How to Tax and Spend 0% of GDP and 100% of GDP Without Affecting Anything

Let’s say I wanted to maintain the status quo distribution and all of the incentives related to it. But I also wanted to do it in a way that cut the tax level to 0% of GDP or 100% of GDP. How would I do that? It’s simple.

0% of GDP
Right now the government spends using United States Dollars. But you could change that. Instead of spending USD, it could finance all of its operations (including transfers and paying employees) by issuing transferable refundable tax credits. Instead of sending an elderly person a Social Security check equal to $12,000, for instance, the government would send out a transferable refundable tax credit equal to $12,000 (or possibly somewhat higher given the transaction costs the recipient will have to suffer). The recipient of that tax credit could then sell it to someone else (because it is transferable) and thus convert it into cash.

Come tax time, the refundable tax credits would be redeemed by whoever had them. The government would also levy taxes equal to its outstanding refundable tax credits. Thus, the refundable tax credits would entirely offset the revenue raised by the taxes and the tax level would be 0% of GDP.

100% of GDP
Right now the government permits some money to go out as market income untaxed. But you could change that. Instead of taxing a percentage of income, it could tax 100% of income and then send out welfare benefits to each individual in order to achieve it’s preferred distribution. Instead of taxing someone’s market income 30%, the government would tax it at 100% and then pay out a cash welfare benefit equal to 70% of that individual’s market income.

Because every market action leads to the same distributive result, this should not affect incentives at all. That is, a person who is taxed 30% right now should behave the same way if they are taxed 100% and then given a welfare payment equal to 70%. In both systems, their market action is generating the same outcome and so they should be willing to do it identically.

Thus, I have just demonstrated how I could keep the status quo distribution intact and all of the incentives contained within it while also switching us to a 0% of GDP tax level or a 100% of GDP tax level.

Nontax Revenue Accounts for 92% of Growth in “Taxpayer Money” Since 1970

The last few months has been an banner one for the incoherent but totally hegemonic concept of “taxpayer money.” I’ve been meaning to talk about this funny concept for some time now and so I figure I should take advantage of these events to do that here.

Bobby Jindal

In Louisiana, Bobby Jindal wanted to increase revenue without raising taxes. This is because the state desperately needs more revenue, but Jindal’s political commitments make it impossible for him to raise taxes.

So here is how Jindal sensibly handled this conundrum:

  1. Charge public college students a new fee. ($350 million of nontax revenue).
  2. Provide public college students a new tax credit. ($350 million tax reduction).
  3. Increase taxes elsewhere. ($350 million tax increase).

So you see (2) and (3) offset each other, meaning that there is no net increase in taxes. Thus, under normal revenue accounting, Jindal raises $350 million solely with nontax revenue, i.e. the fee in (1). The new revenue increase is, I guess, not “taxpayer money.”

The consensus liberal take seems to be that this is a ridiculous gimmick, as if there is deep down some sort of obvious taxing going on here. But these objections operate of the assumption that taxation is a clear concept when it very much isn’t. Jindal’s move is not obscuring a tax; rather, it is taking advantage of the fact that “tax” is not a coherent thing and is interchangeable with almost any other distributive move.

For example, suppose the US government wanted to balance a revenue shortfall by reducing the disposable income of old-age pensioners. It could do so by cutting Social Security benefits, which is scored as a spending reduction that has nothing to do with taxes. Or it could do so by increasing taxes on Social Security benefits, which would be a tax increase. These are distributively identical actions, but one is a tax and the other isn’t. If you think Jindal’s student fee scheme is really deep down a tax, then mustn’t you also think Social Security benefit cuts are deep down a tax? Wouldn’t such a cut be an assault on “taxpayers”?

Jordan Weissman

If you think I am nitpicking here, consider another recent and interesting use of “taxpayer,” this time from Slate’s Jordan Weissman:

Lee Siegel is an award-winning critic and an unrepentant leech. After pursuing not one, not two, but three degrees from an Ivy League university, he chose to default on his student loans at taxpayer expense, because he felt that paying them back would have hampered his ambitions of becoming a writer.

Yesterday Weissman did it again:

Earlier this month, Bloomberg turned up Laura Strong, for instance, a part-time therapist and yoga instructor who took out $245,000 for a Ph.D. in psychology, toward which she is contributing $100 dollars a month under Pay as You Earn. Two decades from now, taxpayers will likely have to eat whatever she doesn’t pay off.

What’s interesting here is that, at least in most cases, student loan defaults are fully accounted for in the interest rate charged by the Department of Education. That is, the DOE sets the interest rate on student loans at such a level that the student loan system is “self-funding” even with defaults here and there. The incidence of the cost of defaults is borne, not by the “taxpayer” in some quasi-normal sense of the word, but rather by other student debtors who pay somewhat higher interest.

Of course, all government revenue is fungible, and so in some sense the student loan interest bucket is not meaningfully distinguishable from the official tax revenue buckets, and so maybe there is some sort of “taxpayer expense” going on here. Though, if we use fungibility in this way to access the word “taxpayer” in situations like this, then that also means money spent on roads is coming at “student debtor expense,” for student debtors also contribute to government revenue.

More Generally

Aside from these two examples, what’s bugged me generally is the way in which “taxpayer money” is used totally inaccurately as a substitute for “public money.” For starters, “taxpayer money” is an obvious oxymoron because any money that is taxed is, by definition not the money of the person we call the “taxpayer.” The laws determine what money belongs to who and tax laws clearly demarcate that the money does not belong to the taxed.

Aside from that point though, the usage is inaccurate because not all government revenue comes from taxes. A good deal of it comes from fines, licenses, fees, tolls, lotteries etc. And this is increasingly true. According to the OECD, 20% of total US government (local, state, and federal) revenue in 2013 was nontax revenue. This means 1 in 5 public dollars do not come from what gets categorized in national accounting as “taxes.”

Moreover, the share of government revenue coming from nontax sources has been on the rise:

Between 1970 and 2013, US government revenue increased from 29.7% of GDP to 33.2%. Nontax revenue growth accounts for nearly all of that revenue growth (92%). Or to put it in the confusing parlance of pundits: nontaxes account for almost all of the growth in taxpayer money. Heh.


When talking about “taxes” versus “fees” versus “cuts,” it can become very difficult practically to avoid picking one and going with it for descriptive purposes. This is the hazard of the way we’ve decided to carve up specific distributive categories. Some halfway accurate usage is thus understandable.

What’s not understandable is why “taxpayer money” (or similar) gets so much use, despite the fact that it is an objectively inaccurate descriptor. That is to say, the phrase literally is lying about the origin or whatever of the money being talked about, 1/5th of which is from nontax sources. “Taxpayer money” serves certain ideological ends well, which explains some of its usage, but it serves truthful description purposes extremely poorly. Meanwhile “public money” is a 100% accurate description.

Low Hanging Tax Code Simplification

Like everyone else in policy circles, I think the tax code as it is currently constructed is a complicated joke. There is a lot that can be said for how to fix it, some of which I don’t know enough to comment about. One thing I do know enough to comment about is how to fix the child-related parts of the tax code. Here is how I would do that (figures from 2013).

  1. Eliminate the Child Tax Credit (save $54.15 billion).
  2. Eliminate the Dependent Exemption (save $39.61 billion).
  3. Eliminate the Head of Household Filing Status (save $5.92 billion).
  4. Eliminate the Child and Dependent Care Tax Credit (save $3.51 billion).
  5. Reform the Earned Income Tax Credit so that it pays the same benefit to all qualifying workers regardless of how many children they care for (savings unknown).
  6. Create a Child Benefit (aka Child Allowance) program that pays out a universal monthly benefit per child to every family.

The basic thrust of this reform is to convert all of the code-complicating child benefits into a universal monthly benefit. In addition to the straight deletions from the code, this also reforms the EITC in a way that makes it way more understandable. Right now, the EITC benefit looks like this:

You got that? You understand also how that interacts with other tax code provisions? Probably not. Now imagine how well something that looks like that can incentivize people to work, which is the purported goal of the program. The Reformed EITC would have one line that goes up in a trapezoid shape, instead of this mess. Further, by detaching the EITC from child benefits, it makes it easier to eventually transform it into a monthly benefit as well, instead of the yearly windfall that it currently is.

This reform could be done in a “budget neutral” fashion. You’d just set the Child Benefit level to be equal overall to the money saved from the tax code changes. It can also be done, as I’d prefer, in a way that “spends” more money overall. You’d just set the Child Benefit level higher than the money saved from the tax code changes.