Interactive companion · working paper Preliminary
Doubling Luxembourg's highways
What would residents of the Greater Region pay for twice the capacity on every highway link inside Luxembourg? This dashboard shows the model's answer, together with the traffic, speeds and vehicle-kilometres behind it, under four ways the economy can adjust.
Companion to Raian Kudashev and Pierre M. Picard, Welfare Effects of Congestion in Luxembourg and the Greater Region (working paper). Every number is read from the saved model outputs; nothing is re-estimated in the browser.
Preliminary results. These are results from a working paper; numbers may change as the paper develops. Feedback is very welcome: r.kudashev@vu.nl.
Key model assumptionsclick to expand
- Commuting is by car. Every commuter drives, and road traffic consists of commuters only: public transport, freight and non-work trips are not modelled. This is not far from reality for cross-border work trips: 83% of commuters from France's Grand Est region to Luxembourg drive to work (INSEE, 2021 census, reported September 2026).
- Drivers choose routes and congestion responds. Drivers pick among many routes, favouring faster ones; travel time on each road rises with its traffic relative to its capacity. Doubling a highway's capacity therefore attracts traffic from other roads.
- Homes and offices share one floorspace market. In each place, residential and commercial users compete for the same buildings at one price, and the building stock does not change: nothing new is built in the counterfactuals. When better access raises firms' demand for space in Luxembourg City, they bid up its price and floorspace shifts from homes to offices, so the city gains jobs and loses residents.
- Land is owned by the residents of each country. Land income in a country is paid out equally to that country's residents, not to outside landlords. Higher land values in Luxembourg therefore benefit residents of Luxembourg.
- Labour income is taxed where it is earned. As under the prevailing tax treaties, wages are taxed in the country of work, while consumption and housing are taxed in the country of residence. Tax rates do not change; each government spends its revenue on public goods for its own residents.
- Productivity and amenities are fixed. Places do not become more productive or more pleasant as jobs or residents move in: there are no agglomeration effects.
- Costs are not included. Willingness to pay measures the gross benefit of the extra capacity. Construction and maintenance costs of the wider highways are not subtracted.
In the open city, expected utility is pinned to the outside option and total population adjusts instead, so willingness to pay is not a welfare measure there. Its values are shown for completeness only.
Who gains, by country
Willingness to pay is the income a resident would give up each year to keep the doubled highways. Tax revenue follows workplaces, so Luxembourg collects most of the fiscal gain even though residents across the border gain as well.
Willingness to pay per resident
€ per resident per year, by country of residence
Change in tax revenue
€ billion per year, by country
What the willingness to pay is made of
€ per resident per year, by country of residence and channel
Traffic and vehicle-kilometres
Capacity relief shortens trips at first, but once workplaces and homes respond, people commute further. The time saving almost disappears while vehicle-kilometres keep rising, and traffic moves off primary roads onto the highways.
About this dashboard
This page accompanies the working paper by Raian Kudashev and Pierre M. Picard. The quantitative spatial model combines recursive-logit route choice on a simplified road network with location and workplace choice across 1 km cells in Luxembourg, France, Belgium and Germany. Congestion follows a BPR volume-delay function, t = t̄(1 + α(f/k)β), with α and β estimated from traffic-camera data. The counterfactual doubles capacity k on the treated highway links inside Luxembourg.
- Only rerouting: commuter origins and destinations and all economic variables stay at baseline; only routes, flows and travel times respond.
- Fixed residences: residents stay where they live but can change workplace; jobs, wages, floorspace prices and routes adjust.
- Closed city: residents may also move, with total population fixed.
- Open city: total population adjusts too, keeping utility at the outside option.
Credits: route-choice formulation after Fosgerau, Frejinger and Karlström (2013); validation against Duranton and Turner (2011), American Economic Review 101(6), whose commuting-distance, vehicle-km and traffic-diversion elasticities the model reproduces closely (details in the paper); the BPR volume-delay function of the US Bureau of Public Roads. Data sources are listed in the paper. Traffic figures are commuter trips in vehicle equivalents: one modelled commute leg per road commuter.