This paper offers a novel contribution to the literature on Marginal Emission Factors (MEF) by proposing a robust empirical methodology for their estimation across both time and space. Our Autoregressive Integrated Moving Average models with time-effects not only outperforms the established models in the economics literature but it also proves more reliable than variations adopted in the field of engineering. Utilising half-hourly data on carbon emissions and generation in Great Britain, the results allow us to identify a more stable path of MEFs than obtained with existing methodologies. We also estimate marginal emission effects over subsequent time periods (intra-day), rather than focussing only on individual settlement periods (inter-day). This allows us to evaluate the annual cycle of emissions as a result of changes in the economic and social activity which drives demand. Moreover, the reliability of our approach is further confirmed upon exploring the cross-country context. Indeed, our methodology proves reliable when applied to the case of Italy, which is characterised by a different data generation process. Crucially, we provide a more robust basis for valuing actual carbon emission reductions, especially in electricity systems with high penetration of intermittent renewable technologies. • A robust method is proposed to estimate marginal emission factors. • Our ARIMA outperforms established models for MEFs estimation. • Consistent results are obtained when using both UK and Italian data. • We provide a robust basis for valuing actual carbon emission reductions.