Steensland would probably dislike the title of this review. The reason is that he sees the defeat in the U.S. Senate of Nixon’s proposed Family Assistance Plan, which the House of Representatives passed in 1970 by a vote of 243-155, not as dodging a bullet but, on the contrary, as a huge lost opportunity. But one doesn’t have to agree with his perspective to learn from his book.
Steensland gives a detailed account of the various proposals for some version of a GAI [Guaranteed Annual Income], starting with the Johnson administration, through the Nixon, Ford (briefly, given his two years in office), and Carter administrations. Although he is a sociologist, he exhibits a basic understanding of the effects of economic incentives on work and family dissolution. And, refreshingly, his book is relatively free of cheap shots at those with whom he disagrees. He tries hard to understand their views rather than merely dismissing them as unworthy. That makes it relatively easy to judge the arguments of the various players and come to conclusions different from Steensland’s. In particular, I found the arguments against Nixon’s proposal by some of Nixon’s own staff much more convincing than Steensland did.
This is from David R. Henderson, “How We Dodged a Bullet,” my review of Brian Steensland’s The Failed Welfare Revolution: America’s Struggle over Guaranteed Income Policy in Regulation, Spring 2008.
Another excerpt:
You can take one side or the other on each of the above issues. But, as one of Nixon’s main advisers on welfare, Daniel Patrick Moynihan, became famous for saying, “Everyone is entitled to his own opinion, but not to his own facts.” I repeat the quote because Moynihan, in pushing for one version of a plan, told Nixon a whopper. In an April 1969 letter to Nixon, Moynihan wrote, “For two weeks’ growth in the Gross National Product you can all but eliminate family poverty in America.” Um, no. The typical annual growth rate in those days was just above 3%, so two weeks of growth was about 0.12% of GNP. Even the cheapest plan would have cost well over half a percent of GNP, which is more than four times Moynihan’s estimate.
And, finally, the relevance for today’s policy issues:
Because there is a renewed push today for some form of GAI–even some libertarians advocate it–there’s a lesson to be learned from Steensland’s book. Some libertarians claim that we could have a reasonable GAI by cutting all other means-tested benefits. This turns out to be false, as I showed in “A Philosophical Economist’s Case Against a Government-Guaranteed Basic Income” (Independent Review 19[4], 2015). But even if it were true, the Nixon proposal was for a major expansion of welfare spending, with no offsetting cuts in means-tested programs such as housing aid and food stamps. The lesson: libertarians who push for a GAI will end up being in an alliance whose main constituents want more government spending. And that’s what they’ll probably get.
READER COMMENTS
Maniel
Mar 24 2018 at 1:19pm
What is your view of Dr. Friedman’s negative income tax idea?
David R Henderson
Mar 24 2018 at 1:34pm
@Maniel,
What is your view of Dr. Friedman’s negative income tax idea?
Short answer: I’m against it.
Short reasoning: It’s very expensive because the other programs he wanted to get rid of wouldn’t be abolished. Also, the implied implicit marginal tax rates would be very high. Also it would get millions of people on “welfare” who would have never thought to be on welfare. Finally, I like much better the welfare reform law of 1996 that put a 5-year lifetime limit on receiving welfare. That way, you don’t have to worry that the implicit marginal tax rates with our current system are even higher than with the NIT because that’s a problem for only 5 years.
BC
Mar 24 2018 at 3:03pm
I initially found two features of GAI attractive: (1) efficiency of cash over in-kind benefits and (2) lower effective marginal tax rates in phase out region of means-tested benefits. However, I now agree that pure GAI would likely be too expensive for the reasons that David and others have pointed out.
Also, we can already gain much of these two features incrementally. For example, we can convert existing in-kind benefits to vouchers, e.g., school choice and medicare, and convert existing vouchers to cash or something more cash like, e.g., changing food stamps to cash or combining food stamps, housing aid, health insurance subsidies, etc. into one super-voucher that can be spent on any combination of food, housing, health insurance, etc. We can also just phase out means-tested benefits more gradually to trade off effective marginal rates in the phase out region, benefit levels, and marginal rates at the high end. In fact, one can view GAI as stretching the phase out to infinitely high income, which probably is not the best tradeoff.
The fact that it’s politically difficult to even incrementally convert in-kind benefits like K-12 education and Medicare to vouchers seems to me to be evidence that GAI would indeed expand welfare rather than replace it. If we can’t even cut direct government spending on a K-12 school when offset with tuition vouchers, then how could we cut spending on that same school when offset with some general cash benefit like GAI?
Maniel
Mar 24 2018 at 3:12pm
@Prof. Henderson
Short response: I like the idea of a NIT paired with a flat tax because it 1) would include the same incentive to work for all and 2) would replace welfare bureaucracies with one simple system. I accept your observations on the cost side, but a lot depends on where rates and thresholds would be set. In practice though, I agree that “other programs … wouldn’t be abolished.” Sigh!
Thaomas
Mar 24 2018 at 4:58pm
“The typical annual growth rate in those days was just above 3%, so two weeks of growth was about 0.12% of GNP. Even the cheapest plan would have cost well over half a percent of GNP, which is more than four times Moynihan’s estimate.”
I think Monynihan is mixing concepts here. The “cost” of a fiscal proposal is dead-weight loss caused by the the tax that finances it which IS pretty small compared to the real income effects of GDP growth. Now if we are only discussing the size of the transfer from some taxpayers to others with gdp growth as a unit, the comparison is valid though possibly miscalculated as Henderson points out.
I think rather than a GAI, we should have a much higher EITC. This would offset some (or more than all?) of the dead weight loss of whichever tax finances the transfer with the additional work effort incentivized by the EITC.
David R Henderson
Mar 24 2018 at 5:17pm
@BC,
Also, we can already gain much of these two features incrementally. For example, we can convert existing in-kind benefits to vouchers, e.g., school choice and medicare, and convert existing vouchers to cash or something more cash like, e.g., changing food stamps to cash or combining food stamps, housing aid, health insurance subsidies, etc. into one super-voucher that can be spent on any combination of food, housing, health insurance, etc.
Yes. Another candidate is VA benefits.
The fact that it’s politically difficult to even incrementally convert in-kind benefits like K-12 education and Medicare to vouchers seems to me to be evidence that GAI would indeed expand welfare rather than replace it. If we can’t even cut direct government spending on a K-12 school when offset with tuition vouchers, then how could we cut spending on that same school when offset with some general cash benefit like GAI?
Exactly.
@Maniel,
Sigh!
Yes.
@Thaomas,
The “cost” of a fiscal proposal is dead-weight loss caused by the the tax that finances it which IS pretty small compared to the real income effects of GDP growth.
Not correct. The cost includes deadweight loss, which, you might notice, I didn’t include, and so I understated the cost. The cost is not just the deadweight loss. It’s the revenue + deadweight loss.
ColoComment
Mar 24 2018 at 6:00pm
I recently finished this book on the history of U.S. entitlement programs. Great read. Short summary: they never go away and they always expand well beyond the original promise & intent.
https://www.amazon.com/High-Cost-Good-Intentions-Entitlement/dp/1503603547
I would much rather go after fraudulent and/or improper payouts in, and improving the performance of, the systems we already have, before moving to [experimenting with… again?] some other program(s), and starting the cycle all over again.
Or, ditch it all and go back to letting local sources manage local charity & welfare needs.
John Alcorn
Mar 24 2018 at 7:34pm
@ BC,
Thank you for posting a very helpful comment.
Matthias Görgens
Mar 25 2018 at 1:38am
Another thing to keep in mind is that given America’s current restrictive housing situation, a big part of any such assistance will probably accrue to landlords/landowners.
It’s not just ‘tax incidence’ that’s worth studying, but also ‘welfare incidence’.
(However doing a land value tax and distributing the proceeds equally amongst the population would probably work fairly well.)
Matthias Görgens
Mar 25 2018 at 1:38am
Another thing to keep in mind is that given America’s current restrictive housing situation, a big part of any such assistance will probably accrue to landlords/landowners.
It’s not just ‘tax incidence’ that’s worth studying, but also ‘welfare incidence’.
(However doing a land value tax and distributing the proceeds equally amongst the population would probably work fairly well.)
Thaomas
Mar 25 2018 at 10:53pm
Dave,
Do we have different conceptions of cost?
If we tax Peter $1.00 to give Paul a benefit of $0.99, I’d say the cost was $0.01. Do you say it is $1.00 or 1.01?
Greg
Mar 27 2018 at 2:12am
The cost is almost certainly $2-5.
You have not accounted for the IRS bureaucracy, the federal marshall’s service, the federal prison system and the ‘pay your taxes’ propaganda ministry needed to collect $1 from Peter ‘voluntarily’.
If you believe that 99% of our tax dollars are spent on social benefit payments, you are clearly too uninformed to be devising policy just yet.
David R Henderson
Mar 27 2018 at 9:44am
@Thaomas,
If we tax Peter $1.00 to give Paul a benefit of $0.99, I’d say the cost was $0.01. Do you say it is $1.00 or 1.01?
First, it’s not “we.” It’s the government.
But now to your question.
The deadweight loss as a percent of revenue raised, at current tax margins and for most taxes considered, is about 30%. So taxing Peter $1.00 imposes a cost of $1.30.
The benefit is a separate issue. In cost/benefit analysis, we need to keep costs and benefits separate. So the cost of taxing Peter $1.00 is $1.30 and the benefit to Paul is $0.99. So the net cost, not the cost, is $0.31.
Comments are closed.