Schiphol: what if the airport is in the wrong place?

The location assumption hidden in aviation policy

On 14 September 2026, planner Zef Hemel returned publicly to an old provocation: move Schiphol to Flevoland and use its present site for something else. The intuitive appeal is easy to see. Schiphol puts its airport site at 2,787 hectares, almost 28 square kilometres, at a highly connected Randstad transport node in a country with a large housing shortage. Aircraft noise and safety rules constrain land beyond the airport boundary. Lelystad already has an airport, and additional infrastructure and employment there could have regional value.

But Schiphol is not simply 2,787 hectares of runways waiting to be repurposed. It is a mature international hub embedded in the western Randstad, with rail beneath the terminal, a large local passenger market, tens of thousands of workers, cargo and logistics networks, airline facilities, road connections, air-traffic-control systems, security and border operations, and decades of accumulated investment. Moving it would mean recreating or migrating much of that system elsewhere.

The formal Dutch aviation studies reviewed here mostly debate traffic levels, noise, environmental conditions, investment, or supplementary-airport use while holding Schiphol’s location itself fixed. Schiphol is where Schiphol is. Hemel’s provocation matters because it relaxes that assumption. Suppose the Netherlands closed flight operations at Haarlemmermeer, built a replacement hub in Flevoland, and redeveloped the old site - would the alternative use of Schiphol’s land, relief from current noise and spatial restrictions, and any genuinely additional benefits in Flevoland be worth more than the replacement airport and transport system, changed passenger and worker access, transition costs, and environmental burdens created elsewhere?

Based on the evidence currently available, a current-scale Schiphol-to-Flevoland relocation has not been shown to create more Dutch social value than retaining Schiphol. That is not the same as showing that relocation would fail. The public record reviewed for this article contains no integrated, decision-grade project specifying a Schiphol-function Flevoland replacement, the transport network needed to serve it, and whole-site Schiphol redevelopment together. Several quantities that could decide the result therefore remain unavailable at decision-grade resolution.

One easy argument against moving is also too strong. An airport hub is not physically welded to its runways; international experience shows that a functioning connecting network can migrate with a coordinated airport replacement. What remains location-dependent is the Dutch value around that hub: passenger and worker access, local demand, logistics, airline economics, and its relationship with the rest of the Randstad. Operational portability is not welfare portability.

Even if a serious appraisal eventually found relocation not worthwhile, that would not make Schiphol’s present location costless or socially optimal. Keeping the airport at Haarlemmermeer reserves a large, highly connected site in the middle of the Randstad (and some surrounding spatial capacity) for aviation. The opportunity cost is real in mechanism, unquantified in magnitude, and largely held fixed in the formal aviation studies reviewed here.

Lelystad is not a spare Schiphol

Relocation would not mainly involve making the existing Lelystad Airport bigger.

Schiphol handled 68.8 million passengers and 477,552 air transport movements in 2025. Its final traffic review reported connections to 300 destinations and about 1.43 million tonnes of cargo.

The present Lelystad passenger facility belongs to a different category. Its terminal is about 12,000 square metres, with four aircraft stands, and the commercial concept was built around roughly 1.5 million passengers and around 10,000 annual commercial movements. Even this much smaller operation has not yet opened for commercial traffic.

The Dutch government said in April 2026 that it was preparing for 10,000 commercial flights a year from October 2027. That remains a policy trajectory rather than an unconditional opening date. The airport decision must be amended, environmental work remains necessary, and commercial opening requires a nature permit.

A replacement Schiphol would therefore not be today’s Lelystad Airport with a few more gates. It would be a new national airport and transport system in the same broad part of the country.

The scale mismatch clarifies what should actually be compared. There are four useful cases:

• Scenario A, retain and reinvest Schiphol: flight operations remain at Haarlemmermeer, while aviation volume, aircraft, infrastructure, environmental policy, and the network can continue to change.
• Scenario B, full relocation with a broadly comparable hub: Schiphol closes for aviation and a principal Dutch hub is constructed in Flevoland, with a serious attempt to migrate the anchor airline, connecting banks, alliance relationships, baggage, cargo, maintenance, and other hub functions together.
• Scenario C, relocation with deliberate aviation reduction: Schiphol closes, but the replacement is materially smaller than today’s system.
• Scenario D, a materially smaller retained Schiphol: aviation contracts while the airport stays at Haarlemmermeer.

Scenario D matters because Scenario C changes two things at once. It moves the airport and reduces flying. Any benefit that comes merely from fewer flights should not be credited to relocation if the Netherlands could obtain the same benefit while leaving the airport where it is.

Scenario A also cannot be defined as “today’s Schiphol forever.” Even its current legal starting point is less tidy than a headline movement number suggests. A May 2025 amendment to the Luchthavenverkeerbesluit Schiphol (LVB), the airport traffic decree governing Schiphol operations, introduced a ceiling of 478,000 commercial movements per use year and 27,000 at night. The Administrative Jurisdiction Division of the Council of State annulled that amendment on 11 March 2026.

At the 22 September 2026 research cutoff, the previous LVB framework was therefore back in force without a statutory total annual movement ceiling. A provisional 27,000 commercial night-movement limit remained by court order. A broader replacement LVB, which proposed an initial 478,000 movements and a conditional route toward 500,000, was still a draft under procedure.

A separate Natura 2000 authorization track was also unsettled. The District Court of The Hague annulled Schiphol’s 2023 nature permit on 4 June 2025. The appeal did not restore the permit, and the Council of State denied an interim request to suspend the judgment. On 19 December 2025, the competent authority issued a temporary toleration decision. At the research cutoff, Schiphol was therefore operating without a valid nature permit under temporary toleration. Toleration is not a permit, and this legal track is distinct from the LVB.

A counterfactual needs a real baseline. “Keep Schiphol” does not mean freezing one particular movement ceiling, investment plan, or network structure; the airport itself is already a moving target.

What Schiphol’s location is worth to aviation

The same centrality that makes Schiphol’s land attractive for redevelopment also makes the airport unusually accessible.

Schiphol Airport station sits directly beneath the terminal. NS advertises Amsterdam Centraal at about 17 minutes and eight Airport Sprinters per hour on the airport service. Intercity Direct connects Schiphol with Rotterdam Centraal in roughly 26 minutes. The airport is also embedded in the western Randstad motorway system.

That is not a decorative feature of the airport. Ground access is part of the core product.

A 2024 survey by the Netherlands Institute for Transport Policy Analysis, KiM, found that among Dutch residents aged 18 to 80 departing from Schiphol on their most recent air journey, about 34 percent arrived by public transport. Another 29 percent were dropped off by car, 27 percent parked, and 10 percent used a taxi or closed bus service. Those figures do not describe every Schiphol passenger, and they say nothing directly about transfer passengers or employees, but they show why moving the terminal changes more than an aircraft’s point of departure.

Workers are another part of the geography. Older CBS-based evidence used in a Schiphol economic study found that 16 percent of workers in the airport cluster lived in Amsterdam, 12 percent in Haarlemmermeer, and 7 percent in Haarlem, with roughly 70 percent living in the top 20 origin municipalities. Those data describe the 2017 workforce and are too old to forecast a future one, but they show that the airport workforce was geographically concentrated in the current western-Randstad catchment.

Move the airport east and some people would gain. Lelystad, Almere, and parts of the north and east could become better connected to the national airport. Many travelers and workers in Haarlemmermeer, Haarlem, Leiden, The Hague, Rotterdam, and parts of Amsterdam would probably lose unless new transport connections compensated. Utrecht and Amsterdam are especially sensitive to the exact replacement site, rail alignment, stopping pattern, and interchange design.

There is no defensible national average travel-time effect yet. To produce one would require a defined replacement site, a defined road and rail network, origin weights, timetables, and separate treatment of passengers and workers. An arrow pointing from Schiphol to Flevoland on a map cannot do that work.

The air network creates a second kind of accessibility.

Schiphol’s preliminary 2025 traffic split reported 43.6 million origin-and-destination passenger movements and 25.2 million transfer passenger movements. The transfer number demonstrates the scale of Schiphol’s connecting function, but it is easy to misuse. A transfer traveler is counted on the arriving and departing legs, so 25.2 million is not 25.2 million unique foreign people passing through the terminal. More importantly, transfer volume is not a measure of Dutch social welfare.

Transfer passengers can nevertheless matter to Dutch travelers. Their fares can help support routes and frequencies that would not be viable on local demand alone. More passengers feeding a route can support additional departures, better connection banks, or service to a destination that Dutch origin passengers also use.

Dutch guidance for aviation social cost-benefit analysis handles much of this through generalized journey costs and user surplus. Frequency matters because waiting for a badly timed departure is costly. Fares matter quite a lot. Access time, flight time, transfer time, reliability, and the availability of useful destinations also matter. Airline and airport profits belong in producer surplus, not in a separate pile of gross revenue. Wider economic effects count only where they are genuinely additional rather than another name for benefits already experienced by travelers and firms.

A destination count is therefore descriptive rather than an economic valuation. Two airports could serve the same number of destinations while offering very different frequencies, prices, departure times, connections, and reliability.

Cargo requires similar discipline. KiM estimates that in 2022 about 64 percent of Schiphol air-freight tonnage moved on full freighters and 36 percent in passenger aircraft bellies. That makes passenger and freight networks partly complementary, but tonnes of freight are not tonnes of welfare. What matters is what relocation does to shipper costs, handling, reliability, trucking distances, airline economics, and the ability to substitute through other European airports.

At this point the obvious objection to relocation is that all these network relationships took decades to build and would collapse if Schiphol closed. That goes too far. A hub can move. The harder question is how much of the value attached to the hub moves with it.

What Schiphol’s location prevents

If Schiphol’s location is productive infrastructure, it is also an allocation of scarce land.

Schiphol says its entire airport site covers 2,787 hectares, or 27.87 square kilometres. That is the safest current figure for the direct site. A 2019 government quick scan of an offshore-airport concept used a figure of “almost 3,100 hectares” for the Schipholterrein, but did not define the boundary in a way that lets the two numbers be reconciled. Neither number is a measure of land that could actually be built on after closure.

Runways do not turn into serviced housing plots when the last aircraft departs.

Redeveloping the site could require removal or adaptation of extensive pavement, soil and contamination work, water and drainage reconstruction, new streets and utilities, grid capacity, schools, parks, climate adaptation, and years of phasing. No whole-site masterplan currently establishes the net developable share, the number of homes, the redevelopment cost, the absorption rate, or the resulting net social value.

The evidence therefore supports a large physical opportunity, not a housing number.

The airport’s spatial effects also extend outside its perimeter. The Luchthavenindelingbesluit Schiphol (LIB) is the airport spatial-planning decree. It establishes differentiated areas around Schiphol in which noise, external safety, building height, obstacles, and other flight-safety considerations can restrict development. The current consolidated LIB remained in force at the research cutoff.

These restrictions can matter in practice. In Rijsenhout, Haarlemmermeer has pursued plans involving roughly 400 homes in an area affected by LIB restrictions and the need for special approval. But the same case also illustrates why the wider “land unlocked by closing Schiphol” cannot simply be read from a map. Industrial noise and other planning constraints also affect the area.

A legal restriction is not the same thing as a binding economic constraint. Two restriction zones can overlap. A parcel can face an aviation rule and a water constraint at the same time. Existing development may remain. Exceptions may already be possible. Grid limitations, nature protection, infrastructure, ownership, flood risk, or municipal policy may prevent development even after an aviation restriction disappears.

A credible estimate of wider spatial release would therefore require parcel-level or GIS analysis asking which constraint actually binds in each location. That work has not been done at the aggregate scale needed here.

Still, “not quantified” does not mean “zero.” At least some aviation restrictions demonstrably constrain what surrounding municipalities can build. Closing the airport would change those constraints. The opportunity cost is real even though its total area and value are not known.

The same asymmetry appears in the airport’s environmental burden.

Government material for Schiphol’s Balanced Approach procedure used modeled reference quantities of 6,962 dwellings within the 58 dB(A) Lden contour and 111,955 seriously annoyed people within the 48 dB(A) Lden contour. It also modeled 5,750 dwellings within the 48 dB(A) Lnight contour and 24,502 seriously sleep-disturbed people within 40 dB(A) Lnight. These are policy-reference model outputs, not observed counts of individual victims or a simple statement of actual 2025 exposure.

RIVM separately used 2024 exposure and survey data to update airport-specific relationships between aircraft noise, serious annoyance, and sleep disturbance. The important point is not to collapse different kinds of evidence into one giant harm number. It is that current noise around Schiphol is measurable and location-specific.

Local air pollution and ultrafine particles add another burden, although the health evidence requires more care. RIVM’s work supports concern about possible long-term health effects from aircraft ultrafine particles, while retaining uncertainty about causal effects for particular outcomes. RIVM has also cautioned against pretending aircraft noise, local pollution, and ultrafine particles can already be compressed into a single robust cumulative health-risk value.

Relocation would remove these airport sources from Haarlemmermeer, but it would not abolish them. Aircraft, road traffic, ground operations, noise, and safety zones would reappear around the replacement airport. Schiphol imposes real costs because of where it is; a new airport would impose many of the same types of costs because of where it would be. The relevant benefit is the difference.

The bill for moving east

Once Lelystad is treated as the location of a possible replacement rather than an airport already waiting to take over, the physical problem becomes much larger.

A national hub needs a runway and taxiway system designed for peak connection banks and Dutch wind conditions; terminals, baggage, stands, cargo, and maintenance; security, border, and emergency services; air-traffic control and redesigned airspace; fuel, utilities, and environmental mitigation; and roads and high-capacity public transport for passengers and workers. It also needs operational testing and a transition plan linking the old and new airports.

None of this establishes that such a system cannot be built. It establishes what must be costed.

There is currently no defensible euro estimate for a Schiphol-function replacement in Flevoland because there is no defined project from which to estimate quantities. How many runways? With what orientation and spacing? Where exactly? What terminal capacity? What rail alignment? Which roads? What peak-hour assumptions? Which utilities already exist at adequate capacity? Which communities and protected areas fall under the routes?

Without those choices, a capital-cost total would mainly give uncertainty a currency symbol.

The Netherlands has examined other airport-location ideas. The 2019 Quickscan luchthaven in zee, for example, considered moving flight operations to a five-runway artificial island while retaining important landside functions at Schiphol and connecting the two by tunnel and rail. The government discussed an island-and-tunnel cost range of €33 billion to €46 billion for that concept and acknowledged important omitted costs and schedule risks.

That is useful evidence about how many systems a complete relocation can implicate. It is not a price quote for Flevoland. Transplanting the figure would replace one missing estimate with the wrong estimate.

Flevoland also cannot be treated as empty land on which engineering is the only constraint.

The government’s own April 2026 programme for just 10,000 commercial flights at Lelystad says an amended airport decision and nature permit are still required. That does not tell us whether a future hub-scale project could obtain permission. It establishes that the environmental and legal baseline is already constrained at a much smaller scale.

Nitrogen law has also become more demanding. In January 2026, the Council of State’s Pasgeld-West judgment extended its changed treatment of intern salderen, or internal nitrogen offsetting, to spatial plans. In the preliminary screening, the effects of new development must be assessed on their own when deciding whether significant Natura 2000 effects can be excluded. Existing lawful activity cannot simply be netted against the new project at that stage to avoid a fuller assessment. Internal offsetting can still play a role later as mitigation if the legal conditions are met.

That judgment did not concern Lelystad Airport. It certainly did not decide that a Flevoland hub is prohibited. Its relevance is narrower: a large airport and associated infrastructure would need project-specific nature assessment under that screening-and-mitigation sequence.

Military aviation adds another unresolved design problem. Current government policy includes preparation for 2,300 fighter sorties at Lelystad. That number does not prove that military and hub-scale civil aviation are incompatible, nor that they can comfortably coexist. Compatibility depends on runway capacity, routes, airspace, separation, scheduling, noise, and the eventual operational design.

Bird risk, Natura 2000 areas, construction impacts, nitrogen deposition, new road and rail traffic, and route geometry introduce similar questions. Existing Lelystad studies show that environmental effects can depend substantially on which flight routes are used. A Schiphol-scale replacement would require new analysis rather than multiplying the environmental impacts of a 10,000-flight airport.

Flevoland is constrained but unresolved.

The transition would be a project in its own right. Airlines and aircraft would have to move with crews, equipment, maintenance, cargo handling, baggage and IT systems, border and security functions, emergency services, and air-traffic-control procedures. Road and rail connections need to be ready when the replacement airport needs them. The old airport cannot be fully redeveloped while it is still needed to maintain national aviation operations.

Royal Schiphol Group’s current investment programme adds one more tempting but misleading number. Schiphol reported €1.056 billion of investment in 2025 and roughly €10 billion planned through 2035, including overdue maintenance, renewal, and improvement.

A relocation decision would not “save €10 billion.” Money already spent is sunk. Maintenance required to keep Schiphol safe during a long transition would still be required. Some investments might retain value in redevelopment. Other airport-specific projects could genuinely be avoided if a closure decision came early enough.

The correct number is the future expenditure that becomes unnecessary because of relocation. The public record reviewed for this article contains no decision-grade decomposition of the €10 billion programme into sunk, committed, safety-critical, reusable, transition-required, and avoidable components sufficient to determine that amount.

The absence of a replacement cost estimate is therefore not an embarrassing hole that can be filled with a better Google search - it is a consequence of not yet having a project.

Could the hub actually move?

One of the strongest arguments against relocation is that even if the Netherlands can build a new airport, it cannot rebuild the hub.

There is real evidence behind the concern. When an airport actually loses its anchor airline or connecting network, recovery can be difficult.

A study by Kerry Tan and Andrew Samuel examined seven US legacy-airline dehubbing events between 1993 and 2009 and found heterogeneous fare effects that depended heavily on the competitive response. A broader international study of 37 dehubbing events found that seat capacity often remained below pre-event levels for years and recovery as a hub was uncommon.

That evidence establishes a serious risk if the network actually breaks. It does not tell us what happens when a government, airport, and anchor carrier deliberately move the hub together.

Dehubbing and hub relocation are different treatments.

Istanbul provides the clearest modern correction. In 2019, Turkish Airlines transferred its principal operation from Atatürk to the new Istanbul Airport. ACI Europe’s schedule-based analysis subsequently found direct and hub connectivity at the new airport above the prior-year Atatürk comparison in its June snapshot. Turkish Airlines’ November 2019 traffic release also reported January-November international-to-international transfer passengers 4.6 percent above the same period of 2018. That carrier-reported comparison demonstrates continued transfer activity after the move; it is not a causal no-move counterfactual.

Istanbul is not a Dutch forecast. Turkey has a much larger domestic market, the move came with major capacity expansion, and the airline, regulatory, governance, and demand conditions differ. The example establishes something narrower but important: physically replacing an airport does not inherently destroy a connecting network.

Munich supplies a second mechanism. When the city’s airport moved in 1992, years of preparation culminated in an operational transfer of roughly 16 hours. That does not mean transition was free. It shows that building a replacement for years does not require airlines to operate a fragmented hub across two airports for those same years.

A serious Schiphol relocation scenario should therefore assume that the KLM-centered hub, alliance relationships, connecting banks, baggage system, and other airside functions make a coordinated attempt to migrate. Permanent network erosion should then be treated as a possible outcome, not as an assumption smuggled into the starting point.

Some parts of the network are highly portable. Aircraft can move. Schedules can be redesigned. Alliance relationships, crew planning, commercial agreements, connecting-bank structures, and many maintenance or cargo functions are not made of Haarlemmermeer soil.

But an airline also needs passengers.

A traveler beginning a journey in The Hague who faces a substantially longer trip to the airport is not made whole because KLM preserves the same flight number. If moving east reduces the attractiveness of the hub to the large origin-and-destination market around Amsterdam and the western Randstad, the airline may face weaker local demand. Higher operating costs, different airport charges, altered competition, or less attractive schedules could then reduce frequencies or routes even after a technically successful migration.

That distinction matters because transfer traffic and local traffic support the same network. A foreign connecting passenger may not contribute the same Dutch consumer surplus as a Dutch traveler beginning a journey at Schiphol, but the foreign passenger can help make the Dutch traveler’s flight commercially viable.

Ground access and hub economics therefore interact. A rail line that merely replaces some lost airport access has one kind of benefit. A line that also creates useful new east-west journeys for non-airport users has another. Those independent transport benefits could strengthen the relocation case, but only to the extent they are genuinely additional. The analysis cannot charge the entire railway to the airport and then credit the project with every wider rail benefit, or do the reverse.

Cargo has the same location problem in a different form. Airside capacity can move, and European airports can substitute for some cargo activity. But warehouse locations, trucking distances, labor availability, handling reliability, and the relationship between passenger belly capacity and freighters all respond to geography.

The most difficult part of the hub question remains proprietary. Public data do not reveal KLM’s route contribution margins or its equilibrium response to a particular Flevoland airport with a particular rail network, charge structure, movement allocation, and competitive environment. That response cannot be recovered from a destination count.

The research changes an important part of the argument without settling it. “A hub cannot move” is too strong. “A hub can move” is true in a narrower operational sense, but that does not establish that Dutch passengers, workers, shippers, airlines, or taxpayers would retain the same value after it moves.

The comparator that changes the question

The relocation debate changes once the objective changes from “reproduce Schiphol somewhere else” to “operate less aviation.” Suppose fewer flights reduce aircraft noise around Haarlemmermeer, local pollution, climate emissions, and pressure on airport infrastructure. If the same reductions can be obtained by shrinking Schiphol where it is, those are not benefits of moving the airport. They are benefits of contraction. Scenario D is designed to separate the two, and Dutch research already provides some evidence on lower-volume Schiphol scenarios, although not enough to identify an ideal size.

CE Delft’s study for the Municipality of Aalsmeer compared cases with 375,000, 500,000, and 540,000 movements. The study is especially instructive because it had to be rectified in 2022. The correction removed double counting of climate costs that were already reflected in ticket prices under the aviation appraisal method and fixed an error in the population calculation inside the 48 dB noise contour. The corrected results changed materially, particularly for the shrinkage alternative.

That is not a reason to ignore the study. It is a useful demonstration of how sensitive the answer can be to background demand, climate valuation, hub assumptions, movement allocation, and whether Lelystad is operating. It does not establish that 375,000 movements is the correct target.

A newer IenW-commissioned analysis by SEO compared a 460,000-movement alternative with a 500,000 reference. The report explicitly says it is not a new integral social cost-benefit analysis; it scales and updates earlier work to examine orders of magnitude and mechanisms. Under its scenarios, the modeled effect on destination breadth was limited and the researchers did not identify an immediate threat to continuation of the hub role at 460,000 movements.

That supports a modest proposition: some contraction scenarios can coexist with a functioning hub.

A second CE Delft report extends the comparison to deeper contraction. Brede welvaart en de omvang van Schiphol, commissioned by the Maatschappelijke Raad Schiphol, is dated December 2025 and was published by CE Delft in February 2026. It compares constructed steady states at 500,000, 375,000, and 250,000 flights, with a full hub assumed at 500,000, a slimmer hub at 375,000, and no hub at 250,000. It reports lower direct and indirect connectivity as volume falls while arguing that substantial international accessibility can remain in the lower-volume scenarios.

The study is exploratory, not a traditional MKBA. It produces no integrated welfare balance, performs no new case-specific model runs for the three scenarios, and does not model the transition path. Its hub configurations are partly scenario assumptions, not empirically discovered tipping points. Taken with the 460,000 analysis, it strengthens the case for treating deeper contraction seriously within Scenario D, but it establishes neither an optimal flight volume nor a safe minimum.

The way airlines allocate fewer movements matters. They can preserve destinations while cutting frequency. A city served twice a day rather than four times a day still appears on the destination map, but the service is less useful. Schedule delay rises, some same-day business trips become harder, connection banks weaken, and passengers have fewer alternatives when disruption occurs. An airline could instead drop thinner destinations to preserve frequency on stronger routes.

There is no verified Schiphol-wide tipping point at which these adaptations suddenly become “hub collapse.” Route-level economics and minimum-frequency constraints can matter to airline decisions, but the available evidence does not provide a general response curve linking total annual movements to national network value.

Scenario D is therefore not a disguised recommendation for a smaller Schiphol. It is a control.

Scenario D matters most when Scenario C is considered. A smaller airport in Flevoland might need fewer gates, stands, and terminal capacity than a current-scale replacement. Yet it still requires much of the fixed institutional and physical platform: runways, air-traffic control, security, border processing, utilities, road and rail access, emergency services, environmental approval, transition, and decommissioning at the old site.

If the long-run aviation system is materially smaller, the relevant comparison is no longer a large Schiphol against a large new hub. It becomes a smaller Schiphol where it is against paying the fixed cost of moving a smaller system elsewhere.

That could strengthen either side of the relocation balance. A smaller replacement could reduce construction requirements and new externalities. But it also makes the fixed costs of relocation harder to spread over the aviation activity that remains.

The same logic applies to climate. At broadly comparable traffic levels, changing the location of the runways does not eliminate aviation’s climate emissions. A new airport also creates construction emissions and changes passenger and worker surface travel. Climate gains generated by fewer flights belong to both Scenarios C and D, not uniquely to relocation.

Noise works similarly. Closing Schiphol would eliminate aircraft noise at Haarlemmermeer, which is a location benefit. Reducing the number or pattern of flights can also reduce noise without closing the site, which is a contraction benefit. A replacement hub would meanwhile create a new noise geography in Flevoland.

Separating those mechanisms prevents an attractive project from receiving credit twice for the same change.

What the balance sheet can and cannot say

The welfare calculation is straightforward to state: a relocation creates value if the net benefits from reusing the old site, releasing other spatial constraints, reducing existing externalities, and producing any genuinely additional regional or transport gains exceed the resource cost of the new airport and transport system, conversion of the old site, transition, new externalities, and any persistent loss of passenger, worker, cargo, or network accessibility.

The difficulty is that several of its largest terms are still blank.

Potential gains

A central positive unknown is the alternative use of the current Schiphol site.

A 2,787-hectare site at a major Randstad transport node plainly has uses other than aviation. The existing railway and motorway connections make the land interesting precisely because they also make the current airport valuable. If a redevelopment plan could create large amounts of housing, employment, public space, and other urban uses after paying for demolition, remediation, water management, utilities, streets, public facilities, and long phasing, the net benefit could be substantial.

But the word net is carrying a lot of heft in the previous sentence. There is no defensible estimate yet of the developable share, the eventual mix of uses, the redevelopment cost, or the social value after those costs.

Some surrounding land could also gain development options as LIB restrictions lose their airport purpose. Rijsenhout demonstrates that this mechanism is real. No defensible aggregate exists for the number of hectares affected, because an aviation restriction is not always the binding constraint.

Relocation would remove existing aircraft noise and airport-related local pollution and external-safety burdens from Haarlemmermeer. The benefit is the burden avoided there minus the burden created at the new location, plus any difference caused by the distribution of population and activity. Present Schiphol exposure can be measured. A future Flevoland exposure pattern cannot be measured until there are runways and routes to model.

New transport infrastructure could also produce benefits unrelated to airport users. A new east-west rail connection might improve journeys within the Netherlands even for people who never board an aircraft. If those benefits survive a proper transport appraisal, they belong in the balance. They should not be assumed merely because a line is new.

Flevoland could gain employment, infrastructure, and accessibility. Some of that might increase national productivity by correcting genuine labor-market or agglomeration frictions. Some would simply redistribute activity from the western Randstad. Regional gains can matter greatly to the people and governments involved without becoming an additional national welfare benefit.

Finally, some future Schiphol investment could become unnecessary. The relevant amount is the genuinely avoidable airport-specific spending, not the entire €10 billion programme.

Potential costs

The most obvious negative unknown is the replacement system itself.

The airport, surface transport, utilities, environmental mitigation, and transition have not been designed to a level at which a reliable lifecycle resource cost can be estimated. That could be the largest single cost in the appraisal.

Moving the hub also changes access. Some eastern and northern travelers gain, while many western and southern users probably lose unless new transport compensates. The national effect depends on how many travelers start where, what travel time means to them, what the new network looks like, and how airlines react to changed local demand.

Workers face a related problem. An airside job that follows the airport is not a “job saved” or a new national job merely because it appears in Flevoland. The worker may now have a much longer commute, change residence, leave the employer, or be replaced. Firms incur recruitment and matching costs. New Flevoland workers may gain opportunities. National welfare depends on those adjustments, not on adding up job announcements on either side of the map.

Transition creates temporary costs even if steady-state hub performance eventually recovers. Equipment and systems have to move. Staff must be trained. Operations have to be tested. Baggage, cargo, security, border control, fuel, maintenance, air-traffic control, and surface transport all have to work when the switch occurs. Disruption during this phase should not be confused with permanent network damage, but it is not free.

Persistent network erosion is possible as well. It should enter the appraisal only through a mechanism: weaker local demand, higher airline cost, poorer ground access, altered competition, reduced frequency, or some other reason why the post-move equilibrium is worse. Dehubbing studies cannot simply supply a percentage loss for this purpose.

The new location creates environmental costs of its own. Noise, local pollution, safety contours, nature effects, nitrogen deposition, construction emissions, road traffic, and ecological disturbance do not disappear when the airport changes postcode. Their scale cannot be known until the project is defined.

Then there is project risk. Building a national airport and transport network before the old site can be redeveloped creates large, partly irreversible commitments. Aviation demand, aircraft technology, climate policy, the hub network, and the housing market could look different by the time both halves of the project are complete.

What is not a benefit merely because it is large

Several impressive numbers should not be added to either side of the welfare account.

Gross airport turnover is not social welfare. Gross cargo value is not social welfare. The number of Schiphol jobs is not the number of jobs the Netherlands loses if the airport moves. The number of jobs created in Flevoland is not automatically a national gain.

Transfer passengers matter through their effect on routes, frequencies, airline economics, and any Dutch producer surplus, not because each transfer movement carries a fixed Dutch benefit.

Destination count has similar limitations. A route served once daily and one served six times daily each contribute one destination to the total.

The €10 billion investment programme is not a relocation dividend. Only future spending genuinely skirted because of closure may be tallied as a true expense avoided.

And land-sale proceeds are not a second benefit on top of putting the site to more valuable use. If planning conversion makes public land more valuable, Royal Schiphol Group’s public ownership structure could make it easier for governments to capture some of that uplift and use it to finance infrastructure. That may be important for borrowing, taxes, timing, and political feasibility.

But selling a valuable asset does not manufacture another layer of social surplus. The welfare gain comes from the better use of resources, net of the cost of achieving it. Counting both the full redevelopment benefit and the full sale receipt would count much of the same change twice.

Future redevelopment receipts could nevertheless help finance relocation, but the replacement airport and required transport capacity must be financed and available before full closure, while old-site receipts can arise only after redevelopment can begin. That sequencing is a cash-flow and fiscal constraint, not a separate social welfare cost equal to the financing requirement.

Connectivity creates another double-counting hazard. If better frequency saves passengers time and that benefit is already included in generalized journey costs, a separate “connectivity premium” cannot be added for the same improvement. Peer-reviewed work in other settings has found causal effects of air service on local or regional economic activity, including Blonigen and Cristea (2015) and Campante and Yanagizawa-Drott (2018). Those findings do not automatically become additional Dutch national welfare in a Schiphol relocation appraisal. Any wider effect would still need to be shown to remain after direct passenger and firm benefits have already been counted.

A giant balance sheet assembled from every large number associated with an airport can look sophisticated while being economically meaningless.

The missing numbers are the result

A decision-grade analysis would need quantities that the current public record does not provide.

There is no estimate of the net developable share of the Schiphol site under a real masterplan; there is no aggregate estimate of wider land materially released from airport restrictions after other constraints are considered; there is no net redevelopment value after conversion costs and delay; there is no conceptual design and lifecycle cost for a Schiphol-function airport in Flevoland; there is no national demand-weighted estimate of the passenger and worker access change under a specified transport network; there is no reliable public estimate of KLM’s persistent network response to that specific airport.

There are no hub-scale Flevoland noise, nature, nitrogen, bird-risk, or external-safety contours, because there is no hub-scale design to model. Civil-military airspace compatibility remains a design question. Transition cost is unquantified. The avoidable share of Schiphol’s future investment programme is unknown, as is the socially preferred long-run aviation volume.

The list of necessary yet unavailable data inputs seems endless. These are not minor caveats surrounding an otherwise complete calculation. They include some of the terms most capable of changing its sign.

This creates an evidentiary asymmetry that favors neither conclusion automatically. Schiphol exists, so its infrastructure, passengers, travel times, noise contours, workforce, investments, and legal regime can be observed. A Flevoland replacement does not exist, so many of its costs are described as engineering requirements rather than invoices. But the same is true of the positive side: a new city on the Schiphol site does not exist either, so its net land value, housing capacity, infrastructure requirements, and wider spatial benefits are also hypothetical.

That asymmetry is the strongest case for keeping the relocation question open.

A real project could discover that the redevelopment dividend is much larger than current evidence can show. A new rail system might create substantial independent transport benefits. Detailed ecological work might find manageable constraints. A coordinated airline migration might preserve the hub more successfully than skeptics assume. The access penalty might prove smaller after a well-designed rail network is included.

A real project could just as easily discover the opposite: expensive remediation, limited developable land, severe transport costs, difficult environmental mitigation, weaker local demand, high airport charges, civil-military conflicts, or a transition programme that consumes much of the theoretical gain.

At the moment, both are hypotheses.

That is why the absence of a defensible net present value is not a failure of the analysis. Producing one anyway would be.

What the thought experiment reveals

On the present evidence, a current-scale relocation of Schiphol to Flevoland is not demonstrated to create more Dutch social value than retaining the airport at Haarlemmermeer. The replacement is not an enlarged version of today’s Lelystad Airport but a new national airport, transport network, utility platform, institutional system, and transition programme. Its cost and environmental feasibility are unresolved, while the value of the old site has not been specified well enough to show that the redevelopment dividend would outweigh those costs.

That is not the same conclusion as demonstrating that relocation would fail. A functioning hub can physically migrate, so fragmentation should not be assumed merely because the runways move. The harder risks are location-specific: Dutch passenger and worker access, local origin-and-destination demand, airline economics, cargo geography, the replacement system, and the environmental burdens transferred to Flevoland.

Nor would an unfavorable relocation appraisal establish that Schiphol’s current location is socially optimal. The airport’s 2,787-hectare site has an alternative-use opportunity cost, aviation rules restrict some development beyond it, and residents around Schiphol bear measurable noise and other local burdens. The strongest objection is that these opportunity costs may be real without being decisive: the airport itself may make unusually productive use of this central location; redevelopment may be expensive and slow; many surrounding restrictions may not be binding; and housing may be added elsewhere at lower social cost. That remains possible, but it can only be tested after the alternative use is specified seriously enough to compare with the airport.

If a future Netherlands wants materially less aviation, Scenario D changes the comparison again. The relevant question may be not “Schiphol versus a new Schiphol in Flevoland” but “a smaller Schiphol where it is versus the fixed cost of moving a smaller aviation system elsewhere.” Existing evidence suggests that some modest reductions in movements can coexist with continued hub operation. It does not identify the ideal number of flights, establish how deep contraction would affect the network, or settle the environmental and accessibility trade-offs. It is enough, however, to rule out a convenient false choice between recreating today’s Schiphol somewhere else and preserving today’s Schiphol indefinitely.

A serious relocation appraisal would therefore have to define both sides before claiming an answer: a conceptual Flevoland airport and transport design, a transition plan, a whole-site Schiphol redevelopment plan, wider spatial release, national passenger and worker accessibility, airline and cargo responses, hub-scale environmental effects, and a consistent aviation-volume assumption. It would also need to compare relocation at that volume with retaining Schiphol at the same or a meaningfully comparable volume. Only then could the two alternatives be valued without quietly changing the counterfactual halfway through the calculation.

Hemel’s proposal is consequential and should be taken seriously because it exposes something ordinary aviation policy tends to hide. Choosing to keep the airport at Haarlemmermeer is itself a spatial choice, even when no government explicitly makes it anew each year.

The current evidence is maddeningly inconclusive - it does not show that Schiphol is in the wrong place, nor does it prove that it is in the right place. As this review has outlined, the current evidence only demonstrates that treating Schiphol’s location as a fixed, costless fact leaves a material part of the policy question unmeasured.