疫情后不確定經濟時期的抵押貸款信用風險轉移（CRT）.pdf

《疫情后不確定經濟時期的抵押貸款信用風險轉移（CRT）.pdf》由會員分享，可在線閱讀，更多相關《疫情后不確定經濟時期的抵押貸款信用風險轉移（CRT）.pdf（51頁珍藏版）》請在三個皮匠報告上搜索。

1、Michael C.Schmitz,FCAS,MAAA-Principal and Consulting Actuary,MillimanAlan Tiernan,FIA,FCAS,FSAI-Chief Actuary,Global Mortgage Group Arch Capital Group Ltd.Jonathan Berenbom,FCAS,MAAA-Managing Director,Guy Carpenter2022 CAS Seminar on Reinsurance June 6,2023These slides are for general information/ed

2、ucational purposes only and shall not be considered as specific advice.As such,no action or decision should be taken solely on the basis of the information set out herein without obtaining specific advice from a qualified advisor.Copyright Milliman,Inc.2023.All rights reserved.Mortgage Credit Risk T

3、ransfer(CRT)while emerging from a PandemicMillimanPrivate Mortgage Insurance(PMI)the basicsTracking a changing economic environmentGSE loan performance updateLong-term view of mortgage credit riskReinsurer PML in uncertain times234Poll Key=QSXQJLabel=Mortgage/CRT MarketBasics of Private Mortgage Ins

4、urance(PMI)Front-end risk transferCovers lender/investor for financial loss if borrower defaultsGenerally required if(loan 80%x property value)What makes it different:Long term contracts with pricing set up frontReserves only cover delinquent loans claims vs curesBut ultimate losses have substantial

5、 tail riskMismatch between revenue and expenseTherefore contingency reserves and capital requirementsReserving:short-ish tail Pricing/PDR/Capital Adequacy:long tail!Q/S and XOL but function similarly for PMIERs capital reliefCRT shares these risk characteristics5Milliman M-PIReEnd-to-end CRT and mor

6、tgage reinsurance analytics software solutionKey Product Features1.Contains all required data for complete and thorough CRT analysis including reinsurance and private transactions:Eliminates the requirement for time-consuming data collection,processing,and report generation2.100%Cloud-based product

7、with all required integrated modules for CRT analysis:No need to purchase multiple licenses from multiple providers and try to stich the pitches together3.Real-time updates to platform for all users:All users received continuous updates to the platform;clients do not need to download or update softw

8、are 4.Stochastic portfolio analysis can be performed within minutes for a new portfolio:M-PIRe was built by a development team with expertise in mortgage analysis and high-frequency trading.5.Coverage for multiple exposure types:GSE CRT,MI Reinsurance,MI ILN,Multifamily,and non-US risk transfer deal

9、s7Moodys E Interest Rates:10-Year Treasury Constant Maturities,(%p.a.,NSA)Economics post-pandemic:10-Year Treasury80.000.501.001.502.002.503.003.504.004.502018Q12018Q22018Q32018Q42019Q12019Q22019Q32019Q42020Q12020Q22020Q32020Q42021Q12021Q22021Q32021Q42022Q12022Q22022Q32022Q42023Q12023Q22023Q32023Q42

10、024Q12024Q22024Q32024Q42025Q12025Q22025Q32025Q42026Q12026Q22026Q32026Q42027Q12027Q22027Q32027Q42028Q12028Q22028Q32028Q410-year Treasury RateMoodys E10-Year Treasury RateForecastActuals To-DateBaseline-February 2022S3 Modest Recession-February 2022S4 Protracted Slump-February 2022910Poll Key=WMXFBLab

11、el=10 year TreasuryMoodys E Interest Rates:10-Year Treasury Constant Maturities,(%p.a.,NSA)Economics post-pandemic:10-Year Treasury110.000.501.001.502.002.503.003.504.004.502018Q12018Q22018Q32018Q42019Q12019Q22019Q32019Q42020Q12020Q22020Q32020Q42021Q12021Q22021Q32021Q42022Q12022Q22022Q32022Q42023Q12

12、023Q22023Q32023Q42024Q12024Q22024Q32024Q42025Q12025Q22025Q32025Q42026Q12026Q22026Q32026Q42027Q12027Q22027Q32027Q42028Q12028Q22028Q32028Q410-year Treasury RateMoodys E10-Year Treasury RateForecastBaseline-February 2022Baseline-May 2023S3 Modest Recession-February 2022S3 Modest Recession-May 2023S4 Pr

13、otracted Slump-February 2022S4 Protracted Slump-May 2023Moodys E Household Survey:Unemployment Rate,(%,SA)Economics post-pandemic:Unemployment Rate12024681012142018Q12018Q22018Q32018Q42019Q12019Q22019Q32019Q42020Q12020Q22020Q32020Q42021Q12021Q22021Q32021Q42022Q12022Q22022Q32022Q42023Q12023Q22023Q320

14、23Q42024Q12024Q22024Q32024Q42025Q12025Q22025Q32025Q42026Q12026Q22026Q32026Q42027Q12027Q22027Q32027Q42028Q12028Q22028Q32028Q4Unemployment RateMoodys EUnemployment RateNational ForecastActuals To-DateBaseline-February 2022S3 Modest Recession-February 2022S4 Protracted Slump-February 2022Moodys E House

15、hold Survey:Unemployment Rate,(%,SA)Economics post-pandemic:Unemployment Rate13024681012142018Q12018Q22018Q32018Q42019Q12019Q22019Q32019Q42020Q12020Q22020Q32020Q42021Q12021Q22021Q32021Q42022Q12022Q22022Q32022Q42023Q12023Q22023Q32023Q42024Q12024Q22024Q32024Q42025Q12025Q22025Q32025Q42026Q12026Q22026Q3

16、2026Q42027Q12027Q22027Q32027Q42028Q12028Q22028Q32028Q4Unemployment RateMoodys EUnemployment RateNational ForecastBaseline-February 2022Baseline-May 2023S3 Modest Recession-February 2022S3 Modest Recession-May 2023S4 Protracted Slump-February 2022S4 Protracted Slump-May 2023Economics post-pandemic:Ho

17、me Prices(FHFA)14Moodys E FHFA All Transactions Home Price Index,(Index 1980Q1=100,NSA)3003504004505005506006502018Q12018Q22018Q32018Q42019Q12019Q22019Q32019Q42020Q12020Q22020Q32020Q42021Q12021Q22021Q32021Q42022Q12022Q22022Q32022Q42023Q12023Q22023Q32023Q42024Q12024Q22024Q32024Q42025Q12025Q22025Q3202

18、5Q42026Q12026Q22026Q32026Q42027Q12027Q22027Q32027Q42028Q12028Q22028Q32028Q4HPIMoodys EFHFA All Transactions Home Price IndexNational ForecastActuals To-DateBaseline-February 2022S3 Modest Recession-February 2022S4 Protracted Slump-February 20221516Poll Key=PHYZKLabel=Cumulative home price changeEcon

19、omics post-pandemic:Home Prices(FHFA)17Moodys E FHFA All Transactions Home Price Index,(Index 1980Q1=100,NSA)3003504004505005506006502018Q12018Q22018Q32018Q42019Q12019Q22019Q32019Q42020Q12020Q22020Q32020Q42021Q12021Q22021Q32021Q42022Q12022Q22022Q32022Q42023Q12023Q22023Q32023Q42024Q12024Q22024Q32024Q

20、42025Q12025Q22025Q32025Q42026Q12026Q22026Q32026Q42027Q12027Q22027Q32027Q42028Q12028Q22028Q32028Q4HPIMoodys EFHFA All Transactions Home Price IndexNational ForecastBaseline-February 2022Baseline-May 2023S3 Modest Recession-February 2022S3 Modest Recession-May 2023S4 Protracted Slump-February 2022S4 P

21、rotracted Slump-May 2023Milliman M-PIRe-GSE Acquisitions by Loan Purpose-Data through March 2023Purchase vs Refi Market18Milliman M-PIRe-GSE Acquisitions by FICO-Data through March 2023FICO Trends19Milliman M-PIRe-Annualized Prepayment Rate-Select ACIS transactions,2019 through 2022-Data through Apr

22、il 2023Prepayment Speeds20Milliman M-PIRe Current and Historical Delinquencies-Select ACIS transactions,2019 through 2022-Data through April 2023Delinquency Rates21Milliman M-PIRe Cumulative Loss Rate-Select ACIS transactions,2019 through 2022-Data through April 2023Net Loss Rates22Mortgage Insuranc

23、e CyclesHousing Economics Managing the Supply and Demand CurveHistorical performance fundamentally impacted more by economics than creditHowever,credit shifts fundamentally linked by product changes as a results of economics:No or negative amortization featuresLimited termsDownpayment/CLTVProperty T

24、ypesAvoidance of“financing ones way into a home”Oil PatchFinancial Crisis24EconomicsHouse Prices vs.Personal IncomeSource:Milliman M-PIReMortgage credit risk reinsurance is structured and modeled quite differently than traditional P&C exposures.Therefore,the PML approach is often unique for mortgage

25、 reinsurance.1What considerations are there to ensure that the measure of risk is consistent across the enterpriseAligning magnitude of stress with firm-wide risk tolerancesAligning underlying modeling drivers to maintain consistency across firmE.g.,if using a deterministic scenario,is it internally

26、 consistent with other similar business lines(e.g.,unemployment rate shock)Should the PML decline over-time if there is a strong housing market or strong performance to-date?Should the PML increase over-time if there is an adverse housing market or sudden increase in defaults?2How will this be used

27、to make business decisions and/or impact financialsWill this figure be the constraint on capacity/new writings?Should historical cashflows be considered?Will this figure impact financial statements?Ultimate vs.Accident Year PMLMortgage has unique/prolonged loss developmentIs the business concerned w

28、ith summarizing lifetime losses in one figure or appreciating how losses develop over-timeMeasuring Gross or Net LossUnderstanding how the PML changes period to period and key drivers of changeConsiderations for Defining a Mortgage PMLSeparate from the business considerations/implications of the PML

29、,several different analytical techniques/choices can be defended when choosing a methodology.DeterministicDefining macroeconomic forecast inputsSimplicity/InterpretabilitySenior stakeholders can easily understand the inputs and it does not change over timeCalibrationGiven inputs are selected,it may

30、be difficult to understand what VaR this representsStochasticSimulating future mortgage performanceConsistencyIf other business lines use a 99th pctl PML,you can select the same for mortgageStabilityIf the distribution is conditioned on to-date performance,it may change month-to-monthPML Analytical

31、MethodsDesign of a PML ScenarioSingle ScenarioUsing a single scenario to set PMLSimplicity/InterpretabilitySenior stakeholders can easily understand a single scenarioMarried to one viewLayering-in conservatism and exploring results sensitivity becomes difficult without other benchmarks to consultMul

32、tiple ScenariosUsing an average of multiple scenarios or stochastic trials to set PMLFlexibilityMultiple stress dynamics can be averaged/weighted to account for different downside risksStabilityE.g.,TVaR vs.VaR;avg over part of the dist.272023 Arch Capital Group Ltd.All rights reserved.272023 Arch C

33、apital Group Ltd.All rights reserved.CAS Seminar on ReinsuranceGlobal Perspective:Mortgage Insurance in AustraliaAlan Tiernan,FIA,FCAS,FSAIChief Actuary,Global Mortgage GroupArch Capital Group Ltd.282023 Arch Capital Group Ltd.All rights reserved.28Discussion Topics1.Mortgage Insurance in Australia

34、vs.the U.S.How do the markets and products compare structurally?2.Historical Economic Trends and PerformanceDiscussion of how Australia has experienced a relatively“stress-free”history,even during the Global Financial Crisis(GFC).3.Current State and OutlookHow are the economic factors that affect ho

35、using trending now,and how does that compare to the U.S.?292023 Arch Capital Group Ltd.All rights reserved.29CharacteristicMortgage Interest RatesPredominately fixed-ratePredominately variable-rateFull Recourse LendingNoYesMI Requirement Required for conventional mortgages with 80%LTV(i.e.Loan is 80

36、%of property value)in order to be sold to the Government Sponsored Entities(GSEs),also known as Fannie Mae and Freddie Mac May be significantly transferred later via GSE Credit Risk Transfer(CRT)programs Not a formal requirement but market convention where LTV is 80%MI Market Size2022 Gross Earned P

37、remium of US$5.8B2022 Gross Written Premium of US$0.6BMI Coverage Level(%of loan amount insured)Approximately 25%coverage100%coverageMI PremiumPrimarily monthly premium businessPrimarily single premium businessMI RelationshipsLenders have relationships with multiple insurersLenders typically have a

38、direct relationship with one insurerMI Market players6 primary mortgage insurers(Arch,Essent,NMI,Enact,Radian,MGIC)4 main players,each having a relationship with one of the 4 major banks,which control 80%of the lending market.(CBA-Helia,Westpac-Arch,NAB-QBE,ANZ-own captive)Reinsurance ParticipationR

39、einsurance participation on GSE CRT program(ACIS and CIRT)and reinsurance(XOL and QS)of primary mortgage insurersReinsurance capacity(XOL and QS)provided to primary mortgage insurersCapital Markets ParticipationCapital markets participation on GSE CRT(STACR and CAS)and Mortgage insurance linked note

40、s(MILNs)from primary mortgage insurersMarket not developedThe Mortgage Insurance market in Australia exhibits key structural differences to the U.S.marketMarket Comparison302023 Arch Capital Group Ltd.All rights reserved.300246810121416Unemployment RateAusUSSources:US:US Bureau of Labor Statistics,A

41、ustralia:Australian Bureau of StatisticsCOVID peak:Aus 7.5%U.S.14.7%GFC peak:Aus 5.9%U.S.10.0%Historical LowAustralia historically has seen very low unemployment rate volatilityDuring the GFC,unemployment peaked at 5.9%(June 09)vs.10%(October 09)in the U.S.Australia did not experience a large econom

42、ic downturn during the GFC.However,the pace of economic growth slowed significantly and the unemployment rate rose amid heightened uncertainty.Australias economy was buoyed at the time by large resource exports to China(i.e.,mining boom).The COVID-19 peak in Australia was a 7.5%unemployment rate(Jul

43、y 20),approximately half of the U.S.peak of 14.7%(April 20).Unemployment in both markets is currently at historically low levels(3.5%).Limited Economic Stress312023 Arch Capital Group Ltd.All rights reserved.31Limited Economic StressPeak-to-trough national home price decline in Australia from GFC wa

44、s 8%(March 08 to March 09)vs.19%in US(March 07 to June 12).Both countries have seen a significant run-up in home prices following the pandemic.Although periods of smaller home price corrections have been more prevalent,this has not been historically correlated with elevated delinquency experience.Th

45、is may be because of the full recourse nature of Australian lending,which potentially mitigates against the likelihood of strategic borrower default.Sources:US:US Federal Housing Finance Agency All-Transactions House Price IndexAustralia:Australian Bureau of Statistics,CoreLogic Hedonic Home Value I

46、ndex(monthly)1.01.21.41.61.82.02.22.42.62.83.0Sep-03Jun-04Mar-05Dec-05Sep-06Jun-07Mar-08Dec-08Sep-09Jun-10Mar-11Dec-11Sep-12Jun-13Mar-14Dec-14Sep-15Jun-16Mar-17Dec-17Sep-18Jun-19Mar-20Dec-20Sep-21Jun-22Mar-23HPI(Rescaled to Sep-03 levels)AusUSHome price declines in Australia were significantly less

47、and more short-lived during the GFC,although periods of smaller home price corrections are historically more prevalentLimited Economic Stress322023 Arch Capital Group Ltd.All rights reserved.322010201120122013201420152016201720182019202020212022Aus 27.0%33.0%51.8%30.4%10.0%19.4%27.8%33.6%39.9%44.4%8

48、1.4%-2.7%-18.1%US173.2%183.2%177.2%54.4%61.6%38.0%26.4%16.1%9.4%11.1%38.0%7.0%-17.6%-25%0%25%50%75%100%125%150%175%200%MI Industry Calendar Year Net Loss RatiosAusUSSources:US:Company Filings and Public ReleasesAustralia:APRA Quarterly General Insurance Institution-Level StatisticsLower economic vol

49、atility and lack of significant stress events in the Australian market drives more stable MI industry performance over time Mortgage Insurance is a capital intensive“cat-like”line of business,so stable economic performance has supported the MI industry in Australia.Loss experience during GFC was sta

50、rkly different.There has been somewhat of an upward trend since 2014 in Australia due to late emergence of losses from regional mining towns following the end of the mining boom,but experience has been favorable on average.2021 and 2022 have seen significant favorable reserve releases in both market

51、s related to COVID delinquencies,as insurers quickly increased reserves in 2020.From an actuarial perspective,Australias lower historical volatility and lack of meaningful stress events significantly reduces the quantum of data available for use in predictive modeling.This is compounded by a lack of

52、 publicly available data,unlike in the U.S.(i.e.GSE data).Limited Economic Stress=More Stable Historical Performance332023 Arch Capital Group Ltd.All rights reserved.33Interest rates have increased significantly over the past year,consistent with U.S.Australian borrowers are potentially impacted mor

53、e adversely by interest rate increasesImpact of rising ratesMonthly Mortgage PaymentNeutralPrimarily a fixed rate marketNegativePrimarily a variable rate marketAffordabilityNegativeNegativeThere are signs of an uptick in delinquencies,but this is relative to an historically low levelThe following se

54、rve as mitigants against borrower stress from recent rate hikes:The Australia regulator(APRA)requires new borrowers to prove they can service their mortgages at an interest rate 300bps in excess of the initial rate.Households are generally ahead on their repayments(almost 40%of loans were two or mor

55、e years ahead as of Dec.2022).Historically low rate of unemployment.But affordability has certainly been impacted0.00.51.01.52.02.53.03.54.04.55.0Apr-14Oct-14Apr-15Oct-15Apr-16Oct-16Apr-17Oct-17Apr-18Oct-18Apr-19Oct-19Apr-20Oct-20Apr-21Oct-21Apr-22Oct-22Apr-23AUS 10yr Govt Bond(%)RBA Cash Rate Targe

56、t(%)Source:Reserve Bank of AustraliaSource:S&P Global Ratings Australian RMBS Prime SPIN Data0.0%0.5%1.0%1.5%2.0%Mar-96Jan-97Nov-97Sep-98Jul-99May-00Mar-01Jan-02Nov-02Sep-03Jul-04May-05Mar-06Jan-07Nov-07Sep-08Jul-09May-10Mar-11Jan-12Nov-12Sep-13Jul-14May-15Mar-16Jan-17Nov-17Sep-18Jul-19May-20Mar-21J

57、an-22Nov-22Delinquency Rates90+61-9031-60Recent Rate Rises Similar to U.S.,but Potentially Different Impact342023 Arch Capital Group Ltd.All rights reserved.34Reduced affordability due to rising rates has resulted in home price declines of 8.5%nationally since the peak in April 2022House Price Appre

58、ciation(HPA)Source:CoreLogic.Hedonic Home Value Index(monthly)86909498102106110114118BrisbaneCanberraPerthAdelaideNationalSydneyMelbourneHobartDarwinPeak to trough declines in dwelling valuesSource:CoreLogic.RegionCOVID trough to peak growthDecline from recent peakMonth of peakSydney27.7%-12.3%Jan-2

59、2Melbourne17.3%-9.1%Feb-22Brisbane42.7%-10.9%Jun-22Adelaide44.7%-2.4%Jul-22Perth25.9%-0.4%Jul-22Hobart37.7%-12.9%May-22Darwin31.1%-2.0%Aug-22Canberra38.3%-9.5%Jun-22Combined capitals25.5%-9.0%Apr-22Combined regionals41.6%-7.5%Jun-22National28.6%-8.5%Apr-22Home prices remain well above pre-COVID leve

60、ls.This home price correction has hit Australia sooner than the U.S.A 10-month streak of home price reductions was broken in March 2023 with national dwelling values rising by 0.6%,led by a 1.4%increase in Sydney.It remains to be seen if this represents a turning point.Factors favoring home prices i

61、n Australia are generally consistent with U.S.experience and include:Low supply:Active listings down 3%year-over-year and 26%below 5-year average.Low rental vacancy rates:National vacancy rate of 0.8%,which is a record low.Low unemployment.Renewed international in-migration since COVID.Home Price Co

62、rrection has Bitten Earlier in Australia352023 Arch Capital Group Ltd.All rights reserved.35Understanding key similarities and differences between Australian and U.S.Mortgage Insurance markets.The Australian economy has experienced significantly less volatility over time,which has supported favorabl

63、e and consistent Mortgage Insurance industry performance.But the lack of volatility presents actuarial challenges in that a true stress event is not present in historical data.The Australian economy is seeing similar trends to the U.S.currently:Interest rates have risen,which has the potential to im

64、pact borrowers ability to repay in Australia,as well as impacting affordability.Home price declines have already begun,and although it is too early to say if they are at an end,various economic factors are a definite counter force,consistent with the U.S.Summary of Takeaways362023 Arch Capital Group

65、 Ltd.All rights reserved.36For more information,please reach out to:Alan TiernanChief Actuary,Global Mortgage GroupArch Capital Group Ltd._+1 441 278 A business of Marsh McLennanMORTGAGE CREDIT RISK TRANSFER THE REINSURANCE MARKETJonathan Berenbom,FCAS,MAAAManaging Director,Guy CarpenterCARe Seminar

66、June 6,2023Discussion Topics Mortgage Credit Risk Transfer(CRT)Market Updates Private Mortgage Insurers(MIs)Government Sponsored Enterprises(GSEs)PML Considerations for Reinsurers38Industry Dynamics Continue to Lead to MI Growth39Even as NIW Declined Sharply,Moderating Toward Pre-COVID LevelsBecause

67、 Persistency Increased in Rising Rate EnvironmentMI Industry Saw Strong IIF GrowthIn Force Persistency30Y Mortgage RateQ1 MI IIF grew 8%YoY,even as NIW declined substantiallySharp rise in persistency supported growth,as prepays materially lower given lack of borrower rate incentiveHigher persistency

68、 leads to longer premium streams for MIs and their reinsurers for their in-force book$0B$200B$400B$600B$800B$1,000B$1,200B$1,400B$1,600BMar-18Jul-18Nov-18Mar-19Jul-19Nov-19Mar-20Jul-20Nov-20Mar-21Jul-21Nov-21Mar-22Jul-22Nov-22Mar-23MI Insurance In ForceArchEnactEssentMGICNMIRadian-80%-60%-40%-20%0%2

69、0%40%60%80%$0B$20B$40B$60B$80B$100B$120B$140B$160B$180B$200BMar-18Sep-18Mar-19Sep-19Mar-20Sep-20Mar-21Sep-21Mar-22Sep-22Mar-23YoY ChangeQuarterly MI Industry NIW2.5%3.0%3.5%4.0%4.5%5.0%5.5%6.0%6.5%7.0%50%55%60%65%70%75%80%85%90%95%100%Mar-18Aug-18Jan-19Jun-19Nov-19Apr-20Sep-20Feb-21Jul-21Dec-21May-2

70、2Oct-22Mar-2330Y Mortgage Rate.Book Persistency0%10%20%30%40%50%60%70%80%90%100%$0.0T$0.5T$1.0T$1.5T$2.0T$2.5T$3.0T$3.5T$4.0T$4.5T$5.0T20052006200720082009201020112012201320142015201620172018201920202021202220232024Purchase Loan ShareUS Mortgage OriginationsOrigination Projections Support Continued

71、Growth40MBAForecastPurchase Loan ShareThough originations to materially decline,purchase loan volumes expected to be largest in 16 years,except for 2020-2022MI penetration rates for purchase loans in mid 20%s vs.low/mid single-digits for refis purchase loan volumes drive NIWLow projected refis to su

72、pport persistencyRefinancePurchaseSource:Mortgage Bankers Association,MBA April 17 Mortgage Finance Forecast1 1 2 2 3 1 3 5 5 6 1$0B$2B$4B$6B$8B$10B201820192020202120222023MI CRT IssuanceMIs Expected to Continue Elevated Reinsurance CRT Issuance4124%17%29%23%74%76%83%71%77%26%0%20%40%60%80%100%20182

73、019202020212022Share of MI CRT Issuance$3BAllocation to Reinsurance Market Grew in 2022Projected 2023 Issuance Near 2022 LevelReinsuranceCapital MarketsRecord 2022 Issuance Impacting Available CapacityReinsurers GSE CRT portfolios have materially de-risked,as 43%of limit has runoff since program inc

74、eptionAll vintages 2020 and prior have virtually no expected losses in both severe stress scenariosNotes:Modeled results are based on M-PIRe MPM_2022.1 model as of May 2023,assuming an economic callVintageOriginalLimitRemainingLimit2007Replay-35%HPI2013$.1 B$0.0 B$0.0 B$0.0 B2014$.9 B$0.2 B$0.0 B$0.

75、0 B2015$3.8 B$0.5 B$0.0 B$0.0 B2016$4.5 B$0.6 B$0.0 B$0.0 B2017$5.3 B$0.4 B$0.0 B$0.0 B2018$5.1 B$1.2 B$0.0 B$0.0 B2019$4.8 B$1.8 B$0.0 B$0.0 B2020$6.4 B$2.6 B$0.0 B$0.0 B2021$10.4 B$9.3 B$1.4 B$2.6 B2022$14.0 B$13.9 B$3.6 B$6.1 B2023$2.3 B$2.3B$1.0 B$1.6 BTotal$57.6 B$32.8 B$6.0 B$10.3 B$0.0 B$2.0

76、B$4.0 B$6.0 B$8.0 B$10.0 B$12.0 B$14.0 B$16.0 B20132014201520162017201820192020202120222023Original ExposureRemaining ExposureLifetime Losses(2007 Replay)Lifetime Losses(-35%HPI)GSE Reinsurance Transactions42Mortgage PMLs Offer Unique Challenges Reinsurance contracts are long-duration,covering losse

77、s arising from delinquencies over 10-12 years In most economic environments,limits amortize more slowly than risk Most reinsurers measure their aggregate mortgage capacity by modeling their in-force portfolio Unlike property catastrophe,mortgage stress events occur over a multi-year period The risk

78、of mortgage default is constantly changing,in response to a variety of factors,including Home Price Appreciation/Depreciation since loan inception Macroeconomic outlook(Home prices,interest rates,unemployment,etc.)Ability of borrowers to afford monthly payments Availability of GSE forbearance progra

79、ms Reinsurers use a variety of modeling approaches Deterministic(i.e.2007 Replay scenario or home price decline of X%)or stochastic(i.e.99th%ile of loss distribution)Single scenario or a blend of multiple scenarios Losses over the treaty term or the maximum in any one year43Economic IndicatorMortgag

80、eImpactModeled LossImpactHousingPricesLoan-to-Value RatiosInterest RatesPrepay-mentsModeled LossMacro Environment Leading to Modeled Loss Increases44Inception of CRT(2013)2Q 2022Economic IndicatorMortgageImpactModeled LossImpactHousingPricesLoan-to-Value RatiosInterest RatesPrepay-mentsModeled Loss2

81、Q 2022-PresentPMLs Impacted by Modeled Loss Volatility After peaking in May 2022,home prices began to decline nationally for the first time since 2012 Mortgage rates increased from 3.5%in January 2022 to 5.0%in April and 6.9%in October In a volatile economic environment,economic forecasts changed si

82、gnificantly from one month to the next Stochastic models centered around an economic forecast experienced similar monthly volatility Increasing PMLs pose challenges for reinsurance underwriters Difficult to explain Long duration contracts limit ability to resize portfolio Additional capacity may no

83、longer be available Risk of further increases may limit willingness to deploy capacity Both deterministic and stochastic approaches need to contemplate how to appropriately reflect changes in economic outlook as well as past home price declines Should a deterministic scenario be tempered to reflect

84、home price declines that have already occurred?Should a change in the baseline economic outlook impact the extreme tail of the loss distribution?45Questions?4647DisclaimerGC Analytics DisclaimerThe data and analysis provided by Guy Carpenter herein or in connection herewith are provided“as is,”witho

85、ut warranty of any kind whether express or implied.The analysis is based upon data provided by the company or obtained from external sources,the accuracy of which has not been independently verified by Guy Carpenter.Neither Guy Carpenter,its affiliates,nor their officers,directors,agents,modelers,or

86、 subcontractors(collectively,“Providers”)guarantee or warrant the correctness,completeness,currentness,merchantability,or fitness for a particular purpose of such data and analysis.The data and analysis is intended to be used solely for the purpose of the companys internal evaluation and the company

87、 shall not disclose the analysis to any third party,except its reinsurers,auditors,rating agencies and regulators,without Guy Carpenters prior written consent.In the event that the company discloses the data and analysis,or any portion thereof,to any permissible third party,the company shall adopt t

88、he data and analysis as its own.In no event will any Provider be liable for loss of profits or any other indirect,special,incidental and/or consequential damage of any kind howsoever incurred or designated,arising from any use of the data and analysis provided herein or in connection herewith.Statem

89、ents or analysis concerning or incorporating tax,accounting or legal matters should be understood to be general observations or applications based solely on our experience as reinsurance brokers and risk consultants and may not be relied upon as tax,accounting or legal advice,which we are not author

90、ized to provide.All such matters should be reviewed with the clients own qualified advisors in these areas.There are many limitations on actuarial or modeling analyses,including uncertainty in the estimates and reliance on data.We will provide additional information regarding these limitations upon

91、request.As with any analysis,the results presented herein are subject to significant variability.While these estimates represent our best professional judgment,it is probable that the actual results will differ from those projected.The degree of such variability could be substantial and could be in

92、either direction from our estimates.Reliances and Limitations/Milliman M-PIRe GuidelinesThese slides are for general informational purposes only.Action should not be taken solely on the basis of the information set out herein without taking specific advice.No portion of this presentation or supporti

93、ng material may be provided to any other party without Millimans prior written consent.Neither the presenter,his employer,nor the Casualty Actuarial Society shall have any responsibility or liability to any person or entity with respect to damages alleged to have been caused directly or indirectly b

94、y the content of this presentation.Any Milliman estimates referenced herein,including M-PIRe estimates,make no provision for extraordinary future emergence of new classes of losses or types of losses not sufficiently represented in historical mortgage performance databases or that are not yet quanti

95、fiable,including the potential impact of the emerging situation regarding the COVID-19 pandemic.There is substantial uncertainty regarding the impact of COVID-19 on the level of mortgage default,prepayment,and severity rates.Exposures,claim frequency,and claim severity will likely be affected in way

96、s we cannot currently estimate.It is important to recognize that actual losses may emerge significantly higher or lower than the estimates in this analysis.It is unknown how the COVID-19 pandemic may affect the timeliness of future default losses.Milliman should be contacted directly in order to lic

97、ense M-PIRe and have a proper understanding of forecasts contained therein and all the inputs,methodologies,assumptions and qualifications and limitations surrounding such estimates.LimitationsThese slides are for general informational purposes only.Action should not be taken solely on the basis of

98、the information set out herein without taking specific advice.Accompanying Oral DiscussionThis document is not complete without the accompanying oral discussion and explanation of the underlying projection methodologies,results and variability.Limited DistributionThis presentation is solely for disc

99、ussion purposes only.This presentation should not be distributed,disclosed or otherwise furnished,in whole or in part,without the express written consent of Milliman.In the event such consent is provided,it must be provided in its entirety.Any reader of this presentation must possess a certain level

100、 of expertise in areas relevant to this analysis to appreciate the significance of the assumptions and the impact of these assumptions on the illustrated results.The reader should be advised by actuaries or other professionals competent in the area of actuarial projections of the type in this presen

101、tation,so as to properly interpret any communicated analytics.49Resources Available to Explore Participating in Future Credit Risk Transfer Transactions Jonathan Berenbom,Managing Director,GC AnalyticsGuy Carpenter|1166 Avenue of the AmericasNew York,NY 10036Direct:917-937-3189|Mobile:646-954-2025Jonathan.Bhttps:/ Schmitz,Principal and Consulting Actuary Milliman|17335 Golf Parkway,Suite 100Brookfield,WI 53045 USADirect:262-796-3323|Mobile:262-352-Alan Tiernan,Chief Actuary,Global Mortgage GroupArch Capital Group Ltd.Direct 1 441 278 9197|Mobile 1 441 591 Thank you