How safety nets emerge naturally from people caring about each other, and how we can use the internet to safely multiply this effect many times over.
Insurance is a huge industry, yet is riddled with social problems from fraud, to price gouging, to unjust rejection of claims. In some cases like medical insurance, claim rejection can mean life or death, and so for those affected, there isn’t a more serious topic. In this article I’m going to describe a solution to the insurance problem. It’s simple in concept, but delicate in execution, because it depends on the informed participation of large social networks.
Care is Natural
In small groups who trust each other like a family or a village, cooperation and help can be completely second nature. This has been called everyday communism, but does not require any kind of government or formal organization. It’s just the social tendency to help what you can.
This is easy to think about in terms of individuals choosing to help one another. It’s obvious that we each have the free will to offer our help whenever we like. There may be social pressures that encourage us to, but at the end of the day, when you lend a hand you’re voter, representative, and contractor all rolled into one. Or maybe you’re the insurance broker, claims agent, and contractor all together. Those examples can seem silly, because the roles of the current system don’t really make sense when dealing with personal relationships, as they are designed for dealing with strangers at arm’s length.
Trust is Transitive
There’s a profound thing that people do all the time, and have done for millennia: We vouch for each other.
When we do this, we accomplish something powerful: We create peace or even cooperation between two strangers, with our own reputation on the line binding them together. If my guest at a party creates major mayhem, I might not be invited to the next one, and I might also be hesitant to invite them again!
The threat of exclusion is the ultimate social danger. There is maybe nothing worse than becoming a pariah. As social creatures, our relationships with the people we love may be the most valuable things in the world to us. Many of us would die for the people we love.
It should be no surprise then that vouching for someone works. In formal terms, I can safely trust you to put me at risk up to how much you value our relationship. Your reputation serves as a kind of collateral, and that’s why in economics, this phenomenon is called social collateral .
I can safely trust you to put me at risk up to how much you value our relationship.
The most important observation here that I want to stress is that this value equation applies at every link in a social graph, and also every segment of links between two people (even if the chain forks and reconnects)! That’s what I mean by trust is transitive. When you invite me to your party, you clearly believe I value our relationship more than I value ruining your place, and when I bring a guest, I believe they value my relationship more than ruining your place, and so by transitivity, you’re able to trust my guest simply by trusting me.
And for the game theory fans out there, this relationship is water-tight: There’s no economically efficient bribe you can offer any of these people to break their social bonds. If you could offer them more money than the relationship is worth, it would also be enough for them to repair the damages they create, by definition, and potentially rip off the briber in the process!
Applying this to the insurance example, while you might eagerly pay all of a spouse’s medical bills, what would it take for you to confidently chip in for a friend? Or a friend of a friend? Or a friend ten times removed? Or an enemy?
It’s easy to understand the trust you have for a friend in need, but if we want to safely exchange our social safety net with people far socially removed from ourselves, we need to understand how much we can safely trust each other, and we need to create tools to put that trust to work for us.
For the easiest example, we will use an arbitrary unit for measuring what is at stake, and for the sake of an intuitive unit, let’s measure it in US Dollars ($). For example, I may unconditionally trust you with $100 of my money, because I believe my relationship is worth more than $100 to you, and so those $100 are at stake. In practice, anything can be lent.
If I had a special kind of bank account, I could give you a debit card that will always allow you to withdraw up to that $100, and I can even give you any context I want, like “This is just for emergencies, please try to minimize using this.” That extra context adds actual and tangible terms under which I might hold the withdraw against you, and since you’re protecting our relationship, those terms are meaningful to you, even outside of a legal context.
Now let’s say you don’t have $100 to spare yourself, but you have a friend who you care for very much, and so you might want to grant them the permission to withdraw the $100 I pledged to you. Will I mind, at that moment of withdraw? Well this is up for us to decide as friends. If we agreed it’s useful for any emergency you see fit, and you’re willing to explain the situation to me, and possibly even promise to pay it back, there are many conditions where I might be perfectly fine with you delegating my trust in you.
So here’s the biggest mental leap: If everyone granted all their friends some spending limit for emergencies, with the understanding that it could be delegated to people they trust, and so forth, a naturally dense social group could develop a robust and resilient support network that does not charge premiums or co-pays.
Debit cards may not make this easy, but cryptocurrency can.
Fair Claim Evaluation
The claims process in insurance today is adversarial: When you apply for insurance, you may be rejected for a pre-existing condition, because the mere risk of paying out is antithetical to accepting your premium payments. When you submit a claim, your rates immediately increase because you’re assumed to be an increased risk. A claims investigator working on behalf of the insurance company looks for any clues that you were looking for any easy pay day, and all for good reason:
You don’t value your insurance company (beyond what they may pay you), and they don’t value you (beyond what they hope you’ll pay them). As strangers, the incentive to cheat is constant, and no matter how strong you make your investigations, there will be cheaters, there will be fraud, there will be rejected claims, and over-payment, and people who go uninsured because they appear too risky.
To contrast, let’s consider a situation where four people have shared an allowance in a chain:
- Alice, a wealthy baroness.
- Bob, her nephew.
- Charlie, Bob’s childhood friend.
- Darlene, Charlie’s coworker at the grain mill.
Imagine for example that each person trusted the next one as follows:
- Alice entrusted her nephew Bob with $20k in case he should ever need it.
- Bob entrusted his friend Charlie with $10k of Alice’s allowance in case Charlie should ever need it.
- Charlie trusted Darlene with $2k of his allowance from Bob.
An accident at the mill takes place, and Darlene finds herself in the hospital, needing an x-ray. She opens an application on her phone that displays her available $2k allowance. The x-ray will cost $500. She types a brief explanation into the app, waves her phone over the cash register, and clicks spend.
The friends all receive a notification of the expense and explanation. At this point they each get an option: Leave it as is, or adjust (either revoking or refreshing) the allowance.
- Charlie has to decide whether he thinks this expense was legitimate or not. Fortunately, he was at the factory, so he knows it was. He adjusts Darlene’s allowance, increasing it by $500 to cover the x-ray.
- Bob has a chance to review the expense, and if he thinks Charlie is colluding with Darlene, he could freeze their allowance, but for now he trusts his old friend, and snoozes the notification.
- Alice doesn’t bother herself with notifications of expenses less than $10k, so she doesn’t get a message in this case, but she regularly reviews the feed of transactions that her vast network of delegated wealth is enabling.
As you can see here, in a peer to peer insurance system like I’m proposing, each mutual link in the social web serves as both the insurance agent and the claim evaluator, but there is a stark difference between their perspective, and that of a judge, jury, or traditional claims adjuster:
These people have no presumption of impartiality or objectivity, quite the opposite: They each have a social bond at stake, and they are each expected to evaluate the claim in the context of this person they know, who they probably have a variety of ways of learning more about.
Even a claim investigator can be bribed, because they only have their job at stake, while the claim could be worth much more than their salary. In this case though, each person has their own assets and relationships at stake, and so no external incentives can irrationally influence their internal logic. The incentives are aligned.
The Road to Cooperation
If this seems fanciful, it should. People have tried to build similar systems before, and given up, proclaiming normal people would never be willing to participate in such a complicated system. I would emphasize that while people may be stubborn learners, we can also be brilliant designers, who lower the barrier to entry, and make new things like this accessible.
It may also seem impractical that a mill worker is four links removed from a baroness, and I grant this is a contrived example, but want to point out a few things: First of all, each of us has more than a single friend, and so there is actually a compounding network of support the more people we know participating in such a system. Also, while the folk wisdom is that any of us can be connected by six degrees of separation, Facebook has demonstrated that in dense digital social networks, this is actually closer to four.
A few things need to happen before this kind of fantasy can become practical: A mobile wallet application needs to enable efficient and scalable sharing of permissions. And once one does, the grassroots work begins of teaching people how to use and leverage this technology to build themselves safety nets. Everyone will be able to participate by helping extend the network: Your own power will grow as you share what you have more widely.
This article was a bit of a fake-out. It talked a lot about insurance, but actually it’s up to the people in the network what their allowances are good for. If they’re industrious, this same framework could be used as an alternative to lending or investing, too. Since each person defines the context and conditions of their social allowance, this framework can describe a peer to peer version of almost any financial relationship.
I used dollars as the units in my examples, but in practice we may not all have many dollars to trust each other with, but we can promise other things, too, and as long as we stay good to our words, digitizing our true capacities could empower groups of people that have strong relationships but little access to capital.