“Famine in North Korea:” An Interactive Review (1 of 3)

noland-haggard.jpgThe time stamp on this post may be the most telling part of it, for I first got my hands on Marcus Noland and Stephan Haggard’s “Famine in North Korea:  Markets, Aid, and Reform” back in late March.  The intervening months have been very busy for me, and the book raised more points of discussion than I can cover here. 

Noland and Haggard  are two of the finest, most respected scholars of all things North Korean and economic, and their work does not make for light reading.  They do not attempt to give us  a sense of what the famine was like or emote  about its human toll.  Their analysis is dry, factual, and as objective as any book on this subject can be.  That approach will stop you at Chapter One if you’re looking for a North Korean “Harvest of Sorrow” or “Hungry Ghosts.”   This is not that  book; instead, it is a  record of economic history, a first draft of a statistical indictment, and the  fusion of several scholarly papers  meticulously  and dryly documenting why this famine didn’t have to happen. 

You will see that I also have some criticisms of this book, but you’ll get an idea of its importance by the amount of thought I’ve put into this review.  At its best, “Famine in North Korea” reflects the meticulous character of the men who wrote it and the depth of their knowledge of  this subject.  That is  more than enough  make this book well worth its price.

As I wrote this, I  had also contacted the authors’ publicists, seeking their comment on this review, and asking  some questions that their book raises.  I am now pleased to report that both Stephan Haggard and Marcus Noland  have read the following review and have provided a response, and I offer them my deepest thanks for honoring my thoughts with theirs.  I will publish this review in two parts.  Their response will be Part 3, so aside from comments and responses, they will have the last word.  With recent flooding raising fears of food insecurity rising in the North, their response carries added relevance today.

I.  Crime

nkfamine_children-40.jpegThe two most important questions I brought to this book were how many died, and why.  Let’s address those questions  in inverse order.  The exceptional  Havel-Wiesel-Bondevik report  does not think that the famine was intentionally engineered like the Ukrainian famine (with the exception of labor camp inmates, whose rations were clearly kept at a starvation level) [55].  I’ve gone back and forth on this question myself. 

Noland and Haggard don’t seem all that certain, either; they struggle with conclusions about intent and  culpability.   This is just  the first of many times in their book when you suspect  Noland and Haggard aren’t necessarily in perfect agreement.   Early on, Noland and Haggard warn us that  “outsiders’ first response to genocides is denial,” [2]  though  they never  quite levy  that charge.  Instead, they  make an  extensive statistical case that removes all reasonable  doubt about the most important question.  The North Korean famine was not ultimately caused by any  lack of resources, but by a misallocation of those resources:

As is immediately obvious, aggregate supply [of food]  exceeds [estimated minimum needs] for the entire decade of the 1990s regardless of the production numbers used (emphasis in original). [Page 46]

Noland and Haggard  believe that  North Korea’s “songbun” system of political castes “did not directly determine access to food, but had powerful indirect effects” such as the right to live and work in places where food was relatively abundant. They document how the regime directed multiple food shipments from the WFP, China, and South Korea to its West Coast, accessible to areas that were politically favored, instead of sending them to the East Coast, where the need was greatest [64-65]. No area was spared completely, of course; even in Pyongyang, the authors report that 4% of citizens were “wasted” from malnutrition [198; see maps on 201-02].  In the end, one hardly needs to know exactly what Kim Jong Il had in mind:

The state’s culpability in this vast misery elevates the North Korean famine to a crime against humanity. [208]

Then there is the question of priorities.  The authors  document how a fraction of the regime’s military spending could have fed feed its people and how it  responded to the arrival of international food aid by purchasing more unspecified “commercial” imports [44-47]. 

The greatest value of this book may be how conclusively it destroys the reprehensible falsehoods spoken in Kim Jong Il’s defense, and in favor of artificially sustaining his rule [67].  Crude apologists  like Christine Ahn parrot the party-line excuse that the weather and an American “embargo” caused the famine.  Sophisticated apologists like John Feffer blame  the weather and North Korea’s lack of  good agricultural land, but try to absolve Kim Jong Il of “culpable slowness” in reacting to the famine.  To Noland and Haggard, the question is more than one of production:

This problem could probably be solved purely by expenditure switching: shifting the composition of imports away from other priorities. But even taking existing expenditure preferences as given, the improvement in the performance of the export sector needed to address this constraint are modest: the annual import shortfall is not large, in the hundreds of millions of dollars. [211]

In plain terms: food is cheaper than Omega watches and missiles [44]. 

II.  Reform

famine.jpgThe most insightful moments of “Famine in North Korea” describe the relationship between markets and hunger, the unequal distribution of hunger, and how the North Korean authorities failed ““ sometimes, it seems, intentionally ““ to make economic adjustments to mitigate this lethal misallocation.  To Noland and Haggard,  North Korean “reform” is a misnomer for  the spontaneous, bottom-up  development of markets by people  who would have starved without them. 

In 2002, the state  temporarily  ratified some of the least reversible changes in an effort to regain control over an economy that was largely extralegal [16].  The tolant mood didn’t last.  By 2005, the regime had regained enough control to transition from half-measures to  drastic reversals and punitive confiscations, though the  ideological purity  of local officials appears to vary.  The most  bizarre scheme?  The coercion  of starving citizens into purchasing bonds that operated more like a lottery, with only a lucky few getting back their money, plus a prize [190].

III.  Statistics

ryongchon-street-scene.jpgNoland and Haggard’s death toll estimate for the  North Korean famine, between 600,000 and 1 million, is  among the lowest offered by experts on the subject.   I don’t believe Noland and Haggard are trying to minimize anything, but  they’re statistical outliers.  Unfortunately, the book never explains how these men, who who deal in statistics with such exceptional sophistication, arrived at their own.

Noland and Haggard  survey, discuss, and criticize other estimates in great detail, but ultimately favor two that are based on age- and gender-specific population models  based on a 1998 WFP nutritional survey and death rate statistics from China’s Great Leap Forward.   Then they stop there, without explaining  how they applied those models to North Korea.  Instead, they summarily conclude that “in our view, the most sophisticated attempts to measure excess deaths put them in the range of 600,000 to 1 million.  [76]    

There are some problems with this model, however.  First, as the authors admit,  North Korea’s economy was largely an industrial one by 1994, while  China’s was mostly agricultural  [211].  Second, one of  the studies they favor  calculates mortality only for  1995  through 2000 [75], thus truncating mortality before and after that period.  Finally,  there are some obvious  questions about  any estimate based on  official DPRK statistics  or the WFP’s 1998  nutritional survey.   Both must have been  subject to North Korean government control and manipulation.  For example, maps on Pages 93 to 95 purport to show the areas where the WFP was permitted access, yet the “access” areas indicated after 1996 include North Korea’s worst concentration camps, including Camp 22, which is easily large enough to show on those maps.  No international inspector has ever been let near those camps; the regime denies that they even exist.  There’s little question that  the mortality rate in those areas would have been both severe and unreported.     

Let’s return to the  problem of using an  agricultural economy’s model for an industrial economy, because the concentration of people away from their food sources poses  what I’ll call the problem of the “goners.”  In any famine, people wander  away from their homes in  search of food, but  the problem is greatly  concentrated for large numbers of people  concentrated in factory towns like this one, as  opposed to smaller concentrations of people surrounded by other farming areas.  According to Andrew Natsios, “goners” was  a  grim  expression  aid workers applied to wandering famine  refugees  few were likely to  survive.   Defector Yomiko Chiba vividly described  the  arrival of scores of goners in  Sinuiju in 1995.  They wandered in from the countryside and began showing up dead on the town’s streets, which concerned local officials.  Chiba was a teacher,  so the regime mobilized her and her students to collect and bury cartfulls of goners of all ages and both genders.   Natsios also  wrote of having  witnessed mass burials of famine victims in unmarked graves and described defector accounts of daily collections of corpses at railroad stations. 

I do not relate these stories simply because they are grim.  I relate them to raise the strong possibility that because their identifies and causes of death  may never have been recorded,  “goners” might not show up in death statistics would likely exceed the number of “goners” in mostly agricultural areas, such as 1950’s China.   Goners from cities  could also  have a  ripple effect  on nearby agricultural areas and  areas along roads, railways, and other escape routes.   The North Korean regime also rounded up “goners” without travel permits and put them in 9/27 camps [170], or  perhaps in institutions such as hospitals and orphanages.  In both cases, mortality in those  institutions  was generally higher than elsewhere  [73].   Family members who died after being left behind by their providers might show up in local death statistics, but if no family members survived at all, they might not be reflected in refugee surveys, on which the authors also rely.  We’re left with too little information about the authors’ favored Chinese models to know if they account for those questions.

Perhaps Noland and Haggard can answer them, but regardless of their final toll, they would make a stronger argument had they done more to explain it.  It’s always to a challenge explaining complex statistical analysis to economic novices like your correspondent.  In the end, the estimate I put the most weight on is the one that’s best explained to me.  Here, I still  credit Andrew Natsios, the former Administrator of USAID, current Special Envoy for Darfur, and a former aid worker in North Korea.  Natsios  estimated the toll at 2.5 million in “The Great North Korean Famine,” a extraordinarily powerful book that will inevitably, and perhaps unfairly,  be compared to this one.  

Noland and Haggard take issue with Natsios’s weighted estimate because it places too much statistical weight on North Hamgyeong, in the far northeast, which was “by consensus, the worst affected province. [75]    That was my guess, too, but  Noland and Haggard  cite a refugee survey’s finding that 62% actually thought that South Hamgyeong province was the worst affected [70], though 60% of those surveyed were from North Hamgyeong.  That makes sense; after all, it’s a much longer walk to China from Hamhung or Hungnam than it is from Chongjin.  And as I’ve mentioned above, North Hamgyeong statistics  almost certainly fail to report death statistics from forced labor camps that occupy a significant portion of the province’s land area.

It’s tempting to declare this whole  discussion hopeless.  Last winter, hundreds died when a heavy snowstorm  cut an entire district off from its food and fuel supplies.  Were these famine victims, too?  I don’t claim to know the answer.  All of these estimates must rely on questionable data, and even if we learn that the actual  toll  matches Noland’s lowest estimate, it still would be the highest death toll caused by government policy toward its own people since the Khmer Rouge fled Phnom Penh.  But there’s nothing academic about this discussion if you’re a policy-maker, a war crimes prosecutor, or a North Korean. Unless you accept Stalin’s maxim that a million deaths is a mere statistic, we are dealing with a discrepancy of at least 1.5 million tragedies. 

[Continued tomorrow …]


  1. Yep, I read a lot of Noland’s work, and it is nothing short of amazing on what kind of data he presents in a very detailed way. You are right, those books are dry and factual, but if one can read through supply/demand side, and a host of other terms, then it is very good reading.

    I am not an economist, but I am very, very interested in how the regime may work, so I made myself understand the odd terms in the papers to gain a better understanding of the economic situation.

    I look forward to the rest of your series.


  2. “Last year, hundreds died when a snow storm cut off food and fuel supplies. Were these famine victims too? I don’t claim to know the answer.” Let me give you the unequivical answer. Yes, they were famine victims. By a government who cared less about delivering food to starving human beings than thier own self preservation. That is also famine related.


  3. Thanks Josh, for the updates – Noland’s and Haggard’s arguments are right as you pointed out – any marketization of the economy was from the bottom up, driven by individuals who acted out of hunger and desperation. Once the DPRK regime regained capacity to do so, it crushed emerging markets –
    All signs seem to indicate that this will continue to be the case unless we do “something”.


  4. Marcus is quoted all the time, and what’s more, almost everyone else who is quoted by the media has been influenced by Marcus’s research in one way or another.