The main problem is that there are many constraints for the material of a dashboard. Namely (just a few that pop to mind - I will update later if I forgotten any important ones):
- good UV resistance
- good thermal stability (low coefficient of expansion)
- high melting point
- good surface
- coming in different shape
- machinable (in order to take a nice shape)
- flame retardant (Pete W addition)
- able to withstand common cleaners like bleach and peroxide (Pete W addition)
- and cheap.
When there are so many constraints its difficult to find a material/solution that satisfies perfectly all needs. (A wise friend said to me constantly "quickly, done correctly, cheaply, pick any two", I guess this is applicable in this instance its just with more constraints)
One final note, since cars are sometimes (IMHO more often than not) sold more on looks rather than the durability of the dashboard (keep in mind that the owner's handling is also a factor here), the compromise tends to be the aesthetics over durability.
Another factor why this is not permanently solved is, that car companies want to "innovate" (notice the ""). Sometimes that means that they try a "new" and "better" material. Sometimes a driver for the decision, (especially in car markets that are potentially huge), they might make a deal to use that "new" and "better" (but unproven) material in their cars for a reduced cost. That benefits the car maker (because they get a material at a lower cost), and the material producers because they find a potentially large market. The problem is that durability testing of the dashboard is not something that can be completed in the timescales that a new car is designed nowadays. So sometimes, the buyer of the car gets the short straw.