I am quite clear that it is not normally tested - so the potential risks have almost certainly not been explored during flight test of your C182 or numerous similar non-aerobatic types.
I'm not an aeronautical engineer or anything, but I find myself in agreement with you here.
The certification standards are so that below Va you have to be able to put in full deflection of a single control, without the aircraft breaking up (plus a suitable safety margin). But there is no requirement whatsoever for the airframe to be strong enough to handle any rocking forward/backward or side/to/side of any control to the limit of its travel. Something a few Airbus pilots found out.
http://www.ntsb.gov/doclib/reports/2004/AAR0404.pdf
In a "falling leaf" maneuver, it may well be that the vertical G forces are very limited - probably around 1G during the maneuver, and maybe 2.5G during the recovery. So I'm quite sure the wings won't come off. But at the same time you are using full left-to-right-and-back rudder to keep the aircraft level. Granted, you do so at Vs and not at Va, so the loads will be relatively light, but still there is no requirement whatsoever that the airframe is able to sustain this. So essentially you are a test pilot.
And it's not just the tail. Also consider what the fuel in your wing tanks of your PA28, and the wings themselves are experiencing when you start yawing left-right-left-right. Wings are immensely strong in the vertical, but not designed for any significant horizontal load.
In my opinion the falling leaf is a suitable maneuver for advanced pilot training, to help somebody get rid of "lazy feet" or to demonstrate edge-of-the-envelope flying. But it should be treated just like spin training: Suitable (aerobatic) aircraft, suitable instructor.