Milton Friedman on Donahue 1979 (2/5)

Do Tax Cuts “Starve the Beast”?

January 7, 2013 by Dan Mitchell

There’s a debate among policy wonks about whether a no-tax-hike policy is an effective way of restraining the burden of government spending.

At the risk of over-simplifying, the folks who support the “starve the beast” theory argue that there are political and/or economic limits to government borrowing, so if you don’t let politicians tax more, you indirectly impose a cap on total spending (outlays = tax revenue + borrowing limit). We’ll call this the STB approach, for obvious reasons.

Critics of the theory, by contrast, say that a low-tax policy creates fiscal illusion by making government spending seem artificially cheap. After all, standard microeconomic analysis tells us that people will demand more of something when the perceived price is low (get a $1 of spending for 80 cents of tax = recipe for higher outlays). We’ll call this the “pay for government” approach, or PFG.

There’s almost surely some truth to both arguments, but the real issue if whether one effect is dominant – particularly in the long run. In other words, should supporters of small government fight tax increases? Or welcome them?

I’ve never studied this issue, but my gut instinct has been on the “STB” side of the debate. Here are a few of the reasons.

1. The politicians and interest groups that favor bigger government seem especially anxious to convince anti-tax lawmakers to change their minds. If nothing else, that suggests higher taxes would “feed the beast.” I suppose this could be a clever example of reverse psychology, but something tells me that Harry Reid and Nancy Pelosi lack the cleverness and subtlety to pull off that kind of trick.
2. The people who pay for government generally aren’t the ones who reap the benefits. And if you keep increasing taxes on the “rich,” as Obama proposes, why would that affect the preferences of the rest of the population? Especially the huge chunk of the population that doesn’t pay income tax? Simply stated, the PFG approach incorrectly assumes that payers and payees are the same.
3. Casual empiricism certainly suggests that higher taxes are associated with more government, not less red ink. We see this, for instance, in the evidence I recently shared from Europe. Taxes have jumped in recent decades, but government debt also has climbed, which implies all additional revenue was spent, and then some.
4. Just look at the real world, specifically the fiscal crisis in nations such as Greece. At the risk of stating the obvious, the recent events in Europe confirm that there does come a point when governments lose the ability to borrow. So if taxpayers somehow can prevent politicians from seizing more money, there is a de facto limit on government spending.

Seems like the STB approach makes sense, but not everyone thinks my theoretical musings and generic observations are all that’s needed to settle an argument.

Particularly when there are some very sensible people on the other side. The late Bill Niskanen wrote in the 2006 Cato Journal that:

There are three major problems with the starve-the-beast argument: (1) it is not a plausible economic theory; (2) it is inconsistent with the facts; and (3) it has diverted attention away from the political reforms needed to limit government growth.

I fully agree with Bill that there should be much more focus on restraining the growth of government, so there’s no disagreement on his third point. I think he’s wrong on the first point because half the population no longer pays federal income tax and the top 20 percent pay the lion’s share, but that’s a bit of a judgment call.

What about the facts? Bill does some regression analysis for the 1949-2005 period, where he looks at the change in federal spending as a share of GDP and tests its relationship with the level of tax receipts as a share of GDP, the change in the unemployment rate, and the change in interest payments (the latter two variables are there to hopefully wash out the effects of the business cycle and to limit the analysis to the spending that lawmakers actually can control).

Bill crunches the numbers and concludes:

For no extended period did these estimates reveal a significant positive relation between the change in federal spending as a percent of GDP and the level of federal receipts as a percent of GDP, the necessary condition for the starve-the-beast hypothesis to be confirmed.

Moreover, Bill even found evidence for the PFG approach when he looked solely at the 1981-2005 period.

A 1 percentage point increase in current federal receipts as a share of GDP apparently reduces the change in current federal spending as a share of GDP by about one-seventh of 1 percent a year indefinitely.

I don’t doubt that Bill’s numbers are sound. Indeed, Cato Adjunct Scholar Michael New re-crunched the numbers for the Cato Journal in 2009 and produced similar findings, even when looking only at non-defense discretionary spending.

But I don’t find this research very compelling, and it’s not just because I’m from Austrian school, which sometimes has a reputation for being skeptical about empirical analysis.

Here are some reasons why I’m not convinced, and even the biggest quant jocks in the world should share these concerns.

1. Is 57 years of data (1949-2005) or 25 years of data (1981-2005) really enough to draw any sweeping conclusions, particularly when there could be many other factors involved? We would be very reluctant to jump to conclusions about the demand for Big Macs by interviewing a handful of customers and looking at just three variables.
2. More important, why didn’t Bill measure changes in spending against legislated tax changes? After all, lawmakers rarely pay attention to tax receipts as a share of GDP, and that variable rarely if ever is part of the lawmaking process. But politicians are acutely aware of whether they are voting to either reduce taxes or increase them.
3. And why use spending as a share of GDP rather than nominal spending or inflation-adjusted spending, particularly since Congress votes to spend specific amounts of money, not for outlays as a percent of economic output.
4. Equally perplexing, why didn’t Bill include lags in his research? I’m not aware of any STB proponents who claim that there’s an instantaneous impact. Instead, they argue that long-term limits on revenue can impose long-run restraints on spending.

To be fair, Bill was breaking some new ground. There was not a lot of empirical analysis to that point, so there was no right or wrong way to test the relationship between taxing and spending. Niskanen picked one approach, and it’s the role of subsequent researchers to poke and prod the results and contemplate alternatives.

That’s exactly what Christina Romer and David Romer did in their article that appeared in the 2009 Brookings Papers on Economic Activity. They investigated the data from several angles and decided it made the most sense to look at legislated tax changes and look at the long-run impact on spending. And, in an attempt to test the STB hypothesis, they looked solely at major tax bills designed to reduce government revenue.

That’s the good news. The bad news is that this gave them only four pieces of data – the Revenue Act of 1948, the Kennedy tax cuts, the Reagan tax cuts, and the 2001/2003 Bush tax cuts.

Setting aside this problem of limited data, what did Romer and Romer discover? Their headline results were similar to Niskanen’s.

The results provide no support for the hypothesis that tax cuts restrain government spending.

That sounds like bad news for STB advocates. But if you dig into their findings, you find out that the real problem is that politicians can’t resist the temptation to feed the beast.

…roughly three-quarters of a long-run tax cut is typically undone by legislated tax increases of various sorts within five years. …The fact that policymakers have been able to largely reverse tax cuts helps to explain why the cuts have not reduced spending.

In other words, you can’t starve the beast if you don’t maintain the diet.

Which is basically what other economists concluded when analyzing the work of Romer and Romer. Here’s what Steven Davis of the University of Chicago wrote.

…if it takes 5 years for a new policymaker to reverse a previous tax cut, so that it remains in effect for 10 years rather than 5, the starve-thebeast effect roughly doubles. In the extreme case where tax cuts cannot be reversed, government spending cuts must eventually absorb the entire adjustment. Clearly, then, tax cuts can produce large starve-the-beast effects if they are sufficiently sticky.

And Jeffrey Miron of Harvard University had a similar interpretation.

…concerns over letting children play with matches—that is, giving politicians access to increased tax revenue—are valid. Thus, advocates of small government would seem to have good reason to oppose tax increases.

All things considered, I think that STB is correct.

But I’ll close by returning to one of Bill Niskanen’s points. He warned that the focus on tax limitation was harmful because it “diverted attention away from the political reforms needed to limit government growth.”

I fully agree. Too many politicians focus on the easy – and more politically popular – job of fighting tax increases. But then they fail to support measures to restrain the burden of government spending.

Or, as we saw during the Bush years, they cut taxes and then opened the spigot on the spending side of the fiscal equation. No wonder Romer and Romer found that tax cuts generally are reversed. Tax cuts are difficult to maintain and preserve if they are simply gimmicks put in place by feckless politicians.

P.S. Another interesting tidbit is that Romer and Romer acknowledge the Laffer Curve.

We also find that the overall rebound in revenue exceeds the portion due to legislated changes. The key source of the nonlegislated change in revenue is almost certainly the effect of the tax cut on economic activity.

Too bad Christina Romer didn’t share that insight with the President when she was at the Council of Economic Advisers.