The most commonly asked question about the distributional effect of the value-added tax is whether it is a regressive tax. The apparently simple question is actually not so simple to answer.

The issue is related to various other considerations such as rate differentiation, exemptions, alternatives to VAT and also other measures to promote progressivity like pro-poor expenditure policy.

The usual definition of regressiveness is the following: When the ratio of consumption tax to income increases with higher income levels, it is called a progressive tax. When this ratio is proportionate, it is called a proportionate tax and when this ratio decreases it is called a regressive tax.

A single rate of VAT with no zero-rate -- except for exports, and few exemptions -- must mean that the VAT payments by low-income households will be a higher proportion of their incomes than payments by higher income households.

Or in other words, the VAT will be regressive (Alan A. Tait, VAT Policy Issues, p 5 in VAT Administrative and Policy Issues edited by Alan A. Tait, IMF, 1991).

What is important is that VAT is not regressive if tax payments are expressed as a percentage not of income but consumption. If consumption is used as the denominator, then by definition the impact would be proportional.

The burden of a VAT levied at uniform rate would also be largely proportional if the denominator is lifetime income instead of annual income because many income recipients are only temporarily in the lower income brackets.

They move into middle or upper income brackets as their earnings increase (Sijbren Cnossen, Key Questions in Considering Value Added Tax for Central and Eastern European Countries, VAT Monitor, November 1992, p.7).

The more important question to ask is not whether VAT is regressive but whether VAT is more regressive than the alternative indirect taxes, namely, sales, excise and turnover tax (not income tax because that is a direct tax).

Indirect taxes generally are regressive as they continue to impinge on the lives of many poor people in limited but potentially important ways (Richard M. Bird, A New Look at Indirect Taxation in Developing Countries, World Development, Vol.15 No.9, p.1158, 1987).

VAT is just as regressive as excise duty or sales tax, not more than that. Other taxes are not only regressive, but also cascading.

It is potentially misleading to focus on the distributional aspects of VAT in isolation. What affects poverty and fairness is not the impact of any particular tax but the impact of the tax system as a whole.

So VAT should not to be seen in isolation but in combination, that is, as a part of a total tax structure, including the excise on luxuries and demerit goods, income tax and expenditure policy. What has to be judged is whether the tax structure as a whole is regressive or not.

So the conclusion is that in the usual sense VAT is regressive. It is, however, not more regressive than all other indirect taxes. It can be made less regressive by giving exemptions and zero-rates for items mainly consumed by the poor. But the administrative cost is high and it may not be commensurate with the welfare gain, which is also not so definite.

The redistributional aspect of VAT has to be seen in combination with excise duty, income tax and a suitable expenditure policy. What has to be seen is that the whole tax structure is progressive in that it raises enough revenue to make sufficient funds available for pro-poor welfare expenditure.