That's probably good for the robo-callers. Just like with horrible spelling/grammar errors in spam emails, they don't mind if the smart folks can figure out their pattern. They are after the suckers born every minute.
* The fact that many geographic numbers are in fact non-geographic means that the idea of being "in" an area has long since gone out of the window.
* It's not done to avoid extra charges. Call cost is determined by routing and tarrif, not caller ID. It's done pretty much for the reason that bonobo_34 supposed.
BUT.... I've had the opposite effect where they spoof to where I physically was regardless of my phone number.
I recently contacted a phishing prevention company for a quote. They knew where I was physically located, and called me using a spam number of my local-actual code. I told them flat out that was a bad idea and to not do that. Nothing made me lose initial trust more than a legit company using spoofing tactics.
It was the company that bought Wombat Anti-Phishing. They even told me they didn't do it to "prove a point" but rather just they get people to "pick up the phone more often".
The “feature” started innocuously enough.
When a call comes in from those ACs (the one corresponding to the phone), I reject it. Very easy!
Something that works for me is to have a phone number in a low-value area code. My Seattle and New York numbers get robocalls. My phone numbers in certain other parts of the country are left alone.
It seems most spam calls either fake a source number that's in my phone's area code, or use a completely random source. Almost any inbound call I actually expect without an address-book entry would come from my (new) local area code, so I can pick up a call if (in current area code) OR (in address book). So in essence, my current area code is a 3-digit passcode.
Perhaps a nice trick if one is allocating a new VoIP number -- at the cost of looking like a non-local when giving it out, you can effectively screen for actual local calls...