While it may not be economical for you personally to figure out the fault in your $500 TV, it might very well be economical for a sufficiently large repair business if access to schematics was easy.

There is no reason why you couldn't have a repair chain that can massively optimize the repair process by both collecting information about typical failures of particular products and by developing tools for diagnosing in particular products as well as their own spare parts, so they could offer things like free analysis for common problems and fixed-price repairs for those problems, for example.

Repair doesn't need to have a 100% success rate to be useful, but success rate and whether repair is economical depends heavily on availability of information. Every hour that has to be spent reverse engineering a device before you can repair it makes a few more percent of defects non-economical to repair.

Also, mind you that if it were an engineering goal, the increased use of processors in devices instead of specific circuitry could actually help a lot with repairs. While you might have a harder time figuring out what's going on with a scope probe, a processor in the system could actually help you with diagnosing problems, isolating faults, generating test signals , whatever. Instead, JTAG is disabled in sold products so you can't even use the diagnostic functionality that is built into the device.