Managing Climate Risk with AI

Artificial intelligence, natural language processing and big data can better assess financial losses from climate change.

In September 2018, Typhoon Mangkhut battered Hong Kong and southern China, causing power failures, business and school closures, and the emergency evacuation of 2.5 million people. Total damage in China came to an estimated US$1.96 billion (RMB13.68 billion). In recent years, catastrophic events such as this have been on the rise due to the effects of climate change, and it is insurance companies that have taken the lead to assess and manage the ensuing economic fallout. As more companies seek ways to predict and hedge against risks from climate change, few industries have as much experience and know-how as the insurance sector.

Ping An - one of China’s largest insurance, banking and financial services companies—has long understood the risk of climate change on its businesses and is tackling this challenge with advanced technology. The company’s life insurance and property and casualty insurance units have started using artificial intelligence to monitor climate change and environmental quality in China. By accumulating data and employing AI algorithms, Ping An has been able to build models that can assess business risks from climate change with far better accuracy. The company is also embedding environmental, social and governance (ESG) factors into its investments, using AI to analyze data about companies and hedge against risks from climate change.

“As a global, leading integrated financial services group, Ping An has always responded proactively to business risks and operational risks arising from climate change,” says Richard Sheng, Ping An Group Board Secretary. As one of China’s leading financial conglomerates, Ping An achieved US$18.6 billion (RMB132.96 billion) in operating profit in 2019 and has an investment portfolio of insurance funds of US$449.4 billion (RMB3.21 trillion) at the end of 2019.

Ping An adopts ESG-governance structure integration as an approach to standardize management. The board of directors is fully accountable for the company’s management of ESG and climate change risks.

USING AI TO ASSESS INSURANCE RISK

To minimize damage from catastrophic events, Ping An developed an internal digital risk system that can assess extreme weather events, such as typhoons, heavy rain or floods, by analyzing a variety of sources, including satellite images, data from drones, information from Internet of Things devices and location-based services positioning. The system combines more than 1.4 billion data points from disaster and meteorological science, among other sources, as well as Ping An’s own underwriting and claim records. Together, the system delivers a more thorough risk assessment that can help prevent or decrease financial losses.

As Typhoon Mangkhut approached China, Ping An was able to use its digital risk system to draw on its own data as well as data from the Chinese national meteorological system to provide warnings to 8,000 customers in affected areas. Ping An sent SMS about when the storm was likely to hit as well as tips on how people could prepare. The company also harnessed data to tailor critical safety information to their clients’ situations.

“As the platform accumulates more data and we refine the models, we can get more accurate alerts before a disaster hits,” says Crystal Geng, Group ESG expert at Ping An.

Ping An is also assessing physical, transitional and other types of risk from climate change in its product development, underwriting and product management.

By incorporating climate-risk factors into the pricing process, Ping An comprehensively considers the impact of climate change on the client’s assets and insurance claims based on the historical data and analysis of its internal digital risk system.

HEDGING INVESTMENT RISK

Gathering ESG data in China and elsewhere can be challenging. Unlike financial data and disclosures, which must adhere to strict reporting standards, ESG reporting can vary significantly among companies. And that’s if companies report that kind of information at all.

The good news is this is starting to change. For the past two years, the Hong Kong Stock Exchange has required ESG disclosure from the 2,500 listed companies. Shanghai and Shenzhen are gearing up to introduce similar ESG disclosure guidelines soon.

Without adequate data, accurate risk analysis is challenging. It’s easy for companies to do simple things like reducing exposure to coal mining or thermal power generation, but a deeper level of information is needed to decrease the risk of climate change in a diversified portfolio.

Ping An is pulling alternative data and using natural language processing to do its own assessment of companies based on various ESG metrics. The company has developed a platform called AI ESG that crawls the web for relevant information about companies. For instance, if an auto company announces it will launch a renewable energy car, that might improve its ESG evaluation rating because it signals the industry is shifting from carbon-based to renewable energy due to climate change.

The AI platform, which assesses both listed and unlisted companies and various asset classes, is for internal use only right now, but there are plans to make it available as a product to other companies later this year.

Richard Sheng, Ping An Group Board Secretary

A SAFER HARBOR

With its broad adoption of AI and its data-rich technology platform, Ping An is a pioneer in installing precautions against investing in industries and regions that are particularly vulnerable to climate change risks. In September 2019, the company became the first Chinese asset owner to sign the Climate Action 100+, a global investor initiative to ensure the world’s largest corporate greenhouse gas emitters take action on climate change. Ping An says the move was a big milestone for ESG investing in China.

By incorporating climate-risk factors into the pricing process, Ping An comprehensively considers the impact of climate change on the client’s assets and insurance claims based on the historical data and analysis of its internal digital risk system.