As globalisation and increasing catastrophes make risks around the world more and more complex, geoinformation and location analytics could open up a whole new vista for the insurance industry which needs to continuously innovate and develop new tools
June 2013. With floods continuing to wreak havoc in central Europe, Aon Benfield, the global reinsurance intermediary, created a scenario model using event footprints to enable insurers to obtain a realistic estimate of their exposure. As the flooding continued, the footprints were updated with new developments.
March 2011. When a 9.0 magnitude earthquake rocked Japan setting off a devastating tsunami, catastrophe risk modelling firm EQECAT needed quick information to update clients on risk exposure. EQECAT’s webinar Catastrophe Watch was ready within a week and attended by a host of primary insurers, reinsurers, brokers, corporations, hedge fund managers and investment firms.
December 2005. A series of explosions hit the London Buncefield oil depot, impacting the neighbouring commercial and residential properties. Within 36 hours, UK insurer RSA provided an estimate of the damage to its policyholders and the likely payout in claims. Months later, with all claims finally in, the estimation was within 7% of the actual amount.
Natural catastrophes cost the global insurance industry $40 billion in 2010 while man-made disasters cost $3 billion. In contrast, the total insured losses for 2009 were $27 billion, according to a report The Reactions Guide to Insurance Markets 2012. In 2012, Superstorm Sandy alone is estimated to have cost around $50 billion in insurance claims.
Accuracy is one of the most fundamental elements for insurance players. The more accurate one is in assessing the risk situation, the more efficient will be its risk management. This also means accurate allocation of reserves and no unnecessary holding up of the capital. With an alarming increase in the number and value of exposure in risk-prone areas around the world, it is becoming critical for insurance companies to adopt better risk management practices and move to risk-based pricing. However, an insurer’s ability to manage risks depends to a large extent on how well it knows the risk situation and the insured area; the primary question being ‘where’.
Location intelligence helps insurance companies to identify, manage and mitigate risks. “It is instrumental in catastrophe management planning and response, risk exposure management, pricing and cost control, and more importantly, financial survival,” says Bill Sinn, Strategic Marketing Director, Insurance and Healthcare Practices, Pitney Bowes.
Risk modelling firms like EQECAT and Aon Benfield, which use multiple data sources such as field surveys, land-use information, digital elevation models, satellite imagery, and develop on them with their internal experts from wide-ranging fields like seismologists, meteorologists, hydrologists, engineers, mathematicians, finance, risk management and insurance professionals, see GIS as the only tool to turn such disparate information into a meaningful solution. “We have shown that by using disparate yet accurate data in our models, we can reduce the uncertainty in loss calculation,” says Chris Ewing, Catastrophe Model Developer at Aon Benfield’s Impact Forecasting team. For instance, following the floods in Europe, Impact Forecasting used footprints based on its field survey, images from the German Aerospace Center and Astrium Services/Infoterra GmbH, and data from SERTIT supplied by PERILS. The hazard was then superimposed onto the insurer’s portfolio to calculate the exposed sum insured.
“GIS is an effective tool in decision mak ing of insurance business as it provides an integrated view of exposure, hazard and business data and helps unlock valuable information hidden in seemingly disparate data by analysing geographic/location linkages,” says Anup Jindal, Chief Operating Officer, RMSI, a geospatial solution provider in this space.
Agrees Iain Willis, Product Manager, EQECAT. “The CatWatch visuals following the Japan earthquake showed where the disaster occurred, the magnitude and the likely insurance exposure. These aspects are of immediate concern for clients, who will ultimately be impacted by these events and therefore, the need to react to the market quickly.”
Esri sees a huge demand for Esri’s Disaster Response Programme, which provides software, data, and technical assistance for disaster response, such as hurricanes. “If you can model the impact in advance, everyone can be better prepared for possibilities,” says Simon Thompson, Director, Global Commercial Solutions, who feels Hurricane Sandy, which struck North America in 2012, was a big lesson. Esri quickly provided authoritative information — including after-the-event imagery — so insurance carriers could overlay these with their own portfolios and perform analysis, adds Thompson, “Our users were able to estimate potential losses, work out how to deploy adjusters, identify risks and work with customers to mitigate losses, as well as help to educate the public.”
Flood footprint of Deggendorf develop by Aon Benfied
Geoinformation tools and analy tics are being widely used by dedicated solution providers like EQECAT, Aon Benfield, Willis, WSN, Insurance Services Office, Property Loss Research Bureau, AIR Worldwide, Risk Management Solutions, as well as pure insurance majors like Zurich Insurance, Allianz, Royal Sun Alliance, Amica and BMS.
The importance of location and geoinformation is underlined by the fact that geographic analysis has come up from being a back-office, ‘after-the-fact’ function in the property and casualty (P&C) insurance industry till a few years back, to a leading role in real-time underwriting decisions. After the record high losses due to natural catastrophes in 2011, carriers and reinsurers are demanding a vastly improved catastrophe risk management methodology, adds Sinn.
An encouraging trend, Thompson sees, is that these solutions are not predominantly GIS-centric, but are coming from a different angle, so often maps are not used. Instead, the outcomes of analysis and spatial data fusion are what the end users see.
GIS for reinsurance
Insurers control their exposure through risk transfer. This is vital to reduce acceptable levels of probability of a severe claim that could threaten its financial stability. Accumulation control, portfolio analysis and catastrophe analysis are integral part of risk management through which insurers apply risk transfer or reinsurance, cession limits, liability limits, etc.
Location is not a new buzzword for the reinsurance sector. Reinsurers and insurance advisory firms have for some time now been depending on geoinformation and catastrophe modelling. But as the world becomes more connected, recent catastrophes demonstrate new dimensions of risk correlation. Losses from the 2011 Thailand floods, for instance, revealed the magnitude of business interruption elsewhere. The floods disrupted automobile and electronics manufacturing in Japan — as a large number of automotive part manufacturers and electronic chipmakers were hit — and caused contingent business interruption losses for thousands of insureds in Europe and North America. Such interconnections add another dimension to catastrophe modelling, notes 2013 Global Insurance Outlook by Ernst & Young, which says increasing natural disasters are reshaping the insurer views of risk and reinsurance, compelling them to cede more risk to reinsurers.
“Geographic information has a strong influence on risk management, mainly risk identification and assessment,” says Andreas Siebert, Head of Munich Re’s Geospatial Solutions Department. Munich Re analyses customer portfolios in combination with natural hazard information or man-made risks. In the case of major catastrophes, it does a lot of loss estimation and claims analysis with geo-tools.
The significance of geospatial in the insurance sector can be gauged from the fact that the world’s biggest reinsurance firm had started way back in 1995 with a local standalone GIS-tool, before making geointelligence an integral part of its underwriting process, mainly for its non-life business such as property, marine, engineering. Siebert says the advantages are on the qualitative side: improved risk transparency and risk dialogue with stakeholders; improved risk modelling and pricing; new perspectives in portfolio steering; better risk control; higher efficiency in daily underwriting process, claims handling and risk mitigation.
Similarly, Swiss Re, the world’s second largest reinsurer, has a catastrophe net work service that disseminates information to direct insurers and other organisations, including risk rates and CRESTA zones.
The London-headquartered Willis Group also uses the whole spectrum of GIS capabilities. The group’s Willis Research Network, world’s largest collaboration between the public, science and financial sector, has dedicated hubs such as climate risk, earthquake risk etc, each of which is essentially geospatial in nature, points out Sharon Palmer, Divisional Director, Global Analytics.
GIS for property & casualty
With Google Earth opening up a whole new world, geoinformation systems have established themselves in the insurance industry more quickly than was anticipated. For an insurance company, especially in the commercial lines for products such as property and fire, having a geo-location capability is becoming a necessity. “It enables the tracking of the risk accumulation at a fine-grained level and provides an ability to do further analytics on whether the company is overexposed in certain geographies,” says Ram Medury, Vice President (Technology) ICICI Lombard, India’s largest private general insurer.
Zurich Insurance uses granular location information primarily to support risk selection, pricing, claims management for personal lines and risk/catastrophe management for commercial lines, reveals Vincent Branch, Chief Underwriting Officer, Motor & Personal Lines. Online resources and integration with non-GIS applications have also opened up a much wider vista for primary insurers.
»Risk management: By combining geographic and location-related data with other business data, organisations can gain critical insights, identify a trend or recognise a pattern amongst customer and demographic information, and are empowered to make better business decisions, notes The Next Big Wave — Location Intelligence in Insurance, a white paper from Cognizant.
“GIS has certainly helped us improve the quality of risk assessment, save time and reduce costs,” says Hiroo Shimada, Manager, Corporate Communications and Investor Relations Group, Tok io Marine Holdings. The Japanese insurance giant uses GIS mainly for risk assessment.
“GIS helps you gain insights by connecting predictive modelling and come out with different ways of managing a risk portfolio,” points out Manuel Z. Rios, President and CEO, American Modern Insurance (AMI), which provides specialty insurance products for homeowners, specialty dwellings, watercraft and collector vehicles etc. AMI, owned by Munich Re and part of the German insurance giant’s primary insurance operation in the US, seems to have learnt early lessons from its parent. It has all its policies geocoded and is now doing risk modelling for entire blocks of its business.
For contingencies like floods, pure street address information could lead to wrong conclusions, as Zurich Insurance learnt early. “In catastrophe risk management, where we rely heavily on models, we need pure location-specific information to better predict our exposure,” says Jean-Pierre Krause, Zurich’s Head of Risk Engineering, Europe & Middle East/Africa.
»Underwriting: Primary insurers need quick and accurate location-based information to come up with fair and competitive pricing models. They need solutions that organise, manage, and analyse extremely large datasets to measure risks such as proximity to hazards and emergency services, traffic patterns, commute routes, assets, and crime, which is how location data allows for better pricing, finds the report Understanding the Value of Location Intelligence Solutions by research firm Celent.
For the commercial property under writer, location-based solutions can map other insured properties in the area. Depending on the appetite for new risk given the aggregate risk for the area, he can then decide to proceed or not. For instance, UK-based Allianz Commercial uses an accumulation and perils management tool to analyse various risk levels against postcodes and produce hazard rating for an individual postcode. This means under writers can make decisions using up to-date information, which could then be combined with other information, such as flood risk areas, contaminated land or potential terrorist targets.
»Claims management: The speed and accuracy of claims processing is always an important factor for P&C insurers, especially during major disasters. New Jersey’s insurance regulator asked insurance companies to supply it with claims data soon after Hurricane Sandy struck in October 2012. It also gathered data on the National Flood Insurance Program claims. The National Association of Insurance Commissioners, an umbrella group for state insurance regulators, helped to crunch the data by county and ZIP codes. Despite use of all the available location intelligence, of the 4 million-plus claims filed in New Jersey alone, nearly 23% has resulted in no payment as of today.
Further, there are other areas where the claims department can act pre-emptively. Location intelligence enables it to accurately forecast the number of adjusters and inspectors required to handle an incident. The forecasting can also help predict if independent inspectors and adjusters are required.
Accurate location information helps insurers to avoid fraudulent claims too. For instance, during Hurricane Katrina, social networks carried the news that homeowners could claim $500 for electrical outages and the claims, many fraudulent, flooded in. Adjusters used mapping to see if a claim was coming from outside the area hit by a power outage.
»Marketing: In addition to risk analysis, geo-solutions also help insurers in marketing pitch like targeting underexposed areas, identifying prospects or change in pricing, write additional policies or depute more inspectors. By combining geographic and location data with others, organisations can gain critical insights, identify a trend or recognise a pattern among customer and demographic information, notes the Cognizant white paper. A typical example here would be insurers immediately reassessing their risk assessment strategies in cities following the 9/11 attacks. For instance, auto insurers who may have considered an underground parking facility safe some years ago today know that such structures are potential terrorist targets.
Geo-analy tics for customer segmentation, market penetration, producer and channel effectiveness help agents better identify and target the best business and consumer sales opportunities. Once the exposure in terms of sum insured and claims are aggregated at a specific administrative level, a thematic map of the sum insured and claims by administrative unit could be prepared. Visualising this geo-tagged exposure and claims against the demographics gives insight for identifying areas for growing business, odd patterns in claims and possible reasons, says RMSI’s Jindal.
»Customer service: A host of insurers are integrating location intelligence services with customer service portals. ICICI Lombard, a motor and health insurance provider, empowers customers with details of nearby garages, hospitals and branches, including real-time driving directions, via the location-sensing capability of the user’s mobile. “For both motor and health, customers get their claims serviced at various garages and hospitals, respectively. This translates into hundreds of locations across India,” explains Medury. During a claim situation, one of the first questions that a customer thinks is ‘Where do I get my car serviced?’ or ‘How do I reach the nearest hospital that is part of my insurance net work?’ Medury says this was the trigger for the company to create this tool.
Insurance companies could also use catastrophe modelling and their client information to send out a warning to policyholders to prepare for a storm or get cars into the garage. Because the system runs in real time, carriers can change plans and priorities as often as a storm changes course.
GIS for auto insurance
Evolving technologies like connected car and telematics are opening hitherto unexplored areas for auto insurance players. Although motor insurance is a mature market and the largest class of business within general and property/casualty insurance in many countries, underwriting profits have remained elusive for the sector for some time now.
Aggressive and unsafe driving can cost companies hundreds of thousands of dollars in employee injuries, lost productivity, fines, asset damage, lit igation, poor fuel economy and excess CO2 emissions.
Telematics devices can wirelessly monitor a vehicle’s geographic coordinates and driver performance in real time and has attracted huge consumer awareness, a Celent report on vehicle data and telematics has found. This has prompted auto companies like General Motors’ to come up with offerings like ‘OnStar’, a subscription-based communication on in-vehicle security, hands-free calling, turn-by-turn navigation, and remote diagnostics. Similarly, the ‘Sync’ service is available on all Ford, Lincoln, and Mercury products. Within Asia-Pacific, ‘GBook Alpha’ is a standard in most Toyota vehicles. Though the adoption rates are quicker in commercial fleets, the trend is pick ing up in personal vehicle space too.
“As specialised analy tics from these applications continue to grow, companies can gain real insight into driver activity. The right system can make a significant improvement to field service and fleet operations,” says Joyce Tam, Director of Product Marketing, Trimble Field Service Management. Trimble caters to fleet management and partners with carriers like Zurich for telematics solutions to help fleet operators mitigate risk and protect their people and assets. Early results on the use of telematics as part of the Zurich Fleet Intelligence proposition indicate a very positive impact on collision rates, adds Branch.
Fleet intelligence solutions can also help customers to understand their drivers’ risks and init iate risk mit igation action in the form of behavioural assessments and t rainings. Interestingly, Krause points out, location-specific information is just one element w ithin the mi x of data that is analysed to get a smarter understanding of risk. “The ‘where’ doesn’t tell us the full story. Of ten it needs to be brought toget her w ith the ‘how ’, behavioural aspects, that drive the risk,” he adds.
Agrees Tam: “Whichever telematics solution is selected, data on its own is of little value — how that data is actually delivered, interpreted and put to use is the key.”
In the compet it ive world of motor insurance, insurers are look ing at such data as a possible mine of information as this could be used to develop accurate pricing and reduce claims losses. For actuaries, such data could help determine new pricing variables and improve current variable quality; for under w riters, track ing data devices can provide validation of annual mileage, commute distances, garaging location, and ot her variables linked to leakage. For marketers, innovat ive products or unusual additional services could be used to reinforce positioning and branding. The data thus collected can also be used in claims assessment process — in case of stolen vehicle track ing or crash data to support claims.
Use of telematics has led to one of the most innovative schemes in motor insurance — pay as you drive. As the name suggests, it allows the policyholder to pay premium based on the car’s usage. The scheme relies upon a device that tracks a vehicle’s position, speed and other information via GPS and transmits the data to a production centre.
GIS for agri insurance
Innovative technologies like precision farming are streamlining crop insurance reporting. While automatic data capture simplifies the time-consuming crop reporting process, the systems deliver accurate planting and harvest information, avoiding mistakes from missed fields or inaccurate reporting.
John Deere, the US-based agricultural machinery giant, has its ow n insurance division as also an automated crop reporting service for its clients. Geospat ial major Trimble recently tied up with Great A merican Insurance Group to help farmers automate their crop insurance reporting under the US Department of Agriculture Federal Crop Insurance Program. The recordkeeping system makes it possible for Trimble’s Farm Works software users to submit their crop insurance records elect ronically to the insurer. Tools like Farm Works can be used to read precision agriculture data from a w ide range of displays, including planting and har vesting data. Once the data is read into the software, it can be submitted elect ronically to crop insurance providers, thus mak ing the process simpler and less t ime consuming, says Scott Nusbaum, Product Manager, Trimble’s Agriculture Division.
Return on investments
The benefits of using location intelligence for the insurance industry are difficult to translate into figures though and have tended to be internal rather than externally published. They can fall into one of the two areas: saving costs or generating new income.
As Rios spells out, RoI is loss avoidance. “I can’t give a number. But if we put together portfolio max imisation and the returns on investment based on cost and loss avoidance, the damages we would have had to pay had this technology not been there is extreme. I think a robust, real-time geospatial programme can get a minimum of 10% improvement in loss cost.”
Esri’s Thompson thinks the RoI is hard to calculate since much of it is intangible. “If you didn’t incur the cost because of certain better decisions, it is not going to show up on your balance sheet.” Buthe believes the consistent double-digit-plus growth in this sector is testament to its awareness of the RoI.
Access to a single tool for the end-to-end risk assessment process can reduce the time taken to process a new application. Providing frontline staff with the automated decision on a new prospect reduces the volume of referrals, saving both the time of the skilled underwriter and the capacity in the team, finds the Celent report. Rios elaborates: “Even if my policies keep growing 10% every year, the staff count could remain the same because the technology takes care of the workload. That is where the heavy RoI kicks in.”
Why is it then that the market not flooded with P&C insurers look ing for GIS solutions?
The reason is cost of investment and focus on core systems modernisation. Based on the data collected in its vendor review process in 2011, Celent found that a GIS project could cost around $1.6 million for single country implementation with 400 users and $600 million in premium income. This is a significant investment, and one that would require board approval and iterations through IT department. And herein, Celent identifies, lies the second problem. Many insurers are some way into their legacy modernisation programme and another large-scale IT investment against this backdrop is not a priority.
Swipe map of Hurricane Sandy
The main issue is convincing the top-level management about the potential of geoinformation. “A lot of them do not understand what longitudes, latitudes and real-time maps got to do with insurance. If they do not see the bigger picture, you do not get any funding,” says Rios, a rare case where the CEO is a geospatial enthusiast.
Also, in countries like India, primary insurers are not capturing structural and occupancy attributes of locations at the level of details that is enough for appropriate risk assessment from natural or man-made hazards, point outs Jindal. This could be because of the inherent lack of awareness about location intelligence in the developing world. Lack of historical data about hazards, absence of standardised data, integration of spatial and non-spatial data are also major challenges. Already, data from government agencies like meteorological departments, disaster management authorities, and security agencies like the police and armed forces are providing crucial layers to the primary data.
What is also required is, to facilitate proper and fair claims settlement in emergency situations, establishment of standards and best practices by the insurance regulators to address grey areas. For instance, insurance claims in the US due to Hurricane Sandy are expected to total $18.8 billion, according to ISO’s PCS unit which tracks catastrophe claims. The cost to private insurers might have been higher, but much of the damage was due to storm surge and subsequent flooding which was not covered by the commercial policies of standard homeowners or most small business owners.
In the US, coverage for floods is available from the federal government’s National Flood Insurance Program, which has been marred by reports of claims rejection following Sandy even after extensive use of geospatial data and technology. In developing countries like India, there is no system to even scientifically estimate the losses, let alone fix relief quantum after natural disasters like the recent flash floods and landslides in the Himalayan state of Uttarakhand.
Population growth, new businesses and wealth creation are driving growth in construction, land development, energy and transportation, all creating a greater need for insurance. Given that, it becomes all the more pertinent for insurance industry to use tools that spell out accurate location-specific requirements and risks. In times to come, new technologies like the emergence of unmanned aerial systems, Cloud, social networks and crowdsourcing are expected to add traction to the acceptance of geospatial data and technology by smaller players and further innovation by existing ones.
“Latest developments from risk modelling companies show that there is a now a full integration of mapping analysis and direct links to the so-called financial modules within the risk analysis tools,” says Siebert, who even sees pure insurance companies expand services on risk management in future.
Moving geospatial data and analysis to the Cloud is mak ing the technology lot more relevant. Questions that might not have had answers only three years ago can now be solved with more data and powerful analy tics, says Thompson, who feels the acceleration is phenomenal, leading to many more geo-centric solutions for insurers.
The insurance industry understood the need for accurate geo-information years back, which saw the establishment of CRESTA in 1977, an independent, non-profit organisation, aimed at creating an efficient, globally uniform system for the exchange of accumulation-risk data. Based on geographically defined accumulation areas, known as CRESTA zones, it now covers 79 countries
Further, there is a huge potential in the health and life insurance market still to be tapped. “GIS is a powerful resource for sales and marketing because it brings together a company’s internal view of the customers’ life with public demographics, lifestyle and location information,” says Thompson. Companies are increasingly creating specialised products based on specific characteristics for each family or household. This wide range of offers benefit both the insurer and insured; things like life and health insurance, mortgage and income protection, auto, travel and even pet insurance. GIS improves the agency role so companies can better understand, price and target products to this sector which globally accounts for more than 60% of all written premiums.
Ruggedised data collection and infield devices have opened up the ability for both sales and customer care work to be done any where with mobile-based location analytics and GIS. New insurance models are arising from the ability to view and analyse telematics data for both commercial fleet and consumer automotive insurance industries.
The ability to aggregate, visualise and manage risks, in some cases, is fundamental to regulatory compliance, overall solvency and long-term viability, says Sinn of Pitney Bowes. “Yet, ithas been difficult for most insurers to gain a single, operational view of risks across their organisations. Operational, data and technology silos have obstructed a 360-degree understanding of the risk, and the ability to act on it. But new technologies now offer a single operational and map-based view of risk in near real-time.”
A demand for such accurate data also opens up the problematic area of geocoding. Although street-level geocoding solutions exist for some countries, it is still an issue particularly in the developing world. Geospatial players see a huge potential there, something the insurance sector is also upbeat about. “It will mean that exposure in emerging markets can be more precisely located,” says EQECAT’s Willis, who also sees a growth area for ‘lightweight’ geo-visualisation tools. “Although GIS desktop tools will continue to be needed for the ‘heavy lifting’ for analytical work, I foresee an increase in the amount of spatially enabled dashboard reports.”
Insurance players also feel the sector is not a priority for the geospatial industry. “Big geospatial companies are focussed on municipalities, governments, mining, agriculture… if they were to dedicate some resources just for us that would be a home run for both. There would be a whole new world, because insurance is big money,” emphasises Rios.
The global economic slowdown, a prolonged period of low interest rates and subsequent anaemic returns highlight a critical mandate for increased under writing excellence for the insurance industry, predicts Deloitte’s 2013 Property and Casualty Insurance Industry Outlook.
The future success of location intelligence in insurance lies in how quickly it realises the opportunity, breaks away from the traditional GIS world and integrates into the key operations of corporations. Because, as Aon’s Ewing says, if you do not know where your risks are located you do not have an accurate view of your risk.