Part of Our World: North Korea’s Exposure to World Commodity Prices
by Sewon Park Harvard College ‘21
ABSTRACT: This paper explores how responsive North Korean production activity is to global commodity prices. I use satellite nighttime lights data, IMF reported world commodity prices, and data from the 38 North Digital Atlas over 1992-2013 to present three main findings. First, luminosity (a proxy for economic activity) in coal-producing counties in North Korea is considerably responsive to increases in global coal prices. The same pattern is revealed for steel. I estimate a 0.957 world price elasticity of coal production and a 2.652 world price elasticity of steel production in North Korea, both statistically significant at the 1% level. This empirically confirms how the North Korean regime has strategically aligned its coal and steel production activity with global coal and steel price fluctuations, respectively. Second, our study finds statistically significant, positive relationships between North Korean coal/steel exports to China and luminosity in coal/steel-producing regions, potentially suggesting that luminosity reflects trade with China as well. Third, I found no statistically significant nonzero relationship between Chinese producer prices and luminosity in commodity-specific producing regions. This indicates that the DPRK-Chinese relationship is far less market-driven and more characterized by geopolitical factors instead. These results demand a recalibration of our understanding of how North Korea maintains its regime, nuclear activities, and evades sanctions through its integrated economy and trade activity.
Introduction
In this paper, I examine how movements in global commodity prices impacted production intensities of corresponding goods in North Korea from 1992-2013. This time period marks a transformative time in North Korean economic history: The Soviet Union collapsed, the country fell into a mass famine, sparking marketization from below, the government began to allow some level of market activity, and critical economic and political linkages with China formed. Contrary to popular belief, North Korea is not a “hermit kingdom” and since the late 1990s has developed global trade networks through a variety of state companies. The North Korean government has taken active steps to increase North Korea’s international trading presence, albeit with a limited number of countries given the nation’s nuclear development activities. How integrated North Korean industries really are to the world economy, however, remains unknown. This limits our understanding of how connected the new North Korean economy is to the rest of the world. Knowing how exposed North Korean markets, production capabilities and export activities are to global prices in turn allows us to better analyze how to implement effective, targeted sanctions measures. I explore how integrated the North Korean economy is with the rest of the world economy by analyzing how North Korean production behavior changes based on fluctuations in corresponding world commodity prices. This paper presents three main findings. First, production intensity in coal-producing counties in North Korea is considerably responsive to increases in global coal prices. The same pattern is revealed for steel. This empirically confirms how the North Korean regime has strategically aligned its coal and steel production activity with global coal and steel price fluctuations, respectively. North Korea is and has been considerably intertwined with and connected to world markets and coal/steel prices. Second, my findings suggest that luminosity may reflect trade with China as well. Third, there is no statistically significant nonzero relationship between Chinese producer prices and luminosity in commodity-specific producing regions. This indicates that the DPRK-Chinese relationship is far less market-driven and more characterized by geopolitical factors instead. I hypothesize that as global prices of coal/steel increased, China bought more North Korean coal/steel because it was offered at lower prices. This illustrates how the North Korean economy has been closer/more intertwined with the rest of the world than we might think: global changes in coal prices impact Chinese importing decisions, which lead to changes in North Korean production and export activities. For the majority of North Korea’s economic history, the juche ideology of nationalist self-reliance has demonized international trade. The country’s external economic relations were historically brittle as the centrally planned state economy and juche rhetoric prevented much international trade (Haggard and Noland, 2017). With the disintegration of the Soviet Union in the 1980s, however, North Korea lost its primary trading partner, which had devastating economic and social consequences. Combined with the failure of its public distribution system, a series of natural disasters, and a mass famine, the North Korean economy collapsed in the 1990s (Park, 2003). It was out of this context that the North Korean economy went through a massive transformation. The country underwent two major changes: the formation of state trading companies that operated on an international scale, and the rise of free market activity within the DPRK’s internal economy. Indeed, in the wake of the famine, collapse of the Soviet bloc, and later on, international sanctions, state trading companies were formed to bring the country out of its downward economic spiral. The DPRK government switched from a “self-sufficient economy” model to a “trade-first policy” to generate foreign currency and secure funds for their own operating budget (Park, 2009). Domestically, given that the public distribution system collapsed and people could no longer rely on the government for basic needs, individuals began to engage in barter and trade for their daily livelihoods. In the following years, the government also provisioned “special economic zones” for limited amounts of market activity and trade. As such, since the late 1990s, even though the North Korean economy was by no means an open international trading partner and still run under a central state-planned system, it had established international state trading companies and supported some level of domestic private market activity. Starting in the 1990s, the Chinese-North Korean political and economic relationship truly began to take form. Figure 1 shows DPRK-PRC total trade from 1992-2013. Initially, in May 1996, Chinese premier Li Peng and North Korean vice premier Hong Song-nam signed an agreement establishing high-level contacts, promising a widening of economic cooperation, and providing large commitments of food aid (Park, 2003). China paid “friendship prices” in buying North Korean goods, translating to implicit food and/or energy aid estimated at nearly $6 billion by 2002 (Park, 2003). Over time, this relationship evolved, with the PRC today importing goods such as coal or steel from North Korea at cheaper prices, given their lower quality and international sanctions limiting North Korea’s available trade partners. Political relationships between the two nations and their working parties have also fluctuated throughout history, impacting economic relations over time. In all, the major trading relationship with China developed over this period, leading to China being by far and away North Korea’s largest trading partner today.
Research on North Korea’s economy has yielded insights into North Korea’s political and economic relationship with China. However, the majority of prior work has been qualitative and anecdotal in nature. Park found through a series of interviews with ROK firms/Chinese businessmen operating ventures in North Korea and North Korean defectors who previously worked in DPRK state trading companies, that sanctions meaning to harm the economy and limit access to WMD-related materials have not worked effectively (Park, 2009). This is because through a variety of trade networks, many involving China, state trading companies have been able to generate funds maintaining the loyalty of North Korean elites and an operating budget to keep the regime afloat (Park, 2009). Despite its international isolation and pressures, the regime is able to survive and cope due to these state trading companies. The trade relationship between China and North Korea has also been carefully analyzed by 38 North, the Korea Economic Institute of America, and other organizations that outline import/export behavior over time, as well as the dependency of the DPRK economy on China. Specific studies on certain commodities, like coal, have tracked how the regime has used trade and exports of that good for sanctions evasion. Pavone and Sun provide a comprehensive analysis of the coal trade and Chinese-North Korean relations as it pertains to coal as a means of sanctions evasion (Pavone and Sun, 2014). Indeed, empirical work on North Korea’s economy and its exposure to the world is limited. The methodologies in the aforementioned studies were primarily empirical and focused on trade reports and expert interviews. Very few papers on the North Korean economy as it pertains to global trade are entirely empirical in nature, to my knowledge. Notably, Lee examined the effect of international sanctions on the regional distribution of economic activity in North Korea over time (Lee, 2017). Using a sanctions index he developed and nighttime lights data from The National Oceanic and Atmospheric Administration under the Defense Meteorological Satellite Program (DMSP), he reported that an increase in sanctions was associated with luminosity increases in the capital city, trade hubs near China, and manufacturing cities. He reveals how China has offset the trade restrictions imposed by other countries and regional inequality has increased as a result of these sanctions (Lee, 2017).
Data
For my luminosity measure, I used Lee’s nighttime lights data. Light intensity is reported in digital numbers ranging from 0 to 63 for each pixel. The data provides a luminosity measure over the years 1992-2013 for around 180 different sigungu (counties). Summary statistics for the luminosity measure are shown in Table 1, with the dlum variable indicating the change in luminosity from the prior year. Lee points to a study by Henderson et al. (2012) which found that “nighttime lights are correlated with GDP with an elasticity of about 0.3” and “nighttime lights is a good proxy for economic output when subnational data are not available” (Lee, 2017).
Then, I used IMF Global Commodity Price Data to get world prices of commodities over time, and China Data Online to get Chinese retail prices and producer price indices of commodities as well. I analyzed a total of nine commodities: coal, steel/iron, soybeans, cotton/textiles, cement, natural gas, timber, fish, and poultry. Chinese prices on soybeans were limited, so that regression is omitted. Figure 2 shows world and Chinese commodity price indicators over time. To get a measure of how exposed output in a certain sigungu might be to changes in world commodity prices, I then recorded the number of production sites in each sigungu for each of the nine commodities. For this data, I used the 38 North Digital Atlas’s information on where specific commodities are produced regionally in North Korea. Figure 3 shows the regional distribution of coal mines in North Korea. Sponsored by the Stimson Center, their digital atlas is arguably the most comprehensive and authoritative public geospatial dataset on production facilities in North Korea. I used their mining, manufacturing/production and agriculture layers, which provides names, descriptions and coordinate locations of different facilities. I matched each facility to its nearest sigungu using QGIS software. From information in the names/descriptions of facilities in the 38 North data, as well as other sources like CIA reports on major energy production facilities in North Korea and USGS surveys, I was able to record the number of coal, steel, cement, soybean, natural gas, textiles, timber, fish, and poultry producing facilities in each sigungu. This data was matched to the luminosity and price data, with each sigungu having the same number of facilities for each commodity throughout the time frame 1992-2013. Because 38 North’s Digital Atlas provided production facilities as seen at a snapshot in time, we do not account for changes over time in building/removal of facilities. Lastly, I used the UN Comtrade dataset to find mirror statistics of North Korean exports to China through reported Chinese imports from the DPRK. This data was limited to only a select few commodities.
Methodology
The base regression that examines how commodity prices over time affect the luminosity of regions in North Korea is:
with i being region, t being time, j being commodity, and Exposure referring to the exposure measure, which in my case is the number of commodity-specific production facilities in that sigungu. This is essentially a measure of the elasticity, as I regress the percent change in luminosity on the percent change in price. I interact the change in price with that exposure measure in a sigungu to analyze how the relationship between change in price and change in luminosity differs for any additional commodityspecific production facility in a specific region. For instance, if a place has no coal mines, its luminosity measure should not change with coal prices. I use year fixed effects, δy , to control for factors that vary across time for the whole country, such as sanctions pressure, global recessions, nation-wide political turmoil, or nation-wide famines. I use county fixed effects, δc , to control for factors that vary across different counties but remain constant over time, like certain regional economy characteristics, geographical proximity to China, region-specific politics, and the like. By taking the year-on-year log differences, I control for any time trends in both luminosity and prices. In addition, I clustered standard errors by county. I ran this base regression for all eight commodities (coal, steel, fish, timber, textiles, natural gas, soybeans, and poultry), regressing both global and Chinese prices separately.
I then included both world and Chinese prices to isolate and separate the effects of each from one another and added the Chinese share of North Korean commodity-specific exports in each year as a control to account for changes in Chinese-North Korean trade patterns. Figure 4 shows how this number has changed over time.
Furthermore, I estimate the same coefficients here but include both coal and steel commodities’ price-exposure interactions. This specification responds to the possibility that counties’ response to a steel price shock is confounded by the fact that global coal prices are correlated with global steel prices. Moreover, given that coal is an input into steel production, an increase in global coal prices could impact North Korean steel production independently of changes in global steel prices. Lastly, I tested how reflective the luminosity measure was as a measure of trade with China by regressing the change in luminosity on the change in commodity-specific trade value with China over time.
Results
How sensitive is North Korean commodity production to the world and Chinese prices of commodities? Table 2 reports the effect of world coal prices on luminosity in counties with coal production facilities. All columns include the Chinese share of North Korean coal exports as a control. Column 1 reports the impact of world coal prices on luminosity independently of the Chinese coal price and Column 2 reports the impact of Chinese coal prices on luminosity independently of the world coal price. The coefficient in Column 1 is 0.216, statistically significant at the 1% level. A percent increase in the world coal price is associated on average with a 0.216 percent increase in luminosity in counties containing coal mines. As such, we can interpret this as the world price elasticity of production being 0.216. North Korean coal production, as measured by luminosity, is indeed sensitive to changes in the world price. The coefficient in Column 2 is not significant, indicating that there is no nonzero relationship between Chinese coal prices and luminosity in coal producing regions. Column 3 includes both world and Chinese coal prices in the regression and yields a coefficient of 0.31 on the regressor world coal price, statistically significant at the 1% level. Here, the world price elasticity of coal production is estimated to be 0.31, a ~43% increase from the coefficient in Column 1, which excludes Chinese coal prices in the specification. The inclusion of Chinese coal prices, along with world coal prices isolates each’s impact on luminosity and increases the magnitude of the world price elasticity of coal production estimate. Still, Column 3 does not yield statistically significant results for the relationship between Chinese coal prices and luminosity. We see a similar trend in Table 3, which reports the relationship between world iron prices and Chinese steel prices on luminosity in steel-producing counties in North Korea. Once again, all columns include the Chinese share of North Korean steel exports as a control. The specification in Column 1 measures the impact of world iron prices, Column 2 measures the impact of Chinese steel prices, and Column 3 includes both world and Chinese prices. The coefficient on world iron prices in Column 1 is 0.533 and is statistically significant at the 1% level, indicating a precisely estimated world price elasticity of North Korean steel production of 0.533. This elasticity estimate decreases slightly in magnitude to 0.497 but remains statistically significant at the 1% level in Column 3, with the inclusion of Chinese prices in the regression equation. Similar to in Table 2, the results for the coefficient on Chinese steel prices are not statistically significant. I performed similar analyses on world and Chinese prices and luminosity in counties with commodity-specific production facilities for fish, timber, textiles, natural gas, soybeans, and poultry. Table 4 shows that none of these commodities yielded statistically significant, nonzero relationships between price and luminosity, indicating that the price elasticity of production, as measured by luminosity, for these commodities is not distinguishable from zero in our dataset. Data availability was also limited for a number of these other commodities.
In Table 5, I include all regressors from Table 2 and Table 3 together and analyze the impact of world coal and steel prices on luminosity in coal and steel producing counties, respectively. I estimate the same coefficients as in Table 2 and Table 3, but now include both commodities’ price-exposure interactions. Column 3 reports that the world price elasticity of coal production is 0.957 and statistically significant at the 1% level when I include the world price, Chinese price and Chinese export share of steel. This is a notable increase from the same coefficient in Table 2, which estimated an elasticity of 0.31. Once steel-related variables are controlled for, the sensitivity of coal production to global coal prices increases considerably. Interestingly, Column 3 also reports that the world price elasticity of steel production is now 2.652 and statistically significant at the 5% level, once controlling for coal-related variables. This is a marked increase from the estimate in Table 2, which was 0.533. This specification reveals that North Korean steel production is remarkably sensitive to global iron prices once coal-related variables are controlled for. To contextualize these elasticity estimates, I compare our regression coefficients to literature on countries’ price elasticity of supply for coal and steel. To my knowledge, such studies are limited. Still, one paper estimates that the supply of US coal has a short run annual elasticity of 0.61 and a long run elasticity of 1.31 (Dahl). The fact that our elasticity estimates come close to this range illustrates that the North Korean coal and steel world price elasticity of production is remarkably high. This is because Dahl’s study estimates the sensitivity of US coal production to US prices, which intuitively is very elastic given that domestic coal production is likely highly responsive to domestic coal prices. That North Korean coal and steel production is similarly responsive in magnitude to world prices of these commodities is thus exceptional. In Table 6, I take a step back and test a more fundamental assumption of this paper. Here, I regress luminosity on commodity specific trade values with China again for counties with varying numbers of production facilities for that specific commodity. I run this regression to test to what extent luminosity reflects production intensity insofar as it pertains to export activity to China. Due to limited data availability, I only ran regressions for coal, steel, timber and fish. Column 1 reports that a percent increase in North Korean exports of coal to China is associated on average with a 0.033 percent increase in luminosity in coal-producing regions, statistically significant at the 1% level. This result supports the assumption that luminosity can appropriately proxy for economic production activity for coal. Similarly, Column 2 shows that a percent increase in steel exports to China is associated on average with a 0.148 percent increase in luminosity in steel-producing regions, but this result is only statistically significant at the 10% level. This provides somewhat weaker evidence for the idea that luminosity is an effective indicator of production and economic activity in steel. Columns 3 and 4 show that there is no relationship between luminosity and trade value for areas producing timber and fish.
Discussion
North Korea is heavily dependent on coal and steel for both domestic energy production and external trade. Indeed, North Korea holds around $6tn, or even $10tn according to some estimates, of untapped deposits in the form of around 200 different minerals (Vella). The coal industry specifically is of particular importance given that the North Korean regime has been using the coal trade to survive despite international sanctions (Pavone and Sun, 2014). Figure 5 shows the fraction of anthracite coal in total DPRK energy exports to China from 1992-2013, and Figure 6 shows DPRK anthracite exports to PRC as a fraction of total production. I found that production intensity in coal-producing regions is highly responsive to global coal prices, with an estimated world price elasticity of coal production of 0.957. A likely interpretation is that coal production intensity has increased because an increase in global coal prices pushes China to purchase cheaper coal from North Korea. This is confirmed by other studies and in-depth analyses of the coal trade with China over this time period that show how China has developed this trade network to exploit cheaper resource deposits in North Korea when other sources internationally become too expensive (Pavone and Sun, 2014). It could also simply be that global coal prices rose during this time period, and this coincided with the growth of the Chinese-North Korean trade relationship, spurring production. Still, because we exploit regional variation in coal production activity within North Korea and use region fixed effects, we control for the evolving relationship between China and North Korea insofar as it affects all the regions the same way. Yes, there may be differences in how the Chinese-North Korean relationship impacted different regions in the DPRK, but I assume that I have sufficient regional variation in the dataset, and the graphical representation of coal mines in North Korea (Figure 2) roughly tell us that the mines are not concentrated around one specific area and are rather spread out. Research on the coal trade in literature primarily identifies North Korea trading cheap coal to China as a relatively recent phenomenon. The most comprehensive report on North Korea and coal, a Harvard Kennedy School report by Gregory Pavone and Jin Sun, points to how the North Korean regime has in recent years used coal “in tremendously sophisticated and strategic ways to thrive despite economic sanctions” (Pavone and Sun, 2014). The authors’ main puzzle is that since 2008, there has been a substantial increase in DPRK-PRC coal trade that coincides with North Korea’s nuclear tests in 2006, 2009, and 2013, as well as the 2010 sinking of the Cheonan (Pavone and Sun, 2014). They explain that this increasing commercial activity came at a strategic time, indicating that North Korea has begun to use the coal trade to evade sanctions as they come. In this sense, the coal trade is a reactionary measure developed relatively recently as a tool of regime survival in difficult times. Indeed, their driving question is: “How is trade increasing dramatically in the midst of destabilizing actions and targeted sanctions” (Pavone and Sun, 2014)? The cheap coal trade with China is described as North Korea’s recent vehicle to evade sanctions and is now a key strategic resource for the regime. My results reveal, however, that the DPRK started increasing production intensity for coal long before 2008, which is when absolute numbers of coal exports to the PRC increased substantially. Moreover, rather than just being a reactionary tool against sanctions, it seems the DPRK has been more deliberate and strategic with its coal trade starting from the 1990s. North Korea increased coal production intensity as global coal prices increased, revealing that since 1992, the regime has strategically raised and lowered production intensity of coal based on a) its world price itself, or b) Chinese demand for cheap coal, which rose with world coal prices. Option a) shows us how strategic, sophisticated and internationally integrated the North Korean regime/economy is, and option b) demonstrates how global coal prices affect North Korean production decisions through Chinese demand. Rather than simply being a relatively recent, new strategy of dealing with geopolitical shocks and international pressure, it seems the North Korean regime has been focusing on coal production for a while, adjusting intensity based on global coal prices and/or Chinese demand for cheap coal. The current trends may be just the tip of the iceberg; a small piece of a larger trend that goes further back – by which the North Korean regime strategically adjusts its coal production. My results illustrate that even though the North Korean economy does not react to world markets the same way that a typical free market, open economy does, it is indirectly affected by global economic factors, like world commodity prices. The mechanism by which global prices affect North Korean production might be increasing Chinese demand for lower cost coal. In this sense, the North Korean economy is closer to international markets than we might think: the DPRK is indirectly intertwined in the world economy and alters its production intensity for coal based on global coal prices, which affect Chinese demand. The implications of this finding are vast: it is empirically proven that North Korea is not a “hermit kingdom.” Fluctuations in global commodities markets actually affect the North Korean economy as well. Furthermore, this study empirically confirms that early sanctions in this time have been ineffective in deterring production and economic activity in the DPRK, particularly with respect to coal.
A similar interpretation can be applied to my finding on steel, where the world price elasticity of steel production is estimated to be 2.652. North Korean steel production is remarkably sensitive to global steel prices. While less research on the North Korean steel trade exists, to my knowledge, it is likely that the underlying mechanics are similar to that of the coal trade given the similarities between the commodities in North Korean abundance and trade. Again, it is plausible that North Korea has strategically adjusted its steel production in response to global steel prices, responding to the world market or Chinese demand for steel reflected in those global prices. This finding on steel further indicates that the North Korean economy is very much responsive to the global coal and steel commodities markets. My findings based on Chinese prices are more puzzling – I found no statistically significant nonzero relationships between prices and luminosity for all eight commodities (coal, steel, fish, timber, textiles, natural gas, soybeans, and poultry). First, my results could be insignificant because of the inherent unreliability of Chinese price data. While China Data Online is a reputable source, the Chinese government is notorious for not publishing accurate price data. The fact that Chinese producer price indices for these commodities are seemingly not associated with production intensity in North Korea for those goods is not consistent with what we know/understand about the North Korean-Chinese trade relationship. Or, it may simply be that the Chinese-North Korean trade relationship is far less driven by market forces than, perhaps, the North Korean-rest of the world trade relationship is. While North Korean production and markets may be sensitive to global commodity prices, the nature of the Chinese-North Korean trading partnership is so driven by internal political factors, domestic politics in both countries, and other non-market decisions that our results are not capturing. It may be that the Chinese government has different reasons for placing restrictions on imports to boost domestic production that might be affecting production intensity in North Korea. China also has other major import partners like Australia, Malaysia, and Indonesia (in the case of coal) to support its domestic demand, so the effect of its import demand on North Korea might be smaller than we thought. In future studies, it will be worthwhile to expand the duration of our dataset. Much of North Korea’s economic opening up and increased trade activity has come after 2013, and with a new round of sanctions as well it may be that the North Korean economy has only become fully susceptible to changes in global prices in more recent years. Further analysis including years until the present would therefore add much to my work. Table 6 reports that there was a positive relationship between luminosity in coal producing regions and coal exports from North Korea to China, statistically significant at the 1% level. Although this result is simply a correlation, not proof of causation, it may suggest that luminosity can appropriately proxy for trade with China. The fact that there is a statistically significant relationship between North Korean exports in coal and steel, respectively, and luminosity in coal and steel producing regions may provide evidence for North Korea’s export relationship with China. In this case, it could be that the relationship between global coal prices and luminosity was driven by greater Chinese demand for cheap North Korean coal/steel when those commodities get more expensive globally. The observed increase in luminosity as a result of global coal prices might truly be reflective of increased trade with China. While such a relationship has frequently been discussed anecdotally and qualitatively in other research, this empirical relationship is notable given its statistical significance. Again, however, it is important to keep in mind that we have merely identified a correlation and cannot establish causality between luminosity and exports to China. Additionally, trade data between North Korea and China is unreliable. I used mirror statistics, inferring North Korea’s commodity exports to China based on Chinese data on imports. There are many problems with using this data -- it is unlikely that the Chinese government properly reports their trade statistics with North Korea, given that the DPRK is heavily sanctioned by the international community. There are most certainly a substantial amount of trade dealings that the Chinese government does not report, and there may even be transactions that North Korean state trading companies make with private Chinese businessmen, who do not inform the government either. With the existence of illegal dealings at the border, under-the-table transactions, and more, it is likely that the trade data I used in the regressions in Table 6 is flawed. Still, in addition to our results, our assumption that production intensity is a sufficient indicator of economic activity and/or increased trade with China is acceptable given that coal is North Korea’s primary export to China, its largest trading partner by far.
Conclusion
North Korea’s place in the global economy has transformed greatly since the 1990s. Much research has been done on North Korea’s nuclear weapons program and its evasion of international sanctions, but there is still considerable uncertainty around just how exposed and integrated the North Korean economy is with world markets. This paper uses nighttime lights data and IMF reported world commodity prices to find that production intensity in regions with coal mines and steel production facilities is remarkably responsive to increases in global coal and steel prices, respectively. I estimate a 0.957 world price elasticity of coal production and a 2.652 world price elasticity of steel production in North Korea. This reveals how the North Korean regime has strategically aligned its coal and steel production activity with global coal and steel price fluctuations in a way that is advantageous to its export of such commodities. Based on Chinese demand for cheap coal and steel, that also depends on global commodity prices, North Korea has been able to adapt and alter its production intensity accordingly. This reveals how North Korea is and for a while has been considerably intertwined and connected with world markets and commodity prices. This demands a recalibration of our views of how North Korea maintains its regime, nuclear activities, and evades sanctions through its integrated economy and trade with China. Our study also finds statistically significant, positive relationships between North Korean coal exports to China and our luminosity measure in coal-producing regions, potentially suggesting that luminosity reflects trade with China as well. A similar positive relationship is found for steel, but is only statistically significant at the 10% level. Still, causation has yet to be established, and there are many data limitations, especially regarding trade statistics reported by China. Based on research, we can reasonably link much of North Korea’s coal and steel producing activities to export with the PRC. Our analysis of the impact of Chinese producer prices on production intensity yielded no statistically significant results for all eight commodities (coal, steel, fish, timber, textiles, natural gas, soybeans, and poultry). Perhaps this indicates that the DPRK-Chinese relationship is far less market-driven and more characterized by geopolitical factors instead. Chinese price data may also be unreliable. Academic research has for the most part moved away from assuming North Korea to be a “hermit kingdom” and in general a greater understanding of the North Korean regime’s strategic nature has been perpetuated in recent years. Still, my findings suggest that the regime’s coal and steel industry strategies are far older and deeper than we might have already imagined -- since the 1990s, the DPRK has been responsive to global coal and steel price fluctuations and adapted its production intensity accordingly. The regime’s exposure to international markets and its ability to adjust to them shows just how sophisticated the regime has been and will likely continue to be even in the face of great geopolitical pressure.
Puzzles regarding the relationship with Chinese producer prices remain, but likely reflect the more political, non-market nature of the trade relationship with the PRC. In any case, deepening our understanding of how responsive the North Korean economy is to world/Chinese price movements helps us understand the strategies and capabilities of the DPRK and, consequently, how to most effectively impose sanctions or other diplomatic measures.