Computers were one of Susan Athey’s first loves. As a child, she had a Radio Shack TRS-80 model at her house and taught herself to code; her earliest publication, as a preteen, was an article for Color Computer Magazine describing how she programmed a sound simulator.

At Duke University, she studied computer science and took a job administering the Unix workstation of an economics professor who soon encouraged her to pursue a PhD in his field. Athey was skeptical at first. But the professor was called to testify before Congress about procurement practices they were studying together, and Athey was impressed by the impact economics could have on policy and industry. She added economics and mathematics to her course of study.

But, in a move that proved prescient, she didn’t drop computer science (she earned degrees in all three disciplines in 1991). Years later — after Athey earned her PhD from Stanford GSB and had entered the academic job market, after she had proven herself as one of the foremost economists studying, among other things, markets, and auctions, and after winning the American Economic Association’s prestigious John Bates Clark Medal for leading economists under the age of 40 — she did a stint as chief economist at Microsoft, where she helped the company develop the advertising platform for its search engine Bing. In so doing, she became one of the first economists to recognize that her field and computer science could be more powerful working hand in hand. Her early experiences in the co-application of both disciplines made her another kind of expert altogether — what has since become known as a tech economist.

Today, Athey, the Economics of Technology Professor at Stanford GSB, is using her expertise to promote the public good. In 2019, she founded the Golub Capital Social Impact Lab, which uses digital technology and social science research to improve the effectiveness of social sector organizations.

The lab’s projects use economic principles and digital technologies — including artificial intelligence and machine learning — to assist social-impact organizations working in education, financial health, government, healthcare, charitable giving, and career transitions.

In 2022, Athey took her work with the social sector a step further, accepting a two-year position as the chief economist for the U.S. Department of Justice’s antitrust division. Among other efforts there, she led the DOJ’s team that, together with the Federal Trade Commission, drafted new Merger Guidelines, which are issued periodically to describe the factors and frameworks government agencies use to review the legality of mergers and acquisitions. She also helped create a new technology policy team by hiring data scientists and bringing in the antitrust division’s first chief technologist.

Motivated by Impact

For more than a decade, Athey’s professional passions have been linked to their potential for impact. She chose to return to Stanford — after six years teaching at Harvard — because of the opportunity for cross-disciplinary collaboration. And she has helped make such collaboration possible. In 2019, she was a founding associate director of the Stanford University Human-Centered Artificial Intelligence Institute. She is also a leader of the Business and Beneficial Technology pillar within Stanford GSB’s newly launched Business, Government, & Society Initiative, which brings together academics, practitioners, and policymakers to address issues such as technology, free markets, and sustainability.

Athey’s Golub Capital Social Impact Lab epitomizes interdisciplinary work, putting students from computer science, engineering, education, and economics backgrounds to work helping partner organizations leverage digital tools and expertise that are generally only available to — or affordable for — large technology companies.

“I like building things that demonstrate how a class of problems can be solved,” Athey says. “If there is a problem worth solving, and I can solve it myself in a particular case, I know there are other people like me who are going to encounter the same problem. Part of the motivation and theory of change of the lab is that we will solve particular problems for particular social-impact organizations but also create the research that will guide others in solving similar problems.”

Some of the lab’s projects include working with PayPal and ImpactMatters (now Charity Navigator) to identify ways to increase contributions to charitable organizations during checkout experiences; helping the creators of a learn-to-read English educational app in India improve student interest and engagement, including through personalized recommendations and a contest that offered free books for students who read the most; making a digital tablet application to help nurses counsel patients about contraception options in Cameroon in partnership with the World Bank; and developing an online program that helps women create portfolios that can increase their chances of being hired by technology companies in Poland.

The lab has also developed CAREER, a generative artificial intelligence model based on 24 million resumes that can predict what type of job will follow a previously held job. Such a model is important for questions in labor economics, including estimating gender or racial differences in unemployment. Previous such predictions were based on much smaller survey data sets.

The model “is a way to study the future of work but also study” foundation models themselves — including how to reduce bias in such models, Athey said at a Brookings event last year.

Athey says each project demonstrates a potential benefit of technology but not all the potential benefits of technology. That’s by design. Progress, she says, should be incremental, and one of the benefits of digital and data-driven work is that it can be: Researchers can test the effects of small tweaks and adjust as they go rather than committing to bet-the-company changes that may not end up solving the original problem.

She described that approach in an interview for Microsoft Research in 2018.

“What we’ve learned from tech companies is that you shouldn’t think about coming up with one grand idea and then go out and spend a year testing that one idea,” she says. “Rather, you iteratively improve… Digital programs are amenable to incremental innovation.”

Athey says some parts of economics are evolving to include a healthy dose of engineering. In the Microsoft Research interview, she described stereotypical economics research as evaluating existing programs and often finding that “stuff doesn’t work.”

“There’s a lot of negativity,” she says. With help from data and machine learning techniques, “my prediction is that economists are going to become more [like] engineers. Instead of complaining that nothing works, we’re going to start building things that do work to achieve economic outcomes…. We’re going to realize that nothing works if it’s one size fits all, but that a lot of things work if they are actually personalized and appropriately delivered.”

Through the lab, Athey also taught a course called Data-Driven Impact. In the class, which is part of the GSB’s Action Learning Program, students designed experiments and data analysis techniques to estimate the impacts of product features for partner organizations.

“Across my experience in the business world, I learned how valuable it is to understand techniques and how they should be applied,” she adds. “There is a shortage of people in the world who both understand technology and how to guide and measure its development to truly benefit its users.”

Athey says she enjoys helping students achieve those skills; her passion for teaching and mentorship is one reason she didn’t accept a permanent, full-time position at Microsoft.

“I’m very motivated by the need to serve, to help other people be productive,” she says. “It’s exhilarating and intoxicating when people have questions they need to know the answers to, and you have the ability to answer them.”

Athey’s lab isn’t the only one that offers students and researchers the chance to partner with the private sector. But Emil Palikot, a postdoctoral researcher who has worked with Athey since 2020, says it is groundbreaking.

“What is unique here is that Susan identifies organizations that have a strong social impact and integrates with them very closely,” he says. “This level of close integration almost never happens because generally, academics don’t have the know-how of how the technology operates.”

Optimizing government efficiency

Athey’s Department of Justice gig is another attempt to use what she knows to make a difference. “If you connect it back to the lab,” she says, “by going there, I’m trying to make the social sector — in this case, the government — more effective by, for example, building better capabilities for analyzing large datasets and understanding algorithms.”

For the updated merger guidelines, Athey helped incorporate new economic analysis — including from academic writing she had done with Fiona Scott Morton, a Yale economist who previously held the chief economist role — that describes some of the behavior unique to digital platforms.

“The agencies are incorporating — more explicitly and to a greater extent — the applied theory and strategy literature developed over the last fifty years,” Scott Morton wrote in a blog post about the guidelines.

The final guidelines were released in December following an extensive public comment period (more than 5,000 comments were submitted), four listening sessions, many panel discussions, and an additional 30,000 comments received after an initial draft was released last summer.

“Synthesizing all of that — listening to and understanding and evaluating it — was a huge intellectual challenge,” Athey says.

The division’s new technology policy team enhances the department’s capabilities to work with large datasets and broadens the type of cases and conduct the antitrust division can analyze.

“Bringing that expertise on board changes the kind of cases we even have the capability to investigate,” says Laura Edelson, who served as the first chief technologist before returning to academia last year. “There were cases that we couldn’t even start without having that expertise available.”

The name of the group the antitrust division’s chief economist oversaw used to be the Economic Analysis Group, but during Athey’s tenure, it was changed to the Expert Analysis Group to acknowledge the inherently interdisciplinary nature of the work.

“It’s a recognition that economics is not the only domain of expertise the division needs in order to investigate its cases,” Edelson says. “That’s exactly the kind of insight that Susan is your person for. She really does think about problems in a complex way.”

A large footprint

Impact is one of the threads connecting the diverse strands of Athey’s career: She pursued economics because of its potential to influence policy and business; she returned to computer science and technology because of its potential to help society; and she has combined those disciplines in her lab and government work to maximize the effects of interventions designed to improve people’s lives.

Nevertheless, Scott Morton says it’s nearly impossible to judge Athey’s impact in economics, the field that made her famous in the first place.

“If you restrict the inquiry to theory, she’s easily one of the top theorists; if you restrict the inquiry to machine learning, she would be one of the most impactful economists in the world; if you confine yourself to economists who have impacted policy, she is one of the most important economists in the world; and if you restrict yourself to economists who have dealt with business strategy, she would be one of the top economists in the world,” Morton says. “I have no idea how to rank one of those areas over another. We won’t know her true impact for 50 years.”