Friday 26 November 2004

Global Quantitative Analysis Stock Selection Model revisited

In this report we analyse the performance of ABN AMRO's Stock Selection Model since inception. We find that the model has shown significant predictive power over the last 18 months, while it has done especially well in terms of identifying outperforming stocks.
Global

Chart 1 : Information coefficient since inception

20

10

-10 Jul 03
Source: ABN AMRO

Oct 03

Jan 04

Apr 04

Jul 04

Oct 04

Picking the winners Our model has been especially successful in identifying stocks that are going to outperform. The model's top octile picks have outperformed the market by on average 1.3% per month since inception. Value and earnings revisions are the driving factors Value and earning revisions are the driving force behind the model's performance, while the momentum factors performance has been relatively disappointing. Dynamic weighting scheme The model uses a dynamic weighting scheme based on future predictions of the information coefficients. These predictions are shown to be accurate and unbiased. Our model has performed approximately 25% better than a comparable model based on static weights.

Analysts
Stefan Hartmann
ABN AMRO Bank N.V. +44 20 7678 7613 stefan.hartmann@uk.abnamro.com

Peter Wesselius
+44 20 7678 7623 peter.wesselius@uk.abnamro.com

Nick Aldred
+44 20 7678 7139 nick.aldred@uk.abnamro.com

Adam Strudwick
+44 20 7678 2760 adam.strudwick@uk.abnamro.com

Disclosures and analyst certifications are at the end of the body of this research. 250 Bishopsgate, London, EC2M 4AA, United Kingdom

www.abnamroresearch.com

Contents
ANA L YS IS

Model success We present the performance of the European Stock Selection Model since inception. The model was developed in 1Q03 and its live performance has been monitored on a monthly basis since May 2003. Factor weight dynamics The ABN AMRO Stock Selection Model uses a dynamic weighting scheme. We make predictions of future information coefficients based on their univariate and multivariate properties. How to use the model The output from the ABN AMRO Stock Selection Model can be used in a variety of ways. Below we give some examples.
T EA M

Advanced execution services - team

10

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

A N A L Y S I S

Model success
We present the performance of the European Stock Selection Model since inception. The model was developed in 1Q03 and its live performance has been monitored on a monthly basis since May 2003. We find that the model has been especially successful in identifying stocks that are going to outperform, and has significant predictive power overall. A more detailed description of the model and its use of fundamental and pricing data can be found in our initial publication on the model1. Over recent years we have seen a significant shift in the importance of fundamental variables. While at the end of the nineties growth stocks excelled and momentumbased variables dominated the stock selection, the impact of value variables was negligible or even negative during that period. In the aftermath of the TMT bubble, the importance of fundamental variables changed dramatically and defensive characteristics and cheapness were rewarded. We developed a stock selection model that varies the weights of the fundamental variables based on forecasts on trends in the underlying variables. The weighting scheme of the variables is thus dynamic and changes on a monthly basis. Table 1 shows the model's monthly information coefficients since inception (May 2003) until October 2004 and the top (bottom) octile returns relative to the market.
Table 1 : Stock Selection Model performance
InformationTopbottom octile coefficient May '03 Jun '03 Jul '03 Aug '03 Sep '03 Oct '03 Nov '03 Dec '03 Jan '04 Feb '04 Mar '04 Apr '04 May '04 Jun '04 Jul '04 Aug '04 Sep '04 Oct '04 Average Standard Deviation Information Ratio
Source: ABN AMRO

Top octile market return 8.9% 3.6% 0.6% 2.1% 0.6% 0.4% -1.1% -1.4% 5.3% 0.2% 0.1% -0.6% 0.6% 1.9% 0.7% 0.4% 1.8% -0.5% 1.3% 2.5% 1.8

Market - bottom octile return 1.2% -1.5% -7.2% -2.7% 2.5% -0.3% 0.2% 3.8% -1.8% -2.7% -0.9% 0.5% 0.2% -0.4% 1.7% 1.4% -0.5% -0.4% -0.4% 2.4% -0.6

return 10.0% 2.1% -6.5% -0.6% 3.2% 0.1% -0.8% 2.3% 3.6% -2.5% -0.8% -0.1% 0.7% 1.5% 2.4% 1.8% 1.4% -1.0% 0.9% 3.3% 1.0

24.7% 10.8% -10.8% -1.8% 14.7% -0.8% -0.9% 12.8% 9.6% -8.9% -3.1% -1.8% 3.2% 9.2% 9.5% 6.9% 6.9% -1.1% 4.4% 8.9% 1.7

The top (bottom) octile returns are calculated assuming an equal-weighted portfolio of the top (bottom) 12.5% stocks. The results in table 1 show that our Stock Selection Model has been especially successful in identifying stocks that are going to outperform. The stocks in the top octile beat the market by on average 1.3% per
1

Global Quantitative Analysis - Separating the winners from the losers, 8 April 2003
GL O B A L Q U A NT IT AT I VE A N AL Y SI S 26 N OV E MB ER 20 04

A N AL Y SI S

month. The model has been somewhat disappointing in identifying the underperforming stocks, as these stocks outperformed the market by on average 0.40%. It is currently not yet clear whether the difference in performance between the top and bottom octile (relative to the market) is caused by a structural change in the market due to the small sample size. In longer-term back-tests we have found that the bottom octile stocks underperform more than the top octile stocks outperform. Overall performance (top-bottom) has been satisfactory with an average monthly return of 0.90% and an annualised information ratio of approximately 1.0. It should be noted that this information ratio can be further improved upon by applying a proper portfolio construction process. Further information about the optimisation process and associated issues can be found in our publication on portfolio construction2.
Chart 2 : Model and factor information coefficients (May 2003-October 2004)
Stock Selection Model
20 IC (%) IC (%) 10 0 -10 Jul Oct Jan Apr Jul Oct 10 IC (%) 0 -10 -20 Jul Oct Jan Apr Jul Oct

Growth
20 10 0 -10 -20 -30 Jul

LT Momentum

Oct

Jan

Apr

Jul

Oct

ST Momentum
20 10 0 -10 -20 -30 Jul Oct Jan Apr Jul Oct 15 IC (%) 5 0 -5 -10 Jul

Earnings Revisions
10 IC (%) 0 -10 -20 -30 Oct Jan Apr Jul Oct Jul Oct

Size

IC (%)

10

Jan

Apr

Jul

Oct

Risk
40 IC (%) IC (%) 20 0 -20 Jul Oct Jan Apr Jul Oct 20

Profitability
20 IC (%) 15 10 5 0 -5 Jul Oct Jan Apr Jul Oct Jul Oct

Value

10 0 -10 -20

Jan

Apr

Jul

Oct

----- = historical long-term average IC Source: ABN AMRO

Chart 2 above shows the information coefficients (IC) for our European Stock Selection Model and the individual component factors from May 2003 until October 2004. The dotted line in each of the charts represents the historical long-term average IC, which we took from our initial publication on the European Stock Selection Model3. We see that although the average Stock Selection Model's IC over the past 18 months has been somewhat lower that its longer-term average (4.4% vs. 7.4%), the model does show significant predictive power. The model's annualised IR is 1.7, measured by its average IC divided by the standard deviation of the IC. The underlying data for chart 1 can be found in table 2.

2
3

Global Quantitative Analysis - Capturing alpha in the real world, 4 August 2004 Global Quantitative Analysis - Separating the winners from the losers, 8 April 2003
GL O B A L Q U A NT IT AT I VE A N AL Y SI S 26 N OV E MB ER 20 04

A N AL Y SI S

Of the eight aggregate factors included in our model, value has been the best performer with an average IC of 4.9%, followed by earnings revisions (3.0%). Small caps stocks have outperformed large caps stocks over the last 18 months, which can be concluded from the very negative average information coefficient for the size factor (-5.8%). Long-term momentum has an average IC of -3.1%, while its longterm average is 3.3%. The lack of a clear market trend over the past two years is a possible reason for this poor factor performance. The ICs of the four remaining factors (growth, short-term momentum, risk and profitability) are on average very close to zero, and apart from diversification, they have contributed little to our model over the past 18 months.
Table 2 : Information coefficients May 2003- October 2004
% SSM Growth Long-term momentum May 2003 Jun 2003 Jul 2003 Aug 2003 Sep 2003 Oct 2003 Nov 2003 Dec 2003 Jan 2004 Feb 2004 Mar 2004 Apr 2004 May 2004 Jun 2004 Jul 2004 Aug 2004 Sep 2004 Oct 2004 Average Standard deviation
Source: ABN AMRO

Short-term momentum 5 -12 24 3 -34 0 1 6 -4 8 13 1 -1 -5 3 8 4 -5 0.8 11.8

Earnings revisions 15 1 -8 -9 0 -2 2 17 2 -6 12 -2 7 4 19 0 7 -2 3.0 8.2

Size

Risk

Profitability

Value

25 11 -11 -2 15 -1 -1 13 10 -9 -3 -2 3 9 10 7 7 -1 4.4 8.9

-17 -6 -19 1 3 -17 4 14 -21 2 16 -2 10 -2 12 7 -4 -6 -1.4 11.4

-19 -25 -30 -20 3 4 3 3 15 -7 -7 -21 1 11 2 16 22 -6 -3.1 15.0

-17 -25 -32 -15 17 -3 -1 5 -20 -2 -6 -4 5 -9 1 -3 1 3 -5.8 12.0

16 17 43 19 -29 40 -17 -16 23 -9 -29 -15 -11 6 -33 -27 4 -3 -1.2 23.4

-22 -8 -21 0 5 -19 -2 6 -22 5 19 8 12 -10 15 24 -1 -9 -1.0 14.2

23 13 8 7 9 2 5 2 12 -1 -4 7 -4 7 -3 5 -5 5 4.9 6.9

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

A N AL Y SI S

Factor weight dynamics

The ABN AMRO Stock Selection Model uses a dynamic weighting scheme. We make predictions of future information coefficients based on their univariate and multivariate properties. The models weights are chosen such that they produce the maximum ratio of average (IC)/standard deviation (IC), thereby taking into account the diversifying effect of having exposure to multiple styles. The actual monthly model weights since inception are shown in table 3. As discussed in the previous section, the short-term momentum factor has been less successful recently and the model has reacted over time by putting less weight on it, reducing it from -40% to -30% over the past 18 months. In contrast, the profitability factor has seen its weight nearly double to 25%, even though the average information coefficient for the factor has been slightly negative. This effect is attributed to the correlation between the factors and the increase in weight is due to its diversifying properties. We have also seen decreases in weight for the long-term momentum, earnings revisions and value factors.
Table 3 : Factor weights
% Growth Long-term momentum Jun 2003 Jul 2003 Aug 2003 Sep 2003 Oct 2003 Nov 2003 Dec 2003 Jan 2004 Feb 2004 Mar 2004 Apr 2004 May 2004 Jun 2004 Jul 2004 Aug 2004 Sep 2004 Oct 2004 Nov 2004 Average
Source: ABN AMRO

Short-term momentum -40 -33 -33 -32 -31 -31 -30 -35 -36 -33 -33 -32 -32 -31 -30 -34 -31 -30 -33

Earnings revisions 60 56 63 63 55 55 50 56 57 55 55 56 54 54 52 53 53 52 55

Size

Risk

Profitability

Value

-11 -8 -15 -14 -12 -17 -12 -10 -12 -11 -11 -11 -10 -10 -9 -8 -9 -8 -11

23 13 9 11 12 16 17 18 18 16 15 15 14 15 15 17 17 16 15

-1 2 -2 -5 -4 -3 -1 0 -1 -1 -2 -3 -3 -3 -3 0 -4 -4 -2

7 8 12 13 11 14 11 10 10 10 10 10 10 10 8 9 8 7 10

13 12 10 17 22 20 21 18 17 17 18 21 23 23 24 19 24 25 19

49 50 54 48 47 47 44 45 46 48 47 46 44 44 42 45 42 42 46

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

A N AL Y SI S

The current factor correlation matrix is shown below. Similar to our experience in earlier back-tests, we see that value and risk are the most diversifying factors, being negatively correlated with most other factors. The most highly positively correlated factors are growth and profitability.
Table 4 : Factor correlation, November 2004
% Growth Long-term momentum Long-term momentum Short-term momentum Earnings revisions Size Risk Profitability Value
Source: ABN AMRO

Short-term momentum

Earnings revisions

Size

Risk

Profitability

18 4 10 7 -19 42 0 35 26 2 -16 15 -16 26 10 0 0 -15 7 -7 11 -10 -5 13 -6 -30 4 0

Table 5 : Performance contribution

Factor Value Earnings revisions Size Growth Risk Profitability Short-term momentum Long-term momentum
Source: ABN AMRO

Score 2.3 1.7 0.1 0.1 -0.1 -0.2 -0.3 -0.5

We have also measured the performance contribution in order to give an indication of the size and sign of the contribution that the individual factors have made to the overall strategy. The factor contribution scores in table 5 have been calculated by multiplying the 18-months average weight with the average IC. We see that value (2.3) and earnings revisions (1.7) have been the driving force behind the model's success, while the two momentum factors (-0.3 and -0.5) have actually made an overall negative performance contribution over the last 18 months. Over the past 18 months the market was dominated by value factors. We find growth-related factors tend to do poorly in this environment. This is reflected in the performance of the long-term momentum factor. When growth stocks do well, momentum investing becomes more effective as growth companies competitive advantages lead to a series of positive earnings surprises over several quarters. However in a value driven environment mean reversion sets in for stocks with strong historical performance as their value rankings deteriorate with price performance. The model has catered for this by lowering the weights on long-term momentum over the past 18 months.

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

A N AL Y SI S

In chart 3 below we show the relationship between the predicted information coefficients and the realised information coefficients (for four) of the factors. As the predicted information coefficients are used to choose the optimal factor weights, it is important that our overall predictions are 'good enough'. For example, we would expect the slope in chart 3, depicting the relationship between the predicted and realised information coefficients, to be close to one for a good model, as well as statistically significant different from zero. This is indeed what we find4. This is further evidence that the information coefficients predictions used in our model are accurate and (statistically) unbiased. Furthermore, we have compared the results of the Stock Selection Model with a model using static weights. These static weights are taken to be the optimal weights at the start of the 18-month live period (from May 2003 in table 3). We find that the risk/return characteristics for the dynamic model are approximately 25% better than for the static model.
Chart 3 : Information coefficients for selected factors May 2003-October 2004

40

ST Momentum Earn.Revision Profitability Value

20 Realised IC (%)

-20

-6

-4

-2

0 Predicted IC (%)

Source: ABN AMRO

Regression results: ActualIC = -1.05% + T-value -1.1 R2 = 0.097

GL O B A L

1.17 * PredictedIC 3.9

Q U A NT IT AT I VE A N AL Y SI S 26 N OV E MB ER 20 04

A N AL Y SI S

How to use the model

The output from the ABN AMRO Stock Selection Model can be used in a variety of ways. Below we give some examples.

Consistency check
Managers use the model to compare their holdings with the rankings of the stock selection model, eg they would be concerned if a larger number of stocks with a low ranking appear in their portfolio. On the other hand, an overlap of stock positions with our buy list would support the choice of the manager. Many investors concentrate on the changes in stock rankings and try to understand the drivers in the revaluations. The model may dislike a stock due to its high valuation or due to negative revisions of forecasted earnings. Fund managers are in a position to understand the reasons for the changes in the stock rankings and should be able to see if their investment case is still in place of if they need to react to changing market conditions.

Stock screen
Quantitative models are often used as part of a qualitative investment process to reduce the universe. The model output is used to reduce the number of potential stocks to a manageable number that can be analysed in full detail.

Additional model
The model can also be used as an additional source of alpha in existing qualitative or quantitative models. The way the Stock Selection Model can add value is twofold: its alpha generation adds to the overall model alpha and it also adds value by diversification. The proportion of the overall model that should be allocated to the Stock Selection Model is dependent on the joint statistical behaviour of the information coefficients.

Main quantitative model

The monthly alphas that are generated by our model can be used as the key source of alpha in the portfolio construction process. We have successfully used these alphas to build long only and long-short portfolios for various tracking errors and portfolio sizes. We explicitly take transaction costs into account in our portfolio construction process, focusing on returns after cost. A detailed discussion on this subject can be found in our Capturing Alpha publication5.

Global Quantitative Analysis - Capturing Alpha in the Real World, 4 August 2004
GL O B A L Q U A NT IT AT I VE A N AL Y SI S 26 N OV E MB ER 20 04

T E A M

Advanced execution services - team

Table 6 : ABN AMRO portfolio product sales contacts
London Gijsbert de Lange Mike de Friend Brian Curran Neill Flack US Brenda Colon Peter Krase Vivek Arora Asia Jason Barrow Kee Meng Tan Lorraine Ho Japan Mimiko Adachi Hiroyuki Yokoyama mimiko.adachi@jp.abnamro.com hiro.yokoyama@jp.abnamro.com +813 5405 6256 +813 5405 6865 jason.barrow@abnamro.com kee.meng.tan@abnamro.com lorraine.ho@abnamro.com +852 2700 5193 +852 2700 5193 +852 2700 5193 brenda.colon@abnamro.com peter.krase@abnamro.com vivek.arora@abnamro.com +1 212 409 7138 +1 212 409 7138 +1 212 409 7138 gijsbert.de.lange@uk.abnamro.com mike.de.friend@uk.abnamro.com brian.p.curran@uk.abnamro.com neill.flack@uk.abnamro.com +44 20 7678 3403 +44 20 7678 1620 +44 20 7678 3698 +44 20 7678 1083

Source: ABN AMRO

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

10

DISCLOSURES

Recommendation structure
Absolute performance recommendation: The target price and absolute recommendation are based on implied upside/downside for the stock relative to the market. A Buy/Sell implies upside/downside of 15% or more; an Add/Reduce implies upside/downside of 5-15%; and a Hold implies less than 5% upside/downside. Sector relative to market: The sector view relative to the market is the responsibility of the strategy team. Overweight (OW)/Underweight (UW) implies upside/downside of 10% or more and Neutral (N) implies less than 10% upside/downside. Asset allocation: The asset allocation is the responsibility of the economics team. The recommended weight (Over, Neutral and Under) for equities, cash and bonds is based on a number of metrics and does not relate to a particular size change in one variable. Target price: The target price is the level the stock should currently trade at if the market were to accept the analyst's view of the stock and if the necessary catalysts were in place to effect this change in perception within the performance horizon. In this way, therefore, the target price abstracts from the need to take a view on the market or sector. If it is felt that the catalysts are not fully in place to effect a re-rating of the stock to its warranted value, the target price will differ from 'fair' value. Performance parameters and horizon: Given the volatility of share prices and our predisposition not to change recommendations frequently, these performance parameters should be interpreted flexibly. Performance in this context only reflects capital appreciation and the horizon is 6-12 months.

Distribution of recommendations
The table opposite shows the distribution of ABN AMRO's recommendations. The first column displays the distribution of recommendations globally and the second column shows the distribution for the region. Numbers in brackets show the percentage for each category where ABN AMRO has an investment banking relationship. In all cases the numbers include both absolute and sector relative recommendations.

Recommendation distribution (as at 9 Nov 2004)

Global total (IB%) Buy Add Hold Reduce Sell Total (IB%) 451 (22) 337 (27) 379 (18) 154 (8) 59 (10) 1380 (20)

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

11

DISCLAIMER
Copyright 2004 ABN AMRO Bank N.V. and affiliated companies ("ABN AMRO"). All rights reserved. This material was prepared by the ABN AMRO affiliate named on the cover or inside cover page. It is provided for informational purposes only and does not constitute an offer to sell or a solicitation to buy any security or other financial instrument. While based on information believed to be reliable, no guarantee is given that it is accurate or complete. While we endeavour to update on a reasonable basis the information and opinions contained herein, there may be regulatory, compliance or other reasons that prevent us from doing so. The opinions, forecasts, assumptions, estimates, derived valuations and target price(s) contained in this material are as of the date indicated and are subject to change at any time without prior notice. The investments referred to may not be suitable for the specific investment objectives, financial situation or individual needs of recipients and should not be relied upon in substitution for the exercise of independent judgement. ABN AMRO may from time to time act as market maker, where permissible under applicable laws, or, as an agent or principal, buy or sell securities, warrants, futures, options, derivatives or other financial instruments referred to herein. ABN AMRO or its officers, directors, employee benefit programmes or employees, including persons involved in the preparation or issuance of this material, may from time to time have long or short positions in securities, warrants, futures, options, derivatives or other financial instruments referred to in this material. ABN AMRO may at any time solicit or provide investment banking, commercial banking, credit, advisory or other services to the issuer of any security referred to herein. Accordingly, information may be available to ABN AMRO, which is not reflected in this material, and ABN AMRO may have acted upon or used the information prior to or immediately following its publication. Within the last three years, ABN AMRO may also have acted as manager or co-manager for a public offering of securities of issuers referred to herein. The stated price of any securities mentioned herein is as of the date indicated and is not a representation that any transaction can be effected at this price. Neither ABN AMRO nor other persons shall be liable for any direct, indirect, special, incidental, consequential, punitive or exemplary damages, including lost profits arising in any way from the information contained in this material. This material is for the use of intended recipients only and the contents may not be reproduced, redistributed, or copied in whole or in part for any purpose without ABN AMRO's prior express consent. In any jurisdiction in which distribution to private/retail customers would require registration or licensing of the distributor which the distributor does not currently have, this document is intended solely for distribution to professional and institutional investors. Australia: Any report referring to equity securities is distributed in Australia by ABN AMRO Equities Australia Ltd (ABN 84 002 768 701, AFS Licence 240530), a participant of the ASX Group. Any report referring to fixed income securities is distributed in Australia by ABN AMRO Bank NV (Australia Branch) (ABN 84 079 478 612, AFS Licence 238266). Australian investors should note that this document was prepared for wholesale investors only. Canada: The securities mentioned in this material are available only in accordance with applicable securities laws and may not be eligible for sale in all jurisdictions. Persons in Canada requiring further information should contact ABN AMRO Incorporated. Hong Kong: This document is being distributed in Hong Kong by, and is attributable to, ABN AMRO Asia Limited which is regulated by the Securities and Futures Commission of Hong Kong. India: Shares traded on stock exchanges within the Republic of India may only be purchased by different categories of resident Indian investors, Foreign Institutional Investors registered with The Securities and Exchange Board of India ("SEBI") or individuals of Indian national origin resident outside India called Non Resident Indians ("NRIs") and Overseas Corporate Bodies ("OCBs"), predominantly owned by such persons or Persons of Indian Origin (PIO). Any recipient of this document wanting additional information or to effect any transaction in Indian securities or financial instrument mentioned herein must do so by contacting a representative of ABN AMRO Asia Equities (India) limited. Italy: Persons in Italy requiring further information should contact ABN AMRO Bank N.V. Milan Branch. Japan: This report is being distributed in Japan by ABN AMRO Securities Japan Ltd to institutional investors only. New Zealand: This document is distributed in New Zealand by ABN AMRO NZ Limited an accredited NZX Firm. Russia: The Russian securities market is associated with several substantial risks, legal, economic and political, and high volatility. There is a relatively high measure of legal uncertainty concerning rights, duties and legal remedies in the Russian Federation. Russian laws and regulations governing investments in securities markets may not be sufficiently developed or may be subject to inconsistent or arbitrary interpretation or application. Russian securities are often not issued in physical form and registration of ownership may not be subject to a centralised system. Registration of ownership of certain types of securities may not be subject to standardised procedures and may even be effected on an ad hoc basis. The value of investments in Russian securities may be affected by fluctuations in available currency rates and exchange control regulations. Singapore: Any report referring to equity securities is distributed in Singapore by ABN AMRO Asia Securities (Singapore) Pte Limited (RCB Regn No. 198703346M) to clients who fall within the description of persons in Regulation 49(5) of the Securities and Futures (Licensing and Conduct of Business) Regulations and Regulations 34 and 35 of the Financial Advisers Regulations. Any report referring to non-equity securities is distributed in Singapore by ABN AMRO Bank NV (Singapore Branch) Limited to clients who fall within the description of persons in Regulations 34 and 35 of the Financial Advisers Regulations. Investors should note that this material was prepared for accredited investors only. Recipients who do not fall within the description of persons under Regulation 49(5) of the Securities and Futures (Licensing and Conduct of Business) Regulations or Regulations 34 and 35 of the Financial Advisers Regulations should seek the advice of their independent financial advisor prior to taking any investment decision based on this document or for any necessary explanation of its contents. United Kingdom: Equity research is distributed in the United Kingdom by ABN AMRO Equities (UK) Limited, which is registered in England (No 2475694), and is authorised and regulated by the Financial Services Authority. All other research is distributed in the United Kingdom by ABN AMRO Bank NV, London Branch, which is authorised by the Dutch Central Bank and by the Financial Services Authority; and regulated by the Financial Services Authority for the conduct of UK business. The investments and services contained herein are not available to private customers in the United Kingdom. United States: Distribution of this document in the United States or to US persons is intended to be solely to major institutional investors as defined in Rule 15a6(a)(2) under the US Securities Act of 1934. All US persons that receive this document by their acceptance thereof represent and agree that they are a major institutional investor and understand the risks involved in executing transactions in securities. Any US recipient of this document wanting additional information or to effect any transaction in any security or financial instrument mentioned herein, must do so by contacting a registered representative of ABN AMRO Incorporated, Park Avenue Plaza, 55 East 52nd Street, New York, N.Y. 10055, US, tel + 1 212 409 1000, fax +1 212 409 5222. - Material means all research information contained in any form including but not limited to hard copy, electronic form, presentations, e-mail, SMS or WAP.

Regulatory disclosures
_________________________________________________________________________________________________________________________________ The research analyst or analysts responsible for the content of this research report certify that: (1) the views expressed and attributed to the research analyst or analysts in the research report accurately reflect their personal opinion(s) about the subject securities and issuers and/or other subject matter as appropriate; and, (2) no part of his or her compensation was, is or will be directly or indirectly related to the specific recommendations or views contained in this research report. On a general basis, the efficacy of recommendations is a factor in the performance appraisals of analysts. _________________________________________________________________________________________________________________________________ For a discussion of the valuation methodologies used to derive our price targets and the risks that could impede their achievement, please refer to our latest published research on those stocks at www.abnamroresearch.com. Disclosures regarding companies covered by ABN AMRO group can be found on ABN AMRO's research website at www.abnamroresearch.com. Should you require additional information please contact the relevant ABN AMRO research team or the author(s) of this report.

GL O B A L

Q U A NT IT AT I VE

A N AL Y SI S

26

N OV E MB ER

20 04

12