In this paper we present the results of a study that aims to establish the potential for high-speed train travel as a substitute for short distance air travel at Amsterdam Airport. We investigated the 13 most important destinations that offer direct flights to and from Amsterdam Airport. Almost 40% of the air passengers travelling to/from these destinations are transfer passengers. Empirical evidence reveals that high-speed trains dominate the market for journeys of 2 hours or less, such as between Paris and Brussels. However, trains claim only a tiny market share of journeys longer than 5 to 6 hours; air travel dominates that market segment. Using these findings, we developed a model to estimate the substitution of air travel with high-speed train travel. The explanatory variables in this model are travel time, daily departure options, fares, and the inconvenience associated with transferring at airports. In a “minimum” scenario, we predict that in 2030 high-speed trains could replace approximately 1.9 million air journeys. This calculation is based on feasible reductions of train travel times and increased train frequencies for part of the rail network. In this scenario, Amsterdam–London accounts for more than three-quarters of the predicted substitution. In a “maximum” scenario, substitution could increase up to 3.7 million air journeys per year, provided that inconveniences for passengers when transferring at airports from plane to train are resolved and train ticket fares are reduced by 20%. These two scenarios imply a reduction of 2.5 to 5% of all flights to/from Amsterdam Airport in 2030.

This paper develops and applies a practical method to estimate the benefits of improved reliability of road networks. We present a general methodology to estimate the scheduling costs due to travel time variability for car travel. In contrast to existing practical methods, we explicitly consider the effect of travel time variability on departure time choices. We focus on situations when only mean delays are known, which is typically the case when standard transport models are used. We first show how travel time variability can be predicted from mean delays. We then estimate the scheduling costs of travellers, taking into account their optimal departure time choice given the estimated travel time variability. We illustrate the methodology for air passengers traveling by car to Amsterdam Schiphol Airport. We find that on average planned improvements in network reliability only lead to a small reduction in access costs per trip in absolute terms, mainly because most air passengers drive to the airport outside peak hours, when travel time variability tends to be low. However, in relative terms the reduction in access costs due to the improvements in network reliability is substantial. In our case we find that for every 1 Euro reduction in travel time costs, there is an additional cost reduction of 0.7 Euro due to lower travel time variability, and hence lower scheduling costs. Ignoring the benefits from improved reliability may therefore lead to a severe underestimation of the total benefits of infrastructure improvements.

The paper addresses the consumer value of changes in service frequency for timetable-based transport systems such as bus, train, ferry and air transport. Instead of using "average waiting times" we propose a more appropriate model specification of the different impacts of timetable changes for individual travellers. This includes a decision whether to plan the journey or not, and (if planning) when to start the journey as a function of the services available. We also address the issue of how to aggregate the individual values across the travelling population to obtain estimates of total welfare. Because there is significant variation in preferences between individual travellers, logsum-type measures are used to derive expected utilities rather than mean values. We end our paper by illustrating the application of our method for a rail service.

Depuis le milieu des années 90, la fréquentation des transports publics en Île-de-France (région parisienne) a considérablement augmenté – de 20 % au cours des dix dernières années seulement. Cette progression, qui figurait parmi les objectifs du plan de mobilité urbaine durable adopté en 2000, n’avait toutefois pas été totalement anticipée. Le renouvellement des infrastructures ferroviaires et du matériel roulant est nécessaire pour faire face à cette situation mais à lui seul, il ne suffira pas. Des investissements majeurs sont prévus afin d’accroître la capacité, en construisant de nouvelles lignes ou en augmentant la capacité des lignes existantes. Le Grand Paris Express est le plus connu de ces projets. Pour l’évaluation socio-économique, il est nécessaire de quantifier l’ensemble des impacts de ces investissements. Cependant, on connaît mal la valeur que les voyageurs accordent à la réduction des niveaux de congestion. Le Syndicat des Transports en Île-de-France a donc confié en 2011 à Significance la mission de réaliser une nouvelle étude sur la perception du confort dans les véhicules de transports publics en général, et plus particulièrement sur la question de l’affluence. L’étude devait couvrir l’ensemble des modes de transports publics franciliens.

This paper describes the results of a research project that aimed to establish passenger values of crowding on public transport services in the Paris region. Qualitative research, stated preference (SP) experiments, and passenger counts and surveys were conducted to obtain such values. A simple method was developed to quantify the passenger benefits of specific public transport projects aiming to reduce crowding on existing lines. This method was applied in a case study to the regional rail (RER) RER Line E extension project. With regard to the value of crowding, the research indicated that the perceived disutility of crowding could be more accurately described as a constant disutility per trip than as a travel time multiplier. However, for ease of application often the multiplier formulation was preferred. When the value of crowding was expressed as a travel time multiplier, values were obtained ranging from 1.0 when all passengers could be seated to 1.7 for standing bus passengers when the vehicles reached their maximum capacity. Also for seated passengers, multipliers well above 1.0 were observed for (highly) congested vehicles (maximum value = 1.5 for bus passengers). These values were applied in a case study that estimated the effects of an extension of the regional rail line RER E in the western direction, partially running parallel to the existing RER Line A. This extension would reduce the current (very) high crowding levels on the RER A and B lines to more moderate levels and generate benefits of about €23 million per year.

Since the mid 90’s, public transport patronage in Ile-de-France (the Paris region) has increased substantially: over the last decade alone a 20% growth was observed. This growth, even though it was an aim of the Sustainable Urban Mobility plan adopted in 2000, was not completely anticipated. Consequently, the capacity is no longer sufficient to meet the demand during the peak hours, particularly on several parts of the network in the dense central area of the region. This results in over-crowded vehicles and long waiting times for passengers at rail platforms and bus stops. The lack of maintenance and modernisation of the transport system causes additional operational difficulties.

The City of Paris, together with surrounding “communes”, created a public authority to investigate the possibility to launch by 2011 of a new transport system: Autolib’. The project is related to the highly successful Velib’ project that was launched in Paris a few years ago. Autolib’ is essentially a system of 4000 “shared” electric cars that can be used for one-way trips of limited distance between 1400 parking points within central Paris and the surrounding regions. Details of the fare system are still being studied, but it is envisioned that the user would pay a subscription fee and a variable cost depending on the duration of use. And importantly: there would be a guaranteed parking space at the destination of the trip. Avis, RATP, SNCF and VINCI Park formed a consortium to bid for the operation of the Autolib’ system. They have commissioned research to estimate the potential demand and revenue for the new Autolib’ service with the highest possible accuracy. This to help them to shape the service in the best possible way, to determine the financial conditions and the economic basis of the project, and also to generate the information and argumentation necessary to maximise the reliability of the system. In the paper we shall briefly introduce the proposed new system, and report the stated choice research that was carried out to estimate the potential demand. The following three experiments were conducted: One stated intentions exercise to measure preferences for different subscription possibilities; One stated choice experiment to assess the propensity to buy a subscription; One stated choice experiment investigating mode choice among three alternatives: chosen mode, best alternative mode, newly proposed mode. In the paper we will describe the chosen methodology, the way in which the results of the three experiments have been integrated, and the lessons that can be learnt for estimating potential demand for new transport modes using stated choice experiments.

We analyze the cost of access travel time variability for air travelers. Reliable access to airports is important since the cost of missing a flight is likely to be high. First, the determinants of the preferred arrival times at airports are analyzed. Second, the willingness to pay (WTP) for reductions in access travel time, early and late arrival time at the airport, and the probability to miss a flight are estimated, using a stated choice experiment. The results indicate that the WTPs are relatively high. Third, a model is developed to calculate the cost of variable travel times for representative air travelers going by car, taking into account travel time cost, scheduling cost and the cost of missing a flight using empirical travel time data. In this model, the value of reliability for air travelers is derived taking “anticipating departure time choice” into account, meaning that travelers determine their departure time from home optimally. Results of the numerical exercise show that the cost of access travel time variability for business travelers are between 0% and 30% of total access travel cost, and for non-business travelers between 0% and 25%. These numbers depend strongly on the time of the day.

This paper analyses the effect of reliability of car access travel time on air passenger decisions, including departure time choice, access mode choice and ultimately departure airport choice. The method proceeds in three steps. Firstly, the generalized costs of passenger access to airports are calculated including the costs of travel time, but also the costs of scheduling and of missing a flight. Other monetary expenses and the costs of parking are also included. In these calculations the authors explicitly model the buffer time that passengers build in to counter the risk of being late due to variations in travel time as a function of traffic congestion. Secondly, the authors simulate how different levels of traffic congestion lead to different travel times and risks of delay for all links inside the road network. These simulations are based upon an analysis of observed variations in car travel times (within and between days) for roads leading to the airport. This analysis enables the authors to develop reasonably accurate predictions of both mean travel times and variations in travel times for road sections under different levels of congestion. Thirdly, the impact of possible road infrastructure improvements on air passenger decisions is analyzed. The authors simulate how increases in road capacity for existing bottlenecks affect the means and the variances of access travel times, and hence the access costs. These are then used as input for an existing airport simulation model which predicts how the market shares of competing access modes are affected, and ultimately even the choice of departure airport. An illustrative case study will be elaborated for air passenger access to Schiphol airport near Amsterdam. The emphasis in this paper is on presenting and illustrating a practical method to include the impact of reliability on air passenger choices, illustrating the importance of short and reliable access travel times to airports, particularly during the morning peak hours.

This discussion paper resulted in a publication in Transportation Research Part B: Methodological (2011). Vol. 45(10), pages 1545-1559. This paper analyses the cost of access travel time variability for air travelers. Reliable access to airports is important since it is likely that the cost of missing a flight is high. First, the determinants of the preferred arrival times at airports are analyzed, including trip purpose, type of airport, flight characteristics, travel experience, type of check-in, need to check-in luggage. Second, the willingness to pay (WTP) for reduction in access travel time, early and late arrival time at the airport, and the probability to miss a flight is estimated using a stated choice experiment. The results indicate that the WTPs are relatively high, which is partially due to the low cost sensitivity of air travelers. Third, a model is developed to calculate the cost of variable travel times for air travelers going by car, taking into account travel time cost, scheduling cost and the cost of missing a flight. In this model, the value of reliability for air travelers is derived taking 'anticipating departure time choice' into account. Results of the numerical exercise show that the cost of access travel time variability for business travelers are between 3-36% of total access travel cost, and for non-business travelers between 3-30%. These numbers depend strongly on the time of the day.

It is a common finding in empirical discrete choice studies that the estimated mean relative values of the coefficients (i.e. WTP's) from multinomial logit (MNL) estimations differ from those calculated using mixed logit estimations, where the mixed logit has the better statistical fit. It is, however, less clear under exactly what circumstances such differences arise, whether they are important, and if they can be seen as biases in the WTP estimates from MNL. We use datasets created by Monte Carlo simulation to test, in a controlled environment, the effects of the different possible sources of bias on the accuracy of WTP's estimated by MNL. Consistent with earlier research we find that random unobserved heterogeneity in the marginal utilities does not in itself biases the MNL estimates. Furthermore, whether or not the unobserved heterogeneity is symmetrically shaped also does not affect the accuracy of the WTP estimates of MNL. However, we find that if two heterogeneous marginal utilities are correlated then the WTP's from MNL may be biased. If the correlation between the marginal utilities is negative, then the bias in the MNL estimate is negative, whereas if the correlation is positive the bias is positive.

This paper studies the choice of type of train ticket using a Stated Preference experiment conducted among current Dutch single ticket travellers. Multinomial logit (MNL), nested logit and mixed logit models are used to analyse the choices of the respondents. The experiment has three alternatives, namely: (1) an unrestricted ticket, (2) a cheaper restricted ticket which has restrictions on travel during the peak and (3) neither the first nor the second alternative. The price elasticities for the unrestricted ticket are rather low (in absolute sense), whereas responses to changes in the price of the restricted ticket are stronger. We find that MNL, compared with mixed logit, underestimates the value of (in-vehicle) travel time and overestimates the WTP’s for the travel moment restrictions. We find that travel cost compensation by the employer substantially decreases the price sensitivities of the respondents. This is an expected, but important finding, as a large share of Dutch commuters receives travel costs compensation.

In The Netherlands, major infrastructure projects are assessed using cost-benefit analysis, following official guidelines. Until recently, the reliability of travel times could not be included in the cost-benefit analysis, because the corresponding monetary valuation was unknown. In recent years, the literature on valuing reliability of travel times was reviewed for the Dutch transport ministry. The outcomes of this were discussed at an expert workshop, which led to an agreement on preliminary monetary values for passenger transport. A key concept is that of the reliability ratio. This is defined as the value of reliability (measured as the standard deviation of travel time) divided by the value of travel time itself. For freight transport a follow-up study was carried out, which transforms the results of earlier stated preference research into a reliability ratio. The paper presents and explains the preliminary values, focussing on the derivation of reliability values for freight transport. It also describes how these values can be used in practical project evaluations.

This paper will discuss how traffic forecasts are widely used for cost-benefit analysis of proposed infrastructure investment projects. Forecasts are by definition uncertain and it is desirable to take this uncertainty into account in the analysis. The uncertainty is not only due to the fact that a model is a simplification of a complex reality, but also to uncertainty in the inputs of the model, such as the expected economic growth. Previous studies have shown that traffic models are not all bad, but the main cause of deviations between traffic forecasts and observed traffic volumes are the errors in these model inputs that really drive the growth of travel. Macro-economic data are not the only inputs for travel models. Another important group of factors that influence traffic forecasts are the policies of other authorities and organizations (both policies that have already been decided upon, but have not yet been implemented, and new future policy decisions) but are external to the decision-maker for a specific project,Uncertainty in these policies can also be an important factor. However, these uncertainties are usually not taken into account in traffic forecasting studies. The Frejus tunnel is an important toll road connection between France and Italy and is part of the trans-European transport network (TEN-T). The operators of the tunnel use traffic forecasts to estimate their future revenues, which in turn are being used to negotiate subsidies from the road authorities in Italy and France. In the forecasts several types of uncertainties are explicitly taken into account: uncertainties in the key drivers of traffic growth, such as gross domestic product (GDP), but also uncertainties caused by policies of other parties affecting traffic volumes. The author have developed a tool to take all these uncertainties into account and present the outcomes in a clear and intuitive way that is useful for stakeholders. Similar tools have also been developed for other projects in The Netherlands and France. In this paper the authors will demonstrate an example of this tool and will discuss the importance of taking uncertainties into account for policy making.