Tax Confusion All Around

Every four years, some marginal GOP candidate resurrects the flat tax idea. This year it is apparently Ted Cruz's turn.

The arguments for how stupid this would be (as well as needlessly redistributive from poor to rich) are well-known by now. What's more remarkable to me at this stage is just how many people are passionately stirred up by what is ultimately a meaningless gimmick.

Under a 20% "flat tax" scheme, someone making $10,000/year pays $2,000 tax while someone making $100,000/year pays $20,000 tax, i.e. $18,000 more tax. It's somehow OK for the richer person to pay $18,000 more in taxes than the poorer person just because he or she makes more money, but having that person pay $18,001, $19,000, or $25,000 more in tax would be an injustice. This makes no sense whatsoever.

Once you've given up on taxing people equal amounts (i.e. decided against using a head tax), you've already basically conceded the necessity of the progressive tax. In fact, a "flat tax" is considered a progressive tax by the normal tax progressivity metrics you see used in international comparisons. The "flat tax" and "progressive tax" debate is not about whether the rich should pay more tax, but about how much more, whether many thousands of dollars more or very many thousands of dollars more.

Progressivity Oversimplified

On the other end of the spectrum of tax confusion, you have this line that has mainly been pushed by the Matthews, Klein, and Yglesias gang about how low-inequality countries don't really rely on progressive taxes, but instead on "extremely progressive social spending." The problem with this line is that it obscures far more than it illuminates, and doesn't really illuminate much at all.

The source that this argument generally relies upon comes from the OECD tax progressivity metric and data. This metric works just like the income gini, but for taxes. So a society that had everyone pay literally the same amount (not rate, but amount) of tax would have a progressivity of 0 (perfectly equal). A society that had a flat tax (same rate for all) would have a progressivity gini that exactly matched its income gini. A society that had one person pay all of the tax would have a progressivity of 1 (perfectly unequal) and so on.

The basic trouble with this is that "taxes paid" is a mushy concept subject to easy gaming. This ends up presenting serious tax progressivity comparison problems.

To see why, it's helpful to have an example to manipulate, and luckily I have granular income distributive data from Finland that provides just that example. In Finland, the bottom decile has the following income situation:

The bottom decile had €3,083 in market income, €10,832 in transfer income, and €1,314 in income taxes, which combine to create a disposable income of €12,601. This only counts income information and so does not count VAT or other taxes. But that's sufficient for the basic point.

For the purposes of measuring progressivity, you'd say the bottom decile paid €1,314 in taxes, which is pretty high by international standards in the context of the income tax distribution. But noting this tells you nearly nothing of any importance.

Finland surely could, if it really wanted to, bring that income tax to €0 by structuring €1,314 of the transfer income paid to the bottom decile as "tax relief." This would change literally nothing, but Finland's income tax progressivity would go way up (especifically if it did this for the other deciles that are net transfer beneficiaries as well). In fact, not only would Finland's tax progressivity go up, but it's overall tax level would fall, and it's overall spending level would fall as well. Again, literally nothing of any significance would change if Finland did this, but its supposedly important fiscal indicators would be dramatically altered.

This is not a tedious point either. This is precisely how the US does things. The so-called 47% who don't pay income tax are largely beneficiaries of "tax relief" of this sort. The dependent exemption, the earned income tax credit, and the child tax credit (among others) "clear out" their tax liability, which has the effect of making tax progressivity high. But if the US implemented those same programs as transfers (child benefit and wage subsidy), everything would stay the same while tax progressivity plummeted and tax and spending levels shot up.

It's not just that Finland (and the other Nordics) fail to extinguish tax liabilities by cleverly disguising transfers as tax relief. It's also that they tax the transfer benefits themselves (at least many of them). So if you get, say, €5,000 in transfer benefits, and you have a 10% income tax rate, you would pay €500 of those benefits in taxes. This might seem like a pointless exercise (why not just reduce the transfer benefit to €4,500 rather than taxing it?), but coupled with a progressive income tax, it's actually a very efficient and simple way to indirectly means-test universal benefits. Under this scheme, those with higher incomes net less of their "universal" transfer benefits than those with lower incomes.

You could replicate this exact sort of means-testing by reducing benefit amounts according to income (i.e. using phaseouts). This would reduce "taxes paid" and also reduce "spending," but it would do nothing real. This is how the US tends to handle benefits: instead of taxing them as income in order to allow its progressive income tax to take a bite out of them, it tends to use phaseouts and direct means-testing. Consequently, it shows up as taxing the bottom less (and spending less), which deceptively bolsters its progressivity score relative to the more Finnish-like approach.

So when someone says the US has very progressive taxes and countries like Finland don't, they aren't actually illuminating much. Taxes can be used as a way of raising revenue or as a way of means-testing benefit distributions. Transfers can be used to increase transfer income or structured so as to reduce tax liability. Depending on which mix of those you choose (among other things), tax progressivity shows up very differently even if there is no actual meaningful difference involved. The US mix makes it very "progressive" while the Finnish mix makes it less so, but a lot of that is just accounting minutiae.

As a last note, it's worth pointing out that these countries don't necessarily have "extremely progressive social spending," if by that you mean that transfer incomes are heavily concentrated on the bottom. Here is transfer income in Finland by decile (horizontal axis is Euros):

The bottom decile actually received the least amount of transfer income of any decile. Transfers are high and fairly even across deciles.

The story changes however when we look at net transfers (transfer income minus income taxes):

If transfer incomes are pretty even, but net transfers look like this, one might actually conclude that it is primarily income tax progressivity that's cutting income inequality in Finland.