Are Small-Scale SVARs Useful for Business Cycle Analysis? Revisiting Nonfundamentalness
指出格兰杰因果检验在SVAR变量存在截面聚合或代理非可观测变量时有问题，提出替代检验方法，并用蒙特卡洛模拟和典型小规模SVAR模型重新评估其非基本性。
Nonfundamentalness arises when current and past values of the observables do not contain enough information to recover structural vector autoregressive (SVAR) disturbances. Using Granger causality tests, the literature suggested that several small-scale SVAR models are nonfundamental and thus not necessarily useful for business cycle analysis. We show that causality tests are problematic when SVAR variables cross-sectionally aggregate the variables of the underlying economy or proxy for nonobservables. We provide an alternative testing procedure, illustrate its properties with Monte Carlo simulations, and re-examine a prototypical small-scale SVAR model.