## 1. Introduction

## Source details

**Canonical URL:** [1. Introduction](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023190-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023190-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023190-print-pdf.pdf.json)

---

### Background and recent episode
- The inflation shock that started in 2021 followed a long period of price stability and surprised forecasters and central banks as inflation climbed higher and proved more persistent than expected over 2022–23.
- In 2022, inflation peaked above 8 percent in OECD countries—its highest rate since 1984—and approached double-digits in emerging market and developing economies for the first time since the 1990s.
- Contributing factors to the 2021–23 surge: Russia’s war on Ukraine (an energy and food price shock), pandemic-related supply chain strains and rotation of consumption from services to goods, faster-than-expected demand recovery (reinforced by demand-stimulating fiscal policies), and nominal wage increases amid unprecedented labor market tightness.

### Policy responses and trade-offs
- Policymakers shifted from accommodative post-pandemic policies to policy tightening, with central banks raising interest rates to suppress fast-growing prices.
- Tightening posed trade-offs:
  - heightened financial stability risks from rapid monetary tightening;
  - fiscal policy risk of adding to price pressures while supporting households and firms facing higher energy and food bills;
  - an accentuated output-inflation tradeoff given the prominent supply-side inflation component.
- Repeated inflation surprises raised concerns about central bank credibility and the limitations of macroeconomic models calibrated to “normal times.”

### Purpose, sample, and core questions
- Main purpose: draw on historical episodes to inform current policy debates.
- Sample: identifies 111 inflation shock episodes in 56 countries during 1973–2014; over half (61 shocks, 55 percent) linked to the 1973 and 1979 oil crises.
- Baseline episode definition: increases in the inflation rate by at least 2 percentage points within a year after periods of relatively stable inflation (ΔΠ_t ≥2%); focus on an 11-year window [T−5, T+5].
- Core questions:
  - (i) How long did inflation take to return to its pre-shock rate (or “resolve”)?
  - (ii) Which policies were associated with resolving inflation?
  - (iii) Did resolving inflation involve sacrifice in terms of growth and unemployment?

### Seven stylized facts (documented)
- 1. Inflation is persistent, especially after a terms-of-trade shock.
- 2. Most unresolved inflation episodes involved “premature celebrations”, where inflation declined initially only to plateau at an elevated level or re-accelerate later.
- 3. Countries that resolved inflation had tighter monetary policy. There is further, weaker evidence for tighter fiscal policy and better inflation anchoring pre-shock.
- 4. Countries that resolved inflation implemented restrictive policies more consistently over time, as captured by the variance of the policy stance over the post-shock years.
- 5. Countries that resolved inflation contained nominal exchange rate depreciation.
- 6. Countries that resolved inflation had lower nominal wage growth. This did not translate into lower real wage growth, as lower nominal wage growth coincided with lower inflation.
- 7. Countries that resolved inflation experienced lower growth in the short-term but not over the 5-year horizon, potentially indicating that the benefits of macroeconomic stability and policy credibility over time offset the costs of a tighter policy stance.

### Caveats and scope
- Analysis focuses on empirical associations and does not ascribe causal interpretation.
- Historical data limitations: much of the analysis focuses on bivariate correlations rather than multivariate causal inference.
- Robustness checks include verification for a sub-sample of smaller inflation shocks (ΔΠ_T <10 percentage points) and for a sub-sample of free-floating exchange rate regimes.
- The paper notes structural differences between today’s economies and those of the 1970s and 1980s (sectoral structure, experience of relative economic stability, higher policy credibility).

### Data and methods (summary)
- Primary data sources: IMF’s World Economic Outlook (WEO) database, supplemented by BIS, Haver Analytics, OECD, Ilzetzki et al. (2019), and “Public Finances in Modern History” (PFMH, Mauro et al., 2015).
- Key variables from WEO: CPI inflation, real GDP growth, unemployment rates, output gap, current account and trade balances.
- Wage data: OECD Employment Outlook database for nominal and real wages.
- Monetary indicators: WEO for broad money growth; short-term interest rates extended using BIS, Haver Analytics, and OECD MEI.
- Proxies and constructed variables:
  - central bank credibility and inflation anchoring proxied by medium-term averages and volatility of CPI inflation (WEO) and the central bank independence index from Romelli (2022);
  - fiscal variables from WEO and PFMH including general government overall and primary fiscal balances, primary expenditures, gross debt-to-GDP ratios; CAPB constructed using Hamilton (2018) and Hodrick and Prescott (1997) filters;
  - exchange rates: BIS data for real and nominal effective exchange rates; Ilzetzki et al. (2019) used to construct exchange rate peg dummy.
- Empirical approach:
  - Difference-in-differences-style comparison between resolved and unresolved episodes using differences between post-shock and pre-shock conditions (pre: average over T−2,T−1; post: average over T,...,T+5).
  - Winsorize data at 90 percent to control for outliers.
  - Statistical test: t-test on coefficient β̂ from regression ΔX_i = a + β D_i + ε_i, with robust standard errors.

*Source: wpiea2023190-print-pdf — 1. Introduction.*

### 1. Introduction ........................................................................................................

### 1. Introduction

### Background and recent episode
- The inflation shock that started in 2021 followed a long period of price stability and surprised forecasters and central banks as inflation climbed higher and proved more persistent than expected over 2022–23.
- In 2022, inflation peaked above 8 percent in OECD countries—its highest rate since 1984—and approached double-digits in emerging market and developing economies for the first time since the 1990s.
- Several factors contributed to the 2021–23 surge: Russia’s war on Ukraine (an energy and food price shock), pandemic-related supply chain strains and rotation of consumption from services to goods, faster-than-expected demand recovery (reinforced by demand-stimulating fiscal policies), and nominal wage increases amid unprecedented labor market tightness.

### Policy responses and trade-offs
- Policymakers shifted from accommodative post-pandemic policies to policy tightening, with central banks raising interest rates to suppress fast-growing prices.
- Tightening posed acute trade-offs: heightened financial stability risks from rapid monetary tightening; fiscal policy risk of adding to price pressures while supporting households and firms facing higher energy and food bills; and an accentuated output-inflation tradeoff given the prominent supply-side inflation component.
- Repeated inflation surprises raised concerns about central bank credibility and the limitations of macroeconomic models calibrated to “normal times.”

### Purpose, sample, and core questions
- Main purpose: draw on historical episodes to inform current policy debates.
- The paper identifies over 100 inflation shock episodes in advanced and emerging economies between 1970 and today; over half of these episodes are linked to the 1973 and 1979 oil crises.
- Inflation shocks are defined in the baseline as increases in the inflation rate by at least 2 percentage points within a year after periods of relatively stable inflation.
- The paper asks three questions:
  - (i) How long did inflation take to return to its pre-shock rate (or “resolve”)?
  - (ii) Which policies were associated with resolving inflation?
  - (iii) Did resolving inflation involve sacrifice in terms of growth and unemployment?

### Seven stylized facts (documented)
- 1. Inflation is persistent, especially after a terms-of-trade shock.
- 2. Most unresolved inflation episodes involved “premature celebrations”, where inflation declined initially only to plateau at an elevated level or re-accelerate later.
- 3. Countries that resolved inflation had tighter monetary policy. There is further, weaker evidence for tighter fiscal policy and better inflation anchoring pre-shock.
- 4. Countries that resolved inflation implemented restrictive policies more consistently over time, as captured by the variance of the policy stance over the post-shock years.
- 5. Countries that resolved inflation contained nominal exchange rate depreciation.
- 6. Countries that resolved inflation had lower nominal wage growth. This did not translate into lower real wage growth, as lower nominal wage growth coincided with lower inflation.
- 7. Countries that resolved inflation experienced lower growth in the short-term but not over the 5-year horizon, potentially indicating that the benefits of macroeconomic stability and policy credibility over time offset the costs of a tighter policy stance.

### Caveats and scope
- Analysis focuses on empirical associations and does not ascribe causal interpretation.
- Due to historical data limitations, much of the analysis focuses on bivariate correlations rather than multivariate causal inference.
- The authors conduct robustness checks to address confounding factors, including verifying results for a sub-sample of smaller inflation shocks and for a sub-sample of free-floating exchange rate regimes.
- The paper notes structural differences between today’s economies and those of the 1970s and 1980s (sectoral structure, experience of relative economic stability, higher policy credibility).

### Related literature (high-level links)
- Contrasts with Cecchetti et al. (2023): this paper focuses on inflation shocks (a larger sample) rather than selected disinflation episodes and finds that successful disinflations are not associated with lower growth or higher unemployment over a 5-year horizon.
- Contrasts with Blanco et al. (2022): that study begins in the 1990s and includes many crisis-related episodes; by contrast the present baseline emphasizes the 1973–79 oil shocks.
- Situates findings within three strands: determinants of inflation (commodity prices, supply disruptions, fiscal/monetary stances, wages, expectations), policies to control inflation (monetary policy, role of fiscal policy), and literature on oil-shock inflation in the 1970s.

### Data and methods (summary)
- Primary data sources: IMF’s World Economic Outlook (WEO) database, supplemented by BIS, Haver Analytics, OECD, Ilzetzki et al. (2019), and “Public Finances in Modern History” (PFMH, Mauro et al., 2015).
- Key variables from WEO: CPI inflation, real GDP growth, unemployment rates, output gap, current account and trade balances.
- Nominal and real wage data: OECD Employment Outlook database.
- Monetary indicators: WEO for broad money growth; short-term interest rates extended using BIS, Haver Analytics, and OECD Main Economic Indicators (MEI).
- Proxies: central bank credibility and inflation anchoring proxied by medium-term averages and volatility of CPI inflation (WEO) and the central bank independence index from Romelli (2022).
- Fiscal variables: combined WEO and PFMH for general government overall and primary fiscal balances, primary expenditures, gross debt-to-GDP ratios; cyclically adjusted primary balances (CAPB) constructed using Hamilton (2018) and Hodrick and Prescott (1997) filters.
- Exchange rates: BIS data for real and nominal effective exchange rates; Ilzetzki et al. (2019) used to construct a dummy for exchange rate pegs.

*Source: wpiea2023190-print-pdf — 1. Introduction.*

### Section 3.2 further elaborates on our selection methodology and the resultant sample of inflation shock episodes.

### wpiea2023190-print-pdf - Section 3.2: Episode Selection and Sample (with related empirical strategy and initial results)

### Episode selection methodology
- Baseline algorithm identifies inflation shock episodes dating back to 1973.
- Step 1: Select country-year pairs where average annual inflation Π_t rises by at least 2 percentage points (i.e., ΔΠ_t ≥2%). Define T as the start year of the inflation shock episode, ΔΠ_T as the inflation shock size, and focus on an 11-year window [T−5, T+5].
- Step 2: Remove episodes in low-income, undiversified, or non-market economies, and episodes where the time window coincides with economic disruption related to an armed conflict or the COVID-19 pandemic. Excluding the pandemic effectively limits coverage to episodes that start prior to 2015.
  - Undiversified defined as: real per capita GDP in PPP averages less than 20 percent of the US per capita GDP in the 5 years prior to the episode start date; or fossil fuels and other commodities are over 20 percent of net exports and/or rents from these amount to more than 20 percent of GDP; or population less than 1 million (not applied to advanced economies such as Cyprus and Malta).
  - Non-market economies include planned economies and those in early stages of transition during 1990–2000.
- Step 3: Remove episodes where post-shock inflation remains low: Π_T <3%, interpreted as re-inflation from a deflationary shock (Π_{T−1} <1% by definition).
- Step 4: Remove episodes where pre-shock inflation is too high: avg(Π_{T−1}, Π_{T−2}) >25%.
- Step 5: Remove episodes where the inflation shock is a reversion to recent high inflation: Π_T < max(Π_{T−2}, Π_{T−3}). (Not applied to episodes dated between 1973–82.)
- Step 6: Remove other episodes where visual inspection suggests persistent inflation volatility or that the episode follows an earlier, larger inflation shock. This final step removes 46 episodes, 35 of which are due to overlaps with other episodes.
- The choice of a 5-year window reflects a trade-off between having a sufficiently wide window and the risk of conflating consecutive inflation shocks and unrelated macroeconomic developments.

### Sample construction and definitions
- The algorithm selects 111 inflation shocks in 56 countries during 1973–2014.
- Of these, 61 shocks (55 percent of the sample) are episodes associated with the 1973 and 1979 oil crises.
- Episodes are categorized as “resolved” if inflation falls to within 1 percentage point of its pre-shock rate by the end of the 5-year window: Π_{T+5} ≤ Π_{T−1} + 1%.
- The dataset excludes episodes that start after 2014 due to excluding COVID-19 overlaps.
- Appendix III lists selected episodes and plots inflation trajectories for each episode; Appendix II details episodes removed under steps 2–6.

### Descriptive statistics (from Table 1 and related text)
- Full sample: No. of episodes = 111.
  - Pre-shock avg. (T−5 to T−1): Mean = 6.8; Median = 5.1; Std. dev. = 5.7.
  - Post-shock avg. (T to T+5): Mean = 11.4; Median = 8.4; Std. dev. = 10.4.
  - Shock size (ΔΠ_T): Mean = 6.4; Median = 4.0; Std. dev. = 8.2.
  - Years to resolve: n.a.
- Inflation resolved (full sample): No. of episodes = 64 (57.7%).
  - Pre-shock avg.: Mean = 6.0; Median = 4.9; Std. dev. = 5.1.
  - Post-shock avg.: Mean = 7.0; Median = 5.1; Std. dev. = 5.0.
  - Shock size: Mean = 4.0; Median = 3.5; Std. dev. = 2.4.
  - Years to resolve: Mean = 3.2; Median = 3.0; Std. dev. = 1.5.
- Inflation unresolved (full sample): No. of episodes = 47 (42.3%).
  - Pre-shock avg.: Mean = 7.7; Median = 6.1; Std. dev. = 6.4.
  - Post-shock avg.: Mean = 17.4; Median = 12.9; Std. dev. = 12.7.
  - Shock size: Mean = 9.6; Median = 5.8; Std. dev. = 11.6.
  - Years to resolve: n.a.
- 1973–79 oil crises sample: No. of episodes = 61.
  - Pre-shock avg.: Mean = 7.4; Median = 6.2; Std. dev. = 4.3.
  - Post-shock avg.: Mean = 14.3; Median = 11.4; Std. dev. = 10.9.
  - Shock size: Mean = 7.2; Median = 4.8; Std. dev. = 9.4.
  - Years to resolve: n.a.
- Inflation resolved (1970s sample): No. of episodes = 29 (47.5%).
  - Pre-shock avg.: Mean = 7.3; Median = 6.6; Std. dev. = 3.6.
  - Post-shock avg.: Mean = 8.8; Median = 8.4; Std. dev. = 3.5.
  - Shock size: Mean = 4.6; Median = 3.9; Std. dev. = 2.9.
  - Years to resolve: Mean = 3.4; Median = 3.0; Std. dev. = 1.1.
- Inflation unresolved (1970s sample): No. of episodes = 32 (52.5%).
  - Pre-shock avg.: Mean = 7.5; Median = 6.2; Std. dev. = 4.9.
  - Post-shock avg.: Mean = 19.4; Median = 15.1; Std. dev. = 12.9.
  - Shock size: Mean = 9.5; Median = 6.1; Std. dev. = 12.3.
  - Years to resolve: n.a.
- The 1970s sub-sample has higher pre- and post-shock inflation and larger shock sizes on average; shocks take longer to resolve and are less likely to be resolved within 5 years.
- The authors note concerns that larger initial shock sizes in unresolved episodes could confound comparisons; Section 5 shows findings are robust to excluding episodes with large shock sizes (i.e., ΔΠ_T ≥10%), after which average shock sizes are comparable between resolved and unresolved episodes.

### Empirical strategy and measurement
- Focus: associations of inflation trajectories with macroeconomic fundamentals and policy settings during inflation shocks.
- Baseline focuses on the inflation shocks that coincide with the 1973–79 oil crises to aid identification and comparability; robustness verified on full sample.
- Data constraints for 1970s sample: although 61 relevant shocks are identified, data availability reduces usable country observations to between 13 to 60 depending on variables, with 42 as the median number of observations.
- Approach: difference-in-differences-style comparison between resolved and unresolved episodes using differences between post-shock and pre-shock conditions.
  - Use averages over years (T−2, T−1) for pre-shock and (T,...,T+5) for post-shock.
  - Winsorize data at 90 percent to control for outliers.
- Statistical test: t-test on coefficient β̂ from regression ΔX_i = a + β D_i + ε_i, where D_i is a dummy for resolved inflation shocks and robust standard errors are used.
- Monetary conditions and policy stance measures:
  - Real short-term rates and broad money growth rates.
  - Deviations of short-term nominal rates from Taylor Rule-implied policy rates using both output and unemployment-gap-based Taylor rule formulations:
    - R_{i,t}^{TR,Ỹ} = R* + φ_Π (Π_{i,t} − Π*) + φ_{Ỹ} Ỹ_{i,t}
    - R_{i,t}^{TR,u} = R* + φ_Π (Π_{i,t} − Π*) + φ_u (u_t − u*)
    - Coefficients calibrated to φ_Π = 1.5 and φ_{Ỹ} = 1.0 based on Hofmann and Bogdanova (2012) and φ_u = 2 φ_{Ỹ} using Okun’s law.
    - Neutral interest rate R*, implied-inflation target Π*, and natural unemployment rate u* are assumed time-invariant and eliminated when taking differences.
  - Real interest rates estimated ex-post as nominal interest rates less inflation (both in year t).
- Fiscal stance captured using primary fiscal balance and the cyclically adjusted primary balance (CAPB); primary expenditure and government debt-to-GDP ratio also used in Appendix IV.
- Inflation anchoring proxied by pre-shock inflation volatility: variance of inflation between T−5 and T−1 normalized by its mean over the same period.
  - Robustness checks include pre-shock medium-term inflation rate (mean inflation) and central bank independence index from Romelli (2022), averaged over T−5 to T−1.
- Consistency of policy response measured by post-shock variance of year-on-year first differences in policy stance variables (excluding variance of the stance itself to avoid capturing steady, progressive tightening).

### Key empirical findings (Section 4: The Seven Stylized Facts — Facts 1–3)
- Fact 1: Inflation persistence
  - Only in under 60 percent of episodes in the full sample (64 out of 111) was inflation resolved within 5 years after a shock.
  - For resolved episodes the mean years to resolve = 3.2 (3.4* for episodes starting in 1973–82); median = 3 (3*).
  - In the 1973–79 oil crises sub-sample less than 50 percent of episodes were resolved in 5 years (29 out of 61) and disinflation took on average 3.5 years.
  - Rapid resolution within one year occurred in 10 percent of episodes (12 out of 111), but 7 of these 12 corresponded to the Global Financial Crisis (GFC) or the Asian Financial Crisis (AFC) where demand collapse produced a rapid disinflation (“hard landing”). The remaining 5 rapid disinflations were associated with anomalous circumstances. These 12 episodes are excluded from the main empirical analysis.
- Fact 2: “Premature celebrations” in unresolved episodes
  - About 90 percent of unresolved episodes (42 out of 47 in the full sample; 28 out of 32 during the 1973–79 oil crises) saw inflation decline materially within the first three years after the initial shock, then plateau or re-accelerate.
  - Definition used: "premature celebrations" are unresolved episodes where inflation declines by at least 1 percentage point or by more than 10 percent of its initial increase within 3 years after the shock. Results are robust to a more stringent definition (decline by 2 percentage points or by 20 percent of the initial increase).
- Fact 3: Monetary policy tightening associated with resolution
  - Focus on 1973–79 oil crises subset for identification; results are robust in the full sample (Section 5).
  - Countries that resolved inflation tightened monetary policy substantially relative to pre-shock stance:
    - On average, countries that resolved inflation raised their effective real short-term interest rate by about 1 percentage point compared to the pre-shock stance.
    - In countries that did not resolve inflation, the real rate was 4.5 percentage points lower on average compared to pre-shock. This difference is statistically significant at the 1 percent level.
  - Broad money growth:
    - Countries that resolved inflation reduced broad money growth by 4 percentage points compared to pre-shock.
    - Countries that did not resolve inflation saw money growth accelerate by 3 percentage points compared to pre-shock. This difference is significant at the 25 percent level.
  - Taylor rule-adjusted results:
    - Countries that resolved inflation tightened their Taylor rule-adjusted interest rate stance by 1 or 2 percentage points (both output gap- and unemployment gap-based formulations).
    - Countries that did not resolve inflation loosened their Taylor rule-adjusted interest rate stance by 8 and 6 percentage points, respectively. These differences are significant at the 1 percent level.
  - Fiscal policy (illustrative, less significant):
    - Countries that resolved inflation also tightened fiscal policy on average, whereas fiscal positions in non-resolving countries were largely unchanged. The difference is statistically insignificant and economically small; results are not robust across fiscal metrics and samples (see Appendix IV).

*Source: wpiea2023190-print-pdf (selected sections 3.2–4) — IMF Working Paper "One Hundred Inflation Shocks: Seven Stylized Facts."*

### Section 5).

### Section 5

### Monetary and fiscal policy stances (Fact 3)
- Countries that resolved inflation implemented tighter macroeconomic policies post-shock:
  - Real short-term rate: resolved vs unresolved difference statistically significant at *** p<0.01.
  - Broad money growth rate: resolved vs unresolved difference statistically significant at ^ p<0.25.
  - Deviation from Taylor Rule (output gap-based): resolved vs unresolved difference statistically significant at *** p<0.01.
  - Deviation from Taylor Rule (unemployment gap-based): resolved vs unresolved difference statistically significant at *** p<0.01.
  - Cyclically adjusted primary balance: reported means/medians shown in Figure 10 (no explicit p-value stated in text).
- Interpretation: better pre-shock inflation anchoring helps but cannot substitute for tighter macroeconomic policies post-shock.

### Inflation anchoring (Fact 2 / illustrative result)
- Pre-shock inflation volatility:
  - Lower inflation volatility pre-shock in countries that resolved inflation; significance at ^ p<0.25 (Panel A, Figure 11).
- Other anchoring proxies:
  - Pre-shock inflation level and central bank independence show no statistically significant difference between resolved and unresolved in the baseline (Panels B and C), though pre-shock inflation becomes more significant in the full sample (Section 5).
- Note: central bank independence index (Romelli, 2022) ranges between 0 and 1; shown means/medians in Figure 11.

### Consistency of policy stance over time (Fact 4)
- Post-shock money growth dynamics (Figure 12):
  - Resolved episodes: money growth rates over T+1,...,T+5 are on average negative and generally stable.
  - Unresolved episodes: higher average money growth with substantial changes over time (high at first, decline in T+2 and T+3, then rebound).
- Variance measures (Figure 13):
  - Var(∆Broad money growth rate): difference between resolved and unresolved statistically significant at ** p<0.05.
  - Var(∆Real short-term rate): differences less significant (no explicit p-value in text).
  - Var(∆CAPB): differences less significant (no explicit p-value in text).
- Methodological note: focus on variance of the first difference to distinguish gradual tightening from inconsistent policy reversals.

### Exchange rates and trade balances (Fact 5)
- Nominal exchange rate stability:
  - Countries that resolved inflation were less likely to abandon their exchange rate peg; likelihood difference statistically significant at * p<0.1 (Panel A, Figure 14).
  - Less likely to experience a large nominal ER depreciation; difference statistically significant at ** p<0.05 (Panel B, Figure 14).
- Real effective exchange rate (Panel C):
  - Maintaining nominal ER stability induced a degree of real appreciation in countries that resolved inflation.
- Trade balance (Panel D):
  - Trade balances deteriorated for all countries after the terms-of-trade shock; no difference in trade balance deterioration between resolved and unresolved episodes despite real appreciation in resolved cases.
- Interpretation: abandoning the peg often follows rapid reserve depletion, which can be induced by overly loose monetary and fiscal policies documented under Fact 3.

### Wages and labor market outcomes (Fact 6)
- Sample and caveat:
  - Wage analysis uses full sample of episodes during 1973–2014 due to data limitations; internationally consistent wage data available for only 13 episodes (footnote 23).
- Nominal wages:
  - Countries that resolved inflation: nominal wage growth moderated after the inflation shock.
  - Countries that did not resolve inflation: accelerated nominal wage growth.
  - Difference statistically significant at ^ p<0.25 (Figure 15).
- Real wages:
  - Real wage growth slowed across both groups, insignificantly more so in countries that resolved inflation.
- Policy implication: be mindful of labor market pressures when formulating inflation responses; slower nominal wage growth may have limited effect on real wages due to its role in fighting inflation.

### Growth and unemployment outcomes (Fact 7)
- Conceptual framing:
  - Tight policies for disinflation may incur short-term costs (lower growth, higher unemployment), but unresolved inflation accumulates real costs over time.
- Five-year horizon (Panel A, Figure 16):
  - No statistically significant difference in real GDP growth between resolved and unresolved episodes over 5 years; CPI inflation rate differences are statistically significant at *** p<0.01.
  - Mean and median output declines are marginally larger for unresolved episodes over the medium term.
- Two-year horizon for smaller shocks (Panel B, Figure 16):
  - In a sub-sample with inflation shock size below 10 percentage points, countries that resolved inflation experienced a larger decline in GDP growth than those that did not; difference statistically significant at ^ p<0.25.
  - CPI inflation rate differences remain statistically significant at *** p<0.01 in this sub-sample.
- Unemployment:
  - No statistically significant difference in unemployment outcomes; sample for unemployment has 29 observations versus 46 for GDP growth (footnote 28).
- Interpretation: short-term Phillips-curve-like trade-off exists; successful disinflations tend to “pay off” in the medium term by avoiding accumulating costs of macroeconomic instability.

### Robustness checks (Section 5)
- Main robustness findings:
  - Full sample results align with baseline 1973–79 oil crises episodes, with two notable differences:
    - Pre-shock inflation volatility association with resolving inflation is stronger in the full sample (significant at 1 percent).
    - Pre-shock inflation rate becomes statistically significant at ^ p<0.25 in the full sample.
  - In the full sample, resolving inflation shocks is associated with:
    - Lower real GDP growth (significant at ^ p<0.25).
    - Looser fiscal policies (significant at 1 percent).
  - These full-sample associations may be influenced by episodes overlapping with the GFC; excluding GFC episodes (full non-GFC sample) makes the GDP-growth relationship insignificant and the fiscal balance relationship less significant (only at ^ p<0.25).
- Episode selection sensitivity:
  - Stricter inflation shock criteria (3 percent increase threshold; narrower 4–20 percent bounds) do not alter the selected set for the 1973–79 oil crises.
  - Reintroducing visually removed episodes yields nearly identical results; only inflation anchoring proxies become statistically insignificant when reintroducing those episodes.
- Identification concerns:
  - Potential sample compositional differences between resolved and unresolved episodes are discussed; Appendix V documents additional robustness checks.

*IMF WORKING PAPERS One Hundred Inflation Shocks: Seven Stylized Facts — Section 5*

### Section 3.2, which indicates that, although pre-shock inflation averages are similar between resolved

### Section 3.2 — Robustness Checks, Case Studies, and Policy Implications

### Robustness checks: excluding large inflation shocks
- Motivation: Resolved and unresolved episodes had similar pre-shock inflation averages, but unresolved episodes were hit by larger inflation shocks in year T. Larger shocks may be harder to resolve and confound associations between resolution and policy stances.
- Sub-sample rule: retain only episodes where inflation rose by less than 10 percentage points in year T (relative to T-1).
- Key data/sample impacts:
  - Difference between average inflation shocks in the restricted sub-sample is about 1 percentage point (as opposed to 5 percentage points in Table 1).
  - Excluding large shocks reduces sample size from 61 to 48 episodes.
  - Data availability further reduces the number of available country observations to between 9 to 47 (from 13 to 60) depending on the variables examined.
  - Median number of observations reduced to 33 (from 42).
- Statistical outcomes (1970s sample, smaller shocks):
  - Results remain nearly identical to baseline in Table IV.1 despite smaller sample sizes reducing statistical power.
  - Loss of statistical significance for:
    - inflation volatility (previously significant at the 25 percent level),
    - variance of the first difference of the real short-term rate (previously significant at the 25 percent level).
  - Higher variance of the first difference of money growth in countries that did not resolve inflation remains statistically significant at the 25 percent level.

### Robustness checks: exchange rate regime and policy space
- Concern: Fixed exchange rate regimes constrain policy space; maintaining a peg in the face of an inflation shock may mechanically adjust monetary aggregates rather than reflect active monetary policy decisions under floats.
- Sample: 1970s sample excluding countries with a fixed exchange rate regime in year T-1.
- Findings:
  - Most baseline results remain robust.
  - Notable change under floating regimes: countries with unresolved inflation have higher variance in the first differences of primary and cyclically-adjusted primary balances; this difference becomes statistically significant at the 25 percent level.
  - Interpretation: consistency in the fiscal stance matters more for fighting inflation under floating exchange rate regimes, consistent with impulse responses to discretionary fiscal policy in open economy New Keynesian models.
- Additional small-sample effects noted:
  - Money growth no longer has a statistically significant association with inflation resolution (only 7 observations for money growth data).
  - Central bank independence becomes statistically significant at the 25 percent level but with an unexpected sign on means (driven by outliers); medians do not show the same effect.

### Robustness checks: shorter post-shock horizon
- Two-year horizon checks:
  - Results are robust over a 2-year horizon (Table V.6).
  - Combining the 2-year horizon with the smaller-shock sample alters results on real GDP growth rate only (interpretation as discussed under Fact 7).

### Illustrative case studies (selected advanced economies, 1973–79 oil crises)
- Purpose: Complement quantitative evidence with qualitative, historical narratives on policy responses and outcomes.
- Overall observation: Countries’ policy responses varied—some tightened policies consistently; others loosened prematurely (“premature celebrations”). Case selection focused on intersection of resolved vs unresolved episodes and tighter vs looser post-peak policies.

- Spain 1973:
  - Inflation increased almost 20 percentage points between 1968 and 1978.
  - Real interest rates declined by about 10 percentage points.
  - Fiscal policy remained neutral for most of the episode, loosening slightly in the last two years.
  - Outcomes: inflation kept accelerating; real GDP growth halved; unemployment rate increased threefold.
  - Institutional context: period of weakening leadership and democratic transition (1973–1978).
  - Resolution: Moncloa Pacts in 1977 and substantial monetary tightening during the 1980s eventually brought inflation under control.

- Canada 1973:
  - Inflation spiked from 3 to 12 percent between 1972 and 1974.
  - Policy path: initial loosening (to protect growth and employment), then tightening of monetary and fiscal policy, followed by loosening again as inflation appeared to recede.
  - Inflation appeared to decrease from 12 to 8 percent in 1975-76, prompting policy loosening ahead of the 1979 election; subsequent reacceleration left the episode unresolved.
  - Outcomes: GDP growth zig-zagged; unemployment doubled from about 4 to 8 percent.

- France 1974:
  - Inflation spiked in 1974; monetary and fiscal policy tightened modestly over the next two years.
  - As inflation seemed to come under control, policy mix loosened again (real interest rates declined; primary fiscal balances shifted to deficits).
  - Disinflation halted and inflation rose again.
  - Outcomes: aside from a brief 1975 contraction, GDP growth remained stable but slightly downward; unemployment rose from about 2 percent to 5 percent.

- United Kingdom 1979:
  - Inflation spiked from 5 to 15 percent after the second oil shock.
  - Monetary policy tightened sharply: real interest rates increased from -5 percent to 7 percent between 1980 and 1984.
  - Fiscal policy largely unchanged (constant primary deficit over the ten-year period).
  - Outcome: monetary tightening succeeded in resolving the inflation shock; GDP growth reverted to pre-shock rates of 3 to 5 percent by the mid-1980s; unemployment rose from 5 to 12 percent in five years.

- Italy 1980:
  - Real interest rates rose from -5 percent to 6 percent (sharp, consistent monetary tightening).
  - Additional factors: end of price indexation practices; growing central bank independence.
  - Fiscal stance: primary balances remained unchanged and settled in negative territory with average primary deficits of around 5 percent per year.
  - Outcome: inflation brought down below pre-shock levels; GDP growth slowed in post-shock years; unemployment did not increase significantly.

- Japan 1980:
  - Inflation doubled from around 4 to 8 percent in 1980.
  - Both monetary and fiscal policies tightened immediately; policies had been restrictive since 1977.
  - Real interest rates remained in restrictive territory post-shock; fiscal policy continued to tighten.
  - Outcome: GDP growth declined to 3 percent in the immediate aftermath, then gradually recovered; unemployment remained stable.

### Conclusions and policy implications (from Section 7)
- Sample scope and findings:
  - Identified over 100 inflation shocks in 56 countries since the 1970s, including over 60 episodes linked to the 1973–79 oil crises.
  - Only in 60 percent of the episodes was inflation resolved within 5 years.
  - Even in successful cases, resolving inflation took on average over 3 years.
  - Success rates were lower and resolution times longer for episodes induced by terms-of-trade shocks during the 1973–79 oil crises.
- Key policy implications:
  - Policy consistency is key: maintain tight monetary and fiscal policies for some time; avoid loosening in response to emerging softer inflation readings (to prevent “premature celebrations”).
  - Although inflation shocks entail costs (higher unemployment and lower real wage growth), successful disinflations were not associated with larger output, employment, or real wage deteriorations over a 5-year horizon—suggesting benefits from policy credibility and macroeconomic stability.

*IMF Working Paper — Section 3.2, extracted from the provided PDF content*

### References

### References and Appendix I: Data sources and summary statistics

### Key references and thematic coverage
- Studies on wage-price dynamics and historical inflation episodes: Alvarez et al. (2022) "Wage-price spirals: What is the historical evidence?", Ball (1994) "What determines the sacrifice ratio?", Blinder (1982) "The anatomy of double-digit inflation in the 1970s", Gordon et al. (1977) "Can the inflation of the 1970s be explained?".
- Monetary policy theory and practice, communication, and tools: Barro & Gordon (1983), Bernanke (2020) "The new tools of monetary policy", Blinder et al. (2008) "Central Bank Communication and Monetary Policy", Clarida, Gali, & Gertler (1999, 2000), Orphanides (2003).
- Recent analyses of post-2020 inflation and inflation forecasting: Benigno & Eggertsson (2023), Bernanke & Blanchard (2023), Blanco, Ottonello, & Ranosova (2022), Koch & Noureldin (2023).
- Supply shocks, energy price shocks, and related policy responses: Ari et al. (2022), Arregui et al. (2022), Celasun et al. (2022), Fornaro & Wolf (2023).
- Fiscal policy, fiscal history, and interactions with inflation: Blanchard (2004), Mauro et al. (2015), Sims (2011), IMF (2023) "Fiscal Monitor: “Inflation and Disinflation: What Role for Fiscal Policy?”".
- Empirical methods and data issues relevant to macroeconomic measurement: Hamilton (2018) on the Hodrick-Prescott filter, Hodrick & Prescott (1997), Havranek & Rusnak (2013) meta-analysis of transmission lags, Ilzetzki, Reinhart, & Rogoff (2019) on exchange arrangements.
- Other relevant literature cited includes research on central bank independence (Romelli 2022), macroprudential policy (De Nicoló, Favara, & Ratnovski 2012), and the effects/costs of inflation and disinflation (Driffill et al. 1990; Tetlow 2022).

### Table I.1 — Summary Statistics, Data Coverage and Data Sources (as reported)
- Notes: BIS = "Bank for International Settlements"; OECD = "Organization for Economic Cooperation and Development" EO = "Employment Outlook"; MEI = "Main Economic Indicators"; PFHM = "Public Finances in Modern History Database"; WEO = "(April 2023) International Monetary Fund World Economic Outlook Database". The CBIE index ranges between 0 (no independence) and 1 (full independence). Source: Authors' estimates.

- Short-term interest rate (%) — OECD MEI: 6.45.15.92610
- Short-term interest rate (%) — Haver Analytics: 7.14.810.5249
- Short-term interest rate (%) — BIS: 6.15.05.2198
- Short-term interest rate (%) — WEO: 8.56.010.29113
- Short-term interest rate (%) — Haver Analytics: 4.43.54.3345

- Broad money growth rate (%) — WEO: 51.613.31,217.2174114

- CPI inflation rate (%) — WEO: 38.05.3806.2196145

- Output gap (% Potential GDP) — WEO: 0.20.05.610439

- Unemployment rate (%) — WEO: 8.57.26.011251

- Fiscal balance (% of GDP) — WEO: -2.1-2.25.4713

- Primary fiscal balance (% of GDP) — PFHM: -0.2-0.113.114452

- Cyclically-adjusted primary balance (Hamilton, % of GDP) — Hamilton filter: -1.3-0.173.714448

- Cyclically-adjusted primary balance (HP, % of GDP) — HP filter: 0.5-0.152.214449

- Primary expenditure (% of GDP) — PFHM: 27.925.318.114452

- Government debt-to-GDP ratio (% of GDP) — WEO: 54.744.744.919112

- Government debt-to-GDP ratio (% of GDP) — PFHM: 51.242.936.614454

- Fixed exchange rate regime dummy — Ilzetzki et al. (2019): 0.51.00.5188188

- NEER (index, 2020 = 100) — BIS: 131.7101.9115.92525

- REER (index, 2020 = 100) — BIS: 103.3100.318.62626

- Trade balance (% of GDP) — WEO: -6.1-2.919.4194132

- Current account balance (% of GDP) — WEO: -2.4-2.514.3196133

- Real GDP Growth Rate (% of GDP) — WEO: 3.83.96.1197148

- Nominal wage growth rate (%) — OECD EO: 4.43.44.7350

- Real wage growth rate (%) — OECD EO: 1.51.23.3350

- Central bank independence index - extended (CBIE) — Romelli (2022): 0.550.520.1715395

*Source: References list and Appendix I (Table I.1) from the chapter "References" and "Appendix I. Data sources and summary statistics" in the provided content.*

### Appendix II. Episode Selection

### Appendix II. Episode Selection

### Episodes Removed due to Country Characteristics
- Small population  
  Anguilla 2003, Anguilla 2006, Anguilla 2010, Antigua and Barbuda 1970, Antigua and Barbuda 1972, Antigua and Barbuda 1976, Antigua and Barbuda 1979, Antigua and Barbuda 1987, Antigua and Barbuda 1990, Antigua and Barbuda 1994, Antigua and Barbuda 1998, Antigua and Barbuda 2001, Antigua and Barbuda 2008, Antigua and Barbuda 2010, Aruba 2008, Aruba 2010, Aruba 2014, Bahamas, The 1972, Bahamas, The 1974, Bahamas, The 1978, Bahamas, The 1991, Bahamas, The 2003, Bahamas, The 2008, Bahrain 1971, Bahrain 1973, Bahrain 1976, Bahrain 1981, Bahrain 1993, Bahrain 1997, Bahrain 2003, Barbados 1970, Barbados 1973, Barbados 1977, Barbados 1979, Barbados 1987, Barbados 1991, Barbados 1995, Barbados 1997, Barbados 1999, Barbados 2005, Barbados 2008, Barbados 2010, Belize 1972, Belize 1977, Belize 1980, Belize 1988, Belize 1990, Belize 1996, Belize 2008, Belize 2010, Bhutan 1970, Bhutan 1973, Bhutan 1979, Bhutan 1982, Bhutan 2011, Botswana 1973, Botswana 1979, Botswana 1981, Brunei Darussalam 1993, Brunei Darussalam 1995, Brunei Darussalam 2003, Cabo Verde 1970, Cabo Verde 1973, Cabo Verde 1975, Cabo Verde 1977, Cabo Verde 1980, Cabo Verde 1986, Cabo Verde 1990, Cabo Verde 1992, Cabo Verde 1995, Cabo Verde 1997, Cabo Verde 2001, Cabo Verde 2005, Cabo Verde 2008, Cabo Verde 2011, Comoros 1970, Comoros 1973, Comoros 1980, Comoros 1982, Comoros 1985, Comoros 1987, Comoros 1989, Comoros 1991, Comoros 1993, Comoros 2000, Comoros 2012, Djibouti 1994, Djibouti 2007, Djibouti 2011, Dominica 1970, Dominica 1973, Dominica 1979, Dominica 1989, Dominica 1991, Dominica 2008, Dominica 2010, Equatorial Guinea 1971, Equatorial Guinea 1973, Equatorial Guinea 1976, Equatorial Guinea 1979, Equatorial Guinea 1981, Equatorial Guinea 1985, Equatorial Guinea 1987, Equatorial Guinea 1993, Equatorial Guinea 1998, Equatorial Guinea 2000, Eswatini 1973, Eswatini 1977, Eswatini 1979, Eswatini 1985, Eswatini 1988, Eswatini 1990, Eswatini 1993, Eswatini 2000, Eswatini 2002, Fiji 1971, Fiji 1980, Fiji 1987, Fiji 1990, Fiji 1998, Fiji 2001, Fiji 2003, Fiji 2007, Fiji 2011, Gabon 1973, Gabon 1980, Gabon 1982, Gabon 1989, Gambia, The 1971, Gambia, The 1974, Gambia, The 1981, Gambia, The 1983, Gambia, The 1986, Gambia, The 1990, Grenada 1979, Grenada 1988, Grenada 1994, Grenada 2008, Grenada 2010, Guinea-Bissau 1970, Guinea-Bissau 1974, Guinea-Bissau 1979, Guinea-Bissau 1983, Guinea-Bissau 1987, Guinea-Bissau 1989, Guinea-Bissau 1991, Guyana 1970, Guyana 1972, Guyana 1978, Guyana 1981, Guyana 1984, Guyana 1987, Guyana 1991, Guyana 1994, Guyana 1999, Guyana 2002, Guyana 2005, Guyana 2007, Kiribati 1980, Kiribati 1986, Kiribati 1988, Kiribati 1991, Kiribati 1997, Kiribati 2001, Kiribati 2007, Kiribati 2011, Kiribati 2014, Macao SAR 2004, Macao SAR 2008, Macao SAR 2011, Maldives 1970, Maldives 1973, Maldives 1976, Maldives 1978, Maldives 1983, Maldives 1986, Maldives 1990, Maldives 1992, Maldives 1995, Maldives 1999, Maldives 2004, Maldives 2007, Maldives 2011, Malta 1977, Malta 1979, Malta 1986, Malta 1990, Malta 1993, Marshall Islands 2008, Marshall Islands 2011, Mauritius 1972, Mauritius 1980, Micronesia 2005, Micronesia 2008, Micronesia 2012, Montenegro, Rep. of 2008, Montenegro, Rep. of 2011, Montserrat 2008, Nauru 2006, Nauru 2009, Nauru 2012, Palau 2005, Palau 2008, Palau 2012, Samoa 1972, Samoa 1977, Samoa 1979, Samoa 1987, Samoa 1989, Samoa 1992, Samoa 1994, Samoa 1996, Samoa 2001, Samoa 2004, Samoa 2009, Samoa 2011, São Tomé and Príncipe 1970, São Tomé and Príncipe 1973, São Tomé and Príncipe 1978, São Tomé and Príncipe 1981, São Tomé and Príncipe 1983, São Tomé and Príncipe 1986, São Tomé and Príncipe 1991, São Tomé and Príncipe 1994, São Tomé and Príncipe 1996, São Tomé and Príncipe 2004, São Tomé and Príncipe 2008, Seychelles 1972, Seychelles 1974, Seychelles 1983, Seychelles 1989, Seychelles 1994, Seychelles 1997, Seychelles 1999, Seychelles 2003, Seychelles 2007, Seychelles 2011, Solomon Islands 1970, Solomon Islands 1974, Solomon Islands 1977, Solomon Islands 1981, Solomon Islands 1984, Solomon Islands 1986, Solomon Islands 1988, Solomon Islands 1991, Solomon Islands 1994, Solomon Islands 1996, Solomon Islands 1998, Solomon Islands 2006, Solomon Islands 2008, Solomon Islands 2011, St. Kitts and Nevis 1978, St. Kitts and Nevis 1980, St. Kitts and Nevis 1989, St. Kitts and Nevis 1997, St. Kitts and Nevis 2006, St. Kitts and Nevis 2011, St. Lucia 1970, St. Lucia 1973, St. Lucia 1978, St. Lucia 1980, St. Lucia 1987, St. Lucia 1989, St. Lucia 1995, St. Lucia 1998, St. Lucia 2000, St. Lucia 2005, St. Lucia 2008, St. Lucia 2010, St. Lucia 2014, St. Vincent and the Grenadines 1976, St. Vincent and the Grenadines 1979, St. Vincent and the Grenadines 1989, St. Vincent and the Grenadines 1996, St. Vincent and the Grenadines 2004, St. Vincent and the Grenadines 2007, St. Vincent and the Grenadines 2011, Suriname 1972, Suriname 1979, Suriname 1985, Suriname 1990, Suriname 1992, Suriname 1997, Suriname 2001, Suriname 2003, Suriname 2005, Suriname 2008, Suriname 2010, Timor-Leste 2003, Timor-Leste 2006, Tonga 1976, Tonga 1981, Tonga 1986, Tonga 1988, Tonga 1991, Tonga 1994, Tonga 1996, Tonga 1998, Tonga 2001, Tonga 2008, Tonga 2011, Tuvalu 2002, Tuvalu 2008, Tuvalu 2011, Vanuatu 1980, Vanuatu 1984, Vanuatu 1986
- Conflict  
  Afghanistan 2003, Afghanistan 2008, Afghanistan 2010, Bosnia and Herzegovina 1997, Bosnia and Herzegovina 1999, Burundi 1987, Burundi 1989, Burundi 1993, Burundi 2000, Burundi 2003, Burundi 2005, Burundi 2007, Cambodia 1988, Cambodia 1993, Central African Republic 2008, Central African Republic 2012, Central African Republic 2014, Chad 1973, Chad 1977, Chad 1984, Chad 1987, Chad 1990, Congo, Democratic Republic of the 1994, Congo, Democratic Republic of the 1996, Croatia 1998, Cyprus 1973, Cyprus 1977, Cyprus 1979, Egypt 1973, Egypt 2007, Egypt 2009, Egypt 2014, Eritrea 1994, Eritrea 1998, Eritrea 2000, Eritrea 2002, Ethiopia 1976, Ethiopia 1982, Ethiopia 1985, Ethiopia 1995, Ethiopia 1998, Ethiopia 2002, Georgia 2005, Georgia 2010, India 1970, India 1973, Iran 1977, Iran 1980, Iran 1986, Iran 1991, Iraq 2006, Israel 1970, Israel 1973, Israel 1977, Jordan 1972, Lebanon 1972, Lebanon 1974, Lebanon 1979, Lebanon 1984, Lebanon 1992, Lebanon 1995, Lebanon 2002, Lebanon 2006, Lebanon 2008, Lebanon 2010, Liberia 2000, Liberia 2005, Liberia 2008, Mali 2012, Mozambique 1981, Mozambique 1986, Mozambique 1992, Mozambique 1994, Pakistan 1970, Pakistan 1973, Rwanda 1987, Rwanda 1990, Rwanda 1993, Serbia 1998, Serbia 2004, Sierra Leone 1989, Sierra Leone 1998, Sierra Leone 2001, Sierra Leone 2003, Sierra Leone 2007, Sri Lanka 1978, Sri Lanka 1980, Sri Lanka 1983, Sri Lanka 1986, Sri Lanka 1988, Sri Lanka 1990, Sri Lanka 1996, Sri Lanka 2001, Sri Lanka 2007, Sri Lanka 2010, Sudan 1972, Sudan 1974, Sudan 1977, Sudan 1982, Sudan 1988, Sudan 1991, Sudan 1994, Sudan 1996, Sudan 2002, Sudan 2004, Sudan 2007, Syria 1970, Syria 1973, Syria 1980, Syria 1984, Syria 1992, Syria 2008, Tanzania 1977, Tanzania 1979, Tanzania 1982, Tanzania 1984, Uganda 1979, Uganda 1983, Uganda 1985, Ukraine 2014, West Bank and Gaza 2002, West Bank and Gaza 2005, West Bank and Gaza 2008, Yemen 2000, Yemen 2008, Yemen 2010
- Commodity exporter  
  Algeria 1973, Algeria 1975, Algeria 1977, Algeria 1981, Algeria 1985, Algeria 1989, Algeria 1991, Algeria 1994, Algeria 2001, Algeria 2003, Algeria 2012, Angola 1972, Angola 1974, Angola 1976, Angola 1979, Angola 1991, Angola 1995, Angola 1999, Azerbaijan 2000, Azerbaijan 2004, Azerbaijan 2007, Azerbaijan 2010, Bahrain 2012, Botswana 1989, Botswana 1992, Botswana 2006, Botswana 2008, Burundi 2011, Chad 2005, Chad 2008, Chad 2011, Congo, Democratic Republic of the 1986, Congo, Democratic Republic of the 1989, Congo, Democratic Republic of the 1991, Congo, Democratic Republic of the 1999, Congo, Democratic Republic of the 2005, Congo, Democratic Republic of the 2007, Congo, Democratic Republic of the 2009, Congo, Republic of 1970, Congo, Republic of 1974, Congo, Republic of 1977, Congo, Republic of 1982, Congo, Republic of 1989, Congo, Republic of 1994, Congo, Republic of 1997, Congo, Republic of 2002, Congo, Republic of 2006, Congo, Republic of 2008, Congo, Republic of 2012, Egypt 1980, Egypt 1982, Ethiopia 2005, Ethiopia 2007, Gabon 1993, Gabon 1997, Gabon 2000, Gabon 2008, Gabon 2014, Guinea 1970, Guinea 1973, Guinea 1978, Guinea 1986, Guinea-Bissau 1995, Indonesia 1979, Iran 1971, Iran 1994, Iran 2002, Iran 2007, Iran 2011, Iraq 2010, Kazakhstan 2000, Kazakhstan 2007, Kuwait 1977, Kuwait 1990, Kuwait 1999, Kuwait 2005, Kuwait 2007, Libya 1971, Libya 1978, Libya 1991, Libya 1994, Libya 2003, Libya 2007, Libya 2011, Malaysia 1973, Malaysia 1977, Malaysia 1980, Malaysia 1989, Mauritania 2010, Mongolia 2007, Mongolia 2010, Mongolia 2012, Nigeria 2001, Nigeria 2005, Nigeria 2008, Norway 2008, Oman 1979, Oman 1985, Oman 1990, Oman 2007, Papua New Guinea 1974, Papua New Guinea 1980, Papua New Guinea 1983, Papua New Guinea 1988, Papua New Guinea 1990, Papua New Guinea 1995, Papua New Guinea 1998, Papua New Guinea 2002, Papua New Guinea 2008, Qatar 1987, Qatar 1994, Qatar 1996, Qatar 2003, Qatar 2010, Saudi Arabia 1971, Saudi Arabia 1973, Saudi Arabia 1980, Saudi Arabia 1988, Saudi Arabia 1991, Saudi Arabia 1993, Saudi Arabia 1995, Saudi Arabia 2007, Sudan 2011, Togo 1974, Togo 1977, Trinidad and Tobago 1972, Trinidad and Tobago 1979, Trinidad and Tobago 1983, Trinidad and Tobago 1987, Trinidad and Tobago 1989, Trinidad and Tobago 1992, Trinidad and Tobago 2001, Trinidad and Tobago 2005, Trinidad and Tobago 2008, Trinidad and Tobago 2010, Trinidad and Tobago 2012, Turkmenistan 1999, Turkmenistan 2001, Turkmenistan 2005, Turkmenistan 2008, Turkmenistan 2010, Uganda 1992, Uganda 2003, United Arab Emirates 1991, United Arab Emirates 2006, Uzbekistan 2001, Uzbekistan 2005, Venezuela 1974, Venezuela 1979, Venezuela 1984, Venezuela 1987, Venezuela 1989, Venezuela 1993, Venezuela 1996, Venezuela 2002, Venezuela 2007, Venezuela 2010, Venezuela 2013, Yemen 1998, Zambia 1971
- Non-market and transition economy  
  Albania 1991, Albania 1996, Armenia 1994, Azerbaijan 1994, Belarus 1992, Belarus 1997, Belarus 2008, Belarus 2011, Bulgaria 1982, Bulgaria 1989, Bulgaria 1994, Bulgaria 1996, China 1985, China 1988, China 1992, China 2003, China 2007, China 2010, Czech Republic 1998, Georgia 1999, Hungary 1973, Hungary 1975, Hungary 1979, Hungary 1982, Hungary 1984, Hungary 1987, Hungary 1990, Hungary 1995, Kyrgyz Republic 1993, Kyrgyz Republic 1999, Lao P.D.R. 1972, Lao P.D.R. 1977, Lao P.D.R. 1980, Lao P.D.R. 1982, Lao P.D.R. 1985, Lao P.D.R. 1988, Lao P.D.R. 1991, Lao P.D.R. 1994, Lao P.D.R. 1998, Lao P.D.R. 2002, Lao P.D.R. 2008, Lao P.D.R. 2010, Lao P.D.R. 2013, Moldova 1999, Mongolia 1992, Mongolia 1996, Poland 1973, Poland 1981, Poland 1984, Poland 1986, Romania 1982, Romania 1990, Romania 1996, Russia 1991, Russia 1998, Slovak Republic 1999, Tajikistan 1995, Ukraine 1993, Ukraine 1999, Uzbekistan 1994, Vietnam 1980, Vietnam 1984, Vietnam 1988, Vietnam 1991, Vietnam 1995, Vietnam 1998, Vietnam 2002, Vietnam 2004, Vietnam 2008, Vietnam 2010
- Below GDP cutoff  
  Albania 2001, Armenia 2001, Armenia 2003, Armenia 2006, Armenia 2008, Armenia 2010, Armenia 2013, Bangladesh 1981, Bangladesh 1995, Bangladesh 1998, Bangladesh 2004, Bangladesh 2008, Benin 1973, Benin 1979, Benin 1982, Benin 1984, Benin 1988, Benin 1992, Benin 1994, Benin 2000, Benin 2005, Benin 2008, Benin 2012, Bolivia 1973, Bolivia 1979, Bolivia 1982, Bolivia1991, Bolivia 1995, Bolivia 1998, Bolivia 2000, Bolivia 2003, Bolivia 2007, Bolivia 2011, Bosnia and Herzegovina 2005, Bosnia and Herzegovina 2008, Bosnia and Herzegovina 2010, Burkina Faso 1973, Burkina Faso 1975, Burkina Faso 1979, Burkina Faso 1982, Burkina Faso 1985, Burkina Faso 1988, Burkina Faso 1991, Burkina Faso 1993, Burkina Faso 2001, Burkina Faso 2005, Burkina Faso 2008, Burkina Faso 2011, Burundi 1973, Burundi 1979, Burundi 1981, Burundi 1983, Cambodia 1997, Cambodia 2004, Cambodia 2008, Cambodia 2010, Cameroon 1973, Cameroon 1979, Cameroon 1982, Cameroon 1992, Cameroon 1994, Cameroon 2001, Cameroon 2006, Cameroon 2008, Central African Republic 1985, Central African Republic 1988, Central African Republic 1992, Central African Republic 1994, Central African Republic 2000, Central African Republic 2003, Central African Republic 2005, Chad 1994, Chad 1996, Chad 2000, China 2004, Congo, Democratic Republic of the 1973, Congo, Democratic Republic of the 1974, Congo, Democratic Republic of the 1979, Congo, Democratic Republic of the 1982, Côte d'Ivoire 1973, Côte d'Ivoire 1986, Côte d'Ivoire 1991, Côte d'Ivoire 1994, Côte d'Ivoire 1997, Côte d'Ivoire 2005, Côte d'Ivoire 2011, Dominican Republic 1973, Dominican Republic 1979, Dominican Republic 1984, Dominican Republic 1987, Dominican Republic 1990, Dominican Republic 1994, Dominican Republic 1997, Dominican Republic 2003, Dominican Republic 2006, Dominican Republic 2008, Ecuador 1973, Ecuador 1992, Ecuador 1997, Ecuador 2008, Egypt 1986, Egypt 1989, Egypt 1992, Egypt 2004, El Salvador 1980, El Salvador 1985, El Salvador 1990, El Salvador 1993, El Salvador 2004, El Salvador 2008, El Salvador 2011, Eritrea 2008, Eritrea 2014, Eswatini 2007, Eswatini 2012, Ethiopia 1973, Ethiopia 1988, Ethiopia 1991, Ethiopia 2011, Gambia, The 1995, Gambia, The 1999, Gambia, The 2001, Gambia, The 2007, Georgia 2014, Ghana 1973, Ghana 1975, Ghana 1981, Ghana 1983, Ghana 1986, Ghana 1990, Ghana 1993, Ghana 1995, Ghana 2000, Ghana 2003, Ghana 2005, Ghana 2008, Ghana 2013, Guatemala 1973, Guatemala 1985, Guatemala 1989, Guatemala 1993, Guatemala 1996, Guatemala 2008, Guatemala 2010, Guinea 1998, Guinea 2000, Guinea 2003, Guinea 2010, Guinea-Bissau 2000, Guinea-Bissau 2004, Guinea-Bissau 2007, Guinea-Bissau 2010, Haiti 1973, Haiti 1979, Haiti 1983, Haiti 1985, Haiti 1988, Haiti 1992, Haiti 2000, Haiti 2003, Haiti 2008, Haiti 2011, Honduras 1974, Honduras 1979, Honduras 1988, Honduras 1993, Honduras 2008, Honduras 2011, India 1979, India 1983, India 1986, India 1990, India 1994, India 1998, India 2006, India 2008, Indonesia 1973, Indonesia 1983, Indonesia 1987, Indonesia 1993, Indonesia 1998, Indonesia 2001, Indonesia 2005, Indonesia 2008, Indonesia 2013, Jamaica 2003, Jamaica 2008, Jamaica 2010, Jamaica 2013, Jordan 1996, Jordan 2006, Jordan 2008, Jordan 2010, Kenya 1973, Kenya 1979, Kenya 1980, Kenya 1982, Kenya 1985, Kenya 1987, Kenya 1990, Kenya 1996, Kenya 2000, Kenya 2003, Kenya 2008, Kenya 2011, Korea 1973, Korea 1974, Korea 1979, Kosovo 2006, Kosovo 2010, Kyrgyz Republic 2007, Kyrgyz Republic 2011, Kyrgyz Republic 2013, Lesotho 1979, Lesotho 1983, Lesotho 1985, Lesotho 1989, Lesotho 1991, Lesotho 1995, Lesotho 2002, Lesotho 2006, Lesotho 2008, Liberia 2014, Madagascar 1973, Madagascar 1974, Madagascar 1979, Madagascar 1986, Madagascar 1988, Madagascar 1990, Madagascar 1992, Madagascar 1994, Madagascar 2000, Madagascar 2002, Madagascar 2004, Malawi 1973, Malawi 1974, Malawi 1979, Malawi 1983, Malawi 1986, Malawi 1992, Malawi 1994, Malawi 1998, Malawi 2005, Malawi 2012, Mali 1973, Mali 1979, Mali 1980, Mali 1983, Mali 1988, Mali 1993, Mali 1998, Mali 2001, Mali 2005, Mali 2008, Mauritania 1992, Mauritania 1995, Mauritania 1998, Mauritania 2004, Mauritius 1985, Mauritius 1988, Mauritius 1991, Moldova 2003, Moldova 2010, Mongolia 2000, Mongolia 2003, Mongolia 2014, Morocco 1973, Morocco 1974, Morocco 1979, Morocco 1981, Morocco 1984, Morocco 1990, Morocco 2002, Morocco 2006, Mozambique 2000, Mozambique 2002, Mozambique 2006, Mozambique 2008, Mozambique 2010, Myanmar 2002, Myanmar 2006, Myanmar 2010, Myanmar 2013, Namibia 1992, Namibia 1994, Namibia 1999, Namibia 2002, Namibia 2006, Namibia 2008, Nepal 1973, Nepal 1979, Nepal 1980, Nepal 1983, Nepal 1986, Nepal 1992, Nepal 1999, Nepal 2006, Nepal 2009, Nicaragua 1995, Nicaragua 1998, Nicaragua 2004, Nicaragua 2008, Nicaragua 2011, Niger 1973, Niger 1975, Niger 1981, Niger 1984, Niger 1988, Niger 1991, Niger 1993, Niger 2000, Niger 2005, Niger 2008, North Macedonia 2000, North Macedonia 2006, North Macedonia 2008, Pakistan 1979, Pakistan 1988, Pakistan 1991, Pakistan 1993, Pakistan 1997, Pakistan 2005, Pakistan 2008, Pakistan 2011, Papua New Guinea 1973, Paraguay 2002, Paraguay 2005, Paraguay 2008, Paraguay 2010, Peru 2003, Peru 2008, Philippines 1973, Philippines 1979, Philippines 1984, Philippines 1987, Philippines 1991, Philippines 1994, Philippines 1998, Philippines 2004, Philippines 2008, Rwanda 1973, Rwanda 1979, Rwanda 1982, Rwanda 2000, Rwanda 2003, Rwanda 2008, Rwanda 2011, Senegal 1973, Senegal 1979, Senegal 1982, Senegal 1988, Senegal 1994, Senegal 2001, Senegal 2007, Senegal 2010, Sierra Leone 1973, Sierra Leone 1974, Sierra Leone 1979, Sierra Leone 1981, Sierra Leone 1985, South Sudan 2014, Sri Lanka 1973, Syria 2001, Syria 2003, Syria 2005, Tajikistan 2000, Tajikistan 2003, Tajikistan 2006, Tajikistan 2011, Tanzania 1973, Tanzania 1990, Tanzania 1993, Tanzania 2006, Tanzania 2008, Tanzania 2011, Thailand 1973, Thailand 1979, Timor-Leste 2010, Togo 1973, Togo 1979, Togo 1986, Togo 1990, Togo 1994, Togo 2001, Togo 2005, Togo 2008, Togo 2011, Tunisia 1973, Tunisia 1975, Tunisia 1979, Tunisia 1982, Tunisia 1987, Tunisia 2006, Uganda 2005, Uganda 2008, Uganda 2011, Ukraine 2003, Ukraine 2007, Uzbekistan 1997, Yemen 1992, Zambia 1975, Zambia 1980, Zambia 1983, Zambia 1985, Zambia 1988, Zambia 1992, Zambia 1996, Zimbabwe 1999, Zimbabwe 2002, Zimbabwe 2006, Zimbabwe 2008

Note: 'Small population' refers to episodes in countries that have less than 1 million population and that are not advanced economies. 'Conflict' refers to episodes that coincide with active military conflicts and other armed hostilities. 'Commodity exporter' refers to episodes in countries where fossil fuel and other natural resources amount to more than 20 percent of net export revenues and/or the rents from these amount to more than 20 percent of GDP. 'Non-market and transition economy' refers to episodes in planned economies as well as post-communist countries undergoing a period of transition between 1990-99. Finally, 'below GDP cutoff' refers to episodes where real GDP per capita PPP averaged less than 20 percent of the US in the 5 years prior to the episode start date T.

### Episodes Removed due to Inflation Trajectory Characteristics
- Below lower inflation threshold (inflation below 3 percent at T; reported inflation rates in parentheses)  
  Belgium 2010 (2.3), Cyprus 2010 (2.6), Czech Republic 2004 (2.8), Estonia 2010 (2.7), Hong Kong SAR 2001 (-1.6), Hong Kong SAR 2004 (-0.4), Ireland 2011 (1.2), Japan 2014 (2.8), Lithuania 2004 (1.2), Luxembourg 2010 (2.8), New Zealand 2000 (2.6), North Macedonia 2010 (1.5), Panama 2005 (2.9), Portugal 2010 (1.4), Singapore 2010 (2.8), Spain 2010 (2.0), United Arab Emirates 2010 (0.9), Zimbabwe 1984 (-1.9), Zimbabwe 1990 (1.3), Zimbabwe 1992 (-3.0)
- Above higher inflation threshold (average inflation in T-1 and T-2 exceeds 25 percent; reported inflation rates in parentheses)  
  Brazil 1979 (41.2), Brazil 1983 (101.2), Brazil 1987 (186.6), Brazil 1992 (1690.3), Chile 1973 (48.9), Colombia 1979 (25.9), Colombia 1990 (27), Ecuador 1987 (25.5), Iceland 1979 (37.3), Iceland 1980 (44.9), Iceland 1983 (50.9), Iceland 1985 (56.7), Israel 1982 (216.7), Jamaica 1994 (49.5), Jamaica 1996 (27.7), Mexico 1982 (27.3), Mexico 1986 (61.6), Mexico 1990 (66.7), Nicaragua 1982 (29.5), Nicaragua 1989 (8942.3), Peru 1981 (62.9), Peru 1983 (69.9), Peru 1985 (110.7), Peru 1987 (120.7), Türkiye 1979 (38.1), Türkiye 1984 (31.2), Türkiye 1987 (39.6), Türkiye 1991 (61.8), Türkiye 1994 (68.1), Türkiye 1997 (84.9), Uruguay 1973 (50.2), Uruguay 1975 (87.1), Uruguay 1979 (54.2), Uruguay 1983 (26.5), Uruguay 1989 (62.9)
- Elevated pre-episode inflation (start date after 1979 where inflation in T-1, T-2 or T-3 is above inflation in period T)  
  Australia 1985, Bulgaria 2000, Bulgaria 2004, Chile 1983, Chile 1989, Colombia 1985, Costa Rica 1994, Dominican Republic 2010, Greece 2010, Hong Kong SAR 1987, Hungary 2004, Iceland 1988, Israel 1989, Israel 2002, Jamaica 1983, Jamaica 2000, Latvia 2011, Mauritius 2004, New Zealand 1987, Norway 1980, Paraguay 1998, Paraguay 2000, Paraguay 2014, Poland 2000, Poland 2004, Portugal 1989, Romania 2008, Serbia 2008, Serbia 2011, Slovak Republic 2003, Slovak Republic 2008, Slovak Republic 2011, Slovenia 2000, Thailand 2008, Thailand 2010, Türkiye 2010, Türkiye 2012

Note: ‘Below lower inflation threshold’ refers to episodes where inflation is below 3 percent at T. ‘Above lower inflation threshold’ refers to episodes where average inflation in T-1 and T-2 exceeds 25 percent. In both categories, the corresponding inflation rate values are reported in parentheses. ‘Elevated pre-episode inflation’ refers to episodes with a start date after 1979 where inflation in T-1, T-2 or T-3 is above the inflation rate in period T.

### Episodes Removed Following Visual Inspection
- Overlap with another episode (episodes dropped in favor of another, overlapping episode selected by the algorithm)  
  Argentina 2005, Australia 2000, Austria 1984, Brazil 2000, Brazil 2008, Costa Rica 1979, Costa Rica 1985, Costa Rica 1990, Costa Rica 2008, Ecuador 1979, Ecuador 1980, France 1991, Greece 1990, Hong Kong SAR 1981, Hong Kong SAR 2011, Iceland 2008, Korea 1990, Latvia 2007, Lithuania 2011, Mauritius 2011, Mexico 1979, New Zealand 1973, New Zealand 1985, Norway 1973, Panama 1974, Panama 2011, Paraguay 1983, Paraguay 1993, Portugal 1983, Singapore 1979, South Africa 2007, Sweden 1973, Sweden 1993, Switzerland 1979, Türkiye 1975
- Inflation volatility caught as episode (cases where the selection algorithm misidentifies fluctuations in the inflation rate as an episode)  
  Chile 1985, Côte d'Ivoire 1979, Guatemala 1979, Hungary 2007, Jordan 1979, Mauritius 1993, Zimbabwe 1986, Zimbabwe 1995
- Elevated pre-episode inflation (cases where the algorithm incorrectly captures the trajectory of a 1973 inflation shock as a new episode)  
  Denmark 1979, Jamaica 1979, Spain 1979

Note: In total, 46 episodes are dropped following visual inspection, of which 35 are due to overlaps with another episode.

*Sources: IMF World Economic Outlook, World Bank World Development Indicators, and IMF staff calculations.*

### Appendix III. Selected Episodes

### Appendix III. Selected Episodes

### Selected Episodes by Sample and Inflation Outcomes
- Episodes associated with the 1973 and 1979 oil crises:
  - Inflation resolved: Austria 1980, Belgium 1974, Belgium 1980, Canada 1981, Costa Rica 1973, Finland 1973, Finland 1980, France 1980, Hong Kong SAR 1973, Ireland 1973, Italy 1980, Japan 1973, Japan 1980, Luxembourg 1974, Malta 1981, Norway 1974, Panama 1973, Panama 1979, Portugal 1974, Singapore 1980, South Africa 1979, Sweden 1974, Sweden 1980, Switzerland 1973, Switzerland 1981, Taiwan Province of China 1979, United Kingdom 1973, United Kingdom 1979, United States 1979
  - Inflation unresolved: Australia 1973, Brazil 1974, Canada 1973, Colombia 1973, Costa Rica 1980, Denmark 1973, Spain 1973, France 1974, Greece 1973, Greece 1979, Hong Kong SAR 1979, Ireland 1979, Iceland 1973, Italy 1973, Jamaica 1973, Mexico 1973, Mexico 1980, Malta 1973, Malaysia 1973, Nicaragua 1979, New Zealand 1974, New Zealand 1980, Peru 1973, Portugal 1981, Paraguay 1973, Paraguay 1979, Singapore 1973, El Salvador 1973, Türkiye 1973, Taiwan Province of China 1973, United States 1973, South Africa 1973
- Other episodes:
  - Inflation resolved: Belgium 2008, Bulgaria 2008, Brazil 2003, Chile 2008, Chile 2014, Colombia 1988, Costa Rica 2004, Croatia 2008, Cyprus 2000, Cyprus 2008, Czech Republic 2008, Estonia 2007, Finland 2008, France 1989, Greece 1986, Iceland 2006, Ireland 2000, Israel 2008, Korea 1998, Korea 2008, Latvia 2004, Lithuania 2007, Luxembourg 1989, Malaysia 1998, Malaysia 2008, Malta 2008, Netherlands 2001, New Zealand 1995, Panama 2008, Paraguay 1990, Russia 2008, Singapore 2008, Sweden 1990, Switzerland 1990, Thailand 1998
  - Inflation unresolved: Argentina 2002, Australia 1995, Costa Rica 1987, Ecuador 1983, Hong Kong SAR 2008, Jamaica 1990, Jordan 1988, Korea 1988, Luxembourg 2000, Mexico 1995, Mauritius 2006, Taiwan Province of China 1989, Uruguay 2002, South Africa 1985, South Africa 2002

### Classification note
- Episodes associated with the 1973 and 1979 oil crises include episodes with start dates between 1973–82 (inclusive).
- The inflation shock is categorized as having been “resolved” if inflation at T+5 is within 1 percentage point of its pre-shock value (i.e., Π_{T+5} ≤ Π_{T−1} + 1%).  
- Sources: IMF World Economic Outlook, and IMF staff calculations.

### Figures
- The source contains Panels A–D presenting selected episodes by sample and inflation outcomes:
  - Panel A. Episodes Associated with the 1973 and 1979 Oil Crises; Inflation Resolved
  - Panel B. Episodes Associated with the 1973 and 1979 Oil Crises; Inflation Unresolved
  - Panel C. Other Selected Episodes; Inflation Resolved
  - Panel D. Other Selected Episodes; Inflation Unresolved

---

### Appendix IV. Baseline Results (1973–79 Oil Crises)
Notes: CAPB refers to cyclically adjusted primary balances. For robustness, two alternative versions of output gaps are applied using the Hodrick Prescott (HP) (1997) and Hamilton (2018) methodologies. The baseline results in Section 4 correspond to CAPB (Hamilton). The real short-term rate is defined as the nominal short-term rate minus the contemporaneous annual inflation rate. Inflation volatility is the variance of inflation between [T−5,T−1] normalized by average inflation over the same period. The likelihood of breaking peg (post-shock) is the share of episodes which had an exchange rate peg before T but not after the start year of the episode. Deviation from Taylor rule (output gap-based) corresponds to a Taylor rule with an inflation response coefficient of 1.5 and output gap response coefficient of 1. In the unemployment-based version, the response to the unemployment rate is calculated by using Okun’s Law which results in a corresponding coefficient of -2. P-values pertain to the difference between means of resolved and unresolved episodes. The CBIE index ranges between 0 (no independence) and 1 (full independence). R-squared values pertain to the regression specification provided in equation (1), Section 3.3.

Key baseline statistics (median, mean, P-value, Obs., R-squared):
- Monetary policy
  - Real short-term rate: 1.17, -3.66, 0.77, -4.56, 0.00, 25, 0.47
  - Nominal short-term rate: 3.32, 3.49, 3.08, 3.67, 0.40, 25, 0.03
  - Broad money growth rate: -2.98, 2.76, -3.92, 3.31, 0.12, 17, 0.12
  - Deviation from Taylor Rule (output gap-based): 1.49, -8.99, 0.93, -7.52, 0.00, 20, 0.48
  - Deviation from Taylor Rule (unemployment gap-based): 2.57, -2.28, 2.19, -5.87, 0.00, 23, 0.36
  - Inflation volatility (pre-shock): 0.81, 1.01, 1.23, 1.72, 0.23, 60, 0.02
  - Inflation rate (pre-shock): 6.29, 6.16, 7.24, 7.51, 0.81, 60, 0.00
  - Central bank independence index (pre-shock): 0.37, 0.46, 0.41, 0.45, 0.33, 47, 0.02
- Fiscal policy
  - Primary fiscal balance: 0.13, -0.11, -0.24, -0.07, 0.77, 46, 0.00
  - Cyclically adjusted primary balance (Hamilton): 0.38, -0.08, 0.14, 0.03, 0.86, 42, 0.00
  - Cyclically adjusted primary balance (HP): 0.17, -0.11, 0.16, 0.02, 0.82, 42, 0.00
  - Primary expenditure: 2.50, 0.79, 2.88, 1.21, 0.03, 46, 0.10
  - Government debt-to-GDP ratio: 4.35, 0.37, 5.21, 2.97, 0.50, 43, 0.01
- Policy consistency
  - Var(∆Real short-term rate): 6.35, 7.20, 11.70, 22.28, 0.21, 30, 0.06
  - Var(∆Broad money growth rate): 51.41, 78.17, 47.11, 235.78, 0.03, 23, 0.12
  - Var(∆CAPB (Hamilton)): 1.80, 3.99, 3.37, 3.97, 0.56, 43, 0.01
  - Var(∆CAPB (HP)): 1.79, 4.18, 3.32, 3.97, 0.52, 43, 0.01
  - Var(∆Primary fiscal balance): 2.13, 4.02, 4.22, 3.94, 0.81, 47, 0.00
  - Corr(CAPB (HP), Real short-term rate): 0.02, 0.16, -0.01, 0.02, 0.89, 26, 0.00
  - Corr(CAPB (HP), Broad money growth rate): 0.25, 0.38, 0.10, 0.06, 0.92, 13, 0.00
  - Corr(CAPB (Hamilton), Real short-term rate): -0.15, 0.10, -0.07, -0.02, 0.82, 26, 0.00
  - Corr(CAPB (Hamilton), Broad money growth rate): 0.26, 0.38, 0.14, 0.07, 0.83, 13, 0.00
- External sector
  - Percent of episodes with ER peg (pre-shock): n.a., n.a., 0.43, 0.50, 0.55, 58, 0.01
  - Likelihood of breaking peg (post-shock): n.a., n.a., 0.54, 0.81, 0.09, 31, 0.10
  - NEER: 6.27, -7.97, 0.52, -10.25, 0.02, 38, 0.13
  - REER: 6.53, 0.20, 2.80, 0.01, 0.39, 40, 0.02
  - Trade balance: -1.17, -0.99, -0.75, -0.66, 0.90, 54, 0.00
  - Current account: -0.98, -1.46, -0.20, -0.90, 0.43, 52, 0.01
- Macro outcomes
  - CPI inflation rate: 1.87, 9.04, 2.18, 10.69, 0.00, 60, 0.36
  - Real GDP growth rate: -1.83, -2.12, -2.08, -2.49, 0.45, 59, 0.01
  - Unemployment rate: 0.51, 1.08, 1.15, 1.28, 0.78, 33, 0.00
  - Output gap: -0.93, -2.50, -0.81, -0.74, 0.95, 25, 0.00

---

### Appendix V. Robustness Results (selected tables and highlights)
Notes: See notes under Appendix Table IV.1. Results that are consistent with the baseline results in Section 4 and Table IV.1 are shown in bold in the source.

- Table V.1. Results from the Full Sample — selected entries (Median, Mean, P-value, Obs., R-squared)
  - Real short-term rate: -0.53, -1.93, -0.66, -3.48, 0.005, 10, 0.19
  - Broad money growth rate: -2.98, 2.02, -6.31, 2.37, 0.004, 6, 0.21
  - Deviation from Taylor Rule (output gap-based): 0.04, -3.01, 0.04, -5.21, 0.004, 3, 0.28
  - Inflation volatility (pre-shock): 0.48, 1.01, 0.93, 1.86, 0.019, 9, 0.08
  - CPI inflation rate: 1.52, 6.47, 1.81, 9.61, 0.009, 9, 0.34
  - Primary expenditure: 2.41, 0.88, 2.75, 1.13, 0.017, 6, 0.10

- Table V.2. Full Sample Excluding GFC Episodes — selected entries (Median, Mean, P-value, Obs., R-squared)
  - Real short-term rate: 0.57, -1.93, 0.27, -3.57, 0.0037, 0.31
  - Deviation from Taylor Rule (output gap-based): 0.79, -3.79, 0.85, -5.72, 0.0029, 0.37
  - CPI inflation rate: 1.75, 7.62, 1.99, 9.97, 0.0081, 0.32
  - Primary expenditure: 2.41, 0.72, 2.77, 1.05, 0.0163, 0.11

- Table V.3. 1973–79 Oil Crises without Adjustments — selected entries
  - Real short-term rate: 0.57, -3.31, 0.33, -4.14, 0.00, 30, 0.40
  - Broad money growth rate: -4.28, 6.32, -4.42, 4.89, 0.01, 24, 0.19
  - CPI inflation rate: 1.78, 9.07, 1.78, 10.42, 0.00, 76, 0.38

- Table V.4. 1973–79 Oil Crises with Inflation Shock Size below 10% — selected entries
  - Real short-term rate: 0.87, -1.96, 0.56, -3.15, 0.01, 21, 0.37
  - Broad money growth rate: -7.07, 10.19, -5.15, 8.94, 0.01, 11, 0.55
  - CPI inflation rate: 1.78, 6.61, 1.91, 8.33, 0.00, 47, 0.35

- Table V.5. 1973–79 Oil Crises without Exchange Rate Pegs — selected entries
  - Real short-term rate: 1.79, -2.96, 1.45, -4.29, 0.00, 18, 0.50
  - Deviation from Taylor Rule (output gap-based): 3.66, -8.99, 2.38, -7.47, 0.00, 15, 0.58
  - CPI inflation rate: 1.39, 9.01, 1.60, 10.35, 0.00, 34, 0.38

- Table V.6. 1973–79 Oil Crises at a 2-Year Horizon — selected entries
  - Real short-term rate: -0.02, -4.33, -0.57, -6.38, 0.00, 25, 0.41
  - CPI inflation rate: 4.40, 10.01, 4.99, 10.60, 0.00, 60, 0.27

- Table V.7. 2-Year Horizon with Inflation Shock Size below 10% — selected entries
  - Real short-term rate: -0.06, -3.42, -0.76, -4.41, 0.012, 10, 0.30
  - CPI inflation rate: 3.63, 6.72, 4.46, 8.17, 0.0047, 0.22

*Source: One Hundred Inflation Shocks: Seven Stylized Facts, Working Paper No. WP/2023/190 (IMF).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023190-print-pdf.pdf_