E-mail spam is mostly a solved problem. The nice thing about TCP/IP is that the other end of the connection is known and you can maintain blacklists of spam-friendly networks. All that remains are compromised computers that rapidly come and go. Since mail is not quite real-time, an SMTP server can refuse to accept messages that are flagged by analysis. Newer mail servers don't have the messy bounce or greymail issues of post-processing.
Telephone spam is a different problem. A call receiver knows absolutely nothing about the caller. Gimmicks like Nomorobo were long ago defeated by using a random (or trusted) local phone number as the Caller ID. I don't hear stories of telcos getting huge fines for supporting illegal telemarketers so keeping Caller ID broken is all profit for them.