Skip to contents

Bauer-Swanson orthogonalised monetary policy surprise

Source: R/data.R
bauer_swanson.Rd

The monthly orthogonalised monetary policy surprise series (MPS_ORTH) from Bauer and Swanson (2023), with the raw MPS alongside. The orthogonalised series removes predictability from public economic information available before each FOMC meeting, isolating a genuinely exogenous monetary-policy innovation.

Usage

bauer_swanson

Format

A data frame with columns:

date

Date. First day of the observation month.

shock

numeric. Orthogonalised MPS, percentage points.

mps_raw

numeric. Raw high-frequency MPS summed within the month, percentage points.

series

character. Series identifier "bauer_swanson".

Source

Bauer, M. D., & Swanson, E. T. (2023). "A Reassessment of Monetary Policy Surprises and High-Frequency Identification." NBER Macroeconomics Annual 37: 87-155. doi:10.1086/723574 . Data: https://www.frbsf.org/research-and-insights/data-and-indicators/monetary-policy-surprises/.

Details

Orthogonalisation. MPS_ORTH is the OLS residual from regressing the raw MPS on six pre-announcement predictors: surprise in the most recent nonfarm payrolls release, trailing 12-month employment growth, log S&P 500 change over the prior three months, change in the 10-year minus 2-year Treasury slope over the same window, log commodity-price- index change, and Bauer-Chernov option-implied 10-year Treasury skewness. See Bauer and Swanson (2023) Appendix Table A.1 for exact predictor definitions and data sources.

Relation to nakamura_steinsson. Both series use a first principal component of tight-window futures surprises. MPS_ORTH additionally removes predictability from public data, which Bauer- Swanson argue isolates the policy shock from the "Fed response to news" that NS attribute to a Fed information effect.

Critique. Hoesch, Rossi, and Sekhposyan (2023) show alternative orthogonalisation choices yield different residuals. If the Fed has any informational advantage, the orthogonalisation throws it away by construction.

Vintage. Bundled from the FRBSF-maintained update covering through December 2023. For a frozen vintage, download directly from the source URL in mp_source().