In his opinion article, Drought-tolerance: a learning challenge for poor farmers, Travis Lybbert points to a growing interest in drought-tolerant (DT) crops in recent years, largely motivated by impending climate change.
In fact, research and development of DT crop varieties have a much longer history. For decades DT crops have provided the basis for viable dryland agriculture in semi-arid regions such as Australia, the Canadian Prairie, parts of Latin America, the Mediterranean and the Midwestern United States.
As early as 1932, the US maize breeder, M.T. Jenkins, described his selection of DT maize hybrids in Agronomy Journal. And the Rockefeller Foundation has been very active in supporting the development of DT rice in South East Asia since the 1990s — before the issue of climate change became such a dominant topic of concern.
Progress in DT breeding has been slow because selection criteria have focused merely on yield, and in most cases high drought tolerance is negatively associated with high yield potential. When developing a DT variety, breeders have difficulties accepting that its yield under normal conditions might be mediocre.
I have met breeders in developing countries who avoid showing the results of their field DT trials to their superiors, because the results look terrible. But if you question a poor drought-prone farmer he will indicate that he prefers a DT variety with a low but stable yield over a variety that yields extremely well in a good year but fails completely in a dry one.
Lybbert's misconceptions about breeding DT varieties and their adoption may stem from questionable assumptions he makes in his model of farmers' decisions to adopt DT crops.  These include the assumption that at extreme drought pressure DT varieties are virtually indistinguishable from non-DT varieties.
It is true that both DT and non-DT crop varieties reduce yield as drought stress increases, but the rate at which DT yields reduce is comparatively smaller. The yields of DT crops are often lower under favourable conditions but they are almost always better as conditions become very dry, and are certainly distinguishably higher than non-DT crop yields.
 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)