They say things like this:
In 2003, to take one example, when President George W. Bush chose to topple Saddam Hussein, he did not appear to fully appreciate either the difference between Sunni and Shiite Muslims or the significance of the fact that Saddam’s regime was led by a Sunni minority that had suppressed the Shiite majority. He failed to heed warnings that the predictable consequence of his actions would be a Shiite-dominated Baghdad beholden to the Shiite champion in the Middle East—Iran.Really? I was a vocal opponent of Bush's invasion and I read every argument against it I could put my hands on, and my memory is that "experts" disagreed vehemently about the significance of the Sunni-Shiite divide. Some thought it would explode, but others called that old-fashioned colonial thinking and said the more sophisticated people of Baghdad had moved beyond it. There absolutely was no consensus among Middle Eastern historians and other experts that deposing Saddam would lead to a worsening Sunni-Shiite conflict.
Start with the issue that the president and his national-security team have been struggling with most: ISIS. Recent statements indicate that the administration tends to see ISIS as essentially a new version of al-Qaeda, and that a top goal of U.S. national-security policy is to decapitate it as al-Qaeda was decapitated with Osama bin Laden’s assassination. But history suggests that ISIS is quite different in structure from al-Qaeda and may even be a classic acephalous network. When we searched for historical analogues to ISIS, we came up with some 50 groups that were similarly brutal, fanatical, and purpose-driven, including the Bolsheviks of the Russian Revolution.Interesting, but I posted here a completely different analysis of the historical analogs of the Islamic State that focused on anti-colonial revolts like that of the Mahdi in the Sudan. Maybe Allison and Ferguson's analogies are better, maybe not. Which historical events provide parallel cases to each other is one of the topics on which historians never agree.
An even worse problem with relying on advice from historians is that most of them are too ideological to be trustworthy. All social scientists have this weakness; as soon as partisan politics rears its head, they toss empiricism aside and start talking nonsense like the Laffer Curve. Actually that isn't fair, there are lots of issues on which most economists (and most historians, and even on a few issues most sociologists) agree because they are firmly founded in fact. But there is always somebody who disagrees. And since Presidents will generally take the advice they find most congenial, they can always find some expert to endorse their own preferences.
Obviously I think our policies would benefit from a proper understanding of history; I write about these issues all the time. I just don't really trust the historical profession to provide useful advice about the big problems facing us. We just don't agree on enough.