The authors narrow the disagreement about the empirical relevance of the liquidity effect. They thoroughly review the empirical literature on the liquidity effect, differentiating between single-equation and systems-modeling approaches. Arguing for a systems approach, they focus their attention on the estimation of vector autoregressions (VARs), which is the most promising tool for identifying a statistically significant and empirically relevant liquidity effect. The authors go on to confirm what previous empirical work suggests, namely, that finding a statistically significant liquidity effect depends critically on the definition of money used. A statistically significant liquidity effect is generally found only with nonborrowed reserves or the ratio of nonborrowed to total reserves. No statistically significant liquidity effect is found using total reserves, the monetary base, or M1. The authors also find that the magnitude of the estimated liquidity effect depends on the sample period.