Nothing is permanent in life.
In 1980 oil and gas was the place to be. If you had a PhD in maths you could develop a reservoir engineering Finite Element package from scratch (cool), now you can download it from the Web.
And the research jobs are not the privilege of North America and Europe any more, by no means.
And not so long ago PhDs wanted to work for themselves..
Career risk is something very few people understand, as I have said before. I saw someone on LinkedIn advise an applied mathematician "you will easily get a job in oil and gas" about 3 years ago, such is people's lack of understanding of change (judging by his profile he graduated in the late 70s). And I can see guys my age telling people the same about QF in 2036, no matter what the market reality is, which says a lot about people - most senior modellers I know complain about this all the time, especially when it is their own firm applying 1980s approaches to 2010s problems (I was in this situation in my first firm when they applied a 1980s insurance approach to their strategy in 2007).
Back to the OP - you are asking the wrong question. Interesting or intellectually stimulating is not what firms want - useful is what they want. With maths careers intellectual stimulation is often incidental and can sometimes happen when there is no marketing or business origination involved, when markets are calm and when complex maths is unavoidable. At least any jobs I did in finance that involved client work were not intellectually stimulating, while middle office roles could involve learning a new trick with correlation, modelling etc.
The thing is I still thrived in some of those jobs that involved marketing - had I followed the "follow your passion" directive I would have gone back into middle office and had a very sharp but narrow skillset. But I learnt a lot of other soft skills and a lot of non-mathematical skills that proved useful in my current career. In fact I worked with an ex-engineer that was one of those guys that had become an expert in Finite Element methods and that kept harping on about developing marketing and soft skills as he had hit a glass ceiling before he moved into banking as he was seen as "too intellectual to become a manager" as he had not looked at these things. He is now a senior partner at an infrastructure fund as his change of strategy and broadening of skills served him better in his new career.
I had a similar experience where I don't deal with clients but where I am a lot more ahead of the political game thanks to picking up skills that don't strictly speaking relate to "intellectual stimulation".
Where people get confused is with things like the difference between being an actuary or an accountant, where I know people with maths background in both that call them boring jobs. The difference is vast - although actuary is reportedly boring most maths guys I know in the field like it as they are usually the most useful on their team and, they already knew a huge chunk of the knowledge when entering the field, they become senior managers very quickly. In contrast any mathematician I know that got sucked into the whole atrocious "accountancy has numbers you'll be brilliant at it!" found that in accountancy they were learning an entire new field while calculators and Excel performed the tiny fraction of their skill set, and to round it all off after all this education (I know one accountant with a PhD in physics) people that did, guess what, an accountancy degree kicked their ass at the job, irrespective of enthusiasm.
Usefulness and seeing the way the tide turns is king NOT intellectual stimulation or even passion. I'm not sure I can comment on research as I have no experience of PhD environments.
