One of the benefits of having a pool of part-time testers was that the people who helped rate really had very little financial stake in the ratings, and as such felt free to give whatever rating they felt was appropriate.
With full-time raters, they have a financial stake in their ratings. This could lead to harsher initial ratings so that companies would have to pay for the re-rate, or unscrupulous publishers/developers might be able to try graft outside the office.
Another major issue here is that with their former method of using part-timers pulled from a volunteer pool who submitted demographic information up front, they could get a decent cross-section of the populance (parents, students, etc.) With the full-time raters, U.S. employment law forbids the ESRB from asking a lot of this demographic information during the employment process.