Mediation analysis in recursive systems of distributed-lag linear regressions

Abstract Recursive systems of linear regressions is a consolidated methodology for mediation analysis, allowing to determine causal effects interest in closed form based on the regression coefficients. In dynamic perspective, distributed-lags can be added each order represent persisting over several periods. However, analysis case challenging, because depend time lag, and general procedure compute their lag distribution coefficients currently missing. this paper, we formalize rules perform recursive distributed-lag regressions, here called Distributed-lag Linear Models (DLRMs). Firstly, Directed Acyclic Graph (DAG) representation DLRM, then DAG-free algorithm proposed improve computational efficiency. Our applied DLRM representing impact pathways agricultural research expenditure towards poverty reduction rural areas.