The Problems of Past Performance

1

• The Problems of Past Performance

11-06-98
2
The Fallacy of Past Performance as A Predictor of
Future Performance

• Percentage of Lipper Funds Performing in the Top
  25 That Repeated in the Top 25 the Following
  Year

Source Lipper Analytical Services, Inc.Funds
included in the Lipper Growth Universe normally
invest in companies whose long-term earnings are
expected to grow significantly faster than the
earnings of the stocks represented in the major
unmanaged indices. Funds included in the Lipper
Small Universe limit their investments to
companies on the basis of the size of the
company. Performance is historical and does not
guarantee future results. Information derived
from sources deemed to be reliable, but its
accuracy cannot be guaranteed.
02-23-99
3

• The Fallacy of Past Performance as a Predictor of
  Future Performance

47.9
1st Quartile
2nd Quartile
52.1
50.0
3rd Quartile
4th Quartile
50.0
1981-1985
1986-1990
1981-1985
1986-1990
Source SEI Investments Funds Evaluation
database.



The database represented institutional money
managers of domestic equity pools of assets.
Performance is historical and does not guarantee
future results. Information from sources deemed
to be reliable, but its accuracy cannot be
guaranteed.
02-17-99
4

• The Fallacy of Past Performance as a Predictor of
  Future Performance

Source Micropal? (excludes international,
balanced and specialty funds).
Performance is historical and
does not guarantee future results.
Information
from sources deemed reliable but its accuracy
cannot be guaranteed.
03-04-03
5

• The Fallacy of Market Timing

CRSP value-weighted index with dividends
reinvested. Compound annual returns assume a 1
transaction cost per portfolio turnover.
Source John D. Stowe, A Market Timing Myth.
Journal of Investing, Winter 2000.


Performance is historical and does not guarantee
future results.
Information from sources
deemed reliable, but its accuracy cannot be
guaranteed.
03-04-03
6
The Power of Diversification
Portfolio 5
Quarterly January 1973 - December 2002
The Sharpe Ratio is a measure of the
risk-adjusted return of an investment. A higher
ratio indicates a greater return for a unit of
risk. The Sharpe Ratio is calculated as the
average annual portfolio return less the average
annual risk-free rate (1 Month T-Bills) divided
by the portfolios annualized standard deviation.
Source Dimensional Fund Advisors Inc. Asset
classes represent a combination of live funds and
simulated strategies. Performance is historical
and does not guarantee future results. Portfolios
shown do not include tax-managed funds. All
Portfolios assume reinvestment of dividends and
capital gains. Annualized from quarterly data.
All portfolios rebalanced quarterly and do not
reflect investment advisory fees and fund
expenses when simulated data is used. Simulated
data is used prior to the inception of the live
portfolios. Simulated data does not reflect
deductions of fund expenses that a client would
pay. Nor do simulated returns represent results
of actual trading. See Sources and Descriptions
of Data to identify which periods are simulated
and which periods contain live data for each data
series. Live data does not reflect all fund
expenses incurred by the portfolios and
incorporates actual trading results. Both
simulated and live data reflect total returns.
Such fees and expenses will affect subsequent
performance. All material represented is compiled
from sources believed to be reliable and current,
but accuracy cannot be guaranteed.
03-04-03
7
Can You Pick the Next Winner? Asset Class Returns
1978-2002
SP 500
US Large Cap Value Stocks
US Micro Cap Stocks
US Small Cap Value Stocks
US REITS
Intl Value Stocks
Intl Small Stocks
Source Dimensional Fund Advisors,
Inc. Information from sources deemed reliable,
but its accuracy cannot be guaranteed.
Performance is historical and does not guarantee
future results.
03-03-03
8
It Doesnt Pay to Chase Yesterdays Winners

• A Morningstar study evaluated 199 no-load growth
  mutual funds for which they had performance data
  for the period 1989-19941
• Average total return for the funds was 12.011
• The average investor, holding funds for only 21
  months, received a return of just 2.021
• A study by Dalbar found that investors in
  stock-owning mutual funds earned on average 10
  less than the funds themselves in each of the 13
  years from 1984-1996.2
• Another study by Morningstar showed that in
  comparing returns for the next one, two and three
  years, the three least popular equity fund
  categories beat the 3 most popular categories
  over 90 of the time.3
• From 1987-1998, Morningstars Manager of the
  Years on average underperformed the SP 500
  Index by over 8 per annum.4
• Morningstars 1999 Manager of the Year, Jim
  Callanan of RS Emerging Growth, didnt fare so
  well the next year. His fund finished down 27
  in 2000.5 The SP 500 fell 9.1.

Popularity based on net cash inflows. 1Beyond
Stocks, John Merrill. 2St. Louis Post-Dispatch,
July 11, 1997. 3The Index Fund Solution, Richard
Evans. 4JohnBogle on Investing, John Bogle. 5Fund
Managers of the Year 2000, Russel Kinnel.
11-20-01
9

• Relying on Trade Publications
• Each year Business Week selects The 100 Best
  Small Companies (for example in its Classes of
  95 and 96)
• For the two-year period ending April 30, 1997,
  the Russell 2000, a benchmark for small
  companies, rose 33.
• Business Week reported that a Class of 1995
  portfolio would have returned -4.5.
• The Class of 1996 showed similar results
• Russell 2000 38
• U.S. Class of 96 2.6
• 10 of 100 lost 70 of their value

Source Business Week, May 27, 1997.
03-31-99
10
Relying on Rating Services

• Morningstars 4- and 5-star funds A Lipper
  Analytical Services study showed that an investor
  purchasing a portfolio consisting of only
  Morningstars 5-star funds, for the period
  1990-94, in each year would have failed to beat
  the portfolios respective asset classes1
• Morningstar Mutual Funds responseThe
  connection between past and future performance
  has not been firmly established by stars,
  historic star ratings or any raw data.1 We
  never intended to suggest that the stars could be
  used to predict short-term returns or to time
  fund purchases. They were just a way to sort
  funds according to past success.2 Our 5-star
  bond funds have posted lower aggregate returns
  than their peers.3
• Conclusion
• Morningstars star ratings have no predictive
  value. Past performance is a very poor predictor
  of future performance4

1Investment Advisor, September 1994. 2Morningstar
Mutual Funds, December 8, 1995. 3Kiplingers
Personal Finance, February 1997. 4Kiplingers
Personal Finance, February 1997.
06-28-99
11

• Relying on Stock Selection
• I believe the search for top-performing stock
  funds is an intellectually discredited exercise
  that will come to be viewed as one of the great
  financial follies of the late 20th Century.
• Jonathan Clements, Columnist Wall Street Journal

Source Wall Street Journal, April 29, 1997.
03-31-99
12
Evidence on Stock Selection Skills

• Economist Terrance Odean studied 100,000 trades
  of individual investors at a major brokerage
  firm.
• His conclusion Individual investors arent as
  bad at stock picking as many people think.
  Theyre worse.
• Buys trailed the market stocks sold outperformed
  after the sale.
• As the time span increased, the more their
  performance trailed the market.
• Investors shot themselves in the foot with their
  trades even before transaction costs and taxes.
• The more investors traded, the worse the results
  were.
• The average individual that trades most often
  underperforms a market index by over 5 p.a. On
  a risk-adjusted basis, the underperformance was
  over 10 p.a.3
• Women outperform men by 1.4 p.a. and single men
  by 2.3 p.a.1
• Investment clubs trailed broad market index by
  3.7 p.a.2

1Wall Street Journal, October 20, 1998
2Wall Street Journal, November 4,
1998 3Trading is Hazardous to Your Wealth The
Common Stock Investment Performance of Individual
Investors, Journal of Finance, April 2000. All
remaining data from Wall Street Journal, May 16,
1997.
03-22-00
13

• Supporting Research
• On Persistence in Mutual Fund Performance
• By Mark M. Carhart, University of Chicago
• Study covered 1961-1993 and 1,892 mutual funds1
• Accounting for common risk factors such as size
  and value, the average fund underperformed by
  1.8 per annum1
• Any persistence in mutual fund performance was
  explained by common factors of expenses,
  transaction costs and exposure to risk factors
• There was little evidence of stock selection
  skills
• Expenses have a 1 x 1 negative impact on returns
• Turnover costs approximate 1 of the assets traded

1Mark Carhart, On Persistence in Mutual Fund
Performance,doctoral diss., University of
Chicago, December 1994.

All
remaining data from Journal of Finance, March
1997.
03-31-99
14
Important Disclosures Regarding Simulated
Strategies
Company logo or name here

• The preceding pages include illustrations of
  returns for the types of portfolios we design for
  clients. The Simulated Strategies may or may not
  be the actual allocation determined to be
  appropriate for any individual clients. Clients
  with the allocations shown may have different
  results based on capital flows, timing of
  rebalancing decisions, fees charged or other
  factors.
• Our investment strategy is based on the
  principles of Modern Portfolio Theory (MPT.) The
  tenets of MPT provide for a passive, long-term,
  buy-and-hold strategy implemented through
  globally diversified portfolios. Mutual funds
  representing asset classes where academic
  research has demonstrated higher expected returns
  for the level of risk taken are combined in a
  single portfolio. Portfolios are constructed in
  a manner to provide diversification for the
  purpose of reducing the risk caused by
  volatility. Portfolios are rebalanced to
  maintain agreed upon asset allocations.
• The historical performance information that
  precedes this slide is provided to demonstrate
  the methodology used in building portfolios using
  the aforementioned investment strategy. This
  information should not be considered as a
  demonstration of actual performance results, and
  should not be interpreted as such. The results
  are based on back-tested data and not actual
  accounts. Past performance is not a guarantee of
  future results. The Simulated Strategies were
  begun in 1996 and have been adjusted on occasion
  over the years which resulted in minor changes in
  the simulated returns (less than 34 bps per
  annum). The investment returns and principal
  value of mutual funds recommended by our firm
  will fluctuate and may be worth more or less than
  their original cost when sold.
• In 1999, tax-managed funds became available for
  several different asset classes. We now use
  tax-managed funds extensively for taxable
  entities. While the tax-managed funds are
  consistent with the passive asset class approach
  we follow, they should not be expected to
  regularly track the performance of corresponding
  taxable funds in the same or similar asset
  classes. As such, the performance of portfolios
  using tax-managed funds will vary from portfolios
  that do not utilize these funds.
• Back-tested data does not represent the impact
  that material economic and market factors might
  have on an investment advisors decision-making
  process if the advisor were actually advising an
  investor. The back-testing of performance
  differs from actual account performance because
  an investment strategy may be adjusted at any
  time, for any reason and can continue to be
  changed until desired or better performance
  results are achieved. The back-tested results
  assume ordinary income and capital gains
  distributions are reinvested, annual rebalancing
  and no income taxes. If performance reflects the
  deduction of an advisory fee (1.85 or less)
  billed quarterly in advance, it is indicated on
  the page. More information about mutual fund
  fees and expenses is available in the prospectus
  for each mutual fund.
• The back-tested data used in creating the
  Simulated Strategies includes simulated data
  where live funds were not in existence to provide
  actual returns. All funds, except Emerging
  Markets Value, are live for five years or more.
  The simulated data was developed in the belief
  that it represents the historic performance of
  the live funds. Live funds will differ from
  simulated strategies. Simulated data does not
  reflect deductions of fund expenses that an
  investor would pay and does not represent results
  of actual trading. Analysis has shown that if
  fund expenses were included in simulated data,
  the return results over either a 10- or 20-year
  period would have been less by .18 per year.
  Live funds may exclude securities with
  characteristics not otherwise excluded in certain
  databases used in some simulated strategies. In
  2003, adjustments were made to the Fama/French
  Value Strategy data. The adjustment impacted only
  20-year returns with lower returns as follows
  100/032 bps lower, 80/20 26 bps lower, 60/4018
  bps lower and 40/6012 bps lower. Sources and
  Descriptions of Data which follow at the end of
  this presentation booklet is an integral part of
  and should be read in conjunction with this
  explanation.
•  

03-13-03
15
Sources and Descriptions of Data
Company logo or name here
01-27-03
Sources and Descriptions of Data supplied by
Dimensional Fund Advisors, Inc.
Information from sources deemed reliable but its
accuracy cannot be guaranteed.
16
Sources and Descriptions of Data
Company logo or name here
01-27-03
Sources and Descriptions of Data supplied by
Dimensional Fund Advisors, Inc.
Information from sources deemed reliable but its
accuracy cannot be guaranteed.