Faculty: Economics and Business Section: Economics and Finance
Thesis Specialization: Finance
Did the referendum of Catalonia about independency had effect on the stock
market of Spain?
Abstract
In this study the effect of the referendum of Catalonia about independency on the Spanish stock index is examined. This is done using event studies. Two different events are examined. One for the announcement of the referendum on 9June 2017, and one for the outcome of the referendum on 1 October 2017. Beside there are made three different categories to check if there was a greater or smaller effect for particular firms. The firms are categorized on the basis of if they are active in the financial sector or not, if their fiscal office is located in Spain and on the basis of high or low research and development expenses. The main result was that during both events the Spanish stock prices were indeed affected. On the other hand, financial firms suffered more from the outcome of the referendum than financial firms. The same applies for Catalonian firms with respect to non-Catalonian and firms with low R&D expenses with respect to firms with high R&D expenses.
Author: Mike Dooijeweerd Supervisor: P.M. Golec
2
Statement of originality
This document is written by Student Mike Dooijeweerd who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Table of Contents
1. Introduction……….4
2. Literature review………..5
3. Methodology………..7
3.1 Data description………..8
3.2 Event study………..9
4. Results………11
4.1 Market results……….11
4.2 Financial versus Non-Financial firms………12
4.3 Catalonian versus Non-Catalonian firms………..13
4.4 Firms with High versus Low research and development
expenses……….14
5. Limitation and discussion……… 14
6. Conclusion………..15
7. Appendices……….16
4 1 Introduction
Investors expect that all market information that is available will be reflected in the stock prices (Brickley, 1983). So when for instance the results of a referendum or election are announced, in most of the times this will have effect on the stock prices in the particular country. Given recent studies, it is not predictable in which way the stock returns will be affected due to political outcomes. In this paper the effect on stock prices due to the announcement and outcome of the referendum in Catalonia is examined. The research question in this paper is: ‘Did the referendum of Catalonia about independency affect the stock market of Spain?’
On the 9th of June the prime minister of Catalonia, Carlos Puigdemont, announced that a referendum on whether Catalonia should be independent from Spain or not will be hold. Before it took place, there was many rumour around the referendum if it will go ahead, because the Spanish government tried to block the holding of the referendum. Despite the insistence of the Spanish government, the referendum was held on the 1st of October 2017. Regardless of the outcome of the referendum, the Spanish government will not agree with secession of Catalonia. After the
referendum it was still unclear what will happen if the outcome was that the people of Catalonia voted for leaving Spain (Anderson, 2017). If Catalonia will become independent, this can have a huge impact in the business sector. On the 4th of October more than 15 big companies, and many more small companies already had left Catalonia with their fiscal head office, only due to the uncertainty around, and the outcome of the referendum. There are different reasons for this. If the secession will take place Catalonian firms will have less access to the European market, which will have two
negative sides: foreign investors will be holding back with investing in Catalonia. Thereby it will be more complicated for firms in Catalonia self to reach their markets (Martin, 2017). This declares that there is a lot of economic uncertainty around the event in Spain, especially in Catalonia.
Before the referendum the latest polls found that around 40 per cent of the voters will vote ‘yes’ on the question if they want Catalonia to be an independent state (Balfour, 2017). This was not far away from the 50+ required per cent. The result of the referendum was quite different from the latest polls, 90 per cent of the 2,2 million voters vote for independency (Van der leeuw & Winkels, 2017). The 2,2 million was 42% of the total amount of people which may had been allowed to vote. The turn-out was not very high, because the Spanish government did everything to stop the people from voting. For instance they deployed police on the streets of Catalonia. Since the amount of ‘’yes-voters’’ was so big, it becomes more and more complicated for the government to block the
separation. Catalonia will just put more and more pressure on the Spanish government with the outcome.
In this thesis there is done research on the abnormal returns of the firms listed on the IBEX-35 using classical event study’s. This is done for two different event days. One for the day the
5
referendum was announced and one for the day when the result of the referendum was announced. After the tests for all the companies together, different categories of the companies are made. The firms are categorized in three different ways. The first one is based on the fact if the firm is
participating in the financial sector or another sector (non-financial). In the second there is been made difference between firms with their fiscal office located in Catalonia and firms with their fiscal office located somewhere else in Spain. The last category is about research and development expenses. Firms are categorized in the high or low research and development expenses category. In the next section the literature review is described. Here some previous researches on referenda and their results are mentioned. Beside some practical e ffects of political uncertainty on stock prices are described. In section 3 the methodology is explained. First a description of the used data in this research in section 3.1. In 3.2 the event study is explained. After the methodology the results are shown in section 4. First the results of the general market are shown in 4.1, afterwards the results of the different categories are in the sections 4.2 4.3 and 4.3. In section 5 the limitations and discussion are described. After the limitations and discussion the conclusion on the research is explained in section 6. All relevant tables are described in the appendix, section 7. Finally the bibliography is stated in the last section.
2 Literature review
In this paragraph, first will be discussed what the consequences for Catalonia would be if they become an independent state. Beside previous referenda in Canada and the United Kingdom and the effects on the stock returns will be mentioned. Further some relevant theory about the effects of political and economic uncertainty in a general form will be discussed
Since 90% of the voters in Catalonia vote for independency, Catalonia insist more than ever on independency. But even though it becomes more and more difficult for the government to block the separation, they still hold on that it is not going to happen. And maybe the government is right, because it will have huge consequences for Catalonia if they secede from Spain. If Catalo nia separate from Spain, it have to leave and re-apply the European union, according to the European
Commission. This procedure to become a member of the EU again can take 10 years. All members of the EU, so also Spain, have to agree with the request. Spain already said they will not agree. EU membership and trade have a positive relation, according to Dringha et al. (2016). So in this period without a EU membership the trade for Catalonia will be lower. Another consequence Catalonia should take into account is the currency, will they start their own currency, or will they go on with the Euro if it is allowed. All these uncertainty can provide to that Catalonians will save more money or maybe withdraw their money from the banks. This can lead to inflation.
6
Even for Spain as a whole the it will have consequences if Catalonia separate. Catalonia is the most prosperous region in Spain. Even only 16% of the total population of Spain lives in Catalonia, it makes around 20% of the total GDP of Spain (Tieleman, 2015). Beside Catalonia is the most
important trade hub of Spain. It trades around 26% of all intra-region trade of Spain and 20% of all foreign Spanish trade (Fazekas, 2016). If Catalonia becomes independent from Spain, trade barriers between the two will raise. In addition the capital of Catalonia, Barcelona, is an important business card for Spain for trade and tourism (Tieleman, 2015).
Referenda about independency have taken place earlier in other countries. In 1995 there was the Quebec referendum about independency from Canada (Beaulieu, Cosset, & Essadam, 2006). Beaulieu et al. (2006) investigate the short-run political uncertainty impact on stock returns of firms in Quebec. They categorized the firms in two categories based on exposure to political risk. The classical event study is used to determine the abnormal returns of the firms. They found an positive affect on stock returns after the outcome of the referendum was become clear. The outcome was different from the outcome in Catalonia, the main part of the voters vote for no independency. Beaulieu et al. think that the ‘no’ vote is associated with reduction of political and economic uncertainty, which led to the positive affect. Another results they found from their research, is that the uncertainty around the referendum had a greater effect on firms with more exposure (domestic firms) to political risk than companies with less exposure (multinationals) to political risk.
Another more recent example, is the Brexit referendum. The outcome of the referendum of leaving the European Union was a surprise if you believe the latest polls prior to the vote where was forecasted an outcome to remain in the European Union (Oehler, Horn, & Wendt, 2017). In the paper of Oehler et al. (2017) where the effect of the Brexit referendum is examined, they investigate the short-term abnormal returns around the time of the referendum was held. The expectation was that if the result of the referendum was leaving the European Union the stock prices will immediately fall. The expectation came true when the FTSE100 closed the first trading day after the result with a loss of more than 3%. Thereby the abnormal returns are showing a decline during the event window of the referendum. Another result that was found is the same as Beaulieu et al . The effect on the returns were smaller on multinational firms than on domestic firms (Oehler, Horn, & Wendt, 2017). In general political and economic uncertainty will have an effect on the stock returns, because stock prices reflect all available information in the market (Brickley, 1983). In a study about political uncertainty and risk premium it had become clear that political news dominate the financial market (Pastor & Veronesi, 2013). In the paper of Pastor & Veronesi they argue that when political uncertainty is high, stocks are more volatile and more correlated, especially when the economy is weak. In this case the risk premium should be higher because investors should be compensated for the risk they take. In this research in the period between the announcement and the date of the
7
referendum it is uncertain what the result will be. According to Pastor & Veronesi in this period the risk premium should be higher. Not all stocks have the same correlation with the market. Stocks that are highly correlated with the market, will react more if uncertainty rises in a country than a firm which is not much correlated with the market of that country. In this case firm-specific information will have more impact on the stock price of the firm (Morck, Schleifer, & Vischny, 1990). For instance health care firms are likely to react less on political uncertainty than firms in the financial sector. The stock price not only change due to political uncertainty. There are many factors that can have influence on the stock prices of securities. Some of the most important factor that influence the stock prices are the interest rates, inflation rates, exchange rates and industrial production. In a study of Cox, Ingersoll & Ross (1985) they found a negative relationship between the stock price and the interest rate. The risk-free interest rate reflects the required return on a risk free investment. According to Cox et al. a change in the interest rate will change the required return that investors expect from their securities. Therefore, if the government sets an higher risk-free rate, the interest rate will increase as well, which ensures an increase in the required rate of return. So a rise in the risk free rate will led to an decline in the stock price (Cox, Ingersoll, & Ross, 1985). The second factor, inflation rates, also has negative correlation with the stock prices, according to a study of Bo ucher (2008). He claims that a higher inflation rate, leads to an increase in the required rate of return, which in his place leads to a lower stock price. The next factor is the exchange rate. Stock prices and exchange rates are positively correlated (Phylaktis & Ravazzolo, 1980). Phylaktis & Ravazzolo found that if the exchange rate changes, this has effect on the international competitiveness and trade balance of the country. They found that an increase in the exchange rate makes the import more expensive. Domestic firms products becomes relatively cheaper, which increases their
competitiveness and profitability. In order of the increasing profitability the stock prices increase. The last factor is industrial production. Chen, Roll & Ross (1986) are claiming a positive relation between stock prices and industrial production, because industrial production is the main part of the national production. Thus if the industrial production rises, the national production rises. In order to this the GDP of the country increases. As a result of the increasing GDP the good’s demand in the country increases, and so the profitability. It becomes more profitable for investors to invest, which lead to an increase in stock prices.
3 Methodology
This chapter will describe the methodology and data which are used during the research. In the first section there will be described which data is used and where the data is obtained. Afterwards the event study, which is used to investigate the effect of the referendum, is extensively described.
8 3.1 Data description
In this research, the Spanish stock market is the market which is examined. Therefore, the IBEX-35 index is used. The IBEX-35 is chosen because it is part of ‘Bolsa de Madrid’, the biggest stock
exchange of Spain. On the IBEX-35 list are the 35 most important companies of Spain. It gives a good indication of the market in Spain. Therefore it is chosen that the IBEX-35 represent the market in this paper. The composition of the 35 companies takes place on the basis of free float market
capitalisation and trade volume. The maximum weight of a company on the list is 20%. The financial sector represent one third part of the index. The energy sector is on the second place with 20% of the index.
In order to test if there was an effect on the Spanish stock market due to the referendum, an event study will be used. To do the event study the stock prices of the firms of the IBEX -35 are obtained from Thomson one. Because an estimation period is needed to calculate the normal performance of the firms, the obtained data starts 250 days before the first event took place (announcement referendum). Usually between 100 and 300 days are enough for the estimation period to predict the normal performance of a firm (Fama, Fisher, Jensen, & Roll, 1969). So in this research, the timeline starts on 23 September 2016. The first event, announcement of referendum, is on June 9th 2017. The second event, result of referendum, is on 2 October 2017. The time line ends in the days after the second event (showed in the figure below).
After the test for the whole index, the firms are categorized on particular characteristics. There is tested if the events had more or less effect on some particular firm characteristics. The firm characteristics are obtained from Wharton WRDS. The firms are categorized in three different ways. The first one is the sector in which the firm is active. There is made a financial category and a non -financial category. This one is chosen because in some previous studies about political uncertainty and the effect on stock prices there are found significant differences between financial and non -financial firms. It will be tested if this was the case in Spain also. The second category has to do with the location of the fiscal office. There is been made a difference in fiscal offices located in Catalonia, and fiscal offices outside of Catalonia. Despite there are only three fiscal offices located in Catalonia, this is an important thing to investigate. This category is obvious. The consequences for Catalonian firms will be larger than for non-Catalonian firms if the separation takes place. With the outcome of 90% advantage for separation, the expectation is that the Catalonian firms are affected more. The last category is about the research and development expenses of the firms. The research and development expenses are normalized. The median is chosen to be the split-off to categorize the
9
firms in high or low research and development expenses. This one is chosen to check whether high R&D expenses indeed can have a positive effect on your firm. In the sense of, the firms with high R&D expenses have done research on the effects a ‘yes’ outcome will have and therefore they have done some adjustments to minimize these effects. In this case the low R&D expenses category will be affected more during the outcome of the referendum than the high R&D category.
In the next section the procedure of conducting an event study will be extensively explained.
3.2 Event study
First thing to do if you are going do to an event study, is to define the event of interest. After that identify which firms are involved in the event. Then determine the period over which the stock prices of the firms will be examined (Mackinlay, 1997). According to Mackinlay (1997) the period of interest always reflect the day of the event and the day after the event. The day before the event may also be of interest. In this research the period of interest consists of three days, one day before and one day after the event. This is called a three-day event window (-1, +1). Identifying which firms are involved in the event can be determined by restriction, such as firms listed on an index or activity in a
particular industry.
To estimate the impact of an event on a firm a measure of the abnormal return is required. According to Mackinlay (1997) the abnormal return is measured as follows:
𝐴𝑅𝑖𝑡= 𝑅𝑖𝑡− 𝐸(𝑅𝑖𝑡|𝑋𝑡)
Where 𝐴𝑅𝑖𝑡 is the abnormal return and 𝑅𝑖𝑡 is the actual return for time period t. 𝐸(𝑅𝑖𝑡|𝑋𝑡) represent the estimated return for time period t. 𝑋𝑡 is a variable which contains the information about the normal return model. There are two common normal return models, the constant mean return model and the market model (Mackinlay, 1997). In the constant return model 𝑋𝑡 is a constant. This model assumes that the mean return of a firm is constant trough time. In the market model 𝑋𝑡 is the
10
market return. The market model, which is used in this research, assumes a stable relationship between the market return and security returns. If the model for normal performance is chosen, the normal performance is identified using an estimation period. As mentioned earlier a good estimation period lies between 100 and 300 days. Usually the estimation period does not overlap the event window, so that the event returns do not influence the normal performance returns. Because the market model predicts that a firm’s return has a relation with the market return, the estimated return of the firm can be calculated based on the assumption of rational expectations and the efficient market hypothesis (Fama, Efficient Capital Markets, 1991). This leads to the following:
𝐸(𝑅𝑖𝑡) = 𝛼 + 𝛽 ∗ 𝑅𝑚𝑘𝑡+ 𝜀, where 𝐸(𝜀) = 0 𝑎𝑛𝑑 𝑣𝑎𝑟(𝜀) = 𝜎2(𝜀𝑖𝑡).
𝑅𝑚𝑘𝑡 is the return of the market index, in this research the IBEX-35 return. The parameters α and β are estimated through an OLS regression over the trading days of the estimation window. After the parameter estimation and the associated predicted returns the abnormal performance can be calculated. Under the null hypothesis, that the referendum had no impact on the stock returns, the following applies: 𝐴𝑅𝑖𝑡~𝑁(0,𝜎2(𝜀
𝑖𝑡)). To analyse the effect on a multiple period event window, the one day abnormal return is not a good indication. Therefore the cumulative abnormal return is calculated. The cumulative abnormal return is the sum of the abnormal returns in the event window: 𝐶𝐴𝑅 = ∑𝑡=2𝑡=1𝐴𝑅𝑖𝑡 where t=1 represent the first day and t=2 the last day of the event window. To check if the abnormal returns of the firms are statistically different from zero the t-statistic is calculated as follows: 𝑡 = 𝐶𝐴𝑅𝜎𝐴𝑅𝑖
√𝑁
, where N is amount of days of the event window (N=3), and 𝜎𝐴𝑅 is the standard deviation of the abnormal returns. The t-statistic is two tailed tested, which give the following rejection zones for different significance levels (table 1).
Significance α(%) Critical t-value 2-tailed 1% ±2.576 5% ±1.960 10% ±1.645 20% ±1.282
Table 1: Critical t-values for different significance levels
11 4 Results
In this section the results of the event studies will be discussed. First the tests on the whole market, and afterwards the tests on the different categories.
4.1 Market results
In order to test the hypothesis that the announcement and the outcome of the referendum both had effect on the stock market of Spain, event studies were conducted. During the first event
announcement of the referendum as you can see in table 2, the average actual returns one day before the event day was around 0.10%, at the event date 0.18% and the day after it was -0.85%. We see that the average one trading day after the announcement was negative.
Day before Day of announcement Day after
Average actual return 0.10% 0.19% -0.85%
Table 2: actual average returns during the announcement of referendum.
Because the actual returns and the predicted returns were calculated, the abnormal returns could be calculated. Table 3 shows an overview of the average abnormal returns of the market of Spain during the announcement of the referendum.
Day before Day of announcement Day after Average abnormal
returns
-0.046% 0.0032%** 0.051%
Table 3: Average abnormal returns during announcement of referendum. **Abnormal returns significant for α=5, 10 and 20%
We see negative average abnormal returns on the day before. The day of the announcement and the (trading) day after there were positive average abnormal returns. If we test the abnormal returns, we see in table 4 that the results were significant at a 20%, 10% and 5% level (table 6). Therefore we can conclude that the results on the abnormal returns are significant, and indeed the announcement of the referendum had an effect on the Spanish stock market. However, in table 3 is shown that the effect is not very large.
The same as above is done for the second event, the outcome of the referendum. This event took place on the 2nd of October. On the first trading day before the event, 29 September, the actual average return was 0.45%. On 2 October we see a negative return of -0.87%, and the day after a positive 0.19%.
12
Day before Day of outcome referendum
Day after
Average actual return 0.45% -0.87% 0.19%
Table 4: actual average returns during the outcome of referendum.
With the Actual returns and predicted returns again the abnormal returns are calculated. In table 6 the overview of the average abnormal returns during the event window for the outcome of the referendum.
Day before Day of outcome referendum
Day after
Average abnormal returns
0.054% 0.015%* 0.016%
Table 5: average abnormal returns Market of Spain during outcome of referendum. *Abnormal returns significant for all levels of α
The average abnormal returns for the three-day event window are all positive. If we test the abnormal return with a t-test, we find that for all levels of significance we can conclude that the hypothesis was right, the outcome of the referendum of Catalonia had an effect on the stock prices of Spain (table 6). Well, it’s surprising that the abnormal returns during the event window are positive. Here also, there is an effect, but the effect on the abnormal returns is not very large. Further research is required to investigate why the abnormal returns aren’t negative during these days.
If we compare table 3 and 5, we see that the effect during the outcome is a little bit larger than during the announcement. Therefore, for the next section where the results of the different categories are shown, only event two, the outcome of the referendum is examined. This is enough to check whether such a referendum had more effect on particular firms. Since the sample is a bit small, there is looked more on the economical magnitude of the results than on the significance. The significance can be seen in tables. First the results of the part were financial firms are tested against non-financial firms.
4.2 Financial versus non-financial firms
In this part the firms are categorized in a category of financial firm or non-financial firm. There were 7 firms that are participating in the financial sector. So 28 firms are non-financial firms. In table 7 the average abnormal returns for the different sectors are shown. As you can see the abnormal returns for the financial firms were negative the day before and the day of the event. The day after they
13
were positive. For the non-financial firms the average abnormal returns were positive during the whole event window. Also the significance of the results is shown in table 7. If we compare the results of the financial and non-financial sector, we see that the effects on the financial sector are relatively larger than on the non-financial sector. Thereby, there were negative abnormal returns for the financial sector. From this we can conclude that there was an effect on both sectors, but the effect for the financial sector was more unpleasant. There can be certain reasons for this. One of them is that the non-financial sector includes different sectors, like the industrial-, energy- and healthcare sector (and more). For instance the healthcare sector will not be negatively affected by such an event as an outcome of a referendum because it is a necessity of life. There are more sectors which are not or not much affected by political outcomes. With further research, the firms can be categorized in more different sectors, to check for which sector such a referendum is a danger. Another fact is that two of the 7 financial firms are located in Catalonia. In the next section becomes clear that this is negatively correlated with the outcome of the referendum. This also can be a reason for the difference with the non-financial sector. Well, for the non-financial sector there is one Catalonian firm, but this is the firm in the healthcare sector, and described above this firm will not be affected by such events. To take account of all these factors, further research is required.
4.3 Catalonian versus non-Catalonian firms
This part is about the Catalonian and non-Catalonian firms. Here the firms are categorized on the basis of if there fiscal office is located in Catalonia or somewhere else in Spain. After some research it is become clear that the fiscal office of only 3 of the 35 firms listed on the IBEX-35 are located in Catalonia. It’s about the companies Banco Sabadell, Caixabank and Grifols. If we look at the average abnormal returns for the Catalonian firms in table 7, the day before the referendum and the day of the referendum the average abnormal returns were negative. The day after they were positive . This can be due to uncertainty in the days before and of the referendum. The abnormal returns of the firms outside of Catalonia were positive through the event. The results are significant (table 7). The average abnormal returns for Catalonian firms are negative in two of the three event days. Since the average abnormal returns for firms outside Catalonia were positive, we can conclude that the firms in Catalonia suffered more from the event. This is not really surprising. In the literature review already some consequences if Catalonia separate are described. Since advantage for leaving was so big, the change they will separate increased. Therefore the consequences like Catalonia also leaves the EU, trade barriers increase, there will be set import tariffs become more realistic. All these things have a negative influence on firms in Catalonia. For the non-Catalonian firms there will not change very much. Therefore, the effect on Catalonian firms was more negative than on non -Catalonian firms.
14
4.4 Firms with high versus low research and development expenses
In this section only 14 of the 35 firms are investigated. This is because the research and development expenses of some firms, especially banks, are not known or that low that they are not relevant to take account of. Because the research and development expenses are in millions, and big firms probably spend a more amount in millions in research and development than small firms, the
expenses are normalized. This is done by dividing the R&D expenses by the current assets of the firm. To categorize the 14 firms in the high or low category, the median is used. In Appendix B you can see which firms are participating in this section. Further you can see in which category they belong. What becomes clear of the results, is that the average abnormal returns for the high R&D category firms are positive during the three days event, and for the low R&D category they are negative (table 7). Again the results were significant. The results show in table 7 show that the low R&D firms suffered more from the outcome of the referendum than the high R&D firms. This is an interesting result. Maybe it can be explained because the firms with high R&D expenses did research of what the consequences will be if Catalonia is going to separate. Maybe this firms already made some adjustments, so that no matter the outcome they will not be negatively affected.
5 limitations and discussion
In the largest sense this research gives a good indication of the effect of the Catalonian referendum on the Spanish stock market. However there are some limitations found during the research. One of the limitations has to do with the categorization of the firms into Catalonian and non-Catalonian firms. There is mentioned that the non-Catalonian firms are located with their fiscal office
somewhere else in Spain. But the fiscal office of the company Arcelor Mittal is located in Luxemburg City. This is not in Spain. Perhaps the Company should be ignored in this section.
Another limitation is that there were only three Catalonian firms. Despite there was an effect on the firms, maybe this is not enough to generalize this for the whole Catalonian market. Add ing some other firms located in Catalonia with their fiscal office, which are not listed on the IBEX -35, could be a solution here. The same thing for the parts with the research and development expenses and the Financial sector. In the part with the R&D expenses only 14 firms are investigated since the data is available for the other firms. For the financial sector there were only 7 firms against 28 non -financial firms. Also here maybe some other firms which are not listed on the IBEX -35 should be added to make a general conclusion.
Further, in the category with financial and non-financial firms, there is only been made a difference between the finance sector. Maybe to get a better indication of which sectors are affected the most during the Catalonian referendum, more sectors should be investigated here.
15 6 Conclusion
The main question of this research was: Did the outcome of the referendum of Catalonia about independency had effect on the stock market of Spain? This is investigated using event studies o n the abnormal returns of the firms on the IBEX-35. The hypothesis was that the referendum had indeed an effect on the stock market of Spain. From the results in this paper we see that the hypothesis was right. Both for the announcement as the outcome of the referendum the results were significant. But the abnormal returns during the events were not always negative. So we cannot conclude that the referendum had an unlikely effect on the Spanish stock market. However, there was found that the effect during the outcome of the referendum was larger than during the announcement.
Another part of the paper was categorization of the firms and check whether the effect differ for the categories. First financial versus non-financial firms. On both categories there was found a significant effect. Since there were found negative abnormal returns in the financial sector in the first two days of the event, and only positive abnormal returns for the non-financial sector, we can conclude that the financial sector suffered more from the outcome of the referendum. This can be due to that the consequences if Catalonia will separate will have more impact on the financial firms. The next category was about firms with their fiscal office located in Catalonia or out there. For both categories there was enough evidence that there was an effect. For the Catalonian firms the abnormal returns were negative during the first two days of the event. Probably this is becaus e there was a lot of rumour in Catalonia during the referendum. The day after the abnormal returns became positive. This can be due to that the Spanish government still holds on with saying that no matter the outcome of the referendum, Catalonia is not leaving Spain. For the non-Catalonian firms the
abnormal returns were positive during the event. We can conclude that the Catalonian firms suffered more from the referendum. This was in line with the expectations.
The last category was based on research and development expenses. The firms were classified in high and low R&D expenses. The result was interesting. For the low R&D expenses category the abnormal returns were negative during the whole event window. For the high R&D expenses category they were positive during the even window. So in this case, the firms with relatively low R&D expenses suffered more from the referendum than the firms with high R&D expenses. Maybe this is not really surprising, but it is surprising the differences were quite big. It looks like the firms with relatively high R&D expenses had done research on what the effects will be if the outcome of the referendum will be a separation and had in advance made some adjustments, so that they will not be negatively affected. Further research is research is required to check whether this was the case or not.
16 7 Appendices
Abertis Gamesa
Acciona Gas natural
Acerinox Grifols
Actividades construccion y servici IAG
AENA Iberdrola
Amadeus Inditex textil
Arcelormittal Indra sistema
Banco Bilbao (BBVA) inmobiliara colonial
Banco Sabadell Mapfre
Banco Santander Mediaset
Bankia Melia
Bankinter Merlin
Caixabank R.E.C.
Cellnex Repsol
Dia Tecnicas Reunidas
Enagas Telefonica
Endesa Viscofan
Ferrovial
17
Firm name Fiscal City Sector
R&D expense/current assets
Abertis Madrid Non-Financial .
Acciona Alcobendas Non-Financial .
Acerinox Madrid Non-Financial 0.00746 (high)
Actividades construccion y
servici Madrid Non-Financial .
AENA Madrid Non-Financial 0.000262 (low)
Amadeus Madrid Non-Financial 0.2672 (high)
Arcelormittal Luxemburg city Non-Financial 0.0107 (high)
Banco Bilbao (BBVA) Madrid financials .
Banco Sabadell Barcelona*. financials .
Banco Santander Madrid financials .
Bankia Valencia financials .
Bankinter Madrid financials .
Caixabank Barcelona* financials .
Cellnex Madrid Non-Financial .
Dia Madrid Non-Financial .
Enagas Madrid Non-Financial 0.00105 (low)
Endesa Madrid Non-Financial 0.00294 (low)
Ferrovial Madrid Non-Financial .
Gamesa Zamudio Non-Financial 0.0101 (high)
Gas natural Madrid Non-Financial .
Grifols Barcelona* Non-Financial 0.0633 (high)
IAG Madrid Non-Financial .
Iberdrola Bilbao Non-Financial 0.0197 (high)
Inditex textil A Coruna Non-Financial .
Indra sistema Alcobendas Non-Financial 0.000648 (low)
inmobiliara colonial Madrid Non-Financial .
Mapfre Majadahonda financials .
Mediaset Madrid Non-Financial .
Melia
Palma de
Mallorca Non-Financial .
Merlin Madrid Non-Financial .
R.E.C. Alcobendas Non-Financial .
Repsol Madrid Non-Financial 0.00458 (low)
Tecnicas Reunidas Madrid Non-Financial 0.00087 (low)
Telefonica Madrid Non-Financial 0.0454 (high)
Viscofan Tajonar Non-Financial 0.00481 (low)
Appendix B: Firm Characteristics
18 Abertis
Acciona Acerinox
Actividades construccion y servici AENA
Amadeus Arcelormittal Banco Bilbao (BBVA) Banco Sabadell Banco Santander Bankia Bankinter Caixabank Cellnex Dia Enagas Endesa 0.5516 (0.041) 0.0648 (0.048) 0.7023 (0.077) -0.0189 (0.061) 0.583 (0.050) 0.5362 (0.049) -0.1373 (0.138) 1.408 (0.0492) 1.398 (0.067) 1.0084 (0.0116) 1.356 (0.814) 0.8702 (0.0396 1.414 (0.0636) 0.5775 (0.0577) 0.735 (0.0568) 0.4699 (0.0465) 0.512 (0.0464) Ferrovial Gamesa Gas natural Grifols IAG Iberdrola Inditex textil Indra sistema inmobiliara colonial Mapfre Mediaset Melia Merlin R.E.C. Repsol Tecnicas Reunidas Telefonica Viscofan 0.785 (0.050) 0.697 (0.079) 0.758 (0.043) 0.439 (0.059) 1.504 (0.321) 0.724 (0.0423) -0.00646 (0.0589) 0.720 (0.079) 0.673 (0.0584) 0.998 (0.056) 0.782 (0.061) 0,778 (0.0515) 0.834 (0.057) 0.479 (0.0433) 0.923 (0.0492) 0.578 (0.070) 1.201 (0.0383) 0.340 (0.060) Appendix C: Estimated beta of each firm. Between brackets the standard error.
19
Announcement: t-value=-2.49 Outcome: t=5.3
α=1%: Do not reject 𝐻0 α=1%: Do reject 𝐻0
α=5%: Do reject 𝐻0 α=5%: Do reject 𝐻0
α=10%: Do reject 𝐻0 α=10%: Do reject 𝐻0
α=20%: Do reject 𝐻0 α=20%: Do reject 𝐻0
Table 6: Hypothesis rejection or not during the two events for the market.
Day before Day of outcome referendum Day after Financial sector -0.15% -1.4%*** 0.49% Non-financial sector 0.010% 0.31%* 0.010% Catalonian Firms -0.515% -1.74%* 0.41% Non-Catalonian firms 0.11% 0.18%* 0.13%
High R&D expenses 0.34% 0.62%* 0.90%
Low R&D expenses -0.18% -0.36%* -0.087%
Table 7: Average abnormal returns during event window of the outcome of the referendum. *Abnormal returns significant for all levels of α
***Abnormal returns significant for α=10% and α=20%
t-statistic Financial sector -1.93 Non-financial sector 6.83
Catalonian Firms -2.95 Non-Catalonian firms 19.9
High R&D expenses 3.85 Low R&D expenses -7.97 Table 8: t-values of the tested sectors.
20 8 Bibliography
Anderson, P. (2017, June 12). Catalonia’s independence referendum: The stage is set for yet
another political and legal battle . Europian Politics and Policy.
Balfour, S. (2017, August 1). Catalonia and Spain: will the referendum on indepence go
ahead? Europian Politics and Policy.
Beaulieu, M.-C., Cosset, J.-C., & Essadam, N. (2006). Political uncertainty and stock market
returns: evidence from the 1995 Quebec referendum. Canadian Journal of
Economics, 621-642.
Boucher, C. (2008). Stock Prices, Inflation and Stock Returns Predictability. SSRN Electronic
Journal.
Brickley, J. (1983). Shareholder wealth, information signaling and the specially designated
dividend: An empirical study. Journal of financial economics, 187-209.
Chen, Roll, & Ross. (1986). Economic forces and the stock market. The journal of business,
383.
Cox, J. C., Ingersoll, J., & Ross, A. (1985). A theory of the term structures of interest rate.
econometrica, 53(2), 385.
Den Haan, W., Ellison, M., Ilzetski, E., McMahon, M., & Reis, R. (2016, februari 25). Center for
Macroeconomics. Retrieved from cfmsurvey.org:
http://cfmsurvey.org/surveys/brexit-and-financial-market-volatility
Dhingra, Ottaviano, Sampson, & Reenen. (March 2016). The consequences of Brexit for UK
trade and living standards.
Fama, E. (1991). Efficient Capital Markets. The journal of finance, 1575.
Fama, E., Fisher, L., Jensen, M., & Roll, R. (1969). The Adjustment of Stock Prices to New
Information. SSRN Electronic Journal.
Fazekas, S. G. (2016). The economic consequences of Catalonia's independence from Spain.
Mackinlay, A. (1997). Event studies in economics and finance. Journal of economics
literature, 13-39.
Martin, W. (2017, September 5). Als Catalonië zich afscheidt van Spanje zijn de gevolgen
misschien wel erger dan de Brexit. Business Insider. Retrieved from
https://www.businessinsider.nl/catalonie-onafhankelijk-spanje-brexit-europese-unie/
Morck, Schleifer, & Vischny. (1990). The stock market and investment: Is the market a
side-show? Brookings papers on economic activity, 157-215.
21
