Regime Change and On-the-Spot Guidance
Last year I wrote a post “Fishing with dynamite” based on a paper by by Kim Insoo and Lee Min Yong in which they argued that the frequency of incidents along the Northern Limit Line was more accurately predicted by the vagaries of the fishing industry than by the strategic calculations of Pyongyang. The two are back with a new paper examining the “predictors” (in their words) of the frequency of on-the-spot guidance visits to military and non-military sites.
Kim and Lee point out that despite the initiation of “military-first” politics, the frequency of Kim Jong-il’s visits to military sites did not trend up, but rather fluctuated over time, as did his visits to non-military sites. So what might explain this pattern?
Kim and Lee argue that if on-the-spot guidance is driven by military-first politics, then it should be correlated with the purposes of military-first politics, namely national defense, consolidation of power, preventing social disruption, and economic reform. The authors then specify proxies for each of these motives. If national defense is key, then on-the-spot guidance at military sites should increase then security threats increase. How can this be measured? Kim and Lee argue that US and South Korean aid to North Korea can be interpreted as an inverse indicator of security threats, i.e. when aid is up, threats are down. Similarly, they proxy the consolidation of power motive with the number of promotions to the rank of general-grade officer. Social disruption could be represented by the number of North Korean refugees fleeing the country, but since reliable data is unavailable, they use the number of refugees arriving in South Korea. Lastly, Kim and Lee associate Chinese food aid with pressure from China to reform the economy, and expect that visits to the military would increase as Kim Jong-il sought to mollify this key constituency. They then proceed to model on-the-spot guidance as a function of these four correlates.
The model under-predicts the frequency of on-the-spot guidance following the onset of the second nuclear crisis in October 2002. The model is more tightly correlated with visits to non-military sites than military sites, with US and South Korean aid and Chinese food aid particularly associated with non-military visits.
Lee and Kim interpret their model causally—i.e. aid generates on-the-spot guidance. It is perhaps just as plausible that that the causality runs in the other direction: on-the-spot guidance to non-military sites is interpreted by foreign observers as a signal of reform, and induces supportive aid or that both the visits and the aid are jointly caused by some third driver (an economic downturn induces greater attention to non-military concern as well as prompts compensatory aid).
Of course this analysis pertains to the behavior of Kim Jong-il. It would be of enormous interest to determine whether these patterns have changed under Kim Jong-un. Statisticians have a term for a statistically significant change in the magnitude or pattern of such correlations: regime change. I look forward to hearing from Kim and Lee in some future paper as to whether North Korea has experienced regime change, at least in a statistical sense, or not.