There's a strong incentive to get a high citation count for your papers. This encourages behavior like the manipulations we're seeing here, on the part of fraudsters. But it also fails to incentivize caution on the part of researchers who cite existing works. If there were a "bamboozled count" that showed how many times a researcher cited a work that was later retracted, that would incentivize people to be a little more cautious, and perhaps avoid citing work of people who are suspected to be fudging the numbers.