National Operations of the NKVD

What was actually being targeted

The logic of the national operations turned not on nationality as such but on real or presumed ties to 'hostile capitalist states', that is, on cross-border diasporic connections. Arrest criteria therefore reached far beyond citizenship to include membership in German, Polish or Latvian cultural societies, correspondence with relatives abroad, visits to consulates, or past residence in a target country. Volga Germans were seen as 'our Germans' and suffered more under the Kulak Operation than under the national rubric, while people who had returned from Harbin, ethnic Russians though they were, were treated as a functional diaspora and targeted as such.

Which operations were there

The national operations did not begin with a single order. From July 1937 to February 1938 a new order or directive came down for one nationality after another, and the Polish Operation (No. 00485) became the procedural template that later operations cited clause by clause. In the table below Nos. 00439, 00485 and 00593 are operational orders, while Nos. 49990 and 50215 were wired directly to the local offices as telegraphed directives. Each number links to the full Korean translation of that document.

The same directive of 14 December 1937 extended the Latvian line to Lithuanians and Bulgarians. No separate founding order has been identified for the Chinese, but mass arrests ran through 1937 and 1938, and in May to July 1938 some 7,900 Chinese were deported from the Far East to Xinjiang. The Korean case is different in kind: the deportation of August to October 1937 shares the targeting logic but was carried out by a resettlement decree of the Council of People's Commissars rather than by arrest and sentence, and is therefore usually classed as an ethnic deportation rather than a national operation.

In total 335,513 people were convicted and 247,157 of them (73.7 per cent) shot. The figures above come from the post-1990 archival record; estimates for the Harbin and Greek operations still vary widely between researchers.

The German operation came first but was initially narrow, covering German nationals in armaments plants and on the railways, and only spread to Soviet citizens of German descent once the Polish order was issued. When the card indexes ran dry, investigators began arresting people for having a German-sounding surname.

The Latvian operation is a case of the provinces moving before the centre. In November 1937 Nasedkin, head of the Smolensk NKVD, reported having uncovered a 'Latvian National Centre' and asked permission to arrest 500 Latvians; Yezhov replied that this was nonsense, that he would clear it with the Central Committee, and that if Latvian blood was to be let then no fewer than 1,500 to 2,000 should be arrested. Arrests began before Directive No. 49990 was issued, and when Nasedkin asked whether people could be arrested without compromising material, he was told to obtain the material during the investigation.

The Greek operation had a particularly high death rate. The directive listed the districts of compact Greek settlement one by one and declared them bases of British, German and Japanese intelligence; in the Kuban 92 to 94 per cent of those arrested were shot. The text of Directive No. 50215 was declassified only in 2006.

Who decided the sentences

Unlike the Kulak Operation (Order No. 00447), which began with numerical quotas, the national operations left the scale of arrests to local discretion but controlled the outcome through the 'album procedure'. Every ten days, local NKVD chiefs and prosecutors (the 'dvoika') bound indictments with recommended sentences into albums and sent them to Moscow, where the Commission of the NKVD and the Prosecutor of the USSR, signed off by Yezhov and Vyshinsky, made the final decision.

Under orders Nos. 00485 and 00593 alone the commission convicted 235,122 people: 172,830 were sentenced to be shot, 9,529 expelled from the USSR and 7,777 exiled internally, while 37,619 cases were returned for further investigation and 12,903 people released. Results diverged sharply by region: in Novosibirsk province the local investigation proposed death for 97 per cent of the 6,067 people it convicted, whereas in Omsk province 43.6 per cent of those processed were sentenced to death. When unresolved albums piled up by September 1938, Order No. 00606 scrapped the procedure and created regional 'special troikas' with two months to clear the backlog.

Why at this moment

The immediate impetus was Stalin's reading of the Spanish Civil War: he feared that rearguard uprisings of the kind seen behind Republican lines could be replicated inside the USSR, which made every diaspora nationality with ties across a hostile border a potential 'fifth column'. On 31 January 1938 the Politburo extended the operations to 15 April and added Bulgarians and Macedonians to the target list, then extended them again in May to 1 August.

Why the toll varied so much by region

The absence of quotas meant that scale was set by the discretion and the cruelty of the local chief. NKVD officers coded arrestees as 'livestock headcount' and executions as 'slaughter'; beatings and torture were routine instruments of confession. In Moscow each department worked to a monthly arrest target of 1,000 to 1,200, and the membership roll of the Latvian club was used directly as an arrest list.

The result was extreme unevenness shaped by local ethnic demography. The operations imprisoned roughly 16 per cent of all Poles in the USSR and, in Karelia, made Finns, 3.2 per cent of the population, more than 40 per cent of all victims of repression.