Decentralized Insurance, a Building Block of DeFi
Risk and Insurance
Risk is an essential part of life. Everything you do involves some sort of risk; your house could be struck by lightning, your car could be rear-ended by a drunk driver, or you could suffer a heart attack out of the blue. I don’t mean to be grim, but hey, it could happen!
While some risks can be controlled by individual behavior, others cannot and despite our best efforts, we are at the mercy of the behavior of others. So, how do we control what isn’t in our control, and what is just random? One way could be to do nothing and hope nothing unfortunate happens. This is called risk avoidance.
Another way could be to mitigate the cost we would bear if something unfortunate does happen. Why cost, though? Well, because the cost can be measured, planned for, and protected against. This measurement, planning, and protection of cost for eventual risks form the principle of insurance.
Insurance is a way to decrease the financial cost of liability should it arise. It does not decrease the likelihood of the liability itself; you can’t decrease the probability of the weather, drunk drivers, or, to an extent, heart attack,(It would be amazing if you could though!). What you can do is decrease the financial impact of these events should they arise.
Since risk is all around us, and we are naturally ‘risk-averse’ beings, insurance is essential to modern society. From insurance against damage to a house during a natural calamity, like a flood or an earthquake, to insurance against a means of livelihood like farmer’s crops, or even insurance against the risk of the cancellation or delay of a flight, there’s insurance for everything because people can find risk in anything and insurance companies can potentially cover the risk for anything. Fun fact, a 45-year old Chinese woman named Li was arrested in April 2020 for claiming $424,000 in flight delay insurance for 900 flights she booked between 2015-2019.
To be fair, not everyone is a Li, and most people who take out insurance policies do so to protect themselves or their assets from future risks. But how exactly does insurance work and why are there so many forms of insurance out there?
Well, the second question has a simple answer – there are so many policies because people see so many risks in everyday life. From the house you live in, to the car you drive, the plane you catch (or not), to your life, everything has an associated risk. With risks comes the need for protection against its inevitable cost. The importance of insurance increases with the cost it covers. Take car insurance, health insurance, or life insurance. Any damage to the asset it covers (your car, your health, and yourself) will prevent future earnings. I mean, you can’t really earn an income if you’re dead.
Back to the first question, how does insurance work? Well, to understand this let’s break it up into a few important parts. This will also help us to understand which parts of insurance may be open to greater efficiencies through blockchain and decentralized finance.
In simple words, insurance companies work according to the same principle as banks. The fundamental function of a bank is receiving deposits from people who can save or lend their money (the lenders) and giving it to people who need to borrow money (borrowers). Banks are paid an interest rate by the borrowers, and banks pay an interest rate to the lenders. The profit banks make is on the difference between the two interest rates.
The assumption that banks go by is not everyone who lends money will withdraw it at the same time. Insurance companies operate along the same principle. However, instead of pooling all the saved money and giving it to those who need it, insurance companies only pool saved money specifically to cover a particular risk and give it to those who bear a financial cost based on the particular risk. Here, not only is the money pooled, but also the risk. This is called the principle.
We’ve all ‘pooled money’ at some point in our lives. For some of us, it’s in the context of a sports-bet made with our friends, betting on the outcome of our favorite sports team in a match or the end of the season. The pooling of insurance works in a similar way. Each player contributes a specific amount of money into the pool. But they don’t do this just once, they do this periodically, monthly, quarterly, yearly, etc. Just like in a sports-bet pool, every participant bets on the outcome of the sports team, in an insurance pool every participant bets on the outcome of the same risk, spread across different participants.
Here the participants face similar risks; home damage, car damage, medical issues, or the all-encompassing, death. The participants pool their money to cover for this risk if (or in some cases, when) it arises. Again, going by the aforementioned banking principle, not everyone will face the same risk at the same time. Hence, not everyone would require their insurance policies to provide coverage at the same time. So, the pool remains intact, or to use a more appropriate term liquid.
Since insurance is essentially financial risk management, the first step is measurement. The amount of money an individual pays every month, quarterly or yearly, for insurance is reflected in their insurance premiums. These premiums are not set in stone and vary depending on what is insured, its value and associated risks. I mean, you can’t insure your life and your flight in the same way, can you?
Take home insurance, there are a host of factors which can potentially affect insurance premiums. From weather, high-value assets within the house, neighborhood crime-rate, age of the building, credit score of the occupants, even your dog’s breed. Everything is a factor in deciding your premium, who knew? Anything which could potentially be a risk factor to the asset (in this case a house), from its structure to the behavior of the occupants (or their pets) is factored into the premiums.
Insurance at play
Let’s look at this in the form of an example. Adam and Ron live in the same town and the same neighborhood. They each take out an insurance policy on their respective homes, which have a market value of $100,000 each. The only difference between Adam’s and Ron’s house is the roof. Adam got his roof replaced five years ago, while Ron’s hasn’t been changed for the past twenty years. During the insurance process, the insurance company determines a higher premium for Ron, say at 12.5%. For Adam, despite the two houses having the same market value, and located in the same neighborhood, his premium is 10%.
The difference is because Adam’s house is less likely to suffer any damage than Ron’s. In the event of a hailstorm damaging both their roofs, the difference in damages will be reflected in the coverage. Adam’s roof is less likely to suffer damage [and hence a lower cost) because it was installed five years ago. Conversely, Ron’s roof is more likely to suffer damage because it was installed twenty years ago. Since Adam paid a lower premium and is facing a lower cost, his coverage would be small. However, Since Ron paid a higher premium, and is facing a higher cost, his coverage would be large. The premiums are a reflection of the risk the asset could face. The coverage is based on the cost to the asset because of the risk faced.
Same kind of insurance, home insurance. Same nature of risk, damage to the home. Same pool but different premiums, coverages, and deductibles.
Centralized Insurance: Specificity and Selection
Could you point out one common theme in the fundamentals of insurance policy? I’ll give you a hint, it’s a theme which the decentralized finance ecosystem is trying to solve for. If you guessed the theme of centralization you are correct! The very idea of ‘pooled money’ for a ‘common risk’ builds the premise of centrality, and in that traditional insurance is vulnerable because it has a central point of control, payout, and failure.
Here, the participants are beholden to the policies of the insurance company. This involves a slew of additional costs and complications.
Problems with Traditional Insurance
Sometimes risk is where you’d least expect it. While insurance companies do calculate periodic premiums based on the risk assessment of the parties they insure, those premiums also include a host of expenses, such as document processing, paper work, fees, employee wages and other administrative costs associated with a traditional financial system.
Further, in order to stay liquid (make sure they have enough funds to pay out claims) an insurance company will make strategic investments (on the funds received as premiums) to grow their net wealth, much like individuals and companies do with extra cash.
Why do they do this though? Good question! Well, because money loses value with time. It’s better to have it invested in something which can increase its value over time. Insurance companies can treat insurance premiums as income and invest it for better returns.
Centralization also presents a problem of selection bias, which is especially common in the insurance world. Not everyone can get the coverage they need and not everyone who has an insured loss gets paid. This is because insurance-providers need data to understand risks and pay claims, and enough risk pools to cover costs.
The failure of selection bias is important, so let’s understand this with the help of an example, with a specialized but increasingly popular form of insurance, especially in the decentralized finance world.
Sundar is a farmer in India who wants to insure his crops against drought during the harvest season. In such a situation, he takes out crop insurance on his estimated produce. For this, the insurance company will look at a number of factors. From weather forecasts, historic crop yield, fertilizer rates, mandated or market selling price, etc. This is done to gauge the payout if a drought occurs based on the harvest and its selling price. Other farmers within Sundar’s community would also take out a similar claim.
Selection bias arises when the drought adversely affects a part of the ‘pool’ and not the entire pool. In this situation, only Sundar’s and his immediate neighbors’ crops get insufficient rainfall. The rest of the community does not, resulting in an unequal outcome. Here, the insurance company would not be willing to give out the full claim to Sundar and his neighbors because the entire pool was unaffected. It should be noted that crop insurance is comprehensive covering several risks from pre-sowing to post-harvesting. Consider this a TL: DR of crop insurance.
In many cases, insurance companies do not even offer crop insurance policies to specific areas. This might happen because the risk is too remote for insurance companies to effectively calculate. In some cases, there may not even be enough people seeking insurance.
Cases like these, either because of specificity of risk versus selection bias or for remote insurance seekers without an insurance avenue, the same problem persists. That is the problem of a central point of control, payout and failure, or the overarching problem of centralized insurance.
In centralized insurance, the resources are “pooled” centrally with an insurance company. This central company:
- Assesses the risk
- Determines the premium
- Determines when the liability is realized
- Assesses the damage
- And eventually offers the payment to the insured
The key to anything decentralized is, as we’ve said before, ‘disintermediation.’ This refers to the replacement of inefficient middleman with efficient, effective, and transparent programs. On the same principle, decentralized insurance can replace some of the functions with code, which will allow for greater efficiency and participation.
Let’s look back at the example of crop insurance. This policy is a fairly significant iteration of decentralized insurance because each intermediary is important to the final calculation.
The liability realization, the actual event (which triggers the insurance payout), would be tracked by an oracle. Put simply an oracle is anything that provides data, any data. It could be weather data, price data, or even social data. For crop insurance, the relevant data points could be anything that could affect the growth of the crops. From the centimeters of rainfall, the variable purchase price-per-crop, yield-rate, to even such things as locust eggs hatchings per-capita. Who knew? These oracles could receive data from measurement devices, which feeds information to both parties (insured and insurer), free of bias.
The insurance payout would emanate from a pool of investors/insurers who each contribute to avoiding the financial cost of the liability of risk. Once the terms are agreed (based on periodic premiums, deductibles, nature of risk, degrees of coverage, measurement, oracle-feed, the trigger for payout, and other potential conditions) the entire policy is wrapped in a smart contract. Put simply, a smart contract is a computer program that is automatically triggered based on conditions agreed upon by the parties to the contract.
Decentralized insurance at play
Looking back at Sundar’s example, how would the payout play out in the decentralized insurance market? In the case of an isolated drought, the rain gauge (measuring rainfall) on Sundar’s farms would be alerted, feeding data to the oracles which would trigger a clause in the smart contract. Sundar would then get an insurance payout based on his periodic contributions (less his deductible) for the crops affected by the drought.
In this particular example, the oracle was tied to rain gauges, but because of the flexibility of the decentralized insurance ecosystem, these terms could be modified based on the type of insurance, and the preferences of the parties. Other farmers could generate the same policy for locust infestations, depending on a measurable metric that could accurately predict future infestations and thus determine a decrease in crop yield. (I’m not sure how one would count locust infestations, I wouldn’t recommend by hand though).
Each relevant detail to the insurance contract, whether it’s the triggered metric, the oracle used, the currency paid-out, the area(s) specified, would algorithmically determine the premium, and based on the final amount and its derivations, the insurance-seeker can choose to buy it or not.
Owing to the disintermediation and detailed nature of the breakdown, the cost decreases, and the transparency increases. Most importantly, it will allow the parties, on either side of the insurance policy, to use it for what it is intended. Insurance policies can be used to hedge against the likelihood of the financial cost of risk, and not to either make a profit or ‘game the system,’ like Li’s ingenious flight-insurance scam. In the decentralized insurance system, there is also a decrease in the time-lag between incurring the financial cost and insurance coverage payout. This would be calculated and paid out immediately based on data provided by the oracle.
Potentially every form of risk and its associated liability could have an insurance policy, based on specific oracles, and the agreed smart contracts. For theft insurance, the parameters could be the address of cryptocurrencies, linked accounts, transfers between exchanges, etc. Hurricane insurance can be based on wind speeds in a particular area, against damage of properties, based on the durability of the structure. Life insurance has the most obvious parameter of all, your pulse.
In terms of a larger picture, decentralized insurance can work better, or rather in a more streamlined manner. This would be based on the ease of data procured by the oracles and the nature of the metric. Some metrics are zero-sum, dead or not, for life insurance, flight delayed or not for flight insurance, collision, or not for car insurance. For other forms of insurance which are based on a degree of risk, damage, and payout, (eg: crop insurance, health insurance, home insurance) the kinks need to be worked out to operate seamlessly in the decentralized finance world. But in time, it will be refined.
Insurance, not a hedge
It may seem like insurance is another word for a hedge, and there are a number of products allowing hedging in the cryptocurrency space. However, decentralized insurance is not one of them. This would not only be oversimplifying the concept of insurance, especially decentralized insurance, but it would also be limiting.
Insurance would be the prevention of a risk owing to uncertainty. That’s not the same thing as hedging. Hedging is a purely investment-oriented term where two different positions are taken so that loss is offset by profit. We’ll be sure to cover hedging in the cryptocurrency space in more detail in a later post!
Ensure the future
While still in its early stage, decentralized insurance is one of the main building blocks to the larger decentralized financial ecosystem. Risk is an inextricable part of any financial economy, and when you have the volatility of cryptocurrencies coupled with the ‘maturing, but not yet there’ internal infrastructure and governance of most decentralized networks, decentralized insurance is all the more important.
Decentralized insurance is key to the larger decentralized finance space. Through it, a robust foundation for the new digital economy will be built, on the back of decentralized currency and blockchain technology. Nothing can guarantee that we’ll all live long and prosper but insurance can guarantee the latter at least.
You can find our videos about Decentralized Insurance here: