Stock Market Reactions of UK Banks During the COVID-19 Crisis

Summary of Contents

Summary of Contents

List of Abbreviations

List of Figures

List of Tables

1 Introduction

2 Hypothesis development
2.1 Literature review
2.1.1 Theoretical framework
2.1.1.1 Market reactions to news
2.1.1.2 Influence of major events on the stock market
2.1.2 Function of banks
2.1.3 The banking system in the UK
2.1.4 Previous studies on the effect of COVID-19 on financial markets
2.2 Global financial of crisis 2007 to 2009
2.2.1 Timeline and causes
2.2.2 Event study description
2.2.3 Event study results
2.3 Hypothesis formulation

3 Event study methodology
3.1 Definition
3.2 Timeline
3.3 Calculating returns
3.4 Measuring normal returns
3.4.1 Constant mean return model
3.4.2 Market model
3.4.3 Economic models
3.5 Abnormal returns
3.5.1 Calculation of abnormal returns with the market model
3.5.2 Calculation of cumulative and average returns
3.6 Significance test
3.7 Robustness checks

4 Event study data
4.1 Historical data
4.2 Event dates

5 Empirical results
5.1 Bad-news events
5.2 Good-news events
5.3 Hypothesis analysis

6 Conclusion

List of Appendices

Appendix

References

Abstract

COVID-19 hit the world economy very hard in nearly every sector. As financial intermediaries, banks face the difficult task dealing with a shrinking economy. Previous studies have shown that events during the COVID-19 pandemic have significantly influenced the stock market. In the financial crisis of 2007 to 2009, bad-news events resulted in a negative abnormal return and good-news events resulted in positive abnormal returns in the banking sector in the United Kingdom.

This thesis aims to determine whether bad-news events resulted in a negative reaction and good-news in a positive reaction in the United Kingdom banking sector, or if there was no significant abnormal reaction at all.

To address this research question, an event study was conducted for four events investigating 11 banks in the United Kingdom as securities and the FTSE 350 as a market index. The results showed an overall significant abnormal reaction; however only bad­news events result in significantly negative abnormal returns, while good-news events did not show consistent results.

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

List of Figures

Figure 1: Overview of financial systems

Figure 2: Event study timeline

List of Tables

Table 1: Bad News: 29.09.2008 - Bailout plan rejection

Table 2: Good News: 10.03.2009 - Citigroup Inc. made profit

Table 3: Bad News: 12.03.2020 - WHO declares COVID-19 a pandemic

Table 4: Bad News: 23.03.2020 - Shutdown in UK

Table 5: Good News: 25.05.2020 - Non-essentials might reopen on June 15th

Table 6: Good News: 15.06.2020 - Non-essential retailers reopen

Introduction

Police-enforced curfews; closed schools, parks and restaurants; never leaving your home without a mask; not being able to visit friends and family - after starting like any other year, 2020 took an unexpected turn due to the outbreak of corona-virus disease 2019 (COVID-19). The respiratory disease COVID-19, caused by a novel strain of the coronavirus family, originated in Wuhan, China and has since reached a dramatic scale, socially and economically, that could not have been predicted.

The economic impact of the virus can be seen in the world’s major stock markets. Between 1 January 2020 and 17 March 2020, the Nikkei, Dow Jones and Financial Times Stock Exchange (FTSE) dropped by 27%, 29% and 33%, respectively¹ (BBC, 2020b). Furthermore, the unemployment rate in the United States of America (USA) reached its highest level since the start of recording by U. S. Bureau of Labour Statistics (2020) at 14.7% in April 2020, rising from 3.6% in January 2020. On the other hand, unemployment rates in the United Kingdom remained stable at 3.9% in the first two quarters of 2020 (Office for National Statistics, 2020). One reason for these stable numbers is the fact that furloughed workers are still counted as employed, despite their wages being paid by the government.

As a situation like this has never occurred before, its development cannot be predicted, and reactions such as lockdowns by the government can happen without premonition. There is uncertainty regarding how long COVID-19 will reduce economic activity and how long the UK government can support weakened businesses as well as its people. It may be only a matter of time before bankruptcies and personal insolvencies rise. Both would result in the inability to pay back bank loans and would lead to higher default rates than expected in the banks’ initial calculations, causing banks to lose profit.

A previous study by Niederhoffer (1971) describes a significant correlation between news and reactions on the stock market, and other studies (Bash, 2020; He, Liu, Wang, & Yu, 2020) have investigated an abnormal return (AR) on the stock market following major events in the COVID-19 pandemic. This thesis specifically investigates the influence of COVID-19 news and events in the UK on the UK banking sector.

This leads to the following research question: do banks’ stock prices react more abnormal to good and bad news in comparison to the market represented by FTSE 350? If so, how do they react?

To determine if banks react with more sensitivity, ARs were calculated using an event study methodology. In total the stock prices of 11 banks in the UK were examined on two good-news and two bad-news events during the COVID-19 crisis.

This thesis is organized as follows. First, the hypothesis is developed by setting the theoretical framework for market reactions to news and the role of banks, reviewing previous studies and conducting a study about the great financial crisis of 2007 to 2009. Next, the methodology of an event study and the calculation of necessary values is described, following the description of the data used in this study. Finally, the empirical results are presented, analysed, and interpreted.

2 Hypothesis development

2.1 Literature review

2.1.1 Theoretical framework

2.1.1.1 Market reactions to news

Looking at the point of view of economics, a market is efficient when the price of a given security fully reflects all information available to the public. Accordingly, stock markets react to deviations of expectations by immediately adjusting stock prices by reflecting the assumption of all participants (Fama, 1970, p. 383). Indeed, a study by Chordia, Roll, and Subrahmanyam (2001, pp. 28-29) showed that market imbalances adjust in a period of 5 to 60 minutes. Fama (1970, p. 383) segments this theory, known as the efficient market hypothesis, in three levels of information influencing the stock market, all assuming efficiency.

The weak form of the efficient marked hypothesis states that current share price is only based on information that historical prices have already consolidated. However it is not possible to predict future stock prices on former patterns (Malkiel, 1991, p. 127).

The semi-strong form of the efficient market hypothesis asserts that all information that is available to the public is priced into the price of a security. In addition to historical prices mentioned in the weak form, financial statements or announcements from companies shape the current price of a security (Clarke, Jandik, & Mandelker, 2001, p. 5).

The strong form claims that the current stock price does not only reflect past information but also future information that is not yet publicly available. According to Fama (1970, p. 410), this non-public information can lead to abnormal returns of a security. Market participants with knowledge about the aforementioned non-public information can therefore react prior to the market.

Event studies rely on the theory that the market is efficient. Furthermore, they presume the semi-strong form of market efficiency (Ang, 2015, p. 181).

2.1.1.2 Influence of major events on the stock market

In addition to the efficient market hypothesis described by Fama (1970), Niederhoffer (1971, p. 193) determined that it is more likely that the stock market has a bigger reaction to world events compared with dates that were randomly selected. In addition, Merrill (1984, p. 19), stated that bad news immediately results in a bad reaction on the stock market, but it often recovers after a short time.

After analysing a sample of 900,000 publicly available news stories Heston and Sinha (2016, pp. 16-19) confirmed Merrill’s findings that news does only influence stock market prices for one to two days. Their research also shows a clear difference in stock market reactions to good and bad news. While good news and events result in a positive reaction in the stock market its reaction to bad news and events is negative.

2.1.2 Function of banks

Banks are generally known as financial intermediaries and being the heart of the financial system. The financial intermediary plays the role of a middleman between borrowers and savers/investors. Their main function is accepting deposits from people or companies that want to save money and lending those deposits to credit users. Besides general banking fees, banks earn money by charging borrowers more in interest than they pay their investors in yield. Banks use many small deposits to finance fewer larger credits. (Mankiw, Taylor, Wagner, & Herrmann, 2012, pp. 682-683).

Abbildung in dieser Leseprobe nicht enthalten

Figure 1: Overview of financial systems (own figure, modified from Allen, Chui, and Maddaloni (2004, p. 491))

Figure 1 further outlines the role of banks in financial markets as intermediaries. While lenders can directly invest into and borrowers can directly obtain money from bond and equity markets, banks help lenders and borrowers access those markets (Allen et al., 2004, p. 491).

In addition, banks can increase or decrease the total amount of money circulation in an economy by lending newly created money from central banks or depositing excess money (Emmerich, 2012).

2.1.3 The banking system in the UK

The UK’s financial system is built on its central bank, the Bank of England. It regulates monetary policies and the banking system, manages national debt and foreign exchange reserves, distributes the currency, and serves as a bank to the government and commercial banks accepting deposits and making loans (Howells & Bain, 2007, p. 52).

In the history of UK banking, two Banking Acts have set the role of the Bank of England following major scandals. The secondary banking crisis of 1973 to 1975 was caused by disproportionally high lending by small banks in relation to their assets to fulfil the need incurred by rising housing demand and prices in prior years (Reid, 1982, pp. 77­79). Due to the 1973 Oil Crisis, housing prices dropped, and houses and land with previously secured loans were worth less than the initial loan, leading banks to near bankruptcy (Reid, 1982, pp. 124-125). This resulted in the Banking Act of 1979, which extended the regulatory power of the Bank of England over banks and created a deposit protection system that demands banks to build reserves of 75% for deposits up to £10,000 (Reid, 1982, p. 197).

In 1984, Johnson Matthey Bankers Limited (JMBL) bank unable to recover £245 million in loans after expanding without significant regulation or even internal supervision. JMBL was not under regulation by the Bank of England because it numbered among the tier-one banks that were not tightly supervised (Moran, pp. 164-165). This changed with the Banking Act of 1987. Initiated because of the JMBL crisis, the Banking Act of 1987 further tightened the regulation by the Bank of England and required all deposit-takers to be authorised (McConnachie, 2009).

The Banking Act of 1998 is the last to mention here. This act placed the Bank of England in charge of determining the key interest rate and handed the task of supervision to the Financial Conduct Authority (FCA) (Rodgers, 1998, p. 93).

The FCA regulates the banking system in the UK by ensuring a certain degree of protection for customers, and probity of the financial market, and emboldening competition benefitting customers (FCA, n.d.a). This is done by requiring a special approval by the FCA for people in controlling functions (FCA, n.d.b, p. 1) or identifying and breaking up potentially harmful structures (FCA, 2018, p. 11).

Today’s Big 4 banks in the UK are Barclays plc (BARC), HSBC Holdings plc (HSBA), Lloyds Banking Group (LLOY) and Royal Bank of Scotland (RBS), which take up 77% of all current accounts in the UK (Connerlly, 2020).

2.1.4 Previous studies on the effect of COVID-19 on financial markets

He et al. (2020, pp. 6-12) investigated the impact the COVID-19 pandemic has had on stock markets as well as spill-over effects on markets that are not directly located in highly affected countries. To do this, they compared the following stock indices: CSI 300 (China), FTSE MIB (Italy), KOSPI (South Korea), CAC-40 (France), SMSI (Spain), DAX-30 (Germany), Nikkei 225 (Japan) and S&P 500 (USA). They found that all investigated markets except the CSI 300 index, underperformed during their chosen event window from 23 January 2020 to 10 March 2020. The authors also showed that these markets performed well before the pandemic and found that there are indeed spill-over effects following events happening in other countries.

Zeren and Hizarci (2020, pp. 81-82) conducted a study to investigate the relationship between daily stock returns of the SSC (China), FTSE MIB, KOSPI, CAC-40, DAX-30 and IBEX35 (Spain) and daily deaths and cases of COVID-19. All markets reacted in a cointegration to total daily death counts, but only the SSE, KOSPI and IBEX35 reacted similarly to the total daily cases, leading to the assumption that death has a higher influence on stock markets than confirmed cases.

Salisu and Akanni (2020, pp. 2311-2314) set themselves the challenge of constructing a global fear index for COVID-19 and analysed the effect that fear and panic has on stock markets. The purpose of their study was to determine how much of the market disturbance can be associated with the pandemic. The authors based the measurement of fear on the mean of reported cases and reported deaths and used a total of 37 stock market indices worldwide. The results indicated that the stock price trends reacted inversely to the fear index in all test countries, revealing a high correlation between the pandemic and worldwide stock market drops and confirmed the study of Zeren and Hizarci (2020).

Bash (2020) conducted an event study to investigate the market reactions after the first confirmed case in a specific country. A total of 30 countries were investigated with the date of their first registered COVID-19 patient. The results revealed significant negative ARs, indicating that the first confirmed case results in abnormally high negative market reactions.

A study by Heyden and Heyden (2020, pp. 4-5) also confirmed the findings of Bash (2020, pp. 35-37) by investigating market reactions of the announcement of the first case and first death. The authors found negative returns following the event date. Collectively, these studies show that COVID-19 impacts stock markets in general and that events happening in other countries impact local stock markets as well due to spill-over effects.

In another paper, Görker, Eren, and Karaca (2020, pp. 23-24) reviewed the ARs of 26 different sectors, including banks, in the Turkish market. They chose the date on which COVID-19 was declared a pandemic by the World Health Organisation (WHO) as the event date and determined a total event window of 41 days, from 20 days prior to, until 20 days after the event. While other sectors such as tourism (-22.28%) and transportation (-27.57%) showed far more distinctive negative ARs, the banking sector also showed an AR of -6.68%.

As seen so far, the COVID-19 crisis has severely affected the economy in all sectors. To help businesses, the government and central banks can use instruments to stabilize, or even boost the weakened sectors. One way of doing this is to lower the value-added tax, as Germany did on 1 July 2020 for a limited time (until 31 December 2020). The goal is to boost consumption with lower end-consumer prices, while the revenue of the selling company remains the same (Bundesregierung, 2020). On a consumer side, low- and middle-income earners would profit the most by lowering the value-added tax (Stefan Bach, Niklas Isaak, & Albert Banal-Estanol, 2017, pp. 629-631).

Banks should profit as well, because struggling businesses may make more sales, more revenue, and more profit, and therefore be more solvent again. This decreases their overall probability of default so that banks do not need to compensate loans in default. To help banks directly, central banks can utilize a method called quantitative easing, where they buy large assets from banks to enable the banks to have more capacities in lending loans . Hartley and Rebucci (2020, pp. 6-7) investigated quantitative easing announcements during the COVID-19 crisis and how they affected 10-year government bond yields of the respective country. The authors found that the impact is predominantly negative, resulting in changes in 10-year government bond yields of -0.23%, -0.29% and -0.31% in a one-day, two-day and three-day windows after the announcement, respectively.

2.2 Global financial of crisis 2007 to 2009

2.2.1 Timeline and causes

The global financial crisis of 2007 to 2009 can be separated into three different components: the housing market, the financial market, and government responses (Fouche, Mukuddem-Petersen, Petersen, & Senosi, 2008, p. 27). Due to low interest rates from 2001 to 2005 in the USA many buildings were either constructed or bought with credit. Banks tended to underestimate the probability of default because of the steady growth over the previous years and believed that a huge recession was not going to happen. This underestimation was reinforced by the assumption that if an economic downfall did occur, the Federal Reserve Bank would step in to protect the market (Goodhart, 2008, pp. 332-333). As a consequence of overbuilding, the housing prices dropped (S&P Dow Jones Indices LLC, 2020) leading to a significant decrease in mortgage cash flow (Financial Crisis Inquiry Commission, 2011, p. 70). Banks had high amounts of debt to maximise the leverage effect during the times of low interest rates. This, in combination with losses on mortgages led to financial difficulties followed bank failures. Lehman Brothers, known for being the biggest investment bank for United States (US) mortgage bonds, was the biggest to fail (Johnson & Mamun, 2012, p. 376). Banks stopped trusting one another and ceased lending money to each other. The total amount of loan yields declined from $700 billion USD in the second quarter of 2007 to under $200 billion USD in the first quarter of 2008 (Ivashina & Scharfstein, 2010, p. 322). The government responses included lowering interest rates, bank bailouts and rescue plans (Fouche et al., 2008, pp. 27-28).

2.2.2 Event study description

This event study follows the method of Fama, Fisher, Jensen, and Roll (1969) to calculate ARs using the market model. Seven banks in the UK and the FTSE350 were selected for retrieving data. In total, an estimation window of 100 days was chosen to determine the factors of the regression that were needed to use the market model. The AR, average abnormal returns (AAR) cumulative abnormal returns (CAR) and cumulative average abnormal returns (CAAR) were tested with a standard t-test to determine their statistical significance. The event window length totals to three days, starting from one day prior to one day after the event date. The following event dates were chosen: 29.09.2008 - Bailout plan rejection²:

To prevent a predicted economic meltdown, the administration of the then president of the USA - George W. Bush - and several lawmakers created a bailout plan to make banks and other financial institutions more willing to lend money again. To do this, the U.S. Treasury would buy assets with a high probability of default totalling to a budget of $700 billion USD. On 29 September 2008, this plan was rejected by the House of Representatives by only 13 votes resulting in a drop in stock market indices of several percent all over the world. As the news had an overall negative impact this, this event is categorised as a bad-news event. 10.03.2009 - Citigroup Inc. made a profit³:

After banks all over the world declared bankruptcy or received huge bailouts by governments a memo from then chief executive officer Vikram Pandit to his employees revealed that Citigroup Inc., an American investment bank and financial service corporation, made a profit in the first two months of 2009. It was the first time since late 2007 that Citigroup Inc. was profitable. This resulted not only in an increase of 11.43% in the price of Citigroup shares but also in an increase in indices worldwide as some analysts saw this as a first sign of the end of the banking crisis. Because of the positive effect that this memo had financially and morally, it is categorised good-news event.

The test sample consists of banks listed on the London Stock Exchange (LSE) in 2008 and 2009. Therefore, this study includes HSBA, LLOY, BARC, RBS, Standard Chartered plc (STAN), Close Brothers Group plc (CBG), and Secure Trust Bank plc (STB), resulting in a total of seven banks. Due to a lack of data, Alliance & Leicester, HBOS, and Northern Rock could not be included, despite being listed on the LSE at the said time.

As this event study is intended to contribute to the development of the hypothesis and is not an element of the initial event study about reaction of banks during news about the COVID-19 pandemic. Therefore, the exact methodology for conducting an event study will be described in a later state of this thesis.

2.2.3 Event study results

The table below shows the banks’ corresponding AR on the event date, AR(0); day prior to the event, AR(-1); and day after the event, AR(+1). The table also shows the CARs for the event date to the day after the event, CAR (0, +1); and the entire event window from the day prior to the day after the initial event, CAR (-1, +1). The significance is denoted as *, ** and *** indicating a significance level of 10%, 5% and 1%, respectively. The corresponding t-values are listed in Appendix 1 and Appendix 2.

Abbildung in dieser Leseprobe nicht enthalten

Table 1: Bad News: 29.09.2008 - Bailout plan rejection

The output table of this event study shows a negative AR on the event date in four out of the seven banks included, with three having a significance level of 1%. The AAR on this day is -3.3681%. The averaged results of the day prior and day after the event show an AAR of -0.0611% and 2.2486%, respectively. This indicates that the event had a disproportionally negative impact on the returns of banks in comparison to the market index, despite not being statistically significant. The increase of the AR from the event date to the day after also shows a recovery of the investigated negative reaction. Still, the sample used achieved cumulative CAAR of -2.7195% over the event window.

Abbildung in dieser Leseprobe nicht enthalten

Table 2: Good News: 10.03.2009 - Citigroup Inc. made profit

The second event study conducted refers to 10 March 2009, when the announcement was leaked that Citigroup Inc. made a profit for the first time in months. On the event date itself, all banks that were taken included recorded a positive AR, ranging from 0.7752% to 9.8139% with a level of significance of at least 5%, totalling to an AAR of 4.6673% at a 1% level of significance. The day prior to the event date shows an AAR of -1.3775%, where five out of eight banks made a negative AR, three with a 1% level of significance. The day after the announcement, the AR of the banks predominantly decreased. Not considering RBS, all ARs on τ + 1 are significant with a level of at least 10%. Despite the considerable decrease in the AAR, the CAAR from τ = 0 to τ + 1 are still positive resulting in 2.8646% in total. The single CAR values are significant at a 1% level, except HSBA. The CAAR in the entire event window showed a change of 1.7659%.

Several findings can be deducted from these two event studies. First and foremost, the small AAR value on τ — 1 and τ + 1 on both studies reveals only minor differences to the market. Further examination of the single values on τ — 1 reveals that despite the rather small AAR the AR of the companies show an unbalanced distribution where no trend can be detected. While some banks achieved a positive AR, others obtained a negative one. On the other hand, the AARs on the event dates reveals positive AARs on good-news events and negative AARs on bad news events.

2.3 Hypothesis formulation

The results of the literature review and the event study on the financial crisis from 2007 to 2009 lead to the following hypotheses will be examined in this study:

Hypothesis 1

H0: UK bank stock prices do not have a more sensitive than the stock market following major events in the COVID-19 crisis.

H1: Stock prices of UK banks have a more sensitive reaction than the stock market following major events in the COVID-19 crisis.

Recalling the studies of He et al. (2020), Heyden and Heyden (2020) and Salisu and Akanni (2020), it can be seen that COVID-19 negatively influenced stock markets worldwide, including spill-over effects to other countries. As banks have the role of financial intermediaries, they interfere with the financial market more directly than other sectors. Therefore, the first hypothesis states that banks have an AR to the stock market following news in the COVID-19 pandemic.

Hypothesis 2.1

H0: Bad-news events do not result in a more negative reaction of bank stocks in UK compared to the market.

H1: Bad-news events result in a more negative reaction of bank stocks in UK compared to the market.

As seen in the study by Görker et al. (2020) the banking sector in Turkey had a significant negative AR when the WHO declared COVID-19 a pandemic. The study performed in this thesis that investigated ARs of banks in UK on 29 September 2008 - the date of the rejection of the US bailout plan - resulted in negative ARs as well. Following this the second hypothesis claims that banks react with a negative AR following bad-news events.

[...]

¹ Indices are normalized to 0% in the beginning of each sample (1 January 2020)

² Articles by CNN Money written by Isidore (2008) (https://money.cnn.com/2008/09/29/news/economy/bailout/) and The New York Times written by Hulse and Herszenhorn (2008) (https://www.nytimes.com/2008/09/30/business/30cong.html) provide proof that the news event was publicly available on the mentioned date.

³ CNN Money (Barr, (2009) (https://money.cnn.com/2009/03/10/news/newsmakers/citi.pandit.fortune/index.htm) and The Guardian (Clark (2009) (https://www.theguardian.com/business/2009/mar/10/pandit-citi-profit-memo )reported about this event on the mentioned date.