Having been involved in exactly these type of decisions for a decent sized manufacturer I have a few comments.
Yes there are economies in using the same parts in multiple places rather than bringing in multiple parts. In practice this generally means parts being used that are better (more uF or higher voltage) than needed if unique parts were specified for every node.
Products are not designed to fall apart one day past warranty, the science isn't that precise. Generally there are only a handful of weak links in a given design that will dominate effective service life, and even then reliability engineering is a bit of a seat of the pants exercise, where there is a dynamic tension between engineering wanting to keep using old proven parts, and purchasing trying to save money by bringing in newer cheaper parts (after passing engineering imposed hurdles to prove they don't suck). In an ideal world these newer cheaper parts are also more reliable, and that has generally been my experience but there are always gotch'as and vendors who lose the recipe after making good parts for years.
Regarding sizing power transformers for audio amps, that could make a chapter in the book all by itself. In general these transformers are more taxed by UL testing parameter than in actual audio use, as it should be. Another quirk about audio amps that is becoming a distant memory as simple class AB fades into historical obscurity, is that FTC regulations required a long 1/3 power preconditioning cycle before power is measured. For those reading along at home, 1/3 power is the worst case for heat sink dissipation in simple class AB so a serious heat load. Another little known fact is transformers put out more power when cold than hot, because wire resistance actually increases with hear rise,,, long story short, the FTC rules forced amps to be over designed for actual use. Modern multi rail amps are dialed in to run relatively cool at 1/3 power mooting the FTC preconditioning test.
Another tidbit about transformers in audio application is that dynamic music does not make a well defined or consistent load. Also clip limiting and amp protection circuitry can reduce stress on the transformer vs an unprotected amp channel or theoretical calculations. I recall being told by a new engineer that the transformer being used in a new version of a popular 300W topbox mixer/amp was inadequate. Since we were shipping thousands of channels of month using this exact transformer and power amp, with near zero service complaints. There was nothing wrong with the young engineers math, but sometimes the real world is easier than the test bench not harder.
I recall when we decided we wanted to increase our warranty guarantee from 3 years to 5 years at my old day job. We did a thorough review of what our actual warranty experience really was model by model. We found we were already fine at 5 years, with the exception of one amp in my product mix that had a previously undiscovered design flaw. It was a low volume seller, so service didn't notice that the claims history was out of whack and flag it for engineering review, they generally respond to how big the stack of repairs is.
Curiously when I told the engineering director for the power amp division what parts were failing, he looked at the schematic and immediately found the mistake... a simple value change later and that amp was rock solid too.
I have been out of this high volume manufacturing mix for several years, and I would be apprehensive these days about offshore contract manufacturing where the contract manufacturer is allowed to source the parts and profit from saving money. Most of the chinese assembly contracts I dealt with, we specified every part.. But then was then,, I don't know how it goes now.
JR