The fallacy of a transaction tax

Bruce Bartlett

6/21/2005 12:00:00 AM - Bruce Bartlett

One reason I thought it was a bad idea to appoint a commission to study tax reform is that its members would waste a lot of time reviewing options that most tax experts rejected long ago.

 I know from experience that everybody and his brother have some brilliant scheme to reform the tax system that may appear superficially plausible. But once it is subjected to careful political and economic analysis, one quickly discovers why no one has ever given it serious consideration.

 Take, for example, the deceptively simple-sounding idea of taxing all financial transactions instead of income, profits or sales. Since the volume of transactions in the economy is many times greater than the gross domestic product, one could theoretically match current federal revenues with a very low tax rate.

 One advocate of a financial transactions tax is economist Edgar Feige, who estimates that a tax of 0.6 percent (60 basis points), with each party to a transaction paying half, would be sufficient to abolish all existing federal taxes. He bases this on an estimated tax base of $337 trillion. By contrast, GDP is only about $12 trillion.

 The tax would be assessed through the banking system, which would deduct the fee on every check, electronic funds transfer or credit card transaction. Currency would be taxed whenever it is deposited or withdrawn. Feige asserts that the collection cost of this tax system would be very low and that it is progressive because the volume of financial transactions rises with income.

 While it is true that a tax of 0.6 percent on a base of $337 trillion would in theory yield $2 trillion -- about what the federal government will raise this year from conventional taxes -- it's important to remember that revenues still must ultimately come out of current production. All goods and services -- including capital investment and intermediate goods -- will have to be taxed dozens of times to get the federal government's current 17 percent share of GDP in taxes.

 However, not all goods and services involve the same number of transactions. Thus, the burden of the tax will vary wildly depending not on the price of a good or its profitability, but solely on whether its production involved many or few financial transactions before it reached the final consumer. Jewelry will bear a low tax, while groceries will be taxed heavily. This makes no sense from a distributional viewpoint.

 Since GDP equals the money supply times the turnover of money -- what economists call velocity -- a fully effective transactions tax will presumably reduce velocity. Consequently, it would be severely deflationary unless the Federal Reserve substantially increased the money supply to compensate. It also means that the tax base will shrink as soon as the tax is imposed.

 A recent study by the International Monetary Fund of real world experiences with financial transactions taxes comes down very hard against them. For starters, by targeting the means of payment, a transactions tax will hinder development of a nation's financial system -- which all economists believe is critical to economic development -- and lead to disintermediation from the banking system. Primitive economies use cash and barter; advanced economies use checks, credit cards and debit cards. So why would we want to encourage people to use primitive financial methods and discourage them from using advanced methods?

 The study notes that wealthy people were able to avoid transactions taxes by using offshore financial institutions and businesses did so by vertically integrating in order to keep transactions in-house and not utilize the banking system. It also found that governments often penalized themselves with such taxes because much financial activity involved trading in government securities. The tax raised interest rates, thereby forcing governments to pay more to finance their debts.

 The problems of transactions taxes have always caused them to break down within a short time. Generally, they have been instituted only in relatively undeveloped nations experiencing extreme financial crises that demanded a large, temporary injection of revenue. There is no reason to think that financial transactions taxes could form the foundation of a tax system. "The use of these taxes should be avoided," the IMF study concluded.

 The nominal rate of taxation is very important, but it is only one factor in determining how burdensome a tax is. If the same income is taxed over and over again, the effective tax rate will be far higher than the statutory rate. Or if the tax base is badly defined, then a low tax rate can be much more debilitating than a higher rate on a proper base.

 The only attraction of a financial transactions tax is the low rate. But it is a chimera and would have many undesirable economic consequences if enacted. We should stick to taxes we better understand and know will work in the real world.