So what you're really saying is "Take the number of [batteries] in the field, A, multiply by the probable rate of failure, B, multiply by the average out-of-court settlement, C. A times B times C equals X. If X is less than the cost of a recall, we don't do one. "
The actual answer is more nuanced.
FAA regulations define 5 levels of critical for safety systems: levels A through E.
Level A is for things that can knock a plane out of the sky when they fail; for example the stall speed alarm.
Level C is for things that can cause injury or at most a single death; for example, the cabin pressurization system
Level E is for things that don't affect flight safety; such as, in-flight entertainment or the microwave in the galley
For reference, I wrote the software for cabin pressurization systems. It's level C (hardware == B), which means that failure in pressurization is an emergency situation, but isn't expected to kill everyone on board. The masks drop and the pilot immediately dives to under 10,000 feet to restore breathable air.
If the cabin fills with smoke, it's not life-threatening per se. The pilot can override the pressurization system and "dump" the cabin atmosphere, and it clears pretty quick. (The captain also dives to under 10,000 feet if necessary.)
The battery catching fire isn't a problem SO LONG AS the fire itself won't cripple the aircraft. The battery underpowering the plane when the alternator dies MAY BE a problem which would kill people.
The people who design these things take these levels into consideration, and the general rule is "fail safe". If you can't "fail safe", then "fail in the least dangerous way". In my experience, the engineer must make many choices when designing an aircraft unit. The answer is always "do it *this* way, because if *that* happens it will be less dangerous.
Let's wait and see what the investigation uncovers. Here are some Cliff notes:
1) Li-Ion batteries might behave differently at altitude (cabin pressure is reduced while flying)
2) The battery may be performing to spec, while trying to compensate for a more dangerous problem
3) Smoke in the cabin is not as dangerous as you might think
4) Things that burn are designed to not damage things when burning
5) People who design aircraft are pretty smart, and have a generally high moral standard.
6) People who investigate aircraft incidents are really, really thorough, and have a good track record.
(Note: Glossing over some details to make an easier read.)