The hardest problem in public policy
By Scott Sumner
A recent post discussing a policy of universal basic income (UBI) is less convincing:
About 40 million Americans live below the poverty line, which is $12,000 for an individual and a little higher for families. Multiplying these out to get $480 billion to end poverty is too high, first because most of these people live in families with each other, and second because most of them already have some income. Let’s halve it to $240 billion.
The top 1% are currently taxed at a rate of 37%, and this brings in about $560 billion. Increasing it to an even 50% would give an extra $200 billion or so, leaving $40 billion to get from random other places, the top 2%, etc. For a basic sanity check, the Bush tax cuts decreased revenue by $180 billion per year. A tax increase of the same scale as the Bush tax cuts would suffice for a basic income that ended poverty.
Isn’t basic income supposed to be universal? Yes, but most serious proposals accept that it will be gradually reabsorbed as higher taxes. People below some income gain money on net, people above the income lose money on net, and there’s some break-even point. I’m proposing the break-even point is somewhere between the poverty line and the top 1% – not in an abrupt way that forms a welfare cliff, but gradually according to the normal progressive tax system, and at a level so that we can imagine it abstractly as transferring money from the top to the bottom, with everyone else ending up about equally well-off, or getting gains and losses that cancel out.
This ignores the concern that higher taxes would stifle the economy, and the concern that the promise of a UBI would make more people quit their jobs and fall into the income stratum that benefits. But it also ignores the hope that lifting everyone out of poverty would obviate some welfare programs, or improve education, or bring other economic benefits. I don’t want to claim to be able to calculate all of these considerations, but order of magnitude estimate, we could give out a UBI sufficient to end poverty with a medium-sized tax increase.
I understand what he is trying to do here, but I think it’s a mistake. Unless I’m mistaken, it would be like describing the cost of the Social Security program not in terms of gross expenditure, but in terms of net redistribution. Or it would be like describing the tax burden of “Medicare for all” in terms of the difference between what people currently pay for their private health care plan and the extra taxes they would pay under the new system.
There are two ways to provide a basic income, which look very different but are effectively identical. In one version, only the poor and the near-poor receive benefits, and those benefits phase out as one’s income rises. In a second version, everyone receives an identical basic income, and the entire program is financed by a progressive income tax (or progressive consumption tax.)
Although these two versions look very different, they are effectively identical. Indeed you can draw up a version of each form of basic income that has exactly identical net benefits, exactly identical net taxes, and exactly identical marginal tax rates at each income level. And yet the gross taxes paid under the two regimes will be vastly different. (I believe Alexander does understand this point.)
In my view, the intellectually honest way of describing the tax cost of a basic income program is to assume that everyone gets the $12,000 income, and the entire program is financed by a progressive tax system. This makes the program look really, really expensive, but that’s because it is really, really expensive.
Today it’s increasingly common to see people gloss over the incentive effects of higher MTRs, as neoliberalism is out of fashion. But imagine someone saying, “Ignoring the incentive effects on agricultural output, Mao’s system of communal agriculture was an effective way of improving equality.” Tens of millions died in that egalitarian experiment, perhaps the worst disaster in human history. On the other hand, the inadequate response of the British government during the Irish potato famine may have been partly motivated by worry about the disincentive effects of charity. This is the hardest problem in public policy.
Obviously a UBI in the United States wouldn’t produce a “Great Leap Forward” style disaster. But the reason that won’t happen is that it’s likely to be financed with a really, really expensive tax increase, a tax increase vastly greater than the (modest) net cost of moving 40 million American up to the poverty line (a cost Alexander describes in the first quoted paragraph). I haven’t studied this issue in detail, but given the US population (330 million, and perhaps half as many households) it’s obvious that a UBI of $12,000 per person and significantly more per household would cost trillions of dollars. There is no plausible income tax increase that could raise that sort of revenue.
Defenders of the UBI would correctly point to the fact that you can supplement the income tax boost with a 20% VAT, and for many middle class families the extra UBI benefits would roughly offset the extra VAT and income tax. But it doesn’t end there. European countries with high levels of government spending and taxes typically have economies where GDP/person is 20% to 40% lower than in the US, mostly due to less work effort, although productivity is also lower in some cases. Basic economic theory predicts that Europe’s high taxes should result in substantially lower GDP/person, mostly due to less work effort, but also less productivity. The UBI can be seen as a gamble that these basic economic theories are wrong, even though the stylized facts strongly support the claim that taxes reduce output in the long run. Even within Europe, less taxed areas such as Switzerland tend to have higher output.
If standard economic theory is correct, then a UBI will not merely redirect money from the rich to the poor, it will dramatically reduce (material) living standards for average Americans. There will be more leisure, but it will be unevenly distributed.
This is not to say a UBI is a bad idea. It seems plausible to me that the case for UBI will gradually increase over time. Perhaps in 100 or 200 years, today’s opponents of the UBI will look mean-spirited or foolish. But that doesn’t mean those future people will be correct about today, any more than our current (negative) views about Bronze Age wars are correct. What do we know about the trade-offs faced by people in the Bronze Age? (Hint: the Malthusian trap meant it was often a choice between death through war and death through starvation.)
Let’s keep studying the UBI, but it’s far too early to try the experiment. Especially with employers currently desperate to hire almost anyone, even people just out of prison.
PS. I said that Alexander is fair-minded. The quotation I provided above is directly preceded by this:
Everything below is epistemic status: wild speculation.
PPS. Presidential candidate Andrew Yang is running on a UBI platform: