According to the local paper today, there are apparently 12,000 library books missing over the last three years. It's not altogether clear exactly what "missing" means, as the article starts off talking about knowing who's got them but not how to get them back, which isn't really "missing" as in 'we don't know where they are'.
When I was working in libraries, people who borrowed stuff from us generally brought it back, and most of them did so on time, although the fines levied on those who didn't were a nice little earner for the Council. The cry of "What!! I could've bought it for that amount" inevitably resulted whenever the charge rose to more than a couple of quid, and the likelihood of getting something back decreased more or less in proportion to the rate at which the fines went up. For a while, I think at the auditors' suggestion, they experimented with taking people to Court, but needless to say the people they chose to sue hadn't got any money, so it was simply an experiment in pouring further money down the drain without getting a result. Peversely enough, the one thing that did seem to work was "sending the boys round": contrary to what you might've expected, most people were quite relieved to get the books off their hands without further ado, which was the main object of the exercise.
On the other hand, the article's a bit vague over whether theft (as in taking a book without actually bothering to "borrow" it first) is included in the figures. It always used to be difficult to quantify this, as it was often simply the fact you couldn't find something when a customer asked for it which prompted the "missing?" categorization. Even then there wasn't any conclusive proof it hadn't simply been put back in the wrong place. I won't go into all the ins and outs of security systems other than to say that it's not really feasible to tag every £2.99 paperback in and out each time - any more than it is for a supermarket to tag every tub of margarine that they sell.
But given that book borrowing from libraries has been declining steadily for a number of years now, it seems logical to suppose that book losses have too, and the comparative figures they've quoted in the article do seem to bear out that inference.