Share this:
09/26/2012 09:33:44

Matthew Elliott: Modeling Networks

Matthew Elliott is Caltech's newest assistant professor of economics. Born in England, he earned his BA and MPhil from Oxford in 2002 and 2004. After receiving his PhD from Stanford in 2011, he spent a year at Microsoft Research in Cambridge, Massachusetts, before arriving in Pasadena this fall.

Elliott's research focuses on mathematically modeling different kinds of networks. For example, in networked markets, the interactions among players are constrained in a way that can be represented as a network. In a labor market, for instance, not everyone can be employed in every job, whether it's because they're not qualified, they don't have the requisite connections to get the job, or they simply aren't aware that there's an opening. Elliott distills this kind of complex system into its mathematical essence, developing theories that can eventually inform policy. His research is part of the Social and Information Science Laboratory, which is funded by the Gordon and Betty Moore Foundation and the Ronald and Maxine Linde Institute for Economic and Management Sciences. He recently spoke a little more about his work.

What's another example of a network market that you're trying to model?

Within this context of network markets, another thing you might think about is the production and sale of natural gas. The producer of natural gas can only sell its gas to another country if there's a pipeline between it and that other country. You can view the pipelines as a network that describes which countries can trade directly with which other countries. And a question you might be interested in is whether the network is built efficiently. Are there good incentives in place? What kind of inefficiencies would you expect and how bad can they be?

Does this analysis occur after the fact, or do you do this before you actually build the gas pipelines?

Most of the analysis is after the fact, and you're trying to explain what's going on. But by being able to explain things that have happened before, and why, you can hopefully understand a little better the problems that will arise in the future and try to avoid some of those inefficiencies.

What excites you most about your job?

I love doing what I do. It's the problem solving. You go to work and your job consists of playing with problems and trying to find solutions to them. I find it pretty remarkable that people pay me to do this—and it's not something I just do in my spare time.

Your research is very theoretical. But do you also work with real-world, empirical data? Or do you pass along your theories to someone else who can apply them?

Somewhere in between. A good example is a project on financial networks that I'm now working on with [Stanford economist and former Caltech professor] Matt Jackson and [MIT postdoc and Caltech graduate (BS '07)] Ben Golub. The idea is to model the network of financial relationships between either banks or countries and to understand the dependencies between them. What we want to know is, if one of those countries or banks receives some shock that's going to cause it to fail, how does that spread through the system? When does one failure lead to a contagion of other failures?

Sounds like something that's quite relevant today.

It's definitely a topical thing, and we're certainly not the only people working on this. A small part of that project is collecting data on European countries and their cross holdings. Then we try to see what our theory has to say when we apply it to the data.

I'm very aware that the research I do is very theoretical, and most of the time it isn't going to specifically be something that policymakers are going to read and take to heart. But I think my work does provide a framework for them to think about problems. I think it's exciting to be able to do that.

Written by Marcus Woo