Recognizing the calamitous erosion of incentives that would be brought about by a straight guaranteed-income plan, some reformers have advocated what they call a “negative income tax.” This proposal was put forward by the prominent economist, Professor Milton Friedman, of the University of Chicago, in his book Capitalism and Freedom, which appeared in 1962. The system he proposed would be administered along with the current income tax system.
Suppose that the poverty-line income were set at $3,000 per “consumer unit” (families or individuals), and suppose that the negative income tax (which is really a subsidy), were a flat rate of 50%. Then every “consumer unit” (this is the statisticians’ technical term) whose income fell below $3,000 would be paid a subsidy of, say, 50% of the difference. If its earned income were $2,000, for example, it would receive $500; if its earned income were $1,000 it would receive $1,000; if its earned income were zero it would receive $1,500.
Professor Friedman freely concedes that his proposal, “like any other measure to relieve poverty … reduces the incentives of those helped to help themselves.” But he argues that “it does not eliminate that incentive entirely, as a system of supplementing incomes up to some fixed minimum would. An extra dollar earned always means more money available for expenditure.”
It is true that a “negative income tax” (which is a misleading name for a tapered-off guaranteed income) would not have quite as destructive an effect on incentives as would a straight guaranteed income. In fact, some thirty years ago I put forward a similar proposal myself. This appeared in an article in The Annalist (a weekly then published by the New York Times) of January 4, 1939. I suggested what I called a “tapering subsidy,” a relief payment that would be reduced by only $1 for every $2 the relief recipient earned by himself.
But I abandoned the proposal when I realized that it leads straight into a dilemma, which is precisely the dilemma of the negative income tax: either it is altogether inadequate at the lower end of the scale of self-earnings, or it is unjustifiably excessive at the higher end. Either it must pay only half an adequate income (by its own definition of “adequate”) to a family that earns no income, or it must pay nearly twice an adequate income to a family that already earns an almost adequate income.
Trick names of this sort corrupt the language and confuse thought. It would hardly clarify matters to call a handout a “negative deprivation” or having your pocket picked “receiving a negative gift.”
The problem that the NIT (negative income tax) evades or glosses over is the problem of the individual or family with zero income. If an individual were given only $300 (the figure suggested in Professor Friedman’s original proposal in 1962), nobody would regard this as nearly adequate — particularly if, as Professor Friedman also proposed, NIT were made a complete substitute for all other forms of relief and welfare. If the NIT payment for a family of zero income is set at $1,700, no advocate of the guaranteed income would regard it as adequate to live on in “decency and dignity.” So if the NIT were ever adopted, the political pressure would be irresistible to make it provide the minimum “poverty-line” income of $3,400 even to families with zero earned income.
The basic subsidy would therefore be as great as under the straight guaranteed income. But if the basic subsidy under NIT to a family with zero income were $3,400, then under the NIT 50% “incentive” formula that family would continue to get some government subsidy until its annual income reached $6,800. But this is higher than the median family income for the whole country in 1963 ($6,637). In brief, this would be fantastically expensive.
In addition, it would raise serious problems of equity. When the subsidized family was earning $6,798 income it would still be getting a $1 subsidy. When it earned $6,802 would it fall off the gravy train entirely, and have to wait until its income fell below $3,400 before it could get on again? And what about the family that had been earning $3,402 all along, and had never got on the gravy train?
The arithmetical dilemma of the NIT has received so little attention from its advocates that I hope I may be forgiven another illustration to show the paradoxical way in which it would work out.
An orthodox relief program would pay the jobless head of a family, say, $60 a week. If he then started to earn something, he would be paid simply the difference between that amount and $60. Under the NIT principle a man who was earning nothing would also receive a relief payment of $60 a week. But if he then earned $30 a week on his own he would still get a $45 payment (reduced by only $1 for every $2 earnings), bringing his total income to $75 a week. If he was later able to earn the full $60 for himself he would still be getting a relief payment of $30 a week, bringing his total income to $90. In fact, even if he succeeded in bringing his total self-earnings to $118 a week he would still be getting $1 a week in relief payment.
He would then be almost twice as well off economically as he would if he had always earned enough — say, $61 a week — not to get on the relief rolls in the first place. This would be clearly inequitable to those who had never got on relief. The incentive to get on relief, and certainly to stay on relief, would be enormously greater under NIT than under the present system.
If we tried to escape this result by using the NIT formula only in part — taking the man off relief, say, as soon as he was himself earning $60 a week — we would get an even more absurd result. When he was earning $58 a week under NIT, he would still be getting $31 a week from the government, making his total income $89. But if he then made the mistake of earning only $2 more he would end up with a net loss of $29 a week. So the negative income tax would create a tremendous positive incentive to get and stay on relief permanently.
The NIT scheme could avoid this preposterous result by paying a man with zero income only, say, $30 a week, or only half as much as its own proponents assume that he needs to live on.
Some readers may think that the dilemma of the NIT scheme can somehow be escaped by changing the percentage by which the relief payment or income supplement is reduced as self-earnings increase. But any change from 50% one way or the other merely reduces one horn of the dilemma by making the other more formidable. If we reduce the government supplement by 75 cents for every dollar of self-earnings, we correspondingly reduce or destroy the incentive for such self-earnings. If we reduce the government supplement by only 25 cents for every dollar of self-earnings, we increase the recipient’s incentive to work and earn, but at the cost of a still more expensive program for the government, and we increase the recipient’s positive incentive to stay on relief because of the violent drop in his income if he ever got off.
If we make the scheme more complicated by, say, reducing the relief payment or supplementary income by only 25 cents for every dollar of the recipient’s first $10 of weekly self-earnings, 50 cents for every dollar of his second $10 of self-earnings, and 75 cents for every dollar of his third $10 of self-earnings, or some similar scheme, we merely pile up an administrative nightmare without solving the basic dilemma. The unpalatable truth seems to be that whenever we try to “increase incentives” by reducing a relief payment by less than a dollar for every additional dollar of self-earnings, we solve an immediate problem at the cost of building up a bigger problem for the future.
In addition to the special dilemma it presents, the NIT retains the fatal defects of the straight guaranteed income. By neglecting the careful applicant-by-applicant investigation of needs and resources made by the ordinary relief system, it would open the government to massive fraud, chiseling, and swindling. And it would also, like the guaranteed income, force the taxpayers to support a man regardless of whether or not he was making any sincere effort to support himself. The government is bound to get into insoluble difficulties if it starts to give money away to “the poor” not only without making sure that they are poor but without bothering to find out the reasons why any particular individual or family is poor.