Jonathan Bernstein is a political scientist. He blogs at A plain blog about politics.

This is a follow-up to something I wrote yesterday about presidents and foreign policy successes, and in particular to some great comments that I recommend to everyone.  After political scientists Richard Skinner and Matt Jarvis added to the point I made that foreign policy achievements don't really help presidents get re-elected, commenter Colby said:

Seems like a good, general rule is that outside of the economy, your policy initiatives are only marginal help if they work, and a giant albatross if they don't. Remember what Homer Simpsons said: trying is the first step toward failure. 

I think that's mostly correct.  Think about health care reform.  Had the bill crashed and burned, the general consensus (with which I agreed) was that Obama and the Democrats would suffer from the reputation of being unable to govern.  Once the bill passed, Obama and the Democrats were at risk of being held responsible for anything that goes wrong in the broad field of health care, whether or not the ACA was responsible.  Indeed, I'm surprised at how little of this has been done so far this summer; Republicans as far as I've seen have mostly continued to fight the same battles they've fought all along, instead of blaming Democrats for every doctor shortage, insurance premium increase, and nasty hangnail, which is what I expected them to do.  Be sure, however, that if health care reform winds up not working that it will hurt Obama and the Democrats.  And if ACA works?  Democrats won't get much credit at the ballot box, not this year, not in 2012, not in 2016 when it's fully implemented.  The debate will move on to how to improve it, and everyone (including voters) will treat the exchanges and the rest of it as just part of the normal furniture, not as a reason to switch one's vote.  Of course, if a policy really works well, then whatever the problem it addressed drops right off the radar screen. 

This brings me to a question: if policy innovation entails major risks for presidents and few electoral rewards, why do they bother?  Why not just restrict yourself to talking about what a great idea school uniforms are, and call it a day?

Of course, they may do it because they believe it's the right thing to do; because they care about their place in history; or because they haven't read the political science literature (and blogs) and mistakenly believe that there are re-election advantages to policy accomplishments.  I don't fully discount those things, but let's just say I'm not convinced that they fully explain the level of presidential involvement in policy initiatives.  So, why do they do it?

One reason, and perhaps the most important one on yesterday's topic of foreign policy and national security, is to avoid policy disasters.  No, Barack Obama will not win any votes if he manages to "win" in Afghanistan, whatever that actually means.  But he'll lose plenty of votes if Americans continue to die there in ever-increasing numbers.  He'll also lose votes if Americans leave, the Taliban takes over and shelters bin Laden, and that allows more devastating terror attacks (followed by another invasion, followed by more US casualties).  Turning closer to far as I know, the administration did an absolutely terrific job of responding to the swine flu last year (although Obama's response to the Fried Chicken Flu left a lot to be desired).  So, who is voting for the Democrats this year or generally supporting Obama because he or she didn't die horribly in the prevented horrible pandemic of 2009-2010?  Well, nobody; no one even remembers last year's flu scare, (perhaps) because successful government action turned it into a non-story.  But if things had gone wrong, people would have blamed Obama and, beyond that, people would have been more unhappy about everything, which tends to hurt incumbents.

Beyond avoiding disasters, however, I think the big answer here is about representation.  If we think of representation as a process, it involves politicians who campaign by making promises, certainly including policy promises.  If you've never done it, I recommend going through PolitiFact's database of the over 500 specific policy promises that Barack Obama made while campaigning in 2007 and 2008.  Campaigning requires making promises...perhaps the better way to say it is that campaigning is making promises.  Not all promises are policy; as I've talked about before, sometimes politicians promise explicitly or implicitly to act in certain ways, to look out for allied groups; a pol can even promise to be a specific person, which implies all sorts of behaviors. 

The need for policy promises, however, is particularly strong in high-profile nomination contests, in which candidates compete for the support of party-aligned groups and even individual voters by making policy commitments.  To run for president as a Democrat in 2008 and have any chance of winning, a candidate needed to pledge to attempt health care reform.  A candidate needed have a plan to fight climate change.  A candidate needed to support card check and other issues that unions cared about.  A candidate needed to oppose the war in Iraq and pledge to end American involvement there.  We talk a lot about how issues don't really matter much in elections, but usually we're talking about general elections, when party voting dominates.  That's not the case in primaries, where voters -- without the help of party cues -- rely on things such as issues, and endorsements by organizations and pols who care about issues. 

Once those politicians are elected, then, they tend to keep their promises.  Partially that's just because they've already set things in motion to keep those promises (for example by hiring staff dedicated to doing so).  Partially it's because pols fear being branded as flip-floppers.  Partially because they don't want to create enemies, and allies scorned can be powerful enemies.  And, related to that, partially because they want to be renominated -- and they want to be renominated by acclamation, without going through the kinds of struggles that Jimmy Carter had in 1980 (when Ted Kennedy almost defeated him for the nomination) or even the annoying problems George H.W. Bush had in 1992 (when Pat Buchanan forced Bush to actively campaign).

Of course, sometimes they do flip, usually because the costs of following through turn out to be higher than the costs of flipping.  And, of course, a lot of times presidents don't actually get done what they promised to get done because they don't have the votes (such as climate change), or because they've promised something that they don't have the capacity of delivering (such as peace between Israel and the Palestinians), or because they've traded it for something else they've promised (such as Obama's broken promise on drug re-importation, which bought him votes of marginal Dems for the overall ACA).  Moreover, there are various degrees of keeping promises...while I don't think he's broken a promise on DADT, Obama has put a whole lot more effort into, say, health care reform than he has into DADT repeal.  And it's sometimes really difficult for outsiders to tell which things are which: have Obama's actions on Gitmo been a flip?  Relative indifference?  Genuine attempt met with failure?  Part of an ongoing fight that may still lead to fulfilling a promise?

The main point, however, is that for presidents, keeping promises involves policy activism.   That's the main reason they do it.  It isn't really to give them something to run on in the next campaign (pols can always find something to run on, regardless of whether they have any real accomplishments or not), and so discovering that general election voters don't care about foreign policy in most cases, for example, should not lead us to believe that presidents shouldn't or won't care about foreign policy.  Presidents get nominated by making promises, and once in office they're more or less trapped into trying to keep them.