Cell phones have encoding/compression delay on the voice part, delays waiting for a transmit slot if the air interface is time division multiplexed, sometimes a jitter buffer to allow for resends/error correction, and often less than ideal routing (lots of people are using numbers from out of state, which may require the audio path to traverse that state).

Originally POTS was circuit switched analog connection between you and the other party --- only delays from the wires, and maybe amplifiers. Nowadays POTS is most likely digitally sampled at the telco side, but each sample is sent individually --- there's no delay to get a large enough buffer to send, because for multiplexed lines each individual line is sampled at the same rate and a frame has one sample from each.