Abstract:Summary Learning about cause and effect is arguably the main goal in applied econometrics. In practice, the validity of these causal inferences is contingent on a number of critical assumptions regarding the type of data that has been collected, and the substantive knowledge that is available about the phenomenon under investigation. For instance, unobserved confounding factors threaten the internal validity of estimates; data availability is often limited to nonrandom, selection-biased samples; causal effects need to be learned from surrogate experiments with imperfect compliance; and causal knowledge has to be extrapolated across structurally heterogeneous populations. A powerful and flexible causal inference framework is required in order to tackle all of these challenges, which plague essentially any data analysis to varying degrees. Building on the structural perspective on causality introduced by Haavelmo (1943) and the graph-theoretic approach proposed by Pearl (1995), the artificial intelligence (AI) literature has developed a wide array of techniques for causal inference that allow us to leverage information from various imperfect, heterogeneous, and biased data sources (Bareinboim and Pearl, 2016). In this paper, we review recent advances made in this literature that have the potential to contribute to econometric methodology along three broad dimensions. First, they provide a unified and comprehensive framework for causal learning, in which the above-mentioned problems can be addressed in generality. Second, due to their origin in AI, they come together with sound, efficient, and complete (to be formally defined) algorithmic criteria for automation of the corresponding identification task. And third, because of the nonparametric description of structural models that graph-theoretic approaches build on, they combine the analytical rigor of structural econometrics with the flexibility of the potential outcomes framework, and thus offer a valuable complement to these two literature streams.

Integrating Structural and Reduced-Form Methods in Empirical Finance

Ensemble Learning with Statistical and Structural Models

Merging Structural and Reduced-Form Models for Forecasting

Epistemic Limits of Empirical Finance: Causal Reductionism and Self-Reference

A Combined Approach to the Inference of Conditional Factor Models

Integrating Factor Models

Invited Commentary—Removing the Boundary Between Structural and Reduced-Form Models

Fixed-Effects in Empirical Accounting Research

Teaching Economics to the Machines

Valuation Algorithms for Structural Models of Financial Interconnectedness

On the Effects of Dimension Reduction Techniques on Some High-Dimensional Problems in Finance

Efficient Likelihood-based Estimation Via Annealing for Dynamic Structural Macrofinance Models

Adaptive Estimation and Uniform Confidence Bands for Nonparametric Structural Functions and Elasticities

Recent methods from statistical inference and machine learning to improve integrative modeling of macromolecular assemblies

Economic Motivations for Limited Dependent and Qualitative Variable Models*

Integrative Data Analysis Where Partial Covariates Have Complex Nonlinear Effects by Using Summary Information from an External Data

Metamodel Estimation Using Integrated Correlation Methods.

Causal inference and data fusion in econometrics

Efficient estimation of true fixed-effects stochastic frontier models

EM estimation of a Structural Equation Model

Estimation of Large Dimensional Conditional Factor Models in Finance