"There's arguments for and against both methods. Back when I first started programming, memory was a luxury (we're talking 80's personal computers here), so it made a great deal of difference. These days, however, choosing an integer type which most closely matches the values stored is largely an academic process which costs developer time"
Actually, memory is even more of a luxury than ever, since your caches have limited size and main memory bandwidth is atrocious, compared to the speed of your execution units. Using 32 bit fields for stuff that safely fits into 16 bits can cost you for that reason alone.