Does Mitt Romney Have a Wealth Problem?

The debate among the Republican candidates over Mitt Romney’s time at Bain Capital has raised again questions about whether Romney’s tenure in the “1%” will damage his campaign.  The Obama team certainly welcomes this debate.  After all, they have been attacking Romney along precisely these lines:

The day after Mr. Romney squeezed out a razor-thin victory in the Iowa caucuses, Mr. Obama’s political brain-trust trained most of its fire on him, painting him as both a Wall Street 1 percent type and an unprincipled flip-flopper.

Some new survey data that Lynn Vavreck and I have gathered in collaboration with YouGov suggests that Romney is vulnerable along precisely these lines.  The survey was conducted nationwide from January 7-10—right about the time that the Republican attacks on Romney’s “vulture capitalism” were crescendoing as the New Hampshire primary approached.

In this poll, we asked respondents:

How well do you think each the following describes Barack Obama/Mitt Romney: very well, somewhat well, not very well, or not well at all?

  • Personally wealthy

  • Cares about people like me

  • Cares about the poor

  • Cares about the middle class

  • Cares about the rich


(The order of the candidates and the order of the descriptions were randomized for each respondent.)

Consider first the responses to “personally wealthy.”


Both candidates are obviously wealthy, and the public tends to see them as such.  Very few people say that “personally wealthy” describes Obama or Romney not very well or not well at all.  That said, the public thinks this description applies much more to Romney than Obama.  Almost three-fourths (72%) say that “personally wealthy” describes Romney “very well.”  Only 45% say that about Obama.

But perhaps personal wealth doesn’t imply anything about who the candidates “care about.”  Unfortunately for Romney, this proves not to be the case:

The graph reports the percent saying “somewhat well” or “very well.”  Here Obama has an advantage on three of these items: “people like me,” the poor, and the middle class.  By contrast, Romney is perceived by more respondents as sympathetic to the wealthy than to any other group: the vast majority (89%) say that “cares about the wealthy” describes him at least somewhat well (in fact, 50% say it describes him “very well”).  But only 55% express similar sentiments about Obama.

Given that political independents are a particular target in presidential campaigns, do their opinions mirror the public’s?  In general, yes, although the contrast between the candidates is not as stark.  Consider “personally wealthy”: 68% of independents think this describes Romney very well; 53% feel similarly about Obama.   Similarly, 84% of independents think that “cares about the wealthy” describes Romney somewhat or very well; 64% say this of Obama.  One silver lining for Romney is that the gap for “cares about the middle class” is small: 41% think this describes Romney somewhat or very well; 44% say this of Obama.

But the data get worse for Romney in another sense: how these indicators are correlated with each other.  The better one thinks “personally wealthy” describes Romney, the better one thinks that “cares about the wealthy” describes him (the correlation is 0.60).  But the same correlation for Obama is much smaller (0.18).  People’s perception that Obama is personally wealth does not translate as strongly into the perception that he cares about the wealthy.

Moreover, people who perceive that Obama cares about the wealthy are actually a bit MORE likely to perceive that he cares about “people like me,” the poor, and the middle class.  The correlations are not always large, but they are positive—e.g., the correlation between believing Obama cares about the wealthy and cares about “people like me” is 0.19.

By contrast, people who perceive that Romney cares about the wealthy are LESS likely to think that he cares about “people like me,” the poor, or the middle class.  For example, the correlation between believing Romney cares about the wealthy and cares about “people like me” is -0.14 vs. the aforementioned +0.19 for Obama.  (All of the correlations I am reporting are statistically significant, by the way.)

So here is the problem that Romney confronts.  Americans perceive him as personally wealthy more than they do Obama.  They perceive him as caring more about the wealthy, but less about “people like me” and the middle class, than does Obama.  Moreover, Obama can “get away with” being perceived as personally wealthy or caring about the wealthy in ways that Romney cannot.  For Americans, Romney’s personal wealth is more intimately tied to the perception that he cares about the wealthy—and this in turn implies that he cares less for the middle class.


(Cross-posted at Model Politics.)

8 Responses to Does Mitt Romney Have a Wealth Problem?

  1. Luis January 13, 2012 at 1:46 am #

    This may sound like sort of a dumb/naive question, but why is this kind of timely, issue-driven research being left to academics? This seems like exactly the kind of question that an even vaguely thoughtful/curious media should be asking themselves. Is polling too expensive? Or are they really not that bright?

    • James W. January 13, 2012 at 2:27 am #

      The media are in business to tell “middle class americans” what to think, not to find out or report on what they think.

  2. Bruce Bartlett January 13, 2012 at 11:29 am #

    Media organizations hate citing any poll data other than their own. This forces them to expend vast resources asking the same identical questions that every other organization asks so that they have time series data. This leaves little money for new or interesting questions. Also, pollsters tend to be thick-headed in my experience.

  3. Larry Bartels January 13, 2012 at 11:46 am #

    How do these perceptions relate to vote intentions? I would guess that Romney’s “wealth problem” mostly matters through its effect on “cares about people like me” (or “out of touch” or some such).

    I wrote a piece many years ago on the impact of candidate traits on presidential election outcomes. (It appeared in a volume edited by Anthony King on _Leaders’ Personalities and the Outcomes of Democratic Elections_.) After some somersaults to allow for the fact that trait ratings are strongly colored by partisanship and current economic conditions, the effects looked fairly small. However, the effects for “really cares about people like me” were pretty consistently larger than those for other traits tapped in NES surveys, including “moral,” “knowledgeable,” “inspiring,” and “provides strong leadership.”

    Incidentally, the current 10-point gap between Obama and Romney on “cares about people like me” is not out of line with typical partisan differences. In the elections I looked at (1984-2000), every Democratic candidate had a more favorable image than his Republican opponent on this dimension. The gaps ranged from 6 or 7 points (for Clinton over Dole in ’96) to 18 points (for Clinton over Bush in ’92). Of course, this year’s gap may widen over the course of the campaign, especially if Romney keeps going on about how much he likes to be able to fire people.

    • Karen Jusko January 13, 2012 at 4:17 pm #

      Larry, John’s post reminded me of your “Homer Gets a Tax Cut” chapter: Romney may not have a wealth problem if “people like me” is understood to be “those with an expectation of being wealthy.” How do these perceptions vary across self-reported social status groups?

  4. Aaron January 13, 2012 at 5:56 pm #

    Is there any effect from question order? Specifically, I’m wondering whether people who were first asked whether Obama and Romney were personally wealthy were less likely to then say Romney and Obama cared about people like them.

  5. Sam Wang January 14, 2012 at 5:08 am #

    The simple idea of calculating correlations is indeed a slight mind-blower, and does raise the question of why it’s not done more often. Probably because most people do not understand correlations. The usual cross-tabs employed by most pollsters are coarse but more understandable to the statistically naive.

    One source of interesting information, though not locked closely to the news cycle, is the Pew Center.

    Translation to voter intention is obviously important here. At the risk of making statements obvious to readers here, what experiment could translate correlations into understanding the direction of causality? The implication here is for Democratic messaging: (1) tell voters Romney is very rich, (2) this will make them think he only cares about the rich, and (3) their intentions will shift.

    In regard to Bruce Bartlett’s point, unwillingness of one media organization to cite another one’s polls has created an opportunity for poll aggregators.

  6. Carl January 23, 2012 at 7:18 pm #

    Mitt Romney for president?