The World Bank's 'Recipe for a Livable Planet’

This post, part of a series on the World Bank and climate finance, reviews the Bank’s “Recipe for a Planet: Achieving net zero emissions in the agrifood system” (henceforth “Recipe”)[1]. It covers Recipe’s findings on mitigating greenhouse gas emissions (GHGs) in agriculture and food systems, sinks of those emissions (where the emissions might be reduced at source or stored, generally known as mitigation), hypothetical mitigation costs, discusses the report’s limitations, and concludes with some suggestions about what the Bank might do to advance decarbonization in agrifood systems.

Recipe is a thorough summary of potential decarbonization in the agriculture and food sectors. The report, however, is less valuable than it should be because it makes no strong recommendations and declares no priorities. The absence of recommendations and priorities is the fault of Bank management which, as usual, fears anything which might reduce lending volume or vex large shareholders, still the only metrics that count to the COP followers who now occupy the Bank’s higher circles or did so in the recent past.

What does Recipe find ?

Recipe finds “Big Opportunities” (p. xxii) in agrifood as: (1) a large potential source of cheap climate actions; (2) an ex-ante benefit-cost ratio (BCR) of 16 on those actions, with benefits accruing to better health, higher incomes, and cheaper environmental goods; (3) greater efficiency in public spending by shifting from carbon subsidies to green investments; and (4) a chance for a “just” green transition in which benefits of decarbonization would accrue disproportionately to the poor.

How does Recipe analyze sources of GHGs in agrifood systems ?

Recipe identifies sources of GHGs from agrifood sectors (AFEs) by regions (Table 2.1) and countries. AFEs are about 31% of global emissions, with East Asia and the Pacific (EAP) and Europe and Central Asia (ECA) contributing more than half of the world total of 51.9 GtCO2eq/yr. Latin America and the Caribbean (LAC) at 64.5% and Sub-Saharan Africa at 64% had the highest shares of AFEs in their respective regional totals.

Recipe estimates GHG sources across activities (Recipe: Figure 2.2, pp. 18-19) from FAOSTAT calculations that are considered reliable (perhaps because they are the only estimates). The dirtiest sectors are in production: farm gate emissions (45% of AFEs, including livestock, manure, soil drainage, rice), and from forest conversion and fires (21%) with the balance of 34% of AFEs found in pre- and post-production. Middle-income countries have the highest shares of AFE and the fastest rates of emissions growth (Figure 2.5, p. 23). Low-income countries, despite having rapid growth in emissions since the 1990s, have less than 6 percent of global AFEs.

The bimodality of sources—47% of AFEs from enteric fermentation, manure, and land use change and 53% from 14—creates a “concentrated source problem” and “a dispersed source problem”. The land use sources, most importantly, are concentrated in a few countries (Brazil, Indonesia, DRC, Colombia, Bolivia, and Peru; Figure 0.8) where protection of forests could save perhaps 3 GtCO2eq/yr.

The two broad classes of sources demand different solutions, something that the Recipe does not fully consider. The concentrated source problem is tractable through policies—stopping deforestation, restricting ruminant production, promoting reforestation—that can achieve big and near-immediate reductions in emissions^[2] with high initial fixed costs and lower subsequent annual costs.

The dispersed source problem has low initial fixed costs (with one indirect exception) and higher annual variable costs. The dispersed source problem requires different policies, especially a constant effort to transfer technology to millions of farmers, processors, traders, and consumers. Examples are using less water in rice cultivation, raising N efficiency on farms, managing crop residues, promoting plant-based diets, cutting food waste, and lowering energy use in food processing and preparation.

The indirect exception to low fixed costs in dispersed GHG sources is in the shift from fossil fuels to renewables in electricity generation. Expanding renewables for electricity has become less expensive in LCOE terms than coal-fired generation and this trend will continue; in some regions, renewables for electricity have become cheaper than natural gas, a trend that will accelerate in low-access regions with abundant solar and wind potential. The report fails to note the positive interactions between mitigation finance for renewable energy and related reductions in dispersed sources (processing, transport, packaging, consumption) and complains that “most climate finance is dedicated to [sectors other than agrifood]” (p. 32).

How does Recipe define mitigation finance needs ?

The mitigation finance estimates in Recipe (Figure 2.14, Table 2.3) are probably too high because of double counting. The degree of double counting in Recipe’s calculated annual mean of US$260 billion from 2025 to 2030 is unverifiable because the ‘specific measures’ under the 6 sub-titles in Table 2.3 are not separated into amounts by country and action (e.g., “x billions for avoided conversion of forestland” in Brazil). The reported spending on sub-titles 2 (“productive and regenerative agriculture”) through 5 (“reducing food loss and waste”) is likely not additive to current spending. Adaptation spending (weather forecasts, air conditioning, fire and flood prevention) can lower mitigation costs (less input use, more cost-effective education and health systems, less carbon released in fires, less vegetation destroyed in floods), so there is an obscured multiplicative effect between mitigation and adaptation that would tend to reduce Recipe’s estimated values for mitigation funding. The report does not describe interactions among more units (heavier cattle, numbers of tractors, tons of fertilizer) and emissions intensity per unit and, if lower emissions intensity is associated with higher profitability per unit, then one may expect the benefits of lower intensity to be cancelled to some extent by greater animal weights, which is much the story of cattle production in the USA.

Another factor in Recipe that tends to inflate mitigation costs is discussing sinks—notably reforestation, livestock production, and manure management—as if they were independent of other changes in greenhouse gas cycles. We know, of course, that per capita income in middle-income countries will continue to grow with corresponding increases in their emissions, especially in electricity, transport, and farm mechanization (China, India, Indonesia, Poland). The same countries will consume more animal products, adding to emissions on farms and in the value chain, and thereby raising temperatures. Ignoring trends in emissions does not change the absolute effects of various ingredients in the Recipe, but it does inflate the relative effects.

What does the Recipe recommend ?

The Recipe recommends nothing specific.^[3] The generic recommendation is to “Scal[e] finance to transform the food system” through “conventional finance” (WBG loans and credits), “market mechanisms”, “innovative tools”, de-risking (notably guarantees and blended finance to “reduce the high upfront cost of low-emission technology”).

The Recipe gives some anodyne advice, having the common WB feature of striking a noble tone while avoiding anything controversial.

• We learn that “high-income countries should lead the way”, that those nations should reduce emissions by promoting renewable energy in AFE while offering more technical and financial support to low and middle countries;

• We are told that “middle-income countries have the greatest potential”, and the use of “cleaner and more energy-efficient production processes” is a good idea; and

• We are further informed that “low-income countries can play a pro-active role [in bypassing] high-emissions development paths” and that “climate-smart agriculture techniques” are appropriate, leaving aside the fact that no one can find replicable evidence about climate-smart agricultural techniques (pp. 121-127 in Recipe) despite years of effort to look for such evidence.

What is the value of Recipe for a Livable Planet ?

Recipe’s value as policy guidance for decarbonization in low- and middle-income countries is diminished by the usual international organization practice of trying to represent all points of view at all times, whether those points of view mean anything or not.

Recipe dithers on land use and forestry. An example (p. 128) is “ … the Democratic Republic of Congo’s populations are among the fastest growing in the world, so economic pressures on the country’s rainforest are increasing … creating alternative sources of income and delinking agriculture and deforestation are key to protecting the country’s forests, as is helping the country to access carbon markets and external technical and resource support to protect the forests. One mechanism is the Central African Forest Initiative (CAFI), which is funded by EU countries to reduce deforestation and forest degradation and through the development of national investment frameworks.” While this sounds noble, the EU’s estimate of CAFI’s decarbonization impact in DRC (the major recipient of CAFI funds) is that the program spent US$191 million to eliminate 2.7 MtCO2eq.^[4]

Recipe reports verifiable and unverifiable results as if they had the same scientific weight. The report (Figure 3.4, p. 77, p. 255) gives reliable data from China^[5] on fertilizer management (by which GHG mitigation would lower input costs, giving a negative carbon price), grazing restrictions, and direct seeding of paddy. Recipe^[6] then blocks its own message by reporting unverified results on emissions following technical change in animal production. An example of unverified results is in reciting the fairy tales of the US cattle lobby about “breeding cattle for low enteric fermentation” (p. 103) or setting out, yet again, the uneconomic case for “more efficient recycling of organic manure” (p. 77) or reporting NDCs as if they had been independently verified (p. 170).

Recipe argues (Figure O.5) for more AFE mitigation finance while showing that some decarbonization technologies do not need incremental finance because they save costs. The chief examples are formulation of fertilizer doses nearer to economic optima and more efficient water and fertilizer use in paddy production (Figure 3.4 for China). Recipe, moreover, does not account for the indirect cost savings from investments in local mitigation (packaging, processing, transport, on-farm energy) in terms of foregone local adaptation costs (flood control, road maintenance).

The report recites the doctrine that MIC and LICs cannot meet their mitigation NDCs for lack of technology. Yet, decarbonization methods in MICs and LICS are simple and cheap, with most investments in known technologies (water use in paddy, higher-quality livestock feed, farm practices to reduce N losses, on-farm energy, crop residue management, product packaging, transport, processing, and retail) and benefit from economies of scale because they can be delivered to many farms with a common message. Indeed, the projects in the WBG “Scorecard” for net emissions^[7] embody local technology models.

Recipe falls for the trick (p. xxiii, p. xli, p. 167, p. 169) that market mechanisms, such as carbon trading, can reduce AFE, despite weak evidence in rural areas (https://www.nature.com/articles/s41467-024-53645-z#citeas) and given allegations of corruption in those mechanisms (https://climatecasechart.com/case/united-states-v-newcombe/ and the widespread attacks on Verra’s reliability as a carbon certifier). Recipe also seems to overstate the carbon benefits of the REDD + mechanism (Duchelle et al, “What is REDD+ achieving on the ground?”,Current Opinion in Environmental Sustainability,https://doi.org/10.1016/j.cosust.2018.07.001). Note also that Recipe pretends that initiatives like CAFI can help countries like DRC access carbon markets when there is no evidence that DRC, or other poor nations, have gained such access.

How can the Bank support decarbonization in agrifood systems ?

A complete reorganization of global climate finance is needed, to which I return in a future post. Here I conclude about what the Bank can do today.

The Bank can advocate carbon and Tobin taxes, which would expand climate finance with low collection costs. Recipe only discusses carbon taxes anecdotally, a gap following from the Bank’s financial interests, which insists on private (or “blended”) finance as a complement to its own project finance. Recipe, moreover, does not mention a financial transactions tax. There has been published reference [8] to possible Bank support for a financial transactions tax at COP 30 and while I suppose that some Bank people have said this in private, it is revealing that a flagship report like Recipe says nothing in public.

The highest potential for agrifood mitigation involving global finance is in forestry and land use. Global efforts (the various green funds, MDBs, REDD) are fragmented, use impact estimates that cannot be replicated, and have achieved modest decarbonization at best. A radical new approach to stop tropical deforestation is needed—buy millions of hectares of tropical forests that are now being destroyed for beef, soy, palm, maize, and cassava —and pay a rent to the local inhabitants for the income lost from destructive activities.^[9] Instead of wasting WBG (and IMF) money on bail outs of private banks and on small vanity schemes, that money could buy Brazilian (Indonesian, Peruvian, Myanmari, Colombian, Bolivian) forests and pay a rent to the people of those countries.

Another high potential return to global finance is in reducing ruminant numbers, something that would cut emissions from enteric fermentation and from manure deposition. There would, moreover, be positive interactions between blocking deforestation and reducing cattle numbers in Brazil. A modest first step would be for the Bank to stop all research and development of ruminant livestock production in projects with WBG support and in Bank funding to the CGIAR. A more consequential step would be allocating finance to IDA countries for grazing restrictions while giving cash compensation to the herding groups.

The Bank can reduce the demand for additional financing by being more aggressive in demanding an end to fossil fuel and fertilizer subsidies in borrower nations, with financing and the public information campaigns focused on production cost savings (e.g., less N wasted), on the gains from diverting subsidies into investment in renewables and in public transport, and on publicizing the regressive incidence of fossil fuel subsidies.

The Bank can shift its own programs from the feeble “climate” measures^[10] in budget support operations—which usually do nothing but allow management and the Board the false comfort that they are doing something in the global heating crisis—into investment projects that do something.

* * *

[1]. https://www.worldbank.org/en/topic/agriculture/publication/recipe-for-livable-planet. I thank a reviewer for comments on an earlier draft.

[2]. Brazilian deforestation follows the nation’s elections, falling when Lula is President and rising when Bolsonaro is President. Something similar can be said of the IMF which speaks frankly to Brazil about its environment when the second term of the Fund’s MD is not in question; and less frankly when a second term is up in the air.

[3]. It should not be too much to ask for specific recommendations from a report with > 100 contributors.

[4]. 2023 CAFI Financial Report (2023): Table 5.1. P. 9; and 2023 CAFI Annual Report: p. 20.

[5]. Recipe (p. 77) cites Nayak, D., et al., Management opportunities to mitigate greenhouse gas emissions from Chinese agriculture. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.035)

[6]. Recipe, Table B.2 “Cost-saving and Low-cost mitigation options in selected MICs and LICs by Sector”.

[7]. https://scorecard.worldbank.org/en/data/result-details/CSC_RES_GHG_EMS_YEAR?orgCode=ALL&refareatype=REGION&refareacode=ACW&age=_T&disability=_T&sex=_T

[8] Miguel Correia, “Financing the costs of tackling climate change: the financial transaction tax at the COP 30”, https://catolicalaw.fd.lisboa.ucp.pt/faculty-knowledge.

[9]. Jose Scheinkman spoke at the 2024 ABCDE on "Carbon pricing and reforestation in tropical forests" of Brazil and argued that “with modest prices for CO2, Brazilian Amazon would produce noticeable CO2 capture.”

[10]. Kenya, Ethiopia, and Ghana are examples.