Signaling, screening, or sunk costs? Experimental evidence on how prices affect agricultural technology adoption in East Africa

Smallholder farmers in sub-Saharan Africa face growing pressure to produce more on less land while contending with worsening impacts of climate change. The need for sustainable intensification has rekindled calls for a “Green Revolution” in the region, centered on the widespread adoption of modern inputs such as hybrid seeds and inorganic fertilizers. But introducing new agricultural technologies is not just a matter of making such inputs available; it also requires convincing farmers to try them.

One common strategy to promote adoption is the distribution of free (or subsidized) starter packs—for example, small quantities of seed and/or fertilizers given to farmers as testers. These trial packs are designed to reduce the risk of experimentation, kick-start investment in improved technologies, and stimulate peer learning. In theory, farmers who successfully try out a new variety may go on to purchase it again, and their neighbors might follow suit. Public agencies also view trial packs as a tool to increase turnover in seed markets and diversify the genetic base of crops.

Despite their popularity, however, the effectiveness of free trial packs is not guaranteed. Critics argue that when people receive goods for free, they may not value them. Reports abound of mosquito nets distributed for malaria prevention ending up as fishing nets, or chlorine intended for water purification used to clean toilets. If farmers do not have to pay for an agricultural input such as seeds of a new variety, they may similarly fail to plant it, use it incorrectly, or even sell it. This concern has given rise to the idea that charging a (small) price could improve outcomes by ensuring the seed are actually used and resources are not wasted, an argument that is often used to guide policy recommendations.

In a recent IFPRI Discussion Paper, we outline an experiment testing this idea. We implemented the design in Uganda and Ethiopia with nearly 1,800 smallholder farmers growing maize, teff, or wheat—finding that in some (but not all) circumstances, charging a price for test packs encouraged more effective adoption.

Potential effects of charging a price

Charging a positive price could increase the appropriate use of a product through three key mechanisms. The first and most obvious is a screening effect: Only those who value the seed will be willing to pay for it, making payment an effective targeting tool. The second is a sunk cost effect: Once money has been spent, people may feel psychologically compelled to use the product, lest the expenditure go to waste. The third is a signaling effect: In cases characterized by asymmetric information, when quality is difficult to assess—such as with seed, where benefits are realized only after harvest—a higher price might signal better quality and encourage use.

The challenge, however, is that all three mechanisms predict the same outcome—higher use at higher prices—making it difficult to disentangle them in observational data. To address this, we designed a three-stage pricing experiment that allows us to identify each effect separately.

The experiment

Our experiment was implemented for teff and wheat in Ethiopia (about 650 and 430 smallholder farmers, respectively; 16 farmers per village from 68 villages in the Amhara region), and maize in Uganda (about 760 smallholders, 10 farmers per village from 76 villages in Eastern Uganda). In the first stage, farmers were offered a seed trial pack at a randomized initial price, which served as a potential quality signal. In the second stage, they engaged in a simple bargaining game to determine their willingness to pay and identify screening effects. In the third stage, a randomly selected subset of farmers received a surprise 100% discount, meaning they paid nothing for the seed despite having agreed to purchase it, whilst others incurred a sunk cost when purchasing the seed.

Results

The results provide clear evidence of screening: Farmers with higher willingness to pay were more likely to use the seed trial pack in the immediate term and adopt the promoted variety in a subsequent season, though the timing varied across crops. For teff, the screening effect was immediate, while for maize and wheat it appeared more strongly in subsequent adoption and production outcomes. This suggests that prices indeed help screen farmers who are motivated and able to experiment with a new variety.

Sunk cost effects, by contrast, were mixed. For teff in Ethiopia, paying a positive price increased commitment to proper use. But for maize in Uganda and wheat in Ethiopia, to our surprise, we found that sunk cost effects were actually negative: Paying reduced appropriate use. It could be that liquidity constraints have limited farmers’ ability to also invest in labor or complementary inputs, and as a result farmers decided not to plant the seed (perhaps selling it to recoup some of the money). This shows that payment does not automatically induce commitment, and in some settings may even discourage experimentation.

Finally, we did not find evidence of signaling effects. Higher initial offer prices did not boost confidence in seed quality; farmers likely trusted the quality of seed coming from a research institution regardless of price.

Key takeaways

Overall, the findings suggest that modest prices can help ensure seed trial packs end up with motivated farmers, but higher prices risk backfiring, especially where liquidity is tight and complementary inputs are essential. Free or very low-cost trial packs may be more effective for encouraging experimentation in such settings. Policymakers should also invest in alternative quality signals, such as certification and branding, rather than assuming that higher prices will be interpreted as indicators of quality.

In the end, seed pricing is not just about cost recovery; pricing can be used as an economic and behavioral tool to shape how new technologies are used. Striking the right balance between accessibility and screening is crucial for leveraging trial packs to promote learning about new varieties and accelerate varietal turnover.

Bjorn Van Campenhout is a Senior Research Fellow with IFPRI’s Innovation Policy and Scaling (IPS) Unit based in Leuven, Belgium; Gashaw T. Abate is a Research Fellow with IFPRI’s Markets, Trade, and Institutions (MTI) Unit based in Washington, D.C.; Liesbeth Colen is a Professor of Nutrition and Agriculture at the University of Göttingen, Germany; Berber Kramer is an MTI Senior Research Fellow based in Nairobi, Kenya. This post is based on research that is not yet peer reviewed. Opinions are the authors’.

This work was supported by the CGIAR Research Initiatives on Market Intelligence and Seed Equal, and the CGIAR Science Program Breeding for Tomorrow.

Reference:
Van Campenhout, Bjorn; Abate, Gashaw T.; Colen, Liesbeth; and Kramer, Berber. 2025. Signaling, screening, or sunk costs? Experimental evidence on how prices affect agricultural technology adoption in East Africa. IFPRI Discussion Paper 2369. Washington, DC: International Food Policy Research Institute. https://hdl.handle.net/10568/177343