The government of North Korea regards economic statistics as state secrets, which makes the country's economy difficult to study. I do careful survey research on the North Korean economy by surveying defectors, Chinese enterprises, and South Korean firms. Still, North Korea is so opaque that when I am asked where I get my data, I normally reply, "I make it up." And I'm only half-joking.
Others seem less than half-serious. Last month, the South Korean news agency Yonhap ran a story about a report from a major South Korean think tank stating that North Korea's GDP grew 4.7 percent in 2011. That think tank, the Hyundai Research Institute, used a combination of United Nations infant mortality data for 198 countries over the 2000-2008 period and North Korean crop data to estimate annual North Korean per capita income. While infant mortality and food availability correlate with income, one cannot meaningfully estimate year-to-year income changes with these two pieces of information alone.
Spurious precision is not unique to the study of North Korea, and this example isn't even the worst -- in the 1990s, two Japanese researchers claimed they had calculated North Korean per capita income down to the dime. But nowhere else is the gap between what we know and what we think we know so wide. I cannot remember the last time the North Koreans published a budget containing levels (i.e., actual numbers) and not percentages (e.g., crop yield grew 5 percent). Nicholas Eberstadt of the American Enterprise Institute once visited the Central Statistics Bureau in Pyongyang; the bureau told him that its statistics were elastic, "like rubber bands."
Whatever your metaphor, North Korean statistics are fragmentary, subject to gross error and even intentional deception. Although one can employ statistical techniques to root out obvious anomalies and impose internal consistency, in the end, analysis of North Korean data is as much about forensic art as science, about acknowledging that some statistics deny precision.
In principle, we should be able to at least get a sense of North Korea's external relations by examining the "mirror statistics" of the country's trading partners -- adding up what other countries say they import from North Korea, for example, to calculate its exports.
But even such an apparently simple exercise is fraught. Every year some country around the world reports an amazing spike in trade or investment with North Korea, consisting mostly of North Korean cell phones and automobiles. Has Apple opened an iPhone factory in Pyongyang? No, someone in the bowels of some national statistical agency has confused North and South Korea. Both Mexico and Austria have done so in the past; Denmark confused the two in its 2009 outward investment data. The disparity in the magnitudes of North and South Korean trade volumes is so vast, with North Korea trading very little and South Korea trading very much, that a single such instance of misidentification can completely distort North Korea's trade statistics. Our best-guess estimate for North Korea's total trade in 2010 is about $7 billion. Mexican trade with South Korea is $15 billion. So if a major trading country confuses the two Koreas, the misreporting can swamp reality.
The most widely cited source on North Korean trade, the South Korean government agency KOTRA, carefully screens the mirror data for such obvious anomalies. But KOTRA excludes the North's trade with South Korea on the constitutional grounds that inter-Korean trade is domestic. Because of budget cuts, or a desire to downplay North Korea's Middle East connections, KOTRA also ignores trade with many Middle Eastern countries like Algeria and Saudi Arabia, both of which report trade with North Korea to the U.N. statistical agencies. As a result, KOTRA greatly exaggerates the prominence of the trade partners that it does record, with important geopolitical implications. The New York Times and the Washington Post, for example, have both reported that China accounts for about 80 percent of North Korea's trade. The actual figure is roughly half as much, which means North Korea is a lot less economically dependent on China than those figures imply.
How about internal data? The most widely cited sources are the South Korean government and the U.N. agencies. The former produces estimates of North Korean national income while the latter produces estimates on things such as agricultural output and food aid requirements. In recent years the U.S. government has downscaled its efforts in these areas and no longer provides much of an alternative to the South Korean and U.N. figures. These analysts face the challenge of analyzing an economy swathed in secrecy, a challenge compounded by the politicization of their efforts. Thus, their estimates come with a wide margin of error.
It is doubtful that anyone, including the people inside the North Korean government, knows the real size of the North Korean economy. Since the centrally planned economy went into decline 20 years ago, a significant share of economic activity has moved into the semi-legal, largely unregulated market economy. North Korea has never been egalitarian, but over the same period, an enormous gulf has opened between lifestyles in the capital, Pyongyang, and conditions in the hinterland. Little hard evidence exists to quantify these developments. The North Korean government restricts aid agencies' access to the population (and even restricts the number of Korean speakers in agencies' resident missions), making conventional assessments impossible. When the U.N. World Food Program began doing household surveys and focus groups, the government quickly shut down the effort, afraid that too much information would leak out.
The South Koreans construct their estimates by making estimates of physical output based on satellite imagery, and possibly spies or informants, and then aggregate the physical output according to a secret input-output table. This estimate is then subject to interagency discussion before the South Korean central bank publicly releases the figures.
This process is not particularly transparent and appears vulnerable to politicization. In 2000, the central bank delayed the announcement of the estimate until one week before the historic summit between South Korean President Kim Dae-jung and North Korean leader Kim Jong Il. The figures implied an extraordinary acceleration of North Korea's growth rate to nearly 7 percent. This had never occurred before and has not been repeated since. Under current South Korean President Lee Myung-bak, a conservative, the central bank's figures imply that the North Korean economy has barely grown at all.
The work of the U.N. agencies appears similarly problematic. Probably the most policy-relevant information released by the United Nations are estimates of North Korea's food aid needs produced by the Food and Agriculture Organization (FAO) and the World Food Program (WFP). On the supply side, the FAO is diplomatically obliged to take seriously the official North Korean production numbers, which appear to exaggerate harvests in good times and understate output when aid is needed. The demand-side calculation of how much food the North Koreans need is open to challenge on a variety of grounds, most notably uncertainty about the centrality of grain in the North Korean diet. The U.N. agencies have become increasingly transparent and open about the possible shortcomings of their calculations; previous reports would include a figure and then merely state that it had not been scientifically observed. Still, the FAO-WFP numbers imply that North Korea has been in continuous famine for a decade, something that no credible observer would claim.
Sadly, we don't even know how many North Koreans there are, how many serve in the armed forces, how many died in the famine, or how many have fled their country. For some purposes this is fine -- we can count their tanks even if we don't know how they pay for them. But for others -- how much aid should we provide, are sanctions likely to work, and ultimately, is time on our side or theirs -- we would be wise to ask the questions "what do we know" and "how do we know it" when formulating policy on North Korea.
Former U.S. Vice President Walter Mondale once counseled me that anyone who claims to be an expert on the North is a liar or a fool. My corollary is: Don't trust any figure on the North Korean economy that comes with a decimal point attached.