Confusing Income Effects and Substitution Effects
with a digression on a universal basic income
Someone argues that, to get more and better medical drugs, we should reduce the unnecessarily high cost of getting new drugs approved. A listener disagrees on the grounds that the companies currently make enough money to afford to fund to research more drugs if they wanted to.
A critic of a negative income tax proposal argues that the effect will be to make poor people less willing to work. A defender of the plan responds that even with what the negative income tax will pay them there will still be things that poor people want and need additional income to buy.
In both cases the response interprets the argument as based on an income effect. In both the response is irrelevant if the argument is actually about a substitution effect.
In the first case, the listener interprets the argument as “drug companies don’t have enough money to afford to research more drugs.” But however much money it has, a drug company will only research a new drug if it believes that future revenue from the drug will more than pay for the cost of developing the drug and getting it approved. Even if it has the money it is not going to throw it away on research that it expects to cost more than it is worth.
If a drug company does have projects that it expects to make money it does not have to pay for them out of revenue — there is a capital market out there for financing profitable projects. The argument against unnecessarily expensive requirements for approval is not that the company needs more money but that the lower the cost of getting a drug approved, the more drugs will be worth researching.1 The rebuttal is irrelevant to that argument.
That is less true in the second case. One reason why a negative income tax might result in poor people working less is it would leave them with less need for money, hence less willing to bear the costs of earning more. That is the income effect. The more important things there are that require additional income to buy the less one would expect that effect to be.
What the argument is missing is the substitution effect, the fact that with a negative tax rate of 50% every dollar the recipient earns reduces his payment from the negative income tax by fifty cents, converting a wage of ten dollars an hour to one of five dollars an hour. A worker is less willing to work for five dollars an hour than for ten, so the negative income tax will result in workers working less.
The point may be made clearer by considering two alternatives to a negative income tax, a universal basic income and an income guarantee. A basic income gives everyone a fixed amount, say ten thousand dollars a year. The income effect is still there but, since the amount received does not depend on how much the recipient earns, it looks as though there is no substitution effect.
An income guarantee pays the difference between the guarantee, say ten thousand dollars a year, and what the recipient earns. If he makes nothing he gets ten thousand, if he makes five thousand he gets five, if he makes ten thousand, or any larger sum, he gets nothing. Now the substitution effect, for any income below ten thousand, is total — for each dollar he earns his payment is reduced by a dollar so he is working for nothing. Which is a good reason not to.
It sounds as though the basic income is the best of the three, but there is a reason I said “it looks as though.” A basic income has to be paid for. A ten thousand dollar basic income going to three hundred and thirty million people costs 3.3 trillion dollars a year, roughly half the total revenue of the federal government. Funding it, assuming the government is not willing to cut other expenditures by that amount, requires the government to find some way of getting more money. 3.3 trillion is about the total amount brought in by the federal income tax. Doubling tax rates to fund a basic income would produce a large substitution effect as taxpayers receiving sixty cents on the dollar instead of eighty substitute away from activities that produce taxable income. And because taxable income would fall, doubling tax rates wouldn’t actually bring in enough money.
A lot of people like the idea of a basic income in the abstract. Few, in my experience, have done the arithmetic. What they are imagining is a guaranteed income high enough that poor people don’t have to work if they don’t want to. Ten thousand a year per person is low for that purpose, barely sufficient, by current standards, for a couple to live on, insufficient for a single individual. But ten thousand is enough to produce serious problems in raising the money.
One possible response is that we can fund a demogrant by cutting other programs that will no longer be necessary. What programs?
Federal Expenditures: 2023
Health insurance programs: 1.6 trillion, half on medicare
Social Security: $1.4 trillion
Defense: 820 billion
Economic security programs:2 545 billion
Veterans and fed retirees benefits: $481 billion
All other programs:3 About 1 trillion
Interest on the debt: $658 billion
Replacing social security by a $10,000 demogrant would be a drastic income reduction for people on social security. Medicare could be replaced by private medical insurance paid out of the demogrant, but that is likely to cost enough to absorb half or more of the demogrant. With Social Security, Health insurance programs, defense, veterans and retiree benefits and interest on the debt eliminated, remaining expenditures are just over a trillion and a half, of which less than half a trillion is in categories for which a demogrant would at least in part substitute.
What about state and local spending? In 2021, total state and local expenditure was 3.7 trillion dollars. A little less than a quarter of that was on public welfare, which could arguably be replaced by a basic income. So if all state and local welfare spending and all federal economic security programs, including unemployment insurance, were eliminated, that would free up enough money to pay about half the cost of a $10,000 basic income. I doubt that many supporters of a basic income would prefer a basic income of $5,000 a year in a world with no welfare payments, unemployment insurance, food stamps, or school meals an improvement on the present situation.
Past posts, sorted by topic
A search bar for past posts and much of my other writing
For evidence for that proposition, see “Evaluation of Consumer Protection Legislation: The 1962 Drug Amendments” by Sam Pelzman.
Unemployment insurance, food stamps, school meals, the refundable part of the Earned Income Tax Credit,…
They include investing in education; investing in basic infrastructure such as roads, bridges, and airports; maintaining natural resources, farms, and the environment; investing in scientific and medical research; enforcing the nation’s laws to promote justice; and other basic duties of the federal government. A very small slice — 1 percent of the budget — goes to programs that operate internationally, including humanitarian aid and the operation of U.S. embassies and consulates.(Center on Budget and Policy Priorities)
Years ago I did the same math, and went from supportive of UBI to against it. It's not going to happen. I had concerns about the disincentive to work and distortionary effects. I was and am pretty confident it will cause a lot of inflation, possibly equal to the amount of the UBI (i.e., we're not producing anything more and people get more money = everything gets more expensive in dollars, or worse, the work disincentive means we make fewer things).
All of those things pale in comparison to the core math problem of funding it. Anything affordable does too little to matter. Anything that does enough to matter isn't affordable.
> there is a reason I said “it looks as though.” A basic income has to be paid for. ... Doubling tax rates to fund a basic income would produce a large substitution effect
I quantified this in a comparison of UBI vs negative income tax a couple months ago: https://governology.substack.com/p/ubi-vs-negative-income-tax . The result is that negative income tax would be massively more efficient than a UBI in our current tax regime (of income taxes).