Content uploaded by Jonathan Benchimol
Author content
All content in this area was uploaded by Jonathan Benchimol on Jul 31, 2023
Content may be subject to copyright.
Stock Market Reactions to Monetary Policy
Surprises Under Uncertainty
Jonathan Benchimol,†Yossi Saadon†and Nimrod Segev†
October 2023
Abstract
This article investigates how uncertainty impacts the effect of monetary
policy surprises on stock returns. Using high-frequency US data, we demon-
strate that stock markets respond more aggressively to monetary policy sur-
prises during periods of high uncertainty. We also show that uncertainty
asymmetrically influences the transmission of positive and negative mone-
tary policy surprises to stock market prices. The amplifying effect of uncer-
tainty is found to be stronger for expansionary shocks than for contractionary
shocks. Our robustness analysis confirms that financial uncertainty has a sig-
nificant role in shaping the influence of monetary policy on the stock market.
Keywords: Monetary policy, Uncertainty, Stock returns, High-frequency data,
Event study.
JEL Codes: E44, E52, E58, G12, G14.
The views expressed in this paper are those of the authors and do not necessarily reflect the
views of the Bank of Israel. We thank Itamar Caspi (discussant), Elyas Elyasiani, Michael Gel-
man, Steven Laufer, and the research seminar participants at the Bank of Israel for their helpful
comments and suggestions.
†Bank of Israel, Jerusalem, Israel. Corresponding author: Nimrod Segev. Email: nim-
rod.segev@boi.org.il
1
1 Introduction
This paper investigates whether the level of financial uncertainty affects the stock
market’s reaction to monetary policy (MP) surprises. Many studies have investi-
gated the connection between MP and stock markets.1Economic theory suggests
several channels through which MP surprises affect stock prices. First, unexpected
changes in the policy rate may have a direct effect through discount rates. That
is, an increase in the policy rate reduces the present discounted value of firms’
expected earnings and dividends, therefore harming equity valuation. Second,
MP surprises may indirectly impact stock price valuation by signaling information
about the economic outlook and future path of MP. For example, Nakamura and
Steinsson (2018) document that in the US, MP surprises convey information to fi-
nancial markets about the Federal Reserve’s (Fed) internal forecast, which in turn
causes asset prices to react.
The level of uncertainty in financial markets may also influence stock prices
and their reaction to MP shocks. Theoretically, Bansal and Yaron (2004) show
that the equity risk premium2is a function of uncertainty, with a fall in the latter
triggering an immediate increase in stock prices. Lakdawala and Schaffer (2019)
present a theoretical model to examine the connection between MP and the stock
market, with a focus on the impact of shocks that reveal information about eco-
nomic activity (“Delphic” shocks). They build on the same framework as our pa-
per (Bernanke and Kuttner, 2005) by using high-frequency market data to identify
MP surprises. Their model decomposes the MP surprise measure into an exoge-
nous and a Delphic component. The exogenous component represents shocks un-
related to economic activity, while the Delphic component captures shocks that
reveal private information the Fed possesses. They construct a measure of pri-
vate information by combining market survey data with the Fed’s internal fore-
casts to show that such Delphic shocks reveal that the Fed’s private information
about underlying macro fundamentals also affects the market’s perceived uncer-
tainty about the future, thus highlighting the unique uncertainty dimension of MP
transmission to the stock market. Our approach is close to their theoretical model,
although we investigate the effects of uncertainty states rather than uncertainty
shocks of the transmission of MP surprises on stock returns.
On the empirical front, Bekaert et al. (2013) show that MP decisions have im-
portant effects on financial uncertainty. Gu et al. (2018) show that uncertainty
decreases after Federal Open Market Committee (FOMC) announcements, specif-
1See for example, Bernanke and Kuttner (2005) and Gürkaynak et al. (2005), among many oth-
ers.
2The equity risk premium refers to the excess return investors can potentially earn by investing
in the stock market over a risk-free rate. It compensates for the higher risk associated with equity
investments. The size of the premium varies based on portfolio risk and changes over time.
3
ically when they are accompanied by the release of the Summary of Economic Pro-
jections. They argue that the decrease in uncertainty following the announcement
can help explain positive post-announcement stock market returns. In a recent pa-
per, Cieslak et al. (2019) suggest a new channel through which the Fed affects stock
prices, which they term the “downside-risk channel”. They provide evidence that
Fed officials use systematic informal communication with the media and financial
sector to reduce uncertainty and downside risk via an implicit promise to act if the
financial system is under stress. The channel is in line with the “Fed put” view,3
according to which the Fed reacts to falling stock prices by providing (or promis-
ing) MP accommodation. Kroencke et al. (2021) provide evidence for a “risk shift”
following FOMC announcements, suggesting that there is an “uncertainty chan-
nel” through which MP announcements affect stock valuation by increasing or
decreasing market participants’ uncertainty.4
Several recent papers specifically focus on the impact of MP uncertainty. For
example, Kurov and Stan (2018) examine the effects of economic news and MP
uncertainty on MP expectations. They find that economic news significantly in-
fluence interest rates and the risk premium in times of high MP uncertainty. Con-
sequently, the response of stocks to economic news weakens during periods of ele-
vated MP uncertainty. Bauer et al. (2022) emphasize the crucial role of uncertainty
about future policy rates in transmitting MP to financial markets. They demon-
strate that FOMC announcements systematically resolve uncertainty, which sub-
sequently gradually ramps up again. Changes in MP uncertainty around FOMC
announcements, often driven by forward guidance, have pronounced effects on
asset prices that are distinct from conventional policy surprises.
Also related are papers that examine how MP shocks impact investors uncer-
tainty and sentiment. For example, Guo et al. (2021) focus on the pre-FOMC an-
nouncement drift, investigating the influence of investor’ sentiment and economic
policy uncertainty on stock prices before FFR announcements. In months charac-
terized by high sentiment, which may correspond to low uncertainty periods, they
find a positive drift in the S&P 500 index during the pre-FOMC window. However,
this positive drift is absent in months of low sentiment (high uncertainty). Build-
3The “Fed put view” refers to Fed MP accommodations in response to significant declines in the
stock market. This view suggests that the Fed is willing to ease MP to mitigate the negative effects
of stock market slumps on the economy. This policy approach refers to the Fed’s willingness to
provide a safety net for the stock market.
4The “uncertainty channel” suggests that the level of uncertainty in financial markets signif-
icantly influences how the stock market reacts to unexpected changes in MP. Supported by eco-
nomic theory and empirical evidence, reductions in financial uncertainty following MP announce-
ments can explain positive stock market returns. Central banks also use communication strategies
to reduce uncertainty and downside risk, reinforcing the “uncertainty channel.” The level of finan-
cial uncertainty plays a critical role in determining the stock market’s reaction to MP decisions.
Specifically, accommodative MP surprises can boost the stock market during periods of high un-
certainty, while restrictive MP surprises may not have the same impact (For more details on these
asymmetric effects, see Section 4).
4
ing upon their previous work, Guo et al. (2022) explore the interaction between
investor sentiment and MP news. They find that the stock market reacts strongly
to MP surprises only during sentiment-waning phases, which are distinct from
recessionary or bear market episodes.
Overall, these papers indicate that the impact of MP on financial uncertainty
and the equity risk premium might be a crucial channel through which MP affects
stock prices.5Our paper contributes to this literature by unveiling this relation-
ship’s state-dependent nature and nonlinear dynamics. In particular, we reveal
that heightened uncertainty magnifies the impact of MP shocks solely in the case
of negative shocks and under non-extreme circumstances. These findings suggest
that certain channels elucidated in the literature and empirical observations may
only hold relevance under specific conditions, highlighting the need for a nuanced
understanding of this complex interaction.
Specifically, we examine if the connection between MP and stock returns de-
pends on the direction of the policy decision–accommodative vs. restrictive. We
expect that financial uncertainty plays a different role in accommodative and re-
strictive policy surprises for several reasons. First, according to the studies on
the “fed put view,” MP may have an asymmetric impact on stock markets since
markets expect an accommodative policy stance during bad times but do not ex-
pect tightening during good times (Mishkin, 2017). An unexpected loosening of
monetary conditions during periods of high uncertainty may signal to the mar-
kets that the Fed is committed to acting aggressively to ease financial conditions.
However, an unexpected tightening may not have the same impact since markets
may interpret the shock as a positive signal about economic health and not as the
Fed attempting to cool the stock market. Second, a reduction in uncertainty fol-
lowing the MP announcement may positively affect stock prices, regardless of the
direction of the policy decision. Therefore, the post-announcement reduction in
uncertainty will strengthen the direct impact of accommodative MP decisions and
weaken the direct effects of contractionary ones. Consequently, we expect that if
the amplifying effect of financial uncertainty exists, it will be stronger for accom-
modative MP surprises than for contractionary shocks.
To test these predictions, we follow previous literature6and use an event study
methodology with high-frequency US data to examine how uncertainty affects the
reaction of stock market returns to MP decisions. This methodology allows us to
identify the immediate effect of an unanticipated MP decision, while mitigating
any simultaneity concerns between stock markets and policy changes. We find
evidence that the impact of MP on stock markets depends on the level of uncer-
5Unanticipated changes in the MP rate may exacerbate (alleviate) financial uncertainty, decreas-
ing (increasing) equity risk premiums, and causing contemporaneous stock prices to decline (rise).
6From Bernanke and Kuttner (2005) and Gürkaynak et al. (2005), to Tsai (2014), Lucca and
Moench (2015), and Cieslak et al. (2019), among many others.
5
tainty. Specifically, we show that during periods of high uncertainty, MP surprises
induce a stronger reaction by stock markets.
On the one hand, we find that accommodative MP surprises can boost the stock
market during periods of high uncertainty but is mostly ineffective when uncer-
tainty is low. On the other hand, restrictive MP is ineffective when uncertainty is
high but has a significant negative relationship with the stock market when un-
certainty is low. Overall, the results suggest that reducing the level and price of
uncertainty is an important mechanism through which MP affects stock markets.
This mechanism is critical to investors in periods of high financial uncertainty. To
strengthen the interpretation of the results, we also investigate possible asymmet-
rical MP influences on financial uncertainty. Looking at the change in uncertainty
days around an FOMC announcement, we show that uncertainty declines follow-
ing a restrictive MP announcement. The decline in uncertainty is stronger and
more significant during periods of mid-high levels of financial uncertainty, sug-
gesting that the uncertainty-based connection (channel) between MP and stock
prices is mostly relevant during periods of high uncertainty. We also find that the
response of stock returns is asymmetric to positive and negative policy changes.
This paper makes several contributions to the literature. First, it advances the
literature on the state dependency of the stock market response to MP surprises.
Previous studies have shown that MP surprises tend to have a stronger impact on
stock markets during periods of extreme market conditions such as recessions or
bear markets (Basistha and Kurov, 2008; Jansen and Tsai, 2010; Kurov, 2010; Kon-
tonikas et al., 2013). In line with these studies, we show that periods with elevated
levels of uncertainty are characterized by a stronger reaction of the stock market
to MP surprises, specifically accommodative surprises. Second, we show that fi-
nancial uncertainty can help explain the time-varying response of stock prices to
MP shocks documented by previous studies (Galí and Gambetti, 2015; Jansen and
Zervou, 2017). Finally, the paper contributes to a series of recent studies that use
the stock market’s response to MP announcements to decompose MP surprises
into exogenous and information shocks (Jaroci´nski and Karadi, 2020). This paper
highlights a possible shortcoming of such an identification: Since the interpreta-
tion of a MP shock by financial markets might be state-dependent on the level of
stress (uncertainty), two MP shocks can result in different market reactions even
when they are similar in magnitude and information.
From a policy perspective, the results highlight the possible adverse effects of
failing to monitor uncertainty when making MP decisions. For example, when
uncertainty in the financial markets is high, a strong accommodative policy sur-
prise may induce an over-aggressive reaction by the markets, which could lead to
excessive risk-taking and a higher risk of asset price bubbles. These unintended
consequences may lead to additional interventions and weaken the central bank’s
6
credibility. In line with Kurov and Gu (2016), we find that MP can also be helpful
to boost equity prices, especially during periods of financial stress. However, the
results also suggest that restrictive MP will be an ineffective tool to cool down the
stock market if policymakers are concerned that stock prices are deviating from
fundamentals, in line with Galí and Gambetti (2015).
The paper is structured as follows. Section 2 describes the empirical model,
and Section 3 presents the high-frequency data. Section 4 describes the empirical
findings and Section 5 the robustness and sensitivity tests. Section 6 draws some
policy implications, and Section 7 concludes.
2 Empirical Methodology
This section first presents a standard event-study specification following Kuttner
(2001). Specifically, we regress the S&P 500 stock return during a narrow window
around each FOMC announcement on the unexpected change in the federal funds
rate (FFR):
∆rt=α+β∆is
t+εt, (1)
where rtis the log change in the S&P 500 stock index between 10 minutes before
and 20 minutes after the FOMC announcement, ∆is
tis the surprise change in the
FFR, and εtis a stochastic error term that represents the effects unrelated to the
FOMC announcement that influence the stock index. αand βare the estimated
parameters.
This methodology mitigates endogeneity concerns. All the public information
available at the beginning of a narrow window is already incorporated into fi-
nancial markets. In these regressions, the error terms only contain information
revealed in a very narrow window. This methodology identifies a “pure” MP
shock, which is assumed to be orthogonal to this limited amount of information
(Bernanke and Kuttner, 2005; Gürkaynak et al., 2005).
To investigate the impact of financial uncertainty on stock markets’ sensitivity
to MP, we re-estimate Eq. 1 with an interaction term between the unexpected
policy change and the measure of financial market uncertainty:
∆rt=α+β1Ut1+β2∆is
t+β3∆is
tUt1+εt, (2)
where Ut1is the measure of uncertainty in financial markets, lagged one pe-
riod.7We next examine the possible asymmetric impact of market uncertainty by
estimating Eq. 2 separately for negative and positive policy surprises. Note that
according to the expected effect through the discount rate, for both sub-samples
7In the baseline specification, we use a lagged daily measure of financial uncertainty explained
in Section 3.
7
the coefficients on the positive and negative policy surprises should be negative,
such that stock prices drop following MP tightening (positive shock) and rise after
monetary conditions are loosened (negative shock).
We next examine the possible asymmetric impact of market uncertainty for
negative and positive policy surprises using the following specification:
∆rt=α+β1Ut1+εt
+β
2∆is
t+β
3∆is
tUt1D
t(3)
+β+
2∆is
t+β+
3∆is
tUt1D+
t,
where D
t(D+
t) is a dummy variable taking the value one if the surprise FFR
change is negative (positive) and 0 otherwise. The coefficient β
2and β+
2measures
the response of stock market prices to negative and positive surprises, respec-
tively. β
3and β+
3measure the influence of uncertainty on negative and positive
surprises, respectively. According to the expected effect through the discount rate,
the coefficients on positive (β+
2) and negative (β
2) policy surprises should both be
negative, such that stock prices drop following MP tightening (positive shock) and
rise after monetary conditions are loosened (negative shock).
Finally, we consider the possible nonlinear impact of financial uncertainty on
the relationship between MP and stock markets. Kontonikas et al. (2013) show that
stock prices typically increase as a response to unexpected FFR cute in the U.S.,
and that the reaction is stronger during “bad times" such as recessions and bear
markets. However, stocks did not react positively to expansionary shocks during
the GFC since they were perceived as a signal for worsening future economic con-
ditions. Therefore, extreme levels of financial stress may induce a non-positive
reaction of stocks to expansionary FFR shocks. We address this issue of nonlinear-
ity using a quantile regression model. Specifically, the quantile regression model
is expressed as follows:
rt=α(τ)+β(τ)
1Ut1+β(τ)
2∆is
t+β(τ)
3∆is
tUt1+εt(4)
where τ2[0, 1]. The regression coefficients depend on the τth quantile regression
function of stock returns.
3 Data
Based on an updated version of the Gürkaynak et al. (2005) dataset, we utilize
asset-price changes for 240 FOMC announcements from 1990 to 2016.8However,
our sample starts in 1994 since market participants had to infer the policy change
8We are grateful to Refet Gürkaynak for providing us with the data.
8
by observing open market operations and movements in the FFR before 1994, as a
press statement did not accompany FOMC decisions.
Previous studies suggest that the Global Financial Crisis (GFC) caused a struc-
tural change in the relationship between stock prices and MP (Kontonikas et al.,
2013). The GFC induced panic and a “flight to safety” reaction, which changed
the stock market’s response to policy changes. Additionally, the period following
the crisis featured unprecedentedly low policy rates and unconventional policy
measures. We account for these possible structural breaks by also reporting re-
sults for a sub-sample ending in December 2007. We use both scheduled (183) and
unscheduled (9) FOMC meetings.9Following much of the related literature, we
exclude the 9/17/2001 and 3/18/2009 target rate announcements.10
Our baseline measure of stock price returns is the log change in the S&P 500
stock index between 10 minutes before and 20 minutes after the announcement. To
measure MP surprises, we follow Jaroci´nski and Karadi (2020) and use the change
in the 3-month FFR futures traded on the Chicago Board of Trade in the window
around the FOMC policy decisions. Following the seminal work of Bloom (2009),
our primary measure of financial uncertainty is the Chicago Board Options Ex-
change Volatility Index (VIX). The VIX is commonly used as a measure of global
financial uncertainty, while local uncertainty measures (Jurado et al., 2015; Baker
et al., 2016, 2019) are used in Section 5. However, we acknowledge the potential
limitations of using the VIX as a proxy for financial uncertainty.11
Summary statistics for the full and pre-GFC samples and positive and negative
MP surprises are presented in Table 1.
9Results are insensitive to using only the scheduled meetings. See Section 5.
10The 9/17/2001 announcement occurred on the first trading day following the September 11
attacks. On 3/18/2009, the FOMC made its first announcement regarding purchases of longer-
term Treasuries and expansion of MBS purchases.
11Sentiment may also affect the direction of price reaction. For example, Stambaugh et al. (2015)
show that overpricing is higher following periods of high sentiment, and Mian and Sankaragu-
ruswamy (2012) reveal that sensitivity to good (bad) earnings is higher during high (low) senti-
ment periods. While the VIX may be related to sentiment, they are not the same concept. Sentiment
refers to the overall emotional state of investors, whereas the VIX is a measure of implied volatility
based on option prices. Therefore, our use of the VIX, which differs from investor sentiment, also
contributes to the literature. In addition, Bekaert et al. (2013) and Drechsler (2013) have discussed
and contributed to considering the variance risk premium as a measure of ambiguity. This sug-
gests that the VIX may be inappropriate for proxying financial uncertainty. Therefore, one could
consider alternative measures of uncertainty, such as the variance risk premium, to gain a more
comprehensive understanding of market conditions or sentiment that might yield results similar
to those obtained using the VIX and provide directions for further research.
9
Table 1: Descriptive Statistics
Full sample Obs. Mean St. Dev. Median Min. Max.
S&P 500 return 192 0.001 0.644 0.051 1.880 4.076
MP surprise 192 0.009 0.054 0 0.370 0.120
Uncertainty (VIX) 192 20.938 8.613 19.355 10.710 66.960
Pre-GFC
S&P 500 return 118 0.043 0.682 0.129 1.597 4.076
MP surprise 118 0.011 0.061 0 0.370 0.120
Uncertainty (VIX) 118 20.237 7.036 19.905 10.710 39.680
Positive surprise
S&P 500 return 58 0.223 0.527 0.239 1.597 1.665
MP surprise (+) 58 0.031 0.031 0.020 0.005 0.120
Uncertainty (VIX) 58 20.785 6.914 20.400 11.460 37.950
Negative surprise
S&P 500 return 81 0.177 0.789 0.139 1.880 4.076
MP surprise (-) 81 0.043 0.063 0.020 0.370 0.005
Uncertainty (VIX) 81 22.741 10.303 21.480 10.710 66.960
Notes: This table reports descriptive statistics for the full sample (February 1994-December 2016)
and the pre-GFC sample (February 1994-December 2007). S&P 500 return is the log change in the
S&P 500 index in the 30-minute window around the FOMC announcement. MP surprise is the
unexpected change in the FFR target rate, and Uncertainty is measured using the VIX index.
10
4 Results
4.1 Baseline Analysis
Fig. 1 presents a simple scatter plot of the change in the stock market against the
MP surprises in the 192 announcements, looking separately at periods of high and
low uncertainty.12 The negative relationship between MP surprises and the stock
market return is visible for periods of both low and high uncertainty. However,
this negative relationship appears stronger when financial uncertainty is high,
providing suggestive evidence that financial uncertainty can amplify the reaction
of stock markets to MP shocks.
Figure 1: MP Surprises and Stock Prices
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
−2
0
2
4
−0.3 −0.2 −0.1 0.0 0.1
MP surprise, %
S&P500 index return (30−minute window), %
Uncertainty ●High Low
Notes: The figure shows a scatter plot of the change in the S&P 500 stock index between 10 minutes
before and 20 minutes after a MP announcement, against the surprise component of the policy
decision. The level of financial uncertainty splits observations, measured using the VIX at the
beginning of the announcement, relative to the full sample median.
Table 2 reports the results of estimating Eq. 1 and Eq. 2 for all types of MP sur-
12A point is defined as low uncertainty if the VIX at the beginning of the decision day was below
the full sample median. Otherwise, it is defined as high uncertainty.
11
prises, and for positive and negative MP surprises. Columns (1) and (3) confirm
previous studies that show that MP surprises significantly impact stock market re-
turns. The negative coefficients indicate that an unexpected increase in the policy
rate induces a fall in stock prices as predicted by the direct impact of policy rates
through the discount rate.
Estimates of Eq. 2 appear in columns (2), (4) and (6) of Table 2, which allow
for the interaction with financial uncertainty. The coefficients on the interaction
term are negative and significant for both samples but not for positive surprises,
suggesting that the impact of MP surprises is significant during periods of high
uncertainty and is asymmetrically effective for negative surprises than for posi-
tive.
Comparing the results from the two samples, the magnitude of the interaction
term is larger in the pre-GFC period, in line with Kontonikas et al. (2013), who
found that the GFC weakened the negative relationship between MP surprises and
stock markets. Additionally, for the full sample, the magnitude of the coefficient
is larger for the negative policy shocks, in line with Chuliá et al. (2010), who find
that the response of stock returns to negative policy surprises is stronger and more
significant than it is to positive changes.
As explained in the introduction, several possible explanations exist for the
asymmetric impact of uncertainty. One reason is the asymmetric impact of the
“Fed put view,” which postulates that market participants interpret accommoda-
tive MP during periods of heightened uncertainty such as the Fed doing “what-
ever it takes” to help markets, while restrictive shocks are interpreted as a positive
signal regarding the economic outlook. Another possible explanation is that both
positive and negative MP announcements decrease the level of uncertainty since
they both reveal new information. The drop in uncertainty reduces the equity risk
premium, which positively impacts stock prices, thereby mitigating the direct ef-
fect of restrictive MP and strengthening the direct effects of accommodative mon-
etary conditions. We examine these different explanations formally in the next
section. Overall, the results suggest that MP’s impact on stock markets depends
on the level of financial uncertainty, and that the relationship could shift given the
direction of the policy change.
The asymmetric nature of this uncertainty channel is also highlighted in Table
3, presenting the results from estimating Eq. 3, which considers the asymmetric
effect of positive and negative policy surprises.
Table 3 displays the estimation results over the full and pre-GFC samples.
Columns (1) and (3) present results without the interaction with uncertainty for
the full and pre-GFC samples. Both the positive policy surprise coefficient (β+
2)
and the negative policy surprise coefficient (β
2) are negative and significant in
both samples. This implies that, in general, stock prices experience negative co-
12
Table 2: Responses of Stock Returns to Unexpected FFR Changes: Controlling for
Uncertainty.
S&P 500 index return (30 min. window)
All surprises Positive surprise Negative surprise
(1) (2) (3) (4) (5) (6)
Full sample
MP surprise 6.186 1.918 6.089 2.618 6.690 2.804
(1.071) (1.800) (1.973) (6.052) (1.358) (1.984)
Uncertainty 0.021 0.003 0.045
(0.006) (0.013) (0.008)
MP surprise 0.178 0.144 0.382
Uncertainty (0.059) (0.276) (0.050)
Observations 192 192 58 58 81 81
Adjusted R20.264 0.326 0.110 0.080 0.273 0.454
Pre-GFC sample
MP surprise 6.957 0.028 4.302 6.064 7.337 4.961
(1.200) (2.089) (1.623) (5.879) (1.489) (3.396)
Uncertainty 0.018 0.008 0.044
(0.007) (0.013) (0.009)
MP surprise 0.290 0.5170.497
Uncertainty (0.100) (0.273) (0.147)
Observations 118 118 39 39 53 53
Adjusted R20.380 0.421 0.048 0.039 0.359 0.461
Post-GFC sample
MP surprise 3.740 4.124 7.461 10.894 3.761 9.791
(2.396) (5.324) (3.720) (40.544) (3.768) (6.734)
Uncertainty 0.028 0.001 0.048
(0.009) (0.025) (0.011)
MP surprise 0.2640.113 0.455
Uncertainty (0.140) (1.299) (0.133)
Observations 74 74 19 19 28 28
Adjusted R20.056 0.235 0.137 0.023 0.026 0.496
Notes: This table reports the estimates of the empirical specification described in Eq. 1 and Eq.
2 where the dependent variable is the S&P 500 measured in a 30-minute window around FOMC
announcements. MP surprise is the unexpected change in the FFR target rate, and Uncertainty is
the one-day lagged VIX index. White heteroskedasticity-consistent standard errors are reported in
parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
13
Table 3: Asymmetric Responses to Positive and Negative Surprises
S&P 500 index return (30 min. window)
Full sample Pre-GFC
(1) (2) (3) (4)
Positive MP surprise 5.684 12.563 4.924 7.238
(1.458) (4.355) (1.457) (4.789)
Negative MP surprise 6.329 0.423 7.452 0.899
(1.330) (1.982) (1.468) (2.723)
Uncertainty 0.027 0.023
(0.007) (0.007)
Positive MP surprise 0.3300.118
Uncertainty (0.197) (0.240)
Negative MP surprise 0.250 0.338
Uncertainty (0.049) (0.119)
Observations 192 192 118 118
Adjusted R20.261 0.336 0.380 0.418
Notes: This table reports the estimates of the empirical specification described in Eq. 3, where
the dependent variable is the S&P 500 measured in a 30-minute window around FOMC announce-
ments. MP surprise is the unexpected change in the FFR target rate, and Uncertainty is the one-day
lagged VIX index. White heteroskedasticity-consistent standard errors are reported in parentheses.
***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
14
movement with both positive and negative MP surprises. That is, an unexpected
increase in the policy rate induces a fall in prices, while an unexpected decrease is
related to an increase in prices.
Columns (2) and (4) present the results while also considering the impact of
uncertainty. Interestingly, the effect of uncertainty significantly differs between
positive and negative policy surprises. For the positive MP surprises, the policy
coefficient (β+
2) is negative in both samples and significant for the full sample.
The interaction with the uncertainty measure (β+
3) is positive for both samples
and significant at a 10% level only for the full sample. The policy coefficient (β
2)
is positive and significant for the negative policy surprise, while the interaction
term (β
3) is negative and significant in both columns.
The results demonstrate an interesting asymmetry between periods of low and
high financial uncertainty: During periods of low uncertainty, positive policy sur-
prises will harm stock returns as expected by standard theory, but negative policy
surprises will be ineffective. However, during periods of high uncertainty, the
relationship changes, with negative policy changes inducing a strong positive re-
action by the stock market and positive changes having little or even a counter-
intuitive positive effect. These results align with the hypothesis that financial un-
certainty has an amplifying impact on accommodative MP shocks and a mitigat-
ing effect or no effect on restrictive MP shocks.
4.2 Monetary Policy and Financial Uncertainty
The previous section established that periods of high financial uncertainty are as-
sociated with a stronger reaction of stock prices to MP surprises, and that this rela-
tion is especially relevant for accommodative policy changes (asymmetry). There-
fore, the results suggest that MP decisions are at least partly transmitted to the
stock market through their effect on uncertainty, and that this channel is stronger
for periods of high uncertainty. In this section, we provide some evidence for this
interpretation by examining how uncertainty in financial markets changes in the
days around the policy decision.
Fig. 2 presents the average change in the VIX index in trading days around the
FOMC decisions. Specifically, the figure displays the cumulative average changes
in the natural log of VIX relative to the value of the announcement day in the
five days before and after the announcement with 90% confidence intervals. The
full sample is split into six groups based on the level of uncertainty (High/Low)
and the direction of the policy change (positive/negative/neutral). An FOMC
announcement is characterized as part of the “High uncertainty” level if the VIX
at the beginning of the announcement day is higher than the full sample median,
and as part of the “Low uncertainty” level otherwise. As before, the direction
of the policy change is measured using the change in the 3-month FFR futures
15
around the FOMC decision.
Fig. 2 illustrates that the decline in the VIX following FOMC announcement
days is mostly associated with negative, i.e., accommodative policy changes. This
result is in line with Äijö and Vähämaa (2011) and Fernandez-Perez et al. (2017),
who find that the VIX declines following FOMC announcements, and this relation-
ship is mostly driven by negative surprises.Panels (c) and (d) add another layer to
these studies by distinguishing between negative surprises in periods of low and
high uncertainty. The panels show that the negative impact on uncertainty fol-
lowing negative surprises is stronger and more significant during periods of high
uncertainty. Therefore, Fig. 2 suggests that the results presented in the previous
sections are at least partly driven by a reduction in financial stress following an
accommodative MP surprise, and that this amplification channel is most relevant
when the level of uncertainty is high.
4.3 Nonlinearities
The possible nonlinear impact of financial uncertainty on the relationship between
stocks and MP is further investigated using quantile regression estimation. The
quantile estimates of β(τ)
3from estimating Eq. 4 over the full and pre-GFC sam-
ples are presented in Fig. 3. The plots also show the 10% confidence intervals and
the OLS estimated parameter with its corresponding 90% confidence band. The
x-axis shows the quantile, and the y-axis displays the size of the estimated coeffi-
cient. The negative interaction between MP surprises and uncertainty is statisti-
cally significant in the mid-upper range of the return distribution (Panel A, Fig. 3).
Interestingly, at the low quantiles the interaction between MP and uncertainty is
positive (albeit not significant) suggesting a positive relationship between MP and
stock prices during periods of negative returns and high uncertainty, as was the
case during the great recession. The quantile regression results indicate that the
role of financial uncertainty in amplifying the sensitivity of stocks to MP shocks
is state-dependent. That is, during periods of “normal" (around the median) re-
turns, more uncertainty will straighten the negative relationship between stocks
and policy rates. However, under extreme market conditions, more uncertainty
can result in a weaker or even positive relationship between stocks and interest
rates. As with the baseline estimation, the results are overall consistent for the full
sample and the pre-GFC period.
Estimating the quantile regression separately for the positive (Panel C) and
negative (Panel D) surprises, it seems that the exemplifying impact of uncertainty
is much stronger and more significant for the negative surprises versus the posi-
tive surprises. Again, the results align with the hypothesis that the level of uncer-
tainty will not significantly impact the connection between positive MP surprises
and will strengthen the impact of negative surprises. The quantile regression also
16
Figure 2: Changes in Uncertainty around Monetary Policy Decisions
−8
−4
0
4
8
−5 −4 −3 −2 −1 0 1 2 3 4 5
Trading days around FOMC meeting
Change in VIX (%)
a. Low uncertainty + Positive MP surprise
−8
−4
0
4
8
−5 −4 −3 −2 −1 0 1 2 3 4 5
Trading days around FOMC meeting
Change in VIX (%)
b. High uncertainty + Positive MP surprise
−8
−4
0
4
8
−5 −4 −3 −2 −1 0 1 2 3 4 5
Trading days around FOMC meeting
Change in VIX (%)
c. Low uncertainty + Negative MP surprise
−8
−4
0
4
8
−5 −4 −3 −2 −1 0 1 2 3 4 5
Trading days around FOMC meeting
Change in VIX (%)
d. High uncertainty + Negative MP surprise
−8
−4
0
4
8
−5 −4 −3 −2 −1 0 1 2 3 4 5
Trading days around FOMC meeting
Change in VIX (%)
e. Low uncertainty + Neutral MP surprise
−5
0
5
−5 −4 −3 −2 −1 0 1 2 3 4 5
Trading days around FOMC meeting
Change in VIX (%)
f. High uncertainty + Neutral MP surprise
Notes: This figure shows the average change in MP uncertainty on trading days around the FOMC
decisions, relative to the day before the FOMC decision day. The full sample includes 192 FOMC
decisions between February 1994 and December 2016. The full sample is split into six samples by
the level of financial uncertainty and direction of the policy change. The level of uncertainty is
measured using the VIX at the beginning of the announcement day relative to the sample median
(High/Low uncertainty). The direction of the policy change (positive/negative MP surprise) is
measured as the change in the 3-month FFR futures around the FOMC decision. The shaded gray
region shows 90% bootstrap confidence intervals. Number of announcements per panel: (a), 28;
(b), 30; (c), 34; (d), 47; (e), 34; and (f), 19.
17
shows that the amplifying impact is specifically relevant for periods of "normal"
to high returns on stock markets.
Figure 3: Quantile Regression Estimates
Panel A: Full sample Panel B: Pre-GFC
0.2 0.4 0.6 0.8
−0.6 −0.2 0.0 0.2 0.4
Estimate of β3
(τ)
0.2 0.4 0.6 0.8
−1.0 −0.5 0.0
Estimate of β3
(τ)
Panel C: Positive surprise Panel D: Negative surprise
0.2 0.4 0.6 0.8
−3 −2 −1 0 1 2
Estimate of β3
(τ)
0.2 0.4 0.6 0.8
−2.0 −1.5 −1.0 −0.5 0.0
Estimate of β3
(τ)
Notes: Quantile regression estimates for β(τ)
3in Eq. 4, where τ=0.1, 0.2, , 0.9. The shaded areas
represent the 90% confidence intervals. The solid red line represents the OLS estimated parameter
with corresponding 90% confidence intervals.
5 Robustness and Sensitivity Checks
In this section, we show that our conclusions are robust to several sensitivity tests.
5.1 Alternative Measures of Uncertainty
One might be concerned that a broader measure should represent financial market
uncertainty than the one addressed by the implied volatility of the stock market. A
18
further concern is that higher volatility may be critical in high-frequency analysis.
Finally, it is possible that the results found in the paper are not specific to financial
uncertainty but also hold for other category-specific measures.13
To address these concerns, we test if the results hold when using alternative
financial and nonfinancial uncertainty measures. For alternative measures of fi-
nancial uncertainty, we use the newspaper-based Equity Market Volatility (EMV)
measure developed by Baker et al. (2019) and the financial uncertainty measure
developed by Jurado et al. (2015). For nonfinancial measures, we use the macro
and real uncertainty indices also developed by Jurado et al. (2015). The empirical
specification is the same as Eq. 2 with Ut1corresponding to one of the alternative
uncertainty measures. Note that since the alternative measures are only available
monthly we use the measure in the month before each announcement.14
Results when using the alternative financial uncertainty measures are presented
in Table 4. Panel A presents results for all surprises, Panel B for the positive MP
surprises, and Panel C for negative MP surprises. For each panel, column (1)
presents the results using the EMV uncertainty measure, and columns (2) to (4)
present the results for Jurado et al. (2015). For each measure of uncertainty, we
present results for the full sample.15 For negative MP surprises, the interaction
between the MP surprise and the financial uncertainty measures is negative and
highly significant for all uncertainty measures. Thus, the results align with the
baseline specification, suggesting they are not only driven by VIX.
5.2 Recession and Bear Markets
As explained in the introduction, many papers have found that the impact of MP
on stock prices is stronger during “bad times”, such as recessions or bear mar-
kets.16 Since these periods are associated with higher levels of financial uncer-
tainty, an additional concern is that the results presented in this paper are picking
up the same impact as these previous studies. Furthermore, suppose the results
are specific to extreme macroeconomic and financial conditions. In that case, it
raises questions about interpreting the specific role of uncertainty, as these peri-
ods may have other unique characteristics that may cause the stock market to be
more sensitive to MP changes. Consequently, we re-estimate Eq. 1 and Eq. 2, ex-
13Unlike analyzing financial uncertainty, some studies analyze other uncertainty types. For in-
stance, Bauer et al. (2022) show that low MP uncertainty increases the effect of MP surprises on
asset prices.
14These measures are regularly updated. The EMV index is available at poli-
cyuncertainty.com/EMV_monthly.html. Jurado et al. (2015) measures are available at
sydneyludvigson.com/data-and-appendixes.
15Pre- and Post-GFC results are available upon request.
16Central banks often change policy rates in the face of high financial stress (Baxa et al., 2013).
19
Table 4: Alternative Measures of Uncertainty
S&P 500 index return (30 min. window)
Equity Market Financial Macroeconomic Real Economic
Volatility Uncertainty Uncertainty Activity Uncertainty
(1) (2) (3) (4)
Panel A. All surprises
MP surprise 1.575 1.581 2.433 1.026
(2.041) (3.544) (4.456) (6.547)
Uncertainty 0.017 0.499 0.634 0.195
(0.006) (0.239) (0.567) (1.103)
MP surprise 0.174 7.673 5.518 11.333
Uncertainty (0.063) (3.585) (6.079) (10.055)
Observations 192 192 192 192
Adjusted R20.310 0.287 0.265 0.260
Panel B. Positive surprise
MP surprise 3.690 0.569 6.460 22.071
(5.071) (7.547) (11.975) (11.547)
Uncertainty 0.001 0.077 0.019 4.163
(0.011) (0.511) (1.749) (1.637)
MP surprise 0.105 5.594 0.505 45.061
Uncertainty (0.222) (7.286) (17.624) (17.120)
Observations 58 58 58 58
Adjusted R20.080 0.083 0.077 0.127
Panel C. Negative surprise
MP surprise 3.415 13.212 4.564 14.858
(3.280) (4.800) (4.777) (6.707)
Uncertainty 0.041 1.743 1.8282.801
(0.009) (0.411) (1.078) (1.971)
MP surprise 0.377 19.530 16.394 33.790
Uncertainty (0.098) (4.529) (5.981) (9.817)
Observations 81 81 81 81
Adjusted R20.423 0.385 0.283 0.280
Notes: This table reports the estimates of the empirical specification described in Eq. 2, where the
dependent variable is the S&P 500 measured in a 30-minute window around FOMC announce-
ments. The full sample (February 1994-December 2016) is used in Panel A, while positive and
negative MP surprises are used in Panels B and C, respectively. Uncertainty is the one-month-
lagged financial uncertainty index constructed by Baker et al. (2019) in column (1), and the finan-
cial, macro and real uncertainty indices developed by Jurado et al. (2015) in columns (2),(3) and
(4). White heteroskedasticity-consistent standard errors are reported in parentheses. ***, **, and *
indicate significance at the 1%, 5%, and 10% level, respectively.
20
cluding any announcement that happened during a recession or a bear market.17
Table 5 presents the results where the NBER recession periods are excluded
(Panel A), and where bear market states are excluded (Panel B). Results are con-
sistent with the baseline estimation. This suggests that the impact of uncertainty
revealed in this paper is not specific to extreme market conditions. That is, the
uncertainty level also significantly impacts the connection between negative MP
surprises and stock markets in “normal time”.
5.3 Non-Scheduled Announcements
We also check if a few specific unscheduled policy announcements drive the re-
sults. It is possible that these announcements, which were not anticipated by the
markets, caused large movements that may have significantly impacted the results
and interpretation. Table 6 presents the results of estimating Eq. 1 and Eq. 2 with
the scheduled announcement sample only. While some of the significance is lost
for the full sample, the results are robust overall, with a negative coefficient on the
interaction between negative MP surprises and the level of uncertainty.
5.4 Comparative Analysis
This section aims to investigate whether Israeli data also supports the baseline US
results. To that end, we use the same empirical specification described in Section 2
to test the reaction of Israeli stocks to the Bank of Israel (BoI) policy surprises. Our
primary measure of stock return is the log change in Tel Aviv 100 (TA-100) stock
index. The measure of BoI policy shocks is borrowed from Kutai (2020), who cal-
culated the surprise component of the interest rate change using the change in the
1-month Telbor rate in a 24-hour window around each BoI policy decision. Finally,
we use the daily implied volatility of TA-25 index options to measure financial un-
certainty in Israel.
One issue with Israeli data is that until April 2014, BoI interest rate announce-
ments were at 17:30, after the Tel Aviv stock market close. As a result, until April
2014, the impact of the policy change was only reflected a day after the announce-
ment. We, therefore, use the overnight returns (i.e., the log difference between
the opening prices of the day after the announcement and the closing prices of
the announcement day). Alternatively, we also consider the return over a full-day
window, which starts at the close of the announcement day and ends at the close
of the next trading day. Our final Israeli sample includes 75 policy announcements
17Recessions are identified using the contraction period identified by the NBER dates. The dates
of bull and bear market states are from the website accompanying Zakamulin (2017) available at
vzakamulin.weebly.com/the-book.html. Specifically, we take the dates that are found using the
method defined by Pagan and Sossounov (2003).
21
Table 5: Controlling for Bad and Good Times
S&P 500 index return (30 min. window)
Full sample Positive surprise Negative surprise
(1) (2) (3) (4) (5) (6)
Panel A. Excluding recessions
MP surprise 7.568 0.981 6.722 0.861 8.271 3.326
(1.884) (2.665) (2.120) (5.791) (2.817) (4.194)
Uncertainty 0.018 0.001 0.037
(0.006) (0.010) (0.007)
MP surprise 0.277 0.376 0.453
Uncertainty (0.128) (0.254) (0.168)
Observations 170 170 51 51 68 68
Adjusted R20.259 0.309 0.109 0.098 0.267 0.382
Panel B. Excluding bear market
MP surprise 6.214 2.345 9.085 1.462 5.022 3.906
(1.199) (2.331) (3.483) (8.912) (1.800) (2.372)
Uncertainty 0.009 0.013 0.028
(0.006) (0.014) (0.012)
MP surprise 0.157 0.491 0.356
Uncertainty (0.072) (0.356) (0.066)
Observations 146 146 44 44 54 54
Adjusted R20.221 0.231 0.130 0.106 0.206 0.302
Notes: This table reports the estimates of the empirical specification described in Eq. 1 and Eq.
2, excluding NBER recession periods (columns (1) and (2)) and bear market states (columns (3)
and (4)). The dependent variable is the S&P 500 measured in a 30-minute window around FOMC
announcements. MP surprise is the unexpected change in the FFR target rate, and Uncertainty is
the one-day lagged VIX index. White heteroskedasticity-consistent standard errors are reported in
parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
22
Table 6: Excluding Unscheduled Announcements
S&P 500 index return (30 min. window)
Full sample Positive Surprise Negative Surprise
(1) (2) (3) (4) (5) (6)
MP surprise 5.231 1.770 8.233 1.050 4.556 4.421
(1.013) (2.386) (2.217) (8.726) (1.572) (1.869)
Uncertainty 0.023 0.008 0.050
(0.006) (0.013) (0.007)
MP surprise 0.1560.375 0.406
Uncertainty (0.091) (0.341) (0.047)
Observations 183 183 55 55 75 75
Adjusted R20.154 0.253 0.148 0.132 0.105 0.475
Notes: This table reports the estimates of the empirical specification described in Eq. 3, where
the dependent variable is the S&P 500 measured in a 30-minute window around FOMC announce-
ments. MP surprise is the unexpected change in the FFR target rate, and Uncertainty is the one-day
lagged VIX index. White heteroskedasticity-consistent standard errors are reported in parentheses.
***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
23
between October 2006 to April 2014.
Figure 4: MP Surprises and Stock Prices in Israel
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
−2
0
2
4
−0.4 −0.2 0.0 0.2
MP surprise, %
TA−100 index return (Overnight window), %
Uncertainty ●High Low
Notes: The figure shows a scatter plot of the change in the TA-100 stock index between the closing
prices of the announcement day and the opening prices of the day after the announcement against
the surprise component of the policy decision. The level of financial uncertainty splits observa-
tions, measured using the daily implied volatility of the TA-25 index option on the day before the
announcement relative to the full sample median.
Fig. 4 presents a scatter plot of the Israeli data along the lines of Fig. 1. That
is, it presents the change in the stock market against the MP surprise in the 75
announcements, looking separately at periods of high and low uncertainty.18 In
line with the US data, the negative relationship between MP surprises and the
stock market return is stronger when financial uncertainty is high.
Table 7 presents the baseline results with Israeli data. The level of financial un-
certainty appears related to not only the magnitude of the stock market response
to policy surprises but also the direction of the response. The results suggest that
while periods of medium-to-high levels of uncertainty are characterized by a neg-
18A point is defined as low uncertainty if the VIX on the day before the decision was below the
full sample median, and high otherwise.
24
Table 7: Responses of Stock Returns to Unexpected Interest Rate Changes: Con-
trolling for Uncertainty in Israel
TA-100 index return
Overnight window One day window
(1) (2) (3) (4)
MP surprise 3.167 4.752 1.751 5.968
(0.906) (1.413) (1.211) (2.417)
Uncertainty 0.025 0.023
(0.020) (0.027)
MP surprise 0.398 0.386
Uncertainty (0.083) (0.136)
Observations 75 75 75 75
Adjusted R20.190 0.451 0.024 0.164
Notes: This table reports the estimates of the empirical specification described in Eq. 1 and Eq.
2 using Israeli data. The dependent variable is the TA-100 index return in an overnight window
(columns (1) and (2)) and one-day window (columns (3) and (4)) around BoI announcements. MP
surprise is the unexpected change in the interest rate, and Uncertainty is the one-day lagged TA-
25 implied volatility index. White heteroskedasticity-consistent standard errors are reported in
parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
25
ative relation between unexpected interest rate changes and stock prices, the rela-
tion is positive during periods of low uncertainty. We also examine the possible
asymmetry response to positive and negative shocks.
The results for the asymmetric response are presented in Table 8. Like the US,
the results show that uncertainty plays a much larger role in accommodative pol-
icy changes than in restrictive changes. This suggests that, in Israel, MP shocks
will have the most significant impact on stock prices during periods of high fi-
nancial uncertainty and when the policy stance is accommodative. The online
appendix contains several robustness tests for the Israeli data. We show that re-
sults are robust to using the Tel Aviv 25 stock price as an alternative stock price
index. We also show that the results are robust to using a survey-based measure
of MP shocks, calculated as the difference between the actual MP decisions and
the mean of experts’ forecasts.
Table 8: Asymmetric Responses to Positive and Negative Surprises in Israel
TA-100 index return
Overnight window One day window
(1) (2) (3) (4)
Positive MP surprise 2.124 2.948 0.028 5.052
(0.796) (6.957) (1.152) (7.764)
Negative MP surprise 3.711 5.176 2.649 7.056
(1.414) (1.537) (1.861) (2.858)
Uncertainty 0.028 0.039
(0.023) (0.034)
Positive MP surprise 0.305 0.303
Uncertainty (0.441) (0.472)
Negative MP surprise 0.414 0.465
Uncertainty (0.093) (0.158)
Observations 75 75 75 75
Adjusted R20.186 0.436 0.023 0.155
Notes: This table reports the estimates of the empirical specification described in Eq. 3 using Israeli
data. The dependent variable is the TA-100 index return in an overnight window (columns (1) and
(2)) and one-day window (columns (3) and (4)) around BoI announcements. MP surprise is the
unexpected interest rate change, and Uncertainty is the one-day-lagged TA-25 implied volatility
index. White heteroskedasticity-consistent standard errors are reported in parentheses. ***, **, and
* indicate significance at the 1%, 5%, and 10% level, respectively.
26
6 Policy Implications
Our results highlight the importance of monitoring uncertainty and the stock mar-
ket before MP decisions. For instance, a sharp and surprising decrease in the pol-
icy rate when uncertainty in the financial markets is high may trigger an aggres-
sive reaction on the part of stock market prices. This could increase excessive
risk-taking and asset price bubble risk. As financial stability is one of the primary
objectives of any central bank,19 such unintended consequences that may lead to
complementary conventional or unconventional interventions may weaken the
central bank’s credibility. We highlighted the possible adverse effects of failing to
monitor uncertainty by policymakers.
MP surprises can also be useful to boost equity prices, especially under finan-
cial stress periods (Kurov and Gu, 2016). Stock markets react more positively after
a negative MP surprise under high uncertainty, which could be related to mar-
ket makers’ knowledge in Fig. 2, panel (d). Consequently, uncertainty and the
risk premium will decrease, and stock market prices will rise. Alternatively, the
public, and thus market makers, expect less restrictive policies in good times, as
displayed in Fig. 2, panel (c).
Interestingly, our results also suggest that restrictive MP will be an ineffective
tool to cool down stock market prices if policymakers are concerned that they
deviate from fundamentals, in line with Galí and Gambetti (2015).
MP committees sometimes decide to surprise the market. In most cases, the
MP committee knows its announcement will surprise market players. Although
the outcomes of MP surprises are undetermined, they are part of the policy toolkit
and instruments central bank committees have at their disposal. Such a commit-
tee is committed to the central bank’s objectives (usually, inflation target, output
growth, and financial stability). As such, their decision to surprise is not necessar-
ily intentional. Our results allow the policymaker to better understand the trans-
mission channel and determine these outcomes on stock market prices depending
on the financial uncertainty level of the financial markets.
While the following policy recommendations depend on the central bank’s
credibility, limiting their frequent use, this instrument may also influence real ac-
tivity in two ways. First, MP surprises may change public opinion and expec-
tations, which influence uncertainty levels and, thus, stock prices (Di Bella and
Grigoli, 2019). Second, the relation between stock market returns and real activity
assessed in the literature (Fama, 1990; Schwert, 1990; Jay Choi et al., 1999), and our
results, provide a new perspective to policymakers.20 Depending on the financial
19Although the complex trade-off between monetary and financial stability (De Graeve et al.,
2008), MP tightening surprises do not necessarily reduce the systemic risk (Laséen et al., 2017).
20Recent studies have shown that the effect of MP shocks on real economic activity is lower
in the US and the eurozone during periods of high uncertainty (Aastveit et al., 2017; Pellegrino,
27
uncertainty level, they can use MP surprises to influence future (real) activity. In
both cases, our paper provides background for MP committees regarding how and
when to use the MP surprise instrument to influence the real economy.
Finally, our results could highlight a liquidity transmission channel. Accord-
ing to Nagel (2012), the return from liquidity provision is highly predictable with
the VIX index. In response to FOMC announcements, Chung et al. (2013) report
that liquidity disruption lasts for approximately 1.5 hours, with variability propor-
tional to the information content of the FOMC announcement, larger effects being
associated with unscheduled announcements and scheduled announcements with
larger policy surprises. Following a substantial market downturn, Hameed et al.
(2010) find that bid-ask spreads and reversal strategy returns increase weeks fol-
lowing large stock market declines. The FOMC announcement effects associated
with extremely elevated VIX levels do not necessarily support the "downside risk
channel" hypothesis. In this scenario, liquidity evaporation may be the dominant
mechanism.
7 Conclusion
State-of-the-art shock identification and high-frequency stock market data are used
to analyze the effect of MP on stock prices under financial uncertainty. Specifically,
this paper investigates to what extent the level of financial uncertainty influences
the effect of MP surprises on the US stock market. The results suggest that, in
general, MP shocks have a more considerable impact on the stock market during
uncertain times. This finding aligns with recent studies that suggest that MP af-
fects stock prices through its effect on uncertainty and the equity risk premium.
That is, the evidence points to an uncertainty channel of the transmission of MP
to the stock market.
The response of stock prices to MP shocks is found to be asymmetric concern-
ing the direction of the policy change–restrictive vs. accommodative. While fi-
nancial uncertainty is related to a stronger reaction to accommodative shocks, it
has little impact on the response to positive surprises. The movements of the VIX
around the days on which FOMC decisions are announced suggest that the un-
certainty channel is indeed most relevant for periods of high uncertainty. Overall,
the results indicate that monitoring market conditions and the level of financial
uncertainty is crucial if policymakers are concerned with the reaction of the stock
market.
2018). However, we focus on the role of financial uncertainty in the transmission of MP to the stock
market, i.e., whether the state of uncertainty affects the reaction of equity prices to MP shocks.
28
References
Aastveit, K. A., Natvik, G. J., Sola, S., 2017. Economic uncertainty andthe influence
of monetary policy. Journal of International Money and Finance 76 (C), 50–67.
Äijö, J., Vähämaa, S., 2011. The Fed’s policy decisions and implied volatility. Jour-
nal of Futures Markets 31 (10), 995–1010.
Baker, S. R., Bloom, N., Davis, S. J., 2016. Measuring economic policy uncertainty.
Quarterly Journal of Economics 131 (4), 1593–1636.
Baker, S. R., Bloom, N., Davis, S. J., Kost, K. J., 2019. Policy news and stock market
volatility. NBER Working Papers 25720, National Bureau of Economic Research.
Bansal, R., Yaron, A., 2004. Risks for the long run: a potential resolution of asset
pricing puzzles. Journal of Finance 59 (4), 1481–1509.
Basistha, A., Kurov, A., 2008. Macroeconomic cycles and the stock market’s reac-
tion to monetary policy. Journal of Banking & Finance 32 (12), 2606–2616.
Bauer, M. D., Lakdawala, A. K., Mueller, P., 2022. Market-based monetary policy
uncertainty. Economic Journal 132 (644), 1290–1308.
Baxa, J., Horváth, R., Vašíˇcek, B., 2013. Time-varying monetary-policy rules and
financial stress: does financial instability matter for monetary policy? Journal of
Financial Stability 9 (1), 117–138.
Bekaert, G., Hoerova, M., Lo Duca, M., 2013. Risk, uncertainty and monetary pol-
icy. Journal of Monetary Economics 60 (7), 771–788.
Bernanke, B. S., Kuttner, K. N., 2005. What explains the stock market’s reaction to
Federal Reserve policy? Journal of Finance 60 (3), 1221–1257.
Bloom, N., 2009. The impact of uncertainty shocks. Econometrica 77 (3), 623–685.
Chuliá, H., Martens, M., Dijk, D. v., 2010. Asymmetric effects of federal funds tar-
get rate changes on S&P100 stock returns, volatilities and correlations. Journal
of Banking & Finance 34 (4), 834–839.
Chung, K. H., Elder, J., Kim, J.-C., 2013. Liquidity and information flow around
monetary policy announcement. Journal of Money, Credit and Banking 45 (5),
781–820.
Cieslak, A., Morse, A., Vissing-Jorgensen, A., 2019. Stock returns over the FOMC
cycle. Journal of Finance 74 (5), 2201–2248.
29
De Graeve, F., Kick, T., Koetter, M., 2008. Monetary policy and financial
(in)stability: An integrated micro-macro approach. Journal of Financial Stability
4 (3), 205–231.
Di Bella, G., Grigoli, F., 2019. Optimism, pessimism, and short-term fluctuations.
Journal of Macroeconomics 60 (C), 79–96.
Drechsler, I., 2013. Uncertainty, time-varying fear, and asset prices. Journal of Fi-
nance 68 (5), 1843–1889.
Fama, E. F., 1990. Stock returns, expected returns, and real activity. Journal of
Finance 45 (4), 1089–1108.
Fernandez-Perez, A., Frijns, B., Tourani-Rad, A., 2017. When no news is good news
- The decrease in investor fear after the FOMC announcement. Journal of Em-
pirical Finance 41 (C), 187–199.
Galí, J., Gambetti, L., 2015. The effects of monetary policy on stock market bubbles:
some evidence. American Economic Journal: Macroeconomics 7 (1), 233–257.
Gu, C., Kurov, A., Wolfe, M. H., 2018. Relief rallies after FOMC announcements as
a resolution of uncertainty. Journal of Empirical Finance 49 (C), 1–18.
Guo, H., Hung, C.-H. D., Kontonikas, A., 2021. Investor sentiment and the pre-
FOMC announcement drift. Finance Research Letters 38 (C).
Guo, H., Hung, C.-H. D., Kontonikas, A., 2022. The Fed and the stock market: A
tale of sentiment states. Journal of International Money and Finance 128 (C).
Gürkaynak, R. S., Sack, B., Swanson, E., 2005. Do actions speak louder than words?
The response of asset prices to monetary policy actions and statements. Interna-
tional Journal of Central Banking 1 (1), 55–93.
Hameed, A., Kang, W., Viswanathan, S., 2010. Stock market declines and liquidity.
Journal of Finance 65 (1), 257–293.
Jansen, D. W., Tsai, C.-L., 2010. Monetary policy and stock returns: Financing con-
straints and asymmetries in bull and bear markets. Journal of Empirical Finance
17 (5), 981–990.
Jansen, D. W., Zervou, A., 2017. The time varying effect of monetary policy on
stock returns. Economics Letters 160 (C), 54–58.
Jaroci´nski, M., Karadi, P., 2020. Deconstructing monetary policy surprises: the role
of information shocks. American Economic Journal: Macroeconomics 12 (2).
30
Jay Choi, J., Hauser, S., Kopecky, K. J., 1999. Does the stock market predict real
activity? Time series evidence from the G-7 countries. Journal of Banking &
Finance 23 (12), 1771–1792.
Jurado, K., Ludvigson, S. C., Ng, S., 2015. Measuring uncertainty. American Eco-
nomic Review 105 (3), 1177–1216.
Kontonikas, A., MacDonald, R., Saggu, A., 2013. Stock market reaction to fed
funds rate surprises: State dependence and the financial crisis. Journal of Bank-
ing & Finance 37 (11), 4025–4037.
Kroencke, T. A., Schmeling, M., Schrimpf, A., 2021. The FOMC risk shift. Journal
of Monetary Economics 120 (C), 21–39.
Kurov, A., 2010. Investor sentiment and the stock market’s reaction to monetary
policy. Journal of Banking & Finance 34 (1), 139–149.
Kurov, A., Gu, C., 2016. Monetary policy and stock prices: does the "Fed put" work
when it is most needed? Journal of Futures Markets 36 (12), 1210–1230.
Kurov, A., Stan, R., 2018. Monetary policy uncertainty and the market reaction to
macroeconomic news. Journal of Banking & Finance 86 (C), 127–142.
Kutai, A., 2020. Measuring the effect of forward guidance in small open
economies: the case of Israel. mimeo, Bank of Israel.
Kuttner, K. N., 2001. Monetary policy surprises and interest rates: Evidence from
the Fed funds futures market. Journal of Monetary Economics 47 (3), 523–544.
Lakdawala, A., Schaffer, M., 2019. Federal reserve private information and the
stock market. Journal of Banking & Finance 106 (C), 34–49.
Laséen, S., Pescatori, A., Turunen, J., 2017. Systemic risk: a new trade-off for mon-
etary policy? Journal of Financial Stability 32 (C), 70–85.
Lucca, D. O., Moench, E., 2015. The pre-FOMC announcement drift. Journal of
Finance 70 (1), 329–371.
Mian, G. M., Sankaraguruswamy, S., 2012. Investor sententiment and stock market
response to earnings news. Accounting Review 87 (4), 1357–1384.
Mishkin, F. S., 2017. Rethinking monetary policy after the crisis. Journal of Inter-
national Money and Finance 73 (PB), 252–274.
Nagel, S., 2012. Evaporating liquidity. Review of Financial Studies 25 (7), 2005–
2039.
31
Nakamura, E., Steinsson, J., 2018. High-frequency identification of monetary non-
neutrality: the information effect. Quarterly Journal of Economics 133 (3), 1283–
1330.
Pagan, A. R., Sossounov, K. A., 2003. A simple framework for analysing bull and
bear markets. Journal of Applied Econometrics 18 (1), 23–46.
Pellegrino, G., 2018. Uncertainty and the real effects of monetary policy shocks in
the Euro area. Economics Letters 162 (C), 177–181.
Schwert, G. W., 1990. Stock returns and real activity: a century of evidence. Journal
of Finance 45 (4), 1237–1257.
Stambaugh, R. F., Yu, J., Yuan, Y., 2015. Arbitrage asymmetry and the idiosyncratic
volatility puzzle. Journal of Finance 70 (5), 1903–1948.
Tsai, C.-L., 2014. The effects of monetary policy on stock returns: Financing con-
straints and "informative" and "uninformative" FOMC statements. International
Review of Economics & Finance 29 (C), 273–290.
Zakamulin, V., 2017. Market timing with moving averages: Anatomy and perfor-
mance of trading rules. New York, NY: Palgrave Macmillan.
32
