Fund Cashflow Model by ort32188

VIEWS: 4 PAGES: 11

More Info
									                                  ValueData7   1




Index-linked longevity risk transfer
reduced basis risk with sociodemographic
parameter


Dr. Albert Jürgen Enders
Managing Director
ValueData7
                                                                   ValueData7          2




Model of longevity risk profile of a young pension
plan evaluated with actual ex

 Pension fund cashflow profile in € mio                             Hedging of
                                                                    Interest risk
                                                                    and Inflation
                              2% ex (unexpected)
                                                                    risk is common
                              improvements p.a. effects
                              25% increase in liabilities
€ mio100                      after 10 years
                                                                    What about
                                                                    longevity risk ?



 € mio50




     2009   2019    2029    2039      2049       2059       2069



                                                                                       2
                                      ValueData7   3




Xpect Indices of Deutsche Börse are
published monthly




                                                   3
                                                                                   ValueData7                            4




Modeling the Xpect generation life table
is based on the last 30 years qx trend                                                                   Year of Birth
          1875 1876 1877                   1907              1976                  2003 2004 2005 2006
      0
      1
      2
      3




                           realised data
                                                  30 years

Age


                                                                    modeled data

   97
   98
   99
  100




  Only realised qx per age from the last 30 years are considered for modeling
                         the Xpect generation life table

                                                                                                                         4
                                                                                              ValueData7                    5




        These actual ex represents longevity changes of the
        population near-time
 Remaining life expectancy for all                               Remaining life expectancy for all
 respectively 65 year old men in Germany                         respectively 65 year old women in Germany
       ex                                                              ex
  30                                                              30




  25                                                              25




  20                                                              20




  15                                                              15
    1940    1950      1960    1970    1980    1990   2000           1940    1950      1960    1970    1980    1990   2000

                   Xpect generation table 2005                                     Xpect generation table 2005
                   Xpect generation table 2004                                     Xpect generation table 2004
                   destatis* generation table 2004                                 destatis* generation table 2004

(*) destatis = Statistisches Bundesamt Deutschland (German Federal Statistical Bureau)
                                                                                                                            5
                                                            ValueData7      6




   Parametric Indices as underlying support longevity
   risk transfer, but basis risk is still an issue
                                               Hedging on a
                                               parametric index

                                               = standardised index hedge
                                               = value hedge
                         Xpect Indices         = cheaper, more liquid

                                         basis risk



                 covered risk

Hedging on
indemnity risk

= tailored hedge
= cash flow hedge
= expensive (structural costs)
                                                                            6
                                             ValueData7             7




We selected the CHAID approach to get
sociodemographic segment qx indices

                        To get the relevant attributes a chi-square
                        independence test is applied
                        400.000 dead records in relation to 70
                        mio live records – comparison of deads
                        with alives per year of birth
                        The parameter with the highest ratio (chi
                        square ratio) is selected
                        Result: cohort and gender specific qx
                        prognosis based on residential quarter
                        parameters




                                                                      7
                                                                            ValueData7            8




  Female born 1943-1947 (61-65 years old)
  Relevance of parameter value to qx prognosis
Priority 1   Geldanleger

             AC Nielsen Freizeitmilieus_Wahrscheinlichkeit Traditionelle
Priority 2   Anzahl Gewerbehaushalte

             Technik
             Wohnadresse ist Senioren- oder Pflegeeinrichtung                 the higher the
Priority 3   Distanz zur nächsten Autobahn                                     parameter value,
             Distanz zum nächsten Kernkraftwerk                               the higher qx
             Direktversicherte
                                                                              the higher the
             TV Lotto                                                         parameter value,
             TV-Magazin                                                       the lower qx
Priority 4   Lifestyle
             Anzahl Haushalte
             Distanz zum nächsten Park
             Yellowpress
             Distanz zum nächsten ICE-Bahnhof
             AC Nielsen LOHAS Typologie_Reife LOHAS
             Familienanteil
Priority 5   Arbeitslosenquote (Gemeindeebene)
             Einwohnerdichte
             Finanzinteressierte
             AC Nielsen Freizeitmilieus_Wahrscheinlichkeit Intellektuelle
                                                                                                 ValueData7                          9



Index results: Male, 64 years old, born 1944
                                                              Xpect customised cohort index 
                                                               (socio demographic factors)
                                                100,000
                                                 90,000
                                                 80,000




                         Numbers of survivors
                                                 70,000
                                                 60,000
                                                 50,000
                                                 40,000
                                                 30,000
                                                 20,000
                                                 10,000
                                                      0
                                                          1    11   21     31    41        51   61    71     81    91    101   111

                                                                                                age

                                                              Xpect cohort index (base)           Xpect cohort index (best)
                                                              Xpect cohort index (worst)



                           Adjustments were based on the assumption
                           that the effects are weaker before the point
                           of measurement and stronger afterwards.




                                                                                                                                     9
                                                                                                                                            ValueData7 10

                                              Index Forward Curves (with or without
                                              sociodemographic parameter) are the underlyings for
                                              Longevity Index Products e.g Forwards, zerobonds
                                                                                Xpect Forward Cohort DE M 1920‐1939
                                      45000



                                      40000

                                                                                                                  a change in qx resp. ex per
                                      35000
                                                                                                               cohort changes the index
Number of normalised survivors (lx)




                                      30000                                                                    forward curve
                                      25000                                                                       In other words: a monthly
                                                                                                               change in qx resp. ex changes
                                      20000
                                                                                                               the index respresenting the
                                      15000                                                                    longevity liabilities of a portfolio
                                      10000



                                       5000



                                          0
                                              2008   2013             2018           2023               2028           2033          2038              2043   2048

                                                            Cohort Value (Feb '09)   Cohort value (jan '09)    Cohort value (2008)   Cohort Value (2007)             10
                               ValueData7 11




Thank you for your attention

Questions ?



Dr. Albert Jürgen Enders
Managing Director
ValueData7 GmbH
D-61462 Königstein
juergen.enders@valuedata7.de

								
To top