Dr. Annika Camehl
Period: December 2025 – September 2026
Funded by: NWO
This project develops a novel econometric framework to improve structural identification in macroeconomic time-series models by systematically integrating micro-level causal evidence with aggregate data. Micro-level evidence provides a natural complementary source of information, as empirical designs based on cross-sectional exposure, difference-in-differences, or event studies exploit heterogeneity across firms, regions, or households to estimate causal effects of similar shocks. However, its relationship to aggregate structural parameters is not straightforward as such estimates may not map cleanly into aggregate dynamics. This project therefore asks how disaggregated micro data can be used to inform structural parameters in the presence of potential misspecification.
To address this question, the project focuses on structural vector autoregressions (SVARs). It proposes a composite likelihood approach that links micro-level transmission mechanisms to aggregate impulse responses while explicitly allowing for imperfect alignment between the two. Rather than treating micro estimates as fixed calibrations or imposing them as strong priors, the method incorporates them as probabilistic restrictions. This preserves uncertainty, accommodates potential misspecification, and avoids distortions that arise when reduced-form micro evidence is mapped too rigidly into macro parameters.
By jointly estimating micro and macro information, the framework strengthens identification in settings where traditional approaches struggle, such as weakly identified models. It improves statistical efficiency, sharpens inference, and reduces issues like multimodality in parameter estimates. The method is flexible and can incorporate evidence from a wide range of empirical designs.
Beyond its methodological contribution, the project provides a practical tool for analyzing the macroeconomic effects of rare but high-impact events. By using rich micro-level data, it enables more credible estimation of aggregate responses to shocks such as natural disasters, trade disruptions, or abrupt policy changes, where macro data alone is often insufficient.
