For most of history, centralized administration faced a constraint no amount of political authority could remove.
The centre could not process enough information.
To administer a complex society, it had to observe conditions it could not see directly, transmit that information, integrate it into a decision and send the decision back — all before reality changed.
Beyond a certain scale, the arithmetic stopped working.
Decentralization was one solution.
Local institutions could see things the centre could not.
But something important is changing.
The arithmetic is weakening
Machine processing attacks several parts of the old constraint simultaneously.
Information already generated digitally can be gathered cheaply.
Transmission is nearly instantaneous.
Structured data can be processed at enormous scale.
Digital identity can connect a person to credentials and entitlements across systems.
The European Digital Identity Wallet is one current example.
Every EU member state is required to provide a certified wallet, creating infrastructure intended to make identity and credential verification interoperable across a market of roughly 440 million people.
The broader significance is not the wallet itself.
It is what the infrastructure represents.
Administrative capacity that once required enormous human bureaucracies can increasingly be bought.
But not all capacity can be bought
This produces a distinction we think will matter increasingly.
Infrastructure — identity rails, registries, compute and payment systems — can often be acquired through procurement in several years.
Judgment cannot.
Assessment, adjudication, inspection and professional discretion depend on trained people and institutional practice.
They may require five, ten or fifteen years to build.
This explains an apparent contradiction.
The same institution can simplify aggressively in one domain while building far greater centralized capability in another.
It subtracts complexity where capacity must be trained.
It acquires complexity where capacity can be bought.
The next divide
We therefore do not expect every administration to become equally capable.
Four complementary inputs matter:
Structured data.
Identity infrastructure.
Compute and the electricity underneath it.
Standards and legal authority to connect them.
Because these inputs complement rather than replace one another, capability may arrive as a threshold.
Below it, additional investment produces limited results.
Above it, the entire stack suddenly becomes usable.
Our base scenario assigns a 50% probability to divergent capability — a minority of administrations crossing this threshold while others continue simplifying because they cannot.
This suggests a different way of comparing jurisdictions.
Administrative capability may increasingly depend less on conventional governance rankings and more on the infrastructure underneath government.
Then comes the harder problem
More processing capacity solves one problem.
It does not solve another.
Distinguishing is not correcting.
A system may become extremely good at recognizing millions of different cases while still applying the wrong model to them.
Correction requires feedback.
Someone or something must be capable of telling the system that reality does not match its model.
Historically, some of that feedback came from expensive human intermediaries:
local officials with discretion;
people who saw the rule and the individual case simultaneously;
appeals requiring decisions to be defended;
professionals who noticed exceptions the central model did not anticipate.
Automation can remove the reason these intermediaries existed.
And when they disappear, some of the feedback they carried disappears with them.
The efficiency gain will be measurable.
The lost correction mechanism often will not be.
What this means
For individuals, check what important institutions already hold about you — and whether it is correct. An error that once encountered a human intermediary may increasingly propagate automatically through multiple decisions.
Maintain a second route where possible. A manual alternative can remain legally available long after it has become practically unusable.
For business, assume identity and credential verification become progressively cheaper. Ask what parts of your revenue exist because verification is currently difficult — and what new products become possible when it is easy.
For capital, treat administrative infrastructure as a jurisdictional variable alongside taxation, political stability and rule of law. Two countries with similar statutes but radically different registries and identity infrastructure can produce very different operating environments.
The historical constraint on centralized administration was processing capacity.
That constraint is beginning to weaken.
Which leaves us with a more difficult question:
The question is no longer whether the centre can process enough.
It is whether it can still find out that it is wrong.
Read the full analysis:
WHEN CAPACITY CONCENTRATES
What changes when the historical limit on the centre stops binding
THRIVE IN CHAOS
Decision Intelligence for an Uncertain World
Analysis → Forecast → Recommendations
Signal → Meaning → Action → Stability
Forecasts are probability-based analytical assessments, not certainties. This material supports independent judgment and does not constitute financial, legal or investment advice.