The bus always seemed like the oddest part to zero in on. By analogy, an Opteron in 2003 was a 64 bit CPU with a 32 bit HyperTransport bus, but no one called an Opteron system 32 bit. The width of a particular internal implementation detail is a strange duck IMO.
All ALU operations are also more expensive with 32 bit operands. So 16 bit data bus, 24 bit address bus. Slower arithmetic with 32 bit operands. I never though of it as a 32 bit CPU.
I think part of it was that to hardware companies the bus width is actually extremely important - the whole system is built around it, and the programming model the software guys work with less so.
And then the other part of it is the marketing angle: everyone knew full 32-bit inside and out chips were just on the horizon. Downplaying the 68k’s 32-bitness would give them a selling point for the 68020.