So-called single points of failure “are undesirable in any system whose goal is high availability”; so teaches Wikipedia.
Now, systems for social security are, on the one hand, designed with this in mind: resources that need to be highly available in individual life, such as health care and nutrition, should not depend on the individual’s ability to pay for them, as this would be a SPOF. But on the other hand: in the Western “developed world”, social security systems are central, government based systems, which makes these systems fail if the state fails to pay for them. Which is a very real danger, as can be seen from the German pension insurance, which crashed for demographic and other reasons.
Again, Wikipedia teaches us the principles of reliability engineering to avoid SPOFs: reduced complexity, redundancy, diversity (of implementation), and transparency (Wikipedia article “Single Point of Failure”). All of these are not, or only to a low degree, implemented in these centralized Western social security systems. The extended family was able to perform the task better, as it was a highly redundant system, existing in tens of thousands of instances in a society. However, this type of family does no longer exist in our society.
What to do? How about experimenting with the idea of the “small autarkic community” as a shared risk community. This is not about anonymous insurance, as this is doomed to fraud and inefficiency: it lacks for example tools to really educate and motivate people to lead a more healthy lifestyle. The autarchic community is about the smallest thinkable group that can bear all risks of individual life on its own; which might be about 100 people. They would live in one house together and care for each other in all aspects of the shared risks, like caring for the sick and elderly, supporting the unemployed financially (and emotionally), etc.. One could enter at all times if one is admitted (depending on ones currrent issues, and the ability of the community to bear them, and the number of free places). And one could leave at any time, to switch to a like community at another place or back into government-backed system. But one could not be kicked out of the shared risk community; which is necessary to be a true “social insurance” type of community.
Such a 100-person community could take over other tasks that are simplified by load balancing or risk sharing in a community. Like child care. This would then be called “human crechès“. Yes, why not use this animal concept in human culture also. This is way less expensive than government’s child care units like kindergartens etc.. It can also be implemented in even smaller groups: for example, five families could form a crechè for the time before the parents return from work. Which would mean that from each couple, one person has to take one day off in two weeks to organize the crechè at this family’s home.