Share this

The Alliance for a Green Revolution in Africa (AGRA) has been under fire over the last year after our research revealed that the billion-dollar agency had made little progress toward its stated goals of doubling yields and incomes for 30 million small-scale farming households while halving food insecurity. Since the publication of that research in July 2020, as an academic working paper and a related report, False Promises, AGRA has failed to provide evidence to refute our findings, withholding outcome monitoring reports after requests by African organizations.

Many hoped AGRA’s 2020 Annual Report, published July 12 with a companion report on “Emerging Results 2017-21,” would finally offer some evidence of its impacts. After reviewing the 66-page annual report and the 37-page companion document, I can report that AGRA provides some data but no convincing evidence of progress toward these three topline goals. The document confines itself largely to reporting not on its 15 years of work but its most recent 2017-2021 strategic plan. The evidence base for the reporting is unclear but it is undoubtedly thin; AGRA has published only one set of Outcome Monitoring reports, based on 2019 surveys.

The lack of accountability to its goals is particularly troubling for two reasons. First, the original endpoint for achieving them was 2020, which was then extended to 2021. This leaves African governments and farmers as well as AGRA’s donors — most notably the Gates Foundation, Rockefeller Foundation, U.S. Agency for International Development (USAID), UK Agency for International Development (UKAID) and the German development agency BMZ — with no compelling evidence of AGRA’s impacts.

Second, AGRA is now formulating its new strategic plan; it has been drafted and approved by the Board of Directors and is awaiting an implementation plan. That plan will require new funding commitments, and anonymous sources indicate that AGRA will seek another $1 billion in funding through 2030. AGRA’s failure to provide evidence of progress means donors will be asked to continue their support without any assurance that such aid has been effective. As I wrote earlier this month, they will be asked to “throw good money after bad.”

I was asked by the U.S. Right to Know (USRTK), a transparency organization that has been tracking this controversy, to review AGRA’s 2020 Annual Report and companion document to assess whether it provides the kinds of evidence that have been lacking to date. Below are my findings, which can be summarized as follows:

  • AGRA provides no evidence of its effectiveness in raising yields, incomes and food security since its founding in 2006; in fact, it fails to offer any information about its first 10 years of work, reporting as if the initiative just started in 2017.
  • AGRA does report on yields, incomes and food security, but the data comes from a mix of sources, including “rapid assessments” in an indeterminate number of countries with an indeterminate number of farmers. As such, the data lacks validity.
  • The data presentation is misleading, clearly intended to cherry-pick success stories in selected countries and crops without even pretending to put such outcomes in a larger context.
  • Claimed improvements in food security suffer from these same deficiencies. The progress flies in the face of hunger estimates from a recent FAO report, released on the same day AGRA published its annual report. The FAO reported a jump of 44 million undernourished people in Sub-Saharan Africa to an alarming 264 million. The new AGRA documents are disturbingly tone-deaf about the dire and worsening conditions for poor Africans.
  • AGRA’s stated monitoring methodology is deeply flawed, ensuring that future claims of progress will be based on unreliable data collected on selected crops over too short a time period to offer valid results. AGRA’s inability to account for its first 10 years of work renders the organization’s claims of impact anecdotal and impossible to verify. As such, donors should reconsider their continued support for such a poorly run, unaccountable and ineffective organization.

Vague data from undocumented sources

The Annual Report is the fourteenth since AGRA was initiated by the Gates Foundation and Rockefeller Foundation in 2006 and all share this same problem of vague data and undocumented sources. This year’s report, titled “Nourishing Change Across African Agriculture,” focuses on the organization’s current 2017-21 strategic plan, which AGRA Board Chair H.E. Hailemariam Dessalegn (former Prime Minister of Ethiopia) characterizes as “an integrated delivery model to catalyze agricultural transformation.” AGRA’s goal is to catalyze private sector and government capacities through public-private partnerships to improve the delivery of commercial seeds, fertilizers and other inputs and achieve the kind of productivity revolution AGRA promises.

As such, AGRA reports more on “transformation processes” than it does on outcomes. For AGRA, many of those processes are the outcomes: the increasing availability of certified seeds, the rising number of Village-Based Assistants as private extension agents, the rapid approval of policy reforms to speed the delivery of commercial seeds and fertilizers, the capital catalyzed by AGRA for small and medium scale agribusinesses. There is little attention to the outcomes for farmers beyond the numbers of farmers “reached” by AGRA, with no clarity on the extent of those farmers’ engagement nor the impacts on their farming. The companion document on “Emerging Trends” offers little more data or clarity.

One brief presentation on page 10 of the Annual Report, shown on the next page, summarizes the 2017-21 impact on smallholder farmers. It illustrates many of the limitations of AGRA’s presentation of outcomes. The sources of the data are unclear, the number of farmers reached by what interventions is not disclosed, and the percentages are chosen to overstate AGRA’s impacts.


summary of the 2017-21 impact on smallholder farmers


What is wrong with this picture from AGRA?

  1. “Regional impact?” The map shows AGRA’s current 11 priority countries and then the claimed “regional impact,” though rarely is any evidence presented of such impact. In general AGRA assumes that if commercial seeds and fertilizers are more widely available this will spill over into regional economies. A more accurate map of AGRA’s true areas of influence would likely be small blotches within the borders of its 11 focus countries.
  2. Number of farmers reached: The claim of 10.1 million farmers reached directly by AGRA is not documented and fails to identify how they engaged with AGRA. Some may have attended a training, others perhaps sold to an AGRA-supported buyer, others simply had easier access to commercial seeds and fertilizers, and they are considered “beneficiaries” even if they did not buy any. AGRA should disclose how many farmers were reached by what AGRA intervention.
  3. No point of comparison: The next four categories — technology adoption, seed use (“replacing open-pollinated varieties” with commercial hybrids), access to credit and producing more than a subsistence — offer percentages with nothing to compare them to. Is this an improvement over time? How much of an improvement? Over how long a period? We do not know.
  4. Number of farmers unclear: Nor do we know how many farmers are represented by these percentages. The presentation implies that those are percentages of the 10.1 million farmers reached, but that is unlikely to be true and we do not know what is.
  5. Painting a rosy picture: The last figure in the graphic is particularly misleading, suggesting that two-thirds of farmers (the 10.1 million?) earned higher incomes in 2020 than they did in 2017 from selling surplus crops, presumably the product of higher yields from AGRA inputs. From their own data, it would be more accurate to say that nearly two-thirds of farmers saw little or no income growth. Consider:
    1. Later in the report AGRA states that “60% of farmers reached by AGRA have adopted new farming practices.” (page 15) This contradicts the 76% figure in the graphic above. It means that 40% have not adopted, perhaps a testament to the technology being expensive and unproductive;
    2. AGRA reports that “60% of farmers with surplus had higher incomes; over 70% had significant income growth, by 20% to over 80% vs. 2017.” (page 15) Again, the picture is far less rosy. We do not know how many farmers they are counting, perhaps the 87% cited in the earlier graphic. If so, 13% did not even produce a surplus;
    3. and only 60% of those with surpluses had higher incomes, so just 52% (60% of 87) of farmers — barely half — saw higher incomes;
    4. Only 70% of those saw significant increases, so only 36% of all farmers (70% of 52) saw significantly higher incomes;
    5. Put another way, 64% of farmers saw zero or only small income gains.

The fact that many farmers do not see any benefits is important. One of the findings in the case study research for the False Promises report was that some farmers are buying inputs at least partially on credit, and some are finding themselves in debt when yields fail to rise enough to cover input costs. This is a problem that is endemic to Green Revolution programs, from India to Africa.

Rising crop yields? Picking cherries

One common fallacy in data presentation is to offer selected success stories (“cherry-picking”) without revealing the larger trends or documenting the sources of the data or the time periods they cover. AGRA’s presentation on productivity increases exhibits all of these failures, with the clear intent to present progress toward the goal of doubling yields for 30 million farmers. The graphic below is from page 16, prefaced by the statement: “We can see evidence of higher yields across different countries and crops.” Indeed, they show different countries and crops and they show yield increases. But:

  1. AGRA does not identify the number of farmers represented.
  2. AGRA does not offer any characterization of broader yield trends beyond the selected crops, an important point. My research showed better results for supported crops like maize and terrible results for crops not favored by AGRA or government input-subsidy schemes. (For example, across all AGRA countries millet yields declined 21%.)
  3. AGRA does not identify the time period of the yield growth, presumably since 2017.
  4. The figures certainly do not represent national trends. Consider the figures for these same crops and countries in our well-documented national-level data from FAO from a baseline of 2004-6 to an endpoint of 2016-18. I also include my more comprehensive Staple Yield Index which captures a range of staple crops in each country.


Critique of productivity increases


In other words, these are not just cherry-picked figures. Those cherries were picked after an indeterminate growth period and from orchards that may have been very small, i.e., from a small number of farmers unrepresentative of national trends.

Again, in yield growth AGRA presents unconvincing data that it is approaching its goal of doubling productivity for the majority of smallholder farmers in AGRA countries. For AGRA as a whole, our data stands unrefuted that over a 12-year period staple yields grew only 18%, not the promised 100%.

Unsubstantiated claims on food security

The final goal against which AGRA fails to convincingly document progress is halving hunger. As we show in our own research, across AGRA’s 13 countries there was a 31% increase in the number of people considered severely undernourished by the FAO over 12 years through 2018. That is a far cry from the promised 50% decrease. Now, after a year that saw dramatic hunger caused by COVID-19 and its related economic impacts, we would not expect to see much progress, though AGRA devoted considerable resources to COVID-19 relief, much to its credit.

Using selective and poorly sourced data, AGRA reports dramatic success since 2017. They report only on the unsourced 87% of farmers with a surplus in 2020, so we do not know how many farmers they include. Using one common metric — the number of months a family’s food production lasts — they report that 66% of those with a surplus saw an increase of at least four months in 2020 compared to 2017. That would indeed be a positive outcome, though it does not easily correlate with their goal of halving hunger. Again, we do not know where this data comes from, nor how many countries or farmers it covers.

What we do know is that such figures are wildly out of line with national trends, documented most recently by FAO in its annual State of Food Insecurity report. At a global level, those figures are alarming, showing an increase of as much as 25% in the number of undernourished to 811 million people, an increase of up to 165 million in one year. It is the fifth straight year in which the numbers rose.

Sub-Saharan Africa as a whole saw 264 million chronically undernourished people, a jump of 44 million just since 2019. FAO reports undernourishment by country as a three-year average, so the published data on AGRA countries do not reveal separate 2020 impacts. The 2018-20 average for all AGRA countries remained largely unchanged from the 2016-18 levels we calculated in last year’s report, with 128 million residents in AGRA’s 13 focus countries (including Niger and Zambia) suffering chronic and severe hunger. That remains a major failing for AGRA, which held out the goal of halving food insecurity not just for the farmers it worked with but for a larger group of 20 countries.


Supposed food security impact


Other observations

Beyond the limited and misleading presentation related to AGRA’s topline goals, I can offer additional observations:

  • To AGRA’s credit, it devoted $11 million to efforts to mitigate the destructive effects of COVID-19, support that was surely needed in such a rapid-onset crisis.
  • AGRA frequently stresses its focus on “farming systems,” acknowledging at one point that it initially had too much of a “technology-focused approach.” (page 13) There is attention to language. For example: “Like any ecosystem, the parts of the agricultural sector must work in concert.” They are clearly referring to functioning markets, not functioning ecosystems.
  • AGRA is very proud of its 32,000 Village-based Assistants (VBAs), privately funded extension agents trained to complement governments’ meager extension services. AGRA’s own mid-term evaluation warned that this system was unsustainable since the VBAs would not continue to work when funding is withdrawn and the government is not able to pick up the tab.
  • There is a constant focus on hybrid maize seed, and even the explicit intent to replace open-pollinated varieties. This combines with a focus on domestic production of hybrid seeds by local agribusiness firms. In one illustrative story, Rwanda proclaims “self-sufficiency” — not in food, but in hybrid maize seed production. AGRA’s obsessive focus on replacing farm-saved seed with commercial varieties, which must be purchased every year, is one of its most objectionable activities.
  • Much of AGRA’s work on “farming systems” involves lobbying national governments to change laws and regulations to allow easier entry and distribution of commercial seeds, fertilizers and other inputs. One of their advertised successes in the annual report is to have “reduced the time required for the policy reform process by 50%.” (page 31) Many Africans see this not as a measure of administrative efficiency but of lobbying muscle.
  • AGRA has traditionally taken the position that it opposes government input-subsidy programs as market-distorting, even though such programs provide direct financial support to farmers to purchase the inputs AGRA is promoting. It is safe to say there would be very little technology adoption without subsidies. In this report, AGRA seems to support “efficient subsidy systems” and claims it has worked with governments to improve them, no longer pretending to bite the hand that feeds their Green Revolution.
  • At one point in the report AGRA President Agnes Kalibata claims AGRA has “touched in some way” 44 million smallholder farmers in its lifetime. She offers no evidence for this, of course. But one of AGRA’s criticisms of our research is that it is unreasonable to use national-level data as an indicator of progress when AGRA is working with only a subset of farmers. Its target of 30 million farmers was already a substantial majority of smallholders in AGRA’s 13 countries, according to the most comprehensive academic survey available. That is why we thought our methodology was justified. 44 million would represent an overwhelming majority, further validating our claim that national-level data is indeed indicative of AGRA’s progress.

AGRA’s flawed monitoring methodology

As I pointed out in my analysis of AGRA’s Outcomes Monitoring reports, AGRA is now stating that it will rely on three years of such data, from 2019, early 2021 and 2022, to evaluate its progress. This will not generate reliable data. Here is how they explain it on their website:

“The first wave of the outcome surveys carried out in 2019 has provided the first data point in systems assessments and household surveys. Due to the COVID-19 pandemic in 2020, the second data point planned for April 2020 could not proceed. The next round was commissioned in November 2020, implemented through the first half of 2021. The third and final wave of the outcome panel surveys may be carried out through the first half of 2022 depending on AGRA’s next strategy re-investment decisions, providing three data points to initially assess AGRA’s contribution through its programmes.”

It is common in agricultural research to measure progress over time by using three-year averages for starting and ending points. This is what I did in my AGRA research, using a baseline of 2004-6 for the pre-AGRA baseline and 2016-18 for the end point, based on the latest data available. Climatic, agronomic and other variations can significantly affect outcomes data in any given year. AGRA is failing to account for this in its monitoring plan, which is partly the result of AGRA having done a very poor job monitoring and evaluating its first 10 years of work. It now scrambles to assemble a quick progress report.

Ironically, glowing proclamations about the original Green Revolution in India in the 1960s and 1970s suffered from precisely this error of relying on a skewed baseline. The two years before the introduction of Green Revolution inputs were drought years with severe impacts on production. The rains returned as the Green Revolution technology arrived. Production returned to near-normal levels, but they appeared to be miraculous in comparison to two unusually unproductive years. In fact, when one compares yield growth for wheat in the 10-year period before the introduction of the Green Revolution with the 10-year period after, yields grew faster before than they did after. (Read more on this here.)

AGRA’s four-year time period, 2019-22, is both too short to measure progress and fails to account for seasonal variability year-to-year. As such, the outcomes data they are promising for 2022 will not produce reliable measures of impact.

The other telling flaw in AGRA’s outcome monitoring is that for farmer impact they are relying on household surveys of 1,000 AGRA beneficiary farmers growing a given crop, using the 2019 surveys as a baseline and comparing survey data from 2021 and 2022. That means AGRA is not monitoring progress among other farmers growing other crops. Of the 14 surveys reported in their 2020 Outcome Monitoring reports, nine were for the supported crops of maize (six) and rice (three). Another two were for soybeans, with two for beans and one for cowpeas. This means:

  1. Over four years their monitoring will heavily favor supported crops such as maize. Our research showed that support for maize, especially government seed and fertilizer subsidies, was undermining other crops such as millet, which saw a 24% decline in production and a 21% decline in yields in AGRA countries.
  2. Surveying only favored and supported crops will bias the monitoring results upward. Rwanda, for example, tripled maize production and increased yields 66%, but overall its staple yields languished, rising just 24% as traditional crops lost land and investment to maize. Most important, the number of undernourished people jumped 40%. A survey of Rwandan maize farmers would be very misleading on its own.

AGRA’s goal was to double food crop productivity, not just productivity in one or two crops.

Given AGRA’s poor record of monitoring and evaluation since its inception in 2006, it is ironic that one of AGRA’s claims of impact in the area of “state capabilities” reads: “We also supported the development and improvement of nine national agricultural monitoring and evaluation systems….” (page 57)

Financial report

The financial documentation in the annual report shows $93,703 in 2020 contributions with $52,728 in grants. Another $16,000 goes to other program costs. Overhead is a relatively high $27,135, 28% of total expenditures of $96,025. As a percentage of direct expenses (a common way of assessing nonprofit overhead burdens), overhead is a very high 39% of direct expenses.

As usual for AGRA, there is no breakdown of contributions by donor or even category of donor. Such information is not made public in its tax filings either. Its donors are listed, in a manner of speaking, on the final page of the report. A page of logos present “Resource Partners,” a mixture of bilateral aid donors, private foundations, corporate donors and international agencies. Again, the lack of transparency is troubling and beneath professional standards for nonprofit organizations. We know from our research that over AGRA’s lifetime the largest funder by far is the Gates Foundation (about $650 million of a roughly $1 billion budget), followed by the Rockefeller Foundation, USAID and UKAID (contributing maybe $75-$120 million each), followed by German aid organization BMZ (perhaps $11 million in recent years). The Netherlands and Norway’s NORAD are listed as donors as well, but NORAD told the Alliance for Food Sovereignty in Africa that it no longer supports AGRA. So, AGRA’s report on who their donors are also seems to be incorrect.

The Gates Foundation’s dominant role is of course obscured by this presentation. This seems intentional. All questions I have directed to the foundation have been referred to AGRA, and in one in-person interview the program officer refused to discuss AGRA, saying that the organization is not the foundation’s responsibility. AGRA also commonly distances itself from BMGF, insisting that AGRA is an “African institution” and even disputing the fact that AGRA was started by the Gates and Rockefeller foundations. One fingerprint in the Annual Report is telling: it shows that AGRA is registered as a nonprofit in Olympia, Washington, a short drive from BMGF headquarters in Seattle.

AGRA's registered address

Conclusion: Donors should demand better

AGRA’s 2020 Annual Report and its companion “Emerging Trends” report provide no convincing evidence that AGRA is making significant progress in its original goals of doubling yields and incomes for 30 million small-scale farm families while halving food insecurity by 2020. The new documents make an effort to assess progress on these topline goals, but the data sources are not disclosed and AGRA’s presentation is selective and misleading. AGRA’s plan for monitoring progress for its 2017-21 strategy is deeply flawed and guaranteed to provide a favorable picture of AGRA’s impacts.

The new reports certainly fail to refute our findings in the comprehensive review of AGRA’s progress toward its goals using national-level data:

  • Yields for a basket of staple crops grew just 18% over 12 years through 2018, far below the goal of doubling productivity, a 100% increase.
  • There was no sign of significant increases in farmer incomes thanks to rising yields and marketable surpluses. Overall, poverty remained endemic in most AGRA countries.
  • The attention to favored crops such as maize, supported by government subsidies for the purchase of Green Revolution inputs, resulted in a decline in the land and resources devoted to key staples such as millet, sorghum and sweet potato. This had negative impacts on soil fertility, as well as nutritional diversity.
  • Hunger rose dramatically, with the number of undernourished people increasing 31% across AGRA countries, not decreasing 50% as promised by AGRA.

AGRA’s donors should reconsider their support for such an unsuccessful and unaccountable initiative. As AGRA prepares a drive to raise an additional $1 billion through 2030, donors should do their own rigorous assessments of aid effectiveness. They should shift their funding to agroecology and other low-cost, low-input systems. These systems have shown far better results, raising yields across a range of food crops, increasing productivity over time as soil fertility improves, raising incomes and reducing risk for farmers by cutting input costs, and improving food security and nutrition from a diverse array of crops.

AGRA’s continued failure to report accurately on progress toward its goals, and its apparent failure to achieve them, represent a challenge to the upcoming U.N. Food Systems Summit, led by AGRA President Agnes Kalibata. By many accounts, the summit is preparing to endorse a set of business-as-usual “innovations” rather than breaking with floundering programs such as AGRA to explore promising new strategies to achieve zero hunger by 2030.


Download a PDF of the policy brief.

Filed under