Moody's Corporation · NYSE:MCO
Moody's — Moody's Analytics — Technical Presentation (April 2021)
39 slides from Moody's's technical presentation, April 2021
Slide 1: It's Not All About the Economy: Climate Change and U.K. and U.S. Mortgage Credit Risk
It’s Not All About the Economy: Climate Change and the U.K. and U.S. Mortgage Credit Risk Dr Petr Zemcik • Dr Pouyan Mashayekh • Cecilia Bocchio • Muhammad Jafree April 2021
Slide 2: Moody's Analytics independence disclaimer
Moody's Analytics operates independently of the credit ratings activities of Moody's Investors Service. We do not comment on credit ratings or potential rating changes, and no opinion or analysis you hear during this presentation can be assumed to reflect those of the ratings agency. Climate-Adjuste
Slide 3: Today's Speakers
Today’s Speakers Predictive Analytics Muhammad Jafree Petr Zemcik, PhD Pouyan Mashayekh, PhD Cecilia Bocchio Assistant Director Senior Director Senior Director Assistant Director Climate-Adjusted Credit Risk Metrics for Residential Mortgages 3
Slide 4: Portfolio Analyser (PA) Suite of Models
Portfolio Analyser (PA) Suite of Models Retail loan-level econometric models for credit and impairment metrics Delinquency/Arrears (Flow rates or migration matrices) R e t a i l As s e t C l a s s e s Default Rates & PDs Mortgages Overdrafts (Dynamic term-structures) (With relevant sub-categories) (
Slide 5: Agenda
Agenda 1. From climate to credit risk: Methodological Challenges 2. Impact of frequency and severity of natural disasters on the PD for U.S. mortgages 3. Simulations of natural disasters in credit models using the U.S. Mortgage Portfolio Analyzer 4. The effect of climate scenarios on the U.K. mortga
Slide 6: From climate to credit risk: Methodological challenges
1 From climate to credit risk: Methodological challenges
Slide 7: Climate Risk Adjustment for PD/LGD
Climate Risk Adjustment for PD/LGD Mortgages Model Factors Model Structure PD/LGD Panel logit model of the form Adjustment Mortgage Portfolio Mortgage Portfolio Characteristics Pr 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 Analyser 𝑁𝑁 𝑁𝑁𝑗𝑗 𝑁𝑁 f-1( PD 09-2020) = f-1(6.5%) α+∑ 𝑖𝑖 β𝑖𝑖 𝑥𝑥𝑖𝑖+∑ µ 𝑥𝑥…
Slide 8: Mortgage Portfolio Analyzer
Mortgage Portfolio Analyzer INPUTS MODELS OUTPUTS Forecasts, Account-level, Economic Data Default/Prepayment Staging & forward-looking, (Scenarios) Impairment scenario-conditional Calculations cash-flows LGD Scenario projections Loan Level Data Risk simulations, for risk metrics such cash-flow as EC
Slide 9: Moody's Analytics Climate Risk Scenarios
Moody’s Analytics Climate Risk Scenarios Climate Risk Variables Macro-financial Variables Physical Variables Macroeconomic Variables Global & regional temperature pathways GDP & unemployment Climate-related perils (e.g. flood, subsidence) Inflation & central bank rates Longevity Corporate profits &
Slide 10: 427 Data – Physical Risk Score
427 Data – Physical Risk Score On-Demand Scoring Key Features of 427 Data Facility Risk Score Score single assets or large portfolios via an interactive, browser-based application or API Analyze data and underlying climate indicators via multiple visualizations Score thousands of properties in minut
Slide 11: Impact of frequency and severity of natural disasters on the PD for U.S. mortgages
Impact of frequency and 2 severity of natural disasters on the PD for U.S. mortgages
Slide 12: Monthly default rate: Hurricane Harvey
Monthly default rate 0.00% 0.10% 0.20% 0.30% 0.40% 0.50% 0.60% 0.70% 0.80% 200402 200405 200408 200411 200502 200505 200508 200511 200602 200605 200608 200611 200702 200705 200708 200711 200802 200805 200808 Hurricane Harvey 200811 200902 200905 200908 200911 Mortgage default rate in Texas 201002 20
Slide 13: Monthly default rate: Katrina
Monthly default rate 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 200402 200408 200502 200508 200602 200608 Katrina 200702 200708 200802 200808 200902 200908 201002 MS 201008 201102 201108 201202 201208 201302 201308 201402 201408 201502 201508 201602 201608 201702 201708 201802 Mortgage default
Slide 14: Federal Emergency Management Agency (FEMA) data
Federal Emergency Management Agency (FEMA) data Close to 13,000 events for 24 different event types: Year, month, State, number of deaths & the estimated amount of damage to property incurred by the weather event Event Level Event Type 1 flood 2 hurricane_typhoon 3 wildfire 4 tornado_waterspout 5 ts
Slide 15: Top 9 natural disasters
Top 9 natural disasters 56 natural disasters with more than $500M estimated damage state yrm damage property deaths Disaster TX 201708 $ 43,704,268,000 67 Harvey NJ 201210 $ 20,950,000,000 2 Sandy PR 201709 $ 19,018,177,000 20 Irma CA 201811 $ 17,000,000,000 86 Wildfire LA 200508 $ 16,933,030,000 81
Slide 16: Multiplier added in the model
There are many events with significant impacts on PD Multipliers in FRM pd model 40 35 30 Multiplier added in the model 25 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Event # Climate-Adjusted Credit Risk Metr
Slide 17: Simulations of natural disasters using the U.S. Mortgage Portfolio Analyzer
Simulations of natural 3 disasters in credit models using the U.S. Mortgage Portfolio Analyzer
Slide 18: The distribution of damages in California
The distribution of damages in California Wildfire Flood, Hurricane, Tornado &Heavy Rain .15 .2 .15 .1 Density Density .1 .05 .05 0 0 5 10 15 20 25 5 10 15 20 logdamage logdamage Climate-Adjusted Credit Risk Metrics for Residential Mortgages 18
Slide 19: The distribution of damages in Wyoming
The distribution of damages in Wyoming Wildfire Flood, Hurricane, Tornado &Heavy Rain .25 .25 .2 .2 .15 .15 Density Density .1 .1 .05 .05 0 0 8 10 12 14 16 6 8 10 12 14 16 logdamage logdamage Climate-Adjusted Credit Risk Metrics for Residential Mortgages 19
Slide 20: Portfolio: 2688 loans, Total Balance = $803M
Portfolio: 2688 loans, Total Balance = $803M FICO # % Exposure Loan Size # % Exposure >= 790 537 19.90 $ 159,792,000 >= 368,500 181 8.48 $ 68,108,000 >= 750 331 12.53 $ 100,668,000 >= 352,000 327 14.66 $ 117,744,000 >= 710 360 13.24 $ 106,328,000 >= 335,500 247 10.53 $ 84,544,000 >= 670 356 13.31 $
Slide 21: Simulation results for loans in TX
Simulation results for loans in TX; Without natural disasters (above) and with natural disasters (below) Portfolio Loss Distribution Aggregate Statistics Num ber of Sim ulations 10,000 10.0000% Mean (Portfolio EL) 5.8959% 9.0000% Standard Deviation 2.6908 8.0000% Interquartile Range 3.3077 7.0000% S
Slide 22: Simulation results for loans in WY
Simulation results for loans in WY; Without natural disasters (above) and with natural disasters (below) Portfolio Loss Distribution Aggregate Statistics Num ber of Sim ulations 10,000 8.0000% Mean (Portfolio EL) 4.4758% 7.0000% Standard Deviation 2.0276 Interquartile Range 2.3861 6.0000% Skewnes s
Slide 23: Simulation results for loans in CA
Simulation results for loans in CA; Without natural disasters (above) and with natural disasters (below) Portfolio Loss Distribution Aggregate Statistics Num ber of Sim ulations 10,000 25.0000% Mean (Portfolio EL) 2.3528% Standard Deviation 2.0021 20.0000% Interquartile Range 1.7891 Skewnes s 3.8773
Slide 24: The effect of climate scenarios on U.K. mortgages credit parameters
The effect of climate 4 scenarios on the U.K. mortgages credit parameters
Slide 25: Mortgage Characteristics
Mortgage Characteristics Geographic region 18.0% December 2020 snapshot 15.0% Accounts Exposure Origination year 12.0% 2019 % 9.0% 2018 2017 6.0% 2016 » 126,396 mortgages totaling 2015 3.0% an outstanding balance of 2014 2013 0.0% £12,149m. 2012 2011 » There are 877 under- Originatuon Year 2010 perf
Slide 26: Climate Change Scenarios for the UK
% 0.0 2.0 4.0 6.0 8.0 -4.0 -2.0 2020Q1 2021Q3 2023Q1 2024Q3 HPI, % change yr ago 2026Q1 2027Q3 2029Q1 2030Q3 2032Q1 2033Q3 2035Q1 2036Q3 2038Q1 Current Policy 2039Q3 2041Q1 2042Q3 2044Q1 2045Q3 2047Q1 2048Q3 Early Policy 2050Q1 2051Q3 2053Q1 2054Q3 2056Q1 2057Q3 Late Policy 2059Q1 2060Q3 % 3.0 4.0 5
Slide 27: Impact of Climate Scenarios on Credit Risk
Impact of Climate Scenarios on Credit Risk Expected Losses, % Annualised Conditional PD, % 0.04 0.8 0.7 0.6 0.03 0.5 % 0.4 % 0.3 0.02 0.2 0.1 Current Policy Early Policy Late Policy 0.01 0 2021M1 2021M8 2022M3 2023M5 2024M7 2025M2 2025M9 2026M4 2027M6 2028M1 2028M8 2029M3 2030M5 2031M7 2032M2 2032M9
Slide 28: Usage of location-specific climate risk scores to forecast climate-adjusted credit risk metrics
Usage of location-specific 5 climate risk scores to forecast climate-adjusted credit risk metrics
Slide 29: Location-Specific Climate Risk
Location-Specific Climate Risk Modelling Methodology 427 Climate Risk Scores Obtain climate risk scores at the facility level. 427 returns six climate event risk scores based on the geographic location provided. Climate Scenarios Build country-specific climate scenarios (e.g., NGFS Current Policy, N
Slide 30: Physical Climate Risk Score Data: UK Sample
Physical Climate Risk Score 427 Data » 1,000 random names of the portfolio and ran the physical risk climate scores within the Floods Score Sea Level Rise Score Water Stress Score on-demand scoring application » Six climate hazards assessed at the UK Sample (1000 facilities) facility-level: floods,
Slide 31: Credit Risk Impact of Flooding Exposure
Credit Risk Impact of Flooding Exposure A Worked Example on a UK Residential Mortgage Assume 7 loans with the same characteristics (exposure, LTV, interest rate, DPD status, etc.) but located in different areas of the UK. Climate-Adjusted Credit Risk Metrics for Residential Mortgages 31
Slide 32: Linking Probability of Climate Events to Credit Risk
Linking Probability of Climate Events to Credit Risk Example for flood events Return period of inundation 1-in-5 years When will the event occur? 1-in-10 years » Impact on macroeconomic drivers (e.g., 427 Score 1-in-20 years 1-in-50 years Climate Event HPI). 1-in-75 years Simulation 1-in-100 years »
Slide 33: Linking Severity of Climate Events to Credit Risk
Linking Severity of Climate Events to Credit Risk Impact of flooding in HPI 1. Analyse 70+ flooding events in the UK. 2. Classify the events based on severity. 3. Analyse the impact on HPI after the event is observed by severity buckets. Fixed effect model 𝑦𝑦𝑖𝑖𝑖𝑖 = 𝛼𝛼 + 𝜆𝜆𝑡𝑡 + 𝑐𝑐𝑖𝑖…
Slide 34: Impact of Flooding in Location-specific HPI
Impact of Flooding in Location-specific HPI HPI forecast before a flood event 10.00% Baseline Early Policy 8.00% 6.00% 4.00% YoY 2.00% 0.00% -2.00% -4.00% -6.00% 2020Q1 2020Q3 2021Q1 2021Q3 2022Q1 2022Q3 2023Q1 2023Q3 2024Q1 2024Q3 2025Q1 2025Q3 2026Q1 2026Q3 2027Q1 2027Q3 2028Q1 2028Q3 2029Q1 2029Q
Slide 35: Impact of Flooding Events in LGD
Impact of Flooding Events in LGD Location-specific LGD Climate-Adjusted LGD by Severity of Flood Event, peak value 10.0 Baseline Early Policy 8.0 6.0 % 4.0 2.0 0.0 Severity 1 Severity 2 Severity 3 Severity 4 Severity 5 Severity 6 Severity 7 Less severe More severe Climate-Adjusted Credit Risk Metric
Slide 36: Climate-Adjusted Credit Risk Approach: Summary
Climate-Adjusted Credit Risk Approach: Summary Scenario Loss Distribution Standard Model Portfolio NO Frequency Dummy HPI(S), UR(S) YES Severity PDs (S), LGDs (S) PDs, LGDs Climate-Adjusted Credit Risk Metrics for Residential Mortgages 36
Slide 37: Q&A
Q&A Additional questions? Email us at help@economy.com Additional Analysis, Whitepapers, Case Studies and Solution Information: www.economy.com/products/consumer-credit-analytics/portfolio-analyzer Climate-Adjusted Credit Risk Metrics for Residential Mortgages 37
Slide 38: Contact Us: moodysanalytics.com
Contact Us: Content Solutions - Economics & Business Analytics Offices West Chester, EBA-HQ London Prague Singapore +1.610.235.5299 +44.20.7772.5454 +420 (234) 747-532 +65.6511.4400 121 North Walnut Street, Suite 500 One Canada Square / Canary Wharf Pernerova 691/42 6 Shenton Way West Chester PA 193
Slide 39: Climate-Adjusted Credit Risk Metrics for Residential Mortgages
© 2021 Moody’s Corporation, Moody’s Investors Service, Inc., Moody’s Analytics, Inc. and/or their licensors and affiliates (collectively, “MOODY’S”). All To the extent permitted by law, MOODY’S and its directors, officers, employees, agents, representatives, licensors and suppliers disclaim liabilit
Moody's Analytics — Technical Presentation (April 2021) on 6ix