Why Consumption Taxes are Fine

In my last post, I said that it would be good if the US imposed a consumption tax such as the value-added tax (VAT). Critics generally say that these kinds of taxes are bad because they are “regressive.” While it is true that they are regressive under the way that word is generally used, that entire way of thinking about taxes is confused and muddled (as I’ve discussed previously in the case of the Nordics).

The standard response to those who raise the regressivity objection is to say that it just depends on how the proceeds of the consumption tax are spent. This is true, but only because it is true of all taxes. Even progressive taxes are only good if the proceeds are spent well. You could spend them on bad things and even in ways that make inequality worse.

The bigger problem with the regressivity objection, in my view, is that dividing taxes paid by income seems to obscure the more important point. What really matters in all of this is how many dollars you are scraping from poor, middle class, and rich people. Consumption taxes scrape more dollars from people who consume more and it is the rich who consume more.

According to the 2015 Consumer Expenditure Survey, the richest income quintile consumes an average of $110,424 while the poorest income quintile consumes an average of just $24,355.

Screen Shot 2017-04-05 at 10.46.56 PM

A 10 percent consumption tax would thus draw $2,435 from the poorest quintile and $11,042 from the richest quintile. Which is to say that such a tax draws 4.5 times as much money from the rich as the poor.

Whether the money drawn from the consumption tax ultimately reduces inequality does depend on how it is spent, but it is not like it needs to be spent in an especially “progressive” way, i.e. in a way that is heavily targeted towards the poor. Even if you spent this money in a way that benefited all quintiles equally, you’d still see a pretty significant net swing.

Screen Shot 2017-04-05 at 11.00.48 PM

Of course, this graph features a rather simplistic analysis as it assumes consumption does not change at all in response to the tax and benefit reforms. But even if the precise figures would be somewhat different in a real life implementation, the basic pattern of the graph above would still hold.

None of this is to say that a consumption tax is better than other taxes. The US has plenty of room to increase its tax level by upping income taxes, and so that’s a natural place to look. But to say consumption taxes are bad generally is pretty clearly mistaken.

  • Dave Rohlfing

    I agree with the point that it is important how taxes are spent. The problem that might exist (without exemptions/credits, etc.) is the taxation of basic goods. If you tax everything at the same rate, a low income consumer and a high income consumer pay the same amount of tax on a gallon of milk, a sweater, or a car. This is where the unfairness (and regressiveness) comes in. It is true that the high income person will probably pay a greater amount of tax, but will probably pay a smaller percentage of their income in taxes than the low income person, who is spending all of their money on taxable stuff.

    One way this gets dealt with is to have certain goods exempted from the tax – food, clothes, etc. Another is to allow a deduction on income taxes for other taxes paid, such as sales (or other consumption) taxes.

  • While the notion of “regressive” vs. “progressive” taxation might be muddled indeed when it comes to funding welfare (or other direct transfer of wealths), when it comes to funding public projects in which it is more difficult to assess the value per person, such militaries, police departments, fire departments, basic infrastructure, etc., it makes sense to think about whether individuals are “paying their fair share,” in which case the progressive and regressive labels seem appropriate.

  • oie

    (Reposting this as it was meant as a reply to Sam Hopkins.)
    There is nothing inappropriate with the labels per se. The issue is if
    the regressivity of some tax is an objection against that tax. I take
    Matt’s post to rebut that objection. I don’t see how the cases you
    mention changes that. Suppose some regressive VAT funds are spent
    on fire departments and basic infrastructure that serve the needs of
    the general public. What is the problem? Isn’t the important thing the
    overall effect of the whole system of taxes, transfers and public
    services?

  • The point is with welfare spending it is easy to argue, as Matt has, that poor people pay $X in tax, and get > $X out from almost any conceivable tax system. But imagine instead a scenario far different from welfare: a bunch of aliens are about to attack earth, and we need to raise a fixed amount, 1 trillion dollars or something, to fend them off. Everyone clearly “benefits” from this government action, but it’s very hard to say that any given person gets $X of benefit from it. If the question is, how do we fairly distribute the burden of this alien defense tax, I think poor people would be right in claiming that a consumption tax would be very regressive, i.e., very unfair.

  • oie

    The alien case is weird but fun so lets roll with it.

    Are you assuming that the background set of taxes and policies are constant? Then the only choice seem to be between getting the extra 1 trillion through (a) a regressive VAT or (b) through non-regressive, perhaps progressive, taxation. In that case I agree it is unfair to pic alternative (a). But it is unfair because of the outcome that the tax plus the background set produces. Both alternatives fend off the attack but (a) generates a total pattern of distribution that is more unequal than that from (b). That is the problem, not the regressivity per se.

    To argue that point I make use of a version of the case where the only two action alternatives are (c) 1 trillion dollars through progressive taxation or (d) 1 trillion dollars through regressive VAT plus a compensatory shift in other taxes in policies that in total yield the exact same outcome in terms of resource inequality as alternative (c). In that modified case I see no reason to pick (c) over (d) if everything else is really equal. Because it is the total outcome that matters and the total outcome is the same. To support that conclusion further imagine a modification where (c) has some other drawback compared to (d). For example with (c) one extra person will lose a leg in the alien war. Then I think it is clear that we should do (d). It seems wrong to let someone lose their leg over inconsequential regressivity.

    So far principled argument. In practical political activism it might still sometimes make sense to campaign in terms of problems with specific taxes, for example shape a campaign around increasing progressivity in some tax. Perhaps drawing attention to the immediate effects of regressivity can be useful there in some cases. But I suspect the best long term strategy there too is to draw attention to the big picture. That is needed anyway to counter the entrenched libertarian mythology about natural property rights.

  • oie

    There is nothing inappropriate with the labels per se. The issue is if the regressivity of some tax is an objection against that tax. I take Matt’s post to rebut that objection. I don’t see how the cases you mention really changes that. Suppose some regressive VAT funds are spent on fire departments and basic infrastructure that serve the needs of the general public. What is the problem? Isn’t the important thing the overall effect of the whole system of taxes, transfers and public services?

  • Yes, but I don’t see much Venn overlap between the people wanting to do spending in a progressive way and the people slinging Newspeak concepts like the so-called “fair” tax.

  • Sales tax is inherently regressive compared to income tax, and I find the “consumption tax” framing more than a little troubling as it reeks of producerism, which has become something of a disease of late. Nevertheless I once heard of an idea (from Jerry Springer, of all people) that is simultaneously in some sense a “consumption tax” and also in some sense non-regressive. You apply a progressive (hopefully steeply progressive) rate bracket to “consumption,” understood to be income minus deposits (and purchase of investments) plus withdrawals (and sale of investments). I heard this on the radio back when Air America was a thing. Seems Wayback captured it from a blog Springer had back then:

    https://web-beta.archive.org/web/20050312042945/http://www.springerontheradio.com:80/story/2005/3/7/11201/46353

  • Eric Stephens

    “The bigger problem…is that dividing taxes paid by income seems to obscure the more important point. What really matters in all of this is how many dollars you are scraping from poor, middle class, and rich people.”

    This strikes me as an odd justification for a consumption tax. Is it not the case that a flat-rate (or even severely regressive) income tax would scrape more money (in absolute terms) from rich people than poor people?