## 1. Distribution of Tax Reforms

## Source details

**Canonical URL:** [1. Distribution of Tax Reforms](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019298-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019298-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019298-print-pdf.pdf.json)

---

### I. Introduction and motivation
- Context: Large fiscal deficits and concerns about debt sustainability in many advanced economies after the Global Financial Crisis increased pressure to adopt fiscal consolidation measures.
- Conventional wisdom and tradeoff:
  - Fiscal adjustment is costly for elected leaders because voters penalize them at the polls.
  - Tax-based consolidations are associated with larger output declines than spending-based measures.
- Political heterogeneity:
  - Benefits and costs of spending cuts and tax increases are rarely evenly distributed across the electorate; heterogeneous impacts make consolidation politically costly and can explain delayed fiscal adjustments.
- Key empirical questions:
  1. How do tax reforms affect the probability of reelection of the incumbent government party or its leader?
  2. How do electoral costs vary depending on the type and design of tax policy measures?
  3. How do initial political and economic conditions underpinning tax consolidations impact reelection outcomes?

### II. Data and empirical approach
- Dataset and sample:
  - Narrative dataset of tax-based fiscal consolidations compiled by Dabla-Norris and Lima (2018) at a quarterly frequency.
  - Ten advanced economies: Australia, Austria, Canada, Germany, France, Italy, Portugal, Spain, the United Kingdom, and the United States.
  - Coverage: 1980Q1–2016Q4 (last forty years); precise announcement and implementation dates, magnitudes, and motivations of reforms.
- Tax reform classification and measures:
  - Grouped into personal (PIT), corporate (CIT), and indirect tax reforms (mainly value added, sales, and excises).
  - Reform indicator R: total expected revenue yield (in percent of GDP) at announcement date; cumulative expected yields used for multiyear reforms.
- Estimation strategy:
  - Baseline panel probit with dependent variable y_it = incumbent party or leader reelected (binary).
  - Main regressor T_it−k = direct and indirect tax reforms announced during current term (before reelection).
  - Controls Z_it: contemporaneous and lagged changes in structural fiscal balance, real GDP per capita growth (two years before election), initial public debt level, dummy for financial and debt crises, political orientation (left/right/center).
  - Balanced panel: 10 countries, 1988Q1–2016Q4; cluster standard errors at time level; include election-fixed effects.
- Robustness and extensions:
  - Separate tax types (CIT, PIT, indirect); interact reform effects with political/economic conditions; IV approach; specifications restricted to election quarters; alternative databases (Alesina, Favero, and Giavazzi 2015).

### III. Main findings — baseline results
- Overarching result:
  - Tax reforms are politically costly; tax-based consolidations significantly lower the probability of reelection of the incumbent government party and its leader.
- Baseline marginal effects (Table 1):
  - A one percentage point of GDP tax consolidation significantly lowers probability of reelection of the incumbent government by about 8 percentage points.
  - Leader reelection probability declines by about 7 percentage points for a similar reform.
  - A 1-percentage point tax consolidation lowers the likelihood of maintaining similar popular support in the next elections (statistically significant).
- Other control variable marginal effects (Table 1):
  - Structural fiscal balance (change over 2 years): -0.041*** (ruling party), -0.214*** (party leader), -0.009*** (number of votes).
  - Initial public debt level (percent of GDP): -0.141** (ruling party), -0.047 (party leader), 0.065*** (number of votes).
  - Financial crises: -0.079* (ruling party), 0.049 (party leader), 0.051*** (number of votes).
  - Real GDP per capita (log, change over 2 year): 0.059*** (ruling party), 0.085*** (party leader), -0.011*** (number of votes).
  - Political orientation (1-right; 0-left): 0.041 (ruling party), 0.011 (party leader), -0.052*** (number of votes).
  - Observations: 860 (ruling party and leader), 480 (number of votes). (Pseudo) R-squared: 0.074, 0.050, 0.030 respectively.
- Interpretation:
  - Improvement in structural fiscal balance during incumbent’s term lowers reelection probability.
  - A 1 percentage point increase in real GDP per capita in two years before reelection increases likelihood of ruling party reelection by about 5 percentage points.
  - Voters penalize ruling party when there is a financial crisis during incumbent’s term.

### IV. Robustness checks — battery of tests
- Endogeneity and selection:
  - Strong governments (first-election vote share ≥ 35 percent) engage more in tax-based consolidation than weak governments.
  - Table 2 frequency of tax reforms — Strong: 0.81, Weak: 0.19; only consolidation reforms — Strong: 0.79, Weak: 0.21. Implication: bias likely towards zero; true negative electoral effects could be larger than baseline estimates.
- Controls and alternative estimators:
  - Controlling for spending-based consolidations: baseline results broadly unchanged.
  - IV approach (tax policy reforms instrument for tax-to-GDP ratio): IV results broadly align with baseline.
  - Election-quarter-only specification: results broadly similar though sample reduces.
- Focus on tax increases:
  - One percentage point of GDP tax-based consolidation (tax increases) lowers reelection probability by 24 percentage points (Table 3, column V).
  - Large announced tax increases (top fifty percent) have more detrimental effects: -0.259*** reported (Table 3, column VI).
- Additional robustness:
  - Results robust to controls for political support, cumulative GDP growth per capita over term, GDP growth in election year, initial tax system characteristics, implementation lags, changes in government spending.
  - Analysis repeated using Alesina, Favero, and Giavazzi (2015) database for 16 countries, 1980–2014: findings robust (Annex, Table A4).

### V. Which tax reforms are costlier? — design and composition
A. Reform types (Table 4)
- Marginal effects:
  - Indirect tax consolidation: a 1 percentage point of GDP indirect tax consolidation lowers reelection probability by about 12 percentage points.
  - Direct tax consolidation: about 6 percentage points (direct tax coefficient -0.072**).
  - CIT reforms: lower reelection probability by about 20 percentage points (coefficient -0.207**).
  - PIT reforms: coefficient -0.023 (not statistically significant).
  - CIT impact statistically different from PIT at the 10 percent level; PIT effects also statistically different from indirect tax effects.
- Interpretation:
  - CIT effect potentially large because corporate interest groups are better organized and politically strong.
  - Non-significant PIT effect may reflect offsetting measures (allowances, deductions, tax credits).

B. Reform composition — frontloaded versus backloaded (Table 5)
- Definitions:
  - Frontloaded multi-year reform: at least 50 percent of expected revenue yield occurs in first year; one-year reforms considered frontloaded.
- Marginal effects (selected):
  - Frontloaded tax reforms (all types): -0.148**.
  - Non-frontloaded tax reforms (all types): -0.077**.
  - Frontloaded CIT: -0.222 (not significant given standard error).
  - Frontloaded PIT: -0.463**.
  - Frontloaded indirect tax: -0.260**.
  - Non-frontloaded CIT: -0.191; non-frontloaded PIT: -0.028; non-frontloaded indirect tax: -0.115**.
  - Observations: 860; (Pseudo) R-squared range 0.080–0.098.
- Interpretation:
  - Frontloaded reforms are generally costlier.
  - Frontloaded PIT reforms lower reelection probability by over 40 percentage points.

C. Motivation behind reforms (Table 6)
- Coding:
  - Motivations: consolidation (reduce existing deficits/debt), long-term growth, other (administrative).
- Marginal effects:
  - Consolidation-motivated tax reforms: coefficient -0.483*** (a one percentage point of GDP consolidation lowers reelection probability by about 45 percentage points).
  - Long-term growth-motivated tax reforms: coefficient -0.094 (not economically significant).
  - Observations: 860; (Pseudo) R-squared: 0.090.
- Interpretation:
  - Consolidation tax reforms entail substantially larger electoral penalties than reforms framed as improving long-term growth.

### VI. Timing, political conditions, and economic context
A. Timing of reforms (Table 7)
- Definition:
  - Dummy = 1 when lag between announcement and election date is at least 8 quarters (2 years) before the election; 0 otherwise.
- Marginal effects (selected):
  - Tax Reforms Closer to Reelection:
    - All Reforms: -0.155** (0.064)
    - CIT: -0.308** (0.136)
    - PIT: -0.160* (0.094)
    - Indirect Tax: -0.328*** (0.104)
  - Tax Reforms Further from Reelection:
    - All Reforms: -0.041 (0.027)
    - CIT: -0.199* (0.114)
    - PIT: -0.012 (0.347)
    - Indirect Tax: -0.090 (0.060)
  - Observations: 860; (Pseudo) R-squared: All Reforms 0.182, CIT 0.077, PIT 0.078, Indirect Tax 0.071.
- Interpretation:
  - PIT reforms have no impact when announced ≥8 quarters before the new mandate but are heavily penalized when announced in the run-up to the next election.
  - Indirect tax reforms have larger impacts close to elections.
  - CIT consolidations are costly regardless of timing.

B. Political conditions (Table 8 and Table 9)
- Political capital (share of votes at first election; split at country-specific median):
  - Tax Reforms with Lower Political Support:
    - All Reforms: -0.171*** (0.038)
    - CIT: -0.271** (0.107)
    - PIT: -0.268** (0.027)
    - Indirect Tax: -0.207** (0.056)
  - Tax Reforms with Higher Political Support:
    - All Reforms: -0.038 (0.054)
    - CIT: -0.129 (0.199)
    - PIT: -0.029 (0.045)
    - Indirect Tax: -0.012 (0.136)
  - Observations: 860.
  - Interpretation: Electoral costs are not material for governments with strong political capital.
- Political orientation (dummy = 1 for left-wing governments; 0 for right-wing) — marginal effects:
  - Left-wing governments:
    - All Reforms: -0.040 (0.063)
    - CIT: -0.213** (0.101)
    - PIT: 0.063 (0.073)
    - Indirect Tax: -0.137** (0.060)
  - Right-wing governments:
    - All Reforms: -0.098*** (0.030)
    - CIT: -0.324* (0.169)
    - PIT: -0.107* (0.060)
    - Indirect Tax: -0.164 (0.116)
  - Observations: 860.
  - Interpretation:
    - CIT reforms lower reelection probability by about 32 percentage points when government is right-wing and by about 20 percentage points for left-wing governments.
    - PIT reforms lower probability by 10 percentage points for right-wing governments; effect insignificant for left-wing governments.

C. Economic factors — recessions and tax progressivity (Table 10 and Table 11)
- Recessions (GDP growth below 25th percentile):
  - Tax Reforms implemented during Recessions:
    - All Reforms: -0.215** (0.109)
    - CIT: -0.213 (0.264)
    - PIT: -0.365** (0.061)
    - Indirect Tax: -0.378 (0.267)
  - Tax Reforms during Normal Times:
    - All Reforms: -0.041* (0.0198)
    - CIT: -0.073 (0.049)
    - PIT: -0.043 (0.034)
    - Indirect Tax: -0.058 (0.035)
  - Observations: 860.
  - Interpretation: Probability of reelection is significantly lower when reforms are announced during recessions; PIT reforms particularly penalized (about 36 percentage points).
- Tax progressivity (PIT progressivity indicator):
  - High progressivity (above 25th percentile):
    - All Reforms: -0.034* (0.021)
    - PIT: -0.040 (0.037)
  - Low progressivity (below 25th percentile):
    - All Reforms: -0.355*** (0.126)
    - PIT: -0.395*** (0.087)
  - Observations: 860.
  - Interpretation: Electoral penalty for tax reforms is higher in countries with low levels of tax progressivity; PIT reforms in low-progressivity countries reduce likelihood of reelection by around 40 percentage points.

### VII. Conclusions and implications
- Summary of empirical conclusions:
  - Tax-based consolidations reduce public support for incumbent governments; political costs vary by reform type, timing, motivation, and political/economic context.
  - Indirect tax reforms generally entail larger electoral costs than direct tax measures.
  - Within direct taxes, CIT reforms carry higher electoral costs than PIT changes.
  - Frontloaded consolidations, especially frontloaded PIT reforms, are more costly than gradual reforms.
  - Consolidation-motivated tax increases (aimed at lowering deficits and debt) entail substantially larger electoral costs than reforms framed as improving long-term growth.
  - Electoral costs are, on average, lower for left-wing governments, for governments with stronger political capital, and if announced earlier in the election cycle.
  - Reforms announced during recessions and in countries with low tax progressivity are penalized most severely.
- Policy implications and design considerations:
  - Indirect taxes and reforms affecting organized interest groups carry higher electoral costs and may require complementary measures (timing, communication, compensatory policies) to mitigate political backlash.
  - Early-in-mandate reforms and attention to political capital can reduce electoral risk for certain reform types, notably PIT reforms.
  - Policymakers face tradeoffs between necessary fiscal consolidation and electoral risks of tax-based measures.
- Suggested avenues for future research (as stated in the source):
  - Role of structural factors such as labor market and product rigidities and their interaction with tax reforms.
  - Differential impact of tax base versus rate measures targeting different voter segments.
  - Interaction between PIT reforms, redistribution, and inequality.

### VIII. Selected summary statistics (Annex A2 and annex tables)
- Elections and reelection counts:
  - Number of Elections (actual number): 89
  - Government Party Reelected (actual number): 49
  - Leader Reelected (actual number): 39
  - Government Party Reelected with at least the same number of votes as in the previous elections (actual number): 27
  - # of Left-Wing Government Ruling Parties per Country (actual number): 41
  - # of Right-Wing Government Ruling Parties per Country (actual number): 48
- Other variables:
  - Public Support (Number of Votes of Government Ruling Party at the Parliament at the Time of the Reform, in quarters): Mean 40.94, Standard Deviation 5.89
  - Lag between Reform Announcement Tax Reforms and Reelections (in quarters): Mean 8.27, Standard Deviation 5.95
- Tax reforms (means and standard deviations, in percent of GDP):
  - All Reforms:
    - CIT reforms: Mean 0.09, Standard Deviation 0.77
    - PIT reforms: Mean 0.06, Standard Deviation 1.33
    - Indirect tax reforms: Mean 0.54, Standard Deviation 0.99
    - Total tax reforms: Mean 0.69, Standard Deviation 1.70
  - Only Tax Increases:
    - CIT reforms: Mean 0.40, Standard Deviation 0.79
    - PIT reforms: Mean 0.72, Standard Deviation 0.91
    - Indirect tax reforms: Mean 0.90, Standard Deviation 0.95
    - Total tax reforms: Mean 1.32, Standard Deviation 1.47
- Other economy and policy indicators (means and standard deviations):
  - PIT Progressivity (percent): Mean 25.8, Standard Deviation 9.3
  - Real GDP Growth (percent): Mean 2.1, Standard Deviation 1.9
  - Real GDP Growth (change over 2 years): Mean 1.32, Standard Deviation 1.39
  - Public debt-to-GDP (percent): Mean 53.5, Standard Deviation 30.5
  - Structural balance-to-GDP (change over 2 years): Mean -0.21, Standard Deviation 2.52
- Alternative database results (Annex A3, A4, A5, A6 highlights):
  - Alesina, Favero and Giavazzi (2015) Database (Observations: 339):
    - All Reforms (I): -0.247** (0.124)
    - Direct Tax (II): -0.159 (0.105)
    - Indirect Tax (IV): -0.276*** (0.103)
  - Differential impact: Tax increases versus tax decreases (Observations: 860):
    - Tax increases — All Reforms (I): -0.165*** (0.058); PIT (III): -0.228*** (0.071); Indirect Tax (IV): -0.229*** (0.073)
    - Tax decreases — All Reforms (I): -0.052 (0.073); PIT (III): 0.007 (0.086)
  - Alternative performance measures (Observations: 860/854):
    - Tax Reforms marginal effects: (I) -0.084*** (0.029); (II) -0.086*** (0.029); (III) -0.090*** (0.029)
  - Robustness to clustering choices (Observations: 860) yields broadly consistent marginal effects around -0.084*** (standard errors reported in annex).

*Source: IMF Working Paper — "1. Distribution of Tax Reforms" (extracted content).*

### 1. Distribution of Tax Reforms ____________________________________________________________________ 9

### 1. Distribution of Tax Reforms

### I. Introduction and Motivation
- Context: Large fiscal deficits and concerns about debt sustainability in many advanced economies after the Global Financial Crisis increased pressure to adopt fiscal consolidation measures.
- Conventional wisdom: Fiscal adjustment is costly for elected leaders because voters penalize them at the polls.
- Economic tradeoff: Fiscal consolidation measures improve stabilization and growth in the long-term but can entail short-term economic costs. Evidence suggests tax-based consolidations are associated with larger output declines than spending-based measures.
- Political heterogeneity: Benefits and costs of spending cuts and tax increases are rarely evenly distributed across the electorate; heterogeneous impacts make consolidation politically costly and can explain delayed fiscal adjustments.
- Prior evidence and debate:
  - Some studies find voters do not systematically punish governments for consolidations (Alesina, Carloni, and Lecce 2013).
  - Other work finds tax-based consolidations have larger electoral impacts than spending-based consolidations (Alesina et al., forthcoming).
  - Political consequences likely depend on composition of tax reforms and on political and economic conditions (Ilzetzki 2018).
- Key empirical questions addressed:
  1. How do tax reforms affect the probability of reelection of the incumbent government party or its leader?
  2. How do electoral costs vary depending on the type and design of tax policy measures?
  3. How do initial political and economic conditions underpinning tax consolidations impact reelection outcomes?

### II. Data and Empirical Approach
- Dataset: A narrative dataset of tax-based fiscal consolidations compiled by Dabla-Norris and Lima (2018) at a quarterly frequency.
- Sample: 10 advanced economies (Australia, Austria, Canada, Germany, France, Italy, Portugal, Spain, the United Kingdom, and the United States).
- Coverage: Last forty years, with precise announcement and implementation dates of direct and indirect tax changes, magnitudes of announced and implemented reforms, and motivations behind each tax policy action.
- Estimation strategy: Exploits the granularity of reform timing (quarterly data) and maps tax policy reforms to electoral outcomes; estimations include all quarters, not only election quarters, to exploit granularity and avoid aggregation of tax reform indicators.
- Robustness checks: Also test specification restricted to election quarters (Table 2 referenced in source) and a battery of additional robustness tests and alternative specifications (detailed in annexes).

### III. Main Findings
- Overarching result:
  - Tax reforms are politically costly; tax-based consolidations significantly lower the probability of reelection of the incumbent government party and its leader.
- Heterogeneity of costs:
  - Broad-based indirect tax reforms and reforms affecting specific interest groups (e.g., corporate income tax reforms) are associated with higher electoral costs.
  - Consolidation tax measures primarily aimed at reducing existing fiscal deficits and debt are costlier than tax increases motivated by long-term growth or administrative efficiency considerations.
- Timing and design effects:
  - Political costs are lower for reforms implemented earlier in the government’s mandate when the incumbent government has higher political capital.
  - Electoral outcomes are materially impacted by frontloaded consolidations, particularly in the case of personal income taxes (PIT).
- Role of political support and initial economic conditions:
  - Electoral costs of tax-based consolidations are higher if the ruling coalition is right-wing.
  - Electoral costs are higher if the government has weak public support.
  - Electoral costs are higher when tax reforms are implemented during recessions.
  - Personal income tax changes are electorally salient if reforms are announced during recessions, and if the tax system is less progressive.
- Robustness:
  - Results are robust to a battery of robustness tests and alternative specifications.

### IV. Implications and Interpretive Points
- Policy tradeoffs: Policymakers face a tradeoff between necessary fiscal consolidation and the electoral risks of implementing tax-based measures, with the political cost varying by tax type, timing, motivation, and political/economic context.
- Design considerations:
  - Indirect taxes and reforms affecting organized interest groups carry higher electoral costs and may require complementary measures (timing, communication, compensatory policies) to mitigate political backlash.
  - Early-in-mandate reforms and attention to political capital dynamics can reduce electoral risk for certain reform types, notably PIT reforms.
- Political economy: The partisan orientation of the ruling coalition and prevailing macroeconomic conditions (e.g., recession) importantly condition the electoral consequences of tax reforms.

*Source: IMF Working Paper — 1. Distribution of Tax Reforms (extracted content).*

### Section III presents the baseline results and a battery of robustness checks. Section IV examines

### Section III: Baseline Results and Robustness Checks

### II. Empirical approach and data
- Tax reform dataset:
  - Narrative database of tax reforms identified by Dabla-Norris and Lima (2018) for 10 OECD countries, 1980Q1–2016Q4.
  - Contains announcement and implementation dates of individual direct and indirect tax reforms, expected annual revenue yields using forecasts produced by country authorities, and specific motivation (e.g., long-term growth, administrative considerations, consolidation).
  - Announcement date defined as first formal proposal (senior official communication or draft legislation); implementation date is first moment revenue impacts are non-zero.
- Tax reform classification:
  - Grouped into personal (PIT), corporate (CIT), and indirect tax reforms (mainly value added, sales, and excises).
  - Magnitude, composition (frontloaded versus backloaded), and timing recorded.
- Strategies to minimize endogeneity:
  - Focus on consolidation reforms introduced for reasons other than current stabilization.
  - Assess direction of bias from governments implementing reforms based on initial public support (see Section III.B).
  - Include battery of controls and robustness checks; distinguish consolidation versus long-term growth motivations.
  - Additional robustness: use annual fiscal consolidation measures constructed by Alesina, Favero, and Giavazzi (2015).

### Reform episodes and summary statistics
- Reform indicator R:
  - Takes value of the total expected revenue yield (in percent of GDP) at announcement date; zero otherwise.
  - Uses cumulative expected revenue yields for multiyear reforms; announcement date of first reform in a package used when reforms are announced simultaneously.
- Distribution and summary:
  - Among tax consolidation measures: 33 percent are PIT reforms, 30 percent CIT reforms, and 30 percent indirect tax consolidations.
  - Average expected revenue yield is about 0.7 percent of GDP; around 1.3 percent of GDP for tax increases (Annex A2).
  - Expected revenue yields of indirect tax consolidations are slightly higher than for direct tax reforms.
  - Almost half of consolidation reforms aimed to address long term structural challenges; reforms triggered by high deficit and debt concerns accounted for 38 percent of consolidation episodes.
  - Majority of tax reforms announced shortly after previous elections; frequency decreases closer to reelection date.

### Reelection outcomes and sample
- Three reelection outcome definitions:
  - Binary indicator: incumbent ruling (majority) government party reelected (1) or not (0); dummy applies for all quarters between “first election” and reelection.
  - Leader reelection: similar indicator for leader of ruling party, including only observations where leader ran for reelection.
  - Vote share: government party reelected with same or larger number of votes (1) versus reelected with lower number of votes (0).
- Data sources: Global Elections Database, Database of Political Institutions, National Election Database (Annex A1).
- Sample size and frequencies:
  - Total sample: 89 election episodes.
  - Government ruling party reelected about 55 percent of the time.
  - Party leader reelected about 44 percent of the time.

### B. Empirical framework
- Baseline panel probit specification (Equation (1)):
  - Dependent variable y_it = political dummy (incumbent party or leader reelected = 1).
  - Main regressor: T_it−k = direct and indirect tax reforms announced during current term (before reelection).
  - Controls Z_it include contemporaneous and lagged changes in structural fiscal balance, real GDP per capita growth during the term (specifically two years before election), initial public debt level, dummy for financial and debt crises, political orientation (left/right/center).
  - Balanced panel: 10 countries, 1988Q1–2016Q4.
  - Cluster standard errors at time level; include election-fixed effects e_t.
- Extensions:
  - Equation (2): separate tax reforms into types (CIT, PIT, indirect) to estimate differential impacts (β1,...,βN).
  - Equation (3): interact reform effects with political/economic conditions using binary split D_it^l / D_it^h where indicator is below/above prevailing average (examples: political orientation, vote share, business cycle).

### III.A Baseline results — Tax reforms are politically costly
- Main baseline marginal effects (Table 1):
  - A one percentage point of GDP tax consolidation significantly lowers probability of reelection of the incumbent government by about 8 percentage points.
  - Leader reelection probability declines by about 7 percentage points for a similar reform.
  - A 1-percentage point tax consolidation lowers the likelihood of maintaining similar popular support in the next elections (statistically significant).
- Other control variable effects (Table 1):
  - Structural fiscal balance (change over 2 years): coefficient -0.041*** (ruling party), -0.214*** (party leader), -0.009*** (number of votes).
  - Initial public debt level (percent of GDP): -0.141** (ruling party), -0.047 (party leader), 0.065*** (number of votes).
  - Financial crises: -0.079* (ruling party), 0.049 (party leader), 0.051*** (number of votes).
  - Real GDP per capita (log, change over 2 year): 0.059*** (ruling party), 0.085*** (party leader), -0.011*** (number of votes).
  - Political orientation (1-right; 0-left): 0.041 (ruling party), 0.011 (party leader), -0.052*** (number of votes).
  - Observations: 860 (ruling party and leader), 480 (number of votes).
  - (Pseudo) R-squared: 0.074, 0.050, 0.030 respectively.
- Interpretation:
  - Improvement in structural fiscal balance (fiscal contraction) during incumbent’s term lowers reelection probability.
  - A 1 percentage point increase in real GDP per capita in two years before reelection increases likelihood of the ruling party being reelected by about 5 percentage points.
  - Voters penalize ruling party when there is a financial crisis during incumbent’s term.

### III.B Robustness checks — battery of tests
- Dependent variable focus: likelihood government party reelected for the remainder of analyses.
- Endogeneity bias:
  - Strong governments (first-election vote share ≥ 35 percent) engage more in tax-based consolidation than weak governments.
  - Table 2: frequency of tax reforms — Strong: 0.81, Weak: 0.19; only consolidation reforms — Strong: 0.79, Weak: 0.21.
  - Implication: bias likely towards zero; true negative electoral effects could be larger than baseline estimates.
- Spending-based consolidations:
  - Control for change in fiscal balance and include announced spending-based consolidations during incumbent’s term; baseline results remain broadly unchanged (Table 3, column II).
- Measurement error:
  - Use IV approach (tax policy reforms instrument for tax-to-GDP ratio) following Mertens and Ravn (2012) and Dabla-Norris and Lima (2018); IV results broadly align with baseline (Table 3, column III).
- Sample structure:
  - Re-estimate using only election quarters (drop non-election quarters); results broadly similar but sample reduces substantially (Table 3, column IV).
- Focus on tax increases:
  - Separate measures aimed at increasing tax revenues; results (Table 3, column V) show a one percentage point of GDP tax-based consolidation lowers reelection probability by 24 percentage points.
  - Large announced tax increases (top fifty percent) have more detrimental effects (Table 3, column VI): -0.259*** reported.
- Additional controls:
  - Include political support for ruling party, cumulative GDP growth per capita over term, GDP growth in election year, initial tax system characteristics, implementation lags, changes in government spending; results broadly consistent with baseline.
  - Repeat analysis using tax reforms identified by Alesina, Favero, and Giavazzi (2015) for panel of 16 countries, 1980–2014; findings robust (Annex, Table A4).

### IV. Which tax reforms are costlier? (design and composition)
A. Reform types: PIT, CIT, and indirect tax reforms
- Section findings (Table 4):
  - Indirect tax consolidation: a 1 percentage point of GDP indirect tax consolidation lowers reelection probability of incumbent government party by about 12 percentage points.
  - Direct tax consolidation effect substantially smaller: about 6 percentage points (direct tax reforms coefficient -0.072**).
  - CIT reforms: lower reelection probability by about 20 percentage points (coefficient -0.207**).
  - PIT reforms: coefficient -0.023 (not statistically significant).
  - Effects: CIT impact statistically different from PIT at the 10 percent level; PIT effects also statistically different from indirect tax effects.
- Interpretation:
  - Larger effect of CIT possibly due to corporate interest groups being better organized and politically strong.
  - Non-significant PIT effect may reflect offsetting measures (allowances, deductions, tax credits, special treatments) that target specific groups.
  - Appendix evidence: indirect tax and PIT increases hold political costs, while tax decreases are benign electorally; for indirect taxes, voters reward lower taxes.

B. Reform composition (frontloaded vs backloaded)
- Definition:
  - Frontloaded multi-year reform: at least 50 percent of expected revenue yield occurs in first year; all one-year reforms considered frontloaded.
- Results (Table 5):
  - Frontloaded tax reforms (all types): -0.148** marginal effect on reelection probability.
  - Non-frontloaded tax reforms (all types): -0.077**.
  - Frontloaded CIT: -0.222 (not significant given standard error reported).
  - Frontloaded PIT: -0.463**.
  - Frontloaded indirect tax: -0.260**.
  - Non-frontloaded CIT: -0.191; non-frontloaded PIT: -0.028; non-frontloaded indirect tax: -0.115**.
  - Observations: 860 for all columns; (Pseudo) R-squared range 0.080–0.098.
- Interpretation:
  - Frontloaded reforms are generally costlier.
  - Especially salient: frontloaded PIT reforms lower reelection probability by over 40 percentage points.
  - PIT represents between a third and a half of national income in OECD countries (Piketty and Saez, 2012); three quarters of PIT collection borne on labor income—which may explain large electoral costs when frontloaded and visible.

C. Motivation behind reforms
- Three-step approach:
  1. Use Dabla-Norris and Lima (2018) motivation coding: long-term growth vs consolidation (lower existing deficits/debt) vs other.
  2. Focus on tax increases among these reforms.
  3. Test whether consolidation-motivated tax measures entail larger electoral costs than other motivations.
- Results (Table 6):
  - Consolidation-motivated tax reforms: coefficient -0.483*** (a one percentage point of GDP consolidation lowers reelection probability by about 45 percentage points).
  - Long-term growth-motivated tax reforms: coefficient -0.094 (not economically significant).
  - Observations: 860; (Pseudo) R-squared: 0.090.
  - Test for difference in coefficients between consolidation and long-term growth: Wald ch2-statistic (7.8), significant at *** level.
- Interpretation:
  - Consolidation tax reforms have substantially larger electoral penalties than reforms presented as improving long-term growth.
  - Impact of other (administrative/other) measures on electoral outcomes not statistically significant (results available upon request).
  - Implementation lag tests (whether announcement before and implementation after reelection matters) were not statistically significant (results available upon request).

*Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019298-print-pdf.pdf*

### introduction impact electoral outcomes. First, we analyze the role of political capital and political

### introduction impact electoral outcomes. First, we analyze the role of political capital and political

### A. Timing of Reforms
- Literature: timing of reforms in the electoral cycle affects election outcomes; linked to electoral myopia and higher political support earlier in incumbent’s term.
- Empirical test: dummy = 1 when lag between announcement and election date is at least 8 quarters (2 years) before the election; 0 otherwise.
- Key findings (Table 7, marginal effects):
  - Tax Reforms Closer to Reelection:
    - All Reforms: -0.155** (0.064)
    - CIT: -0.308** (0.136)
    - PIT: -0.160* (0.094)
    - Indirect Tax: -0.328*** (0.104)
  - Tax Reforms Further from Reelection:
    - All Reforms: -0.041 (0.027)
    - CIT: -0.199* (0.114)
    - PIT: -0.012 (0.347)
    - Indirect Tax: -0.090 (0.060)
  - Observations: 860 across columns; Controls: Yes; Time FE: Yes; Clustering: Yes.
  - (Pseudo) R-squared:
    - All Reforms: 0.182
    - CIT: 0.077
    - PIT: 0.078
    - Indirect Tax: 0.071
  - Wald test ch2-statistics (in parentheses) reported for differences; Note: * p< 0.10, ** p<0.05, ***p<0.01.
- Interpretation:
  - PIT reforms have no impact on reelection odds when announced ≥8 quarters before the new mandate but are heavily penalized (probability of reelection falls by almost 15 percentage points) when announced in the run-up to the next election.
  - Indirect tax reforms have larger impacts close to elections.
  - CIT consolidation reforms are costly regardless of timing.

### B. Role of Political Conditions
- Political Capital
  - Proxy: share of votes incumbent government party has in Parliament after the first elections; government has more (less) political capital if share is above (below) the country-specific median.
  - Dummy 퐷 indicates less (1) vs more (0) political capital.
  - Key findings (Table 8, marginal effects):
    - Tax Reforms with Lower Political Support:
      - All Reforms: -0.171*** (0.038)
      - CIT: -0.271** (0.107)
      - PIT: -0.268** (0.027)
      - Indirect Tax: -0.207** (0.056)
    - Tax Reforms with Higher Political Support:
      - All Reforms: -0.038 (0.054)
      - CIT: -0.129 (0.199)
      - PIT: -0.029 (0.045)
      - Indirect Tax: -0.012 (0.136)
    - Observations: 860; Controls: Yes; Time FE: Yes; Clustering: Yes.
    - (Pseudo) R-squared:
      - All Reforms: 0.083
      - CIT: 0.078
      - PIT: 0.078
      - Indirect Tax: 0.080
    - Wald test ch2-statistics in parentheses; note significance levels as above.
  - Interpretation:
    - Electoral costs are not material for governments with strong political capital.
    - PIT reforms do not affect election outcomes when the government has strong policy support.

- Political Orientation and Preferences
  - Test: dummy = 1 for left-wing governments; 0 for right-wing.
  - Key findings (Table 9, marginal effects):
    - Reforms Implemented by Left-wing Governments:
      - All Reforms: -0.040 (0.063)
      - CIT: -0.213** (0.101)
      - PIT: 0.063 (0.073)
      - Indirect Tax: -0.137** (0.060)
    - Reforms Implemented by Right-wing Governments:
      - All Reforms: -0.098*** (0.030)
      - CIT: -0.324* (0.169)
      - PIT: -0.107* (0.060)
      - Indirect Tax: -0.164 (0.116)
    - Observations: 860; Controls: Yes; Time FE: Yes; Clustering: Yes.
    - (Pseudo) R-squared:
      - All Reforms: 0.18
      - CIT: 0.173
      - PIT: 0.078
      - Indirect Tax: 0.179
    - Wald test ch2-statistics reported in parentheses; significance levels noted.
  - Interpretation and magnitudes:
    - CIT reforms (column II) lower probability of reelection by about 32 percentage points when government is right-wing and by 20 percent, on average, for left-wing governments.
    - Indirect tax increases announced by left-wing governments are electorally costly.
    - PIT reforms lower probability of reelection by 10 percentage points for right-wing governments; effect statistically insignificant for left-wing governments.

### C. Role of Economic Factors
- Economic Conditions (Growth Regimes)
  - Recession defined when GDP growth is below 25th percentile of country-specific distribution; dummy 퐷 = 1 when below 25th percentile.
  - Key findings (Table 10, marginal effects):
    - Tax Reforms implemented during Recessions:
      - All Reforms: -0.215** (0.109)
      - CIT: -0.213 (0.264)
      - PIT: -0.365** (0.061)
      - Indirect Tax: -0.378 (0.267)
    - Tax Reforms implemented during “Normal Times”:
      - All Reforms: -0.041* (0.0198)
      - CIT: -0.073 (0.049)
      - PIT: -0.043 (0.034)
      - Indirect Tax: -0.058 (0.035)
    - Observations: 860; Controls: Yes; Time FE: Yes; Clustering: Yes.
    - (Pseudo) R-squared: 0.1430 (All Reforms), 0.1420 (CIT), 0.1430 (PIT), 0.143 (Indirect Tax).
    - Wald test ch2-statistics in parentheses; significance levels noted.
  - Interpretation:
    - Probability of reelection is significantly lower when reforms are announced during recessions.
    - For PIT reforms, likelihood of reelection is reduced by about 36 percentage points when implemented during recessions.

- Tax Progressivity
  - PIT progressivity indicator: 1 - (100 - marginal tax rate)/(100 - average tax rate); identify high PIT progressivity countries as above 25th percentile.
  - Test: dummy 퐷 = 1 when PIT progressivity indicator is above 25th percentile.
  - Key findings (Table 11, marginal effects):
    - Tax Reforms in Countries with Higher Tax Progressivity:
      - All Reforms: -0.034* (0.021)
      - PIT: -0.040 (0.037)
    - Tax Reforms in Countries with Lower Tax Progressivity:
      - All Reforms: -0.355*** (0.126)
      - PIT: -0.395*** (0.087)
    - Observations: 860; Controls: Yes; Time FE: Yes; Clustering: Yes.
    - (Pseudo) R-squared:
      - All Reforms: 0.4348
      - PIT: 0.4276
    - Wald test ch2-statistics in parentheses; significance levels noted.
  - Interpretation:
    - Electoral penalty for tax reforms is higher in countries with low levels of tax progressivity.
    - PIT reforms in low-progressivity countries reduce likelihood of reelection by around 40 percentage points.

### VI. Conclusions
- Overall findings:
  - Tax-based consolidations affect public support for governments; political costs depend on reform design and political/economic conditions.
  - Indirect tax reforms generally entail larger electoral costs than direct tax measures.
  - Within direct taxes, electoral costs are higher for CIT reforms than for PIT changes.
  - Frontloaded consolidations, especially for PIT reforms, are more costly than gradual reforms.
  - Consolidation measures aimed at lowering deficits and debt entail larger electoral costs than those aimed at improving long-term growth prospects.
  - Electoral costs are, on average, lower for left-wing governments, for governments with stronger political capital, and if announced earlier in the election cycle.
  - Reforms undertaken in the earlier period of an incumbent’s term do not affect election prospects.
  - Voters penalize governments for tax consolidations during recessions (particularly PIT reforms) and in countries with low tax progressivity.
- Suggested avenues for future research (as stated in the source):
  - Role of structural factors such as labor market and product rigidities and their interaction with tax reforms.
  - Differential impact of tax base versus rate measures targeting different voter segments.
  - Interaction between PIT reforms, redistribution, and inequality.

*Source: wpiea2019298-print-pdf - introduction impact electoral outcomes. First, we analyze the role of political capital and political*

### Annex A2. Selected Summary Statistics

### Annex A2. Selected Summary Statistics

### Summary statistics (Table A2)
- Number of Elections (actual number): 89
- Government Party Reelected (actual number): 49
- Leader Reelected (actual number): 39
- Government Party Reelected with at least the same number of votes as in the previous elections (actual number): 27
- # of Left-Wing Government Ruling Parties per Country (actual number): 41
- # of Right-Wing Government Ruling Parties per Country (actual number): 48

Other variables
- Public Support (Number of Votes of Government Ruling Party at the Parliament at the Time of the Reform, in quarters): Mean 40.94, Standard Deviation 5.89
- Lag between Reform Announcement Tax Reforms and Reelections (in quarters): Mean 8.27, Standard Deviation 5.95

Tax reforms (means and standard deviations, in percent of GDP)
- All Reforms:
  - CIT reforms: Mean 0.09, Standard Deviation 0.77
  - PIT reforms: Mean 0.06, Standard Deviation 1.33
  - Indirect tax reforms: Mean 0.54, Standard Deviation 0.99
  - Total tax reforms: Mean 0.69, Standard Deviation 1.70
- Only Tax Increases:
  - CIT reforms: Mean 0.40, Standard Deviation 0.79
  - PIT reforms: Mean 0.72, Standard Deviation 0.91
  - Indirect tax reforms: Mean 0.90, Standard Deviation 0.95
  - Total tax reforms: Mean 1.32, Standard Deviation 1.47

Other economy and policy indicators
- PIT Progressivity (percent): Mean 25.8, Standard Deviation 9.3
- Real GDP Growth (percent): Mean 2.1, Standard Deviation 1.9
- Real GDP Growth (change over 2 years): Mean 1.32, Standard Deviation 1.39
- Public debt-to-GDP (percent): Mean 53.5, Standard Deviation 30.5
- Structural balance-to-GDP (change over 2 years): Mean -0.21, Standard Deviation 2.52

*Source: Annex A2 Table A2.*

### Effects of tax reforms on reelection — Alternative database (Annex A3, Table A3)
- Data: Alesina, Favero and Giavazzi (2015) Database; Observations: 339
- Marginal effects of a one percentage point of GDP tax consolidation on the probability of reelection:
  - All Reforms (I): -0.247** (0.124)
  - Direct Tax (II): -0.159 (0.105)
  - Indirect Tax (IV): -0.276*** (0.103)
- Controls: Yes
- Time FE: Yes
- Clustering: Yes
- (Pseudo) R-squared: 0.097 (All Reforms), 0.103 (Direct Tax)
- Test for the difference in coefficients: (0.7)
- Note: * p< 0.10, ** p<0.05, *** p<0.01. Coefficients interpreted as impact of a one percentage point of GDP tax consolidation on probability of reelection. Wald test ch2-statistics in parentheses.

### Differential impact: Tax increases versus tax decreases (Annex A4, Table A4)
- Observations: 860 (for All Reforms, CIT, PIT, Indirect Tax)
- Marginal effects (impact of announcement of tax increases or decreases on probability of reelection):
  - Tax increases:
    - All Reforms (I): -0.165*** (0.058)
    - CIT (II): -0.125* (0.070)
    - PIT (III): -0.228*** (0.071)
    - Indirect Tax (IV): -0.229*** (0.073)
  - Tax decreases:
    - All Reforms (I): -0.052 (0.073)
    - CIT (II): -0.111 (0.082)
    - PIT (III): 0.007 (0.086)
    - Indirect Tax (IV): -0.166* (0.095)
- Controls: Yes; Time FE: Yes; Clustering: Yes
- (Pseudo) R-squared: 0.074 (All Reforms), 0.069 (CIT), 0.075 (PIT), 0.079 (Indirect Tax)
- Tests for difference in coefficients (Wald test ch2-statistics in parentheses):
  - All Reforms: (1.70) — ** indicates significance at p<0.05
  - CIT: (0.02)
  - PIT: (5.84)
  - Indirect Tax: (0.31)
- Note: * p< 0.10, ** p<0.05, *** p<0.01. Results based on Equation (3) with reform indicator =1 for tax increases and zero otherwise.

### Alternative economic performance measures (Annex A5, Table A5)
- Observations and specifications:
  - Columns (I), (II), (III)
  - Observations: 860 (I), 860 (II), 854 (III)
  - Time FE: Yes; Clustering by time: Yes
  - (Pseudo) R-squared: 0.071 (I), 0.071 (II), 0.067 (III)
- Marginal effects (impact of a one percentage point of GDP tax consolidation on probability of reelection):
  - Tax Reforms:
    - (I): -0.084*** (0.029)
    - (II): -0.086*** (0.029)
    - (III): -0.090*** (0.029)
  - Structural fiscal balance (change 2 years):
    - (I): -0.041*** (0.006)
    - (II): -0.039*** (0.006)
    - (III): -0.035*** (0.006)
  - Initial public debt level (percent of GDP):
    - (I): -0.141** (0.062)
    - (II): -0.178** (0.062)
    - (III): -0.189** (0.061)
  - Financial crises:
    - (I): -0.079* (0.492)
    - (II): -0.105** (0.053)
    - (III): -0.126** (0.053)
  - Real GDP per capita (log, change over 2 year):
    - (II): 0.059*** (0.012)
    - (III): 0.059*** (0.012)
  - Real GDP growth in the election year:
    - (III): 0.034*** (0.009)
  - Real GDP growth 6 month ahead of the elections:
    - (III): 0.041*** (0.013)
  - Political Orientation of the Government (1-right; 0-left):
    - (I): 0.041* (0.284)
    - (II): 0.050* (0.290)
    - (III): 0.064* (0.028)
- Note: * p< 0.10, ** p<0.05, *** p<0.01.

### Robustness: Alternative standard errors clustering (Annex A6, Table A6)
- Observations: 860 in all specifications
- Time FE: Yes
- Clustering variations and marginal effects:
  - Clustering by time (Column I):
    - Tax Reforms: -0.084*** (0.029)
    - Structural fiscal balance (change 2 years): -0.041*** (0.006)
    - Initial public debt level (percent of GDP): -0.141** (0.062)
    - Financial crises: -0.079* (0.492)
    - Real GDP per capita (log, change over 2 year): 0.059*** (0.012)
    - Political Orientation of the Government (1-right; 0-left): 0.041 (0.284)
    - (Pseudo) R-squared: 0.074
  - Clustering by time and elections (Column II):
    - Tax Reforms: -0.084*** (0.028)
    - Structural fiscal balance (change 2 years): -0.041*** (0.006)
    - Initial public debt level (percent of GDP): -0.014** (0.062)
    - Financial crises: -0.079 (0.050)
    - Real GDP per capita (log, change over 2 year): 0.059*** (0.012)
    - Political Orientation of the Government (1-right; 0-left): 0.420 (0.029)
    - (Pseudo) R-squared: 0.074
  - Clustering by time and country (Column III):
    - Tax Reforms: -0.084*** (0.027)
    - Structural fiscal balance (change 2 years): -0.041*** (0.007)
    - Initial public debt level (percent of GDP): -0.014** (0.062)
    - Financial crises: -0.079 (0.062)
    - Real GDP per capita (log, change over 2 year): 0.059*** (0.014)
    - Political Orientation of the Government (1-right; 0-left): 0.422 (0.054)
    - (Pseudo) R-squared: 0.074
- Note: * p< 0.10, ** p<0.05, *** p<0.01. Coefficients interpreted as impact of a one percentage point of GDP tax consolidation on probability of reelection.

*Source: Annex A2–A6, "Selected Summary Statistics" and accompanying tables from the provided content.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019298-print-pdf.pdf_