AI strengthens the case for a Universal Basic Income (UBI), but not necessarily as a substitute for paid employment. Back in the 1990s, I independently reinvented the UBI without knowing that the idea already existed and already had a name! But here's the rub: I thought of the UBI not as a job substitute, but as a job creator.
How so? A UBI, as the ‘U’ implies, is payable to wage-earners, so that their wages only need to supplement the UBI. Thus a UBI is, among other things, a wage subsidy. And indeed, given the social costs of unemployment, the optimal (‘Pigovian’) tax rate on the employer-employee relationship is negative: for maximum economic efficiency, at least for typical wage levels, hiring should be subsidized, not taxed.
To see how the subsidy works, suppose you want to set the minimum wage so that it covers the cost of living. Without a UBI, the minimum wage must cover the whole cost. With a UBI, the minimum wage only needs to cover the shortfall left by the UBI. Hence employers find it more attractive to hire workers, because the marginal cost is reduced, and less attractive to fire workers, because the marginal saving is reduced.
Or suppose you have a minimum wage that doesn't cover the cost of living, or no minimum wage at all. Without a UBI, lower wages make it more attractive to hire workers at a given level of demand for the employers' products, but also reduce that level of demand, so that they don't necessarily yield full employment. Adding a UBI restores demand for the employers' products and relieves the poverty of the workers.
As long as the UBI isn't funded by any additional tax on labor (at least up to a certain wage level), it doesn't reduce the attractiveness of hiring (up to that wage level). So how should it be funded? We already have two ingredients of the answer:
- To the extent that the UBI replaces part of the wage bill, it can be funded by turning that part of the wage bill into part of the tax bill.
- By creating more jobs, a UBI creates a bigger economy, with more capacity to pay for things—such as a UBI.
But further ingredients come to mind:
- The economy expands not only because employers become more willing to hire, but also because individuals become more willing to upskill, or start businesses. For as much as the UBI reduces the marginal cost of hiring, it reduces the marginal cost of taking time out to build a new career, moderating the risk and increasing the capacity to persevere.
- Less poverty means fewer poverty-driven evils that politicians will be obliged, or at least tempted, to throw money at.
- A UBI is not a net addition to a welfare system; the rest of the system is scaled down as special-purpose welfare payments are replaced by, or reduced to supplements to, the UBI.
- Tax reform can minimize the size of the UBI that needs to be funded on-budget, in two ways. First, make it uneconomic to own residential land without housing people on it. This minimizes rents (relative to capacity to pay), hence the size of the UBI needed to address the cost of living. Second, make it uneconomic to own commercial/industrial land without employing people on it. This minimizes the size of the UBI needed to give full employment. I have likened these tax arrangements to “An OFF-BUDGET Universal Basic Income”; but they are also needed to prevent any on-budget UBI from being competed away in higher rents.
- There are ways of making a Basic Income (BI) less than universal, hence cheaper, without defeating its purposes of creating jobs and alleviating poverty. An income test would defeat those purposes by denying the UBI to workers and the self-employed. But an assets test would be compatible, especially if it were limited to kinds of assets that you can't produce, so that it can't deter their production. And an activity test would be compatible provided that it recognized a sufficiently wide range of activities (paid employment, self-employment, education/training, and, preferably, productive non-market activities such as child-rearing and housekeeping) and had appropriate exemptions (e.g. for the old and the disabled). Hence I don't include the word universal in the title.
By the way, wherever there is an income tax, income testing is a fraud. An income-tested benefit is equivalent to a non-income-tested benefit plus a tax at the same rate(s) on the same base as the income test. You can't do anything with income testing that you can't do with income taxation, except hide the obscene effective marginal tax rates due to the combination of taxing the earned income and withdrawing the benefit. If income testing were reclassified as taxation, so that the effective marginal rates became the official rates, public opinion would demand that the rates be smoothed in a revenue-neutral manner; and the ensuing reduction in disincentives would expand the economy.
Inevitably someone will say “Wouldn't the reduction in unemployment be inflationary?” No, because it would be accompanied by—indeed caused by—a reduction in the marginal cost of labor.
Inevitably someone will say “Doesn't a Basic Income destroy the incentive to work?” No!—because you don't lose the BI if you start working! If the legislators really wanted to destroy the incentive to work, they would invent a benefit that is paid to the unemployed but withdrawn when they start earning... Oh, wait, that's exactly what happens with unemployment benefits in most developed countries! Hence the aforesaid “obscene effective marginal tax rates”.
But even that all-out war on the work ethic has not succeeded. The unemployed still want to increase their income above the unemployment benefit, even if their participation tax rate (the fraction of their total earnings eaten by taxation, withdrawal of benefits, and work-related expenses) is absurdly high. Then the wage-fixing system enters the picture. The more employers are required to compensate workers for the loss of benefits, the less attractive it will be to hire workers at a given level of demand for the employers' products. The less the employers are required to compensate workers, the less demand there will be. Milton Friedman wasn't wrong when he infamously wrote, concerning unemployment benefits, that “We have increased the demand for unemployment, and the supply of unemployed has risen to meet that demand.” He just left out an essential part of the story: through the wage-fixing system and the retail market, the disincentive to work becomes a disincentive to hire.
The obvious remedy, wherever there is an unemployment benefit, is to stop clawing back the unemployment benefit in response to earned income—that is, to turn the unemployment benefit into a Basic Income, presumably with an activity test, and presumably with an assets test, but not with an income test.
To his credit, Friedman did mention something else that reduces the reward for working and increases the cost of hiring, namely taxation of labor, via income tax and/or payroll tax. This tendency to tax desirable activities, while undertaxing economic rents and negative externalities (including those caused by hoarding land), leads to a more respectable argument against a Basic Income: Shouldn't we clean up the way we raise revenue before we invent more ways to spend it?
Fair question. But if all options for mere tax reform are exhausted, it will still be true that hiring should be subsidized, not taxed. If such a subsidy is combined with a safety net for people without jobs, the resulting coverage will resemble that of a Basic Income (albeit not a universal one). In so far as this Basic Income operates as a subsidy for hiring, its purpose will be not to eliminate the need for jobs, but to satisfy that need.
And if we do all that and still don't get full employment, then we'll know how much we can blame AI.
