Robert Merton on the Limits of Quantitative Models

Well it also kind of starts to talk about, and we may come back to this later, starts to address this issue of models, and that they're inherently, to use your words, incomplete. And so there are people walking around with models that they thought were much more complete than they turned out to be. Oh, absolutely, well, models, you'll often hear people say during a crisis or something, these were bad models, or good models. And someone says, is that a good model? And that sounds like a good question, a reasonable question. It actually isn't really well posed. You need a triplet: a model, the user of the model, and its application. You cannot judge the model in the abstract. And the reason is, as you said to me, every model is an abstraction from complex reality. There is no complete model. In fact, there's Godel's theorem or something, shows you can't overdo it. Anyway, so if they're incomplete, that means a couple of things. They don't capture everything. But two, that means you have to make judgements. Because, any science, physics, chemistry, biology, and finance, finance is a science, it qualifies. All of them have the feature, that no matter how quantitative they are, how mathematical they are, and how much they rely focused on data, and all these very seeming precise answers, they all have what I call the art of the science. And the art of the science is the abstraction. Because when you look at a problem in the real world, what you're trying to say is, what model can I create, that will give me, if I solve it, insights into the way that part of the world really works? Not hypothetically, but how it really works. There is no formula for that. Right. That's judgment. Because somehow, if you look at the stream of models that have lasted or ideas that have preserved, or ideas that were continued or things that influenced the developments of the future, there is no simple pattern. Sometimes the models are very simple, and sometimes they're rather complex. So it isn't simplicity or complexity, it's the right tool for the right job. And then in doing that, that also applies in the world of practice. Because if people say oh no, I don't believe in these models. I just use my intuition and history. Yep. That's a model. So you can't get away from it. And understanding that is critical to having an understanding of how science progresses, how science can influence practice, and what's good practice.