AI Cannot Outperform Markets: Why Economic Complexity Defies Computation
Whenever significant advancements in computational power emerge, advocates for centralized economic planning resurface with renewed vigor. They argue that sophisticated artificial intelligence (AI) systems could finally optimize tax rates, ensure adequate production to meet societal needs, and allocate resources to maximize overall well-being more effectively than decentralized markets. This perspective has gained theoretical prominence multiple times throughout history—first in the early 20th century, then with the mid-century rise of modern computing and operations research, and now again with the impressive progress in AI technology.
The False Premise of Economic Computation
However, this line of thinking rests on a fundamentally flawed assumption: that an economy is merely a computational problem waiting to be solved with accurate equations, sufficient data, and adequate processing power. In reality, the economy is not a vast set of equations but rather a complex discovery process that unfolds in real time through the interactions of countless individuals. Even the most powerful computers available today cannot fully capture or replicate this dynamic process.
As argued in a recent paper for the Montreal Economic Institute, this critical error was understood as far back as the 18th century by Adam Smith. In his seminal work The Wealth of Nations, which recently celebrated its 250th anniversary, Smith observed that producing even simple goods requires the cooperation of so many different participants that the complete network of exchanges would "exceed all computation." For instance, creating a woollen coat involves farmers, spinners, dyers, merchants, shippers, and numerous other actors just to transform raw materials into a market-ready product.
The Invisible Hand and Tacit Knowledge
This immense complexity does not prevent the coat from being produced. Smith's crucial insight was that no single mind directs every step of production from raising sheep to selling finished garments. Instead, production becomes possible through the spontaneous cooperation of the many hands and minds that constitute what Smith famously described as the "invisible hand" of the market.
Italian economist Vilfredo Pareto expanded on this concept in the late 19th century, noting that coordinating even a modest economy and matching resources to uses and preferences would create an explosion in the number of equations requiring solution. While today's computers can handle quintillions of computations per second—far beyond anything Pareto could have imagined—this computational power alone cannot overcome the fundamental limitations identified by Nobel laureate economist Friedrich Hayek.
Hayek explained that the problem extends beyond merely decentralized knowledge spread across millions of individuals. Much of this knowledge is tacit—local shopkeepers' understanding of their customers' buying habits cannot be reduced to a single data point for AI models. Similarly, we cannot predict the emergence of entrepreneurs who dream up entirely new products that previously did not exist.
The Indispensable Role of Prices
Most importantly, Hayek emphasized the phenomenon of prices as indispensable signals that guide economic decision-making. Prices are neither fixed in stone nor arbitrarily determined. Instead, they emerge organically from real exchanges between buyers and sellers. When the price of wheat rises, for example, this reflects competition for limited supply among market participants.
This price increase serves multiple crucial functions: it signals relative scarcity, provides incentives to adjust consumption patterns and conserve resources, encourages the search for substitutes, stimulates increased production, and fosters innovation. These dynamic price mechanisms operate in real time through market interactions, creating a discovery process that no computational model can fully replicate or anticipate.
As powerful and helpful as AI can be for improving logistics, managing inventories more efficiently, and analyzing market trends, it remains just that—a tool. The economy's inherent complexity, reliance on tacit knowledge, and dynamic price discovery processes ensure that markets will continue to outperform even the most sophisticated computational approaches to economic planning.
