ThestudyisbasedindatafromtheCentralStatisticalOffice(GUS)data-
bases.Theremainingpartofthestudyisstructuredasfollows:Thesecondsec-
tiondescribesthedefinitionoflifeexpectancymeasureandtheideaofthetheo-
reticalconceptunderlyingthedeterminantsofmortality.Thenextsection
presentsthebasicdataonthelifeexpectancyofwomenandmeninPoland.The
theoryoffixedeffectspaneldatamodelsandtheseeminglyunrelated(SUR)
estimationofsetofmodelsarediscussedinsection4.Theempiricalanalysisis
describedinsection5.Itsfirstpartdescribestheadoptedresearchassumptions,
whilethesecondsectiondiscussestheresults.Thefinalsectionconcludes.
2.Lifeexpectancyanditsdeterminants
–theoreticalissues
2.1.Mathematicaldefinition
RandomvariableTxrepresentsthecompletefuturelifetimeforalifeofex-
actage.Then,completeexpectationoflifeforalifeofagexisexpectedvalueof
therandomvariableTx:
ex=E(Tx)=∫t⋅tpx
o
Ż
d
(1)
wherepx
t
i
i
representsprobabilitythatapersonagedxsurvivesatleasttfurther
years.
Forpracticalreasons,exactagesareseldomused,andageisexpressedin
completedyears.Therefore,foradiscreterandomvariableKxrepresentingthe
totalnumberofyearsremainingforapersonagedxyears,completeexpectation
oflifetakestheform
2:
ex=E(Kx)=
l
2
+∑
Ż
kŻo
k⋅k+l
px
(2)
Modelingandforecastinglifeexpectancyisimplicitlythesameasmodeling
mortalityrates(andvice-versa).Therefore,inrecentdecadesthedevelopmentof
variousmeasuresbasedonaveragelifeexpectancyhasbeenobserved.Theyare
usedtoanalyzenotonlytheagingprocessofthepopulation,butalsotomodelthe
healthofthepopulationandthelevelofwell-being.Themostpopularpopulation
healthmeasuresusinglifeexpectancyare:YearsofLifeLost(YLL),YearsLived
withDisability(YLD),DisabilityAdjustedLifeYears(DALY),HealthyLifeYears
(HLY),HealthAdjustedLifeExpectancy(HALE),DisabilityAdjustedLifeExpec-
2
Thisformulaisaccurateifdeathsaredistributeduniformlyoveragivenyear.
13