Back Matter
  • 1 0000000404811396 Monetary Fund

Annex 1. A Framework for VAT Refund Management

Annex Figure 1.1.
Annex Figure 1.1.

A Framework for VAT Refund Management (Detailed Steps)

Citation: IMF How To Notes 2021, 004; 10.5089/9781513577043.061.A999

Source: IMF staff.Note: TSA = treasury single account; VAT = value-added tax.


This note was prepared by Mario Pessoa, Andrew Kazora Okello, Artur Swistak, Muyangwa Muyangwa, Virginia Alonso-Albarran, and Vincent Koukpaizan. It was reviewed by Gerd Schwartz, Michael Keen, Victoria Perry, Katherine Baer, Manal Fouad, Andrea Lemgruber, Andrew Masters, Debra Adams, Eric Hutton, Therese Van der Poel, Enrique Rojas, Yasemin Hurcan, Abdoulahi Mfombouot, Racheeda Boukezia, Majdeline El Rayess, Xavier Rame, David Gentry, and Lisette Atiyeh. This note is dedicated to the memory of our wonderful colleague, Mario Pessoa, who passed away before its completion.


The focus throughout this note is on the “invoice-credit” method, which is applied almost universally. Under this form of the VAT, a taxpayer has the right to credit the input VAT (s)he pays against the output VAT (s)he receives. In most cases, taxpayers have more than enough output VAT to offset their input VAT credits, which leaves them with positive net VAT liabilities. However, this is not always the case; taxpayers with little or no output VAT (for example, exporters or those investing heavily in business equipment or premises during the taxable period) may be left with negative VAT liabilities.


Reported VAT refund fraud illustrates the scope of this problem. For example, in 2013, the total VAT collection gap for European Union (EU) member states was almost EUR 170 billion, of which the estimated cross-border (refund) fraud accounted for around EUR 50 billion. The balance was attributed to other types of VAT evasion, legal avoidance, and unpaid VAT liabilities caused by insolvencies. In Australia, a 2013 investigation into a VAT refund fraud involving gold revealed that the Australian treasury lost about AUD 700 million over a five-year period through fraudulent VAT refund claims. This loss represented about 5.5 percent of net VAT receipts.


The latter part is addressed more comprehensively in “Combatting VAT Refund Fraud” (Andrew and Baer, forthcoming). The note discusses the nature and extent of VAT refund fraud in selected EU countries. It argues that this type of noncompliance requires the tax administration to adopt a coordinated strategy and deploy a range of countermeasures. While the note focuses on the EU, several of the lessons therein apply to countries at different income levels.


Harrison and Krelove (2005) examine the refund approaches of tax administrations in 36 developing, transitioning, and developed countries. They also evaluate the effectiveness of these approaches and suggest a best practice model that accounts for compliance issues faced by countries at different developmental stages. See also Keen and Smith (2007), and Chapter 15 of Ebrill and others (2001).


The countries are Australia, Bolivia, Botswana, Costa Rica, Ecuador, Guatemala, Indonesia, Jamaica, Kenya, Madagascar, Mauritius, Morocco, Peru, Portugal, Sierra Leone, Uruguay, Zambia, and Zimbabwe.


These include the Tax Administration Diagnostic Assessment Tool (TADAT), Revenue Administration Fiscal Information Tool/ International Survey on Revenue Administration (ISORA), and Revenue Administration Gap Analysis Program (RA-GAP).


This category also includes governments, NGOs, and other entities deemed to be final consumers, for example, small traders or VAT-exempt businesses.


Other VAT accounting methods, such as the subtraction method, are possible, and would also generate VAT refunds, albeit at lower levels (Zee 1995). Because that method is not widely used, this discussion of the VAT refund mechanism and its management focuses on the invoice-credit method.


Good practice (resulting in an “A” score) for VAT refunds—as defined by TADAT’s standardized assessment criteria—is to pay 90 percent of VAT refund claims within 30 days.


VAT withholding and reverse charge mechanisms differ. In the case of VAT withholding, a buyer retains a portion of VAT on his purchases (usually at a prescribed percentage of gross payment, for example, 10 percent) and pays this amount directly to the tax administration (on behalf of the seller). Depending on the amount withheld, the seller may need to seek a refund from the tax administration for the overpaid amount. Under the reverse charge mechanism, the buyer is responsible for accounting for and remitting the VAT associated with the transaction.


This approach is preferable to zero-rating at the point of sale, as it helps limit VAT system abuse.


Turkey uses a positive list of cases for which excess input tax credits may be refunded (those resulting from exports, domestic supplies subject to zero or reduced rates, or supplies subject to VAT withholding). Other cases, such as those resulting from investment or inventory buildup, do not qualify for refunds and must be carried forward indefinitely.


These limit VAT refunds by eliminating input VAT (exemptions granted on imports or zero-rating domestic purchases of business inputs) or payment requirements. In the case of VAT deferral, VAT due on importation is deferred to the subsequent VAT return; it is recognized as both input and output VAT, eliminating the need to pay first and then request a refund. In the case of a deemed VAT scheme, the buyer does not pay VAT charged on the supply, though it is assumed (deemed) that the supplier has collected it (a solution akin to zero-rating). VAT grouping limits the need for refunds if the group includes entities that are in refund positions, for example, exporters, and entities with positive VAT liability (in such a setting, the group may have enough output VAT to offset input VAT).


Whether zero-rating, exemptions, waivers, or other (less common) types of VAT breaks.


Interestingly, countries that have adopted VAT reliefs on the input side often use other policy and compliance measures, which result in an increased numbers of VAT refunds.


See formula in Ebrill and others (2001), p. 158.


Sources: ISORA and IMF survey data for 2014 and 2015. Definitions: LIC refers to low-income countries; LMIC to lower-middle income countries; UMIC to upper-middle income countries; and HIC to high-income countries.


ISORA’s second round did not cover Asian countries, except for those that are OECD members.


According to TADAT benchmarks, good practice involves refunding 90 percent of claims within 30 days.


RA-GAP is an IMF tool that enables countries to estimate the sizes of their tax gaps (differences between potential and actual taxes). To calculate the VAT gap, detailed data on VAT liabilities, payments, credits, and refunds are required for all VAT payers across all sectors of economic activity. This exercise often provides insight into VAT performance. To date, IMF staff has conducted 30 VAT gap estimates across Fund member countries.


Analyses of VAT credit stocks and refunds is based on RA-GAP data.


A TSA system represents a number of interlinked government bank accounts, including accounts with zero balances and one main bank account (at the central bank) into which all revenues are deposited, and from which all expenditures are disbursed.


The discussions in this note revolve around TSA arrangements, but apply equally to countries without TSAs. Countries that do not have, or are transitioning to, TSA systems often have banking arrangements that allow revenue transfers from commercial banks to central banks, with the commercial banks serving as intermediaries.


A good self-assessment system also requires the tax administration to provide effective services that help taxpayers understand the law and their obligations, thereby minimizing unintentional mistakes.


For example, during the 12-month period covered by the respective TADATs, about 10 percent of claims (by value) were rejected in Guatemala and Jordan; 23 percent were rejected in Kenya; and 34 percent were rejected in Tanzania.


Kenya, Morocco, Zambia, and Zimbabwe require the following evidence to support applications: copies of invoices of all input VATs being claimed, copies of output taxes charged, and proof of export.


The Australian Tax Office’s (ATO) experience is illustrative: in the 2000s, taxpayers attempting to test the VAT refund system began presenting numerous small refund claims in the hopes that such claims would “fly under the radar” of the administration’s risk filters. However, ATO officials detected this trend and began monitoring it closely, eventually adjusting the risk parameters to thwart such activities.


In Cameroon, VAT refund claim stocks were paid off using measures that were established to avoid further accumulation.


For example, the UK’s financial statement notes that, “where output tax is charged and input VAT is recoverable, the amounts are stated net of VAT. The net amount due to, or from, HM Revenue and Customs in respect of VAT is included within receivables and payables in the Statement of Financial Position” (Whole of Government Accounts: year ended 31 March 2012).

How to Manage Value-Added Tax Refunds
Author: Mario Pessoa, Andrew Okello, Artur Swistak, Muyangwa Muyangwa, Virginia Alonso-Albarran, and Vincent de Paul Koukpaizan