Increasing investments in One Health to reduce risks of emerging infectious diseases at the source © 2022 International Bank for Reconstruction and Development/The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions: This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2022. “Increasing investments in One Health to reduce risks of emerging infectious diseases at the source.” Washington, DC: World Bank. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org December, 2022 Table of Contents Introduction................................................................... 4 4. What elements should be considered for the development of an investment 1. What is the trend in emergence of framework?..........................................................55 infectious diseases?.............................................. 5 4.1. Mobilizing public resources...................................... 56 1.1. Trends in outbreaks and cases.................................... 5 A. Domestic resource mobilization.................................56 1.2. Limitations and comparability of datasets...........9 B. Multilateral development banks/Official 2. What drives emergence of infectious development assistance................................................ 57 diseases?................................................................ 11 C. Financing One Health and health security............. 58 2.1. Drivers identified from individual spillover 4.2. Public investments to mobilize private events for specific diseases........................................ 12 investment........................................................................ 61 A. Zoonotic influenza viruses..............................................16 4.3. Enabling environment for a B. Coronaviruses......................................................................17 One Health-aligned economy................................... 62 A. Realign expenditures.......................................................63 C. Ebola and Marburg viruses............................................ 18 B. Real sector policies and regulations..........................64 D. Lassa virus............................................................................ 18 C. Financial sector policies and regulations................64 E. Hendra and Nipah virus (Henipaviruses)................. 18 2.2. Mitigation and adaptation for disease References................................................................... 66 emergence............................................................................... 19 Annexes.........................................................................70 3. What are the costs of prevention?...................21 Annex 1: Classification of EID risk............................70 3.1. What is the cost to avoid deforestation?............ 21 Annex 2: Hectares of annual avoided forest A. Analysis of avoided deforestation costs...................21 loss by 2030 if 50 percent of B. Proposed approach and results on the costs of deforestation is avoided.....................................72 avoided deforestation.....................................................23 Annex 3: Hot spot countries by income C. Analysis of restoration costs........................................24 classification......................................................... 75 D. Discussion............................................................................ 27 Annex 4: Cost of biosecurity per farmer 3.2. What is the cost to prevent diseases and total................................................................. 77 through the food systems?....................................... 28 A. Risk framework for viral zoonotic outbreaks..........31 B. Reducing the risk of infections through the food supply chain..............................................................32 C. Estimating costs of on-farm biosecurity.................49 D. External co-benefits of pandemic risk management...................................................................... 50 E. Discussion............................................................................52 Page 1 ACKNOWLEDGMENTS This report was written by a World Bank team co-led by Sulzhan Bali, Garo Batmanian, and Franck Berthe, with contributions from Juan David Alduncin, Lucia Patricia Avila Bedregal, Marco Boggero, Luis Constantino, Felipe F. Dizon, Luis Diego Herrera Garcia, Kenan Karakulah, Catherine Machalaba, Simmy Martin Jain, Lisa Li Xi Lau, Samantha Elizabeth Power, Gayatri Rao Sanku, Sarah Louise Shanks, Hina Khan Sherwani, Shweta Sinha, Yurie Tanimichi Hoberg, Eva Teekens, Aditya Babu Upadhyaula, Alex Vaval Pierre Charles, and Leonardo Viotti. The team benefited from regular interactions with the One Health extended team of the World Bank including Mark E. Cackler, Ana Cristina Canales Gomez, Claire Chase, Gina Cosentino, Stephen Geoffrey Dorey, Sambe Duale, Lisa Farroway, Pierre Gerber, Artavazd Hakobyan, Hikuepi Katjiuongua, Anna Elisabeth Larsen, Sitaramachandra Machiraju, Daniel Mira-Salama, Tamer Samah Rabie, Yoshini Naomi Rupasinghe, Tahira Syed, Dipti Thapa, Mariela Huelden Varas, and Shiyong Wang. The team is grateful for the peer review provided by Priya Basu, Richard Damania, Benoit Bosquet, and Magnus Lindelow. The team is also grateful for the guidance and advice of the External Advisory Panel: Bernice Dahn (former Minister of Health, Republic of Liberia), Victor Dzau (President, National Academy of Medicine), Catherine Geslain-Laneelle (Director, Council of the European Union), Amanda Glassman (Vice-President, Center for Global Development), and Cecilia Mundaca Shah (Director, Global Health, United Nations Foundation). The authors received helpful advice and comments from Gaël Giraud (Georgetown University); William B. Karesh, Noam Ross, and Emma Mendelsohn (EcoHealth Alliance); Ben Oppenheim (Metabiota); and Richard Seifman. In addition, the team is greatly indebted to colleagues from the Quadripartite alliance for their advice and helpful comments during this research: Emily Tagliaro and Chadia Wannous (World Organisation for Animal Health), Katrin Taylor and Katinka DeBalogh (Food and Agriculture Organization of the UN), Margarita Meldon (United Nations Environment Programme), and Leen Meulenbergs and Stephane de la Rocque (World Health Organization). Martien van Nieuwkoop and Juergen Voegele provided strategic guidance and substantive input throughout the preparation of the report, with support from Geeta Sethi, Mark E. Cackler, William Sutton, and Julian Lampietti. Finally, Venkatakrishnan Ramachandran, Pawan Sachdeva, Michelle Rebosio, and Beaulah Noble provided impeccable administrative support for which the team is grateful. The report was funded by the Federal Ministry for Economic Cooperation and Development (BMZ, Germany) as part of the World Bank’s Food Systems 2030 Trust Fund program on One Health. The authors thank Karen Schneider for editing and Jay Groff for publication layout and design. They provided invaluable support for turning a manuscript into a finalized report. Page 2 List of figures Figure 24. Deficit and excesses in animal protein consumption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 Figure 1. Evolution of the number of outbreaks (a) and Figure 25. Co-benefits from less food loss and waste in the reported human cases per year (b), 1980–2013. . . . . . . . . . . . . . 5 chicken supply chain (UK parameters). . . . . . . . . . . . . . . . . . . . . . . 51 Figure 2. Evolution of outbreaks based on zoonotic Figure 26. Financing policies for One Health. . . . . . . . . . . . . . . . 64 versus non-zoonotic infections (a), regions (b), type of infectious agent (c), and vector-borne versus Figure 27. Global trends in EIDs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 non-vector-borne infections (d), 1980–2013. . . . . . . . . . . . . . . . . 6 Figure 28. Risk of zoonotic diseases by country. . . . . . . . . . . . . 71 Figure 3. Evolution of share of total cases based on four classifications, 1980−2013. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 4. Evolution of Intensity of outbreaks, 1980–2013. . . . 8 List of tables Figure 5. Evolution of zoonotic outbreaks and reported Table 1. Comparison between datasets considered. . . . . . . . . 10 cases in %, 1980–2013. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Table 2. Missing or improbable values in datasets. . . . . . . . . . . 10 Figure 6. Intensity of zoonotic outbreaks, 1980–2013. . . . . . . . 9 Table 3. Domains of drivers associated with spillover events. . . 13 Figure 7. Hectares of forest used to estimate the costs of avoided deforestation—intersection between high Table 4. Drivers and consequences of recorded spillover deforestation risk by 2030, biodiversity density, and events in humans for the zoonotic pathogens under the accessibility.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 WHO R&D blueprint and PEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 8. Areas with restoration benefit-cost ratio Table 5. Differences between mitigation and adaptation greater than 3 intersecting with high biodiversity in for infectious disease emergence. . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Assam State, India. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Table 6. Annual cost of halving deforestation - by risk of Figure 9. Areas with restoration benefit-cost ratio zoonotic spillover. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 greater than 3 and high biodiversity in Liberia. . . . . . . . . . . . . . . 27 Table 7. Restoration costs for Assam state, India, for Figure 10. Areas with restoration benefit-cost ratio 13,688 ha of high restoration potential. . . . . . . . . . . . . . . . . . . . . 26 greater than 3 and high biodiversity in Vietnam. . . . . . . . . . . . . 27 Table 8. Restoration costs for Liberia (2017 US$), for Figure 11. The food system in the context of other systems. . 29 97,237 ha of high restoration potential. . . . . . . . . . . . . . . . . . . . . 26 Figure 12. Global poultry and pig population. . . . . . . . . . . . . . . . 30 Table 9. Restoration costs for Vietnam (2017 US$), for 73,375 ha of high restoration potential. . . . . . . . . . . . . . . . . . . . . 26 Figure 13. Number of animals per thousand humans . . . . . . . . 31 Table 10. Summary of chicken production systems. . . . . . . . . 33 Figure 14. Framework for identifying hot spots for risk of zoonotic viral epidemics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Table 11. Summary of pig production systems. . . . . . . . . . . . . . 34 Figure 15. High-risk areas for pathogen transmission Table 12. Burden of zoonotic diseases. . . . . . . . . . . . . . . . . . . . . . 45 between wildlife, livestock, and humans. . . . . . . . . . . . . . . . . . . . 36 Table 13. Impacts of reducing food loss and waste of Figure 16. South-East Asia and Indonesia—very high-risk chicken by 50 percent in processing (slaughtering). . . . . . . . 47 areas for viral species jumping (in red). . . . . . . . . . . . . . . . . . . . . . 36 Table 14. Impacts of reducing food loss and waste of Figure 17. Eastern and Western Africa - very high-risk areas for chicken by 50 percent at all stages of the supply chain. . . . . 47 viral species jumping (in red) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Table 15. Impacts on policy goals of taxes on chicken meat. . 50 Figure 18. Density of poor livestock keepers. . . . . . . . . . . . . . . . 38 Table 16. Annual cost of biosecurity—lower bounds, point Figure 19. Global meat estimates and projections – estimates, and upper bounds—for low- and lower-middle- 1961–2050 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 income countries in hot spots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 20. Protein efficiency of dairy and meat production. . 42 Table 17. One Health financing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Figure 21. Productivity of meat production in poultry Table 18. Total cost of biosecurity—lower bounds, point and pigs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 estimates, and upper bounds—for low- and lower-middle- income countries in hot spots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Figure 22. Losses estimates in the poultry supply chain in the UK. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Table 19. Per farmer cost of biosecurity—lower bounds, point estimates and upper bounds—for low- and lower- Figure 23. Animal protein consumption in relation to middle-income countries in hot spots. . . . . . . . . . . . . . . . . . . . . . . 77 National Recommended Diets thresholds. . . . . . . . . . . . . . . . . . 48 Page 3 Introduction COVID-19 leaves a global reminder that the world is year. And most outbreaks were, and remain, infections vulnerable to pandemic risk. The burden of infectious caused by bacteria and viruses. Researchers have diseases continues to grow, and humanity faces more identified several broad drivers of the emergence of outbreaks, some with the potential to become infectious diseases. For the eight disease groups pandemic. The cost of COVID-19 on the global examined in this technical report, specific practices economy is estimated at trillions of dollars, and future associated with spillover events typically fell into two pandemics will only increase costs on response, for broad sets of drivers: land use changes and food which the world must turn the focus toward prevention. systems. Understanding these factors of disease The One Health approach offers a practical and emergence, and the costs associated, will allow for successful framework to reduce pandemic risks at the development of risk reduction measures at different source. It is the necessary foundation to increase global levels. Therefore, this document describes an approach health security. One Health recognizes that human and to estimate the cost of avoided deforestation and animal health are interdependent and bound to the forest restoration by 2030 for countries at high risk of health of the ecosystems in which they exist. zoonotic spillovers using readily available global datasets and proposes additional analytical steps. This In October 2022, the World Bank published its flagship document also focuses on the costs of reducing the risk report ‘Putting Pandemics Behind Us’ for reducing the of infections through the food supply chain, addressing risk of spillover to humans, domestic animals, and biosecurity. wildlife. This report reflects on current gaps in system capacity, mandates, and resourcing. Ideas for an This report is the backbone of the World Bank Flagship investment framework are proposed to prioritize Report (Putting Pandemics Behind Us) and provides pandemic prevention and take a risk-based approach detailed analysis on key research questions: Is there a focused on spillover interfaces that increase contact trend to accelerated emergence of infectious diseases? between wildlife, livestock, and human populations. What drives emergence of infectious diseases? What is Suggested no-regrets interventions to reduce risk along the cost to avoid deforestation? And what is the cost to food systems pathways are discussed. Practical actions prevent diseases through the food systems? This targeted to national governments, international technical report provides complementary evidence on financial institutions, and the Quadripartite are why and how to mobilize One Health finance and suggested to set the path toward a safer world with increase investments to reduce pandemic risks at the expected co-benefits for sustainable development, source. biodiversity, and climate action that can be generated under a One Health approach. Since 1980, the number of outbreaks per year as well as the number of reported cases have been steadily increasing. Between 1980 and 2012, the number of outbreaks increased at an average of 6.7 percent per Page 4 What is the trend in emergence of infectious diseases? 1.1 TRENDS IN OUTBREAKS AND CASES Since 1980, the number of outbreaks per year has been steadily increasing (Figure 1.a). Between 1980 and 2012, the number of outbreaks increased at an average of 6.7 percent per year. There were 92 outbreaks recorded in 1980, compared to 579 in 2012. There is a 2016, followed by a decline. 1 al. 2014 dataset that is publicly available). Although not seen in our dataset, graphs from Morand (2020) show a stagnation of the yearly outbreaks between 2013 and Simultaneously, we have observed an increase in the number of reported cases per year. Figure 1.b shows the decline in the number of outbreaks between 2012 and evolution of reported cases per year in base 10 2013. The steep decline observed in Figure 1.a could logarithmic scale. The spikes indicate a significant stem from the fact that outbreak data for 2013 span increase of reported cases compared to the previous only until July 2013. This is, however, confirmed by year. Generally, there has been an increase in the size of Morand (2020), who also uses data from Global these spikes since 1995. Infectious Diseases and Epidemiology Network (GIDEON)from 1980 to 2019, whereas our analysis The majority of outbreaks were, and remain, non- looks at the 1980–2013 period (based on the Smith et vector-borne diseases (Figure 2.d). Infections caused FIGURE 1. Evolution of the number of outbreaks (a) and reported human cases per year (b), 1980–2013 Source: World Bank developed graphs. Data source: Smith et al. 2014 Note: (a) shows the evolution of the number of outbreaks and (b) shows the evolution of reported cases in base 10 log scale. Page 5 by bacteria and viruses were the main source of outbreaks (29.5 percent). After seeing a decrease in the outbreaks. However, viruses rather than bacteria share of outbreaks between 1980 and 1999, Africa has account for an increasing share of outbreaks (Figure experienced an increase from less than 8 percent in 2.c). Zoonoses (diseases that originate from animals) 2000 to more than 10 percent in 2010. South account for 56 percent of the outbreaks, with the share America’s share of outbreaks slightly varied at around 9 increasing between 1980 and 2009 and slightly percent before 2000 followed by a slight decline (at 8 decreasing afterward (Figure 2.a). Every region in the percent) until 2004. Its share of outbreaks increased world has experienced an increase in the number of sharply after 2010 to reach nearly 13 percent of total outbreaks (Figure 2.b). From 1980 to 1984, North outbreaks. America accounted for the highest share of outbreaks (35 percent) compared to 25 percent and 20 percent The distribution of outbreaks based on the for Europe and Asia and Pacific, respectively. From 1985 classifications in Figure 2 is not completely parallel to to 1989, Europe accounted for the highest share of the distribution of reported cases using those same classifications, the only exception being the FIGURE 2. Evolution of outbreaks based on zoonotic versus human-specific infections (a), regions (b), type of infectious agent (c), and vector-borne versus non-vector- borne infections (d), 1980–2013 Source: World Bank developed graphs. Data source: Smith et al.2014 Note: The observations are grouped by demi-decade, such that 1980 comprises outbreaks from 1980 until 1984, 1985 comprises outbreaks from 1985 until 1989, and so on. The share of outbreaks caused by zoonoses is displayed in percentages in (a). Page 6 classification by vector-borne versus non-vector-borne Figure 4 plots the average number of cases per diseases. Although zoonotic outbreaks have outpaced outbreak per year in thousands. The size and color of non-zoonotic outbreaks, their share of the total cases the circle corresponds to the number of outbreaks reported shows an opposite trend (Figure 3.a). The observed in a given year. The bigger the circle, the more distribution of cases based on region varies slightly the outbreaks observed in that year. Figure 4 shows from that of outbreaks. For example, compared to other that despite the increase in outbreaks in the last 30 regions, the Asia and Pacific region accounts for the years, their intensity (measured by the average number highest share of reported cases for the whole period. of cases per outbreak) has been declining on average. Furthermore, Africa accounts for a greater share of However, looking at the whole period hides differences cases than it does for outbreaks, during the period, in trends observed. If we separate the period in two, whereas for North America the opposite is observed 1980–1999 and 2000–2013, we see that in the first (Figure 3.b). Lastly, although viruses account for a period, intensity per outbreak is indeed decreasing slightly lower proportion of outbreaks than bacteria, (shown by the blue regression line), whereas in the they account for the vast majority of reported cases second period, intensity is sharply increasing (red (except for the 1990–1994 period) (Figure 3.c). regression line). FIGURE 3. Evolution of share of total cases based on four classifications, 1980−2013 Source: World Bank developed graphs. Data source: Smith et al.2014 Note: (a) Host type, (b) Region, (c) Pathogen Taxonomy, (d) Transmission type. The observations are grouped by demi-decade, such that 1980 comprises outbreaks from 1980 until 1984, 1985 comprises outbreaks from 1985 until 1989, and so on. Page 7 FIGURE 4. Evolution of Intensity of outbreaks, 1980–2013 World Bank developed graphs. Data source: Smith et al.2014 Note: The two lines (blue and red) are linear regression lines for the periods 1980–1999 and 2000– 2013, respectively. The color of the circles varies depending on the number of outbreaks: circles are dark green if the number of outbreaks is lower than 200, light green if the number of outbreaks is lower than 400, and orange if the number of outbreaks is lower than 600. The standard errors are represented by the gray area around the regression lines. Figure 5 focuses on zoonoses only and looks at progressively increasing but still only makes up about outbreaks (a) and reported cases (b) as a share of 40 percent of total outbreaks from 2010 to 2013. In total zoonotic outbreaks and reported cases between terms of total zoonotic reported cases, it is interesting 1980 and 2013. Comparing viral zoonoses and to note that although viral zoonoses outbreaks are the nonviral zoonoses (meaning bacterial, parasitic, and minority, they represent a higher share of reported fungal zoonoses), viral zoonoses represent a minority cases and even the majority of cases from 2005 to of total zoonotic outbreaks (<30 percent between 2013. Intuitively, it means that viral zoonotic 1980 and 2004). Since 2005, that share has been outbreaks tend to spread more easily than other FIGURE 5. Evolution of zoonotic outbreaks and reported cases in %, 1980–2013 Note: The observations are grouped by demi-decade, such that 1980 comprises outbreaks from 1980 until 1984, 1985 comprises outbreaks from 1985 until 1989, and so on. Page 8 FIGURE 6. Intensity of zoonotic outbreaks, 1980–2013 World Bank developed graphs. Data source: Smith et al.2014 zoonotic outbreaks and that the spread has increased likely that numbers for Africa and South America are in the past 20 years. underestimated. Another important source of bias comes from the distribution of the population. It is Figure 6, which looks at the intensity of zoonotic likely that regions experiencing the highest share of outbreaks, confirms the statement that viral zoonotic cases are also those with the largest populations. This outbreaks tend to spread more easily than other could explain why the Asia and Pacific region accounts zoonotic outbreaks and that there has been an increase for the highest share of cases for the studied period. in the spread in the past 20 years. On average, viral zoonotic outbreaks have more cases than other Several datasets on infectious disease exist and have zoonotic outbreaks (light blue regression line is above supported analyses of trends over time. This study the light red one). Furthermore, although the average considered the datasets from Jones et al. (2008), number of cases per outbreak decreased for both viral Smith et al. (2014), Marani et al. (2021), Stephens et al. and nonviral zoonotic outbreaks between 1980 and (2021), and Bernstein et al. (2022). Table 1 shows a 1999, this is not the case after 2000 for viral zoonoses, comparison of the considered datasets based on some which have seen a sharp increase in the average number variables of interest. The Marani, Stephens, and of cases per outbreak. Bernstein datasets consider a longer and more recent period, which gives a broader and more recent window 1.2 LIMITATIONS AND COMPARABILITY of observation of the evolution of infectious diseases. OF DATASETS The Jones, Marani, and Bernstein datasets include information on deaths, which is an important limitation As with all data on infectious diseases, a potential in the Smith and Stephens datasets. Jones and limitation is that some regions are overrepresented in Stephens have included extensive information on the global incidence of infectious diseases because of drivers of outbreaks and more variables than Marani, better reporting and monitoring and more frequent Smith, and Bernstein. testing. In that sense, when looking at the baseline and evolution of outbreaks and deaths based on region, it is Page 9 TABLE 1. Comparison between datasets considered Jones et al. Marani et al. Smith et al. Stephens et Bernstein et (2008) (2021) (2014) al. (2021) al. (2022)a Number of observations 335 539 12,102 300 Not stated Number of variables 106 10 8 64 7 Time frame 1940–2004 1 1500–2018 1980–2013 1974–2017 1918–2020 Source Various Various GIDEON GIDEON Various Information on deaths Yes (incomplete) Yes (incomplete) No No Yes Information on cases Yes No Yes Yes No Information on drivers Yes Yes No Yes No Information on location Yes Yes Yes Yes No Information on host type Yes No Yes Yes Yes Information on pathogen type Yes No Yes Yes (partial) — Note: a. The information displayed here is based on Table S4, available in the supplementary material provided by the authors. The dataset used for their analysis was not made public and therefore we largely rely on the content of the paper. While each dataset has its own limitations, the Smith 200 randomly and then the 100 largest zoonotic dataset offered several strengths. For example, Smith’s outbreaks in the dataset. Bernstein focuses on novel data come from a single reliable source compared to viral zoonoses that have caused at least 10 deaths. various sources (typically compiled through literature Lastly, the Smith dataset has complete information for reviews) for all but Stephens’ data, which also use 12,047 outbreaks (99.5 percent); the remaining 55 GIDEON data. Furthermore, Smith considers every outbreaks have missing data for one or more variables. outbreak between 1980 and 2013, while the others This level of completeness is not reached by Marani and consider subsets of episodes. Jones focuses on EID Jones especially for information on deaths and cases. ‘events,’ which are defined as “the first temporal For example, 30.6 percent of observations in the Jones emergence of a pathogen in a human population which dataset have missing values for the number of deaths was related to the increase in distribution, increase in and 28.8 percent in the Marani dataset. incidence or increase in virulence or other factor which led to that pathogen being classed as an emerging TABLE 2. Missing or improbable values in datasets disease,” thereby not including subsequent spillover events. Marani focuses on EID epidemics, which are Number of Number of infected deaths defined based on three criteria: no overlap in different Jones et al. (2008) 44 107 epidemic episodes of the same disease, no ongoing (12.6%) (30.6%) epidemics (such as AIDS or Malaria), and no epidemics Marani et al. (2021) Not available 155 ended by the introduction of vaccines. Those (28.8%) differences in the criteria of selection make comparing Stephens et al. (2021) Not available Not available and merging these datasets difficult. Stephens Bernstein et al. (2022) Not available Nonea excluded protozoal pathogens (such as malaria) and viruses and bacteria in which transmission is unknown Note: a. Numbers could not be verified (dataset is not publicly available). in the time range of outbreaks; based on this, a subset of 300 outbreaks (of a total of 8,000) were selected: 1 The original Jones dataset stopped in 2004. Page 10 What drives emergence of infectious diseases? The causal pathways leading to initial spillover events, spreading to become epidemics, and in some cases leading to pandemics, are most often complex, involving a mix of factors that shape risk and increase vulnerability. Researchers and experts have used a variety of terms to describe factors that may be associated (whether causal or correlated) with with the conditions that create risk (such as change in food systems as a response to increased consumption of animal products). 2 Researchers have identified several broad domains of drivers of the emergence of infectious diseases, which have been continuously refined over the years. In 1992, emergence and spillover events, such as risk factors, a report by the Institute of Medicine (IOM) (now the US determinants, and drivers. In this report, we will use the National Academies), Emerging Infections: Microbial term ‘drivers’. Drivers may not be mutually exclusive and Threats to Health in the United States, identified six may co-occur (Semenza et al. 2016), potentially with a key drivers (IOM et al. 1992). In a subsequent IOM greater or lesser role depending on the specific context. report, this initial list was expanded to 13 drivers, Relevant drivers may range from practices linked to corresponding to a greater understanding of the nature transmission pathways (for example, aerosol inhalation of microbial threats (Box 1) (IOM et al. 2003). There, through animal butchering) to broad trends associated the drivers encompass four main domains: (a) genetic and biological; (b) physical environmental; (c) ecological; and (d) social, political, and economic. The 2003 report recognized that the list of drivers would BOX 1. Drivers for emergence likely expand in the future, based on the growing understanding of the complexity of infectious disease emergence. • Microbial adaptation and change • Human susceptibility to infection This list of drivers of emergence has been the basis for • Climate and weather subsequent analyses of EID events. A 2008 analysis of the first reported emergence event for zoonotic, vector- • Changing ecosystems borne, and drug-resistant infections occurring between • Human demographics and behavior 1940 and 2004 (n = 335 EID events) also examined their • Economic development and land use driver (Jones et al. 2008). The authors reclassified • International travel and commerce several drivers from the 2003 IOM report, adding new • Technology and industry categories of drivers, including ‘Agricultural industry • Breakdown of public health measures changes’, ‘Medical industry changes’, ‘Food industry • Poverty and social inequality changes’. ‘Land use changes,’ and ‘Bushmeat’. In a • War and famine subsequent analysis of zoonotic emergence events (n = • Lack of political will 183), Loh et al. (2015) analyzed the number of events • Intent to harm attributed to each primary driver, finding that land use change, agriculture industry change, and international Source: IOM et al. 2003. travel and commerce were the top three primary drivers. Page 11 Additional analyses have also examined predictors of Organization (FAO) of the United Nations, the emergence and spillover events based on previous Program on Emerging Diseases (ProMED), and outbreak occurrences, including those linked to spatial national public health agencies. Available information and temporal environmental and demographic trends. varied by disease and event. Researchers obtained For example, after adjusting for reporting effort additional event information from scientific literature, (primarily underreporting by some national authorities), including case reports and review papers, to identify risk of EID event occurrence from wildlife was associated spillover drivers. This analysis was focused on with forested tropical regions experiencing land use spillover factors leading to the index case (the first changes and high mammal species richness (Allen et al. known case in an outbreak); factors associated with 2017). Numerous studies link increases in zoonotic and further spread in human populations were not vector-borne disease occurrence with forest ecosystem examined. Events without definitive or high- change, though this may involve multiple types of such confidence attribution to specific drivers (for example, change. A recent study found that increases in outbreaks based on ecological and epidemiological investigation) of zoonotic and vector-borne diseases from 1990 to were recorded as inconclusive. Events with spread 2016 were associated with deforestation in tropical linked to prior outbreaks (signifying spread or areas as well as reforestation in temperate countries imported cases) were grouped to the origin event. (Morand and Lajaunie 2021). Events are usually counted by public health agencies as unique based on their defined outbreak period, In addition to the first reported emergence event of a which accounts for the typical incubation period for a zoonotic disease, important information can be gained given disease. Genomic sequencing has allowed for from subsequent spillover events (sometimes called increased precision regarding whether events are reemergence events). Understanding the factors likely to stem from new spillover occurrences or be leading to spillover events, including both proximal risk linked to prior outbreaks (for example, from factors and broad, large-scale changes, is important for unrecognized continuing low-level transmission or developing prevention strategies. Epidemiological and persistent infection in a survivor). In some cases, ecological investigations have shed light on specific multiple spillover events were suspected during a practices and conditions associated with individual given outbreak, and it is also probable that some events that can guide risk reduction. small-scale outbreaks went undetected. Thus, the actual number of spillover events may be higher. The Pandemic Emergency Financing Facility (PEF) and the World Health Organization (WHO) Research and Development (R&D) Blueprint created a database of 2.1 DRIVERS IDENTIFIED FROM outbreak events for prioritized zoonotic-origin INDIVIDUAL SPILLOVER EVENTS diseases and pathogen groups (basing their selection FOR SPECIFIC DISEASES on “severe emerging diseases for which there are Spillover events spark individual cases and may result in insufficient or no existing medical countermeasures or onward spread in human populations, in some cases pipelines to produce them”). They did not examine leading to major epidemics and pandemics. For the diseases with vector-borne origins or transmission eight disease groups examined in this analysis, specific (for example, Rift Valley fever, Crimean-Congo practices associated with spillover events typically fell hemorrhagic fever, and Zika virus). Events occurring into two broad sets of drivers: land use changes and through March 2022 were counted and defined as food systems (Table 3). Specific practices leading to those causing human illness. Event information was spillover facilitated close direct or indirect contact with sourced from publicly available outbreak records by animals or infectious animal materials and were linked the US Centers for Disease Control and Prevention to a range of purposes and sectors (for example, food (CDC), the WHO, the Food and Agriculture Page 12 TABLE 3. Domains of drivers associated with spillover events Drivers domain Land use change Food systems Other Specific drivers • Cave or forest encroachment • Livestock production practices • Handling animal/carcass • Deforestation • Livestock trade biosafety practices (for example, disease investigation) • Mining practices • Live animal markets • Laboratory biosafety practices • Tourism practices • Orchards • Urbanization and dwelling • Wildlife hunting or butchering • Consumption of raw date palm quality • Wildlife trade juice • Industry practices • Exposure or consumption of infected meat acquisition or production, food consumption, tourism, change, with effects on bat-roosting behavior that and mining). In some cases, multiple practices linked to facilitates viral spillover. These include distributed spillover were identified for a given event; therefore, the populations of Pteropus medius bats in areas with high number of specific practices can exceed the number of human population density and feeding on cultivated spillover events (Table 4). These provide a sense of the food resources (McKee et al. 2021). Recent context for spillover, including how risk is shaped and deforestation (for example, within the past two years) how risk reduction measures can be targeted, and are has been found to precede some outbreaks of Ebola examined in further detail in the following section. A virus in rainforest biomes (Olivero et al. 2017). major limitation of available data is that spillover practices may be attributed for an outbreak event, but A. Zoonotic influenza viruses the total number of primary spillover cases (for example, multiple spillover events in a given outbreak The genetic composition of influenza viruses makes period) are not always quantified. Overall case and them prone to evolutionary changes (called death counts include primary cases from spillover and ‘reassortments’ of their gene segments), which can lead onward spread (Table 4). In some outbreak events, the to more pathogenic strains. Zoonotic spillover to spillover origin was inconclusive, totaling 42 events of humans has primarily been associated with Ebola, Marburg, and Nipah viruses alone. transmission from livestock (poultry and pigs). While Table 4 examines only direct human disease and Zoonotic spillover has been recorded for several strains mortality impacts from each disease, indirect costs can of low and highly pathogenic avian influenza viruses. be substantial, as seen from the wide-ranging The majority of human cases are associated with highly consequences of the severe acute respiratory syndrome pathogenic H5N1 and H7N9 circulation in poultry, with (SARS) pandemic and West African Ebola epidemic. transmission to humans largely via poultry rearing and live bird markets. The mixing of wild birds and domestic In addition to the factors linked to spillover that have poultry, which in some regions has increased with been identified in the epidemiological and ecological expansion or intensification of poultry production investigations of individual outbreak events, several systems, including around areas frequented by papers have analyzed environmental factors across waterfowl, has created opportunities for new, highly multiple diseases and spillover events. For example, in pathogenic avian influenza strains. These events point Bangladesh, an association has been detected between to the role of livestock production intensification in spillover events and several hundred years of land use creating conditions for pathogen amplification and Page 13 TABLE 4: Drivers and consequences of recorded spillover events in humans for the zoonotic pathogens under the WHO R&D Blueprint and Pandemic Emergency Financing Facility. Compiled March 2022. Based on publicly-available event attribution and case information. A single event may be linked to multiple factors (captured as specific practices and environmental trends). Methodology: Because of variation in reporting format over time, including reporting of total cases in a given epidemic period or year, the number of unique spillover events may be higher. Case estimates vary slightly by source, likely reflecting varying case criteria (e.g., suspected vs. laboratory-confirmed infections). Number of spillovers reflect rough estimates; multiple spillover events may potentially be linked to one defined outbreak event in a population. KEY: SPILLOVER EVENTS # Ebola viruses 32 YEAR FIRST DETECTED Specific Practices Other Relevant Environmental 1976 (Number of Spillover Events) b Trends Reported UNIQUE SPILLOVER EVENTSa 32 outbreak events Wildlife hunting or butchering (9); Recent deforestation RECORDED CASES Laboratory biosafety practices (3); 34,762 Handling animal/carcass biosafety (1); RECORDED DEATHS 15,343 Forest encroachment (1); Inconclusive (18) Hendra virus 5 FIRST YEAR DETECTED Specific Practices Other Relevant Environmental 1994 (Number of Spillover Events)b Trends Reported UNIQUE SPILLOVER EVENTSa 5 outbreak events Horse industry practices (5); El Niño Southern Oscillation (ENSO) event RECORDED CASES Handling animal/carcass biosafety (5) 7 RECORDED DEATHS: 3 Page 14 Table 4 continued KEY: SPILLOVER EVENTS # Lassa virus NON-SPECIFIC FIRST YEAR DETECTED Specific Practices Other Relevant Environmental 1969 (Number of Spillover Events)b Trends Reported UNIQUE SPILLOVER EVENTSa Millions of spilloversc Urbanization and dwelling quality; Rodent reservoir more abundant near crop RECORDED CASES Wildlife hunting or butchering production than in forested Unknown (>100,000 cases/year) areas RECORDED DEATHS Unknown (~5,000 deaths/year) Marburg virus 14 YEAR FIRST DETECTED: Specific Practices 1967 (Number of Spillover Events)b UNIQUE SPILLOVER EVENTSa 14 outbreak events Cave encroachment (5); RECORDED CASES Tourism practices (4); 471 Laboratory biosafety practices (2); RECORDED DEATHS 377 Mining practices (2); Wildlife trade (1); Wildlife hunting or slaughter (1); Inconclusive (5) Middle East Respiratory Syndrome coronavirus NON-SPECIFIC YEAR FIRST DETECTED: Specific Practices 2012 (Number of Spillover Events)b UNIQUE SPILLOVER EVENTSa Hundreds of spilloversc Livestock production practices RECORDED CASES 2,585 RECORDED DEATHS 931 Page 15 Table 4 continued KEY: SPILLOVER EVENTS # Nipah virus 29 YEAR FIRST DETECTED: Specific Practices Other Relevant Environmental 1998 (Number of Spillover Events)b Trends Reported UNIQUE SPILLOVER EVENTSa Consumption of raw date palm juice (7); Forest fires; 29 outbreak events; Hundreds of spilloversc Slaughter, exposure or consumption El Niño Southern RECORDED CASES of infected meat (2); Oscillation-induced 707 drought; Livestock production practices (1); RECORDED DEATHS Bat-pig mixing (poor 412 Deforestation (1); biosecurity); Orchard practices (1); Bat-human proximity linked to fragmented bat habitats Livestock trade (1); Inconclusive (19) SARS 1 YEAR FIRST DETECTED: Specific Practices 2003 (Number of Spillover Events)b UNIQUE SPILLOVER EVENTSa 1 outbreak event Wildlife trade RECORDED CASES 8,096 RECORDED DEATHS 774 Zoonotic influenza virus NON-SPECIFIC YEAR FIRST DETECTED: Specific Practices Other Relevant Environmental 1958 (Number of Spillover Events)b Trends Reported UNIQUE SPILLOVER EVENTSa Livestock production practices Wild bird-livestock mixing Hundreds of thousands of (mainly linked to poultry or pig rearing (poor biosecurity and spilloversc and live animal markets) poultry rearing in/around RECORDED CASES wetland or flyway areas) >3,150 plus >60 million in 2009 H1N1 pandemic RECORDED DEATHS >1,120 plus >150,000 in 2009 H1N1 pandemic Page 16 potential spread to humans and to the importance of B. Coronaviruses biosecurity in animal production systems to minimize wildlife-livestock interactions. Preceding the emergence of COVID-19, coronaviruses have been associated with two other major emergence The 2009 H1N1 pandemic, caused by a variant unique events in humans: the SARS pandemic and outbreaks of from H1N1 seasonal influenza viruses, was found to be a MERS. quadruple reassortment of genes previously individually detected in pigs. While the precise conditions leading The emergence of the SARS-Coronavirus (SARS-CoV) to the evolution of this variant and its spillover to in 2002 was initially linked to wildlife trade based on humans have not been determined, the variant is evidence of infection in those with occupational ties to considered of zoonotic origin, with pigs likely infected wildlife transportation, markets, and slaughterhouses from several sources (birds, humans, and other pigs). (for example, WHO 2003; Xu et al. 2004). Subsequent Because pigs are susceptible to infection from multiple sampling and screening efforts over more than a sources, they likely acted as a mixing vessel for this new decade have indicated that bats were likely the source variant and its onward spillover to humans (CDC of SARS-CoV, potentially with direct spread to humans 2009). Occasional human cases of infection with H3 or via an intermediate species (Hu et al. 2017). Markets and other H1 swine influenza viruses have also been selling wildlife likely played a role either in spillover or detected, with spillover documented in pig production amplification, which allowed for further spread. settings (Ma, Kahn, and Richt 2009). Proximity to wildlife, associated with encroachment BOX 2. Ebola virus In the context of the Ebola crisis in West Africa, a review of scientific literature examined the reported drivers of Ebola virus spillover events and their connections. The analysis found that 40 drivers played a role (Grotto and Ricci 2015). Of these, hunting, deforestation/forest fragmentation, and demographic changes of wildlife were found to be the central drivers. When grouped as social, technological, economic, environmental, and political (STEEP), the most common drivers were social, economic, and environmental in nature. The analysis also emphasized the importance of context, finding some differences in the drivers for transmission from the suspected reservoir (for example, fruit bats) compared to those from intermediate host species (such as great apes), likely reflecting how people interact with different species and the resulting opportunities for spillover. In addition to identifying reported drivers, the study examined the connections between them, finding 142 links. Among the most frequent links were deforestation/forest fragmentation leading to ecosystem changes, livelihood resilience leading to hunting, and hunting leading to butchering and preparing wildlife (for example, for food). Such interactions indicate that often multiple drivers contribute to the conditions for spillover, and therefore a level of complexity, while also reinforcing that some drivers may be more modifiable to reduce spillover risk. Page 17 TABLE 5. Differences between mitigation and adaptation for infectious disease emergence Mitigation Adaptation Prevention Preparedness and resilience Examples Examples • Risk reduction of spillovers to avoid new emergence events • Early detection of outbreaks • Animal and environmental health services for detection • Health care and public health services to effectively treat and containment of pandemic threats in animals patients and contain spread in human populations • Alternative livelihoods and safer practices • Coping and safety nets to restore livelihoods and overcome other societal disruptions Note: As with climate action, efforts should include both mitigation and adaptation to reduce the frequency and impact of disease events. into wildlife habitat, and wildlife trade practices are precede human cases in some epidemics, suggesting a implicated as the likely drivers. sentinel role (Leroy et al. 2004). See Box 2 for case study. Analyses of the locations of index cases suggest The first cases of MERS were detected in 2012. Although an association with recent deforestation and forest the virus (MERS-CoV) may have spilled over from fragmentation (Olivero et al. 2017; Rulli et al. 2017). another species at some point, dromedary camels are considered the major reservoir host for MERS-CoV and a Human infections with Marburg virus, a related virus in source for infection in humans (WHO 2019). Onward the filovirus family, have mainly been linked to cave human-human spread has primarily been documented in tourism and mining, through contact with infected bats. household and health care settings. The majority of The virus was first detected when human cases cases have occurred in the Middle East, primarily in Saudi occurred following the handling and slaughter of Arabia. Travel-associated cases have also occurred, with African Green Monkeys imported for laboratory a total of 27 countries reporting a MERS case as of 2022. activities in Germany and the former Yugoslavia MERS-CoV does not appear to cause illness in adult (Brauburger et al. 2012). camels. Limited research has indicated that other domestic animals can be infected (OIE 2019), though D. Lassa virus whether they have the potential to play a role in maintaining viral circulation or spread to humans is not Lassa virus is considered endemic in Nigeria, Guinea, known. Livestock production, specifically via camel Liberia, and Sierra Leone, although it has been detected rearing, camel slaughter, or consumption of infectious more widely in West Africa. Spillover to humans is camel products, is considered the source of spillover associated with infected urine and feces from the events into human populations (Hui et al. 2018). rodent reservoir, Mastomys natalensis, typically via inhalation in homes, ingestion of contaminated food, C. Ebola and Marburg viruses and rodent hunting and consumption practices. Human-human transmission is generally uncommon, Out of six known Ebola virus species, four have been though community- and health care-associated spread found to affect humans (Zaire, Bundibugyo, Sudan, and does occur. The local burden can be significant; Taï Forest). Zoonotic transmission occurs through approximately 10–16 percent of people hospitalized in contact with infected animals, including fruit bats, parts of Liberia and Sierra Leone each year have Lassa chimpanzees, gorillas, monkeys, forest antelopes, and fever (CDC 2019). The rodent reservoir is typically porcupines (WHO 2021). Wildlife deaths were found to Page 18 found near agricultural fields and in residential areas, Hendra virus, a related virus, was first detected in 1994, with much lower densities in forest areas (Gibb et al. with human cases linked to contact with infected 2017). Infections usually occur during the dry season, horses in Australia. Following infection by fruit bats, followed by reproduction of the reservoir in the rainy horses amplify the virus, transmitting onward to other season; thus, higher temperatures and increased rainfall horses and humans. Although human cases remain rare, from climate change are expected to expand the range at least 50 bat-horse transmission events have been of the reservoir in some areas (Gibb et al. 2017). detected (Kessler et al. 2018). Contact between fruit bats and horses is thought to have increased during El E. Hendra and Nipah virus (Henipaviruses) Niño-related drought, among other factors affecting bat food availability. The range of some Pteropid bats The first recorded Nipah virus cases involved spillover associated with infection appears to be expanding with events from pigs to humans in Malaysia, following climate change (Yuen et al. 2021). infection and amplification in pigs through indirect contact with infected flying fox (Pteropid) bats. While 2.2 MITIGATION AND ADAPTATION FOR bats were present in the outbreak area before the DISEASE EMERGENCE event, environmental conditions are thought to have played a role in their abundance and availability as food, Understanding the proximal and broad factors in with forest fires in Southeast Asia and the 1997–1998 disease emergence allows for development of risk ENSO event potentially driving bats to feed in orchard reduction measures at different levels. For example, in areas where pig farming was also present (Chua, Chua, response to the Nipah virus crisis, the Malaysian and Wang 2002). Following importation of infected government designated areas safe for pig farming, pigs from Malaysia, human cases were also reported in encouraging alternative livelihoods in locations outside slaughterhouse workers in Singapore. Human cases of these areas. Guidelines were developed to promote have been reported in the Philippines, linked to the distancing between fruit orchards and pig-rearing areas slaughter, exposure, or consumption of infected horse on farms to prevent pigs from coming into contact with meat in 2014 (Skowron et al. 2022). Since 2001, cases bat-contaminated fruit. Several practical measures have been reported in Bangladesh and India; in have also been taken to reduce the risk of Marburg virus Bangladesh, these have been linked to consumption of spillover, ranging from mining site closures to safe infected raw date palm juice, which can be tourism practices in areas where spillover has contaminated by fruit bat urine, feces, or saliva during previously occurred. In Queen Elizabeth National Park’s the juice harvesting process. Prior landscape-level Maramagambo Forest, a glass enclosure was installed changes in forest cover are thought to affect the near a bat cave to allow tourists to view bat activity current distribution and feeding patterns of P. medius while avoiding direct exposure to potentially infectious bats in the country, which facilitates proximity to materials (Sun 2018). The safe viewing area provides an human populations (McKee et al. 2021). Although added benefit for conservation by reducing human cases have been reported in only five countries, anthropogenic disturbance of a sensitive wildlife Nipah virus has been found in several additional habitat while protecting the vital ecosystem services countries in Southeast Asia (Wacharapluesadee et al. that bats provide to the surrounding area. 2021). Page 19 Page 20 What are the costs of prevention? 3.1 WHAT IS THE COST TO AVOID DEFORESTATION? Deforestation, forest degradation, and forest fragmentation are considered primary drivers of EIDs linked to wildlife due to increased human-wildlife interactions and increased populations of species that 3 recently, Bernstein et al. (2022) (who build on Dobson et al. 2020) estimate the global costs of avoided deforestation. The annual cost of reducing deforestation by half by 2030 in the most critical parts of the tropics (for example, where the risk of spillover is higher) is estimated between US$1.5 billion and US$9.6 billion (in 2020 US$). These estimates provide host pathogen which could spill over to humans. a baseline to compare our own set of cost estimates. As Avoiding deforestation (and the many aspects of described below, we use a different set of data and ecosystem change and encroachment associated with assumptions to estimate the costs of avoided it) and restoring degraded forests are part of primary deforestation and propose a methodology to estimate prevention of disease emergence. By targeting these the costs of restoration. actions, policy makers can reduce the likelihood of pathogen spillover from wildlife (from wildlife to people A. Analysis of avoided deforestation costs or from wildlife to domestic animals to people) and reduce pandemic risk. High-resolution global forest data and deforestation risk models This document describes an approach to estimate the annual cost of avoided deforestation and forest Data on forest area and forest loss are the starting restoration by 2030 for countries at high risk of point for calculating the cost of avoided deforestation. zoonotic spillovers using readily available global We estimate avoided deforestation costs by using datasets and proposes additional analytical steps. For global high-resolution data, including data from Hansen this exercise, we used data from the World Bank’s et al. (2013) and Hewson et al. (2019). Changing the Wealth of Nations 2021, high-resolution Hansen et al. (2013) provide information on global forest cover data from Hansen et al. (2013), high- forest loss and gain with accuracy, consistency, and resolution deforestation risk data from Hewson et al. spatial and temporal resolution previously available (2019), biodiversity spatial layers from Jenkins, Pimm, only for isolated places and times. Hansen et al. (2013) and Joppa (2013) and Pimm et al. (2014) based on present an annual map of the world’s forests at a IUCN and Birdlife International data, and accessibility resolution the size of a baseball diamond. This dataset data from Nelson et al. (2019). For forest restoration, does not distinguish natural forests from plantations or we used results from se.plan, a global spatial tool that other tree cover. Moreover, its approach to estimate assesses the suitability of restoration efforts and forest gain is not consistent with the approach to estimates its benefits and associated costs. calculate forest loss; thus, a net forest loss is not Few studies have proposed a methodology to estimate estimated. the costs of actions to prevent zoonotic spillovers in Recent studies provide datasets and projections of the land use sector. Dobson et al. (2020) and, more hectares at risk of deforestation at the global level. Page 21 Hewson et al. (2019) used spatially explicit, globally change on future crop yield growth rates based on consistent variables and global historical tree cover loss country-specific crop yield growth rates estimated at from Hansen et al. (2013) to analyze how global- and the grid-cell level, accounting for the impacts of future regional-scale variables contributed to historical tree changes in precipitation, temperature, and degradation cover loss and to model future risks of tree cover loss, (driven by salinization, unsustainable irrigation, and based on a business-as-usual scenario. The outputs of erosion). Future crop production is based on these models are available in 1 km resolution spatial projections of the yields of 10 major crops, which layers. The data give a probability score [0,1] of losing together comprise 83 percent of calories produced on tree cover until 2029. The study uses two different cropland. Fixed growth rates are used for pastureland models, one global and the other a combination of six production. regional models. These data have been used for different applications such as Koh et al. (2021) on Cropland rents are estimated per crop product as carbon prospecting in tropical forests. production multiplied by the unit price multiplied by the rental rate. Data of crop and livestock production Busch and Engelman (2017), who provide the come from FAO statistics. For crops, the rental rates are underlying deforestation data for Dobson et al. (2020) constant over time and crop products are region also use Hansen et al. (2013) data between 2001 and specific (Evenson and Fuglie 2010). Pastureland rents 2012, along with information on topography, are also estimated per livestock product as production accessibility, protected status, potential agricultural multiplied by the unit price multiplied by the rental rate. revenue, and other variables, to project tropical However, rents from livestock products are different for deforestation from 2016 to 2050. livestock raised in extensive versus intensive production systems. Intensive systems are In this exercise we use a combination of Hansen et al. characterized by high output of animal products per (2013) and Hewson et al. (2019) data. unit surface area, and extensive systems use land areas of low production and under conditions of moderate Opportunity cost data grazing. For livestock raised in extensive production The cost of avoided deforestation is assumed to be systems, the rental rate is assumed to be twice that for equal to the opportunity cost of land in forests, intensive systems. The same regional rental rates assuming this cost would need to be covered to assumed for crop products are assumed for livestock maintain forest standing in the future. In this analysis, products in intensive systems. Therefore, when the opportunity cost per hectare of forest is based on calculating pastureland rent, the rent is weighted Changing Wealth of Nations (CWON) 2021 estimates according to the country’s share of livestock production (World Bank 2021a) of protected areas’ wealth derived in extensive systems and intensive systems. from a combination of World Development Indicators Se.plan’s data on opportunity for low- and middle- and FAO data and complemented with opportunity income countries are based on survey data and cost data for low- and middle-income countries from econometric modeling. Like CWON, se.plan assumes the se.plan. that the alternative land use would be some form of In CWON 2021, opportunity costs of forest land in agriculture, either cropland or pastureland. It sets the protected areas are estimated as the lowest of the opportunity cost of potential restoration sites equal to returns to cropland and pastureland, using the annual the value of cropland for all sites where crops can be flow of rents the land generates from crop and livestock grown, with the opportunity cost for any remaining production and taking the present value of such rents in sites set equal to the value of pastureland. Se.plan uses the future. In CWON 2021, cropland future rents a number of spatially disaggregated data on production account for the effect of soil degradation and climate per hectare, revenues, and input costs from FAO, Page 22 International Food Policy Research Institute (IFPRI), wildlife to humans are likely to occur overlays European Space Agency, and other sources. Rental deforestation risk, high biodiversity, and accessibility rates are based on an econometric model relating data. We estimate the annual cost of avoided profits to fixed inputs including cultivated area. All deforestation for each of the zoonotic spillover risk details of the methodology implemented by se.plan can groups developed by Allen et al. (2017) (see Appendix 1 be found in the tool’s documentation. for more details on the approach developed by Allen et al. 2017). In our analysis both CWON 2021 and se.plan data are used for low- and middle-income countries to provide a We combined Hansen et al. (2013) forest data with the range of possible costs of avoided deforestation (a Hewson et al. (2019) predictions of deforestation risk lower and upper bound). For high-income countries by 2030, using the results only from their global model only CWON 2021 data are used, given this information and pixels with a high probability of deforestation. We is not available in se.plan. compared results using forest plots with a likelihood of deforestation greater than 0.7. The total area of forest Biodiversity data at risk is adjusted based on the Hansen et al. (2013) high-resolution data. Hectares at risk are estimated We focus the analysis on the main groups of species based on tree cover in 2000 minus tree cover loss up which are known to host potential pathogens. Data for to 2020. the biodiversity of priority clades—birds (no seabirds) and mammals in the orders Chiroptera (bats), Rodentia, Then we overlay the hectares at risk by combining and Primates—on a global scale are derived at a Hansen et al. (2013) data and Hewson et al. (2019) data resolution of 10 km from biodiversitymapping.org at the 0.7 risk threshold with the layer of high (Jenkins, Pimm, and Joppa 2013; Pimm et al. 2014) and biodiversity in birds, bats, primates, or rodents and were updated for March 2018. The data were accessibility. Based on this intersection of layers, Figure resampled to match the data of tree cover loss risk until 7 shows the hectares of forest used to calculate the 2030 in Hewson et al. (2019) at 1 km resolution. Pixels costs of avoided deforestation. were classified as having high biodiversity if they are in a region (10 km pixel) containing several species above We assume that 50 percent of deforestation is avoided the 85th global percentile in any selected clade. in areas of high risk. The lower and upper bounds are provided by the variation between CWON 2021 and Accessibility data se.plan opportunity cost estimates. Spatially explicit accessibility indicators for 2015 are Appendix 2 presents the number of hectares of avoided derived from Nelson et al. (2019). The data are at a 30 deforestation by country annually, assuming that 50 arc seconds spatial resolution (approximately 1 km at percent of the total predicted forest loss is prevented. the equator) and were projected to match the data of tree cover loss risk until 2030 in Hewson et al. (2019) The annual cost of avoided deforestation in countries at 1 km resolution. We selected areas as ‘accessible’ if at high risk of spillovers is between US$3.16 and they are within five hours or less of a human settlement US$4.43 billion if forest plots with deforestation risk between 100,000 and 200,000 people. greater than 0.7, high biodiversity, and accessibility are considered. These results assume that half of the B. Proposed approach and results on the costs projected deforestation is avoided (about 1.5 million of avoided deforestation ha per year in high-risk countries using the 0.7 threshold). Our approach to estimate the annual costs of avoided deforestation in areas where zoonotic spillovers from Page 23 C. Analysis of restoration costs • They include areas where tree cover can potentially occur under current climatic conditions, as Restoration costs are estimated using se.plan, a determined by Bastin et al. (2019). spatially explicit online tool designed to support forest • Their current tree cover, as measured by the restoration planning decisions developed by the UN European Space Agency’s Copernicus Programme FAO in collaboration with Spatial Informatics Group (Buchhorn et al. 2020), is less than their potential (SIG), SilvaCarbon, Peking University, and Duke tree cover. University. This tool identifies locations where the • They are not in urban use. benefits of forest restoration are high relative to restoration costs, subject to biophysical and The tool labels areas that satisfy these criteria as socioeconomic constraints that users impose to define potential restoration sites. It treats each grid cell as an the areas where restoration is allowable. It also independent restoration planning unit, with its own provides information on the benefits and costs of potential to provide restoration benefits and to include restoration. restoration costs and risks. The tool includes only areas with potential for The tool provides information on four benefit restoration that satisfy the following four criteria: categories: biodiversity conservation, carbon • They are in countries or territories of Africa, the sequestration, local livelihoods, and wood production. Near East, Asia and Pacific, and Latin America and The benefit variable prioritized in this study is the Caribbean that the World Bank classified biodiversity conservation. Se.plan includes two as low- or middle-income countries or territories indicators for biodiversity conservation and local during most years from 2000 to 2020. livelihoods and one indicator each for carbon FIGURE 7. Hectares of forest used to estimate the costs of avoided deforestation—intersection between high deforestation risk by 2030, biodiversity density, and accessibility. Page 24 TABLE 6. Annual cost of halving deforestation - by risk of zoonotic spillover High risk Medium risk Low risk Non-classified Total   Avoided forest loss (ha/year) 1,466,991 3,012,040 254,397 260,370 4,993,799 Cost (billions 2020 US$/year) 3.16–4.43 6.02–6.35 0.20–0.46 0.47–0.47 10.44–11.13 Cost (2020 US$/year/ha) 2,155–3,023 2,000–2,109 794–1,820 1,789–1,793 2,091–2,228 sequestration and wood production. The following are se.plan, the opportunity cost refers to the value of land the two indicators for biodiversity conservation: if it is not restored to forest.  The tool assumes that the • Number of endangered species2: Total number of alternative land use would be some form of agriculture, critically endangered and endangered3 amphibians, either cropland or pasture. It sets the opportunity cost reptiles, birds, and mammals whose ranges overlap a of potential restoration sites equal to the value of site. Sites with a larger number of critically cropland for all sites where crops can be grown, with endangered and endangered species are ones where the opportunity cost for any remaining sites set equal successful forest restoration can potentially contribute to reducing a larger number of extinctions. to the value of pasture. Sites that cannot be used as either cropland or pasture are assigned an opportunity • Biodiversity intactness index (BII) (Newbold et al. cost of zero. Implementation costs refer to the expense 2016): This indicator describes the average abundance of a large and diverse set of organisms in of activities required to regenerate forests on cleared a geographical area, relative to the set of originally land. They include both initial expenses incurred in the present species. The tool subtracts the BII from 100 first year of restoration (establishment costs), including to measure the gap between full intactness and activities such as site preparation, planting, and fencing, current intactness. Sites with a larger BII gap are and expenses associated with monitoring, protection, ones where successful forest restoration can and other activities during the subsequent three to five potentially contribute to reducing a larger gap. years that are required to enable the regenerated stand Se.plan allows rating of the relative importance of these to reach the ‘free-to-grow’ stage (operating costs). benefits and then calculates an index that indicates Se.plan assumes that implementation costs include each grid cell’s relative restoration value. This planting expenses on all sites. This assumption might restoration value index is a weighted average of the not be valid on sites where natural regeneration is benefits, with user ratings serving as the weights. It feasible. To account for this possibility, the tool includes therefore accounts for not only the potential of a grid a data layer that predicts the variability of natural cell to provide each benefit but also for the relative regeneration success.4 The tool calculates an importance that a user assigns to each benefit. It is approximate benefit-cost ratio for each site by dividing scaled from 1 (low restoration value) to 5 (high the restoration value index by the restoration cost and restoration value). For this study, the BII was selected converting the resulting number to a scale from 1 (small as the indicator with the highest importance. ratio) to 5 (large ratio). Sites with a higher ratio are the Forest restoration incurs two broad categories of ones that are predicted to be more suitable for costs–opportunity costs and implementation costs. In restoration. 2 World Bank, which processed over 25,000 species range maps from (a) IUCN, The IUCN Red List of Threatened Species, https://www.iucnredlist.org, and (b) BirdLife International, Data Zone. http://datazone.birdlife.org/species/requestdis. 3 Please see se.plan documentation for more details on these categories. 4 See https://docs.sepal.io/en/latest/modules/dwn/seplan.html?highlight=se.plan#cost-data-layers for more details on data sources used for assessing opportunity and implementation costs. Page 25 We overlay the results of se.plan for hectares with a TABLE 8. Restoration costs for Liberia (2017 US$), benefit-to-cost ratio greater than 3, that is, high for 97,237 ha of high restoration potential restoration potential, with the biodiversity layer identifying high density of birds, bats, rodents, and Cost category 2017 US$ primates. Results of this tool are presented in Tables Land opportunity cost (US$/ha) 1,299.98 7–9 and Figures 8–10 for three key areas of study: Implementation cost (US$/ha) 796.33 Assam State in India (13,688 ha), Liberia (97,237 ha), Implementation cost for highly suitable 77,432,299.57 and Vietnam (73,375 ha). areas (US$) Land opportunity cost for highly suitable 126,406,574.52 areas (US$) TABLE 7. Restoration costs for Assam state, India, TABLE 9. Restoration costs for Vietnam (2017 for 13,688 ha of high restoration potential US$), for 73,375 ha of high restoration potential Cost category US$ Cost category 2017 US$ Land opportunity cost (US$/ha) 3,333.54 Land opportunity cost (US$/ha) 4,534.46 Implementation cost (US$/ha) 1,408.88 Implementation cost (US$/ha) 1,475.34 Implementation cost for highly suitable Implementation cost for highly suitable 108,252,831.47 19,284,739.29 areas (US$) areas (US$) Land opportunity cost for highly suitable Land opportunity cost for highly suitable 332,715,981.80 45,629,561.21 areas (US$) areas (US$) FIGURE 8. Areas with restoration benefit-cost ratio greater than 3 intersecting with high biodiversity in Assam State, India Page 26 FIGURE 9. Areas with restoration FIGURE 10. Areas with restoration benefit-cost ratio greater than 3 benefit-cost ratio greater than 3 and high biodiversity in Liberia and high biodiversity in Vietnam D. Discussion This study presents an initial approach to estimate the annual costs of avoided deforestation and restoration by 2030 to prevent emerging zoonotic spillovers using high-resolution global datasets. Additional analytical steps could follow this analysis to improve the resolution of cost estimates of avoided deforestation and restoration potential and to better target areas where zoonotic spillover can be prevented. Our approach allows to spatially locate where avoided deforestation and restoration actions can be It is important to highlight that in forest plots that are implemented to not only identify how much it would at risk but that fall under protected area management, cost but where to do it. Additionally, it could help the opportunity costs presented here could be an identify overlap with investments in nature-based overstatement of the true conservation costs. In this solutions that generate other forms of benefits. In other case the financial (management) costs of conservation words, this approach could help decision-making at the could be a better approximation. As a point of country level with spatially located cost-benefit comparison, Waldron et al. (2020) estimate current metrics. global spending on protected areas at about US$24.3 billion per year. The authors also estimate annual costs Page 27 to give minimum adequate funding to existing deforestation prevention and restoration should be protected areas at US$67.64 billion per year. Using the focused for preventing emerging zoonotic spillovers. latter figure, the average financial cost per hectare of forest protection would be about US$40 (based on the On forest restoration, which is an important addition to total number of hectares in protected areas by 2018 the analysis of actions to prevent zoonotic spillovers from CWON 2021). This is a much lower cost per from land use changes, future work could generate hectare compared to the average opportunity costs per estimates of restoration costs for all high-risk countries hectare. That said, of the hectares identified as according to Allen et al. (2017) using different selection priorities for prevention of zoonotic spillovers in this criteria, spatial constraints, and the se.plan tool to study, only 10.5 percent overlap with protected areas, provide reasonable bounds for these estimates. which means opportunity costs may be a relevant measure to estimate the cost of implementing 3.2 WHAT IS THE COST TO PREVENT conservation actions in these areas. DISEASES THROUGH THE FOOD SYSTEMS? Not considered in the analysis thus far are the ecosystem services provided by forests being Food systems expose many opportunities for protected or restored. While these would depend on interactions between viruses and humans. They the specific locations that are targeted, based on embrace the entire range of actors and their interlinked CWON 2021 data, the global net present value of the value-adding activities along a complex supply chain, ecosystem services associated with the hectares of from farm to fork to landfill. A chicken or a pig is avoided deforestation considered in this analysis produced at the farm, aggregated with other animals, (including three ecosystem services: water, recreation, transported, slaughtered, processed, distributed, and and non-wood forest products) is about US$6.7 billion consumed and some of its parts discarded in a landfill, per year. While this does not represent the whole sewage, or nature. Food products originate from spectrum of services provided by forests (for example, agriculture (including livestock), forestry, fisheries, and not including carbon sequestration, pollination, or food industries and are embedded in the broader avoided costs from pandemics), these figures show the economic, societal, and natural environments (Von importance of considering benefits next to costs when Braun et al. 2021). As can be seen from Figure 11, the assessing investments in avoided deforestation. food system spans across several critical drivers that are related to increasing risk of EIDs. The global databases used in this study can be used to conduct additional analysis complementary to the one Viral species jumping and human infection can occur presented in this note. Country-level models of risk of at any stage of the food supply chain, following deforestation can be developed for key high-risk human exposure to animals (wildlife and livestock). countries using a range of control variables such as More than 60 percent of EIDs are zoonotic (that is, an human settlement data, measures of economic activity animal incubates the organism, which then transmits (for example, nighttime light data), distance to official into human populations) (Taylor, Latham, and and unofficial roads, land use type, and other key Woolhouse 2001). Different animals can be at the geographical determinants such as slope and biome. A source of infection: wildlife (for example, bats, rodents, combination of theory-based models and machine- primates, and wild birds), livestock (for example, learning models can be used to validate the model poultry, pigs, and camels), and blood sucking insects predictions using historical trends at the country level. (intermediate species or vectors, for example, Results of these models can be overlaid with high- mosquitos, fleas, and ticks). The infectious agent also resolution biodiversity data as done in this analysis as varies (for example, bacteria, virus, parasites, and fungi). well as land use data to better target where Page 28 FIGURE 11. The food system in the context of other systems Income & Health Employment Ecology & Systems Climate Systems Inclustion Consumption, Agriculture Nutrition & Safety | Availability | Diversity and Food Health Industries Economic & Inclustion Science & Governance Markets, Innovation Systems Infrastructure, & Systems Services Source: Adapted from Von Braun et al. 2021. Note: Yellow circled area is the food system. Among animals, livestock deserves special attention countries. Although the volume of consumption of because it is associated with constant, daily, high wildlife products for food is at least an order of contact between humans and animals. This leads to a magnitude lower than it is for domestic livestock particularly high risk of spillover of pathogens from (Karesh and Cook 2005), wildlife may contain a larger animals to humans. Given the high contact nature of diversity of potential pathogens. Moreover, wildlife livestock animals and humans, livestock animals, in fact, hunting is typically not directly addressed in national play an ‘amplifier role’ in terms of dramatically agriculture, food security, or animal health policies, increasing the likelihood of pathogen spillover (see perpetuating the risk of infections. While most Figure 14). Opportunities for disease transmission to countries have services that cover domestic animal other animals or humans along the supply chain can health, in many countries their reach is neither occur at many points: at the farm, transportation, sufficient nor does it cover wildlife, and many slaughterhouses, wet markets and other retail outlets, pathogens become endemic. consumption, and disposition. The farm (or livestock production unit) is of particular concern due to the In fact, while the next pandemic is expected to be of frequency and proximity of encounters between viral origin, nonviral zoonotic pathogens are also humans, livestock, and wildlife hosts. important. Historically, discounting the COVID pandemic and HIV, the costs of nonviral infections have Wildlife brought into the food supply chain is also a dwarfed the costs of viral infections, affecting large risk factor. Meat hunting, consumption, and trade predominantly the poor. Livestock is exceedingly are an important economic and cultural activity in many important for poor people. Some estimates suggest Page 29 nearly 1 billion people living on less than US$2 per day impacts of endemic zoonoses. Understanding factors are dependent to some extent on livestock (Staal et al. that distinguish typical localized outbreaks from large 2009). Over 600 million are in South Asia, mostly in regional epidemics and pandemics is important India. Sub-Saharan Africa has over 300 million poor (Stephens et al. 2021). livestock keepers, concentrated in East and West Africa, with fewer in Southern and Central Africa. In Poultry and pigs increasingly dominate in terms of poor countries, livestock provides 6 to 36 percent of number of animals kept. In 2018, an estimated 69 protein intake. COVID, however, has turned the tables. billion chickens, 1.5 billion pigs, 656 million turkeys, 574 While the costs of past viruses have been modest, the million sheep, 479 million goats, and 302 million cattle COVID pandemic alone has created 517 million were killed for meat production (Roser and Ritchie infections and 6.3 million deaths (Worldometers 2022), 2013). About 85 percent of all domestic animals alive even with aggressive health sector responses to it. are now pigs or poultry Robinson et al. 2011. Zoonotic food-borne pathogens are markedly higher in poultry The way these costs are distributed is also different. and pigs than in small and large ruminants. This Nonviral infection costs are primarily borne by poor suggests that as monogastric systems expand, so may people in developing countries, where there is lesser food-borne diseases. As disease transmission is capacity to monitor, manage, and remediate infections. dependent on numbers and contact rates and Costs of viral pandemics may be distributed differently, monogastric systems are in higher numbers and more with infections spreading far from first transmission, intensive systems, they are more important in disease and with rich people, and richer countries, sharing a emergence. larger burden of disease. While endemic zoonoses cause on average 2.7 million deaths a year, mostly Increased livestock numbers coupled with among poorer people in poorer countries, COVID has intensification, which increases animal density, are averaged 3 million deaths per year, mostly in the rich an important risk factor for increased disease and middle-income world. Consequently, donors and transmission. More livestock creates more rich decision-makers are often more concerned with opportunities for viral transmission. While in 1961 there these emerging diseases, whose impacts on poor were about 1,409 poultry types per 1,000 humans, by farmers could be orders of magnitude less than the 2018 the ratio was 3,367. With human population FIGURE 12. Global poultry and pig population Source: Our World in Data 2022. Page 30 FiGURE 13. Number of animals per thousand humans 4,000.0 3,500.0 3,000.0 2,500.0 2,000.0 1,500.0 1,000.0 500.0 0.0 1961 1966 1971 1976 1981 1986 1991 1996 2001 2006 2011 2016 Poultry/Humans Pigs/Humans Cattle/humans Source: Our World in Data 2022. expected to increase by 2.3 billion in the next 30 years, A. Risk framework for viral zoonotic outbreaks infection risks from poultry will also increase under a business-as-usual scenario. In contrast, the pig-human Many authors have written on the risks of zoonotic ratio has varied little, from 131 pigs per 1,000 humans in viral outbreaks reaching pandemic proportions. The 1961 to 128 in 2018, probably reflecting the FLURISK initiative (De Nardi et al. 2014) developed a intensification of pig production. Selection, breeding, risk assessment methodological framework for and management for increased productivity in livestock potential pandemic influenza strains. While the project create host populations conducive to pathogen dates from 2011 and is focused on influenza strains, its evolution and persistence (through lack of genetic framework is simple, easy to explain, and therefore diversity, high numbers and contact opportunities, powerful to guide thinking about virological, stress-induced immunosuppression, and other factors). epidemiological, and ecological risk factors for an This provides opportunity for ‘wild’ microorganisms to animal virus to jump species and infect humans. invade and amplify or for livestock pathogens to evolve There are two critical factors in zoonotic transfer of to new and more pathogenic forms. In addition, pathogens. The first is the ability of the virus to infect corollaries of intensification such as high pest densities, humans, a function of its genetic and biochemical extensive transportation networks, sale of live animals characteristics. The second is the opportunity of the for food and pets, landscape modification, poor waste virus to do so, a function of the frequency of management, and juxtaposition of agriculture or interactions between humans and infected animals, recreation with wildlife contribute to ‘emergence’ and wild or domestic, and the susceptibility of those shifting virulence of diseases (Grace et al. 2011). humans to infection. The ability of the virus to infect Furthermore, much of the current intensification is humans is related to virological or phenotypic driven by rising demand in developing countries, where characteristics, mainly (a) receptor preference, (b) the systems for disease control and reporting are presence of mutations in the ribonucleoprotein relatively weak Page 31 complex, (c) reassortment within the virus, and (d) humans as a critical determinant of risk of relatedness to strains already circulating in humans that opportunities for transmission. Opportunities are also may have endowed them with immune protection. The likely to depend on the proximity of animals and opportunity for the virus to jump to humans is a humans when encounters take place. That is, one can function of the frequency and proximity of encounters have high human and animal density, say per square between human and animal hosts. kilometer, but the species may be physically separated, for example, through enclosures or other barriers. Proper detection of viruses with the intrinsic Under these circumstances, some of the encounters, abilities—attachment, mutations, reassortment—to even if less frequent, may bring humans and animals infect humans is not easy. It is limited by the extent of nearer, increasing again the risks of infection. Some current scientific knowledge. It is an unpredictable, humans are more susceptible to infection, especially natural process in which it is not possible to account for those in close and frequent contact with animals. This all uncertainties (De Nardi et al. 2014). For example, the is particularly so when there are direct ‘transactions’ mechanisms by which specific viral mutations occur, between humans and animals, for example, through and their impact, are not fully understood. Detection of hunting, caving, feeding, petting, cleaning, slaughtering, able viruses requires specialized laboratories that can transporting, retailing, consuming, or disposing. map the genetic characteristics of the virus and rigorous collection and sampling of animals, consistent The interaction between animals and humans, and with approved scientific and statistical methods. the wider epidemiological context in which the virus Collection and sampling are particularly costly if not is present, is therefore paramount when assessing impossible in remote and wild areas in developing zoonotic risks. Several authors argue that the virus countries, the least capable of doing so. Wild carriers of intrinsic abilities are not the most important viruses may be difficult to reach. Livestock at higher risk component driving risk. For example, influenza viruses is likely to be those in poorer areas where government derived from mammals and birds represent a significant services are not fully present. Allen et al. 2017, for public health threat globally because influenza viruses example, suggests that tropical forests (broadleaf have exceptional capacity for genetic recombination, cover) are an important indicator of outbreak risk: hence the ability to infect. But even here, in the “After accounting for reporting effort, we show that presence of virulent strains, the dominant risk zoonotic EID risk is elevated in forested tropical regions parameter is the number of extensively reared domestic experiencing land-use changes and where wildlife chickens. That is, the location of the virus, determining biodiversity (mammal species richness) is high.” opportunity, appears to dominate the innate ability of a virus to cause human infection as a source of risk (De Given the difficulties in detecting viruses with the Nardi et al. 2014). ability to jump to humans, the alternative course of action is to reduce the opportunities for the viruses B. Reducing the risk of infections through the to do so. Opportunities are likely to depend on factors food supply chain such as wildlife, poultry, or pig density relative to human population density. Where there are more animals and i. Zeroing in on hot spots humans stationed at a given location, encounters are The opportunity or frequency of encounters more likely. Records of known human infections with between the virus and humans depends on the H5N1 influenza virus show that this is most likely to density of humans and hosts. The characteristics of occur in regions with high human and poultry densities the location where encounters take place, for example, (De Nardi et al. 2014). Likewise, high density of wild farming system, underly successful transmissions mammal hosts is also a critical risk factor. We therefore (Tables 10 and 11). In intensive, quasi-industrial, animal propose looking at the relative densities of hosts and Page 32 FIGURE 14. Framework for identifying hot spots for risk of zoonotic viral epidemics KEY: transmissions • Meat relative prices • Land use changes opportunities drivers • Income growth Livestock/ Livestock Health humans density Systems density Livestock/ Ecology & wildlife Human Climate density density Systems • Urbanization • Migration Economic & Wildlife/ • Population Growth Wildlife Governance humans density Systems density • Deforestation • Hunting • Trade TABLE 10. Summary of chicken production systems System Housing Characteristics Broilers Broilers assumed to be primarily loosely housed on Fully market-oriented; high capital input requirements litter, with automatic feed and water provision (including Infrastructure, buildings, equipment); high level of overall flock productivity; purchased non-local feed or on-farm intensively produced feed Layers Layers housed in a variety of cage, barn and free-range Fully market-oriented; high capital input requirements systems, with automatic feed and water provision (including infrastructure, buildings and equipment); high level of overall flock productivity; purchased non-local feed or on-farm intensively produced feed Backyard Simple housing using local wood, bamboo, clay, leaf Animals producing meat and eggs for the owner and material and handmade construction resources for local market, living freely. Diet consists of swill and supports (columns, rafters, roof frame) plus scrap wire scavenging (20 to 40 percent) and locally-produced netting walls and scrap iron for roof. When cages are feeds (60 to 80 percent) used, these are made of local material or scrap wire Source: FAO. Page 33 production, one human may interact with a large vectors including wild birds, rodents, and insects. This number of hosts, say, chicken or pigs, some of which facilitates disease transmission from wildlife to may be infected. In extensive, backyard farming livestock, from livestock to livestock, and from livestock systems, one human may interact with just an to humans. A typical example of zoonotic disease individual host. However, because there will be many affecting smallholders is trichinosis, a parasitic disease such backyard units, a large number of hosts will circulating in wild and domestic animals such as rats, interact with a large number of humans (De Nardi et pigs, and wild pigs and occasionally infecting human al. 2014). Or in other words, consider a simple through the consumption of inadequately cooked, hypothetical situation of 100 chickens in backyard infected pork. Another example is avian influenza systems versus 100 in intensive systems and that each occurrence in smallholders in Southeast Asia, day one human interacts with one chicken. Assume potentially via mixing of wild and domestic birds and that 10 percent are infected in each system and that a amplification and spread of highly pathogenic strains in single encounter with a human will pass on the virus. poultry with production and public health impacts. In backyard systems, 10 humans will be infected with Proximity due to occupational exposure to an infected certainty from that encounter, but in intensive animal population was identified as being correlated systems there may be only a 10 percent chance that a with human infection, and there were indicators that human will be infected the day the encounter takes working with backyard poultry increased the risk to place. humans. Small-scale livestock production systems are usually Intensive systems, on the other hand, benefiting from linked to poor hygiene and low biosecurity with few economies of scale, are more likely to adopt barriers to prevent contacts between animals, biosecurity measures. The use of personal protective humans, and wildlife. Particular biosecurity challenges equipment, such as masks, gloves, or body covers; arise from the fact that animals are reared outdoors, enclosures and fencing; mechanized feeding; and which increases the chances of contact with disease specialized disposal may be financially justified where TABLE 11. Summary of pig production systems System Housing Characteristics Industrial Fully enclosed: slatted concrete floor, steel roof and Fully market-oriented; high capital input requirements support, brick, concrete, steel or wood walls (including infrastructure, buildings, equipment); high level of overall herd performance; purchased non-local feed in diet or on-farm intensively produced feed Intermediate Partially enclosed: no walls (or made of a local Fully market-oriented; medium capital input material if present), solid concrete floor, steel roof requirements; reduced level of overall herd and support performance (compared with industrial); locally sourced feed materials constitute 30 to 50 percent of the ration Backyard Partially enclosed: no concrete floor, or if any Mainly subsistence driven or for local markets; level of pavement is present, made with local material. Roof capital inputs reduced to the minimum; herd and support made of local materials (e.g., mud bricks, performance lower than in commercial systems; feed thatch, timber) contains maximum 20 percent of purchased non-local feed; high shares of swill, scavenging and locally sourced feeds Source: FAO. Page 34 there are many chickens in close quarters. It may appear Nor are all pathogens the same. In fact, livestock outrageously expensive, though, where there are only a intensification is likely to have different impacts on the small number of chickens, such as in backyard farming. key zoonoses depending on their epidemiology. Of the priority zoonoses, viral and bacterial, evaluated by the Nevertheless, intensive systems pose their own risks. International Livestock Research Institute (ILRI) in While virus diversity may be less in intensive than in 2012, nine are likely to become more of a problem with backyard systems, if an aggressive virus hits the intensification, four are likely to decrease, and for the intensive unit, close quarters, high density, and similar remaining there is no clear link. Whether disease genetic content of livestock are likely to facilitate the increases or decreases also depends on the type of quick spread of the infection. That is the rationale for intensification and other factors. Where intensification the quick and overwhelming recourse to culling of occurs in close association with wildlife, risks for millions of chickens, when avian flu is suspected in the disease spillover are higher. If intensification is flock. accompanied by improvements in biosecurity and a disease control program, then diseases such as bovine Moreover, not all intensive systems look the same. In tuberculosis (TB) may decrease. Other diseases tend to fact, we are talking about a gradient, a range of systems increase in the early stages of intensification but may going from the subsistence backyard farm to medium- decrease thereafter. scale higher intensity systems to factory farms with sometimes millions of chickens or thousands of pigs There are other transmission hot spots along the (Tables 10 and 11). Poor practices remain prevalent for supply chain, where encounters are frequent and most of the path toward full intensification. In extreme proximity is high. One is wet markets, which can be cases, animals are raised on top of their own excrement, seen as an extreme situation of the extensive breathing poor air quality with high levels of dust, (backyard) farm model, with many humans interacting aerosol particles, and gaseous ammonia, and are under with one potentially infected animal. The frequency of continued stress due to overcrowded housing, encounters will also be affected by mobility. An infected predisposing them to respiratory conditions. host that moves intentionally (for example, migration) or unintentionally (for example, transported by humans) Furthermore, along with their close living quarters is likely to have more and closer encounters with other and lack of clean areas, weakened immune systems livestock and humans along the route. from lack of exercise and nutrition and excess use of antibiotics render animals more susceptible to Capacity to conduct disease detection and viral disease. Although antibiotics were originally used to discovery in people, livestock, and wildlife is lowest in treat infections from various pathogens, they are also countries with the highest mammalian biodiversity used improperly to prevent infections and as additives (and therefore viral diversity). Many viruses may for feed and biocides for growth promotion. Given the remain undiscovered, even if they have already emerged density of animals, good animal husbandry practices, repeatedly into human populations in these regions. biosecurity, and vaccinations are needed in these Bats have a statistically significant higher proportion of farming systems, but unfortunately this is often zoonoses than any other mammalian order, even though compensated by the excessive use of antimicrobials, bat viruses are not historically well studied relative to driving antimicrobial resistance. Therefore, special other mammalian host groups. attention is needed for the proper, prudent use of antimicrobials and to protect against antimicrobial The best, and perhaps the only one viable, strategy in resistance, likely to be lagging in low- and middle- face of huge uncertainties about future zoonoses and intensive systems without economies of scale. pandemic risk is to focus on hot spots. Hot spots will help direct prevention effort and financial assistance to Page 35 FIGURE 15. High-risk areas for pathogen transmission between wildlife, livestock, and humans higher risk areas, thus avoiding wasting resources in FIGURE 16. South-East Asia and Indonesia— lower risk locations and the escalation of prevention very high-risk areas for viral species costs. Hot spots can be defined as areas where several jumping (in red) factors converge to create high-risk disease emergence and spread zones and where weakening or deteriorating public and animal health systems and inappropriate institutional conditions mean that disease has a high chance of becoming entrenched. These factors include the presence of dense human and wild mammal populations; increased contact among human, wild, and domestic animals; changing farming systems; and areas of poor sanitary conditions, especially in peri-urban settlements and farming communities. The relevant transmission paths are first, from wildlife (bats, primates, rodents, and wildfowl) to other wildlife, livestock, or humans; second, from livestock to wildlife, other livestock, or humans; and third, from humans to other humans, livestock, or wildlife (Figure 15). The frequency and closeness of encounters is likely to be closely related to wildlife, livestock, and human density. Higher densities will create more opportunities for transmission. Page 36 And where are the hottest hot spots, FIGURE 17. Eastern and Western Africa - very high-risk areas for viral species jumping (in red) where the three densities intersect? Absent other criteria, we suggest that where the three densities intersect marks the highest risk hot spots (hottest hot spots), while where only two densities intersect marks the lesser risk hot spots. The hottest hot spots are emerging in Asia and Africa (in red in Figures 16 and 17). In Asia the most susceptible countries and regions are Myanmar, Thailand, Malaysia, and Vietnam and the islands of Sumatra, Java, and Sulawesi. Vietnam appears particularly high risk with many red Accessibility also matters, that is, whether encounters spots. In Africa, the most susceptible countries are happen at close quarters or not. In fact, transmission Ghana, Togo, Benin, and Nigeria in West Africa and may still occur at low densities if the encounters are Uganda, Kenya, Rwanda, Burundi, and Tanzania in East with proximity, for example, in wildlife hunting. It also Africa. matters whether there are barriers to encounters, such The presence of biosecurity measures, such as as in farming systems with enclosures that limit enclosures and fences, and the use of personal movement of livestock and therefore interactions with protective equipment also need to be considered. wildlife or those that adopt relevant biosecurity They are a major factor in impeding the virus from measures that create barriers for viruses to jump jumping species and spreading. The density of poor species. livestock keepers who are more likely to lack High-risk hot spots from a food system perspective are appropriate animal health measures can indicate where those where high densities intersect. There are four biosecurity coverage may be weaker. While dated, relative densities to consider: wildlife and livestock; Figure 18 suggests that the regions less likely to resist wildlife and humans; livestock and humans; and the infections are South Asia, Southeast Asia, the combined livestock, wildlife, and human density. High- Philippines, Indonesia, and parts of Eastern and Central risk regions are those where at least two of the density Africa. They therefore coincide to a large extent with risk parameters intersect—wildlife biodiversity the hottest hot spots and with some of the lesser hot (susceptible hosts – primates, bats, rodents, wildfowl), spots. While pig production is highly concentrated in livestock (poultry and pigs), and humans—and where Southeastern China and coincides with high human there are few constraints on proximity, such as natural or numbers, the density of poor livestock keepers is much artificial barriers (enclosures and fences). Figure 15 lower, more likely because biosafety measures are in illustrates the method. Wildlife to human risk (in blue) place. It could therefore be taken out of the list of hot prevails, from left to right, in Mexico and Central America, spot countries in a first approximation. On the other Colombia, Brazil, West and Eastern Africa, Southeast hand, although India and Bangladesh exhibit lower Asia, Sumatra, and Borneo. The livestock to human risk wildlife to human risk, they have high livestock to (in yellow) prevails in parts of West and East Africa, India, human risk, compounded by a large concentration of Bangladesh, parts of Southeast Asia, Southeastern poor livestock keepers and should join the high-risk hot China, the Philippines, and the island of Java. spots list. Page 37 development; however, despite caveats various conclusions can be drawn. In terms of numbers of livestock keepers, the critical regions remain South Asia and sub-Saharan Africa. The mixed farming systems (crop and livestock) contain large numbers of poor (over 1 billion), and numbers of poor people dependent to some extent on livestock in these systems are considerable. Mixed rainfed systems have more poor livestock keepers than mixed irrigated systems. Rangeland systems have least absolute numbers of the poor but the poor in this system have highest dependency on livestock. Almost half of the poor in rangeland systems are located in sub-Saharan Africa. Figure 2.4 Density of poor livestock keepers in developing countries based on national data (updated diseases with epidemic potential at FIGURE 18. Density of poor livestock keepers March 2012) national, regional, and international levels ii. Limiting encounters and proximity How livestock operations are managed (on farms, slaughterhouses, and markets) to minimize risks to animal and human health (that is, biosecurity) determines the risk level that livestock poses in the emergence (or reemergence) and transmission of infectious diseases to humans. Some changes are evident since the map of 2002. There has been a marked decrease in the density Despite its importance, a policy of poor livestock keepers in South America and SE Asia. framework and investments needed for 25 optimal biosecurity measures typically remain underfunded or lacking in most developing countries. Focusing on In these hot spots, the strategy should focus on two biosecurity in livestock operations is particularly goals: first, reducing the number of encounters and important now because growth in consumption of meat proximity between humans, wildlife, and pigs and over the next decade is projected to increase by 12 poultry, and second, setting up barriers to impede percent by 2029 when compared to the base period transmission when close encounters happen. One (2017–2019). Unlike many high-income countries that should consider an investment and policy framework to are reaching saturation levels in terms of per capita reduce the risk of transmission. consumption, this is driven by trends in developing An investment framework for reducing the risk of countries where aggregate consumption of meat has EIDs must balance the twin objectives of (a) primary been on a continuous upward trajectory, driven by prevention, which attempts to mitigate the population and income. underlying factors that lead to infection spillover Reducing the risks of transmission in hot spots calls from animals to humans and (b) preparedness for adequate policy, regulatory, institutional, and including early detection (given that prevention investment interventions to limit the number and measures are not foolproof). Priority areas for primary proximity of encounters between wildlife, livestock, prevention measures related to the food system are and humans. Petrovan et al. (2021) offered an centered around improving biosecurity in livestock exhaustive list of potential interventions that should be operations (small-scale and large-scale intensive applied in hot spots. Interventions to limit encounters operations) as well as relevant policies such as land use and proximity fall into two categories: (a) enclosures, or land zoning to guide the location of livestock walls, and fences and (b) restricted proximity and operations far away from humans and wildlife. They mobility. also include healthy diet policies that relate to animal protein consumption. Priority areas for preparedness Enclosures through fencing or other barriers, including early detection related to the food system are artificial or natural, can reduce the number of the establishment of global risk-based surveillance encounters between wildlife and livestock, on one systems for existing and emerging infectious zoonotic hand, and between livestock and humans, on the other hand. Free roaming of livestock can put them in Page 38 direct contact with infected hosts, wild or domestic. • Restricting the long-distance transport of high-risk Interventions should include the following: species by using species quotas, licensing traders, or export certification; regulating maximum allowed • Setting up enclosures to prevent livestock roaming animal density for animals simultaneously housed and fencing to enable density controls in crop areas or transported considering the feasibility of traded • Zoning land use to separate human settlements livestock licenses; minimizing international from livestock farming and limiting the urban transport of live livestock; and building the capacity sprawl into natural areas with host biodiversity of airport, seaport, border, and cargo personnel for presence safe, secure, and scientific handling of wildlife and • Protecting and fencing off areas used by domestic animals and animal products during susceptible hosts, such as bats in caves or transit checks important roosting tress, in proximity of farmland • Introducing mandatory separation at markets or • Restricting visits to high stocking density livestock shops between live animals of different species, units that house similar species or species with between live domesticated and wild animals, and high disease transmission risks between live animals and animal parts • Regulating and controlling logging, mining, and • Preventing consumption of raw animal products touristic activities to avoid disturbing wildlife and (for example, raw meat in restaurants, raw animal triggering contact and migrations. parts used in traditional medicine) • Increasing awareness of laws and penalties for Restricted proximity and mobility should help trade of high-risk species at places of sale (that is, reduce the likelihood of close encounters between posters at known wildlife markets). wildlife, livestock, and humans. Close interactions are more prone to facilitating pathogen transmission than iii. Creating barriers to species jumping and interactions at a distance. Interventions should include infections the following: Additionally, biosecurity measures should help reduce • Changing grazing practices from nomadic and free the risks of viral species jumping between wildlife, roaming to more sedentary management and livestock, or humans when they come in close contact indoors livestock systems where there are risks of infected hosts and in particular restricting spread with each other, directly or indirectly through body of livestock near tropical forests parts, fluids, or air. These would fall into three main • Configuring farms to minimize livestock exposure groups: sanitary, surveillance, and veterinary services. to humans, agricultural pests (for example, rodents and crows), neighboring livestock, and wildlife; Sanitary measures would reduce the survival rates of separating poultry and pigs; and creating sheltered viruses and other pathogens. These would include the spaces for livestock feed and water to avoid following: contamination from wildlife • Ensure that water sources are not contaminated by • Controlling the population of wild birds such as manure or slurry (stored and freshly spread); crows, seagulls, and pigeons and the presence of guarantee fresh drinking water for animals; dispose rodents such as rats and mice in domestic animal of farm waste products (bedding, slurry) away from farms livestock, humans, and watercourses. • Preventing mixing of domesticated and wild • Introduce licensing and monitoring of sanitary sourced or farmed wildlife animals in farms and practices, especially where high-risk species are discouraging or making more secure the backyard involved. farming system of semi-wild animals (for example, • Clean and disinfect clothing, equipment, and use fencing to avoid regular contact with humans) vehicles before and after contact with animals and and educating local communities on risky behaviors in particular ensure good hygiene for visits to places such as keeping primates as pets or destroying with high wild animal densities, in particular for bat-roosting areas high-risk species. Page 39 • Introduce and enforce health protocols for farm Surveillance should include the following: workers and promote good hygiene such as • Enforce animal health monitoring, especially for regularly brushing and disinfecting footwear, using notifiable and emerging diseases and implement personal protective equipment and strict sanitary regular animal health surveillance programs and measures wherever possible for all farm workers early warning systems to detect zoonotic and in particular for personnel in direct contact pathogens promptly. with livestock. • Conduct risk assessment at the wildlife-livestock- • Use adequate ventilation, waste disposal, and human interfaces to inform the type of emergency compartments inside farm systems to minimize response or longer-term planning for prevention disease transmission. and control of zoonotic pathogens. • Prevent symptomatic people from entering farms • Educate and train workers to identify early signs of (for example, people with flu-like symptoms on pig diseases and have action protocols when farms) and enforce a minimum symptom-free symptoms are detected. period before entering. • Enforce laws to ensure mandatory quarantine for • Minimize new animal introductions (both number exported and imported domestic and wild animals. of sources and individuals), check their health status, isolate them, and use separate equipment • Apply whole supply chain traceability methods (for for these animals before introducing them to others example, tamper-proof boxes, electronic tags, on the farm. physical marking with barcodes, blockchain) to show that shipments have undergone quarantine/ • Introduce licensing or certification systems for the pathogen detection/disinfection at each step. transport of live animals or animal parts, to ensure adherence to hygiene and welfare standards, • Introduce mandatory testing (additional screening introduce standards for regular checking and safe of viral pathogens in high-risk taxa or those utilized removal of dead or sick individuals during in high volumes) and health surveillance of species transport, including their bedding materials. considered for human consumption or use. • Introduce safe and hygienic animal handling and • Enable easy anonymous modes of reporting of slaughter standards (for example, for waste wildlife for sale or violations of restrictions in hygiene disposal of slaughter by-products), and train regulations, for example through reporting hotlines, licensed processors of both live animals and animal and make such rules widely visible; and build products (for example, tannery and wool enforcement capacity and forensic accounting ability industries). to audit marketplaces and vendors. • Introduce a licensing or certification system for any Veterinary measures refer to establishing animal vendor or processor of live animals or animal parts, health standards, enforcing or offering incentives for with regular checks and training, to ensure laws and standards are adhered to and that quotas can be implementation, and providing remedies when monitored. infections happen such as vaccines and treatment. • Ensure all places of sale that deal with live animals, These would include the following: raw meat, or animal parts have standardized • Develop and enforce minimum health, welfare, and sanitary control. sanitary standards for animal rearing that minimize • Introduce mandatory pasteurization methods for stress-induced immunosuppression and reduce the raw animal products from non-domesticated spread of disease into and among captive individuals. animals (for example, for traditional medicine or • Conduct gap analysis to estimate the additional food). resources needed to bring veterinary services at par with the zoonotic risks. Surveillance aims at detecting any dangerous virus or • Develop and implement a veterinary health plan. other pathogen as early as possible so that • Ensure regular veterinary visits and preventative care. remediation measures can be adopted before the • House sick animals in isolated areas. outbreaks become more difficult to control. Page 40 iv. Increasing productivity to reduce livestock nefarious impacts that extensive cattle rearing has on the numbers natural environment and greenhouse gas emissions (GHGEs). The result will be further upward pressures on Livestock production is exploding, driven by demand the magnitudes of poultry and pig stocks. from increasing population and incomes. On one hand, this will increase the number of potential hosts and Increasing consumption of poultry and pork is a need encounters between hosts and humans, calling for limiting for the world, as alternative sources of protein are or even reducing the numbers and densities of the animal less efficient (Figure 20). With more than 1 billion and human stock. On the other hand, one-third of the people undernourished, animal protein consumption world population is still malnourished, lacking a sufficient per capita needs to increase. The alternative of intake of proteins and nutrients. Animal protein will ruminant meat, which requires much more land and continue to have a critical role to play and will be in even other inputs per kg of meat produced triggers more demand in the future as an additional 2.7 billion deforestation and is more carbon intensive due to people yet to be born join at the table. This highlights a methane emissions, is just not feasible given the fundamental trade-off between limiting the number of burgeoning population. livestock, to rein in on epidemics and pandemics, while at the same time increasing the amount of animal protein in To address this trade-off between reducing the the food system to address food insecurity and number of livestock hosts on one hand and increasing malnutrition. The trade-off is particularly difficult for supply of protein from poultry and pork on the other poultry and pigs, which are two of the most important hand, the livestock sector needs to increase its sources of animal protein for the poorest quintiles of the productivity. Not all productivity measures are the populations. Compounding this challenge, consumers are same though. What matters here is to increase the being urged to shift meat consumption away from animal protein that reaches consumers while reducing ruminants and toward poultry and pork because of health the number of animals, since the number of animals issues—cardiovascular diseases from beef diets—and the drives the frequency of encounters with humans or FIGURE 19. Global meat estimates and projections – 1961–2050 Source: Adapted from Our World in Data. Database from FAOSTAT and Alexandratos and Bruinsma 2012 Page 41 FIGURE 20. Protein efficiency of dairy and meat production Source: Adapted from Our World in Data, based on Alexander et al 2016. wildlife and opportunities for pathogen transfer and animals per unit of land not by increasing the amount of species jumping. The productivity measure of interest protein from each animal. In fact, meat yields per animal is the animal protein that gets consumed (preferably by increased only by 30 percent for poultry and 25 undernourished people) at the end of the supply chain, percent for pigs in the last 60 years, much behind other per animal that enters the food supply chain, starting at productivity gains in food production (Figure 21). It is the farm. the animal numbers that matter for disease transmission, not how ‘fat’ they are, so it would have There are two ways of increasing productivity as been better to increase productivity by increasing the defined here. One is to increase the amount of animal kilograms of protein per animal rather than the number protein travelling through the supply chain and reaching of animals (Figure 21). consumers while reducing the number of animals. The other way one is to rein in on food loss and waste in the Fortunately, the technology to increase edible livestock supply chain. The number of animals is likely protein per animal is available. In some developed to increase anyway, driven by demand from population countries the size of an average chicken increased and income growth. However, it may increase less under fourfold since 1950. However, these gains were not a productivity growth scenario than in a business-as- widely shared. Or in other words, there is a large usual scenario. It is in this sense that the number of variation in animal protein productivity from very low to animals is less in a high productivity than in a low very high across the world. Therefore, the small average productivity world, productivity as defined above. gains in productivity per animal of only 30 percent (poultry) reflect the combined contributions from low One way to increase the productivity is to increase and high productivity farms. The good news is that animal size, which is likely to be closely correlated since the technology is available but not widely used, with protein content. Unfortunately, intensification of there are many opportunities for improvement. Breed, poultry and pigs appears to be, on average, feedstock, and animal health are all measures predominantly done by increasing the number of conducive to higher protein efficiency. Page 42 FIGURE 21. Productivity of meat production in poultry and pigs Poultry meat per animal, 1961-2020 Pigs meat per animal, 1961-2020 1.6 kg 80 kg World 1.4 kg 70 kg 1.2 kg 60 kg 1 kg 50 kg 0.8 kg 40 kg 0.6 kg 30 kg 0.4 kg 20 kg 0.2 kg 10 kg 0 kg 0 kg 1961 1970 1980 1990 2000 2010 2020 1961 1970 1980 1990 2000 2010 2020 Source: Food and Agriculture Organization of the United Nations OurWorldInData.org/meat-production • CC BY Source: Adapted from Our World in Data 2020, based on FAOSTAT 2020 Another path for reining in on ever-expanding market value, and food insecurity, to name a few. livestock numbers is to reduce food loss and waste in Regardless, the economic and social impacts of the animal supply chain. Livestock is one of the food livestock diseases have been recognized globally. A commodities most prone to food loss and waste as it focus on prevention implies quantifying the economic travels from the farm to the consumer. The percentage impacts of livestock disease outbreaks (Baratt et al. of livestock slaughtered, culled, or dying from disease is 2019). Table 12 shows an estimation of costs for huge, exceeding 40 percent in some countries. different diseases. Reducing disease to keep animals in the food supply chain would reduce the number of animals, as There are ample opportunities for reducing the producers would not need to maintain such large herds burden from domestic animal disease, thus helping to edge against disease loss. Or in other words, the keep more animal protein in the food supply chain, yields in animal protein would increase, since much eventually resulting in fewer animals. About 12 fewer would die before reaching consumers. percent of animals have infections with brucellosis, reducing production by 8 percent. About 10 percent Much of the food loss and waste (FLW) in the of livestock in Africa are infected with livestock supply chain reflect losses to animal trypanosomiasis, reducing production by 15 percent. diseases, which lead to death, culling, or About 7 percent of livestock are currently infected slaughtering and discarding of sick animals, with TB, reducing production by 6 percent (3–10 voluntarily or involuntarily. Reducing meat loss and percent of human TB cases may be caused by zoonotic waste along the supply chain would put more animal TB). About 17 percent of smallholder pigs show signs protein at the consumer’s dinner table without of current infection with cysticercosis, reducing their increasing the number of animals at the farm. Animal value and creating the enormous burden of human diseases represent threats not only to livestock but cysticercosis. About 27 percent of livestock show also to the environment, public health, and the signs of current or past infection with bacterial economy. There are many associated losses linked to food-borne disease, a major source of food animal diseases, such as increased mortality, reduced contamination and illness in people. About 26 percent productivity, control costs, loss in trade, decreased of livestock show signs of current or past infection Page 43 with leptospirosis, reducing production and acting as To highlight the likely effects of less food loss and a reservoir for infection. About 25 percent of livestock waste in the meat supply chain, we use a partial show signs of current or past infection with Q fever equilibrium model disaggregating eight stages of the and are a major source of infection of farmers and meat supply chain from farm to fork to landfill. These consumers. Endemic zoonoses, present in many places stages comprise farming, THS (transportation, handling, and affecting many people and animals, are and storing, such as in a barn or corral), processing (for responsible for the great majority of human cases of example, slaughtering), retailing, hospitality (hotels, illness and deaths as well as the greatest reduction in restaurants, institutions), home consumption, away from livestock production (Grace et al. 2012). home consumption, and disposal. The computational framework is described in World Bank (2020). We apply There is little comprehensive information on mortality this framework to the chicken subsector in the UK, since among smallholder pigs, but mortality is often high it is the only country with detailed data on food loss and among pre-weaned piglets in smallholder systems waste on a number of food commodities including (around one-fifth) and very high losses occur during chicken. The chicken loss rate in the UK is a staggering outbreaks of African swine fever and other epidemics 58 percent of animal meat produced, but this figure (Wabacha et al. 2004). Some of the annual livestock comprises both loss to disease and loss of edible body losses are due to noninfectious causes (mainly parts that might have been recovered. Most losses are at accidents, poisoning, predation, and malnutrition) the consumer level (38 percent of the tonnage of chicken while other losses are due to non-notifiable diseases that enters the supply chain) and at the processor level (such as endoparasites), but farmers and experts (22.5 percent). The modeling approach is flexible and agree that the 87 notifiable diseases are among the allows for different assumptions on the structure of the most important causes of mortality for livestock in economy, especially in its openness to trade and supply Africa. The authors of the report consider that at least and demand elasticities. We are interested in two results: 50 percent of mortality is attributable to notifiable (a) how a reduction in disease reflected in fewer animals diseases (Otte and Chilonda 2002). slaughtered and more animal meat reaching the consumer levels would affect policy indicators of interest How less loss and waste in the livestock sector such as farm production, farm welfare, consumer food would play throughout the supply chain is not security, and GHGEs and (b) how a shift in consumption evident. As more animal protein is put back into the away from meat would trickle down through the same food system, it will affect prices, demands, and policy indicators. This second aspect is discussed in the supplies of not only farmers and consumers but also next section; here the interest is on reducing losses due of all stages of the complex supply chain. For example, to animal disease. less loss and waste from less disease would make livestock rearing more profitable, and this could We first seek to understand how a reduction in food trigger larger herds and flocks, as producers shift to or loss and waste in the UK’s chicken processing might expand livestock farming to seize the greater income trickle throughout the chicken economy. We consider opportunities. This is in fact an empirical question, both a closed (no trade) and a small open chicken (trade and the reactions of the meat supply chain as costs, but no price setting) economy structure and report results supplies, demands, and prices change could go in any in Table 13 and Figure 22. We further note that the model direction. reports results in percentage change, and although the baseline values (in absolute terms) might differ greatly across countries, the percentage change is more robust and less influenced by country-level data. For example, Page 44 TABLE 12. Burden of zoonotic diseases Cost Human Human Livestock Livestock Sector Disease Time frame cases deaths cases deaths (US$) Rift Valley Livestock 2005–2022 5,228 987 77,064 19,257 Losses of 109 fever (cattle, (2000– (2000– (sheep, (sheep, million in 1998– sheep, and 2019) 2019) goats, cattle, goats, cattle, 1999 (Horn of goat camels, camels, Africa); 32 million production buffaloes, buffaloes, in losses in and trade) rabbits) rabbits) 2006–2007 (Kenya) Pandemic Livestock 2009–2010 60,000,000 151,700– 14,864 3,475 45–55 billion; influenza H1N1 (pig in the US 575,400 (pigs) (pigs) Losses in Mexico 2009 and production) alone globally to tourism (2.8 swine billion) and pork influenzas industry (27 million) Bovine TB Livestock 2005–2022 140,000 12,000 per 2,928,398 2,755,750 3 billion annually (cattle per year year (cattle, (cattle, (global) production sheep, sheep, and trade) goats, goats, camels, camels, swine, swine, rabbits) rabbits) Brucellosis Livestock 2005–2022 381 in EU — 3,504,062 2,767,651 3.43 billion (cattle, (2017) (buffaloes, (buffaloes, (in livestock sheep, goat, 57,222 in camels, camels, populations in and pig China cattle, cattle, India) production (2014) equids, equids, and trade) goats, goats, sheep, and sheep, and swine) swine) Nipah virus Agriculture 1998–2022 707a 412 Thousands >1 million 625 million (Livestock (pigs) and crops) Avian influenza Livestock 2005–2022 486 282 204,086,351 513,190,241 20+ billion from (high (poultry (poultry) (poultry) production losses pathogenicity production) due to H5N1 and low (2000s) pathogenic) MERS Livestock 2012–2022 2,585 931 207 0 12 billion (in (camel Republic of Korea) production and trade) Note: a. Human cases of Nipah virus include all events, including those not linked to livestock. Unless otherwise stated, livestock cases and deaths reflect officially reported numbers for the 2005–2021 reporting period (posted to the World Animal Health Information System [WAHIS] as of early April 2022). Livestock deaths include animal mortality from the disease itself, animals killed and disposed of for disease control, and livestock slaughtered for disease control. Page 45 two different countries may report production of 1 million Our second goal is to see if reducing food loss and and 100,000 metric tons of chicken production, waste in the chicken supply chain, not just by respectively; the baseline values and the model predict a reducing disease but also addressing other sources of decline of 10 percent in production for both countries. losses, would have the desired effects of reducing the This would translate in 100,000 chickens lost for the first number of chickens. For this we look at the combined country and 10,000 for the second. effects of cutting losses by 50 percent at all stages of the supply chain, although processing and consumption The processor level is where slaughtering takes place, are where most losses occur. For a closed economy so this would cover most chicken killed because of mimicking the world at large, farm production would disease (Figure 22). We consider a hypothetical decline decline by 12 percent implying a small number of in slaughtering losses of 50 percent because of chickens. Farm welfare would decline from lower sales diseased animals for illustration purposes only (Table and so would consumer prices, improving food security. 13). We dare to interpret the closed economy results as Home consumption would increase by 38 percent, thus mimicking the world at large, since for the world there is making more animal protein available from a small no external trade and the small open economy for a number of chickens, since fewer chickens and chicken country in the world that trades but whose trade is not parts are discarded. large enough to affect global prices. While we interpret those results assuming a closed Our first goal, from a pandemic risk perspective, is to economy such as the entire world, smaller open reduce the number of chickens at the farm level. We economies within the world might fare differently, therefore look at how a 50 percent reduction of chicken since trading is now possible (Table 14). The results losses at the processing (slaughtering) stage would will depend on whether the economy is a net exporter affect farm production (number of chickens). We confirm or importer of chicken or if it switches from exporter to that the impacts on production are in fact negative. As importer and vice versa. The UK is a net importer of more chickens reach the consumer at the end of the chicken. For a country like the UK, open to trade, supply chain, home consumption increases (7 percent), reducing food loss and waste of chicken at all stages of consumer prices decline (−8 percent), and farm sales the supply chain would increase production and sales, decline as well because slaughterhouses (processing) and therefore number of chickens, since domestic and consumers are buying fewer chickens, since fewer production would substitute for imports which would are being discarded by them. Thus, the size of flock at the decline by 67 percent. But since this would happen at farm level also declines, helping reduce the number of the expense of imports, which would decline, the chickens—one influencer of pandemic risk. FIGURE 22. Losses estimates in the poultry supply chain in the UK Production Transport, Processing Wholesale Consumers handling, and retail storage (THS) Chicken 58% 4.3% 1% 22.5% 7.6% 38% net loss rate Page 46 TABLE 13. Impacts of reducing food loss and waste TABLE 14. Impacts of reducing food loss and waste of chicken by 50 percent in processing of chicken by 50 percent at all stages of the (slaughtering) supply chain Closed Open Closed Small open Indicator Indicator economy (%) economy (%) economy (%) economy (%) Farm production −6 4 Farm production −12 3 Farm sales −6 4 Farm sales −10 6 Farm welfare −17 13 Farm welfare −35 12 Farm sales price −12 8 Farm sales price −26 5 Processor sales 7 2 Processor sales 4 −5 Home consumption 7 2 Home consumption 38 24 Consumer price −8 −2 Consumer price −16 −7 Imports n.a. −40 Imports n.a. −67 GHGEs −7 −11 GHGEs −15 −23 number of chickens would be less in the UK trading consumption. An increase in consumption of animal partners, since their exports to the UK would decline. protein can contribute both positively and negatively to nutrition and health. Animal protein contains nutrients The above results suggest that strategies to reduce with a high bioavailability that, especially in areas with the number of chickens by reducing disease (losses at low nutrient diversity, are key for addressing slaughtering or culling) or food loss and waste of both malnutrition. Overnourishment with animal protein, on entire chicken and body parts would help reduce the the other hand, is associated with obesity, cancer, and number of chickens. However, while this result might cardiovascular diseases, main causes of premature mimic a world where all countries reduce chicken loss death, loss of productivity and incomes, and health and waste by 50 percent, impacts are distributed based costs (Matena, L. S. 2019). on whether they are net exporters or net importers of chicken. In general, simulations of lower food loss and Underconsumption of animal protein prevails in the waste in other countries and for a range of food low-income quintiles of the population in Africa and commodities suggest that reducing food loss and waste Asia, whereas the highest overconsumption happens tends to boost the domestic economy and reduce the in North America, Europe, Oceania, and parts of Latin need for trade. Notably, less food loss and waste also America and among the better-off inhabitants in both reduces GHGEs, both in a closed and open economy— Africa and Asia. Matena, L. S. (2019), in a painstaking an environmental benefit. analysis, used National Recommended Diets from several countries as thresholds to estimate the amounts of v. Reducing excess meat consumption overconsumption and underconsumption of animal protein. Meat consumption above the recommended The explosive growth of livestock numbers, a source thresholds is considered luxury consumption. For the of pandemic risk, can also be tamed by influencing world at large and grouping all income classes, she found consumer diets. According to diet guidelines animal a global gap in consumption of dairy and eggs but an protein is important, but excess consumption, excess consumption of animal meat products. Gaps and particularly of beef, can have deleterious health excess are distributed differently. For example, while impacts. Meat consumption is both a necessity and a dairy and meat are consumed excessively in Europe, dairy luxury. Currently, livestock provides 18 percent of the seems to be less of a problem than meat in North calories and 25 percent of the protein for human Page 47 America. Overconsumption of meat is also a problem in surplus meat in the world to overcome the gap in meat Latin America. Dairy and eggs are, however, much below consumption that harms poorer people in poorer what they should be in Africa and parts of Asia. From a countries. The situation is different for dairy and eggs, pandemic point of view, these results suggest the need where eliminating luxury consumption would not be to shift animal protein production away from meat and sufficient to offer the rest of the world the dairy and toward dairy and eggs. The net effect is likely to be a eggs that they need. This suggests shifting some of the reduction in the global number of poultry and pigs, the stock of animals assigned to meat production toward more susceptible hosts for a dangerous virus, as well as dairy and eggs. While this shift might reduce the cattle, since fewer animals might be needed for dairy and number of animals, meeting the world’s current and eggs than for meat. upcoming need for dairy and eggs might require an increase in the number of animals. This increase would In the low-income countries and low-income be less, however, than if the same share of animals was quintiles, there is little luxury consumption. The kept in meat production for luxury consumption. luxury consumption in these countries is only in the higher-income classes and mostly for meat. In high- These results signal the opportunities for reducing income countries, almost all people are in the highest meat consumption, which might also lead to a smaller income classes and luxury consumption was observed number of animals. While this reduction might help for all commodities. The luxury consumption for meat reduce the risks of pandemics originating in domestic was the largest, followed by that of dairy and eggs animals, it would also bring substantial benefits in (Figure 24). reducing GHGEs and pressure on natural resources such as forests and water as well as in improving diets The gap in consumption for meat is smaller than the to reduce cardiovascular disease. While beef has the sum of luxury consumption and the consumption most deleterious impacts on health and environment, waste. Or in other words, were rich consumers to abide pigs and poultry compete with human beings for by recommended guidelines, there would be enough human-edible feed materials (Herrero et al. 2013; Van FIGURE 23. Animal protein consumption in relation to National Recommended Diets (NRDs) thresholds 200,000 Global consumption (kton/year) 150 ,0 00 100,000 Consumpti on waste Luxury consumption 50,0 0 0 Consumpti on wi thin NRDs - Low Low Low High High High Middle Middle Middle Aspi ring Aspi ring Aspi ring Dairy Eggs Meat Page 48 FIGURE 24. Deficit and excesses in animal protein consumption 100,000 50,000 Global food quantity (kton/year) 0 -50,000 Luxury consumption -100,000 -150,000 Consumption waste -200,000 -250,000 Gap in consumption -300,000 Net consumption -350,000 gap -400,000 -450,000 -500,000 Dairy Eggs Meat Zanten et al. 2016); thus, their numbers also harm food C. Estimating costs of on-farm biosecurity security. The cost of biosecurity is estimated using the cost of To assess the impacts of policies to reduce meat biosecurity per beneficiary (farmers, livestock consumption, we turn again to the food supply chain households) from World Bank projects in low-income model for chicken in the UK. Suppose a 20 percent tax and lower-middle-income countries. The cost of on poultry is applied (Table 15). The tax at the farm biosecurity is then calculated for hot spots (countries would reduce production by 4 percent while a tax on with a combined risk of human and livestock [chickens/ consumption would cut chicken production by 10 pigs] interaction) in low-income and lower-middle- percent. This would reduce farm production and income countries (Annex 3). Finally, the per farmer cost therefore number of animals. Chicken consumption is extrapolated to the total number of poor livestock would decline as prices increase. Farm welfare would keepers, based on data from the FAO Robinson et al. also decline. However, these results do not consider 2011.5 Bounds were estimated by adding and how the proceeds from the tax might be used. These subtracting 1 standard deviation from the point could be redistributed back to those affected, such as estimate. farmers and consumers, to reduce income losses. As long as the redistribution is not tied to output, it would The annual cost (expressed in 2019 purchasing power likely help shift production and consumption toward parity [PPP]) of biosecurity for both low-income and commodities less associated with pandemic risk. lower-middle-income countries is estimated at US$5 billion, over the next 10 years (Table 16). The cost of biosecurity varies significantly between low-income countries and lower-middle-income countries (US$0.2 5 The combined risk of human-livestock interaction is calculated by including countries with above 85th density percentile worldwide (250 per km2 for pigs; 7,536 per km2 for chicken; and 200 people per km2 for human population). Page 49 private good—or to other individuals in the same TABLE 15. Impacts on policy goals of taxes on chicken meat jurisdiction who do not get infected because the disease does not spread—a subnational, national, or regional 20% tax on public good. This reflects a lower capacity of the 20% tax at Policy objective consumption pathogen to spread. Prevention and control of highly the farm (%) (%) infectious viruses with pandemic potential is mostly a Farm production −4 −10 global public good. All humans benefit from not being (number of animals) infected, given the risk of a global infection. This is Farm welfare −12 −27 because an episode of infection in one country can easily Chicken consumption −4 −10 spread to other countries, as the experience with Consumer chicken price 5 12 COVID-19 shows. Virus infections tend to spread quickly GHGEs −4 −10 from its original location, for example, avian influenza, SARS, and MERS. In contrast for bacteria and fungi infections, although they can and do spread across billion and US$4.8 billion, respectively). Annex 3 shows borders, most impact is felt domestically. that countries categorized as high risk in Allen et al. (2017) are also hot spots for human-livestock While a viral pandemic is the scariest, other endemic interaction in most cases (with the exception of Bhutan, livestock pathogens cause large losses to farmers, in the Democratic Republic of Congo, and Kenya). Total particular poorer ones who may lack the means to cost of biosecurity and the cost per farmer is outlined adopt biosecurity measures. As discussed earlier, the in Annex 4. distribution of costs of zoonotic pathogens differs, with the richer world worrying about viral pandemics and D. External co-benefits of pandemic risk the less developed world preoccupied with endemic management diseases that may be taxing their livestock sector and While zoonotic pathogens are numerous, they also cause poverty. differ in economic impact. An important dimension of Fortunately, many of the actions to reduce viral risks economic impact is the extent of externalities an action are also measures that will help reduce the burden of generates. In some cases, zoonotic disease may not be diseases of nonviral pathogens, not only in human contagious beyond the animal keeper who is the only losses of health and lives but also incomes and one bearing losses. In other cases, contagium may be productivity. Or in other words, a program to reduce limited and slow and avoid epidemic proportions but the risk of viral zoonoses will bring large co-benefits by may spread to other animals or humans in the same reducing the burden of disease from nonviral sources. jurisdiction. And in the extreme case of pandemics, it These co-benefits are mostly in the form of national or may quickly spread across borders and continents or subnational public goods, accruing to jurisdictions in even the world as happened with COVID. the form of fewer infections, or of private goods, when These different levels of external costs translate into the owner or keeper of the infected livestock incurs different categories of prevention goods and services lower health and loss costs. according to their reach. Consequently, the type of good Reducing food loss and waste from disease and other or benefit from prevention and control is also different causes in meat supply chains will bring about for different pathogens. Control of bacteria and other dividends for national policy goals. A simulation with pathogens assumes mainly the form of private or national UK parameters, absent equivalent data from other public goods, in the sense that most benefits accrue to countries, and assuming a small open economy, the first individuals who are susceptible to infection at exemplifies how these co-benefits might be distributed. the farm level but who do not suffer from the infection—a Page 50 The global public good nature of pandemic risk TABLE 16. Annual cost of biosecurity—lower bounds, point estimates, and upper bounds—for reduction requires global financing of national animal low- and lower-middle-income countries in hot health systems, which is in the country’s interest as spots well because reducing viral risk will also decrease the burden of other endemic diseases that generate Cost (US$, billion 2019 Upper Point Lower domestic or private costs. It puts the onus on national PPP/year) bound estimate bound and subnational governments of assuming the costs of Low-income countries 0.4 0.2 0.5 some of the lockdown and biosecurity interventions. It Lower-middle-income 7.5 4.8 2.0 is in the interest of governments to improve veterinary countries Low-income and and animal health services. Most governments will face lower-middle-income 8.0 5.0 2.0 urgent competing needs that may limit their ability to countries invest in animal health, but since pandemic risk management could also reduce the impact of domestic pathogens, governments will have an incentive to adopt Farmer welfare improves from more sales, food security measures that benefit the world at large. also improves from lower chicken prices for consumers. Imports decline, which might be relevant for countries Grant financing for livestock keepers from global experiencing a trade imbalance or for those concerned sources is also necessary despite the private and with geopolitical dimensions and seeking to reduce domestic public good nature of improving animal dependency on untrustworthy supply chains due to health. This is for two reasons. First, most livestock war and COVID. Natural resource stress on land and keepers will lack the means to cover the costs of water increases if the reduction in chicken loss and biosecurity because of their poverty and exclusion from waste is at the farmer, transportation, and, especially, credit systems and would need targeted financing to processing levels, but this is compensated by cover the financial gap. Second, it is unlikely that a focus reductions at the retail and consumer levels. GHG on private benefits, that is, accruing to livestock emissions decline, a global public good. keepers, would generate the kind of investments needed to also address the risk of global pandemics. FIGURE 25. Co-benefits from less food loss and waste in the chicken supply chain (UK parameters) United Kingdom – Small Open Economy - Chicken Natural Total Food Farmer Welfare Food Security Imports Resource GHG Emissions Waste Stress 50% reduction at production 50% reduction Legend at THS Positive Impact < 5% 50% reduction Positive Impact ≥ 5% at processor Negative Impact < 5% 50% reduction at retail Negative Impact ≥ 5% Negligible Impact < 1% 50% reduction at consumer Direction of Impact Source: Adapted from World Bank 2020b Page 51 Global financing would therefore have two objectives, an between species; (c) to set up barriers to species economic efficiency objective and a distributional jumping; (d) to reduce loss of livestock from disease objective. Meeting the economic efficiency objective is and other causes throughout the food chain; and (e) to assuming its responsibility, for the global public good nature, influence diets away from meat, primarily luxury of reducing the risks of a global pandemic, which could harm consumption in developed, richer countries, and toward citizens in every country in the world and create widespread dairy and eggs in developing countries. economic loss, as COVID did. Meeting the distributional objective is offering resources, subsidized or on a cost- High-risk hot spots are the locations where livestock recovery basis, to underfunded national governments and and wildlife, wildlife and humans, or livestock and poor livestock keepers to enable them to take actions at human densities are highest. We suggest a mapping their levels for addressing endemic disease. These two approach to identify those hot spots where pairs of objectives imply alternative sources of financing as well as high density intersect. In addition, the mapping should the rules under which they should be delivered. consider the type of farming that is present. Where there is a higher density of poor livestock keepers, it is E. Discussion less likely that biosecurity measures have been put in place. A mapping approach is feasible since digital In this report we deliberately make some strong technologies and satellite imagery permit a large scale assumptions and propose hypotheses on how to of disaggregation, down to areas of 5 km2 or less. A mitigate the risks of a future pandemic of zoonotic mapping approach based on densities is also easy to origin infecting through the food system. To make the understand, interpret, and explain to policy makers with challenge more manageable, we recommend focusing little background on virology and epidemiology, helping on viral pathogens from wild bats, primates, rodents, or with buy-in and action from the leadership, more than wildfowl that will infect humans through an compensating for errors from less precision. We intermediate livestock species, pigs, or poultry. We also illustrate the application of the mapping approach to postulate that detecting viruses with the ability to the world at large but warn that for meaningful policy infect humans is too much of a challenge to be input it would need to be carried out at the country addressed quickly and broadly. Not minimizing the level. We conclude that high-risk zones are Southeast importance of detection work, we recommend focusing and South Asia (in particular India and Bangladesh), on limiting the opportunities for the viruses to jump Indonesia (Sumatra, Java, and Sulawesi), and parts of species. We also postulate that species jumping will be West and East Africa. While most pigs are concentrated easier under high densities of wildlife, livestock, and in Southeastern China, overlapping with high human humans in a given location, which will increase the densities, the lower density of poor livestock keepers frequency and proximity of encounters between makes it more likely that at least some biosecurity species where jumping can take place. We recognize measures are being followed, reducing infection risk. that there are drivers triggering the increase in animal and human densities, primarily human population To reduce the frequency and proximity of encounters growth, expansion of urbanization, human migration, between wildlife, livestock, and humans, we propose meat prices, income growth, land use changes, measures equivalent to COVID lockdowns, now deforestation, hunting, and animal trade. including the animal realm. Encounters and proximity are to be addressed through enclosures for both We propose a five-pronged strategy to mitigate the livestock and wildlife, changes in farm design to risks of zoonotic viruses jumping species and separate species, progression from migratory and eventually reaching humans. The five components are nomadic livestock rearing toward more sedentary (a) to focus on high-risk hot spots locations; (b) to forms, zoning to separate animal from human reduce the frequency and proximity of encounters Page 52 concentrations, and appropriate barriers and length of due to trade effects, so that although the overall result transportation to limit encounters en route. is a decline in numbers, some countries might see an increase in numbers of animals due to increased To make species jumping more difficult when close exports or substitution for imports. contact happens, we propose biosecurity measures that can create barriers to stop or destroy the Finally, we turn to the challenge of reducing meat viruses. These include measures equivalent to some consumption and number of animals, by influencing adopted under COVID: personal protective equipment diets. We note that 30 percent of meat consumption is such as masks and suits; handwashing and sanitizing, in luxury consumption, that is, consumption that is above particular for humans in close contact with animals the nationally recommended diet guidelines. This luxury such as livestock keepers; disinfection and ventilation; consumption is concentrated in the rich world, although testing; surveillance; and vaccination. It would also rich consumers in poor countries also share the burden. imply biosecurity measures throughout the rest of the In the poorer world and poorer income groups, there is supply chain, in particular in transportation, a deficit of meat consumption instead, and an even slaughterhouses, and wet markets. larger deficit in dairy and eggs, which are more protein efficient. These results suggest increasing consumption Cognizant that the world needs animal protein to of diary and eggs by poorer people in poorer countries address rampant undernutrition and future demands by promoting this subsector while reducing meat and from increasing populations, we look for ways to dairy luxury consumption in the richer world. Shifting generate more animal protein from a smaller number animals from meat to dairy and eggs would likely reduce of animals. Here we propose a focus on increasing the number of live animals. Slowing luxury meat livestock productivity, defined as the amount of edible consumption in richer countries would greatly reduce animal protein that is consumed by humans per animals the number of animals farmed. produced at the farm. We argue that the strategy should focus on reducing food loss and waste in the We extrapolate the costs of the second and third livestock supply chain. We further suggest that there components of the strategy: enclosures and are two types of food loss and waste to consider. One is biosecurity. We use the study ‘World Bank (2012)’ and the loss due to animal diseases which are apply a rate of inflation to update the costs of predominantly caused by nonviral outbreaks. Losses of biosecurity. Using FAO data and the hot spots livestock to disease are huge, exceeding 40 percent of identified, we suggest about 1 million small and all meat produced in most countries and causing backyard farms should have the means to set up farmers to increase the sizes of herds and flocks, enclosures to limit opportunities and proximity of therefore animal numbers and densities, to edge encounters. We estimate that US$1.5 billion would be against risk from disease. We suggest that most actions needed with a life span of 10 years. The total costs of to address risks of virus pandemics, such as on-farm upgrading biosecurity and enclosures in high-risk spots biosecurity, will also help dampen the risks of nonviral would therefore amount to about US$450 million a diseases that kill domestic animals. The other one is to year. reduce other meat food loss and waste, not because of disease but because of wasteful management We argue that there are several important co- approaches that may discard edible body parts. We benefits from the proposed five-pronged strategy; suggest, through a partial equilibrium model using UK first reducing viral risks also reduces nonviral risk data for chicken, that reducing meat loss and waste in which is the predominant cause of animal loss and the world as a whole would in fact reduce the number burden of disease affecting mainly poorer people in of animals but that this reduction would comprise developing countries. For example, while average combined increases and declines in different countries deaths per year due to COVID amount to about 3 Page 53 million people, 2.5 million people on average die from And finally, reducing luxury meat diets and shifting nonviral zoonotic, especially the poorer people in the herds and flocks toward dairy and eggs would not poorer countries. only improve the health of the human population but also cut down on the disturbing negative effects that Second, setting up enclosures to limit encounters livestock has on nature. Deforestation and biodiversity would signal property rights where enclosures such loss would decline. Feedstock that currently is used for as fences coincide with property boundaries. While poultry and pigs would be made more available for this should be avoided where there are disputes, World human consumption, helping rein in on food security. Bank experience with land titling projects suggests disputes in only 10 percent of properties. More secure In conclusion, the One Health approach to mitigating property rights would enable more investment at the pandemic risk has many components that contribute farm, both in biosecurity and in productivity. to good national development. The fact that the world at large, fearing another pandemic like COVID, is Third, increasing animal protein efficiency or interested in reducing this risk by putting in financial productivity would make more animal protein resources is also an opportunity for countries to available for the poorer countries and segments of practice good development that contributes to growth, society without increasing the number of animals. incomes, and the well-being of the poorest sectors of The burden of disease associated with malnutrition society. The challenge for national governments is to would decline and labor productivity and incomes bring together the numerous institutions in charge of would increase. the many components of a One Health risk mitigating strategy, with each adopting and implementing the Fourth, reducing food loss and waste in the supply strategy as its own. It is critical that Ministers of chain would have positive environmental Finance or Planning be given a central role in externalities. It would reduce the emissions from coordinating the diverse mix of institutions, since it decomposition of organic matter as well as pressures extends far beyond a focus on human and animal on biodiversity, forests, water, or oceans since less health. production is necessary. Page 54 What elements should be considered for the development of an investment framework? This section explores how to finance forest conservation, landscape restoration, and sustainable food systems transformation to prevent the emergence of infectious diseases and reduce pandemic risk. Financing pandemic prevention will require both more funding and a 4 in pandemic prevention even after the current crisis has passed, breaking the cycle of panic and neglect. Funding to address pandemics has so far been grossly inadequate, fragmented, overly reliant on ad hoc contributions, misdirected, and far too focused on concerted effort to reduce the drivers of deforestation and unsustainable food systems through structural response. Compared to the cost of managing economic changes to more sustainable production and pandemics, relatively modest investments in prevention consumption practices. The actions recommended in capacities can improve the resilience of health systems, this section will help address crises beyond pandemics, make investments in preparedness and response more particularly nature loss and climate change. successful, drastically reduce the need for response, and lessen the broader economic and social impacts of This financing approach aims to prevent the emergence pandemics. of infectious diseases at the source and complements the need to strengthen preparedness—as a capacity to Governments, and Ministries of Finance (MoFs) in respond and mobilize funds swiftly to meet the surge in particular, can use their policy, regulation, planning, and times of need during an outbreak before it becomes a budgeting levers to drive structural economic reform that pandemic. It also complements the need to improve the reduces economic drivers of deforestation and resilience of health systems and their capacity to unsustainable food production. One Health finance is deliver in times of crises. The proposed approach aligns compatible with and reinforces existing agendas. For with the 2014 Independent Evaluation Group (IEG) example, the draft of the post-2020 global biodiversity recommendation to follow “the example of the Bank’s framework, which is expected to be adopted at the successful shift in approach to natural disasters, moving fifteenth meeting of the Conference of the Parties from a responsive approach using emergency (COP-15) to the Convention on Biological Diversity (CBD) instruments to one that favors preemptive risk in 2022, calls for urgent action to transform the economic, reduction and risk management through regular social, and financial models so that the trends that have country programs and operations.” exacerbated nature loss stabilize by 2030 and allow for the recovery of natural ecosystems in the following 20 One Health finance is defined as the practice of raising, years. Implementing this requires considering future risks directing, and managing capital and using financial and associated with nature loss—including pandemics—and economic policy and regulation to reduce health risks at systematically accounting for the value of nature in the human-animal-ecosystem interface. It aims to reduce decisions at all levels and across all sectors. pressure on public health systems through primary prevention and to improve global health security. This The proposed framework includes mobilizing domestic approach will enable governments to strategically invest resources and overseas development assistance and strategically deploying public resources to maximize Page 55 public goods. Additionally, the One Health Financing pandemic prevention in budget allocation processes Framework provides recommendations for approaches (WBG 2017). to private capital mobilization. As long as financial flows in a country’s economy are not aligned with One Health A. Domestic resource mobilization outcomes, no number of public resources will be sufficient to counter the impact of economic activity To contribute to developing resilient domestic finance driving pandemic risk. The enabling environment for prevention, One Health Taxes could help (economic and financial policy and regulation as well as governments finance One Health prevention, real sector policy and regulation) must therefore be surveillance, and risk assessment. Earmarked taxes harnessed to reduce the drivers of pandemic risk and to facilitate a link between activities that contribute to align financial flows with forest conservation, landscape pandemic risk and those that mitigate that risk and will restoration, and food systems transformation. help achieve greater allocative efficiency. They can also create a connection between use of, or pollution of, global public goods and activities to maintain and 4.1 MOBILIZING PUBLIC RESOURCES restore those goods (Costanza et al. 2021). Tax policies The G20 Joint Finance and Health Task Force (2022) should be accompanied by efforts to build greater estimated the annual financing need for the future awareness of the risks of infectious disease outbreaks Pandemic Preparedness and Response (PPR) system at among private sector leaders. In addition to stimulating US$31.1 billion, consistent with the estimate of the G20 companies to improve their own internal prevention High-Level Independent Panel (2021). Considering current measures, such awareness building could make and expected domestic and international financing for business leaders less resistant to taxes or regulations PPR, it is estimated that at least an additional US$10.5 related to reinforcing pandemic prevention and more billion per year in international financing will be needed to inclined to work with governments to mitigate the risks fund a fit-for-purpose PPR architecture. A high-level (WBG 2017). estimate of overall PPR needs and gaps amounts to an annual investment of US$4.7 billion. To fill these funding Considering the risk related to spillover from domestic gaps, governments will need to mobilize domestic animals raised for food production, a national One resources, strategically deploy official development Health Tax could be levied through the meat value assistance (ODA), and identify strategic public chains in the 33 countries classified at high risk of investments to maximize public prevention benefits, in zoonotic diseases in the global hot spot map of addition to mobilizing private investment and phasing out zoonotic EID risk index (Allen et al. 2017). The tax can economic activities working against prevention goals. be partly conceived as Meat Tax or Animal Protein Tax, but One Health Taxes should be tailored to the With an estimated domestic financing gap of US$1.7 externality of the risk of zoonotic spillover and not to billion for pandemic prevention financing,6 there is a carbon emissions. Ideally, the tax could be applied on need to increase domestic resources through a mix of the value chain of meat across one or more of the four fiscal policy instruments. For many countries, financing groups classification proposed earlier (zoonotic prevention through the domestic public sector budget pathogens from wildlife, zoonotic pathogens from can guarantee sustained funding. Yet, this requires domestic [livestock] animals, drug-resistant pathogens, ensuring sufficient priority is attached to investing in or vector-borne pathogens).7 6 “Analysis of Pandemic Preparedness and Response (PPR) architecture, financing needs, gaps and mechanisms.” Prepared for the G20 Joint Finance & Health Task Force. Paper prepared by the World Health Organization & the World Bank. March 13, 2020. 7 Taking stock of the announcements of meat taxes at the United Nations Climate Change Conference (UNFCCC) COP26 and the current circumstantial limitations from the spike in global food prices, the proposal for a global One Health Tax could be accompanied by further analytics. At a later stage, such a tax could be linked to the wider global costs associated with a national emergence of zoonotic diseases, and the additional climate and health co-benefits could be outlined. Page 56 Select countries to pilot the tax could be identified revenue. Feebates involve a system of fees and rebates considering risk hot spots, impact on inequality, other applied to landowners according to a basic formula that constraints, and historical context. For example, in promotes forest carbon storage, as proposed in high-risk contexts like Vietnam, past interventions Designing Fiscal Instrument for Sustainable Forests reduced negative environmental impact and improved (2021). Investing in One Health Feebates could be food safety along the pig and poultry value chains while associated with climate finance co-benefits. The One significantly improving incomes of the supported Health methodology on deforestation puts forward an farmers (Livestock Competitiveness and Food Safety estimate of the costs of avoided deforestation at Project; LIFSAP-Vietnam). In such countries, a fiscal national and subnational scales, for each of the 33 levy to reinforce prevention and safety could potentially countries at high risk of zoonotic spillover. Assuming be adopted with less opposition. Similarly, in Tanzania, deforestation is reduced by 50 percent, the cost of the first country in the WHO African region to develop avoided deforestation in countries at high risk of a costed National Action Plan for Health Security, a spillovers is between US$8.68 and US$28.5 billion (see public awareness campaign included lobbying for methodology, attachments). sustained and adequate domestic funding for prevention (WBG 2017). Taking stock of the current Subsidies for ecological forest management and limitations due to the spike in food price, at a later regenerative agricultural policies can provide incentives stage, a global One Health Tax could be conceived as to timber companies and agricultural producers to linked to the wider global costs associated with the transition to One Health-aligned practices. The national emergence of zoonotic diseases rather than government may implement payments for ecosystem the national externalities. services scheme to pay private landowners for the services their investments generate. The government The experience of selected taxes on timber extraction, could also provide technical assistance support to water extraction, or fertilizer or pesticide uses, and landowners and companies that want to transition their other soil and water pollution shows the potential to practices (United States Department of Agriculture – create disincentives to economic activity, which USDA). generate revenue and also contribute to forest health. For example, Costa Rica’s Payments for Environmental B. Multilateral development banks/Official Services (PES) program pays private landowners for development assistance forest maintenance and recovery activities and received In addition to the mobilization of domestic resources, 89 percent of its funding from a fossil fuel tax and 7.5 there is a key role for multilateral development banks percent from a water tax in 2018 (Costanza et al. 2021). (MDBs) and ODA to support investment in the three The program is credited with substantially contributing key pandemic prevention activities discussed. to the halt and reversal of deforestation in the country. Multilateral and bilateral development partners can Funds raised through taxes on timber and water can incentivize investments in pandemic prevention at the also be earmarked for prevention activities. Such taxes country level through various financial instruments, are currently, and have been, used historically to control including through increasing concessionality of lending unsustainable resource consumption. Denmark has for pandemic prevention (for example, World Bank rate levied taxes on fertilizer and pesticides since the buydowns) and increasing grants. The World Bank’s 1990s.8 International Development Association (IDA) is Fees and charges for use of, or resources provision uniquely positioned to support the poorest countries, from, public lands are another potential source of and its new policy commitment aims to strengthen 8 These policies need to be aligned with real sector regulation that limits timber extraction in old growth forests and key biodiversity areas and regulation that phases out use of the most harmful chemical fertilizers and pesticides. Page 57 health security by improving pandemic preparedness Following countries’ commitments to set aside 30 and prevention at the nexus of human, animal, and percent of their territorial lands and waters by 2030, ecosystem health, including zoonotic diseases and MDBs could support countries to identify priority areas antimicrobial resistance, in at least 20 IDA countries. for conservation and restoration and to provide much needed funding to execute these activities. Additionally, One Health finance can be catalyzed through existing with a global public goods mandate, MDBs could play a global trust funds for long-term and gap financing role in protecting the oceans—the health of which is solutions (for example, Global Environment Facility, critical for terrestrial ecosystems and contributes to Global Climate Facility, Climate Investment Funds, pandemic prevention. FoodSystems2030, Health Emergency Preparedness and Response Umbrella Trust Fund (HEPRTF), REDD, C. Financing One Health and health security UN MPTF for biodiversity for health and pandemic prevention, PROGREEN, PROBLUE, Global Program for Governments and MoFs should allocate specific Sustainability, and the Global Program on Nature Based resources to catalyze, sustain, and monitor the national Solutions). Trust funds have played a key role in One Health financing architecture, an exercise supporting analytic work, policy dialogue, and co- inextricably associated with the effort to improve and financing World Bank projects. track health security expenditures and improve global health security. For many countries, development assistance can act as an effective catalyst to influence policy reform and Governments should support voluntary best practice implementation of programs aligned with the three key initiatives (for example, Joint External Evaluation prevention actions: forest conservation, landscape alliance and Global Health Security Agenda), and restoration, and food systems transformation. investments should be tied to performance indicators Financing should encourage countries to close any and the One Health Joint Plan of Action of the identified gaps and be tied to performance indicators Quadripartite. and the One Health Joint Plan of Action of the Quadripartite.9 This is a global and local effort that should include support and monitoring of national action plans for While MDBs already have extensive portfolios of lending health security—country-owned, multiyear, planning and advisory projects supporting forest conservation, processes that are based on a One Health, all-hazards, landscape restoration, and food systems transformation, whole-of-government approach. The plans capture the G20 High-Level Independent Panel has national priorities for health security and support the recommended that the mandates of MDBs are adjusted implementation of International Health Regulations so that they are empowered to invest in the provision of capacities. global public goods. No country is incentivized to invest in global public goods to provide critical ecosystem services (such as disease prevention) to the world at the socially optimal level, when a large proportion of the benefits of these ecosystem services are captured by citizens in other countries. The difficulty for governments to prioritize environmental investments over social investments is particularly clear in low-income countries or countries that still have high rates of poverty. 9 WHO, the World Organisation for Animal Health (WOAH, founded as OIE), FAO, and the United Nations Environment Programme (UNEP). Page 58 TABLE 17. One Health financing Tool Strategies Advantages Risks and disadvantages New finance Taxes Fiscal policy to support One Can be tailored to respond to a Enforcement; relies on tax Health finance and policy national externality discipline and collection objectives and incentivize governments with national Can contribute to discourage Potential domestic fiscal commitments, including unsafe practices in high-risk opposition the One Health Tax countries Can raise new resources Tariffs and trade policy Incentives to support One Can potentially support a Enforcement Health finance and policy sanctioning mechanism for objectives underinvesting and high-risk, Volatility in multilateral trade countries10 policy Debt Lending operations (including Independent source of Requires collateral and long-term concessional financing revenue stream lending, for example, Development Policy Operations - DPOs, IDA loans) Available to all member Repayment risk to be screened for One Heatlh countries in good standing risks during preparation11 Bonds (for example, bonds Independent source of issuance for agriculture financing finance suppliers, and agribusinesses, agriculture investment funds, sustainability-linked bonds) International Monetary Fund Independent source of Health conditionality rapid financing facilities financing (triggers) (Resilience and Sustainability Trust loans) Crisis Response Window One Health prevention window Only IDA-eligible countries (CRW) early response could be added: pre-allocated financing12 Contingent Emergency Response Component (CERC) CRW in Project Appraisal CRW experience in health Document (PAD) could include crisis limited that project shows medium- term contribution to building Health conditionality resilience and prevention (triggers) 10 Options are being considered, including on the model of the Cross-Border Adjustment Mechanism. Cooperation in selected areas of international animal trade depends upon geopolitical opportunities and must respect cultural traditions and local livelihoods 11 Policy commitment on One Health is already a requirement for all IDA operations. IDA’s regional window has been a useful instrument to strengthen PPR in general at the country level and through regional institutions and to support cross-country collaboration. 12 As an exception to CRW rules, used for health emergencies since IDA 17. Page 59 Table 17 continued Tool Strategies Advantages Risks and disadvantages Grants Financial support to projects Increases the financial Availability is limited and (for example, Global attractiveness of selected amount not likely to cover Environment Facility (GEF), projects that might otherwise high investment costs Global Agriculture and Food not be economically feasible Security Program (GAFSP), ProClean, grants to small businesses for investing in nature). Gap funding (for example, Gap Potentially geared toward Fund for Urban Development) specific technical innovation, global and regional reach, integrating public health and veterinary science IDA co-financing grants that Short-term accommodation for encourage use of countries’ building capacity core IDA allocation for One Health Both repurposing (with no direct fiscal implications) or new finance Payments for ecosystem Payments for conservation Increases returns of Relies on local institutions, services programs efforts, tree planting, investments in conservation implementation, and improved agricultural and restorations enforcement capacities management, and so on Supports upstream environmental investments for prevention Public-private Financial and policy support Flexible model accommodates Direct lending by partnerships for targeted investments multiple instruments development banks may be required Agribusiness insurance and Gaps remain in the financial lending products infrastructure in many countries, including for Voluntary best practice Further support to existing agribusiness; risk of benefits initiatives voluntary best practice accruing to larger private initiatives (for example, JEE players rather than alliance and GHSA) smallholders Space for new ad hoc voluntary Potential risk of standards best practice initiatives, proliferation including around domestic wildlife trade,13 congruent with existing initiatives 13 Options are being considered stemming from the increased agreement that existing wildlife trade regulations, such as Convention on International Trade in Endangered Species (CITES) were not designed specifically for disease risk reduction Page 60 4.2 PUBLIC INVESTMENTS TO MOBILIZE They can also set up natural capital or regenerative agriculture business incubator or accelerator programs. PRIVATE INVESTMENT This could even be done through a national green bank Given the significant funding gap outlined in the section or development bank. Additionally, governments can above, it is critical that governments not only make give these public banks a policy mandate to invest a strategic investments in the provision of public goods certain amount of their funding in mobilizing private but also make strategic public investments to mobilize investment in prevention activities. private investment in pandemic prevention. Public investments can play an important role in co-funding, A One Health fund could be set up to support farmers de-risking, and funding the development of business in the transition to sustainable or regenerative cases, including through providing technical assistance agricultural practices. For efficiency, funding could be to strengthen business models. channeled through rural development programs, as the EU has through the European Agricultural Fund for The private sector is becoming more aware of Rural Development. opportunities to invest in projects contributing to forest conservation, landscape restoration, and food One Health perspectives can continue to accompany systems transformation. Investors are developing their reflections on insurance solutions and risk pooling, own nature business incubator and accelerator including in the process of developing post-disaster programs and are making pledges for investment in actions plans geared toward prevention. The process of these assets. developing such post-disaster actions plans and identifying related costs can also generate risk One Health can provide unique integrated animal health information and create incentives to step up data information to model risk factors and support investments in prevention and adaptation to reduce private initiatives (see Box 3). Governments can risks in the first place. support private investment in pandemic prevention activities by providing blended finance, better It has been shown that insurance companies can biophysical and economic data on natural capital, and potentially play a significant role in stimulating targeted funds to provide grants to small businesses investment in preparedness, but more work is needed contributing to prevention. on prevention. The most powerful benefits of insurance from a risk reduction perspective continue to be in (a) the additional insights into risk drivers and mitigants, BOX 3. Rationale for private sector One Health solutions provide rigorous integrated animal health data insurance solutions information to model risk factors; improved safety standards can provide triggers for private insurance providers, targeted at the national and subnational levels, in the agricultural value chain. The policies and guidelines as well as risk-based surveillance, inspection, monitoring and preventive regulatory protocols along the value chains could substantially diminish the risk of infections. At the same time, on the model of food system index insurance, the safety standards could themselves create incentives to reduce moral hazard and offer a business proposition with true value to private insurers. Page 61 generated by insurance providers and (b) the incentives risk that have been discussed in this report. created for governments and private sector companies, Governments can do this through realigning which will be incentivized to take action to reduce the expenditures, implementing real sector policy and risks and thus the premiums (WBG 2017). regulation to directly influence private sector behavior, and implementing financial sector policy and regulation Sustainability and nature-related data provision can be to change where capital flows in the economy. achieved through encouraging or mandating corporate sustainability reporting for corporations over a certain The widespread degradation of nature, one of the key size (see Box 4) and through natural capital accounting drivers of pandemics, is the result of an array of indirect and the provision of cost-benefit analysis and modeling socioeconomic drivers of change that have accelerated tools and analysis. Governments can ensure that in recent decades. To effectively manage pandemic- national data are accessible to the public and in a related risks, development strategies and national plans usable format—particularly for subnational need to articulate a whole-of-economy approach to governments and the private sector to consider in their reverse nature loss and transform food systems in an planning, operations, and decision-making and also for integrated way. The potential scale of tail risks from citizens who play a crucial role in ensuring good pandemic-related loss could be significant enough to environmental quality in their communities. Having justify policy action to mitigate nature-related risks, spatial Necessary Condition Analysis data that can be even without full quantification of impacts (Kedward, disaggregated is crucial to enabling such application. Ryan-Collins, and Chenet 2020). The National Governments can also encourage corporate natural Biodiversity Strategies and Action Plans (NBSAPs), capital accounting and offer training and access to tools which will be updated following the adoption of the to private companies or nongovernmental organizations post-2020 global biodiversity framework, and the interested in using national NCA data or developing Nationally Determined Contributions (NDCs) under the their own accounts. Paris Agreement offer an opportunity to lay out national strategies, plans, and programs for forest Governments can also make strategic anchor conservation, landscape restoration, and food systems investments in protected areas that provide ecosystem transformation across different economic sectors. services that create opportunities for private investment in the surrounding landscape or seascape. While there are many synergies between pandemic For example, a public investment in a protected forest prevention, nature conservation and restoration, and can enable the ecotourism, hunting, and agricultural climate action, there are also some trade-offs, which sectors to invest around the protected areas and should be carefully managed. For example, food engage in sustainable economic activity that monetizes systems transformation to reduce pandemic risk might the ecosystem services provided by the forest and involve the development of food transportation contributes to their sustained provision. infrastructure in the global south to reduce food loss and waste. However, the development of such 4.3 ENABLING ENVIRONMENT FOR A infrastructure could lead to nature loss, including of ONE HEALTH-ALIGNED ECONOMY forests, and could contribute to pandemic risk. The development of clean energy infrastructure and mass Governments must help drive structural changes that transportation is likely to lead to increased mining, create an enabling environment for a One Health- which could also lead to further nature loss. Holistic aligned economy. Governments have a role to play in planning and analysis, including considering risks that structuring markets that drive capital away from cannot be usefully quantified, should be pursued to destructive and extractive activity and toward nature- drive efficient national planning. positive activity to reduce the key drivers of pandemic Page 62 It is important for governments to be aware of, and financing provides windows that are automatically strategically manage, trade-offs associated with the triggered to provide swift, scaled-up access to funds. time differences between the costs and benefits of CRW early response financing involves preparedness prevention activities. Some benefits from preventative plans before the emergence of a crisis but greater action might only materialize over the medium term emphasis on prevention is still lacking. while political costs and financial costs are felt more quickly. Ensuring a just, equitable, smooth transition to A specific effort in communication and awareness a One Health-aligned economy will be critical to raising should include influencing external financing building and maintaining political support for this instruments (for example, GEF, GAFSP). One Health agenda. Furthermore, there is a need for global succeeded in combating river blindness during the cooperation on governance and financial support for 1970s and avian influenza during the 2000s, and more the conservation and restoration of global public good can be achieved if effective communication on ecosystems such as rainforests, as all countries depend prevention is accounted for. on them. Trade policy, while not covered in this section, New instruments could potentially leverage prevention is also a critical tool that governments can use to align objectives, such as the green and sustainability labeled their economies with One Health goals. bonds, a tool that governments are increasingly using to fund nature-related investment activities. One example A. Realign expenditures is Chile that recently became the first country to issue a Governments and MoFs should be aware of new sovereign sustainability-linked bond (SLB) tied to financing opportunities made available or prepared by meeting its emissions reduction targets. A range of the International Financial Institutions. nature-related key performance indicators (KPIs) could be aligned to One Health principles and used for In this direction, the World Bank Group (WBG) has sovereign SLBs. Building on the ‘SLB principles’ set out supported policy and institutional reforms related to by the International Capital Market Association, a the PPR agenda, at several levels. First, through DPOs World Bank report provides a framework for assessing and Catastrophe-Deferred Drawdown Options (CAT- the suitability of nature-related KPIs.14 DDOs), and although an explicit focus on disease outbreak preparedness in DPOs remains rare, a growing Public, philanthropic, and private sector investors take number of DPOs and CAT-DDOs focus on climate on many forms, from institutional investors to funds, to change, deforestation, and crisis and risk management local banks and microfinance institutes. It is only by capacity, and hence support PPR indirectly by involving them all from the onset that a truly holistic addressing both drivers of disease outbreaks and financing approach can be developed for One Health. response capacity. One Health can make DPOs more Many of the projects eligible for biodiversity or climate effective and help them grow, across a wide range of finance lack proven business cases and need grants or themes such as climate change mitigation and donations to create these business cases after which adaptation, green growth, natural resource they are equipped to raise funds from investors seeking management, disaster risk management, forestry, water a return. This is what is often referred to as integrated resource management, environmental policy, and capital or blended finance—making use of public or others. philanthropic funding to catalyze private sector investment in sustainable development. New fast-tracked surge financing from the IFIs in response to a pandemic has been launched and could be part of One Health. The CRW early response 14 The metrics include whether a potential indicator is sufficiently robust, properly interpreted, aligned with the country context, and credibly ambitious. Page 63 FIGURE 26. Financing policies for One Health Financing Green Increase return by better monetizing ecosystem services Positive Conservation project Conservation project Project where BES are Impact on biodiversity and ecosystem services impact on without cashflow which using program-related monetized, or co-benefits requires grant funding or investment or blended are created through BES public support; Develop- finance business activities ment project with biodiversity component No impact Development project Development project Project with proper risk with negative cashflow using program-related mitigation practices or Greening on BES which does not affect investment or blended biodiversity offset Finance BES with proper risk finance with proper risk mitigation practices mitigation practices Increase positive impact Development project Development project Traditional investment by better Negative internalizing without proper risk using program-related without proper risk impact on mitigation practices investment or blended mitigation practices environmental BES finance without proper and social risks risk mitigation practices and benefits Negative Below Market Rate Market Rate Concessional Commercial Risk-Adjusted Return Source: World Bank 2020a. Note: BES = Biodiversity and ecosystem services. The figure assumes that projects comply with national environmental regulations. Standards refer to International Finance Corporation (IFC) performance standards or other widely accepted market standards. B. Real sector policies and regulations rules. Real sector policies correct externalities and other market failures, thus leveling the playing field on Real sector policies directly influence the behavior of which private actors operate (World Bank 2020a). firms that operate in productive sectors. These policies are critical because they change the incentives of the C. Financial sector policies and regulations industries and value chains that are driving the loss and degradation of ecosystems. They include (a) standards To mobilize private investment in pandemic prevention, and regulations, often termed command and control governments should take steps to drive both ‘financing policies; (b) pricing policies, which include fiscal policies green’15 and ‘greening finance’16 within their countries, governing taxes and subsidies, fees, and payment for aligned with One Health objectives. ecosystem services; and (c) information disclosure 15 Financing green is increasing financial flows to projects that contribute—or intend to contribute—to the conservation, sustainable use, and restoration of biodiversity and ecosystems and their services to people. 16 Greening finance is directing financial flows away from projects with negative impacts on biodiversity and ecosystems to projects that mitigate negative impact and/or pursue positive environmental impacts as co-benefits. Page 64 BOX 4. Sustainability reporting Following widespread adoption of the Task Force on Climate-related Financial Disclosures (TCFD) four pillars framework, climate-related reporting has become mandatory in a growing number of countries. One Health can provide unique value to the ongoing efforts to create both sovereign and corporate sustainability reporting. The International Financial Reporting Standards (IFRS) Foundation’s International Sustainability Standards Board (ISSB) has recently developed prototype climate and sustainability disclosure standards that it hopes will be adopted as baseline standards globally. As companies better assess and manage pandemic prevention risks and opportunities, markets can drive more sustainable behavior, corporations will be incentivized to invest in more sustainable production and operations—contributing to forest conservation, landscape restoration, and food systems transformation. One Health is crucial for both corporate and sovereign reporting. There is also progress on sovereign sustainability reporting, which could help bolster corporate sustainability reporting by providing critical contextual information. Following a call by the World Bank for the International Public Sector Accounting Standards Board (IPSASB) to develop guidance for sovereign sustainability reporting in its report launched in January 2022, Sovereign Climate and Nature Reporting: Proposal for a Risks and Opportunities Disclosure Framework, IPSASB launched a consultation report in May 2022, Advancing Public Sector Sustainability Reporting. Both reports lay out the case for sovereigns reporting on their climate and nature-related risks and opportunities and how they are managing them to sovereign lenders. Pandemic prevention measures could become a key component of the information sovereigns disclose to their investors in this reporting. France has already introduced mandatory reporting on financial risks related to biodiversity loss as well as dependencies on and impacts to biodiversity. From 2023, the EU sustainability reporting framework requires reporting on alignment with all six environmental objectives of the EU taxonomy, including the protection and restoration of biodiversity and ecosystems. One Health risk analysis integrates national natural Financial institutions’ net-zero commitments also capital accounting principles. Governments can encourage greater investment in natural capital support investment by financial institutions in projects businesses and projects, and governments have a role contributing to conservation, restoration, and to play in increasing the impact and transparency of sustainable use of nature by providing blended finance these commitments. Ministries of Finance can also and better biophysical and economic data on natural encourage and support the alignment of financial capital. This can be done by encouraging or mandating instruments with international principles and standards disclosure through natural capital accounting and the to ensure the transparency and credibility of green and provision of cost-benefit analysis and modeling tools. sustainable-labeled instruments. Page 65 References Chua, K. B., B. H. Chua, and C. W. Wang. 2002. “Anthropogenic Deforestation, El Niño and the Emergence of Nipah Virus in Malaysia.” Malays J Pathol 24 (1): 15–21. Allen, T., K. Murray, C. Zambrana-Torrelio, S. Morse, C. Rondinini, and M. Di Marco, N. Breit, K.J. Olival, and P. Costanza, R., Paul W.B. Atkins, Marcello Hernandez-Blanco, Daszak. 2017. “Global Hotspots and Correlates of Emerging and Ida Kubiszewski. 2021. “Common Asset Trusts to Zoonotic Diseases.” Nature Communications 8: 1–10 https:// Effectively Steward Natural Capital and Ecosystem Services doi.org/10.1038/s41467-017-00923-8. at Multiple Scales.” Journal of Environmental Management 280: 111801. Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050: the 2012 revision (Vol. 12, No. 3). FAO, De Nardi, M. , A. Hill, S. von Dobschuetz, O. Munoz, R. Rome: ESA Working paper. Available at: http://www.fao.org/ Kosmider, T. Dewe, K. Harris, G. Freidl, K. Stevens, K. van der docrep/016/ap106e/ap106e.pdf Meulen, K.D.C. Stäerk, A. Breed, A. Meijer, M. Koopmans, A. Havelaar, S. van der Werf, J. Banks, B. Wieland, K. van Reeth, Alexander et al. (2016). Human appropriation of land for food: G. Dauphin, I. Capua, the FLURISK consortium. 2014. the role of diet. Global Environmental Change. “Development of a risk assessment methodological framework for potentially pandemic influenza strains Baratt, Alyson, Karl M. Rich, Jude I. Eze, Thibaud Porphyre, (FLURISK)”. EFSA Supporting Publications: Volume 11, Issue George J. Gunn, and Alistar W. Stott. 2019. “Framework for 5. URL: https://doi.org/10.2903/sp.efsa.2014.EN-571 Estimating Indirect Costs in Animal Health Using Time Series Analysis.” Vet. Sci. 6: 190. https://doi.org/10.3389/ Dobson, A., S. Pimm, L. Hannah, L. Kaufman, J. Ahumada, A. fvets.2019.00190. Ando, et al. 2020. “Ecology and Economics for Pandemic Prevention.” Science 80 (369): 379–381 Bastin, J. F., Y. Finegold, C. Garcia, D. Mollicone, M. Rezende, D. Routh, C. M. Zohner, and T. W. Crowther. 2019. “The Global Evenson, R. E., and K. O. Fuglie. 2010. “Technology Capital: Tree Restoration Potential.” Science 365 (6448): 76–79. The Price of Admission to the Growth Club.” Journal of https://www.science.org/doi/10.1126/science.aax0848. Productivity Analysis 33: 173–90. Bernstein, A. S., A. W. Ando, T. Loch-Temzelides, M. M. Vale, B. GHS Index. 2019. “Global Health Security Index.” https:// V. Li, H. Li, ... and A. P. Dobson. 2022. “The Costs and Benefits www.ghsindex.org/. of Primary Prevention of Zoonotic Pandemics.” Science Advances 8 (5): eabl4183. Gibb, R., L. M. Moses, D. W. Redding, and K. E. Jones. 2017. “Understanding the Cryptic Nature of Lassa Fever in West Brauburger, K., A. J. Hume, E. Mühlberger, and J. Olejnik. 2012. Africa.” Pathog Glob Health 111 (6): 276–288. doi:10.1080/2 “Forty-Five Years of Marburg Virus Research. Viruses 4 (10): 0477724.2017.1369643. 1878–1927. doi:10.3390/v4101878. Grace D., F. Mutua, P. Ochungo, R. Kruska, K. Jones, L. Brierley, Buchhorn, M., M. Lesiv, N.-E. Tsendbazar, M. Herold, L. Bertels, L. Lapar, M. Said, M. Herrero, PM Phuc, NB Thao, I. Akuku and and B. Smets. 2020. “Copernicus Global Land Cover Layers— F. Ogutu. 2012. “Mapping of poverty and likely zoonoses Collection 2.” Remote Sens 12: 1044. https://doi. hotspots.” Zoonoses Project 4. Report to the UK Department org/10.3390/rs12061044. for International Development. Nairobi, Kenya: ILRI Busch, J., and J. Engelmann. 2017. “Cost-Effectiveness of Grotto, E., and D. Ricci. 2015. “Identification and Analysis of Reducing Emissions from Tropical Deforestation, 2016– the Main Drivers for Ebola Virus Spillover.” EFSA Journal 12 2050.” Environmental Research Letters 13: 015001. (7): 860E. https://efsa.onlinelibrary.wiley.com/ doi/10.2903/sp.efsa.2015.EN-860. 11th International Symposium on Veterinary Epidemiology and Economics, Cairns, Australia. New Zealand Veterinary Hansen, M. C., P. V. Potapov, R. Moore, M. Hancher, S. A. Association, Wellington, New Zealand. Turubanova, A. Tyukavina, D. Thau, S. V. Stehman, S. J. Goetz, T. R. Loveland, A. Kommareddy, A. Egorov, L. Chini, C. O. CDC (Centers for Disease Control and Prevention). 2009. Justice, and J. R. G. Townshend. 2013. “High-Resolution “Origin of 2009 H1N1 Flu (Swine Flu): Questions and Global Maps of 21st-Century Forest Cover Change.” Science Answers.” https://www.cdc.gov/h1n1flu/information_h1n1_ 342: 850–53. http://earthenginepartners.appspot.com/ virus_qa.htm. science-2013-global-forest. CDC. 2019. “Lassa Fever.” https://www.cdc.gov/vhf/lassa/ index.html. Page 66 Herrero, M. and P. Thornton. 2013. “Livestock and global Kedward, K., J. Ryan-Collins, and H. Chenet. 2020. “Managing change: Emerging issues for sustainable food systems”. Nature-related Financial Risks: A Precautionary Policy Edited by W.C. Clark, Harvard University. Approach for Central Banks and Financial Supervisors.” Working Paper Series (IIPP WP 2020-09), UCL Institute for Herrero, M. P. Havlík, H. Valin, A. Notenbaert, M.C. Rufino,  Innovation and Public Purpose. https://www.ucl.ac.uk/ P.K. Thornton, M. Blümmel, F. Weiss, D. Grace, and M. bartlett/public-purpose/wp2020-09. Obersteiner. 2013. “Biomass use, production, feed efficiencies, and greenhouse gas emissions from global Kessler, M. K., D. J. Becker, A. J. Peel, et al. 2018. “Changing livestock systems”. Edited by W.C. Clark, Harvard University. Resource Landscapes and Spillover of Henipaviruses.” Ann N https://doi.org/10.1073/pnas.1308149110 Y Acad Sci 1429 (1): 78–99. doi:10.1111/nyas.13910. Hewson, J., S. C. Crema, M. González-Roglich, K. Tabor, and C. Koh, L. P., Y. Zeng, T. V. Sarira, et al. 2021. “Carbon A. Harvey. 2019. “New 1 km Resolution Datasets of Global Prospecting in Tropical Forests for Climate Change and Regional Risks of Tree Cover Loss.”  Land 8: 14. https:// Mitigation.” Nat Commun 12: 1271. https://doi.org/10.1038/ doi.org/10.3390/land8010014. s41467-021-21560-2. Hu, B., L.-P. Zeng, X.-L. Yang, X.-Y. Ge, W. Zhang, B. Li, et al. Leroy, E. M., P. Rouquet, P. Formenty, S. Souquière, A. 2017. “Discovery of a Rich Gene Pool of Bat SARS-Related Kilbourne, J. M. Froment, M. Bermejo, S. Smit, W. Karesh, R. Coronaviruses Provides New Insights into the Origin of SARS Swanepoel, S. R. Zaki, and P. E. Rollin. 2004. “Multiple Ebola Coronavirus.” PLoS Pathog 13 (11): e1006698. https://doi. Virus Transmission Events and Rapid Decline of Central org/10.1371/journal.ppat.1006698. African Wildlife.” Science 303 (5656): 387–90. doi:10.1126/ science.1092528. Hui, D. S., E. I. Azhar, Y. J. Kim, Z. A. Memish, M. D. Oh, and A. Zumla. 2018. “Middle East Respiratory Syndrome Loh, E. H., K. J. Olival, C. Zambrana-Torrelio, T. L. Bogich, C. K. Coronavirus: Risk Factors and Determinants of Primary, Johnson, J. A. K. Mazet, and P. Daszak, et al. 2015. “Targeting Household, and Nosocomial Transmission.” Lancet Infect Dis Transmission Pathways for Emerging Zoonotic Disease 18 (8): e217–e227. doi:10.1016/S1473-3099(18)30127-0. Surveillance and Control.” Vector Borne and Zoonotic Diseases 15 (7): 432–437. IEG (Independent Evaluation Group). 2014. Responding to Global Public Bads: Learning from Evaluation of the World Ma, W., R. E. Kahn, and J. A. Richt. 2009. “The Pig as a Mixing Bank Experience with Avian Influenza 2006-13. Washington, Vessel for Influenza Viruses: Human and Veterinary DC: World Bank. Implications.” J Mol Genet Med 3 (1): 158–166. IOM (Institute of Medicine (US) Committee on Emerging Marani, M., G. G. Katul, W. K. Pan, and A. J. Parolari. 2021. Microbial Threats to Health), J. Lederberg, R. E. Shope, S. C. “Intensity and Frequency of Extreme Novel Epidemics.” Oaks Jr, eds. 1992. Emerging Infections: Microbial Threats to Proceedings of the National Academy of Sciences 118 (35): Health in the United States. Washington, DC: National e2105482118. Academies Press. Matena, L. S. 2019. The Contribution of Animal Source Food IOM, M. S. Smolinski, M. A. Hamburg, J. Lederberg, eds. 2003. to Food Security. Internship Report. World Bank. Microbial Threats to Health: Emergence, Detection, and Response. Washington, DC: National Academies Press. McKee, C. D., A. Islam, S. P. Luby, et al. 2021. “The Ecology of Nipah Virus in Bangladesh: A Nexus of Land-Use Change and Jenkins, C. N., S. L. Pimm, and L. N. Joppa. 2013. “Global Opportunistic Feeding Behavior in Bats.” Viruses 13 (2): 169. Patterns of Terrestrial Vertebrate Diversity and doi:10.3390/v13020169. Conservation.” PNAS 110 (28): E2602-E2610. doi:10.1073/ pnas.1302251110. Mills, J. N., A. N. Alim, M. L. Bunning, et al. 2009. “Nipah Virus Infection in Dogs, Malaysia, 1999.” Emerg Infect Dis 15 (6): Jones, K. E., N. G. Patel, M. A. Levy, A. Storeygard, D.Balk, J. L. 950–952. doi:10.3201/eid1506.080453. Gittleman, and P. Daszak. 2008. “Global Trends in Emerging Infectious Diseases.” Nature 451 (7181): 990–993. https:// Morand, S. 2020. “Emerging Diseases, Livestock Expansion doi.org/10.1038/nature06536. and Biodiversity Loss Are Positively Related at Global Scale.” Biological Conservation 248: 108707. doi:10.1016/j. Karesh, W. B., and R. A. Cook. 2005. “The Human-Animal biocon.2020.108707. Link, One World—One Health.” Foreign Aff 84: 38–50. Page 67 Morand, S., and C. Lajaunie. 2021. “Outbreaks of Vector- Rich, K. M., and F. Wanyoike. 2010. “An Assessment of the Borne and Zoonotic Diseases Are Associated with Changes in Regional and National Socio- Economic Impacts of the 2007 Forest Cover and Oil Palm Expansion at Global Scale.” Rift Valley Fever Outbreak in Kenya.” Am. J. Trop. Med. Hyg. Frontiers in Veterinary Science 8: 230. doi:10.3389/ 83 (Suppl. 2): 52–57. doi:10.4269/ajtmh.2010.09-0291. fvets.2021.661063. Roser, Max, and Ritchie Hannah. 2013. “Food Supply.” Nelson, A., D. J. Weiss, J. van Etten, et al. 2019. “A Suite of OurWorldInData.org. https://ourworldindata.org/food- Global Accessibility Indicators.” Sci Data 6: 266. https://doi. supply. org/10.1038/s41597-019-0265-5. Rulli, M., M. Santini, D. Hayman, et al. 2017. “The Nexus Newbold, T., L. Hudson, A. Arnell, et al. 2016. “Dataset: Global between Forest Fragmentation in Africa and Ebola Virus Map of the Biodiversity Intactness Index, from Newbold et Disease Outbreaks.” Sci Rep 7: 41613. https://doi. al., 2016, Science.” Natural History Museum Data Portal. org/10.1038/srep41613. https://doi.org/10.5519/0009936. Semenza, J. C., E. Lindgren, L. Balkanyi, et al. 2016. OIE (World Organisation for Animal Health). 2019. “Update “Determinants and Drivers of Infectious Disease Threat January 2019: Questions & Answers on Middle East Events in Europe.” Emerg Infect Dis 22 (4): 581–589. Respiratory Syndrome Coronavirus (MERS-CoV).” https:// doi:10.3201/eid2204. www.oie.int/app/ uploads/2021/03/q-a-mers-cov-en-update-jan2019.pdf. SEPAL. “Methodology.” https://docs.sepal.io/en/latest/ modules/dwn/seplan.html?highlight=se.plan. Olivero, J., J. E. Fa, R. Real, et al. 2017. “Recent Loss of Closed Forests Is Associated with Ebola Virus Disease Skowron, K., J. Bauza-Kaszewska, K. Grudlewska-Buda, N. Outbreaks.” Sci Rep 7: 14291. https://doi.org/10.1038/ Wiktorczyk-Kapischke, M. Zacharski, Z. Bernaciak, and E. s41598-017-14727-9. Gospodarek-Komkowska. 2022. “Nipah Virus-Another Threat from the World of Zoonotic Viruses.” Front Microbiol 12: Otte, M.J. and Chilonda, P. 2002. “Cattle and Small Ruminant 811157. doi:10.3389/fmicb.2021.811157. Production Systems in Sub-Saharan Africa. A Systematic Review.” Smith K., M. Goldberg, S. Rosenthal, L. Carlson, J. Chen, C. Chen, and S. Ramachandran. 2014. “Global Rise in Human Paton, N. I., Y. S. Leo, S. R. Zaki, A. P. Auchus, K. E. Lee, A. E. Infectious Diseases Outbreaks.” Journal of the Royal Society Ling, S. K. Chew, B. Ang, P. E. Rollin, T. Umapathi, I. Sng, C. C. Interface 11: 20140950. https://doi.org/10.1098/ Lee, E. Lim, and T. G. Ksiazek. 1999. “Outbreak of Nipah-Virus rsif.2014.0950. Infection among Abattoir Workers in Singapore.” Lancet 354 (9186): 1253–6. doi:10.1016/S0140-6736(99)04379-2. Staal S., J. Poole, I. Baltenweck, J. Mwacharo, A. Notenbaert, T. Randolph, W. Thorpe, J. Nzuma and M. Herrero. 2009. Petrovan S.O., D.C. Aldridge,H. Bartlett,A.J. Bladon,H. “Strategic investment in livestock development as a vehicle Booth,S. Broad,D.M. Broom,N.D. Burgess, S. Cleaveland, A.A. for rural livelihoods.” Bill and Melinda Gates Foundation – ILRI Cunningham, M. Ferri, A. Hinsley, F. Hua, A.C. Hughes, K. Knowledge Generation Project Report. International Jones, M. Kelly, G. Mayes, M. Radakovic, C.A. Ugwu, N. Uddin, Livestock Research Institute, Nairobi, Kenya, 78pp D. Veríssimo, C. Walzer, T.B. White, J.L. Wood, W.J. Sutherland. 2021. “Post-COVID-19: A solution scan of Stephens, P. R., N. Gottdenker, A. M. Schatz, J. P. Schmidt, and options for preventing future zoonotic epidemics”. Wiley J. M. Drake. 2021. “Characteristics of the 100 Largest Modern Online Library, Biological reviews, Vol. 96, Issue 6, Pg. Zoonotic Disease Outbreaks.” Philosophical Transactions of 2694-2715. https://doi.org/10.1111/brv.12774 the Royal Society B 376 (1837): 20200535. Pimm, S. L., C. N. Jenkins, R. Abell, T. M. Brooks, J. L. Sun, L. H. 2018. “On a Bat’s Wing and a Prayer.” https://www. Gittleman, L. N. Joppa, P. H. Raven, C. M. Roberts, and J. O. washingtonpost.com/news/national/wp/2018/12/13/ Sexton. 2014. “The Biodiversity of Species and Their Rates of feature/these-bats-carry-the-lethal-marburg-virus-and- Extinction, Distribution, and Protection.” Science 344 scientists-are-tracking-them-to-try-to-stop-its-spread/. (6187): 1246752. Taylor, L. H., S. M. Latham, and M. E. J. Woolhouse. 2001. “Risk Factors for Human Disease Emergence.” Philos Trans R Soc Lond B Biol Sci 356: 983–89. https://pubmed.ncbi.nlm.nih. gov/11516376/ Page 68 Van Zanten, H.H.E., B.G. Meerburg, P. Bikker, M. Herrero and Xu, R. H., J. F. He, M. R. Evans, et al. 2004. “Epidemiologic I.J.M. de Boer. 2016. “Opinion paper: The role of livestock in a Clues to SARS Origin in China.” Emerg Infect Dis 10 (6): sustainable diet: A land-use perspective.” Animal Volume 10, 1030–1037. doi:10.3201/eid1006.030852. Issue 4, 2016. Pg. 547-549. https://doi.org/10.1017/ S1751731115002694 Yuen, K. Y., N. S. Fraser, J. Henning, K. Halpin, J. S. Gibson, L. Betzien, and A. J. Stewart. 2021. “Hendra Virus: Epidemiology Von Braun, Afsana, Fresco, Hassan, and Torero. 2021. “Food Dynamics in Relation to Climate Change, Diagnostic Tests, Systems – Definition, Concept and Application for the UN and Control Measures.” One Health 12: 100207. doi:10.1016/j. Food Systems Summit.” https://knowledge4policy.ec. onehlt.2020.100207. europa.eu/sites/default/files/food_systems_concept_ paper_scientific_group_-_draft_oct_261.pdf. Wabacha, J.K., J.M. Maribei, C. Mulei, M.N. Kyule, K.H. Zessin, W. Kosura. 2004. “Health and production measures for smallholder pig production in Kikuyu Division, central Kenya.” Preventive Veterinary Medicice 63(3-4): 197-210. https:// doi.org/10.1016/j.prevetmed.2004.02.006 Wacharapluesadee, S., S. Ghai, P. Duengkae, et al. 2021. “Two Decades of One Health Surveillance of Nipah Virus in Thailand.” One Health Outlook 3: 12. https://doi. org/10.1186/s42522-021-00044-9. WHO (World Health Organization). 2003. Consensus Document on the Epidemiology of Severe Acute Respiratory Syndrome (SARS). https://apps.who.int/iris/ handle/10665/70863. WHO. 2019. Global Preparedness Monitoring Board. A World at Risk: Annual Report on Global Preparedness for Health Emergencies. Geneva: World Health Organization. WHO. 2021. Ebola Virus Disease. https://www.who.int/ news-room/fact-sheets/detail/ebola-virus-disease. World Bank. 2012. “People, Pathogens and our Planet. Volume 2. The Economics of One Health”. World Bank. 2018. Operational Framework for Strengthening Human, Animal, and Environmental Public Health Systems at Their Interface. https://documents1. worldbank.org/curated/en/703711517234402168/ pdf/123023-REVISED-PUBLIC-World-Bank-One-Health- Framework-2018.pdf. World Bank, 2020a. Mobilizing Private Finance for Nature. World Bank. 2020b. Addressing Food Loss and Waste : A Global Problem with Local Solutions. World Bank, Washington, DC. © World Bank. https://openknowledge. worldbank.org/handle/10986/34521 License: CC BY 3.0 IGO World Bank. 2021a. The Changing Wealth of Nations 2021: Managing Assets for the Future. Washington, DC: World Bank. doi:10.1596/978-1-4648-1590-4. Page 69 Annexes resistant pathogens, and vector-borne pathogens), and model the spatial variation of the events. To model the potential risk of disease emergence, the four groups are ANNEX 1: CLASSIFICATION OF EID RISK regressed as a function of human population density There are two key papers that classify EIDs. Jones et al. and growth, latitude, rainfall, and mammalian wildlife (2008) analyze a database of 335 EID events (origins species richness. of EIDs) between 1940 and 2004. Allen et al. (2017) According to the study, zoonoses dominate EID events builds on Jones et al. (2008), improving the predictor (60.3 percent of EIDs). In addition, pathogens with a set to better target underlying mechanisms specifically wildlife origin caused 71.8 percent of these zoonotic for EID events for wildlife zoonoses through 2008 (n = events—for example, the emergence of Nipah virus in 224, with n = 147 ultimately analyzed based on available Perak, Malaysia, and SARS-Coronavirus in Guangdong covariate information for 1970–2008). Province, China. The study suggests that EID origins are Jones et al. (2008) collected biological, temporal, and significantly associated with socioeconomic, spatial data on human EID events from the literature of environmental, and ecological drivers. In addition, the 1940 to 2004 and analyzed them. The authors divide authors found that disease emergence is substantially a the EID events into four groups, one of which is EIDs of product of anthropogenic and demographic changes wildlife origin, which are responsible for nearly all recent and is a hidden cost of human economic development. pandemics (the others being EIDs from animals other Furthermore, mammalian wildlife host species richness than wildlife – for example, domestic animals), drug- (with main groups of hosts being non-human primates, FIGURE 27. Global trends in EIDs Source: Jones et al. 2008. Note: Maps are derived for EID events caused by (a) zoonotic pathogens from wildlife, (b) zoonotic pathogens from animals other than wildlife, (c) drug-resistant pathogens, and (d) vector-borne pathogens. Page 70 rodents, and bats) is a significant predictor of the EIDs from the last few decades and are responsible for emergence of zoonotic EIDs with a wildlife origin, with Ebola, MERS, and almost every recent pandemic. no role for human population growth, latitude, or rainfall. Finally, the study provides a basis for developing Allen et al. (2017) developed a spatial model to describe a predictive model for emerging disease hot spots. the global spatial patterns of zoonotic EID events. They find that the risk of disease emergence is elevated in According to Allen et al. (2017), the work by Jones et al. tropical forest regions experiencing anthropic land use (2008) has some limitations. They argue that the lack changes related to agricultural practices and where of specificity of the predictors limits the paper’s mammal biodiversity is high. Like Jones et al. (2008), mechanistic inference. Therefore, Allen et al. (2017) they found that EID events are observed predominantly claimed that improving the predictor set to better in developed countries, where surveillance is strongest target underlying mechanisms could improve model (reporting bias). However, they also found that the performance and the ability to explain spatial variation predicted risk is higher in tropical, developing countries. in EID risk. To that end, the authors updated a global database of EID events; created a novel measure of Allen et al. (2017) presented a new global hot spot map reporting effort; and created regression tree models to of spatial variation in their zoonotic EID risk index. analyze the demographic, environmental, and biological According to this, the authors classified countries into correlates of their occurrence. They intended to better three categories of emerging zoonotic diseases: low analyze the mechanistic underpinnings of disease risk, medium risk, and high risk (Figure 28), classifying emergence for zoonotic EID events of wildlife origin, 33 countries as having a high risk of emerging zoonotic while addressing some methodological limitations of diseases, 56 countries as medium risk, and 41 countries Jones et al. (2008). The authors focus on these EID as low risk. events, which constitute the majority of high-impact FIGURE 28. Risk of zoonotic diseases by country Source: Allen et al. 2017. Page 71 ANNEX 2: HECTARES OF ANNUAL High deforestation AVOIDED FOREST LOSS BY 2030 IF ISO Spillover risk risk (> 0.7), 50 PERCENT OF DEFORESTATION IS Country code category biodiversity AVOIDED density, and accessibility Angola AGO Low risk 48,668 High Azerbaijan AZE Low risk 0 deforestation Belarus BLR Low risk 0 ISO Spillover risk risk (> 0.7), Country Bolivia BOL Low risk 49,576 code category biodiversity density, and Brunei accessibility BRN Low risk Darussalam 355 Bangladesh BGD High risk 0 Bulgaria BGR Low risk 0 Belgium BEL High risk 0 Burkina Faso BFA Low risk 6,128 Bhutan BTN High risk 1 Canada CAN Low risk 0 Burundi BDI High risk 9,169 Chile CHL Low risk 0 Cambodia KHM High risk 38,866 Costa Rica CRI Low risk 3,489 China CHN High risk 34,112 Djibouti DJI Low risk 0 Congo, Dem. Rep. COD High risk 12,757 Estonia EST Low risk 0 Egypt, Arab Rep. EGY High risk 0 Eswatini SWZ Low risk 984 Ethiopia ETH High risk 9,910 Gabon GAB Low risk 740 Germany DEU High risk 0 Gambia, The GMB Low risk 1,051 Ghana GHA High risk 198,306 Georgia GEO Low risk 0 Haiti HTI High risk 0 Greece GRC Low risk 0 India IND High risk 884 Guinea GIN Low risk 74,552 Indonesia IDN High risk 415,672 Guinea-Bissau GNB Low risk 23,856 Italy ITA High risk 0 Ireland IRL Low risk 0 Japan JPN High risk 0 Kazakhstan KAZ Low risk 0 Kenya KEN High risk 70,905 Latvia LVA Low risk 0 Korea, Rep. KOR High risk 0 Lesotho LSO Low risk 0 Malawi MWI High risk 118,813 Liberia LBR Low risk 673 Malaysia MYS High risk 43,128 Lithuania LTU Low risk 0 Myanmar MMR High risk 50,399 Moldova MDA Low risk 0 Nepal NPL High risk 103 Mongolia MNG Low risk 0 Netherlands NLD High risk 0 New Zealand NZL Low risk 0 Nigeria NGA High risk 204,538 Oman OMN Low risk 0 Pakistan PAK High risk 0 Papua New PNG Low risk Philippines PHL High risk 6,177 Guinea 0 Rwanda RWA High risk 12,682 Paraguay PRY Low risk 42,649 Singapore SGP High risk 3 Romania ROU Low risk 0 Sri Lanka LKA High risk 0 Serbia SRB Low risk 0 Thailand THA High risk 91,877 Sierra Leone SLE Low risk 1,675 Uganda UGA High risk 90,728 Somalia SOM Low risk 0 United Kingdom GBR High risk 0 Sweden SWE Low risk 0 Vietnam VNM High risk 57,962 Turkmenistan TKM Low risk 0 Page 72 High High deforestation deforestation ISO Spillover risk risk (> 0.7), ISO Spillover risk risk (> 0.7), Country Country code category biodiversity code category biodiversity density, and density, and accessibility accessibility Ukraine UKR Low risk 0 Nicaragua NIC Medium risk 24,760 United Arab Niger NER Medium risk 0 ARE Low risk Emirates 0 Peru PER Medium risk 10,587 Uruguay URY Low risk 0 Poland POL Medium risk 0 Uzbekistan UZB Low risk 0 Portugal PRT Medium risk 0 Afghanistan AFG Medium risk 0 Russian RUS Medium risk Algeria DZA Medium risk 0 Federation 0 Argentina ARG Medium risk 1,437 Saudi Arabia SAU Medium risk 0 Australia AUS Medium risk 0 Senegal SEN Medium risk 3,689 Austria AUT Medium risk 0 Slovak Republic SVK Medium risk 0 Benin BEN Medium risk 124,589 South Africa ZAF Medium risk 84,027 Brazil BRA Medium risk 1,510,000 Spain ESP Medium risk 0 Cameroon CMR Medium risk 1,867 Sudan SDN Medium risk 0 Colombia COL Medium risk 42,743 Syrian Arab SYR Medium risk Côte d’Ivoire CIV Medium risk 188,939 Republic 0 Croatia HRV Medium risk 0 Tajikistan TJK Medium risk 0 Cuba CUB Medium risk 0 Tanzania TZA Medium risk 145,869 Czech Republic CZE Medium risk 0 Timor-Leste TLS Medium risk 0 Dominican Togo TGO Medium risk 70,170 DOM Medium risk Republic 0 Tunisia TUN Medium risk 0 Ecuador ECU Medium risk 61,490 Turkey TUR Medium risk 0 El Salvador SLV Medium risk 5,646 United States USA Medium risk 0 Eritrea ERI Medium risk 0 Venezuela, RB VEN Medium risk 213,467 France FRA Medium risk 0 Yemen, Rep. YEM Medium risk 0 Guatemala GTM Medium risk 11,959 Zimbabwe ZWE Medium risk 191,206 Honduras HND Medium risk 11,457 Albania ALB Unclassified 0 Iran, Islamic Rep. IRN Medium risk 0 American Samoa ASM Unclassified 0 Iraq IRQ Medium risk 0 Andorra AND Unclassified 0 Israel ISR Medium risk 0 Antigua and ATG Unclassified Jamaica JAM Medium risk 0 Barbuda 0 Jordan JOR Medium risk 0 Armenia ARM Unclassified 0 Kyrgyz Republic KGZ Medium risk 0 Aruba ABW Unclassified 0 Lao PDR LAO Medium risk 44,721 Bahamas, The BHS Unclassified 0 Lebanon LBN Medium risk 0 Bahrain BHR Unclassified 0 Madagascar MDG Medium risk 557 Barbados BRB Unclassified 0 Mali MLI Medium risk 573 Belize BLZ Unclassified 4,083 Mexico MEX Medium risk 71,174 Bermuda BMU Unclassified 0 Morocco MAR Medium risk 0 Bosnia and BIH Unclassified Herzegovina 0 Mozambique MOZ Medium risk 191,114 Page 73 High High deforestation deforestation ISO Spillover risk risk (> 0.7), ISO Spillover risk risk (> 0.7), Country Country code category biodiversity code category biodiversity density, and density, and accessibility accessibility Botswana BWA Unclassified 713 Micronesia, Fed. FSM Unclassified British Virgin Sts. 0 VGB Unclassified Islands 0 Monaco MCO Unclassified 0 Cabo Verde CPV Unclassified 0 Montenegro MNE Unclassified 0 Cayman Islands CYM Unclassified 0 Namibia NAM Unclassified 1,020 Central African Nauru NRU Unclassified 0 CAF Unclassified Republic 317 New Caledonia NCL Unclassified 0 Chad TCD Unclassified 0 Northern Mariana MNP Unclassified Comoros COM Unclassified 0 Islands 0 Congo, Rep. COG Unclassified 503 Norway NOR Unclassified 0 Cyprus CYP Unclassified 0 Palau PLW Unclassified 0 Denmark DNK Unclassified 0 Panama PAN Unclassified 4,897 Dominica DMA Unclassified 0 Qatar QAT Unclassified 0 Equatorial Guinea GNQ Unclassified 0 Samoa WSM Unclassified 0 Faroe Islands FRO Unclassified 0 San Marino SMR Unclassified 0 Fiji FJI Unclassified 0 Sao Tome and STP Unclassified Finland FIN Unclassified 0 Principe 0 French Polynesia PYF Unclassified 0 Seychelles SYC Unclassified 0 Gibraltar GIB Unclassified 0 Slovenia SVN Unclassified 0 Greenland GRL Unclassified 0 Solomon Islands SLB Unclassified 0 Grenada GRD Unclassified 0 South Sudan SSD Unclassified 82,258 Guam GUM Unclassified 0 St. Kitts and Nevis KNA Unclassified 0 Guyana GUY Unclassified 0 St. Lucia LCA Unclassified 0 Hungary HUN Unclassified 0 St. Martin (French MAF Unclassified part) 0 Iceland ISL Unclassified 0 St. Vincent and Isle of Man IMN Unclassified 0 VCT Unclassified the Grenadines 0 Kiribati KIR Unclassified 0 Suriname SUR Unclassified 0 Korea, Dem. PRK Unclassified Switzerland CHE Unclassified 0 People’s Rep. 0 Tonga TON Unclassified 0 Kuwait KWT Unclassified 0 Trinidad and Libya LBY Unclassified 0 TTO Unclassified Tobago 679 Liechtenstein LIE Unclassified 0 Turks and Caicos Luxembourg LUX Unclassified 0 TCA Unclassified Islands 0 Macedonia, FYR MKD Unclassified 0 Tuvalu TUV Unclassified 0 Maldives MDV Unclassified 0 Vanuatu VUT Unclassified 0 Malta MLT Unclassified 0 Virgin Islands (US) VIR Unclassified 0 Marshall Islands MHL Unclassified 0 West Bank and PSE Unclassified Mauritania MRT Unclassified 0 Gaza 0 Mauritius MUS Unclassified 0 Zambia ZMB Unclassified 165,902 Page 74 ANNEX 3: HOT SPOT COUNTRIES BY INCOME CLASSIFICATION Above Number of Above Above population Spillover risk ISO World Bank income poor population livestock density and category Country code classification (2022) livestock density density livestock (Allen et al. keepers threshold threshold density 2017). threshold Afghanistan AFG Low income 6,224,614 No No No Medium risk Algeria DZA Lower middle income 2,164,658 No No No Medium risk Angola AGO Lower middle income 2,525,497 No No No Low risk Bangladesh BGD Lower middle income 17,036,489 Yes Yes Yes High risk Belize BLZ Lower middle income 31,840 No No No Unclassified Benin BEN Lower middle income 1,645,098 No Yes No Medium risk Bhutan BTN Lower middle income 71,402 No No No High risk Bolivia BOL Lower middle income 1,219,699 No Yes No Low risk Burkina Faso BFA Low income 4,316,528 No Yes No Low risk Burundi BDI Low income 3,057,901 Yes Yes Yes High risk Cabo Verde CPV Lower middle income 71,666 No Yes No Unclassified Cambodia KHM Lower middle income 2,303,289 No Yes No High risk Cameroon CMR Lower middle income 2,665,167 No Yes No Medium risk Central African CAF Low income 1,509,706 No No No Unclassified Republic Chad TCD Low income 2,809,253 No No No Unclassified Comoros COM Lower middle income 98,348 Yes Yes Yes Unclassified Congo, Dem. Rep. COD Low income 330,039 No No No High risk Côte d’Ivoire CIV Lower middle income 3,988,271 No Yes No Medium risk Djibouti DJI Lower middle income 96,199 No No No Low risk Egypt, Arab Rep. EGY Lower middle income 4,350,392 Yes Yes Yes High risk El Salvador SLV Lower middle income 353,526 Yes Yes Yes Medium risk Eritrea ERI Low income 1,510,546 No No No Medium risk Ethiopia ETH Low income 21,007,007 No No No High risk Gambia, The GMB Low income 280,196 Yes No No Low risk Ghana GHA Lower middle income 2,968,231 No Yes No High risk Guinea GIN Low income 438,417 No No No Low risk Guinea-Bissau GNB Low income 1,740,295 No Yes No Low risk Haiti HTI Lower middle income 1,768,778 Yes Yes Yes High risk Honduras HND Lower middle income 1,188,881 No Yes No Medium risk India IND Lower middle income 106,835,505 Yes Yes Yes High risk Indonesia IDN Lower middle income 12,096,968 No Yes No High risk Iran, Islamic Rep. IRN Lower middle income 5,060,129 No Yes No Medium risk Kenya KEN Lower middle income 8,990,264 No No No High risk Korea, Dem. PRK Low income 290,773 Yes Yes Yes Unclassified People’s Rep. Kyrgyz Republic KGZ Lower middle income 1,278,444 No No No Medium risk Page 75 Above Number of Above Above population Spillover risk ISO World Bank income poor population livestock density and category Country code classification (2022) livestock density density livestock (Allen et al. keepers threshold threshold density 2017). threshold Lao PDR LAO Lower middle income 1,056,997 No Yes No Medium risk Lesotho LSO Lower middle income 560,750 No No No Low risk Liberia LBR Low income 575,051 No No No Unclassified Madagascar MDG Low income 4,299,601 No No No Medium risk Malawi MWI Low income 4,419,066 Yes Yes Yes High risk Mali MLI Low income 4,437,546 No No No Medium risk Mauritania MRT Lower middle income 964,323 No No No Unclassified Mongolia MNG Lower middle income 611,436 No No No Low risk Morocco MAR Lower middle income 2,307,367 No Yes No Medium risk Mozambique MOZ Low income 4,518,368 No No No Medium risk Myanmar MMR Lower middle income 4,676,487 No Yes No High risk Nepal NPL Lower middle income 3,119,872 Yes Yes Yes High risk Nicaragua NIC Lower middle income 772,789 No Yes No Medium risk Niger NER Low income 6,048,070 No No No Medium risk Nigeria NGA Lower middle income 18,121,312 Yes Yes Yes High risk Pakistan PAK Lower middle income 17,159,433 Yes Yes Yes High risk Papua New Guinea PNG Lower middle income 955,295 No Yes No Low risk Philippines PHL Lower middle income 8,523,519 Yes Yes Yes High risk Rwanda RWA Low income 3,093,717 Yes Yes Yes High risk Senegal SEN Lower middle income 1,954,876 No Yes No Medium risk Sierra Leone SLE Low income 1,789,964 No Yes No Low risk Sri Lanka LKA Lower middle income 732,642 Yes Yes Yes High risk Sudan SDN Low income 10,863,322 No No No Medium risk Syrian Arab Republic SYR Low income 1,920,158 No Yes No Medium risk Tajikistan TJK Lower middle income 1,721,384 No No No Medium risk Timor-Leste TLS Lower middle income 202,767 No Yes No Medium risk Togo TGO Low income 836,067 No Yes No Medium risk Tunisia TUN Lower middle income 288,941 No Yes No Medium risk Uganda UGA Low income 6,354,153 Yes Yes Yes High risk Ukraine UKR Lower middle income 2,126,405 No Yes No Low risk Uzbekistan UZB Lower middle income 2,706,459 No No No Low risk Vietnam VNM Lower middle income 9,224,629 Yes Yes Yes High risk Yemen, Rep. YEM Low income 5,264,035 No Yes No Medium risk Zambia ZMB Lower middle income 4,197,681 No No No Unclassified Zimbabwe ZWE Lower middle income 2,474,258 No No No Medium risk Page 76 ANNEX 4: COST OF BIOSECURITY PER FARMER AND TOTAL TABLE 18. Total cost of biosecurity—lower bounds, point estimates, and upper bounds—for low- and lower-middle-income countries in hot spots Total cost (billion 2019 PPP$) Upper bound Point estimate Lower bound Low-income countries 4.4 2.5 5.1 Lower-middle-income countries 75.3 47.6 20.0 Low-income and lower-middle income countries 80.0 50.1 20.4 TABLE 19. Per farmer cost of biosecurity—lower bounds, point estimates and upper bounds—for low- and lower-middle-income countries in hot spots Cost per farmer Upper bound Point estimate Lower bound Low-income countries 256.8 143.0 29.2 Lower-middle-income countries 401.7 254.0 106.2 Page 77