Much economic activity in low-income countries takes place in groups whose members are associated through social networks. Researchers in the Indian state of Karnataka conducted a randomized evaluation to study whether collective incentive payments to village cooperatives of dairy farmers could increase milk quality and reduce microbial contamination. Incentive payments increased cleanliness and point to the importance of leadership and social networks in cooperative production. However, some village elites declined payments to the larger community when they cannot control information disclosure.

Policy Issue: 

In cooperative agriculture, farmers organize into groups (cooperatives) to jointly purchase materials and market their products. Around the world, cooperative agriculture has been promoted to increase the benefits farmers experience from their investments and improve the livelihoods of poor rural households. In rural areas throughout low- and middle-income countries, cooperative agriculture is a popular way to gain access to markets. Smallholder producers organize into cooperatives and similar arrangements. This way, they can purchase supplies in bulk, and thus at lower prices, and sell their collective outputs to a broader market. While cooperative arrangements between farmers allow for increased market access, the pooling of produced goods can make it hard to track individual farmers’ production practices. This can make it challenging to enforce regulation such as hygiene standards. How do financial incentives applied at the agricultural cooperative-level affect the quality of products and cooperative revenue? Further, how does public disclosure of information on incentives affect its role in improving quality?

Context of the Evaluation:

Cooperative structures for milk production are present in many South Asian nations, including in the Indian state of Karnataka where this evaluation took place. The Karnataka Milk Federation (KMF) comprises over 2.4 million dairy producers in over 22,000 villages. Each farmer in the KMF owns one or two milk-producing cows and earns between 20 percent and 30 percent of their income from dairy activities. Village-level cooperative societies aggregate the dairy output from smallholder farmers in the village. In total, the KMF collects, processes, and distributes 2 million gallons of milk per day.  The pooled milk is then delivered to processing facilities for packaging and sale. The KMF distributes dairy products nationally with farmers cooperatives receiving the surplus profits. Under the cooperative structure, raw milk production is highly decentralized into village-level units, while processing, packaging, and retail are centralized through state-level facilities.

The cooperative’s milk quality has implications for farmers’ revenues. Microbial contamination of milk lowers the overall profitability, and thus dairy farmers’ incomes, by limiting the potential uses of raw milk. Due to different pasteurization methods, raw milk with a high microbial load is only suitable for sale as liquid milk. Cleaner raw milk is suitable for more profitable products such as cheese, yogurt, and milk sweets. Cleanliness at the point of collection and the time passed between collection and refrigeration are the main determinants of the milk’s microbial contamination.

Both farmers and the KMF can limit the microbial contamination of milk by sticking to basic sanitary practices such as washing hands before milking and sterilizing both individual and cooperative collection equipment.  However, the decentralized cooperative structure makes enforcing sanitary practices on an individual level hard. Milk from different farms is poured together for transit, so tracing back the milk quality to individual farmers becomes impossible.

Details of the Intervention:

From 2014 to 2015, KMF partnered with researchers to conduct a randomized evaluation of financial incentives for dairy cooperatives to reduce microbial contamination in milk.

Depending on the quality of milk, cooperatives could receive an incentive payment between 0 and 2,000 Indian rupees (about US$ 40 at the time of the evaluation). Cooperatives received milk quality testing and incentive payments once every two weeks. The average KMF dairy cooperative generates daily revenues of 5,600 rupees. An average cooperative would thus receive a 36 percent increase in daily revenue from the financial incentive if it produced the highest quality milk.

Researchers assigned each of the 51 dairy cooperatives with a total of 2,859 members participating in the study to one of three groups:

• Incentive payment, publicly announced (13 villages): A subset of cooperative members as well as the president and secretary of the cooperative learned about the existence of incentive payments for low levels of microbial contamination.
• Incentive payment, announced to leadership (19 villages): Only the local cooperative secretary and president learned about the existence of incentive payments for low levels of microbial contamination. The president and secretary could choose to share this information with other members of the cooperative at their discretion.
• No incentive payment (19 villages): Cooperatives in this group received no incentive payment and thus served as the comparison group.

Researchers collected samples of pooled milk from village cans as the basis for quality assessment. Any incentive payments were made into the cooperative bank account. The cooperative then decided how to split and distribute the incentive payment between the cooperative management and individual farmers.

Results and Policy Lessons:

Results indicate that incentive payments for village cooperatives substantially reduced the microbial contamination of milk.  Overall, the results indicate that there is sufficient information and means of enforcement within villages for the cooperatives to improve hygiene practices and take advantage of opportunities presented to the whole group. Incentives appeared to be more effective when communicated only to cooperative leadership rather than announced publicly to all cooperative members. In contexts where local elites experienced disadvantages from public information disclosure, they sometimes overturned the program’s main objective to improve overall quality. Therefore, programs aimed at improving social welfare need to consider local social structures when deciding on the level of transparency.

Cleanliness: Microbial contamination decreased among village cooperatives that received incentive payments. Average cleanliness improved by 0.64 standard deviations in villages where incentive payments were announced to cooperative leadership and by 0.32 standard deviations in villages where the incentive payments were publicly announced. Overall, the improvements in milk quality would generate an 81 percent increase in the fraction of raw milk suitable for more profitable processing such as cheese or yogurt. Quality improvements appeared larger in villages where incentive payments were officially announced to the cooperative leadership only.

Cooperative revenue: In villages that did not receive public information about incentive payments, greater sanitation among cooperatives generated roughly an additional 100 rupees (US$2 at the time of study) per milk collection day for the cooperative. Compared to the average daily revenue of a DCS cooperative, 5,600 rupees (US$112), this value amounts to a 1 percent increase in revenue over the two-week testing window.

Quantity: In the villages where incentive payments were publicly announced, dairy farmers increased their production by nearly 16 percent. While there was no added incentive payment for an increase in quantity, farmers producing higher-quality milk could also increase their revenue as milk prices were paid by quantity.

Forgone payments: In the second half of the study, seven out of 22 cooperative secretaries opted to forego incentive payments to the cooperative entirely rather than accept a publicly announced payment. Researchers denied the cooperatives’ previous requests for payment to be made to the DCS account without public knowledge. This may, in part, explain the differences in payments received by the public and the privately announced payment group. The managers most likely to opt out of payment are those lower in social status within the village social network, perhaps because they have less influence over the allocation of DCS profits.

There may have been other costs associated with the publicly announced payments that managers wanted to avoid. For example, milk farmers may misunderstand public announcements about payment and conclude they are owed more than the DCS receives as surplus. Announcements about cleanliness payments may also reveal undesirable information about DCS finances.