Joint Research Centre

Transcription

1 Joint Research Centre the European Commission's in-house science service Serving society Stimulating innovation Supporting legislation The use of microsimulation model for fiscal policy analysis: Evidence from the EU Salvador Barrios and Anne van Bruggen Seville, 23 September 2016

2 Outline 1. Policy context 2. The EUROMOD model & the European Semester 3. Dynamic scoring of tax reforms in the EU 4. Summary and conclusions

3 1.Policy context

4 1. Policy context Subsidiarity: the Union shall only act if the objective can be better achieved at Union level than at Member State level. Proportionality: the action shall not exceed what is necessary to achieve the objective. Hence: tax policy is only harmonised to a limited extent to ensure the proper functioning of the EU's internal market.

5 Tax revenue, % GDP (2014)

6 Labour/consumption tax revenue % total tax revenue (2014)

7 EU's role in national tax policy and the usefulness of microsimulations I Stability and Growth Pact: rules to ensure that countries in the European Union pursue sound public finances. Member States decide specific measures. Commission assesses the revenue impact. Existing tool box of macro-estimates and country-specific expertise now complemented by microsimulations provided by the JRC.

8 EU's role in national tax policy and the usefulness of microsimulations II Economic policy coordination ("European Semester"): multilateral surveillance to boost growth and employment at the national level. European Commission provides country-specific analysis and recommendations including in the area of taxation. JRC makes a crucial contribution through microsimulations that assess the impact of reforms on revenue, labour supply and income distribution.

9 2. The EUROMOD model & the European Semester

10 What is EUROMOD? 1. The EUROMOD model and the European Semester Microsimulation model developed by University of Essex + 28 National teams & financed by the European Commission (DG EMPL). Plans to transfer to the JRC. Tax systems (PIT and SSC) are coded and validated with less than a year lag Latest version includes 2016 systems (all MS) and 2014 or 2012 survey data. Standardised coding across member states. EUROMOD accounts for the interactions between the various components of the tax and benefit systems. Macro-validation of EUROMOD outputs by comparison with national external statistics

11 1. The EUROMOD model and the European Semester Data used in EUROMOD EU-SILC cross-section data (2012) harmonised by EUROSTAT +/- ½ million individuals / households covered: Demographic, labour market, gross market income, pensions, Social transfers Some expenditures included (housing cost, payments for life inscurance ) Uprating factors (CPI, avg. earnings increase, legal changes in benefit amounts) update the values of non-simulated income variables. Survey data: need to account for sample design (Goedemé, 2013).

12 1. The EUROMOD model and the European Semester Output from EUROMOD Simulated tax liabilities, benefit entitlements, disposable income. Macro-results: budgetary and distributional impact of reforms (incl. poverty indicators) and effect on work incentives

13 1. The EUROMOD model and the European Semester Extension of EUROMOD to include consumption taxation Including consumption taxation (VAT and excises). 2016: +4 MS 2017:+10 MS (project with Un. Leuven & Essex) Imputation of consumption expenditures in EU-SILC (parametric estimation or statistical matching) using EUROSTAT Household Budget Survey Combined analysis (incl. Welfare) of direct and indirect taxes.

14 1. The EUROMOD model and the European Semester Extension to account for behavioural reaction Use EU-SILC data to estimate labour supply function with intensive and extensive margins (project with ZEW- Mannheim) Estimate logit model for married couples and single individuals (by gender). Tax benefit function based on EUROMOD evaluated at discrete points of the labour supply choice set.

15 1. The EUROMOD model and the European Semester Examples of policy questions (European Semester 2015) Budgetary and equity impact of personal income tax reform in Spain 2016? Would a refundable in-work tax credit reduce the tax wedge of low-income earners in the Czech republic? Budgetary and equity impact of 80-euros tax credit in Italy? What is the budgetary and equity impact of reforming tax rates and tresholds of PIT in Austria?

16 JRC-EUROMOD web interface

17 Table 1. Budgetary impact (EUR) Budgetary impact Baseline Reform Difference Total Total Total Standard error 95% confidence interval % of baselin Lower bound Upper bound e Income tax (Einkommenssteuer) 30,761,834,345 29,714,849,521-1,046,984,824 46,205,933-1,137,564, ,405, Tax on investment income (Kapitalertragssteuer) 604,297, ,297, Total taxes 31,366,132,322 30,319,147,498-1,046,984,824 46,205,933-1,137,564, ,405, SIC employee 21,337,055,521 21,337,055, SIC employer 25,779,158,497 25,779,158, SIC self-employed 4,440,458,866 4,440,458, Total SIC 51,556,672,884 51,556,672, Pensions 44,516,840,048 44,516,840, Means tested benefits 4,526,113,168 4,517,219,433-8,893,735 1,484,328-11,803,531-5,983, Non-means tested benefits 9,099,074,221 9,099,074, Total benefits 58,142,027,437 58,133,133,702-8,893,735 1,484,328-11,803,531-5,983, Totals are aggregate figures for the whole population. All values are annual.

18 Distributional effects Table 4. Equivalised disposable income by deciles (EUR) Baseline Reform Difference Decile Mean Mean Mean Standard 95% confidence interval % of error Lower bound Upper bound baseline 1 10,637 10, ,968 14, ,536 16, ,822 19, ,144 21, ,672 23, ,471 26, ,883 30, ,269 35, ,117 55, All 25,083 25, Means are calculated at household level for the whole population. Equivalised values are calculated by applying the modified OECD equivalence scale (1 for the household head, 0.5 for other persons aged 14 or more and 0.3 for children under 14). All values are annual.

19 Impact on poverty measures Table 7. At-risk-of-poverty rates (%) for different income variables Baseline Reform Difference Rate Rate Rate Standard 95% confidence interval error Lower bound Upper bound Equivalised original income Equivalised original income + benefits Equivalised disposable income At-risk-of-poverty rates are calculated as the share of individuals whose equivalised income is below the poverty line. Poverty line is EUR (60% of the median equivalised disposable income in the baseline). Equivalised values are calculated by applying the modified OECD equivalence scale (1 for the household head, 0.5 for other persons aged 14 or more and 0.3 for children under 14). All values are annual.

20 3. Dynamic scoring of tax reforms in the EU

21 Dynamic scoring: objectives 1. Use EUROMOD to obtain precise modelling of tax reforms to assess impact of discretionary tax measures. 2. Use EUROMOD-labour supply model to account for the behavioural reaction to tax changes. 3. Combine EUROMOD-labour supply with DSGE model (QUEST, DG ECFIN) to account for interaction of reforms with the rest of the economy and macro-trajectories of prices and income to adjust micro-data

22 Dynamic scoring of tax reforms in the EU: methodological steps MICRO MACRO MICRO EUROMOD + LS Model QUEST EUROMOD RESULTS CALIBRATION Elasticities (behaviour) Non-participation rate (behaviour) Gross wages POLICY SHOCK Implicit tax rates (without behaviour) UPRATING FACTORS Gross wages CPI INPUT FILE Employmentt (weights of employed / rest of population) BASELINE REFORM WITHOUT BEHAVIOUR (direct fiscal impact) REFORM WITH BEHAVIOUR (medium-term projection)

23 Dynamic scoring: model consistency Labour supply function - macro & micro levels Optimal labour supply results from the maximization of an utility function Common determinants of labour supply: gross wages and the tax-benefit system, i.e., on disposable income Tax incidence (QUEST) consistent with micro-model Tax incidence depends on the sign and magnitude of the elasticities of labour supply and demand Inverse relationship between the change in total tax burden and net wages Positive relationship between the change in total tax burden and total compensation of employees

24 bil. euros bil. euros Dynamic scoring- example: reduction of SIC employees by 4 ppt in Belgium bil. euros bil. euros Total tax revenues Baseline (no policy change) Tax policy change, no behavioural reaction Tax policy change, including behavioural reaction SIC Employees Baseline (no policy change) Tax policy change, no behavioural reaction Tax policy change, including behavioural reaction Personal Income Taxes SIC Employers Baseline (no policy change) Tax policy change, no behavioural reaction Tax policy change, including behavioural reaction Baseline (no policy change) Tax policy change, no behavioural reaction Tax policy change, including behavioural reaction

25 Dynamic scoring: distributional effects on disposable income of reduction of 4 ppt in SIC in Belgium 6,600 10,000 Decile 1 Decile 2 11,600 Decile 3 9,800 11,400 11,200 6,400 9,600 11,000 12,600 Decile 4 13,600 Decile 5 14,400 Decile 6 12,400 13,400 14,200 12,200 13,200 14,000 12,000 13,000 13,800 16,200 16,000 15,800 15,600 15,400 15,200 Decile 7 17,600 17,400 17,200 17,000 16,800 16,600 Decile 8 20,200 20,000 19,800 19,600 19,400 19,200 19,000 Decile 9 28,500 28,000 27,500 27,000 26,500 26,000 Decile 10

26 Summary and extensions Until recently the European Commission made only a macro-assessment of the impact of tax reforms. Microsimulation model only used recently. EUROMOD is workhorse model for JRC analysis of tax reforms. EUROMOD used extensively for the assessment of actual or hypothetical reforms, including for PIT, SSC and recently consumption taxation. Ongoing extensions should better cover tax expenditures (linked to personal expenses) and wealth taxation Use of administrative tax data?

27 Thank you!