Insurers getting tougher on cyber risk during pandemic

The coronavirus outbreak has heightened activities from cyber criminals, leading to insurers scrutinising cyber coverage risks, said Marsh cyber practice leader Asia Pacific Kelly Butler, during a webcast to clients.

“Unlike other classes, there has not been a move to COVID-19 related exclusions, but insurers are looking more closely at each firm’s situation and how well-placed they are in the new environment,” she said. “Cyber policies are very broad. What I anticipate is the coverage itself won’t change but they will very much look at an individual risk and decide the coverage they will put on any of those individual risks.”

She added that insurers would be looking at policyholders’ overall business resilience in much more details come renewal. They will also be finding out how the world’s increased spike of dependence on technology will affect the ability to respond if there is a breach or disruption in services.

She said that while Marsh is very closely monitoring cyber cover being provided from a global perspective, there have not been any discussions with insurers regarding exclusions.

Marsh head of strategic risk (NSW/ Queensland) Ben Crowther says firms now may have most of their employees working remotely under arrangements organised rapidly, and many people are using their own home equipment and infrastructure. Increased interest in clicking on latest information on the outbreak is also increasing the risks of phishing attacks.

“There is an incredible amount of cyber activity going on at the moment, all because of this crisis, and hackers and cyber actors making the most of this,” he said. “The level of sophistication around cyber security threats is high, and often there is the risk information can be stealthily collected for months before a breach is recognised.”

The widespread use of remote offices brought about by global quarantine has generated discussion over whether organisations will completely revert to previous ways of operating as the pandemic subsides.

“I think the new normal is likely to involve a lot more of what we are doing right now,” Mr Crowther said. “Some of those risks associated from working from home are still going to be as valid.”

Heightened COVID-19 cyber risks

CFC Underwriting has also issued a warning on new cyber scams exploiting the COVID-19 pandemic to target businesses.

“Since countries around the world went into lockdown, the types of incidents that our cyber claims team are dealing with shows that while there hasn’t yet been a change in frequency of attacks, the likelihood of companies falling victim to these scams in a vulnerable and remote working scenario are escalated in comparison to what we were experiencing pre-COVID-19,” said CFC cyber development leader Lindsey Nelson.

“This new era of home working couldn’t be a better situation for cyber criminals. Employees are working on potentially insecure devices and businesses may not have implemented any additional training to help them spot things like phishing links that play on, for example, human curiosity about coronavirus.”

The company’s cyber response team has observed fake websites offering medical equipment and information on COVID-19, and fake government agency emails and social media posts.

These fake sites, emails and posts contain links that trick users into giving up personal information, install ransomware and lead to fraudulent payments.

“With initial efforts being focused on the employees of the company working remotely, but not necessarily securely, it’s very possible that hackers have already penetrated mailboxes through business email compromise scams and are simply lingering, waiting for the right opportunity to strike. This means we won’t see the true implications of these attacks until a few weeks or even months down the line,” said Ms Nelson.

Japan’s industry adjusts to low interest rates

The cancellation of the EAIC 2020 is truly unfortunate. The role of insurance is now in the spotlight around the world and its role in society is growing. We hope that we can overcome these difficulties early and the next time we meet, we would like to have a great discussion about the development of insurance business.

We would like to report on the overview of Japanese life/general insurance in last few years, especially FY 2018.

Life insurance

For life insurance, the situation is steady despite very low interest rates. This is the result of adapting to changes in the market. For example, we have continued to create new products over the past few years, such as health-promoting insurance and dementia insurance.

Responding to low interest rates is a common challenge for many Asian countries, therefore we believe that Japan’s experience will be useful not only in product development, but also in asset management, risk management and other areas.

In individual insurance in FY 2018, the number of new policies and the amount of benefits significantly increased due to strong sales in whole life, term life and cancer insurance. Considering this situation, premium income in FY 2018 increased compared to the previous year. The total premium income in FY 2018 was JPY33.91tn and it was the first increase in the last three years.

Ordinary profits were JPY3.01tn, which increased for two consecutive years. The basic profits which represent the profitability of life insurance companies, were JPY3.77tn which increased for three consecutive years.

Total insurance benefits paid in FY 2018 was JPY28.78tn.

Looking at insurance claims, annuities and health benefits, insurance claims decreased for 11 consecutive years since 2008 due to the decrease in number of existing postal life insurance policies. Annuities had been increasing continuously until FY 2014, however, it has been decreasing since FY 2015. Maturity benefits increased for the first time in three years.

Among all the insurance claims paid, the number of policies and the amount of death benefits were 1.14m and JPY3.08tn. As for the maturity benefits, 1.24m and JPY2.81tn respectively. Death/ maturity benefits increased for three consecutive years.

Among all the health benefits paid, the number and the amount for hospitalisation benefits were 7.24m, and JPY715.9bn, surgery benefits were 4.41m and JPY444.9bn in which both benefits rose in numbers and amounts. Hospitalisation and surgery benefits increased for 14 consecutive years.

General insurance

Recently Japan has faced numerous large-scale natural disasters such as earthquakes, typhoons, and torrential rains and the general insurance claim payments regarding natural disasters has been increasing.

We recognise the increasingly important role the general insurance industry plays in supporting citizens to prepare for natural disasters that is increasing in severity in the recent years.

Net claims paid on all classes of insurance during FY 2018 amounted to JPY5,324.2bn, an increase of 13.2% compared with the previous fiscal year. This was due to an increase in earthquake insurance claim payments for earthquake with its epicentre in the northern part of Northern Osaka and 2018 Hokkaido Eastern Iburi earthquake, as well as increased fire insurance claim payments caused by domestic natural disasters such as Heavy Rain Event of July 2018, typhoon No. 21 (Jebi), and typhoon No. 24 (Trami), etc.

Net premiums written for FY 2018 came to JPY8,392.8bn, (up 0.1%, compared with the previous fiscal year) due to an increase in premium for fire insurance mainly consisting of earthquake insurance and new types of insurance.

Due to increased gross profit on asset management, ordinary profits for FY 2018 increased to JPY864.3bn (up 6.4%, compared to the previous fiscal year). Net profits for the current year decreased to JPY676.5bn (down 0.3%, compared with the previous fiscal year) due to decrease in special profits, etc.

All these figures are based on materials published by The Life Insurance Association of Japan and The General Insurance Association of Japan.

(Re)insurance in the time of COVID-19

The coronavirus pandemic has upended human life and economic activity as we know it. At the time of writing, 9.4m COVID-19 cases have been reported globally and almost half a million people have lost their lives to the virus. The accompanying economic losses are already immense, and while the pandemic is ongoing, there will be much uncertainty over the eventual numbers of cases and deaths.

Industries and supply chains have been adversely impacted by the lockdowns in many jurisdictions, and the ramifications can be felt across the globe. The Asian Development Bank (ADB) estimates the global economy could suffer losses of up to $8.8tn – or almost 10% of global gross domestic product – as a result of the pandemic. Additionally, as many as 242m people worldwide could lose their jobs with their overall income being reduced by $1.2tn.

Collaboration is imperative to mitigate the scale of the financial challenge that confronts the world, clearly suggesting that various stakeholders from the private insurance industry, governments, international think tanks and others need to come together and pool their capabilities and resources to alleviate the economic pain societies endure when risks such as pandemics materialise. We need to bridge the gap between insured loss and economic loss.

Pandemic funding and risk pools

Risks become amplified without proper levels of insurance coverage, often leading to the perpetuation of economic downturns. Insurance companies do not have the resources to provide pandemic coverage, leaving businesses vulnerable to what are termed ‘acts of God.’ The crises that result require a multi-stakeholder approach. A suitable solution could be the creation of a ‘pandemic pool’ similar in function to existing terrorism pools.

The foundation of the initiative could be spearheaded by the respective national government and designed in cooperation with the local insurance industry and international capacity providers. The pool could be administered by the insurance industry but given the size of the risk in relation to available market capacity, a government back-stop is key along with additional capacity support from the international reinsurance and capital markets. Government participation is vital as it helps in distributing the risk component and in creating a safety net for insureds to provide pandemic coverage. Most importantly, the pool allows for the quantification of risk from a reinsurance standpoint, which means reinsurers can support coverage in a limited nature as the amount reinsured is capped at the total size of the pool.

Issues with pandemic-risk modelling

Pandemics are neither as clear as earthquakes and typhoons nor as untraceable as terrorist attacks and financial crises. Epidemiological modelling is in the middle of these two extremes, like the modelling of large-scale cyber attacks and cyber risks. On the one hand, some rules can be followed, and the accumulation of data can help us analyse regularly on area such as mortality.

On the other hand, the economic loss can be affected by human factors such as the use of prevention and control measures, the effectiveness of these measures and the government’s strength in the implementation. These factors cannot be captured by a model. So, it is difficult to be categorised into the third risk pool, which is about the economic loss.

Designing pandemic-related products

From the perspective of the overall insurance industry, we should be more cautious about handling pandemics. There are two reasons: First, at this stage, major insurers are struggling to survive. Next, everyone will face a long-term low-interest-rate environment, which will have a huge impact on the financial stability and capital adequacy of each company. And the impact of the economic downturn on demand has caused the entire industry to enter a truly long-term depression state.

It is time to consider how to launch new pandemic-related products, accept new risks at this moment. In the long term, the insurance industry and governments could discuss insurance coverage, liability and product design with regard to the pandemics.

COVID-19 update – Manila, Hong Kong and Brunei

Manila has a total number of 155,256 COVID-19 cases as of 22 September with 2481 fatalities and 127,536 recoveries to date. The capital city has been experiencing a surge in cases from July with some days seeing over 6000 new cases.

Manila is considered to be the hotspot of the COVID-19 outbreak in the Philippines and is the biggest contributor to confirmed cases in the country.

To tackle a rising number of cases, the government enforced some of the world’s most prolonged and strictest quarantine restrictions from 15 March in Manila which is referred to as ‘general community quarantine’ (GCQ).

During this period, domestic travel was stopped and orders to stay home were imposed with companies and offices expected to adopt the work-from-home policy. Schools were closed and large social gatherings were all cancelled.

Currently, the GCQ has been extended till 30 September but some restrictions have been eased.

Most businesses, including dine-in services, has been allowed to reopen and religious services are permitted with a maximum attendance of 10 people.

Manila residents are also required to wear masks in public and observe onemetre social distancing while children, the elderly and pregnant women are urged to stay at home.

Hong Kong has a total number of 5038 cases as of 22 September with 103 fatalities and 4717 recoveries to date. The territory faced its worst wave of COVID-19 infections in end-July with the government warning its hospital system could face collapse.

Currently, Hong Kong has mostly contained its latest wave of the COVID-19 cases but extended social distancing measures till 29 September.

When COVID-19 cases first emerged in Hong Kong in late January, the territory kept its infection numbers down and was able to avoid extreme lockdown measures.

However, Hong Kong experienced a ‘second wave’ in March with growing imported cases. This led to the territory introducing strict border controls and banning all non-residents from entering its borders from overseas.

On 18 September, bars, pools and theme parks were allowed to reopen in Hong Kong. Bars were allowed to remain open until midnight and dine-in services at restaurants were extended by two hours until 12 am. However, the limit on public gatherings is still kept at four people.

The territory continues to prohibit transit flights to mainland China and restrict border crossings with mainland China remain restricted.

Brunei has a total number of 145 cases as of 23 September with three fatalities and 142 recoveries to date. The country is one of the few countries to avoid a massive surge in cases and deaths. It has reported no new COVID-19 cases for 19 consecutive days now.

A large majority of COVID-19 cases in Brunei were reported in March.

To curb the spread of COVID-19, the country enforced strict containment measures. It also banned both local and foreigners alike from leaving from 15 March onwards. Any travellers entering the country are required to take a COVID-19 test on arrival and quarantine for 14 days at a governmentdesignated facility at their own expense.

From 27 July, Brunei’s government eased COVID-19 restrictions allowing businesses to resume operations. Schools, markets, restaurants, sports facilities, and places of worship are now allowed to operate at full capacity. Public gatherings of up to 100 people are also permitted.

*Data has been sourced from the Johns Hopkins Coronavirus Resource Center as well as the World Health Organization.

The bias in AI algorithms

When AI was first introduced into business, it brought with it the promise of greater processing, efficiency and objectivity. While it has delivered on two of those three promises, the objectivity of AI algorithms is still open to debate.

On 7 November last year, tech entrepreneur David Hansson wrote a scathing Twitter post calling out Apple Card’s algorithm for being sexist. Apple Card is a credit card created by Apple in partnership with Goldman Sachs.

His tweet was prompted by Apple Card’s denial of a request by his wife for an increase to her credit limit – while he had a credit line 20 times higher than hers, based on the company’s algorithm. The couple live in a community property state and share all assets as a married couple.

In addition, a research paper published in Science journal in October last year found significant racial biases in algorithms that were widely used in US health systems. The research found that black people were considerably sicker than white people at a given risk score. Unequal access to healthcare in the US means less money is spent caring for black patients than white patients who have the same level of healthcare needs, and the bias occurs because the algorithm uses health costs as a proxy for healthcare needs.

In the end the algorithm erroneously concludes that black patients are healthier than equally sick white patients because it bases its decision making on healthcare costs. Fixing this disparity would increase the percentage of black people receiving additional help to match their actual healthcare needs.

While there is no direct evidence of such bias in AI in insurance, it is something that insurers should be mindful of as they come to rely on such technology more and more.

How did the bias arise?

AI was created to be different from other algorithms and should have the ability to learn from the data it is fed. The problem arises when that data contains biases to begin with.

A report by the Chartered Insurance Institute (CII) said that, “artificial intelligence, with its algorithms taught on historic data, is going to learn the gender biases ingrained in so much of our historic data and then propagate this into the automated decisions that insurers will be making in underwriting, claims and marketing. The algorithmic machines will learn societal biases.”

However, an article in MIT Technology Review suggested that the problem could start even before the algorithms learn from biased data. Depending on what an algorithm was created to do, it might make decisions that prioritise business or profitmaking objectives at the expense of fairness and non-discrimination.

Considerations for insurers

Given that biases in AI could lead to it profiling people of different races and genders differently, core operations in insurance such as underwriting and claims could be affected.

“Females, in roles that algorithms associate more often with males could find their policies being underwritten differently, or their claims looked at more closely,” said the CII report. It also brought up how gender bias in insurance decisions is illegal, just as it is in analogue insurance decisions.

Another report by Cognizant highlighted the “significant reputational damage and loss of shareholder value that can result from instances of inappropriate or biased AI decisions”.

But AI should not be written off so quickly. As pointed out in an article from the Insurance Information Institute, it will be much easier to eliminate bias in AI than it is in people. AI algorithms are just software that can be fixed with patches and updates.

At the same time, there is no magic bullet to this problem. Eliminating bias in AI will require concerted efforts to raise awareness of the bias and actively factor in the possibility of biases when creating algorithms.