Concentration in Seed Markets

7. Policy responses

Abstract

The debate regarding concentration in seed markets has focused on the recent mergers and how competition authorities should respond. This chapter discusses the concerns, methods and policy responses of competition authorities around the world. It then outlines several complementary policy options available to policy makers to guarantee an innovative and competitive seed industry. These include avoiding unnecessary barriers to entry, facilitating access to intellectual property and genetic resources, and stimulating public and private R&D.

7.1. Competition authorities and the seed industry

In November 2017, the OECD distributed a brief questionnaire to competition authorities of several countries with important seed markets (Box 7.1). This questionnaire asked whether the competition authority was, or had been, investigating any of the recent mergers and acquisitions in the seed and agrochemical industry. If so, they were asked to indicate the key concerns these had raised; the measures and methods used to evaluate the impact of the transactions; potential remedies (such as divestitures) considered; and whether there had been any coordination with counterparts in other countries before a final decision was reached.

In view of ongoing investigations, some competition authorities declined to answer the questionnaire. This was the case for the United States and the European Commission. One country provided input under the condition of anonymity. Responses were received from Argentina, Brazil, Canada, Chile, Japan, Mexico, New Zealand, the Russian Federation, South Africa, and Korea. In general, all competition authorities declined to answer questions regarding ongoing investigations; in these cases, additional information was obtained from public statements after the merger investigations had been completed.

Box 7.1. OECD questionnaire for competition authorities

In November 2017, the OECD sent a questionnaire to 15 competition authorities to learn about how they are responding to consolidation in the seed industry. The competition authorities contacted were those of Argentina, Australia, Brazil, Canada, Chile, the European Union, India, Indonesia, Japan, Mexico, New Zealand, Korea, the Russian Federation, South Africa, and the United States. No response was received from competition authorities in India and Indonesia. The competition authorities in the United States and the European Union also declined to comment at the time given the ongoing Bayer-Monsanto investigation, while several other competition authorities completed the questionnaire without commenting on ongoing cases. After the approval of the Bayer-Monsanto merger in the spring of 2018, public statements by competition authorities have been used to complement the information from the questionnaires.

The questionnaire included the following questions:

• Has the Competition Authority formally investigated any of the mergers currently underway in the global seed industry (Dow-DuPont; ChemChina-Syngenta; Bayer-Monsanto; any others)?

• If so, what were the key concerns of the Competition Authority? (e.g. impact on prices; impact on variety/choice; impact on innovation; potential barriers to entry).

• What were the measures used by the Competition Authority? (e.g. C4 concentration ratios; Hirschman-Herfindahl Index; existence of a sufficient number of competing firms with R&D capabilities). What were the analytical methods used by the Competition Authority to assess the impact of the transaction on the market? (e.g. a formal merger simulation; an analysis of the R&D pipeline of the parties; a consultation with stakeholders).

• If the Competition Authority found any competition concern, did it consider the possibility to accept remedies? What types of remedies (if any) were considered? (e.g. divestiture of certain businesses or R&D capabilities; mandatory licensing of intellectual property).

• Did the Competition Authority coordinate with Competition Authorities in other countries? If so, which ones and on which issues?

The questionnaire asked competition authorities to provide contact details of experts who would be willing to provide further information. The OECD gratefully acknowledges the co-operation of competition authorities in providing background and information for the purpose of this study.

Concerns

In several cases, the competition authorities noted that the mergers would have no noticeable impact on the seed markets. The ChemChina-Syngenta transaction in particular was mentioned as having little or no effect on seed markets. (As noted previously, this transaction did raise concerns around competition in agrochemical products, and both the US and European competition authorities required the divestiture of several pesticide products.)

In Korea, none of the three transactions were considered to have a meaningful effect on the market. For the DowDuPont merger, the Korea Fair Trade Commission imposed remedies for a (non-agricultural) petrochemical product only, while the other transactions did not even meet the requirements for notification to the competition authorities. In New Zealand, authorities have been monitoring the global mergers but have not initiated any formal investigations.

The DowDuPont merger was also judged to pose no risk in the Russian Federation Chile, Mexico and Japan. In Canada, the competition authority was concerned only about potential effects on the agrochemical sector.

In South Africa, the Competition Commission concluded that the DowDuPont merger could negatively affect maize seed prices, choice, and product choice, as well as posing a risk to innovation and entry into the industry.

Similarly, in Brazil the competition authority identified risks of a DowDuPont merger for competition in maize seeds. In some segments, the combined market share of DowDuPont would have reached 40-50% or higher. DowDuPont’s combined germplasm banks are well-adapted to Brazilian soil and weather conditions, while there were concerns around the capacity of other players in the market to effectively compete with DowDuPont after the merger. In addition, the competition authority noted the high barriers and long time span required for new players to enter the market. In the end, the DowDuPont transaction was approved conditionally on the divestiture of a significant part of Dow’s Brazilian business and R&D capabilities.

As noted previously, US and European competition authorities had concerns around Dow and DuPont’s agrochemical business and required divestitures, most notably of DuPont’s global R&D organisation.

The Bayer-Monsanto merger raised more concerns. In addition to effects in seed markets, several competition authorities were worried about the market for broad-spectrum herbicides, where Bayer’s glufosinate-based product is an important alternative to Monsanto’s glyphosate-based herbicide. Moreover, some competition authorities also feared a loss of competition in the emerging field of digital agriculture.

In South Africa, the Competition Commission identified potential harmful effects of a Bayer-Monsanto merger on cotton seed prices and product choice, as well as the potential for reduced innovation in biotechnology traits, especially for maize.

In Brazil, the competition authority indicated risks of a Bayer-Monsanto merger through both horizontal and non-horizontal effects. Horizontally, there was a risk of creating a large combined market share in cotton and soybean regarding biotechnology traits as well as in terms of production and commercialisation of varieties. Non-horizontally, the competition authority noted that the merger would contribute to a further vertical integration of biotechnology and seeds markets. Without remedies, the merged firm would have been able to control other firms’ access to biotechnology and could adopt commercial practices to restrict competitors’ development. Moreover, the merged firm would have been able to use its strong market position to achieve prominence in complementary products. In this way, the combined firm would have been able to raise barriers to entry for new firms by making it necessary to enter different market simultaneously. The competition authority thus concluded that the merger would have impacts on prices, product choice and innovation for cotton and soybean seed if no remedies were imposed.

In the United States, negative effects on competition were foreseen in cotton, canola, soybean, and vegetable seed markets, as well as in seed treatment. In the European Union, areas of concern were cotton, rapeseed, and vegetable seeds, in herbicide tolerance and insect resistance traits, and in seed treatment. In Mexico, Bayer and Monsanto would have held a near-monopoly on GM cotton seed and a dominant position in vegetable seed. In Canada, the Bayer-Monsanto merger was judged to reduce competition in seeds and traits for canola and soybean, as well as carrot seeds and seed treatments against nematodes.

The European Commission expressed concern with regard to seed markets for vegetables, rapeseed, and cotton, as well as GM traits, non-selective herbicides, seed treatment, and digital agriculture (European Commission, 2018[134]).

The competition authorities in the Russian Federation expressed concerns about the Bayer-Monsanto merger, noting that it would take place after two other transactions had already reshaped the global industry. The competition authority noted Bayer and Monsanto’s exclusive technologies and strong positions in plant breeding as well as in digital and precision agriculture. As a result, the competition authority considered that the merger posed threats for competition by creating and strengthening barriers to entry (including regarding exclusive digital platforms for precision agriculture), by increasing incentives for anticompetitive concerted behaviour, and by creating the possibility for the combined firm to abuse its market power. The Russian competition authority ultimately approved the deal conditional on a number of remedies (discussed further below).

Measures and methods

The competition authorities generally reported using very similar analytical approaches to evaluating the mergers.

Most competition authorities start by analysing the effect of mergers on combined market shares, as well as on concentration measures such as the four-firm concentration ratio and/or the Hirschman-Herfindahl Index. Some competition authorities (e.g. Chile) emphasised that these techniques are merely a first approximation or “screening”. In terms of quantitative measures, some competition authorities (e.g. Canada) sometimes use more advanced analytical techniques such as regression analysis and merger simulation.1 Consultations with stakeholders were mentioned by nearly all competition authorities. Such consultations typically involve competitors and customers as well as other interested parties.

Several competition authorities (e.g. South Africa, Brazil, the Russian Federation, Mexico) also explicitly mentioned an analysis of the R&D pipelines of the merging firms and/or an evaluation of the viability of alternatives to the merging firms’ products, including foreign competition. The competition authority for New Zealand noted that, while it is not formally investigating the mergers, previous analyses of the seed industry have focused on the ability and incentive of other players (notably foreign firms) to enter the market.

Remedies

The term “remedies” refers to measures used by competition agencies to resolve and prevent harm to the competitive process that may result from a merger (OECD, 2011[90]). Typically, a distinction is made between structural and behavioural remedies. A structural remedy requires the divestiture of an asset, while a behavioural remedy imposes an obligation to engage in, or refrain from, a certain conduct. The sale of a part of the business or the transfer or licensing of intellectual property rights are considered structural remedies. Examples of behavioural remedies include non-discrimination obligations, transparency provisions, or limitations on what firms may include in contracts with customers.

In response to the questionnaire, most competition authorities did not mention any remedies regarding the ChemChina-Syngenta transaction, as there was only limited overlap in the activities of both firms. The main exception concerned several pesticide products where ChemChina’s Adama subsidiary was an important generic competitor to Syngenta’s products. Both US and European authorities required divestitures in this area.

The DowDuPont merger attracted scrutiny in several jurisdictions (including the European Union, the United States and Canada) regarding potential overlap on herbicides, insecticides and fungicides as well as some non-agricultural products, leading to divestitures in these areas. In other jurisdictions, there were additional concerns regarding seed and GM markets, with corresponding remedies. South Africa required DuPont to license intellectual property rights for maize seed to third parties. In Brazil, the remedies included the divestiture of large parts of Dow’s Brazilian business and R&D capabilities, particularly regarding the maize seed business. (Dow subsequently sold its Brazilian maize seed business to the Chinese firm Citic Agri Fund Management for USD 1.1 billion).

In the Bayer-Monsanto case, Bayer committed to selling a number of businesses to BASF. These include its global business in glufosinate tolerance traits and glufosinate herbicide (known under the “LibertyLink” and “Liberty” brands), its global vegetable seeds business and almost its entire field crop seed business (including rapeseed, cotton, soybean, and wheat). In addition, Bayer sold some seed-treatment assets and its digital agriculture portfolio to BASF, as well as three alternatives for glyphosate currently under development.

These divestitures addressed a large part of the competition concerns across jurisdictions, but some competition authorities required additional measures. In Russia, the competition agency required Bayer and Monsanto to provide access to technologies and databases under non-exclusive licenses and non-discriminatory access of seed producers, crop protection producers, and other parties to Bayer and Monsanto’s digital platforms for precision agriculture. Similar measures were required by authorities in China (Reuters, 2018[158]). In Chile, Bayer is barred from offering exclusivity-related rebates or from using tying or bundling with respect to Roundup herbicide (McLennan, 2018[159]). In India, the Competition Commission required the firm to provide broad-based, non-exclusive licensing of GM and non-GM traits and non-selective herbicides. The firm is also required to provide non-exclusive access to data, platforms and applications related to digital agriculture, and needs to grant the Government of India free access to Indian agro-climatic data for public purposes (CCI, 2018[160]).

Competition authorities therefore used a broad range of tools to remedy possible risks to competition. In addition to requiring the divestiture of certain product lines, whole businesses, or even global R&D organisations, some jurisdictions have used requirements around licensing, non-discrimination and commercial practices.

Coordination

Given the global nature of the mergers, competition authorities engaged in international coordination. Such coordination was highlighted by most of the competition agencies. In its press release announcing the approval of the Bayer-Monsanto merger, for instance, the European Commission explicitly mentioned co-operation with competition authorities in Australia, Brazil, Canada, China, India, South Africa, and the United States (European Commission, 2018[134]).

Topics of coordination in the seed mergers included market definition, approaches on the assessment of the consequences of mergers, and possible remedies. Coordination on remedies is particularly useful, as it allows firms to propose a logically consistent set of divestitures, which minimises transaction costs. In the Bayer-Monsanto merger, Bayer was able to propose the sale of coherent business assets (e.g. its global glufosinate business, its global vegetable seed business) to address concerns in various jurisdictions, rather than having to negotiate a patchwork of measures with different competition authorities.

Competition authorities often cooperate on mergers of multinationals, and such co-operation is beneficial for the competition authorities as well as for the merging parties. However, successful coordination requires the co-operation of the merging parties themselves. For instance, documents provided to the competition authority in one country typically fall under strict confidentiality rules, and merging parties need to agree to a waiver to allow the competition authority to share this information with other competition authorities (Competition Bureau Canada, 2014[161]).

Successful co-operation may also require institutional changes. In Canada, a change to the Competition Act in 2009 aligned the Canadian merger review process more closely with US procedures (Competition Bureau Canada, 2014[161]). In Argentina, prior to the adoption of a new competition law in May 2018, the merger review process worked on an ex post basis, where companies notified the competition authority only after the mergers were concluded (and hence approved by competition authorities in other jurisdictions). The new competition law changes the merger regime to an ex ante system, which will allow better co-operation with other competition authorities.

Conclusion

The results of the survey among competition authorities broadly confirm the insights presented earlier. To a large extent, the mergers combine complementary firms, with limited concerns about horizontal effects in many markets. In some markets (e.g. Korea, New Zealand) there was no formal investigation of potential impacts on seed markets, as the merging firms do not constitute a significant presence in the market. In other markets, concerns were raised about competition in certain segments or products. The importance of innovation was highlighted by the fact that many competition authorities analysed the R&D pipelines of merging firms as part of their investigation. Both horizontal and non-horizontal effects were mentioned by competition authorities as potential areas of concern. Competition authorities tend to coordinate with each other about the remedies needed to resolve these issues. These were typically structural remedies such as the sale of certain businesses or products.

In general, competition authorities appear to have a broad range of tools to prevent or limit risks to competition of a merger, from challenging a merger completely to requesting targeted structural or behavioural remedies. Such competition policy in the narrow sense is not the only way to safeguard or stimulate competition, however. Laws and regulations often inadvertently restrict competition, as highlighted in the OECD’s Competition Assessment Toolkit (OECD, 2016[162]). Conversely, other policies may stimulate competition by creating a more level playing field. Attention to the competition impact of such policies is complementary to competition policy in the narrow sense. Competition policy is often restricted to a “reactive” role, for instance intervening when market power is abused. Complementary policy options allow a more “proactive” approach and can prevent competition problems from emerging in the first place. Some complementary policy options are presented in the next section.

7.2. Complementary policy options

Several policy options exist to safeguard or maintain competition in the seed industry. This chapter reviews some possibilities. Given the complexity of the seed industry, and the differences in organisation of the sector across different countries, not all of these options are equally feasible or desirable for all countries. However, they do provide some suggestions regarding policy levers which can influence competition in the industry. A first set of policy options aims to avoid or reduce barriers to entry caused either by regulatory requirements. A second set of policy options enables entry by facilitating access to intellectual property and genetic resources. A final set of options aims to stimulate R&D through either public investments or through other policies that stimulate private R&D.2

Avoiding unnecessary regulatory barriers to entry

While a sound regulatory framework is necessary to ensure markets function properly, regulation may also inadvertently create transaction costs and barriers to entry. An important goal for policy makers is therefore to evaluate how to achieve valid policy objectives (e.g. in terms of protecting human health and the environment) without unduly restricting competition or innovation (OECD, 2016[162]). At the same time, the regulatory framework must also convince consumers of its effectiveness to ensure consumer acceptance, as a weak regulatory framework could lead to a lack of public trust which would in turn reduce demand and hence innovation.

In the context of innovation in seed markets, two relevant dimensions involve the regulatory environment for New Plant Breeding Techniques and the costs of regulatory science.

Regulatory environment for NPBT

Plant breeding techniques have seen rapid developments in the last few years. Several so-called New Plant Breeding Techniques (NPBT) have emerged, which potentially enable more precise, faster, and therefore more cost-effective plant breeding. NPBT refers to a broad set of different techniques, including cisgenesis, intragenesis, genome editing through Site-Directed Nuclease (SDN), and reverse breeding, among others. A brief introduction is provided in Box 7.2, based on Schaart et al. (2015[11]).

For several of the techniques, the resulting product may be similar to varieties generated using traditional breeding techniques, but some processes may rely on genetic modification along the way. For instance, cisgenesis introduces a new gene into the genome of a plant, similar to transgenic genetic engineering; but in contrast with transgenic techniques, the new gene is taken from the same plant species. The resulting genome could therefore potentially also arise through traditional plant breeding. Similarly, SDN techniques can lead to the “knock-out” of a gene or its targeted modification or replacement, both of which could be achieved through traditional plant breeding (e.g. through mutagenesis).

From a policy perspective, a key question on NPBT is therefore under which regulatory regime these techniques (or more properly speaking, the products which result from them) should fall (Laaninen, 2016[163]). It is important for policy makers to strike the right balance between providing sufficient safeguards to protect public health and the environment while avoiding excessive regulatory burdens that would slow or stall these important innovations, and while ensuring consumer trust in the regulatory framework. These questions are studied in ongoing OECD work in the Environment directorate under the auspices of the Joint Meeting of the Chemicals Committee and the Working Party on Chemicals, Pesticides and Biotechnology.

In March 2018, the United States Department of Agriculture announced that it would not regulate “plants that could otherwise have been developed through traditional breeding techniques,” provided that they are not “plant pests or developed using plant pests” (USDA, 2018[164]). This statement implies that new plant breeding techniques such as genome editing will not fall under the same level of regulatory scrutiny used for genetically modified organisms.

In the European Union, it was unclear whether NPBT fell under the existing legislation on genetically modified organisms (EU Directive 2001/18/EU on Deliberate Release of Genetically Modified Organisms). The existing legislation lists a number of techniques excluded from the regulation, including mutagenesis. However, the text does not precisely define mutagenesis, while the definition of a genetically modified organism itself is also somewhat ambiguous (Eriksson et al., 2018[165]). On 25 July 2018 the European Court of Justice clarified the interpretation of the Directive. The Court ruled that varieties obtained using the new plant breeding techniques are GMOs and hence fall under the same regulatory framework (Court of Justice of the European Union, 2018[166]). In reaching this conclusion, the Court deviated from the Opinion of the Advocate General (Court of Justice of the European Union, 2018[167]).

Box 7.2. New plant breeding techniques

The term “New Plant Breeding Techniques” covers a broad range of tools and techniques. Based on the characteristics of the final product after breeding, there are three broad possibilities:

• Plants with new genes

• Plants without new genes but with modifications of genes

• Plants without new genes and without modifications of genes

Several techniques result in plant varieties which could in principle also be obtained through traditional breeding techniques, although NPBTs are potentially faster and cheaper than traditional techniques. However, during the breeding process, NPBTs often use steps that could be interpreted as genetic modification, as (recombinant) DNA is introduced into the plant.

Cisgenesis is a technique similar to transgenic genetic modification, but using only genes of the species itself or from a closely related species. The technique can be used to transfer e.g. disease-resistance genes from wild relatives into commercially produced species. This can be done using traditional methods, but cisgenesis is potentially faster, as it can achieve the desired outcome without having to go through several generations of cross-breeding to introduce the gene into the commercially produced species. Hence, in principle, a cisgenic plant could also be achieved through traditional breeding.

Intragenesis is comparable to cisgenesis in its reliance on genetic material from the same or a closely related species. However, in contrast to cisgenesis, intragenesis introduces new combinations of elements of different genes. Changing these combinations can for instance combine a desirable gene with a stronger “promoter gene” that enhances the expression of the desirable gene. While this process still relies on genetic material of the same or a closely related species, an intragenic plant could not be achieved through normal breeding, as the recombination of elements is highly unlikely to occur in a traditional breeding process.

Genome editing through Site-Directed Nuclease (SDN) techniques (also called site-specific or sequence-specific nuclease techniques) comprises a number of different specific technologies with a common operating principle: exploiting the fact that the biological mechanism to repair broken DNA is itself relatively inaccurate. All SDN techniques therefore first deliberately create a lesion at a specific location in plant DNA. In some cases, the plant fails to accurately repair the lesion, causing the targeted gene to be “knocked out.” This application of SDN is referred to as SDN-1.

The plant’s own DNA repair machinery can also use a copy of the gene as a template for repair. A second application of SDN, referred to as SDN-2, exploits this fact by providing a modified copy of the gene as a template, inducing the plant to create the desired modifications during the repair process.

Finally, SDN-3 uses the same principle as SDN-2 but induces the plant to insert a large fragment (e.g. a complete extra gene) instead of merely a modified copy of the original gene. This latter approach can be used as a way to introduce new genes (whether transgenic, cisgenic or intragenic).

SDN focuses on a specific target DNA sequence and can hence be used as a precise tool to knock out (SDN-1), mutate (SDN-2), or replace (SDN-3) specific genes. Plants resulting from SDN-1 and SDN-2 could in principle be obtained through traditional methods using mutagenesis.¹ Compared with traditional techniques, SDN has the advantage of producing only a change at the desired location, whereas mutagenesis can create random mutations throughout the genome and therefore may require several generations of breeding to remove the undesired mutations.

SDN covers four main technologies: Zinc Finger Nucleases (ZFN), Meganucleases, TALENs, and CRISPR-Cas. One version of the latter, CRISPR-Cas9, appears to be particularly promising as it is relatively cheap, easy to implement, and versatile (Bortesi and Fischer, 2015[168]) (Belhaj et al., 2015[169]). While not an SDN-technology, Oligonucleotide-directed mutagenesis (ODM) works along similar principles as SDN-1 and SDN-2.

In contrast with the above techniques, RNA-dependent DNA methylation (RdDM) does not change the DNA sequence of the plant. Rather, it exploits an innate defense mechanism of plants which blocks the transcription of foreign DNA (e.g. from viruses). This mechanism is triggered by the presence of short double-stranded RNA molecules, as these could be degradation products of viral origin. The mechanism then locates the matching DNA sequence and attaches a methyl group to it. This deactivates the expression of the DNA while leaving the DNA sequence itself unchanged. The principle behind RdDM is to “trick” this defence mechanism into methylating a part of its own DNA. As this does not change the DNA of the plant, it is an epigenetic modification. However, the methylation is not stable; it tends to fade out over generations.

Reverse breeding solves the difficulty of reproducing a heterozygous offspring plant (i.e. a plant where chromosome pairs are not homogeneous). Such a plant cannot be reproduced from seed, as the next generation will be heterogeneous; the unique combination of characteristics is lost. Currently, such plants are reproduced vegetatively. Reverse breeding instead attempts to “reverse engineer” homozygous parental lines from which the heterozygous plant would be the offspring, similar to the way F1 hybrids are the offspring of true-breeding parental lines. The homozygous parental lines can be reproduced from seed, and the offspring plants will each time have the same desired genetic make-up. Reverse breeding uses a GM step to suppress recombination of chromosomes, although the gene used for this genetic modification is crossed out during the process. The end product does not contain genetically modified DNA.

Induced early flowering is a genetic modification used to speed up the process of plant breeding with fruit trees. For example, an apple tree takes five or six years before first flowering. This greatly slows down cross-breeding. By using genetic modification, an apple tree can be created which flowers within a year. This allows plant breeders to perform one breeding cycle per year. In the final step, the early flowering genes are crossed out, resulting in the desired variety, without genetically modified sequences in the final DNA.

Grafting on a GM rootstock involves letting the top part of a non-GM plant grow on a rootstock which has been genetically modified to improve characteristics such as resistance to soil-borne diseases.

As these examples show, the range of New Plant Breeding Techniques encompasses a broad range of techniques that can result in plants: with small genetic modifications, plants with new genes (transgenic, cisgenic, or intragenic); with unchanged DNA but where some genes are disabled through epigenetic modification; or without modifications to DNA but where genetic modification was used as a step in the breeding process. Policy discussions of the NPBT need to take into account the diverse nature of the techniques.

___________________________________

1. Mutagenesis is an approach where the plant breeder creates variation by deliberately triggering random mutations in the DNA of the species, for instance by exposing seeds to chemicals or radiation. The technique was developed in the first half of the 20^th century and widely used by the early 1990s. By 1990, around one thousand commercially-available crop varieties had been created through induced mutation, mostly via gamma rays and X-rays (Kingsbury, 2009[30]). A drawback of mutation breeding is that it often generates undesired mutations, which subsequently need to be bred out again.

Source: Schaart et al. (2015[11]).

Cost of regulatory science

The high costs of regulatory science for genetically modified varieties have been noted earlier. While these costs by themselves do not explain the current merger wave, it is plausible that high costs associated with regulatory science reduce the rate of innovation by increasing the cost of introducing new traits and genetically modified varieties. Miller and Bradford (2010[53]) have argued that high regulatory costs explain the relative scarcity of GM traits for specialty crops, as opposed to large-scale commodity crops.

One solution to reduce such regulatory burdens to innovation without reducing safety standards would be to let public institutions take care of some or all regulatory science. Such an approach would have the benefit of drastically reducing the fixed cost associated with the introduction of new genetically modified varieties, by 26% according to some estimates (Phillips McDougall, 2011[52]). Such an approach would stimulate innovation and would also enable the development of new GM varieties by smaller firms without the financial resources and regulatory expertise of large firms.

There are precedents for letting public institutions take care of the regulatory science. In jurisdictions where tests for Value for Cultivation and Use (VCU) are a requirement for marketing a new variety, the necessary tests are usually performed by public research institutions.

A clear drawback of this proposal would be the budgetary cost. Public agencies might be able to partly recover their costs using a levy on the sales of approved GM varieties, similar to commodity checkoff programs used to fund generic marketing. However, such a levy would reduce the expected revenues of new GM varieties, which could partly offset the benefit of lower regulatory costs in stimulating innovation. A more targeted approach could be to allow small and medium-sized plant breeders to rely on public institutes for assistance with the necessary tests and data generation. A related option is to provide subsidies, grants or loans for small and medium-sized plant breeders to cover some of the costs of regulatory science.

Other ways of reducing regulatory costs without weakening regulatory standards include simplifying regulatory processes and providing advice on the regulatory approval process tailored to small and medium-sized plant breeders, e.g. through liaison officers.

Facilitating access to intellectual property and genetic resources

Access to intellectual property and genetic resources has been mentioned as a major concern regarding market concentration in the seed industry. Yet, several policy options exist to facilitate access.

Licensing of proprietary genetic material

Under the UPOV convention, which forms the basis of legislation on plant breeders’ rights in many countries, an exemption is provided for other plant breeders to use varieties with a PBR in breeding programmes. This exemption provides other breeders with non-exclusive access to proprietary germplasm as long as this germplasm is protected only by plant breeders’ rights. However, specific genetic traits (and in some jurisdictions, including the United States, plant varieties as such) can be protected with a patent. In contrast with plant breeders’ rights, patents do not always provide a breeder’s exemption (Box 3.1). In that case, a plant breeder who wishes to include patented genetic material may need to negotiate a licensing agreement with the owner.

The risk exists that firms refuse to license their patented genetic material to a competitor. Moreover, for smaller firms, the transaction costs of negotiating a licensing fee may be considerable. Both to stimulate innovation and to safeguard competition, licensing should be made easily available. In particular, licensing should ideally happen in a non-discriminatory way, at low transaction costs, while providing sufficient financial rewards for the original innovator to stimulate investments in R&D.

A promising development is the emergence of “patent clearinghouses” in plant breeding. In November 2014, eleven leading companies from the vegetable plant breeding industry (jointly representing more than 50% of the global vegetable seed market) created the International Licensing Platform – Vegetables, with support from the Dutch government (Bruins (2015[170]), Kock & ten Have (2016[171])). The goal is to provide an easy way for plant breeders worldwide to access each other’s patented genetic traits for a reasonable licensing fee.

Members of the ILP make all their patents related to vegetable traits available to other members under the conditions of the ILP. When a breeder wants to use a trait patented by another ILP member, the parties engage in bilateral negotiations. If no agreement is reached within three months, both parties submit their own license fee proposal to independent experts. The experts can only choose the most reasonable proposal among the two (and hence cannot formulate a compromise). This method of decision-making, known as “baseball arbitration,” has the benefit of forcing both parties to formulate a reasonable proposal from the outset. The license fee chosen by the experts is subsequently communicated to all other ILP members to increase transparency.

This innovative approach allows access to traits at modest transaction costs, yet maintains incentives for innovation through the licensing fee. The ILP therefore provides “free access but not for free.” Moreover, the ILP uses a most favoured nation-style approach where a firm is required to grant a license to a member under the best terms it has granted to other members.3

A “patent clearinghouse” could be set up among private-sector actors without requiring any formal change in public policies. But, as was the case with the ILP, governments can play a facilitating role, for instance by taking the initiative or by encouraging private-sector actors throughout the process of setting up a platform.

Such initiatives deserve further study and would clearly help avoiding transaction costs for firms willing to share access to proprietary genetic material, but would be of little use when a dominant firm refuses to license its intellectual property to competitors. Under some circumstances, this might be considered an abuse of a dominant position in the European Union, and competition law may require compulsory licensing (Slaughter and May, 2016[172]). In the United States, compulsory licensing is not uncommon as part of merger remedies, but relatively rare outside of merger contexts (Delrahim (2004[173]), Pate (2005[174])). One notable example in recent years was the Microsoft case, where the US Department of Justice required Microsoft to license a range of intellectual property rights needed to create products interoperable with Microsoft Windows (Page and Childers, 2009[175]). In the seed industry, mandatory licensing has been used in the context of mergers. For example, when Monsanto acquired DeKalb in 1998, competition authorities required Monsanto to allow competitors access to patented technology on agrobacterium transformation of maize, as well as Monsanto’s stock of maize germplasm obtained during the earlier acquisition of Holden’s Foundation Seed company.

These examples all involve competition authorities acting in reaction to threats to competition. It is unclear whether policy makers in other domains are able to rely on similar tools to stimulate licensing. Outside of competition policy, the only prominent example of compulsory licensing involves provisions under the TRIPS agreement, which allow countries to procure “generic” versions of patented products under certain conditions. These rules are most commonly used for pharmaceuticals (Reichman, 2009[176]).

Some have proposed a more widespread use of compulsory licensing to ensure that plant genetic material remains easily available. Bjørnstad (2016[177]) has suggested an approach whereby plant breeders are allowed to use patented plant material, but are required to declare the patents used when they register a new variety. A fair royalty fee could be established by a system such as the ILP Vegetables example. This approach could also be used for patented processes and methods, as well as for material obtained through the International Treaty or the Nagoya Protocol (Box 7.3). Bjørnstad (2016[177]) emphasises the many outstanding questions around this proposal, including compatibility with existing laws and the TRIPS agreement.

Several jurisdictions already seem to have the necessary legal framework in place. In the European Union, the “biotech directive” (Directive 98/44/EC) provides for the possibility of compulsory cross-licensing (Article 12). The possibility exists for plant breeders who “cannot acquire or exploit a plant variety right without infringing a prior patent” and symmetrically also for holders of a patent concerning a biotechnological invention who cannot exploit it without infringing a prior plant variety right. The EU regulation requires that the applicant had unsuccessfully applied for a contractual license and that the new plant variety or invention constitutes considerable technical progress. The EU Directive requires Member States to ensure that in such cases a non-exclusive compulsory license can be granted in exchange for an appropriate royalty.4 A similar provision exists in Swiss law (Article 22a of the Swiss Plant Variety Protection Law and Article 36a of the Patent Law). If an applicant has unsuccessfully applied for a contractual license, a compulsory license can be granted by a court. So far, this possibility has never been used.

Rules regarding intellectual property need to strike a careful balance between providing incentives for investments in research and providing access to intellectual property. Proposals to facilitate licensing of intellectual property while remunerating the original inventor could benefit both competition and innovation in plant breeding, and therefore deserve further study.

Access to international germplasm

Access to genetic resources is essential for plant breeding, and it is therefore an important policy goal to maintain genetic diversity as well as providing easy access to this diversity. This creates a global policy challenge. Most countries grow food crops which originated in other countries or even other continents (Khoury et al., 2016[178]). Wild ancestral relatives or landraces which could be useful in plant breeding are therefore likely to be located outside of national borders (ISF (2012[179]), BSPB (2014[65])). This raises questions about how to ensure access to genetic resources, how to stimulate the preservation and sustainable use of genetic diversity, and how to organise benefit-sharing arising from the use of genetic resources, some of which may be the result of centuries of cultivation efforts by farmers.

These questions are governed by two international treaties (Box 7.3). The Convention on Biodiversity (1992) is the “default” treaty governing genetic resources in general for countries that have ratified it. The Convention recognises national governments’ sovereign rights over these resources, as well as the authority to determine access to them. The Nagoya Protocol (2010) further specifies the implementation of the “access and benefit-sharing” provisions of the Convention on Biodiversity.

For many agricultural crops, however, the relevant framework is provided by the International Treaty on Plant Genetic Resources in Food and Agriculture, which was adopted in 2001 and entered into force in 2004. For a list of 64 crops (known as “Annex I” crops) parties to the treaty agree to make genetic resources from local, national and international gene banks in the public domain available through a standardised procedure. Plant breeders consider that the International Treaty offers a more efficient access to genetic material for plant breeding, thanks to its standardised Material Transfer Agreements. For this reason, the International Seed Federation (2012[179]) has suggested delegating the Access and Benefit Sharing responsibilities for all genetic resources for plant breeding to the authority managing genetic resources under the International Treaty. One possibility could be expanding the list of crops in Annex I to cover all crops where breeding occurs, as well as other genetic resources used in breeding these crops (ISF, 2012[179]).

An important emerging policy issue concerns access to digital sequence information (DSI) (Halewood et al., 2018[180]). The current governance of genetic resources is based on access to physical material, and genetic information is likely to pose serious governance challenges within the current framework (Welch et al. (2017[181]). For instance, in its current formulation the treaties do not cover genetic information, so the terms and conditions for access to genetic resources and benefit sharing mechanisms do not apply. At the same time, an overly strict application of the existing legal access and benefit-sharing framework to digital sequence information risks creating excessive costs for researchers and plant breeders and significant obstacles to further research (Marden, 2018[182]). The International Seed Federation (2018[183]), for instance, is strongly opposed to regulating the access and utilisation of DSI.

In addition to ensuring an efficient and equitable system for accessing genetic resources, an important task is the preservation of these genetic resources. At a global level, the gene banks of the CGIAR Research Centers in particular play a significant role in conserving and providing access to the world’s genetic resources. At present, these gene banks contain 750 000 accessions, about 12% of all plant genetic resources conserved in gene banks.5 Between 2012 and 2016, the different gene banks of CGIAR distributed almost 600 000 samples of their accessions. In recent years, most Centres have reported an increase in distribution. For instance, between 1985 and 2009 nine CGIAR gene banks registered an annual rate of distribution of about 40 000 samples; the annual rate was 92 000 samples per year for 2012-2014. CGIAR expects this upward trend to continue given the rise of new technologies (CGIAR, 2017[184]). Of the materials reported as exchanged through the multilateral system of the International Treaty, more than half are genetic materials managed by CGIAR gene banks, in particular CIMMYT (accounting for 30% of all materials) and ICARDA (23%) (FAO, 2017[185]).6 In other words, a large part of the benefits of the International Treaty derive from the participation of the CGIAR gene banks, which constitute an international public good.7

The continued funding of these gene banks should therefore be an important priority for policy makers, especially since the benefit-cost ratio of doing so appears overwhelmingly positive while the magnitude of investments is modest. The annual costs of operating the CGIAR gene banks are estimated at around USD 6 million. As Koo et al. (2003[186]) note, these costs have to be compared with benefits estimated in the tens of billions of dollars due to higher crop yields made possible by the existence of these gene banks.

To organise funding in perpetuity, the United Nations Food and Agriculture Organization (FAO) and CGIAR have jointly established the Global Crop Diversity Trust in 2004. The goal of this Trust is to fund not only collections such as those of the CGIAR, but also those preserved by countries and regions, as well as maintaining a global last-resort seed vault, the Svalbard Global Seed Vault, located in Spitsbergen (Norway) and managed in collaboration with the Norwegian government and the Nordic Genetic Resource Center.8 The Crop Trust’s goal is to secure an endowment of USD 850 million overall, of which some USD 300 million has been secured already (Burwood-Taylor, 2016[187]). Conservation and sustainable use of global genetic resources is a crucial public good, and contributing to these initiatives should have a high priority for policy makers.

Box 7.3. International agreements on plant genetic resources

Two international treaties govern questions of access and benefit sharing (ABS) of genetic resources. The first is the Convention on Biodiversity (CBD), which entered into force in 1993. This convention recognises sovereign rights over genetic resources and the authority of national governments to determine access to those resources. The CBD also tasks governments with the responsibility of preserving genetic diversity. The Nagoya Protocol, which entered into force in 2014, is a supplementary agreement to the CBD and provides (among other things) a framework for the implementation of the “fair and equitable sharing of benefits arising out of the utilization of genetic resources,” one of the three objectives of the CBD.

Countries adopting the Nagoya Protocol commit to setting up a transparent, non-arbitrary process for access to genetic resources, as well as domestic benefit-sharing measures. The Nagoya Protocol also sets up an “Access and Benefit-sharing Clearing-House” to share information on topics such as domestic regulatory ABS requirements. The United States has not ratified the CBD, limiting the global impact of these provisions.

The second treaty is the International Treaty on Plant Genetic Resources for Food and Agriculture (ITPGRFA), which entered into force in 2004. This treaty is a more specialised agreement focusing specifically on plant genetic resources used in food and agriculture, with the Nagoya Protocol serving as a “default” for countries that have not ratified the International Treaty (but which have ratified the Nagoya Protocol).

The ITPGRFA sets up a system to govern access and benefit sharing of genetic resources for a specific list of 64 crops known as “Annex I” crops. This list of crops includes the major food crops (e.g. rice, maize, wheat, potatoes) but excludes some other major agricultural commodities, notably soybeans.

In terms of access, countries ratifying the ITPGRFA agree to make available their genetic diversity as well as information about the crops stored in gene banks (Hufler, 2009). Under the ITPGRFA, material in local, national and international gene banks, including the collections of the CGIAR, are put in a Multilateral System for Access and Benefit-Sharing. This global pool of genetic material is made available to users (e.g. plant breeders) under terms and conditions of a Standard Material Transfer Agreement.

In terms of benefit sharing, users accessing genetic materials through the ITPGRFA agree to share benefits from their use, following the conditions stipulated in the Standard Material Transfer Agreement. In particular, if users of genetic materials create products which have restrictions on further research and breeding, they agree to pay an “equitable share” of commercial benefits into a Benefit-Sharing Fund to support conservation and sustainable use projects, especially in developing countries (FAO, 2017). In addition, users agree to share non-monetary benefits through exchange of information, technology transfer and capacity building, among others.

The ITPGRFA manages the Annex I plant genetic resources as a pooled good, using the Multilateral System and the Standard Material Transfer Agreement as a low-cost way to enable access and benefit sharing. The advantage of this system is that it avoids bilateral negotiations between users and individual owners, thus greatly reducing transaction costs (Hufler, 2009).

Since its introduction, the ITPGRFA has enabled 4.2 million exchanges of genetic material through almost 60 000 Standard Material Transfer Agreements. The largest share of genetic resources (2.1 million accessions, half of the total) has been sent by countries in Latin America and the Caribbean, while most of the genetic resources transferred so far have gone to Asia (1.3 million accessions) (FAO, 2017). In addition to crops covered by Annex I of the ITPGRFA, the Multilateral System can also be used for other crops using a similar Material Transfer Agreement. Overall, around 4% of accessions to genetic resources have been for non-Annex I crops.

Despite the large number of successful accessions, problems remain. Bjørnstad et al. (2013[188]) sent seed requests to 121 countries that are Contracting Parties. They received no response from 54 countries, mainly in Africa and Latin America and the Caribbean, and concluded that the “facilitated access” promised by the ITPGRFA is not straightforward.

Open source germplasm

Some actors have proposed open source initiatives for access to germplasm, analogous to open source software (Kloppenburg, 2014[189]). In the most commonly used approaches to open source software, users can access the source code of the software, modify it, and build derived applications as long as those modifications or derived applications are distributed under similar open source permissions. The result is the creation of intellectual property in a “protected commons.” While there is no individual owner of the intellectual property, open source products cannot be appropriated as intellectual property by others. In such a setup, private seed companies could still multiply and sell open source seed, but would not be able to use intellectual property rights law to prevent competitors from selling the same seed or further improving the open source variety. Advocates of open source seed see this as a method to avoid the appropriation of germplasm by major corporations, and cite its potential value for safeguarding genetic resources in developing countries in particular. Some organisations have already emerged to promote such “open source seed” (Open Source Seed Initiative, 2016[190]).

An interesting characteristic of open source initiatives from a policy point of view is that they do not require any changes in the legal or institutional framework. Open source software operates using a “general public license” which relies on standard contract law; open source initiatives for germplasm can similarly be implemented when initial owners provide free access to germplasm under such a license. The only policy question in this setup relates to whether public plant breeding initiatives and other public organisations such as the CGIAR should adopt an open source approach to the distribution of their germplasm.

However, an important drawback of open source initiatives for germplasm is its incompatibility with other intellectual property rights regimes. For instance, the Open Source Seed Initiative (OSSI) allows certain restrictions and allows benefit-sharing arrangements (similar to royalty payments), but only when these restrictions are limited to the contracting parties. Restrictions (e.g. on seed saving or further breeding) cannot be passed on to customers under this system, and this makes “OSSI-Pledged” seed incompatible with current intellectual property rights for plant varieties. For advocates such as Kloppenburg (2014[189]) this is precisely one of the attractive properties of open source germplasm, as it allows the creation of “a mechanism for germplasm exchange that allows sharing among those who will reciprocally share, but excludes those who will not” (p. 1238). The open source seed system would therefore essentially be a parallel system to the commercial seed system, with little or no exchange of germplasm between the two domains. This limits the availability of germplasm in both systems. If public organisations such as CGIAR adopted this approach for distributing their germplasm, they would thus deprive private-sector plant breeders from an important source of genetic material. At the same time, given the long time lags and capital-intensive nature of plant breeding, it is doubtful whether an open-source approach could provide sufficient incentives for private investments in plant breeding.

Creating a generic market for GM traits

The expiry of patents can lead to the emergence of a generic market, as in pharmaceuticals or agricultural chemicals. A similar process could occur for patented genetic material, in particular for patented GM traits. Allowing the emergence of a generic market for GM traits would be beneficial as it would allow off-patent GM traits to become available to farmers at lower costs, while stimulating innovative firms to continue to invest in R&D (rather than benefiting from high prices for GM traits long after patents have expired).

However, some regulatory obstacles may prevent the emergence of such a generic market. Seed companies seeking to sell generic GM seed need to obtain regulatory approvals before the seed can be sold. The firms therefore need to submit a “data package” to the regulatory authorities, which requires either access to the original data submitted by the patent holder, or the ability to experiment on the patented material. However, patent laws may make it difficult to conduct research on the patented seed or trait for these purposes. The pharmaceutical industry has resolved this problem by creating an exemption under patent laws to allow collection of data for regulatory approvals before the patent has expired. Introducing a similar exemption for utility patents on biotech traits would be important to stimulate a generic market for GM traits (Schenkelaars, de Vriend and Kalaitzandonakes, 2011[29]).9

In the United States, private-sector actors have taken the initiative to introduce the “Ag Accord,” a pair of agreements setting up a legal framework for dealing with the expiry of patented GM traits: the Generic Event Marketability and Access Agreement (GEMAA) and the Data Use and Compensation Agreement (DUCA).10

The GEMAA was signed by large seed and biotech firms (DowDuPont, Bayer, BASF, Monsanto) and several farmer organisations (e.g. the National Corn Growers Association). It sets up an orderly process for dealing with patent expiration of biotechnology traits by creating clarity over who has responsibility for maintaining regulatory authorisations, and who has access to the proprietary data and other information necessary to obtain these authorisations (in the terminology of the GEMAA called the “proprietary regulatory package”).

Under the GEMAA, when a patent expires its owner can choose to maintain regulatory responsibility, seek to share responsibility, or discontinue. Under the first option, the original patent holder continues to be responsible for managing regulatory authorisations, at no cost to users of the (now-generic) biotechnology trait, but without providing access to the regulatory package. Under the second option, the firm can seek other firms willing to share the responsibility. These other firms then obtain access to the regulatory package. The third option, discontinuing regulatory responsibilities, builds in a seven-year transition period and starts a negotiation process with other firms for taking over the regulatory responsibilities and access to the regulatory package.

The DUCA is similar to the GEMAA, but provides an alternative mechanism for access to proprietary data and compensation for this access to the original patent owner. At present, this mechanism has not been implemented.

As with the International Licensing Platform – Vegetables discussed earlier, the “Ag Accord” is an agreement among private-sector actors and does not rely on any changes in public policies. But again, governments can play a facilitating role by encouraging such initiatives and bringing together the relevant stakeholders.

Stimulating public and private R&D

The policy options identified so far aim to stimulate competition and innovation by avoiding or reducing regulatory barriers and facilitating access to intellectual property and genetic material. Several other options exist to more directly stimulate innovation, including through public R&D.

Investing in public R&D

A large body of literature has documented high rates of return from public investments in agricultural R&D (e.g. Hurley et al. (2016[191]), Alston et al. (2010[192]), Alston et al. (2000[193]), Alston et al. (2000[194]) and Andersen (2015[195])). Given high levels of private R&D in plant breeding, what role should the public sector play? A useful conceptual framework to study the complementary roles of public and private R&D is the “Stokes-Ruttan paradigm” described by Heisey and Fuglie (2018[196]).

In this paradigm, types of research are classified according to whether the research is fundamental or applied, and whether there are explicit considerations for commercial use or not. This creates four possibilities. “Bohr’s Quadrant” includes fundamental research undertaken without any direct commercial objective; “Pasteur’s Quadrant” includes fundamental research with an ulterior commercial objective; “Edison’s Quadrant” contains applied research with a commercial objective; while the fourth category, “Rickover’s Quadrant,” includes applied research without a commercial objective.11

The private sector will tend to focus its efforts on Edison’s Quadrant, provided that its investments in R&D can be recovered through either sufficient protection of intellectual property rights or biological mechanisms (such as hybrid varieties). The historical role of the public sector in applied plant breeding was often due to a lack of private interest. Where the private sector has sufficient incentives and resources to engage in R&D, the public sector has tended to retreat from direct plant breeding (Heisey and Fuglie, 2018[196]).

It is possible in theory to maintain public plant breeding in Edison’s Quadrant strategically as a check on private-sector plant breeders. Kloppenburg (1988[14]) provided an early criticism of increasing concentration in the US seed industry and the diminishing role of public plant breeding. He advocated continued public efforts to develop and release finished plant varieties (as opposed to the trend towards fundamental, non-applied research), among other reasons to provide a “restraint on the activities of private industry” (p.285). The availability of affordable and high-quality public varieties would then serve as a check on the exercise of market power by large firms. However, where sufficient competition and innovation in a crop variety exists, public plant breeding would be a duplication of efforts, or might lead to a crowding-out effect where increased public plant breeding leads to a reduction in private-sector investments.

Public R&D may be more useful for exploring the potential of orphan crops or underutilised crops, for which there is less competition and/or less interest from private-sector breeders overall. Public R&D could also develop new varieties with positive environmental externalities. Similarly, public R&D could focus on biofortification, the practice of breeding new varieties with a better micronutrient content to combat malnourishment in the developing world.12 These are examples of applied research without a commercial objective (“Rickover’s Quadrant”) where public R&D has a role to play.

Public R&D might be similarly useful in Bohr’s Quadrant (fundamental research without a commercial objective) and in Pasteur’s Quadrant (fundamental research with a commercial objective), where the private sector will tend to underinvest. Public R&D efforts here could focus on pre-competitive research, e.g. on new breeding methods and tools, or on developing whole genome maps. By making the fruits of this research available in the public domain, public R&D can stimulate competition between different firms for the applications resulting from these more fundamental techniques and insights. Continued investments in such R&D might be useful to ensure that innovative techniques and tools (such as those underpinning the New Plant Breeding Techniques) remain accessible. Such an approach would also remove a “winner takes all” dynamic where the firm which first develops a new technique can in theory block its use by others.13

Public-private partnerships

There is a fundamental tension between stimulating private R&D and stimulating the widespread adoption of a useful innovation. With private funding of R&D through intellectual property rights, the innovator can temporarily charge a monopoly price for the innovation. This provides incentives for private R&D investments, but it limits broad and easy access to the innovation. On the other hand, if R&D is financed by the public sector, the innovation could be made available more cheaply, but the question is to what extent public funds should be allocated to it, and whether there is a way to make the ultimate beneficiaries of innovation pay for the R&D efforts. Public-private partnerships can under some conditions reconcile these conflicting goals. An interesting example in the case of plant breeding is industry-directed, levy-funded research such as the Saskatchewan Pulse Growers described by Gray (2012[197]).14

Saskatchewan Pulse Growers (SPG) represents over 18 000 pulse crop producers in the province of Saskatchewan (Canada) and is directed by seven farmers, elected by their peers. The scheme uses a mandatory 1% check-off (levy) on the value of the gross sale of all pulse crops, which is collected similar to a sales tax. An independent body appointed by the Ministry of Agriculture and Food supervises the activities. The revenues from the check-off are used to fund various projects such as research on pulse breeding, the provision of royalty-free seed, agronomic research, a Pulse Production Manual and efforts to increase domestic and international demand. Plant breeding funding from the SPG is directed to the University of Saskatchewan’s Crop Development Centre. The SPG also engages with private plant breeders. For instance, private firms may be given the right to distribute new varieties abroad in return for royalties; SPG also grants access to BASF for certain lentil varieties so the firm can incorporate its herbicide-tolerant traits. The SPG has been highly successful; Saskatchewan pea yields have increased by 40% in two decades and the internal rate of return on SPG investments has been estimated at 20% per year.

Compared with IP-protected private research, levy-funded research such as that by the SPG has three main advantages. First, if research is financed with a levy on farmers’ total output (regardless of the varieties used), royalties for the use of a new variety can be set at zero. In this way, farmers can adopt new, improved varieties more quickly and at a lower cost than under a ‘pure’ private system. Second, if all growers are paying for research, it makes sense for them to invest in R&D which would not be protectable under IP rules (e.g. research on agronomic best practices), and which would therefore not be provided by the private sector. Third, if this research is commissioned or undertaken by a single industry association, it avoids a possible duplication of research effort across different private plant breeders.

Levy-funded research also has several advantages compared with publicly funded research. First, as funds are managed by farmer associations, the system gives voice to end users. This ensures that research will benefit those who are paying for it. Second, as a semi-private initiative it may be easier to enter agreements with the private sector (e.g. by outsourcing certain types of research or by licensing intellectual property). Third, compared to public R&D, this organisational set-up does not require the use of public funds. Benefits are more likely to accrue when the levy system is mandatory. If the system is voluntary, some farmers may choose not to pay the levy and instead free-ride on the contributions of others.

Levy-funded research is not the only possible model for such public-private partnerships. In Australia, the Grains Research and Development Corporation was created in 1990 to provide levy-funded R&D (Gray, 2012[197]). Over time, the GRDC has explored new institutional mechanisms to organise R&D, in particular the development of for-profit partnerships with multinationals such as Monsanto, Syngenta, and Vilmorin (Limagrain). Such a setup can potentially tap private investment as well as managerial and technological capabilities while providing a check on over-pricing. An important institutional innovation was the introduction of end-point royalties, collected when the farmer sells the final product instead of when the farmer buys seed. Similar to the levy used by the SPG, end-point royalties ensure that breeders receive revenues even when farmers rely on farm-saved seed of the protected variety, while simultaneously lowering the initial cost for farmers and sharing the production risk between farmers and breeders.15

Despite these promises, farmer-funded research has often had disappointing results. The United Kingdom and the Netherlands have not had much success in stimulating R&D through such systems; in the United States, most producer levies are used for advertising and promotion instead of research to improve productivity. Successful examples often appear to rely on a combination of farmer contributions with matching public funds, to provide incentives to farmer associations to invest sufficiently in R&D (Heisey and Fuglie, 2018[196]).

7.3. Conclusion

The increasing levels of concentration in seed markets have led to concerns about the potential impact on prices, product choice, and innovation in the industry. As this chapter has illustrated, the recent mergers have been scrutinised by competition authorities to avoid harmful effects. Decisions by competition authorities, however, are not the only factor affecting seed markets; several other policies can influence competition and innovation in this sector. These policy options can be summarised as avoiding regulatory barriers to entry, facilitating access to intellectual property and genetic resources, and stimulating both public and private R&D.

Notes

← 1.  For an overview of analytical methods used in merger analysis, see OECD (2011[241]).

← 2.  As noted earlier, a broad range of policies affect the seed sector. Countries differ in the extent to which they have an integrated seed policy. Developing countries in particular often have explicit seed policies formulated in relation to strategies for agricultural development and food security. These policies may cover varietal development, seed production (multiplication), quality certification, agricultural extension, and/or stimulating the growth of a domestic seed industry. Some developed countries (e.g. Switzerland) similarly have a strategy for the development of their seed industry. In other countries, aspects of seed policy may be distributed over different agencies or ministries.

← 3.  In addition to patented traits, a second pillar of the ILP also covers patents on plant varieties through a mutual “non-assert” clause, effectively creating the equivalent of a breeder’s exemption for patented varieties. An ILP member can use a patented variety of another ILP member free of charge as the basis for further breeding, as long as the patent owner is notified and as long as the new variety is sufficiently different from the protected variety.

← 4.  A Commission Notice (2016/C 411/03) regarding this Directive notes the difficulty in demonstrating technical progress, and further notes that the European Commission may study this issue further. Bjørnstad (2016[178]) suggests that registration on the National List (after VCU testing) could be used as sufficient proof of such progress, so that the compulsory license can be granted almost automatically when a variety passes the VCU test.

← 5.  An accession refers to a “distinct, uniquely identifiable sample of seeds representing a cultivar, breeding line or a population, which is maintained in storage for conservation and use” (FAO, 2018[242]).

← 6.  However, it should be noted that nationally operated gene banks are not obliged to report on exchanges.

← 7.  A detailed overview of conservation and use of genetic resources, including information on the state of different collections worldwide, is provided by the Second Report on the State of the World’s Plant Genetic Resources for Food and Agriculture, published by FAO (FAO, 2010[237]).

← 8.  The global back-up already proved its usefulness. ICARDA, one of the CGIAR centres with a regional seed bank, was initially located in Aleppo in Syria. Due to the conflict in Syria, ICARDA moved to Lebanon and withdrew some of its back-up samples from the Svalbard vault to reconstitute the collections (Regan, 2015[247]).

← 9.  As noted previously, countries differ in the scope and strength of their patent law; in some jurisdictions research on patented products may be allowed under a “research exemption.”

← 10.  See www.agaccord.org.

← 11.  The quadrant is named after Admiral Rickover, who led efforts to develop a practical nuclear power plant to power naval vessels.

← 12.  See, for example, HarvestPlus, a joint initiative of the International Centre for Tropical Agriculture (CIAT) and the International Food Policy Research Institute (IFPRI), http://www.harvestplus.org/.

← 13.  Another mechanism through which public R&D can stimulate private R&D is by training future private-sector breeders. Public-sector R&D often takes place at universities, where researchers may also have teaching duties as part of their appointment. In the United States, some public-private partnerships exist between large plant breeding firms and universities, with training of future plant breeders as one of the objectives (Stephen Malone, personal communication).

← 14.  For a broader discussion of farmer-funded R&D, see Alston et al. (2012[253]).

← 15.  In 2015, Canada amended its Plant Breeders’ Rights Act to bring it into conformity with the 1991 UPOV Act. This reform includes the possibility to introduce regulations to improve remuneration mechanisms for breeders by placing conditions or restrictions on the use of farm-saved seed. The Canadian grain industry has suggested using an end-point royalty system or royalty collection on farm-saved seed via production contracts. At the time of writing, the Government of Canada has not yet decided how to proceed.