June 29, 2018

Understanding Holodomor loss numbers


The controversy about the number of Holodomor losses has divided parts of our communities, both in Ukraine and abroad. Some support the 7 million to 10 million estimate, while others consider 4 million as a more valid figure. The first group relies on the research spearheaded by Prof. Volodymyr Serhiychuk from the National Taras Shevchenko University and his associates; the second group relies on research conducted by a team of Ukrainian and U.S. demographers and historians at the Ptoukha Institute of Demography and Social Studies and the Institute of History in Ukraine, and the University of North Carolina at Chapel Hill in the U.S. 

Evidence is provided on the misleading origin of the 7 million estimate and problems with the different attempts to justify it. The choice is between a figure based on research that ignores universally accepted scholarly principles and can be easily discredited due to its unrealistic implications, and a figure based on research conducted according to Western scientific standards, independently reviewed by experts and published in Western peer-reviewed journals and books.

It is important to clarify first our understanding of the Holodomor and Holodomor losses. The Holodomor is the Famine in Soviet Ukraine during 1932-1934 (it has been shown that there were also Holodomor-related deaths in 1934). Holodomor losses are deaths that occurred during this period in addition to the “normal” deaths that would have occurred had there been no Famine. Losses of Ukrainians outside of Soviet Ukraine are reported separately.

What is the origin of the 7 million losses figure? Prof. Serhiychuk cites two sources in support of the 7 million figure: a) estimates by German diplomats in Ukraine in the 1930s and an article by the Ukrainian agronomist-statistician S. Sosnovyj published in the Kharkiv newspaper Nova Ukraina in 1942 and reprinted in the diaspora newspaper Ukrayinski Visti in 1950. Subjective estimates by foreign diplomats in Ukraine and witnesses of the Holodomor have little statistical value. Sosnovyj’s was probably the first estimate based on an analysis of demographic statistics and is much more credible.

Someone in the diaspora picked up the following sentence from Sosnovyj’s article: “Thus, it seems the population shortfall in Ukraine that resulted from the 1932-1933 Famine comes to 7,465,000 individuals,” without a careful reading of the whole article, and a new figure for Holodomor losses was born. 

Prof. Serhiychuk uses the following quote from the Sosnovyj article as evidence for his “at least 7 million” estimate: “… of the total 7.5 million population deficit caused by the famine, 4.8 million died due to hunger (1.5 million in 1932 and 3.3 million in 1933), and 2.7 million are due to decline of population growth after the Famine.” In spite of this citation, he uses the 7.5 million figure instead of Sosnovyj’s 4.8 million. 

He also ignores quotes in the same article that further contradict his “at least 7 million” estimate: “We’ve heard of significantly larger figures. Six, eight and more million have been claimed. Undoubtedly these are exaggerated numbers. When considering only those who died in 1932-1933, and not counting the decrease in population growth after 1933, then the number of deaths caused directly by the Famine is, as mentioned earlier, close to 4.8 million, possibly 5 million people.”

Thus, the 7 million figure was originally based on a misunderstanding and later by attempts to have it legitimized with a misrepresentation of Sosnovyj’s analysis.

Problems with the methodology of trying to justify the 7 million figure have been analyzed by several scholars. Three of these problems are described briefly here: a) basic estimation method; b) faulty statistics; c) claim that our estimate missed many deaths.

a) The estimation of the 7 million losses consists basically in taking the difference of two populations at different times. This difference includes, besides losses, deaths due to natural causes and births and, if net migration (outmigrants minus inmigrants) is not zero, this difference distorts the estimation of losses. If net migration is positive, outmigrants are counted as deaths or losses and losses are overestimated; if net migration is negative, losses are underestimated.

b) A key figure used to justify the seven million estimate is the 32.7 million population of Ukraine for January 1, 1932, taken from the 1933 Statistical-Economic Compendium. However, as stated in the compendium’s introduction: “All the work on this compendium was done under an urgent timetable… and it was not possible to thoroughly verify this information on site… we cannot guarantee full confidence in and completeness of the indicators published in this compendium.” The reliability of this figure is questionable. It is a well-known fact among demographers and statisticians that yearly statistical compendiums contain approximate data; the only proven method for a reliable estimation of the size of a country’s population is a census.

c) The claim that the Ukrainian-U.S. team failed to include a whole list of deaths like unregistered deaths in peasant’s houses and on city streets, shootings in prisons, deaths of children in orphanages, corpses in anonymous graves, etc. is incorrect. It is a hopeless task to come up with a comprehensive list of all the possible kinds of deaths and to try to count them. Demographic techniques were applied to estimate the total number of deaths during the Holodomor period, which includes large numbers of unregistered deaths. For example, the number of registered rural deaths in 1933 was 1.6 million, while the estimated total is 3.8 million deaths.

Besides specific problems with the attempts to justify the 7 million estimate, the basic research approach used by Prof. Serhiychuk has serious structural flaws. First, more than 20 international scholars (Frank Lorimer, Robert Conquest, Massimo Livi-Bacci, Jacques Vallin, etc.) have tried to estimate Holodomor losses during the last 30 years, and practically all their estimates are in the 2.8 million to 5 million range. However, all this research has been dismissed without any analysis. This violates a basic prerequisite of any research that requires a systematic review of what has been done so far. 

Second, this research was dismissed with the explanation that it is based on documents from Moscow archives that contain falsified data. This implies that: a) all research about the Soviet Union in the last 50 or more years is incorrect or at least suspect; b) most, if not all, Soviet statistics were falsified. The first implication challenges the validity of a significant part of research on the Soviet Union in the last half century; the second one defies common sense. 

To review all the numbers produced by a government on a daily basis and falsify most of them would require an enormous bureaucratic apparatus. Data were falsified in the Soviet Union only if there was an important reason for this; most demographic statistics were collected and analyzed professionally and not tampered with. As an example, the 1937 census was conducted correctly and documented the extraordinary population losses caused by the 1932-1933 Famine. The results of the census surprised the Soviet regime, but it was too late to try to falsify its results. The census was declared defective and access to the data closed until 1988. During the next two years a sophisticated program of falsification of the 1939 census data was prepared, and the results were falsified before they were published.

Third, many Holodomor-related documents were deliberately destroyed in Ukraine in June of 1941, among them most of 1930-1939 birth and deaths summary registration reports for Ukraine, but copies of these reports can be found in Moscow archives and they were not falsified. These data are essential for the estimation of Holodomor losses. 

Fourth, usually scientists start with a hypothesis and then try to validate or reject it using rigorous data-based analysis. Proponents of the 7 million estimate reject a priori the possibility that the number of Holodomor losses may be smaller than 7 million. 

Fifth, the use of “scientific” conferences with papers supporting predetermined results and then “approving” these results by resolutions is contrary to scholarly practice in the West. Such conferences are not scientific events; they are political or ideological events and may result in unscientific conclusions. For example, the following statement is found in the resolutions of one of the conferences in support of the 7 million to 10 million estimate, held in Kyiv on October 4, 2016: “Also demographer’s definition of ‘excess deaths’ as Holodomor losses (3,942.5 thousand persons) was criticized by presenters, as well as their non-inclusion of 1,608.6 thousand, according to them, natural deaths during the Holodomor period, in the total number of losses….” The logical implication of this statement is that, had there been no Famine there would have been zero deaths in Ukraine in 1932 and 1933. 

On the other hand, in his article in this newspaper on February 11, (https://www.ukrweekly.com/uwwp/holodomor-the-ukrainian-genocide-remembering-and-counting-the-losses/Serhijchuk), Serhiychuk admits that deaths due to natural causes are not part of Holodomor losses: “…the total number of losses during the Holodomor was 7,040,800 people. Assuming some 1 million died in 1932-1933 from natural causes, then 6,040,800 was the loss from the Holodomor in the Ukrainian SSR alone.”

Sixth, the proposed strategy of estimating the number of Holodomor losses has an intrinsic contradiction. On the one hand, global estimates of Holodomor losses are derived by taking the difference of two population figures, while at the same time it is stated that the only way to come up with an exact estimate is by adding all the different kinds of deaths, from excavating corpses in common graves to counting the number of peasants shot while trying to escape to the West across the Dnister River. 

Seventh, regarding the alleged 3 million losses among Ukrainians in Russia, there were 7 million Ukrainians in Russia in 1926. If one assumes that 1932-1933 Famine losses for all of them were as high as Holodomor losses in rural Ukraine, 16 percent, an unrealistically high assumption, one would have 1.1 million losses of Ukrainians in Russia, not 3 million. 

Eighth, besides being based on faulty methodology and data, the 7 million to 10 million estimate of losses during 1932-1933 (7 million in Ukraine and 3 million Ukrainians in Russia), has several unrealistic implications:

a) Seven million losses is equivalent to 23 percent of the total population of Ukraine in 1933. This is a staggering level of mortality for a country, exceeding by far the highest levels of losses in any man-made famine and is not credible. 

b) The extra 3 million deaths (7 million minus our estimate of 4 million) during 1932-1933 have to be compensated by a corresponding increase in population between 1934 and 1939, as the 1939 census population is fixed. Possible scenarios for this compensation are: Ukrainian women produced an additional 3 million births between 1934 and 1939 and/or 3 million persons (mainly Russians) migrated to Ukraine during this period. The first scenario implies that after 1936 each woman in Ukraine would have to have on average 20 births, which exceeds the biological maximum of 17 births per woman. Census data clearly discredit the second scenario. Russians settled almost exclusively in cities of Ukraine during this period. There were 1.3 million Russians in urban areas of Ukraine in 1926 and 2.6 million in 1939 – a far cry from 3 million.

Finally, Prof. Serhiychuk keeps insisting on the 7 million to 10 million estimate, while he has proposed new lower figures in his February 11 article in this newspaper: “Why do we insist today that the loss figure is at least 7 million people? Because more than 6 million died from the Holodomor in the Ukrainian SSR alone. …At least another 1 million Ukrainians died as a result of Famine outside Ukraine, the Kuban region and the like.” Thus, all of a sudden there are 6 million instead of 7 million losses in Ukraine and 1 million instead of 3 million losses among Ukrainians in Russia.

The irony is that there is no need to have an unrealistically high number of losses in order to document the magnitude of the Holodomor tragedy. Our research has produced a number of quantitative indicators that characterize the Holodomor as one of the worst man-made famines in human history: a) the 4 million losses is equivalent to 13 percent of Ukraine’s 1933 population, one of the highest levels of mortality among all man-made famines; b) in some oblasts this percent approached 25 percent and in some raions more than 40 percent; c) close to 85 percent of all rural excess deaths occurred in the first seven months of 1933; d) the number of losses in rural Ukraine increased 10 times during this period and in some oblasts this increase was 15-fold; e) at the peak of the Holodomor, June of 1933, the daily average number of excess deaths was 28,000; f) more than 30 percent or 1.2 million of the Holodomor losses were children under age 10. All these indicators provide conclusive evidence of a man-made famine with staggering consequences.

In sum, although the number of losses is not a necessary condition for recognizing the Holodomor as genocide, it has important symbolic and political value. The adoption of a figure that has no scientific validity and implies evident demographic anomalies is counterproductive. Additional documents may provide elements for refining our estimate by a maximum variation of +/- 10 percent, but a 75 percent increase (from 4 million to 7 million) is demographically implausible and, as was pointed out, is contradicted by unrealistic implications. A figure validated by the international scientific establishment provides a solid and credible basis for the Holodomor narrative. Quantitative indicators based on this figure prove beyond reasonable doubt that the Holodomor was one of the worst tragedies of the 20th century. 

Prof. Serhiychuk’s discovery of new documents is very valuable, as it provides new insights on the tragedy of the Holodomor. However, practically all of these documents have no relationship to the problem of estimating Holodomor losses. As our experience has shown, a more productive collaboration between demographers and historians is that of complementarity. Each discipline has its unique strengths and their combined efforts provide the basis for a more productive research strategy. 

A fixation on the number of Holodomor losses has important drawbacks. The absolute number of losses is not the best indicator for establishing the magnitude of Holodomor’s tragedy, as it is related to the size of the country’s population. A more adequate indicator is the relative number of losses in terms of percent of the relevant population. For example, the Ukrainian-U.S. team has estimated as 3.3 million the number of 1932-1934 Famine losses in Russia, and 3.9 million in Ukraine. Does this mean that the magnitude of the Famine was similar in Ukraine and Russia? Absolutely not. The estimate for Russia is equivalent to three percent of its 1933 population, while the percent for Ukraine is 13 percent – that is, almost four times higher. 

To cite another example, if one goes by absolute numbers of losses the Famine in Communist China caused by Mao’s Great Leap Forward acquires extraordinary proportions due to China’s extremely large population size, while the magnitude of the Armenian genocide seems minuscule in comparison. But if one standardizes these losses by the relevant population sizes, the magnitude of Armenia’s genocide becomes as significant as the Chinese famine.

The government of Ukraine and our diaspora communities are faced with a difficult decision. It is understandable that some persons accustomed to the 7 million figure have a hard time accepting a lower figure. Their sincerity and patriotism are unquestionable; by the same token, it is inappropriate to label as unpatriotic persons who believe that scientific methods should take precedence over ideology when dealing with such a sensitive and important issue as the number of Holodomor losses. History is full of myths. Often, later research based on more detailed data provides more accurate information and rectifies some of these myths without diminishing the importance of the historical events related to these myths.

It is standard practice in Western countries that when a government has to make a decision on a scientific matter, it asks its National Academy of Sciences to convene a panel of experts on the subject to study the problem and make a scientifically based recommendation to the government. This would be a rational solution to this very important issue.

AUTHOR’S NOTE: For the interested reader, the full list of publications by the Ukrainian-U.S. team, in English and Ukrainian, can be accessed on the website http://www.inform-decisions.com/holodomor. 

Oleh Wolowyna is research fellow at the Center for Slavic, Eurasian and East European Studies at the University of North Carolina in Chapel Hill.