Tax Competition Greenfield Investment versus Mergers and Acquisitions

Tax Competition Greenfield Investment versus Mergers and Aquisitions JOHANNES BECKER CLEMENS FUEST CESIFO WORKING PAPER NO. 2247 CATEGORY 1: PUBLIC FINANCE MARCH 2008 An eletroni version of the paper may be downloaded from the SSRN website: www.ssrn.om from the RePE website: www.repe.org from the CESifo website: Twww.CESifo-group.org/wpT

CESifo Working Paper No. 2247 Tax Competition Greenfield Investment versus Mergers and Aquisitions Abstrat In this paper, we analyze tax ompetition in a model where investor firms have the hoie between two types of investment, greenfield investment and mergers and aquisitions. We show that the oexistene of these two types of investment intensifies tax ompetition in omparison to the ase where there is only greenfield investment. If a speifi tax on aquisitions is available, this result hanges. Then, tax ompetition is mitigated ompared to the pure greenfield ase. The existene of an aquisition tax may even lead to orporate overtaxation. JEL Code: H25, F23. Keywords: orporate taxation, mergers and aquisitions, tax ompetition. Johannes Beker Cologne Center for Publi Eonomis University of Cologne Albertus-Magnus-Platz 50923 Cologne Germany johannes.beker@uni-koeln.de Clemens Fuest Cologne Center for Publi Eonomis University of Cologne Albertus-Magnus-Platz 50923 Cologne Germany lemens.fuest@uni-koeln.de This version: 27th February 2008 We thank Steve Bond, Harry Huizinga, Peter Neary and Andreas Wagener as well as partiipants at researh workshops in Oxford, Tilburg and Vienna for very helpful omments. The usual dislaimer applies. We gratefully aknowledge finanial support from the Deutshe Forshungsgemeinshaft (DFG), Grant No. FU 442/3-1.

1 Introdution The inreasing mobility of apital and the growing importane of multinational rms have given rise to an intensive politial and aademi debate on tax ompetition and the e ets of taxes on ross-border apital ows. A large theoretial and empirial literature has emerged whih has signi antly improved our understanding of this issue. A harateristi of this literature is that it fouses almost entirely on green eld investment. Building a new plant, however, is not the only way to realize an investment projet. As an alternative, the investor may purhase an existing rm. Empirially, mergers and aquisitions (m&a) play an important role. Figure 1 displays the total volume of all m&a transations worldwide and in Europe (left ordinate) over time. Mergers and aquisitions ome in waves with a peak of more than three trillion US dollars in 2000 and falling to only one trillion in 2002. The ordinate on the right depits the fration of national m&a, i.e. transations where aquirer and vendor are within the same borders, and the sum of national and intraregional m&a. In ontrast to the high volatility in volumes, these frations stay virtually onstant over time. What is the di erene between green eld investment and mergers and aquisitions from a tax poliy perspetive? In standard models of tax ompetition 1, investment is modelled as a redistribution of net savings aross ountries. However, as reently pointed out by Desai and Hines (2004), m&a do not imply a reloation of orporate apital but rather a hange in ownership and ontrol rights. Clearly, rms whih onsider a new investment projet often have the hoie between di erent types of investment, inluding the aquisition of an existing rm or a green eld investment. It is the purpose of this paper to analyse the impliations of this hoie for the welfare e ets of tax ompetition. How would we expet mergers and aquisitions to a et tax ompetition? Intuitively, one ould argue that aquisitions are less tax sensitive than green eld investment beause taxes are likely to be apitalized in the purhase prie of immobile assets. This might suggest that the existene of m&a investment mitigates tax ompetition. 1 Standard referenes are Musgrave (1969) Feldstein and Hartman (1979), Wilson (1986), Zodrow and Mieszkowski (1986), Bond and Samuelson (1989), Buovetsky and Wilson (1991). 1

In this paper, we develop a simple theoretial framework whih allows to explore how the oexistene of m&a and green eld investment a ets orporate tax ompetition. We assume that investor rms onsidering a new projet rstly sreen the market for existing rms whih are suitable as aquisition targets. If they do not nd an adequate existing rm, they build a new plant (green eld investment). We thus onsider a setting where green eld and m&a investments are substitutes. In suh a framework, taxes may distort both the deision on the overall number of projets and the hoie to realize these projets as green eld investments or on the basis of aquisitions. 4,000 1 3,500 National + intraregional m&a 0.9 0.8 3,000 M&a volume in billion US dollars 2,500 2,000 1,500 1,000 500 World Europe National m&a 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Perentage of regional distribution 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 0 Figure 1: M&a volume worldwide (data soure: Thompson Finanial) Our results do not on rm the view that the existene of mergers and aquisitions investment mitigates tax ompetition. To the ontrary, in the baseline version of our model, we show that tax ompetition is intensi ed. The reason is that, due to the existene of m&a investment, green eld investment beomes more tax sensitive. If a ountry inreases its taxes, it does not only lose marginal green eld investment projets, but intra-marginal green eld projets are replaed 2

by aquisitions of existing rms. This redues the number of new projets in the ountry and, as a onsequene, total tax revenue. Put di erently, the introdution of m&a into the standard tax ompetition model generates a seond sal externality whih points in the same diretion as the one known from the standard model. An interesting impliation of this result is that high-tax ountries are predited to have more m&a projets than low-tax ountries. But this is not to their advantage beause, at the margin, green eld projets generate more tax revenue than m&a projets. Things are di erent, though, if one takes into aount that tax poliy may disriminate between m&a and green eld projets. In this ase, orporate tax ompetition is mitigated due to the existene of aquisition taxes. We demonstrate that there may even be equilibria where orporate taxes beome too high in the sense that a oordinated inrease of orporate tax rates redues welfare. Furthermore, there is a potential for welfare enhaning oordination of the tax treatment of aquisitions. This also sheds light on attempts by the European Union to oordinate the tax treatment of ross-border m&a. 2 In this paper, we fous on domesti m&a transations in the presene of an international market for portfolio apital. We do so in order to relate our analysis as losely as possible to the standard literature on (harmful) tax ompetition and the underprovision of publi goods, see the literature ited in footnote 1. Aounting for international transations, i.e. ross-border green eld and m&a projets, ompliates the analysis by raising issues like repatriation tax shemes, foreign rm ownership e ets, ompetition between domesti and foreign investors et. These are all important aspets of international taxation, but inluding them would ompliate our analysis without hanging the main insights. We disuss some of these issues in the extensions setion (3.2). Moreover, although rossborder transations are more often debated, the bulk of transations still takes plae within national borders, as gure 1 shows. Our analysis demonstrates that suh transations are important for international tax ompetition, as well. In the publi nane literature, Devereux (1990) is one of the rst to shift 2 EC level oordination is mainly onerned about disrimination of border rossing relative to national transations whereas our argument for oordination also applies to purely national transations. 3

the fous from apital to ownership alloation. He does not refer expliitly to mergers and aquisitons but points out that tax distortions to ownership may be important if apital produtivity depends on ownership. The paper introdues the onept of apital ownership neutrality as a property of international tax systems whih avoid distortions in ownership. Gordon and Bovenberg (1996) also onsider tax poliy in a model where investors may aquire existing rms. But they onentrate on problems of asymmetri information, and a hange in ownership does not a et the produtivity of existing rms. Fuest and Huber (2004) analyze tax poliy in a model where rms may be sold to foreign investors, but they fous on the integration of personal and orporate inome taxes and do not disuss tax ompetition. Moreover, Desai and Hines (2004), as mentioned above, argue that apital ows in the form of M&A are likely to have impliations for tax poliy whih di er from the impliations of the standard apital mobility model. Their main point is that US taxation of foreign soure inome is likely to distort ownership patterns and to put US rms at a disadvantage when ompeting for foreign aquisitions. They propose to exempt foreign soure inome from domesti taxation. 3 In Beker and Fuest (2007b), we analyze this argument and show that exemption is an appropriate poliy hoie when ownership advantage is a publi good within the rm, but is dominated in welfare terms by a ross-border ash- ow tax system. Hau er and Shulte (2007) onsider tax inentives in a model where mergers and aquisitions an take plae within and aross borders. They show that ownership patterns are highly important for the welfare impliations of tax poliy hoies. There is also a growing literature dealing with the impat of globalisation on mergers and aquisitions, see e.g. Neary (2007). This literature analyses mergers of rms operating in imperfetly ompetitive markets. In this paper, we deliberately abstrat from imperfet ompetition and the question of how m&a investment a ets market strutures and trade patterns, mainly beause we want to keep our approah as lose as possible to standard models of tax ompetition. Nevertheless, we will disuss this issue further in the extensions setion (3.3). Apart from this, the present paper is related to two strands of literature. 3 See also Desai and Hines (2003) and the debate between Grubert (2005) and Desai and Hines (2005). 4

Firstly, there are some reent theoretial papers on merger poliy, e.g. Hau er and Nielsen (forthoming), as well as on M&A and trade poliy, e.g. Huk and Konrad (2004). Empirial evidene on M&A is reported in Andrade, Mithell and Sta ord (2001). A seond strand of literature deals with apital mobility and tax ompetition. 4 There is a broad empirial literature on the impat of taxes on investment and apital ows, whih is surveyed by Hines (1999) and Devereux (2007). However, virtually all studies treat investment ows as if they were green- eld projets. Combining these two literatures raises the question of how taxation a ets M&A ativity. As Auerbah and Slemrod (1997) and Kaplan (1989) suggest, taxes may be of ruial importane for M&A investment. There are some papers disussing the impat of the U.S. tax reform on aquisitions of US rms by foreign investors. Here, the main idea is that the e etive inrease in the tax burden aused by the 1986 tax reform indued investors loated in ountries with foreign tax redit regimes to take over U.S. rms beause the higher US taxes were redited against home ountry taxes (Sholes and Wolfson (1990), Collins, Kemsley and Shakelford (1995)). Swenson (1994) applies the same argument to US inbound foreign diret investment and nds robust evidene supporting the hypothesis. In a reent paper, Huizinga and Voget (2006) study the empirial impat of international taxation shemes on M&A ativity. The remainder of the paper is set up as follows. In setion 2, we present the model and the main results. In setion 3, we onsider some extensions. Setion 4 disusses some poliy impliations and onludes. 2 The model In this setion, we desribe the setup of the model and derive a benhmark result whih is based on green eld investment. Then, we introdue the opportunity for investor rms to aquire existing rms. 4 For a reent survey see e.g. Fuest, Huber and Mintz (2005). 5

2.1 The setup The world onsists of n idential open eonomies. Eah ountry is populated by a representative household whih lives for two periods. The utility funtion of the representative domesti household of ountry i is given by U (C 1 ; C 2 ; G) = u(c 1 )+ C 2 + h(g), 5 where C 1 and C 2 are onsumption levels in the rst and the seond period and G is a publi onsumption good provided by the government in period 2. For notational onveniene, we omit the ountry index unless misunderstandings may arise. The funtions u(c 1 ) and h(g) are stritly onave, with u 0 > 0, u 00 < 0 and h 0 > 0, h 00 0. In period 1, the household has an endowment of E units of a numeraire good. This numeraire good may be transformed into the private onsumption good and the publi onsumption good on a one to one basis. Households may borrow or lend in the international apital market at the interest rate r. There are no residene based taxes on apital inome. Households are also endowed with m existing, immobile rms. We refer to these rms as target rms, as opposed to investor rms whih will be introdued below. Target rms are endowed with immobile apital goods from investment in previous periods. They do not onsider new investment opportunities, but they may be sold to investor rms. If a target rm is not sold to an investor rm, it yields an after tax pro t (1 ) in period 2, where is the orporate tax rate. Thus, under their initial owners, all target rms are assumed to yield the same pro ts. However, we assume that they di er in their suitability as aquisition targets. This will be explained in greater detail below. Next to the target rms, there is a large number of investor rms. For notational onveniene, we normalize their number to unity. The representative investor rm is also owned by the domesti household. 6 The investor rm onsiders a set of investment projets in its ountry of residene. In the seond period, eah projet j yields a projet-spei pre-tax return denoted by j. is assumed to be uniformly distributed over the interval [ ; + ]. The distribution funtion is 5 We use this quasilinear utility funtion beause it eliminates inome e ets on savings whih would ompliate the analysis without adding further insights. 6 Thus, there is no ross-border investment in the strit sense. However, in the absene of repatriation taxes, the results derived in this model arry over to the ase of ross-border investment, but get more omplex due to the inentive to tax pro ts aruing to foreign owners, as will be disussed in setion 3. 6

denoted by () : The ost of investment in period 1 annot be deduted from the orporate tax base in period 2. Thus, the after tax ash ow generated by projet j in period 2 is given by j (1 ). These projets may be arried out as green eld investments or as aquisitions, whih means that the level of j does not depend on the type of transation. The ost of investment may di er, though. If a projet is arried out as a green eld investment, it requires the investment of one unit of the numeraire good in period 1. If the projet is arried out on the basis of an aquisition, rather than a green eld investment, the investor rm has to aquire an existing target rm in period 1. We will proeed as follows. As a rst step, we assume that all projets are green eld investment projets. In a seond step, we introdue the opportunity to aquire existing rms and thus allow for green eld and m&a investment to oexist. 2.2 Green eld investment Assume that all projets are arried out as green eld investments. In this ase, the investor rm will arry out all investment projets whose return exeeds a ritial value. The rm will hoose this ritial value so as to maximize its market value V gf, whih is given by " (1 + r) V gf + Z + # d = Z + (1 ) d (1) The supersript gf denotes the pure green eld ase. Maximizing V gf over the uto value yields the result that, not surprisingly, investment is dereasing in the interest rate and the orporate tax rate: = 1 + r 1 (2) In the rst period, the household nanes green eld investment of the domesti investor rm. In addition, the household may borrow (S > 0) or lend (S < 0) in the international redit market. The household s budget onstraint is C gf 1 = E S Z + d (3) 7

In the seond period, the household reeives inome from savings, pro t inome from ongoing rms and pro t inome from the investor rm. The budget onstraint is given by Z + C gf 2 = S(1 + r) + m (1 ) + (1 ) d (4) Optimal savings of the domesti households imply u 0 (C 1 ) = 1 + r (5) The budget onstraint of the government is " G gf = m + Z + # d (6) Consider next the determination of the interest rate in the international apital market. Capital market equilibrium implies nx S i = 0 (7) i=1 Equations (5) and (7) determine S i, 8i = 1; :::; n, and r, for given values of i. Straightforward omparative stati analysis yields dr d = @d 1 < 0 (8) d @ d where = P h i 1 1 < 0, i.e. an inrease in the tax rate in ountry u i00 (C1 i) (1 i ) d leads to a deline in the interest rate. 7 2.3 Tax ompetition and tax oordination with green eld investment Under tax ompetition, the domesti government maximizes domesti welfare W = u(c 1 ) + C 2 + h(g) subjet to the onstraints in (3)-(6) and takes the tax poliy of 7 Note that if n! 1, then! 1 and dr d i! 0. 8

the other ountries as given. The rst order ondition for the optimal tax poliy of the domesti ountry an be written as " Z # @W + @ = (h0 1) m + d where ^ = arg max W and @ h 0^ @ + @W @ = 0 (9) @W = S @ h0^ (10) with @ = 1. The government faes a trade-o between raising tax revenue 1 for publi goods provision and distorting investment. In a symmetri equilibrium, with S = 0, an underprovision of publi goods relative to a rst best equilibrium ours, i.e. h 0 > 1. 8 How does a simultaneous hange in all orporate tax rates, departing from the equilibrium without oordination, a et global welfare? The hange in welfare of an individual ountry d an be formulated as dw d = @W d Xn 1 d^ d + @ d i=1 @W d @ i d^ i : (11) The optimal tax poliy under tax ompetition implies @W d @ d = 0. A hange in the tax rate of other ountries, though, does a et welfare in ountry d beause it a ets the interest rate in the world apital market. Using (8), (10) and the symmetry property of the equilibrium under tax ompetition, the overall welfare e et an be written as dw d = @W d Xn 1 i=1 i d^ i = h 0 ^ n 1 @ i 1 ^ ( ) 2 n u 00 d^ > 0 (12) u 00 (1 ^) whih implies that a oordinated inrease in the orporate tax rate inreases 8 The rst order ondition for the optimal tax rate an be written as @W @ = [h 0 1] hm + R i + @ d h0^ @ [1 + @ @ ] = 0. Using the expressions derived above, this simpli es to @W @ = (h0 1) hm + R i h i + u d 1 00 @ n[u 00 (1 )] h0^ @ = 0, so that h 0 > 1: 9

welfare. Proposition 1 A oordinated inrease in all orporate tax rates, departing from the equilibrium under tax ompetition, inreases welfare. This result is well known from the literature on tax ompetition with green eld investment. The undertaxation result ours beause orporate tax uts give rise to negative sal externalities on other ountries. 9 It serves as a benhmark for the analysis of tax ompetition in the presene of mergers and aquisitions in the following setion. 2.4 Adding mergers and aquisitions We now allow rms to hoose between aquisitions and green eld investment as possible ways to realize their projet. An aquisition is a substitute for a green- eld investment, but we assume that it is an imperfet substitute. Existing rms with ongoing prodution di er in their suitability as target rms. We model this as follows. If an investor rm deides to arry out a projet on the basis of an aquisition of target rm g, rather than a green eld investment, there is an output loss in period 2 denoted by k g. The variable k g is assumed to be uniformly distributed over the interval [0; k + ]. The distribution funtion is denoted by (k). The underlying idea is that green eld investment allows the investor rm to set up its fatory and hoose a labour fore exatly as it suits its interests whereas existing rms will not exatly math the investor s needs. 10 What are the tax impliations of an aquisition? We assume that the proeeds from selling a rm are untaxed and that the aquiring rm annot write o the purhase prie. This omes lose to the usual tax treatment of a share deal, as opposed to an asset deal. We will disuss the robustness of our results with respet to this assumption in the extensions setion (3.1). In addition, there is a disriminatory tax on aquisitions whih allows ountries to tax green eld investment and aquisitions investment di erently. In real world tax systems, suh 9 The onept of sal externalities is explained in detail in Buovetsky and Wilson (1991). 10 Of ourse, it may also our that existing rms have unique assets whih make them more suitable than a newly reated rm. It would be straightforward to inlude this ase by allowing for a negative k. 10

a disrimination an be ahieved by designing rules for the transfer of reserves, inter-ompany dividends, the depreiation of goodwill et. We summarize this in a tax on aquisitions denoted by. The aquisition prie is determined as follows. Sine some rms are more suitable as aquisitions targets than others, investor rms will also be willing to pay a higher prie for them. However, the prie a vendor may harge is limited by the fat that the investor rm may always hoose a green eld investment. In equilibrium, aquisition pries will be suh that the representative investor rm is indi erent between the two options. This is the ase if the prie of rm g, P (k g ); satis es: ( j k g ) (1 ) 1 + r whih an be rearranged to P (k g ) = j (1 ) 1 + r 1 (13) P (k g ) = 1 k g (1 ) + 1 + r (14) The initial owner of target rm g will sell the rm if the prie is at least as high as the present value of the pro t the rm will make if it is not sold. 11 requires P (k g ) (1 ) 1 + r This 0 (15) It follows that the initial owners of all target rms haraterized by a k g satisfying k g k will sell their rms, where k is given by k = 1 + r 1 (16) k thus haraterizes the marginal aquisition, where the vendor is just indifferent between selling and not selling the rm. Aquisitions will only our if k > 0. In the following, we will fous on equilibria where some aquisitions take plae. 12 Note further that the tax system is neutral with respet to the number 11 The highest possible aquisition prie is P (0) = 1 (if = 0). This implies that the original investment ost to reate the target rm, net of output generated in previous periods, must have been lower than 1. 12 An equilibrium where no green eld investment ours is also possible, but will be negleted in the following. 11

of aquisitions if = (1 + r), see (16). In the absene of an aquisitions tax, the orporate tax distorts the hoie between aquisitions and green eld investment in favour of aquisitions, i.e. k is higher than in the absene of orporate taxes. Figure 2 illustrates the model. The rst margin, whih determines the overall number of investment projets, is de ned by = 1+r. For a given r, neither 1 the initial pro t level nor the aquisition tax will a et the total number of investment projets arried out in the ountry under onsideration. Note, though, that hanges in all these parameters may a et the equilibrium interest rate r. The seond margin is de ned by (16) and determines the number of projets realized on the basis of aquisitions, whih is given by R k dk. k θ k + π + 1 τ 1+ r 1 τ m&a seond margin greenfield projets first margin number of projets Figure 2: Two deision margins. Not surprisingly, an inrease in the aquisition tax redues the number of aquisitions. In ontrast, an inrease in the orporate tax rate.p. leads to an inrease in the number of aquisitions: @k @ = 1 + r (1 ) 2 > 0 (17) The reason is that the higher orporate tax is apitalized in the purhase prie for existing (immobile) rms whereas the prie of new apital does not hange 12

(given r). Thus, aquisitions beome more attrative relative to green eld investment. An inrease in the interest rate also leads to more aquisitions and less green eld investment. The reason is that a higher interest rate means that new apital beomes more expensive. This inreases the inentives to use old apital. The value of the rm is now given by = (1 + r) Z + " V + Z + (1 ) d d + Z k 0 Z k 0 (P (k) 1) dk # ( + (1 ) k) dk (18) In period 1, expenditure for green eld investment is equal to the overall number of projets whih are arried out, R + d. For eah aquisition, the rm has to pay the aquisition prie P (k), but it an redue its expenditure on new apital by one unit. In period 2, the ash ow is redued by the tax on aquisitions and the after-tax output loss (1 ) k. Competition between investors drives up the pries of target rms so that investor rms are indi erent between aquisitions and green eld investment. As a result, the equilibrium value of the investor rm does not depend on the mix between green eld investment and aquisitions. Using P (k g ) = 1+r kg (1 ) 1+r, the rm value equation boils down to (1 + r) " V + Z + d # = Z + (1 ) d: (19) This re ets that the surplus reated by using existing rms rather than new apital fully arues to the initial owners of the target rms. The maximization of V over the total number of projets yields = 1 + r 1 (20) Thus, the marginal projet is a green eld projet in the sense that the overall number of investment projets realized in the ountry under onsideration is determined by the ost of green eld investment. The mix between green eld projets 13

and aquisitions depends on the availability of suitable target rms. More formally, the number of aquisitions is determined by (16). The remaining projets are realized as green eld investments. Consider next the budget onstraints of the private household and the government. The budget onstraint of the domesti household in period 1 an be written as C 1 = E S Z + d Z k 0 dk! (21) Compared to the pure green eld ase, the household an redue the expenditure for investment in period 1 by using old rather than new apital, i.e. by inreasing the number of aquisitions. The seond period budget onstraint is given by C 2 = S(1 + r) + m Z k Z + + (1 ) d 0 dk (1 ) (22) Z k 0 ( + (1 ) k) dk Here, the the existene of aquisitions a ets onsumption opportunities as follows. The seond term on the right hand side of (22) re ets that the household s inome from ongoing rms is now smaller beause some of them have been aquired by the investor rm. The third and the fourth terms represent the pro ts from new investment (based either on aquisitions or green eld projets) net of aquisition taxes and the output losses due to k. The publi setor budget onstraint now beomes G = " m Z k 0 Z # + Z k dk + d + ( k) dk (23) 0 Eah additional aquisition inreases tax revenue by the aquisition tax and redues it through the output loss k and by dereasing the number of ongoing rms, so that tax revenue delines by ( + k ), at the margin. 14

2.5 Capital market equilibrium Given the funtions k i = k i (r; i ; i ) and i = i(r; i ), i = 1:::n, implied by (16) and (20), the apital market equilibrium is determined by the rst order onditions for optimal savings u i0 = 1 + r, i = 1:::n, and the redit market equilibrium ondition P i Si = 0. These n + 1 equations determine optimal savings S i, i = 1:::n, and the interest rate r, for given values of the tax instruments i and i, i = 1:::n. Thus, the interest rate in the international apital market an be expressed as a funtion r = r( 1 ::: n ; 1 ::: n ). Standard omparative stati analysis yields @k = i + @ i 1 < 0 @ i @ i @ i (24) = @k 1 > 0 @ i @ i (25) where = P n 1 2 i=1 u i00 (1 i < 0. Equation (24) shows that, as expeted, an inrease in the tax rate i redues the interest rate. An inrease in i, in ontrast, inreases the interest rate beause it redues the number of aquisitions while the overall number of projets arried out in ountry i remains onstant. As a result, apital demand for green eld investment in ountry i inreases, and this drives up the interest rate. 2.6 Tax ompetition Again, we assume that ountries set their tax poliy to maximize the welfare of the representative domesti household and take the tax poliy of the other ountries as given. In the presene of aquisitions, the rst order ondition for the optimal orporate tax rate is given by @W @ = (h 0 1) " m h 0 ^ ( + k ) Z k k Z + ( + k) dk + @k @ ^ + ^ @ @ d + @W # @ (26) 15

where ^ = arg max W and ^ = arg max W, as de ned below. How does the oexistene of green eld investment and aquisitions a et the optimal tax poliy? An inrease in the orporate tax raises revenue, as re eted by the rst term on the right hand side of (26), and hanges the orporate tax base, as the seond term indiates. In ontrast to the ase of pure green eld investment, there is a seond margin whih a ets the tax base. An inrease in the orporate tax indues rms to replae green eld investment by aquisitions. For a given interest rate, the e et on tax revenue is equal to ^ ( + k @k ) ^. Finally, tax poliy a ets the interest rate. This is aptured by the term @W @W = S h0 ^ ( + k ) ^ @k @ @ with @ + ^ whih an be simpli ed to @W = S h0 2^(1+r) ^ (1 ^) 2. What is the di erene between this expression and the one from the pure green eld ase? Next to the e et on (27) interest inome (re eted by S), a rise in r lowers the total number of projets realized in the domesti ountry. This is re eted by an inrease in the uto level. In addition, the number of aquisitions inreases (k rises), whih a ets tax revenue as disussed above. How are taxes on aquisitions set in a tax ompetition equilibrium? The rst order ondition for the optimal tax on aquisitions is given by: Z @W k @k @ = (h0 1) dk h ^ 0 ( + k ) ^ 0 @ ^ + @W Using (27) and k = 1+r, equation (28) an be rewritten as 1 Z @W k @ = (h0 1) dk 0 h 0 (1 ) 2 (^ The optimal level of is given by: ^ (1 + r)) 1 (1 ^) ^ 2 h 0 Z 1 k = dk + ^ (1 + r) 1 @ h 0 0 @ = 0 (28) + ^ (1 + r) = 0: @ @ (29)! 1 2 @ 1 @ (30) 16

Sine dr < 0; 5 for all n and d u00, the right hand side of (30) is unambiguously positive. It thus turns out that the aquisitions tax whih emerges under tax ompetition is positive. However, it is ambiguous whether or not the tax system as a whole disriminates aquisitions relative to green eld investment (as mentioned above, neutrality requires ^ = ^(1 + r)). 13 2.7 Tax Coordination Is there any sope for welfare enhaning tax oordination? Consider rst a oordinated hange in, departing from the equilibrium under tax ompetition and holding onstant the aquisition tax. The e et on the welfare of the ountry under onsideration is given by dw d = @W d Xn 1 d^ d + @ d i=1 @W d @ i d^ i (31) Given that @W d @ d = 0 holds in the equilibrium under tax ompetition, and using the symmetry property S i = 0 8i, the welfare e et an be expressed as dw d = @W d Xn 1 i=1 d^ i = h ^ 0 ( + k ) ^ @k @ Xn 1 + ^ @ i i=1 @ i d^ i (32) The rst term in the parentheses on the right hand side of (32) reveals that the existene of mergers and aquisitions gives rise to an additional sal externality of orporate tax uts. A orporate tax ut in other ountries inreases the interest rate. This leads to an inrease in aquisitions ( @k > 0) or, more preisely, to a substitution of green eld investment by aquisitions. The inrease in aquisitions may inrease or derease tax revenue, depending on whether ( + k ) positive or negative. If = 0, (16) implies that tax revenue delines as green eld investment is replaed by aquisitions. In this ase, a negative sal externality arises, whih reinfores the standard externality due to the deline in the overall number of projets. The latter is aptured by the seond term in the parentheses 13 It is straightforward to show that ^ > ^(1 + r) emerges if the number of ountries n is large, whih implies that @ onverges to zero. is 17

in (32). This implies that the possibility of replaing green eld investment by aquisitions (and vie versa) unambiguously intensi es tax ompetition if there is no aquisitions tax. In ontrast, if there is suh a tax, the situation is di erent beause the sign of the sal externality aused by the existene of aquisitions beomes ambiguous. Equation (32) an be rearranged to dw d = h 0 (^ 2^ (1 + r)) (1 ) 2 Xn 1 i=1 @ i d^ i (33) Given (24), it is easy to show that this expression may in general be positive or negative. We summarize this in Proposition 2 In the absene of a tax on aquisitions, the possibility of replaing aquisitions by green eld investment gives rise to an additional negative sal externality of orporate tax uts. Tax ompetition is intensi ed. A oordinated inrease of the orporate tax inreases welfare. Proposition 3 In the presene of an aquisitions tax, a oordinated inrease of orporate tax rates is welfare enhaning if ^ < 2^ (1 + r) and redues welfare if ^ > 2^ (1 + r). If ^ = 2^ (1 + r), the di erent sal externalities ompensate eah other and oordination does not a et welfare. Proposition 3 implies that the question of whether orporate tax rates are too high or too low under tax ompetition depends on the level of the aquisition tax. As pointed out in the preeding setion, ^ is unambiguously positive in the tax ompetition equilibrium, but whether it exeeds 2^ (1 + r) is, in general, ambiguous. Given that the standard model with only green eld investment unambiguously leads to undertaxation, the question arises whether a negative welfare e et of a oordinated orporate inome tax inrease is possible. The appendix provides an example showing that parameter ranges exist where ^ 2^ (1 + r) > 0 holds in the equilibrium under tax ompetition. We may thus state: Proposition 4 The opportunity to levy a tax on aquisitions mitigates tax ompetition. In the presene of an aquisitions tax, a oordinated inrease in orporate tax rates may redue welfare. 18

The result in proposition 4 shows that taking into aount the existene of m&a investment in the analysis of tax ompetition is important beause one of the benhmark results in the theory of orporate tax ompetition - the nding that tax ompetition leads to an undertaxation of orporate pro ts, is alled into question. The eonomi explanation for this result is the following. A tax ut in ountry i drives up the interest rate r. This will redue the level of green eld investment in all other ountries. Sine the taxes on the marginal green eld investment are positive, tax revenue and, hene, welfare in these ountries delines. But at the same time, the higher interest rate leads to an inrease in the number of aquisitions. If the tax on aquisitons is su iently high, this has a positive impat on overall tax revenue. The seond sal externality may dominate the rst, so that the sal externalities of orporate tax uts may in fat be positive in our model. Is there any sope for welfare enhaning tax oordination of aquisition taxes? The welfare e et of a oordinated tax hange, departing from the equilibrium under tax ompetition and holding onstant the orporate tax, is given by dw d = @W Xn 1 i=1 Using (27) and (25), this an be expressed as: dw d = h 0 (^ We may therefore state 2^ (1 + r)) (1 ^) 2 n 1 n @ i d^ i (34) u 00 2u 00 (1 ^) d^ (35) Proposition 5 Departing from the equilibrium under tax ompetition, a oordinated redution in the tax on aquisitions inreases (redues) welfare if ^ < 2^ (1 + r) (^ > 2^ (1 + r)). The welfare e ets of tax oordination of both and depend on whether ^ is larger or smaller than 2^ (1 + r). This is not surprising beause in our model sal externalities are transmitted through the interest rate in the international apital market. If ^ < 2^ (1 + r), tax hanges whih drive up the interest rate give rise to negative sal externalities and vie versa. 19

3 Extensions In this setion, we onsider three extensions of the above presented model. In subsetion 3.1, we show that our results are robust to modi ations in the tax treatment of aquisitions. Subsetion 3.2 disusses the impliations of ross-border aquisitions. In subsetion 3.3, we provide a brief disussion of imperfet ompetition and its importane for the analysis of mergers and aquisitions. 3.1 Taxation of apital gains and dedutibility of aquisition expenditures An important but ertainly restritive assumption we have made is that the revenue from selling the rm, whih arues to the initial owners, is not subjet to tax, and the investor rm annot dedut the purhase prie. The tax onsequenes of aquisitions are important in our model beause it is relevant for a key e et whih drives our results: the nding that a higher orporate inome tax indues rms to replae green eld investment by aquisitions. The question is how robust this result is. An alternative approah would be to assume that the vendor has to pay tax on the revenue from selling the rm while the aquiring rm may dedut the purhase prie. This would be a simple way of modelling the usual tax treatment of an asset deal, as opposed to a share deal. In this ase, the investor rm will be indi erent between aquiring any rm g and making a green eld investment if ( j k g ) (1 ) 1 + r The initial owners will sell their rm if P (k g ) (1 ) = j (1 ) 1 + r 1: (36) P (k g ) (1 ) (1 ) 1 + r 0 (37) This implies that all target rms haraterized by a level of k satisfying k k will sell their rms, where k is given by k = 1 + r 1 (38) 20

whih is idential to the expression in equation (16). The reason is that, ompared to the baseline version of our model, the tax disadvantage of the vendor is exatly equivalent to the tax advantage of the buyer. It is straightforward to show that our results are robust with respet to di erent ways of treating aquisitions for tax purposes, provided that the vendor and the seller are treated symmetrially. Situations where this is not the ase are aptured by our parameter. 3.2 Cross-border m&a investment So far, our analysis has been restrited to national aquisitions. What happens if we allow for ross-border aquisitions in our model? The simplest way of introduing ross-border aquisitions is to assume that the investor rms from one ountry also untertake green eld investments and aquisitions in other ountries. Sine the loation of the headquarter of the investor rms does not play any eonomi role in our model, this would be equivalent to assuming that investor rms whih untertake projets in ountry i are owned by residents of some other ountry i. This would hange our results only in so far that the desire to tax pro ts aruing to foreign residents would be added as an additional motive to tax orporate pro ts. It is well known that this may lead to orporate overtaxation under tax ompetition (Huizinga and Nielsen (1997)). 14 Of ourse, ross-border aquisitions would also raise issues like double taxation agreements or pro t shifting. We are on- dent that adding these elements to the model would not hange the key insights provided by the analysis, but a thorough analysis of these issues may nonetheless be valuable. We leave these questions for future researh. 3.3 Imperfet Competition among Firms Another limitation of our analysis is that we abstrat from what is widely seen as an important fator for mergers and aquisitions: the existene of imperfet ompetition among rms. We have deliberately done so in order to keep our model as lose as possible to the standard model of (harmful) tax ompetition. An alternative approah to analyse the role of m&a investment for tax ompetition would 14 In an earlier version of this paper, we onsidered a framework with border rossing aquisitions. The results of the analysis are available from the authors on request. 21

be to start with a model of oligopolisti ompetition and add intergovernmental sal ompetition to it. Several additional issues would arise in this ase. Firstly, m&a investment may hange the number of rms in the market. If one rm whih is ative in the market aquires another ative rm, onsumers may be negatively a eted by inreasing pries. However, if the merger paradox applies, mergers will only arise if they give rise to synergies, whih has further impliations for both onsumers and the s. It is even possible that pries deline, due to lower marginal osts of the newly reated rm. Of ourse, green eld investment may also hange the number of rms in a market. In general, e ets of investment on the intensity of ompetition in markets for private goods are not only an issue for tax poliy but also for ompetition poliy and merger ontrol. Seondly, if the desire to redue ompetition is a fator driving m&a investment, green eld investment would not be onsidered as a substitute. Possibly, a framework without green eld investment would be appropriate. There is no doubt that these issues are worth to be investigated in models of sal ompetition. But doing so would divert attention from the fous of this paper. In addition, the e ets arising in our model are also likely to be relevant in models whih do aount for imperfet ompetition among rms. 4 Disussion and onluding remarks This paper departs from the observation that the literature on international tax ompetition has negleted the role of mergers and aquisitions. It mainly fouses on green eld investment although the former type of investment is empirially at least as important as the latter. We therefore suggest a simple framework whih introdues mergers and aquisitions into a standard tax ompetition model. Investor rms hoose between aquiring existing rms or realizing green eld projets. We show that, if we abstrat from the possibility of levying a spei tax on aquisitions, the introdution of m&a investment intensi es tax ompetition beause it gives rise to an additional negative sal externality of orporate tax uts. Interestingly, an inrease in orporate taxes raises the number of aquisitions in a ountry but redues the total number of investment projets. Therefore, a tax inrease does not only a et the quantity of investment but also its omposition 22

and, hene, its quality in terms of welfare. 15 If an aquisition tax is available, unoordinated poliies lead to a positive tax in our model. Whether the tax system as a whole disriminates aquisitions relative to green eld investment in a tax ompetition equilibrium is ambiguous. If the number of ountries is large, a systemati disrimination of aquisitions relative to green eld investment emerges. The existene of the aquisition tax implies that the sal externalities of orporate tax rate hanges are di erent. If aquisition taxes are su iently high in the unoordinated equilibrium, the sal externality whih arises due to the existene of m&a investment beomes positive. As a result, orporate tax ompetition is mitigated, and it may even be that overtaxation ours. In terms of poliy impliations, our analysis draws attention to the fat that orporate tax oordination whih fouses on the (tax inlusive) ost of apital for green eld investment is inomplete. The tax treatment of aquisitions is an important fator as well. Clearly, it has to be taken into aount that our model only highlights a rather spei aspet of mergers and aquisitions: the possibility of replaing a green eld investment by an aquisition. There are many other fators driving mergers and aquisitions investment, and these fators are likely to be relevant for the workings of tax ompetition as well. This is an agenda for future researh. Referenes Andrade, G., Mithell, M. and Sta ord, E. (2001). New Evidene and Perspetives on Mergers, Journal of Eonomi Perspetives 15(2): 103 120. Auerbah, A. J. and Slemrod, J. (1997). The Eonomi E ets of the Tax Reform At of 1986, Journal of Eonomi Literature 35(June): 12 13. Beker, J. and Fuest, C. (2007a). Quantity versus Quality - The Composition E et of Corporate Taxation on Foreign Diret Investment, CESifo Working Paper No. 2126. 15 The idea that taxation may a et not only the quantity of investment but also its omposition and, hene, its quality, is developed in greater detail in Beker and Fuest (2007a). 23

Beker, J. and Fuest, C. (2007b). Taxing Foreign Pro ts with International Mergers and Aquisitions, Oxford University Center for Business Taxation Working Papers Series No. 07/19. Bond, E. W. and Samuelson, L. (1989). Strategi Behavior and the Rules for International Taxation of Capital, Eonomi Journal 99: 1099 1111. Buovetsky, S. and Wilson, J. (1991). Tax Competition with Two Tax Instruments, Regional Siene and Urban Eonomis 21: 333 350. Collins, J., Kemsley, D. and Shakelford, D. (1995). Tax Reform and Foreign Aquisitions: A Miroanalysis, National Tax Journal 48(1): 1 21. Desai, M. A. and Hines, J. R. (2003). Evaluating International Tax Reform, National Tax Journal 56(3): 487 502. Desai, M. A. and Hines, J. R. (2004). Old Rules and New Realities: Corporate Tax Poliy in a Global Setting, National Tax Journal 57(4): 937 60. Desai, M. A. and Hines, J. R. (2005). Old Rules and New Realities: Corporate Tax Poliy in a Global Setting: Reply to Grubert, National Tax Journal 58(2): 275 278. Devereux, M. P. (1990). Capital Export Neutrality, Capital Import Neutrality, Capital Ownership Neutrality and All That, Unpublished Working Paper. Devereux, M. P. (2007). The Impat of Taxation on the Loation of Capital, Firms and Pro t: A Survey of Empirial Evidene, Oxford University Centre for Business Taxation Working Paper Series No. 07/2. Feldstein, M. and Hartman, D. (1979). The Optimal Taxation of Foreign Soure Investment Inome, Quarterly Journal of Eonomis 93(4): 613 629. Fuest, C. and Huber, B. (2004). Why Do Countries Combine the Exemption System for the Taxation of Foreign Pro ts with Domesti Double Taxation Relief?, Journal of International Eonomis 62(1): 219 231. 24

Fuest, C., Huber, B. and Mintz, J. (2005). Capital Mobility and Tax Competition, Foundations and Trends in Miroeonomis 1(1): 1 62. Gordon, R. H. and Bovenberg, A. L. (1996). Why is Capital So Immobile Internationally? Possible Explanations and Impliations for Capital Inome Taxation, Amerian Eonomi Review 86: 1057 1075. Grubert, H. (2005). Comment on Desai and Hines, "Old Rules and New Realities: Corporate Tax Poliy in a Global Setting.", National Tax Journal 58(2): 263 274. Hau er, A. and Nielsen, S. B. (forthoming). Merger Poliy to Promote Global Players? A Simple Model, Oxford Eonomi Papers. Hau er, A. and Shulte, C. (2007). Merger Poliy and Tax Competition, Working Paper. Hines, J. R. (1999). Lessons from Behavioral Responses to International Taxation, National Tax Journal 52(2): 305 322. Huk, S. and Konrad, K. A. (2004). Merger Pro tability and Trade Poliy, Sandinavian Journal of Eonomis 106(1): 107 122. Huizinga, H. and Voget, J. (2006). International Taxation and the Diretion and Volume of Cross-Border M&As, CEPR Disussion Paper No. 5974. Kaplan, S. (1989). Management Buyouts: Evidene on Taxes as a Soure of Value, Journal of Finane 44(3): 611 632. Musgrave, P. B. (1969). United States Taxation of Foreign Investment Inome: Issues and Arguments, Cambridge (MA), International Tax Program, Harvard Law Shool. Neary, J. P. (2007). Cross-Border Mergers as Instruments of Comparative Advantage, Review of Eonomi Studies 74(4): 1229 57. Sholes, M. S. and Wolfson, M. A. (1990). The E ets of Changes in Tax Laws on Corporate Reorganization Ativity, Journal of Business 63(1): S141 164. 25

Swenson, D. L. (1994). The Impat of U.S. Tax Reform on Foreign Diret Investment in the United States, Journal of Publi Eonomis 54(2): 243 266. Wilson, J. D. (1986). A Theory of Interregional Tax Competition, Journal of Urban Eonomis 19: 296 315. Zodrow, G. R. and Mieszkowski, P. (1986). Pigou, Tiebout, Property Taxation, and the Underprovision of Loal Publi Goods, Journal of Urban Eonomis 19: 356 370. Appendix: An example for orporate overtaxation due to the existene of m&a In this appendix, we show that the tax ompetition equilibrium may indeed imply > 2 (1 + r), so that orporate overtaxation (i.e. a positive welfare e et of orporate tax rate redutions) emerges. We do so by providing a simple example whih makes the following assumptions. 16 The number of ountries n is large, so that @ i onverges to zero. The representative household s utility funtion takes the quadrati form u(c 1 ) + C 2 = (a bc 1 ) C 1 + C 2 (39) where a and b are positive parameters. Assume further that h 0 = > 0 and h 00 = 0. and k are uniformly distributed, with = 0. is normalized to zero. With the budget onstraints given by (21) and (22), optimal period 1 onsumption is given by C 1 = a whih implies savings of (1+r) 2b S = E Z + Z k d + 0 dk a (1 + r) 2b (40) Due to symmetry, S = 0 holds in equilibrium. Therefore, (40) determines r. The total number of projets is given by + m and the total number of + aquisitions is k m. Note further that k2 m = R k kdk and (+ )( + + ) m = k + 2k + 0 2 R + + d. 16 A more detailed desription of how we derived the results in this example is available upon request. 26

Equations (26), (28) and (40) determine the equilibrium values of, and r. Starting with the optimality ondition for, we an solve for and replae it in (40). Then, the resulting expression for r is replaed in (26). Assuming parameter values for, m, E, k +, +, a and b, an be determined iteratively. If = 1:25, m = 20, E = 10, k + = 1, + = 3, a = 10 and b = 0:5, the equilibrium values are ^ = 0:35, ^ = 0:97, r = 0:19 (41) 17 whih implies 2^ (1 + r) < ^. 17 Note that these parameter values also imply 1+r >, whih makes sure that an equilibrium with a positive number of aquisitions is onsidered. 27

CESifo Working Paper Series for full list see Twww.esifo-group.org/wpT (address: Poshingerstr. 5, 81679 Munih, Germany, offie@esifo.de) 2184 Pierre Dubois, Bruno Jullien and Thierry Magna, Formal and Informal Risk Sharing in LDCs: Theory and Empirial Evidene, January 2008 2185 Roel M. W. J. Beetsma, Ward E. Romp and Siert J. Vos, Intergenerational Risk Sharing, Pensions and Endogenous Labor Supply in General Equilibrium, January 2008 2186 Lans Bovenberg and Coen Teulings, Rhineland Exit?, January 2008 2187 Wolfgang Leininger and Axel Okenfels, The Penalty-Duel and Institutional Design: Is there a Neeskens-Effet?, January 2008 2188 Sándor Csengődi and Dieter M. Urban, Foreign Takeovers and Wage Dispersion in Hungary, January 2008 2189 Joerg Baten and Andreas Böhm, Trends of Children s Height and Parental Unemployment: A Large-Sale Anthropometri Study on Eastern Germany, 1994 2006, January 2008 2190 Chris van Klaveren, Bernard van Praag and Henriette Maassen van den Brink, A Publi Good Version of the Colletive Household Model: An Empirial Approah with an Appliation to British Household Data, January 2008 2191 Harry Garretsen and Jolanda Peeters, FDI and the Relevane of Spatial Linkages: Do third Country Effets Matter for Duth FDI?, January 2008 2192 Jan Boukaert, Hans Degryse and Theon van Dijk, Prie Disrimination Bans on Dominant Firms, January 2008 2193 M. Hashem Pesaran, L. Vanessa Smith and Takashi Yamagata, Panel Unit Root Tests in the Presene of a Multifator Error Struture, January 2008 2194 Tomer Blumkin, Bradley J. Ruffle and Yosef Ganun, Are Inome and Consumption Taxes ever really Equivalent? Evidene from a Real-Effort Experiment with Real Goods, January 2008 2195 Mika Widgrén, The Impat of Counil s Internal Deision-Making Rules on the Future EU, January 2008 2196 Antonis Adam, Margarita Katsimi and Thomas Moutos, Inequality and the Import Demand Funtion, January 2008 2197 Helmut Seitz, Demorati Partiipation and the Size of Regions: An Empirial Study Using Data on German Counties, January 2008