Yeah, it's a shame that the phrase cost benefit analysis has gotten a bad reputation.

Well, it's not "no" safety devices. There are strict rules about how high bus seats have to be, and how rigid so as to limit how far kids can fly in an accident, and to minimize injury when they get fired face first into the seat in front of them.
They also rely on the bus being much taller, and having far more mass than most vehicles, so kids are less likely to get tossed around.

Like it said in the linked document, perfect seatbelt use was anticipated to save two lives a year in the US amongst school children, but spending that money on pick-up and drop-off safety would save more lives, and anything that made busses less available would cost them.

It's largely moot, since more recent data says that it's now worth it, and so the recommendation has shifted to adding them, although it's not mandatory since it's not a super powerful advantage.

It's seriously not about fire safety, the data is right there as well as their rationale. The addition of seatbelts would save lives from bus accidents, but likely increase fatalities from decreased ridership.
NHTSA believes the cost in lives and dollars isn't justified given the data.

Also, it looks like they reevaluated, and now believe that they are worth it given new information.