This study is concerned with the analysis of the long-run substitution pattern of primary energy sources for the Italian economy, over the period of 1960-1994. A neoclassical model, set in the cost function approach, has been used to retrieve the energy inputs derived demand functions, via Shephard's lemma, using a translog cost function specification. Four primary energy sources have been considered, namely, oil, electricity, natural gas and coal. Recent advances in time series econometric theory have provided tools devices for modelling long-run equilibrium relationships and their associated short-run dynamics jointly. The Engle and Granger (1987) and the Engle and Yoo (1989) cointegration approach has been utilised in this study to estimate the long-run share relationships, while the general to specific methodology has been followed to derive error correction formulations for the adjustment processes. Extensions to time-varying parameter cointegration, carried out in the framework of the structural time series approach, have also been considered. The applications of traditional and time-varying parameter cointegration to the Italian energy market are the main sources of originality of this work. The study is divided into three main parts. The first part introduces the economic and econometric frameworks employed in the analysis. The second part is concerned with the actual empirical analysis. This consists of data description, the structural time series approach and the application of traditional and time-varying parameter cointegration theory to estimate a derived factor demand model. Finally, the third part summarises and discusses the results of the analysis.