Several recent articles have made the case that monetary policy should not ignore information about the stance of monetary policy found in a properly specified measure of the money supply. Properly specified means (1) appropriately grouping monetary assets into aggregates based on how money-like they are and (2) recognizing that there are both retail and institutional money assets. Standard measures of money like M2 ignore both issues. The graph above shows monetary aggregates from the Center for Financial Stability that do account for both.
Peter Ireland made the case for these measures at the most recent Shadow Open Market Committee while Doug Irwin did so on the pages of the Financial Times. Steve Hanke, meanwhile, invokes them to warn that the Fed may be undermining their growth by imposing stricter bank capital requirements. My own thoughts on properly measuring money can be found to right under the "Safe Assets, Money, and the Great Recession" header.
Update I: Just to be clear, I am not advocating money supply targeting. I still prefer a NGDP level target.
Update II: Michael Belongia reminds me that the targeting the money supply--an intermediate target--should not be confused with the ultimate goal of monetary policy, stable NGDP growth. Point taken. See his paper with Peter Ireland where they show how the Fed could implement NGDP targeting by stabilizing a Divisia measure of the money supply.