Sparsity-induced identification of factor-augmented VAR models
提出一种正则化因子增强VAR模型，通过因子载荷的稀疏性实现因子识别，并用货币政策冲击实证验证了模型效果。
This paper introduces a regularized factor-augmented vector autoregressive (RFAVAR) model which incorporates sparsity in the factor loadings. Within this framework, the factors can load on a subset of variables, thereby enabling factor identification and enhancing their economic interpretation. The proposed RFAVAR model allows to investigate the effects of structural shocks on economically interpretable factors and on all observed time series included in the model. We prove consistency for the estimators of the factor loadings, the covariance matrix of the idiosyncratic component, the factors, and the autoregressive parameters in the dynamic model. In an empirical application, we examine the effects of a monetary policy shock on a broad range of economically relevant variables. The identification of this shock is accomplished through a joint identification of the factor model and the structural innovations in the VAR model. The obtained impulse response functions align with the established economic rationale.