Presentation is loading. Please wait.

Presentation is loading. Please wait.

A Framework for Pension Policy Analysis in Ireland: PENMOD, a Dynamic Simulation Model T. Callan, J. van de Ven and C. Keane.

Similar presentations


Presentation on theme: "A Framework for Pension Policy Analysis in Ireland: PENMOD, a Dynamic Simulation Model T. Callan, J. van de Ven and C. Keane."— Presentation transcript:

1 A Framework for Pension Policy Analysis in Ireland: PENMOD, a Dynamic Simulation Model T. Callan, J. van de Ven and C. Keane

2 Analysing pension policy: the need for a dynamic approach Some key features about pensions Contributions (DC) or entitlements (DB) build up over time Entitlement to pension starts at a particular age True of State Pension and of private/occupational pensions Pensions are payable from pension age till death Sometimes pensions payable to surviving spouse Decisions about providing for a pension are influenced by a range of circumstances: age, accrued wealth, the retirement benefits system

3 Pension policy Policy regarding State pension State pension age Level of payment Policy regarding private pensions Tax treatment of employee and employer contributions Standard fund threshold Impact of changes in these aspects of pension policy unfolds over time

4 Dynamic simulation approach needed No actual data source can track the implications of current or future policy changes This requires simulation – analysis using past data and a “what if” approach Most simulation models are based on cross- sectional (“snapshot”) data Some insights can be gained from such models Many issues would benefit from a dynamic approach (“movies” instead of “snapshots”)

5 Dynamic cohort microsimulation Representative cohort of individuals Simulate key elements of their lifetime experience Age, marriage, divorce, children, death Labour market participation, wages Decisions on savings and pensions Tax and welfare rules Strict limits on degree of detail in order to be able to simulate over full lifetime (to age 120)

6 From Static to Dynamic Modelling Static microsimulation (e.g., SWITCH) Allows detailed modelling of tax and welfare rules Relates to “real” data Dynamic cohort simulation To cope with analysis over time, need to use strategic simplifications for tax, welfare and pensions Based on “synthetic” data Which is better? Balance of advantage depends on issue studied

7 Strategic Simplifications: Model Scope Exclusions Public sector employees Self-employed Persons with illness or disability, affecting benefits eligibility and employment Age 20-120 The precise time of death is uncertain Labour supply Full-time work, part-time work, not at work

8 Strategic Simplifications: State Pensions State Contributory Pension All aged above State Pension Age assumed eligible Therefore State Non-Contributory Pension not needed But could allow for a means-tested top-up of the State Contributory Pension Key parameters (can be changed) State Pension Age Level of State Contributory Pension

9 Strategic simplifications: Welfare Cannot include full range of welfare schemes For those of working age: Floor to income provided by Jobseekers’ Allowance (JSA) Child benefit Qualified (dependant) child addition Family Income Supplement (FIS) if in work One-Parent Family payment (OPFP)

10 Strategic Simplifications: Tax/PRSI/Levies System basics Tax bands, rates, personal and PAYE credits PRSI, levies – allow for exemption limit, allowance, ceiling and rate Special attention given to potential tax treatments of pensions “Old system”: EET Potential new system – NPF – tax relief at a single hybrid rate (exact rate can be specified)

11 Matching the model to survey data Calibration or estimation? Both approaches involve the same basic structure: 1) estimate parameters that are observable 2) adjust unobserved parameters to match simulated population characteristics to survey data Being the first model of its type for Ireland, we calibrate unobserved parameters here

12 Matching the model to survey data Data from the 2005 population cross-section Consumption/saving & labour/leisure decisions observed in a single policy environment Consider cohort aged 20 in 2005 Cross-sectional data adjusted to reflect reasonable expectations regarding longevity and wage growth Official projections for improvements in longevity Observed wage growth over the full period for which data are available (1977 to present)

13 Calibrating the model – step 1 Observable model parameters estimated in the first stage include: Life expectancy Private sector pensions Allow for 3 separate schemes distinguished by their rates of employer and employee contributions Taxes and benefits Reflect policy in 2005 Household demographics Relationship status and the numbers of dependant children

14 Calibrating the model – step 2 Parameters adjusted in the second stage include: Wages Preferences across time: the discount rate, risk aversion Preferences within a period: the value of leisure relative to consumption Adjusted to match age specific population characteristics generated by the model to survey data: Employment income, employment, consumption

15 Calibrating the model – step 2 Employment Income – Single Adults

16 Calibrating the model – step 2 Employment Income – Adult Couples

17 Calibrating the model – step 2 Full-time Employed – Single Adults

18 Calibrating the model – step 2 Full-time Employed –Adult Couples

19 Calibrating the model – step 2 Not Employed – Single Adults

20 Calibrating the model – step 2 Not Employed –Adult Couples

21 Calibrating the model – step 2 Consumption – Single Adults

22 Calibrating the model – step 2 Consumption – Adult Couples

23 Consumption


Download ppt "A Framework for Pension Policy Analysis in Ireland: PENMOD, a Dynamic Simulation Model T. Callan, J. van de Ven and C. Keane."

Similar presentations


Ads by Google