The Tartan

Last Tuesday, Patrick Harker, the president and chief executive officer of the Federal Reserve Bank of Philadelphia (Philadelphia Fed), spoke about the impacts of AI on financial technology (fintech) and the economy.

The net economic effects of AI are a focus area of the Federal Reserve Bank’s attention. The job of the Philadelphia Fed extends beyond just adjusting the interest rate – its main goal is to stabilize prices and maximize employment. How transformative AI technology affects the social infrastructure and the resulting impacts on economic growth potential are key questions the Philadelphia Fed aims to investigate and formulate public policies around. 

A more efficient payment system is one of the most revolutionary changes that can be fueled by AI technology in the fintech area. However, Harker says, “we have to look at questions of whether these [applications], even in their created efficiencies, actually provide consumers with end-user benefits.” Harker believes this question is still unsettled, pointing to research found in “The Great Reversal,” a book by economist Thomas Philippon.

 “The cost to the consumer remains roughly constant at two percent, which creates a puzzle,” Harker says. “If we have so much more invested in the intermediary systems in the name of efficiency and speed of the financial system, why does the cost on the consumer end remain so steady?” 

The disproportionate benefits of AI, as revealed by a study on payment solutions, shed doubt on the real economic benefit of financial technology advancements. Dr. Harker further talked about how AI technology could reinforce structural inequalities. For instance, a potential application of AI can be evaluating housing applicants by filtering out the personal attributes that should not be taken into consideration in the decision process. However, “even here we found a challenge, algorithms are not perfect, errors and input can exacerbate biases that can fan out across multiple applicants.” In other words, the usage of AI can cause more severe economic inequalities by having data inputs that are already biased. 

When it comes to research, the impact of AI appears to be more positive and optimistic, in contrast to the uncertain effects on the economy. The most profound impact of AI is in the area of data analysis — it greatly expands the capacity of research endeavors by allowing researchers to deploy massive unstructured historical data accumulated over a span of more than a hundred years instead of decades when developing economic models. 

The Center for the Restoration of Economic Data exemplifies research endeavors that employ advanced AI technology. As “small decisions can generate ripple effects,” AI technology can help to capture events that trigger nuanced and persistent impacts, but are often overlooked by human eyes. For example, in housing, Philadelphia Fed senior advisor and research fellow Larry Santucci and his team developed a housing map that traced the growth and spread of racial covenants — clauses in property deeds that prohibit the purchase and occupation of a property based on race — in Philadelphia. Although such racial covenants were outlawed in 1968, their rippling effects spread across the economic landscape, so much so that now, Philadelphia remains the nation’s poorest large city. Increasing the scope and magnitude of data in economic studies helps to fill blind spots in understanding how the economy and market evolved. 

A key takeaway from the lecture is that efficiencies and advancements in some areas do not always bring net positive economic impacts. In today’s fast-changing world, the ability to constantly adjust and learn is the only certain solution to facing challenges brought by changing technology.