This paper draws attention to two problems associated with the use of instrumental variables. First, using potential instruments that explain little of the variation in the endogenous explanatory variables can lead to large inconsistencies of the IV estimates even if only a weak relationship exists between the instruments and the residual. Second, finite-sample IV estimates are biased in the same direction as OLS estimates with the biases converging as the r-squared between the instruments and the explanatory variable approaches zero. To gauge the severity of both problems, we suggest that both the partial r-squared and the F statistic on the exlcuded instruments from the first stage be reported as approximate guides to the quality of the IV estimates.
Sign up for Brad Delong's (general) mailing list