This observation in The Economist caught my eye:
Contrary to popular belief, there is some evidence for the idea that Americans might quite like some more government. Another consistent finding in political science is that voters are ideological conservatives but operational liberals. Small government is more popular than big government in theory, but voters do not like spending cuts. “The public mostly agrees with the Republicans in philosophical terms and with the Democrats in policy terms,” write David Hopkins and Matt Grossmann in “Asymmetric Politics”, the best recent book about how the two parties became what they are.
This reminds me of something Eliezer Yudkowsky talks about in his book entitled Inadequate Equilibria. He suggests that people often have two views on an issue, the view they express when they think directly about the issue, and the view they express when they take into account their likely biases. Thus my “inside view” might be that I’m an above average driver. But I also know that the vast majority of people regard themselves as above average drivers. So my outside view is that I’m probably just an average driver. The outside view is from 64,000 feet up, where I look just like any other ordinary person. What makes me think I’m a better than average driver, if most other people also have that perception?
When I went to the University of Chicago, I learned that lots of government policies that sound good actually have hidden downsides due to the “unseen” effects, to use Bastiat’s term. While most people have not studied economics at the University of Chicago, I believe that lots of people of my (boomer) generation have this sort of general view:
Back in 1965, LBJ’s “War on Poverty” seemed like a great idea. But after 50 years it doesn’t seem to have solved the problems in the inner city, or in rural West Virginia, or on Indian reservations in South Dakota. Somehow these government programs are not as effective as they sound. Government is a relatively ineffective instrument for solving complex problems.
At the same time, these people have not studied economics at Chicago, and therefore have a hard time articulating the specific ways that various incentive effects might undermine the effectiveness of a given progressive policy. Thus when asked about any specific program that sounds good on utilitarian grounds, such as doubling the minimum wage or guaranteed health care for everyone, they have trouble saying no (inside view). But when asked more generally about big government, they are skeptical (outside view). Their outside view reflects an tacit assumption that their inside view is likely to be too optimistic, just as in 1965.
Most importantly, I think it’s a big mistake (for either party) to assume that either view is the true opinion of the public. There is no such thing as “actual” public opinion, there is only the way that people vote in elections, whether for candidates or on referendum issues. That’s the only reality that counts. Public opinion depends on how poll questions are framed.
For Democrats, the challenge is to get voters thinking in inside view terms, and for the GOP it’s getting them to think in outside view terms. (Or at least it was when the GOP still saw itself as the small government party. Now things are less clear.)
PS. Speaking of poll results, voters were asked what fraction of Republicans make over $250,000/year. The average person estimated that 38.2% of Republican made more than that amount. The study says the actual figure is 2.2% (I suspect it’s more like 3%.) Think about that disparity for a moment. Either Americans know shockingly little about the country they live in, or there’s some sort of weird innumeracy that makes Americans unable to articulate their perceptions using actual numbers. Don’t get me wrong, I expected the public’s estimate to be higher than the actual, but that gap is just bizarre.
Now assume that Americans are asked a poll question about the “distribution of income”. Given their absurdly inaccurate estimate of how much other people make, should we take those poll results seriously? What happens to the welfare of someone making $80,000, if we make incomes more equal? It depends on whether you wrongly think the average income is way above $80,000, or correctly assume it’s below $80,000.
PPS. A particularly amusing example of trying to pin down “public opinion” is with polls on immigration. This Gallup poll shows Americans are amazing pro-immigration on all sorts of questions:
A record-high 75% of Americans, including majorities of all party groups, think immigration is a good thing for the U.S. — up slightly from 71% last year. Just 19% of the public considers immigration a bad thing.
And an anti-immigration spokesman counters with this factoid:
Asked an open-ended question about how many immigrants the U.S. should admit each year, 81% gave a figure that is less than the approximately one million we currently admit each year.
If Americans think 38% of Republicans earn over $250,000, God help us trying to figure out how Americans interpret “1 million immigrants”. What if the question is rephrased as “What percent of the US population should be admitted each year?” (Roughly one third of one percent is the current flow.)
READER COMMENTS
Alan Goldhammer
Jul 29 2018 at 6:40pm
Scott, thanks for the notice about ‘Asymmetric Politics.’ I was one course short of a double major in chemistry and political science as an undergrad and still read a lot of contemporary poli sci works (my favorite of the last couple of years is Achen and Bartels, ‘Democracy for Realists: Why Elections Do Not Produce Responsive Government’). One of the courses I took dealt with public opinion and we did some field experiments in polling (this was in 1967) and how one frames and presents the question is critical. I’m not surprised at all by the two results you noted. In particular, probably just over 80% would not have a grasp of income ranges of Americans.
Benjamin Cole
Jul 29 2018 at 8:05pm
Yes, sadly, it appears innumeracy is the norm.
People are funny. They can be against “socialized medicine” yet for communist healthcare (the VA and military healthcare facilities).
The great unwashed and the elites seem to share these afflictions…
Thaomas
Jul 30 2018 at 8:26am
The miss-perception about average Republicans’ income is possibly excused by the by the tax polices their representatives in congress follow.
E. Harding
Jul 30 2018 at 10:54pm
Strong post, Sumner.
Jacob Egner
Jul 31 2018 at 12:52pm
I think we should embrace a lot of the caveats of interpreting poll results as explained by Scott Alexander. A good rule of thumb is that at least 4% of poll results are likely to be trolling/error. And far beyond 4% of responses are likely to be people using the poll to cheer on their allies and boo their adversaries.
And people are just sloppy. What incentives does a person have to sit down and really crunch numbers in their head for a poll? For instance, there was a poll that asked “What percent of Democrats are gay?” and the responses were absurdly high. There might have been trolling and cheering/booing, but I imagine a lot of people had mentally morphed the question to roughly “What percent of gays are Democrats?”.
My suggestion is: nooooo.
Dr. Scott Sumner: If you’ve read Scott Alexander’s article that I linked to, I’d love to hear a few words on what you thought of it. Thanks.
robc
Aug 1 2018 at 9:00am
If they paid the respondents based on the accuracy of their answers on the factual questions, it would get them to think before they answered.
That would be more of an economic experiment than a poll.
Robert EV
Aug 3 2018 at 10:23am
Eh. not necessarily.
I took part in ones of those “watch some TV ads, answer some questions, and get a couple bucks” surveys and completely blanked while watching the ad.
I’d guess that money is less important than time and indicated seriousness when getting people to try to guess right. And even then those on the sociopathic spectrum (which includes trolls and jerks) aren’t going to give a damn.
Comments are closed.