Research interests

  My research focuses on the theory, inference, and application of causal analysis in dynamic econometric systems. In particular,

  The theory and inference of Granger causality.

  Identification of policy shocks in structural VAR models.

  Causal and counterfactual analysis in DSGEs.

  Rank testing theory and applications

Working papers

The Stochastic Wald Test (Job Market Paper)

  This paper develops a general and very simple framework for tests of rank. The framework allows for tests of rank to be based on virtually any reduced rank approximation. We show that all existing tests of rank, including cointegration rank tests, are, either exactly or asymptotically, of the form of a stochastic Wald test where the deterministic constraint in the Wald (1943) statistic is replaced by its estimator. Thus every rank test is asymptotically equivalent to a Wald test. We provide conditions that allow the researcher to ascertain when this equivalence holds under very general conditions, which greatly simplies the asymptotics of these tests. The framework allows for many new tests of rank as well as ones based on the new HAC robust hypothesis testing theory developed by Nicholas Kiefer and Timothy Vogelsang. An application is given to subspace causality testing in time series, where it is shown that the HAC robust test performs comparably to the bootstrap but is more accurate than the bootstrap in terms of accuracy of rank detection.

Causality Along Subspaces: Theory (Revise & Resubmit, Journal of Econometrics)

  This paper extends previous notions of causality to take into account the subspaces along which causality occurs as well as long run causality. The properties of these new notions of causality are extensively studied for a wide variety of time series processes. The paper then proves that the notions of stability, cointegration, and controllability can all be recast under the single framework of causality.

Causality Along Subspaces: Inference

  This paper extends the analysis of Dufour et al. (2006) in the direction of the more general notions of causality introduced by Al-Sadoon (2009). In particular, we propose new tests for Granger non-causality along subspaces in VAR processes at various horizons. The methodology is illustrated by reexamining the Bernanke & Mihov (1998) data set for forecast horizons of 1-24 months. We find that monetary policy predicts output growth and inflation only along a single direction in output-growth-inflation space for most of the forecast horizons we consider. We also find that the variations of non-borrowed reserves and the federal funds rate along certain directions have no predictive power for output growth and inflation at most of the horizons considered. Thus there is substantial scope for subspace causality analysis in macro data.

Matlab Code for "Causality Along Subspaces: Inference"

  This code allows the user to conduct subspace causality testing according to the theory laid out in Al-Sadoon (2010).