From Economic Primitives to Econometric Modeling: A Time-Series Study of Population Dynamics

Taiwan has one of the lowest fertility rates in the world and is aging rapidly, which makes it important to understand the socioeconomic factors behind its declining births. To move beyond the familiar observation that economic conditions affect childbearing, this study borrows the cognitive systems framework as an interpretive lens, combining Kahneman's (2011) System 1 and System 2 with the recently proposed System 3 of socially conditioned, norm-driven behavior (Dyachenko et al., 2025). Using monthly data for Taiwan from 2007 to 2024, this study applies a four-step time-series strategy to examine how three channels, namely changes in unemployment, rent, and marriage, are related to year-on-year changes in monthly births. The strategy moves from a contemporaneous baseline model to specifications that successively add month fixed effects, theoretically motivated lags, and Newey-West robust standard errors. In contemporaneous models, none of the three variables is significantly associated with births. Once the explanatory variables are lagged by ten to twelve months, however, all three channels become significant, with unemployment and rent negatively related to subsequent birth growth and marriage positively related to it. These associations remain after accounting for autocorrelation and heteroskedasticity. Read through the lens, this pattern suggests that childbearing is not an impulsive System 1 reaction, but a deliberate System 2 process embedded in a System 3 environment of marriage norms and institutions. By incorporating recent data from the 2020s and examining several socioeconomic channels within a single framework, the study provides updated and exploratory evidence on Taiwan's fertility decline.