The Perception of Brexit Uncertainty and How it Affects Markets

Priberny, Christopher; Kreuzer, Christian; Huther, Johannes Article The Perception of Brexit Uncertainty and How it Affects Markets Credit and Capital Markets – Kredit und Kapital Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Priberny, Christopher; Kreuzer, Christian; Huther, Johannes (2024) : The Perception of Brexit Uncertainty and How it Affects Markets, Credit and Capital Markets – Kredit und Kapital, ISSN 2199-1235, Duncker & Humblot, Berlin, Vol. 57, Iss. 1/4, pp. 31-47, https://doi.org/10.3790/ccm.2025.1457101 This Version is available at: https://hdl.handle.net/10419/324958 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ The Perception of Brexit Uncertainty and How it Affects Markets Christopher Priberny*, Christian Kreuzer**, and Johannes Huther*** Abstract We empirically study the perception of political uncertainty by UK’s stock markets, covering the entire Brexit period from January 2013 to March 2020. We find that indices dominated by the largest capitalized companies anticipate negatively perceived events already prior to the actual event, whereas positive events only effect them on the event day or following. In contrast, the FTSE 250, composed of medium-sized companies, tends to move prior to positively perceived events. Furthermore, we investigate the daily perception of Brexit measured by a metric based on Google Trends. Our results show that perception significantly affects all major UK indices. Keywords: Brexit, event study, perception, Google Trends, GJR-GARCH, United Kingdom JEL Classification: G14, G11, G39, H87 I. Introduction The Brexit vote and the resulting decision of the United Kingdom to withdraw from the European Union is unique in terms of various issues. The possible future effects in terms of trade barriers and tariffs, free trade agreements, freedom * Prof. Dr. Christopher Priberny: Corresponding author, Deutsche Bundesbank University of Applied Sciences, 57627 Hachenburg, Germany and Department of Finance, University of Regensburg, 93040 Regensburg, Germany. E-Mail: [email protected]. ** Dr. Christian Kreuzer: Department of Finance, University of Regensburg, 93040 Regensburg, Germany. E-Mail: [email protected]. *** Johannes Huther: Deutsche Bundesbank, 60006 Frankfurt am Main, Germany. E-Mail: [email protected]. The contributions to this study represent the authors’ personal opinions and do not necessarily reflect the views of Deutsche Bundesbank. Acknowledgement: We are grateful to an anonymous referee and the participants of the World Finance Conference 2022 for helpful comments and valuable suggestions. Credit and Capital Markets, 57 (2024) 1 – 4: 31 – 47 https://doi.org/10.3790/ccm.2025.1457101 Scientific Papers Open Access– Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0). Duncker & Humblot · Berlin 32 Christopher Priberny, Christian Kreuzer, and Johannes Huther Credit and Capital Markets, 57 (2024) 1 – 4 of movement, and economic decoupling could hardly be assessed in the stormy political period after the referendum on June 23rd, 2016. Uncertainties resulting from the surprising outcome of the Brexit referendum were also reflected in the global financial markets. As a first reaction, many international stock markets suffered significant losses on the following day. Brühl (2018) highlights the role of London in clearing euro-denominated OTC derivatives and implications after Brexit. However, there are indeed many more events related to this period of political distress that are closely related to Brexit and have caused significant reactions on the stock market. In this study, we examine the impact of major political events in the context of Brexit on financial markets. For a holistic financial view, we cover the entire Brexit period from January 2013 to March 2020 and analyze positively or negatively perceived events separately. Furthermore, we show how the daily perception of political uncertainty, proxied by a metric based on Google Trends affects UK stock indices. We contribute to a literature strand that deals with stock market reactions of various Brexit events (Breinlich etal., 2018; Hudson etal. 2020; Ramiah etal. 2017; Shahzad etal. 2019). As a matter of fact, early studies (Breinlich etal. 2018; Ramiah etal. 2017; Shahzad etal. 2019) comprise only the outcome of the Brexit referendum and some close events and only show a limited picture of the whole Brexit process. To our best knowledge, we are the first to investigate Brexit-related events in the entire time frame between January 2013 and March 2020. Furthermore, we extend previous studies by considering the timing of market reactions, avoiding overlapping event windows and heeding whether an event is perceived as positive or negative by the market. In particular, we analyze certain Brexit events by using a GJR-GARCH Model. We find differences in the impact of various Brexit events depending on both whether they are perceived as positive or negative on the financial market as well as on the companies’ size. In the case of positively perceived events, we find — in particular for indices dominated by larger sized companies — significant influences from the event day and the following ones. In contrast, significant influences can already be identified in the preceding days of negatively perceived events. Additionally, we consider the whole process of Brexit as a phase of uncertainty. Therefore, we measure the public sentiment of political uncertainty by an innovative measure based on Google Trends data. We find, the daily sentiment has a significant effect on stock returns. The rest of the article is organized as follows: In Section II we present relevant literature and develop hypotheses. In Section III we address data sources as well as methodological approaches. The results are presented in Section IV. SectionV concludes this work. The Perception of Brexit Uncertainty and How it Affects Markets 33 Credit and Capital Markets, 57 (2024) 1 – 4 II. Literature & Hypotheses Concerning the existing academic literature, one field deals with general economic consequences of Brexit (Born etal. 2019; Hosoe 2018; Jackson/Shepotylo 2018; Steinberg 2019). Furthermore, there are studies focusing on the economic performance measured by the gross domestic product (GDP) (Born etal. 2019; Hosoe 2018) as well as there is a literature strand addressing implications on welfare (Jackson/Shepotylo 2018; Steinberg 2019). Another strand of literature investigates the impacts on the volatility of stock and exchange markets (Adesina 2017; Belke etal. 2018; Qiao etal. 2021). In this regard, Belke etal. (2018) assess interactions of the UK’s political uncertainty on the economy and the volatility on financial markets, while Adesina (2017) investigates the effect of the Brexit vote on the persistence of volatility. Moreover, Belke etal. (2018) show that political uncertainty will on the one hand continue to cause instability in key financial markets and on the other hand has the potential to damage the economy not only in the UK but also in European countries. Adesina (2017) indicates a significant increase in volatility persistence for stock markets, but a decrease of volatility persistence in foreign exchange markets. Qiao etal. (2021) investigate the impact on the US stock market and show increased volatility of S&P 500 returns even before Brexit. Besides, some authors analyze the impact of economic policy uncertainty (Armelius etal. 2017; Ko/Lee 2015; Nilavongse etal. 2020; Phan etal. 2019; Yung/ Root 2019) on the economy and stock markets. In particular, Nilavongse etal. (2020) evaluate implications of economic policy uncertainty shocks on the UK economy and find that Brexit uncertainty caused a massive depreciation of the British pound. Moreover, Braun/Zenker (2022) discuss the relations between the trust in EU and UK government and perceived uncertainty and identify strong country differences for soft and hard Brexit scenarios. Few authors examine the relational dynamics and cross-correlation between the UK and European markets before and after the Brexit referendum (Bashir etal. 2019; Guedes etal. 2019). In detail, Bashir etal. (2019) conclude that after the Brexit referendum most EU financial markets tend to show a negative correlation with the UK market in the long term, whereas Guedes etal. (2019) show a decrease in cross-correlation. Besides, Ayadi (2021) investigates the transmission of shocks caused by the Brexit across international equity markets to detect contagion effects. Another literature strand deals with stock market reactions of various Brexit events. Most of these studies focus on the UK market (Breinlich etal. 2018; Hudson etal. 2020; Ramiah etal. 2017; Shahzad etal. 2019). In particular, Ramiah etal. (2017) examine how the outcome of the Brexit referendum affects UK industry sectors and find a negative impact on the banking, travel, and leisure in- 34 Christopher Priberny, Christian Kreuzer, and Johannes Huther Credit and Capital Markets, 57 (2024) 1 – 4 dustry. However, as an early study it is limited to the Brexit vote itself and therefore to a rather small time window from June to July 2016. Furthermore, Breinlich et al. (2018) analyze reactions of stocks to three events around the referendum on EU membership. They show that initial stock movements are driven by fear of economic recession and sterling depreciation following the referendum as well as of potential changes to the UK-EU trade relations. Shahzad etal. (2019) investigate a chain of pre- and post-Brexit referendum events. They find not only an initially negative market reaction attributed to the Brexit referendum, but also positive reactions to post-Brexit referendum events up to March 2017 as future economic relations of the United Kingdom with EU began to take a shape. Hudson etal. (2020) analyze the impact of Brexit events on 34 British financial (sub-)indices in the mid of the whole Brexit phase until April 2017, applying a GJR-GARCH framework. They conclude that, depending on the business sector, new information regarding Brexit is quickly incorporated into market prices and could be widely explained by rational asset pricing models. However, they do not distinguish between positively and negatively perceived events, which might have resulted in many of their event coefficients being erroneously insignificant. Moreover, they use event windows of five days before and after each event, which results in overlapping time windows with regard to their event dataset and may cause some bias. While academic literature investigating political uncertainty is well studied (Julio/Yook 2012; Kelly etal. 2016; Li etal. 2022; Obenpong Kwabi etal. 2024; Pástor/Veronesi 2013), literature focusing on political uncertainty resulting from the Brexit Referendum is quite evolving. Manasse etal. (2024) investigate potential linkage of political risk on British pound exchange rates and find that the probability of Brexit predicts a depreciation of the pound and also that political risk is linked to exchange rates. Cucinelli etal. (2020) perform an event study solely on three Brexit events and find, that investors only price the days before the referendum as an event of political uncertainty. Hill etal. (2019) analyze the cross-sectional determinants of UK firms’ exposure to theBrexitevent and find that internationalization moderates potential Brexit exposure. When eyeing on the history of the whole phase of Brexit, you can identify events, that have caused a higher level of political uncertainty, as well as events showing a relief of political tensions. In line with early studies (Breinlich etal. 2018; Ramiah etal. 2017), we expect events, that caused increased political uncertainty to have a negative effect on stock markets. The effect is economically sound, as many companies in the UK show close relations with the EU and uncertainty about the future legal framework of those bonds have severe effects on their business models. The Perception of Brexit Uncertainty and How it Affects Markets 35 Credit and Capital Markets, 57 (2024) 1 – 4 Hypothesis 1: Political uncertainty has a negative influence on stock market returns. Following that implication, we also expect differences regarding the size of the companies. Largely capitalized companies are usually internationally oriented and often operate in a network of branches located in different countries. One can assume, that those players might have developed alternative plans on how to react when the final political decision about a specific form of Brexit is made. This might not hold for smaller and medium sized firms, as the setup cost for assuring excess to the European market even in the case of a hard Brexit are rather high. Therefore, we expect stock returns of medium sized companies to act more anxiously on news, that increase political uncertainty. Hypothesis 2: Medium sized firms are affected earlier by political uncertainty in the context of Brexit then other companies. However, the period following the Brexit referendum might also be seen as a whole phase of political uncertainty by the market, rather than being driven by specific events. Therefore we expect the daily public sentiment regarding political uncertainty1, whether it is positive or negative, to affect stock returns. Hypothesis 3: The market sentiment regarding political uncertainty has an impact on stock returns. III. Data & Methodology 1. Data For our analyses we use a unique dataset retrieved from three different data sources. a) Events Similarly to Hudson etal. (2020), the starting point for our thorough selection of events is about one year before Prime Minister David Cameron promised the EU referendum in case of his reelection. However, we use a larger time frame covering all relevant Brexit events from January 2013 (David Cameron advocates for a referendum) until the United Kingdom finally left the European Union on 31 January 2020. Furthermore, we neglect some events compared to Hudson etal. (2020) in order to avoid overlapping event windows. 1 Proxied by a Google trends metric. 36 Christopher Priberny, Christian Kreuzer, and Johannes Huther Credit and Capital Markets, 57 (2024) 1 – 4 We identify appropriate events in accordance with five principles: (1) Widespread mass media coverage of the respective event is necessary. As a suitable proxy we use peaks in Google Trends regarding the keywords2 ‘Exit’ and ‘Brexit’. (2) The coverage needs to have a sufficient impact on the emotions of a large part of the population. (3) It is sufficient that the emotional impacts on the population do not have different directions and thus offset each other but correlate across the majority of the population and are likely to affect market sentiment as well as asset prices. (4) If news of an event are published on a non-trading day or after time of closing, the following trading day is considered as the event day. (5) We select the main event and drop the secondary event whenever two events are closely together, in order to avoid overlapping event windows. As a result, we identify 33 Brexit-related events from January 2013 to March 2020, presented in Table 5 in the Appendix. b) Market returns Our second dataset consists of daily log returns (rt), covering the indices FTSE350, FTSE 100, as well as FTSE 250 and FTSE All-Share. All log returns were derived from daily performance indices retrieved from LSEG Datastream (www.lseg.com, formerly Refinitiv), covering the period between January 20123 and February 2020. c) Google Trends The last source are Google Trends data (www.google.com). However, obtaining and using Google Trends data is not straightforward. Google limits the frequency of Trends data available for individual download according to the period of interest. This means short periods like a month provide daily measures whereas longer periods only provide data on a monthly basis. Note that Google Trends data is indexed separately in each time period for which the data is downloaded. E.g. if January 2020 is selected, we receive 31 daily scores ranging from 100 to 0, whereas the day with the most Google searches regarding the re- 2 The term ‘Brexit’ became a common phrase in Feb 2016. Therefore, we apply ‘Exit’ as keyword until Jan 2017 and ‘Brexit’ since Feb 2016. In the overlapping window we use the mean of both Google trend series. 3 The additional log returns covering the year 2012 are necessary for calibrating the model. The Perception of Brexit Uncertainty and How it Affects Markets 37 Credit and Capital Markets, 57 (2024) 1 – 4 spective keyword is allocated the value 100. In order to derive a comparable daily metric for the whole period under consideration, we collect Google Trends data in the United Kingdom first on a monthly basis for the entire period under consideration and then daily data using a rolling time window for each month of the period under consideration. Next, we calculate a comparable daily metric GTt in the following way: =× 1 10,000 daily t tmon t GT GT GT where daily t GT notes the daily Google Trends value on a monthly basis and mon t GT the referring monthly value over the entire period. For better readability, we divide the resulting measure by 10,000. The descriptive statistics of market returns and Google Trends are shown in Table 1. A comparison of Google Trends data and Brexit events is presented in Figure 1 (see Appendix). It shows that all selected events are in line with spikes in the Google Trends data. Table 1 Descriptive statistics NMean SD Min Median Max FTSE 350 2,094 0.0003 0.0081 –0.0464 0.0005 0.0346 FTSE 100 2,094 0.0003 0.0083 –0.0478 0.0005 0.0352 FTSE 250 2,094 0.0005 0.0081 –0.0746 0.0007 0.0410 FTSE All-Share 2,094 0.0003 0.0079 –0.0463 0.0005 0.0341 Google Trends 2,094 0.1039 0.0873 0.0010 0.0885 1.0000 Notes: This table presents the mean, standard deviation, minimum, and maximum values of our dataset. Our sample ranges from January 2012 to February 2020. 2. Methodology a) Brexit events The application of classical event study methodology in the spirit of Fama etal. (1969), which focuses on the calculation of cumulative abnormal returns, is not appropriate for our setting. The reason is that while we focus on market returns there is no suitable benchmark that is not influenced by the Brexit events themselves. Hence we utilize the approach introduced by Sun/Tong (2010) and Glosten etal. (1993) to apply a GJR-GARCH model (see also Hudson etal. 2020; Priberny 2023). The GJR-GARCH model is suitable for our setting, as it allows error terms to deviate in an asymmetric way around events and therefore con- 38 Christopher Priberny, Christian Kreuzer, and Johannes Huther Credit and Capital Markets, 57 (2024) 1 – 4 trols for heteroscedasticity inherent in periods of stress and relief on the markets. In particular, we apply the following framework: ( ) αα α αε -+ = =- = + + +× + åå  53 0 1, 2, 13 t kt k k t k t t kk r r EC β β βε β βε - +- -- =- =+ + + + å 3 22 0112 3, 4 1 11 3 . t t k tk t tt k hh E I In this model, the logarithmic return for the market index on day t is represented by t r while - tk r comprises the k ’th previous daily market return, which controls for auto-correlation. By analyzing ACF plots and tests we identify 5 lags as suitable for our setting. t C represents a vector of controls (see Table 2 for details). + tk E denotes dummy variables addressing a window ±3 days around event days. Therefore, on event day =1 t E , and consequently 0 otherwise. Furthermore, ε t describes the residual for asset i at time t . In the second equation, t h represents the conditional variance of ε t as a proxy for market risk. -1t I denotes another dummy variable and equals 1 , if ε - < 1 0 t, and 0 otherwise. However, if there exists an event day return effect, the regression coefficient α 2 ,0 will be statistically significant. b) Daily Sentiment As addressed by Hypothesis 3, the whole period following the Brexit referendum might be seen as a phase of uncertainty by the market. Thus, we present additional analyses that are not limited to specific events. Therefore, we examine how the markets react to the public interest in Brexit, measured by Google Trends. In doing so, we are the first to examine the effect of Google Trends values on Brexit events, using a comparable daily metric. For this purpose, we adjusted the GJR-GARCH model in the following way: α α αδ α ε - = = + + +× + å 5 0 1, 2 1 ˆ t kt k t t t t k r r TC β β βε βε -- -- =+ + + 22 0112 4 1 11 tt t tt hh I where t T is the Google Trends value variable and δ ˆ t is a sign function which returns the sign of the market return on day t . Note that the integration of this factor is necessary to differentiate between days with mostly positive or negative information. However, we are thus restricted to analyzing the magnitude of the perception, as we can no longer investigate its direction. The Perception of Brexit Uncertainty and How it Affects Markets 45 Credit and Capital Markets, 57 (2024) 1 – 4 Ko, J. H./Lee, C. M. (2015): International economic policy uncertainty and stock prices: Wavelet approach. Economics Letters 134, 118 – 122. Kwabi, F. O./Adegbite, E./Ezeani, E./Wonu, C./Mumbi, H. (2024). Political uncertainty and stock market liquidity, size, and transaction cost: The role of institutional quality. International Journal of Finance & Economics 29, 2030 – 2048. Li, Q./Maydew, E. L./Willis, R. H./Xu, L. (2022): Corporate tax behavior and political uncertainty: Evidence from national elections around the world. Journal of Business Finance & Accounting 49, 1605 – 1641. Manasse, P./Moramarco, G./Trigilia, G. (2024): Exchange rates and political uncertainty: The brexit case. Economica 91, 621 – 652. Nilavongse, R./Rubaszek, M./Uddin, G. S. (2020): Economic policy uncertainty shocks, economic activity, and exchange rate adjustments. Economics Letters 186, 108765. Pástor, L./Veronesi, P. (2013): Political uncertainty and risk premia. Journal of Financial Economics 110, 520 – 545. Phan, H. V./Nguyen, N. H./Nguyen, H. T./Hegde, S. (2019): Policy uncertainty and firm cash holdings. Journal of Business Research 95, 71 – 82. Priberny, C. (2023): The Impact of COVID-19 on Demand and Lending Behavior in Prosocial P2P Lending. Credit and Capital Markets 56, 5 – 26. Qiao, K./Liu, Z./Huang, B./Sun, Y./Wang, S. (2021): Brexit and its impact on the US stock market. Journal of Systems Science and Complexity, 1 – 19. Ramiah, V./Pham, H. N. A./Moosa, I. (2017): The sectoral effects of brexit on the british economy: early evidence from the reaction of the stock market. Applied Economics 49, 2508 – 2514. Shahzad, K./Rubbaniy, G./Lensvelt, M./Bhatti, T. (2019): UK’s stock market reaction to brexit process: A tale of two halves. Economic Modelling 80, 275 – 283. Steinberg, J. B. (2019): Brexit and the macroeconomic impact of trade policy uncertainty. Journal of International Economics 117, 175 – 195. Sun, Q./Tong, W. H. (2010): Risk and the january effect. Journal of Banking & Finance 34, 965 – 974. Yung, K./Root, A. (2019): Policy uncertainty and earnings management: International evidence. Journal of Business Research 100, 255 – 267. 46 Christopher Priberny, Christian Kreuzer, and Johannes Huther Credit and Capital Markets, 57 (2024) 1 – 4 Appendix Table 5 Description of Brexit events Date Description of event Sentiment 01/23/13 David Cameron: Pro referendum + 05/08/15 UK 2015 General Election + 05/27/15 EU Referendum bill unveiled – 01/05/16 (Conservative) Ministers are allowed to campaign for either side in the referendum. – 02/02/16 European Council publishes a draft blueprint for the proposed changes to the UK’s membership of the EU. – 02/22/16 Announcement of referendum date + 06/16/16 Labour Party MP Jo Cox, a supporter of remaining in the EU, was murdered. + 06/24/16 Referendum + 07/12/16 Theresa May will become Prime Minister on 13/07/2016. + 10/03/16 Theresa May confirms that she will trigger Article 50 notice of Lisbon Treaty in March 2017. + 11/23/16 The UK’s Chancellor of the Exchequer, outlines his financial plans. + 01/17/17 May sets out plan for Brexit at Lancaster House. – 01/24/17 Supreme court: Parliament must be allowed to vote. – 03/20/17 Triggering of Article 50 announced. – 03/29/17 Triggering of Article 50: Two-year period for exit nego tiations begins. – 04/18/17 Prime Minister May calls snap general election. – 06/09/17 2017 General Election – 12/08/17 Joint report proposes solutions for Irish border. + 11/14/18 May and EU publish withdrawal agreement. – 12/13/18 May wins vote of confidence. – 01/16/19 May loses meaningful vote. + 03/13/19 May loses 2nd meaningful vote. + 03/29/19 Brexit Day 1: Theresa May loses 3rd meaningful vote. + 04/10/19 EU agrees to extension II. + The Perception of Brexit Uncertainty and How it Affects Markets 47 Credit and Capital Markets, 57 (2024) 1 – 4 Date Description of event Sentiment 05/24/19 May announces resignation. – 07/23/19 Boris Johnson wins race. + 08/29/19 Proroguing of Parliament + 09/24/19 Prorogation unlawful + 10/03/19 Johnson outlines proposal in parliament. – 10/17/19 New agreement with EU – 10/30/19 General election called. + 12/13/19 2019 General election + 02/03/20 Brexit Day 3 + Note: This plot compares Google Trends data (graph) to key Brexit events (lines). Figure 1: Google trends and Brexit events