Here’s a nice exchange between Agnes Callard and Russ Roberts on EconTalk. Remember — the defense resource allocation process, the PPBE, was explicitly created on neoclassical economic theory. So this isn’t just theoretical, but has real impacts on the way defense weapons programs are decided upon.
Agnes Callard: … once I say ‘rational choice,’ you’re immediately in the framework where somebody has like a few different options. Those options somehow are magically pre-articulated for them. Like, we’re automatically in the supermarket somehow, right? A supermarket of choices.
And, the question of their rationality is the question: Which of those things do they pick? And, in addition to the choices being given to them, somehow magically, all the information that they will ever have about these choices is given to them. And, some of that information might be information about what they don’t know. But that’s also given, fixed.
Russ Roberts: I think that the view of the economic project is–I used to like it. I don’t like it so much anymore. I think it’s a–there’s something robotic about it that I only appreciated after I read James Buchanan, at my friend and former colleague Don Boudreaux’s encouragement, where he says, ‘That’s not economics, that’s just engineering. You give me my preferences and tell me my relative values of these different things and then you show me the prices and then I’ve got an engineering problem.’ That’s not what life is about. Life is so much richer and more complicated… And, the economic view, by the way, of the firm, of a company, is: Given that it’s making widgets, how many widgets should it make? That’s the theory of the firm. When in fact, if you ask anybody in business or think about it for a minute, the real goal of, the challenge a firm faces, is trying to figure out which market it’s in. Which kind of widget it should be making. Not how many. How many is a narrow, sterile problem.
Agnes Callard: Right. So, maybe my view about the economic agent is a little higher than yours, which is that I think that that’s a real and legitimate form of rationality. And, I’m sure you don’t deny that–right?–to figure out how many widgets you should make. And, I describe it as reasoning from value. Right?
So, suppose you do have all the information. What should you do? Suppose you do know what your options are and you have these desires. Like, then, often, the answer is very easy, okay? But, you know, what I’m trying to bring out is, like, I can grant all that and I can grant that that covers an important range of phenomena, but there’s still this question, like, ‘Where did we get these preferences? Where did they come from?’ Right?
And it’s not true that they just pop into existence full-grown like Athena from the head of Zeus, right? There’s a story that can be told and telling that story, and managing that story is super-important to us.
And so, there’s another thing called reasoning towards value, and that’s the process I’m trying to describe. One of the sort of projects of my book is to say, ‘Don’t assimilate these two things. They’re not the same.’ We can’t understand reasoning towards value as a special funny case of reasoning from value. There really is an important independence; and that independence has to come from the fact that reasoning towards value is a learning process. It’s not a process you make from whatever information you have. You make it towards a better informational state. You’re trying to learn, and you’re trying to come to value. You’re moving towards having preferences. You’re not moving, sort of jumping off from them.
Air Force acquisition chief William Roper is certainly trying to do what Callard says is reason towards value through a learning process. We see that approach on Advanced Battle Management System, which he says is not going to be run as a regular major defense acquisition program. It will do iterative deployments of increasing capabilities, largely by setting standards for interoperability.
But the DoD’s process of decision-making treats it more like a case of reasoning from value, as if the entire set of user requirements and technical feasibility were completely given at the start of the calculation.
Leave a Reply