So I am currently one year through a maths PhD, but that's about to change - I'm going to convert it to a Masters, and then do my maths PhD overseas (probably at Princeton). Here's my thoughts:
In Australia, the standard postgraduate scholarship is an Australian Postgraduate Award, which is $22,800 per year for three years (the standard length of a PhD), plus six months more if needed. I also got an $8,000 per year top up on top of that (my department just hands them out for no reason... weird), and I tutor some courses, which also gets me a bit more money. Overall, I probably earned a bit more than $35,000 last year, with only the tutoring income being taxed. This was certainly more than enough to live on (I definitely could've survived on just the $22,800), and I paid $200 a week rent and bought my own food and all that jazz.
Academia is probably not a career you want to do if you really value earning as much money as possible. The income at the top is more than enough to live on (full professors earn $130,000 or so a year, department heads etc even more), though it takes a while to get there (and obviously you don't earn very much at all while doing a PhD or a postdoc). It's certainly more than you'd get as a teacher, say, but a lot of industry jobs would be higher. On the other hand, you have a bit more freedom in what you can work on as an academic compared to working in the industry.
Job security depends on your area of study and where you work. In the US, academics want tenure track jobs so they can get tenure and hence have strong job security - if you're in a tenure track job (i.e. assistant professor), then you want to do as much research as possible and publish lots. In Australia, it's a bit different: we don't exactly have tenure like they do in the US, so the job security at the top probably isn't quite as good, but you earn more at the bottom. Once again though, you have to publish lots to move up.
This last point is a big reason why most people don't research things that aren't in vogue (unless they have tenure, in which case they're less afraid to work in less popular areas) - you want to have well-received publications so you can get promoted, and the best way to do that is work in areas that have lots of people who'll read your stuff. That being said, this seems to be less of a problem in mathematics, because it's much less likely that publications will be controversial and shunned - in (pure) maths, proofs are either right or wrong (and to be published, they're nearly always right), so the worst that can happen is that someone can find your paper boring.