## wp18220

## Source details

**Canonical URL:** [wp18220](https://www.imf.org/-/media/files/publications/wp/2018/wp18220.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18220.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18220.pdf.json)

---

### Sample and sources (sections 2.1)
- Sample and scope
  - Start point: narrative dataset constructed by Devries et al. (2011) and extended by Alesina et al. (2015).
  - Focus: tax changes during years of fiscal consolidation in OECD countries.
  - Coverage: 10 OECD countries from 1978 to 2014.
  - Dataset constructed: nearly 2,500 exogenous tax changes for 10 OECD countries during fiscal consolidation episodes.
  - Final dataset: 2,445 exogenous tax measures, of which 1,585 had a long-run motivation, and 860 had a consolidation motivation.
  - Countries covered: Australia, Austria, Canada, France, Germany, Italy, Portugal, Spain, the United Kingdom, and the United States.
- Sources and validation
  - Primary sources: budget documents; reports from the Ministry of Finance and tax authorities; technical reports and notes produced during the legislative procedure; parliamentary debate records.
  - Secondary sources (used when necessary): IMF staff reports; OECD Economic Surveys; Stability and Growth Pact documents; national newspapers; International Bureau of Fiscal Documentation (IBFD).
  - Country-specific sources: Joint Committee on Taxation and Congressional Budget Office (U.S.); Ministry of Finance reports, Bank of Italy Economic Bulletin series, and Stability and Growth Pact documents (Italy).
  - Cross-checking: measures compared against other narrative studies and cross-country narrative datasets (e.g., Riera-Crichton et al. (2016)); justifications recorded when judgments differ.
- Advantages of the dataset
  - Greater detail than Devries et al. (2011) and Alesina et al. (2015) on tax types and whether measures were rate or base changes.
  - Records ex ante magnitudes and details of each tax measure, including forecast horizon (1 to 10 years).
  - Dates recorded: exact announcement and implementation dates, enabling use of quarterly data and identification of anticipation effects.
  - Motivation of each measure recorded to identify exogenous shifts.

### Key empirical findings (contextual)
- Main empirical result: tax base changes during consolidations appear to have a smaller impact on output and employment than tax rate changes of a similar size.
- Estimated cumulative tax rate multipliers: 1.2 (one-year cumulative), rising to 1.6 (two years).
- Estimated cumulative tax base multipliers: 0.3 (one-year cumulative), 0.4 (two years); these estimates are not statistically significant.
- Tax multiplier estimates are smaller than in the literature because estimation controls for base changes when examining rate changes.
- Robustness: findings are robust to alternative controls, estimation methodologies, and using annual instead of quarterly data.

### Coding tax measure types (section 2.2)
- Tax classification
  - Tax classification followed: OECD Revenue Statistics (2016 edition).
  - Broad tax categories assigned to each measure: personal income tax, corporate income tax, social security contributions, payroll tax, property and wealth tax, value added tax, sales tax, and other taxes on goods and services (e.g., excises).
- Rate vs. base coding (definition)
  - Tax rate measure: any change to the main statutory tax rate inclusive of surtax (examples: change to the standard VAT rate, the top corporate tax rate, any rate in personal income tax schedule).
  - Tax base measure: any change to the legal definition of the base to which the statutory rate is applied, or any change to provisions that modify the resulting tax liability.
  - Classification source for rate/base: following Amaglobeli et al. (2018).
- Examples of base measures by tax type
  - Personal income tax base measures: changes to personal allowances (including the zero bracket amount, if any), married couple and other specific allowances and deductions, tax credits (e.g., child tax credit), special treatment of specific income types or taxpayers (e.g., reduced tax rates on capital income or capital gains, income tax rates for foreign residents).
  - Corporate income tax base measures: changes to corporate tax credits (e.g., R&D or investment credits), depreciation schedules, loss carryforward and carryback rules, thin capitalization rules, transfer pricing provisions, controlled foreign corporation rules, tax treatment of foreign companies, reduced corporate tax rates for small businesses or specific industries.
  - Value added tax base measures: changes to reduced or increased VAT rates, the type of goods and services they apply to, and exemption of certain goods and services from VAT.
- Exclusions from base-change classification
  - Statutory inflation adjustments (indexation) are not considered tax base measures unless there is a deviation from the statutory indexation rule that leaves tax rates unchanged (such deviations are considered base measures).
  - Excluded actions: tax amnesties; pure revenue administration measures (changes in minimum VAT thresholds, penalties and fees, enforcement/collection improvements); changes that only alter the payment date of taxes.

### Size and timing of tax changes (section 2.3)
- Recorded magnitudes
  - Expected revenue impact recorded using forecasts produced by country authorities at announcement.
  - Forecasts expressed in local currency relative to an unchanged policies baseline.
  - Forecast horizon: can range from one to ten years ahead, depending on country and year.
  - When measures change future liabilities in stages, revenue impact of each step recorded separately when available.
- Announcement vs. implementation dates
  - Announcement date: defined as the first time the measure was formally proposed by the government; in some cases date of signature or promulgation of the final bill is used.
  - Implementation date: first moment when revenue impacts are non-zero; often coincides with beginning of next fiscal year.
  - Purpose of distinction: announcement dates capture when changes to future tax liabilities become certain (a news shock); implementation dates capture when disposable income actually changes.
  - Exclusions: actions that extend or renew existing taxes are excluded because announcement dates are often difficult to establish.

### Motivation (section 2.4)
- Endogenous vs. exogenous coding (methodology)
  - Methodology follows Romer and Romer (2010) in differentiating endogenous and exogenous tax changes.
  - Endogenous tax changes: measures taken to offset factors that push growth away from normal (examples: personal income tax cuts during a recession; excise changes to mitigate oil-price volatility; measures to finance a specific increase in government spending).
  - Exogenous tax changes: not taken in response to contemporaneous economic shocks, falling into two categories:
    - Measures primarily aimed at increasing long-run growth, competitiveness, protecting tax revenue, or increasing tax system efficiency (example: Australian income tax reform announced September 1985; corporate tax reforms such as UK 2010).
    - Measures taken primarily to reduce an existing fiscal deficit or ensure public debt sustainability.
- Treatment of deficit-reduction measures
  - Convention adopted: classify tax measures that aimed to reduce an existing fiscal deficit as exogenous (following Devries et al. (2011)).
  - Rationale and caveat: difficult to separate measures motivated by inherited budget deficits from those responding to current economic conditions; political cycle and framing can affect stated motivation.
  - Empirical check: Granger causality tests indicate tax measures with a consolidation motivation are uncorrelated with past business cycle conditions.

### Example: The UK June 2010 Budget (section 2.5)
- Overview and main themes
  - Budget presented to Parliament on June 6th, 2010; second budget of 2010 and first budget of the new Conservative-Liberal Democrats coalition government after the May 2010 general election.
  - Three main themes:
    - Eliminate the structural fiscal deficit by 2014-15 through additional tax and spending consolidation.
    - Reform corporate income tax to promote enterprise and sustainable growth.
    - Promote fairness in direct taxation (redistribution).
- Selected coded tax measures (values in billions of pound sterling; t0 = fiscal year 2010-11, t1 = 2011-12, t5 = five years out)
  - Increase main VAT rate to 20% / VAT Rate / 2.9 / 12.1 / 13.5 / 06/22/10 / 01/04/11 / Consolidation
  - Decrease CIT rate to 24% over 4 years / CIT Rate / 0 / -0.38 / -4.1 / 06/22/10 / 04/01/11 / Long-Run
  - Lower capital & investment allowances / CIT Base / 0 / 0 / 2.7 / 06/22/10 / 04/01/12 / Long-Run
  - Increase personal allowance by £1,000 / PIT Base / 0 / -3.3 / -3.8 / 06/22/10 / 04/06/11 / Long-Run
  - Increase capital gains tax rate to 28% / PIT Base / 0 / 0.7 / 0.9 / 06/22/10 / 06/23/10 / Long-Run
  - Increase child tax credit / PIT Base / 0 / -1.2 / -2.0 / 06/22/10 / 04/06/11 / Spending-Driven
- Aggregate totals reported for the June 2010 Budget
  - Total June 2010 Budget: 2.8 (t0) , 6.3 (t1) , 7.8 (t5)  [billions of pound sterling]
  - % GDP: 0.2 (t0) , 0.5 (t1) , 0.6 (t5)
- Policy details, motivations, and coding decisions
  - VAT increase:
    - Main standard rate raised from 17.5 to 20 per cent from 4 January 2011.
    - Coded as motivated by deficit reduction (Consolidation).
    - Expected to raise £2.9 billion in the 2010-11 fiscal year, and about £12.1 billion in the first full year of implementation (one percent of GDP).
  - Corporate income tax reform:
    - Top corporation tax rate reduced from 28 to 24 percent over four years starting in April 2011.
    - Announced decreases in capital (depreciation) and investment allowances from April 2012.
    - Reform aimed at increasing investment and job creation in future years; coded as Long-Run even though aggregate effect was revenue-decreasing.
  - Personal taxation and redistribution:
    - Personal allowance raised by £1,000 starting in April 2011; coded as Long-Run.
    - Capital gains tax rate increased to 28 percent; coded as PIT Base and Long-Run.
  - Child tax credit and linked spending action:
    - Increase in child tax credit financed by freezing child benefit over three years; classified as Spending-Driven.
- Additional items and counts
  - In total, 29 tax measures are identified in the June 2010 Budget.
  - Other significant tax measures mentioned include: a higher threshold for employer social security contributions; a new bank levy; an increase in insurance premium tax; a one-year freeze in council tax; a reduction in the CIT rate for small businesses; additional PIT base broadening measures.

### Dynamic effects of tax changes (section 4.1 and overview of 4.2–4.3)
- All tax types: response to a 1 percent of GDP tax increase (16-quarter horizon)
  - First-stage (tax-to-GDP ratio instrumented by narrative tax shock):
    - Point estimates positive throughout; response initially builds up consistent with announced intention.
    - Improvement in tax-to-GDP ratio then plateaus with a point estimate slightly below one.
    - Some evidence that some of the improvement in tax ratios is partially undone after a few years.
  - Second-stage (GDP instrumented by the narrative tax shock):
    - An announced increase in taxes of 1 percent of GDP leads to:
      - Decrease in output of about 0.3 percent in the first quarter.
      - Peak decline of about 1 percent eight quarters after announcement.
    - Results consistent with Romer and Romer (2010); Cloyne (2013); Alesina et al. (2015).
    - Results are robust to a battery of checks.
- Tax rate versus tax base changes: GDP and components (scaled to a 1 percent of GDP tax increase)
  - GDP responses
    - Tax rate shock:
      - Decline of about 0.4 percent in the first quarter.
      - Peak decline of 2 percent after about two years.
      - Impact persistent and remains significant even four years after announcement.
    - Tax base shock:
      - Impact on GDP is neither economically nor statistically significant.
    - Both rate and base shocks lead to a significant increase in tax revenue; increases are persistent but not permanent (tend to become smaller and non-significant after 12 quarters).
  - Consumption and investment (percent of GDP)
    - Private consumption:
      - After a tax rate shock: steep decline, decreasing by about 1 percent of GDP after about eight quarters, then recovering slightly.
      - After a tax base shock: more muted response; point estimates indicate a decline of about 0.3 percent of GDP a year after announcement, but not statistically significant.
    - Private investment:
      - After a tax rate shock: investment decreases by about 1 percent of GDP after about six quarters, then partially recovers.
      - After a tax base shock: point estimates indicate no decline in the first two quarters, and a decline of about 0.5 percent of GDP after about six quarters (not statistically significant).
  - Labor market
    - Unemployment rate:
      - Tax increases via rate hikes: significant increase starting at around 0.1 percent in the first quarter; builds up to about 1 percent after three years.
      - Tax increases via base broadening: smaller impact; no initial impact; point estimate rises slowly to about 0.6 percent in the medium-term, estimates statistically insignificant.
    - Average wages per employed worker:
      - Tax rate increases: average wages decline; point estimates indicate a statistically significant decline of around 2 percentage points after two years, appearing persistent.
      - Tax base increases: no similar decline observed in the first two years; point estimates remain close to zero; some evidence of a medium-term decline but not statistically significant.
- Translation into tax multipliers (overview)
  - Two multiplier estimation approaches: contemporaneous multiplier and cumulative multiplier.
  - Instrumentation: change in tax instrumented with relevant narrative tax shock; baseline controls included; first-stage F-statistics typically large.
  - Key findings (Table (3) summary):
    - Overall tax multiplier (all tax types):
      - About 0.8 to 0.9 after one year (4 quarters).
      - About 1 to 1.5 after two years (8 quarters).
    - Rate versus base multipliers:
      - Rate changes:
        - Multiplier about 1.2 after one year.
        - Reaches between 1.5 and 2 after two years (depending on specification).
      - Base changes:
        - Point estimates around 0.3 after one year.
        - Rise to about 0.4 or 0.5 after two years but not statistically different from zero.
    - Difference between rate and base multipliers robust to methodology.
- Robustness checks summary
  - Excluding all controls: tax rate multiplier about 1 (versus about 1.3 in baseline); tax base multiplier remains about 0.3; estimates less precise.
  - Controlling for contemporaneous government spending shocks: multipliers essentially unchanged.
  - Broader definition of narrative shocks: estimated multipliers slightly smaller (e.g., 0.7 using all exogenous tax shocks vs. 0.9 baseline), key findings unchanged.
  - Excluding temporary measures: estimated multipliers essentially unchanged.
  - OLS using narrative tax shocks directly: estimated multipliers slightly larger than baseline IV results; OLS still finds sizable difference between rate and base multipliers.
  - Annual data: large difference between rate and base multipliers persists; somewhat bigger multipliers; first-stage F-statistics substantially smaller with annual data.

### Dynamic effects of specific tax changes (section 5.2)
- Personal Income Tax (PIT)
  - Shocks scaled to a 1 percent increase in the effective PIT rate.
  - Output response to a PIT rate hike:
    - Fall in output on impact; builds and peaks at around 2 percent of GDP after about eight quarters.
  - Output response to a PIT base increase:
    - Smaller impact; no impact in initial quarters; small subsequent negative impact statistically insignificant.
  - Cumulative tax multipliers:
    - One-year cumulative multiplier for a PIT rate shock: 1.3.
    - One-year cumulative multiplier for a PIT base shock: 0.
    - Two-year PIT rate multiplier: 1.5.
    - Two-year PIT base multiplier: 0.4.
- Corporate Income Tax (CIT)
  - Shocks scaled to a 1 percent increase in the effective CIT rate.
  - Output response to a CIT rate hike:
    - Initial decline on announcement; builds and peaks at about 2.2 percent of GDP after about two years.
    - Medium-term impact imprecisely estimated with large confidence interval.
  - Output response to a CIT base increase:
    - Negative but smaller impact; typically insignificant in most quarters.
  - Note: estimated impact of CIT shocks on tax revenue is close to zero, implying limited sense in computing tax multipliers for these shocks.
- Value Added Tax (VAT)
  - Shocks equivalent to a 1 percent increase in the effective VAT rate.
  - Base changes for VAT: few instances exist; estimation of base-change effects not possible.
  - Output response to a VAT rate hike:
    - Small but negative initial impact; peaks at around 1.8 percent of GDP after seven quarters, then recovers slightly.
  - Cumulative multipliers:
    - One-year cumulative multiplier for a VAT rate shock: 2.2.
    - Two-year VAT rate multiplier: 2.6.
  - Comparison to literature:
    - Riera-Crichton et al. (2016) finds a one-year multiplier close to 4 for VAT rate shocks.
    - Gunter et al. (2017) find a two-year VAT rate multiplier of 2.1 in industrial countries.
  - Note on mechanism: larger VAT multiplier driven by a slower response of VAT collections in the first six quarters compared to quicker pickup in PIT revenue after a PIT rate hike.
- Synthesis and policy implications
  - Overall finding: changes in taxes have significant macroeconomic effects that vary with composition; tax base changes lead to smaller output and employment declines than tax rate increases.
  - Aggregate cumulative multipliers:
    - Estimated cumulative tax rate multiplier one year after announcement: 1.2.
    - Estimated cumulative tax base multiplier one year after announcement: 0.3 (not statistically significant).
    - Multipliers after two years: tax rate multiplier 1.6; tax base multiplier 0.4.
  - Policy implication: evidence supports policy advice that base broadening can raise additional revenue while being less detrimental to growth.
- Directions for future research
  - Explore variation in rate and base multipliers with initial conditions such as the state of the economy or features of the tax system (e.g., initial tax rate or overall tax structure).
  - Expand the dataset to include additional countries and non-consolidation years.
  - Classify base measures in more detail to indicate whether they would be expected to increase tax efficiency, enabling analysis of macroeconomic impact of different types of base measures and identification of those more likely to be beneficial during consolidations.

### Data and appendix highlights (from References and Appendix A–B)
- Dataset coverage and construction
  - Quarterly dataset for 10 OECD countries from 1975 to 2016.
  - Country note: for Germany, use West Germany data from 1979 to 1990 (Destatis) and whole Germany from 1991 onwards; regressions include a dummy equal to one for Germany in all quarters after 1991Q1 inclusive, and zero otherwise.
- Macroeconomic and tax sources
  - GDP, employment, and interest rates: OECD Economic Outlook Nr. 102 (November 2017) and OECD.Stat.
  - Tax revenue by tax type: national statistics offices and tax authorities; checked against annual OECD Tax Revenue dataset.
  - Statutory tax rates: OECD Tax Database, European Commission for VAT; supplemented with national sources.
- Tax base definitions used
  - VAT Base = Consumption − VAT revenue.
  - PIT Base = Household gross disposable income + Current taxes on income and wealth (excludes social security contributions).
  - CIT Base = Gross operating surplus of financial and nonfinancial corporations (does not subtract consumption of fixed capital).
- Table A.2 (summary statistics; quarterly sample 1978–2014; values in real per capita 2011 PPP dollars unless noted)
  - Macro (selected)
    - GDP p.c.: Mean 33,270; Std. Dev. 7,079; 10% 24,721; 25% 28,110; 50% 32,953; 75% 37,893; 90% 42,282; Obs 1,358
    - Consumption p.c.: Mean 19,086; Std. Dev. 4,567; 10% 13,617; 25% 16,125; 50% 18,900; 75% 21,477; 90% 23,223; Obs 1,358
    - Investment p.c.: Mean 7,079; Std. Dev. 1,943; 10% 4,794; 25% 5,644; 50% 6,823; 75% 8,304; 90% 9,718; Obs 1,358
  - Fiscal (selected)
    - Tax p.c.: Mean 8,193; Std. Dev. 2,232; 10% 5,300; 25% 6,391; 50% 8,150; 75% 9,800; 90% 11,045; Obs 1,358
    - Government gross debt p.c.: Mean 17,250; Std. Dev. 10,685; 10% 5,614; 25% 9,380; 50% 14,538; 75% 22,952; 90% 34,064; Obs 1,355
  - Personal Income Tax (selected)
    - PIT base p.c.: Mean 24,101; Std. Dev. 5,803; 10% 16,978; 25% 20,080; 50% 23,725; 75% 27,356; 90% 30,403; Obs 1,326
    - PIT revenue p.c.: Mean 2,874; Std. Dev. 1,006; 10% 1,474; 25% 2,103; 50% 2,867; 75% 3,591; 90% 4,184; Obs 1,358
    - PIT statutory (top) rate (%): 47 10 34 40 48 53 60; Obs 1,358
    - PIT effective rate (%): 12 4 7 10 12 13 16; Obs 1,326
  - Corporate Income Tax (selected)
    - CIT base p.c.: Mean 7,239; Std. Dev. 2,730; 10% 4,274; 25% 5,402; 50% 6,803; 75% 8,541; 90% 10,403; Obs 1,326
    - CIT revenue p.c.: Mean 841; Std. Dev. 439; 10% 351; 25% 514; 50% 770; 75% 1,060; 90% 1,477; Obs 1,358
    - CIT statutory (top) rate (%): 35 10 24 30 34 39 50; Obs 1,358
    - CIT effective rate (%): 12 5 6 8 12 16 20; Obs 1,326
  - Value Added Tax (selected)
    - VAT base p.c.: Mean 17,879; Std. Dev. 4,553; 10% 12,784; 25% 15,124; 50% 17,239; 75% 19,565; 90% 22,161; Obs 1,358
    - VAT revenue p.c.: Mean 1,553; Std. Dev. 959; Obs 1,358
- Appendix figures and tests
  - Figure A.1: Fiscal responses to rate or base increases (responses to a 1 percent of GDP tax increase; 68 and 90 percent confidence intervals).
  - Robustness figures: annual data and OLS estimation panels (Figure A.3, A.4).
  - First-stage responses for specific taxes (Figure A.5).
  - Granger-causality tests reported in Table A.3 and joint tests in Table A.4 indicate limited predictability of narrative tax shocks by past macro conditions.

*Source: wp18220 - sections 2.1–2.4, 2.5, 4.1, 5.2, Appendix A–B.*

### 2.1  Sample and sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   4

### 2.1  Sample and sources

### Sample and scope
- Start point: narrative dataset constructed by Devries et al. (2011) and extended by Alesina et al. (2015).
- Focus: tax changes during years of fiscal consolidation in OECD countries.
- Coverage: 10 OECD countries from 1978 to 2014.
- Dataset constructed: nearly 2,500 exogenous tax changes for 10 OECD countries during fiscal consolidation episodes.
- Final dataset: 2,445 exogenous tax measures, of which 1,585 had a long-run motivation, and 860 had a consolidation motivation.
- Countries covered: Australia, Austria, Canada, France, Germany, Italy, Portugal, Spain, the United Kingdom, and the United States.

### Sources and validation
- Primary sources: budget documents; reports from the Ministry of Finance and tax authorities; technical reports and notes produced during the legislative procedure; parliamentary debate records.
- Secondary sources (used when necessary): IMF staff reports; OECD Economic Surveys; Stability and Growth Pact documents; national newspapers; International Bureau of Fiscal Documentation (IBFD).
- Country-specific sources: Joint Committee on Taxation and Congressional Budget Office (U.S.); Ministry of Finance reports, Bank of Italy Economic Bulletin series, and Stability and Growth Pact documents (Italy).
- Cross-checking: measures compared against other narrative studies and cross-country narrative datasets (e.g., Riera-Crichton et al. (2016)); justifications recorded when judgments differ.

### Advantages of the dataset
- Greater detail than Devries et al. (2011) and Alesina et al. (2015) on tax types and whether measures were rate or base changes.
- Records ex ante magnitudes and details of each tax measure, including forecast horizon (1 to 10 years).
- Dates recorded: exact announcement and implementation dates, enabling use of quarterly data and identification of anticipation effects.
- Motivation of each measure recorded to identify exogenous shifts.

---

### Key empirical findings summarized in introduction (contextual)
- Main empirical result: tax base changes during consolidations appear to have a smaller impact on output and employment than tax rate changes of a similar size.
- Estimated cumulative tax rate multipliers: 1.2 (one-year cumulative), rising to 1.6 (two years).
- Estimated cumulative tax base multipliers: 0.3 (one-year cumulative), 0.4 (two years); these estimates are not statistically significant.
- Tax multiplier estimates are smaller than in the literature because estimation controls for base changes when examining rate changes.
- Robustness: findings are robust to alternative controls, estimation methodologies, and using annual instead of quarterly data.

---

### 2.2  Coding tax measure types

### Tax classification
- Tax classification followed: OECD Revenue Statistics (2016 edition).
- Broad tax categories assigned to each measure: personal income tax, corporate income tax, social security contributions, payroll tax, property and wealth tax, value added tax, sales tax, and other taxes on goods and services (e.g., excises).

### Rate vs. base coding (definition)
- Tax rate measure: any change to the main statutory tax rate inclusive of surtax (examples: change to the standard VAT rate, the top corporate tax rate, any rate in personal income tax schedule).
- Tax base measure: any change to the legal definition of the base to which the statutory rate is applied, or any change to provisions that modify the resulting tax liability (examples vary by tax type; see below).
- Classification source for rate/base: following Amaglobeli et al. (2018).

### Examples of base measures by tax type
- Personal income tax base measures: changes to personal allowances (including the zero bracket amount, if any), married couple and other specific allowances and deductions, tax credits (e.g., child tax credit), special treatment of specific income types or taxpayers (e.g., reduced tax rates on capital income or capital gains, income tax rates for foreign residents).
- Corporate income tax base measures: changes to corporate tax credits (e.g., R&D or investment credits), depreciation schedules, loss carryforward and carryback rules, thin capitalization rules, transfer pricing provisions, controlled foreign corporation rules, tax treatment of foreign companies, reduced corporate tax rates for small businesses or specific industries.
- Value added tax base measures: changes to reduced or increased VAT rates, the type of goods and services they apply to, and exemption of certain goods and services from VAT.

### Exclusions from base-change classification
- Statutory inflation adjustments (indexation) are not considered tax base measures unless there is a deviation from the statutory indexation rule that leaves tax rates unchanged (such deviations are considered base measures).
- Excluded actions: tax amnesties; pure revenue administration measures (changes in minimum VAT thresholds, penalties and fees, enforcement/collection improvements); changes that only alter the payment date of taxes.

---

### 2.3  Size and timing of tax changes

### Recorded magnitudes
- Expected revenue impact recorded using forecasts produced by country authorities at announcement.
- Forecasts expressed in local currency relative to an unchanged policies baseline.
- Forecast horizon: can range from one to ten years ahead, depending on country and year.
- When measures change future liabilities in stages, revenue impact of each step recorded separately when available.

### Announcement vs. implementation dates
- Announcement date: defined as the first time the measure was formally proposed by the government (formal communication of Prime Minister or Ministry of Finance, or introduction of draft legislation to Parliament). In some cases where legislative alteration was significant, the date of signature or promulgation of the final bill is used.
- Implementation date: first moment when revenue impacts are non-zero; often coincides with beginning of next fiscal year.
- Purpose of distinction: announcement dates capture when changes to future tax liabilities become certain (a news shock); implementation dates capture when disposable income actually changes.
- Exclusions: actions that extend or renew existing taxes (e.g., annual renewal of excise taxes, extensions of existing tax allowances and credits) are excluded because their announcement dates are often difficult to establish.

---

### 2.4  Motivation

### Endogenous vs. exogenous coding (methodology)
- Methodology follows Romer and Romer (2010) in differentiating endogenous and exogenous tax changes.
- Endogenous tax changes: measures taken to offset factors that push growth away from normal (examples: personal income tax cuts during a recession; excise changes to mitigate oil-price volatility; measures to finance a specific increase in government spending).
- Exogenous tax changes: not taken in response to contemporaneous economic shocks, falling into two categories:
  - Measures primarily aimed at increasing long-run growth, competitiveness, protecting tax revenue, or increasing tax system efficiency (example: Australian income tax reform announced September 1985; corporate tax reforms such as UK 2010).
  - Measures taken primarily to reduce an existing fiscal deficit or ensure public debt sustainability.

### Treatment of deficit-reduction measures
- Convention adopted: classify tax measures that aimed to reduce an existing fiscal deficit as exogenous (following Devries et al. (2011)).
- Rationale and caveat: difficult to separate measures motivated by inherited budget deficits from those responding to current economic conditions; political cycle and framing can affect stated motivation.
- Empirical check: Granger causality tests indicate tax measures with a consolidation motivation are uncorrelated with past business cycle conditions.

---

*Source: wp18220 - sections 2.1–2.4 (Narrative dataset: sample, coding, size/timing, and motivation).*

### 2.5  Example: The UK June 2010 Budget

### 2.5  Example: The UK June 2010 Budget

### Overview and main themes
- Budget presented to Parliament on June 6th, 2010; second budget of 2010 and first budget of the new Conservative-Liberal Democrats coalition government after the May 2010 general election.
- Three main themes:
  - Eliminate the structural fiscal deficit by 2014-15 through additional tax and spending consolidation.
  - Reform corporate income tax to promote enterprise and sustainable growth.
  - Promote fairness in direct taxation (redistribution).

### Selected coded tax measures (Table summary)
- Note: Estimated revenue impacts are in billions of pound sterling, and by fiscal year, with t0 representing fiscal year 2010-11, t1 representing fiscal year 2011-12, and so on. Data taken from the June 2010 and March 2011 Financial Statement and Budget Reports.
- Selected measures (measure / tax type / t0 / t1 / t5 / announcement date / implementation date / motivation):
  - Increase main VAT rate to 20% / VAT Rate / 2.9 / 12.1 / 13.5 / 06/22/10 / 01/04/11 / Consolidation
  - Decrease CIT rate to 24% over 4 years / CIT Rate / 0 / -0.38 / -4.1 / 06/22/10 / 04/01/11 / Long-Run
  - Lower capital & investment allowances / CIT Base / 0 / 0 / 2.7 / 06/22/10 / 04/01/12 / Long-Run
  - Increase personal allowance by £1,000 / PIT Base / 0 / -3.3 / -3.8 / 06/22/10 / 04/06/11 / Long-Run
  - Increase capital gains tax rate to 28% / PIT Base / 0 / 0.7 / 0.9 / 06/22/10 / 06/23/10 / Long-Run
  - Increase child tax credit / PIT Base / 0 / -1.2 / -2.0 / 06/22/10 / 04/06/11 / Spending-Driven
- Aggregate totals reported for the June 2010 Budget:
  - Total June 2010 Budget: 2.8 (t0) , 6.3 (t1) , 7.8 (t5)  [billions of pound sterling]
  - % GDP: 0.2 (t0) , 0.5 (t1) , 0.6 (t5)

### Policy details, motivations, and coding decisions
- VAT increase:
  - Main standard rate raised from 17.5 to 20 per cent from 4 January 2011.
  - Coded as motivated by deficit reduction (Consolidation).
  - Expected to raise £2.9 billion in the 2010-11 fiscal year, and about £12.1 billion in the first full year of implementation (one percent of GDP).
- Corporate income tax reform:
  - Top corporation tax rate reduced from 28 to 24 percent over four years starting in April 2011.
  - Announced decreases in capital (depreciation) and investment allowances from April 2012.
  - Reform aimed at increasing investment and job creation in future years; coded as Long-Run even though aggregate effect was revenue-decreasing.
  - Extended in subsequent years, eventually reducing the corporate tax rate to 20 percent by 2015 (as described in the narrative).
- Personal taxation and redistribution:
  - Personal allowance raised by £1,000 starting in April 2011 to reduce tax burden on lower-income taxpayers; coded as Long-Run.
  - Capital gains tax rate increased to 28 percent; coded as a personal income tax base measure because it changed treatment of capital gains without changing main PIT rate schedule; coded Long-Run.
- Child tax credit and linked spending action:
  - Increase in child tax credit above inflation financed by freezing child benefit over three years (explicitly linked tax and spending action).
  - Given the explicit link in the narrative account, the increase in the child tax credit is classified as Spending-Driven.

### Additional items and counts
- In total, 29 tax measures are identified in the June 2010 Budget.
- Other significant tax measures mentioned include:
  - A higher threshold for employer social security contributions
  - A new bank levy
  - An increase in insurance premium tax
  - A one-year freeze in council tax
  - A reduction in the CIT rate for small businesses
  - Additional PIT base broadening measures

*Italic: Source content — "2.5 Example: The UK June 2010 Budget" (excerpt from provided PDF content).*

### 4.1  Dynamic effects of tax changes

### 4.1  Dynamic effects of tax changes

### All tax types: response to a 1 percent of GDP tax increase
- Impulse responses present the effect of a shock equivalent to a tax increase worth 1 percent of GDP over a forecast horizon of 16 quarters, with 68 and 90 percent confidence intervals.
- First-stage (tax-to-GDP ratio instrumented by narrative tax shock):
  - Point estimates are positive throughout; response initially builds up over time consistent with announced intention.
  - Improvement in the tax-to-GDP ratio then plateaus with a point estimate slightly below one.
  - Some evidence that some of the improvement in tax ratios is partially undone after a few years.
- Second-stage (GDP instrumented by the narrative tax shock):
  - An announced increase in taxes of 1 percent of GDP leads to:
    - Decrease in output of about 0.3 percent in the first quarter.
    - Peak decline of about 1 percent eight quarters after announcement.
  - Results consistent with Romer and Romer (2010); Cloyne (2013); Alesina et al. (2015).
  - Results are robust to a battery of checks.

### Tax rate versus tax base changes: GDP and components
- Estimation approach:
  - Rate changes: change in tax-to-GDP ratio instrumented by narrative shock for rate changes only; control for current and lagged narrative tax base shocks.
  - Base changes: change in tax-to-GDP ratio instrumented by narrative shock for base changes only; control for current and lagged narrative tax rate shocks.
  - All responses scaled to represent a 1 percent of GDP tax increase enacted through rate or base measures.
- GDP responses:
  - Tax rate shock:
    - Decline of about 0.4 percent in the first quarter.
    - Peak decline of 2 percent after about two years.
    - Impact is persistent and remains significant even four years after announcement.
  - Tax base shock:
    - Impact on GDP is neither economically nor statistically significant.
  - Both rate and base shocks lead to a significant increase in tax revenue; increases are persistent but not permanent (tend to become smaller and non-significant after 12 quarters).

### Consumption and investment responses (scaled as percent of GDP)
- Definitions:
  - Variables scaled so that Y_{i,t+h} − Y_{i,t−1} = (x_{i,t+h} − x_{i,t−1}) / GDP_{i,t−1}, x ∈ {C, I, NX}.
- Private consumption:
  - After a tax rate shock:
    - Steep decline, decreasing by about 1 percent of GDP after about eight quarters, then recovering slightly.
  - After a tax base shock:
    - More muted response; point estimates indicate a decline of about 0.3 percent of GDP a year after announcement, but not statistically significant.
- Private investment (gross fixed capital formation):
  - After a tax rate shock:
    - Investment decreases by about 1 percent of GDP after about six quarters, then partially recovers.
  - After a tax base shock:
    - Point estimates indicate no decline in the first two quarters, and a decline of about 0.5 percent of GDP after about six quarters, though remaining not statistically significant.

### Labor market responses to a 1 percent of GDP tax increase
- Unemployment rate:
  - Tax increases via rate hikes:
    - Significant increase starting at around 0.1 percent in the first quarter.
    - Builds up to about 1 percent after three years.
  - Tax increases via base broadening:
    - Smaller impact; no initial impact on unemployment.
    - Point estimate increases slowly to about 0.6 percent in the medium-term, but estimates remain statistically insignificant in all quarters.
- Average wages per employed worker:
  - Tax rate increases:
    - Average wages decline; point estimates indicate a statistically significant decline of around 2 percentage points after two years, appearing persistent.
  - Tax base increases:
    - No similar decline observed in the first two years; point estimates remain close to zero.
    - Some evidence of a decline in the medium-term after a base shock, but not statistically significant.

### Translation into tax multipliers (overview from 4.2)
- Two multiplier estimation approaches:
  - Contemporaneous multiplier (regresses change in output h quarters ahead on change in tax-to-GDP ratio over same horizon).
  - Cumulative multiplier (regresses cumulative change in output relative to quarter before shock on cumulative change in tax ratio).
- Instrumentation:
  - Change in tax instrumented with the relevant narrative tax shock; baseline controls included.
  - First-stage F-statistics typically large, suggesting no weak instruments problem.
- Key findings (Table (3) summary):
  - Overall tax multiplier (all tax types):
    - About 0.8 to 0.9 after one year (4 quarters).
    - About 1 to 1.5 after two years (8 quarters).
  - Rate versus base multipliers:
    - Rate changes:
      - Multiplier about 1.2 after one year.
      - Reaches between 1.5 and 2 after two years (depending on specification).
    - Base changes:
      - More muted impact on output.
      - Point estimates around 0.3 after one year.
      - Rise to about 0.4 or 0.5 after two years but not statistically different from zero.
  - Difference between rate and base multipliers is robust to methodology; both contemporaneous and cumulative methods deliver similar results.

### Robustness checks (summary from 4.3)
- Excluding all controls:
  - Estimated tax rate multiplier about 1 (versus about 1.3 in baseline).
  - Tax base multiplier remains about 0.3.
  - Estimates less precise without controls.
- Controlling for contemporaneous government spending shocks:
  - Inclusion of narrative annual government spending shocks does not drive differences between rate and base multipliers; estimated multipliers essentially unchanged.
  - Observed government purchases react very little to the tax shock measures.
- Broader definition of narrative shocks (including long-run motivations announced during consolidation years):
  - Estimated multipliers slightly smaller (e.g., 0.7 using all exogenous tax shocks vs. 0.9 baseline), but key findings unchanged.
- Excluding temporary measures:
  - Estimated multipliers essentially unchanged.
- OLS using narrative tax shocks directly:
  - Estimated multipliers slightly larger than baseline IV results; OLS may be biased upward.
  - OLS still finds sizable difference between rate and base multipliers.
- Annual data (instead of quarterly):
  - Large difference between rate and base multipliers persists.
  - Somewhat bigger multipliers compared to baseline.
  - First-stage F-statistics substantially smaller with annual data, which could signal weak-instrument bias.
  - Annual responses: tax base increases have smaller impact on GDP and estimates not significantly different from zero; tax rate increases lead to decline in GDP of 1 percent in the first year and peak decline slightly over 2 percent in the medium-term.
- Overall implication:
  - Using quarterly data reduces concerns about weak instrument bias and improves estimation efficiency.

*Source: wp18220 - 4.1  Dynamic effects of tax changes*

### 5.2  Dynamic effects of specific tax changes

### 5.2  Dynamic effects of specific tax changes

### Personal Income Tax (PIT)
- Shocks scaled to correspond to a 1 percent increase in the effective PIT rate.
- Output response to a PIT rate hike:
  - Fall in output on impact.
  - Builds and peaks at around 2 percent of GDP after about eight quarters.
- Output response to a PIT base increase:
  - Smaller impact than a rate hike.
  - No impact in the initial quarters, and a small subsequent negative impact that remains statistically insignificant.
- Cumulative tax multipliers (cumulative output loss divided by cumulative increase in PIT revenue):
  - One-year cumulative multiplier for a PIT rate shock: 1.3.
  - One-year cumulative multiplier for a PIT base shock: 0.
  - Two-year PIT rate multiplier: 1.5.
  - Two-year PIT base multiplier: 0.4.

### Corporate Income Tax (CIT)
- Shocks scaled to represent a 1 percent increase in the effective CIT rate.
- Output response to a CIT rate hike:
  - Initial decline in output on announcement.
  - Builds up and peaks at about 2.2 percent of GDP after about two years.
  - Medium-term impact is imprecisely estimated and has a large confidence interval.
- Output response to a CIT base increase:
  - Negative but smaller impact on output, typically insignificant in most quarters.
- Comparison to other literature:
  - Mertens and Ravn (2013) find a permanent increase in output of around 0.5 percent following a 1 percent cut in the effective CIT rate (comparable to the output response we estimate for a CIT base shock, but smaller than the response to CIT rate shocks).
  - Estimated impact of CIT shocks on tax revenue is close to zero, implying limited sense in computing tax multipliers for these shocks.

### Value Added Tax (VAT)
- Shocks equivalent to a 1 percent increase in the effective VAT rate.
- Base changes for VAT: few instances exist, making estimation of base-change effects not possible.
- Output response to a VAT rate hike:
  - Small but negative initial impact.
  - Peaks at around 1.8 percent of GDP after seven quarters, then recovers slightly.
- Cumulative multipliers:
  - One-year cumulative multiplier for a VAT rate shock: 2.2.
  - Two-year VAT rate multiplier: 2.6.
- Comparison to literature:
  - Riera-Crichton et al. (2016) finds a one-year multiplier close to 4 for VAT rate shocks.
  - Gunter et al. (2017) find a two-year VAT rate multiplier of 2.1 in industrial countries.
- Note on why VAT multiplier is larger than PIT multiplier:
  - Larger VAT multiplier driven by a slower response of VAT collections in the first six quarters compared to quicker pickup in PIT revenue after a PIT rate hike.

### Synthesis and Policy Implications
- Overall finding:
  - Changes in taxes have significant macroeconomic effects that vary with composition.
  - Tax base changes lead to smaller output and employment declines than tax rate increases.
- Aggregate cumulative multipliers:
  - Estimated cumulative tax rate multiplier one year after announcement: 1.2.
  - Estimated cumulative tax base multiplier one year after announcement: 0.3 (not statistically significant).
  - Multipliers after two years: tax rate multiplier 1.6; tax base multiplier 0.4.
- Policy implication:
  - Evidence lends support to the policy advice that base broadening can raise additional revenue while being less detrimental to growth.

### Directions for Future Research
- Explore variation in rate and base multipliers with initial conditions such as the state of the economy or features of the tax system (e.g., initial tax rate or overall tax structure).
- Expand the dataset to include additional countries and non-consolidation years.
- Classify base measures in more detail to indicate whether they would be expected to increase tax efficiency, enabling analysis of the macroeconomic impact of different types of base measures and identification of those more likely to be beneficial during consolidations.

*Source: IMF Working Paper — section 5.2 "Dynamic effects of specific tax changes".*

### References

### wp18220 - References

### References
- Bibliographic list of cited works on fiscal policy, tax multipliers, narrative approaches, and tax databases, including authors such as Alesina, Auerbach, Barro, Blanchard, Devries, Guajardo, Jordà, Mertens, Nakamura, Ramey, Romer, and others.  
- Topics covered by citations: output effects of fiscal consolidation, measurement of output responses to fiscal policy, narrative tax shock identification, datasets of fiscal consolidation and tax policy measures, corporate tax competition, and methods for estimating impulse responses (local projections).

### A  Data — Dataset construction and definitions
- Coverage:
  - Quarterly dataset for 10 OECD countries from 1975 to 2016.
  - Summary statistics shown in Table (A.2).
  - Country-specific note: for Germany, use West Germany data from 1979 to 1990 (Destatis) and whole Germany from 1991 onwards; regressions include a dummy that takes value one for Germany in all quarters after 1991Q1 inclusive, and zero otherwise.
- Macroeconomic indicators:
  - GDP, employment, and interest rates sourced from OECD Economic Outlook Nr. 102 (November 2017) and OECD.Stat.
  - Investment proxied by gross fixed capital formation.
  - Short-term interest rates: prevailing 90 day or 3-month interbank rates.
  - Long-term interest rates: yields on 10-year government bonds.
  - For Germany, interest rate series from St. Louis FRED database where necessary.
  - Population data from UN Population Statistics, complemented with WEO (for 2016) and Statistisches Jahrbuch der Bundesrepublik Deutschland for West Germany.
- Tax data:
  - Tax revenue by tax type from national statistics offices and tax authorities; checked against annual OECD Tax Revenue dataset.
  - Focus on changes in personal income, corporate income, and value added taxes due to limited cross-country quarterly data for some tax types.
  - Use only actual tax collections; avoid interpolated or imputed quarterly series (resulting tax revenue panel is unbalanced).
- Statutory tax rates:
  - Primary sources: OECD Tax Database for personal and corporate income taxes, European Commission for value added taxes; supplemented with national sources when necessary.
  - Converted to quarterly frequency using exact dates of tax rate changes and convention that a tax rate changes in a quarter if changed at any time during that quarter.
- Tax bases — exact definitions used:
  - VAT Base = Consumption − VAT revenue
    - VAT base defined as household consumption expenditure by resident households exclusive of VAT.
  - PIT Base = Household gross disposable income + Current taxes on income and wealth
    - PIT base includes labor income for dependent workers, self-employment income, income from unincorporated businesses, and capital income paid to households through rents, interest, dividends, and realized capital gains.
    - Social security contributions are excluded from the PIT base.
  - CIT Base = Gross operating surplus of financial and nonfinancial corporations
    - Approximates corporate profits before tax and dividend distribution.
    - Does not subtract consumption of fixed capital from the tax base.
    - Rationale: depreciation methods for CIT differ from national accounts; methodologies not harmonized; measures of depreciation in the corporate sector missing for some countries/years.
  - Note on OECD code 6100: other taxes paid solely by businesses treated as corporate income taxes (affects IRAP in Italy).
- Government debt:
  - Defined as gross government debt outstanding, measured at the federal or general government level.
  - For countries lacking quarterly early-year debt figures, annual series are linearly interpolated.

### Table A.1: Data sources (variable categories and sources)
- Macro variables (GDP, Consumption, Investment, Govt. Purchases, Exports, Imports): OECD.
- Corporate GOS, Household Income, Wages and Salaries: National Sources.
- Interest Rates: OECD EO 102 (Except West Germany).
- Labor Market (Employment, Unemployment, Labor Force, Hours Worked): OECD EO 102 (Except West Germany).
- Population: UN / WEO (Interpolated Annual data).
- Fiscal (Govt. Debt, Tax Revenue): OECD, National Sources.
- Tax Rates: OECD, National Sources, EU Commission.

### B  Other Figures and Tables — Figures, robustness checks, and statistical tables
- Figures (descriptions and notes):
  - Figure A.1: Fiscal responses to rate or base increases (panels for Rate increase and Base increase). Notes: darker (lighter) areas show 68 percent (90 percent) confidence intervals. Standard errors two-way clustered by country and time. Estimates show response to shock equivalent to a 1 percent of GDP tax increase.
  - Figure A.3: Robustness check using annual data (panels for Rate increase, Base increase, GDP). Notes: darker (lighter) areas show 68 percent (90 percent) confidence intervals. Standard errors clustered by country and year. Estimates show response to shock equivalent to a 1 percent of GDP tax increase.
  - Figure A.4: Robustness check using OLS estimation (panels for Rate increase, Base increase on GDP). Notes: darker (lighter) areas show 68 percent (90 percent) confidence intervals. Standard errors clustered by country and year. Estimates show response to shock equivalent to a 1 percent of GDP tax increase.
  - Figure A.5: First-stage responses for specific taxes (panels include PIT Statutory Rate, PIT Effective Rate, CIT Statutory Rate, CIT Effective Rate, VAT Statutory Rate). Notes: darker (lighter) areas show 68 percent (90 percent) confidence intervals. Standard errors two-way clustered by country and time. Responses to shocks equivalent to a 1 percent increase in the effective tax rate.
- Table A.2: Summary Statistics (quarterly sample of 10 OECD countries between 1978 and 2014; values in real per capita 2011 PPP dollars unless noted)
  - Macro
    - GDP p.c.: Mean 33,270; Std. Dev. 7,079; 10% 24,721; 25% 28,110; 50% 32,953; 75% 37,893; 90% 42,282; Obs 1,358
    - Consumption p.c.: Mean 19,086; Std. Dev. 4,567; 10% 13,617; 25% 16,125; 50% 18,900; 75% 21,477; 90% 23,223; Obs 1,358
    - Investment p.c.: Mean 7,079; Std. Dev. 1,943; 10% 4,794; 25% 5,644; 50% 6,823; 75% 8,304; 90% 9,718; Obs 1,358
    - Exports p.c.: Mean 9,622; Std. Dev. 8,265; 10% 3,577; 25% 5,000; 50% 7,690; 75% 10,246; 90% 17,278; Obs 1,358
    - Imports p.c.: Mean 9,262; Std. Dev. 7,216; 10% 3,209; 25% 4,971; 50% 8,016; 75% 10,383; 90% 15,733; Obs 1,358
  - Fiscal
    - Tax p.c.: Mean 8,193; Std. Dev. 2,232; 10% 5,300; 25% 6,391; 50% 8,150; 75% 9,800; 90% 11,045; Obs 1,358
    - Government purchases p.c.: Mean 6,720; Std. Dev. 1,175; 10% 5,175; 25% 5,758; 50% 6,857; 75% 7,649; 90% 8,204; Obs 1,358
    - Government gross debt p.c.: Mean 17,250; Std. Dev. 10,685; 10% 5,614; 25% 9,380; 50% 14,538; 75% 22,952; 90% 34,064; Obs 1,355
  - Personal Income Tax
    - PIT base p.c.: Mean 24,101; Std. Dev. 5,803; 10% 16,978; 25% 20,080; 50% 23,725; 75% 27,356; 90% 30,403; Obs 1,326
    - PIT revenue p.c.: Mean 2,874; Std. Dev. 1,006; 10% 1,474; 25% 2,103; 50% 2,867; 75% 3,591; 90% 4,184; Obs 1,358
    - PIT statutory (top) rate (%): 47 10 34 40 48 53 60; Obs 1,358
    - PIT effective rate (%): 12 4 7 10 12 13 16; Obs 1,326
  - Corporate Income Tax
    - CIT base p.c.: Mean 7,239; Std. Dev. 2,730; 10% 4,274; 25% 5,402; 50% 6,803; 75% 8,541; 90% 10,403; Obs 1,326
    - CIT revenue p.c.: Mean 841; Std. Dev. 439; 10% 351; 25% 514; 50% 770; 75% 1,060; 90% 1,477; Obs 1,358
    - CIT statutory (top) rate (%): 35 10 24 30 34 39 50; Obs 1,358
    - CIT effective rate (%): 12 5 6 8 12 16 20; Obs 1,326
  - Value Added Tax
    - VAT base p.c.: Mean 17,879; Std. Dev. 4,553; 10% 12,784; 25% 15,124; 50% 17,239; 75% 19,565; 90% 22,161; Obs 1,358
    - VAT revenue p.c.: Mean 1,553; Std. Dev. 959; 10% 0879; 25% 1,692; 50% 2,225; 75% 2,687; Obs 1,358
    - VAT statutory (standard) rate (%): (blank entries) 14 80 71 72 0 21; Obs 1,358
    - VAT effective rate (%): 12 4 6 9 12 15 17; Obs 1,102
  - Note: This table presents summary statistics for the quarterly sample of 10 OECD countries between 1978 and 2014. Values are in real per capita 2011 PPP dollars, unless otherwise noted.
- Table A.3: Granger-causality tests by tax type
  - Tests use Equation (1) with four lags; each entry is the F-statistic for the null that the lags of the regressor are jointly equal to zero, with p-values in brackets. Standard errors two-way clustered by country and time. Significance notation: *0.10, **0.05, ***0.01.
  - Selected entries (F-statistics with p-values in parentheses):
    - Personal Income Tax — Rate: 1.86 (0.20); Base: 0.77 (0.57)
    - Corporate Income Tax — Base: 1.53 (0.27); Base (last column) 9.76*** (0.00)
    - Value Added Tax — Rate: 1.32 (0.33); Rate (second column) 3.57* (0.05)
  - Country FE: Yes; Time FE: Yes.
- Table A.4: Joint Granger-causality tests
  - Joint tests using Equation (1) with two or four lags. Dependent variables: narrative tax shocks (Total, Consolidation, Long-Run, Rate, Base). Regressors include output growth, inflation, interest rate, debt ratio, change in tax and government spending ratios, and lags of the narrative tax shock. Standard errors clustered by time. Significance notation: *0.10, **0.05, ***0.01.
  - Results:
    - 2 Lags — Total: 1.53 (0.12); Consolidation: 1.26 (0.25); Long-Run: 0.58 (0.86); Rate: 1.06 (0.40); Base: 1.40 (0.17)
    - 4 Lags — Total: 1.35 (0.14); Consolidation: 1.27 (0.19); Long-Run: 0.42 (0.99); Rate: 0.86 (0.66); Base: 1.39 (0.12)
  - Country FE: Yes; Time FE: Yes.

*Italic: Source document: wp18220 - References*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18220.pdf_