ch4中财非寿险精算(韦晓).pptxVIP

Ch4.AggregatelossmodelsIntroduction(9.1and9.2)Generalpropertyofcollectiveriskmodel(9.2)ComputationthecollectiveriskmodelDirectcalculate(9.4)Approximation(9.5)Recursivemethod(9.6)Inversionmethod(9.8)Calculationwithapproximationdistribution(9.9)Individualriskmodel(9.11)Ch4.1AggregatelossmodelsTheaimofinsuranceistocombineindividualriskintoanaggregaterisktofurtherdiverselossintheriskpool.It’simportanttomodeltheaggregatelossforinsurancebusiness.AnexamplefromindividualrisktoaggregateriskseeLossModelspage199,example9.1.Ch4.1AggregatelossmodelsCh4.2CollectiveriskmodelGeneralpropertyofcollectiveriskmodelPfPgfMgfMeanVarianceBookP203-P205Section9.3Ch4.3.1DirectcalculateP217Example9.7exponentialseverity+NegativebinomialfrequencyP218Example9.8exponentialseverity+anyfrequencyP219Example9.9MixedErlangseverity+anyfrequencyP220Example9.10SeverityclosedunderconvolutionCh4.3.2ApproximationNormaldistributionapproximationLognormaldistributionapproximationCh4.3.3RecursivemethodSeveritydistributionisdefinedonintegerandwith(a,b,1)classfrequencydistribution(Th9.8)with(a,b,0)classfrequencydistribution(Co9.9)withcompoundfrequencymodelswithpremierandsecondarydistributionfrom(a,b,1)and(a,b,0)class,theaggregatedis.canbederivedP228Example9.12Ch4.3.3RecursivemethodForcompoundfrequencymodelswithpgfandseveritywithpgfThenthepgfofaggregateclaimsSisAnditsprobabilityfunctionthecanbederivedasStep1calculatebyStep2userecursiveformulatoderiveStep3calculatebyStep4userecursiveformulatoderivefromCh4.3.3RecursivemethodSeveritydistributioniscontinuousfor(a,b,1)classfrequency,thepfofaggregatesatisfiesVolterraintegralequationfor(a,b,0)classfrequency,thesameVolterraintegralequationisalsoderived.Note:numericalsolutionforVolterraintegralequationisavailablebutcomplicated.Ch4.3.3RecursivemethodOthersolutionforcontinuityseverityMethodofrounding(massdispersal)Methodoflocalmomentmatchinglocatepointmassesatpointssothatthefirstp(1or2)momentarepreserved.ThenCh4.3.4InversionmethodsFastFourierTransform(P242)Directnumericalinve