Economic growth projections are fundamental to long term investment planning by businesses and government. There is an extensive and well developed literature concerned with methodologies for projecting economic growth over short time horizons, but over long horizons, where the goal is to quantify the implications of a particular scenario, the literature is limited. This paper provides a review of publicly available projections of GDP per capita over long time horizons and compares them with projections from a multi-sector model of the world economy called the G-Cubed model. Although there appears to be some general agreement over methodology across the reviewed sources, the projections vary considerably, particularly for developing regions over long time horizons, highlighting the importance and influence of alternative model methodology and assumptions. The comparison with the G-Cubed model builds on previous research that has highlighted the importance of a detailed disaggregated approach to projecting output and productivity that accounts for the dynamic interactions between sectors and across economies. Long term economic issues are becoming increasingly important and economic growth projections are fundamental to both the design and the assessment of long term economic policy alternatives. In addition, our ability to analyse the impact of long term policy alternatives is linked to our ability to model and understand the impact of global economic shocks. Models need to be clear in their methodological design and assumptions, but they also need to be sufficiently complex to account for the important sectoral relationships and international linkages that drive economic growth.