E

very year at the end of June, The Frank Russell Company

reconstitutes its series of equity indexes. This process
typically results in significant changes to both the capi-
talization and style indexes. The 2004 reconstitution
took place on June 25, the last Friday in June. With more
than $350 billion indexed to Russell strategies, this annual
trade is a significant event that draws the attention of both
index managers and other market participants. Index manager
trading alone is estimated to be more than $30 billion. This
years reconstitution was larger than in past years due to the
increased value of the assets managed to these benchmarks.
Additionally, several changes were instituted this year. As in
most years, early predictions of the modifications to the index
changed as market dynamics made the 2004 reconstitution
a constantly evolving process.
2004 Changes
This year The Frank Russell Company introduced a provisional
index in an attempt to get index managers to rebalance early
and reduce the market impact on the effective date. The
provisional index was announced on June 14 and reflected
all of the changes that would be official on June 25. While
the concept behind introducing a provisional index is to provide
an available benchmark for those seeking to rebalance early,
there was not much of a noticeable impact. This likely was
due to the fact that assets managed
within fund vehicles, as compared
to separate accounts, would still be
measured against official benchmarks,
which would not reflect changes until
the effective date. A second change
instituted this year was to hold the reconstitution on the last
Friday of June rather than on the last business day of the
month. This was done to provide additional time for trade
settlement purposes. Finally, the effective prices for NASDAQ
stocks were based on the new NASDAQ Closing Cross
mechanism. Because NASDAQ is a dealer market, there has
never been an effective auction mechanism for determining
closing prices. In April 2004, NASDAQ introduced a new
closing cross to facilitate the ability of investors to execute at
closing prices. Prior to this, there was no way to guarantee
an execution of a NASDAQ stock at closing prices, although
some broker-dealers provided this as an accommodation for
index managers. Based on hourly volume analysis, approxi-
mately 30-40% of this years trade, both in listed and NASDAQ
stocks, was executed at closing prices. This is indicative of
the desire of index managers to limit risk by trading at close
to the effective price.
Turnover Analysis
Capitalization turnover in the large cap core indexes was
slightly higher than last year, while turnover in the small cap
Russell 2000 index was lower. Within the style indexes, large
cap turnover was higher primarily due to the reclassification
of GE from 100% growth to predominantly value.
2004 Russell
Reconstitution Recap
Alexander J. Matturri, Jr. CFA
Director, Global Equity Index
Quantitative Management
Northern Trust Global Investments
July 2004

2
Review of Northern Trust Global
Investments Reconstitution Strategy
In order to successfully manage the reconstitution process,
we started reviewing the anticipated changes to the Russell
indexes in January, a full six months ahead of the effective
date. In addition to constantly reviewing the preliminary list
of changes, we analyzed various trading alternatives and the
anticipated market impact and transaction costs of each
alternative. A key input to this analysis is an understanding
of the potential trading patterns of non-index fund market
participants such as hedge funds, proprietary trading desks
and aggressive active managers. These market participants
have been accounting for a much greater portion of the trading
associated with the annual rebalancing as they seek to profit
from anticipated index fund flows. In order to minimize
potential tracking error, index fund managers must trade
near the close on the effective date. These other market
participants attempt to profit by providing the necessary
liquidity needed by the index fund managers. The net
effect from these non-index fund market participants is
the erosion of index fund investor wealth as they take
advantage of short-term supply/demand imbalances.
Beginning in March, we interviewed all the major
broker-dealers to determine which firms would be able to
execute our strategy most effectively. By May we started to
formulate our initial trading strategy, and we finalized broker
selection and documentation by the end of the first week of
June. Our trading strategy was continuously updated until
very close to the reconstitution effective date to take into
account changing market dynamics. Vital to the review of
broker-dealers was a detailed study of our operational needs
and those of our clients custodians. The ability of a broker
to settle the thousands of transactions involved on time was
an important consideration in the selection process. Throughout
this review process was an understanding of the investment,
legal and administrative risks involved and how to control
these risks. Based on this review process we developed a trading
strategy that would, within a risk-controlled environment,
seek to minimize the wealth erosion that has historically
arisen due to reconstitutions.
After a detailed analysis of the anticipated changes and
the expected market dynamics, the decision was made to
enter into principal transactions with certain key brokers. In
these transactions, the broker guaranteed the closing price
of stocks that needed to be traded, net of any commissions,
with a profit split on any outperformance accruing to our
funds. This approach provided us with no market risk yet the
opportunity to profit from any outperformance associated
with their trading activity. Given the potential risks, we felt
this more conservative approach was prudent this year.
As has been our approach during the past few reconsti-
tutions, we decided not to use internal crossing as an integral
part of the implementation strategy. We believe that for
information-based trades such as the Russell reconstitution,
trading in the open market rather than crossing provides the
optimal solution. In these exposed or information-based trades,
crossing (which can result in buying at the expected high
and selling at the expected low) is not always in our clients
best interests. By splitting the sells and the buys one can try
to achieve the best possible execution price for all investors.
A variation to direct agency trading in the market is to utilize
a broker-dealers ability to execute the trade in a principal
capacity. In this situation, the broker guarantees, at a mini-
mum, the close net of commissions with our funds retaining
a majority of any outperformance relative to the close. The
broker has an incentive to execute better than the close due
to its participation in the profit split. By using a principal
trade, we are guaranteed an execution no worse than a cross
while retaining the ability to outperform.
Our analysis of the market environment determined
that demand by non-index managers would be a significant
portion of this years trading once again. In addition, we
expected that much of the trading was going to be concen-
trated on the last day and more precisely, in the last few
hours prior to the close on June 25. The rationale for this
view included the performance patterns of last years trades
(when those who rebalanced early were penalized) and this
years volume trends through mid-June, which indicated
muted rebalancing activity.
2004 Russell Reconstitution One-Way Turnover
2004 2003
Russell 3000 2.8% 2.1%
Russell 1000 2.9% 2.6%
Russell 1000 Value 13.3% 10.2%
Russell 1000 Growth 13.5% 9.6%
Russell 2000 18.8% 19.8%
Russell 2000 Value 28.6% 34.8%
Russell 2000 Growth 32.2% 36.7%
Source: Goldman Sachs

3
Changes to the Russell 1000 & Russell 2000 Indexes
2004 2003 2002 2001 2000
New Additions to R1000 16 7 21 24 76
New Additions to R2000 308 282 378 495 545
Deletions from R2000 205 180 222 276 322
R1000 to R2000 67 83 105 110 139
R2000 to R1000 67 94 136 151 118
Source: Lehman Brothers & Goldman Sachs
Index Construction Methodology
The Russell indexes are capitalization-weighted equity indexes
covering approximately 98% of the investable market cap in
the United States. At the time of the annual reconstitution, all
the stocks in the Frank Russell universe are ranked by market
capitalization using the May 31 cutoff date. The largest 1000
companies become the Russell 1000 index while the next
2000 companies comprise the Russell 2000 index. The
combined list makes up the Russell 3000 index.
The Russell style indexes are also reconstituted at the end
of June, using a proprietary methodology. Each capitalization
index is divided into a growth and a value style index. At the
time of the reconstitution, approximately 50% of the capi-
talization weight of each index is assigned to a style index.
Approximately 30% of the companies in each capitalization
index are represented in both style indexes, while the remaining
70% are represented in one or the other style index, but not
both. Companies are deemed to be either growth or value
based on their adjusted price-to-book ratio and their long-term
earnings growth rate as determined by I/B/E/S. This process
contrasts with the methodology used by Standard & Poors
which bases its style determination strictly on price-to-book
and uses a mutually exclusive policy which places a company
in one or the other index, but not both.
Review of Index Changes
As of the June 25 effective date there were 16 new direct
additions to the Russell 1000 index and 67 companies
migrating up from the Russell 2000 index. This compares
to 2002 when there were only seven new additions and
94 issues migrating up from the Russell 2000 index.
Sector Profiles
Within the large cap core and style indexes there werent any
major shifts in sector weightings from our earlier forecast. In
the large cap indexes, the Financial Services sector continues
to be the largest weight in the Russell 1000 index and Russell
1000 Value index at 22.52% and 33.64%, respectively. The
Consumer Staples sector had the largest drop in the core index,
dropping 0.22%, while the Producer Durables sector had the
greatest increase at 0.15%. In the Russell 1000 Value index,
the Other sector was up the most (+5.08%) while Financial
Services had the largest decline (-1.42%). The increase in the
Other sector can be attributed primarily to the reassignment
of General Electric (which is classified in Other), from 100%
growth to predominantly value. The reverse took place in
the Russell 1000 Growth index as Other declined the most
(-4.98%) in response to this move. The General Electric
weighting was spread across the other sectors of the Russell
1000 Growth index with the Consumer Discretionary sector
gaining the most (+1.61%). The Health Care sector contin-
ues to be the largest weighting in the large cap growth index
at 24.31%.
Changes from our preliminary analysis of the small
capitalization indexes, while greater than in the large cap
indexes, were still generally close to our original projections.
Similar to the large cap indexes, the Financial Services sector
will continue to be the largest weight, accounting for 22.54%
of the Russell 2000 index and 33.89% of the Russell 2000
Value index. The increase in weighting for Financial Services
in both the Russell 2000 index (+0.89%) and Russell 2000
Value index (+3.06%) was the greatest for each index. Within
the core index, the Producer Durables sector declined the most
(-0.69%) while in the Russell 2000 Value index, Producer
Durables (-2.57%) and Consumer Discretionary (-2.29%)
were the largest decreases. In the Russell 2000 Growth index,
Consumer Discretionary increased the most (+1.75%) while
Health Care decreased the most (-1.62%). Consumer
Discretionary will now be the largest weight in the small
capitalization growth index at 22.67%, overtaking Health
Care, which will account for 21.38%.

4
Core Indexes Sector Weights
RUSSELL 1000 RUSSELL 2000
Sector Pre-Recon. Post-Recon. Change Pre-Recon. Post-Recon. Change
Auto & Transportation 2.27% 2.31% 0.04% 4.35% 4.42% 0.07%
Consumer Discretionary 13.94% 14.05% 0.11% 19.00% 18.68% -0.32%
Consumer Staples 7.43% 7.21% -0.22% 1.87% 1.77% -0.10%
Financial Services 22.66% 22.52% -0.14% 21.65% 22.54% 0.89%
Health Care 13.93% 13.87% -0.06% 13.55% 12.89% -0.66%
Integrated Oils 4.61% 4.47% -0.14% 0.06% 0.13% 0.07%
Materials & Processing 3.52% 3.50% -0.02% 9.02% 9.33% 0.31%
Other 4.49% 4.56% 0.07% 0.28% 0.40% 0.12%
Other Energy 1.78% 1.88% 0.10% 4.42% 4.40% -0.02%
Producer Durables 4.07% 4.21% 0.14% 8.47% 7.78% -0.69%
Technology 14.49% 14.57% 0.08% 13.27% 13.57% 0.30%
Utilities 6.81% 6.84% 0.03% 4.05% 4.08% 0.03%
Source: Goldman Sachs
Style Indexes Sector Weights
RUSSELL 1000 VALUE RUSSELL 1000 GROWTH
Sector Pre-Recon. Post-Recon. Change Pre-Recon. Post-Recon. Change
Auto & Transportation 2.82% 2.59% -0.23% 1.72% 2.02% 0.30%
Consumer Discretionary 10.56% 9.29% -1.27% 17.30% 18.91% 1.61%
Consumer Staples 5.96% 5.35% -0.61% 8.90% 9.11% 0.21%
Financial Services 35.07% 33.64% -1.43% 10.30% 11.16% 0.86%
Health Care 3.50% 3.66% 0.16% 24.33% 24.31% -0.02%
Integrated Oils 9.23% 8.84% -0.39% 0.00% 0.00% 0.00%
Materials & Processing 5.73% 5.42% -0.31% 1.32% 1.54% 0.22%
Other 1.68% 6.77% 5.09% 7.28% 2.30% -4.98%
Other Energy 2.27% 2.48% 0.21% 1.30% 1.27% -0.03%
Producer Durables 4.51% 3.61% -0.90% 3.63% 4.83% 1.20%
Technology 6.59% 6.14% -0.45% 22.37% 23.18% 0.81%
Utilities 12.07% 12.21% 0.14% 1.57% 1.36% -0.21%
Source: Goldman Sachs
RUSSELL 2000 VALUE RUSSELL 2000 GROWTH
Sector Pre-Recon. Post-Recon. Change Pre-Recon. Post-Recon. Change
Auto & Transportation 5.26% 5.39% 0.14% 3.42% 3.40% -0.02%
Consumer Discretionary 17.15% 14.86% -2.28% 20.92% 22.67% 1.75%
Consumer Staples 2.65% 2.74% 0.09% 1.07% 0.75% -0.32%
Financial Services 30.83% 33.89% 3.06% 12.17% 10.65% -1.52%
Health Care 4.39% 4.80% 0.41% 23.00% 21.38% -1.62%
Integrated Oils 0.06% 0.04% -0.02% 0.06% 0.22% 0.16%
Materials & Processing 12.91% 12.27% -0.64% 5.00% 6.26% 1.26%
Other 0.50% 0.70% 0.20% 0.06% 0.08% 0.02%
Other Energy 5.02% 4.95% -0.07% 3.80% 3.83% 0.03%
Producer Durables 8.84% 6.27% -2.57% 8.10% 9.37% 1.27%
Technology 6.45% 6.97% 0.53% 20.31% 20.48% 0.17%
Utilities 5.95% 7.11% 1.16% 2.09% 0.91% -1.18%
Source: Goldman Sachs

5
Performance Review of the 2004
Reconstitution Trade
The 2004 Reconstitution trade performance was in line with
our expectations. During the month of June, performance of
the various segments was generally in line with rebalancing
flows. The buys went up relative to their benchmark while
the sales went down. The big exception was the new additions
to the Russell 2000 index. Throughout the full month, the
Russell 2000 index additions declined 2.8% relative to the
index, with most of the underperformance coming in the
last week when they declined 1.6%. Typically, this segment
posts positive relative performance as investors need to buy
these stocks. This years performance was worse than 2003
when the Russell 2000 index additions were up 1.7% on a
relative basis during June.
On the large cap side, the 16 stocks making up the
Russell 1000 index additions were up 3.1% relative to the
index during the last week, and up 8.1% for the full month.
The rebalance date saw positive relative performance for the
Russell 1000 index additions while the Russell 2000 index
additions declined 1.3%. Overall, as previously stated, trading
volume was light as market participants waited until very
close to the effective date to rebalance.
Intra-Day Performance June 25
Last years reconstitution was affected by the strong relative
performance of smaller capitalization and lower-quality issues.
Counter to expectations and historical trends, issues being
dropped from the Russell 2000 index in 2003 outperformed
the benchmark during the final week (+2.4%) and for the
entire month of June (+5.1%). This year a more traditional
pattern returned as stocks dropped from the Russell 2000
index declined 3.4% during the month, and 4.1% during
the last week.
PERFORMANCE PERFORMANCE
RELATIVE TO THE RUSSELL 2000 RELATIVE TO THE RUSSELL 1000
Period Year R2 Adds R2 Deletes Adds vs. Deletes R1 Adds R2 to R1 Shifts R1 to R2 Shifts
5/31 to Friday before 2000 14.4% -7.6% 22.0% 14.9% 17.6% 0.2%
last week of June 2001 -3.4% -4.9% 1.5% -7.5% 0.1% -8.4%
2002 3.4% -4.4% 7.8% 4.4% 3.4% -9.3%
2003 2.5% 2.0% 0.5% 1.1% -3.6% 3.6%
2004 0.5% -0.6% 1.1% 5.6% -2.6% 0.8%
Month of June 2000 19.8% -9.0% 28.8% 20.7% 15.1% -3.7%
2001 -0.4% -10.6% 10.2% 2.3% -0.4% 5.0%
2002 6.5% -13.7% 20.2% 5.2% -0.7% -3.6%
2003 1.7% 5.1% -3.4% 6.1% -3.2% 4.3%
2004 -2.8% -3.4% 0.6% 8.4% 1.6% 0.4%
Last week of June 2000 5.4% -1.4% 6.8% 5.8% -2.5% -3.8%
2001 3.3% -5.9% 9.1% 10.9% -0.6% 15.1%
2002 3.0% -9.2% 12.2% 0.6% -2.8% 1.1%
2003 -0.8% 2.4% -3.2% 2.8% 1.3% 0.7%
2004 -1.6% -4.1% 2.5% 3.1% 3.0% 0.0%
Rebalance Day 2000 -2.6% 1.1% -3.7% 2.1% 2.6% -6.6%
2001 6.1% -6.8% 12.9% 8.4% -1.8% 9.4%
2002 0.6% -5.7% 6.3% 3.5% -0.2% 2.5%
2003 -0.1% 0.2% -0.3% 2.0% -0.9% 0.0%
2004 -1.3% -2.5% 1.2% 2.3% 2.5% -1.0%
Expected Flow Buy Sell Buy Buy Sell Buy
Source: Goldman Sachs

Northern Trust Global Investments
The Northern Trust Company
50 South LaSalle Street
Chicago, Illinois 60675
www.northerntrust.com
Q7731 (07/04)
Northern Trust Global Investments comprises Northern Trust Investments, N.A., Northern Trust Global Investments (Europe), Ltd.,
Northern Trust Global Investments Japan, K.K., the investment advisor division of The Northern Trust Company and Northern Trust
Global Advisors, Inc. and its subsidiaries to offer investment products and services to personal and institutional markets. This infor-
mation has been obtained from sources believed to be reliable, but its accuracy and completeness are not guaranteed. Any opinions
expressed herein are subject to change at any time without notice. This report is for informational purposes only.
To discuss the 2004 Russell Reconstitution or any other investment issue
you may have, please contact your Investment Relationship Manager.