Do Americans Really Want to Cut the Deficit?

by John Sides on June 23, 2010 · 8 comments

in Public opinion

Andy notes this post by Ben Somberg, who is fussing at the Washington Post for this paragraph:

bq. If Congress doesn’t provide additional stimulus spending, economists inside and outside the administration warn that the nation risks a prolonged period of high unemployment or, more frightening, a descent back into recession. But a competing threat — the exploding federal budget deficit — seems to be resonating more powerfully in Congress and among voters.

Somberg objects to “among voters,” citing numerous polls showing that the public prioritizes the economy and jobs over the deficit. He is correct. You can peruse some data here. So the first point is: Americans are concerned about the deficit, but not _that_ concerned.

The second point: Americans don’t embrace spending cuts or higher taxes — the steps necessary to cut the deficit. This does _not_ mean that Americans simply want a free lunch. Several studies have found that, when confronted directly with trade-offs, very few Americans simply insist that we somehow cut taxes and increase spending simultanously. In this 1985 piece (gated), Susan Welch writes:

bq. We have found that support for expansion of public services is high. We have shown that the inconsistency of citizens wanting more for less, more services and less spending, probably is a paradox for only a minority. Most citizens are willing to raise additional revenue to pay for these services, or at least to reallocate from less desired to more desired services…Thus we have found that in the abstract at least, not as many Americans are looking for a “free lunch” as a simple comparison of taxing and spending preferences would suggest.

Similarly, in this 1998 piece (gated), John Mark Hansen examines preferences over taxes, spending, and deficits and finds that the public is remarkably consistent:

bq. The public has the ability to make budget policy choices with reasonable discernment. Allowed by survey questioners to be practical in their views, ordinary Americans are practical in their views. They have well-formed and well-behaved preferences. Ordinary citizens may not have preferences over spending, taxes, and deficits that experts consider “correct,” but their opinions are coherent.

So the problem is not inconsistency or incoherence. The problem is that there is no large majority of Americans who favors cutting the deficit over other alternatives. An April 2010 CBS/NYT poll asked “If you had to choose, would you prefer reducing the federal budget deficit or cutting taxes?” 45% said cut the deficit and 47% said cut taxes. This same poll also asked respodents to choose between two options:

bq. 1) The federal government should spend money to create jobs, even if it means increasing the budget deficit.

bq. 2) The federal government should not spend money to create jobs and should instead focus on reducing he budget deficit.

50% chose spending money on jobs even if it increases the deficit. 42% wanted to reduce the deficit.

It’s also important to remember that it is difficult to find majorities of the public willing to cut any particular category of spending.

A third and final point: Americans don’t punish presidents for large deficits. They punish them for bad economies.

Let’s assume for the moment that voters would punish them for deficits much like they punish them for bad economies. If so, then it would be changes in the deficit or debt in the year prior to the election that mattered because voters are myopic about these things. For the sake of illustration, I will rely on the percentage change from the previous year in the gross federal debt held by the public (data here).

If you plot the incumbent party’s percentage of the vote against the percentage change in the deficit, this is what you get:

There is only a modest negative relationship. Comparing the historical low (2000) to the historical high (1976), the incumbent party’s predicted vote share would decline by only about 5 points. This contrasts with the large and well-known effects of economic growth (data here):

If you regress the incumbent party’s vote share on changes in the debt and in disposable income, the debt has an statistically insignificant effect while income has a large, positive, and statistically significant effect.

None of this means that Obama or members of Congress won’t act to reduce the deficit. They are (sort of) trying to do that now. But, contra the Washington Post, let’s not fool ourselves about the public: they don’t prioritize deficit reduction and don’t punish politicians who increase the deficit.

{ 5 comments }

Joel June 23, 2010 at 10:40 am

regarding coherence: if A and B are logically incompatible, and you tell me to choose between them, my choice of one or the other does not mean i am being logically consistent, only that i can follow instructions.

i have trouble swallowing polls that are interpreted as suggesting that many people are willing to pay more taxes. aside from myself, i have never engaged anyone in conversation who would say such a thing (and i do ask – never have me over for dinner).

now, if we simply mean that most people want lower taxes, and don’t much care about what that does to the deficit, well sure. but that hardly speaks to coherence, it only speaks to how people want to pay lower taxes and don’t much care about anything else.

Eric June 23, 2010 at 11:06 am

Is there any good solid research that shows that politicians actually get punished for raising taxes? Maybe it’s different if they raise taxes on a small population as opposed to a larger population.

William Ockham June 24, 2010 at 10:50 am

I agree with the overall thrust of this post, but every time I see that bottom chart (the one that proves that economic performance matters), I’m struck by a couple of things. First, if you look at the years that are above the line and above 50%, they all have one thing in common: there was an incumbent on the ballot. Second, if you look at the two big below the line outliers (we’ll ignore 2000 for a moment), they are the two years with missing incumbents (i.e. Truman and Johnson both could have run for another turn, but didn’t). Those two facts lead me to believe that the formula doesn’t properly account for the positive power of personal incumbency. I say positive power because when you get to the region where the formula predicts the party in power will lose, personal incumbency doesn’t seem to help at all (it might actually hurt). Once you think of it that way, you start to see that Bush’s narrow victory in 2004 is really an outlier. According to the formula and adjusted for personal incumbency, Bush should have earned at least 2% more of the vote. That makes 1996, 2000, and 2004 significant. For a variety of other reasons, I think the predictive power of this formula ended in 1992 with the third-party candidacy of Perot. Of course, you poli ci types will never accept that because if 1996 and 2000 didn’t shake your faith in this formula, nothing will. Every conceivable outcome of presidential elections now lies within the acceptable level of variation. The slope of that line will never change because it’s determined by a few election results that are unlikely to be repeated.

John Sides June 24, 2010 at 2:14 pm

William: I think there is some effect of incumbency, above and beyond the economy. I don’t have the statistical evidence at my fingertips but I know that some scholars (e.g., James Campbell) make that argument.

As for 1952 and 1968, the usual explanation is that the public is punishing the incumbent party for casualties in war. One of the links in the post is to Doug Hibbs’ webpage, and he presents some evidence for this.

In general, I tend to emphasize the relationship between the economy and election outcomes simply because it often gets lost amidst media narratives. I certainly don’t think that the relationship is deterministic.

John Sides June 24, 2010 at 2:22 pm

William: I just did a quick little analysis and found that there is an effect of incumbency in addition to the economy. The size of the effect depends a bit on whether LBJ and Ford are considered incumbents in 1964 and 1976, respectively — i.e., whether incumbency implies that you were elected. But in either case, there is evidence that incumbents do about 3-4 points better than challengers, all else equal.

Comments on this entry are closed.

{ 3 trackbacks }

Previous post:

Next post: