WWF applied its technical expertise by identifying those crops and basins facing the most significant water risks globally. 

To achieve this, WWF assessed the water risks of all crops and river basins included in the Agricultural Water Risk Assessment Add- On to the Water Risk Filter. This includes 405 river basins and 122 crops (over 15,000 crop and basin combinations) and was used to identify and rank which crop-river basin combinations face the highest overall water risks globally. A focus was placed on crops that are irrigation dependent, grown in water stressed areas, and globally traded. The data sets used in the assessment are all from publicly available sources, and include the latest globally comprehensive data on water scarcity, drought, crop water use, production area, and pollution levels, among other factors.

After identifying the crop-river basin combinations exposed to water stress, WWF aggregated these crop-basin combinations into crop-country combinations based on respective crop production percentages per river basin. This was to make the prioritisation of geographies more userfriendly for investors; however, both sets of data have been developed and presented in the final assessment. WWF’s final phase in this process involved designing a structure which allowed sorting and filtering the crop-country combinations by crop production volumes, export volumes, and export value. Applying such underlying economic data to the crop-country combinations enabled the narrowing down of the large list of combinations to the most significant global crops in terms of both economic importance and water risk.

The outcome of this process was an indicative list of 25 key crop-country combinations which are both highly exposed to water risks and of high economic importance; these included wheat in Bangladesh as well as grapes in Australia. This list was provided to PwC for alignment with their ESCHER tool, along with the universe of 78 target companies compiled by the steering committee, for their research into the water risk exposure of individual companies.

PwC - Identifying companies exposed to water risks 

The PRI commissioned PwC Germany to both identify the most appropriate companies for inclusion in the collaborative engagement, and to provide information which will enable the group to hold meaningful dialogue with investee companies on their exposure to water risks in their agricultural supply chains. In particular, PwC was tasked with identifying the most exposed companies according to WWF’s crop-country combinations list. Additionally, PwC shared data on the water management practices of the researched companies, in collaboration with oekom research.

The PwC ESCHER and supply chain model

For the research, PwC applied its ‘ESCHER approach’ which estimates water consumption in water stressed regions across the world. The Efficient Supply Chain Economic & Environmental Reporting (ESCHER) model provides a multi-criteria analysis of portfolios covering the complete upstream value chain. ESCHER can be described as an Environmentally Extended Input-Output Model; a stateof- the-art model to assess and estimate direct as well as indirect effects of separate sectors and regions. It is a statistical approach which was developed to shine light into global supply chains and procurement.

ESCHER applies Leontief’s Input-Output Theory to assess the worldwide interaction of 57 sectors in 129 regions. The model shows intra and inter-regional linkages between markets and market participants, as well as the resulting feedback effects.

To quantify the supply chain effects of separate sectors and regions across the world, ESCHER is based on an assumption that global foreign trade activities, as registered through the Global Trade Analysis Project (GTAP), estimate individual companies’ procurement activities. GTAP includes trade activities at a global and regional / country level, bilateral trade patterns, as well as production, consumption and intermediate use of commodities and services, all of which form the basis for water consumption estimates. These estimates are calculated by the system which are then normalised as water consumption per monetary unit.

To reflect the regional characteristics of water, the inputoutput approach was extended using data from WaterGAP3 – a global water resource model developed by the University of Kassel.24 WaterGAP3 simulates hydrologic processes for finite areas of land across the globe (global geographic rasters of 5 arc minutes, i.e. approximately 6 x 9 km in Europe). The water scarcity data from WaterGAP3 is used in ESCHER to model consumption-to-availability ratios for different crops in individual regions around the world. The consumption-to-availability ratio is then used as an indicator of areas under water stress (0 – 0.05 no or low water stress; 0.05 – 0.2 medium water stress; 0.2 – 0.3 high water stress; > 0.3 severe water stress). Such data can then be aggregated for each country to calculate the percentage of that country experiencing water stress for a given crop; which in turn is matched to the consumption of water for individual supply chains being modelled. Note that this step in the ESCHER approach effectively replicates work done by WWF using their Water Risk Filter earlier in the research process. This is because the PwC ESCHER model was already based on WaterGAP3 data and could not be readily modified to factor in WWF’s research.

Limitations of ESCHER and supply chain modelling

The more detailed information on supplier structures, the higher the granularity of ESCHER outputs. Research completed for the PRI was based on publically available information on supplier structures. This means that supplier patterns for each company are based on average sourcing structures for the relevant sector, complemented by additional information on company structures and sales information, to calculate water demand globally. With nonpublic information, such as detailed supplier information from companies’ procurement systems, ESCHER produces more detailed and accurate outputs.

It is important to note that modelling through an inputoutput system such as ESCHER, with limited transparency into the specifics of individual company supply chains, does not provide robust data on actual individual company water risks. Without knowing (confidential) detail on supply chains, it is only possible to estimate, which can be done using relevant but never fully correlated alternative data, such as trade flows and hydrological modelling systems. The output from ESCHER therefore provides investors with a first estimate of risk exposures for the modelled companies, but not empirical data on water consumption and water risk. Instead, the output can be used for prioritisation of engagement and as a tool for improving understanding of supply chain risks for both companies and investors.

Data on management control 

For the research, PwC also provided data on management control of individual companies from oekom research. This data is based on publicly available company documentation and external sources such as NGOs, government bodies, trade unions and the media.

Oekom data rating risk management was available for 48 companies in the food, beverage and agricultural products sectors and for 10 apparel companies. No data on retailers was available.

Each company with data available was given an overall rating of between 1 and 4 to summarise their performance on water management. Scores of 1 are equivalent to a D- (poor performance), while scores of 4 indicate an A+ (excellent performance). Specific oekom data used for this research included ‘policy on sustainable water use in the company and its supply chain’ and ‘measures to ensure sustainable water use in the company and its supply chain’ as indicators of overall company water risk management, including in the supply chain. In addition to oekom research data, PwC and the investor steering committee also consulted company disclosures to CDP Water. This information, like the oekom research, was not exhaustive, covering only 19 of the 78 companies in the universe in the 2013 CDP dataset (the most recent at time of writing). There was some limited detail on company sourcing structures available in the CDP reporting but again, this did not include all companies in the target universe.

Download the full report