Repairing the tax code after Trump

A better, more progressive code is within reach, but Democrats should avoid unwise pledges

Barack Obama and Joe Biden both ran for president on a pledge not to raise taxes on anyone making under a certain threshold — $250,000 a year for Obama and $450,000 a year for Biden. This is a reasonable pragmatic response to the reality that Americans don’t want to pay higher taxes but Democrats do want to increase spending.

Wonks hate these pledges, though, because sticking to them can force politicians into some odd contortions and keep good ideas off the table.

Obama also found himself arguably and repeatedly in violation of his own pledge, and on the core question of the expiring Bush tax cuts, he had to settle for the $450,000 threshold rather than the $250,000 one that he ran on. I know this frustrated the White House, not only because they wanted more revenue, but because raising the threshold from $250,000 to $450,000 meant that a person with a seven-figure income derived much more benefit from the policy than someone earning $300,000.

This is what happens when you base all your tax arguments on arbitrary thresholds.

That said, you can’t take the politics out of politics, and the pledge worked well enough for its intended purposes. But precisely because it worked pretty well, there’s a strong chance that Democrats running in 2028 will offer similar pledges. I think that would be a big mistake for three reasons:

1. The actual fiscal situation of the United States has gotten much worse since Obama ran and requires a different way of thinking about fiscal policy.

2. Trump’s gimmicks — like a tax deduction for car loan interest payments or for overtime pay — are eroding the core tax base.

3. Trump’s tariffs are creating an incredibly regressive and economically destructive form of tax revenue that Democrats should replace.

The upshot is that repeating Obama/Biden-style pledges will be more damaging under the present circumstances, and also that the tariffs create an opportunity to be more creative with revenue.

I’m hoping that in the 2028 cycle, Democrats abandon their habit of releasing super-specific policy blueprints altogether, but especially on taxes. They should instead acknowledge that the details of legislation are determined by Congress and no one on the campaign trail actually knows who will even be in Congress after the election. What voters want from the president isn’t “plans,” it’s clear goals. And the goal on taxes should be to bring down the cost of living by narrowing the budget deficit while making the tax code more progressive — not that literally zero non-rich people will end up paying more.

The tariff dilemma

Trump’s trade policies seem to be unpopular with an electorate that desperately wants politicians to focus on bringing down the cost of living.

That said, I am astounded, given everything that we know about the politics of taxation, that the backlash to an enormous, broad-based tax increase hasn’t been much larger. To some extent, Trump benefits from the fact that people associate Republicans with tax cuts, so the idea that he is, on net, raising taxes on lower income people doesn’t really scan. But it’s also true that people seem to put a lot of stock in legal tax incidence. When you ring items up at the cash register, there’s no “tariff” line on the receipt.

Still, the fact is that tariffs are a huge, regressive tax increase.

On the campaign trail, Kamala Harris characterized tariffs as a kind of national sales tax. And while I appreciate Harris’s effort to draw an analogy with some explanatory power, it’s important to understand that a tariff is not a consumption tax. Here are a few of the many differences:

1. A tariff falls on capital goods, like machines you might buy to put in your factory.

2. A tariff also falls on production inputs, the parts and raw materials that finished products are made out of.

3. A tariff falls only on goods, not services.

4. A tariff falls only on imported goods, a narrower tax base than that of a goods tax.

The first two points mean that high tariffs are taxes on investing in productive activity in the United States. Point three means that tariffs are much more regressive than something like a European VAT (ultimately a big European-style welfare state requires something like this) or Canada’s Goods and Services Tax. That’s because while rich people certainly buy more goods than poor people, the richer you get, the more your consumption bundle tilts toward services. A tariff on imported tomatoes, for example, raises the price of home cooked meals or cheap fast food much more than it raises the price of fine dining, because the further up the fanciness spectrum you go, the less direct food costs determine the price of the meal. If you buy a soccer ball for your kid, the price is going up thanks to tariffs. If buy tickets to the World Cup finals, there’s no tariff on that.

And finally, point four means that a tariff is significantly more distortionary than a broad consumption tax. The revenue raised by a 10 percent tariff on imported cars could be generated by a much lower tax on all cars.

The regressive nature of tariffs is clearly one reason Trump likes them. The whole name of the game for him is finding ways to be a “populist” without backing away from the Republican Party’s fundamental commitment to plutocratic economic policy. Taxing imported goods is a way of raising revenue that falls overwhelmingly on low-income people.

Swiss cheese tax base

As he’s been raising taxes in a regressive and economically destructive way, Trump has also been chipping away at the income tax base through gimmicky policies with no real economic merit.

The general idea of tax policy is that while raising revenue is good, it’s important to do so in a way that doesn’t create huge problems of economic incentives.

From a growth and incentives standpoint, what does “no tax on tips” accomplish? The policy puts a small thumb on the scale in favor of tip-based labor compensation models rather than normal wages and salaries. But why? A few years back, anti-tip think pieces were all the rage. I mostly stayed out of it — tipping norms are basically fine, if a little weird. But tipping is certainly not so good that we need to restructure the tax code to encourage companies to compensate their employees that way.

Similarly with “no tax on overtime.” Sometimes employers pay bonus wages to get people to work extra hours. Other times, when employers need more hours of work, they add more employees. Which option you pick is going to depend on various considerations, and now one of those considerations is that Trump has created arbitrarily more favorable tax treatment for “pay extra in the form of overtime” relative to “pay extra in the form of more hiring.” Like a cut in the tax rates, this gimmick deprives the government of potentially useful revenue. But unlike a rate cut, it doesn’t really improve incentives to work and invest. It just introduces small biases in decision-making.

The auto loan interest rate thing is even worse, essentially encouraging people to go marginally deeper into debt by buying slightly more expensive cars.

A car is a necessary expense for most people. On the other hand, walking around any parking lot in America, it’s obvious that the vast majority of the population is not engaged in subsistence levels of vehicle consumption. Most people spend a lot of time in their cars, so they’re generally inclined to splurge a bit on making their vehicle a bit nicer than it really needs to be. That’s fine. But financing personal consumption with debt, as most car-buyers do, is a bit risky and unwise. Most people would be better off economizing more on early-life car purchases and having more money down the road, if for nothing else, than to eventually buy fancier cars.

These policies are all bad ideas, and they’re all somewhat regressive. But you can’t repeal them consistently with a Biden-style tax pledge.

Lots of good taxes are broad-based

Another, more hopeful, story is playing out in New York City. Congestion pricing was very unpopular in advance of its rollout but has become much more popular since it was implemented — the same trajectory we saw in Stockholm, Oslo, and London.

A major component of the case for congestion pricing is that, yes, it raises revenue, but its impact on incentives is beneficial. And critically, it’s not just that it’s beneficial to pedestrians who enjoy the reduced traffic noise and don’t need to pay the fee.

Once a year, I used to stand in line on Free Cone Day at Ben & Jerry’s. Ice cream is tasty, and free is an appealing price. And as an occasional excursion with friends, standing in a long line and chatting is fun. But it would actually be terrible if the only way to get an ice cream cone was to stand on a line that long. Of course free ice cream is great! But 95 percent of the time, a rational person would prefer to pay a modest fee to get ice cream at a reasonable pace than to stand on a long marketing gimmick line in order to get one for free. And that’s even more true if you assume that in the absence of charging customers, ice cream parlors would be kept in business thanks to generalized income taxes.

It’s just genuinely better for almost everyone for roads to be paid with user fees that offset congestion and wear and tear. The biggest problem with NYC congestion pricing is that it’s not applied broadly enough.

We should be shifting our tax code to rely more heavily on taxing things like congestion, pollution, marijuana, and alcohol. I think Paul Romer’s idea for a progressive tax on digital ad revenue is badly underrated. Everyone more or less agrees that we could almost all stand to spend a little less time scrolling. Directly taxing the business model that leads to so much investment in the compulsive scrolling industry seems like a big win.

Bharat Ramamurti wrote an interesting piece about the policy response to AI’s impact on the labor market. He floated the idea of taxing automation directly to fund investments in things like child care:

For example, South Korea has implemented a reduced tax deduction for automation investments that directly replace human workers, using the resulting revenue to fund expansions of their caregiving workforce. There are multiple ways that this tax could be designed, including a blanket corporate tax rate increase, a tax specifically on the AI firms most poised to benefit from this technology, a tax on corporate subscriptions to AI models, or a tax on firms that displace workers in favor of automation or AI.

This is an interesting thought, but it seems like a mistake to me. In some spheres, after all, automation could be a huge boon for society — there’s no need to put a special tax on it. But I think AI does raise the stakes on taxing social harms. If electricity usage spikes thanks to AI, it’s more important than ever to tax pollution. If a broad carbon tax is outside the realm of political possibility, maybe we could pair regulatory moves that are friendly to the deployment of new energy with a sector-specific carbon tax on data centers. An influx of self-driving cars will make it crucial to tax road use properly.

In all these things, tax progressivity matters, but it can’t be the only consideration. We need an overall good tax code.

It all starts with tariff repeal

What makes some of this workable is, again, that tariffs are so insanely regressive. If you cancel most of Trump’s tariffs and implement statutory safeguards against a reprise of this style of government, you could replace them with almost anything and that would be more progressive.

Classic progressive ideas like ending the “stepped-up basis” loophole and subjecting private equity limited partner earnings to normal income tax should, of course, be on the table. But they shouldn’t be the only things on the table; they should be there to ensure an overall progressive structure to a general tax reform, the main point of which is to re-broaden the tax base. On average, the rich will pay more and the non-rich won’t. But it is true that there will be specific instances of non-rich people paying more. That’s the price for implementing a rational system. There are just too many people in the country — and too much variation in circumstance — to keep promising that literally nobody in the middle class will ever face a tax hike.

Finally, I do think it’s worth recalling that while Obama ran on a strict tax pledge, he raised taxes on cigarettes. The Affordable Care Act included things like a new tax on tanning salons. It also, more consequentially, featured an individual mandate that the administration defended in court as a tax.

I’m not going to tell you that all of these were good ideas, but some of them were, and they all at least arguably violated the pledge.

I’d like to see politicians avoid making unwise promises because lying to the public is bad. But conditional on having made an unwise promise, it’s better to break the promise than to make bad policies. The fiscal circumstances of the country seem to be getting more dire, and it’s completely unclear what Congress may look like in the future. The next president needs to preserve maximum optionality, then see what she can possibly do with Congress to get more revenue in ways that have constructive, rather than destructive, consequences for the economy. That can and should include a more progressive code. But it cannot include an ironclad commitment to zero taxes on the bottom 95 percent of the income distribution.