Corporate Electoral Donations and Tax Aggressiveness

Objective: To investigate whether companies that donate to winning electoral campaigns are more aggressive in terms of tax planning than companies that do not make these contributions. The relationship between politicians and companies may be signaled by political connections in which companies try to get political benefits in exchange for providing politicians with campaign financing. Our hypothesis is that a quid pro quo occurs in which these companies benefit from favorable tax treatment that reduces their relative tax burden. Methodology: The focus of this study is donations which were made in the presidential elections of 2010 and 2014. The sample covers the period between 2010 and 2016 for companies listed on the B3, using proxies for tax aggressiveness computed based on Value Added Reporting. Through linear regressions we have tested whether the companies that made these campaign contributions tend to have a lower tax burden. Results: The proposed hypothesis was confirmed, revealing that a political connection between campaign donations reduces the tax burden for donating companies during the years following the election. These donations appear to depict an environment characterized by an exchange of favors, in which these donating companies exhibit greater tax aggressiveness as compared to non-donating companies. Originality/Value: The current study deals with a subject that has not yet been examined empirically in Brazil, and reinforces the position adopted by the Supreme Court in prohibiting campaign donations to inhibit quid pro quo practices. The study offers additional arguments for the criminalization of the so-called “second set of books�? used to record electoral campaign contributions.


INTRODUCTION
The campaign financing literature presents evidence that receiving campaign donations is essential to conducting a successful electoral campaign. (Samuels, 2001;Speck, 2005;Figueiredo Filho, 2009;Mancuso, 2015). Until recently this has also been the case in Brazil, with the largest source of campaign contributions coming from corporate donations (Cervi, 2010).
In this sense, Camilo et al. (2012) explain that this practice is a tactic that companies employ that depends on public agents to obtain benefits, and thus political agents seek to strengthen their ties with companies with the primary goal of obtaining financing for their campaigns.
Thus, through political connections designed to collect financial resources, this practice of lobbying arises, which is directly related to the actions of private managers in relation to politicians, seeking to obtain benefits through these political agents by influencing the legislative process or collecting financing or benefits (Watts & Zimmerman, 1978).
To Silva et al. (2015), lobbying is a form of dialogue, convincing or debate designed to orient decision making regarding a given subject within public or private environments, looking for benefits within the area of this activity. The practice of lobbying is associated with groups that have economic objectives (Carmo, 2014). This fact is due to the dominant role of the State in the economy and tax planning, making lobbying one of the main instruments to make political agents listen to their constituents, influencing them to make decisions in favor of various particular interests within society (Carmo, 2014).
In this manner, political connections are widely debated in the international literature (Claessens et al 2008;Cooper et al 2010;He et al ,2014;So, 2009;Li & Zhao, 2015), even though their relationship with Tax Aggressiveness has not been encountered in the Brazilian literature, specifically relating them with campaign contributions.
Electronic copy available at: https://ssrn.com/abstract=3361766 In establishing connections through campaign financing, the company makes financial resources available believing that it will yield a positive financial return and this fact leads to serious problems, since the objectives of companies may diverge from those of society and even democratic principles, that, in theory, should guide the behavior of political agents (Camilo et al., 2012). This is why this study seeks to verify whether there is a relationship between Tax Aggressiveness and Political Connections between companies that make financial donations to political campaigns.  (Bazuchi, 2013).
This study is important in that it demonstrates that when corporations have political relationships, they exhibit more tax aggressiveness depending on this connection. This study intends to contribute to this subject given that in recent years the electoral process has been concerned with the possibility of corruption and manipulations involving companies linked to political agents. Our hypothesis is that there is a quid pro quo established between firms and politicians, in which companies benefit from favorable tax treatment after campaign donations that leads them to have a relatively lower tax burden. Companies that make these contributions should thus display greater tax aggressiveness than others, in order to receive particular tax benefits and incentives such as the REFIS Refinancing Program.
The material used in this study is data related to campaign donations obtained from the TSE Supreme Electoral Court for the period from 2010 to 2014, featuring firms listed on the Bovespa stock exchange for the presidential election.
Two metrics for tax aggressiveness were used to measure tax aggressiveness in addition to control variables to reduce any bias in the proposed regression. The data was collected from the TSE website, the Economatica system, and the B3 Stock Exchange portal.
This study is divided into five sections. After the introduction, we have section two which covers its theoretical references and a review of the literature. The following sections first describe the methodology used and then the analysis of the regression results. Finally, we present our conclusions and bibliographic references.
Political Connections according to Marcon, Camilo, and Bandeira-de-Mello (2012), may be characterized as the relationship between political agents and companies, in which both seek to further their common interests. Thus, Political Connections are a practice that companies which depend on financial resources use to attain their objectives, seeking to expand their ties with politicians, who in turn are interested in the funding of their electoral campaigns (Pinheiro et al., 2016).
During recent decades, even though democracy and the economic system have evolved, mainly due to control mechanisms and the publication of public and private acts (Bazuchi, 2013), organizations still need to diminish the uncertainties they face and increase their funding to remain competitive, and thus they often use Political Connections to strengthen their ties to public agents (Brey, 2011).
Through political ties, for example, companies can secure conditions that are more favorable to the obtaining of concessions from public institutions, compared to other private financial institutions (Bazuchi, 2013).
The advantages of political connections may be lower than their costs, but the dominant literature understands that connections can be favorable for companies, because political agents, directly or indirectly, have the power to interfere with competitiveness and competition, through the creation of limitations for competing firms (Brey, 2014;Brey et al, 2012;Izer & Alakent, 2012).
There is no definition within Brazilian legislation of what Abusive, Aggressive or Pure Tax Planning is. Thus, it is inevitable that there is doubt as to whether something can be done or not to diminish, remove or delay the incidence of taxes, with there being pressure to confront abusive tax behavior, taking into account the interests of states in collecting tax revenues (Martinez, 2017).
Analyzing political connections as a form of linking companies to public agents, the principle of resource dependence points to the relevance of associating companies with external causalities as a way to guarantee benefits for companies, such as financial resources, influence, governmental income and protection (Camilo et al, 2012). Another advantage of these political ties is the ability of governments to program their budgets, using their discretion with certain public spending to favor companies (Amore & Bennedsen, 2013). To Jackowicz et al (2014), institutional, economic and political spaces are fundamental points in the conception of political connections and the actions of companies. In practice, they detect that political connections in an organization are something onerous, and for technical reasons and in order to simplify methodologies, various authors adopt the term to designate the presence of a political agent on the board of directors or the leadership of companies (Infante & Piazza, 2014). However, this theory does not take into account possible indirect political ties that companies can possess.
In the literature there is no consensus to indicate which is the best way to identify a politically connected firm (Jackowicz et al, 2014). In the current study, taking into account the obstacles to determining political ties in Brazilian companies, they will be evaluated through official campaign donations in the years 2010 and 2014.
In the framework produced by Claessens et al (2008), it has been verified that companies that make campaign contributions obtained greater returns than other firms, and even succeeded in increasing their bank financing with public institutions. Campaign finance has received more attention in recent years, because the practices that the parties have used to maintain themselves have become small in relation to the resources obtained from companies through electoral donations (Lodoño & Zovatto, 2014). In this sense, it is important to observe that after the election, elected candidates end up having to "return" donations made by supporters through specific interests. (Horochovski et al 2016).
In a study similar to this one addressing political connections and company behavior, Based on the literature, we propose investigating the following research hypothesis: " . Companies that make electoral campaign donations tend to practice a greater degree of tax aggressiveness.
The methodology used in this investigation will next be discussed, and its overall objective is to analyze whether companies that realize electoral campaign donations are different from other firms in terms of their tax aggressiveness.

METHODOLOGY
This study uses a sample of Brazilian firms listed on the B3 stock exchange, or in other words, 529 companies that financed victorious presidential electoral campaigns. The time periods selected range from 2010 to 2016 (7 periods).
Due to the particularities of tax norms for state firms and also those of the financial sector, the number of firms was reduced. Also excluded from this analysis were all the firms that did not have published data for one of the periods analyzed, because they were not listed anymore on the B3, or they did not present compatible financial information.  Source: Prepared by the authors Figure 2 relates the metrics which were used to verify tax aggressiveness, the control variables, and also the proxies for political connections. The first metric used in this study is the VAT, which to Martinez (2015) not only covers taxes on profits, but also specific taxes on revenues, which represent most of the tax burden of a company.
The second metric applied in this work to measure tax aggressiveness is Cash_VAT, which is used to calculate, based on the variation in taxes paid on the Balance Sheet, the sum of taxes effectively paid by companies during a given year.
The Donation variable, is a dummy variable assumed to have a value of 1 when the company made a donation to a winning candidate or coalition. For control purposes, we examine just those who donated to the winning presidential candidate   (2015), Size can interfere with the degree of tax aggressiveness. To measure a company's performance, the Delta ROA was used (Mathieu & Zajac, 1990). And the last variable, according to Gong and Yuching (2011), which can directly influence tax aggressiveness is Leverage.
In terms of measuring the strength of Political Connections of the companies that will be analyzed, we use the studies of Claessens, Feijen and Laeven (2008)  After the data is collected the results can be interpreted, but first descriptive statistics need to be presented for the variables used in this analysis.  Table 1 it is possible to see that all of the tax aggressiveness proxies presented positive averages, and they were: 34.6% for VAT and 48.7% for Cash VAT. Just 1.8% of the company-years donated to the winning party or candidate in the following election. Table 2 exhibits the correlation coefficients for the study variables.  Table 2 presents the correlation coefficients between the variables used in this study.
It's possible to observe significant and positive correlations between leverage and the Cash_VAT proxies for aggressiveness, which suggests that the greater the leverage, the less a company is predisposed to practice tax aggressiveness, remembering that the greater the metrics for tax aggressiveness, the lesser the tax aggressiveness will be. The correlations of the Size variable were significant for all the variables used in this study (except ΔROA). It should be noted that the correlation between Size and the Cash-VAT metric is negative, therefore the larger the company, the smaller Cash_VAT will be, indicating greater tax aggressiveness by the company. The correlation between the Donation dummy and Size was significant and positive, indicating that the companies that donate to parties and/or candidates have greater total assets than those that do not.
This section displays the analyses of the relationship between tax aggressiveness and donations made to the elected presidential parties and candidates.

Tax Aggressiveness -Continuous Variables
The table below presents the panel regression with fixed effects for companies, verifying the relationship between donations and tax aggressiveness as measured by the proxies VAT and Cash_VAT.  Robust Standard Error in parentheses *** p<0.01, ** p<0.05, * p<0.1 Analyzing the results of the table above, it can be affirmed that companies that made donations had lower average VAT and Cash VAT values than companies that did not make donations, that is donating firms exhibited more tax aggressiveness on average when compared with companies that did not make donations, thus verifying the study's hypothesis.
The Size control variable was significant at 1% with a negative coefficient, which was consistent with the results found in Table 2. In addition, it's important to emphasize that the constants of all four models were positive and significant, which shows the consistency of the results. It should be mentioned that we opted for a fixed effects model presentation, given the results of the Hausmann test.

Tax Aggressiveness -Discrete variables
This subsection presents panel data fixed effects logistic regressions for companies, which verified the relationship between donations and tax aggressiveness measured through qualitative proxies. Table 4 exhibits panel data fixed effects logistic regressions for companies, verifying the relationship between donations and tax aggressiveness measured by qualitative proxies that assume a value of 1 for high tax aggressiveness (values for VAT and Cash VAT below the first quartile) and 0 otherwise (values for VAT and Cash VAT above the first quartile). Note: Standard error between parentheses. *** p<0.01, ** p<0.05, * p<0.1 -The variable HTA_VAT (high tax aggressiveness measured by VAT) is a dummy which assumes a value of 1 for observations below the first quartile for VAT, and 0 for observations above the first quartile of VAT; HTA_Cash_VAT (high tax aggressiveness measured by Cash VAT) is a dummy which assumes a value of 1 for observations below the first quartile of Cash VAT, and 0 for observations above the first quartile of Cash VAT; Donation is a dummy that assumes a value of 1 when the company donated to the elected candidate or party and 0 otherwise; Size is the logarithm of the total assets, with 1% Winsorized on both sides; Δ ROA is the difference between ROA at time t and ROA at time t-1, with 1% Winsorized on both sides; Leverage is 1% Winsorized on both sides.
Analyzing the model, it is possible to affirm that companies that made donations have a greater probability of belonging to the group of companies which have high tax aggressiveness (the 20% which are most aggressive) compared to companies that did not make these donations, thus donating firms exhibit greater tax aggressiveness, which confirms the study hypothesis. It should be noted that the model in attributing a value of 1 for HTA_VAT and HTA_Cash_VAT in the observations of the more aggressive companies, and 0 for the others, identifies those qualities that define companies as more aggressive in terms of taxes.  (1) Note: The variable LTA_VAT (low tax aggressiveness measured by VAT) is a dummy that assumes a value of 1 for observations above the third quartile of VAT and 0 for observations below the third quartile of VAT; LTA_Cash_VAT (low tax aggressiveness measured by Cash_VAT) is a dummy which assumes a value of 1 for the observations above the third quartile of the Cash VAT and 0 for the observations below the third quartile of the Cash_VAT; Donation is a dummy that assumes a value of 1 when the company/year donated to the elected candidate or party, and 0 otherwise.
Based on the three models that converge, it is possible to affirm that companies that made donations have a lower probability of belonging to the group of the 20% of companies with the least tax aggressiveness compared to companies that did not make donations, that is donating firms had greater tax aggressiveness, verifying the research hypothesis. It should be added that even though they are not documented in the tables, additional tests were performed to assure the robustness of the statistics, among which should be highlighted: i. the Jarque-Bera (JB) normality test, indicating that the residuals have a normal distribution; ii. The Factor Inflation Variance (FIV) test which presented elevated values close to 4,000, which however were below the limits that would be characterized as serious problems of multi-collinearity and iii. the Breusch-Godfrey (BG) test, which found that there is no autocorrelation among the residuals.

Conclusion
This study was designed to study whether the companies listed on the B3  This information constitutes a significant finding due its novel nature within the Brazilian context with there being no similar studies up until now, as well as the fact that the results are interesting for investors and regulatory agents, as well as the community in general, because they help in analyzing and forecasting a company's tax conduct. This study contributes by emphasizing the relationship between corporate electoral donations and tax aggressiveness, providing information that can help identify whether a company's tax planning is conducted based on its political connections.
It should be emphasized that there appears to be the implication that donations are promoted in an environment characterized by an exchange of favors and potential corruption.
This reinforces the position adopted by the Supreme Court in prohibiting corporate campaign donations to inhibit quid pro quo practices. This study offers, even in an indirect fashion, an additional argument to support the criminalization of the so-called "second set of books" for electoral campaigns. If companies already display tax aggressiveness within a context of transparency in terms of donations, it is logical to presume that in an environment characterized by a lack of transparency, donations will be made with few democratic objectives in mind, and will rather be designed to obtain various advantages including tax advantages, without revealing to third parties that there is a possible conflict of interest between donating companies and the benefits offered by public agents.
This being so, the modification of the federal legislation of Law nº 9.504/1997, known as the Election Law through the electoral reform of Law nº 13.165/2015, to prohibit corporate donations in electoral campaigns is very opportune. In this sense, this study supports the urgency of criminalizing clandestine campaign donations to avoid potential quid pro quo practices between public agents and companies.
In terms of the limitations of this work, we have the impossibility of obtaining data for campaign contributions of companies that are not listed on the B3. Thus, this study is limited to openly traded companies, thus excluding companies outside of this capital market. Finally, areas for future study could include segregating the analysis by election and noting whether reelections condition these results. Studies could also be made investigating the tax aggressiveness of companies that donate to various candidates in relation to those who just donate to the winning candidates, or even those who donate consistently to the same candidate in every election.