Garett Jones  

Starve the Beast Mark II: Deficits, Debt, and Spendthrift Grasshoppers

Hummel on Bernanke's New Centr... Yglesias on Freedom in the 50 ...
Milton Friedman's theory of Starve the Beast said that there was a "politically acceptable deficit," so a one-time ex nihilo cut in taxes would have to cause a spending cut to keep the deficit stable.  My commenters have revised (or maybe just clarified) this theory, drawing partly on academic work.  The revised version of Starve the Beast (STB) says that ex nihilo tax cuts cause spending cuts not because of deficits, but because of the large debt that the deficits eventually build up: STB kicks in someday.  That's an improvement in the Starve The Beast story.  

Still, I offer three critiques of STB: 

1.  As this author notes, if a history of deficits increase tolerance for deficits then the politically acceptable deficit expands, and the door opens for extra spending.  Starve the Beast ironically grows the appetite and so makes future spending easier.  We should worry about electing tax-cutting Grasshoppers for that reason.  

2 STB still ignores the fact that tax cutters are typically across-the-board Grasshoppers: Tax cutters are spending exploders.  Tax cuts are not exogenous: The kind of politicians who cut taxes are usually the kind of politicians who boost spending.  

3.  If #2 is true, then any simple STB story runs into a tension: Is the spending created by the tax-cutting Grasshopper a bigger problem than the spending you're hoping to prevent with your beast-starving deficits?  The Grasshopper's extra spending now is a fact, while the feared extra spending of the future is still just a theory...supporters of limited government should be wary of a politician who says "I'm going to spend like crazy but I'll prevent the next politician from spending even more." 

This blurring of "deficits" with "tax cuts" is an important part of the failure of any simple version of Starve the Beast.  If we focus on actually existing politicians rather than hypotheticals, I think we'll see that the "deficits" created by the alleged Beast Starvers involve a lot of movement on the spending side of the ledger. 

Comments and Sharing

COMMENTS (7 to date)
Ken Plahn writes:

Totally agree. Getting comfortable with higher debt levels is a good point.

Using a household analogy, one could force a reduction in future spending by paying less on bills now. Debt would rise and you would be forced to reduce spending. One would likely also become comfortable with higher debt levels.

James A. Donald writes:

How can you deduce that raising taxes cuts spending, when, in the midst of the sequester, government spending continues to explode in the sense of new hires, more jobs for the boys, ever more grants for ever sillier projects?

The sequester cuts are typically something short term and silly, but each new hire creates a permanent new charge upon the state that will eventually lead to a gigantic pension obligation, each new grant creates or increases an interest group that will forever grow.

Because government employees are fire proof, each new job for the boys should be costed at not what it costs today, but what it will cost all the way to retirement, similarly each new disability SSSI. If costed on that basis, government growth has arguably increased during the sequester.

What arguably is happening is that civil servants are reacting to the sequester by giving long term benefits to their pals and political allies in place of short term handouts to their pals and political allies, moving government expansion off the books - much as the fed has collected a vast pile of dud mortgages which it keeps on its books at face value.

Garett Jones writes:

Good analogy, Ken.....

Chris Koresko writes:

As John Cochrane points out, the deficit is a function of more than tax rates and spending. Economic growth rates can have a huge impact over the long haul.

So if your Grasshoppers are conservatives who cut tax rates but also deregulate, push through international trade deals, etc., which tend to boost growth, then they may be much more acceptable than liberal Grasshoppers who add regulation and slow or halt trade deals. Yes?

On a side note: I recently read a claim that the majority of the growth in the federal deficit under Reagan was due to the increased cost of servicing the existing debt. The servicing costs rose because the Fed raised interest rates in a (successful) attempt to get inflation under control. So maybe Reagan wasn't so much of a Grasshopper as the STB studies assumed. I don't know how much their conclusions relied on that one data point, but I'm under the impression that Reagan's economic policies were heavily influenced by Friedman's ideas. So his term might have looked like an ideal test case for STB.

sourcreamus writes:

There is a political explanation rather than a personality type explanation. Democrats want 1. high spending 2. high taxes and 3. low deficits. Republicans want 1.low taxes 2.low taxes, and 3. low deficits. Neither side has enough strength to pass its agenda and so must compromise. Increasing taxes or lowering spending by themselves mean that one party has lost. Increasing spending and lower taxes mean both parties have partial wins. Likewise lower spending and higher taxes mean both parties have partial wins.

Jim Glass writes:

I really don't understand what the purported argument is about.

Strave the Beast is *not* a "theory", it is a policy prescription. IF one believes that government has grown too large, into negative returns -- as Coase put it, after examining govt programs for decades in the Journal of Law and Econ -- then one wants to limit its revenue to restrain its size, at least its growth in size.

If this is the inferior policy then the superior policy to restrain govt growth must be to give the govt more money to spend. Really???

Someone is going to have to explain that to me.

What has been the real-world result of revenue constraints? Again, three serious examples...

1) Social Security running out of revenue in 1983 resulted in a deal that was 50% benefit cuts, including (hidden) means testing by making benefits subject to income tax (but with the tax remitted to the SSA, not general revenue -- a benefit cut). Result: A smaller program and less money being paid by the poor to the rich. This was *good*.

2) The deficits of the post-Reagan years politically drove the enactment of the BEA and its paygo rules, as I mentioned before, which were a real constraint on spending growth. Which was *good*.

Then the deficit ending and the surplus arriving caused the end of paygo restrictions and much easier spending compounding through the next several years. As CBO head Penner stated. If that wasn't deficits restraining spending exactly as Friedman described, what was it???

3) Today's sequester. It's reducing spending. And when I pick up the NY Times yesterday, it reports that the sequester is driving bipartisan Medicare reform to reduce cost, as part of a deal to resolve the budget fight (for the moment). This is *good*.

All these gov-growth restraining, efficiency enhancing *goods* came from revenue restraint and hitting the "acceptable deficit" limit, just as Friedman said. With greater revenue (SS being fed general revenue, without the Reagan and Bush-Obama tax cuts) they wouldn't have happened.

Why don't you want such good things to happen?

How can the alternative to restricting revenue, that is *not* restricting revenue(!), be a better policy for limiting the size and growth rate of government?

I think the story should end right there, until that question is answered.

But just for the fun of argument, I'll run on...

Milton Friedman's theory of Starve the Beast said that there was a "politically acceptable deficit," so a one-time ex nihilo cut in taxes would have to cause a spending cut to keep the deficit stable

Where did he ever say that last part? You don't quote him saying that, you link to Ponnuru writing about Chait for some reason. And even so, Ramesh says...

Milton Friedman correctly noted that total federal spending must equal federal revenues plus the maximum politically acceptable deficit..."


"...But it does not follow that reducing revenues would reduce spending."

Well not immediately, that fiscal year or the next. But where did Milton ever say that? That's a total straw man. Yet isn't that what you are arguing against when you claim he said a "cut in taxes would have to cause a spending cut to keep the deficit stable".

C'mon, Friedman was no naif. He lived through the Reagan tax cuts and knew they did *not* cause spending cuts to keep the deficit stable. You're claiming he was not merely a naif about what would happen but also insensate about what actually did.

He never said the Bush tax cuts would reduce spending then but explicitly that they would restrain growth *later* -- and he certainly never "predicted" that when deficits are **above** the politically acceptable level in a fiscal/debt crisis that tax increases to close them would instead be spent on "a binge of bridge-building, school-expanding, and needless ditch-digging".

Friedman said "In the long run government will spend whatever the tax system will raise, plus as much more as it can get away with." In the long run. And he was right, see #1 #2, #3 above.

You yourself quoted him as saying "resulting deficits will be an effective restraint on the spending propensities of the executive". Now, it is a pretty big jump from restraining spending propensities in the future to forcing current spending cuts. No? It seems you are arguing against a Straw Milton.

Another point:

STB still ignores the fact that tax cutters are typically across-the-board Grasshoppers.

Why pick on "tax cutters"? Non-tax cutters who create $100 trillion of unfunded entitlements liabilities aren't Grasshoppers?

StB as policy is derived from the "theory" that ALL politicians are Grasshoppers. As I noted before, there are **no** Ant politicians or governments. Where in the OECD is the Ant government that hasn't run 40 years of deficits to wind up with massive unfunded entitlements coming at it?? Name those Ants.

If govt quickly currently spends all new revenue as it comes in then not restricting revenue will simply make it grow faster. QED. For "small govt"-ers restraining revenue then becomes logically unavoidable as policy.

If you wish to attack StB, here is the "theory" point where to do it. Show govt *doesn't* rapidly currently consume new revenue as it comes in -- the examples of the Social Securty surplus, circa 2000 fiscal surplus, Europeans VATs (and rest of all history) aside. Give us a "business-like" govt run like a fiscally sound, optimally-sized and run for greatest productivity business, and then StB becomes moot. Good luck.

Last point:

if a history of deficits increase tolerance for deficits then the politically acceptable deficit expands, and the door opens for extra spending.

A very fair point to have argued 50 years and $100 trillion of unfunded entitlements ago. But this horse is that long out of the barn. FDR in fact made this argument in 1940 when he vetoed Congress's converting Social Security from funded to paygo. But Congress overrode the veto. This argument was lost in politics then.

As a result, we now are on course to repeat #1 above on a humongously larger scale circa 2030. As in 1983, the lower the revenue-spending line going in, the lower (and better quality) will be the (increased) spending line coming out. And the higher the revenue line going in, the higher (and lower quality) the spending line coming out.

And that really dwarfs all other fiscal concerns at this point.

But really -- it is very easy to shoot at StB as policy, because it is very ugly in "business-like" terms, **when you don't consider the alternative**. But politics isn't business. And tax-hikers aren't ants. The only alternative is more grasshoppers given even more money to spend to make govt larger.

To claim that for people who want to *restrain* the size of govt limiting its revenue is the *inferior* policy, one must explain how the alternative of giving govt *more* revenue is the better way to *reduce* its size. I think it's a lot easier to argue with a Straw Milton than to do that.

Brian writes:

"Tax cutters are spending exploders. Tax cuts are not exogenous: The kind of politicians who cut taxes are usually the kind of politicians who boost spending."


I don't think this point is correct. Let's devise a test. Tax cutting reduces government tax revenue, at least in the short run. That's the idea behind starve the beast (STB). You claim that STB politicians increase spending more, despite lower revenue.

We can test this by looking at average annual federal spending increases relative to revenue. Your grasshopper/ant theory says that ratio will be huge for grasshoppers and small for ants. In contrast, the STB theory says the ratio won't vary much because lower revenues go with lower spending increases (held down by the threat of exploding deficits/debt) and higher revenues with higher spending increases (feed the beast-FTB).

Let's see what actually happens based on government revenue data ( We can average the spending increase to revenue ratio over the fiscal years of each presidency. These are the results.


76 - 80_____Carter_____________14.1%
81 - 88_____Reagan_____________9.3%
89 - 92_____Bush 1_____________8.0%
93 - 00_____Clinton_____________3.6%
01 - 08_____Bush 2_____________7.2%
09 - 12*____Obama_____________9.4%

These data agree that Clinton looks like an ant ("the era of big government is over") but says Carter (70%+ maximum marginal tax rate) looks like a grasshopper. It doesn't fit your theory. Moreover, Reagan, the Bushes, and Obama are nearly indistinguishable, with Bush 2 being the most antlike of the bunch (contrary to your claim).

The small variation in the ratio supports the STB theory. This is especially true if we consider that Carter's high ratio may have been driven by high inflation and Clinton's aided by low inflation. In addition, Reagan, Bush 2, and Obama had to contend with recessions, which drive up the ratio, while Clinton did not. Taking all that into account, it's hard to see any ant/grasshopper activity at all. It looks, instead, like STB/FTB holds sway.

Comments for this entry have been closed
Return to top