## _wp13111 - 1. Average VAT Revenue by Income Group (1993–2012)

## Source details

**Canonical URL:** [_wp13111 - 1. Average VAT Revenue by Income Group (1993–2012)](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13111.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13111.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13111.pdf.json)

---

### I. Introduction — purpose and scope
- Purpose:
  - Develop descriptive tools to understand developments and differences in value added tax (VAT) performance.
  - Tools intended to identify problems and opportunities (anatomy/diagnosis), not to prescribe specific policy remedies.
- Focus of the decomposition:
  - VAT revenue decomposed into elements tied to the standard rate, the average propensity to consume, and C-efficiency.
  - C-efficiency further decomposed into a “policy gap” (rate differentiation and exemptions) and a “compliance” gap (implementation weakness).
- Empirical coverage:
  - Broad decomposition applied to the universe of VATs over 1993–2012.
  - Narrower, more speculative coverage used for policy vs. compliance gap analysis for subsets of countries (notably EU countries).

### II. Key empirical observations on VAT revenue trends
- Sample and period:
  - Twenty years: 1993–2012.
  - Total VATs in place by end-2012: 150.
- Changes in VAT revenues (VAT revenue measured in percent of GDP), averaged by income group:
  - High income group: increase from a little under 7 percent to a little over 7 percent (described as only a modest increase).
  - Upper middle income: increase is marked; these countries now raise about as much from the VAT as do high income countries.
  - Low income: VAT revenue has about doubled since the mid-1990s.
- Regional patterns:
  - Western Hemisphere: VAT revenues have increased substantially.
  - Middle East and Central Asia: VAT revenues have increased substantially over the last ten years or so.
  - Sub-Saharan Africa: VAT revenues have increased, though less steadily.
- VAT rate adjustments around the 2008 crisis:
  - Two years before the crisis: only one EU country increased its standard rate.
  - Two years after the crisis: 13 (of the 27) EU countries increased their standard rate.
- Broad conclusion from decompositions:
  - Changes in VAT revenues over the sample period have been driven much less by changes in the standard rate than by changes in C-efficiency.

### III. Conceptual decomposition and focus on C-efficiency
- Decomposition components:
  - Standard VAT rate (θ0).
  - Average propensity to consume (C/Y).
  - C-efficiency: departure from a perfectly enforced tax at a uniform rate on all consumption.
- C-efficiency decomposition:
  - Policy gap (P): impact of rate differentiation and exemptions (design).
  - Compliance gap (Γ): shortfall from imperfect implementation (enforcement/administration).
- Empirical findings (subset analysis, EU countries):
  - Policy gaps are in almost all cases far larger than compliance gaps.
  - Policy gaps reflect both rate differentiation and exemptions to differing degrees across countries.

### IV. Analytical role and limitations of the tools
- Role:
  - Identify whether revenue gains or losses originate from rates, structure (policy), or compliance (implementation).
  - Highlight the growing importance of reduced rates and non-compliance as standard rates rise.
- Limitations:
  - Essentially descriptive; do not specify precise policy responses.
  - Detailed decomposition (policy vs compliance gaps) requires narrower coverage and is more speculative.

### V. Mechanics of the decomposition and quantitative impressions
- VAT revenue V (percent of GDP Y) proximate drivers:
  - Standard rate (θ0).
  - C-efficiency (VAT revenue divided by the product of the standard rate and consumption).
  - Share of consumption in GDP (C/Y).
- Proportional-change translation:
  - Proportional difference in VAT revenues ≈ sum of proportionate differences in the three components above.
- Two empirical impressions:
  - Changes in C-efficiency have been the most powerful immediate driver of changes in VAT revenues.
  - Changes in the standard rate directly account for a relatively small part of VAT revenue changes; changes in average propensity to consume account for even less (with exceptions around the 2008 crisis in high income countries).
- Example (high income countries, 1996–2000):
  - VAT revenue increased on average by about 3 percent, composed of:
    - 1 percent increase attributable to a higher standard rate,
    - 2.6 percent increase from greater C-efficiency,
    - 0.6 percent reduction from a fall in private consumption relative to GDP.

### VI. Understanding C-efficiency: concept, measurement, and interpretation
- Definition and intuition:
  - C-efficiency compares actual VAT revenue with revenue that would be raised if the standard rate applied uniformly to all consumption and were perfectly enforced.
  - Numerical intuition example: if C-efficiency is 60 percent and standard rate is 10 percent, extending the standard rate to all consumption would increase revenue by two-thirds and could alternatively be achieved by a uniform rate of 6 percent.
- Conceptual limitations:
  - No deep welfare basis: increasing C-efficiency toward 100 percent does not necessarily imply a better VAT; policy choices can raise C-efficiency while undermining VAT logic (e.g., denying refunds to exporters, exempting intermediates).
  - C-efficiency changes combine effects on deadweight loss and welfare; Appendix 2 formalizes that C-inefficiency is approximately the sum of a reduction in deadweight loss and an associated welfare loss when moving toward a uniform-rate system.
  - The single-rate benchmark is pragmatic but its appropriateness varies by country.
- Measurement issues and nuances:
  - Treatment of purchases by non-residents can materially affect small tourism-intensive economies.
  - National accounts consumption may diverge from the VAT-relevant consumption base (owner-occupied housing imputation example).
  - Inclusion of government final consumption (valued by cost of production) in denominators tends to over-estimate the potential VAT base.
  - Mandatory exemptions (e.g., in the EU) create interpretational choices; for simplicity, mandatory exemptions are not excluded in these analytics.

### VII. Formalization and the gaps framework
- Extended VAT revenue expression uses:
  - Effective rates on taxed consumption items (captures production-chain collection and unrecovered VAT).
  - True consumption of items included in the potential base (including government consumption).
- C-efficiency = policy gap (P) × compliance gap (Γ), where:
  - Policy gap (P): how design (rate differentiation and exemptions) reduces revenue relative to a single-rate VAT applied to all consumption (assuming full compliance).
  - Compliance gap (Γ): shortfall from imperfect implementation (zero if implementation is perfect).
- Policy gap decomposition:
  - Rate gap (r): impact of statutory rate differentiation (non-negative if consumption-weighted average rate ≤ standard rate).
  - Exemption gap (x): captures effects of exemptions, including unrecovered VAT on intermediates (tends to reduce measured policy gap) and inclusion of public-good-type consumption in denominators (tends to increase measured policy gap). The sign of x is ambiguous.
- Asymmetry note:
  - Policy gap computed assuming perfect compliance (uses true consumption).
  - Compliance gap computed relative to actual policy in place (uses effective rates). This asymmetry affects how changes show up in each gap.

### VIII. Measuring the compliance gap (“VAT gaps”)
- Two broad approaches:
  - Top-down: estimate theoretical liability via consumption disaggregation (household surveys plus national accounts adjustments for unrecoverable input VAT and small traders). Widely used and feasible with routine data.
  - Bottom-up: aggregate operationally detected unpaid liabilities from audits and enforcement activity; requires richer administrative data.
- Practical measurement issues:
  - Cash vs. accrual receipts: cash collections include past liabilities and amounts destined for refunds; accrual measures avoid some timing issues but can hide uncollected accrued revenues.
  - Treatment of avoidance and definitional choices affects what the compliance gap captures and how bases in denominators are adjusted.
  - Decomposition of compliance gap can distinguish on-time collection from subsequently collected amounts to gauge administrative performance.

### IX. Illustration: EU/OECD example (2006 data)
- Combining OECD (2012) C-efficiency estimates for 2006 and Reckon (2009) compliance gaps for 2006 yields policy gaps as residuals; two lessons:
  - Policy gap > Compliance gap in almost all reported countries (Greece an exception).
  - Policy gaps vary widely despite common EU rules.
- Selected country entries (values preserved exactly as in source):
  - Austria: C-efficiency 59, Compliance gap 14, Policy gap 31, Rate differentiation 18 (23), Exemptions 17 (11)
  - Belgium: C-efficiency 52, Compliance gap 11, Policy gap 42, Rate differentiation 22 (30), Exemptions 25 (17)
  - Denmark: C-efficiency 64, Compliance gap 4, Policy gap 33, Rate differentiation 0 (10), Exemptions 33 (26)
  - Finland: C-efficiency 61, Compliance gap 5, Policy gap 36, Rate differentiation 12 (33), Exemptions 27 (17)
  - France: C-efficiency 51, Compliance gap 7, Policy gap 45, Rate differentiation 26 (30), Exemptions 26 (22)
  - Germany: C-efficiency 57, Compliance gap 10, Policy gap 37, Rate differentiation 12 (18), Exemptions 28 (22)
  - Greece: C-efficiency 47, Compliance gap 30, Policy gap 33, Rate differentiation 30 (26), Exemptions 4 (9)
  - Ireland: C-efficiency 66, Compliance gap 2, Policy gap 33, Rate differentiation 24 (38), Exemptions 12 (-0.09)
  - Italy: C-efficiency 43, Compliance gap 22, Policy gap 45, Rate differentiation 26 (30), Exemptions 26 (21)
  - Luxembourg: C-efficiency 87, Compliance gap 1, Policy gap 12, Rate differentiation 30 (34), Exemptions -26 (-32)
  - Netherlands: C-efficiency 60, Compliance gap 3, Policy gap 38, Rate differentiation 24 (31), Exemptions 19 (11)
  - Portugal: C-efficiency 53, Compliance gap 4, Policy gap 45, Rate differentiation 25 (36), Exemptions 27 (14)
  - Spain: C-efficiency 57, Compliance gap 2, Policy gap 29, Rate differentiation 33 (31), Exemptions -6 (-3)
  - Sweden: C-efficiency 56, Compliance gap 3, Policy gap 42, Rate differentiation 19 (22), Exemptions 29 (26)
  - United Kingdom: C-efficiency 48, Compliance gap 17, Policy gap 42, Rate differentiation 21 (31), Exemptions 27 (17)
- Notes on the decomposition:
  - Two approaches produce bracketed alternative rate/exemption numbers (Mathis (2004) and Borselli, Chiri, and Romagno (2012)); the approach using 2011 weighted averages gives systematically higher rate gaps and lower exemption gaps.
  - Exemption gaps are mostly positive and often large; negative exemption gaps (e.g., Luxembourg) reflect special national features.

### X. Key analytical conclusions and policy implications
- Main findings:
  - C-efficiency is the principal proximate driver of changes in VAT revenue.
  - Policy gap (design) and compliance gap (implementation) are distinct; both matter and their relative importance differs across country groups.
  - In the EU/OECD sample, policy gaps are larger than compliance gaps; implementation gaps may be more important in many developing countries.
- Interpretational caveats:
  - C-efficiency is a useful diagnostic for potential base-broadening gains but has no simple welfare interpretation.
  - Measurement choices (treatment of public consumption, non-resident purchases, timing of receipts, inclusion of avoidance) materially affect estimates.
- Practical recommendations implied:
  - Use the gaps framework to structure diagnostics: estimate C-efficiency, then infer policy or compliance gaps using available auxiliary information.
  - Prefer time-series tracking within countries to monitor changes, given cross-country comparability issues.
  - Undertake further empirical work to identify determinants of VAT performance and complement anatomy with deeper analysis of VAT chains and administrative functioning.

### XI. Appendix 1 — Data (sources, classifications, and summary statistics)
- Data sources:
  - VAT revenue: IMF Tax Policy Revenue Mobilization Database.
  - Standard VAT rates: IMF Tax Policy VAT Rates Database and the International Bureau of Fiscal Documentation.
  - GDP (NGDP) and Total final consumption (NC): World Economic Outlook, measured in local currency units at current market prices.
- Income group classification (World Bank Income Classification, per capita GNI at July 1st, 2012):
  - 26 Low Income countries: GNI per capita at $1025 or less.
  - 40 Lower Middle Income countries: GNI per capita from $1026 to $4035.
  - 46 Upper Middle Income countries: GNI per capita from $4036 to $12475.
  - 38 High Income countries: GNI per capita of $12476 or more.
- Regional classification: IMF area departments — Africa, Asia and Pacific, Europe, Middle Eastern and Central Asian (including North Africa), and Western Hemisphere.
- C-efficiency calculation:
  - C-efficiency = VAT revenue ÷ (standard rate × final consumption expenditure less VAT revenue).
  - Availability of C-efficiency observations constrained by availability of VAT revenue data.

- Appendix Table 1 summary statistics (1993–2012):
  - VAT revenue (in percent of GDP)
    - Observations: 1192
    - Mean: 6.04
    - Std. Dev.: 2.52
    - Min.: 0.02
    - Max.: 16.02
  - Total final consumption
    - Observations: 2870
    - Mean: 82.89
    - Std. Dev.: 17.22
    - Min.: 10.37
    - Max.: 475.54
  - Standard rate
    - Observations: 2514
    - Mean: 15.88
    - Std. Dev.: 4.88
    - Min.: 1.50
    - Max.: 35.00
  - C-efficiency
    - Observations: 1136
    - Mean: 0.52
    - Std. Dev.: 0.19
    - Min.: 0.002
    - Max.: 1.23
- Number of countries with a VAT (as of 2012: 150):
  - 26 low income
  - 40 lower middle income
  - 46 upper middle income
  - 38 high income

*Source: _wp13111 - 1. Average VAT Revenue by Income Group (1993–2012) and Appendix 1.*

### 1. Average VAT Revenue by Income Group .............................................................................4

### _wp13111 - 1. Average VAT Revenue by Income Group .............................................................................4

### I. Introduction — purpose and scope
- The paper develops descriptive tools for understanding developments and differences in the performance of the value added tax (VAT).
- Tools are intended to identify problems and opportunities (anatomy/diagnosis), not to prescribe specific policy remedies.
- Focus areas:
  - Decomposition of VAT revenue into elements tied to standard rate, average propensity to consume, and C-efficiency.
  - Decomposition of C-efficiency into a “policy gap” (rate differentiation and exemptions) and a “compliance” gap (implementation weakness).
- Empirical coverage:
  - Broad decomposition applied to the universe of VATs over the last twenty years (1993–2012).
  - More speculative, narrower coverage when analyzing policy and compliance gaps for subsets of countries (notably EU countries).

### II. Key empirical observations on VAT revenue trends
- Sample and period:
  - Twenty years: 1993–2012.
  - Total VATs in place by the end of the sample period: 150.
- Changes in VAT revenues (VAT revenue measured in percent of GDP), averaged by income group:
  - High income group: increase from a little under 7 percent to a little over 7 percent (described as only a modest increase).
  - Upper middle income: increase is marked; these countries now raise about as much from the VAT as do high income countries.
  - Low income: VAT revenue has about doubled since the mid-1990s.
- Regional patterns (summary):
  - Western Hemisphere (the Americas and Caribbean): VAT revenues have increased substantially.
  - Middle East and Central Asia: VAT revenues have increased substantially over the last ten years or so.
  - Sub-Saharan Africa: VAT revenues have increased, though less steadily.
- VAT rate adjustments around the 2008 crisis:
  - In the two years before the crisis, only one European Union (EU) country increased its standard rate.
  - In the two years after the crisis, 13 (of the 27) EU countries increased their standard rate.
- Broad conclusion from decompositions:
  - Changes in VAT revenues over the sample period have been driven much less by changes in the standard rate than by changes in C-efficiency.

### III. Conceptual decomposition and focus on C-efficiency
- Decomposition for VAT revenue links revenue to:
  - The standard VAT rate.
  - The average propensity to consume.
  - C-efficiency: indicator of departure from a perfectly enforced tax at a uniform rate on all consumption.
- C-efficiency further decomposed into:
  - Policy gap: effects of rate differentiation and exemptions.
  - Compliance gap: imperfect implementation and enforcement.
- Empirical findings (qualitative and from subset analysis):
  - For a subset of EU countries, policy gaps are in almost all cases far larger than compliance gaps.
  - Policy gaps reflect, to differing degrees across countries, the impact of both rate differentiation and exemptions.

### IV. Analytical role and limitations of the tools
- Role:
  - Tools identify where revenue gains or losses originate (rates vs. structure vs. compliance).
  - Useful for highlighting the growing importance of reduced rates and non-compliance as standard rates rise.
- Limitations:
  - Essentially descriptive; do not specify precise policy responses.
  - More detailed decomposition (policy vs compliance gaps) requires narrower country coverage and becomes more speculative.

*Source: _wp13111 - 1. Average VAT Revenue by Income Group (1993–2012), extracted from the provided PDF content.*

### Appendix 1 describes the data and sources used in this paper. All averages, of course, are over only countries

### _wp13111 - Appendix 1 describes the data and sources used in this paper. All averages, of course, are over only countries with a VAT at that time.

### Changes in VAT revenues and rates — empirical patterns
- Average standard VAT rates have trended downward since the mid-1990s in most groups, with the notable exception of recent increases in high income countries of Europe.
- In the lowest income countries, the average standard rate has fallen by around 2 percentage points over the last twenty years.
- Trend decreases in average standard rates are reported for sub-Saharan Africa and Middle East and Central Asia, while average VAT revenues in sub-Saharan Africa and low income countries have increased even as standard rates fell.
- Example decomposition (high income countries, 1996–2000): VAT revenue increased on average by about 3 percent, composed of:
  - 1 percent increase attributable to a higher standard rate,
  - 2.6 percent increase from greater C-efficiency,
  - 0.6 percent reduction from a fall in private consumption relative to GDP.

### Decomposing VAT revenue (mechanics and main components)
- VAT revenue V as percent of GDP Y is decomposed into three proximate drivers:
  - the standard rate (θ0),
  - C-efficiency (ratio of VAT revenue to the product of the standard rate and consumption),
  - the share of consumption in GDP (C/Y).
- Translating into proportional changes, the proportional difference in VAT revenues is the sum of proportionate differences in the three components (standard rate, C-efficiency, consumption share).
- Two empirical impressions:
  - Changes in C-efficiency have been the most powerful immediate driver of changes in VAT revenues.
  - Changes in the standard rate directly account for a relatively small part of VAT revenue changes; changes in the average propensity to consume account for even less (with exceptions around the 2008 crisis in high income countries).

### Understanding C-efficiency: concept, measurement, and interpretation
- Definition and intuition:
  - C-efficiency compares actual VAT revenue with revenue that would be raised if the standard rate applied uniformly to all consumption and were perfectly enforced.
  - Example numerical intuition: if C-efficiency is 60 percent and standard rate is 10 percent, extending the standard rate effectively to all consumption would increase revenue by two-thirds and could alternatively be achieved by a uniform rate of 6 percent.
- Conceptual limitations:
  - No deep welfare basis: increasing C-efficiency toward 100 percent does not necessarily imply a better VAT; it can be increased by policy choices (e.g., denying refunds to exporters, introducing exemptions for intermediate goods) that may undermine VAT logic.
  - C-efficiency changes combine effects on deadweight loss and welfare; Appendix 2 (referenced) formalizes that C-inefficiency is approximately the sum of a reduction in deadweight loss and an associated welfare loss when moving toward a uniform-rate system.
  - Uniform taxation as a reference point is pragmatic: the single-rate benchmark is common among practitioners, but its appropriateness may vary across country contexts.
- Measurement issues and nuances:
  - Treatment of purchases by non-residents: national practice generally excludes non-resident purchases from C but often does not exclude VAT revenue from sales to non-residents; this can materially affect small tourism-intensive economies.
  - National accounts consumption may diverge from the VAT-relevant consumption base (for example, owner-occupied housing is imputed in accounts but rarely practically taxable).
  - Public sector consumption valuation: national accounts treat government final consumption as consumption proxied by cost of production (including labor and capital costs). Using final consumption expenditure (aggregated over government, households, and non-profits) in denominators tends to over-estimate potential VAT base by including non-commodity costs of producing public goods.
  - Mandatory exemptions (e.g., in the EU) pose interpretational choices; for simplicity, mandatory exemptions are not excluded in the analytics here.

### Formalization and gaps framework
- Extended expression for VAT revenue uses:
  - effective rates on taxed consumption items (captures revenue collected through production chains and unrecovered VAT),
  - true consumption of items included in the potential base (including government consumption items discussed above).
- C-efficiency can be written as the product of two multiplicative components:
  - a policy gap (P) — how much design (rate differentiation and exemptions) reduces revenue relative to a single-rate VAT applied to all consumption (assuming full compliance);
  - a compliance gap (Γ) — the shortfall from imperfect implementation (zero if implementation is perfect).
- Policy gap decomposition:
  - Rate gap (r): impact of statutory rate differentiation (non-negative if consumption-weighted average rate ≤ standard rate).
  - Exemption gap (x): captures effects of exemptions, including unrecovered VAT on intermediates (which tends to reduce the measured policy gap) and inclusion of public-good-type consumption in denominators (which tends to increase the measured policy gap). The sign of x is ambiguous in principle.
- Asymmetry: policy gap is computed assuming perfect compliance (uses true consumption), while compliance gap is computed relative to the actual policy in place (uses effective rates). This asymmetry affects how policy or implementation changes will show up in each gap.

### Measuring the compliance gap (“VAT gaps”)
- Two broad approaches:
  - Top-down: estimates theoretical liability via consumption disaggregation (household surveys plus national accounts adjustments for unrecoverable input VAT and small traders). Widely used and feasible with routine data.
  - Bottom-up: aggregates operationally detected unpaid liabilities from audits and enforcement activity; provides more direct insight into remedial measures but requires richer administrative data.
- Practical issues in defining and measuring VAT receipts:
  - Cash vs. accrual: cash collections include past liabilities and may include amounts destined for refunds; accrual measures avoid some timing issues but can hide uncollected accrued revenues.
  - Treatment of avoidance and definitional choices (e.g., including avoidance in the compliance gap) affect what the compliance gap captures and whether associated bases in denominators need adjustment.
- Decomposition of compliance gap can be informative: for example, distinguish revenue collected on time from that collected subsequently to gauge administrative performance.

### Illustration (EU/OECD example, 2006 data)
- Combining OECD (2012) C-efficiency estimates for 2006 and Reckon (2009) compliance gaps for 2006 yields policy gaps as residuals; two lessons:
  - In all reported countries the policy gap is larger than the compliance gap — in most cases much larger (Greece an exception).
  - Policy gaps vary widely across countries despite common EU rules.
- Table 1 highlights (selected entries preserved exactly as in source):
  - Austria: C-efficiency 59, Compliance gap 14, Policy gap 31, Rate differentiation 18 (23), Exemptions 17 (11)
  - Belgium: C-efficiency 52, Compliance gap 11, Policy gap 42, Rate differentiation 22 (30), Exemptions 25 (17)
  - Denmark: C-efficiency 64, Compliance gap 4, Policy gap 33, Rate differentiation 0 (10), Exemptions 33 (26)
  - Finland: C-efficiency 61, Compliance gap 5, Policy gap 36, Rate differentiation 12 (33), Exemptions 27 (17)
  - France: C-efficiency 51, Compliance gap 7, Policy gap 45, Rate differentiation 26 (30), Exemptions 26 (22)
  - Germany: C-efficiency 57, Compliance gap 10, Policy gap 37, Rate differentiation 12 (18), Exemptions 28 (22)
  - Greece: C-efficiency 47, Compliance gap 30, Policy gap 33, Rate differentiation 30 (26), Exemptions 4 (9)
  - Ireland: C-efficiency 66, Compliance gap 2, Policy gap 33, Rate differentiation 24 (38), Exemptions 12 (-0.09)
  - Italy: C-efficiency 43, Compliance gap 22, Policy gap 45, Rate differentiation 26 (30), Exemptions 26 (21)
  - Luxembourg: C-efficiency 87, Compliance gap 1, Policy gap 12, Rate differentiation 30 (34), Exemptions -26 (-32)
  - Netherlands: C-efficiency 60, Compliance gap 3, Policy gap 38, Rate differentiation 24 (31), Exemptions 19 (11)
  - Portugal: C-efficiency 53, Compliance gap 4, Policy gap 45, Rate differentiation 25 (36), Exemptions 27 (14)
  - Spain: C-efficiency 57, Compliance gap 2, Policy gap 29, Rate differentiation 33 (31), Exemptions -6 (-3)
  - Sweden: C-efficiency 56, Compliance gap 3, Policy gap 42, Rate differentiation 19 (22), Exemptions 29 (26)
  - United Kingdom: C-efficiency 48, Compliance gap 17, Policy gap 42, Rate differentiation 21 (31), Exemptions 27 (17)
- Notes on the illustration:
  - Policy gaps exceed compliance gaps in almost all listed countries, suggesting design (rate differentiation and exemptions) explains much C-inefficiency in these cases.
  - Decomposition into rate and exemption gaps uses two approaches (Mathis (2004) and Borselli, Chiri, and Romagno (2012) data) producing bracketed alternative numbers; the approach using 2011 weighted averages gives systematically higher rate gaps and lower exemption gaps.
  - Exemption gaps are mostly positive and often large; some negative exemption gaps (e.g., Luxembourg) reflect special national features.

### Key analytical conclusions and policy implications
- C-efficiency is the principal proximate driver of changes in VAT revenue; understanding changes in C-efficiency is essential for understanding VAT revenue developments.
- The policy gap (design) and the compliance gap (implementation) are distinct and both matter; their relative importance differs across country groups (policy gaps larger in the EU/OECD sample; implementation gaps likely more important in many developing countries).
- Care is required in interpreting C-efficiency:
  - It is a useful diagnostic for potential base-broadening gains but has no simple welfare interpretation.
  - Measurement choices (treatment of public consumption, non-resident purchases, timing of receipts, inclusion of avoidance) materially affect estimates.
- Practical recommendations implied by the analysis:
  - Use the gaps framework to structure diagnostics: estimate C-efficiency, then infer policy or compliance gaps using available auxiliary information.
  - Prefer time-series tracking within countries (to monitor changes) given cross-country comparability issues arising from methodological differences.
  - Further empirical work is needed to identify determinants of VAT performance and to complement the superficial anatomy with deeper analysis of VAT chains and administrative functioning.

*Appendix 1 of the paper (content unit: _wp13111).*

### Appendix 1. Data

### Appendix 1. Data

### Sources and classifications
- VAT revenue is from the IMF Tax Policy Revenue Mobilization Database.
- Standard rates of VAT are from IMF Tax Policy VAT Rates Database and the International Bureau of Fiscal Documentation.
- GDP is from the World Economic Outlook, series name NGDP, measured in local currency units at current market prices.
- Total final consumption expenditure is from the World Economic Outlook, series name NC, measured in local currency units at current market prices.
- Countries are placed in income groups by the World Bank Income Classification, according to their per capita GNI at July 1st, 2012:
  - The 26 Low Income countries have GNI per capita at $1025 or less.
  - 40 Lower Middle Income countries have GNI per capita from $1026 to $4035.
  - The 46 Upper Middle Income countries have GNI per capita from $4036 to $12475.
  - The 38 High Income countries have GNI per capita of $12476 or more.
- Countries are classified into regions as in the IMF area departments: Africa, Asia and Pacific, Europe, Middle Eastern and Central Asian (including North Africa), and Western Hemisphere. Details are in the Tax Policy VAT Rates database.
- C-efficiency is calculated by dividing VAT revenue by the product of the standard rate and final consumption expenditure less VAT revenue. Generally, the availability of observations on C-efficiency data is constrained by the availability of VAT revenue data.

### Description — summary statistics (Appendix Table 1, 1993–2012)
- VAT revenue (in percent of GDP)
  - Observations: 1192
  - Mean: 6.04
  - Std. Dev.: 2.52
  - Min.: 0.02
  - Max.: 16.02
- Total final consumption
  - Observations: 2870
  - Mean: 82.89
  - Std. Dev.: 17.22
  - Min.: 10.37
  - Max.: 475.54
- Standard rate
  - Observations: 2514
  - Mean: 15.88
  - Std. Dev.: 4.88
  - Min.: 1.50
  - Max.: 35.00
- C-efficiency
  - Observations: 1136
  - Mean: 0.52
  - Std. Dev.: 0.19
  - Min.: 0.002
  - Max.: 1.23
- Note: VAT revenue is reported in percent of GDP.

### Number of countries with a VAT (Appendix Figure 1)
- As of 2012, 150 countries have a VAT:
  - 26 low income
  - 40 lower middle income
  - 46 upper middle income
  - 38 high income

### C-efficiency and data constraints
- C-efficiency formula (as described): VAT revenue divided by (standard rate × final consumption expenditure less VAT revenue).
- Availability of C-efficiency observations is generally constrained by the availability of VAT revenue data.

*Source: _wp13111 - Appendix 1. Data*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13111.pdf_