Advertisement

International Economics and Economic Policy

, Volume 16, Issue 1, pp 161–192 | Cite as

The impact of Brexit news on British pound exchange rates

  • Arthur KorusEmail author
  • Kaan Celebi
Original Paper

Abstract

Using event-study techniques, we investigate the impact of Brexit-related events on the spot exchange rate of the British pound against the euro and the US dollar. We want to find out whether Brexit-related news, including the Brexit referendum itself, has an impact on British pound exchange rates. By splitting our Brexit-related events into ‘good’ Brexit news and ‘bad’ Brexit news, we find that Brexit news has an impact on British pound exchange rates. Bad Brexit news is associated with a depreciation of the British pound against the euro and the US dollar whereas ‘good’ Brexit news appreciates the Pound against the euro. Furthermore, our empirical results suggest that market participants display a delayed reaction to bad Brexit news. As the referendum has clearly a significant impact on both British pound/euro and British pound/US dollar exchange rate volatility, the impact of Brexit news is only for the British pound/euro exchange rate volatility measurable. Besides the asymmetric volatility pattern towards positive and negative shocks in general, we find that the statistically significance and the magnitude of the impact of good Brexit news is higher than these of bad Brexit news. Concerning the British pound/US dollar exchange rate volatility, our results display a weak presence of volatility asymmetry in terms of shocks and good/bad Brexit news, respectively.

Keywords

Spot exchange rate Brexit Event-study EGARCH 

JEL classifications

C 32 C58 F 31 G15 

1 Introduction

In this paper we study the impact of the Brexit process on the spot exchange rate of the Pound sterling against the euro by using event-study techniques. However, we also analyse the effects of Brexit-related events on the spot exchange rate of the British pound against the US dollar. Hence, we want to find out whether Brexit-related news, including the Brexit referendum itself, has an impact on the external value and volatility of the British pound. On 23 June, 2016, the United Kingdom (UK) voted to leave the European Union (EU), a process which has become well-known as Brexit. Just one day after the referendum, stock market indices in the UK decreased sharply, uncertainty increased, and the Pound depreciated heavily against major currencies. The Pound depreciated against the US dollar by nearly 8 % and against the euro by 6 %, respectively. Brexit also triggered a more aggressive expansionary monetary policy by the Bank of England. Interest rate cuts and additional non-standard monetary policy measures conducted by the Bank of England further depreciated the Pound against major currencies. Undoubtedly, Brexit represents a significant shock to the economy of the United Kingdom. The Brexit process has already had significant short-term effects on financial markets and is expected to have long-term effects on real economic activity through increasing barriers to trade, labour and capital mobility, and hence through declining productivity in the United Kingdom (Gourinchas and Hale 2017).

Many studies focus on the long-term impact of Brexit on real economic activity in the UK. For example, Ottaviano et al. (2014) predict that increasing trade barriers would result in a 1.1% to 3.1% reduction of GDP by using a quantitative static general equilibrium trade model. Dhingra et al. (2017) estimate that Brexit would lead to a permanent 1.3% to 2.7% decline in UK consumption per capita using a quantitative trade model. By using gravity equations for bilateral trade, Dhingra et al. (2017) find that Brexit may lead to a reduction of UK income per capita of between 6.3 and 9.4%. The general finding of studies, which focus on the impact of higher trade costs due to Brexit, is that welfare in the United Kingdom will drop. Other studies focus on the impact of Brexit on foreign direct investment (FDI) and immigration. Baier and Welfens (2018), for example, use the gravity equation for FDI flows and find that Brexit will lead to a reduction of FDI inflows to the UK by 42%. Portes and Forte (2017) elaborate on the impact of lower immigration from the EU on the UK’s GDP per capita. They find that restricting immigration from the EU may reduce GDP per capita in the UK by between 0.9% and 3.4% until 2030.

There are also studies which focus on the short-run impact of Brexit on financial markets. In particular, these studies investigate the effects of Brexit events on stock markets, uncertainty (Baker et al. 2016) and exchange rates. Davies and Studnicka (2017) analyse the influence of Brexit-related events on stock movements in the UK by using event study methods. They find that the announcement of the referendum’s result led to a sharp decline of the FTSE 350. According to their results, the FTSE 350 lost 7 % of its value in the first two days after the announcement of the referendum result. Moreover, they find that stock markets in the UK reacted promptly and persistently. They also show that after the Brexit referendum, the share price of a firm heavily depended on its global value chain structure and export performance, respectively. By conducting an event study, Ramiah et al. (2017) find that stock prices of financials were particularly affected by the announcement of the referendum results. Caporale et al. (2018) apply long-memory techniques to elaborate on the impact of Brexit on the FTSE 100 Implied Volatility Index and on the British pound’s implied volatilities vis-à-vis the other main currencies. They find that Brexit led to significant changes in the degree of persistence of the FTSE 100 Implied Volatility Index and on the British pound’s implied volatility vis-à-vis the euro and the US dollar, respectively.

The Bank of England (2016a) elaborate on the impact of the Brexit referendum on the Sterling effective exchange rate index (ERI). The Bank of England (2016a) finds that the Sterling ERI was significantly affected by a measure of the referendum-related news flow, accounting for half of the depreciation of the British pound. Krause et al. (2016) investigate the impact of Brexit on the British pound by using poll survey data. They find that poll results pointing toward a Brexit led to a depreciation of the British pound against the euro and the Swiss franc, respectively. Arnorsson and Zoega (2016) also use poll survey data and find that a one percentage point increase in the proportion of respondents who want the UK to leave the EU depreciated the pound vis-à-vis the euro by 1.1%. By using betting odds, the findings of Belke et al. (2016) indicate that a one percentage point increase of the Brexit probability decreased the external value of the British pound by 0.12%. Additionally by focusing on poll survey data, Belke et al. (2016) find that the value of the pound declined by 0.23%.

By focusing on the impact of Brexit-related events on the spot exchange rate of the British Pound against the euro, our paper is related to the studies which investigate the short-term impact of Brexit on financial markets and in particular to the studies which elaborate on the impact of Brexit on the external value and volatility of the British pound. However, we also elaborate on the impact of Brexit news on the British pound/US dollar exchange rate. Our study differs from the literature which investigates the impact of Brexit on the value of British pound in at least two ways. First, the aforementioned studies focus on the impact of the Brexit referendum on the value of the British pound, whereas we consider 16 additional Brexit-related events. Second, we also consider the impact of Brexit-related news on the exchange rate volatility of the British pound/Euro exchange rate and the British pound/US dollar exchange rate, respectively.

Our results show that Brexit-related events as a whole have no statistically significant impact on the spot exchange rate of the British pound against the euro and the US dollar, respectively. However, by splitting our 17 Brexit-related events into ‘good’ Brexit news and ‘bad’ Brexit news, we find an impact of Brexit news on the British pound/euro exchange rate. Our results show that bad news is associated with a depreciation of the British pound against the euro and that good news led to an appreciation of the British pound vis-à-vis the euro. Furthermore, our results suggest that the spot exchange rate of the British pound against the US dollar reacted only to the release of bad Brexit news. Moreover, we find that market participants display a delayed reaction to bad and good Brexit news, respectively. Additionally, our results indicate that the impact of bad Brexit news on the spot exchange rate of the British pound against the euro is more persistent than the effect of good Brexit news.

Our results indicate that Brexit-related events as a whole have a significant impact on the British pound/euro exchange rate volatility; whereas the significant impact on the volatility of the exchange rate vis-à-vis the US dollar cannot be clearly shown. Besides the asymmetric volatility pattern towards positive and negative shocks in general, we find an asymmetric behaviour of market participants towards good and bad Brexit news in terms of the British pound/euro exchange rate volatility, namely that the statistical significance and the magnitude of the impact of good Brexit news is clearly higher than those of bad Brexit news. For the British pound/US dollar exchange rate volatility, our results indicate a weak presence of volatility asymmetry in terms of both shocks and good/bad Brexit news.

The remainder of the paper is organised as follows. Section 2 describes our data and our event identification method. In sections 3 and 4 we present our empirical approaches and discuss our empirical findings for the mean and volatility of the exchange rates. Finally, section 5 concludes.

2 Data description and events

The sample data set ranges from 1 January, 2013, to 30 March, 2018. Daily data for the spot exchange rate of the British pound against the euro is obtained from Datastream. For our regression analysis, we include three-month OIS rates, 10-year government benchmark bond yields, benchmark stock market indices and a measure for macroeconomic surprise for the United Kingdom, the Eurozone, and the US, respectively. All data is collected from Datastream. Additionally, we control for the impact of unconventional monetary policies conducted by the Bank of England (BoE), the European Central Bank (ECB), and the US Federal Reserve (FED) on the British pound/euro exchange rate and the British pound/US dollar exchange rate, respectively. In order to identify events associated with announcements of unconventional monetary policies, we look at all BoE news releases, ECB press releases, and FED press releases from 1 January, 2013, to 30 March, 2018. Additionally, we try to identify unconventional monetary policy events by reading the Financial Times. We include an unconventional monetary policy announcement in our event list whenever an announcement is covered in the press releases of each central bank and on the first three pages of the Financial Times. Hence, we follow the event identification method of Fratzscher et al. (2014). Furthermore, we use 10-year government bond futures as a market-based measure for monetary policy surprise. This data is also obtained from Datastream. Descriptive statistics are provided in Table 1.
Table 1

Descriptive statistics

 

Mean

Median

Max

Min

Std. dev.

ΔPound/Euro

0.0001

0.0000

0.0622

−0.0198

0.0055

ΔPound/USD

0.0001

0.0000

0.0840

−0.0298

0.0058

ΔUK OIS rate (3 M)

0.0001

0.0000

0.1079

−0.0884

0.0059

ΔEA OIS rate (3 M)

−0.0003

0.0000

0.0600

−0.0620

0.0065

ΔUS OIS rate (3 M)

0.0012

0.0010

0.0370

−0.0560

0.0058

ΔUK benchmark bond yield (10y)

−0.0003

−0.0008

0.1776

−0.2855

0.0460

ΔEA benchmark bond yield (10y)

−0.0013

−0.0032

0.2004

−0.1243

0.0342

ΔUS benchmark bond yield (10y)

0.0007

−0.0002

0.2155

−0.1603

0.0433

ΔFTSE 100

0.0001

0.0002

0.0352

−0.0478

0.0084

ΔEurostoxx 50

0.0002

0.0002

0.0460

−0.0901

0.0113

ΔS&P 500

0.0005

0.0003

0.0383

−0.0418

0.0077

ΔUK CESI

−0.0251

−0.1000

31.4000

−21.1000

4.7716

ΔEA CESI

−0.0363

−0.2000

36.5000

−33.2000

4.7368

ΔUS CESI

0.0072

0.0000

23.5000

−27.7000

4.3689

UK monetary surprise

0.0061

0.0000

6.4345

−4.3746

1.0000

EA monetary surprise

0.0188

0.0559

4.3044

−8.3665

1.0000

US monetary surprise

−0.0228

0.0000

4.1906

−4.7774

1.0000

Concerning the identification of events associated with Brexit, we also follow the identification method of Fratzscher et al. (2014). The overall aim of this identification method is to identify only truly relevant event days. First, we look at the Brexit event list published by the British parliament (Walker 2018). From this event list, we only take Brexit-related events which are defined by Walker (2018) as events leading to the UK’s exit from the European Union. Second, we determine if these events were covered by the Financial Times on the first three pages of that paper on the next day. Third, we include an event in our list of event days, if this event can be unambiguously identified as either good news or bad news. By reading the Financial Times thoroughly, we are able to distinguish between bad and good news associated with the Brexit process. In order to classify Brexit news into good news and bad news, we search for signalling words.1 If an event fulfils all these restrictions, it will be included in our list of event days. With our identification strategy of Brexit-related events, we find a total of 17 event dates as illustrated in Table 2. Moreover, we find a total of nine bad Brexit event days and eight good Brexit event days.2
Table 2

Event Days from Walker (2018) and Financial Times Headlines

Date

Event

Type of Event

23.01.2013

Announcement of an in-out referendum

bad

22.02.2016

The Prime Minister announces the EU referendum date

bad

23.06.2016

UK holds referendum on its membership of the EU

bad

13.07.2016

Theresa May becomes the new UK Prime Minister

bad

02.10.2016

Article 50 will be triggered before the end of March 2017

bad

03.11.2016

High court rules UK parliament must have a say

good

17.01.2017

UK Prime Minister gives her Lancaster House speech

good

24.01.2017

Supreme Court rejects the UK Government’s appeal of the Gina Miller case

good

29.03.2017

UK Prime Minister triggers Article 50 of the Treaty on European Union

good

30.03.2017

UK Government publishes the Great Repeal Bill White Paper

bad

18.04.2017

UK Prime Minister calls a General Election

good

08.06.2017

UK General Election results

bad

19.06.2017

First round of UK-EU exit negotiations begin

bad

22.09.2017

UK Prime Minister delivers her key Brexit speech in Florence

good

20.10.2017

European Council meeting to assess progress on the first phase of Brexit negotiations

good

08.12.2017

Joint Report concludes Phase 1 of negotiations and both sides move to Phase 2

good

02.03.2018

Prime Minister gives a speech at Mansion House

bad

Walker (2018) and Financial Times, pp.1–3

Figure 1 shows the behaviour of the spot exchange rate of the British pound from January 2016 to March 2018. The British pound depreciated heavily against the euro following the announcement of the referendum result. On 23 June, 2016, the British pound depreciated against the euro by more than 6 % (from 0.77 to 0.82). The British pound also lost against the US dollar and the Yen one day after the referendum. On the day of the announcement of the referendum result, the British pound depreciated against the US dollar by 8 % and against the Yen by nearly 12%. The depreciation of the British pound against the euro continued until November 2016. During that time, two additional bad Brexit events led to a deprecation of the British pound against the euro. The election of Theresa May as Prime Minister on 13 July was perceived as bad Brexit news by market participants because on that day Theresa May stated that there will be no second referendum. This statement shattered the hopes of the market participants that it might not come to a Brexit and created significant medium-term uncertainty. The Brexit-related event on 2 October, 2016, was perceived as bad Brexit news by financial markets because the Brexit plans set out by Theresa May on that day increased, according to the opinion of the financial markets, the probability of a hard Brexit. This protracted depreciation of the British pound vis-à-vis the major currencies may have been beneficial for UK exports but it was also a sign of higher political uncertainty due to Brexit (Baker et al. 2016).
Fig. 1

Spot Exchange Rate of the British Pound against the Euro (British Pounds Per Unit of Foreign Currency). Source: Datastream and Walker (2018). Notes: The black line reflects the spot exchange rate of the Pound against the euro and the vertical red lines indicate Brexit-related events

From November 2016 until the end of 2016, the British pound appreciated against the euro remarkably after the High Court in the UK ruled that only parliament had the authority to trigger Article 50 of the Treaty on the EU (which would begin the official withdrawal negotiation period) on 3 November 2016. We suspect that at this time market participants expected that Brexit would have less severe implications for the economy of the UK. The spot exchange rate of the British pound against the euro stagnated between January 2017 and May 2017. However, on 29 March 2017 the British pound reached a seven-week high against the euro. On that day, Article 50 was triggered by Prime Minister May with parliamentary consent. Hence, the actual triggering of the Article 50 was read as positive news whereas the announcement that the Article 50 will be triggered was classified, as mentioned before, as bad Brexit news by the markets. Afterwards the British pound depreciated against the euro until September 2017 mainly due to two Brexit-related events which encouraged investors to expect that Brexit may lead to severe welfare losses in the UK. From September 2017 to October 2017 the British pound appreciated against the euro again. Two Brexit-related events seem to be responsible for the appreciation of the British pound against the euro. Firstly, the key Brexit speech given on 22 September, 2017, by Prime Minister Theresa May has led to a perception amongst market participants that the UK government is seeking to move forward and to avoid a hard Brexit. Secondly, during the European Council meeting on 20 October, 2017, the EU exploited the political weakness of the UK’s prime minister in order to avoid a cliff-edged Brexit. From November 2017 to March 2018 the spot exchange rate of the British pound against the euro was relatively stable. The British pound appreciated against the euro when a deal on the UK’s exit terms was achieved. The overall market opinion was that the divorce agreement was a signal that the UK government was indeed on course for a soft Brexit. The speech of the Prime Minister May on 2 March, 2018, is classified as bad Brexit news mainly because of the Prime Minister’s statement that the level of access to EU markets will be less than currently. Less access to EU markets should be associated with lower expected GDP growth rates in the UK and hence with a depreciation of the British pound.

Overall, it seems that the spot exchange rate of the British pound against the euro is driven by Brexit-related events. Moreover, descriptive statistics suggest that on the one hand there are Brexit events which lead to a depreciation of the British pound against the euro while, on the other hand, other Brexit-related events are associated with an appreciation of the British pound against the euro.

3 The impact of Brexit on foreign exchange returns

3.1 Event study methodology

Our event study analysis, through which we try to detect the impact of Brexit-related events on the British pound/euro exchange rate, uses the ordinary least squares estimation. We use Newey-West adjusted standard errors because we have detected autocorrelation by conducting Breusch Pagan LM tests. Firstly, we estimate the following equation using daily data from 1 January, 2013, to 30 March, 2018:
$$ \varDelta {E}_t=\alpha +\gamma Brexi{t}_t+{\beta}_1\varDelta OI{S_t}^{UK}+{\beta}_2\varDelta OI{S_t}^{EA}+{\beta}_3\varDelta Bon{d_t}^{UK}+{\beta}_4\varDelta Bon{d_t}^{EA}+{\beta}_5\varDelta FTS{E}_t+{\beta}_6\varDelta Stox{x}_t+{\varepsilon}_t, $$
(1)
where ΔEt is our dependent variable of interest and represents the change of the exchange rate of the British pound vis-à-vis the euro. Brexitt is our independent variable of interest. This vector is an event-dummy which captures our identified Brexit events. The dummy variable takes the value of one on all days identified as having had Brexit-related events according to our event identification methodology. As previously mentioned, we have identified 17 events related to the Brexit process. In principle, Brexit news should be associated with lower future growth expectations and thus with a depreciation of the Pound against the euro. Gourinchas and Hale (2017) argue that Brexit could be considered as a de-globalization shock for the UK economy. Brexit could lead to increasing barriers to trade in goods and services, to labour and to capital mobility. Hence, it can be expected that the economy of the United Kingdom will become less specialized and less efficient. Thus, Brexit could be associated with lower aggregate productivity and hence lower than expected GDP growth rates in the future. According to the monetary exchange rate models, lower growth expectations in the future are associated with a depreciation of the British pound against the euro today. Moreover, Brexit may trigger a more aggressive expansionary monetary policy by the Bank of England. Hence, interest rate cuts and additional non-standard monetary policy measures conducted by the Bank of England may further depreciate the British pound against the euro (Gourinchas and Hale 2017).

However we expect that not all Brexit-related news in our event list are associated with a depreciation of the British pound against the euro. There may be some items of Brexit news which lead to an appreciation of the British pound vis-à-vis the euro because this news might implicate that Brexit will not occur or that its impact on the economy would not be as severe as predicted. Therefore, we divide our events into good news and bad news in several sub-sample analyses. Good news items are defined as news which are associated with less severe implications for the economy of the United Kingdom and hence with an appreciation of the British pound against the euro. Bad news items are defined as events which impose a drag on the UK economy and thus we expect that bad Brexit news will lead to a depreciation of the British pound vis-à-vis the euro. We split up our Brexit events into good Brexit news and bad Brexit news by reading the Financial Times thoroughly. By reading the Brexit-related articles published by the Financial Times we are able to detect whether market participants assessed the associated Brexit event as being either good or bad for the UK economy. We have identified nine bad Brexit news events and eight good Brexit news events.

The variable OIStUK reflects the three-month OIS rate in the UK and controls for the impact of the Bank of England’s conventional monetary policies on the British pound/euro exchange rate. OIStEA stands for the three-month OIS rate in the Eurozone and displays the effect of conventional monetary policies conducted by the ECB on our dependent variable of interest. The variables BondtUK and BondtEA are the 10-year government bond yields in the UK and the Eurozone, respectively. We control for market participants’ expectations about future economic developments in the UK and the Eurozone by including long-term interest rates into our baseline specification. The variable FTSEt is the FTSE 100 index and the variable Stoxxt reflects the Euro Stoxx 50 index. By including benchmark stock market indices, we control for the impact of market risk or financial market uncertainty on the British pound/euro exchange rate. Additionally, stock market indices also reflect expectations about future economic growth rates. In other specifications of our empirical model, we additionally control for the influence of macroeconomic news on the spot exchange rate. We use the United Kingdom series and the Eurozone series of the Citigroup economic surprise index (CESI) to control for the impact of news stemming from macroeconomic data. Moreover, we control for the impact of the BoEs (NSMPBOE), ECBs (NSMPECB) and FEDs (NSMPFED) non-standard monetary policy announcements on the British pound/euro exchange rate. We introduce event-dummies in order to capture announcements of unconventional monetary policies and interact them with a market-based measure of monetary policy surprise. The unconventional monetary policy dummies take the value of one on the day of an announcement and zero elsewhere.

We also estimate the impact of Brexit-related news on the spot exchange rate of the British pound against the US dollar. For this estimation, the three-month US OIS rate, the yield on 10-year US government bonds, the S&P 500 and the US series of the Citigroup economic surprise index are used as controls. However, in this study we focus on the impact of Brexit-related news on the British pound/euro exchange rate because trade linkages between the UK and the Eurozone are stronger than trade ties between the UK and the US.

The variables used in our regressions are defined as follows: Spot exchange rates of the British pound and the stock market indices are expressed in percentage rates of change. The three-month OIS rates, the 10-year government bond yields and the Citigroup economic surprise indices are expressed as percentage point changes. For the event dummies, we use one-day event windows.

3.2 Results

3.2.1 Spot exchange rate of the British pound against the euro

First, we investigate whether our Brexit event-dummy, including all Brexit-related events, has a significant impact on the spot exchange rate of the British pound against the euro. We expect that the coefficient of our event-dummy shows a positive sign, indicating that Brexit news is associated with a depreciation of the British pound vis-à-vis the euro. Our findings, using the HAC estimator, are presented in Table 3. Several specifications are developed, all of which include the three-month OIS rates, the 10-year benchmark bond yields and the benchmark stock market indices. In model (2) we additionally control for the influence of macroeconomic news on our dependent variable of interest. In specification (3), we control for the impact of unconventional monetary policies on the spot exchange rate. In model (4), both the macroeconomic news variables and the unconventional monetary policy event-dummies are included.
Table 3

Impact of Brexit Events on the Spot Exchange Rate of the British Pound against the Euro (1-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0001

0.0001

0.0001

0.0001

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit news (all events)

0.0003

0.0007

0.0004

0.0008

(0.0019)

(0.0018)

(0.0011)

(0.0018)

UK OIS rate (3 m)

−0.2032***

−0.1990***

−0.1991***

−0.1936***

(0.0665)

(0.0678)

(0.0226)

(0.0715)

EA OIS rate (3 m)

0.0878***

0.0912***

0.0747***

0.0784***

(0.0223)

(0.0224)

(0.0196)

(0.0197)

UK benchmark bond yield (10y)

−0.0233***

−0.0202***

−0.0225***

−0.0192***

(0.0061)

(0.0056)

(0.0041)

(0.0056)

EA benchmark bond yield (10y)

0.0192**

0.0159**

0.0158***

0.0126*

(0.0077)

(0.0071)

(0.0051)

(0.0074)

FTSE 100

0.3075***

0.3065***

0.3088***

0.3073***

(0.0385)

(0.0349)

(0.0268)

(0.0347)

Eurostoxx 50

−0.2973***

−0.3023***

−0.2971***

−0.3020***

(0.0330)

(0.0295)

(0.0214)

(0.0294)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

EA CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0008

0.0004

  

(0.0005)

(0.0008)

NSMPECB*bund

  

−0.0018***

−0.0018***

  

(0.0005)

(0.0006)

NSMPFED*tnote

  

−0.0009

−0.0007

  

(0.0006)

(0.0005)

No. obs.

1368

1368

1368

1368

R-squared

0.29

0.31

0.30

0.32

Adjusted R-squared

0.28

0.30

0.29

0.31

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Our results, presented in Table 3, show that the Brexit dummy, including all events, has no statistically significant impact on the British pound/euro exchange rate in any of the four specifications. Hence, we suspect that our Brexit event-dummy contains both bad news and good news. As mentioned previously, we expect that bad news is associated with a depreciation of the British pound against the euro, while good news should lead to an appreciation of the British pound vis-à-vis the euro.

In a next step, we split our Brexit event-dummy into bad news and good news. Hence, we create an event-dummy, which captures bad Brexit news, and a dummy variable which consists of good Brexit news. We suspect, based on the results shown in Table 3 and our considerations in Section 2, that positively and negatively surprising Brexit news differ in sign. It could also be the case that market participants’ reactions to Brexit events differ in size between bad Brexit news and good Brexit news.

The results reported in Table 4 show that our Brexit bad news event-dummy and Brexit good news event-dummy both have a statistically significant impact on the British pound/euro exchange rate.3 The Brexit bad news event-dummy is statistically significant at the 5 % level in every specification. Moreover, we find that bad news is associated with a depreciation of the British pound against the euro. The four models show a positive coefficient sign of the Brexit bad news dummy variable. Furthermore, our results indicate that the coefficient size of the Brexit bad news event-dummy variable differs between our four specifications.
Table 4

Impact of Bad and Good Brexit News on the Spot Exchange Rate of the British Pound against the Euro (1-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0001

0.0001

0.0001

0.0001

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit bad news

0.0048**

0.0056**

0.0049**

0.0057**

(0.0021)

(0.0022)

(0.0023)

(0.0022)

Brexit good news

−0.0045***

−0.0047***

−0.0044***

−0.0046***

(0.0016)

(0.0015)

(0.0017)

(0.0014)

UK OIS rate (3 m)

−0.1972***

−0.1921***

−0.1929***

−0.1864***

(0.0552)

(0.0615)

(0.0666)

(0.0650)

EA OIS rate (3 m)

0.0877***

0.0913***

0.0746***

0.0786***

(0.0221)

(0.0222)

(0.0196)

(0.0193)

UK benchmark bond yield (10y)

−0.0228***

−0.0198***

−0.0220***

−0.0186***

(0.0063)

(0.0056)

(0.0057)

(0.0056)

EA benchmark bond yield (10y)

0.0192**

0.0156**

0.0158**

0.0124*

(0.0079)

(0.0073)

(0.0076)

(0.0074)

FTSE 100

0.2993***

0.2977***

0.3006***

0.2984***

(0.0393)

(0.0341)

(0.0366)

(0.0348)

Eurostoxx 50

−0.2912***

−0.2958***

−0.2909***

−0.2956***

(0.0330)

(0.0285)

(0.0297)

(0.0288)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

EA CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0008

0.0010

  

(0.0008)

(0.0008)

NSMPECB*bund

  

−0.0018***

−0.0018***

  

(0.0007)

(0.0007)

NSMPFED*tnote

  

−0.0008

−0.0007

  

(0.0005)

(0.0005)

No. obs.

1368

1368

1368

1368

R-squared

0.30

0.32

0.31

0.33

Adjusted R-squared

0.29

0.31

0.30

0.32

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

All four models reported in Table 4 indicate that the Brexit good news event-dummy variable is highly statistically significant at the 1 % level. Hence, the statistical significance does not differ between our four baseline specifications. The coefficient sign of the Brexit good news dummy variable is continuously negative. Therefore, good Brexit news is associated with an appreciation of the British pound vis-à-vis the euro. Moreover, our results show that the coefficient size of the Brexit good news event-dummy variable differs only slightly between our four models. Overall, the empirical results shown in Table 4 confirm our presumption that Brexit news has to be split into bad Brexit news and good Brexit news as negatively and positively surprising Brexit-related news differ in sign. Furthermore, the coefficient signs of our event-dummies show the expected signs, namely, that bad news is associated with an increase of the British pound/euro exchange rate and that good news leads to an appreciation of the British pound against the euro.

Moreover, we find that the impact of bad news and good news on our dependent variable of interest seems to differ in size. In all specifications reported in Table 4, the coefficient of the Brexit bad news event-dummy variable is larger in absolute values than the coefficient of the Brexit good news dummy variable. However, by applying Wald coefficient tests, we find that the impact of bad Brexit-related events and good Brexit news on the spot exchange rate of the British pound against the euro doesn’t differ in magnitude. Hence, we find that the effect of bad news and good news on the British pound/euro exchange rate only differs in sign.

Furthermore, we additionally check for delayed market reactions by lagging our Brexit-related event dummies (see Tables 5 and 6). There are at least two reasons why market participants may respond in a delayed manner to Brexit-related news. First, it could be the case that market participants need a prolonged time to fully process Brexit-related news. Second, Brexit-related news could create uncertainty about the correct pricing of the British pound against other major currencies. Firstly, we want to check whether the British pound/euro exchange rate has a delayed reaction to bad Brexit news. We therefore additionally introduce a one-day, two-day, and three-day lagged Brexit bad news event dummy. Hence, we lag our Brexit bad news event dummy by one day, two and three days. Table 5 shows the impact of our lagged Brexit bad news event dummies on our dependent variable of interest.
Table 5

Impact of Bad Brexit News on the Spot Exchange Rate of the British Pound against the Euro (4-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0001

0.0001

0.0000

0.0000

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit bad news t

0.0048**

0.0055**

0.0049**

0.0056**

(0.0021)

(0.0022)

(0.0023)

(0.0022)

Brexit bad news t-1

0.0042*

0.0041**

0.0043**

0.0043**

(0.0029)

(0.0020)

(0.0022)

(0.0021)

Brexit bad news t-2

0.0027**

0.0028**

0.0028**

0.0028**

(0.0014)

(0.0011)

(0.0013)

(0.0012)

Brexit bad news t-3

−0.0011

−0.0008

−0.0011

−0.0008

(0.0015)

(0.0013)

(0.0014)

(0.0013)

UK OIS rate (3 m)

−0.2032***

−0.1985***

−0.1988***

−0.1927***

(0.0511)

(0.0544)

(0.0590)

(0.0576)

EA OIS rate (3 m)

0.0861***

0.0895***

0.0727***

0.0765***

(0.0222)

(0.0225)

(0.0199)

(0.0197)

UK benchmark bond yield (10y)

−0.0230***

−0.0199***

−0.0221***

−0.0187***

(0.0062)

(0.0055)

(0.0057)

(0.0055)

EA benchmark bond yield (10y)

0.0190**

0.0155**

0.0155**

0.0122

(0.0080)

(0.0073)

(0.0077)

(0.0075)

FTSE 100

0.3013***

0.2993***

0.3023***

0.2998***

(0.0391)

(0.0345)

(0.0369)

(0.0349)

Eurostoxx 50

−0.2922***

−0.2967***

−0.2918***

−0.2963***

(0.0329)

(0.0286)

(0.0300)

(0.0289)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

EA CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0009

0.0011

  

(0.0008)

(0.0008)

NSMPECB*bund

  

−0.0019***

−0.0018***

  

(0.0007)

(0.0007)

NSMPFED*tnote

  

−0.0009

−0.0007

  

(0.0005)

(0.0005)

No. obs.

1366

1366

1366

1366

R-squared

0.30

0.32

0.31

0.33

Adjusted R-squared

0.29

0.31

0.30

0.32

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Table 6

Impact of Good Brexit News on the Spot Exchange Rate of the British Pound against the Euro (4-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0002

0.0002

0.0002

0.0002

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit good news t

−0.0046***

−0.0047***

−0.0045***

−0.0046***

(0.0017)

(0.0016)

(0.0019)

(0.0015)

Brexit good news t-1

−0.0024**

−0.0018

−0.0024*

−0.0017

(0.0010)

(0.0013)

(0.0012)

(0.0013)

Brexit good news t-2

0.0002

0.0004

0.0003

0.0005

(0.0018)

(0.0016)

(0.0017)

(0.0016)

Brexit good news t-3

0.0001

−0.0005

0.0001

−0.0004

(0.0014)

(0.0012)

(0.0012)

(0.0012)

UK OIS rate (3 m)

−0.2024***

−0.1987***

−0.1987***

−0.1938***

(0.0648)

(0.0674)

(0.0713)

(0.0724)

EA OIS rate (3 m)

0.0876***

0.0908***

0.0747***

0.0781***

(0.0224)

(0.0226)

(0.0206)

(0.0201)

UK benchmark bond yield (10y)

−0.0243***

−0.0213***

−0.0236***

−0.0203***

(0.0063)

(0.0056)

(0.0056)

(0.0057)

EA benchmark bond yield (10y)

0.0203***

0.0170**

0.0170**

0.0137*

(0.0079)

(0.0071)

(0.0072)

(0.0073)

FTSE 100

0.3014***

0.3008***

0.3030***

0.3019***

(0.0382)

(0.0348)

(0.0348)

(0.0339)

Eurostoxx 50

−0.2937***

−0.2992***

−0.2936***

−0.2991***

(0.0334)

(0.0298)

(0.0294)

(0.0293)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

EA CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0007

0.0009

  

(0.0008)

(0.0008)

NSMPECB*bund

  

−0.0018***

−0.0018***

  

(0.0007)

(0.0006)

NSMPFED*tnote

  

−0.0009

−0.0007

  

(0.0006)

(0.0005)

No. obs

1366

1366

1366

1366

R-squared

0.29

0.31

0.31

0.32

Adjusted R-squared

0.29

0.31

0.30

0.32

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

The results suggest that the British pound/euro exchange rate shows a delayed response to bad Brexit news up to two days after the event. In all four models reported in Table 5, the one-day lagged Brexit bad news dummy variable has a statistically significant impact at the 5 % level on the British pound/euro exchange rate. Bad Brexit news released at time t-1 is associated with a depreciation of the British pound against the euro at time t. The coefficient size of the one-day lagged bad Brexit dummy is almost identical to the coefficient size of the bad Brexit event-dummy variable at time t.

Moreover, we find that the two-day lagged Brexit bad news event-dummy has a statistically significant impact on our dependent variable of interest. Our results suggest that bad Brexit news is associated with a depreciation of the British pound against the euro two days after its release. However, the coefficient size of the two-day lagged bad Brexit event-dummy is clearly smaller than the coefficient size of the one-day lagged dummy variable and the event-dummy at time t. Moreover, we find that the three-day lagged Brexit bad news event-dummy has no significant impact on the British pound/euro exchange rate.

Secondly, we investigate whether the British pound/euro exchange rate responds in a delayed manner to good Brexit news (see Table 6). Therefore, we lag our Brexit good news event-dummy by one, two or three days. The results reported in Table 6 indicate that the statistical significance concerning the delayed reaction of the British pound/euro exchange rate to good Brexit news is mixed. In models (1) and (3) the Brexit good news event-dummy lagged by one day has a statistically significant impact on the British pound/euro exchange rate, whereas in model (2) and model (4) the one-day lagged event-dummy becomes insignificant. Hence, the inclusion of our macroeconomic news variables makes the impact of the Brexit good news dummy lagged by one-day on our dependent variable of interest insignificant.

The coefficient sign of the one-day lagged Brexit good news event dummy indicates that good Brexit news released at time t-1 is associated with an appreciation of the British pound/euro exchange rate. Moreover, we find that good Brexit news released at times t-2 and t-3 have no statistically significant impact on the British pound/euro exchange rate.

3.2.2 Spot exchange rate of the British pound against the US Dollar

In this section, we study whether Brexit-related events affect the spot exchange rate of the British pound against the US dollar. Firstly, we investigate whether our Brexit event-dummy, including all Brexit-related events, has a statistically significant impact on the British pound/US dollar exchange rate. The results reported in Table 7 indicate that the Brexit dummy, including all Brexit-related events, has no statistically significant impact on the spot exchange rate of the British pound against the US dollar. The coefficient of our Brexit event dummy is not statistically significant in any of the four specifications. Hence, our Brexit dummy, including all events, does not affect the British pound/euro exchange rate and the British pound/US dollar exchange rate, respectively.
Table 7

Impact of Brexit Events on the Spot Exchange Rate of the British Pound against the US Dollar (1-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0001

0.0001

0.0001

0.0001

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit (all events)

0.0029

0.0031

0.0030

0.0031

(0.0037)

(0.0037)

(0.0037)

(0.0037)

UK OIS rate (3 m

−0.2785**

−0.2791**

−0.2799**

−0.2800**

(0.1095)

(0.1099)

(0.1120)

(0.1133)

US OIS rate (3 m)

0.0015

−0.0001

0.0021

−0.0008

(0.0357)

(0.0359)

(0.0359)

(0.0364)

UK benchmark bond yield (10y)

−0.0365***

−0.0345***

−0.0354***

−0.0328***

(0.0048)

(0.0049)

(0.0051)

(0.0052)

US benchmark bond yield (10y)

0.0387***

0.0361***

0.0370***

0.0338***

(0.0052)

(0.0053)

(0.0055)

(0.0056)

FTSE 100

0.1479***

0.1407***

0.1456***

0.1374***

(0.0356)

(0.0345)

(0.0354)

(0.0347)

S&P 500

−0.1998***

−0.1936***

−0.1969***

−0.1892***

(0.0355)

(0.0352)

(0.0359)

0.035721

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

US CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0001

0.0002

  

(0.0010)

(0.0010)

NSMPFED*tnote

  

−0.0010**

−0.0013***

  

(0.0004)

(0.0004)

NSMPECB*bund

  

0.0007

0.0007

  

(0.0005)

(0.0005)

No. obs.

1368

1368

1368

1368

R-squared

0.22

0.23

0.22

0.23

Adjusted R-squared

0.21

0.22

0.21

0.23

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Next, we want to find out whether our Brexit bad news event-dummy and Brexit good news event-dummy both have a statistically significant impact on the British pound/US dollar exchange rate. We suspect that bad Brexit news is associated with a depreciation of the British pound against the US dollar and that good Brexit news leads to an appreciation of the British pound vis-à-vis the US dollar. The results shown in Table 8 indicate that bad Brexit news has indeed affected the spot exchange rate of the British pound against the US dollar.
Table 8

Impact of Bad and Good Brexit News on the Spot Exchange Rate of the British Pound against the US dollar (1-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0001

0.0001

0.0001

0.0001

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit bad news

0.0101**

0.0102**

0.0101**

0.0103**

(0.0050)

(0.0050)

(0.0049)

(0.0050)

Brexit good news

−0.0055

−0.0053

−0.0055

−0.0053

(0.0034)

(0.0035)

(0.0035)

(0.0035)

UK OIS rate (3 m)

−0.2671***

−0.2676***

−0.2683***

−0.2684***

(0.0887)

(0.0977)

(0.0959)

(0.1007)

US OIS rate (3 m)

0.0034

0.0010

0.0040

0.0012

(0.0331)

(0.0341)

(0.0341)

(0.0346)

UK benchmark bond yield (10y)

−0.0353***

−0.0332***

−0.0341***

−0.0315***

(0.0049)

(0.0047)

(0.0049)

(0.0049)

US benchmark bond yield (10y)

0.0379***

0.0353***

0.0362***

0.0330***

(0.0053)

(0.0052)

(0.0054)

(0.0055)

FTSE 100

0.1435***

0.1361***

0.1411***

0.1329***

(0.0384)

(0.0345)

(0.0368)

(0.0346)

S&P 500

−0.1957***

−0.1894***

−0.1926***

−0.1849***

(0.0341)

(0.0337)

(0.0344)

(0.0341)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

US CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0001

0.0002

  

(0.0010)

(0.0010)

NSMPFED*tnote

  

−0.0010**

−0.0013***

  

(0.0004)

(0.0004)

NSMPECB*bund

  

0.0008

0.0007

  

(0.0005)

(0.0005)

No. obs.

1368

1368

1368

1368

R-squared

0.24

0.25

0.24

0.25

Adjusted R-squared

0.23

0.24

0.24

0.25

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

The Brexit bad news event-dummy is statistically significant at the 5 % level in all four specifications.4 The coefficient of the Brexit bad news event-dummy shows the expected sign. Thus, bad Brexit news is associated with an increase of the British pound/US dollar exchange rate. The sign of the coefficient of the Brexit good news dummy variable indicates that good Brexit news is associated with an appreciation of the British pound against the US dollar. However, the Brexit good news dummy is not statistically significant in any specification reported in Table 8. Hence, our empirical results suggest that good Brexit news has an impact on the spot exchange rate of the British pound against the euro but not an effect on the British pound/US dollar exchange rate.

Next, we check whether market participant’s show delayed reactions to bad Brexit news and good Brexit news, respectively. As before, we use a four-day event window in order to control for delayed market reactions. First, we want to find out whether the British pound/US dollar exchange rate displays a delayed response to bad Brexit news. Table 9 reports the results of the event study controlling for delayed market reactions to bad Brexit news. The results indicate that the British pound/US dollar exchange rate shows a delayed reaction to bad Brexit news.
Table 9

Impact of Bad Brexit News on the Spot Exchange Rate of the British Pound against the US Dollar (4-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0001

0.0001

0.0001

0.0000

(0.0001)

(0.0001)

(0.0001)

(0.0001)

Brexit bad news t

0.0101**

0.0102**

0.0101**

0.0103**

(0.0049)

(0.0050)

(0.0049)

(0.0050)

Brexit bad news t-1

0.0069**

0.0071**

0.0070**

0.0072*

(0.0030)

(0.0028)

(0.0029)

(0.0028)

Brexit bad news t-2

−0.0009

−0.0009

−0.0009

−0.0009

(0.0015)

(0.0015)

(0.0015)

(0.0015)

Brexit bad news t-3

−0.0002

−0.0001

−0.0001

0.0001

(0.0014)

(0.0013)

(0.0014)

(0.0013)

UK OIS rate (3 m)

−0.2784***

−0.2794***

−0.2791***

−0.2796***

(0.0794)

(0.0870)

(0.0855)

(0.0897)

US OIS rate (3 m)

0.0035

0.0009

0.0037

0.0005

(0.0338)

(0.0349)

(0.0351)

(0.0355)

UK benchmark bond yield (10y)

−0.0353***

−0.0332***

−0.0339***

−0.0313***

(0.0047)

(0.0045)

(0.0047)

(0.0046)

US benchmark bond yield (10y)

0.0378***

0.0352***

0.0363***

0.0330***

(0.0054)

(0.0052)

(0.0055)

(0.0055)

FTSE 100

0.1451***

0.1371***

0.1425***

0.1337***

(0.0377)

(0.0333)

(0.0363)

(0.0336)

S&P 500

−0.1949***

−0.1885***

−0.1922***

−0.1843***

(0.0340)

(0.0336)

(0.0343)

(0.0340)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

US CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

0.0002

0.0004

  

(0.0010)

(0.0010)

NSMPFED*tnote

  

−0.0009**

−0.0012***

  

(0.0004)

(0.0004)

NSMPECB*bund

  

0.0007

0.0007

  

(0.0005)

(0.0005)

No. obs.

1366

1366

1366

1366

R-squared

0.24

0.26

0.25

0.26

Adjusted R-squared

0.24

0.25

0.24

0.25

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

The results indicate that market participants exhibited a delayed reaction to bad Brexit news up to one day after the event. In all four specifications reported in Table 9, the bad Brexit event-dummy has a statistically significant impact on the spot exchange rate of the British pound against the US dollar. Bad Brexit news released at time t-1 is associated with a depreciation of the British pound against the US dollar. Note that the coefficient size of the one-day lagged bad Brexit event-dummy is visibly smaller than the coefficient size of the bad Brexit event-dummy variable at time t. Moreover, we find that the two-day lagged bad Brexit dummy variable does not affect our dependent variable of interest. Hence, we find that the effect of bad Brexit news on the British pound/US dollar exchange rate fades away one day after the event has happened. Whereas the spot exchange rate of the British pound against the euro displays a delayed reaction to bad Brexit news up to two days after the event. Furthermore, we find no statistically significant impact of the three-day lagged bad Brexit dummy variable on our dependent variable of interest.

Secondly, we investigate whether the market participants displayed a delayed reaction to good Brexit news. The results reported in Table 10 suggest that the spot exchange rate of the British pound against the US dollar do not react in a delayed manner to good Brexit events. The one-, two-, and three-day lagged good Brexit event dummies have no statistically significant impact on the British pound/US dollar exchange rate, respectively. Hence, in the case of the British Pound/US dollar exchange rate, our empirical results indicate that market participants displayed only a delayed market reaction to bad Brexit news. Thus, it seems that only bad Brexit news has affected the spot exchange rate of the British pound against the US dollar. So, our results suggest that market participants perceive that bad Brexit news is a good compass for the future economic relationship between the United Kingdom and the Eurozone and between the UK and the US, respectively. Furthermore, we conclude that our selected good Brexit events seem to be only important for the future economic relationship between the UK and the Eurozone.
Table 10

Impact of Good Brexit News on the Spot Exchange Rate of the British Pound against the US Dollar (4-day Window)

VARIABLES

(1)

(2)

(3)

(4)

Constant

0.0002

0.0002

0.0002

0.0002

(0.0002)

(0.0001)

(0.0002)

(0.0001)

Brexit good news t

−0.0055

−0.0053

−0.0055

−0.0053

(0.0039)

(0.0036)

(0.0039)

(0.0035)

Brexit good news t-1

0.0007

0.0010

0.0008

0.0010

(0.0020)

(0.0019)

(0.0020)

(0.0018)

Brexit good news t-2

0.0013

0.0013

0.0013

0.0013

(0.0022)

(0.0021)

(0.0021)

(0.0021)

Brexit good news t-3

−0.0002

−0.0007

−0.0001

−0.0006

(0.0014)

(0.0015)

(0.0013)

(0.0014)

UK OIS rate (3 m)

−0.2812**

−0.2819**

−0.2832**

−0.2835***

(0.1152)

(0.1157)

(0.1211)

(0.1185)

US OIS rate (3 m)

0.0027

−0.0001

0.0032

0.0000

(0.0380)

(0.0368)

(0.0362)

(0.0370)

UK benchmark bond yield (10y)

−0.0374***

−0.0354***

−0.0364***

−0.0340***

(0.0049)

(0.0050)

(0.0052)

(0.0054)

US benchmark bond yield (10y)

0.0386***

0.0361***

0.0372***

0.0341***

(0.0053)

(0.0052)

(0.0054)

(0.0055)

FTSE 100

0.1436***

0.1362***

0.1417***

0.133624

(0.0283)

(0.0326)

(0.0299)

(0.0334)

S&P 500

−0.1982***

−0.1923***

−0.1961***

−0.1888***

(0.0330)

(0.0350)

(0.0351)

(0.0358)

UK CESI

 

−0.0001***

 

−0.0001***

 

(0.0000)

 

(0.0000)

US CESI

 

0.0001***

 

0.0001***

 

(0.0000)

 

(0.0000)

NSMPBOE*lgilt

  

−0.0001

0.0001

  

(0.0011)

(0.0011)

NSMPFED*tnote

  

−0.0009**

−0.0011***

  

(0.0004)

(0.0004)

NSMPECB*bund

  

0.0007

0.0007

  

(0.0005)

(0.0005)

No. obs.

1366

1366

1366

1366

R-squared

0.22

0.23

0.22

0.24

Adjusted R-squared

0.22

0.22

0.22

0.23

Newey-West standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

4 The impact of Brexit on volatilities of foreign exchange returns

4.1 Econometric model

Considering that the market participants react on the Brexit news in the mean, we also suppose an impact of Brexit on the volatility of exchange returns. To check this presumption and its strength, we conduct an EGARCH(p,q,s) model as follows, also using daily data from 1 January, 2013, to 30 March, 2018 (Nelson 1991):
$$ {\Delta E}_t={\beta}_0+{\varepsilon}_t\kern0.5em \mathrm{with}\kern0.5em {\varepsilon}_t\sim \mathrm{N}\left(0,{\sigma}_t^2\right) $$
(2)
$$ \ln \left({\sigma}_t^2\right)={\alpha}_0+{\sum}_{i=1}^p{\alpha}_i\left|\frac{\varepsilon_{t-i}}{\sigma_{t-i}}\right|+{\sum}_{j=1}^s{\theta}_j\frac{\varepsilon_{t-j}}{\sigma_{t-j}}+{\sum}_{k=1}^q{\delta}_k{\sigma}_{t-k}^2+\gamma {Referendum}_t+\omega {Brexit}_t, $$
(3)

where on the left-hand side of the mean equation (2) ∆Et represents the percentage change of the exchange rate of the British pound vis-à-vis the euro and US dollar, respectively. On the left-hand side of the variance equation (3) \( \ln \left({\sigma}_t^2\right) \) is the logarithm of the conditional variance of the exchange rate returns and is modelled by the symmetric ARCH \( \left[{\sum}_{i=1}^p{\alpha}_i\left|\frac{\varepsilon_{t-i}}{\sigma_{t-i}}\right|\right] \), the asymmetric ARCH \( \left[{\sum}_{j=1}^s{\theta}_j\frac{\varepsilon_{t-j}}{\sigma_{t-j}}\right] \), the generalized ARCH \( \left[{\sum}_{k=1}^q{\delta}_k{\sigma}_{t-k}^2\right] \), the Referendumt dummy variable, which takes the value one on the day after the vote, and the Brexitt dummy variable, which is in our focus.5 Besides capturing heteroskedastic patterns in the exchange rate returns, the benefit of this logarithmized model is the removal of the restrictive non-negativity constraint of the conventional GARCH model, so that the variance coefficients – like Brexitt – are allowed to be negative. Since individual Brexit news could have different impact extents on the volatility of the exchange rate and lead to residuals that are not conditionally normally distributed, we also use the Huber-White sandwich estimator to gain robust covariances. For more adequate modelling of detected fat tails in the exchange rate return time series, we perform the model with Student’s t distribution. For the EGARCH(p,q,s) order selection, we use the Akaike information criterion and check for autocorrelations in the squared residuals using the Ljung-Box Q-statistics and the ARCH-LM test.

Basically, the presence of Brexit news raises uncertainty and risk among market participants as described in Section 3.1 and should increase the conditional variance of the model, meaning that γ and ω in equation (3) show a positive and significant value. As before, we also run the EGARCH model for the good and bad Brexit dummies and check for asymmetric behaviour of market participants with regard to Brexit news. Additionally, the EGARCH model allows us to check for a possible asymmetric reaction of market participants towards other innovations by applying the Wald test for the sum of the asymmetric ARCH coefficients θj. In the literature, the asymmetric reaction of market participants is called the leverage effect, since negative innovations show a higher impact on the volatility of assets than positive innovations (Pagan and Schwert 1990; Engle and Ng 1993). Based on this consideration, we expect for the spot exchange rates higher volatilities in depreciation periods. This asymmetric behaviour could be also present in case of good and bad Brexit news, because in cases of bad Brexit news market participants could perceive a higher probability of the Brexit and fear of the negative consequences on the British economy. However, the fact, that only two of seventeen Brexit news observations are located before the referendum, could lead to the opposite behaviour: After the vote to exit the EU, market participants anticipate the worst scenario of a hard Brexit. As a conclusion, bad Brexit news appearing after the vote could lead to no further changes of expectations and thus no major changes in the market. Good Brexit news, however, increases the hope for a soft Brexit or even of the possibility of the UK not leaving the EU after all. In this case, the expectations of market participants would change and lead to market adjustments with higher volatilities than in case of bad Brexit news.

Since the referendum fundamentally changes the economic framework between the UK and other EU countries, and has only an indirect impact on the UK’s bilateral economic relations with the US, the effect of Brexit should be greater on British pound/euro exchange rate volatility than on the British pound/US dollar volatility. To check this consideration, and to compare the impact of Brexit news on both exchange rates, we also perform the EGARCH model for the British pound/US dollar exchange rate volatility.

4.2 Results for the volatility

4.2.1 Spot exchange rate of the British pound against the euro

At first, we perform our model for the volatility of the spot exchange rate of the British pound against the euro using all Brexit-related news. As presented in Table 11, the EGARCH(p,q,s) specifications with the lowest Akaike values show that for all cases, the Brexit referendum dummy has a significant and positive impact on the variance of the British pound/euro exchange rate. This result is not very surprising, since the vote changes the framework conditions for all sectors in the domestic British economy and the UK’s foreign trade. More interesting is the finding that in all four specifications, the coefficient of the remaining Brexit news ω also shows a highly significant and positive impact on the exchange rate variance. Overall, the results indicate the increasing effect of Brexit as a whole on the volatility of the British pound/euro exchange rate and show robust coefficients. Furthermore, in all model setups, the sum of the asymmetric coefficients \( {\sum}_{j=1}^s{\theta}_j \) is statistically significant at the 5 % level according to the Wald test. As expected, this result indicates the presence of an asymmetric behaviour of market participants towards shocks in the exchange rate market, similar to the leverage effect: Volatility tends to rise higher in cases of an (unexpected) depreciation of the pound than in cases of an (unexpected) appreciation.
Table 11

Impact of Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the Euro

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(4,4)

(4,5)

(5,5)

Asymmetric order (s)

(3)

(3)

(3)

(3)

Variance equation:

 Brexit referendum (γ)

3.1920**

3.2307**

3.2577**

3.2549**

(1.4694)

(1.4699)

(1.3826)

(1.3969)

 Brexit news (all events) (ω)

0.9653***

1.0107***

1.0216***

1.0208***

(0.2843)

(0.2812)

(0.2900)

(0.2929)

 Sum of the asym. Coeff. \( \left({\sum}_{j=1}^s{\theta}_j\right) \)

0 .1322**

0.1400**

0.1380**

0.1379**

(0.0553)

(0.0567)

(0.0546)

(0.0547)

Akaike IC

−7.7702

−7.7696

−7.7682

−7.7667

Schwarz IC

−7.7129

−7.7085

−7.7033

−7.6980

Hannan-Quinn IC

−7.7487

−7.7467

−7.7439

−7.7410

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

After the detection of the asymmetric behaviour towards innovations, we check whether market participants also show asymmetric behaviour towards bad and good Brexit news. As in Section 3.2, we run the EGARCH model using the bad and good Brexit news dummies and test whether their impact on the variance differ in terms of their significance, magnitude and sign. Tables 12 and 13 show the results of the EGARCH estimations with bad and good Brexit news implemented separately. At first sight, we can see that all model variations except for specification (4) of Table 12 indicate the presence of asymmetric behaviour towards positive and negative shocks in the manner of the leverage effect, as shown above. The Brexit referendum dummy has a significant impact on the volatility in all specifications in Table 12 with the exception of model (1). Only for the first specification does the bad Brexit news dummy show a significant impact on the volatility. However, the results in Table 13 reveal that in all specifications the good Brexit dummy has a significant impact on the volatility, even if in two specifications the significance level is under 10%. Comparing the results in Tables 12 and 13, it seems like the impact significance and also the impact magnitude of good Brexit news is higher than the impact of bad Brexit news, with the exception of the referendum. Nevertheless, the results are not robust with respect to the lag specification of the EGARCH model. The reason for that could be the small number of days with Brexit news after the splitting of the sample into good and bad Brexit dummy variables.
Table 12

Impact of Bad Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the Euro

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(3,3)

(2,3)

(2,3)

Asymmetric order (s)

(3)

(3)

(3)

(1)

Variance equation:

 Brexit referendum (γ)

3.3824

3.7352***

3.8012***

3.6809***

(2.5991)

(0.8080)

(0.8347)

(0.9298)

 Brexit bad news (ω)

0.9670**

0.3325

0.1776

0.0890

(0.3768)

(0.4303)

(0.4515)

(0.5104)

 Sum of the asym. Coeff. \( \left({\sum}_{j=1}^s{\theta}_j\right) \)

0.1029**

0.0907**

0.0876**

0.0550

(0.0512)

(0.0408)

(0.0382)

(0.0379)

Akaike IC

−7.7615

−7.7610

−7.7605

−7.7571

Schwarz IC

−7.7042

−7.7076

−7.7109

−7.7151

Hannan-Quinn IC

−7.7401

−7.7410

−7.7420

−7.7414

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Table 13

Impact of Good Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the Euro

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,3)

(2,3)

(4,4)

(4,5)

Asymmetric order (s)

(3)

(3)

(3)

(3)

Variance equation:

 Brexit referendum (γ)

4.0720***

4.1302***

3.5535**

3.3897**

(0.7294)

(0.7642)

(1.4332)

(1.4272)

 Brexit good news (ω)

0.7148*

0.6972*

1.1116***

1.2822***

(0.3768)

(0.4289)

(0.3579)

(0.3883)

 Sum of the asym. Coeff. \( \left({\sum}_{j=1}^s{\theta}_j\right) \)

0.1044**

0.1040**

0.1471***

0.1407***

(0.0464)

(0.0414)

(0.0525)

(0.0441)

Akaike IC

−7.7636

−7.7638

−7.7636

−7.7629

Schwarz IC

−7.7102

−7.7142

−7.7025

−7.6980

Hannan-Quinn IC

−7.7436

−7.7453

−7.7408

−7.7386

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

To construct a more robust framework, we include both good and bad Brexit dummy variables in the EGARCH model simultaneously. Table 14 exhibits the results of this approach. Besides specification (4), the coefficients and their significance become more robust and confirm the previous findings. As above, the sum of the asymmetric coefficients \( {\sum}_{j=1}^s{\theta}_j \) and the Brexit referendum dummy variable have a significant impact on the volatility. In particular, the coefficients and the statistical significance of the bad and good Brexit dummy variables become more evident and show an asymmetric impact on the volatility of the exchange rate. First, the good Brexit news dummy is statistically significant at the 1 % level in models (1), (2) and (3) and significant at the 10 % level in model (4). In contrast to that, the bad Brexit dummy variable is significant in models (1), (2) and (3) only at the 10 % level and shows no significant impact in model (4). Moreover, in all specifications the magnitudes of the good Brexit coefficient (ωgood) are clearly higher than the coefficient of the bad Brexit variable (ωbad). To summarize these results: In contrast to the asymmetric behaviour towards other innovations, market participants react more strongly to good Brexit news than to bad Brexit news, without taking into account the referendum effect. Apparently, the participants already expect the exit of the UK from the EU and thus are pricing in at least a soft Brexit, so that bad Brexit news has a minor impact on the volatility compared to good Brexit news.6
Table 14

Impact of Good and Bad Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the Euro

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(4,4)

(4,5)

(3,3)

Asymmetric order (s)

(3)

(3)

(3)

(3)

Variance equation:

 Brexit referendum (γ)

3.5008***

3.5242***

3.5398***

4.0846**

(1.3058)

(1.1648)

(1.1618)

(0.7442)

 Brexit bad news (ωbad)

0.7180*

0.7649*

0.7449*

0.3721

(0.4059)

(0.3964)

(0.4240)

(0.4425)

 Brexit good news (ωgood)

1.0813***

1.1268***

1.1653***

0.7294*

(0.3781)

(0.3758)

(0.4507)

(0.4447)

 Sum of the asym. Coeff. \( \left({\sum}_{j=1}^s{\theta}_j\right) \)

0.1338**

0.1426***

0.1404***

0.1057**

(0.0534)

(0.0543)

(0.0525)

(0.0444)

Akaike IC

−7.7657

−7.7654

−7.7640

−7.7632

Schwarz IC

−7.7046

−7.7005

−7.6953

−7.7059

Hannan-Quinn IC

−7.7428

−7.7411

−7.7382

−7.7417

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

4.2.2 Spot exchange rate of the British pound against the USD

In this section, we rerun the EGARCH approach of Section 4.2.1. in using the spot exchange rate of the British pound against the US dollar. Our aim is to reveal if Brexit news has a crucial impact on the volatility of the US dollar exchange rate and if volatility patterns with respect to bad and good Brexit news exist and indeed differ from the findings for the British pound/euro exchange rate volatility. As above, we start with the EGARCH model using the entire data set of Brexit news, the results of which is presented in Table 15. The referendum dummy variable shows a significant impact at the 1 % level in specifications (1) and (3) and in specification (2) a significant impact at the significance level of 5 %. However, we find no significant impact of the referendum dummy in model (4). Surprisingly, the coefficient of the Brexit news dummy ω shows a significant impact on the exchange rate volatility only for specification (3) at the 10 % level. So, regarding these results, the statistical significance of the impact of the Brexit news on the exchange rate is rather thin in this setup, except for the impact of the referendum. Nevertheless, in all four specifications we can measure that the volatility tends to increase more strongly in periods with depreciations than in periods with appreciations, although the statistical significance is only at the 10% level.
Table 15

Impact of Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the US Dollar

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(4,4)

(4,5)

(2,3)

Asymmetric order (s)

(3)

(3)

(3)

(3)

Variance equation:

 Brexit referendum (γ)

3.1285 ***

3.1095**

4.1145***

2.2014

(1.1582)

(1.3063)

(0.9022)

(1.8418)

 Brexit news (all events) (ω)

0.7443

0.7118

0.6684*

0.5573

(0.4798)

(0.5076)

(0.3741)

(0.6182)

 Sum of the asym. Coeff. \( \left(\sum \limits_{j=1}^s{\theta}_j\right) \)

0.1141*

0.1232*

0.1356*

0.0732*

(0.0644)

(0.0673)

(0.0740)

(0.0434)

Akaike IC

−7.7622

−7.7611

−7.7599

−7.7599

Schwarz IC

−7.7049

−7.7000

−7.6951

−7.7103

Hannan-Quinn IC

−7.7408

−7.7382

−7.7357

−7.7414

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Tables 16 and 17 show the results of the bad and good Brexit news, respectively, firstly implemented separately in the EGARCH model. The referendum dummy shows for all specifications a significant impact on the exchange rate volatility. More interesting is the finding, contrary to the results in Section 4.2.1, that good Brexit news has no impact in all specifications and the impact of the bad Brexit news is statistically measurable only in model (3) at the 5 % level and in model (4) at the 10 % level.
Table 16

Impact of Bad Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the US Dollar

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(4,4)

(4,5)

(5,5)

Asymmetric order (s)

(3)

(3)

(3)

(3)

Variance equation:

 Brexit referendum (γ)

2.9728**

2.9680**

4.0121***

3.5042***

(1.1899)

(1.2732)

(1.0234)

(1.0935)

 Brexit bad news (ω)

0.8150

0.8066

0.7742**

0.9318*

(0.5121)

(0.5186)

(0.3451)

(0.5482)

 Sum of the asym. Coeff. \( \left({\sum}_{j=1}^s{\theta}_j\right) \)

0.1174**

0.1210**

0.1435*

0.1137*

(0.0577)

(0.0567)

(0.0739)

(0.0657)

Akaike IC

−7.7610

−7.7598

−7.7575

−7.7566

Schwarz IC

−7.7038

−7.6987

−7.6926

−7.6879

Hannan-Quinn IC

−7.7396

−7.7369

−7.7332

−7.7309

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Table 17

Impact of Good Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the US dollar

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(4,4)

(4,5)

(5,5)

Asymmetric order (s)

(3)

(3)

(1)

(3)

Variance equation:

 Brexit referendum (γ)

2.9685**

2.8925**

4.1504***

2.9268*

(1.1885)

(1.4182)

(1.0443)

(1.7822)

 Brexit good news (ω)

0.5316

0.5254

0.8481

0.5297

(0.5586)

(0.4909)

(0.5508)

(0.5617)

 Sum of the asym. Coeff. \( \left(\sum \limits_{j=1}^s{\theta}_j\right) \)

0.1280**

0.1367**

0.0285

0.1393**

(0.0544)

(0.0540)

0.1138

(0.0601)

Akaike IC

−7.7593

−7.7585

−7.7566

−7.7556

Schwarz IC

−7.7021

−7.6974

−7.6994

−7.6869

Hannan-Quinn IC

−7.7379

−7.7356

−7.7352

−7.7299

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

The results displayed in Table 18, with the good and bad Brexit dummy variables implemented simultaneously into the EGARCH model, confirm that the volatility patterns towards Brexit news and innovations, which we show for the British pound/euro exchange rate, are identifiable only in case of the British Pound/US dollar exchange rate.
Table 18

Impact of Good and Bad Brexit News on the Variance of the Spot Exchange Rate of the British Pound against the US Dollar

 

Specification

(1)

(2)

(3)

(4)

Order of GARCH (p,q)

(3,4)

(4,4)

(3,4)

(4,5)

Asymmetric order (s)

(3)

(3)

(2)

(1)

Variance equation:

 Brexit referendum (γ)

3.2298***

3.2391***

2.6569

4.1690***

(1.0781)

(1.2216)

(3.8926)

(0.8729)

 Brexit bad news (ωbad)

0.9464*

0.9228

0.7599

0.8399**

(0.5449)

(0.5962)

(1.2616)

(0.3536)

 Brexit good news (ωgood)

0.6871

0.6874

0.6727

0.6967

(0.6253)

(0.6239)

(1.2234)

(0.5374)

 Sum of the asym. Coeff. \( \left({\sum}_{j=1}^s{\theta}_j\right) \)

0.1147*

0.1227*

0.0875

0.0433

(0.0644)

(0.0661)

(0.1294)

(0.0352)

Akaike IC

−7.7617

−7.7606

−7.7606

−7.7595

Schwarz IC

−7.7006

−7.6957

−7.7033

−7.6984

Hannan-Quinn IC

−7.7388

−7.7363

−7.7392

−7.7366

Standard errors in parentheses. ***, ** and * display significance at the 1%, 5% and 10% level, respectively

Looking at the big picture, the results for the British pound/US dollar exchange rate volatility fit the expectations mentioned in Section 3.1, namely that in this case Brexit news plays a relatively minor role compared with the impact on the Pound/euro volatility.

5 Conclusion

This study investigates the impact of Brexit-related events on the spot exchange rate of the British pound against the euro. With our identification strategy, we find a total of 17 Brexit-related events. Hence, we consider the influence of the Brexit referendum and other Brexit- related events on the British pound/euro exchange rate. Moreover, we split our Brexit-related events into bad and good Brexit news, respectively. We find, that bad Brexit news is associated with a depreciation of the British pound against the euro and that good Brexit news resulted in an appreciation of the British pound vis-à-vis the euro. Furthermore, our results suggest that market participants displayed a delayed reaction to bad and good news because our findings show that the coefficients of the Brexit news dummy variables are clearly larger in the two-day window than in the one-day window. Our results also show that the impact of bad Brexit news is more persistent than the effect of good Brexit news on the spot exchange rate of the British pound against the euro. We find that bad Brexit events have an impact on the spot exchange rate of the British pound against the euro two days following the announcement of bad Brexit news, whereas good Brexit events have no influence on the British pound/euro exchange rate two days following the announcement of good Brexit news.

Concerning the effect of Brexit-related news on the spot exchange rate of the British pound against the US dollar, we find that only bad Brexit news has a statistically significant impact on the British pound/US dollar exchange rate. Hence, good Brexit news events seem to have no impact on the pricing of the British pound against the US dollar. Furthermore, our empirical results suggest that the spot exchange rate of the British pound against the US dollar reacts in a delayed manner to bad Brexit events but does not exhibit a delayed reaction to good Brexit news. Overall, our empirical results indicate that the British Pound/US dollar exchange rate was affected by bad Brexit news whereas the British pound/euro exchange rate was influenced by both bad and good Brexit news, respectively.

Our findings also suggest that the British pound/euro exchange rate and the spot exchange rate of the British pound against the US dollar is more strongly affected by bad news than by good news, respectively. Hence, in sum, Brexit-related news resulted in a depreciation of the British pound against the euro and the US dollar, respectively. The depreciation of the British pound against these two major currencies might dampen the expected negative impact of Brexit on the GDP growth rate in the UK. It seems that UK exports were strengthened by the depreciation of the British pound against the euro and the US dollar due to Brexit-related events. However, the depreciation of the British pound generated higher inflation rates and hence lower real wages and thus lower household consumption in the UK (Bank of England 2016b). However, it seems that the deprecation of the British pound helped the UK economy as the GDP growth rate in 2017 was higher than previously forecasted by the Bank of England (Bank of England 2016b).

Using the EGARCH model, we show that the Brexit referendum and other Brexit-related events have a significant impact on the volatility of the British pound/euro exchange rate. As discussed, we use both bad and good Brexit news and illustrate that market participants behave asymmetrically towards good and bad Brexit news in terms of the British pound/euro exchange rate volatility. The statistical significance and the magnitude of the impact of good Brexit news is clearly higher than those of bad Brexit news. In stark contrast to that finding, we show that negative shocks have a greater impact on the volatility of the British pound/euro exchange rate than positive shocks. Finally, we show that with the exception of the referendum event, the evidence for the impact of Brexit news on the British pound/US dollar exchange rate volatility is either scarce or poor.

Our empirical results also indicate that the communication strategy of the UK government could be improved. Bad Brexit news and good Brexit-related events led to an increase of the exchange rate volatility of the British pound/euro exchange rate and hence higher economic uncertainty. A smoother communication strategy could at least have avoided a rise of the exchange rate volatility of the British pound/euro exchange rate due to statements concerning good Brexit news. A higher exchange rate volatility may have detrimental effects on economic growth in the UK as a higher exchange rate volatility could be associated with lower business investments due to the “waiting effect” (Darby et al. 1999) and lower trade due to higher transaction costs. There also seems to be a negative relationship between exchange rate volatility and job creation (Belke and Kaas 2004). Hence, the UK government should avoid thoughtless Brexit statements that trigger a higher exchange rate volatility of the British pound because this may have severe effects on both the GDP growth rate and on job creation in the UK. So, the information content of a Brexit-related statement of the UK government should be detailed in order to avoid financial turmoil. Furthermore, the outcome of the Brexit negotiations between the UK and the EU should be well communicated to market participants in order to avoid a sudden and strong rise of the exchange rate volatility of the British pound/euro exchange rate and hence an increase of economic uncertainty.

Footnotes

  1. 1.

    See Appendix 1 Table 19.

  2. 2.

    Note, that similar events are not necessarily classified into the same news group.

  3. 3.

    Our empirical results also suggest that bad Brexit news, excluding the announcement of the referendum result, has a statistically significant impact on the spot exchange rate of the British pound against the euro. Hence, we find that other bad Brexit-related events, excluding the referendum, also affected the British pound/euro exchange rate. Results are set out in Table 20 in Appendix 2.

  4. 4.

    Here, we also find that bad Brexit-related events, excluding the Brexit referendum, have a statistically significant impact on the spot exchange rate of the British pound against the US dollar. The results are reported in Table 21 in Appendix 2.

  5. 5.

    Like ARMA models, GARCH models are very sensitive to outliers, since outlying values bias the estimation repeatedly due to the autoregressive progress. Since the British pound depreciated sharply against the euro following the Brexit referendum, we implement in the variance equation the referendum dummy variable, which absorbs the external shock of the vote. So Referendumtconsists of zeros except for the 24 June 2016, at which it takes the value one. To avoid double counting, we replace the value one at the 24 June 2016 in the Brexitt dummy variable with a zero. The stark depreciation of the British pound against the US dollar after the vote persists over two working days. Thus, vis-à-vis the British pound/US dollar exchange rate volatility, the referendum dummy takes the value one on the 24 and 27 June 2016.

  6. 6.

    In addition, this finding is consistent with the results of Section 3.2.1 (see Table 3), where the coefficient of good Brexit news has a higher significance level than the coefficient of bad Brexit news.

Notes

Acknowledgements

This paper is part of EIIW research funded by the Deutsche Bundesbank. While the authors gratefully acknowledge funding from the Deutsche Bundesbank within the project “The Influence of Brexit on the EU28: Banking and Capital Market Adjustments as well as Direct Investment Dynamics in the Eurozone and other EU Countries”, opinions expressed within represent those of the authors and do not reflect the views of the Deutsche Bundesbank or its staff. We gratefully acknowledge editorial assistance by David Hanrahan (EIIW), data assistance by Lev Nazarov and Valeryia Siarheyeva, comments by Prof. Dr. Andre Jungmittag, Frankfurt University of Applied Sciences and IPTS Sevilla, Prof. Dr. Frank Bohn, Radboud University, Prof. Dr. Volker Clausen, University of Duisburg-Essen, Tobias Gruhle, Johannes Gutenberg-University, EIIW seminar participants and participants of the 14th CEUS Workshop on European Economics. Comments from Alexandra Dumitru, Rabobank, Manfred Kremer, European Central Bank, Prof. Dr. Paul J.J. Welfens, University of Wuppertal, Andrew Mullineux, University of Birmingham, and Christopher Thiem, University of Duisburg-Essen, at the EIIW workshop, March 16, 2018, are appreciated. The usual disclaimer applies.

References

  1. Arnorsson A, Zoega G (2016) On the causes of Brexit. CESifo Working Paper 6056Google Scholar
  2. Baier F, Welfens PJJ (2018) BREXIT and FDI: key issues and new empirical findings. EIIW discussion paper no. 241Google Scholar
  3. Baker SR, Bloom N, Davis SJ (2016) Policy uncertainty: trying to estimate the uncertainty impact of Brexit. Presentation. http://www.policyuncertainty.com/media/Brexit_Discussion.pdf
  4. Bank of England (2016a) Inflation report May 2016. LondonGoogle Scholar
  5. Bank of England (2016b) Inflation report November 2016. LondonGoogle Scholar
  6. Belke A, Dubova I, Osowski T (2016) Policy uncertainty and international financial markets: the case of Brexit. Ruhr Econ Pap 657Google Scholar
  7. Belke A, Kaas L (2004) The Impact of Exchange Rate Volatility on European and U.S. Labor markets. Journal of Economic Asymmetries 1(2):21–50Google Scholar
  8. Caporale GM, Gil-Alana L, Trani T (2018) Brexit and uncertainty in financial markets. DIW Discussion Papers 1719Google Scholar
  9. Darby J, Hughes Hallet A, Ireland J, Piscitelli L (1999) The impact of exchange rate uncertainty on the level of investment. Econ J 109(454, Conference Papers):C55–C67CrossRefGoogle Scholar
  10. Davies RB, Studnicka Z (2017) The heterogeneous impact of Brexit: early indications from the FTSE. CESifo Working Papers 6478Google Scholar
  11. Dhingra S, Huang H, Ottaviano G, Pessoa JP, Sampson T, Van Reenen J (2017) The costs and benefits of leaving the EU: trade effects. CEP Discussion Paper 1478Google Scholar
  12. Engle RF, Ng VK (1993) Measuring and testing the impact of news on volatility. J Financ 48(5):1749–1778CrossRefGoogle Scholar
  13. Fratzscher M, Lo Duca M, Straub R (2014) ECB unconventional monetary policy actions: market impact, international spillovers and transmission channels. In: International Monetary FundGoogle Scholar
  14. Gourinchas PO, Hale G (2017) Brexit: whither the pound? FRBSF Economic Letter 2017–11Google Scholar
  15. Krause T, Noth F, Tonzer L (2016) Brexit (probability) and effects on financial market stability. IWH online, 5/2016, Halle Institute for Economic ResearchGoogle Scholar
  16. Nelson DB (1991) Conditional Heteroskedasticity in asset returns: a new approach. Econometrica 59(2):347–370CrossRefGoogle Scholar
  17. Ottaviano G, Pessoa JP, Sampson T (2014) The costs and benefits of leaving the EU. Technical paper, CEPGoogle Scholar
  18. Pagan AR, Schwert GW (1990) Alternative models for conditional stock volatility. J Econ 45(1–2):267–290CrossRefGoogle Scholar
  19. Portes J, Forte G (2017) The economic impact of Brexit-induced reductions in migration. Oxf Rev Econ Policy 33(Supplement 1):31–S44CrossRefGoogle Scholar
  20. Ramiaha V, Phama HNA, Moosab I (2017) The sectoral effects of Brexit on the British economy: early evidence from the reaction of the stock market. Appl Econ 49(26):2508–2514CrossRefGoogle Scholar
  21. Walker N (2018) Brexit timeline: events leading to the UK’s exit from the European Union. Briefing paper no. 07960, house of commons libraryGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Authors and Affiliations

  1. 1.EIIW at the University of WuppertalWuppertalGermany
  2. 2.Schumpeter School of Business and EconomicsWuppertalGermany
  3. 3.Frankfurt University of Applied SciencesFrankfurtGermany

Personalised recommendations