Endowment fund - Trucks in a crop field

Return on investment in genebanks

While the value of crop diversity is not disputed, it has multiple components and can be complex to quantify. We focus here on two broad sources of economic value of crop diversity: use and option value. “Use value” refers to the ability of crop diversity to provide yield (including yield stability) and non-yield (e.g. nutritional, environmental) benefits. “Option value” is associated with retaining potentially valuable but unknown genes and traits within a crop diversity collection, which may be discovered and provide use value in the future. This category of value can be equated with the insurance provided by crop diversity against future unpredictable challenges, such as new pests and diseases and evolving market conditions.

The value of crop improvement

The monetary value of the current and past use of diversity to improve yield, nutrition and resistance to pests and diseases has been documented a number of times in the literature because it is the least complicated to calculate. Raitzer and Kelley (2008) provide a meta-analysis of the relative benefits and costs of CGIAR research investment (Table 1), and find aggregate benefit-cost ratios ranging from 1.94 to 17.26, confirming significant net benefits and robustly suggesting that the CGIAR research programs have been a productive investment. Renkow and Byerlee (2010) confirm large returns from investment in CGIAR research programs, particularly on crop genetic improvement, which have had the most profound documented positive impacts globally.

Further, there have been a number of attempts to quantify the specific contribution to crop improvement of a particular subset of crop diversity: crop wild relatives. Table 2 gives examples of the economic value of genetic contributions from wild relatives to the improvement of a number of different crops. Table 3 provides examples of the range of valuable traits transferred from wild species to improved varieties of rice.


Table 1. Plausible aggregate benefit estimates by crop and geographical region

Crop Region Reference million US$, 1990 values
Barley Global (Aw-Hassan et al., 2003) 330
Beans Global (N.L. Johnson et al., 2003b) 590
Cassava Global (S. Johnson et al., 2003) 230
Maize Global (Morris, 2002) 440
Wheat – spring bread Global (Byerlee and Traxler, 1995) 9,750
Wheat – spring bread Global (Heisey et al., 2002) 880
Rice (IRRI) Asia (Hossain et al., 2003) 4,310
Rice (CIAT) Latin America (Sanint and Wood, 1998) 8,280
Rice (WARDA) West Africa (Dalton and Guei, 2003) 150

Source: Raitzer and Kelley (2008)


Table 2. Estimates of the economic value of genetic contributions of crop wild relatives (CWR)

Crop Region Value, adj. 2012 US$ Parameters Reference
Wheat Global 107 million Annual benefits from disease resistance introgressed from wild wheat species (Witt, 1985)
Tomato Global 16 million Annual contribution of genes from wild tomato species Lycopersicon chmielewski (Iltis, 1988)
Coffee Global 1.660 billion Net present value of wild coffee genetic resources (Hein and Gatzweiler, 2006)
Sunflower Global 273-392 million Annual contributions of the wildsunflower gene pool (Hunter and Heywood, 2011)
Tomato Global 255 million Annual contributions of a wild tomato species providing a 2.4% increase in solids content (Hunter and Heywood, 2011)
Multiple US 712 million Annual contributions of CWR to US economy from domestic and imported sources (Prescott-Allen and Prescott-Allen, 1986)
Multiple US 28.61 billion Annual contributions of CWR to US economy (Pimentel et al., 1997)
Multiple Global 164.5 billion Annual contributions of CWR to world economy (Pimentel et al., 1997)
Multiple Global 1.686 billion Annual value of increase in cropproductivity because of CWR genetic contributions (NRC, 1991)
Multiple Global 68 billion* Current value of genetic contributions from wild gene pools of 32 crops (PwC, 2013)

Source: Tyack and Dempewolf (2015)


Table 3. Examples of CWR traits transferred to rice (Oryza sativa)

  Crop wild relative Traits
1 O. australiensis Resistance to: bacterial blight, brown plant-hopper
2 O. brachyantha Bacterial blight resistance
3 O. glaberrima Nutritional and grain quality improvement, acidity and iron and aluminium toxicity tolerance, resistance to nematodes
4 O. glumaepatula Cytoplasmic male sterility, yield improvement
5 O. grandiglumis Improved grain quality
6 O. longistaminata Drought resistance, yield improvement, resistance to: grassy stunt virus, bacterial blight, yellow stem
7 O. minuta Improved agronomic traits, resistance to: bacterial blight, blast, brown plant hopper, white-backed plant hopper
8 O. nivara Cytoplasmic male sterility, resistance to: bacterial blight, grassy stunt virus, leaf-folder, tungro, yellow stem borer
9 O. officinalis Resistance to: bacterial blight, brown plant-hopper, grassy stunt virus, white-backed plant hopper
10 O. perennis Cytoplasmic male sterility
11 O. ridleyi High acidic-sulphate content soil tolerance
12 O. rufipogon Aluminium toxicity tolerance, cytoplasmic male sterility, drought resistance, yield improvement, acid sulphate soil tolerance, cold tolerance, resistance to: bacterial blight , brown plant-hopper, grassy stunt virus, leaf-folder, rice stripe necrosis virus, soil-borne diseases, tungro, white-backed plant hopper, yellow stem borer
13 O. sativa f. spontanea Cytoplasmic male sterility
14 O. latifolia Resistance to: bacterial blight, yellow stem borer
15 Zizania latifolia Improved grain quality, resistance to: blast, sheath blight

Source: (Maxted and Kell, 2009)

Yield stability

Crop improvement has contributed not only to the level of crop yields, but also to their stability. Yield stability is especially critical for farmers in vulnerable, marginal situations (FAO, 2015; Harvey et al., 2014; Heisey and Rubenstein, 2015). Gollin (2006) finds declining variability of maize and wheat yields in developing countries, as measured by the coefficient of variation around trends over the past 40 years. The result is strongly associated with the spread of improved varieties, even after controlling for increased use of irrigation and other inputs. The annual value of benefits from improved yield stability are estimated at USD149 for maize and USD143 million for wheat, which exceed the total annual spending on breeding research on these crops. Other studies put values on long-standing efforts in breeding for disease and pest resistance. Marasas et al. (2004) estimate that CIMMYT’s work on maintaining leaf rust resistance has generated USD5.4 billion (in 1990 dollars), at a benefit-cost ratio of 27:1. Dubin and Brennan (2009) put the global benefits of resistance to all types of wheat rusts between USD600 million and USD2 billion per year (in 2006 dollars). More importantly, agricultural systems globally have largely avoided major crop failures, in part because more frequent turnover of varieties has brought new sources of resistance (Renkow and Byerlee, 2010).



Nutrition enhancement

The majority of poor households rely on staple crops for their nutrient, and not just calorie, needs. This recognition necessitates an increased emphasis on the nutrient content of high-yielding varieties of cereals, pulses, roots and tubers. HarvestPlus (2014) highlights the condition of “hidden hunger,” whereby more than two billion people in the world do not get enough essential vitamins and minerals – such as vitamin A, zinc, and iron – because more nutritious foods are too expensive or simply unavailable. Through biofortification, HarvestPlus and its partners have developed new varieties of staple food crops that contain higher amounts of key nutrients. The focus on biofortification of staple crops is advantageous because such crops can reach marginal communities often missed in public nutrition interventions. Moreover, biofortification is cost-effective, as it requires one up-front investment, and is sustainable, because consumers eat staple foods on a regular basis.


A study by Gannon et al. (2014) in Zambia found significant increases in total body stores of vitamin A for groups that received biofortified orange maize. In Mozambique, Low et al. (2007) showed significantly increased intake of vitamin A among young children in households receiving orange-fleshed sweet potatoes combined with extension advice on nutrition. A follow-up study by Jones and de Brauw (2015) found evidence of reduced duration of diarrhea through agricultural interventions that promoted the consumption of orange-fleshed sweet potato. In 2013, the first zinc-rich rice variety “BRRI dhan 62”, developed from zinc-rich parental germplasm produced at IRRI and advanced by the Bangladesh Rice Research Institute with support from HarvestPlus, was released in Bangladesh. In Ethiopia, Ghana, India, Mexico and Nicaragua, quality protein maize (QPM), a pioneering technology developed by the International Maize and Wheat Improvement Centre (CIMMYT), is now being widely promoted in response to the problem of putting affordable protein within the reach and means of smallholder farmers. While impact studies have not considered aggregate adoption and long run use, this type of work is likely to accelerate with the scaling up of biofortification research by CGIAR and partners (Renkow and Byerlee, 2010).



Environmental impacts

Crop genetic improvement has also resulted in increases in resource-use efficiency in farms, minimizing adverse pressures on the environment. While the early varieties of the Green Revolution were input intensive, there has been a shift towards improved varieties that require less pesticide, fertilizer, water, labor, and indeed, land. Genetic improvement of crops under resource-scarce conditions will be an important avenue to improve food security given increasing population demands and changing climate conditions (Hall and Richards, 2013). Studies by Byerlee et al. (2014) and Villoria et al. (2014) show net saving of land due to agricultural innovations and that technology-driven intensification at a global level minimizes expansion of agriculture into marginal lands. Crop genetic improvement also provides opportunity to identify and select for physiological and morphological traits that increase the efficiency of water use and yield under water-limited conditions (Ito et al., 1999; Richards et al., 2002). For example, IRRI scientists have identified several key regions of the rice genome associated with drought tolerance and improved grain yield under drought conditions (Kumar et al., 2014). Drought-tolerant varieties released in India, the Philippines and Nepal show promising yield advantages of 0.8-1.2 tons per hectare over drought-susceptible ones (IRRI, 2015).

A meta-analysis by Klümper and Qaim (2014) covering 147 publications finds 37% reduction in chemical pesticide use and 22% increase in yields from adoption of improved crop varieties. Together, the yield gains and cost savings have resulted in a 68% increase in farm profits. At IRRI’s rice research farms in the Philippines, insecticide use has been reduced by 96% between 1993 and 2008 (Hamilton, 2008). In a similar vein, nitrogen-fertilizer efficiency of maize in the U.S. has increased by 36% in the past 21 years due to public sector research and extension (Tilman et al., 2002).

Option value

Option value is associated with the unknowable future role of crop diversity as a source of valuable genes and traits that have yet to be discovered. A large collection of crop diversity has high option value as insurance against future, unanticipated challenges, such as new pests and diseases, changing environments and evolving consumer needs. Smale and Hanson (2010) show that the option value of large collections of crop diversity is greatest when the chances of finding a specific trait are slim but the economic return on discovery is significant, or when the trait of interest is found in a tiny part of the genepool, such as a subset of landraces or crop wild relatives from a particular geographic location. The availability of large amounts of characterization and evaluation data is a particular added value of large international collections, facilitating the identification of exactly what is needed by the user.

For example, there is a near absence of resistance to Russian wheat aphid in materials originating outside Central Asia (Robinson and Skovmand, 1992). Using CIMMYT’s data on germplasm search costs and areas planted to susceptible wheat, Gollin et. Al. (2000) simulate various scenarios depending on the time lag from discovery of resistant material to farm-level adoption and estimate net benefits ranging from USD1.2 to USD166 million. Zohrabian et al. (2003) calculate a benefit-cost ratio for investing in an additional accession to prevent losses from a single pest in the range of 36 to 61. This confirms that the expected marginal benefit from exploring an additional unimproved genebank accession for breeding resistant varieties of soybean more than covers the costs of acquiring and conserving such collections in the U.S. national plant germplasm system.

Option value is also highest when genebanks maintain material that is held nowhere else. Much of the material in the international collections was collected decades ago, before the modernization of agriculture around the world. We know that some of that diversity can no longer be found in the field. Perhaps most importantly, however, the value of collections is influenced by the geographic origin of the material. Genetic diversity within a crop is richest in its center of origin. Multiple crops have centers of origin concentrated in a relatively small number of specific geographical areas worldwide. Coupled with the globalization of agriculture and food systems, this has resulted in all countries being interdependent for crop diversity. Such interdependence is a key reason why large international genebanks have particularly high option value (Box 2).

Isolating the contribution of genebanks

Few studies have succeeded in attributing a dollar value to the specific contribution made by genebanks to crop improvement. The difficulty of this task leads Smale (2006) to conclude that the economic benefits of using crop diversity in breeding can probably not be calculated with accuracy, but are so great that they certainly far outweigh the costs (Box 3) of long-term conservation and maintenance in genebanks. What has been made clear in recent studies is that the availability of diverse germplasm for evaluation in multiple environments plays a key role in the success of crop improvement programs. A study commissioned by CGIAR’s Standing Panel on Impact Assessment estimates that a total of about USD800 million (in 2002 dollars) was spent by CGIAR on germplasm collecting, conservation, characterizing and evaluation (GCCCE) activities in the years from 1970 to 2010 (Robinson and Srinivasan, 2013). A 2012 study of rice varietal releases, for example, reveals that 100% of IRRI rice varieties and 90% of rice varieties released by national programs had at least one IRRI genebank accession in their pedigrees (CGIAR, 2013). Similarly, N.L. Johnson et al. (2003a) show that nearly 60% of the 411 bean varieties released since 1976 contain materials from the genebank of the International Center for Tropical Agriculture (CIAT). A significant share of the impact of CGIAR research and breeding can plausibly be attributed to the international genebanks. The results from detailed case studies – rice in Asia (Box 4), the cassava Kasetsart 50 (KU 50) in Thailand (Box 5), and the potato Cooperation 88 in China (Box 6) – suggest that the benefits accrued from the adoption of a few improved varieties alone would together pay for CGIAR’s 40-year investment in GCCCE activities.

Beyond the research programs on crop improvement, CGIAR has also been instrumental in helping to rebuild agricultural systems in at least 47 developing countries affected by conflicts and natural disasters across Asia, Africa and Latin America, including through the restoration of crop diversity (Varma and Winslow, 2005). The economic value of such contributions has not been estimated, but this is a clear additional benefit beyond the contribution of crop diversity to improvement programs. Moeller and Stannard (2013) document the assistance given to countries in Asia after the 2004 tsunami, which drew attention to salinity problems in paddy rice cultivation. The International Potato Center (CIP) has been working for many years with the communities of the Parque de la Papa in Peru to restore lost ancestral potato varieties (CIP, 2012). The cultural importance of such interventions is impossible to value in dollar terms.

Box 2. Global interdependence in genetic resources

Global interdependence supports the rationale for considering crop diversity as a global public good and offers a strong argument for a more comprehensive participation of countries in the Multilateral System of Access and Benefit-Sharing of the International Treaty (Halewood et al., 2013). Khoury et al. (2014b) provide a dynamic estimation of countries’ interdependence in crop diversity from 1961 to 2009. They find that countries strongly depend on crops whose genetic diversity originates from foreign regions, both in their food supply (with an average 66% dependence on foreign crops for calories, 67% for protein, 74% for fat and 69% for food weight across countries worldwide) and production systems (71% for production quantity, 64% for harvested area and 73% for production value). Dependence on foreign crops is highest in countries that are geographically isolated or located at a great distance from the primary regions of diversity of major staple crops, such as Australia and New Zealand, the Indian Ocean Islands, the Caribbean, southern South America, North America, southern Africa and northern Europe. While these countries are generally in temperate climates, some continental tropical regions, such as Central Africa, also have very high levels of dependence. Moreover, the dependence on crops that originated in other regions has increased over time. Countries with the greatest increases in dependence over the past 50 years were located in Africa; West, South, Southeast and East Asia; Central America and Mexico; and Andean and tropical South America.

Galluzzi et al. (2015) analyzed international movements of crop diversity facilitated by the genebanks of seven CGIAR centers from 1985 to 2009. This study also showed strong global interdependence, with dozens of countries both contributing to, and benefitting from, the international genebanks. Similarly, Halewood et al. (2013) show that both developed and developing countries are net recipients of crop diversity, receiving more diversity than they contribute to others through the international genebanks. The top providers of crop diversity are developing countries in important centers of crop origin, domestication or diversification. However, many top recipients are also developing countries in centers of origin or diversity of crops. Further, in both developed and developing nations, the main recipients of CGIAR germplasm are public institutions, including national agricultural research centers, national genebanks and universities. The analysis also found differences in the types of materials provided by developed and developing countries. While developed countries provide an overall lower quantity of materials compared to developing countries, they contribute a proportionally higher share of advanced materials, on which some formal research, pre-breeding or other form of improvement has been conducted. In the end, the studies confirm that no country is self-sufficient for the crop diversity needed in agricultural production.

Box 3. Costs of genebanks

The operational costs of a number of international genebanks have been studied in detail by Hawtin et al. (2011), Koo et al. (2003) and Pardey et al. (2001), among others. The costs of operation may be kept as low as US$ 1 per accession if a genebank undertakes only the minimum activity required to keep seeds in a cold room or freezer. However, other important costs must be considered, such as health testing, disease cleaning, information management and characterization. Horna et al. (2010) establish that the reproductive biology of the crop is the major determinant of the scale of the costs, varying widely between outcrossing and self-pollinated species and between seed and vegetatively propagated crops.

The most recent costing study (Hawtin et al., 2011) estimates per-accession annual operating costs (not including capital costs) at US$ 3 for wheat, compared to US$ 33 for tropical forages, with an overall average of US$ 12 for seed accessions. Clonal crops are substantially more expensive to conserve and distribute because of the labor-intensive nature of field and in vitro conservation; the average per accession cost is at least ten times higher than for seed crops. A rational duplication of conserved samples must also be considered for breeding and research institutes in different countries to have ready access to popular material (Hodgkin et al., 1992; van Treuren et al., 2010).

Considering the scale of the cost of conserving 7.4 million accessions worldwide, while improving the coverage of numerous genepools, especially with regard to crop wild relatives which contain considerable amounts of untapped genetic diversity, the rationalization of conservation efforts should be a global priority (Engels and Visser, 2003; FAO, 2010).

Ex situ collections remain the best and most cost-efficient way to study, store, document, share, pass on, and make available the widest possible crop diversity to the widest possible audience of users, researchers, and breeders. (Lusty et al., 2014)

Box 4. IRRI and rice in Asia

As the staple food for more than 3.5 billion people worldwide, rice is one of the best documented crops. In Asia, where many people eat rice two or three times a day, rice contributes 30% to 70% of calorie intake. Future crop failures due to extreme weather events or pest and disease outbreaks could spell disaster for millions. Evenson and Gollin (1997) trace the genealogies of rice varieties released by national programs and IRRI from 1965 to 1990. They estimate the value of a landrace added to the IRRI genebank to be as high as US$ 50 million (in 1990 dollars), and an addition of 1,000 catalogued accessions to be associated with the release of 5.8 additional varieties, which would generate a present value income stream of US$ 325 million (in 1990 dollars), assuming a delay of 10 years and a 10% discount rate.

Box 5. CIAT and cassava in Thailand and Vietnam

Kasetsart 50 (KU 50) is a high-yielding cassava variety developed through collaboration between CIAT and the Department of Agriculture and Kasetsart University in Thailand. KU 50 is currently grown on over one million hectares in Thailand and Vietnam and has also been adopted in Indonesia and Cambodia. It was developed to escape the poor yields associated with the narrow genetic base of established cassava varieties in Thailand. KU 50 has also been successfully used as a parent in crosses that have produced several cultivars that are currently being adopted in Southeast Asia.

KU 50’s pedigree shows it to be a selection from hybrid seed produced from a cross between Rayong 1 and Rayong 90, the latter of which was the product of a cross between CMC 76 and V 43 (Figure 2). CMC 76, a key node in the pedigree of KU 50, came frCIAT and cassava in Thailand and Vietnamom the CIAT genebank (collected in Venezuela in 1967) and was selected by CIAT cassava breeders during their evaluation of genebank accessions. Extensive programs of evaluation and selection, conducted over 30 years at many sites in Colombia and Thailand, led to the eventual development of KU 50 and other high-yielding cassava varieties.

It is estimated that the aggregate economic benefits accruing from adoption of KU 50 exceed USD44 million in Thailand (where it was released in 1992) and USD53 million in Vietnam (where it was released in 1995), at adoption levels of 60% and 75%, respectively. Moreover, KU 50 has had a substantial impact on poverty alleviation in Thailand and Vietnam through the surpluses accruing to cassava growers. Such an impact would have been very difficult to achieve without the use of cassava germplasm conserved in the CIAT genebank in Colombia. No other institute would have been able to screen such a broad range of cassava genetic diversity, and thus identify and provide the key accession CMC 76.

Source: Robinson and Srinivasan (2013)

Box 6. CIP and potato in China

Cooperation 88 (C88) is a high yielding potato variety developed by the Chinese national agricultural research system and CIP to improve late blight resistance in potato adapted to sub-tropical highlands. It is grown on about 400,000 hectares in five provinces of southwestern China, where it has replaced Mira, a variety of German origin which has become increasingly susceptible to late blight and viruses. Late blight, a fungal pathogen, is considered the most serious threat to potato production, accounting for more than USD 1 billion each year in lost production and costs of control.

Breeders from CIP and Yunnan Normal University jointly evaluated the germplasm for the maternal parent of C88, while the male parent was derived from potato crosses made in the Philippines. The potato seed was evaluated in China, and after five years of trials and selection, clone #88 was identified as a high-yielding and late blight resistant variant which was adapted to longer day growing conditions. The variety was named Cooperation 88 and launched in 1996. A large part of this genebank material was collected in the center of origin of potato in the South American Andes, and thus provided the unique possibility to substantially broaden the genetic basis of potato in China.

It is estimated that the economic benefits accruing from C88 in China at the level of adoption in 2010 were USD350 million, and will increase to USD465 million per year if farm-level adoption increases to 600,000 hectares. More than half of the economic benefits are estimated to accrue to the poor (between USD 192 and USD 256 million). Moreover, C88 has stimulated growth in the potato processing industry because of its suitability for both table and chipping purposes. Such benefits would not have been possible without the use of germplasm conserved in the genebank at CIP.

Source: Robinson and Srinivasan (2013)