Evolution of trade theories and technical change endogeneity in computable general equilibrium & integrated assessment models

Global trade modelling has relied heavily on the Armington assumption. While analytically simple, this fails to capture firm-level productivity dynamics or the economic impact of new export varieties. Consequently, computational economists are increasingly exploring modern trade theories like the Krugman and Melitz models: Krugman emphasises demand-side variety but faces limitations in complex multi-region applications due to symmetric variety assumptions, while Melitz is far more comprehensive, integrating both demand-side variety and supply-side productivity effects. Despite their benefits, integrating these micro-founded models into large-scale quantitative systems such as integrated assessment and computable general equilibrium models remains computationally demanding. Likewise, representation of technological change has been transitioning from purely exogenous assumptions toward endogenisation; today, large-scale macroeconomic models increasingly use endogenous mechanisms based on learning-by-doing (where cumulative production increases efficiency) and R&D-driven knowledge stock accumulation. Recent literature highlights the critical importance of knowledge spillovers transmitted across sectors and borders via trade, foreign direct investment, and new product varieties. This intersection of trade and technology is recognised as a central driver of embodied technology transfer. Additionally, advanced models employ directed technical change to endogenise innovation toward specific economic factors in response to emerging policy and price feedback loops. Ultimately, incorporating these dynamic, firm-level trade structures and endogenous innovation mechanisms provides a more realistic, empirically consistent framework for evaluating modern economic policies. In this context, we carry out systematic literature reviews across titles, abstracts, and keywords in Scopus, on trade theory and endogenous technological learning, to capture the state-of-the-art while assessing how economic modelling has advanced in time. We also carry out a targeted survey across EU-funded research projects that have implemented or experimented with relevant modelling developments and innovations, alongside a structured exploration of well-known model inventories and platforms. We synthesise our findings into actionable takeaways.