One of the things that struck me as implausible when reading Earth Abides was that decades after the collapse of civilisation, the protagonists were still eating food out of cans they scavenged from old warehouses. “That can’t be right,” I thought. Yesterday and today, though, I did some research – this was driven by my starting to play All Things Zombie again, finally catching a repeat of Life After People on TV, and wondering how long our preserved food would last before completely non-survivalist types such as myself would have to learn farming.
It turns out that Earth Abides is correct, much to my surprise. The US Army is cited on a number of websites as studying this sort of thing, and determining that canned food is still good after 46 years in some cases – the record seems to be tinned veal from 1824, still edible when opened in 1938. The limiting factor seems to be that it starts to taste funny, and the testers begin to comment that they wouldn’t eat it even in an emergency, despite the fact that medical analysis shows it still to be nutritious (although vitamins A and C are lost after a few years).
Dried goods last nearly as long; things like rice, pasta, and flour are good for 25-30 years if stored properly, and salt, sugar and honey seem to last literally forever. (Or in my cupboards, until the ants find a way in.)
With the exception of some anti-malaria drugs, which fail stability tests after a few years, drugs seem to last a long time too.
Admittedly, these sites are secondary sources; the original reports are supposedly buried somewhere in the US DoD and DFA websites – I can’t find them, but since this is idle curiousity rather than life or death, I’ve given up after a few hours’ cursory search.
So now, apart from the obvious manufacturing need to have a steady throughput of products rather than one bulk order every decade, I wonder why we have such short shelf lives on most products?
Halfway Station 
The reason for what you consider to be a short shelf life is because there is no control over how the product is handled once it leaves the store. However there is still consequences for the producer and retailer if the product causes a health issue. For example you could buy a can of spam and take it straight home and put it in a cool dark cupboard, or you could instead leave it in the boot of your car having it heat up and then cool down repeditively.
So the food technologists utilise statistical methods to come up with a magic date to put on the packet which is most likely to be in the zone where problems that they can be sued for won’t occur.
Hmm, that makes sense. Thank you.