## _wp1071 - Section VI presents the results for VAT C-efficiency, and section VII for PIT and SSC

## Source details

**Canonical URL:** [_wp1071 - Section VI presents the results for VAT C-efficiency, and section VII for PIT and SSC](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1071.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1071.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1071.pdf.json)

---

### Stylized facts
- Tax collections and the growth rates of their bases are correlated over the business cycle; changes in tax collections tend to be more pronounced than changes in bases.
- Baltic countries example:
  - Real GDP contracted by 21¾ percent in Latvia and 17 percent in Estonia over 2008–09, and by 15 percent in Lithuania in 2009; prior expansions over 2005–07: Latvia 36½ percent, Estonia 29 percent, Lithuania 27¾ percent.
  - VAT growth was lower than private consumption growth during the 2008–09 contractions and higher during 2005–07 expansions.
- Cross-country correlation:
  - For EU countries, 1995–2008: output gap coefficient 0.9, p-value less than 0.01; 405 observations.

### Data
- Datasets:
  - Annual 1995–2008 for 32 EU countries.
  - Annual 1995–2008 for 84 advanced and developing economies.
  - Quarterly Q1 1999–Q1 2009 for 37 advanced and developing economies.
- Data sources:
  - VAT, PIT, SSC: Eurostat general government surveys (EU), IMF GFS and IMF staff country data (non-EU).
  - VAT standard rates: IMF Fiscal Affairs Department database.
  - PIT and SSC rates: Eurostat.
  - GDP, private consumption, wages and salaries: IMF WEO.
  - Control of tax evasion: IMD World Competitiveness Online.
  - Law and order: International Country Risk Guide.
  - Share of necessity goods in consumption: Eurostat.

### Methodology
- Core fixed-effects panel regression (equation (1)):
  - taxefficiency_it = α_i + β gap_it + ε_it
  - taxefficiency is the tax revenue efficiency ratio; gap is the output gap.
- Extensions and controls:
  - Interaction with a “bad times” dummy (bad times when actual real GDP growth < potential) to test asymmetry.
  - Interaction with an “advanced economy” dummy to test heterogeneity.
  - Additional covariates X_it: share of necessity goods in consumption, control of tax evasion, law and order, (lagged) tax burden.
- VAT revenue elasticity specification (equation (3)):
  - log(tax revenue)_it = α_i + β1 log(tax base)_it + β2 taxrate_it + β3 gap_it + ε_it
  - taxrate is the standard VAT rate.

### Estimation results — VAT C-efficiency (quarterly and annual)
- Quarterly results (Table 1):
  - Full sample (column 1): Gap coefficient = 0.286*** (0.061).
    - Interpretation: A one percentage point increase in the output gap in one quarter → increase in VAT C-efficiency by 0.29 percentage point.
  - Bad times (column 2): Gap = 0.332*** (0.085).
  - Good times (column 3): Gap = 0.184* (0.100).
  - Interacted model (column 4):
    - Gap = 0.190** (0.096).
    - Bad times dummy = -1.874*** (0.555).
    - Bad times*Gap = 0.145 (0.129) — not statistically significant.
  - Observations: Full sample 1,117; Bad times 454; Good times 661.
  - Number of countries: 37 (full), 36 (bad times).
- Annual results (Table 2):
  - EU countries (column 5): Gap = 0.904*** (0.229).
  - Broader set (column 7): Gap = 1.166*** (0.193).
  - Interacted models (columns 6 and 8): no significant difference in slope between good and bad times.
  - Observations: EU sample 405; broader set 674.
  - Number of countries: 32 (EU), 84 (broader set).
  - Illustrative implication: If output declines to 5 percent below potential → VAT C-efficiency falls by almost 6 percentage points (from 62 to 56 percent) based on average economy example.
- Advanced vs developing economies (Table 3):
  - Full sample (column 7 repeated): Gap = 1.166*** (0.193).
  - Advanced (column 9): Gap = 1.019*** (0.274).
  - Developing (column 10): Gap = 1.274*** (0.277).
  - Interacted (column 11): Gap = 1.274*** (0.269); Advanced*Gap = -0.254 (0.389) — interaction not significant.
  - Observations: Full 674; Advanced 374; Developing 300.
  - Number of countries: Full 84; Advanced 31; Developing 53.
  - Conclusion: Differences between advanced and developing economies in gap coefficient are not statistically significant.

### Channels: shifts in consumption patterns and tax evasion (Table 4)
- Share of necessity goods (proxy: share of food and non-alcoholic beverages in total consumption) — 25-country sample:
  - Column 12 (baseline for these 25): Gap = 1.010*** (0.288).
  - Column 13 (add share of necessity goods): Gap = 0.562** (0.269); Share of necessity goods = -2.654*** (0.400).
    - Interpretation: As the share of necessity goods rises, VAT C-efficiency declines; controlling for consumption shifts reduces the output gap coefficient from 1.01 to 0.56.
- Control of tax evasion (IMD indicator):
  - Column 14: Gap = 0.961*** (0.281); Control of tax evasion = 1.810*** (0.518).
  - Column 15 (both controls): Gap = 0.565** (0.268); Share of necessity goods = -2.491*** (0.391); Control of tax evasion = 0.906** (0.411).
- Law & order and tax burden controls (column 16):
  - Gap = 0.629** (0.297).
  - Law & order = 3.114*** (0.917).
  - Tax burden (lag) = -0.8305** (0.3959).
- Observations and fit:
  - Observations: 251 for each specification in Table 4; Number of countries: 25.
  - R-squared range in Table 4: 0.077 to 0.329.
- Interpretation:
  - Shifts toward necessity goods and weaker control of tax evasion during downturns are key channels reducing VAT C-efficiency.
  - Stronger legal/institutional framework (law & order) is positively correlated with VAT C-efficiency.
  - Higher (lagged) tax burden is correlated with lower VAT C-efficiency.

### Estimation results — PIT and SSC efficiency (Table 5)
- PIT efficiency (column 17):
  - Gap = 1.428* (0.822).
  - Constant = 56.20*** (0.830).
  - Observations = 157; Number of countries = 13.
  - R-squared = 0.038.
- SSC efficiency (column 18):
  - Gap = 1.133** (0.572).
  - Constant = 74.21*** (0.627).
  - Observations = 258; Number of countries = 20.
  - R-squared = 0.019.
- Notes:
  - PIT and SSC efficiency both positively correlated with the output gap.
  - Interacted models: SSC gap remains significant; no statistically significant difference in level and slope between good and bad times for SSC; for PIT, no significant correlation in interacted model.

### Estimation results — VAT revenue elasticity (Table 6 and Table 7)
- VAT revenue elasticity with respect to private consumption (Table 6):
  - Full sample (column 19): Consumption (logs) = 1.073*** (0.0264).
  - Advanced (column 20): Consumption = 1.103*** (0.0330).
  - Developing (column 21): Consumption = 1.041*** (0.0361).
- VAT rate coefficients:
  - Full sample: 0.0288*** (0.00806).
  - Advanced: 0.0438*** (0.00780).
  - Developing: -0.0334** (0.0157).
- Gap coefficients (Table 6):
  - Full sample: Gap = 0.0175*** (0.00337).
  - Advanced: Gap = 0.0112*** (0.00410).
  - Developing: Gap = 0.0216*** (0.00505).
- Observations and fit:
  - Observations: Full 674; Advanced 374; Developing 300.
  - R-squared: Full 0.862; Advanced 0.914; Developing 0.838.
- Interpretation:
  - Consumption elasticity ≈ 1.1 (full sample): a one percent increase in consumption → 1.1 percent increase in VAT revenue.
  - Gap coefficient interpretation: As dependent variable is in logs and gap not, coefficient multiplied by 100; a one percentage point increase in the output gap → 1.75 percent increase in VAT revenue over a year (consistent with the VAT C-efficiency finding).
- Illustrative summary (Table 7):
  - Change in output gap 1 percentage point:
    - Change in VAT revenues (All countries) = 1.75 percent; Advanced = 1.12 percent; Developing = 2.16 percent.
    - Change in VAT C-efficiency (All) = 1.17 percentage point; Advanced = 1.02 pp; Developing = 1.27 pp.
  - Change in output gap 5 percentage points:
    - Change in VAT revenues (All) = 8.75 percent; Advanced = 5.6 percent; Developing = 10.8 percent.
    - Change in VAT C-efficiency (All) = 5.83 percentage points; Advanced = 5.10 pp; Developing = 6.37 pp.

### Policy implications and conclusions
- Main quantitative estimates:
  - A one percentage point increase in the output gap corresponds to:
    - 1¼ percentage point increase in VAT C-efficiency.
    - 1.75 percent increase in VAT revenue collections.
  - Results are consistent across quarterly and annual data, across advanced and developing economies, and in both good and bad times (bad times defined by growth below potential).
- Key channels to consider in forecasting and policy:
  - Shifts in consumption toward necessity goods (often taxed at lower or zero VAT rates) during downturns reduce VAT C-efficiency.
  - Changes in tax evasion (worse during downturns) reduce VAT C-efficiency.
  - Strong institutional underpinnings (law & order, tax administration control) improve VAT C-efficiency.
  - Higher (lagged) overall tax burden is associated with lower VAT C-efficiency.
- Policy recommendation:
  - During major booms and sharp downturns, policymakers should incorporate the effects of the economic cycle on tax revenue efficiency — beyond long-run revenue elasticities — to improve revenue forecasting and the calculation of structural fiscal balances, thereby informing counter-cyclical fiscal policy effectiveness.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1071.pdf*

### Section VI presents the results for VAT C-efficiency, and section VII for PIT and SSC

### _wp1071 - Section VI presents the results for VAT C-efficiency, and section VII for PIT and SSC

### Stylized facts
- Tax collections and the growth rates of their bases are correlated over the business cycle; changes in tax collections tend to be more pronounced than changes in bases.
- Baltic countries example (Estonia, Latvia, Lithuania):
  - Real GDP contracted by 21¾ percent in Latvia and 17 percent in Estonia over 2008–09, and by 15 percent in Lithuania in 2009; prior expansions over 2005–07: Latvia 36½ percent, Estonia 29 percent, Lithuania 27¾ percent.
  - VAT growth was lower than private consumption growth during the 2008–09 contractions and higher during 2005–07 expansions (Figures 1a and 1b).
- Cross-country correlation:
  - For EU countries, 1995–2008: output gap coefficient 0.9, p-value less than 0.01; 405 observations (Figure 2).

### Data
- Three datasets:
  - Annual 1995–2008 for 32 EU countries.
  - Annual 1995–2008 for 84 advanced and developing economies.
  - Quarterly Q1 1999–Q1 2009 for 37 advanced and developing economies.
- Sources:
  - VAT, PIT, SSC: Eurostat general government surveys (EU), IMF GFS and IMF staff country data (non-EU).
  - VAT standard rates: IMF Fiscal Affairs Department database.
  - PIT and SSC rates: Eurostat.
  - GDP, private consumption, wages and salaries: IMF WEO.
  - Control of tax evasion: IMD World Competitiveness Online.
  - Law and order: International Country Risk Guide.
  - Share of necessity goods in consumption: Eurostat.

### Methodology
- Core fixed-effects panel regression (equation (1)):
  - taxefficiency_it = α_i + β gap_it + ε_it
  - taxefficiency is the tax revenue efficiency ratio; gap is the output gap.
- Extensions:
  - Interaction with a “bad times” dummy (bad times when actual real GDP growth < potential) to test asymmetry.
  - Interaction with an “advanced economy” dummy to test heterogeneity.
  - Additional covariates X_it: share of necessity goods in consumption, control of tax evasion, law and order, (lagged) tax burden.
- VAT revenue elasticity model (equation (3)):
  - log(tax revenue)_it = α_i + β1 log(tax base)_it + β2 taxrate_it + β3 gap_it + ε_it
  - taxrate is the standard VAT rate.

### Estimation results — VAT C-efficiency (quarterly and annual)
- Quarterly (Table 1):
  - Full sample (column 1): Gap coefficient = 0.286*** (0.061).
    - Interpretation: A one percentage point increase in the output gap in one quarter → increase in VAT C-efficiency by 0.29 percentage point.
  - Bad times (column 2): Gap = 0.332*** (0.085).
  - Good times (column 3): Gap = 0.184* (0.100).
  - Interacted model (column 4):
    - Gap = 0.190** (0.096).
    - Bad times dummy = -1.874*** (0.555).
    - Bad times*Gap = 0.145 (0.129) — not statistically significant.
  - Observations: Full sample 1,117; Bad times 454; Good times 661.
  - Number of countries: 37 (full), 36 (bad times).
- Annual (Table 2):
  - EU countries (column 5): Gap = 0.904*** (0.229).
  - Broader set (column 7): Gap = 1.166*** (0.193).
  - Interacted models (columns 6 and 8): no significant difference in slope between good and bad times.
  - Observations: EU sample 405; broader set 674.
  - Number of countries: 32 (EU), 84 (broader set).
  - Implication example: If output declines to 5 percent below potential → VAT C-efficiency falls by almost 6 percentage points (from 62 to 56 percent) based on average economy example.
- Advanced vs developing economies (Table 3):
  - Full sample (repeated column 7): Gap = 1.166*** (0.193).
  - Advanced (column 9): Gap = 1.019*** (0.274).
  - Developing (column 10): Gap = 1.274*** (0.277).
  - Interacted (column 11): Gap = 1.274*** (0.269); Advanced*Gap = -0.254 (0.389) — interaction not significant.
  - Observations: Full 674; Advanced 374; Developing 300.
  - Number of countries: Full 84; Advanced 31; Developing 53.
  - Conclusion: Differences between advanced and developing economies in gap coefficient are not statistically significant.

### Channels: shifts in consumption patterns and tax evasion (Table 4)
- Share of necessity goods (proxy: share of food and non-alcoholic beverages in total consumption) — 25-country sample:
  - Column 12 (baseline for these 25): Gap = 1.010*** (0.288).
  - Column 13 (add share of necessity goods): Gap = 0.562** (0.269); Share of necessity goods = -2.654*** (0.400).
    - Interpretation: As the share of necessity goods rises, VAT C-efficiency declines.
    - Output gap coefficient declines from 1.01 to 0.56 when controlling for consumption shifts.
- Control of tax evasion (IMD indicator):
  - Column 14: Gap = 0.961*** (0.281); Control of tax evasion = 1.810*** (0.518).
  - Column 15 (both controls): Gap = 0.565** (0.268); Share of necessity goods = -2.491*** (0.391); Control of tax evasion = 0.906** (0.411).
  - Column 16 (law & order and lagged tax burden):
    - Gap = 0.629** (0.297).
    - Law & order = 3.114*** (0.917).
    - Tax burden (lag) = -0.8305** (0.3959).
- Observations: 251 for each specification in Table 4; Number of countries: 25.
- R-squared range in Table 4: 0.077 to 0.329.
- Interpretation:
  - Shifts toward necessity goods and weaker control of tax evasion during downturns are key channels reducing VAT C-efficiency.
  - Stronger legal/institutional framework (law & order) is positively correlated with VAT C-efficiency.
  - Higher (lagged) tax burden is correlated with lower VAT C-efficiency.

### Estimation results — PIT and SSC efficiency (Table 5)
- PIT efficiency (column 17):
  - Gap = 1.428* (0.822).
  - Constant = 56.20*** (0.830).
  - Observations = 157; Number of countries = 13.
  - R-squared = 0.038.
- SSC efficiency (column 18):
  - Gap = 1.133** (0.572).
  - Constant = 74.21*** (0.627).
  - Observations = 258; Number of countries = 20.
  - R-squared = 0.019.
- Notes:
  - PIT and SSC efficiency both positively correlated with the output gap.
  - Interacted models: SSC gap remains significant; no statistically significant difference in level and slope between good and bad times for SSC; for PIT, no significant correlation in interacted model.

### Estimation results — VAT revenue elasticity (Table 6 and Table 7)
- VAT revenue elasticity with respect to private consumption (Table 6):
  - Full sample (column 19): Consumption (logs) = 1.073*** (0.0264).
  - Advanced (column 20): Consumption = 1.103*** (0.0330).
  - Developing (column 21): Consumption = 1.041*** (0.0361).
  - VAT rate coefficients:
    - Full sample: 0.0288*** (0.00806).
    - Advanced: 0.0438*** (0.00780).
    - Developing: -0.0334** (0.0157).
  - Gap coefficients:
    - Full sample: Gap = 0.0175*** (0.00337).
    - Advanced: Gap = 0.0112*** (0.00410).
    - Developing: Gap = 0.0216*** (0.00505).
  - Observations: Full 674; Advanced 374; Developing 300.
  - R-squared: Full 0.862; Advanced 0.914; Developing 0.838.
- Interpretation:
  - Consumption elasticity ≈ 1.1 (full sample), i.e., one percent increase in consumption → 1.1 percent increase in VAT revenue.
  - Gap coefficient interpretation: As dependent variable is in logs and gap not, coefficient multiplied by 100; a one percentage point increase in the output gap → 1.75 percent increase in VAT revenue over a year (consistent with the VAT C-efficiency finding).
- Illustrative summary (Table 7):
  - Change in output gap 1 percentage point:
    - Change in VAT revenues (All countries) = 1.75 percent; Advanced = 1.12 percent; Developing = 2.16 percent.
    - Change in VAT C-efficiency (All) = 1.17 percentage point; Advanced = 1.02 pp; Developing = 1.27 pp.
  - Change in output gap 5 percentage points:
    - Change in VAT revenues (All) = 8.75 percent; Advanced = 5.6 percent; Developing = 10.8 percent.
    - Change in VAT C-efficiency (All) = 5.83 percentage points; Advanced = 5.10 pp; Developing = 6.37 pp.

### Policy implications and conclusions
- Main quantitative estimates:
  - A one percentage point increase in the output gap corresponds to:
    - 1¼ percentage point increase in VAT C-efficiency (equivalently stated in text).
    - 1.75 percent increase in VAT revenue collections.
  - These results are consistent across quarterly and annual data, across advanced and developing economies, and in both good and bad times (bad times defined by growth below potential).
- Key channels to consider in forecasting and policy:
  - Shifts in consumption toward necessity goods (often taxed at lower or zero VAT rates) during downturns reduce VAT C-efficiency.
  - Changes in tax evasion (worse during downturns) reduce VAT C-efficiency.
  - Strong institutional underpinnings (law & order, tax administration control) improve VAT C-efficiency.
  - Higher (lagged) overall tax burden is associated with lower VAT C-efficiency.
- Policy recommendation:
  - During major booms and sharp downturns, policymakers should incorporate the effects of the economic cycle on tax revenue efficiency — beyond long-run revenue elasticities — to improve revenue forecasting and the calculation of structural fiscal balances, thereby informing counter-cyclical fiscal policy effectiveness.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1071.pdf*

### References

### _wp1071 - References

### Listed references
- Agha, Ali, and Jonathan Haughton, 1996, “Designing VAT Systems: Some Efficiency Considerations,” Review of Economics and Statistics, Vol. 78, No. 3, pp. 303-08.
- Aizenman, Joshua, and Yothin Jinjarak, 2008, “The Collection Efficiency of the Value Added Tax: Theory and International Evidence,” The Journal of International Trade & Economic Development, Vol. 17, No. 3, pp. 391-410.
- Brondolo, John, 2009, Collecting Taxes during an Economic Crisis: Challenges and Policy Options, IMF Staff Position Note, SPN/09/17 (Washington: International Monetary Fund).
- Cai, Hongbin, and Qiao Liu, 2009, “Competition and Corporate Tax Avoidance: Evidence from Chinese Industrial Firms,” The Economic Journal, Vol. 119, Issue 537, pp. 764-95.
- De Mello, Luiz, 2009, “Avoiding the Value-added Tax: Theory and Cross-country Evidence,” Public Finance Review, Vol. 37, No. 1, pp. 27-46.
- Eble, Stephanie, Cemile Sancak, and Ricardo Velloso, 2009, “Revenue Forecasting in Crisis Countries: Preliminary Lessons from Six Program Countries,” FAD, IMF, mimeo.
- Ebrill, Liam, Michael Keen, Jean-Paul Bodin, and Victoria Summers, 2001, The Modern VAT (Washington: International Monetary Fund).
- Plumley, Alan H., 1996, “The Determinants of Individual Income Tax Compliance: Estimating the Impacts of Tax Policy, Enforcement, and IRS Responsiveness,” Department of the Treasury, Internal Revenue Service, Publication 1916 (Rev. 11-96).

*Source: _wp1071 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1071.pdf_