## _wp13151

## Source details

**Canonical URL:** [_wp13151](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13151.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13151.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13151.pdf.json)

---

### I. Introduction — context and motivation
- The Great Recession of 2007-09 contributed to a rise in public debt in advanced economies from 70 percent of GDP in 2007 to about 100 percent of GDP in 2011.
- Financial crises typically lead to a deterioration of fiscal positions; Furceri and Zdzienicka (2013) find that the debt-to-GDP ratio has typically increased by about 35 percentage points compared to pre-crisis trends, with the effect lasting for about 10 years.
- Policymakers have pursued spending and tax-based consolidation measures to reduce deficits and debt (IMF Fiscal Monitor, 2012; IMF 2012).
- Prior literature indicates fiscal consolidation is associated with increased poverty and income inequality (Smeeding, 2000; Mulas-Granados, 2005; Agnello and Sousa, 2012; Bova and others, 2013).

### II. Main empirical findings (short- and medium-term distributional effects)
- Sample and episodes:
  - 173 episodes of fiscal consolidation for 17 OECD economies during 1978–2009 (Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, the United Kingdom, and the United States).
  - Magnitude of consolidation episodes ranges between 0.1 and about 5 percent of GDP, with an average of about 1 percent of GDP.
- Key quantitative results:
  - Fiscal consolidation episodes have increased inequality by 0.1 percentage point (about 0.4 percent) in the very short term (1 year after the episode).
  - Fiscal consolidation episodes have increased inequality by 0.9 percentage point (about 3.4 percent) over the medium term.
  - Fiscal consolidation episodes led to a significant and long-lasting fall in the wage income share of about 0.8 percentage point of GDP.
  - Fiscal consolidation episodes raised long-term unemployment by about 0.5 percent over the medium term.
- Descriptive patterns:
  - Cumulative change in the Gini: average increase of about 0.3 percentage point in the short term (two years after) and about 1.5 percentage points in the medium term (8 years after).
  - Wage income share: already on a declining trend, but decreased more rapidly after consolidation episodes.

### III. Data and definitions
- Dependent variables in regressions:
  - Gini coefficient for disposable income from the Standardized World Income Inequality Database (SWIID).
  - Shares of wage and profit in GDP from the OECD Analytical Database.
  - Short-term unemployment (lasting less than six months) and long-term unemployment (lasting more than six months) from the OECD Analytical Database.
- Fiscal consolidation episodes:
  - Taken from Devries et al. (2011) database.
  - Consolidation measured via changes in the cyclically adjusted primary balance (CAPB), focusing on policy actions—tax hikes and/or spending cuts—intended to reduce the budget deficit (narrative approach).
  - Noted limitations of CAPB: measurement errors from cyclical adjustments, failure to strip out asset price or commodity price movements, and inclusion of policy tightenings motivated by demand management rather than consolidation intent.
- Appendix reference:
  - Table A1 contains descriptive statistics and sources.

### IV. Empirical methodology
- Approach:
  - Local projections method (Jorda, 2005) to estimate impulse response functions (IRFs) directly.
  - For each future period k (k = 1,..8), estimated changes in the dependent variable on annual data using Equation (1) with a consolidation dummy P_{i,t} that equals 1 at the start of a consolidation episode.
- Specification details:
  - Country fixed effects included.
  - Time trend included.
  - Number of lags l chosen equal to two; results robust to different lag choices.
  - Estimation using the panel-corrected standard error (PCSE) estimator (Beck and Katz, 1995).
  - IRFs obtained by plotting estimated β_k coefficients for k = 0,1,..8 with confidence bands computed from standard deviations of β_k.
- Identification and robustness:
  - Reverse causality addressed by estimating effects in years following consolidation episodes.
  - Finite sample bias of dynamic panel is in the order of 1/T, where T in the sample is 32.
  - Alternative ARDL approach discussed but not used because ARDL-derived IRFs can be sensitive to lag choice and may impose permanent effects; local projections avoid these problems.
  - Robustness checks for different controls and different lags are reported.

### V. Results — interpretation and heterogeneity
- Average magnitude considered: consolidation episodes of about 1 percent of GDP.
- Distributional impacts:
  - Short-run: Gini up ~0.1 percentage point (about 0.4 percent) one year after the episode.
  - Medium-run: Gini up ~0.9 percentage point (about 3.4 percent).
- Labor-income effects:
  - Wage income share declines by about 0.8 percentage point of GDP, long-lasting.
  - Differential effects observed between wage vs. profit and rent income.
- Unemployment:
  - Long-term unemployment increases by about 0.5 percent over the medium term.
- Spending vs. tax-based consolidations:
  - Spending-based programs have produced larger distributional effects than tax-based programs in the medium term.

### VI. Robustness checks and identification
- Re-estimating Equation (1) with time fixed effects leaves results statistically significant and broadly unchanged.
- Re-estimating Equation (1) without country fixed effects shows the bias is negligible (difference in point estimate small and not statistically significant).
- Augmenting Equation (1) to control for: (i) contemporaneous and past crises episodes (banking and currency crises); (ii) change in economic activity (proxied by real GDP growth); and (iii) change in unemployment confirms robustness of the results.
- Re-estimating Equation (1) for different lags (l) of changes in the Gini coefficient:
  - The medium-term effect ranges from 0.8 percentage point in the case of five lags to about 1 percentage point in the case of zero lags.

### VII. Spending-based versus taxes-based consolidation
- Literature context:
  - Taxes-based consolidations are typically more contractionary than spending-based consolidations (Guajardo et al. (2011) find an effect of -1.3 percent on output after two years for taxes-based programs versus -0.3 percent for spending-based programs).
  - Most direct redistributive impact of fiscal policy in advanced economies has been achieved through the expenditure side—especially non-means-tested transfers (Bastagli et al. 2012).
- Empirical results:
  - Spending and tax-based programs have a similar effect over the short term and the medium term (interpret with caution because most past adjustments involved both types).
  - When separately estimating episodes where taxes-based adjustments have been larger than spending adjustments and vice versa, the medium-term effect on income inequality is:
    - about 1 percentage point for spending-based consolidations;
    - 0.6 percentage point for taxes-based measures.
- Note: the average magnitude of both spending and taxes-based consolidation is about 1 percent of GDP.

### VIII. Wage versus profit and rent income
- Estimating the wage income share (W) shows fiscal consolidation measures typically reduce the wage share of income.
- Fiscal consolidations have a larger negative effect on the level of (inflation-adjusted) wage income than on the level of (inflation-adjusted) profit and rent income.
- Estimating episodes where taxes-based adjustments exceed spending adjustments and vice versa indicates the level of wage income has typically fallen more for spending-based consolidations.
- Possible channels:
  - Direct channel: public sector wage cuts in some plans.
  - Indirect channel: consolidations increase unemployment, particularly the share of long-term unemployed.

### IX. Short-term versus long-term unemployment
- Equations (3) and (4) estimated separately for short-term unemployment (Su) and long-term unemployment (Lu) show:
  - Fiscal consolidations typically lead to a significant and long-lasting increase in long-term unemployment.
  - Fiscal consolidations do not have significant effects on short-term unemployment.
- Implications and risks:
  - Fiscal consolidations add to the burden on the long-term unemployed—a group already large in many OECD countries after the Great Recession.
  - Job loss is associated with persistent earnings loss, adverse health impacts, and declines in children’s academic performance and earnings potential; these adverse impacts are exacerbated the longer unemployment lasts.
  - Long spells of unemployment reduce the odds of being rehired. Example (United States, estimates as of March 2012):
    - a person unemployed for over six months had only a 1 in 10 chance of being rehired in the next month;
    - a person unemployed less than a month had 1 in 3 odds of being rehired in the next month.
  - Long-term unemployment risks entrenching unemployment as a structural problem through “hysteresis” (Blanchard and Summers, 1986).
  - Long-term unemployment also threatens social cohesion; surveys show negative political effects and increased desire for a rogue leader, particularly for the long-term unemployed.

### X. Main findings, scope, and sample
- Using episodes of fiscal consolidation for a sample of 17 OECD countries over the period 1978-2009, the paper finds that fiscal consolidation has typically led to:
  - a significant and persistent increase in inequality;
  - declines in wage income and in the wage share of income;
  - increases in long-term unemployment.
- Robustness: results are robust across specifications with time fixed effects, exclusion of country fixed effects, additional controls for crises and macro variables, and different lag structures.

### XI. Policy implications
- Fiscal consolidation has non-negligible and persistent distributional consequences: increased income inequality, reduced wage shares, and higher long-term unemployment.
- The composition of consolidation matters: spending-based adjustments are associated with larger adverse distributional outcomes than tax-based adjustments.
- Measurement and identification caveats:
  - CAPB-based measures can misclassify episodes due to cyclical adjustment errors and failure to capture policymakers’ intent; narrative approaches focusing on policy actions can mitigate some problems.
  - Robustness checks and endogeneity analyses support the validity of the estimated effects.
- Recommended approaches:
  - Approve fiscal measures that reduce deficits in the future—when global recovery is more robust—for example, linking statutory retirement ages to life expectancy and improving entitlement program efficiency.
  - Avoid unduly hasty consolidations that pose risks to the recovery; countries with scope should opt for a slower pace of consolidation combined with policies to support growth.
  - Fiscal consolidation plans should specify how policies respond to shocks (for example, shielding unemployment benefits from cuts if growth is slower than assumed), permitting some flexibility while credibly preserving medium-term consolidation objectives.

*Source: _wp13151 - References................................................................................................. 12*

### References................................................................................................. 12

### _wp13151 - References................................................................................................. 12

### I. Introduction — context and motivation
- The Great Recession of 2007-09 contributed to a rise in public debt in advanced economies from 70 percent of GDP in 2007 to about 100 percent of GDP in 2011.
- Financial crises typically lead to a deterioration of fiscal positions; Furceri and Zdzienicka (2013) find that the debt-to-GDP ratio has typically increased by about 35 percentage points compared to pre-crisis trends, with the effect lasting for about 10 years.
- Policymakers have pursued spending and tax-based consolidation measures to reduce deficits and debt (IMF Fiscal Monitor, 2012; IMF 2012).
- Prior literature indicates fiscal consolidation is associated with increased poverty and income inequality (Smeeding, 2000; Mulas-Granados, 2005; Agnello and Sousa, 2012; Bova and others, 2013).

### II. Main empirical findings (short- and medium-term distributional effects)
- Sample and episodes:
  - 173 episodes of fiscal consolidation for 17 OECD economies during 1978–2009 (Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, the United Kingdom, and the United States).
  - Magnitude of consolidation episodes ranges between 0.1 and about 5 percent of GDP, with an average of about 1 percent of GDP.
- Key quantitative results:
  - Fiscal consolidation episodes have increased inequality by 0.1 percentage point (about 0.4 percent) in the very short term (1 year after the episode).
  - Fiscal consolidation episodes have increased inequality by 0.9 percentage point (about 3.4 percent) over the medium term.
  - Fiscal consolidation episodes led to a significant and long-lasting fall in the wage income share of about 0.8 percentage point of GDP.
  - Fiscal consolidation episodes raised long-term unemployment by about 0.5 percent over the medium term.
- Patterns observed in descriptive analysis:
  - Cumulative change in the Gini: average increase of about 0.3 percentage point in the short term (two years after) and about 1.5 percentage points in the medium term (8 years after).
  - Wage income share: already on a declining trend, but decreased more rapidly after consolidation episodes.

### III. Data and definitions
- Dependent variables in regressions:
  - Gini coefficient for disposable income from the Standardized World Income Inequality Database (SWIID).
  - Shares of wage and profit in GDP from the OECD Analytical Database.
  - Short-term unemployment (lasting less than six months) and long-term unemployment (lasting more than six months) from the OECD Analytical Database.
- Fiscal consolidation episodes:
  - Taken from Devries et al. (2011) database.
  - Consolidation measured via changes in the cyclically adjusted primary balance (CAPB), focusing on policy actions—tax hikes and/or spending cuts—intended to reduce the budget deficit (narrative approach).
  - Noted limitations of CAPB: measurement errors from cyclical adjustments, failure to strip out asset price or commodity price movements, and inclusion of policy tightenings motivated by demand management rather than consolidation intent.
- Appendix reference:
  - Table A1 contains descriptive statistics and sources.

### IV. Empirical methodology
- Approach:
  - Local projections method (Jorda, 2005) to estimate impulse response functions (IRFs) directly.
  - For each future period k (k = 1,..8), estimated changes in the dependent variable on annual data using Equation (1) with a consolidation dummy P_{i,t} that equals 1 at the start of a consolidation episode.
- Specification details:
  - Country fixed effects included.
  - Time trend included.
  - Number of lags l chosen equal to two; results robust to different lag choices.
  - Estimation using the panel-corrected standard error (PCSE) estimator (Beck and Katz, 1995).
  - IRFs obtained by plotting estimated β_k coefficients for k = 0,1,..8 with confidence bands computed from standard deviations of β_k.
- Identification and robustness:
  - Reverse causality addressed by estimating effects in years following consolidation episodes.
  - Finite sample bias of dynamic panel is in the order of 1/T, where T in the sample is 32.
  - Alternative ARDL approach discussed but not used because ARDL-derived IRFs can be sensitive to lag choice and may impose permanent effects; local projections avoid these problems.
  - Robustness checks for different controls and different lags are reported (figures referenced).

### V. Results — interpretation and heterogeneity
- Average magnitude considered: consolidation episodes of about 1 percent of GDP.
- Distributional impacts:
  - Short-run: Gini up ~0.1 percentage point (about 0.4 percent) one year after the episode.
  - Medium-run: Gini up ~0.9 percentage point (about 3.4 percent).
- Labor-income effects:
  - Wage income share declines by about 0.8 percentage point of GDP, long-lasting.
  - Differential effects observed between wage vs. profit and rent income (figures referenced).
- Unemployment:
  - Long-term unemployment increases by about 0.5 percent over the medium term.
- Spending vs. tax-based consolidations:
  - Spending-based programs have produced larger distributional effects than tax-based programs (figures and robustness checks referenced).

### VI. Policy implications and interpretation
- Fiscal consolidation has non-negligible and persistent distributional consequences: increased income inequality, reduced wage shares, and higher long-term unemployment.
- The composition of consolidation matters: spending-based adjustments are associated with larger adverse distributional outcomes than tax-based adjustments.
- Measurement and identification caveats:
  - CAPB-based measures can misclassify episodes due to cyclical adjustment errors and failure to capture policymakers’ intent; narrative approaches focusing on policy actions can mitigate some problems.
  - Robustness checks and endogeneity analyses support the validity of the estimated effects.

*Source: _wp13151 - References................................................................................................. 12*

### 3.4 percent) in the medium term—8 years after the occurrence of the consolidation episode.

### Distributional effects of fiscal consolidation

### Robustness checks and identification
- Re-estimating Equation (1) with time fixed effects to control for specific time shocks (such as those affecting world interest rates) leaves results statistically significant and broadly unchanged.
- Re-estimating Equation (1) without country fixed effects (to address the potential bias from the interaction of fixed effects and country-specific arrival rates of consolidation episodes) shows the bias is negligible (the difference in the point estimate is small and not statistically significant).
- Augmenting Equation (1) to control for: (i) contemporaneous and past crises episodes (banking and currency crises); (ii) change in economic activity (proxied by real GDP growth); and (iii) change in unemployment confirms robustness of the results.
- Re-estimating Equation (1) for different lags (l) of changes in the Gini coefficient:
  - The medium-term effect ranges from 0.8 percentage point in the case of five lags to about 1 percentage point in the case of zero lags.

### Spending-based versus taxes-based consolidation
- Literature context: taxes-based consolidations are typically more contractionary than spending-based consolidations (Guajardo et al. (2011) find an effect of -1.3 percent on output after two years for taxes-based programs versus -0.3 percent for spending-based programs).
- Most direct redistributive impact of fiscal policy in advanced economies has been achieved through the expenditure side—especially non-means-tested transfers (Bastagli et al. 2012).
- Empirical results (Equation (1) estimated separately for taxes and spending-based adjustments):
  - Spending and tax-based programs have a similar effect over the short term and the medium term (interpret with caution because most past adjustments involved both types).
  - When separately estimating episodes where taxes-based adjustments have been larger than spending adjustments and vice versa, the medium-term effect on income inequality is:
    - about 1 percentage point for spending-based consolidations;
    - 0.6 percentage point for taxes-based measures.
- Note: the average magnitude of both spending and taxes-based consolidation is about 1 percent of GDP.

### Wage versus profit and rent income
- Equation (2) estimated for the share of wage income in GDP (W) shows fiscal consolidation measures typically reduce the wage share of income.
- Fiscal consolidations have a larger negative effect on the level of (inflation-adjusted) wage income than on the level of (inflation-adjusted) profit and rent income.
- Estimating Equation (2) for episodes where taxes-based adjustments exceed spending adjustments and vice versa indicates the level of wage income has typically fallen more for spending-based consolidations.
- Possible channels:
  - Direct channel: public sector wage cuts in some plans.
  - Indirect channel: consolidations increase unemployment, particularly the share of long-term unemployed.

### Short-term versus long-term unemployment
- Equations (3) and (4) estimated separately for short-term unemployment (Su) and long-term unemployment (Lu) show:
  - Fiscal consolidations typically lead to a significant and long-lasting increase in long-term unemployment.
  - Fiscal consolidations do not have significant effects on short-term unemployment.
- Implications and risks:
  - Fiscal consolidations add to the burden on the long-term unemployed—a group already large in many OECD countries after the Great Recession.
  - Job loss is associated with persistent earnings loss, adverse health impacts, and declines in children’s academic performance and earnings potential; these adverse impacts are exacerbated the longer unemployment lasts.
  - Long spells of unemployment reduce the odds of being rehired. Example (United States, estimates as of March 2012):
    - a person unemployed for over six months had only a 1 in 10 chance of being rehired in the next month;
    - a person unemployed less than a month had 1 in 3 odds of being rehired in the next month.
  - Long-term unemployment risks entrenching unemployment as a structural problem through “hysteresis” (Blanchard and Summers, 1986).
  - Long-term unemployment also threatens social cohesion; surveys show negative political effects and increased desire for a rogue leader, particularly for the long-term unemployed.

### Main findings, scope, and sample
- Using episodes of fiscal consolidation for a sample of 17 OECD countries over the period 1978-2009, the paper finds that fiscal consolidation has typically led to:
  - a significant and persistent increase in inequality;
  - declines in wage income and in the wage share of income;
  - increases in long-term unemployment.
- Robustness: results are robust across specifications with time fixed effects, exclusion of country fixed effects, additional controls for crises and macro variables, and different lag structures.

### Policy implications
- Governments should pay attention to the composition and design of fiscal measures to lessen distributional impacts or offset them through other measures.
- Distributional effects of consolidation must be balanced against potential longer-term benefits (e.g., lower interest payments permitting cuts to distortionary taxes).
- Recommended approaches:
  - Approve fiscal measures that reduce deficits in the future—when global recovery is more robust—for example, linking statutory retirement ages to life expectancy and improving entitlement program efficiency.
  - Avoid unduly hasty consolidations that pose risks to the recovery; countries with scope should opt for a slower pace of consolidation combined with policies to support growth.
  - Fiscal consolidation plans should specify how policies respond to shocks (for example, shielding unemployment benefits from cuts if growth is slower than assumed), permitting some flexibility while credibly preserving medium-term consolidation objectives.

*Source: _wp13151 - 3.4 percent) in the medium term—8 years after the occurrence of the consolidation episode.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13151.pdf_