Most energy intensive industries are high carbon dioxide (CO2) emitters. As CO2 emissions show large discrepancies between regions, it is important to test for spatial dependence when estimating emissions. The present study thus examines spatial dependence of CO2 emissions for the twenty-seven governorates of Egypt over the period 2007-2016. Determinants of CO2 emissions include road network density, investments in fixed assets, the structure of economic activity, the share of females, and the various levels of education in the regions’ respective population. Incorporating the spatial dimension, as well as using female and educational levels amongst the determinants, are contributions to research done on CO2 emissions in Egypt. Estimation results indicate that CO2 emissions across governorates are spatially-random rather than spatially-dependent. The study therefore uses a generalized method of moments (GMM) dynamic panel model. The lag of CO2 emissions per capita, net fixed capital formation, share of illiterates in the working age population, and the share of services in economic activity are significant and have positive effects (increasing emissions). The share of females in working age population, and the share of agriculture in economic activity are significant and have negative effects (diminishing emissions). Policy implications based on the study results are given.