1. Jonathan has a great post on "non-market" activity here. Although it's a minor point in the post he suggests I mean "non-price" rather than "non-market". This is interesting... probably an improvement but it kind of depends on what you mean by "price" and for that matter "market". Jeff Sessions in his head a price that he would take for a vote on immigration reform, for example. In this "political market" as some have referred to it there's some quid pro quo that could get Sessions to vote for it - presumably quite high. Other politicians have other prices. You may respond that it's not a monetary price, but there aren't monetary prices in cases of barter either and we usually consider those to be markets. So I'm not sure "non-market" is as inferior to non-price as Jonathan suggests, but at the very least talking about these issues helps us clarify what we're talking about in certain cases! It also helps us think about what modeling methods are amenable to what cases. Household economics, for example, often does not involve exchanges but it does involve utility maximization and payoffs. So you end up seeing a lot of game theory and optimization in that literature, but not many supply and demand curves. I'm not an expert on public choice, but I imagine more traditional supply and demand reasoning might be applicable. Another dimension we can think about here is personal vs. impersonal exchange relations.
2. Another point that came up in the comments is that Rand liked big corporations. Sure, fine. This is something I said on Facebook that I probably should have said here: I don't know Rand and if you want to knock Krugman for tying this decision to the guy's Randian sympathies than by all means ding Krugman for that. That seems to me to miss the major point of the post, which was not to be a treatise on Randianism but instead to highlight the importance of taking Coase and Williamson seriously.