The Referendum Scotland Won and Nobody Noticed

At the Aberdeen Exhibition and Conference Centre, on the night of 18 September 2014, the ballot boxes came in past midnight and kept coming until nearly six. Counting agents from Yes Scotland and Better Together sat opposite each other at the same trestle tables, hundreds of independent observers logged every tray, and officers from Police Scotland stood by doors that, for once, nobody needed guarding. Mary Pitcaithly, the Chief Counting Officer, would later say she was satisfied every count had been conducted properly, a sentence that sounds like nothing until you remember it needed saying at all, because for six weeks half the country had genuinely believed the other half might cheat. Turnout across Scotland reached 84.6%, the highest for any nationwide vote since women got the franchise in 1918. In Aberdeenshire it hit 87.2%. Sixteen and seventeen-year-olds had registered to vote for the first time, 109,533 of them, casting a ballot on their own country's future before most could legally buy a drink in the bar where the results were probably being watched.

The record, once the boxes stopped coming, was unambiguous. No won by 10.6 points: 2,001,926 votes to 1,617,989, or 55.3% to 44.7%. No carried 28 of Scotland's 32 council areas, including Edinburgh. Yes carried four, including Glasgow, the largest city in the country voting to leave it. In the Borders, No took two-thirds of the vote outright.

By any normal electoral arithmetic, this was not close. It was, in the word politicians reached for immediately, decisive.

Twelve years on, the word has not aged well. The political scientist Michael Keating called the outcome an apparent paradox at the time, a decisive result that decisively failed to settle anything, and it has held up better than most academic predictions do. The losing side, structurally, is in ruder health than the winners. The 45% has become an achieved baseline of support rather than a defeat to recover from; opinion in Scotland has continued to track toward maximum devolution short of outright independence, a fairly precise description of the settlement nobody actually offered. The winners spent the following decade trying to make good on a promise made in the final fortnight of panic, the one delivered on the front page of the Daily Record over three party leaders' signatures, for something close to federalism. What they delivered instead was the Smith Commission.

It is worth being exact about what the Smith Commission was, because the exactness is the point. It was not Parliament, which never voted on its terms before it reported, nor the electorate, who had just voted and would not be asked again. It was ten politicians from five parties, convened under Lord Smith of Kelvin, given ten weeks to produce recommendations on further devolution, which they duly did, and which were then processed into the Scotland Act 2016 largely as drafted. This is not a criticism of Smith personally, who did the job he was handed with more competence than the job deserved. It is an observation about where the decision actually got made: in a commission assembled at speed to answer a question the referendum had just, supposedly, closed. Who decided how much power Scotland would get? A panel appointed in the ten weeks after the vote, not the eighteen months of campaigning that preceded it.

This is the pattern the referendum entrenched rather than resolved, and it is now visible in the government's Northern Downing Street initiative, an executive outpost intended to answer the same complaint the Yes campaign weaponised so effectively: that decisions about a place are made a long way from the people who live in it. The instinct behind No 10 North is not dishonest. Regional disparity in England is real, measurable, and older than devolution itself.

But look at what the instinct produces. An office. A satellite of the centre, staffed by the centre's people, answerable in the end to the centre's Treasury, whose purpose is to make the executive look present in Darlington or Manchester while the actual decisions stay where they always were. It is Smith's logic in a different postcode: convene something quickly, call it devolution, retain the power.

The confident reading of 2014 is not stupid, and it was roughly the government's own case at the time. The referendum produced immediate legislative action: the Scotland Act followed within two years, and Holyrood now controls income tax bands, welfare top-ups and a fair slice of its own borrowing, more power than Scotland has held since 1707. Turnout of 84.6% is the kind of democratic engagement most Western nations would sell their statistics agency for. On its own terms, Union preserved, further powers delivered, question asked and answered, the case that 2014 worked as designed is not a fringe argument. Much of it is true.

What it misses is what Keating identified within months. Engagement is not the same as settlement, and a mandate delivered through panic rather than negotiation tends to be kept in outline and hollowed out in practice. The Vow promised something close to federalism; Smith delivered a schedule of reserved and devolved competences drafted by ten people in ten weeks; the underlying grievance, that power sits somewhere you cannot reach it however you vote, simply migrated south, first to the Northern Powerhouse, now to an office in Whitehall's northern annexe with a grander name.

There is a version of English devolution being built right now that repeats, almost exactly, the shape of the answer Scotland was given in 2014: fast, technocratic, generous in the small print, stingy in the substance, and designed by people nobody elected to satisfy a question the electorate has already asked twice. The referendum did not fail to settle the constitutional question. It answered it, correctly, and then the state spent a decade proving the question wasn't really about the constitution at all. It was about who gets to decide, and where — a panel of ten, ten weeks, no ballot required.