In the context of writing about anonymous negative leaks emanating from Hillary Clinton's campaign, Brad DeLong posits a typology of staff:

There are two kinds of people who get involved in politics--those who care about the substance of policy, and those who want to get White House Mess privileges, or as a consolation prize become media celebrities. The first kind--the policy people--will be loyal to a politician as long as he or she is trying his or her best to achieve the shared policy goals. The second kind--the spinmasters--will be loyal to a politician as long as he or she is a winner who favors them. If a politician stops looking like a winner, or if a politician starts favoring others for what they hoped would be their west wing job, they will jump ship as fast as they can--and you will start seeing the "infighting" stories.

The moral? A politician with an ideological policy compass is best off not hiring spinmasters as his or her senior aides. Hire people who care about the substance of policy instead.

Let's formalize this.

Let win = p(candidate) = ex ante probability of a candidate winning an election

Let policy = 1/the square of (wonk's ideal policy point - candidate's preferred policy point) = the ideological proximity between candidate and staffer

A wonk will maximize his/her utility by maximizing win*policy.

A policy wonk will not rarely stab a candidate in the back during a campaign -- but their prior choice of who to back is, in fact, quite strategic.

[Does this really matter? Doesn't a presidential nominee simply consolidate the cream of the crop after the primary season?--ed. Based on the last two election cycles, my answer would be no. Kerry did not do this in 2004. As for 2008, let's outsource to Michael Hirsch:

They were devotees of the cult of Clinton. Greg Craig was Bill Clinton's lawyer, defending him on TV against impeachment charges. Susan Rice was a protégée of Madeleine Albright, the 42nd president's secretary of State. Anthony Lake was Clinton's personal foreign-policy consigliere, his first-term national-security adviser. Now, however, Craig, Rice and Lake are all top advisers to Hillary Clinton's main rival, Barack Obama. In an increasingly bitter fight for the best and brightest policy advisers of Clinton's presidency, these defectors are aggressively recruiting junior- and midlevel officials from his administration.

That's provoked anger in Hillary's camp—and, Obama aides charge, threats of retaliation if she wins the nomination. "The word some people are hearing is, 'You're making a big mistake. If Barack Obama wins, he'll welcome all into his administration. But if Hillary wins, that's not going to be the case'," says a midlevel Obama adviser who would speak about infighting only on the condition of anonymity. The main Hillary enforcer has been Richard Holbrooke, Bill Clinton's former U.N. ambassador (and an aspiring secretary of State), says a senior adviser to Obama, who would also discuss personnel matters only anonymously. "I have had at least two people directly tell me they had been told by Holbrooke if they went with Obama, the Hillary people would not forget and forgive," says the adviser, refusing to identify them.]

In some respects, the sensibility behind the behaviorist critique of economics is one shared by all the Obama wonks, whether they're domestic policy nerds or grizzled foreign policy hands. Despite Obama's reputation for grandiose rhetoric and utopian hope-mongering, the Obamanauts aren't radicals--far from it. They're pragmatists--people who, when an existing paradigm clashes with reality, opt to tweak that paradigm rather than replace it wholesale. As [Richard] Thaler puts it, "Physics with friction is not as beautiful. But you need it to get rockets off the ground." It might as well be the motto for Obama's entire policy shop.

The Clintonites were moderates, but they were also ideological. They explicitly rejected the liberalism of the 1970s and '80s. The Obamanauts are decidedly non-ideological. They occasionally reach out to progressive think tanks like the Economic Policy Institute, but they also come from a world-- academic economics--whose inhabitants generally lean right. (And economists at the University of Chicago lean righter than most.) As a result, they tend to be just as comfortable with ideological diversity as the candidate they advise.