Peter Feaver advances “a novel hypothesis”:http://shadow.foreignpolicy.com/posts/2010/03/23/has_obama_lost_his_silver_tongue.
bq. This pattern of comparatively low scores for the handling of signature domestic policy priorities and higher scores for the handling of foreign policy may be due to several factors. Perhaps the public just disapproves of Obama’s health and economic policies and approves of the national security policies. Or perhaps the public approves of the way Obama has pursued more of a bipartisan policy on national security than he has on health care, which passed on a pure partisan basis. Note that the Republicans, who were quite loud in shouting “No” on Obamacare have been the loudest “Yes” voices on Afghanistan. Perhaps the low numbers are just the direct result of all of the partisan shouting. Or perhaps Obama’s numbers on domestic policy are contaminated by the public’s total disdain for Congress, which has approval numbers in the low teens. Perhaps the public and the media have been so focused on health care that neither has not paid much attention to the wars and if they did they might not like what they see there. Perhaps the president is still benefiting from a commander-in-chief halo.
bq. If I were in the White House, however, I would be concerned about yet another possible explanation: perhaps the more the president talks about an issue the more he drives his own numbers on that issue down. I would worry about that because as a national security policy person, I do not want the president’s political advisors to have a perverse incentive to avoid talking about the war. There are other costs, not directly measured in public opinion polls, when a president avoids the national security issue. … I suspect that in the coming months that commitment will be tested by developments on the ground in [Iraq and Afghanistan] that cannot be ignored, not by the media, not by the public, and thus not by the president. It would be paradoxical and problematic for the Obama Team if they discovered that having the president speak to those issues seemed to undermine public support for them, at least on the margins.
Peter Feaver is a reasonably well-known political scientist, working in the political science department at Duke University. He is also someone of recognizable partisan inclinations, having served for a spell in the Bush White House. This certainly isn’t a problem as such – I have recognizable partisan inclinations (which I try not to indulge on this blog) myself. But it is a problem if it interferes with purportedly political scientific analysis, or, worse, if it becomes a substitute for such analysis. And … well … how can I put this best … Either Feaver has identified an important new effect, which overturns the existing political science consensus that Presidential rhetoric has “no significant consequences”:http://www.amazon.com/Deaf-Ears-Limits-Bully-Pulpit/dp/0300115814 for public opinion. Or he is allowing his personal druthers and biases – Obama has a reverse Midas Touch! Everything he touches turns to dreck! – to substitute for actual analysis.
Given the dubious evidence and complete lack of any even halfway plausible general causal mechanism underlying Feaver’s claim, I’m inclined to favor the latter rather than the former hypothesis. Perhaps he’ll prove me wrong with a bang-up article for the _American Political Science Review._ What is most remarkable, perhaps, is Feaver’s failure to consider selection effects. A _much_ more plausible account would reverse the causality – it is not that Obama’s speeches make political topics more controversial, but that more controversial topics are more likely to be the subject of presidential speeches. This account would have an obvious causal mechanism – presidents are more likely to want to devote their speeches to topics that they feel they need to justify their position on, perhaps to elites than to waste their time talking about non-controversial issues. This would take account of the obvious fact that the President chooses the topics he (or, one day, she) speaks on – and that the choice of topic is very unlikely indeed to be uncorrelated with the degree of controversy surrounding that topic.[1] It wouldn’t have any obvious partisan implications, nor be as ‘counter-intuitive’ – but it strikes me as much more likely to be good political science.
fn1. To put it another way, if there is a strong argument and/or evidence that the President’s choice of speaking topic _is_ uncorrelated with the degree of controversy of that topic, I’d love to see it.