“USA250: The Story of the World’s Greatest Economy” is a yearlong WSJ series examining America’s first 250 years. Read more about it from Editor in Chief Emma Tucker.
Much has been said about what artificial intelligence could mean for jobs, corporate business models, productivity and economic growth.
But what does this technological leap mean for the field of economics? In particular, how might AI impact economic policymaking? The short answer: It will make economic policymaking more accurate, particularly when it comes to the work of central banks.
For decades, institutions such as the U.S. Federal Reserve have made enormously consequential decisions—raising or lowering interest rates—based on incomplete and often delayed information.
The Fed must adhere to mandates of full employment and price stability, but inflation readings arrive weeks after the fact. Employment statistics are revised months later. This leaves policymakers operating in a world of uncertainty, interpreting imperfect signals and using models undermined by gaps in real-time knowledge. The result is that central banks sometimes wait too long to raise or cut interest rates in the face of price changes in the economy.
AI has the potential to change that.
AI-led systems will be able to process and continuously update vast information sets—ranging from consumer prices and wage settlements, to financial transactions and supply-chain activity. This will allow policymakers to observe economic dynamics as they unfold rather than long after the fact. In practice, this could materially transform—for the better—the policy decisions that are made.
In addition to providing real-time data, AI could significantly improve central banks’ models of how the economy works and the complex interrelationships between economic variables. AI-driven analysis, for example, could more quickly and accurately show the impact of a quarter-percentage-point rate increase on growth, prices and inflation, and the broader flow through into the economy. The Bank of England has noted that generative AI introduces “fundamental changes” to the way data is used and has already significantly increased the size and complexity of the models it uses.
End of assumptions?
AI’s implications for economics could extend far beyond policymaking, however.
AI could upend the discipline of economics itself, as its ability to collate and analyze vast quantities of data could reduce or in some cases eliminate the need for economists to make assumptions about individuals, businesses or markets when modeling the economy.
Take the “aggregate utility function,” or the idea that economists can combine the preferences and desires of millions of individuals into a single, representative and aggregated measure of overall utility of a society, or its living standards, commonly measured as GDP per capita. This type of extrapolation and analysis has been a cornerstone of economics research for decades.
AI, however, may make it possible to gather and analyze previously unimaginable amounts of data on individual utility and decision-making. By leveraging AI, economists will be able to replicate the complex economic system at a granular level rather than relying on population or cohort averages, or data extrapolated from smaller samples. This will enable economists to spot emerging risks faster and more accurately, and to a degree they haven’t been able to do before.
Then there is the notion of the “rational actor,” with logical or determinable preferences.
Neoclassical economic theory assumes how this rational individual behaves in response to prices and incentives, and this forms a basis to model the economy. Economists assume, for example, that consumers with limited income will allocate their budget to maximize utility—choosing the combination of goods that delivers the greatest satisfaction per unit of cost. So if a household is choosing between chicken and beef, and beef becomes more expensive while household income remains the same, the rational consumer will purchase less beef and more chicken because chicken now provides better value for money. Economics also assumes that we can learn something about our economy as a whole by analyzing and aggregating these decisions.
But if AI can collate and analyze data to capture individual preferences at scale we may no longer need the assumption of the rational actor. We would no longer have to make assumptions about individuals’ preferences because we would know them with certainty.
Big data and advanced analytics have taken us part of the way on this journey. But AI will take us to the point where the preferences of each individual, at any given moment in time, are known. Essentially, AI could push economics closer to its “ideal,” free of herculean assumptions. In such a world, economic predictions will no longer be weakened by delayed or incomplete information.
What the future holds
If AI is going to render many of the governing assumptions of economics moot and help us realize the intellectual goal of the discipline itself—to be able to accurately model the economy—where will that leave economists?
AI could in fact make the field of economics more relevant, not less.
AI might take some of the data collection and extrapolation work of economics away from economists. But it also could pave the way for economists to produce faster, high-quality analysis as their ability to anticipate economic turning points—from financial crises to inflationary pressures, and possibly to identify systematic risks building in the economy—improves with wider access to better information. That ultimately will lead to better decision-making by consumers and businesses, as well as policymakers.
Dambisa Moyo is a global economist and a member of the U.K.’s House of Lords. She can be reached at reports@wsj.com.