Stock returns over the FOMC cycle

by on August 28, 2014 at 2:31 am in Economics, Uncategorized | Permalink

I find this paper (pdf), by Anna Cieslak, Adair Morse, and Annette Vissing-Jorgensen, quite scary.  Do you?:

We document that since 1994 the US equity premium follows an alternating weekly pattern measured in FOMC cycle time, i.e. in time since the last Federal Open Market Committee meeting. The equity premium is earned entirely in weeks 0, 2, 4 and 6 in FOMC cycle time (with week 0 starting the day before a scheduled FOMC announcement day). We show that this pattern is likely to reflect a risk premium for news (about monetary policy or the macro economy) coming from the Federal Reserve: (1) The FOMC calendar is quite irregular and changes across sub-periods over which our finding is robust.  (2) Even weeks in FOMC cycle time do not line up with other macro releases. (3) Volatility in the fed funds futures market and the federal funds market (but not to the same extent in other markets) peaks during even weeks in FOMC cycle time. (4) Information processing/decision making within the Fed tends to happen bi-weekly in FOMC cycle time: Before 1994, when changes to the Fed funds target in between meetings were common, they disproportionately took place during even weeks in FOMC cycle time. In addition, after 2001 Board of Governors discount rate meetings (at which the board aggregates policy requests from regional federal reserve banks and receives staff briefings) tend to take place bi-weekly in FOMC cycle time. As for how the information gets from the Federal Reserve to the market, we rule out the Federal Reserve signaling policy via open market operations post-1994. Furthermore, the high return weeks do not systematically line up with official information  releases from the Federal Reserve or with the frequency of speeches by Fed officials.  We end with a discussion of quiet policy communications and unofficial information flows.

Say it ain’t so!

The pointer is from BH, a loyal MR commentator.

Habib August 28, 2014 at 2:43 am

WOW *speechless*

JVM August 28, 2014 at 3:06 am

Implications:

a) Scott Sumner is right that Fed policy is extremely important to the functioning of the macro economy

and

b) FOMC discretion may be sub-optimal: there *is* some “correct” policy that markets would greatly prefer. Scott thinks it’s NGDP targeting and presents some persuasive evidence, but of course if the Fed were to commit to the wrong policy that could be disastrous in itself, so maybe discretion is the best we can hope for.

One interesting perspective is to view FOMC discretion as the last vestige of central planning in our economy. We’ve banished it in every other area, but it remains very difficult to move off of it for money, despite the great importance of handling it well. Free banking provides one potential alternative there, I’d be interested to hear people’s ideas on whether it’s really feasible to set it up in a way that would be preferable to the current system.

BC August 28, 2014 at 7:31 am

I used to be skeptical of Sumner’s claim that the stock market was a good indicator of the stance of monetary policy. In my mind, TIPS spreads are obviously tied to market expectations of near-term inflation as well as how the Fed’s responses to future events will affect future inflation (for longer-dated TIPS). The stock market, however, seemed to reflect lots of factors, not just NGDP expectations. I am less skeptical of Sumner’s claim now.

“…if the Fed were to commit to the wrong policy that could be disastrous in itself, so maybe discretion is the best we can hope for.”
But, if Fed communications policy is effective, then the FOMC calendar shouldn’t be effecting the markets in this way. The market should already be anticipating the Fed’s actions. For example, long-term TIPS spreads seem to move less than short-term spreads. That indicates that the Fed is effective in communicating how it will offset future events’ impact on inflation.

“We’ve banished [central planning] in every other area…”
What is your definition of central planning? If the central government plans out what your health insurance covers and how much you pay for it, is that central planning?

In honor of the start of college football, when the Fed determines the money supply, three things can happen and two of them are bad [http://quoteinvestigator.com/2011/10/04/pass-3-things/]: (1) the Fed can be tighter than market demand and cause a recession, (2) the Fed can be looser than the market demands and cause inflation, or (3) the Fed can meet market demand. This paper seems to demonstrate a market-perceived risk that (3) won’t happen.

Cal August 28, 2014 at 3:54 am

This reinforces that 2012 paper by Lucca and Moench which the authors cite here as a starting point – and that fact in itself is pretty scary since surely the effect should have disappeared in the two years since we all learnt about it (EMH etc).

So my conclusions from this, unsatisfyingly, are:
1) Everyone actually just loves buying stocks on a creepily regular cycle after being reminded the FOMC exists; or
2) Whole lotta insider trading; or
3) The market moves in roughly fortnightly cycles anyway, and the original ‘day 0′ FOMC bump is the real culprit.

China Cat August 28, 2014 at 12:21 pm

+1

Anon. August 28, 2014 at 12:32 pm

The 3rd possibility is addressed: the FOMC schedule is irregular, so that’s not a possibility.

Doug August 28, 2014 at 12:38 pm

It doesn’t have to *imply* any of those things. Check out Figures 12 and 13 of the research paper. The alternating cycle isn’t a response of the equity market to the original FOMC meeting. It’s because Federal Reserve protocol typically only works on alternating weeks in the cycle. The sizable bulk of discount rate meetings and inter-meeting target changes occur on the alternating weeks.

Also while it may seem that the paper implies that the Fed totally drives the market, that’s not necessarily the case. The paper found a difference of 0.27% weekly returns between the two respective parts of the cycle. Weekly equity market volatility is around 2.7% over the period. Fed lifecycle timing only explains 1% of the variance of weekly equity returns. That’s a large enough magnitude to wipe out the equity premium, but that’s only because the premium is so tiny relative to the volatility. There are many other things we know that wipe out the equity premium. Excluding overnight returns, only holding stocks from close to open, gives you zero premium. Excluding earnings announcements (which happen per stock even less than FOMC) wipes out the equity premium.

Yet none of these factors, are “silver bullets” in the sense that they completely explain the stock market.

rayward August 28, 2014 at 8:22 am

With all those technical stock traders, index funds, and program trading, not to mention those complex algorithms that the quants develop, and so few fundamental traders, it’s not surprising that stocks move like an enormous ship in the sea, sort of like the Titanic.

Jason August 28, 2014 at 8:28 am

The paper was presented at the NBER Summer Institute, and several people in the audience were regular participants at such discount rate meetings. They agreed that there was no plausible way these obscure meetings could have such an effect on markets. The authors could not come up with any other explanation, either.

Frankly, the results scream “data mining” inspired by the Lucca and Moensch paper. Here’s an experiment for you: generate some random numbers, smooth them using a 5-day moving average, and see how easy it is to get “cycles.”

Please don’t enable people who run a million regressions, find some correlation, and trumpet its “significance” from the rooftops.

Alexei Sadeski August 28, 2014 at 8:30 am

Perhaps all of those senators and congresspeople taking full legal advantage of their knowledge.

Scott Sumner August 28, 2014 at 8:35 am

BC, Just to be clear, I do NOT think the stock market is a good indicator of the stance of monetary policy. It’s an indicator, but not at all reliable. It’s most useful at the point in time when the Fed releases important news, and when AD is clearly suboptimal. In the 1970s (when AD was too high) it’s a more ambiguous indicator.

Bill August 28, 2014 at 8:37 am

That there can be a premia, and that the premia is disclosed after the announcement, actually makes me more confident that the FED is holding its cards close to the chest, and not disclosing or leaking information to favored entities for them to take advantage of it.

Or, to put it a different way, there must exist a premia for inside information as to the Fed’s course of conduct.

If there wasn’t a premia, that might mean someone is leaking to friends.

Steve Sailer August 28, 2014 at 9:57 am

I was taught the Efficient Markets Hypothesis at MBA school in 1980-82 and the notion that you can’t beat the stock market in the long run without inside information seemed pretty reasonable to me because in the early 1980s, there weren’t many hedge fund billionaires. So, the stock market seemed pretty efficient and honest.

Now, however, there are of lot of stock market billionaires. What does that imply?

Laura August 28, 2014 at 10:02 am

Thirty years of inflation.

Alexei Sadeski August 28, 2014 at 10:19 am

Lots of stock market billionaires implies lots more of people who went broke playing the market.

John Smith August 28, 2014 at 10:52 am

This is in-line with Steve’s insinuation.

Steve Sailer August 28, 2014 at 11:27 am

“Where are the customers’ yachts?”

Brian August 28, 2014 at 11:41 am

No, the implication is that the more people who get into the market, even if it’s entirely random who wins and loses (strong EMH) the more people you’ll see on long winning streaks who will believe (and make others believe) that they have some special talent for winning (and can extract large fees and salaries for their management services). What you don’t see is all the others on long losing streaks because they’re all out of the market. So seeing a proliferation of winners when we’ve seen a proliferation of players isn’t necessarily a sign of anything.

byomtov August 28, 2014 at 1:17 pm

Yes.

Set enough people to coin-tosssing and some will get 10 or 15 heads in a row.

Brian Donohue August 28, 2014 at 10:32 am

A little knowledge is a dangerous thing?

John Smith August 28, 2014 at 10:53 am

Or a little more knowledge is a valuable thing!

Medici1 August 28, 2014 at 10:52 am

Wall Street maven Leon Cooperman once told a roomful of MBA students that he made his money in the inefficient parts of the market, and gave a nod to the professors ‘of one of the homes of the Efficient Market Theory’ in the room. Since then, the creativity on the Street has cranked up to find more ways to seek rents or otherwise exploit inefficiencies. Combine that with regulatory capture and the Greenspan Put for unique opportunities to enrich oneself.

john August 28, 2014 at 10:58 am

The EMH was a bundle of two claims, only now being split. One is market unpredictability (yes) and the other was a proposed mechanism, efficiency (no). Whether or not this paper on information flows is directly disproving efficiency, we know by many other means that efficiency is missing. The market “turns its attention” in a way that is both unpredictable and inefficient. When will people stop loving Facebook? Unknowable and inefficient.

Doug August 28, 2014 at 12:57 pm

In 1980 it was *much* easier to make money trading, even very simple systems like pairs trading, trend following, and put-call parity arbitrage were making people double-digit returns with no few to any down months. In contrast the market today is ruthlessly efficient. It is incredibly hard to consistently generate excess return, and even many of the smartest people on Earth fail to do so.

The difference is that the market is much more liquid. In 1980 the daily average volume for the S&P 500 index was 45 million shares a day. In 2013 the ADV was 3,358 million shares a day (plus the share prices are higher, so the total dollar volume is even higher). Say you have some magic system where everyday I tell you 1 stock that will be in the upper quartile of returns. You’re going to to out and try to buy as many shares as you can. How much money you can make is a function of how many shares you can buy, which is a function of how much liquidity exists in the market. Relative to 1980, today you can trade about *1,000 times* more capital.

In short today finance is much more of a winner take all system. Everybody’s fighting ruthlessly just to keep their heads above the water. But the few people that can consistently generate excess returns in this thunder dome style tournament, have access to enormous amounts of liquidity (and awesome technology to scale up) that they can use to build massive fortunes.

Ray Lopez August 28, 2014 at 2:40 pm

Insider trading, as anybody who has ever worked on Wall Street can tell you. In fact, I know of people who used to brag about inside information coming from the SEC. It may have been bravado, but it may be a smoke/fire thing. “Furthermore, the high return weeks do not systematically line up with official information releases from the Federal Reserve or with the frequency of speeches by Fed officials. We end with a discussion of quiet policy communications and unofficial information flows . ” – is it a coincidence that Hank Paulson, ex of Goldman Sachs, got bailouts for all the counterparties to GS just in time so GS was preserved, at the same time he got phone calls from his buddies at GS and the same time that Paulson predicted the world would come to an end unless bailouts were instituted? That the bailouts worked out (save Chrysler and a few others including Fannie and Freddie Mac, which have still not been reformed and could go bust again) and have largely been paid back is small comfort to the taxpayers who helped preserve these corrupt organizations. Capitalism is not a one-way bet, aided and abetted by government.

Really Curious August 28, 2014 at 10:50 pm

Oh, come on, ever hear of in-sample overfitting? next thing we know it will be weeks 1,3,5,9

Comments on this entry are closed.

Previous post:

Next post: