> the researchers carried out robustness tests with various variables including the strength of IP enforcement, political factors, and economic development.
So they did take into account some factors like economic development, which makes this more interesting. However, why not analyze w.r.t. to the education levels of the area as well, since that would also affect the region's 'collective IQ' and probably indicate how many people are too poor to afford education, not to mention buying software. I searched the paper and 'education' as a search term only appeared twice, both times in references (so, its never mentioned in the actual text).
The researchers probably didn't consider the prevailing views of property rights or agreement with international treaties either. In some countries, its just more culturally acceptable to share software which makes it easier to average people to do so without feeling guilty. Other countries might have low income groups that are forced to buy licenses because their leadership got some incentives for IP treaties and are eager to show their enforcement to attract more FDI. Also, often countries with limited software exposure don't even know about good alternatives so its basically, either pirate or buy what everyone else uses.