A totally off-the-news post about taxes

How I think it should work in theory

Every once in a while someone asks me what I think about optimal tax theory or how I think an idealized tax system would be designed.

I personally do not love this subject because I don’t think abstract considerations really tell you that much about tax policy. I also think that taxes are kind of overrated as a policy issue. That’s not to say taxes aren’t important. But the issue garners a ton of attention from the political system for the boring, obvious reason that an incredible amount of short-term money moves around based on tax changes. Looking at the broad sweep of economic development, though, it’s both true that redistribution has done much less than economic growth to ameliorate poverty and that countries never really rocket to prosperity because they enacted right-wing tax policy.

But my grandfather was an economist, and one of the things he told me is that when you see a business practice that doesn’t seem to make sense, it’s usually either fraud or motivated by tax avoidance.

And I do think that once your eyes are open, you see a fair amount of static inefficiency stemming from tax avoidance. Business people accept a lot of de facto compensation in the form of excessive travel expenses rather than taxable salary, for example. This ultimately makes us poorer than we’d be in an optimal tax situation. More consequentially, providing health care to people largely through a murky tax subsidy leads to overconsumption of health resources by some, even while many others struggle with lack of access.

We also miss a lot of opportunities to make tax avoidance work for us by taxing things that are harmful. Alcohol taxes should be considerably higher. Marijuana should be legal in more places, but it’s undertaxed in the places where it is legal. Greenhouse gas emissions should be taxed. Sweeteners should be taxed. The point of this kind of thing isn’t to make it impossible for anyone to afford to buy Twizzlers, a can of beer, or some charcoal for the grill — that stuff would become considerably more expensive under my dream tax code, but other taxes would be lower, so you’d have the cash to buy the beer if you want it. But taxes promote avoidant behavior at the margin, so relative to today’s average consumption bundle, we’d see a shift to less consumption of intoxicants, sweeteners, and highly polluting activity, and probably more spending on various kinds of services.

Beyond that, the broad goal is to tax consumption rather than investment, but then find ways to build progressivity into it.

The case for taxing consumption

Saying that we should tax consumption rather than savings sounds good to a lot of people. But saying we should tax labor income at a higher rate than investment income sounds bad to a lot of people — you get Warren Buffet saying he shouldn’t be paying a lower tax rate than his secretary. And a lot of people are very angry that hedge fund managers get to classify a portion of their compensation as capital gains and avail themselves of a low tax rate.

So without getting too deep into the economic modeling, it’s worth starting with an intuitive case against capital taxation.

Jim and Bob are both dentists living in Kerr County, Texas. They both have two kids who attend local public schools and then the University of Texas. They both own homes that are pretty big but not especially lavish, and they drive normal rich guy trucks but nothing wild. They also don’t get a lot of time off of work, so their vacation time is precious. Jim and his family react to that by taking very lavish, expensive vacations every year. Bob tends to spend his vacations on low-key trips visiting family in Dallas and Houston. The upshot is that Bob has a higher savings rate than Jim. And because both men are well-paid dentists, even at Jim’s relatively low savings rate, he’s maxing out his retirement accounts. Over the years, Bob accumulates a bunch of stock which pays dividends that he reinvests. So eventually Bob ends up with a substantially higher income than Jim thanks to the dividends. He pays more in taxes than Jim does (because those dividends are taxed), but he pays a lower tax rate than Jim (because the dividends are taxed at a preferential rate).

• On the one hand, this is regressive — we are taxing the richer of the two rich dentists at a lower rate than the less-rich one.

• On the other hand, Bob is paying more taxes than Jim purely because he was more frugal over the course of his life. He could have sent his kids to private school or bought weird exotic cars or done any of a dozen other things with the money he saved by not taking fancy vacations.

You can play Intuition War all day long as to which is fairer. But the one really important issue here is whether, at the margin, it is good to encourage more Bob-like behavior or more Jim-like behavior. And I think the answer is Bob-like behavior. A society with a higher savings rate is one in which it is cheaper to obtain financial capital in order to make investments in tangible capital. And a society with a larger stock of tangible capital will be a richer society over the long run.

I could give you a bunch of reasons why I don’t think this is particularly important in practice for the contemporary United States, but my most basic one is just “look around.” Over the past quarter-century interest rates have consistently been very low, and obtaining capital to make an investment hasn’t been a major challenge. So I think it makes a lot more sense to focus on direct barriers to investment — zoning, permitting, and NIMBY stuff in general — rather than on providing tax windfalls to rich people. But in principle, I do think the case for taxing consumption rather than investment holds. With a big exception.

We should tax land value

About one-quarter of Americans’ assets are real estate, and a majority of the value of that real estate is the value of land rather than the value of the structures that are on it.

And while talking about this a lot is associated with being a crank on the internet, I think it’s true that Henry George was right all along and there ought to be heavy taxes levied on land value. George and other classical economists thought of the world as having three major factors of production — land, labor, and capital — implicitly imagining the working farm as the quintessential business.

Neoclassical economics tends to talk in terms of a simple dichotomy of labor and capital. Here they work implicitly with the idea of the industrial age factory, where location is irrelevant and you just put it somewhere that the land is cheap. Neoclassical income is either consumed or saved, and whatever is saved (by definition) is invested in growing the nation’s capital stock. But in reality, a non-trivial fraction of savings goes to bidding up the price of land. In today’s terms, that doesn’t mean bidding up the price of farmland. It means bidding up the price of beach and mountainside locations for vacation houses. It means bidding up the prices of single-family lots in the desirable suburbs with the “good schools.” It means bidding up the prices of townhouses in the cutest historic urban neighborhoods.

None of this bidding-up of prices adds to the capital stock in a way that complements labor, so tax code tweaks designed to promote saving and investment will fail to the extent that the money is plowed into land instead.

So while for reasons of measurement error and practicality I don’t think we should go for George’s full 100 percent land value tax, the taxation here ought to be heavy in pursuit of the dual goals of raising revenue and keeping land cheap.

How to tax consumption

My preferred framework for a progressive consumption tax is the one that Cornell University economist Robert Frank has outlined. It would work more or less exactly the same as today’s income tax, but with two major changes:

• There would be no differentiation based on the source of income — whether it’s wages, capital gains, dividends, interest, rent, whatever — it all goes into the “income” bucket.

• Instead of a little form where you list your 401(k) contributions and deduct them from your income, you’d deduct all contributions to any kind of savings vehicle (bank account, brokerage, whatever).

• This is separate from optimal tax theory, but the way it ought to work is the IRS sends you a pre-filled form saying what they think you owe based on what’s been reported to them by employers and financial institutions. In 80-90 percent of cases, you could just check “yes, okay, that’s right” rather than needing to do your taxes.

Here’s how Frank describes it:

Families would report their incomes and their annual savings to the IRS, just as many now do with 401(k) and other similar retirement savings accounts. Their taxable consumption would then be calculated as income minus savings minus a large standard deduction–say, $30,000 for a family of four. For example, a family that earned $50,000 and saved $5,000 during a given tax year would have taxable consumption of $50,000–$5,000–$30,000, or $15,000 total. Tax rates on taxable consumption would start off low–say, 10 percent for the first $30,000 of taxable consumption. Under the consumption tax, this family would owe $1,500, about half of what it would pay under the current income tax.

There are smart people who exploit accounting identities to come up with other ways to administer a progressive consumption tax, of which the one I’m most familiar with is the X-Tax. My concern with these ideas is that while it’s true in an economics textbook that a tax on wages is equivalent to a tax on consumption, in practice I think this would create large undesirable incentives for highly paid professionals to organize their lives around becoming the owners of independent small businesses rather than salaried employees of big companies. Right now in the national accounts, Kate and I would show up as earning capital income in virtue of our ownership of Slow Boring. But in the real world, the amount of capital associated with Slow Boring is minimal and our income derives from the work we do writing and editing and managing the site.

So I think you need to try to tax people based on actual consumption patterns rather than sources of income.

Key to the progressive consumption tax is that the marginal rates would get really high. Unlike with a conventional income tax, there’s no reason the rates can’t go above 100 percent. I think it’s also important to classify interpersonal financial transfers — gifts, etc. — as consumption for tax purposes because otherwise it becomes a huge loophole to exploit.

Other taxes

For regulatory purposes, the U.S. government designates an official “social cost of carbon” (SCC) which is supposed to represent the external harms caused by CO2 emissions. The Trump administration set this to a laughable $1 per ton while Biden has pushed it back up to $51. Some environmental economists think the right answer is $258.

That’s an annoyingly wide dispersion of estimates, but in principle we shouldn’t just be using this number for arcane regulatory math; we should actually be charging it as a fee. If we did, it would be a lot easier to do the right thing and relax about regulatory curbs on oil and gas drilling because we would be formally taxing the externality. Since the science and economics of the appropriate social cost shift around, I would suggest formally linking carbon tax revenue to the standard deduction threshold for the progressive consumption tax. In other words, updating the science to tweak the SCC down should automatically trigger a fiscally equivalent reduction in the standard deduction. And the reverse should trigger an increase in the standard deduction.

You’re obviously never going to have a world where something like SCC is decided as a purely technical matter with no politics involved. But you can at least try to set the parameters so that the incentives mostly point in the direction of trying to get the technical question right rather than back-filling the technical analysis to reach the desired conclusion.

There ought to be higher alcohol taxes. And I would make this a two-stage tax with a fee levied per milliliter of alcohol and also an ad valorem tax on the price of the bottle. The idea, in other words, is to both raise the floor price of booze but also to especially soak rich people with a taste for fancy drinks.