Lesson 2.3 Responsible Data Sharing in Agricultural Value Chains
Learning outcomes
At the end of this lesson, learners will (be able to):
Address the ethical and legal sensitivities of data-driven services where farmers’ data is shared through the value chain through awareness of legislation, consideration of contractual best practices and adoption of trust-building equitable business models
Negotiate fair conditions for the sharing of farmers’ data
Devise fair governance models of data sharing platforms and ways farmers can have a voice
Address issues of data availability through awareness of public data policies and strategies around private-sector data of public interest
Consider and contribute to the development of international guidelines
1. Introduction
This lesson outlines the policy spaces and instruments to be considered when dealing with farmers’ data sharing, especially addressing the issues. The policy spaces that are relevant here are different from those relevant for the open data lifecycle (see Lesson 2.1), as data shared along the value chain is normally not open and not designed for public use but for mutual transactions for the provision of specific services, which raises a different range of issues, mostly contractual. The issue of personal data is dealt with more in detail in the next lesson.
On the one hand, there is legislation. Some examples of relevant legislation areas will be outlined briefly, but it must be noted that many important aspects of agricultural data sharing (attribution, access, portability, interoperability, benefits, risk of lock-in) are not covered by legislation. Examples of legislation areas that are (tangentially) relevant are: digital policies in general and personal data protection laws in particular (see next lesson), open data policies, agricultural policies – especially with respect to the data sharing entailed by subsidies scheme, competition law, unfair trade practices.
Beyond legislation, there are self-regulatory instruments that can help negotiate fair conditions for this type of data sharing: this lesson will illustrate existing examples of such instruments, like codes of conduct and guidelines agreed upon by different actors in the value chain, and their potential role in making data sharing fairer. You will also see that some experts also envisage collaboration at the global level to create international guidelines for agricultural data sharing.
Finally, it is important to highlight the importance of governance of the components that form the “data ecosystem” that supports smallholders. The data platforms in the ecosystem can be managed by different actors and with different purposes: in most cases, farm data is still managed on technology providers’ platforms, but there are examples of farm data cooperatives and the potential role of Trust Centers.
2. Legal and policy spaces for addressing the challenges of farm data sharing
While there isn’t a dedicated policy space for agricultural data sharing, there are other policy spaces to be considered to have the full picture of instruments to use to ensure fairness of farm data sharing.
For instance, important issues like data protection, data portability, data standards, access to data, public data/open data, movement of data across countries are addressed by digital strategies, which are nowadays a common component of public policies, and related legislation, while issues related to extension services, interventions on agricultural value chains, subsidy compliance data etc. are normally addressed by agricultural policies. More indirectly, aspects like unfair data practices and power imbalances are under the influence of competition law in general and legislation on unfair trade practices in particular, although the specific data dimension is not addressed yet. More specific aspects of contractual fairness and fair data governance are addressed by self-regulatory instruments like codes of conduct or governance models like Private-Public Partnerships (PPPs).
Although public policies do not address agricultural data sharing explicitly and do not offer solutions for most of the issues highlighted in the previous lessons, it is useful to be aware of the existing policy spaces to understand where in the future these issues might be addressed and to be able to influence these policies and push for instance for a better coverage of: a) the data dimension in agricultural policies; and b) the value chain and data asymmetry dimension in digital strategies.
After a brief overview on policy spaces in general, in the next sections this lesson will focus on codes of conduct and innovative data governance options as potential stakeholder-led and bottom-up solutions to a more equitable data sharing in agriculture.
Table 1 Examples of policy spaces relevant for farm data sharing, linked to the issues highlighted in the previous lesson[1]
Data asymmetry/ conflict
Data flows/ technologies
Risks for actors
Policy space
Within private sector
Within value chain
● Farmers vs. Agricultural Technology Providers (ATPs)
● Farmers vs. processors and distributors
● Farmers vs. financial institutions
Farm data > ATP Precision agriculture
Farm data > supply chain IoT in some cases
Farm data > banks So far, submission, now initial use of satellite / sensor data
Very little raw data towards the farmer, only in the form of services / information, often through the ATP
Farmers:
● excessive data transparency, sensitive data more exposed than other actors’
● lock-in
● unfair contractual practices
● no capacity to process the data
● sharing more than receiving
Data protection (Data ownership / rights, business data) (IPR and copyright)
Agricultural policies (Interventions on agricultural value chain, extension services, subsidies for digital technologies, possible use of blockchain data)
Public data policies (More data for farmers to empower them in the value chain)
Contract law & Codes of conduct (Data rights, data sharing guidelines, contractual power)
Within same segment of value chain
● ATPs
● Data processors
● Any actor that aggregates data
Little flow, concentration Cloud, big data
New companies, small companies: unfair competition
Farmers: lock-in, unfair contractual practices
Competition law (Avoid monopoly, facilitate market entry, unfair trade practices)
Public data policies (More provision for levelling the playing field)
Private vs. public interest
Farmers vs. government
e.g. subsidies management
Farm data > farm registries
Farm data > ag policy monitoring
So far, submission, now initial use of satellite / sensor data
Farmers: excessive data transparency, sensitive data more exposed than other actors
Data privacy policies (Data ownership / rights, sensitive business data; use of public data for administrative simplification)
Agricultural policies (Consider sensitive data)
Private sector data holders vs. government
e.g. public-interest data, SDG monitoring, now also raw data for AI
Private sector data à Open data systems Very little so far, manual submission, APIs
Rare transfer of raw data so far, now initial idea of sharing raw IoT data
Private sector: risk of losing the competitive advantage of exclusive access to the data they have
Public data policies (Incentives for private sector to share, identification of pre-competitive spaces)
Public-private partnerships (PP data platforms, data sharing agreements)
Codes of conduct (Private sector data sharing agreements)
Public data for private sector
Public data to level the playing field, foster fair competition and stimulate innovation
Public data à private sector applications Open data, big data
Government: high investment, lack of high-impact/real-time data
Public data policies (prioritize high-impact data, big data platforms, real-time data)
PPPs (share investment, prioritize)
For all data asymmetry issues, given the impact on fundamental rights and the cross-border nature of data flows
International Treaties (idea of an IT on agricultural data flows)
2.1. Digital strategies
2.1.1. Data protection laws and non-personal data rights
Data protection used to be addressed as part of privacy law, or, in the case of consumer data protection, under trade law. Nowadays, the core issues of data protection concern online or digitally transmitted data and in countries or regions where there is a digital strategy, data protection is often addressed under that policy area.
Data protection laws normally protect personal data. For an in-depth analysis of the dimension of personal data in farm data sharing, see Lesson 2.4. Other types of data that are commonly protected by laws do not fall under the space of digital strategies, as those laws apply to all types of resources, also non-digital: for instance, copyright or trade secret protection (see paragraph below on Trade and competition law).
As long as non-personal data are not protected by copyright or trade secret, there is no legal framework under which it can be protected, so the rights we illustrated in the previous lesson (access, control, portability) are not enforced. In practice, there seems to be a tendency to extend some rights typical of personal data protection, like access, control or portability rights, also to non-personal-sensitive data. This can happen in contractual practice or, as we’ll see in the next chapter, in agreed codes of conduct, but is not covered by legislation.
One important aspect of personal data protection that is rarely extended to other sensitive data and that is very relevant for farm data is the right to portability. The right to retrieve and reuse the data, sometimes granted in contracts, is only a part of the full implementation of portability: in order to be reused, data should be standardized and interoperable, while currently machine-generated data is almost always in a format that is compatible only with the machinery and software sold by the technology provider. This is related to the issue of lack of interoperability between machinery of different brands.
For both agricultural machine and data interoperability, solutions come more from international technical organizations (like ISO, Agricultural Industry Electronics Foundation (AEF) or AgGateway) than from public policy. The so-called ISOBUS standard (ISO standard 11783) has become the de-facto interoperability standard between tractors and equipment from different manufacturers, while AgGateway and the Open Ag Data Alliance (OADA) provide good practices and specifications for farm data standardization
Another aspect that digital policies often cover, and is relevant for farm data, are data localization requirements: quite commonly, considering the different data protection regimes in different countries, such policies prescribe that initial collection, processing, and storage of data (primarily about citizens) occur first within the national boundaries, and in some cases that data about citizens stay in the country, or may be transferred only to countries that have the same level of data protection. Again, this applies normally to personal data and some other sensitive data (relevant for taxation or justice), but some countries have applied it more broadly. It is something important to consider, especially because it affects cloud services and is considered as a protective measure (and ultimately, in the case of commercial data services like agricultural data services, as a form of vendor lock-in). In some trade agreements, this is considered in contrast with the fair competition: in the European Union, the recent “Regulation on a framework for the free flow of non-personal data in the EU” explicitly prohibits national governments from creating unjustified data localization requirements.
As it appears from this brief overview, public policies don’t address most of the issues described in the previous lesson. See the section on codes of conduct below for a self-regulatory industry-led approach.
2.1.2. Open data policies
As mentioned in the previous lesson, in many countries (41 in 2016 according to the Open Data Barometer [2]) there are policies that prescribe that public-sector data should be open and reusable. While many of these policies are similar in approach and objectives, priorities and actual implementation are very different in each country.
It is important that service providers and farmers/Farmers’ Organizations (FOs) are aware of the open data policies (if any) and data publication status in their country in order to assess the availability of free open data and therefore be able to determine the feasibility of services (which may need additional paid data). It is also necessary to understand the licensing clauses and the interoperability standards used by the open data platform to determine the legal and technical feasibility of services.
Besides the obvious recommendation of consulting national laws, there are some international initiatives that try to keep track of the implementation of open data policies worldwide:
The Open Government Partnership (OGP) is a formal partnership with specific eligibility criteria, among which “an access to information law that guarantees the public’s right to information and access to government data is essential to the spirit and practice of open government.” If a country is a member of the OGP (79 countries are members so far), this gives good leverage to agricultural stakeholders and farmers organizations in particular to advocate for the publication of data useful for agriculture.
The Open Data Charter is an initiative that invites national and local governments to adopt a common charter of 6 principles. Public data has to be open by default; timely and comprehensive; accessible and usable; comparable and interoperable; for improved governance and citizen engagement; for inclusive development and innovation. The Charter has been adopted by 22 countries. Again, if a country has adopted the charter, agricultural stakeholders can leverage this to advocate for the publication of data useful for agriculture, especially pushing for the principle “for inclusive development and innovation”.
You can look up the ranking of your country in one of the indexes created by different international initiatives that assess the level of openness of public data, like the Open Data Watch, the Open Data Barometer (not actively maintained), the OECD Index of Open-Useful-Reusable Government Data (OURIndex) for OECD countries, the European Data Portal “European Open Data Maturity report” for EU countries, or, for the more general dimension of right to information, the Global Right to Information Rating (RTI).
In some cases, open data sources don’t have to necessarily be from the farmer’s country: there may be open data from other countries that either are not geospatially sensitive (pest treatment, some general aspects of crop growth) or cover different countries/regions or have global coverage (e.g. many services rely on weather data from the NASA).
Open data policies are very important to enable equitable data sharing, however:
Policies rarely address the two issues reported in the previous lesson (usefulness and usability of the data), especially the prioritization of high-impact datasets and the challenge of providing real-time dynamic data.
Normally open data policies do not address issues of data asymmetries and how public data can counterbalance data concentrations and contribute to levelling the playing field for new actors.
These issues have been very recently addressed by European Union policy makers in policy briefs (see the 2018 EC Communication “Towards a common European data space”) and partly in the new Public Sector Information Directive (which focuses on reusability and impact of data and encourages the identification and prioritization of high-value datasets and the publication of real-time data). In general, it is not easy to find advanced open data policies that foresee public real-time dynamic data and prioritization processes for high-impact datasets tailored to the needs of farmers. However, it appears that things are moving:
In some countries, prioritization “models” or processes are becoming part of open data strategies and include demand from / consultation with the stakeholders (two examples: public engagement and prioritization methodologies in the US open data project and the prioritization model in the Open Data Strategy of Macedonia).
Regarding real-time data, some developed countries have started recommending the publication of real-time data, often limitedly to transport data; among less developed countries, there are for instance new draft open data policies in Tunisia, Ghana or Ethiopia planning the design and implementation of a data inventory that includes the periodicity criterion, and making data update mandatory based on the periodicity.
While a good part of the data that governments already opened or might be expected to be asked to open for the benefit of farmers is quite static or changes over longer periods (soil maps, cadaster data), there are additional types of data that governments may be asked to collect and open that are very sensitive to timeliness and periodic if not real-time update: granular weather forecasts, market data (and price information at all stages), pest early warnings. These types of data are nowadays more often covered by private sector services. Governments could either start collecting this data as a public service or could explore ways to induce private companies to share it. See the next section on this.
Private sector data sharing
As mentioned in the previous lesson, a big amount of high public interest is held by the private sector. In particular, data that can be of high value to farmers is collected or aggregated by private companies (e.g. reliable weather data, market data, precision agriculture aggregated data on soil, water, use of fertilizers and pesticides...).
Many governments are trying to negotiate the publication of private-sector data of public interest and to explore grounds on which the private sector might be willing to share data, both with other businesses and with the government, both for boosting innovation and for public interest. The difficulty is to strike a balance between the privacy/business value of this data and its public interest or social responsibility value.
The paths that have been proposed so far are:
Just claiming public interest based on specific criteria. Some examples: (a) the European Statistical System suggests providing a clear legal framework recognizing “a general principle of access to privately held data of public interest”[3]; (b) claiming public interest based on the level of public or collective contribution to the value of certain private data assets[4]; (c) the enforcing in public contracts the open access publication of all data generated with public money (following the now broadly adopted approach of open science, enforcing the publication of all publicly funded research data as open data); (d) the enforcement of open data publication of data generated by companies that provide public services (as in the French "données d'intérêt général"[5] (data of public interest) policy).
Leveraging the sense of social responsibility of companies, e.g. creating social certification schemes or leveraging “data philanthropy”. Initiatives such as DataKind and the Global Partnership for Sustainable Development Data champion the use of private sector data for social and humanitarian purposes, while Data Collaboratives proposes “a new form of collaboration, beyond the public-private partnership model, in which participants from different sectors exchange their data to create public value”.
Identifying pre-competitive spaces for sharing private sector data, for instance companies sharing early-stage data, without much added value, to be combined with other datasets for new insights; or sharing data for improved value chain efficiencies.
Public/private partnerships where benefits for the private partners are identified and compensation for data may be considered.
General “crowd-sourcing” of public data from citizens and companies, like in France, where the government opens its national open data portal to anyone to publish open data sets[6].
The relevance of this trend towards private-sector data sharing for agriculture, and for farmers in particular, is still very low because so far there are examples in other sectors (transport, telecommunications...), but not many in agriculture. Some private-sector data shared publicly is already useful for agriculture (weather data, transport data) and a few agricultural input and technology companies have started publishing some datasets (e.g. Syngenta with the Good Growth program, which publishes datasets of productivity and soil data from the field). The potential would be very high if more companies started sharing data on precision agriculture aggregated data on soil, water, use of fertilizers and pesticides, plant and animal health etc.
2.2. Agricultural policies and other policy spaces
2.2.1. Agricultural policies
Agricultural policies do not have the same coverage in all countries. Some aspects that are often covered and are relevant for farm data are:
Subsidies and the related data conferment for subsidy eligibility and compliance. In some countries, governments have started or are considering accepting data from new data-driven technologies for subsidy compliance. Two examples are: the EU, now allowing the use of Sentinel data for compliance evidence for payments under the Common Agricultural Policy (CAP) scheme; and India, where the government and an input supplier are implementing a Proof-of-Concept application using blockchain technology for fertilizer subsidy management. An example of an interactive platform where farmers confer this data directly is the SNS system in Rwanda, where farmers provide data to the government to get subsidized inputs. Issues have been raised regarding privacy for this type of data conferment also: for instance, three German farmers won a case against the EU regarding the publication of data of farms that were not firms[7], and the EU Common Agricultural Policy underwent an assessment of the European Data Protection Supervisor, with a positive outcome.
Agricultural advisory services, which can counterbalance the dependence of farmers from private services and can include the provision of public data services.
Agricultural information systems, like market observatories, farm registries, plant variety databases, animal monitoring/tracking systems.
(In some cases) rebalancing agri-food value chains, strengthening the position of farmers and strengthening cooperation among farmers: this could be a space for unfair data practices in the value chain and for supporting data cooperatives.
It is evident from this overview that some data aspects of agricultural policies overlap with aspects covered by digital policies. It is not straightforward to say how many of the farm data issues are / should be addressed in agricultural policies and how many in digital strategies: it is interesting that the FAO e-agriculture project, which supports countries in developing policies on ICTs in the rural domain, invites to promote national e-agriculture strategies “as part of national ICT and/or agriculture strategies” and to “map the relevant existing policy environment that can be sometimes fragmented, within the agriculture and information sectors”.
Agricultural public data services, market observatories, farm registries and crop/animal monitoring/tracking systems may be part of general digital strategies, but issues that are very specific to agriculture and to the value chain dynamics may require a dedicated policy space.
In theory, the agricultural policy could be a space where to address issues of farm data and data asymmetries in agricultural value chains, extending and specifying more general digital policy provisions on data rights (as, in a different way, bottom-up way, do the codes of conduct illustrated in the next chapter) or open data.
2.2.2. IPR and copyright law
IPR is an overarching term for a wide variety of different legal instruments. IPRs protect the results of intellectual efforts or, if you wish products of the human mind. It is a broad concept, as indicated in the following diagram.
Table 2 The table below summarizes a number of areas that are relevant to data[8]
Type of law
What does it protect?
Differences between legislations
Applicable to data?
Patent law
Inventions
Most legislations protect inventions
No, but data may underlie patent applications
Copyright law
Creative, intellectual, artistic works
Generally, legislations protect copyrights
Yes
Database law
Effort to compile data collections
EU legislations and Mexico; in some countries (e.g. India, South Africa) seen as part of copyright
Yes
Trademarks and ‘trade dress’
Signs, names and expressions that identify marketable products or services
Generally, legislations protect trademarks
No, but there are concerns that such rights may be infringed, when reusing data from the private sector
Breeders’ rights
Plant cultivars and animal breeds
In most legislations, breeders’ rights are protected, but the way cultivars or breeds are registered varies
No, but data may underlie registrations
IPRs may vary in different national legislations, but there are international treaties with which signatory countries’ legislation has to comply and which they must enforce. Examples for such treaties are the Berne Convention and the World Trade Organization’s ‘TRIPS’ agreement for patents.
Copyright and database rights are the most relevant property rights in relation to data: they apply mainly when there is either a clear creative effort in the creation of an artifact (copyright) or a clear compilation effort (database copyright), so in the context of agricultural data they can apply (if producers decide to apply them) to compiled datasets for which an intellectual and unique effort in the design or collection of data can be demonstrated. They don’t apply for instance to raw data.
2.2.3. Competition law
Issues of monopoly and concentration of power in the same sector fall under legislation on fair competition and trade laws. Data concentration is not an infringement of competition rules, only its abuse is, for instance using a dominant position for price discrimination, lock-in, denial of service etc. It is important to be aware of the competition law and in particular legislation on unfair trade practices that are applicable in the farmer’s jurisdiction.
3. Focus on self-regulatory instruments: codes of conduct
As seen in the previous lesson, legislation doesn’t address or solve many of the challenges described in the previous lesson, in particular rights on non-personal data in data value chains.
While laws and regulations that govern personal data are becoming more and more common, legislation still doesn’t cover data flows in many industries where different actors in the value chain need to share data and at the same time protect all involved from the risks of data sharing. Data in these value chains is currently governed through private data contracts or licensing agreements, which are normally very complex and on which data producers have very little negotiating power.
Examples of the current variety of common contractual practices on farm data are:
(a) Data ownership: there may be no clauses on data ownership, or clauses stating that IoT generated data belongs to the IoT producer, or, in some cases, clauses stating that raw IoT data generated on the farm belongs to the farmer, while processed and aggregated data belongs to the technology provider.
(b) Data reuse: in most cases, either uses of farm data are not clarified and data is subject to unlimited reuse, or uses of farm data are clarified, but not negotiable; in some cases, need for consent from the farmer is required for reuse.
In this legislative void, self-regulatory instruments have started to emerge to set common standards for data sharing contracts. These instruments have taken slightly different shapes and names (codes of conduct, voluntary guidelines, principles): we will hereafter just call them “codes” for ease of reference. Codes provide principles that the signatories/subscribers/members agree to apply in their contracts.
In the agricultural sector, there are three codes that have been published recently and are known in the community of experts worldwide; they are, in chronological order:
American Farm Bureau Federation’s Privacy and Security Principles for Farm Data (2014). A set of principles [7] around consent and disclosure in farm data sharing, providing companies that collect and analyze farm data (Agricultural Technology Providers, ATPs) with a few generic principles that should be applied in contracts.
New Zealand Farm Data Code of Practice (2014). A set of guidelines [6] for data sharing in the New Zealand agriculture industry.
EU Code of Conduct on Agricultural Data Sharing by Contractual Agreement (2018). The EU Code [5] focuses on contractual agreements and provides guidance on the use of agricultural data, particularly data rights, access rights and re-use rights. Its aim is to create trust between the partners.
The three codes have some common aspects: they all have a self-regulatory and voluntary nature; they’re principle-based (they focus on the outcome of ag-data practices rather than the exact process or actions by which this is to be achieved [8]); they have been prepared by a combination of stakeholders (different combinations of farmers’ associations, Agricultural Technology Providers (ATPs), machinery suppliers and other input suppliers); they revolve around three core common points: consent, disclosure and transparency.
Table 3 Summary of the key points of the three codes (in bold, the points that are specific to one code)
US
NZ
EU
Farmers are the owners of farm data and continue to be the owners of non-aggregated farm data down the line
Responsibility of service providers to inform farmers that their data are being collected, and how they are used
Collection and reuse require consent from farmer; do nothing without the consent of the farmer
Right to retrieve own data for storage or use in other systems
Make disclosures to primary producers and other end users about the rights that the parties have
Disclose practices and policies around: data rights, data processing and sharing, data storage and security
Implement practices to ensure data is managed according to agreed terms and for agreed purposes, and accessible under appropriate terms and conditions
The data originator continues to be the owner of the data down the line and can determine who can access data and use it
Originator’s right to know the purpose of data collection and sharing
Collection and reuse require consent and reuse is subject to purpose limitation
Right to retrieve their data down the line
Originators’ right to benefit from their data (even financially)
Aggregated data belongs to the aggregator
It is interesting to notice that most of the rights attributed to the farmer in these codes are an extension of the rights attributed to the data subject by personal data protection laws. The most important points of these codes, which address some of the issues identified in the previous lesson, are:
Data ownership assertions. The EU and US codes consider the farmers as the “owners” of information generated on their farms and as such entitled to decide on data use and sharing with the other stakeholders; they recognize the "data generating" role also of the precision agriculture system, but still considering the farmer as the owner. A particularly interesting concept in the EU code is that of the “data originator”: “The data originator of all the data generated during the operation is the one who has created/collected this data either by technical means (e.g. agricultural machinery, electronic data processing programs), by themselves or who has commissioned data providers for this purpose”: this definition avoids the complications of the ownership concept and bypasses also the issues related to the figure of the farmer as either the cultivator or the land owner: the originator is the person who collected the data or commissioned the data collection. The NZ code doesn't assert any ownership rights (if anything, given the fact that it is agribusinesses that have to disclose which rights are asserted on the data, they might assert their own ownership rights).
Rights to access and control. For the EU and US codes, collection, access and use of farm data should be allowed only with the explicit consent of the farmer and the farmer maintains control of the data down the line, while the NZ code leaves it to the agribusiness to decide and disclose to primary producers what rights the organization asserts in relation to the data and what rights the primary producer has in relation to the data. In all three codes, control down the line also means that no reuse of the data is allowed for different purposes than those that had been originally agreed (purpose limitation).
Transparency and choice. All three codes require that farmers be informed that their data is being collected, for what purposes and how it will be used, and that they be allowed to opt out of the agreement and halt the collection.
Disclosure. All three codes prevent agribusinesses from disclosing non-aggregated farm data to third parties without the farmer's consent and without the same bounding legal conditions as the agribusiness has with the farmer.
Retrieval/portability. All three codes require that farmers be able to retrieve their data for storage or use in other systems. As for standards and interoperability, the EU and the NZ codes mention that data should be made available in a structured, frequently used and machine-readable format.
3.1.1. Certification
The US and the NZ codes foresee some form of certification:
The US code is associated with the Ag-Data Transparency Evaluator, a process to certify those Ag Tech providers whose contracts comply with the code and award them with the Ag Data Transparent Seal of Approval.
The NZ code provides a compliance checklist, which is then evaluated by a review panel: compliance is awarded by an annual licence and certificate as well as the NZ Farm Farm Data Code trademark to use.
A data certification scheme can enhance trust because producers are assured that an independent and objective party has evaluated the provider’s practices and deemed them worthy of certification.
3.1.2 The role of farmers’ organizations
The organizations producing a code should carefully consider the balance of the perspectives represented. In particular, farmers’ associations should negotiate for the most vulnerable actors, those who risk the most from data sharing and might therefore be most reluctant to share. Endorsement/co-creation of codes by farmer-led associations can ensure that the farmers’ perspective becomes central.
The existing codes, although co-written by farmers’ associations, have the declared objective of gaining producers’ trust for agribusinesses, so they seem to reflect the perspective of agribusinesses (basically, the impression is that codes include what agribusinesses are ready to accept)
Regarding the target audience of these codes, it is important to note that the existing farm data codes do not have farmers or Farmers’ Organizations as primary target audience – not to mention smallholder farmers – but rather the agribusinesses and agtech companies that work with farmers and use their data. So, while being prepared by bodies that represent also farmers (so far, big farmers’ associations of developed countries) and indirectly raising farmers’ awareness of their data rights, they are not written primarily for farmers. This is an important point for farmers’ organizations: they have an important role in making farmers aware of the codes and for instance assessing contracts against the codes for their farmers.
3.1.3. Advantages of codes of conduct
Codes are not mature enough and their adoption is not broad enough to evaluate their success so far. One study on agricultural-data codes of practice [8] identifies some key positive aspects of codes:
(1) They build trust.
(2) They fill normative gaps.
(3) They simplify the assessment of behaviours (like other forms of accreditation when companies want to demonstrate compliance with social responsibility requirements; this is true especially if they’re accompanied by some form of certification).
(4) They build awareness (among technology providers as well as farmers).
(5) They foster participation and inclusiveness (codes of conduct are normally co-developed by different organizations representing the concerned stakeholders; this in turn fosters trust and increases credibility).
3.1.4 International guidelines
Existing codes of conduct have regional or national coverage, which makes sense considering that they concern contractual practices and are quite sensitive to local contract laws. However, we already saw that they share many common points, which indicates that there may be a need for some general guidelines worldwide.
Considering the cross-border nature of agri-food systems, there have been repeated suggestions from policy studies (from the Directorate-General for Parliamentary Research Services[33], from [1], and very recently the plea from the GFFA Communiqué[34]) to coordinate guidelines at the international level, perhaps under the umbrella of the UN and more precisely, as suggested by GFFA, the Food and Agriculture Organization of the United Nations.
Such coordination could lead to international voluntary guidelines, or a set of standards, or an international Agreement or Treaty (on the model of the International Treaty on Plant Genetic Resources for Food and Agriculture).
4. Focus on governance options for a “data ecosystem” for farm data
Agri-food data ecosystems are a combination of governance (from policies to laws, codes of conduct, community norms...), institutions and infrastructures dedicated to the management and flows of agri-food data, as well as the actors providing and using the data.
We will focus here on governance options for the data infrastructure, more precisely the data platforms where farm data is stored. The data platforms in the ecosystem can be managed by different actors and with different purposes: in most cases, farm data is still managed on technology providers’ platforms, but some new platforms have recently been launched for farm data to be shared independently. Many experts agree that the use of independent platforms should be encouraged. "Farmers, consultants, advisers, and related companies need a data infrastructure that can collect, store, visualise, exchange, analyse and use large amounts of data, and they require a legal framework to deal with the ownership and the use of data outside of the farm premises" [9].
The governance of such platforms is key to make them really "trusted" platforms. There have been suggestions in this regard in different policy recommendations, foreseeing either public governance or stakeholder governance:
4.1 Public-sector-led data platforms
Regarding the role that the public sector could have in the provision of trust-enabling platforms, such as blockchain-based platforms and e-infrastructures for data collaboration among farmers and other actors, there doesn’t seem to be any explicit policy, There have been suggestions of independent, farmer-centric data repositories under public governance, which could be either general or organized by scope (commodity-specific, value-chain segment-specific...). We are not aware of any example of public-sector-led collaborative or interactive farm data platforms, although some of the stakeholder-led platforms listed in the next chapter are supported, endorsed (like JoinData in the Netherlands) or partly funded (like AgBox in Canada or the Fiji Crop and Livestock Council) by governments.
Similar approaches are: (a) government-led platforms for interactive conferment of subsidy-compliance data: see the chapter above on agricultural policies and the example of the SNS system in Rwanda; (b) although not necessarily collaborative or interactive, databases of farmers’ profiles maintained by governments, which often include a lot of farm data (see the lesson on farmer profiling): as an example, Rwanda again is planning to put in place a national farmer digital profile platform; and (c) government-led market/price observatories, where data are contributed by producers’ associations.
4.2 Stakeholder-led data platforms
Trust in non-public data platforms can be built if platforms are governed by a trusted organization of network members. Examples can be data platforms governed by farmers' aggregations or consortia including other value-chain actors as also in any form of "data cooperative" owned by its membership (see a few examples below). The bodies governing these platforms should be recognized as "Trust Organizations" that are entitled to verify, validate and authenticate data flowing as also assure fair, just, inclusive and equitable data and information flows in agri-food systems [1].
Governance models for these trusted platforms would be based on negotiation, transparency, innovative business models and would facilitate equitable flows of agri-food data. Public/private partnerships could also be considered for both the governance and the funding of such data platforms. There are already a few examples of stakeholder-led platforms:
In the Netherlands, the Dutch JoinData platform[35] allows agricultural actors to share data on the basis of clear agreements about access to and use of the data. JoinData is not a public initiative, it's a cooperative, but the government sees it as an example of a good type of agreement that can work for sharing private data.
In Jamaica, the Slash-Roots Foundation is currently working on a project to take the data from the Farmer’s Registry in Jamaica and turn it into a platform for transactions [10].
In the US, a few years ago the Iowa Farm Bureau had already proposed a farmer-controlled data warehouse. Recently, the Grower Information Services Cooperative (GiSC) and the Ag Data Coalition have created the AgXchange platform[36], a “grower-owned and governed data cooperative” whose vision is to provide cooperative members with an independent data platform, state-of-the-art tools for decision making and a market for farm data. GiSC also partnered with Farmobile[37], an “independent farm data company” that provides a collect–share–monetize strategy for farm data and a technology to read and harmonize all data from the farm independent of farm equipment brands.
In Canada, AgBox[38], managed by a consortium of actors and partially funded by the government, is envisioned as a farmer owned data cooperative, giving farmers a confident and secure Canadian blockchain platform for the storage of on-farm data, featuring data connections to several precision farming data platforms.
In the EU, the EU Declaration encourages the creation of "a European data space for smart agri-food applications" and mentions the revision of the PSI Directive. The 2017 study from the European Parliamentary Research Service [9] recommends an EU-wide independent, farmer-centric data repository.
Some other stakeholder-led platforms are more focused on managing databases of farmers’ profiles and not (perhaps yet) on letting them share farm data, but they’re still good examples of how producers’ associations can manage data platforms and can substitute (or act as intermediaries with) governments in managing farmers’ registries. For instance, the Fiji Crop and Livestock Council (FCLC), made up of commodity associations and supported by the government and the EU, manages the farmers’ registry for all the commodity associations. Or in Colombia, the Colombian Coffee Growers Federation channels government subsidies to farmers and maintains a geospatial database with profiles of more than 520,000 coffee growers and their farms.
Some of the existing platforms are owned by farmers. Farmers’ associations or co-operatives as trust organizations can have an essential role in shepherding farmers' data, negotiating access to other actors' data and ensuring equitable data flows.
5. Summary
In this lesson you have been introduced to policy spaces and instruments that address or can address some key challenges of data sharing in agricultural value chains, especially for small farmers.
We have seen important public policy spaces, primarily digital strategies, which are relevant under two aspects: (a) data sharing safeguards – nowadays most data protection laws, as well as laws on data localization, are formulated under digital strategies; (b) access to data – open data policies play a very important role in providing data to less resourced actors, and some recent trends in most advanced open data policies go in the direction of providing data with more impact potential: “high-value” data based on industry demand and even data from the private sector.
We have also seen that in the area of digital policies, for issues of portability and interoperability of data across systems standardization organizations and industry collaboration have a stronger role than public policies.
We have briefly indicated some aspects of agricultural policies that can be relevant for data sharing, like data conferment for subsidies compliance, agricultural advisory services and agricultural information systems like market observatories, farm registries, plant variety databases, animal monitoring/tracking systems.
However, we have noted that certain key challenges regarding data ownership, data control and bargaining power – in general, trust issues among actors in value chains – have been better addressed by stakeholder-led initiatives, like codes of conduct or data platforms with a trusted governance.
Codes of conduct prepared in a participatory way by different stakeholders including farmers’ representatives build trust through the provision of agreed guidelines on how digital agriculture contracts should address farmers’ rights on farm data. Building trust is also the objective of platforms managed either directly by farmers (or farmer-led associations) or by third parties with transparent governance.
Bibliography
Maru, A. et al. 2018. Digital and Data-Driven Agriculture: Harnessing the Power of Data for Smallholders.: Global Forum on Agricultural Research and Innovation, Rome, Italy. https://cgspace.cgiar.org/handle/10568/92477
de Beer, J. 2017. Ownership of Open Data: Governance Options for Agriculture and Nutrition.: GODAN, Wallingford, UK.
Jellema, A. E. 2015. Open data and smallholder food and nutritional security (CTA Working Paper 15/01). Wageningen: CTA. Retrieved from https://cgspace.cgiar.org/handle/10568/75490
FAO. 2018. Status of Implementation of e-Agriculture in Central and Eastern Europe and Central Asia. Insights from selected countries in Europe and Central Asia. FAO, Budapest. 52pp. Available online: http://www.fao.org/3/I8303EN/i8303en.pdf
COPA-COGECA, C. (2018). EU Code of conduct on agricultural data sharing by contractual agreement. Retrieved from https://copa-cogeca.eu/img/user/files/EU%20CODE/EU_Code_2018_web_version.pdf
Farm Data Accreditation Ltd. (2015). Farm Data Code of Practice. Version 1.1. New Zealand: Farm Data Accreditation Ltd. Retrieved from http://www.farmdatacode.org.nz/wp-content/uploads/2016/03/Farm-Data-Code-of-Practice-Version-1.1_lowres_singles.pdf
American Farm Bureau Federation. (2016). Privacy and Security Principles for Farm Data. US Farm Bureau. Retrieved from https://www.fb.org/issues/technology/data-privacy/privacy-and-security-principles-for-farm-data
Sanderson, J. E. 2018. What's behind the ag-data logo? An examination of voluntary agricultural-data codes of practice. International Journal of Rural Law and Policy (1). Available online: https://epress.lib.uts.edu.au/journals/index.php/ijrlp/article/view/6043
European Parliamentary Research Service 2017. Precision agriculture in Europe. Legal, social and ethical considerations, Available online: http://www.europarl.europa.eu/RegData/etudes/STUD/2017/603207/EPRS_STU(2017)603207_EN.pdf
Ferris, L and Rahman, Z. 2016. Responsible Data in Agriculture, [Online].https://www.godan.info/sites/default/files/documents/Godan_Responsible_Data_in_Agriculture_Publication_lowres.pdf
Dorward, A., and Chirwa, E. (2010). The Farm Input Subsidy Programme (FISP) 2009/10 : a review of its implementation and impact. Retrieved from https://eprints.soas.ac.uk/16734/1/FIPS%202009_10%20Review%20Final.pdf
Kuteya, A. N., Lukama, C., Chapoto, A., and Malata, V. (2016). Lessons Learnt from the Implementation of the E-voucher Pilot. Indaba Agricultural Policy Research Institute Policy Brief 81. Available online: http://www.renapri.org/wp-content/uploads/2017/01/IAPRI_PB81_Oct2016.pdf
Ladele, A.A and Oyelami, B. O. 2015. Incidence of sharp practices in growth enhancement support scheme redemption centres of Oyo State. Nigerian Journal of Rural Sociology, 16(1), 76–82. Available online: https://www.researchgate.net/publication/320466698_Incidence_of_sharp_practices_in_Growth_Enhancement_Scheme_redemption_centres_of_Oyo_State
Footnotes
[1] Table from background research funded by the World Bank (slightly adjusted)
[2] http://www.aef-online.org/
[3] http://aggatewayglobal.net/
[4] https://www.isobus.net/isobus/
[6] https://en.wikipedia.org/wiki/Data_localization
[7] https://eur-lex.europa.eu/eli/reg/2018/1807/oj
[8] Download the data from https://opendatabarometer.org/4thedition/data/ to check the values related to “existing OD policy”.
[9] https://www.opengovpartnership.org/process/joining-ogp/eligibility-criteria/
[10] https://opendatacharter.net/principles/
[11] https://odin.opendatawatch.com/report/rankings
[12] http://www.oecd.org/gov/digital-government/open-government-data.htm
[13]https://www.europeandataportal.eu/sites/default/files/edp_landscaping_insight_report_n4_2018.pdf
[14] https://www.rti-rating.org/
[15] https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52018DC0232&from=EN
[16] https://eur-lex.europa.eu/eli/dir/2019/1024/oj
[17] https://project-open-data.cio.gov/prioritization/
[18]http://mioa.gov.mk/sites/default/files/pbl_files/documents/strategies/open_data_strategy_en.pdf
[19] https://public.sbc4d.com/2017/Ethiopia%20Open%20Data%20SBC4D.pdf
[21] https://www.technologyreview.com/s/611489/lets-make-private-data-into-a-public-good/
[22] https://www.gouvernement.fr/action/l-ouverture-des-donnees-publiques
[23] https://www.datakind.org/
[24] http://www.data4sdgs.org/
[25] http://datacollaboratives.org/
[26] https://doc.data.gouv.fr/a-propos/que-publier-et-comment-le-publier/
[27] http://opendata.syngenta.agroknow.com/the-good-growth-plan-progress-data
[30] https://www.smartnkunganire.rw/
[31] https://www.bbc.com/news/world-europe-11724893
[32] WIPO Intellectual Property Handbook: Policy, Law and Use (2004). WIPO, Geneva, Switzerland. ISBN 92-805-1291-7 available at http://www.wipo.int/about-ip/en/iprm/
[35] https://www.join-data.nl/?lang=en#
Last updated
