Quantum BiasFOMC Week Trading: Does the Fed Cycle Still Move the Nasdaq?
In this article
No. A well-known 2019 study found that U.S. stocks earned their whole equity premium in the even weeks after each Fed meeting. On the Nasdaq-100 from January 2017 to October 2026, even weeks actually returned less than odd weeks, and the gap is no different from luck.
| Question | Data | Answer |
|---|---|---|
| Are even weeks of the FOMC cycle stronger for the Nasdaq-100? | QQQ weekly total returns, Jan 2017 – Oct 2026, 78 scheduled Fed meetings | No: +0.31% vs +0.53% a week, t = −0.9 |
The study we tested
Anna Cieslak, Adair Morse and Annette Vissing-Jorgensen published Stock Returns over the FOMC Cycle in the Journal of Finance (vol. 74, no. 5, 2019). Using 184 scheduled Fed meetings from 1994 to 2016, they counted weeks from each meeting: week 0 around the announcement, then weeks 1 to 6. Even weeks (0, 2, 4, 6) averaged 0.33% to 0.60% of excess return. Odd weeks were slightly negative. Their explanation: news about Fed policy hits the market on a regular, every-other-week rhythm.
Our question is simple. Their data stops in 2016. Did the pattern keep going on the index NQ traders follow?
What we found
Exact values from table 2 in this article. Bar lengths share a zero baseline; no new calculation or simulated price path. Read the table and its stated sample and limits in this article.
- Even weeks: 267 weeks, average +0.31%.
- Odd weeks: 232 weeks, average +0.53%.
- Difference: −0.23 points a week, t-statistic −0.92. Not significant, and pointing the wrong way.
- Share of the total return earned in even weeks: 38%, not 100%.
Could it be luck either way?
We shifted the even/odd labels by a random number of weeks, 10,000 times, keeping the returns in order. In 80% of those shuffles, "even" weeks beat "odd" weeks by at least as much as the real even weeks did. Put simply, the real FOMC calendar doesn't do anything a random calendar wouldn't.
What a trading rule would have earned
A rule that holds QQQ only in even weeks and Treasury bills in odd weeks switches about 48 times a year.
| Annual return | Worst drawdown | |
|---|---|---|
| Buy and hold QQQ (all weeks) | 21.5% | −35.1% |
| Even weeks only, before costs | 9.0% | −23.1% |
| Even weeks only, 0.02% cost per switch | 8.0% | — |
| Even weeks only, 0.05% cost per switch | 6.4% | — |
The rule cut drawdowns because it was out of the market half the time. It also gave up more than half the return.
What this means if you trade NQ
The Fed still matters on decision day: the New York Fed's own update found a pre-announcement drift that held through 2018 at press-conference meetings. What we couldn't find is a weekly rhythm between meetings. On this evidence, where a week sits in the Fed calendar isn't a reason to lean bullish or bearish.
Limits
- Calendar weeks, not trading days. The study counts trading days from the announcement (week 0 = days −1 to 3). We use Friday-to-Friday weeks, which overlap the study's windows by about four days out of five. Daily data would make the test exact.
- QQQ, not the NQ contract. Same index, but QQQ includes dividends and only trades during the cash session.
- Raw returns, not excess returns. The Treasury bill rate is the same for even and odd weeks, so it doesn't change the difference.
- Long gaps left out. We left out 11 weeks that fell more than six weeks after a meeting, mostly in spring 2020, when emergency actions replaced the March meeting.
- One period. Nine years is long for a weekly test, but it's still one market regime.
Method and sources
- QQQ weekly prices: Alpha Vantage
TIME_SERIES_WEEKLY, adjusted by us for the March 20, 2000 two-for-one split; dividends added in the week of their ex-date (Alpha VantageDIVIDENDS). - Fed meeting dates: Federal Reserve FOMC calendars, scheduled meetings only.
- Treasury bills: 3-month yield, Alpha Vantage
TREASURY_YIELD. - Original study: Cieslak, Morse and Vissing-Jorgensen, Stock Returns over the FOMC Cycle, Journal of Finance, 2019.
Research and education only. Not financial advice. Past results do not predict future returns. A backtest is not real performance.
The week just closed, summarized, and the economic calendar for the week ahead.