Preliminary-Data Error and Monetary Aggregate Targeting

Preliminary monetary aggregate data are subject to subsequent revision because of errors from seasonal and nonseasonal sources that will only be known later. This article examines the characteristics of these revision errors and analyzes their possible effect on monetary policy. It is found that false signals, in the sense that the preliminary M1 growth rate falls within the tolerance band set monthly by the Federal Open Market Committee while the final revised figure lies outside the tolerance limits, or vice versa, occur about two-fifths of the time. It is shown that this proportion can be halved by the adoption of improved seasonal adjustment procedures.