Transport models and extreme scenarios

Transcription

1 concept Christian Steinsland Lasse Fridstrøm Transport models and extreme scenarios A test of the NTM5 model Concept report no. 41 concept

2 concept Christian Steinsland Lasse Fridstrøm Transport models and extreme scenarios A test of the NTM5 model Concept report no. 41

3 Concept Research Programme 2014 Concept report no. 41 Transport models and extreme scenarios A test of the NTM5 model English summary Christian Steinsland Transportøkonomisk institutt Lasse Fridstrøm Transportøkonomisk institutt Original title: Transportmodeller på randen. En utforsking av NTM5-modellens anvendelsesområde. For the full report in Norwegian, see Concept website ISSN: (papirversjon) ISSN: (nettversjon) ISBN: (papirversjon) ISBN: (nettversjon) Date: 12 June 2014 Publisher: Concept Research Programme Norwegian University of Science and Technology 7491 NTNU Trondheim Norway Responsibility for the information in the reports produced on behalf of the Concept Research Programme is on the commissioned party. Views and conclusions is on account of the authors and not necessarily identical to the views of the Concept Research Programme. All contributions are reviewed in a peer review process.

4 2 The national Norwegian long distance travel demand model has been subjected to various stress tests. We explore the predicted effects of a 12 fuel price, of a 100 kilometres per litre fuel efficiency, and of a drastically improved road infrastructure between major cities. The model appears to handle fairly large changes in input rather well. However, the most extreme scenarios do not engender credible results. As the fuel cost of travel is reduced to 1/7, car travellers are predicted to choose radically more distant destinations, as if the time cost of travel were of only minor importance. The NTM5 model for long distance travel demand in Norway Covering trips longer than 100 km one way, the national Norwegian travel demand model for long distance trips, NTM5, is part of a larger modelling apparatus available to transport planners in Norway. Being fully segmented by travel purpose, the model predicts long haul travel frequency, mode and destination choice in a network with 1428 zones. The underlying behavioural choice models have been estimated on data from the national travel survey of 1998, while network data have, in this application, been updated to 2010 (Figure E.1). Scenarios The purpose of the exercise is to explore whether the model can provide useful predictions when input variables are extrapolated far outside their observed empirical range. To this end, we have run, inter alia, the following scenarios: Reference (business-as-usual) Extreme fuel efficiency in passenger cars: 0,01 litres per km Extreme fuel price: NOK 100 per litre, i. e. more than 12, due (a) to a sudden increase in the fuel tax paid by private motorists, or (b) to a higher crude oil price. Motorways allowing for 120 km/h travel speed between the four major cities. In the extreme fuel efficiency scenario, the overall long haul trip frequency increases by 14 per cent, while aggregate passenger kilometres travelled go up by 38 per cent. For the car mode, the corresponding rates of impact come out at 33 and 97 per cent. In the Oslo-Bergen and Oslo-Trondheim corridors, long distance travel by car is predicted to grow by 264 and 248 per cent, respectively.

5 Figure E.1: NTM5 model network. Roads marked in red, railroads in black, sea routes in blue, air routes in grey. Certain Swedish roads are included, as they sometimes offer faster routes on Norwegian origin-destination pairs. The extreme fuel price scenario produces very different results depending on the origin of the price increase. If the price increase results from a sudden increase in

6 4 the fuel tax, so that car owners do not have time to replace their vehicles by more energy efficient ones, and so that other modes of travel escape an increase in energy costs, vehicle kilometres travelled on long haul trips by car fall by 53 per cent. If, on the other hand, one imagines that the fuel price increase is due to a 20-fold surge in the world market price of crude oil, which has developed over a certain period of time, the effect will be very different. In this case, one must assume that cars as well as coaches, airplanes and ships, being exposed to rising energy costs, have become drastically more energy efficient than today. In summary, we assume that air fares increase by 146 per cent, coach fares by 61 per cent and boat fares by 72 per cent. The mean per kilometre fuel consumption of private cars is assumed to decrease to 0,01 litre, i.e. to roughly 1/7 of today s level. In this case, overall long distance travel demand, as measured in passenger kilometres, goes down by 14 per cent. The air mode shrinks by 56 per cent, while the car mode actually grows by 8 per cent, as the energy efficiency improvement more or less offsets the fuel price increase. While the cities of Bergen, Trondheim and Stavanger are no more than 500 km by road apart from Oslo, the present state of the national road network does not allow for shorter travel times by car than 6-8 hours. If, on the other hand, we imagine four-lane motorways allowing for 120 km/h travel speed throughout the Oslo- Bergen, Oslo-Trondheim and Oslo-Stavanger corridors (see Figure E.2), the model predicts a 5 per cent increase in national overall long distance travel demand and a 12 per cent increase in passenger kilometres travelled by car. In the Oslo-Bergen and Oslo-Trondheim corridors specifically, long distance car travel demand is predicted to grow by per cent. In the Bergen-Stavanger corridor, where no road improvement is assumed to take place, overall long distance travel demand is predicted to shrink by 5-6 per cent, as is also the vehicle kilometres travelled by car.

7 Figure E.2: Alignment of hypothetical motorways between four major cities (in green). Assessment Extreme input assumptions must be expected to produce extreme output. In view of the rather drastic changes in relative prices and travel times that underlie our scenarios, the aggregate changes in travel demand, compared to the reference scenario, do not appear unreasonable. However, when a closer look is cast on the less aggregate origin-destination flows, which add up to the national totals, a number of puzzling results come to light.

8 6 In the extreme fuel efficiency scenario, a major shift in destination choice is predicted to take place, so as to increase the average length of haul (on trips longer than 100 km) by 48 per cent. According to the model, when fuel becomes sufficiently cheap, road users find it worthwhile to travel considerably longer for the same purpose, implicitly accepting an increase in their time costs more or less in proportion to the added journey length. This observation gives rise to a suspicion that the generalised cost functions implicit in the mode and destination choice models places too little weight on the time component and too much weight on the out-of-pocket expenditure. On account of this, destination choice comes out as quite elastic in response to changes in fuel cost. Such a diagnosis seems reasonable in view of the rather obsolete travel behaviour data underlying the model. Since 1998, subjective values of time have probably increased considerably among the Norwegian population, in line with rising income. Another indication of this apparently exaggerated elasticity is visible in the travel demand predictions for specific corridors. In the extreme fuel price scenario a, overall long distance travel demand in the Oslo-Bergen and Oslo-Trondheim corridors increase by 22 and 18 per cent, respectively, although the car mode shrinks by 35 and 40 per cent. The coach, rail and air mode expand by 59, 47 and 50 per cent, respectively, in the Oslo-Bergen corridor. Similar changes take place in the Oslo-Trondheim corridor. When the car mode becomes exorbitantly expensive, the model predicts a change in mode and destination choice favouring destinations that are easily accessible by other modes than the car. Unlike most other destinations in Norway, those located in the two main corridors are often served by a number of different modes. The model assumes these destinations to become comparatively much more attractive when private car use becomes very expensive. In a situation with a very high kilometre cost of car use, one would expect the occupancy rate to increase. A number of car trips will not take place unless there are several persons to share the cost. But the NTM5 model does not contain any behavioural relations capturing this phenomenon. Occupancy rates are set at a fixed level within each travel purpose and distance band. In the NTM5 model, driver s license holding and car ownership are set at certain levels. But when large changes in exogenous variables occur, one might want to take account of the medium and long term effects on car ownership, which in turn affects trip frequency and mode and destination choice. The absence of such

9 relationships makes the model less useful for the study of scenarios that are radically different from the present situation. Last, but not least, a major drawback of the Norwegian travel demand apparatus is the arbitrary distinction made between trips shorter or longer than 100 km. The shorter trips are handled by a set of regional travel demand models (RTM), with no interaction with NTM5. Separate coefficients and implicit values of time are estimated for the two distance bands. The modelling system ignores the fact that trips 95 and 105 km long may have a number of common characteristics. More seriously, it ignores the fact that destinations 95 and 105 km away from the origin may well be in competition with each other. In the NTM5 model, nearby destinations simply do not exist.

10 Concept report series Paper version ISSN Web version ISSN Norwegian English summary Report Title Author No 1 No 2 No 3 No 4 No 5 No 6 No 7 No 8 No 9 No 10 No 11 No 12 No 13 Styring av prosjektporteføljer i staten. Usikkerhetsavsetning på porteføljenivå Project Portfolio Management. Estimating Provisions for Uncertainty at Portfolio Level. Statlig styring av prosjektledelse. Empiri og økonomiske prinsipper. Economic Incentives in Public Project Management Beslutningsunderlag og beslutninger i store statlige investeringsprosjekt Decisions and the Basis for Decisions in Major Public Investment Projects Konseptutvikling og evaluering i store statlige investeringsprosjekt Concept Development and Evaluation in Major Public Investment Projects Bedre behovsanalyser. Erfaringer og anbefalinger om behovsanalyser i store offentlige investeringsprosjekt Needs Analysis in Major Public Investment Projects. Lessons and Recommendations Målformulering i store statlige investeringsprosjekt Alignment of Objectives in Major Public Investment Projects Hvordan trur vi at det blir? Effektvurderinger av store offentlige prosjekt Up-front Conjecture of Anticipated Effects of Major Public Investment Projects Realopsjoner og fleksibilitet i store offentlige investeringsprosjekt Real Options and Flexibility in Major Public Investment Projects Bedre utforming av store offentlige investeringsprosjekter. Vurdering av behov, mål og effekt i tidligfasen Improved Design of Public Investment Projects. Up-front Appraisal of Needs, Objectives and Effects Usikkerhetsanalyse Kontekst og grunnlag Uncertainty Analysis Context and Foundations Usikkerhetsanalyse Modellering, estimering og beregning Uncertainty Analysis Modeling, Estimation and Calculation Metoder for usikkerhetsanalyse Uncertainty Analysis Methodology Usikkerhetsanalyse Feilkilder i metode og beregning Uncertainty Analysis Methodological Errors in Data and Stein Berntsen and Thorleif Sunde Dag Morten Dalen, Ola Lædre og Christian Riis Stein V. Larsen, Eilif Holte and Sverre Haanæs Hege Gry Solheim, Erik Dammen, Håvard O. Skaldebø, Eystein Myking, Elisabeth K. Svendsen and Paul Torgersen Petter Næss Ole Jonny Klakegg Nils Olsson Kjell Arne Brekke Petter Næss med bidrag fra Kjell Arne Brekke, Nils Olsson and Ole Jonny Klakegg Kjell Austeng, Olav Torp, Jon Terje Midtbø, Ingemund Jordanger and Ole M Magnussen Frode Drevland, Kjell Austeng and Olav Torp Kjell Austeng, Jon Terje Midtbø, Vidar Helland, Olav Torp and Ingemund Jordanger Kjell Austeng, Vibeke Binz and Frode Drevland

11 Concept report series Paper version ISSN Web version ISSN Norwegian English summary Report Title Author Analysis No 14 No 15 No 16 No 17 No 18 No 19 No 20 No 21 No 22 No 23 No 24 No 25 Positiv usikkerhet og økt verdiskaping Positive Uncertainty and Increasing Return on Investments Kostnadsusikkerhet i store statlige investeringsprosjekter; Empiriske studier basert på KS2 Cost Uncertainty in Large Public Investment Projects. Empirical Studies Kontrahering i prosjektets tidligfase. Forsvarets anskaffelser. Procurement in a Project s Early Phases. Defense Aquisitions Beslutninger på svakt informasjonsgrunnlag. Tilnærminger og utfordringer i prosjekters tidlige fase Decisions Based on Scant Information. Challenges and Tools During the Front-end Phases of Projects Flermålsanalyser i store statlige investeringsprosjekt Multi-Criteria Decision Analysis In Major Public Investment Projects Effektvurdering av store statlige investeringsprosjekter Impact Assessment of Major Public Investment Projects Investorers vurdering av prosjekters godhet Investors Appraisal of Project Feasibility Logisk minimalisme, rasjonalitet - og de avgjørende valg Major Projects: Logical Minimalism, Rationality and Grand Choices Miljøøkonomi og samfunnsøkonomisk lønnsomhet Environmental Economics and Economic Viability The Norwegian Front-End Governance Regime of Major Public Projects A Theoretically Based Analysis and Evaluation Markedsorienterte styringsmetoder i miljøpolitikken Market oriented approaches to environmental policy Regime for planlegging og beslutning i sykehusprosjekter Planning and Decision Making in Hospital Projects. Lessons with the Norwegian Governance Scheme. Ingemund Jordanger Olav Torp (red.), Ole M Magnussen, Nils Olsson and Ole Jonny Klakegg Erik N. Warberg Kjell Sunnevåg (red.) Ingemund Jordanger, Stein Malerud, Harald Minken and Arvid Strand Bjørn Andersen, Svein Bråthen, Tom Fagerhaug, Ola Nafstad, Petter Næss and Nils Olsson Nils Olsson, Stein Frydenberg, Erik W. Jakobsen, Svein Arne Jessen, Roger Sørheim and Lillian Waagø Knut Samset, Arvid Strand and Vincent F. Hendricks Kåre P. Hagen Tom Christensen Kåre P. Hagen Asmund Myrbostad, Tarald Rohde, Pål Martinussen and Marte Lauvsnes No 26 Politisk styring, lokal rasjonalitet og komplekse koalisjoner. Tidligfaseprosessen i store offentlige investeringsprosjekter Political Control, Local Rationality and Complex Coalitions. Focus on the Front-End of Large Public Investment Projects Erik Whist and Tom Christensen

12 Concept report series Paper version ISSN Web version ISSN Norwegian English summary Report Title Author No 27 No 28 No 29 No 30 No 31 No 32 No 33 No 34 No 35 No 36 No 37 Verdsetting av fremtiden. Tidshorisont og diskonteringsrenter Valuing the future. Time Horizon and Discount Rates Fjorden, byen og operaen. En evaluering av Bjørvikautbyggingen i et beslutningsteoretisk perspektiv The Fjord, the City and the Opera. An Evaluation of Bjørvika Urban Development Levedyktighet og investeringstiltak. Erfaringer fra kvalitetssikring av statlige investeringsprosjekter Sustainability and Public Investments. Lessons from Major Public Investment Projects Etterevaluering av statlige investeringsprosjekter. Konklusjoner, erfaringer og råd basert på pilotevaluering av fire prosjekter Evaluating Public Investment Projects. Lessons and Advice from a Meta-Evaluation of Four Projects Store statlige investeringers betydning for konkurranse- og markedsutviklingen. Håndtering av konkurransemessige problemstillinger i utredningsfasen Major Public Investments' Impact on Competition. How to Deal with Competition Issues as Part of the Project Appraisal Analyse av systematisk usikkerhet i norsk økonomi. Analysis of Systematic Uncertainty in the Norwegian Economy. Planprosesser, beregningsverktøy og bruk av nyttekostnadsanalyser i vegsektoren. En sammenlikning av praksis i Norge og Sverige. Planning, Analytic Tools and the Use of Cost-Benefit Analysis in the Transport Sector in Norway and Sweden. Mulighetsrommet. En studie om konseptutredninger og konseptvalg The Opportunity Space. A Study of Conceptual Appraisals and the Choice of Conceptual Solutions. Statens prosjektmodell. Bedre kostnadsstyring. Erfaringer med de første investeringstiltakene som har vært gjennom ekstern kvalitetssikring Investing for Impact. Lessons with the Norwegian State Project Model and the First Investment Projects that Have Been Subjected to External Quality Assurance Bruk av karbonpriser i praktiske samfunnsøkonomiske analyser. En oversikt over praksis fra analyser av statlige investeringsprosjekter under KVU-/KS1-ordningen. Kåre P. Hagen Erik Whist and Tom Christensen Ola Lædre, Gro Holst Volden and Tore Haavaldsen Gro Holst Volden and Knut Samset Asbjørn Englund, Harald Bergh, Aleksander Møll and Ove Skaug Halsos Haakon Vennemo, Michael Hoel and Henning Wahlquist Morten Welde, Jonas Eliasson, James Odec and, Maria Börjesson Knut Samset, Bjørn Andersen and Kjell Austeng Knut Samset and Gro Holst Volden Knut Samset and Gro Holst Volden Gro Holst Volden

13 Concept report series Paper version ISSN Web version ISSN Norwegian English summary Report Title Author No 38 No 39 No 40 No 41 Use of Carbon Prices in Cost-Benefit Analysis. Practices in Project Appraisals of Major Public Investment Projects under the Norwegian State Project Model Ikke-prissatte virkninger i samfunnsøkonomisk analyse. Praksis og erfaringer i statlige investeringsprosjekter Non-Monetized Impacts in Economic Analysis. Practice and Lessons from Public Investment Projects Lav prising store valg. En studie av underestimering av kostnader i prosjekters tidligfase Low estimates high stakes. A study of underestimation of costs in projects' earliest phase Mot sin hensikt. Perverse insentiver om offentlige investeringsprosjekter som ikke forplikter Perverse incentives and counterproductive investments. Public funding without liabilities for the recipients Transportmodeller på randen. En utforsking av NTM5- modellens anvendelsesområde Transport models and extreme scenarios. A test of the NTM5 model Heidi Bull-Berg, Gro Holst Volden and Inger Lise Tyholt Grindvoll Morten Welde, Knut Samset, Bjørn Andersen and Kjell Austeng Knut Samset, Gro Holst Volden, Morten Welde and Heidi Bull- Berg Christian Steinsland and Lasse Fridstrøm

14 Concept report no Forskningsprogrammet Concept skal utvikle kunnskap som sikrer bedre ressursutnytting og effekt av store, statlige investeringer. Programmet driver følgeforskning knyttet til de største statlige investeringsprosjektene over en rekke år. En skal trekke erfaringer fra disse som kan bedre utformingen og kvalitetssikringen av nye investeringsprosjekter før de settes i gang. The Concept research program aims to develop know-how to help make more efficient use of resources and improve the effect of major public investments. The Program is designed to follow up on the largest public projects over a period of several years, and help improve design and quality assurance of future public projects before they are formally approved. Concept er lokalisert ved Norges teknisk- naturvitenskapelige universitet i Trondheim (NTNU), ved Fakultet for ingeniørvitenskap og teknologi. Programmet samarbeider med ledende norske og internasjonale fagmiljøer og universiteter, og er finansiert av Finansdepartementet. The program is based at The Norwegian University of Science and Technology (NTNU), Faculty of Engineering Science and Technology. It cooperates with key Norwegian and international professional institutions and universities, and is financed by the Norwegian Ministry of Finance. Address: The Concept Research Program Høgskoleringen 7A N-7491 NTNU Trondheim NORWAY ISSN: (papirversjon) ISSN: (nettversjon) ISBN: (papirversjon) ISBN: (nettversjon)