It does work.  It's been done somewhere in California on a small scale, IIRC.  It's not done more because the initial investment is high.
I don't understand why the initial investment should be so high.  Conversion of heat to electricity is a mature technology, and aimable mirrors don't seem very expensive.
The higher the steam temperature the better your thermal efficiency for your generating cycle.  It takes a lot of mirrors to get high temperatures.
Mirrors are expensive actually.  They are basically a pane of  glass with an aluminium coating and an anti-reflective coating on top of that.  The electrical actuators used to track the sun are also expensive.  Futhermore, mirrors don't work when it's cloudy.  That limits concentrating systems to locations like the Sonata and Sahara deserts.  Concentrators can be highly effective in those high insolation environments.

Photovoltaics have no moving parts so they don't require much maintenance.  You just have to hose off the bird droppings every once and a while.  

I interviewed for a job as startup engineer on the "Power Tower" in the Mohave Desert run by Southern California Edison back in the late '70s.

What a loser!  I declined as the performance was so marginal. It's still there - you can see it from the freeway or if you take Amtrak's Southwest Chief:
http://www.amtrak.com/servlet/ContentServer?pagename=Amtrak/am2Route/Horizontal_Route_Page&c=am2 Route&cid=1081442673827&ssid=132