MAS Value/Inst 153% -2.1% 88% Amer Cnt Ultra/Inv 26.5% 2.34
Baron Asset Fund 152% -3.4% 73% New Economy 24.8% 2.3
Fidelity Value 152% -4.1% 79% Vanguard Windsor II 20.7% 2.28
Fidelity Advisor Equity Growth 151% 0.8% 90% MFS Research Fund/A 23.7% 2.26
Oakmark Fund 151% -2.7% 80% Growth Fund of America 26.5% 2.25
Fidelity Contrafund 150% 2.2% 81% Fidelity Disciplined Equity 23.0% 2.25
Large Value_US 150% 0.0% 100% Vanguard Morgan Growth 26.9% 2.24
Amer Cnt Growth/Inv 150% -4.5% 85% Putnam Growth & Income/A 19.8% 2.18
Fidelity Low Priced Stock 149% -1.6% 74% MFS Emerging Growth 25.4% 2.15
Capital Income Builder 149% 0.3% 79% Federated Amer Ldrs/A 21.3% 2.15
Founders Growth 148% -0.1% 88% Merrill Basic Value/A 20.1% 2.11
AIM Value Fund/A 147% -1.3% 86% American Mutual 17.9% 2.07
Fidelity Growth Company 147% 1.8% 82% Franklin Stgc Sm Cap Gr/A 27.3% 2.07
Fidelity Aggressive Growth 147% 6.2% 68% Cap Research AMCAP 23.3% 2.04
T Rowe Price Science&Tech 147% 3.7% 61% T Rowe Price Growth & Inc 18.3% 2.04
MFS Mass Inv Growth Stock/A 146% 4.9% 89% Fidelity Equity Income II 19.3% 2.03
Income Fund of America 146% 1.5% 94% Fidelity Equity Income 20.1% 2
Fidelity OTC Portfolio 146% 2.6% 74% AXP Stock 19.2% 2
AXP Growth 146% -0.5% 85% T Rowe Price Equity Inc 19.0% 1.9
Amer Cnt Ultra/Inv 146% -4.2% 82% Dodge & Cox Stock 20.8% 1.89
Putnam New Opportunity/A 145% 0.7% 74% Fidelity Retirement Growth 21.0% 1.87
Small Value_US 145% 0.0% 100% Franklin Mutual 17.7% 1.84
Putnam Vista/A 145% 0.0% 77% Aggressive Value_US 21.0% 1.76
Vanguard U#S# Growth 144% -1.5% 96% T Rowe Price Spect Growth 18.7% 1.74
Putnam Voyager/A 144% 1.3% 86% Neuberger Partners Fund 19.7% 1.73
Amer Cent Value 144% 1.3% 90% T Rowe Price Mid-Cap Grth 23.8% 1.69
Putnam OTC Emerg Grth/A 143% -3.6% 55% Capital Income Builder 14.5% 1.69
Merrill Growth Fund/A 143% -9.5% 44% Baron Asset Fund 20.2% 1.66
PBHG Growth 143% -13.7% 53% Scudder Growth & Income 18.1% 1.63
Franklin Balance Sheet 142% -0.8% 65% Income Fund of America 14.7% 1.58
Cap Research AMCAP 142% -0.5% 91% Vanguard Windsor 17.4% 1.57
MFS Emerging Growth 141% -2.1% 72% Small Growth_US 22.1% 1.49
New Economy 141% 2.2% 85% Vanguard Idx Ext Mkt/Inv 20.4% 1.49
Vanguard Idx Ext Mkt/Inv 141% -2.0% 86% PBHG Growth 19.9% 1.44
Vanguard PRIMECAP 141% 5.2% 82% T Rowe Price New Amer Gr 21.1% 1.43
Growth Fund of America 139% 1.5% 86% AIM Constellation Fund/A 19.7% 1.32
INVESCO Dynamics 139% 5.3% 78% T Rowe Price New Horizons 19.9% 1.25
Janus Inv Twenty 139% 6.0% 79% Neuberger Genesis Fund 16.3% 1.22
Vanguard Morgan Growth 136% 2.4% 93% Fidelity Low Priced Stock 15.4% 1.21
AIM Constellation Fund/A 133% -6.0% 79% Amer Cent Value 17.3% 1.12
T Rowe Price New Amer Gr 133% -4.1% 78% Prudential Equity Inc/B 15.6% 1
Vanguard Explorer 130% -9.1% 71% MAS Value/Inst 15.7% 0.97
Franklin Stgc Sm Cap Gr/A 130% 3.2% 65% Small Value_US 16.2% 0.95
T Rowe Price Mid-Cap Grth 130% 0.9% 79% Franklin Balance Sheet 13.3% 0.94
T Rowe Price New Horizons 130% -4.9% 72% Neuberger Guardian Fund 14.9% 0.91
Small Growth_US 128% 0.0% 100% Oakmark Fund 14.1% 0.81
Aggressive Growth_US 127% 0.0% 100% Vanguard Explorer 15.4% 0.78
MFS Capital Opport Fund/A 124% 9.9% 86% Fidelity Value 13.9% 0.72
Large Growth_US 68% 0.0% 100% Merrill Growth Fund/A 12.6% 0.61
Bond_US 1% 0.0% 100% US_Bond 7.9% -0.17
3 Month T-bills_US -99% 0.0% 100% Money market 5.2% -1.99

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.  Based solely on its raw returns for the last five years, Washington Mutual would not have made the first, or even the second quartile in this analysis.  Using manager data to calculate a Sortino Ratio (shown in red) over a five year interval it ranked 18th.   But, based on twenty years of style data, Washington Mutual has 93% more upside potential than downside risk (U-P ratio = 193%).  The U-P ratio  presumes the investor has an MAR of 8.6%.  Notice the T-bill rate shown here has 99% more downside risk than upside potential.  It is not a low risk asset for investors who must earn 8.6% or more to accomplish their goal.

 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.  For details on the Omega return see the article by that title under published papers.

 

Omega Excess 2.gif (16483 bytes)

 

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.  For example, Windsor fund is overweighted in finance stocks and tobacco, causing it to underperform the passive benchmark.  

 The U-P ratio identifies which passive set of indexes could provide the most upside potential for the downside risk.  The Omega excess identifies who should beat the passive indexes. Those funds with an Omega excess of 100 basis points or more are shown in blue. 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 article titled "Managing Uncertainty" on this web page.

 

 The Allocation Question:

The performance analysis above 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 shown in Table 2 assumes one ranks by the U-P ratio.  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.  Unlike models that hold risk constant and add alpha, SAM ensures that the mix of styles will be held constant while adding Omega.  

 

                                                                         Table 2

Style allocation.gif (11173 bytes)

Sam only allows those funds that are in the upper half and have an Omega excess of 100 basis points or more to be considered in the final solution.   Suppose one allowed Janus to be considered in the asset allocation.  The results are shown below.

  Table 3

Janus allocation.gif (12810 bytes)

 

American Mutual is replaced by Janus and other positions are reduced to allow a 4.7% allocation to Janus.  This demonstrates that SAM can accommodate both reversion-to-the-mean and momentum philosophies while maintaining a well diversified portfolio. 

Although I believe strongly that international stocks should be a part of most portfolios, to simplify the analysis, I did not include them.

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 SetSAM, a proprietary product of the Pension Research Institute that provides the input to SAM. 

The purpose of this paper is to present research ideas.   None of the information shown in this research paper should be considered an endorsement or recommendation of any mutual fund or investment strategy.  Many assumptions were made that could invalidate the results.  Anyone acting on the statistics shown here does so at their own risk.