Meeting-Day Microstructure and Reaction-Function Inference

Scheduled meetings sit at the intersection of three distinct risk premia: directional rate risk priced through SOFR (formerly Eurodollar) futures and options, volatility risk priced through swaption straddles, and reaction-function uncertainty embedded in the cross-section of forward path implieds. Lucca and Moench (2015) document the pre-FOMC announcement drift — roughly 80% of the unconditional US equity risk premium since 1994 has accrued in the 24 hours preceding scheduled FOMC announcements — a phenomenon robust across subsamples and not fully explained by standard risk-factor models. The blackout-period information asymmetry binds Committee members from speaking in the seven calendar days preceding each meeting, concentrating signal in the statement itself and the subsequent press conference.

SOFR option skew and 1m1m vs 2m1m vol term structure typically steepen into FOMC weeks, with implied moves of 6-10bp on the policy rate during transition regimes; the realized-to-implied ratio compresses on dovish surprises and expands on hawkish ones. Reading the SEP requires going beyond the median dot: the dispersion of individual projections (interquartile range), the central tendency excluding the three highest and lowest, and the year-end levels for 2026 and the long-run estimate jointly contain more information than any single point. The dot plot's signal-to-noise has degraded post-2022 as participants explicitly note conditioning on differing assumptions about r* and productivity. Cross-bank disclosure norms differ materially: the ECB publishes meeting accounts (not minutes) with a roughly four-week lag and provides no equivalent of the SEP; the BoE publishes the full MPC vote breakdown with the statement; the BoJ historically released Outlook Reports four times a year with median projections but minimal dispersion data, complicating reaction-function inference for yen-rate traders.