That would be a very difficult answer to quantify. The answer may be indicated with other related statistics. There are documented UK studies that indicate in the public sector, CCTV reduces overall crime by about 16%, car theft in particular by over 25%, and violent crime, hardly any at all. The UK of course has a far more advanced implementation policy for CCTV than most civilized countries so it would logically follow that the effects have been measured in numerous studies.
In the residential market, it is a proven fact that burglar alarms do reduce overall crime, especially in neighborhoods that have a dense installation base. Using alarm registrations statistics and reports of crimes to LE in those areas, they do deter crime without displacing the criminal activity to other nearby residences or neighborhoods. We also have actuarial evidence offered by insurance companies that offer between 15-20 reduction of insurance premiums for installed, monitored burglar systems. They don't offer that because they are nice people or as a public service. They do it because their numbers tell them so. In short, according to some meaningful, measurable methods, security systems to prevent and/or act to mitigate residential crime.
Both of these share at least two common traits. 1. Signage. A system that is combined with public signage is far more effective in prevention for obvious reasons. 2. Response. In areas where crime gets noticed and reported, by either a public CCTV or private alarm system, Law Enforcement response and follow up is essential in creating at least the perception of an effort to investigate, solve and prosecute criminals.
There is also a difference between the two that is important. Public perception is that street CCTV systems do at least infer that someone is watching the cameras - that the area is under direct surveillance. Private CCTV systems do not share that trait, although there have been largely unsuccessful efforts to increase that ability.
In my opinion only, video systems that have some ability to "mimic" alarm systems (summon LE), and can offer footage of the crime would have significant value. The industry is getting there. Many Central Stations already offer video alarm verification. How much benefit is almost impossible to measure given the young age of the market and the lack of statistics. I have no doubt reporting will get better and statistics become available as the market and the equipment matures.