Issues in the Design of
a Value-added Tax
In appraising the economic effects of the value-added tax and thus the desirability of an American VAT, it is generally satisfactory to consider whether the United States should adopt a tax with the economic effects of a federal retail sales tax. Although the two taxes and their economic effects would not be exactly equivalent, they would be sufficiently similar to make the comparison valid for most purposes. Having said that, however, I must emphasize that the administrative techniques employed to collect a VAT do matter. First, the VAT treats capital investment, services, and international trade more satisfactorily than a retail sales tax, at least as levied by the states; this is explained further in chapter 7.
Second, the textbook descriptions of the subtraction and credit methods presented in chapter 3 obscure important differences between the two methods. This chapter explains that in its simplest form—what I call the naive subtraction method—the subtraction method is truly satisfactory only if the VAT is levied at a single rate on a comprehensive tax base. 1 Once multiple rates and exclusions from the tax base enter the picture, this system breaks down.
Some of the defects of the naive subtraction method can be remedied by adopting a more sophisticated version that allows deductions only for purchases on which tax has been paid. But even this sophisticated subtraction-method VAT, which is essentially equivalent to a credit-method VAT, functions satisfactorily only if levied at a uniform rate (with at most a zero rate in addition to the uniform positive rate). Given the likelihood that political pressure for multiple rates would be irresistible, this conclusion throws considerable doubt on the wisdom of enacting a subtraction-method VAT, such as the business transfer tax (BTT) proposed by Senator William Roth.As the Treasury Department concluded in its 1984 assessment of VAT: "While there are different forms of value-added tax and alternative