Whose speeches impact European markets: ECB's or the national central banks'?

• Published date: 01 November 2022
• Author: Abhinav Anand,Sankarshan Basu,Jalaj Pathak,Ashok Thampy
• Date: 01 November 2022
• DOI: http://doi.org/10.1111/eufm.12334

Eur Financ Manag. 2022;28:1413–1476. wileyonlinelibrary.com/journal/eufm © 2021 John Wiley & Sons Ltd.

|

1413

DOI: 10.1111/eufm.12334

ORIGINAL ARTICLE

Whose speeches impact European markets:

ECB's or the national central banks'?

Abhinav Anand

1

|Sankarshan Basu

1

|Jalaj Pathak

2

|

Ashok Thampy

1

1

Department of Finance and Accounting,

IIM Bangalore, Bangalore, Karnataka,

India

2

Department of Finance and Accounting,

Great Lakes Institute of Management,

Chennai, Tamil Nadu, India

Correspondence

Jalaj Pathak, Department of Finance and

Accounting, Great Lakes Institute of

Management, Chennai, Tamil Nadu

603102, India.

Email: jalaj.p@greatlakes.edu.in

Abstract

We quantify the tone from the speeches of the Eur-

opean Central Bank (ECB) as well as that from the

national central banks of six leading European

nations and analyze its role in explaining the returns

of their respective stock market indices. Using recent

innovations in financial text analysis, we find evi-

dence that except for France, all nations' stock indices

are significantly associated with the tone of speeches

delivered by either the national banks or the ECB

(or both). For France, the national stock index vola-

tility is found to be associated with its national central

bank speech tone.

KEYWORDS

central bank communication, central bank speech, European

Central Bank, financial text analysis, tone analysis

JEL CLASSIFICATION

G14, G18, G28, G41

1|INTRODUCTION

Although the effect of central bank policies on stock markets is an actively studied area of

research, the European Union (EU) provides a unique opportunity to compare the market

impact of actions undertaken by the national central bank (NCB) to that of a supranational

entity: the European Central Bank (ECB). In this study, we quantify the effect of speeches

EUROPEAN

FINANCIAL MANAGEMENT

We would like to thank the editor, John Doukas, and an anonymous referee for their insightful comments which

helped to improve the quality of the paper.

delivered by both NCB as well as those by the ECB on the respective national stock market

indices of six leading European nations: France, Germany, Italy, Spain, Ireland and Finland.

Central bank communication has been found to be significantly associated with an array of

economic variables such as interest rate (Demiralp & Jorda, 2004; Kohn & Sack, 2003; Lucca &

Trebbi, 2009; Smales & Apergis, 2017); money supply (Gerlach, 2007); currency market

(Dossani, 2018) and stock return and volatility (Apergis & Pragidis, 2019; Ehrmann &

Fratzscher, 2004; Savor & Wilson, 2013; Schmeling & Wagner, 2019). Hubert and Labondance

(2021) report that that the US Federal Open Market Committee (FOMC) statements explain

monetary surprises beyond policy announcements. Baranowski et al. (2021) use survey data to

analyze the expectation channel of monetary policy and find a significant impact on the interest

rate and inflation expectations. As specified by Schmeling and Wagner (2019), the central bank

communication impacts market expectations and thus can be associated with the market

return. This could be due to the information content of the communication which in turn

impacts the asset prices (Savor & Wilson, 2013). On similar lines, Doukas and Han (2021)

report that the sentiment impacts the beta and the market risk premium of the conditional

version of the Capital Asset Pricing Model (CAPM).

Although a majority of the previous literature has examined press releases and FOMC

statements (Gonzalez & Tadle, 2021; Hansen & McMahon, 2016; Lucca & Trebbi, 2009), we

examine another important yet understudied tool in the central bank communication toolkit—

speeches delivered by senior functionaries of central banks. Central bank speeches intend to

convey to the markets, its desires and wishes, regarding future paths of relevant policy variables

such as inflation, short‐term rates, unemployment numbers and so forth. For functionaries of

the central bank, it is vital that their message broadcasted to the market participants in the form

of speeches reaches their intended audience; and (even more importantly) is interpreted in the

same way that the bank intends them to be interpreted. Any miscommunication or mis-

interpretation in this regard can prove costly to the economy which can then be used to infer

that the central bank failed in fulfilling its mandate. Thus, a study of how central bank speeches

impact markets is vital for the central banker who wishes to convey accurate, unambiguous

information to market participants. Moreover, insofar as central bank communication can itself

be used for policy implementation as suggested in Guthrie and Wright (2000), any evidence

which connects the impact of central bank speeches to movements in the market helps the

central bank in gauging whether it is successfully transmitting its message.

For the investors also, the applications of studies connecting central bank speech impact to

the markets are essential. For example, a large part of central banks' mandate is to implement

the nation's monetary policy. In particular, several studies cited above show central bank

communication to be significantly associated with interest rates and inflation expectations.

1

Stock market securities are generally priced at a premium to risk‐free assets whose yields are

directly influenced by both benchmark interest rates as well as future inflation expectations—

an important subject matter for about 35% speeches on average for our study. Clearly, investors

should (and do) pay a lot of attention to speeches delivered by central banks.

The methodology used to extract the tone of central bank communication is borrowed from

Anand et al. (2021), using polar words (negative/positive) from the Loughran and McDonald

Dictionary (LM hereafter) (Loughran & McDonald, 2011) and polar phrases extracted in line

1

In our study, 14%–55% of different nations' central bank speeches are focused on interest rates or related variables such

as money supply, short‐term rates, quantitative easing and so forth.

1414

|

EUROPEAN

FINANCIAL MANAGEMENT

ANAND ET AL.

with Apel and Grimaldi (2014) and Apergis and Pragidis (2019), along with appropriately

weighed valence shifters (adjectives and adverbs) which modify the meaning of words but have

not been given weightage in the LM dictionary.

To the best of our knowledge, this is the first study which analyzes speeches delivered by the

European NCBs as well as those by the ECB and compares their putative influence on the

movement of European stock markets of their member nations. Further, our adaptation of

the novel tone quantification methodology of Anand et al. (2021), which proposes the usage of the

sentence as a unit of analysis and assigns proper weights to valence shifters—adverbs and

adjectives which modify the meaning of a sentence (such as ‘but’,‘slight’,‘very’,‘despite’,etc.)—

has not been applied to study central bank speeches in prior studies.

One of several, nonequivalent ways of categorizing existing literature on central bank

communication is by means of its tone quantification methodology. On the basis of this cri-

terion, there are two popular techniques for analyzing communication from central banks. The

first category includes the studies in which the central bank's communications' reaction is

quantified into a dummy classification (e.g., +1, 0, −1) based on the authors' subjective as-

sessment or a dictionary‐based analysis of its content by the researcher. For example, Guthrie

and Wright (2000) use central bank communication to show how central bank statements

(rather than open market operations) can be used to implement monetary policy in New

Zealand. The communication is classified into categories (+1, 0, −1) based on the authors'

subjective assessment and it is shown that the communication, rather than open market op-

erations causes the large changes in interest rates. The second category includes studies that

analyze the importance of speech days based on a dummy variable for the presence/absence of

speech. For example, Savor and Wilson (2013) show how macroeconomic announcements

affect market returns and the Sharpe ratio.

However, there are drawbacks to both categories of studies. For the first category, if the

communication is classified on the basis of researchers' intent, the results cannot be agreed

upon to be standard. On similar lines, the second category of studies classify the commu-

nication on the basis of its presence/absence and ignore its content.

The methodology of tone quantification, using polar dictionaries, ngram phrases and/or

‘bag‐of‐words’approach largely overcomes the limitations pertaining to the above two strands

of literature on central bank communication. Moreover, in this paper, we further improve the

tone quantification process by following the novel approach introduced in Anand et al. (2021)

and divide a speech into a set of sentences and extract the tone for each sentence considering

both the polar words (negative/positive) from the LM dictionary as well as ngram phrases using

the approach specified in Apel and Grimaldi (2014). These polar words/phrases are then used

in conjunction with adverbs and adjectives (valence shifters) to extract the accurate tone of the

central bank communication (Polanyi & Zaenen, 2006; Schulder et al., 2018).

We employ the union of two lexicons in this study: the LM dictionary for financial text

proposed by Loughran and McDonald (2011) and the Apel and Grimaldi (2014) dictionary

which characterizes the tone quantification with respect to central bank communication. Their

joint usage ensures that our tone quantification method assigns proper weights to words used

from the perspective of both financial and central bank text classification.

The valence shifters can be divided into four categories: adversative conjunction (e.g.,

‘although’,‘however’), negator (e.g., ‘nor’,‘not’), amplifier (e.g., ‘very’) and de‐amplifier (e.g.,

‘few’) and can alter the tone of the sentence. For example, for the sentence below: (taken from a

speech given by a member of German Central bank on 28 April 2012).

ANAND ET AL.EUROPEAN

FINANCIAL MANAGEMENT

|

1415