The Tractors Are Back. The Nitrogen Constraint Never Left.

On 27 July, around 20 tractors gathered near the Merwedebrug, with other farmers demonstrating along the A16 near the Moerdijk Bridge. They displayed upside-down Dutch flags and signs asking for help but stayed off the motorway and did not block the bridges. Two days later, Agractie and Farmers Defence Force said they would not discuss the government’s nitrogen package with Agriculture Minister Jaimi van Essen unless he first withdrew it. The protest tactics were restrained; the precondition for talks was not. (Omroep Brabant; Nieuwe Oogst)

That distinction broadly matches the public response. A Hart van Nederland panel survey of 3,268 participants found that 68% understood why farmers wanted to protest. Demonstrations at designated locations were considered acceptable by 93%, and protests at municipal or provincial buildings by 84%. Only 25% accepted blocking motorways or provincial roads. The panel is not a substitute for a probability-sample national poll. Its internal contrast is still useful - sympathy for farmers does not automatically extend to every tactic. (Hart van Nederland)

The immediate object of the protests is the package announced by the Jetten cabinet on 26 June. It proposes cutting agricultural ammonia emissions by 42% to 46% by 2035 relative to 2019. It also proposes farm-level emissions standards, stronger linkage between dairy production and land, additional rules around roughly 100 nitrogen-sensitive nature areas, and substantial spending on farm adaptation, extensification, closure, relocation, and nature restoration. Industry and mobility would receive sectoral reduction objectives as well. The cabinet presented the package as its route out of the stikstofslot - the “nitrogen lock” constraining permits for farms, housing, infrastructure, energy projects, and industrial investment. (Rijksoverheid; Tweede Kamer)

It has not yet delivered that outcome. Its central components still require legislation, implementing rules, area-level decisions, monitoring, and legally defensible application to particular Natura 2000 sites. The government has announced a destination and committed serious money. It has not created a bankable stock of permits.

The protest is the immediate confrontation. The harder question is the one the Netherlands has spent seven years circling: what would a nitrogen settlement have to accomplish if it were genuinely lawful, ecologically effective, and economically coherent - and is the agricultural sector prepared to participate in one?

Farmers have strong reasons to distrust the state. The Programme Approach to Nitrogen collapsed in court after government had relied on future ecological gains to authorize present activity. The resulting legal uncertainty has proved particularly damaging for the PAS-melders, businesses that had operated under the notification regime the state itself created. Later attempts to legalize them have been slow enough that the government was still designing a successor programme in 2026. That record does not establish that every farmer-side objection is correct, but it does establish that a fresh ministerial assurance is not a risk-free asset. (Raad van State; Rijksoverheid)

Distrust does not alter atmospheric chemistry, restore damaged habitats, or amend Article 6 of the EU Habitats Directive, however. Agriculture is not the only source of reactive nitrogen, and a fair settlement must impose visible obligations elsewhere. Yet agriculture accounts for nearly 90% of Dutch ammonia emissions and is the largest domestic contributor to average nitrogen deposition on Natura 2000 sites. Taken together with the legal duty to prevent deterioration, those facts support a difficult but defensible conclusion - a credible settlement cannot avoid substantial, verifiable agricultural reductions. (Emissieregistratie; CLO source shares)

What remains negotiable is the route. The current package combines proposed farm standards, feed and manure measures, land-based dairy production, spatial targeting, voluntary exits, and nature restoration. Other lawful options include relocation, different treatment of livestock rights, and, as a last resort in narrowly defined cases, compulsory acquisition with full compensation and legal protection. The mix, pace, geography, and distribution of costs are political choices. Preserving every existing production model is not a plausible baseline.

Nor is “agriculture” one negotiating party. In a July 2025 joint proposal, LTO, NAJK, provinces, municipalities, and water boards accepted an agricultural reduction objective of 42% to 46% by 2035, tied to accountable sectoral and farm-level goals. LTO later objected strongly to the June 2026 package’s zoning and land-linkage rules. NAJK welcomed movement on goal-based steering and permitting while calling important parts of the package inadequately justified. Agractie and FDF currently demand withdrawal before talks. These positions are not interchangeable. There is no single farmer position for the government to accept or reject. (LTO joint proposal; LTO response; NAJK; Nieuwe Oogst)

One word, several mechanisms

The Dutch political word stikstof compresses several substances and processes into one convenient noun. The atmosphere is mostly nitrogen gas, which is not the policy problem. The relevant reactive compounds are mainly ammonia, or ammoniak (NH₃), associated predominantly with agriculture, and nitrogen oxides, or stikstofoxiden (NOₓ), produced by combustion sources such as traffic, shipping, industry, and energy use. (RIVM definitions)

An emission is the amount released by a source. A concentration is the amount present in the air. Deposition is the quantity that eventually lands on vegetation or soil. Ecological effects arise after that deposition interacts with the habitat’s condition, soil, and water system. A modelled deposition estimate is not an emission measurement, and neither is automatically a legal limit. Confusing these categories makes it possible to move from a valid criticism of one small project-level estimate to the much less defensible conclusion that the national problem has been invented. (RIVM definitions)

The source data leave little doubt about ammonia. In 2024, Dutch agriculture emitted 101.1 kilotonnes of ammonia out of a national total of 112.8 kilotonnes—about 89.6%. That does not mean agriculture caused 89.6% of deposition in every protected area. Emissions travel, transform, and deposit differently; Dutch emissions cross borders and foreign emissions arrive here. It does mean that a national ammonia strategy from which agriculture is largely absent would be an unusually elaborate way of avoiding the principal source. (Emissieregistratie)

The deposition figures tell a related but different story. For 2023, the Compendium voor de Leefomgeving estimated that Dutch agriculture contributed 47% of average nitrogen deposition on Dutch Natura 2000 areas. Foreign sources contributed 36%, Dutch mobility 12%, Dutch industry 2%, and other Dutch sources 3%. These are modelled national averages, not site-specific verdicts. Even with that qualification, agriculture was neither the whole problem nor an incidental contributor. It was the largest domestic source by a wide margin. (CLO source shares)

The foreign share is large enough to sound exculpatory. It is less useful as an exemption. The Netherlands cannot directly regulate a German farm or a Belgian motorway, but it remains responsible for the domestic measures needed to protect Dutch Natura 2000 sites. Dutch pollution also travels outward: the same official indicator estimates that the Netherlands exports roughly four times as much ammonia deposition as it imports. Cross-border pollution is an argument for coordinated European reduction, not for waiting until every neighbour moves first. (CLO source shares)

National averages also conceal geography. In a 2025 technical note based on 2022 emissions, RIVM estimated that agricultural and industrial sources within 500 metres of nitrogen-sensitive Natura 2000 areas contributed about 82 moles of nitrogen per hectare per year to average deposition on those areas. Extending the zone to 1,000 metres increased the contribution to about 127 moles. RIVM described the exercise as indicative and based on limited data; it was not a project-level assessment of every farm inside a circle. (RIVM zones)

Sources within those 1,000-metre zones accounted for about 9% of average total deposition on nitrogen-sensitive nature. The remainder includes Dutch sources farther away and foreign sources. That result cuts in two directions. It supports stronger attention to sources near vulnerable habitats, but it does not establish that a one-kilometre policy could replace all wider reductions. Location matters. So does the load arriving from beyond the immediate neighbourhood.

A 2026 RIVM review gives targeted policy stronger support than the zone averages alone. RIVM reviewed scenario calculations prepared by Gispoint and concluded that reducing emissions closer to nature areas can produce more deposition reduction than an equivalent generic reduction. Targeting can therefore reduce how much has to happen elsewhere. RIVM also found that some Gispoint scenarios understated the emissions reduction needed when recalculated with newer data; in one 2030 scenario, roughly 10 additional kilotonnes of ammonia reduction were required. Targeting may be more efficient, but its apparent precision still depends on the data, boundaries and measures used. (RIVM targeted scenarios)

The ecological mechanism is more specific than “nitrogen is bad for nature.” Excess deposition can fertilize nutrient-poor habitats, acidify soils, and alter which plants can compete. As plant communities change, the habitat available to dependent animals changes with them. Nitrogen is also rarely the only pressure. Drying, altered water systems, poor water quality, and habitat fragmentation can interact with eutrophication and acidification. This is why source reduction and habitat restoration are complements: lower deposition does not repair every hydrological problem, while restoration cannot indefinitely compensate for an excessive atmospheric load. (CLO ecological pressure)

The kritische depositiewaarde (KDW), or critical deposition value, estimates the annual deposition above which a significant risk of harm to a habitat type cannot be ruled out. It is an ecological risk indicator, not a project-specific verdict and not an EU-prescribed emissions ceiling. Exceeding the KDW does not prove that one additional farm caused the habitat’s condition, and falling below it does not prove that every other ecological pressure has been solved, but widespread exceedance is evidence that nitrogen remains a material pressure, and it matters when government must demonstrate that it is preventing deterioration.

In 2023, deposition was below the KDW on 30% of the nitrogen-sensitive habitat area assessed by RIVM. Under the policies sufficiently concrete for its 2025 forecast, RIVM projected 32% to 34% below the KDW in 2030, against the then-current statutory target of 50%, and 33% to 39% in 2035, against a target of 74%. Forecasts and statutes can change. The scale of that gap is harder to wave away. (RIVM monitor)

Farmers are entitled to point out that agricultural ammonia emissions fell from 336.6 kilotonnes in 1990 to 101.1 kilotonnes in 2024, a decline of about 70%. The inventory does not apportion that fall among regulation, technology, herd changes and management, but it establishes an important baseline. Past reductions should shape the fairness of the next stage; they do not determine whether the remaining load is sufficient. (Emissieregistratie)

The legal lock is not a model setting

The principal legal constraint comes from Article 6 of the EU Habitats Directive. Member states must establish conservation measures and avoid deterioration of protected habitats. A plan or project likely to have a significant effect must undergo an appropriate assessment. The competent authority may approve it only after establishing that it will not adversely affect the integrity of the site, unless the strict derogation route in Article 6(4) is satisfied. (Habitats Directive; IPLO assessment rules)

This does not prescribe a Dutch livestock number, write the KDW into EU law, or require a particular atmospheric model. It does dictate the result the permit system must protect. Dutch legislators retain substantial discretion over national targets, farm standards, zoning, compensation, restoration, and the distribution of burdens. They do not have discretion to declare that deterioration no longer counts because the political argument has become exhausting.

For most Natura 2000 activities, the province is the competent authority, although the national government or another authority decides in specified cases. The authority must apply the Omgevingswet, the Dutch law governing the physical environment, consistently with EU law. Courts do not normally choose which farm should change or write an agricultural transition plan. They review whether the authority had a sufficient legal and scientific basis for the decision before it. (IPLO assessment rules)

The modern crisis began with the collapse of the Programma Aanpak Stikstof (PAS), the Programme Approach to Nitrogen. The PAS permitted activities partly on the expectation that future source measures and restoration would later create ecological improvement. In practical terms, it treated anticipated recovery as present permitting currency.

The Court of Justice of the European Union scrutinized that construction in 2018. On 29 May 2019, the Administrative Jurisdiction Division of the Council of State - the Netherlands’ highest general administrative court - ruled that the PAS could no longer serve as a basis for permission. Expected future gains were not sufficiently certain to justify present authorization. The state had tried to spend the improvement before it existed. (Raad van State: PAS)

That holding explains much of what followed. A cabinet cannot reopen permitting merely by announcing a national ambition. Measures relied upon in a permit must be sufficiently certain and legally usable. Improvement already needed to meet an existing conservation duty cannot automatically be counted again as mitigation for a new project. A measure that might be adopted later cannot simply be booked today. And the same ecological gain cannot be allocated to several projects at once, however tidy the spreadsheet may look.

The December 2024 decisions on intern salderen, or internal offsetting, tightened this logic. Internal offsetting compares the emissions from a proposed activity with emissions from the previously permitted activity at the same site. The Council of State ruled that this comparison could no longer be used at the preliminary screening stage to conclude that no nature permit was needed. It could still be considered later in the permit assessment, under stricter conditions. The court also provided transitional rules, but the change increased the legal exposure of activities undertaken under the earlier interpretation. (Raad van State: internal offsetting)

The attraction of a rekenkundige ondergrens - a numerical calculation threshold - is therefore understandable. AERIUS combines data and atmospheric models to estimate deposition from projects, and those estimates have uncertainty. RIVM’s own work distinguishes uncertainty in national totals from the generally larger and more context-dependent uncertainty around individual sources. It also warns that there is no single universal error margin that can simply be attached to every project result. (RIVM model uncertainty)

That matters for fairness and administration. A model output is not a direct measurement, and the consequences for an applicant can be more discrete than the scientific estimate. Rules should not present every decimal place as a personal note from the atmosphere. But uncertainty does not make every small contribution zero, nor does it invalidate the broader emissions inventory or national source pattern.

The institutional response is also changing. On 1 July 2026, RIVM said it would retain management of AERIUS temporarily, but under conditions intended to separate its scientific role more clearly from the execution of policy choices. Scientific advice and ministerial decisions are to be published separately, while RIVM and the ministry evaluate the role of model calculations in permitting and improve how uncertainty and calculation resolution are described. That is not an admission that AERIUS is worthless. It is recognition that model governance matters when technical outputs carry legal consequences. (RIVM AERIUS governance)

In May 2025, the Council of State’s advisory division assessed a proposed threshold of one mole per hectare per year and found the design legally vulnerable. It accepted that very small contributions could not be distinguished from zero with the claimed measurement certainty but emphasized that inability to measure a source separately did not mean the source caused no deposition. It also pointed to accumulation: many individually small contributions can add up to a material load. (Raad van State: threshold)

A threshold may still form part of a lawful system, particularly if it rests on an adequate scientific basis and a credible package controlling cumulative effects. What the advice rejected was the idea that legal and ecological space could be created merely by changing where the calculator stops displaying a result.

The KDW debate follows the same pattern. Critics are correct that the KDW is not the text of Article 6 and should not be treated as an automatic project ban. The Dutch legislature could replace KDW-based national targets with a different architecture. But a replacement would still have to show how the state will prevent deterioration, improve habitat condition and assess projects lawfully.

In December 2025, the Council of State advised against submitting the then-proposed replacement bill in its existing form. The bill would have removed statutory targets requiring specified shares of nitrogen-sensitive habitat to be below the KDW and replaced them with an open-ended 2035 emissions objective. The advisory division did not say emissions steering was impermissible. It said the proposal lacked adequate guarantees, intermediate obligations, and a convincing explanation of how it would meet the nature duties. (Raad van State: statutory targets)

The January 2025 Greenpeace judgment added another constraint. The District Court of The Hague held that the state was acting unlawfully by failing to stop deterioration and by being very unlikely to meet its own 2030 target. It ordered the state to ensure that 50% of nitrogen-sensitive habitat area was below the KDW by 31 December 2030, with priority for the most overloaded areas, and attached a €10 million penalty for non-compliance. The State appealed in April 2025. The appeal means the ruling should not be described as final appellate law; the district court nevertheless declared its order provisionally enforceable. (Rechtspraak; State appeal)

None of this means that building, farming, or infrastructure is categorically impossible. It means permission is difficult where a protected site is overloaded, and the authority cannot exclude harm or rely on secured, additional improvement. The practical exit is a combination of real source reductions, defensible restoration, credible monitoring, and rules for allocating any resulting permit space without double counting.

Even then, one tonne of reduced emissions does not automatically become one tonne of immediately usable permission. The legal assessment is site-specific. Authorities must consider the affected habitat, the durability and additionality of measures, and whether the improvement is already needed to meet an existing duty. This is why the June package may improve the route to permitting without having reopened it merely by announcement.

The state taught farmers not to trust it

The PAS is the clearest example: government allowed activity on the strength of expected future improvement, and the courts rejected that legal basis. The notification regime left PAS-melders exposed even though they had followed the rules then presented to them. By January 2026, the ministry was still reporting on a successor to the legalization programme, and in April it published a new draft programme. A state with that record cannot reasonably demand uncomplicated confidence in its next design. (Rijksoverheid: PAS-melders)

Nitrogen policy also operates on businesses whose physical assets and land use cannot be rearranged on a ministerial timetable. A livestock-housing or manure system installed for one regulatory regime may not deliver what a later regime requires. That does not prove the original investment was prudent, or that government must preserve its value indefinitely. It does explain why apparently technical changes can become household-level transition costs.

“Voluntary” closure should therefore be described honestly. A farmer may formally choose whether to accept a buyout. If the alternative is a business facing tighter permit, land, or emissions constraints, the choice may feel heavily structured by public policy. That does not make a buyout illegitimate. It means compensation and procedure should recognize that the state has narrowed the available options.

The government also has a sequencing problem. Farmers may need to invest in lower-emission systems while the permit route for those investments remains uncertain. LTO’s response to the June package accepted the need for a substantial agricultural contribution but argued that emissions reduction and permission to invest must advance together. That is a serious objection. A modernization programme that cannot tell businesses whether modernization can be permitted is not yet an investment strategy. (LTO response)

These complaints support a demanding transition settlement: clear standards, usable schemes, fair payment for closure and relocation, and particular attention to younger farmers whose investment horizon is longest. NAJK is right about at least one administrative point: the nominal size of a budget matters less than whether the instruments arrive on time, have intelligible eligibility rules and can actually be used. That is NAJK’s position, not independent proof that every scheme will fail. It is nevertheless a warning the government has earned. (NAJK)

These complaints justify compensation and procedural reliability. They do not establish a right to preserve every production model where an activity’s environmental claim is too large for its location.

Nor did farmers create the institutional failure alone. The state designed PAS, Parliament enacted the legal framework, authorities administered permits, and courts applied the Habitats Directive and Dutch law when those arrangements were challenged. The relevant responsibility is systemic without being equal; government bears special responsibility for rules it created, while the agricultural source contribution remains central to the remedy.

The June package itself recognizes that primary producers cannot carry every market adjustment alone. It proposes binding agreements with supermarkets and processors by April 2027 to support demand for more sustainable products, with statutory measures from 2029 if agreements are not reached. That does not prove that retailers or processors caused the nitrogen crisis. It does support a narrower principle - if government wants farmers to change production systems, demand-side and supply-chain arrangements belong in the transition rather than being treated as somebody else’s problem. (Rijksoverheid)

The state’s conduct has earned suspicion. What follows is not a right to regulatory stasis, which no democratic government can guarantee. A more defensible demand is that the new standards be legally anchored, monitorable, funded and designed to survive foreseeable judicial review. That is a high bar. After the PAS, it is not an unreasonable one.

BBB’s meaningful but incomplete test

The rise of the BoerBurgerBeweging, or Farmer–Citizen Movement, appeared to test the claim that the nitrogen crisis mainly reflected the wrong people controlling policy. BBB entered the Schoof cabinet when it was sworn in on 2 July 2024 and held the agriculture ministry through Minister Femke Wiersma. The party therefore had more than an outside platform from which to criticize earlier governments. (Schoof cabinet)

During that period, the ministry pursued several ideas associated with the farmer backlash: farm-level goal steering, voluntary measures, a numerical calculation threshold, and replacement of KDW-based statutory targets with emissions-oriented goals. These were not frivolous questions. Outcome standards can give farms more flexibility than prescribing one stable design, and a defensible threshold could reduce false precision around tiny project contributions.

The governing experience nevertheless exposed the difference between changing the policy frame and changing the constraint. The proposed one-mole threshold was judged legally vulnerable in the form presented. The first KDW-replacement bill was found to lack the guarantees needed to show timely compliance with the nature duties. Neither result proves that every possible version of emissions steering or threshold relief must fail. Both show that renaming the system is not enough. (Raad van State: threshold; Raad van State: statutory targets)

BBB’s test was incomplete. The Schoof cabinet lost the PVV ministers on 3 June 2025 and continued in caretaker form until the Jetten cabinet took office on 23 February 2026. A short and unstable coalition inheriting years of legal damage is not a controlled experiment in which every BBB idea received ideal implementation. (Schoof cabinet; Jetten cabinet)

But it was meaningful. The agriculture ministry placed the threshold and target redesign before the institutions that had to assess them. The objections were not produced by a shortage of sympathy for farmers; they arose because the proposals still had to address cumulative effects, deterioration, and nature recovery.

By July 2026, BBB was no longer part of the executive. The Jetten government is a D66-VVD-CDA coalition, and the Tweede Kamer faction register lists BBB with three seats. The June package is therefore not BBB policy. BBB is entitled to oppose it. Its governing record remains relevant for a narrower reason - political victory over one nitrogen plan did not produce a mechanism through which every painful choice could be avoided. (Jetten cabinet; Tweede Kamer factions)

Farm organizations do not have to accept the Jetten package as written. They do have to explain what replaces its reductions. “Innovation” names a category of instruments, not a verified aggregate result. “Measure the actual state of nature” is sensible, but measurement does not itself stop deterioration. “Remove the KDW” may alter Dutch target architecture; it does not remove the Habitats Directive. The closer a movement comes to governing, the less work the word “different” can do.

A serious package that is not yet a settlement

The June package is more candid than many earlier approaches about the scale of the task. It proposes agricultural ammonia emissions 42% to 46% below 2019 levels by 2035. It would establish a farm-level dairy standard for 2035, with standards for pigs, poultry, and calves to be developed earlier, and allocates €2 billion for measures including feed and livestock-housing changes. The remaining agricultural reduction is expected to come from less manure application, extensification, voluntary closure, and partial cancellation of livestock production rights when they are transferred outside a family. (Rijksoverheid)

For dairy farming, the cabinet proposes grondgebondenheid, stronger linkage between livestock and available land, through a maximum stocking density of 2.6 grootvee-eenheden per hectare in 2035. A grootvee-eenheid is a standardized livestock unit, not necessarily one literal cow. (CBS) The cabinet presents the rule as a route toward more land-based production and less manure pressure, with some scope for cooperation between dairy and arable farms. (Rijksoverheid)

Around 100 nitrogen-sensitive areas, additional rules would apply. About 15 are expected to receive 1,000-metre zones and the remainder 500-metre zones. The cabinet has set aside €9 billion for measures including agricultural land revaluation, extensive operations and conversion to organic production. It also proposes an approach for major industrial sources in the zones. (Rijksoverheid)

Nature restoration and other sectors are included. The package assigns €2.2 billion to nature management and restoration. Industry and mobility are each expected to achieve a 50% nitrogen reduction by 2035, much of it through existing climate and environmental policy, supplemented by €250 million in nitrogen-specific measures. The cabinet describes the wider package as a €20 billion commitment. (Rijksoverheid)

The package attempts to recognize prior performance by crediting farms that have already reduced emissions rather than prescribing identical equipment. The principle is attractive, but a farm-level system must specify what is measured or calculated, over what period, with what uncertainty, who verifies compliance and what happens when a credited technique underperforms. Without those safeguards, the state risks counting reductions before they are sufficiently secured.

The stocking-density proposal also deserves more than reflexive approval or rejection. A uniform limit will not impose an identical adjustment on an extensive dairy farm with ample land, an intensive business in a high-price land market and a mixed operation able to cooperate with arable farms. NAJK argues that the 2.6-unit rule is an input prescription insufficiently tied to the environmental result. The cabinet may have an answer, but it must show the causal link and the distributional effects rather than treating a national number as self-explanatory. (NAJK)

The zones present a similar problem. RIVM’s work supports the premise that proximity can make targeted reduction more effective. It does not validate every announced radius or every restriction in advance. The government must show how its general method translates into proportionate measures for particular areas. A 500-metre or 1,000-metre boundary may be appropriate; the roundness of the number is not the evidence.

The opposite simplification is also wrong. Zoning is not arbitrary merely because deposition is modelled. Official scenario work indicates that targeted reductions can reduce the total adjustment needed elsewhere. The defensible criticism is therefore not that location is irrelevant, but that location-specific policy needs transparent ecological, emissions, and proportionality reasoning. (RIVM targeted scenarios)

The largest weakness is the gap between the cabinet’s headline and the package’s legal status. “Taking the Netherlands off the nitrogen lock” is a claimed outcome. The official announcement itself says new nitrogen goals still require a bill and that the measures will be anchored in a programme. The Council of State has already explained why a threshold and redesigned statutory targets need stronger foundations. The cabinet may be assembling those foundations. It should not describe the building as occupied. (Rijksoverheid; Tweede Kamer)

Money does not close that gap by itself. Twenty billion euros is a substantial fiscal commitment, but it does not establish that standards can be verified, that provinces can process the resulting decisions on schedule, that relocation opportunities will exist, or that every restoration measure will be legally additional. These are not predictions of failure, however; they are implementation tests the package must still pass.

What the farmers’ case gets right

Agriculture’s record of past reduction is the easiest part of the farmers’ case to verify. Agricultural ammonia emissions have fallen by about 70% since 1990. The claim that other sectors have escaped equivalent pressure requires a more exact comparison. Agriculture dominates ammonia and is the largest domestic contributor to average Natura 2000 deposition; traffic and industry are more important NOₓ sources. Equal percentage cuts across unlike pollutants and locations might look even-handed while producing an uneven ecological result. (Emissieregistratie; CLO source shares)

Model uncertainty is a serious objection when a small project-level estimate can determine the fate of an individual farm. At small increments, the scientific and procedural concern is real. AERIUS results are modelled estimates, and RIVM acknowledges that uncertainty for individual sources is context-dependent and can be substantial. (RIVM model uncertainty)

The boundary is that project-level uncertainty does not erase robust larger-scale findings. It does not make the national ammonia inventory disappear, turn agriculture’s 47% average deposition share into an unknowable quantity, or show that location has no effect. A sensible response is to improve models, publish uncertainty more clearly, and design a lawful approach to contributions judged negligible within a controlled cumulative system. Treating every uncertain contribution as zero would merely convert uncertainty into a policy assumption.

The KDW objection is also partly right. A project does not become unlawful merely because a national map shows an exceedance, and the legislature may choose another target architecture. But the replacement must still show whether habitat condition is deteriorating, whether measures are working, and whether project approval is compatible with site integrity. Removing the dashboard warning is not a repair strategy.

Innovation deserves the same conditional answer: it could achieve more than government models assume. Feed, manure, and housing measures are included in the cabinet’s package, and farm-level standards could encourage combinations ministries would not design themselves. But a credible innovation pathway requires verified performance over time and a response when delivery falls short. The dispute is not “technology or livestock reduction.” It is how much verified technology can deliver, by when, and what covers the remainder. (Rijksoverheid)

Targeted action has a genuine empirical case. A targeted approach near sensitive areas can be more efficient than a flat national reduction. RIVM’s review of the Gispoint scenarios supports this. Reducing emissions near nature can generate more deposition reduction, allowing less intervention elsewhere. That is a significant point in favour of spatial differentiation. (RIVM targeted scenarios)

It is not evidence that a ring around each site can carry the whole strategy. The earlier zone note found that sources within 1,000 metres accounted for about 9% of average total deposition, while the scenario review cautioned that some targeted calculations understated the reduction needed when newer data were used. The strongest conclusion is additive - targeted action can reduce the total cost and disruption of a broader strategy, but the amount and location of the remaining reduction must still be demonstrated. (RIVM zones; RIVM targeted scenarios)

The arguments about rural communities, food security, and production leakage are harder to quantify, not easier to dismiss. A substantial contraction of livestock production could affect employment, suppliers, land markets, and the continuity of farm businesses. Some production might move abroad, potentially reducing the European or global environmental benefit. Dutch agriculture also has cultural and strategic value not captured by a narrow farm-income measure.

Those are serious considerations, but the research record does not establish a reliable leakage estimate for the June package or a single quantitative measure of the sector’s non-market strategic value. They should therefore shape sensitivity analysis, regional support, and policy design - not be asserted as either negligible or decisive. The right question is not whether leakage exists in theory, but how much a particular measure causes, where production moves, and whether the receiving jurisdiction has higher or lower total impacts.

Cultural value does not settle every land-use decision. Arable farming, horticulture, extensive grazing, intensive livestock production, and dairy farms near protected heathland have different nitrogen profiles. Respecting agriculture does not require treating every configuration as equally sustainable.

The strongest objection to the June package is institutional: government cannot demand trust while compensation, permits, and detailed rules remain unresolved. Farmers are being asked to plan against standards maturing in 2035 while some norms and schemes still have to be designed. The state has not earned the benefit of vague implementation.

That supports a demand for synchronized legislation, funding, permitting, and transition options. It does not support indefinite suspension of the aggregate result. A farm organization can reasonably say, “We will not sign until compensation and legal safeguards are concrete.” It is harder to defend, “We will not discuss implementation until the package is withdrawn, and we will specify a sufficient replacement later.”

LTO and NAJK show that agricultural opposition need not take that form. In July 2025, they joined public authorities in supporting a 42% to 46% agricultural reduction objective by 2035 under accountable sectoral and farm-level goals. Both now seek major changes to the June package. Their prior commitment nevertheless shows that participation in a large reduction pathway is not inherently incompatible with representing farmers. (LTO joint proposal; LTO response; NAJK)

Agractie and FDF are entitled to organize against the package. Their withdrawal-first position is also evidence of their position, not evidence that all Dutch farmers share it. The available record does not establish the organizations’ numerical representativeness across agricultural subsectors. What it does establish is that their current demand does not yet answer the governing question - which enforceable combination would they accept that produces comparable ecological improvement, survives Article 6 review, and can be implemented before further delay narrows the options again?

Without that answer, a new wave of farmers’ protests may defeat a package. But more protests cannot settle the crisis.

What a credible settlement requires

A durable settlement begins by abandoning the promise that every constituency can receive its preferred outcome. Habitats cannot deteriorate indefinitely; permits cannot rest on unsecured future measures; and no agricultural organization can insist that every livestock business remain unchanged regardless of location, emissions, or social cost.

Start with sequence. Government should identify area-specific conservation needs, set credible emissions paths, enact the standards and safeguards needed to deliver them, fund the measures, and build the monitoring system before allocating large amounts of new permit space. Where earlier permits are possible, the state must show that the underlying reductions are secured, additional, and not already required or promised elsewhere. The PAS failed because the sequence ran in the other direction.

The agricultural pathway must then be differentiated. Farm-level outcome standards are preferable where emissions can be measured or estimated reliably enough for enforcement. They need transparent methods, uncertainty rules, independent verification, and a fallback when performance fails. Input rules may remain necessary where outcomes cannot yet be monitored credibly, but they should not become permanent merely because they are more convenient to administer.

Technology should receive a full opportunity, not an advance guarantee. Feed and manure measures, verified housing systems, extensification, relocation, and voluntary closure will contribute in different places. None is the universal answer.

The package’s combination of farm standards, stocking-density rules, rights cancellation, and closures would probably reduce livestock numbers in at least some regions or subsectors. The available evidence does not provide a verified national herd forecast, so the scale should not be presented as certain. Government should nevertheless be candid that structural change is a likely consequence rather than hiding it behind a sequence of technical nouns.

Compulsory acquisition should remain a last resort, used narrowly where a source has a material relationship with an acutely vulnerable site, voluntary routes have failed, and delay would frustrate the public duty. “Last resort” should not mean waiting until every deadline has been missed. It should mean a demanding necessity test, fair compensation, and full legal protection.

Spatial policy must follow evidence rather than cartographic convenience. The most overloaded habitats should receive priority under the 2025 district-court order, while the design of zones should also reflect source patterns, habitat condition, and the likely effect of measures. A 500-metre boundary may be suitable in one place and inadequate or excessive in another. Government needs a defensible method, not an identical answer.

Nature restoration must proceed alongside source reduction. Some sites need hydrological repair, better buffers, or changed management. Those measures affect how lower deposition translates into recovery. They should not be used to pretend emissions no longer matter, just as emissions reduction should not be booked as automatic habitat restoration.

The transition contract must also be credible to the people expected to act. Closure and relocation schemes must offer fair compensation and usable procedures. Extensification and environmental services need durable revenue arrangements rather than a temporary grant followed by exposure to the same production incentives. Younger farmers, tenants, workers, and communities may bear losses that do not appear in the owner’s buyout and should be considered explicitly.

The package’s proposed agreements with supermarkets and processors are therefore relevant, but their role should be stated precisely. The evidence supports requiring supermarkets and processors to help create demand for lower-impact production; it does not support blaming them, without further evidence, for the whole historical crisis. Public support can also be conditioned so that payments for environmental services reach the farms actually changing land use.

Other sectors must face credible burdens of their own. Industry and mobility should face measures connected to their actual contributions and affected locations. Existing climate measures may legitimately produce nitrogen benefits, but government should distinguish reductions already expected from genuinely additional nitrogen policy. Visible obligations matter for legitimacy even where agriculture’s required contribution remains larger.

Fairness does not mean identical percentage cuts for symmetry. Activities differ in pollutant, location, duration, and social return. Once ecological space exists and legal criteria are met, policy should compare the value generated, the capacity consumed and the available alternatives - not preserve historical allocations by default.

All of this depends on administrative realism. Provinces and implementing agencies need staff, data, common methods, and legal authority. Buyouts must be processed, land and relocation opportunities identified, restoration contracted, emissions verified, and permits defended. The June package does not prove that this capacity is absent, but it also does not create it automatically by appropriating money.

Agricultural organizations should have a substantial role in designing implementation. Farmers possess practical knowledge about how measures interact on different businesses, and consultation can reduce needless losses. Participation cannot become a veto over the aggregate outcome, however, any more than environmental organizations should receive a veto over every instrument or permit.

Government, for its part, must stop disguising the direction of travel. Parts of livestock agriculture are likely to contract or be transformed. Some farms will close; others will become more extensive, relocate, change production, or earn more from environmental services. Intensive production may continue where location and verified performance make it compatible with the limits. Ministers should state the range honestly and negotiate how the costs are shared.

Is agriculture prepared to participate in such a settlement? There is no single sectoral answer. LTO and NAJK have accepted large reductions under demanding conditions and continue to engage while opposing parts of the package. Agractie and FDF are mobilizing around withdrawal before negotiation. That position may work as protest strategy, but it has not yet matured into a settlement strategy.

Peaceful protest remains legitimate. Government deserves scrutiny over every model assumption, zone, deadline, and compensation rule. On the other hand, dangerous obstruction deserves no special immunity, and the Hart panel suggests public tolerance falls sharply when protest closes major roads. More fundamentally, repeated rejection cannot function as an indefinite sectoral veto over compliance with law or protection of a shared natural asset.

The government’s responsibilities are correspondingly heavy. It must substantiate the package’s ecological effect, correct weak measures, secure the legal foundation, fund the transition, impose real obligations outside agriculture, and avoid promising permits before it can defend them. It must make political choices openly rather than allowing litigation and regulatory attrition to perform the restructuring by default.

A settlement remains possible because the mix, geography, timetable, compensation, and design of farm standards remain open. Those choices can reduce unnecessary harm, not eliminate every loss or preserve every business model.

Continued delay carries risks in a predictable direction. Vulnerable habitats remain under pressure; permits remain exposed to litigation; farmers defer investments under continuing uncertainty; and in the absence of progress, later governments will face tighter deadlines with ever fewer options. The exact pace of each consequence is uncertain. The mechanism is not.

Farmers are right that the state cannot demand trust it has not earned. Distrust, however justified, does not suspend the underlying obligation. The Netherlands can choose how to reduce nitrogen, where to act first, and how to divide the bill. But it cannot use the same environmental space twice.