The forces defining the dynamics of a system expose systemic interconnections between the involved variables, often creating multiple interactions that cannot be adequately represented by isolated models. This research develops an integrated statistical framework based on Full Causal Effect (FCE) modelling and systems of simultaneous linear equations (SLE) to investigate how explanatory variables interact during linked events. The technique presented here can be applied to several different systems; the only condition is that the system to be analysed involves interaction between its variables.
This analysis focuses on the influence of the COVID-19 pandemic across fourteen major economies—the United States, China, Japan, Germany, India, the United Kingdom, France, Italy, Canada, Brazil, Australia, Spain, Mexico and the Netherlands—collectively representing approximately 75% of global GDP in the first quarter of 2020.
The research was developed through three complementary stages. First, epidemiological dynamics were obtained using compartmental pandemic models, including SIRDi formulations incorporating transmission, recovery, mortality, quarantines and immunisation effects. Analytical solutions and Particle Swarm Optimisation were used to solve the model. The pandemic analysis permitted the estimation and examination of spreading parameters and the evaluation of how changes in transmission and containment conditions influenced the timing and magnitude of the infection spread. Second, the resulting epidemiological indicators were integrated with macroeconomic and industrial variables. Combinations of instruments and predictors were evaluated using the Generalised Method of Moments, with statistically admissible equations identified through Hansen’s J tests. Valid equations were subsequently assembled into simultaneous linear systems and evaluated through least squares optimisation.
The framework was verified through Seemingly Unrelated Regression (SUR), allowing equations with correlated residual structures to be identified and combined into compatible systems of equations. Benchmark comparisons with reduced form dynamic panel models, conventional SUR and lagged specifications were used to examine the consistency of the statistical relationships. Finally, the industrial sector evidence was synthesised through a structured literature review, screening records, assessing full text studies and retaining those studies for relevant analysis of construction disruption, supply chains, labour, project performance, digital transformation and resilience.
The first level of analysis established causal relationships between pandemic and economic variables affecting construction productivity at global and continental levels. Recovery rate, share prices, foreign direct investment (FDI) and total spending generally supported industrial productivity, whereas inflation, mortality and pandemic related variables frequently increased disruption. The relative risk index (RRI) and relative productivity of the industry displayed a clear inverse pattern: risk increased while productivity declined during the first quarters of 2020, followed by a progressive reduction in risk and recovery in productivity. Polynomial models relating RRI to productivity provided close agreement between observed and forecast productivity, with coefficients of determination of 0.960 globally, 0.978 in America, 0.980 in Asia, 0.969 in Europe and 0.968 in Oceania.
The second level of analysis extended the framework from individual causal relationships and industrial productivity to contemporaneous systems of equations describing multiple variables simultaneously. The results demonstrated that epidemiological shocks propagated through employment, consumption, investment, trade, prices and financial markets, producing interconnected feedback mechanisms. Recovery, transmission and mortality rates generated effects on employment, inflation and private spending, while mobility, liquidity, government spending and financial market variables modified the transmission of the shock through the wider economy. The analysis also identified country level differences in causal structures. For example, employment in the United States was strongly associated with liquidity and recovery conditions; FDI formed an important pathway to exports in China; fiscal expenditure was associated with exports and private spending in Italy; and pandemic and financial variables played different roles across European, Asian and American economies. The relationships also changed over time: transmission and pandemic growth were particularly influential during the initial outbreak, recovery and mobility became more important during vaccination and reopening, while inflation, liquidity and private spending became increasingly influential during the later recovery period.
The third level of analysis consisted of a literature review which further demonstrated that industrial sector events resulted from interacting health, labour, financial, supply chain, regulatory and technological mechanisms rather than from isolated events. The evidence indicates that the mechanisms change between onset, adaptation and recovery phases and vary across countries and industrial subsectors. Supply chain flexibility, workforce adaptation, digitalisation, health and safety measures, public investment and emergency preparedness were consistently identified as mechanisms supporting continuity. The review provides a theoretical and empirical basis for applying FCE modelling for resilience and for developing adaptive strategies under future pandemics, geopolitical disruptions and other systemic events.
The findings demonstrate that pandemic resilience depends on the coordinated behaviour of health and economic systems rather than on isolated interventions. The proposed FCE–GMM–SUR–SLE framework provides a means of identifying multiple statistically admissible causal pathways, examining how their configurations change across time and space, and translating these relationships into risk and productivity measures. The research provides a mathematical foundation for dynamic risk assessment, causal optimisation and adaptive decision making under systemic shocks, with potential applications beyond COVID-19 to future pandemics, geopolitical disruptions, supply-chain failures, climate related events and other complex social and natural processes.
