Stock Returns over the FOMC Cycle

• Author: ADAIR MORSE,ANNETTE VISSING‐JORGENSEN,ANNA CIESLAK
• DOI: http://doi.org/10.1111/jofi.12818
• Published date: 01 October 2019
• Date: 01 October 2019

THE JOURNAL OF FINANCE •VOL. LXXIV, NO. 5 •OCTOBER 2019

Stock Returns over the FOMC Cycle

ANNA CIESLAK, ADAIR MORSE, and ANNETTE VISSING-JORGENSEN∗

ABSTRACT

We document that since 1994, the equity premium is earned entirely in weeks 0, 2,

4, and 6 in Federal Open Market Committee (FOMC) cycle time, that is, even weeks

starting from the last FOMC meeting. We causally tie this fact to the Fed by study-

ing intermeeting target changes, Fed funds futures, and internal Board of Governors

meetings. The Fed has affected the stock market via unexpectedly accommodating

policy, leading to large reductions in the equity premium. Evidence suggests system-

atic informal communication of Fed officials with the media and financial sector as a

channel through which news about monetary policy has reached the market.

DOES THE FEDERAL RESERVE HAVE a substantial impact on the stock market?

How much of realized stock returns can be attributed to the Fed, if any? And

what is the economic mechanism behind any impact? Prior work on these

questions focuses on Federal Open Market Committee (FOMC) announcement

days (starting with Bernanke and Kuttner (2005)) or the 24-hour period just

before scheduled FOMC announcements (Lucca and Moench (2015)). However,

monetary policy news may not arrive only on FOMC announcement dates. If

so, the overall impact of the Fed on the stock market could be much larger

than previously thought. From Kuttner (2001), who uses federal funds futures

data to decompose target rate changes on FOMC announcement dates into

∗Anna Cieslak is with Duke University and CEPR. Adair Morse is with the University of Califor-

nia Berkeley and NBER. Annette Vissing-Jorgensen is with the University of California Berkeley

and NBER. We thank seminar and conference participants at Harvard Business School, Federal

Reserve Bank of Boston, Federal Reserve Board, NBER EASE meeting, NBER Summer Institute

Monetary Economics meeting, NBER Asset Pricing meeting, London School of Economics, Uni-

versity of Minnesota, MIT Sloan, University of Michigan, Michigan State University, Washington

University St. Louis, European Central Bank, University of California Berkeley (Haas and Eco-

nomics), Stanford University, NYU Stern, Duke University (Economics), Columbia, ASU Sonoran

conference, BIS Research Network meeting, Texas A&M, Dartmouth, Bocconi, University of Am-

sterdam, University of Washington, Goethe University Frankfurt, Frankfurt School of Finance &

Management, University of Zurich, AQR, Singapore Management University,Nanyang Technolog-

ical University, Hong Kong University, HKUST, Vanderbilt, Nationalbanken, American Finance

Association, European Finance Association, Western Finance Association, Midwest Finance As-

sociation, Society of Economic Dynamics, CEPR ESSFM in Gerzensee, BlackRock, Universidad

Cat´

olica de Chile Finance Conference, the Financial Intermediation Research Society Conference,

Central Bank Workshop on Market Microstructure, and the Red Rock Finance Conference. Wealso

thank Tim Loughran, David Romer, Christina Romer, Andrew Rose, Ken Singleton, Julio Ruitort,

Kosuke Aoki, Pavel Savor, Lars Svensson, Michela Verardo, Brian Weller, Mungo Wilson, Bryan

Kelly, David Reeb, and various current and former Federal Reserve officials for their help and

feedback. We have no conflicts of interest to disclose.

DOI: 10.1111/jofi.12818

2201

2202 The Journal of Finance R

expected and surprise components, we know that the news driving investors’

updates of monetary policy expectations do not principally come out at the time

of FOMC statements.1Furthermore, with eight scheduled FOMC meetings per

year, focusing only on FOMC announcement dates could lead to small-sample

issues.2

To overcome these challenges, in this paper we study the evolution of stock

returns over the full cycle of days between scheduled FOMC meetings. Our

main finding is that in the period from 1994 to 2016, the average excess re-

turn on stocks over Treasury bills follows a biweekly pattern over the FOMC

cycle. In particular, the equity premium over this 23-year period was earned

entirely in weeks 0, 2, 4, and 6 in FOMC cycle time, which we refer to as “even

weeks,” where week 0 of the FOMC cycle starts the day before a scheduled

FOMC announcement day. Moreover, average excess returns are statistically

significantly higher in even weeks than in odd weeks in FOMC cycle time, with

this fact holding up in the 2014 to 2016 period following the first draft of our

paper, which covered the 1994 to 2013 period. The pre-FOMC announcement

drift of Lucca and Moench (2015), resulting in average stock returns of about

0.5% over the 2 pm to 2 pm period prior to scheduled FOMC announcements, is

thus part of a broader biweekly pattern in stock returns over the FOMC cycle.

To support the view that the biweekly cycle in excess returns on stocks over

Treasury bills is causally driven by monetary policy news from the Fed, we

provide four pieces of evidence. First, changes to the federal funds (Fed funds)

target tend to take place in even weeks in FOMC cycle time. This is the case

both in the 1994 to 2016 period, which has only seven intermeeting changes and

in the period from 1982 to 1993, during which time about two-thirds of target

changes took place between scheduled FOMC meetings. Second, on average,

Fed funds futures rates decline in even weeks, consistent with unexpectedly

accommodating news arriving in even weeks. Third, stock returns are par-

ticularly high on the even-week days that also have more Fed information

production and decision making. Specifically, we document higher even-week

returns on days that follow board meetings of the Fed’s governors. We also

present narrative evidence suggesting that these board meetings (also called

“discount rate meetings”) are an important venue for the free exchange and

aggregation of monetary policy views among the chair, vice chair, and gover-

nors. Fourth, we document that about half of the high even-week stock returns

arises due to stock market mean-reversion in even weeks after significant stock

market declines. This pattern fits a “Fed put” interpretation, meaning that the

Fed reacts to low stock returns by providing (a promise of) monetary policy

accommodation, with this accommodation being unexpectedly strong in our

1The surprise component tends to be small in Kuttner’s data for 1994 to June 2008. For example,

on the 25 occasions when the FOMC increased the federal funds rate target by 25 basis points, the

market expectation on average was off by only 2 basis points from the realized change (the average

absolute value of the surprise).

2As an example of potential small-sample issues, in Table IA.1, in the Internet Appendix, which

may be found in the online version of this article, we show that neither the main result of Bernanke

and Kuttner (2005) nor that of Lucca and Moench (2015) is significant in postpublication data.

Stock Returns over the FOMC Cycle 2203

sample. In support of this interpretation, we show that the intermeeting target

changes over the 1994 to 2016 period were preceded by low stock returns. We

also show that high even-week stock returns that follow market declines predict

future reductions in the Fed funds target. Return patterns are thus consistent

with a commonly mentioned narrative of Fed decision making. We rule out the

possibility that the high even-week returns could be driven by other regular

economic events, notably reserve maintenance periods, macroeconomic news

releases or corporate earnings announcements.

Turning to how the Fed drives the stock market, we argue that even-week

changes in Fed funds futures rates are economically too small to account for

most of the high even-week stock returns. Using equity premium estimates

from Martin (2017), we instead provide evidence that even-week reductions in

the equity premium are large and can account for most of the high realized

even-week stock returns. This suggests that the Fed reduces the amount or

price of uncertainty, thereby lowering the equity premium. One way how the

Fed may have reduced uncertainty is through reducing downside risk via a

promise to act (i.e., to provide stimulus) as needed.

The channel through which the Fed reduces uncertainty helps explain why

reductions in Fed funds futures are modest relative to either the reduction in

the equity risk premium or the average even-week stock returns. In particular,

the Fed’s promise to act aggressively if needed leads markets to expect a lower

Fed funds rate in bad states of the world. However, if this promise succeeds in

shifting the probability distribution across economic states to the right (making

bad states less likely), the net effect on expected Fed funds rates could be small.

By contrast, for stocks, the two effects of reduced downside risk work in the

same direction, with both lower Fed funds rates in bad states of the world and

an improved distribution across states being good news for stocks.

This “downside-risk channel” appears particularly important for explaining

even-week mean-reversion in stocks—the Fed put. Specifically, the equity pre-

mium falls on even-week days that follow stock market declines, while Fed

funds futures rates do not fall on these days. The fact that high stock returns

on even-week days following stock market declines predict target changes de-

spite no reduction in Fed fund futures rates implies that the economy turned

out worse and monetary policy easier ex post than expected ex ante. This inter-

pretation is in line with evidence based on short-rate expectations from both

private-sector surveys and the Fed’s own forecasts in Cieslak (2018).

The Fed downside-risk channel adds a new dimension to the importance of

Fed communication suggesting that to some extent forward guidance is about

an unlimited promise to act as needed, similar to Draghi’s “whatever it takes”

speech in the ECB context. For evidence on the stock market rally following that

speech, see Krishnamurthy, Nagel, and Vissing-Jorgensen (2017). It is inter-

esting to note that while Draghi’s speech in Europe was sufficient to calm asset

markets, the ECB—just like the U.S. Fed following stock market declines—

eventually had to take additional steps to increase inflation and growth (the

ECB started quantitative easing [QE] in 2015, which eventually amounted to

over 20% of GDP).