Leader: Matteo Pelagatti - UNIMIB; Other collaborator(s): Fabrizio Culotta - UNIMIB
The Forecasting task is aimed at deriving projections for a set of socioeconomic variables starting from official ISTAT demographic projections over the period 2022-2080. The output covers:
Among different available statistical methods, this task is performed by following a dynamic microsimulation approach. In doing so, the Italian population is simulated at the individual level (micro units). Each micro unit is characterized by a set of socioeconomic variables (e.g. education, household type, health and labour market status). Transition probabilities are estimated from available microdata. Output is obtained by aggregating individual outcomes.
The resulting model can be used as a simulation platform to project the effects of alternative assumptions and policy scenarios for the Italian ageing context in the short, medium and long term.
Literature review
The literature on dynamic microsimulation models is reviewed. The aim of the survey is to identify recent international and national contributions on the topic of population ageing as reference point for the modelling exercise.
AGIT: the demographic module
The statistical model underlying the projections of family dynamics, care needs and labour market participation derived from ISTAT demographic projections is named AGIT (AGeing ITaly, l’Italia che invecchia). Its goal is to project the implications of population ageing in Italy.
The first step is to build the demographic module of AGIT. In doing so, the ISTAT methodology is closely followed to replicate the official demographic projections over the years 2022-2080. Inputs are ISTAT time series on age-specific fertility rates (1977-2019), sex-by-age mortality (1974-2019), and migration rates (2015-2019). The dynamics of fertility, mortality and migration forces give rise to the dynamics of births, deaths and migration flows. All together, they characterize the overall population dynamics and its ageing process.
Conferences
Conceptualization, literature review, first replicative results are presented for the first Age-It Conference (September 2023, Florence) and for ninth World Congress of the International Microsimulation Association (January 2024, Vienna).
Main policy implication
The microsimulation model AGIT represents a valid policy tool aimed at analyzing and projecting in the future (up to 2080) alternative scenarios and cost-benefit analysis of specific public interventions. This will enable policymakers to have a quantitative comprehension of the main implications of population aging in Italy and, along a data-driven approach, to formulate adaptive and corrective interventions.
AGIT: beyond the demographic module
The model population is enriched along other socioeconomic dimensions regarding household type, health status and labour market participation. The software R, package MicSim, is used for the microsimulation in continuous time. But when the complexity of the model is relatively high, an entry-level toolkit is no more adequate. This drawback motivates the adoption of a specific microsimulation programming language, ModGen (Model Generator), rather than rely on specific R packages.
AGIT: an Italian adaption of microWELT
The choice of ModGen is also related to the possibility of using microWELT, a well-established and fully-documented dynamic microsimulation model in continuous time covering the full individual life (Spielauer et al., 2020). A visiting period is spent at WIFO to practice (Vienna, June-July 2024).
AGIT starts from microWELT 1.0, calibrated to the Italian case using ISTAT demographic projections and SILC (Survey on Income and Living Conditions) microdata to calibrate the initial population and estimate transition probabilities across states (e.g. education, partnership and marital status).
Conferences
Results from the demographic module are presented at the second Age-It conference (May 2024, Venice). They mainly regard age-specific fertility rates, sex-age specific mortality and net migration projected profiles. Results from the new version of AGIT, built now in ModGen, are presented for the first Spoke-1 Conference (October 2024, Rome). They contain projections on health status and labor force participation disaggregated by sex, age, education and partnership status.
Main policy implication
Population ageing deteriorates the health condition of an older and older Italian population and affects the dynamics of family formation, resulting in higher portion of single individuals without children.
Distinguishing by partnership type (single or in partnership, with or without children), first results show that kinlessness plays a role in identifying individuals declaring bad health, especially among women. At first instance, they can be considered as potential care receivers who will not receive informal assistance at home from their partners nor children.
Projecting care needs, gaps and expenditure
Modelling care needs, namely Long-Term Care (LTC) needs, requires identifying the target population affected by physical and mental limitations. The population in need of care is estimated from SHARE microdata (Survey on Health, Ageing and Retirement in Europe). Based on the number and type of limitations in base daily activities (ADLs) and instrumental ADLs (IADLs), LTC needs are modelled at the extensive margin. Furthermore, by converting LTC needs in hours of required LTC assistance, focus is extended towards the intensive margin. Finally, the LTC gap is obtained as difference between required and supplied assistance.
Modelling of LTC care needs in hours, with all forms of LTC services (nursing homes, formal and informal home care), is proposed as contribution together with the WIFO modelling team during the second visiting period at WIFO (October-December 2024, Vienna).
The proposed modelling framework accommodates for the projection of LTC expenditure based on official national data from the System of Health Account (OECD).
Main policy implication
Population ageing leads to an increase of the share of population exposed to the risk of suffering from physical and mental limitations. This increases the level of required assistance in ADLs and IADLs. In Italy, since most of LTC assistance is often covered informally by household and non-household members, increasing singleness and childlessness will raise the additional required amount of LTC assistance. Public intervention should target vulnerable individuals whose LTC needs are unmet.
Coming soon