Sukumar Mukhopadhyay: Is VAT a regressive tax

There is prevailing idea that VAT is regressive. In effect it is or is not is not so simple to pronounce. Many writers have expressed different  opinions on it. So it is better to begin by defining it first.

The usual definition of regressivity is the following: When the ratio of consumption tax to income increases with higher income levels, it is called progressive tax. When this ratio is proportionate it is called 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 incomes than payments by higher income households. That is, the VAT will be regressive .

The design of tax is based on the premise that tax must be fair.  Since moral values change over time, the concept of fairness is not set in stone. The definition of fair is less clear and the interpretation of the concept varies from person to person and from time to time. 

Therefore, it is logical that fairness must be interpreted relative to the extent a person derives benefits from the society, relative to his wealth and relative to what others contribute.   In Europe, when the norm was the absolute power of the ruler, the concept that taxation should be just and fair was not quite the rule.  The rule depended on the discretion of the ruler.  Only in the 18th Century during the Age of Enlightenment, were ideas of fairness in taxation postulated. 

In England, Adam Smith laid down his concept of fairness in four canons, the first of which stated that taxes are fair when the tax burden is universal and applies equally to all in society.     Smith’s idea on fairness, including the notion that the equal must be treated equally and the unequal must be treated unequally in proportion to their individual inequality, became an ethical guideline for the next century, though Smith based his conclusion on an economic approach to taxation and never spoke of equity or justice.

John Stuart Mill and Edwin Seligman argued later that Smith’s approach was not defensible because the benefit of individual members of society could not be accurately measured.  The idea, however, prevailed in subsequent discussions on tax over a period of time that though the benefits of individual members could not be accurately measured, the concept of Adam Smith about fairness of tax remained.  The concept of proportional tax is fundamentally same as Adam Smith idea of fairness. 

High cost of attaining progressivity

There is a great administrative cost for trying to get progressivity by rate differentiation. The advantage of progressivity gained by a whole host of exemptions, zero rating and multiple rates is almost not certainly sufficient to offset the disadvantages associated with departure from a uniform rate and comprehensive base. Exemptions and differential rates distort consumer and producer choices, make for administrative complexity (and hence increased administrative and compliance cost) and raise questions of interpretation that often must be resolved in courts . Any equity gains from differentiation of rates are likely to be more than outweighed by the additional administrative cost entailed. Exemption introduces cascading leading to major distortion and lack of neutrality, which is the virtue of VAT. In Turkey to achieve progressivity in 1991, the total VAT collection, as a percentage of total tax revenue itself was only 24% while half of it was given back as rebate to the lower income groups. This was done through a cumbersome mechanism that required monthly representation of receipts, their verifications, huge bureaucracy and a very high compliance cost .

Empirical Verification of VAT’s Regressivity

            Here is a study on VAT regressivity keeping consumption constant and making it variable thereafter.    In its most basic model, the pro-regressive argument distinguishes between, on the one hand, a low-income category, whose representatives spend their entire income (Y) on consumption (c), i.e. Y=C and, on the other hand, middle-income and high-income categories spending the same amount of money on consumptive expenditure and saving the remainder (Y=C+S; C being constant).  If we take the families of low income, middle income and high income and assume that they all spend the same absolute amount on consumption, their tax burden, as a percentage of their income, is as follows:

The tax burden as a percentage of disposable income is by far the highest for the low income group, and regression is rather steep: as income increases, the tax burden diminishes from 6% to 1.2%

However, the postulate of the regressive model that consumption is a constant is not reliable at all.  The fact is, that people who earn a higher income spend more money, but not necessarily all of it.  Consumption is not a constant but a variable determined by long-term income perspective.  This is relevant because increased consumption is accompanied by increased taxation.

Level of expenditure is variable

If we assume that the families in the middle and high-income groups spend an increasing amount on consumption and particularly on conspicuous, the regression becomes rather moderate:

On the basis such a realistic assumption that consumptive expenditure varies with income, in the sense that expenditure will increase when income increases, though not at the same rate, a VAT remains regressive in nature, but a more nuanced picture appears, showing that VAT has a moderate regression rate. 

Under an income tax, the savings in these examples would have been taxed as earned income.  Under a VAT, savings remain untaxed since they are not used for consumptive expenditure.  The exclusion of savings in the model has a direct influence on the degree of VAT’s regressivity.  A VAT is regressive to the extent that the saving ratio increases with income.