Who lost the most marks when cheating was stopped? Those of lower socio-economic status. Because those of such status generally engage in higher rates of criminal behavior [1], why is this presumed to be surprising?
It seems to me that the basis of the author's misunderstanding that leads to their surprise is the narrative of wealthy people being more likely to be corrupt on a per capita basis, which clearly runs counter to the statistics which demonstrate those of less economic means committing more crimes per capita.
It also runs counter to the fact that those of greater means have greater scores due to greater means, and thus less likely to need to cheat to have good scores.