Market reactions to investment and divestment announcements in the European football industry

(1)

announcements in the European football industry

University of Groningen

Faculty of Economics and Business

MSc: Business Administration, Finance

Profile: Corporate Financial Management

January 2009

Author: Floris de Kort

Hereweg 44

9725 AE Groningen 06-44400792

fdekort@hotmail.com

Student ID: 1384783

(2)

Abstract

This study examines the market reactions of the stock market to investment and divestment announcements in the European football industry. The sample consists of 257 events, football transfers, from twenty listed, European, clubs in the period 01/2001 - 08/2008. The methodology used is an event study. The results show that for the whole sample significant positive abnormal returns are created. These significant results confirm the shareholder maximization theory in football. Both buying and selling transfers create positive abnormal returns, but solely the selling results are significant. Cross-sectional analyses show significant findings for transfers with players playing for their national team and for transfers with higher ranked clubs. These findings confirm the rational view of investors, who are looking for investments that give more certainty for generating immediate future cash flows. Interaction analysis results in a positive coefficient for the relationship between buying transfers and the quality of the other club involved. This indicates the ability of the market to distinguish between good and bad investments. The results on the interaction analysis show no confirmation of the growth opportunity theory in football. In general this study finds, despites the lack of evidence for the growth opportunity theory, support for some rationality among investors in European football clubs.

Key words: Market reaction, football, investment announcements, event study, stock returns

(3)

1. Introduction

Nowadays football clubs have become large, internationally orientated, corporations. Manchester United, the most valued football club in the world, had in 2006 an estimated value of $1.373 billion (Forbes, 2006). Furthermore, in the last years, the transfer prices of players have rosen significantly. In England in 2007, all clubs participating in the four professional football competitions together set a new record of total budget spend on transfers in one transfer window. Set on 741 million Euros, the old record of 450 million was easily beaten (Voetbal International, 2007).

This research will focus on the announcements of these transfers of football players. In this study, an announcement of a transfer will be seen as an investment announcement or capital expenditure, when it relates to a player being bought by a specific club, or as a divestment, when it concerns players being sold by the selected clubs. Since the world of sport is a world of emotions, the rationality of the investors in football is investigated. This research will be done in the light of the observed rationality by investors in others businesses and the irrationality that surrounds football stocks (Edmans, 2007; Duque and Ferreira, 2005).

Shareholder maximization theory (Afshar et al., 1992; McConnel and Muscarella, 1995; Woolridge and Snow, 1990) would indicate positive reactions to both selling and buying transfers, since the intention of the managers of the firms is to maximize the shareholder value. Furthermore, these announcements would be seen as new information for the market (Chan, 1990). Therefore, the market would react positive to both investments and divestments.

(5)

on the contrary, will then be valued higher than the divestment of a younger player since with the loss of a younger player, his growth potential is lost.

Other literature (Chan et al., 1995) shows the ability of the market to distinguish between good and bad investments of firms. When a firm makes a bad investment, the stock return of the involved firms decreases. For a good investment, on the other hand, the stock returns increases. When linked to this paper, this would indicate that investments in players that come from better clubs based on the statistical all time club ranking of IFFHS and players that already have made their appearance in their national team, would be valued higher than those players that do not have achieved that yet. Besides this, the theory of Burton (1999), relates positive market reactions to the cash flow generating possibility of the investment. Both national players and players from better clubs seem to give more certainty for cash flows, since these players already have performed on a high level.

In this paper the assumed rationality is tested. The described factors, age of player, whether it is a nationality player or not and the quality of the club, will be tested and the reaction of the market to certain investments will be evaluated. This will be done by performing an event study, followed by cross sectional, interaction and explanation variable analysis.

(6)

The sample of this study consists of the events of twenty different listed football clubs from seven different European countries. The results obtained from this study, however, should be interpreted with some caution, because of the limited sample size of 257 events only.

The results of this paper show certain rationality among investors. Both selling and buying transfers show positive cumulative abnormal returns, but these numbers are only significant for selling transfers. Transfers involving national players and better clubs are higher valued than the transfers with non national players and worse clubs. The correlation between the sale of a player and the age of the specific player is found to be negative. This would indicate that the market rather keeps its older players in the club than sell them to other clubs. However, in that same analysis a negative coefficient, -0,0001, is found for the buying transfers, which confirm, the growth opportunity theory. Since the coefficient is so small and the result insignificant, this result does not find any support for the growth opportunity theory in football.

(7)

2. Literature Review

This literature review can be split into three different cornerstones. First of all, existing literature on the market reaction to investment and divestment announcements will be discussed. This will be followed by a description of the stock market reactions of listed football clubs on certain events. Subsequently, the third part of the literature review consists of the link between investor behavior and on-field sport results.

2.1 Investment and divestment announcements

Former studies have researched the effects of capital expenditure and investment announcements. A study of McConnell and Muscarella (1985) found evidence for the value maximization hypothesis. This hypothesis states that, due to market forces, managers make capital expenditure decisions that maximize the value of the shareholders. In their sample of industrial firms, announcements of capital expenditure are associated with significant increases in market value of the stocks. Furthermore, announcements of decreases in planned capital expenditure are associated with statistically significant decreases of that market value. These reactions are consistent with the motive of managers to maximize the value of the firm by solely investing in positive net present value projects.

(8)

In earlier research, Chan et al. (1990) concluded that high-technology firms are associated with significant positive abnormal returns, on the market for the announcement of increasing R&D expenditures. Apparently, the announcements are seen as new information that has a positive impact on the share value. Furthermore, these positive returns apply for firms that suffer earning declines. This suggests that investors look further than the short term earnings and have patience when it comes to strategic investments and the valuation of a firm. Cross-sectional analyses reveal that R&D intensity is positively correlated to the stock price reaction, whereas other variables, like the size of the announced investment, industry concentration or the firm’s dominance in its industry, explain less of the differences in the cross-sectional analysis.

In a more recent study, Jones et al. (2004) find generally positive, but small, market-adjusted abnormal returns for investments. They make a distinction in types of investments and find the level of abnormal returns to vary according to these different types. In particular, they find that the market reacts more positive to investments that ‘create’ growth options, than to investments that are labeled as ‘exercising’ investment opportunities. Besides this difference, they find a variable reaction when the size of the company is taken into account. Large companies tend to experience smaller responses to announcements than small companies do. Project size is also a factor that has a significant positive impact on the level of abnormal returns generated by the market after an announcement.

Research by Brailsford and Yeoh (2004) also confirms this growth theory. Their research provides an examination of the market response of capital expenditure in the context of agency problems, created by differences in growth and free cash flow environments. Their results show that growth opportunities are singularly the most important variable for explaining the market reaction to physical asset expenditure announcements. Cash flow itself seems to be a less relevant variable in this context of growth and it has a more interacting role with growth such that free cash flows become relevant.

(9)

announcements depends on two factors. The first factor is whether the project is carried out by a single firm or in a joint venture structure, and the second is whether the investments generate cash immediately or after a period of time.

Their results indicate that only announcements of joint ventures have positive market reactions. According to the researchers, these announcements surprise the market and thus the prospects of the firm. In their sample, the size of the firm does not influence the abnormal return, which indicates that the market is not differently surprised when news of investments is issued by either large publicly traded firms or smaller firms.

In contrast with the results of Burton et al. (1999), Akbar et al. (2008) find the market to react positive to capital expenditure announcements, regardless of the type of investment or project. Investors reassess the market value of a firm that makes a public announcement of capital expenditure. The results confirm the theory that managers seek to maximize the market value of the firm by making capital expenditure decisions. This research solely investigates the reactions to the announcement and ignores whether the intended investments were executed or not.

In a study of Chung et al. (1998) the impact of corporate capital expenditure decisions on share prices is examined. They argue that the reaction of the market depends more on the quality of the investment opportunity than on the type of investment or the type of firm. An increase in capital spending results in positive market reactions among firms with a Tobin’s Q larger than unity. Announcements of increases in capital expenditure, by high-technology firms, result in negative share price changes when the market perceives the quality of the investment to be poor. Announcements of decreases in capital spending by low-technology firms have a negative impact on the share price. These results are in line with the belief that growth prospects are more important for the market reaction than industry affiliation is.

(10)

as they need to justify their expenditures. This could lead to a negative stock return, subsequent to the investments, when investors fail to appreciate the management’s incentive to oversell their firms in these situations. This negative reaction of the market is consistent with the hypothesis that investors tend to under react to the empire building implications of increased investment expenditures.

A specific event study on the impact of IT-investments (Dos Santos et al., 1993) reveals that, on average, by the whole sample, as well as by industry specific samples, these investments are zero net present value investments. There is no specific reaction of the market to these announcements and no abnormal returns were found. Cross sectional analysis, however, shows that solely those announcements in the sample that regard innovative investments, increase the value of the firm and create a return. This result links innovative IT-investments to competitive advantages for firms.

Instead of investments, Afshar et al. (1992) study the impact of announcements of divestments by firms on the market. They find a positive market reaction to the announcement of a divestment, which is in line with the shareholder value maximization theory. Abnormal returns are highest when the completion of the sell-off is announced and the price declared. If only the intention to divest is announced and no specific price is mentioned, there is no significant market reaction. The relative size of the divestment has a direct positive relationship with the magnitude of the abnormal return.

A comparative study of Wright and Ferris (1997) finds contradictive results. In their South African sample, divestment announcements are followed by negative excess results. These results support the premise that noneconomic pressures may influence managerial strategies rather than value-enhancement goals. Divestment decisions might be motivated by the self-interest of senior managers, which represents the manifestation of an agency problem.

(11)

after the announcement of the divestment. They find, however, a significant difference between healthy and financially distressed firms, indicating that less healthy firms are more pressed for cash to ease their liquidity problems.

This overview of conducted research on market reactions to announcements of investments and divestments shows different results. Most studies find positive market reactions to such announcements, where specific factors influence these positive reactions. Growth opportunities are valued positively by the market, whereas empire building and self-interest of managers are valued negatively.

2.2 Stock market returns of listed football clubs

Several researches examine the different external factors which influence stock returns of listed sport clubs. In many of these studies, the conclusion is that the results of sport games are seen as new information to the market, which should and will be incorporated by the market into the share price.

Scholtens and Peenstra (2008) perform an event study on the results of eight football clubs across Europe and their relationship with stock prices. They find that the stock market responds positively to victories of the specific clubs and negatively to losses. The responses to losses are ‘stronger’ than the responses to wins. More important games, like international European matches, have stronger results than regular national league matches. In addition to this, their last finding is a stronger reaction of share prices to unexpected results in international matches than to expected results in these matches. This phenomenon was not applicable for national games, since the reaction to unexpected and expected results was the same there.

(12)

shows a stronger reaction to a loss than to a win. Besides this, for most clubs in the samples of both Scholtens and Peenstra (2008) and this study, the expectation of the market is to win every single match, since it concerns the best football clubs in the world (IFFHS).

Renneboog and Vanbrabant (2000) perform an event study on eight British football clubs. They study the weekly performances of the clubs on the field and the influence of these game results on the stock price return. They find a significant abnormal return of approximately 1%, when the result was a victory. For losses and draws however, they find negative abnormal returns of respectively 1,4% and 0,6%, indicating a stronger reaction to losses than wins, which is in line with the prospect theory.

A study of Batyrbekov (2007) gives the same results. In an event study with over 2000 events, he concludes that investors, considering the expected reaction, overreact to the performance of the club on the field. Batyrbekov (2007) discovers a significantly positive abnormal return of 1,1% after a win and a significant negative abnormal return of 1,2% after a loss. These results are even stronger for important international games.

Palomino et al. (2005) incorporate the betting odds, given by broker Ladbrokes, in their study on professional football clubs listed on the London Stock Exchange. Their results show a fast reaction to the news of game results, especially when these results were positive. They find no evidence for abnormal returns on the trading days following the release of betting information, although this information is a very good prediction of future performance on the field. Their findings are consistent with theory that indicates an under reaction to public information and the impact of the level of salience of information on the speed at which the financial markets process information.

Besides football, in other sports, researches on sport performances and their link to stock performance are conducted as well.

(13)

an asymmetrically response of investors to wins and losses. Losses significantly influence the share price, where wins do not. This is again in line with the prospect theory discussed above (Tversky and Kahneman, 1979). In important playoff games a stronger reaction to either wins or losses is discovered compared to the reaction on regular-season games.

A research by Gannon et al. (2006) is closely related to the football stock market reaction on events. They perform an event study on the impact of the winning bid announcements for live broadcasting rights of the English Premier league on the stocks of six involved football clubs and the successful bidder BSkyB. The bids involved in this study were from 1996 and 2000. Their results show that the reaction to the 1996 bid was neutral from investors in BSkyB and positive among investors in the six football clubs. This indicates that the clubs valuated the broadcasting rights lower than the actual bid. In 2000 the reaction of the investors in the clubs was muted, while the share of BSkyB increased in price, indicating a possible underpayment of the broadcasting rights. In contrast to earlier described literature, these reactions show a more rational approach of investors to football stocks and these valuations.

2.3 Investor behavior related to sport results

(14)

that investors in football and sport stocks are emotional investors, which do not react rational to certain events.

Contrasting to the study of Edmans et al. (2007) are the results of a comparable study performed by Boyle and Walter (2003). Their study focuses on the stock market in New Zealand, where the single dominant sport is rugby. For rugby, the primary contests are international in nature and could, therefore, have a large impact on national investor behavior. Hence, Boyle and Warner (2003) investigate the influence of the successes of the national rugby on the local stock return in New Zealand. However, they find no evidence for any irrational behavior of investors. Therefore, they conclude that irrational behavior of investors, regarding sports issues, is at best transitory.

2.4 The motive of the investor in football stocks

In a study on two Portuguese listed football clubs, FC Porto and Sporting Lisboa, by Duque and Ferreira (2005), the emotional link between football and stock ownership is mentioned in the two explanations given for trade in football stocks:

1. The irrational esteem of supporters 2. The economic rationale of any investment

Where by the trade of ‘normal’ stocks the economic rationale is often the only motive for trade, with football stocks an emotional aspect should be taken into account, since owners of football stocks, which are often supporters, could be more emotionally linked to their stocks.

In line with this is the view of Sloane (1971), who states that it is ‘quite apparent that directors and shareholders invest money in football clubs not just because of expectations of pecuniary outcome, but for psychological reasons like the urge for power, the desire of prestige, the propensity of group identification and the related feeling of group loyalty.’ This statement indicates a less rational economic approach of investors in football.

(15)

on the stock market reaction to events in the football world and the last two sections relate to irrational behavior of investors related to on field sport results and the ownership of sport stocks.

(16)

3. Research design

In this research design, the research question, the variables involved in the analyses and the hypotheses, will be discussed.

3.1 Research question

For this study the following research question is formulated:

In which way is the behavior of football investors and the reaction on investment and divestment announcements qualified as rational?

3.2 Hypotheses

To answer this research question a number of hypotheses are formulated.

H₀ The announcements of both the purchase and sale of football players create the same

reaction of investors to a major football transfer.

According to the shareholder value maximization theory (Woolridge and Snow, 1990; McConnel and Muscarella, 1985), one would expect the market to react positively to announcements of both investment and divestment decisions. These decisions would increase the firm’s market value and thus the value for the shareholders.

(17)

A third theory, the rational expectation theory (Woolridge and Snow, 1990), would suggest no reaction of the market at all. The rational expectation theory indicates that investments create at best temporary competitive advantages and therefore the market will not react quickly or strongly to these investment announcements.

The expectation is to find a positive reaction of the market to transfers. This would be in line with the findings and theory of McConnell and Muscarella (1985) that support the hypothesis that managers always seek value maximization.

This study will take different characteristics of the specific events into account. These characteristics are the following:

1. Age of the player

2. Quality of the former/destination club 3. National team player

The age of the players, in years, is measured on the date of the announcement of the transfer. The quality of the former/destination club is based on the all time club ranking of the International Federation of Football History and Statistics (IFFHS). This is a statistical ranking of all worldwide football clubs, based on their performances in the last seventeen years. The third variable concerns the fact whether the player is at the moment of the transfer, active in the national team of his country. With the help of some football databases (Soccerbase, Transfermarkt), these data were found.

With these variables in mind, three alternative hypotheses are developed. H 1 is

formulated as follows:

H₁ The age of a football player being transferred does not influence the reaction of

investors on the announcements of a major football transfer.

(18)

significantly greater abnormal returns for the investments that ‘create’ growth options, than for the investments that ‘exercise’ investment options. Therefore the rational expectation of the market would be that it reacts stronger to younger players, since these transfers are more in line with the ‘creation’ of growth options, whereas the transfers of older players are more similar to the ‘exercise’ of investment options. Such a positive reaction to younger players would be considered as rational and a negative reaction as irrational.

Subsequently, H₂ andH₃ concern the quality of the other club involved in the transfers and the fact whether the specific player is a national player.

H₂ The quality of the former/destination club of a football player being transferred does

not influence the reaction of investors on the announcements of a major football transfer.

H₃ The fact whether a football player is a player for the national team or not does not

influence the reaction of investors on the announcements of a major football transfer.

The same expectation as for age might be applicable for these two other hypotheses. If the player is from a club of higher quality, one could expect a less strong reaction, than for players from clubs with lower quality, because the latter has more growth opportunities. Players that do not yet play for their national team have more growth opportunities, than players that already do. However, since the development and growth opportunities has more to with age than with the quality of the club and whether the player is a national player or not, we link these hypotheses to a different theory; the immediately generating cash flows theory and the certainty of these cash flows (Burton, 1999).

(19)

investments in national players and players from better clubs as better investments than investments in players that do not play in their national team or come from a worse club. Therefore, the rational reaction of the market would be positive to both transfers of national players and players from better clubs.

3.3 Control variables

Some control variables are included in the models testing the hypotheses in order to check for club characteristics that could influence the reaction of the market. Below these control variables are shown.

1. Size of club 2. Mispricing

3. Amount of transfer fee

The size of the club is measured by the value of the total assets of the club. For mispricing, the market to book ratio is used. These values are derived from the database Amadeus. The transfer fee is measured in Euros and is defined as the fee the club has announced in public that it is willing to pay for a player. In the analyses the natural logarithm of both the total assets and the transfer fee is used.

These control variables are described and used in earlier research as well. Jones et al. (2004) finds the size of the firm to correlate negatively with the abnormal return of the market. The larger a firm, the less influence an investment or divestment announcement will have on the market (Jones et al., 2004).

(20)

can be seen as an exception. This control variable is included in this study to check for the influence of the variable in the investments in the football industry.

(21)

4. Methodology

To test the formulated hypotheses, according to existing literature, an event study would be appropriate to apply. This study will be conducted in the way MacKinlay (1997) and Brown and Warner (1985) describe it.

4.1 Timeline of event study

The timeline of an event study is as follows. An estimation window is used to calculate the expected return for the days in the event window.

The estimation window used in this study exists of 250 days, from t₋₂₈₀ until t₋₃₀. This is a large, but necessary, estimation window, because of the low trading level of some of the involved shares. With this large window, the estimation of the expected return is more adequate. With the ending of the estimation window thirty days before the event date, the influence of the event does not affect the methodology.

4.2 The event study methodology

The two event windows in the analysis will exist of respectively two days (t and₋₁ t₀) and five days (t₋₂, t₂ ). The actual return of the individual shares is calculated with the following formula: ) 1 ( ) 1 ( ) ( ) ( − − − = t Pi t Pi t Pi t Ri (1)

In which Ri(t)denotes the actual return of share I on time t, Pi(t)denotes the closing price of a share on day t and Pi(t−1)denotes the closing price of a share the day before t.

(22)

The returns of the stocks of the clubs and the indexes are corrected for dividends and stock splits.

The expected return is calculated by using two market indices. The first one is a regional market index and the second one the Dow Jones EuroSTOXX Football index, an index that includes all European football clubs that are listed (STOXX). An advantage of using an extra industry index is that it allows for reducing the variance of the abnormal return through explaining more of the variation in the normal return than the market index (Batyrbekov, 2007). it Mt DJSFt n R R Rit E( )=α +β₁ +β₂ +ε (2) 0 ) ( it = E ε Var(εit)=σεi2

Where E(Rit)is the expected return of stock i on time t. α, β₁ and β₂ are estimated values from OLS regressions of the estimation period (t₋₂₈₀− t₋₃₀) of event t, R_DJSFt and

Mt

R are the actual return of the Football Index on time t and the actual return of the regional market index on time t respectively. The regressions are performed in the statistical program of Eviews 6.0.

Subsequently, the difference between the expected and the realized return is the abnormal return ( ARit ). ) ( ) (t E Rit Ri ARit= − (3)

Hence, the average abnormal return is calculated by:

∑

= = N i it ARi t N AR 1 ) ( 1 (4)

And furthermore the variance of this average abnormal return is:

(23)

Following this calculation the average cumulative abnormal return (CAR) can be derived by:

∑

= = 2 1 ) , (₁ ₂ t t t it AR t t CAR (6)

Where CAR(t₁,t₂)denotes the average cumulative abnormal return for periodt₁,t₂.

Before checking the significance of the findings, the standard deviation of the average CAR should be calculated. Subsequently, a student’s t-test, which can be applied, due to the size of the sample (Brown and Warner, 1985), can be performed.

∑

= = 2 1 ) , 2 1, ) ( ( t t t t i AR VAR t t CAR S (7)

The t-statistic on these coefficients can be used to determine whether the CARs are significantly different from zero.

) , ( ) , ( 2 1 2 1 t t CAR S t t CAR = θ (8)

This event study methodology measures the reaction of the market after an event. The abnormal return of the share, and thus the under- or overvaluation of the player by the club according to the market, can be derived.

To test the equality of the medians the Wilcoxon/ Mann Whitney U is performed. This value is derived by:

2 ) 1 ( ₁ 1 1 1 + − =R n n U (9)

(24)

4.3 Regression model

After performing the equality analysis regression models are used to check for significant influences of different variables. In these analyses all involved variables are taken into account.

At first, regressions are performed with the five day event window CAR as dependent variable and the control variables and one dummy as independent variables. This can be illustrated by the following formula:

ε β β β β β β α + + + + + + + = −2,2) 1 1 2 2 3 3 4 4 5 5 6 6 ( *X *X *X *X *X *X CAR_t _n (10)

where αn and allβ’s are estimators from OLS regressions and ε is the error term of the

regression. X_i are the different variables from the study, taken into account, i.e. age of player or market to book ratio of club. All the exact formulas for the different regression models can be found in appendix E.

4.4 Interaction model

After performing these regressions, an interaction analysis is done to check for interaction. The following model is formulated to test for this interaction:

ε β β β β β β β β α + + + + + + + + + = − ) * ( * * * * * * * * 2 1 8 7 7 6 6 5 5 4 4 3 3 2 2 1 1 ) 2 , 2 ( X X X X X X X X X CARt n (11)

where αn and allβ’s are estimators from OLS regressions and ε is the error term of the

regression. X are the different variables taken into account in the study, i.e. age of player or profit margin of club. The interaction effect is the (X₁*X₂)part. This interaction is done for all three dummies; age of player, nationality dummy and quality of club dummy.

(25)

ε β β β β β β β β β β α + + + + + + + + + + + = − ) * ( * ) * ( * ) * ( * * * * * * * * 4 1 10 3 1 9 2 1 8 7 7 6 6 5 5 4 4 3 3 2 2 1 1 ) 2 , 2 ( X X X X X X X X X X X X X CAR_t _n (12)

where X₁ is the buy/sell dummy and X₂ X₃ X₄ are the independent variables: age of player, nationality dummy and quality of club dummy.

4.5 Explanatory variable analysis

(26)

5. Data

The data section of this paper consists of a discussion of the data selection and the data description.

5.1 Data selection

The sample of this research involves solely football clubs from Europe. The sample consists of twenty different clubs which, except for Leeds United, currently perform on the highest national level. Leeds United though, has played on the highest national level in the time period of this sample. These clubs come from seven different countries: The Netherlands, Germany, England, Scotland, Italy, Turkey and Portugal.

The clubs are selected on the basis of their presence on a European Stock Exchange in the period 2001-2008 and the presence of one transfer or more in this period with a fee of at least €5.000.000.

For calculating the abnormal returns of the stocks, six different market indices are used; one for each individual country, where the Scottish and English clubs have the same market index. The following market indices are used: Netherlands (Dow Jones Netherlands), England and Scotland (FTSE All Share Index), Portugal (Dow Jones Portugal Index), Italy (Dow Jones Italy Index), Turkey (Dow Jones Turkey) and Germany (Dow Jones Germany).

(27)

region. The index represents the breadth and depth of the European football industry (STOXX).

Market data are gathered from the DataStream computer of the University of Groningen and the transfer events are retrieved from the German football website

www.transfermarkt.de and the international database Soccerbase. Dates of the announcement of the transfers are available from databases Soccerbase, Lexus Nexus and other football websites like Voetbal International. Financial characteristics of the specific clubs like size, profit margin and gearing ratio are found in the database AMADEUS that is available for students of the University of Groningen. The ranking of the clubs is available from the website of the IFFHS.

5.2 Data description

Table 1 provides an overview of all the involved events in this study. A distinction in buying and selling transfers, split out per club and per season, can be found in appendix A and B. Some clubs, like Galatasaray, went public after 2001, so for those clubs the early years in the sample are not applicable. Other clubs, like Manchester United, were delisted during the 2001-2008 period, so for those clubs some years are left out as well. As can be seen, the split over the different years is relatively small, with numbers varying from 21 to 39. Between clubs, the difference is larger, due to the difference in activity when it comes to transfers.

Table 1 Overview of number of events per club per year

Number of events 2001 2002 2003 2004 2005 2006 2007 2008 Total

(28)

6. Results

The section of the results is split into an overview of the descriptive statistics and

multicollinearity, followed by the results of the performed event study and, subsequently, the results of the cross sectional, regression and interaction analyses.

6.1 Descriptive statistics

For the whole sample, using the statistical package of Eviews 6.0 the following descriptive statistics for all involved variables are found.

Table 2 Descriptive statistics of all variables

Descriptives Age AR (t0) CAR (t-1,0) CAR (t-2,2) Transfer fee (€) Total assets (€) M to B

Observations 257 257 257 257 257 193 204 Mean 24,13 0,33% 0,67% 0,78% € 12.480.603 € 240.365.900 3,74 Median 24,00 0,04% 0,12% 0,19% € 9.900.000 € 238.559.200 1,77 Maximum 33,00 0,25% 0,37% 0,29% € 48.000.000 € 506.767.100 69,57 Minimum 16,00 -1,60% -0,18% -0,25% € 5.000.000 € 17.571.810 -15,44 Std. Dev. 3,26 0,03 0,04 0,05 € 10.106.994 € 109.619 14,08 Skewness -0,09 1,81 2,64 0,57 1,91 0,30 5,50 Kurtosis 2,36 22,19 30,24 10,32 6,86 2,53 34,92 Jarque-Bera 4,46 4.084,89 8.243,73 588,45 316,76 4,69 9.688,00 Probability 0,09 0,00 0,00 0,00 0,00 0,10 0,00

The age concerns the age of the player involved, AR is the abnormal return for an event window of one day and the CAR is the cumulative abnormal return for respectively two and five days. Transfer fee is the fee paid in the transfer and total assets is a proxy for the size of the club. M to B denotes the market to book ratio of the club.

Unfortunately, as can be derived from table 2, data were not available for all clubs. For some events the total assets and market to book ratio were not found. For that same reason, the variables leverage and profit margin were excluded in this study.

(29)

day event window it is 0,78%. The significance of these findings will be discussed in the following section.

6.2 Multicollinearity

In appendix D a correlation table with all coefficients is included for checking on multicollinearity. Brooks (2002) and Hill et al. (2001) give as an indication for collinearity a coefficient of 0,8. The only variables that do have such a high correlation are the CAR’s from the different event windows. Since these variables are never taken together in an analysis, the problem of multicollinearity does not exist in this study.

6.3 Buying vs. Selling

The first analysis of the results of this study concern the differences in reaction of the market between the two types of events; (1) buying and (2) selling events. From the total sample of 257 events, 103 events concern selling transfers and the other 154 events regard buying transfers.

As described in the methodology, the significance of these descriptive statistics is tested by a student’s t test. The following values were found.

Table 3 Significance of buying vs. selling Overview CAR's

Event window CAR (%) t-stat CAR (%) t-stat CAR (%) t-stat

t (0) 0,33% 0,94 0,23% 0,50 0,47% 0,90

t(-1,0) 0,67% 1,94** 0,62% 1,34 0,75% 1,45

t(-2,2) 0,78% 2,25** 0,52% 1,11 1,18% 2,26**

All transfers (257) Buying transfers (154) Selling transfers (103)

*,**,*** is significant on respectively 10%, 5%, and 1%

(30)

event window of five days, a positive abnormal return of 0,78% was found. This is in line with the shareholder maximization theory (McConnel and Muscarella, 1985), which indicates that managers solely undertake projects and make investments that maximize the value of the firm and that firms only undertake divestments, whereby the value of the firm in general increases. This positive abnormal return shows the appreciation of the investors for the investments and divestments executed by the clubs. Besides that, these results indicate that the football industry does not suffer from any agency problems, as described by Wright and Ferris (1997). The managers of the clubs seem to act in the best interest of the shareholders.

For comparing the data from the two samples, an univariate ANOVA analysis is performed for both the median and the mean of both samples. This analysis gives the following results.

Table 4 Test of equality on means and medians

Test of equality CAR t(-1,0) CAR t(-2,2)

Test P-value P-value

t-test (Mean) 0,803 0,338

Wilcoxon/Mann-Whitney (Median) 0,867 0,393

Table 4 gives an overview of the test of equality on the means and medians of the two compared series. Although for the t-test the p-value is decreasing with the larger event window, the findings for both event windows are not significant on a 10%-level. This indicates that the null hypothesis, of equal means for the CAR’s, cannot be rejected, indicating the means of the sample groups are equal.

Furthermore, the test of equality of the medians for the two subsamples, again, show decreasing p-values with an enlarging event window. These numbers however, cannot reject the hypothesis that both medians of the two series are equal.

6.4 Cross sectional analysis

(31)

Table 5 Significance of national vs. non national players Overview CAR's

Event window CAR (%) t-stat CAR (%) t-stat

t (0) 0,60% 1,44 -0,31% -0,49

t(-1,0) 1,05% 2,51** -0,20% -0,31

t(-2,2) 1,04% 2,49** 0,18% 0,29

All national players (180) All non-national players (77)

One of the factors being studied in this research is whether a football player is, at the date of the announcement of the transfer, active in his national team or not. In table5, selling and buying transfers are combined and then a split in the sample is made between national and non-national players. As can be seen from table 5, there is a significant positive CAR, for both two and five day event windows, respectively 1,05% and 1,04%, when it regards national players. For non-national players no significant results are found. These results are in contrast with the growth theory, indicating that there should be a stronger reaction to investment announcements that ‘create’ growth opportunities than those investments that do not (Jones, 2004).

These findings are, however, in line with the immediate generating cash flow theory (Burton, 1999). Players, that already have made their appearance in their national squad, have already performed on a high level and therefore, give more guarantees for good performances. Besides that, players from national teams enjoy more media coverage and will therefore, generate more and faster merchandising revenues. Subsequently, cash flows are generated more quickly.

Table 6 Test of equality of means

Test of equality CAR t(-1,0) CAR t(-2,2)

Test (257 observations) P-value P-value

t-test (Mean) 0,03 0,25

(32)

Table 6 shows the results of the tests of equality on the means and medians of the two samples. For the smaller event window a significant result is found, on the test of equality of the means. This would indicate that the means of both samples would differ. For the larger event window however, this result does not show up, which means the equality hypothesis cannot be rejected. The test on the medians shows no significant findings, indicating that the medians of both subsamples are equal.

Next, for a different analysis the sample of transfers is divided into two groups, based on the quality of the other club involved in the transfer. According to a statistical ranking of IFFHS all clubs are ranked. All transfers that involve a club with a higher ranking on the list are put together in one group: the transfers with better clubs group. All transfers that involve clubs, lower ranked by IFFHS, were included in the other group: the transfers with worse clubs group. The results of this analysis are found in table 9.

Table 7 Significance of better clubs vs. worse clubs

Overview CAR's

Event window CAR (%) t-stat CAR (%) t-stat

t (0) 0,89% 1,44 0,01% 0,02

t(-1,0) 1,29% 2,07** 0,34% 0,80

t(-2,2) 1,51% 2,43** 0,39% 0,92

Transfers with better clubs (88) Transfers with worse clubs (169)

The group of transfers dealing with lower ranked clubs has lower CAR’s for all event windows and, more importantly, show no significant findings. The group with the higher ranked clubs, however, does have significant results. This indicates that the market responds positively to transfers where better clubs are involved. The market responds with a 1,51% higher return in a five day event window, when a specific club executes a transfer with a better club.

(33)

distinction between good and bad investments. Investments in players from better clubs are valued as good investments.

The event study is followed by a test of equality on both the means and medians of the two samples. The results can be found in tables 10 and 11 below.

Table 8 Test of equality of means and medians

Test of equality CAR t(01,0) Car t(-2,2)

Test (257 observations) P-value P-value

t-test (Mean) 0,09 0,12

Wilcoxon/Mann-Whitney (Median) 0,15 0,13

The test on equality of the means, provided in table 8, shows significant findings on a 10% level for the two day event window, but not for the five day event window. This indicates that the hypothesis of equal means for the large event window cannot be rejected. For the smaller event window, however, this hypothesis can be rejected.

The results on the test on equality of the median, also provided in table 8, are for both event windows not significant, so the equality hypothesis cannot be rejected.

6.5 Regression and interaction analysis

An interaction analysis is executed to check for interaction between different variables. Before performing the interaction model, regressions are performed as described in the methodology. Afterwards, the interaction model is performed in several ways, with different interactions between variables, as described in the methodology. All results can be found in table 9 on page 35. Models 8, 9, 10 and 11 concern the interaction analyses.

(34)

and 4, indicating that the amount of the fee has a positive influence on the abnormal return on the market.

The results in the interaction models 8,9,10 and 11 show low interaction coefficients, but, more importantly, for some relationships significant results. Interesting to see, is the result of the interaction model where the quality of club dummy is involved. In model 9, the coefficient for the dummy itself is significantly positive, 0,026, but when combined in the interaction with the buy/sell dummy it shows a significant negative coefficient, -0,029. This finding indicates a difference in market reaction to announcements for buy and sell events, related to the quality of the club that is involved. This finding will be further discussed in the explanatory variable analysis, where difference between buys and sells are studied.

Model 8 shows a negative coefficient, -,004, for the interaction between buy/sell and the age of the player. This result indicates a negative reaction to the sell of both younger and older players, but a stronger negative reaction to older players. In model 11, where all interactions are included, this negative coefficient holds, although both the value and the t-value decrease.

This negative relationship between age and market reaction can be classified as an irrational reaction, when compared with the growth opportunity theory (Jones, 2004), since this theory would argue that the sale of an older player would be more positive than the sale of a younger player, because then no growth potentials are lost. Caution should be taken though, when looking at these results, since the theory of Jones (2004) concerns investments and the results here concern divestments. Therefore, for a more thorough look, a distinction is made between buy and sell in the next section and a further analysis is performed for checking for the unexpected negative relationship between age and buy/sell dummy.

(35)

6.6 Robustness check

(36)

(37)

6.7 Sample split in age

To explain the unexpected, irrational, negative coefficient between sell transfers and the age of players the total sample of selling transfers is split into two categories. The first category refers to all players younger than 27 and the second category consists of players of 27 and older.

Table 10 Average transfer fee split in age segments selling transfers

Descriptive statistic Average transfer fee Age <27 € 14.025.859 Age ≥27 € 14.052.564

Surprisingly, table 13 shows that the average transfer fee for both samples is almost the same. However, the expectation was that older players, with more experience, would have a higher value than their younger colleagues. Since the average transfer fees are comparable, this might explain the more negative reaction to the sale of older players. Although this results is not in line with the growth opportunity theory, with this explanation the investor reaction seems rational and explained.

6.8 Explanatory variable analysis

(38)

Table 11 Comparison between buying and selling events and their explanatory variables

Comparison buy/sell explanation variables

Buy Sell Buy Sell Buy Sell Buy Sell Observations 97 63 97 63 97 63 97 63

Size -0,015 -0,021 -0,014 -0,028 -0,015 -0,028 -0,014 -0,021

t-value -1,50 -1,09 -1,41 -1,63 -1,55 -1,63 -1,44 -1,06

Transfer fee 0,021 0,005 0,019 0,002 0,019 0,001 0,017 0,003

t-value 2,14 ** 0,33 1,92 * 0,15 1,88 * 0,10 1,68 * 0,18

Market to book ratio 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000

t-value 0,67 0,30 0,39 0,40 0,74 0,38 0,40 0,25 Age 0,001 -0,003 -0,0001 -0,003 t-value 0,48 -0,92 -0,06 -0,93 Quality dummy 0,024 0,002 0,023 0,002 t-value 2,00 ** 0,11 1,82 * 0,12 Nationality dummy 0,013 0,011 0,010 0,012 t-value 1,09 0,51 0,78 0,58

Model 1 Model 2 Model 3 Model 4

Size is the natural logarithm of total assets, transferfee is the natural logarithm of the transferfee, quality dummy is 0 for lower quality and 1 for higher quality, nationality dummy is 0 for non-national players and 1 for national players.

Although not all results in table 11 are significant, it is interesting to see that the quality of club dummy in model 4 shows a significant positive coefficient of 0,023 for buying transfers and a much smaller and non-significant coefficient, 0,002, for selling transfers. The quality of the other club involved in the transfer seems much more important for buying events than it is for selling events. These results are in line with theory of Chan (1995); investors in football are apparently able to make distinctions in good and bad investments. Investment announcements for investments from better clubs on the IFFHS ranking are valued positive by the market.

(39)

7. Conclusion

The intention of this paper was to provide insight into the stock market reaction of investors to the announcements of major football transfers in Europe and whether this reaction could be qualified as rational. Such a study, focused on the rationality of the investors in football, linked to the investment and divestment announcement effects of transfers, is unique in its kind. Earlier studies have solely emphasized on the influence of game results on the stock market and on the psychological influence of national sport success on the stock market.

The sample of this study consists of 257 announcements of football transfers, both buying and selling, with a minimum transfer fee of €5.000.000. The methodology followed in this research is in line with the event study methodology as Brown and Warner (1985) and MacKinlay (1997) describe it. An event window of five days t(-2,2) is used to calculate the short-term abnormal returns in investment and divestment announcements. The results of this study in general, show rational behavior by the investors. Significantly positive Cumulative Abnormal Returns (CAR) were found for all investment announcements together, for both a two day and a five day event window of respectively 0,67% and 0,78%.

This CAR is even larger for a five day event window for only the selling transfers with 1,18%. These results support the shareholder maximization theory. Financial announcements of football clubs in general, and divestment announcements in particular, generate positive abnormal returns. This indicates that the market values these activities as value adding and reacts rational to these announcements. Another implication of these results is the absence of agency problems in the football industry. Management seems to act in the best interest of the shareholder.

(40)

contradicting with the growth opportunity theory of Jones (2004), they can be explained as a rational reaction. National players have already proved their quality and are therefore an investment that gives more certainty for generating cash flows (Burton, 1999).

Interestingly, both buying and selling national players have higher CARs than the buying or selling non-national players, indicating the buy/sell characteristic does not influence the reaction of investors.

These findings are in line with the conclusion of Chan et al. (1995), who stated that investors are able to distinguish between good and bad investments and divestments.

The same conclusion can be drawn from the analysis on the quality of the club. The announcements of transfers with clubs, having a higher position on the IFFHS ranking, create significant positive abnormal returns, where the announcements of transfers with clubs, lower on that ranking, create no significant abnormal returns.

An interesting similarity between the two dummies, respectively nationality and quality of club, is the relatively high (0,28 and 0,23) correlation with the amount of the transfer fee announced. This correlation confirms the assumption that transfers with national players and with better clubs, concern transfers with better players, since clubs are willing to pay a higher transfer fee for better players. Subsequently, this leads to a stronger market reaction.

When it comes to quality of the investment the market reacts stronger on higher quality investments, which is in line with former literature (Chan et al., 1995).

(41)

Furthermore, the reaction of the market to buys and sells is different, when the explanatory variables are compared. The quality of the other club involved in the transfer is of higher importance for investors when the transfer is a buy than for a sell. This can be explained in line with earlier results of Chan et al. (1995), who found that the market is able to distinguish between good and investments. The market appreciates the buy from a better club, considering this would be a good investment.

(42)

8. Limitations and further research

In this last part of this paper the limitations of the study and the field for further research will be discussed.

8.1 Limitations

Although this research is performed with the highest possible accuracy, some limitations of the study remain. One of these limitations concerns the data. This study researched the announcement effect of investments and divestments of football clubs. Subsequently, the announcement date of each single event must be found. After very careful and extensive research, for most events, this date was found. For some events, though, this date was untraceable, so for those events a careful assumption of the announcement date was made, by checking the rumors in the market and the actual date of the deal.

A second limitation is in the sample size. Since this study solely observes the transfers with a minimum of €5.000.000 in Europe, in the time period from 01/2001 till 08/2008, the sample size is limited to 257. When enlarging the time period or the geographical scan, or by lowering the minimum amount of transfer fee, the sample size could be extended and the results could be interpreted more universal.

A recent trend in football that jeopardizes this research field is the trend of delisting. In Great Britain giant football clubs, like Chelsea, Manchester United and Manchester City have been recently delisted, due to the takeover of the club by immensely rich individuals. This limits the investment opportunities in football. In November 2008 some rumors about Ajax, considering delisting, entered the market. However, the Dutch football club waited with further steps due to the bad stock exchange climate, which is now suffering from the current credit crunch (Voetbal International, 2008). Although this trend is not a concrete limitation for this paper, it is a fact to take into account when reading the paper.

(43)

is limited. Further research may use more control variables. Variables like profit and leverage are interesting variables to add. In this research, considering the databases available, it was unfortunately not possible to find enough relevant data for those variables.

8.2 Further research

Since this study is the first in its kind, it is a good precedent for following studies in this field. The rationality of investors in general is in an emerging research field and to combine this with an emotional world, like the world of sports, is a challenging combination.

A further research could compare different sports to check for differences in market reaction and rationality or a specific cross-country analysis to check for differences among cultures. Hence, a cross-continent analysis could be performed.

Beyond the scope of this research, but a topic that could be researched, is to check for the quality of the undertaken investments. This research indicates that investors are able to make a distinction between good and bad players. Further studies may investigate the question if the performances of a player, being qualified by investors as valuable, after a transfer, are as good as the investors predicted or do investors overvalue certain players?

Furthermore, more research into the agency problems of football clubs is required. Although this paper finds support for the shareholder maximization theory, more research may investigate what the motives of listed football clubs are for buying and selling players. Is there solely a financial motive or are there other factors involved?

(44)

References:

Afshar, K., Taffler R. and P. Sudarsanam, (1992), ‘The effect of corporate divestments on shareholder wealth: The UK experience’, Journal of Banking and Finance, 16, 115-135.

Akbar, S., Ali Shah, S. and I. Saadi, (2008), ‘Stock market reaction to capital expenditure announcements by UK firms’, Applied Financial Economics, 18, 617-627.

Baker, M., Stein, J. and J. Wurgler, 2003, ‘When does the market matter? Stock prices and the investment of equity-dependent firms,’ The Quarterly Journal of Economics, 118, 969-1005.

Batyrbekov, K.Y., (2007), ‘Soccer Stocks: Market Reaction to Game Results of Professional Soccer Franchises’, Undergraduate thesis, Harvard College.

Boyle, G. and B. Walter, (2003), ‘Reflected glory and failure: international sporting success and the stock market’, Applied Financial Economics, 13, 225-236.

Brailsford, T. and D. Yeoh, (2004), ‘Agency problems and capital expenditure announcements’, Journal of Business, 77, 223-256.

Brambor, T., Clark, W.R. and M. Golder, (2006),’Understanding interaction models: improving empirical analysis’, Policital analysis, 14, 63-82.

C. Brooks, (2002), Introductory econometrics for finance, Cambridge University Press, seventh edition 2006.

Brown, G. and J. C. Hartzell, (2001), ‘Market reaction to public information: The atypical case of the Boston Celtics.’ Journal of Financial Economics, 60, 333-370.

Brown, S., and J. Warner, (1980), Measuring Security Price Performance, Journal of

(45)

Burton, B., Lonie, A.and D. Power, (1999), ‘The stock market reaction to investment announcements: The case of individual capital expenditure’, Journal of Business Finance

& Accounting, 26.

Chan, S., Gau, G. and K. Wang, (1995), ‘Stock market reaction to capital investment decisions: Evidence from business relocations’. Journal of Financial and Quantitative

Analysis, 30.

Chan, S., Martin J. and J. Kensinger (1990), ‘Corporate research and development expenditure and share value’, Journal of Financial Economics, 26, 255-276.

Chung, K., Wright, P. and C. Charoengwong, (1998), ‘Investment opportunities and market reaction to capital expenditure decisions’, Journal of Banking & Finance, 22, 41-60.

Dos Santos, B., Peffers, K. and D. Mauer, (1993), ‘The impact of information technology investment announcements on the market value of the firm’, Information systems

research, 4, 1-23.

Duque, J. and N.A. Ferreira, (2005), ‘Explaining share price performance of football clubs listed on the Euronext Lisbon’, Universidade Tecnica de Lisboa Business Administration Working Paper No. 05-01.

Edmans, A., Garcia, D. and O. Norli, (2007), ‘Sports sentiments and stock returns’,

Journal of Finance, 62, 1967-1998.

Gannon, J., Evans K. and J. Goddard, (2006), ‘The stock market effects of the sale of live broadcasting rights for English premiership football: An event study’, Journal of Sports

(46)

Hill, R.C., Griffiths, W. E. and G. Judge, (2001), Undergraduate econometrics, John Wiley and sons, 2nd edition.

IFFHS, (2008), All time club ranking,

http://www.iffhs.de/?3d4d443d0b803e8b40384c00205fdcdc3bfcdc0aec70aeedbe1a

Jones, E., Danbolt, J. and I. Hirst, (2004), ‘Company investment announcements and the market value of the firm’, European Journal of Finance, 10, 437-452.

Lexus Nexus, 2008, University of Groningen, Information database.

Loughran, T. and J.R. Ritter, 1995, “The new issues puzzle,” The Journal of Finance 50, 23-51.

A. C. MacKinlay, (1997), ‘Event studies in economics and finance’, Journal of

Economic Literature, 35, 13-39.

P. Malatesta, (1983), ‘The wealth effect of merger activity and the objective functions of merging firms’, Journal of Financial Economics, 11, 155-181.

McConnel J. and C. Muscarella, (1985), ‘Corporate capital expenditure decisions and the market value of the firm’, Journal of Financial Economics, 14, 399-422.

Palomino, F., Renneboog, L., and C. Zhang, (2005), ‘Stock price Reactions to Short-lived Public Information: The Case of Betting Odds’, ECGI - Finance Working Paper, 81

Renneboog, L., B. Vanbrabant, (2000). ‘Share price reactions to sporty performances of soccer clubs listed on the London Stock Exchange and the AIM’, Center for Economic

(47)

Saadouni, B., Briston, R. J. and C. A. Mallin, (1996), Security price reaction to divestments by healthy and financially distressed firms: the case of MBOs, Applied

Financial Economics, 6, 85-90.

Scholtens, B. and W. Peenstra, (2008), ‘Scoring on the stock exchange? The effect of football matches on stock market returns: an event study, Applied Economics, 1-7.

Soccerbase (2008), www.soccerbase.com,

P. Sloane, (1971), ‘The economics of professional football: The football club as utility maximizer’, Scottish journal of political economy, 17, 121-146.

Stoxx.com. (2008), ‘Dow Jones Stoxx Football Index’

http://www.stoxx.com/indices/index_information.html?symbol=FCTP

H. Sun, (1994), ‘The relationship between the valuation effect of equity financing and firm-specific characteristics’, Journal of Economics & Finance, 18, 55-67.

Titman, S., Wei, K. and F. Xie, (2004), ‘Capital investments and stock returns’, Journal

of Financial & Quantitative Analysis, 39, 677-700.

Tversky, A. and D. Kahneman, (1979), ‘Prospect theory: An analysis of decision under risk’, Econometrica, 47, 263-291.

Transfermarkt.de (2008), Overview all transfers from European football clubs.

Voetbal International, (2008), www.vi.nl, ‘Ajax ziet vooralsnog af van beëindiging beursnotering’, 18-11-2008.

(48)

Woolridge J. and C. Snow (1990), ‘Stock market reaction to strategic investment decisions’, Strategic Management Journal, 11, 353-363.

Wright, P. and S. Ferris (1997), ‘Research notes and communications agency conflict and corporate strategy: The effect of divestment on corporate value’, Strategic Management