Practical Knowledge Cannot Be Centralized

The most consequential knowledge in any complex system is distributed, tacit, and impossible to aggregate which is why centralized planning consistently fails and why local autonomy consistently outperforms.

"Nobel Laureate Friedrich Hayek eloquently makes this case in his book, The Fatal Conceit. He argues that formal planning methodologies do not capture what really drives a competitive economy, in particular the information processed through decisions made daily by millions of buyers and sellers."

Both Hayek and James C. Scott arrive at the same conclusion from opposite directions. Hayek, the economist, argued that the price system processes more information than any central planner could ever gather that the economy is essentially a distributed computation, and replacing it with a central bureau is like replacing the internet with a single computer. Scott, the anthropologist, documented how peasant farmers, forest dwellers, and urban neighborhoods develop intricate knowledge systems which crops to rotate, which paths flood, which neighbors can be trusted that no state bureaucracy can replicate.

Boyd's military philosophy reinforces this from yet another angle. The German Blitzkrieg succeeded because it pushed decision-making authority down to the lowest competent level. The Auftragstaktik system mission-type orders that specified what to achieve but not how worked because frontline soldiers had the practical knowledge to adapt to rapidly changing conditions. The French, with their centralized command structure, had to pass information up and orders down, losing precious time at every step. As General Gaedcke noted, if he had tried to write everything down, "we would have been too late with every attack we ever attempted."

The pattern holds across domains: Soviet collective farming destroyed the metis of individual farmers. High-modernist forestry replaced diverse ecosystems with monocultures that collapsed. Urban renewal projects demolished neighborhoods whose social capital had developed over generations. In each case, the knowledge that made the system work was precisely the knowledge that could not be seen from above, reported in a spreadsheet, or replicated by decree.

Takeaway: Any system that requires all important knowledge to flow through a central node will be outperformed by one that trusts distributed actors to use what they know because the most important knowledge is the kind that resists transmission.


See also: High Modernism Fails Because It Ignores Metis | Seeing Like a State Means Missing What Matters | Legibility Kills What It Tries to Measure