Japan: No Risk (Management), No Reward

Yasunori Araga, managing director – RMS Japan

Christine Ziehmann, vice president – Model Product Management, RMS

On June 14, RMS Japan welcomed 119 insurance professionals to its fourth In:Site conference — a record total, high-up on the twenty-sixth floor of the Sanno Park Tower in Tokyo, with the RMS Japan headquarters also based in this building.

Yasunori Araga, managing director of RMS Japan opened the afternoon event, and introduced an agenda that was highly relevant to the core concerns of RMS clients, from domestic issues such as earthquake risk in Japan, to the flooding and related losses following Hurricane Harvey in 2017. The agenda also explored the impact of modeling on one of the oldest lines of business — marine, and right through to one of the newest with cyber. The event attracted insurance professionals from multiple departments of insurance companies, such as reinsurance, risk management, property underwriting, casualty underwriting, marine underwriting, international and more.

Getting a Complete Picture on Japan Earthquake

With Japan still recovering from two major seismic events in seven years, with Tōhoku in 2011 and Kumamoto in 2016, delegates listened attentively to Chesley Williams, senior director, Model Product Management at RMS, who explained the science behind the recently released RMS Japan Earthquake and Tsunami High Definition (HD) model. She showed how RMS incorporated the latest science from the 2017 National Seismic Hazard Maps published by the Headquarters for Earthquake Research Promotion (HERP), and also how these two events generated billions of Yen in claims data, which were then used for model building and calibration.

Providing comprehensive modeling including shake, tsunami and fire following earthquake, the model also includes the “anti-correlated” perils of landslides — which tend to occur in mountainous regions. Liquefaction, more common in lowlands along rivers with a particularly high risk in the Kanto Plain area, is also included, with the model allowing for very differentiated risk selection, underwriting and portfolio management.

Chesley’s presentation was followed by a highly interesting talk by Professor Midorikawa, who highlighted how recent earthquakes have revealed the limitations of the current seismic hazard evaluation methodologies. He shared new approaches that have been discussed at HERP for improving these methodologies ready for use in earthquake modeling in the region.

Fixated on Flood

Japanese insurance companies were impacted by Hurricanes Harvey, Irma and Maria in 2017, collectively known as HIM, with Hurricane Harvey creating record precipitation, flooding, and losses.  With the U.S. state-backed National Flood Insurance Program (NFIP) largely covering the risk, only a very small fraction of the economic loss from Harvey will be covered by the private insurance market. As a result of this state intervention, Munich Re recently raised an interesting point, that solutions for flood modeling have typically not provided a complete risk picture. Hurricane models simulate losses from wind and storm surge — but not losses from flooding resulting from torrential rain. Pure flood models often do not include tropical cyclones, so for each approach, one component of the risk is missing. Munich Re concluded that insurance industry expansion into the U.S. flood insurance market will require more complete modeling solutions.

RMS has been building such a complete solution with its RMS U.S. Inland Flood HD model which was presented by Christine Ziehmann, vice president of Model Product Management at RMS. Using the RMS model shows that tropical cyclone flooding contributes about 25 percent of the annual average loss from flood losses nationwide and rising to 40 percent in the hurricane-affected states. RMS will be releasing this model coupled with the North Atlantic Hurricane model to offer a comprehensive and consistent solution to modeling flood in the United States.

New Approaches for the Oldest Line of Insurance

Events such as Hurricane Sandy in 2012 and the blast at the Port of Tjanjin in 2015 caused huge marine cargo losses across the industry and heightened awareness of the marine cargo risk globally as accumulations can be enormous in these global huge container ports. Manabu Masuda, senior director of Model Development for RMS presented the capabilities of the RMS Marine Cargo and Specie model, an industry first, and showcased real-world examples comparing the simple workarounds previously available with marine model results. The ability to locate the risks precisely within the port using highly detailed port shapefiles, to specify the exposure through many different product and storage types, to add information about packaging and protection and to specify cargo transit times all lead to much more realistic and differentiated results.

In contrast, cyber insurance is still in its infancy although as Tom Harvey, senior product manager at RMS explained there has been substantial growth in Japan domestic market over the last two years. This has been fueled by growing awareness of both the number of attacks and the substantial costs incurred by businesses following a cyberattack. RMS has adopted a multidimensional approach to identify and quantify key risk variables that determine the frequency and severity of cyberattacks.

This includes a thorough assessment of threat-actor groups, historical cyber incidents, corporate IT infrastructure and human vulnerabilities, digital assets at risk, and the interplay with insurance contract terms. This proprietary research is being released within the RMS third-generation Cyber Model which can be used to quantify both individual and portfolio-level risks providing the modeling required for insurers to grow their cyber business. The audience was particularly interested in the quantification of silent exposure, which continues to be a substantial concern across the market.

No Risk Modeling, No Reward

The event concluded with a very interesting dialog between Bernard Katanko and Masaaki Tanaka, both from McKinsey, discussing the strategic challenges global property and casualty business face today, and the importance of effective risk management practices in response to these challenges.

The insurance industry is all about risk taking, with a “no risk, no reward” mantra, but our speakers stressed that to be competitive, companies need to be smart about risk taking. Models provide the ability for deep risk differentiation by location, by peril, by sub-peril, by exposure types, and so on. Coupling this with a HD modelling framework also gives the ability to price any policy no matter how complex. Even after Japanese insurance market deregulation that largely ended in 2001, there is still a remarkable uniformity in terms of policies and premiums, but there are now the first signs of companies breaking away from this market tradition and embracing risk based pricing. The new insurance contracts standard IFRS 17, will be another driver for more detailed understanding of risk by peril and ultimately the need for modeling and better risk management. So, it would be more correct to say, “no risk modeling, no reward”.

In concluding the event, it was expressed that Japan is the first insurance market in the world with access to RMS HD models for all major perils, presenting a great opportunity to the market which has started to process their exposure data to be used in HD models. As the Japan market adopts HD models, RMS remains committed to releasing more detailed and better differentiating models and to expand modeling to areas such as agricultural risk and wildfire.