Compromise capitals

Why is the seat of government so often a modest town rather than the country's greatest city?

Ottawa, Canberra, Bern, Brasília, Abuja, Washington: none was its country's leading city when it was chosen, and several were barely towns. The pattern feels like a rule waiting to be stated. But before stating it, count the other way. London, Paris, Tokyo, Moscow, Seoul, Cairo, Bangkok, Mexico City and Jakarta are all capitals and the largest city, and worldwide they are the clear majority. So the modest capital is the exception — and the useful question is not why capitals are small but why the exception clusters where it does.

Look at where it clusters and the answer nearly announces itself. Almost every modest capital belongs to a country assembled out of units that already existed separately: colonies, cantons, provinces, kingdoms, each with its own court and its own leading city. Countries that grew outward from a single core — France from Paris, England from London — never produced this pattern at all.

Why should the assembled kind behave differently? Ask what hosting the government is worth. Salaries, ministries, contracts, embassies, land values, and above all proximity: whoever is close to the rule-makers is heard first and cheapest. That is not a one-off gift but a permanent stream, attached to a place forever.

Now put yourself in the founding room. Each unit is being asked to join, and each can walk out. If the capital goes in a rival's leading city, you have signed away a permanent advantage to the party you were most worried about. Your ranking of the options is not "which is best" but "anywhere but theirs" — and every rival's ranking says the same. Every prominent candidate has at least one determined opponent, and unanimity is the price of union.

Here is the strange turn. Usually a group that must agree without discussing terms gravitates to whatever option stands out — the obvious meeting point, the round number, the biggest city. Prominence is what makes coordination possible. But in this game prominence is exactly what disqualifies a site, because prominence means it already belongs to someone. The one property every party can accept is advantages nobody, and that property is possessed only by a place with no standing at all. So the bargain does not land on the second city, which still belongs to someone; it lands on a small town, a boundary, or an empty plain.

That explains the compromise capitals. It does not explain all of them, because a second and quite different motive keeps turning up. Governments sitting in the largest city are within walking distance of the largest crowd that can assemble outside their windows. Rulers have moved for that reason — to a purpose-built court, an inland town, a walled administrative city with no public in it. This is not neutrality among rivals; it is distance from the governed, and it is chosen by strong rulers rather than negotiated by weak ones.

Is that distance costly? Studies comparing US states, and then countries, found that more isolated capitals go with weaker accountability — less newspaper attention to state government, lower turnout, more corruption — the argument being that scrutiny depends on how many people are near enough to notice. It is a real finding from a small literature, contested rather than settled, since isolation is entangled with a great deal else.

One thing the founding logic does not survive. The advantage the founders fought over does not disappear once the site is chosen; it simply moves. Brasília and Abuja grew from nothing into millions precisely because government is a growth engine. The neutral site becomes a prize — solving the founding generation's problem by creating a smaller version of it for later ones.

Three distinctions matter, and they are routinely blurred.

A compromise capital was neutral at the founding: Washington on a boundary river, Ottawa on the line between English and French Canada, Bern chosen over Zurich and Lucerne, Canberra fixed by law inside New South Wales but far from Sydney.

An escape capital was chosen to put distance between government and a great city's population or its foreign entanglements. Ankara over Istanbul is the clearest case, and Naypyidaw the most extreme.

A relic capital was never modest by design — it simply stayed where it was while another city outgrew it. Wellington did not shrink; Auckland grew. The Hague never was the largest Dutch city and never intended to be. Reading a relic as a compromise invents a bargain that never happened.

And a caution about the sources. The compromise story is tidy and gets told even where the record shows something messier: land speculation, a monarch arbitrating on thin advice, a single leader with a plan and no obligation to negotiate. Neutrality was often the constraint on the choice rather than the reason for it.

• Why concentrating a state's administration in one city tends to make that city primate, and what the compromise capitals avoided by not doing so.
• How federal countries divide the functions of a capital — legislature, courts, central bank — across several cities, and whether that settles the same bargain more cheaply.

Every term the collection defines is gathered in the glossary.