ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Towards GST: Choices and Trade-offs

What kind of Goods and Services Tax is under contemplation? Will it only replace CENVAT and the service taxes levied by the centre, leaving, the state VATs untouched? Or will it also displace the VATs now being levied by the states? These are important issues which have a vital bearing not only on the shape of trade taxation in the country but also on the federal structure of the polity envisaged in the Constitution and call for a more careful consideration than has taken place, so that choices are made with a clear-eyed view of the trade-offs.

Budget 2006

Towards GST: Choices and Trade-offs

What kind of Goods and Services Tax is under contemplation? Will it only replace CENVAT and the service taxes levied by the centre, leaving, the state VATs untouched? Or will it also displace the VATs now being levied by the states? These are important issues which have a vital bearing not only on the shape of trade taxation in the country but also on the federal structure of the polity envisaged in the Constitution and call for a more careful consideration than has taken place, so that choices are made with a clear-eyed view of the trade-offs.


Why GST?

he announcement by union finance minister Chidambaram in this year’s budget speech of his intent to introduce a national Goods and Services Tax (GST) from 2010 has been enthusiastically welcomed by many, especially corporate businesses. This is understandable.

For even after more than two decades of reforms, the task of reforming domestic trade taxes and moving towards a system of value added tax (VAT) which was considered necessary for the economy to function efficiently and smoothly remains incomplete; a full-fledged VAT is still not in place. Neither CENVAT, the current name for excise duties levied by the union government which is purportedly built on the VAT principle nor the VATs that have come into operation in the majority of the states since April 2005 in replacement of their sales taxes bear all the attributes associated with a good VAT.1 Excises are a tax on manufacturing and suffer all the limitations of a manufacturer level tax (definitional, correct value determination and so on), despite heroic efforts to get around them. While the rates have been compressed into a general rate of 16 per cent, selective exemptions/concessions continue resulting in effect in multiplicity of rates. Services are also being taxed by the centre but selectively and not in an integrated fashion with goods. While cascading is sought to be alleviated by allowing credit for the tax paid on the purchase of services used as inputs against CENVAT payable on the manufactured product, the two taxes remain separate and administrated under different statutes. Crediting of service tax against CENVAT is riddled with rules that are far from simple.

On the states’ side, considerable progress has been made towards evolving a harmonised system of VAT. Even so deficiencies and irritants remain. Efforts for harmonisation with the prescription of a uniform standard floor rate notwithstanding, the structure of the state VATs remains diverse and flawed, with many states deviating from the agreed rate scheme unilaterally. Their rate structure pays no heed to the basic tenet of VAT, viz, taxation of inputs and final products at the same rate. Industrial inputs are taxed at a lower rate than what is applicable generally and a large number of commodities are taxed at a rate much lower than the general rate. Besides, the base of state VATs suffers from a grave deficiency; it does not include services because of constitutional limitations. There are also sundry taxes like stamp duty on transfer of property and entry of goods in “local areas” and on a few services like entertainment which ought to have been merged in VAT. Then there is the tax on interstate sales – the infamous central sales tax (CST).

While several of these anachronisms are on their way to being phased out, it is widely felt that only a unified tax on all or nearly all goods and services based on the VAT principle can remove the infirmities of the tax system and enable Indian industry and the economy to flourish in a fast globalising world. Not surprisingly, the idea of a unified VAT on goods and services mooted by the Kelkar Task Force (KTF) on the implementation of the FRBM Act received widespread support and endorsement from the Twelfth Finance Commission. With the line between goods and services getting blurred separate taxation of services has become untenable. The finance minister was not far from wrong when he said “It is my sense that there is a large consensus that the country should move towards a national level Goods and Services Tax (GST)…”.

But what kind of national VAT in the form of GST is under contemplation? Will it be in replacement of only CENVAT and the service taxes levied by the centre leaving the state VATs untouched? Or will it also displace the VATs now being levied by the states? The answers to these questions have a vital bearing not only on the shape of trade taxation in the country and how they impact on households and businesses, but also on the federal structure of the polity envisaged in the Constitution and call for a more careful consideration than has taken place so that choices are made with a clear-eyed view of the tradeoffs. This note seeks to discuss briefly the merits and disadvantages of the alternatives and the trade-offs involved, and explore a workable alternative for India.


A Single National VAT

While it is difficult to make out exactly what the finance minister had in mind while talking about a “national level GST”, the presumption (and wish) among businesses and even many economic journalists seems to be that what is being contemplated is a single, unified tax on goods and services levied and administered only at the central level. For them, VAT levied at the sub-national level with 29 states in the fray, each with its own laws and procedures is simply not on any more. A unified single market within Indian economy cannot function smoothly, it is

Economic and Political Weekly April 8, 2006

argued, unless the trade tax system is also unified at the national level. The question then boils down to this: should the contemplated national GST be the only tax on domestic trade in goods and services with arrangement for revenue sharing with the states? The finance minister’s budget statement setting a target date for ushering in “a national level GST that should be shared between the centre and the states” has raised expectations that this is the direction in which the reforms will be moving.

Apart from the obvious economic benefits from having a single VAT at the national level, for a long time it was believed even by experts that in a federal country VAT is best administered at the national level. Operating a destination-based VAT in a large federal country presents acute difficulty. In the absence of border controls, as is the case in federations, zero rating of interstate sales which is required to operate such a system opens up opportunities for fraud. Several federal countries (Germany, Austria, Switzerland, Belgium and since 2000, Australia) have their VAT levied nationally with arrangements for revenue sharing. A little reflection would, however, show that however neat and attractive each may seem, a VAT at the central level only is neither practicable nor desirable for India. It is not necessary either.

First of all, it needs to be realised that the idea of an exclusively national VAT will be a non-starter in India as it would require the states’ sales tax administration to be wound up or taken over by the centre. It is extremely doubtful if the states can be persuaded to accept such a proposal, even if they agree to the extension of union tax powers to enable the centre to tax sales through all stages under the “Grand Bargain” proposed by the KTF. One also wonders whether it will be possible for a central tax agency, however well endowed, to administer a tax with “such local moorings as sales tax” as the Taxation Enquiry Commission of 1953-54 said while turning down the demand for the centralisation of sales tax from trade and industry.2 Reportedly, the number of taxpayers that the central excise (CE) department currently handles at present does not exceed one lakh and service tax assessees number about four lakhs. The number of registered dealers for state VATs/sales tax on the other hand would not be less than 35 lakhs, going by available figures. It is difficult to visualise the central excise department handling such a large number of taxpayers spread across the country.

More importantly, taking away the powers of sales taxation which is believed to be the dominant source of revenue for sub-national governments from the states would be detrimental not only to their fiscal autonomy but also accountability as that would make the states merely spending agencies with little responsibility for the funds they spend. It cannot be gainsaid that the levy of GST in Australia only at the national level, however efficient from the economic angle, negates the rationale of federal governance marking the near elimination of “the last vestiges of state fiscal autonomy in that country” as Richard Bird put it.3 Even though the revenue is supposed to flow back to the states under the equalisation arrangements, it has greatly widened vertical fiscal imbalance in the Australian federation [Collins 2000]. Revenues that flow from the centre cannot be regarded as something that the recipient governments are politically responsible for raising. Politicians and bureaucrats may find it convenient to avoid the onus of raising the monies they like to spend, but that way lies the road to fiscal irresponsibility.

Lastly, for the efficient implementation of a destination-based VAT, centralisation is not essential. While for a long time it was believed that VAT is a tax that is best administered at the national level, because of problem of zero rating cross border trade, recent discussion in the literature shows that expert opinion has veered away from this view and several models are now available that can make it possible to operate VAT at the sub-national level efficiently.4

For all these reasons the option of a single GST at the national level should be ruled out and alternatives looked for.


Dual VAT

The KTF, while putting forward the idea of a unified GST, also did not envisage the tax to be only a central levy. In their scheme, there would be a state component, presumably implying that it would partake of the character of what is called a “dual VAT”. It is acknowledged that a well designed and well administered dual VAT can contribute to strengthening intergovernmental fiscal relations [World Bank 2004]. If, however, the GST is to be levied at two levels, one at the centre and another at the states, there are several alternatives.

Basically these are: (i) completely independent VATs at the two levels, (ii) a tax regime whereby each level of government sets its own rates independently but on a similar base and under close administrative cooperation, and (iii) a single “joint” VAT with a state component in the rate and some of the revenue flowing to the states under an agreed arrangement or formula [Bird and Gendron 2001].5

The system of VAT in Brazil, the pioneer in having VAT as a major tax instrument, is of type (i). VAT is levied at the federal, state and also local level. Despite some central control over tax on interstate trade, under the Brazilian system the states have considerable autonomy in running their VAT. Few however regard the Brazilian system as an example to emulate because of the many problems it has encountered. It is only categories (ii) and (iii) that seem to merit consideration.

In some respects the KTF model falls in category (ii) in that it envisions the GST to be levied at a rate made up of two components, one the central part (CGST) and the other, the states part (SGST). The key features of the dual VAT contemplated in the KTF report relevant in the present context are: there would be three rates apart from “zero” viz, a standard rate of 20 per cent (12 per cent for the centre and 8 per cent for the states), a lower rate of 6 per cent and a maximum rate of 20 per cent. Exemptions would be few and limited mainly to food, medical care, education, residential housing and certain financial services. Central excises could be levied on products like petroleum, natural gas and tobacco.

Since the introduction of such a tax will require the centre to allow the states to tax services and the states to accept extension of the tax powers of the centre to tax sales at all stages, the task force proposed a “grand bargain” between the two levels, referred to earlier, whereby

  • Both levels will have concurrent but independent jurisdiction over a largely common tax base and the tax at both levels will extend to final consumers covering both goods and services.
  • The existing octroi/entry tax, central sales tax, states sales tax, stamp duties, and other cascading taxes and fees will go.
  • Both levels will have power to fix the rates but there would be one rate for all states and rate setting will be coordinated between the two levels.
  • Thus, clearly the KTF wanted the unified tax to be administered at two levels and

    Economic and Political Weekly April 8, 2006 not one at the national level. This is evidenced also by the suggestion in the KTF report for a clearing house mechanism to settle the dues of states arising from credit for tax allowed in the state of destination in interstate sales for the tax collected in the state of origin. (Whether such a mechanism can work in a country like India – it was considered infeasible in the European Union – is another matter.)

    However, the KTF has ignored the federalism angle in prescribing a uniform rate in all states and a rate scheme whereby the major part of GST revenue will go to the centre. Presumably, the SGST would be set on the basis of consultation or “consensus” as has happened for state VATs but no state would be able to go beyond it if it so desires. While there is a strong case to harmonise the base and the procedures among the states expert opinion is almost unanimous that the essence of a tax power lies in the power to fix the rates and so sub-national governments must have autonomy in the matter of rates subject only to a floor as in the European Union. Unless this is allowed the “grand bargain” contemplated by the KTF would operate to the disadvantage of the states. The vertical imbalance that marks the finances of the centre and the states will accentuate sharply under the KTF dispensation [Shah 2004]. Rough computations indicate that the proportion of tax revenue raised by the states in the total tax revenue of the government (centre and states combined) which stands at around 33 per cent at present would come down to barely 15 per cent or so. That would grievously weaken the financial autonomy of the states and increase their dependence on the centre.

    Hence while looking for a suitable dual VAT model for India one has to explore the feasibility of only alternative (ii). With a concurrent central GST it may not be necessary to go in for a compensating VAT or CVAT of the kind proposed by Varsano and McLure (2000a) as another model of a dual VAT. Thus one is left with only alternative (ii). But how feasible is it to operate a dual VAT of the kind envisaged in model (ii) in a federal polity as diverse as India?

    Canadian Dual VAT6

    Experience shows that it is perfectly possible to operate a dual VAT in a federation given some harmonisation (not total uniformity) and coordination between the two levels. Since 1991 Canada is having a federal level VAT in the form of GST while the provinces have their own sales tax or VAT. In fact, Canada runs more than one model of dual VAT/sales tax. Quebec has its own VAT, the Quebec sales tax (QST), levied at the rate of 7.5 per cent on the GST (7 per cent) inclusive of price. Ontario runs a retail sales tax at 8 per cent exclusive of the GST. The differences in the base that marked the QST and GST initially have now mostly gone. Thus a uniform VAT base is established for both federal and provincial VAT though differential treatment is allowed at the level of the final consumer. Initially, there were some differences in the input tax credits. These have been ironed out but some restrictions are still there in effect on input tax credits for large firms under QST. While such differences in the base are not desirable, the GST-QST model shows that these can be accommodated within limits under a dual VAT system.

    The rates of the two taxes are set independently by the respective governments. The tax bases are also determined independently; however, they are essentially the same. Right from the beginning the two taxes are being collected by the revenue department of Quebec according to rules set by Ottawa. Taxes on interprovincial sales are dealt with on the “deferred payment basis” as in EU, that is, the sale from a registered vendor from Quebec to a registered counterpart in say Ontario is zero-rated for QST though not for GST. This in a way serves to maintain the VAT chain and helps enforcement. The overriding federal GST acts as an enforcement mechanism. Audit is conducted in consultation between the two authorities. The system is reported to be working fine at the technical level despite the political differences between the two governments. After reviewing the QST-GST system, Bird and Gendron (2001) conclude: “Overall however the QST and GST, as they exist, constitute an operational ‘dual’ or ‘concurrent’ VAT system – with essentially none of the problems usually associated with such systems”.

    Proposals for reform of the Brazilian VATs also seek to replace the federal and state VATs by a dual VAT. The local level VAT which also operates in Brazil will be replaced by a retail sales tax at the municipal level. There will be uniform legal norms across the country; however the states will have power to legislate on local “specificities”. Tax rates both federal and state will be set by a federal law but states will have the power to vary the rate within a given band (15 to 20 per cent). The taxes will operate on the destination principle [Varsano 2004].

    The idea of a parliamentary VAT law for the states may not be acceptable in India. However, harmonisation of the base and procedures can and should be established through a consultative body like the Empowered Committee. But, the decisions reached at such a body should be enforceable with sanction which can be invoked not by going to the central government but by creating a separate judicial body as is done in EU through the European Court of Justice.

    Another model of “dual” VAT operating in Canada since 1997 is the harmonised sales tax or HST. Under this system, operative in three maritime provinces the tax is levied on a harmonised base in place of the earlier federal and provincial sales taxes at a federal-provincial rate of 15 per cent. Of this 7 per cent is the federal GST and 8 per cent represents provincial sales tax. The tax is administered federally. Although the HST limits individual provincial autonomy any change in either the base or the rate of tax requires the unanimous agreement of the provinces.

    Options for India

    The HST no doubt serves to reduce the costs of both compliance and administration. However the economic rationale for requiring uniform rates across provinces is questionable. For, as McLure puts it, “…tax harmonisation which may be needed to simplify compliance and administration, should not extend to the choice of tax rates” [Mclure 2000b]. The QST-GST arrangement as it operates now seems superior providing almost an “ideal solution”. The base should be harmonised but total rate uniformity need not be there across states. Only there should be just one

    Economic and Political Weekly

    Available from

    Star News Agency

    Mahendra Chambers, Magazine Market 146, D N Road Mumbai - 400 001

    Economic and Political Weekly April 8, 2006

    rate in each jurisdiction. There should also be a mechanism (missing in the EU) to enable changes to be made in the consumer codes whenever required.

    It is worth noting that the proposal for a national sales tax put forward by the federal authorities in Canada in 1991 to introduce a nationwide VAT in place of the sales taxes prevailing at the two levels fell through because of opposition from the provinces. The proposal for a common system of VAT for the EU proposed by the European Commission in 1996 also has not materialised since it was seen as curbing the fiscal sovereignty of the member states. No one can deny the need for harmony though as argued in Cnossen (1990) there can be a case for some tax diversity as well. The advent of e-commerce and digital products is posing a threat to state or even national sovereignty in taxation. The challenge lies in reconciling the conflict between the need for harmony and preservation of federalism with its manifold advantages in its true spirit.

    As and when its federal level GST comes into being, India should be in a position to replicate the QST-GST model. It needs to be considered however whether the two taxes should be administered by tax authorities at the two levels independently with concurrent jurisdictions as implied in the KTF report or whether it would be possible to have the two taxes administered by state governments as in Quebec. Implementation of the tax by two independent authorities running in parallel may be vexatious and troublesome. But, the Quebec model of state governments administering both CGST and SGST would not seem to be feasible in India. In that case there seems to be no alternative but to have the CGST and SGST to be administered at two levels of government leaving the task of administration at the ground level to the states. With a harmonised base and procedures that may not be an impossible task.

    One way to administer a dual VAT in India at two levels could be to require vendors to charge both the taxes on the invoice and remit the tax realised to the respective governments as is done for the federal GST and the provincial sales tax in the Canadian provinces having their own sales tax.7 Of course there has to be nationwide agreement on the base and the procedures – like in the Sixth Directive of the European Commission – and coordinated audit. It may not be necessary for the central authorities to directly involve themselves in the administration of the tax in all stages or for all dealers. Every dealer liable to register for both CGST and SGST may be required to file a common return with the state tax authorities with an additional page for CGST. With a Tax Information Exchange (TIE) network now under construction the central tax agencies may be able to access any return whenever felt needed. If the SGST is levied on the CGST inclusive price, there will be an incentive for the states to see that the CGST is duly paid. Some states, if they so desire should be given the option of harmonising their VAT and having it administered by the centre as in the maritime provinces of Canada.

    Concluding Observations

    To conclude, while there are significant advantages both economic and administrative in having only a single national VAT, it also involves costs that are both economic and political. Taking away the most important tax powers of the states goes against the tenets of decentralisation, which works best when the authorities that spend bear responsibility for raising the funds they spend. That impacts adversely on the efficient organisation of the public sector. Then there are political costs in weakening the sub-national governments. As Buchanan argues, efficiency should not be viewed only in the narrow economic sense. The imperatives of “political efficiency” should not be lost sight of [Buchanan 1999]. Trading off political efficiency against economic gains would be short-sighted.

    Centralisation of sales taxation is not essential either. Central and state taxes can exist side by side as the experience of Canada with VAT and of the US with income taxes show. Given the limitations of resources of the central tax department and the experience of the states in implementing sales tax, however deficient, every effort should be made to get the best out of the two. A dual VAT – a variant of the QST-GST model of Canada –seems to be the most promising alternative with elbow room to the states in the matter of rates and to a limited extent in the base.

    Administration at the field level is best left with the states. Audit has to be a joint endeavour but each authority should have the power to institute audit in cases selected by them. Trust and coordination between the centre and states tax authorities would however be essential for smooth implementation of a dual VAT. Give the experience with the working of the Empowered Committee that should not be an impossible task. There should however be an independent quasi-judicial body to enforce agreements between the centre and the states and among the states. Thoughtfully, the Constitution provides for the creation of such a body under Article 307.




    1 For a succinct description of these attributes, see Cnossen (1998).

    2 Report of the Taxation Enquiry Commission 1953-54, Vol III.

    3 In a recent communication.

    4 See for a symposium on the models, articles in the International Tax and Public Finance, December 2000.

    5 Bird and Gendron (2001) add a fourth category whereby a compensating central VAT (CVAT) is levied along with the state VATs to help track interstate sales. For an elaboration of CVAT, see McLure (2000b).

    6 This section draws liberally on Bird and Gendron (2001).

    7 This resembles the dual VAT scheme proposed in Poddar (1990).


    Bird, Richard and Pierre-Pascal Gendron (2001): ‘VATs in Federal Countries: International Experience and Emerging Possibilities’, Bulletin, International Bureau of Fiscal Documentation, July.

    Buchanan, James (1999): ‘Public Finance and Public Choice’, MIT Press.

    Cnossen, Sijbren (1990): ‘The Case for Tax Diversity in the European Community’, European Economic Review, 34.

    – (1998): ‘Global Trends and Issues in Value Added Taxation’, International Tax and Public Finance, 5.

    Collins, D J (2000): ‘The Impact of the GST Package on Commonwealth-State Financial Relations’ (Australian Tax Foundation, Research Study No 34).

    McLure, Charles (Jr) (2000a): ‘Implementing Subnational Value Added Taxes on Internal Trade: The Compensating VAT’, International Tax and Public Finance.

    – (2000b): ‘Tax Assignment and Sub-national Fiscal Autonomy’, International Bureau of Fiscal Documentation, December.

    Poddar, Satya N (1990): ‘Options for a VAT at the State Level’ in Shoup Gillis and Sicat (eds), Value Added Taxation in Developing Countries, World Bank.

    Shah, Anwar (2004): ‘Comments on KTF Proposals’ in World Bank.

    Varsano, Ricardo (2004): ‘Brazil’s Experience with an Interstate VAT’ in World Bank (2004).

    World Bank (2004): ‘Report to the Kelkar Task Force on a Proposed Dual VAT for India’, October.

    Economic and Political Weekly April 8, 2006

    To read the full text Login

    Get instant access

    New 3 Month Subscription
    to Digital Archives at

    ₹826for India

    $50for overseas users


    (-) Hide

    EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

    Back to Top