## _wp11158

## Source details

**Canonical URL:** [_wp11158](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11158.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11158.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11158.pdf.json)

---

### Introduction and research question
- Literature is divided on short-term effects of budget deficit reduction.
- Standard Keynesian implication: cutting government spending or raising taxes has contractionary effects on aggregate demand in the short term.
- Alternative theoretical possibility (Blanchard (1990)): fiscal consolidation can be expansionary in the short term if it:
  - Reduces the need for a larger adjustment later, or
  - Signals substantial tax cuts in the future, thereby raising households’ expected future disposable income and increasing investor confidence.
- This phenomenon is referred to as “expansionary fiscal contraction” or “expansionary austerity.”

### Identification concerns with conventional (CAPB-based) approach
- Conventional identification: discretionary changes measured by changes in the cyclically-adjusted primary balance (CAPB).
- Key measurement and identification problems:
  - Nonpolicy changes (e.g., asset price movements, one-off accounting operations) enter CAPB and can correlate with private demand, biasing estimates toward finding expansionary effects.
  - Reverse causality / forward-looking policy responses: CAPB rises may reflect deliberate anti-overheating actions, biasing results.
  - Structural VARs may still ignore nonpolicy changes and forward-looking responses, especially at annual frequency.

### Methodological response: historical (action-based) approach
- Strategy: identify discretionary fiscal changes from contemporaneous policy documents motivated by reducing the budget deficit (plausibly exogenous to near-term business cycle), following Romer and Romer (2010) logic.
- Sources consulted: Budget Speeches, Budgets, central bank reports, IMF Staff Reports, IMF Recent Economic Developments reports, OECD Economic Surveys, country-specific sources.
- Focus on contemporaneous estimates of budgetary impact contained in these sources.
- Emphasized need to rule out:
  - Nonpolicy changes embedded in cyclically-adjusted fiscal measures.
  - Forward-looking policy responses correlated with prospective macroeconomic conditions.

### Dataset and summary statistics
- Sample coverage: 17 OECD countries over 1978-2009: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Netherlands, Portugal, Spain, Sweden, the United Kingdom, and the United States.
- Action-based fiscal adjustments:
  - Total number of fiscal policy changes: 173.
  - Mean budgetary impact: 0.99 percent of GDP.
  - Standard deviation: 0.94 percentage point.
  - Range: –0.75 percent of GDP to 4.74 percent of GDP.
  - Note: Negative values arise when temporary consolidation measures expire (example: a one-year tax of $1 gives budgetary impact $1 then –$1 next year).

### Baseline empirical specification
- Autoregressive panel regression with country and year fixed effects.
- Key regressor: action-based fiscal consolidations (∆Z) measured in percent of GDP.
- Lag order in baseline: 2 (selected by information criteria and serial correlation checks).
- Cumulated impulse responses recover level effects on log real private consumption and log real GDP to a permanent 1 percent of GDP fiscal consolidation.
- Comparison: same specification re-estimated replacing action-based series with change in CAPB-to-GDP ratio.

### Main empirical findings (baseline action-based series)
- A 1 percent of GDP fiscal consolidation reduces real private consumption over the next two years by 0.75 percent (peak within two years; t-statistic = –4.36).
- Real GDP declines by 0.62 percent within two years (t-statistic = –3.82).
- For large fiscal adjustments (> 1.5 percent of GDP):
  - Marginal effect per additional 1 percent of GDP on private consumption: –0.55 percent within two years (t-statistic = –2.97).
  - For GDP the effect is –0.50 percent within two years (t-statistic = –2.88).

### CAPB-based estimates (same sample and specification)
- A 1 percent of GDP rise in the CAPB-to-GDP ratio is associated with an increase in private consumption of 0.37 percent within two years (t-statistic = 4.30).
- The estimated impact on GDP is 0.29 percent within two years (t-statistic = 3.69).
- CAPB-based results for large adjustments are mildly expansionary and not always statistically significant.

### Robustness checks and additional outcomes
- Static model (no lags): private consumption effect –0.83 (standard error 0.15), GDP –0.65 (standard error 0.14).
- Longer lag structure (four lags): private consumption –0.81, GDP –0.64.
- Private investment (logs): baseline effect –1.48 percent (standard error 0.73; t-statistic = –2.02).
- Unemployment rate: baseline rise of 0.32 percentage point (standard error 0.09; t-statistic = 3.44).
- Sensitivity to outliers:
  - Dropping extreme 10 percent: private consumption –0.72 (s.e. 0.22); GDP –0.81 (s.e. 0.20).
  - Dropping observations by Cook’s distance > 4/N: private consumption –0.73 (s.e. 0.15); GDP –0.57 (s.e. 0.13).
  - Excluding one country at a time: peak effects remain significantly contractionary and estimates are within one standard error of baseline (baseline private consumption –0.75*** (0.17); baseline GDP –0.62*** (0.16)).
- Instrumental variables (two-stage least squares) using action-based series as instrument for change in CAPB:
  - First stages are strong (F-statistics on excluded instruments have p-values well below 1 percent).
  - TSLS estimates: a 1 percentage point rise in CAPB-to-GDP ratio (instrumented) reduces private consumption by 0.98 percent within two years (t-statistic = –2.49) and reduces real GDP by 0.80 percent within two years (t-statistic = –2.57).
- Alternative estimation methods:
  - Two-variable VAR: private consumption effect –0.96 percent (t-statistic = –4.72), GDP –0.73 percent (t-statistic = –3.92).
  - Arellano-Bond (1991): private consumption –0.70*** (0.18); GDP –0.62*** (0.17).
- Additional controls (three lags included when used) broadly preserve results:
  - Government debt-to-GDP ratio: private consumption –0.60*** (0.22); GDP –0.58** (0.23).
  - Institutional Investor Ratings (perceived future sovereign default risk): private consumption –0.60*** (0.18); GDP –0.50*** (0.17).
  - Sovereign bond yield (10-year): private consumption –0.70*** (0.17); GDP –0.58*** (0.16).
  - Increase in old-age dependency ratio: private consumption –0.74*** (0.17); GDP –0.61*** (0.16).

### Trade, exchange rate, and external offsetting
- Net exports contribution:
  - A 1 percent of GDP fiscal consolidation raises net exports’ contribution to GDP by 0.47 percentage points of GDP (t-statistic = 4.46).
  - Real exports rise by 0.91 percent (t-statistic = 2.44).
  - Real imports fall by 1.18 percent (t-statistic = –2.92).
- Real effective exchange rate:
  - Depreciates by 1.57 percent (t = –4.13) within one year following a 1 percent of GDP fiscal consolidation.
  - Nominal policy rate falls by 19 basis points (t-statistic = 1.35) following consolidation.
- Consistency with conventional elasticities:
  - Conventional elasticities imply 0.94 percent on exports and 1.22 percent on imports given 1.57 percent real depreciation; close to estimated 0.91 and 1.18 percent respectively.
- Exchange rate regime differences:
  - Pegged regimes: 1 percent of GDP consolidation reduces real GDP by 0.84 percent (t-statistic = –4.15) within two years.
  - Floating regimes: same consolidation reduces real GDP by 0.33 percent (t-statistic = –1.35) within two years.
  - Interpretation: offsetting net-exports channel is less potent in pegged regimes; larger fiscal multipliers in fixed-exchange-rate regimes.

### Composition: spending-based vs tax-based consolidations
- Definitions:
  - “Tax-based”: contribution of tax hikes > contribution of spending cuts.
  - “Spending-based”: contribution of spending cuts > contribution of tax hikes.
- Estimated effects per 1 percent of GDP consolidation (within two years):
  - Spending-based:
    - Private consumption: –0.57 (t-statistic = 2.95).
    - Real GDP: –0.43 (t-statistic = –2.38).
  - Tax-based:
    - Private consumption: –1.35 (t-statistic = –4.52).
    - Real GDP: –1.29 (t-statistic = –4.67).
- Large spending-based consolidations (> 1.5 percent of GDP) show similar results to broader spending-based sample—little evidence of expansionary effects even for large spending-based retrenchments.
- Monetary policy differences:
  - Interest-rate response difference between spending-based and tax-based consolidations is 71 basis points in the first year (t-statistic = 3.38), with greater monetary easing after spending-based adjustments.
  - Difference in real GDP fall (tax-based minus spending-based): 0.33 percentage point in the first year (t-statistic = 1.82) and 0.83 percentage point in the second year (t-statistic = 2.73).
  - To attribute entire first-year output difference to monetary policy: implied effect = 0.33/0.71 = 0.46 percent per 100 basis points; second-year implied effect = 0.83/0.71 = 1.17 percent per 100 basis points.
- Interpretation: much of the difference in output performance appears attributable to different monetary policy responses; results do not support a reverse “balanced-budget multiplier” stimulation.

### Perceived sovereign risk
- Sample split using Institutional Investor Ratings (scale 0 to 100; median = 85) measured in the year before consolidation.
- Estimated effects per 1 percent of GDP consolidation (within two years):
  - High-risk (above-median perceived sovereign credit risk):
    - Private consumption: –0.62 percent (t-statistic = –3.07).
    - Real GDP: –0.53 percent (t-statistic = –2.81).
  - Low-risk (below-median perceived sovereign credit risk):
    - Private consumption: –1.11 percent (t-statistic = –3.65).
    - Real GDP: –0.80 percent (t-statistic = –2.83).
- Interpretation: consolidations preceded by high perceived sovereign default risk are somewhat less contractionary, consistent with confidence or credibility effects mitigating negative impacts; nonetheless effects remain contractionary.

### Aggregate effects and mechanisms summary
- Baseline action-based estimates:
  - A 1 percent of GDP fiscal consolidation reduces real private consumption by 0.75 percent within two years.
  - A 1 percent of GDP fiscal consolidation reduces real GDP by 0.62 percent within two years (t-statistic = –5.42 reported in heading context).
- Net exports partially offset domestic demand contraction; excluding net exports, domestic output contraction is substantially greater.
- Exchange rate depreciation and resulting trade effects explain part of the mitigation.
- Monetary easing and composition of consolidation (spending vs tax) materially affect outcomes.

### Cases and comparative assessment: CAPB vs action-based measures
- Examination of 13 largest disagreements between CAPB-based and action-based approaches:
  - 12 cases where identifiable economic or budgetary developments cause the CAPB-based measure to inaccurately identify the size of deficit-driven fiscal consolidation and largely explain the gap.
  - 1 case (Italy in 1993) where a sharp economic contraction could plausibly have caused the CAPB-based approach to be highly inaccurate.
- Selected case illustrations where CAPB overstates deficit-driven consolidation due to one-offs or cyclical factors:
  - Germany (1996): CAPB rise of 6.8 percentage points in 1995 driven by one-time transfer of Treuhand and East German housing debt.
  - Japan (1999): CAPB rise of 4.9 percentage points driven by one-time capital transfer of about ¥24.3 trillion (4.8 percent of GDP).
  - Finland (2000): CAPB rise of 4.1 percentage points driven by asset-price-related revenue gains and counter-overheating measures.
  - Denmark (1986): CAPB increase of 3.6 percentage points reflects anti-overheating tightening and one-time payment timing.
  - Belgium (1984): CAPB increase 4.7 percentage points, but policy measures amounted to 0.7 percent of GDP; one-time capital transfer explained much of discrepancy.
  - Japan (2006): CAPB rise of 4.1 percentage points but action-based measures of 0.7 percent of GDP; asset operations explain much of CAPB increase.
- Selected case illustrations where CAPB understates deficit-driven consolidation:
  - Ireland (2009): CAPB-to-GDP fall of 4.4 percentage points (Alesina and Ardagna computation) while action-based consolidation measures totaled 4.7 percent of GDP; crisis-driven revenue declines and banking-sector support explain CAPB understatement.
  - Germany (1995): CAPB-to-GDP fall of 6.7 percentage points while action-based consolidation measures worth 1.1 percent of GDP were recorded; one-time capital transfer explains bulk of CAPB fall.
  - Japan (1998): CAPB-to-GDP fall of 5.4 percentage points while action-based measures 0.5 percent of GDP; large accounting operation and stimulus explain CAPB decline.
  - Spain (1993): CAPB-to-GDP fall of 3.3 percentage points; action-based consolidation measures 1.1 percent of GDP; recession and one-off financial transaction explain discrepancy.
  - Italy (1993): CAPB-to-GDP implied consolidation 4.3 percent of GDP (4.5–0.2) with policy records reporting more than 4.0 percent of GDP in measures; sharp contraction associated with ERM crisis likely caused CAPB inaccuracy.
- Authors find no cases where action-based measure appears less accurate.

### Conclusions and suggested extensions
- Using action-based, contemporaneous-document-identified, deficit-driven fiscal adjustments yields consistently negative short-term effects of fiscal consolidation on private domestic demand and GDP.
- CAPB-based approaches can be misled by nonpolicy factors, one-off accounting operations, or policy actions taken in response to cyclical/outlook considerations; instrumenting CAPB with action-based series reverses CAPB-based expansionary results.
- Main conclusion: biases associated with cyclically-adjusted data may be substantial; action-based identification provides evidence that fiscal consolidations are typically contractionary in the short term.
- Suggested extensions:
  - Explore the effect of fiscal consolidation on the current account.
  - Collect data on fiscal expansions motivated by supply-side reforms to identify impacts of long-run growth–oriented policy changes and to increase the sample of policy changes.

*Source: _wp11158 - Section II and related excerpts from the provided IMF working-paper content.*

### References .............................................................................................................

### _wp11158 - References .............................................................................................................

### Introduction and research question
- The literature is divided regarding the short-term effects of budget deficit reduction.
- Standard Keynesian implication: cutting government spending or raising taxes has contractionary effects on aggregate demand in the short term.
- Alternative theoretical possibility (Blanchard (1990)): fiscal consolidation can be expansionary in the short term if it:
  - Reduces the need for a larger adjustment later, or
  - Signals substantial tax cuts in the future, thereby raising households’ expected future disposable income and increasing investor confidence.
- This phenomenon is referred to as “expansionary fiscal contraction” or “expansionary austerity.”

### Empirical background and prior findings
- Seminal empirical contributions report evidence consistent with expansionary fiscal contractions:
  - Giavazzi and Pagano (1990, 1996): fiscal consolidations sometimes correlate with expansions in private consumption within one year; evidence based on case studies and regressions of private consumption on cyclically-adjusted government revenue and spending for a panel of OECD economies.
  - Alesina and Perotti (1997): case studies showing fiscal consolidations sometimes correlate with rapid output growth, particularly when consolidations are implemented by cutting government spending rather than by increasing taxes.
  - Subsequent research, including Alesina and Ardagna (2010), confirms these findings using larger samples of countries and years.

### Identification concerns and potential biases in the conventional approach
- Conventional identification: discretionary changes in fiscal policy measured by changes in the cyclically-adjusted primary balance (CAPB) or other cyclically-adjusted fiscal variables.
- Key measurement and identification problems highlighted:
  - Changes in cyclically-adjusted fiscal variables often include non-policy changes correlated with other developments affecting economic activity. Example: a stock market boom can improve the CAPB by increasing capital gains and cyclically-adjusted tax revenues, while simultaneously affecting private consumption and investment.
  - This measurement error can bias analyses toward downplaying contractionary effects of deliberate fiscal consolidation.
  - Reverse causality concern: a rise in the CAPB may reflect a government’s deliberate decision to raise taxes or cut spending to restrain domestic demand and reduce overheating risk; using CAPB increases to measure consolidation could therefore bias results toward supporting expansionary fiscal contractions.
- Structural VAR approach (e.g., Blanchard and Perotti (2002)):
  - Uses institutional information to distinguish automatic changes from discretionary fiscal changes.
  - Still assumes that, after controlling for lags of output growth, changes in government revenue and spending are uncorrelated with other short-term developments affecting economic activity.
  - Thus it may still ignore nonpolicy changes in cyclically-adjusted fiscal data and forward-looking policy responses to prospective conditions—issues particularly relevant at the annual frequency used in many multi-country studies.

### Methodological response in this paper
- To mitigate the identification problems, the paper examines economic activity following discretionary fiscal changes that historical sources suggest are not correlated with the short-term domestic economic outlook.
- Strategy:
  - Consult a wide range of contemporaneous policy documents to identify fiscal consolidations motivated by a desire to reduce the budget deficit rather than to restrain domestic demand in an overheated economy.
  - Such fiscal actions are viewed, following Romer and Romer (2010), as responses to past decisions and economic conditions rather than to prospective conditions.
  - These actions are therefore unlikely to be systematically correlated with other short-term developments affecting output and are considered valid for estimating the short-term effects of fiscal consolidation on economic activity.

### Notable methodological emphasis
- The paper emphasizes that properly identifying discretionary fiscal consolidations requires ruling out:
  - Nonpolicy changes embedded in cyclically-adjusted fiscal measures.
  - Forward-looking policy responses correlated with prospective macroeconomic conditions.
- The historical-document approach is proposed as a way to obtain more credible identification of discretionary fiscal consolidations for short-term impact analysis.

*Source: _wp11158 - References .............................................................................................................*

### Section II provides further details on how we identify discretionary adjustments in fiscal

### 1.05 percent of GDP (t-statistic = –5.42) within two years. Thus, the effect on total GDP

### _wp11158 - 1.05 percent of GDP (t-statistic = –5.42) within two years. Thus, the effect on total GDP

### Estimated aggregate effects and mechanisms
- Baseline estimated effects (from fiscal actions identified in contemporaneous policy documents):
  - A 1 percent of GDP fiscal consolidation reduces real private consumption by 0.75 percent within two years.
  - A 1 percent of GDP fiscal consolidation reduces real GDP by 0.62 percent within two years (t-statistic = –5.42 reported in heading context).
- Net exports partially offset the contraction in domestic demand; the reported effect on total GDP combines a contraction in domestic demand with a partially offsetting rise in net exports.
- When excluding the rise in net exports, the contraction in domestic output is substantially greater than in the baseline.

### Exchange rate adjustment and trade channel
- Real effective (trade-weighted) exchange rate response:
  - Depreciates by 1.57 percent (t = –4.13) within one year following a 1 percent of GDP fiscal consolidation.
- Short-term composition of the real exchange rate adjustment:
  - Occurs primarily through a fall in the nominal value of the currency (nominal exchange rate), not through an adjustment in domestic prices relative to trading partners (domestic price differential).
- Monetary policy response associated with currency depreciation:
  - Nominal policy rate falls by 19 basis points (t-statistic = 1.35) following consolidation.
- Real export and import responses are broadly consistent with conventional elasticities:
  - Reference conventional elasticities: within two years, a 1 percent real depreciation should raise exports by 0.60 percent and reduce imports by 0.78 percent.
  - Given the estimated real depreciation of 1.57 percent, conventional elasticities imply impacts of 0.94 percent on exports (1.57×0.60) and 1.22 percent on imports (1.57×0.78), close to the paper’s estimated effects of 0.91 percent (exports) and 1.18 percent (imports).

### Role of exchange rate regime (pegged vs floating)
- Aggregate output effects by regime:
  - In pegged exchange rate regimes, a 1 percent of GDP fiscal consolidation reduces real GDP by 0.84 percent (t-statistic = –4.15) within two years.
  - In floating regimes, the same consolidation reduces real GDP by 0.33 percent (t-statistic = –1.35) within two years.
- Exchange rate and trade responses by regime:
  - The real effective exchange rate depreciates more in the short term and real exports expand more in floating regimes than in pegged regimes.
- Interpretation:
  - The offsetting net-exports channel is less potent in economies with pegged exchange rates, consistent with larger fiscal multipliers in fixed-exchange-rate regimes.

### Role of composition: spending-based vs tax-based consolidations
- Definition of types:
  - “Tax-based”: consolidations where the contribution of tax hikes to the adjustment is greater than that of spending cuts.
  - “Spending-based”: consolidations where the contribution of spending cuts is greater than that of tax hikes.
- Estimated effects (per 1 percent of GDP consolidation):
  - Spending-based:
    - Private consumption: –0.57 (t-statistic = 2.95) within two years.
    - Real GDP: –0.43 (t-statistic = –2.38) within two years.
  - Tax-based:
    - Private consumption: –1.35 (t-statistic = –4.52) within two years.
    - Real GDP: –1.29 (t-statistic = –4.67) within two years.
- Large spending-based consolidations (greater than 1.5 percent of GDP) show virtually identical results to the broader spending-based sample—i.e., little evidence of expansionary effects even for large spending-based retrenchments.
- Monetary policy differences and contribution to outcomes:
  - Difference in interest-rate responses between spending-based and tax-based consolidations is 71 basis points in the first year (t-statistic = 3.38), with greater monetary easing after spending-based adjustments.
  - Difference in real GDP fall (tax-based minus spending-based):
    - 0.33 percentage point in the first year (t-statistic = 1.82).
    - 0.83 percentage point in the second year (t-statistic = 2.73).
  - To attribute the entire first-year output difference to monetary policy, a 100 basis point rise in interest rates would need to reduce output by 0.46 percent (0.33/0.71) in the first year; for the second year the implied effect is 1.17 percent (0.83/0.71).
  - These implied interest-rate effects fall within the range of estimates in the empirical literature cited.
- Interpretation:
  - Much of the difference in output performance between spending- and tax-based consolidations appears attributable to different monetary policy responses—central banks ease more after spending-based cuts.
  - Results do not support the reverse “balanced-budget multiplier” notion that equal reductions in spending and taxes would stimulate activity.

### Role of perceived sovereign risk
- Sample split by perceived sovereign default risk (Institutional Investor Ratings index; scale 0 to 100; median = 85):
  - High-risk: above-median perceived sovereign credit risk in the year before consolidation.
  - Low-risk: below-median perceived sovereign credit risk.
- Estimated effects (per 1 percent of GDP consolidation), within two years:
  - High-risk group:
    - Private consumption: –0.62 percent (t-statistic = –3.07).
    - Real GDP: –0.53 percent (t-statistic = –2.81).
  - Low-risk group:
    - Private consumption: –1.11 percent (t-statistic = –3.65).
    - Real GDP: –0.80 percent (t-statistic = –2.83).
- Interpretation:
  - Consolidations preceded by high perceived sovereign default risk are somewhat less contractionary, consistent with confidence or credibility effects that mitigate negative impacts when perceived sovereign risk is high.
  - Even for the high-risk group, the effects are contractionary and not statistically distinguishable from the low-risk group in the sense of reversing contraction.

### Conclusions and extensions
- Main findings:
  - Using fiscal actions identified from contemporaneous policy documents, fiscal consolidation is generally contractionary: a 1 percent of GDP consolidation reduces private consumption by 0.75 percent and real GDP by 0.62 percent within two years.
  - Results are robust across tests and hold even for cases where expansionary effects might be expected (large spending-based consolidations, consolidations in high perceived sovereign risk settings).
  - A fall in the value of the currency and a rise in net exports mitigate some of the contraction; this mitigation is weaker under pegged exchange rate regimes.
- Methodological note:
  - Results contrast with those based on cyclically-adjusted fiscal data (CAPB-based approach), which can bias estimates toward overstating expansionary effects.
- Suggested extensions:
  - Explore the effect of fiscal consolidation on the current account.
  - Collect data on fiscal expansions motivated by supply-side reforms to identify impacts of long-run growth–oriented policy changes and to increase the sample of policy changes.

*Source: Excerpt from the referenced IMF working-paper content provided in the prompt.*

### 6.8 percentage points in 1995. As the 1996 IMF Recent Economic Developments reports (p.

### _wp11158 - 6.8 percentage points in 1995. As the 1996 IMF Recent Economic Developments reports (p.

### Cases where CAPB increases overstate deficit-driven fiscal consolidation
- Germany (1996)
  - CAPB-to-GDP: rise implied by accounting: 6.8 percentage points in 1995 due to a one-time transfer of Treuhand and East German housing debt to the general government amounting to 6.8 percent of GDP (1996 IMF Recent Economic Developments, p. 18).
  - General government deficit: rose to 9.7 percent of GDP in 1995 (OECD Economic Outlook Database No. 88, code NLGQ); in 1996 the deficit fell to 3.3 percent of GDP implying a change of 6.3 percentage points in 1996 (9.7–3.3).
  - Interpretation: the sharp increase in the CAPB in 1996 reflects the end of a one-time large accounting operation rather than deficit-driven fiscal consolidation; CAPB-based measure therefore inaccurately identifies consolidation size.

- Japan (1999)
  - CAPB-to-GDP: rise of 4.9 percentage points.
  - Policy record: no evidence of deficit-reduction measures; fiscal consolidation program initiated in 1997 was suspended end-1998 and no measures to cut the budget deficit until 2002.
  - One-time operation: one-time capital transfer in 1998 to Japan National Railway of about ¥24.3 trillion (4.8 percent of GDP), implying a change in the general government CAPB of about 4.8 percentage points in 1999.
  - Interpretation: CAPB increase driven by one-time capital transfer; discrepancy of 4.9 percent of GDP between CAPB and policy actions.

- Finland (2000)
  - CAPB-to-GDP: increase of 4.1 percentage points.
  - Drivers: exceptional rise in revenue during asset price boom and spending restraint motivated by preventing overheating rather than deficit reduction.
  - Asset price movements: real stock prices rose by 70 percent in 1999 and by 86 percent in 2000 (Haver Analytics Database).
  - Revenue specifics: rise in general government tax revenues on income and wealth of 23 per cent in 2000; extraordinary dividend and property income rise of 1.1 percent of GDP; taxes and proceeds from stock options and capital gains paid by households rose by about EUR 1.1 billion in 2000 (0.83 percent of GDP). Extraordinary revenue gains unrelated to policy: about 1.93 percent of GDP (1.1+0.83).
  - Policy tax cut: deliberate income tax cut cost about EUR 1 billion (0.75 percent of GDP) in 2000.
  - Spending restraint: cuts in 2000 amounted to about 2 percent of GDP; IMF staff estimate fiscal withdrawal in 2000 of some 2 percentage points of GDP.
  - Interpretation: increase in CAPB largely due to cyclical/one-off revenues and countercyclical spending restraint to prevent overheating, not deficit-motivated consolidation; discrepancy of 4.1 percent of GDP.

- Denmark (1986)
  - CAPB-to-GDP: increase of 3.6 percentage points in 1986.
  - Prior measures: deficit-reduction measures in 1985 included one-time bringing forward of payments, producing a negative impact of DKr 5 billion (0.7 percent of GDP) in 1986.
  - Additional factors: fiscal tightening in 1986 to curb overheating (packages in December 1985, mid-March 1986, and October 1986), measures estimated to strengthen budget by some 2½ percent of GDP; large capital gains in 1985 increased corporate tax revenue in 1986.
  - Interpretation: CAPB increase reflects anti-overheating tightening and exceptional revenues; do not classify as deficit-reduction consolidation. Discrepancy: effectively 4.3 percent of GDP (3.6+0.7).

- Belgium (1984)
  - CAPB-to-GDP: increase reported at 4.7 percentage points; OECD CAPB series (code NLGXQA) rises by 4.1 percentage points in 1984.
  - Policy record: fiscal consolidation measures amounting to 0.7 percent of GDP.
  - Accounting/one-off: a one-time capital transfer in 1983 produced an impact on CAPB change of –2.1 percentage points in 1983 and +2.1 percentage points in 1984 when it ended; this accounted for about two-thirds (2.1/3.4) of the remaining discrepancy.
  - Interpretation: CAPB-based measure overstates deficit-driven consolidation; discrepancy of 4.0 percent of GDP (4.7–0.7), with 0.6 percentage points due to CAPB version differences.

- Japan (2006)
  - CAPB-to-GDP: increase of 4.1 percentage points.
  - Policy record: consolidation measures of 0.7 percent of GDP; OECD CAPB series reports only a 3.2 percentage point rise in 2006, giving 0.9 percentage point of discrepancy from CAPB version differences.
  - One-time asset operations: receipts from corporate pension funds, privatization of highway corporations, transfer of reserve fund from Fiscal Loan Fund Special Account improved fiscal balance in 2006 but unrelated to tax hikes or spending cuts. OECD estimates CAPB increase excluding these asset operations at only 0.4 percentage point in 2006.
  - Interpretation: asset operations explain much of CAPB increase; adjusted CAPB close to action-based 0.7 percent of GDP.

- Finland (1988)
  - CAPB-to-GDP: increase of 3.4 percentage points.
  - Drivers: tightening in 1988 motivated by restraining domestic demand and influencing wage expectations; measures included limiting inflation adjustment of tax brackets, allowing no inflation adjustment of tax allowances, raising fuel taxes and indirect taxes, and further indirect tax increases in January 1988; monetary policy tightened.
  - One-off factors: end in 1988 of one-time fiscal stimulus measures introduced in 1987 of about 1 percent of GDP.
  - Interpretation: CAPB increase due to countercyclical tightening to curb overheating and end of one-off stimulus; not classified as deficit-reduction consolidation.

### Cases where deficit-driven fiscal consolidation exceeds CAPB change
- Ireland (2009)
  - CAPB-to-GDP: fall of 4.4 percentage points (Alesina and Ardagna computation); OECD CAPB series (code NLGQA) falls by 2.7 percentage points, leaving 1.7 percentage points due to CAPB version differences.
  - Historical (action-based) consolidation measures: 4.7 percent of GDP.
  - Discrepancy: 9.1 percentage points (4.4+4.7) relative to CAPB fall; remaining unexplained gap 7.4 percentage points after CAPB version differences.
  - Key factors reducing CAPB: sharp recession in 2009 with real stock prices down 44 percent and house prices down 20 percent (Haver Analytics Database); decline in capital taxes by about 55 percent in 2009 to €1.5 billion representing a fall in cyclically-adjusted revenue of 1.2 percent of GDP; cyclically-adjusted indirect tax-to-GDP ratio fell by 1.1 percentage points despite deliberate VAT and excise hikes; discretionary banking-sector support had budgetary impact of 2.5 percent of GDP in 2009 (Anglo Irish Bank support €4 billion, 2.5 percent of GDP).
  - Adjusted CAPB calculation presented: excluding cyclically-adjusted revenue decline and banking support, CAPB-to-GDP ratio rises by 2.1 percentage points (–2.7+1.2+1.1+2.5), indicating a large fiscal consolidation closer to the action-based estimate.
  - Interpretation: cyclically driven revenue declines and crisis-related banking support caused CAPB to understate deficit-driven consolidation; specific economic and budgetary developments explain the discrepancy.

- Germany (1995)
  - CAPB-to-GDP: fall of 6.7 percentage points.
  - Action-based consolidation: policy record indicates consolidation measures worth 1.1 percent of GDP.
  - Discrepancy: 7.8 percent of GDP (6.7+1.1).
  - Driver: one-time capital transfer of Treuhand and East German housing debt to general government produced a change in general government CAPB of –6.8 percentage points in 1995 (1996 IMF Recent Economic Developments, p. 18).
  - Interpretation: one-time transfer explains bulk of CAPB fall; CAPB-based approach inaccurately identifies consolidation size.

- Japan (1998)
  - CAPB-to-GDP: fall of 5.4 percentage points.
  - Action-based consolidation measures: 0.5 percent of GDP.
  - Discrepancy: 5.9 percent of GDP (5.4+0.5).
  - Drivers: assumption by general government of Japan National Railway financial debt of about ¥24.3 trillion (4.8 percent of GDP) in 1998 causing a CAPB change of –4.8 percent of GDP; April 1998 fiscal stimulus package in response to recession also negatively impacted CAPB.
  - Interpretation: large accounting operation and stimulus unrelated to deficit reduction caused CAPB decline; action-based measure better captures deficit-driven consolidation.

- Spain (1993)
  - CAPB-to-GDP: fall of 3.3 percentage points (Alesina and Ardagna computation); OECD CAPB series (code NLGXQA) falls only by 0.6 percent of GDP, so 2.7 percentage points of discrepancy due to CAPB version differences.
  - Action-based consolidation measures: total of 1.1 percent of GDP.
  - Discrepancy: 4.4 percent of GDP (3.3+1.1).
  - Drivers: sharp recession tied to ERM crisis (GDP contracted by 1 percent; unemployment rose about 4 percentage points in 1993); one-time financial transaction with budgetary cost of Ptas 935 billion (1.5 percent of GDP) in 1993 to cover earlier operations.
  - Interpretation: recession-induced CAPB fall and one-off transaction explain why CAPB underestimates consolidation.

- Italy (1993)
  - CAPB-to-GDP: increase of 0.2 percentage points in 1993.
  - Action-based consolidation: large consolidation reported of 4.5 percent of GDP.
  - Discrepancy: 4.3 percent of GDP (4.5–0.2).
  - (Source text truncated at this point; context indicates a substantial divergence between CAPB change and action-based measures for Italy 1993.)

*Italic: Source content from the provided PDF content unit.*

### 4.3 percent of GDP (4.5–0.2). A plausible reason for this discrepancy is that there was a

### _wp11158 - 4.3 percent of GDP (4.5–0.2). A plausible reason for this discrepancy is that there was a

### Discrepancy between CAPB-based and action-based measures
- The CAPB-to-GDP change implied a fiscal consolidation of 4.3 percent of GDP (4.5–0.2).
- A plausible reason for the discrepancy is a sharp economic contraction in 1993 associated with the ERM crisis, which would cause the CAPB-based approach to be inaccurate.
- The CAPB-to-GDP ratio from the OECD Economic Outlook Database No. 88 (code NLGXQA) increases by 1.8 percentage points, consistent with a large fiscal consolidation in 1993.
- Several policy records report larger measures: OECD Economic Surveys 1994 (pp. 44-47), 1993 IMF Recent Economic Developments (pp. 7-8), and the 1993 Banca d’Italia Annual Report (p. 145) all report fiscal consolidation measures in 1993 amounted to more than 4.0 percent of GDP.
- During sharp recessions, cyclical adjustment techniques tend to allocate part of the fiscal worsening due to automatic stabilizers to a fall in the CAPB, causing the increase in the CAPB to understate the size of fiscal consolidation measures.

### Case study: Italy, 1993
- There was a sharp economic contraction in 1993 associated with the ERM crisis.
- This contraction could plausibly have caused the CAPB-based approach to be highly inaccurate, explaining part of the gap between CAPB-based estimates and recorded policy measures exceeding 4.0 percent of GDP.

### Comparative assessment of the two approaches
- Examination of the 13 largest disagreements between the CAPB-based approach and the action-based approach yields:
  - 12 cases where identifiable economic or budgetary developments cause the CAPB-based measure to inaccurately identify the size of deficit-driven fiscal consolidation and largely explain the gap between the two measures.
  - 1 case (Italy in 1993) where a sharp economic contraction could plausibly have caused the CAPB-based approach to be highly inaccurate.
- The authors find no cases where there is evidence that the action-based measure was less accurate.

### Key statistics and counts
- CAPB-to-GDP implied consolidation: 4.3 percent of GDP (4.5–0.2).
- CAPB-to-GDP increase from OECD database: 1.8 percentage points.
- Reported fiscal consolidation measures in 1993 (policy records): more than 4.0 percent of GDP.
- Number of largest disagreements examined: 13.
- Cases where CAPB-based measure misidentified consolidation due to identifiable factors: 12.
- Single case highlighted for sharp contraction: Italy in 1993.

*Source: Excerpt from IMF Working Paper content.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11158.pdf_