I'd imagine it's a mixture of Cost, Time and Laziness.
For absolutely everything to be put into a game so you have the highest amount of detail, you would have to consider that all those small details have to be put together in a games development plan. Such plans are used so that multiple people can work towards the completion of the same program, and they are also helpful in the sense that plans can be made generic and then applied to any programming language.
Also some of those "extras" are fiddley and therefore some designers would stay clear.
Another aspect is the time it would take to place those extras into the plan, and debug them during programming. Most of the mainstream games have a development Gant chart rigged that will have a specified development time. All those graphical extra's could potentially push the development time period up, the longer the time period, the more out of touch with the current software and hardware development the game becomes.
Lastly is cost, all that extra coding is costly if it's not applied to the cost at the very beginning of the development stages.
However it is possible that such pieces can be worked on with the generation of programs version numbers, thats why in some games a patch might add graphics or objects into a program, because those extras were placed into the Retro-Development Queue which is near the end of a Systems Life Cycle.