Credit Policy. Special Comment. Default and Recovery Rates of Asia-Pacific Corporate Bond and Loan Issuers, Excluding Japan, 1990-H1 2009

Transcription

1 Special Comment Mood s Global Credit Polic August 2009 Table of Contents: Summar 1 Data and Methodolog 2 Issuer Characteristics 3 Trends in Credit Qualit: Transition Matrices and Rating Changes 6 Historical Corporate Defaults 8 Cumulative Default Rates 10 Regional Differences in Rating Transitions and Cumulative Default Rates 12 Rating Performance Measures 14 Recover Rates in Default 15 Credit Loss Rates 15 Mood s Related Research 17 Appendix I 18 Appendix II 22 Appendix III 26 Appendix IV 27 Analst Contacts: New York Elena Duggar Analst Kenneth Emer Senior Vice President Daniel Gates Team Managing Director Hong Kong Clara Lau Group Credit Officer Asia Pacific Default and Recover Rates of Asia-Pacific Corporate Bond and Loan Issuers, Excluding Japan, 1990-H Summar This report is Mood s third stud of Asia-Pacific (ex-japan) corporate bond and loan issuers and their historical credit performance, migration, default, and recover experience. Broad conclusions include the following: The Asia-Pacific component of Mood s-rated corporate universe grew rapidl over the past 19 ears, from 68 rated bond and loan issuers at the end of 1990 to 356 issuers at the end of the first half of The speculative-grade share of these issuers also increased during this time period, from virtuall zero in 1990 to over 20% at the end of the first half of On average, default rates b rating categor in the Asia-Pacific region are broadl similar to their global counterparts. The rating distributions of these two groups differ significantl, however, with the Asia-Pacific region having a greater share of higher-rated corporate bond and loan issuers. As a result, when aggregating across all rated issuers, overall historical default rates are slightl lower in the Asia-Pacific region than globall. After several ears of no default experience, defaults picked up in 2008 and the first half of There were 11 rated and 3 unrated defaults during 2008-H1 2009, affecting a total of $3.1bn of debt. Historicall, the majorit of the region s defaults took place during the Asian financial crisis of Indonesia accounted for the largest number of defaults and Korea had the largest default volume. Sdne Brian Cahill Managing Director Asia Pacific

2 The Asia-Pacific region has experienced similar average recover rates on defaulted debt to the global averages. However, this result is based on a relativel small sample of available Asia-Pacific loan and bond recoveries. Rating accurac as measured b the cumulative accurac profile is modestl higher for credits in the Asia- Pacific region than for the global aggregates. Rating volatilit as measured b frequenc of all rating actions indicates that Asia-Pacific ratings have less stabilit than their global counterparts. Data and Methodolog This report is Mood s third detailed examination of the credit and default experience of Asia-Pacific (ex- Japan) corporate bond and loan issuers. 1 The stud examines the rating histories and default experience of 671 Asia-Pacific corporate issuers which have had Mood s-rated local and/or foreign currenc debt outstanding within the 1990-H period. 2 Although Mood s has rated corporate issuers in the Asia-Pacific region as earl as 1950, the report focuses on the modern era of corporate bond and loan issuance in this region, 1990 to present. 3 The basic unit of stud in this report is the corporate bond or loan issuer s individual rating histor. Mood s issuer ratings provide a rank-ordering of firms credit risks, which generall does not var with either the size or number of bonds that a firm ma have outstanding. Additionall, in most cases, when an issuer defaults on one of its bonds or loans, it defaults on all of them. Therefore, Mood s normall reports its default statistics on an issuer-weighted basis, i.e. the fraction of issuers that default, rather than on an issue- or volumeweighted basis. Furthermore, to the extent that historical default experience is predictive of future experience, issuer-weighted statistics are likel to provide the most reliable guide to future default experience. If instead we were to weight our statistics b the number of debt issues or their par amounts, we would place undue emphasis on the particular historical experiences of a few large or ver frequent issuers of corporate bonds. For the purpose of this stud, corporate bond ratings on either domestic- or foreign-currenc denominated debt are considered. We exclude an debt that was backed b a guarantor outside the corporate famil (e.g. a bond insurance compan). 4 We also exclude sovereign and sub-sovereign debt issuers. The data for this stud is from Mood s global default database which covers defaults b both Mood s rated and unrated issuers worldwide. Mood s has compiled this information using a variet of sources, including various print and online publishing sources, press releases, press clippings, internal memoranda, and records of analsts contact with rated issuers. Mood s definition of default includes three tpes of default events: A missed or delaed disbursement of interest or principal pament; Bankruptc filing or legal receivership b the issuer; A distressed exchange whereb: (i) an issuer offers creditors a new or restructured debt, or a new package of securities, cash or assets, that amount to a diminished financial obligation relative to the original obligation and (ii) the exchange has the effect of allowing the issuer to avoid a bankruptc or pament default Japan is excluded from the analsis because the rating transition and default experience of issuers in Japan differs markedl from that of issuers from other regions. The relativel low incidence of Mood s-rated defaults in Japan is quite striking. Onl two Japanese corporate issuers rated b Mood s Mcal Corporation and New Cit Residence Investment Corporation have defaulted on their bonds since This extraordinaril low incidence of default on rated bonds can be explained b two factors: (1) higher credit qualit since access to the bond market b speculative-grade companies is ver limited in Japan compared to other countries; and (2) sstemic support from, amongst others, bank lenders and the government for companies facing financial distress, which has averted man bond defaults. This stud focuses on issuers which have had Mood s-rated debt outstanding during the period of stud. In addition, Mood s also has issuer ratings or corporate famil ratings on issuers which either did not have Mood s-rated debt at an time during the 1990-H period or had onl national scale ratings. We begin our stud in 1990 due to the relativel small number of rated issuers before this date. However, prior to 1990, two significant defaults occurred in this region, both in New Zealand b issuers with initial ratings of Aa3 DFC Financial (Overseas) Ltd. and DFC Overseas Investment Ltd. When available, an issuer s senior unsecured rating is used as the indicator of the credit qualit. In the absence of such a rating, Mood s infers an equivalent senior unsecured rating from the issuer s other rated debt obligations. 2 August 2009 Special Comment -

3 This definition is intended to capture events that change the relationship between the debt-holder and debt issuer from the relationship which was originall contracted, and which subject the debt-holder to an economic loss. We seek to identif onl those economic losses that are the result of a credit event. Technical defaults (covenant violations, etc.) are not included in Mood s definition of default. 5 For the purposes of this stud, we measure recover rates on defaulted bonds and loans using market bid prices quoted one month after default. Price data are derived from various market sources, including Bloomberg, Reuters and Interactive Data Corporation. For the sake of consistenc, the volume figures for loan and bond issuers in different countries have been converted to US dollars from their respective local currencies at the prevailing exchange rates at the time of default. Issuer Characteristics Fueled b rapid regional economic growth during the 1980s and a decline in bank lending capacit in the 1990s, an increasing number of Asia-Pacific corporations issued debt in domestic and international bond markets and obtained Mood s credit ratings. At the end of 1990, there were 68 Mood s-rated loan and bond issuers in the Asia-Pacific region that excluded Japan. This number has increased dramaticall since then, reaching 356 issuers b the end of the first half of 2009 (see Exhibit 1). Despite the Asian economic crisis that hit the financial markets in , the growth of Mood s-rated corporate loan and bond issuance has continued, although at a slower pace in recent ears. As the rated debt market has grown during the past two decades, so has the share of the speculative-grade portion of the market. The vast majorit of the issuers in the Asia-Pacific region were rated investment grade in 1990, while b the end of the first half of 2009 the share of speculative-grade issuers was 21%. In 1998, at the height of the Asian financial crisis, the speculative-grade share of the market reached a peak of 42%. Man of those speculative-grade issuers defaulted during the crisis, which brought the share down to its present level, still much higher than the share at the beginning of the sample period. Exhibit 1: Growth of Mood s-rated Issuers in the Asia-Pacific Bond and Loan Market (excluding Japan) Number of Rated Issuers Year Speculative grade Investment grade Exhibit 2 presents information on the industr composition of Asia-Pacific issuers as of the first half of The majorit of issuers fall into three broad industr categories: industrial sector (43.3%), financial sector (42.1%) and public utilities and transportation companies (14.6%). 5 It is important to note that economic losses suffered b debt-holders due to changes in market conditions and/or market prices are not considered defaults as long as the terms of the obligation are being met. Additionall, missed paments which are cured within the contractuall-specified grace period are not considered a default. 3 August 2009 Special Comment -

4 Exhibit 2: Distribution of Asia-Pacific (ex-japan) Corporate Issuers b Broad Sector as of June 2009 Broad Industr Number of Issuers Share Banking % Finance % Industrial % Insurance 5 1.4% Other non-bank 8 2.2% Public utilit % Real estate finance 3 0.8% Securities 2 0.6% Transportation % Total % Exhibit 3 presents the domiciles of Mood s-rated Asia-Pacific corporate issuers. Approximatel one-third of Asia-Pacific domiciled issuers that currentl have rated debt outstanding are located in Australia. Other countries with a significant share include Hong Kong, Singapore and Korea. Indonesia previousl had a large presence in this market, but man of its issuers defaulted during the Asian financial crisis of Exhibit 3: Distribution of Asia-Pacific (ex-japan) Corporate Issuers b Countr of Domicile as of June 2009 Countr of Domicile Number of Issuers Share Australia % China % Hong Kong % India % Indonesia % Korea % Malasia % Mongolia 2 0.6% New Zealand % Pakistan 1 0.3% Philippines 6 1.7% Singapore % Taiwan 3 0.8% Thailand 7 2.0% Total % Exhibit 4 shows the rating distribution of Asia-Pacific issuers and changes in the rating distribution over time. In 1990, 98.5% of the issuers were rated investment grade. However, the investment-grade share of issuers has decreased to 78.9% at June 30, And, within the investment-grade and speculative-grade sectors, the average rating has drifted downward. Much of the decline in average credit qualit can be explained b the increase in the number of non-financial corporates relative to financial corporates in the overall rated population and the increased access of lower-rated issuers to the international bond and loan markets. 4 August 2009 Special Comment -

5 Exhibit 4: Distribution of Period-End Asia-Pacific (ex-japan) Corporate Issuers b Whole Letter Rating (number of issuers and percent of total) Year H Aaa Aa A Baa Ba B Caa-C Investment grade Speculative grade All Year H Aaa 16.2% 5.0% 5.0% 3.8% 2.8% Aa 55.9% 21.6% 21.6% 16.7% 18.0% A 19.1% 43.9% 43.9% 38.1% 34.6% Baa 7.4% 20.9% 20.9% 25.8% 23.6% Ba 0.0% 6.5% 6.5% 8.2% 7.3% B 1.5% 2.2% 2.2% 6.6% 7.9% Caa-C 0.0% 0.0% 0.0% 0.9% 5.9% Investment grade 98.5% 91.4% 91.4% 84.3% 78.9% Speculative grade 1.5% 8.6% 8.6% 15.7% 21.1% The rating distributions of Asia-Pacific (ex-japan) and global corporate issuers as of June 2009 are compared in Exhibit 5. We see that compared to the global corporate sample the share of Asia-Pacific issuers is higher in the investment grade categories, especiall the Aa and A rating categories, and lower in speculative grade categories. As of June 2009, the median rating in the Asia-Pacific region was A compared to Baa for the global sample. Exhibit 5: Rating Distribution of Asia-Pacific (ex-japan) and Global Corporate Issuers, June % Percentage of Rated Issuers 35% 30% 25% 20% 15% 10% 5% 0% Aaa Aa A Baa Ba B Caa-C Rating Categor Asia-Pacific Global 5 August 2009 Special Comment -

6 Trends in Credit Qualit: Transition Matrices and Rating Changes Exhibit 6 displas the historical average frequenc of alpha-numeric rating changes for Asia-Pacific (ex-japan) and the global universe of rated corporate issuers over the period 1990-H For example, a categor of 0 indicates no rating change over the twelve-month period. A categor of -1 indicates a single-notch alphanumeric rating downgrade, while +2 indicates a two-notch alpha-numeric rating upgrade. The vertical axis indicates the percentage of issuers in each categor. Exhibit 6: Annual Frequenc of Alpha-Numeric Rating Changes, 1990-H % 80% 73.1% 76.4% 60% 40% 20% 0% 8.8% 3.1% 6.8% 7.7% 7.0% 3.8% 0.9% 2.3% 3.4% 3.4% 2.2% 0.9% -3 and below and above Asia-Pacific Global Mood s ratings are generall ver stable over a one-ear horizon. Despite the Asian financial crisis, ratings stabilit exhibited b Asia-Pacific issuers is similar to global issuers the Asia-Pacific region has issuers that are almost as stable as their global counterparts (73.1% vs. 76.4%). On average, Asia-Pacific issuers experienced a 15.6% probabilit (6.8% downgrade + 8.8% upgrade) of a single alpha-numeric rating change over a one-ear horizon. Changes in excess of a single alpha-numeric rating change, whether upgrades or downgrades, have been extremel infrequent over a one-ear horizon. Rating migration matrices present a more complete picture of changes in rating qualit over time. Exhibit 7 shows average annual, whole-letter rating migration rates since Each cell in the matrix shows the weighted average fraction of issuers who held a given row s rating at the beginning of the measurement period and the column rating it held at the end of the period, including defaults and withdrawn ratings (WR). 6 The weights correspond to the size (number of issuers) of the annual cohorts. Global statistics are presented alongside for comparison. The largest values in the transition matrix are along the diagonal, reflecting the fact that the most likel rating for an issuer at the end of a given ear is the rating with which it began the ear. For example, an issuer domiciled in the Asia-Pacific region who was rated Aaa at the beginning of the time period has a 79.5% historical probabilit of remaining Aaa one ear later. 6 Ratings are tpicall withdrawn when all of an issuer s rated debt matures, is called or converted, or is retired through some other orderl market function (e.g. M&A). 6 August 2009 Special Comment -

7 Exhibit 7: Average One-Year Rating Migration Rates for Asia-Pacific (ex- Japan) and Globall, 1990-H Asia-Pacific (ex-japan) Rating to: Rating from: Aaa Aa A Baa Ba B Caa-C D WR Aaa 79.50% 15.43% 0.05% 0.00% 0.00% 0.00% 0.00% 0.00% 5.01% Aa 1.61% 85.83% 5.90% 0.07% 0.02% 0.00% 0.00% 0.00% 6.57% A 0.00% 2.82% 85.41% 4.43% 1.82% 0.11% 0.00% 0.00% 5.42% Baa 0.00% 0.09% 8.54% 77.35% 4.51% 1.14% 0.42% 0.23% 7.72% Ba 0.00% 0.00% 0.04% 10.79% 70.18% 6.16% 3.06% 1.05% 8.72% B 0.00% 0.00% 0.00% 0.00% 13.88% 62.73% 9.34% 6.76% 7.41% Caa-C 0.00% 0.00% 0.00% 0.09% 0.35% 9.98% 62.35% 18.13% 9.11% Global Rating to: Rating from: Aaa Aa A Baa Ba B Caa-C D WR Aaa 86.56% 8.99% 0.34% 0.02% 0.06% 0.01% 0.00% 0.00% 4.02% Aa 0.92% 85.88% 7.45% 0.28% 0.04% 0.01% 0.02% 0.02% 5.39% A 0.07% 2.84% 86.53% 5.09% 0.38% 0.08% 0.03% 0.05% 4.94% Baa 0.04% 0.20% 4.64% 84.29% 3.91% 0.81% 0.30% 0.18% 5.63% Ba 0.00% 0.06% 0.40% 6.21% 73.25% 8.01% 0.88% 0.83% 10.36% B 0.01% 0.04% 0.14% 0.35% 4.99% 72.85% 6.87% 3.69% 11.04% Caa-C 0.00% 0.02% 0.05% 0.20% 0.46% 8.09% 62.96% 15.00% 13.23% B 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 ear. For example, an Aarated issuer in the Asia-Pacific region had a historical probabilit of 5.9% of being downgraded to an A rating within one ear and a 1.6% historical probabilit of being upgraded to Aaa over the same period. The farther one moves awa from the diagonal, the smaller the migration rates, reflecting a relativel low historical probabilit of issuers moving across more than one rating categor during the course of a ear. As shown in Exhibit 7, issuers in the Asia-Pacific region are slightl less stable than their global counterparts. For example, on average, 15.5% of Aaa-rated issuers in the Asia-Pacific region have been downgraded within one ear, as opposed to 9.4% globall. Similar trends persist in other rating categories, however the slightl lower stabilit of Asia-Pacific issuers is primaril due to a higher probabilit of upgrade. For instance, the percentage of upgrades from the B-rating categor to the Ba rating categor was 13.9% in the Asia-Pacific region, about three times larger than the global historical average. Rating drift, the net percentage of upgrades relative to downgrades, provides a useful indicator of changes in aggregate credit qualit. Positive drift indicates overall improvement in credit qualit, while negative drift signals overall deterioration in credit qualit. Alternativel, no drift (when upgrades and downgrades are roughl equal for a given time period) indicates a stable credit environment. Exhibit 8 compares the annual rating drift rates (upgrade rate minus downgrade rate divided b total rating count outstanding) for the Asia-Pacific issuers excluding Japan with the ratings drift of issuers globall. On average, upgrades and downgrades in Asia-Pacific have more or less offset each other across the sample. As a result, the average rating drift is slightl negative at -4.1% relative to -3.3% globall. 7 August 2009 Special Comment -

8 Exhibit 8: Annual Rating Drift for Asia-Pacific (ex-japan) and Globall, H (upgrades minus downgrades divided b total rating count outstanding) 50% 40% 30% 20% Rating Drift 10% 0% -10% -20% -30% -40% -50% Year Asia-Pacific Global As is shown in the exhibit, credit qualit trends for Asia-Pacific corporate issuers have been more variable than for global issuers and moderatel correlated with global credit trends. The ratings drift in Asia-Pacific is a result of several factors, including the Asian financial crisis of (negative drift) and recent efforts b companies to restructure their debt and repair balance sheets during the past few ears of buoant economic growth (positive drift). Since the second half of 2008, the credit ccles of the two regions have converged on a downward path, consistent with our expectations for a continued gradual softening in credit qualit and a growing number of defaults globall. Historical Corporate Defaults In the period between 1990 and June 2009, 216 issuers of corporate bonds and loans in the Asia-Pacific region defaulted on a total of $50.1 billion of bonds and loans. Of these, 49 issuers were rated b Mood s with a total default volume of $15.6 billion. Exhibit 9 presents a chronological summar of the number of defaults and the volume of defaulted debt for Asia-Pacific issuers, both rated and unrated. In addition, Appendix II lists the defaulted issuers chronologicall. 7 Not surprisingl, given the Asian financial crisis of , the ears are the ears with the largest number of defaults. Defaults have decreased dramaticall since In fact, there was no rated default in the Asia-Pacific region from 2003 until 2008, although there have been unrated defaults throughout this time period. The number of defaults picked up slightl in 2008 and the first half of 2009, with a total of 11 rated and 3 unrated issuers defaulting on $3.1bn of debt. In fact the default volume during the first half of 2009 alread represents the highest default volume observed since the end of the Asian crisis in Our default studies focus on Mood s-rated defaults. We present available information on unrated defaults as well in this section, but we caution that coverage of unrated defaults is incomplete and likel inconsistent over time. 8 August 2009 Special Comment -

9 Exhibit 9: Asia-Pacific (ex-japan) Corporate Default Counts and Volumes (Millions of $), Bonds and Loans, as of June 2009 Year Rated Count Unrated Count Total Count Rated Volume Unrated Volume Total Volume $200 $345 $ $0 $0 $ $0 $0 $ $0 $0 $ $0 $0 $ $0 $0 $ $0 $201 $ $614 $2,549 $3, $1,444 $9,825 $11, $2,109 $8,849 $10, $1,048 $7,977 $9, $4,749 $1,935 $6, $2,010 $214 $2, $374 $377 $ $0 $105 $ $0 $523 $ $0 $1,783 $1, $0 $82 $ $288 $61 $349 H $2,791 $0 $2,791 Total $15,627 $34,825 $50,452 Exhibit 10 presents the geographical distribution of defaulters for the period under analsis. During the period from 1990 H1 2009, Indonesia had the largest number of rated and unrated defaults, where a total of 62 issuers (17 rated and 45 unrated) defaulted on more than $10 billion worth of bonds and loans. Korea, on the other hand, had the largest default volume during the period, where a total of 43 issuers defaulted on more than $18 billion of debt. Most of these defaults occurred during the Asian financial crisis. Exhibit 10: Geographical Distribution of Asia-Pacific (ex-japan) Corporate Defaulters, Counts and Volumes (Millions of $), 1990-H Countr Rated Count Unrated Count Total Count Rated Volume Unrated Volume Total Volume Australia $2,795 $726 $3,521 China $2,203 $524 $2,728 Hong Kong $2,408 $2,536 $4,945 India $0 $510 $510 Indonesia $5,269 $5,303 $10,572 Korea $2,506 $16,158 $18,664 Malasia $0 $2,357 $2,357 Philippines $258 $414 $672 Singapore $51 $414 $465 Taiwan $0 $421 $421 Thailand $135 $5,462 $5,597 Total $15,627 $34,825 $50,452 9 August 2009 Special Comment -

10 Exhibit 11 presents the broad industr distribution of defaults during the period of our analsis. Most of the defaults 77% of default counts and 78% of the default volume - occurred in the broad categor of industrial companies. This is followed b bank and other financial companies, which together account for 18% of the default count and 19% of the default volume. Finall, utilit and transportation companies together represent 5% of the count and 3% of the default volume. Exhibit 11: Distribution of Asia-Pacific (ex-japan) Corporate Defaulters b Broad Sector, Counts and Volumes (Millions of $), 1990-H Year Rated Count Unrated Count Total Count Rated Volume Unrated Volume Total Volume Banking $519 $1,548 $2,068 Finance $0 $1,397 $1,397 Industrial $14,394 $25,090 $39,484 Other Non-Bank $0 $6,073 $6,073 Public Utilit $154 $0 $154 Real Estate Finance $0 $100 $100 Transportation $559 $618 $1,176 Total $15,627 $34,825 $50,452 Cumulative Default Rates Exhibit 12 presents average annual speculative-grade default rates for Mood s-rated Asia-Pacific (ex-japan) issuers. We clarif here that Exhibit 9 (Chronological Summar of Defaults) lists both rated and unrated default counts, while Exhibit 12 plots the default rate for Mood s-rated issuers onl. Exhibit 12: Global and Asia-Pacific (ex-japan) Rated Issuer-Weighted Speculative-Grade Default Rates, 1996 H % 20% 15% 10% 5% 0% Jan-96 Jul-96 Jan-97 Jul-97 Jan-98 Jul-98 Jan-99 Jul-99 Jan-00 Jul-00 Jan-01 Jul-01 Jan-02 Jul-02 Jan-03 Jul-03 Jan-04 Jul-04 Jan-05 Jul-05 Jan-06 Jul-06 Jan-07 Jul-07 Jan-08 Jul-08 Jan-09 Jul-09 Global Speculative-Grade Default Rate Asia-Pacific Speculative-Grade Default Rate Between 1990 and 1996, there were no speculative-grade defaults in the Asia-Pacific region, and therefore the speculative-grade default rate is zero. The speculative-grade default rate peaked twice during the Asian financial crisis once in 1998 and again in There were no rated defaults in the four ears preceding the 2008 global financial crisis, and therefore the speculative-grade default rate for issuers domiciled in the Asia- Pacific region is zero. With the pick up in defaults over 2008 and 2009, we see the default rate for Asia-Pacific 10 August 2009 Special Comment -

11 issuers rising along with the global default rate. In June 2009, the issuer-weighted speculative-grade default rate for Asia-Pacific (ex-japan) issuers stood at 10.5%, similar to the global speculative-grade default rate of 10.3%. Investors are also interested in default rates for investment horizons longer than one ear. Exhibit 13 presents one-ear through ten-ear issuer-weighted average cumulative default rates for Asia-Pacific (ex-japan) issuers. Also presented are the global statistics for comparison. As in our other default studies, historical cumulative default rates are calculated b averaging the experiences of issuer cohorts formed at monthl frequencies. 8 B forming and tracking such cohorts of all Mood s-rated issuers at the beginning of ever month, we replicate the experience of a portfolio of both seasoned and new-issue bonds purchased in an given month. The dnamic nature of the cohorts allows the estimation of cumulative default risk over multi-ear horizons. It also allows for the comparison and averaging of default rates over different periods. Appendix I at the end of the report described the methodolog for estimating cumulative default rates. Exhibit 13: Asia-Pacific (ex-japan) and Global Issuer-Weighted Cumulative Default Rates, 1990-H Asia-Pacific (ex-japan) Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 Year 9 Year 10 Aaa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Aa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% A 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Baa 0.237% 0.600% 1.094% 1.869% 2.627% 3.175% 3.393% 3.393% 3.393% 3.393% Ba 1.099% 3.029% 5.297% 8.170% % % % % % % B 7.016% % % % % % % % % % Caa-C % % % % % % % % % % Investment Grade 0.067% 0.169% 0.304% 0.513% 0.711% 0.844% 0.892% 0.892% 0.892% 0.892% Speculative Grade 5.041% % % % % % % % % % All Asia-Pacific 1.062% 2.124% 3.226% 4.137% 4.775% 5.191% 5.420% 5.420% 5.420% 5.420% Global Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 Year 9 Year 10 Aaa 0.000% 0.022% 0.022% 0.022% 0.022% 0.022% 0.022% 0.022% 0.022% 0.022% Aa 0.018% 0.044% 0.056% 0.079% 0.109% 0.130% 0.150% 0.160% 0.162% 0.195% A 0.047% 0.136% 0.255% 0.393% 0.579% 0.781% 1.002% 1.272% 1.553% 1.821% Baa 0.185% 0.489% 0.854% 1.267% 1.717% 2.188% 2.655% 3.148% 3.688% 4.328% Ba 0.873% 2.359% 4.178% 6.187% 7.849% 9.601% % % % % B 3.907% 9.346% % % % % % % % % Caa-C % % % % % % % % % % Investment Grade 0.083% 0.225% 0.394% 0.587% 0.810% 1.041% 1.275% 1.532% 1.801% 2.094% Speculative Grade 4.418% 9.110% % % % % % % % % All Global 1.617% 3.287% 4.817% 6.103% 7.184% 8.141% 8.993% 9.752% % % As Exhibit 13 illustrates, Mood s ratings powerfull ran order default risk at all horizons both in Asia-Pacific and globall, as the probabilit of default rises with lower ratings. A comparison between Asia-Pacific and global default rates show that, on average, default rates are broadl similar b rating categor. Indeed, for shorter time horizons, Asia-Pacific default rates in the Baa-C categories 8 Monthl cohorts have the advantage of capturing rating changes that occur within a calendar ear. The default rates are calculated based on cohorts of all issuers holding a given estimated senior unsecured rating at the start of a given month. The cohorts are dnamic in that the change based on whether these issuers leave the cohort due to default or non credit-related reasons (e.g. maturing of debt). While the cohort frequenc is monthl, the accumulation periodicit remains 12 months, so that we track default rates over horizons of one ear, two ears, three ears, etc. 11 August 2009 Special Comment -

12 are slightl higher than for the global sample, largel reflecting the impact of the defaults in the pulp and paper industries during However the differences narrow down and reverse as the time horizon lengthens. The clustering of corporate defaults in the Asia-Pacific region around the ears of the Asian financial crisis also explains the slightl higher shorter-horizon Caa-C categor default rate. Even though the investment grade and speculative grade default rates are broadl similar between the Asia- Pacific and the global sample, the overall default rate is lower in the Asia-Pacific region. This is due to the difference in the rating distributions of the two samples of issuers. As previousl illustrated in Exhibit 5, the Asia-Pacific region has a greater share of higher-rated corporate bond and loan issuers compared to the global sample. As a result, when aggregating across all rated issuers, historical default rates are actuall lower in the Asia-Pacific region than globall. Regional Differences in Rating Transitions and Cumulative Default Rates Until now, we have discussed the Asia-Pacific region as a single portfolio of issuers operating in similar capital markets. However, issuers in different countries in the region operate under different bankruptc and regulator regimes, commercial lending environments, and macro-economies. As such, investors ma be interested in defaults and rating transitions at a more specific level. In this section, we create sub-samples based on a more natural pooling of issuers to account for differences in capital markets. Exhibit 14: Average One-Year Rating Transitions within the Asia-Pacific Region, 1990-H Australia and New Zealand Rating to: Rating from: Aaa Aa A Baa Ba B Caa-C D WR Aaa 79.71% 15.61% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 4.67% Aa 1.75% 86.83% 4.19% 0.09% 0.03% 0.00% 0.00% 0.00% 7.12% A 0.00% 2.89% 86.55% 5.03% 0.32% 0.10% 0.00% 0.00% 5.11% Baa 0.00% 0.19% 7.33% 82.69% 2.56% 0.48% 0.00% 0.00% 6.75% Ba 0.00% 0.12% 0.00% 6.67% 65.05% 11.53% 4.85% 1.33% 10.44% B 0.00% 0.00% 0.00% 0.00% 4.37% 61.93% 15.85% 6.01% 11.84% Caa-C 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 16.05% 80.25% 3.70% Asia-Pacific excluding Japan, Australia and New Zealand Rating to: Rating from: Aaa Aa A Baa Ba B Caa-C D WR Aaa 77.87% 14.04% 0.43% 0.00% 0.00% 0.00% 0.00% 0.00% 7.66% Aa 0.97% 81.19% 13.83% 0.00% 0.00% 0.00% 0.00% 0.00% 4.00% A 0.00% 2.75% 84.34% 3.87% 3.23% 0.12% 0.00% 0.00% 5.70% Baa 0.00% 0.02% 9.34% 73.86% 5.79% 1.57% 0.69% 0.38% 8.36% Ba 0.00% 0.00% 0.02% 11.52% 71.10% 5.20% 2.74% 1.00% 8.41% B 0.00% 0.00% 0.00% 0.00% 16.11% 62.93% 7.71% 6.94% 6.31% Caa-C 0.00% 0.00% 0.00% 0.09% 0.38% 10.74% 65.88% 13.38% 9.52% Within the Asia-Pacific region, we focus on two distinct sub-regions: 1. Australia and New Zealand 12 August 2009 Special Comment -

13 2. Rest of Asia Pacific. 9 Exhibit 14 shows the different rating transitions for these sub-regions. Investment-grade issuers in Australia and New Zealand are slightl more stable than those in the rest of the Asia-Pacific region in the investmentgrade rating categories. However, speculative-grade issuers in Australia and New Zealand are slightl less stable. There are ver few Caa-C rated issuers in Australia and New Zealand, and those specific issuers had a ver high one-ear default rate. Exhibit 15 compares the cumulative default rates between Australia and New Zealand and the rest of the Asia- Pacific region. These default rates show that credit qualit is ver strong for Australia-New Zealand investment-grade issuers there are no defaults within the 10-ear horizon. However, the riskiest of the speculative-grade issuers, in the Caa-C rating categor, are risk indeed there has been a 100% default rate b ear 2. For B-rated issuers, there is a 100% default rate b ear 6. However, the extremel small number of speculative-grade issuers in Australia and New Zealand is driving these anomalous results. In an given cohort, there were no more than a few issuers both in the Caa-C and in the B rating categories. The rest of the Asia-Pacific region has cumulative default rates that are fairl similar to the complete Asia-Pacific region (see Exhibit 13). Exhibit 15: Issuer-Weighted Cumulative Default Rates b Region, 1990-H Australia and New Zealand Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 Year 9 Year 10 Aaa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Aa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% A 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Baa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Ba 1.408% 3.522% 8.176% % % % % % % % B 6.389% % % % % % % % % % Caa-C % % % % % % % % % % Investment Grade 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Speculative Grade 7.916% % % % % % % % % % All Australia and New Zealand 0.503% 0.976% 1.458% 1.833% 2.236% 2.359% 2.359% 2.359% 2.359% 2.359% Rest of Asia-Pacific (excl. Japan, Australia, and New Zealand) Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 Year 9 Year 10 Aaa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Aa 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% A 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% Baa 0.393% 0.995% 1.812% 3.094% 4.335% 5.207% 5.547% 5.547% 5.547% 5.547% Ba 1.044% 2.942% 4.825% 7.484% 9.132% % % % % % B 7.168% % % % % % % % % % Caa-C % % % % % % % % % % Investment Grade 0.155% 0.393% 0.721% 1.243% 1.755% 2.111% 2.243% 2.243% 2.243% 2.243% Speculative Grade 4.516% 9.065% % % % % % % % % All Rest of Asia- Pacific 1.587% 3.220% 4.954% 6.444% 7.351% 8.116% 8.626% 8.626% 8.626% 8.626% 9 As mentioned in the introduction, we have eliminated Japan from our analsis because of its man differences from the rest of the region. For interested readers, we provide cumulative default rates and rating transitions for Japan, as well as for the Asia-Pacific region including Japan in Appendix IV. 13 August 2009 Special Comment -

14 Rating Performance Measures The default rates presented above demonstrate that Mood s Asia-Pacific corporate ratings are correlated with subsequent default experience at long investment horizons. One of the desirable properties of an efficient rating sstem is its abilit to separate the low credit risks from the high credit risks. A ke metric designed to measure the relative accurac of a rating sstem is the cumulative accurac profile (CAP). The CAP curve is constructed b plotting, for each rating categor, the proportion of defaults accounted for b issuers with the same or lower rating against the proportion of all issuers with the same or lower rating. Exhibit 16 presents the one-ear-ahead horizon CAP curves for Asia-Pacific (ex-japan) and for global ratings observed between 1990 and H The CAP curve is useful for making a visual assessment of the information content embedded in the relative ranking of credit risk provided b a set of ratings. A rating sstem that possessed no predictive information about default risk would lie on the 45-degree line. The further the CAP curve bows towards the top left corner, the better it is at identifing defaults in the riskiest rating categories. The CAP plots reveal that historicall both global and Asia-Pacific ratings have done a good job rank-ordering one-ear default risk. Across regions, issuers historicall have carried low ratings one ear prior to default. For example, 95% of defaulters in Asia-Pacific carried a rating of Ba1 or lower 12 months before default, while Ba1 rated issuers or below accounted for approximatel 97% of defaulting issuers globall. Exhibit 16: One-Year Cumulative Accurac Profiles, 1990-H % 90% Cumulative Proportion of Defaulters 80% 70% 60% 50% 40% 30% 20% 10% 45-degree line Asia-Pacific Global 0% 0% 20% 40% 60% 80% 100% Cumulative Proportion of Issuers A summar measure of rating accurac that compresses the information depicted in the CAP curve into a single summar statistics is the accurac ratio (AR). The AR is the ratio of the area between the CAP curve and the 45-degree line to the total area above the 45-degree line. The AR lies between minus one and plus one (or -100% and +100%), similar to a correlation statistics. As can be inferred b the CAP curves in Exhibit 16, Mood s Asia-Pacific ratings have had modestl higher accurac ratios than their global counterparts. The historical average one-ear accurac ratio for Asia-Pacific ratings is 95.2% for the 1990-H period, compared to 90.6% for global ratings during the same period. 14 August 2009 Special Comment -

15 Recover Rates in Default Mood s ratings are statements about expected credit losses, which equals the probabilit of default times the expected loss severit rate (i.e., one minus the expected recover rate) given default. Since the probabilit of default is usuall the same across the various obligations of an issuer, differences in ratings across an issuer s capital structure generall reflect different expectations of recover rates in the event of default. Commensurate with our practice in our global corporate default stud, in Exhibit 17, we estimate recover rates using 30-da post-default bid prices on defaulted debt, although no trades ma have taken place at these prices. Appendix III provides a list of recover prices for defaulted debt in the Asia-Pacific region. Exhibit 17 presents the average recover rates for the Asia-Pacific region and for the global sample, with one caveat that the number of observations in Asia-Pacific is quite small compared to the global universe. We see that issuer-weighted recover rates at the senior unsecured level, where data is more abundant, compare closel with their global counterparts. The monotonic relationship between instrument seniorit and recover rate is clearl visible in the global sample. The higher recover rate shown for Subordinated instruments in the Asia-Pacific region is based on a ver small sample size. Exhibit 17: Historical Recover Rates for Global and Asia-Pacific (ex-japan) Issues b Lien Position (rated and unrated issuers), 1990-H Global Instrument Lien position Issuer-weighted recover rate Number of issuers Loan Sr. Secured Bond Sr. Secured Bond Sr. Unsecured Bond Subordinated Asia-Pacific (ex- Japan) Instrument Lien position Issuer-weighted recover rate Number of issuers Loan Sr. Secured Bond Sr. Secured Bond Sr. Unsecured Bond Subordinated Credit Loss Rates In the previous sections we examined the two main components of credit loss: the probabilit of default and the severit of loss given default (one minus recover rate). In this section, we bring these two components together to discuss credit losses and demonstrate that Mood s ratings effectivel differentiate credit loss rates. Mood s ratings process is designed to produce a consistent measure of relative credit risk, which in large part is determined b credit losses. Credit losses are defined mathematicall as follows: Credit Loss Rate= (Default Frequenc)*(1-Recover Rate) Credit losses can be defined as the loss incurred in total return b a bu-and-hold bond portfolio due to default. Exhibit 18 presents average annual credit losses for portfolios based on Mood s broad letter rating categories. 15 August 2009 Special Comment -

16 Exhibit 18: One-Year Credit Losses for Asia-Pacific (ex-japan) and Global Issuers, 1990-H % 12% 10% 8% 6% 4% 2% 0% Baa Ba B Caa-C Speculative- Grade Global Asia-Pacific Average speculative-grade one-ear credit loss rates across the two regions are ver similar. In the lower part of the speculative-grade range, issuers in the Asia-Pacific region excluding Japan have slightl higher credit loss rates than the complete global sample. At the aggregate speculative-grade level, Asia-Pacific loss rates are ver similar to their global counterpart, reflecting the similarities in both aggregate default and recover rates. Importantl, these results signal that Mood s ratings in Asia-Pacific are consistent with Mood s ratings globall as indicators of credit loss rates. The highest-risk portfolios (with an average rating of Caa-C) have the highest credit losses and the lowest-risk portfolios (rated Baa) have the lowest credit losses. The amount of credit loss increases exponentiall as we go down the rating scale. 16 August 2009 Special Comment -

17 Mood s Related Research Special Comments: Corporate Default and Recover Rates, , Februar 2009 (114844) European Corporate Default and Recover Rates, , Ma 2009 (115292) Latin American Corporate Default and Recover Rates, 1990-H1 2009, Jul 2009 (118744) Emerging Market Corporate and Sub-Sovereign Defaults and Sovereign Crises: Perspectives on Countr Risk, Februar 2009 (113931) Sovereign Default and Recover Rates, , March 2009 (115232) Sovereign Defaults and Interference: Perspectives on Government Risks, August 2008 (110114) Strong Loan Issuance in Recent Years Signals Low Recover Prospects for Loans and Bonds of Defaulted U.S. Corporate Issuers, June 2008 (109457) Guide to Mood s Default Research: June 2009 Update, June 2009 (118044) Measuring Corporate Default Rates, November 2006 (100779) Mood s Senior Ratings Algorithm and Estimated Senior Ratings, Februar 2009 (114614) Rating Methodolog: Mood s Approach to Evaluating Distressed Exchanges, March 2009 (115337) A Guide to Mood's Sovereign Ratings, December 2008 (98177) Sovereign Bond Ratings, September 2008 (109490) To access an of these reports, click on the entr above. Note that these references are current as of the date of publication of this report and that more recent reports ma be available. All research ma not be available to all clients. 17 August 2009 Special Comment -

18 Appendix I: Methodolog and Data Sources Methodolog Changes in This Year s Report The historical data contained in Mood s 2008 global default stud includes a number of revisions relative to previous ears publications. Mood s occasionall discovers historical defaults, leading to minor revisions of the historical data. As alwas, the data contained in the most recentl published Mood s default stud supersedes the data published in previous reports. Change in Senior Rating Algorithm In Januar 2009, we applied a smoothing procedure in the senior rating algorithm to estimate ever issuer s senior unsecured ratings. This enhancement aims at ironing out artificial rating changes that are not resulting from the change of an issuer s credit qualit, but due to the switch of reference debt based on which the estimated senior unsecured rating is derived from. The implementation of smoothing procedure is particularl important to structural breaks when Mood s introduced new debt ratings, refined its rating scale and adjusted its notching practices. For example, the introduction of issuer ratings on Oct 25th 1999 will result in numerous artificial rating changes if the smoothing procedure is not in place. For other times, the impact of smoothing is modest. For instance, onl 3% of Mood s-rated issuers observe adjustments in their estimated senior unsecured ratings (mostl b one notch) on Januar 1st, Removal of Special Purpose Vehicles In Januar 2009, we removed substantial number of special purpose vehicles from Hartford Financial Services Group, Inc., Principal Financial Group, Inc., and Protective Life Corporation. These SPVs are non-insurance conduits and mostl Aa3 or A1 credits. None of them have defaulted. Rating Migration Calculations A rating migration matrix completel summarizes changes in credit ratings over a given time horizon. The cells of the matrix are discrete-time estimates of rating migration probabilities. The show rate of rating change measured at two points in time; e.g. the start and end of one ear. When calculating credit rating migration rates, Mood's treats rating changes, rating withdrawals, and defaults as mutuall exclusive states. For example, an issuer that is downgraded on the da of default is counted onl as a default, not a downgrade or a downgrade and default. Similarl, an issuer whose rating is withdrawn on the da of default is counted as a default, not a withdrawal. The probabilit that an issuer s rating i held on cohort date will transition to rating j (which includes default and withdrawal) over a time horizon T is calculated as: nij ( T ) p ij ( T ) = n (0) The weighted average rating migration rate for all cohorts in the historical data set Y is calculated as: p ( T ) = ij Y Y n n ij i ( T ) (0) i 18 August 2009 Special Comment -

19 Default Rate Calculations Mood s method for calculating cumulative default rates is a discrete-time approximation of the nonparametric continuous-time hazard rate approach. 10 Cumulative default rates are calculated from marginal default rates, which represent the probabilit that an issuer that has survived in the sample through a particular date will default over the next time interval (tpicall, one ear) following that date. The marginal default rate for issuers holding rating i on cohort date is the ratio of the number of defaulting issuers in period t divided b the number of issuers exposed to the risk of default in period t: d i ( t) = n i ( t) x i xi ( t 1) ( t) [ w ( t) + w ( t 1 2 i i The denominator of the marginal default rate adjusts for defaults that occur prior to time interval t, as well as rating withdrawals, w(t), that occur in periods prior to interval t and a small adjustment for withdrawals that occur in time interval t. Cumulative default rates for investment horizons of length T, D(T), are built up from the marginal default rates: D i ( T) = 1 T t= 1 [1 d Or, expanding the above equation (and dropping indices for brevit): i ( t)] D ( T ) = d(1) + d(2)[1 d(1)] + d(3)[(1 d(1))(1 d(2))] d( T)( 1)] T 1 t=1 [1 d( t)]) In the first time period, a fraction of the credit exposures in the cohort either defaults or survives. The credit exposures that survive period one ma then go on to default or survive in period two; those that survive period two ma go on to default or survive in period three, etc. Because the time periods are non-overlapping and the probabilit of default in each period is assumed to be independent, the T-period cumulative default rate is defined as one minus the product of the T marginal survival rates. The calculation of the average cumulative default rate for rating class i, D i (T ), is derived from the weighted average marginal default rates, d i (t), calculated from all the available cohort marginal default rates in the historical data set Y: D ( T ) = 1 i T t= 1 [1 d i ( t)] where d ( t) = i Y Y x n i i ( t) ( t) 10 See Mood s Global Credit Research Measuring Corporate Default Rates, November August 2009 Special Comment -

20 The default rates calculated in multi-ear rating migration matrices are not comparable to those calculated using Mood s discrete-time hazard rate method described above. Rating migration matrices account for rating withdrawals separatel (in the column labeled WR) while the hazard rate method incrementall adjusts the denominator of the marginal default rate to remove rating withdrawals. Occasionall, withdrawal-adjusted rating migration matrices are calculated as follows: p ij ( T ) * pij ( T ) = (1 p ( T )) Using this method, all issuers whose ratings are withdrawn are removed. In effect, data for issuers whose ratings are withdrawn is completel discarded. This method generall ields higher default rate estimates than the hazard rate method. Hence, Mood s hazard rate-derived default rate estimates la between unadjusted migration matrix-derived default rates and withdrawal-adjusted migration matrix-derived default rates. Recover Rate Calculations Mood's estimates defaulted debt recover rates using market bid prices observed roughl 30 das after the date of default. Recover rates are measured as the ratio of price to par value. Using post-default prices to measure recover (or its complement, loss severit) parallels common practice in the credit default swaps market. Moreover, recover rates measured in this wa are most relevant for cash bond investors who liquidate their holdings shortl after default as often required b their portfolio governance rules or their own investment objectives. For investors holding defaulted securities until ultimate resolution, prices observed shortl after default are generall accepted as the market's estimate of discounted expected ultimate recover rates. The alternative approach of directl measuring ultimate realized recoveries presents a number of estimation challenges, including the appropriate discount rate to appl to cash-flows and valuing the assets used to pa creditor claims, which ma include illiquid new securities (e.g., equit and derivative instruments) as well as phsical assets. The table below defines the various was Mood's calculates recover rates. 11 For the purposes of measuring expected credit loss rates, we rel on issuer-weighted mean recover rates. iw Statistic Issuer-Weighted Mean Recover Rates Value-Weighted Mean Recover Rates Issuer-Weighted Median Recover Rates Issue-Weighted Mean Recover Rates Definition The are derived b estimating mean recover rates for each issuer, then averaging them across issuers. The are useful for predicting recover rates for portfolios that are well diversified across issuers. The represent the average of recover rates on all defaulted issuers, weighted b the face value of those issues. These estimates are useful for predicting recover rates on the market portfolio. The are estimated as median of issuer-weighted recover rates and are used for predicting the most likel recover rate for a randoml selected issuer. The are estimated using recover rates for each issue and taking the average of all issues. While this measure is widel reported, it is useful onl for predicting the average recover rate on a portfolio of default bonds diversified across issues but without reference to issuer or issue size. 11 This table is derived from Mood s Special Comment Recover Rates on Defaulted Bonds and Preferred Stocks, , December August 2009 Special Comment -