Causality and aggregation in economics: the use of high dimensional panel data in micro-econometrics and macro-econometrics

This study proposes one plausible procedure to address two methodological issues,which are common in micro- and macro- econometric analyses, for the full realization ofresearch potential brought by recently available high dimensional data. To address the issue ofhow to infer the causal structure from empirical regularities, graphical causal models areproposed to inductively infer causal structure from non-temporal and non-experimental data.However, the (probabilistic) stability condition for the graphical causal models can be violatedfor high dimensional data, given that close co-movements and thus near deterministic relationsare oftentimes observed among variables in high dimensional data. Aggregation methods areproposed as one possible way to address this matter, allowing one to infer causal relationshipsamong disaggregated variables based on aggregated variables. Aggregation methods also arehelpful to address the issue of how to incorporate a large information set into an empirical model,given that econometric considerations, such as degrees-of-freedom and multicollinearity, requirean economy of parameters in empirical models. However, actual aggregation requires legitimateclassifications for interpretable and consistent aggregation.Based on the generalized condition for the consistent and interpretable aggregationderived from aggregation theory and statistical dimensional methods, we propose plausiblemethodological procedure to consistently address the two related issues of causal inference andactual aggregation procedures. Additional issues for empirical studies of micro-economics andmacro-economics are also discussed. The proposed procedure provides an inductive guidance forthe specification issues among the direct, inverse, and mixed demand systems and an inversedemand system, which is statistically supported, is identified for the consumer behavior of softdrink consumption. The proposed procedure also provides ways to incorporate large informationset into an empirical model with allowing structural understanding of U.S. macro-economy, which was difficult to obtain based on the previously used factor augmented vectorautoregressive (FAVAR) framework. The empirical results suggest the plausibility of theproposed method to incorporate large information sets into empirical studies by inductivelyaddressing multicollinearity problem in high dimensional data.