Continental Transportation Systems

A continental transportation system is not a mode. It is the rail, road, inland-water and air networks that share a landmass, the transfer points between them, and the institutions that price, permit and assemble the right of way. The question that matters is rarely the one that gets asked. It is not which mode is best; it is: given a network that already exists, what is the return on completing a missing link, on adding capacity to one that exists, and on introducing a new mode — and against what comparison?

Three findings organise this brief. First, the unit costs on which every continental comparison rests are systematically scope-corrupted, and the corruption runs in a documented direction: qualifiers get dropped, price years vanish, and modes get swapped. Second, a corridor volume figure quoted without its subsidy is a measure of fiscal commitment rather than of commercial demand, and the transport literature has produced the number that proves it. Third — the brief's analytic spine — no new continental mode clears the bar because its promoters compare it against the incumbent as it is today, while the incumbent keeps improving along a cost curve nobody has characterised. The missing artefact is not a technology. It is a forward cost curve for intermodal rail and road haulage.

The unusual edge in this brief's dependency graph is deliberate. Continental economics are increasingly set at the transfer points rather than on the links — terminal handling, gauge breaks, border processing, crane cycles, dwell — and the transfer points are exactly where automation lands first. That is why this brief waits on Autonomous Supply Chains rather than on a civil-engineering result.

Sourcing, declared at the front. The research behind this brief ran with no web search, with Semantic Scholar returning HTTP 429 on every attempt, and with a provenance gate that refused Wikipedia on every attempt and every government, agency and multilateral domain. Crossref was the workhorse. Of the thirty-five sources assembled for this subject, two were read in full, thirteen are publisher-deposited abstracts, and twenty are verified bibliographic records whose text was never opened. No Channel Tunnel outturn cost, no Gotthard modal-shift outturn, no HS2 estimate history and no Chinese network figures appear below, because they live in agency documents this brief could not reach. Where a number is missing, the brief says so rather than supplying it from memory.

Established The most-quoted capital-cost band in rail has five qualifiers, and all five are routinely dropped. The source is Flyvbjerg, Bruzelius and van Wee in EJTIR 8(1), 2008. The deposited abstract reads, verbatim: capital costs per route-kilometre vary highly between projects, and “looking at European projects and excluding outliers, the total capital costs per route-kilometre (including stations and rolling stock) lie mainly between US$50-100 million (2002 prices).” Established Count them: urban rail — metro and light rail, not intercity high-speed rail; European projects, not a global sample; outliers excluded, which removes precisely the tails where the interesting projects live; including stations and rolling stock; and 2002 prices, roughly a quarter-century stale and almost never re-based. Established It is then compared against high-speed-rail figures that are usually infrastructure-only, in a different decade, in a different mode. Frontier That is a scope mismatch, not a rounding problem, and it inflates the apparent gap in whichever direction flatters the party quoting it.

Established Two hard facts anchor what actually drives the spread, and neither is nationality. The Taiwan high-speed line runs 345 km with 252 km of continuous viaduct — roughly three-quarters of the alignment elevated — reported in a Joint Rail Conference paper on its seismic-safety and passenger-comfort design. A per-kilometre figure for that line and a per-kilometre figure for a line built at grade across a plain are not the same quantity measured twice; they are two different quantities. Frontier Construction method moves the same variable almost as hard: a schedule and productivity comparison on an identical 5 km viaduct with 125 spans reports 732 days for full-span launching against 1,350 days for span-by-span, a factor approaching two on identical scope. Established Terrain class, structural fraction, construction method, land-acquisition regime and labour cost dominate the variance, and none of them appear in the headline number. Speculative Cross-country league tables of cost per kilometre are therefore artefacts of assembly rather than evidence about engineering.

Established The measured record on forecast accuracy is real, and its canonical citation is formally invalid. The safe, unretracted ancestor is Skamris and Flyvbjerg in Transportation Research Record 1518, 1996, on a sample of Danish bridge and tunnel projects: “Construction costs tend to be underestimated and those of traffic, overestimated,” with cost overruns of 50 to 100 per cent common and traffic forecasts off by 20 to 60 per cent. Established The paper that made the finding famous — Flyvbjerg, Skamris Holm and Buhl, Transport Reviews 26(1), 2006 — carries the title string “RETRACTED: Inaccuracy in Traffic Forecasts” in Crossref, with a retraction relation dated 21 February 2012 and a separate one-page notice by the journal's then editor. Frontier A retraction is not a finding of falsity, and the reason for this one could not be established here; Megaproject Governance carries the full treatment. Established What matters for continental planning is blunt: the most repeated statistic in transport appraisal travels on a formally withdrawn citation, and essentially nobody downstream checks.

Established The prescribed remedy is contested by two peer-reviewed results that point in opposite directions. Reference class forecasting — estimating from the outturn distribution of comparable past projects rather than from the bottom up — is presented in policy documents as settled. Established For: Park, EJTIR 21(2), 2021, across 107 major projects, reports average cost overrun falling from 38 per cent to 5 per cent after the United Kingdom's 2003 mandate, with the UK exceeding its targeted probability of on-budget completion by 12 per cent while the US underperformed by 17 per cent. Established Against: Themsen, Financial Accountability & Management, 2019, a longitudinal case study of a 23.6-billion-kroner Danish megaproject that applied the method and still failed, argues in the abstract's own words that reference class forecasting “does not lead to more accurate cost estimates.” Frontier Both stand. Frontier Non-Western evidence exists and is thinner than it should be: Liu and colleagues measured optimism bias by field research and interviews across more than thirty long bridges in China, using the Hangzhou Bay Bridge to show how overrun expectation, implicit benefits and operating duration move net present value. Speculative The honest position: one strong aggregate result in one jurisdiction, one documented failure in a specific large case, and disagreement even among advocates about whether the method works by de-biasing estimators or by raising the political cost of publishing a low number.

Established On modal economics itself — the cost, energy and time comparison between rail, road, inland water and air — this brief must be careful about what it is entitled to say. The reference external-cost comparisons for truck and rail freight exist and are verified to exist: Forkenbrock's 1998 University of Iowa study and Austin's 2019 chapter in U.S. Freight Rail Economics and Policy. Neither text was obtained. No per-tonne-kilometre external cost figure appears in this brief, by mode or otherwise. Established What can be stated is the shape of the problem. Krakutovski and colleagues put the definitional statement plainly: “External effects occur when production or consumption activities affect welfare without compensation.” Frontier That sentence has an underappreciated consequence: valuing an externality is a normative choice about whose welfare counts, not a measurement, and more data does not converge the estimates. Established The World Bank's treatment of rail freight in emerging economies encodes its own finding in its title — the volume is about how to regain modal share and reverse the declining trend — which is a statement that rail freight share has been falling in most economies outside the North American bulk-haul and the Chinese and Russian state-rail cases. Frontier And the current state of the art in modelling the road-rail split, Rosell, Codina and Montero's combined modal-split and traffic-assignment model, matters for a reason that is easy to miss: it shows how much the answer depends on network assignment rather than on mode-level cost comparison. Speculative An argument about which mode is cheaper per tonne-kilometre may be an argument about a number the assignment problem renders close to irrelevant.

Established The largest natural experiment in modal economics is North American freight deregulation, and its record is genuinely two-sided. The Staggers Rail Act, signed in 1980, removed most of the rate and route regulation under which North American freight railroads had operated for a century, and the four decades since are the longest continuous natural experiment in modal economics anywhere. Established A thirty-year assessment by Eakin, Bozzo, Meitzen and Schoech is emphatic that the act “has lived up to its promise, delivering early, substantially, and over a long period of time” — and was produced in the regulatory-economics consulting sphere, which is context a reader is entitled to before treating it as neutral. Established The rate-series analysis that would settle the trajectory independently, Dennis in Transportation Research Part E, 2001, is verified to exist and was not read here. Frontier The counterweight is a whole literature: the Journal of Law and Economics study of the act's effects on grain transportation exists because captive shippers — those with one railroad and no barge alternative — experienced deregulation differently from shippers with options. Speculative Both things are true at once. An aggregate productivity gain and a distributional loss concentrated on captive origins are not in conflict, and a continental network's political durability depends more on the second than its economics does on the first.

Established Where continental integration is actually blocked, the blockage is usually institutional rather than physical. Kotthaus, writing on the Trans-European Transport Network, describes an instrument supporting “seamless, safe, and sustainable mobility of persons and goods” that represents “a delicate balance between EU and national authorities” for infrastructure funding. Established The governance sentence is the finding: Europe's continental network is limited less by engineering than by the division of competence between the union and its member states. Established On the assembly side, Boardman, Moore and Vining state the corridor argument in one clause — “a single entity” must “assemble the required rights of way.” Frontier And the strongest recent evidence that assembly and operation are separable comes from Spain: an ex-post evaluation of open-access competition on Spanish high-speed rail reports that competition is associated with fare reductions “averaging 30%” and total high-speed traffic increasing “by more than 75%.” Frontier The state assembled the corridor; the market operated on it; the measured outcome improved. Speculative That is the most useful available synthesis on whether continental networks are inherently state projects: assembly and operation are different problems and should be answered separately.

Established The number that should accompany every Belt-and-Road volume figure is a subsidy, and it has been modelled. Feng, Zhang, Liu and Fan, applying a game-theoretic subsidy model to the China Railway Express under the Belt and Road Initiative, put the optimal subsidy at between USD 2,000 and USD 2,500 per forty-foot equivalent unit. That is a very large fraction of a container's total transport cost on that corridor. Frontier It converts the headline train counts into something else entirely: at that level, a corridor volume figure measures fiscal commitment, not commercial demand. Established The general lesson generalises past China. A corridor volume figure without its subsidy is not an economic fact. Established The physical constraint on the same corridor is documented separately — Abramović and Šugić analyse “different track gauges that cause loading limits and loading gauge issues” as the organising problem for China-EU container trains — and the two constraints interact, because a gauge break adds handling cost that the subsidy then has to cover. Speculative Longer treatment of the Eurasian corridors belongs to Intercontinental Rail Systems.

Established Finally, the oldest and least answered question in the subject. Fogel's 1964 Railroads and American Economic Growth made the counterfactual argument that the American railroad was not indispensable to nineteenth-century growth — that its social saving over the next-best alternative, canals and wagons, was small relative to national product. Established The monograph carries no DOI and was not read here; the specific social-saving percentage is therefore not printed. Frontier What survives regardless of the magnitude is the methodological demand: name the next-best alternative and price it. Speculative Sixty years and a Nobel Prize later, continental link proposals are still justified by development rhetoric that never performs that step.

Established The megawatt charger is not a battery problem. It is an arithmetic consequence of a labour regulation, and that is the cleanest way into heavy freight electrification. Frontier Measured Class 8 tractor consumption in real duty cycles clusters around 1.7 to 2.2 kWh per mile, on fleet demonstration data from small samples rather than from a standardised test. Established United States hours-of-service rules require a 30-minute break after eight cumulative hours of driving, inside an eleven-hour driving limit. Established Put the two together: at 2 kWh per mile, a megawatt delivered for the length of a mandated break is 500 kWh, or about 250 miles of range. That single line explains the entire megawatt-charging category. The power rating is set by how long a driver is legally required to stop, not by what a cell can absorb, and any charging standard rated materially below it forces the truck to stop for reasons the regulation did not require.

Established The connector standard exists and is specified well above what any vehicle currently draws. The Megawatt Charging System developed through CharIN and carried into SAE J3271 and parallel international work targets up to 1,250 V and 3,000 A, about 3.75 MW at the coupler. Frontier Ratification status and the pace of vehicle-side adoption are the moving parts, not the electrical target. Established Public money has followed the hardware: the United States Department of Energy has funded heavy-duty charging demonstration at the scale of tens of millions of dollars per programme. Speculative Read against this brief's spine, that is a familiar shape — the demonstration is not the constraint here either. A coupler rated at 3.75 MW is useless to a fleet that cannot get 7 MW of service to its yard, and the interesting part of this subject is on the other side of the meter.

Frontier Two numbers decide depot economics, and neither is the price of a battery. The first is interconnection: a twenty-truck depot charging simultaneously at 350 kW draws 7 MW, which is a small industrial plant rather than a commercial building, and the distribution upgrade and utility process behind a new multi-megawatt service is measured in years rather than months in most jurisdictions. Established The second is the demand charge. Commercial electricity tariffs bill a facility on its peak kilowatts as well as its kilowatt-hours, commonly in the range of ten to twenty-five dollars per kW per month, so a 1 MW peak is a fixed monthly bill of ten to twenty-five thousand dollars before a single kilowatt-hour is sold. Frontier That fixed cost divided by throughput is the whole of the economics: at low charger utilisation the effective cost per kWh can exceed diesel on an energy basis, and at high utilisation it falls well below it. A dollars-per-kWh figure quoted without its utilisation assumption is not an economic fact, in exactly the sense in which a corridor volume figure without its subsidy is not one.

Frontier The most important open result in this subject is one this brief cannot state, and the reason is instructive. Wallimann, von Arx and Hesse have a 2025 working paper applying causal identification to the question the modal-shift literature normally answers by assertion: did the Gotthard Base Tunnel actually take traffic off the parallel road? Established The Gotthard is the cleanest natural experiment in continental freight policy that exists — a very large rail investment, a defined Alpine corridor, an explicit federal modal-shift target, and a parallel road with measurable counts. Established The bibliographic record is verified; the text was not obtained; no result from it appears here. Speculative A field in which the single best-identified test of its central policy claim is a working paper that most practitioners have not read is a field whose confidence is not earned by its evidence.

Frontier Automated highways have a live and underrated engineering objection, and it is structural rather than computational. The deployment problem was formalised as far back as Hall and Tsao in 1997 and has not been solved since. Established The modern version, truck platooning, gets its benefit from short inter-truck gaps. Established Ling and colleagues, in Sensors 22(20), 2022, report that the “load effect of ATP on bridges increases with the increase in the gross vehicle mass and the truck platooning size but decreases with the increasing inter-truck spacing.” Frontier Read that carefully: the efficiency gain and the infrastructure liability are the same parameter. Closing the gap is what produces the saving and what concentrates axle load on spans that were not designed for it. Speculative The public-side retrofit cost of the existing span stock is the term nobody puts in the business case, and until someone does, platooning's net continental benefit is unsigned rather than positive.

Established The platooning benefit case has already migrated once, and the field's own reference volume says so out loud. Atasayar, Blass and Kaiser record that “the focus of the discussion changed over time,” with a shift in stakeholders: “while the initial intention was reduction in fuel costs, the claimed benefits have now diversified.” Frontier That is a documented instance of a pattern rather than a general suspicion — when a technology's original justification proves smaller than hoped, the justification set expands rather than the programme contracting. Frontier A companion chapter in the same volume adds the human-factors constraint that is missing from the aerodynamics: legislative readiness, infrastructure readiness, and “car drivers' subjective tolerance of gap sizes between trucks.” Speculative The gap that is aerodynamically optimal may be socially intolerable to the surrounding traffic, which makes the binding headway a regulatory and behavioural variable rather than a physical one.

Frontier Emissions comparisons in this literature are drifting in a specific and checkable way. Li's 2025 bi-level multi-objective optimisation of freight on high-speed rail reports that optimised services with carbon-tax policies can achieve a 56.97 per cent reduction in CO2 emissions — compared to air freight only, under an assumed tax regime, inside a model. Established Quoted without those three qualifiers it becomes a claim about rail against road that the paper does not make. Speculative The hard cases for modal shift — low rail share, short hauls, nowhere for a line-haul advantage to accumulate — are where the emissions arithmetic should be tested, and are where it is tested least.

Frontier Two live continental programmes are running experiments nobody designed. Ukraine's proposal to integrate its railway network at 1,435 mm gauge into the Trans-European Transport Network is the largest gauge-conversion programme proposed in Europe since the Iberian conversion, it is war-driven, and it is almost entirely absent from Anglophone discussion. Frontier It is also a revealed preference: decision-makers closest to the problem believe the physical standard is the binding constraint and is worth billions to change. Frontier Meanwhile the China-Europe rail service is being reshaped by geopolitics rather than by economics — Sun and Tomanek's terminal case study records that “geopolitical tensions and sanctions linked to the Russo-Ukrainian War have constrained its development.” Speculative Together these are the strongest available evidence that the resilience argument for redundant continental links is not merely rhetorical, and the weakest available evidence about how to price it.

Frontier And there is a software-layer answer with a real programme behind it. The H2020 MyCorridor project pursued Mobility-as-a-Service across a multimodal European cross-border corridor, aiming to “facilitate sustainable travel in urban and interurban areas and across borders by replacing private vehicle ownership by private vehicle use.” Speculative The strong version of the claim — that the next continental mode will be a software layer rather than a physical one — rests on the proposition that the utilisation gap on existing continental assets is large enough that better allocation substitutes for capacity. Established That is measurable from existing operating data and nobody has published the continental-scale version. Established The limit is also clear: software cannot relieve a gauge break, a bridge load limit, or a terminal's crane cycle time.

Frontier If road freight decarbonises at the tailpipe, what is left of the modal-shift case? This brief already holds the external-cost literature most rail-freight advocacy rests on, and a large share of the quantified externality in it is combustion. Electrify the tractor and that share does not vanish — generation still emits, tyre and brake wear still make particulates — but it shrinks a great deal, and the residual case for shifting freight to rail falls back onto congestion, pavement and structure damage, and land take. Speculative Those are real and they are smaller, which would make three continents' modal-shift policies substantially less valuable at the moment they are being funded. Handwave Nobody has recomputed the external-cost coefficients for an electrified road fleet and published the result, so every comparison in circulation assumes a diesel counterfactual that policy is trying to abolish.

Frontier Payload decides whether that arithmetic is even available. A 500-mile pack at 2 kWh per mile is 1,000 kWh; at pack-level densities near 150 Wh per kg that is roughly 6,700 kg of battery, against limits of 80,000 lb in the United States and 40 tonnes for most European combinations. Established Both jurisdictions grant zero-emission vehicles a weight allowance, and both allowances offset part of the penalty rather than all of it. Frontier They also push axle loads up, and pavement damage rises very steeply with axle load. Speculative Electrification trades a fuel externality for a pavement externality, and no corridor appraisal in this brief's evidence base prices both.

The workback plan below runs to a target most people in the field would call impossible: a genuinely new continental transport mode entering commercial service and displacing an incumbent on unsubsidised cost, rather than being sustained by state purchase. Eight results would have to land, and they are not equally hard.

Frontier (1) Establish whether large rail investments actually shift mode — a causal estimate of road-traffic reduction attributable to the Gotthard Base Tunnel against a credible counterfactual. In progress. Established (2) Separate generated from redistributed traffic by ex-post decomposition on at least three corridors with pre-opening controls; the method exists in the ex-post-evaluation literature and has never been executed at continental scale. Established (3) Price the gauge and standards break: a shadow price per container derived from dwell time, handling cost and reliability penalty. Not attempted, and tractable with existing corridor data. Frontier (4) Fix the external-cost vector: an agreed per-tonne-kilometre external cost by mode with uncertainty bounds. Contested, and partly normative rather than empirical. Established (5) Establish the incumbent's projected cost curve — forward cost per tonne-kilometre for intermodal rail and road freight to 2050, including automation and decarbonisation. This artefact does not exist. Established (6) Compute the crossover volume at which a candidate mode falls below that projected curve; requires (4) and (5). Frontier (7) Test right-of-way assembly at that volume: can a single entity assemble the corridor, at what cost, under what governance. Speculative (8) Demonstrate three years of audited unsubsidised operating surplus on a commercial segment. Never achieved by any post-1970 new mode.

Established Link 5 binds, and the discussion almost never goes there. It usually goes to 7 — land assembly, politics, eminent domain — because that is where the drama is. Frontier But the reason no new continental mode clears the bar is upstream of assembly: promoters compare their candidate to the incumbent as it stands today, while intermodal rail and road haulage keep improving along a curve nobody has published. Established Until a defensible forward cost curve exists, every crossover-volume calculation in this field is unfalsifiable, which is precisely why these arguments never resolve and why both sides can cite real numbers indefinitely. Speculative This is cheap, unglamorous, entirely feasible work — a decade of operating data, a stated automation and decarbonisation scenario, and published assumptions — and it gates everything else in the cluster.

Frontier Link 3 is the cheapest and the most neglected. A shadow price per container for a gauge break is computable from dwell time, transhipment handling cost and a reliability penalty, on corridors whose operating data already exists. Established It would decide whether Ukraine's conversion programme is rational, whether the Iberian conversion was, and what the Eurasian corridor's physical friction actually costs. Speculative Nobody in the Anglophone literature appears to have computed it, which is a statement about what the field finds interesting rather than about what is hard.

Frontier Corridor charging has the opposite bottleneck to depot charging, and confusing the two is the commonest error in this literature. Depot charging is an interconnection and tariff problem: trucks return nightly, utilisation is predictable, and the economics work as soon as the service exists. Corridor charging is a utilisation problem: a trunk-route site costs millions and earns nothing until enough megawatt-capable vehicles pass it, and those vehicles are not sold in volume until the sites exist. Established The joint ventures the incumbent manufacturers formed to build those networks are an admission that no single party will carry that gap alone. Speculative Their announced site counts are targets and should be read as such; this brief records no independently verified utilisation figure for any public heavy-duty charging site anywhere.

Established This brief waits on one result from another brief, and it is not a civil-engineering result. Autonomous Supply Chains holds the measured record on automation at the transfer points — warehouse, yard, terminal, line-haul handoff — and the continental question depends on it, because the cost of a continental movement accumulates disproportionately at the transfers rather than along the links. Frontier A corridor study of regional multimodality names its binding obstacles as “infrastructural constraints, inadequate handling equipment, organisational inefficiencies”: two of the three are not infrastructure at all. Speculative If terminal automation lowers handling cost and dwell materially, rail's structural disadvantage against road — terminal cost and schedule inflexibility, not line-haul price — shrinks without a metre of new track. Speculative If it does not, no amount of line-haul investment recovers the share, and the modal-shift policies of three continents are being argued on the wrong variable.

Established Three further dependencies run outward rather than in. Future Ports and Shipping carries the maritime leg and the terminal-automation evidence that the transfer-point argument leans on. Established Megaproject Governance carries the forecast-accuracy and retraction material in full, including what a retraction does and does not establish. Frontier High Speed Transit Networks carries the passenger-side density-and-distance economics that this brief treats only as one modal option among four, and Intercontinental Rail Systems carries the Eurasian corridor and the fixed-link cases whose subsidy and gauge arithmetic appear here only in summary.

Speculative One dependency has no brief and no owner: a published, contestable forward cost curve for the incumbent modes. Established It is not a scientific result, which is why no research funder owns it; it is not a regulatory filing, which is why no agency produces it; and it is not proprietary, which is why no carrier has reason to publish one. Speculative It is nonetheless the single artefact that would make the rest of this subject decidable, and its absence is better explained by incentives than by difficulty.

Established The Gotthard decomposition, done properly and published. Take the causal estimate of road-traffic change attributable to the tunnel and decompose it into traffic substituted from road and traffic generated by the improvement. The methodological template already exists in the ex-post review literature on mode substitution and induced demand following high-speed rail; it has never been applied at continental freight scale. Frontier The distinction decides both the emissions case and the economic case and is elided in nearly all advocacy.

Established The gauge-break shadow price. For a defined China-EU or Ukraine-EU corridor, measure the per-container cost of a gauge break as the sum of dwell time, transhipment handling cost and reliability penalty, with a stated distribution rather than a mean. Frontier Compare it against the amortised cost of conversion at the corridor's actual volume. Speculative That comparison is computable with data that already exists, it decides whether the Ukrainian conversion programme is rational, and nobody in the Anglophone literature appears to have computed it. It is the most tractable open question in this brief.

Established The bridge-loading retrofit census. For a national highway network, take the platooning load-effect relationship — load rises with gross vehicle mass and platoon size, falls with inter-truck spacing — and compute the fraction of the existing span stock that fails at the headways platooning requires for its claimed fuel saving. Frontier Publish the retrofit cost against the fuel saving at the same headway. Speculative If the retrofit cost exceeds the saving, automated highways are a capacity technology with a fuel penalty rather than an efficiency technology, which is a different policy object entirely.

Established The forward cost curve. Publish, with assumptions exposed and open to attack, cost per tonne-kilometre for intermodal rail and for road haulage to 2050 under stated automation, electrification and carbon-price scenarios. Speculative Then and only then compute a crossover volume for any candidate mode. Established Everything in section 4 that follows link 5 is blocked on this and on nothing else. Frontier The experiment is not technically demanding; it is institutionally awkward, because a published curve makes its author falsifiable and every party to the argument currently prefers being unfalsifiable. Speculative A funder that wanted to change this subject would commission the curve, not another mode study.

Established Electrification is where continental interoperability breaks in the least glamorous way. Gauge, loading gauge and electrification system differ across European and Eurasian networks, and the practical limits on a through-running continental train are set by that stack rather than by traction technology. Established The count of distinct electrification systems in Europe is frequently quoted and frequently wrong; this brief does not print a count, because it could not verify one. Frontier What is verified is the direction of decision-maker belief: the Ukrainian programme proposes converting to 1,435 mm rather than working around the break, which is a bet that the standard, not the rolling stock, is the constraint.

Established Structural fraction, not route length, drives civil cost. Taiwan's 345 km line with 252 km of continuous viaduct is the clean demonstration. Established And method matters nearly as much as geometry — 732 days against 1,350 for full-span launching versus span-by-span on an identical 5 km, 125-span viaduct. Speculative A continental programme that standardises structural type and launching method across thousands of kilometres is buying a schedule effect larger than most technology choices deliver, and it is the kind of gain that never appears in a mode comparison because it is not a mode.

Established The transfer point is an engineering object with a cycle time. Crane cycle, yard dwell, chassis availability and gate throughput set the intermodal penalty that a carbon price cannot touch. Frontier Handling equipment adequacy is named directly in the multimodality literature as a binding obstacle alongside organisational inefficiency. Speculative The engineering programme that would most change continental modal share is therefore not a new mode and not a new line; it is a terminal that transfers a container in a fraction of the current cycle at a fraction of the current cost. Speculative That is the same object Robotics in Infrastructure and Autonomous Supply Chains approach from the machine side.

Established The nearest neighbour is the automation of the transfer, not the automation of the vehicle. Autonomous Supply Chains and Future Ports and Shipping together hold the terminal and yard evidence that decides whether intermodal rail's terminal penalty is a fixed feature of the technology or a solvable engineering problem. Frontier Driverless line-haul trucking is adjacent in the other direction: it lowers the incumbent's forward cost curve, which raises the bar for every candidate new mode rather than lowering it. Speculative That inversion is worth stating plainly. Progress in road automation makes the case for new continental rail modes weaker, not stronger.

Established Energy Corridors shares this brief's central institutional problem exactly — one right-of-way assembly, many potential users, and a governance question about who holds the corridor. Frontier The multi-modal corridor hypothesis, that assembling one right of way for several infrastructures is cheaper than assembling several, has a clear mechanism and no empirical test; Northern Development Corridors treats it where it matters most.

Established Infrastructure Resilience holds the redundancy and disruption framing that the resilience argument in section 3 leans on. Established High Speed Transit Networks and Intercontinental Rail Systems carry the passenger and fixed-link cases respectively. Speculative And the appraisal literature is adjacent in an uncomfortable way: the study of how cost-overrun findings became canonical is a study of citation sociology, and it belongs in the toolkit of anyone assembling a continental business case from secondary sources. Frontier Inland waterways sit adjacent and largely unexamined here, because the comparative cost and energy data for barge against rail against road on a continental scale live in agency series this brief could not reach; the omission is a gap, not a judgement about the mode.

Established Every continental network in existence was assembled by a state. The United States interstate system, the Chinese high-speed network, the Trans-European Transport Network, the Soviet and Russian rail systems: all state-planned and state-financed. Established The reason is stated cleanly in the corridor literature — “a single entity” must “assemble the required rights of way” — and no market mechanism assembles a linear easement across thousands of holdouts. Frontier But assembly is not operation. The North American freight rail network is privately owned and operated and outperforms most state systems on freight productivity, and the Spanish open-access evidence — roughly 30 per cent average fare reduction and more than 75 per cent traffic growth — shows the split working in the same asset. Speculative The defensible institutional design that follows is: the state assembles and owns the corridor; competitive operators run on it; and the two decisions are made by different bodies on different criteria.

Established Competence, not capital, is Europe's binding constraint. The Trans-European Transport Network is described by its own legal commentary as “a delicate balance between EU and national authorities” on infrastructure funding. Frontier A continental network whose funding instrument is a negotiated balance between nineteen-plus sovereign budget authorities will complete links in the order that the balance permits, which is not the order that network value would dictate. Established Border management is a first-class corridor component in the World Bank's own corridor toolkit rather than an afterthought bolted to a transport plan. Speculative A brief that treats continental transport as a civil-engineering subject will systematically mis-rank its own priorities.

Established The subsidy question is institutional before it is economic. A modelled optimal subsidy of USD 2,000 to 2,500 per FEU on a flagship corridor is not a scandal; it is a policy choice with a stated price. Frontier What is defective is the reporting convention that publishes the volume and omits the price. Speculative The minimal institutional reform available to any continental programme — cheap, immediate, and resisted everywhere — is to require that every published corridor volume figure carry, in the same table, the direction split, the loading basis, and the subsidy per unit.

Established Europe has legislated megawatt charging into the continental corridor, which makes it the only place where this is a transport-network question rather than a fleet-procurement question. The alternative fuels infrastructure regulation obliges Member States to build heavy-duty recharging pools along the trans-European transport network on fixed spacing — as drafted, of the order of 1,400 kW at roughly 120 km intervals on the core network at the mid-decade milestone, tightening toward 3,600 kW at roughly 60 km by 2030. Established The heavy-duty carbon dioxide standards supply the demand side: 45% by 2030, 65% by 2035 and 90% by 2040 against a 2019 reference. Frontier North America has no equivalent corridor mandate, and its strongest instrument — a state fleet rule reaching drayage fleets — lost its federal waiver route in 2025. The corridor is the unit of analysis in Europe because a regulation made it one, and is not one in North America because nothing did.

Established Externality valuation is a distributional choice wearing a technical costume. If external effects are those that affect welfare without compensation, then pricing them requires deciding whose welfare enters the sum and at what weight. Frontier That is why decades of additional data have not converged the per-tonne-kilometre external cost estimates, and why a brief that promises convergence with better measurement is promising something the definition forbids.

Established Deregulation's aggregate gain and its distributional loss are both real. Captive shippers — a grain elevator with one railroad and no barge alternative — are the standing counter-case to the North American freight success story, and the sector literature exists because their experience differed. Speculative A continental network that optimises aggregate ton-mile cost will concentrate its losses on exactly the origins with the fewest alternatives, which are typically the least politically able to object and the most economically dependent on the link.

Frontier Connection is not convergence. The decade-long study of high-speed rail expansion in the Yangtze River Delta across forty cities found accessibility improving sharply — roughly a 50 per cent gain, with “closer intercity connections” — while the difference in urban intensity between cities in 2018 was “significantly greater” than in 2009. Central hubs concentrated the advantage. Speculative The development case routinely made for continental links — that connecting a periphery to a core develops the periphery — is the case this evidence most directly undercuts, and the mechanism is not mysterious: a faster link moves activity toward whichever end already has the agglomeration.

Established And the assembly that makes continental networks possible is a coercive act. A single entity assembling a right of way across thousands of holdings is exercising expropriation, and the ethical weight of that falls on people who receive compensation set by the expropriator. Speculative The corridor concept's efficiency argument — assemble once for many infrastructures — concentrates that coercion into fewer, larger acts, which is either a mitigation or an aggravation depending on whether consent costs are additive. Nobody has measured it.

Established The Fogel question is the civilizational question and it remains unanswered in the only form that matters. Whatever the true social saving of the American railroad over canals and wagons, the methodological demand stands: name the next-best alternative and price it. Speculative Continental transport is the domain where civilizations most reliably mistake large for indispensable, and the discipline of the counterfactual is the only defence.

Frontier The objective function may be wrong. The case for resilience over efficiency in continental networks has real evidence behind it — a pandemic demand shock and a war-driven rerouting of Eurasian freight both showed single-corridor dependence failing — and one fatal weakness: resilience is unpriceable ex ante, which makes it available as a justification for any redundant link anyone wants to build. Established A defensible price for corridor optionality does not exist. Speculative State that as unsolved rather than resolving it in either direction, because the alternative is either a network that fails under shock or a network that builds everything.

Speculative There is a real possibility that continental transport integration is finished as a frontier subject. Containerised intermodal freight, trunk aviation and a mature highway network already deliver near-universal continental reachability at declining real cost, and every proposed new mode either duplicates an existing function at higher capital intensity or occupies a niche too narrow to amortise a network. Frontier The evidence against that comfortable conclusion is that rail freight share has been falling for decades despite sustained policy effort, which is not what a satisfactory equilibrium looks like, and that the Gotthard exists because the previous system could not deliver an Alpine modal shift. Speculative Both readings are live. The measurement that separates them is the same one that link 5 demands.

These horizons track measurements and institutional decisions rather than demonstrations, because in continental transport the demonstrations have never been the constraint.

• 10 yr: Frontier The Gotthard causal estimate is published, replicated or contradicted, and becomes the reference result for whether large rail investments shift mode. Frontier Truck platooning operates on a growing but jurisdictionally patchy set of corridors at headways set by regulation and driver tolerance rather than by aerodynamics, and the bridge-retrofit liability is quantified for at least one national network. Speculative Ukraine's gauge conversion begins in earnest or is deferred, and either way generates the first serious corridor-level shadow price for a gauge break. Speculative The forward cost curve for intermodal freight is still unpublished, and continues to be the reason new-mode arguments do not resolve.
• 25 yr: Speculative Terminal automation, not line-haul technology, has moved intermodal rail's competitive position measurably in at least one continental market, and the movement is visible in modal share rather than in press releases. Speculative Corridor volume reporting conventions have improved — direction split, loading basis and subsidy per unit published together — because a major funder required it, not because the field reformed itself. Frontier Open-access operation on state-assembled corridors is the default institutional form for new continental passenger capacity, on the Spanish evidence. Speculative No new continental freight mode has entered unsubsidised commercial service.
• 50 yr: Speculative Continental networks are differentiated by terminal capability rather than by route length or top speed, and the countries that invested in transfer rather than in track are ahead on freight cost. Speculative Automated road freight has lowered the incumbent's cost curve enough that the crossover volume for any exotic mode has moved further out, not closer. Handwave A genuinely new mode carries meaningful continental freight somewhere, and it is a niche — high-value, time-critical, short-corridor — rather than a replacement for anything.
• 100 / 250+ yr: Handwave The distinction between a transport network and an energy network has eroded, with corridors assembled once and carrying whatever flows are worth carrying. Handwave The binding continental constraint is neither steel nor software but the right-of-way assembly problem, unchanged since the nineteenth century and still solved only by states. Handwave And the forward cost curve is still not published, because it never was anyone's job.

Frontier One freight-electrification line belongs in the ten-year window and it is a measurement, not a demonstration. Either a regulator or a utility publishes multi-megawatt interconnection lead times and charger utilisation as a series, and depot and corridor economics become arguable with numbers, or neither does and the subject stays a contest between manufacturer targets and advocacy modelling. Speculative The European corridor mandate resolves in the same window into built pools with measurable usage or a compliance exercise with idle hardware.

• Depends on This brief depends on Autonomous Supply Chains for a result it cannot produce itself, and the edge is unusual for a transport brief on purpose. The cost of a continental movement accumulates disproportionately at the transfer points rather than along the links: terminal handling, yard dwell, gauge breaks, border processing, crane cycles. Rail's structural disadvantage against road in most markets is terminal cost and schedule inflexibility, not line-haul price, which is why a carbon price does not close it and why a faster line-haul does not either. The transfer point is also where automation lands first, because it is enclosed, repetitive and already instrumented. So the question of whether continental modal share can move at all reduces, to a first approximation, to whether terminal and yard automation lowers handling cost and dwell materially — and that is the measured record another brief holds. If the answer there is negative or unproven, the modal-shift policies of three continents are being argued on the wrong variable.
• Requires (not on this map) Three inputs this brief needs and cannot produce, all of them disclosures or decisions rather than research results. First, a published interconnection lead time for multi-megawatt depot service: heavy freight electrification is gated by how long a utility takes to energise a seven-megawatt yard, and no jurisdiction publishes that interval as a series. Established Second, the forward cost curve for intermodal rail and road haulage that this brief names as missing, made worse by an electrified road cost that swings with charger utilisation. Third, corridor volume reporting that publishes direction split, loading basis and subsidy per unit together. Speculative None of the three is hard; each is simply nobody's job.
• Enables What this brief would enable, if its links landed, is narrower and more useful than new-mode advocacy suggests. A published forward cost curve for intermodal rail and road haulage would make every crossover-volume calculation in the cluster falsifiable for the first time, which would settle arguments about high-speed freight, vacuum-tube transport and automated highways that currently cannot be settled at all. A corridor-level shadow price for a gauge break would tell Ukraine, the Iberian peninsula and the Eurasian corridors whether conversion is rational, using data that already exists. And a bridge-retrofit census against platooning headways would convert automated highways from a contested efficiency claim into a costed capacity option, which is a thing a legislature can actually decide about.
• Adjacent Adjacent work runs in both directions. The passenger-side density-and-distance economics belong to the high-speed brief and the fixed-link engineering to the intercontinental one; this brief treats both as modal options rather than as subjects. The appraisal and forecast-accuracy literature, including what a retracted canonical citation does to a field, belongs to megaproject governance. Energy corridors share this brief's central institutional problem exactly — one right-of-way assembly, many users, and an unsettled question about who holds the corridor — and the multi-modal corridor hypothesis they both rest on has a clear mechanism and no empirical test. Robotics in infrastructure approaches the terminal from the machine side, and the resilience framing that section 11 declines to resolve is held by the infrastructure-resilience brief.

Established “Cost per route-kilometre figures are comparable across projects.” They are not, and the canonical source says so in its own abstract. The US$50-100 million per route-kilometre band is urban rail, European projects, 2002 prices, outliers excluded, and includes stations and rolling stock. Established Most figures it gets compared against share none of those five properties, and the most common comparison — against an infrastructure-only high-speed-rail figure in a later price year — violates four of them at once. Frontier Enthusiast-side note, because this section is not for one side. The error is committed as often by rail advocates comparing a cheap foreign line to an expensive domestic one as by opponents doing the reverse. Established Terrain class, viaduct and tunnel fraction, land-acquisition regime and labour cost dominate the variance, and Taiwan's 345 km line with 252 km of continuous viaduct is the demonstration. Speculative A per-kilometre number without terrain class, scope definition and price year is not evidence about anything.

Established “N thousand China-Europe trains ran last year, therefore the corridor works.” Three defects, and they compound. Direction: trains per year does not distinguish westbound from eastbound, and the imbalance drives empty repositioning, which is where the money goes. Loading: a train is not a TEU and a TEU is not a loaded TEU; one FEU is roughly two TEU, and conflating the units halves or doubles the figure. Established Subsidy: the modelled optimal subsidy is USD 2,000 to 2,500 per FEU. Frontier At that level the volume figure measures fiscal commitment rather than commercial demand, and the general rule follows: a corridor volume figure without its subsidy is not an economic fact. Speculative This applies to every subsidised corridor on every continent, not only to this one.

Established “Rail forecasts are biased and road forecasts are honest.” The mode-asymmetry claim traces to a paper that was retracted in 2012. Established The live source is Næss, Flyvbjerg and Buhl's 2006 companion, whose title poses the comparison as a question rather than answering it, and whose text was not obtained here. Frontier And the intentionality reading — that forecast error reflects strategic misrepresentation rather than difficulty — is contested outright by Dufour, Steane and Wong, whose contextualist analysis of a Hong Kong container-handling case challenges the assumption that forecast error reflects intentional bias. Speculative Strategic misrepresentation is an inference from outcome data, not an observation of intent, and the distinction matters because the two diagnoses prescribe opposite remedies: one wants better methods, the other wants different incentives.

Established “A retraction means the finding was false.” It does not. A retraction is a statement about a publication, not a verdict on a result, and reasons range from duplicate publication to data error to misconduct. Established The retraction of the canonical traffic-forecast paper is a matter of record with a dated notice; the reason could not be established here and is not asserted. Frontier The unretracted 1996 ancestor reports the substantive finding directly, which is why this brief cites that one. Speculative The interesting fact is not the retraction but that a result probably substantially right has travelled for over a decade on a formally withdrawn citation with nobody downstream checking — a fact about how citation works, not about traffic forecasting.

Established “Reference class forecasting fixed the cost-overrun problem.” One strong aggregate result across 107 projects with overruns falling from 38 per cent to 5 per cent, and one strong negative case study of a 23.6-billion-kroner project that applied the method and failed. Frontier Not settled, and presented as settled in policy documents. Speculative There is also a published study of how these statistics became canonical, which is where to check whether the headline numbers were transmitted faithfully — a question the field asks about other literatures more readily than about its own.

Established “Emissions comparisons between modes are like-for-like.” The 56.97 per cent CO2 reduction figure from the 2025 high-speed-freight optimisation is against air freight only, under an assumed carbon-tax regime, inside a model. Quoted without those qualifiers it becomes a claim about rail against road that its authors did not make. Frontier Such comparisons also require separating traffic a new mode took from traffic it created: induced trips add emissions, only substituted trips subtract.

Established “Platooning is close to deployment and the case is stable.” The field's own reference volume records that the claimed benefits “have now diversified” from the original fuel-cost rationale, and separately that bridge loading worsens as platoon gaps close. Frontier The deployment case has been re-argued, not confirmed. Speculative When a programme's stated output becomes the re-framing of questions, its deployment case has weakened whatever its other merits.

Established “Externality pricing is a technical question awaiting better data.” Externalities are effects on welfare without compensation, which makes their valuation a normative choice about whose welfare counts. Frontier More data does not converge the estimates, and a policy that waits for convergence waits forever.

Established “Railroads were indispensable to continental development.” Fogel's counterfactual is sixty years old, was awarded a Nobel Prize, and is routinely ignored when new continental links are justified by development rhetoric. Speculative Whatever one concludes about the magnitude of the social saving, the methodological demand — name the next-best alternative and price it — is unanswerable and almost never answered.

Handwave “A new continental mode will make the existing ones obsolete.” The strongest version of this claim — vacuum tubes, freight maglev, dedicated automated freeways — fails on a comparison it never makes. Established The correct comparator is the incumbent's improved version at the crossover volume, not the incumbent as it stands. Frontier Every proposal in this cluster is evaluated against a static incumbent, and the incumbent is not static: containerisation, terminal automation, road-freight automation and electrification are all moving its cost curve downward at rates nobody has published. Speculative Make that the design rule and most exotic proposals fail earlier, more honestly, and for a reason their promoters can actually test.