We tend to think of economic planning as a problem of collecting enough information and using it to make better decisions. If only we had all the relevant data, we could calculate the best way to allocate resources, organise production, and meet people’s needs.
In 1945, economist Friedrich Hayek challenged this assumption in an essay that remains relevant to some of the biggest questions in economics, technology and artificial intelligence.
His argument begins with a deceptively simple observation: the knowledge needed to make good decisions is scattered across society. A shopkeeper knows something about local demand. A factory manager knows something about a production bottleneck. A customer knows something about a need that may never appear in any formal dataset. Much of this knowledge is local, temporary, and hard to communicate in full.
The problem, Hayek argues, is not simply how to process information. It is how to coordinate decisions when nobody possesses all the knowledge required to make them.
He uses the price system to explore how people can adjust their behaviour to changing circumstances without needing to know everything that caused those changes in the first place.
The essay is about economics, but its implications reach much further. It raises questions about centralised planning, organisational decision-making, data-driven management, and the assumption that greater computational power necessarily produces better decisions.
The problem isn’t just how to calculate the right answer, but how to use knowledge that no one person possesses.
Read: The Use of Knowledge in Society
Friedrich Hayek, The American Economic Review, September 1945.
Why RareRead picked it
Hayek makes a distinction that is easy to miss: a problem of calculation is not the same as a problem of knowledge. Better computation can help us process what we know, but it cannot automatically give us knowledge that is dispersed, changing or never fully articulated.
That distinction changes how we frame the problem before we even begin looking for a solution.
Nearly eight decades later, as we build systems designed to collect more data, automate more decisions and optimise more of society, the question Hayek raised remains worth considering.
What if the hardest problem isn’t calculating the right answer, but knowing enough to calculate it in the first place?