70 öre på kronan

Sweden has run this arithmetic for sixty years

A samlad effektbedömning, a consolidated impact assessment, is a standardized form. Every page carries the object number, the case officer and a status line reading granskad och godkänd av Trafikverket, reviewed and approved, with a date. Inside sit tables of present values, row by row: passengers, freight, traffic safety, climate, health, landscape, each cell colour-coded. An analysis-result box condenses it all to four lines: net present value, the benefit-cost ratio called NNK, effects not calculated, and final assessed profitability. In the Hässleholm–Lund assessment the box reads minus 22,132 million kronor, minus 0.78, negligible, robustly unprofitable. The whole archive lies open on Trafikverket's website, sorted by region.

The ratio is built to a rulebook called ASEK. The benefit side is forecasts converted to kronor, with travel-time gains carrying 60 to 80 percent of the quantified value, followed by traffic safety and emissions, all discounted at 3.5 percent over a sixty-year horizon. The cost side is the construction money multiplied by 1.2, because a tax-financed krona is assumed to cost society 1.20 kronor once the distortions of collecting it are counted. Trafikverket's own guidance then grades the result: profitable above plus 0.1, unprofitable below minus 0.1.

Sweden has been refining this instrument for sixty years. Vägverket began building economic decision support for road investments in the 1960s and 70s, the ASEK working group has managed the calculation values since the 1990s, and since 2010 the method lives at Trafikverket, with an inter-agency advisory group and a scientific council. The method even audits itself. The Hässleholm–Lund assessment flags that a discovered error in the national forecast model inflates predicted rail travel:

Det har upptäckts fel i prognosmodellen Sampers som ger ett för högt resande med tåg … vilket innebär att de beräknade nyttorna i denna kalkyl kan vara överskattade.

The minus 0.78 is, if anything, generous.

Seventy öre, by the state's own count

The portfolio figure comes from the proposal Trafikverket delivered on 30 September 2025, page 23: the named investments whose profitability has been assessed average an NNK of minus 0.3, which Trafikanalys translates to a net present value of roughly minus 50 billion kronor. The same page splits the average in two. Five objects inherited as bound by earlier government decisions, Göteborg–Borås, Hässleholm–Lund, Norrbotniabanan, four tracks through Uppsala and Sydostlänken, average minus 0.8. The remaining named investments average plus 1.0. On 24 April 2026 the government fixed the plan at 1,171 billion kronor and presented it at a press conference four days later; the object list shifted at the edges, and no recalculated ratio for the fixed plan has been published.

Stockholms Handelskammare, which ran its own review of the plan that spring, translated the ratio into öre:

Det innebär att varje investerad krona skapar samhällsnytta värd 70 öre.

For scale: HS2, the railway Britain treats as a national embarrassment, converts to plus 0.2 on the Swedish ruler. Norway is planning Ringeriksbanen at minus 0.7, and Crossrail in London was approved at plus 1.0, where the British definition of high value for money begins.

Three confusions to avoid. The figure is monetized benefit, travel time, safety and emissions priced at the state's own calculation values, not cash back to the treasury. It covers the plan's named new builds; maintenance and minor upgrades, roughly half the money, are profitable and not included, and inside the new builds it is the five bound giants at minus 0.8 that pull a plus 1.0 remainder under water. And it is not Riksrevisionen's minus 0.39, a median for objects newly added to the previous plan, 2022 to 2033, a group whose average was at the same time plus 0.39: a different statistic about a different plan. Use either, but say which.

The roads are the state's most audited money

Before the conclusion becomes that infrastructure is waste, turn the comparison around. The infrastructure plan is the only one of the state's large money flows that gets a standardized calculation in advance at all. The ROT renovation deduction costs about twelve billion kronor a year; when Riksrevisionen audited it after the fact, it found the deduction far from self-financing, with at most 3,500 full-time equivalents of added labour supply and no measurable effect on the buyers' incomes, and recommended shrinking it. The interest deduction cost 61 billion kronor last year, by the National Financial Management Authority's reckoning, and no calculation has ever been made.

The seventy-öre krona survives that comparison surprisingly well: transparently computed, object by object, by a published method, and once the money is spent there is at least a road to show for it. The scandal is not that Sweden counts badly. Sweden counts better than almost anyone. The scandal is that the count carries no weight: the road is audited in advance and built anyway; the deductions are never audited and roll on anyway.

Once, the number said no

How little the count weighs shows in the selection. Trafikverket runs the ratio both on the projects that make the plan and on the candidates that are turned away, and Handelskammaren's review found the two groups land on the same minus 0.3. Whatever decides which objects get in, it is not the arithmetic. Riksrevisionen has called entry into the plan a needle's eye: once through, a project is in practice built whatever happens to its cost. Since 2010, thirteen objects have left the plan without being built, none of them over cost.

There is one exception. In the proposal, Trafikverket struck an Ostkustbanan stage, Kubikenborg–Dingersjö south of Sundsvall, 3.8 billion kronor, robustly unprofitable at minus 0.9, citing deficient societal-economic efficiency: the one place in 251 pages where the number was allowed to act as a gate. On 24 April the government put the stage back, with a press release celebrating the investment in Västernorrland. The proposal came in September, the reversal in April: the gate had held for seven months.

On 16 June 2026 the riksdag debated the plan for a full day. The word nettonuvärdeskvot was never uttered, according to the protocol, and neither was 70 öre. The adjective, on the other hand, was everywhere. Sten Bergheden of the Moderates credited the government with not building unprofitable high-speed rail, Malin Östh of the Left attacked Östlig förbindelse, a motorway classed as robustly unprofitable and sitting in the plan regardless, and Aylin Nouri of the Social Democrats drew her own conclusion: revise the models until they capture defence, the climate and growth.

The rest of the portfolio already clears the bar

The ratio is not fate, and the proposal itself says so: outside the five bound giants, the named investments average plus 1.0, the road objects 1.4, comfortably past the profitability bar. The previous plan sat at minus 0.2, and the slide is Nordic-wide and some fifteen years old. What sinks the average is the megaprojects' cost side, and it grows during the planning years, in the investigations, design rounds and reviews that fill the decade before construction, while the construction sites themselves deliver roughly on budget. None of this argues for investigating less; the assessments are the part of the system that worked. The cost sits in what every added year of producing them lays onto the projects that wait.

The size of the lever has been estimated. In 2022 the construction sector produced 4 percent less per worked hour than it had 25 years earlier; hold the escalation down, Innovationsföretagen calculates, and the money already committed would stretch roughly 225 billion kronor further over the plan period. Projects that are unprofitable on paper today would then land above the bar without a single forecast changing. And the decade where that is decided does not sit in the ministry: roughly half of the 450 assessments behind the last plan revision were bought from consultancies, the investigations and design rounds behind them are drawn up the same way, and the pace of the plan's most expensive years is set at the desks where those documents are made, most of which stand with the consultancies and contractors.

The method, for its part, keeps its side of the bargain. On 4 May 2026, ten days after the government fixed the plan, the next edition of the rulebook came into force: ASEK 8.1, revised values, a scientific council behind every figure. Somewhere in the archive a case officer is already grading the next object, filling in the four lines, marking the file granskad och godkänd. Whether the answer is read before the decision is not in the rulebook, and never has been.