The transceivers are potentially capable of stronger drive than the processor.

However, the transceivers will have increased propagation delay compared to the mux.

On the other hand, the mux introduces additional capacitance and resistance which (in my opinion) is more likely to become a problem before the latency of the mux does.

But the mux is a single-chip solution whereas the transceiver is two. Since this is an SPI bus that does not require a transceiver (like RS-232, RS-422, or RS-485) then the fact the mux is single-chip probably wins out.

For interfaces that do require a transceiver anyways, you are better off with a second transceiver since you end up with a two-chip solution either way. You also do not want to use a mux with interfaces that have ruggedized transceivers (ESD, high common mode voltage, etc) since the you are unlikely to find a mux that is just as rugged.

The transceivers are potentially capable of stronger drive than the processor.

However, the transceivers will have increased propagation delay compared to the mux.

On the other hand, the mux introduces additional capacitance and resistance which (in my opinion) is more likely to become a problem before the latency of the mux does.

But the mux is a single-chip solution whereas the transceiver is two. Since this is an SPI bus that does not require a transceiver IC (like RS-232, RS-422, or RS-485) then the fact the mux is single-chip solution probably wins out.

For interfaces that require a ruggedized transceiver IC anyways (like RS-232, RS-422, or RS-485) you do not want to use a mux if you can help it because you are unlikely to find a mux with that is just as rugged as the transceivers (i.e. ESD protection, fault protection, high common-mode tolerance, etc). Not to mention that if a transceiver IC is needed anyways, you end up with two ICs either way (either transceiver+mux or transceiver+transceiver) so there is no chip count advantage to going with a mux in this case.