Thanks for the kind words from those who commented.
I have not looked at a lot of amplifier schematics, hardly any recently (a lot of manufacturers included them but few do now), so do not know which amplifiers might have regulated output rails. You can add Sanders to the list, however. In general regulating the output rails is a mixed blessing. Output is more stabilized, but at a cost in components, heat, and max output headroom. I have had amps with and without and am not sure I would say it really mattered in the end, assuming a decent power supply design (sometimes a flawed assumption). Remember IME the main difference is reduced max output, something you are far more likely to measure on the bench than hear in the real world.
Some AVRs, pro amps, and other components use switching power supplies and those are essentially regulated by design.
Anytime you over-stress a component you are shortening its life, whether voltage too high, too much current, too hot (or even too cold), etc. Like micrstrip said in his excllent summary, it depends upon the design margin. That said, it would be a pretty poor design that had less than 10 % margin, so I would expect it to work OK. Transformer insulation breakdown ratings are usually much higher; active devices may have lower margin if you go the active (e.g. UPS) route but again I would expect at least 20 %. However, I am not a UPS designer by any stretch of the imagination.
Most problems I have seen with transformers (power or output) are due to over-current (e.g. during loud passages). When the core saturates it wreaks havoc on the output. In my world we try to keep the peak current through an inductor (transformer or choke/filter bead) below 1/4 to 1/2 its max current rating.
