Supplementary Material: Identification of Partially Observed Linear Causal Models Jeffrey Adams 1, Niels Richard Hansen

Let us present the complete theorem first, and then give its proof. We are now ready to present Theorem 1. Theorem 1 But since F induces a different DAG, F is not identified up to trivialities. Proposition 4. F or any graph G there exists F F There are two cases to consider. The backward direction is obvious. This follows from definitions and acyclicity.1.4.5 Proof of Theorem 3 Theorem 3. Then F is identifiable up to trivialities.