Lesson 2.2 Challenges for Smallholders in Data Value Chains

Learning outcomes

At the end of this lesson, learners will (be able to):

• Address and/or raise awareness of the specific data challenges that smallholder farmers face

• Assess the ethical and legal sensitivities of data-driven services where farmers’ data is shared through the value chain

• Evaluate issues of data ownership and data rights regarding farm data

• Assess the different roles of public and private data sources and the different challenges in reusing them in services for farmers

1. Introduction

This lesson aims at making both service providers and farmers’ organizations aware of the challenges and risks that smallholders face with the data flows in different value chains (for example through precision agriculture or with any actor that needs to profile them, like farmers’ associations or governments).

Understanding these challenges is essential to be able to create services and negotiate business models that meet farmers’ needs and address their concerns.

For smallholders, the two main challenges are: (a) to gain access to relevant and usable data and services; and (b) to make sure that any data they share does not actually weaken their position in the value chain (and ideally that sharing data actually benefits them). In both directions of data sharing, smallholder farmers face big data asymmetries in relation to the other actors in the value chain. This lesson will illustrate these data asymmetries and the related ethical, legal and policy issues.

The figure below, from the GFAR/GODAN/CTA paper “Digital and Data-Driven Agriculture: Harnessing the Power of Data for Smallholders”[1], illustrates the different streams of data from and to the farm (the fourth stream, completely off farm, is not relevant for our topic) and the related types of challenge farmers are facing.

2. Data asymmetries and power imbalances

The data challenges that we describe in the following chapters have probably always existed: farmers always needed to find good sources of information for decision making and they always shared information about their farm in order to get advice while trying not to lose any competitive edge. However, such challenges have been amplified in recent years by new technologies that collect and process data at drastically higher volumes and velocity, and by the difficulty of tracking data and data rights across the many flows and transformations generated by ICTs. In addition to these data flow challenges, there are also socio-economic factors that add to the complexity: the ownership and administration of such technologies and related amounts of data are of course linked to the power dynamics already present in value chains.

Such power dynamics are not the same in all regions of the world. They depend on economic and political conditions that determine the structure of the value chains. However, with the market economy and similar liberalist policies adopted widely, there tend to be similar trends worldwide. Power imbalances are especially strong in regions where the vast majority of farms are small family farms and where other actors in the value chain are stronger, for instance where actors upstream of the farm (e.g. input providers, technology providers) or downstream of the farm (processors, distributors, retailers) are more concentrated in large companies, often multinationals. Concentration means less competition and more bargaining power. Besides, there is the growing phenomenon of “vertical integration”, by which big companies integrate other steps of the supply/value chain, thus becoming in some cases both a supplier and a buyer for the farmer (if the production itself is not integrated as well, for example through contract farming). Vertical integration brings less dependence on other actors, even more bargaining power and better knowledge of the whole value chain.

The reason why digital technologies and related data services have come to play an important role in these dynamics is that nowadays they are mainly administered by large companies (concentrated or vertically integrated) that sell them to farmers with much less bargaining power. This has added data asymmetries to the already existing value-chain disparities. Besides the fact that data, software and equipment are managed by the technology provider, smallholder farmers would not anyway have the skills to operate the equipment and infrastructure, and to collect, manage and give informed access to their data.

”Data asymmetry” means a disparity in access to data and knowledge, as well as a disparity in how much data each actor shares and how dependent s/he is on other actors’ data. Farmers share a lot of valuable data with different actors in different value chains (supply chain, farmers’ organizations, finance, government data collection for subsidy schemes…) and often have little access to other valuable data and knowledge, even to the knowledge generated by their data once aggregated. At the same time, their data, aggregated and combined with other data from the value chain, gives data collectors precious knowledge and foresight to make decisions and a privileged position to tailor other actors’ needs and to influence important aspects of the value chain (e.g., the seeds market, prices in general - including price discrimination, supply chain disruptions, etc.).

In the next sections, an overview will be given of the main challenges that farmers face, focusing on a few that more strongly affect service providers when dealing with farmers’ concerns.

3. Challenges with farm data sharing

As mentioned above, farmers share a lot of valuable data with several other actors in different data value chains, e.g. with technology providers for precision agriculture decision support systems, with suppliers and distributors for data exchange in the supply chain, with farmers’ associations for the purpose of registration and service provision, with banks for financial assessment, with governments for subsidies eligibility and compliance etc. An overview of data flows is provided in the figure below, extending Figure 4 in Lesson 1.1. Dashed lines mean that the specific data flow is still not very common but is being experimented with. Note that this figure is about data shared for obtaining services, so some data flows like supply chain flows are not included.

Figure 2 Farm data shared with other actors

This data sharing is more and more often managed through digital technologies. New digital technologies – such as sensors, drones and the Internet of Things (IoT) in general or the blockchain – create an automatic flow of data from the farmer to the data collectors, over which the farmer has very little control. Besides, such flows are designed in ways that are not clearly explained in contracts, and contracts are often non-negotiable. For this reason, farmers are often wary about the use of technologies like drones or sensors whose data is under the control of the technology provider.

Table 1 gives a quick summary of challenges presented by sharing farm data, followed by a more in-depth analysis.

Table 1 Challenges of sharing farm data through the value chain[2]

Challenge	Specific relevance to the farmer
Risk of unfair data practices	● Uncertainty on the ownership of data collected through digital technologies and related rights on these data ● Lack of legal protection for sensitive non-personal data ● Issues of data privacy, security, manipulation, veracity, validation, liability ● Lack of awareness of and consent from the farmer ● Monetization (e.g. actors down the line in the value chain reusing acquired data for commercial purposes) ● Lack of clear legal framework for new ICTs (especially blockchain and IoT)	● The farmer is in most cases the primary originator and the subject of the data and therefore the most exposed to data rights uncertainty, data manipulation, veracity and liability. ● Farmer in a weak contractual position, often not aware of data reuse down the line. ● Monetization: the farmer is the actor that generates most of the data and profits the least from it, while farm data is easily monetized by other actors.
Risk of data power imbalances → Risk of widening digital and socio-economic gaps	● Unfair competition (data giving some actors more knowledge and a privileged position to sell tailored services, risk of lock-in) ● Risk of “excessive transparency” of weak actors’ data ● Unbalanced data value chains and different degrees of dependence on external data sharing ● Risk of concentration of power: ○ Cost of infrastructure (telecoms, secure protocols, “ledgers”, clouds…): risk of natural monopoly for big actors and first movers ○ Possible unfair trading practices (lock-in, price discrimination, opaque algorithms hiding biased decision making and lock-in mechanisms)	● Actors upstream and downstream of the farmer have more knowledge about the market and about farmers’ needs; they can sell farmers tailored technologies and products and potentially lock them in ● Excessive transparency is a phrase coined by[3] regarding the excessive and not always justified amount of data shared by farmers with government (but it applies to transparency towards the rest of the value chain as well) ● Big multinational consortia and vertical segments of the agricultural value chain are data self-sufficient and do not need to share ● Farmers in a weak position to negotiate ● Technologies too expensive for small farmers ● Risk of opaque (biased) algorithms removing all decisional power from farmers (devaluation and loss of farmer’s knowledge) ● Risk of infringing on farmers' and/or indigenous rights (traditional knowledge, indigenous seeds…)

3.1 Legal uncertainty and unfair data practices

3.1.1. Sensitive data

Besides personal data, which is normally protected by legislation, there are other types of data shared by farmers that are of a sensitive nature, such as confidential data concerning specific farming techniques. Please see Lesson 2.4 for more detail on the topic of personal data. Besides the fact that precision agriculture equipment can reveal all details about farming conditions and techniques and other potentially sensitive business data, there is also data collected by other actors, especially the government, such as agricultural censuses, satellite data and geospatial data in general. Combining all this data, a lot of information about a farm and its activities can be inferred.

Currently there is no legal protection for this type of sensitive non-personal data, unless they are classified as trade secrets, in which case they should not be shared. (See also section 3.2.1 below on excessive transparency of farmers’ data.)

In the next lesson on policies and governance, you’ll see some attempts at self-regulation that aim at giving the farmer more control on sensitive data, extending some of the principles of personal data protection (e.g. consent, access, purpose limitation, portability, disclosure) to farm data.

3.1.2. Ownership, access and control rights

It is common to read that farm data belongs to the farmer. Sometimes this is even stated in contracts. However, “ownership” is a legal assertion and ownership of data is not addressed by legislation (except for copyright for datasets as intellectual products). This is partly due to the peculiar nature of data compared to other goods that can be owned. In legal terms, it is non-rivalrous: the same data can be in different places and be owned by different people, because when data is copied or migrated to other platforms, it remains itself. In addition, there is a difference between data collected in a structured dataset which can be considered as an intellectual product by law and raw data as individual, unstructured bits before they’re collected and made sense of. These raw data are more assimilable to facts, for which no copyright and no ownership is legally applicable.

Thus, the concept of ownership is not strictly applicable to farm data. Besides the lack of legal applicability of the concept to raw data in general, machine-generated farm data presents additional complexities: (a) it is generated on the farm and is about the farm, but is generated by machines without the intervention of the farmer (so the farmer is not considered the generator or collector); (b) it is raw data, so not an intellectual product, but is then transmitted and processed and combined with other data in aggregated datasets which are intellectual products (and can therefore be “owned”).

It may be useful to also point out additional difficulties in using the concept of ownership:

• Depending on the type of data, some of it may fall under other established ownership-like rights (with precedence on any contractual statement of “ownership”): copyright or database rights (in the case of “datasets”), personal data, property rights, trade secrets and patents[4]. Copyright and database rights are often applied to aggregated data, which would technically belong to the company responsible for its collection and processing (see [5] and [6]).

• Focusing on the traditional ownership concept in farm-data value chains that require data sharing and data transformation and aggregation may be counterproductive and restrict the flow of data.

• It is not obvious whether the ownership of farm data should be attributed to the farmer or to the landowner, when they’re not the same person.

This does not mean that in the future there may not be new legislation addressing all the difficult aspects of data ownership (and in particular raw data and machine-generated data). But at the moment “legal and regulatory frameworks around agricultural data ownership remain piecemeal and ad hoc”[7].

The next lesson will explain that the current predominant attitude towards data ownership in agriculture follows the approach of recognizing “attribution” (the word ownership is used reluctantly, for the reasons mentioned above) of farm data to the farmer (or, as in the EU code of conduct illustrated later, the “data originator”).

In the absence of a legal framework for farm data ownership, and considering the difficulties above, many experts agree that it is much more important to clarify rights of access and control on the subsequent versions of processed data (see for instance [8]). In a way, if ownership cannot be asserted legally, it can be attributed “by proxy” by defining other key aspects of ownership, like control and access.

Right to access: This attributes the right to “see” and retrieve the data at any stage, even when collected and processed by other actors. For instance, with IoT equipment collecting data, manufacturers may or may not grant data access and reuse rights to the user of the object: a farmer using a drone to get images and data from his field may only have the right to use the drone software and not the right to actually “see” or store the underlying data. Quite often, once aggregated and in further stages of reuse, data is no longer retrievable by the farmer (as noted in [9]).

Right to control: This attributes the right to decide on the sharing and further reuse of the data. Regarding further reuses of data, under personal data protection laws, it is very common to apply the principle of purpose limitation (no reuse for purposes other than those to which the right-holder has originally consented) and this principle is sometimes recognized also for non-personal data.

However, its implementation seems difficult in the management of digital agriculture data, where data needs to be transformed and combined with other data in order to be useful for decision making. For instance, if a farmer gives consent to a company to use farm data on soil, crop growth and pests in aggregation with data coming from other farmers in order for the company to provide back production advice, if at a certain point the company wants to share this data with an input supplier to get recommendations on fertilizers and pesticides, the company should again ask the farmer for consent for the new data sharing purpose.

Since there is no legal framework for these rights either, except when they’re about personal data, in practice the definition of rights is currently left to contractual agreements. Contracts quite often recognize data “ownership” to the farmer, but without specifying specific access and control rights: since legislation doesn’t define data ownership rights, the recognition of ownership in contracts without access and control rights doesn’t guarantee the enforceability of any rights.

Regarding contracts as an instrument, it has been observed that they’re not the ideal solution: on the one hand, contract law is not harmonized internationally; on the other hand, contractual negotiations may not be a suitable solution for stakeholders that have very little bargaining power[10]. However, contractual agreements may work if farmers negotiate collectively and/or if contracts are based on a strongly endorsed code of conduct.

3.1.3. Data portability and interoperability

An important aspect of the right to access and control the data is the right to exchange the same data again with other actors. For example, a farmer using precision agriculture equipment that collects data on soil properties, irrigation, weather and crop health, may want to share this data with an insurance company for negotiating better premiums or with a bank for demonstrating the viability of his business. Precision agriculture systems do not always allow the information to be repackaged for further sharing.

A more specific and technical aspect of the right to reuse and re-exchange the data is the technical implementation of data portability (the ability to port the data from one system/provider to another). Without this feature, there is little freedom to switch providers, which again weakens the farmer’s bargaining power and is a limitation of fair competition in general. While again this right is normally recognized for personal data, it isn’t always recognized for non-personal data.

The issue of data portability is linked also to the issue of interoperability between farm instruments and tractors and the data they generate (often only compatible with other machinery of the same brand). Beyond the right to portability itself, the lack of interoperability between pieces of agricultural equipment and software systems across the value chain also contributes to locking farmers into one technological solution. While interoperability standards exist (see the next lesson), they are not legally enforced.

3.1.4. Liability, veracity

With large amounts of data collected and transmitted by machines and used to make decisions throughout the value chain, one single error, transmitting incorrect or intentionally manipulated data, can have a potentially disastrous domino effect (as noted in [11]). It is true that such risks can actually be mitigated by data-driven technologies themselves: for instance, data collected by sensors and drones is more accurate and reliable data than data collected manually, and blockchain technology can ensure that data are not manipulated in subsequent transactions. However, the legal value of data collected by IoT equipment is not universally accepted and legislation is still not clear about liability in the case of damages caused by incorrect IoT behaviour.

One aspect that farmers may want to consider is the legal framework (still not stable) and more importantly the contractual clauses on their potential liability for data generated by digital technologies on their farm or for their data in aggregated datasets.

3.1.5. Monetization

Data monetization is the profits that farm data can generate for data processors once aggregated, if they are sold or reused in paid services. An example would be a precision agriculture provider selling aggregated data on farm soil, crop growth and crop pests to agricultural input suppliers, who would be able to sell tailored fertilizers and pesticides.

Monetization is not forbidden. However, it is useful to consider a couple of aspects that may be covered in contracts:

• If the contract includes a purpose limitation clause (see above: data cannot be used for purposes different from the ones initially agreed upon) and monetization was not agreed upon, in theory data cannot be monetized without asking again for consent from the farmer.

• It has been often claimed that any financial benefit generated thanks to data contributed by the farmer should be shared with the farmer, or that farmers should in some way benefit from it as well. However, this hasn’t been successfully tried so far (although there are a few initial examples in the US and in Canada of platforms for farmers to share and sell their data, see next lesson). The main difficulties are (along the lines of the Open Data Institute’s reasoning around personal data, see [12]): (a) while the total value of all farm data from all farmers is high, the value of data from the individual farmer would probably be very small; (b) mainly poor farmers would resort to selling data, while richer farmers would maintain the privilege of data control.

Given these two challenges, the idea of collective platforms for farm data sharing and selling could work, if they reach a critical mass (so that aggregated data can become interesting for buyers) and are governed transparently, but there doesn’t seem to be a market yet (see [13]).

3.2 Data power imbalances

3.2.1. Excessive transparency of farmers’ data

Whether it is for receiving advice and services, or for the sake of certification, or for ensuring traceability, or for demonstrating compliance for subsidies, farmers have to share a lot of potentially business-sensitive data. A position paper of the German Agricultural Society mentions the impression of farmers’ “excessive transparency vis-à-vis the public authorities” (but the same could be said of transparency towards other actors). As the same position paper notes, it is as though farmers were not considered “on an equal footing with other economic operators whose operational and business data are recognized as worthy of protection”[14].

As we mentioned before, some of these business data might even be considered trade secrets. As noted in another study[15], “details on soil fertility and crop yield have historically been considered akin to a trade secret for farmers, and suddenly this information is being gathered under the guise of technology and miracle yield improvements”. This statement may be considered too strong, as farmers have always shared data in some way, especially if this could benefit them, for instance to get back advice. With precision agriculture, data is shared for the same reason: to gain something back. Perhaps what has changed is that rather than being shared selectively and on a case by case basis, data is "taken" by technology providers automatically directly from machinery at a rate and at a level of precision that practically reveals to the technology provider everything about the farmer's practices, even more than the farmer himself knows.

In any case, the fact itself that certain data are necessary for the functioning of precision agriculture or for obtaining subsidies or for certification makes it impossible to recognize them as trade secrets: data underlying a trade secret should remain secret and should never become readily accessible to third parties (as noted in [16]).

3.2.2. Risk of natural monopoly

The sector of agricultural technology providers lends itself to the risk of natural monopoly. It isn’t only for the cost of the infrastructure needed by certain digital technologies (not all of them: some digital technologies don’t have high costs and can actually facilitate the entry of new actors in the market), but especially for the advantage of having accumulated huge amounts of data before other competitors: network effects and switching costs can create barriers to scale for new entrants[17]. Besides, the fact that in many developed countries agriculture technology providers are becoming bigger and fewer increases the risk of monopolies (more precisely, oligopolies).

3.2.3. Imbalance of contractual power and risk of unfair trade/ data practices

The risk of oligopolies also affects the farmer’s freedom of choice and his contractual power. The farmer may become dependent on the provider or be subject to unfair practices because of his lack of choice. Companies with more data and more insights than all the others can enact “anti-competitive practices including price discrimination and speculations in commodity markets that may affect food security”[18].

3.2.4. Risk of lock-in and biased algorithms

A risk that many farmers perceive is the opaqueness and potential bias of the algorithms used to process their data and provide advice. Since the algorithms by which decision support tools produce advice are almost always closed, the farmer can easily feel that he cannot exercise control over the decision-making process (as noted in footnote 18). Potential consequences of this that have to be considered are: (a) the devaluation of farmers’ knowledge and the weakening of his decisional power; and (b) the potential bias of algorithms, which can lock farmers into the solutions chosen by the service provider. It has also been noted (see footnote 18) that impartial advisory services provided by governments or farmers’ organizations could counterbalance the domination of private-sector agronomic advice, but thus far they cannot offer the same level of tailored advice and are not well organized.

4. Challenges with accessing and reusing necessary data

This section of the lesson illustrates the main challenges in accessing and reusing necessary data. Related lessons are: Lesson 1.1, providing examples of data that are useful for farmers and the potential of sharing this data; and Unit 3, which will focus on using data and will provide advice and pointers for finding, assessing and reusing data for farmers. This lesson only provides a general overview: issues of availability and accessibility of data for farmers are mainly a matter for policy makers, who are not the main intended audience for this course.

Farmers need a great amount of data (or better, since farmers don’t use data directly, they need services that in turn need to access a great amount of data) that is generated or aggregated by other actors. Examples are weather forecasts, climatic data, market prices, crop growth data, pest alerts etc. (See Lesson 1.1, in particular “Identifying key datasets in farming crop cycles”.) This data is usually owned, managed and controlled by a third party and made available, directly or through intermediaries, to farmers (and their representatives). It can be managed by governments or, more often, by private companies. (See footnote 1 for a detailed analysis of the challenges related to this type of data.) For farmers (and subsequently for those who provide services for the farmer), this type of data presents the common challenges of availability (is it available? From whom?), accessibility (is it free? Is it open?), reusability (can it be reused in other services? Is it interoperable? Are there licenses?) and quality (is it reliable? Does it fit the purpose?), many of which are introduced in Unit 3.

Since this Unit is about data in the value chains, in this section we will focus on challenges related to the relative weight of different actors in the value chain, their willingness to share, their dominating position and the role of the public sector in providing data to level the playing field.

Below is Table 2 with a quick summary of challenges related to the availability of data for farmers and its accessibility and use by the farmer.

Table 2 Challenges for farmers to access relevant data and services[19]

Challenge	Specific relevance to the farmer
Relevance of and lack of access to private sector’s data	● Lack of access to private sector’s public-interest data; the private sector often holds the highest-impact datasets ● Lack of incentives for private sector to share data publicly ● Cost of private sector’s data ● Lack of public alternatives to private sector’s data and services ● Lack of public data as a level playing field for smaller service providers	● Lack of access to private sector’s agricultural data of public interest (e.g. product tracking, sensor data, prices…) which would help foresee market crises, epidemics etc. ● Farmers dependent on private sector’s data and services, with a few big providers and no competitive market
Little use and usability of publicly available data	● Difficulties in reuse: lack of comprehensive coverage, quality, veracity, standardization; lack of applicability and fitness for use ● Lack of publicly available high-impact datasets ● Public data not responding to needs ● Lack of real-time data	● Public agriculture-related datasets not much reused by private sector ● Public data not responding to the needs of the farmer ● Lack of agricultural real-time data (real-time data very relevant for agriculture; e.g. on product safety status through the supply chain, or pest alerts)

4.1 Accessing private-sector data

The private sector nowadays holds a huge amount of data, most of which has a very high public interest. In particular, data that can be of high value to farmers is collected or aggregated by private companies (e.g. reliable weather data, timely market data, precision agriculture aggregated data – soil, water, use of fertilizers and pesticides, plant and animal health, sales data, product tracking data...). In some cases, this data (or more often, related services) can be purchased, but often at prices that are too high for smallholder farmers. In most cases, only services are sold, not the raw data or data at the level of aggregation needed by other actors to gain insights or build services. If governments had access to such data, they could offer better services to farmers.

A related issue is the little amount of business-to-business data sharing, especially between big data holders and small companies: such exchange would encourage innovation and lower the barrier for new market entrants, therefore increasing competition and offering more choice to farmers (avoiding monopoly and lock-in). At the moment, it seems incentives are lacking for the private sector to share data: some ideas that have been put forward to encourage this data sharing are pre-competitive spaces, innovative business models, leveraging social responsibility.

4.2 Accessing public data

4.2.1. Open data for farmers

In many countries (41, according to the Open Data Barometer[20]) there are policies prescribing that public sector data – or, better, data that is of the nature of a public good – should be open and reusable. The objective of these policies in most cases is that of providing free useful data for the development of innovative services. However, due to difficulties in collecting/digitizing the data and unclear criteria for the prioritization of types of data, not all sectors are equally covered by the provision of open data: even in countries where agriculture represents a big percentage of the GDP agricultural data is only now starting to be explicitly mentioned in open data policies (see for example the FAO report on e-agriculture policies in Eastern Europe and Central Asia[21]).

Types of data that are useful in agriculture and are traditionally prioritized in open data policies are: geospatial data, soil data/soil maps, cadaster data, sometimes weather data (although the private sector has a much bigger role here), more recently price data. However, even this data is not always published and not always in a useful way (see below); besides, other data that would be especially useful for farmers is normally not published: agronomic data (e.g. crop growth data, pest and disease management data), value chain data, land productivity data...) (see footnote 2).

Open data that is actually useful for the farmer would have an important role in mitigating data asymmetries, as it would provide data that the less resourced actors, like small farmers, cannot access or can only get from expensive providers, and it would also contribute to levelling the playing field for small companies that could reuse the data to offer competitive services to farmers (again making farmers less dependent on big concentrated providers).

In order to achieve this, public data would have to include data that is really needed by the industry, in this case agriculture, to support businesses and boost innovation; in other words, data that can make a big impact on the agricultural value chain (which would entail an impact assessment exercise to determine which types of data are most needed by the agricultural sector, and by farmers in particular). However, so far, the public sector hasn’t engaged much in impact assessments or consultations with society and industry to determine what data is needed and not enough "high-value" datasets (datasets that would have a high impact) are published. This is why private sector data (see above) is still essential for the development of high-quality agricultural data services.

The major challenges why public open data doesn’t make the desired impact are:

• Little usefulness/use of public data: Data publicly available may not be the most useful for farmers: e.g. governments may have price time series but not real-time price data (which clearly has more value).

• Limited usability of data: Public data doesn’t always meet the needs of the various audiences that need it: e.g. data needed by farmers (or their service providers) may be available but not at a suitable level of granularity or standardization.

An additional source of data and services that is becoming more and more relevant is farmers’ associations: they can have the same role as public data in supporting farmers, especially in combination with farmers’ profiling, which allows them to tailor services to the needs of their members. See more in Lesson 1.2 on farmer profiling.

4.2.2. Data for subsidy schemes

At the intersection between farmers’ data sharing and public open data is the bi-directional data flow around subsidized input schemes: this is data in the public sector to which farmers contribute directly. Farmers share data with the government (on their expenses, their farming practices, management of natural resources etc.) as evidence for subsidies' payments. Some of the compliance data is aggregated in statistics for monitoring the results of the subsidy scheme and becomes public data.

Traditionally, collection of this data has been done via paper forms, or more recently with electronic forms manually filled in by farmers. However, some countries are starting to allow automatic subsidy payments based on digitally collected and submitted data, for administrative simplification and lower cost of transactions. See more about this in the next lesson, section 2.2.

5. Summary

This lesson analyzed the main challenges that farmers, especially smallholders, face in accessing necessary data and information for their farming activity and in sharing their farm data for different purposes (decision support, certification, subsidies compliance…). These challenges are linked to existing power imbalances in agricultural value chains: on the one hand existing imbalances are reflected in new data asymmetries and on the other hand new data asymmetries amplify the existing imbalances. Under this respect, it was explained that public data could have a role in levelling the playing ground in the value chain and attenuate the imbalances.

Regarding data sharing, the main challenge for farmers is to share the data that they need to share for specific purposes without losing their competitiveness and without weakening their position in the value chain. In this sense, it is important that technology and service providers consider a few major issues in building their services and negotiating their contracts (and that farmers and their representatives voice the importance of these issues): the definition of data ownership/attribution rights, or more appropriately “ownership-like” rights like the rights to data access and control; the right and technical feasibility of data portability; issues of liability; risks of lock-in and other unfair practices arising from data concentrations.

Regarding access to necessary data, farmers, their representatives and the technology/service providers have to understand the challenges of accessing public and private-sector data. Currently, public data and services (as well as those provided by farmers’ associations), which could help mitigate data access asymmetries, are often not sufficiently useful or reusable and cannot compete with private-sector data and services, which on the other hand are expensive and potentially biased. Besides, attempts to encourage the private sector to share data haven’t been very successful so far.

There is clearly a need for measures (policies, platforms, guidelines) that build trust between the actors in the value chain, especially from the weaker actors, the small farmers, and mitigate the data asymmetries. The next lesson will illustrate some policy spaces and some stakeholder-led instruments that address or can help address these issues.

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. [Online]. https://hdl.handle.net/10568/92477

DLG e.V. 2018. Digital Agriculture: A DLG position paper, [Online]. https://www.dlg.org/en/agriculture/topics/a-dlg-position-paper/

de Beer, J. 2017. Ownership of Open Data: Governance Options for Agriculture and Nutrition.: GODAN, Wallingford, UK.

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

Rasmussen, N. 2016. From precision agriculture to market manipulation: a new frontier in the legal community, Minnesota Journal of Law, Science & Technology, 17(1), 9. Available online: https://scholarship.law.umn.edu/cgi/viewcontent.cgi?referer=&httpsredir=1&article=1008&context=mjlst

World Bank Group, 2019 Future of Food : Harnessing Digital Technologies to Improve Food System Outcomes, World Bank, Washington DC, USA. [Online]. http://hdl.handle.net/10986/31565

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

van der Wees, A. 2017. H2020 – CREATE-IoT Project. Deliverable 05.05. Legal IoT Framework (Initial). Available online: https://european-iot-pilots.eu/wp-content/uploads/2018/02/D05_05_WP05_H2020_CREATE-IoT_Final.pdf

Tennison, J. 2018. Getting paid for personal data won’t make things better [blog]. Open Data Institute. Available online: https://theodi.org/article/jeni-tennison-getting-paid-for-personal-data-wont-make-things-better/

Bloch, S. 2018. The New Food Economy. Available online: https://newfoodeconomy.org/farmobile-farm-data/

I. M. Carbonell, I.M. 2016. The ethics of big data in big agriculture, Internet Policy Review, 5,(1), DOI: 10.14763/2016.1.405. Available online: https://policyreview.info/articles/analysis/ethics-big-data-big-agriculture

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

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

European Parliamentary Research Service 2017. Precision agriculture and the future of farming in Europe. Scientific foresight study, Available online: https://publications.europa.eu/en/publication-detail/-/publication/40fe549e-cb49-11e7-a5d5-01aa75ed71a1/language-en

Footnotes

[1] A. et al Maru, Digital and Data-Driven Agriculture: Harnessing the Power of Data for Smallholders.: Global Forum on Agricultural Research and Innovation, 2018. [Online]. https://hdl.handle.net/10568/92477

[2] Table from background research funded by the World Bank (slightly adjusted)

[3] DLG e.V., "Digital Agriculture: A DLG position paper," 2018. [Online]. https://www.dlg.org/en/agriculture/topics/a-dlg-position-paper/

[4] J. de Beer, Ownership of Open Data: Governance Options for Agriculture and Nutrition.: GODAN, 2017.

[5] L. Ferris et al., "Responsible Data in Agriculture," 2016. [Online]. https://www.godan.info/sites/default/files/documents/Godan_Responsible_Data_in_Agriculture_Publication_lowres.pdf

[6] N. Rasmussen, "From Precision Agriculture to Market Manipulation: A New Frontier in the Legal Community," Minnesota Journal of Law, Science & Technology, vol. 17, no. 1, 2016. [Online]. https://scholarship.law.umn.edu/cgi/viewcontent.cgi?referer=&httpsredir=1&article=1008&context=mjlst

[7] World Bank Group, "Future of Food : Harnessing Digital Technologies to Improve Food System Outcomes," Washington, DC, 2019. [Online]. http://hdl.handle.net/10986/31565

[8] 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(01). Retrieved from https://epress.lib.uts.edu.au/journals/index.php/ijrlp/article/view/6043

[9] European Parliamentary Research Service, "Precision agriculture in Europe. Legal, social and ethical considerations," 2017. [Online]. http://www.europarl.europa.eu/RegData/etudes/STUD/2017/603207/EPRS_STU(2017)603207_EN.pdf

[10] J. de Beer, Ownership of Open Data: Governance Options for Agriculture and Nutrition.: GODAN, 2017

[11] Van der Wees, A. (2017). H2020 – CREATE-IoT Project. Deliverable 05.05. Legal IoT Framework (Initial). Retrieved from https://european-iot-pilots.eu/wp-content/uploads/2018/02/D05_05_WP05_H2020_CREATE-IoT_Final.pdf

[12] J. Tennison. (2018, Mar.) Open Data Institute. [Online]. https://theodi.org/article/jeni-tennison-getting-paid-for-personal-data-wont-make-things-better/

[13] S. Bloch. (2018, July) The New Food Economy. [Online]. https://newfoodeconomy.org/farmobile-farm-data/

[14] DLG e.V., "Digital Agriculture: A DLG position paper," 2018. [Online]. https://www.dlg.org/en/agriculture/topics/a-dlg-position-paper/

[15] I. M. Carbonell, "The ethics of big data in big agriculture," Internet Policy Review, vol. 5, no. 1, 2016. [Online]. https://policyreview.info/articles/analysis/ethics-big-data-big-agriculture

[16] van der Wees, A. (2017). H2020 – CREATE-IoT Project. Deliverable 05.05. Legal IoT Framework (Initial). Retrieved from https://european-iot-pilots.eu/wp-content/uploads/2018/02/D05_05_WP05_H2020_CREATE-IoT_Final.pdf

[17] World Bank Group, "Future of Food : Harnessing Digital Technologies to Improve Food System Outcomes," Washington, DC, 2019. [Online]. http://hdl.handle.net/10986/31565

[18] European Parliamentary Research Service, "Precision agriculture in Europe. Legal, social and ethical considerations," 2017. [Online]. http://www.europarl.europa.eu/RegData/etudes/STUD/2017/603207/EPRS_STU(2017)603207_EN.pdf

[19] Table from background research funded by the World Bank (slightly adjusted).

[20] The data can be downloaded from https://opendatabarometer.org/4thedition/data/ to check the values related to “existing OD policy”.

[21] 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. Budapest. Retrieved from http://www.fao.org/3/I8303EN/i8303en.pdf