Corporate Tax Avoidance of Private Equity-backed Firms in the United States and the European Union.

(1)

1

Corporate Tax Avoidance of Private Equity-backed Firms in the

United States and the European Union.

Peter Eskes

Student number: 1789023

Master Thesis

University of Groningen

Faculty of Economics and Business

MSc. Finance

Version date: 10 January 2014

(2)

2

ABSTRACT

___________________________________________________________________________ In this thesis, the tax avoidance activities of firms that are owned by Private Equity (PE) firms are investigated. As tax avoidance is hard to measure, I use two tax avoidance proxies: the book-tax difference and the effective tax rate. To investigate the impact of PE ownership, I collected firm-year observations for the three years before and three years after the year of delisting. Prior research (for instance Badertscher et al., 2009) has studied the corporate tax avoidance of PE-backed firms that are located in the U.S. My study includes both the U.S. as well as the E.U. market, therefore a comparison is made between these two markets. The results indicate that PE-backed firms in the U.S. engage in less nonconforming tax avoidance in private years compared to public years. This is explained by the higher leverage ratios and interest deductibility that result in fewer incentives to avoid taxes. In contrast with the U.S. market, PE-backed firms in the E.U. engage in more nonconforming tax avoidance in private years. Interestingly, the leverage ratios of E.U.-firms after the delisting did not increase substantially. This result strengthens the idea of a negative relationship between leverage and tax avoidance. The conclusion is that the U.S. market and the E.U. market show significant differences with respect to tax avoidance of PE-backed firms.

___________________________________________________________________________ Keywords: ownership structure, leverage, buyout, delisting, taxation, taxes, private equity, tax planning

(3)

3

1. Introduction

“Tax structuring is key to private equity success” is the title of a paper, written by Lars Jonsson (2005) during a big private equity wave in March 2005. He states (p. 1): “putting together tax-effective private equity transactions in Europe can be complex”. However, most of the time, tax leakage can be mitigated and sometimes entirely eliminated by pre- and/or post-acquisition tax planning. In this paper I am going to investigate whether private equity firms create shareholder value when they use tax planning strategies to manage the tax liabilities of their portfolio firms.

To introduce the reader into the topic of this paper, section 1 of this chapter starts with an introduction to private equity. This will explain what private equity firms do and how they are organised. Moreover, it will provide characteristics of the private equity transactions. The second section explains why private equity firms have been characterized as ‘greedy’ but it will also discuss the proponents of private equity. The next two sections discuss the importance of tax and tax planning for firms, private equity firms included. The final section contains a short introduction to the methodology and data used in this paper.

Important to note: the investigated tax planning strategies are tax avoidance strategies and not tax evasion strategies. The difference between these two types is that tax evasion is by definition illegal while tax avoidance is the legal use of tax laws to reduce the tax burden. Tax evasion is often associated with (unaccepted) dishonest tax reporting while tax avoidance is linked to a more commonplace use of a tax regime to reduce the amount of tax payable. However, tax avoidance is nowadays less accepted from a moral point of view than ten years ago.

1.1. Private Equity

(6)

6

leveraged buyouts (LBOs) where they typically get majority control of an existing or mature firm (Kaplan and Strömberg, 2008).

In a typical leveraged buyout a company is acquired by a specialized investment firm using a relatively small portion of equity and a relatively large portion (60 to 90 percent) of outside debt financing, resulting in highly leveraged portfolio companies (Kaplan and Strömberg, 2008).

A typical PE firm is organized as a limited liability corporation or a partnership and raises equity capital through a private equity fund to finance the remaining 10 to 40 percent of the buy-out price (Kaplan and Strömberg, 2008). Most private equity funds are “close-end” vehicles in which investors provide a certain amount of money to pay for investments in companies as well as management fees to the PE firm. Investors cannot withdraw their funds until the fund is terminated (Kaplan and Strömberg, 2008).

The private equity funds are organized as limited partnerships where general partners are charged with managing the fund while the limited partners provide most of the capital (Badertscher et al., 2009). Institutional investors, such as corporate and public pension funds, insurance companies, as well as wealthy individuals are typical examples of limited partners. The private equity firm serves as the fund’s general partner. The limited partners have little to say in how the PE firm deploys the investment funds, as long as the basic covenants of the fund agreement are followed (Kaplan and Strömberg, 2008).

Private equity funds have a limited contractual lifetime, usually ten to thirteen years. After investing the capital committed to the fund into companies, the PE fund has usually five to eight years to return the capital to its investors. Therefore, exiting the investments is an important aspect of the private equity process (Kaplan and Strömberg, 2008).

(7)

7

In the further sections of this paper I will use the term private equity firm instead of private equity fund because the firm is responsible for managing the fund and the fund can be seen as an acquisition vehicle where the limited partners invest in.

1.2. Criticism and proponents

A huge fraction of the worldwide historic buyout activity took place from 2005 through June 2007, representing 43% of the total real value of the transactions from 1984 to 2007. In these 2.5 years more than 5,000 transactions were reported with a combined estimated enterprise value of $ 1,563,250. Though, private equity transactions are very cyclical, the last decades there have been several peaks and bottoms. The value of transactions peaked in 1988, plunged during the early 1990s, rose and peaked in the later 1990s, dropped in the early 2000s and increased spectacularly from 2004 to 2007 (Kaplan and Strömberg, 2008).

Katz (2009, p. 624) notes that “The rapid growth and increasing globalization of the PE industry has raised demands for increased regulation and disclosure within the sector due to concerns regarding anticompetitive behaviour, excessive tax benefits, and stock manipulation”. In several (news)papers (for instance the Wall Street Journal of October 10, 2006) and magazines (for instance Forbes, December 10, 2007) PE-backed IPOs (also known as reverse-LBOs) have been subject of public scrutiny. The media accused PE firm owners and managers as having excessively low tax rates. Therefore, private equity firms have been characterized as ‘greedy’ and ‘self-serving’ (Badertscher et al., 2009). Additionally, critics oppose that PE firms aggressively manage their tax liabilities and those of their portfolio companies (Behind the Buyouts, 2007. http://gobnf.org/i/wog/behindthebuyouts.pdf). This paper investigates this latter allegation.

(8)

8

(because the companies are not listed anymore) reduces management’s incentive to manipulate short-term performance (Kaplan and Strömberg, 2008).

The results in prior research suggest that effective monitoring, combined with PE firms’ financial, governance, and operational strategies, as well as reputational considerations, have a positive impact on PE-backed firms’ long-term financial performance, as well as financial reporting quality (e.g. Cao and Lerner, 2009; Kaplan and Strömberg, 2008; Acharya et al., 2009; Katz, 2009). Prior research has used different methodologies, proxies and time periods. The ratio of cash flow to sales, the ratio of operating income to sales, total factor productivity and industry-adjusted stock performance after an IPO are some examples of proxies that have been used to measure the long-term financial performance of PE-backed firms.

1.3. Taxes

Managing taxes is important for financial performance and increasing enterprise value since taxes represent a significant cost to a company. Taxes are a reduction in cash flows available to the firm and shareholders. This is why there are firm’s and shareholders’ incentives to reduce taxes through tax (aggressive) strategies.

The U.S. Department of the Treasury (1999, p.7) notes that “the focus on increasing corporate earnings and managing the reported effective tax rates has caused many corporations to treat their in-house tax departments as profit centres.” The view of corporate accountants goes even further. A high effective tax rate (ETR) is a sign of weakness (Lisowsky, 2010).

Managers promoting tax avoidance have a significant effect on their firms’ reported earnings since reductions in traditional effective tax rates translate directly into increased reported earnings. However, tax planning does not always lead to shareholder value maximization as there are potential costs of being tax aggressive, including reputational costs and non-tax costs arising from agency conflicts or from manager’s hidden actions (Chen et al., 2008).

(9)

9

Rent extraction takes place when decision makers choose for non-value maximizing activities at the expense of shareholders (Chen et al., 2008). This happens most of the time for luxury things like airplanes or luxury hotel bookings. Desai and Dharmapala (2004, 2006) indicate that tax avoidance activities, such as creating complex structures involving tax-indifferent related parties, are designed to obscure the underlying intent. For this reason, these tax avoidance activities are hard to detect by the Internal Revenue Service (IRS). Moreover, this obscure nature makes it easier for decision makers to hide rent extraction activities.

1.4. Tax planning

Managing taxes via tax planning can take place in two different ways, namely book-tax conforming book-tax planning and book-book-tax nonconforming book-tax planning. Tax planning that reduces a firm’s income tax but not its financial income is called book-tax nonconforming tax planning. This form of tax planning generates temporary or permanent book-tax differences. On the other hand, book-tax conforming tax planning, including the tax benefits of debt financing, is tax planning that affects book and taxable income in a similar manner (Badertscher et al., 2009).

The first question that I raise is: “Is the debt tax shield used in LBOs the only form of value creation for the highly leveraged portfolio firms or do PE firms implement (aggressive) tax planning strategies to (further) decrease the taxes payable?” In particular, I investigate whether there are tax planning differences between private (non listed) and publicly-traded (listed) companies that are majority-owned by PE firms (‘majority PE-backed firms’). To what extend can these differences be explained by book-tax conforming or book-tax nonconforming tax strategies?

1.5. Methodology and data

(10)

10

Previous research has investigated the United States market (Badertscher et al., 2009) or a specific country like Sweden (Bergström et al., 2007) or France (Boucly et al., 2008). This study distinguishes itself with investigating both the U.S. as well as countries that are part of the European Union. Therefore it is possible to make a comparison between these two markets, as far as I know is this the first time such a comparison of tax avoidance by PE-backed firms is made.

The sample in this study allows for a direct comparison of tax avoidance activity across the public and private phases of the same firms in different industries and different countries. It consists of firms that are linked to private equity firms and that are delisted in the period 2001-2010. The delisting should be the result of a private equity leveraged buy-out. To investigate this, I used SEC announcements and the database Zephyr. The companies are selected by availability of financial data.

Prior research has used data from Compustat to investigate the U.S. market. Due to the unavailability of European data in Compustat, I use Orbis to get European data as well as U.S. data. The methodology I am going to use is a pooled least squares regression to estimate whether the tax avoidance proxies can be explained by a private firm-year indicator variable. I use several variables, used in previous research and indicated in the literature, to control for effects that are not the result of the delisting and taking over of private equity firms.

This study should be of interest to tax regulators, the OECD and other parties that are concerned with the tax practises of firms that are owned by private equity. Moreover, my results should be of interest to critics that accuse PE firms of taking excessively advantage of tax avoidance, at least for the U.S. market. The results indicate that U.S. PE-backed firms engage in less nonconforming tax planning while E.U. PE-backed firms engage in more nonconforming tax planning although European PE-backed firms do not make excessively use of the debt tax shield.

(11)

11

2. Prior Literature

This chapter provides an overview of the previous research on tax avoidance and shareholder value creation. In the first section, I describe some recent work of the Organisation for Economic Co-operation and Development (OECD) and their contribution in targeting tax avoidance. Governments lose considerable corporate tax revenue due to base erosion and profit shifting (BEPS) by multinational enterprises. The OECD describes the effects for governments and provides an action plan to address BEPS. The OECD states that the ETR is attractive to investigate BEPS activities.

Hybrid mismatch arrangements using hybrid entities or hybrid instruments are examples of BEPS strategies. These arrangements are classified by the OECD as aggressive tax planning. However, these activities increase the shareholder value of the firm. Hence, the second section is dedicated to the relationship between tax avoidance and shareholder value. When this relationship is clear, previous research on tax avoidance is discussed in section three. In the final section I will describe my hypotheses.

2.1. OECD

2.1.1. Base Erosion and Profit Shifting

The OECD is an international economic organisation to stimulate economic progress and world trade. The OECD “uses its wealth of information on a broad range of topics to help governments foster prosperity and fight poverty through economic growth and financial stability” (OECD website, http://www.oecd.org/about/whatwedoandhow/). The OECD is well known for its collaboration regarding taxation, this has encouraged the growth of a global web of bilateral tax treaties. Its model tax convention serves as a template for bilateral negotiations regarding tax cooperation and coordination.

(12)

12

The OECD noticed the growing perception of governments losing considerable corporate tax revenue because of tax planning aimed at shifting profits in ways that erode the taxable base to locations where they are subject to a more favourable tax treatment.

The BEPS Report reflects the view that national tax laws have not kept pace with the globalisation of corporations and the digital economy, in particular in the area of intangibles and the development of e-commerce. MNEs can exploit gaps to artificially reduce their taxes while base erosion constitutes a serious risk to tax revenues, fairness and sovereignty for many countries. The BEPS Report concludes that there is a need for increased transparency on the effective tax rate of MNEs. Current rules provide opportunities to associate more profits with legal constructs together with intangible rights and obligations. Besides that, intra-group risk shifting has the result of reducing the share of profits associated with substantive operations. The BEPS Report supports the idea that the allocation of taxation rights should follow the location of “real business activities”. However, intellectual property rights can be shifted, in an early stage, to a group member that is located in an off-shore tax haven. In a later stage, when these rights contribute considerably to the reported income, a high portion of the profit is shifted to this off-shore tax haven without any real business activities. Contractual risk shifting is considered as means to allocate income arbitrarily (Schön, 2009).

In response to the recommendation of the BEPS Report to develop an action plan to address BEPS issues in a comprehensive manner, the OECD launched its Action Plan on Base Erosion and Profit Shifting in July 2013. This action plan “offers a global roadmap that will allow governments to collect the tax revenue they need to serve their citizens. It also gives businesses the certainty they need to invest and grow” (Public Press release, OECD,

http://www.oecd.org/ctp/closing-tax-gaps-oecd-launches-action-plan-on-base-erosion-and-profit-shifting.htm). The Action Plan aims to solve the problem of international tax rules, that make sure that businesses are not confronted with double taxation (pay taxes in two countries), which rules are unfortunately now being abused to permit double non-taxation (not paying any taxes in both countries).

(13)

13

The (backward-looking) ETR is attractive to investigate BEPS activities because it is based on measures of actual taxes paid, and therefore capturing the range of factors impacting actual tax liability. However, an extremely low ETR can be the result of aggressive tax planning strategies or the achievement of the government policy that a given incentive was meant to promote, for instance the R&D tax credit. The OECD notes that, in principle, consistently measured ETRs could provide useful indications of whether BEPS is indeed taking place.

2.1.2. International (aggressive) tax planning

Lisowsky (2010, p. 1698) numerates three ways in which tax sheltering can be achieved. “First, they artificially inflate or shift asset bases to other (typically tax-indifferent) parties upon sale to decrease gains subject to tax or shift gains to a lightly taxed jurisdiction or party. Second, they accelerate or inflate expenses to decrease current taxable income. Third, they defer revenue to understate taxable income.” Remarkably, these three ways correspond with the four parts that are essential for successful tax planning, according to the OECD.

In the eyes of the OECD, any international tax planning will need to incorporate a number of co-ordinated strategies to be successful. These strategies can be broken down into four parts:

- Minimisation of taxation in a foreign operating or source country. - Low or no withholding tax at source.

- Low or no taxation at the level of the recipient.

- No current taxation of the low-taxed profits (at the level of the ultimate parent).

Aggressive tax planning issues being addressed in the BEPS Report are: avoidance of withholding tax at source, artificial interest deduction, debt-push down and circumvention of CFC (controlled foreign company) legislation and thin capitalisation rules. This could be achieved by using transfer pricing, leverage, low-taxed branch of a foreign company, hybrid entities, hybrid financial instruments, conduit companies and/or derivatives.

(14)

14

and host countries. This corresponds with the main goal of tax shelters: to lower the corporate tax liability by taking advantage of discontinuities in the tax law.

Several countries, for instance France, respond to these tax avoidance strategies by introducing anti-thin capitalisation rules. These rules restrict the deductibility of net interest expenses. However, the Netherlands abolished its thin capitalisation rules from 1 January 2013. Instead, they introduced a specific anti-abuse provision. This new provision aims to limit the deductibility of excessive interest expenses. These interest expenses are related to debts that are incurred to acquire or expand participations in subsidiaries that qualify for the participation exemption. The provision should only work in case of abuse. Moreover, the OECD model tax treaty contains several articles which have a subject-to-tax clause. For instance, the business profits of a company of a contracting state "shall be taxable only in that State" unless the company carries on business in the other contracting state through a permanent establishment situated therein (Article 7 (1) OECD model tax treaty). These clauses are not only helpful for avoiding double taxation but also to fight tax avoidance.

2.1.3. Conforming versus nonconforming tax planning

Tax planning that reduces a firm’s income tax but not its financial (or book) income is called book-tax nonconforming tax planning. This form of tax planning generates temporary or permanent book-tax differences. On the other hand, book-tax conforming tax planning, including the tax benefits of debt financing, is tax planning that affects book and taxable income in a similar manner. The difference between taxable income and book income is the way it is derived. Taxable income is calculated by applying tax law and fiscal standards while book income is calculated using (financial) accounting standards such as U.S. Generally Accepted Accounting Principles (U.S. GAAP) or International Financial Reporting Standards (IFRS).

Creating permanent book-tax differences is the most valuable form of value creation. Temporary book-tax differences have only a liquidity benefit. The temporary book-tax differences result in higher actual cash flows while future cash flows will be lower when the book-tax differences are reversed. Due to discounting, the present value of cash flows will be higher. In the case of permanent book-tax differences, these differences are not reversed resulting in higher actual cash flows and no lower future cash flows.

(15)

15

by Badertscher, Katz and Rego (2009). Badertscher et al. investigated 227 U.S. private firms with public debt that have both private firm-years and public firm-years in the period 1980-2005. Their main results indicate that privately held majority PE-backed firms pay 15 cents less income tax per dollar of EBIT than publicly held majority PE-backed firms, while the private firms engage in less nonconforming tax planning.

2.1.4. Hybrid Mismatch Arrangements

On 5 March 2012, the OECD released a report entitled Hybrid Mismatch Arrangements: Tax Policy and Compliance Issues (the Hybrid Report). Hybrid entities and hybrid instruments are noticed in the BEPS report as examples of aggressive tax planning arrangements. The Hybrid Report describes the most common types of hybrid mismatch arrangements and the effects they aim to achieve. These arrangements can be used to exploit national differences in the tax treatment of instruments, entities or transfers. They “achieve results such as (i) the multiple deduction of the same expense in different countries, (ii) the deduction of a payment in the country of the payer without a corresponding inclusion in the country of the payee and (iii) multiple tax credits for a single amount of foreign tax paid” (the Hybrid Report, p. 12).

A hybrid entity is an entity that is treated as transparent or disregarded for tax purposes in one country while the same entity is non-transparent in another country. An instrument that is treated differently for tax purposes in the countries involved is called a hybrid instrument. A sale and repurchase agreement concerning shares can be an example of a hybrid transfer. In country A the transaction is treated as a sale and a repurchase of the shares, while in country B it is treated as a loan with the shares serving as collateral.

The Hybrid Report presents examples of double deduction, deduction / no inclusion, and foreign tax credit generator schemes. Hereafter, the same examples are provided. Therefore, I refer to the Hybrid Report for more (background) information. The wavy blue lines in the next three figures indicate a national border.

(16)

16

FIGURE 1

Double deduction of interest structure

Figure 1 provides a structure to create a deduction of interest at A Co as well as at B Co. The Hybrid Entity is transparent for country A tax purposes. Its interest expenses are therefore allocated to A Co. Since Hybrid Entity is not transparent (disregarded) for country B tax purposes, Hybrid entity is subject to corporate income tax in country B. Its interest expenses can be used to offset income of B Co since Hybrid Entity and B Co are part of a group tax regime. A double deduction of interest expenses is the result.

The parent company in country A (“A Co”) holds all or almost all equity interests in an entity (“Hybrid Entity”) that is treated as transparent for country A tax purposes and as non-transparent for country B tax purposes. This Hybrid Entity holds all or almost all equity interests in the operating company in country B (“B Co”). Next, Hybrid Entity borrows from a third party to inject the loan amount as equity into B Co (or buy the shares in B Co).

(17)

17

Another effect of hybrid mismatch arrangements can be a deduction in one country but the avoidance of a corresponding inclusion in the taxable income in another country. Such a deduction / no inclusion scheme can be accomplished with the next structure (Figure 2):

FIGURE 2

Deduction / no inclusion structure

Figure 2 provides a structure to create a deduction of interest at B Co while there is no inclusion for the recipient (A Co). A hybrid instrument is used, this instrument is treated as equity in country A but as debt in country B. The interest expenses are therefore deductible for B Co while the receipts of A Co qualify as dividends. Since capital gains and dividends are generally exempt from corporate income tax, there is no corresponding inclusion in country A. A net deduction with no corresponding inclusion is the result.

This scheme uses a hybrid instrument: an instrument that is treated as equity in country A but as debt in country B. Payments that are made under this instrument qualify as interest expenses for B Co under country B tax law. Hence they are deductible for B Co. In contrast, the receipts of A Co qualify as dividends for A Co under country A tax law. Generally, capital gains and dividends derived from a qualifying subsidiary are exempt from corporate income tax in country A. Resulting in a net deduction in country B and no corresponding inclusion in country A.

(18)

18

FIGURE 3

Foreign tax credit generator structure

Figure 3 provides a structure for B Co to claim a foreign tax credit for the corporate income tax paid in country A by Special Purpose Vehicle (SPV). Therefore A Co enters into a “Repo” agreement: A Co sells its shares and enters into a repurchase agreement with B Co. In country A this agreement is treated as a loan by B Co to A Co that is collateralised with SPV shares. A Co is therefore seen as the holder of the shares in SPV. Country A applies a method that allows A Co to receive the dividends effectively tax-free. Since the agreement is treated as a loan, A Co can claim a deduction for the interest expenses. In contrast, the agreement is treated as a sale and repurchase in country B. B Co qualifies as the holder of SPV shares. Since the dividends are exempt, B Co receives the dividends tax-free. Moreover, B Co is able to claim a foreign tax credit for the corporate income tax paid by SPV in country A since country B has an indirect foreign tax credit regime. The result is a foreign tax credit generator.

(19)

19

time (indicated with (2) in the figure), SPV earns income and pays corporate income tax to country A. Moreover, SPV pays out dividends to B Co, which serves as B Co’s fee/compensation in the transaction.

This scheme is a foreign tax credit generator since the repo is treated differently between countries. In country B it is treated as a sale and a repurchase. B Co qualifies as the owner of the SPV (preferred) shares and the recipient of the dividends. Moreover, B Co is able to claim a foreign tax credit for the corporate income tax paid by SPV in country A since country B has an indirect foreign tax credit regime.

In contrast, for country A tax purposes, the repo is treated as a loan by B Co to A Co that is collateralised with the SPV shares. A Co is thus seen as the owner of the SPV (preferred) shares and as the recipient of the dividends. “Country A applies an exemption for dividends received by B Co, or an indirect foreign tax credit regime that allows A Co to claim a tax credit for the corporate income tax paid by SPV, in any case a method that allows A Co to receive the dividends effectively tax-free” (the Hybrid Report, p. 10). Since the repo is treated as a loan, A Co can claim a deduction for the interest expenses. The interest expenses are equal to the dividend payments (the remuneration for B Co).

In the Hybrid Report, the OECD describes four possible domestic law options for countries concerned with hybrid mismatch arrangements. The first solution would be harmonisation of domestic laws. However, the elimination of differences in the tax treatment of entities, instruments and transfers does not seem possible. This option is only mentioned for the sake of completeness. Another option that can be effective in addressing some hybrid mismatch arrangements would be general anti-avoidance rules. Some judicial doctrines that are often included in these rules are “economic substance”, “abuse of law” and “business purpose”. However, frequently there needs to be a direct link between the transactions and the avoidance of that particular tax. This makes the application of general anti-avoidance rules in hybrid mismatch cases difficult.

(20)

20

mismatches between countries is eliminated by linking the domestic tax treatment of an entity, instrument or transfer involving a foreign country to the tax treatment in the foreign country. The OECD notes (p. 15) that “domestic law rules which link the tax treatment of an entity, instrument or transfer in the country concerned to the tax treatment in another country appear to hold significant potential as a tool to address hybrid mismatch arrangements that are viewed as inappropriate”.

2.2. Theory on shareholder value through tax strategies

The tax revenue of governments is decreased due to double non-taxation, artificial interest deduction, avoidance of withholding tax at source and other aggressive tax planning issues. However taxes represent a cost to the firm; a decrease of the effective tax rate results in an increase of after-tax earnings thereby boosting firm value.

2.2.1. Effective tax rate

Since firm values are often based on the firm’s reported earnings, an increase in after-tax earnings boosts the firm value. A reduction in the effective after-tax rate translates directly into increased reported earnings. For example, a reduction of 5 %-point of the effective tax rate, from 35 % statutory tax rate (US) to an effective tax rate of 30 %, would result in an 7.7 percent increase in after-tax earnings (5/65 = 7.7 %).

2.2.2. Debt tax shield

The first part of this subsection is concerned with the theoretical research on debt tax shields while the second part discusses the empirical research. The debt tax shield has been subject of decades of debate regarding firm valuation and the cost of capital. Modigliani and Miller (1963) (MM), first hypothesized that the tax benefits decrease the cost of using debt capital and increase firm value. Miller (1977) reasoned that the tax benefits of debt were (partially) distributed to creditors since creditors were compensated for the personal tax disadvantage of debt by higher interest rates. DeAngelo and Masulis (1980) proposed that at least some of the tax benefits are offset by the financial distress costs of debt.

(21)

21

considerable amount. However, Cooper and Nyborg (2005) have explained that the mistake in Fernandez’ paper comes from mixing the Miles and Ezzell and Miller and Modigliani leverage policies (Miles and Ezzell, 1980; Modigliani and Miller, 1963).

Obviously, the amount of debt should not exceed that amount at which the interest is still deductible. Loss carry forwards and carry backs should be taken into account when this amount is determined. In theory, when a firm earns a lot of profit, the whole firm should be financed with debt to maximize the debt tax shield. This suggests that firm value (shareholder value) is positively associated with (more) debt.

Empirical results: using cross-sectional regressions to estimate the value of the debt tax shield, Kemsley and Nissim (2002) find that firm value is a positive, strong function of debt. Their estimated value for the debt tax shield is 40 percent (10 percent) of debt balances (firm value).

Not only is the relationship between debt and firm value investigated, the relationship between debt and corporate taxes has been subject of research. Various empirical studies in the 1980s tested whether non-debt tax shields, such as investment tax credits or depreciation, reduce the tendency to use debt tax shields. However, none of these studies found significant tax effects. In contrast, empirical studies developed in the 1980s find significant cross-sectional evidence that high marginal tax rates promote the use of debt.

Examples of research are studies of MacKie-Mason (1990) and Graham (1996, 1999). In these studies is empirically proven that debt policy is a function of corporate taxes. Moreover, they show that a portion of the gross corporate tax benefits of debt is offset by the personal tax disadvantage of debt, but the magnitude is limited.

However, theoretical research indicates that high debt balances are also associated with costs. As explained before, costs of debt include personal taxes (Miller, 1977). Moreover, debt can cause financial distress costs (Scott, 1976) or a debt overhang problem (Myers, 1977).

(22)

22

positively sloped. This result is indicative of debt costs that increase with the amount of debt used.

They also estimate the net benefits of debt financing by integrating the area between the marginal cost and benefit curves. They estimate the all-in cost of debt to be 6.9% of firm value while the estimated gross benefit (including the tax benefit) is 10.4%. This results in a net debt benefit of 3.5% of firm value. The net debt benefit depends on the location of the cost function. This location varies with firm characteristics such as size, asset tangibility, asset collateral and cash flows.

Korteweg (2010) estimates the net benefits to leverage by estimating how observed variation in stock and bond betas and valuations can be explained by variation in leverage and other variables. Therefore he collected monthly debt and equity values for 290 U.S. firms in 30 industries. His results are consistent with those of Van Binsbergen, Graham and Yang. The median net benefits of debt are about 4% relative to total firm value.

Van Binsbergen, Graham and Yang (2010, p. 1) also note that “the cost of being overlevered is asymmetrically higher than the cost of being underlevered”. This shows the riskiness of (too much) debt and rejects the idea of ‘more is better’. A theory that states that more debt is always lead to higher firm values is therefore not true.

2.3. Prior literature on tax avoidance

Since tax planning can decrease the effective tax rate or increase the debt tax shield, tax planning can increase firm value. However, tax avoidance has some disadvantages as well. This section provides prior literature on tax avoidance. It starts with the history of tax avoidance research. Secondly, the benefits and costs are discussed. It is shown why tax avoidance is more beneficial for PE-backed firms than for non PE-backed firms. Section 3 provides information about book-tax conforming and non-conforming tax planning and why I chose to investigate nonconforming tax planning.

2.3.1. History and definition

(23)

23

In one of the earliest studies of firms’ overall tax avoidance, 30 years ago, Zimmerman (1983) examined whether firm size is related to the effective tax rate. The ETR is the ratio of total tax expense to pre-tax book income in any year. In the years that followed, more studies investigated the determinants of annual effective tax rates while also other proxies for tax avoidance aroused.

Not only tax avoidance has been investigated, recent literature is also concerned with aggressive tax planning and tax shelters. Can we conclude that recent research is sundry or are they investigating the same thing?

Frank et al. (2009, p. 469) define aggressive tax reporting as “downward manipulation of taxable income through tax planning that may or may not be considered fraudulent tax evasion”. Badertscher et al. (2009) describe aggressive tax strategies as strategies that are considered abusive by the Internal Revenue Service (IRS) and the Treasure Department, and give the example of sale-in-lease-out transactions. Lisowsky (2010) describe tax shelters as methods to lower the corporate tax liability by taking advantage of discontinuities in the tax law. Tax shelters are illegal when they are created for the sole aim of evading tax. In these cases they do not exhibit “economic substance” or a “business purpose”.

While a clear definition of tax avoidance is missing, it has been a subject for several studies. Most of the studies have in common that they are investigating the factors why some companies engage more in tax avoidance than other firms.

2.3.2. Benefits and costs

The most obvious benefit of aggressive tax planning is greater tax savings. While these tax savings swell into the accounts of shareholders, managers can be compensated, directly or indirectly, for their efforts in effective tax management. Next to this potential benefit for decision makers there is another way for managers to benefit from tax aggressiveness. Rent extraction, which refers to non-value maximizing activities like related party transactions or conceited consumptions, can be both enabled and masked by opaque tax avoidance activities (Chen et al., 2008).

(24)

24

Since a benefit for the firm (‘s shareholders) is often related to a cost to another party, for instance the tax revenue authority, I take the perspective of the company to describe the costs of being tax aggressive.

The most obvious cost of tax aggressiveness is the potential fine imposed by a tax revenue authority. However, Desai and Dharmapala (2006) argue that there is another significant non-tax cost. They state that shareholders will protect themselves by pricing down the firm when they cannot distinguish whether the managers avoid taxes to increase shareholder value or use it for rent extraction. That investors are aware of the potential managerial self-dealing problem is proved by Desai, Dyck and Zingales (2007). Tax enforcement in Russia increased after Putin’s election in 2000. Their case study on Sibneft, the 5th largest Russian integrated oil company, provided the first indication of an improved return for outside shareholders. However, Sibneft’s experience was not unique. Desai, Dyck and Zingales (2007, p. 33) provide empirical evidence that the “increased tax enforcement in Russia enhanced the value of targeted companies and reduced their control premia”. Firms that were affected by increased tax enforcement in Russia encountered an increase in market value.

Private equity firms are both the managers of the firm as well as the shareholders of the firm. It might be expected that they will be monitored by the limited partners of the fund in such a way that rent extraction will not excessively take place. However, private equity firms have a relatively short investment horizon (Kaplan and Strömberg, 2008). This short period is used to increase the shareholder value as much as possible. Since the firm value can be calculated as a multiple of earnings before interest, depreciation and amortization (EBITDA), which will be lower in the case of rent extraction, it is in the interest of PE firms not to make use of excessively high rent extraction practices.

The short investment period has also the consequence that the private equity firms are not imposed by the potential penalty of the tax revenue authority. Generally, there is a long lag between the design and implementation of tax aggressive transactions and the detection by the authority (Chen et al., 2008).

(25)

25

These ownership characteristics predict that shareholders of PE-backed firms have higher benefits of tax aggressiveness and have less (non-tax) costs than non PE-backed firms. This implies that PE-backed firms could be more tax aggressive than non PE-backed firms.

The ownership characteristics of private equity are not the only things that have got in impact on tax aggressiveness. The management of the firm can also have an effect on the firm’s tax avoidance. A typical executive board of a firm will not have the specific knowledge to understand and implement various difficult tax strategies. The CEO, for example, is likely to have knowledge of the competitive landscape and positioning of the firm. However, he/she is probably not a tax expert, sometimes not even in the possession of a financial background (Dyreng, Hanlon and Maydew, 2008).

Nevertheless, a CEO can affect tax avoidance by setting the tone at the top (Dyreng, Hanlon and Maydew, 2008). This can result in more emphasis on the tax area of the firm or it can change the relative emphasis of the firm’s internal and external (tax) advisors.

Dyreng, Hanlon and Maydew (2008) investigated the effects of managers on corporate tax avoidance. Their results show that the extent to which firms achieve tax savings depends to a great extent on how much involvement the firm’s tax advisors, both internal as well as external, have in the firm’s operational and financial decisions. They also indicate that managers play a significant role in determining the level of tax avoidance that firms undertake. Managers with an MBA or law degree are more likely to be associated with a lower cash effective tax rate.

In accordance with the results of Dyreng, Hanlon and Maydew is the statement of Badertscher et al (2009). They describe PE firms as firms with sophisticated managers that have accounting, management consulting or investment banking backgrounds. They assert (p. 12) that “Managers with these backgrounds are more likely to facilitate and promote aggressive tax avoidance at portfolio firms than managers with other backgrounds (e.g., engineering or product development)”. Since higher educated managers are associated with lower cash effective tax rates and tax sheltering requires sophistication to be successful, the statement of Badertscher et al. can be true.

(26)

26

aspects of a LBO by private equity. This way of thinking leads toward the idea that PE-backed firms employ more in tax planning activities in private years than in public years.

There are more differences than leverage ratios between public and private firms. These differences can be explanatory for differences in tax avoidance. Private firms have different executive compensation schemes so that managers of private firms fixate less on income reported in financial statements than managers at public firms. This theory is validated by Penno and Simon (1986), Klassen (1997) and Mikhail (1999). Moreover, private firms have less financial reporting requirements. This leads to lower financial reporting costs which can be expected to result in more tax planning.

Cloyd et al. (1996) and Mills and Newberry (2001) state that the greater financial reporting pressure to which public firms are generally subject causes their managers to further fixate on financial statement income, resulting in less tax planning at public firms than private firms. However, recent empirical research (Desai 2003; Graham and Tucker 2006; Frank et al. 2009; Wilson 2009) provides evidence of book-tax nonconforming tax strategies, while the arguments of Cloyd et al. and Mills and Newberry assume that firms engage in book-tax conforming tax planning.

To clear things up, Cloyd et al. and Mills and Newberry assume that tax planning reduces financial income while recent research indicates that tax planning can lower a firm’s taxable income but not its financial income, by using nonconforming tax strategies. Therefore, the prediction that public firms engage in less tax avoidance strategies than private firms may not hold empirically in my sample.

2.4. Hypotheses development

Taking into account the ideas of more tax avoidance due to the need to serve the high interest burdens and (private equity) ownership characteristics, like the short investment horizon (value creation focus and small chance of getting caught) and managers that bring expertise, lead to the first hypothesis

(27)

27

The use of a high portion of debt in a leveraged buy-out, results in a high tax deduction. As explained above, the interest burden can result in an increased focus on tax avoidance. On the contrary, this deduction leads to a lower income before taxes. Therefore, the need to avoid taxes can be lower. If the profit is low or even zero, there is not much left to avoid.

Graham and Tucker (2006) use a unique sample of 44 tax shelter cases in the U.S. to investigate the magnitude of tax shelter activity and whether the shelters are non-debt tax shields. They examine whether participating in a shelter is related to corporate debt policy. They provide evidence that the firms in their sample use less debt when they engage in tax sheltering.

However, they note that their results should be interpreted as showing correlations, not causality. So, a firm might resort to sheltering after finding out that it is unable to issue much debt (for what reason), or it might use less debt after having first established tax shelters. Or the other way around, a company that has a high debt level is using less tax shelters.

This is in line with the conclusion of Badertscher et al. (2009) that privately-held PE-backed firms in the U.S. engage in significantly less nonconforming tax planning. They suggest that their lower income taxes payable are the result of conforming tax planning, such as debt financing, and that this reduces the need to engage in nonconforming tax planning. The second hypothesis to test is

(28)

28

3. Methodology

This chapter provides in three sections the methodology that is used in this study. Section 1 provides a description of the tax avoidance proxies used in prior literature and the proxies I use in my study. The second section provides the regression equation and an explanation of the dependent, independent and control variables. The last section discusses the methodology of time series tests.

3.1. Tax avoidance proxies

Chen, Chen, Cheng and Shevlin (2008) investigated whether family firms are more tax aggressive than non-family firms, using four proxies of tax aggressiveness. The first proxy they employed is the effective tax rate (ETR), it uses the traditional definition of total tax expense (current and deferred) divided by pre-tax income. The ETR is computed on an annual basis. This proxy reflects tax aggressive planning through book-tax differences. Examples of such tax planning are investments in tax havens with lower foreign tax rates, investments in tax exempt or tax favoured assets, and participation in tax shelters that give rise to losses for tax purposes but not for book purposes.

Firm value is often calculated as a multiple of EBITDA or by using the discounted cash flow model. Therefore, the reported (book) income is an important determinant of firm value. Since a private equity buy out is often followed by an IPO (reverse-LBO) or a secondary sale, the reported income is crucial for PE managers. As the tax expense is divided by pre-tax income, the ETR proxy takes the managerial focus on reported (book) income into account. Hence, the ETR proxy is a relevant proxy for tax avoidance of PE-backed firms.

However, the ETR contains some measurement error since it does not control for book-tax differences that are the result of nondiscretionary items unrelated to tax planning (for example intangibles, depreciable and amortizable assets). Therefore I use several control variables that are explained carefully in section 3.2.2.

(29)

29

time intervals. However, due to insufficient time series data for each firm to follow this methodology, the CETR is often (for instance by Chen et al. (2008)) computed on an annual basis.

Dyreng et al. (2008) appropriately note that tax avoidance might not pinpoint tax aggressiveness. Avoidance can include legitimate deductions, while aggressiveness is a subset of avoidance because aggressiveness can also include prohibited activities such as tax sheltering. Nonetheless, this proxy has been used in several academic studies as a proxy for tax aggressiveness, for example in Chen et al. (2008).

Various studies indicate that large positive book-tax differences are associated with tax avoidance activity or with actual cases of tax sheltering (Wilson, 2009). However, since the book-tax difference can be a result of both tax planning and earnings management, book-tax differences can be a noisy proxy of tax planning activities. Next to that, Manzon and Plesko (2002) and Hanlon (2003) identify firm specific characteristics, like large capital expenditures, associated with book-tax differences (due to depreciation) that are not necessarily reflective of corporate tax planning. Earnings management and innate firm characteristics are the primary determinants of these book-tax differences, not tax planning activities.

These problems are solved by using control variables in the regression equation. Abnormal accruals (ABACCR) control for earnings management while assets (ASSETS) and sales growth (SALESGR) control for more depreciation of large and growing companies. Lastly, INTANG controls for differences in accounting and tax reporting standards of intangibles.

The contingent tax liability reserve, or tax cushion, has been studied as proxy of income smoothing and tax aggressiveness. The tax cushion of a firm is calculated as the difference between current tax expense and Total Tax after Credits adjusted for the tax benefit of stock options. Gleason and Mills (2002) estimated the level of cushion attributable to federal income taxes using private IRS (Internal Revenue Service) data and find that the contingent tax liability reserve is increasing in expected tax losses. If firms record cushion in accordance with Generally Accepted Accounting Principles (GAAP), firms who engage in more aggressive tax planning are expected to record more cushion.

(30)

30

related to tax shelter usage. He suggests that a nontrivial portion of the tax cushion likely reflects tax sheltering transactions.

Nevertheless, for calculating the tax cushion, tax return data from the IRS (or a comparable tax revenue authority) is required. Since this data is not publicly available, this measure of tax avoidance cannot be used in this paper.

As explained above, prior research has led to different proxies for tax avoidance. Including the effective tax rate, the cash effective tax rate, the book-tax difference and the contingent tax liability reserve (or tax cushion).

Due to unavailability of cash flow statements, detailed income statements and detailed balance sheets for private firms in the database Orbis, I can only employ the effective tax rate (ETR) proxy and the book-tax difference (BTD) proxy. Cash flow statements are required for getting data about actual taxes paid (CETR), detailed income statements and detailed balance sheets are needed for calculating other tax avoidance proxies. These other tax avoidance proxies (such as the Desai and Dharmapala & Manzon Plesko book-tax difference measures) are not mentioned here for the sake of brevity.

3.2. Regression equation

3.2.1. Dependent and independent variables

I examine the tax planning of firms that are majority-owned by PE firms during two distinct ownership phases. The first phase is the private phase; in this phase, the firm’s equity is privately-held. The second phase is the public phase, where the equity is publicly-traded. I estimate equation (1) while holding certain firm characteristics constant across the different ownership phases. The ‘full’ model that is used to investigate the U.S. market is given as:

(1)

(31)

31

If private firm-years exhibit more tax avoidance than public firm-years, then the coefficient on PRIVATE should be positive when BTD is the proxy of tax avoidance and negative when ETR is the proxy of tax avoidance.

3.2.2. Control variables

I control for firm characteristics that are related with my tax aggressive measures, as documented by prior research (Mills 1998; Manzon and Plesko 2002; Rego 2003; Chen et al. 2008; Dyreng et al. 2008; Frank et al. 2008; Wilson 2009).

In the next paragraphs I will explain the control variables of the ‘full’ model. The first set of factors that affect a firm’s tax avoidance activity, which includes LOSS, NOL and LEV, controls for a firm’s need to plan tax. Since more profitable firms and firms with less loss carry forward tend to have higher effective rates, these firms have greater incentive to plan tax. Furthermore, private firm-years exhibit greater losses than public firm-years (Katz 2009). For these reasons, LOSS and NOL are included to capture a firm’s profitability and the presence of net operating loss carry forwards (NOL) respectively. Leverage (LEV) is included because firms with greater leverage have more tax benefits of debt financing and have therefore less need to tax plan.

I also interact PRIVATE with LOSS, and LEV to further control for the significantly different average rates of these control variables in public vs. private firm-years.

The fourth control variable (INTANG) captures differences in book and tax reporting that can affect my nonconforming measures of tax avoidance. I include intangible assets (INTANG) in my regression to control for differential accounting rules for book and tax purposes of intangible assets.

Thirdly, I control for sales growth (SALESGR) as growing firms may make more investments in depreciable assets that generate timing differences in the recognition of expenses. Besides via SALESGR I control for firm size (ASSETS), since large firms likely enjoy economies of scale in tax planning.

(32)

32

recognition and earnings management by including ABACCR. This is the amount of abnormal accruals after controlling for conservatism (see Ball and Shivakumar 2006).

I use the Modified Jones model (1991) to determine the amount of abnormal accruals. Total accruals are calculated as the difference between net income and cash flows from operations (CFO) (Jones, 1991). Jones (1991) shows that large positive accruals could be an indication for earnings management but does not necessarily have to be the result of earnings management. Therefore, a distinction has to be made between non-discretionary or normal accruals and discretionary or abnormal accruals. The difference between these two types is that non-discretionary accruals can be explained by the firm’s performance or market environment, while the abnormal accruals are the result of earnings manipulation.

In 1991, Jones proposed a model to extract the non-discretionary part from the total accruals. To reduce heteroscedasticity, she scales all variables in her model by lagged assets. This model has been modified by Dechow et al. (1995) which is called the Modified Jones (1991) model. This model is the commonly used model to identify earnings management. The Modified Jones (1991) model is further explained in Appendix B.

3.2.3. The restricted model

Thanks to the general consensus in the U.S. on publishing annual reports on the website of a company, I got the data of all control variables for the companies that are active in the U.S. market. Most of the data that was still missing after inspecting the publicly available annual reports, I was able to obtain by using SEC filings. Due to unavailability of detailed income statements, detailed balance sheets and cash flow statements for the E.U. market, equation 1 is only estimated for the U.S. market.

That publishing of annual reports on the website of the company is not accustomed in the E.U. is affirmed by the observation that none of the E.U. companies included in the sample provides publicly available information on the Internet. Contacting these companies with the question of providing these statements for a research purpose did not result in the expected/hoped outcomes. Therefore, a ‘restricted’ model to investigate the E.U. market is given as:

(33)

33

The reader should be aware that the control variables NOL and ABACCR are excluded in the ‘restricted’ model. Firstly, this restricted model is used to investigate the U.S. market. After comparing the results of this restricted model with the results of the full model, something can be said about the validity of the restricted model. In case of a valid model, the restricted model can be used to investigate the E.U. market.

I note that adding variables that are related to standard methods of tax planning (intangible assets, loss carry forwards, etc) simply controls for the average tax effects associated with those variables. Controlling for these variables purges out the average ETR and BTD effect but leaves much room to capture firm’s tax planning activities.

3.3 Time-series tests

The full and restricted models are estimated using ordinary least squares (OLS) time-series regressions. The data consists of multiple firm-year observations for each firm. Due to unavailable firm-year observations, the data is structured as an unbalanced panel. The advantage of panel data is the possibility of testing with cross-sectional and/or period fixed or random effects. Firm fixed effects dummies are not included in the restricted and the full model to reduce endogeneity. In Section 5.3, the robustness test section, I investigate whether cross-sectional fixed effects should be included.

The regression residuals are tested for heteroscedasticity and autocorrelation, at a 5% significance level. If the regression residuals do not have a constant variance, they are said to be heteroscedastic. Heteroscedasticity is tested using White’s (1980) test, which tests the null hypothesis of no heteroscedasticity against heteroscedasticity of unknown, general form. If the ordinary least squares method is used when the errors are heteroscedastic, the estimators will still give unbiased coefficient estimates. However, the standard errors could be wrong resulting in inferences that could be misleading. The reason for this is that the residual variance does appear in the formulae for the coefficient variances. Therefore, in the presence of heteroscedasticity, I use White’s (1980) heteroscedasticity consistent covariance estimates to prevent the coefficient estimates to become inefficient.

(34)

34

using OLS when autocorrelation is present are similar to those of ignoring heteroscedasticity. Wrong inferences could be made on coefficients because the estimates are inefficient (but still unbiased). Since the Breusch-Godfrey test did not detect any autocorrelation, the Newey and West’s (1987) heteroscedasticity and autocorrelation consistent covariance estimates are not used.

Individual t-statistics of the factor coefficients are used to see whether the coefficients on the variables are significantly distinguishable from zero. I evaluate the models on the significance of the factor coefficients, at a 5% significance level. However, the results indicate with *, ** or *** whether a coefficient is significant at the 10%, 5% or 1% level respectively.

(35)

35

4. Data

Chapter 4 holds three sections. Section 1 contains the sample selection procedure and the sample composition. The second section provides descriptive statistics and the final section provides a first graphical indication of differences in tax avoidance between public years and private years.

4.1. Sample selection

Since I investigate both the U.S. market as well as the E.U. market, I cannot follow a similar approach as Katz (2009) and Badertscher et al. (2009). They use the database Datastream for investigating the U.S. market. Orbis is the only database, which I have access to, that is in the possession of financial data of European firms that are not listed. However, the detailed financial statements data is restricted to the last 10 years. As a result, I select the firm-year observations of firms that are delisted in the period 2001-2010.

For several firms is additional (less detailed) data available for the years 1998 and later, this extra data is needed since I require at least data availability of two years before and two years after the year of delisting.

Hence, I select all firm-year observations in Orbis in any of the years from 1998 through 2012 that satisfy the following criteria: (1) the firm is located in the U.S. or in the countries that are part of the E.U. (2) the firm is linked to private equity shareholders, (3) the firm is not a subsidiary of another firm, (4) the firm is a formerly publicly listed company (the firm is delisted), (5) data is available at least two years before (public years) and after (private years) the year of delisting, (6) the firm is not a financial institution or in a regulated industry (SIC codes 6000-6999 and 4800-4900).

I investigated every delisting whether this event was the result of a private equity leveraged buy-out, therefore I used SEC announcements and the database Zephyr. Firms that delisted, for instance in 2007, and that were acquired in 2010 by private equity firms were not selected for the final sample because the delisting was not the result of a private equity LBO.

(36)

36

Table 1 reports the sample composition. My sample consists of 19 firms that are located in the U.S. and 21 firms that are located in the E.U. These firms have both private and public ownership phases. As can be seen in Table 1, most of the firms have seven firm-year observations (three years before delisting, year of delisting and three years after delisting). The firms that have only six firm-year observations are delisted in 2010, therefore only two firm-year observations after the year of delisting are available.

TABLE 1 Sample Composition

The total sample consists of 40 firms that are delisted in the period 2001-2010. This panel presents the number of firms that are located in each country that is represented in the sample. Moreover, the number of firm-year observations per country is exhibited.

Country Code No. of firms Firm-year observations

Bulgaria BG 1 7 Denmark DK 1 7 France FR 4 28 Germany DE 2 14 Great Britain GB 8 54 Italy IT 2 14 Spain ES 1 7 Sweden SE 2 14 E.U. 21 145 U.S. 19 131 Total 40 271 4.2. Descriptive statistics

Table 2 presents the descriptive statistics of firm characteristics, tax avoidance proxies and control variables for the U.S. market. These statistics are presented separately for private firm-years and public firm-years. Consistent with prior literature (Badertscher et al., 2009), I find significant (at 5% or 1% level) differences between means in almost all variables.

(37)

37

with prior literature, the difference in the means of SALESGR is almost zero and therefore not significant.

I document that private firm-years are more likely to have current year losses (LOSS) and are more likely to report net operating loss carry forwards (NOL). Consistent with the association of private equity, the leverage ratio of private firms is significantly higher for private firm-years than of public firm-years. On average, it is approximately three times as high (0.697 vs. 0.242).

(38)

38

TABLE 2

Descriptive statistics for private vs. public years of PE-backed firms in the U.S.

The total sample consists of 271 year observations of 40 firms that are delisted in the period 2001-2010. The number of firms that are located in the U.S. is 19 (131 firm-year observations) (see Table 1). This table presents characteristics of the tax avoidance proxies: book-tax differences (BTD) and effective tax rate (ETR) for the U.S. firms. Furthermore, characteristics of control variables: leverage ratio (LEV), indicator of current year loss (LOSS), indicator of net operating loss carry forwards (NOL), intangibles (INTANG), sales growth (SALESGR), abnormal total accruals (ABACCR) and the natural logarithm of assets (ASSETS) are presented. The number of firm-year observations and means are presented for both private firm-years as for public firm-years. The difference in means (medians) is calculated as the public mean (median) minus private mean (median).

Private Public Difference between

# Obs. Mean Median Std. Dev. # Obs. Mean Median Std. Dev. Means Medians

BTD 54 -0.036 -0.003 0.101 57 0.018 0.006 0.066 0.054 *** 0.009 *** ETR 23 0.283 0.252 0.223 54 0.372 0.361 0.334 0.089 0.109 ** LEV 55 0.697 0.532 0.445 56 0.242 0.239 0.175 -0.455 *** -0.293 *** LOSS 55 0.582 1.000 0.498 57 0.053 0.000 0.225 -0.529 *** -1.000 *** NOL 53 0.943 1.000 0.233 56 0.750 1.000 0.437 -0.193 ** 0.000 * INTANG 54 0.580 0.599 0.650 55 0.342 0.353 0.276 -0.238 ** -0.246 ** SALESGR 55 0.108 0.050 0.617 57 0.116 0.079 0.171 0.008 0.029 ** ABACCR 55 -0.045 -0.025 0.113 57 0.007 -0.004 0.084 0.051 ** -0.029 ** ASSETS 55 15.299 15.225 1.192 56 14.675 14.482 1.185 -3.985 ** -0.743 ***

Corporate Tax Avoidance of Private Equity-backed Firms in the United States and the European Union.