EFFECT OF TAX AVOIDANCE ON SHAREHOLDER VALUE
Case study: Panama Papers and Paradise Papers
Name:
Alexander Kool
Student Number:
10809848
E-mail:
Alexander_kool@live.nl
Study:
Economics and Business
Track:
Finance and Organisation
Thesis coordinator:
Ieva Sakalauskaite
Statement of Originality
This document is written by Student Alexander Kool who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents
List of abbreviations ... 4
1. Introduction ... 5
2. Literature Review... 7
2.1 Background theory ... 7
2.1.1 Tax avoidance vs tax evasion ... 7
2.1.2 Panama Papers and Paradise Papers ... 7
2.2 Related literature ... 8
3. Methodology ... 10
3.1 Method ... 10
3.2 Model and Hypothesis ... 11
3.3 Data ... 13
3.4 Results ... 13
3.4.1 Panama Papers ... 13
3.4.2 Paradise Papers ... 14
3.5 Discussion ... 16
4. Conclusion ... 17
4.1 Summary and conclusion ... 17
4.2 Recommendations for further research ... 17
Reference list ... 18
Literature ... 18
Newspaper Articles ... 19
Appendix ... 20
List of abbreviations
Abbreviation
Definition
CAR Cumulative Abnormal Return
ETR Effective Tax Rate
FTSE Financial Times Stock Exchange
ICIJ International Consortium of Investigative Journalists
OECD Organisation for Economic Coordination and Development
OLS Ordinary Least Squares
1. Introduction
On the 18th of December 2017, furniture and home accessories brand Ikea made media headlines.
The headlines were not made because a new type of furniture was invented. It was made because Ikea was being investigated by the European Commission for tax avoidance. The organization was being accused of not paying its fair share of taxes (FD, 2017).
Ikea is not the only firm to publicly be accused of avoiding taxes. APRIL, a large pulp-and-paper manufacturing company was able to cut down large parts of the Indonesian tropical rainforests, due to its ability to avoid paying taxes (Alecci, 2017). This firm had help from large banks as Credit Suisse and ABN Amro as well as law-firms Appleby and Estera, who provided corporate structures to facilitate the shifting of income from high tax regions to low tax regions.
According to Freedman (1970) tax avoidance is simply the shifting of wealth from the state to the shareholders. Therefore, tax avoidance is in the best interest of the owners of the firm. It is possible however, that with increasing public awareness, tax avoidance is not a simple transfer of wealth anymore. An organization accused of avoiding taxes, might be confronted with negative publicity. This publicity, in turn could affect sales and eventually profitability. Another problem of tax avoidance arises when firms get caught. An example is technology firm Apple, who has been
sentenced by the European Commission to the payment of €13bln in back taxes (Farrel, & McDonald, 2016).
Public awareness increased when tax avoidance made headlines in April 2016, due to leaking of the Panama Papers. This publication exposed the tax avoidance strategies of 214,488 firms
worldwide. In November 2017, the publication of the Paradise Papers exposed the avoidance strategies of another 160,000 entities. The publications present an opportunity to investigate how shareholders respond to widespread news about tax avoidance.
It is publicly seen as unfair to avoid taxes (OECD, 2013, p. 9). Therefore, the costs of tax avoidance could outweigh the benefits. To research this, the following proposition is made:
To what extent does an accusation of tax avoidance influence shareholder value in the Dutch market?
Although tax avoidance itself might be positive for shareholders, the accusation of tax avoidance could have negative implications for firm value, due to the associated costs. The costs might in turn, outweigh the benefits. The value of an organization pursuing an aggressive tax strategy, i.e.
attempting to lower their effective tax rate (ETR) by avoiding taxes, could therefore be negatively affected. This assumption can be researched using the reaction of stock prices on the time of publication of the Panama Papers and the Paradise Papers.
To quantitatively research how tax avoidance affects firm value, an event study is performed for both publications. The hypothesis of this paper is:
𝐻0: 𝐴𝑛 𝑎𝑐𝑐𝑢𝑠𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡𝑎𝑥 𝑎𝑣𝑜𝑖𝑑𝑎𝑛𝑐𝑒 𝑑𝑜𝑒𝑠 𝑛𝑜𝑡 𝑖𝑛𝑓𝑙𝑢𝑒𝑛𝑐𝑒 𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟 𝑣𝑎𝑙𝑢𝑒
𝐻1: 𝐴𝑛 𝑎𝑐𝑐𝑢𝑠𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡𝑎𝑥 𝑎𝑣𝑜𝑖𝑑𝑎𝑛𝑐𝑒 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒𝑙𝑦 𝑖𝑛𝑓𝑙𝑢𝑒𝑛𝑐𝑒𝑠 𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟 𝑣𝑎𝑙𝑢𝑒
This hypothesis is researched using an empirical model. The cumulative abnormal return (CAR) is used as a proxy to measure the reaction of a firm’s stock price. The effective tax rate is an indication of tax aggressiveness. A low ETR indicates the possibility a firm is avoiding taxes. Additional variables are added to control for affects beyond the relationship between the cumulative abnormal return and a firm’s effective tax rate. These variables are institutional ownership, firm size, deferred tax assets, tax shield and industry.
Not all Dutch firms are however mentioned in the Panama Papers or the Paradise Papers. In fact, only seven of the twenty-five largest Dutch firms, making up around 33 percent of the AEX index are mentioned. As the accusations are large-scale, organizations not mentioned in the leaks can still be affected. Firms with a low ETR might still react negatively, due to shareholders anticipating negative consequences of being part of the next publication. Therefore, all Dutch firms are included in the sample.
The remainder of the paper is structured as follows: Section 2 will display related literature on the effect of tax avoidance on firm value. The empirical analysis is performed in section 3. Here the model is introduced and the regression analysis is performed. The results are subsequently analysed. Finally, this paper will be concluded in section 4.
2. Literature Review
2.1 Background theory
2.1.1 Tax avoidance vs tax evasion
There is a subtle difference between tax avoidance and tax evasion. Organizations avoiding taxes are within their legal rights to do so, whereas firms that evade taxes are not. Firms caught evading taxes are subject to legal penalties. The term avoidance is defined by the OECD as the arrangement of a taxpayer’s taxes to reduce their tax liability, which is in contradiction with the intent of the law. Although an entity is within the boundaries of the law, and thus acts legally, it is possible to take actions which were not intended by the authorities. Evasion on the other hand, is defined as any illegal arrangement where the liability to tax is either hidden or ignored. Not only hiding income, but also hiding or ignoring information that could lead to taxation is considered evasion. It is important for organizations to report all ways in which they avoid paying taxes, as withholding this information can result in their actions becoming illegal. This paper solely focusses on the accusation of tax avoidance. A good indicator is a firm’s ETR, which is transparent information and therefore easy to collect. Tax evasion is illegal and information is therefore often not publicly available.
2.1.2 Panama Papers and Paradise Papers
The Panama Papers and the Paradise Papers were both published by the International Consortium of Investigative Journalists (ICIJ). These two leaks are the largest public accusations of tax avoidance ever recorded. The documents known as the Panama Papers, were leaked by an anonymous source to the German newspaper Suddeutsche Zeitung on April 3rd 2016. The Paradise Papers were
subsequently leaked to the Suddeutsche Zeitung by an anonymous source on November 5th 2017. The
documents reveal how big banks and law firms provide financial anonymity to wealthy individuals and organizations through hard to trace firms in offshore tax havens. Through these structures, tax can be legally avoided.
The Panama Papers consist of 11.5 million documents, revealing information on
214,488 offshore entities in over 200 countries. The data revealed the secret owners of bank accounts and firms in 21 offshore jurisdictions, including the Bahama’s, British Virgin Islands and Panama. Central in the publication was the Panamanian law firm Mossack Fonseca. This law firm creates shell companies and corporate structures, through which wealth is transferred to tax shelters.
The publication of the Panama Papers made global headlines. Politicians and public officials who had been mentioned were subjected to public scrutiny. The tax authorities of all OECD countries have announced that they will be investigating the publication. Although corporate structures are legal, many tax authorities require firms to communicate how wealth is earned and shifted within the corporation. Organizations have possibly failed to do this. The Panama Papers namely reveal how the
Netherlands is used as a hub, to pass through wealth from foreign destinations to offshore tax havens. The Netherlands is of interest to foreign investors due to the stable democracy, well defined property rights and the possibility of corporate structures (Kleinnijenhuis, 2016).
The publication of the Paradise Papers consists of 13.4 million documents, revealing information on an additional 160,000 offshore entities. The data revealed the owners of offshore entities in 19 jurisdictions, including Bermuda and Guernsey. The publication of the Paradise Papers had bigger implications for the Netherlands. The leak revealed that large corporations as Nike, Tesla and Uber used the Netherlands as a hub to shift profits to offshore locations. Furthermore, seven AEX listed firms, including AEGON, NN and Shell were mentioned. ABN AMRO helped facilitate
corporate structures that helped APRIL devastate the tropical rainforest in Indonesia. The Dutch tax authorities also came under scrutiny. Tax deals made between the Dutch tax authorities and
multinationals might not have been legitimate. Consumer goods corporation Procter and Gamble was for example able to cut its Dutch tax bill by €169 million due to agreements with the tax authorities. In response to this, the Dutch government ordered 4000 rulings to be investigated.
2.2 Related literature
Many studies have researched how tax avoidance influences shareholder value. The traditional view states that tax avoidance is beneficial to shareholders as it entails transferring wealth from the state to shareholders (Freedman, 1970). The market views firms with a low ETR as more cost efficient (Swenson, 1999). Recent literature indicate however that tax avoidance is not costless. There are both financial- and non-financial costs associated with tax planning. This can affect the favourability for shareholders as costs may outweigh the benefits.
Financial costs consist of implementation costs and the chance of incurring legal penalties. Implementation costs range from time and effort to costs of setting up corporate structures (Chen et al., 2010, p. 42). Legal penalties are incurred when an organization are not transparent enough about the way they shift wealth to low tax jurisdictions. Non-financial costs include agency costs and reputational damage. Agency costs occur when executives perform in a way that is harmful for shareholders. Dyreng, Hanlon and Maydew (2010, pp. 1185-1187) find evidence that individual executives have a big influence on the level of tax avoidance within a firm. This is due to personal characteristics. Moreover, it is found that the alignment of managerial compensation influences tax aggressiveness. When the incentives of executives align with the incentives of the shareholders, managers are less prone to shelter income (Desai, & Dharmapala, 2006, p. 166). Finally, the attitude of an organization regarding corporate social responsibility (CSR) is related to the level of corporate tax avoidance (Hoi, Wu, & Zhang, 2013, p. 2025). Firms with irresponsible CSR activities are more prone to engage in tax planning. It seems then that well-governed firms are less likely to engage in tax avoiding activities. This is confirmed by Desai and Dharmapala (2009, p. 542).
Reputational costs also influence the level of tax avoidance. Chen et al. (2010, p. 42, p. 60) show that family owned firms, where reputation is perceived to be very important, are less likely to engage in tax planning. Not only tax avoidance itself, but also the mere accusation is harmful for these businesses. Family owned firms are willing to forgo tax benefits in order to reduce the risk of a tax related lawsuit. Dyreng et al. (2016, p. 179) finds some evidence to infer that high tax-related media coverage increases a firm’s ETR in a sample of FTSE 100 companies. This difference is however not statistically significant. Choy et al. (2017, p. 216) shows that accused FTSE 100 firms decrease in value after an accusation. Hanlon and Slemrod (2009, p. 139) also find a decrease in stock price for US firms after an accusation of tax avoidance. Not only actual tax avoidance, but news relating to possible tax planning is thus negatively received by the market. Lastly, Austin and Wilson (2017, p. 89) conclude that organizations with valuable consumer brands are less likely to engage in tax planning. Especially when a firm is easily identified with a brand. They question however if organizations actually incur any reputational costs due to tax avoidance. There is currently no existing literature quantitatively relating tax avoidance to reputational damage.
Firms that avoid taxes are often large. Domestic firms are unable to avoid taxes, as there is no alternative tax jurisdiction to which wealth can be shifted. There is thus a relationship between firm size and tax avoidance. This proposition is confirmed by Rego (2003) who finds that multinationals with more extensive foreign operations, report lower worldwide ETRs. Multinationals have more incentives and resources to pursue an aggressive tax strategy. However, Morck and Yeung (1991) do not find empirical evidence that firm value increases due to avoidance possibilities created by
extensive foreign operations. As the likelihood of tax avoidance increases with firm size, it is possible the opposite is true.
Literature has been published on the effects of the publication of the Panama Papers. The value of organizations mentioned in the Panama Papers decreased by 0.7% relative to firms that were not mentioned (O’Donovan, Wagner, & Zeume, 2017, p. 3, p. 28). Firms subsequently respond to an accusation by becoming less tax aggressive. Bauckloh et al. (2017, p. 21) perform an event study after the publication of the Panama Papers, to capture the effect of a state aid investigation announcement. They find that the stock price of the scrutinized firms decreased. The competitive tax advantages of rulings are shown to diminish after an allegation. Finally, little literature has yet been published concerning the publication of the Paradise Papers. The effect on shareholder value is yet to be revealed.
3. Methodology
3.1 Method
This paper examines the effect of an accusation of tax avoidance on firm value. This is done by analysing the reaction of stock prices to the leaking of the Panama Papers and Paradise Papers. Although a firm is accused of tax avoidance, this does not necessarily mean it actually avoids taxes. An accusation is enough to affect shareholder value (Hanlon, & Slemrod, 2009, p. 139). The setting for this research is the Netherlands, as no related literature has yet covered this market. Moreover, the Netherlands is accused of being a tax haven (Berkhout, 2016, p. 4). Do large-scale accusations of tax avoidance affect the domestic market in a country which facilitates tax avoidance differently?
To evaluate how the reaction of shareholders is influenced by the leaks, an event study is performed for both the publication of the Panama Papers and Paradise Papers, respectively. Efficient markets incorporate all available information in the pricing of assets. By determining the change of a stock price over a small period, the effects of the publication can be isolated.
An event study investigates if the return of a stock performs extraordinarily around the time of an event (Fisher, & Roll, 1969, p. 3). To determine if a stock performs differently, the abnormal return is calculated by subtracting the expected return from the realized return using the formula:
𝐴𝑅𝑖,𝑡= 𝑅𝑖,𝑡− 𝐸𝑅𝑖,𝑡
To capture the entire abnormality, the abnormal return is calculated over a small period. This is done to include any delayed effects. The timeframe should be long enough to capture delays and short enough to isolate the effect of the leaks. This paper assumes the day of publication and the subsequent three days to be sufficient. The expected return is calculated using the average return over a longer period. To avoid incorporating effects of the Panama Papers into the expected returns for the Paradise Papers, a period of 9 months (197 observations) is chosen. To treat both events equally, a nine-month period (197 observations) is also used to calculate the expected return for the Panama Papers sample. The CAR is calculated using the following formula:
∑ 𝐴𝑅𝑖,𝑡=0+. . +𝐴𝑅𝑖,𝑡=3
𝑛
𝑖=1
Following the calculation of the CAR for all firms in the sample, a test is performed. This t-test determines if the returns are statistically significantly abnormal. The null hypothesis assumes that the return is normal. The alternative assumes the return is abnormal:
𝐻0: 𝐶𝐴𝑅𝑖 = 0 𝐻1: 𝐶𝐴𝑅𝑖 ≠ 0
To test how an accusation of tax avoidance is related to shareholder value, an OLS regression is performed. A scatterplot is drawn to determine if the relationship is linear (see appendix I). From the scatterplot, it is seen that the residuals for both samples are heteroskedastic. Therefore, robust standard errors are used. A regression is performed for both the Panama Papers and Paradise Papers.
The sample includes all Dutch firms. Not all companies in the sample are however mentioned in the leaks. It is possible that firms with low ETRs are not mentioned, but are affected by the large-scale accusations of tax avoidance. Shareholders might fear for the probability that the firm is included in a next publication. Thus, to determine the reaction of the entire market, all firms are included.
3.2 Model and Hypothesis
Tax avoidance might not only be the shifting of wealth from the state to shareholders. As both financial and non-financial costs of tax planning can outweigh the benefits due to accusations, it may not be in the best interest for shareholders to pursue an aggressive tax strategy. After the publication of the Panama Papers and the Paradise Papers, public awareness has increased. Firms with a low ETR can be seen as actively avoiding taxes and incur reputational damage because of subsequent scrutiny. The following hypothesis is formulated:
𝐻0: 𝐴𝑛 𝑎𝑐𝑐𝑢𝑠𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡𝑎𝑥 𝑎𝑣𝑜𝑖𝑑𝑎𝑛𝑐𝑒 𝑑𝑜𝑒𝑠 𝑛𝑜𝑡 𝑖𝑛𝑓𝑙𝑢𝑒𝑛𝑐𝑒 𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟 𝑣𝑎𝑙𝑢𝑒 𝐻1: 𝐴𝑛 𝑎𝑐𝑐𝑢𝑠𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡𝑎𝑥 𝑎𝑣𝑜𝑖𝑑𝑎𝑛𝑐𝑒 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒𝑙𝑦 𝑖𝑛𝑓𝑙𝑢𝑒𝑛𝑐𝑒𝑠 𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟 𝑣𝑎𝑙𝑢𝑒 To evaluate this hypothesis, an OLS regression is performed. The empirical model used to test this proposition is given by the equation:
𝐶𝐴𝑅𝑖 = 𝛽0+ 𝛽1𝐸𝑇𝑅 + 𝛽2𝐼𝑂 + 𝛽3𝑆𝑖𝑧𝑒 + 𝛽4𝐷𝑇𝐴 + 𝛽5𝑇𝑆 + 𝛽6𝐼𝑛𝑑 + 𝜀𝑖
Shareholder value is measured using the CAR as dependent variable. The effective tax rate is chosen as a proxy for tax avoidance following the paper of Hanlon and Slemrod (2009). They dislike the alternative, the book-tax gap, used by Desai and Dharmapala (2009), because it can be influenced by many factors. The ETR can reflect how successful a firm is at avoiding taxes. A highly tax aggressive firm will subsequently have a low ETR, making this variable a good proxy. The hypothesis tested states:
𝐻0: 𝛽1= 0
𝐻1: 𝛽1> 0
A positive ß1 infers a negative influence on CAR. This is because reducing ETRi implies a negative
change (-∆) to the variable. Although the ETR is the main variable of interest, there are more factors that influence CAR. To control for these influences, additional variables are added. The descriptive statistics of all variables can be found in table I.
Firstly, the variable institutional ownership (IO) is added. When institutional owners are present, agency costs are expected to be lower, because owners will be more inclined to monitor managers. Desai and Dharmapala (2009, 539, p. 542) subsequently find that well-governed firm engage in less tax planning. A well-governed firm is thus expected to have a higher ETR. This in turn should have a positive effect on CAR.
Rego (2003) finds a relationship between foreign operations and tax planning, resulting in lower reported ETRs. Bigger firms have more incentives and opportunities to engage in tax planning. Morck and Yeung (1991) test to find a positive relationship between avoidance opportunities and firm value. The results they find are not statistically significant. It is possible then, that size negatively affects CAR, as the probability of multinationals avoiding taxes is greater than for local firms. Therefore, the variable Sizei is added to the regression.
Due to accounting techniques, an organization can legally lower or higher its ETR without actually attempting to avoid paying taxes. As deferred tax assets lower a firm’s ETR in the subsequent year, the variable DTAi (measured by the deferred tax assets of the firm) is added to control for this
effect. A high DTA is an indicator that the ETR of an organization can be expected to be lower. This variable is expected to have a positive influence on CAR, as it signals that a firm is cost-efficient, not tax aggressive. Tax deferred to later years result in higher profits for the current book year.
In the Netherlands, it is possible to deduct debt from income before it is taxed. A highly levered firm will have an advantage compared to equity-financed firms as its taxable income is deduced. This is known as a tax shield. It is possible then, that equity-financed firms will be more inclined to engage in tax planning as a tax shield is not available. As a result, the ETR of equity financed firms will be lower than highly leveraged firms. The variable tax shield (TS), measured by the ratio between long term debt over total capital, is therefore added. The smaller this ratio is, the more inclined a firm will be to avoid taxes, negatively influencing CAR.
Finally, the industry wherein an organization operates, can affect a firm’s attitude toward tax avoidance. Austin and Wilson (2017, p. 89) find that firms with valuable consumer brands are more aware of possible reputational damage. Organizations in industries close to consumers, will thus be
Table I – Descriptive statistics
Panama Papers Observations Mean Std. Dev. Min Max CAR 86 -0.007244 0.02054 -0.05945 0.04088
Effective Tax Rate 86 0.2119 0.07372 0.01 0.37
Institutional Ownership 86 0.05814 0.2354 0 1
Size 86 3.68e+07 5.74e+07 213,579 2.34e+08
Deferred Tax Asset 86 1,627,989 3,740,815 0 1.44e+07
Tax Shield 86 35.8957 18.3479 0 78.15
Industry 86 88.6279 36.8890 33 161
Paradise Papers Observations Mean Std. Dev. Min Max CAR 89 -0.009499 0.03162 -0.08491 0.07940
Effective Tax Rate 89 0.2187 0.07249 0.01 0.37
Institutional Ownership 89 0.05618 0.2316 0 1
Size 89 3.56e+07 5.22e+07 213,579 2.34e+08
Deferred Tax Asset 89 1,477,127 3,415,916 0 1.44e+07
Tax Shield 89 35.8754 17.3575 0 78.15
Industry 89 90.0674 36.2748 33 161
The variables are as defined in the text. The regression uses the natural logarithm of the variables Size and deferred tax asset. The observations include all Dutch firms, not only those mentioned in the respective papers.
less likely to engage in tax planning and show a higher ETR. As with firm size, industry is also connected to the extensiveness of foreign operations. Highly internationally oriented industries are expected to pursue an aggressive tax strategy, resulting in a lower ETR. This is due to the possibilities of shifting wealth to a different jurisdiction.
3.3 Data
The data was collected from the DatasStream database. Only Dutch firms were included in the sample. A total sample size of 1827 observations was primarily collected. All stocks that did not have changing returns at least once every five days were dropped from the sample. This was done to avoid including too many zero observations, which could bias the sample. Furthermore, all observations with incomplete data were excluded. Finally, two outliers were dropped from the sample. This was done because these observations influenced the regression dramatically. These two observations were also incidental and did not occur in previous years. A regression was run with a sample size of 86 observations for the event study of the Panama Papers and 89 for the Paradise Papers.
3.4 Results
3.4.1 Panama Papers
To determine if the CAR is statistically significantly abnormal, a t-test is performed. The results for the publication of the panama papers are shown table II. The CAR is negative and significantly different from zero. This indicates that shareholder value diminished following the publication of the Panama Papers. Tax avoidance accusations are thus frowned upon by investors.
To estimate how ETR effects CAR, an OLS regression analysis is performed. Before the results are estimated, a scatterplot is drawn. The scatterplot firstly indicates heteroscedasticity (see appendix I). Also, it is seen that the Panama Papers sample is best approximated using a quadratic model. The results of the regression for the Panama Papers can be found in table III.
The first regression (I) is run using only the independent variable ETR. The coefficient ß1 is
negative and significant at the one percent level with a p-value of 0.006. This result contradicts the hypothesis that shareholder value is negatively affected by an accusation of tax avoidance. The R2 is
low, indicating that the model does not fit the data very well. Therefore, additional variables are added. This is done in regression II. The coefficient ß1 remains negative and significant at the one
percent level. Tax avoidance itself is positive for shareholders, as wealth is added to the firm. This proposition is reflected in the results. As ETR decreases, CAR increases. The costs of tax avoidance
Table II –two-sided t-test of the cumulative abnormal returns
CAR Obs. Mean Std. Err. T statistic df P-value [95% Conf. Interval]
Panama Papers 86 -0.00724 0.00222 -3.2707 85 0.0016 -0.0117 -0.00284
Note: The CAR represents the cumulative abnormal return. This measure was calculated using the stock prices of all observations in the sample. The t-test is two-sided. The degrees of freedom are measured by n-1. The rejection region at the five percent level is 1.960 and at the one percent level is 2.576.
accusations might not be fully reflected in the regression yet. Therefore, an industry dummy is included. Operating in global industries can affect the opportunities and incentives to avoid taxes. This can increase agency costs and reputational damage. The results are shown in regression III.
Table III – Effect of tax avoidance on shareholder value: OLS results Panama Papers Dependent Variable (I) Robust std. errors CARi (II) Robust std. errors CARi (III) Robust std. errors CARi (IV) Robust std. errors CARi Constant 0.0077 (0.0066) -0.0838*** (0.0183) 0.2673 (0.2955) 0.1575 (0.3037)
Effective tax rate -0.0703*** (0.0294) -0.1285*** (0.0272) 0.0301 (0.1343) 0.4721*** (0.1707) Institutional ownership 0.0140** (0.0063) -0.0067 (0.0244) 0.0189 (0.0252) Ln (Size) 0.0076*** (0.0018) -0.0124 (0.0103) -0.0122 (0.0096)
Ln (Deferred tax debit) -0.0014
(0.0015) -0.0093 (0.0138) -0.00054 (0.0144) Tax Shield -0.00013 (0.00014) 0.0015 (0.0017) 0.00007 (0.00177)
Effective tax rate ^2 -1.2849***
(0.3498)
Industry NO NO YES YES
R2 0.0636 0.2854 0.5466 0.5530
Observations 86 85 85 85
Note: The dependent variable CAR is the cumulative abnormal return as defined in section 3.2. The sample is taken during an event study, using Dutch firms. The variables institutional ownership is binary, with a value of either 0 or 1. The industry variable sorts the sample into 20 different categories, the numerical values are ordinal. The squared value of the effective tax rate is included due to model misspecification. The parentheses *, ** and *** denote the significance levels 10%, 5% and 1%, respectively.
The results indicate that ETR is positively related to CAR. With a p-value of 0.824 the regression is highly insignificant however. This is due to model misspecification. When adding the quadratic term ETR2 to the model in regression IV, the results are improved. The model is non-linear as most observations are negative between the 25th and 75th quantile. The coefficient ß
1 is positive and
significant at the one percent level with a p-value of 0.008. Interestingly, the squared coefficient of ETR, ß6 is largely negative and highly significant with a p-value of 0.001. Due to this large effect of
the quadratic term, a change in the firm’s ETR with -1 means an increase of CAR with 0.8128! This result contradicts the theory that shareholder value is negatively influenced by the accusation of tax avoidance. A logical reason why tax avoidance is positive for shareholder value, is because the costs of reputational damage are not severe enough. As none of the AEX indexed firms are mentioned, this result is not surprising. The Panama Papers did not have an average negative influence on firms in the Netherlands.
3.4.2 Paradise Papers
To test for abnormality, a t-test is also performed for the Paradise Papers sample. The results are shown in table IV. The CAR is also here, negative and highly significant. This indicates that
following the publication of the Paradise Papers, firm value decreased. This is in line with the expectation that public awareness diminishes the positive effect of tax avoidance.
To determine the effect of ETR on the decrease of CAR, an OLS regression is also run for the Paradise Papers. From the scatterplot, it is visible that the sample is linear and heteroskedastic (see appendix I). The results are shown in table V. The first regression (I) is run using only the
independent variable ETR. The coefficient ß1 is positive, but insignificant. The R2 is close to zero,
indicating a very poor fit of the model. Additional variables are added in regression II. ETR remains insignificant. Including the effect of industry increases the R2 to 0.8099 indicating a good fit of the
model. With a p-value of 0.001 ETR becomes significant at the one percent level. The coefficient ß1 is
positive, indicating a negative effect on CAR. As the costs of tax avoidance accusations increase after the publication of the Paradise Papers, the benefit of tax avoidance is diminished. This result is in line with the hypothesis.
Table V – Effect of tax avoidance on shareholder value: OLS results Paradise Papers Dependent Variable (I) Robust std. errors CARi (II) Robust std. errors CARi (III) Robust std. errors CARi Constant -0.0142 (0.0097) -0.0380 (0.0278) 0.3186*** (0.1187)
Effective Tax Rate 0.0213
(0.0407) -0.0467 (0.0372) 0.2162*** (0.0650) Institutional Ownership 0.0318*** (0.0109) 0.0167 (0.0113) Ln (Size) 0.0047* (0.0027) -0.0241*** (0.0065)
Ln (Deferred tax debit) -0.0048**
(0.0021) 0.0011 (0.0041) Tax Shield 0.00056*** (0.00015) 0.00041 (0.00052) Industry NO NO YES R2 0.0024 0.1713 0.8099 Observations 89 88 88
Note: The dependent variable CAR is the cumulative abnormal return as defined in section 3.2. The sample is taken during an event study, using Dutch firms. The variables institutional ownership is binary, with a value of either 0 or 1. The industry variable sorts the sample into 20 different categories, the numerical values are ordinal. The parentheses *, ** and *** denote the significance levels 10%, 5% and 1%, respectively.
The results of the Panama Papers and Paradise Papers are not similar. The Panama Papers show that tax avoidance can have a positive influence on shareholder value. A one percent decrease in the ETR, increase CAR by 0.81 percent. The Paradise Papers show the opposite. A decrease in the ETR by one percent decreases CAR by 0.22 percent. The number of Dutch firms mentioned in the Paradise Papers is much larger than in the Panama Papers. Therefore, it is to be expected that the
Table IV –two-sided t-test of the cumulative abnormal returns
CAR Obs. Mean Std. Err. T statistic df P-value [95% Conf. Interval]
Paradise Papers 89 -0.00950 0.00335 -2.8339 88 0.0057 -0.0162 -0.00284
Note: The CAR represents the cumulative abnormal return. This measure was calculated using the stock prices of all observations in the sample. The t-test is two-sided. The degrees of freedom are measured by n-1. The rejection region at the five percent level is 1.960 and at the one percent level is 2.576.
Dutch market is more severely influenced by the publication of the Paradise Papers. These results also show, that organizations in markets which were not as severely mentioned in a publication still benefit more from pursuing an aggressive tax strategy.
Finally, in the Panama Papers sample, firm size is positive, though not significant. This result is in line with the research of Morck and Yeung (1991). The variable lnSizei becomes negative and
highly significant, with a p-value of 0.001, in the regression of the Paradise Papers. This result shows that tax avoidance possibilities decrease firm value, rather than increase it. Multinationals are thus expected by investors to engage in tax planning as they have the resources to do it.
3.5 Discussion
The publication of the Panama Papers and Paradise Papers has increased the public awareness of corporate tax avoidance. Tax avoidance itself is beneficial for shareholders. An increase in public awareness can however raise the costs associated with tax avoidance. Including an accusation of engaging in tax planning. These costs can outweigh the benefits.
The publication of the Panama Papers had little implications for the Dutch market, as few well know Dutch firms were mentioned. The costs associated with tax avoidance are in these circumstances lower than the benefits. An aggressive tax strategy is therefore optimal for
shareholders, as value is added to the firm. When a market is affected, as the Dutch market was with the publication of the Paradise Papers, the costs of an accusation of tax avoidance diminishes the advantages. These results are in line with the findings of Hanlon and Slemrod (2009) for US firms, and the findings of Choy et al. (2017) for British firms. That the Netherlands is perceived to be a tax haven, does not influence the reaction of shareholders.
4. Conclusion
4.1 Summary and conclusion
In April 2016 and subsequently in November 2017, two leaks concerning corporate tax avoidance were published. These leaks are the Panama Papers and Paradise Papers. Public interest increased after the publication and accused firms faced possible reputational repercussions. Tax avoidance is beneficial to shareholders at it shifts wealth from the state to investors. An accusation can however increase the costs of tax avoidance to such an extent, that engaging in tax planning becomes costly.
In this paper, an event study is performed for both the publication of the Panama Papers as for the publication of the Paradise Papers. The effect of an accusation of tax avoidance on shareholder value is investigated. As the accusations are large-scale, firms not mentioned by the leaks might also be affected. Shareholder value is approximated using the cumulative abnormal return for Dutch stocks. Tax avoidance is estimated using the effective tax rate of each firm.
The results indicate that Dutch firms are not negatively influenced by the publication of the Panama Papers. The coefficient of ETR is significant at the one percent level. Dutch organizations are negatively influenced by the publication of the Paradise Papers. With a p-value of 0.1 percent, this result is significant at the one percent level. It becomes clear that costs associated with accusations of tax planning, can diminish the positive effects of tax avoidance. Following these results, it can be concluded that the costs of an accusation of tax avoidance outweigh the benefits when many firms within the market are accused. In markets with less attention to tax avoidance, engaging in tax planning remains most favourable option for shareholders.
4.2 Recommendations for further research
This paper has attempted to empirically add to the discussion into the effect of tax avoidance on shareholder value. There are however limitations to the research performed in this paper. Firstly, this paper has focussed on the Dutch market. The external validity to countries with bigger economies is therefore limited. The sample size of both the Panama Papers and Paradise Papers is also very small. This can affect the validity of the results. Moreover, future research is recommended using a database with more complete observations. Furthermore, all Dutch firms are included in the sample. It would be interesting to compare the market as a whole to a sample including only accused firms. This can isolate the effect of an accusation more precisely. Finally, the public attention to tax avoidance might have only a temporary effect. As the publications are recent, long term effects will need to be studied to determine if tax avoidance is indeed not in the best interest of shareholders anymore.
Reference list
Literature
Alecci, S. (2017). Leaked records reveal offshore’s role in forest destruction. ICIJ. Retrieved from: https://www.icij.org/investigations/paradise-papers/leaked-records-reveal-offshores-role-in-forest-destruction/
Austin, C. R., & Wilson, R. J. (2017). An Examination of Reputational Costs and Tax Avoidance Evidence from Firms with Valuable Consumer Brands. The Journal of the American Taxation
Association. Vol. 39(1), pp. 67-93.
Bauckloh, T., Hardeck, I., Wittenstein, P., & Zwergel, B. (2017). Do U.S. Multinational Firms Benefit from Competitive Tax Advantages through Advanced Tax Rulings in Europe? ResearchGate – Working Paper.
Berkhout, E. (2016). Nederland Belastingparadijs. Oxfam Novib, Den Haag.
Chen, S., Chen, X., Cheng, Q., & Shevlin, T. (2010). Are family firms more tax aggressive than non-family firms? Journal of Financial Economics. Vol. 95, pp. 41-61.
Choy, S. K., Lai, T. K., & Ng, T. (2017). Do tax havens create firm value? Journal of Corporate
Finance. Vol. 42, pp. 198-220.
Desai, M. A., & Dharmapala, D. (2009). Corporate Tax Avoidance and Firm Value. The review of
Economics and Statistics, Vol. 91(3), pp. 537-546.
Desai, M., & Dharmapala, D. (2006). Corporate tax avoidance and high- powered incentives. Journal
of Financial Economics. Vol. 79, pp. 145–179.
Dyreng, S. D., Hanlon, M., & Maydew, E. L. (2010). The Accounting Review. Vol. 85(4), pp. 1163-1189.
Dyreng, S., Hoopes, J., & Wilde, J. (2016). Public pressure and corporate tax behavior. Journal of
Accounting Research. Vol. 54 (1), pp. 147–186.
Fisher, L. J. M., & Roll, R. (1969). The Adjustment of Stock Prices to New Information. International
Economic Review. Vol. 10, pp. 1-21.
Friedman, M. (1970). The social responsibility of business is to increase profits. New York Times
Magazine (September 13), pp. 32–33.
Hanlon, M., & Slemrod, J. (2009). What does tax aggressiveness signal? Evidence from stock price reactions to news about tax shelter involvement. Journal of Public Economics, Vol 93, pp. 126-141.
Hoi, C. K., Wu, Q., & Zhang, H. (2013). Is Corporate Social Responsibility (CSR) Associated with Tax avoidance? Evidence from Irresponsible CSR Activities. The Accounting Review. Vol. 88 (6), pp. 2025-2059.
Morck, R., & Yeung, B. (1991). Why Investors Value Multinationality. The Journal of Business. Vol. 64 (2), pp. 165-187.
O’Donovan, J., Wagner, H. F., & Zeume, S. (2017). The Value of Offshore Secrets: Evidence from the Panama Papers. HKUST Finance Symposium 2017.
OECD. (2013). Action Plan on Base Erosion and Profit Shifting. OECD Publishing. Retrieved from: http://dx.doi.org/10.1787/9789264202719-en
Rego, S. O. (2003). Avoidance Activitites of U.S. Multinational Corporations. Canadian Academic Accounting Association. Vol. 20 (4), pp. 805-833.
Stock, J. H., & Watson, M. W. (2015). Introduction to Econometrics: Updated Third Edition. Pearson
Education Limited. England, Edinburgh Gate.
Swenson, C. (1999). Increasing stock market value by reducing effective tax rates. Tax Notes. Vol. 83 (June 7), pp. 1503-1505.
Newspaper Articles
Farrel, S., & McDonald, H. (2016). Apple ordered to pay €13bn after EU rules Ireland broke state aid laws. The Guardian, 30 August 2016. Retrieved from:
https://www.theguardian.com/business/2016/aug/30/apple-pay-back-taxes-eu-ruling-ireland-state-aid
FD. (2017). Brussel onderzoekt belastingontwijking Ikea via Nederland. FD, December 18 2017. Retrieved from:
https://fd.nl/economie-politiek/1233027/brussel-onderzoekt-belastingontwijking-ikea
Kleinnijenhuis, J. (2016). Nederland heeft een probleem. Trouw, April 10 2016. Retrieved from: https://www.trouw.nl/home/nederland-heeft-een-probleem~ad138c6e/
Appendix
Appendix I
Scatterplot of the Panama Papers sample. The data is best described using a linear model. The variables have been scaled by 100 percent to make the graph easier to interpret.
Below is the scatterplot for the Paradise Papers sample. The first graph predicts the linear model. The second approaches the data using a quadratic model. As can be seen, the quadratic model fits the data better. The data is also here, scaled by 100 percent for interpretive reasons.
-1 0 -5 0 5 10 0 10 20 30 40 ETR100
Paradise100 Fitted values
-1 0 -5 0 5 10 0 10 20 30 40 ETR100
Paradise100 Fitted values
-6 -4 -2 0 2 4 0 10 20 30 40 ETR100
