SPECIAL COMMENT
Table of Contents:
SUMMARY 1
INTRODUCTION 2
DEFAULT PACE ROSE SHARPLY IN 2015 2 DEFAULT RATE EXPECTED TO RISE IN
2016 9
RATING ACCURACY METRICS 13 MOODY’S RELATED RESEARCH 15 METHODOLOGY AND DATA SOURCES 16 RATING ACCURACY: REDESIGNED SRA VERSUS PREVIOUS SRA 17 GUIDE TO DATA TABLES AND CHARTS 21 DATA TABLES AND CHARTS 22
Analyst Contacts:
NEW YORK +1.212.553.1653 Sharon Ou +1.212.553.4403
Vice President - Senior Credit Officer
[email protected] John Kennedy +1.212.553. 1303 Associate Analyst [email protected] Yang Liu +1.212.553.2961 Associate Analyst [email protected] Varun Agarwal +1.212.553.4899
Assistant Vice President – Analyst
Sumair Irfan +1.212.553.1035
Associate Analyst
Kumar Kanthan +1.212.553.1428
Senior Vice President
» contacts continued on the last page
Annual Default Study: Corporate Default and
Recovery Rates, 1920-2015
Summary
» At the forecasted 4.0%, the default rate for Moody’s speculative-grade issuers in 2016 is
expected to approach the historical average of 4.2% since 1983. The predicted 4.0% continues
a negative trend in the spec-grade default rate, which increased substantially from 1.9% in
2014 to 3.5% in 2015.
» The default rate for all of Moody’s-rated corporate issuers is also expected to rise to 2.1% in
2016, the highest rate since the 2008-2009 global financial crisis. This will be a second
consecutive annual increase in default rates from 0.9% in 2014 and 1.7% in 2015.
» The 2.1% predicted default rate implies 138 defaults for 2016, which if realized, will result in a
more than 30% increase in defaults from 2015.
1»
Commodity prices that will remain at low levels, an economic cycle that is getting long in the
tooth, and increased investor risk aversion as reflected in widening high yield spreads are key
players this time around. Further interest rate hikes this year by the Fed may exacerbate the
default situation. Our default rate forecasting model confirms that commodity sectors will
remain in significant distress in 2016.
» The default tally was 109 in 2015, almost doubling the prior year’s 55. Not surprisingly, default
volume was also higher in 2015 ($97.9 billion) relative to 2014 ($71.4 billion); several sizeable
defaults by the likes of Caesars Entertainment ($20.5 billion), Alpha Natural Resources ($4.5
billion) and Samson Investment Company ($4.2 billion) added to the default volume.
» By region, default counts roughly doubled from 2014 to 2015 in all regions except Middle East
& Africa. Of those 109 defaults in 2015, 62 were in North America and 27 were in Europe. The
default count in Latin America was also notable as it was near the level seen in the previous
crisis.
» Not surprisingly, credit quality weakened in 2015 as downgrades outpaced upgrades. As a
result, the rating drift, measured in notches, fell to -8.5% from the prior year’s -0.7%. Quality
deterioration was most significant in Oil & Gas (-70.1%) and Metals & Mining (-65.6%).
1 This 2.1% default rate speaks to issuers in the January 1, 2016 cohort. In 2015, the comparable count was 103 when restricted to the January 1 cohort. Including companies outside of the January 1 cohort, the default count was 109 in 2015.
THIS REPORT WAS REPUBLISHED ON 10 MAY 2016 WHEN WE ADDED ANNUAL DEFAULT STUDY TO THE REPORT TITLE. ANOTHER CHANGE WAS MADE IN EXHIBIT 31 WHERE WE REMOVED THE Caa1 AND Caa3 DEFAULT RATES PRIOR TO 1997 WHEN MOODY’S DID NOT HAVE ALPHANUMERIC RATINGS WITHIN Caa.
»
Measured by post-default trading prices, the issuer-weighted average recovery rate for senior unsecured
bonds was 37.9% in 2015, lower than 2014’s 46.4%.
» Rating accuracy measures like Average Default Position (“AP”) generally show that Moody’s ratings
have been successfully rank ordering credit risk. The average one-year AP since 1983 stands at 91.5%.
This report comprises Moody's 29
thannual default study, in which we update statistics on the default, loss
and rating transition experience of corporate bond loan and deposit issuers for 2015, as well as for the
historical period since 1920. This study covers financial institutions, corporates, and regulated utilities that
have long-term debt ratings.
Introduction
This report is primarily concerned with the statistical documentation of corporate defaults among
Moody's-rated long-term debt issuers and the performance of Moody's ratings for the year 2015, as well as the
historical period since 1920. The first section of this study reviews some of the major trends characterizing
defaults, recoveries and rating actions in 2015, including analyses by industry and geography. The second
section discusses the default rate outlook and the drivers behind our forecasts. The final section examines
the performance of Moody’s ratings since 1983.
Results presented in this study are based on a proprietary database of ratings and defaults for industrial and
transportation companies, utilities and financial institutions. This database covers the credit experiences of
over 20,000 corporate issuers that had long-term rated bonds, loans and/or deposits between 1920 and
2015. This report also introduces some methodology changes including 1) an expansion of the universe of
this study to include banks with only deposit ratings, 2) a modified algorithm that estimates the senior
unsecured or equivalent ratings for corporate issuers and 3) an updated treatment of defaulted issuers that
allows them to re-enter various cohorts in a more appropriate timely manner. These changes are described
in details in the Methodology and Data Source section.
Default pace rose sharply in 2015
Default counts doubled
The pace of defaults increased rapidly in 2015, sending the default tally up to a level not seen since 2009.
Worldwide, 109 Moody’s-rated corporate issuers defaulted in 2015, almost doubling 2014’s 55.
2Defaults
were roughly evenly paced in the first three quarters with an average of 25 defaults per quarter. However, in
the fourth quarter, default counts jumped to 35, driven by continued low oil prices. Measured by volume,
about $97.9 billion of debt went into default in 2015, comprising of $77.5 billion in bonds and $20.4 billion
in loans. In comparison, the default volume was $71.4 billion in 2014 consisting of $39.4 billion in bonds,
$28.2 billion in loans and $3.7 billion in deposits. The largest 2015 default was Caesars Entertainment
Operating Company Inc., which filed for bankruptcy in January. With over $20 billion in debt at default,
Caesars was the seventh largest defaulter in history among Moody’s-rated non-financial corporations.
3Like any other year, most of the 2015 corporate defaults were found in North America and Europe, which is
not surprising as most of the Moody’s rated issuers are domiciled in these two regions. Of the 109 defaults,
62 (57%) were in North America. Europe had 27 (25%) while the default count in Latin America (10) rose to
near-crisis level. Compared to 2014, default counts have roughly doubled in all regions except Middle East &
2 These default counts include issuers outside of the January 1st cohorts. The list of the 2015 defaults can be found in Exhibit 15 in the appendix. 3 The largest defaulter remains General Motors, which had roughly $50 billion of debt at default.
This publication does not announce a credit rating action. For any credit ratings referenced in this
publication, please see the ratings tab on the issuer/entity page on
www.moodys.com for the most updated credit rating action information and rating history.
Africa, which recorded only one default in both years. In terms of volume, $73.6 billion (~75%) of defaulted
debt was from North America, followed by Europe, which contributed $15.3 billion (16%).
4By default type, 40.3% of last year’s defaults were distressed exchanges. The rest were roughly split
between bankruptcies (30.3%) and payment defaults (29.4%). Exhibit 1 presents the annual default counts
and defaulted debt volumes for the period 1970-2015.
EXHIBIT 1
Defaults jumped in 2015
Default rate highest in six years
As the default count nearly doubled in 2015 from 2014, the issuer-weighted annual default rate jumped to
1.7% in 2015 from 0.9% in 2014 (see Exhibit 2). This marked the highest annual rate since 2009. Among
speculative-grade issuers, the default rate showed a similar trend—rising to 3.5% from 1.9% and also
reaching the highest level since the 2008-2009 global financial crisis. To put these in historical prospective,
the one year default rate has averaged 1.6% for all-rated issuers and 4.2% for speculative-grade credits
since 1983.
5Measured on a dollar volume basis, Moody’s corporate bond default rates tell a similar story.
For all of Moody’s-rated issuers, the volume-weighted default rates finished 2015 at 0.8%, twice as high as
2014’s 0.4%. The increase in the dollar-weighted bond default rate was driven by more sizable defaults in
2015 than in 2014. Among speculative-grade issuers, the comparable rate rose to 3.4% from 1.7%.
4 2015 also recorded a sovereign default by the Government of Ukraine as it missed the principal and interest payments on several bonds during the process of a broader debt restructuring that is aimed at extending debt maturities and reducing debt burden.
5 These averages are weighted by cohort size.
0 50 100 150 200 250 300 350 400 19 70 19 71 19 72 19 73 19 74 19 75 19 76 19 77 19 78 19 79 19 80 19 81 19 82 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15
EXHIBIT 2
Default rate highest since 2009
Credit quality deteriorated especially in commodity sectors
Credit quality among Moody’s-rated issuers weakened in 2015 with rating downgrades outpacing upgrades.
Rating drift, which is defined as the average upgraded notches minus the average downgraded notches per
issuer, fell to -8.5% in 2015 from -0.7% in 2014 (see Exhibit 3). During 2015, credit quality deteriorated
throughout the year except during the second quarter. Among the four quarters, rating drift fell the most in
the last quarter from -2.9% to -4.8%. In the second quarter, rating drift was positive at 1.2%. It should be
noted that the majority of the second quarter’s rating actions (both upgrades and downgrades) were in the
banking sector, following Moody’s updated banking methodology. Excluding banks, second quarter’s rating
drift would have been -0.4%.
EXHIBIT 3
Rating Drift Fell in 2015
Across sectors, we find that credit deterioration was most severe in commodity sectors, i.e. Oil & Gas and
Metals & Mining (see Exhibit 4). Specifically, Oil & Gas’ rating drift fell to -70.1% of a rating notch in 2015
from -2.4% a year earlier. Similarly, the comparable metric declined to -65.6% from -11.1% for Metals &
Mining. Both sectors’ one-year rating drifts were significantly below their historical averages of -8.2% and
-18.2%, respectively. On the other hand, Consumer Transportation benefited from low commodity prices
and saw its rating drift jump to 31.3% in 2015 from -20.3% in 2014. Similarly, the rating drift for
automotive issuers rose to 14.1% from -0.8%. Historically, these two sectors’ drift both averaged -21%.
0% 2% 4% 6% 8% 10% 12% 14% 16% 18% Allcorp Spec-grade -60.0% -50.0% -40.0% -30.0% -20.0% -10.0% 0.0% 10.0% 20.0% 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15
EXHIBIT 4
Credit Quality Weakened Most in Commodity Sectors
-80.0% -60.0% -40.0% -20.0% 0.0% 20.0% 40.0% 60.0%
transportation: consumer automotive forest products & paper beverage, food, & tobacco fire: finance fire: real estate fire: insurance utilities: water consumer goods: durable retail utilities: electric wholesale hotel, gaming, & leisure transportation: cargo sovereign & public finance chemicals, plastics, & rubber banking utilities: oil & gas media: diversified & production media: broadcasting & subscription containers, packaging, & glass telecommunications healthcare & pharmaceuticals high tech industries services: business environmental industries media: advertising, printing & publishing energy: electricity capital equipment construction & building aerospace & defense services: consumer consumer goods: non-durable metals & mining energy: oil & gas
Oil & Gas under stress
Unlike 2009, when defaults were widely spread among a variety of industries, last year’s defaults reflected
sector-specific problems, leading to a high concentration of defaults in particular pockets. Oil & Gas
recorded 30 defaults and contributed the largest proportion of corporate defaults, comprising 28% of the
total count.
6Metals & Mining had 13 defaults and was the second biggest contributor among non-financial
sectors. Among financial institutions, the banking sector was the most troubled as it observed 20 defaults.
The vast majority of those bank defaults were from Greece, Ukraine and Russia, reflecting economic stress in
those countries.
When measured by default volume, Oil & Gas topped the list again by accounting for 32% of the total, with
the next highest shares coming from Hotel, Gaming, & Leisure (21%)
7and Metals & Mining (14%). Besides
Caesars, 2015 has observed other sizable defaults including Samson Investment Company ($4 billion) and
Alpha Natural Resources ($4 billion). Exhibit 5 shows the distribution of 2015 defaults by specific industries.
EXHIBIT 5Oil & Gas Accounts for Roughly 30% of Defaults
Although the Oil & Gas sector accounted for 28% of defaults last year, it was not the sector with the
highest rate of default: that distinction belongs to Metals & Mining, which had a 6.5% default rate in 2015.
Oil & Gas followed right behind with a rate of 6.3% (see Exhibit 6).
6 Including financial and non-financial defaults.
7 Mainly driven by Caesars Entertainment Operating Company.
0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0%
energy: oil & gas banking metals & mining services: business construction & building retail beverage, food, & tobacco consumer goods: non-durable hotel, gaming, & leisure aerospace & defense automotive fire: insurance forest products & paper services: consumer consumer goods: durable environmental industries fire: real estate high tech industries media: advertising, printing & publishing media: broadcasting & subscription media: diversified & production telecommunications utilities: electric wholesale
EXHIBIT 6
Default Rates by Sector
Industry Default Rates* Industry Default Rates*
metals & mining 6.5%
fire: insurance 0.8%
energy: oil & gas 6.3%
telecommunications 0.5%
consumer goods: non-durable 4.7%
fire: real estate 0.5%
environmental industries 4.3%
high tech industries 0.5%
forest products & paper 4.2%
utilities: electric 0.3%
media: diversified & production 3.6%
automotive 0.0%
services: consumer 2.8%
capital equipment 0.0%
aerospace & defense 2.6%
chemicals, plastics, & rubber 0.0%
construction & building 2.4%
containers, packaging, & glass 0.0%
hotel, gaming, & leisure 2.3%
energy: electricity 0.0%
services: business 2.3%
fire: finance 0.0%
media: advertising, printing & publishing 2.2%
healthcare & pharmaceuticals 0.0%
Retail 2.1%
sovereign & public finance 0.0%
consumer goods: durable 2.0%
transportation: cargo 0.0%
beverage, food, & tobacco 2.0%
transportation: consumer 0.0%
wholesale 1.7%
utilities: oil & gas 0.0%
Banking 1.5%
utilities: water 0.0%
media: broadcasting & subscription 1.0%
* Issuer-weighted
Recovery rate lower in 2015
In Exhibit 7, we present the average recovery rates for debt defaulted in the past two years and put them in
context with the historical averages.
8The table shows that over the past three decades recovery rates were
generally correlated with the priority of claim in the capital structure, with a higher priority of claim
enjoying a higher average rate of recovery.
9For example, first lien bank loans’ recovery rates average 66.6%
on an issuer-weighted basis and 62.3% on a dollar volume basis; both are highest of all defaulted
instruments. This is not surprising given their secured nature, seniority in capital structure and strong
protective covenants.
Exhibit 7 further shows that 2015’s recovery rates were in large part lower than the comparable rates in
2014. This reinforces the findings in our prior studies that recovery rates and default rates have been
negatively correlated. For example, the issuer-weighted recovery rate for senior unsecured bonds was 37.9%
in 2015, down from 46.4% in 2014. Measured by dollar volume, the comparable rate was 33.3% in 2015
versus 40.8% in 2014. To put this in historical prospective, the senior unsecured bond recovery rates in 2015
were similar to their long term averages.
8 In Exhibit 7, we use market prices to proxy recoveries.
9 Average recovery rates of senior unsecured bonds and other debts can be based on different defaulters because some defaulters may have senior unsecured bonds and no other debts. Among those 2015 defaulters that have recovery estimates on both senior unsecured and subordinated bonds, the recovery estimates for the senior unsecured bonds are consistently higher than the subordinated bonds of the same issuers. Please see Exhibit 19 for more details.
EXHIBIT 7
Exhibit 7 - Average corporate debt recovery rates measured by trading prices
Panel A Recoveries
Issuer-weighted recoveries Volume-weighted recoveries
Lien Position 2015 2014 1983-2015 2015 2014 1983-2015
1st Lien Bank Loan 63.4% 78.4% 66.6% 52.0% 80.6% 62.3%
2nd Lien Bank Loan 32.1% 10.5% 31.8% 21.5% 10.5% 27.6%
Sr. Unsecured Bank Loan n.a. n.a. 47.1% n.a. n.a. 40.2%
1st Lien Bond 53.5% 73.6% 53.4% 58.2% 86.5% 53.4%
2nd Lien Bond 26.0% 51.0% 49.7% 20.6% 75.5% 47.4%
Sr. Unsecured Bond 37.9% 46.4% 37.6% 33.3% 40.8% 33.7%
Sr. Subordinated Bond 36.6% 39.1% 31.1% 20.3% 24.3% 25.8%
Subordinated Bond 58.5% 38.8% 31.9% 56.8% 38.0% 27.1%
Jr. Subordinated Bond 14.0% n.a. 24.2% 14.0% n.a. 17.1%
Panel B Observation counts
Issuer count Dollar volume (in billions of USD)
Lien Position 2015 2014 1983-2015 2015 2014 1983-2015
1st Lien Bank Loan 15 9 444 6.4 24.8 260.4
2nd Lien Bank Loan 2 1 66 1.3 0.2 11.4
Sr. Unsecured Bank Loan 0 0 67 0.0 0.0 33.8
1st Lien Bond 27 10 290 17.4 10.0 120.0 2nd Lien Bond 6 8 44 6.9 6.6 24.8 Sr. Unsecured Bond 63 26 932 44.7 19.7 629.1 Sr. Subordinated Bond 5 3 510 1.1 1.0 112.4 Subordinated Bond 6 1 411 2.2 0.6 80.3 Jr. Subordinated Bond 1 0 22 0.0 0.0 3.0
The above recovery data are based on trading prices at or post default.
10An alternative recovery measure is
based on ultimate recoveries, or the value creditors realize at the resolution of a default event. For example,
for issuers filing for bankruptcy, the ultimate recovery is the present value of the cash and/or securities that
the creditors actually receive when the issuer exits bankruptcy, typically 1-2 years following the initial
default date.
11In Exhibit 8, we present data on ultimate recovery rates for North American companies included in Moody’s
Ultimate Recovery Database (“URD”).
12The average “firm-wide” recovery rate
13for the fourteen default
resolutions was 55.7% in 2015 compared to 71.0 % for the 11 companies that emerged from default in
2014. During both 2015 and 2014, the family recovery rates exceeded the historical average rate of 55.1%.
The higher family recovery rates can be mostly attributed to the fact that out of the 14 default resolutions
in 2015, 57% of them (eight) were distressed exchanges and 29% (four) were pre-arranged bankruptcies,
which are historically characterized by higher family recovery rates than regular bankruptcies. Looking at
10 For distressed exchanges, we take trading prices at default. For other types of defaults, we take trading prices approximately one month after default. 11 For details, see Moody’s Special Comment Moody’s Ultimate Recovery Database, April 2007.
12 The analysis on ultimate recovery is provided by David Keisman and Julia Chursin. The data are from Moody’s Ultimate Recovery Database, which includes robust detailed recovery information for over 5,200 loans and bonds from more than 1,000 North American corporate defaulters since 1987.
recovery rates by lien positions, the table clearly reflects recovery rates’ correlation with instruments’
priorities of claims, with recovery rates falling as we descend the capital structure.
EXHIBIT 8
Average Corporate Debt Recovery Rates Measured by Ultimate Recoveries, 1987-2015*
Emergence Year Default Year
Lien Position 2015 2014 1987-2015 2015 2014 1987-2015
Loans 100.0% 83.6% 80.4% 100.0% 79.8% 80.4%
Senior Secured Bonds 81.3% 58.6% 63.3% 77.3% 75.8% 63.3%
Senior Unsecured Bonds 41.8% 50.7% 48.8% 42.8% 33.9% 48.8%
Subordinated Bonds 31.2% 0.0% 28.2% 2.6% 44.1% 28.2%
*2015 Loans' recovery rate is based on 11 observations (by year of default) and 14 observations (by year of emergence).
Ten out of 14 loans that emerged in 2015 were not part of distressed exchanges, and hence realized full recovery by definition, and thus contributed to a higher than average overall recovery rate in 2015 (by year of emergence). In case of "by year of default" sample of loans, similarly, 8 out of 11 loans were not restructured in a distressed exchange and realized 100% recovery.
Unusually high recoveries of 2015 Sr. Secured bonds stem from a small sample of 8 observations by year of default and 13 by year of emergence. These debt instruments were parts of only 6 defaults and 9 default resolutions last year. In both cases samples consisted of mainly distressed exchanges and prepackaged bankruptcies, which on average result in higher recovery rates for investors.
Moody’s credit ratings are opinions of relative expected credit losses, which are a function of both the
probability of default and severity of default (“LGD”). Exhibit 9 shows annual average credit loss rates from
1983 through 2015 for Moody’s-rated corporate issuers. The chart indicates that the average credit loss rate
among all Moody’s-rated issuers rose to 1.0% in 2015 from 0.5% in 2014. Historically, the average annual
credit loss rate for Moody’s-rated issuers is 0.9% since 1983.
EXHIBIT 9
Annual Credit Loss Rates Rose in 2015
Default Rate Expected to Rise in 2016
Default rate approaching historical average in a year
In the past six years, the global economy has enjoyed a benign default environment as accommodative
monetary policies have fueled the corporate debt market with abundant liquidity, allowing many low-rated
issuers to refinance when needed. The party is likely coming to an end soon as the global default rate is
expected to approach the historical average mark by the end of 2016. Based on Moody’s default rate
0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15
forecast model, the default rate for all of Moody’s-rated issuers is expected to rise to 2.1% by the end of this
year, which—if realized—will surpass the average of 1.6% since 1983. Among speculative-grade issuers, the
default rate is predicted to climb to 4.0%, just below the historical average of 4.2% (see Exhibit 10). These
forecasts are made under our baseline scenario, which does not expect a global economic recession. In
addition, we assume the US high yield spread will widen from the current level of 660 bps to 750 bps at the
end of 2016 whereas the unemployment rate will ease from 5.0% to 4.8%.
14EXHIBIT 10
Benign Default Environment Ending Soon
Upward pressure from both credit red flags and widening spreads
Most drivers behind the rise of defaults in 2015 will likely continue into 2016, putting upward pressure on
the global default rates. Commodity prices are expected to stagnate at low levels and rating actions are
anticipated to remain downward biased. High yield spreads are assumed to increase further and be more
volatile in 2016 due to concerns of a prolonged energy crisis, more possible interest rate actions by the Fed
and the fear of a worse-than-expected economic slowdown in China. On the other hand, a stable US
economy, together with global accommodative monetary policies, should provide counteracting forces and
prevent the default rate from rising sharply.
On the rating side, credit deterioration has surfaced in at least two dimensions. First, we observe a rapid
increase in the number of low-rated new issuers entering the corporate debt market. Between 2010 and
2015, the number of new issuers with Caa
15or below ratings averaged 235 per year, representing more than
a third of the entire rated universe.
16Before the global financial crisis, that number was much lower at 94
during 2004-2006, which accounted for 15% of the rated portfolio. Similar patterns were found before the
2001 and 2008 economic cycle turns, and we believe the recent new Caa issuers will once again sow default
seeds for 2016-2017. Second, while roughly 70% of Moody’s-rated issuers had stable outlooks at the
beginning of 2016, we expect rating actions to remain downward biased this year as the number of
unfavorable watchlist/outlook assignments more than doubled the favorable ones. Specifically, 7.5% of
issuers were put on watch for possible downgrade and 13.6% of issuers had negative outlooks. This
compares with 0.9% of issuers on watch for upgrade and 7.3% with positive outlooks. Among Caa-C issuers,
14 The spreads are option adjusted.
15 The ratings refer to senior unsecured or equivalent ratings, which are often times lower than the corporate family ratings. Therefore, those newly rated issuers include some companies with single B corporate family ratings.
16 These are issuer counts, not corporate family counts.
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0%
Actual-Allcorp Baseline forecast-Allcorp Pessimistic Forecast-Allcorp
the most vulnerable credits, 5.3% were on watch for downgrade and 23.6% had negative outlooks. On the
other hand, only 0.6% were on watch for possible upgrade and 5.9% had positive outlooks.
On the macro side, the global default rate forecast is driven by the level and change in the US high yield
spread, which has been a leading indicator historically. The spread has widened noticeably from 483 bps to
660 bps in 2015. It is expected to continue its upward trend in 2016 and more than offset the gradual
improvement in the US unemployment rate.
Commodity sectors to remain in trouble but stress will also grow outside of those sectors
Across industries, we anticipate the stress among Oil & Gas issuers to continue, and this sector will likely
remain the biggest contributor of defaults in 2016. Exhibit 11 shows the baseline one-year global default
rate forecasts by industry.
17In this chart, we apply the same economic assumption to all industries, so the
only factor driving the different forecasted default rates is the underlying rating histories and current ratings
of the issuers in those industries. Among the 35 industries, our model expects 20 of them to have higher
default rates in 2016 relative to their long-term averages.
Measured by default rate, the most troubled sector is Metals & Mining (7.0%). The next highest forecast
belongs to Oil & Gas (6.0%). These forecasts, if realized, will be more than double and triple their historical
averages, respectively.
Exhibit 11 also shows the default rate forecast implied by Expected Default Frequency (EDF).
18As shown in
the chart, EDF-implied ratings indicate a worse default rate outlook for commodity sectors than those
suggested by the Credit Transition Model. This certainly reflects the recent volatility in the equity market
that may not be fully captured in the high yield spread or the rating factors.
19In addition, the EDF-implied
default rate may also be higher as that portfolio includes some unrated issuers which may have relatively
weaker credit qualities. Nevertheless, EDF implied forecasts reinforce the high default risk for both Oil & Gas
and Metals & Mining in 2016.
17 The industry default rate forecasts include both investment-grade and speculative-grade issuers. 18 Note that the EDF-implied default rate forecast includes rated and unrated issuers.
EXHIBIT 11
One-Year Corporate Default Rate Forecasts by Industry
As mentioned before, the rise in 2015’s default rate was mainly driven by commodity sectors. Excluding Oil
& Gas and Metals & Mining, the global default rate remained low for most of the year. As shown in Exhibit
12, while the default rate for the entire Moody’s-rated universe jumped from 0.9% in 2014 to 1.7% in 2015,
the rate only rose from 0.8% to 1.1% when we exclude issuers in the Oil & Gas and Metals & Mining
sectors. Looking into 2016, we expect stress to also grow outside of the commodity sectors. Excluding Oil &
Gas and Metals & Mining issuers, the global default rate is anticipated to climb to 1.6% by the end of 2016.
0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0%
Metals & Mining Energy: Oil & Gas Environmental Industries Wholesale Aerospace & Defense Media: Advertising, Printing & Publishing Forest Products & Paper Services: Consumer Consumer goods: durable Services: Business Retail High Tech Industries Transportation: Cargo Hotel, Gaming, & Leisure Containers, Packaging, & Glass Media: Broadcasting & Subscription Capital Equipment Construction & Building Consumer goods: non-durable Healthcare & Pharmaceuticals FIRE: Finance Chemicals, Plastics, & Rubber Automotive Telecommunications Energy: Electricity Beverage, Food, & Tobacco Media: Diversified & Production Transportation: Consumer Banking FIRE: Insurance FIRE: Real Estate Utilities: Oil & Gas Utilities: Water Utilities: Electric Sovereign & Public Finance
EXHIBIT 12
Stress to Grow Outside of Commodity Sectors
While our baseline forecast is likely the one that is going to unfold, we acknowledge there is risk associated
with the global economic conditions, especially outside of the US. Indeed, in just one and a half months,
most countries have seen a 20% plummet in the stock markets, which reflects worries over prolonged
commodity stress, uncertainties from geopolitical developments in Russia, the recession in the economies
of several emerging market countries and the fear of a hard landing in the Chinese economy. Although we
estimate only a 4% chance our pessimistic scenario being realized, the worldwide economy could contract
with the unemployment rate rising to 9.4% and the high yield bond spread widening to 1530 bps. In that
case, the global default rate is expected to rise to 6.4% for all rated issuers and 11.6% for speculative-grade
issuers (see Exhibit 10).
Rating Accuracy Metrics
Moody’s ratings have historically proven to be effective predictors of default. This can be seen in Exhibit 13,
which plots the median ratings of roughly 2,000 corporate issuers that defaulted from 1983 to 2015. The
chart demonstrates that, historically, Moody’s-rated issuers have been downgraded to the B1 level as early
as five years prior to default. The comparable rating was lower at B2 among issuers that defaulted in 2015.
The median rating one year prior to default was Caa2 among last year’s defaulters, two notches lower than
that rating measured over the entire period 1983-2015.
0.0% 0.5% 1.0% 1.5% 2.0% 2.5%
EXHIBIT 13
Median Ratings Prior to Default, 2015 vs. Long-Term Average
The evolution of median ratings presented in Exhibit 13 above demonstrates that Moody’s corporate ratings
are correlated with subsequent default experience. To further demonstrate the ability of ratings to separate
issuers with low credit risk from those with high credit risk, we use the Average Position of defaults (“AP”) to
evaluate the accuracy of Moody’s ordinal rating systems (see Exhibit 14).
20AP measures the average
position for defaulters with position defined as the percentage of issuers with higher or equal ratings. A
greater AP indicates a more discriminatory rating system as there are more issuers rated higher than the
defaulters, or equivalently that defaulters are generally found in lower rating categories. Exhibit 14 reveals
that between 1983 and 2015, the Average Position of defaults has been consistently high during the entire
period, with an average of 91.5% for the one-year horizon and 86.6% for the five-year horizon. This
indicates that Moody’s ratings have been effective in predicting defaults over both the short- and long-term
periods. The lowest one-year AP was observed in 2008 when Lehman Brothers and several other high grade
financial institutions failed. Since then, the AP has quickly recovered and reached 87.9% in 2015.
EXHIBIT 14
One- and Five-Year Accuracy Default Position by Cohort Year, 1983-2015
20 For a detailed discussion of average default position and the mathematical derivation of accuracy ratio from the average default position, please refer to Moody’s Special Comment, Measuring Ratings Accuracy Using Average Default Position, Feb 2011.
13 14 15 16 17 18 19 20 21 0 5 10 15 20 25 30 35 40 45 50 55 60
Months prior to default
2015 1983-2015 Ba3 B1 B2 B3 Caa1 Caa2 Caa3 Ca C 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 1-Year 5-Year
Moody’s Related Research
Special Comments and Sector In-Depth Reports:
»
Annual Default Study 1920-2014, March 2015 (179348)
»
Default Research: Industry Credit Risk: Recent Trends for Global Non-Financial Corporations, January
2016 (1012379)
»
European Corporate Default and Recovery Rates, 1985–2015H1, December 2015 (1012344)
»
Leveraged Finance Interest - North American Edition, January 2016 (187109)
»
Glossary of Moody’s Ratings Performance Metrics, September 2011 (135451)
»
High Yield Interest – European Edition, January 2016 (187108)
»
High Yield Interest – Asian Edition, January 2016 (187181)
»
High Yield Interest – Latin American Edition, January 2016 (187100)
»
SGL Monitor: LSI Gains Point to Rising Default Risks as 2016 Gets Underway, January 2016 (187113)
»
Moody's B3 Negative and Lower Corporate Ratings List: Oil & Gas Sends List to Six-Year High, Fueling
Forecast for More Defaults, January 2016 (1012945)
»
Global Oil and Natural Gas Industry: Increased Supply and Concerns About Demand Growth Drive
Prices Yet Lower, January 2016 (1014345)
»
Growing Stress in Commodity Sectors Is a Credit Hazard for 2016, December 2015 (1009103)
»
US Corporate Defaults & Recoveries - Oil and Gas: The Bad, Ugly and Good, May 2015 (1004376)
»
Corporate Default and Recoveries - US: What May Happen in the Next Default Cycle Given Falling
Credit Quality, August 2015 (1004664)
»
Moody's Ultimate Recovery Database: Lessons from 1,000 Corporate Defaults, November 2011
(137405)
»
Introducing Moody’s Credit Transition Model, August 2007 (104290)
»
Measuring Ratings Accuracy Using Average Default Position, February 2011 (129451)
2016 Outlooks:
»
Oil and Gas Industry – Global: 2016 Outlook – All Regions and Sectors Facing Lower-for-Longer
Environment (Presentation), December 2015 (185710)
»
Non-Financial Corporates – North America: 2016 Outlook – Outlook Stable Overall, But Investors
Becoming More Cautious (Presentation), December 2015 (185690)
»
Non-Financial Corporates - EMEA: 2016 Outlook - Stable Despite Slower Growth Prospects
(Presentation), December 2015 (174787)
»
Non-Financial Corporates – Asia (ex-Japan) - 2016 Outlook - Lower Growth and High Leverage Cast
Shadow over Corporates (Presentation), November 2015 (186277)
»
Banks - Global: 2016 Outlook - Macro Challenges, Regulation Will Offset Improved Fundamentals
(Presentation), December 2015 (186447)
»
Global Macro Outlook 2015-17: Lacklustre Global Economic Recovery Through 2017 Diminishes
Resilience to Shocks, November 2015 (1009471)
Methodology and Data Sources
Moody’s Definition of Default
Moody’s definition of default is applicable only to debt or debt-like obligations (e.g., swap agreements). For
details, please refer to
Moody’s Rating Symbols and Definitions.
Methodology
The methodology used in this study can be found in the
Glossary of Moody’s Ratings Performance Metrics.
The Glossary report is a technical paper that explains how Moody’s calculates default rates, transition rates
and rating performance metrics in detail.
Changes in this Year’s Report
Moody’s occasionally discovers historical defaults, leading to minor revisions of the historical data. As
always, the data contained in the most recently published Moody’s default study supersede the data
published in previous reports. This year, Moody’s revised its senior ratings algorithm and modified the rule
that determines when an issuer can return to a cohort post default.
Appendix: Description of New Senior Ratings Algorithm and New Default Treatment
The default and ratings performance statistics presented in this study are based on entity-level senior rating
histories produced by Moody’s Senior Ratings Algorithm (SRA).
21The SRA takes a Moody’s-rated entity’s
actual senior unsecured rating (or senior rating) history when one exists and estimates such when the entity
does not have any rated senior unsecured debt in all or part of its rating history. These estimated ratings
allow Moody’s to meaningfully compare credit quality across entities, regardless of their capital structures.
The process of estimating an entity’s senior rating has three broad steps. In the first step, notching rules are
created based on the average notch difference in ratings between each class of debt and the senior
unsecured debt of an entity. In the second step, the entity’s reference credit—the debt class rating that has
the highest priority—is selected. This is accomplished by ranking each class of rating on the basis of its
ability to predict the senior rating; this ranking is referred to as the priority of the notching rule. In the third
step, the reference credit’s rating is adjusted by the number of notches based on its corresponding notching
rule to estimate the entity’s senior rating.
22We have redesigned the SRA so that notching rules are determined dynamically and are consistent with
Moody’s current rating practices.
23,24The redesigned algorithm allows dynamic changes in notching rules
and their priority, whereas in its previous iteration, the algorithm used static notching rules that were
updated intermittently. We have also expanded our universe of issuers to include banks that only have
deposit ratings. As a result, we now include deposit defaults into our default statistics as opposed to only
bond and loan defaults in prior years’ studies. Finally, we have modified the rules by which a defaulted entity
can reenter a cohort for the purposes of the default and ratings performance statistics.
25
21 Throughout this appendix, we take entity and entity-level to refer to obligor and obligor-level.
22 By entity-level rating, we mean ratings assigned to entities as a whole, as opposed to ratings assigned to specific debt instruments. 23 We adopted the new algorithm starting with Moody’s September Default Report (October 2015).
24 We described the previous iteration of the SRA in Moody’s Senior Ratings Algorithm & Estimated Senior Ratings (February 2009). An updated publication reflecting the redesigned SRA is forthcoming.
25 We relax the conditions under which defaulted entities whose resolution dates are unknown are reintroduced back into our study sample. Previously, we waited until an entity’s estimated senior rating was upgraded to B3 or higher. In some cases, this treatment had the unintended consequence of keeping low-rated defaulted entities out of the sample for prolonged periods of time. Our new rule incorporates a time element in addition to the B3 rating threshold; now, even if its
These updates to our approach have the following impact:
26»
A modest increase in cohort sizes over time
27» A modest increase in the number of investment grade default events and as a result some slight
increase in investment-grade default rates and small decrease in average default positions
28»
A slight decrease in cumulative default rates for speculative grade rating categories
» A noticeable rise in the proportion of Caa-C ratings
29and a commensurate fall in the proportion of
single B ratings for recent cohorts.
Our new approach, while slightly more complex, has a few distinct advantages over its predecessor. First
and foremost, the redesigned SRA estimates senior ratings assigned by Moody’s analysts more accurately
than before.
30Second, because the redesigned model is dynamic, it reflects evolving notching trends in real
time, rather than ex post. The chart below shows that barring a couple of instances (for example the period
between 2003 and 2005), the redesigned SRA more accurately estimates senior ratings compared to its
previous iteration. In addition, this new approach expands the coverage of entities and default events,
resulting in more robust default and ratings performance measurements. Finally, these gains do not come at
the expense of introducing artificial rating volatility into the estimated entity-level rating histories.
Rating Accuracy: Redesigned SRA versus Previous SRA
Rating Accuracy: Redesigned SRA versus Previous SRA
Below, we describe in greater detail the various steps of the redesigned SRA, and where applicable, we
highlight the major differences vis-à-vis our previous model.
Procedurally, the SRA performs the following six steps:
rating has not been upgraded to B3 or higher, a defaulted entity will be reinstated into the study sample after one week for distressed exchange events, one year for missed interest or principal payment events and five years for bankruptcy events.
26 We have recalculated the historical measures presented in this study, incorporating all the changes mentioned above. Entity ratings prior to April 26, 1982 are calculated using the previous SRA. Hence, historical measures prior to 1982 have not changed since last year’s study.
27 The redesigned SRA estimates entity-level ratings from a broader universe of rating classes (e.g. bank deposits and industrial revenue bonds), which has resulted in larger cohort sizes over time. In addition, the modified “cohort reentry rule” has also contributed to increased cohort sizes. The increase of cohort size is generally small with most of the additional companies in the investment grade universe.
28 In recent years, a few financial institutions (mostly European banks) have selectively defaulted on subordinated debt only while continuing to pay senior debt obligations. Investment grade default counts in the redesigned SRA have increased in large part due to our treatment of these selective defaults. The previous SRA would force the defaulted subordinated debt to be the reference debt, resulting in an artificially lowered estimated senior rating. This approach was intended to produce an estimated rating that indicated the default risk of the issuer, even though it may not have been the truest estimate of the entity’s senior rating. In the redesigned SRA, we make no such adjustment for selective defaults and consequently observe higher investment grade default counts.
29 This has also led to a noticeable drop in the Caa-C default rates.
30 In this context, rating accuracy refers to the ability of the model to correctly estimate a Moody’s assigned senior rating when using non-senior ratings as inputs.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 19 82 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 20 15 20 16 % o f Es tim ate d Se ni or R ati ng s Eq ua l to As sig ne d Se nio r R at in g Previous Redesigned Source: Moody's
1. Filter out ineligible credits
2. Compute aggregated credit group histories
3. Infer notching rules
4. Select the reference credit group
5. Smooth artificial changes in the estimated senior rating
6. Remove entities without debt ratings
1.
Filter out ineligible credits
We form a universe of credits from all of Moody’s public, monitored, global scale long-term ratings, with the
exceptions of structured finance ratings, short-term ratings, modeled ratings, defeased ratings and
externally-backed ratings.
31From this universe, we infer notching rules and select reference credits, as
described below.
2.
Compute aggregated credit group histories
Each credit from among this broad universe is characterized along the following dimensions:
»
Class of debt or entity-level rating (e.g., regular bond, bank loan, first mortgage bond, Issuer Rating,
Corporate Family Rating)
32» Seniority (e.g., senior unsecured, senior secured, subordinated)
» Backing status (not backed, internally backed, externally backed)
» Currency type (e.g., local currency, foreign currency)
The SRA aims to estimate the hypothetical senior unsecured, non-backed, local currency, regular bond
rating—in other words, the benchmark rating—for each entity. For entities that already have a benchmark
rating, no estimation is required.
The four dimensions listed above are the salient factors affecting a particular credit’s rating in relation to
other credits in an entity’s capital structure. We expect that for a given entity, credits that match along
these four dimensions (referred to as credit groups) are homogenous and share the same rating at a fixed
moment in time. In cases where ratings from the same credit group differ for a given entity, we calculate the
median rating at each point in time.
33In this way, we construct sanitized credit rating histories for each
entity and refer to them as aggregated credit group histories.
3.
Infer notching rules
In the next step of the SRA, we infer notching rules from each credit group to the benchmark rating. A
notching rule is an abstraction for the prevailing notch difference that exists between a given credit group
rating and the benchmark rating of the same entity as a function of time, credit group rating level, region
and sector.
34For each credit group and among the set of entities that have both the credit group rating and
31 Our previous SRA excluded certain types of debts that are now included in the redesigned SRA, such as industrial revenue bonds. 32 Moody’s Rating Symbols and Definitions (January 2016) provides detailed definitions for Issuer Ratings and Corporate Family Ratings.
33 If there are an even number of credits in the credit group, we select the worse rating among the two middlemost ratings. We refer to this as the Median-Worst algorithm.
34 Even though we refer to “notching rules”, in no way are we suggesting that Moody’s rating analysts rigidly follow these rules in practice. Our algorithm only seeks to determine the “average” notching observed within a particular region and sector.
the benchmark rating, we compute the most frequently observed notch difference as a function of time, the
credit group’s rating and the entity’s region and sector. If there is not sufficient consensus for a particular
rule, no rule can be formed.
35Rather than relying on static notching rules as the previous SRA did, the
redesigned SRA allows historical ratings to drive the formation of notching rules, which can evolve over
time.
4.
Select the reference credit group
Having inferred notching rules, we then select the reference credit group for each entity at all points in time.
The reference credit group is the one whose notching rule most reliably predicts the benchmark rating. In
other words, it is the credit group that has the highest priority. A notching rule’s priority is broadly based on
two factors:
» How consistent is the credit group’s notching from the benchmark rating?
» How targeted is the pool of entities from which the notching rule was formed?
The more consistent a credit group’s notching is, the higher priority it will be assigned. For two notching
rules that are equally consistent, the rule that is formed from a more targeted pool of entities (with respect
to the entities for which we are selecting the reference credit group) will be assigned a higher priority.
36We have revised how we choose the reference credit group for entities that selectively default on
subordinated debt, but continue to pay on senior debt obligations. The previous SRA chose the defaulted
subordinated debt as the reference credit group when the gap between the senior- and subordinated debt
ratings becomes wider than historical standards. This resulted in an artificially lowered rating, and was
intended to balance two competing considerations: (1) estimating entity-level ratings, assuming that default
risk is shared evenly across the entity’s capital structure, and (2) reflecting Moody’s ratings’ true
discriminatory power by referencing the subordinated debt rating. In the redesigned SRA, we make no such
adjustment for selective defaults.
37,38Once we have determined the reference credit group for a particular entity, we apply its notching rule to the
reference credit group’s rating to derive the estimated senior rating. In this way, we construct entity-level
estimated senior rating histories at all points in time.
5.
Smooth artificial changes in the estimated senior rating
It is possible that an entity’s estimated senior rating history created in the previous step contains artificial
rating changes—that is, rating changes that are unsupported by the entity’s underlying aggregated credit
group histories.
39Artificial rating changes can be introduced due to changes in the reference credit group or
changes in the notching rule. To remove these artificial rating changes, we apply a remedial smoothing
procedure, by which we mean the process of shifting the entity’s estimated senior rating history either prior
35 In order for a rule to be formed, we require two conditions be satisfied: (1) at least 50% of all entities must have the same notching, and (2) there must be at least 10 entities that have the same notching.
36 For ease of exposition, this discussion of priority has been simplified. In our forthcoming publication on the redesigned SRA, we will expand on this topic in more detail.
37 In the future, we may choose to publish our default studies at the debt-class level. For example, we may choose to study subordinated debt ratings for financial institutions and consider only defaults on subordinated debts. This approach would avoid the issue highlighted here because both ratings and defaults would be measured on the same class of debt.
38 The previous approach led to a more correct rating accuracy measurement at the expense of introducing inaccurate ratings while the new approach has the opposite effect. We plan to address this issue in the future by doing studies at the instrument level or by censoring non senior unsecured defaults when doing senior unsecured entity-level studies.
to or after an artificial rating change by the same magnitude as the intended artificial change. This has the
effect of eliminating the artificial rating change at the cost of distorting the entity’s rating level.
The redesigned SRA allows artificial changes in an entity’s estimated senior rating history to be smoothed
either back in time or into the future. If the priority of the current notching rule is higher than that of the
previous notching rule (with respect to a particular artificial rating change), the entity’s estimated senior
rating is smoothed back in time, meaning the previous rating is replaced with the current rating. Conversely,
if the priority of the current notching rule is lower than that of the previous notching rule, the entity’s
estimated senior rating is smoothed into the future, meaning the current rating is replaced with the previous
rating.
406.
Remove entities without debt ratings
After smoothing is completed, the final step in the redesigned SRA is to remove entities from the data
sample for the periods of time when they do not have rated debt obligations—namely, bonds, loans or bank
deposits. Gaps in an entity’s estimated senior rating history that are due to the entity not having debt
ratings are treated as rating withdrawals. For this study, we only include entities that have debt ratings in
our data sample because only for these entities do we track default events with a high degree of confidence.
As a result, our default statistics remain unbiased and accurately reflect the true credit risk observed in the
corporate universe.
41Data Sources
Moody’s bases the results of this study on its proprietary database of ratings and defaults for corporate
bond and loan issuers. Municipal and sub-sovereign debt issuers, structured finance securities, private
placements and issuers with only short-term debt ratings are excluded unless otherwise noted. In total,
Moody’s data covers the credit experiences of over 20,000 corporate issuers that sold long-term public debt
at some time between 1920 and 2015. As of January 1, 2015, over 6,000 corporate issuers held a Moody’s
long-term bond, loan, deposit or corporate family rating.
Moody’s database of corporate defaults covers more than 3,000 long-term bond and loan defaults by
issuers both rated and non-rated by Moody’s. Additional data sources, such as Barclay’s Fixed Income Index
data, supplemented Moody’s proprietary data in the construction of the aggregate dollar volume-weighted
default rates. Defaulted bond pricing data was derived from Bloomberg, Reuters, IDC and TRACE. The
majority of these market quotes represent an actual bid on the debt instrument, although no trade may
have occurred at that price. Over the 1982-2015 period, the dataset includes post-default prices for
approximately 5,000 defaulted instruments issued by over 1,700 defaulting corporations. Moody’s makes
the 1970-2015 credit rating, default and recovery rate data used in this study available through its Default
and Recovery Database (DRD).
40 The concept of smoothing is vast and complex, and cannot be covered adequately in this appendix. In our forthcoming publication on the redesigned SRA, we will discuss smoothing in greater detail.
41 Suppose we included entities that only were assigned entity-level ratings in our data sample. Our default statistics would then understate the true default risk because we would be including entities that do not even have the possibility of defaulting.
Guide to Data Tables and Charts
In this section, we briefly describe the interpretation of some of the Exhibits contained in this report. Exhibit
13 was derived by mapping Moody’s ratings to a linear scale, then taking the median values of the
numerically mapped ratings.
Exhibit 21 shows average senior unsecured recovery rates by letter rating and year prior to default. Each cell
in the table indicates the average recovery rate on senior unsecured bonds with a specific rating within T
years of default. For example, the 36.6% two-year B recovery rate reported in Exhibit 21 indicates the
average recovery rate on B- rated issues that default at some time within a two-year period, not recovery
rate for issuers rated B exactly two years before default. Together with issuer-weighted average cumulative
default rates, these multi-period recovery estimates are used to calculate cumulative expected credit loss
rates, as presented in Exhibit 22.
Exhibits 32 through 37 show issuer-weighted historical average default rates by rating category over various
investment horizons. These data were generated by averaging the multi-year default rates of cohorts
formed at monthly intervals. In addition to their being statements of historical fact, these data are also
useful proxies for expected default rates. For example, over a five-year period a portfolio of B-rated issuers
defaulted at a 22.2% average rate between 1983 and 2015 (see Exhibit 34). For an investor with a five-year
exposure to a B-rated debt obligation or counterparty, this estimate also happens to be the best estimate of
the expected risk of default for a B-rated exposure based on the available historical data, particularly over
long investment horizons.
Exhibit 40 shows average cumulative volume-weighted default rates by rating category. Whereas
issuer-based default rates weight each issuer equally, these data weight each issuer by the total volume of
defaulted debt; larger defaults receive relatively more weight. Average default rates based on debt volume
affected are less suitable estimates of expected default risk. One reason is that issuer default volumes vary
considerably over time. On average, a leveraged corporate issuer defaults on approximately $300 million of
bonds. However, that total has been as high as $30 billion (WorldCom). Issuer-based default rates receive
particular emphasis in the rating process because the expected likelihood of default of a debt issuer holding
a given rating is expected to be the same regardless of differences in the nominal sizes of the exposures.
Exhibit 41 shows the cumulative issuer-weighted historical default rates of cohorts formed between the
years 1970 and 2015 (January 1 of each year). These data are a subset of the data used to calculate the
issuer-weighted averages shown in Exhibits 32 through 34 (which, again, are based on cohorts formed at
monthly time intervals). The default rates in Exhibit 41 may be useful for scenario analysis. For example, if
one believed that future default rates for a given pool of issuers will behave as they did in, say, 1997, then
one can use the January 1, 1997 cohort cumulative default rates as proxies for expected default rates.
Data Tables and Charts
EXHIBIT 15
Moody’s-Rated 2015 Corporate Bond and Loan Defaults
1,2,3Company Country Default Type Month (US$ mil) Bonds (US$ mil) Loans In Jan 2015 cohort?
Affinion Group Holdings, Inc. United States Distressed Exchange November 585 yes
Agropecuaria Nossa Senhora do Carmo S.A. Brazil Missed Interest Payment February 735 yes
Allied Nevada Gold Corp. United States Prepackaged Chapter 11 March 400 75 yes
Alpha Bank AE Greece Bank Holiday June 1069 yes
Alpha Natural Resources, Inc United States Distressed Exchange April 596 0 yes
Alpha Natural Resources, Inc United States Chapter 11 August 2807 1056 no
Altegrity, Inc. United States Missed Interest Payment January 1420 275 yes
American Apparel, Inc. United States Prepackaged Chapter 11 October 214 25 yes
American Eagle Energy Corporation United States Missed Interest Payment March 175 yes
American Energy - Woodford, LLC United States Distressed Exchange June 340 yes
Anchor Hocking, LLC United States Prepackaged Chapter 11 April 249 yes
Attica Bank S.A. Greece Bank Holiday June 0 yes
Automotores Gildemeister S.A. Chile Missed Interest Payment December 400 yes
BANIF-Banco Internacional do Funchal, S.A. Portugal Placed Under Administration December 539 yes
Bank Finance and Credit, OJSC Ukraine Bankruptcy September 0 yes
BANK RSB 24 (JSC) Russia Placed Under Administration November 23 yes
Bank Uralsib Russia Distressed Exchange November 77 0 yes
Berau Coal Energy TBK (P.T.) Indonesia Bankruptcy July 950 8 yes
Black Elk Energy Offshore Operations, LLC United States Chapter 11 September 139 no
Caesars Entertainment Operating Company,
Inc. United States Prepackaged Chapter 11 January 16394 4091 yes
California Resources Corp. United States Distressed Exchange December 2813 0 yes
CFG Investment S.A.C. Peru Liquidated November 288 550 yes
Chassix Holdings, Inc. United States Prepackaged Chapter 11 March 2 no
Chassix Inc. United States Prepackaged Chapter 11 March 375 no
CHC Group Ltd. Canada Distressed Exchange August 41 yes
Chesapeake Energy Corporation United States Distressed Exchange December 3694 0 yes
China Fishery Group Limited Hong Kong Liquidated November 0 0 yes
Colt Defense LLC United States Chapter 11 June 249 70 yes
Comstock Resources, Inc. United States Distressed Exchange September 101 0 yes
Connacher Oil and Gas Limited Canada Missed Interest Payment March 831 146 yes
CORE Entertainment Inc. United States Missed Interest Payment July 160 yes
Delta Bank Belarus Seized By Regulators March yes
Dex Media, Inc. United States Missed Interest Payment October 270 yes
Doral Financial Corporation United States Chapter 11 March 170 yes
Drill Rigs Holdings Inc. Marshall Islands Distressed Exchange December 156 yes
EXHIBIT 15
Moody’s-Rated 2015 Corporate Bond and Loan Defaults
1,2,3Company Country Default Type Month (US$ mil) Bonds (US$ mil) Loans In Jan 2015 cohort?
Edcon Holdings Limited South Africa Distressed Exchange July 456 yes
Edmentum, Inc. United States Distressed Exchange June 361 yes
Education Management LLC United States Distressed Exchange January 217 1333 yes
ELO Touch Solutions, Inc. United States Distressed Exchange September 15 yes
Emeco Holdings Limited Australia Distressed Exchange December 0 yes
Emeco Pty Limited Australia Distressed Exchange December 52 0 yes
Empresas ICA, S.A.B. de C.V. Mexico Missed Interest Payment December 700 0 yes
Energy XXI Gulf Coast, Inc. United States Distressed Exchange September 428 0 yes
Essar Steel Algoma Inc. Canada Bankruptcy November 375 375 yes
Eurobank Ergasias S.A. Greece Bank Holiday June 1299 yes
EXCO Resources, Inc. United States Distressed Exchange October 577 0 yes
Ferrexpo Plc Switzerland Distressed Exchange February 500 yes
General Shopping Brasil S.A. Brazil Distressed Exchange October 86 yes
Getty Images, Inc. United States Distressed Exchange December 240 0 yes
Glorious Property Holdings Limited China Missed Principal And Interest Payments June 0 482 yes
Goodrich Petroleum Corporation United States Distressed Exchange September 213 0 yes
Great Atlantic & Pacific Tea Co., Inc. (The) United States Chapter 11 July 270 no
Halcon Resources Corporation United States Distressed Exchange April 1793 0 yes
Heckler & Koch GmbH Germany Distressed Exchange November 48 yes
Hercules Offshore, Inc. United States Prepackaged Chapter 11 August 1211 yes
Hidili Industry International Development
Ltd China Missed Principal And Interest Payments December 183 yes
Kaisa Group Holdings Ltd China Missed Interest Payment April 1292 0 yes
Liberty Tire Recycling Holdco, LLC United States Distressed Exchange March 225 yes
Lightstream Resources Ltd Canada Distressed Exchange July 465 0 yes
Linn Energy, LLC United States Distressed Exchange November 1999 0 yes
Logan's Roadhouse Inc. United States Distressed Exchange October 107 0 yes
Magnetation LLC United States Chapter 11 May 425 0 yes
Magnum Hunter Resources Corporation United States Chapter 11 December 600 397 yes
Maxcom Telecomunicaciones, S.A.B. de C.V. Mexico Distressed Exchange September 37 yes
Mega Energia Locacao e Admin. de Bens S.A. Brazil Missed Interest Payment January 27 yes
Metinvest B.V. Netherlands Missed Principal And Interest Payments March 114 113 yes
Midstates Petroleum Company Inc. United States Distressed Exchange May 630 yes
Millennium Health, LLC United States Prepackaged Chapter 11 November 1775 yes
MMM Holdings, LLC Puerto Rico Missed Interest Payment September 475 yes
Molycorp, Inc. United States Chapter 11 June 1405 50 yes
National Bank of Greece S.A. Greece Bank Holiday June 708 yes
Norske Skogindustrier ASA Norway Distressed Exchange February 774 yes
