Working Paper Series no. 1017. This paper investigates how investors process central bank announcements when present and future policy signals conflict. We classify FOMC statements by whether signals about the policy path reverse current-decision news. Reversal statements account for nearly half of meetings and drive a disproportionate share of monetary policy effects on long-term interest rates, while non-reversal meetings move stock prices and short-term yields. This fundamental heterogeneity extends to the ECB and Bank of England. Further evidence suggests reversal statements convey information about the pace of future policy adjustments, with content in the joint distribution of present and future signals, not their individual levels.This paper investigates how investors process central bank announcements when present and future policy signals conflict. We classify FOMC statements by whether signals about the policy path reverse current-decision news. Reversal statements account for nearly half of meetings and drive a disproportionate share of monetary policy effects on long-term interest rates, while non-reversal meetings move stock prices and short-term yields. This fundamental heterogeneity extends to the ECB and Bank of England. Further evidence suggests reversal statements convey information about the pace of future policy adjustments, with content in the joint distribution of present and future signals, not their individual levels.
Figure 1. Long-term interest rates and monetary surprises
Non-Technical Summary
This paper investigates how investors process Federal Reserve’s policy announcements when the central bank simultaneously conveys conflicting signals about the present and the future stance of monetary policy. We propose a novel classification of policy announcements, over the period from 1994 to 2026, that identifies two distinct statement types.
Central bank announcements convey two types of information at once: news about the current interest rate decision and signals about the future path of policy. These two pieces of information may not point in the same direction. A central bank can raise rates today while signaling that future hikes will be fewer or slower than expected or hold rates unchanged while suggesting that the pace of future tightening will be faster than anticipated. We call such announcements reversal statements. All others are direction-preserving statements.
This novel classification is based on the interplay between the Target factor (news about the current rate decision) and the Path factor (signals about the future policy path orthogonal to the current decision) from Gürkaynak, Sack, and Swanson (2005). A reversal statement is one where the Path surprise moves in the opposite direction to the Target surprise and dominates it in magnitude, so that the net signal is primarily forward-looking. This classification is simple and transparent, and two model-free alternatives that bypass the factor model entirely deliver the same results.
The main finding is stark. Reversal statements account for nearly half of all scheduled FOMC meetings and drive a disproportionate share of monetary policy's effect on long-term interest rates. A 10 basis point monetary surprise from a reversal statement moves 10-year Treasury yields by 6.7 basis points, more than twice the effect from non-reversal statements. Direction-preserving statements, by contrast, primarily move stock prices and short-term rates but have little effect on long-term yields. This heterogeneity extends to the ECB and the Bank of England, confirming it is not specific to FOMC communication.
We then investigate what information reversal statements convey. Decomposing yields into real rates and inflation compensation shows that the reversal effect operates entirely through real interest rates, ruling out news about the inflation target or the central bank's reaction function. Decomposing further into expectations hypothesis and term premium components reveals that the effect operates primarily through revisions in expected future short rates, with a secondary contribution from the term premium. Forward rate evidence confirms that reversal statements generate a sharp revision in policy expectations at me²dium-term horizons that declines and becomes insignificant at the 10-year horizon, a rotation in the expected rate path rather than a parallel shift, consistent with investors learning about the pace of future rate adjustments. Option-implied uncertainty measures do not respond differentially on reversal days, indicating that the term premium effect reflects duration risk associated with the change in the shape of the rate path, not an increase in uncertainty about its level.
These findings have direct implications for the identification of monetary policy effects. Standard event-study analyses pool together two qualitatively different types of statements, introducing an aggregation bias. The information that matters for long-term rates lies in the joint distribution of present and future policy signals, not their individual levels.
Keywords: Monetary Policy Surprises, Policy Expectations, Yield Curve, Identification, Policy Signals
Codes JEL : E43, E52, E58, G12
Download the full publication
Updated on the 12th of June 2026