The present paper considers an Italian dataset with an annual frequency from 1861 to 2000. It implements Granger non-causality tests between energy consumption and output contrasting methods allowing for structural change with those imposing parameter stability throughout the sample. Though some econometric details can differ, results have clear policy implications. Energy conservation policies hasten an underlying tendency of the economy towards a more efficient use of fossil fuels. The abandonment of traditional energy carriers was a positive change. The challenge for renewable energy is to diversify among different sources and to overcome possible social acceptance problems.