Skip to contents

The "pure" monetary policy shock from Jarocinski and Karadi (2020), identified via sign restrictions on the joint response of short-term interest rates and stock prices around FOMC announcements. The median decomposition allows the MP and information shocks to co-occur. Monthly US series from the authors' maintained update.

Usage

jarocinski_karadi_mp

Format

A data frame with columns:

date

Date. First day of the observation month.

shock

numeric. Pure MP shock (MP_median), percentage points.

series

character. Series identifier "jarocinski_karadi_mp".

Source

Jarocinski, M., & Karadi, P. (2020). "Deconstructing Monetary Policy Surprises: The Role of Information Shocks." American Economic Journal: Macroeconomics 12(2): 1-43. doi:10.1257/mac.20180090 . Updated data: https://github.com/marekjarocinski/jkshocks_update_fed_202401.

Details

Identification. Two high-frequency surprises enter: the 3-month fed-funds futures and the S&P 500, both in 30-minute windows around FOMC announcements. A 2-shock SVAR is identified by sign restrictions: a positive MP shock raises rates and lowers stocks (negative co-movement); a positive CB-information shock raises both (positive co-movement).

Median vs poor-man's decomposition. The "poor-man's" variant sorts events by sign pattern and assigns each surprise wholly to one shock. The median variant solves the set-identified problem and picks the rotation whose impulse responses lie at the median of admissible rotations; both shocks can co-occur at every event. mpshock uses the median version (MP_median, CBI_median). Results are set-identified, not point-identified: users should report robustness across rotations.

Critical follow-up. Acosta (2023, "Perceived Causes of Monetary Policy Surprises") argues that the rate-stock sign pattern is a weak discriminator between policy and information shocks because the two are typically negatively correlated at high frequency regardless of shock type. Bauer and Swanson (2023) argue the information shock reflects omitted pre-announcement data rather than genuine Fed private information. Users estimating information-effect IRFs should report robustness to these critiques.