The Value of a Second Way Across

A bridge carries traffic while giving travelers an alternative to another crossing, a service whose value becomes easier to appreciate when one connection is no longer available. On an ordinary day, traffic can be counted fairly straightforwardly; the contribution of the alternative is harder to observe, even when it helps the network function.

Mannheim and Ludwigshafen provide a useful case because the two German cities face each other across the Rhine, which also forms the boundary between Baden-Württemberg and Rhineland-Palatinate. Their labor markets overlap extensively, while their road networks connect through a small number of crossings, two of which serve the central urban area. Since early August 2026, damage at the Ludwigshafen bridgehead has closed the northern of these, the Kurt Schumacher Bridge, to motor vehicles.

The tempting question is how much this closure costs. The available evidence supports a more useful starting point: which relationships depend on the crossing, what would have to happen for the remaining network to absorb its traffic, and which outcomes would distinguish successful adaptation from a loss of access?

Those questions lead to a different assessment than either a dramatic congestion forecast or a reassuring photograph of an empty road. Almost 29,000 recorded commuting relationships connect the two cities. Dependence on the neighboring labor market is markedly asymmetric. And even substantial diversion away from the remaining central road bridge could leave it with a large additional daily traffic load. None of these findings, however, identifies the closure's realized economic cost.

The missing connection is specific

The closure concerns the Kurt Schumacher Bridge and its northern bridgehead in Ludwigshafen. Tram services across that bridge had already stopped in June. Walking and cycling remain possible. The Konrad Adenauer Bridge provides the other central road crossing; the A6 and its Theodor Heuss Bridge provide a northern alternative. Rail has separate crossing infrastructure. A lost road connection therefore changes the network differently for different modes. City of Mannheim, July 30, 2026

The surrounding network has also been changing, with Ludwigshafen's rebuilt southern elevated road opening to traffic on July 4. Further connections became available during August, including the Mundenheimer Straße access ramp on August 31, while other work and restrictions continued. The treatment in a before-and-after study would consequently be a sequence of network changes, unless the analyst took care to separate them. Ludwigshafen construction updates

This matters for the counterfactual. Continuing to operate an unsafe structure is not an economically meaningful policy alternative. The relevant comparisons concern different safe construction schedules, access arrangements and replacement services. An estimate that treats an unsafe bridge remaining open as an available option would begin with the wrong decision problem.

Start with the accounting

An account of normal traffic volumes reported by SWR, attributed to Mannheim's construction administration, places the two central road bridges at approximately 120,000 vehicles per day in total. About 55,000 used the northern crossing, implying a southern reference volume of about 65,000. The north therefore represented roughly 46 percent of these two bridges' combined reference traffic. This is not a measure of regional road capacity. SWR traffic report

Suppose half of the former northern passages join the southern crossing. Suppose the other half is accommodated outside that southern daily vehicle total, through other routes, other modes, or trips no longer made. The south would then carry 65,000 plus 27,500 passages: 92,500 in total, an increase of about 42 percent over the reference.

The calculation conserves traffic within an explicitly defined scenario, without estimating where drivers have actually gone or how much additional travel time 92,500 passages would produce. Daily volumes cannot reveal the concentration of demand within the peak, the effect of an intersection, or the amount of usable road capacity. Those require observations and a calibrated network model.

The accounting nevertheless disciplines the discussion. Although finding another solution for half the displaced traffic sounds substantial, the half that remains is also substantial relative to the southern crossing's previous load. The accounting makes the relationship between these two quantities explicit.

Reverse the calculation and it becomes a question for traffic management. If the southern daily total is to remain within 30 percent of its reference, approximately 64.5 percent of former northern passages must be accommodated outside that total. At a 20 percent increase, the required share rises to 76.4 percent. These are conditions set by the reader, not engineering limits inferred from the data.

Moving a journey from 8 a.m. to 10 a.m. on the same bridge would not reduce its daily total. It might help the peak considerably, but that is a different mechanism. The distinction is deliberately preserved in the interactive exercise. Otherwise, an apparently simple diversion assumption would quietly combine changes in daily demand with changes in its timing.

Access belongs to households and firms on both banks

The official Länder commuter accounts record 28,845 potential commuters between Mannheim and Ludwigshafen in 2024. Of these, 17,641 lived in Ludwigshafen and worked in Mannheim; 11,204 made the opposite residential-workplace connection. These are recorded employment relationships between places, which should not be converted into daily vehicle counts without information on attendance, mode choice and occupancy. Official 2024 commuter flows

The appropriate denominator changes the interpretation. Adding internal and outbound commuters gives 88,684 recorded employed residents in Ludwigshafen. Mannheim is the workplace of 19.9 percent of them. In the opposite direction, Ludwigshafen accounts for 6.5 percent of Mannheim's 172,289 recorded employed residents. Restricting the denominator to outbound commuters produces shares of 38.8 and 16.2 percent, respectively. The first comparison describes dependence across the recorded resident workforce; the second describes the concentration of external workplaces. Rhineland-Palatinate denominators, Baden-Württemberg denominators

The asymmetry does not establish that one municipality bears a correspondingly larger economic loss. The household encountering a longer journey may be in Ludwigshafen, while the employer having difficulty staffing a shift is in Mannheim. Because the relationship has two ends, each jurisdiction's administrative boundary cuts through the economic benefit of making it work.

This is one reason the value of a crossing is difficult to locate in a single budget. A transport improvement can expand the set of jobs a household can realistically consider and the set of workers a firm can reach. The present data do not estimate that expansion, but they identify a substantial set of existing relationships for which reliable access already matters.

Those relationships were also becoming more numerous before the current closure. Between 2021 and 2024, the combined count rose by 1,293, or 4.7 percent. Ludwigshafen-to-Mannheim relationships increased by 1,361; the opposite direction declined by 68. The first direction thus grew by 8.4 percent, while the second was broadly stable, ending 0.6 percent below its 2021 level.

The trend predates the closure, so it cannot be attributed to a 2026 event. Since the source files contain only the ten leading destinations or origins for each municipality, an absent edge cannot be interpreted as zero. The study therefore uses the directly published city-pair relationship rather than summing the visible edges into a purportedly complete regional labor market. 2021 commuter file, 2022, 2023, 2024

A border is a poor substitute for a journey

The earlier version of this research asked whether the Rhine acted as an invisible border tax. Geography and administration coincide here, making a separate state-border effect difficult to identify. A strong version of the claim would also have to survive inconvenient comparisons.

One such comparison comes from the archived timetable analysis. Starting at Mannheim's main station, the stored connections were used to calculate the earliest available arrival for each of 180 possible start minutes between 6:30 and 9:29 a.m. on September 16. The archived feed was dated August 24, 2026. Waiting counts as part of elapsed time.

In that calculation, the median scheduled elapsed time to Ludwigshafen Mitte was nine minutes. To Mannheim Marktplatz it was eleven. Ludwigshafen Mitte could be reached within twelve minutes for 164 of the 180 possible starting minutes, or 91.1 percent. Crossing the river was competitive with a journey remaining inside Mannheim. Archived routing evidence and reproducible calculation

The result is deliberately retained as an archival finding because a fresh query of the original routing service was unavailable during the update; it cannot therefore serve as a currently verified travel recommendation. It includes neither real-time delays nor journeys from an arbitrary home to the main station or from Ludwigshafen Mitte to an arbitrary workplace.

The limitation points toward the useful unit of analysis: the complete journey at the required time. A direct rail connection may be an excellent substitute when the endpoints align with it, while offering little help to someone whose shift starts before the feeder service operates. Distance to a station, transfer reliability, access at the destination and the ability to change working hours determine how usable the nominal alternative is.

The historical road evidence is similarly specific. An INOVAPLAN accessibility study queried Google Maps travel-time estimates at fifteen-minute intervals over two weeks in January 2020. On its short Konrad Adenauer Bridge section toward Ludwigshafen, 26 percent of queries exceeded that section's own mean time by more than 30 percent. The two Kurt Schumacher sections recorded 11 and 9 percent; the southern section toward Mannheim recorded 1 percent. MRN/IHK/INOVAPLAN 2020, Table 14

These are old estimates for short sections of different lengths, each compared with its own mean. They are not current congestion rates or observations of independent travelers. The same report's regional comparison also placed the morning premium relative to the best travel time at 18 percent for cross-Rhine routes and 25 percent for routes within Baden-Württemberg. The route sets differ, so this does not establish a causal advantage of crossing the river. It does rule out treating the river as a sufficient explanation of every difficult journey.

Less traffic can mean several things

Observed traffic is an outcome of both the opportunities available and the choices people make in response. After a closure, a lower count at a particular location could reflect successful mode substitution. It could also reflect longer detours, fewer visits, postponed appointments or working arrangements that people would not otherwise have chosen.

For that reason, the absence of a visible queue cannot by itself establish the absence of a cost. Equally, displaced vehicle passages cannot simply be multiplied by an assumed delay and called a measured welfare loss. Some journeys may improve, some may disappear at little cost, and others may impose substantial inconvenience. The distribution of changes matters alongside their average.

A useful monitoring design would therefore combine direction-specific traffic and travel times with public-transport arrival reliability, active-travel counts and repeated surveys of complete work journeys. The network's construction state should be recorded alongside holidays, weather and major incidents. Comparisons need consistent weekdays and time windows. A control route affected by the same diversion is not an unaffected control simply because it does not cross the closed bridge.

The objective is not to prescribe an elaborate model before knowing what data exist. It is to make a subsequent estimate answer a defined question. A model of peak road delay, a measure of access to jobs and an estimate of household time costs are different objects. They should not acquire the same label merely because they concern the same closure.

A transparent scale calculation

Consider 10,000 affected person-trips per day, each taking ten additional minutes. Together they generate roughly 1,667 additional person-hours daily. At an assumed €20 per person-hour over 250 assumed days, the annualized time value is approximately €8.3 million.

Every input in the example is an assumption, with the 10,000 person-trips specified independently of the 55,000 vehicle passages. A return journey counts twice if both legs incur the stated delay, while the illustrative €20 hourly value and 250 days define how the resulting time is valued. The sensitivity table displays exact outcomes under alternative assumptions; it is not a confidence interval around an estimated closure cost.

The example provides a way to ask whether accelerating a safe repair or improving a replacement connection could be worth its cost. Under the stated assumptions, one-twelfth of the modeled duration corresponds to about €0.7 million in gross time value. A real proposal would have to establish how much of the assumed burden it actually removes, for how long, and at what implementation cost.

Other effects remain outside this arithmetic. Unreliable travel may require additional buffers, while missed connections or inflexible schedules may constrain the feasible alternatives. Conversely, an hour of personal time saved does not automatically become an hour of paid production. Time value is not interchangeable with lost GDP, and a scenario total should not acquire that interpretation as it moves into a headline.

What the second crossing was providing

The case establishes substantial integration, asymmetric dependence and a demanding arithmetic of diversion. It also establishes the limits of a broad border narrative. Some cross-river journeys have useful rail alternatives, and historical road performance varied sharply by section and direction.

What remains unmeasured here is the realized distribution of delay, adaptation and lost access since the closure, and that gap should determine the next measurement effort. A more confident adjective would do little to resolve it.

For infrastructure policy, the broader implication is that an alternative connection can have value even when another route remains open. Its contribution depends on whether it provides a workable substitute when the network changes. This makes redundancy an economic question about access, reliability and the range of feasible choices, as well as an engineering question about structures.

The Rhine still separates two municipalities and two states. Whether it separates their labor markets more strongly now depends on the journeys people can actually complete. A successful response to the closure would preserve those possibilities, especially for people whose working lives leave them little room to improvise.

Methods and source boundaries

The commuter calculations use the official 2021–2024 state flow files and 2024 municipal totals. Employed residents are represented by internal plus outbound potential commuters. Published top-ten relationships are not a complete origin-destination matrix. Historical road indicators come from Tables 2, 13 and 14 of the 2020 MRN/IHK/INOVAPLAN study. Archived timetable itineraries are retained with their feed date, query URLs, scheduled status and recomputed start-minute results. They are not real-time observations or a newly verified September 4 timetable.

Traffic accounting uses N = 55,000 and S = 65,000 reference vehicle passages per day. If a is the share of former northern passages accommodated outside the southern daily total, the scenario gives S′ = S + N(1 − a). A chosen proportional southern increase g requires a ≥ max(0, 1 − gS/N). Neither expression estimates capacity or congestion. Time value is Q × d / 60 × D × v, where Q is affected person-trips per day, d is additional minutes per trip, D is assumed days and v is euros per person-hour. All inputs remain independently adjustable. The evidence package includes exact values, sources, acquisition receipts, calculations and validation records.