Measuring regional inequality by internet car price advertisements : evidence for Germany

We suggest an alternative indicator based on the car sales price placed on the Internet for measuring economic inequality among regions. The regional data on car prices in Germany were downloaded from two specialised websites http://www.mobile.de and http://www.autoscout24.de in December 2011. The corresponding number of unique car price observations downloaded from each website is 914,105 and 802,047. The following information was recorded: make, model, ZIP code, mileage, engine volume in liters and cubic centimeters, type of transmission (manual, automatic, etc.), year of the first registration, and offer price. The ZIP code information was used to find the geographical coordinates (latitude and longitude) of each car's seller. Then, the price data were assigned to the respective NUTS1 and NUTS2 regions, given the information on their borders. The shapefile containing the geographical information on the regional borders was taken from the Eurostat. Using Germany as an example we illustrate that our estimates of regional income levels as well as of Gini indices display high, positive correlation with the official figures. This implies that the observed car prices can serve as a reasonably good proxy for income levels. In contrast to the traditional measures, our data can be fast and inexpensively retrieved from the web, and more importantly allow to estimate Gini indices at the NUTS2 level-something that has never been done for Germany before. We conclude that our approach to measuring regional inequality is a useful alternative source of information that could complement officially available measures.