Using the Scientific Process to Lift North Americans Out of Poverty – Professor Simon Quach on His Approach to Economics and the Effects of ChatGPT on Labor Market Outcomes
My first glimpse into Professor Simon Quach’s approach to economics came at USC’s Café Literatea over cups of coffee. Previewing his Spring semester course — Econ 471: Economics of Labor Markets and Human Capital — Professor Quach told me that he designed his class to “introduce students to the scientific process of economics.” After speaking with Dr. Quach again this Spring, I recognized his enthusiasm for the power of economics. Explaining that he sees “beauty in the scientific process of economics,” Professor Quach shared that he designed his course to enable students to use empirical tools and better understand the discipline.
“The Labor Market Effects of Expanding Overtime” — Putting Empirical Methods to Work
Noting that the beauty of economics lies beyond classical theory, Professor Quach described the “clever ways economists think of theory and test them with data.” He explained that the scientific process is more than an art; it is a tool to discover how to better help people.
While Dr. Quach respects classical theories as ways to explain the world, he cautioned that “they are not always correct.” For him, it is the labor economist’s job to test theories with data and to reevaluate those theories if they prove incorrect. This approach has motivated him to use labor market policy changes as natural experiments to learn how to lift individuals in North America up out of poverty.
Professor Quach entered the Ph.D. program at Princeton University with this guiding question. After switching from development economics to labor economics, Dr. Quach focused on identifying “what policies we can implement to help those that have not gained as much from the economy.” With this focus in mind, Dr. Quach examined how increasing overtime pay to higher-income thresholds affects labor market outcomes for lower-income individuals. He explained:
Most people assume that salary workers do not get any overtime; they get a fixed salary no matter what. In the US, legally, salary workers who earn below a legislative earnings cutoff are qualified for overtime, but historically that cutoff has not changed much since the 1980s. Since 2016, the federal cutoff has only been $23,000 a year. California, New York, Alaska, and Maine have thresholds that are way above the federal limit and change frequently. My paper analyzes these policies.
The results of his study revealed both intended and unintended consequences. On the one hand, his experiment revealed that an unexpected chain reaction — ignited by expanding overtime — boosts workers’ wages. However, he also found that expanding overtime may come at a slight cost:
I found that the policies succeeded in raising the workers’ incomes. What happens is not due to increased income from overtime pay but rather employers saying, “I do not want to bother keeping track of these workers’ hours and paying overtime; I am just going to raise their salaries above a certain cutoff.” So, income rises, which is great for a select group of workers. It appears from the data that there is some evidence that there is a slight decrease in employment. The result is interesting because one of the initial intentions of overtime’s passage in the Great Depression was that overtime would raise employment. The idea was that if we incentivize companies to lower workers’ hours to forty, they will hire more employees to replace them. In this instance, that theory does not hold.
A Pandora’s Box or a Compliment? Identifying the Effects of AI and ChatGPT on Labor Market Outcomes
Dr. Quach noted that although labor economists have heavily researched how the rollout of artificial intelligence and ChatGPT will affect labor markets, it is “unclear what will happen”. However, he identified a few things to look for in the future. Mentioning that AI could “open up a new pandora’s box of jobs being replaced,” Professor Quach said:
There is concern that robots are replacing medium-skilled labor, such as people in factories with routine jobs, because robots are good at repeating an action over and over. As a result, economists thought non-routine jobs, like computer programming and research, were safe. Non-routine jobs that are lower-paid, such as service-sector jobs that require interactions with people, tend to be non-routine. However, with the advancement of AI, some of those jobs could be replaced. Sometimes if you have a complaint, you chat with a robot — which is not that great all the time — instead of a person. Now, with the launch of ChatGPT, you could have a decent conversation, so I imagine some potential for disruption in this sector of the service economy.
As to high-skilled labor, Professor Quach used his job as an economist to lay out two different possibilities. Noting that “ChatGPT could take my job,” Dr. Quach said, on the other hand, it could augment high-skilled labor. He explained that economists invest considerable time writing code for their research but that there is potential for ChatGPT to lower this. A colleague told him that after giving ChatGPT their code, it returned a faster copy. Therefore, ChatGPT could complement economic research by making the process more efficient and productive.
Thank you to Professor Quach for being the first subject of our new interview series.
