Abstract:
This paper explores the relationship between per capita carbon
dioxide emissions and per capita income for Sri Lanka, Japan and United
States. Unit root tests carried out showed that these variables are nonstationary.
Since non-stationary data outperform regression models
developed with them, regression models were developed in this study
with stationarized variables. For the best fitted models for Sri Lanka,
Japan and United States, it was found that the per capita emissions were
driven mainly by its own autoregressive term, GDP per capita and its
autoregressive term. For the best fitted models, statistical characteristics
such as Mean Square Error, Akaike Information Criterion and Bayesian
Information Criterion took the least values among the models studied.
Also, the residuals of the best fitted models were found to posses the
characteristics of white noise and they were normally distributed.
