Connecting the dots: geospatial big data for sustainable finance

Location always mattered, both to measure impacts and risks; this is particularly true in sustainable finance, which made of the “where” an eligibility criterion in SDG finance[1]. Location always mattered, both to measure impacts and risks; this is particularly true in sustainable finance, which made of the “where” an eligibility criterion in SDG financea. Exposure to natural hazards, reliance on ecosystems, fulfillment of basic social needs, logistics chokepoints, are known to be influenced by geography, from the outset. Henceforth, access to geocoded addresses of millions of facilities, for hundreds of thousands of firms, is a game changer. Wide geospatial analysis across loan books or investment portfolios is now within reach.

Value chains dots are increasingly connected, among themselves, and to cartographic layers provided by scientific bodies, satellites or drones. When hard assets’ latitude and longitude coordinates are matched with cartographic layers – ex: heatmaps on land-cover/use, protected areas, wildfire hazards, water stress- risk analysis can take a new dimension. Shortly, this qualified data may support the integration of newly unveiled risk exposure into credit risk assessments, pricing models, covenant design, and capital allocation.   

This geodata supply & analytics breakthrough sketches new financial uses cases. “Geo-bigdata” [2] uptake is spurred by supervisory pressure (ECB & EBA’s stress testing and scenario analysis), technological progress (drones, satellite imagery resolution, machine learning and NLP[2]) and the rise of nature on the political and business agendas (incl. zero deforestation pledges). Point assets, but also linear or polygon assets (i.e. for spatially extended sites) are now geocoded, with additional characteristics such as physical form (ex: height, materials) or functional types (ex: office, warehouse).

Prior to Geo-bigdata, analysis was solely micro, or macro and top-down. Thanks to MSCI, S&P and the likes, micro analysis can become meso and then macro, through a bottom-up consolidation logic, which was previously inaccessible to financiers. Moreover, the detection of critical situations at facility level – ex: on deforestation risks– is eased through automated screening of large facilities samples, favoring financiers’ engagement with assets operators.

Nonetheless, despite progress, corporate asset data availability remains incomplete. Inventories of factories, pipelines, mines, power plants, warehouses widely differ by sector and geographies. In emerging markets, where property registries and disclosure often lack, coverage is limited with high geocoding error rates. Mapping from facility to legal entity is further hindered by JVs or SPVs structures.

Another challenge lies on information shortage regarding the operating patterns, and productive or economic importance of each asset—actual outputs, physical flows, asset efficiency, revenue/EBITDA by site. Hence, geospatial analysis predominantly captures inherent risks exposure overlooking actual hedging or risk-mitigation actions[3].

In some sectors, this lack of information on the criticality of individual assets obfuscates consolidation. The computation of vulnerabilities or impacts from micro to meso, and to macro, is then less rigorous and reliable. However, no one pretends that spatial information is self-sufficient. Building proxies is feasible through blending with other data streams on E&S governance, sub-ESG ratings, controversies screening, products labelling or certifications (ex: ISO).  Impact investing could be another winner of the geo-bigdata revolution. Territorial context is key to impact measurement and delivery, therefore, higher frequency asset and area monitoring should spur impact finance.