Abstract :
Recentempiricalevidencesupportstheviewthattheincomeprocesshasanindividual-
specificgrowthratecomponent[Baker,M.,1997.Growth-rateheterogeneityandthe
covariancestructureoflife-cycleearnings.JournalofLaborEconomics15,338–375;
Guvenen, F.,2007b.Learningyourearning:Arelaborincomeshocksreallyvery
persistent?AmericanEconomicReview97,687–712;Huggett,M.,Ventura,G.,Yaron,A.,
2007.Sourcesoflife-cycleinequality.Workingpaper,UniversityofPennsylvania].
Moreover,theindividual-specificgrowthcomponentmaybestochastic.Motivatedby
theseempiricalobservations,Istudyanindividual’soptimalconsumption–savingand
portfoliochoiceproblemwhenhedoesnotobservehisincomegrowth.Asinstandard
incomefluctuationproblems,theindividualcannotfullyinsurehimselfagainstincome
shocks. Inadditiontothestandardincome-risk-inducedprecautionarysavingdemand,
theindividualalsohaslearning-inducedprecautionarysavingdemand,whichisgreater
whenbeliefismoreuncertain.Withconstantunobservedincomegrowth,changesin
beliefarenotpredictable.However,withstationarystochasticincomegrowth,beliefis
no longeramartingale.Meanreversionofbeliefreduceshedgingdemandonaverage
and inturnmitigatestheimpactofestimationriskonconsumption–savingand
portfoliodecisions
