First, keep to your time. The audience has been pretty patient and want to ask questions. You are in the way of that. However, you are also far more likely to have read the paper and be equipped to ask good questions, so do not surrender your time either.
Second, if there are ways that the papers talk together, suggest what they are and what each paper could gain from the others. If the papers are completely unrelated (happens all the time, especially at the ISA), do not force it. If the papers do not engage each other, then do not make them do so.
Third, consider what would make the papers better. Not what would the paper look like if you wrote it (a favorite stance by discussants and commenters at job talks). But what are a few things the author could do to make the argument more compelling, the research design better, the findings clearer, whatever. Just two or three suggestions/criticisms per paper. If there is a common weakness in the papers, then you can use that in your general discussion rather than repeating 3-6 times.
Fourth, people tend to rush through conclusions, so you might want to suggest what the policy implications might be or where the next step in the research should go.
- Do not expect a paper to be more than what the author wants it to be.
- Give each paper equal time. Just because one paper is better than another or is written by a bigger name does not mean that you should play favorites.
- Again, be brief. Do not try to provide the same kind of review that one does when reviewing manuscripts for journals--not enough time. There is also an audience out there.