Skip to main content

The dynamic between airlines and high-speed trains in Europe

Final Report Summary - DATE (The dynamic between airlines and high-speed trains in Europe)

The DATE (Dynamic between Airplanes and high-speed Train modes in Europe) project aimed to evaluate the potential for high-speed rail (HSR) to cut air transport flows and the conditions required to support it in order to achieve ‘greener’ medium-distance mobilities in Europe.
We started with a critical analysis of the literature on the relationships between HSR and air services with a special focus on mode substitution (namely, passengers changing from one transport mode to another) and on induced demand effects (that is, new traffic generated by the HSR). Indeed, to date, relatively little is known about the nature of the demand for HSR after inauguration of the services, despite the nearly 50-year experience of HSR. This is a real lacuna given the scale of HSR construction around the world, the amount of resources committed to it, the desired accessibility, economic, and environmental effects associated with HSR development and the relatively poor track-record of forecasts demand for HSR services. Such a review faced main methodological challenges and existing evidence that is largely influenced by route-specific characteristics. However, the information we gathered from numerous sources allowed us to draw several conclusions.
First, acknowledging the limitation of the measurement of induced demand, it can be concluded with caution that a few years after the introduction of HSR, on average, about 20% of the demand is new, induced demand (with large standard deviation). Thus, it is reasonable to assume, for planning purposes (especially when modelling/forecasting is not possible) that induced demand ranges from 10 to 20% between two and four years after the introduction of HSR.
Given the level of induced demand, it can be assumed that on average about 80% of the passengers using HSR a few years after its inauguration will be passengers that previously used other modes to travel on the route, including conventional rail. The level of mode substitution varies between routes along with the composition of the passengers shifting to HSR. When considering the demand for HSR, mode substitution is mainly from conventional rail to HSR. It is clear that the level of substitution from conventional rail to HSR is likely to be high, yet the implications that this will have on conventional rail services and use at the network level as well as on overall rail use overall (combined high-speed and conventional rail) is unclear. In a policy context of promoting rail use on the expense of car and air modes, this is an important matter for further research and a subject that does not get enough attention in the context of HSR research. Investments in HSR can have an adverse impact on the conventional rail network due to both reduction in demand and reduction in investments in the conventional network and deteriorating service as a result.
On some routes, however, the effect of mode substitution can be larger for air to HSR substitution, especially in the case of uncompetitive previous rail services (e.g. Madrid-Seville) and sea crossing (Channel tunnel). It also appears than the decrease of air passengers might be larger than the decrease in air services. In other words, the environmental benefit of mode substitution may be smaller than the passenger dynamics suggest. The introduction of HSR services can also result in closing air services, especially in the case of shorter journeys. In some cases, the airlines might as a result decide to cooperate with the HSR operator and offer their service by HSR instead of aircraft. Road transport is usually the third most important source of HSR passengers after conventional rail and air transport, but it is more difficult to conclude on the effect of mode substitution on demand for car travel. This effect likely depends largely on the ease to access the HSR station. In general, and all other things being equal, as route distance increases, HSR gains market share from road transport but starts to lose market share to air transport at some point it.

We then started the econometric analyses to assess the impacts of European HSR on air services. Most published works have focused on the expected intermodal competition, considered demand (i.e. passengers) alone, and covered few city pairs, if not only one. In contrast, our research was fundamentally an ex-post analysis of the current air services under the influence of HSR. Furthermore, it considered all routes where direct HSR services compete with air services, including those where airlines exited the market following intermodal competition. This led us to consider 161 city pairs EU-wide. We used both transport-related data and ESPON geo-economic data that make it possible to compare cities on an EU basis. Finally, we used specific methods that guarantee robust results. This also involved revisiting the geography of European low-cost airline (LCA) networks, because the LCAs were suspected to affect the competition between HSR and air transport. We found that most LCA services mainly serve large cities and are short-haul (55% of LCA seats are less than 1,000 km and 37% less than 750 km), suggesting than HSR and LCAs might be partly focused on similar markets.
The first series of models aims to analyse the impact of HSR on the current (2012) provision of air services. It was found that air services are indeed affected by HSR travel time: they are more air services if HSR travel time is longer. Such a result was expected based on previously published ex-post evidence. However, to our knowledge, it has never been found through an EU-wide quantitative analysis covering one hundred routes. We also found that HSR travel time has a similar impact on both airline seats and the number of flights with a similar magnitude. This involves that in our sample, intermodal competition does not lead airlines following frequency-oriented strategies to maintain their competitive position. This confirms that within a free competition between modes, HSR helps to restrict the provision of air services when travel time is not too excessive. In contrast, HSR frequency is found to have only limited impact on air services.
At the route level, hubbing strategies led by the airlines have the opposite effect from HSR, as hubs involve more air services concentrated on a central airport. This raises the issue of whether airports should be served more by HSR, with the latter replacing short-haul flights to feed longer ones. On the one hand, this suggests that there is potential for additional mode substitution through either competition or integration if HSR served the airports. On the other hand, the success of such practice is subject to various market and technical factors, including the range of airline connections offered, the geography of HSR routes versus airports locations, schedule optimisation, degree of commercial and technical integration, etc. Furthermore, at congested airports, the freed slots may be reused for other flights, potentially long-haul ones with larger environmental impact. Even without such rebound effect, greenhouse gas emissions prevented thanks to HSR are small compared with emissions from long-haul flights. Also, servicing airports instead of city centres may mean higher car use for access and egress journeys. Finally, within our sample, we did not detect much specific competition from the low-cost airlines, all other things being equal.

The second series of models aims to analyse the impact of HSR on the change in air services, namely, a focus on before/after HSR provisions for air services. This stage is still ongoing, because it takes much more time than expected to gather the data related to more than two decades. However, arrangements have been made to make it possible to finalise this part of the work by autumn, before further dissemination. This will answer the question of whether European HSR services have been helpful environmentally-speaking and what the related-factors of (un)success are.

As a provisional conclusion, high-speed rail does have an impact on the provision of air services notably due to attractive travel times. Ongoing works will reveal more based on a long-term ex-post analysis. Nevertheless, induced traffic and mode substitution from traditional rail should not be underestimated. Furthermore, the potential for additional HSR services throughout European is significantly restricted by traffic density, because only routes with enough traffic can justify the high investment costs. We had a preliminary look at this issue just before starting the DATE project, and we will try to go further in a subsequent research project.