Appendix
Appendix A: GHG Factors for U.S. Food Production and Surplus Disposal
ReFED developed the following weighted average GHG factors (Table A1-A5) for each sector by using the individual food category GHG factors developed by Quantis (Table A6) as proxies for different food types in each sector. After assigning the Quantis factors as proxies (e.g., Bananas were used as a proxy for all heavily imported tropical fruits), ReFED used the surplus tonnage results from the Food Waste Monitor to weight and aggregate the factors to less granular food types (e.g., Produce). This was also useful for developing a single ‘Standard Mix’ GHG factor for each sector, as this is one of the most common requests from businesses in cases when their waste data may not be broken down into multiple food types. Negative GHG values indicate a GHG reduction.
Raw Data and Documentation for Weighted Average GHG Factors (the individual Quantis factors with additional decimal places can also be found here):
https://refed-roadmap.s3-us-west-2.amazonaws.com/public_documentation/Documentation_Solutions_ GHGFactors.xlsx
N/A = “Not Applicable”
Table A1. Farm (Produce Only) Weighted Average GHG Factors for Food Production and Disposal
*Donations numbers account for transportation to a food bank plus storage, and they assume that every ton of food donated results in one less ton of production to meet food demand. Donations numbers also assume that 4.2% of food donated to food banks actually gets landfilled [:ref:``] as opposed to distributed to people as intended.
**Industrial Use numbers were estimated by modeling the impacts of rendering.
***ReFED reused the Land Application numbers to estimate the impacts of Dumping. More research is needed to account for the differences in emissions between the two destinations.
Table A2. Manufacturing Weighted Average GHG Factors for Food Production and Disposal
*Donations numbers account for transportation to a food bank plus storage, and they assume that every ton of food donated results in one less ton of production to meet food demand. Donations numbers also assume that 4.2% of food donated to food banks actually gets landfilled [:ref:``] as opposed to distributed to people as intended.
**Industrial Use numbers were estimated by modeling the impacts of rendering.
***ReFED reused the Land Application numbers to estimate the impacts of Dumping. More research is needed to account for the differences in emissions between the two destinations.
Table A3. Retail Weighted Average GHG Factors for Food Production and Disposal
*Donations numbers account for transportation to a food bank plus storage, and they assume that every ton of food donated results in one less ton of production to meet food demand. Donations numbers also assume that 4.2% of food donated to food banks actually gets landfilled [:ref:``] as opposed to distributed to people as intended.
**Industrial Use numbers were estimated by modeling the impacts of rendering.
Table A4. Foodservice Weighted Average GHG Factors for Food Production and Disposal
*Donations numbers account for transportation to a food bank plus storage, and they assume that every ton of food donated results in one less ton of production to meet food demand. Donations numbers also assume that 4.2% of food donated to food banks actually gets landfilled [:ref:``] as opposed to distributed to people as intended.
**Industrial Use numbers were estimated by modeling the impacts of rendering.
Table A5. Residential Weighted Average GHG Factors for Food Production and Disposal
*Donations numbers account for transportation to a food bank plus storage, and they assume that every ton of food donated results in one less ton of production to meet food demand. Donations numbers also assume that 4.2% of food donated to food banks actually gets landfilled [:ref:``] as opposed to distributed to people as intended.
**Industrial Use numbers were estimated by modeling the impacts of rendering.
Quantis developed the following GHG factors. See the methodology [:ref:
][:ref:] to learn more.Table A6. Individual Food Category GHG Factors for Food Production and Disposal
Appendix B: Water Footprint Factors for U.S. Food Production
ReFED developed the following weighted average water factors by using individual food category water factors developed by the Water Footprint Network (WFN)[:ref:
][:ref:] as proxies for different food types. The WFN factors include water use throughout the supply chain (farm to end-of-life) and are not broken down by supply chain stage (sector). Therefore, these factors do not account for differences between sectors (e.g., Water use for manufacturing is embedded in the water factors and is used to estimate the water footprint of farm surplus.). The WFN factors used are specific to the United States. ReFED chose to only include WFN’s blue water footprint factors [:ref:``]After assigning the WFN factors as proxies (e.g., Wheat bread was used as a proxy for bread and bakery items), ReFED used USDA survey farm production tonnages7 and Nielsen IQ grocery retail sales8 to weight and aggregate them to less granular food types (e.g., Produce, Breads & Bakery). This was also useful for developing a single ‘Standard Mix’ water factor for each sector, as this is one of the most common requests from businesses in cases when their waste data may not be broken down into multiple food types. With additional research, future iterations of this work could take a more robust approach similar to the previous section on Greenhouse Gas Emissions (Appendix A) so that the water factors vary by sector and destination.
Table B1. Weighted Average Water Footprint Factors for U.S. Food Production
Appendix C: ReFED Food Destinations in Order of Priority
Table C1. Food Destination Order of Priority
ORDER OF PRIORITY |
FOOD DESTINATION |
|---|---|
1 |
Prevention |
2 |
Donations |
3 |
Animal Feed |
4 |
Industrial Uses |
5 |
Composting |
6 |
Anaerobic Digestion |
7 |
Not Harvested |
8 |
Land Application |
9 |
Incineration |
10 |
Landfill |
11 |
Sewer |
12 |
Dumping |
Appendix D: Example Calculation of Status Quo GHG Footprint
The solution ‘Donation Storage Handling & Capacity’ sends food to the destination ‘Donations’. This food would have otherwise gone to destinations that are “worse” than donations. Everything from Animal Feed and below is considered “worse” than donations according to ReFED’s Food Destination priorities (See Appendix C).
Table D1. Example calculation of the status quo GHG footprint per ton of Prepared Food surplus in the Florida Family Casual Foodservice sector in 2019:
*Status quo tons sent to each destination was determined from the Food Waste Monitor: refed.com/ insights-engine/food-waste-monitor
**These factors were derived from research by Quantis and can be found in Table A4 in Appendix A. ReFED is assuming that one ton of food prevention or donations results in one less ton of food production, which cancels out upstream emissions.
***These factors can also be found in Table A4 in Appendix A. Negative GHG values indicate a GHG reduction.
Appendix E: Example Calculation of Status Quo Water Footprint
The solution ‘Donation Storage Handling & Capacity’ sends food to the destination ‘Donations’. This food would have otherwise gone to destinations that are “worse” than donations. Everything from Animal Feed and below is considered “worse” than donations according to ReFED’s Food Destination priorities (See Appendix C).
Table E1. Example calculation of the status quo water footprint per ton of Prepared Food surplus in the Florida Family Casual Foodservice sector in 2019:
*Status quo tons sent to each destination was determined from the Food Waste Monitor: refed.com/ insights-engine/food-waste-monitor
**These factors were derived from Water Footprint Network [:ref:
][:ref:] data and can be found in Table B1 in Appendix B. ReFED is assuming that one ton of food prevention or donations results in one less ton of food production, which cancels out upstream water use.***Water footprint factors have not yet been developed for food destinations (e.g., water use required to compost food), so this is not yet accounted for in ReFED’s modeling.
Appendix F: Job Creation Potential
ReFED determined potential job creation by reviewing current employment numbers of food waste solution providers, where organizations have provided data that fell under the buckets of Prevention, Rescue, and Recycling.
Table F. Job Creation Potential
SOLUTION OR CATEGORY |
JOBS PER THOUSAND TONS |
|---|---|
Prevention Solutions |
__job_var__ |
Rescue Solutions |
__job_var__ |
Centralized Composting, Community Composting |
__job_var__ |
Centralized Anaerobic Digestion, Co-digestion at Wastewater Treatment Plants |
__job_var__ |
Consumer Education Campaigns |
__job_var__ |
Prevention solutions assumption (1.5165 jobs / thousand tons) was determined with current employment data of 40+ solution providers at the earlier stages of development. This figure was determined by dividing the sum of jobs with the sum of tons. The data set included solutions at varying levels of maturity which would factor in scaled solutions (which may yield lower jobs per thousand tons). Given the diverse set of business models that can be found in Prevention solutions such as software, hardware, or service-based, the job estimate was made to be from a conservative perspective.
Rescue solutions assumption (3.72 jobs / thousand tons) was determined with current employment data of a group of food banks varying in size (volume of food distributed) from local to larger organizations. The average employee per thousand tons was determined with this dataset. Note this does not include volunteers that they contribute a significant amount of labor to food banks. These organizations require more employees per thousand tons as the work tends to be more manual and processing-related.
Centralized and community composting assumption (1.03 jobs / thousand tons) was determined through the work of the Institute for Local Self-Reliance (ILSR), 2013 [:ref:``]. These reflect jobs created at composting sites and due to compost use. Note: this information is based on the state of Maryland.
Centralized Anaerobic Digestion and Co-digestion at Wastewater Treatment Plants (1.026 jobs / thousand tons) was implied using the data from ReFED’s 2016 Roadmap [:ref:``] of over 1,900 jobs created through anaerobic digestion facilities (excluding potentially hundreds of additional jobs related to composting digestate from these facilities). This information was also based on the ILSR data mentioned above for centralized and community composting.
Consumer Education Campaigns assumption (0.379125 jobs / thousand tons) was determined using the jobs / thousand tons of prevention solutions with an applied discount of 75%. It was the expectation that implementation of Consumer Education Campaigns would not require as many jobs as other Prevention solutions on a per-ton basis.
Appendix G: Implementation Order
Business can choose to manage surplus food in many different ways. For example, changes could be made to prevent the surplus in the first place, it could be donated, or it could be recycled via composting, animal feed, etc. When businesses prevent surplus or donate food, that leaves less surplus available to be recycled. In order to model this relationship between solutions and to avoid double counting the reduction potential of multiple solutions working in tandem, ReFED implemented a “waterfall” approach. In this approach, solutions are ordered and modeled one after the other so that food surplus reduced by the first solution is subtracted from the total surplus available to be addressed by the next solution. Solutions are ordered according to their destination priority as documented in Appendix C (e.g. prevention before donations, donations before recycling), their chronological intervention point within the supply chain, and their net financial benefit. This waterfall ordering only takes place when multiple solutions address a single cause of surplus in a specific sector, because each cause is modeled independently and represents a discrete quantity of surplus. See Table G1 below for specific examples.
Table G1. Solution Implementation Order
Appendix H: Solutions Modeling Assumptions
Overall Approach
To create the ReFED Solutions Database, ReFED, Juniata, and Deloitte first established a list of approximately 80 solutions that contribute to food waste reduction. We then conducted a literature review and outreach to dozens of practitioners, experts, and solution providers to determine diversion rates, costs, and benefits for each solution. Below are the data sources and assumptions used for each of the 42 solutions that we were able to model. Our goal is to improve the information behind this model over time, and we welcome input to that end.
Applicable Sectors and Causes
Various solutions can only be applied to certain portions of surplus food. For instance, a restaurant may have overproduction in its kitchen and plate waste in the front of the house. A donation solution could only be applied to the overproduction, not the plate waste. For each of our solutions, we establish both the applicable sectors (e.g., Foodservice, in this example) and the applicable causes (e.g., Overproduction) and then apply the diversion rate only to the quantities of surplus food estimated for those sectors and causes.
Because distribution happens throughout a food product supply chain, ReFED included food distribution surplus in each relevant sector (e.g. manufactured food products shipped and then rejected by a retail buyer are included as manufacturing sector surplus). The only distribution surplus that is not included in ReFED’s model is food lost during transit and during storage at third-party distribution facilities.
Diversion Rates
Diversion rates were derived from the best sources available and applied only to the sectors and causes as described above. Where diversion rates were provided by a solution provider directly, a 25% “discount” was applied to account for case studies and results typically being selected to demonstrate best results.
“Waterfall” Implementation Order
In many cases, if a business has surplus food, a variety of things could happen to that food. Changes could be made within business operations to prevent surplus or that food could be donated, for instance. In order to avoid double counting the same food, our model implements solutions in order, removing food saved by that solution from the total that the next solution considers. Solutions are ordered considering their position within the EPA food waste hierarchy [:ref:``], their logical implementation order, and their net financial benefit.
It’s important to note that this happens within the applicable causes. For instance, the Imperfect & Surplus Produce Channels solution is applied to the produce that is Left Behind After Harvest (Marketable). Then the Gleaning solution is applied to the amount left over in that Left Behind After Harvest (Marketable) cause category. This would not affect Assisted Distressed Sales, however, since that solution is not applied to the same cause category. Implementation order can be seen in Appendix G above.
Cost and Benefit Assumptions
Our analysis was conducted on a per-ton-diverted basis. All costs and benefits were estimated at that level.
Solutions were considered over a 10-year timeframe. Upfront costs were divided over the tons diverted over that 10 year period and all costs and benefits are considered through a lens of the net present value over those 10 years, with a 4% discount rate. The values provided in the tables below represent the costs/benefits before the NPV has been applied.
To generate estimates of the costs and revenue/savings for various solutions, the following assumptions were made throughout:
Tip Fee Savings - The value of avoided landfill tip fees were derived from the Environmental Research & Education Foundation (EREF) 2022 “Analysis of MSW Landfill Tipping Fees” report [:ref:``]. EREF maintains a database of Municipal Solid Waste (MSW) landfills across the United States. This database was used to draw a sample of active facilities for analysis of MSW landfill (MSWLF) tipping fees. MSWLF tip fee data were compiled by geographic region and basic statistical data were computed. For 2022, the national MSW landfill tip fee average was [tip_fee_var].
Wholesale and retail price assumptions - Retail prices were derived from the 2021 Nielsen IQ retail sales data set [:ref:
]. Wholesale prices were derived from that same data set, assuming gross margins as reported by the U.S. Census Bureau Annual Retail Trade Survey [:ref:]. Importantly, prices for food cost savings for each solution are derived from an average of the product type categories that are considered applicable. So, if a solution only applies to produce and dairy, the food costs for only those two product types would be averaged to generate estimates of food cost savings. The price estimates are as follows:
Table H1. Retail and Wholesale Price Assumptions
FOOD TYPE |
WHOLESALE |
RETAIL |
|---|---|---|
Standard Retail Mix* |
$__price_var__ |
$__price_var__ |
Ready to drink beverages |
$__price_var__ |
$__price_var__ |
Breads & Bakery |
$__price_var__ |
$__price_var__ |
Dairy & Eggs |
$__price_var__ |
$__price_var__ |
Dry Goods |
$__price_var__ |
$__price_var__ |
Fresh Meat & Seafood |
$__price_var__ |
$__price_var__ |
Produce |
$__price_var__ |
$__price_var__ |
Frozen |
$__price_var__ |
$__price_var__ |
Prepared Foods |
$__price_var__ |
$__price_var__ |
Food service |
$__price_var__ |
$__price_var__ |
* Standard Retail Mix of food and beverage products sold at grocery stores in the U.S. according to data from Nielsen IQ [:ref:``].
Data Quality Scores
Data in the field of food waste reduction is challenging. In many cases, only case studies or anecdotal evidence is available, while in others third-party, peer-reviewed academic studies have been performed or many proof points are available. In modeling our solutions, we aimed to get the best data we could, but recognize that significant assumptions and extrapolations are involved. We therefore developed a Data Quality Rubric to rank our sources and how we were using them. Scores are included below for each solution. A full description of the rubric can be found in Appendix I.
Financing
Effective action against food waste requires a smart matching of the correct type of capital with the appropriate opportunity, and in many cases, multiple types of capital are required to fund food waste reduction solutions from conception to adoption. ReFED’s Insights Engine and Roadmap to 2030 calculated the total financing required for each solution across nine sources of capital; allocating the quantified investment required from the Solutions Database to various capital types. ReFED’s intent is that this can galvanize the funding required to fill financing gaps and achieve the benefits highlighted in the Insights Engine.
ReFED’s analysis first starts by acknowledging that there are different capital types - each with varying goals and investment theses. As a result, certain types of capital are more appropriate depending on the financing opportunity and can depend on a variety of factors including, but not limited to growth potential, market size, solution maturity, and business model. The chart below defines the nine sources of capital analyzed and their expected rates of return.
Table H2. Capital Types
SOURCE |
DEFINITION |
RATE OF RETURN |
|---|---|---|
Tax Incentives |
Tax incentives and deductions related to donations. R&D tax credits are not factored in this analysis.
|
__percent_var__%
|
Non-Government Grants |
Philanthropic grants from non-government sources, including high networth individuals, family offices, and foundations.
|
__percent_var__%
|
Impact-First Investments |
Investments that seek some sort of financial return, but are willing to accept more risk or potentially lower returns in pursuit of measurable social or environmental impact. Examples include low- or no-interest loans, loan guarantees, variable payment options, program-related investments (PRIs), etc.
|
__percent_var__%
|
Venture Capital |
A type of financing that investors provide to startup companies and other for-profit businesses that are believed to have long-term, high growth potential. Investors in this asset class have a perceived higher risk as companies are at an earlier stage and therefore require a high rate of return.
|
__percent_var__%
|
Private Equity |
Composed of funds and investors that directly finance private companies. Organizations receiving this type of capital are established organizations or ones requiring growth equity.
|
__percent_var__%
|
Corporate Finance and Spending |
Spending by for-profit corporations with the intent to return the cost of capital. Examples include spending on solutions (through paying solution providers or internally developing capabilities) and corporate acquisitions (M&A). Marketing type spending (non-foundation spending) would be considered part of this category as an operating cost despite not directly leading to market returns. Additionally, ReFED has considered traditional lending (leases, working capital loans) as part of corporate finance and spending.
|
__percent_var__%
|
Government Project Finance |
Direct municipal, state, or federal project financing.
|
__percent_var__%
|
Commercial Project Finance |
Financing provided for projects with the cash flows of the specific project paying down the project loan. This is sourced from for-profit financiers.
|
__percent_var__%
|
Note: there are types of capital that are hybrids or exist outside of the types listed above. Mission Related Investments (MRIs), for instance, would fall under the broad definition of “Impact Investments’’, but require market-rate returns. Therefore, it could arguably be a form of venture capital or its own capital type.
ReFED’s proposed allocation of capital for each solution was determined by analyzing historical funding, stakeholder feedback, and industry knowledge. This exercise provides a rough, directional estimate of the total amount of funding needed for each solution, by funding source, and in aggregate. The proposed allocations are not meant to be prescriptive, as actual financing is highly dependent on funder interest and relative costs of capital. As external market and environmental factors change – a national spotlight on food waste, for example – funding availability may shift to favor more or less expensive forms of financing.
First, in order to estimate allocations of financing across capital types, desk research and analysis was conducted and applied to each solution according to what typical organizational and business model (e.g., for-profit or non-profit) exist in each solution bucket, historical examples of funding, level of maturity for each solution, and if the solution is asset light or requires significant capital expenditure/ infrastructure spending. For example, Manufacturing Byproduct Utilization (Upcycling) is a nascent, yet growing solution often adopted using for-profit business models within large corporate entities or startups in the early stages of maturity, and requiring a large amount of capital expenditure.
According to these factors, solutions were allocated a qualitative weighting of 0-Low, Low, Medium, Medium-High, High, and All for the amount of capital required by each capital type which had corresponding numerical weightings.
Lastly, ReFED sought and received feedback from 15+ capital providers (including foundations, impact investors, venture capitalists, private equity firms, and institutional investors) and food businesses on the proposed weights, methodology, and appropriateness of finance amount by capital type.
The table below represents the results of this exercise, including the resulting recommended financing mix for each solution, and the assumptions underscoring this analysis.
ReFED was particularly interested in the concept of catalytic capital as a way to influence further capital entering the food waste space. According to The MacArthur Foundation, catalytic capital is defined as “investment capital that is patient, risk-tolerant, concessionary, and flexible in ways that differ from conventional investment” and “is an essential tool to bridge capital gaps and achieve breadth and depth of impact, while complementing conventional investing.” ReFED has measured catalytic capital by totaling Non-Government Grants, Government Grants, and Impact-First Investments. Additionally, incubators, accelerators, and challenge platforms that provide funding, as well as seed/angel rounds can be considered catalytic.
Table H3. Financing Breakdown