Generating the Wealth of Nations 5: Real GDP per Capita in North and South Korea

Acemoglu, Johnson, and Robinson (DJR) (in one of the recommended readings, "Institutions as the Fundamental Cause  of Long-Run Growth") use North Korea and South Korea as one of their "test cases" to explore the effects of institutions (and institutional change) on economic development.  Here, graphically, are the data from the Maddison Project:

 

(Click the charts to enlarge them.)

The first chart simply shows the levels of real GDP per capita in the Maddison Project database.  The second chart is North Korean real GDP per capita relative to South Korea's.

I have a little trouble believing the data.  First, if DJR are correct about the role of institutions, it's odd that it takes two decades for any difference to become apparent in the data.  Second, it's even harder to believe that real GDP per was essentially constant in North Korea from 1973 to 1991 and then (again, at a much lower level) from 1996 to 2010.

I'm willing to believe that the institutional regime change in North Korea (and in South Korea) mattered.  But I'm hard pressed to accept these data as being accurate.