OT : Strange production method

This has nothign to do with vintage computers, really. It related to a telepone, and not a particularly old one... But since some of you have knoweldge of production methods, I wonder if you have any thoughs on this. I have a basic 2-piece landline telephone here. Inside is a PCB with qutie a few discrete components (transsitors diodes, passives) and a single IC, which is clearly the dialer. Now the PCB has pads for an 18 pin DIL pacakge, but that's not waht's fitted. Instead there are 1 9-pin headers osldered to the main PCB. On tol fo those is an other little PCB connecting to the 'top 16' pins -- pins 8 and 9 of the origianl DIL position are not connected -- but then they go nowhere o nthe main PCB either. On this litle PCB which is fitted track-side up is a single epoxy-capped IC. I traced out the connections to the dialer keypad, ceramin resonator, poerr and ground in the hope I could identify the device. I then desoldered the header pins from the pin PCB and pulled the assembly out. On the underside (plain side) of this little IC-carrier is silk-screened 'SC91710A'. I ahve typed that into datsheetarchive. It exists as an IC. An 18 pin DIL IC that's a telephone dialer circuit. And all the pins I'd traced match up perfectly. Pins 8 and 9 are for a handsfree function that's not implemented here. So this telephone was clearly designed to use that dialer IC in the DIL package. My question is why was this subassembly made and fitted? I can understand that a driect-on-board IC is cheaper than fitting a DIL pacakge in many cases (for all I find such things objectionable!). But I can't beelive making up this daughterboard and fittign it with the header strips is cheaper than a DIL packaged IC -- is it?