An Empirical Test for Inter-State Carbon Dioxide Emissions Leakage from the Regional Greenhouse Gas Initiative
Andrew Kindle, Rensselaer Polytechnic Institute Economics, Phone 1-585-506-5525, E-mail:
Daniel Shawhan, Rensselaer Polytechnic Institute Economics, Phone 1-518-331-6186, E-mail:
Michael Swider, New York Independent System Operator, E-mail:
Overview
The Regional Greenhouse Gas Initiative (RGGI) is a carbon dioxide (CO2) cap-and-trade program that applies to the emissions of the electric power industry in nine states in the northeastern United States. Participating states require all electric generation units with capacities greater than 25 megawatts (MW) to obtain and turn in one permit or “allowance” for each ton of their CO2 emissions. The allowances can be bought and sold in trading markets. The need for these allowances increases the marginal cost per MW-hour (MWh) of emitting generators, in proportion to their CO2 emissions per MWh. If this cost of RGGI compliance causes participating states to import more power from non-participating states and provinces, and if the incremental generation in those non-participating states produces CO2 emissions, the result is a transfer of emissions production from participating to non-participating states, which can be called“inter-regional emissions leakage”. This analysis of inter-regional emissions leakage is applicable on regional and global scales as it could imply reductions in the effectiveness of a cap-and-trade program covering one region or country and not covering an adjacent, interconnected region or country.
Methodology
Pennsylvania has an abundance of coal fired power plants and is not part of the RGGI program but is bordered on three sides by states that were part of the program during our analyzed time period. To test for leakage we examine whether a higher RGGI allowance price (zero before the program, fluctuating after) is associated with higher Pennsylvania CO2 emissions or higher power flows from Pennsylvania to the neighboring RGGI state of New York. These two states share a long border. We use hourly data covering the time span of January 2006 to September 2010 which comes from the EPA, EIA, and proprietary sources. We estimate econometric models with ARMA disturbances for three dependent variables: Pennsylvania CO2 emissions, day ahead scheduled flows from Pennsylvania to New York, and real time flows from Pennsylvania to New York. We control for the effects of other variables such as New York and Pennsylvania hourly load, fuel prices, SO2 and NOx allowance prices, and non-emitting generation.
Results
We find no evidence of leakage. Neither Mmodels of Pennsylvania CO2 emissions indicatenor emission leakage. In fact, a higher RGGI allowance price is associated with lower Pennsylvania CO2 emissions. Mmodels of the power flow from Pennsylvania to New York do not find a statistically significantly association between the RGGI allowance price and power flows across the border. All of the models fit the data well with R2 values for the emissions models around 0.9, and for the flows models in the range of 0.7 to 0.8.
Conclusions
Based upon the present analysis, there is no evidence of emissions leakage to Pennsylvania due to the RGGI program. Given an increase in the marginal cost of fuel burning generation in New York and not in Pennsylvania from the RGGI allowance program there is an economic incentive for a change in cross border flows. This means that the RGGI allowance price was not high enough to impact flows or the lack of variation in the allowance price (cents at a time) resulted in such small impacts on the cross border flows as to make the impact undetectable. The model for CO2 emissions in Pennsylvania actually indicates negative emissions leakage as indicated by the significant and negative coefficient on the RGGI allowance price variable. Such negative leakage could result from a change in the cross-border scheduling of power flows by the system operators with responsibility for Pennsylvania and the neighboring RGGI states.