Ideas Made to Matter
Why a leading economist is embracing machine learning
By
Smart watches that track health, websites that anticipate purchases, and voice-recognition systems that respond to commands are a few of the ways machine learning has transformed daily living.
Yet many economists haven’t embraced the artificial intelligence technology because the predictive algorithms that make Siri smarter over time can’t answer questions about correlation and causation yet. However, Stanford University Graduate School of Business professor Susan Athey recently told MIT Sloan’s Andrew McAfee, the co-director of the MIT Initiative on the Digital Economy, she believes economists are gradually acknowledging machine learning’s potential to transform the way they work.
In a recent appearance on the “Mind and Machines” podcast, Athey talked with McAfee about how economists can use machine learning effectively, what she’s doing to create algorithms to answer questions of correlation and causation, how women can excel in high-tech and academic jobs, and how she creates a work-life balance.
On the benefits of machine learning
For decades, economists have built their assumptions about prices, wages, and inflation on data sets only as large as they or their research assistants could calculate. Machine learning has the potential to dramatically enlarge those data sets and allow economists to test their models faster than ever.
“It’s going to really improve the way that we can do empirical work in economics,” Athey said.
On economists’ resistance to machine learning
Athey has received a “lot of pushback” from economists who say machine learning algorithms can’t determine whether a connection between a statistically linked pattern is a coincidence or a cause-and-effect relationship. Economists, she said, should embrace the benefits of effectively having a robotic research assistant. A machine learning algorithm allowed her to study cell phone mobile location data from millions of customers to see where people ate lunch. Large data sets are useful when tackling questions around urban planning and poverty. In the meantime, Athey and her Stanford colleagues are working to create algorithms that can interpret causal relationships.
On interest in machine learning for economics
There is a gradual acceptance of applying machine learning to economics. Athey said enrollment in her machine learning course at Stanford has more than doubled in three years.
On being a female leader in the tech industry
Athey was consulting chief economist for Microsoft Corp., and she has served on the boards of Expedia, Rover, and Ripple. She told McAfee women typically fare better at companies where lines of authority are transparent, evaluation criteria are clear, and where office politics and fraternization matter less. She advised younger women to develop unique talent as a way to land at an organization where they’ll thrive.
On work-life balance
Athey said she’s relentless about delegating things that aren't important and won’t compromise when it comes to spending personal time with her family. “I've taken a million ‘red-eyes’ so I can be back for breakfast. I've turned down a lot of travel, especially international travel, to be with my kids.”
