Title: Impact and Adoption of Proprietary Seed Technologies in Developing Countries
Abstract: In the 1980s, many developing countries began to liberalize their seed markets. This has led to an increased prevalence of proprietary seed technologies, including genetically modified (GM) crops. The implications for smallholder farmers are the subject of a controversial debate. Critics argue that privatization reduces the quality and affordability of seed for smallholder farmers. Proponents claim that proprietary seed technologies can raise agricultural productivity and incomes. In this dissertation we examine the impacts of proprietary seeds on smallholder farmers, with Bacillus thuringiensis (Bt) cotton in India and hybrid maize in Tanzania serving as two relevant cases. The third focus of this study relates to the adoption of proprietary seed technologies, again taking the case of hybrid maize in Tanzania. As a result of foreign investment, an Indian seed firm transferred the Bt technology into a small number of cotton hybrids in the mid-1990s. These hybrids were approved for cultivation in 2002. Over the ensuing decade, multiple other seed firms incorporated the Bt technology into their own breeding programs. By 2011 about 90% of the Indian cotton acreage was under Bt, grown by 7 million farmers. Bt cotton is resistant to bollworm, an important insect pest. Bt can thus reduce the need for chemical insecticides, and also increase effective yield. This would reduce pesticide cost and increase revenue, which might lead to higher farm income. But if Bt seed prices are high and productivity effects low, the income effect could be small or even negative. Many impact studies of Bt have been published, but selection bias, impact dynamics and possible consequences beyond the plot level were not sufficiently taken into account. Here, we analyze panel data covering the years 2002-2008. Using panel models with farm and household fixed effects, we control for selection bias and look at the impact dynamics over time. We estimate that Bt increased yield by 24% and profit by 50% in 2002-2008, and household living standard by 18% in 2006-2008. Furthermore, there are indications that yield and profit effects of Bt increased over time. We also observe that insecticide use on Bt and conventional plots decreased over time, probably a sign of large-scale suppression of bollworm populations by Bt. In Tanzania, seed market liberalization has attracted several private seed firms since the early 1990s that have been selling hybrid maize. Hybrids can confer a yield advantage over nonhybrids, raising revenue and farm income. On the other hand, if hybrid seeds are more expensive than nonhybrid seeds, the income effect might not be positive. In addition, a nonhybrid with superior germplasm could even achieve higher yield than some hybrids. While there are studies that evaluate the impact of hybrid maize on farm productivity in Africa, they usually do not control for potentially confounding factors, such as soil quality or fertilizer use. Moreover, hardly any study on modern maize varieties in Africa has looked beyond farm productivity to analyze impacts on household living standard. Analyzing cross-sectional household survey data collected in the north and east of Tanzania, we find that hybrids raise yield by 50-60%. As a result, hybrid adopters realize higher profits. The benefits mostly occur in the north, where hybrids also increase household living standard by 15%. In the east, hybrids confer no gain in yield, profit or living standard. But since the vast majority of hybrid adopters are located in the north, most adopters benefit from hybrids. We also find that the yield impact of hybrids is not dependent on the use of fertilizer, irrigation, pesticides, or intercropping. The adoption rate of hybrid maize in Tanzania is low and one possible reason is that nonadopters are constrained, for example by lack of information about hybrids. In our sample, 31% of farmers adopt hybrids. In the north, most farmers are aware of hybrids and 49% are adopters. In the east, only 50% are aware and 12% adopters of hybrids. Using the average treatment effect framework, we find that closure of the exposure gap would raise the overall adoption rate to 45%. However, the majority of new adopters would be in the east, where hybrids have a negligible impact on yield and profit. In the north, where the benefits of hybrid adoption are more significant, most farmers are already aware of hybrids. We also observe that information about hybrids spreads through extension and farmer networks in the north but not in the east. Aside from the information factor, we find no evidence for constraints related to risk or credit. Therefore it is likely that adoption and awareness of hybrids are primarily driven by the benefits hybrids offer relative to nonhybrids. We conclude that the private sector can deliver improved seed technologies to smallholder farmers in developing countries, and that improved seed technologies, including GM crops, can benefit these farmers. Accordingly, liberalization of seed markets should be promoted in developing countries. Second, regulations that impede the availability of seed technologies must consider the potential benefits these technologies offer to farmers. Finally, policymakers and donors must weigh alleviating adoption constraints against technology improvements in their funding decisions.