A bun fight? Glastonbury tickets? The Taylor Swift presale?
£233 million gone in under six hours.
What happened on the ground
Window 2 opened at 10am. Within two hours Defra emailed to say a quarter of the funding had been allocated, then half had gone 90 minutes later, 75% gone by the afternoon and just before 4pm the window was closed with all the money spent.
Today’s reports from the farmers themselves have been hard to read. Many livestock farmers came back from the fields to find they had missed the boat. Others who tried to apply while it was open reported technical problems as thousands submitted at once. Even the best prepared were caught out.
Joe Stanley at the GWCT Allerton Project reported his team could not submit their long-planned, carefully developed application because of RPA portal problems and hours-long waits to speak to RPA staff. Natalie Gaibani of Strutt & Parker said her team hit a range of errors that made an already stressful process harder.
The CLA’s Joe Evans summed up the day as “a scramble for thousands”. In Devon, councillor Cheryl Cottle-Hunkin argued the scheme had effectively become a lottery, with money going to whoever got through a failing online system first rather than to the greatest need or the best outcomes. Another local voice put it as: farmers “should not have to race each other for vital support”.
If you follow farmers on social media, you will have seen many at their desks, screenshots of pages failing to load and timed out applications.
What went wrong, and why?
The race was primed. Defra said back in February it could not set a fixed end date because that depended on how many farmers applied. Window 1 was around 75% allocated by 13 July, and advisers warned farms to be ready to move, because hesitating risked missing funding they were counting on. The NFU had repeatedly warned the budget was insufficient. Every signal told farmers the same thing: this is scarce, and the fast will win. The starting pistol was set.
Scarcity plus a coordination problem makes a stampede. Once everyone believes everyone else will rush, rushing is the only rational move. Add financial strain, withdrawn funding, rising input prices and a bad harvest all adding demand on funds, and you get a scarcity mindset. Financial pressure measurably reduces cognitive capacity. Tellingly, one of the landmark studies showed this in farmers, who performed worse on cognitive tests before harvest, when money was tight, than after it (Mani et al., 2013). Attention narrows to “get in now”, leaving little room for considered choices about which actions actually suit the farm or the time to complete the application well.
The progress updates became a countdown clock. The 25%, 50% and 75% updates were well meant as transparency. Behaviourally, they work like the “only 2 rooms left!” banner on a hotel booking site when there is only one hotel. They amplify urgency and pull in anyone still deliberating.
The say–do gap. As the window opened, the RPA’s chief executive said the focus was on clear guidance, predictable processes and “the confidence to apply when they are ready”. Unfortunately, the design of the application did the opporite. Punishing deliberation and rewarding speed. When the message and the mechanism contradict each other, people believe the mechanism, and they remember the message.
Friction landed on the wrong people. A 10am Tuesday opening in late September falls in the middle of drilling, livestock work and market days. Portal glitches, phone queues and patchy rural broadband decided outcomes as much as eligibility did. This is what Herd and Moynihan (2018) call administrative burden. The learning, compliance and psychological costs of getting through a process are never neutral: they fall hardest on those with the least time, support and capacity, and in effect they decide who receives public money. The safety nets were hidden too: the two exceptions both required farmers to have contacted the RPA before the window closed, a rule many people only learned after the fact.
Loss aversion was triggered at full strength. Voiding half-completed applications turned hours of invested effort into a confiscation. Losses weigh roughly twice as heavily as equivalent gains, well documented in Kahneman and Tversky, 1979, so taking away something people had already started building feels far worse than a simple “not this time”. That asymmetry is where much of the anger is coming from.
In fairness to Defra…
None of this means Defra or the RPA acted in bad faith, delivering in this situation is never going to keep everyone hapy and SFI is a cash-limited scheme. Defra cannot write an open-ended cheque against demand-led uptake, and the budget was stretched thin to accomodate rather than cut: the £233m combined £180m allocated to Window 2, an extra £50m announced by the Prime Minister in August for farmers facing drought and rising costs, and £3m left over from Window 1.
We saw efforts to spread the money fairly across applicants, no SFI26 agreement can be worth more than £100,000 a year, and each farm business can hold only one agreement, so the funding reaches more farms. Defra also listened on timing. Around 21,000 farms with agreements expiring by the end of February 2027 could apply early for their next agreement, a fix the NFU had lobbied for.
First-come-first-served also has an administrative logic. It is simple, it avoids subjective scoring, and it gets money moving quickly, all of which matter when you are accountable for public funds and working with the farming calendar. The RPA is working with legacy systems that were likely designed when Windows ’95 was launched (note this is unconfirmed and assumed by the author!) under intense load, and the demand shows how much the sector values the scheme….if values is the right word to use here. Looking ahead, ministers have publicly committed to keeping SFI’s main design stable for the rest of this Parliament, which is exactly the kind of predictability farmers have asked for….even if we don’t know who will be in or out and when.
The minister described this year’s scheme as “simpler and fairer”, and the application rules may well be.
The trouble is that fairness is judged on the whole experience, not the rulebook. On Tuesday, the release process undid much of the good work that went into the scheme and it is going to take a herculean effort to re-gain the ground. Here is where I feel it has set us back in terms of framing, behaviour and reputation…
Farmer trust takes another gut punch. Research on procedural justice shows people will accept a disappointing outcome if the process feels fair, transparent and respectful. When the process feels unfair, their willingness to cooperate with the institution behind it falls, whatever the outcome (Tyler, 2006). This time many farmers got neither a good outcome nor a fair process. Trust is a policy asset that has expereinced considerable erosion under this government and it decides whether farmers engage with the next scheme, the next consultation and the next data request…and we have a lot to be asking of farmers if upcoming reports are indicating correctly.
Stewardship becomes a transaction. Many farmers enter environmental schemes partly from intrinsic motivation, a sense of responsibility for the land. A review of conservation policy found that financial incentives can crowd out that motivation when they are framed or delivered in ways that feel controlling or purely transactional (Rode, Gómez-Baggethun and Krause, 2015). A race recasts environmental action as a smash-and-grab, which is exactly the framing that risks eroding the motivation you most want to keep and the impacts and outcomes you are seeking become less likely.
The money may go to the wrong places. A speed-based allocation favours the digitally connected, the agent-supported and the well-staffed (Herd and Moynihan, 2018). Those are not necessarily the farms with the greatest need or the best environmental outcomes. There is also a quality risk: applications assembled in a rush may choose actions that are quick to select rather than right for the land, which stores up compliance headaches later…which I am sure we will come to see as a beast of many heads in due course.
Advisers and wellbeing suffer too. Agents who promised clients they’d get them in are carrying the reputational fallout. Farm families who spent a day watching loading screens instead of working, and still missed out, are carrying the stress into an already tough season. The stress and mental heath of farming is not in a place where this should be taken lightly.
The race gets worse next time. Farmers have now learned that speed is everything. Next round, expect earlier preparation, more agent dependency (thus more disparity), more pressure at opening and an even faster sell-out, unless the design changes.
What could we design differently next time?
Behavioural science doesn’t create budget out of thin air so we can’t expect miracles, but it can change how a limited budget is experienced by the audience, which is important when there isn’t enough money to go round.
- Replace the race with a fixed window and a transparent allocation. This is probably the most important one for me even if it comes with admin and cost. Keep the window open for two to four weeks, then allocate using published criteria, pro-rata scaling, or an openly declared ballot (a ballot being the least cost option). Explicit randomness is often seen as fairer than an accidental lottery decided by broadband speed, and a visibly fair process protects trust even for those who miss out (Tyler, 2006).
- Run an expression-of-interest stage first. Pre-registration gives Defra hard demand data to size each window, allows forms to be pre-populated from RPA records, and removes the incentive to pile in when the buzzer sounds. It also strips out much of the administrative burden for both sides before the pressure is on (Herd and Moynihan, 2018).
- Set expectations in plain numbers. Tell farmers something like “This budget will fund roughly X agreements; we expect around Y applications.” An honest reference point narrows the gap between expectation and outcome, and that gap is where anger lives.
- Protect effort already invested. Timestamp applications when they are started, or give in-progress applications a grace period to finish. That removes the sense of loss that makes expired applications so painful exacerbated by a ‘near miss’ feeling.
- Rethink the live updates. Share demand information before the window opens so farmers can plan, rather than sounding a countdown during it that adds fuel to the fire.
- Brief trusted messengers early. Farmers weigh information by who it comes from. Research on English agri-environmental advice shows trust sits with advisers farmers already know and see as independent (Sutherland et al., 2013). Agents, the NFU, the CLA and other farming organisations should hold the same clear script as Defra ahead of opening and be able to work from a position of understanding, not just for their benefit but to protect and enhance advisory networks ahead of the next ask.
- Soften the landing for those who miss out. possible but I know probably unlikely though there is a little precedent for this based on who could apply previously. Guaranteed priority or a fast-track in 2027 turns a loss into a deferral and keeps people engaged rather than resentful. It also signals that environmental delivery is valued, rather than just being a prize for the quickest (Rode, Gómez-Baggethun and Krause, 2015).
- Test the process, the message and the outcome. Load-test the portal, user-test the journey with farmers and agents, and run a behavioural pre-mortem: “It’s 4pm, the window’s shut and the industry is furious. Why?” Most of what went wrong on Tuesday would have come up in that conversation if it were had.
Turning support for nature and sustainable food production into a scramble has once again devalued the food system, broken the bond between government and farming and left many farmers in the dark about how they will be able to deliver either and keep farming at a time when food imports tip 70% and geopolitical unrest reaches record highs and global weather events require the word ‘super’ to be prefixed to their name.
If anyone needs me, I’ll be in the cupboard with the gin.
I would be really interested to hear views from those applying, or who supported those applying on where they think the process can be improved, not just from a behavioural perspective but more widely. This process will be re-run and it is important not to dwell on what went wrong for too long when there are so many ways it could be improved.
Just a final note to say that anyone in farming who is facing this as a challenging time, please reach out to RABI or FCN for guidance on where to seek help and support.
References (for the nerds!)
Herd, P. and Moynihan, D.P. (2018) Administrative Burden: Policymaking by Other Means. New York: Russell Sage Foundation.
Kahneman, D. and Tversky, A. (1979) ‘Prospect theory: an analysis of decision under risk’, Econometrica, 47(2), pp. 263–291.
Mani, A., Mullainathan, S., Shafir, E. and Zhao, J. (2013) ‘Poverty impedes cognitive function’, Science, 341(6149), pp. 976–980.
Rode, J., Gómez-Baggethun, E. and Krause, T. (2015) ‘Motivation crowding by economic incentives in conservation policy: a review of the empirical evidence’, Ecological Economics, 117, pp. 270–282.
Sutherland, L.-A., Mills, J., Ingram, J., Burton, R.J.F., Dwyer, J. and Blackstock, K. (2013) ‘Considering the source: commercialisation and trust in agri-environmental information and advisory services in England’, Journal of Environmental Management, 118, pp. 96–105.
Tyler, T.R. (2006) Why People Obey the Law. Princeton, NJ: Princeton University Press.