McBride couches his endorsement in the context of historic moments where being wrong impacted people's reputations going forward: Kudlow saying No Bubble Here in 2005, the (second) war in Iraq. He believes that supporting Trump will be another such litmus test, and despite being a registered Republican, he is supporting HRC.
I was surprised to read this excellent post, largely because (as McBride notes) he does not often get into politics. He makes his case succinctly and well, and the future will show whether he is, once again, correct in believing that supporting the de facto Republican nominee at this point will look Real Bad in the future.
McBride does not mention what about Trump has led him to this conclusion (he's spoiled for choice, really). However, I would point to this article as some excellent reasons (skip the video):
Without getting into the question of whether printing money is a bad idea (I would imagine many economists and informed amateurs might want to know how much money you were planning on creating and what you were then going to do with it before weighing in definitively), the zooming back and forth between positions as incompatible and extreme as "we'll pay it back -- at a discount" and "we'll just print whatever we need" is disorienting and hardly inspires trust, which is, in the end, sort of important when it comes to money.