Consumption taxes are not a major topic in US politics because they are rarely proposed, especially on the federal level. When consumption taxes are brought up, the discussion tends to devolve into a debate about whether they are regressive or not. As with most progressivity or regressivity debates, this one tends to be muddled and confused because we don’t actually use those two words consistently across different policy topics.

In the context of tax policy, the way you determine whether a tax is progressive or regressive is by adding up how much tax each person is being charged and then dividing that amount by their income. If the resulting percentages go up as incomes go up, then that is considered a progressive tax. If the percentages go down as incomes go up, then that is considered a regressive tax. If the percentages are the same across the income distribution, then that is considered a flat tax.

In the context of spending policy, the way you determine whether a spending program is progressive or regressive is by adding up how much money each person gets from the program. You do not divide that amount by income. Instead, you just look directly at the dollar amounts. If those amounts go up as incomes go up, then that is considered a regressive program. If those amounts go down as incomes go up, then that is considered a progressive program. If the amounts are the same across the income distribution, then that is considered a flat benefit.

The fact that the same words have different meanings in different contexts is an unfortunate feature of the way language works, but it is especially unfortunate in this context because “taxes” and “spending” are fungible categories and it really would be better if we could talk about them with a common set of word meanings.

When it comes to consumption taxes, there are two facts that are pertinent to understanding whether they are progressive or regressive:

1. Richer people consume more than poorer people.
2. Poorer people spend a larger share of their income on consumption than richer people.

Because richer people consume more than poorer people, taxing consumption results in richer people paying more consumption tax than poorer people pay. But because poorer people spend a larger share of their income on consumption than richer people, taxing consumption results in poorer people paying a higher percentage of their income toward consumption tax than richer people pay.

So if you divide the dollar amounts involved by income, as is typical in tax policy, consumption taxes are regressive. If you do not divide the dollar amounts involved by income, as is typical in spending policy, consumption taxes are progressive. Because a consumption tax is a tax, under common usage, it is correctly described as “regressive.”

But there is actually an exception to the way we use these words in the context of tax policy that is worth dwelling upon a bit. When we are talking about raising taxes, we use “regressive” and “progressive” in the tax policy way already discussed. But when we are talking about cutting taxes, we actually use the words the same way that we use them when talking about spending policies. This is presumably because cutting taxes kind of feels like spending money because, relative to the baseline where we don’t cut taxes, it kind of is.

This came up recently in a Finnish debate about the application of the country’s 24 percent value-added tax (VAT) to electricity. After Russia invaded Ukraine, electricity prices across Europe skyrocketed, including in Finland. This naturally led lawmakers to consider various ways of providing consumers relief from the energy price shock. One such proposal was to temporarily reduce the VAT on electricity from 24 percent to 10 percent.

Is reducing the VAT on electricity by 14 percentage points progressive or regressive? As always, the question turns on whether you divide the tax savings by each person’s income or not.

If you look only at electricity consumption, you find that rich households in Finland consume five times as much electricity as poor households do. But, if you divide that electricity consumption by income, you find that poor people spend about four times as much on electricity as rich households do as a share of their respective incomes.

If we apply the spending-policy meanings of the words, we would say that cutting the VAT on electricity is very regressive: it gives five euros to the rich for every one euro it gives to the poor. But if we apply the tax-policy meanings of the words, we would say that it is very progressive: the tax savings for the poor are four times what they are for the rich after dividing by income.

Because this is a tax cut and not a tax increase, and is specifically a tax cut aimed at achieving a spending-program type goal (providing electricity subsidies), it appears to have been analyzed by at least some Finnish commentators using the spending-policy word meanings and was thus criticized as horribly regressive.

When the words are used this way, and they often are, you can wind up in a somewhat hilarious linguistic quagmire where it is “regressive” to implement a consumption tax and “regressive” to repeal it. Thus, in Finland, the 24 percent VAT is “regressive” because the poor spend more of their income on consumption than the rich. But also repealing the 24 percent VAT is “regressive” because the rich receive far more euros from such a repeal than the poor do.

If we put the imprecision of language aside — and stop putting so much weight on what label can be applied to what policy — the actual distributive stakes of a consumption tax are fairly straightforward. Because consumption taxes take more dollars from the rich than the poor, they are generally good, especially since the public programs they help finance tend to provide at least as much benefit to the poor as the rich. But insofar as other kinds of taxes — like income and wealth taxes — take even more dollars from the rich than the poor, those taxes are better than consumption taxes, holding all else equal.

Of course, ultimately, it is not really possible to analyze one piece of an overall distributive system and decide whether it is itself good or bad. What matters is whether the system as a whole achieves your overall distributive goals. Put differently: distributive justice can only really be coherently evaluated at the level of the overall system, not at the level of each particular institution in that system.

So try not to get yourself too worked up about each and every fiscal policy measure and whether, and under what meanings, that measure is progressive or regressive. Such debates are equivocal at best, irrelevant at worst.