## wpiea2023178-print-pdf

## Source details

**Canonical URL:** [wpiea2023178-print-pdf](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023178-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023178-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023178-print-pdf.pdf.json)

---

### Inflation surge and macro context
- World average annual inflation increased from 3.2 percent in 2020 to 8.7 percent in 2022.
- Advanced economies: inflation increased from 0.7 percent to 7.3 percent between 2020 and 2022.
- Emerging market and developing economies: inflation increased from 5.2 percent to 9.8 percent between 2020 and 2022.
- Euro area inflation increased from 0.3 percent in 2020 to 8.4 percent in 2022, having averaged only 1 percent between 2015 and 2019.
- US Federal Reserve hiking cycle started in March 2022 with 12-month headline inflation already exceeding 8.5 percent; over the following 14 months the Federal Reserve increased its policy rate by 500 basis points.
- ECB hiking cycle started in July 2022 with 12-month headline inflation exceeding 8 percent; by June 2023 it had tightened policy rates by a cumulative 400 basis points.
- April 2023 IMF World Economic Outlook forecast:
  - Global headline inflation projected to decrease to 6.6 percent year-on-year in Q2 2023.
  - Inflation excluding food and energy declining to 5.9 percent.
  - In the United States and the euro area, the corresponding inflation Charts are 4.6 percent and 5.1 percent, and 6.3 percent and 6.5 percent, respectively.

### Unconventional fiscal policy (UFP) measures in Europe — description and costs
- Many European countries implemented fiscal measures (transfers, energy subsidies, tax cuts) to contain energy prices for households and businesses.
- Types of measures:
  - Price-suppressing measures: caps on energy retail prices; cuts to excise duties or value-added tax rates on energy products; block tariffs.
  - Non–price-suppressing measures: energy vouchers, lump-sum income tax credits, cash transfers.
- Targeting and scale:
  - Most measures were largely untargeted to prioritize speed and reach.
  - Price-suppressing and untargeted measures account for about 45 percent of fiscal outlays for households and SMEs in EU countries.
  - Targeted measures account for about 24 percent of total fiscal measures in the euro area.
  - Projected spending for 2023 is 2.0 percent of GDP.
  - Total measures include fiscal support targeted at households and SMEs at 2.7 percent of GDP.
  - Fiscal costs budgeted for energy support to firms amounted to 0.6 percent of GDP in the euro area.
  - Estimated size of overall energy fiscal support aggregating over all measures is about 1.3 percent of euro area GDP (text breaks off).

### Empirical approach and main simulation findings (FSGM)
- Model: IMF’s Flexible System of Global Models (FSGM) calibrated to unwind supply-side energy price factors in 2023 and 2024 consistent with futures markets.
- Baseline simulation (difference between with-UFP and without-UFP):
  - Fiscal policy lowers inflation in the euro area by 0.9 percentage point in 2022 and by 0.5 percentage point in 2023.
  - UFP raises inflation by 1.5 percentage point in 2024 as subsidies are withdrawn.
  - Cumulated impact of UFP measures on the price level is slightly positive (about 0.1 percentage point).
  - Absolute deviations from the no-shock baseline are 20 percent smaller for inflation and 30 percent smaller for output with UFP measures.
  - UFP stabilizes output: supporting it in 2022 (especially in France) and moderating it in 2024.
- Financing scenarios:
  - Deficit-financed measures vs deficit-neutral (offset by cuts to government consumption):
    - Financing via reduced government consumption increases the negative impact on inflation from –0.9 percentage point to –1.0 percentage point in 2022 and from –0.5 percentage point to –0.8 percentage point in 2023.
    - Output under deficit-neutral spending cuts is about 0.9 percent lower in 2023 compared with the no-shock baseline.
  - Conventional fiscal consolidation (cuts totalling 3.3 percent of GDP across 2022–23) reduces inflation by –0.1 percentage point in 2022 and –0.3 percentage point in 2023, while reducing output by 1.1 percent in 2022 and 0.9 percent in 2023.
- Sensitivities:
  - If energy supply is twice as elastic, impact of fiscal energy subsidies on inflation increases in absolute value from –0.9 percentage point in 2022 to –1.0 percentage point.
  - Demand offset from income support: combining fiscal expansion size (1.3 percent of GDP) with literature elasticities implies an offsetting effect of about 1 percent on energy prices via demand within a year, corresponding to a rise in overall consumer prices of less than 0.1 percent.

### Empirical Phillips-curve analysis, nonlinearities, and decomposition
- Framework: Phillips-curve with expected inflation, labor market tightness, and headline-inflation shocks; allows nonlinearities via squared and cubic terms.
- Key euro area empirical findings:
  - Baseline regression (1999–2023, Table 1 column 4):
    - U gap coefficient: -0.520*** (Newey-West s.e. 0.159)
    - H coefficient: 0.587*** (0.097)
    - H-squared: 0.172*** (0.029)
    - Observations: 292
    - Rbar-squared: 0.725
  - Since 2020, absolute size of headline-inflation shocks (12-month average) is 1.7 percentage points versus 0.5 percentage point pre–COVID-19 (2009–19); larger shocks show stronger pass-through (H-squared positive and often significant).
  - Weighted median inflation improves fit (adjusted R-squared 73 percent for baseline weighted median vs 38 percent when using XFE inflation).
- Cross-country contrasts:
  - Euro area: most of the rise in core inflation since January 2021 is due to pass-through from headline-inflation shocks.
    - Of the 5.0 percentage point rise in core inflation between January 2021 and April 2023: 3.8 percentage points due to pass-through; 0.2 percentage point due to labor market tightness; 0.5 percentage point from higher expectations; 0.5 percentage point residual.
  - United States: rise in core inflation driven mainly by labor market tightness.
    - Of the 4.9 percentage point rise between January 2021 and April 2023: 3.4 percentage points from labor market tightness; 0.5 percentage point from pass-through; 0.6 percentage point from higher expectations; 0.4 percentage point residual.
  - US shows stronger nonlinearities: pass-through of positive headline shocks raises core inflation significantly while negative shocks have modest effects; labor-market effects steepen sharply at low unemployment (example slope from –0.3 at 8 percent unemployment to –2 at 3.5 percent unemployment in prepandemic US Beveridge relation).

### Counterfactual analysis and quantified effects of UFP measures
- Retail energy price counterfactuals:
  - Without energy price measures, average energy retail prices in 2022 would have been about 19 percent higher than observed, increasing by about 65 percent over 2021 (actual increase 46 percent); peak in October 2022 would have been 31 percent higher.
- Counterfactual headline inflation (euro area):
  - In absence of energy measures, 12-month euro area headline inflation would have peaked at 13.7 percent in October 2022 (3.2 percentage points higher than actual peak of 10.6 percent).
  - On average in 2022, headline inflation is higher by 2.2 percentage points in the counterfactual without measures:
    - About 0.8 percentage point direct effect from headline-inflation shocks.
    - 1.2 percentage points from pass-through into core inflation.
    - Offset from weaker activity less than 0.1 percentage point.
  - Longer-term inflation expectations rise by 0.2 percentage point in absence of energy measures.
- Aggregate 2021–24 effects (stable energy-price scenario, headline shocks set to zero from May 2023):
  - Simulated headline inflation path with measures: April 2023 = 7.0 percent; December 2023 = 2.8 percent; December 2024 = 2.1 percent.
  - Counterfactual without measures: April 2023 = 8.1 percent; December 2024 = 1.6 percent.
  - Absolute deviations from target during 2021–24 are 24 percent smaller with the energy price measures than without them.
  - Average inflation in 2021–24 is 0.5 percentage point lower with measures.
  - Implied cumulative (price level) reduction about 2.0 percent.
- Persistently high energy-price scenario (market prices remain at October 2022 peak):
  - With measures, inflation averages 8.1 percent in 2023 (vs. 5.5 percent under favorable scenario).
  - With measures, headline inflation reaches 6.0 percent by December 2023 (vs. 2.8 percent).
  - With measures, headline inflation reaches 3.1 percent by December 2024 (vs. 2.1 percent).
  - Absolute deviations from target 11 percent smaller with measures (vs. 24 percent in favorable scenario).
  - In 2024, inflation is on average 1.6 percentage points higher than without measures (favorable scenario: 0.7 percentage point higher).
  - Across 2021–24, inflation is on average 0.4 percentage point lower than without measures; cumulative (price level) reduction about 1.6 percent.
- US overheated counterfactual applying euro-area style measures:
  - Headline inflation on average in 2022 initially 1.2 percentage points below actual.
  - By April 2023, inflation exceeds actual by about 1.6 percentage points due to demand stimulus.

### Mechanisms, state dependence, and policy implications
- Mechanisms:
  - UFP neutralizes part of the headline energy-price increase (shifts Phillips curve back) but is expansionary and shifts economy rightward along the Phillips curve.
  - Net inflation outcome depends on initial overheating: if economy not too hot, UFP can reduce inflation; if economy hot, UFP can increase inflation.
- State dependence:
  - Effectiveness of UFP is state dependent (transitory energy shock and low initial overheating are favorable).
  - Euro area outcome likely reflected favorable timing (“luck”) given energy prices receded and economies were not strongly overheated.
- Policy risks:
  - Persistent energy shocks reduce UFP effectiveness and raise fiscal costs.
  - Mismeasurement of slack or unexpected persistence can make UFP counterproductive.
  - Simultaneous suppression across segmented energy markets may yield minimal headline inflation impact.
- Preferable design features:
  - At minimum, UFP measures should be fiscally neutral.
  - Target support to lower-income households to improve efficiency.
  - Preserve price signals at the margin (non-linear or block subsidies) to allow demand compression and avoid shortages.
  - Coordination across countries in integrated markets and budget-compensating measures can improve outcomes.

### Selected exact numeric facts and parameters preserved
- Assumed fiscal costs for energy measures: 1.3 percent of GDP in 2022; 2.0 percent of GDP in 2023.
- Counterfactual retail energy prices: average 19 percent higher in 2022 without measures; peak 31 percent higher in October 2022.
- Expectations anchoring coefficient: 훾 = 0.984 (estimated Q1 2020–Q1 2023).
- Stable energy-price simulated headline inflation: April 2023 = 7.0 percent; December 2023 = 2.8 percent; December 2024 = 2.1 percent.
- Counterfactual without measures (stable-energy): April 2023 = 8.1 percent; December 2024 = 1.6 percent.
- Persistently high energy scenario: inflation averages 8.1 percent in 2023 with measures; December 2023 = 6.0 percent with measures; December 2024 = 3.1 percent with measures.
- US overheated case: initial 2022 average shortfall 1.2 percentage points; by April 2023 overshoot about 1.6 percentage points.
- Euro area baseline Phillips-curve regression (1999–2023, Table 1 column 4):
  - U gap = -0.520*** (0.159)
  - H = 0.587*** (0.097)
  - H-squared = 0.172*** (0.029)
  - Observations = 292
  - Rbar-squared = 0.725

*Source: wpiea2023178-print-pdf*

### Introduction ...........................................................................................................

### Introduction

### Inflation surge and macro context
- World average annual inflation increased from 3.2 percent in 2020 to 8.7 percent in 2022.
- For advanced economies, inflation increased from 0.7 percent to 7.3 percent between 2020 and 2022.
- For emerging market and developing economies, inflation increased from 5.2 percent to 9.8 percent between 2020 and 2022.
- Euro area inflation increased from 0.3 percent in 2020 to 8.4 percent in 2022, having averaged only 1 percent between 2015 and 2019.
- US Federal Reserve hiking cycle started in March 2022 with 12-month headline inflation already exceeding 8.5 percent; over the following 14 months the Federal Reserve increased its policy rate by 500 basis points.
- ECB hiking cycle started in July 2022 with 12-month headline inflation exceeding 8 percent; by June 2023 it had tightened policy rates by a cumulative 400 basis points.
- April 2023 IMF World Economic Outlook forecast:
  - Global headline inflation projected to decrease to 6.6 percent year-on-year in Q2 2023.
  - Inflation excluding food and energy declining to 5.9 percent.
  - In the United States and the euro area, the corresponding inflation Charts are 4.6 percent and 5.1 percent, and 6.3 percent and 6.5 percent, respectively.

### Fiscal policy questions and rationale
- Textbook view: tighter fiscal policy can compress demand, reinforce credibility of disinflation strategy, and alleviate pressures on external and fiscal accounts when used alongside monetary tightening.
- In a monetary union (where monetary policy addresses average price pressures), fiscal tightening can play a helpful role (see Erceg and Lindé 2013; Chen and others 2022 referenced in the source).

### Unconventional fiscal policy (UFP) measures enacted in Europe
- Many European countries implemented fiscal measures aimed at directly countering the rise in energy prices using transfers, energy subsidies, and tax cuts to contain energy prices for households and businesses.
- Economists’ reactions were generally skeptical:
  - Concerns that reducing energy prices would reduce incentives to conserve energy and keep energy demand too high.
  - Measures were often poorly targeted.
  - Significant budgetary cost, estimated at 3.3 percent of GDP for the euro area, could delay debt reduction.
  - If the energy shock were permanent, UFP would be fiscally unsustainable and only delay higher inflation; if transitory, they would smooth inflation over time (reducing it now and increasing it later).
  - The fiscal impulse could stoke aggregate demand and underlying (core) inflation, hampering central bank efforts.

### Empirical approach and key simulation findings
- The paper assesses the effects of such fiscal measures using the IMF’s Flexible System of Global Models (FSGM).
- Simulations assume supply-side factors that caused the spike in international energy prices observed in 2022 unwind in 2023 and 2024, consistent with current futures markets.
- Baseline calibration simulation results:
  - The deficit-financed unconventional measures contribute to stabilizing inflation and output during 2022–24.
  - The measures reduce inflation by about 0.9 percentage point in 2022.
  - The measures reduce inflation by 0.5 percentage point in 2023.
  - The measures raise inflation by 1.5 percentage point in 2024.
  - Net effect: very modest cumulative positive effect; the energy measures smooth the path of inflation—reducing the overshoot in 2022–23 and preventing an undershoot in 2024.
- The paper also considers a simulation where the impact of UFP on the deficit is offset by cuts to government consumption; according to the source, the inflation outcomes are largely unchanged.

### Perspectives from economists and surveys
- In a June 2022 Chicago Booth poll of European economists, most respondents disagreed with the statement “Fiscal measures putting a cap on consumer energy prices would be a more appropriate immediate response to increased inflation in the euro area than raising interest rates.”
- Olivier Blanchard—one of the economists who agreed—argued: “This is a case where a larger fiscal deficit can make the job of monetary policy easier” (Chicago Booth 2022).

*Source: wpiea2023178-print-pdf - Introduction*

### 0.1 percentage point and 0.3 percentage point in 2022 and 2023, respectively. The overall cumulative effect on

### wpiea2023178-print-pdf - 0.1 percentage point and 0.3 percentage point in 2022 and 2023, respectively. The overall cumulative effect on

### Key findings and quantitative results
- UFP (unconventional fiscal policy) can smooth inflation in response to a temporary energy price shock while leaving cumulative inflation largely unchanged.
- The inflation-reducing impact of fiscal tightening in the model class is modest: a fiscal consolidation of 3.3 percent of GDP results in a cumulative inflation (price level) reduction of 0.6 percent (implying about a 0.2 percentage point of inflation reduction per 1 percentage point of GDP in fiscal consolidation).
- Comparison with other studies: Coenen and others (2012) find a temporary fiscal stimulus of 1 percent of GDP in the euro area raises inflation by 0.1 to 0.3 percentage point, consistent with the reverse implication of the present result.
- Empirical conclusion for the euro area:
  - Without energy price measures, euro area headline inflation would have been higher in 2022 by about 2 percentage points.
  - About one-third to one-half of this difference reflects the direct impact of UFP measures on headline inflation; much of the remainder reflects a lower pass-through into core inflation.
  - Net effect of UFP measures: reduce inflation by about 0.5 percentage point in 2021-24 and keep it nearer to the target.
  - Longer-term inflation expectations (ECB’s Survey of Professional Forecasters): would have reached 2.5 percent by the end of 2022, 0.3 percentage point higher than the observed 2.2 percent, in the absence of UFP.

### Mechanisms, nonlinearities, and state dependence
- Phillips curve nonlinearities: the slope steepens as the output gap increases, so the inflation response to demand policies is state dependent.
  - On the flatter part of the curve (economy not too hot), fiscal or monetary consolidation reduces inflation but not by much.
  - On the steeper part of the curve (economy running “hot”), aggregate demand policies are more effective at reducing inflation.
- Energy price shocks shift the Phillips curve up; UFP that neutralizes part of the energy-price increase shifts it back toward the original position.
  - Because UFP is expansionary, it shifts the economy rightward along the Phillips curve; the net inflation outcome depends on initial overheating:
    - If economy is not too hot, UFP can reduce inflation (aggregate demand effects modest).
    - If economy is hot, UFP can increase inflation (aggregate demand effects dominate).
- Empirical decomposition:
  - Headline inflation = core inflation (weighted median inflation) + headline-inflation shocks (deviations of headline from core).
  - For the euro area, much of the increase in core inflation reflects pass-through of headline-inflation shocks; absent those shocks, core would have been more stable.
  - For the United States, most of the increase in core inflation would have occurred even without headline-inflation shocks, reflecting significant overheating.

### Counterfactuals and cross-country comparisons
- Euro area counterfactuals: UFP measures reduced headline inflation by about 2 percentage points in 2022 relative to a no-UFP scenario; net 2021-24 reduction about 0.5 percentage point.
- Persistence of the energy shock matters:
  - In a counterfactual with energy prices remaining at peak 2022 levels and fiscal measures gradually unwound in 2023, headline inflation, pass-through to core, and inflation expectations are substantially higher; energy measures provide less than half as much stabilization compared to the favorable scenario with declining energy prices.
  - Example market prices cited: May 2022 spot Brent crude ~123 US dollars per barrel; TTF gas ~30 US dollars per MMBTU. Forward prices priced Brent at $98 and TTF at $25 for May 2023 delivery. As of May 2023, spot prices were $73 dollars for Brent and $8.4 for TTF.
- US counterfactual applying euro-area style measures:
  - Headline inflation would have been lower by 1.2 percentage points on average in 2022 but would have drifted upward in late 2022 and exceeded actual level by about 1.6 percentage points by April 2023.
  - Using deficit-financed, price-suppressing measures in a significantly overheated economy can add to inflation rather than contain it.

### Policy implications and recommendations
- UFP measures can achieve some inflation reduction and help keep inflation nearer the target under favorable conditions, but they are risky and should be deployed with caution.
- Two conditional aspects that made UFP more successful in the euro area:
  - The energy shock proved more transitory than initially expected (luck).
  - European economies were not strongly overheated to start with.
- Risks and cautions:
  - Temporariness of energy price shocks is hard to determine in real time; persistent shocks would reduce the effectiveness of UFP and require more costly fiscal interventions to avoid persistent inflation.
  - If the economy is overheated, UFP can be counterproductive and raise core inflation.
  - Mismeasurement of economic slack or unexpectedly persistent price shocks increases the risk that UFP exacerbates price pressures.
- Preferable design features:
  - At minimum, UFP measures should be fiscally neutral.
  - Coordination among countries is important where energy markets are segmented; simultaneous suppression attempts across a segmented market may yield minimal impact on headline inflation.
  - Designing measures to preserve price signals at the margin (e.g., non-linear or block subsidies) can allow demand compression, avoid shortages, and minimize fiscal exposure.

### Fiscal cost and scale
- The estimated size of overall energy fiscal support aggregating over all measures to provide relief to households and businesses is about 1.3 percent of euro area GDP (text breaks off here).

### Monetary policy context
- Central banks engaged in a large and synchronized tightening cycle:
  - September 2022: more than 19 policy rate hikes by advanced economies and more than 11 by emerging market central banks.
  - Federal Reserve path: initial 25 basis point hike in March 2022; 50 basis points in May; five successive 75 basis point hikes from June to November 2022; slowed to 50 basis points in December 2022 and 25 basis points from February to May; paused in June 2023.
  - ECB path: 50 basis points in July 2022 (from negative to zero), 75 basis points in September and October, then 50 basis points between December and March 2023, and 25 basis points in May and June.
  - Net outcomes: Federal funds rate raised from close to zero to more than 5 percent over the past 15 months; ECB deposit facility rate raised from a negative level to 3.5 percent.

### Literature connections (selected)
- Nonlinear Phillips curve and state dependence:
  - Ball, Leigh, and Mishra (2022): related empirical decomposition for the United States emphasizing labor-market tightness and nonlinearities.
  - Benigno and Eggertsson (2023): search-and-matching new Keynesian model where supply curve steepens with tight labor markets.
  - Baqaee and Farhi (2022): input-output models with sectoral supply shocks yielding nonlinearities.
- Fiscal policy potency:
  - Model-based results (Christiano, Eichenbaum, and Rebelo 2011; Erceg and Linde 2013) suggest fiscal policy is more powerful when monetary policy is constrained.
  - Chen and others (2022) find monetary policy more potent than fiscal policy at reducing inflation for advanced economies.
  - Beyer and others (forthcoming): a 1 percent of GDP fiscal consolidation in both 2023 and 2024 reduces core inflation by a modest 0.15–0.25 percentage point in the first two years.
- Evidence on transfers and prices:
  - Gourinchas and others (2021): large transfer policies during COVID-19 had a small impact on output (fiscal multiplier 0.06) but a large impact on prices.

*Source: wpiea2023178-print-pdf - 0.1 percentage point and 0.3 percentage point in 2022 and 2023, respectively. The overall cumulative effect on (IMF PDF content provided).*

### 2022. The projected spending for 2023 is 2.0 percent of GDP. The total set of measures can be broken down

### wpiea2023178-print-pdf - 2022. The projected spending for 2023 is 2.0 percent of GDP. The total set of measures can be broken down

### Overview of fiscal measures
- Projected spending for 2023 is 2.0 percent of GDP.
- Total measures include fiscal support targeted at households and small and medium-sized enterprises (SMEs) at 2.7 percent of GDP.
- Fiscal costs budgeted for energy support to firms amounted to 0.6 percent of GDP in the euro area.
- Some countries’ totals are net of windfall profit taxes collected from firms (example noted: energy producers in France).
- Note: some economies also introduced energy support measures in 2021.

### Classification and targeting of measures
- Price-suppressing measures: caps on energy retail prices; cuts to excise duties or value-added tax rates on energy products; block tariffs (discounted tariff for a limited consumption volume).
- Non–price-suppressing measures: energy vouchers, lump-sum income tax credits, cash transfers.
- Targeting:
  - Most measures were largely untargeted to prioritize speed and reach.
  - Price-suppressing and untargeted measures account for about 45 percent of fiscal outlays for households and SMEs in EU countries (Arregui and others 2022).
  - Targeted measures (price suppressing or not) account for about 24 percent of total fiscal measures in the euro area.
  - Examples of targeted measures: expansions of existing social assistance programs; progressive or taxable transfers.
- Country examples:
  - France, Greece, and the Netherlands enacted extensive price controls or reduced value-added tax or other energy-specific tax rates.
  - Germany and Ireland relied more on transfers.

### Cross-country correlations and limitations
- Chart 6 finding: both total and untargeted price-suppressing measures increase weakly with initial exposure to the energy shock, measured by the share of natural gas in primary energy consumption.
- Raw correlation between fiscal costs of measures and variation in inflation is weak across countries, stronger among larger euro area countries.
- Countries with larger fiscal price measures do not show more firmly anchored inflation expectations (based on Consensus Economics surveys and household surveys) for three- and five-year ahead expectations.
- These associations are subject to endogeneity and do not indicate causal efficacy on headline inflation or inflation expectations.

### Model-based assessment (FSGM) — setup and main results
- Model: IMF’s FSGM (multi-region semi-structural model) for the euro area; includes commodities production, consumption, trade, hand-to-mouth and Ricardian households, and a range of fiscal policy tools.
- Energy index: average of oil and natural gas prices weighted by their share in end-use energy consumption; futures use energy price assumptions underlying April 2023 IMF World Economic Outlook.
- Indirect effects: model expanded to include indirect effects of energy prices on core inflation via supply chain effects calibrated from literature.
- Main simulation results (difference between with-UFP and without-UFP measures):
  - Fiscal policy lowers inflation in the euro area by 0.9 percentage point in 2022 and by 0.5 percentage point in 2023 relative to a counterfactual with the same energy shock but without UFP measures.
  - About one-third of the negative impact on headline inflation stems from reduced core inflation.
  - As energy subsidies are withdrawn in 2024, UFP measures raise inflation by 1.5 percentage points in 2024.
  - Cumulated impact of UFP measures on the price level is slightly positive (about 0.1 percentage point).
  - Absolute deviations from the no-shock baseline are 20 percent smaller for inflation and 30 percent smaller for output with the UFP measures.
  - The UFP measures stabilize output: supporting it in 2022 (especially in France) and moderating it in 2024.

### Extensions and scenarios examined
- Deficit-financing versus deficit-neutral scenario:
  - Financing measures through reduced government spending (cutting government consumption) increases the negative impact on inflation from –0.9 percentage point to –1.0 percentage point in 2022 and from –0.5 percentage point to –0.8 percentage point in 2023.
  - Output under deficit-neutral spending cuts is significantly lower: about 0.9 percent lower in 2023 compared with the no-shock baseline (panel b reference).
- Conventional fiscal consolidation (deficit-neutralizing cuts to government consumption):
  - Cuts amount to 1.3 percent of GDP in 2022 and a further 2.0 percent of GDP in 2023 (total 3.3 percent of GDP within two years).
  - Reduce output level by 1.1 percent in 2022 and 0.9 percent in 2023.
  - Effect on inflation: –0.1 percentage point in 2022 and –0.3 percentage point in 2023.
  - Achieving a reduction in inflation in 2022 comparable with UFP via conventional tightening would require significantly larger tightening with larger output losses.
- Energy supply elasticity sensitivity:
  - If energy supply is twice as elastic, impact of fiscal energy subsidies on inflation increases (in absolute value) from –0.9 percentage point in 2022 to –1.0 percentage point.
- Demand offset from income support:
  - Combining fiscal expansion size (1.3 percent of GDP) with literature elasticities implies an offsetting effect of about 1 percent on energy prices via the demand channel within a year, corresponding to a rise in overall consumer prices of less than 0.1 percent.

### Phillips curve analysis and nonlinearities — euro area vs US
- Framework: Phillips-curve focusing on expected inflation, labor market tightness, and headline-inflation shocks; allows for nonlinearities in labor market tightness and past headline shocks.
- Measures used:
  - Euro area expected inflation: ECB Survey of Professional Forecasters five-year-ahead HICP inflation forecast.
  - Euro area labor market tightness: unemployment gap (deviation from natural rate using IMF staff estimates from April 2023 WEO); unemployment has averaged 6.7 percent since start of 2022 vs historical average 9.5 percent (1999–2019) and IMF natural rate estimate of 7.0 percent for 2022.
  - US labor market tightness: ratio of job vacancies to unemployed (V/U) over current and previous 11 months, used because of a shift in the US Beveridge curve.
  - Past headline-inflation shocks: 12-month average of headline-inflation deviations; nonlinearities allowed via squared and cubic terms.
- Empirical findings for the euro area:
  - Sample: monthly data January 1999–April 2023; results reported for full sample and pre–COVID-19 sample (ending December 2019).
  - Pre–COVID-19 linear specification: unemployment gap and pass-through terms strongly statistically significant and match Ball and Mazumder (2021).
  - Since 2020, absolute size of headline-inflation shocks (12-month average) is 1.7 percentage points versus 0.5 percentage point pre–COVID-19 (2009–19).
  - Including COVID-19 period shows evidence of stronger pass-through; coefficient more than doubles.
  - Allowing for nonlinearities (squared terms) implies large shocks during COVID-19 had especially powerful pass-through effects.
  - For the unemployment gap, results do not indicate a strong degree of nonlinearity over observed sample range.
- Robustness:
  - Results robust to excluding outliers and including higher-order nonlinearities (cubed terms).
  - Using weighted median inflation is important: adjusted R-squared 73 percent for baseline weighted median inflation vs 38 percent when using XFE inflation (inflation excluding food and energy).

### Key takeaways
- UFP measures can reallocate inflation over time, smoothing a large temporary energy price shock:
  - Cumulated reduction in inflation (effect on the price level) is around 1.4 percent for the euro area in 2022–23, followed by a more than offsetting increase in 2024.
- The inflation-reducing impact of conventional fiscal tightening is modest in the model and requires large fiscal adjustments for comparable effects, entailing larger output losses.
- Nonlinearities in the Phillips curve, especially pass-through from large headline shocks, are important for understanding inflation dynamics; linear model assumptions may understate impacts.
- Empirical evidence highlights stronger pass-through from large headline shocks since 2020 and supports using weighted median inflation for better fit.

*Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023178-print-pdf.pdf*

### Annex Table 2 reports additional estimation results with the degree of labor market tightness measured by the

### wpiea2023178-print-pdf - Annex Table 2 reports additional estimation results with the degree of labor market tightness measured by the

### Robustness of labor market tightness measures and samples
- Annex Table 2 reports additional estimation results with the degree of labor market tightness measured by the ratio of vacancies to unemployed (V/U).
- Data for V/U are available for a shorter sample starting in 2006 and at the quarterly frequency.
- The estimation sample here starts in 2007 given our use of the four-quarter-average of labor market tightness in our specifications.
- Results summary:
  - The results with V/U are less stable than with the unemployment gap.
  - For the pre-COVID-19 sample, the V/U terms are statistically insignificant.
  - For the full sample through the first quarter of 2023, the V/U terms are statistically significant but lose significance and change sign when outlier observations are discarded.
  - By contrast, the results based on the unemployment gap for this quarterly sample are robust and similar to the baseline results in Tables 1 and 2, including when discarding outliers.
  - Estimation results based on the output gap (as measured in the IMF World Economic Outlook database) indicate a similarly strong fit as do the baseline results based on the unemployment gap reported in Table 1.
    - The Phillips curve slope estimate for the output gap is positive, at about 0.2, for both the prepandemic sample and for the full sample ending in 2023, and indicate little evidence of nonlinearity.
    - The estimation results for the headline-inflation shock terms in the output gap-based specification are statistically significant and similar to those obtained in the baseline specification based on the unemployment gap.
    - Results for the additional empirical exercises reported in this section are also similar when using the output gap as the measure of economic slack.

### Nonlinearities and asymmetries in the Phillips curve: United States vs. euro area
- United States:
  - Strong evidence of nonlinearities in the pass-through of headline-inflation shocks:
    - Positive shocks significantly raise core inflation.
    - Negative shocks have modest, statistically insignificant effects on core inflation.
  - Greater evidence of significant nonlinearities in the effects of labor market tightness on core inflation.
  - Chart 12 and Annex Table A1 (column 2) / Beveridge curves imply a highly nonlinear slope function for the unemployment rate:
    - For the prepandemic period, the slope of the US relation is about –0.3 at 8 percent unemployment but steepens to –2 at 3.5 percent unemployment.
    - Overheating has pushed the United States onto the steep part of the Phillips curve.
- Euro area:
  - Relationship with unemployment shows less nonlinearity than for the United States.
  - The slope when the unemployment rate is 8 percent is about –0.4, implying that a 1 percentage point rise in unemployment comes with a 0.4 percentage point fall in core inflation.
  - The slope steepens to about –0.7 when the unemployment rate declines to 6 percent.
  - The euro area might have a similarly steep part of the curve, but this remains uncertain as there has not yet been sufficient overheating to reveal it.

### Implications for policy and inflation dynamics
- Policy implications:
  - Policy tightening that cools demand can potentially achieve larger inflation reductions in a more overheated economy.
  - Increasing demand in an already overheated economy comes with greater inflation risks.
- Pass-through from headline to core inflation:
  - Both economies display a strong degree of pass-through from headline-inflation shocks to core inflation, with an asymmetry:
    - Positive headline-inflation shocks have a strong increasing relation with core inflation.
    - Negative headline shocks have an insignificant relation with core inflation.
  - More research is needed on the sources of such asymmetries.

### Drivers of recent euro area and US core inflation (fitted values comparison)
- Methodology notes:
  - Chart 13 plots fitted values from the respective Phillips curve models, along with actual core inflation.
  - Fitted values are converted from monthly annualized inflation gaps into 12-month inflation rates and are augmented by the level of longer-term expected inflation expectations.
  - For both economies, the resulting fitted values of 12-month core inflation are close to their actual values.
- Euro area finding:
  - The rise in euro area core inflation since January 2021 primarily reflects pass-through from past headline-inflation shocks, not economic overheating.
  - Of the 5.0 percentage point total rise in core inflation from January 2021 to April 2023, ... (content ends in source).

*Source: wpiea2023178-print-pdf - Annex Table 2 reports additional estimation results with the degree of labor market tightness measured by the (canonical IMF PDF).*

### 3.8 percentage points (three-quarters of the rise) are due to pass-through; 0.2 percentage point is due to labor

### wpiea2023178-print-pdf - 3.8 percentage points (three-quarters of the rise) are due to pass-through; 0.2 percentage point is due to labor

### Drivers of the rise in core inflation
- Euro area: of the 5.0 percentage point rise in core inflation between January 2021 and April 2023:
  - 3.8 percentage points are due to pass-through (three-quarters of the rise).
  - 0.2 percentage point is due to labor market tightness.
  - 0.5 percentage point comes from higher inflation expectations.
  - 0.5 percentage point is a residual.
- United States: of the 4.9 percentage point difference between core inflation in January 2021 and April 2023:
  - 3.4 percentage points come from labor market tightness (70 percent of the rise).
  - 0.5 percentage point comes from pass-through from past headline shocks.
  - 0.6 percentage point contribution from higher expectations.
  - 0.4 percentage point residual.
- Interpretation:
  - During the initial rise in core inflation, pass-through from past headline-inflation shocks plays a dominant role.
  - During 2022, labor market tightness became the principal factor explaining core inflation, particularly in the United States.

### Accounting for the rise in headline inflation
- Euro area headline inflation at peak (October 2022):
  - Headline inflation = 10.6 percent, which is 9.7 percentage points higher than January 2021.
  - 4.8 percentage points reflect the estimated direct contribution of energy price inflation to headline-inflation shocks.
  - 4.2 percentage points reflect the associated pass-through effects into core inflation.
  - Total contribution from energy shocks and pass-through = 9.0 percentage points (93 percent of the total rise).
  - With energy price inflation subsiding in late 2022 and early 2023, most of the remaining rise compared with January 2021 reflects pass-through into core inflation.
- United States headline inflation at peak (June 2022):
  - Headline inflation is 7.5 percentage points higher than January 2021.
  - 2.8 percentage points reflect the estimated direct contribution of energy price inflation to headline-inflation shocks.
  - 0.6 percentage point reflects associated pass-through effects.
  - Headline-inflation shocks from other industries and their pass-through together account for 2.3 percentage points.
  - Labor market tightness accounts for 0.8 percentage point of the rise through June 2022.
  - By April 2023, with headline-inflation shocks subsiding and turning negative, the rise in inflation compared with January 2021 fully reflects labor market tightness.

### Counterfactual analysis: constructing headline-inflation shocks without UFP measures
- Method overview (three steps):
  1. Construct monthly series of counterfactual headline-inflation shocks in the absence of UFP measures using the gap between “market” energy retail prices (from official sources) and actual retail prices.
  2. Derive counterfactual path of the core inflation gap by:
     - Applying the monthly euro area Phillips curve (Table 1, column 4) to account for nonlinear pass-through from headline-inflation shocks to core.
     - Deriving the impact from an assumed unemployment path in the absence of the measures using the budgetary cost of the energy measures and assumptions on fiscal multipliers and Okun’s law.
  3. Calculate the counterfactual effect of the measures on inflation expectations using the counterfactual core inflation gap and an estimated process for longer-term inflation expectations that allows feedback from headline inflation.
- Data and assumptions for market energy prices:
  - Gas: use CRE indicative market retail prices as measure of counterfactual retail gas price growth for the euro area.
  - Electricity: use CRE theoretical retail electricity prices for France and IMF staff estimates for Spain to construct counterfactual retail electricity prices.
- Key calculated implications for retail energy prices:
  - In the absence of the energy price measures, the level of energy retail prices in the euro area on average in 2022 would have been about 19 percent higher than observed, increasing by about 65 percent over the average level in 2021, instead of by the actual 46 percent.
  - The peak of energy prices in October 2022 would have been 31 percent higher.

### Counterfactual effects on the core inflation gap through demand channel
- Fiscal cost assumptions for energy measures (from Arregui and others (2022)):
  - 1.3 percent of GDP in 2022.
  - 2.0 percent of GDP in 2023.
- Assumed fiscal multiplier and Okun’s law (conservative/high-end assumptions):
  - Average peak fiscal multiplier on GDP of 1 after 12 months, fading to zero after 24 months (implying a 0.5 multiplier on average over the year).
  - Okun’s law coefficient of –0.5.
- Implications for unemployment in the absence of measures:
  - Unemployment would have been 0.7 percentage point higher at the end of 2022.
  - Unemployment would have been 1.0 percentage point higher at the end of 2023.
  - Monthly unemployment path obtained by interpolating between year-end values; 12-month average of this counterfactual unemployment path added to actual unemployment to compute impact on core inflation gap via the Phillips curve.
- Note: these assumptions tend to err on the side of overstating the measures’ effects on inflation via the aggregate demand channel; even so, the inflation effects through this channel are modest.

### Counterfactual inflation-expectation process
- Expectation evolution specification (Ball, Leigh, and Mishra (2022) style):
  - 휋_t^e = 훾 휋_{t-1}^e + (1−훾) 휋_t  (with adaptations for nowcasts and an error term in estimation).
- Estimated 훾 (anchoring parameter) using ECB SPF five-year-ahead series:
  - 0.992 for the period Q3 2008–Q4 2019.
  - 0.984 for the period Q1 2020–Q1 2023.
  - 0.985 for the period Q1 2001–Q2 2008 (noted for context).
- Interpretation:
  - Longer-term inflation expectations have responded modestly to near-term expectations.
  - Anchoring weakened slightly in the pandemic period relative to pre-pandemic.

### Counterfactual simulation results: effects of energy support measures on headline inflation
- Peak effect (euro area):
  - In the absence of energy measures, 12-month euro area headline inflation would have peaked at 13.7 percent in October 2022.
  - This is 3.2 percentage points higher than the actual peak of 10.6 percent.
- Average 2022 effect:
  - On average in 2022, headline inflation is higher by 2.2 percentage points in the counterfactual without measures.
  - Composition of the 2.2 percentage point difference:
    - About 0.8 percentage point reflects the direct effect of the headline-inflation shocks.
    - 1.2 percentage points come from pass-through effects into core inflation (greater than one-for-one due to steep pass-through slope under large positive headline shocks).
    - Offset from weaker economic activity in absence of measures on core inflation is negative but modest (less than 0.1 percentage point).
  - Longer-term inflation expectations rise by 0.2 percentage point in the absence of the energy measures.

### Scenarios for future inflation (framework and inputs)
- Approach:
  - Use empirical Phillips curve framework to predict conditional paths of inflation given IMF staff forecasts of the unemployment gap and different assumptions about future energy prices.
- Unemployment inputs:
  - Use unemployment forecasts from the April 2023 World Economic Outlook.
  - Quarterly forecasts peak at 6.9 percent in Q1 2023.
  - Natural rate of unemployment forecasts remain near 7.0 percent in 2023 and 2024.
  - Construct a monthly unemployment gap path by assigning each quarterly unemployment forecast to the second month of each quarter and interpolating between months, starting with the actual unemployment rate.

*Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023178-print-pdf.pdf*

### 6.5 percent in April 2023. For the natural rate, we assign the estimates to June of each year and interpolate

### Future Inflation Paths with Stable Energy Prices

### Methodology and key assumptions
- Unemployment and natural rate treatment:
  - Assign natural rate estimates to June of each year and interpolate between available observations (Annex Chart A2).
  - Based on interpolated monthly paths, compute the 12-month average of the unemployment gap needed for the Phillips curve model.
- Headline-inflation shocks and energy prices:
  - Simplifying assumption: no further headline-inflation shocks stem from energy prices or other sectors starting from May 2023.
  - Set headline-inflation shocks to zero for future months for this scenario.
  - Resulting 12-month headline shock (H), given zero monthly shocks starting in May 2023 and the earlier actual shocks, declines to zero in April 2024.
  - Uncertainty note: future headline shocks could be adverse (e.g., intensification of the war in Ukraine) or disinflationary (e.g., further falls in energy prices).
- Alternative energy-price counterfactual:
  - Scenario where market energy prices remain at October 2022 peak levels and actual (regulated) energy prices rise linearly to market levels by December 2023 from October 2022 levels––a total rise of 52 percent.
- Expectations anchoring:
  - Solve jointly for the path of inflation expectations and headline inflation, holding the degree of expectations anchoring at estimated level with coefficient 훾 = 0.984.

### Stable energy prices scenario (no further headline shocks from May 2023)
- Simulated headline inflation path (IMF staff assumptions):
  - April 2023 level: 7.0 percent.
  - December 2023: 2.8 percent.
  - December 2024: 2.1 percent.
- Counterfactual without UFP (unconventional fiscal policy) energy measures (market energy prices converge to regulated levels by December 2023 and remain thereafter):
  - Counterfactual April 2023 level: 8.1 percent.
  - Counterfactual December 2024: 1.6 percent.
- Comparative outcomes and aggregate effects (2021–24):
  - Absolute deviations of inflation from target during 2021–24 are 24 percent smaller with the energy price measures than without them.
  - Average inflation in 2021–24 is 0.5 percentage point lower with the measures.
  - Implied cumulative (price level) reduction of about 2.0 percent.

### Persistently high energy prices scenario (market energy prices remain at October 2022 peak)
- Aggregate and timing effects:
  - With measures, inflation averages 8.1 percent in 2023 instead of 5.5 percent under the more favorable (declining energy) scenario.
  - With measures, headline inflation reaches 6.0 percent by December 2023 instead of 2.8 percent.
  - With measures, headline inflation reaches 3.1 percent by December 2024 instead of 2.1 percent.
- Counterfactual without energy measures in this scenario:
  - Declines more gradually; no undershoot of the 2 percent inflation target.
- Comparative stabilization effects (2021–24):
  - Absolute deviations of inflation from target are 11 percent smaller with the energy price measures than without them (versus 24 percent in the favorable scenario).
  - In 2024, inflation is on average 1.6 percentage points higher than without the measures (favorable scenario: 0.7 percentage point higher than without the measures).
  - Across 2021–24 in this scenario, inflation is on average 0.4 percentage point lower than without the measures, implying a cumulative (price level) reduction of about 1.6 percent.
- Interpretation:
  - In this “unlucky” case, the energy price measures deliver less than half as much inflation stabilization during 2021–24 as in the favorable scenario.

### Inflation paths in an overheated economy (US counterfactual experiment)
- Experiment design:
  - Apply the euro area counterfactual headline-inflation shocks and unemployment effects from UFP measures during 2022–23 to the United States over the same period but with opposite sign to simulate efforts to reduce inflation below actual levels.
- Results:
  - Initially lower inflation: on average in 2022, headline inflation is 1.2 percentage points below the actual level.
  - Rapid reversal due to demand stimulus: by the end of 2022 inflation drifts upward; by April 2023 inflation exceeds the actual level by about 1.6 percentage points.
- Policy implication:
  - Implementing UFP measures in an overheated economy risks stoking demand and producing significantly higher inflation, complicating monetary policy’s disinflation efforts.

### Synthesis of model findings and policy recommendations
- Complementary modeling approaches:
  - Semi-structural large-scale New Keynesian model: UFP measures can help smooth inflation facing large transitory energy shocks; conventional fiscal policy would have been of little help in curbing inflation given low fiscal multipliers in a high inflation environment and a modest Phillips curve slope.
  - Empirical Phillips curve framework with nonlinearities: accounts for surge in euro area and US inflation; finds UFP measures reduced euro area inflation by 2.2 percentage points on average in 2022 through three channels—direct headline shock reduction, indirect core inflation pass-through, and better-anchored longer-term inflation expectations.
- State dependence and generalizability:
  - Effectiveness of UFP measures is “state dependent.” If energy shocks are persistent or the economy is overheating, outcomes can be adverse.
  - Euro area outcome may reflect favorable timing (“luck”) given energy prices receded; results do not generalize automatically to other contexts (e.g., an overheated US economy).
- Fiscal design and targeting:
  - UFP measures imposed fiscal costs; more effectiveness and lower fiscal costs could be achieved if accompanied by budget-compensating measures or more targeted deployment.
  - Deficit-neutralizing spending cuts would have modestly lowered inflation further.
  - Targeting support to lower-income households is a more efficient use of public funds to limit incidence of energy price shocks.
- Questions for future research:
  - Effects of price-suppressing measures on energy demand and wholesale prices given low short-run demand and supply elasticities.
  - Welfare-based assessments of UFP policies across countries.
  - Optimal design of UFP in a monetary union with integrated energy markets, accounting for asymmetric shocks across countries.
  - Whether subsidies should target household energy prices or firms’ energy inputs to better stabilize consumer prices and upstream costs.
  - Whether non price-distorting measures (e.g., lump-sum transfers) can reduce inflation by alleviating cost increases without distorting prices.

### Selected exact numeric facts and model parameters (preserved)
- Natural rate assignment: estimates assigned to June of each year and interpolated.
- Headline-inflation shocks set to zero starting in May 2023; 12-month headline shock (H) declines to zero in April 2024 under that assumption.
- Assumed linear rise of regulated energy prices to market levels by December 2023 from October 2022 levels: total rise of 52 percent.
- Expectations anchoring coefficient: 훾 = 0.984.
- Simulated headline inflation path with stable energy prices:
  - April 2023: 7.0 percent.
  - December 2023: 2.8 percent.
  - December 2024: 2.1 percent.
- Counterfactual without measures (stable-energy scenario):
  - April 2023: 8.1 percent.
  - December 2024: 1.6 percent.
- Aggregate outcomes 2021–24 (stable-energy scenario):
  - Absolute deviations from target 24 percent smaller with measures.
  - Average inflation 0.5 percentage point lower with measures.
  - Cumulative (price level) reduction about 2.0 percent.
- Persistently high energy scenario:
  - Inflation averages 8.1 percent in 2023 with measures (vs. 5.5 percent under favorable scenario).
  - December 2023: 6.0 percent with measures (vs. 2.8 percent).
  - December 2024: 3.1 percent with measures (vs. 2.1 percent).
  - Absolute deviations from target 11 percent smaller with measures (vs. 24 percent in favorable scenario).
  - In 2024, inflation on average 1.6 percentage points higher than without measures (favorable scenario: 0.7 percentage point higher).
  - Inflation on average 0.4 percentage point lower than without measures overall; cumulative (price level) reduction about 1.6 percent.
- US overheated case:
  - In 2022, headline inflation on average 1.2 percentage points below actual initially.
  - By April 2023, inflation exceeds actual by about 1.6 percentage points.
- Paper’s aggregate comparison (summary):
  - With measures: euro area headline inflation declines from 7.0 percent in April 2023 to 2.8 percent by December 2023 and to 2.1 percent by December 2024.
  - Without these measures: inflation would have peaked at 13.7 percent in October 2022 before declining to 8.1 percent in April 2023 and to 1.7 percent in December 2023.

*Source: wpiea2023178-print-pdf (IMF staff analysis and simulations).*

### 2.3 percent target for CPI based on 2 percent PCE target reported on Federal Reserve Bank of Atlanta Underlying Inflatio

### wpiea2023178-print-pdf - 2.3 percent target for CPI based on 2 percent PCE target reported on Federal Reserve Bank of Atlanta Underlying Inflation

### Euro Area Phillips Curve Estimates (Table 1; Annex Tables)
- Dependent variable: Median CPI Inflation Gap (core inflation minus expected inflation), with core measured by monthly annualized weighted median HICP inflation and expected inflation by ECB Survey of Professional Forecasters (SPF) five-year-ahead forecast of headline inflation.
- Baseline (1999-2023, Table 1 column 4) estimates:
  - U gap coefficient: -0.520*** (Newey-West s.e. 0.159)
  - U gap-squared: 0.090 (0.068)
  - H coefficient: 0.587*** (0.097)
  - H-squared: 0.172*** (0.029)
  - Observations: 292
  - Rbar-squared: 0.725
- Robustness and extensions (Annex Table 1, sample 1999-2023):
  - Baseline U gap: -0.520*** (0.159)
  - Variants show U gap ranging from -0.432*** to -0.673** across columns and specifications.
  - H coefficients across specifications: 0.279 to 0.651***; H-squared positive and often statistically significant (e.g., 0.250*** in column 6).
  - Rbar-squared varies across specifications (e.g., 0.379 to 0.731).
- Quarterly data extensions (Annex Table 2, 2007 onwards):
  - U gap coefficients significant (e.g., -0.561***, -0.653***)
  - H coefficients significant (e.g., 0.343*** to 0.545***)
  - Rbar-squared reported up to 0.860 in some quarterly specifications.

### Estimated Inflation Gap as Function of Unemployment and Headline Shocks (Chart 12)
- Chart reports estimated relations using Table 1 (euro area) and updated US estimates (Annex Table 1, US Beveridge curve variants).
- Key modeling elements:
  - Inflation gap denotes monthly annualized median HICP inflation minus longer-term inflation expectations.
  - Relations reported for values of U and H observed in the euro area and US samples.
  - Bands report 95 percent confidence intervals.

### Predictions and Fitted Values for Core Inflation (Charts 13–14)
- Chart 13: Fitted values for core inflation (12-month weighted median) for 2020–23:
  - Euro Area: fitted values based on Table 1 (column 4).
  - United States: fitted values based on Annex Table 1 (US specification updating Ball, Leigh and Mishra (2022)).
  - Fitted inflation-gap estimates converted into 12-month rates and summed with longer-term expectations.
  - Panels include "Assuming no headline shocks" counterfactuals.
- Chart 14: Decomposition/Accounting for rise in headline inflation (12-month rate):
  - Components shown: Headline, Headline shock: energy, Headline shock: other, Expectations, Labor market tightness, Passthrough: energy, Passthrough: other, Residual.
  - Separate panels for Euro Area and United States.

### Energy Price Measures: Actual vs Counterfactual (Charts 15, 17–19)
- Chart 15 (Euro Area energy price levels; Index, January 2020 = 100):
  - "Actual" indicates HICP for Electricity, Gas and Other Fuels series.
  - Counterfactual path constructed without energy price measures.
- Chart 17 (Headline inflation: actual and counterfactual without energy measures; 12-month rate):
  - Horizontal dashes show 2 percent target for HICP inflation.
  - Counterfactuals include: direct effect; add pass-through effect; add demand effect; add expectations drift.
- Chart 18 (Scenario: Headline Inflation with and without Energy Measures; Percent; 12-month rate):
  - Horizontal dashes show 2 percent target for HICP inflation.
  - Two-panel presentation: headline-inflation paths with measures and without measures for 2020–2025; marginal effect of energy measures (percentage points) plotted for 2021–2025.
  - Chart reports average gap between paths.
- Chart 19 (Alternative scenario with energy prices staying high):
  - Vertical line indicates October 2022 after which market energy prices assumed to be constant with actual retail prices rising linearly to their market price level by December 2023.
  - Horizontal dashes show 2 percent target for HICP inflation.
  - Series plotted: With measures, Without measures, With measures (alternative), Without measures (alternative); marginal effect and average gap shown.

### US Headline Inflation and Unconventional Fiscal Measures (Chart 20)
- Chart 20 (US Headline Inflation: Actual and Counterfactual with Unconventional Fiscal Measures; Percent; 12-month rate):
  - Horizontal dashes show 2.3 percent target for CPI based on 2 percent PCE target reported on Federal Reserve Bank of Atlanta Underlying Inflation Dashboard.
  - Counterfactual decomposition steps: add direct effect; add pass-through effect; add demand effect; add expectations drift.
  - Period plotted: 2020–2023 with actual and counterfactual fitted series.

### Annex: Additional Figures and Tables
- Annex Chart 1: Marginal effect of deficit-financed energy fiscal measures on headline inflation — baseline vs alternative with elastic energy supply (deviation from counterfactual path without energy measures; percentage points). Simulations based on IMF Flexible System of Global Models.
- Annex Chart 2: Euro Area unemployment — actual rate and estimated natural rate (Percent of labor force). Vertical line indicates April 2023 (IMF World Economic Outlook forecast).
- Annex Table 3 (US Phillips Curve estimates; median CPI Inflation Gap):
  - Sample 1985-2019 and 1985-2023 updates:
    - V/U coefficients: 9.626** and 7.821*** (s.e. 4.272 and 2.138)
    - V/U-squared: -11.014* and -8.133*** (6.384 and 2.342)
    - V/U-cubed: 4.512 and 3.151*** (2.926 and 0.746)
    - H and H-polynomial terms: H not significant in these columns (e.g., 0.010 and 0.033), but H-squared and H-cubed often significant.
    - Constants: -2.770*** and -2.438***.
    - Observations: 420 and 460; Rbar-squared: 0.273 and 0.615.

### Key quantitative references preserved from the source
- 2.3 percent target for CPI based on 2 percent PCE target (Federal Reserve Bank of Atlanta Underlying Inflation Dashboard) used in Chart 20.
- Euro area baseline regression (1999-2023, Table 1 column 4): U gap -0.520***; H 0.587***; H-squared 0.172***; Observations 292; Rbar-squared 0.725.
- Annex Table 1 baseline sample size: Observations 292; Rbar-squared 0.725.
- Annex Table 2 quarterly sample observations: 52, 65, 60, 52, 65, 59; Rbar-squared up to 0.860.
- Annex Table 3 US sample observations: 420 and 460; Rbar-squared 0.273 and 0.615.

*Source: Unconventional Fiscal Policy in Times of High Inflation, Working Paper No. WP/2023/178 (excerpts from provided PDF).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023178-print-pdf.pdf_