Factoring in the Micro: A Transaction‐Level Dynamic Factor Approach to the Decomposition of Export Volatility

This paper analyzes the export volatility sources estimating a dynamic factor modelon transaction-level data. Utilizing an exhaustive dataset of French export transactionsfrom 1993 to 2017, we reconstruct the latent factors space associated with globaland destination-specific macroeconomic shocks through a Quasi-Maximum likelihoodapproach which allows accommodating both the high share of missing values and thehigh dimensionality of the microeconomic time series. The estimated parameters are thenused to derive a volatility decomposition of the aggregate and firm-level export growthrates, highlighting structural spatial patterns and the role of geographical diversificationin mitigating export risks.