The determinants of CO₂ emissions in MENA countries: a responsiveness scores approach

ABSTRACT

This paper examines the relationship among carbon dioxide (CO₂) emissions, GDP, and energy in the Middle East and North Africa (MENA) countries by using a Responsiveness Scores (RS) approach. Empirical results over the period 1971–2013 suggest that GDP per capita and energy consumption show positive RSs, while trade and urban population negative ones. Moreover, energy consumption and urban population reveal moderate increasing returns to scale, while GDP per capita exhibits decreasing positive returns. Furthermore, three-way factors analysis sets out that most of the countries lays on regions with moderate negative Total Responsiveness Scores (TRS). This means that when all factors are jointly increased, CO₂ emissions have a moderate decrease. In addition, some GCC countries present a different pattern compared to the average pattern of MENA countries. Finally, radar plots indicate that, overall, RS pattern over factors is moderately heterogeneous within GCC countries, with larger variability appearing in the response to urban population and GDP.

KEYWORDS:

• Carbon dioxide emissions
• energy consumption
• economic growth
• responsiveness scores
• MENA

Supplemental material

Supplementary data for this article can be accessed here

Notes

1. It is true that, under certain conditions, a simple bivariate regression could not lead to biased estimates of the coefficients (intercept and slope). However, as the Simpson’s paradox (Simpson 1951) makes clear, not taking into account the presence of possible ‘confounders’ can lead to paradoxical conclusions such as a positive effect of x on y in sample subgroups, and a negative one in the entire sample.

2. As beta-coefficients are measured in standard deviations, they can be compared. The meaning of a beta-coefficient is straightforward: suppose that in a regression of y on x the beta is found to be equal to 0.3, then it means that one standard deviation increase in x leads to a 0.3 standard deviation increase in the predicted y, with all other variables in the model held constant (ceteris paribus assumption).

3. http://databank.worldbank.org/data/reports.aspx?source=world-development-indicators.