The Heterogeneous Effects of Government Size and Press Freedom on Corruption in Sub-Saharan Africa: Method of Moment Quantile Regression Approach

Abstract

Although an active body of literature exists on the relationship between government size and corruption, different findings proliferate the literature due to the nuances of geographical context, methods, and variables employed. The paper employs the novel method of moment quantile regression with fixed effects over the period 1984–2018 to investigate the dynamic impact of government size and press freedom on corruption in Sub-Saharan African (SSA) countries. The main contribution is to examine the impact of government size and press freedom across the conditional distribution of corruption in SSA. The findings show that the impacts of government size and press freedom are heterogeneous across the conditional quantile distribution of corruption. Among the most corrupt countries in SSA, larger government size and press freedom significantly increase corruption. Some policy recommendations are offered based on these findings.

Exploring the nexus between natural resource depletion, renewable energy use, and environmental degradation in sub-Saharan Africa

“…It is clear that the use of a linear regression model tells us the average/mean relationship between the explanatory variables (i.e., natural resource depletion, renewable energy consumption) and the dependent variable (environmental degradation). Their estimation method relies on the dependent variable’s central distribution tendency without integrating for the upper and lower ranges (see Amegavi 2022 ). This also means that the linear regression estimation method does not take into account countries with higher/lower (natural resource depletion, renewable energy, and environmental degradation) than medium countries.…”

“…This also means that the linear regression estimation method does not take into account countries with higher/lower (natural resource depletion, renewable energy, and environmental degradation) than medium countries. This can cause overestimation or underestimation of regression coefficients (Sarkodie and Strezov 2019 ), as all data cannot be fitted to reflect reality and distorting some important information (see Amegavi 2022 ). For this reason, the panel quantile regression approach is used to address the limitations of the standard linear regression technique.…”

“…In this regard, our study employs Powell ( 2020 ) generalized quantile regression to address the limitations of the standard linear regression technique (i.e., ARDL, FMOLS, DOLS, AMG, GMM estimator, DID and PSM approaches). The quantile regression method is based on the assumption that the influence of the explanatory variables varies along the conditional distribution of the dependent variable (Amegavi 2022 ; Koenker and Bassett 1978).…”

Exploring the nexus between natural resources depletion, renewable energy and environmental degradation in sub-Saharan Africa: A panel quantile regression

“…is clear the use of a linear regression model tells us the average/mean relationship between the explanatory variables (i.e., natural resource depletion, renewable energy consumption) and the dependent variables. Their estimation method relies on the dependent variable's central distribution tendency without integrating for the upper and lower ranges (See, Amegavi, 2022). This also means that the linear regression estimation method does not take into account countries with higher/lower (natural resource depletion, renewable energy, and environmental degradation) than medium countries.…”

“…In the estimation process, the panel quantiles divide the data into nine different quantiles (10th, 20th, 30 th, 40th, 50th, 60th, 70th, 80th, and 90th ) to explore the nexus between natural resources depletion, renewable energy and environmental degradation while controlling other covariates such as trade, industrialization and economic growth. In other words, the panel quantiles show an observation of data into intervals values, whereas the country performance indicates the magnitude at a median for all countries at 50% quantile compared to other countries (Amegavi, 2022). Therefore, countries with lower and higher quantiles than the median (50th Quantiles) can be described as having worse or better performance.…”