Applications of Granular Macro-Network Models: US-China Trade War and Covid-19 Impact

This thesis applies Granular Macro-Network Models to analyse the impacts of two major recent economic events – the US-China trade war and the COVID-19 pandemic – on sectoral and total GDP in the US and Chinese economies. Using the OECD Inter-country Input-Output (ICIO) database along with Leontief inverse coefficients and the Ghosh model, supply and demand shocks are introduced to estimate changes in economic output. A critical methodological contribution is developing the partial extraction method based on the hypothetical extraction method (HEM) to trace intermediate goods shock propagation (an advancement over traditional GDP models focused solely on final demand). The author also categorises tariff data and ICIO data by matching 6-digit Harmony System (HS) codes to input-output sectors and calculating sector-level weighted average tariffs. In Chapter 1, the trade war analysis, import demand changes from tariffs and elasticities are modelled. Chapter 2 extends this by assessing three trade response strategies – foreign trade diversion, domestic import substitution, and a mixed approach of both. Chapter 3 applies similar Leontief and Ghosh models to estimate COVID-19 shutdowns, adding empirically derived lockdown constraints differentiated by severity, lockdown duration, and fiscal interventions. Across analyses, results highlight the significance of interconnected production structures in propagating sectoral shocks. The applied models estimate granular national and industry-level impacts by quantifying total and sectoral GDP changes. This demonstrates how supply/demand disruptions to one sector can widely transmit through integrated macro-network models, capturing intermediate interdependencies absent in traditional GDP frameworks. This novel approach provides robust analytic capabilities for crisis scenario modelling and policy analysis focused explicitly on the interconnected intermediate good trade network – an essential contrast from existing final demand-centric GDP impact analyses. The predictive capabilities exhibited by this model suggest its potential for application to additional economic crisis situations that may arise in the future.