In prison, late 1938
about 1 million
In camps, late 1938 (assuming 5 million in camp at the end of 1936)
about 8 million
I also concluded, from much Soviet and other testimony, that not more than 10 percent of those then in camp survived.
My estimates were based on thirty-odd sources, mostly unofficial—though including considerations of the 1959 census and of the secret NKVD section of the 1941 State plan. A number of further unofficial testimonies of some value have meanwhile emerged, confirming the earlier. But they, and the fifteen pages of calculations and considerations in The Great Terror, have been made superfluous by Soviet figures given in 1987–1989. However, we may look briefly at the categories:
1. Arrests. My approximately 7 million was derived from (among several solid sources) analyses by Alexander Weissberg and other ex-prisoners of the numbers arrested in the catchment areas of various prisons, and based on prison documents warranting high confidence. Full Soviet figures have not been given. But those provided for the Kursk province2 imply a total of about 8 million for the USSR as a whole.
2. Executed: approximately 1 million. This would be the result for the USSR as a whole for those shot by Troikas alone, if figures now given for Uzbekistan are taken as typical.3 Figures from Irkutsk imply over 1.5 million.4 The evidence of the Kuropaty mass graves in Minsk and the Bukovnya mass graves near Kiev suggest higher figures still. My estimate was based in part on relatives’ reports of the proportion of death sentences to other sentences. But while realizing that “10 years without the right of correspondence” was often a euphemism for the death sentence, we did not then know that it always was.5 I, and Roy Medvedev (who also gives the 1 million figure), will probably turn out to have underestimated this.
3. Died in camps, 1937 and 1938: approximately 2 million. This was based on prisoners’ reports of the death rate and (taken together with the execution figures) on Yugoslav sources. It would include those executed in camps, who do not figure in the execution estimate above (on the evidence of Kolyma and Bamlag, this may be approximately 600,000 to 700,000).
4. In prison, late 1938: 1 million. This rough estimate was based on the known prison accommodation and many reports of the level of overcrowding, together with figures of inmates for particular prisons. These last are now confirmed in Soviet accounts.6
5. In camps, late 1938: approximately 8 million. This is based in the main on multifarious reports from ex-prisoners. The assumption that 5 million were already in camps at the beginning of 1937 was derived merely from the estimates by other students of the matter, and may be too high. I should be inclined to reduce the 8 million at the end of 1938 to 7 million, or even a little less. Such a figure is consonant with the 12 million now given in Moscow for the camp population in 1952.7
The above are, in any case, only approximations.
The Great Terror was only peripherally concerned with the total casualties of the Stalin epoch. But it reckoned the dead as no fewer than 20 million. This figure is now given in the USSR. And the general total of “repressed” is now stated (e.g., in the new high-school textbooks) as around 40 million, about half of them in the peasant terror of 1929 to 1933 and the other half from 1937 to 1953.8
Thus it would now be accepted almost everywhere that the estimates given above cannot be far wrong. But it is just worth recording that, though it has long been clear that the victims ran into the millions or tens of millions, Western misconceptions of the sort we discussed in Chapter 15 recently had a brief revival. Over the years, a vast amount of true information had established itself in the Western consciousness. But by an inexplicable development, when such notions no longer seemed possible, a few Western Sovietologists began to assert that the Terror had claimed far fewer victims, and that ordinary life was not affected. The writer of a Western Sovietological textbook concerned to reduce the estimates to, as he put it, a few hundred thousand or even a few tens of thousands, wrote, “Surely we don’t want to hypothesise 3 million executions or prison deaths in 1937–1938 or anything like this figure, or we are assuming most improbable percentages of men dying.”9 The key word here is “improbable.” The Stalin epoch is replete with what appear as improbabilities to minds unfitted to deal with the phenomena. Similarly with the argument that Stalin could not have killed millions of peasants, since that would have been “economically counterproductive.” Following such leads, a new group of Westerners came forward, with singularly bad timing, in the mid-1980s and told us (in the words of one of them) that the terror had only killed “thousands” and imprisoned “many thousands.”10 Such views could only be formed by ignoring, or actively rejecting, the earlier evidence. This was accomplished by saying that those who produced it were opposed to Stalin and Stalinism, and therefore prejudiced, and that some of the material was secondhand. Thus it was not merely a matter of mistaken assessment of the evidence. It was, contrary to the duties of a historian, a refusal to face it.
There were even demographers who, among other errors, accepted the faked census of 1939. A Soviet demographer, deploring this, explains that that census was unacceptable on three grounds. First (as I had already registered in Chapter 16 of The Harvest of Sorrow), the earlier 1937 census had been suppressed and the Census Board shot as spies who had “exerted themselves to diminish the population,” thus providing a certain incentive to their successors to find higher figures than were justified. Second, the 1939 totals were announced by Stalin before the new Census Board had examined the material. Third, censuses of the period omitted the deaths of those who had “died in custody.”11 It is unfortunate that implausibly benign assessments should have appeared in one or two Western textbooks and periodicals and that students should thus have been methodically misinformed. Soviet professionals with whom I have discussed this are, naturally, especially outraged. And, indeed, it is worth noting of glasnost that it has produced a number of articles attacking Western apologists for the regime, from the Webbs on.
Anything like complete accuracy on the casualty figures is probably unattainable. As a Soviet analysis puts it, some records were lost, or never existed. In addition, when it comes to the terror-famine of 1932–1933, it will be almost impossible to sort out the infant deaths from those unborn owing to the decline in birth rate, since registration of births and deaths ceased in the affected areas over the critical period (moreover, under the then procedures, infants dying within a few days of birth were counted as unborn). Nevertheless, it now seems that further examination of the data will not go far from the estimates we now have except, perhaps, to show them to be understated. For example, Sergo Mikoyan, son of the Politburo member, has recently given from his father’s unpublished memoirs a figure reported to the Politburo by the KGB on Khrushchev’s orders in the 1960s: of, between 1 January 1935 and 22 June 1941, just under 20 million arrests and 7 million deaths.12 The respected A. Adamovich has lately criticized me in a historians’ “round table” in Literaturnaya gazeta: “always lowering the numbers of the repressed, he is simply unable to understand the true size of these fearful figures, to understand that one’s own government could so torment the people.” It is true that I always described my figures as conservative; but hitherto, I have been more used to objectors finding them unbelievably large.
In any case, the sheer magnitudes of the Stalin holocaust are now beyond doubt.
In Voprosy istorii, in December 1988, a group of Soviet historians which included the long-rejected Roy Medvedev discussed the Stalin regime; Medvedev argued that once such fearful numbers were reached, it was also a question of the “quality” as well as the “quantity.” It was a matter of