Drought-tolerant crops could improve food security — if researchers take downstream adoption challenges seriously, says Travis Lybbert.
Few aims have attracted as much attention and investment from private, public, academic and philanthropic sectors in recent years as drought tolerance (DT) in agriculture. In the past decade, more than US$1 billion has been spent on DT research and investment shows no signs of letting up.
With climate change, growing water insecurity and renewed concerns about food security in the wake of recent price spikes, the potential welfare gains from effective DT crops are enormous.
In rainfed regions of Australia and North America, investments in DT are expected to bring large private profits. Among the poor in developing dryland areas, gains from DT could make the difference between survival and starvation. During a drought, DT could limit catastrophic losses and help households recover more quickly.
Many proponents argue that adopting DT varieties may also allow poor farmers to become more entrepreneurial and diversify their livelihoods.
All these prospective DT benefits not only hinge on transferring lab results to farmers' fields, but also on farmers being able to see these benefits for themselves — which may be particularly tricky for smallholder farmers.
Tough to test
Public institutes and private firms release DT varieties only after they have proven themselves in experimental trials. Even in very controlled settings, breeders struggle to stress their test varieties with the right amount of drought at the right time, but the difficulties don't end there.
A breeder may be satisfied that a DT variety outperforms conventional crops, but poor farmers in difficult growing conditions will, rightly, insist on comparing varieties themselves.
Yet smallholder farmers — who typically face poor soils, erratic weather, and limited or no access to irrigation and other inputs — often lack the control over conditions required to perceive subtle differences between competing varieties.
This is precisely why private firms often can't afford to target smallholders as their clientele: this 'background noise' can make it tough for them to see the difference between a new variety and an old one.
Traits that confer truly dramatic benefits can outcompete this background noise. Bt cotton, for example, has been rapidly adopted by poor farmers in India because its benefits are almost impossible to miss (even with counterfeit Bt seeds in circulation). This is particularly true in extreme cases — indeed, the higher the bollworm pressure the more exaggerated the relative performance of Bt cotton.
In contrast, the relative benefits of DT peak in just the right drought conditions, then quickly fade with increasing drought pressure. These benefits are also much less uniform and observable as they depend on microclimates, rainfall timing, and soil topography and composition.
In a recent paper, we modelled the differences between farmers' decisions to adopt Bt and DT. Our model predicts that the diffusion of DT will be four times slower than Bt crops. 
The model also shows that vulnerable farmers — the professed target clientele of many public or public–private DT research efforts — take four times longer to reach 90 per cent diffusion than their less vulnerable peers. This is because the vulnerable farmers are highly sensitive to extreme drought. DT crops do not fare well in extreme drought: when the rains fail and households are really suffering from the broader impacts of drought, DT yields may also fail to deliver.
Furthermore, although DT research is often motivated by impending climate change, the more frequent extreme events predicted by most climate models may actually slow DT adoption.
These learning complications are surmountable, but downstream challenges must be taken seriously.
Bundling DT with other improvements that offer unconditional benefits, such as early maturation, could speed adoption. A functional agro-services sector and regulatory environment could also alleviate some of the learning problems by improving the flow of information to farmers through effective extension, variety labelling and certification.
And pricing will be key. DT diffusion is likely to be especially sluggish among vulnerable farmers if they have to pay a premium for DT seeds, highlighting the importance of royalty-free, humanitarian uses of intellectual property in existing and future public–private partnerships.
DT certainly has the potential to help poor rural households cope with and recover from drought but developing effective DT traits in laboratories and test plots is only part of the solution.
To clear the path to widespread adoption among poor farmers, we must take seriously the quandary of a smallholder farmer in drought-prone Africa trying to figure out whether his neighbour's DT maize really did better than his own.
Travis J. Lybbert is assistant professor of agricultural and resource economics at University California, Davis in the United States.
See Letter to the editor.
 Lybbert, T.J. and Bell, A. Stochastic Benefit Streams, Learning and Technology Diffusion: Why Drought Tolerance is not the new Bt. AgBioForum 13(1) (2010)