T-DYMM: Background and Challenges Intermediate Conference Rome 10 th May 2011 Simone Tedeschi FGB-Fondazione Giacomo Brodolini
Outline Institutional framework and motivations An overview of Dynamic Microsimulation Models (DMSM) The structure of the model The way forward (migration, behavioral responses and the fiscal module) Conclusions 2 Simone Tedeschi-FGB
THE INSTITUTIONAL FRAMEWORK AND MOTIVATIONS Dealing with the EU pension framework Adequacy and sustainability are two sides of the same coin. Addressing concerns in pension adequacy can contribute to achieve the goals of Europe 2020 (higher and better quality employment and positive transitions, reduction of poverty) Reviewing the EU framework to better respond to changes in labour markets (EC, 2010, Green Paper on Pensions). 3 Simone Tedeschi-FGB
THE INSTITUTIONAL FRAMEWORK AND MOTIVATIONS Italy: the reforming process After two decades of intense reforms, the Italian pension system is fully consistent with European guidelines. The latest reform, enacted in July 2010 (Law 122/2010), has: tightened the exit window mechanism, equalized the statutory retirement age of women and men in the public sector introduced an automatic mechanism for adjusting the minimum age requirements to changes in life expectancy. However, the Italian system needs to be assessed with reference to structural changes of the labor market and in the aftermath of the financial crisis. Assessing adequacy calls for quantitative tools based on micro data 5 Simone Tedeschi-FGB
THE INSTITUTIONAL FRAMEWORK AND MOTIVATIONS The project aims: ADSILC and TDYMM The project is based on two main activities: Developing a unique and innovative dataset by matching, through individual fiscal codes, administrative longitudinal data coming from INPS (National Institute of Social Security) with survey data produced by ISTAT (National Institute of Statistics). Providing preliminary studies through the analysis of these original datasets. Developing a dynamic micro-simulation model for Italy. Endowing the Ministry with a policy evaluation tool. Including in T-DYMM a specific and simplified module accounting for the Italian tax and benefits system. 6 Simone Tedeschi-FGB
AN OVERVIEW OF DMSM The Model LIAM MIDAS-IT ADSILC T-DYMM EconLav (taxation) New possibilities in estimating heterogeneous trajectories in a segmented labour market 7 Simone Tedeschi-FGB med./long run distributional analyses
AN OVERVIEW OF DMSM A classification of microsimulation models Start Standard Simulation Models Micro Simulation Models Static Dynamic Static Ageing Dynamic Ageing Cross-sectional ageing Longitudinal Ageing T-DYMM Base-data: population Base-data: cohort Open Closed 8 Simone Tedeschi-FGB
AN OVERVIEW OF DMSM Dynamic MSM Use panel data = past information to simulate future life cycle = transitions Transitions can be semi-probabilistic, deterministic, behavioural high degree of institutional details (policy parameters) The aim is to generate a panel of future synthetic data Dynamics is achieved essentially by updating attributes of each micro-unit for each time interval (Caldwell, 1990) 9 Simone Tedeschi-FGB
AN OVERVIEW OF DMSM Alignment The aim of alignment is to ensure that output aggregates match external control totals The reason this is done is that MSM are not in General Equilibrium and data taken from relatively short periods of time may not fully reflect dynamics over time. For these and other the reasons dynamic MSMs may not be able forecast aggregate characteristics of population well MSMs exchange the loss in fully accounting for average effects with the gain in explicitly allowing for agents heterogeneity So far, alignment was the solution. In the next future integration with CGE models another option 10 Simone Tedeschi-FGB 10
0 density.1.2.3.4 what is endogenous and what is aligned -5 0 5 x f(x_t) f(x_t+i) without alignment Mainly, changes in 2 (3, 4 ) central moments are endogenous study also within group inequality over time 11 Simone Tedeschi-FGB
0 density.1.2.3 what is endogenous and what is aligned -5 0 5 10 x f(x_t+i)without align. f(x_t+i)alignned Mainly, changes in means is driven by alignment = exogenous, Only some processes are aligned, to embody the macro dimension. 12 Simone Tedeschi-FGB
THE STRUCTURE OF THE MODEL The Structure of T-DYMM Start Base Population: IT-SILC (2005) linked with Administrative data (INPS)=ADSILC DEMOGRAPHY-EDUCATION Mortality Fertility Education (three levels) Cohabitation and Marriage Separation or Divorce Children leaving home Migration (on working) Model population at time t Next year (t=t+1) Parameters = estimated coefficients Alignement AWG 2010 PENSIONS Old age pension Survival pension Disability Social assistance Transitions into/within the LM Occupational status Sector of employment Income generation (gross earnings) gross of taxes LABOUR MARKET Alignement AWG 2010 RETIREMENT DECISION Aggregation Simulated Panel 2006-2060 End No Yes Simulation year < 2060 Model population at time (t+1) 13 Simone Tedeschi-FGB Yes loop to 3 No Behavioural Response? (further develop.) FISCALITY (ECONLAV) Input gross labour + pension incomes Social security contributions Main deductions Personal income tax Tax credits Regional taxes Output After tax incomes Parameters = Policy Rules (current Vs counterfactual
THE STRUCTURE OF THE MODEL The demographic module Mortality, fertility and migration: AWG- projections (2008 2010) Education Step 1: Monte Carlo-routine assigns an educational attainment level to every 10-year old. Step 2: The individual enters the labour market at an age determined by the level of eduation. The marriage market module (from MIDAS): logit model that estimates the probability of marriage between pairs of individuals 14 Simone Tedeschi-FGB
THE STRUCTURE OF THE MODEL The Labour Market Module Transitions among occupational statuses and persistence Estimation of gross labour incomes controlling for ISTAT socio-demo characteristics Focus on atypical careers both in terms of transitions and wages to be further improved by estimating discrete duration models on retrospective/prosepective ADSILC 15 Simone Tedeschi-FGB
THE STRUCTURE OF THE MODEL Adjusting AWG projections for the recent crisis 2,3 2,1 1,9 1,7 1,5 1,3 1,1 0,9 0,7 AWG 2008 adjusted AWG 2008 macro alignment of Real GDP 16 Simone Tedeschi-FGB
employment rate THE STRUCTURE OF THE MODEL Adjusting AWG projections for the recent crisis employment rates males 0,92 0,90 0,88 0,86 0,84 0,82 0,80 0,78 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 17 Simone Tedeschi-FGB 30-34 adjust. 35-39 adjust. 30-34 AWG 35-39 AWG micro alignment: the model endougenously determines who is more at risk of an event, alignment sets the totals
Pension module As a result of the 1995 reform, in Italy there are currently three different public pension schemes which differ with respect to the pension formula. Workers enrolment to such schemes depends on their seniority at 1995, according to the following rules: Individuals with a seniority of at least 18 years at that date receive a benefit fully earnings related (so called retributivo). Individuals entered in labour market since 1996 receive a benefit entirely based on the NDC scheme (so called contributivo). Individuals with less than 18 years of seniority receive a mixed benefit computed pro quota by a weighted average of pension benefits resulting in earnings related and NDC schemes, where weights are years worked until and after 1995 respectively. Because of the specific features of the Italian pension system, in order to compute benefits many retrospective information on seniority, past earnings and employment statuses are needed. in the new database such information is available and does not need to be simulated 18 Simone Tedeschi-FGB
Innovative feautures Main features THE STRUCTURE OF THE MODEL The pension module Classification of individuals in the different pension schemes (retributivo, misto, contributivo) and funds (Fondo pensione lavoratori dipendenti, Autonomi, Gestione separata, etc.) Specification of requirements for eligibility to retirement benefits according to the reform process Computing the different kinds of retirement benefits (IVS) Better prediction of future incomes/pensions thanks to the availability of occupational history (INPS) at micro level linked to socio-demographic characteristics (IT SILC) Within group distributional analysis Adequacy analysis Introducing behavioral responses 19 Simone Tedeschi-FGB
THE STRUCTURE OF THE MODEL The Fiscal module The experience of Econlav allows to include gross revenues as an input in the labour market for wages estimation. In EconLav the Italian system of tax and benefits is reproduced with an high level of details. As distributional analysis focuses on net revenues and pensions, Econlav has been used also at the end of the estimation process to compute net values. The idea forward is to endogenize the fiscal module and base the retirement decisions on net incomes instead of on gross revenues. The availability of ADSILC allows also accounting for family income, instead of personal income only, as reference for retirement decision 20 Simone Tedeschi-FGB
THE WAY FORWARD Behavioural Response on retirement decision A desirable feature often ignored in dynamic MSMs is the ability to include feedback loops so that behaviour can respond to changes in public policy In order to be able to simulate behaviour[ ], the model needs to be able to call a policy simulation routine a number of times to quantify the financial impact of alternative choices on the decision in questions (such as the choice to work or retire). As a result the software framework must be designed to be able to incorporate feedback loops (O Donoghue, 2009 ) incorporating behavioural response can be computationally expensive, but with the next software (LIAM2) and higher computational power now available will soon be feasible 21 Simone Tedeschi-FGB 21
CONCLUSIONS Road map May 2011: o a version 0 of T-DTMM has started to be tested October 2011: o Complete pension module to simulate 3 regimes + other components o Improve labour market estimates by using appropriate methods accounting for high heterogeneity in working trajectories o Present first simulation results in the long run 2012?: o extension to migration o behavioural response on retirement decision 22 Simone Tedeschi-FGB
23 Simone Tedeschi-FGB Grazie per l attenzione
THE WAY FORWARD Migration The development of this important sub-module asks for: Some change in the simulation platform Deeper empirical analysis of demographic behav s as well as specific patterns in the labour market Hypotheses on back migration 24 Simone Tedeschi-FGB
Marriage Market (LIAM-MIDAS) If an individual is selected to marry or form a partnership a process is needed to determine which spouse they will take. The process used here is to take the characteristics of the individual chosen to marry and the characteristics of each possible spouse and determine the likelihood of a match. this is done using a logit model that estimates the probability of marriage between pairs of individuals. The module itself forms a matrix of the characteristics of the n men and n women selected to marry. It estimates a probability for each pairing and assigns a match to the couples with the highest probability of marrying. In order to avoid strange matches (Bouffard et al. (2001) ), the user can create a super-set of potential male partners, so that rather the last female in the marriage pool being forced to select an unlikely match, there remain a number of males to choose from. The logic is that those in the centre of the data are average people who are more likely to find a good match than someone at the extremes. 25 Simone Tedeschi-FGB 25
Married Cohabiting The marriage market person is selected to find a partner marriage market -link individuals -create a new household Marriage Marriage or cohabitation? Cohabitation Divorce? Separation? yes No No Yes Yes Marriage? Divorced Married No Cohabiting Single 26 Simone Tedeschi-FGB
The labour market module IN WORK No Yes WIDOW(ER)S PENSION BENEFIT UNEMPLOYMENT Disability benefit DISABILITY PUBLIC SECTOR PENSION BENEFIT EARLY-RETIREMENT, UNEMPLOYMENT FOR OLDER WORKERS, RETIREMENT Ye s Working pensioner? Early retirement benefit, unemployment benefit for older workers, No PRIVATE SECTOR PENSION BENEFIT SELF-EMPLOYED PENSION BENEFIT 27 Simone Tedeschi-FGB OTHER INACTIVE Yet to be implemented
The labour market module IN WORK Yes Yes Logit + (alignment) Yes PUBLIC SECTOR No Yes PUBLIC SECTOR PENSION SCHEME EMPLOYEE Perm or Temp No A-TYPICAL (Cococo, Co.pro) PRIVATE SECTOR PENSION SCHEME No No SELF EMPLOYED SELF-EMPLOYED PENSION SCHEME MONTHS OF WORK 28 Simone Tedeschi-FGB MONTHLY WAGE (Gross of taxes and contributions) FE panel estimator (+ alignment) WIDOW(ER)S PENSION BENEFIT