## EXECUTIVE SUMMARY (South Africa) — Content unit: _cr15180

## Source details

**Canonical URL:** [EXECUTIVE SUMMARY (South Africa) — Content unit: _cr15180](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15180.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15180.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15180.pdf.json)

---

### Main findings
- RA-GAP methodology applied to South Africa for the period 2007–12 to estimate the VAT compliance gap using a top-down value-added approach.
- Compliance gap estimated to be between 5 percent and 10 percent of potential VAT revenues during 2007-12, peaking in 2008 and 2009.
- Compliance gap increased to 10 percent of potential revenue in 2009 and gradually decreased to 6 percent of potential revenue by 2012 (around the same level as 2007).
- Calculated gap numbers will be revised upward once recent national accounts revisions are incorporated, by 1-3 percent.
- Estimated gap is low by international standards, below typically observed levels in European and Latin American countries.
- VAT policy gap calculated to be between 27 percent and 33 percent of the theoretical potential VAT during 2007–2012.
- VAT tax expenditure calculated to be 15-18 percent of the potential VAT under normative policy.
- Although policy gaps are higher than compliance gaps, scope for additional revenue by changing VAT policy structure looks limited.

### Observations on trends and drivers
- Declining trend of VAT c-efficiency ratio after 2007 explained by increase of ‘non-taxable’ final consumption both by policy and by character among total final consumption.
- C-efficiency ratio level is high among peer countries but fluctuated with a downward trend after 2007.
- Overall reduction in c-efficiency compared to pre-2007 due to changes in non-taxable consumption and VAT tax expenditures (changes to composition of GDP and final consumption).
- Economic downturn between 2008 and 2009 resulted in increase of the compliance gap, with subsequent decrease.
- Cash collection numbers more volatile than accrued collections, exacerbating movements in c-efficiency ratio.
- Calculated gaps generally consistent with internal SARS estimates using a demand approach between 2007 and 2012; SARS net cash collection estimates are more volatile, RA-GAP accrual-based estimates relatively stable.
- Combined SARS and RA-GAP estimates show a declining trend of compliance gaps in years after 2002.

### Observations and possible follow-up action (recommendations)
- SARS should continue to monitor the VAT compliance gap to evaluate performance and inform strategic tax decisions.
- SARS should use revised supply-use tables due to be released in February 2015 to update its VAT gap estimate and sectoral composition.
- SARS could consider broadening tax gap analysis to include other major taxes.
- SARS should further integrate revenue and national compliance analyses to support systemic compliance risk management.
- Scope for more detailed analysis of revenues from individual industry sectors and taxpayer segments to support strategic risk analysis.

### Background and methodology overview
- RA-GAP provides quantitative analysis of the gap between potential revenues and actual collection (the compliance gap) using a top-down value-added approach enabling sectoral breakdown.
- Two main components:
  - Estimate potential net VAT collections for a given period.
  - Determine accrued net VAT collections for that period.
- Compliance gap = potential net VAT collections − accrued net VAT collections.
- Policy gap defined by comparing potential VAT under current policy (PV1) to theoretical VAT from final consumption at the standard rate (PV3); tax expenditure = difference between PV2 (normative policy) and PV1.

### Value-Added Tax revenue performance (2002–2013 highlights)
- VAT revenues increased in nominal terms from 70 billion R to 238 billion R between FY2002 and FY2013.
- Total VAT revenues = net domestic VAT (net domestic cash collections) + import VAT collected by Customs.
- Imports relative to household final consumption averaged 50 percent between 2002 and 2013.
- Between 2008 and 2011: domestic VAT collections fell then recovered in 2008 and 2011; import VAT significantly declined between 2009 and 2010.
- VAT revenues as percent of GDP increased from 2002, peaking at 7.3 percent in 2006; fell to 6.0 percent in 2009; returned to 6.9 percent in 2013.
- No significant VAT policy changes during the decade; standard rate remained 14 percent.
- Fluctuations in VAT revenues relative to GDP likely attributable to changes in tax base and taxpayers’ compliance.

### C-efficiency ratio in South Africa
- C-efficiency ratio measures actual VAT collections relative to amount under perfect enforcement at the standard rate on final consumption.
- C-efficiency roughly equates to the product of the compliance gap and the policy gap.
- Average c-efficiency ratio in South Africa between 2007 and 2013 is 63.6 percent.
- This is among the highest in Sub-Saharan Africa for the same period.
- High c-efficiency partly reflects limited exempted and zero-rated goods and services.
- Major VAT exemptions: education, passenger transport (road or rail), financial services (except fees), rent for residential accommodation.
- Zero-rated items include goods used for farming activities (6 Items), some foodstuffs (19 Items), fuels (fuel levy imposed) and kerosene, and local government services.
- Yearly changes in c-efficiency decomposed into: compliance gap changes; policy gap (tax expenditures) changes; changes in share of non-taxable consumption; timing effects of cash payments/refunds.
- Fall in c-efficiency in 2009 largely attributed to increased compliance gap and timing impacts in VAT refunds.

### Estimating the VAT compliance gap (2007–2012 application)
- RA-GAP used supply and use tables and national accounts to estimate potential VAT revenues; actual collections based on SARS tax returns and payment data on an accruals basis as at August 2014 (diesel refund claims added back).
- Compliance gap pattern 2007–2012: hump-shaped—grew significantly from 2007 to 2009 (10 percent in 2009) then reverted to same level as 2006 by 2012 (6 percent).
- Observed increase in compliance gap explains much of decline in VAT collections as percent of GDP, especially 2008–2009.
- Accrued VAT collections show smoother trend than net cash collections, particularly in 2009; timing of payments and refunds drives cash vs accrued differences.
- For import VAT, monthly cash collection aggregated by SARS used as actual collections.

### Effects of the recent GDP revision (Stats SA, November 27th 2014)
- Yearly nominal GDP level revisions (Previous → Revised → Difference):
  - 2006: 1,767 → 1,839 → 4.1%
  - 2007: 2,016 → 2,110 → 4.7%
  - 2008: 2,256 → 2,369 → 5.0%
  - 2009: 2,408 → 2,508 → 4.2%
  - 2010: 2,674 → 2,748 → 2.8%
  - 2011: 2,933 → 3,025 → 3.1%
  - 2012: 3,139 → 3,263 → 4.0%
  - 2013: 3,385 → 3,534 → 4.4%
- Implications:
  - 2008 SNA implementation does not materially affect estimated potential VAT revenues (capitalization and redefinitions account for about 1.2 percent of previous GDP).
  - Revised supply and use tables (including new 2013 tables) will raise potential VAT revenues by 2-4 percent relative to report numbers, raising estimated compliance gaps by 1-3 percent; trends and general levels of gaps will not significantly change.

### The assessment gap and collection gap (2007–2012)
- Definitions:
  - Collection gap: actual collections − declared/assessed VAT (identified portion of compliance gap).
  - Assessment gap: declared/assessed VAT − potential VAT (unidentified portion).
- Findings:
  - Collection gap gradually grew during 2007–2012.
  - Assessment gap first increased sharply then fell back to less than its former level.
  - Growing collection gap may reflect first-in-first-out payment prioritization and risk of rising taxpayers’ arrears stock.

### Changes in potential revenues and actual collections (drivers)
- Potential VAT revenues as percent of GDP declining since at least 2008; decline attributed to tax base changes rather than policy changes.
- Major contributors:
  - Increased non-taxable GDP components, notably government final consumption (18.8 percent of GDP in 2007 → 22.2 percent of GDP in 2013).
  - Increased share of petroleum products in household consumption, which are zero-rated for VAT (separate specific petroleum levy exists).
- Decline in imports in 2009 and 2010 temporarily reduced potential VAT revenues relative to GDP; on average import changes matched zero-rated exports, canceling effects.
- Accrued collections smoother than net cash collections; higher refunds in 2009 reflected excess credits declared in 2008.

### Sectoral distribution and decomposition issues
- Main contributors to VAT collections:
  - Sector 8 (Financial intermediation, insurance, real-estate and business services): 19.4 percent of economic activity in 2013.
  - Sector 6 (Wholesale and retail trade, catering and accommodation): 13.7 percent.
  - Sector 3 (Manufacturing): 13.0 percent.
  - Sector 2 (Mining and quarrying): 7.7 percent (records negative VAT due to large zero-rated exports).
- Sectoral comparability issues:
  1. Different classification systems between national accounts and tax collection.
  2. Limited detailed sector export, import, and capital formation data.
  3. Limited availability of sector import VAT collections data.
- Sector allocation of actual collections based on VAT vendor classification by Stats SA where possible; SARS classification used otherwise; accrued to tax periods (differences from Tax Statistics 2014).
- Large vendors classified as sector 8 may span multiple sectors, potentially understating other sectors’ actual collections.
- Export/import sector data limitations: export data linked to VAT reference numbers available only after 2012; some Customs export declarations lack VAT numbers.
- Recommendation: establish common sector classification and consistent decompositions of export and import; investigate discrepancies between Customs export declarations and VAT zero-rating declarations.

### Estimation and evaluation of the policy gap (definitions and measures)
- Policy gap: PV3 (theoretical revenue from final consumption taxed at standard rate) − PV1 (potential VAT under current policy); indicates VAT policy efficiency (not expected immediate revenue gain if implemented).
- VAT tax expenditure: PV2 (normative policy with single standard rate and minimal exemptions/reliefs) − PV1; PV2 assumes exempt: financial services, residential housing rent, public administration services; PV2 assumes zero-rated: exports.
- Policy gap results:
  - Policy gap equivalent to "2.7-3.5 percent of GDP".
  - Policy gap equals "25-32 percent of theoretical potential VAT".
  - Average policy gap in European countries is "41 percent" (comparison).
  - VAT tax expenditure "1.1-1.5 percent of GDP".
- Policy implication: VAT tax expenditure higher than compliance gap; greater potential for revenue mobilization by restricting exemptions/zero-rates than by only improving administration, but overall policy gap low by international standards so room limited.

### Decomposition of c-efficiency changes after 2007
- Four components:
  - Changes in compliance gap.
  - Changes in VAT tax expenditure.
  - Changes in share of non-taxable consumption.
  - Timing effects of payment and refund delays (cash effects).
- Observed pattern:
  - 2007→2009 sharp c-efficiency decline explained by increasing compliance gap plus payment/refund timing effects.
  - Subsequent c-efficiency increase as compliance and cash effects abated, despite continued widening from VAT tax expenditure and other base effects.

### Comparison of RA-GAP results with SARS analysis (methodological differences)
- SARS national compliance analysis (2002–2012) used net cash collection basis and demand-side statistics; RA-GAP uses accrual allocation and supply-use tables.
- SARS gap fell from nearly "30 percent" in 2002 to "10 percent" in 2005, then fluctuated between "5 percent and 17 percent".
- Two major methodological differences:
  1. SARS uses net cash collection by fiscal year; RA-GAP uses accrual allocation by tax period.
  2. SARS uses demand-side statistics for potential VAT; RA-GAP uses supply and use tables.
- Consequences:
  - Net cash collection basis yields large fluctuations, especially 2009 and 2011; accrual approach links receipts to economic activity more clearly and reduces timing volatility.
  - RA-GAP accrual approach preferred for presenting underlying taxpayer compliance trends.

### Identified issues in SARS potential VAT estimates and recommended corrections
- SARS method: sum of VAT liabilities from final demand, intermediate consumption, and capital formation using national accounts and AFS.
- Identified issues:
  - Final consumption by non-residents treated inconsistently through time; recommendation: do not add or subtract to avoid step change around 2007–2008.
  - Exemption ratios for financial services changed qualitatively; recommendation: reconsider exemption ratios over time.
  - Double counting in informal sector (sales by informal non-registrants not excluded while intermediate consumption included) causes overestimation.
  - Undercoverage of capital formation (only residential dwellings included) causes underestimation; recommend including capital formation by entities producing exempt products (government, financial sector) and capital formation related to passenger motor vehicles.
- Net effect: Overestimation of final consumption component outweighs underestimation of capital formation, making SARS potential VAT slightly higher than RA-GAP after adjustments; corrections recommended to capture trends and stabilize estimates.

### Compliance management reforms and observed impacts
- VAT compliance gap in South Africa low by international standards:
  - Average VAT compliance gap in EU member states in 2012 estimated at "16 percent".
  - Average VAT gap for Latin American countries estimated at "27 percent" for 2006-2010.
- SARS compliance reforms phases:
  - 1998–2002: Structural reforms—consolidation, centralization, standardized services, HR autonomy.
  - 2001–2006: Rolling compliance campaigns (Woodmead, Siyakha 1 & 2), private-sector experts, Large Business Centre (2005).
  - Since 2006: Modernization/automation, national compliance analysis introduced FY 2008/09, national risk profiles, enhanced LBC analysis, VAT registration threshold increase from "R300,000 pa" to "R1 million pa".
- Observed impacts:
  - VAT gap trends align with reforms: high gap (25–30 percent) pre-2002, steady decrease to under 10 percent during 2002–2006, increase in 2008–2009 (financial crisis), return to under 10 percent and broadly level 2007–2012.
  - SARS reports improved service times, better audit results from automated risk processes, greater data efficiency reducing auditor routine checking time.

### SARS compliance monitoring, indicators and recent performance (post-2011)
- National compliance analysis (‘pillars of compliance’) monitors:
  - registration, filing compliance, payment compliance and credit returns, declaration compliance (including audit results), net tax receipts.
- Filing compliance (2008/09 – 2013/14):
  - Non-filing by active taxpayers increased by 13.3 percent.
  - Overall filing rate fell from 79.1 percent to 65.8 percent.
  - SARS attributes part of decline to increasingly inactive taxpayers recorded as active.
- Payment compliance (2008/09 – 2013/14):
  - Payment compliance rose from 81.1 percent to 85.1 percent.
  - No penalties for late filing of VAT returns; penalties exist for late/non-payment.
- VAT credit returns (2008/09 – 2013/14):
  - Proportion of credit returns fell from 21 percent to 18.4 percent.
  - Average value of excess credits rose from R182,735 to R317,869.
  - Total value of credit returns reached R157 million in FY 2013/14.
  - Increase in excess credits attributed to increasing imports.
- Audit counts and yield:
  - Audits fell from 7,612 in FY 2011/12 to 3,565 in FY 2013/14.
  - Percentage of audits finding inaccurate declarations rose from 51.7 percent to 62 percent.
  - Overall yield from audits rose from R4.5 billion to R6.1 billion; average yield per audit increased strongly.
- VAT registration threshold change and micro-businesses:
  - Threshold increased from R300,000 pa to R1 million pa in FY 2007/08.
  - Deregistration campaign reduced registered micro businesses; prior to rise, businesses with turnovers < R1 million received net VAT repayments ~2-3 percent of total VAT receipts.
  - After threshold rise, net repayment to such businesses rose to ~5 percent of total VAT receipts in FY 2008/09, then fell to negligible levels, reducing observed VAT gap by ~2-3 percent.
  - SARS noted refund increase to micro businesses in FY 2013/14 to about 5 percent of total VAT receipts.
- Non-VAT registered businesses (2001–2013):
  - Counts (thousands): 2,258 (2001), 1,668 (2005), 1,144 (2009), 1,517 (2013).
  - Business-size distribution (percent by turnover band, 2001, 2005, 2009, 2013):
    - R0 - R100: 10.8, 11.4, 8.8, 8.2
    - R101 - R350: 29.5, 21.4, 14.5, 11.0
    - R351 - R750: 21.4, 23.3, 18.7, 13.8
    - R751 - R1 500: 15.3, 17.0, 19.9, 19.3
    - R1 501 - R3 000: 8.5, 12.2, 15.8, 18.5
    - R3 001 - R6 000: 6.0, 7.0, 10.8, 14.6
    - R6 001 plus: 3.7, 7.4, 11.5, 14.6
    - Unspecified: 4.9, 0.3, 0.0, 0.0
  - Majority (85 percent in 2013) reported annualized turnovers less than R72 thousand, well below registration threshold of R1 million.

### Recommendations (operational and analytic)
- SARS should continue to estimate the VAT compliance gap and further integrate revenue and compliance analyses.
- Use compliance analysis to inform revenue analysis and vice versa; update compliance gap estimates as updated data become available.
- Monitor detailed time series of sectoral and demographic aggregate returns and payments to identify changing trends, step changes, and spikes indicating emerging compliance risks.
- Bring together revenue and compliance analysts with operational experts for integrated monitoring to aid compliance risk management and fiscal analysis.
- Consider extending integrated revenue and compliance analysis approach to other taxes.

---

### Appendix I — RA-GAP model and methodology (summary)
- RA-GAP is a top-down value-added model estimating taxable value-added across sectors using supply-use or input-output tables; advantages: covers all compliance losses and allows comparison to tax expenditures.
- Potential net VAT for a sector computed from sectoral imports, output, exports, intermediate demand, investment, statutory VAT rates by commodity, allowed input tax credit proportions, proportion of output by registered businesses, and exempt output proportions.
- Accrued collections computed by reallocating cash payments and refunds to tax periods: accrued VAT = customs collections for period + payments received for period − excess credit accrued for period + payments offset by excess credit.
- Two RA-GAP compliance measures:
  1. Compliance gap at time of filing (captures voluntary compliance; static).
  2. Compliance gap at time of estimation (uses latest available data; changes over time).
- Reporting conventions: gap usually reported as (CPV − AV)/CPV and also as percent of GDP.

### Appendix II — Application to South Africa (data and parameters)
- Potential VAT (PV1 and PV2) calculated for each year 2007–2012 using supply and use tables (62 sectors and 104 commodities) with policy parameters.
- PV1 uses ratios for exempted and zero-rated products based on IES2010/11.
- Selected exemption ratios (examples):
  - CPC_239 Food n.e.c.: 0.0637
  - CPC_64 Passenger transport services: 0.9097*
  - CPC_711 Financial services: 1
  - CPC_72 Real estate services: 0.9818
  - CPC_92 Education services: 0.9690
  - Note: For CPC_64, production by I42 and I61 treated as exempted.
- Selected zero-rating ratios (examples):
  - CPC_01 Products of agriculture: 0.8799
  - CPC_02 Live animals and animal products (excluding meat): 0.5048
  - CPC_212 Prepared and preserved fish: 0.4730
  - CPC_231 Grain mill products: 0.7427
  - CPC_33 Coke oven products, refined petroleum products: 0.9792
  - CPC_233 Preparations used in animal feeding: 1* (only intermediate consumption treated as zero-rated)
  - CPC_346 Fertilizers and pesticides: 1* (only intermediate consumption treated as zero-rated)
- Government treatment:
  - Government final consumption regarded as non-taxable; public administration services by LG zero-rated (can claim input tax); services by CSG treated as exempt (cannot claim input tax).
- Informal sector (I62) production regarded as exempt in the model.
- Restrictions on input tax credits: accommodation, catering services, and purchase of motor vehicle (except I42 and I51) not allowed.
- Sector allocation of capital formation uses National Bank data; import/export allocation uses Customs data in 2012.
- Actual collections measurement:
  - Aggregate cash collection from SARS monthly data; accrued collection = payments reallocated to tax period + excess credit declared in tax period + customs payments in each calendar year.
  - Diesel refund adjustments: before FY2013 refunds recorded as reductions of VAT net payments were added back; after FY2013 diesel claims included as separate item so adjustment not necessary.
- Data sources listed for potential VAT and accrued collections (supply and use tables 2007-2012, IES2010/11, Customs declarations 2010–2013, SARS payments/refunds/returns/audit datasets).

### Appendix III — Comparison with SARS national compliance analysis
- SARS national compliance analysis (introduced FY 2008/09) monitors ‘pillars of compliance’: registrations, filing compliance, payment compliance and excess credit returns, declaration compliance.
- Registrations: new registrations fell from 64.4 thousand in 2008/09 to 26.6 thousand in 2009/10; total VAT register 662 thousand vendors (~4 percent churn).
- Filing compliance: declined by 13.3 percent since FY 2008/09; on-time filing steady at 55-60 percent; overall filing fell from 79.1 percent to 65.8 percent; LBC filing compliance 87.9 percent; lowest sector: construction at 45.2 percent.
- Payment compliance: ~85 percent overall; LBC payment compliance 95.7 percent; highest noncompliance in public administration (10.3 percent) and construction (7.2 percent).
- Credit returns: proportion decreased, value increased; refunds to public administration rose with investment.
- Audits: number decreased (7,612 → 3,565), yield increased (R4.5 billion → R6.1 billion); targeted audits found inaccuracies in 62 percent of cases with average yield R1.7 million per audit.
- Declaration compliance: highest inaccuracies in public administration (71.1 percent) and construction (64.8 percent); highest yielding noncompliance behaviors include under-declaration of standard rated outputs and over-declared zero rated outputs (exports).

*Source: _cr15180 — RA-GAP staff estimates and analysis as presented in the EXECUTIVE SUMMARY and related chapters.*

### EXECUTIVE SUMMARY ___________________________________________________________________________ 5

### EXECUTIVE SUMMARY

### Main findings
- The RA-GAP methodology was applied to South Africa for the period 2007–12 to estimate the VAT compliance gap using a top-down value-added approach.
- The compliance gap is estimated to be between 5 percent and 10 percent of potential VAT revenues during the period 2007-12, peaking in 2008 and 2009.
- The compliance gap increased to 10 percent of potential revenue in 2009, when the global financial crisis severely hit the South African economy, and has since gradually decreased to the same level as 2007.
- The calculated gap numbers will be revised upward once the recent revisions of national accounts data in South Africa are incorporated, by 1-3 percent.
- The estimated gap is low by international standards, below the typically observed levels in European and Latin American countries.
- The VAT policy gap is calculated to be between 27 percent and 33 percent of the theoretical potential VAT during 2007–2012.
- VAT tax expenditure is calculated to be 15-18 percent of the potential VAT under normative policy.
- Although policy gaps are higher than compliance gaps, the room for additional revenue by changing VAT policy structure looks limited.

### Observations on trends and drivers
- The recent declining trend of VAT c-efficiency ratio after 2007 can be explained by the increase of ‘non-taxable’ final consumption both by policy and by character among total final consumption.
- The level of c-efficiency ratio is high among peer countries but has fluctuated with a downward trend after 2007.
- The overall reduction in c-efficiency levels compared to the period before 2007 is due to changes in non-taxable consumption and VAT tax expenditures (changes to the composition of GDP and final consumption).
- The economic downturn between 2008 and 2009 resulted in an increase of the compliance gap, but it subsequently decreased.
- Cash collection numbers have been more volatile than accrued collections, exacerbating movements in the c-efficiency ratio.
- Calculated gaps are generally consistent with internal estimates by SARS using a demand approach between 2007 and 2012; SARS estimates using net cash collection data are more volatile, while RA-GAP estimates based on accrued collections are relatively stable.
- Combined estimates by SARS and RA-GAP show a declining trend of compliance gaps in the years after 2002.

### Observations and possible follow-up action (recommendations)
- SARS should continue to monitor the VAT compliance gap as a means of evaluating its performance, and to inform strategic decisions about tax.
- SARS should take the opportunity of revised supply-use tables due to be released in February 2015 to update its estimate of the VAT gap, and its sectoral composition.
- SARS could consider broadening its tax gap analysis to include other major taxes.
- SARS should further integrate its revenue and national compliance analyses, to support systemic compliance risk management.
- There is scope for more detailed analysis of revenues from individual industry sectors and taxpayer segments to support strategic risk analysis.

---

### Background and methodology overview
- The RA-GAP program provides a comprehensive quantitative analysis of the gap between potential revenues and actual collection, known as the compliance gap.
- The RA-GAP model uses a value-added approach that allows a breakdown of the VAT compliance gap by sector of economic activity.
- The RA-GAP methodology employs a top-down approach for estimating the potential VAT base, using statistical data on value-added generated in each sector.
- Two main components for estimating the VAT compliance gap:
  - Estimate potential net VAT collections for a given period.
  - Determine accrued net VAT collections for that period.
- The difference between potential net VAT collections and accrued net VAT collections equals the compliance gap.
- Potential VAT revenue under current policy is compared to theoretical VAT revenue calculated from final consumption and the standard rate to define the policy gap (difference between PV3 and PV1 relative to PV3). Tax expenditure is calculated as the difference between PV2 and PV1 relative to PV2.

### Value-Added Tax revenue performance (2002–2013 highlights)
- VAT revenues increased in nominal terms from 70 billion R to 238 billion R between FY2002 and FY2013, largely owing to economic growth and inflation.
- Total VAT revenues = net domestic VAT (net domestic cash collections) + import VAT collected by Customs.
- Imports relative to household final consumption averaged 50 percent between 2002 and 2013, making import VAT collections similar in level to net domestic VAT.
- Between 2008 and 2011, there were fluctuations: in 2008 and 2011 domestic VAT collections fell then recovered; import VAT significantly declined between 2009 and 2010.
- VAT revenues as a percent of GDP increased from 2002, peaking at 7.3 percent in 2006.
- In 2009, the ratio fell to 6.0 percent of GDP, returning to 6.9 percent in 2013.
- No significant changes in VAT policy (rate or exemption) occurred during this decade; standard rate remained 14 percent.
- Fluctuations in VAT revenues relative to GDP are likely attributable to changes in the tax base and changes in taxpayers’ compliance.

### C-efficiency ratio in South Africa
- The c-efficiency ratio measures the ratio of actual VAT collections to the amount that would be collected under a perfectly enforced tax levied at the standard rate on overall final consumption.
- C-efficiency ratio roughly equates to the product of the compliance gap and the policy gap.
- The average c-efficiency ratio in South Africa between 2007 and 2013 is 63.6 percent.
- This result is among the highest in Sub-Saharan African countries over the same period.
- High c-efficiency partly reflects South Africa’s limited number of exempted and zero-rated goods and services.
- Major VAT exemptions include education, passenger transport (road or rail), financial services (except fees), and rent for residential accommodation.
- Zero-rated items include goods used for farming activities (6 Items), some foodstuffs (19 Items), fuels (on which fuel levy is imposed) and kerosene, and local government services.
- Yearly changes in c-efficiency can be decomposed into: changes in the compliance gap; changes in the policy gap (tax expenditures); changes in the share of non-taxable consumption; and timing effects of cash payments and refunds.
- The fall in c-efficiency in 2009 can be largely attributed to an increase in the compliance gap and timing impacts in VAT refunds.

*Source: RA-GAP staff estimates and analysis as presented in the EXECUTIVE SUMMARY (South Africa).*

### 8. Estimating the VAT compliance gap. The VAT compliance gap for a particular year is the

### 8. Estimating the VAT compliance gap. The VAT compliance gap for a particular year is the

### A. VAT Compliance Gap in South Africa
- Definition: The VAT compliance gap for a particular year is the difference between revenues actually collected and the potential revenues that could have been collected given the policy framework that was in place during that year.
- Methodology:
  - RA-GAP approach used to estimate the compliance gap for the years 2007 to 2012.
  - Potential VAT revenues estimated using detailed national accounts data provided by Stats SA.
  - VAT actual collections based on tax returns and payment data between 2007 and 2012 provided by SARS, calculated on an accruals basis as at August 2014.
  - Adjustments in measuring actual collections required adding back diesel refund claims.
- Main findings:
  - The compliance gap is hump-shaped over 2007–2012: it grew significantly from 2007 to 2009 and has since reverted to the same level as 2006.
  - Estimated compliance gap increased to 10 percent of potential revenue in 2009.
  - The compliance gap gradually decreased to 6 percent of potential revenue by 2012, which is around the same level as 2007.
  - The observed increase in the compliance gap explains much of the observed decline in VAT collections as a percent of GDP, particularly the sharp increase between 2008 and 2009.
- Data notes:
  - The general method for accrual tax collection calculation is described in Appendix I, section C; specific method and data for measuring VAT collections in South Africa are in Appendix II, section C.

### Box: The Effects of the Recent Revision of Gross Domestic Products in South Africa
- Stats SA revised GDP numbers on November 27th 2014 due to implementation of the 2008 SNA and inclusion of new data sources.
- Yearly nominal GDP level revisions (Previous GDP (billion R) → Revised GDP (billion R) → Difference (%)):
  - 2006: 1,767 → 1,839 → 4.1%
  - 2007: 2,016 → 2,110 → 4.7%
  - 2008: 2,256 → 2,369 → 5.0%
  - 2009: 2,408 → 2,508 → 4.2%
  - 2010: 2,674 → 2,748 → 2.8%
  - 2011: 2,933 → 3,025 → 3.1%
  - 2012: 3,139 → 3,263 → 4.0%
  - 2013: 3,385 → 3,534 → 4.4%
- Implications:
  - Implementation of 2008 SNA does not affect estimated potential VAT revenues materially because changes arise from capitalization and redefinitions that do not alter the VAT tax base magnitude (these changes account for about 1.2 percent of previous GDP numbers).
  - Revised supply and use tables (including new 2013 tables) will raise potential VAT revenues by 2-4 percent relative to the numbers in this report, which would raise estimated compliance gap numbers by 1-3 percent; neither the trend nor the general level of gaps will significantly change.

### B. The Assessment and Collection Gaps
- Definitions:
  - Collection gap: difference between actual VAT collections and the total amount of VAT declared or assessed as due from taxpayers (identified portion of compliance gap).
  - Assessment gap: difference between the amount of VAT declared or assessed and potential VAT (unidentified portion of compliance gap).
- Findings over 2007–2012:
  - Collection gap gradually grew during the period.
  - Assessment gap first increased sharply and then fell back to less than its former level.
  - The increase in the collections gap widens differences between declared/assessed VAT and collected VAT year by year; this may reflect first-in-first-out payment prioritization for older liabilities and a risk of growing stock of outstanding taxpayers’ arrears.

### C. Changes in Potential Revenues and Actual Collections
- Changes in potential VAT revenues:
  - Potential VAT revenues as a percent of GDP have been declining since at least 2008 (calculated from national accounts data).
  - No significant VAT policy changes occurred during this period; decline attributed to changes in the tax base.
  - Major contributors to the decline:
    - Increased non-taxable GDP components, notably services provided by government: government final consumption grew from 18.8 percent of GDP in 2007 to 22.2 percent of GDP in 2013.
    - Increased share of petroleum products in household consumption, which are zero-rated for VAT (noting a separate specific levy on petroleum products exists in South Africa).
  - The decline in imports in 2009 and 2010 temporarily reduced potential VAT revenues relative to GDP due to the global financial crisis; on average over the whole period, changes in imports matched changes in zero-rated exports, canceling effects on overall potential VAT revenues.
- Changes in actual VAT collections:
  - Accrued VAT collections show a much smoother trend than net cash collections between 2007 and 2012, especially in 2009.
  - Timing of payments and refunds drives differences between accrued and cash collections: higher refunds in 2009 reflected declared excess credit for tax periods in 2008 when economic activity was higher; when the economy recovered, adverse effects occurred in 2011 and 2013 with refunds for prior periods lower than in later years.
  - For import VAT, total monthly cash collection aggregated by SARS was used as actual VAT collections.

### D. Actual Collections and Potential VAT by Sector
- Concentration of VAT collections:
  - Sector ‘8. Financial intermediation, insurance, real-estate and business services’ and sector ’6. Wholesale and retail trade, catering and accommodation’ are the main contributors of VAT collections.
  - Sector shares of economic activity in 2013: sector 8 = 19.4 percent; sector 6 = 13.7 percent; sector 3 (Manufacturing) = 13.0 percent; sector 2 (Mining and quarrying) = 7.7 percent.
  - Sector 2 records negative VAT due to a large volume of zero-rated exports.
- Decomposition and comparability issues:
  - Overall potential VAT can be decomposed into sectors, but direct comparison with sector collections to infer sector compliance gaps is difficult due to:
    1. Different classification systems between national accounts (supply and use tables) and tax collection.
    2. Limited availability of detailed sector export, import, and capital formation data.
    3. Limited availability of sector import VAT collections data.
  - Sector allocation of actual VAT collections is based on VAT vendor classification provided by Stats SA where possible; for vendors without Stats SA classification, SARS classification applied. Values are accrued to tax periods; therefore sector collection values differ from those in ‘Tax Statistics 2014’.
  - Large VAT vendors classified as sector 8 (SARS classification) may have operations spanning multiple sectors, causing actual collections attributed to sector 8 to include VAT from other sectors’ activity — this may make actual collections appear smaller than potential revenues in those other sectors.
  - Export and import sector data limitations:
    - Export data associated with VAT reference numbers are available only after 2012; some Customs export declarations lack VAT declaration numbers and cannot be classified into sectors.
    - In some sectors (notably mining and quarrying, sector 2), allocated export values are substantially smaller than declared zero-rated goods and services, producing larger potential VAT estimates relative to actual collections.
- Recommendation for sector analysis:
  - To specify sector compliance gap numbers appropriately, a common sector classification and consistent sector decompositions of export and import are necessary.
  - Despite overall potential revenue and compliance gap estimates being robust to detailed sector information, sector-level potential revenues and gaps heavily rely on consistent sector data; current detailed sector information is not sufficiently available.
  - Suggested follow-up: investigate discrepancies between Customs export declarations and VAT zero-rating declarations (for example, possible holding-company purchase patterns that reallocate VAT attribution across sectors).

### Estimation and Evaluation of the Policy Gap
#### A. Definitions and Methodologies
- Purpose: VAT policy gap and tax expenditure are calculated as the difference between theoretical revenue given a hypothetical policy framework and potential revenue under the current policy (Potential VAT under current policy: PV1), to show the efficiency of VAT policy.
- Interpretation caveat: Policy gap is not an indicator of the size of expected additional revenues following transition to the normative framework (taxpayers’ behavior is not assumed to change); it is an indicator of VAT policy efficiency.
- Measures:
  - Policy gap: difference between theoretical revenue from final consumption taxed at the standard rate (PV3) and Potential VAT under current policy (PV1). This is a “theoretical” potential as taxing all final consumption at a standard rate may be difficult to implement; comparable to the c-efficiency ratio.
  - VAT tax expenditure: difference between theoretical revenue under a specific normative VAT structure with a single standard rate and a necessary minimum of exemptions and reliefs (Potential VAT under normative policy: PV2) and Potential VAT under current policy (PV1).
    - Under PV2, assumed exempt: financial services, residential housing rent, public administration services.
    - Under PV2, assumed zero-rated: exports.
    - VAT tax expenditure mainly captures effects of other exemptions and zero-rating policies (foodstuffs, agricultural related goods, fuels, local government activities).

*Source: IMF staff estimates and analysis as presented in the chapter.*

### 26. The level of the VAT policy gap in South Africa is quite low by international standards.

### 26. The level of the VAT policy gap in South Africa is quite low by international standards.

### Key findings on level and composition of VAT gaps
- Policy gap in South Africa is equivalent to "2.7-3.5 percent of GDP".
- Policy gap equals "25-32 percent of theoretical potential VAT".
- By comparison, the average policy gap in European countries is "41 percent".
- VAT tax expenditure has been "1.1-1.5 percent of GDP".
- The VAT compliance gap is lower than the VAT tax expenditure, implying greater potential for revenue mobilization by reducing exemptions/zero-rates than by only improving administration.

### Trends and drivers, 2007–2012
- Policy gap and VAT tax expenditure increased over 2007-2012.
- No significant changes in VAT policy over this period; increases in the policy gap are attributable to increases in economic activities that are not taxable by policy choice.
- Main contributors to growth in the policy gap:
  - Increasing share of zero-rated local government activities leading to rising input tax credit.
  - Increasing share of final consumption taken by petroleum products due to price increases and inelastic demand.
- Note: Concept of final consumption in national accounts includes items difficult to levy VAT on (such as financial intermediation), and imputed transactions (such as government final consumption and imputed rents).

### VAT tax expenditure versus compliance gap
- VAT tax expenditure is higher than the compliance gap.
- Policy implication: Potential scope for additional revenue mobilization by restricting exemptions and zero-rates (reducing the policy gap) is greater than by increasing effectiveness of revenue administration (reducing the compliance gap).
- Caveat: The policy gap is quite low by international standards, so actual scope for base erosion reduction may be limited.
- C-efficiency ratio defined as actual VAT collections divided by theoretical potential revenues.

### Decomposition of changes in c-efficiency after 2007
- Yearly fluctuations of c-efficiency ratios after 2007 are decomposed into four components:
  - Changes in the compliance gap.
  - Changes in VAT tax expenditure.
  - Changes in the share of non-taxable consumption (difference between theoretical potential VAT and potential VAT under normative policy).
  - Timing effects of variations in delays in making payments and refunds (cash effects).
- Observed pattern:
  - Sharp decline of c-efficiency ratios from 2007 to 2009 explained by increasing compliance gap, with amplification from increasing delays in payments and decreasing delays in refunds.
  - Subsequent increase in c-efficiency as compliance and cash effects abated, even though VAT tax expenditure and other base effects continued to widen.
- Implication: Decline of both c-efficiency and VAT receipts as a percent of GDP driven by increases in activities outside the VAT base or relieved by policy.

### Comparison of RA-GAP results with SARS analysis (methodological differences and implications)
- SARS’ internal VAT gap calculation (2002–2012) used potential VAT revenues from statistical data and defined VAT gap as the ratio difference between actual tax revenues (net cash collection) and potential VAT revenues.
  - SARS-calculated compliance gap fell from nearly "30 percent" in 2002 to "10 percent" in 2005, then fluctuated between "5 percent and 17 percent" of potential VAT revenues.
- Two major methodological differences between SARS and RA-GAP:
  1. SARS uses actual collections based on net cash collection data (by fiscal year); RA-GAP allocates individual payments and refunds to the period for which they were made (accrual approach).
  2. SARS calculates potential VAT revenues from demand-side statistics only (mainly final consumption); RA-GAP uses supply and use tables.
- Consequences:
  - Net cash collection basis results in large fluctuations in calculated gap numbers, especially in 2009 and 2011; accrual allocation provides a clearer link between receipts and underlying economic activity and reduces timing volatility.
  - RA-GAP’s accrual approach is preferred to present underlying trends in taxpayers’ compliance more clearly.

### Identified data and modeling issues in SARS potential VAT estimates and recommended corrections
- SARS potential VAT estimation details:
  - Sum of three demand components: VAT liabilities from final demand, intermediate consumption, and capital formation.
  - Final demand used taxable household final consumption, intermediate consumption by government and informal sectors using national accounts use tables and IES ratios (IES2005/06 and IES2010/11).
  - Intermediate consumption and capital formation for the private sector (except informal) drawn from Annual Financial Survey (AFS).
- Identified issues and recommended corrections:
  - Treatment of final consumption by non-residents inconsistent through time: subtracted after 2008, added before 2007. Correction: final consumption by non-residents should be neither subtracted nor added to potential VAT revenues to eliminate a step change around 2007–2008.
  - Exemption ratios for financial services changed qualitatively (assumed 53 percent taxable up to 2007 vs. no taxable household final consumption after 2008). Recommendation: reconsider exemption ratios over time to avoid inconsistent assumptions.
  - Double counting of VAT liabilities in the informal sector leads to overestimation of potential VAT: sales by informal non-registrants were not excluded from taxable final consumption while their intermediate consumption was included, causing double counting.
  - Undercoverage of capital formation in SARS model causes underestimation of potential VAT from capital formation: SARS included only capital formation in residential dwellings; should include capital formation by entities producing exempted products (including government and financial sector) and capital formation related to passenger motor vehicles given significance.
- Net effect: Overestimation of final consumption component outweighs underestimation of capital formation, causing SARS potential VAT to be slightly higher than RA-GAP after data-treatment adjustments. Corrections are recommended to capture trends (e.g., decline in capital formation since 2008) and to stabilize potential VAT estimates.

### Compliance management reforms and observed impact
- Estimated VAT compliance gap in South Africa is low by international standards:
  - Average VAT compliance gap in EU member states in 2012 estimated at "16 percent" of potential revenues.
  - Average VAT gap for Latin American countries estimated at "27 percent" for 2006-2010.
- SARS estimate indicates a strong sustained improvement in VAT compliance over 2002-2007: gap decreased from around "30 percent" in 2002 to less than "10 percent" in 2007.
- SARS compliance management reforms (three broad phases):
  - 1998–2002: Structural reforms—consolidation into a single revenue authority, centralized process functions, standardized service offerings, greater autonomy in HR policies.
  - 2001–2006: Rolling compliance campaigns and initiatives (e.g., Woodmead compliance campaign; Siyakha 1 and 2; hiring private-sector tax compliance experts; Large Business Centre created in 2005).
  - Since 2006: Modernization and automation of administrative processes, national compliance analysis introduced in FY 2008/09, national risk profiles for all taxpayers except LBC, enhanced analysis for LBC taxpayers, and policy measures to reduce compliance burdens (VAT registration threshold increased from "R300,000 pa" to "R1 million pa").
- Observed impacts consistent with SARS evaluations:
  - VAT gap trends align with timing of reforms: administrative disturbance prior to 2002 with high gap (25–30 percent); steady decrease to under 10 percent during 2002–2006 compliance initiatives; increase in 2008–2009 (likely due to financial crisis) then return to under 10 percent and broadly level over 2007–2012.
  - SARS reports improved service times, better audit results from automated risk processes, and greater data efficiency reducing auditor time on routine checking.

*Source: Staff estimates and SARS data as presented in the chapter.*

### 46. SARS are now using the national compliance model to monitor current compliance

### 46. SARS are now using the national compliance model to monitor current compliance

### Compliance monitoring and indicators
- The program of automation and modernization of processes included implementing the national compliance analysis indicators, referred to within SARS as, ‘the pillars of compliance’.
- Key performance indicators being monitored:
  - registration
  - filing compliance
  - payment compliance and credit returns
  - declaration compliance including audit results
  - net tax receipts
- Overall, the results of these indicators are consistent with a level tax gap since 2011 and increasing efficiency within SARS.

### VAT filing, payment, and credits
- Filing compliance (over 2008/09 – 2013/14):
  - Non-filing by active taxpayers has increased by 13.3 percent.
  - Overall filing rate (on time and late) fell from 79.1 percent to 65.8 percent.
  - SARS believes this may be partly due to increasing numbers of taxpayers recorded as active that are actually inactive.
- Payment compliance (over 2008/09 – 2013/14):
  - Payment compliance has risen from 81.1 percent to 85.1 percent.
  - NB: there are no penalties for late filing of VAT returns, but there are penalties for late or non-payment.
- Correlation:
  - Analysis of individual taxpayer segments shows a positive correlation between turnover and these compliance indicators.
- VAT credit returns (2008/09 to 2013/14):
  - Proportion of credit returns fell from 21 percent to 18.4 percent of all VAT returns.
  - Average value of excess credits rose from R182,735 to R317,869.
  - Total value of credit returns reached R157 million in FY 2013/14.
  - The increase in overall excess credits is attributed to increasing import in South Africa.

### Audit coverage and yield
- Audit counts (excluding verification checks):
  - Audits fell from 7,612 in FY 2011/12 to 3,565 in FY 2013/14.
- Audit findings and yield:
  - Percentage of audits finding inaccurate declarations rose from 51.7 percent to 62 percent.
  - Overall yield from audits rose from R4.5 billion to R6.1 billion.
  - The average yield per audit has risen very strongly.
  - Note: the increase in FY 2013/14 may be partly due to higher levels of noncompliance; there is no estimate yet of the VAT gap for that year.
  - These results indicate better targeting of audit resources and more effective auditing.

### Effects of the VAT registration threshold change and micro-businesses
- Policy change:
  - In FY 2007/08, the registration threshold for VAT was increased from R300,000 pa to R1 million pa.
  - A sustained campaign to de-register micro businesses trading below the threshold was implemented (though such traders retained the right to be registered on a voluntary basis).
- Impacts:
  - Noticeable reduction in the number of registered micro businesses.
  - Prior to the rise in the threshold, businesses with annual turnovers of less than R1 million had been receiving overall net VAT repayments worth around 2-3 percent of total VAT receipts.
  - Following the increase in the threshold, net repayment to such businesses initially rose to around 5 percent of total VAT receipts in FY 2008/09 as businesses liable for net VAT payments deregistered.
  - Net repayments to micro businesses fell to negligible levels in subsequent years, reducing the observed VAT gap by around 2-3 percent.
  - NB: SARS noted an increase in refunds paid to micro businesses in FY 2013/14, to about 5 percent of total VAT receipts.
- SARS response to compliance risk:
  - Implemented tighter controls on VAT registrations, including the scrapping of a planned introduction of online registrations.

### Non-VAT registered businesses (2001–2013)
- Survey notes:
  - Stats SA conducts regular surveys of employers and the self-employed at four year intervals.
  - The most recent survey shows a decline in the number of individuals running businesses trading below the VAT registration threshold from 2.3 million to 1.1 million in 2009, then an increase to 1.5 million in 2013.
  - There is a view that the 2001 result cannot be compared with other years because of a different sampling method, and the downward trend of non-VAT registrants could be exaggerated.
- Business-size distribution (percent by turnover band, 2001, 2005, 2009, 2013):
  - R0 - R100: 10.8, 11.4, 8.8, 8.2
  - R101 - R350: 29.5, 21.4, 14.5, 11.0
  - R351 - R750: 21.4, 23.3, 18.7, 13.8
  - R751 - R1 500: 15.3, 17.0, 19.9, 19.3
  - R1 501 - R3 000: 8.5, 12.2, 15.8, 18.5
  - R3 001 - R6 000: 6.0, 7.0, 10.8, 14.6
  - R6 001 plus: 3.7, 7.4, 11.5, 14.6
  - Unspecified: 4.9, 0.3, 0.0, 0.0
- Number of businesses (thousands):
  - 2,258 (2001), 1,668 (2005), 1,144 (2009), 1,517 (2013)
- Additional note:
  - Over 2001-2013 the average size of non-registered businesses has increased, but the great majority (85 percent in 2013) reported annualized turnovers less than R72 thousand, well below the registration threshold of R1 million.

### Recommendations
- SARS should continue to estimate the VAT compliance gap in South Africa, and further integrate their revenue analysis and compliance analysis.
- Both the revenue analysis and national compliance analysis currently being conducted by SARS represent good practice; SARS should review and update their compliance gap estimates as updated data becomes available.
- There is scope for more systemic integration of these analyses:
  - Use the compliance analysis to inform the revenue analysis.
  - Use the revenue analysis—particularly, a regularly updated compliance gap analysis—to set the strategic context for the national compliance analysis.
- SARS could monitor detailed time series of sectoral and other demographic aggregate returns and payments data to identify changing trends, step changes and spikes that might indicate new or emerging compliance risks.
  - Such monitoring should bring together both revenue and compliance analysts and operational experts.
  - This process should aid both better compliance risk management and better fiscal analysis.
- SARS could consider extending this integrated revenue and compliance analysis approach to other taxes.

*Source: _cr15180 - 46. SARS are now using the national compliance model to monitor current compliance*

### Appendix I. The RA-GAP Model and Methodology

### Appendix I. The RA-GAP Model and Methodology

### A. Introduction
- Purpose:
  - RA-GAP estimates potential tax revenues from macroeconomic data and measures the compliance gap by comparing potential VAT collections under current tax rules with full compliance to actual VAT collections.
  - The methodology breaks down revenues into economic sectors and traces gaps over time to identify whether fluctuations are due to changes in potential revenue or compliance issues in specific sectors.
- Conceptual distinctions:
  - Compliance gap: difference between potential revenue under current law with full compliance and actual revenue; treated as the representative indicator in RA-GAP.
  - Policy gap: supplementary indicator based on hypothetical tax legislation and analyses of effects due to changes in tax policies; used to compare magnitude with the compliance gap.
- Approach:
  - Top-down estimation: compares actual VAT collections to potential VAT collections estimated from macroeconomic statistics covering the whole VAT tax base. Advantages: (a) should cover all compliance losses whether or not separately identified; (b) results can be compared to costs of tax expenditures and reliefs.
  - Bottom-up approach: can supplement the top-down approach by estimating losses of individual behavioral components to help identify drivers of the total gap.
- Basis for VAT base estimation:
  - RA-GAP focuses on sectoral value added (output minus input) as VAT tax bases, enabling direct reflection of exemptions and different tax rates and facilitating sectoral matching of potential and actual revenues.
  - Alternative approaches (e.g., household surveys, demand data) may be preferable where data quality warrants; all macro-statistical approaches have error margins and risk overstating potential VAT because avoidance and legal interpretation issues may reduce revenue but appear as non-compliance without specific adjustments.
- Allocation of actual collections:
  - Use of cash collections linked to the periods in which tax due actually accrued is required; RA-GAP uses cash collection data reconciled to accrued-period activity by leveraging individual tax return and payment records.
  - Sectoral sorting of tax collection data enables identification of sectors with larger compliance gaps and insight into nature/location of non-compliance.
  - Assessment data (amounts assessed but not yet collected) should be considered to analyze causes of changes in compliance and needs for system improvements.

### B. Measuring Potential Revenues for a Value-Added Tax
- Model design:
  - The model estimates taxable value-added across all sectors using supply-use tables or input-output tables to compute tax on imports by sector plus tax on sectoral output less input tax credits due to the sector. (See footnote 32.)
- The value-added based potential revenues model is:
 ܸܲ
௦
ൌ	෍
ሺ
ܯ
௖
௦
߬ൈ
௖
ሻ
ݎൈ
௦
௖
	൅	൥෍
ሺ
ܻ
௖
௦
ܺെ
௖
௦
ሻ
߬ൈ
௖
௖
ݎ൩ൈ
௦
		െ		൥෍
ሺ
ܰ
௖
௦
ܫ൅
௖
௦
ሻ
߬ൈ
௖
௖
ݎ൩ൈ
௦
ൈሺ1െ݁
௦
ሻൈ	ߟ
௖
௦
- Definitions of model variables:
  - ܸܲ
௦
 = the potential net VAT for a sector.
  - ܯ
௖
௦
 = imports by sector s of commodity c.
  - ܻ
௖
௦
 = output by sector s of commodity c.
  - ܺ
௖
௦
 = exports by sector s of commodity c.
  - ܰ
௖
௦
 = intermediate demand (consumption) by sector s of commodity c.
  - ܫ
௖
௦
 = investment by sector s of commodity c.
  - ߬
௖
 = the VAT rate that applies to commodity c (zero if zero-rated or exempt).
  - ߟ
௖
௦
 = the proportion of input tax credits for commodity c by sector s allowed to be claimed.
  - ݎ
௦
 = the proportion of output for a sector produced by registered businesses.
  - ݁
௦
 = the proportion of output for a sector which is exempt output.
- Data sourcing and determination:
  - Y, X, M, N, and I: obtained from components in statistical supply-use (or input-output) tables; X and M (external trades) require adjustments (see below).
  - ߬
௖
 (policy variable 1): obtained from the tax rate structure for each commodity except trade services; for hypothetical reference tax structure the standard rate is assigned to the full vector ߬
௖
 except supplies typically exempted internationally (margin-based financial services, life insurance, and residential rents).
  - ߟ
௖
௦
 (policy variable 2): determined by statutory limitations on input tax credits; defaults to 1; for reference tax structure all values set to 1.
  - ݎ
௦
: estimated in conjunction with authorities, using business licensing or Customs transactions data. (See footnote 33.)
  - ݁
௦
: proportion taxable is function of ߬
௖
; computed as ݁
௦
 = ∑(ܻ
௖
௦
′߬ൈ
௖
)/∑(ܻ
௖
௦
) where ′߬
௖
 distinguishes exempt (′߬
௖
 = 1) from taxable (′߬
௖
 = 0). (See footnote 34.)
- Adjustments for X and M:
  - Exports must be adjusted to remove domestic consumption by non-nationals; imports must be adjusted to remove consumption abroad by nationals. (See footnote 35.)
- Trade sector statutory rate:
  - Weighted average statutory rate for retail and wholesale trade services (߬
்
) determined by trade margins by commodity type:
  ߬
்
 = [∑(߬
௖
ᇲ
ܭൈ
௖
ᇲ
)] / [∑(ܭ
௖
ᇲ
)]
  - ߬
௖
ᇲ
 = statutory rate for commodity ܿ
ᇱ
 (includes all commodities but trade services); ܭ
௖
ᇲ
 = trade margins for commodity ܿ
ᇱ
.
- Accommodating policy complexities:
  - Sector-specific tax rates: can use sector-by-commodity matrix of tax rates ߬
௖
௦
 instead of commodity vector ߬
௖
; imports continue to use ߬
௖
. ݁
௦
 calculation adjusted to include ܻ
௖
௦
߬ൈ
௖
௦ᇱ
 term where ߬
௖
௦ᇱ
 is a matrix indicating exempt commodities for a sector.
  - Transaction-specific treatments:
    - Taxpayer-to-taxpayer transactions: either split commodity into components (requires adding new commodity to supply-use tables and data on transaction values) or ignore (no net impact on overall gap; affects sectoral allocation).
    - Taxpayer-to-final consumer transactions: final potential VAT from retail sector reduced by external estimates of tax expenditure cost.
- Noted model limitations and assumptions:
  - Macro-statistics have error margins and may overestimate potential VAT by not capturing avoidance/legal interpretation differences—these may be included in compliance gap though they require policy or litigation responses.

### C. Measuring Actual Collections
- Concept:
  - RA-GAP measures actual tax collections from the same economic activities used to estimate potential revenues, requiring reallocation of cash collection data into periods when tax due actually accrued to derive "accrued collections." (See footnote 36.)
- Accrued collections formula:
  ܸܣ
௦
	ൌ	ܥ
௦
ܲ	൅
௦
ܴ	െ
௦
ܱܲ	ሺ൅	
௦
ሻ
- Definitions of accrued collections variables:
  - ܸܣ
௦
 = accrued VAT collections for the period.
  - ܥ
௦
 = collections at customs in the period.
  - ܲ
௦
 = payments received for the period.
  - ܴ
௦
 = excess credit accrued for the period.
  - ܱܲ
௦
 = payments offset by excess credit (excess credit carried forward to offset tax due, or excess credit accrued for the period used to offset tax owing for past periods).
- Data sources and required fields:
  - ܥ
௦
 (customs collections): from customs declaration database; requires value of VAT payments on imports, date of entry for declaration/payment, and sector of taxpayer making the declaration.
  - ܲ
௦
 (payments received): from payments transaction database; requires value of VAT payments (exclusive of interest or penalties), date of payment, tax period for which payment applies, and sector of taxpayer.
  - ܴ
௦
 (excess credit accrued): from tax returns database; requires value of excess credit, tax period the excess credit return was submitted for, date of filing, and sector of the taxpayer. (See footnote 37.)
  - ܱܲ
௦
 (payments offset by excess credit): applies where excess credit can be carried forward or used to offset past liabilities; data from tax return database with related tax period and sector. (See footnote 39.)
- Additional nuances:
  - Proper measurement of excess credit may require recomputing net tax for the period from fundamental return line items: output tax on supplies made in the period + self-assessed VAT on imports − VAT paid on inputs used in making taxable supplies. This recomputation may need to be done taxpayer by taxpayer. (See footnote 38.)
  - If excess credit is used to offset other tax obligations, it should be recognized as a reduction in net VAT collections.

### D. Measuring and Reporting the Compliance Gap
- Definition:
  - Compliance gap = current potential collections (from B) − actual collections (accrued collections from C). As accrued collections change over time, the gap value changes.
- Two standardized RA-GAP measures for comparability:
  1. Compliance gap at the time of filing:
     - Measured at the original filing/payment deadline.
     - In accrued collections, ܲ
௦
, ܴ
௦
, and ܱܲ
௦
 are filtered to select payments and returns received before their appropriate deadlines.
     - Tax return data for ܴ
௦
 and ܱܲ
௦
 are the data as originally submitted by the taxpayer. This measure is intended to capture voluntary compliance and will not change over time. (See footnote 40.)
  2. Compliance gap at the time of estimation:
     - Measured using the latest available data for returns filed, assessment values, and collection and refund payment values, ideally annually at the anniversary of the last filing/payment deadline for a tax year.
     - Data for ܲ
௦
 filtered to select payments made by the measurement date; tax returns data for ܴ
௦
 and ܱܲ
௦
 are the current assessed values as of that date. This value will change year to year; the measured value at a point in time is static for that comparison. Comparing changes over time provides insights into collection performance. (See footnote 41.)
- Reporting conventions:
  - RA-GAP commonly reports the compliance gap as CPV − AV divided by CPV (i.e., the compliance gap as a percentage of current potential revenues) to facilitate comparisons over time and across jurisdictions.
  - Compliance gaps are also expressed as percentages of GDP to provide common basis for comparison with economic activity and policy gap magnitudes. (See footnote 42.)

*Source: Appendix I. The RA-GAP Model and Methodology (Content unit: _cr15180).*

### Appendix II. Application of RA-GAP Model to South Africa

### Appendix II. Application of RA-GAP Model to South Africa

### A. Introduction
- Potential VAT revenues (PV1 and PV2) are calculated for each year between 2007 and 2012 using supply and use tables (62 sectors and 104 commodities) of each year with the policy parameters specified in section B.
- Actual VAT collections for each year are calculated as the sum of:
  - [1] payments reallocated to the tax period starting each calendar year,
  - [2] excess credit declared in the tax period starting each calendar year, and
  - [3] payment at Customs during each calendar year.

### B. RA-GAP model for potential VAT revenues
- PV1 (potential VAT revenues under current policy structure) uses ratios to specify amounts of exempted and zero-rated products based on consumption share in the IES2010/11.

- Exemptions (CPC codes and ratios)
  - CPC_239 Food n.e.c. 0.0637
  - CPC_64 Passenger transport services 0.9097*
  - CPC_711 Financial services 1
  - CPC_72 Real estate services 0.9818
  - CPC_92 Education services 0.9690
  - Note: For CPC_64, the production by I42 (Land transport, transport via pipe lines) and I61 (Other activities) is treated as exempted.

- Zero-rating (CPC codes and ratios)
  - CPC_01 Products of agriculture, horticulture and market gardening 0.8799
  - CPC_02 Live animals and animal products (excluding meat) 0.5048
  - CPC_212 Prepared and preserved fish 0.4730
  - CPC_213 Prepared and preserved vegetables 0.2090
  - CPC_215 Animal and vegetables oils and fats 0.2924
  - CPC_22 Dairy products and egg products 0.3796
  - CPC_231 Grain mill products 0.7427
  - CPC_233 Preparations used in animal feeding 1*
  - CPC_234 Bakery products 0.4924
  - CPC_33 Coke oven products, refined petroleum products 0.9792
  - CPC_346 Fertilizers and pesticides 1*
  - Note: For CPC_233 and CPC_346, only transactions for intermediate consumption are treated as zero-rated.

- Treatment of government
  - Public administration services produced by the government are used in three categories: government final consumption, final consumption by households, and intermediate consumption.
  - Government final consumption is regarded as non-taxable (notional value of services provided to general public without actual transactions).
  - Other transactions (final consumption by households and intermediate consumption) are regarded as taxable transactions.
  - Public administration services produced by LG are zero-rated and can claim input tax credit for their purchase.
  - Services produced by CSG are treated as exempted and cannot claim input tax credit.
  - Production of public administration services by the government sector (I55) is divided into production by central and state governments (CSG) and production by local governments (LG) according to the ratio of intermediate consumption for each year.

- Treatment of informal sector
  - In supply and use table, Sector I62 (non-observed, informal, non-profit, households) represents economic activities by small entities estimated from data of business entities under the threshold of VAT; production of I62 is regarded as exempted in the model.
  - Production by I62 can be used in three categories: final consumption of households, intermediate consumption, and capital formation.
  - It is assumed trade services are directly used by final consumers, and other goods and services are used primarily as intermediate consumption and capital formation.

- Treatment of restriction on input tax credits
  - Input tax credit for accommodation, catering services and purchase of motor vehicle (except for I42 (land transportation services) and I51 (renting of machinery and equipment)) are not allowed.

- Sector allocation of capital formation, import and export
  - Capital formation is allocated to sectors using National Bank’s data for sector capital formation.
  - Import and export are allocated to sectors using Customs data in 2012.

### C. Measurement of actual collections for VAT
- Aggregate cash collection
  - Cash collection data are based on monthly collection data provided by SARS and consist of gross cash collection for domestic VAT, refund, and cash collection for import VAT.

- Accrued collection
  - Accrued collection is calculated as the sum of:
    - [1] payments reallocated to the tax period starting each calendar year,
    - [2] excess credit declared in the tax period starting each calendar year, and
    - [3] payment at Customs during each calendar year.
  - Reallocation of payments to tax period is based on individual payment data from 2007 to 2012, and reallocation of excess credit is based on individual VAT declaration data for the period.

- Treatment of diesel refund
  - Diesel refund system allowed some specific sectors to claim refunds of fuel levy and Road Accident Fund levy as credits against their VAT liability.
  - Before FY2013, such refunds had been recorded as reductions of VAT net payments; amounts subtracted from reported net VAT collections were added back to actual collections to show underlying VAT revenue trends.
  - After FY2013, these claims have been included as a separate item in tax collections, so adjustment of reported figures is not necessary.

### D. Data used in the RA-GAP model
- For estimating Potential VAT revenues:
  - Supply and use tables, 2007 - 2012 (Stats SA; provided in August 2014)
  - Quarterly Bulletin, June 2014 (South African Reserve Bank; downloaded in August 2014)
  - Income and Expenditure Surveys (IES) 2010/11
  - Customs declaration for Trade Statistics 2010 – 2013 (SARS; provided in August 2014)

- Tax collections data (Aggregate):
  - Tax Statistics 2008 – 2014 (published by NT and SARS)
  - Monthly cash collection of VAT 2005/06 – 2013/14 (SARS; provided in August 2014)
  - Diesel refund data 2008 – 2013 (SARS; provided in September 2014)

- For calculating Accrued VAT collections (Anonymized):
  - VAT Payments and Refunds from January 2007 to July 2014 (SARS; provided in August 2014)
  - VAT Returns from January 2006 to July 2014 (SARS; provided in August 2014)
  - VAT Audit from 2007 to 2013 (SARS; provided in August 2014)
  - Customs Import VAT declaration from 2010 to 2013 (SARS; provided in August 2014)

---

### Appendix III. Comparison of RA-GAP Results with Analysis by SARS

### SARS national compliance analysis (overview)
- Introduced in FY 2008/09 as a performance management framework monitoring key performance indicators of compliance for VAT, collectively known as ‘the pillars of compliance’:
  - Registrations
  - Filing compliance
  - Payment compliance and excess credit returns
  - Declaration compliance
- The national compliance analysis also assesses compliance risk by industrial sector. VAT receipts are monitored by SARS’ revenue analysis team.

### Registrations
- New registrations declined sharply in FY 2009/10 and remain low relative to the taxpayer register, with a small rise in 2013/14.
- New registrations fell from 64.4 thousand in 2008/09 to 26.6 thousand in 2009/10.
- Registrations continued to decline, rising slightly in 2013/14; they now stand at 26.6 thousand, out of a total VAT register of 662 thousand vendors, about 4 percent.
- SARS attributes low churn to tightened controls on new registrations and a high registration threshold of R1 million.

### Filing compliance
- Filing compliance has declined by 13.3 percent since FY 2008/09.
- Filing on time rates have remained at 55-60 percent of active VAT taxpayers since FY 2008/09.
- Overall filing compliance declined from 79.1 percent to 65.8 percent.
- Taxpayers with higher turnovers are more compliant than those with lower turnovers; filing compliance for LBC vendors is 87.9 percent.
- Industrial sector with lowest filing compliance: construction at 45.2 percent.

### Payment compliance
- Payment compliance is higher than filing compliance, at 85 percent.
- Payment compliance has remained fairly steady at around 80-85 percent.
- Vendors with higher turnovers are more compliant; payment compliance of LBC vendors is 95.7 percent.
- Highest noncompliance found in:
  - public administration (10.3 percent)
  - construction (7.2 percent)

### Credit returns
- Credit returns as a proportion of VAT returns have decreased slightly since FY 2008/09, but the value of such returns has increased.
- Decline in the number of credit returns may be caused by the registration threshold increase.
- Increase in the overall value of refunds claimed is attributable to increases in refunds to public administration reflecting increased government investment.

### Audits
- Number of audits decreased since FY 2011/12, but overall yield from audits increased.
- Targeted audits found inaccuracies in 62 percent of cases with an average yield of R1.7 million per audit.
- Number of audits decreased from 7,612 in FY 2011/12 to 3,565 in FY 2013/14.
- Total yield from audits increased from R4.5 million to R6.1 million over the same period.
- Observation: Could indicate better targeted and more effective audits or greater levels of noncompliance (VAT gap for 2013/14 has not been estimated).

### Declaration compliance and sector risks
- Highest levels of declaration inaccuracy:
  - public administration: 71.1 percent
  - construction: 64.8 percent
- Highest yielding noncompliance behaviors:
  - under-declaration of standard rated outputs
  - over-declared zero rated outputs (exports)
  - undisclosed income on secondary bank accounts
- Highest risk sectors: construction, personal and household services, and agencies and other services (unchanged from FY 2012/13).
- Of these three, construction and agencies and other services are in the high volume, high revenue category of industries in South Africa.

*Appendix II. Application of RA-GAP Model to South Africa*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2015/_cr15180.pdf_