• No results found

Do contributions to political election campaigns by S&P500 companies generate abnormal stock returns? : a comparison of abnormal returns over the market and over the business industry of the contributing company, using

N/A
N/A
Protected

Academic year: 2021

Share "Do contributions to political election campaigns by S&P500 companies generate abnormal stock returns? : a comparison of abnormal returns over the market and over the business industry of the contributing company, using "

Copied!
29
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Do contributions to political election campaigns by S&P500

companies generate abnormal stock returns?

A comparison of abnormal returns over the market and over the business industry of the contributing company, using evidence provided by U.S. federal election cycles from 1990 until 2010.

Bachelor thesis Economics and Business, University of Amsterdam

Student name Tobias Stornebrink Program Economics and Business

Student number 6144063 Track Finance and Organization

Supervisor Natalya Martynova Date February 2014

ABSTRACT

Previous empirical studies indicate abnormal stock performance over the market for publicly traded U.S. companies, politically connected through campaign contributions. Using an original dataset containing relevant information on all U.S. election cycles between 1990 and 2010, this paper compares abnormal stock returns over the market with abnormal stock returns over applicable business industries after election result announcement for S&P500 companies, contributing to U.S. political election campaigns. Univariate results found regarding market adjusted abnormal returns are in concordance with previous research. Univariate analysis indicates industry adjusted abnormal returns are statistically significant lower compared to market adjusted abnormal returns. Additionally, these industry adjusted abnormal returns fail to be statistically significant. Multivariate analysis of the entire panel dataset indicates the contribution behavior of S&P500 companies does not have any statistically significant post-election effect on stock returns.

(2)

2

I. INTRODUCTION

Several recent studies indicate a positive relation between corporate politically connectedness through political campaign contributions and corporate stock and accounting performance. This relation is shown not only in countries with high levels of governmental corruption, but as well in the U.S., which is generally judged to have a stable and fair government (Kaufmann et al., 2003).

Explaining the origin of the relation between stock performance and political campaign contributions by U.S. corporations, one could roughly divide the existing literature into two strands. The first strand, consisting of mainly theoretical papers, argues political campaign contributing companies will receive individual political favors from government. Anticipating on these individual favors, following the announcement of favorable election results abnormal stock performance should thus be generated by political campaign contributing companies.

The second strand of literature, consisting of mainly empirical papers, argues that political campaign contributing companies merely make campaign contributions to political candidates they beforehand know to propagate ideas and plans potentially beneficial for the company. Thus, this second strand of literature opposes the idea of political campaign contributions as a bid to individual favors, but argues political campaign contributions by corporations merely are a means to support a political candidate propagating potentially beneficial ideas.

Empirical research by Knight (2006) shows business industry wide abnormal stock returns after a Republican election win. After this Republican election win industries classified by Knight (2006) to benefit from a Republican administration show positive abnormal returns over the market, whilst industries classified to benefit from a Democratic administration show negative abnormal returns over the market (Knight, 2006). Thus, the question is raised whether the by previous research reported abnormal stock returns over the market for U.S. companies politically connected through political campaign contributions, are indeed a result of these donations, or merely are a result of an industry specific effect. The intention of this paper is to address this question, making use of an original dataset, containing information on campaign contributions and stock returns of S&P500 companies, on every election cycle, both Presidential and Congressional, in the period of 1990 until 2010.

For every election cycle in the 1990 - 2010 period, all campaign contributions by S&P500 companies are gathered. Next, cumulative abnormal returns (CARs) are calculated for the contributing companies after the announcement of each election result in the sample period. These CARs are both determined using market normal returns, and using applicable business industry normal returns.

In an univariate setting, linear regression analysis of the panel data in the original dataset of this paper shows statistically significant positive abnormal returns over the market for companies

(3)

3

primarily contributing to Republican election campaign, after the announcement of a Republican election win. After the announcement of a Democratic election win these companies are shown to exhibit negative abnormal returns over the market. These results are in concordance with previous related empirical studies. However, performing the same univariate analysis on the industry weighted CARs, the abnormal returns become statistically insignificant. Additionally, industry weighted CARs are shown to be statistically significant lower than their market weighted peers.

Comparable results are reported in a multivariate regression analysis of the single election cycle of 2004. The influence of explanatory variables indicating the campaign contribution behavior of a company are shown to be of statistically significant influence on the market weighted CAR. The influence on of these explanatory variables on the industry weighted CAR is substantially lower, and not statistically significant.

Hence, this paper provides clear empirical evidence abnormal returns of political campaign contributing companies after election result announcements do not exist when abnormal returns are calculated using business industry specific normal returns. Thus, political campaign contributing companies do not show abnormal returns over the market for making campaign contributions, but merely show abnormal returns over the market since they are part of a specific business industry that is believed to perform in a certain way under the election winning political party. Based on this result, this paper believes to contribute to the strand of literature opposing the idea of U.S. companies attaining individual governmental favors by making contributions to U.S. political election campaigns.

Performing linear multivariate regression analysis on the entire panel dataset, this paper finds explanatory variables indicating the political campaign contribution behavior of a company do not exhibit statistically significant influence on the CAR of a contributing company following election result announcements. This influence is found to be not statistically significant when CARs are calculated using market normal returns, nor when CARs are calculated using applicable business industry specific normal returns. Thus, analysis of the entire panel dataset shows no statistically significant influence of the political campaign contributing behavior of a S&P500 company on the stock performance of that company following announcements of election results. From this, it could be implied that political campaign contributions by companies simply do not have any effect on the stock performance of a company in the U.S. Or additionally, it could be implied that campaign contributions are no accurate measure of the politically connectedness of a U.S. company. Further research could supply answers on this matter.

The remainder of this paper is organized as follows: section II. provides an overview of related previous empirical and theoretical literature. Section III. explains the research methodology applicable in this paper. Section IV. contains a description of the original dataset used in this paper.

(4)

4

Section V. presents the results of the empirical research conducted in this paper. A short discussion of these results is provided in section V. as well. Section VI. concludes. Except for table 1, which is included in section IV., all tables reporting results are added to this paper as appendices.

II. LITERATURE REVIEW

A substantial amount of studies has been conducted, explaining the relation between corporate political connectedness through political election campaign contributions and accounting and stock performance of these politically connected companies. As does this present paper, a substantial branch of these studies focuses on the merits of making political campaign contributions for contributing companies. As mentioned in the previous section, this branch of literature covering the question why companies make political election campaign contributions can roughly be divided into two opposing strands.

The first strand of literature explaining why companies make political campaign contributions consists of mainly theoretical studies. These theoretical studies generally consider corporate political campaign contributions to be a bid by contributing companies, on buying governmental favors. By making political campaign contributions, this theoretical strand of literature argues, corporations attain individual, company specific favors from legislators. Hence, companies are induced to make political campaign contributions.

The second strand of literature explaining why companies make political campaign contributions consists of mainly empirical studies. Generally, these empirical studies oppose the idea supplied by the theoretical literature mentioned above, of companies attaining individual favors from government by making political campaign contributions. In contrast to this theoretical idea, most empirical studies conclude companies make political campaign contributions to political candidates they beforehand know to propagate ideas potentially beneficial to the company.

As mentioned above, most theoretical papers explaining the political campaign contributing behavior of corporations, indicate corporate political campaign contributions to be a bid by a corporation, attempting to attain individual governmental favors. This predominant idea in theoretical studies related to this paper is stated in a literature review conducted by Stratmann (2005). As stated by Stratmann (2005), most models used in theoretical studies explaining why corporations make political campaign contributions, assume political campaign contributions by corporations to be an attempt by the contributing corporation to enforce favors by the supported political candidate.

A study by Coate (2004a) supplies a theoretical model indicating the possibility for political candidates of offering policy favors to corporations, in order to receive campaign contributions from those corporations in return. Coate (2004a) argues political candidates should be able to attract

(5)

5

additional campaign financing by offering additional policy favors to corporations. Based on the theoretical model provided by Coate (2004a), political campaign contributions by corporations are thus a means to attain political favors by those corporations. The more campaign contributions a corporation makes, the more political favors it should expect to receive. The theoretical model provided by Coate (2004a) is in concordance with the literature review conducted by Stratmann (2005), stating corporate political campaign contributions generally are viewed in theoretical literature to be a bid on political favors.

Similar views on corporate political campaign contributions are provided by theoretical studies by Grossman and Helpman (1994) and Ashworth (2006). Grossman and Helpman (1994), define corporate contributions to political election campaigns as an attempt by the contributing corporation to influence governmental policy. Grossmann and Helpman (1994) view campaign contributions by corporations as an auction bid, trying to buy protection from the receiving political candidate. Hence, as suggested by Coate (2004a), in their theoretical model Grossmann and Helpman (1994) suggest increasing contributions to political election campaign by a corporation, should result in increasing governmental policy favors supplied to that corporation. Similar conclusions are made by Ashworth (2006). In addition, Ashworth (2006) claims the idea of corporations buying policy favors by making contributions to election campaigns to be not purely academic. Referring to the Bipartisan Campaign Reform Act of 2002, Ashworth (2006) claims fear for policy influencing corporations has its effect on campaign financing laws.

As mentioned above, a strand of mainly theoretical studies explains political campaign contributions by corporations to be an attempt of the contributing corporation to attain policy favors provided by the supported political candidate. However, as explained earlier in this section, a strand of mostly empirical literature appears to oppose this theoretical idea. Alternative empirical explanations of political campaign contributions by corporations is supplied by e.g. Bronars and Lott (1997) and Ansolabehere et al. (2003).

Bronars and Lott (1997) conduct an empirical study covering the last term U.S. legislators are in office, due to upcoming retirement. Due to their upcoming retirement, these legislators no longer have to keep their reelection in mind, and thus are no longer bound by the influence of the threat of potential loss of contributions to their individual reelection campaign. Thus, Bronars and Lott (1997) argue, campaign contributions should not influence the voting behavior of these legislators.

Comparing to previous office terms, Bronars and Lott (1997) conclude that legislators do not change their voting behavior when arrived in their last office term before retirement. Thus, while the threat of a loss in campaign contributions is eliminated, legislators do not change their voting behavior. From this, Bronars and Lott (1997) conclude that political campaign contributions by corporations do not influence policy decisions made by the supported legislators. Hence, companies

(6)

6

do not make political campaign contributions to influence governmental policies, but merely make contributions to politicians they beforehand know to value the same strand of policies as beneficial for the company. In concordance with Bronars and Lott (1997), Ansolabehere et al. (2003) perform empirical research on corporate contributions to U.S. political campaigns and find no evidence that contributions influence the voting behavior of legislators.

In addition to the strand of empirical studies mentioned above, Coate (2004b) provides a theoretical study using a theoretical model opposing the idea of corporate political campaign contributions as a means for corporations to buy governmental policy favors. As opposed to pursuing policy favors, in the theoretical model presented by Coate (2004b) corporations are believed to make campaign contributions to political candidates propagating ideas and plans valued by the contributing company. Thus, in concordance with the empirical indications mentioned by Bronars and Lott (1997) and Ansolabehere et al. (2003), the theoretical study by Coate (2004b) assumes companies do not make political campaign contributions in order to attain individual policy favors, but merely make contributions to politicians they beforehand know to value a strand of policies and ideas potentially beneficial to the company.

As mentioned earlier in this section, a strand of empirical literature exists opposing the theoretical idea of corporations attempting to purchase governmental policy favors by making political campaign contributions. In addition, a substantial amount of empirical studies have been conducted showing a positive relation between firm value and the politically connectedness of a firm. Whilst not explaining the origin of this relation, the empirical indication of the existence of a relation between corporate politically connectedness and corporate stock performance, is highly relevant for this paper.

Some of these empirical studies indicating a relation between corporate politically connectedness and corporate stock performance have been conducted in developing countries as Indonesia. Fisman (2001) identifies Indonesian companies that are connected to the Indonesian government, using the Suharto Dependency Index. Following the death of President Suharto, Fisman (2001) shows a significantly larger decline in equity value of a company the more the company is connected to the Indonesian government. A similar empirical study has been conducted for companies in the U.S., following the unexpected death of the Democratic Senator for the State of Washington, Henry Jackson in 1983. After the death of Senator Jackson, Roberts (1990) finds significant negative abnormal returns over the market for companies that contributed to the election campaign of the Senator.

Jayachandran (2006) assesses the announcement effect of Senator Jim Jeffords leaving the Republican party to become an Independent. Due to Jeffords leaving the Republican party, the Democratic party gained the majority of the U.S. Senate. As did Roberts (1990) following the

(7)

7

announcement of the dead of Senator Jackson, Javochandran shows significant negative (positive) abnormal stock returns for companies which contributed to Republican (Democratic) campaigns in the previous election cycle, following the announcement of Senator Jim Jeffords leaving the Republican party.

Several recent studies show a more direct positive relation between a company’s politically connectedness and its stock market and accounting performance in the U.S., highly relevant for this paper. For example, Cooper et al. (2010) show a positive, significant correlation between corporate contributions to U.S. political campaigns and future stock returns for publicly listed companies in the period 1979 - 2004. Goldman et al. (2009) show a positive effect of politically connected board members on firm value of S&P500 companies. Boubakri et al. (2009) show that investors require a lower cost of capital for politically connected firms than for their unconnected peers.

Using data from the Federal Election Commission (F.E.C.), Cooper et al. (2010) create their own original dataset; capturing 70% of all campaign contributions made by U.S. publicly traded companies in the period 1979 - 2004. Cooper et al. (2010) measure politically connectedness by the number of political candidates a firm supports. Using panel data regression analysis Cooper et al. (2010) show a significant positive relation between future abnormal returns and the number of candidates a firm supports. This relation is stronger for firms supporting more powerful candidates and for firms supporting the same candidates over a longer period of time. When the dataset is split between Democratic and Republican candidates, Cooper et al. (2010) show a stronger effect for contributions made to a Democratic candidate. However, Cooper et al. (2010) do not explain the origin of the contribution effect they describe.

Goldman et al. (2009) define the politically connectedness of a company by the contributions a company makes to political election campaigns, as does this paper. In addition, Goldman et al. (2009) define corporate politically connectedness by the political background of the board members of a corporation. Using a hand collected dataset on board compositions of S&P500 companies, Goldman et al. (2009) perform an event study on the Republican win of the 2000 U.S. Presidential election. In the period following the election, stock of S&P500 companies classified as having a Republican board are shown to significantly outperform stock of S&P500 companies classified as having a Democratic board. Additionally, portfolios of Republican classified companies show significant positive cumulative abnormal returns over the market, after the Republican Presidential nomination.

In addition to a relation between firm value and corporate politically connectedness through campaign contributions and politically connected board members, in their paper Goldman et al. (2009) show a significant positive effect on the stock price of a S&P500 company following the nomination announcement of a new politically connected board member. This announcement effect

(8)

8

is reported for both Democratic and Republican connected board member nominees. Testing for non-connected nominees, Goldman et al. (2009) find that the announcement of the nomination of a non-politically connected board member does not generate any significant abnormal stock returns. From this, Goldman et al. (2009) conclude that the board of directors can add value to a company by the political connections it provides.

Measuring the politically connectedness of a company by political campaign contributions, Goldman et al. (2009) show statistically significant positive abnormal stock returns over the market for S&P500 companies donating primarily to the Republican election campaign, after the 2000 Republican Presidential nomination. Companies donating primarily to the Democratic election campaign are shown to display negative abnormal stock returns over the market after the same Republican Presidential nomination. These results are in concordance with the conclusions by Cooper

et al. (2010), which are derived from a larger sample of contributing firms and elections.

Both the results by Cooper et al. (2010) and Goldman et al. (2009) suggest a relation between company value and its politically connectedness by contributions to political campaigns. However, instead of by a direct relation between campaign contributions and company performance, this paper intends to explain the relation found by Cooper et al. (2010) and Goldman et al. (2009) by making use of evidence presented in a previous empirical study conducted by Knight (2006).

During the six months before the 2000 U.S. Presidential election, Knight (2006) shows business industry wide reactions of equity prices to the changing probability of a Republican (George W. Bush) Presidential nomination. Data on the probability of a Republican win are derived from the

Iowa Electronic Market, an on-line future market run by the University of Iowa, where

contract-payoffs are based on current events, like political announcements and election outcomes (IEM, 2014).

Knight (2006) argues the specific policies proposed by the Democrats and the Republicans have their effect on specific corporate industries. According to Knight (2006), industries that would benefit from the policies proposed by the Republican party increase in equity value when the probability of a Republican Presidential nomination increases. Industries that would benefit from the policies proposed by the Democratic party decrease in equity value when the probability of a Republican Presidential nomination increases. Illustrating this business industry wide reaction on political election results, Knight (2006) shows the entire tobacco industry displays a positive return of 13% after a Republican Presidential nomination, whilst the alternative energy industry shows a negative return of 16% after that same Republican Presidential nomination.

The results described by Knight (2006) suggest industry wide stock market reactions to political election outcomes. Knight (2006) contributes these stock market reactions to policy platforms advocated by a political candidate which will potentially benefit or harm an industry as a

(9)

9

whole. This paper expects to show that the connection between contributions to political campaigns by corporations and the equity prices of those corporations, described by for example Cooper et al. (2010) and Goldman et al. (2009), can alternatively be explained by the results of Knight (2006). Thus, this paper expects to show stock price reactions after election outcomes merely result from industry expectations about expected policy changes, instead of from the politically connectedness of an individual company through campaign contributions.

The recent papers by Cooper et al. (2010) and Goldman et al. (2009) both show significant abnormal positive stock returns over the market for individual politically connected companies. As previously been done by Goldman et al. (2009) for the 2000 U.S. Presidential election, this paper as well expects to show significant abnormal stock returns over the market after the announcement of election results, for companies that are politically connected through political campaign contributions. Compared to the paper by Goldman et al. (2009), this paper will use a much broader range of elections.

Additionally, based on the results by Knight (2006) explained above, this paper expects to find campaign contributing companies do not show abnormal stock returns after the announcement of political election results, relatively to their specific business industry return. A similar result is found by Goldman et al. (2009). For the 2000 U.S. Republican Presidential nomination Goldman et al. (2009) show that no significant positive (negative) abnormal stock returns exists for S&P500 companies contributing to the Republican (Democratic) Presidential election campaign, when comparing with their industry peers. In addition to the paper by Goldman et al. (2009) this paper will use a broader range of elections. Thus, this paper will be able to draw more generalized conclusions.

Based on arguments explained above, this paper expects to find corporations making contributions to political election campaigns, do not show abnormal returns after election result announcements, when these abnormal returns are calculated using applicable business industry normal returns. When campaign contributing companies do not show abnormal returns over their business industry peers, no company-specific governmental favors should result from corporate political campaign contributions. Thus, this paper expect to contribute to the strand of literature mentioned earlier in this section, opposing the idea of corporations intending to buy specific individual political favors by making political campaign contributions.

III. METHODOLOGY

As explained in the previous section, this paper expects to show significant abnormal returns over the market after the announcement of federal election results in the period 1990-2010, for S&P500 companies politically connected through campaign contributions. Based on previous research these abnormal returns are expected to be positive when the election result is favorable for the political

(10)

10

party a company is most connected to, and negative otherwise. In addition, as stated in the previous section, the expectation of this paper is to contribute to the strand of literature opposing the idea that companies are able to attain specific individual political favors, by contributing to political election campaigns. For this, this paper expects to show no abnormal returns exist for individual companies after the announcement of an election result, when correcting the return of a contributing company with the return of its specific business industry.

Both hypotheses are tested by conducting an event study as described by MacKinlay (1997). This paper conducts a panel data regression analysis, using the following regression equation. The announcement of every separate election cycle result during the period 1990-2010 is included in the analysis of this paper as a separate event. The explanatory variables in the regression equation are explained further in this section. Year fixed effects are assumed in the regression analysis. Fixed effects are also assumed in the highly related regression analysis performed in the paper by Cooper

et al. (2010). Cooper et al. (2010) conduct robustness checks using other model specifications,

leaving their results to remain unchanged.

REGRESSION EQUATION

CAR(%) = β₀ + β₁dDonDem + β₂dDonWeakRep + β₃dDonStrongRep + β₄dREPSWIN + β₅(dREPSWIN x dDonDem) + β₆(dREPSWIN x dDonWeakRep) + β₇(dREPSWIN x dDonStrongRep) + β₈*dCongressional + β₉*SIZE + β₁₀*BtM + β₁₁*Employees + β₁₂*DON/REV α + u

The dependent variable in the regression equation stated above is formed by the cumulative abnormal return (CAR) of a political campaign contributing company, following the announcement of the result of a U.S. Presidential or Congressional election during the period 1990-2010. Abnormal returns for each event are calculated separately using market normal returns and applicable business industry normal returns. Market adjusted returns are calculated using the value-weighted index from the CRSP database, applicable for the specific event. As shown by Brown and Warner (1980) the value-weighted index and the equally-weighted index from the CRSP database are expected to give similar results. Additionally, Goldman et al. (2009) perform their analysis related to this paper using both the value-weighted and the equally-weighted index from the CRSP database to calculate abnormal returns. Goldman et al. (2009) report no significant difference in results when using the value-weighted or the equally-weighted index. Abnormal returns over specific business industries are determined using the value-weighted industry returns, as stated in the Fama-French 30 Industry Portfolio.

In order to assess the existence of abnormal returns, when comparing to the specific business industry of a political campaign contributing company, theses contributing companies are

(11)

11

classified in business industries. These business industries are defined using the Fama-French 30 Industry Portfolios, published on the website of Kenneth R. French at Darthmouth University. The Fama-French 30 Industry Portfolios are determined using the Standard Industrial Classification (SIC) codes granted to a company in the Compustat database, whenever available. When Compustat SIC codes are not available, the Fama-French 30 Industry Portfolios identifies the business industry of a company by its SIC code in the CRSP database. In analogy with the Fama-French 30 Industry Portfolios this paper will identify the specific business industry of the companies in its original dataset by the SIC codes of those companies in the Compustat database. For each company in the dataset of this paper, SIC codes and the associated Fama-French 30 Industry Portfolios business industry classification are determined for each election cycle separately.

Possible abnormal returns are calculated for each event, using an event window of 1 day after, until 3 days after the event. In the analysis related to this paper, Goldman et al. (2009) use an event window of (+1,+3) days after the event as well. Goldman et al. (2009) perform robustness checks of the analysis related to this paper, using different event windows. Results derived from the analysis using these different event windows remain similar.

Abnormal returns over the market and over applicable business industries after the announcement of an election result are combined to form cumulative abnormal returns for each company in the sample. CARs are formed for the event window, for each event. Thus, each company maximally has a set of 10 CARs over the market and 10 CARs over the business industry of the company, depending on the amount of election cycles it is included in the sample.

All cumulative abnormal returns of included companies are combined in one panel dataset. The CARs in this dataset are checked for significance and their origin is explained using the regression equation stated above. The combined panel dataset of CARs being the dependent variable in this regression analysis. CARs weighted by the market normal return and CARs weighted by the applicable business industry of a contributing company are analyzed separately. Differences in results when using the market weighted or industry weighted CARs are analyzed for statistical significance.

This paper finds that most companies in its original dataset contributing to political campaigns, are contributing to campaigns of both major U.S. political parties. This is in concordance with earlier empirical research, e.g. by Cooper et al. (2010). The fact that most politically contributing companies donate to both the Republicans and the Democrats implies the formation of different contribution portfolios.

As done by Goldman et al. (2009), this paper assigns contributing companies to contribution portfolios based on their relative contribution behavior. The contribution portfolios are incorporated

(12)

12

into the regression equation using the dummy variables dDonDem, dDonWeakRep and

dDonStrongRep. These dummy variables are granted a value of ‘1’ when applicable.

Political campaign contributing companies are assigned to a contribution portfolio based on the relative share of their total campaign contributions to Democratic and Republican election campaigns. Companies are assigned to contribution portfolio DonDem when their relative share of campaign contributions to Democratic election campaigns exceeds their relative share of campaign contributions to Republican election campaigns. All other companies are assigned to contribution portfolios DonWeakRep and DonStrongRep. Contribution portfolios DonWeakRep and DonStrongRep thus contain companies relatively predominantly making contributions to Republican election campaigns. Contribution portfolios DonWeakRep and DonStrongRep contain an equal amount of companies. Contribution portfolio DonStrongRep contains those companies with the highest relative share of contributions to Republican election campaigns.

This paper includes the variable DON/REV in its regression analysis. DON/REV is an explanatory variable, covering the political contributions made by a company relative to the revenue made by that company during the applicable election cycle. The variable DON/REV is measured as total contributions made by a company during an election cycle, divided by the total revenue of that company during the same election cycle.

This paper claims to be the first to include a variable such as DON/REV, compared to related empirical studies. As shown in previous studies, corporate contributions are an important source of campaign finance for political candidates (Snyder, 1990). Combining this with the theory that corporations can influence governmental policies by making political campaign contributions (Grossmann and Helpman, 1994), elected candidates should be expected to make particularly large favors to their biggest contributors. Since companies donating a large share of their revenue to election campaigns, potentially can become larger contributors in case their revenue increases, increasing revenue of those companies could be beneficially for contribution receiving candidates. If this theory were to hold, increasing contributions relative to revenue should lead to increasing cumulative abnormal returns after the announcement of an election result.

The dummy variable dREPSWIN measures the effect of a Republican election win on the dependent variable. This dummy variable dREPSWIN is combined with the dummy variables indicating the contribution portfolio of a contributing company. These combined explanatory variables are used to measure the effect of a Republican election win on the cumulative abnormal return of a contributing company, assigned to a certain contribution portfolio.

Cooper et al. (2010) indicate a stronger influence of political campaign contributions on stock returns for companies contributing to Democratic election campaigns compared to companies contributing to Republican election campaigns. Therefore, as a robustness check, the regression

(13)

13

analysis described in this section will additionally be performed replacing the dummy variable

dREPSWIN, with a similar dummy variable dDEMSWIN, indicating a Democratic election win.

The nature of an election cycle, Presidential or Congressional, is included in the regression equation using the dummy variable dCongressional. dCongressional is appointed a value of ‘1’ when a cumulative abnormal return is measured during an announcement event following a Congressional election cycle.

The explanatory variables in the equation, SIZE, BtM and Employees describe the general company characteristics market capitalization, book to market ratio and the number of employees, of the companies in the sample. These variables are derived from earlier empirical studies related to this study (Cooper, 2010) (Goldman, 2009).

IV. DATA DESCRIPTION

This paper uses data on political campaign contributions by S&P500 companies for the period 1990 – 2010. Data on campaign contributions is derived from the website of the Center for Responsive Politics (CRP). According to the Federal Election Campaign Act, funds raised by and spent on political campaigns are to be publicly disclosed via the Federal Election Commission (F.E.C.) (F.E.C., 2014). The CRP collects, processes and publishes data on corporate political campaign contributions made publicly available through the F.E.C..

This paper intends to evaluate the effect of politically connectedness through political campaign contributions on the stock price of S&P500 companies for the period 1990 – 2010. To take into account the changing nature of the S&P500 index a random selection of two historical S&P500 indexes is made. The S&P500 index of January 1995 and the S&P500 index of January 2005 are merged to form the database containing the contributing companies in the sample of this paper.

After merging the two selected historical S&P500 indexes, CUSIP identifiers are used to check for duplicate companies in the two index lists. Deleting duplicate CUSIP identifiers results in a list of 725 unique companies. The CRP does not provide contribution data on companies that are presently an affiliate of another company. Hence, all companies that presently are an affiliate of another company are excluded from the sample. Additionally, companies not clearly identifiable in the CRP database and companies on which the CRP database does not contain sufficient information are excluded from the sample. After these exclusions, a database of 618 unique companies remains.

Unfortunately, the CRP database only contains information on companies which were politically active through lobbying or through political campaign contributions in the 2012 election cycle. On all other companies, the CRP database does not document political contribution profiles. Due to this lack of available data, additional exclusions have to be made, ultimately resulting in a dataset of 413 unique sample companies. This set of exclusions due to a lack of available information

(14)

14

may lead to the exclusion of companies which did make political contributions in the period 1990 – 2010, but were not politically active in 2012.

The Bipartisan Campaign Reform Act of 2002 banned all soft money donation to federal political election campaigns (F.E.C., 2014). However, the database of the CRP does include soft money contributions on election cycles before 2002.

Table 1 provides core descriptive statistics of the contributing companies in the dataset of this paper. As described earlier in this paper, most campaign contributing companies in the original dataset of this paper make campaign contributions to both major U.S. political parties. Taking all contributions by all contributing company in the dataset, covering the entire sample period, 11% of all specific contribution preferences of a company, is a 100% relative donation to either the Democratic or Republican election campaign. This is in concordance with empirical evidence presented in previous papers (Cooper et al., 2010) (Goldman et al., 2009).

Table 1

Descriptive statistics of different contribution portfolios

Portfolio N Average total donation (x1000$) Average % to Republican Campaign Minimum total donation (x1000$) Maximum total donation (x1000$) a) Analysis with 3 contribution portfolios

DonDem 1094 372 30,4 0,20 7191

DonWeakRep 1282 504 63,6 0,45 9355

DonStrongRep 1271 200 89,3 0,20 4003

b) Analysis with 4 contribution portfolios

DonStrongDem 549 363 30,0 0,20 5159

DonWeakDem 545 380 30,1 0,30 7191

DonWeakRep 1282 504 63,6 0,45 9355

DonStrongRep 1271 200 89,3 0,20 4003

Most companies contributing to political election campaigns, contribute to election campaigns of both the Republican party and the Democratic party. However, the relative contributions made to the election campaigns of the Republicans and the Democrats do differ among the contributing companies in the sample. This implies the formation of contribution portfolios, as done by Goldman et al. (2009). These contribution portfolios are determined for every election cycle in the sample period separately. Contributing companies are assigned to a contribution portfolio based on the relative share of their contributions to a certain political party, compared to their total contributions in a specific election cycle.

(15)

15

In table 1, DonDem comprises the contribution portfolio of contributing companies relatively contributing more to Democratic election campaigns compared to Republican election campaigns. Companies contributing more to Republican election campaigns, are assigned to contribution portfolios DonWeakRep and DonStrongRep. Each of these contribution portfolios contains an equal amount of contributing companies, with DonStrongRep containing the contributing companies relatively donating the highest share of their total donations to the Republicans.

As a robustness check, this paper includes a second analysis in which contributing companies are spread over four different contribution portfolios. In this analysis, contributing companies contributing relatively predominantly to Democratic election campaigns, are spread over contribution portfolios DonWeakDem and DonStrongDem in analogy with the spread of companies contributing predominantly to Republican election campaigns over contribution portfolios DonWeakRep and DonStrongRep, explained above. Core descriptive statistics of this additional analysis using four contribution portfolios are stated in Table 1.

V. RESULTS AND DISCUSSION

UNIVARIATE RESULTS

This subsection describes stock price reactions after the announcement of political election results for S&P500 companies politically connected through political campaign contributions, using a univariate approach. Results derived using a multivariate approach are discussed in the following subsection.

As stated in table 2, contributing companies in the dataset assigned to Republican oriented contribution portfolios DonWeakRep and DonStrongRep experience positive abnormal returns over the market after the announcement of a Republican election win. The abnormal returns over the market after a Republican win for companies in contribution portfolios DonWeakRep and DonStrongRep are highly statistically significant. The more predominantly Republican contribution portfolio DonStrongRep clearly shows higher abnormal returns after a Republican election win than does the more moderately Republican oriented contribution portfolio DonWeakRep. These results are in concordance with previous related studies by e.g. Goldman et al. (2009).

Comparing their post-announcement returns with their specific business industry returns, contributing companies assigned to contribution portfolios DonWeakRep and DonStrongRep still show positive abnormal returns after the announcement of a Republican election win. However, as opposed to the related market weighted abnormal returns, these abnormal returns over the applicable business industry are not statistically significant for contribution portfolio DonStrongRep and notably less statistically significant for contribution portfolio DonWeakRep. Additionally, contribution portfolios DonWeakRep and DonStrongRep abnormal returns over specific business

(16)

16

industries after the announcement of a Republican election win are considerably lower compared to the related abnormal returns over the market. The difference between abnormal returns over the market and over applicable business industries after election result announcements is stated in table 4. As shown in table 4, the difference between post-announcement abnormal returns over the market and over the industry after a Republican election win is highly statistically significantly different from zero for contribution portfolio DonStrongRep.

As stated in table 2, after a Democratic election win, contribution portfolio DonWeakRep and DonStrongRep show negative cumulative abnormal returns over the market. Contribution portfolio DonDem shows a slightly positive cumulative abnormal return over the market in the case of a Democratic election win. However, these abnormal returns over the market after the announcement of a Democratic election win are not statistically significant. Similarly, abnormal returns over applicable business industries are statistically insignificant after a Democratic election win, for all contribution portfolios.

Insignificant abnormal returns after a Democratic election win can possibly be explained by the relatively lower amount of companies in the sample of this study, primarily contributing to Democratic election campaigns. The amount of companies that has a distinct relative Democratic contribution preference is even smaller. These contribution preferences are found as well in previous related research by Goldman et al. (2009). In the paper by Goldman et al. (2009), abnormal returns over the market after a Republican election win are found to be negative for companies primarily contributing to Democratic election campaigns. However, in concordance with results in this paper, these negative abnormal returns are not statistically significant.

In addition to the low amount of companies with a Democratic contribution preference, the statistical insignificance of abnormal returns after a Democratic election win can possibly be explained by results found by Cooper et al. (2010). Cooper et al. (2010) perform empirical research on the effect of politically connectedness through campaign contributions on the stock prices of those politically connected companies. They report an incremental abnormal return for companies contributing to Democratic election campaigns, controlling for Republican contribution effects. Hence, since most companies in the sample of this paper make political campaign contribution to both major U.S. political parties, after a Democratic election win the expected negative abnormal returns over the market for contribution portfolios DonWeakRep and DonStrongRep could be offset by the incremental Democratic contribution effect reported by Cooper et al. (2010). Thereby causing statistically insignificant abnormal returns after a Democratic election win.

In their analysis related to this paper, Goldman et al. (2009) use evidence from a single election cycle. In order to make a more direct comparison of results in this paper, with results in the highly related analysis by Goldman et al. (2009), table 3 states cumulative abnormal returns for

(17)

17

different contribution portfolios after the 2004 Republican election win. Information on all other election cycles in the original dataset of this paper is excluded for the 2004 election cycle analysis in table 3.

In concordance with results found by Goldman et al. (2010), table 3 states positive abnormal returns over the market for contribution portfolios DonWeakRep and DonStrongRep after the announcement of the Republican election win. The abnormal return over the market of contribution portfolio DonDem is negative after the same announcement. The abnormal returns of contribution portfolios DonWeakRep and DonStrongRep are statistically significant at the 1% level.

In addition to the empirical analysis performed by Goldman et al. (2009), this paper checks for the existence of cumulative abnormal returns after the announcement of election results over the business industry of a political campaign contributing company. As shown in table 3 for the 2004 election cycle, abnormal returns for contribution portfolios DonWeakRep and DonStrongRep are considerably lower when calculated over applicable business industries instead of over the market. The abnormal return of contribution portfolio DonWeakRep becomes statistically insignificant when calculated over applicable business industries. The abnormal return of contribution portfolio DonStrongRep becomes statistically significant at a lower level when calculated over the industry. Abnormal returns for the Democratic contribution portfolio DonDem are not statistically significant when calculated over the market, nor when calculated over the industry. This could possibly be due to the same reasons explained above for statistically insignificant abnormal returns of contribution portfolio DonDem in table 2.

The differences in abnormal returns of contribution portfolios DonWeakRep and DonStrongRep when calculated over the market and over applicable business industries are reported in table 4 for the announcement of the 2004 Republican election win. As stated in table 4, this difference is statistically significant for both contribution portfolio DonWeakRep and DonStrongRep.

In a univariate setting, this paper reports statistically significant abnormal returns over the market after the announcement of a Republican election win for companies relatively contributing primarily to Republican election campaigns. However, abnormal returns over the specific business industry of the contributing companies are reported to be statistically significant lower for the same companies after the same election result announcement. These results hold both for a multiyear panel data analysis of all election cycles in the period 1990 - 2010 and for an analysis of the single election cycle of 2004.

Abnormal returns after the announcement of a Democratic election win are reported to be not statistically significant for all contribution portfolios. This result holds for both the panel data analysis of the complete sample and for the analysis of the single 2004 election cycle. The statistically insignificance of these abnormal returns could possibly be due to the relatively low number of

(18)

18

companies in the dataset primarily contributing to Democratic election campaigns, or to the incremental Democratic contribution effect reported by Cooper et al. (2010).

MULTIVARIATE RESULTS

This section describes stock price reactions after the announcement of political election results for S&P500 companies politically connected through political campaign contributions, using a multivariate approach. Results are reported in tables 5, 6 and 7.

Table 5 reports results for various models assuming a Republican election win. Table 6 does the same for a Democratic election win. Both tables show comparable results. Based on previous related empirical research by e.g. Goldman et al. (2009), Republican oriented contributing portfolios DonWeakRep and DonStrongRep are expected to show statistically significant positive (negative) influence on the cumulative abnormal returns over the market following the announcement of a Republican (Democratic) election win. Democratic oriented contribution portfolio DonDem is expected to show negative (positive) influence on the cumulative abnormal returns over the market following the announcement of a Republican (Democratic) election win.

Contradicting expectations derived from Goldman et al. (2009), as stated in tables 5 and 6, this paper shows when using a panel data linear regression approach with a large set of election cycles, no statistically significant influence on abnormal returns over the market exist for any contribution portfolio. Since in both table 5 and table 6 no statistically significant influence is reported for explanatory variables indicating contribution portfolios, the political campaign contributing behavior of a company is not of influence on the return of that company after the announcement of a federal election result, regardless of the nature of that election result.

Panel b) in both table 5 and table 6 reports panel data linear regression results, appointing cumulative abnormal returns over specific business industries after the announcement of election results to be the dependent variable. As stated in table 5 and 6, when using the industry weighted CAR(%)s as the dependent variable, the influence of the independent contribution portfolio variables remains statistically insignificant. The influence of the independent election result variables

dREPSWIN and dDEMSWIN becomes less and statistically significant at a lower level, when compared

to the same panel data regression analysis using the market weighted CAR(%)s as the dependent variable.

In panel a) and panel b) of both table 5 and table 6, the explanatory variable DON/REV is reported to have no statistically significant influence on the dependent variable. In section III of this paper, the theory is presented based on theoretical studies, an increasing ratio between total political campaign donations made by a company and total revenues of that company in the same election cycle, should lead to an increasing abnormal return following election result

(19)

19

announcements. Since the variable DON/REV is shown to have no statistically significant influence on abnormal returns following election result announcements, this theory presented in section III of this paper, is rejected.

Table 7 reports multivariate linear regression results using cumulative abnormal returns after the announcement of the 2004 election result as the dependent variable. Data on all other election cycles is excluded from the analysis performed in table 7. Panel a) in table 7 states results for multiple regression models using the market weighted CAR(%) as the dependent variable. As opposed to in the panel data results in table 5 and table 6, the independent variables indicating contribution portfolios are statistically significant in panel a) of single election cycle table 7.

In concordance with the related previous empirical research by Goldman et al. (2009), panel

a) of table 7 shows a statistically significant positive influence on the cumulative abnormal return

over the market for the Republican oriented contribution portfolio DonRep, after the announcement of the 2004 Republican election win. The influence on the cumulative abnormal return for the Democratic oriented contribution portfolio DonDem is reported to be negative and highly statistically significant, as in the previous research by Goldman et al. (2009).

However, as stated in table 7, the regression models in panel a) of table 7 all show a very low R-squared, indicating a low representation of the variance of the dependent variable by the regression model (Cameron and Windmeijer, 1997). Low R-squared values are reported by the related regression models by Goldman et al. (2009) as well.

Panel b) of table 7 reports linear regression models using the applicable business industry weighted CAR(%) after the announcement of the 2004 Republican election win as the dependent variable. Compared to results in panel a) of table 7, the influence of the explanatory contribution portfolio variables dDonRep and dDonDem on the CAR(%) is substantially smaller in panel b). Additionally, the explanatory contribution portfolio variables dDonRep and dDonDem are statistically insignificant in industry weighted panel b), whereas they are highly statistically significant in market weighted panel a).

Hence, panel 7 indicates results comparable to those mentioned in the section of this paper presenting univariate results. In the univariate section CAR(%)s are reported to be statistically significantly lower when calculated over applicable business industry, compared to when calculated over the market. Industry weighted CAR(%)s are often reported to be statistically insignificant whereas market weighted CAR(%) are found to be highly statistically significant. Similarly, in multivariate table 7, the influence of explanatory contributing portfolio variables dDonRep and

dDonDem on the industry weighted CAR(%)s is reported to be substantially lower and statistically

insignificant compared to their statistically highly significant influence on the market weighted CAR(%)s.

(20)

20

ROBUSTNESS CHECKS

All analysis performed in this study, are conducted using an event window of (+1,+3) days after each election result announcement in the sample period of 1990-2010. All analysis conducted in this paper are using CRSP value-weighted normal returns to calculated abnormal returns. Related analysis in the study by Goldman et al. (2009) use these model specifications as well. Goldman et al. perform robustness checks using different event windows and calculating abnormal returns using CRSP equally-weighted normal returns. Goldman et al. (2009) report no difference in results between their original model and models analyzed as robustness checks.

The multivariate panel data linear regression analysis in this model assumes fixed effects. Related analysis performed by Cooper et al. (2010) assumes fixed effects as well. Cooper et al. (2010) perform robustness checks using alternative model specifications. No differences in results are reported.

All analysis performed in this paper assign campaign contributing companies to contribution portfolios based on the relative share of their contributions to election campaigns of each major U.S. political party. In their related analysis, Goldman et al. (2010) use the same assigning criteria. Goldman et al. (2010) perform their analysis using different assigning criteria as well. Results are reported to remain similar. This paper provides a robustness check using different cut-offs assigning companies to contribution portfolios, delivering similar results.

VI. CONCLUSION

Performing multivariate panel data linear regression on the complete original dataset, covering all U.S. election cycles from 1990 until 2010, this paper indicates no statistically significant influence of U.S. corporate political campaign contributions on the stock price of those U.S. campaign contributing companies after the announcement of U.S. election results exist. This holds both when market weighted cumulative abnormal returns after election result announcements are analyzed and when industry weighted cumulative abnormal returns after election result announcements are analyzed.

Using panel data regression on the entire original dataset, this paper shows no statistically significant influence of corporate political campaign contributions on the stock prices of campaign contributing companies, following election result announcement. This could possibly imply that campaign contributions are no accurate indicator of the politically connectedness of a corporation. However, this could also imply that the politically connectedness of a company simply does not have any effect on the stock performance of that company.

Future empirical papers could perform similar research as is conducted in this paper, using other definitions of the politically connectedness a company. E.g., future papers could measure the

(21)

21

politically connectedness of a company by the amount of political connected members of its board of directors.

In concordance with earlier related research, using alternative empirical settings, this paper does shows statistically significant abnormal stock price reactions compared to the market, in the period following the announcement of U.S. political election results for S&P500 companies politically connected trough political campaign contributions. Abnormal stock price reactions are shown for panel data multi election cycle univariate regression, single election cycle univariate regression and single election cycle multivariate regression analysis.

Comparing the post-election result announcement returns of political campaign contributing companies with their specific business industry return, this paper indicates no statistically significant abnormal returns exist. Hence, this paper expands on the evidence presented by Knight (2006), by providing clear empirical evidence abnormal returns of campaign contributing companies after the announcement of election results merely are based on a industry specific, general return.

This paper shows U.S. companies, politically connected through political campaign contributions do not generate industry weighted abnormal returns after the announcement of U.S. election results. From this, this paper concludes corporate contributions to U.S. political federal election campaigns do not result in individual governmental favors supplied to the contributing corporation. This paper indicates abnormal returns over the market after election result announcements shown by previous studies to be merely the result of business industry wide expectations of the implied policy changes, caused by the political party winning the election. Due to this result, this paper believes to contribute to the strand of empirical literature opposing the mainly theoretical idea of U.S. companies making political campaign contributions in order to attain specific individual political favors.

(22)

22

REFERENCES

Ansolabehere, S., de Figueiredo, J., Snyder, J. (2003), Why is there so little money in U.S. politics? The

Journal of Economic Perspectives, 17: 105-130

Ashworth, S (2006), Campaign finance and voter welfare with entrenched incumbents. American

Political Science Review, 100: 55-68

Boubakri, N., Guedhami, O., Mishra, D., Saffar, W. (2012), Political connections and the cost of equity capital. The Journal of Corporate Finance, 18: 541-559

Bronars, S. and Lott, J. (1997), Do campaign donations alter how a politician votes? Or, do donors support candidates who value the same things that they do? Journal of Law and Economics, 40: 317:350

Brown, S. and Warner, J. (1980), Measuring security price performance. Journal of Financial

Economics, 8: 205-258

Cameron, C. and Windmeijer F. (1997), An R-squared measure of goodness of fit for some common nonlinear regression models. Journal of Econometrics, 77: 329-342

Coate, S. (2004a), Pareto-improving campaign policy. The American Economic Review, 94: 628-655

Coate, S. (2004b), Political competition with campaign contributions and informative advertising.

Journal of the European Economic Association, 2: 772-804

Cooper, M., Gulen, H., Ovtchinnikov, A. (2010), Corporate political contributions and stock returns.

The Journal of Finance, 65: 687–724

Federal Election Commission. About the FEC. fec.gov. Retrieved February 5, 2014, from http://www.fec.gov/about.shtml

Fisman, R. (2001), Estimating the value of political connections. The American Economic Review, 91: 1095-1102

(23)

23

Goldman, E., Rocholl, J., So, J. (2009), Do politicaly connected boards affect firm value?. The Review

of Financial Studies, 22: 2331-2360

Grossman, G. and Helpman, E. (1994), Protection for sale. The American Economic Review, 84: 833-850

Iowa Electronic Markets. About the IEM. Retrieved February 5, 2014, from

http://tippie.uiowa.edu/iem/about

Jayachandran, S. (2006), The Jeffords effect. Journal of Law and Economics, 49: 397-425

Kaufmann, D., Kraay, A., Mastruzzi, M. (2003), Governance matters III: Governance indicators for 1996-2002. The World Bank, working paper

Knight, B. (2006), Are policy platforms capitalized into equity prices? Evidence from the Bush/Gore 2000 Presidential Election. Journal of Public Economics, 90: 751-773

MacKinlay, A. (1997), Event studies in economics and finance. Journal of Economic Literature, 35: 13-39

Roberts, B. (1990), A dead senator tells no lies: Seniority and the distribution of federal benefits.

American Journal of Political Science, 34: 31-58

Snyder, J. (1990), Campaign contributions as investments: The House of Representatives. Journal of

Political Economy, 98: 1195-1227

Stratmann, T. (2005), Some talk: Money on politics. A (partial) review of the literature. Public

(24)

24

Table 2: CAR(%)s of different contribution portfolios

Standard deviations are stated in parentheses. *, ** and *** denote statistical significance at the 10%, 5% and 1% level

respectively.

Corrected by market return

Corrected by industry return

Republican election win

Democratic election win

Republican election win

Democratic election win

Portfolio

Value weighted CAR(%)

(+1,+3)

Value weighted CAR(%)

(+1,+3)

Value weighted CAR(%)

(+1,+3)

Value weighted CAR(%)

(+1,+3)

a) Analysis with 3 contribution portfolios

DonDem

0,07

(3,94)

0,0006

(4,79)

0,13

(3,91)

0,05

(4,47)

DonWeakRep

0,39

(4,29***)

-0,28

(4,61)

0,26

(3,91*)

-0,26

(4,35)

DonStrongRep

0,58

(3,92***)

-0,13

(4,15)

0,18

(3,68)

-0,12

(3,95)

b) Analysis with 4 contribution portfolios

DonStrongDem

0,00

(3,39)

-0,12

(5,14)

0,09

(3,43)

-0,21

(4,83)

DonWeakDem

0,15

(4,44)

0,12

(4,41)

0,18

(4,35)

0,30

(4,08)

DonWeakRep

0,39

(4,29***)

-0,28

(4,60)

0,26

(3,91*)

-0,26

(4,35)

DonStrongRep

0,58

(3,92***)

-0,13

(4,15)

0,018

(3,68)

-0,12

(3,95)

(25)

25

Table 3: Descriptive statistics and CAR(%)s of different contribution portfolios for election cycle 2004

Standard deviations are stated in parentheses. *, ** and *** denote statistical significance at the 10%, 5% and 1% level respectively.

Corrected by market

return

Corrected by industry

return

Portfolio

N

Average total

donation

(x1000$)

Average % to

Republican

Campaign

Minimum total

donation

(x1000$)

Maximum total

donation

(x1000$)

Value weighted

CAR(%) (+1,+3)

Value weighted

CAR(%) (+1,+3)

a) Analysis with 3 contribution portfolios

DonDem

99

444

28,9

0,3

6425

-0,13

(2,87)

-0,12

(2,75)

DonWeakRep

141

541

64,0

1,75

4018

0,80

(2,95***)

0,35

(2,87)

DonStrongRep

140

250

86,4

0,5

2430

1,17

(3,20***)

0,56

(3,12**)

b) Analysis with 4 contribution portfolios

DonStrongDem 50

424

28,0

2,0

3761

-0,10

(2,59)

-0,10

(2,48)

DonWeakDem

49

465

29,9

0,3

6425

-0,16

(3,15)

-0,14

(3,02)

DonWeakRep

141

541

64,0

1,75

4018

0,80

(2,95***)

0,35

(2,87)

DonStrongRep

140

250

86,4

0,5

2430

1,17

(3,20***)

0,56

(3,12**)

Referenties

GERELATEERDE DOCUMENTEN

An alarming finding from our study is that a large proportion of COVID- 19 trials test the same treatments or drugs, creating a thicket of redundant, uncoordinated, and

The temperature and degree of cure distributions inside the processing material have been calculated using the developed thermo-chemical numerical process models and subsequently

The reformulation as a Mealy Machine can be done in di fferent ways, in particular, the higher order functions present in the Haskell definitions may be executed over space or

45 Nu het EHRM in deze zaak geen schending van artikel 6 lid 1 EVRM aanneemt, terwijl de nationale rechter zich niet over de evenredigheid van de sanctie had kunnen uitlaten, kan

The final results of this research could not confirm that when an organization uses CRM and celebrity endorsement to communicate to their consumer, it will

Despite their advantages, thermoplastic composites exhibit strong time-dependent behavior in the matrix-dominated transverse and off-axis loading: strength depends on the

Table 9 shows that only the difference in average (median) return between the High Score and Value portfolio (Adjusted Low Score and value) is significantly different from

If I find evidence for the situation presented in figure 2 and the difference in announcement returns between high market- to-book cash acquirers and low market-to-book share