Matt Peterman — Social Proof Insurance and Dispute Resolution
Full video and transcript below.
On April 10th, Mattereum hosted the third Internet of Agreements® (IoA) conference at the Google Campus in London. IoA® is a vision for global supply chains and logistics, integrating national laws and regulation with international commerce through the application of technology such as blockchains and smart contracts.
InsurePal CEO Matt Peterman presented his project’s approach to reinforcing trust and reducing uncertainty in business transactions through a peer-to-peer insurance platform using blockchain and smart contracts. At the heart of InsurePal’s design is “social proof,” whereby friends and associates can endorse an individual’s credibility, facing reputational or financial penalties if that individual fails to perform. Such a system could be used to increase trust in business deals (endorsing an individual’s trustworthiness), or to lower the cost of motor insurance (endorsing an individual’s driving skills).
Peterman invited Mattereum CEO Vinay Gupta on stage to help provide an overview of the dispute resolution process. A Ricardian contract, a combination of a natural language contract and executable code, forms the basis of a financial agreement over which the parties might come into dispute. A mediator can help the parties resolve the matter amongst themselves in the event of a mere misunderstanding or mistake. If the dispute is not resolved, then the matter goes to binding arbitration. Gupta then elaborated on how a P2P insurance platform would need to scale to keep a sustainable balance between the size of the insurance pool and the cost borne by the individual.
Matt: How many of you would vouch for your friend with £20, hands up?
Comment: It depends on the friend. [laughter]
Matt: Okay, so we have around 30% of people saying that vouching for others could be a solution for doing identity. I’ll start with a short video, maybe it’s the easiest explanation for what we are trying to do.
On one side, we are trying to move the existing insurance models, where everything is held centralised. As Michael said at the beginning, when Equifax came to UK, the same Equifax that holds most of the data of their clients is centralised. What we are trying to do is to put some of that data away from the insurers.
On this slide we have two ways of let’s say underwriting each person. I can check your Facebook, I can maybe even use Cambridge Analytica to do that [laughter], I can put some gadgets in your car, I can… My background was fraud detection in large insurers, so we usually checked when there was some fraudulent claims, we check their Facebook profiles, we check Google Links, we check any other social media to see the clients. The other way is we could also do it is ask friends whether they would endorse them. I saw that only 30% of people are willing to trust their friends, but that’s still a good amount, and we are trying to build a model based on that.
Here is how it works. If Vinay has car insurance for £1,000, I say, “I know him, I know he’s not a good driver, but I would still vouch for him with around £200 in order to reduce his premium,” and in that sense I am giving additional information to insurers and helping them to… On one side I’m reducing the premium for Vinay, and I am helping to do a better underwriting. You can do the same thing for other things also. We currently have around five clients using this kind of a social proofing for gig economies. For example, we have a client in Switzerland where if you rent an apartment, since they don’t know you, they ask that you put at least three rents in advance in a deposit, which you get back when you return the keys. What we’re offering is that you ask three or four of your friends whether they’d endorse you, whether they’d put their credit cards, we just check their credit cards whether they’re valid, and if the cards are valid then most friends would vouch for the same amount.
On one side that brings cheaper premiums, improved profitability, and it also gives benefits to society.
Blockchain can enable the expansion of that model to a whole network, so a kind of a recursive, transitive deductible which can be used if we have a lot of identities or a lot of clients in a space.
In the case that I vouch for somebody and there’s a dispute, then I always call Vinay to help us, so I would ask maybe to help how we resolve some of the smart contracts which we have, let’s say of the insurer or in gig economies, how to enable them to do it quicker. Vinay, can you help us?
Vinay: Sure. Do you need me to judge something?
Matt: Basically, currently most of the insurers are resolving the contracts in court. Some of the contracts could be also resolved easier through smart contracts. However, as I heard today, up until now only one smart contract has been resolved in the UK courts.
Vinay: Yes, this would be Ian’s contract that you’re referring to, from the 90s. There’s no doubt at all that arbitration is binding, it’s a very old, very stable process, there’s no doubt at all that arbitration works. The tricky part is connecting the smart contract to the arbitrator. Because the smart contract is just a chunk of code that lives on the blockchain, and there’s nothing in that chunk of code which can necessarily even be tied to a legal persona. So if this code is signed with a key, how do we know that that key is your key? Because if you give a copy of that key to somebody else, either one of you could be the person that signed the document. Tracing the necessary chains of legal cause and effect between a smart contract and a nominated individual who can then consent to the arbitration, all of that stuff is the tricky, step-by-step legal procedural work that you have to do, if you want to be able to use binding arbitration to settle these kinds of disputes.
Matt: How would we resolve disputes when people are vouching for each other’s identities?
Vinay: The simple case is that you first take the money; if something goes wrong, the first thing you do is you take the money, and if they decide that they would like their money back, now they’re the one that raises the dispute. You have things that looks like terms of service agreements, which say, “In the event of a dispute being raised, you first talk to a neutral third party that will try and have a dialogue about what went wrong, and that neutral third party will tend to resolve it in a way that everybody consents to,” so you have mediation step, presumably. In most cases the mediation step will resolve the problem, because it will be a simple misunderstanding.
But remember that you are holding onto the money at this point, so if the mediation step doesn’t work, you can basically shrug your shoulders and your customer will be like, Damn it, this is not right, this is just not how it’s going to be,” and at this point they can escalate to a more formal process, at which point the decisions become very binding, they have full legal force, and as a result people tend to do things like bring their lawyers. This is where there may be a mismatch between the sums of money in play and the degree of inconvenience required to have a litigation, and the bridge for this mismatch is of course mutualisation: everybody pays an additional 25 cents on their transaction fee, that money goes into a bucket, and that bucket then pays for the legal expenses of the parties in the event that a formal arbitration occurs.
Matt: So basically for each premium we would add 25 cents for that part?
Vinay: And you wind up with a kind of bucket of money which is like a legal defence fund, which is used to pay the costs of full process to do asset recovery. Because the thing is that the integrity of the process is much more important than the efficiency of the process. We’re all used to completely unaccountable dispute resolution mechanisms like eBay, where if something goes wrong on eBay, your chance of actually sorting it out is pretty low. Theoretically they have mechanisms in place for sorting it out, but practically speaking those mechanisms are kind of weird and opaque and it’s not like you could just talk to somebody and explain what happened; the decision making event tends to favour the buyer, but the procedural knowledge is all with the sellers, so you wind up with two unequal forces contesting.
All of those mechanisms… You just kind of play the odds and think that it will all be okay, and you continue to trade on eBay and once in a while you lose some value, but the general feeling that you get is that eBay is kind of a sketchy environment to do business in. If the alternative was that they had very, very high-quality arbitration, so that when something went wrong it was handled by a really competent authority, even if the transaction fees went a little higher, we would all feel much better about eBay. And this is how it works in high trust societies, everybody pays taxes and the taxes pay for the courts, and there’s been an enormous amount of work in the UK on streamlining the cost of core processes, not just for small claims but further up by doing things like changing the way that case management is done. So rather than having an environment where people could spend an infinite amount of money on lawyers, there are things like caps on legal fees that’ll be awarded if the dispute is ruled against somebody, and all of these mechanisms have been done to try and streamline these processes.
Matt: But isn’t it true that people are let’s say… We see in insurance that for small sums they usually don’t go to court, because they simply say it’s a lost battle.
Vinay: This is why mediation is important. If you have a mediation process where a neutral third party comes in and tries to make it work, in most cases just having a neutral third party there will resolve the problem. Somebody will look at it, and the agreement might be “Oh yeah, we have made a mistake here,” or “Why don’t we split this with you 50–50?” and you can resolve it, because the mediator is backed up by arbitration. If the mediator can’t resolve the problem, there’s always a way of escalating, and because we all know there’s a way of escalating, we know that justice is available, and at that point we are more willing to trust the processes at the lower level. If the mediator’s word is final and there’s no way of going above the mediator’s head, then how do you know if you’re going to get a bad mediator? You tend not to trust the process. But if you know that if the mediator does the wrong thing you have access to a higher-level system that will correct their errors, you come in with an expectation that the mediator will behave more professionally.
So the more rigid, solid process you have at the very top level, that kind of flows down through the rest of the systems. Even the front level technical support folks, if they know that it is technical support, then mediation, then maybe expert determination, then arbitration, you kind of have this tiered response, you know for sure that your technical support people understand that what they’re doing matters, so the first point of contact where something is going to get resolved is dealing with the extension of the authority of the ultimate arbitrators, rather than just the internal processes inside of a company. It’s very, very much like the kind of processes that give us trust in the national courts, and it’s that form of trust, this trust in a central authority that really understands the law and will administer justice fairly, that we want to bring into the blockchain space.
Matt: What we’re seeing is we have a lot of requests from so-called gig economies, where people are saying, “In order to onboard somebody to a renting platform,” they say, “If you bring additional social proof, we will put you on the platform,” but still on top of it, if I rent your car and if I crash it and the cost is 70,000 euros, your friends will vouch for 5,000, but the rest has to be covered by insurance.
Vinay: Yes, absolutely. This is the question about how big is the risk pool. If you have these kind of approaches where the risk pool is basically five people, you can’t get the kind of system performance that you want. You need a much, much larger risk pool, here’s your risk pool; if that network goes out to be 50,000, you would have the size of a typical insurance pool. If it’s five million people, that would be something like house insurance, and when the pools get really large, that extra 25 cents or the $100/year you charge somebody for insurance turns into such a lot of money that it covers almost all risks.
As Michael Mainelli was saying earlier on, there’s a whole question about scale. You need network effects to get these things working, because you have to push over the hump where you don’t really have enough insurance to make it work because the volatility is too large. You need to get lots of people with the same kinds of risks in a bucket, and that’s how you begin to smooth the risk. If you have something like that… If you’re dealing with $70,000 worth of risk, you’re going to need tens of thousands of people before you could split the risk so finely that nobody notices. How much insurance claim can you pay in a year before you feel the pain, maybe $100–200? So a $70,000 risk is going to have to be spread over 350 people, and you can sort of estimate the kind of risk tolerance of each individual and say, “How much is in the buckets?”
Does everybody know how Lloyd’s of London works? You have this kind of social network model, and at the top end of that you have a bunch of people who have yachts and country homes and Rolls Royces, and if something goes badly enough wrong that these people are asked to pay the claims that Lloyd’s is generating, these people begin to have to sell their country homes and their Rolls Royces. There was a famous bloodbath — how long ago was the bloodbath, 15–20 years, something like that? — where there was just a series of unfortunate events, and an enormous number of wealthy English names got hit very, very hard, and that caused a bunch of restructuring of risk at Lloyd’s.
The notion is always that you’re going to have to basically spread the risk thinly enough that each individual could tolerate it, and that applies as much to legal fees as it applies to the actual settlements. Because the difference between you being fined and you having to pay an expensive lawyer, it’s all money coming out of your pocket. So you need to find ways of pooling your risk of access to the expensive lawyers and pooling the risk of the actual payments.
Matt: That’s why in terms of scaling we think that the insurance business is one of the ways to scale. Because to onboard large amount of clients, insurance is one way. We see in the UK that if we offer a client an insurance premium that is at least £80 cheaper, they might consider to go and use the social proof insurance. We also have really good feedback from Dutch insurers, maybe because you had this event in Lloyd’s.
Vinay: Yeah, we’ve been warming them up for a while.
Matt: That’s the concept in short. If there any questions, I would be glad to answer them.
Vinay: I think there’s one thing that we should talk about here a little, which is the sort of notion that we’re up-ramping towards the future, there are a series of small steps: the first insurance product, there will be a bunch of disputes which will happen on the blockchain, and Ricardian contracts and dispute resolution will all kick in… All of these kinds of processes are steps towards a long-term equilibrium, and the long-term equilibrium is that you get a vastly higher quality of life, because you get much less uninsured risk.
All of these little risks, like you drop your cell phone and it breaks: if you had immediate replacement, where somebody will bicycle to you with the exact same phone and when it arrives your backups are restored, suddenly this is no longer a big deal, and the price of that service might only be £15–20/year, as long as it’s spread over a wide enough risk pool. And life is filled with crazy little inconveniences like this that are huge suckers of time and money when they happen to you, but they happen infrequently enough that they’re insurable. I like to think that we’re basically building out, you guys are leading on this process, a world in which basically if you have a little additional money, you can smooth out nearly all of the random little inconveniences, and that’s a combination of payment and service.
Matt: That’s a good point, and that’s also why we have a lot of interest from so-called gig economies, where people are using let’s say tools or cars for a few minutes or for an hour, renting them out, and to smoothen that process of onboarding and insurance, you need to have a different way to approach it.
Question: I have a question about the insurance side of this, because I think traditionally insurers have been very good at taking premiums, and then they’ll often so quick to pay out when an insurable event happens. Underlying insurance is this idea of good faith, because you can’t DD every piece of information, so you rely on the information that you’d been given in good faith. What happens when your third-party endorsement is given by somebody, and it turns out that that is incorrect? Does that affect your ability to pay out on the policy? How are you going to streamline this, to make certain that there’s an indemnity, if the social endorsement is incorrect, but will you still pay out?
Matt: Well, I would start from maybe from a different angle. Today if you pay £100 of insurance, you pay, in motor insurance you pay around £60 for claims, and out of that is 10–15% fraudulent. What we are trying is to reduce that part at least a little. There will still be cases of course when people will be have fake identities or will not pay, but that’s usually… We see that reduction of the claims by using social proof is between 20 and 40 percent. No system is ideal, our system isn’t ideal, but it can bring back some of roots of the insurance, that at the beginning it was mutual insurance among people, people insured among each other.
Vinay: This is where the peer-to-peer nature of the blockchain comes in. Because the overheads in organising an insurance pool peer to peer, where they’re insuring each other for 100,000 people, overwhelm existing approaches and you wind up with a central third party. If you wind up with basically protocol-based projects, you could build very large risk pools with absolutely minimal third party that really just does standards processes and dispute resolution. We do dispute resolution, you do the standards and product design, and you wind up with very, very thin intermediaries. I think that that potentially produces much, much more fluid and elastic and functional insurance markets.
All materials from the conference: http://internetofagreements.com/identity/
Join the telegram https://t.me/mattereum