- Email: [email protected]
Economic integration and the profitability of cross-border mergers and acquisitions
European Economic Review 48 (2004) 1211 – 1226 www.elsevier.com/locate/econbase
Economic integration and the pro!tability of cross-border mergers and acquisitions Kjetil Bjorvatn∗ Department of Economics, Norwegian School of Economics and Business Administration, Helleveien 30, 5045 Bergen, Norway Received 14 February 2003; accepted 29 March 2004
Abstract The 1990s was a decade of increased economic integration. The decade also witnessed a sharp increase in cross-border mergers and acquisitions. From a theoretical perspective, the increase in international mergers in more integrated economies is rather puzzling. It is a well-established result that due to the “business stealing e3ect”, mergers in integrated markets are not likely to be pro!table. A reasonable conjecture would therefore be that closer integration of markets would reduce the attractiveness of cross-border mergers and acquisitions. The present paper demonstrates that this is not necessarily the case: Economic integration may trigger cross-border acquisitions by reducing the business stealing e3ect and by reducing the reservation price of the target !rm. The paper thus provides explanations to the observed increase in cross-border mergers in a world of more integrated economies. c 2004 Elsevier B.V. All rights reserved. JEL classi#cation: F15; F21; F23; L12; L13 Keywords: Economic integration; Mergers and acquisitions; Trade; Foreign direct investment
1. Introduction Cross-border mergers and acquisitions (M&As) increased sharply during the 1990s. UNCTAD reports that the value of cross-border M&As rose from less than $100 billion in the late 1980s to $720 billion in 1999. Within Europe alone, the value of cross-border acquisitions reached $498 billion in 1999, a 75% increase from the year before. 1 ∗
Tel.: +47-55-95-95-85; fax: +47-55-95-95-43. E-mail address: [email protected] (K. Bjorvatn). 1 In the following, the terms M&A, merger, acquisition, and takeover will be used interchangeably. The great majority of transactions classi!ed as cross-border M&As are in fact acquisitions, see UNCTAD (2000, Chapter IV). c 2004 Elsevier B.V. All rights reserved. 0014-2921/$ - see front matter doi:10.1016/j.euroecorev.2004.03.007
1212
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
Indeed, most of the growth in international production during the 1990s has been driven by cross-border M&As, in 1999 accounting for more than 80% of foreign direct investment Gows. Together with this dramatic increase in international acquisitions, the 1990s saw a liberalization of international trade and closer regional integration. Examples of trade liberalization include the conclusion of the Uruguay round and the establishment of the single-market in the European Community. Moreover, UNCTAD (2000, p. 146) reports that in the period 1991–1999 close to one thousand regulatory changes facilitating foreign direct investment Gows were made in over 100 countries. Many observers refer to economic integration as an important reason for the expansion of international M&As. For instance, on p. 20 in the overview of the UNCTAD 2000 report, it is stated that: “Trade liberalization and regional integration e3orts have added an impetus to cross-border M&As by setting the scene for more intense competition...” From a theoretical perspective, the link between economic integration and pro!tability of cross-border M&As is far from trivial. It is well known from the literature that in a fully integrated market it is generally more pro!table to be outside a merger than to participate in it, see Stigler (1950). Moreover, Salant et al. (1983) show that in a Cournot model with symmetric !rms, a merger involving less than 80 percent of the industry will not be pro!table. The reason is that the outside !rms will expand their production and thereby “steal business” from the merging parties. A reasonable conjecture would therefore be that economic integration would reduce the pro!tability of cross-border mergers. 2 Although cross-border acquisitions is an empirically important phenomenon, it has received relatively little attention in the literature on trade and investment and likewise in the literature on mergers. Exceptions include Kabiraj and Chaudhuri (1999), Head and Ries (1997), Barros and Cabral (1994) and Horn and Levinshon (2001), which analyse the welfare e3ects of mergers and derive policy implications. More recently, the positive issue of equilibrium market structure in an international context has been analysed by Horn and Persson (2001), NorbKack and Persson (2004), and GKorg (2000). The present paper is closely related to Horn and Persson (2001) who focus on international versus national mergers. In their paper, the main advantage of a cross-border merger is that it provides access to a foreign market, while a national merger reduces the competitive pressure in the domestic market. Their main result is that an increase in trade costs may increase the pro!tability of domestic mergers relative to cross-border mergers. The intuition is basically that when trade costs are high, a domestic merger results in very limited international competition, which is a more pro!table venture than a cross-border merger resulting in tough duopoly competition in both markets. When trade costs are low, national mergers do not reduce the competitive pressure to any signi!cant extent. The market access argument then dominates, resulting in cross-border mergers. 2
For recent contributions to merger formation in a closed economy, see Barros (1998), Gowrisankaran (1999), Tombak (2002), Inderst and Wey (2002), and Lommerud et al. (2004). Deneckere and Davidson (1985) demonstrate that in an oligopoly with Bertrand-competition, mergers of any size are pro!table.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1213
Horn and Persson explain how a reduction in international trade costs can trigger cross-border acquisitions. This is the main ambition of the present paper, too. However, the modelling approach and the mechanisms di3er. Horn and Persson use a more general criterion for merger incentives than that of Salant et al. (1983), and allow for more than one merger to take place at the time. The present paper employs a traditional merger model in an open economy setting. This has the advantage of simplicity as well as the fact that it builds on, and thereby hopefully sheds light on, well-established theory. 3 The present paper demonstrates that increased competition following a process of economic integration may increase the pro!tability of cross-border acquisitions. The paper shows exactly when this is likely to be the case, and the underlying mechanisms driving the result. At the same time, the paper also shows that economic integration not necessarily increases the pro!tability of international mergers, and provides the details about when this is true and why. Moreover, the paper shows that economic integration, in the form of lower trade costs, also may lead to less intense post-merger competition, which increases the pro!tability of a merger. The paper is organized as follows. Section 2 presents the model. The analysis of the benchmark case of symmetric !rms is presented in Section 3. Section 4 introduces two cases with asymmetries in !rms’ characteristics. Section 5 concludes. 2. The model Consider a market where demand for the homogeneous good q is given by q = 1 − p;
(1)
where p is the price. The market is supplied by local and foreign !rms. With sj denoting marginal costs for !rm j and qj its supply, operating pro!ts for this !rm are given by j = (p − sj )qj :
(2)
To study the issue at hand, it suOces to consider the situation where, prior to a merger, there are three !rms in the economy, call them a; b and c. 4 We know from Salant et al. (1983) that in a closed economy context with Cournot-competition, a merger between a pair of !rms in the symmetric triopoly case is not pro!table. In line with the literature, I ignore the trivial case of the three !rms merging to form a 3 Horn and Persson (2001) consider only exports. NorbK ack and Persson (2004), on the other hand, allow for both exports and green!eld investment in their analysis of privatization. However, since the reservation price of the privatized !rm is zero, they do not consider the possibility of acquisitions not taking place. Finally, GKorg (2000) analyses the choice between acquisition and green!eld, and therefore abstracts from exports as possible entry mode. By considering only a two-!rm case, he also does not address the issue of an acquisition being unpro!table. 4 Salant et al. (1983) show that adding more !rms reduces the pro!tability of a merger to the merging parties. The mechanisms determining the pro!tability of cross-border mergers in a three-!rm model would survive if we added more !rms to the analysis.
1214
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
monopoly. We can think of such monopolization as being prohibited by competition policy. Let c be the target !rm, located in country C. In the benchmark version of the model, !rms a and b are both located outside country C. These two !rms, which we shall sometimes refer to as the “foreign” !rms or the “active” !rms, are considering how to enter that market. A foreign !rm has three modes of servicing market C. First, it may acquire the !rm already located there. This is the acquisition strategy A. Second, it may enter by investing in a new production plant at a !xed cost f. This is the green!eld strategy G. Third, it may choose exports at a per unit trade cost t, which I shall call strategy X . Finally, the !rm may choose not to serve the market at all, which I call strategy 0. The sequence of moves is as follows. At stage one, a and b simultaneously decide whether or not to invest, and in case of investment, whether to choose strategy A or G. At stage two, there is production and sales, with Cournot-competition between the !rms. j Let YZ represent !rm j’s operating pro!ts in market C if it chooses strategy Y and c represent its rival chooses strategy Z in the absence of a merger. Similarly, let ZY !rm c’s operating pro!ts if the foreign !rms choose strategies Z and Y . In the case of merger, let ˆj+c represent the post-merger pro!t of the merged entity when the outside Y !rm chooses strategy Y , and let ˆkY be the post-merger pro!ts of the outside !rm, !rm k, when it chooses Y as a response to an acquisition by !rm j. In what now follows, I de!ne the relevant operating pro!ts under the various entry strategies. All !rms have identical technologies, and marginal production costs are normalized to zero. 5 If j does not enter the market, its pro!ts are clearly zero: 0j = 0:
(3)
With only two !rms active in market C, both with local production, operating pro!ts are 1 j c k G0 = G0 = ˆj+c (4) G = ˆG = : 9 This market structure is relevant if one of the two active !rms makes a green!eld investment and the other chooses to stay out, or in the case of merger followed by green!eld investment by the outside !rm. A local producer facing competition from another local producer and a foreign exporter receives operating pro!ts equal to (1 + t)2 ; (5) 16 while operating pro!ts in a local triopoly are 1 j c = GG = : (6) GG 16 A !rm exporting to a market with only one local producer receives pro!ts given by j c GX = GX =
Xj 0 = ˆkX = 5
(1 − 2t)2 : 9
Section 4.1 brieGy discusses asymmetries in marginal production costs.
(7)
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1215
This market structure is relevant in the no-merger case where one !rm exports and the other responds by not servicing the market, and for an outside exporter in case of merger. If no merger takes place and both foreign !rms export to market C, their operating pro!ts are given by j XX =
(1 − 2t)2 : 16
(8)
A foreign exporter facing competition from two local producers receives j XG =
(1 − 3t)2 : 16
(9)
The monopoly payo3, which is relevant for the acquiring !rm if the outside !rm stays out, or for the target !rm if both of the two active !rms stay out of the market, is given by 1 c = 00 = : ˆj+c 0 4
(10)
A local producer facing a single, foreign based competitor makes operating pro!ts given by c ˆj+c X = X 0 =
(1 + t)2 : 9
(11)
Finally, in the absence of a merger, if the two active !rms choose exports, the target !rm makes pro!ts equal to c XX =
(1 + 2t)2 : 16
(12)
Following Salant et al. (1983), I focus on the pro!tability of a merger. We know from the study by Salant et al. that in a Cournot-oligopoly, it is typically more pro!table to be outside the merger than to take part in it. This is due to the business stealing e3ect discussed in the introduction: As the merged entity raises the price to exploit the increased market power that follows from the merger, the outside !rm will respond by expanding its output. In the present model, too, it is generally the case that it is more pro!table to be the outside !rm. However, as will be clari!ed at the end of Section 3, for high entry costs it may be more pro!table to be inside the merger than to be an outsider. When it is more pro!table to be outside the merger, there is no reason to believe that a bidding contest for the target !rm will arise. On the other hand, if both !rms prefer to be inside the merger, we would expect to see a bidding contest for the target !rm. I do not address the issue of equilibrium acquisition price in the present study. Rather, I focus on the question of whether a merger is pro!table or not. The assumption is that if an acquisition is pro!table, it will take place. The pro!tability of a merger between !rm j and c is given by gj = ˆj+c − j − c :
(13)
1216
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
Here, ˆj+c denotes the post-merger operating pro!ts of the merged entity, j is j’s pro!ts in the absence of a merger (net of investment cost f in case the best alternative to acquisition is green!eld investment), and c is the target !rm’s pro!ts in the absence of a merger. The condition gj ¿ 0 implies that the pro!ts of the merged entity, ˆj+c , exceed those of the merging parties in the absence of a merger, (j +c ). In terms of an acquisition, (ˆj+c − j ) represents the increase in pro!ts for the acquiring !rm j resulting from the merger, and thus j’s maximum willingness to pay for !rm c. The reservation price of the target !rm is given by c , which is the pro!t it would make in the absence of a merger. Clearly, for an acquisition to be pro!table, the willingness to pay on the part of the acquiring !rm must exceed the reservation price of the target !rm. Notice that the pro!tability of the outside !rm does not a3ect the merger decision. 6 It suOces to know that at least one of the two active !rms can bene!t from acquiring c. If gj ¿ 0 for both !rms, either a or b could be the acquiring !rm, and I will not deal with the issue of which !rm will be inside and which !rm will be outside the merger. I will, however, clarify when it is more pro!table to be inside the merger and when it is more pro!table to be the outside !rm. Economic integration a3ects the pro!tability of a cross-border merger through three channels, given by the three terms on the right hand side of (13). First, economic integration a3ects the pro!ts of the merged entity, ˆj+c , by a3ecting the marginal sales costs of the rival !rm and hence the degree of competition facing the merged entity. In case the rival is an exporter, economic integration in the form of lower trade costs leads to a reduction in sales costs, and hence tougher competition for the merged entity. In other words, this increases the “business stealing” e3ect, which clearly makes a merger less pro!table. Economic integration may also cause a change in the rival’s entry mode. A reduction in trade costs makes exports the more likely entry mode, whereas a reduction in green!eld investment costs makes a green!eld investment more likely. Note that a reduction in trade costs that changes the rival’s entry mode from green!eld to exports reduces the competitive pressure on the merged !rm, and therefore increases the pro!tability of merger. On the other hand, a reduction in green!eld investment costs that causes the rival to change from exports to green!eld, increases the competitive pressure on the merged !rm and thus reduces the pro!tability of merger. Hence, the e3ect of economic integration on the !rst term on the right hand side of (13) depends on whether it causes a shift in entry mode or not, and whether economic integration is in the form of lower trade costs or lower green!eld investment costs. Second, economic integration raises the pro!tability of the alternatives to cross-border acquisition as entry modes into a foreign market. This increases the second term in (13), j , and reduces the pro!tability of a cross-border merger. Third, economic integration may a3ect the third term of (13), namely the reservation price of the target !rm, c . As in the case of ˆj+c , the e3ect of economic integration on c depends on whether the 6 Two articles that focus on the way in which the pro!tability of the outside !rm may a3ect the acquisition processes are Inderst and Wey (2002) and Fridolfsson and Stennek (2000). The former discusses how the free rider problem in mergers a3ects the probability of a merger taking place, while the second analyses how the free rider problem may delay mergers.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1217
foreign !rms are exporters or green!eld investors, and whether economic integration causes a shift in entry mode or not. A reduction in green!eld investment costs may increase the competitive pressure in the market in the absence of a merger by inducing !rms to enter by investment rather than exports. This would reduce the reservation price of c and thereby increase the pro!tability of an international merger. Economic integration in the form of lower trade costs may have a similar e3ect, as the marginal sales cost of the exporters goes down. But if a reduction in trade costs induces a shift from green!eld investment to exports, this increases the reservation price of the target !rm and thus reduces the pro!tability of an acquisition. In sum, the net e3ect of economic integration on the pro!tability of cross-border acquisitions is not evident. In the analysis that follows I show in more detail how the pro!tability of a cross-border acquisition depends on trade costs and green!eld investment costs, and demonstrate that economic integration may indeed trigger international mergers, and when and why this takes place. 3. Analysis: The benchmark case In the benchmark case !rm a and b are identical in all respects. Asymmetries between !rms a and b are discussed in Section 4. Fig. 1 illustrates the equilibrium market outcomes in the benchmark scenario. The horizontal axis measures per unit trade costs, t, and is bounded above by t = 12 , at which point pro!ts for two foreign exporters j ¿ 0 ⇒ t 6 12 . The vertical axis measures the size of the !xed green!eld are zero; XX investment costs, f. The length of this axis is de!ned by the condition that a green!eld investment should be pro!table when there is only one !rm operating in the foreign j market; G0 − f ¿ 0 ⇒ f 6 19 . The capital letters in Fig. 1 indicate equilibrium entry strategies. XX means that both foreign !rms choose exports; GG that both choose green!eld investment; G0 that one foreign !rm chooses green!eld investment, the other !rm staying out of the market; XG that one !rm exports and the other invests; AX (XX ) means that one !rm acquires c, the response of the other being exports, with the no-merger market structure being characterized by exports from both foreign !rms; AX (GX ) that one !rm acquires c, the response of the other being exports, with the no-merger market structure being characterized by exports by one !rm and green!eld investment by the other; AG (XG) that one !rm acquires c and the other chooses green!eld investment, the no-merger market structure being characterized by exports by one !rm and green!eld investment by the other; AG (GG) that one !rm acquires c and the other chooses green!eld investment, with the no-merger market structure being characterized by green!eld investment by both foreign !rms. The shaded areas in the !gure illustrate combinations of !xed green!eld costs and trade costs for which a merger is unpro!table. In order to analyse the mechanisms involved, it is instructive to start by investigating the equilibrium entry modes in the absence of merger. In Fig. 1, line i shows the critical level of green!eld costs for which a !rm is indi3erent between staying out of the market, given by pro!ts 0j , and investing green!eld in the market side by side j − f. This critical level of !xed cost can with another green!eld investor, given by GG
1218
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
0.12 AX(XX)
v
AX(GX) iv
vi G0
0.08
ii XX
AG(XG)
f
XX
i
iii 0.04
AG(GG) viii XG
vii GG
0 0.1
0.2
0.3
0.4
0.5
t Fig. 1. The benchmark case.
be found as 1 : (14) 16 Similarly, line ii shows the critical level of trade costs for which a !rm is indi3erent between staying out of the market and exporting to the market, given that the rival active !rm has chosen green!eld: 1 j ii : 0j = XG ⇒t= : (15) 3 The iii-line shows combinations of !xed investment costs and trade costs for which a !rm is indi3erent between green!eld investment and exports, given that the other active !rm has invested in the market: 3 9 2 j j − f = XG ⇒f= t− (16) iii : GG t : 6 16 Line iv represents combinations of !xed investment costs and trade costs for which a !rm is indi3erent between green!eld investment and exports, given that the rival !rm is an exporter: 3 3 2 j j iv : GX − f = XX ⇒f= t− (17) t : 8 16 Finally, the v-line shows combinations of investment costs and trade costs above which an entrant, in case he meets a single competitor in the foreign market, chooses exports and below which he chooses green!eld. This information is relevant if the rival chooses to stay out of the market or acquires !rm c. 4 4 j v : Xj 0 = G0 − f ⇒ f = t − t2: (18) 9 9 i:
j −f ⇒f= 0j = GG
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1219
The model without mergers is standard, and the results are as one should expect. Basically, when trade costs are low relative to green!eld costs the foreign !rms choose exports, and when trade costs are high, they choose green!eld. For intermediate levels of green!eld costs there is room for only one green!eld investor; when trade costs are relatively low, the optimal response of the rival is to choose exports, whereas if trade costs are high, the rival stays out of the market. We now have the information we need on the no-merger market structure to analyse the main issue, namely the pro!tability of merger. The vi-line shows the critical level of trade costs below which a merger is unpro!table, given that in the absence of a merger, both active !rms would have chosen exports and that the response of the outside !rm to a merger is also exports (i.e. that we are to the left of the v-line): 1 j c : (19) vi : ˆj+c X − XX − XX ¡ 0 ⇒ t ¡ 14 The vii-line shows the critical level of green!eld costs below which a merger is unpro!table, given that in the absence of a merger, both active !rms would have chosen green!eld investment (i.e. that we are to the right of the iii-line) and that the response of the outside !rm to a merger is also green!eld investment: 1 j c vii : ˆj+c : (20) G − (GG − f) − GG ¡ 0 ⇒ f ¡ 72 The viii-line shows the critical level of trade costs below which a merger is unpro!table, given that the rival !rm has chosen green!eld, and that !rm j’s choice is whether to export to the market or to acquire the foreign !rm: 1 j c viii : ˆj+c : (21) G − XG − XG ¡ 0 ⇒ t ¡ 15 The main reason why a merger is not pro!table for low trade costs or low green!eld costs, as de!ned in (19)–(21), is business stealing by the outside !rm. The same e3ect that makes mergers unpro!table in a single-market setting also dominates when markets are almost perfectly integrated. In (13), this means that ˆj+c is fairly low. If a merger leads to green!eld investment by the outside !rm, while the no-merger scenario is characterized by exports by both !rms, a merger is never pro!table. Hence, the area above the iv-line and below the v-line, which we shall de!ne as (iv; v), is shaded: j c (iv; v) : ˆj+c G − XX − XX ¡ 0:
(22)
Finally, if the rival !rm is established with a green!eld investment and the choice of !rm j is whether to acquire c or to stay out of the market, an acquisition yields zero pro!ts. This is true in the area above the i-line, to the right of the ii-line, and below the v-line, de!ned as (i; ii; v): (i; ii; v) :
j c ˆj+c G − 0 − G0 = 0:
(23)
The main reason why acquiring c is not pro!table for intermediate levels of green!eld costs and trade costs, as de!ned by (22) and (23), is that the competitive pressure is not very tough in the absence of a merger, and so the reservation price of the
1220
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
target !rm is relatively high. In (i; ii; v) the no-merger market structure is characterized by duopoly competition between c and a green!eld investor. Since an acquisition in this case would not change the equilibrium market structure, the willingness to pay for c exactly equals c’s reservation price. In (iv; v) the no-merger market structure is characterized by exports by both active !rms, and with fairly high trade costs, the reservation price is also high. It is clear from Fig. 1 and the discussion above that economic integration not necessarily increases the pro!tability of a cross-border merger. The most interesting result, however, is that economic integration may trigger cross-border acquisitions and this is what I choose to highlight here. The main result of the paper can be stated as: Proposition 1. Economic integration, in the form of reduced trade costs and/or reduced green#eld investment costs, may trigger a cross-border acquisition through two channels. First, by intensifying competition in the no-merger case, and thus reducing the reservation price of the target #rm. Second, by reducing the competitive pressure in the post-merger case, and thus increasing the willingness to pay on the part of the acquiring #rm. Starting from “high” entry costs, more speci!cally the shaded area marked with G0 in Fig. 1, a reduction in trade costs such that we cross the ii-line or a reduction in green!eld costs such that we cross the i-line turns the pro!tability of a merger from negative to positive. The reason is that economic integration changes the no-merger market structure from duopoly to triopoly, thus reducing the reservation price of the target !rm. Starting from “intermediate” entry costs, more speci!cally the shaded area marked XX in Fig. 1, a reduction in green!eld costs such that we cross the iv-line induces one of the foreign !rms to choose green!eld investment as entry mode in the no-merger case. Again, this reduces the reservation price of the target !rm and turns the profitability of a merger from negative to positive. Starting from the same area, a reduction in trade costs such that we cross the v-curve reduces the competitive pressure on the merged entity and thus induces a cross-border merger. The reason is that the optimal post-merger entry mode of the outside !rm changes from green!eld to exports. The e3ect of economic integration on the pro!tability of a merger is explored in more detail in Fig. 2. Here, I consider the e3ect of a reduction in trade costs on the pro!tability of a merger for a given level of green!eld costs, f = 0:07. The reservation price of the target !rm, c , is illustrated with a bold line, and the willingness to pay of !rm j; ˆj+c − j , is illustrated with a !ne line. Region (i; ii; v) is characterized by duopoly, with one foreign investor competing with !rm c. We know from (23) that in this region, the reservation price of !rm c equals j c the willingness to pay on part of the active !rm; ˆj+c G − 0 = G0 . A reduction in t such 1 that we cross the ii-line, de!ned by t = 3 , leads to a reduction in c . The reason is that the no-merger market structure changes from duopoly to triopoly, with one foreign investor and one foreign exporter. Intuitively, by causing more intense competition in the market, economic integration raises the gains from increased market concentration through merger.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1221
0.14
^j πj+c- Π 0.12
0.1
πc
0.08
0.06
vi
v
iv
ii
(i,ii,v)
0.04 0.1
0.2
0.3
0.4
0.5
t Fig. 2. Reservation price and willingness to pay.
Consider next a reduction of t such that we cross the iv-line. This changes the sign of gj from positive to negative. The reason is basically that economic integration in this case reduces the intensity of competition in the absence of a merger, by changing the market structure from one investor and one exporter to a situation in which both active !rms export. This reduces the pro!tability of a merger by increasing c and j , thus shifting both curves in Fig. 2. A further reduction in t such that we cross the v-line again changes the sign of gj , this time from negative to positive. The reason is that economic integration now leads to a radical reduction in post-merger competition, as the outside !rm changes the entry strategy from green!eld investment to exports. The resulting reduction in the business stealing e3ect by the outsider leads to a sharp increase in ˆj+c , and hence a positive shift in the willingness to pay function illustrated in Fig. 2. Finally, a reduction in t such that we cross the vi-line makes a merger unpro!table by increasing the business stealing e3ect. As is evident from Fig. 2, to the left of the vi-line, this e3ect dominates the fall in c caused by lower trade costs and hence tougher competition. Of course, with zero trade costs we are e3ectively in a perfectly integrated market, and we know from the literature that a merger between two !rms would not pro!table in that case. Fig. 2 illustrates economic integration in the form of a reduction in trade costs. Similar mechanisms can, however be derived by considering a reduction in green!eld costs, f. Consider, for instance a reduction in f such that we cross the v-line and enter region (i; ii; v) in Fig. 1. This changes the sign of gj from positive to negative by changing the entry mode of the outside !rm from exports to green!eld investment, thus reducing ˆj+c . A continued reduction in f such that we cross the i-line allows pro!table
1222
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
investment by both active !rms. This makes the merger pro!table by intensifying competition in the no-merger case and thus reducing the reservation price of the target !rm. A reduction in f such that we cross the vii-line makes a merger unpro!table by making the alternative to an acquisition as entry mode more pro!table for !rm j, hence increasing j . Given that a merger is pro!table, is it more pro!table to be inside or outside the merger? This is an interesting question, although not at the core of the present analysis. It is straightforward to demonstrate that: Proposition 2. Given that the acquisition price is de#ned by the reservation price of the target #rm, the pro#ts of the inside #rm dominate those of the outside #rm when trade costs and green#eld investment costs are relatively high. Comparing ˆk , which is the post-merger pro!ts of the outside !rm, with ˆj+c − c , which is the post-merger pro!ts of the acquiring !rm, net of the acquisition price 3 c ; ˆj+c − c ¿ ˆk is satis!ed only for t ¿ 14 above the v-curve. Intuitively, when exports is the relevant alternative to acquisition as entry mode into market C, high trade costs reduces both the business stealing e3ect and the pro!ts associated with the alternative entry mode. In combination, this makes it more pro!table to be an insider than an outsider to the merger. When both a and b wish to acquire c, it is reasonable to assume that a bidding contest will arise, thus raising the acquisition price. In a standard auction model we would expect the equilibrium price of the asset to be such that the bidding !rms are indi3erent between acquiring c or exporting to the market. If this is the case, then naturally the post-merger pro!ts of the inside !rm, net of the acquisition price, will be equal to those of the outside !rm. 4. Extensions: Asymmetries So far in the analysis, I have assumed that the two active !rms are entirely similar. In what now follows, I allow for asymmetries between the active !rms along two dimensions. First, in terms of cost eOciency. Second, in terms of their initial location. There are of course a number of ways in which the three !rms could di3er, even restricting our attention to eOciency and location. Rather than carrying out a full analysis of how asymmetries in these two dimensions would a3ect the model, I use a couple examples that capture some interesting e3ects. 4.1. Technological asymmetry Consider !rst the case where there is one technological leader, !rm a, with the other two !rms having identical, but inferior technology to that of a. Maintaining the assumption of zero marginal costs for a, let ¿ 0 represent the common marginal production costs of b and c, which can also be interpreted as the technology gap between the more and less advanced !rms. Given that !rm a can apply its superior technology to c in case of acquisition, this modi!cation of the model clearly increases the pro!tability of a merger between a and c. It does so by adding an eOciency gain
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1223
to the merger. In fact, if the technology gap is suOciently large, there is an incentive for a to acquire c even for zero trade costs and zero green!eld costs. We can easily see this by noting that the marginal production cost a3ects pro!ts exactly as the trade cost t does. Let t = 0 or f = 0 so that we are e3ectively in a single-market situation. In this case, a merger between a and c is pro!table to the merging parties if (1 − 2)2 (1 + )2 (1 + 2)2 ¿ ⇒ ¿ 0:07; − 16 16 9
(24)
where the !rst term represents the post-merger pro!ts of the merged entity, the second term the no-merger pro!ts of !rm a and the third term the no-merger pro!ts of !rm c. Hence, for ¿ 0:07, !rm a will acquire c even for zero trade costs or zero green!eld costs. The result that asymmetries in technology can make a merger pro!table even in a closed economy setting is not novel, see for instance Barros (1998). 4.2. Location asymmetry Consider next asymmetry in the initial location of the two active !rms. Assume that initially, i.e. prior to any investment decision, b and c are located in the same market with a being the only foreign !rm. In this case, market entry is of course an issue only for !rm a. 7 As in the benchmark case, the !rms are assumed to have identical marginal production costs. This modi!cation of the analysis allows us to deal with the interesting issue of the nationality of the acquiring !rm. I shall refer to this scenario as the “nationality” case. When should we expect to see cross-border mergers and when are mergers between two !rms located in the same market more pro!table? Fig. 3 illustrates the outcome of this scenario. The de!nitions of the lines dividing the various market structures in Fig. 3 are the same as in the benchmark scenario. X means that !rm a chooses exports; G that a chooses green!eld investment; 0 that a stays out of the market; Aa (X ) means that a acquires c, the alternative to acquisition for a being exports; Aa (G) that a acquires c, the alternative being green!eld investment; Ab X means that b acquires c, the response of a being exports. The no-merger equilibrium market structure is relatively straightforward. Again, low trade costs relative to green!eld costs induces a to export, indicated by X in Fig. 3, whereas high trade costs relative to green!eld costs induces investment, G. When the cost of both kinds of entry are high, !rm a stays out, indicated by 0. As in the benchmark scenario, Fig. 3 shows that economic integration may trigger cross-border acquisitions. The acquisition of c by a is indicated by Aa . It can be shown that for “low” entry costs, it is never pro!table for !rm b acquire c. This is clearly the case under the v-curve: Here, an acquisition by b would be followed by a green!eld investment by a. Hence, !rm b has nothing to gain from acquiring c itself rather than leaving the acquisition to b, since in either case the post-merger market structure will be characterized by local duopoly competition. 7 The general insights derived from this analysis would apply also in a situation where b faces lower trade costs than a, and not necessarily zero trade costs as in the present example.
1224
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
0.12
ix
AbX v
vi ii
0.08
0
Aa(X)
f
X i
iii Aa(G)
0.04
vii G
0 0.1
0.2
0.3
0.4
0.5
t Fig. 3. The nationality case.
The situation is slightly di3erent above the v-line. In this case, if b acquires c, the response of a would be exports. As such, this is bene!cial to !rm b who prefers to have a at a distance from the market. However, it is possible to show that for “low” trade costs, more speci!cally to the left of the ii-line, this is not a pro!table option for !rm b. To see this, note that above the v-line and to the left of the ii-line, the no-merger entry mode of !rm a is exports. Hence, the reservation price of !rm c c c is given by GX . The value added from a merger by b is therefore ˆbX − GX . The b alternative for b is to let a acquire c, resulting in pro!ts G0 for b. It can be shown that c b ˆbX − GX − G0 ¡0
for all t ¡ 0:5:
(25)
On the other hand, when entry costs are “high”, i.e. to the right of the ii-line and above the v-curve, it may indeed be pro!table for b acquire c. For a, an acquisition is here not a pro!table option. If b does not acquire c, !rm a stays out of the market. c Hence, the reservation price of c is given by G0 . In Fig. 3, the critical level of trade costs is given by the ix-line, which can be found as c b ix : ˆbX − G0 − G0 ¡ 0 ⇒ t ¡ 0:41:
(26)
Hence, for t ¿ 0:41 and above the v-curve, !rm b !nds it pro!table to acquire c. Intuitively, the gain to the local !rm, !rm b, from this merger is that it replaces a local competitor with a foreign based exporter. For suOciently high trade costs, the business stealing e3ect is suOciently small to make the merger pro!table. In Fig. 3, the equilibrium market structure in this case is indicated by Ab X , i.e. acquisition by b followed by exports by a.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1225
5. Concluding remarks The results presented here demonstrate that the relation between economic integration and the pro!tability of cross-border mergers is complex. One possibility is that economic integration may trigger cross-border mergers, contrary to what one perhaps would think from reading the theoretical literature on mergers in closed economies. The reason is that economic integration may intensify the pre-merger competition in the market, thereby reducing the reservation price of the target !rm. In addition, economic integration in the form of lower trade costs may reduce the post-merger business stealing e3ect as the outside !rm chooses exports rather than green!eld investment. The benchmark version of the model deals with the standard case of symmetric !rms. An extension to the model deals with asymmetries between !rms. It shows that a technologically advantaged !rm has a stronger incentive to acquire a foreign !rm than its less advanced rival, given that technology can freely be applied to the target !rm. Finally, I demonstrate that when there is an asymmetry in the location of the two active !rms, entry costs is an argument in favor of cross-border acquisitions rather than national mergers. However, if entry costs are suOciently high, the price of the target !rm may be too high for the foreign !rm to !nd an acquisition pro!table. In this case, a merger between the local !rms is more likely. There are certainly a number issues related to how asymmetries between !rms a3ect equilibrium entry modes that are not explored in the present paper. Does a technological leadership necessarily raise the pro!tability of acquisition as entry mode? What role does the possibility of transferring technology from the mother !rm to the acquired !rm play in the entry decision? How would trade costs a3ect the analysis if the acquisition is an export platform investment rather than a market seeking investment, i.e. if the acquired plant is used primarily for export purposes? These interesting questions are left for future research. Acknowledgements I would like to thank Lars Persson, Lars SHrgard and two anonymous referees for useful comments. Remaining errors are mine. References Barros, P., 1998. Endogenous mergers and size asymmetry of merger participants. Economics Letters 60, 113–119. Barros, P., Cabral, L., 1994. Merger policy in open economies. European Economic Review 38, 1041–1055. Deneckere, R., Davidson, C., 1985. Incentives to form coalitions with Bertrand competition. Rand Journal of Economics 16 (4), 473–486. Fridolfsson, S.-O., Stennek, J., 2000. Should mergers be controlled? IUI Working Paper No. 541. The Research Institute of Industrial Economics, Stockholm. GKorg, H., 2000. Analysing foreign market entry: the choice between green!eld investment and acquisitions. Journal of Economic Studies 27 (3), 165–181. Gowrisankaran, G., 1999. A dynamic model of endogenous horizontal mergers. Rand Journal of Economics 30 (1), 56–83.
1226
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
Head, K., Ries, J., 1997. International mergers and welfare under decentralized competition policy. The Canadian Journal of Economics 30 (4), 1104–1123. Horn, H., Levinshon, J., 2001. Merger policies and trade liberalization. Economic Journal 111 (470), 244– 276. Horn, H., Persson, L., 2001. The equilibrium ownership of an international oligopoly. Journal of International Economics 53, 307–333. Inderst, R., Wey, C., 2002. The incentives for takeover in oligopoly. CEPR Discussion Paper 3163, Centre for Economic Policy Research, London. Kabiraj, T., Chaudhuri, M., 1999. On the welfare analysis of a cross-border merger. The Journal of International Trade and Economic Development 8 (2), 195–207. Lommerud, K.E., Straume, O.R., SHrgard, L., 2004. Downstream merger with upstream market power. European Economic Review, Forthcoming. NorbKack, P.J., Persson, L., 2004. Privatization and foreign competition. Journal of International Economics 62 (2), 409–416. Salant, S.W., Switzer, S., Reynolds, R.J., 1983. Losses due to merger: the e3ects of an exogenous change in industry structure on Cournot-Nash equilibrium. Quarterly Journal of Economics 98, 185–199. Stigler, G.J., 1950. Monopoly and oligopoly by merger. American Economic Review 40, 23–34. Tombak, M.M., 2002. Mergers to monopoly. Journal of Economics and Management Strategy 11 (3), 513– 546. UNCTAD, 2000. World Investment Report 2000: Cross-border Mergers and Acquisitions and Development, United Nations.
Economic integration and the pro!tability of cross-border mergers and acquisitions Kjetil Bjorvatn∗ Department of Economics, Norwegian School of Economics and Business Administration, Helleveien 30, 5045 Bergen, Norway Received 14 February 2003; accepted 29 March 2004
Abstract The 1990s was a decade of increased economic integration. The decade also witnessed a sharp increase in cross-border mergers and acquisitions. From a theoretical perspective, the increase in international mergers in more integrated economies is rather puzzling. It is a well-established result that due to the “business stealing e3ect”, mergers in integrated markets are not likely to be pro!table. A reasonable conjecture would therefore be that closer integration of markets would reduce the attractiveness of cross-border mergers and acquisitions. The present paper demonstrates that this is not necessarily the case: Economic integration may trigger cross-border acquisitions by reducing the business stealing e3ect and by reducing the reservation price of the target !rm. The paper thus provides explanations to the observed increase in cross-border mergers in a world of more integrated economies. c 2004 Elsevier B.V. All rights reserved. JEL classi#cation: F15; F21; F23; L12; L13 Keywords: Economic integration; Mergers and acquisitions; Trade; Foreign direct investment
1. Introduction Cross-border mergers and acquisitions (M&As) increased sharply during the 1990s. UNCTAD reports that the value of cross-border M&As rose from less than $100 billion in the late 1980s to $720 billion in 1999. Within Europe alone, the value of cross-border acquisitions reached $498 billion in 1999, a 75% increase from the year before. 1 ∗
Tel.: +47-55-95-95-85; fax: +47-55-95-95-43. E-mail address: [email protected] (K. Bjorvatn). 1 In the following, the terms M&A, merger, acquisition, and takeover will be used interchangeably. The great majority of transactions classi!ed as cross-border M&As are in fact acquisitions, see UNCTAD (2000, Chapter IV). c 2004 Elsevier B.V. All rights reserved. 0014-2921/$ - see front matter doi:10.1016/j.euroecorev.2004.03.007
1212
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
Indeed, most of the growth in international production during the 1990s has been driven by cross-border M&As, in 1999 accounting for more than 80% of foreign direct investment Gows. Together with this dramatic increase in international acquisitions, the 1990s saw a liberalization of international trade and closer regional integration. Examples of trade liberalization include the conclusion of the Uruguay round and the establishment of the single-market in the European Community. Moreover, UNCTAD (2000, p. 146) reports that in the period 1991–1999 close to one thousand regulatory changes facilitating foreign direct investment Gows were made in over 100 countries. Many observers refer to economic integration as an important reason for the expansion of international M&As. For instance, on p. 20 in the overview of the UNCTAD 2000 report, it is stated that: “Trade liberalization and regional integration e3orts have added an impetus to cross-border M&As by setting the scene for more intense competition...” From a theoretical perspective, the link between economic integration and pro!tability of cross-border M&As is far from trivial. It is well known from the literature that in a fully integrated market it is generally more pro!table to be outside a merger than to participate in it, see Stigler (1950). Moreover, Salant et al. (1983) show that in a Cournot model with symmetric !rms, a merger involving less than 80 percent of the industry will not be pro!table. The reason is that the outside !rms will expand their production and thereby “steal business” from the merging parties. A reasonable conjecture would therefore be that economic integration would reduce the pro!tability of cross-border mergers. 2 Although cross-border acquisitions is an empirically important phenomenon, it has received relatively little attention in the literature on trade and investment and likewise in the literature on mergers. Exceptions include Kabiraj and Chaudhuri (1999), Head and Ries (1997), Barros and Cabral (1994) and Horn and Levinshon (2001), which analyse the welfare e3ects of mergers and derive policy implications. More recently, the positive issue of equilibrium market structure in an international context has been analysed by Horn and Persson (2001), NorbKack and Persson (2004), and GKorg (2000). The present paper is closely related to Horn and Persson (2001) who focus on international versus national mergers. In their paper, the main advantage of a cross-border merger is that it provides access to a foreign market, while a national merger reduces the competitive pressure in the domestic market. Their main result is that an increase in trade costs may increase the pro!tability of domestic mergers relative to cross-border mergers. The intuition is basically that when trade costs are high, a domestic merger results in very limited international competition, which is a more pro!table venture than a cross-border merger resulting in tough duopoly competition in both markets. When trade costs are low, national mergers do not reduce the competitive pressure to any signi!cant extent. The market access argument then dominates, resulting in cross-border mergers. 2
For recent contributions to merger formation in a closed economy, see Barros (1998), Gowrisankaran (1999), Tombak (2002), Inderst and Wey (2002), and Lommerud et al. (2004). Deneckere and Davidson (1985) demonstrate that in an oligopoly with Bertrand-competition, mergers of any size are pro!table.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1213
Horn and Persson explain how a reduction in international trade costs can trigger cross-border acquisitions. This is the main ambition of the present paper, too. However, the modelling approach and the mechanisms di3er. Horn and Persson use a more general criterion for merger incentives than that of Salant et al. (1983), and allow for more than one merger to take place at the time. The present paper employs a traditional merger model in an open economy setting. This has the advantage of simplicity as well as the fact that it builds on, and thereby hopefully sheds light on, well-established theory. 3 The present paper demonstrates that increased competition following a process of economic integration may increase the pro!tability of cross-border acquisitions. The paper shows exactly when this is likely to be the case, and the underlying mechanisms driving the result. At the same time, the paper also shows that economic integration not necessarily increases the pro!tability of international mergers, and provides the details about when this is true and why. Moreover, the paper shows that economic integration, in the form of lower trade costs, also may lead to less intense post-merger competition, which increases the pro!tability of a merger. The paper is organized as follows. Section 2 presents the model. The analysis of the benchmark case of symmetric !rms is presented in Section 3. Section 4 introduces two cases with asymmetries in !rms’ characteristics. Section 5 concludes. 2. The model Consider a market where demand for the homogeneous good q is given by q = 1 − p;
(1)
where p is the price. The market is supplied by local and foreign !rms. With sj denoting marginal costs for !rm j and qj its supply, operating pro!ts for this !rm are given by j = (p − sj )qj :
(2)
To study the issue at hand, it suOces to consider the situation where, prior to a merger, there are three !rms in the economy, call them a; b and c. 4 We know from Salant et al. (1983) that in a closed economy context with Cournot-competition, a merger between a pair of !rms in the symmetric triopoly case is not pro!table. In line with the literature, I ignore the trivial case of the three !rms merging to form a 3 Horn and Persson (2001) consider only exports. NorbK ack and Persson (2004), on the other hand, allow for both exports and green!eld investment in their analysis of privatization. However, since the reservation price of the privatized !rm is zero, they do not consider the possibility of acquisitions not taking place. Finally, GKorg (2000) analyses the choice between acquisition and green!eld, and therefore abstracts from exports as possible entry mode. By considering only a two-!rm case, he also does not address the issue of an acquisition being unpro!table. 4 Salant et al. (1983) show that adding more !rms reduces the pro!tability of a merger to the merging parties. The mechanisms determining the pro!tability of cross-border mergers in a three-!rm model would survive if we added more !rms to the analysis.
1214
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
monopoly. We can think of such monopolization as being prohibited by competition policy. Let c be the target !rm, located in country C. In the benchmark version of the model, !rms a and b are both located outside country C. These two !rms, which we shall sometimes refer to as the “foreign” !rms or the “active” !rms, are considering how to enter that market. A foreign !rm has three modes of servicing market C. First, it may acquire the !rm already located there. This is the acquisition strategy A. Second, it may enter by investing in a new production plant at a !xed cost f. This is the green!eld strategy G. Third, it may choose exports at a per unit trade cost t, which I shall call strategy X . Finally, the !rm may choose not to serve the market at all, which I call strategy 0. The sequence of moves is as follows. At stage one, a and b simultaneously decide whether or not to invest, and in case of investment, whether to choose strategy A or G. At stage two, there is production and sales, with Cournot-competition between the !rms. j Let YZ represent !rm j’s operating pro!ts in market C if it chooses strategy Y and c represent its rival chooses strategy Z in the absence of a merger. Similarly, let ZY !rm c’s operating pro!ts if the foreign !rms choose strategies Z and Y . In the case of merger, let ˆj+c represent the post-merger pro!t of the merged entity when the outside Y !rm chooses strategy Y , and let ˆkY be the post-merger pro!ts of the outside !rm, !rm k, when it chooses Y as a response to an acquisition by !rm j. In what now follows, I de!ne the relevant operating pro!ts under the various entry strategies. All !rms have identical technologies, and marginal production costs are normalized to zero. 5 If j does not enter the market, its pro!ts are clearly zero: 0j = 0:
(3)
With only two !rms active in market C, both with local production, operating pro!ts are 1 j c k G0 = G0 = ˆj+c (4) G = ˆG = : 9 This market structure is relevant if one of the two active !rms makes a green!eld investment and the other chooses to stay out, or in the case of merger followed by green!eld investment by the outside !rm. A local producer facing competition from another local producer and a foreign exporter receives operating pro!ts equal to (1 + t)2 ; (5) 16 while operating pro!ts in a local triopoly are 1 j c = GG = : (6) GG 16 A !rm exporting to a market with only one local producer receives pro!ts given by j c GX = GX =
Xj 0 = ˆkX = 5
(1 − 2t)2 : 9
Section 4.1 brieGy discusses asymmetries in marginal production costs.
(7)
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1215
This market structure is relevant in the no-merger case where one !rm exports and the other responds by not servicing the market, and for an outside exporter in case of merger. If no merger takes place and both foreign !rms export to market C, their operating pro!ts are given by j XX =
(1 − 2t)2 : 16
(8)
A foreign exporter facing competition from two local producers receives j XG =
(1 − 3t)2 : 16
(9)
The monopoly payo3, which is relevant for the acquiring !rm if the outside !rm stays out, or for the target !rm if both of the two active !rms stay out of the market, is given by 1 c = 00 = : ˆj+c 0 4
(10)
A local producer facing a single, foreign based competitor makes operating pro!ts given by c ˆj+c X = X 0 =
(1 + t)2 : 9
(11)
Finally, in the absence of a merger, if the two active !rms choose exports, the target !rm makes pro!ts equal to c XX =
(1 + 2t)2 : 16
(12)
Following Salant et al. (1983), I focus on the pro!tability of a merger. We know from the study by Salant et al. that in a Cournot-oligopoly, it is typically more pro!table to be outside the merger than to take part in it. This is due to the business stealing e3ect discussed in the introduction: As the merged entity raises the price to exploit the increased market power that follows from the merger, the outside !rm will respond by expanding its output. In the present model, too, it is generally the case that it is more pro!table to be the outside !rm. However, as will be clari!ed at the end of Section 3, for high entry costs it may be more pro!table to be inside the merger than to be an outsider. When it is more pro!table to be outside the merger, there is no reason to believe that a bidding contest for the target !rm will arise. On the other hand, if both !rms prefer to be inside the merger, we would expect to see a bidding contest for the target !rm. I do not address the issue of equilibrium acquisition price in the present study. Rather, I focus on the question of whether a merger is pro!table or not. The assumption is that if an acquisition is pro!table, it will take place. The pro!tability of a merger between !rm j and c is given by gj = ˆj+c − j − c :
(13)
1216
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
Here, ˆj+c denotes the post-merger operating pro!ts of the merged entity, j is j’s pro!ts in the absence of a merger (net of investment cost f in case the best alternative to acquisition is green!eld investment), and c is the target !rm’s pro!ts in the absence of a merger. The condition gj ¿ 0 implies that the pro!ts of the merged entity, ˆj+c , exceed those of the merging parties in the absence of a merger, (j +c ). In terms of an acquisition, (ˆj+c − j ) represents the increase in pro!ts for the acquiring !rm j resulting from the merger, and thus j’s maximum willingness to pay for !rm c. The reservation price of the target !rm is given by c , which is the pro!t it would make in the absence of a merger. Clearly, for an acquisition to be pro!table, the willingness to pay on the part of the acquiring !rm must exceed the reservation price of the target !rm. Notice that the pro!tability of the outside !rm does not a3ect the merger decision. 6 It suOces to know that at least one of the two active !rms can bene!t from acquiring c. If gj ¿ 0 for both !rms, either a or b could be the acquiring !rm, and I will not deal with the issue of which !rm will be inside and which !rm will be outside the merger. I will, however, clarify when it is more pro!table to be inside the merger and when it is more pro!table to be the outside !rm. Economic integration a3ects the pro!tability of a cross-border merger through three channels, given by the three terms on the right hand side of (13). First, economic integration a3ects the pro!ts of the merged entity, ˆj+c , by a3ecting the marginal sales costs of the rival !rm and hence the degree of competition facing the merged entity. In case the rival is an exporter, economic integration in the form of lower trade costs leads to a reduction in sales costs, and hence tougher competition for the merged entity. In other words, this increases the “business stealing” e3ect, which clearly makes a merger less pro!table. Economic integration may also cause a change in the rival’s entry mode. A reduction in trade costs makes exports the more likely entry mode, whereas a reduction in green!eld investment costs makes a green!eld investment more likely. Note that a reduction in trade costs that changes the rival’s entry mode from green!eld to exports reduces the competitive pressure on the merged !rm, and therefore increases the pro!tability of merger. On the other hand, a reduction in green!eld investment costs that causes the rival to change from exports to green!eld, increases the competitive pressure on the merged !rm and thus reduces the pro!tability of merger. Hence, the e3ect of economic integration on the !rst term on the right hand side of (13) depends on whether it causes a shift in entry mode or not, and whether economic integration is in the form of lower trade costs or lower green!eld investment costs. Second, economic integration raises the pro!tability of the alternatives to cross-border acquisition as entry modes into a foreign market. This increases the second term in (13), j , and reduces the pro!tability of a cross-border merger. Third, economic integration may a3ect the third term of (13), namely the reservation price of the target !rm, c . As in the case of ˆj+c , the e3ect of economic integration on c depends on whether the 6 Two articles that focus on the way in which the pro!tability of the outside !rm may a3ect the acquisition processes are Inderst and Wey (2002) and Fridolfsson and Stennek (2000). The former discusses how the free rider problem in mergers a3ects the probability of a merger taking place, while the second analyses how the free rider problem may delay mergers.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1217
foreign !rms are exporters or green!eld investors, and whether economic integration causes a shift in entry mode or not. A reduction in green!eld investment costs may increase the competitive pressure in the market in the absence of a merger by inducing !rms to enter by investment rather than exports. This would reduce the reservation price of c and thereby increase the pro!tability of an international merger. Economic integration in the form of lower trade costs may have a similar e3ect, as the marginal sales cost of the exporters goes down. But if a reduction in trade costs induces a shift from green!eld investment to exports, this increases the reservation price of the target !rm and thus reduces the pro!tability of an acquisition. In sum, the net e3ect of economic integration on the pro!tability of cross-border acquisitions is not evident. In the analysis that follows I show in more detail how the pro!tability of a cross-border acquisition depends on trade costs and green!eld investment costs, and demonstrate that economic integration may indeed trigger international mergers, and when and why this takes place. 3. Analysis: The benchmark case In the benchmark case !rm a and b are identical in all respects. Asymmetries between !rms a and b are discussed in Section 4. Fig. 1 illustrates the equilibrium market outcomes in the benchmark scenario. The horizontal axis measures per unit trade costs, t, and is bounded above by t = 12 , at which point pro!ts for two foreign exporters j ¿ 0 ⇒ t 6 12 . The vertical axis measures the size of the !xed green!eld are zero; XX investment costs, f. The length of this axis is de!ned by the condition that a green!eld investment should be pro!table when there is only one !rm operating in the foreign j market; G0 − f ¿ 0 ⇒ f 6 19 . The capital letters in Fig. 1 indicate equilibrium entry strategies. XX means that both foreign !rms choose exports; GG that both choose green!eld investment; G0 that one foreign !rm chooses green!eld investment, the other !rm staying out of the market; XG that one !rm exports and the other invests; AX (XX ) means that one !rm acquires c, the response of the other being exports, with the no-merger market structure being characterized by exports from both foreign !rms; AX (GX ) that one !rm acquires c, the response of the other being exports, with the no-merger market structure being characterized by exports by one !rm and green!eld investment by the other; AG (XG) that one !rm acquires c and the other chooses green!eld investment, the no-merger market structure being characterized by exports by one !rm and green!eld investment by the other; AG (GG) that one !rm acquires c and the other chooses green!eld investment, with the no-merger market structure being characterized by green!eld investment by both foreign !rms. The shaded areas in the !gure illustrate combinations of !xed green!eld costs and trade costs for which a merger is unpro!table. In order to analyse the mechanisms involved, it is instructive to start by investigating the equilibrium entry modes in the absence of merger. In Fig. 1, line i shows the critical level of green!eld costs for which a !rm is indi3erent between staying out of the market, given by pro!ts 0j , and investing green!eld in the market side by side j − f. This critical level of !xed cost can with another green!eld investor, given by GG
1218
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
0.12 AX(XX)
v
AX(GX) iv
vi G0
0.08
ii XX
AG(XG)
f
XX
i
iii 0.04
AG(GG) viii XG
vii GG
0 0.1
0.2
0.3
0.4
0.5
t Fig. 1. The benchmark case.
be found as 1 : (14) 16 Similarly, line ii shows the critical level of trade costs for which a !rm is indi3erent between staying out of the market and exporting to the market, given that the rival active !rm has chosen green!eld: 1 j ii : 0j = XG ⇒t= : (15) 3 The iii-line shows combinations of !xed investment costs and trade costs for which a !rm is indi3erent between green!eld investment and exports, given that the other active !rm has invested in the market: 3 9 2 j j − f = XG ⇒f= t− (16) iii : GG t : 6 16 Line iv represents combinations of !xed investment costs and trade costs for which a !rm is indi3erent between green!eld investment and exports, given that the rival !rm is an exporter: 3 3 2 j j iv : GX − f = XX ⇒f= t− (17) t : 8 16 Finally, the v-line shows combinations of investment costs and trade costs above which an entrant, in case he meets a single competitor in the foreign market, chooses exports and below which he chooses green!eld. This information is relevant if the rival chooses to stay out of the market or acquires !rm c. 4 4 j v : Xj 0 = G0 − f ⇒ f = t − t2: (18) 9 9 i:
j −f ⇒f= 0j = GG
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1219
The model without mergers is standard, and the results are as one should expect. Basically, when trade costs are low relative to green!eld costs the foreign !rms choose exports, and when trade costs are high, they choose green!eld. For intermediate levels of green!eld costs there is room for only one green!eld investor; when trade costs are relatively low, the optimal response of the rival is to choose exports, whereas if trade costs are high, the rival stays out of the market. We now have the information we need on the no-merger market structure to analyse the main issue, namely the pro!tability of merger. The vi-line shows the critical level of trade costs below which a merger is unpro!table, given that in the absence of a merger, both active !rms would have chosen exports and that the response of the outside !rm to a merger is also exports (i.e. that we are to the left of the v-line): 1 j c : (19) vi : ˆj+c X − XX − XX ¡ 0 ⇒ t ¡ 14 The vii-line shows the critical level of green!eld costs below which a merger is unpro!table, given that in the absence of a merger, both active !rms would have chosen green!eld investment (i.e. that we are to the right of the iii-line) and that the response of the outside !rm to a merger is also green!eld investment: 1 j c vii : ˆj+c : (20) G − (GG − f) − GG ¡ 0 ⇒ f ¡ 72 The viii-line shows the critical level of trade costs below which a merger is unpro!table, given that the rival !rm has chosen green!eld, and that !rm j’s choice is whether to export to the market or to acquire the foreign !rm: 1 j c viii : ˆj+c : (21) G − XG − XG ¡ 0 ⇒ t ¡ 15 The main reason why a merger is not pro!table for low trade costs or low green!eld costs, as de!ned in (19)–(21), is business stealing by the outside !rm. The same e3ect that makes mergers unpro!table in a single-market setting also dominates when markets are almost perfectly integrated. In (13), this means that ˆj+c is fairly low. If a merger leads to green!eld investment by the outside !rm, while the no-merger scenario is characterized by exports by both !rms, a merger is never pro!table. Hence, the area above the iv-line and below the v-line, which we shall de!ne as (iv; v), is shaded: j c (iv; v) : ˆj+c G − XX − XX ¡ 0:
(22)
Finally, if the rival !rm is established with a green!eld investment and the choice of !rm j is whether to acquire c or to stay out of the market, an acquisition yields zero pro!ts. This is true in the area above the i-line, to the right of the ii-line, and below the v-line, de!ned as (i; ii; v): (i; ii; v) :
j c ˆj+c G − 0 − G0 = 0:
(23)
The main reason why acquiring c is not pro!table for intermediate levels of green!eld costs and trade costs, as de!ned by (22) and (23), is that the competitive pressure is not very tough in the absence of a merger, and so the reservation price of the
1220
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
target !rm is relatively high. In (i; ii; v) the no-merger market structure is characterized by duopoly competition between c and a green!eld investor. Since an acquisition in this case would not change the equilibrium market structure, the willingness to pay for c exactly equals c’s reservation price. In (iv; v) the no-merger market structure is characterized by exports by both active !rms, and with fairly high trade costs, the reservation price is also high. It is clear from Fig. 1 and the discussion above that economic integration not necessarily increases the pro!tability of a cross-border merger. The most interesting result, however, is that economic integration may trigger cross-border acquisitions and this is what I choose to highlight here. The main result of the paper can be stated as: Proposition 1. Economic integration, in the form of reduced trade costs and/or reduced green#eld investment costs, may trigger a cross-border acquisition through two channels. First, by intensifying competition in the no-merger case, and thus reducing the reservation price of the target #rm. Second, by reducing the competitive pressure in the post-merger case, and thus increasing the willingness to pay on the part of the acquiring #rm. Starting from “high” entry costs, more speci!cally the shaded area marked with G0 in Fig. 1, a reduction in trade costs such that we cross the ii-line or a reduction in green!eld costs such that we cross the i-line turns the pro!tability of a merger from negative to positive. The reason is that economic integration changes the no-merger market structure from duopoly to triopoly, thus reducing the reservation price of the target !rm. Starting from “intermediate” entry costs, more speci!cally the shaded area marked XX in Fig. 1, a reduction in green!eld costs such that we cross the iv-line induces one of the foreign !rms to choose green!eld investment as entry mode in the no-merger case. Again, this reduces the reservation price of the target !rm and turns the profitability of a merger from negative to positive. Starting from the same area, a reduction in trade costs such that we cross the v-curve reduces the competitive pressure on the merged entity and thus induces a cross-border merger. The reason is that the optimal post-merger entry mode of the outside !rm changes from green!eld to exports. The e3ect of economic integration on the pro!tability of a merger is explored in more detail in Fig. 2. Here, I consider the e3ect of a reduction in trade costs on the pro!tability of a merger for a given level of green!eld costs, f = 0:07. The reservation price of the target !rm, c , is illustrated with a bold line, and the willingness to pay of !rm j; ˆj+c − j , is illustrated with a !ne line. Region (i; ii; v) is characterized by duopoly, with one foreign investor competing with !rm c. We know from (23) that in this region, the reservation price of !rm c equals j c the willingness to pay on part of the active !rm; ˆj+c G − 0 = G0 . A reduction in t such 1 that we cross the ii-line, de!ned by t = 3 , leads to a reduction in c . The reason is that the no-merger market structure changes from duopoly to triopoly, with one foreign investor and one foreign exporter. Intuitively, by causing more intense competition in the market, economic integration raises the gains from increased market concentration through merger.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1221
0.14
^j πj+c- Π 0.12
0.1
πc
0.08
0.06
vi
v
iv
ii
(i,ii,v)
0.04 0.1
0.2
0.3
0.4
0.5
t Fig. 2. Reservation price and willingness to pay.
Consider next a reduction of t such that we cross the iv-line. This changes the sign of gj from positive to negative. The reason is basically that economic integration in this case reduces the intensity of competition in the absence of a merger, by changing the market structure from one investor and one exporter to a situation in which both active !rms export. This reduces the pro!tability of a merger by increasing c and j , thus shifting both curves in Fig. 2. A further reduction in t such that we cross the v-line again changes the sign of gj , this time from negative to positive. The reason is that economic integration now leads to a radical reduction in post-merger competition, as the outside !rm changes the entry strategy from green!eld investment to exports. The resulting reduction in the business stealing e3ect by the outsider leads to a sharp increase in ˆj+c , and hence a positive shift in the willingness to pay function illustrated in Fig. 2. Finally, a reduction in t such that we cross the vi-line makes a merger unpro!table by increasing the business stealing e3ect. As is evident from Fig. 2, to the left of the vi-line, this e3ect dominates the fall in c caused by lower trade costs and hence tougher competition. Of course, with zero trade costs we are e3ectively in a perfectly integrated market, and we know from the literature that a merger between two !rms would not pro!table in that case. Fig. 2 illustrates economic integration in the form of a reduction in trade costs. Similar mechanisms can, however be derived by considering a reduction in green!eld costs, f. Consider, for instance a reduction in f such that we cross the v-line and enter region (i; ii; v) in Fig. 1. This changes the sign of gj from positive to negative by changing the entry mode of the outside !rm from exports to green!eld investment, thus reducing ˆj+c . A continued reduction in f such that we cross the i-line allows pro!table
1222
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
investment by both active !rms. This makes the merger pro!table by intensifying competition in the no-merger case and thus reducing the reservation price of the target !rm. A reduction in f such that we cross the vii-line makes a merger unpro!table by making the alternative to an acquisition as entry mode more pro!table for !rm j, hence increasing j . Given that a merger is pro!table, is it more pro!table to be inside or outside the merger? This is an interesting question, although not at the core of the present analysis. It is straightforward to demonstrate that: Proposition 2. Given that the acquisition price is de#ned by the reservation price of the target #rm, the pro#ts of the inside #rm dominate those of the outside #rm when trade costs and green#eld investment costs are relatively high. Comparing ˆk , which is the post-merger pro!ts of the outside !rm, with ˆj+c − c , which is the post-merger pro!ts of the acquiring !rm, net of the acquisition price 3 c ; ˆj+c − c ¿ ˆk is satis!ed only for t ¿ 14 above the v-curve. Intuitively, when exports is the relevant alternative to acquisition as entry mode into market C, high trade costs reduces both the business stealing e3ect and the pro!ts associated with the alternative entry mode. In combination, this makes it more pro!table to be an insider than an outsider to the merger. When both a and b wish to acquire c, it is reasonable to assume that a bidding contest will arise, thus raising the acquisition price. In a standard auction model we would expect the equilibrium price of the asset to be such that the bidding !rms are indi3erent between acquiring c or exporting to the market. If this is the case, then naturally the post-merger pro!ts of the inside !rm, net of the acquisition price, will be equal to those of the outside !rm. 4. Extensions: Asymmetries So far in the analysis, I have assumed that the two active !rms are entirely similar. In what now follows, I allow for asymmetries between the active !rms along two dimensions. First, in terms of cost eOciency. Second, in terms of their initial location. There are of course a number of ways in which the three !rms could di3er, even restricting our attention to eOciency and location. Rather than carrying out a full analysis of how asymmetries in these two dimensions would a3ect the model, I use a couple examples that capture some interesting e3ects. 4.1. Technological asymmetry Consider !rst the case where there is one technological leader, !rm a, with the other two !rms having identical, but inferior technology to that of a. Maintaining the assumption of zero marginal costs for a, let ¿ 0 represent the common marginal production costs of b and c, which can also be interpreted as the technology gap between the more and less advanced !rms. Given that !rm a can apply its superior technology to c in case of acquisition, this modi!cation of the model clearly increases the pro!tability of a merger between a and c. It does so by adding an eOciency gain
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1223
to the merger. In fact, if the technology gap is suOciently large, there is an incentive for a to acquire c even for zero trade costs and zero green!eld costs. We can easily see this by noting that the marginal production cost a3ects pro!ts exactly as the trade cost t does. Let t = 0 or f = 0 so that we are e3ectively in a single-market situation. In this case, a merger between a and c is pro!table to the merging parties if (1 − 2)2 (1 + )2 (1 + 2)2 ¿ ⇒ ¿ 0:07; − 16 16 9
(24)
where the !rst term represents the post-merger pro!ts of the merged entity, the second term the no-merger pro!ts of !rm a and the third term the no-merger pro!ts of !rm c. Hence, for ¿ 0:07, !rm a will acquire c even for zero trade costs or zero green!eld costs. The result that asymmetries in technology can make a merger pro!table even in a closed economy setting is not novel, see for instance Barros (1998). 4.2. Location asymmetry Consider next asymmetry in the initial location of the two active !rms. Assume that initially, i.e. prior to any investment decision, b and c are located in the same market with a being the only foreign !rm. In this case, market entry is of course an issue only for !rm a. 7 As in the benchmark case, the !rms are assumed to have identical marginal production costs. This modi!cation of the analysis allows us to deal with the interesting issue of the nationality of the acquiring !rm. I shall refer to this scenario as the “nationality” case. When should we expect to see cross-border mergers and when are mergers between two !rms located in the same market more pro!table? Fig. 3 illustrates the outcome of this scenario. The de!nitions of the lines dividing the various market structures in Fig. 3 are the same as in the benchmark scenario. X means that !rm a chooses exports; G that a chooses green!eld investment; 0 that a stays out of the market; Aa (X ) means that a acquires c, the alternative to acquisition for a being exports; Aa (G) that a acquires c, the alternative being green!eld investment; Ab X means that b acquires c, the response of a being exports. The no-merger equilibrium market structure is relatively straightforward. Again, low trade costs relative to green!eld costs induces a to export, indicated by X in Fig. 3, whereas high trade costs relative to green!eld costs induces investment, G. When the cost of both kinds of entry are high, !rm a stays out, indicated by 0. As in the benchmark scenario, Fig. 3 shows that economic integration may trigger cross-border acquisitions. The acquisition of c by a is indicated by Aa . It can be shown that for “low” entry costs, it is never pro!table for !rm b acquire c. This is clearly the case under the v-curve: Here, an acquisition by b would be followed by a green!eld investment by a. Hence, !rm b has nothing to gain from acquiring c itself rather than leaving the acquisition to b, since in either case the post-merger market structure will be characterized by local duopoly competition. 7 The general insights derived from this analysis would apply also in a situation where b faces lower trade costs than a, and not necessarily zero trade costs as in the present example.
1224
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
0.12
ix
AbX v
vi ii
0.08
0
Aa(X)
f
X i
iii Aa(G)
0.04
vii G
0 0.1
0.2
0.3
0.4
0.5
t Fig. 3. The nationality case.
The situation is slightly di3erent above the v-line. In this case, if b acquires c, the response of a would be exports. As such, this is bene!cial to !rm b who prefers to have a at a distance from the market. However, it is possible to show that for “low” trade costs, more speci!cally to the left of the ii-line, this is not a pro!table option for !rm b. To see this, note that above the v-line and to the left of the ii-line, the no-merger entry mode of !rm a is exports. Hence, the reservation price of !rm c c c is given by GX . The value added from a merger by b is therefore ˆbX − GX . The b alternative for b is to let a acquire c, resulting in pro!ts G0 for b. It can be shown that c b ˆbX − GX − G0 ¡0
for all t ¡ 0:5:
(25)
On the other hand, when entry costs are “high”, i.e. to the right of the ii-line and above the v-curve, it may indeed be pro!table for b acquire c. For a, an acquisition is here not a pro!table option. If b does not acquire c, !rm a stays out of the market. c Hence, the reservation price of c is given by G0 . In Fig. 3, the critical level of trade costs is given by the ix-line, which can be found as c b ix : ˆbX − G0 − G0 ¡ 0 ⇒ t ¡ 0:41:
(26)
Hence, for t ¿ 0:41 and above the v-curve, !rm b !nds it pro!table to acquire c. Intuitively, the gain to the local !rm, !rm b, from this merger is that it replaces a local competitor with a foreign based exporter. For suOciently high trade costs, the business stealing e3ect is suOciently small to make the merger pro!table. In Fig. 3, the equilibrium market structure in this case is indicated by Ab X , i.e. acquisition by b followed by exports by a.
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
1225
5. Concluding remarks The results presented here demonstrate that the relation between economic integration and the pro!tability of cross-border mergers is complex. One possibility is that economic integration may trigger cross-border mergers, contrary to what one perhaps would think from reading the theoretical literature on mergers in closed economies. The reason is that economic integration may intensify the pre-merger competition in the market, thereby reducing the reservation price of the target !rm. In addition, economic integration in the form of lower trade costs may reduce the post-merger business stealing e3ect as the outside !rm chooses exports rather than green!eld investment. The benchmark version of the model deals with the standard case of symmetric !rms. An extension to the model deals with asymmetries between !rms. It shows that a technologically advantaged !rm has a stronger incentive to acquire a foreign !rm than its less advanced rival, given that technology can freely be applied to the target !rm. Finally, I demonstrate that when there is an asymmetry in the location of the two active !rms, entry costs is an argument in favor of cross-border acquisitions rather than national mergers. However, if entry costs are suOciently high, the price of the target !rm may be too high for the foreign !rm to !nd an acquisition pro!table. In this case, a merger between the local !rms is more likely. There are certainly a number issues related to how asymmetries between !rms a3ect equilibrium entry modes that are not explored in the present paper. Does a technological leadership necessarily raise the pro!tability of acquisition as entry mode? What role does the possibility of transferring technology from the mother !rm to the acquired !rm play in the entry decision? How would trade costs a3ect the analysis if the acquisition is an export platform investment rather than a market seeking investment, i.e. if the acquired plant is used primarily for export purposes? These interesting questions are left for future research. Acknowledgements I would like to thank Lars Persson, Lars SHrgard and two anonymous referees for useful comments. Remaining errors are mine. References Barros, P., 1998. Endogenous mergers and size asymmetry of merger participants. Economics Letters 60, 113–119. Barros, P., Cabral, L., 1994. Merger policy in open economies. European Economic Review 38, 1041–1055. Deneckere, R., Davidson, C., 1985. Incentives to form coalitions with Bertrand competition. Rand Journal of Economics 16 (4), 473–486. Fridolfsson, S.-O., Stennek, J., 2000. Should mergers be controlled? IUI Working Paper No. 541. The Research Institute of Industrial Economics, Stockholm. GKorg, H., 2000. Analysing foreign market entry: the choice between green!eld investment and acquisitions. Journal of Economic Studies 27 (3), 165–181. Gowrisankaran, G., 1999. A dynamic model of endogenous horizontal mergers. Rand Journal of Economics 30 (1), 56–83.
1226
K. Bjorvatn / European Economic Review 48 (2004) 1211 – 1226
Head, K., Ries, J., 1997. International mergers and welfare under decentralized competition policy. The Canadian Journal of Economics 30 (4), 1104–1123. Horn, H., Levinshon, J., 2001. Merger policies and trade liberalization. Economic Journal 111 (470), 244– 276. Horn, H., Persson, L., 2001. The equilibrium ownership of an international oligopoly. Journal of International Economics 53, 307–333. Inderst, R., Wey, C., 2002. The incentives for takeover in oligopoly. CEPR Discussion Paper 3163, Centre for Economic Policy Research, London. Kabiraj, T., Chaudhuri, M., 1999. On the welfare analysis of a cross-border merger. The Journal of International Trade and Economic Development 8 (2), 195–207. Lommerud, K.E., Straume, O.R., SHrgard, L., 2004. Downstream merger with upstream market power. European Economic Review, Forthcoming. NorbKack, P.J., Persson, L., 2004. Privatization and foreign competition. Journal of International Economics 62 (2), 409–416. Salant, S.W., Switzer, S., Reynolds, R.J., 1983. Losses due to merger: the e3ects of an exogenous change in industry structure on Cournot-Nash equilibrium. Quarterly Journal of Economics 98, 185–199. Stigler, G.J., 1950. Monopoly and oligopoly by merger. American Economic Review 40, 23–34. Tombak, M.M., 2002. Mergers to monopoly. Journal of Economics and Management Strategy 11 (3), 513– 546. UNCTAD, 2000. World Investment Report 2000: Cross-border Mergers and Acquisitions and Development, United Nations.