The Kyoto Protocol and Carbon Dioxide Emissions in New Zealand: A Synthetic Control Approach
On the 11th of December 1997 the International Community signed the Kyoto Protocol: an international environmental treaty that commits countries to reducing their greenhouse gas (GHG) emissions to mitigate the effects of human activity driven climate change. The Kyoto Protocol imposes individual GHG emissions reductions targets on developed countries for commitment period one (2008-2012). Reduction targets amount to an aggregate 5% reduction in GHG emissions for participating countries when compared to 1990 levels of GHG emissions. The Kyoto Protocol is criticised as insufficient, with criticisms focusing on its structure. The inclusion of flexibility mechanisms, unrestricted international emissions trading and the large endowment of emissions credits given to former Soviet Union countries are said to have created compliance costs that fail to encourage any real decrease in emissions. The withdrawal of the United States from the Kyoto Protocol in 2001, the largest GHG emitter at the time, furthered the worries that the Kyoto Protocol would result in “business-as-usual” emissions. I analyse the effects of a legally binding emissions reductions target on the carbon dioxide (CO_2) emissions of New Zealand. Formally I seek to answer if the legally binding emissions reductions targets of the Kyoto Protocol reduce the carbon dioxide emissions of New Zealand? I employ the Synthetic Control Method (SCM), in which a weighted average of an untreated series is used as a counterfactual estimate to a treated series, to estimate the CO_2 emissions of New Zealand had it not joined the Kyoto Protocol. Furthermore, I extended the set of variables previously used to estimate causal effects of the Kyoto Protocol to include variables that are considered crucial in forecasting GHG emissions in Computable General Equilibrium (CGE) analyses. I then seek to answer a second question, are causal analysis specifications for the Kyoto protocol improved be including variables used to forecast emissions in CGE modelling? My results provide no statistically significant evidence that legally binding emissions targets result in New Zealand experiencing a reduction in CO_2 emissions. Additionally, I show that the inclusion of common variables in CGE literature, when estimating Synthetic Controls for climate policy, provides a better pre-treatment fit and bring results closer to statistical significance. Furthermore, my results show that, contrary to previous work, the use of US state level data is not always preferable to country level data when using the SCM to estimate causal effects of climate policy.