Costs of cars are among the most relevant factors influencing travel behavior. However, there is a lack of data about the true costs of car ownership and specifically on how these costs are distributed across different vehicles and across the population. This paper presents a multistage method for imputing car costs by cost item in a German national travel survey data set. Based on vehicle information reported by survey participants, we assign costs to each of the three thousand cars in the data set using the most comprehensive German vehicle cost data base. In addition to combining different data sets, we use model based imputation methods. In order to validate the average costs for private vehicles we analyze the German income and expenditure survey EVS. The average total cost of ownership for a private car in Germany is about 310 Euros per month. This translates to about 30 Eurocents per auto-km. About one third of the costs are fuel, another third is depreciation, and the rest are other mainly fixed costs (insurance, tax, repair and maintenance). However, the cost distribution is strongly skewed with a long tail to the right. He ... mehrnce, the majority of motorists pay less than average for their private vehicles while few pay more and evidently some pay a lot more. This imputation approach delivers unprecedented vehicle cost information in particular with regard to the distribution of vehicle costs. Such data is key for understanding the fundamentals of mobility choices.