StanCollender'sCapitalGainsandGames Washington, Wall Street and Everything in Between

The Idiocy of Starve-the-Beast Theory

27 Nov 2010
Posted by Bruce Bartlett
A prime reason why we have a budget deficit problem in this country is because Republicans almost universally believe in a nonsensical idea called starve the beast (STB). By this theory, the one and only thing they need to do to be fiscally responsible is to cut taxes. They need not lift a finger to cut spending because it will magically come down, just as a child will reduce her spending if her allowance is cut — the precise analogy used by Ronald Reagan to defend this doctrine in a Feb. 5, 1981, address to the nation.
It ought to be obvious from the experience of the George W. Bush administration that cutting taxes has no effect whatsoever even on restraining spending, let alone actually bringing it down. Just to remind people, Bush inherited a budget surplus of 1.3 percent of the gross domestic product from Bill Clinton in fiscal year 2001. The previous year, revenues had been 20.6 percent of GDP, spending had been 18.2 percent, and there had been a budget surplus of 2.4 percent.
When Bush took office in January 2001, we were already well into fiscal year 2001, which began on Oct. 1, 2000. He immediately pushed for a huge tax cut, which Congress enacted. In 2002 and 2003, Bush demanded still more tax cuts, even as the economy showed no signs of having been stimulated by his previous tax cuts. The tax cuts and the slow economy caused revenues to evaporate. By 2004, they were down to 16.1 percent of GDP. The postwar average is about 18.5 percent of GDP.
Spending did not fall in response to the STB decimation of federal revenues; in fact, spending rose from 18.2 percent of GDP in 2001 to 19.6 percent in 2004, and would continue to rise to 20.7 percent of GDP in 2008. Insofar as the Bush administration was a test of STB, the evidence clearly shows not only that the theory doesn't work at all, but is in fact perverse.
Of course, spending rose partly because of terrorist attacks and wars in Iraq and Afghanistan. While it's true enough that these events caused an increase, they do not come close to justifying a rise in spending equal to 2.5 percent of GDP. Even if one believes that things like the Transportation Security Administration were worth the money, it doesn't necessarily follow that they justify putting all the expense on the national credit card. Bush could have proposed spending cuts or tax increases to pay for these things and prevent a run-up in the debt. But he didn't. Throughout his administration he didn't veto a single spending bill until well into his second term.
Nor was Bush's budgetary profligacy limited to programs that could be justified, however loosely, on national security grounds. As I detailed last week, he and a Republican Congress created a massive new entitlement program, Medicare Part D, to buy the votes of seniors and buy themselves reelection in 2004. Among those voting for this monstrosity were many Republicans still in Congress today who are unjustly considered to be staunch fiscal conservatives, including incoming Speaker of the House John Boehner, House Majority Leader Eric Cantor, and House Budget Committee chairman Paul Ryan.
Because of its obvious ridiculousness, one seldom hears conservatives say openly that tax cuts automatically reduce spending. But it still underpins the entire Republican budget strategy — tax cuts never have to be paid for, no meaningful spending cuts are ever put forward, earmarks and foreign aid are said to be the primary sources of budget deficits, and similar absurdities.
But there is a flip side to STB at work as well. If tax cuts starve the beast, then it logically follows that tax increases must feed the beast. This variation of STB was on full display in a Nov. 21 Wall Street Journal op-ed article by Wall Street Journal editorial writer and Republican operative Steve Moore, who founded the Club for Growth, which gives vast sums to Republican candidates, and Ohio University economist Richard Vedder.
The Moore-Vedder article argues strenuously that tax increases must never be considered no matter how big the deficit is. The reason, based on research Vedder has been updating since the 1980s, is that tax increases always feed the beast, leading to spending increases larger than the tax increase. Originally, he said that spending would rise $1.58 for every dollar of tax increase, leading to an increase in the deficit rather than a reduction. Vedder now says that spending only rises $1.17 for every dollar of tax increase.
By this logic, the tax increase enacted in 1993, which raised the top federal income tax rate to 39.6 percent from 31 percent, should have caused a massive increase in the federal budget deficit. In fact, it did not. Spending was 22.1 percent of GDP in 1992 and it fell every year of the Clinton administration, to 21.4 percent of GDP in 1993, 21 percent in 1994, 20.6 percent in 1995, 20.2 percent in 1996, 19.5 percent in 1997, 19.1 percent in 1998, 18.5 percent in 1999, and 18.2 percent in 2000.
And contrary to another commonly-held Republican idea — that all tax increases reduce revenue via the Laffer Curve — revenues rose from 17.5 percent of GDP in 1992 to 20.6 percent in 2000.
According to Republican mythology, repeated by Moore and Vedder, the budget was balanced only because Republicans got control of Congress in the 1994 elections. But the deficit had already shrunk from 4.7 percent of GDP in 1992 to 2.9 percent in 1994. Budget experts who don't shill for the Republican Party generally agree that the budget reforms and tax increases of 1990 and 1993 — which were both enacted against strenuous opposition from almost every Republican in Congress — deserve the bulk of the credit for bringing down spending and the deficit with tough budget enforcement rules and higher taxes.
Moore and Vedder simply deny that the 1990 and 1993 tax increases had anything to do with the budget surpluses that emerged in 1998, even though higher federal revenues contributed 44 percent to the improvement in the deficit. Between 1992 and 1998, spending fell by 3 percent of GDP and revenues rose 2.4 percent.
To the extent that there was an increase in revenues, Moore and Vedder imply that it was due to a Laffer Curve effect from the cut in the maximum capital gains tax rate that Republicans achieved in 1997 — an argument first made by Republican speechwriter Clark Judge in a Wall Street Journal op-ed on Jan. 10.
To be sure, there is good research showing that a cut in the capital gains tax may raise short-term revenues due to an unlocking effect as people realize gains on assets they may have owned for many years. But this is a one-time effect. To raise revenues for more than a couple of years, a lower capital gains rate would have to raise investment and economic growth, which it undoubtedly does to some extent. However, it's implausible to attribute all of the rapid growth in the late 1990s to lower capital gains taxes. The keys were development of the Internet, which long predated the 1997 capital gains rate cut, and an easy money policy by the Federal Reserve.
In any case, Treasury Department data show that higher capital gains tax revenues contributed at most 0.33 percent to the 1.4 percent-of-GDP rise in revenues between 1997 and 2000. Of course, they contributed nothing to the decline in spending from 19.5 percent of GDP to 18.2 percent. Thus even taking the Republican argument at face value, higher capital gains revenues contributed at most 6 percent to the deficit improvement between 1997 and 2000.
Starve the beast is a crackpot theory, and its flip side that higher taxes invariably feed the beast is no better. They are just self-serving rationalizations for Republican budgetary irresponsibility.
Note: A few years ago, I went into great detail explaining the origin and development of STB in an academic journal. My article is available online for those with an interest in the gory details. In July, I posted a bibliography of more recent academic research in The Fiscal Times, all of which shows no evidence whatsoever that tax cuts reduce spending. More recently, the International Monetary Fund has confirmed this conclusion in a September working paper.

What a load of ****

Pure strawman, nothing more. Go interview a senior Republican congressional leader and you'll get the actual nuance of the "starve the beast" slogan, not the silly version you've turned it into.

Yes please, post the

Yes please, post the counter-argument here. Because it's entirely *not* self-obvious.


Go interview a senior Republican congressional leader and you'll get the actual nuance of the "starve the beast" slogan

Bruce has posted as least as sensible a version as I've heard from any sitting Republican. If you think it's a strawman argument, then post the information here.

You Rock Bruce

Thanks for pointing out, once again, that the emperor has no clothes.

Govt. impeding Govt.?

Such underhanded policies as these always seemed to me like they should be illegal.

If you have a problem with a government policy (e.g., spending), and you are a government representative with the ability to instigate change in policy, do so directly, with votes, representation, and legislation.

Instead, the sore loser takes the easy way out to get his/her way, working to undermine legislation that is purportedly the will of the country. Sore loser works against the country, until the country is brought to its knees to what sore loser wanted in the first place. Sore loser is too lazy to fight for his/her beliefs in a straightforward way that requires telling the truth to the country.

It's like a fireman not liking the brand of one of the hoses on the firetruck, so he cuts that hose until enough houses burn down and that brand hose gets replaced.

Logic and systematic thinking

Logic and systematic thinking - I'd expect nothing more from Bruce. Excellent article.
makita bhp452

What are you really trying to say?

This post is perplexing and confusing.

Milton Friedman often said: "Over a long period of time, government will spend whatever the tax system raises, plus as much more as they can get away with. That's why you have had universal deficits."

He was talking about revenue constraining spending, not tax rates. Now you are saying he was wrong (and "stupid"), because if revenue does indeed constrain spending then there you are -- with the beast being contrained by its feeding.

Yet everything you write actually confirms what Friedman said. Nothing contradicts it. E.g.:

"It ought to be obvious from the experience of the George W. Bush administration that cutting taxes has no effect whatsoever even on restraining spending, let alone actually bringing it down. Just to remind people, Bush inherited a budget surplus..."

What's a surplus? Free revenue! And Bush spent it as you detail -- and just like Friedman said.

"tax increase enacted in 1993 ... the top federal income tax rate to 39.6 percent from 31 percent... [yet spending fell]"

All during a period of revenue shortfall, and thus paygo budget restraints. But what happened as soon as the budget surplus came in? Creating new free revenue? Brad DeLong asked this of Ruddy Penner, then head of CBO. Quote:

Brad praises how well and honest the budgetary process worked in the early and mid-1990s, and what an important role the Budget Enforcement Act [BEA] and its Pay-As-You-Go restrictions [PAYGO] played...

Rudy: ... One substantive point: You rightly praised how the BEA worked from 1990 thru 1997. But then it broke down completely...

Brad: So what happened at the beginning of 1998 to change things so completely? That's still not clear to me...

Rudy: I believe it was the surplus. PAYGO was originally designed to stop tax and entitlement policy from increasing the deficit. (Really, to preserve the gains from the 1990 budget agreement.) After 1997, it had the effect of preventing any reduction of the surplus and that didn't make much sense...

To me, that sure reads like Penner saying that the spending beast *was* being starved *until* the unexpected free revenue, the surplus, came in. Then the constraints went off, promptly. How do you read it? Revenue is the key, not tax rates.

"Starve the beast is a crackpot theory, and its flip side that higher taxes invariably feed the beast is no better."

Yes -- higher revenue (not taxes) -- and we have a great natural experiment as to this: the Social Security surplus. It came in unexpectedly by accident (contrary to all mythology, the Greenspan Commission did not create it on purpose). Did Congress save it? Or even limit its spending of this gift to its face amount?

To quote Kent Smetters (.pdf):
"We find that there is no empirical evidence supporting the claim that trust fund assets have reduced the level of debt held by the public. In fact, the evidence suggests just the opposite: trust fund assets have probably increased the level of debt held by the public....

"... each dollar of Social Security surplus appears to have actually increased the debt held by the public in the past by $1.76."

A free revenue lunch came in, and the beast feasted. No?

"one seldom hears conservatives say openly that tax cuts automatically reduce spending"

True. And beating them up for what they don't say seems rather disingenuous.

OTOH, are you really saying that politicians don't regularly boost spending when new free revenue becomes available to spend? Because that's an interesting claim worth backing up.

If you really think Friedman was "stupid" for saying they do, and that Penner and Smetters were wrong in giving examples of it happening in actual fact on a big scale, please explain. (Things like the Laffer Curve have nothing to do with the issue.)

For one thing, why (as Friedman noted) have all the developed nations as a group -- no matter what their individual level of tax collections, high or low --for the last two generations near universally run long-term deficits at an average long-term rate of around 3% of GDP, their maximum sustainable amount ("as much more as they can get away with")?

Advocates of STB think a debt

Advocates of STB think a debt crises will force Washington to cut spending.

What they do not tell you is anything about that crises.

That is because deficits do not hurt government, rather they hurt the private economy.

So the STB strategy really is to destroy the capitalist economy to create their desired crises.

Apart from soak the rich tax

Apart from soak the rich tax policies, the only other thing that will reduce the deficit is to have some sort of job guarantee program with some minimum wage levels. There are multiple advantages. The no. of tax payees increase and hence tax collection. The will be reduced drain due to unemployment insurance and entitlements. The additional manpower can be utilized for rebuilding the public infrastructure at competitive labor rates.

That's a silly proposal

Who pays for this 'guaranteed job at a guaranteed minimum wage'? Far from being a deficit reducer, this is just welfare disguised. Since transfer payments is what is blowing up the budget, you've done nothing other than deluding yourself into believing that this accomplishes anything.

Supply Side Money Machine

A VAT on consumption is a more efficient way to raise taxes than taxes on production and investment. The ability to raise more taxes with less distortion makes the VAT a money machine that government will use to increase spending. In the same way, selling bonds to willing buyers is a better way to finance government then taxing production and investment. Supply side tax cuts will increase the supply of capital and reduce interest rates and increase the spending power of both the private and public sectors. Government will use the greater supply of capital as a money machine, just like a VAT. This is why starve-the-beast does not work.

What we need is a better way of controlling spending than deficit demagoguery.

Temporary increases in tax

Temporary increases in tax revenue after tax cuts remind me of the western electric experiments - turn up the lights and productivity increases, turn down the lights and productivity increases again. In that case, rotate capital gains tax rates regularly to unlock invested capital for redeployment to more productive uses.

The logic of this analysis

The logic of this analysis holds up only if it can be proved that government spending would still be exactly the same at any point during the time frames examined, had tax rates and revenue been higher than actual. This Bruce does not prove, or even attempt to prove.

Put another way, in his angry way Bruce does prove that lower taxes do not necessarily result in less spending. But he fails to make any case (or even to attempt to make the case) for his master assertion that lower taxes don't result in less spending than in a higher tax (presumably, higher revenue) regime.

I think that would be an almost impossible case to make for two reasons. First, there is no way to definitively know what government spending would have been with no Bush tax cuts. Second, I think on a gut level most of us probably believe that, given more money to spend, Bush would have spent a greater percentage of GDP than he did, and Obama would have done the same. But there is no way to truly know, which is why I find such self-righteousness as Bruce's somewhat amusing.

I'd also like to point out that GDP returned quickly to solid growth following the Bush tax cuts, despite the violent popping of the Internet bubble and 9/11. See: ( While stimulus spending has not had the same efficacy, in the face of the financial bubble explosiion (admittedly a more serious bubble by far, but one uncomplicated by a horrific 9/11-style attack.) I might also direct the reader's attention to the early 80s and a similar surge in GDP following the first Reagan tax cuts.


Blurgh! So the economy returned to solid growth after the Bush tax cuts because of the Bush tax cuts? Why because nothing else was happening at the time? Say something like flooding the economy with money from China through the real estate market? Oh no, that's right, that wouldn't fit the outcome you NEED to arrive at. \

In both bubble explosions,

In both bubble explosions, the Fed increased the money supply. Indeed, more so now than in the early 2000s. The main policy difference was tax cuts under Bush and stimulus under Obama. But let's call things like they are. A certain group believes that government spending increases economic activity best and another group believes tax-cuts do. Both groups hate each other and are expert in framing reality to "get the results they want." And then to angrily denounce the other group as "idiots." So let's just stipulate that before discussing anything further.

Magical Thinking

The logic of this analysis holds up only if it can be proved that government spending would still be exactly the same at any point during the time frames examined, had tax rates and revenue been higher than actual. This Bruce does not prove, or even attempt to prove.

I think you are completely wrong about this. To view this is simple logical terms, we can start with Bartlett's reduction of the STB theory, which is:

"A decrease in taxes always leads to a decrease in spending due to the declining revenues."

A valid counterexample would then be a case where a decrease in taxes led to an increase in spending. Now you can claim that his interpretation of the STB theory is wrong, and that the STB theory is actually:

"A decrease in taxes always leads to a decrease in spending when compared to leaving taxes the same or raising them, because decreasing taxes leads to less revenues."

But the problem with that, as Bartlett pointed out, is that under Clinton the tax rate was held the same for many years and yet the spending side decreased (it went from 22.1% of GDP in '92 to 18.2% in '02). This is in direct contrast to the period after the Bush tax cuts where the spending side increased from 18.2% of GDP in 2001 to 19.6% in 2004, and would continue to rise to 20.7 percent of GDP in 2008.

So taxes were lowered, and spending increased in '01-'08 which is a clear counter example to Bartlett's view on STB theory, and taxes were static or raised from '92 to '02 and yet spending decreased, which is a counterexample to the second view on STB theory.

Now your side claim, that spending would have been even higher had we increased taxes more during the Bush years, and even lower had we decreased taxes under the Clinton years may still stand, but that claim is not only unverifiable, but it doesn't even sound like a useful theory for governance. More than that, it seems less likely than the view that I hold: that spending levels have more to do with the needs and wants of the populace than what the tax rate is at any given period.

Straw Man

In response.... And also more info on why Bruce's understanding of STB is lacking.

"A decrease in taxes always leads to a decrease in spending due to the declining revenues."

I think this is simply a way to set up the philosophy so that you can make it take a fall. I think a better and more fair way of characterizing STB would be, "Less money to spend will result in less money being spent."

Also -- think about time horizons. Those are key. If you understood, you would know that STB-ers (like Reagan) were thinking in long time horizons -- consistently less revenue over time would in the long run result in a smaller government even while having little or no impact in the short run. I actually do believe that big deficits are part of the plan. Simply put, cuts in revenues would not in the short term result in cuts in spending, because the government and its special-interest parasites are addicted to spending and entitlements are holy. Therefore, spending would continue to grow, even with tax cuts, thus causing the deficit to grow, and it would be the deficit that would then FORCE the government to cut spending and would indeed be the only way to force government to cut spending. The mechanism is not Tax Cut = immediate spending cut, but Tax Cut = Deficit = Eventual spending cut. So in that view, all is going according to plan and you just need to expand your time horizon to see it.

Pertaining to the last two points: those are my points--that it's impossible to know what spending would have been with different tax rates in the past. Therefore to make a claim that tax cuts had no impact on spending is impossible and therefore a debunking of STB by looking to history in such a simplistic manner as whether spending was going up or down is not going to work. (By the way, pertaining to Barlett's argument as a whole--Let's say I made the following argument: even when we were putting people in jail in the 70s and 80s in NYC crime was still rising. Therefore, there is no evidence that putting people in jail deters crime, since crime still existed and was indeed rising.)

I do think there is plenty of real-time evidence for the efficacy of STB -- if you look at the panicked moves now to cut spending (including freezing federal worker pay) we can see that tax-cut fueled deficits are actually powering either a decline in spending or a decline in the rate of growth and have made the size of government, its role and its budget at the center of the national conversation. STB is indeed FORCING the government to cut, via the mechanism of deficits.

say what you like about the tenets of National Socialism...

"Republicans almost universally believe in a nonsensical idea called starve the beast"

I don't see any evidence for the assertion that Republicans have views on policy. They stand simply for a series of psychological resentments, and for redistributing income upwards.

And... um... reducing executive power (unless of course the president is a Republican), a more humble foreign policy (unless of course the president is a Republican), an aversion to new spending programs and deficits (unless of course the president is a Republican), judicial restraint (unless of course they like the result), and federalism (unless of course the president is a Republican).

"And contrary to another commonly-held Republican idea — that all tax increases reduce revenue via the Laffer Curve — revenues rose from 17.5 percent of GDP in 1992 to 20.6 percent in 2000."

Yes, that is one belief that Republicans have. Of course, it's false: "'Federal revenue is lower today than it would have been without the [Bush] tax cuts. There's really no dispute among economists about that,' said Alan D. Viard, a former Bush White House economist now at the nonpartisan American Enterprise Institute. 'It's logically possible' that a tax cut could spur sufficient economic growth to pay for itself, Viard said. 'But there's no evidence that these tax cuts would come anywhere close to that.' Economists at the nonpartisan Congressional Budget Office and in the Treasury Department have reached the same conclusion." George W. Bush claimed otherwise, and was entirely wrong.

(Because the media is terrible at reporting the news, the headline of that article read, "Lower Deficit Sparks Debate Over Tax Cuts' Role." There's no dispute over the facts of the situation, but, because Republicans say false things, the media reports it as a "debate" rather than openly informing readers about the news. It's kinda like post-modernism, where all beliefs are equally valid, except it destroys any semblance of rational policy, rather than obscuring critical thinking about literature).

Honestly: can you look back at the past 20 years and identify one issue, other than the need to reduce gov't revenues regardless of context, that could be considered a "Republican belief"?

What the Republicans Believe

Hey Elvis, you forgot one. Republicans believe in state's rights (as long as the states are right).

Bruce, Quite a lot that seems


Quite a lot that seems invalid about your analysis.

Others have pointed out correctly that (1) the independent variable in question is revenue levels, not tax rates, and (2) the question isn't if spending went up or down, but whether it was higher or lower than it would have been at different levels of revenue. Admittedly this is a tougher question to answer, since it's a counterfactual, but that doesn't mean it makes sense to substitute something else that's easier to measure, and as I think you know, at least a handful of economists have applied more sophisticated analytical methodologies (e.g., Granger causality and all sorts of complex algebra -- all that stuff the PhD economists do) to provide their answers to this question. I don't have links handy, but based on the conclusions of those analyses (inclusing meta analyses) it seemed to me (a non-economist) that they generally rejected the hypothesis that a dollar more in revenue has been associated with a full dollar more in spending, but there's a decent/good chance it has been associated with some incremental spending (between $0 and $1).

I'll add what I see as a #3 in the list of flaws in your methodology: You use as your spending metric spending levels as a percent of GDP. I guess that works better for the case you'd like to present, but shouldn't at least our default
metric for the dependent variable (spending) be simply spending (dollars) in real terms? If the argument can be made that the baseline for some type of spending should correspond to some extent to growth in GDP, that case0should have to be made, not assumed, correct. If I made 20% more income this year vs. last year and I spent 10% more on vacations, by your metric I've "cut" my spending on vacations because my spending as a percent of my income has decreased.

% of GDP

Citing budget numbers as a share of GDP seems to me to be an expeditious and simple way of controlling for inflation and population growth. I strongly suspect that if you take out the accounting for GDP and replace it with accounting for inflation and population growth, you'll see pretty much the same thing.

Your vacation analogy is entirely fallacious. The OP proposes comparing spending to GDP. Your analogy compares spending to revenue. Ooops. If the country made 20% more this year at large, it doesn't mean that the government did, especially not per-capita or in constant dollars. It might mean that there are 10% more of us earning incomes denominated in (whatever-it-works-out-to)% inflated dollars.

Recent comments


Order from Amazon


Creative Commons LicenseThe content of is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License. Need permissions beyond the scope of this license? Please submit a request here.