I can see reasons for them not doing this. One would be that they probably don't know. When the pools are looking for a block, they are, I believe, looking at many differnt combinations of transactions at the same time across their supporting miners. Another, they don't know what they will use next until they see what was used by the successfully found blocks.
None of that is correct.
These things are done by computers, there isn't some pool operater sitting there deciding which one to use.
Wouldn't that imply it is easier and fully automated for a pool to know and report what tx are in the current working block?