## 1. The VAT Impact on Low Income Individuals in LICs and EMEs

## Source details

**Canonical URL:** [1. The VAT Impact on Low Income Individuals in LICs and EMEs](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021040-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021040-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021040-print-pdf.pdf.json)

---

### I. Introduction and objectives
- VAT presence and share:
  - VAT is present in more than 160 countries worldwide, representing 26, 33, and 31 percent of total tax in-take in developed, emerging markets, and low-income countries respectively.
- Two main challenges in LICs and EMEs:
  - Non-compliance and evasion as sources of revenue leakage.
  - The impact of the VAT on low-income individuals.
- Tools analyzed:
  - Tools to encourage issuance of invoices or use of payment systems (improve VAT control).
  - Tools to reduce the VAT impact on low-income individuals (PIT credit, VAT rate reduction, targeted transfers).
- Aims:
  - Provide a clear characterization of these tools and their variants.
  - Offer guidance based on empirical studies and novel data analysis.

### II. Tools to encourage invoicing or formal payment systems — VAT paid on personal consumption creditable/deductible against the PIT
- Countries allowing VAT on personal consumption as a credit or deduction against the PIT: Bolivia, Ecuador, Guatemala, Paraguay.
- Selected country parameters (as presented):
  - Bolivia: VAT rate 13; PIT threshold (USD) 7,200; PIT rates (%) 13; VAT Deduction or credit: Credit.
  - Ecuador: VAT rate 12; PIT threshold (USD) 11,290; PIT rates (%) 0; 5; 10; 12; 15; 20; 25; 30; 35; VAT Deduction up to USD 14,703.
  - Guatemala: VAT rate 12; PIT threshold (USD) 6,275; PIT rates (%) 5; 7; Deduction up to USD 2,000.
  - Paraguay: VAT rate 10; PIT threshold (USD) 17,278; PIT rates (%) 8; 9; 10; No deduction limit.
- Design variants:
  - Credit of VAT paid on personal consumption against PIT liability (e.g., Bolivia’s VCR).
  - Deduction of personal consumption from PIT taxable base (e.g., Paraguay, Ecuador, Guatemala historically).

### III. Theoretical analysis — expected effects and distortions
- Rationale and mechanism:
  - A comprehensive PIT with credit for VAT would tax income without additional tax on consumption; VAT on registered trader’s goods creditable; consumption from unregistered traders yields no VAT credit.
- Distributional and behavioral distortions:
  - Personal allowances in PIT mean high-income individuals pay PIT while low-income individuals often only pay VAT, potentially increasing regressiveness.
  - Low-income individuals with no PIT liability cannot benefit from VAT credits/deductions, incentivizing purchases from informal establishments.
  - The policy is distortive: encourages current consumption and can reduce savings.
  - Under progressive PIT schedules, deductions are more valuable at higher marginal rates, increasing regressivity.
- Administrative concerns:
  - Difficulty controlling large numbers of invoices deducted against PIT; increased risk of fake invoices.
  - High control costs and significant allocation of tax administration resources to invoice verification.
  - Bolivia example: Bolivia PIT (VCR) revenue reduced from 0.27 percent of GDP in 2005 to 0.15 percent in 2012.

### IV. Empirical findings — revenue and distributional impacts (Ecuador, Guatemala, Bolivia)
- Cross-country tax revenue context (percent of GDP):
  - Bolivia: Total Tax Revenue 23.8; PIT Revenue 0.2; VAT Revenue 7.8; PIT and VAT Revenue 8.0.
  - Ecuador: Total Tax Revenue 14.3; PIT Revenue 1.0; VAT Revenue 6.5; PIT and VAT Revenue 7.5.
  - Guatemala: Total Tax Revenue 10.4; PIT Revenue 0.4; VAT Revenue 4.8; PIT and VAT Revenue 5.2.
  - Paraguay: Total Tax Revenue 10.3; PIT Revenue 0.1; VAT Revenue 5.1; PIT and VAT Revenue 5.2.
  - Lower middle income: Total Tax Revenue 16.7; PIT Revenue 2.6; VAT Revenue 6.0; PIT and VAT Revenue 8.6.
  - Upper middle income: Total Tax Revenue 18.5; PIT Revenue 2.6; VAT Revenue 5.9; PIT and VAT Revenue 8.5.
  - High income: Total Tax Revenue 21.7; PIT Revenue 7.2; VAT Revenue 7.0; PIT and VAT Revenue 14.2.
  - Africa: Total Tax Revenue 16.6; PIT Revenue 2.9; VAT Revenue 5.3; PIT and VAT Revenue 8.3.
  - Asia Pacific: Total Tax Revenue 17.6; PIT Revenue 3.0; VAT Revenue 5.0; PIT and VAT Revenue 8.0.
  - Europe: Total Tax Revenue 25.1; PIT Revenue 7.1; VAT Revenue 8.1; PIT and VAT Revenue 15.2.
  - Middle East and Central Asia: Total Tax Revenue 13.1; PIT Revenue 2.5; VAT Revenue 6.7; PIT and VAT Revenue 9.2.
  - Western Hemisphere: Total Tax Revenue 18.4; PIT Revenue 2.9; VAT Revenue 5.3; PIT and VAT Revenue 8.3.
- Ecuador — microsimulation results (2012/13 household survey updated; calibrated to 2019 PIT revenue):
  - Eliminating the PIT deduction of personal consumption would increase PIT revenue by 0.4 percentage points of GDP.
  - Eliminating the deduction improves the Reynolds-Smolensky after-tax income Gini by 0.5 Gini points.
  - Scenario reallocating additional revenue to first three deciles yields an improvement of 1.6 Gini points.
  - Table 3 summary for Ecuador:
    - Current PIT system: PIT revenue (% of GDP) 0.9.
    - After policy change: PIT revenue (% of GDP) 1.3.
    - Net Effect: 0.4.
    - After-tax Income Gini: Current PIT system 50.1.
    - After elimination of the deduction of personal consumption: 49.6 (improvement 0.5).
    - After elimination and redistribution to poorest three deciles: 48.52 (improvement 1.6).
- Guatemala — VAT credit elimination (May 2012) evidence:
  - VAT credit eliminated May 2012 (replaced by fixed deduction about US$ 2,000).
  - VAT revenue (percent of GDP) began decreasing in 2011 and continued until 2016 while VAT credit was substituted in 2012.
  - Econometric analysis summary:
    - Change in VAT revenue as share of GDP was not statistically different from zero after the policy change once controlling for household final consumption, imports, and tax expenditures.
    - Changes in VAT collection were statistically explained by changes in imports and final consumption, not by the reform.
- Bolivia — administrative outcome:
  - Bolivia’s VCR allowed employees to claim VAT paid on any personal consumption against VCR liability; an additional credit equivalent to 13 percent of two minimum wages if purchases from small taxpayers who cannot issue VAT invoices.
  - Bolivia experienced a decline in PIT (VCR) revenue from 0.27 percent of GDP (2005) to 0.15 percent (2012), mainly due to problems of control.

### V. Policy implications and recommendations (based on analysis)
- Key insights:
  - Allowing VAT paid as credit or deduction against PIT tends to favor better-off taxpayers and can reduce redistributive capacity.
  - The tool is distortive of consumption-savings decisions and can incentivize consumption from informal providers.
  - Administrative control is costly and susceptible to fraud (fake invoices), reducing PIT revenue and complicating administration priorities.
  - Empirical evidence does not show robust improvements in VAT revenue from PIT-based invoice-incentive tools once other factors (imports, consumption, exemptions) are considered.
- Suggested policy directions:
  - Eliminate VAT-as-PIT-credit/deduction to regain revenue and reduce distortions where PIT rates are already high and administrative costs are substantial.
  - To protect low-income individuals prefer well targeted transfers rather than PIT deductions or credits for personal consumption.
  - Example: In Ecuador, elimination of the deduction coupled with targeted transfers to the three lowest income deciles yields stronger improvements in equality (1.6 Gini points) than retaining the deduction.
  - Avoid use of VAT credit or personal expenses as deduction of the PIT except when targeted toward low-income individuals or with a maximum amount (as in Portugal).
  - Avoid reductions in the VAT rate to encourage use of credit and debit cards: they are regressive, expensive, and lack evidence of improving VAT revenue.
  - Prefer targeted low-income programs (e.g., Uruguay and Canada approaches) when administrative capacity allows and when low-income purchases occur in formal markets or consumed goods/services are not VAT-exempted.
  - Maintain lottery programs only if benefits (increased invoicing, reduced non-compliance) significantly exceed costs; eliminate when objectives are achieved (example: Korea’s cancellation in 2015).

### Non-compliance in Guatemala (post-reform findings)
- Finding: Non-compliance VAT rates have been rising in Guatemala since the policy change took place (time span referenced: 2008–2018).
- Econometric result: The rise in non-compliance was not statistically explained by the reform dummy once controls included final consumption, imports, and tax expenditures.
- Additional observations:
  - Non-compliance rates had been higher during the period the VAT was allowed as a PIT credit.
  - The increase in non-compliance is more likely related to corruption scandals affecting the Guatemalan government and increased smuggling.

### The Reduction of the VAT Rate for Using Electronic Means of Payments
- Policy description: Reduced VAT rates for purchases using electronic payments to discourage cash use, foster financial transactions through banks, and facilitate tax control.
- Equity concern: The tool is regressive because the poorest—often outside the banking system—are left out.
- Country experiences:
  - Argentina: VAT rate reduction of 5 percentage points for debit card purchases until 2016 and 3 percentage points for credit card payments until 2003; both benefits eliminated because they entailed revenue losses and were regressive.
  - Honduras: Refund equivalent to 8 percent of the VAT included in the invoice; benefit remains despite favoring richer households.
  - Pakistan: VAT rate reduced from 15 to 12 percent for payments with debit or credit cards; benefit remains despite favoring richer households.
- Gender note: In Honduras and Pakistan, male individuals have greater access to debit and credit cards, potentially enabling greater benefit for men.

### Korea: Cash Receipt System (2005–2015) — design and outcomes
- Design:
  - Incentives for merchants and unregistered final consumers to report cash sales electronically via seller terminals generating cash receipts sent to National Tax Service.
  - Vendor benefit: credit 1 percent of cash turnover against VAT liability for each receipt issued (until 2009); from 2010 a fixed amount was granted per receipt.
  - Consumer benefit: credit against income tax liability if merchant name and income tax file number provided.
- Outcomes:
  - Program successful in reducing underreported cash sales; cancelled in 2015.
  - Context: in 2014, credit cards accounted for over 64.1 per cent of consumption expenditure; debit cards accounted for 12.4 per cent (Krever, 2014).
  - Corporate adoption: corporate taxpayers using E-invoice or cash receipt system increased to 99.4 percent of total CIT taxpayers in 2014; sales issued with these systems increased to 99.9 percent of total sales.
  - Survey: 69.4 percent of tax practitioners and business taxpayer employees agreed or strongly agreed that tax evasion would have declined.

### Lottery Programs — rationale, costs, and country evidence
- Mechanism: Consumers request invoices in exchange for lottery entries; may include VAT refund or rate reduction.
- Costs and benefits:
  - Costs: prizes, revenue loss from VAT rate reduction, advertisement, administrative costs.
  - Benefits: improved monitoring of transactions via invoices and potential increases in VAT and CIT revenues.
  - Consumer cost: welfare loss due to time invested.
- Country evidence highlights:
  - Brazil and Mongolia: VAT refunds within programs complicate attribution; VAT refunds reach about 20 percent of total VAT paid in Mongolia and about 20 percent of VAT paid by final sale establishment in Brazil.
    - Brazil participation: started 67,000 (2008), peaked 277,000 (April 2015), declined to around 100,000 after refund cutbacks.
    - Naritomi (2019): VAT revenue increased by 9.3 percent net of rewards in 2008–2011 (did not include administrative and welfare costs).
    - Mattos et al (2013): limited impact; no robust effect on total tax collection.
  - Slovakia: modest impact; fiscal impact estimated about EUR 8 million annually (extrapolated); total cost in 2014 EUR 1.6 million including marketing.
  - China: sales tax revenue significantly higher (over 17.1 percent) in districts with lottery; real growth rates of sales tax and total tax revenues over 21.5 percent and over 10.4 percent, respectively, in areas with the lottery (Wan, 2010).
  - Portugal: VAT non-compliance decreased from 16 to 12 percent between 2013 and 2014 after lottery introduction, but causality not established amid other reforms; invoices with consumer taxpayer number sent to tax authorities increased by 36.3 percent in 2014 and total sales invoices increased by 51.2 percent one year later.
  - Mongolia (2016): app-based system with 2 percent refund; VAT revenues increased to 6.6 percent of GDP in 2019 from 4.8 percent in 2016 after program introduction, but concurrent large GDP and mining investment growth complicate attribution; government prize and admin costs were high.
- Program durability: effectiveness often declined when prizes and advertisement decreased; some programs eliminated when objectives achieved or costs exceeded benefits.

### Tools to Reduce the Impact of the VAT on Low-Income Individuals — Tax Credit to Low Income Taxpayers
- Rationale: VAT perceived as regressive because poorer households consume larger share of income; measures include reduced rates, exemptions, VAT thresholds, and targeted VAT credits.
- VAT credit feature:
  - Targets exclusively low-income households or “deserving” populations; more targeted than reduced rates, VAT thresholds, or refundable VAT available to all PIT taxpayers.
- Canada — GST/HST credit example:
  - Only low-income PIT taxpayers may claim GST/HST credit; paid quarterly to individuals and families with low incomes.
  - CRA automatically estimates credit when taxpayers file tax returns online.
  - Estimation factors: number of eligible children, adjusted family net income (considering rent, energy, property taxes), residence, and provincial benefits.
  - Example amounts (October 2019, Ontario):
    - Couple with total annual income $25,731 CAD and two children (ages 4 and 8): GST credit about $221.5 CAD per quarter (plus $102.7 CAD for Ontario Sales Tax).
    - Same family with income $70,800 CAD: federal GST credit about $59.0 CAD and Ontario Sales Tax credit about $31.9 CAD.
  - Design note: Credit amount is not based on GST/HST paid and decreases as income increases.
  - Refundability: The GST/HST credit is refundable; requires near universal filing but does not require PIT liability.
- Effectiveness evidence:
  - Godbout et al, 2011 findings:
    - Without the credit, GST/HST are almost proportional to household income in Canada; with the credit, the effective tax rate rises with income.
    - The tax credit is more effective for improving progressivity of consumption taxes than zero-rating of basic groceries.
- Implementation considerations:
  - Canada’s GST/HST credit operates with high labor formalization and effective control.
  - Most LICs and EMEs need stronger social safety nets and taxpayer registers to control such tools.
  - Uruguay cited as an EME example that can implement targeted VAT-credit-like measures.
- Relative impact point:
  - VAT has a larger impact on low-income individuals in developed countries than in LICs and EMEs because:
    - Higher proportion of consumption from self-production among the poorest.
    - Many purchases in informal markets.
    - Small taxpayers often registered in VAT-exempt special schemes.
  - This affects determination of VAT refund levels.

### Box 1 — VAT incidence, targeted programs, and Guatemala empirical analysis
- VAT incidence mechanisms reducing regressivity:
  - Exemptions/reduced rates on staple goods; high informal/untaxed purchases by poor; self-production of food not taxed.
- Country evidence:
  - Jenkins et al, 2006 (Dominican Republic): VAT burden progressive over all quintiles of household expenditure.
  - Mariscal and Verner, 2017 (Mexico) and Fabrizio et al, 2017 (Uganda): VAT not regressive.
  - Other cited results: neutral impact in Honduras (Fabrizio et al, 2017), Namibia (World Bank, 2017), Ethiopia (Hills et al, 2017), Togo (IMF, 2017).
- Targeted refund/credit mechanisms — case studies:
  - Uruguay — Programa Tarjeta Uruguay Social:
    - Operating since May 2006; transfers via magnetic card to 60 thousand households.
    - Monthly amounts vary; in 2019 a household with two children received URU$ 1,610 (about US$ 40.0) or URU$ 3,220 (about USD 80) per month for the worst-off households.
    - Beneficiaries selected using Critical Deficiency Index (ICC); purchases exclude tobacco and alcoholic beverages.
  - Thailand — State Welfare Smartcard (2018):
    - From November 2018 to April 2019, cardholders had 5 percentage points of the 7 percent VAT rate reimbursed up to USD 16 per month.
    - After launch there were 14.5 million smartcard holders.
    - Government transferred around 1,168 thousand USD to e-wallets and around 223 thousand USD to savings accounts (November 2018–April 2019).
    - From November 2019 to September 2020 (11-month period), government transferred around USD 2,665 thousand to e-wallets on the State Welfare Card.
  - Portugal — VAT creditable to PIT with limits:
    - Credit equal to 15 percent of VAT incurred in selected services with a maximum of EUR 250 per family member.
    - Eligible VAT must come from electronic invoices for specified services; design targets activities prone to cash transactions to reduce regressiveness and encourage invoice requests.
    - Distribution example: ETR for individual with income EUR 7,091 decreases from 14.5 to 11.0 percent; for income EUR 200,000 reduction is 0.1 percentage points.
- Guatemala empirical analysis (Appendix I) key points:
  - Decree 10-2012 (April 2012): VAT invoices no longer creditable against PIT for salaried workers; VAT invoices up to Q.12,000 (USD 1,530) allowed as deduction; PIT standard deduction increased to Q.48,000 (USD 6,120); combined total deduction Q.60,000 (USD 7,650) per year.
  - Observed post-reform trends: VAT non-compliance rates increased and VAT revenue declined, but regression analysis finds no statistical evidence that the reform dummy explains changes in VAT revenue or non-compliance once imports and final consumption are controlled for.
  - Regression framework and results highlights:
    - Annual regressions: percent change in final consumption coeff 0.993* (0.546); percent change in imports coeff 0.707*** (0.12); Dummy -0.00261 (0.0148).
    - Monthly regressions: percent change in IMAE coeff 2.444*** (0.498); Dummy 2.949 (7.181).
    - VAT non-compliance regressions: coefficients statistically insignificant; reform dummy not significant.

### Appendix II — Impact on income distribution of exemptions and reduced rates (summary)
- Main arguments:
  - VAT often characterized as regressive because propensity to save rises with income.
  - Reduced rates and exemptions are poorly targeted: they often reduce regressivity on average but transfer larger absolute benefits to higher-income households.
  - Equity objectives are better achieved via spending instruments financed by VAT revenue.
- Empirical targeting evidence (quintile shares of benefits from differentiated rates / exemptions examples):
  - 20 OECD countries: quintile shares 13.5; 17.3; 20.0; 22.5; 26.7.
  - Honduras: 7.3; 12.4; 17.1; 22.5; 40.7.
  - Nicaragua: 4.0; 7.0; 10.2; 23.7; 55.1.
  - Colombia: 5.6; 9.2; 13.3; 21.0; 50.9.
  - Dominican Republic: 10.3; 14.0; 16.8; 21.8; 37.0.
  - Mexico (zero rate on basic foods only): 10.8; 17.2; 21.2; 23.1; 27.9.
  - South Korea: 6.3; 13.8; 20.2; 25.7; 33.9.
  - Morocco: 14.5; 12.0; 15.0; 19.7; 38.8.
  - Tunisia: 7.9; 12.7; 17.0; 23.2; 39.2.
- Factors moderating measured regressivity:
  - VAT registration thresholds (small retailers exempt) and self-produced food reduce measured regressivity where data permit.
  - Top decile may have less untaxed consumption than bottom decile (Bachas, 2019).
  - Measuring incidence on aggregate consumption rather than current income can yield VAT neutrality or slight progressivity (Alavoutunki, Haapanen, Pirttila, 2019).
- Findings from country CEQ studies (selected):
  - Cameroon (2018): small adverse impact on distribution; removal of VAT exemptions needed to boost inequality-reducing transfers.
  - Ghana (2018): VAT reduced rates and exemptions costly and mildly progressive; cash transfers fairly well targeted.
  - Ethiopia (2017): VAT reduced rates and exemptions costly and regressive; cash transfers poorly targeted.
  - South Africa (2017): VAT mildly regressive; social spending progressive and significant.
  - Togo, Uganda, Zambia, Nigeria, Senegal: varied results but common theme that social transfers are generally more effective for equity than VAT rate differentiation.
- Policy implications:
  - Reduced rates and exemptions are blunt, poorly targeted redistribution tools; they often favor higher-income households in absolute terms.
  - Equity better pursued via well-targeted cash and in-kind transfers financed by VAT revenue.
  - Policymakers should assess cost and targeting efficiency of VAT rate differentiation relative to direct transfers and account for informal-sector dynamics and self-consumption when evaluating VAT incidence.

*Source: IMF staff (content from "1. The VAT Impact on Low Income Individuals in LICs and EMEs").*

### 1. The VAT Impact on Low Income Individuals in LICs and EMEs ____________________________________ 19

### 1. The VAT Impact on Low Income Individuals in LICs and EMEs

### I. Introduction and objectives
- VAT is present in more than 160 countries worldwide, representing 26, 33, and 31 percent of total tax in-take in developed, emerging markets, and low-income countries respectively.
- Two main challenges in LICs and EMEs:
  - Non-compliance and evasion as sources of revenue leakage.
  - The impact of the VAT on low-income individuals.
- Paper analyzes two groups of tools:
  - Tools to encourage issuance of invoices or use of payment systems (to improve VAT control).
  - Tools to reduce the VAT impact on low-income individuals (PIT credit, VAT rate reduction, targeted transfers).
- Aims:
  - Provide a clear characterization of these tools and their variants.
  - Offer guidance based on empirical studies and novel data analysis.

### II. Tools to encourage invoicing or formal payment systems — VAT paid on personal consumption creditable/deductible against the PIT
- Countries allowing VAT on personal consumption as a credit or deduction against the PIT include: Bolivia, Ecuador, Guatemala, Paraguay.
- Table 1 (selected country details):
  - Bolivia: VAT rate 13; PIT threshold (USD) 7,200; PIT rates (%) 13; VAT Deduction or credit: Credit.
  - Ecuador: VAT rate 12; PIT threshold (USD) 11,290; PIT rates (%) 0; 5; 10; 12; 15; 20; 25; 30; 35; VAT Deduction up to USD 14,703.
  - Guatemala: VAT rate 12; PIT threshold (USD) 6,275; PIT rates (%) 5; 7; Deduction up to USD 2,000.
  - Paraguay: VAT rate 10; PIT threshold (USD) 17,278; PIT rates (%) 8; 9; 10; No deduction limit.
- Design variants:
  - Credit of VAT paid on personal consumption against PIT liability (e.g., Bolivia’s VCR).
  - Deduction of personal consumption from PIT taxable base (e.g., Paraguay, Ecuador, Guatemala historically).

### III. Theoretical analysis — expected effects and distortions
- Theoretical rationale:
  - A comprehensive PIT with credit for VAT would tax income without additional tax on consumption; VAT on registered trader’s goods creditable; consumption from unregistered traders yields no VAT credit.
- Practical and distributive concerns:
  - Personal allowances in PIT mean high-income individuals pay PIT while low-income individuals often only pay VAT, potentially increasing regressiveness.
  - Low-income individuals with no PIT liability cannot benefit from VAT credits/deductions, incentivizing purchases from informal establishments.
  - The policy is distortive: encourages current consumption and can reduce savings.
  - Under progressive PIT schedules, deductions are more valuable at higher marginal rates, increasing regressivity.
- Administrative and revenue control issues:
  - Difficulty controlling large numbers of invoices deducted against PIT; increased risk of fake invoices.
  - High control costs and significant allocation of tax administration resources to invoice verification.
  - Example: Bolivia PIT (VCR) revenue reduced from 0.27 percent of GDP in 2005 to 0.15 percent in 2012 (Figure 1).

### IV. Empirical findings — revenue and distributional impacts (Ecuador, Guatemala, Bolivia)
- Cross-country tax revenue context (Table 2: Percentage of GDP):
  - Bolivia: Total Tax Revenue 23.8; PIT Revenue 0.2; VAT Revenue 7.8; PIT and VAT Revenue 8.0.
  - Ecuador: Total Tax Revenue 14.3; PIT Revenue 1.0; VAT Revenue 6.5; PIT and VAT Revenue 7.5.
  - Guatemala: Total Tax Revenue 10.4; PIT Revenue 0.4; VAT Revenue 4.8; PIT and VAT Revenue 5.2.
  - Paraguay: Total Tax Revenue 10.3; PIT Revenue 0.1; VAT Revenue 5.1; PIT and VAT Revenue 5.2.
  - Lower middle income: Total Tax Revenue 16.7; PIT Revenue 2.6; VAT Revenue 6.0; PIT and VAT Revenue 8.6.
  - Upper middle income: Total Tax Revenue 18.5; PIT Revenue 2.6; VAT Revenue 5.9; PIT and VAT Revenue 8.5.
  - High income: Total Tax Revenue 21.7; PIT Revenue 7.2; VAT Revenue 7.0; PIT and VAT Revenue 14.2.
  - Africa: Total Tax Revenue 16.6; PIT Revenue 2.9; VAT Revenue 5.3; PIT and VAT Revenue 8.3.
  - Asia Pacific: Total Tax Revenue 17.6; PIT Revenue 3.0; VAT Revenue 5.0; PIT and VAT Revenue 8.0.
  - Europe: Total Tax Revenue 25.1; PIT Revenue 7.1; VAT Revenue 8.1; PIT and VAT Revenue 15.2.
  - Middle East and Central Asia: Total Tax Revenue 13.1; PIT Revenue 2.5; VAT Revenue 6.7; PIT and VAT Revenue 9.2.
  - Western Hemisphere: Total Tax Revenue 18.4; PIT Revenue 2.9; VAT Revenue 5.3; PIT and VAT Revenue 8.3.
- Ecuador — impact of deducting personal consumption from PIT:
  - Deductions reduce redistributive capacity of the tax system; better-off taxpayers benefit most.
  - Microsimulation results (based on 2012/13 household survey updated and tax returns, calibrated to 2019 PIT revenue):
    - Eliminating the PIT deduction of personal consumption would increase PIT revenue by 0.4 percentage points of GDP.
    - Eliminating the deduction improves the Reynolds-Smolensky after-tax income Gini by 0.5 Gini points.
    - Scenario reallocating additional revenue to first three deciles yields an improvement of 1.6 Gini points.
  - Table 3 — Ecuador: The Revenue and Equity Impacts of Eliminating the PIT deduction allowed for Personal Expenses:
    - Current PIT system: PIT revenue (% of GDP) 0.9.
    - After policy change: PIT revenue (% of GDP) 1.3.
    - Net Effect: 0.4.
    - After-tax Income Gini: Current PIT system 50.1.
    - After elimination of the deduction of personal consumption: 49.6 (improvement 0.5).
    - After elimination and redistribution to poorest three deciles: 48.52 (improvement 1.6).
- Guatemala — impact on VAT revenue from removing VAT credit against the PIT:
  - VAT credit eliminated May 2012 (replaced by fixed deduction about US$ 2,000).
  - VAT revenue (percent of GDP) began decreasing in 2011 and continued until 2016 while VAT credit was substituted in 2012 (Figure 2).
  - Econometric analysis (Appendix I summary):
    - Change in VAT revenue as share of GDP was not statistically different from zero after the policy change once controlling for household final consumption, imports, and tax expenditures.
    - Changes in VAT collection were statistically explained by changes in imports and final consumption, not by the reform.
    - Adding VAT exemptions as control improved model fit but exemptions were not statistically significant; imports remained the significant variable in regression specifications.
- Bolivia — administrative outcome example:
  - Bolivia’s VCR design allowed employees to claim VAT paid on any personal consumption against VCR liability; an additional credit equivalent to 13 percent of two minimum wages if purchases from small taxpayers who cannot issue VAT invoices.
  - Bolivia experienced a decline in PIT (VCR) revenue from 0.27 percent of GDP (2005) to 0.15 percent (2012), mainly due to problems of control (Figure 1).

### V. Policy implications and recommendations (based on analysis)
- Key policy insights:
  - Allowing VAT paid as credit or deduction against PIT tends to favor better-off taxpayers and can reduce the redistributive capacity of the tax system.
  - The tool is distortive of consumption-savings decisions and can incentivize consumption from informal providers.
  - Administrative control is costly and susceptible to fraud (fake invoices), reducing PIT revenue and complicating tax administration priorities.
  - Empirical evidence does not show robust improvements in VAT revenue from these PIT-based invoice-incentive tools once other factors (imports, consumption, exemptions) are considered.
- Suggested policy directions:
  - Elimination of the VAT-as-PIT-credit/deduction appears a priori as a better alternative to regain revenue and reduce distortions, particularly where PIT rates are already high and administrative costs are substantial.
  - If the objective is to protect low-income individuals, prefer implementing well targeted transfers rather than PIT deductions or credits for personal consumption.
  - Example: In Ecuador, elimination of the deduction coupled with targeted transfers to the three lowest income deciles yields stronger improvements in equality (1.6 Gini points) than retaining the deduction.

*Source: IMF staff (content from "1. The VAT Impact on Low Income Individuals in LICs and EMEs").*

### 21.      Although non-compliance VAT rates have been rising in Guatemala since the policy

### 21.      Although non-compliance VAT rates have been rising in Guatemala since the policy change took place

### Non-compliance in Guatemala
- Finding: Non-compliance VAT rates have been rising in Guatemala since the policy change took place (Figure 2).
- Econometric result: The econometric analysis did not reveal that the rise in non-compliance was caused by the policy change; the increase after the reform was passed is not captured by the after-reform dummy variable (Appendix I, Table 2).
- Additional observations:
  - (i) The non-compliance rates had been higher during the period the VAT was allowed as a PIT credit.
  - (ii) The increase in non-compliance would be more likely related to several corruption scandals that had the Guatemalan government over the years following the reform and the increase of smuggling.
- Chart notes (from figure captions):
  - Axes and series mentioned: Total VAT (% of GDP) - left axis; VAT Non-complaince rate (percent) - right axis.
  - Time span referenced in figures: 2008–2018.

### The Reduction of the VAT Rate for Using Electronic Means of Payments
- Policy description: Some countries have allowed a reduced rate applied on goods and services purchased using electronic forms of payments to discourage cash use, foster financial transactions through banks, and facilitate control for the tax administration.
- Equity concern: The tool is regressive because individuals who are out of the banking system, who tend to be the poorest, will be left out.
- Country experiences:
  - Argentina:
    - Granted a VAT rate reduction of 5 percentage points for final consumer purchases paid with debit card until 2016 and 3 percentage points rate reduction for payments with credit card until 2003.
    - Both benefits were eliminated because they entailed revenue losses and were regressive (they benefited a larger share of better-off households, Table 4).
  - Honduras:
    - The refund is equivalent to 8 percent of the VAT included in the invoice.
    - Benefit for payments with debit or credit cards remains in place despite benefiting the rich more than the poor (Table 4).
  - Pakistan:
    - VAT rate is reduced from 15 to 12 percent for payments with debit or credit cards.
    - Benefit remains in place despite benefiting the rich more than the poor (Table 4).
- Gender note: In Honduras and Pakistan, male individuals have greater access to debit and credit cards, potentially enabling them to benefit from the VAT rate reduction more than female individuals.

### Korea: Cash Receipt System (2005–2015)
- Program design:
  - Incentives for merchants and unregistered final consumers to report electronically sales in cash via a seller credit card terminal generating a cash receipt and sending sale information to the National Tax Service.
  - Vendor benefit: credit 1 percent of their cash turnover against VAT liability for each receipt issued (until 2009); starting 2010 a fixed amount was granted for any cash receipt or electronic invoice to compensate for the cost of the Internet or outsourcing E-invoice issuing.
  - Consumer benefit: credit against income tax liability if they provided the name of the merchant and their income tax file number.
- Cost constraint: The main constraint was the need to keep the benefit to vendors and final consumers lower than the forgone VAT revenue if the sale was not reported.
- Outcomes:
  - The program was successful in reducing the number of underreported cash sales and was canceled in 2015.
  - Contextual factors supporting success: since 1997, credit cards have been the primary payment method for private consumption in Korea.
    - In 2014, credit cards accounted for over 64.1 per cent of consumption expenditure, with debit cards accounting for a further 12.4 per cent of consumption expenditure (Krever, 2014).
  - Corporate adoption:
    - Corporate taxpayers who use E-invoice or the ‘cash receipt system’ increased to 99.4 percent of total CIT taxpayers in 2014.
    - Sales issued with these systems increased to 99.9 percent of total sales.
  - Survey evidence: 69.4 percent of tax practitioners and employees of business taxpayers agreed or strongly agreed that tax evasion would have declined.

### Lottery Programs
- Policy description: Lottery programs encourage final consumers to request invoices from suppliers that can be exchanged for lottery entries; in some countries, this includes a VAT refund or VAT rate reduction.
- Mechanism: Incentives for consumers to request invoices force sellers to record the sale and charge VAT, shifting focus from punishment to rewards for compliance.
- Government costs and benefits:
  - Costs: prizes, revenue loss from VAT rate reduction, advertisement, administrative costs.
  - Benefits: improved monitoring of economic transactions via invoices and potentially higher VAT and CIT revenues.
  - Consumer cost: welfare loss due to time invested in participation.
- Country experiences and evidence:
  - Brazil and Mongolia:
    - Lottery programs include VAT refunds, making it difficult to isolate lottery effects from VAT reduction/refund effects.
    - Criticism: VAT refund components benefit mainly better-off households.
    - Cost: VAT refunds reach about 20 percent of the total VAT paid in Mongolia and 20 percent of the VAT paid by the final sale establishment in Brazil (about 1 percent of the total value of the purchase).
    - Administrative costs include implementation, advertisement, customer services, and bank fees for deposit of refunds.
    - Participation dynamics in Brazil:
      - Program started with 67 thousand participants in 2008, peaked at 277 thousand in April 2015, then declined especially since August 2015 when the refund to final consumers decreased to 20 percent from 30 percent of the VAT paid by the final sale establishment, to be nowadays around 100 thousand.
    - Empirical evaluations:
      - Naritomi 2019: VAT revenue increased by 9.3 percent net of rewards in 2008–2011 (did not include administrative and welfare costs).
      - Mattos et al (2013): limited impact; no robust effect on total tax collection nor sectors other than services; evidence of a non-linear effect in time; smaller effect in highly industrialized state of Sao Paulo.
  - Slovakia:
    - Introduced lottery in 2013; impact modest.
    - Fiscal impact estimated at about EUR 8 million annually (extrapolating from last quarter of 2013, focusing on restaurants and retailers).
    - Total cost in 2014 was EUR 1.6 million (including marketing); overall cost high compared to extra revenue.
  - China:
    - Empirical examination using six-year data from 37 districts in Beijing and Tianjin indicated sales tax revenue was significantly higher (over 17.1 percent) in areas where the lottery took place.
    - Real growth rates of sales tax and total tax revenues were significantly higher (over 21.5 percent and over 10.4 percent, respectively) in areas with the lottery relative to areas without it (Wan, 2010).
  - Portugal:
    - VAT non-compliance decreased from 16 to 12 percent between 2013 (lottery introduced) and 2014.
    - No causality established; decrease coincided with broad set of policies including electronic invoicing and the lottery program “Fatura da Sorte”.
    - Governmental reporting: in 2014 the number of sales invoices with the consumer taxpayer number sent to tax authorities increased by 36.3 percent compared to the previous year; one year later, the total number of sales invoices increased by 51.2 percent.
  - Mongolia (program introduced 2016):
    - Technology-intensive system requiring registration, app download, and barcode scanning; app also allowed consumers to claim a refund of 2 percent of the value of their purchase.
    - VAT revenues increased to 6.6 percent of GDP in 2019 from 4.8 percent in 2016 after program introduction.
    - Caveats: simultaneous large increase in investment in the mining sector and real GDP growth (real GDP grew 12 percent and nominal GDP 26 percent) between 2016 and 2018 could also have contributed to VAT revenue increases; direct government costs on prizes and administration were high and may offset revenue gains.
- Program durability: Usefulness short-lived in some countries; participation and effectiveness declined when prizes and advertisement decreased (Argentina eliminated after four years; Brazil participation declined after refunds were reduced).

### Tools to Reduce the Impact of the VAT on Low-Income Individuals — Tax Credit to Low Income Taxpayers
- Rationale: VAT often perceived as regressive relative to income because poorer households consume a larger share of their total income; several measures address the impact, including reduced rates, exemptions, VAT thresholds, and targeted VAT credits.
- VAT credit feature: Uses different criteria to target exclusively low income households or “deserving” populations; more targeted than reduced rates, VAT thresholds, or refundable VAT to all PIT taxpayers.
- Canada — GST/HST credit (example):
  - Only low-income PIT taxpayers may claim a General Sales Tax/Harmonized Sales Tax (GST/HST) credit.
  - The GST/HST credit is paid four times a year to individuals and families with low incomes to help offset all or part of the GST/HST that they pay.
  - Filing: When taxpayers fill online the tax return, the Canadian Revenue Agency (CRA) automatically estimates the amount of the credit.
  - Factors in estimating credit amount: number of eligible children registered for the GST/HST credit, adjusted family net income (considering expenses such as rent, energy, and property taxes), residence of the taxpayer, and benefits received from the province of residence.
  - Example amounts (October 2019, Ontario):
    - For a couple with total annual income of $25,731 Canadian Dollars (CAD) with two children of 4 and 8 years residing in Ontario, the GST credit reached about $221.5 CAD by quarter (plus $102.7 CAD for the Ontario Sales Tax).
    - For the same family with income $70,800 CAD, the tax credit decreases to about $59.0 CAD for the federal GST and $31.9 for the Ontario Sales Tax.
  - Design note: The amount of the credit is not based on the amount of GST/HST paid because it decreases when income increases.
  - Refundability: The Canadian PIT credit for GST is refundable, so requires near universal filing but does not require that any PIT be owing.
- Effectiveness evidence:
  - Godbout, L., et al, 2011 found:
    - (i) Without the credit, GST/HST are almost proportional to household income in Canada; with the credit, the effective tax rate rises with income.
    - (ii) The tax credit is a more effective tool for improving the progressivity of consumption taxes than zero-rating of basic groceries.
  - Authors’ view: The VAT credit in Canada is targeted to low-income taxpayers; zero-rates and exemptions mostly benefit high income individuals (see Appendix II).
- Implementation considerations:
  - The GST/HST credit in Canada applies in a context of high labor formalization and effective control.
  - Most LICs and EMEs that apply the VAT credit against the PIT need to improve their social safety net and taxpayer-register database to control this kind of tool.
  - Uruguay is cited as an example of an EME that can implement this and offers lessons for other countries.
  - Note on relative impact: The VAT has a larger impact on low-income individuals in developed countries than in LICs and EMEs, where first deciles of income distribution are less affected by the VAT because:
    - a higher proportion of their consumption comes from self-production,
    - most of their purchases are made in informal markets,
    - small taxpayers are often registered in special schemes exempted from the VAT (see Box 1).
  - This is important when determining the level of the VAT refund.

*Source: IMF staff based on the provided PDF content.*

### Box 1. The VAT Impact on Low-Income Individuals in LICs and EMEs

### Box 1. The VAT Impact on Low-Income Individuals in LICs and EMEs

### VAT incidence: overview and mechanisms
- The VAT may not be as regressive in practice as often supposed because:
  - Exemptions or reduced tax rates on food and other goods that make up the consumption basket of low-income individuals.
  - In LICs and EMEs most purchases happen in informal markets or are bought from very small retailers who are exempted from the VAT.
  - Consumption of self-produced food is out of the scope of the VAT (the proportion of self-produced food is higher among low-income individuals than among rich).

### Evidence from country studies and literature
- Jenkins et al, 2006 (Dominican Republic):
  - Using a dataset covering 2,042 goods and services and the type of establishment where items were purchased, the study found the burden of the VAT is progressive over all quintiles of household expenditure.
  - If the VAT base is made comprehensive, the estimated incidence remains progressive over all quintiles.
- Other studies and examples:
  - Mariscal and Verner, 2017 (Mexico) and Fabrizio et al, 2017 (Uganda) reached similar conclusions that VAT is not regressive.
  - Some studies found VAT neutral or that VAT is preferred due to efficiency advantages, especially when complemented with expenditure measures.
  - Footnote examples of neutral impact: Fabrizio et al (2017) for Honduras (increase of the VAT rate from 15 to 18 percent would have a neutral impact); same conclusions cited for Namibia (World Bank, 2017); Ethiopia (Hills and others, 2017); and Togo (IMF, 2017).

### Case studies of targeted refund/credit mechanisms
- Uruguay — Social Card Program (Programa Tarjeta Uruguay Social):
  - Operating since May 2006; monetary transfer via magnetic card to 60 thousand households in the worst socioeconomic situation (Uruguay population is about 3.5 million inhabitants).
  - Transfers usable at the Network of Solidarity Shops (about 1,000 shops).
  - Monthly amount varies with household composition (e.g., number and ages of individuals).
  - The 30 thousand households in the worst conditions receive double amount.
  - In 2019, a household with two children received URU$ 1,610 (about US$ 40.0) or URU$ 3,220 (about USD 80) per month.
  - Beneficiaries are selected using the Critical Deficiency Index (ICC) collected by the Ministry of Social Development; purchases exclude tobacco, cigarettes, and alcoholic beverages.
- Thailand — State Welfare Smartcard (introduced 2018):
  - From November 2018 to April 2019, smartcard holders had 5 percentage points of the 7 percent VAT rate reimbursed to the consumer up to USD 16 per month (1 percentage point going to the National Savings Fund or holder’s savings account and the rest to the Revenue Department).
  - After launch there were 14.5 million smartcard holders.
  - Government transferred around 1,168 thousand USD to e-wallets on the State Welfare Card and around 223 thousand USD to cardholder savings accounts.
  - Supplemental note: From November 2019 to September 2020 (11-month period), cardholders were reimbursed 5 percent of the amount of VAT; during this period government transferred around USD 2,665 thousand to the e-Wallet on the State Welfare Card.
- Portugal — VAT paid creditable to PIT with limits:
  - A credit equal to 15 percent of the VAT incurred in selected services with a maximum of EUR 250 per family member (taxpayer and dependents) is available (article 78-F of the CIRS).
  - Eligible VAT must proceed from an electronic invoice for specified services: maintenance and repair of motor vehicles and motorcycles; accommodation, catering and the like; hairdressing salons and beauty institutes; and veterinary activities. Total VAT on public transportation is also deductible within the EUR 250 limit.
  - The design targets activities prone to cash transactions to (i) reduce regressiveness of the VAT and (ii) encourage consumers to request invoices.
  - Distributional example: the Effective Tax Rate (ETR) for an individual with income of EUR 7,091 decreases from 14.5 to 11.0 percent while for an individual with income of EUR 200,000 the reduction is only of 0.1 percentage points.

### Empirical analysis: Guatemala (Appendix I)
- Policy change (April 2012, Decree 10-2012):
  - VAT invoices would no longer be creditable against the PIT liability for salaried workers; instead VAT invoices up to Q.12,000 (USD 1,530) allowed as a deduction from PIT taxable income.
  - PIT standard deduction increased to Q.48,000 (USD 6,120) from Q.36,000 (USD 4,590) prior to the reform; combined total deduction available to salaried workers became Q.60,000 (USD 7,650) per year.
- Observed trends after reform:
  - VAT non-compliance rates increased while VAT revenue declined, but no statistical evidence that the policy change directly affected VAT revenue collection or non-compliance rate; trends better explained by a decline in imports and contemporaneous political scandals.
- Regression framework:
  - Static contemporaneous linear regressions on annual, quarterly, and monthly data for 2009–2019.
  - Main specification: ΔVATrev t = β0 + β1 ΔX t + β2 d t + β3 tt t + u
    - ΔVATrev: year-on-year percent change in VAT revenue as a share of GDP.
    - X: covariates (final household consumption, imports, monthly index of economic activity depending on periodicity).
    - d t: dummy for reform enforcement.
    - tt: seasonal trend for quarterly and monthly data.
- Key regression results (Table I.1):
  - Annual regressions:
    - Percent change in final consumption: 0.993* (0.546)
    - Percent change in imports: 0.707*** (0.12)
    - Dummy (0 before reform): -0.00261 (0.0148)
    - VAT tax expenditures (as % of GDP): -0.0176 (0.0967) [included in column (2)]
    - Observations: 14 (column 1); 10 (column 2)
    - R-squared: 0.787 (column 1); 0.844 (column 2)
  - Quarterly regression (column 3):
    - Dummy: -0.161 (2.949)
    - Observations: 41
    - R-squared: 0.764
  - Monthly regression (column 4):
    - Percent change in the Monthly Index of Economic Activity (IMAE): 2.444*** (0.498)
    - Dummy: 2.949 (7.181)
    - Observations: 184
    - R-squared: 0.539
  - Standard errors reported in parentheses; significance: *** p<0.01, ** p<0.05, * p<0.1.
- VAT non-compliance regressions (Table I.2):
  - Dependent variable: Percent change in VAT non-compliance rate.
  - Coefficients (column 1):
    - Percent change in final consumption: 0.79 (4.569)
    - Percent change in imports: 0.354 (2.410)
    - Dummy (0 before reform): 0.0649 (0.0842)
    - Percent change in VAT tax expenditures (as % of GDP): -1.059 (1.649)
    - Constant: 0.0154 (0.181)
    - Observations: 7
    - R-squared: 0.981
  - Interpretation: coefficients statistically insignificant; the increase in non-compliance after the reform is not captured by the reform dummy.

### Conclusions and policy recommendations (extracted)
- Distinction between two tool groups:
  - Group 1: tools to encourage issuance of invoices to final consumers to control VAT and other taxes — these measures tend to be regressive, impose compliance burdens, and can complicate administration (including use of fake invoices).
  - Group 2: tools to reduce the VAT impact on low-income individuals — these measures are progressive when well targeted and affect a reduced number of people, simplifying administrative control.
- Specific findings on using VAT credits/deductions against PIT:
  - In Ecuador, using VAT as a credit or deduction of the PIT:
    - Reduces the redistributive capacity of the Ecuadorian tax system (measured by the Reynolds-Smolensky index).
    - Absent this measure and keeping everything else constant, PIT revenue would increase by 0.4 percentage points of GDP and the Gini coefficient would improve by 0.5 Gini points.
  - In Guatemala, elimination of the VAT credit did not have a negative impact on VAT revenue.
  - For Ecuador, Bohne A. 2018 finds no evidence that providers who increased formal invoices increased the amount of tax paid.
- Policy recommendations:
  - Avoid use of VAT credit or personal expenses as deduction of the PIT except when targeted toward low-income individuals or with a maximum amount (as in Portugal).
  - Avoid reductions in the VAT rate to encourage use of credit and debit cards: they are regressive, expensive, and lack evidence of improving VAT revenue.
  - Prefer targeted low-income programs (e.g., Uruguay and Canada approaches) to reduce VAT’s distributional impacts when administrative capacity allows and when low-income purchases occur in formal markets or consumed goods/services are not VAT-exempted.
  - Maintain lottery programs to encourage formalization only if their benefits (increase invoicing, reduction of non-compliance) significantly exceed costs; eliminate them when objectives are achieved (example: Korea’s cash receipt system).
- Final empirical takeaway for Guatemala:
  - Changes in VAT collection are statistically explained by changes in imports and changes in final consumption (imports, especially, since VAT on imports represents more than 50 percent of total VAT revenue); no statistical evidence that the 2012 reform dummy explains changes in VAT revenue or non-compliance.

*Source: Box 1. The VAT Impact on Low-Income Individuals in LICs and EMEs (wpiea2021040-print-pdf).*

### Appendix II. The Impact on Income Distribution of Exemptions and Reduced Rates of

### Appendix II. The Impact on Income Distribution of Exemptions and Reduced Rates of Consumption Taxes

### Summary of main arguments
- VAT is often characterized as regressive because propensity to save increases with income, so the tax burden as a share of current income is higher at lower levels of the income distribution and decreases as income rises.
- Reduced rates and exemptions are widely used to mitigate VAT regressivity but are poorly targeted distributional devices: they often reduce regressivity on average but transfer a larger share of benefits to higher-income households in absolute terms.
- The VAT’s distributional effect must be appraised in the context of the overall tax and spending system: VAT is an efficient revenue-raising tax but, being transaction-based, is hard to tailor for equity; equity objectives are mainly achieved via spending instruments.

### Empirical evidence on targeting of reduced rates and exemptions
- OECD and country-sample findings show the largest share of benefits from differentiated VAT rates accrues to the highest income quintile.
- Examples from the tabulated distributions (Differentiated Rates; Exemptions and Differentiated Rates) show concentration of benefits in top quintiles:
  - 20 OECD countries: 13.5, 17.3, 20.0, 22.5, 26.7, Total 100.0
  - Honduras: 7.3, 12.4, 17.1, 22.5, 40.7, Total 100.0
  - Nicaragua: 4.0, 7.0, 10.2, 23.7, 55.1, Total 100.0
  - Colombia: 5.6, 9.2, 13.3, 21.0, 50.9, Total 100.0
  - Dominican Republic: 10.3, 14.0, 16.8, 21.8, 37.0, Total 100.0
  - Mexico (includes only the zero rate on basic foods): 10.8, 17.2, 21.2, 23.1, 27.9, Total 100.0
  - South Korea: 6.3, 13.8, 20.2, 25.7, 33.9, Total 100.0
  - Morocco: 14.5, 12.0, 15.0, 19.7, 38.8, Total 100.0
  - Tunisia: 7.9, 12.7, 17.0, 23.2, 39.2, Total 100.0

- The pattern repeats across OECD, Latin America, Africa, and South Korea: the fifth quintile benefits the most from exempted goods and reduced rates in many cases.

### Factors moderating measured regressivity
- Several effects not always captured in VAT incidence analyses can reduce measured regressivity:
  - VAT registration thresholds exempt sales by small retailers—where the poorest often buy—leading to a substantial moderating impact where data are available.
  - Food production for self or home consumption is untaxed, with similar moderating effects.
  - Noncompliance/untaxed consumption is uneven: Bachas (2019) finds the top decile has "27 percent less untaxed consumption than the bottom one."
  - When measured on aggregate consumption rather than current income, VAT can be close to neutral or even slightly progressive due to untaxed consumption. Alavoutunki, Haapanen, and Pirttila (2019) find VAT adoption "has not led to increased inequality when measured based on consumption," whereas measuring based on disposable income suggests VAT introduction raises inequality.

### Findings from recent fiscal-incidence studies (selected conclusions)
- Cameroon (2018, CEQ, IMF): Small adverse impact on income distribution; removal of VAT exemptions needed to boost inequality-reducing transfers. Redistributive impact of cash and indirect transfers is weak; in-kind transfers have larger impact on poverty and inequality reduction.
- Ghana (2018, CEQ): VAT reduced rates and exemptions costly and mildly progressive (yet, not well targeted towards the poor). Cash transfers are fairly well targeted and have positive redistributive effect.
- Eswatini (2017, CEQ, WB, IMF): Social transfers generally progressive. Step-up in cash transfers and means-testing needed to further reduce inequality.
- Ethiopia (2017, CEQ): VAT reduced rates and exemptions costly and regressive (not equalizing). Cash transfers are poorly targeted and have limited redistributive effect.
- Namibia (2016, CEQ, WB, IMF): Large social assistance program(s) but poorly targeted.
- Nigeria (2018, CEQ, IMF): Increasing VAT revenue progressive (reduce income inequality) but increase poverty. Social programs of limited coverage; widening and scaling up social safety nets would significantly reduce poverty and reduce inequality.
- Senegal (2018, CEQ): VAT reduced rates and exemptions costly and neutral (not equalizing). Cash transfers are fairly well targeted and have positive redistributive effect.
- South Africa (2017, CEQ, WB): VAT mildly regressive. Social spending progressive and significantly reducing inequality; targeting of social grants improved in recent years; quality of in-kind transfers needs improving.
- Togo (2017, CEQ, IMF): VAT is progressive but increases poverty rate of poor and near-poor. Social spending low and poorly targeted; more VAT revenue needed to finance desired social programs.
- Uganda (2017, CEQ): VAT marginally reducing inequality. Social spending, except for spending on tertiary education, reduces inequality but is very small.
- Zambia (2017, CEQ, WB, IMF): VAT reduced rates and exemptions costly and regressive (not equalizing). Cash transfers are fairly well targeted and have very positive redistributive effect.

### Policy implications and recommendations
- Reduced rates and exemptions are a blunt tool for redistribution: they are "poorly targeted" ways of helping vulnerable groups and often benefit higher-income households more in absolute terms.
- Equity objectives are better pursued through spending instruments (social assistance, cash transfers, in-kind transfers) and through well-targeted social programs financed by VAT revenue.
- Where social spending is weak or poorly targeted, increasing VAT revenue and reallocating it to well-designed social programs can reduce inequality and poverty; several country studies indicate the need to widen and scale up social safety nets to improve poverty outcomes.
- Policymakers should consider:
  - Assessing the cost and targeting efficiency of VAT reduced rates and exemptions relative to direct transfers.
  - Strengthening targeting and coverage of cash and in-kind transfers to maximize redistributive impacts.
  - Accounting for informal-sector dynamics, registration thresholds, and untaxed self-consumption when evaluating VAT incidence.

*Source: Appendix II. The Impact on Income Distribution of Exemptions and Reduced Rates of Consumption Taxes (excerpt).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021040-print-pdf.pdf_