The Hidden Exposures Among Equity Style Indices presented to the Chicago Professional Risk Managers

1
The Hidden Exposures Among Equity Style
Indicespresented to theChicago Professional
Risk Managers International AssociationbyDr.
Peter Jankovskis, Ph.D., Director of Research

• January 9, 2003

A Different Way of Thinking
2
Introduction

• Sharp swings in the relative performance of
  growth and value managers in the US have focused
  attention on the subject of investment style risk
• A better understanding of style risk can be
  achieved by expanding the traditional
  growth/value analysis to include additional
  factors

3
OakBrook Style Index Construction

• Evaluate stocks along three lines
• Capitalization
• Value/Growth
• Price to earnings ratios
• Price to book ratios, modified to eliminate
  accounting biases
• Stability/Variability
• Average of absolute monthly returns (3 year
  period)
• Average of absolute daily returns (2 year period)
• Average trading volume relative to free float (2
  year period)

4
The OakBrook Style Indices
5
Large Cap Style Index Characteristics

• Variable Growth aggressive growth stocks
  dominated by technology firms
• Stable Growth stocks with dependable growth,
  moderate P/E features consumer and drug
  companies
• Variable Value classic cyclical stocks, such as
  banks and auto manufacturers
• Stable Value stable commodities, mainly oils
  utilities

6
Significance of Large Cap Variable vs. Stable
See performance disclosure in appendix for
details. Results are simulated and historical
relationships may fail to hold in the future.
7
Significance of Large Cap Variable vs. Stable

• Long run return differences between variable and
  stable are larger than those between growth and
  value
• Annualized tracking error of SP/BARRA Growth vs.
  Value 9.23
• Annualized tracking error of Oakbrook Variable
  vs. Stable 15.82

Based on monthly index returns from January
1985 to December 2002. See performance
disclosure in appendix for details. Results are
simulated and historical relationships may fail
to hold in the future.
8
LargeCap Style Indices Characteristics
(12/31/02)

• See performance disclosure at the end of the
  appendix for details.
• The risk measures used in the table above are
  defined as follows
• Stdev. is the annualized standard deviation of
  quarterly returns from January 1998 through
  December 2002.
• Beta is the beta of quarterly return relative to
  the SP 500.
• Tracking is the annualized tracking error of
  quarterly return relative to the SP 500.
• R-square is the regression r-square of quarterly
  return relative to the SP 500.

9
Analytic Applications

• Increase understanding of macroeconomic forces
  driving returns
• Identify multiple growth and value manager styles
• Distinguish style drift from benchmark drift
• Identify style biases
• Clarify the growth/value small cap style anomaly

10
Macroeconomic Forces Driving Returns

Impact of Macroeconomic Factors on Excess Return
Growth and Value Stereotypes
SP/BARRA
SP/BARRA Macro Factor
Growth Value
Strong profit growth -
Steep yield curve
-
Strong equity returns
- Rising long bond yields
-
11
Macroeconomic Forces Driving Returns

• Some growth and value stereotypes may be variable
  and stable biases in disguise

Impact of Macroeconomic Factors on Excess Return
After Considering Variability and
Stability Macro Factor
Growth Value
Strong profit growth -
Steep yield curve
-
Variable1
Stable2 Strong equity returns
- Rising
long bond yields
-
1 - Positive relationship for Variable Growth and
Variable Value 2 - Negative relationship for
Stable Growth and Stable Value
12
Identifying Manager Styles

• Use return-based style analysis
• Describes a managers returns as a combination of
  the returns on various benchmark style indices
• Some of the benchmarks will have a stronger
  correlation with the managers returns than
  others
• The index with the strongest correlation with a
  managers returns identifies the managers style

13
Identifying Manager Style Russell Indices
Value managers that appear to have the same
style...
14
Identifying Manager Style OakBrook Indices
may in fact be very different.
15
Identifying Manager Style Russell Indices
Similarly, growth managers may appear to have the
same style...
16
Identifying Manager Style OakBrook Indices
but may turn out to be very different.
17
Evaluating Manager Style Drift

• Requiring managers to maintain consistent styles
  is a critical part of managing portfolio style
  risk
• Return-based style analysis can be used to
  determine when a managers style has drifted
• perform several analyses, each using a different
  return period
• examine the shift in style from period to period

18
Evaluating Style Drift Russell Indices
Using the Russell Indices, this manager appears
to have changed style in recent years.
19
Evaluating Style Drift OakBrook Indices
However, this manager may have been more
consistent than the Russell Indices would
indicate.
20
Benchmark Drift Can Distort Results
The Russell Growth Index has become more
variable, requiring managers to exhibit both
growth and variability to be classified as
growth.
21
Drift in the SP/BARRA Indices
Within the SP/BARRA style family, there has also
been a substantial increase in variability.
22
Drift in the SP 500 Index
The equally weighted SP 500 has grown less
variable over time, suggesting a few large stocks
are responsible for the increase in the
variability of the cap weighted index.
23
Identifying Large Cap Style Biases
Style analysis was used to classify large cap US
equity managers in the PSN data base from January
1997 through December 2002. The results were as
follows.
24
Style Biases in the Marketplace Style
Classification of Large Cap US Equity
Managers January 1997 - December 2001
Excludes 676.2 billion of assets under
management in 65 SP 500 Index and Enhanced SP
500 Index Funds classified as growth funds by the
returns based analysis. Also excludes 48.9
billion invested in 3 funds designed to track the
US equity market that were classified as variable
growth funds.
25
Small Cap Style Indices

• The stable/variable approach yields interesting
  results in the small cap space as well.

26
Small Cap Growth/Value Return Anomaly

• Value stocks appear to offer superior long run
  returns
• The Russell 2000 Value Index out performed the
  Russell 2000 Growth Index by an average of 5.80
  annually from 1979 through 2002
• The SP 600 Value Index has out performed the SP
  600 Growth Index by an average of 4.53 annually
  from 1994 through 2002
• Stable stocks appear to offer higher returns than
  value
• OakBrook Small Cap Stable Index outperformed SP
  600 Value by an average of 3.15 annually from
  1996-2002
• A focus on stability may enhance the value
  return characteristics of a small cap index

Please see performance disclosures in the
appendix.
27
Significance of Small Cap Variable vs. Stable
See performance disclosure in appendix for
details. Results are simulated and historical
relationships may fail to hold in the future.
28
Significance of Small Cap Variable vs. Stable

• Average return difference between variable and
  stable is larger than that between growth and
  value
• Annualized tracking error of SP 600/BARRA Growth
  vs. Value 13.76
• Annualized tracking error of OakBrook Small Cap
  Variable vs. Stable 24.36

Based on monthly index returns from January
1996 to December 2002. See performance
disclosure in appendix for details. Results are
simulated and historical relationships may fail
to hold in the future.
29
SmallCap Style Indices Characteristics
(12/31/02)

• See performance disclosure at the end of the
  appendix for details.
• The risk measures used in the table above are
  defined as follows
• Stdev. is the annualized standard deviation of
  quarterly returns from January 1998 through
  December 2002.
• Beta is the beta of quarterly return relative to
  the SP 600.
• Tracking is the annualized tracking error of
  quarterly return relative to the SP 600.
• R-square is the regression r-square of quarterly
  return relative to the SP 600.

30
Portfolio Applications

• Eliminate style biases
• Add a style tilt

31
Completeness Funds

• One way to eliminate an aggregate style bias or
  create a style tilt is to add a completeness fund
  to a portfolio
• The ability to distinguish between stable and
  variable stocks often results in extremely
  efficient completeness funds, whether used to
  eliminate style biases or create them

32
Eliminate Style Biases
33
Add a Style Tilt
34
Summary

• Benefits of the Stable/Variable Approach
• Increased understanding of macro forces driving
  returns
• Sharper identification of manager style
• Better understanding of style drift and benchmark
  drift
• Improved diversification of portfolios
• Possible return enhancement / risk reduction

35
Appendix
36
OakBrook LargeCap Style Indices Returns
(1985-2002)
See performance disclosure at the end of the
appendix for details.
37
OakBrook SmallCap Style Indices Returns
(1996-2002)
Based on a computer simulation of OakBrooks
SmallCap Style Indices.
38
Performance Disclosure - OakBrook Style Indices
Performance histories for OakBrooks Large Cap
Style Indices have been calculated going back to
January, 1985. The simulated results are based on
total return calculations which include a
reinvestment of accrued dividends and price
changes. As with other indices, returns are
gross of all fees and expenses, including
management fees, brokers commissions and market
impact costs. The indices were rebalanced
bi-annually, at the end of each June and
December, with explicit controls to limit index
turnover. During each rebalancing, firms were
placed in indices based on a computer model which
evaluated four years of historical information.
Programming of the computer model was completed
in May of 1999, based on data through December of
1998. Performance histories for OakBrooks
Small Cap Style indices have been calculated
going back to January, 1996. The simulated
results are based on total return calculations
which include a reinvestment of accrued dividends
and price changes. As with other indices,
returns are gross of all fees and expenses,
including management fees, brokers commissions
and market impact costs. The indices were
rebalanced bi-annually, at the end of each June
and December. During each rebalancing, firms
were placed in the index based on a computer
model which evaluated three years of historical
information. Programming of the computer model
was completed in September of 2001, based on data
through June of 2001. As with any study
conducted after the fact, results are subject to
back testing bias that can inflate returns. The
results presented do not represent the actual
performance of any OakBrook client portfolio and
were achieved by means of the retroactive
application of a program that was designed with
the benefit of hindsight. Results may not
reflect the impact of material market or economic
factors which occurred during the back test
period. OakBrook Investments, LLC has not managed
portfolios using the methodology shown and the
results presented should not be considered a
substitute for the investment performance of
actual OakBrook accounts. Results should not be
considered indicative of the skill of OakBrooks
portfolio managers. The back test results shown
are in no way indicative of the future
performance of the OakBrook Style Indices or any
accounts managed to track those indices and
clients may experience a loss while investing in
these indices. Total fees are expected to be
less than 45 basis points per year. The indices
shown are not managed portfolios. The back
testing is for investment strategies that may or
may not be utilized in advising OakBrooks
advisory clients. OakBrooks clients who do not
utilize the strategies reflected by the Style
Indices results may receive investment advice
and obtain performance results which may differ
from those clients accounts utilizing such Style
Indices. The investment returns for OakBrooks
clients, and clients of the predecessor advisors
with which OakBrook portfolio managers were
affiliated, may be materially different from the
investment results portrayed in the Style
Indices.
39
Performance Disclosure - Other Indices
The Russell 1000 Growth and Value Indices are
registered trademarks of the Frank Russell
Company. The Growth and Value Indices are
unmanaged portfolios derived from the Russell
1000, which is itself an unmanaged portfolio of
the 1000 largest US stocks, as ranked by market
capitalization at the end of May each year. The
performance numbers cited are total returns
including the reinvestment of accrued dividends,
earnings and price changes. As with other
indices, returns are gross of all fees and
expenses, including management fees, brokers
commissions and market impact costs. OakBrook
Investments is not affiliated in any way with the
Frank Russell Company. The Russell 2000 Growth
and Value Indices are registered trademarks of
the Frank Russell Company. The Growth and Value
Indices are unmanaged portfolios derived from the
Russell 2000, which is itself an unmanaged
portfolio of 2000 US stocks with market
capitalizations that rank them below the largest
1000 but within the largest 3000. The
performance numbers cited are total returns
including the reinvestment of accrued dividends,
earnings and price changes. As with other
indices, returns are gross of all fees and
expenses, including management fees, brokers
commissions and market impact costs. The
SP/BARRA Indices are registered trademarks of
the McGraw-Hill Companies, Inc. and BARRA
RogersCasey. The SP/BARRA Indices are unmanaged
portfolios derived from the SP 500 Index, which
is itself an unmanaged portfolio of 500 widely
held stocks. The performance numbers cited are
total returns including the reinvestment of
accrued dividends, earnings and price changes. As
with other indices, returns are gross of all fees
and expenses, including management fees, brokers
commissions and market impact costs. OakBrook
Investments is not affiliated in any way with
either McGraw-Hill Companies, Inc. or BARRA
RogersCasey. The SP 600/BARRA Indices are
registered trademarks of the McGraw-Hill
Companies, Inc. and BARRA RogersCasey. The SP
600/BARRA Indices are unmanaged portfolios
derived from the SP 600 Index, which is itself
an unmanaged portfolio of 600 small
capitalization stocks. The performance numbers
cited are total returns including the
reinvestment of accrued dividends, earnings and
price changes. As with other indices, returns are
gross of all fees and expenses, including
management fees, brokers commissions and market
impact costs.