If you're testing your drug against a placebo, and the drug has a negative side effect, then you should try to make sure the placebo has the same/similar side effect. The whole point of double blind testing against a placebo is to test against something that appears similar.

Remember doctors aren't supposed to know what patients in a trial are on drugs vs placebo (that's the 'double' in double blind, the patient and the doctor is blind). If the real drug has a horrible side effect and the placebo doesn't, then the doctor will be able to figure out what patient is getting what, and you no longer have a double blind test.