The government is working on a digital 'stack' of agricultural datasets, with its core as land records. Applications built over the stack will provide farmers with recommendations on which seeds to buy, and best practices to maximise their yield, along with updates on weather, agricultural credit, insurance and more. This will also help increase farmers' income and improve the efficiency of the agricultural sector, the government has said.
But such a centralised stack will use old and inaccurate land records; farmers' personal and financial details will be used without a strong data protection law; and rural areas have a low level of digital literacy. Hence, experts told IndiaSpend, such an 'AgriStack' is problematic.
The proposal comes against the backdrop of the government backtracking on three farm laws passed in September 2020, which the government had said would help in its target of doubling farmers' income by 2022.
What is AgriStack?
As a collection of digital databases, AgriStack would have some core features including a unique farmer identity number for each farmer, and some building blocks such as data on weather, the newest science and research on agriculture, agricultural commodity prices in India and abroad, information and access to central government schemes, agricultural regulations and permissions. In June 2021, the government shared a paper on the India Digital Ecosystem Architecture (IDEA), detailing the architecture for AgriStack.
There are existing technologies and applications, such as Kisan Suvidha, which provide agriculture-related information from existing schemes. But these do not integrate databases, said experts. IDEA /AgriStack is billed as the underlying platform (like Android or iOS) over which apps like Kisan Suvidha and others will run, said Nachiket Udupa, a member of the Mazdoor Kisan Shakti Sangathan (MKSS), a labour and farmers' rights organisation in Rajsamand district in Rajasthan.
IDEA is expected to ensure that state and central data will be independent parts of the architecture, but they will not exist in silos, and can be used together to give farmers the information they require.
The aim of AgriStack is to ensure that farmers take informed decisions on what crops to grow and when, which seeds to buy, and how to maximise yield. Players in the agricultural supply chain can precisely plan their production and logistics; farmers know when to sell their produce, where and at what price; and farmers can benefit from new technologies and data protection. The overall objective is that India moves to precision farming--the use of all agricultural inputs in a specific, measured quantity for high yields.
An AgriStack that can bring all stakeholders in the agri-value chain on a common platform "would be a game changer", said G.R. Chintala, chairman of the National Bank for Agriculture and Rural Development (NABARD), in a speech in September 2020.
The stack could also evaluate the creditworthiness of agricultural borrowers, said the 2019 Report of the Steering Committee on fintech-related issues. "This stack can include a farmer's borrowing history, land ownership data, cropping pattern and income data, among other information."
It puts the "farmer in the centre of the agriculture ecosystem,", the ministry said in a consultation paper on IDEA on June 1, 2021.
But not everyone agrees on the government's motives. The government has two reasons to move to IDEA, said Udupa. "There is an impression, which is not necessarily true, that combining databases will lead to better targeted delivery of government programmes and, therefore, savings for the exchequer," Udupa said. "The second reason is to open a new market for businesses. Sadly, neither of these reasons has much to do with what is best for the farmer."
Currently, the government has signed Memorandums of Understanding (MOUs) with private companies including Microsoft, ITC, Cisco, and Amazon among others, to develop Proof of Concepts (POCs) on feasible ways to develop AgriStack.
Old land records, landlessness limit AgriStack accuracy, usefulness
Land records, that are at the core of AgriStack, are either dated, incomplete, erroneous or non-existent in India.
Though you need technology like AgriStack/IDEA for identifying farmers, said G.V. Ramanjaneyulu, executive director of the Centre for Sustainable Agriculture, an agriculture research non-profit in Secunderabad, the problem is of "not having a proper record of land ownership".
In the 13 years up until 2021, as a result of the Digital India Land Records Modernisation Programme, 92% of India's villages had computerised land records. But government data show that only 68% of cadastral maps in villages (which record the area, ownership and value of land) had been digitised.
The accuracy of the maps is also suspect. The "average age of village/cadastral maps available in most of the states is more than 50 years", according to a 2009 government report, State Agrarian Relations and the Unfinished Task in Land Reforms, and "were mostly prepared during the British regime".
States have neglected updating the records through surveys, and maps have not been used to establish actual boundaries on the ground, noted a September 2017 PRS Legislative report. Property sketches thus do not always match textual records. Further, of all cadastral maps, only 53% were linked to a Record of Rights (ROR) which show land transactions, and less than half the villages issue digitally-signed RORs.
"We are not against digitisation, but against how it is being done," said Udupa. "There are landless livestock farmers, beekeepers, fishers and others [who are part of the agriculture ecosystem]… Reducing IDEA to only people who own land on record is very limiting."
Government data show that nearly 8.2% of rural households are landless, while more than half of agricultural workers are labourers.
"If land record is the basis, there is no clarity on how tenant farmers will be recognised," said R. Ramakumar, economist and professor at the Tata Institute of Social Sciences in Mumbai. Across India, government data show that tenant holdings--that is, land farmed by those who don't own it--increased from 9.9% to 17.3% in 16 years to 2018-19. In Andhra Pradesh, that figure is as high as 42%.
Low digital literacy, small farm sizes
Even if AgriStack becomes a reality, farmers might not be able to use it. In rural areas, between January and March 2021, there were just 36.2 internet subscribers for every 100 people, compared to 60.7 in urban India. Fewer than four in ten (38%) households across India are digitally literate, according to a March 2021 analysis by Venugopal Mothkoor and Fatima Mumtaz of the Niti Aayog. In rural areas, this proportion is lower, at 25%.
And then there is the problem of small landholdings: Of 93.1 million agricultural households, nearly nine in 10 own small landholdings of up to 2 hectares. And for all these reasons, experts believe AgriStack might not be useful in India.
"AgriStack tries to superimpose the nature of running large farms of the west into the small farms of India," said Ramakumar. "This may not be realistic. In India, agriculture is an unorganised and informal sector."
Data privacy and protection
Then come the issues of data privacy, in the absence of a data protection law and high levels of digital illiteracy.
IT firms see an opportunity in controlling agricultural data, particularly because food is a basic need, said Ramanjaneyulu. "There is a need for agriculture-specific data protection regulations and safeguards," said Rohin Garg, policy counsel for regulation and social welfare at the Internet Freedom Foundation (IFF), "so that farmers can control the data they share, and ensure that companies do not access core or raw data which may lead to [their] exploitation".
The report of the joint Parliamentary committee on the Personal Data Protection Bill, 2019 allows for some exceptions from data protection, including non-personal data which is defined as data without personally identifiable information. "Agriculture or farm data may be considered as non-personal data," said Garg.
The lack of data privacy and protection statutes means that private companies that have MOUs for pilots in various districts will decide which data are non-personal. For instance, Microsoft's Standard Operating Procedure, accessed through a Right To Information request by IFF, says, that its team shall ensure that data captured are "verified and reconciled 100%", but there is no clarity on how they would do so, whether by matching with official government data or in some other manner, said Garg.
IndiaSpend has written to Microsoft for their comments on agricultural data privacy and safeguards, and on verifying and correcting discrepancies in data collected during surveys. We will update the story when we receive a response.
"Identifying large MNCs [multinational companies] and asking them to harvest data could have been better served if the process was more open and included smaller firms that work around agriculture data," said Ramakumar.
Ideally, this has to be pushed by the public sector. If not, the government must ensure that the entire interface is in government control, said Garg. "Private entities will have to work in a strictly regulated environment."
IndiaSpend has written to senior officials in the agriculture ministry for their comments on data protection, and pilot projects by private companies. We will update the story when we receive a response.
Over and above all of this sits the question of federalism, given that agriculture is a state subject whereas AgriStack seeks to impose Central command and control. "AgriStack will centralise data, which otherwise would have been property of states," said Ramakumar.
There are "preliminary anecdotal reports which indicate that the Union Government is denying states access to the aggregated database, whereas states are the ones which have provided the databases in the first instance", farmers' groups highlighted in June 2021.
Further, the IDEA proposal of the government suggested that a portion of central assistance to states be linked to "progress achieved in implementing IDEA".
"Incentivisation may work because some states will do better than others, and the process can be adopted by other states," said Ramanjanyelu. But the IDEA model should allow states the flexibility to make changes based on their local context, rather than mandating that all states follow the exact same model, he added.
AgriStack tries to create a small farm into an economic unit by creating a unique ID and linking Aadhaar and transfer of farm subsidies and other benefits, said Ramakumar. "I do not think that our agriculture can be transformed by such digital short-cuts."
As many as 91 organisations working on farmers' rights and on digital rights of citizens have expressed concerns about IDEA, and asked that its implementation be stopped pending consultations with all stakeholders.
"...there is still a need to build into the design ways and means for farmers to participate in the development of AgriStack," said Udupa, adding that there should be better grievance-redressal and accountability measures to ensure that the interests of farmers are not compromised.