Hi-Stat Vox No. 20 (September 29, 2011)

Examining North Korean Statistical Information


Adjunct Associate Professor, Institute of Economic Research, Hitotsubashi University

Photo : Ho Il MOON


Recently, the Society of Korean Historical Studies (Chosen-shi Kenkyu Kai) in Japan published An Introduction to the History of Korea (Chosen-shi Kenkyu Nyumon), presenting the state of research in Japan on the two Koreas. With regard to the economy, for example, the book highlighted that “research in Japan on North Korea's economy generally leaves a lot to be desired” and that “partly as a result of the lack of information resources, the North Korean economy is a field that has been more or less forgotten by researchers.” These observations apply not only to the economy and, as pointed out by, for example, Wada (1998) already more than a decade ago, research on North Korea has been scanty overall.

Yet, from a security perspective, from the perspective of historical accounts, and from the perspective of an eventual normalization of relations, North Korea cannot be ignored1. Thus, there is a need to gain an understanding of North Korea.

Summarizing the situation, one could say that although there is “demand” for research on North Korea, the quality of the “supply” of such research is not very high. This lack of sufficient research has resulted in a situation where much of the statistical information circulating on North Korea has not been properly verified and is therefore inaccurate or incorrect. It is against this background that in this column I would like to present two concrete examples from my Ph.D. dissertation (Moon 2008): figures on the number of North Korean military personnel and figures on the number of famine victims.

Number of military personnel

Since the country established its armed forces in 1963, North Korea has not published the number of military personnel. While the late chairman Kim Il Sung claimed in 1963 that North Korea had the highest ratio of military personnel in the population overall among socialist countries, he did not mention the absolute number. For that reason, it has been necessary to rely on estimates of the number of North Korean military personnel.

A representative example is The Military Balance published annually by the International Institute of Strategic Studies in Britain. The data in The Military Balance are also formally used by the Japanese government. For example, looking at the Ministry of Foreign Affair's web page on North Korea – http://www.mofa.go.jp/mofaj/area/n_korea/ (in Japanese) – this gives the following numbers on North Korea's military strength: the ground forces have 1,020,000 personnel, the navy 60,000, and the air force 11,000, with the 2011 edition of The Military Balance cited as the source for this information. However, no clear explanation is provided in The Military Balance as to how these figures were obtained and the Japanese government also probably simply quoted them without proper verification.

In actual fact, North Korea has indirectly “published” the number of military personnel. However, it appears that this is not well known. Specifically, North Korea has implemented two population censuses – one at the end of 1993 and one on October 1, 2008. The censuses were supported by the United Nations Population Fund and South Korea's Foundation for Inter-Korean Cooperation, in return for which North Korea was obliged to report the census data. Looking at the published aggregate results of the censuses, an oddity emerges: the sum of the populations by administrative district and the figure for the total population do not match up. The population overall is greater than the sum of the populations by administrative district. The discrepancy in the 1993 census is 691,027 persons, while in the 2008 census, it is 702,373 persons. This means that roughly 700,000 people are not registered in administrative districts.

The reason for this, and hence the discrepancy, is North Korea's citizen registration system. North Korea did not adopt the Family Registration Law from the Japanese colonial period and instead has built a personal registration system consisting of a citizen identification (ID). The citizen ID is a document issued by the public security office (corresponding to the police station in Japan) of the area in which someone resides.

However, according to the Citizen Registration Law, not all citizens are registered in this manner. Article 13 of the Citizen Registration Law stipulates that the citizen ID is to be returned to the public security office where the citizen resides if he or she enlists in the Korean People's Army or Korean People's Guard, or a public security or national security institution, or if his/her citizenship ceases because of death or mental illness. This explains the discrepancy between the total population in the censuses and the sum of the populations by administrative region and implies that the difference of about 700,000 persons consists of those enlisted in the army or public or national security institutions or suffering mental illness.

Looking at the demographic characteristics of these 700,000 persons, it becomes clear that it is highly likely that a majority of them are in the Korean People's Army. The figure below shows the sex ratio by age constructed from the population data by administrative region from the 1993 census. This shows that at the time of birth, there are slightly more boys than girls (the ratio is around 1.05). Moreover, because men have a higher mortality than women the older they get, from around age 30, the sex ratio declines first gradually and then at a faster pace. However, as can be clearly seen in the figure, there exists a sharp dent between age 16 to 26. This dent is entirely due to a decline in the number of men, which make up the numerator of the ratio (with no change in the denominator, i.e., the number of women).

Figure: Sex ratio by age (1993 census)

Age 16, when the dent starts, coincides with the age when compulsory education in North Korea ends. North Korea's education system consists of 11 years of free compulsory education, which starts at age 5 and comprises 1 year of kindergarten, 4 years of primary education, and 6 years of secondary education. When compulsory education ends, some students go on to university, some find employment, and some join the armed forces. And, as mentioned above, according to the Citizen Registration Law, those enlisted in the armed forces are excluded from the citizen registration (meaning that they are excluded from the population by administrative district). Moreover, the majority of those enlisted in the armed forces are men. Consequently, most of the 700,000 persons not included in the population by administrative region are members of the armed forces. If, for example, we assumed that the majority was excluded due to mental illness or death, it would be difficult to explain why the sex ratio returned to close to 1 by age 26.

Of course, the 700,000 people may also include persons that are not military personnel. However, what we can say is that this number of 700,000 persons forms the upper limit for estimates of the number of North Korean military personnel. And this, in turn, implies that the figure presented in The Military Balance is an overestimate.

The original purpose of Moon (2008) had been to construct demographic life tables for North Korea. For this purpose, the reasons for the discrepancy between the sum of the populations by administrative region and the total population as well as the demographic characteristics of the “missing population” were examined. As it turns out, the results of this research received attention not so much from other demographers but from those doing research on North Korea's military. Miyamoto (2009), for example, criticized the estimates published in The Military Balance referring to Moon (2008).

Famine victims

Another area in which there has been considerable debate and in which available data are extremely scanty is the number of victims of the famines North Korea experienced in the latter half of the 1990s due to food shortages and general economic malaise. The first figures regarding the scale of the famine that circulated were that “several million people” had died from starvation. Although there were numerous sources for such claims, the first to provide a statistical underpinning based on a survey and bestow credibility on such claims is South Korea's Korean Buddhist Sharing Movement (KBSM).

The KBSM repeatedly interviewed food refugees hiding in China's border regions with North Korea and based on these interviews tried to estimate the scale of the famine. These estimates were publicized a number of times, and at first it was reported that in the roughly three years from the floods in 1995 until late 1997, about 2 million people had died from starvation (Kyunghyang Shinmun, November 11, 1997), which was later revised to around 3 million (Jiji Journal, March 19, 1998) and finally to 3.5 million (KBSM, “Report based on interviews with 1,694 North Korean food refugees,” December 1998 (in Korean); Jung Koo Kang, KBSM et al., Outcry of the North Korean People, 1999, Jeong-To Press (South Korea)).

This claim that 3.5 million North Koreans had starved to death subsequently received support from a number of testimonies and studies. For example, in 1998, the Director of the International Institute of the Juche Idea, Jang Yeop Hwang, who had fled to South Korea in February 1997, published his personal notes, which contain the following passage (Hwang 1998): “In November 1996, I was extremely concerned about the economic situation and asked a cadre in charge of agricultural statistics and food supply problems how many people had died of starvation. He replied: ‘In 1995, about 500,000 people, including 50,000 party cadres, had died of starvation. And in 1996, an estimated 1 million people starved to death.’ He added: “If there is no international aid, in 1997, a further 2 million people will die of starvation.’” That the total is identical to the 3.5 million stated by DBSM is oddly consistent with the figures quoted by the KBSM.

Yet, there are two major problems regarding the numbers presented by Hwang. First, there are questions with regard to who the “cadre in charge of agricultural statistics and food supply problems” could be. Statistical affairs in North Korea generally fall under the aegis of the Central Statistics Bureau; however, static statistics such as the total population as well as vital statistics such as mortality are aggregated by the Population Division. This Population Division is a relatively new government organization that was only established when the 1993 population census was conducted. Therefore, while Hwang's reference to the words of “a cadre” give the impression that the figures are based on primary sources and can viewed at a glimpse, that “cadre” in fact was “in charge of agricultural statistics and food supply problems” and not primary sources.

Second, there are questions concerning how the number of deaths from starvation mentioned in the story of “the cadre” were calculated. From the perspective of demography, it is difficult to calculate the number of famine victims in the midst of a famine. One reason is that even when there is a famine, almost no one dies directly from starvation. In many cases, death takes the form of a weakening of physical strength and the immune system, and victims die from illness such as an infectious disease. Yet, death through illness occurs even when there is no famine. In practice, therefore, it is difficult to distinguish between deaths due to starvation and deaths due to illness. Moreover, even if it were clear that someone had died from starvation, the death certificate will rarely give “death from starvation” as the cause of death. The International Classification of Disease (IDC) does not include “starvation” as a cause of death, and deaths in North Korea, too, are classified according to the IDC. Furthermore, past international experience shows that it usually takes several years from the outbreak of a famine until the total number of deaths due to starvation can be established, and doing so requires taking factors such as population structure into account. It is therefore inconceivable that in 1997, “a cadre” would or could have already calculated the number of deaths due to starvation in North Korea.

Let us return to the claim by KBSM and examine how the organization estimated the “3.5 million deaths from starvation.” Although the relevant page is no longer available, the KBSM at the time published the contents of its interview survey on its website. According to this information, the survey consisted of interviews with food refugees, based on which family structures by age and sex as well as the mortality rate and cause of death were reconstructed. The estimate of famine victims overall was then obtained by simply applying the figures thus obtained to North Korea overall. The problem with this methodology is that even if the information provided by the refugees was truthful, it is questionable whether this information can be applied as it is to North Korea overall.

Among other things, the survey results showed the composition of interviewees by birthplace, indicating that 80% of the refugees interviewed came from North and South Hamgyong provinces in the northeast of North Korea. Yet, even before 1995, this was the region of North Korea with the highest mortality rate from famine. According to the 1993 census, i.e., before the 1995 floods, the crude death rate (number of deaths per 1,000 people) in North Hamgyong province was 6.5‰ and that of South Hamgyong province 6.4‰, considerably exceeding the average for the country as a whole of 5.5‰. This indicates that the survey by KBSM was subject to a selection bias resulting in an upward bias in the estimation of deaths from starvation. Moreover, recent local studies by the United Nations and others highlight that, compared with the rest of the country, the basis for food production in the northeast is weak to begin with and that the northeast is prone to famine. Thus, the survey by KBSM focused on interviewees from the region with the highest mortality rate and the most prone to famine and then simply extrapolated the results to the whole country.

Taking a very different approach, Moon (2008) estimates the number of famine-related deaths (1995–2000) at 336,000 persons. This estimate is based on the concept of “excess mortality,” which is obtained by subtracting the mortality rate at normal times from the mortality rate at the time of the famine, and can be seen as an unambiguous gauge of the scale of the number of famine-related deaths. This estimate, it appears, has also been accepted by South Korean statistical authorities. In August 2010, National Statistics Office (NSO) asked me for a copy of my population estimates, which I provided, and, on November 22, 2010, published its “North Korea Population Estimates 1993-2055,” which put the number of famine victims at an estimated 336,000 persons. This number is exactly the same as the estimate in my dissertation and it therefore appears that my research contributed to their work.


While in the final analysis Moon (2008) appears to have contributed to highlighting various errors in the statistical information available on North Korea, the original purpose of the study had been quite different.

When seeking to investigate issues related to North Korea, one has little choice but to rely on officially published materials. However, such materials are produced by the ruling authorities, and there is therefore a danger that investigations using such official materials produce a biased image of North Korea produced by the rulers. Even historical research on North Korea conducted in Japan displays certain biases reflecting history as written by the rulers.

Yet, as E.A. Wrigley (1969: 12), an authority in the field of historical demography, has observed: “Historical demography deals with all men and women, not simply those who were powerful, well-born, wealthy or literate. By analysing parish registers, listings of inhabitants, returns made to census authorities and the like, we can look into the lives of ordinary people in the past, comparing peasant with gentleman, miner with clothmaker, countryman with city dweller, and so on. Where the necessary records are preserved there is a chance to get down the roots of society almost as a social anthropologist tries to gain an insight into a contemporary community by listening to its members tell of the great events of their lives of birth, marriage and death and the cluster of social attitudes, customs and sanctions which relate to them. The parish registers and other local documents can often provide eloquent witness to the effects on the lives of ordinary people of local economic or social conditions even though no dialogue with the dead is possible.”

By observing the life of the common people of North Korea from a demographic perspective, I would like to examine North Korea from the side of the ruled – this is the ultimate aim of my research on North Korean demography. Moon (2008) will soon be published as Demographic Trends in the Democratic People's Republic of Korea (in Japanese, published by Akashi Shoten), which includes revisions and updates based on the subsequently conducted 2008 census. I hope that this book transmits more clearly the aims and results of my research looking at North Korea from the side of the ruled.


  1. Tatsuo Endo, the Japanese ambassador in charge of negotiations for normalizing diplomatic relations with North Korea, for example, has given the following three reasons why it was desirable to normalize relations with North Korea: (1) to draw a clear line under the unhappy past between Japan and North Korea; (2) to guarantee Japan's security; and (3) because the two are geographically and historically close countries (Endo 2004).


Choson-shi Kenkyu Kai (ed.) (2011), An Introduction to Research on Korean History, University of Nagoya Press, Nagoya (in Japanese).

Tetsuya Endo (2004) “Negotiations on the normalization of diplomatic relations between Japan and North Korea: A consideration of their purpose and issues,” Sekai Shuho, September 28 (in Japanese).

Wada, Haruki (1998) North Korea: The Present Situation of a Commando Unit Nation, Iwanami Shoten, Tokyo (in Japanese).

Miyamoto, Satoru (2009) “North Korea's military strength seen from the 2008 population census,” The Japan Institute of International Affairs, online: http://www.jiia.or.jp/column/200903/02-miyamoto_satoru.html (in Japanese).

Moon, Ho Il (2008) Demographic Trends in the Democratic People's Republic of Korea: Changes in Demographic Trends and Their Reasons, Ph.D. dissertation, Hitotsubashi University.

Hwang, Jang Yeop (1998) North Korea: Truth and Fiction, Kobunsha, Tokyo (in Japanese).

Wrigley, Edward A. (1969) Population and History, Weidenfeld and Nicolson, London.

Original text in Japanese