1. October 8, 2010 ModellingConsumerBehaviour in AirportSelection A System Dynamics perspective on the Dutch (and German) ticket tax Bart Steverink – bart@bsoms.com INFRAday 2010, Berlin The Netherlands Institute for Transport Policy Analysis (KiM) Faculty of Technology, Policy and Management
2. October 8, 2010 Contents 1 Introduction & ProblemDescription Research Questions & Methodology ModellingMethod Overview of the Model Results Conclusions Further Research 2 3 2
6. The Netherlands Institute for Transport Policy Analysis (KiM) initiated an ex-post analysis in January 2010
7.
8. October 8, 2010 Research Questions Whichmethod is suitable to model consumerbehaviour and airportcompetition? Which factors and mechanismsinfluenced the redistribution of passengers as a result of the ticket tax? Ifthere is a structuralchange, whatcanbedone to influencethischange? 6 Method Mechanism Policy
9. October 8, 2010 Methodology Literature study/interviews How decisions are made? What determines airport quality? How do people obtain information? Computer Simulation Model Reduce complexity How do the different concepts from literature interact over time? How can the choice be influenced? 7
11. October 8, 2010 Modelling MethodComparison of Modelling Methods
12.
13. Previous studies
14. Airport statistics
15. KiM Airport Choice Survey 201010
16. October 8, 2010 Overview of the ModelSelection of airports 3 Airports selected for case study: Eindhoven Weeze (Brussels South) Charleroi Because: Overlapping catchment areas Very similar type of services “Footloose” character of low-cost carriers Target market relatively price sensitive 11
17. October 8, 2010 Overview of the ModelTotal system (simplified) 12 Demand Airlines Consumers Utility perception Price Optimal use of fleet Information transfer Awareness # of flights Supply
18. October 8, 2010 Overview of the ModelConsumers 3 pieces of information required to make a decision: Preferences Alternatives (airports) Attributes of alternatives: Price airline ticket Number of flights Access time Learning by using, talking and marketing Forgetting as a result of lack of exposure 13
19. October 8, 2010 Overview of the ModelAirlines Optimize fleet utilization by: Adding flights in case of high demand Reducing flights when demand is low Adjusting price according to demand 14
21. October 8, 2010 ResultsComparison with Empirical data Growth Eindhoven compared to same month last year 80 70 60 50 40 30 20 10 0 -10 -20 Tax % real model 16
22. October 8, 2010 ResultsEffect of marketing 17 Awareness of Dutch people of Weeze % of people aware Boost Normal
23. October 8, 2010 ResultsInfluence of Awareness 18 Dutch people going to Weeze % of Dutch Tax No tax
24. October 8, 2010 Conclusions SD is a suitable method but has limitations Transfer of information by marketing, talking to friends and neighbours and by direct experience. A structural effect of the tax is seen if: The supply can grow fast enough Enough attention is generated Policies that influence information availability are effective 19 Method Mechanism Policy
25. October 8, 2010 Further ResearchEffects of the German ticket tax Somepreliminaryconclusionsbasedon the model: Dutch passengerswill return to Dutch airports, reactionfromGermanpassengers low, due to low awareness and threshold of flightsupply Re-implementation of Dutch taxreducesnumber of Dutch passengersreturning to NL and increasestraffic at Belgianairports Opportunitiesfornear border airportslikeMaastricht-Aachen and (maybe) Twente 20
26. Further ResearchEffects of the German ticket tax October 8, 2010 21 Percentage of population (NL and D) choosingairport Weeze (from NL) Eindhoven (from D)
27. Further ResearchEffects of the German ticket tax October 8, 2010 22 Percentage of Dutch leisurepassengersreturning to EHN NL and GermanTax Re-impl. Dutch Tax
28. Further ResearchEffects of the German ticket tax October 8, 2010 23 Percentage of Dutch choosing Charleroi Re-impl. Dutch Tax NL and GermanTax
29.
30. October 8, 2010 Further Research Twodirectionsforfurther research: Simplifystructure and expandboundaries More numericalprecision 25
Editor's Notes
Start with introducing:Welcome for comingMyselfThe reason we are here is to present the result of the research to obtain the degree of master of scienceAsk the audience first if they knew about the taxSecond if they know about this german airport weeze
Introduce the three parts
Lower curve: Eindhoven users starting to choose for weeze, rapid increase when implemented and rapid decrease when abolished. However, awareness decreases the speed of decline. Shows the difference in speed between mechanisms of learning (direct experience and wom)
Dutch PeoplechoosingWeezefrom the NetherlandsGermanPeoplechoosing Eindhoven fromGermany3 cases, No tax, justdutch, Dutch and afterGerman
Fraction of Dutch Weeze Leisure travelersreturning to Eindhoven in 3 casesJust dutchtaxDutch and afterGermandtaxDutch and afterGermand and thenreimplementation of Dutch tax in 2012
Fraction of Dutch choosing Charleroi in three casesJust Dutch taxDutch and GermantaxDutch and German and reimplementation Dutch tax