I completely agree with you Nickterry that one cannot state that all EC reports contain exaggerated data. The figures for smaller towns seem to be fairly exact. However, the often very rough guesstimates for larger towns actually make up the significant part of the grand totals. I am also not sure whether there was any significant correction of figures done after March 1946. The first official public disclosure of EC numbers was, as far as I know, in a Pravda article of 24.3.1969, and these numbers were identical with those of 1946.
It is indeed the towns that are the problem. I'm exceptionally lucky in studying Belorussia and Russia since these were simply not as urbanised a region, and thus the exaggerations in the towns are counterbalanced by the
underestimates of famine mortality in the countryside. But it's precisely for this reason that one has to wonder about the Ukraine, since in many areas the urbanisation was higher.
I think that demographic data might help to evaluate the accuracy of EC figures. For the city of Lvov, for example, the total number of civilian losses is given by the EC as 200,000 [Shevjakov, Sociologicheskije issledovanija 1992 (11)]. The population of the city, according to archival data, was as follows [Honigsman, Juden in der Westukraine]:
6.41 – 373,800
11.41 – 327,400
6.42 – 263,900
9.42 – 224,600
3.43 – 271,200
11.43 – 265,500
This decline of roughly 100,000 was not caused by killings alone but also by evacuations, deportations, decline in natural growth rate, etc. I am also not sure whether Jewish victims were exclusively included in the EC figure of civilian losses. Jews deported from the city might have been included in the figure of deportees
This is one possible approach, however one must also pay attention to what camps were located in the cities. Most larger cities were the sites of SD prisons, GFP prisons, police/auxiliary police holding cells, forced labour camps, also in other cases transit camps (like Brest). Not to mention resettlement camps for evacuees from the frontline zone. Just to give an example: the jail in Parichi rayon, a typical southern Belorussian district, was filled with arrestees by the Ordnungsdienst; it then passed on well over half of a thousand prisoners to Bobruisk. Multiply that by several times for the district, and you do indeed have thousands of prisoners being fed into the district towns.
Also, especially after 1942 and the disappearence of much of the Jewish population, peasants are being sucked in to replace the now 'missing' part of the urban workforce.
The population turbulence of the cities under the occupation was, in short, fearsome. However, I do not know the exact extent in Ukraine, whereas for Belorussia and Russia it is something I have observed close-up with the primary sources and is also recognised in the literature (Gerlach, Kalkulierte Morde).
Again, my experiences with Russia/Belorussia indicate that it is precisely
inside these camps that the EC lacked sufficienct insight to get things right. Dare I mention Maly Trostinets claims of 206,000 or even half a million? At least that figure is NOT double-counted inside Minsk province reports.
Re: Lvov, one would of course need to ascertain population before occupation and after liberation. I would assume major evacuations in early 1944 before July. Some of the absolute loss therefore should be transferred to the deportations category. However, you'd have to go case by case through every district to see whether there was extensive repression in the surrounding area, or if the town was used as a transit centre.
(the Soviet figures of "Deportees to Germany" do not only refer to deportees for labour in Germany but include all losses due to forced and voluntary migration, inside and out of the USSR, including service in the German army, resettlement of Volksdeutsche, and even refugees – see [Naumov, Novaja i noveishaja istorija 1996 (2)]).
This is exactly right; it however means the March 1946 figures are an underestimate of the total population displacement. I can document well over a million evacuees from central sector alone between 1942 and 1944.
A good way to distinguish between losses caused under German and Soviet rule would be to compare demographic data for the different periods of the war. I agree that such data is hard to find in the literature. It is apparently available in local archives, as seen above on the example of Lvov.
Kharkov and Kiev have good data from the occupation, available in the USHMM via the local archives, also used extensively by Hilberg and Berkhoff. There is even mortality/birth data for these towns.
I also believe that somewhere, I have a rudimentary census of the RK Ukraine which would be interesting to compare with the 1939 census.
To come back to the Baltic states, there is, for example, some population data known for Estonia [Parming, Population Studies 1972; Tsenzy 1945-1951 gg., Vol. 1]:
1.1.40 – 1,122,000
1.1.41 – 1,117,000 = - 5000
1.1.42 – 1,018,000 = - 99,000
1.1.43 – 1,016,000 = - 2,000
1.1.44 – 1,000,000 = - 16,000
1.1.45 – 854,000 = - 146,000
There are basically two periods of major population declines – that of 1941 (Soviet deportations and evacuations) and that of 1944 (mainly refugees). Apart from this, these Estonian numbers do not really reflect major population losses due to war crimes, starvation, etc. It has indeed been estimated by an Estonian Commission that some 6,000 ethnic Estonians, 2,000 Jews and Roma, and 1,000 others were killed during the German occupation [
http://www.historycommission.ee/temp/conclusions.htm].
I would be very much interested in getting more of such detailed demographic data for other regions of the USSR which were much more affected by German and/or Soviet crimes
The data for Estonia is interesting, thanks. I've added some figures for the reductions in absolute terms, one of course must know the pre-war birth rate to deduce the actual population deficit, demographers can work out to what extent the lack of growth was from unborn children, migrations or deaths.
I mentioned Anton Weiss-Wendt's PhD thesis which you can order via Digital Dissertations if your university had access to it, he has managed to document a higher total by a factor of 2 to 4, and is obviously aware of all the other surveys (he is Estonian by origin himself). However, it is generally agreed that Estonia suffered 'the least' of all three Baltic states; the number of Estonian SS/Police killed in action was also around 8,000, of whom many would have become first 'migrants' then died elsewhere, i.e. were non-returning migrants.
I'd be curious myself for demographic data on Lithuania and Latvia.
For other regions - the Population Studies article and Vallin book (latter in French) has a wealth of information for the Ukraine but is national level only, no regional breakdowns. Belarusian demography does touch on these issues somewhat e.g. references to child mortality, but that's citing via other works, the one I want is in the university stacks and I wasn't in town long enough last week to get it.
To conclude for the moment - I think the EC figures are broadly indicative not of deaths by shootings alone, but deaths by shootings plus a soft number which represents civilian deaths in places of confinement other than POW camps. Such deaths are factored into some but not all EC reports. On top of this mortality which corresponds to murder 1st degree and murder 2nd degree, there is the mortality stemming from starvation and disease among the 'free' population, which is in effect manslaughter.
To take the Ukraine example, of 7 million deaths indicated by the overall demographic data, at least 5 million are non-combatant deaths and thus fall on the manslaughter-2nd degree-1st degree continuum. It's a matter of judgement and further research whether one says 3 million murders and 2 million manslaughters, or 2 million murders and 3 million manslaughters, or 2 million murders 1st degree, up to 1 million murders 2nd degree, and over 2 million manslaughters. Any which way one is still talking telephone numbers.