The problem is ultimately managerial in nature. Every manager needs to improve metrics over the last guy (and over their own historical performance). So every year, you’ve got to come up with a “new idea” to improve your figures. Installing all these gates to reduce the incidence of loss makes a key metric improve, even if it hurts revenues overall. So you get points for “fixing a problem” while your real performance is occluded behind the aggregate rise/fall of sales in your region.

And since locked cabinets are a fixed cost that (on paper) increases the value of your unit’s capital, rather than a recurring expense, you can occlude the price of the “improvement” behind the rising book value of your real estate.

In the end, these devices are a consequence of our accounting mechanics and our profit-oriented rewards. Even if implementing a policy is bad in fact, we can make it look good on paper. And since all our superiors see are the metrics on that paper, the success/failure of our business isn’t realized for years after we’ve received our bonus and our promotion and the shitty storefront becomes someone else’s problem.