Insight, Adaptation
European crop forecasts have been cut for every major summer crop, and a Commission-backed forum has concluded that nature-based adaptation needs to move past pilots. Demonstrations cannot tell you what will scale, because of who volunteers for them.
Two European items landed this week that belong in the same argument.
The Commission's latest MARS assessment finds that persistent heat and water deficits have substantially reduced summer crop prospects across western and central Europe, with forecasts cut for all major summer crops and some falling as much as 14 per cent below their five year averages, alongside deteriorating pasture and fodder availability.
Separately, a Commission-backed Adaptation and Nature Forum on 27 August brought together more than 450 researchers, local authorities, practitioners and policymakers around results from LIFE projects, reaching a conclusion worth quoting in substance: Europe has accumulated enough demonstrations, and the next challenge is deployment at scale, including collaboration with farmers, landowners, communities and municipalities, and demonstrating social as well as environmental benefits.
Put those together and Europe's adaptation problem has changed character. It is no longer a future risk requiring more evidence that it exists. It is an operating constraint requiring evidence about what people will actually do.
The forum's conclusion is right and slightly imprecise, and the imprecision matters. Europe does not have enough demonstrations. It has enough feasibility evidence and almost no adoption evidence, and the reason is structural: a demonstration project is close to incapable of producing the second.
There is a well developed literature on what happens to an intervention between a successful trial and a scaled programme. The finding is consistent enough to have a name, the voltage drop, and the mechanisms are specific rather than mysterious.
Participant selection. Demonstration sites are not random. They are farms whose operators volunteered, which means farms with time, curiosity, some financial cushion, an existing relationship with an advisory service, and usually a successor. Every one of those characteristics is correlated with the capacity to adopt. The demonstration therefore measures the effect of the practice on the population least representative of the one that has to adopt it next.
Implementer selection. The project team running a LIFE demonstration is motivated, expert and present. At scale the same practice is delivered by an ordinary advisory service with a caseload, or by nobody.
Support intensity. Pilots come with agronomic support, monitoring, and frequently with the capital cost covered. The scaled version comes with a leaflet and a co-financing rate.
Duration. A three or five year project ends before the practice has met its first genuinely bad year, which is the only real test of an adaptation.
General equilibrium effects. A water saving practice adopted by one farm in a catchment behaves entirely differently from the same practice adopted by all of them, because the water saved by the first is available to the others.
None of this is a criticism of demonstration projects, which do exactly what they were designed to do. They establish that a practice can work under favourable conditions with competent support. That is necessary information and it is not adoption evidence, and treating one as the other is the error that produces a scaled programme with disappointing uptake and no explanation for it.
The uncomfortable observation is that the questions the forum identified as Europe's next challenge have been studied intensively for forty years, in development economics, about farmers in Asia and Africa.
Feder, Just and Zilberman's 1985 survey of agricultural innovation adoption in developing countries set out the constraint list that European adaptation research is now assembling from scratch: limited access to information, lack of credit, risk aversion, inadequate farm size, tenure arrangements that misalign incentives, insufficient human capital, labour shortages that prevent timely operations, unreliable supply of complementary inputs, and poor transport infrastructure.
Read that list against a Mediterranean farm facing drought and almost every item transfers. Read it against the specific adaptations on offer, and the mapping is close to exact.
Their second point is subtler and more useful. They argued that treating adoption as a binary, adopted or not adopted, is inappropriate where the real decision is continuous. Farmers adopt a share of a practice, on a share of their land, for a share of the season. A drought tolerant variety on the two worst fields is a different decision from full conversion, and a survey that asks whether a farmer has adopted will record both as identical or, worse, record partial adoption as failure.
A more recent meta-analysis of the empirical adoption literature confirms the durable findings and adds a correction worth noting. Technologies assumed to be scale-neutral, such as improved varieties and chemical inputs, show significant effect sizes associated with scale, which suggests fixed costs are attached to them, whether in transport, in the knowledge required to use them, or in the complementary equipment they need. Practices are rarely as scale-neutral as their agronomy suggests.
This is where European and emerging market adaptation research have something real to say to each other, and where the comparison is not decorative. The mechanisms of non-adoption are similar. The institutional settings, the subsidy environment and the advisory systems are different. That combination is exactly what makes comparative work informative rather than merely interesting.
Each row is a reason the pilot result will not reproduce at scale.
None of these concern the practice itself, which is why more agronomic evidence does not close the gap.
Here is the part that the phrase "social as well as environmental benefits" is circling without landing on.
Adaptive capacity is correlated with farm size, capital and tenure security. Irrigation upgrades, water storage, precision equipment, shade infrastructure, insurance and improved genetics all require capital, and several require secure long term control of the land. The farms best able to adapt are the larger, better capitalised, owner occupied ones.
Which means that drought does not simply reduce output. It redistributes land.
A sequence of poor seasons removes the marginal producers first, and the assets are acquired by the neighbours who adapted. This is not a hypothetical mechanism; it is how agricultural consolidation has proceeded through every climatic shock on record. An adaptation programme that improves average adoption while being taken up disproportionately by the already capable does not slow that process. It accelerates it, and reports success while doing so.
That is the distributional question a European adaptation evaluation should be built to answer, and it requires knowing which farms adopted rather than how many. Those are different datasets and only one of them is usually collected.
The same structure appears wherever the capacity to adopt is unequal. It is the argument we made about rice methane practice and about who can afford to plant a tree, arriving in the Mediterranean.
Sample the non-adopters, deliberately and disproportionately. The information about why a practice does not scale is held entirely by the people who did not take it up, and they are the hardest group to recruit and the least likely to appear in a project's own dataset.
Measure adoption as a continuous variable. What share of land, for what share of the rotation, for how many seasons. Binary adoption questions discard most of the signal and misclassify partial adopters as failures.
Follow to the first bad year. An adaptation is only tested by the conditions it was adopted against. A study that ends before a drought has measured intention.
Record the conversion factors, not just the outcome. Water rights, tenure, farm size, succession, debt, labour availability, advisory relationship, and previous experience of a scheme that did not deliver. These determine adoption more reliably than attitudes to climate risk, which is what surveys tend to measure because it is easy to ask.
Compare across institutional settings, not just across farms. A Mediterranean farm and an East African or Southeast Asian one facing comparable water stress differ in subsidy regime, advisory system, tenure law and insurance availability. That variation is the analytical resource, and it is not available to a study that stays inside one policy regime.
Europe knows a great deal about what farmers could do and very little about what they will do, and the second is not obtainable from more of the first. Demonstration projects have done their job. What follows requires a different instrument and a different population: the farms that did not volunteer, studied for long enough to meet a bad season, with the reasons for non-adoption treated as the finding rather than as a disappointing response rate.
The forum was right that Europe should stop demonstrating. It should be equally clear about what it needs to start doing instead, because the thing that comes after a demonstration is not a bigger demonstration.
The Lab works on this in regenerative agriculture and climate and ecosystems, through field research designed around non-adopters rather than participants, across the regions where the comparison to European conditions is informative rather than ornamental.
If you are building an adaptation work package and need the adoption evidence rather than more feasibility evidence, tell us what you need to know.
This is an independent insight piece by Transitions Lab. For the Lab's applied work, see Regenerative Agriculture & Agri-Tech and Climate & Ecosystems. See also Distance, Work, Reward on rice methane practice and the geometry of adoption, The Right That Matters Is to the Tree on why capital and permission decide who plants, The Fire Was Put Out by People Making a Living on why grazing pilots do not become adoption without the livelihood underneath, A Recycled Material Is a Promise on the pattern of demonstration versus adoption running through European circular research, The Municipality Is the Instrument on the last-mile implementing capacity that scaling assumes, The Village Twenty Kilometres Off the Road on sampling the people a programme did not reach, The Trough Before the Dividend on why case-study economics are conditional on persistence and miss the farms that reverted, and Adoption Is Not the End of the Research on why uptake has to be treated as a research question rather than a communications budget. To discuss a study, see Contact.