یک متاآنالیز از کشش قیمتی تقاضا برای حمل و نقل در چارچوب تعادل عمومی

Price elasticities of transport demand are an important tool to assess the impacts of pricing policies. Empirical research on these elasticities leads to a rather wide range of outcomes. There is obviously a need for a more rigorous methodological framework. This paper provides a new integrative approach to the estimation of price elasticities whilst taking into account any combination of characteristics of transport demand found in various empirical studies. To this end, we apply meta-analysis to this set of studies. From the various modelling approaches that underlie these studies we develop an overlapping general equilibrium framework that provides a meta-regression equation relating the price elasticity estimate to the study characteristics.

The last decades have witnessed an explosive growth in the demand for transport all over the world following an increase in economic activities in the EU, the USA, and Asian countries like Japan and Korea. Statistics show a structural rise in mobility of travellers, while the liberalisation of trade worldwide (WTO) and in the EU has caused a drastic increase in transboundary freight transport [see for figures, Reggiani et al. (2001)]. There is apparently no natural limit to the rise in physical movements of people and commodities. Both local and global transport is on a rising edge. This explosive growth in transport has clearly an important environmental impact in the form of heavily increased pollution, more accidents, noise, and congestion, causing national as well as international governmental bodies to worry about the sustainability of their transport systems [see for an overview, Nijkamp and Pepping (1998)]. In many cases transport is not an aim in itself, but it derives its value from production of other consumption aims. Therefore, it is plausible to assume that an increase in economic activities causes an increase in transport demand. The cost of transport modes tends not to reflect the true costs they inflict on society, because environmental costs are not accounted for. These costs are not equal for all transport modes and, hence, the distribution of demand for transport over various modes is not efficient from a societal viewpoint. Governments are, therefore, trying to include the environmental costs of each transport mode into its price hoping to obtain a more efficient distribution of transport demand over the various transport modes. This redistribution of transport demand may then lead to a substantial decrease in the environmental impact of the country's transport system. ‘Fair pricing’ has become even a policy objective with a view to better incorporation of externalities in transport policy. Nijkamp and Pepping (1998) mention the price elasticity of transport demand as the most important parameter to understand how pricing policies will affect transport demand. This price elasticity is defined as the relative change in demand for a given mode induced by a relative change in price. It is no surprise that in the last few years several studies in European countries have assessed price elasticities of demand in the transport sector, which has led to a great diversity of empirical results. Nijkamp and Pepping (1998) state that most of these investigations have been made on a non-controlled basis, resulting in a rather feeble comparability of the results from these studies. The well-known survey of Oum et al. (1992) argues that even across-the-board generalisations about transport demand are impossible. Competition between modes, routes or firms, and site-specific conditions give rise to a wide range of price elasticities. Factors such as time horizon, degree of aggregation, functional specification and the like, used in these studies turn out to have a significant bearing on elasticity estimates. Despite the variety in background of these elasticity estimates, Nijkamp and Pepping (1998) consider it sensible to analyse the differences in statistical results in order to identify commonalties and site-specific differences more precisely, as it would allow for more transferability of results under varying quasi-experimental conditions. In their view, meta-analysis may play an important role in this framework. The authors carry out a comparative analysis of various elasticity values of demand for transport that are being used in various member states of the European Union. Their comparative analysis is based on a recently developed approach called ‘rough set analysis’. Meta-analysis has originally been developed as a tool for comparing and synthesising results from various studies in the natural sciences and, has in the past decade, also become popular in experimental psychology and medicine. We refer to Hedges and Olkin, 1985 and Wolf, 1986, and more recently, Cooper and Hedges (1994), which have become authoritative resources of meta-analysis in these subject areas. Behavioural studies in these sciences tend to be more similar and often consist of comparing experimental results. This semi-controlled experimentation leads to a similarity in studies, which is then lost in the economic sciences due to the existence of the ceteris–paribus condition as argued in Bal and Nijkamp (1999). This applies, in particular, to the assessment of price-elasticities in transport due to the factors already mentioned in Oum et al. (1992). A meta-analysis of studies that assess the value of price-elasticities of transport demand should take into account the various underlying data and behavioural models pertaining to transport. The question is now whether it is possible to develop a common frame of reference (or a benchmark) to compare different elasticity studies. Kremers et al. (1999) suggest that a general equilibrium model including transport and its impact on the environment, can serve as an encapsulating model for meta-analysis in the field of transport policy analysis in relation to environmental economics. Following their approach, we aim to present here a general equilibrium model that is able to encapsulate, in principle, the various price elasticity studies. This controlled basis is necessary for comparing the studies by providing a framework in the form of an ‘envelope’ model. Such a model is then also able to offer the foundation for an empirical estimation of price elasticities in the transport sector. The next section describes the general equilibrium model framework that is thought to act as an envelope for the models in the various studies underlying a meta-analysis of price elasticities of transport demand. These models also set the limitations of the framework. Any bias in the literature on this field, thus, results in a bias in the general equilibrium framework. It is the introduction of an enveloping general equilibrium framework as the ‘controlled’ basis for the comparison of possibly very different economic studies, that is new in our approach to applying meta-analysis in economic sciences. In the following section, this framework is used to derive the common characteristics of these models, which influence the estimation of a price elasticity of transport demand. We concentrate our meta-analysis on own-price elasticities. The main result of this section is a set of regression equations that links the various price elasticities of transport demand to these common characteristics. This set of regression equations then form the analytical basis of the actual empirical meta-analysis applied in Section 4. This meta-analysis considers the issues of combining and comparing the study results as described, for example by Hedges and Olkin (1985) and Wolf (1986).

Meta-analysis is currently a popular tool and field of research in experimental sciences such as medicine or psychology, but it hasn't fully reached the economic sciences yet. This is not surprising considering the fact that in order to successfully apply meta-analytical methods, the underlying studies are assumed to be comparable. Due to the underlying ceteris–paribus condition in many economic studies, this assumption turns out to be violated in economics. In this paper, we have considered a meta-analysis of price elasticities of transport demand in order to start up research for the possible application of meta-analytical techniques in the economic sciences. The estimation of such price elasticities is characterised by the use of very different underlying models of transport demand. This paper distinguished among microeconometric, micro-economic, and discrete choice models as frequently applied models in this area. We have, therefore, introduced the idea of constructing a general equilibrium framework that encapsulates all the models used in the studies that underlie our meta-analysis. As such the general equilibrium framework makes all these studies comparable. From this common reference framework, we have derived the model characteristics, which have different values for each study. Together with the data characteristics underlying each study and the model characteristics, these model characteristics created a meta-regression equation to estimate the price elasticities of transport demand (see Appendix A). It turns out that the use of a micro-economic model instead of the reference micro-econometric model has a significant influence on the estimation of price elasticity. Micro-economic models as used, for example, by Friedlaender and Spady (1980), explicitly regard transport as a demand derived from the economic activities of the consumer or producer. These models are much closer to the general equilibrium framework introduced in Section 2 than the other two dominant model types. Micro-economic models not only cover most of the economic factors that are essential to the determination of transport demand, but it also incorporates the economic interactions. Discrete choice models only consider mode choice and, therefore, disregard the impact on the volume of transport. Micro-econometric models only provide a reduced form equation relating transport demand to a number of characteristics, but leave out their economic interactions. This results in significantly more elastic price elasticity when applying micro-economic models. Air transport forms a significant modal characteristic. Air transport demand turns out to be more sensitive to price changes than the other modes of transport, i.e. rail and road, considered in our meta-analysis. This may be due to the different types of economic activities associated with each type of transport. In view of the previous discussion with respect to the use of models, this dependency on the consumer or producer's economic activities suggests a preference for micro-economic models as the most appropriate type of model to underlie the estimation of price elasticity. The meta regression results show that it differs significantly for the estimation of a price elasticity of transport demand whether the data are obtained from transport demand on an urban or national scale. Using transport demand on an urban scale results in a lower estimate of price elasticity. This seems to be a result of availability alternatives. On an urban scale, the consumer or producer has more alternatives to choose from than on a national scale, which makes them change more rapidly in response to price changes.