The Price System: Lessons from “The Use of Knowledge in Society”

I recently came across an article titled “The Use of Knowledge in Society.” Its outline caught my interest, and I later learnt that it’s considered “one of the most significant essays in economics”. Written by a well-known economist, the text occasionally uses some language that’s difficult to read, but the ideas are pretty clear.

Centralized vs. Decentralized Planning

The article examines two economic planning models: centralized/socialist and decentralized/capitalist. A key concept discussed is the “price system” - which is basically the idea that we use prices on things to exchange stuff amongst people. While it’s something we encounter daily, I had never thought deeply about how it works or its significance.

The price system uses prices to coordinate the exchange of goods and services. When prices change, they signal what is needed or not needed, guiding decisions and resource allocation across society. This process happens without centralized oversight, making it an efficient way of organizing economic activity.

A Different Perspective: Termites and Coordination

An unusual comparison in the essay is the system used by termites for coordination between themselves. Unlike humans, termites don’t rely on prices - at least not that we know of - but they still manage a pretty complex system effectively. This is used as an example for an alternative way to think about large-scale organization and highlights the uniqueness of the price system.

Prices as Tools for Communication

The essay talks about how prices are used as a decentralized way of communicating knowledge. Changes in prices reflect the needs and priorities of different parts of the society and individual members use this information to make decisions that keep resources flowing where they are needed most.

Two Types of Knowledge in Economics

The essay also discusses two kinds of knowledge relevant to planning:

1. Long Term Knowledge: This is stable, widely recognized information like physical laws or mathematical principles. It is relatively straightforward for planners to incorporate because it is well-established and doesn’t change quickly. So if they make an assumption based on such knowledge it holds true within the dynamics of the model at any point in the future. He makes a point that we have come to equate this kind of knowledge with “Scientific Knowledge” which ends up occupying the whole idea of knowledge for a lot of people.

2. Local and Unstructured Knowledge: This is specific, ever-changing information held by individuals about particular circumstances at a certain time and place. Examples include:

   • A real estate agent’s awareness of upcoming property listings.
   • A shipper’s knowledge of available capacity on departing ships.
   • A trader’s insight into regional price differences for the same product.

Although this local knowledge is crucial for resource allocation, it is difficult to centralize or use effectively in a top-down planning system.

Bias Toward Different Types of Knowledge

The essay notes that society often views those who profit from local knowledge, such as traders or middlemen, less favorably than those who benefit from the longer term kind of knowledge, like engineers or scientists. This bias exists even though both types of knowledge play essential roles in the economy.

Challenges of Centralized Planning

The author explains why centralized planning cannot fully incorporate local knowledge:

1. Accessibility: Local knowledge is distributed across many individuals and is often hard to gather and communicate to a central authority.
2. Constant Change: Even if centralized planners obtain this knowledge, it may quickly become outdated, making it hard to act on in a timely way.

While centralized systems may excel at applying scientific knowledge, they cannot effectively manage the unstructured, temporal and localized nature of the other kind of dynamic knowledge.

Potential Role of AI

This discussion feels relevant again today with the rise of advanced AI. Large Language Models (LLMs) can process vast amounts of unstructured data, potentially helping planners identify and act on micro-information more effectively. While AI doesn’t truly “understand” the information, its ability to analyze large datasets quickly opens up new possibilities for addressing some limitations of centralized planning.

Conclusion

“The Use of Knowledge in Society” provides an interesting perspective on the role knowledge plays in a economy and how it flows in the society. In this context the price system acts as a mechanism for decentralized coordination.