A mass point vehicle scrappage model

In air quality modeling, vehicle survival rates, also frequently referred to as retention rates, are very important for modeling mobile emissions inventories. The vehicle survival rate determines how many vehicles in the current fleet will survive to a future year, which, when added to new vehicle sales, provides a vehicle population forecast for a given year. In this study, we specify a mass point duration model that agencies like the California Air Resources Board (CARB) and the Environmental Protection Agency (EPA) can use, which will significantly improve the current practice for estimating vehicle survival rates. The data used for the model represents snapshots of passenger car smog check data collected between 1998 and 2002. We use a stratified sampling procedure to select a sample of 678 observations from the original vehicle population of 2.2 million for model specification. Using the sample, we applied a duration model of the Weibull form with two mass points to approximate the unobserved heterogeneity to estimate vehicle survival rates. The results of our vehicle scrappage model are consistent with evidence from the literature suggesting that vehicle age is an important variable influencing the scrappage decision. However, our study also shows that there are other variables that appear equally influential. To compare our model results with the CARB model, we also calculated the survival function and compared the average survival probability by age group to the survival rates used by CARB. From this, we observe that incorporating other variables in addition to vehicle age generally suggests a higher survival rate for the same age group compared to the survival rates used by CARB. (C) 2004 Elsevier Ltd. All rights reserved.