In 2020, TRIS Rating rated and publicly announced the ratings of 224 issuers. The issuers can be categorized as 166 non-financial institution issuers (non-FI), 51 financial institutions (FI), and three structured finance issuers. The four remaining issuers were in the ?Government? sector, including one sovereign, one local government, and two supranational organizations.
However, this study does not cover structured finance issuers and issuers in the Government sector. We also exclude eight non-FI and three FI issuers that issued only guaranteed bonds. Thus, the corporate default study included 206 issuers, comprising 158 non-FIs and 48 FIs.
There were two defaults in 2020. The cumulative number of defaulters since 1994 has increased to 24 (19 issuers defaulted while still having ratings with TRIS Rating; the remaining five issuers defaulted after withdrawing their ratings). The 206 issuers in 2020 included 13 new issuers and six withdrawn issuers. The one-year stability rate of publicly announced ratings in 2020 (excluding 13 new issuers and six withdrawers) was 81.28%.
Rating actions during the year included six upgrades, 27 downgrades (excluding two issuers that withdrew their ratings after downgrades) and two defaults. The ratio of downgrades and defaults to upgrades was 4.83 times, up from 0.83 times in 2019. There were 31 changes in rating outlook, six upwards, and 25 downwards. Fifteen companies were placed on CreditAlert during the year, comprising one ?positive?, and 14 ?negative? CreditAlerts. Most CreditAlerts were resolved during the year. Only one issuer was downgraded and remained on ?negative? CreditAlert at the end of 2020.
The one-, two-, and three-year cumulative default rates during 1994-2020 increased slightly to 1.005%, 2.172%, and 3.181% from 0.998%, 2.172%, and 3.146%, respectively, during 1994-2019. Please note that the 1994-2019 figures have been changed slightly from the last year report since we removed the data of three FIs for whom we no longer assign shadow ratings.
Due to the impact from the Coronavirus Disease 2019 (COVID-19) pandemic, new corporate bonds issued and registered with the Thai Bond Market Association (ThaiBMA) declined by 36.2% year-on-year (y-o-y) to THB659.87 billion. Unrated bonds accounted for 18.60% of the total amount of bonds issued in 2020, increasing from 16.60% in 2019. Issuers in five industries, financial institutions, real estate development, food, energy, as well as commerce, together issued almost 67% of total bond issuances in 2020.
However, the value of outstanding long-term corporate debentures at the end of 2020 was rather flat, increasing by 0.32% y-o-y to THB3.38 trillion. The proportion of non-rated bonds to total outstanding bonds increased to around 15.13%, from 13.64% in 2019.
?
CORPORATE DEFAULT STUDY
RATING ACTIONS IN 2020
The corporate default study is based on 206 issuers, including 158 non-FIs and 48 FIs. There were six upgrades and 27 downgrades (excluding two issuers that withdrew their ratings after the downgrades). The upgrades comprised five non-FI and one FI issuers. Four FI and 23 corporate issuers were downgraded. The downgrade to upgrade ratio increased to 4.83 times in 2020, from 0.83 times in 2019. There were two defaults, comprising one FI and one non-FI issuer.
The one-year stability rate of publicly announced ratings in 2020 (excluding new issuers, rating withdrawals, and defaults) was 81.28%. There were 31 changes in outlook, comprising six upward revisions; and 25 downward outlook revisions. Fifteen companies were placed on CreditAlert during the year, comprising one ?positive?, and 14 ?negative? CreditAlerts. Only one ?negative? CreditAlert remained unresolved at the end of 2020.
? Ratings were mainly in the ?A? and ?BBB? categories
At the end of 2020, companies rated in the ?A? and ?BBB? categories made up the largest proportion of
TRIS Rating?s portfolio, accounting for 34.34% and 38.89% of publicly announced ratings (excluding withdrawals and defaults), respectively. The ratings of 13 new issuers were distributed across several rating categories: one ?B?, two ?BB?, five ?BBB?, two ?A?, one ?AA, and two ?AAA? ratings. Issuers rated in the lower ranges (i.e., ?BB?, ?B?, and ?C?) have consistently accounted for a small proportion of the rated companies. However, the number of issuers in these categories has increased over time. At the end of 2020, 20 issuers were rated below ?BBB-?, accounting for 10.10% of publicly announced ratings (excluding withdrawals and defaults).
? Cumulative default rates increased slightly
We calculated the average cumulative default rates for each rating category to estimate the probability of default during a specified time period after a company was rated. Despite an increase in the sample size, the growing number of defaulted issuers in 2020 has caused the one-, two-, and three-year average cumulative default rates during 1994-2020 to increase slightly from the period during 1994-2019. The one-, two-, and three-year cumulative default rates during 1994-2020 increased to 1.005%, 2.172%, and 3.181%, from 0.998%, 2.172%, and 3.146%, respectively, during 1994-2019.
Corporate Rating Transitions (1994-2020)
A rating transition is the probability of a given issuer rating moving to another rating category within a specified time period. Generally, the ratings of investment-grade issuers are more likely to remain at the same level over a one-year period than the ratings of non-investment grade issuers. The highlighted cells in Table 5 contain the stability rates of each rating category. For example, the stability rate for the ?AAA? issuers is 95.45%.
The rating stability of the investment grade companies exceeded 90%. For the ?A? rating category, 94.55% of the issuers in this category had their ratings maintained at this level in 2020. Around 2.92% of the ?A? rated issuers were upgraded to ?AA?, while 2.15% were downgraded to ?BBB?. However, the rating stability of the ?AA? rated issuers was lower than the rating stability of the ?A? rated issuers. This was due to the relatively small sample size of issuers in the ?AA? rating category. In 2020, there were 19 ?AA? rated issuers, compared with 68 ?A? rated issuers.
As credit ratings should reflect risk of default, the higher the rating, the lower the probability of default. However, due to both the small sample size and the severity, plus widespread financial crisis that led to multiple defaults in the financial sector in 1997, the default rate of the ?AA? rating category is abnormally higher than the default rate of the ?A? rating category.
One-Year Relative Corporate Ratings Performance
To measure the relative accuracy of ratings assigned by TRIS Rating, we focus on the relation between credit ratings (ranked from the highest rating, ?AAA?, to the lowest, ?C?) and the default rates of issuers in each rating category. Normally, a higher-rated entity should have a lower default probability relative to a lower-rated entity.
TRIS Rating measures rating performance or rating accuracy by plotting the cumulative proportion of a universe of rated issuers (ordered from the lowest rating, ?C?, to the highest rating, ?AAA?) against the cumulative proportion of defaulted issuers across all rating categories, which are also ranked from the lowest to the highest rating. This curve is called the cumulative accuracy profile (CAP) curve, also known as the ?Lorenz Curve?.
Chart 4 depicts the one-year relative corporate ratings performance, based on 1,891 observations of issuers rated by TRIS Rating during 1994-2020. The upper curve (as represented by the left end of the horizontal axis), or the ideal curve, is derived from the assumption that defaults occur only among the lowest-rated entities. The middle curve, or the CAP curve, is derived from the actual default rate of each rating category, drawing from the 1,891 observations of issuers rated by TRIS Rating during 1994-2020. The lower curve is a random curve. The random curve assumes that the assigned ratings have no relation with the default rates. Therefore, the cumulative percentage share of defaulters grows at the same rate as the cumulative percentage share of rated issuers. Generally, the closer the CAP curve resembles the ideal curve, the greater the accuracy of the rating model.
The CAP curve is based to calculate the accuracy ratio or the ?Gini Coefficient?. The closer the accuracy ratio is to one, the greater the rating accuracy it reflects of the rating model. The formula used to calculate the accuracy ratio is:
Accuracy ratio = area between CAP curve and random curve (Y)/area between ideal curve and random curve (X+Y)
If the credit ratings have no correlation with the defaulting cohorts, the CAP will resemble the random curve and the accuracy ratio will be equal to zero (0). On the contrary, if all defaults are concentrated among the lowest-rated issuers, the CAP curve will resemble the ideal curve and the accuracy ratio should be equal or close to one (1). If the accuracy ratio equals one, the assigned ratings are perfectly accurate.
From the 1,891 observations of issuers rated by TRIS Rating during 1994-2020, there were 19 observations in which an issuer defaulted in a one-year observation period. The default rate was 1.005%, a slight increase from 0.998% during 1994-2019. From the CAP curve, issuers rated at ?BBB+? category and below represent 43.5% of the overall observations. However, 84.2% of all defaulters (16 out of 19 defaulters) were in this group.
The accuracy ratio, calculated from the observations during 1994-2020, is equal to 0.54, lower than 0.57 obtained in the previous assessment covering 1994-2019. The lower accuracy ratio was due to the default of Thai Airways International PLC (THAI), which was rated at ?A/Stable? at the end of 2019. We classified THAI as a government-related entity, thus, its rating had incorporated the expected extraordinary support from the government. However, the slump in the airline industry due to the COVID-19 pandemic coupled with the heavy debt burden of THAI caused the government to decide to let THAI enter the bankruptcy process. The relatively low accuracy ratios are also attributed to the small number of observations and the financial crisis faced by all issuers in 1997.
If we use observations during the last 10 years (2010-2020), the accuracy ratio improves slightly to 0.56 but dropped sharply from 0.73 during 2009-2019. There were 1,272 observations in this cohort and only five observations defaulted during this period. This implies an overall default rate of 0.39%, leaving the remaining 99.61% of the observations with no defaults.
STRUCTURED FINANCE DEFAULT STUDY
There were only seven structured finance transactions rated by TRIS Rating. However, four were fully guaranteed by the originators and one transaction was partially guaranteed by the originator. These transactions are not included in this study. The two remaining transactions are LSPV Co., Ltd. and DAD SPV Co., Ltd. The first transaction, LSPV, is involved with an inventory securitization. This issue was rated ?A-? in 1999 and was fully redeemed in 2002. The second transaction, DAD SPV, is a securitization program backed by a 30-year lease and service payment agreement from the Treasury Department. The rating of the second transaction has been maintained at ?AAA?.
Appendix I
1. Methodology and Definitions
1.1 Definition of Default
TRIS Rating assigns a ?Default (D)? or ?Selective Default (SD)? rating to an entity or a company on the date the entity or the company misses a payment of a financial obligation, according to the terms and conditions stipulated in the borrowing agreement, irrespective of whether the financial obligation issue is rated or unrated.
1.2 Cumulative Default Rates
The default rate is the number of defaulted issuers as a percentage of the total number of issuers in each rating category. Therefore, the default rate represents the default probability of companies in each particular rating category. The cumulative default rate tends to rise over time.
For example, the three-year cumulative default rate of any particular rating category is the probability that the companies rated in that category will default within three years. The average three-year cumulative default rate is computed by subtracting the average three-year cumulative survival rate from 100%. The average three-year cumulative survival rate is derived by multiplying the first-year survival rate by the second-year rate and the third-year rate. The survival rate for any given year is calculated by subtracting the default rate of that year from 100%.
1.3 Rating Transition Rates
The rating transition rate is the percentage of the issuer ratings changing from a particular rating category at the beginning of a given year to another rating category by the end of that year. To compute a one-year rating transition rate, issuers rated in each rating category at the beginning of the year are tracked for any rating changes by the end of the calendar year.
2. Scope
2.1 Credit ratings include:
Corporate Ratings
2.1.1 For corporate ratings, the ratings used are the ratings of entities (companies or issuers) rather than ratings of the debenture issues (or debentures). The reason is to simplify the default rate calculation process, particularly in cases in which a company has issued several debentures. The different debenture issues might receive different ratings due to different priorities of claims and different expected losses in the case of default.
2.1.2 In the case that the issuer wants to publicly announce only its issue rating, TRIS Rating may also assign a shadow rating to the issuer. Previously, the shadow rating was assigned internally and used in the default study. However, due to the discontinuation of information, TRIS Rating will no longer include the shadow rating in the default study. Therefore, starting from this year, we have excluded from our default study all shadow ratings assigned to three issuers during 2004-2020, 2013-2020, and 2018-2020, respectively.
2.1.3 The period of analysis covers ratings from the first year of TRIS Rating?s operation in 1993 until year-end 2020. The number of rated companies at the end of each year will be recorded as the static pool for the following year. For example, rated clients at the end of 1993 are recorded as the 1994 pool.
Structured Finance Ratings
2.1.4 TRIS Rating also provides the one-year rating transition rates of structured finance securities. For the ratings of structured finance securities, TRIS Rating uses the ratings of the debentures or a series of debentures issued under the same program.
2.1.5 TRIS Rating will include rating transition rates of structured finance securities, e.g., asset-backed securities (ABS), collateralized debt obligations (CDO), commercial mortgage-backed securities (CMBS), and residential mortgage-backed securities (RMBS).
2.2 Credit ratings exclude:
2.2.1 Ratings that are not publicly announced
Ratings assigned by TRIS Rating can be categorized into those that are publicly announced and those that are kept private, based on the issuers? wishes.
2.2.2 Selected structured finance ratings
This category includes ratings of project finance instruments, such as Khanom Electricity Generating
Co., Ltd. (KEGCO), and partially- or fully-guaranteed debentures.
2.2.3 Local government ratings
This category includes the rating of Bangkok Metropolitan Administration (BMA).
2.2.4 Ratings that are withdrawn in the specified period
A company that was initially rated by TRIS Rating in mid-1994 but withdrew its rating in 1997 will be included in the static pools for 1995 and 1996 but not for 1997.
2.2.5 Supranational and sovereign ratings
This category includes the rating of the Lao People?s Democratic Republic (Lao PDR), Neighboring Countries Economic Development Cooperation Agency (NEDA), and Credit Guarantee and Investment Facility (CGIF).
2.3 Data used to calculate default rates
Static pools are established to represent the sample groups. In any given year, a static pool includes all entities with active ratings at the beginning of a year that remain rating clients at the end of that year. For example, there were 20 issuers rated by TRIS Rating on 1 January 1995 and all 20 issuers had remained clients through 31 December 1995. The 1995 static pool comprised 20 issuers. The default records of these 20 issuers are tracked in each subsequent year.
In any given year, the pool is static because no issuer is taken out of the pool even though the issuer may subsequently withdraw its rating. For example, Dhana Siam Securities Co., Ltd. (DS) was initially rated in 1993 but withdrew its rating in 1997, shut down operations, and then defaulted on 14 August 1998. In this circumstance, DS was included in the static pool for 1994, 1995, and 1996, but not for 1997. The subsequent default of DS in 1998 was counted as a two-year default for the 1996 static pool, a three-year default for the 1995 static pool, and a four-year default for the 1994 static pool.
3. Database Limitations
The corporate debentures market in Thailand is in the developing stage. The Thai bond market is largely dominated by debt instruments issued by the government, the Bank of Thailand (BOT), and state enterprises. These debt instruments are not required by law to have credit ratings. As a result, TRIS Rating has considerably fewer clients than the long-established international rating agencies.
One problem with the limited sample size is that it exaggerates the default rate statistics because the number of observations in each rating category is used as the denominator to calculate the default rate. Thus, the fewer the observations in any particular rating category, the higher the default rate.
4. Impact from the Financial Crisis on Cumulative Default Rates
The financial crisis in 1997 and 1998 forced the government to shift to a managed float exchange rate system. This action raised the value of foreign denominated debts in terms of local currency. The credit risks of many financial and non-financial companies rose significantly as a result. As shown in Table 3, the annual default rates of the companies rated by TRIS Rating in 1997 and 1998 were unusually high at 35% and 30%, respectively. The annual default rate of 33% in the ?AA? rating category in 1997 was the result of a default by a FI that was ordered by the BOT to cease operations. The default rate is thus overstated because of the relatively small number of rated issuers in that particular rating category. In 1997, there were only three companies in the ?AA? rating category and 10 companies rated ?BBB?. The default of one company rated ?AA? and five companies rated ?BBB? made the annual default rates equal to 33% and 50% in these two rating categories in 1997. Five out of six defaulting issuers in 1997 were finance companies that defaulted after they were ordered to cease operations by the BOT.