An assortment of senior figures in the reinsurance industry piled on the criticism of risk modellers following 2017’s natural disasters, but RMS is looking to set the record straight regarding how their models held up against the historic catastrophes. “We never changed our estimates for Hurricanes Harvey, Irma and Maria, and when the PCS numbers came out, it proved that our numbers were right as well,” said Mohsen Rahnama, senior executive for RMS’ models and data products, as well as chief risk modelling officer of the firm. “We’re not trying to capture a headline,” he said. “What we are looking to do is help our clients at a critical time. Our job is to understand the characteristics of an event, and capture as much information as we can.” According to Rahnama, catastrophes should be looked at in terms of the hazard, exposure, vulnerability of an affected area, and reconnaissance soon after an event. “Every event is unique, and as a result what we do has a lot of uncertainty, but we must never compromise the accuracy of the information we have compiled. What good is a loss estimate that contains a large gap?” Rahnama noted that RMS’ modelled loss estimates have real financial implications for clients, especially in the alternative capital markets. “We have a responsibility to the market, and we have a responsibility to our client. I asked a client, what would you like to hear? Would you like a result in one week? Two weeks? Of course, accuracy is important, and speed is important. He told me ‘I can hold up my board for one week, the second week is good, but it has to be the right range. I don't need a $40bn range.’ That is what our clients are saying,” Rahnama told Reactions. Michael Pritula, president of RMS mirrored Rahnama’s sentiment concerning the accuracy of RMS’ models. “We track our estimates and compare them with final loss statistics, and we have been quite pleased with the accuracy of our models and believe that they have held up pretty well.” “We are a little slower to release loss estimates following an event, despite the demand for estimates just a few days after a catastrophe. I do understand that it is essentially a marketing opportunity for those firms that can release their estimates first. Still, we prefer to take our time, and in fact, we could be even slower still,” Pritula told Reactions. One of the more specific criticisms brought against the risk modelling market was that there is not enough attention paid to regions outside of the mainland US, particularly in regards to Puerto Rico. But as Michael Young, senior director of model product management at RMS, explained there is simply more data coming out of the mainland US due to a lack of unified building codes throughout the Caribbean, as well as less claims data due to lower insurance penetration. But that did not stop RMS from making an accurate estimate of losses following Hurricane Maria. “The challenge for a modeller is that there isn't a body to go to say ‘show me the overarching building codes’. You literally have to go to each island and that’s one of the key things that we did, and it was a vital element of us making an accurate estimate for Hurricane Maria in Puerto Rico. We had been to the island, and talked to engineers, insurance agents, real estate agents-we really did understand what was on the ground, and more importantly not only what was on the ground but what was insured,” he told Reactions, rebuffing the idea that RMS was not as stringent when it came to modelling perils outside of the mainland US.