In view of the previous comments about the maximum voltage rating of modern resistors, I tried an experiment.
My previous figure of 250V was derived from the data for the Philips CR25 series which nobody, I'm sure, would describe as 'modern', so I headed for Farnell's website (other suppliers are available) because I know they have data sheets readily available.
I took 5% carbon film resistors as a typical choice for most day to day use and looked at what they had.
Their main choices centre on the Multicomp range, to which the following applies.
Starting with the smallest and working up, we have the following:
Type/Power Rating/Max Voltage/Available Resistance Range
MCRE 0.125W Series 200V max, 1R0 - 1M0
MCF 0.25W Series 250V max, 1R0 - 10M
MCCFR 0.5W Series 350V max, 0R5 - 10M
MCF 1W Series 500V max, 1R0 - 10M
Pretty much as I expected and not the serious down rating warned of earlier in this thread. So, a modern 0.5W carbon film resistor can be expected to do everything its composite carbon predecessor was capable of.
I did find one inconsistency, though. There is another 1W range from the same manufacturer and a 2W range in the same series. If you look at Farnell's 'More Details' link, both series are stated to be 350V max yet, if you read the manufacturer's data sheet, it clearly says 500V!
MCCFR 1W Series 350V max, (Farnell) 500V max, (Multicomp), 4R7 - 1M0
MCCFR 2W Series 350V max, (Farnell) 500V max, (Multicomp), 4R7 - 1M0
Personally, I'd believe the manufacturer.
The data sheets also give the formula for RCWV (Rated Continuous Working Voltage) which saved me having to work it out!
RCWV = Square Root (Rated Power x Resistance Value)
From this a compiled the following chart showing maximum voltage for each power range against resistance (to keep it within a reasonable length I've pruned the list of values down to 5 per decade)
Note how the maximum resistance limit for maximum voltage for all
types is the same (or very close) to the value I previously calculated!
On the basis of this I think it is fair to say that the humble, cheap, 0.5W carbon film resistor is perfectly adequate for all but the most demanding of applications you may come across - assuming, of course, that the manufacturer hasn't specified a component with a higher rating!
If the value is above 220k though, you might want to double check whether an uprated component might be justified - which is why I've highlighted that section of the chart - but that should only be necessary occasionally.