ABSTRACT The production network, formed by inter-industrial input-output linkages, propagates and amplifies shocks. Accurately measuring the propagation effect is crucial for risk management. However, most empirical research measuring the effect rely on specific assumptions on parameters and production function forms, such as Cobb-Douglas, constant elasticity of substitution (CES) functions, leading to varying results depending on the setting of assumptions. This paper derives a flexible modified measuring method that does not make assumptions on specific production function forms, reducing subjectivity in the measuring results. The new method indicates that the propagation effect mainly depends on the heterogeneity of Domar weights across sectors in the production network, consistent with our empirical findings from the United States, China, Japan, Germany, and the United Kingdom. Due to stronger heterogeneity in its production networks, China has a significantly higher propagation effect compared to the other four countries. Additionally, we found that the weighted degree of production networks follows exponential distributions instead of the commonly assumed power-law distributions. This study contributes to a quantitative understanding of systemic risk within production networks.