“…6 We indicate by 7 Obviously, the size of these vectors is allowed to vary over both time and cohort. From a longitudinal point of view, we indicate by i,h ,ȳ i,h andp i,h respectively lifetime income, average working and average pension income received by the ith income unit of cohort h. From a cross-section point of view, instead, we indicate by µ(y t ) and µ(p t ) respectively the average working and pension income among living individuals at time t. Henceforth, we refer to working and pension income respectively in 6 If each individual is allowed to participate in both funded (λ i ) and unfunded schemes (1− λ i ) at the corresponding rates of return (r,g), then, unless of particular circumstances (λ i = λ ∀ i) the comparison between actual and virtual income distributions as obtained under actuarial fairness would naturally entail some fictitious intra-generational redistribution, whose origins are indeed of the inter-generational kind (Disney 2004). 7 The distribution of virtual pension incomes is constructed assuming no individual behavioral response under the null solidarity hypothesis.…”