The usual reason for this is to simplify the address generation in the microcontroller. The FLASH probably only output full 16 bit words. Therefore, if you "throw away" A0 on the microcontroller, you do not have to worry about odd addresses and byte boundaries.
The first location of in the flash is at address 0x0000 and also at address 0x0001 from the perspective of the microcontroller. Discarding A0 means any computed address doe not need to checked for its byte boundary. If the computed adddress is odd, you automatically round down, when dropping A0.
On systems that deal with only 32 bit words, we often drop off the last two bits of the address bus for the same reason.