Estimering av efterfrågan på långväga persontransporter : En ekonomisk-teoretisk belysning av gängse modeller samt en ny ansats till uppskattning av efterfrågesamband

Sammanfattning: This dissertation deals with the estimation of intercity travel demand. The purpose is to discuss the derivations and characteristics of existing models, to derive a demand model based on microeconomic theory and to estimate the parameters of the model in order to gain information on important characteristics of the transport system.The assumptions in the derivation of common travel demand models from microeconomic theory (gravity models, entropy maximising models and logit models) are often made only to support existing models and the economic implications of the assumptions are not fully discussed. In many cases this results in models with characteristics that are questionable in the light of economic theory. In other cases theoretically possible but restrictive characteristics are given in advance rather than determined in the estimation process.The intercity travel demand model proposed in this study is based on the general assumption that travel demand can be derived from the demand for visits to other cities. As people living in different parts of a country experience different prices or "generalised transport costs" for visits to the cities, it is possible to estimate demand functions for visits to each city from the use of cross-sectional data, which is normally not possible for other utilities. From these demand functions, demand functions for separate transport relations can be derived. Necessary conditions concerning various demand elasticities are formulated and their consequences for the mathematical specification of the model are discussed. Of particular interest are those parts of the model that link together different levels. In order not to specify the cost function in advance, a Box-Cox formulation is used. The result is a demand model which determines the total number of trips made from each city, as well as the distribution among destinations and travel modes.A simplified version of the model, not including distribution among travel modes, is estimated with Swedish data for the period 1990-1994. Estimations are made for the total number of trips as well as separately for business trips and private trips. The following results are the most interesting. There is strong support for the cost expression in the model to approach a power function, contrary to the exponential function that is given in advance in many other models. In the long run the demand for trips between two cities is elastic with respect to the generalised transport cost. Elasticities estimated range from minus one to minus two. About 80% of the additional trips caused by a decrease in costs are redistributed from other destinations, however, showing that in the long run the possibilities of substitution among destinations are good. Consequently, the relative effects of a general change in transport costs between all city pairs are smaller. Elasticities found in the study lie between -0,3 (private trips) and -0,5 (business trips).