There's no harm in using a capacitor whose voltage rating is far above
the circuit voltage.

There is one peculiarity of electrolytics that should be kept in mind,
however, and maybe it's what you're thinking of. The dielectric in an
electrolytic capacitor is formed and maintained by leakage current. The
dielectric layer grows until the leakage current drops to a very low
value, at which point it quits getting thicker, for practical purposes.
If an electrolytic capacitor is run at a low DC voltage for a very long
time, the dielectric layer slowly dissolves or degrades. (The length of
time depends on temperature and other factors, and in my experience
typically takes years for significant degradation.) If it gets thin
enough for leakage to occur, it grows back until the leakage drops to a
very low value -- but no thicker. So eventually, a 25 volt electrolytic,
say, run at 5 volts DC for a very long time, becomes a 5 volt
electrolytic because its dielectric is just thick enough to sustain that
voltage with minimal leakage. If it ever again needs to sustain 25
volts, it has to be refomred by gradually raising the voltage across it
to allow the dielectric to reform without undue heating. It's this same
phenomenon that causes old, unused electrolytics to get leaky and
require reforming.

It's considered good design practice to never expose an electrolytic
capacitor to more than 80% of its rated voltage, and some guidelines
recommend less.

Roy Lewallen, W7EL

Paul Burridge wrote:

Indeed. Is it correct, however, that one should not go *too much*
higher than the maximum anticipated voltage? I read somewhere that
using (for example) 25V electrolytics in a point in a circuit where
there was only a few volts of PD could result in some sort of problem.
:-/