While I generally agree with ChipH's response, I must iterate that telephone line audio is more difficult to access remotely than IP data. Certainly not impossible, and it is trivial if you have physical access to the local loop of the target.
And all encryption technologies are susceptible to man-in-the-middle attacks. Let's face it, how frequently do you verify that the cert that you received is the actual cert you expected?
But in all of these cases, the data is far more likely to be accessed when it is at rest, than in transit. If I am interested in your banking information, I am more likely to attack the bank's server or your home machine than to attempt to decipher the encypted data in transit or to tap your phone line. Mainly because I get more information for less work (your banking information, plus anyone else's in the database).
And deciphering modem data is non-trivial at rates greater than 2400 bps because of the phase-shift-keying that is in use. To decipher it properly, you need to be in the 4-wire portion of the network, or you need to install your own hybrid. The data traveling in each direction uses the same frequency space, so you must separate the two directional components. Only the actual subscriber loop is 2-wire, so now you must either collect the data after the SLC (where it is multiplexed to a DS-1 rate) or in the CO switch itself.
pansophic
