Bank robbery stats; bank branch business

I suppose I could have looked these up myself, or I could link to the primary source, but why do the research when "Ask Doug" is willing to do it for me? The question for Doug? How many bank robberies actually take place? The answer:
I found the Federal Bureau of Investigation’s Bank Crime Statistics for the second quarter of this year. During those three months (April through June), there were 1,278 bank robberies in the United States. That’s an average of 14 per day.
That's kind of staggering, especially considering the stats that follow (i.e., robbing a bank isn't a good way to make money). That's roughly 5,000 robberies a year. And, according to a comment left on this excellent examination of whether bank branches are going the way of the dinosaur (they're not - yet), there are some 100,000 bank and credit union branches in the United States. So about five percent of all branches are hit each year (I'm sure some are hit twice and I'm sure location matters in risk assessment, but that's a good honest number to use if you're pitching a video system to a bank branch, who all already have video systems anyway, but still, you want to sell them a new one). It's no wonder security is a concern. But how good is the security at these branches? Is the ROI real on the security system?
Robbers got away with $9.5 million, and law enforcement so far recovered $1.46 million of that, from 28 percent of the robberies that ended up with money actually being stolen.
So, they got away with an average of $7,433, and you've got a 28 percent chance of recovering the money (obviously, this again depends on location, etc.). Seems like the bank robbers aren't doing a whole lot of risk assessment (they've got a 40 percent chance of being caught, apparently). That's a pretty paltry sum in exchange for certain jail time (I'm guessing crystal meth impairs judgment, yes?) 40 percent of the time. So bank branches stand to lose an average of $7,433 per robbery (which they probably write off right away, so I'm not going to count the money they get back), and that's bound to happen about once every 20 years, going by the math alone.
In almost all the robberies, alarm systems were activated and surveillance cameras were on.
So, the question is, if alarm systems and surveillance cameras result in our current situation, a loss of $7,433 every 20 years, what would happen if bank branches didn't invest in those things? Would that increase to $7,000 every year? Every five years? How many of the 40 percent get caught because of surveillance footage and how many get caught because they're stupid and hide the money in their couch and their buddy rats them out? Just some numbers to get me thinking, I guess.


Banks usually factor in other costs including fraud losses, loss of reputation/customers from publicity of robberies, employee turnover/counseling costs, etc.