Carbon Taxes and CO2 Emissions: A Replication of Andersson (American Economic Journal: Economic Policy, 2019)

Do carbon taxes reduce CO2 emissions in the countries that adopt it? Andersson (2019. Carbon taxes and CO2 emissions: Sweden as a case study. American Economic Journal: Economic Policy, 11(4), 1–30) provides a clear, affirmative answer. His article has been widely cited as evidence that carbon taxes “work.” To check whether the estimates from Andersson (2019) are reliable, I replicate his article using its publicly available data and codes. I modify his synthetic control method (SCM) by using a more restricted set of control units (excluding one potentially treated unit). I also use a more efficient methodology, the Prais-Winsten estimator, to estimate price effects on gasoline consumption. In addition, I compute prediction intervals (PIs) and add these to the SCM estimates, using the newly developed scpi R package. My best estimate is that carbon taxes reduced per-capita CO2 emissions in Sweden’s transport sector by 7.7%, confirming Andersson’s main finding. I then extend Andersson’s approach to the Norwegian transport sector, estimating a smaller reduction of 2.4%. However, this effect falls within the PIs of the estimates assuming no carbon taxes. When I extend the analysis to the national level in Sweden, I estimate wide PIs and obtain inconclusive results.