Distributed Insurance Network

[dingdongbitcoin]

This is just something I've being thinking about for a while now. It's a lot easier to get it if you drop any type of knowledge you have about how an insurance product works.

Beginning with a story:

"There was this little village with this really smart girl. She said that each family should put some money into a pot. When one family needs money to take care of the dead, they'll be able to take it out of the pot."

That's basically life insurance in a distributed way. Everyone knew everyone, and there wasn't insurance companies running the show. Although we probably all know how their reserves went...

With the age of internet, and the distributed ledger-of-account technology behind Bitcoin, I think it's feasible to replace the role of an insurance company with a piece of software. I'm not saying it should nor better, but feasible.


Case for whole Life insurance

The system would run as a public service, starting with a simple model having mortality as the only variable. Here's a starting model we can use. There are also stochastic methods if we want to add more variables. I think it's possible to trench the computation time for that in exchange for fees.

Gompertz–Makeham law of mortality
http://en.wikipedia.org/wiki/Gompertz%E2%80%93Makeham_law_of_mortality

Stochastic modelling (insurance)
http://en.wikipedia.org/wiki/Stochastic_modelling_(insurance)

Since we don't have an insurance company managing the products, I'd imagine reserves required to be lower, leading to a lower monthly fee. If fees are collected in bitcoins, and we have an inherently deflationary asset; simply holding would generate "interest", that could be used as risk free interest.

Actuarial reserves
http://en.wikipedia.org/wiki/Actuarial_reserves

Each new policy holder unique address to send their "bitcoins" and the system monitors the accounts to notice any lapses. Special keys will be distributed to reliable 3rd party claim assessment agents to sign off on claims. There will also be a referral system so that  a % of new entrant's payments will go to the seller of the policy.

Iceland would be a great place to test this, since death are public data. Whole life is used here because its simple to check if someone's dead. I'd imagine with the advent of GPS tracking, auto insurance could be programmed this way as well.

Having an attractive and fraud resistant incentive system for the involvement of 3rd party claim assessment and underwriters would be challenging. While acknowledging this fact, my focus is to bring about a working prototype of this technology.

I would really appreciate any feedback, and questions. I will answer them the best I can. They will also help me distill my thoughts.

Thanks for reading!

[vm_mpn]

Yes, thought about this for a while, more in terms of FDIC type of insurance paid by established exchanges and crypto financial firms to Bitcoin Foundation, and in turn Bitcoin Foundation as an insurer would guarantee each user account of the insured institutions up to X amount of coins. I guess this would ultimately affect fees those financial firms will charge their customers and will make it much harder for new exchanges and financial firms to compete with BF proven and established institutions... However, this step will also stabilize and will add respectability to the currency which seems to be bouncing from crisis to crisis for the last few years.

After Mt. GOX (Gold Member) fiasco, if Bitcoin Foundation is to have any relevance should be thinking about this as well.

As to your idea - it's great, providing you will find brave and competent enough player/s to take this huge task on... I'm getting dizzy by just thinking of the amount of false claims these insurers will be faced with if they are not diligent and careful. 

Sorry for mumbling - I'm just a simple tech dude and have very little knowledge on this subject.   

[dingdongbitcoin]

Just found these guys,

https://bitcointalk.org/index.php?topic=130137.0

If they work out then fraud would be more difficult

From a design perspective, one requirement of payout would be that the accounts have received payment 12-24 consecutive months. After that, you can have a regression curve to test the fraud rate at additional number of months, and have that as a variable when reiterating the model.