In the late 1970s, some conservatives began arguing that tax cuts pressure legislators to slash spending, while tax increases merely give politicians more money to spend. Since tax cuts take away politicians’ “allowance,” the theory went, enacting them would force legislators to pare back spending.
Four decades later, “starve the beast” continues to drive conservatives’ support for tax relief, as well as their opposition to tax increases as a deficit-reduction strategy. But has this theory held up in practice? The Manhattan Institute’s Chris Pope and Brian Riedl debate this crucial fiscal policy question.
Chris Pope: Politicians like spending other people’s money and will do it as often as they can. But while public spending can be popular, broad-based tax increases are one of the most unpopular things a politician can do. That’s why they’re rare. The exception—the tax hikes of the early 1940s, which expanded the share of households paying income taxes from 6 percent to 72 percent—was prompted by the need to finance the Second World War. From 1936 to 1945, federal revenues surged from 5 percent to 20 percent of GDP.
As peace returned, defense spending fell, but taxes weren’t cut back. This helped Democrats dominate Congress for a generation, since it let them boost domestic spending without needing to raise taxes to pay for those programs. The persistence of wartime taxation generally enabled politicians to make indiscriminate expenditures. As John F. Kennedy’s economic advisor, Walter Heller, argued: when the “pie grows bigger, there is less reason to quarrel over who gets the biggest slices.”
By Jimmy Carter’s administration, defense spending had been slashed, but federal revenue kept growing, as rapid inflation pushed people into higher tax brackets. To stop these stealth increases, Ronald Reagan cut marginal rates and indexed tax brackets. In his first Oval Office address, the Gipper argued against “those who told us that taxes couldn’t be cut until spending was reduced.” Instead of lecturing Congress “about extravagance until we run out of voice and breath,” he suggested “simply reducing their allowance.”
This approach has largely succeeded. Federal domestic spending, which surged from 9 percent of GDP in 1962 to 16 percent in 1981, remained only 18 percent in 2018. Low interest rates enabled a bipartisan spending surge during Covid, but people wouldn’t even be talking about cutting spending if tax revenues had kept up. And though Joe Biden’s agenda has been very costly, he was forced to abandon four-fifths of the spending he proposed in the Build Back Better Act—entirely due to lack of revenues.
Brian Riedl: Your assertion is one side of the “starve the beast” theory, which claims that tax cuts motivate politicians to enact spending cuts, while tax hikes motivate them to pursue spending hikes. This argument’s logical chain is that, when Congress cuts taxes, the resulting deficits quickly induce a public backlash and demands for offsetting spending cuts. The flip side of this theory argues that any tax-increase revenues will be plowed into new spending rather than left for deficit reduction. Thus, the most effective way to slash government spending is by first aggressively cutting taxes.
This theory contradicts both price and public-choice theory. Tax cuts lower the “price” that voters pay for government spending and lead them to want more “free lunch.” If you want voters to oppose spending hikes, make them pay for that spending with accompanying taxes.
It is ahistorical to believe that abandoning all fiscal responsibility on one side of the federal budget will cause Congress and voters to demand even tighter fiscal responsibility on the other side. Lawmakers who cut taxes—typically backed up with “deficits don’t matter” rhetoric—cannot credibly turn around and make a deficit-focused argument to slash popular Medicare benefits or education spending. They’ve already surrendered any deficit credibility. This is why GOP lawmakers who had focused obsessively on deficits during the early 2010s suddenly stopped talking about them following Donald Trump’s 2017 tax cuts.
Instead, Congress and voters rotate between eras of “deficit panic” (with tax hikes and spending cuts) and “deficit dismissiveness” (with tax cuts and spending hikes). The historical record is clear:
The 1981 Reagan tax cuts were followed by a defense-driven spending spree with no significant spending cuts.
The 1990 Bush tax hikes came with significant spending reforms, and, after the 1993 Clinton tax hikes, were followed by the largest spending decline since the 1950s, culminating in the 1998-2001 budget surpluses.
The 2001 Bush tax cuts ended that era of fiscal consolidation and were followed by a historic spending surge of war spending, “compassionate conservative” domestic spending, and new entitlement programs.
The 2017 Trump tax cuts came with a GOP-led 13 percent discretionary spending hike in one year, and the past seven years have seen the largest spending spree since World War II.
“Starve the beast” fails both as theory and in practice.
Chris Pope: Large tax cuts don’t lower the price of spending at the margin, because Congress typically makes spending decisions independently of major tax legislation. The big Reagan, Bush, and Trump tax-cut bills did not include significant spending.
What, for politicians , does set the price of spending? The prospect of pain-inducing budget deficits. When the economy is not running at capacity, the country can incur substantial deficits without pushing up inflation. We typically see this in recessions, when congressmen in both parties spend without fear of the consequences. We also saw this in the 2000 election, when a budget surplus encouraged Bush and Gore to compete over who could promise seniors the most expensive prescription drug benefit.
Reagan’s tax cuts created big budget deficits, which made spending increases politically painful for a generation. If you look at appropriated expenditures, it’s clear that the Gipper turned the tide; non-defense discretionary spending had grown from 3.3 percent of GDP in 1962 to 5.1 percent in 1980, and fell to 3.3 percent in 1989. It hasn’t come close to Carter-era levels since, remaining at only 3.4 percent of GDP as of 2023.
World events sometimes necessitate military build-ups, but such spending tends to quickly fall back when circumstances change: defense spending in 2023 (3 percent of GDP) was much lower than in Carter’s last full year (4.8 percent), despite bursts of rearmament under Reagan and George W. Bush.
Deficits create pressure on politicians, and this pressure has been responsible for many of America’s major spending reforms. The bipartisan Gramm-Rudman-Hollings Balanced Budget Act of 1985, which established caps on discretionary spending, would never have become law otherwise. The same is true of the 1990 bill establishing pay-as-you-go (PAYGO) budgetary rules, which require spending increases to be paid for, and the 2011 Budget Control Act, with sequestration provisions that trigger automatic cuts if Congress exceeds spending caps. Democratic politicians have repeatedly cut Medicare because it was the only source of funds for other programs they supported.
Entitlement reforms don’t happen because pro-cut legislators build up credibility to enact them. They happen because pro-spending legislators feel forced into it by a lack of revenue.
Brian Riedl: True, the Reagan, Bush, and Trump tax cuts did not contain major spending sections—they just set the stage for the spending sprees that followed. The 1981 Reagan tax cuts probably make the best case for starve-the-beast, and even that 1 percent-of-GDP decline in tax receipts between 1980 and 1990 was merely met with spending remaining at its elevated 21.2 percent-of-GDP level. And defense build-ups absolutely count as spending hikes, particularly during peacetime. Shifting the bulk of discretionary spending expansions from domestic to defense programs is not starving the beast.
As stated previously, the 2001 Bush tax cuts were immediately followed with new war spending, annual nondefense discretionary spending hikes as high as 13 percent, a massive new Medicare drug entitlement, and a doubling of farm subsidies. Republicans had no credible argument against this spending after Vice President Dick Cheney declared “deficits don’t matter” during the 2001 tax cuts debate.
And I’ll reiterate that the 2008–2016 GOP deficit-hawk era (under a Democratic president) ended abruptly after President Trump and Republicans cut taxes in 2017. Shortly after these reductions, a unified GOP government abandoned the Budget Control Act spending limits andhiked discretionary appropriations by 13 percent in one year (2018)and then abandoned all Tea Party–era spending-cut proposals. I had worked in the Senate and continued to speak with lawmakers after 2017. Several GOP lawmakers told me that they could no longer bring up deficits at voter town halls without voters screaming “fine, start with repealing your tax cuts.”
Finally, it’s a major reach to credit the modest 1990 PAYGO law as a starve-the-beast victory when it didn’t occur until nine years (and $2 trillion more in debt) after the 1981 tax cuts. Declaring PAYGO a tax-cut victory is especially strained considering that the bill was enacted alongside major tax increases. Similarly, the Budget Control Act, with its sequestration provisions, wasn’t passed until a decade (and $7 trillion in debt) after the 2001 Bush tax cuts. And that law was motivated by the Tea Party response to the recent Great Recession bailouts and stimulus spending, not the 2001 tax cuts.
If we insist on defining starve-the-beast deficit reduction victories as trillions in tax cuts and decade-long spending sprees, eventually followed by modest (and often bypassed) budget-process reforms that ultimately saved little money, then I’m not sure we can afford many more such victories. But if you want spending restraint, you won’t beat the 3.5 percent-of-GDP spending decline—culminating in a historic balanced budget—that followed the 1990 and 1993 tax increases.
Chris Pope: While it’s clear that both defense and domestic discretionary spending constitute a significantly reduced share of GDP in the revenue-starved post-Reagan era, you correctly note that overall spending is up a bit. This, of course, is due to entitlements.
Tax cuts’ influence on entitlement spending is hard to assess because such expenditures are set by law to grow automatically over time as the population ages, prices rise, medical capabilities expand, states raise matched contributions, and recessions swell eligibility. To examine the relationship between tax cuts and entitlement spending, we need to focus on when Congress makes statutory changes.
In the pre-Reagan decades, fast-growing revenues fueled relentless entitlement expansion. Congress created and repeatedly expanded Medicare, Medicaid, disability-benefit, and food-stamp programs, without offsetting spending cuts. Between 1950 and 1971, Congress increased Social Security benefits every election year, and it indexed them to inflation in 1975.
Congress bailed out Medicare in 1977 with a 25 percent payroll-tax hike, but the program ran out of money again in 1983. That year, Reagan refused to raise payroll taxes further, which prompted Congress to cut Social Security benefits.
Since then, Congress has offset most entitlement expansions with cuts elsewhere. Federal entitlement spending as a share of GDP in 2023 was substantially less than the Congressional Budget Office had predicted in 2000, despite the passage of Medicare Part D, Obamacare, and Covid-relief packages.
You noted that Reagan’s tax cuts have had the most demonstrable anti-spending effect. That’s because they were many times larger than any other tax change. Reagan’s 1981 tax bill reduced federal revenue by 4 percent of GDP in 1985 (an amount that increased every year) because it ended the revenue-generating “bracket creep” that Congress had counted on to balance the budget. All fiscal politics since Reagan’s reforms has been mere tinkering within the revenue regime he created.
It’s true that the 1990 budget agreement eroded the revenue reduction from the Reagan cuts a bit (by an average of 0.5 percent of GDP over four years), but the policy was part of a bipartisan package, including spending reforms, to deal with the Reagan deficits. Congress’s 1993 tax increase was similarly modest, and its main contribution to the late 1990s spending restraint was to cost the Democrats their 40-year hold on the House of Representatives.
And while George W. Bush was no model of fiscal restraint, both his spending increases and tax cuts were the products of campaigning in an era of surpluses. To assess the impact of his tax cuts on our debate, the key question is: Would Congress have spent more or less without them?
We can’t directly observe that, or any other, counterfactual just by looking at the single case of the U.S. federal government. But comparative studies are suggestive.
In state and local governments, for example, revenue growth appears to lead to expenditure growth, but shifts in spending do not seem to alter taxes. When the federal government gives grants, studies show that states use the money to boost spending, rather than to reduce their own taxes. Internationally, countries with less efficient tax systems have lower spending levels, while those with broader tax bases have larger welfare states.
Finally, it’s significant that big-government advocates uniformly oppose major tax cuts. Why doesn’t America have Medicare for All? As you once correctly observed: “The obvious answer is because the American people would never accept a $32 trillion tax increase.”
Brian Riedl: Happy to wrap up this helpful discussion. First, incorporating pre-1980 legislative behavior into this debate is tricky because, back then, much of the tax code and some major entitlement benefits (such as Social Security) were not automatically indexed for annual inflation. Consequently, Congress had to pass regular “tax cut” packages simply to inflation-adjust the brackets, and many of the entitlement-expansion bills were passed for a similar purpose. Legislation in the 1970s and early 1980s has rendered most of those subsequent inflation adjustments automatic. This means that comparing fiscal-policy bills before and after 1980 may reveal less than we think.
I will also push back on the claim that most entitlement expansions have been offset since the 1980s. While you’re right about Congress offsetting most reforms to Medicare payment rates, large non-offset entitlement expansions include trillions of dollars for Medicare Part D, farm subsidies, student loans, steep veterans’ benefit expansions, emergency spending in recessions and pandemics, corporate bailouts, SNAP expansions, CHIP, child-credit outlays, and EITC outlays. The easily waived PAYGO rules you offered in defense of the starve-the-beast legacy prevented none of those deficit-financed entitlement expansions.
Also, I don’t agree with classifying the 1983 Social Security deal as a starve-the-beast victory when the bill paired benefit cuts with tax increases. Yes, conservatives preferred this outcome to one entirely consisting of tax hikes, but that does not relate to the starve-the-beast idea that tax cuts somehow drove these spending reductions.
You note that we do not have counterfactuals for how spending trends would have proceeded without these tax changes. That’s true. But we do have: 1) both price theory and public-choice theory debunking starve-the-beast in theory; 2) 40 years of near-perfect correlations in which tax cuts were followed by large spending sprees, and tax hikes were followed by sharp federal-spending reductions; and 3) lawmakers themselves asserting that earlier tax cuts turned their deficit-based arguments against spending increases into a political punchline.
Theory, practice, and lawmaker testimonials all agree that starve the beast does not work, as even President Reagan’s own top economist eventually conceded. In the imperfect world of social-science research, that’s as solid as we can get.
Photo: Douglas Rissing / iStock / Getty Images Plus