The Stockholm Congestion-Charging Trial 2006: Overview of Effects (Eliasson et al.)

Summary

"The Stockholm congestion–charging trial 2006: Overview of effects" (Transportation Research Part A 43(3), 2009, pp. 240–250), by the trial's own expert-committee evaluators Jonas Eliasson, Lars Hultkrantz, Lena Nerhagen and Lena Smidfelt Rosqvist, is the summary evaluation of what is the cleanest city-scale congestion-pricing episode on record. From 3 January to 31 July 2006 Stockholm ran a full-scale, time-differentiated cordon toll as a trial — deliberately staged so citizens could experience the system before a referendum on 17 September 2006 decided whether to make it permanent (it was reintroduced permanently from August 2007). Because it was a genuine before/during/after switch with a subsequent vote, Stockholm is the piece of evidence transport economists reach for first. This paper is distinct from the two Eliasson works already cited on the outcome page (his 2009 cost–benefit analysis and the 2014 "overview" working paper): it is the multi-author Overview of effects paper, and it is the source for the trial's headline traffic, travel-time, and emissions numbers.

Key Findings

• Cordon traffic fell far more than the target. The toll was set to hit a target of reducing car traffic across the cordon by 10–15%; in fact the reduction "stabilised after the first month at around 22%" during charged hours versus the previous year, and the authors note "effects on congestion reduction were larger than anticipated."[1]
• The cut was broad, not just at the boundary. Reductions were largest in the afternoon peak (−23%, 16:00–18:00) and mid-day (−22%), somewhat lower in the morning peak (−18%); crucially, "the trial showed that a single-cordon toll could affect traffic within a large area," with volumes declining well outside the toll zone.[1]
• Travel-time delays fell sharply. As traffic thinned, "travel times for vehicle traffic declined substantially" — queuing/delay time fell by roughly one-third in the morning peak and one-half in the afternoon/evening peak, and reliability improved (the worst-decile trip durations fell to a third or less of their pre-trial values).[1]
• Emissions fell with the traffic. CO₂ emissions from inner-city traffic were estimated to fall about 14% (traffic within the cordon fell ~16%), alongside reductions in NOₓ and exhaust particles concentrated in the most densely populated areas.[1]
• It penciled out. A permanent scheme was calculated to "yield a considerable social surplus net of investment and operational costs" — an annual social surplus of about SEK 650 million net of operating costs (before counting ~SEK 1,900 million of investment/startup cost), with a social pay-off time of about 4 years.[1]
• The politics reversed once the benefits were felt. "Improvements in travel times were large enough to be perceived by the general public. This was pivotal to the radical change of public attitudes… that resulted in a positive outcome of a subsequent referendum." Roughly 92,000 car trips were "evicted" from the cordon, about half shifting to public transit.[1]

What It Supports

• Congestion pricing reduces traffic and congestion — this is the primary-evaluator source for Stockholm, the strongest of the outcome's three quasi-experiments: a target-beating ~22% traffic cut, large delay reductions, measured emission falls, and a positive cost–benefit balance, all from measured trial data rather than model forecasts.
• Congestion pricing — corroborates and sources the concept page's Stockholm section (the ~20–22% cordon reduction and 30–50% congestion reduction figures) with the evaluation team's own overview.

What It Cuts Against / Honest Limits

• Not a designed experiment. The authors are candid: "being a political rather than scientific project, the trial did not have an experimental or quasi-experimental design," so evaluation relies on before/during/after comparison rather than a clean control — a caveat the cleaner Milan discontinuity (Gibson & Carnovale) helps address.[1]
• Exemptions blunt the price signal. Exemptions (including for eco-fuel cars and Lidingö bypass traffic) made nearly 30% of passages free of charge, muting both revenue and the incentive.[1]
• Distributional texture. Inner-city residents on average paid more and saw smaller travel-time gains than outer-area drivers, yet supported the charge — the paper reads this as evidence some benefits were under-counted, but it is also the distributional wrinkle the outcome page keeps honest about.[1]
• Externality first. As with London, the primary justification is externality correction; the road-space-rent framing is this wiki's Geoist reading of the same charge.

Bears On

• Outcome: Congestion pricing reduces traffic and congestion
• Concept: Congestion Pricing
• Research: The London Congestion Charge (Leape 2006) — the sibling city-scale case
• Objection: Homevoters will block LVT — Stockholm's opinion reversal is the road-pricing analogue of the political-economy problem

See Also

• Congestion pricing reduces traffic and congestion
• Congestion Pricing
• The London Congestion Charge (Leape 2006)
• Road pricing, driver behaviour and air pollution: Milan (Gibson & Carnovale)

Sources

1. Jonas Eliasson, Lars Hultkrantz, Lena Nerhagen & Lena Smidfelt Rosqvist (2009), "The Stockholm congestion – charging trial 2006: Overview of effects," Transportation Research Part A 43(3), 240–250. DOI 10.1016/j.tra.2008.09.007 — used for (full text fetched and read this session, author-hosted PDF) the trial dates and referendum, the ~22% cordon reduction (target 10–15%), the −23%/−22%/−18% peak/mid-day/morning splits, the one-third-to-one-half delay reductions, the ~14% inner-city CO₂ reduction, the ~SEK 650m annual social surplus and ~4-year payoff, the ~30%-of-passages-free exemption figure, the ~92,000 evicted trips, and the verbatim framing quotations. Author PDF · RePEc