This study examines the impacts of the population age groups of 20-34, 35-49, 50-64 and 65-79 on the per capita energy use of the oil-exporting countries of Commonwealth Independent States: Azerbaijan, Kazakhstan and Russia employing the modified-STIRPAT framework. Considering that estimations using non-stationary data may yield spurious results, unlike many prior STIRPAT studies, we explore integration and cointegration properties of the data and then estimate long- and short-run elasticities as well as speed of adjustment coefficients. Since our time series analysis covers only 23 observations (1990-2012), as a robustness check, we also conduct panel data analysis by pooling the mentioned countries data with that for members of Organization of Petroleum Exporting Countries. We apply the Autoregressive Distributed Lags Bounds Testing approach in the time series analysis and Pooled Mean Group estimator in the panel analysis, both are superior in small samples. The findings from the time series analysis are supported by those from the panel data analysis. According to the results, there is cointegrated relationship among the variables. The age groups together with affluence and oil prices have statistically significant impacts on the per capita energy use in the selected countries. Moreover, we find the speed of adjustments exhibiting different magnitudes for different countries depending on which population age group is considered. The findings suggest that policymakers should pay special attention to the population age groups of 35-49 and 50-64, as they have a large effect on per capita energy use. Since these groups are the main part of the working age population, increase in their energy consumption is likely to lead to economic growth. Furthermore, the policymakers should take into consideration the finding that speed of adjustments towards an equilibrium path is quite high. It implies that any policy related shocks to the per capita energy use relationship could disappear within a year or even sooner. © 2015 Elsevier B.V.