The empirical analysis of the economic interactions between factors of production, output and corresponding prices has received much attention over the last two decades. Most contributions in this area have agreed on the neoclassical principle of a representative optimizing firm and typically use theory-based structural equation models (SEM). A popular alternative to SEM is given by the vector autoregression (VAR) methodology. The most recent attempts to link the SEM approach with VAR analysis in the area of factor demands concentrate on single-equation models, whereas no effort has been devoted to compare these alternative approaches when a firm is assumed to face a multi-factor technology and to decide simultaneously the optimal quantity for each input. This paper bridges this gap. First, we illustrate how the SEM and the VAR approaches can both represent valid alternatives to model systems of dynamic factor demands. Second, we show how to apply both methodologies to estimate dynamic factor demands derived from a cost-minimizing capital-labour-energy-materials (KLEM) technology with adjustment costs (ADC) on the quasi-fixed capital factor. Third, we explain how to use both models to calculate some widely accepted indicators of the production structure of an economic sector, such as price and quantity elasticities, and alternative measures of ADC. In particular, we propose and discuss some theoretical and empirical justifications of the differences between observed elasticities, measures of ADC, and the assumption of exogeneity of output and/or input prices. Finally, we offer some suggestions for the applied researcher.