## Annex I. Data Sources

## Source details

**Canonical URL:** [Annex I. Data Sources](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023030-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023030-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023030-print-pdf.pdf.json)

---

### Main objectives and scope
- Aim: Shed light on Europe’s post-COVID-19 inflation puzzle by exploring empirically the factors that contributed to cross-country dispersion in inflation and persistent overshooting through the lens of the standard Phillips curve.
- Focus: Differences in Phillips curves across countries and potential shifts in the inflation process since the onset of the COVID-19 pandemic.
- Sample period: 2000Q1-22Q2.
- Country coverage:
  - 24 advanced European economies (AE): Austria, Belgium, Cyprus, Czech Republic, Germany, Denmark, Estonia, Finland, France, Greece, Ireland, Israel, Italy, Lithuania, Latvia, Netherlands, Norway, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, and United Kingdom.
  - 7 emerging European economies (EE): Bulgaria, Croatia, Hungary, Poland, Romania, Russia and Türkiye.

### Key empirical framework and data construction
- Baseline model: New Keynesian Phillips curve augmented with:
  - lagged and expected inflation,
  - domestic economic slack (unemployment gap or output gap),
  - contemporaneous and four quarterly lags of energy and food price growth expressed in domestic currency and weighted by CPI shares,
  - an external price pressure index (extP) capturing import-weighted PPI changes of trading partners and relative bilateral exchange rate changes minus country i GDP deflator change,
  - country fixed effects and an error term.
- Inflation variable: quarter-over-quarter annualized core (headline) inflation in country i in quarter t.
- Inflation expectations: three-year-ahead inflation expectations (robust to using 1-year ahead).
- Slack measure: deviation from the Hodrick-Prescott (HP)-filtered unemployment rate (alternatively log output) with smoothing parameter of 1600.
- External price pressure (extP) defined as:
  - import-share weighted percent change in trading partners’ PPI,
  - plus import-share weighted relative changes in bilateral exchange rates against the US dollar,
  - minus percent change in country i GDP deflator.
- Global commodity prices: converted to local currency using bilateral US dollar exchange rates and weighted by energy/food share in country i's consumption basket.
- Lags: contemporaneous energy and food price increases plus four quarterly lags to capture gradual passthrough (food passthrough slower; lagged food significant up to fourth lag).
- Identification constraint: benchmark imposes constraint that sum of coefficients on past and expected inflation rates (β1+β2) equals one. Relaxing this restriction yields similar findings.

### Principal empirical findings
- Cross-country inflation outcomes:
  - Inflation exceeded 5 percent in almost all European economies by late summer 2022.
  - By summer 2022, headline inflation reached 20–25 percent in the Baltic countries, three to four times the rate in the lowest-inflation countries in the euro area.
  - Average core inflation in emerging European economies exceeded 20 percent by Q2 2022 — a 12 percentage points gap vis-à-vis advanced European countries, compared to a 2-3 percentage points gap prior to the pandemic.
- Benchmark Phillips-curve estimates (core inflation, panel):
  - A 1 percentage point increase in unemployment above its HP trend is associated with a decline in core inflation of 0.4 percentage points in Europe.
  - A 1 percentage point rise in inflation expectations is significantly associated with a 0.6 percentage point increase in core inflation in Europe.
- Heterogeneity between AE and EE:
  - The slope of the Phillips curve (coefficient on domestic slack) is steeper in EE than in AE.
  - Inflation expectations coefficient: 0.4 in EE versus 0.7 in AE.
  - Inflation in EE responds more strongly to foreign price developments, especially global food prices.
  - These differences contributed to a larger passthrough of global commodity price shocks and a larger inflation surge in EE in 2021-22.
- Changes since the pandemic:
  - Evidence that inflation in Europe has become increasingly backward looking and more sensitive to food price shocks since the onset of the COVID-19 pandemic.
  - Higher vacancy-to-unemployment ratios by end-2022 suggest labor markets may have been tighter than unemployment dynamics alone indicate, contributing to larger-than-expected inflation.

### Robustness and alternative specifications
- Robustness checks:
  - Alternative estimators: OLS without constraint on expected and lagged inflation coefficients, median regression to reduce influence of extremes, and use of output gaps instead of unemployment gaps.
  - Across specifications, findings are robust: steeper Phillips curve slope in EE, larger role of lagged inflation, and larger role of foreign price developments in EE than in AE.
  - Note: coefficients on economic slack are not statistically significant for EE when measured by the output gap.
  - Relaxing the (β1+β2)=1 restriction produces similar conclusions on slack, external pressures and commodity prices, though in some cases β1 may exceed 1.

### Policy-relevant implications
- Divergence risk: Differences in Phillips curves point to a risk of persistent inflation divergence between euro area economies, with significantly higher inflation in newer member states (e.g., the Baltic countries).
- Monetary policy calibration:
  - Outside the euro area, emerging market economies may require a more forceful policy response to tame inflationary pressures due to steeper slack sensitivity and greater passthrough from external prices.
  - The increasing backward-looking nature of inflation warrants careful monitoring as it could signal changes in price-setting and wage bargaining, increasing risks of feedback loops between wages and prices and potential de-anchoring of inflation expectations.
- Forecasting and model choice: The heightened persistence and stronger passthrough of external commodity prices help explain repeated underprediction of inflation by models based on historical relationships; modelers and policymakers should account for evolving Phillips-curve parameters and external-price sensitivity.

### Benchmark Phillips-curve estimation (selected coefficients, core inflation)
- Sample: Unbalanced panel, 2000Q1-2022Q2. Sum of coefficients on lagged and expected inflation constrained to be one.
- Unemployment gap:
  - (1) -0.374*** (0.078)
  - (2) -0.337*** (0.097)
  - (3) -0.676*** (0.179)
- Lag of inflation:
  - (1) 0.431*** (0.127)
  - (2) 0.287* (0.161)
  - (3) 0.581*** (0.105)
- Inflation expectations:
  - (1) 0.569*** (0.127)
  - (2) 0.713*** (0.161)
  - (3) 0.419*** (0.105)
- Lag of external price pressure:
  - (1) 0.020*** (0.006)
  - (2) 0.009* (0.005)
  - (3) 0.037** (0.015)
- Food prices (lags 0–4) display significant positive coefficients (examples):
  - Lag 0: (1) 0.127*** (0.034); (2) 0.065*** (0.015); (3) 0.181*** (0.054)
  - Lag 3: (1) 0.070*** (0.021); (2) 0.077*** (0.018); (3) 0.065** (0.032)
- Energy prices:
  - (1) 0.016 (0.011)
  - (2) 0.032*** (0.008)
  - (3) 0.021 (0.022)
- Observations:
  - (1) 2,210
  - (2) 1,707
  - (3) 503

### Contribution to inflation dynamics (Section 3 — selected findings)
- Estimation approach:
  - Country-by-country Phillips curves for countries with at least 30 quarters.
  - Contributions C_{x,i,t} computed as C_{x,i,t} = beta_hat_{x,i} * x_{i,t} + beta_hat_{lag} * C_{x,i,t-1}.
  - Cross-country average contributions presented as deviations from the inflation target (for economies without explicit inflation targets, a three percent target is assumed).
- Drivers of the 2021-22 inflation surge:
  - Foreign price developments were a major driver in both AE and EE.
  - Domestic factors, including inflation expectations, played modest roles; inflation expectations remained well-anchored.
  - Commodity price surge, particularly in 2022, is a key foreign-driven factor.
  - Diminishing economic slack contributed negligibly, given the estimated flatness of Phillips curves, particularly in AE.
- Predicted deviations from target in 2022Q2 (Phillips curve simulations):
  - Average AE: predicted deviation of core inflation from target was 2¾ percentage points.
  - Average EE: predicted deviation of core inflation from target was 6 percentage points.
- Unexplained component:
  - The standard Phillips-curve model can account for at most two-thirds of the surge in inflation; residuals (actual minus model-predicted inflation) during 2022 are positive and very sizable.
  - Possible contributors to unexplained component:
    - Mismeasured economic slack (unemployment or output gaps poor proxies amid supply shocks).
    - High vacancy-to-unemployment ratios at end-2022 indicating tighter labor markets.
    - Unprecedented supply bottlenecks.
    - Discretionary policies tended to contain rather than amplify inflation and thus cannot explain the positive unexplained component.
  - Illustrative calculations suggest widespread input and labor shortages might have raised core inflation by up to 1.5 percentage point on average in Europe.
- Evidence of shifts:
  - Rolling panel estimates show increasing coefficient on lagged core (headline) inflation in recent rolling samples beginning with the 2018Q1-2021Q4 window.
  - Passthrough of global food prices increased sharply; energy passthrough rose more moderately and earlier.
  - Re-estimation with a post-pandemic dummy (quarters after 2019Q4) confirms statistically significant increases in inflation persistence and sensitivity to external price pressures.
  - High-inflation periods (inflation > 60th percentile) are associated with higher coefficient on past inflation, lower coefficient on inflation expectations, and greater sensitivity to food price shocks; effects stronger in EE.

### Global energy and food price inflation (Section 4 — summary)
- Contribution charts:
  - Contributions to core and headline inflation reported in percentage points, quarter-over-quarter annualized; 4 quarter moving average of quarter-over-quarter annualized.
  - Key components: Residual; Energy price; Food price; External price pressure; Unemployment gap; Expected inflation.
  - Time span in figures: 2000Q1–2022Q1 (axes label 2000Q1, 2005Q1, 2010Q1, 2015Q1, 2020Q1) and 2019–2022 panels for higher-frequency view.
- Headline inflation benchmark coefficients (selected, quarter-over-quarter annualized):
  - Unemployment gap:
    - (1) -0.412*** (0.072)
    - (2) -0.272*** (0.058)
    - (3) -0.781*** (0.230)
  - Lag of inflation:
    - (1) 0.460*** (0.065)
    - (2) 0.492*** (0.040)
    - (3) 0.450*** (0.130)
  - Inflation expectations:
    - (1) 0.540*** (0.065)
    - (2) 0.508*** (0.040)
    - (3) 0.550*** (0.130)
  - Food prices Lag 0:
    - (1) 0.159*** (0.039)
    - (2) 0.082*** (0.015)
    - (3) 0.231*** (0.067)
  - Energy prices:
    - (1) 0.152*** (0.014)
    - (2) 0.172*** (0.010)
    - (3) 0.140*** (0.027)
- Data sources listed in annex figures:
  - Core/Headline CPI: Haver Analytics
  - Three-year-ahead inflation expectations: Consensus Economics
  - Commodity price (food and energy): IMF, International Financial Statistics
  - Bilateral exchange rate against the US dollar: IMF, International Financial Statistics
  - External price pressure: IMF staff calculations
  - Producer price index: Haver Analytics
  - Bilateral exports and imports: IMF, Direction of Trade Statistics
  - Real GDP: IMF, World Economic Outlook Database
  - Unemployment rate: IMF, World Economic Outlook Database
  - Inflation target: Central Banks

*IMF Working Paper: The 2020-2022 Inflation Surge Across Europe: A Phillips-Curve-Based Dissection (excerpt).*

### Annex I. Data Sources ..................................................................................................

### Annex I. Data Sources

### Main objectives and scope
- Aim: Shed light on Europe’s post-COVID-19 inflation puzzle by exploring empirically the factors that contributed to cross-country dispersion in inflation and persistent overshooting through the lens of the standard Phillips curve.
- Focus: Differences in Phillips curves across countries and potential shifts in the inflation process since the onset of the COVID-19 pandemic.
- Sample period: 2000Q1-22Q2.
- Country coverage:
  - 24 advanced European economies (AE): Austria, Belgium, Cyprus, Czech Republic, Germany, Denmark, Estonia, Finland, France, Greece, Ireland, Israel, Italy, Lithuania, Latvia, Netherlands, Norway, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, and United Kingdom.
  - 7 emerging European economies (EE): Bulgaria, Croatia, Hungary, Poland, Romania, Russia and Türkiye.

### Key empirical framework and data construction
- Baseline model: New Keynesian Phillips curve augmented with:
  - lagged and expected inflation,
  - domestic economic slack (unemployment gap or output gap),
  - contemporaneous and four quarterly lags of energy and food price growth expressed in domestic currency and weighted by CPI shares,
  - an external price pressure index (extP) capturing import-weighted PPI changes of trading partners and relative bilateral exchange rate changes minus country i GDP deflator change,
  - country fixed effects and an error term.
- Inflation variable: quarter-over-quarter annualized core (headline) inflation in country i in quarter t.
- Inflation expectations: three-year-ahead inflation expectations (robust to using 1-year ahead).
- Slack measure: deviation from the Hodrick-Prescott (HP)-filtered unemployment rate (alternatively log output) with smoothing parameter of 1600.
- External price pressure (extP) defined as:
  - import-share weighted percent change in trading partners’ PPI,
  - plus import-share weighted relative changes in bilateral exchange rates against the US dollar,
  - minus percent change in country i GDP deflator.
- Global commodity prices: converted to local currency using bilateral US dollar exchange rates and weighted by energy/food share in country i's consumption basket.
- Lags: contemporaneous energy and food price increases plus four quarterly lags to capture gradual passthrough (food passthrough slower; lagged food significant up to fourth lag).
- Identification constraint: benchmark imposes constraint that sum of coefficients on past and expected inflation rates (β1+β2) equals one (Galí and Gertler, 1999). Relaxing this restriction yields similar findings (see robustness).

### Principal empirical findings
- Cross-country inflation outcomes:
  - Inflation exceeded 5 percent in almost all European economies by late summer 2022.
  - By summer 2022, headline inflation reached 20–25 percent in the Baltic countries, three to four times the rate in the lowest-inflation countries in the euro area.
  - Average core inflation in emerging European economies exceeded 20 percent by Q2 2022 — a 12 percentage points gap vis-à-vis advanced European countries, compared to a 2-3 percentage points gap prior to the pandemic.
- Benchmark Phillips-curve estimates (core inflation, panel):
  - A 1 percentage point increase in unemployment above its HP trend is associated with a decline in core inflation of 0.4 percentage points in Europe.
  - A 1 percentage point rise in inflation expectations is significantly associated with a 0.6 percentage point increase in core inflation in Europe.
- Heterogeneity between AE and EE:
  - The slope of the Phillips curve (coefficient on domestic slack) is steeper in EE than in AE.
  - Inflation expectations coefficient: 0.4 in EE versus 0.7 in AE (i.e., price-setting is less forward-looking in EE).
  - Inflation in EE responds more strongly to foreign price developments, especially global food prices.
  - These differences contributed to a larger passthrough of global commodity price shocks and a larger inflation surge in EE in 2021-22.
- Changes since the pandemic:
  - Evidence that inflation in Europe has become increasingly backward looking and more sensitive to food price shocks since the onset of the COVID-19 pandemic, consistent with higher persistence of inflation and greater passthrough of global food prices during past high-inflation episodes.
  - Higher vacancy-to-unemployment ratios by end-2022 suggest labor markets may have been tighter than unemployment dynamics alone indicate, contributing to larger-than-expected inflation.

### Robustness and alternative specifications
- Robustness checks:
  - Alternative estimators: OLS without constraint on expected and lagged inflation coefficients, median regression to reduce influence of extremes, and use of output gaps instead of unemployment gaps.
  - Across specifications, findings are robust: steeper Phillips curve slope in EE, larger role of lagged inflation, and larger role of foreign price developments in EE than in AE.
  - Note: coefficients on economic slack are not statistically significant for EE when measured by the output gap.
  - Relaxing the (β1+β2)=1 restriction produces similar conclusions on slack, external pressures and commodity prices, though in some cases β1 may exceed 1.

### Policy-relevant implications
- Divergence risk: Differences in Phillips curves point to a risk of persistent inflation divergence between euro area economies, with significantly higher inflation in newer member states (e.g., the Baltic countries).
- Monetary policy calibration:
  - Outside the euro area, emerging market economies may require a more forceful policy response to tame inflationary pressures due to steeper slack sensitivity and greater passthrough from external prices.
  - The increasing backward-looking nature of inflation warrants careful monitoring as it could signal changes in price-setting and wage bargaining, increasing risks of feedback loops between wages and prices and potential de-anchoring of inflation expectations.
- Forecasting and model choice: The heightened persistence and stronger passthrough of external commodity prices help explain repeated underprediction of inflation by models based on historical relationships; modelers and policymakers should account for evolving Phillips-curve parameters and external-price sensitivity.

_Italic: Source — IMF Working Paper excerpt (text provided)._

### 3.  Contribution to Inflation Dynamics

### 3.  Contribution to Inflation Dynamics

### Methodology and decomposition approach
- Phillips curves are estimated separately for each country with at least 30 quarters of data to obtain country-specific coefficients.
- Country-specific coefficients are used to compute Phillips-curve-predicted inflation and the contribution from each regressor to inflation in each quarter, following Yellen (2015).
- Because the Phillips curves include a lagged inflation term, dynamic simulations are performed to attribute past inflation fluctuations to movements in independent variables.
- The contribution of independent variables x to inflation dynamics in country i at time t, C_{x,i,t}, is computed as:
  - C_{x,i,t} = beta_hat_{x,i} * x_{i,t} + beta_hat_{lag} * C_{x,i,t-1}
  - where beta_hat_{x,i} is the coefficient on variable x from the country-by-country Phillips curve, and beta_hat_{lag} is the coefficient on lagged inflation.
- Cross-country average contributions to core and headline inflation are presented as deviations from the inflation target (for economies without explicit inflation targets, a three percent target is assumed).

### Main empirical findings on drivers of the 2021-22 inflation surge
- Foreign price developments were a major driver of rising inflation during 2021-22 in both advanced Europe (AE) and emerging Europe (EE).
- Domestic factors, including inflation expectations, played modest roles; inflation expectations remained well-anchored.
- The dramatic rise in commodity prices, particularly in 2022, is a key factor behind the foreign-price-driven contribution.
- The role of diminishing economic slack (captured by unemployment or output gaps) is estimated to be negligible given the estimated flatness of Phillips curves, particularly in AE.
- The Phillips curve simulations predict substantially higher inflation on average for EE compared to AE:
  - For the average AE, the predicted deviation of core inflation from target in 2022Q2 was 2¾ percentage points.
  - For the average EE, the predicted deviation of core inflation from target in 2022Q2 was 6 percentage points.
  - These differences persist even after factoring in the higher average inflation target in EE compared to AE.
- The larger predicted deviation in EE is consistent with greater sensitivity of inflation to commodity prices (especially food, but also fuel) and external price pressures (including exchange rate depreciation and its passthrough) in EE relative to AE.

### Unexplained component and potential sources
- The standard Phillips-curve model can account for at most two-thirds of the surge in inflation; estimated residuals (actual minus model-predicted inflation) during 2022 are positive and very sizable.
- Possible explanations for the large unexplained component include:
  - Mismeasured (over-estimated) economic slack: conventional unemployment or output gaps may be poor proxies during extraordinary supply shocks and structural labor market changes (e.g., pandemic-driven declines in labor supply).
  - High vacancy-to-unemployment ratios at end-2022 suggest labor markets may have been tighter than unemployment alone indicates.
  - Unprecedented supply bottlenecks, poorly captured by conventional Phillips curve models, may have contributed to price pressures.
  - Discretionary policies (temporary tax changes, price caps) tended to contain rather than amplify inflation and thus cannot explain the positive unexplained component.
- Illustrative calculations (noted in the source) suggest that widespread reported input and labor shortages might have raised core inflation by up to 1.5 percentage point on average in Europe.
- Cross-country heterogeneity in unexplained inflation reflects variation in the impact of these factors across countries.

### Evidence of shifts in Phillips curve relationships
- Rolling panel estimates for 28 countries over 16 quarters indicate possible post-pandemic shifts in Phillips curve coefficients:
  - The coefficient on lagged core (headline) inflation shows signs of increasing in the last 3 rolling sample periods, starting from the 2018Q1-2021Q4 window, suggesting inflation may have become more backward-looking.
  - Passthrough of global food prices to domestic inflation appears to have increased sharply.
  - The rise in passthrough of energy prices to core inflation is more moderate and started earlier during the pandemic.
  - These patterns are particularly pronounced in the EE sample.
- Re-estimation with interactions for a post-pandemic dummy (equal to 1 for quarters after 2019Q4) confirms statistically significant changes in inflation persistence and sensitivity to external price pressures (summarized in Table 3 of the source).
- Periods of high inflation (defined as inflation exceeding the 60th percentile observed in each country) are associated with:
  - A significantly higher coefficient on past inflation.
  - A correspondingly lower coefficient on inflation expectations.
  - Greater sensitivity of inflation to food price shocks.
  - These effects are more pronounced for the EE sample and align with historical evidence that inflation following surges tends to be more persistent.

### Implications and concluding observations
- Differences in sensitivity of inflation to traditional drivers help explain heterogeneity between AE and EE: inflation in EE is more sensitive to domestic slack and external price pressures, contributing to larger passthrough of global commodity shocks and higher inflation in EE.
- Across Europe, a sizable share of the 2022 inflation uptick cannot be explained by conventional drivers; inflation became more backward-looking and more sensitive to food price shocks after the onset of the COVID-19 pandemic.
- Models based on historical relationships with conventional drivers consistently underpredicted inflation outturns during 2021-22.
- It is too early to judge whether observed shifts signal long-lasting structural changes in the inflation process, but suggestive signs are worrisome:
  - Increased influence of past inflation could make reducing inflation from the multidecade highs of 2022 harder and costlier.
  - Greater passthrough of commodity prices to core inflation increases vulnerability to further adverse supply shocks such as new geopolitical or extreme weather events.

*Source: IMF Working Paper — "The 2020-2022 Inflation Surge Across Europe: A Phillips-Curve-Based Dissection", Section 3 (Contribution to Inflation Dynamics).*

### 4. Global energy and food price inflation

### 4. Global energy and food price inflation

### Contribution to inflation dynamics (figures)
- Charts report contributions to core and headline inflation (percentage points, quarter-over-quarter annualized; 4 quarter moving average of quarter-over-quarter annualized).
- Bars represent simple average contribution of each factor across advanced and emerging economies in Europe. Contributions computed from dynamic simulations of country-by-country Phillips curve regressions.
- Key labeled components in charts: Residual; Energy price; Food price; External price pressure; Unemployment gap; Expected inflation.
- Time span shown in figures covers 2000Q1–2022Q1 (figure axis labels include 2000Q1, 2005Q1, 2010Q1, 2015Q1, 2020Q1) and 2019–2022 period for higher-frequency panels.

### Benchmark Phillips Curve estimation (Table 1: dependent variable = Core inflation, quarter-over-quarter annualized)
- Sample: Unbalanced panel data for 24 advanced and 7 emerging economies in Europe, 2000Q1-2022Q2. Sum of coefficients on lagged and expected inflation constrained to be one. Robust standard errors in parentheses.
- Selected estimated coefficients (column layout: (1) All, (2) AE, (3) EE, (4) All with EE interactions):
  - Unemployment gap:
    - (1) -0.374*** (0.078)
    - (2) -0.337*** (0.097)
    - (3) -0.676*** (0.179)
    - (4) -0.337*** (0.097)
  - Unemployment gap x EE dummy:
    - (4) -0.339* (0.203)
  - Lag of inflation:
    - (1) 0.431*** (0.127)
    - (2) 0.287* (0.161)
    - (3) 0.581*** (0.105)
    - (4) 0.287* (0.162)
  - Lag of inflation x EE dummy:
    - (4) 0.295 (0.192)
  - Inflation expectations:
    - (1) 0.569*** (0.127)
    - (2) 0.713*** (0.161)
    - (3) 0.419*** (0.105)
    - (4) 0.713*** (0.162)
  - Inflation expectations x EE dummy:
    - (4) -0.295 (0.192)
  - Lag of external price pressure:
    - (1) 0.020*** (0.006)
    - (2) 0.009* (0.005)
    - (3) 0.037** (0.015)
    - (4) 0.009* (0.005)
  - Lag of external price pressure x EE dummy:
    - (4) 0.028* (0.015)
  - Food prices (lags 0–4):
    - Lag 0:
      - (1) 0.127*** (0.034)
      - (2) 0.065*** (0.015)
      - (3) 0.181*** (0.054)
      - (4) 0.065*** (0.015)
    - Lag 0 x EE dummy:
      - (4) 0.115** (0.056)
    - Lag 1:
      - (1) 0.078*** (0.020)
      - (2) 0.054*** (0.020)
      - (3) 0.075*** (0.027)
      - (4) 0.054*** (0.020)
    - Lag 1 x EE dummy:
      - (4) 0.020 (0.033)
    - Lag 2:
      - (1) 0.032 (0.022)
      - (2) 0.045*** (0.015)
      - (3) -0.001 (0.030)
      - (4) 0.045*** (0.015)
    - Lag 2 x EE dummy:
      - (4) -0.045 (0.033)
    - Lag 3:
      - (1) 0.070*** (0.021)
      - (2) 0.077*** (0.018)
      - (3) 0.065** (0.032)
      - (4) 0.077*** (0.018)
    - Lag 3 x EE dummy:
      - (4) 0.012 (0.037)
    - Lag 4:
      - (1) 0.042** (0.017)
      - (2) 0.040** (0.019)
      - (3) 0.053** (0.026)
      - (4) 0.040** (0.019)
    - Lag 4 x EE dummy:
      - (4) 0.012 (0.032)
  - Energy prices:
    - (1) 0.016 (0.011)
    - (2) 0.032*** (0.008)
    - (3) 0.021 (0.022)
    - (4) 0.032*** (0.008)
  - Energy prices x EE dummy:
    - (4) -0.011 (0.024)
- Observations:
  - (1) 2,210
  - (2) 1,707
  - (3) 503
  - (4) 2,210
- Country FE: Yes for all columns. Time FE: No for all columns.

### Robustness checks (Table 2)
- Dependent variable: Core inflation, quarter-over-quarter annualized. Advanced Europe and Emerging Europe samples reported across columns with alternate specifications (Const Reg, OLS, Median Reg, Output Gap).
- Selected coefficients (representative entries):
  - Unemployment Gap:
    - Column (1) -0.337*** (0.097)
    - Column (2) -0.305*** (0.064)
    - Column (3) -0.194*** (0.027)
    - Column (4) -0.676*** (0.179)
    - Column (5) -0.706** (0.244)
    - Column (6) -0.382*** (0.078)
  - Output Gap (when included):
    - 0.117*** (0.036)
    - 0.097 (0.076)
  - Lag of inflation:
    - Examples: 0.287* (0.161); 0.264* (0.129); 0.451*** (0.023); 0.310* (0.165); 0.581*** (0.105); 0.543*** (0.045); 0.495*** (0.046); 0.601*** (0.108)
  - Inflation expectations:
    - Examples: 0.713*** (0.161); 1.424** (0.528); 0.644*** (0.089); 0.690*** (0.165); 0.419*** (0.105); 0.656*** (0.089); 0.702*** (0.128); 0.399*** (0.108)
  - Lag of external price pressure:
    - Examples: 0.009* (0.005); 0.011** (0.005); 0.007*** (0.002); 0.005 (0.005); 0.037** (0.015); 0.044* (0.023); 0.011 (0.009); 0.030** (0.014)
  - Food prices (Lag 0 to Lag 4) and Energy Price also display consistent positive coefficients across robustness specifications (see table for exact values).
- Observations reported across columns (examples): 1,707; 503.
- Note: Standard errors corrected for heteroscedasticity and autocorrelations. All regressions include country fixed effects.

### Pre vs Post Covid comparisons (Table 3: Core inflation)
- Dependent variable: Core inflation, quarter-over-quarter annualized. Columns compare full sample and interactions with Post-covid dummy.
- Selected coefficients (panel):
  - Unemployment gap:
    - Pre/Post combined: -0.374*** (0.078); AE: -0.337*** (0.097); EE: -0.676*** (0.179)
    - Post-covid interactions: Unemployment gap x Post-covid dummy examples: 0.018 (0.226); 0.192 (0.187); -0.406 (0.955)
  - Lag of inflation:
    - Examples: 0.431*** (0.127); 0.287* (0.161); 0.581*** (0.105)
    - Lag x Post-covid dummy: 0.551** (0.256); 0.339* (0.176); 0.908** (0.394)
  - Inflation expectations:
    - Examples: 0.569*** (0.127); 0.713*** (0.161); 0.419*** (0.105)
    - Expectations x Post-covid dummy: -0.551** (0.256); -0.339* (0.176); -0.908** (0.394)
  - Lag of external price pressure:
    - Examples: 0.020*** (0.006); 0.009* (0.005); 0.037** (0.015)
    - Lag x Post-covid dummy: -0.037*** (0.014); -0.019** (0.010); -0.006 (0.036)
  - Food price Lag 0:
    - (1) 0.127*** (0.034); (2) 0.065*** (0.015); (3) 0.181*** (0.054)
    - Lag 0 x Post-covid dummy examples: 0.056 (0.098); -0.051 (0.039); 0.119 (0.133)
  - Energy prices:
    - Baseline: 0.016 (0.011); AE: 0.032*** (0.008); EE: 0.021 (0.022)
    - Energy prices x Post-covid dummy examples: 0.034 (0.041); 0.078*** (0.022); -0.050 (0.067)
  - Post-covid dummy:
    - Examples: -0.153 (0.347); -0.070 (0.263); 0.153 (0.923)
- Observations: 2,210 (All), 1,707 (AE), 503 (EE).

### High vs Low inflation periods (Table 4: Core inflation)
- Dependent variable: Core inflation, quarter-over-quarter annualized. High inflation dummy = 1 when core inflation exceeds 60th percentile of historical values for the country.
- Key coefficients:
  - Unemployment gap:
    - Baseline: -0.374*** (0.078); AE: -0.337*** (0.097); EE: -0.676*** (0.179)
    - Columns (4)–(6) (high-inflation interactions): -0.288*** (0.048); -0.278*** (0.048); -0.376*** (0.132)
  - Unemployment gap x high inflation dummy:
    - Examples: 0.274** (0.132); 0.433** (0.173); -0.235 (0.370)
  - Lag of inflation:
    - Examples baseline: 0.431*** (0.127); 0.287* (0.161); 0.581*** (0.105)
    - Lag x high inflation dummy examples: 0.527*** (0.121); 0.663*** (0.126); 0.269* (0.143)
  - Inflation expectations:
    - Baseline: 0.569*** (0.127); 0.713*** (0.161); 0.419*** (0.105)
    - Expectations x high inflation dummy:
      - -0.527*** (0.121); -0.663*** (0.126); -0.269* (0.143)
  - Lag of external price pressure:
    - Baseline examples: 0.020*** (0.006); 0.009* (0.005); 0.037** (0.015)
    - Lag x high inflation dummy examples: 0.009 (0.013); -0.013* (0.007); 0.044* (0.027)
  - Food prices and Energy prices display differential effects in high inflation regimes; examples:
    - Food price Lag 0 x high inflation dummy: 0.206*** (0.055); 0.089*** (0.018); 0.259*** (0.078)
    - Energy prices x high inflation dummy: -0.038* (0.023); -0.009 (0.015); -0.007 (0.039)
  - High inflation dummy coefficients:
    - Examples: 1.755*** (0.170); 1.947*** (0.134); 1.624*** (0.555)
- Observations: 2,210; AE: 1,707; EE: 503. Country FE: Yes. Time FE: No.

### Headline inflation results and annex (Tables A2–A4; Figure A1)
- Benchmark and robustness tables for headline inflation mirror core inflation structure; dependent variable = Headline inflation, quarter-over-quarter annualized. Sample: 24 advanced and 7 emerging economies in Europe, 2000Q1-2022Q2.
- Selected headline regression coefficients (Table A2):
  - Unemployment gap:
    - (1) -0.412*** (0.072)
    - (2) -0.272*** (0.058)
    - (3) -0.781*** (0.230)
    - (4) -0.272*** (0.058)
  - Lag of inflation:
    - (1) 0.460*** (0.065)
    - (2) 0.492*** (0.040)
    - (3) 0.450*** (0.130)
    - (4) 0.492*** (0.041)
  - Inflation expectations:
    - (1) 0.540*** (0.065)
    - (2) 0.508*** (0.040)
    - (3) 0.550*** (0.130)
    - (4) 0.508*** (0.041)
  - Lag of external price pressure:
    - (1) 0.018* (0.010)
    - (2) 0.006 (0.007)
    - (3) 0.036 (0.024)
    - (4) 0.006 (0.007)
  - Food prices Lag 0:
    - (1) 0.159*** (0.039)
    - (2) 0.082*** (0.015)
    - (3) 0.231*** (0.067)
    - (4) 0.082*** (0.015)
  - Energy prices:
    - (1) 0.152*** (0.014)
    - (2) 0.172*** (0.010)
    - (3) 0.140*** (0.027)
    - (4) 0.172*** (0.010)
- Figures in Annex I and II present contributions to headline inflation analogously to core inflation charts; variable sources listed (Core/Headline CPI: Haver Analytics; Three-year-ahead inflation expectations: Consensus Economics; Commodity price (food and energy): IMF, International Financial Statistics; Bilateral exchange rate against the US dollar: IMF, International Financial Statistics; External price pressure: IMF staff calculations; Producer price index: Haver Analytics; Bilateral exports and imports: IMF, Direction of Trade Statistics; Real GDP: IMF, World Economic Outlook Database; Unemployment rate: IMF, World Economic Outlook Database; Inflation target: Central Banks).

*IMF Working Paper: The 2020-2022 Inflation Surge Across Europe: A Phillips-Curve-Based Dissection (chapter: 4. Global energy and food price inflation).*

### References

### References

### Inflation dynamics and the Phillips curve
- Abdih, Yasser, Li Lin, and Anne-Charlotte Paret, 2018. “Understanding Euro Area Inflation Dynamics: Why So Low for So Long?”, IMF Working Paper 18/25, International Monetary Fund, Washington DC.  
- Ball, Laurence. 2006. “Has Globalization Changed Inflation?” NBER Working Paper 12687, National Bureau of Economic Research, Cambridge, MA.  
- Ball, Laurence, Daniel Leigh, and Prachi Mishra. 2022. “Understanding US Inflation during the COVID Era.” BPEA Article, September 7, 2022.  
- Bems, Rudolfs, Francesca Caselli, Francesco Grigoli, and Bertrand Gruss. 2021. “Expectations’ Anchoring and Inflation Persistence.” Journal of International Economics 132(2021): 103516.  
- Bems, Rudolfs, Francesca Caselli, Francesco Grigoli, and Bertrand Gruss. 2022. “Is Inflation Domestic or Global? Evidence from Emerging Markets.” International Journal of Central Banking 18 (4): 125-163.  
- Crump, Richard K., Stefano Eusepi, Marc Giannoni, and Ayşegül Şahin. 2022. “The Unemployment-Inflation Trade-off Revisited: The Phillips Curve in COVID Times.” NBER Working Paper 29785, National Bureau of Economic Research, Cambridge, MA.  
- Coibion, Olivier and Yuriy Gorodnichenko. 2015. “Is the Phillips Curve Alive and Well after All? Inflation Expectations and the Missing Disinflation”. American Economic Journal: Macroeconomics 7(1): 197–232.  
- Coibion, Olivier, Yuriy Gorodnichenko, and Mauricio Ulate. 2019. “Is Inflation Just Around the Corner? The Phillips Curve and Global Inflationary Pressures.” AEA Papers and Proceedings 109: 465-69.  
- Forbes, Kristin. 2020. “Inflation Dynamics: Dead, Dormant, Or Determined Abroad?” Brookings Papers on Economic Activity, Fall 2019 Meetings:. 257-319.  
- Hooper, Peter, Frederic Mishkin, and Amir Sufi. 2020. “Prospects for Inflation in a High-Pressure Economy: Is the Phillips Curve Dead or Is It Just Hibernating?” Research in Economics 74 (1): 26–62.  
- McGregor, Thomas, and Frederik Toscani. 2022. “A Bottom-Up Reduced Form Phillips Curve for the Euro Area.” IMF Working Paper 22/260, International Monetary Fund, Washington, DC.  
- Schmitt-Grohé, Stephanie, and Martín Uribe. 2022. “What Do Long Data Tell Us About the Inflation Hike Post COVID-19 Pandemic?,” NBER Working Paper 30357, National Bureau of Economic Research, Cambridge, MA.  
- European Central Bank. 2011. Inflation in the Euro Area and the United States: An Assessment Based on the Phillips Curve, Monthly Bulletin, Box 1 (June).  
- European Central Bank. 2017. Domestic and Global Drivers of Inflation in the Euro Area, Economic Bulletin, Article, Issue 4.

### Global factors, globalization, and international transmission
- Auer, Raphael, Claudio Borio, and Andrew Filardo. 2017. “The Globalisation of Inflation: the Growing Importance of Global Value Chains.” BIS Working Paper No 602.  
- Borio, Claudio, and Andrew Filardo. 2007. “Globalization and Inflation: New Cross-Country Evidence on the Global Determinants of Domestic Inflation.” BIS Working Paper 227, Bank for International Settlements, Basel.  
- Borio, Claudio. 2017. “Through the Looking Glass.” OMFIF City Lecture, September 22.  
- Ball, Laurence. 2006. “Has Globalization Changed Inflation?” NBER Working Paper 12687, National Bureau of Economic Research, Cambridge, MA.  
- Calza, Alessandro. 2008. “Globalisation, Domestic Inflation and Global Output Gaps—Evidence from the Euro Area,” ECB Working Paper 890, European Central Bank, Frankfurt.  
- Di Giovanni, Julian, Sebnem Kalemli-Özcan, Alvaro Silva, and Muhammed A. Yildirim. 2022. “Global Supply Chain Pressures, International Trade, and Inflation,” NBER Working Paper 30240, National Bureau of Economic Research, Cambridge, MA.  
- Forbes, Kristin. 2020. “Inflation Dynamics: Dead, Dormant, Or Determined Abroad?” Brookings Papers on Economic Activity, Fall 2019 Meetings:. 257-319.  

### Energy, oil prices, and supply bottlenecks
- Ari, Anil, Nicolas Arregui, Simon Black, Oya Celasun, Dora Iakova, Aiko Mineshima, Victor Mylonas, Ian Parry, Iulia Teodoru, and Karlygash Zhunussova. 2022. “Surging Energy Prices in Europe in the Aftermath of the War: How to Support the Vulnerable and Speed up the Transition Away from Fossil Fuels.” IMF Working Paper 22/152, International Monetary Fund, Washington DC.  
- Baba, Chikako, and Jaewoo Lee. 2022. “Second-Round Effects of Oil Price Shocks—Implications for Europe’s Inflation Outlook.” IMF Working Paper 22/173, International Monetary Fund, Washington DC.  
- Conflitti, Cristina, and Matteo Luciani. 2019. “Oil Price Pass-through into Core Inflation.” The Energy Journal 40 (6): 221-47.  
- Kilian, Lutz, and Xiaoqing Zhou. 2021. “The Impact of Rising Oil Prices on US Inflation and Inflation Expectations in 2020–23.” Federal Reserve Bank of Dallas Working Paper 2116, Federal Reserve Bank of Dallas, Dallas.  
- Celasun, Oya, Niels-Jakob Hansen, Aiko Mineshima, Mariano Spector, and Jing Zhou. 2022. “Supply Bottlenecks: Where, Why, How Much, and What Next?” IMF Working Paper 22/31, International Monetary Fund, Washington, DC.  
- Bank for International Settlements (BIS). 2022. “Inflation: A Look under the Hood.” In Annual Economic Report June 2022, 41–73. Basel, Switzerland: Bank for International Settlements.

### Labor market, wages, and long COVID
- Amiti, Mary, Sebastian Heise, Fatih Karahan, and Ayşegül Şahin. 2022. “Pass-through of Wages and Import Prices Has Increased in the Post-COVID Period.” Liberty Street Economics, Federal Reserve Bank of New York, August 23.  
- Amiti, Mary, Sebastian Heise, and Aidan Wang. 2021. “High Import Prices along the Global Supply Chain Feed through to US Domestic Prices.” Liberty Street Economics, Federal Reserve Bank of New York. November 8, 2021.  
- Boranova, Vizhdan, Raju Huidrom, Sylwia Nowak, Petia Topalova, Volodymyr Tulin, and Richard Varghese. 2021. “Wage Growth and Inflation in Europe: A Puzzle?” Oxford Economic Papers 73 (4): 1427–53.  
- Duval, Romain, Yi Ji, Longji Li, Myrto Oikonomou, Carlo Pizzinelli, Ippei Shibata, Alessandra Sozzi, and Marina M. Tavares. 2022. “Labor Market Tightness in Advanced Economies.” IMF Staff Discussion Note 2022/001, International Monetary Fund, Washington, DC.  
- Faberman, R. Jason, Andreas I. Mueller, and Ayşegül Şahin. 2022. “Has the Willingness to Work Fallen during the Covid Pandemic?” NBER Working Paper 29784, National Bureau of Economic Research, Cambridge, MA.  
- Bach, Katie. 2022. “New Data Shows Long COVID is Keeping as Many as 4 Million People Out of Work.” Brookings Report, August 24, 2022.  
- Waters, Tom, and Thomas Wernham. 2022. “Long COVID and the Labour Market.” IFS Briefing Note BN246, Institute for Fiscal Studies, London.  
- Schwartzman, Felipe, and Sonya Ravindranath Waddell. 2022. “Are firms factoring increasing inflation into their prices?”, Federal Reserve Bank of Richmond Economic Brief, no 22-08.

### Central banks, institutions, and policy perspectives
- Carney, Mark. 2017. “[De]Globalisation and Inflation.” Speech at the 2017 IMF Michel Camdessus Central Banking Lecture (September 18).  
- Yellen, Janet L. 2015. “Inflation Dynamics and Monetary Policy.” Speech at the Philip Gamble Memorial Lecture, University of Massachusetts, Amherst, MA, September 24.  
- International Monetary Fund. 2018. “Challenges for Monetary Policy in Emerging Markets as Global Financial Conditions Normalize.” Chapter 3 in World Economic Outlook, October 2018, International Monetary Fund, Washington, DC.  
- International Monetary Fund. 2021. “Inflation Scares.” Chapter 2 in World Economic Outlook, October 2021, International Monetary Fund, Washington, DC.  
- International Monetary Fund. 2022a. “Inflation in Europe: Assessment, Risks, and Policy Implications.” Chapter 2 in European Regional Economic Outlook, October 2022, International Monetary Fund, Washington, DC.  
- International Monetary Fund. 2022b. “Wage Dynamics Post-COVID-19 and Wage-Price Spiral Risks.” Chapter 2 in World Economic Outlook, October 2022, International Monetary Fund, Washington, DC.  
- Gopinath, Gita. 2022. “How Will the Pandemic and War Shape Future Monetary Policy.” PowerPoint presentation presented at the Jackson Hole Economic Policy Symposium, Jackson Hole, WY, August 26.  
- Chahad, Mohammed, Anna-Camilla Hofmann-Drahonsky, Baptiste Meunier, Adrian Page and Marcel Tirpák. 2022. “What Explains Recent Errors in the Inflation Projections of Eurosystem and ECB staff?”, Box 5 in ECB Economic Bulletin, Issue 3/2022.

### Other empirical and methodological contributions
- Blanco, Andrés, Pablo Ottonello, and Tereza Ranosova. 2022. “The Dynamics of Large Inflation Surges.” NBER Working Paper 30555, National Bureau of Economic Research, Cambridge, MA.  
- Calza, Alessandro. 2008. “Globalisation, Domestic Inflation and Global Output Gaps—Evidence from the Euro Area,” ECB Working Paper 890, European Central Bank, Frankfurt.  
- Caporale, Guglielmo Maria, Juan Infante, Luis Gil-Alana and Raquel Ayestaran. 2022. “Inflation Persistence in Europe: The Effects of the Covid-19 Pandemic and of the Russia-Ukraine War.” CESifo Working Paper No. 10071, Munich, Germany.  
- Ciccarelli, Matteo and Chiara Osbat. 2017. “Low Inflation in the Euro Area: Causes and Consequences.” Occasional Paper Series No 181, European Central Bank, Frankfurt.  
- Koester, Gerritt, Eliza Lis, Christiane Nickel, Chiara Osbat, and Frank Smets. 2021. “Understanding Low Inflation in the Euro Area from 2013 to 2019: Cyclical and Structural drivers”, ECB Occasional Paper 280, European Central Bank, Frankfurt.  
- Hilscher, Jens, Alon Raviv, and Ricardo Reis. 2022. “How Likely is an Inflation Disaster?” Unpublished

*Source: References section of wpiea2023030-print-pdf*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023030-print-pdf.pdf_