Quant Memo
Core

The Pre-FOMC Announcement Drift

US equities have historically drifted upward in the hours before scheduled Federal Reserve rate announcements, a pattern that shows up regardless of what the Fed actually decides — one of the more studied and hardest-to-explain anomalies in equity returns.

Prerequisites: Economic Data Releases and Market Reaction

The Federal Open Market Committee announces its interest-rate decision at a scheduled time, eight times a year. You would expect the S&P 500 to move only once the decision is actually known — up if the Fed is more dovish than expected, down if more hawkish. Instead, researchers documenting decades of data found the index has tended to drift upward in the roughly 24 hours before the announcement even lands, regardless of which way the decision eventually goes. This is the pre-FOMC announcement drift, and it remained a mostly unexplained regularity for years after it was first documented.

US equities have historically shown a positive average return in the day or so leading up to a scheduled FOMC announcement, independent of the announcement's actual content — a pattern hard to square with the idea that prices only move on new information, since the drift happens before any information is released.

What the pattern looks like

Lucca and Moench's original study found that a large share of the equity risk premium earned over the entire sample period was concentrated in the 24 hours before scheduled FOMC meetings — a tiny fraction of total trading days accounting for a disproportionate share of long-run equity returns. The effect showed up whether the eventual decision was a hike, a cut, or no change, which rules out the simple explanation that markets were merely front-running a known, predictable outcome.

hours relative to announcement → announcement (2pm ET) pre-announcement drift
The bulk of the historical drift accrues in the roughly 24 hours before the announcement, not after it.

Worked example

Across a sample of FOMC meetings, suppose the average S&P 500 return in the 24 hours before the 2pm announcement is +0.3%, compared with an average return of roughly +0.03% on a randomly chosen 24-hour window elsewhere in the sample — about ten times the typical daily drift, concentrated into a small number of scheduled dates each year. With 8 meetings a year, that's roughly 8 days out of 252 trading days (about 3% of the calendar) capturing a return contribution that, compounded, has historically rivaled a meaningful fraction of the market's total annual excess return over cash.

What this means in practice

Because the drift is a statistical pattern discovered after the fact in decades of data, systematic traders who try to harvest it face two real risks: the position is a leveraged bet on a handful of calendar days rather than a diversified strategy, and like many documented anomalies, its size has reportedly shrunk somewhat since publication — consistent with either the effect being partly arbitraged away once well known, or the original finding being partly a product of the specific sample studied.

A pattern discovered by scanning historical returns for what happened around known calendar dates is especially exposed to data-mining risk — with only a handful of FOMC meetings per year, decades of data still means a fairly small number of independent observations, so out-of-sample decay is the expected outcome, not a surprise.

Related concepts

Practice in interviews

Further reading

  • Lucca and Moench, 'The Pre-FOMC Announcement Drift', Journal of Finance
ShareTwitterLinkedIn