Alternative Tax-Benefit Strategies to Support Children in

Download Report

Transcript Alternative Tax-Benefit Strategies to Support Children in

Cross-National Microsimulation using
EUROMOD
How does it work and what can it do?
Holly Sutherland
Institute for Social and Economic Research, University of Essex, UK
QMSS conference 20-23 June 2007, Prague, Czech Republic
2
Outline
• What are microsimulation models?
• Cross-national microsimulation and EUROMOD
• More about EUROMOD
–
–
–
–
–
–
Framework, design and data
Comparability across countries
What can it used for?
Some examples
Becoming a EUROMOD user
Current limitations and future plans
3
Microsimulation models
• “Static” and “dynamic” microsimulation
– Static “morning after”
– Dynamic: long term, behavioural, both
• Microsimulation models provide data for analysis
• Tax-benefit models are “static” and deal with household
income, re-calculating income components (cash benefit
entitlements and personal tax liabilities) under alternative
scenarios
• “Organised arithmetic” (lots of it…)
• Not so much a “method” but more of a multi-purpose “tool”
that might be combined with other methods/tools
4
Tax-benefit models
• Consistent results for income-related indicators, including:
–
–
–
–
–
budgetary effects
income distributions (poverty and inequality indicators)
redistributive effects
gainers and losers
indicators of work incentives (RRs METRs)
• Budget constraints (=incomes under alternative conditions) for
modelling individual choices at the micro level
• Due to changes in
– Tax and benefit policies (actual, proposed or illustrative)
– Household/individual characteristics or macro conditions
5
What are the main inputs?
1. Microdata on households and individuals, from income
surveys or administrative data
2. Coded policy rules
3. An interface or framework for changing the rules and
assumptions, usually via parameters
 Contrast with analysis of income data taken directly from a
survey
–
–
–
–
more detail on components of taxes/benefits; can be more up-to-date;
both net and gross income
“what if” questions
tested, consistent, documented, organised, “branded”
replicable results…multi use and multi user
6
A multi-country model for cross-national research?
• Consistent results across countries
– Comparative analysis of the effects of policies and policy reforms
– EU-level outputs
– Policy learning across countries: “policy swapping”
 Understanding the effects of tax-benefit systems on different populations
 “Borrowing” policies that seem effective in one country (e.g. UK WFTC)
• But policy structures are very different across countries in several
dimensions:
– Policy design (e.g. rates, brackets, scales, limits)
– Unit of assessment (e.g. family vs. household)
– Definition of personal characteristics (e.g. who counts as a child?)
– Income assessments (e.g. tax base)
– Interaction with other policies (e.g. taxable or not)
• Data sources and requirements are different too.
7
EUROMOD
• EUROMOD is a multi-country tax-benefit model: unique
• It was built because of difficulties in making national model calculations
comparable, funded by a series of EC FP projects (1998-2008)
• National models exist in most of the EU15 and some of the 12 NMS
• Typically much more flexible than national models but covering the taxbenefit systems in less detail or more selectively
 National models for national analysis
 EUROMOD for comparative/comparable multi-country analysis
• Provides national tax-ben modelling capacity in/for Austria, Luxembourg,
Portugal and Greece
• Currently EU15 and working on 12 NMS (PL, HU, SI, EE soon); common
policy years 1998, 2001, 2003 (some), various other years for selected
countries
8
EUROMOD structure
Policy rules
(e.g., income tax)
Code
Parameters
(e.g., tax schedule)
(e.g., tax rates)
Micro output
Runinterface
Model input data
(e.g., imputed gross
income)
Original data
(e.g., ECHP, FES)
(e.g., net income)
Tools
(e.g., summary
statistics)
Specific analysis
9
EUROMOD base datasets for EU15 (June 2007)
Country
Base Dataset for EUROMOD
Date of collection
Austria
Austrian version of EU-SILC
2004
Belgium
Panel Survey on Belgian Households
2002
Denmark
European Community Household Panel
1995
Finland
Income distribution survey
2001
France
Budget de Famille
2001/2
Germany
German Socio-Economic Panel
2002
Greece
Household Budget Survey
2004/5
Ireland
Living in Ireland Survey
1994
Italy
Survey of Households Income and Wealth
1996
Luxembourg PSELL-2
2001
Netherlands Sociaal-economisch panelonderzoek
2000
Portugal
European Community Household Panel
2001
Spain
European Community Household Panel
2000
Sweden
Income distribution survey
2001
UK
Family Expenditure Survey
2000/1
10
Key EUROMOD data requirements
• A recent, representative sample of households, weighted to population level
and to correct for non-response
• Income variables
–
–
–
–
–
No missing values
gross of taxes and contributions
by source of income (e.g. national benefits, not harmonised categories)
by individual recipient (unless naturally household-based e.g. housing benefit)
reference period(s) corresponding to tax-benefit assessment periods
• Other variables
– housing costs and other expenditures affecting tax liabilities or benefit
entitlements (mortgage interest, child care payments…)
– other info affecting tax liabilities or benefit entitlements (hours of work, civil
servant status, private pension contributions...)
– individual characteristics and within-household relationships in the same
period(s) as the income information
11
Database construction
To date
• Responsibility of national teams, including choice of source of data
• Core common variables, plus necessary others
• Imputations and adjustments, including net-to-gross income imputation
• Updating factors to take incomes (by source) from the data year (income
reference period) to the policy year (1998, 2001, 2003…)
In the future
• EU-SILC with significant amount of common (centralised) data
manipulation, done regularly.
• Deficiencies for this purpose of Eurostat EU-SILC are outweighed by scale
economies, simplification of user access and strategic advantages.
12
Extent of parameterisation (flexibility)
In national models:
• monetary elements, tax rates etc
• eligibility conditions (partly)
• income definitions (partly)
In EUROMOD:
• monetary elements, tax rates etc
• eligibility conditions
• income definitions
• unit of assessment
• interaction of policy elements
• order of policies
Main menu
13
Policies
14
15
Comparability of results across countries
•
•
•
•
Input data quality, representativeness and coverage considerations
Data reference date; income reference period
A common framework for doing equivalent things across countries
Output income concepts (ie scope of the income calculation) can be defined
by the user; defaults are offered
• A huge range of user options and choices “comparability through
flexibility”
• Harmonisation of input data can be a hindrance
– e.g. ECHP and SILC benefit categories
• Comparable outputs often require inputs that are different across countries.
• Involvement of national teams in model construction, development and
collaborative research applications
• Documentation to aid interpretation of results
– Country Reports (validation)
• A tension between quality in any one country and comparability
16
Comparability of results across countries
•
•
•
•
Input data quality, representativeness and coverage considerations
Data reference date; income reference period
A common framework for doing equivalent things across countries
Output income concepts (ie scope of the income calculation) can be defined
by the user; defaults are offered
• A huge range of user options and choices “comparability through
flexibility”
• Harmonisation of input data can be a hindrance
– e.g. ECHP and SILC benefit categories
• Comparable outputs often require inputs that are different across countries.
• Involvement of national teams in model construction, development and
collaborative research applications
• Documentation to aid interpretation of results
– Country Reports (validation)
17
Examples of EUROMOD use
• Two examples from my own work
– Measuring child-contingent incomes from taxes and benefits
– A policy swap
• EUROMOD WPs illustrate other types of model use
– www.iser.essex.ac.uk/msu/emod/publications/emodwp.php
18
Spending on child contingent net transfers and tax concessions
Per child spending as a % of per-capita hhold disposable income 2001
25%
20%
Benefits
Taxes
Net effect
15%
10%
5%
0%
-5%
LU
AT
BE
IR
UK
FR
GE
DK
FI
SW
NL
PT
IT
GR
SP
EUROMOD Working
Paper EM4/04
19
The distributional effect of a budget-neutral UK 2003 system of childtargeted cash support
Shares of spending and shares of children by decile group of household income
25%
25%
20%
20%
Austria
25%
Spain
UK
20%
15%
15%
15%
10%
10%
10%
5%
5%
5%
0%
0%
0%
1 2 3 4 5 6 7 8 9 10
EUROMOD Working Paper
EM10/05
1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10
20
Examples of EUROMOD use
EM3/02 P Feres, H Immervoll, C Lietz, H Levy, D Mantovani and H Sutherland:
Indicators for Social Inclusion in the European Union: how responsive are they to
macro-level changes?
EM3/03 H Immervoll and C O'Donoghue: Employment Transitions in 13 European
Countries. Levels, Distributions and Determining Factors of Net Replacement
Rates
EM1/04 H Immervoll, H Jacobsen Kleven, C Thustrup Kreiner, E Saez: Welfare
Reform in European Countries: A Micro-Simulation Analysis
EM5/05 D Mantovani, F Papadopoulos, H Sutherland and P Tsakloglou: Pension
Incomes in the European Union: Policy Reform Strategies in Comparative
Perspective
EM2/06 O Bargain and K Orsini: Beans for Breakfast? How Exportable is the British
Workfare model?
EM4/06 H Levy, C Lietz and H Sutherland: A Basic Income for Europe's Children?
EM2/07 M Matsaganis and M Flevotomou: The impact of mortgage tax relief in the
Netherlands, Sweden, Finland, Italy and Greece
21
Becoming a EUROMOD user
• Training and support
– Essex Summer School course
– Other courses?
– ECASS visits to Essex (for non UK-based people)
http://www.iser.essex.ac.uk/ecass/
•
•
•
•
The model itself, documentation etc can be downloaded free
Data access is an issue
Have a good idea of what you want to use it for
Get in touch. [email protected]
22
Current limitations and future plans
• Age of underlying data in some countries *
• Direct taxes and cash benefits only; no indirect taxes or non-cash benefits *
• Most contributory benefits are not fully simulated *
• Static calculations only *
• Benefit take up is assumed to be 100%; no tax evasion*
• Tax benefit systems for 1998, 2001, 2003 (some)*
• EU15 only; 4 NMS being incorporated (EE, PL, HU, SI) *
• Permission to access some datasets is restricted *
* work is underway or planned to remedy these limitations….
23
http://www.iser.essex.ac.uk/msu/emod/