Andre Mouton

Office Kirby 202A

Welcome!

I'm an Assistant Professor of Economics at Wake Forest University. I joined Wake after receiving my PhD from Carnegie Mellon in 2022.

I study macro-labor topics related to technological change and inequality. My research uses micro-data to examine how structural features of labor markets -- such as tasks and occupations -- inform macroeconomic relationships, like automation's effect on wages and employment.

Recently, I've been working on the development of methodology/resources for measuring labor's exposure to technological innovation. I finetune a "small" language model (Sentence-BERT) and use it to (1) match USPTO patent applications with ONET occupational task statements, and (2) perform on-demand classification of patents and tasks. This allows exposure measures to be quickly estimated for user-defined technologies and task categories, and at a fine cross-sectional and temporal level. Some early results:

Data: aggregate, wage quantiles (quarterly), 4-digit SOC (annual)

The working paper (available here) shows that with training, Sentence-BERT is able to achieve near-GPT levels of matching/classification accuracy -- all while being simpler to use than older NLP methods, free and open-source, and able to run on a laptop. This drastically lowers the barriers to working with large, textual datasets.

Custom measures can be generated in a few minutes with the accompanying Python package (link here). The matched dataset will be updated quarterly, though it should be noted that patent applications are subject to an 18-month publication lag.

Google Sites

Report abuse