Today, calling an A500 32-bit is correct, but when they were released they were classed as 16-bit. In your oppinion then, the Sega Megadrive was 32-bit (it used a 68000), even though it states 16-bit.
Why? Because in those ye oldie days a processor's bitness was based not on it's resgisters but the width of it's data bus. The 68000 ran 32-bit software but had a 16-bit data bus, hence a 16-bit processor.
Since then there have been more anomalies, the 386SX 32-bit regisers, 16-bit data bus, Pentium 32-bit registers 64-bit data bus. Because of things like this AFIAK we now think in terms of register width.
srg