DYNAMO: Dynamic Automobile Market Model for the Netherlands



DYNAMO: Dynamic Automobile Market Model for the Netherlands

Authors

Henk Meurs, Rinus Haaijer, MuConsult, NL; Remko Smit, AVV, Ministry of Transport, NL; Karst Geurs, MNP, NL

Description

The paper describes a new dynamic car market model for The Netherlands including demand for cars, car- and fuel types and car usage as well as supply side developments

Abstract

The paper presents the new model for the car market in the Netherlands, replacing the older Facts-model. The model is able to simulate the dynamics of the car market with respect to the number of cars, type and fuel characteristics of the cars, car usage and car prices. The output is a function of demographic, socio-economic developments, technological developments and sensitive to various policies such as those affecting fixed and variable cost of cars, scrappage policies and so on. Most recent, an environmental module is added allowing assessment of emission effects of future developments in car fleets in The Netherlands. The result is one of the most comprehensive operational car market models nowadays available capable to meet demanding complex policy issues

The model system has a demand and supply side as well as an equilibrium module. The demand side consists of a set of (related) discrete choice models describing choices with respect to the number of cars in the household and vehicle and fuel type choices, as well as models for vehicle mileage. A separate module for business cars is implemented. The supply side reflects developments in technology of cars, including fuel efficiency and changing offers in types. The equilibrium module uses the price variable on the second hand market to bring demand and supply together. A scrappage module estimates the fraction of cars of specific categories that will be scrapped at certain prices. The model also allows for import and export of cars from and to The Netherlands. The model is dynamic with prices of previous years affecting current decisions with respect to car ownership.

The core of the model is a so-called base-year 2003 matrix of car types by household characteristics. In total 71 household types were distinguished as well as 120 car types. Data for this matrix were received from the car registration office in The Netherlands as well as from the Dutch Mobility Survey (OVG/MON). The model modules described above are used to forecast changes in this matrix on a year-by year basis.

The models within the modules were estimated using different data sources. We used RP-data from OVG/MON as well as SP-data to gain coefficients associated with price variables in the model. In order to find coefficients associated with dynamic components in the model we used OVG/MON as pooled cross-section time series data. All models within the modules were extensively tested and validated.

The integral model is validated using data for the periode 1990-2003 and applied for the period 2004-2040. Results demonstrate that the model was able to reproduces historic developments in the car market in The Netherlands.

Specific policies have been analyzed and will be presented in the paper, including variabilisation of taxes on cars from fixed taxed to usage-dependent taxes and scrappage policies. Outcomes demonstrate the model capabilities.

Publisher

Association for European Transport