Honaker on Remote Sensing and Sanctions Enforcement



Satellite imagery has played a role in tracking the prison camp system and nuclear programs, but somewhat less has been done to monitor the economy. The Journal of International Policy Solutions—which comes out of the Graduate School of Global Policy and Strategy where I teach—has an interesting piece in the current issue (.pdf) on remote sensing. Framed in terms of sanctions enforcement, the implications of the piece are much wider, effectively providing ways of measuring the level of economic activity. One key finding: Rason and Kaesong do not appear to be generating significant backward and forward linkages.

The piece considers three distinct indicators, and provides a “how to” on using them:

  • Images from the Landsat Top of Atmosphere (TOA) Percentile Composite, the annual Landsat 8 Normalized Difference Vegetation Index (NDVI) and a random forest classifier provided by Google Earth are used to detect changes in built-up areas between 2000 to 2014.
  • Images from the DMSP-OLS Nighttime Lights (US National Oceanic and Atmospheric Association) are used to measure changes in luminosity (a measure Travis Pope and I played with; see here, here, and here)
  • Forest cover images from 2000 to 2015 were obtained from the Hansen Global Forest Change v1.2 dataset to evaluate local rates of deforestation.

These three somewhat measures of economic activity—built-up area, light and deforestation—are used to measure the broader economic effects of the Rason and Kaesong economic zones, the only two that appear to be functioning. At Kaesong, increases in built-up area are heavily concentrated around the park itself, with only some expansion at the edges of the city of Kaesong. The latter probably reflects the growth in population associated with the large number of workers and their families associated with the zone when it was operating. Kaesong also saw deforestation in the mountains north of the zone, which could have reflected demand for lumber at the site or an alternative hypothesis: private plots, with demand for their output possibly driven by the incomes generated by the zone. Not surprisingly, luminosity around the zone itself shows a sharp increase as KIC came online, but when looked at together there is not strong evidence of robust linkages to the local economy.

In Rason the areas of buildup are more modest, and the exercise also shows some deforestation outside the city that appeared to correlate with it being granted the status of a directly-governed city. Increases in luminosity are visible but more modest.  

Honaker’s conclusion: “allowing increased foreign development has yet to deliver tangible benefits for the immediate areas surrounding the zones. It may be the case that whatever benefits that have materialized are being funneled to the capital, Pyongyang, or that foreign investment has simply failed to bring any benefit whatsoever.” This comports with Marc Noland’s long-standing observation that not only is foreign investment scarce—due both to sanctions and a weak property rights regime—but where it has come in, backward and forward linkages are modest as well.

A lot more work would be required to calibrate this to activity elsewhere in the country, but the advantage of the paper is that it provides a very clear discussion of methods and is thus open to replication. If you are looking for a way to conduct cutting-edge research on North Korea, the applications are endless, from the growth of Pyongyang and other cities, to the effects of infrastructure, to regional differences in growth rates. Kudos to Honaker.

More From

Related Topics