This could be bit late...but this is a famous question :-D
I think its very personal choice, but I would still like (cant control myself in getting involved in this) to make some general comment. You said ckt design, I think that means analog ckt design. Coz digital ckt design happens at a much higher level of abstraction. Anyway...
Analog Ckt design is still done best in industry, if you consider 'best' to be what eventually works without any hidden weaknesses! But as already pointed out something grand new rarely happens! Its almost in all cases some tweaking to previous designs to obtain something new. But you get to work on which 'works' in Silicon. No fuss! Since a company is a business unit, its driven by business requirement. If a company can make millions of dollar by selling same product years after year they will do that (similar to EDA industry)! But thankfully that doesnt happen, there is always competition to come up with a better chip. Now better could mean a lot of things -- cheaper per functionality, better spec, more compatible with the end-customer. The last point really determines a lot many projects to be killed, simply because the customer company doesnt want to change their board :-O
In these cases even a better spec design may not win the business!! But since some smart customers like to change, chip design companies also need to come up with better soln. So broadly in any such company there are 2 design grps -- i) product line & ii) futuristic design grp. Most of us land in the former as this is the grp that keep the money flowing for the company, where the 2nd grp mostly investigates on future trends (many say they waste the money the former earns
). For the 1st one you dont need a PhD. You need to understand circuits. You learn a lot initially to understand the design flow (if you are in digital grp, thats the heart of the whole thing), you work on real ckts, and different ckts teach you abt working principle of different ckts. This is nice! But to enter the latter grp you need to have specific knowledge on something. You will be hired mostly for some strategic reason --- in other words you better have a PhD or have good work exp on that specific topic.
One more thing, if you enter the first grp after doing a PhD, you may see a lower-degree person is mentoring you, as doing a PhD on some specific ckt doesnt make u an immediate pro in all the kinds of ckt that this grp is involved in designing.
One needs to understand simulation tools as well, I have seen good ckt designers struggle with simulation tools and mangers dont like that :-D Your qualification and enormous knowledge on something (which u gathered during ur phd) may not rescue from this day-to-day stuff.
I generally wud suggest to spend some time in product line in the beginning to understand your interest, because more often than not, we dont know what we want. I am saying this more, because the question you asked itself reflects this uncertainty. But if you are more interested in obtaining the degree then go ahead, as things get a bit tougher in later stages...and you may lose the motivation!
cheers!
Rajdeep