A brief panorama of the Brazilian tax reform proposition
On 3 April 2019, Deputy Baleia Rossi submitted to the Brazilian Congress the proposition of constitutional amendment n. 45/20191, which suggests pro- found changes to the Brazilian tax system, establishing a new value added tax. The proposition was drafted by a think tank called CCiF – Centro de Cidadania Fiscal, composed by academics from the areas of Law and Economy, and spon- sored by some of the biggest companies that operate in Brazil. At Congress, the Deputies are currently hearing specialists on taxation. After that, the proposition shall be voted both at the Chamber of Deputies and at the Senate, before it can enter in force.
To understand the depth of the reform, it should be emphasised that Brazil is a Federation, where Federal, States and local governments can impose taxes over consumption. Because of this superposition of tax competences, there are currently five relevant taxes that are charged on consumption in the country. Most of them are structured in a way to foster neutrality throughout the consump- tion chain, even if not always through the system of invoice (credit) method. The Federal government receives three of these taxes: a levy on the production of in- dustrialized goods, called IPI – Imposto sobre Produtos Industrializados, and two taxes on the revenues obtained from entrepreneurial activity, destined to sustain the social security system (PIS/PASEP and COFINS). The States charge a value added tax, called ICMS – Imposto sobre Operações relativas a Circulação de Merca- dorias e Prestação de Serviços de Transporte Interestadual e Intermunicipal e de Comu- nicação. Finally, local governments receive a services tax, called ISSQN – Imposto sobre Serviços de Qualquer Natureza, which is not neutral, since it does not give credits to the buyer. The proposition of constitutional amendment n. 45/2019 re- places these taxes by a single VAT.
In a far more modest proposal, on 21 July 2020 the Brazilian President sub- mitted to Congress a bill that creates a federal VAT, called CBS – Contribuição Social sobre Operações com Bens e Serviços2. CBS would replace two of the three
1 BRASIL. Proposta de Emenda Constitucional n. 45/2019. Available at: www.camara.leg.br/propo- sicoesWeb/prop_mostrarintegra?codteor=1728369&filename=PEC+45/2019. Accessed: 14 Au- gust 2019. For an update on the processing of this proposition, see: www.camara.leg.br/proposi- coesWeb/fichadetramitacao?idProposicao=2196833. Accessed: 31 August 2019.
2 BRASIL. Projeto de Lei n. 3.887/2020. Available at: www.camara.leg.br/proposicoesWeb/fichade- tramitacao?idProposicao=2258196. Accessed: 29 August 2020.
federal taxes (PIS/PASEP and COFINS), and would not alter the power the States’ and local governments’ have to impose their own taxes. Unifying these two feder- al taxes represents some simplification of the tax system. Moreover, the new tax is restructured, in a way that it is collected with greater uniformity among differ- ent economic sectors. Nonetheless, with this proposal most of the complexity from the Brazilian tax system remains.
This complexity arises, mainly, from the superposition of tax competences among Federal, States and local governments, a fact that is well known and inter- nationally recognized. Brazil is ranked 100 in the Tax Complexity Index, which aims to rank 100 countries, from the country which has the least complex tax system for multinational corporations (Singapore) to the country that has the most complex system (Brazil)3. As a consequence, Brazilians dedicate almost
2.000 working hours per year to prepare and pay their taxes – making Brazil, by far, the most time-consuming country –, according to the Doing Business Project from the WORLD BANK (2019).
Among the several attempts to simplify the Brazilian tax system, the propo- sition of constitutional amendment n. 45/2019, which replaces the five taxes men- tioned above by a value added tax, called IBS – Imposto sobre Bens e Servicos, Por- tuguese for GST – Goods and Services Tax, is one of the most recent efforts4. It re- ceives wide support, but, as one would expect, some specific economic sectors worry that the change would represent an increase in their own taxation. More- over, some States and local government fear losing their power to tax, due to the centralization in a single VAT (BATISTA JR. & MARINHO, 2020, p. 780).
Actually, this new tax shall be centrally collected, having uniform tax obliga- tions all over the country. Evidently, from the taxpayer’s perspective, the replace- ment of five taxes by IBS and the centralization of the taxation power represent a deep simplification of the tax system. One example can clarify this impact: today, a Brazilian chain of shops, that sells products all over the country, has different tax obligations, according to the State where its shops are established. Consider- ing that Brazil has 26 States and the Federal District, this chain of shops might have 27 slightly different ICMS’s obligations, since this tax is established by the States. On top of that, the chain of shops must also collect the taxes due to the Federal government. IBS certainly represents an improvement from this perspec- tive, as it proposes to unify the taxation over consumption.
On the other hand, the centralization of IBS’s rule making and taxation authority alters the level of autonomy States and local governments currently have, since they would no longer be able to establish their own legislation on the
3 The Global MNC Tax Complexity Project. Available at: www.taxcomplexity.org/. Accessed: 14 August 2019.
4 RONZANI (2008, p. 186-8) gives a brief panorama on the Brazilian attempts to reform its tax system, explaining the main innovations of each constitutional amendment proposition since the Constitution from 1988.
taxes they collect, neither would have direct enforcement power over the taxpayers. In fact, according to the constitutional amendment n. 45/2019, IBS’s proceeds would be split between Federal, States and local governments, but all the legislation and enforcement is centralized in a single tax authority (BATISTA, 2020, p. 115).
This model of centralized collection, with division of proceeds, is not new to Brazil: the tax imposed over small businesses – called Simples Nacional – is already charged in this way. The division of Simples Nacional’s proceeds is automated, via an electronic system, securing that States and local governments actually receive the appropriate amounts. A similar system is predicted for IBS, which would be overseen by the IBS’s Management Committee, composed by representatives from the Federal, States and local governments. The composition of this Manage- ment Committee is critical to guarantee States’ and local governments’ interests are adequately protected on the new taxation model (TORRES JR. & BARBOSA, 2020, p. 298).
Since the Management Committee is not yet clearly designed, States and local governments are still uncertain about the tax reform proposition. Unlike Simples Nacional, IBS represents a huge parcel of governments’ budget, making the proposed alteration of tax authority a sensible issue in Brazil. States and local governments fear losing the ability to control their finances, as they will no longer be responsible for the collection of taxes over consumption of goods or services (BATISTA, 2020, p. 116).
Effectively, the tax reform proposition represents a profound alteration of the Brazilian tax system. ICMS, which is currently charged by the States, account- ed for 20,73% of all taxes collected in Brazil in 2017 (for comparison, Income Taxes – both over business and personal – represented 18,22%); ISSQN, imposed by the local governments, represented 2,65% (RECEITA FEDERAL DO BRASIL, 2017). Centralizing this power to tax deeply alters the equilibrium between the entities of the Brazilian Federation, and this is a point of strong criticism the tax reform proposition is facing (SOUZA, 2019; ÁVILA, 2018; RIBEIRO, 2019; BA- TISTA JR. & MARINHO, 2020, p. 782; MATA & CARVALHO, 2020, p. 608-9).
On the other hand, de SANTI, de PAULA JR., SANTIN and CYPRIANO (2019) think the tax reform proposition foster the equilibrium between the Bra- zilian federal entities, since each State and local government would have the pow- er to establish the tax rate they deem adequate to sustain their expenditure needs. In a similar way, APPY & LONGO (2020, p. 247-9) sustain that the concentration of the power to tax has operational advantages to the Federal, the States and the local governments.
Moreover, the unification of legislation would strongly diminish the disputes between Brazilian entities (the so called “fiscal war”). At present, in an attempt to attract investment, States and local governments usually give fiscal incentives to new businesses, lowering the tax rates of ICMS and ISSQN. This would no longer be viable if IBS is approved, for two main reasons: the new levy is proposed with
a single rate, regardless the economic activity which is carried out; and IBS adopts the destination principle.
Considering this later perspective, and bearing in mind that the current tax system is not operational, the major alteration proposed by the constitutional amendment n. 45/2019 is welcome. In fact, the present complexity arises, mostly, from the high degree of autonomy that was historically given to States and local governments to establish their own tax rules (CCiF, 2019). This autonomy is strongly limited by the proposition of constitutional reform. Thus, if the tax re- form is adopted, the Brazilian taxation system would become aligned to the ma- jority of the VATs charged on other countries, which usually impose this tax at the national level. In this sense, EBRILL et al (2001, p. 7) recall that it is very difficult to structure a VAT at the subnational level, and the Brazilian experience seems to be a good example of this statement.
It is worth mentioning that a reform as broad as the proposed in Brazil was undertaken by India in 2017, replacing 20 consumption taxes that used to be charged by the Central and States governments by a new value added tax5. Previ- ous to the tax reform, these Indian entities charged diverse taxes, such as VATs, services taxes and excise duties. All were replaced by a dual VAT, which is inte- grated, but collected separately to the Central and the States governments. The simplification of the Indian system can be a good source of comparison and inspi- ration for the Brazilian tax reformers, since it shows that a broad reform is feasi- ble, even in a complex federation such as the ones from Brazil and India.
Nonetheless, even if a broad tax reform is welcome, due to the simplification of the system, this is not to say that the proposition of constitutional amendment has no flaws. Apart from the centralization of tax power – which was briefly de- bated above, but is beyond the scope of this paper – there are specific problems that relate to the taxation of digital products, shown below. Part 2 of the paper states that, due to the particularities of Brazilian legal theory, the name given to the proposed tax – Goods and Services Tax – can leave a loophole for intangible goods or services, such as software, apps or streaming services. Part 3 shows that the tax reform proposition does not clearly address the taxation of these intangi- bles.
IBS: the name given to the tax and the loophole for digital products
The proposition of constitutional amendment n. 45/2019 aims to create a value added tax, as clearly stated on its justification6. Nonetheless, instead of
5 India GST. Available at: www.avalara.com/vatlive/en/country-guides/asia/india.html. Accessed: 26 August 2019.
6 “IBS will have the characteristics of a good value added tax (VAT), a model adopted by most cou- ntries for taxing the consumption of goods and services” (“O IBS terá as características de um bom imposto sobre o valor adicionado (IVA), modelo adotado pela maioria dos países para a tri- butação do consumo de bens e serviços”). Proposta de Emenda Constitucional n. 45/2019 (n. 1).
naming the tax by the traditional initials VAT, the proposition designates it as Goods and Services Tax (GST, or IBS, in Portuguese). It does so following the exam- ples from the New Zealand’s GST, 19867; the Australian GST, 20008; and even the recent Indian GST, established in 20179.
The name GST is perceived, by the authors of the Brazilian tax reform proposition, as more modern than VAT. This change of terminology is also a rec- ognition to the fact that VATs are not usually charged over the value added (EBRILL et al, 2001, p. 1) but over the amount declared on the invoice, allowing credits from the inputs (TAIT, 1988, ch. 1). For this reason, some countries de- parted from the classical terminology (MOREIRA, 2019, p. 24-5), avoiding criti- cisms such as CARRAZZA’s (2015, p. 448), who considers that ICMS, a tax collect- ed via invoice (credit) method, is not a VAT, since the tax base is not the actual value added on the sale.
However, it should be pointed out that there are countries which are cur- rently establishing new taxes on consumption, and are still naming them VAT. This is the case with the members of the Gulf Cooperation Council, which signed a Value Added Tax framework treaty in 2016, calling for each member to impose a 5% VAT on designated goods and services (REGAN, 2019). In this sense, the classical name VAT is still up to date.
Moreover, establishing that the tax bases are “goods or services” already seems an aged concept, in light of the digital economy (SZELBRACIKOWSKI, 2019, p. 158). Actually, there is no clear frontier among goods and services in most information technology (IT) products, as recognized by the OECD (2017, p. 39), and detailed by LAMENSCH (2015, p. 15-18). This certainly represents a problem in tax systems that differentiate the taxation of these bases, such as the Brazilian tax system (MIGUEL & OZAI, 2016). As detailed by CASTELLO (2021,
p. 155-8), the Brazilian Supreme Court has spent the past twenty years debating whether the software industry should pay ICMS to the States or ISSQN to the lo- cal governments10. It has recently decided that these intangibles are subject to the services tax. Before the Supreme Court’s decision, both States and local govern- ments attempted to tax IT, always with big complexity and poor results, since no consensus was found on the adequate classification of these products.
When ICMS was levied, IT companies argued that their products were not goods (merchandises), and, therefore, should not pay this tax. GRECO (2000, p.
7 New Zealand GST. Available at: www.avalara.com/vatlive/en/country-guides/oceana/new-zealand. html. Accessed: 26 August 2019.
8 Australian GST. Available at: www.avalara.com/vatlive/en/country-guides/oceana/australia.html. Accessed: 26 August 2019.
9 India GST (n. 5).
10 The main case is the judicial review undertaken by the Supreme Court at: BRASIL. Supremo Tribunal Federal, ADI-MC n. 1.945, Rel. Min. Cármen Lúcia, Rel. p/ Acórdão Min. Dias Toffoli. DJe 20 mai. 2021.
81-95) made a good analysis on the subtleties of this subject, and concluded that, in some cases, software can be considered goods, for ICMS’s purposes. But the IT sector has certainly evolved sharply in the past 20 years. At present, the Supreme Court has ruled that these intangibles could, in theory, be considered merchan- dise. Nonetheless, software licensing could not be subject to ICMS, because there is not an actual transfer of property rights on the licensing procedure (CASTEL- LO, 2021, p. 157).
When ISSQN was charged, on the other hand, the digital industry would sustain that their products were not services, since there was no direct human activity needed to provide them (LUMMERTZ, 2017). In other cases, ISSQN could not be imposed, since there was no legislation explicitly establishing that the digital service in question was subject to the services tax (MOREIRA, 2008). As a consequence, most authors would classify the trade of IT as a license to use an intellectual property right, which, in their view, is not taxable as a good, nor as a service (CARRAZZA, 2015, p. 190-200; SZELBRACIKOWSKI, 2020, p.152; AM- ARAL, 2018. p. 141; PADILHA, 2016; UHDRE, 2019, p. 177). Due to this blurred
boundary between the concepts of goods and services on the IT market, States and local governments were drown into litigation. They sometimes attempted to collect both ICMS and ISSQN on the same transaction (BRIGAGÃO, 2019, p. 395), and seldom imposed any consumption tax on the sector.
And, contrary to what one would expect, this conflict would not vanish if IBS is approved. Even if Brazil starts to charge the same tax over both goods and ser- vices, there is still a loophole for intangibles. In fact, if one follows the theory which states that the trade of IT is a license to use an intellectual property right, this trade cannot be considered neither a trade in services, nor a trade in goods. It is not considered a trade in services, since there is no direct human activity in- volved in the supply. And it is not considered a trade in goods, as there is no actu- al transfer of property rights. Thus, following the position of the majority of Bra- zilian authors that already wrote on this subject, it would be easy to sustain that the trade of digital intangibles is out of the scope of a Goods and Services Tax.
However, CORREIA NETO, AFONSO and FUCK (2020, p. 41-61) observe a tendency to classify the trade of intangibles as a trade in services. In fact, this tendency is made evident by the Supreme Court’s decision that submitted the software industry to ISSQN. In doing so, it aligned the Brazilian system to the best international practice, since both OECD and the European Union classify the trade of intangibles as trade in services. Following this reasoning, the prob- lem of naming the new VAT as “Goods and Services Tax” would not be grave.
Moreover, to prevent the trade of IT from falling outside the scope of IBS, the drafters of the proposition of constitutional amendment n. 45/2019 suggested the following text for article 152-A, § 1st, I, of the amended Brazilian Constitution: “the goods and services tax: I – shall be levied on: a) intangibles; b) transfer or licencing of property rights; c) leasing of goods; d) imports of tangible or intangi-
ble goods, services or property rights”11. Clearly, in enumerating the situations when IBS can be imposed, the proposition of constitutional amendment seeks to avoid the development of legal theories such as the one just pointed out. GRU- PENMACHER & RODRIGUES (2020, p. 261) consider that this legal text is suffi- cient to guarantee the imposition of IBS over digital intangibles.
Nonetheless, this constitutional provision turns the power to tax inflexible, limited to the forms of consumption we know today. If new legal arrangements to transfer goods, services or property rights are created, they are likely to fall out- side the scope of IBS. Consequently, a constitutional provision containing a gen- eral clause, with open language, would be preferable12. But a general clause that simply establishes a goods and services tax would not endure the theory which sus- tains that IT is neither a good, nor a service. In fact, due to the Brazilian legal traditional (LUKIC, 2020), the drafters of the tax reform proposition appropri- ately felt the need to list, at the Brazilian Constitution, all the tax bases.
The bill that proposes CBS, the federal VAT, has a broader language. It es- tablishes, on Article One, that the tax bases are goods or services. Nonetheless, anticipating that there might be a debate on the appropriate classification of dig- ital goods, it defines, on article 61, § 1, that the licencing of property rights is considered a service. In doing so, the bill departs from the Brazilian legal tradi- tion, and aligns the legislation to OECD’s (2017, p. 39) recommendation. This legal text is preferable, but still does not prevent litigation on the concept of “ser- vices”, for tax purposes. CBS is created by an ordinary bill, with no constitutional status, and one could question its validity, in light of the Brazilian Constitution.
If the new tax is named value added tax, on the other hand, the problem of defining what can be considered a service is not likely to arise. In fact, a value added tax can be charged over digital services or products, regardless their clas- sification as goods, services or property rights, simply because there is a value added when the transaction occurs. In this sense, due to the particularities of the Brazilian legal tradition, it seems preferable to adopt the classical terminology, VAT, than the newer GST (CALIENDO, 2020, p. 853; CASTELLO, 2021, p. 277).
In short, one of the fragilities the proposition of constitutional amendment
n. 45/2019 has is its name: the name of the tax leaves digital services or products in an ambiguous situation, since they are considered, by some Brazilian authors, as neither a good, nor a service, but a license to use an intellectual property right. Consequently, litigation on the ability to charge IBS on IT goods is likely to arise. A simple way of preventing this litigation would be to adopt the classical terminol- ogy – Value Added Tax – establishing a broad tax base.
11 Proposta de Emenda Constitucional n. 45/2019 (n. 1).
12 On the use of general clauses or “principles-based legislation” by Tax Law, see: FREEDMAN (2010, p. 717); JONES (1996, p. 63); RIBEIRO (2002, p. 313).
IBS and cross-border digital transactions
Secondly, and more critical than the first point, there is no provision on the proposition of constitutional amendment n. 45/2019 addressing the cross-border consumption of IT. The proposition simply announces that the tax is collected on the “imports of tangible or intangible goods, services or property rights”13, but it does not establish how this collection shall occur. This is a serious omission: the establishment of a new legal framework for taxation on consumption without con- fronting the cross-border trade of IT does not seem advisable, due to the growing economic relevance of digital trade (OECD, 2015, p. 131; OECD, 2019).
It can be argued that the details of cross border taxation do not need to be done at a constitutional provision, and should be left to secondary legislation. Even so, a debate on how the secondary legislation will address this issue is advis- able, especially because this taxation is still a matter of international apprehen- sion, and there is no consensus on the ability of a VAT to effectively tax digital transactions. On this point, two main concerns arise: it is not always easy to deter- mine where the digital transaction is carried out; and it is difficult to enforce VAT on cross-border transactions.
Place where the digital transaction occurs
On the debate about the place where the transaction is carried out, CCiF’s White Paper on IBS secondary legislation (2020) is a good starting point. Accord- ing to CCiF’s proposal, when there is a remote trade on services or intangibles, the place where the consumer is established/resident is considered the place where the transaction has occurred (CCiF, 2020, p. 28). However, it is not always easy to establish where the consumer is resident. To do so, CCiF (2020, p. 27) proposes the identification of at least two of the following proofs of address: address pro- vided by the consumer; IP address from the device used to hire the service; geo- location of the device used to hire the service; address provided to the payment method; place where the telephone used to celebrate the contract is located; and address provided to the tax authority.
It is not clear on the White Paper who is responsible for determining the correct place where the transaction is carried out. Shall it be the consumer, the supplier or the tax authority? When the consumer is not registered to pay IBS, he cannot be responsible for determining the place where the transaction occurs, since he has no direct link with the tax authority. On the other hand, holding the supplier liable to define the place of the transaction can demand that it provides to the tax authorities information that it does not possess. The tax authority, on its turn, usually relies on the information provided by the supplier to access tax liabilities. Consequently, if the tax authority is responsible for determining the place of the transaction, it would need a system that cross-checks, on real time, all
13 Proposta de Emenda Constitucional n. 45/2019 (n. 1), article 152-A, § 1º, I, d.
the data mentioned above, and this system can raise privacy issues. As one can see, the criteria to establish where the transaction is carried out is complex, and one might need a sophisticated IT system for this definition. Moreover, the struc- ture of IBS can impose an even greater burden on the supplier.
In fact, one needs to bear in mind that, according to article 152-A, § 2º, I, States and local governments have the power to establish the rates of IBS, when the supplies occur in their own territory. This power is given to protect the Bra- zilian Federation: even if the entities have no freedom to establish the legal de- sign of the tax, they have the power to define the tax rate that is adequate to face their governmental expenditure. In this fashion, the project of constitutional amendment seeks to diminish the criticisms the tax reform proposition is facing, due to the concentration of power at the Federal level, as pointed out in Part I of this paper.
But the power to establish different tax rates means that a supplier estab- lished in State A, providing digital services to customers resident at States B and C, might have different tax obligations. This is an obstacle to cross-border trade: establishing where the customer resides becomes highly relevant, not only for in- ternational transactions, but also for trade occurring inside Brazil, since the sup- plier will pay different taxes according to the residence of the customer. If the supplier is required to determine where the e-customer is resident, it will have an “unrealistic compliance burden”, as pointed out by LAMENSCH (2015, p. 24), when writing on the European VAT context.
LAMENSCH (2015, p. 23) considers this burden unrealistic because the cus- tomer can inform the wrong address to the supplier, aiming to lower his/her VAT obligation. Consequently, if the foreign company has the burden to prove that the address given by the customer of digital services is correct; it can face future lia- bilities, due to wrong information provided by the customer. With IBS, the “unre- alistic compliance burden” envisioned by Lamensch (2015, p. 24) seems worst, not only because it is not limited to foreign trade, but also due to the existence of various tax rates, according to States’ and local governments’ needs.
In brief, to prevent this problem – and make cross-border transactions of digital services operational – the Brazilian Tax Authority needs to develop a sys- tem that is secure and informative to the tax payer, so that the supplier is not misled by wrong information provided by the customer. Moreover, if the burden of determining the place where the transaction is carried out is imposed on the supplier, the tax authority needs to clearly establish how the supplier shall verify the customer’s address. Additionally, the tax authority shall not aim to collect VAT from the supplier, if it has undertaken the appropriate verification measures (OECD, 2019A, p. 70).
The enforcement of IBS on cross-border transactions
