+1 for what he wrote. Sometimes failures don't catch fie, but will get very, very hot. About forty years ago I saw a smoking tantalum power supply capacitor which had shorted while I was trying to figure out why the circuit was unstable. It had not caught fire, but was well on its way. That day the capacitor gods smiled upon me. I've also seen cans and PIOs explode, streaming foil like confetti throughout the cabinet. Your luck may considerably vary, as
@merlynski sets forth.
BUT, having said that, sometimes the voltage ratings on the part are far greater than what is required for the reasons I above elucidated. An amplifier with, say, 40 VDC rails can't go to 250 VDC. The value was chosen based upon price, stocking, and other considerations which may have been external to the circuit parameters. Tinkering with values like this, however, requires some understanding of the circuit and this is not a sport for amateurs.
Most, but not all, of the difference in rating is in the dielectric, but sometimes the electrolyte and construction change depending upon the intended use.
Capacitors intended for power supplies, for example, often deliberately have higher ESR to damp out ripple current and prevent standing waves. This is one of the issues in having high-frequency motherboards or phones with many decoupling capacitors. The parallel layout reduces ESR to a vanishingly small number, and standing waves can arise in the groundplane. Plus the issues of having tank circuits created by mismatches in L and C.
My point is not that such effects occur in audio frequency devices, but they can if high-speed digital logic is present and using the same ground configuration, but that capacitors are different depending upon the intended use.