Many years ago, I learned that the Faroe Islands has a peculiar process for compensating workers. There, employers pay each worker’s entire paycheck to the tax authority, which removes any taxes owed and then remits the remainder to each worker’s linked bank account. As part of this process, the tax authority also rolls in any welfare payments an individual is owed when making its periodic payments.

I found this system fascinating for many reasons:

1. It has clear benefits for employers as they no longer have to manage tax withholdings or hire a third-party firm to do it for them.
2. It has clear benefits for individuals as they never overpay or underpay their taxes and therefore don’t have to deal with any end-of-year tax reconciliation.
3. It has clear benefits for managing the economy because it generates nearly real-time, whole-economy data about employment, wages, and incomes.
4. It has clear benefits for welfare administration, as it automatically collects the information that goes into the calculation of certain benefits like unemployment insurance. It also makes it trivial for the welfare state to reach every person in the country, such as in the event of an acute recession where stimulus may be necessary.

Beyond these practical advantages, the idea of a central income distribution institution (CIDI) — i.e. a government entity that all income payments are routed through, even factor income payments like wages, dividends, and interest — is useful for thinking through certain intractable philosophical, accounting, and conceptual debates that frequently pop up in the economic policy discourse.

The reason a CIDI is useful for this purpose is that it makes it easy to cut through a lot of the mental fuzziness that drives so many of those debates. This is probably most true in the debate about means-tested welfare benefits versus universal welfare benefits.

There is a general consensus in the policy world that means-tested benefits cost less than universal benefits. This is demonstrably false and is based on accounting conventions that consider the effective marginal tax rates (EMTRs) imposed by universal programs to be tax-increasers while considering the EMTRs imposed by means-tested programs to be spending-reducers. When you compare universal programs and means-tested programs that are EMTR-equivalent, while also looking through the misleading accounting conventions used to score them, you see that they differ only in that administering a means-tested program is harder, costlier, and more error-prone.

No matter how many times you try to say this, many people just cannot get their head around it and are naturally skeptical that virtually every person in the budget policy world, including the budget scorekeepers at the Congressional Budget Office, are making such a simple mistake. But if incomes were administered through a CIDI in the way they mostly are in the Faroe Islands, the phaseouts of means-tested programs and the taxes of universal programs would no longer be merely economically equivalent. They would be administratively indistinguishable.

To see why this is the case, recall that each person’s disposable income is determined by the following formula:

Disposable Income = Factor Income + Transfer Income – Tax

Factor income refers to payments made to the factors of production, which is labor income for work (e.g. wages) and capital income for capital (e.g. dividends). Transfer income refers to payments made pursuant to certain welfare programs, like old-age, disability, unemployment, and child benefits. Tax refers to tax.

The way a means-tested program works is by reducing each person’s transfer income according to how much factor income they have. The way a universal program works is by increasing each person’s tax according to how much factor income they have.

These net out to the same thing — a $100 reduction in transfer income has the same impact on a person’s disposable income as a $100 increase in tax — but in the absence of a CIDI, they look and (apparently) feel very different. Specifically, in the absence of a CIDI, a universal program requires the depositing of transfer income into the bank accounts of rich people and the payment of taxes by those same people, while a means-tested program avoids both things. Trying to avoid those two things ends up being more complex and thus more costly and error-prone, but it confuses people into thinking that it lowers taxes and spending while also sticking it to the rich.

But once we bring in a CIDI, these illusions about means-tested programs fall apart. In a system with a CIDI, each person’s factor income and transfer income is combined together and taxes are deducted. The net of all of that is then paid out as a single transaction to each individual every week. Thus, in this world, neither employers nor individuals pay a separate tax in the form of a specific tax transaction, as all factor income is routed through the CIDI. Also in this world, there are no separate transfer payment transactions: the CIDI rolls transfer payments into its single disposable income transaction.

These unique characteristics of a CIDI system make it so that reducing a person’s transfer income based on their factor income (means-tested phaseouts) is exactly the same thing as increasing a person’s tax based on their factor income (universal taxes). In a dialectical masterstroke, the CIDI resolves the contradiction by making the two kinds of program designs completely identical. In this world, people fond of means-testing could happily conceptualize the degree to which increases in disposable income are made to lag increases in factor income as a phaseout, and people fond of universalism could happily conceptualize the same thing as a tax.

Indeed, if you wanted to, you could even maintain two separate versions of the computer program used to calculate each person’s disposable income payment: one where taxes rise when factor incomes rise and the other where transfer incomes fall when factor incomes rise. These programs would feature the exact same EMTRs and generate the exact same output, making it impossible for anyone to say whether a phaseout or tax has really been used.

Beyond this conceptual fun, the CIDI approach to income administration would also practically resolve all of the problems with means-tested programs, which is what actually makes them so bad.

Right now, we don’t have a central income distribution institution in the United States, and it seems doubtful we’ll get one any time soon. Operating within that constraint, universal programs that rely on taxes rather than phaseouts to achieve desired EMTRs are clearly superior to means-tested programs, and so we should continue to advocate that welfare policy move in that direction.