Raising taxes is hard

It's important to set priorities and make good choices with public money

Illinois has a 4.95 percent flat income tax rate. In 2020, the state held a ballot initiative on whether to amend the state constitution to allow for the imposition of a progressive income tax. It got 46.7 percent of the vote in a year when Joe Biden got 57.4 percent. Four years later, they held another ballot initiative. This was an “advisory initiative” on the idea of creating an additional 3 percent tax on income above $1 million, with the revenue earmarked for the purpose of reducing property taxes. This one passed with 60.8 percent of the vote in a year when Kamala Harris got 54 percent.

These are, I think, useful facts to assimilate when pondering the politics of taxation and specifically the politics of taxing the rich.

When Trump tries to pass a bill that simultaneously adds trillions to the deficit while cutting Medicaid and nutrition assistance, Democrats are happy to lambaste it as also featuring huge regressive giveaways to the rich. And of course, something like this Yale Budget Lab analysis showing the hideously unfair combined effect of the One Beautiful Bill and Trump’s tariffs is catnip for Democrats.

But it remains the case that if Republicans ever wanted to shrink the state by just passing a broad-based tax cut, I’m not sure what Democrats would say. Even when Donald Trump came up with a couple of gimmicky, populist-sounding ideas like “no tax on tips” and “no tax on overtime,” Democrats found themselves with no counterargument they were prepared to defend.

Note that even a broad-based tax cut can be somewhat regressive. If you take the current 12 percent income tax bracket and lower the rate to 10 percent, that delivers a bigger tax cut to people in the top fifty percent of the income distribution than to people in the bottom and does nothing at all for the poor. So I’d be against it personally, but I’m not sure that Democrats would have a leg to stand on politically. It’s regressive, but not in the “only billionaires benefit from this” sense.

Conversely, as you saw in that first Illinois ballot initiative, even taxing the rich isn’t a total no-brainer politically. I don’t know exactly what the Biden/no crossover voters who killed that package were thinking. One concern might be that the legislature would end up defining “the rich” more expansively than some voters are comfortable with. Another concern might be that hiking taxes on the rich would be bad for the business climate in the state, which would have negative downstream consequences. But I think these voters may also have been driven by a vague sense (shared even by many Democrats) that the state government in Illinois is not necessarily a responsible steward of their tax dollars. That’s why when it was reformulated as “tax the rich to cut your taxes,” it went from underperforming the national Democratic Party ticket to overperforming it.

Voters don’t like the Republican Party’s obsession with catering to rich people, but they also don’t like paying taxes.

As a result, I think it’s important for progressives not only to keep making the case for revenue, but to be more thoughtful about the opportunity cost of spending money on ideas that are just okay.

War is the health of the state

Not only is raising taxes difficult now, as far as I can tell, it has always been hard. I often hear people talking as if there were some golden age of social democratic politics in the United States, when bold elected officials immune to donor interests were happy to stick it to the wealthy. But if you look at the actual history of the federal tax share of the economy, I think you see a different story. Taxes soared during World War II and have been bouncing around in a pretty narrow range since then.

Note that while taxes were initially raised to finance the war effort, they didn’t fall when the war was over. Initially, this was because the postwar demobilization was followed almost immediately by the Korean War. And then after Korea, policymakers decided that the US needed to break with precedent and maintain a large peacetime military to prosecute the Cold War. But the defense share of GDP has been generally falling since Korea. Even the surge in military spending associated with the Vietnam War left us below the Eisenhower level, and the defense spending hikes associated with Ronald Reagan and the War on Terror were similarly below the prior peak.

By the time Joe Biden was president, some people were treating the cost of supplying weapons to Ukraine as some kind of insane burden on the taxpayer, even though total military spending was at a record low.

But even since 2000 or so, total defense spending has been so low that it’s hard to generate significant extra funds for anything else by cutting it. The phrase “peace dividend” comes from the end of the Cold War, but by the numbers, the biggest peace dividend came from the mid-1950s through the mid-1970s or so. This was actually the height of the Cold War, but also a period when a lot of big new government programs — notably Medicare and Medicaid — were created and when Social Security became more generous.

This tends to be a bit underrated, but the tax code has generally trended in a more progressive direction over the past 25 years. When Republicans are in office, they pass tax cuts that reduce everyone’s taxes — but they reduce rich people’s taxes the most. Then when Democrats are in office, they raise taxes — but only on rich people. The upshot is that even though neither party really favors the “prioritize middle class tax cuts” policy agenda, the public is sort of getting what it wants via ping pong. Republicans know they can’t sell tax cuts for the rich without cutting the middle class in on the deal, and even though Democrats’ preference is higher revenue, they know they can’t sell tax hikes that include the non-rich. This is part of a trend where after the huge increase in inequality in the 80s and 90s, we’ve seen a trend toward generally falling inequality in more recent times. But these basic dynamics are obviously a serious challenge for any major social democratic project that’s counting on large increases in spending.

Targeting matters in the real world

I think these facts are important background for understanding certain things that people like to argue about on the internet.

For example, the problem with promising to cut bus fares to zero is that while, in most cases,¹ this would be an improvement for riders, all evidence indicates that bus riders are more worried about the frequency and reliability of the service than its price.

If you want to be completely abstract about it, this doesn’t matter. What matters is that any given level of transit service requires a financing mechanism. That mechanism can be either fares or general tax revenue or some mix of the two. Arguably, cities that have crowding during peak times should charge fares then to manage demand, but off-peak transit and transit in cities with lower ridership should be free. All the revenue should come from taxes, because in a behavioral sense, cities want to encourage transit ridership, which reduces local air pollution and traffic congestion. Fares make sense if and only if crowding is a concern, otherwise you should rely on general taxation.

The problem is that in the real world, raising taxes is hard. There are a lot of things that a city might do with an increased dollop of tax revenue. They might run the buses more frequently. They might hire more police officers. They might renovate some parks. A prudent political movement needs to give some consideration to what it chooses to prioritize.

This particular free bus thing is big on the internet because it annoys transit nerds, but it’s pretty trivial in the grand scheme of things.

A much bigger deal is the question of means-tested versus universal programs. This is an important question to wrestle with, because the proponents of universality are, in my view, completely correct. There are tons of advantages to structuring programs like public schools, where everyone is eligible regardless of need, rather than having income verification and phase-out schedules. America’s habit of sharply means-testing many of its most important social programs generates a lot of problems:

• The working poor often face very high de facto marginal tax rates.

• Because being married is more cost-effective than being a single parent, means-testing often creates marriage penalties.

• When you create an eligibility check, many people who are factually eligible will nonetheless fail to qualify due to administrative burdens.

• Means-testing tends to create a lot of political anxiety about deserving versus undeserving poor, whereas nobody seems to think we need work requirements to visit a public park.

• Ideally, when a program is used by everyone, that generates a lot of scrutiny and buy-in to try to make it into something that works well, rather than an isolated and marginalized thing serving only the poor.

Three cheers for universalism!

The problem is that if you took Medicaid, SNAP, EITC, CTC, and other means-tested programs and insisted on eliminating the income phase-outs, the programs would all need to become dramatically stingier. Of course, you could say, “No, don’t make the programs stingier, just raise taxes and spend more money!” But that’s hard. In practice, if progressives had spent the past 50 years insisting on universalism in all things, the result would be a lower level of spending on the poor. Robert Greenstein makes a strong case that willingness to opt for means-testing to save money has worked pretty well over the decades to build programs that are durable and helpful to low-income households. And if you measure by actual consumption (rather than an official poverty rate that excludes the impact of in-kind transfers), low-income households really are better off than they were a generation ago.

But you could also push this logic too far in the other direction. If you means-test a program to the extent that you create massive disincentives to work, you’re going to create giant substantive problems.² What’s more, over-focusing on any specific poverty threshold would lead you to design policy that is narrowly cost-effective from the standpoint of the threshold you picked, but not necessarily a good idea all things considered. This is all why, in principle, a universal program is best. The point, though, is that there’s a difference between what it might make sense to do if we could transform public opinion and what makes sense as good advice for politicians to follow. Recall that we started with a referendum result in Illinois. The difficulty with raising taxes isn’t timid Democratic Party politicians, it’s that Democrats’ own voters are somewhat more skeptical of raising taxes than the politicians are.

The inescapable need to set priorities

The interminable Build Back Better debates of 2021-2022 feel like a dozen lifetimes ago.

But the signature feature of those arguments was that neither Joe Biden nor Chuck Schumer nor Nancy Pelosi felt willing or empowered to set priorities, and the Democratic caucuses on Capitol Hill had no mechanisms for internal consensus-building. Beyond the various problems with Joe Manchin, House Democrats couldn’t agree on a formula to raise taxes enough to cover all of their own priorities. What they wrote instead was a weird zombie bill that claimed to be fully paid-for only because it featured lots of temporary programs. This never made sense as a legislative agenda relative to spending the same amount of money on a smaller number of permanent programs — ideally, good permanent programs.

To accomplish that, though, Democrats would need to decide which programs are better (or more important) and which are worse (or less important).

Even when BBB got cut down to size by Manchin, Democrats continued to struggle with priority-setting. The climate titles included money to sustain R&D and early deployment in potentially game-changing technology (very high impact), money to accelerate the deployment of already mature technologies (significantly lower impact), and a chunk of change for (borderline useless) “green bank” schemes. That’s not the greatest sin in the world; nobody writes a perfect bill. But at a time when Democrats were trying to show the country they were taking inflation seriously, it was dumb to spend money on things with very low efficacy.

The country as a whole has moved into a world where budget deficits increasingly do matter, where the Social Security Trust Fund is eight years from insolvency, and where the international situation militates in favor of higher levels of military spending rather than lower. That doesn’t mean “never spend money on domestic programs” — far from it. But it does mean that the money is probably going to be hard to find. That means how much good you do is going to be heavily influenced not just by what you say yes to, but by whether you make smart choices about what to say no to. Getting there is going to require a major mindset shift, but it’s incredibly important.

1. In some situations, it might exacerbate crowding.

2. This was a big part of the story with the old AFDC cash welfare program. The “reform” solution that congressional Republicans hit on and Bill Clinton signed was bad. But the program itself was genuinely a case of terrible design.