With the introduction of new technologies such as fertiliser and hybrid seeds, agricultural productivity has experienced an unprecedented rise in the past decades in almost all parts of the world. Not so in Sub-Saharan Africa, however, where yields have remained stagnant and adoption of modern agricultural technologies is woefully lagging behind.
Why? Data from farm demonstrations in Sub-Saharan Africa certainly shows that yield returns to modern inputs are as high as in other parts of the world – a finding that puts paid to the explanation that soil and climatic conditions in Sub-Saharan Africa are not conducive to the use of modern technologies.
Why then do farmers not plant modern seeds and apply fertiliser on their fields? One piece of the puzzle appears to be that average yields obtained on the demonstration plots under highly controlled conditions are simply not representative of the yields and – more importantly – the returns farmers can achieve on their own plots.
Esther Duflo, Michael Kremer and Jonathan Robinson conducted an experiment in Western Kenya to measure the return to fertiliser on farmer’s own plots. They found that fertiliser had large positive returns on average, but that the official package of hybrid seeds and fertiliser recommended by the Ministry of Agriculture yielded negative returns.
The profitability of modern inputs also varies a lot across farmers. In particular, there is evidence that farmers are quite rational in their decision to adopt modern inputs or not. Tavneet Suri studies the adoption of hybrid maize in a sample of Kenyan farmers and finds that those farmers who reportedly adopt hybrid maize have positive returns to doing so, while those who do not have zero returns. There is however a group of farmers who are estimated to have high gross returns, but nevertheless do not choose to adopt.
Given these findings, we want to know how yields on demonstration plots can be high, yet low – and variable – on farmer’s plots. A potential solution to this puzzle is that inputs available in the market simply do not yield very high returns, in other words, technology available to farmers is not of high quality, and therefore yields lower returns than those seen in controlled farm experiments. Certainly, anecdotes abound that both seeds and fertilisers available in the market are of poor quality – perhaps because of poor storage and handling or outright fraud.
Yet, we simply do not know how widespread and serious a problem this is. To learn whether poor quality fertiliser and hybrid seeds can help explain why farmers choose to farm with traditional methods, we sent mystery shoppers to a random sample of local agriculture retailers in two of the main maize-growing regions of Uganda and purchased fertiliser and hybrid seeds.
Collecting evidence in the field and a large scale agricultural experiment
Measuring the nutrient content of the fertiliser in the laboratory we found that:
- 30% of the nutrient was missing;
- There was substantial variation in quality that was not linked to prices.
Bad fertilisers were thus just as costly as good ones.
We also conducted a large-scale agricultural experiment, where we planted the hybrid seeds brought from the retailers and compared yields to traditional seeds and authentic hybrid seeds purchased directly from the producer and experimented with fertilisers with different nutrient content – randomly assigning combinations of fertiliser and seed to different plots.
Given the data generated by the agricultural experiment, we could compare the yields of hybrid seeds bought from retailers, from the producer (authentic) and traditional farmer seeds. Specifically, by comparing different moments of these yield distribution, we concluded that the quality of a bag of hybrid seeds bought from a local retailer is only as good as a bag mixed 50:50 with authentic hybrid and traditional seeds.
The experimental data also allows us to estimate yields as a function of the nutrients (the percentage of nitrogen) contained in fertiliser. There are large and significant returns to fertiliser – for each additional percent nitrogen, yields increase by roughly 50kg/hectare. Relating this to the average nutrient level found in the fertiliser bought from local retailers, many farmers are losing almost one tonne of maize per hectare compared to authentic inputs, due to poor quality fertiliser – a loss that is approximately 50% of the revenue from traditional farming.
Combining the experimental yield data with information on the price of maize, inputs and labor costs, we could also calculate whether investing in fertilizer and hybrid seeds available in the market is profitable. The answer was a resounding ‘no’. Overall, only 20% of fertiliser samples applied to hybrid seeds bought locally were profitable and only 1% of samples yielded a return above 10%. In contrast, if inputs were authentic, the returns are uniformly positive and all samples would have yielded a return in excess of 30%. Hence, a rational profit-maximising farmer would simply not find these technologies profitable.
Why are the low-quality inputs on the market?
The first question should really be how such a low quality and low adoption equilibrium can persist and, second, how one can break out of it. Our data does not just give us average yields, but also allows us to estimate the whole distribution of yields on small-scale experimental plots.
- This shows that yields vary hugely – even for a given level of fertiliser and seed quality;
- In such an environment, it is difficult even for rational farmers to infer the quality of inputs from the size of their harvest and we show in ongoing work that an equilibrium in which sellers sell low quality inputs and few farmers buy can persist.
In our view, the solution to breaking such an equilibrium, might lie in integrating farmers meaningfully into markets and supply chains. If farmers sell to agricultural trading companies then the latter have incentives to supply high quality inputs as that increases their revenue. Whether such market integration can solve the problem of low productivity in agriculture is the subject of ongoing research.
Duflo, E, M Kremer, J Robinson (2008), “How High Are Rates of Return to Fertilizer? Evidence from Field Experiments in Kenya”, American Economic Review 98(2): 482-88.
Suri, T (2011), “Selection and Comparative Advantage in Technology Adoption", Econometrica, 79 (1): 159-209.
Bold, T, K Kaizzi, J Svensson and D Yanagizawa-Drott (2015), “Low Quality, Low Returns, Low Adoption: Evidence from the Market for Fertilizer and Hybrid Seed in Uganda”, Harvard Kennedy School WP.