Statistical Analysis of the U.S. CO2 Emissions Using State-Level Data

Loading...
Thumbnail Image

Date

2010-12-09

Advisors

Journal Title

Journal ISSN

Volume Title

Repository Usage Stats

280
views
768
downloads

Abstract

Past statistical modeling of carbon dioxide (CO2) emissions has primarily followed the reduced-form specification in the Environmental Kuznets Curve literature. This traditional approach relies on the assumption of a homogenous slope relationship between CO2 emissions and economic growth across different geographical/political regions. This study uses a panel data comprised of the fifty U.S. states and the District of Columbia and challenges the homogenous slope assumption. By an innovative multilevel modeling approach, it is possible to model heterogeneous slopes for each state. Moreover, the multilevel models can partially explain the variation in slopes by incorporating state economic compositions as state-level explanatory variables.

The modeling results of this paper show that six out of ten of the major economic sectors are significant state-level explanatory variables of CO2-GDP slope variations with their explanatory capability ranked from high to low as: services, transportation, finance, agriculture, energy- intensive manufacture, and mining. Two of these six sectors, finance and services, have negative effects on the CO2-GDP slopes, while the other four sectors have positive effects. This study demonstrates the feasibility of studying CO2 emission trends using multilevel models; however, this study also has some weaknesses in that it fails to account for certain state-level covariates and temporal changes of economic compositions. Future research should aim to overcome these shortcomings.

Description

Provenance

Subjects

Citation

Citation

Kuo, Peng-Yu (2010). Statistical Analysis of the U.S. CO2 Emissions Using State-Level Data. Master's project, Duke University. Retrieved from https://hdl.handle.net/10161/2866.


Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.