A few weeks ago I read this report by Federal Reserve economists (oxymoron?). The authors summarize it as follows:
We show that since 1994, more than 80 percent of the equity premium on U.S. stocks has been earned over the twenty-four hours preceding scheduled Federal Open Market Committee (FOMC) announcements (which occur only eight times a year)
The pertinent point of this research can be shown in the graph showing S&P 500 returns with and without capital gains from returns surrounding the Fed rate announcement.
Source: The Puzzling Pre-FOMC Announcement "Drift" by David Lucca & Emanuel Moench
Without the capital gains returns from the time surrounding the FOMC meeting, over the last 18 years, the S&P 500 would have returned 650 points less. The authors also provide regression results, adjusted for dividends and "risk free" returns, which show statistical significance at the 1% level. The data was from 1994 - present vs. 1973-1993. What changed in 1994 was that the Fed began to announce policy changes on the day of their meeting and not months afterward.
Here is a link for past FOMC meeting dates from 2012-2007, and also includes future FOMC meetings in 2012 & 2013. Returns from FOMC meeting dates were calculated for the S&P 500 over this time frame for each meeting, both scheduled and unscheduled.
On January 3, 2007 the S&P 500 was at 1416.60 and on July 13, 2012 it was at 1356.78, a capital loss of about of about -4.3% that does not include dividends. The average daily return over this period was -.0031%.
The time period where the FOMC met, both scheduled and unscheduled, returned just over 25.5%. The average trade was .425%. If there was a two-day meeting only the return on the day of the announcement was reported. All returns are buying the day before the announcement, on the close, and selling the day of the announcement, on the close. The stats for all trades are shown below:
Now, a reader might point out that one cannot know about "unannounced" FOMC meetings, although one might hear rumors beforehand, so I divided up FOMC meetings between scheduled and unscheduled and the results were surprising.
The results are actually better when only scheduled FOMC meetings are included. The average trade increases to over .7% and the total return increases to 31%. Here are the stats for the unscheduled FOMC meetings:
Since 2007 a trader could buy the S&P 500 on the close, before a known FOMC announcement, and receive an average return that is significantly greater than zero. A trader can use the SPY ETF, futures, or options on either, to position themselves to take advantage of this FOMC bias. If one uses options it would most likely be best to use spread strategies, since implied volatility usually falls after an FOMC announcement.
The article also provides information about international stocks:
While we do not find similar responses of major international stock indexes ahead of their respective central bank monetary policy announcements, we observe that several indexes do display a pre-FOMC announcement drift, as the chart below shows. Cumulative returns rise for the British FTSE 100, German DAX, French CAC 40, Swiss SMI, Spanish IBEX, and Canadian TSE index when each exchange is open for trading over windows of time around each FOMC announcement in our sample.
The authors refer to the up move bias in stock indexes as "drift" and conclude, "the drift remains a puzzle."
There should not be much of a puzzle. Investor psychology and expectations have responded to an easy money policy by the Fed over the last 2 decades. Even international investors and traders know who has been calling the shots given the central bank's power to set benchmark rates and bailout whomever they see fit. Amazingly, the conspiracy theorists might have been handed the empirical "smoking gun" by the very institution they criticize.
I did preliminary research into the returns of other asset classes. Below are the results for gold, using the ETF GLD. On average the return was positive, however, it failed to achieve statistical significance.
I also looked into the returns of long term US Bonds, by researching the returns of the ETF TLT. There was a very slight drop on average, but with no statistical significance.
Finally I looked at the returns of euro currency using the ETF FXE. Here, I found a "close" enough positive return which was statistically significant (p-value of .051). Fortunately for traders the bulk of these returns, where "significance" was found, were in the scheduled FOMC meetings.
I then turned my research to different sectors and segments of the stock market that can be traded via ETFs, and just for good measure threw in Apple's stock performance (AAPL). Below are the results for only the known meetings.
The financial sector, XLF, had a higher average return of 1.28% over the time period but with more volatility. The emerging markets ETF, EEM, also provided a higher return at .94% an average trade. Apple had an average trade of 1.13%. The consumer staples sector, XLP, had less than half of the average SPY trade at .33%.
The biggest surprise came with the results trading the Nasdaq 100 ETF, QQQ, since it returned .81% on average, higher than SPY, with a slightly lower standard deviation. Statistical tests did not show this difference to be significant and probably more data is needed. That said, most traders would trade an instrument that has a higher average trade historically. The 95% confidence interval of the trades in the QQQ is higher than SPY, with the skewness being more positive too.
This technique could easily be overlaid atop a portfolio. In just the last year, from June 2011 - June 2012, the SPY has returned 5.45% before transaction costs.
Disclosure: I have no positions in SPY, GLD, TLT, AAPL, EEM, QQQ, XLF, XLP but may initiate a Long or Short or Options position in any of these over the next 72 hours. I have an Options position in FXE.