Can Blockchain Improve Insurance?

Andrew W. Singer

|

February 1, 2019

blockchain insurance industry

Fraudulent claims continue to plague the global insurance market, but proponents of blockchain technology insist that a solution may soon be at hand. When implementation of these decentralized digital ledgers becomes widespread, some believe fraud could be reduced significantly.

An estimated 5% to 10% of all insurance claims are fraudulent, costing U.S. non-health insurers more than $40 billion a year, according to a 2017 McKinsey & Company report, which suggested that “by serving as a cross-industry, distributed registry of external and customer data, blockchain can be used to identify fraud.” Indeed, a 2018 Boston Consulting Group (BCG) report predicted that an “all-blockchain” auto insurer could lower its total operating ratio by 10 to 13 points compared with a traditional carrier, two points of which would be from fraud detection alone.

Others contend that new insurance markets can be accessed using blockchain technology, especially in regions that exhibit high rates of corruption, because blockchains provide more reliable alternatives to current registries. In other words, blockchains could not only increase profitability by reducing current fraud, the technology could grow revenues by tapping under-served, fraud-ridden markets.

In practice, however, reducing insurance fraud through blockchain technology may not be quite so easy. “There’s this notion that technology can solve social problems, but it is naïve to think that any technology can completely eliminate fraud,” said Stephan Karpischek, CEO of Etherisc, an insurtech firm that recently built a blockchain platform for flight delay insurance. While fraud is simply a function of human nature, blockchain technology has elements that can support its detection and mitigation, which is no small thing, he said.

How Blockchain Works


Blockchain is often misunderstood. Most frequently, it is confused with the cryptocurrency bitcoin, which is only one application for blockchain—that is, the ability to store and transfer value. Blockchain is much broader, and with the addition of smart contract technology, there are far more possibilities than just cryptocurrencies.

A blockchain is nothing more than a distributed, digital ledger. It has been compared to a dusty, leather-bound ledger in a Dickensian counting house that holds records of every key transaction, but with this distinction: The same ledger is held simultaneously on thousands of different computers (or “nodes”) in multiple locations with different owners, and the moment a line is added to one ledger, it appears on all ledgers.

The blockchain can also be programmed with “smart contracts,” a set of conditions recorded on the blockchain, so that transactions automatically trigger when certain conditions are met. Smart contracts can be used to automate payouts for insurance claims. For example, in the case of flight delay insurance, when a flight is delayed by a certain amount of time—say, more the two hours—it will trigger a payment under the policy.

To visualize how this might deter insurance fraud, put yourself in the shoes of a fraudster, said Karpischek. You have an encrypted identity on the Ethereum blockchain (the most widely used platform) where you are known by a 20-byte encrypted address such as “0xd2a3d9f938e13cd947ec05abc7fe734df8dd826. This maintains your anonymity under normal circumstances. But in a case of suspected fraud, the government or an insurance company could uncover your physical identity.

“If you defraud someone, the government has all the evidence,” Karpischek said, because all of your transactions are recorded on the blockchain, and they cannot be erased. (They are “immutable” in the blockchain argot.) “The government has a perfect trail.”

This could discourage would-be fraudsters. “It’s like robbing a bank and showing your passport to the [bank’s] security camera,” Karpischek said. In practice, it is not so much that the blockchain technology hinders fraud, but that the technology makes it easier to trace, potentially making prospective bad guys think twice before stealing an identity or faking a car accident.

The bitcoin experience is instructive in this regard. Bitcoin was the first widely distributed blockchain application. At the beginning, in 2009, bitcoin was used by some bad actors who traded in illicit items like drugs, weapons and stolen credit cards. Now, some of those unsavory bitcoin transactions committed years ago are being linked to criminal activity. Service firms like Elliptic have sprung up to identify illicit activities conducted through cryptocurrencies and provide actionable intelligence to financial institutions and governments.

If you are “tainted,” you are marked forever, said Karpischek, and you can be refused licenses, loans, coverages or accreditations, making a blockchain’s immutability a potentially huge deterrent to fraud.

Why Not a Centralized Database?


Still, are blockchains really necessary to deter insurance fraud? Would it be more effective to put all insurance data on a single centralized database, like a national fraud registry? After all, it would run faster than a public blockchain, which needs to reproduce every transaction on tens of thousands of separate nodes.

A centralized database or registry could have positive effects on fraud reduction, said Johannes-Tobias Lorenz, senior partner at McKinsey, but there are certain advantages that are unique to the blockchain. First, a blockchain’s decentralized validation allows it to operate in situations of potential distrust, such as when it is unclear who should be the trusted party actually running the centralized database. Second, the redundancies of the distributed ledger make manipulation of data much more difficult compared with a centralized database.

The promise of blockchain technology, especially with its smart contract functionality, is that insurers no longer need to simply trust one another. “When correctly deployed, [blockchain] will enable sharing of data between insurers and other agencies,” said Leanne Kemp, CEO of Everledger. “This will add significantly to each insurer’s ability to combat fraud across all lines of personal and commercial insurance.”

There are four specific areas where blockchains can make a difference in countering fraud:

  1. Identity. Blockchain technology can transform the way insurance companies manage identities and personal information, said Michael Mainelli, executive chairman of Z/Yen Group. If a fraudster claims insurance benefits for a dead person, for example, the insurance company may not even know that the individual has died. If the insurer had immediate access to death certificates on a blockchain, however, the fraudster would have to ply his trade elsewhere.

  2. Provenance of property and assets. Fraudulent jewelry claims alone cost the insurance industry $2 billion a year. Appraisals can be forged, appraisers and owners can collude on evaluations, and low-value jewelry can be substituted for high-value jewelry. Everledger recently built a provenance proof blockchain for the jewelry industry that tracks individual gemstones like diamonds from the mine to the end consumer, disclosing to all participants the (encrypted) identity and location of each entity involved in the stone’s supply chain. Authenticating provenance is key, Kemp said, because “having transparency builds trust among stakeholders and mitigates fraud.
    Everledger is also expanding into other areas, collaborating with research company True Image Solution on evaluation and restoration of cultural heritage objects on a tool that combines blockchain technology with forensic artwork reporting. “By digitally issuing and managing forensic reports encrypted on the blockchain, there is a permanent, tracked record of an artwork’s authentication and certification,” Kemp said. “This information can be made available—through permissioned access—to the insurance companies that require it.”

  3. Decentralized validation. Blockchains use validation algorithms to confirm transactions and events. This makes collaboration among insurance partners possible, even if uncertainty exists about the severity of an event. “In the case of smart contracts, for example, a claim payout would only be triggered if all nodes in the blockchain confirm a specific event, like a flight delay or a natural catastrophe,” Lorenz said. “Since the nodes most likely use diverse data sources, manipulation of events becomes much more difficult.”

  4. Trade and logistics. Fraudulent certificates of insurance remain a big industry issue, said Mainelli. If insurance certificates were entered on a blockchain, counterfeiting scams would become less tenable.


Most agree that blockchain needs to be combined with other technologies to reduce fraud. At a minimum, this means a layer for the smart contract, but it can also involve other things like telemetric data from a car, or the car owner’s speeding tickets and car repair history, according to BCG.

The Limitations of Blockchain


Blockchain is not a panacea. It will not eliminate fraud altogether. It has limitations, particularly the very big assumption that the data entered on the chain is sound.

Any software system with bad inputs will generate bad outputs. If a consignment of horse meat is falsely labeled as 100% beef, then you can track it meticulously on a blockchain all the way to the cash register at the grocery store and you still will not discover the fraud. This is the “garbage in, garbage out” conundrum, and it is particularly acute for blockchains because entries are supposed to be unalterable.

The insurance industry will not be able to dispense with fraud investigators anytime soon. One day, perhaps, a P/C carrier will be able to confirm the occurrence and severity of an event like a car crash through the multiple sensors within cars, as well as closed-circuit television and “nodes” like other cars and police reports, said Lorenz, but that is unlikely within the next five years.

“There will always be a need for investigators,” Kemp said, but blockchain technology can still make a difference by identifying which transactions warrant the attention of fraud departments, while allowing legitimate claimants straight-through processing.

Scaling remains a challenge for smart-contract enabled blockchains. They cannot handle many transactions at present—too much data and transactions slow to a crawl. In a 2018 report, McKinsey predicted that blockchain is still three to five years away from feasibility at scale, “primarily because of the difficulty of resolving the ‘coopetition’ paradox to establish common standards.” That is, natural competitors must agree on governance decisions around how the system, data and investment will be led and managed.

Some specific questions also need to be resolved with regard to governance, said Indranil Nath, vice president at DXC Technology, namely: Will systems be completely open, or will they use permissioned-based access? What are the principles for suitability in interacting with the ledger? How will information be shared among blockchain systems (also known as interoperability)?

Removing the Human Element


If properly applied, Mainelli believes blockchain could eventually reduce insurance fraud by as much as 20% to 40%. This assumes, of course, that the data is sound. Other experts suggest that the usefulness of blockchain is contingent on it being combined with other technologies like smart contracts, and the latest forensic approaches to stemming fraud.

Much insurance blockchain activity is currently concentrated on parametric insurance applications and catastrophe bonds, such as hurricane coverage in which a specific weather statistic like wind speeds in a certain locale can trigger a claims payment. This has certain advantages: There are no claims disputes and the data cannot be questioned as you cannot really argue with objective weather station data, Karpischek said.

While this is admittedly a small part of the overall insurance industry, perhaps it points to the real promise of blockchain—removing the human element from the claims payment process. The insurance company cannot deny you a payout if your airline flight is three hours late because the smart contract receives flight delay data directly from the air controller and a payment is triggered automatically when a delay exceeds two hours. Conversely, policyholders do not have to look for documentation to file a claim and they will be not be tempted to file a false claim. Everything will be automated.
Andrew W. Singer is a freelance writer based in New York City.