Erasmus University Rotterdam - EUR



Erasmus University RotterdamThe expected effect of Fyra on air transport – learning from the past experiences from ICEMaster ThesisMaster in Economics and BusinessSpecialization in Urban, Port and Transport EconomicsStudent: Ziye YaoStudent number: 311602Thesis supervisor: Dr. Peran van ReevenDepartment of Applied EconomicsErasmus School of EconomicsErasmus University RotterdamDate: 17 Nov 2011AbstractThe high-speed train Fyra was introduced to Schiphol on September 2009. Since then, there have been many discussions about the possible effects of Fyra on air transport. Some argued it would increase the air transport while others argued the different way. This paper will discuss the possible future effect of Fyra by comparing Fyra with the ICE for Frankfurt airport. ICE was introduced ten years earlier than Fyra, and Frankfurt airport has the similar size and function as Schiphol airport. A difference-in-difference (DID) estimation will be performed with Schiphol airport as a control group to address the effect of the ICE on air transport. By comparing Fyra to ICE with the result from the DID estimation, along with some evidences and past researches for both airport, it would be able to estimate the expected effect of Fyra on air transport: Fyra is expected to reduce the air transport, but it would be socially beneficial. Furthermore, the effect of Fyra on air transport will be less significant than the effect ICE.AcknowledgmentI would like to express my gratitude to all the people that helped me and supported me with my thesis.Firstly, I would like to thank my thesis supervisor, Dr. Peran van Reeven for all the help and support. All the helpful discussion session with him and all the valuable suggestions and guidance from him was very important for me to complete this paper.Secondly, I would like to thank my friend: Lei Shi, Man Xu, Yaxian, Wu and Yiming Zhong. For all the support and comfort they gave me during the writing process.Finally, I would like to appreciate all the love and support from my parents. Table of content TOC \o "1-3" 1.Introduction PAGEREF _Toc183114112 \h 51.1 Introduction PAGEREF _Toc183114113 \h 51.2 Problem Statement PAGEREF _Toc183114114 \h 61.3 Methodology PAGEREF _Toc183114115 \h 61.4 List of Chapters PAGEREF _Toc183114116 \h 62.Literature Review PAGEREF _Toc183114117 \h 72.1 Reasons why HST should increase air transport PAGEREF _Toc183114118 \h 72.2 Reasons why HST should decrease air transport PAGEREF _Toc183114119 \h 92.2.1 Social advantages of the decrease in air transport PAGEREF _Toc183114120 \h 132.3 Overview on the arguments from literatures PAGEREF _Toc183114121 \h 143.Case Study on ICE / Frankfurt airport PAGEREF _Toc183114122 \h 163.1 Information on Frankfurt airport PAGEREF _Toc183114123 \h 163.2 Information on ICE PAGEREF _Toc183114124 \h 173.2.1 The travel time of ICE PAGEREF _Toc183114125 \h 193.2.2 The price of ICE PAGEREF _Toc183114126 \h 203.3 Impact of ICE on air transport PAGEREF _Toc183114127 \h 203.3.1 Impact on the accessibility of the airport PAGEREF _Toc183114128 \h 213.3.2 Impact on the air/rail integration PAGEREF _Toc183114129 \h 213.3.3 Impact on the short-haul flights PAGEREF _Toc183114130 \h 224.Case study on Fyra / Schiphol airport and the comparison between the two case studies PAGEREF _Toc183114131 \h 244.1 Information on Schiphol airport PAGEREF _Toc183114132 \h 244.2 Information on Fyra PAGEREF _Toc183114133 \h 254.2.1 The travel time of Fyra PAGEREF _Toc183114134 \h 264.2.2 The price of Fyra PAGEREF _Toc183114135 \h 274.3 Impact of Fyra on air transport PAGEREF _Toc183114136 \h 284.3.1 Impact on the accessibility of the airport PAGEREF _Toc183114137 \h 294.3.2 Impact on the air/rail integration PAGEREF _Toc183114138 \h 294.3.3 Impact on short-haul flights PAGEREF _Toc183114139 \h 294.4 Comparison between the two case study PAGEREF _Toc183114140 \h 314.4.1 Comparison between the two airports PAGEREF _Toc183114141 \h 314.4.2 Comparison between ICE and Fyra PAGEREF _Toc183114142 \h 325.Data Analysis PAGEREF _Toc183114143 \h 345.1 Methodology and Data selection PAGEREF _Toc183114144 \h 345.2 DID regression for the effect of HST with one single ICE dummy PAGEREF _Toc183114145 \h 375.3 DID regression for the effect of HST with separate ICE dummies PAGEREF _Toc183114146 \h 396.Conclusion PAGEREF _Toc183114147 \h 426.1 The effect ICE on air transport PAGEREF _Toc183114148 \h 436.2 Estimation of the future effect of Fyra PAGEREF _Toc183114149 \h 446.3 Limitations and future researches PAGEREF _Toc183114150 \h 447.Reference: PAGEREF _Toc183114151 \h 467.1 Literature: PAGEREF _Toc183114152 \h 467.2 Websites: PAGEREF _Toc183114153 \h 48Introduction1.1 IntroductionThe high-speed train services have been introduced in many countries during the past decades. It offers a faster and more efficient choice for the travellers compare to the traditional train services. Furthermore, many airports such as Schiphol airport, Frankfurt airport and Heathrow airport have introduced the high-speed train service to the airport to improve the intermodal transportation system at the airport and to increase the capacity of transporting air passengers to different destinations. The high-speed train service “Fyra” was introduced in September 2009 which runs between Amsterdam, Schiphol, Rotterdam and Breda. The whole route of Fyra is not yet completed; it will extend its service to Antwerp and Brussels in the near future. Furthermore, the operator of Fyra- NS Hispeed is also planning to increase the speed of Fyra in the near futuer. The Fyra service is new to Schiphol, so there have not been many discussions - both positive and negative about the possible effects of having Fyra on air transport. Thus, it would be interesting to see how Fyra would affect the air transport. Furthermore, it will be too limited to examine Fyra alone to investigate the impact of Fyra. Thus, it is necessary to choose another airport as a reference. In this case, Frankfurt airport will be a good choice. Frankfurt airport had longer history of high-speed train (the Intercity Express-ICE) services for more than 10 years and the integration of air and rail was considered successful. Furthermore, Frankfurt airport has the similar size and function as Schiphol airport. Therefore, it would also be interesting to compare the real case between ICE and Fyra. Learning from the past experiences of ICE and Frankfurt airport would give more evidences to forecast the possible impact of Fyra in the future. 1.2 Problem StatementAfter the introduction of Fyra to Schiphol, there have been quite a few debates on whether it is worth the huge amount of investment. Some argued that Fyra is a revolution in the traditional transport market and it would benefit the Schiphol airport while others believed it was a failure that it rarely had any actual influence on Schiphol. This paper is going to examine what are the possible effects of Fyra on air transport. 1.3 MethodologyThis paper will address the research question by literature review, case study and also data analysis. Firstly, the literature review would give a general discussion on the impacts of HST to air transport. Secondly, a case study on both ICE/Frankfurt airport and Fyra/Schiphol airport will give evidences on what happened with the two airports after HST was introduced. Thirdly, a difference-in-difference estimation will be done. The DID analysis estimates the net effect of a treatment to a certain group by comparing the treatment group to a control group. In this case, it could address the past effect of HST on air transport, and by comparing the control group Schiphol to the treatment group Frankfurt group we could estimate the future effect of Fyra on air transport. 1.4 List of ChaptersChapter 2 will discuss the different arguments from various literatures about the effect of HST on air transport. Following the literature review the case study on ICE/Frankfurt airport will be given in Chapter 3. Chapter 4 will continue with another case study on Fyra/Schiphol and the comparison between the two case studies. In Chapter 5, the methodology and result of the DID estimation will be presented. The paper will end with Chapter 6 which gives the conclusion of this study with a few limitations and possible future researches.Literature ReviewThere have been many discussions regarding the possible effects of HST on air transport. Some argued that it would increase air transport while other suggested the opposite way. The main arguments and reasoning for both sides will be discussed in this chapter.2.1 Reasons why HST should increase air transportSeveral literatures argued that HST would increase the air transport. Two main reasons were mentioned the most while researchers argued that the high-speed train could increase the air transport: the HST could improve the accessibility of the airport, thus attract more passengers; or it could also act as a complement for the airplane thus feed more passengers to the airport.According to Gelhusen and Wilken (2006), air passengers will tend to choose an airport with relatively good train service accessibility. The high-speed train brings the airport a better accessibility and connection and thus it would make the airport more preferred by the travellers. Similar arguments were also made by Lopez-Pita and Anton (2003), they claimed that the high-speed train services would bring more passengers to the airport and presented the case for the Lyon- Roissy and Satolas airport (now the Lyon - Saint-Exupery airport). After the high-speed train service started, it became much easier and faster for passengers to access the airport, and that had a huge impact on the passenger flows. For the year after the high-speed train service was connected to the airport, almost 3.5 percent of the air passengers took the high-speed train to access the airport, which is almost 1.5 million passengers. Furthermore, this effect is expected to develop in the future. Between 8.5% and 13.5% air passengers are expected to take the high-speed train by 2005, and high-speed train are anticipated to bring around 3 to 6 million more passengers to the airport every year. Other than improving the accessibility of the airport, the high-speed train service has also been claimed to have “complementary effect” to airplanes and airlines which could also increase the air transport. It was suggested that the high-speed train could offer an intermodal connection between air and rail, or act as a “feeder” for airplanes thus attract more travellers. As Terpstra and Lijesen (2011) stated, the high-speed train could be complementary to air services by acting as a “feeder” that brings passengers to the airports with a shorter travel time. Especially for the bigger airports, such as Madrid airport- the biggest airport in Spain, which gained 1.5% more market shares after the first year which the high-speed train was introduced. Givoni and Banister (2006) also mentioned the complementary effect of the high-speed train that it could bring air passengers to a hub airport instead of an airplane. He suggested that having high-speed train connections would make the airport a better choice for a hub airport. If the airport has a rather good railway access with a fast and convenient transfer option between rail and air, airlines could then include the high-speed train into their own services to transport air passengers from different locations without an airport access. The high-speed train services could also be brought into locations that already have airport accesses. Either way, the high-speed train could offer a faster access to the airport thus make the airport a more attractive choice. The airlines offering the rail-air connection services would also be a more attractive choice for the travellers. Moreover, Grimme (2007) suggested that the efficiency and easiness to access the airport brought by intermodal service between air and high-speed rail would make the airport more attractive to air passengers. In fact, airlines and travel agencies were already using the intermodal service as a marketing campaign. Train tickets and air tickets were sold together with a certain amount of discount.2.2 Reasons why HST should decrease air transportOther than the arguments that the high-speed train would increase the air transport, there are also arguments that the high-speed train could decrease the air transport. Various scholars suggested that the high-speed train has substitution effect on airplanes besides the complementary effect which would take away passengers from airplanes especially for the short-haul flights.While Givoni and Banister (2006) talked about how the airport could be benefit from the high-speed trains, Givoni (2006) also suggested that the high-speed train could be a perfect substitute of the airplane within short distances since it could offer less travel times. The high-speed train would take away some passengers from the airplane for the hub access. Since the high-speed train offers fast and direct service to the airport, passengers taking airplanes for transfer flights before may change their preference and switch to take the high-speed train. Grimme (2007) also claimed that the fast developing intermodal service between air and high-speed rail made high-speed rail an attractive substitute for short-haul flights. The high-speed train would encourage the air passengers to take the train other than a feeder flight within a short distance for transfer flights. Similar ideas were also found in various literatures. Gonza?lez-Savignat(2004) claimed that the high-speed train is becoming a great alternative for airplanes within a certain range of distances. He suggested that the high-speed trains would have a strong impact on the demand of air travel in the future, an important share of current air passengers will be attract to the new high-speed train services. Similar arguments were also expressed by de Rus and Nash (2007) that the high-speed train might decrease the air transport demand with a case study of the Spanish transport market. He introduced the case of the impact of high-speed train to the air demand for the Madrid-Seville which was part of the first high-speed line in Spain. The introduction of the Madrid-Seville line added another option for the passengers for the Madrid-Seville corridor. Among all the existing means of transportation for the Madrid-Seville line such as airplane and normal trains, the high-speed train had the lowest generalized cost but costs more travel time compare to the airplane. However, despite the extra travelling time, the high-speed train still took 50% passenger from the air services for the Madrid-Seville line. The airport Seville suffered almost 25% reduction in usage since it had a heavy amount of a air services for the Madrid-Seville line. Another paper wrote by López-Pita and Robusté (2005) also argued that as the travel time were considerably reduced for travelling by train after the high-speed train services were introduced, the railway again became more attractive to the passengers compare to the air services. Thus, there was a substantial change in the market share of air and rail as table 2.1 shows. Table 2.1 Railway Market Share Compared to Airline Market Share on Madrid—Seville Route Means of TransportMarket share without high-speed train (1992)Market share with high-speed train19941996199820002003Air7120.118.417.916.615.9Rail2979.981.682.183.484.1Source: López-Pita and Robusté (2005)Same as the Madrid-Seville line, another high-speed train line: Madrid –Barcelona which was brought into use few years after the Madrid -Seville line also had a significant impact on the air transport for these two destinations. As López-Pita and Robusté (2005) stated in his paper that the appearance of high-speed train for the Madrid-Barcelona would increase the market share of railway from 11% to nearly 53% to 63% percent in the near future. Consequently, the market share of airlines would drop rapidly from 89% to a much lower level around 36% to 47 %. Before the high-speed train was introduced, the airline was dominating the transport market, but the high-speed train took more than half of the passengers from the airlines which reduced their profit dramatically. Park and Ha (2006) argued that high-speed train is extremely competitive in the air transport market within a distance of 500 km which follows the evidence for the Seoul-Daegu high-speed rail line started in 2004. As shown in table 2.2, both the number of aircraft operations and number of passengers dropped around 72% only one year after the high-speed train (KTX-Korea Train Express) was introduced. Table 2.2 Changes of air traffic volumes on Seoul Gimpo and Daegu air routeSeoul-DaeguNo. of aircraft operationsNo. of passengersApril and May 20032180227,698April and May 200459963,315% of changes-72.5-71.3Source: Park and Ha (2006)Nevertheless, several important factors were mentioned that determine the competitiveness of high-speed trains comparing to aircrafts. Gonza?lez-Savignat(2004) suggested that the impact of the high-speed train services to air services will decrease as the travel distances and travel time increases. Thus, he suggested that the high-speed train would only remain absolutely competitive to airplanes when the travel time is less than three hours. For longer journeys, the airplanes will start to regain more and more competitiveness as the travel time and distance increases. López-Pita and Robusté (2005) also found out that as the journey time decreases the market share of railway would increase as a result (refer Figure 2.1). Figure 2.1. Effect of Train Journey Time on the Proportion of Air and Rail Travelers Source: López-Pita and Robusté (2005)Park and Ha (2006) conducted a survey that showed the competitiveness of high-speed train is negatively related to the ticket prices as table 2.3 shows. Furthermore, Gonza?lez-Savignat (2004) also claimed that the purpose of the travel also have an effect on the competitive of high-speed train. The leisure travellers will be more strongly affected by the ticket price than the business travellers. Thus, the price of the high-speed train might have a greater effect on the leisure travellers. Table 2.3 Probability of transport modal choice based on the high-speed train fares Fare of KTX (W)Probability of choosing airProbability of choosing high-speed train35,0000.1420.85840,0000.2020.79845,0000.2080.72050,0000.3740.626Source: Park and Ha (2006) 2.2.1 Social advantages of the decrease in air transportAlthough the high-speed train could have substitution effect on the air transport, it was suggested that the high-speed train taking away air passengers could also have positive impacts on the environment. Since the high-speed train is more environmental friendly than the air transport, it substituting air transport would result in less pollution. Nash (1991) suggested three environmental advantages for high-speed train compare to the airplane: less land consumption, less noise and less energy consumption. The construction of an airport need a lot of land for the runways, since the air transport demand is increasing every year, it is likely that there will be expansions of existing airports or constructions of new airports in the near future. As the high-speed train could be a complement of air transport and eases the pressure of airports, it could delay or even prevent the construction or expansions of airports and thus save the land also reduce the noise production. Furthermore, Nash claimed that high-speed trains consume less energy than aircrafts. On average, every one hundred passengers for aircrafts consume 5.8 liters of petroleum per kilometer when the aircrafts are fully loaded. On the other hands, every one hundred passengers for high-speed trains only consume 1.0 liters of petroleum per kilometer for the same situation. Givoni and Banister (2006) predicted there would be less emission as a result of the integration of rail and air. Both local air pollution and climate change situation will be improved. Evidences were given with the case of the emission reduction after the high-speed train services started on Heathrow-Paris route. Other than the environmental advantages, Givoni (2006) also suggested that as the high-speed train could free the runway capacity for long-haul flights by substituting the short-haul flights. For example, the high-speed train freed around 20% of the runway capacity for London Heathrow airport. This might improve the congestion problem at the airport. 2.3 Overview on the arguments from literaturesTable 2.4 gives a summary on the effects of HST on air transport suggested in different literatures.Table 2.4 Summary of the literaturesLiteratureResearch MethodEffect on air transportMain arguments/EvidencesLopez-Pita and Anton (2003)Descriptive /Case study+HST brings 3.5% of the air passengers (1.5 million) to the Lyon - Saint-Exupery airport after the first year, and 8.5%-13.5% air passengers (3-6 million) per year after five years.Terpstra and Lijesen (2011)Descriptive /Case study+Madrid airport gained 1.5% more market share after the first year HST was introduced. Givoni and Banister (2006)Descriptive /Case study+/-HST has both complementary and substitution effect on the air transport depends on the characteristics of flights/HST seized 71% of the market share on London-Paris route in 2005. Givoni (2006)Descriptive /Case study-High-speed trains have substitute effects to airplanes that might take away passengers from airlines/HST frees 20% of the runway in Heathrow. Grimme (2007)Descriptive /Case study+/-6.5% of the air passengers for Frankfurt airport used the Rail&Fly service /Air passengers dropped from 150,000 to 50,000 for the Cologne-Frankfurt line after HST was introduced.Gonza?lez-Savignat(2004)Survey/Demand Model-58% of the leisure travellers and 39% of the business travellers are expected to switch from airplanes to HST for Madrid-Barcelona line.; the competitiveness of high-speed train will reduce as travel distance and time increases.de Rus and Nash(2007)Descriptive /Case study-HST reduced 50% of the demand for Madrid-Sevilla route. López-Pita and Robusté (2005)Descriptive /Case study/Demand model-HST took 55% of passengers of the Madrid –Sevilla route from the airplane 10 years after it was built./ HST is estimated to take 42%-53% of the passengers from airplanes for Madrid-Barcelona line. Park and Ha (2006)SP - stated preference analysis and OLS-The analysis estimated 14% passengers would choose air over HST train while the actual data showed 28% passengers would prefer air transport.As the table 2.4 shows, both positive and negative effects of high-speed train on air transport were mentioned. Some scholars even mentioned both positive and negative effects in one paper concerning different situations. Furthermore, most of the researches done regarding the effect of high-speed train on air transport were only case studies and descriptive reasoning. Moreover, most of the studies were only concerning about the effect of HST on one specific air route. There were only few literatures discussed about the overall effect of HST on air transport. Thus, it is not possible to determine the effect of Fyra on air transport just based on the literature review. Case Study on ICE / Frankfurt airportIn this chapter, a case study will be given for ICE and Frankfurt airport. 3.1 Information on Frankfurt airportFrankfurt airport locates about 12 kilometers away from the city center. It is the third busiest airport in Europe and the biggest airport in Germany ranked by yearly passengers. According to the data from Airport Council International, Frankfurt airport had 53,009,221 passengers (both terminal and transit) in year 2010 and 50,932,840 passengers in year 2009. The growth in passenger numbers was around 4.1% for year 2009-1010. More than 50% of the passengers of Frankfurt airport are transfer passengers, which makes Frankfurt airport a big hub airport (Gelhusen and Wilken, 2006). Around 115 airlines run at Frankfurt airport. It is also the primary hub and operation base for Lufthansa airline- the biggest and flag airline in Germany. Before the high-speed train (ICE) was introduced in Frankfurt airport, the only train service at the airport was the RegionalBahn (a type of local stop train) and S-Bahn (a railway system combines the city center and suburban areas). Limited RegionalBahn trains were running through Frankfurt airport, connecting the airport to Frankfurt city center and some of the nearby cities such as Mainz (about 30km away from the airport) and Koblenz (around 120 km away from the airport). The farthest city the regional train from the airport connects is Saarbrücken which is about 180 kilometers away from the airport. The regional stop trains normally runs on an hourly base, or with even less frequency. On the other hand, the S-Bahn connects the airport to closer destinations such as Wiesbaden (about 30km away from the airport) and Offenbach (about 20 km away from the airport) with a higher frequency (per 15 minutes). Other than the train connections, the Frankfurt airport is also accessible by other means of transportation such as car, bus and taxi. Frankfurt airport is connected to various highways including A3, A67 and A5. It takes about 10-15 minutes to drive from the airport to Frankfurt city center. There are also bus connections to some close destinations. 3.2 Information on ICETo improve the accessibility of the airport, and to meet the increasing demand for the airport transportation, the Frankfurt am Main Flughafen Fernbahnhof (Frankfurt Airport long distance railway station) was planned. It was opened in the end on 27 May 1999 which brought the high-speed train service –Intercity Express (ICE) to Frankfurt Airport. ICE is a high-speed train service running by the Deutsche Bahn (DB), the German national railway operator. ICE offers high-speed train services from Frankfurt airport to various destinations in Germany and neighbor countries. Table 3.1 lists all the ICE connections through Frankfurt airport and the main cities they stop by at the moment. Figure 3.1 shows the ICE route for 2011, the red-circled stop is the Frankfurt airport long-distance station. There were fewer lines and frequencies when the ICE service was first introduced. Originally there were only two ICE lines connects to cities farthest to Hannover, Stuttgart and Nurnberg. Table 3.1 Main cities on the train route of ICE LineMain cities on the ICE routeICE 20Hamburg; Hannover; Frankfurt; Mannheim; Basel; (Zurich)ICE 22(Kiel) Hamburg; Hannover; Frankfurt; Mannheim; (Heidelberg); Stuttgart ICE 31Kiel; Hamburg; Dortmund; Duisburg; Dusseldorf; K?ln; Koblenz; Frankfurt; Mannheim; Basel SBBICE 41(Dortmund); Essen; Düsseldorf; Duisburg; Cologne; Frankfurt;; Nurnberg; Munich ICE 42Dortmund; Duisburg; Dusseldorf; Cologne; Stuttgart; Munich ICE 43Cologne; Mannheim; Freiburg; Basel?ICE 49Cologne; Limburg; Frankfurt ICE 50Dresden; Leipzig; Frankfurt; (Wiesbaden)ICE 78Amsterdam; Duisburg; DusseldorfICE 79Brussels; Aachen; Cologne; FrankfurtICE 91(Dortmund; Cologne); Frankfurt; Nurnberg; Vienna(Source: DB Bahn)Figure 3.1 The ICE route for 2011Source: MFO.de3.2.1 The travel time of ICEThe high-speed train service shortens the travel time from various destinations to the airport. Table 3.2 shows the travel time from the airport to various destinations before and after ICE was introduced. Furthermore, the travel time does not include the waiting time for transfer. As there was no direct connections to the destinations before ICE was introduced, passengers has to stop at Frankfurt main station first and then wait for the train services to their final destinations, some of the journey even requires more than one changes. Thus, account for the waiting time, ICE could actually save more time than table 3.2 shows. Table 3.2 Travel Times before and after the high-speed train started its service at Frankfurt Airport DestinationTravel time (BEFORE)Travel time (AFTER)Amsterdam (the Netherlands)4h 45min3h 56minBasel(Switzerland)3h 00min2h 53minBrussels(Belgium)5h 00 min2h 52minLondon (UK)7h 30min5h 30min (BY Eurostar)Munich (Germany)3h 55min3h 32minHanover (Germany)3h 15min2h 35minStuttgart(Germany)1h 25min1h 13minBonn/Sieburg (Germany)1h 48min0h 40minCologne (Germany)2h 00min0h 40 minSource: Payne (1999) and DB website3.2.2 The price of ICEThe price of ICE is higher than the normal train. Table 3.3 shows the price difference between the normal trains and the high-speed train (ICE) from airport to different destinations. It is obvious that passengers need to pay much more to take the ICE train than the normal ones. Especially for the domestic passengers, the differences in ticket costs are significant even with double amounts, comparing to the amount of time saved (refer Table 3.2). Thus, passengers may still choose the normal/ local train over the high-speed train because of the higher price of the high-speed train.Table 3.3 Price difference between normal train and ICE in EURO priceDestinationTicket Price (Normal Train)Ticket Price (high-speed Train)Amsterdam (the Netherlands)79111.20Basel(Switzerland)7250.9Brussels(Belgium)10182Munich (Germany)4291Hanover (Germany)4283Stuttgart(Germany)33.6057Bonn/Sieburg (Germany)30.1068Cologne33.8061Source: DB website3.3 Impact of ICE on air transportHigh-speed trains are suggested to have various impacts on air transport. Most frequently mentioned effects are on the accessibility of airport, air/rail integration and short-haul flights. This section will access the impact of ICE on air transport based on these three aspects.3.3.1 Impact on the accessibility of the airportThe new Frankfurt Airport long distance railway station and the high-speed service offers better access to the airport from various locations in Germany. As mentioned earlier, the airport was not very well connected to train services before the new long distance railway station was built and ICE was introduced. It was very inconvenient and inefficient for the passengers to travel to the airport by train especially if they have much luggage. However, after the Frankfurt Airport long distance railway station was built, air passengers can connect their flight directly to high-speed trains bringing them to various destinations including most of the major cities in Germany and even neighbor countries such as Netherlands and Switzerland. Similarly, it makes it easier and faster for passengers to access the airport from different locations. The improved accessibility would help Frankfurt Airport to become a global gateway. Payne (1999) believed an outstanding and competitive transport infrastructure is crucial for an airport in the fast-growing globalism world. The good connection between air and rail would encourage Frankfurt airport to become more important and competitive in the European transport network.3.3.2 Impact on the air/rail integrationThe introduction of ICE at Frankfurt airport enabled the integration between high-speed train and airplanes, which developed the intermodal system between rail and air. According to Freitag (2006) ,about 83 airlines and 42 package tour operators were selling packages of air tickets and train tickets together in 2006 , the passengers buying this kind of package would get a discount on the train ticket price . This co-called Rail&Fly package had attracted 1.6 milliin passengers in 2005 which is about 6.5% of all passengers departing from Frankfurt Airport. This trend is expected to be continuously increasing in the future. Furthermore, Lufthansa airline took the advantage of the high-speed train and offered the AIRail service with Deutsche Bahn AG and Fraport AG together. According to the information on Frankfurt Airport Website, the AIRail system coordinates the train schedule with the Lufthansa flight and made the check-in procedure faster with Lufthansa check-in counters at the train station in Cologne, Siegburg/Bonn or Stuttgart. The luggage can also be checked through the counters at the train statin. A total of 27 airlines had made agreements with Lufthansa to use the AIRail service to attract customers. The integration between high-speed rail and air definitely offers Frankfurt airport a competitive advantage while competing with other airports for long-distance flights and intercontinental flights. For example, Lufthansa can now compete with KLM for flights to China as they could offer Rail&Fly packages directly from Amsterdam. Thus, the ICE could bring the airport more passengers. 3.3.3 Impact on the short-haul flightsAs ICE shortens the travel time, it will become a strong competitor for the airline, especially for the short-haul flights. Taking account into all the extra activities needed for taking the airplane such as check-in, security-check and boarding, most of the passengers will choose ICE over the plane for short-haul destinations as it is faster and more convenient. Furthermore, as the introduction of Rail&Fly ticket, and the AIRail service, it is possible that the airlines may use high-speed train to substitute the short-haul flights as a feeder thus cut some of the flights. The case study for Cologne-Frankfurt line will discuss more on this matter. The Cologne Frankfurt high-speed rail line taking away passengers from airplane could be a very good example for this effect.Example: The Cologne-Frankfurt high-speed rail lineIn August 2002, the Cologne-Frankfurt high-speed rail line was officially brought into use. It is the fastest high-speed line in Germany connecting Cologne and Frankfurt. The creation of this high-speed rail line shortened the travel time between Cologne and Frankfurt Airport from two hours to only 50 minutes. Thus, the ICE became a more attractive option than the airplane thus took away a significant amount of passengers. Figure 3.2 shows the difference between the number of seats offered and the number of passengers actually flying between Cologne and Frankfurt from 1989 to 2006. Figure 3.2 the difference between the number of seats offered and the number of passengers actually flying between Cologne and FrankfurtSource: Grimme (2007) It is quite obvious from the graph that after the introduction of Cologne-Frankfurt high-speed rail line, the demand for air services between Cologne and Frankfurt dropped dramatically. The number of seats taken was further reduced after the AIRail service was introduced in 2003 (Grimme, 2007). The number of seats offered was decreased from more than 250,000 before 2002 to around only 100,000 in 2003, and with further reduction afterwards. The number of air passengers reduced from around 160,000 before to less than 50,000 in 2003. According to Grimme (2007), the frequency of the air service per day shrunk from a maximum of 7 aircrafts per day to a maximum of 4 aircrafts per day and the size of the aircrafts was also reduced. Furthermore, one can tell from the trend that the air passenger will still be falling in the future. This example showed that the high-speed rail could definitely attract some of the passengers from air services. The number of passengers shifted from air to rail service depends on several factors, mostly the time saved, and also efficiency, price, etc.Case study on Fyra / Schiphol airport and the comparison between the two case studiesIn this chapter, firstly a case study on Fyra and Schiphol airport will be given, and then it will continue with the comparison between the two case studies. 4.1 Information on Schiphol airportSchiphol airport locates around 9 kilometers away from Amsterdam city center. It is the fifth busiest airport in Europe and the biggest and dominate airport in the Netherlands according to the yearly passengers figures from Airport Council International. Schiphol had 45,211,749 terminal and transit passengers in 2010 and 43,570,370 passengers in 2009. There was a 3.8% growth in passenger numbers from year 2009 to 2010. Around 40% of the passengers of Schiphol airport are transfer passengers. There are over 100 airlines operating in Schiphol airport. It is also the hub and operation base for the Royal Dutch airline-KLM. Schiphol Airport already had fairly good train connections before the high-speed train “Fyra” was introduced. It is located on the major train lines of the Netherlands thus there was both a considerable amount of intercity (faster train) and stop trains (slower train) going to most of the major cities or small towns in the Netherlands. Each hour there are frequent train services between Schiphol and the big cities such as Amsterdam and Rotterdam, there are ticket selling points just outside the airport thus it is really convenient for the air passengers to access or leave the Schiphol Airport. Apart from the train service, Schiphol airport was also accessible by other means of transportation such as car and bus. The airport is connected to the motorways A4 and A9, and there are several buses leaving airports every hour to different destination around the airport. 4.2 Information on FyraThe Fyra train service is a high-speed train service planned by the NS (Nederlandse Spoorwegen, the Dutch Railway Company) that runs between Amsterdam, Schiphol Airport, Rotterdam, Breda, Antwerp and Brussels. It was firstly launched in September 2009. The launch of Fyra was the first time that there has been such a fast and efficient connection between Schiphol Airport and some of the major cities in Rotterdam. NS Hispeed claimed that Fyra is a more environmental friendly and energy efficient way than travelling by cars and normal train and it was aiming to attract a large number of travellers and commuters to switch their travelling choice with better service and efficiency (NS Hispeed factsheets, 2009). Fyra only runs between Amsterdam, Schiphol airport and Rotterdam when it was firstly introduced. The service was extended to Breda in April 2011 and the service to Antwerp and Brussels is expected to be available in mid-2012. The NS is also planning to upgrade the train and accelerate the speed in the near future. Figure 4.1 shows the operation route of Fyra after it is completed. Figure 4.1 operation route of FyraSource: Hispeed.nl4.2.1 The travel time of FyraAfter the Fyra was introduced, travel time between Schiphol Airport and its destinations have been shortened. As table 4.1 shows, the travel time between Schiphol and Rotterdam/Breda was almost reduced by half but the travel time between Schiphol and Amsterdam still remained the same. This is quite reasonable considering that there was already a fast intercity connection between Amsterdam and Schiphol Airport. Furthermore, Amsterdam and Schiphol Airport is only 9 kilometers apart thus it is quite hard for the trains to raise its speed. The intercity between Rotterdam and Schiphol Airport have two more stops then the Fyra (Leiden and Den Haag), and to go from Schiphol to Breda passengers need to change the train once in Den Haag, thus shortened travel time between Schiphol Airport and these two cities can be saved from the stop times and transfer times. The current top speed for Fyra is 160 km/h, and after the acceleration of the train speed, the top speed will increase to 250km/h which makes it even faster to travel between Schiphol and Den Haag/Breda (refer Table 4.2). However, the travel time between Amsterdam and Schiphol will still remain the same. Table 4.1 Travel time difference with and without Fyra DestinationTravel time (by intercity)Travel time (by Fyra)Amsterdam17 min17 minRotterdam46 min26 minBreda1h 36 min 53 minSource: NS.nl and Hispeed.nlTable 4.2 Travel time difference with and without Fyra(after speed upgrade) DestinationTravel time (by intercity)Travel time (by Fyra)Amsterdam17 min17 minRotterdam46 min19 minBreda1h 36 min 41 minAntwerp1h 05 min31minBrussels2h 35 min1h 30minSource: NS.nl and Hispeed.nl4.2.2 The price of FyraFyra has a higher price than the normal train. A supplement ticket is needed when the passenger travels with Fyra besides the normal train ticket. When Fyra was introduced, NS Hispeed set up a relatively high supplement fair (refer Table 4.3). Table 4.3 The price of normal train and Fyra in 2009DestinationPrice 2nd class (Normal train)Price 1st class (Normal Train)Price 2nd class (Fyra)Price 1st class (Fyra)Supplement 2nd class (Fyra)Supplement1st class (Fyra)Amsterdam4.006.706.109.302.102.60Rotterdam10.9018.5017.0026.206.107.70(Source: treinreiziger.nl)However, after Fyra was introduced, the result was not as good as predicted. The number of passengers was far away behind expectation. Most of the time Fyra only gets a 50% load on second class while the first class is almost empty, and that is the better scenario. The NS suffered a serious financial difficulty due to the low income of Fyra and they decided to lower the supplement fare of Fyra since the 1st of February 2011. (NOS.nl) As Table 4.4 shows, the supplement fare reduced about 60% compare to the fares before the price cut. However, it is still questionable whether the passengers would be willing to pay extra fairs to save not much time, especially when the normal intercity train has higher frequency than Fyra. Furthermore, it would be almost certain that the passenger would choose the travel by Fyra between Amsterdam and Schiphol since they have the same travel time but Fyra is more expensive than the normal trains. Table 4.4 The price of normal train and Fyra in 2011 after the price cutDestinationPrice 2nd class (Normal train)Price 1st class (Normal Train)Price 2nd class (Fyra)Price 1st class (Fyra)Supplement 2nd class (Fyra)Supplement1st class (Fyra)Amsterdam3.706.304.407.300.701.00Rotterdam10.7018.2012.8021.002.102.80Breda16.9028.7020.3033.103.404.40(Source: Hispped.nl and NS.nl)4.3 Impact of Fyra on air transportThis section will address the impact of Fyra on air transport based on the accessibility of airport, air/rail integration and short-haul flights.4.3.1 Impact on the accessibility of the airportThe Fyra was not likely to have a significant impact on the accessibility of Schiphol airport considering the fact that Schiphol already had very good train connections. As previously mentioned in 4.1, the train connection at Schiphol airport was already quite sophisticated and mature before Fyra was introduced. Passengers could access Schiphol airport by intercity or stop train from various locations in the Netherlands. There were already direct train connections from Schiphol airport to the major cities, such as Rotterdam and Amsterdam. As Fyra was planned for the big cities which already had a good and comparatively fast (refer Table 4.1) train service, the introduction of Fyra did not have a significant impact on the development of airport. Since there are also direct train service to Antwerp and Brussels, even after the Fyra extend its service to these two cities, it is still not expected to have any major affects over the development of Schiphol airport. 4.3.2 Impact on the air/rail integrationFyra did not really improve the intermodal service of air and rail. There is currently no Air&Rail package offered by any airlines or travel agencies in Schiphol since Fyra only connects short distance destinations. Furthermore, the air and rail integration was already quite convenient before Fyra was introduced, and most of the passengers still choose to access the airport by normal trains. Therefore, Fyra did not make a big difference. 4.3.3 Impact on short-haul flightsFyra may not have any significant effect on short-haul flights at all. At the moment, Fyra only have three destinations: Amsterdam, Rotterdam and Breda. None of those destinations would attract many passengers to travel by airplane, not to mention that Breda does not even have an airport. However, there are currently other kinds of high-speed trains running in Schiphol airport. Considering Fyra will extend its service to Brussels and Antwerp, it is worth take a look at the effects of those high-speed trains on short-haul flights. There are currently several numbers of Thalys running through Schiphol each day connecting to Antwerp, Brussels and Paris. Furthermore, passengers could access the ICE service at Amsterdam Central station that connects to Germany cities such as Dusseldorf and Frankfurt. Table 4.5 shows the travel time difference between high-speed train, normal train and airplane (including the time needed for all the processes to get on the airplane such as check-in, boarding, to and go from the city center) from Amsterdam central station to the destinations that can be reached by high-speed train at the moment. Comparing to travel by airplane, travelling by high-speed train saves only a little amount of time, or nearly same amount of time. Count in all the other factors discussed before that might affect the travellers’ decision- less steps required (Check-in, boarding, travel to/out of the city center, etc) and cheaper price, it is reasonable to estimate that some of the air passengers will start to travel by the high-speed train instead of airplane. Table 4.5 travel time from Amsterdam Central Station to HST destinationsDestinationTravel time (High speed train 2010)Travel time (Normal Train)Travel time (Airplane)Brussels1h 44min2h 51min2h 30minParis3h 13 min4h 09min3h 30minFrankfurt3h 56min3h 56min4h 10minSource: Terpstra and Lijesen (2011)Jorritsma (2010) stated in his paper that 5.7 million journeys to France and Belgium was made in 2010 after the Thalys was introduced according to NS Internationaal. Jorritsma (2010) also mentioned that Muconsult (2007) used a logistic regression model to estimate to possible impact of high-speed train competing with airplane. He estimated that around 1.6 million potential passengers would choose to take high-speed train instead in 2020 for the routes Amsterdam-Paris/London/Brussels. About 16,000 flights per year (assuming 100 passengers per flight) will be reduced on Schiphol in 2020 which is approximately 2.5% of the total flights operated at Schiphol.4.4 Comparison between the two case study4.4.1 Comparison between the two airportsFrankfurt and Schiphol airport have the similar situations in terms of size, function and growth rate (refer Table 4.6): they have the similar yearly passenger numbers; they are both the biggest airports in its countries; they are only two places different in the airport ranking; they are both the hubs for the national airlines; around half of the passengers were transfer passengers. Thus, the two airports are quite comparable. Furthermore, the two airports had the similar accessibility by car and bus before HST was introduced. However, the two had different conditions concerning the accessibility by train before the high-speed train was introduced: Schiphol airport had better train connections (refer Table 4.7) Considering Frankfurt had the ICE in 1999 and Schiphol only had Fyra in 2008, it is quite reasonable that the Frankfurt would have a worse train connection back then, but this difference should be taken into consideration in the later discussion and analysis. Table 4.6 Summary of the situation of Frankfurt and Schiphol airportAirportAirport rankingYearly Passenger (2010)Growth rate (2009-2010)Hub airport for airlines?Percentage transfer passengersFrankfurt353,009,2214.1%Yes (Lufthansa)50%Schiphol545,211,7493.8%Yes (KLM)42%Table 4.7 Accessibility of the airport before HSTAirportAccessibility (Car)Accessibility (Bus)Accessibility (Train)FrankfurtConnects to 3 highwaysFair amount and frequency of buses to near destinationsOnly RegionalBahn and S-BahnSchipholConnects to 2 highwaysFair amount and frequency of buses to near destinationsBoth intercity and stop trains.4.4.2 Comparison between ICE and FyraICE and Fyra both shortened the travel time from various destinations to the airports while ICE had a stronger effect on travel time than Fyra. Furthermore, they both have higher prices than the normal train. However, there are also quite a few differences between ICE and Fyra. Firstly, ICE connects to both domestic and international destinations while Fyra only runs inside the Netherlands. Moreover, ICE has dozens of various destinations in every part of Germany while Fyra only connects to three destinations. Secondly, Ice had a positive influence on both accessibility of the airport and air/rail integration but Fyra had almost no impact on these two aspects. Finally, ICE took a considerable amount of air passengers from short-haul flights. On the other hand, Fyra did not have significant effects on short-haul flights since the destination of Fyra were too close to the airport for the passengers to take airplanes in the first place. Nevertheless, when Fyra extend its destinations to Antwerp and Brussels it might have some impact on short-haul flights. Table 4.8 Summary of the impact of HST on different aspectsHigh-speed trainTravel time Travel costAccessibility of the airportAir/rail integrationShort-haul flightsICEShortenedIncreasedImprovedImprovedTook a significant amount of passengers awayFyraLess shortenedIncreasedNo significant impactNo significant impactNo significant impactThe case study could give us some insights on what happened to the airports after the high-speed train was launched. However, it is still quite limited to conclude the effects of high-speed trains just based on these facts. Therefore, it is necessary to collect some actual data about the two airports and perform some analysis to dig deeper into the real scenarios. Data Analysis5.1 Methodology and Data selectionTwo indicators are chosen to examine the effect of high-speed trains on airports: air transport movements and monthly air passengers since they represent the most direct effect on the passenger flows and airlines. To evaluate the change in the indicators before and after the high-speed train was introduced, difference-in difference (DID) estimation would be the most appropriate method. The difference-in difference estimator evaluates the impact of a treatment (in this case having the high-speed train service) by comparing the outcomes for two groups. Group 1 (the treatment group) is exposed to a treatment in the second period but not the first period and group 2 (the control group) is not exposed to the treatment for both time periods. In this case, Schiphol will be the control group for Frankfurt since it only had less than two years of high-speed train service. By comparing the outcomes (the two indicators) for Schiphol and Frankfurt airport, we will be able to see how high-speed train affected the air transport in the long run, thus estimate Fyra it would affect the air transport in the future.The model of DID estimation is given by the following equation (Albouy, D, n.d.):Yi =α+βTi +γti +δ(Ti ·ti)+εi Where Yi is the outcome of the treatment (two indicators), Ti is the dummy that group dummy and β accounts the average permanent differences between the two groups. ti is the time dummy and γ tells the common time trend for the two groups. (Ti ·ti) is the interaction variable for the group dummy and time dummy, and δ represents the actual effect of the treatment (high-speed train). By performing a regression analysis, we could get the DID estimator. The two indicators will be treated as dependent variables to see how they would react on the changes in the high-speed train services and time. A natural logarithm will be taken on both of the dependent variables to remove heteroscedasticity problem for the regression. Since the values of the indicators are too big, the variance of the error term for the regression analysis would not be constant and thus the standard errors for the test will be biased. Taking the natural logarithm of the variables would smaller the variances and thus makes the analysis easier and more accurate. Seven independent variables will be included in different regressions based on different purpose:Frankfurt Airport dummy: this variable will evaluate the difference between Frankfurt airport and Schiphol airport regards the indicators in average. ICE dummy (later noted as ICE): this variable will show how the ICE service would affect the dependent variables.Fyra dummy (later noted as Fyra): this variable will show how the Fyra service would affect the dependent variables. This dummy shows the immediate effect of Fyra. ICE dummy * time (later noted as ICE*date): this is an interaction variable of ICE dummy and dates. It shows the continuous effect of ICE on the dependent variables. Fyra dummy * time (later noted as Fyra*date): this is an interaction variable of Fyra dummy and dates. It shows the continuous effect of Fyra on the dependent variables. Time dummy for each year (later noted as i.year): it would show how the dependent variables would change along the years, thus exclude the effect of year on the indicators. The first year 1980 will be used as a reference.Time dummy for each month (later noted as i.month): it would show the trend of the dependent variables for different months, thus exclude the effect of year on the indicators. January of each year will be used as a reference.To check if all the independent variables can be put in the same regression analysis, first a correlation between the variables should be tested. As a result, the correlation between ICE and ICE*date; Fyra and Fyra*date is too high to be put in the same regression, thus two different regressions should be done regarding immediate effects and continuous effects. Table 5.1 summarizes the independent variables and dependent variables for each regression. Table 5.1 Dependent and independent variables chosen for the regression analysis Immediate or Continuous effect Dependent variableIndependent variables (ICE effect for the whole time period)Independent variables (ICE effect for separate time periods)Immediate effectLog Air Transport MovementsFrankfurt airport dummy , ICE, Fyra , i.year, i.monthFrankfurt airport dummy , ICE 1st yr, ICE 2nd yr, ICE 3rd year onwards, Fyra , i.year, i.monthLog Monthly air passengersContinuous effectLog Air Transport MovementsFrankfurt airport dummy , ICE *date, Fyra *date, i.year, i.monthFrankfurt airport dummy , ICE 1st yr, ICE 2nd yr, ICE 3rd year onwards*date, Fyra , i.year, i.monthLog Monthly air passengersFurthermore, in this case with the Schiphol and Frankfurt airport, the situation is a little different from the theory. As Schiphol, the control group already had Fyra for two years, and there are no control group with no HST connection, the previous DID estimation may not be very solid. Thus, a different regression could be done to improve the above model. To access the possible future effect of Fyra, we could separate the ICE dummy into dummies for each year during the first few years. Thus, there will be five dummies for ICE instead of one in the following regression tests: ICE 1st year, ICE 2nd year, ICE 3rd year, ICE 4th year and ICE 5th year and onwards. This will enable us to examine the effect of ICE for the first four years separately and for the fifth year and onwards. Therefore, we can predict the possible future effect of Fyra according to the trend of the change in effects of ICE on indicators. Regressions with one single ICE dummy to examine the overall effect of ICE and possibility the effect of Fyra on the indicators will be discussed in section 5.2. And regressions with separate ICE dummies will be discussed in section 5.3. 5.2 DID regression for the effect of HST with one single ICE dummyFour different regression analyses were done with single ICE dummy, to examine both the immediate and continuous effect of HST on the two indicators. The results are summarized in Table 5.2. The first number in each cell is the coefficient for the various dummies. The “*” represents significance level of the dummies where “***” means that the dummy is significant at a 0.1% level, “**” means that the dummy is significant at a 1% level and “*” means that the dummy is significant at a 5% level. The numbers in the brackets are the robust standard error. Table 5.2 Summary of the result of the DID regression (with single ICE dummy)Air transport movement (immediate effect)Air transport movement (continuous effect)Number of monthly passengers(immediate effect)Number of monthly passengers (continuous effect)Frankfurt airport0.220***(0.0082)0.220***(0.0080)0.371***(0.0101)0.374***(0.0099)ICE -0.080***(0.009)-0.190***(0.012)ICE*date-0.00015***(0.000017)-0.00036***(0.000021)Fyra-0.050 **(0.016)No significant effectFyra*date-0.000097***(0.000279)No significant effectConstant9.53***(0.014)9.52***(0.013)13.62***(0.015)13.62***(0.015)R20.9800.9800.9850.985Number of observations: 600Note that the year dummy and month dummy are also included in the regression analysis, the result is not shown in the table since they are not very relevant for the discussion. Furthermore, the test results for Schiphol airport maybe biased since there is no control group with no HST connection. However, these results will still be discussed just as a reference.For the continuous effect of HST on air transport movement, we can conclude that having a high–speed train service would having a slightly negative continuous effect for both Frankfurt (-0.015%) and Schiphol Airport (-0.0097%). Furthermore, ICE and Fyra also have a negative immediate effect on the air transport movement, with the coefficient for ICE around -7.95% and the coefficient for fyra around -5.06%. Overall, the regression analysis showed a negative relationship between having a high-speed train and air transport movements for both airports. This might be resulted by airlines cutting short-haul flights. For both Fyra and ICE, the absolute values of the coefficients are quite small, indicating that the air transport movements are only slightly influenced by the high-speed train. The possible explanation could be that the airlines may launch new air (long-haul) routes as the cut of short-haul flights frees runway capacity. Furthermore, the air transport movement data also contains data for cargo planes and it won’t be affected by the high-speed train. The test result for effects on Fyra on air transport movement is significant, although the coefficient is very small. This result is not consistent with what expected in the case study. However, the air transport movement data is not for passenger transportation alone, and Schiphol also had other high-speed train running (Thalys) at the same time. All these elements would have affect the result of the analysis. Thus, it could still be the case that Fyra does not have very much effect on air transport movements for passengers. Nevertheless, the air transport movement contains information on both cargo transportation and passenger transportation. Since the cargo transportation is unlikely to be affected by the high-speed train, it might be more relevant to take a look at the other indicator: monthly air passengers to address the effect of HST on air transportation. The regression result suggested that ICE has both negative continuous (-0.036%) and immediate (-19%) effects on the number of monthly air passengers while Fyra does not have a significant impact on the number of monthly passengers.The regression result showed a bigger effect of ICE on air transport movement for the first two year than the time period after the third year. Furthermore, the immediate effect of the whole period (-7.95%) is a bit stronger than the immediate effect from third year onwards (-7.52%).The ICE had stronger effect on the number of monthly air passengers than the air transport movement, proves the number of monthly air passengers is a more relevant indicator since it only shows the outcome on the air passengers. The main reason of the negative relationship between ICE and monthly air passenger numbers is that the ICE would take away passengers from short-haul flights. On the other hand, Fyra does not have significant effect on the number of monthly air passenger. It is consistent with the estimation in the case study. There are two possible reasons to explain this result: firstly, Fyra will not take passengers away from the short-haul flights. Secondly, the passengers may still choose to take normal trains other than Fyra since the normal trains offers similar time and lower price. 5.3 DID regression for the effect of HST with separate ICE dummiesAnother four different regression analyses were done in this section to examine the effect of ICE for the each of the first few years and also long-term effect of ICE. By comparing to the trend in the effect of ICE we can thus predict the future effect of Fyra. The results are summarized in Table 5.3. The first number in each cell is the coefficient for the various dummies. All the dummies are significant at a 0.1% level and are represented by “***”. The numbers in the brackets are the robust standard error. Table 5.3 Summary of the result of the DID regression (with separate ICE dummies)Air transport movement (immediate effect)Air transport movement (continuous effect)Number of monthly passengers(immediate effect)Number of monthly passengers (continuous effect)Frankfurt airport0.220***(0.0082)0.220***(0.0082)0.373***(0.0098)0.373***(0.0098)ICE 1st year-0.089***(0.013)-0.090***(0.013)-0.117***(0.016)-0.118***(0.016)ICE 2nd year-0.108***(0.012)-0.108***(0.012)-0.122***(0.017)-0.123***(0.017)ICE 3rd year-0.114***(0.13)-0.114***(0.13)-0.199***(0.019)-0.199***(0.019)ICE 4th year-0.071***(0.014)-0.071***(0.014)-0.203***(0.017)-0.201***(0.017)ICE 5th year and onwards-0.069***(0.0094)-0.215***(0.012)ICE 5th year and onwards*date-0.00012***(0.0000017)-0.00039***(0.000020)Constant9.52***(0.013)9.52***(0.013)13.62***(0.015)13.62***(0.015)R20.9800.9800.9860.986Number of observations: 600Note that the independent variables Fyra, Fyra*date, year dummy and month dummy were also included in the regression, but the results are not reported in the table since they are not quite relevant to the analysis. Furthermore, the ICE dummies for each of the first four years will not be tested for continuous effect since they only represent the difference for one year. Thus, it would not not comparable to the continuous effect on the dummy “ICE 5th year and onwards”. Therefore, the following analysis will be focused on the coefficients for immediate effect to observe a trend in the change in effects of ICE. The regression result for immediate effect of ICE on immediate effect for different time periods showed an increase trend for negative effects during the first three years (from -9% to -11.4%), and then from the fourth year after ICE was introduced, the strength of the negative effects appeared a decreasing trend (from -11.4% to -6.9%). Possible explanation for the trend could be that airlines react to the high-speed train firstly by cutting the short-haul flights, but later they might announce new air routes. The same trend can be applied to the Fyra case in the future after it extends its service to Brussels. There might be an increase trend of the effect of Fyra in the beginning, after few years the effect will be less strong. Furthermore, the effect of Fyra might be less strong than ICE since Fyra has much less destinations than ICE. However, as explained earlier in section 5.2, the air transport movement is not the best indicator for the change in air transport as it contains information on both cargo transportation and passenger transportation. Thus, it might be more relevant to look at the other indicator.The regression result showed an increasing trend in the negative effect of ICE on monthly air passenger numbers from the first year to the fifth year and onwards (from -11.7% to -22.5%). Furthermore, the increasing trend in the strength of negative effect became more and more steady as time passes. Three possible reasons could cause the increasing trend in the effect of ICE on monthly air passengers. Firstly, the passengers might need time to get to know and adapt the new ICE train. Secondly, the airlines might also need time to improve the intermodal system. Thirdly, the ICE train developed more and more destinations and routes over time, thus might take away more passengers as it expand its services. Furthermore, after the passengers and airlines have been fully adapted to the new ICE service, and after the ICE network became mature, the effect of ICE will become less violent. This might explain why the growth in the negative effect of ICE on air passengers became steadier after a few years.The future effect of Fyra on monthly air passengers is very likely to have the same trend as ICE. The effect might become stronger in the next few years. Furthermore, after Fyra expands to Brussels and Antwerp in the future, the effect of Fyra are expected to have a significant growth. However, considering that Fyra much less destinations, the effect of Fyra on passengers will not be as strong as ICE. ConclusionThere have been many discussions about the effect of Fyra on air transport but there are no solid conclusions yet. This paper examines the possible effect of Fyra on air transport from different aspects. The literature review offered different point of views on the effects of high-speed trains. Some argued that it would increase the air transport while others argue that it would decrease the air transport. Furthermore, there are also arguments that even though the high-speed train might decrease the air transport but it would improve the environment and congestion at the airport. Most of the literatures only contain descriptive reasoning or case studies on a certain air route and it is hard to draw a conclusion on the effect of Fyra just based on the literature view. Thus it was necessary to continue the research with case study and data analysis. Frankfurt airport was selected to compare with Schiphol airport since they have the similar size and function. Furthermore, Schiphol airport was treated as the control group to see the future effects of Fyra on air transport. Following conclusions was drawn after the case study and analysis:6.1 The effect ICE on air transportConclude from the case study, ICE was introduced ten years earlier than Fyra. It runs through Frankfurt airport to almost every part of Germany, and also neighbor countries. There were quite a few evidences suggest that ICE improved the accessibility of the airport and intermodal system between rail/air of the airport, but it also took away passengers from short-haul flights. The result of the DID regression were consistent with the evidences in the case study. ICE had negative effects on both air transport movements and monthly air passengers, especially on the monthly air passengers. It might be caused by ICE taking away short-haul air passengers. Although ICE had a negative effect on air transport, it was still a necessary and beneficial investment, for the following reasons:Firstly, Frankfurt airport had a worse train connection twelve years ago compare to what Schiphol had two years ago. Thus, ICE was much more essential to the airport at that time to improve the accessibility and air/rail integration. Furthermore, as ICE has many long-distance destinations it would offer a faster and more convenient train service to the travellers. Secondly, as ICE substituted for a considerable amount of short-haul flights, it could be more environmental friendly. ICE produces fewer emissions and has more capacities than the airplane. Furthermore, it could free the airport runway, thus enable the airport to plan more long-distance flights and it could improve the congestion at the airport. Thirdly, ICE offered great champion opportunities for both the airport and the airlines. The introduction of AIRail and Air&Rail service increase the competitiveness of both the airport and airlines to some extent. Overall, ICE did have a strong effect on the air transport. The introduction of ICE was overall socially beneficial and crucial even though it reduces air transport. 6.2 Estimation of the future effect of FyraLearning from the past experience with ICE, Fyra is very likely to have more effect on air transport in the future when it extends its service to Beligum and when travellers and airlines start to adapt to the service. However, the effect will most likely to be small for several reasons: Firstly, before Fyra was introduced, there was already a mature train system in Schiphol airport, and another high-speed train Thalys is already running through Schiphol airport to Brussels. Thus, even if Fyra extends its service to Brussels it might not make a very big difference. Secondly, as the size of Germany is much bigger than the Netherlands, the destinations of ICE are mostly hours away from Frankfurt Airport. Compare to ICE, the destinations of Fyra was relatively close to the Airport. The farthest destination of Fyra at the moment is Breda with a journey time less than one hour. It would not make much a difference for the travellers. Therefore, to conclude: Fyra is expected to have increasing stronger effects on air transport in the future, it might reduce air transport but there will be social benefits. However, the effects are not expected to be very significant. Compare to ICE, Fyra is expected to be less socially beneficial and less crucial to the air transport. 6.3 Limitations and future researchesThis paper gives a more solid evaluation on the effect of ICE and Fyra on air transport based on evidences and data analysis than the previous researches. However, there are still a few limitations.Firstly, while doing the DID estimation, Schiphol, as the control group already had two years of experience with HST. In theory, the control group for DID estimation should not be exposed to the treatment at all. Thus, this might affect the accuracy of the DID estimation. For future research, one could include one more control group with an airport haven’t had any high-speed train connections. Secondly, the indicator “air transport movements” include both the cargo and passenger transportations. As the high-speed train does not affect cargo movements, the result might have less relation with the effect of HST. Thus, for future research, one could exclude the cargo out of the air transport movement.Reference:7.1 Literature:Albouy, D. n.d. Program Evaluation and the Difference in Difference Estimator. Section notes. Economics 131. University of California, Berkeley< Rus G. and Nash C. 2007. In what circumstances is investment in HSR worthwhile?. ITS Working Paper 590, Institute for Transport Studies University of Leeds.de Rus, G. and Inglada, V. 1997. Cost-benefit analysis of the high-speed train in Spain. The Annals of Regional Science, 31, pp. 175-188.Givoni, M. 2006. Development and impact of the Modern High-speed Train: A Review. Transport Reviews 26 (5): 593-611.Givoni M. 2007. Environmental benefits from mode substitution: Comparison of the environmental impact from aircraft and high-speed train operations. International Journal of Sustainable Transportation 1 (4): 209-230.Givoni, M., Banister, D. 2006. Airline and railway integration. Transport Policy, 13 (5): 386-397Gonzalez-Savignat M. 2004. Competition in air transport: The case of the high speed train. Journal of Transport Economics and Policy 38(1): 77-108.Grimme, W. 2007. Experiences with Advanced Air-Rail Passenger Intermodality – The case of Germany. DLR Working PaperGrimme, W. 2006. Air/Rail Intermodality – Recent Experiences from Germany. Aerlines Magazine, issue 34 , Jorritsma, P. (2009), “Substitution opportunities of High Speed Train for air transport”, Aerlines Magazine, issue 43. López-Pita, A. and Robustè, F. (2005) Impact of High-Speed lines in relation to very high frequency air services, Journal of Public Transportation, 8(2), 17-35.López-Pita, A., and F. Robusté. 2003. The effects of high-speed rail on the reduction of air congestion. Journal of Public Transportation 6 (1): 37–52.Nash, C. A. 1991. The case for high speed rail. Working Paper 323. Institute for Transport Studies, The University of Leeds. Park, Y. & Ha, H. 2006. Analysis of the impact of high-speed railroad service on air transport demand, Transportation Research Part E: Logistics and Transportation Review, vol. 42, no. 2, pp. 95-104.Terpstra, I., Lijesen, M. 2011. High-Speed Train as a Feeder for Air Transport. Aerlines Magazine. issue 49Wooldridge, J.2007. Lecture 10: Difference-in-Differences Estimation. NBER Summer Institute.< Websites:Airport Council International < >DB -Deutsche Bahn < db.de >Frankfurt Airport < frankfurt->Fraport AG < >Fyra < ;“Heteroscedasticity” - notes from University of Notre Dame < >MFO.de < Nederlandse Omroep Stichting < - Nederlandse Spoorwege: <ns.nl>NS Hispeed <nshispeed.nl>Schiphol Airport <schiphol.nl>Treinreiziger .nl < > ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download