The problem with studying a drug using control groups and placebos is that for some drugs (such as marijuana, LSD, etc.) people will know that they must have gotten a placebo due to the lack of effect.
Researchers will sometimes use a substitute to create some effect that a naive participant may believe to be the real thing, but it can be difficult to be convincing. This can damage the integrity of the control group because they would know they hadn't gotten the real thing (and thus possibly negating the placebo effect).
People who know they got the real thing may see benefits that are actually the placebo effect, and if your control group isn't able to control for this effect, it lowers the quality of your study.
There are also moral issues with tough studies such as last ditch experimental studies on stage IV cancers. Even in a double blind study, the research will know they are giving placebos to some of the participants, and thus have no chance of helping them.
