UBI vs Negative Income Tax

More nuanced than one might think.

Universal Basic Income has its roots in ideas around social security systems for the old and disabled as advocated for by Thomas Paine in the 1700s, the idea of a citizen’s dividend from Henry George in the 1800s, and the negative income tax that Milton Friedman advocated for in the mid 1900s. But UBI has become a popular concept in recent times, having been advocated for by a number of high profile politicians and business men as well as grass roots advocacy groups.

UBI is a simple concept: every person gets a basic income check in the mail or direct deposited into their bank account, regardless of their income or need or circumstances. This would have many benefits in comparison to our current welfare systems that require substantial administrative overhead in evaluating applications for welfare and monitoring welfare recipients, require the recipient to spend the money in specific ways, and have welfare cliffs that disincentivize recipients from increasing their earnings through work. By contrast, UBI would be money anyone could rely on as a baseline, without an application process, without any restrictions on how it could be spent, and without any major distortion to their incentives.

A negative income tax is also supplemental income. For this, a tax-exempt income level would be chosen, say $3000. Anyone who made above $3000 would pay income taxes calculated by using their income minus $3000 as the taxable income. Anyone who made exactly $3000 would pay not taxes at all (ie $3000 - $3000 = 0 tax liability). And anyone that made below $3000 would again have their tax liability equal to their income minus $3000, which would be a negative number. This negative tax liability would indicate a subsidy paid to them by the government. So for example, if someone made $2000 and the negative income tax rate was 50%, then their taxable income would be -$1000 ($2000 - $3000), and that would be multiplied by 50% to get their tax liability of -$500. They would then receive $500 in some form, usually envisioned as monthly or bi-monthly payments.

Its easy to imagine that UBI and negative income tax are basically the same thing with different clothes. After all with a UBI, people who earn income than some level will pay more additional tax than the UBI they receive, and people below that income level will pay less additional tax than the UBI they receive, so that in net, only the people below some income level will receive a net benefit, subsidized by the rest of the tax payers who see a net detriment. This on its surface looks basically identical to the outcome with negative income tax.

The difference is how much money flows through government and how much has to be paid via extra collected taxes. Less money flowing through government means less possibility for administrative waste and corrupt shinanigans. Even for this reason alone, in our current tax regime, I would advocate for a negative income tax over UBI. However, collection of taxes also results in deadweight losses, for things like income tax and sales tax losses are usually estimated at around 30% of the amount collected (tho this is very dependant on the tax rate, as each additional percentage point of tax increases the losses by more than the previous percentage point).

As Micah Erfran mentions here, the taper effect of a negative income tax has exactly the same deadweight loss effects as additional taxes. So deadweight losses occur not only as a result of collecting the tax, but also as a result of giving out the negative income tax subsidy. This, by contrast, does not happen for UBI.

We can calculate an estimate of what we should expect these deadweight losses might be. For a negative income tax with a baseline of $3000 and a negative income tax-rate of 50%, if we have a population of 100 million people, and 10% receive an average of $900/year from negative income tax, that means the other 90% of the population are collectively paying $9 billion/year in additional taxes, which has a deadweight loss of about 30% of that, or $2.7 billion. The deadweight loss of the negative income tax tapering will the same amount, so the total deadweight loss will be double $2.7 billion, or $5.4 billion.

What about for UBI? Well the UBI itself is given out regardless of anyone’s actions, so distributing the UBI does not have any deadweight loss. However, because much more money is distributed, much more money must be collected. The UBI equivalent to the above negative income tax would be a UBI of something like $1800/year/person (it must be about double the equivalent average negative income tax to account for the additional taxes most will pay). For the 100 million person population, that would amount to $180 billion/year, which would result in a total deadweight loss of $54 billion, which is 10 times higher than for a negative income tax. Note that its actually worse than that, since deadweight losses have a super-linear relationship with the amount collected.

So this is a clear win for the negative income tax. We’d see a very similar thing if the UBI or negative income tax was funded with sales tax, which has similar deadweight losses to income tax.

But things are very different if land value tax is used. Since land value taxes don’t result in deadweight losses (and in fact reduce deadweight losses), funding these things with LVT will not have as much loss. We will ignore the beneficial effects of LVT, since you can get those effects without spending them on something like UBI.

In any case, if the funding source is LVT, we eliminate the deadweight loss caused by the tax, and all that’s reminaing is the deadweight loss caused by negative income tax’s tapering. This results in negative income tax causing some deadweight loss, but UBI causing no deadweight loss at all. This means UBI pairs incredibly well with LVT and would be the ideal welfare situation that causes the least distortion of the market.