THE FUTURE TRAJECTORY OF U.S. CO₂ EMISSIONS: THE ROLE OF STATE VS. AGGREGATE INFORMATION*

Abstract

This paper provides comparisons of a variety of time‐series methods for short‐run forecasts of the main greenhouse gas, carbon dioxide, for the United States, using a recently released state‐level data set from 1960–2001. We test the out‐of‐sample performance of univariate and multivariate forecasting models by aggregating state‐level forecasts versus forecasting the aggregate directly. We find evidence that forecasting the disaggregate series and accounting for spatial effects drastically improves forecasting performance under root mean squared forecast error loss. Based on the in‐sample observations we attempt to explain the emergence of voluntary efforts by states to reduce greenhouse gas emissions. We find evidence that states with decreasing per capita emissions and a “greener” median voter are more likely to push toward voluntary cutbacks in emissions.

Forecasting the path of China's CO2 emissions using province-level information

“…The model using aggregate data has a MSE 43.38% higher than the same model estimated using province level data, suggesting large gains from disaggregation, which is consistent with the literature (e.g. Marcellino et al, 2003;Auffhammer and Steinhauser, 2007;Carson et al, 2005).…”

“…Maddison (2006) shows that per capita emissions of criteria pollutants depend on emissions in neighboring countries. Auffhammer and Steinhauser (2007) show that allowing for dependence in aggregate CO 2 emissions across US states uniformly improves model forecast performance for all specifications considered. Without deciding on a specific structure a priori, we let the model selection criterion decide, which model fits the data best.…”

“…The "best" model is chosen based on out of sample forecast performance and a set of information theoretic model selection criteria. Third, we allow for spatial dependence in emissions across provinces, which has been shown to improve forecasts of aggregate variables if there is sufficient heterogeneity at lower levels of aggregation (Marcellino, Stock and Watson, 2003;Auffhammer and Steinhauser, 2007;Carson, Cenesizoglu and Parker, 2005), which we will show is the case here. Finally, our modeling approach uses an annually updated and publicly available source of data, which allows for frequently updateable forecasts and model improvement.…”

Chapter 4. Economic Considerations: Cost-Effective and Efficient Climate Policies

“…This last feature is particularly important given the experience with other carbon markets around the world, which have experienced both volatility and periods of very low prices [32]. Building on the emissions forecasting work in Auffhammer and Steinhauser [33] and Auffhammer and Carson [34], Borenstein et al [24] find that the combination of multiple regulations, such as energy efficiency programs and renewable energy requirements with a market-based regulation such as cap-andtrade increases the likelihood of extreme price outcomes in the cap-and-trade market. One implication is the importance of a robust price-containment mechanism that can ensure a relatively stable carbon price in order to long-run investment in carbon reduction technologies.…”

Peaking China’s CO2 Emissions: Trends to 2030 and Mitigation Potential

“…For example, by using environmental Kuznets curve hypothesis, researchers have analyzed the importance of state-level relationship between emission and income [9], spatial interaction [10] and spatial dependence between states [11,12]. Meta-analysis is another statistical technique to combine the results of several studies that address a set of related research hypotheses [13,14].…”