Default and Recovery Rates of European Corporate Bond Issuers:

Special Comment March 2006 Contact Phone New York Praveen Varma 1.212.553.1653 Richard Cantor David Hamilton London Eric de Bodard 44.20.7772.5454 Alice Keegan Guillaume Menuet Default and Recovery Rates of European Corporate Bond Issuers: 1985-2005 This is Moody's fifth annual study of European corporate bond issuers and their default experience. Broad conclusions include: The downward trend in European corporate defaults continued in 2005 as only two Moody s-rated issuers defaulted on bonds in 2005. The total volume of Moody's-rated defaulted bonds was 290 million in 2005, down from 1.4 billion in 2004. On an issuer-weighted basis, the European speculative-grade default rate declined to 1.1% in 2005 from 1.2% in 2004, 6.0% in 2003 and 20.5% in 2002. On a volume-weighted basis, the European speculativegrade default rate fell to 0.4% in 2005 from 0.9% in 2004, 6.7% in 2003 and 58.2% in 2002. The two defaulting bond issuers in 2005, Concordia Bus AB and Concordia Bus Nordic AB, are affiliates within the same corporate family. Only one issuer, Jarvis PLC of UK, defaulted on syndicated bank loans of about 470 million. Jarvis was not rated by Moody s. The average credit quality of European corporate bond issuers continued to improve, as the upgrade rate exceeded the downgrade rate in 2005 for the second year in a row. Total speculative-grade bond issuance in Europe was about 20 billion in 2005, modestly lower than the record-setting 23 billion issued in 2004. The credit quality of new issuance declined considerably as the share of new speculative-grade issuance rated Caa-C increased from 16% in 2004 to 32% in 2005. Such low quality issuance portends a significant increase in default rates over the next two to three years. Annual Speculative-Grade Default Rates for European Issuers Continue to Decline 70% 60% 50% 40% 30% 20% 10% 0% Jan-00 Sep-00 May-01 Jan-02 Sep-02 May-03 Jan-04 Issuer Weighted Value Weighted Sep-04 May-05

Table of Contents Page Introduction... 5 Database and Methodology... 5 The 2005 European Corporate Default Experience... 7 European Speculative-Grade Bond Default Rates... 8 Cumulative Default Rates... 9 Trends in Credit Quality... 10 Rating Migration Rates... 10 Upgrade-to-Downgrade Ratios... 12 Ratings Performance... 13 Recovery Rates for European Defaulted Instruments...14 Credit Losses... 14 Looking Ahead (2006-2008)... 15 Summary and Conclusions... 16 Related Research... 17 Appendix I - Methodology... 18 Estimated Senior Ratings... 18 Default Rate Calculations... 18 Recovery Rate Estimates... 19 Appendix II Supplemental Default and Transition Rates Tables for European Issuers... 20 2 Moody s Special Comment

Table of Exhibits Page Exhibit 1 - Distribution of European Bond Issuers by Whole Letter Rating...6 Exhibit 2 - Countries of Domicile of European Moody s-rated Corporate Bond Issuers...6 Exhibit 3 European Bond Default Counts and Volumes...7 Exhibit 4 List of the 2005 European Defaulters...7 Exhibit 5 - European Issuer-Weighted and Volume-Weighted Speculative-Grade Bond Default Rates...8 Exhibit 6 - Average Annual Speculative-Grade Default Rates for European and North American Issuers...8 Exhibit 7 - European and North American Issuer-Weighted Cumulative Default Rates, 1985-2005...9 Exhibit 8 - Broad Letter Grade One-Year-Average Ratings Transition for Europe and North America 1985-2005...10 Exhibit 9 - Broad Letter Grade 2005-Ratings Transition for Europe and North America...11 Exhibit 10 Upgrade-to-Downgrade Ratio and Counts of Upgrade and Downgrade for European Issuers...12 Exhibit 11 Upgrade-to-Downgrade Ratios for European and North-American Issuers...12 Exhibit 12 - European Fallen Angles in 2005...13 Exhibit 13-1-Year Horizon Cumulative Accuracy Plots, Europe vs. North America, 1985-2004 - Percent of Population, Ordered by Rating Category...13 Exhibit 14 - Issuer-Weighted Recovery Rates for European and North-American Issuers*...14 Exhibit 15 - One-Year Credit Losses Based on Broad Letter Rating for European and North American Issuers (1985-2005)...15 Exhibit 16 Rise in Deep Speculative-Grade Issuance since 2000...15 Exhibit 17 - Average Annual Default Rates for European and North American Issuers...20 Exhibit 18 - Trailing 12 Months Value Weighted Speculative-Grade Default Rates for Europe...21 Exhibit 19 - Average Annual Transition Rates for European Issuers (1985-2005)...23 Moody s Special Comment 3

PAGE INTENTIONALLY LEFT BLANK

Introduction By all accounts, 2005 was a benign year for corporate bond and loan defaults in Europe, which is consistent with the positive global credit environment. Only two Moody s-rated issuers in Europe defaulted on bonds (both companies are affiliates within the same corporate family), and one company which was not rated by Moody s defaulted on syndicated loans. Following the peak in corporate defaults in 2002, defaults in Europe have been declining and speculative-grade debt issuance has been increasing. With this backdrop, we present Moody s fifth annual report on the credit and default experience of European corporate bond issuers. We examine the ratings transitions, default experience and default rates, recovery rates, and Moody s rating performance, for the 1985-2005 period. Additionally, we discuss trends in debt issuance and changes in credit quality. Database And Methodology Moody s corporate bond default research is derived from Moody's Default Risk Service Database (DRS), a comprehensive database of corporate defaults, credit ratings and defaulted bond prices for industrial and transportation companies, utilities and financial institutions that have issued long-term debt to the public. As of year-end 2005, nearly 1,200 European issuers held a Moody s rating. The basic unit of study in this report is the corporate bond issuer s individual rating history. 1 Moody s intends its ratings to support credit decisions, which do not generally vary with either the size or number of bonds that a firm may have outstanding. In most cases, when an issuer defaults on one of its bonds, it defaults on all of them. Therefore, Moody s normally reports its default statistics on an issuer-weighted basis, i.e., the fraction of issuers that default, rather than on an issue-" or volume-weighted basis. To the extent that historical default experience is predictive of future experience, issuer-weighted statistics are likely to provide the most reliable guides. If we were to weight our statistics by the number of bond issues or their par amounts, we would be placing undue emphasis on the particular historical experiences of a few large or very frequent issuers of corporate bonds. For the purpose of this study, only corporate bond ratings on domestic- or foreign-currency denominated debt are considered. We exclude any debt that was backed by a guarantor outside the corporate family (e.g. a bond insurer). Moody s global default database, of which the data in this study are a subset, covers over 5,000 defaults by issuers both rated and not rated by Moody s worldwide. Moody s has compiled this information using a variety of sources, various print and online publishing sources, press releases, press clippings, internal memoranda, and records of analyst contacts with rated issuers. Moody s default database includes nearly 250 long-term bond defaults by European issuers, both rated and non-rated, since 1980. Moody s Definition of Default Moody's definition of default is designed to capture credit events that result in impairments to bondholders' contractual claims. Consequently, Moody s definition is strict and includes three types of default events: A missed or delayed disbursement of interest and/or principal, including delayed payments made within a grace period; Filing for bankruptcy, administration, legal receivership, or other legal blocks (perhaps by regulators) to the timely payment of interest and/or principal; or A distressed exchange occurs where: (i) the issuer offers bondholders a new security or package of securities that amount to a diminished financial obligation (such as preferred or common stock, or debt with a lower coupon or par amount); or (ii) the exchange had the apparent purpose of helping the borrower avoid imminent default. This definition is intended to capture events that change the relationship between the bondholder and bond issuer from the relationship which was originally contracted, and which subject the bondholder to an economic loss. We seek to identify only those economic losses that are the result of a credit event. 2 Technical defaults (covenant violations, etc.) are not included in Moody s definition of default. 1. When available, an issuer s senior unsecured rating is used as the reference rating. Absent such a rating, Moody s infers an equivalent senior unsecured rating from the issuer s other rated debt obligations. Additionally, when an issuer has rated debt obligations outstanding in domestic and foreign currency, the lower of the two actual or inferred senior unsecured ratings is used. 2. It is important to note that economic losses suffered by bondholders due to changes in market conditions only are not considered defaults. Nor are losses due to price changes associated with deterioration in the credit quality of the issuer, as long as the terms of the obligation are being met. Moody s Special Comment 5

Exhibit 1 provides information on the number of issuers by whole letter rating category at the end of 2005. While the number of investment-grade issuers in Europe rated by Moody s went up by 39, an increase of about 4.8%, the number of speculative grade issuer went up by 24, an increase of 14%. More dramatic was the increase in issuers rated Caa-C, which went up by more than 75%, from 14 to 25. Almost all of the increase in the Caa-C issuers came from new issuers getting Moody s ratings for the first time. Exhibit 1 - Distribution of European Bond Issuers by Whole Letter Rating 1985 1990 1995 2000 2003 2004 2005 Aaa 17 57 52 62 61 65 65 Aa 2 77 154 249 246 253 274 A 10 43 179 330 390 374 398 Baa 1 2 31 131 192 224 223 Ba 1 3 13 35 52 64 72 B 1 1 8 63 68 93 98 Caa-C 0 0 1 20 21 14 25 Investment Grade 30 179 416 772 889 916 960 Speculative Grade 2 4 22 118 141 171 195 All 32 183 438 890 1030 1087 1155 Exhibit 2 presents the countries of domicile for the universe of European firms represented in this report. The determination of domicile is not made simply on the basis of the country of incorporation. Moody s uses a broad set of criteria to determine the effective as opposed to the legal domicile of rated corporate bond issuers. In addition to legal domicile, we also consider the jurisdiction under which a firm might file for bankruptcy and the location of the majority of its revenues and assets. The increasing use of domiciles of convenience such as Bermuda and the growth of transnational corporate families make the determination of the true country of domicile more challenging. Throughout this study, we compare the rating and default experiences of European issuers to those of North American issuers, i.e. corporations that are domiciled in the United States and Canada. Exhibit 2 - Countries of Domicile of European Moody's- Rated Corporate Bond Issuers Countries of Domicile Austria Hungary Poland Belgium Iceland Portugal Bulgaria Ireland Romania Croatia Isle of Man Russia Cyprus Israel Scotland Czech Republic Italy Slovak Republic Denmark Latvia Slovenia Estonia Liechtenstein Spain Finland Lithuania Sweden France Luxembourg Switzerland Germany Moldova United Kingdom Greece Netherlands Guernsey Norway 6 Moody s Special Comment

The 2005 European Corporate Default Experience The number of European defaults, and resultant default rates, continued to decline in 2005. After peaking in 2002, the number of defaults has been declining steadily, and in 2005 there were only two Moody s-rated bond defaults. While the number of rated bond defaults was the same in 2005 and 2004, the default volume in 2005 was much smaller, 290 million compared to 1.4 billion in 2004. Exhibit 3 provides a time series of defaulted volume and number of defaulting bond issuers, both rated and unrated. Exhibit 3 - European Bond Default Counts and Volumes Year Rated Count Unrated Count Total Count Rated Volume Unrated Volume Total Volume 1990 1 0 1 561 0 561 1991 2 1 3 1,333 191 1,524 1992 0 1 1 0 134 134 1993 0 2 2 0 133 133 1994 1 3 4 554 1,423 1,977 1995 0 7 7 0 967 967 1996 0 3 3 0 1,358 1,358 1997 0 3 3 0 183 183 1998 2 3 5 570 402 972 1999 10 4 14 2,354 83 2,437 2000 3 1 4 804 19 823 2001 17 8 25 10,829 511 11,340 2002 29 4 33 43,628 575 44,203 2003 9 8 17 3,511 6,189 9,700 2004 2 3 5 1,400 491 1,891 2005 2 0 2 290 0 290 Total 78 51 129 65,834 12,659 78,493 Exhibit 4 presents the list of European defaulters in 2005. Both Moody s rated bond defaults Concordia Bus A and Concordia Bus Nordic AB began as missed interest payments that were resolved through the completion of distressed exchanges totaling 370 million. Both issuers were members of the same corporate family and are located in Sweden. The only syndicated loan default recorded by Moody s was Jarvis PLC of the UK, which defaulted on 470 million of loans and entered into a distressed exchange with its lenders. Jarvis was not rated by Moody s. Exhibit 4 - List of the 2005 European Defaulters Name of Defaulter Country Default Amount in Million Rated by Moody's Bond Defaults Concordia Bus AB Sweden 130 Yes Concordia Bus Nordic AB Sweden 160 Yes Total Bond Defaults 290 Loan Defaults Jarvis PLC UK 470 No Total Loan Defaults 470 Moody s Special Comment 7

European Speculative-Grade Bond Default Rates The small number of defaults in 2005 continued the downward trend in defaults after the number and volume of defaults peaked in 2002. As a result, both issuer-weighted and value-weighted default rates have declined sharply, as is evident in Exhibit 5 (reproduced on the cover page), which presents the time series of trailing 12-month speculativegrade default rates since January 2000. 70% Exhibit 5 - Trailing 12-Months European Issuer-Weighted and Volume-Weighted Speculative-Grade Bond Default Rates 60% 50% 40% Issuer Weighted Value Weighted 30% 20% 10% 0% Sep-05 May-05 Jan-05 Sep-04 May-04 Jan-04 Sep-03 May-03 Jan-03 Sep-02 May-02 Jan-02 Sep-01 May-01 Jan-01 Sep-00 May-00 Jan-00 The peak in default volume was reached in 2002 when 33 issuers defaulted on over 44 billion. Default activity has declined considerably since hitting that high water mark. The issuer-weighted speculative-grade default rate for European issuers stood at just 1.09% as of December 2005. Similar declines in issuer-weighted speculative-grade default rates have also been observed on a global basis and for issuers located in the US and Canada. On a volumeweighted basis, the decline in the European speculative-grade default rate has been even more extreme, falling from 58.2% in January 2003 to just 0.44% in December 2005 a decline of more than 99%. Exhibit 6 - Annual Issuer-Weighted Speculative-Grade Default Rates for European and North American Issuers 25% 20% Europe North America 15% 10% 5% 0% 2005 2003 2001 1999 1997 1995 1993 1991 1989 1987 1985 8 Moody s Special Comment

Historically, European speculative-grade defaults rates have been more variable than those of North America because the speculative-grade universe itself was very small in Europe prior to 1996. As such, a small number of defaults would result in a very high default rate. For example, the default rate in 1991 was 25% for Europe because there were only eight rated speculative-grade issuers, out of which two defaulted. However, with an increasing number of speculative-grade issuers now issuing debt, the speculative-grade default rate for Europe is becoming a more meaningful indicator of the overall credit environment in Europe. Cumulative Default Rates Investors are also interested in default rates for investment horizons longer than one year. Exhibit 7 presents 1-year through 5-year issuer-weighted average cumulative default rates for European issuers. Also presented are the North American (US and Canada) statistics for comparison. In previous years, Moody's European default study has presented historical average default and rating migration rates that were calculated using cohorts formed at annual frequencies. In this year's report, we increase the underlying sample by using a monthly cohort periodicity. Monthly cohorts have the advantage of capturing rating changes that occur within a calendar year. The default rates are calculated based on cohorts of all issuers holding a given estimated senior rating at the start of a given month. The cohorts are dynamic in that they change based on whether these issuers leave the cohort due to default or non credit-related reasons (e.g., maturing of debt). The dynamic nature of the cohorts allows the estimation of cumulative default risk over multi-year horizons which is free of survivorship bias. It also allows for the comparison and averaging of default rates over different periods. By forming and tracking such cohorts of all Moody's-rated issuers with debt outstanding as of beginning of every month, we replicate the experience of a portfolio of both seasoned and new-issue bonds purchased in a month. While the cohort frequency is monthly, the accumulation periodicity is still 12-months, in that we track default rates over horizons of 1-year, 2-year, 3-year, etc. 3 Exhibit 7 - European and North American Issuer-Weighted Cumulative Default Rates, 1985-2005 Europe Year 1 Year 2 Year 3 Year 4 Year 5 Aaa 0.00% 0.00% 0.00% 0.00% 0.00% Aa 0.00% 0.00% 0.00% 0.00% 0.00% A 0.02% 0.09% 0.14% 0.18% 0.26% Baa 0.28% 0.70% 1.26% 1.82% 2.32% Ba 1.22% 2.78% 4.16% 4.72% 4.97% B 5.07% 13.50% 21.28% 28.51% 35.65% Caa-C 28.73% 42.87% 55.95% 62.15% 62.80% Investment-Grade 0.05% 0.14% 0.23% 0.31% 0.39% Speculative-Grade 6.78% 13.36% 19.33% 23.62% 26.60% All Corporates 0.75% 1.46% 2.02% 2.38% 2.63% North America Year 1 Year 2 Year 3 Year 4 Year 5 Aaa 0.00% 0.00% 0.00% 0.00% 0.04% Aa 0.00% 0.01% 0.06% 0.13% 0.18% A 0.03% 0.14% 0.32% 0.50% 0.67% Baa 0.23% 0.65% 1.17% 1.78% 2.39% Ba 1.44% 4.01% 6.99% 9.92% 12.54% B 5.86% 12.76% 19.29% 25.01% 30.33% Caa-C 20.11% 32.32% 41.97% 50.17% 56.76% Investment-Grade 0.09% 0.28% 0.55% 0.84% 1.12% Speculative-Grade 5.33% 10.81% 15.95% 20.46% 24.42% All Corporates 2.15% 4.34% 6.35% 8.06% 9.49% 3. See Appendix I for the detailed methodology. Moody s Special Comment 9

A comparison between Europe and North America shows that while investment-grade default rates are lower in Europe, average speculative-grade default rates are somewhat higher for each investment horizon from 1 to 5 years. The differences in cumulative default rates between Europe and North America are explained by differences in the distribution of rating in the two regions. Since the proportion of investment-grade issuers is larger in Europe and they have lower default rates when compared to issuers in the North America, the overall default rate has been lower in Europe than in North America. Trends In Credit Quality In addition to the default rates, rating migration rates and upgrade-to-downgrade ratios are also important indicators of changes in credit quality. Rating Migration Rates Exhibit 8 shows average annual, whole-letter rating migration rates since 1985. Each cell in the matrix is the weighted average fraction of issuers who held a given row s rating at the beginning of the measurement period and the column rating at the end of the period, including defaults and withdrawn ratings (WRs). The weights correspond to the size (number of issuers) of the annual cohorts. North American statistics are presented alongside for comparison. The largest values in the transition matrix are along the diagonal, reflecting that the most likely rating for an issuer at the end of a given year during the period under discussion is the rating with which the issuer began the year. For example, a European issuer rated Aaa at the beginning of a given year has an 89.4% historical probability of remaining Aaa at the end of the same year. By contrast, those elements that are off the diagonal reflect transitions to higher (the triangle below the diagonal) or lower (the triangle above the diagonal) rating categories within one year. For example, an Aa-rated issuer has a historical probability of 7.7% of being downgraded to an A rating at the end of one year and a 0.6% historical probability of being upgraded to Aaa over the same period. The farther one moves away from the diagonal, the smaller the migration rates, thereby generally reflecting a relatively low historical probability of issuers moving across more than one rating category during the course of a year. Exhibit 8 - Broad Letter Grade One-Year-Average Ratings Transition for Europe and North America - 1985-2005 Europe Rating from Rating to Aaa Aa A Baa Ba B Caa-C Default WR Aaa 89.4% 7.6% 0.3% 0.0% 0.1% 0.0% 0.0% 0.0% 2.6% Aa 0.6% 88.6% 7.7% 0.1% 0.0% 0.0% 0.0% 0.0% 3.0% A 0.0% 2.7% 88.7% 4.4% 0.2% 0.1% 0.0% 0.0% 3.9% Baa 0.0% 0.3% 6.7% 78.7% 4.5% 1.5% 0.1% 0.1% 8.2% Ba 0.0% 0.0% 0.9% 5.8% 76.3% 9.8% 0.5% 1.0% 5.7% B 0.0% 0.0% 0.3% 1.0% 6.6% 67.4% 9.4% 3.4% 12.0% Caa-C 0.0% 0.0% 0.0% 0.0% 0.3% 19.1% 47.3% 20.3% 13.0% North America Rating From Rating to Aaa Aa A Baa Ba B Caa-C Default WR Aaa 82.0% 9.8% 1.6% 0.0% 0.0% 0.0% 0.0% 0.0% 6.6% Aa 0.0% 91.7% 2.1% 0.3% 0.3% 0.0% 0.0% 0.0% 5.6% A 0.3% 1.1% 87.9% 4.8% 0.2% 0.0% 0.0% 0.0% 5.7% Baa 0.0% 0.3% 2.2% 87.6% 4.0% 1.0% 0.0% 0.3% 4.8% Ba 0.0% 0.0% 0.8% 7.4% 77.0% 9.0% 0.3% 0.0% 5.6% B 0.0% 0.0% 0.2% 0.0% 6.0% 78.4% 5.0% 1.3% 9.1% Caa-C 0.0% 0.0% 0.4% 0.8% 1.2% 21.4% 56.0% 7.0% 13.2% 10 Moody s Special Comment

European ratings display high degree of stability across various rating categories. On average, 85% of all rated issuers retain their beginning-of-year ratings. European ratings have also exhibited somewhat greater stability than North American ratings. For example, single A-rated European issuers have experienced a 4.4% historical probability of downgrade to a Baa rating, compared to a 4.8% probability for North American issuers. Moreover, most upgrades or downgrades in one year involve a change of just a single whole letter rating category. In contrast to Exhibit 8, which focuses on the long-term historical average rating migration experience, Exhibit 9 summarizes the one-year rating transition experience of European and North American issuers for just the year 2005. Exhibit 9 - Broad Letter Grade 2005-Ratings Transition for Europe and North America Europe Rating From: Rating to Aaa Aa A Baa Ba B Caa-C Default WR Aaa 100.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Aa 0.0% 97.7% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 2.3% A 0.0% 1.7% 87.9% 7.2% 0.0% 0.0% 0.0% 0.0% 3.3% Baa 0.0% 0.0% 11.6% 80.4% 1.7% 1.7% 0.0% 0.0% 4.6% Ba 0.0% 0.0% 1.7% 6.8% 74.6% 5.1% 0.0% 0.0% 11.9% B 0.0% 0.0% 0.0% 0.0% 8.9% 72.2% 3.3% 0.0% 15.6% Caa-C 0.0% 0.0% 0.0% 0.0% 4.0% 32.0% 56.0% 8.0% 0.0% North America Rating From: Rating to Aaa Aa A Baa Ba B Caa-C Default WR Aaa 82.0% 9.8% 1.6% 0.0% 0.0% 0.0% 0.0% 0.0% 6.6% Aa 0.0% 91.7% 2.1% 0.3% 0.3% 0.0% 0.0% 0.0% 5.6% A 0.3% 1.1% 87.9% 4.8% 0.2% 0.0% 0.0% 0.0% 5.7% Baa 0.0% 0.3% 2.2% 87.6% 4.0% 1.0% 0.0% 0.3% 4.8% Ba 0.0% 0.0% 0.8% 7.4% 77.0% 9.0% 0.3% 0.0% 5.6% B 0.0% 0.0% 0.2% 0.0% 6.0% 78.4% 5.0% 1.3% 9.1% Caa-C 0.0% 0.0% 0.4% 0.8% 1.2% 21.4% 56.0% 7.0% 13.2% The credit quality of European issuers continued to improve in 2005, as shown by the number of upgrades for various rating categories. For example, 11.6% of the Baa-rated issuers were upgraded to single-a in 2005, considerably higher than the long-term historical upgrade rate of Baa issues, which is 6.7%. Similarly, 8.9% of the B-rated issuers were upgraded to Ba, compared to 6.6% historically. A large portion (32%) of issuers rated Caa-C was upgraded to B, well above the long-term average of 19.1%. Moody s Special Comment 11

Upgrade-to-Downgrade Ratios Changes in credit quality can also be gauged by the upgrade-to-downgrade ratio. An increase in this ratio indicates improving credit quality, while a decrease is an indicator of deteriorating credit quality. Exhibit 10 presents the time series of this metric. 4 Exhibit 10 - Upgrade-to-Downgrade Ratio and Counts of Upgrade and Downgrade for European Issuers 250 1.6 1.4 Upgrades and Downgrades 200 150 100 50 Upgrades Downgrades Upgrade-to-Downgrade Ratio 1.2 1.0 0.8 0.6 0.4 0.2 Upgrade-to-Downgrade Ratio 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Year 0.0 As evident from the exhibit, average credit quality declined sharply in 2001, 2002, and 2003 in Europe, as it did worldwide. Since that time, upgrades have narrowly exceeded downgrades. In Europe in 2005, for every issuer that was downgraded, roughly 1.1 issuers were upgraded. Moody's believes that the pace of improvement in credit quality will most likely decline in the next couple of quarters, as the share of issuers that are currently on watch for upgrade (2.6%) is considerably lower than the share of issuers on review for possible downgrade (4.1%). The moderation in improvement in European credit quality is more apparent in Exhibit 11, which plots the upgradeto-downgrade ratio on a quarterly basis. In the fourth quarter of 2005, European downgrades greatly exceed upgrades Exhibit 11 - Upgrade-to-Downgrade Ratio for European and North American Issuers 4.50 4.00 3.50 3.00 2.50 North America Europe 2.00 1.50 1.00 0.50 - Q4-05 Q3-05 Q2-05 Q1-05 Q4-04 Q3-04 Q2-04 Q1-04 Q4-03 Q3-03 Q2-03 Q1-03 Q4-02 Q3-02 Q2-02 Q1-02 Q4-01 Q3-01 Q2-01 Q1-01 Q4-00 Q3-00 Q2-00 Q1-00 4. Moody s Investors Service, Special Comment, Moody s Rating Actions, Reviews and Outlooks: Quarterly Update, February 2006 12 Moody s Special Comment

While overall rating quality improved in Europe, five investment-grade issuers (listed in Exhibit 12) were downgraded to speculative grade and became fallen angels in 2005. 5 All except one of these downgrades from investment grade to speculative grade involved movements from Baa3 to Ba1. The exception, TDC A/S, was a three-notch downgrade from Baa1 to Ba1. Exhibit 12 - European Fallen Angels in 2005 Company Country From To Basell Finance Company B.V. Netherlands Baa3 Ba2 Basell America Finance Corporation Netherlands Baa3 Ba2 FKI plc United Kingdom Baa3 Ba2 TDC A/S Denmark Baa1 Ba1 Vauxhall Motors (Finance) PLC United Kingdom Baa3 Ba2 Ratings Performance The default rate analysis presented above demonstrates that Moody s European corporate bond ratings are correlated with subsequent default experience at investment horizons of up to five years. One-year-look-ahead cumulative accuracy profiles (CAPs) plotted in Exhibit 13 provide another set of graphical tools for visualizing the power of ratings to anticipate defaults. A CAP curve is constructed by first sorting the Moody s-rated universe of corporate bond issuers by rating category, from highest to lowest credit risk, and then plotting the share of defaulters accounted for by issuers carry any particular rating or lower. CAP plot for North America is again presented for comparison purposes. Exhibit 13-1-Year Horizon Cumulative Accuracy Plots, Europe vs. North America, 1985-2005 - Percent of Population, Ordered by Rating Category Europe North America 100% Ba3-Baa3 Baa2 Baa1 A3 A2 A1 Aa3 Aa2 Aa1 Aaa 100% Share of Defaults 80% 60% 40% B2 B1 B3 Caa-C Share of Defaults 80% 60% 40% B3 B2 Caa-C Ba1 Baa3 Baa2 Baa1 A3 A2 A1 Aa3Aa2Aa1 Ba2 Aaa Ba3 B1 20% One Year Random Allocation 20% One Year Random Allocation 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Share of Issuers 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Share of Issuers Exhibit 13 shows that, when sorted by risk decile, Moody s European ratings have outperformed ratings in North America at predicting default. For example, nearly 95% of all defaults are captured in the first decile of ratings in Europe, compared to a little over 60% in the North America over the same one-year horizon. The greater power of the European ratings to discriminate default risk follows directly from two features of the data: European defaulters have had lower average ratings than North America defaulters prior to default, and non-defaulting European issuers have been rated higher on average than North America non-defaulters. Given the smaller speculative-grade share of the European corporate market compared to the corporate market in the US, it may in fact have been historically easier to identify likely defaulters in Europe. 5. Three issuers, Ford Capital, B.V and GMAC International Finance B.V. both located in the Netherlands, and GMAC Bank Gmbh, located in Germany, were also downgraded to speculative-grade. However, they are not included in the exhibit as they are financing arms for companies that do a majority of their business in the US. Vauxhall Motors (Finance) PLC, though a subsidiary of a US company (General Motors), is included because its operations are largely European. Moody s Special Comment 13

Recovery Rates For European Defaulted Instruments Moody s ratings are statements about the probability of default and the expected loss severity rate (i.e., one minus the expected recovery rate) in case of default. Since the probability of default is considered to be the same across the various obligations of an issuer, differences in ratings across an issuer s capital structure primarily reflect different expectations of recovery rates in the event of default. Commensurate with our practice in our annual corporate bond default study, in Exhibit 14 we estimate recovery rates using 30-day post-default bid prices on defaulted bonds, though no trades might have taken place at some of these prices. The data set includes both bonds rated by Moody's as well as non-rated issues where available. Since there were no prices available for the only European defaulted bonds of 2005 (Concordia Bus AB and Concordia Bus Nordic AB), Exhibit 14 presents long-term historical averages and detail for 2003 and 2004 only. 6 Exhibit 14 - Issuer-Weighted Recovery Rates for European and North American Issuers 1982-2005* 2003 2004 Instrument Europe North America Europe North America Europe North America Bonds - Senior Secured Bonds 52.7% 62.6% 40.5% 66.4% N.A. 66.0% - Senior Unsecured Bonds 26.0% 38.4% 16.6% 40.3% 95.8% 47.2% - Sr. Subordinated Bonds 40.6% 32.0% N.A. 37.9% N.A. 49.3% - Subordinated Bonds 35.3% 31.0% 8.8% 31.3% 98.8% N.A. - Jr. Subordinated Bonds N.A. 23.9% N.A. N.A. N.A. N.A. All Bonds 32.3% 40.1% 19.9% 39.8% 97.3% 51.4% *The statistics reported for Europe are only for the 1982-2004 period as no prices for European bonds that defaulted in 2005 were available. Credit Losses In previous sections we dealt with two main components of credit loss, the probability of default and the severity of loss given default (one minus recovery rate). We showed that Moody s long-term credit ratings accurately rank order default risk and default severity in a given time period. As the ratings decline, not only does the default probability go up, but the average loss severity increases on a portfolio basis. In this section, we bring these two concepts together to discuss credit losses and how they are affected by default rates and recovery rates. Additionally, we demonstrate that Moody s ratings effectively differentiate credit loss rates. Moody s rating process is designed to produce a consistent measure of relative credit risk, which in large measure is determined by expected credit losses. Expected credit losses are defined mathematically as follows: Credit Loss Rate= (Default Frequency)*(1-Recovery Rate) Credit losses can be defined as the loss incurred in total return by a bond portfolio due to defaults. Exhibit 15 presents average annual credit losses for portfolios based on broad letter rating categories. While credit losses for Ba- and B-rated issuers are fairly similar across the two regions, loss rates for Caa-C rated issuers in Europe are much higher than those for North American issuers due to two reasons: the one-year default rates for the Caa-C rated issuers are higher and recovery rates for defaulted European bonds are lower. 6. A restructuring plan was proposed by the two defaulting issuers where, depending on the participation of bondholders, all bondholders will receive equity worth only 25% of face value of debt in a new company. Additionally, the bondholders will have to forgo any interest accrued and any other claims related to these bond defaults. 14 Moody s Special Comment

Exhibit 15 - One-Year Credit Losses Based on Broad Letter Rating for European and North American Issuers 1985-2005 25% Credit Losses 20% 15% 10% North America Europe 5% 0% Aaa Aa A Baa Ba B Caa-C Broad Letter Rating Looking Ahead (2006-2008) As discussed earlier, the credit environment remained benign in 2005. Defaults were virtually non-existent, and credit quality as measured by upgrade-to-downgrade ratio improved. Borrowers took advantage of this benign environment and issued significant amount of new debt or rolled over debt that matured in 2005. While spreads were marginally wider in 2005 than in 2004, issuance activity continued unabated and near-record amount of speculative debt was issued. By Moody s estimates, about 20 billion of speculative-grade bonds were issued, which is about 3 billion less than the high water mark set in 2004. The quality of new speculative-grade issuance declined sharply in 2005. A benign credit environment, coupled with narrow spreads, allowed many firms to issue debt that was rated in the lower end of the rating spectrum. Exhibit 16 presents European speculative-grade issuance by credit quality over the past ten years. Exhibit 16 - European Speculative-Grade Issuance 100% 80% Ba 100% 80% 60% 60% 40% B 40% 20% 20% Caa Not Rated 0% 0% 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source: Moody's Investors Service compilations After peaking in 1998 and 1999 the B- and Caa-rated issuance had declined significantly until 2003, whereas the share of higher quality speculative-grade debt (Ba-rated) increased substantially in 2001, 2002 and 2003. However, the trend reversed in 2004 when B-rated debt constituted more than 70% of the debt. Financing for bonds in the lowest rating tiers, Caa-rated and even non-rated debt, also rose sharply in 2004. Moody s Special Comment 15

While the volume of speculative-grade issuance declined modestly in 2005, the deep speculative-grade portion of the issuance jumped dramatically. The Caa-rated share of the new issuance increased to over 30% in 2005, representing a 100% increase in share over Caa issuance in 2004 and a 1500% increase over 2003 issuance. The jump in the Caa issuance came at the expense of single-b rated share of issuance, while the Ba-rated issuance shared remained approximately same. Moody s has observed a negative correlation between the quality of issuance and subsequent default rates with a lag of approximately 3-years. This lagged relationship (also known as the seasoning effect) has significant predictive power in Moody s global default rate forecasting model. For this and other reasons, Moody s predicts that the global speculative-grade default rate will rise from 2005's 1.9% to 3.3% in 2006. 7 While Moody s does not forecast default rates regionally, changes in Moody's European default rate are historically positively correlated with changes in the global default rate. Hence, Moody's expects that the European default rate will likely follow the global trend, particularly as a result of the low quality of issuance observed in 2004 and 2005. Summary And Conclusions Commensurate with the benign global credit of the environment last year, issuers in Europe experienced extremely low default activity. Moody s recorded only two rated bond defaults and both issuers were part of the same corporate family. The total defaulted bond volume was 290 million in 2005. Only one loan default (by an issuer not rated by Moody s) was recorded. The issuer-weighted speculative-grade bond default rate declined to about 1.1%, which is even lower than the 1.2% default rate observed in 2004. On a volume-weighted basis, the default rate declined even further, to less than 0.5%. No investment-grade default was recorded in 2005, resulting in the investment-grade default rate of 0% in Europe. In 2005, average credit quality improved in Europe as evident from the ratio of upgrades to downgrades. Five issuers were downgraded from investment-grade to speculative-grade. The improved credit environment allowed many speculative-grade issuers to issue debt in Europe in 2005, as roughly 20 billion of new debt was issued. Over 90% of this debt was rate B or below with Caa rated issuance jumping by more than 100% from its issuance level in 2004. Default activity will likely pick up in coming years, given the high levels of low speculative-grade issuance in 2004 and 2005, perhaps similar to the pattern of issuance and default observed in 1999-2002. 7. Moody s Investors Service Special Comment Default and Recovery Rates of Corporate Bonds Issuers (1920-2005), January 2006 16 Moody s Special Comment

Related Research Special Comments: Predicting Default Rates: A Forecasting Model for Moody's Issuer-Based Default Rates, August 1999 (47729) Measuring the Performance of Corporate Bond Ratings, April 2003 (77916) Default and Recovery Rates of European Corporate Bond Issuers (1985-2004), March 2005 (91623) The Performance of Moody's Corporate Bond Ratings: December 2005 Quarterly Update, January 2006 (96330) Default and Recovery Rates of Corporate Bonds Issuers (1920-2005), February 2006 (96546) Moody's Rating Actions, Reviews and Outlooks: Quarterly Update - Fourth Quarter 2005, February 2006 (96554) To access any of these reports, click on the entry above. Note that these references are current as of the date of publication of this report and that more recent reports may be available. All research may not be available to all clients. Moody s Special Comment 17

Appendix I - Methodology REPRODUCED FROM MOODY S SPECIAL COMMENT DEFAULT AND RECOVERY RATES OF CORPORATE BONDS ISSUERS (1920-2005) FEBRUARY 2006 Estimated Senior Ratings The credit ratings used in Moody's annual default studies do not refer to the ratings of any specific debt obligations. Rather, they are notional, issuer-level ratings derived from each issuer's outstanding rated bonds. The resulting ratings are called estimated senior unsecured ratings or, more concisely, estimated senior ratings. Estimated senior ratings are derived by a statistical algorithm: first, the rating on each company's senior unsecured bonds is used; if an issuer does not have any rated senior unsecured bonds, the estimated senior rating is derived by statistically implying such a rating on the basis of rated subordinated or secured bonds, or senior implied ratings. Deriving estimated senior ratings is necessary because Moody's credit ratings for corporate bonds are based on expected credit loss, which is comprised of (1) the expected probability of default, as well as (2) the expected severity (loss of principal and interest) in the event of default. Therefore, in order to calculate historical default rates by rating category which are estimates of the expected default probability component, one must hold severity considerations constant. In most cases, estimated senior ratings will yield an assessment of default risk that is relatively unaffected by special considerations of collateral or of position within the capital structure. Default Rate Calculations Moody s employs the standard method used by statisticians to analyze cumulative survival (and hence default) probabilities using discrete time data. 8 In calculating default rates for individual cohorts, Moody s adjusts the denominator of the default rate downward for issuers whose ratings are withdrawn to account for survivorship bias. Once an issuer s rating is withdrawn, the observation of its future default experience is treated as a censored observation. 9 By way of example, if while measuring a 10-year CDR an issuer's rating was withdrawn in year 5, certainty regarding default history can only be gained for the period of time between the time the cohort was formed and year 5. The issuer would therefore be considered "exposed" to default for 5 years, not 10 years. Cumulative default rates (CDRs) are calculated from marginal default rates (MDRs), which represent the probability that an issuer that has survived in the sample through a particular date will default over a short interval following that date. By definition, the denominators used to estimate MDRs must adjust for withdrawals and defaults prior to that date. We use the following formal for calculating MDRs, which also makes a very minor adjustment (and inconsequential) adjustment for the withdrawals that occur during the short interval in question. The marginal default rate, d(t), is the ratio of the number of defaulting issuers in period t divided by the number of issuers exposed to the risk of default in period t: x( t) d( t) = n( t) x( t 1) 1 [ w( t) + w( t 1)] 2 Cumulative default rates, D(T), for horizons of length T, for specific cohorts are built up from the MDRs as follows: T 1 1 + (1 d1) d 2 + (1 d1)(1 d 2 ) d3 +... + (1 ) t 1 D T d ( ) = = d t d T 8. The classic reference is Cutler and Ederer, Maximum Utilization of the Life Table Method in Analyzing Survival, Biostatistics, December 1958. This approach is the discrete time analogue of the continuous hazard rate model which is familiar to most economists. 9. Moody s does attempt to follow these companies subsequent experience and occasionally becomes aware of instances that issuers with withdrawn ratings default. It is not possible, however, to use this information in an unbiased way because we do not know what happens to all withdrawn issuers. Moreover, issuers that have their ratings withdrawn because they no longer have any debt outstanding and are liquidated cannot be used to gain insight to the question: what would happen if I have credit exposure to an issuer beyond that date of liquidation? 18 Moody s Special Comment

For example, the two year cumulative default rate is the sum of the marginal default rate in year one (d 1 ) and the marginal default rate in year two (d 2 ) for issuers that survived past year one, (1 d 1 ). When calculating average CDRs over many cohorts, the above formula is also used, but the MDRs are constructing by averaging across cohorts. In this manner, when estimating long-term CDRs, information from recent cohorts need not be ignored entirely. Rather, near-term cohorts are used to obtain more precise estimates of the near-term marginal default rates used in deriving long-term CDRs. Our default rate calculation method has two important advantages over the more simplistic so-called static pool method used by some researchers: If you expect to have a ten-year exposure to an issuer, you want an estimate of the risk of ten years of exposure. Our method produces an unbiased estimate of this risk; whereas, the static pool method produces downwardly biased default rates estimates since it implicitly assumes that issuers that leave the sample would never have defaulted if they had stayed in the sample. If you want to use all the available information in your dataset to estimate long-term CDRs, you need to use the information from recent cohorts, not just cohorts that have history that span the entire horizon of the CDR calculation. Our method using such information in estimating short-term MDRs that enter into the long-term CDR calculations; whereas, the static approach simply ignores that data. Recovery Rate Estimates Moody's generally measures recovery rates in this report using bid prices on defaulted debt obligations observed roughly 30 days after the date of default, specifically measured as a ratio of prices relative to par. Using a post-default price to measure loss severity parallels common practice in the credit default swaps market. Moreover, recovery rates measured in this way are most relevant for the many cash bond investors who liquidate their holdings shortly after default as required by their portfolio governance rules or their own investment objectives. For investors holding defaulted securities until ultimate resolution, prices observed shortly after default are generally accepted as the market's discounted expected ultimate recovery rates. The alternative approach measuring ultimate recoveries realized in bankruptcy resolutions presents a number of estimation challenges including estimation of the discount rate to apply to each cash-flow, assumptions about the value of certain payments made during the process, and estimates of values of securities used to pay creditor claims, which may include illiquid new securities (e.g., equity and derivative instruments) as well as physical assets. The table below defines the various ways Moody's sometimes calculates and presents its recovery rate estimates. For the purposes of measuring expected loss rates, we rely on issuer-weighted mean recovery rates. Term Issuer-Weighted Mean Recovery Rates Value-Weighted Mean Recovery Rates Issuer-Weighted Median Recovery Rates Issue-Weighted Mean Recovery Rates Definition Issuer-weighted mean recovery rates are derived by estimating mean recovery rates for each issuer, then averaging them across issuers. They are useful for predicting recovery rates for portfolios that are well diversified across issuers. Value-weighted recovery rates represent the average of recovery rates on all defaulted issuers, weighted by the face value of those issues. These estimates are useful for predicting recovery rates on the market portfolio. Issuer-weighted median recovery rates are estimated as median of issuer-weighted recovery rates and are used for predicting the most likely recovery rate for a randomly selected issuer. Issue-weighted recovery rates are estimated using recovery rates for each issue and taking the average of all issues. While this measure is widely reported, it is useful only for predicting the average recovery rate on a portfolio of default bonds diversified across issues but without reference to issuer or issue size. Moody s Special Comment 19

20 Moody s Special Comment Appendix II - Supplemental Default and Transition Rates Tables for European Issuers Exhibit 17 - Average Annual Default Rates for European and North American Issuers Europe 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Aaa 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Aa 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% A 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.29% 0.00% 0.00% 0.00% 0.00% Baa 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 1.70% 0.00% 0.00% 0.00% Ba 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 3.92% 0.00% 0.00% 2.82% 10.53% 0.00% 0.00% 0.00% B 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 25.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 9.20% 1.61% 10.39% 9.76% 0.00% 0.00% 0.00% Caa-C 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50.00% 0.00% 0.00% 0.00% 0.00% 33.33% 10.53% 32.56% 56.14% 45.00% 16.00% 8.00% Investment-Grade 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.12% 0.35% 0.00% 0.00% 0.00% Speculative-Grade 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 20.00% 0.00% 0.00% 6.25% 0.00% 0.00% 0.00% 1.82% 11.11% 2.62% 11.94% 20.31% 6.64% 1.25% 1.09% All Corporates 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.48% 0.00% 0.00% 0.27% 0.00% 0.00% 0.00% 0.15% 1.29% 0.35% 1.80% 2.93% 0.90% 0.19% 0.18% North America 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Aaa 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Aa 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% A 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.28% 0.29% 0.00% 0.00% 0.00% Baa 0.00% 1.37% 0.00% 0.00% 0.61% 0.00% 0.29% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.15% 0.56% 0.28% 1.33% 0.00% 0.00% 0.28% Ba 1.42% 2.05% 2.72% 1.26% 2.80% 3.17% 5.41% 0.32% 0.61% 0.28% 0.83% 0.00% 0.24% 0.44% 1.20% 1.06% 1.47% 1.16% 1.12% 0.28% 0.00% B 8.28% 11.80% 6.27% 6.38% 8.71% 16.29% 14.52% 9.22% 5.94% 4.01% 4.91% 1.42% 2.31% 4.62% 5.81% 6.44% 8.87% 4.66% 2.41% 0.81% 1.32% Caa-C 0.00% 26.67% 22.22% 31.58% 25.00% 58.82% 36.84% 26.67% 28.57% 2.74% 12.61% 14.60% 10.07% 15.23% 20.25% 21.85% 33.66% 24.00% 19.86% 13.10% 7.42% Investment-Grade 0.00% 0.35% 0.00% 0.00% 0.17% 0.00% 0.09% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.06% 0.23% 0.22% 0.66% 0.00% 0.00% 0.11% Speculative-Grade 3.71% 5.70% 4.26% 3.63% 5.61% 10.07% 10.40% 5.03% 3.76% 2.01% 3.65% 1.85% 1.96% 3.87% 5.79% 7.31% 10.87% 7.17% 5.23% 2.73% 2.08% All Corporates 1.05% 2.02% 1.61% 1.48% 2.42% 4.16% 3.87% 1.66% 1.25% 0.72% 1.39% 0.70% 0.77% 1.59% 2.55% 3.35% 4.66% 3.18% 2.10% 1.13% 0.91%

Exhibit 18 - Trailing 12 Months Value Weighted Speculative-Grade Default Rates for Europe 12-Months Ending Speculative-Grade Default Volume Speculative-Grade Total Volume Speculative-Grade Default Rate 12/31/05 377 85405 0.44% 11/30/05 1049 85106 1.23% 10/31/05 1049 78483 1.34% 09/30/05 1049 75655 1.39% 08/31/05 1049 74823 1.40% 07/31/05 1049 75350 1.39% 06/30/05 1049 75340 1.39% 05/31/05 1049 81636 1.28% 04/30/05 1049 79628 1.32% 03/31/05 1049 79112 1.33% 02/28/05 1049 78306 1.34% 01/31/05 672 76264 0.88% 12/31/04 1809 78121 2.32% 11/30/04 1138 74813 1.52% 10/31/04 1138 71634 1.59% 09/30/04 1138 71967 1.58% 08/31/04 1138 65698 1.73% 07/31/04 1138 69068 1.65% 06/30/04 1138 67547 1.68% 05/31/04 1710 68311 2.50% 04/30/04 2627 65582 4.01% 03/31/04 2627 63406 4.14% 02/29/04 2777 63775 4.35% 01/31/04 3848 58860 6.54% 12/31/03 3909 57971 6.74% 11/30/03 5798 49179 11.79% 10/31/03 5798 49883 11.62% 09/30/03 5886 43983 13.38% 08/31/03 10430 47445 21.98% 07/31/03 14387 50937 28.24% 06/30/03 15651 45865 34.12% 05/31/03 17259 43355 39.81% 04/30/03 21077 47398 44.47% 03/31/03 29972 56122 53.41% 02/28/03 31060 57795 53.74% 01/31/03 36428 62562 58.23% 12/31/02 23792 49510 48.05% Moody s Special Comment 21