The key difference seems to be German private health insurance contracts are long-term affairs. Multi-year, or perhaps even lifetime? US health insurance contracts are typically a year at a time. So German companies have to reserve for costs projected to occur far in the future because they are liable for them, while US companies have no idea if their customer will still be around in 20 years.
My guess would be there's a healthy dollop of regulation pushing the German insurance market into that shape, otherwise you would probably see short-term insurers outcompeting long-term insurers since they wouldn't have to do old-age reserves and could therefore charge lower premiums. Consumers tend not to be nearly as good at rationally planning for long term expenditures as are actuaries.
I think you are correct in your basic assertion that no one wants manual memory management for its own sake. What they really want is sufficient performance for their use case. The benchmarks you usually see are throughput oriented, and on small heaps. If you have tight latency budgets and/or huge heaps, the performance is not close.
Optional manual memory management sounds great, but I'm skeptical it would work well in practice. The reason is that if the language default is GC, libraries won't be designed for manual memory management, meaning it will be hard for your manual code to interact with data structures created by non-manual parts.
My guess would be there's a healthy dollop of regulation pushing the German insurance market into that shape, otherwise you would probably see short-term insurers outcompeting long-term insurers since they wouldn't have to do old-age reserves and could therefore charge lower premiums. Consumers tend not to be nearly as good at rationally planning for long term expenditures as are actuaries.