Polling, Progressive Taxation and Bloggers

Suzy Khimm, Kevin Drum and Greg Sargent comment on the series of gallup polls dating back to 1992 in which a solid majority of US adults, ranging from 55% to 77%, say that they think the rich pay less than their fair share of taxes. Please click the link to Drum then come back. I don’t want to excerpt. Oddly the fact stressed by Drum (and the only fact noted by Sargent) is the modest decline in this fraction from 77% to 62%since 1992 not the fact that a major party is totally dedicated to reducing taxes which a solid majority think are too low already. Khim is just describing the data. Drum has a long history of warning liberals not to trust the polls which seem to show our countrypeople are liberal (search for “polling literalism”) and a long record of not knowing about, then downplaying the exact polls under discussion.

The change from 1992 to the present came in two or three phases. The third possible phase (stressed by Drum) was an increase from 55% in 2010 to 62% in 2012. I’m not convinced that anything really changed. I think the 2010 sample just happened to be unusual, since other polls from around then show support of 60% or more for higher taxes on the rich. I won’t discuss this further. If we ignore the 2010 poll, we see a decline from 68% in 1994 to 62% in 2012. The null that this decline is entirely due to sampling error is rejected at the 5% level, the standard deviation of the difference due to sampling error is on the order of [root(2*(1/3)*(2/3)/1000)]100% or about 2.5%. But a change (or difference) of 8% is not usually considered worthy of much discussion. There was also a statistically significant change from 77% in 1992 to 68% in 1994. In this case, I do not suspect the 1992 sample as the extreme number is confirmed by something Brad DeLong said in 1994 about internal Clinton Gore 1992 polling.

I have three problems with Drum’s analysis. First he keeps arguing that US public support for more progressive taxes is low or not high enough or not as high as it was. This is of limited interest to readers. Projecting, I think he is being stubborn. Also, he seems to equate “not high enough to get the job done” with “not high” ignoring the huge gap between public and elite opinion on the issue (and the gap between public opinion and elite opinion about public opinion). Among the elite, it is widely agreed that lower tax rates are better and rates much higher than Clinton era rates are unimaginable. The fact which makes me lose my temper (which I am trying to keep now) is that the elite ascribes this view to the general public, asserting without evidence that Americans oppose taxation much more progressive than current taxation considering it class warfare.

Another more useful complaint is to note the tendency to look at trends and ignore levels. 62% support for something is lower, but it is not low. This temptation is understandable as trends are easy to see without defining units. But it isn’t optimal.

Finally, there is a tendency to assume that things were normal back at the beginning of the time series. OK this is the second complaint re warmed, but in 1992 US public opinion about whether the rich were paying their fair share was not ordinary at all. That level is not discussed by any of the three bloggers. The discussion of the difference between the current level and that level is discussed only in terms of factors which increase the current level. This is an aspect of the general tendency to tread trends as if they were levels, but 1992 opinion on the question was extraordinary.

Again I trust my memory of what Brad DeLong told me 18 years about about internal polls which were 2 years old at that time. He claimed that Clinton’s pollster got so tired of solid majorities answering yes when asked if rich peoples’ taxes should be raised in order to reduce the deficit, spend more on education, spend more on this that or the other thing that, finally, he asked if people thought rich peoples’ taxes should be raised to finance more waste fraud and abuse – and a plurality answered yes.

Also I absolutely trust my memory of the discussion of focus group reaction to a debate between Bush and Clinton. They had Bush supporters, Clinton supporters and undecided voters watching with those dial things which allow people to indicate agreement or disagreement. When Clinton said something to the effect that “only the rich have gotten tax cuts” self identified Bush supporters (on average) dialed in agreement. The TV commentators noted that this was very rare. I’ve never heard of such a thing.

In 1992, a solid majority US public was very angry that the rich weren’t paying what that majority considered considered to be a fair share of taxes. A known Philanderer from Arkansas was elected President replacing the man who lead the US and more allies than the US had ever had before to victory. That burning desire for some class war was very nice, but we can perfectly well make do with rather less.

I think the key thing is to deliver on the promise of tax cuts for the non rich as Obama did and Clinton didn’t. Of course, the vast majority don’t know that Obama kept his promise, but anger over his failure to keep the promise is much diminished by the fact that he kept it. Commentators couldn’t stop talking about how Clinton hadn’t kept his word. They don’t want to display ignorance by claiming that Obama didn’t, so they avoid the topic. This could be enough.