This replication package provides code and partial data necessary to replicate the results of this paper.Replicating the results of this paper requires confidential non-public use data from the BLS Consumer Expenditure Survey. Due to confidentiality concerns, the Consumer Expenditure Survey public use files do not include the model name of household vehicles, only the make and year. The model name is a private use variable, makemodel , and the only private variable used in the paper. The vehicle model name is used to determine the fuel efficiency of vehicles owned by households in the sample and determine their assignment to treatment or control groups. Because the assignment into treatment and control groups is based on fuel efficiency being slightly above or below a cutoff, the use of this private variable is central to the results of the paper and they cannot be replicated without it.To aide in understanding how the code runs and results are generated, we have included code to generate a fake version of this private variable. This will allow the file clunkers_main.do to run the entire analysis and generate results, but these results will not match those of the paper because they use a randomly generated variable to assign households to treatment and control groups. Running the code without the BLS private variable is for expositional use only.

Smallest Geographic Unit: state

We begin with respondents to the Consumer Expenditure Survey and then limit the primary analysis to households who own a Clunker or Close-to-Clunker vehicle, as described in the Data and Methodology section of the manuscript referenced below.