Credit Rating Adjustments Prior to Default and Recovery Rates S. B. - - PowerPoint PPT Presentation

▶
credit rating adjustments prior to default and recovery
SMART_READER_LITE
LIVE PREVIEW

Credit Rating Adjustments Prior to Default and Recovery Rates S. B. - - PowerPoint PPT Presentation

Aug 23, 2023 40 likes •319 views

Introduction Hypotheses Models and Results Conclusion Credit Rating Adjustments Prior to Default and Recovery Rates S. B. Bonsall, IV a K. Koharki b K. Muller, III c A. Sikochi d a The Ohio State University b Washington University at St. Louis

slide-1
SLIDE 1

Introduction Hypotheses Models and Results Conclusion

Credit Rating Adjustments Prior to Default and Recovery Rates

  • S. B. Bonsall, IVa
  • K. Koharkib
  • K. Muller, IIIc
  • A. Sikochid

aThe Ohio State University bWashington University at St. Louis cThe Pennsylvania State University dHarvard University

Illinois Young Scholars Symposium, April 1, 2017

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 1 / 25

slide-2
SLIDE 2

Introduction Hypotheses Models and Results Conclusion

Reputational concerns and rating agencies

Critics: issuer-pay model cre- ates conflicts of interest Rating agencies: Reputation mitigates conflicts of interest

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 2 / 25

slide-3
SLIDE 3

Introduction Hypotheses Models and Results Conclusion

Reputational concerns and rating agencies

Critics: issuer-pay model cre- ates conflicts of interest Rating agencies: Reputation mitigates conflicts of interest

1

Do reputational concerns lead credit rating agencies to make credit rating adjustments to reduce rating optimism prior to issuer default?

2

Are credit rating adjustments prior to issuer default informative of lender recovery rates at default?

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 2 / 25

slide-4
SLIDE 4

Introduction Hypotheses Models and Results Conclusion

Credit ratings and ratings adjustments

”quantification [model-based rating] is integral to Moody’s rating analysis [...] However, Moody’s ratings [...] are a product of comprehensive analysis of each individual issue and issuer by experienced, well-informed, impartial credit analysts [subjective adjustments]” (Moody’s Investors Service, 2016) Subjective Adjustments = Actual - Model based Rating. (dubbed ’Rating Optimism’).

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 3 / 25

slide-5
SLIDE 5

Introduction Hypotheses Models and Results Conclusion

Summary of findings

1

For defaulting firms, credit rating adjustments are conservatively assigned by rating agencies

2

Rating adjustments are useful for assessing lender recovery rates

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 4 / 25

slide-6
SLIDE 6

Introduction Hypotheses Models and Results Conclusion

Summary of findings

1

For defaulting firms, credit rating adjustments are conservatively assigned by rating agencies

2

Rating adjustments are useful for assessing lender recovery rates

3

Increased competition leads to rating adjustments that are relatively more optimistic and less accurate of recovery rates

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 4 / 25

slide-7
SLIDE 7

Introduction Hypotheses Models and Results Conclusion

Contribution to existing literature

1

While prior research suggests that rating adjustments are used opportunistically, we highlight that they are also used defensively for issuers approaching default

2

New evidence that subjective rating adjustments to model-based ratings are informative about recovery rates at default

3

Unique setting to highlight instances where rating agencies’ reputational concerns may be greatest

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 5 / 25

slide-8
SLIDE 8

Introduction Hypotheses Models and Results Conclusion

H1: Ratings adjustments and issuer default

Issuer Default Rating Optimism Negative relation (-)

Increased reputational costs from overrating an issuer prior to default Increased ability for investors to ex post assess the bias

  • f credit ratings

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 6 / 25

slide-9
SLIDE 9

Introduction Hypotheses Models and Results Conclusion

H2: Ratings adjustments and recovery rates

Rating Optimism Recovery Rates Positive relation (+)

In the context of default, investors will judge whether assigned credit ratings provide information to predict loan recovery rates. Reputation costs will be higher for rating agencies with less accurate ratings of default recoveries.

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 7 / 25

slide-10
SLIDE 10

Introduction Hypotheses Models and Results Conclusion

Tests for the consequences of competition

Competition Reputational Concerns Negative relation (-)

Credit rating agencies’ adjustments to reduce rating

  • ptimism are lower when rating competition is higher

The predictive ability of credit rating adjustments for recovery rates is lower when rating competition is higher

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 8 / 25

slide-11
SLIDE 11

Introduction Hypotheses Models and Results Conclusion

Hypotheses not obvious for a few reasons

1

Arguably, the credit rating agencies’ most important task is to accurately assess firms’ default risk such that reputational concerns may outweigh conflicts of interests

  • r competition.

2

Greater competition among credit rating agencies can lead incumbents to increase the quality of their assigned ratings (Xia, 2014).

3

Any of these or other explanations could prevent us from finding results consistent with our hypotheses.

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 9 / 25

slide-12
SLIDE 12

Introduction Hypotheses Models and Results Conclusion

Data sources

Sample period covers 1992 - 2015 Moody’s ratings, default dates, price at default, default characteristics from Moody’s Default and Recovery Database (DRD) and www.moodys.com Fitch market share obtained using Fitch ratings in Mergent FISD Firm characteristics obtained from Compustat

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 10 / 25

slide-13
SLIDE 13

Introduction Hypotheses Models and Results Conclusion

Rating Optimism from rating adjustments

Predicted (model-based) rating, cross-sectionally by year

Ratingit = α0 + α1IntCovit + α2Profitit + α3Book Levit + α4Sizeit +α5Debt/EBITDAit + α6NegDebt/EBITDAit + α7EarnVolit +α8Cash/Assetsit + α9ConvDe/Assetsit + α10Rent/Assetsit +α11PPE/Assetsit + α12CAPEX/Assetsit +

j δjIndustryj + uit

Rating Optimism = Actual Rating - Predicted Rating Optimism takes on + (-) values when actual ratings are higher (lower) than predicted ratings Rating = Moody’s historical issuer rating mapped to natural numbers (i.e., C = 1, ..., AA+ = 20, AAA = 21)

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 11 / 25

slide-14
SLIDE 14

Introduction Hypotheses Models and Results Conclusion

Descriptive statistics for rating model

Variable Mean

  • Std. Dev.

Q1 Median Q3 Rating 11.486 4.051 8.000 12.000 15.000 IntCov 10.044 33.188 2.607 4.907 9.282 Profit 0.186 0.645 0.101 0.172 0.286 Book Lev 0.393 0.244 0.237 0.346 0.492 Size 8.287 1.556 7.172 8.216 9.368 Debt/EBITDA 3.724 6.236 1.599 2.894 4.805 NegDebt/EBITDA 0.034 0.182 0.000 0.000 0.000 EarnVol 0.123 1.660 0.013 0.024 0.044 Cash/Assets 0.074 0.094 0.012 0.039 0.099 ConvDe/Assets 0.012 0.044 0.000 0.000 0.000 Rent/Assets 0.016 0.028 0.002 0.008 0.016 PPE/Assets 0.382 0.271 0.342 0.342 0.619 CAPEX/Assets 0.059 0.060 0.022 0.044 0.076

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 12 / 25

slide-15
SLIDE 15

Introduction Hypotheses Models and Results Conclusion

Rating model estimation

(DepVar: = Rating) Coeff. t-stat IntCov 0.0061*** (4.05) Profit 0.0136 (0.51) Book Lev

  • 3.0014***

(-17.53) Size 0.7040*** (23.77) Debt/EBITDA

  • 0.0789***

(-13.66) NegDebt/EBITDA

  • 3.3831***

(-17.66) EarnVol

  • 0.0141

(-1.95) Cash/Assets

  • 1.3708***

(-4.46) ConvDe/Assets

  • 1.5357***

(-3.41) Rent/Assets

  • 5.9437***

(-5.17) PPE/Assets 0.9197*** (4.20) CAPEX/Assets 1.7491*** (2.88) Industry Fixed Effects Yes Observations 26,758 Pseudo R-Squared 0.153

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 13 / 25

slide-16
SLIDE 16

Introduction Hypotheses Models and Results Conclusion

Descriptive statistics for analyses variables

Variable Mean

  • Std. Dev.

Q1 Median Q3 DefaultPrice 41.040 27.027 21.250 34.750 61.880 Optimism

  • 2.510

3.688

  • 4.000
  • 3.000

0.000 PredictedRating 8.472 4.478 5.000 8.000 1.000 Coupon 8.879 2.800 7.400 9.062 10.750 SeniorSecured 0.073 0.260 0.000 0.000 0.000 Subordinated 0.066 0.248 0.000 0.000 0.000 DistressedExchange 0.210 0.407 0.000 0.000 0.000 Chapter11 0.493 0.500 0.000 0.000 1.000 Equity 0.189 0.220 0.059 0.104 0.273 DefaultBarrier 0.351 0.261 0.237 0.299 0.386 LTDIssuance 0.822 0.233 0.748 0.895 0.982 Profitability 0.044 0.114

  • 0.011

0.068 0.113 Intangibility 0.108 0.178 0.000 0.009 0.162 Receivables 0.089 0.088 0.032 0.065 0.126 Log(TotalAssets) 8.080 1.621 7.246 7.796 9.107 Log(Employees) 2.196 1.732 1.035 2.563 3.025

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 14 / 25

slide-17
SLIDE 17

Introduction Hypotheses Models and Results Conclusion

Modeling rating optimism as default nears

OLS Regression, three-month period prior to default

Optimismit = φ1Defaultt−3mo + φ2Defaultt−6mo + φ3Defaultt−9mo +φ4Defaultt−12mo + φ5Defaultt−15mo + φ6Defaultt−18mo +φ7Defaultt−21mo + φ8Defaultt−24mo + ςit Rating Optimism = Actual Rating - Predicted Rating Test within firms (default firms only) and across firms (matched sample firms) We test this over the one (1) and two (2) years prior to default: (1): φ1 − φ4 < 0 & (2) : φ1 − φ8 < 0

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 15 / 25

slide-18
SLIDE 18

Introduction Hypotheses Models and Results Conclusion

Rating optimism decreases prior to default

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 16 / 25

slide-19
SLIDE 19

Introduction Hypotheses Models and Results Conclusion

Rating optimism decreases prior to default

Default only sample (DepVar = Rating Optimism) Coeff. t-stat Defaultt−3mo

  • 1.669***

(-10.15) Defaultt−6mo

  • 1.331***

(-8.14) Defaultt−9mo

  • 1.086***

(-6.71) Defaultt−12mo

  • 1.022***

(-6.29) Defaultt−15mo

  • 1.002***

(-5.98) Defaultt−18mo

  • 0.934***

(-5.39) Defaultt−21mo

  • 0.828***

(-4.93) Defaultt−24mo

  • 0.826***

(-4.92) F-test: Defaultt−3 = Defaultt−12 39.99*** Defaultt−3 = Defaultt−24 44.56***

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 17 / 25

slide-20
SLIDE 20

Introduction Hypotheses Models and Results Conclusion

Rating optimism decreases prior to default

Matched sample (DepVar = Rating Optimism) Default NonDefault Difference Coeff. t-stat Coeff. t-stat Diff. t-stat Defaultt−3mo

  • 1.669

(-10.15)

  • 0.102

(-0.58)

  • 1.566***

(-6.60) Defaultt−6mo

  • 1.331

(-8.14)

  • 0.145

(-0.81)

  • 1.186***

(-4.99) Defaultt−9mo

  • 1.086

(-6.71)

  • 0.124

(-0.71)

  • 0.962***

(-4.10) Defaultt−12mo

  • 1.022

(-6.29)

  • 0.067

(-0.41)

  • 0.955***

(-4.20) Defaultt−15mo

  • 1.002

(-5.98)

  • 0.110

(-0.67)

  • 0.893***

(-3.87) Defaultt−18mo

  • 0.934

(-5.39)

  • 0.140

(-0.88)

  • 0.793***

(-3.40) Defaultt−21mo

  • 0.828

(-4.93)

  • 0.136

(-0.87)

  • 0.692***

(-3.05) Defaultt−24mo

  • 0.826

(-4.92)

  • 0.160

(-1.00)

  • 0.666***

(-2.90) F-test: Defaultt−3 = Defaultt−12 39.99*** 0.17 20.79*** Defaultt−3 = Defaultt−24 44.56*** 0.27 29.11***

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 18 / 25

slide-21
SLIDE 21

Introduction Hypotheses Models and Results Conclusion

Competition on adjustments prior to default

(DepVar = Rating Optimism) Coeff. t-stat Defaultt−3mo

  • 1.612***

(-5.88) Defaultt−6mo

  • 1.281***

(-4.72) Defaultt−9mo

  • 0.875***

(-3.40) Defaultt−12mo

  • 0.839***

(-3.17) Defaultt−15mo

  • 0.878***

(-3.44) Defaultt−18mo

  • 0.741***

(-2.77) Defaultt−21mo

  • 0.530***

(-2.04) Defaultt−24mo

  • 0.504**

(-1.99) FitchMktSh ×Defaultt−3mo 3.039*** (6.22) FitchMktSh ×Defaultt−6mo 2.259*** (3.16) FitchMktSh ×Defaultt−9mo 2.030*** (3.81) FitchMktSh ×Defaultt−12mo 2.111*** (3.12) FitchMktSh ×Defaultt−15mo 1.945*** (3.45) FitchMktSh ×Defaultt−18mo 1.861*** (3.84) FitchMktSh ×Defaultt−21mo 1.646*** (3.34) FitchMktSh ×Defaultt−24mo 2.916*** (5.85) Observations 3,624 Adjusted R2 0.118 F-tests: Defaultt−3mo = Defaultt−24mo 15.34*** Defaultt−3mo = Defaultt−12mo 20.35***

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 19 / 25

slide-22
SLIDE 22

Introduction Hypotheses Models and Results Conclusion

Modeling recovery rates

OLS Regression, rating adjustments and default recoveries

Recoveryit = δ0 + δ1PredictedRating + δ2Optimismit + λjControlj,it + εit Recovery = Default price; price of debt, as % of par, at date for distressed exchanges, or 30 days after all other default types. Controls: default (e.g., default type), debt (e.g., coupon rate, security), and firm characteristics (e.g., firm size, profitability) Ratings and adjustments predict recovery (1), and model based vs rating adjustments (2): (1): δ1 > 0, δ2 > 0 & (2) : δ2 < δ1

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 20 / 25

slide-23
SLIDE 23

Introduction Hypotheses Models and Results Conclusion

Default recoveries and pre-default ratings

(1) (2) Rating at Yeart−1 Rating at Yeart−2 coeff. t-stat coeff. t-stat PredictedRating 2.664*** (3.57) 2.764*** (3.61) Optimism 2.058*** (3.51) 1.747*** (3.15) Control variables Included Included Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Observations 1,126 1,091 Adjusted R-Squared 0.537 0.515 F-test: PredictedRating = Optimism 0.80 2.00

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 21 / 25

slide-24
SLIDE 24

Introduction Hypotheses Models and Results Conclusion

Competition, recoveries and ratings

(1) (2) Rating at Yeart−1 Rating at Yeart−1 coeff. t-stat coeff. t-stat PredictedRating 2.664*** (3.57) 2.865*** (4.34) Optimism 2.058*** (3.51) 2.765*** (4.06) FitchMktSh 11.427 (0.54) Optimism×FitchMktSh

  • 17.609**

(-3.90) Control variables Included Included Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Observations 1,126 1,126 Adjusted R-Squared 0.537 0.542

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 22 / 25

slide-25
SLIDE 25

Introduction Hypotheses Models and Results Conclusion

Controlling for accounting conservatism

(1) (2) Rating at Yeart−1 Rating at Yeart−1 coeff. t-stat coeff. t-stat PredictedRating 2.671*** (3.59) 2.846*** (4.36) Optimism 2.059*** (3.47) 2.699*** (3.97) FitchMktSh 12.113 (0.56) Optimism×FitchMktSh

  • 18.622**

(-4.11) Conservatism 2.131 (1.39) 1.981 (1.28) Control variables Included Included Year Fixed Effects Yes Yes Industry Fixed Effects Yes Yes Observations 1,125 1,125 Adjusted R-Squared 0.539 0.543

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 23 / 25

slide-26
SLIDE 26

Introduction Hypotheses Models and Results Conclusion

Concluding remarks

The subjectivity inherent in soft adjustments can allow credit rating agencies to become more conservative in their credit rating assessments as default approaches. However, the stringency in ratings near default has a reduced impact on creditor recovery rates when rating agency competition increases. Given this, market participants should use caution when using ratings as a barometer for default recoveries in their analysis.

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 24 / 25

slide-27
SLIDE 27

Introduction Hypotheses Models and Results Conclusion

Thank You!

Anywhere Sikochi (Siko) (HBS) Rating Adjustments and Recoveries April 1, 2017 25 / 25