I'm not questioning the data, but this made me think of a few questions. Humor me...
How many crimes could we reasonably expect a CCTV camera to help solve over its lifetime?
Is solving crimes the primary purpose of many / most / all CCTV cameras?
Are all CCTV cameras located where crimes are known to occur?
Are there some places that just never experience a crime? If we don't put CCTV cameras there, will the crime migrate?
Are criminals clever enough to know when they are on CCTV?
Would it make things easier to RFID tag every member of the population so that we didn't have tedious facial matching to do? [just kidding...]
These should lead us to asking the question that really matters..."what is the optimum number of CCTV cameras to achieve the right balance between crime reduction/prevention and privacy?" I don't know that there is an answer - very much in the eye of the beholder, but extrapolation in both directions leads to craziness. The UK is at the leading edge of CCTV deployment - if there's going to be a backlash anywhere - the UK will probably get it first.