Wealth, cost, and misperception: Empirical estimation of three interaction channels in a financial-macroeconomic agent-based model
针对金融宏观经济主体基模型难以用不同频率数据估计的问题，提出混合频率模拟矩方法，基于美国1954-2025年数据量化了股市通过财富效应、成本效应和价格误判效应影响实体经济的三种渠道，并首次提供模型驱动的统计显著性证据。
Financial–macroeconomic agent-based models offer a promising avenue for understanding complex economic interactions, but their empirical estimation is challenging, particularly when the relevant data are observed at different frequencies. We address this gap by developing a mixed-frequency adaptation of the simulated method of moments that evaluates macroeconomic and financial moment conditions at their native quarterly and daily frequencies within a single objective function. Applying it to a stylized integrated agent-based model on U.S. data from 1954 to 2025, we quantify three key interaction channels between the financial and real sectors. The stock market influences the real economy through the wealth effect on consumption and the cost effect on firms’ financing. Conversely, the real economy influences the stock market through a price-misperception effect that shapes fundamental valuations. Our results provide the first model-based empirical evidence that all three channels are statistically significant, offering novel insights into macro-financial dynamics.