This paper from David O. Lucca and Emanuel Moench at the New York Fed, concluding that 80% of excess returns to U.S. equities come in the 24 hours before Fed monetary policy announcements, is pretty amazing. Here is the money chart; what does this tell you about the effect of the Fed's actions on stock prices?
I guess one answer is:
(1) Fed actions push stocks up.
But I submit to you that this answer, by itself, is self-evidently wrong, since the stocks go up before the Fed actions. Two better possibilities are:
(2) The Fed's actions travel back through time to push stocks up, or
(3) The Fed's actions are irrelevant to stock prices, but the warm fuzzy feeling people have when they remember that the Fed exists and takes actions pushes stocks up.
Weirdly severalpeople choose self-evidently incorrect answer (1), while the Fed researchers themselves suggest a bunch of other interpretations like "volatility feedback loops" and illiquidity effects and political risk, though they're not thrilled with any of them.
If the Fed propagates backwards through time, one possibility is I guess insider trading: the Fed's actions generally tend to push up stock prices, and enough people find out about them in advance to do the pushing before the policies are announced. This is a pretty weak answer, though, in part because it's only equities that have abnormal returns before announcement - in particular, the authors find no unusual results on FOMC days for interest rate products, which is what you'd trade if you actually knew secret things about the Fed's plans. (Though they do find Fed effects on foreign stock markets, small world etc.)
Which leaves answer (3): people read a lot about the Fed the day before the Fed announcement, think fondly "oh, yeah, the Fed, I like those guys, good guys," and buy stocks. The positive drift is in part due to the Fed being mostly accomodative in the last 17 years,* and in (likely bigger) part due to the asymmetric effect of the "Bernanke [Greenspan, whatever] put" - as the authors put it, "interest rate policy has sometimes been characterized as having an asymmetric impact on riskier asset values through implicit floors, or so called 'government puts'" - so if you think that the Fed will ease you buy stocks and if you think it will tighten you don't necessarily sell them.**
One model of how that works - recently, anyway - is:
1. Terrible thing happens
2. Stocks go down
3. Goto 1
5. It's the day before the FOMC day
6. "God, with all these terrible things happening, the Fed just has to ease."
7. Stocks go up
Lucca and Moench refer to this, or something like it, as "rational inattention," and are pretty un-thrilled by it: it's fine for retail investors to think like that but seems sort of dumb for everyone to do so. In theory each time a terrible thing happens you should react not only with terror but also with a small compensating expectation that the Fed will ease, and you mostly do,*** so there's no reason to do all your buying the day before the Fed announcement.
One thing you might ask about any economics paper that identifies a colossal and tradeable market inefficiency is: why on earth would you publish this? Here is what I would do if I worked at the New York Fed and ran this model and found these results:
2. Buy stocks 25 hours before FOMC announcements, hold for 24 hours, and sell****
3. Be rich
But they didn't, which is public-spirited of them.***** I guess it's safe to assume that tombstone for "the pre-FOMC announcement drift" will read "1994-2011": if, going forward, you can make a free 3.89% excess return by buying on pre-Fed days, you will, so you can't. And I guess it's also safe to assume that the Fed won't be too broken up about that: while Ben Bernanke probably is happy to support equity prices, he's probably not so thrilled to read headlines about how he's supporting equity prices. This paper's detailing of how much - and how strangely - equity prices are tied to Fed actions is, on that theory, an embarrassment, but it's a one-time embarrassment. And if it eliminates the pre-FOMC-day effect from now on, and replaces it with smaller but more consistent support for prices throughout the year, that's a trade the Fed should be happy to take.
* Incidentally it starts in 1994 because that's when the FOMC started announcing its decisions on set dates.
** A possibly related suggestive nugget:
The point estimates of 62 basis points for the recession dummy and of 33 basis points for the constant imply that pre-FOMC returns have been about 29 basis points higher in recessions than in expansions, indicating that the pre-FOMC announcement returns are somewhat counter-cyclical. However, this difference is only statistically different from zero at the 10% level.
Similar/related effects for "market expects tightening" vs. "market expects easing," all consistent with "if you think that the Fed will ease you buy stocks and if you think it will tighten you don't necessarily sell them."
*** Pretend that here I'm linking to all the times someone has said "stocks rally because terrible [unemployment /ISM/whatever] number means that the Fed is more likely to announce QE17" or whatever.
**** More precisely (from the paper):
The simple strategy that consists in buying the SPX at 2pm the day before a scheduled FOMC announcement and selling fifteen minutes before the announcement while holding cash on all other days would have earned a large annualized Sharpe ratio of 1.14 as reported in the Table.
***** My imaginary drama of Lucca and Moench is that they can't stand each other and William Dudley teamed them up on to work on this paper knowing that if it found anything worthwhile and one quit to go trade on it, the other would publish. Actually maybe the paper originally had three authors and one quit to launch a hedge fund that is now going to have a lot more trouble raising money. All of this is surely false OR IS IT?