I like having a PDA in addition to a mass spec for various reasons:
(1) A PDA spectrum can still be quite diagnostic for a class of chemical. If you are asked, for example, to find all the flavonoids present at significant level in a plant sample, then PDA is very valuable because every flavonoid peak will have a very similar spectrum, and it's much quicker to look at each significant UV peak to check its spectrum looks sensible, than to examine every mass peak to see whether it has a mass and fragmentation pattern appropriate for a flavonoid.
(2) UV absorbance is very stable, while ionisation efficiency changes over the course of running samples. If you cannot obtain isotopically-labelled internal standards (a very common problem!), then external standard quantification based on PDA may be more reliable than external standard quantification based on MS. You can use the MS to confirm identity, and PDA to quantify.
(3) There are many situations where a whole class of compound will have the same response per unit concentration when measured by UV absorbance, but not by MS. This is approximately true of things like anthocyanins, where the visible absorbance changes only slightly when extra groups (glycosylation etc.) are added to the anthocyanin. In these cases, if you can only obtain commercial standards for one or two members of the class of compound, you can still estimate the absolute concentrations of other members by assuming the same response in PDA. You can't do this in mass spec, where the response is likely to be very different.
(4) UV absorbance is a really convenient way to check the performance of your pump, using acetone-spiked solvent versus solvent!
(A little note on dual detector measurements, PDA + MS; I tend to check that the data approximately correlate. If they don't, I assume that the PDA has detected the wrong thing, because PDA isn't as specific as MS. If they do, I assume that PDA has measured the correct thing, and probably more reliably than MS).
Despite what I've written, I do have one instrument without a PDA: they're much more limited when used on high-sensitivity triple quad systems, because the amount of analyte necessary for detection in PDA is so many orders of magnitude above the optimum concentration for MS. It's hard to find a range of concentrations where both detectors are doing a good job.