I think weight has a big effect there - a pallet of just glass bottles would weigh a considerable amount.I also agree.
Several years ago I spoke with a packaging expert who ran his own company supplying packaging, his reply was that most of it was dictated by supermarkets, followed by those transporting food products.
I suggested a good start would be going back in time and introducing milk bottles again as they were glass and a glass milk bottle was used an average 142 times and the customer could be charged a deposit which was returned when they returned the glass packaging.
His response was that glass milk bottles were round and 1 pint, and that supermarkets and logistic companies refused to use them as they didn't fit on a pallet and took up more room when being shipped.
My response was that they could be made square and in a variety of sizes and could be cheaply made to fit standard pallets, he agreed but supermarkets and transport companies didn't like them because if a glass container broke there was a risk of it contaminating other products in the container with broken glass and supermarkets didn't like waste and they would have to scrap a whole batch or pallet of product.
He finished with the irony, he made complete packaging on large plastic rolls and they shipped this out from the UK to several food packing plants around the world by plane.
Leaks are certainly a problem - in the can industry, if a customer finds a leaking can, the whole pallet will be rejected - 2-3000 cans. Can makers pay a lot of money for light detectors, vision systems to detect issues before they make it to the fillers
(And then you go into a plant, and find the operators have turned the system off "because it was rejecting a lot of cans . . ." - and most times, it's because somewhere upstream, the can isn't being made properly, or something is damaging the can as it is in the line, rather than a fault in the detection system. Or you go into a place with claims of no rejects - and due to lack of maintenance, most of the halogen lamps in the light tester have failed - so there isn't enough light for the detectors to "see" any leaks . . . therefore, there are no leaks . . .)