Green risk: Sberbank told how to digitize and calculate climate change

Climate issues are no longer an abstract topic. Climate directly affects every person, regions and states as a whole, as well as the processes and development of any business. Therefore, responsible business is paying more and more attention to the climate agenda.

The climate agenda – both combating the consequences of change and adapting to it – is today inextricably linked with socio-economic development. Tatyana Zavyalova, senior vice president for ESG at the bank, told reporters what climate direction Sber has taken in its updated ESG strategy.

“Today, temperature records are increasing the duration of droughts and water consumption, causing more frequent natural disasters, such as typhoons and forest fires. However, climate change affects more than just natural ecosystems—the agenda is inextricably linked to economic well-being, social satisfaction, and even the pace of technological progress. Warming also affects people’s health: for example, heat stress reduces productivity,” commented Tatyana Zavyalova, senior vice president for ESG at Sberbank.

The climate agenda, in Tatyana Zavyalova’s opinion, is already becoming a cross-cutting theme for almost all sectors of the economy: financing and investment, education and healthcare, agriculture and industry, healthcare and education. Therefore, companies are increasingly taking climate risks into account in their strategic planning.

“We, as a systemically important bank, were one of the first in the financial sector to develop an AI model for forecasting climate risk events and taking into account their economic consequences,” said Tatyana Zavyalova. – For example, with its help, we calculated that on average across the country for the period from 2024 to 2027, annual damage from natural disasters could amount to up to 580 billion rubles in certain scenarios. In addition, such automated risk assessment can be used across industries for strategic planning and development, as well as optimizing infrastructure costs.”

The use of such models allows the bank to accurately calculate credit and insurance risks, and helps clients build an effective strategy for their development. Zavyalova noted that the solution has a short forecasting horizon—only 6–12 months. This allows you to conduct stress testing much more quickly and make informed business decisions. Many similar systems have a forecast period of 10–20 years, which is less efficient in operational processes.

According to Zavyalova, the tool can be useful to a wide range of companies, for example, those involved in maritime transportation. Thus, according to the bank’s estimates, damage from storms as a result of global warming ranges from 2 to 2.5% of revenue due to logistics downtime. And for the oil and gas industry, climate risks increase costs for maintaining infrastructure and developing new fields. In some scenarios, operating costs increase by 5-6% per year due to the need for frequent repairs, including transport infrastructure, which is located in the zone of thawing permafrost.

“With the help of a solution, you can also predict positive changes. For example, to identify those regions and territories where, due to climate change, favorable conditions may arise for the development of agriculture and the forestry industry,” noted Tatyana Zavyalova.

The AI ​​model for forecasting extreme weather events and calculating their economic consequences has already been tested in different geographical zones of Russia. This solution can be adapted for other countries, noted Tatyana Zavyalova. Representatives from Vietnam have already become interested in a similar AI service for monitoring typhoons, which was jointly developed by Sberbank and the Far Eastern Federal University.

“In addition, we taught artificial intelligence to classify ESG risks based on processing data from open sources. Our open language model recognizes text on 47 environmental, social and governance topics and can, for example, help businesses audit contractors for sustainability compliance. For training, we used a dataset of two and a half thousand unique texts; the model can be integrated into any existing mechanisms for assessing ESG risks,” added Zavyalova.

Tatyana Zavyalova also emphasized that the bank, realizing the potential of the synergy of artificial intelligence and sustainable development, plans, as part of the updated strategy, to continue the development of smart technologies that will help in solving both climate, environmental, and socio-economic problems.

Addressing climate issues and mitigating climate risks is a unifying factor on the global stage. Zavyalova noted that as part of the updated ESG strategy, the bank will also continue to build international partnerships in the field of climate and green technologies, including ESG cooperation in the space of the BRICS, SCO and EAEU alliances.

At the end of last year, at the UN international conference – COP28 – Sber proposed creating a Climate Council of the BRICS countries. At the moment, the bank’s experts have formulated the main tasks, functionality and directions of work of the information and coordination center project, and active discussions are already underway with departments on its implementation.

Such structures, according to Zavyalova, are intended to help not only develop mechanisms for cooperation between states in the field of climate change and a sustainable economy at the supranational level, but also create a platform for exchanging experience and ensuring open access to green technologies and innovations on the principle of technological neutrality.