Hailed as the future of cat modelling, GR investigates whether the new software really is a game changer

Beijing floods

Two of the three biggest risk modellers, AIR Worldwide and Eqecat, launched new risk management platforms in January, called Touchstone and RQE respectively.

The third company, Risk Management Solutions, seemed to be late to the party. However, at its Exceedance 2013 client conference last week it unveiled RMS(one), which will be generally available from April 2014.

RMS(one) has been in development for three years. In some ways it is the RMS answer to Touchstone and RQE. However, it also takes the concept of risk management platforms and competition between the main risk modelling firms to a new level.

Here is the GR guide to RMS(one).

What is it?

As its name suggests, RMS(one) is designed to allow customers to run all their exposure and risk management tools from a single interface. This includes both RMS models and analytics and those provided by third parties.

Insurance and reinsurance companies can use RMS(one) to combine exposure data, models, analytics and applications to create their own view of risk.

In keeping with the offerings of some of its rivals, there is a strong focus on visualisation of risk data – presenting it in graphic format so it is easy for all in an organisation to understand – even those without technical modelling expertise.

The models, analytics and apps run in a software environment that RMS calls the analytic operating system (AoS). RMS(one) is only available via the RMS Cloud, which means users access it remotely rather than have technology on-site.

The AoS and RMS Cloud setup will allow users to run new high-definition models. While the results of the models will be the same as those on the current RiskLink modelling platform, the new high-definition models will give users greater visibility of and control over the assumptions the models use to generate their output.

What is different about it?

The most important difference is its acceptance of third-party models, analytics and applications. This introduces a new element of competition to the risk modelling industry. Instead of competing solely on model breadth and quality, risk modellers could now start competing on providing the best risk management platform.

It also acknowledges that no one risk modeller has all the answers. Allowing third parties to develop models and applications for RMS(one) should allow RMS to give its users a more complete view of risk than more closed systems.

Another difference is that unlike its peers’ offerings, which offer a choice of delivery methods, RMS(one) is only available in the cloud. RMS says this is necessary to be able to deliver the computing power and scalability needed to deliver the benefits of RMS(one).

Hemant Shah, RMS

The entire system has also been purpose-built for the insurance industry. RMS chief executive Hemant Shah (pictured) said: “We have not just competently assembled disparate technologies and stitched them together to deliver an answer.”

What problems does it hope to solve?

RMS(one) is designed to solve several issues that have arisen from the industry’s growing use of models.

One is that all the various sources of data a company needs to create its own view of risk can be fragmented. Companies need to draw on their own exposure data, models from various sources and external data. RMS(one) aims to bring them all together.

Also, over the past two years, (re)insurers and modellers alike have been extolling the virtues of a multi-model approach – combining several views of risk. The problem is that not all companies have the resources to implement the modelling platforms of all the various companies.

 

The healthy competition that has existed between AIR and RMS for more than two decades has been good news for the industry and for innovation’

AIR senior vice-president Bill Churney

 

 

 

 

 

 

 

As Lloyd’s insurer Novae’s chief risk officer Ian Hilder puts it: “It is just not practical for us to have ever licensed more than one provider’s cat models.”

RMS(one) could solve this by being a single platform for a variety of different vendor models.

Another issue is that modelled output typically needs condensing and translating for those with less technical backgrounds. With its visualisation tools, RMS(one) aims to make the output accessible to all, and allow chief executives and other decision makers as well as modelling experts ask questions of the data assembled in RMS(one) and get answers about their exposures.

RMS(one) also aims to speed up the process of getting answers from models and data. The firm claims RMS(one) will cut latency from “months to minutes”.

What are clients saying?

It is early days, and clients are still getting to grips with what RMS(one) can do and how it can help their organisations, but there is excitement about the prospects.

(Re)insurers have had to plug the gaps in current modelling offerings by developing their own tools, and for some RMS(one) could promise a neater solution.

Lloyd’s insurer Kiln’s head of insurance operations Robert Stevenson said: “We have hundreds of millions of rows of data and pages of details of the risks we have taken on, but the software is still a 10 to 15-year-old design and needs to catch. Up. What RMS is proposing is that catch-up.”

How have competitors responded?

Eqecat has struck an agreement with RMS to share exposure data formats. This means RMS users will be able to use RQE exposure data in RMS(one).

Eqecat also appears open to discussion for deeper integration. Eqecat senior vice-president and product architect Tom Larsen told GR: “EQECAT is currently engaged in dialogues with our clients and the market to assist in the understanding of market and client needs for model accessibility and usability – including any decision to enhance the RMS(one) platform.”

AIR Worldwide, however, has said it will not allow its models or modelled output to link to RMS(one). AIR senior vice-president Bill Churney said: “The healthy competition that has existed between AIR and RMS for more than two decades has been good news for the industry and for innovation. Therefore, in the interest of nurturing the innovation that has benefitted all of us, we will not allow our intellectual property – our models and model output – to be deployed on a platform or cloud that is operationally controlled by a competitor.”

Others do not share AIR’s concerns however. Three risk modellers – Risk Frontiers, ERN and JBA Risk Management – will allow their models to run natively within RMS(one).

Risk Frontiers managing director John McAneney said that sharing his intellectual property with a competitor was a big concern at first. However, he said: “With the assurances we have had from RMS, our understanding of RMS(one) and the fact that this is an issue for anybody that wants to put company-sensitive information on RMS(one), I am now really confident that our intellectual property is protected.”

What does the change mean for RiskLink users?

RMS(one) will eventually replace RiskLink, RMS’s current modelling platform. RMS insists it will not push clients to move to RMS(one), for example by only making the latest model updates available on it. The models in RMS(One) and RiskLink will be updated in tandem for two years after RMS(one) becomes available. In addition, RMS will continue to support RiskLink as a software application until 2019.

RMS is equally keen to point out that clients will not be forced to use all of RMS(one), and can choose to implement it as they see fit.

Hurricane Sandy

Source: NASA

 

 

Shah told delegates at RMS’s Exceedance 2013 conference that some adopters were going for complete integration, with real-time links between their own systems and RMS(one). But he added: “There will be some who will be using RMS(one) only to run basic straightforward model execution job and not even persist any data in the RMS(one) environment.”

Clients will continue to license models in the same way as they do now, and only pay for the elements of RMS(one) that they use.

What happens next?

While RMS(one)’s offering sounds promising, and seems to address several of the industry’s challenges with using modelled output, much will depend on how useful (re)insurers find it when they are up and running.

If it solves the industry’s problems and becomes a must-have, AIR Worldwide could face heavy pressure to open its models to RMS(one). However, RMS(one) will not be generally available until next year, which gives competitors time to respond.

If another firm is able to create a platform that does a better job of combining disparate models and data sources, RMS could be the one facing pressure to allow its models to run on another platform.