Is pandemic insurance a contradiction in terms?
The answer could lie in public-private risk-sharing mechanisms and innovation
A fast post-Covid economic rebound is highly coveted by businesses, consumers and the government alike – and cancellation insurance could be key. Although many of us are ready to splash out on holidays, sport events and festivals, we are unwilling to gamble and lose on such things in case – heaven forbid – another lockdown hits.
Cancellation insurance policies, nevertheless, are few and far between. You could, for example, insure against the risk of not being able to visit someone if you, your travel companion or a dependant is taken ill by Covid. But coverage against a national or regional lockdown is non-existent.
To be fair to insurers, providing a cover against fire risk while the house is already burning might not appear commercially sound or even sensible. Understandably, the government’s quasi-guarantee that restrictions will irreversibly be lifted by the summer also falls short of convincing insurers to take on such an immense risk.
Even more so because, a year after the breakout of the pandemic, insurers are still left in limbo about the full extent of their Covid-related payouts. The most contentious cause of this ambiguity is the scope of business interruption (BI) policies.
Business interruption (BI) revisited
BI cover typically comes with property insurance. It is meant to keep a business afloat while it’s temporarily closed as a consequence of physical damage, whether caused by fire, flood, theft, vandalism or something else.
Little wonder then that the first sticking point between insurers and authorities representing the insured – such as the FCA in the UK – has been the ability of the coronavirus to cause structural and tangible damage to equipment and properties. Legal cases in the US were built on the virus lingering on surfaces for 28 days, thus necessitating “remediation to clean surfaces.”
What has put defendants to a disadvantage has been the fact that, in the wake of the 2003 SARS epidemic, which in comparison was a minor and rather contained public health crisis, insurers went out of their way to add virus waivers to their BI policies.
About one-third of businesses take out BI covers, and only a fraction of them come with infectious and notifiable disease extensions. Although Covid had been added to the list of notifiable diseases at the beginning of the pandemic last March, unless the cover is for unspecified notifiable diseases, Covid will be an exclusion.
In the US there have been a slew of lawsuits in an attempt to force insurers to pay out for Covid losses, but about four times as many have ruled in favour of the insurers than against them. Currently, seven states also have pending legislation that could get insurers to retroactively pay for BI losses incurred by coronavirus lockdowns. If enacted, they will mandate the redefinition of property damage sustained by SMEs to include the presence of a person infected with Covid-19 in the policyholder’s premises, or in the same municipality or state. 
The Supreme Court in the UK, meanwhile, has also widened the scope of BI policies that have to pay out. It ruled, among other things, that the pandemic, as well as the government and public response to it – that is, the lockdown, whether partial or complete, mandated or only instructed without the force of law – were all perils that will trigger compensation.
According to the FCA’s estimate this ruling will throw a lifeline to approximately 370,000 business, while critics argue that the judgement applies only to a particular wording in certain clauses of some contracts and will only be a drop in the ocean of six million British businesses.
Getting insurers to indemnify losses that they – in many cases – didn’t underwrite at all is a tricky sell. On the other hand, having spent hundreds of billions on protecting jobs and businesses, as well as fighting Covid, the UK government’s “nudge” to the insurance sector to pull its weight sounds like a legitimate endeavour.
But how can insurers – once bitten, twice shy – be incentivised to develop pandemic products in a post-Covid era, when these insurance events are almost impossible to predict and, for the scarcity of data, difficult to model? (The complexities of assessing this type of risk are well demonstrated by the fact that the Economic Forum’s Top 10 list for pandemic preparedness was led by the US and the UK in 2019.) A risk is deemed insurable only if it’s statistically assessable and independent from other risks in terms of time, geography and the type of risk – which a pandemic, by definition and experience, is not.
The new index-based insurance model
Pandemic insurance is not a new concept, though. San Francisco-based disease outbreak prediction start-up Metabiota, having teamed up with Munich Re and brokerage Marsh, put one on the table as early as 2018.
This offering wasn’t the kind of traditional insurance product that, for a set premium, gives the insured protection against losses they incur. Rather, it adopted an alternative, so-called parametric or index-based model which covers the probability of a predefined event occurring: if the water level reaches seven metres, for example, you’ll get paid.
The event triggering payout in a pandemic could include metrics such as mortality rates or border closures in an area. As cover is highly personalised and accommodates the risk appetite of the insured, cover events may vary from policy to policy. And as claim adjustment is eliminated from the process, payout is fast, which is great for the solvency of businesses caught up in a health or natural crisis.
Insurance experts with an orthodox approach argue that the parametric model should be regarded as a financial derivative rather than a type of insurance. Others, despite parametric pandemic insurance’s pre-Covid failure to sell, have high hopes of this new model filling in the grey area of BI insurance that the current pandemic so forcefully exposed.
Innovation network InsTech London’s Parametric Overview 2021 set the coverage that the insurance industry should be able to provide for a future pandemic at $200bn (£143bn), equivalent to the total of all losses insurers pay in a bad year of natural disasters.
But in order for BI insurance to turn the page after Covid and become revitalised, the state as the re-insurer of last resort will most probably have a role to play too. Not unlike after 9/11 – another historic loss event, when the US government passed the Terrorism Insurance Act (TRIA) to ensure the viability of the terrorism insurance market with a federal backstop – insurers will need some state guarantees as incentives to kickstart the pandemic insurance market.
TRIA originally was a three-year federal programme but has been extended several times – most recently in 2019 (until 2027). With a similar safety net provided by governments, insurers will be able to come up with new data analysis and modelling methods, as well as innovative BI and cancellation products that can help the global economy roar back to life. There needs to be a way forward. After all, the Great Fire of London, which devastated central London in 1666 with an estimated loss of £1.5bn in today’s money, resulted in, among other things, the world’s first property insurance policy.
Zita Goldman
Deliver on your digital promises: the Future of Insurance
The future of insurance is already here. Drones can assess accidents without putting agents at risk by capturing footage and inspect damage. Sensors, plug-ins and cameras enable usage-based insurance models, such as offering discounts for safe driving. AI-powered algorithms can deliver claims in seconds.
This tech completely redefines how we think of insurance. It’s exciting stuff, to be sure. But the insurance companies who will thrive in five, 10 or even 30 years from now won’t be distracted by shiny tech – they’ll be focused on delivering a superior human touch, enabled by technology.
Future-proofing your insurance organisation requires you to master the basics and make your customer experience the centre of all you do. In the end, it all comes back to data.
Insurance represents a $1.3 trillion market filled with colossal swirls of processes, people and content. Ninety-five per cent of American homeowners have some form of home insurance. Eighty-seven per cent of car owners have some form of auto insurance. At Adlib alone, we facilitate 90,000 claims per day and 24,000 new policies a month for our enterprise insurance customers. And it’s only going to grow.
You can’t delight without managing your data
Leveraging technology to harness and optimise the collection of people, processes and content is what winners in the insurance market have in common. Unlocking your unstructured data – complex documents such as claims, contracts, proposals, agreements, scanned or uploaded images, and more – allows technology to deliver the essential fuel that powers market-leading customer experiences.
Humanising insurance
The value of all this shiny tech isn’t its “cool factor”. It’s the way it works behind the scenes to humanise the insurance model. When’s the last time you interacted with your insurance company on a good day? Your customers deal with you when life happens: accidents, theft, natural disasters, hospitalisations, death of loved ones.
The best insurance companies will deliver fast, empathetic, light-touch support for customers dealing with emotional life experiences. AI and automation allow insurance companies to let humans do what they do best: be human.
Many of today’s insurance processes are often email-based and highly manual content flows. Eliminate those friction points and increase transparency with the data you already have. It’s not the sexiest or most exciting part of what technology can do, but the best products I know are the ones that just work. You don’t have to think about them. Technology allows you to deliver the right information at the right time to the people and systems that need it, without manually navigating multiple systems. Prioritising digital transformation in insurance allows you to deliver the kind of customer experience that will differentiate your business.
The ability for an organisation to mobilise and connect legacy content and the future nirvana of bots, drones and IoT will be key to delivering the best customer experience now and into the future.
Serve both sides of the customer experience equation
The most common approach to the customer experience focuses on your external customers. But what about your agents and call centre staff that handle everyday challenges? Your internal “customers” are just as much a part of delivering the end result as the technology is. It’s not a wholesale replacement of people – it’s about equipping them with the tools and time they need to do what they do best. And cobbling together legacy systems using duct tape and glue isn’t going to cut it.
Capturing the right data automatically can simplify compliance, improve the customer experience and eliminate tedious data-entry and error-prone manual efforts.
Keeping up with the pace of innovation
Insurance companies need to start thinking of themselves as technology companies that deliver insurance products and services. With greater control over your data, your insurance organisation can: 1. Build products and services that are dynamic 2. Move more quickly to serve your customers 3. Realise the value of new and exciting technologies Once a strong data foundation is in place, you can close the gap between customer expectations and reality, fulfilling the digital promises you make to your customers.
Centres of excellence – quickly becoming beacons of innovation – can help insurance companies quicken their pace of innovation and get one step closer to fulfilling their digital promise.
Adlib can help deliver your digital promise
Adlib bridges these two worlds. At Adlib, we create intelligent data that amplifies human potential and maximises business performance for hundreds of global insurance customers. How do we get there? Our Content Intelligence, Contract Analytics and Content Transformation solutions make it easy to discover, standardise, classify, extract and leverage clean structured data from complex unstructured documents. In doing so, our global customers unlock the power of their data to build seamless customer experiences and achieve a whole new level of performance.
Scott Mackey
|
1. Deloitte: Insurance industry drone use is flying higher and farther 2. McKinsey & Company: Digital ecosystems for insurers: Opportunities through the Internet of Things 3. McKinsey & Company: Insurance 2030 – The impact of AI on the future of insurance 4. Statistica: Value of insurance premiums written in the United States from 2009 to 2019 5. Policygenius: Homeowners insurance statistics |
Spotting liars is hard – but our new method is effective and ethical
Most people lie occasionally. The lies are often trivial and essentially inconsequential – such as pretending to like a tasteless gift. But in other contexts, deception is more serious and can have harmful effects on criminal justice. From a societal perspective, such lying is better detected than ignored and tolerated.
Unfortunately, it is difficult to detect lies accurately. Lie detectors, such as polygraphs, which work by measuring the level of anxiety in a subject while they answer questions, are considered “theoretically weak” and of dubious reliability. This is because, as any traveller who has been questioned by customs officials knows, it’s possible to be anxious without being guilty.
We have developed a new approach to spot liars based on interviewing technique and psychological manipulation, with results just published in the Journal of Applied Research in Memory and Cognition.
Our technique is part of a new generation of cognitive-based lie-detection methods that are being increasingly researched and developed. These approaches postulate that the mental and strategic processes adopted by truth-tellers during interviews differ significantly from those of liars. By using specific techniques, these differences can be amplified and detected.
One such approach is the Asymmetric Information Management (AIM) technique. At its core, it is designed to provide suspects with a clear means to demonstrate their innocence or guilt to investigators by providing detailed information. Small details are the lifeblood of forensic investigations and can provide investigators with facts to check and witnesses to question. Importantly, longer, more detailed statements typically contain more clues to a deception than short statements.
Essentially, the AIM method involves informing suspects of these facts. Specifically, interviewers make it clear to interviewees that if they provide longer, more detailed statements about the event of interest, then the investigator will be better able to detect if they are telling the truth or lying. For truth-tellers, this is good news. For liars, this is less good news.
Indeed, research shows that when suspects are provided with these instructions, they behave differently depending on whether they are telling the truth or not. Truth-tellers typically seek to demonstrate their innocence and commonly provide more detailed information in response to such instructions.
In contrast, liars wish to conceal their guilt. This means they are more likely to strategically withhold information in response to the AIM instructions. Their (totally correct) assumption here is that providing more information will make it easier for the investigator to detect their lie, so instead, they provide less information.
This asymmetry in responses from liars and truth-tellers - from which the AIM technique derives its name - suggests two conclusions. When using the AIM instructions, if the investigator is presented with a potential suspect who is providing lots of detailed information, they are likely to be telling the truth. In contrast, if the potential suspect is lying then the investigator would typically be presented with shorter statements.
The experiment
But how effective is this approach? Preliminary research on the AIM technique has been promising. For our study, we recruited 104 people who were sent on one of two covert missions to different locations in a university to retrieve and/or deposit intelligence material.
All interviewees were then told there had been a data breach in their absence. They were, therefore, a suspect and faced an interview with an independent analyst. Half were told to tell the truth about their mission to convince the interviewer of their innocence. The other half were told that they could not disclose any information about their mission, and that they should come up with a cover story about where they had been at the time and place of the breach to convince the analyst of their innocence.
They were then interviewed, and the AIM technique was used in half of the cases. We found that when the AIM technique was used, it was easier for the interviewer to spot liars. In fact, lie-detection accuracy rates increased from 48% (no AIM) to 81% – with truth-tellers providing more information.
Research is also exploring methods for enhancing the AIM technique using cues which may support truth-tellers to provide even more information. Recalling information can be difficult, and truth-tellers often struggle with their recall.
Memory tools known as “mnemonics” may be able to enhance this process. For example, if a witness of a robbery has provided an initial statement and cannot recall additional information, investigators could use a “change perspective” mnemonic – asking the witness to think about the events from the perspective of someone else (“what would a police officer have seen if they were there”). This can elicit new - previously unreported - information from memory.
If this is the case, our new technique could become even more accurate at being able to detect verbal differences between truth-tellers and liars.
Either way, our method is an ethical, non-accusatory and information-gathering approach to interviewing. The AIM instructions are simple to understand, easy to implement and appear promising. While initially tested for use in police suspect interviews, such instructions could be implemented in a variety of settings, such as insurance-claim settings.
Cody Porter, Senior Teaching Fellow in Psychology and Offending Behaviour, University of Portsmouth
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Evolving through the era of insurance technology
Advances in technology are driving major changes in insurance, but ultimately what consumers want is to pay a fair price to having their losses dealt with quickly and easily.
Thirty five years ago Direct Line was launched as the first ever telephone insurance company, selling directly to consumers. Eighteen years later, in 2002, the first insurance comparison website, Confused.com, enabled us to rapidly find the best price.
But not much happened in insurance for the next 10 years. Innovation had paused. Then suddenly, in 2013, the insurance world discovered its first unicorn, the billion-dollar start-up Zhong An, China’s first fully digital property insurance company. It underwrote over 630 million insurance policies in its first year. Two years later it was valued at $10billion. Insurtech had arrived.
After decades in the shadows, insurance now seemed ripe for disruption. Around the world entrepreneurs young and old, often spurred on by their own poor customer experience buying or claiming on insurance, identified opportunities to take on the incumbent insurers. In the last few years investors have committed more than $5billion to insurtech, creating new insurance companies, or funding the start-ups building the technology, data and analytics to power them.
The need may be there, but it’s proving hard for outsiders to break in, particularly to the established, highly regulated mainstream insurance businesses such as property or motor insurance. Despite the challenges, the shape of insurance is starting to change.
Forty years ago 80 per cent of the value of the largest global companies was defined by their physical assets: property, plant, inventory. Today every major company has massive exposure to technology – both their own and through increasingly widespread connectivity to trading partners. Combine this with the power of social and mainstream media, the value of intellectual property and the escalation and vulnerability of brand value, and the result is that today 80 per cent of these companies’ value is represented by intangible assets. Cyber is the best-known risk, and the fastest growing area of new insurance products. Intangible assets are much harder to insure. New ways to measure risk mean a need for new tools and an opportunity to find new approaches for offering insurance.
What does this mean for the future of insurance? Insurance traditionally has been split across multiple types of risk, broadly categorised by the industry as property, casualty, life and health. For centuries insurers have promised to pay for an infrequent loss in return for smaller regular payments. Each type of cover has been assessed independently, with a broker controlling the relationship between the policy holder and the insurer.
Much may be changing but two characteristics of insurance are likely to stay the same for a long time: we still want to be compensated for our losses and no one enjoys buying insurance. Many of the best-known technology companies have been successful because they have developed products that make our lives easier. The success of companies such as Direct Line, Confused.com and Zhong An suggest that insurance is no different. Expect to see one or more of the following shifts in how we buy and use insurance in the next decade:
Companies self-insure
As companies grow in size they are increasingly able to use their capital base to absorb all but the largest catastrophic losses themselves. With exponentially more data becoming available each year from cheap, widely deployed sensors, companies will be able to perform sophisticated analytics on their own risks. With greater knowledge, and more capital than their insurers, it will make more sense for large organisations to self-insure rather than give away premiums to third parties.
Insurance shrinks then vanishes for most of us
If we are confident we will get compensated for a loss, we don’t care too much who pays us. Already today we can buy a car with insurance built in (Volvo), a train ticket that includes cancellation cover (The Trainline) and a laptop with damage protection (John Lewis). At some point, probably led by Amazon, insurance protection will be bundled in with everything we buy and disappear altogether. What purpose will individual insurers play then?
Or will it be “one insurer, everything covered”?
One of the biggest challenges for many people when buying insurance is understanding what is, and what is not covered. Too often people receive no, or only a partial, payment because their cover was not as comprehensive as they expected. Look out for the first insurer that can offer a “one-stop” solution with no exclusions, giving us the certainty that we are covered for any loss irrespective of whether we damage our car, lose a camera, need healthcare or flood our house.
Insurance will benefit from better data, analytics and technology. But consumers don’t care too much about blockchain, artificial intelligence or machine learning. What we want is to get on with our lives in the confidence that when bad stuff happens, we’ll be taken care of – quickly, fairly and easily. That will be the future of insurance.
Matthew Grant
Matthew is an advisor to some of the fastest-growing data and analytics companies servicing the global insurance market. InsTech London is the largest and most engaged community of people driving innovation in both existing insurers and technology companies.
Cyber-security risks and how to manage them
The Sunburst attack uncovered in December 2020 illustrates the magnitude of the cybersecurity challenge. Hackers were able to breach some of the United States’ top government agencies as well as those of other organisations around the world by compromising updates from one of their software suppliers, SolarWinds. Organisations might make use of hundreds if not thousands of third-party suppliers and contractors through which they could be breached.
While many of these involve third-party suppliers and contractors in the IT realm, in some cases breaches can be caused by third parties that one might not expect. For example, it was reported in 2017 that a casino in Las Vegas was breached through an Internet-connected fish tank. Similarly, in 2014 Target was breached through its air-conditioning supplier. Organisations frequently take on new third-party suppliers and contractors, further compounding the challenge.
The rapidly evolving nature of the risk makes it difficult to assess, and all organisations are currently struggling with how to manage cybersecurity risk. New threat actors and types of attacks regularly emerge. For example, we are seeing the advent of AI-enabled attacks. In one highly publicised instance, cybercriminals tricked an employee into transferring money to them by using AI to imitate the CEO’s voice.
Current approaches to managing cybersecurity risk have significant shortcomings. Risk assessments tend to rely on “risk matrices”, which use a grid to compare the likelihood of the risk and the severity of the impact. The numerical values assigned to the likelihood and severity ratings tend to be ambiguous, meaning that they can assign the same numerical values to threats that are quantitatively quite different . This can cause organisations to incorrectly prioritise threats and thus allocate resources in a suboptimal manner.
A role for cyber-insurance
Increased use of cyber insurance could significantly improve cybersecurity risk management. It helps organisations by transferring the risk to insurance providers. Quite importantly, it could also incentivize organisations to improve their cybersecurity levels, through insurers offering customers a discount in exchange for improving security measures.
Yet cyber insurance is still underdeveloped for a number of reasons. Insurers face difficulty in accurately assessing an organisation’s cybersecurity risk. Unlike in other domains that have similarly elevated levels of risk – such as pandemics – there is limited historical data to draw on. A contributing factor to this lack of data is that organisations are reluctant to disclose that they have been attacked due to reputational concerns. Moreover, an organisation’s risk profile at the time an insurance policy is issued may differ considerably several months later.
These issues are compounded by an acute shortage of experienced cybersecurity underwriters, whose job it is to decide whether to issue a policy to a prospective client.
Another challenge is “accumulation risk”, in which a single incident can spread to other parts of an insurer’s portfolio. It is particularly difficult to evaluate accumulation risk in the cyber realm. In the physical world, a hurricane or other natural disaster may trigger a surge in claims, but these claims are limited to a particular geographic area. In cyberspace, a cyberattack can result in claims around the world. For example, the WannaCry ransomware attack infected some 200,000 computer systems in 150 countries, severely disrupting major organisations such as FedEx and the UK’s National Health Service.
A related issue is “systemic risk”, in which one incident could cause a cascading failure that triggers the collapse of an entire system. For example, a cyberattack that takes down the power grid will impact sectors ranging from transport to communications to healthcare.
Improving our understanding of cyber-risk
To address these issues, in our new book we propose a series of models aimed at helping both organisations and insurers manage cybersecurity risk. They make use of a methodology known as adversarial risk analysis, which makes it possible to better assess the risk that different threat actors pose to an organisation.
These models allow insurers to automatically adjust premiums in response to changes in an insured organisation’s cybersecurity risk. They draw on data provided by third party companies that gather real time information about organisations’ IT infrastructure, security products, and other factors to get a clearer picture of an organisation’s cybersecurity risk at any given point in time. These third-party companies include firms such as SecurityScorecard, Blueliv and BitSight.
One of the models makes it possible to better understand accumulation risk. It does so by breaking out different market segments as separate components in order to isolate, understand, and analyse the accumulation effect of a cyberattack on a given market segment.
The book also describes how some insurers are moving beyond merely selling insurance to assisting customers in improving their cybersecurity readiness. For example, they might share information on security vulnerabilities, assess customers’ IT infrastructure, or help them implement penetration testing of their IT systems and phishing-awareness campaigns aimed at their employees. In addition, they may support customers in responding to cyberattacks, providing crisis management and legal assistance, and helping them get back to business. This is typically accomplished through partnerships with cybersecurity companies, public relations firms and legal firms.
These developments can play an important role in cybersecurity risk management, helping make it possible to create a virtuous cycle where cyber insurance fosters an increase in cybersecurity worldwide.
This article is based in part on our latest book, Security Risk Models for Cyber Insurance, published by Routledge/Taylor & Francis. It grew out of a two-year project funded by the European Union under Horizons 2020, CYBECO (Supporting Cyber Insurance from a Behavioral Choice Perspective).
Created in 2007 to help accelerate and share scientific knowledge on key societal issues, the AXA Research Fund has been supporting nearly 650 projects around the world conducted by researchers from 55 countries. To learn more, visit the site of the Axa Research Fund or follow on Twitter @AXAResearchFund
Caroline Baylon, Security Research and Innovation Lead, AXA and Research Affiliate, Centre for the Study of Existential Risk, University of Cambridge and David Rios Insua, Member of the ICMAT, AXA-ICMAT Chair in Adversarial Risk Analysis and Member of the Spanish Royal Academy of Sciences, AXA Fonds pour la Recherche
This article is republished from The Conversation under a Creative Commons license. Read the original article.