The Verb They Chose

At three minutes past eleven on the night of 20 August 1991, the Estonian Supreme Council did not declare independence. The document is filed as a resolution, and the verb in its first point is kinnitada, to confirm. The claim was that the Republic had never stopped existing and that the previous fifty-one years had been an occupation rather than a country. That argument is why the pre-war gold came back from the Bank of England. It is also why the electorate was smaller by 454,990 people sixteen months later.

The vote closed at three minutes past eleven at night. Ülo Nugis, who was chairing, announced the result, and then he read out the next item of business, which was about transport: "Otsus on vastu võetud. Teadaanne. Saadikuid ootab lossi ees kolm mikrobussi. Nendega viiakse teid koju." The decision is adopted. Announcement. Three minibuses are waiting for deputies in front of the castle. They will take you home.

That is the first sentence of restored Estonian independence, and it is about minibuses. I find it the most Estonian detail in the whole story, and also the most misleading, because the thing they had just voted was not what the minibuses suggest. It was not a country declaring itself into existence at the end of a long night. Look up the instrument in the State Gazette and it is filed as an otsus, a resolution, not a declaration. The verb in its first point is kinnitada. To confirm.

Nobody chooses that verb by accident at eleven at night with Soviet paratroopers in the city. Estonia's answer to the twentieth century was not a declaration of independence. It was a legal argument, made under pressure, that there was nothing to declare, because the Republic of Estonia had never stopped existing and the previous fifty-one years had been an occupation rather than a country.

The argument settled things for my Russian-speaking family that it did not settle for the family next door, and I have spent a fair amount of my life on the receiving end of its second-order effects. I still think it was the right argument. I also think the country has spent thirty-five years not paying for it.

The first vote failed

The version I grew up with has the deputies voting while the tanks close in. The real sequence is less cinematic and more interesting.

On 19 August, the first day of the coup in Moscow, the Supreme Council tried to pass an independence act and could not. The blockage did not come from the pro-Moscow bloc. It came from inside the pro-independence camp: six Christian Democrats and fourteen independent democrats, twenty votes in a chamber where fifty-three was a majority, who thought the timing was suicidal. Illar Hallaste, that day: "Independence? What does it give us? Do you want to let yourselves be occupied a second time?" Part of the delay was pure farce. Several drafting groups had been formed, and each assumed another one was handling independence.

The 234th Airborne Regiment from Pskov crossed the border at Luhamaa that same day and reached Tallinn on the 20th. Volunteers were defending four buildings in the centre of the city with rubber truncheons, gas masks and gravel-loaded lorries parked across the streets. The television tower, which is the image everyone reaches for, had its grounds seized at four in the morning on 21 August, and the defenders barricaded on the twenty-second floor were never dislodged. That is after the vote, not during it.

The deputies were aiming for seventy votes rather than the sixty-nine they got, and when I understood why, the rest of this essay fell into place. Seventy would have been two thirds of the full 105-member chamber, which would have made the act stand as a constitutional amendment inside Soviet law as well, in case the coup succeeded and Soviet law was the law that mattered. Nugis put it more bluntly later: with seventy, the putschists would not have found it so simple, in their own Soviet legal space, to stand us against a wall afterwards. They missed it by one. Thirty-five of the 105 seats were simply empty.

Look at what that chamber was. This was a Soviet-era body, elected in March 1990 by all permanent residents including serving USSR military personnel, with four of its seats reserved for the Soviet army. Its chairman had run the Estonian SSR for seven years. The head of the Estonian Communist Party voted yes. And the paperwork from the decisive days is gone: the Presidium sitting, the council of elders and the drafting committees were never minuted, and the statement by twenty-three deputies that forced the issue survives only as a draft, so nobody can now say for certain who signed it. The founding document of the country I live in has a hole in the middle of it where the minutes should be, and I find that steadies me.

Why they refused the exit that was on offer

There was a legal way out of the Soviet Union, and Estonia would not use it. This is the part I think is genuinely admirable, and it is also the part that later did the damage.

The USSR had passed a secession law on 3 April 1990. It required a two-thirds referendum, a waiting period, separate referendums in autonomous districts, and then a transition of up to five years whose terms Moscow's own Congress of People's Deputies would set. On paper it was a door. In practice it was a door that could be held shut for half a decade, and using it meant conceding the one thing Estonia would not concede. Arno Almann, who was Secretary of the Presidium at the time, put the objection in a single line: a null and void entry into the Soviet Union cannot be followed by a lawful exit from it. Walking through that door would have been an admission that you had legally been inside the building.

So instead of seceding, Estonia built a chain of instruments that all say the same thing in different registers. In November 1989 the Supreme Council annulled the 1940 decision to join the USSR. On the seventieth anniversary of the Tartu Peace Treaty in February 1990 an assembly of deputies declared that treaty still valid. The resolution of 30 March 1990 declared the Soviet state power in Estonia unlawful from the beginning and opened a formal transition period, and the 20 August resolution cites it by name. Then, in October 1992, the newly elected Riigikogu wrote the doctrine down in its cleanest form: the present Republic of Estonia is, as a subject of law, the same state that was proclaimed in 1918. Even the new constitution refuses to be new. The people adopted it, its own closing words say, on the basis of § 1 of the constitution that entered into force in 1938.

None of this would have worked if the West had not spent fifty years quietly agreeing to it. The Welles Declaration of July 1940 committed the United States to not recognising territorial changes achieved by force, and Washington closed its Tallinn legation that August while stating that it did so only under duress. The United States kept recognising the Estonian mission on its territory. The consulate general in New York ran continuously through the entire occupation, and the man who kept it running, Ernst Jaakson, had entered the Estonian foreign service on 20 August 1919. Estonia's independence was restored on 20 August 1991. He served seventy-nine years and died in office at ninety-three. I have never been able to decide whether that shared date is a coincidence worth mentioning or the kind of thing a small country tells itself, and I mention it anyway.

If you want the argument made properly, Lauri Mälksoo's Illegal Annexation and State Continuity is the standard work and the revised edition is free to read.

What the argument was worth

Doctrines are cheap. This one paid out, and the clearest case I know is the gold.

In June 1940 Eesti Pank was ordered to surrender the gold it held in Sweden, Britain and the United States, worth then about forty-one million kroons. Moscow only ever got the Stockholm holding. The rest sat abroad for half a century, and when Estonia needed backing for a new currency in 1992, the depositories paid out to a central bank that had been liquidated in 1940, because the claim was made by the same legal person that had deposited it. The Bank of England returned roughly fifty-two million dollars. Sweden settled in German marks rather than handing back the metal. A last 256.6 kilograms is still at the New York Fed, where Eesti Pank keeps it, in its own words, mainly out of historical interest.

What strikes me is how far the same argument reached. Estonia took no share of Soviet debt, on the reasoning that you cannot inherit from a state you were never part of, and it never joined the CIS. The argument also changed what the rest of the world had to do in the late summer of 1991, and this is the part that gets flattened in the retelling. Iceland moved on 22 August, the Russian SFSR on the 24th, Britain and Sweden on the 27th. The UN admitted Estonia on 17 September. But most Western states were careful not to say they were recognising a new country, because they had never accepted that the old one had ended. They resumed relations. Moscow, for its part, conceded nothing at all: the USSR State Council's resolution of 6 September speaks of the entry of the Republic of Estonia into the USSR, which is the word Estonia had spent two years refusing. Russia has never accepted the occupation. Soviet troops nonetheless left on 31 August 1994.

What it bought, and what it did not

The anniversary version of this argument is weaker than people think.

The economic case does not work the way it is usually told. On the Maddison Project's numbers, Estonian real GDP per head grew about 60% between 1989 and 2022. Belarus, the standing control group for what the other road looked like, grew about 57%. Russia grew about 99%, though most of that is the oil price between 2000 and 2008, and 1989 is a distorted base year for all three. Estonia is far richer than Belarus in level terms and always was, but the growth rates do not carry the story, and Lithuania is an even more awkward case: it started poorest of the three Baltic states, took the opposite decision on citizenship, and has now overtaken Estonia on GDP per head. Convergence with the EU is real and unfinished and currently going backwards. Estonia went from 36.1% of the EU average in 1995 to 79.6% in 2025, but it peaked at 85.2% in 2020 and has fallen in four of the five years since.

Where the comparison does hold is where Estonia started level or behind, and that is bodies. Estonian male life expectancy in 1990 was 64.5 years on the World Bank's series (Eurostat's own figure is 64.7), which was 1.8 years below Belarus on that same World Bank series. Then it fell every year until 1994, when it bottomed at 60.5. That bottom is post-independence, not Soviet-era, and it is the human price of the transition showing up in the only statistic that cannot be spun. From there it climbed to 75.1 by 2024, while Belarus reached 69.7. Measured from 1990, Estonia gained 10.6 years, Belarus 3.4. On the UN's Human Development Index the same reversal shows up: Russia ranked above Estonia in 1990, 0.762 against 0.738, and by 2023 Estonia was 36th in the world and Russia 64th.

So that is the shape of it, dents and all. Independence did not make Estonians grow faster than everyone else. It made them stop dying early, and it bought a set of institutions that the growth figures do not capture.

And it left plenty alone. Estonia has the highest unadjusted gender pay gap in the European Union, 18.8% in 2024. The at-risk-of-poverty rate among the over-65s, a relative measure that sets pensions against a fast-rising median wage, hit 52.3% in 2022 and was still 33.7% in 2025. Someone in Estonia with only basic education can expect to live to 71, against 81.8 for someone with higher education, and for men alone the figure is 67.6. That is an eleven-year gap between the two education groups, inside a country of 1.3 million people, thirty-five years on.

The bill

Now the part the day does not usually mention.

If the Republic never lapsed, then its citizenship never lapsed either, and on 26 February 1992 the Supreme Council followed that logic all the way down. It resolved to apply the 1938 Citizenship Act in the version in force on 16 June 1940, the day before the occupation began. Citizens were those who had been citizens then, plus their descendants. Everyone who had arrived in the intervening half-century, and their children born here, had to apply like any foreigner.

Both of the usual descriptions are wrong, and getting them right is the whole of my quarrel with how this gets discussed. Nobody was stripped of Estonian citizenship in 1992; the people affected had never held it, and their Soviet passports had recorded ethnicity rather than citizenship. That is the answer to the harshest version of the accusation. But "nothing was taken" is equally false, and here is the measure that does the work: the electorate went from 1,144,309 people entitled to vote in March 1991 to 689,319 in the summer of 1992, a gap of 454,990 between two differently defined bodies of voters, in sixteen months. The parliament elected that September was entirely ethnic Estonian.

The March 1991 referendum is what makes this uncomfortable and not just regrettable. That vote is cited in the preamble of the 20 August resolution as its popular mandate. It was open to every permanent resident, citizen or not, and 77.8% voted yes on an 82.9% turnout. American observers on the ground concluded that about half the Russians in Estonia voted for independence. Estonian estimates are lower, around a third, and the difference has never been resolved. Even at the low end, the yes column was carried in part by people who a year later were not on the roll. Narva voted 25% yes and Tartu 86%. The spread tells you what I would want any reader to take from this: nobody's community voted as a bloc, mine included.

There was another road, and until I went looking for this essay I did not know it had been taken off the table on purpose. The Citizenship Commission's draft of 9 September 1991 would have let permanent residents as of 30 March 1990 simply choose Estonian citizenship, with no language examination and no ten-year wait. Once it was published and it became clear how many post-1940 arrivals would qualify, opposition from the Congress of Estonia killed it. The following June, the restored citizenry approved the new constitution by 91.9% and, on the same day, declined by 53.5% to let people who had already applied for citizenship vote that September. The narrowed electorate voted not to widen itself.

The doctrine is not an abstraction, is my point. It is a line on a form. The residence clock for naturalisation starts on 30 March 1990, so a man who arrived in 1961 does not count from 1961; he counts from a date eleven months before the referendum he voted in. Continuity is elegant when you are arguing for the return of gold from the Bank of England. It is something else when it decides which of two neighbours has to sit an exam.

Thirty-five years on, 58,309 people in Estonia hold undetermined citizenship, about 4.3% of the country. That is not the same thing as the 74,278 residents who hold Russian passports, which is a choice. The undetermined figure falls every year, which sounds like progress until you look at how. Naturalisations ran at 22,773 in 1996 and 751 in 2024. In the 2021 census only sixty of the sixty-six thousand then in the category were under fifteen, while 41,326 were over fifty. The problem is being closed as much by death and emigration as by naturalisation.

I have written before about what this feels like from the inside, and about how rights here track citizenship rather than ethnicity. Both of those posts cover the March 2025 constitutional amendment, which took the local vote away from third-country nationals from July 2025 and from stateless residents from March 2026. It is the first time since 1992 that rights here have been subtracted rather than added. What those posts do not have is the arithmetic underneath it. Kohtla-Järve went to the Supreme Court and lost, having calculated that from March 2026 48.32% of its permanent residents would have no vote in their own town. The municipal ballot had been the compensation built into the 1992 settlement by the people who designed it. It has now been withdrawn by their successors, at a moment when the argument for keeping it, that people with no stake in a country are easier for someone else to talk to, is stronger than it has been in decades.

The day was never simple either

One more thing, which I like because it ruins the ceremony a little.

Estonia did not want this holiday, and I had it wrong for most of my life. In the winter of 1994 the Riigikogu voted it down, with the help of Isamaaliit and the ERSP, meaning the parties most committed to the continuity doctrine were the ones least keen to celebrate the day it took effect. For years afterwards the date was mocked on the right as edgaripäev, Edgar's day, because Edgar Savisaar's Centre Party was the one marking it. It became a public holiday only in January 1998. It still sits a rank below 24 February in the law, which reserves the phrase "national day" for the 1918 date and lists 20 August as an ordinary public holiday.

The deputies took it personally. Thirty-six of the sixty-nine met at Pühajärve in 1994 and founded the 20 August Club, whose charter says it dissolves when its last member dies.

And Latvia declared on the 21st, not the 20th, which nobody outside the Baltics ever gets right.

Thirty-five

The arithmetic is why I wanted to write this today rather than in some other year.

The first republic lasted twenty-two years, from 1918 to 1940. The occupation lasted fifty-one, from 1940 to 1991. The restored republic is thirty-five years old as of this evening, which means it has now outlived the original by more than half again, and it needs seventeen more years, until 2043, to have lasted longer than the thing that interrupted it. That is the actual position, and it is why the tone of the day here is closer to relief than to triumph. That surprises visitors. It has never surprised me, or anyone I grew up with.

I think they got it right in that room. The continuity argument is why there is a country here to argue about, why the gold came back, why accession to the EU and NATO was a return rather than a favour, and why a Russian claim on this territory has no legal foothold to stand on. I would make the same argument. Given the fifty-one years, I do not know what else anyone was supposed to do.

But an argument that good has a way of settling questions it was never meant to answer. The doctrine was designed to defeat a claim by Moscow, and it was then used to sort out the people Moscow had sent, most of whom were not sent so much as moved, and some of whom voted for this country before it decided they were not yet part of it. That was a choice. It was made in a particular month, by a particular chamber, against a draft that would have done it differently, and my reading of the record is that any parliament which wanted to could revisit it tomorrow.

The state's position is that it never stopped existing. The people it declined to enrol never stopped living here either. Both of those have been true for thirty-five years, and only one of them has a public holiday.

Enjoyed this? Get the next one.

New essays, in your inbox. Double opt-in, unsubscribe anytime — no tracking pixels.

Prefer RSS? There's a feed →

Related