LOSS-SAMPLER: Tools for quantifying and managing unseen volatility in CER portfolios

气候和天气风险对全球年度GDP的综合威胁超过1300亿美元（剑桥风险研究中心，2020年），使其成为对GDP的任何可想象的威胁中最大的，无论是人为的还是自然的。面对如此严峻的气候和环境风险（CER），巨灾（猫）模型已经成为（再）保险公司不可或缺的风险管理工具，并开始迎来一个新的时代，提高对风险的理解。起源于20世纪80年代末，猫模型在过去的30年里不断发展，以帮助解决全球社会长期可持续性面临的一些最复杂的挑战。在关于气候变化和金融稳定的演讲中，当时的英格兰银行行长直接赞扬了这些模型，指出“英国保险公司经受住了2011年的事件，这是有记录以来保险损失最严重的年份之一。你的模型得到了验证，索赔得到了支付，偿付能力得到了维持”（Mark Carney，2015）。虽然猫模特的出现无疑给社会带来了许多好处，但我们还远未挖掘出它对私营行业和社会的最大价值。我们仍然会被那些灾难性的事件所震惊，这些事件不仅是很容易想象的，而且在当代猫模型中得到了很好的体现，这说明了一项尚未充分发挥其潜力的服务。与许多历史悠久的学科相比，猫模型领域仍然处于新生阶段，这些学科的数据和信息为其提供了支持。这也许可以解释为什么许多潜在价值目前尚未开发。不幸的是，在过去的30年里，猫模型行业主要专注于部署针对狭义风险定义的分析-也就是说，在风险定价和资本管理情况下易于操作的定义。我们创建了一个分析引擎- LOSS-SAMPLER -利用当代猫模型平台来扩展这些狭义的风险定义。这使我们能够开始发现“未知-未知”信息，这些信息存在于猫模型分析的创建过程中，但由于与其应用相关的操作复杂性，这些信息一直被忽视。该分析引擎将不断发展，以应对过去几年需求迅速增长的三个特定业务挑战，即：1.协助评估CAT模型并实施可靠的风险视图2.通知猫风险投资组合管理3.应对这些挑战将有助于在全球金融体系中建立CER抵御能力。

The combined threat of climate and weather perils to annual global GDP is over $130billion (Cambridge Centre for Risk Studies, 2020), making them the largest of any conceivable threat to GDP, whether man-made or natural in origin. In the face of such severe Climate and Environmental Risk (CER), catastrophe (cat)-modelling has become indispensable to (re)insurers as a risk-management tool, and has begun to usher in a new era of improved understanding of risk.Originating in the late 1980s, cat-modelling has evolved over the past ~30years to help tackle some of the most complex challenges facing the long-term sustainability of global society. In his speech on Climate Change and Financial Stability, the then Governor of the Bank of England directly praised the models, stating that "UK insurers withstood the events of 2011, one of the worst years on record for insurance losses. Your models were validated, claims were paid, and solvency was maintained" (Mark Carney, 2015).While there have undoubtedly been many benefits to society from the advent of cat-modelling, we are far from extracting its maximum value to private industry and society. Our capacity to still be shocked by catastrophic events that are not just easily conceivable, but well-represented in contemporary cat-models, speaks of a service that has yet to realise its full potential.The field of cat-modelling is still nascent compared to the many longstanding academic disciplines whose data and information feed its processes. This may explain why much of the potential value currently remains untapped. Somewhat unfortunately, over the past three decades the cat-modelling industry has primarily focused on deploying analytics that target a narrow definition of risk - that is, a definition that is easily operationalised in risk pricing and capital management situations.We have created an analytical engine - LOSS-SAMPLER - that utilises contemporary cat-model platforms to broaden these narrow definitions of risk. This allows us to begin uncovering "unknown-knowns" - information that exists in the creation of cat-model analytics but has historically been overlooked because of operational complexities associated with its application.This analytical engine will be evolved to tackle three specific business challenges for which demand has rapidly increased in the past few years, namely:1. assisting in the evaluation of cat-models and implementing reliable views of risk2. informing cat-risk portfolio management3. quantifying and communicating the impact of climate changeAddressing these challenges will facilitate the building of CER resilience in the global financial system.