Body
It has long been alleged that North Korea has the capacity to impose asymmetric damage on the financial markets of its neighbors through threats and provocations. When I examined this issue around ten years ago, I concluded that the impact of provocations on South Korean and Japanese markets was minor and transitory. In the wake of the fourth nuclear test, both Charlotte Fitzek and Victor Cha and and I re-examined this issue and again reached skeptical conclusions. Indeed, I have argued that due to the thinner markets in North Korea, the strategic use of provocations usually backfires, destabilizing North Korean markets more than South Korean markets.
Due to the thinner markets in North Korea, the strategic use of provocations usually backfires, destabilizing North Korean markets more than South Korean markets.
In the forthcoming issue of the Journal of Asian Economics, Sel Dibooglu and Emrah Cevik once again step into the breach and obtain results more consistent with North Korean capacity to destabilize financial markets narrative. They do time series analysis on monthly data from the NK News North Korean Threat Index which measures the frequency that confrontational terms appear in North Korean state media. The sample period is January 2000 to July 2014. They examine the possible impact of North Korean threats on interest rates, exchange rates, and stock prices in South Korea and Japan.
The results are suggestive, but a little hard to reconcile. They find that in a statistical sense North Korean threats cause stock market movements in both South Korea and Japan, but that the impact is nearly twice as large in Japan, which seems odd: Japanese markets are deeper, and one would expect that South Korea would be under greater direct threat from North Korea. Likewise, North Korean threats move the South Korean won, though the impact disappears after two months. North Korean threats are found to move the Japanese yen, but the effect does not manifest until the second month, and rather than dissipating, the impact in three months out is actually larger than the effects at the two-month mark. North Korean threats are not found to have any impact on interest rates in either market.
In short Dibooglu and Cevik have applied sophisticated time-series techniques to new data, but the results, while suggestive, are hard to accept at face value. One issue might be periodicity: I found that the effects appeared to disappear within a week, so I wonder what the authors are picking up using monthly data. I am not sure what the periodicity of the North Korea Threat Index is, but the financial market data is easily available on a daily basis. If not daily data, a re-examination using weekly data might be worthwhile.
