Decentralization Papers III: Democracy, the Half-Finished Revolution

Democracy is the most successful decentralization project in history: of legitimacy. Its unfinished half is decision-making, which quietly recentralized into parties, bureaucracies, and platforms.

Twenty-five centuries ago, one Greek city ran an experiment that still embarrasses us. Athens filled most public offices not by election but by lottery: the sortition of ordinary citizens into councils and courts, serving short terms, then returning to private life. Magistrates were audited on leaving office. A citizen grown too powerful could be ostracized: banished for a decade by popular vote, not for crimes but for looming. The Athenians had rediscovered, at the scale of a city, the forager insight from Part I. Keeping power distributed takes active, institutional vigilance against anyone who would accumulate it.

Say the caveats plainly: Athenian citizenship excluded women, the enslaved, and resident foreigners (a democracy of a minority). The mechanism, though, deserves separate judgment, and the mechanism was radical: it treated governing as a capacity of ordinary people rather than a profession of specialists. Rotation and lottery made office a duty passing through many hands, not an asset held by few. Rule was something you took turns doing.

Modern representative democracy kept the legitimacy and quietly dropped the mechanism. What we practice is a compression algorithm for scale: millions of preferences compressed, every few years, into a mark beside a name; the elected then decide everything else until the next compression. The American founders were candid about the design; Madison, in Federalist No. 10, praised representation precisely because it would *refine* public views through a chosen body rather than let the public rule directly. Where Athens used office to disperse power, elections select for the ambitious, the funded, and the connected: a filter for exactly the personalities a forager band would have ridiculed into silence.

The drift that follows has a name. Studying political parties, including the most radical and egalitarian of his day, the sociologist Robert Michels concluded that large organizations inexorably concentrate control in a professional leadership whose first priority becomes keeping it. He called it the iron law of oligarchy. Organization itself, he argued, breeds oligarchy: not because leaders are wicked, but because information, agenda, and machinery pool at the top of any permanent structure. Parties, ministries, unions, platforms. The rare exceptions at scale, like the typographers' union Lipset studied with its built-in two-party system, are organizations that fought the law deliberately, which is the point.

Around the elected core grew the unelected mass: career bureaucracies executing policy, central banks setting the price of money, regulators writing the working law, and now private platforms curating the public square that opinion forms in. Each exists for defensible reasons. Together they mean that between elections (which is nearly always) the live decisions of a modern polity are made by people no citizen chose, weighing facts no citizen sees. The ballot decentralized the *source* of authority. The *exercise* of authority re-pooled where it always pools.

And yet democracy's achievement is real and staggering. It made power's legitimacy flow from below; it built the only reliable procedure in history for removing rulers without blood; and where it spread, famine grew rare (no functioning democracy with a free press has suffered one, as Amartya Sen observed); and democracies have almost never gone to war with one another. This series does not join the fashionable despair about democracy. It makes a more precise claim: democracy is a decentralization project that stopped halfway; it distributed the right to *choose* the center, and left the center standing.

Finishing the project means pushing decisions themselves outward, toward the scales where the facts live: the neighborhood, the cooperative, the commons, the protocol. The twentieth century's great tragedy is that its loudest movement for the powerless marched in exactly the opposite direction, concentrating power beyond anything the kings had managed, with a sincerity that makes the failure instructive rather than merely evil. Marx's diagnosis, Lenin's machine, and the knowledge problem that broke it are the subject of Part IV.