Accurate and complete data on the global atmospheric releases (emissions) of individual PCB congeners are essential to study source–receptor relationships and contamination patterns in remote areas, such as the Arctic. Information on the anthropogenic emissions may also be useful for the interpretation of measured levels and patterns of atmospheric PCBs. This study builds upon an accompanying paper, presenting an estimate of the global historical production and consumption of 22 PCB congeners (Breivik K, Sweetman A, Pacyna JM, Jones KC. Towards a global historical emission inventory for selected PCB congeners — a mass balance approach. 1. Global production and consumption. Sci Total Environ, submitted). Here, a dynamic mass balance model is elaborated, parameterised and applied in an attempt to estimate the historical anthropogenic emissions as a direct result of the widespread usage of PCBs for a period of 70 years. This paper presents details of the mass balance approach, along with a discussion of the major uncertainties. It is shown that the diversity of historical usage, disposal and accidental release pathways makes it an extremely difficult task to bridge the gap between consumption and emissions, resulting in an equally complex and diverse true emission pattern. These results may therefore at best represent order-of-magnitude estimates only and the results in absolute terms should be treated with great caution. In spite of these uncertainties, useful information can still be gained. The results suggest that temperature is an extremely important factor controlling both emission amount as well as the emission pattern of PCBs. In particular, the emissions of the more chlorinated (and persistent) PCB congeners appear to be significantly influenced by (uncontrolled) high-temperature sources. As a consequence, it may prove warranted to undertake further measures to avoid that contaminated materials are subject to elevated temperatures and uncontrolled burning. It is furthermore suggested that efforts should be directed towards a better characterisation and quantification of these potentially important release pathways. In addition, alternative approaches to bridge the gap between consumption and emissions in quantitative terms should be considered in order to improve these estimates.