Finextra Research and Responsible Risk today hosted Sustainable Finance Live, the second virtual workshop in a series of events designed to create actionable ESGtech strategies and build an ecosystem of partnerships that will turn strategy into reality.
This workshop details how alternative data from sources such as satellites and sensors can augment traditional risk systems and real-time, forward-looking, data can provide insights for the future of sustainable financing.
What are the issues and opportunities for risk management working with alternative data to inform credit decisions? How can these decisions be quantified against physical and transition risk? Richard Peers, founder of Responsible Risk and contributing editor for Finextra Research, provides the answers.
Satellite data for agricultural risk
Rishikesh Sapre, co-founder of Mantle Labs, kicked proceedings off by describing how Axis Bank uses alternative data to inform its credit risk decision making in agricultural lending.
Axis Bank harnesses Mantle Lab's agritech platform Geobotanics, which aims to provide a risk framework that can spur investment in smallholder farms in Asia and Sub-Saharan Africa.
There are 500 million such farms in these regions, which produce 80% of food for two billion people. These farms struggle to obtain credit from financial institutions however due to lack of visibility over their viability, thus they are generally reliant on government subsidies.
"The platform can enable the creation of affordable products for farmers, ensuring that their business model is viable for financial services companies and provide a global solution without the need for calibration,” Sapre says.
The Geobotanics platform has used satellite data to assess agricultural performance in different areas over the last 10 years to provide a probability of crop failure due to factors such as drought, adverse weather conditions or pests.
The satellite imagery is processed every 10 days, automatically filtering out or highlighting certain higher risk regions when they pass the appropriate threshold.
The probability of crop failure is directly linked to the likelihood of a default on a loan, thus the platform provides a risk capital allocation for banks to use.
Such data can then be used to offer a projection about how regions will perform in the months and years ahead, which helps to provide an end-to-end framework for a bank to manage its agriculture portfolio.
What Geobotanics does, according to Sapre, is “build business layers”, helping to integrate the relevant data into a traditional financial business model.
Mantle Labs regard the platform as a breakthrough for financial inclusion through enabling deployment of profitable farm credit and insurance, easing any dependence on government subsidies, while also providing farmers with cost-effective pricing of financial products and access to complete financial planning and security for their livelihood.
“To sum up, fifteen smallholder loans have been disbursed through Geobotanics just during this 15-minute presentation,” Sapre claims.
Normalising forward-looking data
Filling in contextual concerns, Maya Hennerkes, ESG sector lead, EBRD, follows on from Sapre to explore the issues and opportunities for risk management, working with alternative data to inform credit decisions and why there is a clear need for these decisions be quantified against physical and transition risk.
“Why are we discussing this?” Hennerkes questions. “Because there is still such a significant and widening gap between where we are and the objectives we should be working toward. "We've so far failed to meet our collective financing pledge that accompanied the Paris Agreement, and this isn't necessarily helped by extending the goal of USD100bn annually to 2025."
Hennerkes explains that the Paris Agreement has utterly missed its goal to mobilise $100bn USD of new and additional financing annually by 2020, reaching just $70bn USD in 2019 and gradually less in the years prior. She furthers that the gap of the financing we need to achieve to reach the objectives is only widening each year.
The EU’s Green Deal which calls for funding of around one trillion euros for investments to meet the carbon neutral goal by 2050 is also struggling to show progress, Hennerkes explains.
“We see that the funding just isn’t flowing and the reasoning behind this always seems to come back to the cause of risk and risk management. Traditional institutions, while being encouraged to adopt very innovative technologies have typically looked at track records in order to determine risk, but this simply doesn’t exist with many new technologies.”
Risk therefore presents a fundamental barrier to the requirement of ‘bankability’, which in the traditional sense, means that a given project ticks every box. The larger the investment, the more boxes that need to be ticked.
“The type of institution is also relevant. We the EBRD, are a multilateral development bank, and as a public institution we have a very wide and diverse range of stakeholders with legitimate requests or concerns around how we apply our funding.
Depending on the nature of the institution reaching the threshold of bankability can be a very demanding task.” In order to mitigate the typical ‘backwards looking risk modelling’ prevalent across financial services, Hennerkes explains that the EBRD is increasingly looking to ‘non-financial’ risks to integrate more detailed, future looking data into its modelling.
“Because we need to be integrating this future-looking data such as climate risk prognostics which could, for example, and as Rishi Sapre just discussed, illustrate the impact climate change can bear on certain crops.
“We need to be able to model the risk of innovative technologies themselves which don’t have a track record but potentially have a very high positive impact.”
Adding that there are also well documented shortcomings across the current financial modelling approaches particularly in light of struggling to measure ESG impacts themselves, Hennerkes says that if we were able to measure these externalities accurately we would in fact have a much clearer measure on profitability in addition to the triple bottom line.
“If we can do this at the EBRD than I know others can too – we’re no smarter than everyone else is that’s for sure. We must think about this collectively.”
Giving financial institutions what they want
David Patterson, head of conservation intelligence, WWF then speaks to a recent report the organisation released in collaboration with the World Bank that highlighted how the financial sector can benefit from the emerging field of spatial finance, as the area complements existing ESG data streams and could provide an outline for a robust taxonomy.
Patterson also explores how satellite and sensor data can provide unprecedented insights for investment, risk and governance professionals and what needs to be done to achieve wider adoption. After building a global mapping tool five years ago, the WWF were able to see how the natural world was changing and which companies were causing that impact.
Patterson reveals that this is the information that financial institutions were looking for, but a lack of understanding of spatial finance resulted in frustration. “Spatial finance, at a very basic level, is a geospatial approach where we are defining company’s assets, their locations and their suppliers’ asset locations and comparing those locations against observational data.”
Opining on this point, he provides a definition for asset data. Where “asset data is the information which defines where companies are,” for example, ownership name and a X and Y latitude and longitude at an absolute minimum, in actual fact, asset data “comes with more attributes” such as commodity investment.
Patterson goes on to say that with more data, organisations gain more “intel” but “it’s hard to maintain a global data store when you have hundreds and hundreds of attributes.” Further, asset data is both open and closed, but the best datasets tend to be held in the commercial sphere behind a paywall.
However, as mentioned before, to provide financial institutions with information on which companies are causing changes in the natural world, asset data will need to be compared against observational data. Patterson adds: “It can be remote sensing products, it can be ground data, smart meters, air pollution. […] The challenge is how do we link this together.”
The WWF have a five-tier proposed taxonomy for spatial finance:
Tier 4: Sub-asset level data – Assessment within asset – IoT, smart meters, traditional ESG reporting etc.
- Taking measurements from within the asset itself, for example, a power plant, to extract data that is of a very high temporal resolution.
- Because of this and as it is sector specific, there are legal issues and is reliant on cooperation from operators.
- While this method is insightful and interesting, sector resistance to releasing data will result in complications.
Tier 3: Asset level – Assessment of the asset – GIS overlaps, remote sensing, plus Tier 4
- Assessing the entire asset, for example, a protected area within a World Heritage site, to analyse deforestation, indigenous rights, methane emissions etc.
- This is the most developed part of spatial finance, but a distinction needs to be made around which tools can provide asset data attached to them.
- For those clients that do not have asset datasets, data can be ingested into observational platforms for analysis.
Tier 2: Parent/company level – Summed or aggregated scores for a parent company, based on Tier 3 and 4 results
- At this stage, data is no longer introduced but data is aggregated from Tier 3 and 4 and compared.
- The challenge here is that asset data for all sectors is not available, so therefore, in order to obtain a good understanding of a company’s impact, their operations, their assets and their suppliers’ assets need to be assessed.
- There are a range of commercial tools in the market that fulfil the purpose outlined in Tier 2.
Tier 1: Portfolio level – Summed or aggregated scores for portfolios, based on Tier 2 company scores
- While there are commercial tools that only partially fulfil Tier 1 currently, WWF is looking at selling parent companies up to a portfolio level.
- Some players are also using traditional ESG approaches to fill the gap.
Tier 0: Country level – Summed or aggregated scores for countries, based on Tier 3 and 4 data
- Using the asset data, observational data and data on how they interact, regional aggregation can occur, otherwise known as sovereign debt financing.
Patterson states: “Most of the challenges are not technological. We don't need a magical new type of blockchain to be invented to enable this to happen. Most of the problems are just data-related: we need better data, we need insight on ownership, companies need their asset data and we need the supply chain data.”
He adds: “From my point of view, we also need to improve climate and environmental observational data.”
Finextra and Responsible Risk invite readers to continue this discussion, take these insights and design a solution with peers. Register here for the Co-Creation workshops on 8-9 Dec.