## Inflation Indexation in Public Finances: Annexes, Methodology, Findings (wpiea2023264-print-pdf)

## Source details

**Canonical URL:** [Inflation Indexation in Public Finances: Annexes, Methodology, Findings (wpiea2023264-print-pdf)](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023264-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023264-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023264-print-pdf.pdf.json)

---

### Annexes, Glossary, and Content Inventory
- Annexes included:
  - I. List of Economies and Indexation Practices
  - II. Main Data Sources
  - III. Summary Statistics for the Accounting Exercise
  - IV. Weighted Indexation Values per Economy
- Glossary abbreviations (selected, exact):
  - ADB Asian Development Bank
  - AE Advanced Economies
  - AFR Economies belonging to the African Department at the IMF
  - APD Economies belonging to the Asia and Pacific Department at the IMF
  - ECB European Central Bank
  - EM Emerging Markets
  - EUR Economies belonging to the European Department at the IMF
  - FCS Fragile and Conflicted States
  - GDP Gross Domestic Product
  - IBP International Budget Partnership
  - IMF International Monetary Fund
  - ISSA International Social Security Association
  - LIDC Low-Income and Developing Countries
  - MCD Economies belonging to the Middle East and Central Asia Department at the IMF
  - OBI Open Budget Index
  - OECD Organisation for Economic Cooperation and Development
  - MISSOC Mutual Information System on Social Protection
  - RRC Resource Rich Countries
  - SDS Small and Developing States
  - SSA Social Security Administration
  - WB World Bank
  - WHD Economies belonging to the Western Hemisphere Department at the IMF

### Data compilation, survey design, and coverage
- Two-step data compilation:
  - Review of academic literature and official international and governmental documents.
  - Two IMF desk surveys (sent January 20, 2023; responses by February 22, 2023) with a final internal review in November 2023; information updated for policy changes.
- Survey scope:
  - Surveyed IMF desks of 192 economies in January–February 2023; final internal review in November 2023.
  - Final dataset contains information on latest (as of November 2023) indexation practices for at least 115 and at most 175 economies, depending on the budget item.
  - Dataset includes Hong Kong SAR, Macao SAR, and West Bank Gaza.
- Budget items and sample sizes:
  - Pension indexation: sample = 175 economies.
  - Social assistance programs indexation: sample = 133 economies.
  - Personal income tax threshold indexation: sample = 173 economies.
  - Public sector wages indexation: sample = 121 economies.
- Annex I categorizations (distinct categories, exact):
  - Pension indexation: (i) no automatic indexation; (ii) automatic indexation to prices; (iii) automatic indexation to wages; (iv) automatic mixed (m) indexation to a combination of prices (p), wages (w) etc.
  - Social assistance: (i) no automatic indexation; (ii) automatic indexation to prices; (iii) automatic indexation to other variables.
  - PIT thresholds: (i) ad-hoc adjustment; (ii) automatic threshold adjustment by law; (iii) de facto regular threshold adjustment but with no law.
  - Public wages: (i) no automatic indexation; (ii) automatic indexation to prices; (iii) automatic indexation to other variables.

### Key empirical findings — Pensions
- Sample size: 175 economies.
- Distribution (of 175 economies):
  - 82 economies (47 percent) have no indexation
  - 40 economies (23 percent) index only to prices
  - 15 economies (9 percent) index only to private wages
  - Remaining economies have some form of mixed indexation
- Regional patterns and trends:
  - Europe has the highest degree of pension indexation.
  - Sub-Saharan Africa has the lowest degree of pension indexation.
  - Pension indexation remained almost constant since 2002/2003 in the Americas, Asia, Europe, and the Middle East.
  - Pension indexation meaningfully declined in Sub-Saharan Africa since 2002/2003.
  - Fewer economies index pensions to wages in 2022/23 compared to 2002/03; shift toward less generous mixed and price indexation.

### Key empirical findings — Social assistance programs
- Sample size: 133 economies.
- Distribution (of 133 economies):
  - about 89 economies (65 percent) have no indexation
  - 31 economies (23 percent) have some indexation to prices
  - 15 economies (11 percent) have some indexation to other variables
- Geographic patterns:
  - Most of Sub-Saharan Africa and Asia have not indexed social assistance programs.
  - Indexation to prices or other variables most common in Europe.
  - Most of the developed world including North America, Western Europe, Australia, and China have social assistance programs indexed to prices.
- Explanatory factors:
  - Social assistance typically financed out of general tax revenue (unlike pensions partially financed by contributions).
  - Social assistance programs are often smaller, fragmented, and less well-developed in non-AEs.

### Key empirical findings — Personal Income Tax (PIT) thresholds
- Sample size: 173 economies.
- Distribution (of 173 economies):
  - 18 economies (11 percent) automatically adjust their thresholds
  - 16 economies (10 percent) regularly adjust their thresholds though the process is unclear
  - Majority do not index or automatically adjust PIT brackets
- Geographic patterns:
  - Automatic adjustment most common in the Americas.
  - Europe: many economies have an unclear process.
  - Almost all of Asia and Africa do not regularly adjust PIT thresholds.
- Notable examples:
  - India: has not updated PIT thresholds since FY2016/17.
  - Australia: revised thresholds most recently in 2020-21 with previous revision in 2012-13.
  - United Kingdom: frozen income tax thresholds from 2022 to 2028.

### Key empirical findings — Public sector wages
- Public wages largely unindexed globally; only a few economies index public wages to prices or other variables.
- Examples of indexed public wages:
  - AEs indexing to prices: Belgium, Luxembourg, Malta.
  - Emerging markets indexing public wages: India, Mexico, Türkiye.
  - Eastern European examples indexing to other variables (growth): Slovak Republic, Croatia, Serbia.
- Caveat:
  - Negotiated CPI-linked increases excluded from classification as automatic indexation.

### Box 1 — Indexation of Minimum Wages (selected exact values and patterns)
- Frequency and coverage:
  - 85 economies (representing 54 percent of economies with statutory minimum wages) adjusted minimum wages at least every two years on average during 2010–19 (ILO, 2022).
  - About 50 percent of those 85 economies conducted adjustments every year.
  - During 2010–19:
    - 49 economies adjusted every three to five years.
    - 20 economies adjusted less frequently and/or had no adjustments.
- Regional average intervals (exact):
  - Central Asia and Europe: every 1.9 years.
  - Asia Pacific: 2.7 years.
  - The Americas: 2.9 years.
  - Africa: average interval of 4.7 years.
- By economic status:
  - AEs: on average every 2 years.
  - EMs: on average every 2.5 to 3.7 years.
  - LIDCs: on average every 5 years or more.
- Empirical notes on real effects and inflationary risk:
  - Lindner (2022): in the UK, a 20 percent increase in minimum wages contributed to a rise in inflation of 0.2 percent only.
  - Gautier, Roux and Castillo (2022): minimum wages amplify wage response to past inflation in France microdata.

### Fiscal accounting exercise — conceptual model and empirical illustration
- Conceptual identity:
  - Δ(r−x) = Δπ_def * ε_r − Δπ_cpi * Σ ε_i^x_i
  - Definitions:
    - r = revenue; x = primary expenditure; i indexes expenditure items.
    - Δπ_cpi = CPI inflation surprise.
    - Δπ_def = GDP deflator surprise.
    - ε_r = revenue elasticity to inflation.
    - ε_i^x = primary expenditure item elasticity to inflation.
- Modeling assumptions (exact):
  - Non-indexed primary expenditure is fixed in nominal terms: ε_non-indexed^x = 0.
  - Indexed primary expenditure responds one-to-one to relevant CPI shock: ε_indexed^x = 1 (possibly not in the concurrent year).
- Timing considerations:
  - Many economies update expenditure the year after the inflation surprise; some update within the fiscal year creating immediate fiscal effects.
  - Tax thresholds updated prior to the tax year may miss within-year wage increases from inflation surprises (bracket creep channel).
- Empirical illustrations (selected country highlights and mechanisms):
  - Belgium: effects strongest where wage bill is automatically indexed (health index / pivot index breaches in 2022).
  - Italy: large pension expenditures combined with pension indexation produce large impacts.
  - Türkiye: for many EMs impact small because inflation surprises were smaller, except Türkiye where impacts were notable.
  - Social transfers: generally small fiscal impact because transfers account for a small share of expenditure.
- Accounting exercise caveat:
  - Does not capture second-order effects (e.g., higher income tax revenue from pension indexation expanding tax bases).

### Annex III — Summary statistics for the accounting exercise (exact values)
- Average values:
  - Inflation shock = 5.75
  - Deflator shock = 3.64
  - Wage Bill = 5.95
  - Pensions = 6.84
  - Social Spending = 2.09
- Median values:
  - Inflation shock = 4.50
  - Deflator shock = 1.84
  - Wage Bill = 5.20
  - Pensions = 6.09
  - Social Spending = 1.92
- Percentiles:
  - 25th percentile: Inflation shock = 3.65; Deflator shock = 1.27; Wage Bill = 4.03; Pensions = 4.36; Social Spending = 1.36
  - 75th percentile: Inflation shock = 5.40; Deflator shock = 3.50; Wage Bill = 7.13; Pensions = 8.34; Social Spending = 2.76
- Number of observations:
  - Inflation shock: 9
  - Deflator shock: 12
  - Wage Bill: 4
  - Pensions: 10
  - Social Spending: 5

### Weighted indexation, inflation persistence, and statistical associations (2018–2022)
- Weighted Indexation measure:
  - Constructed as sum of shares of wage bill, pensions, and social assistance in total current spending multiplied by indexation dummies for each item; ranges between 0 and 1.
  - Sample for weighted-indexation analysis: 58 economies (27 AEs, 23 EMs, 8 LIDCs).
  - Average weighted indexation = 0.13.
  - Maximum observed weighted indexation = 59 percent (Cyprus, period analyzed 2018–2020).
- Hurst exponent (persistence) analysis:
  - H = d + 0.5, where d estimated via ARFIMA using monthly inflation data January 2018–December 2022.
  - Interpretation: H > 0.5 = persistence; H = 0.5 = random walk; H < 0.5 = anti-persistent.
  - Correlation results (exact associations reported):
    - 10 percent increase in indexed public spending → Hurst exponent +0.05.
- AR(12) autoregressive analysis:
  - Persistence measured as sum of AR coefficients for first 12 months using monthly inflation data 2018–2022 for 58 economies.
  - Empirical correlation:
    - 10 percent increase in indexed public spending → sum of AR coefficients +0.26.
- Half-life approximation (AR(1) approximation):
  - 20 percent increase in share of indexed spending → half-life of inflation ≈ +1 month.
- Cross-sample summary:
  - Among 53 economies included in Hurst calculations, highest average H observed among AEs and EMs; lowest among LIDCs.
  - Even economies with no indexation tend to have, on average, persistent inflation (y-intercept approximately 0.6 in regression chart).

### Annex IV — Selected weighted indexation values (exact entries)
- Selected Weighted Indexation 2020 ; 3-Year (2018–2020) Weighted Indexation Average:
  - Italy AE: 0.56 ; 0.49
  - Luxembourg AE: 0.53 ; 0.47
  - Cyprus AE: 0.52 ; 0.58
  - Belgium AE: 0.46 ; 0.42
  - Turkey EM: 0.45 ; 0.45
  - Mexico EM: 0.35 ; 0.39
  - Moldova LIDC: 0.32 ; 0.35
  - Belarus EM: 0.28 ; 0.29
  - Finland AE: 0.27 ; 0.19
  - France AE: 0.27 ; 0.28
  - Switzerland AE: 0.26 ; 0.27
  - Croatia EM: 0.24 ; 0.27
  - Slovenia AE: 0.24 ; 0.17
  - Slovak Republic AE: 0.23 ; 0.24
  - Uzbekistan EM: 0.23 ; 0.25
  - Serbia EM: 0.21 ; 0.24
  - Lithuania AE: 0.21 ; 0.15
  - Denmark AE: 0.20 ; 0.14
  - Estonia AE: 0.20 ; 0.14
  - Czech Republic AE: 0.20 ; 0.21
  - Latvia AE: 0.19 ; 0.14
  - United Kingdom AE: 0.17 ; 0.20
  - Sweden AE: 0.16 ; 0.11
  - Netherlands AE: 0.16 ; 0.11
  - United States AE: 0.13 ; 0.15
  - Mauritius EM: 0.12 ; 0.10
  - Chile EM: 0.12 ; 0.13
  - Canada AE: 0.10 ; 0.13
  - Ecuador EM: 0.09 ; 0.09
  - Israel AE: 0.09 ; 0.10
  - Korea AE: 0.09 ; 0.10
  - Hong Kong SAR AE: 0.07 ; 0.10
- Examples of zero weighted indexation entries (exact):
  - El Salvador EM: 0 ; 0
  - Kiribati EM: 0 ; 0
  - Côte d'Ivoire LIDC: 0 ; 0
  - Bulgaria EM: 0 ; 0
  - Spain AE: 0 ; 0
  - Seychelles EM: 0 ; 0
  - Rwanda LIDC: 0 ; 0
  - Myanmar LIDC: 0 ; 0
  - Thailand EM: 0 ; 0
  - Macao, SAR AE: 0 ; 0
  - Cambodia LIDC: 0 ; 0
  - United Arab Emirates EM: 0 ; 0
  - Samoa EM: 0 ; 0
  - Ukraine EM: 0 ; 0
  - Albania EM: 0 ; 0
  - Ireland AE: 0 ; 0
  - Singapore AE: 0 ; 0
  - Mongolia LIDC: 0 ; 0
  - Senegal LIDC: 0 ; 0
  - South Africa EM: 0 ; 0
  - Armenia EM: 0 ; 0
  - Namibia EM: 0 ; 0
  - Austria AE: 0 ; 0
  - Nepal LIDC: 0 ; 0
  - Bosnia and Herzegovina EM: 0 ; 0

### Governance, country characteristics, and indexation associations
- Government effectiveness (Worldwide Governance Indicators: -2.5 to 2.5):
  - Positive correlation between government effectiveness and indexation in the public sector, particularly for social assistance programs and public wages.
- Budget transparency (Open Budget Index, OBI: 0–100):
  - Higher OBI positively correlated with more indexation and more regular (de facto) adjustments.
  - Pensions indexation lowest in economies with lower OBI; economies with OBI in highest quadrant (75-100) have mixed types of pensions indexation.
- Fragile and conflicted states (FCS):
  - Indexation more common in stable economies than in FCS.
- Small Developing States (SDS) and Resource-Rich Countries (RRC):
  - No SDS has indexed PIT thresholds.
  - SDS have higher proportion of pensions indexed to prices and higher degree of indexation of public wages (mostly to variables other than prices).
  - RRCs are less likely to have pensions, social assistance programs, and public wages indexed compared to non-RRCs.
  - Indexation of PITs appears more prevalent in RRCs than non-RRCs.

### Conclusions and policy guidance (exact policy messages extracted)
- Trade-offs:
  - Benefits: preserves purchasing power for targeted groups and programs; can improve economic well-being and reduce poverty over time.
  - Costs/risks: can make inflation more persistent by creating inflationary inertia and raising backward-looking inflation expectations; during acute inflation surprises, indexation can substantially increase budgetary spending and deteriorate fiscal balances; public wage indexation can affect private wage setting and contribute to sticky inflation.
- Policy guidance for reviewing indexation:
  - Decide which groups and programs to protect from income erosion while avoiding policies that entrench inflation.
  - Carefully assess the impact of public wage setting during high inflation episodes on private wages and inflation persistence.
  - Consider potential effects of inflation on the tax system structure and thresholds to manage bracket creep and revenue implications.
- Suggested research extensions:
  - Expand the dataset over time to create a panel dataset.
  - Include additional budget items.
  - Analyze how indexation affects cyclical fiscal responses to inflation, long-term distributional implications, and causal links between indexation and sticky inflation.

*Italic: Source — IMF WORKING PAPERS, Inflation Indexation in Public Finances: A Global Dataset on Current Practices (content unit: wpiea2023264-print-pdf).*

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

### References

### Annexes and Contents
- Annexes listed:
  - I. List of Economies and Indexation Practices
  - II. Main Data Sources
  - III. Summary Statistics for the Accounting Exercise
  - IV. Weighted Indexation Values per Economy

### Glossary
- Abbreviations and agencies defined exactly as listed:
  - ADB Asian Development Bank
  - AE Advanced Economies
  - AFR Economies belonging to the African Department at the IMF
  - APD Economies belonging to the Asia and Pacific Department at the IMF
  - ECB European Central Bank
  - EM Emerging Markets
  - EUR Economies belonging to the European Department at the IMF
  - FCS Fragile and Conflicted States
  - GDP Gross Domestic Product
  - IBP International Budget Partnership
  - IMF International Monetary Fund
  - ISSA International Social Security Association
  - LIDC Low-Income and Developing Countries
  - MCD Economies belonging to the Middle East and Central Asia Department at the IMF
  - OBI Open Budget Index
  - OECD Organisation for Economic Cooperation and Development
  - MISSOC Mutual Information System on Social Protection
  - RRC Resource Rich Countries
  - SDS Small and Developing States
  - SSA Social Security Administration
  - WB World Bank
  - WHD Economies belonging to the Western Hemisphere Department at the IMF

### I. Introduction — Key points and contributions
- Topic and scope:
  - Study focus: indexation of payments and remunerations to the price level or to private (minimum) wages across four budget items: (i) personal income tax brackets; (ii) pensions; (iii) social assistance programs; and (iv) public wages.
  - Claim: dataset is the only existent information about the current status of indexation of budget items with a global scope covering both a revenue and three expenditure items.
- Data collection approach:
  - Two-step preparation: review academic literature and official international and governmental documents; two surveys among IMF country desks of its 190 country-members in February and November 2023, receiving information from 175 of them.
- Literature context:
  - Historical literature concentrated in 1965–1982 (Great Inflation) with references to Barro (1976); Gray (1976); Fisher (1977); Blanchard (1979); empirical work including Fischer (1986); Holland (1995).
  - Pension indexation remained widespread during the Great Moderation (Bank of International Settlements, 2022); Whitehouse (2009) noted pension indexation spread from half-a-dozen OECD countries in the 1960s to virtually the whole OECD.
  - Recent literature renewed by global inflation: Carrillo, Peersman, and Wauters (2022); Gautier, Roux, and Castillo (2022); Beer, Griffiths, and Klemm (2023); Galeano and others (2023); Guerrieri and others (2023).
- Main contributions of the paper:
  - Creation of a new global dataset on current international indexation practices, presentation of descriptive statistics, and correlation analyses (e.g., indexation with budget transparency).
  - Two applications illustrated:
    - An accounting exercise comparing baseline budgets for the 2022 fiscal year with outturns in 12 select economies to quantify contribution of inflation surprises to realized fiscal primary balance through indexation practices. The 12 economies: Belgium, Brazil, France, Germany, Italy, Mexico, South Africa, Tanzania, Türkiye, United Kingdom, and United States.
    - Quantification of the combined degree of indexation of the three spending items for each economy and examination of its correlation with inflation persistence.
- Main findings summarized:
  - Indexation policies vary significantly across economies and regions (AEs, EMs, LIDCs).
  - Indexation is more common on politically salient expenditure items (pensions and social grants) than on taxes.
  - Degree of indexation falls with the level of economic development.
  - Indexation is associated with more effective and transparent governments, and resource-rich economies appear to rely less on indexation.
  - The accounting exercise suggests indexation can have significant impacts on fiscal outcomes during episodes of inflation surprises.
  - Evidence of a positive correlation between the share of spending indexed to inflation and inflation persistency in the sample.

### II. Data Compilation Process — Methods and coverage
- Survey scope and timeline:
  - Surveyed IMF desks of 192 economies in January–February 2023 with a final internal review in November 2023 that resulted in updates reflecting recent policy changes.
  - Note: dataset includes Hong Kong SAR, Macao SAR, and West Bank Gaza.
- A. Review of the literature and official documents — sources and sample sizes by budget item:
  - Pension indexation:
    - Sample: 175 economies on current pension indexation practices in state pensions systems.
    - Sources: OECD (2021); ECB (2022); for most LIDCs and EMs, “Social Security Programs Throughout the World” reports (2002/03 through 2018/19) produced by the United States Social Security Administration (SSA) in collaboration with ISSA (ISSA, 2022).
    - Method: “benefits adjustments” of old-age benefits analyzed to identify regularity; classified as indexed if regular and then categorized by type (prices, wages, mixed/other such as GDP growth); if “benefits adjustments” do not indicate regular adjustments, classified as “no indexation.”
    - Time-series exploration: compiled pension indexation for years 2002/03 and 2018/19 for a sample of 118 economies to explore evolution over two decades; most recent vintage (2018/19) updated with IMF desk survey responses in February and November 2023.
  - Social assistance programs indexation:
    - Sample: 133 economies on current social assistance program indexation practices.
    - Sources: MISSOC (2022) detailed database of 31 economies; OECD (2023); SSA/ISSA “Social Security Programs Throughout the World” (2018/19) for others.
    - Method: focus on “benefits adjustments” of family allowances and guaranteed minimum income; economies with regular adjustments coded as indexed; type of indexation (prices or other variables such as expenditure surveys of low-income households) identified; absence of regular adjustments coded as “no indexation.” Dataset verified and updated via IMF desk survey in February and November 2023.
  - Personal income tax threshold indexation:
    - Sample: 173 economies classified by whether personal income tax bracket adjustments are automatically tied to inflation, regularly adjusted but with an unclear process (de facto), or not adjusted regularly.
    - Sources: Beer, Griffiths, and Klemm (2023) using IBFD country tax profiles (IBFD, 2022); OECD (2023); updated with IMF desk survey in February and November 2023.
  - Public sector wages indexation:
    - Sample: 121 economies.
    - Sources: IMF 2016 survey among IMF country desks on pay setting systems (77 economies); updated and expanded using IMF (2018) survey question “Are base wages indexed?”; ECB (2022) individual country reports; further updated via IMF desk surveys in February and November 2023.
- B. Survey of Indexation Practices among IMF Country Desks — process details:
  - Initial database compiled from literature and documents, then submitted for review by IMF country desks of 192 member economies.
  - Survey sent to all economists working as country desks on January 20, 2023 with a deadline for responses of February 22, 2023.
  - Final internal review in November 2023 produced updates reflecting recent policy changes.
  - Survey instrument presented categories for each budget item; country desks were asked to double-check correctness and complete missing information.

*IMF WORKING PAPERS Inflation Indexation in Public Finances: A Global Dataset on Current Practices — References section and related annex and methodology content (excerpts).*

### Annex I, the distinct categories for each of the budget item were:

### Annex I — Distinct categories for each budget item (from IMF Working Paper: Inflation Indexation in Public Finances: A Global Dataset on Current Practices)

### Indexation categorizations (Annex I)
- Pension indexation categories:
  - (i) no automatic indexation
  - (ii) automatic indexation to prices
  - (iii) automatic indexation to wages
  - (iv) automatic mixed (m) indexation to a combination of prices (p), wages (w) etc.
- Social assistance programs indexation categories:
  - (i) no automatic indexation
  - (ii) automatic indexation to prices
  - (iii) automatic indexation to other variables
- Personal income tax (PIT) threshold indexation categories:
  - (i) ad-hoc adjustment
  - (ii) automatic threshold adjustment by law
  - (iii) de facto regular threshold adjustment but with no law
- Public sector wages indexation categories:
  - (i) no automatic indexation
  - (ii) automatic indexation to prices
  - (iii) automatic indexation to other variables

### Survey coverage and scope
- Survey responses covered at least 115 and at most 175 economies.
- The total number of countries vary across the four budget items, depending on the availability/verification of data.
- The survey captures the broad approach to indexation rather than an exhaustive account of adjustment mechanisms (partial vs full indexation, forward-looking vs backward-looking, frequency, adjusters, etc., were beyond scope).
- The database generally captures both de jure as well as de facto automatic indexation.
- The information does not capture temporary deviations from customary indexation practices.

### Key findings — Pensions
- Sample size: 175 economies.
- Of the 175 economies:
  - 82 economies (47 percent) have no indexation
  - 40 economies (23 percent) index only to prices
  - 15 economies (9 percent) index only to private wages
  - The remaining economies have some form of mixed indexation
- Geographic patterns:
  - Europe is the region with the highest degree of pension indexation.
  - Sub-Saharan Africa is the region with the lowest degree of pension indexation.
- Evolution and trends:
  - Number of economies using some type of pension indexation has remained almost constant since 2002/2003 in the Americas, Asia, Europe, and the Middle East.
  - Pension indexation has meaningfully declined in Sub-Saharan Africa since 2002/2003.
  - Pension indexation has become less generous over time to minimize fiscal burdens and safeguard financial sustainability; fewer economies index pensions to wages in 2022/23 compared to 2002/03, with a shift toward less generous mixed and price indexation.
  - In the current scenario where nominal wage increases have lagged price inflation, price indexation may be costlier than anticipated given that social security contributions track nominal wage growth.

### Mixed indexation examples and heterogeneity (from Annex notes)
- Mixed indexation includes:
  - Weighted indices of prices and wages (e.g., Czech Republic: 50 percent p plus 50 percent w; Slovenia: 40 percent p plus 60 percent w).
  - Mixes including sustainability factors and automatic adjustment mechanisms (AAMs) (examples: Germany with w sustainability factor; UK triple lock: maximum{p,w,2.5 percent}; Sweden and Finland with w plus AAMs; Luxembourg with automatic price indexation when six-month moving average of NICP has increased by 2.5 percent since preceding wage indexation, plus adjustment with respect to evolution of real wages every two years).
  - Indexation varying by age or program (e.g., Georgia: indexation to prices until age 70 and to p plus real GDP growth rate from age 70; Japan: wages until age 68 and prices from age 68 on, plus AAMs).

### Key findings — Social assistance programs
- Sample size: 133 economies.
- Of the 133 economies:
  - about 89 economies (65 percent) have no indexation
  - 31 economies (23 percent) have some indexation to prices
  - 15 economies (11 percent) have some indexation to other variables
- Geographic patterns:
  - Most of Sub-Saharan Africa and Asia have not indexed social assistance programs.
  - Indexation to prices or other variables is most common in Europe.
  - Most of the developed world including North America, Western Europe, Australia, and China have social assistance programs indexed to prices.
- Possible explanatory factors for lower indexation of social assistance programs relative to pensions:
  - Social security part of pensions is at least partly financed by pension contributions; social assistance programs are financed out of general tax revenue.
  - Social assistance programs tend to be smaller, fragmented, and less well-developed in non-AEs compared to pensions.
  - Social assistance programs aim to minimize welfare dependency in the longer term compared to pensions.

### Key findings — Personal Income Tax (PIT) thresholds
- Sample size: 173 economies.
- Of the 173 economies:
  - 18 economies (11 percent) automatically adjust their thresholds
  - 16 economies (10 percent) regularly adjust their thresholds though the process is unclear
  - The majority of governments do not index or automatically adjust PIT brackets
- Geographic patterns:
  - Automatic adjustment of thresholds is most common in the Americas.
  - Most economies in Europe have an unclear process around PIT threshold indexation.
  - Almost all of Asia and Africa do not regularly adjust PIT thresholds.
- Examples:
  - India has not updated its PIT thresholds since FY2016/17.
  - Australia revised its thresholds most recently in 2020-21 with the previous revision in 2012-13.
  - United Kingdom has frozen its income tax thresholds from 2022 to 2028.

### Key findings — Public sector wages
- Public wages are largely unindexed globally; only a few economies index public wages to prices or other variables.
- Examples of indexed public wages:
  - Advanced economies indexing to prices include Belgium, Luxembourg, and Malta.
  - Leading emerging markets indexing public wages include India, Mexico, and Türkiye.
  - Eastern European examples index to other variables (growth): Slovak Republic, Croatia, Serbia.
- Contrasts and caveats:
  - Findings contrast with BIS (2022) which reports indexation practices more prevalent in Latin America relative to Asia.
  - The data does not capture non-automatic wage negotiations that consider past/future inflation; such negotiated CPI-linked increases are excluded from classification as automatic indexation.

### Indexation patterns across economic regions (AEs, EMs, LIDCs)
- General pattern:
  - Indexation is generally most prevalent in AEs, with the majority of these economies indexing pensions and social assistance programs.
  - Emerging markets have a higher proportion of economies that index their public wages.
  - Indexation of PIT thresholds is most common in AEs but still below 50 percent of those economies in the sample.
  - Indexation of PITs is the lowest in LIDCs, where it is below 5 percent of the sample.
- Pension-specific regional shares:
  - Nearly all AEs have some form of pension indexation.
  - About 50 percent of EMs have some form of pension indexation.
  - About 30 percent of LIDCs have some form of pension indexation.
- Temporal regional shifts:
  - Increase in the degree of indexation in AEs over time.
  - Decrease in the degree of indexation in EMs and LIDCs over time.
  - Across regions, shifts are toward less generous indexation with a fewer share of economies indexing to wages.

*Source: IMF Working Papers — Inflation Indexation in Public Finances: A Global Dataset on Current Practices*

### Box 1: Indexation of Minimum Wages

### Box 1: Indexation of Minimum Wages

### Frequency and patterns of minimum wage adjustments
- 85 economies (representing 54 percent of economies with statutory minimum wages around the globe) adjusted their minimum wages at least every two years on average during the decade 2010–19 (ILO, 2022).
- About 50 percent of those 85 economies conducted these adjustments every year (ILO, 2022).
- During 2010–19:
  - 49 economies adjusted their minimum wages every three to five years.
  - 20 economies adjusted less frequently and/or had no adjustments at all.
- Regional average intervals between adjustments:
  - Central Asia and Europe: every 1.9 years.
  - Asia Pacific: 2.7 years.
  - The Americas: 2.9 years.
  - Africa: average interval of 4.7 years.
- By economic status:
  - AEs: on average every 2 years.
  - EMs: on average every 2.5 to 3.7 years.
  - LIDCs: on average every 5 years or more.

### Indexation methods and country examples
- Price developments (CPI) are one of the most considered factors when stakeholders update minimum wages (OECD, 2015, 2022b; Eurofound, 2022).
- Country-specific practices described:
  - Belgium: all wages are indexed to past CPI (excluding alcohol and narcotics) every time that price index increases by 2 percent or more.
  - France: only the minimum wage of the bottom quintile is indexed to past CPI.
  - Luxembourg: wages are automatically indexed to past CPI.
  - Poland: indexes minimum wages to future inflation and corrects it the following year if the inflation forecasts differ from the real price index.
  - Netherlands: the minimum wage is indexed to average wage agreements instead of prices.
- Some economies use minimum wages to index pensions and social benefits (example noted: the Netherlands).

### Real effects, risks, and empirical findings
- Despite prevalence of indexation practices, hikes in minimum wages often fall short of the rise in prices, leading to decreases in real minimum wages (OECD, 2022b).
- Concerns about minimum-wage increases contributing to inflation exist, but empirical magnitudes appear limited in some studies:
  - Lindner (2022): in the UK, a 20 percent increase in minimum wages contributed to a rise in inflation of 0.2 percent only—context: about 5 percent of the workforce is remunerated with the minimum wage (Francis-Devine, 2023).
  - Koester and Grapow (2021): argue the likelihood of a wage-price spiral is relatively limited in the euro area given the current extent of indexation.
  - Gautier, Roux and Castillo (2022): using micro data from France, document that minimum wages contribute to amplifying the response of wages to past inflation.

### Indexation of social assistance and prevalence across income groups
- Indexing social assistance benefits is more common in AEs: over 65 percent of economies in AEs index some or any of such programs to either prices or other variables including average/minimum wages.
- Most economies in both EMs and LIDCs do not index their social benefits.
- Social benefits primarily indexed include disability allowance, old-age allowance, childcare benefits, and unemployment benefits.

### Interaction with governance, transparency, stability, size, and resources
- Government effectiveness:
  - Positive correlation between government effectiveness (Worldwide Governance Indicators scale: -2.5 to 2.5) and indexation in the public sector, particularly for social assistance programs and public wages.
  - Economies with higher government effectiveness scores use indexation more frequently.
- Budget transparency (Open Budget Index, OBI: 0–100):
  - Higher OBI is positively correlated with more indexation and with more regular (de facto) adjustments.
  - Indexation of PIT is more common in economies with higher OBI.
  - Pensions indexation is lowest in economies with lower OBI scores; economies with OBI in the highest quadrant (75-100) have mixed types of pensions indexation.
  - Public wages are less indexed across OBI scores; when indexed they are mostly indexed to variables other than prices.
- Fragile and conflicted states (FCS):
  - Indexation is more common in stable economies than in FCS.
  - FCS classification criteria include CPIA score below 3, presence of UN peace operations in the past 3 years, refugees of 2,000 or more per 10,000 population, or identification as a country in conflict based on conflict deaths (IMF, 2023b).
- Country size (Small Developing States, SDS):
  - No SDS has indexed PIT thresholds.
  - Pensions: higher proportion of SDS have pensions indexed to prices compared to non-SDS.
  - Public wages: higher degree of indexation among SDS than in larger economies; in SDS public wages tend to be indexed to variables other than prices.
- Resource-Rich Countries (RRC; at least 20 percent of exports or fiscal revenues from nonrenewable natural resources):
  - RRCs are less likely to have pensions, social assistance programs, and public wages indexed compared to non-RRCs.
  - Indexation of PITs appears to be more prevalent in RRCs than in non-RRCs.

### Fiscal implications and accounting exercise overview
- The impact of indexation on fiscal outcomes during high inflation depends on frequency, choice of indices, and degree of indexation.
- Accounting model (conceptual summary):
  - Effect of an inflation surprise on nominal revenues and expenditures captured by:
    Δ(r−x) = Δπ_def * ε_r − Δπ_cpi * Σ ε_i^x_i
  - Definitions:
    - r = revenue; x = primary expenditure; i indexes individual expenditure items.
    - Δπ_cpi = CPI inflation surprise.
    - Δπ_def = GDP deflator surprise.
    - ε_r = revenue elasticity to inflation.
    - ε_i^x = primary expenditure item elasticity to inflation.
  - Assumptions:
    - Non-indexed primary expenditure is fixed in nominal terms: ε_non-indexed^x = 0.
    - Indexed primary expenditure responds one-to-one to relevant CPI shock: ε_indexed^x = 1 (possibly not in the concurrent year).
    - Expenditure items focused: wages, pensions, social protection programs indexed to prices.
    - Revenue channel considered: gains from bracket creep in PIT (ε_r > 1).
  - Timing matters: many economies update expenditure the year after the inflation surprise; some update within the fiscal year using outturns or forecasts, creating immediate fiscal effects.
  - Tax thresholds updated prior to the tax year miss within-year wage increases driven by inflation surprises.
- Empirical illustration (select economies, FYs 2022 and 2023):
  - Effects of indexation strongest for economies that automatically index their wage bill (example highlighted: Belgium).
  - Large pension expenditures combined with pension indexation generate large impacts (example highlighted: Italy).
  - For many EMs, the impact of indexation was small because inflation surprises were generally smaller, except in Türkiye.
  - Indexation of social transfers generally has a small fiscal impact because such transfers account for a small share of expenditure.
- Table 1 (referenced) summarizes for selected economies:
  - Which budget items are affected by indexation (Wage bill, Pensions, Social grants, Personal Income Tax Bracket Creep).
  - The relevant inflation measure used for indexation in each case and the timing (concurrent year vs. lagged adjustment).

*IMF Working Paper — Box 1: Indexation of Minimum Wages (excerpt)*

### 2023. The other

### wpiea2023264-print-pdf - 2023. The other

### Methodology and general approach
- Inflation surprises are measured as differences between actual outturns or later forecasts and earlier forecasts (examples: difference between November 2022 and March 2021 projections; difference between June 2022 forecast and December 2021 forecast of HCPI projection of Bank of France).
- Impact on income tax due to bracket creep is estimated by multiplying a bracket creep elasticity by the difference between more recent earnings or deflator projections and earlier projections (examples: projected gains from bracket creep in "Policy decisions at Budget 2021" as a share of average earnings growth; difference between projected deflator for 2022 in the October 2022 WEO compared to the October 2021 WEO).
- Where indexation rules are not automatic, historical practice (e.g., increases in line with inflation) is noted but not assumed automatic unless specified.
- For some countries elasticity proxies are used (examples: bracket creep elasticity proxied by the average of US, UK, Germany, and France; bracket creep elasticity estimated using estimated income losses in the Budgetary Plan 2023).

### Country findings and estimated impacts
- United Kingdom (implicit)
  - Allowances are not automatically linked; practice has been to increase in line with inflation.
  - Impact in FY2022-2023 assessed as income tax brackets were maintained at the FY2021-2022 level.
  - Estimated using projected gains from bracket creep in "Policy decisions at Budget 2021" as a share of average earnings growth projected for this period multiplied by the difference between the most recent average earnings growth projection in November 2022 compared to March 2021 for FY2022-2023 from the Office of Budget Responsibility Economic and Fiscal Outlook.

- Germany
  - No automatic indexation to prices.
  - Impact in 2022 due to higher expected average wage growth.
  - Bracket creep elasticity is estimated using the estimated income losses in the Budgetary Plan 2023 from revising the income thresholds in 2023 and multiplied by the difference in compensation per employee for 2022 estimated in the Budgetary Plan 2023 compared to the projection for 2023.

- France
  - No automatic indexation (noted).
  - Basic pensions (about 67 percent of the total pension bill) are automatically indexed to CPI (excluding tobacco) on 1 January.
  - Inflation shock measured as difference in actual outturn minus the HCPI projection of Bank of France (projection referenced: December 2021).
  - Impact in 2022: social benefits adjusted around mid-year; some are backward-looking (means-tested allowances) while others are forward-looking (family benefits).
  - Proxy for inflation surprise: difference between June 2022 forecast and December 2021 forecast of HCPI projection of Bank of France.
  - Spending on Family Allowances, Disability Benefits and Income Maintenance from the OECD Social Expenditure Database is included.
  - Impact in 2022 due to higher expected average wage growth (bracket creep): bracket creep elasticity estimated using estimated income losses in 2023 from revising income thresholds multiplied by difference in estimated GDP deflator for 2022 in October 2022 WEO compared to October 2021 WEO projection.

- Italy
  - No impact in 2022-23 from public sector wages because unexpected inflation in a given three-year contract period is compensated during the following three-year period; next agreement with the Public Sector Unions Federation covering 2022-2024 will be in May 2024.
  - Impact in 2023 as most pensions (those below a specified threshold) are fully indexed and those pensions higher than a specified threshold are subject to partial indexation.
  - Relevant index: CPI for white- and blue-collar workers, excluding tobacco products.
  - At the aggregate level effective indexation rate is close to 100 percent of the previous-year inflation.
  - Inflation shock calculated as difference between projected increase in benefit size in the 2023 Budget (actual increase) and the HICP projection in December 2021 for 2022 (expected increase prior to inflation surprise).
  - Same approach applied to Pensions.
  - Spending on Family Allowances, Care Allowances, and Disability Benefits is included from the OECD Social Expenditure Database.
  - Impact in 2022 as income tax brackets were maintained at the 2021 level; bracket creep elasticity proxied by the average of US, UK, Germany, and France and multiplied by the difference in projected deflator for 2022 in the October 2022 WEO compared to the October 2021 WEO.

- Belgium
  - Adjustments: pensions adjusted by 2 percent two months after the cost-of-living measure "smoothed health index" reaches a threshold called the “pivot inde ”.
  - Health index definition: equal to the CPI excluding alcohol, tobacco and fossil fuels; smoothed health index is the four-month moving average of the health index.
  - For 2022 the "pivot index" was breached five times in 2022 - the last one in November 2022 compared to a baseline assumption of one breach in 2022 without the inflation surprise.
  - Same treatment as Pensions for other benefits.
  - Spending on Family Allowances and Disability Benefits is included from the OECD Social Expenditure Database.
  - Impact in 2022 due to higher expected average wage growth compared to the increase in tax brackets announced in end-2021.

### Included spending categories and data sources
- OECD Social Expenditure Database used for:
  - Family Allowances
  - Disability Benefits
  - Income Maintenance (France)
  - Care Allowances (Italy)
- Budgetary and forecast sources used for elasticity and impact estimation:
  - "Policy decisions at Budget 2021"
  - Budgetary Plan 2023
  - Office of Budget Responsibility Economic and Fiscal Outlook
  - Bank of France HCPI projections
  - October 2022 WEO and October 2021 WEO comparisons
  - HICP projection in December 2021

*IMF WORKING PAPERS Inflation Indexation in Public Finances: A Global Dataset on Current Practices*

### 2022. The bracket creep elasticity is

### wpiea2023264-print-pdf - 2022. The bracket creep elasticity is

### Bracket creep estimation and country-specific treatments
- Method for estimating bracket creep in many countries:
  - The bracket creep elasticity is proxied by the average (or a proxy economy) and multiplied by the difference in the projected deflator of October 2022 WEO for 2022 compared to the actual increase in tax brackets (or the relevant prior projection).
- Country-specific examples documented:
  - Tanzania:
    - No automatic indexation for pensions, wages, or income tax brackets.
    - Impact in 2022 as income tax brackets were maintained at the 2021 level.
    - Bracket creep elasticity proxied by the measure estimated for South Africa and adjusted by the relative share of personal income tax revenues to GDP of Tanzania compared to South Africa.
    - The estimated elasticity is multiplied by the difference in the projected deflator of October 2022 WEO for 2022 compared to the October 2021 WEO for 2022.
  - Brazil:
    - No automatic indexation for income tax brackets.
    - Impact in 2022 as tax brackets have not been adjusted in 5 years.
    - Bracket creep elasticity proxied by the measure estimated for South Africa and adjusted by the relative personal income tax revenues to GDP of Brazil compared to South Africa.
    - Estimated elasticity multiplied by the difference in the projected deflated of October 2022 WEO for 2022 compared to the October 2021 WEO for 2021.
    - Pension indexation: pensions adjusted in tandem with the minimum wage; the degree of indexation has been decreased for higher pensions with the Budget Law for 2023.
  - Türkiye:
    - Wages adjusted twice a year (January and July), reflecting the previous 6-month inflation rate; impact in 2022 from higher-than-envisaged increase in wages in second-half of year.
    - Pensions adjusted twice a year along with wages of government employees; impact in 2022.
    - No automatic indexation for income tax brackets; bracket creep elasticity proxied by South Africa and adjusted by Turkey’s personal income tax revenues to GDP compared to South Africa; elasticity multiplied by difference in actual public sector wage increase in 2022 compared to projected increase in end-2021.
  - South Africa:
    - No automatic indexation for pensions, wages, or income tax brackets.
    - Impact in FY2022 (April–March) due to higher expected average wage growth compared to the increase in income tax brackets announced in March 2021 for FY2022.
    - Bracket creep elasticity based on bracket creep estimates in the FY2022 budget, multiplied by the difference in the projected deflator for 2022 from the October 2022 WEO compared to the October 2021 WEO.
  - Mexico:
    - Wages and pensions increase by inflation in January of each successive year; impact in 2023.
    - Inflation surprise measured by the difference in estimated inflation for 2022 in November 2022 AIV compared to the projection in November 2021 AIV.
    - No automatic indexation for income tax brackets; impact in 2022 as brackets were maintained at 2021 level; bracket creep elasticity proxied by South Africa and adjusted by relative share of personal income tax revenues to GDP of Mexico compared to South Africa. Inflation surprise measured by difference in the estimated deflator in 2022 from the November 2022 AIV compared to the projection in November 2021 AIV.
  - India:
    - Impact in FY2022/23 (April–March) measured by change in the Dearness Allowance (adjusted in July and January, reflecting previous 6-month inflation rate based on AICPI).
    - Dearness Relief for pensions similarly adjusted in July and January based on AICPI.
    - No automatic indexation for income tax brackets; impact in 2022 as tax brackets were maintained at 2021 level.
    - Bracket creep elasticity proxied by South Africa and adjusted by relative share of personal income tax revenues to GDP of India compared to South Africa.
    - Inflation surprise measured by difference in estimated deflator for 2022 from the October 2022 WEO compared to the projection in October 2021 WEO.
- Methodological notes and proxies:
  - When direct country elasticity estimates are unavailable, the South Africa estimate is frequently used as a proxy and adjusted by the relative share of personal income tax revenues to GDP.
  - In economies where adjustments occur mid-fiscal year, impacts are scaled accordingly (example: US social benefits adjusted in January are weighted by 3/4 to align with fiscal year October–September).

### Dataset coverage and indexation incidence
- Dataset scope and timing:
  - Final dataset contains information on the latest (as of November 2023) indexation practices for at least 115 and at most 175 economies, depending on the budget item analyzed.
- Prevalence by budget item:
  - Income tax thresholds: least prevalent globally; resulting bracket creep contributes to revenue growth.
  - Pensions: most commonly indexed among the budget items analyzed.
  - Transfers to lower-income groups and public sector wages: pensions are followed by transfers and wages in prevalence.
  - Pension indexation has become less prevalent and less generous over time in many economies.
- Cross-item inconsistencies:
  - Some economies index social assistance to prices but do not index public wages (examples in text: US, Canada, Australia).

### Empirical findings on fiscal impact of 2022 inflation surprises
- First-order effect:
  - Indexation typically results in a deterioration of the fiscal balance during periods of inflationary shocks, primarily through additional expenditures.
- Mixed overall impact:
  - Historical analyses cited show mixed outcomes:
    - Garcia-Macia (2023): short-term improvements in fiscal balance after accounting for increase in tax base and delayed catch up of non-indexed expenditures.
    - Bankowski and others (2023): recent 2022–23 surge resulted in adverse impact on euro area budget balance, with spending pressures from indexation more than offsetting revenue gains.
- Figure 12 summary (qualitative):
  - Contributions to FY2022 and FY2023 primary balances from 2022 inflation surprises vary by country and by channel (pension indexation, wage bill indexation, bracket creep). Charts presented ranges from approximately -2.4 to 0.8 percent of GDP for different economies and channels.
- Accounting exercise caveat:
  - The paper’s accounting exercise does not capture second-order effects such as higher income tax revenue from pension indexation expanding tax bases.

### Inflation persistence and indexation: statistical analyses (2018–2022)
- Weighted indexation measure:
  - Weighted Indexation is constructed as the sum of shares of wage bill, pensions, and social assistance in total current spending multiplied by indexation dummies for each item (Equation (2)); ranges between 0 and 1.
  - Sample: 58 economies (27 AEs, 23 EMs, and 8 LIDCs).
  - Average weighted indexation: 0.13.
  - Maximum weighted indexation in sample: Cyprus with 59 percent of its total current spending in a given year having some indexation (period analyzed 2018–2020).
- Hurst Exponent analysis:
  - Hurst exponent H = d + 0.5, where d is the fractional-integration parameter estimated via ARFIMA using monthly inflation data January 2018–December 2022.
  - Interpretation: H > 0.5 indicates persistence; H = 0.5 random walk; H < 0.5 anti-persistent.
  - Sample findings:
    - Among 53 economies included in Hurst calculations, highest average H observed among AEs and EMs; lowest among LIDCs.
    - Correlation: an increase in share of indexed public spending by 10 percent is correlated with an increase in Hurst exponent of 0.05.
    - Even economies with no indexation tend to have, on average, persistent inflation with the y-intercept approximately 0.6 in the regression chart.
- Autoregressive (AR) analysis (AR(12)):
  - Persistence measured as sum of AR coefficients for first 12 months using monthly inflation data 2018–2022 for 58 economies.
  - Interpretation: cumulative AR sum > 1 suggests explosive behavior (strong persistence in increasing direction).
  - Empirical correlation: a rise in share of public spending indexed to inflation by 10 percent augments the sum of AR coefficients by 0.26.
- Half-life approximation (interpretation note):
  - Using an AR(1) approximation, a 20 percent increase in share of indexed spending is associated with an increase in the half-life of inflation by around 1 month.

### Key quantitative values and methodological details
- Sample counts and averages:
  - Total economies in weighted-indexation sample: 58 (27 AEs, 23 EMs, 8 LIDCs).
  - Average weighted indexation: 0.13.
  - Maximum weighted indexation observed: 59 percent (Cyprus, 2018–2020 period).
  - Hurst exponent formula: H = d + 0.5.
- Correlations and elasticities:
  - 10 percent increase in indexed public spending → Hurst exponent +0.05.
  - 10 percent increase in indexed public spending → sum of AR coefficients +0.26.
  - 20 percent increase in indexed spending → half-life of inflation ≈ +1 month (AR(1) approximation).
- Dataset coverage as of November 2023:
  - Between 115 and 175 economies covered depending on budget item.

### Conclusions and policy messages
- Trade-offs of indexation:
  - Benefits:
    - Preserves purchasing power for targeted groups and programs; can improve economic well-being and reduce poverty over time (relevant for social assistance programs).
  - Costs/risks:
    - Can make inflation more persistent by creating inflationary inertia and raising backward-looking inflation expectations.
    - During acute inflation surprises, indexation can substantially increase budgetary spending and deteriorate fiscal balances.
    - Public wage indexation can affect private wage setting and contribute to sticky inflation.
- Policy guidance distilled from analysis:
  - When reviewing automatic or discretionary indexation, policymakers should:
    - Decide which groups and programs to protect from income erosion while avoiding policies that entrench inflation.
    - Carefully assess the impact of public wage setting during high inflation episodes on private wages and inflation persistence.
    - Consider potential effects of inflation on the tax system structure and thresholds to manage bracket creep and revenue implications.
- Suggested extensions for future research:
  - Expand the dataset over time to create a panel dataset.
  - Include additional budget items.
  - Use the dataset to analyze how indexation affects cyclical fiscal responses to inflation, long-term distributional implications, and causal links between indexation and sticky inflation.

*Italic: Source — IMF WORKING PAPERS, Inflation Indexation in Public Finances: A Global Dataset on Current Practices (content unit: wpiea2023264-print-pdf - 2022. The bracket creep elasticity is).*

### Annex I: List of Economies and Indexation

### Annex I: List of Economies and Indexation Practices

### Indexation coding and definitions
- Pension Benefits Indexation codes:
  - 0 - No Automatic Indexation
  - p - Automatic Indexation to prices
  - w - Automatic Indexation to wages
  - m - Automatic Indexation to a combination of prices, wages etc.
- Social Assistance Benefits Indexation codes:
  - 0 - No Automatic Indexation
  - p - Indexation to prices
  - m - Indexation to other variables
- Wage Indexation codes:
  - 0 - No Automatic Indexation
  - p - Automatic Indexation to prices
  - m - Automatic Indexation to other variables
- Personal Income Tax Threshold Indexation:
  - No - Ad hoc adjustment
  - Automatic adjustment - By law
  - Unclear process - De facto Regular adjustment but no law

### Selected country practices (representative excerpts from the list)
- Afghanistan: LIDC; Pension Benefits Indexation = 0; Personal Income Tax = Flat rate PIT.
- Australia: AE; Pension Benefits Indexation = p; Social Assistance = p; Wage Indexation = 0; Personal Income Tax Threshold = No.
- Austria: AE; Pension Benefits Indexation = p; Social Assistance = p; Wage Indexation = 0; Personal Income Tax Threshold = Automatic Adjustment.
- Brazil: EME; Pension Benefits Indexation = m; Social Assistance = 0; Wage Indexation = 0; Personal Income Tax Threshold = No.
- Canada: AE; Pension Benefits Indexation = p; Social Assistance = p; Wage Indexation = 0; Personal Income Tax Threshold = Automatic adjustment.
- Chile: EME; Pension Benefits Indexation = p; Social Assistance = 0; Wage Indexation = 0; Personal Income Tax Threshold = Automatic adjustment.
- China: EME; Pension Benefits Indexation = m; Social Assistance = p; Wage Indexation = No.
- France: AE; Pension Benefits Indexation = p; Social Assistance = p; Wage Indexation = 0; Personal Income Tax Threshold = Unclear process.
- Germany: AE; Pension Benefits Indexation = m; Social Assistance = m; Wage Indexation = 0; Personal Income Tax Threshold = Unclear process.
- India: EME; Pension Benefits Indexation = p; Social Assistance = 0; Wage Indexation = p; Personal Income Tax Threshold = No.
- Mexico: EME; Pension Benefits Indexation = p; Social Assistance = 0; Wage Indexation = p; Personal Income Tax Threshold = Automatic adjustment.
- United States: AE; Pension Benefits Indexation = p; Social Assistance = p; Wage Indexation = 0; Personal Income Tax Threshold = Automatic adjustment.
- Note: For cases where indexation is mentioned as “p/m”, part of the benefits is indexed to prices and some parts indexed to other variables.

(Only a representative subset of entries is shown here; the Annex contains the full list of economies with the codes as defined above.)

### Patterns and notable classifications
- Several economies report "No PIT" or "Flat rate PIT" where personal income tax thresholds are not applicable or are flat-rate systems.
- Many advanced economies (AE) show automatic indexation for pensions and social assistance (commonly coded as p, m, or w) and often have automatic or unclear processes for personal income tax threshold adjustments.
- Numerous low-income and developing countries (LIDC) record 0 (No Automatic Indexation) across pensions, social assistance, and wage indexation columns.
- The dataset records mixed practices within countries (e.g., p/m) where different components follow different indexation rules.

### Annex II: Main Data Sources (methodology notes)
- Primary sources consulted include:
  - US Social Security Administration reports (individual reports 2002–2019) for pension and social assistance benefits.
  - OECD Pension Policy Notes and Reviews and OECD resources.
  - Mutual Information System on Social Protection (MISSOC).
  - International Social Security Association (ISSA) country profiles.
  - IBFD - Tax Research Platform (used for PIT thresholds indexation dataset referenced from a prior IMF publication).
  - IMF Pay Setting Systems Dataset (2016) used as a starting point and verified with IMF country desks.
- The process involved reading individual reports and country profiles to extract indexation practices for pensions, social assistance, and wage-setting, and checking PIT threshold indexation datasets.

### Annex III: Summary Statistics for the Accounting Exercise
- Average values:
  - Inflation shock = 5.75
  - Deflator shock = 3.64
  - Wage Bill = 5.95
  - Pensions = 6.84
  - Social Spending = 2.09
- Median values:
  - Inflation shock = 4.50
  - Deflator shock = 1.84
  - Wage Bill = 5.20
  - Pensions = 6.09
  - Social Spending = 1.92
- Percentiles:
  - 25th percentile: Inflation shock = 3.65; Deflator shock = 1.27; Wage Bill = 4.03; Pensions = 4.36; Social Spending = 1.36
  - 75th percentile: Inflation shock = 5.40; Deflator shock = 3.50; Wage Bill = 7.13; Pensions = 8.34; Social Spending = 2.76
- Number of observations:
  - Inflation shock: 9
  - Deflator shock: 12
  - Wage Bill: 4
  - Pensions: 10
  - Social Spending: 5

### Annex IV: Weighted Indexation Values per Economy
- Selected weighted indexation values (Weighted Indexation 2020; 3-Year (2018–2020) Weighted Indexation Average):
  - Italy AE: 0.56 ; 0.49
  - Luxembourg AE: 0.53 ; 0.47
  - Cyprus AE: 0.52 ; 0.58
  - Belgium AE: 0.46 ; 0.42
  - Turkey EM: 0.45 ; 0.45
  - Mexico EM: 0.35 ; 0.39
  - Moldova LIDC: 0.32 ; 0.35
  - Belarus EM: 0.28 ; 0.29
  - Finland AE: 0.27 ; 0.19
  - France AE: 0.27 ; 0.28
  - Switzerland AE: 0.26 ; 0.27
  - Croatia EM: 0.24 ; 0.27
  - Slovenia AE: 0.24 ; 0.17
  - Slovak Republic AE: 0.23 ; 0.24
  - Uzbekistan EM: 0.23 ; 0.25
  - Serbia EM: 0.21 ; 0.24
  - Lithuania AE: 0.21 ; 0.15
  - Denmark AE: 0.20 ; 0.14
  - Estonia AE: 0.20 ; 0.14
  - Czech Republic AE: 0.20 ; 0.21
  - Latvia AE: 0.19 ; 0.14
  - United Kingdom AE: 0.17 ; 0.20
  - Sweden AE: 0.16 ; 0.11
  - Netherlands AE: 0.16 ; 0.11
  - United States AE: 0.13 ; 0.15
  - Mauritius EM: 0.12 ; 0.10
  - Chile EM: 0.12 ; 0.13
  - Canada AE: 0.10 ; 0.13
  - Ecuador EM: 0.09 ; 0.09
  - Israel AE: 0.09 ; 0.10
  - Korea AE: 0.09 ; 0.10
  - Hong Kong SAR AE: 0.07 ; 0.10
- Economies with zero weighted indexation in the table (examples):
  - El Salvador EM: 0 ; 0
  - Kiribati EM: 0 ; 0
  - Côte d'Ivoire LIDC: 0 ; 0
  - Bulgaria EM: 0 ; 0
  - Spain AE: 0 ; 0
  - Seychelles EM: 0 ; 0
  - Rwanda LIDC: 0 ; 0
  - Myanmar LIDC: 0 ; 0
  - Thailand EM: 0 ; 0
  - Macao, SAR AE: 0 ; 0
  - Cambodia LIDC: 0 ; 0
  - United Arab Emirates EM: 0 ; 0
  - Samoa EM: 0 ; 0
  - Ukraine EM: 0 ; 0
  - Albania EM: 0 ; 0
  - Ireland AE: 0 ; 0
  - Singapore AE: 0 ; 0
  - Mongolia LIDC: 0 ; 0
  - Senegal LIDC: 0 ; 0
  - South Africa EM: 0 ; 0
  - Armenia EM: 0 ; 0
  - Namibia EM: 0 ; 0
  - Austria AE: 0 ; 0
  - Nepal LIDC: 0 ; 0
  - Bosnia and Herzegovina EM: 0 ; 0

### Notes and clarifications from the Annexes
- The dataset records indexation practice heterogeneity across pensions, social assistance, wages, and PIT thresholds with explicit coding for price, wage, mixed, or no automatic indexation.
- The weighted indexation values provide measures for economies in 2020 and a 3-year average (2018–2020) where available; many economies report zero weighted indexation in the table.
- The number of observations used in the accounting exercise varies by variable (see Summary Statistics).

*Inflation Indexation in Public Finances: A Global Dataset on Current Practices — Annex I–IV (extracted tables and notes) — IMF Working Paper No. WP/2023/264*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023264-print-pdf.pdf_