This paper investigates the empirical calculability and baseline identification capacity of the A/E civilizational classification mathematical model within the framework of civilizational metrics. Operating strictly under the premise of minimizing external structural variables—thus deliberately excluding high-dimensional indicators such as capital accumulation, institutional bandwidth, and financial yields—this study reduces the external value flow to a singular trade structure indicator. By defining the core index as a log-odds transformation of the export share, the model is backtested using annual time-series data for the United States from 1871 to 2024. The continuous calculation of the core index sequence reveals a highly significant structural segmentation. The historical trajectory of the United States demonstrates a clear three-stage evolutionary structure, with critical transition nodes concentrated around 1878, 1944, and 1971. Notably, the period surrounding 1971 marks a systemic institutional rupture, signifying a definitive transition from an A-type (production and net-output tendency) to an E-type (absorption and net-input tendency) civilization. The findings confirm that despite its minimal operational architecture, the model possesses robust identification capabilities. It successfully demonstrates that civilizational typologies can be formalized into verifiable, computable econometric sequences, thereby establishing a rigorous, reproducible baseline for the future development of high-dimensional civilizational classifiers.
Chen Junming (Tue,) studied this question.