67 Tax Notes 1382

Tax Notes

JUNE 5, 1995

LENGTH: 761 words

DEPARTMENT: Current and Quotable: Public Comments on Proposed Regulations (COM)

CITE: 67 Tax Notes 1382

HEADLINE: 67 Tax Notes 1382 - CONSUMPTION VS. WAGE TAX.

AUTHOR: Johnson, Calvin H.

University of Texas School of Law

SUMMARY:

The following letter was sent by Calvin H. Johnson, Andrews & Kurth Centennial Professor of Law at the University of Texas School of Law, to Eric Toder, Deputy Assistant Treasury Secretary for Tax

TEXT: 23 MAY 95Dear Eric:

We should be testing the legitimacy of a wage tax according to how well it corresponds to a consumption tax. To do this, we need to abandon the wage tax as the baseline and use consumption tax as the baseline. Considering winners and losers, consumption tax is legitimate while a wage tax is not. A consumption tax is also legitimate when it taxes investors who get higher-than-market-rate returns. These issues are important as we debate whether to move toward use of a consumption or earnings tax.

Assume that twin brothers, Jacob and Esau, work for a year for wages W ($1,000), invest W for n (seven) years at rate r (10%) and consume the output at the end of the last year. The investments are also very volatile: e.g., there is a bet imbedded in the investments by which each $l will either become a payoff of $1.90 or 10 cents with equal odds. The 10-cent payoff leaves penury, but (just) allows survival. Assume indentured servitude is available (slavery in Egypt, perhaps) if either has too little to live on, albeit at considerable pain. We know that Jacob will win the bet and Esau will lose it, but no one knows that in advance.

In absence of tax, Jacob or Esau each have an expectation of ending up with $1,949. Either one expects:

________________________________________________________________________________

(1) 50%*1.90*W(1+r) /n+50%*.10*W(1+r) /n which equals

W*(1+r) /n = $ 1000(1+10%) /7/ = $ 1,949

________________________________________________________________________________

Assume that revenue needs can be met with a tax at rate t (40%). With either a consumption tax on end results or a wage tax of 40% on the $1,000, the expected final position is the same no matter what rate r we assume. With a wage tax, the investable amount W is reduced immediately by (1-t) so that the expectation is

________________________________________________________________________________

(2) W*(1-t)*(1+r) /n*50%*1.90+W*(1-t)*(1+r)

/n*50%*.10 which equals W*(1-t)*(1+r) /n = $ 1,169.

________________________________________________________________________________

With a true consumption or outcomes tax, both the winning and losing outcomes are reduced to 1-t and of course, under the cumulative law of multiplication, the expectation is the same as in (2):

________________________________________________________________________________

(3) W*(1+r) /n*50%*1.90*(1-t)+W*(1+r) /n*50%*.10* /

(1-t) r which equals W*(1+r) /n*(1-t)= $ 1,169.

________________________________________________________________________________

A wage tax will collect $400 times 2 now, and a consumption tax will collect $779 times 2 in seven years, but at r = 10%, those collections have the same value.

Money, however, has a steep-curve diminishing utility, so we cannot fairly assume fungibility of dollars in the hands of winners and losers. Thus Esau, the loser on the bet, has only .10*E(1+r) /n or $195 before tax, just enough to survive. A tax of 40% either at the beginning or the end will send Esau down to $117 and into painful slavery. Jacob, by contrast, has won and has 1.9*$1,949 or $3,703 before tax, which, at the time, bought you ambrosia and nectar. We need to be collecting all of the necessary tax from the winner, Jacob, who will just reduce luxury, and none of it from Esau. Collecting all tax from Jacob will minimize the total harm that tax does to the brothers.

Collecting all tax from Jacob is Pareto improving ex ante. We cannot identify the winner of the bet in advance, but shifting the entire tax burden, including the loser's t*10%*W*(1+r) /n, all onto the winner will improve expected value for both. First if we ignore varying utility of money with different outcomes, shifting all tax to Jacob will not diminish expected value:

________________________________________________________________________________

(4) 50%*[W*(1+r) /n*1.90-t*2*W*(1+r)n]+50%*W* (1+r)

/n*.10 = 55% W*(1+r) /n + 5% W*(1+r) /n = (1-40%)* W*(1+r)

/n= $ 1,169

________________________________________________________________________________

which is the same as expresssions (2) and (3).

When we factor in that Jacob and Esau both know that the tax that reduces ambrosia and nectar hurts less than a tax that forces slavery, both will have a higher expected value from expression (4) than from either (2) or (3) and both will more freely undertake risky investments. The shift of the tax burden to the winner has gotten rid of a scary penalty on risk.

Given the Pareto improvement, we should not object when Jacob, who gets an outcome better than r, pays more tax than Esau, who earns less than r.

Please call if I can provide any further assistance.

________________________________________________________________________________

Sincerely yours,

Calvin H. Johnson

Andrews & Kurth Centennial

Professor of Law

cc: David Bradford

Don Fullerton

________________________________________________________________________________