Famine Deaths, Again

July 30, 2013 6:30 AM

A recent comment by Cormac O’Grada regarding famine deaths in North Korea reminded me how much this issue is misunderstood, even by academic famine specialists. So I thought that I would do a brief re-cap of what we know and what we don’t.

To start, the usual caveat: 25 years ago North Korea was remarkably closed and opaque, and there is simply a lot that we don’t know. A variety of evidence suggests that food availability began declining in the late 1980s; multiple eyewitness accounts attest to a “let’s eat two meals a day campaign” in 1990, in private correspondence, the former East German Ambassador to Pyongyang, Dr. Hans Maretzki, indicated that the campaign started earlier, in the late 1980s.  The onset of the famine is conventionally dated 1994, though there is some argument that mortality rates were already significantly elevated in 1993.  Either date would mean that the famine was already underway when the country experienced floods in 1995. This chronology is critical insofar it undermines the politically convenient depiction of the famine as a product of natural disasters. The famine is generally acknowledged to have ended in 1998.

There have been essentially three waves of estimates of famine deaths. Contemporaneous estimates, sometimes quite high—3.5 million or more were provided by a variety of observers and NGOs (Cormac’s “engagé commentators,”) often with detailed knowledge of the situation on the ground, but lacking professional training in demography or related disciplines. One important exception was the testimony of Hwang Jang-yop, former member of the Standing Committee of the Supreme People’s Assembly, who defected to the South in 1997. Hwang testified repeatedly and in public (go to the 1:12 point in the hyperlinked audio) up until his death in 2010 on the basis of internal Party discussions and documents that “According to the Organization Guidance Bureau official, more than 500 thousand people, including 50 thousand party members, had starved to death in 1995. And as of mid-November 1996, almost 1 million people had already starved to death…That more than 1.5 million people died of hunger from 1995 to 1996 is an irrefutable fact. We do not have accurate data about the situation from 1997 to 1998, but since the food supply did not improved much, it can be deduced that at least a million people have met their deaths every year. According to reports that Chinese press company Xinhua claims it received from officials of the Agricultural Committee in North Korea, a total of 2.8 million people have starved to death at last count at the end of 1997. From this and the irrefutable fact that 1.5 million had died by the end of 1996, we can deduce that another 1.3 million people died in 1997 to add up to a total of 2.8 million deaths by starvation. We have consistently said that more than 1.5 million died between 1995 and 1996 and that 1 million more probably died every year from 1997 to 1998. We never resorted to exaggerating the situation.” Of course, there is no way of verifying Hwang’s claims, and one can speculate on his motives.  Nevertheless, I don’t think that one can dismiss Hwang’s statements out of hand.

The next generation of estimates was done by analysts with greater formal training and more rigorous methodologies. Prominent was work done by Court Robinson and colleagues who analyzed retrospective refugee testimony collected between September 1998 and June 2000 on deaths from selected villages in North Hamgyong province. They concluded that between 1995 and 1997 that 12 percent of the province’s population, or 245,000 people, had died.  Projected onto the whole country (something that Robinson et al. were careful not to do) that would yield an estimate of yield excess deaths of 2.64 million. In my book with Steph Haggard, Famine in North Korea, we argued that for a variety of reasons making this simple extrapolation would be inappropriate, and 2 million excess deaths would seem to be an absolute ceiling, and the actual number was probably much lower.

Demographers Daniel Goodkind and Loraine West, working off the Robinson et al. produced estimates on the range of 605,000-1 million deaths. In his dissertation, Lee Suk, unaware of the Goodkind-West work, using a different methodology came up with estimates also in the 600,000-1 million range. When Steph Haggard and I reviewed this evidence, and additional material we had produced from our own analysis of food availability, we concluded that 600,000- 1 million, or about 3-5 percent of the pre-crisis population, was the most sober, defensible range of estimates.

In a review of our book, Josh Stanton argued that this estimate may have been excessively conservative insofar as death rates may have been higher in South Hamgyong than North Hamgyong. (I can’t help but observing that if you take UNICEF’s latest nutritional survey seriously, South Hamgyong generally scores worse than North Hamgyong—it appears that it is better to be closer to China than Pyongyang.)

“Up to a million deaths” seems to have remained the academic consensus until a series of contributions appeared in the last several years, a paper by Goodkind, West, and Johnson, one produced by South Korean government demographers, another by Spoorenberg and Schwekendiek, and a new paper by Lee Suk, all using variants on the same methodology,.  They take the controversial 1993 census as a starting point; posit a model of population growth; then compare the projections from that exercise to data from the 2008 census.  The shortfall in the actual population numbers from the projections is then allocated over the period 1993-2008, and from that process one obtains estimates of excess deaths during the famine period.  A key point, that comes out most clearly in the paper by Spoorenberg and Schwekendiek, presumably “the two Belgian scholars” to whom Cormac refers in his comment, is that if one allocates few deaths to the famine period (the result that Cormac cites) then by implication one is allocating more deaths to the post-famine period. On the basis of various assumptions, they generate estimates ranging from 237,000 to 420,000 excess deaths. But if these estimates are accepted, then they imply a greater human toll during the 12 years following the famine than during the famine itself! This result comes from some combination of the aftermath of the famine–the long-lived results of chronic malnutrition–and the fact that the economy failed to grow over this period, resulting in a stagnation in life expectancy. Personally, I’m not willing to sign on to these results, but the paper does provide a useful service in reminding us that while the famine may have ended in 1998, its effects continue to reverberate.

The paper that appears to be most informed by actual knowledge of North Korea is by Lee Suk. Somewhat like Spoorenberg and Schwekendiek, Lee treats the entire period as a “slow motion” famine occuring over the whole 15 year sample period, so that his estimates of excess deaths are not directly comparable to the earlier work (including his own) that examined to a more limited period in the 1990s.

Unlike the other papers, Lee takes the problematic nature of the North Korean data seriously, observing that problems of counting males (first identified by Eberstadt and Banister as the “missing male” problem associated with military service) creates anomalies in the apparent cohort-specific survival rates in the two censuses. Lee’s response is to focus on females, where this distortion is likely to be less severe. He calculates excess deaths 1993-2008 among the female population who were over 30 years old in 1993, and finds that deaths were 196,307 higher than expected, or 3.8 percent. If one extrapolates this percentage to the entire population one generates a baseline estimate of 815,000 excess deaths. Lee produces other variants on this calculation which produce estimates ranging from 506,000 to 1,125,000.

So where does this leave us?  Well, much where we started.  Excess deaths during the famine were probably on the order of 600,000-1 million people, or 3-5 percent of the pre-crisis population. I'll leave it up to the reader to decide whether these third wave studies collectively amount to some kind of ersatz rehabilitation of the Kim Jong-il regime.

Add new comment

More From

Marcus Noland Senior Research Staff

More on This Topic