The Mustard seed market in Rajasthan, India, is primarily composed of smallholder farmers who have the option to sell to middlemen at the farm gate or take their harvest to a market. In either transaction there is an unequal power dynamic between the farmer and the buyer, which provides an opportunity for exploitation of the farmer. The situation is compounded by the fact that powerful people at the markets control the equipment used to measure mustard seed oil content (which determines the seed quality). Farmers do not know the quality of their crops before taking them to market; once they bring their crops to the market, they must either sell them or lose the investment of time and money in transportation. The market buyers can take advantage of this by offering low prices and lying to the farmers about the quality of their crops.
We propose a low-cost product that farmers can use to determine the quality of their crops before taking them to market. Two approaches are being considered: 1) when oilseeds are crushed on paper, they create an oil spot. The size of the oil spot is dependent on the amount of oil in the seeds. Using a mobile phone camera and image processing software, the size of the oil spot can be measured and tied to the oil content in the seeds. 2) The second technique relies on the fact that high-oil content seeds have a lower density. Using only a solution of salt and water, the density of the seeds can be measured, which can be tied to the oil content. Preliminary experimental results are promising for both methods, and both methods will be evaluated in the coming months.
The product will be implemented via an app, with minimal external equipment. Depending on the method chosen, the external equipment could be sheets of paper on which to crush seeds, or a see-through container with salt and water. The app will contain the software to process and analyze the data. Three monetization schemes are being considered: pay-per-use, subscription-based, or advertising. In the pay-per-use and subscription models farmers would pay a small fee to have access to the app’s features. In the advertising model, market players would pay to have their offer prices displayed in the app, for farmers to see.
The most important customers are smallholder farmers who grow mustard in the Indian state of Rajasthan. Farmers in Rajasthan make an average of 7350 Rs. per month, or about $3.75 per day. The average farm size in Rajasthan is 3.38 ha, although the women and men we spoke to in communities near Alwar farmed between 1 and 2 hectares.
The customers we are focusing on are the people in the farming household that take harvested crops to the market. Based on discussions with a women's self-help group in Rajasthan, the marketers are generally the men of the household. The reason given was that the women lack the physical strength to lift the heavy gunny sacks. With that being said, as men move to cities women are performing the majority of the farming labour in India. It is likely that women will begin to play a much larger role in marketing in the coming years, so it is important that the project considers both female and male customers.
During discussions about farming practices and rural livelihoods, we asked groups of farmers what the most pressing issues facing their villages were. Their responses varied from lack of access to good inputs to being cheated at the market, to a perception that the government neglects their district. However, a common thread of relative powerlessness links all of their problems: they have no control over when and if quality seed and fertilizer is available, they cannot send their children to good schools because the government does not provide them, and they have no bargaining power with the market players who can unilaterally decide the purchase price of crops.
Of the customer pains identified in our research and discussions with rural Rajasthani farmers, we have chosen to focus on powerlessness in bargaining with middlemen or market buyers. Approximately 46% of India's mustard seed production comes from Rajasthan, but the Rajasthani market is highly unstable and distorted by government incentives that depend on the party in power and whether or not it is an election year. The quality of mustard seed is determined by its oil content - higher oil content is more desirable - and farmers are paid based on the percentage oil. However, the equipment for measuring oil content costs somewhere in the range of $10,000 CAD for a second-hand machine, or anywhere from $20,000 to $60,000 for a new grain analysis computer. The equipment is located at the markets, where it is controlled by market buyers. Farmers complain that they are being cheated by the market buyers, who tell them their grain is of poor quality and hence offer low prices. Farmers have invested time and money into transporting their crops to the market, and generally do not have the option to choose to travel to another market in search of better prices.
We propose a simple, reliable and inexpensive technique for farmers to determine the quality of their seed before travelling to the market. This will help them decide when to travel to the market, or indeed whether or not the trip is worth it. Knowledge about the quality of their crop is a powerful bargaining tool, both at the market and with middlemen. It can allow them to question, with justification, falsified reports of mustard seed quality that they receive from the market buyers. Empirical evidence from India suggests that information can lead to higher offers from middlemen (Mitchell, 2011). Based on a 15% increase in offer price (Svensson & Yanagizawa, 2008), a 3-hectare farmer growing mustard on only 10% of their land could expect to receive an additional 2000 Rs., the equivalent of 8 days' income.
Information about crop quality can also facilitate cooperative marketing. If a group of farmers can verify that their crops are of similar quality, they can combine their harvests and share transportation costs. Rather than each farmer selling their mustard individually (which allows the market buyers to play the farmers against each other) the group can collectively bargain for better prices. A major barrier to collective bargaining identified to our team in India is lack of information about crop quality. Farmers may believe their neighbours' crops are of poor quality and so refuse to combine their harvest. By providing a cheap and easy way to verify crop quality farmers can combine their mustard seed, reduce costs, and improve bargaining power.
Customer relationships will be established through Sehgal Foundation, an NGO we have networked with that works in Rajasthan. While a project member or extension agent will be heavily involved in the initial stages of the project, all usage information will be available in the app, and the project will move to a model in which interaction is mainly through the app. This is especially important in order to ensure sustainability as the project scales. A farmer should be able to download and learn how to use the product with little outside help. Information will be provided both in text format and as videos (in Hindi and other local languages) embedded in the app. Videos will be particularly important for reaching farmers who do not know how to read.
During the development of the product a close relationship will need to be maintained between the project team and customers. Feedback will be requested from farmers and from extension agents, and used to improve the product. As the project moves towards a model with less intensive interaction between customers and extension agents, a feedback mechanism will be incorporated into the phone application. Users will be able to give both qualitative feedback about experiences with the product and report their sales amounts, locations, and prices.
We are currently in the development phase of our product. We are investigating two methods of determining mustard oil content. Both are envisioned to be implemented primarily through an app. The first is crushing mustard seeds on paper to observe the size of the oil patch produced. Images are taken before and after crushing and the area of the oil patch relative to the original volume of the seed is used to estimate the oil content. This method would require some paper and a crushing implement, external to the app. The app will take the images and report the oil content. The second method is based on the fact that seeds with higher oil content are less dense. By observing how many seeds float, and how many sink in a solution of salt water, the density can be determined, and hence the oil content can be estimated. The external materials needed for this method are salt, water, and a see-through container. Both approaches will operate primarily through the phone app, which will contain all the processing software to analyse the results and provide a report of oil content offline.
In the immediate future the key activity is calibrating the techniques against the standard technique for measuring oil content, which is the reliable but expensive Near Infrared Spectroscopy equipment that is not accessible to farmers. The better technique will be selected for development into a product. Preliminary field trials are planned for the next Rabi (winter) harvest period, starting December 2018.
In order to create a viable and sustainable solution that is recognized by all levels of the supply chain, we will need to partner with the Indian Council of Agricultural Research (ICAR) and its subsidiaries, such as the Directorate of Rapeseed-Mustard Research. ICAR's support is necessary to establishing distribution networks through the extension centres (KVKs). We have met with scientists from ICAR and have been given reason to believe they will be willing to support this venture.
While ICAR's support is needed to establish legitimacy, the scientists and staff at research centres are not necessarily the optimal people to promote the product. For this, we have discussed the idea with Mr. Pawan Kumar, the Program Leader of Agricultural Development for Sehgal Foundation. Mr. Kumar has connections in villages where Sehgal Foundation has projects, and has an established rapport with community members.
There are three major areas that need future work:
1) More experiments are needed to verify and calibrate our measurements of oil content. These experiments are underway and we expect to have results by July.
2) We have proposed three separate monetization schemes, and are not sure which would be most appropriate for the Rajasthani context. We need World Vision's expertise in developing a scheme that will be acceptable to farmers in Rajasthan and will ensure the sustainability of our project.
3) Our team has limited expertise in designing products that people with very little education can understand and use. In the development phase we plan to reach out to experts at the University of Toronto in Human Factors Engineering, as well as World Vision, who can advise on how to make the product both attractive and simple to understand.
We are proposing three possible revenue streams. Final decisions on which stream(s) to use and how much to charge will have to be made in consultation with partners in India.
1) Pay-per-use model: Farmers would pay a small fee for every sample of mustard seed analysed. This fee would be small - a few Rupees, and would have to be competitive with charges farmers are required to pay to have their seeds analysed by current methods. Based on the estimated additional profit of 2000 Rs. for a farmer selling mustard seed grown on 0.3 ha of their property, a charge of 10 Rs. per sample seems quite reasonable. Farmers may have to conduct multiple analyses to comply with regulations.
2) Subscription model: The farmer would be able to analyse as many samples as they wish, and would pay an annual subscription fee. The app would include other resources for the farmer (access to information about improved farming techniques, weather, new technologies, etc.) that would make the subscription worthwhile for the farmer year-round.
3) Market price advertisement model: Market buyers would be charged a fee to advertise their offer prices on the app. Farmers would be able to look at the app and see the prices offered at different markets, determine the quality of their seed, and make an appropriate choice. This might have an additional benefit of inducing competition among marketers, raising prices paid to farmers.
Technical development costs: $500. We require some funds to cover the purchase of good samples, and to pay for analysis with standard techniques.
Product Development costs: $ 3000. Development of a mobile phone application will cost some money. The team will create a bare-bones version to start, using online tools such as, appmakr.com. The team will also need to purchase and test various raw materials for the hardware side of the project.
Implementation costs: $10 000: In order to begin the pilot test and implementation the team will need to travel to Rajasthan and purchase enough raw materials to supply the pilot test. A revenue generation model will not be introduced until the product is proven to be effective, so there will be considerable costs in the initial phases of the project.
Ongoing costs: $200/year +$0.2/farmer: Updates will have to be made to the app, raw materials will have to be purchased and distributed.
The exact costs, and the amount of revenue, are at this point highly speculative. An appropriate amount that can be charged of farmers needs to be worked out with our Indian partners. Given the following assumptions:
Revenue/farmer/year: 100 Rs. (2 CAD)
Number of farmers: 50 (year 1), 400 (year 2), 2000 (year 3), 5000 (year 4)
Startup (capital) costs: $13500 (as above)
Ongoing costs: $200/year +$0.2/farmer/year (as above)
We will break even within 4 years and be profitable in the 5th year.
The social mission of our enterprise is to empower rural Rajasthani farmers to improve food security and provide them with the bargaining power to assert socioeconomic self-determination. We believe that information plays a crucial role in empowerment, and that empowerment is critical to food security.
Our direct competition is the system we aim to replace: costly measurement equipment set up at markets, operated by people relative power over the farmers. This system persists because there is no alternative for farmers.
Other interventions that have focused on the impact of information on negotiations have had mixed results: Svenson and Yanagizawa (2008) report that Uganda's Market Information Service increased the offers made by middlemen by 15%. However, Fafchamps and Minten (2012) report no significant impact of Reuters Market Light (an SMS-based market information service) on prices received by farmers in Maharashtra. Muto and Yamano (2009) found that mobile phone coverage in Uganda increased sale prices of bananas, but not of maize. Mitchell (2011) found that middlemen in Gujarat did offer higher prices when farmers were informed of market prices. In general, it seems that the differentiating factor is whether or not the farmer has a choice in who they sell to. If a farmer can convincingly threaten to sell to another market or market buyer, they will receive higher offers (Mitchell, 2011).
In the pilot stage the impact of the product on negotiations will be measured using a randomized control trial. Farmers will report the price they receive at the market over a harvest season, with and without the information provided by the product.
Given the transferability of the technology, a randomized control trial will probably not be possible to evaluate longer-term impacts on income and food security. We will be unable to randomize who receives the technology and who doesn’t, because farmers will be able to learn how to use it from their neighbours. Following the approach of Mendola (2007), Propensity Score Matching can be used to complete an ex-post impact assessment by comparing outcomes for farmers who are similar in most respects.
The major quantitative metric will be the price/quintal of mustard seed brought to the market.
The quantitative techniques outlined above will be complemented by semi-structured interviews and focus groups with farmers throughout the project. Qualitative techniques will be especially important for assessing impact on financial security and financial planning ability.
Strengths & Weaknesses
1) Technical background in Mechanical and Chemical Engineering, including experience in food engineering.
2) Access to University of Toronto resources for conducting experiments and developing product.
3) Understanding of the Indian context: Rahul was born and raised in Delhi, giving insights into how the Indian government and policy landscape works.
1) Lack of experience in product design. The team has little experience in taking products from conception, through design and to production scale.
2) Lack of experience in marketing. The team lacks marketing experience, and will call on World Vision's expertise in this part of the project.
We have assumed:
1) The Indian government policies will not drastically change in the near future, and farmers will continue to be put in a disadvantageous position.
2) The techniques can be developed into a measurement technique that is reliable and accurate enough to be accepted by the government, market buyers, and farmers.
3) Farmers will be able to operationalize the information they receive from the product to bargain for higher prices.
Most Significant Challenge
We anticipate that the most significant challenge will be establishing the legitimacy of the product. Market buyers and farmers have to trust and use the device in order for it to be impactful. Three areas will be important in establishing legitimacy:
1) Government (ICAR) approval. As noted by Mr. Kumar of Sehgal Foundation, the governing scientific body must give its stamp of approval to the device, or it will not be trusted.
2) Product branding. The product must look and feel like a professional measurement system.
3) Timing of release. The product should be tested in the field only when lab experiments have confirmed its accuracy. Testing before it has been developed into a reliable measurement technique will inhibit trust and lead to non-adoption of the technology.