Friday, January 31, 2014

How to Think Like an Economist

Welcome to Economics 101. Please check your politics at the door.

The best piece of advice I can give anyone who is studying economics for the first time is this: leave your politics at the door. It's hard, I know, because economics is a subject that can have a very personal impact on all of us. When biology works against you, you get sick. It's sad, of course, but you can't exactly go and blame nature for doing its thing. When math works against you, you just need to study harder or get help from someone who has more training. When literature works against you, you just find something better to read. But when economics works against you, you're out of a job, you can't pay your mortgage, and -- here's the kicker -- it's probably because someone else did something that, intentionally or unintentionally, has seriously screwed you over.

I think that's what sets economics apart from most other fields of study: the fact that economics studies the way people make choices and how those choices affect other people, including you. In other subjects, you study laws that can't be changed, languages that you may not need to learn, arts that nobody is forcing you to appreciate. But in economics, you're dealing with how other people's behavior affects you, and that is something that we have political resources to do something about. So when most people talk about economics, they're concerned with how to make the world a better place, how to change the state of things, how to make things more fair.

That's not something that economists are particularly concerned about.

Now, don't get me wrong; economists are people too, and of course they care about those issues as well -- on a personal level. But when economists are actually study economics, they're not focused so much on how to improve society as they are focused on understanding why society is the way that it is. Hopefully that understanding can lead to improvements being made, but the economist's job is simply to provide the understanding.

Let's suppose that the rent is too damn high, and you want to have an economist help you solve that problem. What do you expect the economist to do? Well, if she's doing what she's been trained to do, she won't just tell you a policy that will fix the problem. The economist's job is to analyze the housing market, identify factors that are affecting supply and demand, as well as legal constraints that influence how the housing market works, and build a model to explain how all of these conditions interact with each other to determine why rental prices are what they are. Once you have this model and you understand what is going on (more or less; remember that no model is perfect), you can consider policy solutions and the economist should be able to help you predict what consequences those policies might have. The economist might recommend policies to you, but if she does that she is no longer acting as an economist, she's acting as a consultant. There's nothing wrong with that, and it doesn't make her a bad economist -- she's just filling a different role at the moment. What's important is that even if the economist has certain policies she is recommending, she shouldn't let a political agenda drive the reasoning behind her model.

The point of the whole thing is, the economist's job isn't to tell you how the world should be; the economist should be focused on helping you understand why the world is the way it is. Economics, as studied by economists, is more of a branch of applied mathematics than it is a political subject, which is how most people usually discuss it.

In this blog, economics is the means, and artificial intelligence is the end. I really have no intention of discussing current events and policy issues, unless we can build a model to analyze them. All I really care about here is gaining insight from economic models that will help us to program agents who behave more intelligently in a simulated economy.

Get on with it!

I don't plan to do many philosophical posts like this. Moving forward, I plan to alternate between posts like the last one, which show off the results of a simulation, followed by posts like the next one, which show how I did it. It will be a cycle of "here's something cool", followed by "here's how it's done". (Note that my definition of what is cool may or may not match yours.) 

So, where did I come up with that function to determine who gets what when two agents trade in the previous simulation? Coming up next, I'll be talking about utility functions, Edgeworth boxes, contract curves, and general equilibrium theory. This will provide a framework to help us understand the problem our simulated economic agent (economaton?) is trying to solve.


  1. I'm waiting for your next post

  2. Thanks! It's good to know people are reading this thing.

    I did throw up something last night about how to work through the trade function I came up with, but it was basically just a link to some notes that show how to work through these kinds of problems. If you have a hard time following the notes I posted, then post a question in the comments -- it's my way of gauging how familiar my audience is with the material I'm covering.

    As for my next real post (i.e., where I further develop the simulation), I'm working through the math for finding the Contract Curve (set of all equilibrium trades) for two agents in our model. This lets us examine more types of trades, with a particular focus on bargaining power between two agents. I should have that up in a day or two.