Conflicts of interest in analyst’ stock recommendations: Evidence from the Dutch stock exchange

Conflicts of interest in analyst’ stock

recommendations: Evidence from the Dutch

stock exchange

Daniël Huizenga

Studentnumber: 1384589

Master Thesis International Financial Management

University of Groningen

Faculty of Economics and Business

Supervisor: Dr. A. Plantinga

Table of Contents

Abstract ... 3

1. Introduction ... 4

2. Literature Review ... 8

3. Methodology and Data ... 13

3.1 Data ... 13

3.2 Indicators of involvement ... 14

3.3 Methodology ... 16

4. Results ... 18

5. Conclusion and Discussion ... 22

6. References ... 24

Abstract

I investigate the existence of conflicts of interest at analyst recommendations on the Dutch stock exchange in 2002-2008. I make a distinction between analyst firms that perform business with companies they review and analyst firms that perform no business with companies they review. I discover that recommendations of unaffiliated analysts are more optimistic than recommendations of affiliated analysts; this does not correspond with my expectations of the first hypothesis. Moreover, I study recommendations of affiliated analysts on companies that perform business with the analyst firms. I find evidence for optimism in recommendations on companies that perform business with the analyst firms. In addition, I study whether the optimism is related to four indicators of involvement: capital increase, initial public offering (IPO), minority stake and public takeover. I find optimism in recommendations for a minority stake and public takeover. Recommendations related to an IPO and capital increases have no positive bias.

1. Introduction

An abundance of information about shares and stock markets makes it difficult for investors to deal with all the information they gather. For investors it is complicated to find the important information and analyze the performance and possible future actions of a company. As a result, there are analysts who advice investors.

Analysts have an important role for companies and investors. They study companies and industries to forecast the future performance of these companies. Moreover, they inform investors about their findings and give them recommendations. In general they are using a three or five grade scale for stock recommendations. Buy, hold or sell on a three grade scale and strong buy, buy, hold, sell and strong sell on a five grade scale. Investors view analysts as experts and they are eager to obtain recommendations on stocks in their portfolio (IOSCO, 2003).

There are three different types of analysts: sell-side, buy-side and independent analysts. Sell-side analysts work at a research department of a brokerage firm; most brokerage firms are investment banks as well. Investment banks earn money on the commissions they charge for raising money for companies and giving advice to companies when they are involved in, for instance, a public takeover (Newsome, 2005). The research reports and recommendations they provide are in general accessible for all the investors. The banks are preparing reports on companies that their firm covers. However, after the introduction of regulation fair disclosure this has changed; in the literature section this will be discussed. Before the introduction of regulation fair disclosure, many sell-side analysts provide other services than their activities at the research department of firms in the field of investment banking. This could be providing services for companies that they study as an analyst for investors (IOSCO, 2003).

Buy-side analysts are working at money management firms like hedge funds, mutual funds and pension funds. In general their research department is much smaller than the department of an average sell-side analyst. In addition, they do not have the opportunity to get information from different sources like their sell-side colleagues. Another aspect of buy-side analysts’ information is that only portfolio managers of the firm can use it (Groysberg et al., 2008).

and most objective information. When analysts earn a bad reputation for themselves they will not sell their recommendations and reports. Among the different types of analysts, sell-side analysts are dealing with the highest possibility of conflicts of interest. The analysts and their firms provide recommendations and information on companies where they work for and sometimes even perform business with (Ljungqvist et al., 2007; IOSCO, 2003). Hence, I investigate recommendations of sell-side analysts to discover if there are any conflicts of interest.

In prior research on this topic Barber (2006) explains the possibility of conflicts of interest in the United States after studying analyst firms in the mid-2000:

“With the percentage of buy recommendations reaching 74 percent of total outstanding recommendations by mid-2000 and the percentage of sell recommendations falling to 2 percent, allegation arose that analyst’ recommendations did not reflect their true beliefs.”

Some people argue that analyst firms would attract their business by giving recommendations with this level of optimism. In addition, Lin and McNichols (1998) conclude that recommendations of sell-side analysts, working at investment banks, are more optimistic than recommendations of unaffiliated analysts. In contrast, Clarke et al. (2006) conclude that analysts who are working for investment banks are no more optimistic than unaffiliated analysts. Their study focuses on analyst optimism for companies that file for bankruptcy. The focus of prior research is mainly on recommendations of American analyst firms at companies listed on American stock exchanges. In this study, I will focus on recommendations for companies listed on the Dutch stock exchange. I decided to narrow down this study to the Dutch AEX index. The AEX index includes the biggest and most important companies on the Dutch stock exchange. The Dutch stock exchange, in contrast to the American stock exchanges, has low trading volumes and the number of listed companies is smaller than on American stock exchanges. In addition, there are many local analyst firms that provide recommendations for companies on the Dutch stock exchange.

and recommendations of unaffiliated analysts who have no business ties with companies they review.

In my study, I will make a distinction between two types of analyst firms. The first one, full-service investment banks is the main type of firm in my study. Full-full-service investment banks offer activities like being a financial advisor in takeover deals and underwriting. In addition, the banks have a brokerage department and a research department where the analyst works. The second one, non-underwriter banks provide brokerage services and financial advice to their clients but do not underwrite like full-service investment banks (Cowen et al., 2005). I will include analyst firms that do not have a brokerage department as the second type of firm. Analysts at these firms provide only financial advice for their clients.

I will label these two types as affiliated and unaffiliated firms. Affiliated firms sell investment banking services beside the recommendations the analysts provide. Unaffiliated firms do not sell any investment banking services. According to Barber (2006) and Lin and McNicholds (1998) I expect to find more optimism at recommendations of analysts who work for affiliated firms since their firms sell investment banking services at companies they review. I will test the following hypothesis:

Hypothesis 1: Recommendations of affiliated analyst firms are more optimistic than the recommendations of unaffiliated analyst firms

In the methodology and data section I will explain which statistical methods will be applied to test the hypotheses.

After investigating optimism in all the analyst recommendations, the emphasis of the second hypothesis is on recommendations of affiliated analyst firms. Affiliated analyst firms give recommendations about companies where they sell and do not sell investment banking services. In the second hypothesis I will investigate if the recommendations on companies, in which they do sell investment bank services, are biased in comparison with companies in which they do not sell these services.

Two types of companies have been identified:

• Related companies: the companies that perform business with analyst firms.

With the following hypothesis I expect to find a positive bias in the recommendations on related companies:

Hypothesis 2: Recommendations of affiliated analyst firms on companies in which they sell investment banking services are positively biased

For the following part of my study I will only use recommendations of affiliated analysts on related companies. I will select four indicators of involvement; these indicators are a capital increase, initial public offering (IPO), minority stake and public takeover.

Prior research concerning these indicators of involvement focus on IPO’s. Companies need an investment bank to raise money and to guide the IPO process. During and just after the IPO the companies need a liquid market for their stocks. Thus, when analysts give optimistic recommendations on these companies, investors are more eager to buy these stocks (Cowen et al., 2006). Houston et al. (2006) investigate offer prices of IPO’s in relation to target prices that affiliated analysts give in the period just after the IPO. The affiliated analysts recommend a target price that is 275% above the offer price. A contribution to the literature in this field of research is investigating the joint effect of the four indicators of involvement on analyst recommendations.

For instance, recommendations on companies that are involved in an IPO are more optimistic than recommendations on companies that are involved in a capital increase. I expect to find optimism at recommendations related to the indicators of involvement. This leads to the following research question:

Are recommendations more optimistic at specific indicators of involvement?

2. Literature Review

The International Organization of Securities Commissions claims that conflicts of interest mainly arise because of the all-embracing work field of investment banking. The three main divisions of an investment bank according to Michaely and Womack (1999) are: proprietary trading, brokerage services and a corporate finance division.

These divisions are working together. This means that an analyst is sometimes working for a specific company and for the same company he gives independent recommendations. The analysts’ job security can depend on selling shares for companies he has to analyze as well. This means that an analyst could not operate as an objective person without the influence of the investment bank and the companies they recommend. Therefore, the analyst has to work without the influence of the investment bank and other companies. If this is not the case, investors can not trust the recommendations of the analyst.

Agrawal and Chen (2008) describe when analysts work for a firm that is highly dependent on investment banking services they face pressure to be optimistic in their recommendations. The reason for putting the analysts under pressure is that the firm would like to sell investment banking services to the company that an analyst reviews. In other words, investment bankers need analysts to obtain deals with companies that the analyst reviews. Analysts who attract clients like institutional buyers with their influence create business for other parts of the bank (Hong and Kubik, 2003).

An additional potential conflict of interest is the pressure that the brokerage department of the investment bank puts on the analyst. In several companies the salary or bonuses of analysts are related to trading volumes. As a result, there is a high chance analysts would like to increase the trading volumes of stocks. Another conflict of interest is the so called “bribery hypothesis”. This hypothesis argues that companies and institutional investors will take their business to another investment bank if the analyst gives recommendations which they do not like. As a result, analysts have to make decisions in favor of clients of the investment bank to receive their bonuses and their job security in the future (Kuo, 2007).

while providing high quality recommendations, Jacob et al. (2008) find that analysts working at investment banking firms provide better quality recommendations. Their explanation is the financial support of the research department from other investment bank activities. Hong and Kubik (2003) claim that sell-side analysts working at large investment banks in general earn higher salaries than independent and buy-side analysts. The higher salary leads to higher quality research reports and recommendations at companies.

Furthermore, the analyst gets superior information from the investment bank activities on companies and he has a better access to the management and other important sources at the company he reviews. Hong and Kubik (2003) describe that the quality of recommendations is important for the career of the analyst. Thus, an analyst must make good quality recommendations to get influence among the institutional buyers and fund managers. When their recommendations are of a low quality, the investment bank will not generate business from the buy side.

Westphal and Clement (2008) state that conflicts of interest arise not from the field of investment banking but from pressure of the companies that analysts review. Their research focuses on social and political factors that contribute to optimism in analyst recommendations. In addition, the relation between analysts and employees of the company they review is important in this context. They conclude that reciprocity has influence on the recommendations of analysts. Reciprocity is the exchange of favors in a relationship. The favors can be different but must be of the same value. In this context, employees of companies that analysts review use the concept of reciprocity as a tactic to influence analyst recommendations. For instance, employees could use their social status and network to introduce the analyst in exclusive organizations, recommend him for a job or boost his career by giving important information about their industry and company.

The goal of giving these favors to the analyst is that the employee of a company expects a favor in return from the analyst in the form of positive recommendations. Analysts that receive favors are more inclined to give positive recommendations. When they do not give positive recommendations, the reciprocity in favors is turning negative from the side of the employees.

possibilities of companies. As a result, analysts review the announcements of alliances as market signals and in this way alliances influence the recommendations of the analysts.

Chan et al. (2007) study earnings surprises of U.S. companies from 1984-2004. There are earning surprises when the earnings of a company in a quarterly or annual report are higher or lower than the estimates of the analysts. The estimates of the analysts are important for the share price of the company. If the estimates are below the level of the earning there is a high chance that the share price will drop.

Chan et al. (2007) discover a shift in positive earnings surprises from 48.88 percent in the late 1980s to 75.59 percent in the late 1990s. However, there is strong evidence that these results arise from a strategic decision making vision of the analysts. There is a case in which three months before the annual report the earnings of a company were below the level the analysts predicted. The analysts decided to adjust their estimates downward and after the revision the earnings of the company were positive in the annual report.

Consistent with the results of Chan et al. (2007), Hovakimian and Saenyasiri (2007) find for the period of 1984-2006 that analysts make too optimistic earnings estimates. The reason of these too optimistic estimates can be the information an analyst receives of a company. When the analyst is negative about the future performance of a company, the management is less likely to inform the analyst with the necessary information for their analysis. In addition, this can be a possible reason for the higher levels of positive earnings surprises (Chang et al., 2008).

The existence of conflicts of interest in the field of investment banking is an important topic in the United States. This leads to the introduction of Chinese walls and the Regulation Fair Disclosure (Reg FD). Chinese walls are information barriers between different departments of banks. The goal is to decrease conflicts of interest and insider trading. Gorman (2004) argues that Chinese walls decrease the insider trading to some extent, however, there exist many possibilities for employees for insider trading. Hence, Chinese walls cannot reduce all the conflicts of interest. The best solution will be to give investors the opportunity to choose if they rely on an analyst or not.

The reason of the SEC to introduce this new legislation is the existence of “selective disclosure”. Selective disclosure arises when companies provide non-public information to investor groups like institutional investors and pension funds (Ferreira and Smith 2006; Roy 2002; Herrmann et al. 2008).

Roy (2002) explains arguments that are not in favor of selective disclosure. First, there would be less investor confidence in a fair market when a small group of investors can trade with information that other investors not have. Second, the chance of a positive bias in analyst recommendations is reasonable. Analysts must write positive reports about companies that would not accept negative recommendations. A consequence of a negative recommendation could be no access to important information.

Herrmann et al. (2008) describe arguments that support the introduction of Reg FD. The access of all the investors to the same information at the same time would lead to an efficient market with less reliance on information of the management of the firm. In addition, analysts will not provide too optimistic information to their clients. However, they describe that the deletion of private communication leads to reports of analyst with less value for investors. Another negative effect of Reg FD is that firms disclose the information not to both the analyst and investors anymore. In addition, a research of the CFA institute concludes that some types of information are less available in the market. These types are earnings guidance, internal operations information, costs, pricing and sales information (Ferreira and Smith, 2006). Critics of the regulation believe that it leads to more volatile markets.

Another interesting topic is described in the article of Chang et al. (2008); they conclude that investors react to the recommendations of analysts. Analysts have access to all the important information, study it and give a recommendation to the investors. It is easy for investors to have a format that guide them by their decision making process to invest in a company. In this process, investors have to be aware of the distribution and timing of analyst’ recommendations. Panchenko (2007) notices that information is available for all the investors on the market. Nevertheless, he argues that there is an essential timing aspect for this information. At first, institutional investors receive information of analyst firms. After that, they publish the information to the media, who introduce it to all the investors.

analyst. On the other hand, when an analyst recommends a company negatively there will be a higher response rate to the recommendations from investors.

In contrast to this research, Francis and Soffer (1997) find that investors react more on positive than on negative recommendations of analysts’ reports through the fact that: “Buys are less informative than sells about the analysts’ beliefs about share values. The reason for thinking this is because analysts propensity to issue favourable stock recommendations and optimistic forecasts are well documented.”

Barber et al. (2001) study the analyst recommendations from another investors’ perspective. They study the profitability of the investors’ portfolio when they buy or sell stocks following analyst recommendations. It is hard for investors to make profit due to the semi-strong form of market efficiency. Nevertheless, analyst firms spend a lot of money on providing their recommendations. This means that they believe the use of their recommendations is profitable for investors. Barber et al. (2001) find significant abnormal returns. However, after accounting for transactions costs, the abnormal net returns were not greater than zero.

3. Methodology and Data

3.1 Data

My main sources for seeking analyst recommendations are the internet sites www.guruwatch.nl and www.analist.be. In table 1, I present an overview of the total number of analyst recommendations collected for the years 2002-2008. In total, I collect 5677 recommendations of which the largest part (55.57 percent) are “buys” and the second most collected recommendations are “holds” (34.49 percent). “Sells” (9.95 percent) are the smallest amount of collected recommendations of analyst firms. The sample consists of stocks belonging to the AEX index. In addition, I found data for all the 25 companies that are part of the AEX index at this moment (See Appendix: A). I found data for 4 companies that were removed from the AEX index during 2003-2008 (Appendix: B). These companies are Teleatlas, Hagemeyer, Van der Moolen and LogicaCMG.

There have been many changes in the AEX index the last five years and I decided to collect the recommendations of all the companies that were part of the AEX index in the last five years. These companies are removed from the index for several reasons like being part of a merger or acquisition or becoming too small.

In addition, I also used part of the dataset of Heijdenrijk and Plantinga (2001) based on information of analysts in the Dutch magazine Beursbelangen. However, due to mergers, acquisitions and the collapse of analyst firms I decided to use only a small part of their data collection.

Table 1: Summary Statistics of the sample data

Count Percent

Number of recommendations 5677

Number of analyst firms 14

Number of reviewed companies 29

Buys 3155 55,5

Holds 1958 34,5

Sells 565 10,0

In addition, these analyst firms are all part of the top five largest analyst firms in the investment banking sector (Zephyr, 30-10-2008).

Table 2: Summary Statistics of the fourteen analyst firms

Analyst Firm Recommendations Buy Hold Sell

ABN AMRO 405 240 125 36 Delta Lloyd 278 195 34 49 Fortis 381 243 114 24 Goldman Sachs 326 127 153 46 Iris 562 271 226 65 ING 516 270 206 40 JP Morgan 385 178 135 72 Kempen & Co 268 159 87 22 Lehman Brothers 251 100 101 50 Morgan Stanley 177 88 65 24 Rabo Securities 775 470 289 16 SNS 621 424 168 30 Theodoor Gilissen 245 157 65 23 Van Lanschot 487 233 186 68 Total 5677 3155 1958 565

Notice that all the institutes give more buy recommendations then hold and sell recommendations. Thereby, all the institutes give more hold than sell recommendations.

3.2 Indicators of involvement

Table 3 presents nine affiliated analyst firms and their recommendations. In addition, I selected recommendations on related companies. Recommendations on these companies have a high possibility of conflicts of interest.

Table 3: Recommendations of affiliated analyst firms

In the second column (recommendations) are all the recommendations of the affiliated analyst firms presented. In the third column (related) are the recommendations of affiliated analyst firms on related companies presented and sorted out by buy, hold and sell. In Appendix C is a specified table of related recommendations arranged by company name

Analyst Firm Recommendations Related Buy Hold Sell

ABN AMRO 405 173 116 56 1 Fortis 381 54 39 15 0 Goldman Sachs 326 104 42 58 4 ING 516 207 145 60 2 JP Morgan 385 73 49 21 3 Kempen&Co 268 16 5 11 0 Lehman Brothers 251 24 17 6 1 Morgan Stanley 177 101 59 30 12 Rabobank 775 160 107 50 3 Total 3484 912 579 307 26

Table 4 presents the number of deals of affiliated analyst firms arranged by the four indicators of involvement. A minority stake and public takeover are the main services of analyst firms to the companies. There is less business activity in the field of a capital increase and an IPO. In addition, the IPO of TomTom is responsible for five of the seven IPO observations. The roles of the analyst firms at the IPO of TomTom was being a (co-)lead manager, underwriter, financial advisor or broker. When TomTom adds more (co-)lead managers to the deal, they increase the analyst coverage for the period just after the IPO (Bradley et al., 2008).

Table 4: The services of analyst firms arranged by the four indicators of involvement

3.3 Methodology

Every analyst firm uses other terms for their recommendations; I decided to select three categories of analyst recommendations: buy, hold and sell. For instance, I classified strong buy, buy, moderate buy, overweight, outperform, add and accumulate in the same category (buy).

For the first hypothesis, I will investigate the optimism in the recommendations of affiliated analyst firms and not affiliated analyst firms. An affiliated firm has been involved in a transaction between an analyst firm and an AEX company on at least one of the four indicators of involvement. For testing the hypothesis I will make a distinction between optimism at buy and sell recommendations. I will not test optimism at hold recommendations. When analysts give hold recommendations they are not using their optimism that could lead to buy or sell decisions of investors. By giving buy and sell recommendations they activate the behavior of investors to buy or sell stocks. The reason for making a distinction between buy and sell recommendations is the high percentage of buy recommendations and the low percentage of sell recommendations. By applying two different tests I can identify differences in buy and sell optimism in the recommendations of analyst firms. For testing the first hypothesis I will use the following distributions:

F1: Buy recommendations from unaffiliated analyst firms

F2: Sell recommendations from unaffiliated analyst firms

G1: Buy recommendations from affiliated analyst firms

G2: Sell recommendations from affiliated analyst firms

buy and sell recommendations and performing two tests. This leads to the following explanation of hypothesis 2 at buy and sell recommendations.

F1: Buy recommendations from affiliated analyst firms about non-related companies

F2: Sell recommendations from affiliated analyst firms about non-related companies

G1: Buy recommendations from affiliated analyst firms about related companies

G2: Sell recommendations from affiliated analyst firms about related companies

I will test whether there is a bias in the recommendations of the distributions of Gi and Fi. To

determine the results of the two hypotheses a Chi-Square test is applied with one degree of freedom. The Chi-Square tests whether or not there is an independent relation between two variables and for testing the level of significance I will use the Pearson Chi-Square and the Likelihood-ratio Chi Square.

To test my research question, I will use the four indicators of involvement as the independent variables. I will study the relation of the four indicators of involvement with the dependent variable: buy. I will analyze the data by using a probit regression. The regression model estimated for this study is as follows:

BUY

β

+ β

+ β

+ β

+ β4

Public Takeover

+

4. Results

Table 4: Results for hypothesis 1 – difference in distributions of buy recommendations from unaffiliated and affiliated analyst firms

The dataset includes all the collected recommendations; these are recommendations of affiliated and unaffiliated analyst firms

Buy Not Buy Total

Unaffiliated analyst firms 1280 (58%) 913 2193

Affiliated analyst firms 1875 (54%) 1609 3484

Total 2522 3155 5677

Significance level

Pearson Chi-Square 0,001

Likelihood Ratio 0,001

Table 4 presents the results of the first hypothesis. The number of buy recommendations for affiliated and unaffiliated analyst firms is higher than not buy recommendations. To clarify: not buy recommendations could be either hold or sell.

The outcome of the Chi Square test reports that unaffiliated analyst firms are more optimistic in providing buy recommendations than affiliated ones. The significance level of this outcome is at a 1 percent level.

Table 5: Results for hypothesis 1 – difference in distributions of sell recommendations from unaffiliated and affiliated analyst firms

The dataset includes all the collected recommendations; these are recommendations of affiliated and unaffiliated analyst firms

Sell Not Sell Total

Unaffiliated analyst firms 235 (11%) 1958 2193

Affiliated analyst firms 330 (9%) 3154 3484

Total 565 3154 5677

Significance level

Pearson Chi-Square 0,127

Likelihood Ratio 0,129

Hypothesis 1 predicts optimism in recommendations of affiliated analyst firms as opposed to unaffiliated analyst firms. However, recommendations of unaffiliated analyst firms are more optimistic. This result is in line with previous research of Cowen et al. (2006). They identify less optimism of analysts working at full-service investment banks in relation to analyst who work at non-underwriter banks. This is in contrast with my expectations and the research of Lin and McNicholds (1998), who discover a positive bias at recommendations of sell-side analysts in comparison with unaffiliated analysts.

In addition, the results regarding to optimism in sell recommendations does not give a significant outcome. To conclude, this means that I reject hypothesis 1.

Table 6: Results for hypothesis 2 – the difference in distributions of buy recommendations from affiliated analyst firms at non-related and related companies

I use a dataset consisting of 3507 observations, which are only recommendations of affiliated analyst firms

Buy Not Buy Total

Non-related recommendations 1307 (50%) 1288 2595 Related recommendations 582 (64%) 330 912 Total 1889 1618 3507 Significance level Pearson Chi-Square 0,000 Likelihood Ratio 0,000

Table 6 presents results for my second hypothesis. The total number of recommendations on non-related companies is more than the amount of recommendations on related companies. This means that analyst firms provide more recommendations on companies in which they do not sell investment banking services. However, there is more optimism in recommendations on related companies. This outcome is significant at a 1 percent level.

Table 7: Results for hypothesis 2 – the difference in distributions of sell recommendations from affiliated analyst firms at non-related and related companies

I use a dataset consisting of 3507 observations, which are only recommendations of affiliated analyst firms

Sell Not Sell Total

Table 7 present results of the sell recommendations on related and non-related companies. The recommendations on related companies are more optimistic than on non-related companies. There are 24 sells of in total 889 recommendations at related companies. On non-related companies there are 306 sells of in total 2595 recommendations. The outcome of this test is significant at a 0 percent level.

According to hypothesis 2 recommendations on related companies are more optimistic. The outcome of the test shows a positive bias for both the buy and sell recommendations. This in line with the research of Ljungqvist et al. (2007), they find a positive bias in analyst recommendations for companies that perform business with investment banks. To conclude, this means that hypothesis 2 is not rejected.

Table 8: Results of the probit regression

Buy Recommendations Constant Capital Increase Initial Public Offering Minority Stake Public Takeover (1) (2) (3) (4) Buy 0,108 0,11 -0,034* 0,367*** 0,238* R-Squared 0,019 Number of observations 911 Significance level: p < 0,01*** p < 0,05** p <0,1*

Table 8 presents the results of the probit regression. I performed a Huber/White test to investigate the robustness of the results, since error terms suffer from heteroskedasticity. Hence, I prefer to report results obtained from using Huber/White consistent estimators. I also performed a logistic regression (results not reported) and get no significantly deviated results. This means the results are robust even when using different regression methods.

The coefficients of the independent variables have a positive relation for a capital increase (CI), minority stake (MS) and public takeover (PT) on the dependent variable buy. There is a small negative coefficient for the relation of an initial public offering (IPO) on the dependent variable. The relations are significant for IPO and PT on a ten percent level and for MS on a 1 percent level. The relation of CI is not significant.

5. Conclusion and Discussion

This study analyzes the existence of conflicts of interest in analysts’ recommendations at companies on the Dutch stock exchange for the years 2002-2008. The analysts that I review are working at a full-service investment bank, non-underwriter banks and banks or institutes with only a research department. In the literature review many researchers identify conflicts of interest among analysts who work at investment banks. In this study I identify conflicts of interest by studying optimism in analyst recommendations. The introduction of Regulation Fair Disclosure and the existence of Chinese walls in all types of banks should prevent analysts of pressure from their own banks to be more optimistic in their recommendations than their colleagues of other analyst firms. Despite the regulations, I find evidence of a positive bias at analyst recommendations. Although, it is possible that the introduction of Regulation Fair Disclosure reduces the positive bias in analyst recommendations.

The first hypothesis is rejected; this is supported by not finding a positive bias in the recommendations of affiliated analyst firms as opposed to unaffiliated analyst firms. Affiliated analyst firms provide investment banking services to companies included in their recommendations. Unaffiliated firms do not provide investment banking services to companies they recommend to investors. In contrast to my expectations, unaffiliated analyst firms are more optimistic in giving buy recommendations. This indicates no empirical evidence for a positive bias in recommendations of affiliated analyst firms.

The second hypothesis focuses only on recommendations of affiliated analyst firms. I study the existence of a positive bias in recommendations on companies where affiliated analyst firms provide investment banking services. I find a positive bias in buy and sell recommendations on related companies as opposed to the recommendations on non-related companies. This positive bias indicates the existence of conflicts of interest at affiliated analyst firms.

When affiliated analyst firms take a minority stake at a company they are partly owner of that company. As a shareholder they have influence on the decision-making process in these companies. Their goal is to sell this minority stake with a profit. This an explanation of the positive bias in analyst recommendations when they have a minority stake in a company.

To conclude, I find evidence for the existence of conflicts of interest in stock recommendations on the Dutch stock exchange. Analyst recommendations on Dutch AEX companies are too optimistic. Furthermore, I conclude that recommendations on companies with the indicators of involvement minority stake and public takeover are drivers of this optimism.

Except my contribution to the literature, my study offers possibilities for further research directed towards other markets. Second, future research in stock recommendations from the four indicators of involvement is possible.

6. References

Agrawal, A and M. Chen, 2008, “Do Analyst Conflicts Matter? Evidence from Stock Recommendations”, Journal of Law and Economics, Vol. 51, pp. 503-537

Barber, B. et al., 2001, “Can Investors Profit from the Prophets? Security Analyst Recommendations and Stock Returns” The Journal of Finance, Vol. 56, %o.2, pp. 531-563

Barber, B. et al., 2006, “Buys, holds, and sells: The distribution of investment banks’ stock ratings and the implications for the profitability of analysts’ recommendations”, Journal of Accounting and Economics, Vol. 42, pp. 87-117

Bradley, D. et al., 2008, “Analyst Behavior Following IPOs: The “Bubble Period” Evidence”, The Review of Financial Studies, Vol. 8, %o. 1

Chan, L. et al., 2007, “Analyst’ Conflicts of Interest and Biases in Earning Forecasts”, Journal of financial and quantitative analysis, Vol. 42, %o.4, pp. 893-914

Chang, M. et al., 2008, “The Influence of Analyst and Management Forecasts on Investor Decision Making: An Experimental Approach”, Australian Journal of Management, Vol. 33, %o.1

Clarke, J. et al., 2006, “Are Analyst Recommendations Biased? Evidence from Corporate Bankruptcies”, Journal of Financial and Quantitative Analysis, Vol. 41, %o. 1, pp. 169-196

Cowen, A. et al., 2006, “Which types of analyst firms are more optimistic?” Journal of Accounting and Economics, Vol. 41, pp. 119-146

Ferreira, E. and S. Smith, 2006, “Effect of Reg FD on Information in Analysts’ Rating Changes”, Financial Analysts Journal, Vol. 62, %o. 3, pp. 44-57

Gorman, C., 2004, “Are Chinese walls the best solution to the problems of insider trading and conflicts of interest in broker-dealers?”, Fordham Journal of Corporate & Financial Law, Vol.IX

Groysberg, B. et al., 2008, “Buy-Side vs. Sell-Side Analysts’ Earnings Forecast” Financial Analysts Journal, Vol. 64, %o. 4, pp. 25-39

Heijdenrijk, E. and A. Plantinga, 2002, “Beleggingsadviezen in Nederland”, University of Groningen

Herrmann et al., 2008, “International Diversification and Forecast Optimism: The Effects of Reg FD”, Accounting Horizons, Vol. 22, no. 2, pp 179-197

Hong, H. and J. Kubik, 2003, “Analyzing the Analysts: Career Concerns and Biased Earnings Forecast”, The Journal of Finance, Vol. LVIII, %o. 1, pp. 313-351

Houston, J. et al., 2006, “What a Difference a Month Makes: Stock Analyst Valuations Following Initial Public Offerings”, Journal of Financial and Quantitative Analysis, Vol. 41, %o. 1, pp. 111-137

Hovakimian, A. and E. Saenyasiri, 2007, “Conflicts of interest and analyst behaviour: Evidence from recent changes in regulation”, Baruch College

Jacob, J. et al., 2008, “Do Non-Investment Bank Analysts Make Better Earnings Forecast?” Journal of Accounting, Auditing and Finance

Jensen, M., 2004, “Who gets Wall Street’s attention? How alliance announcements and alliance density affect analyst coverage”, Strategic Organization, Vol. 2 pp. 293-312

Lin, H., and M. McNichols, 1998, “Underwriting Relationships, “Analysts' Earnings Forecasts and Investment Recommendations”, Journal of Accounting and Economics, Vol. 25, pp. 101-127

Ljungqvist, A. et al., 2007, “Conflicts of interest in sell-side research and the moderating role of institutional investors”, Journal of Financial Economics, Vol. 85, pp. 420-456

Mackinley, C., 1997, “Event Studies in Economics and Finance”, Journal of Economic Literature, Vol. XXXV (March 1997), pp. 13–39

Michaely, R. and K. Womack, 1999, “Conflict of Interest and the Credibility of Underwriter Analyst Recommendations”, The review of financial studies, Vol. 12, %o. 4, pp. 653-686

Newsome, P., 2005, “Ethical Issues Facing Stock Analysts”, The Geneva Papers, Vol.30, pp. 451-456

Panchenko, V., 2007, “Impact of Analysts’ Recommendations on Stock Performance”, The European Journal of Finance, Vol. 13, %o. 2, pp. 165-179

Roy, R., 2002, “Market Efficiency Effects of Regulation Fair Disclosure”, The Leonard N. Stern School of Business

Technical committee of the international organization of securities commissions (IOSCO), 2003, “Report on Analysts Conflict of Interest”.

7. Appendices

Appendix A: List of AEX companies

Aegon Philips, Kon. Ahold, Kon. Randstad Akzo Nobel Reed Elsevier ArcelorMittal Royal Dutch Shell A ASML Holding SBM Offshore BAM Groep, Kon. TNT

Corio TomTom

DSM, Kon. Unibail-Rodamco Fortis Unilever cert. Fugro USG People Heineken Wereldhave ING Groep Wolters Kluwer KPN, Kon.

Appendix B: List of removed companies at the AEX index (2003-2008)

Corporate Express July-08

Teleatlas June-08

Vedior May-08

Hagemeyer March-08

Royal Numico November-07

ABN AMRO October-07

Rodamco Europe --> take over: Unibail-Rodamco June-07 Buhrmann: name change to Corporate Express April-05

Getronics March-07

VNU May-06

Versatel October-05

Koninklijke P&O Nedlloyd August-05 TPG: name change to TNT May-05 IHC Calland: name change to SBM Offshore May-05

Van der Moolen March-05

Gucci Group May-04