I have blogged earlier about how the incentives are stacked in favor of being corrupt. As the graphic below shows, it just pays, and pays big-time, to be corrupt.
Clearly, the probability of a corrupt person being identified, charged, found guilty, penalized, and his ill-gotten wealth appropriated is about as much as him being fatally hit by fragment from a falling meteor! Even assuming the officer is found guilty and penalized, thereby inflicting some shame and minor pecuniary cost (in relation to the rents), in all likelihood he retains his rents. The massive stakes (in terms of the magnitude of rents) and declining social stigma associated with being perceived as corrupt further distorts incentives.
In the circumstances, one possible strategy to deter rent-seeking, especially at higher levels, is to target the wealth amassed by officials. Once the preliminary investigations reveal some form of rent-seeking, the investigations should simultaneously target the assets of the official and his family. This strategy has atleast two advantages. One, in so far as it is outcomes focussed, targeting the real gains from rent-seeking, it appears more incentive compatible as a deterrent. Two, it may be legally less difficult to establish disproportionate wealth (even with benamis etc) than establish procedural irregularities.