This research paper presents the results of a joint Australian Bureau of Statistics and Bureau of Transport and Regional Economics project which investigated the disparities between the Survey of Motor Vehicle Use (SMVU) estimates of total petrol consumption and annual automotive gasoline sales data, published by the Department of Industry, Tourism and Resources.
The quality of SMVU estimates are subject to debate amongst transport analysts. Since 1991 the petrol consumption estimates for registered vehicle use have been lower than Total Petrol Sales (TPS), by between 5 and 15 per cent. This disparity could not to be explained by sampling error, petrol consumed in off-road uses, or by unregistered and out-of-scope motor vehicle use.
The investigation focused on the 1998-2003 surveys. Two broad aspects of data quality were investigated: (i) differences in scope between the Survey of Motor Vehicle Use (SMVU) and automotive gasoline sales collections; and (ii) coverage of the SMVU. The purpose of the analysis was to identify where scope and coverage factors may impact upon the disparity between SMVU petrol consumption estimates and total petrol sales.
The investigation concludes that significant amounts of petrol are consumed by out-of-scope vehicles and off-road usage. Regression analysis was used to identify if the SMVU methodology is adequately accounting for the distance traveled by newest vehicles in the population, and concludes that the reduced availability of new vehicles in the sample acts to reduce distance traveled and petrol consumption estimates.
The paper concludes that the majority of the disparity between the SMVU and TPS collections is generally accounted for by petrol use in out-of-scope vehicles and off-road vehicles and machinery and by survey frame procedures (particularly the underestimation of new vehicle usage).