Sources: Ben Katzeff Silberstein on North Korean Markets

October 7, 2015 7:00 AM

We love nothing more than innovative new research on North Korea, and Ben Silberstein does not disappoint with a new paper from the US-Korea Institute at SAIS on Growth and Geography of Markets in North Korea: New Evidence from Satellite Imagery. The study uses Google Earth’s “yardstick” function to measure the growth of markets over time in 12 cities. Although the data goes back as far as 2003 for some locations, there is a more-or-less consistent panel of data only for 2008-2014. The markets measured are almost certainly an underestimation of total market activity because they focus on government-maintained market structures, as identified in the painstaking work of Curtis Melvin at North Korea Economy Watch, and his earlier North Korea Uncovered project. However, this choice was probably the right one: the episodic nature of the satellite imagery would make it difficult to capture more informal jangmadang or to draw any meaningful inferences about them.

Silberstein finds that the biggest changes in market size were in Sinuiju—no surprise—but also Haeju and Hyesan. Given the long time frame, the annualized growth of markets over the period elsewhere in the country is close to zero, however, and that is if we exclude the collapse in Pyongsong’s square footage following the closure of its large wholesale market in 2010 (which the satellite imagery does in fact capture). Silberstein puts a positive spin on this finding. In places where data permits a consideration of the period Marc Noland and I identify a reform in reverse in the second half of the 1990s, market space did not in fact contract. But an alternative spin would be that despite the current focus on de facto marketization, it does not generally appear to have occurred through the expansion of formal market structures.

If we move from a metric of sheer growth to a normalization of the data by population (square feet of market space/city population), somewhat different patterns jump out. First, the correlation between population size and aggregate market size is weak. To my surprise all cities in the south of the country have higher square feet per capita of market space than those in the north. 
I was also surprised that the cities with the largest aggregate market size per capita are in the western part of the country; an alternative hypothesis would be that distance from Pyongyang was a plus for market development. Silberstein tests the proposition and finds that the correlation between distance from Pyongyang and aggregate market size (which we can imagine might be positive [spillover] or negative [tighter central government control] is weak).

In line with the Sinuiju finding is another that is obvious on reflection: that cities with larger than average per capita market are ports (in descending order, Nampo, Haeju, and Chongjin). This suggests that proximity to the Chinese border is capturing something else: underlying transportation infrastructure, which could be deepened at the ocean ports as well as the land ports along the border.

It is important to emphasize that the market space variable is only a very imperfect proxy for market activity, and in fact may not be related to it at all. But Silberstein’s measure is a proxy for the commitment to structured market spaces. The biggest question is the extent to which local initiative is playing a role in the observed divergence across cities. If so, it would comport with a point we have made before: that marketization from below will depend not only on central government acquiescence but on local officials experimenting with a different model.

Chapeau to Silberstein and Melvin for this work.


Rudiger Frank

I agree, good work.
From my last trip to NK three weeks ago, I can add that indeed some of the markets were not active while I passed by, which might mean there were closed for good, or just temporarily, for example due to ongoing flood recovery. The market in Rason, which I was able to visit, has been moved into a regular and I dare say modern building, actually a couple of them, and is thus not recognizable as a market by satellites anymore.
The disadvantage of satellite analysis is that is will miss such changes; the advantage is that, unlike anecdotal evidence coming from field visits, it can be done systematically.


Silberstein also has a recent article in 38 North, here:

More From

Stephan Haggard Senior Research Staff