The Great European Fire-Sale?
Evidence of fire-sale acquisitions in the EMU during the last crises
Thijs Kruimer*
Master Thesis
Groningen, August 17
th2015
Supervisor: Dr. H. Vrolijk
Co-assessor: Dr. W. Westerman
University of Groningen
Faculty of Economics and Business
MSc International Financial Management
University of Uppsala
Faculty of Social Sciences
MSc Business and Economics
*Student number: s1819712 Amstelstraat 9 9725 KT GRONINGEN
2
The Great European fire-sale:
Evidence of fire-sale acquisitions in the EMU during the last crises
ABSTRACT
During previous crises, foreign investors increased their acquisitions in crisis struck economies, a phenomenon which is referred to as fire-sale acquisitions. In this paper I investigate if similar acquisitions patterns can be observed in the EMU during the global and euro-crisis. This paper is the first to extensively investigate the share of cross-border acquisitions from different country regions. I find that during the euro-crisis there is evidence of fire-sale acquisitions in the EMU. These findings are more outspoken for EMU countries hit hardest by the crisis but also hold for EMU countries like Germany.I find that similar patterns can be observed during the global-crisis. These results are robust to alternative empirical specifications, different crisis demarcations, and inclusion of macroeconomic controls.
Part two of this study investigates the role of legal- and financial environments of crisis-target countries. For the legal environment no clear results come forth, but a better developed financial environment seems to protect insiders against outsiders during the euro-crisis.
Key words: Cross-border acquisitions, fire-sale, euro-crisis, Eurozone, legal environment, financial environment
3 Table of content 1 Introduction ... 4 2 Theoretical background ... 5 2.1 Fire-sale literature ... 5 2.1.1 Fire-sales of assets ... 6 2.1.2 Fire-sale FDI ... 7
2.1.3 Fire-sale FDI and the euro-crisis... 9
2.2 Fire-sale acquisition patterns ... 10
2.3 Acquisitions targets during crises and the legal and financial environment ... 14
2.3.1 Acquisitions and the legal and financial environment ... 14
2.3.2 The legal environment ... 15
2.3.3 The financial environment and likely targets ... 16
3 Data and methodology ... 17
3.1 Data ... 17
3.2 Methodology ... 20
3.2.1 The Two-Sample t-Test Assuming Unequal Variances ... 20
3.2.2 The Linear Probability Model ... 21
3.3 Data set summaries ... 24
4 Empirical Results ... 26
4.1 Cross-border acquisitions in the EMU... 27
4.1.1 The share of cross-border acquisitions in the EMU ... 27
4.1.2 The cross-border acquisitions shares of different groups in the EMU ... 30
4.2 PIIGS, EMU* and Germany ... 31
4.3 Results for the legal and financial environment ... 33
5 Discussion ... 35
5.1 Examining possible patterns of fire-sale FDI ... 35
5.1.1 Fire-sale acquisition evidence during the global-crisis ... 36
5.1.2 Fire-sale acquisition evidence during the euro-crisis ... 37
5.2 The legal and financial environment ... 38
5.3 Contribution and practical relevancy ... 38
5.4 Limitations ... 40
6 Conclusions and future research ... 40
7 Appendices ... 43
4
1 Introduction
The recent Global Financial Crisis (global-crisis) and the European Sovereign Debt Crisis (euro-crisis) have strongly affected the global economy. Where the global-crisis affected most of the developed economies, the euro-crisis particularly manifested itself in the European Monetary Union (EMU). The euro-crisis most severely affected those crisis-stricken countries identified by Mundell (2012) as the ‘PIIGS’, an acronym for Portugal, Ireland, Italy, Greece and Spain. Of these PIIGS, Greece was affected most harsh and in order to prevent its default the country accepted several rescue packages from the International Monetary Fund (IMF) and the EU. The loan agreement included Greece’s commitment to spending reforms and austerity measures. With a recently new bailout package on the way, the Greece government for now has adverted default. However, most likely the financial hardship will continue for the country while its governments unsuccessfully tried to renegotiate with the IMF and EU for more lenient loan conditions (Beekman, 2014; Chorley, 2015; Reiner, 2015; Schmidt, 2014; Zadelhof, 2014; Trayner et al., 2015).
Although, the worst of the crises seems over for most countries in Europe, the effects of it, especially on Greece, are still strongly felt and dominate the media to this day. The euro-crises sparked what the media called ‘a fire-sale’: a large number of assets sales at low prices. Large newspapers opened with headlines such as: “Bargain Hunting in Greece” by the Wall Street Journal, referring to the cheap sale of Greece state-owned assets (Lawton & Stevens, 2011); “Portugal to hold fire-sale of state assets” by The Guardian (Tremlett, 2012); “Everything Must Go! The Great European Fire Sale” as published by The Independent (Bawden & Cooper, 2012); and more recently The Great Greece Fire Sale (Rankin and Smith, 2015).
The PIIGS might have been hit the hardest; nonetheless the entire European business environment suffered from decreased investment possibilities and lower growth rates. Credit supply within the Eurozone has sharply fallen and is slowly recovering. Simultaneously, business investments in the EU dropped by 20 percent from 2008 to 2009 (Camerinelli, 2012). These financial problems led to a decline in the number of acquisitions within the EMU. It is in this context that I investigate how the two financial crises affected acquisition patterns compared with the pre-crisis period into the EMU.
5
narrative of insiders and outsiders. To the best of my knowledge, this has not been done before so extensively. As one of the first, I will also provide a start to investigate how legal and financial environments matter in explaining acquisition patterns during a fire-sale.
By looking into a sample of over 97 thousand acquisitions deals in the EMU during the period from 1999 to 2013, I am able to find evidence supporting the theory of fire-sale FDI during the euro-crisis in the EMU and suggest that fire-sale acquisition patterns can also be observed during the global-crisis. Additionally, I investigate the relation between the legal and financial environment in target countries with the likelihood that acquisitions are cross-border during crises. Consistent results are not found for the legal environment but evidence on the financial environment suggests a negative relation with the likelihood that an EMU firm is acquired by a firms from outside the EMU.
The rest of this paper is organized as follows; in chapter 2 the relevant literature of fire-sale FDI is discussed and expectations and hypotheses are presented; chapter 3 described the data and methodology; chapter 4 includes the statistical results of my data analyses; chapter 5 discusses the outcomes and contribution of these outcome and the limitations of this study; and chapter 6 concludes.
2 Theoretical background
This chapter has three main sections: Section 2.1 provides an overview of the fire-sale literature for my study; Section 2.2 presents expectations regarding fire-sale acquisition patterns; Section 2.3 introduces the legal- and financial environment and their relating hypotheses.
2.1 Fire-sale literature
6 2.1.1 Fire-sales of assets
Schleifer & Vishny (1992; 2011) describe the existence of fire-sales in the market for real and financial assets as a result of liquidity constraints. They borrow the term fire-sale to indicate the cut-rate prices which are associated with smoke damaged merchandise sold after a fire. These cut-cut-rate prices can also be observed in asset markets troubled by financial distress.
Schleifer & Vishny (1992; 2011) make a distinction between insiders and outsiders. Insiders are firms operating in the same line of business, e.g. competitors, whereas outsiders are in other lines of business. Under normal, i.e. non-crisis, conditions assets with a degree of specificity can be sold to these insiders who are willing to pay a price close to the asset its value at best use. Outsiders are firms who are not able to correctly value the specificity of an asset, or have no use for it, and are thus willing to purchase the specific asset only at a lower price.
There are situations in which a firm might be forced to sell some of its specialized assets to improve its liquidity position, e.g. to meet its payment requirements or because of collateral borrowing contracts. In such a situation the firm would want to sell these assets to insiders, since they are willing to pay a price close to the value of the asset at its best use. However, if these insiders are also financially constrained, e.g. due to an industry-wide crisis, only outsiders might be able to buy the assets, but at a lower price. This dislocation of the asset price to its value at best use is strengthened if other insiders are also forcibly selling similar assets. This leads to an increased supply of the specialized assets, which becomes further undervalued, triggering a downward spiral of decreasing prices and increased supply. The problems the troubled firm face can become even more severe when these assets serve as collateral for an asset-backed loan, which makes attracting credit more difficult. This phenomenon of industry wide forced asset sales at a dislocated price is referred to as a fire-sale.
7 2.1.2 Fire-sale FDI
Fire-sale FDI, in short, is the phenomenon that FDI increases during crises. Krugman (2000) coined the term fire-sale FDI, in the wake of the Asian crisis. In this paper Krugman noticed that a “massive flight of short-term capital and large-scale sell-offs of foreign equity holdings, has at the same time been accompanied by a wave of inward direct investment” (p. 43). This observation, he argues, also seems to hold for other crises, such as the Latin American Crisis of 1995. These opposing capital flows seem counter intuitive, since they normally move in the same direction. Arguably, as foreign investors and creditors lose confidence in the economy of the crisis-stricken country, this loss of confidence should also be exhibited in the FDI flows.
Krugman’s (2000) theory of fire-sale FDI has similarities with how Schleifer & Vishny (1992; 2011) use the term fire-sale. They both argue that because of financial constraints assets can become undervalued. The theory of fire-sale FDI predicts that when domestic sectors face financial constraints due to tighter domestic capital markets, the prices of the domestic firms fall. Foreign investors have access to capital markets and do not face serious constraints. This leads to opportunities for foreign investors to buy domestic firms at an undervalued price.
An important distinction with the fire-sale of assets theory is that in the theory of fire-sale FDI identifies insiders and outsiders on a geographical level, rather than on the industry level. In the insider/outsider terminology of Schleifer & Vishny (1992; 2011) insiders are identified as efficient buyers, and outsiders are identified as asset buyers who are not able to use the asset as efficiently. The nationality of the target and acquirer does not play a role. In the fire-sale FDI theory all firms from a crisis stricken country, or region, are identified as the insiders while all the foreign buyers, or all buyers from outside the crisis region, are identified as the outsiders. The possible difference arising is that according to the theory of fire-sale FDI an outsider might well be identified as an industry insider according to Schleifer & Vishny (1992; 2011). An outsider in the fire-sale FDI literature might even be a technologically superior firm which can use the specific asset more efficiently than an insider firm (Aguiar & Gopinath, 2005; Alquist et al. (2013).
8
“flipping the deal” which can be observed during fire-sale FDI. Alquist et al. (2013) are the first to test if acquisitions are made for these short-run motives but find little evidence, suggesting that fire-sale outsiders might predominantly be industry insiders from other countries.
Studies into the workings of fire-sale FDI are a rather new field of research. Since FDI includes more than only acquisitions flows, most empirical studies investigate changes in acquisitions shares and look for evidence of fire-sale acquisitions. The most notable studies which provide empirical proof of increases in the share of outsider acquisitions in crisis times are the papers by Aguiar & Gopinath (2005), Acharya et al. (2011) and Alquist et al. (2013). These studies investigate Krugman (2000) his observations on rising FDI, or acquisitions, in crisis regions and simultaneously falling short-term capital and large-scale sell-offs of foreign equity holdings, i.e. portfolio investments, in the Asian Crisis of 1997.
In the Asian crisis all firms from Asian countries which suffered because of the crisis are insiders and all investing firms from outside this region are outsiders. Aguiar & Gopinath (2005) find that between 1996 and 1998 the number of domestic acquisitions decreased by 27 percent, while the number of foreign acquisitions increased by 91 percent. This increase in the share of cross-border acquisitions is accompanied by decreases in the value of foreign portfolio investments to around 0 and withdraws of international bank credit (p. 1-2). Moreover, Acharya et al. (2011) and Alquist et al. (2013) also find supporting evidence that indicates that the share of acquisitions made by outsiders during economic crises in Asia, Latin America and South Africa increases while foreign portfolio investments decreased significantly.
Bogack & Noy (2012), analysing all crises for the period 1987-2009 in developing economies, find results which do not support the theory of fire-sale FDI. Poulsen & Hufbauer (2011) find contradictory evidence for developed and developing economies in their study1.With respect to crises in developing economies they find increases in FDI inflows, which conflicts with Bogack & Noy (2012, p. 26-27), while during crisis in developed economies they find a decrease in FDI inflows. With respect to the theory of fire-sale FDI this observation does not say anything about the share of cross-border
1
9
acquisitions. Desbordes and Wei (2012) even argue that the theory of fire-sale FDI could only hold in developing countries.
2.1.3 Fire-sale FDI and the euro-crisis
Recent studies by Munnichs (2014) and Weitzel et al. (2014) also investigate evidence for fire-sale acquisitions within the EU during the recent crises. Munnichs (2014) finds support for the theory of fire-sale FDI in the PIIGS during the euro-crisis. He finds evidence that during the euro-crisis firms from the U.S. and what he calls “Rest of the World”, i.e. outsiders, increased their share of M&A’s in the PIIGS, while insiders firms from the original Eurozone countries minus the PIIGS, i.e. insiders, did not increase their share. The paper by Weitzel et al. (2014) investigates acquisition patterns within the EU without making a distinction between the global-crisis and the euro-crisis. They do not find convincing evidence of fire-sale acquisitions, and thus the fire-sale FDI theory.
My study differs from these two studies on several important aspects. With respect to Munnichs (2014) which only studies the PIIGS, my study includes all EMU countries because the whole EMU region suffered during the euro-crisis. I also investigate the acquisition patterns of the PIIGS and the rest of the EMU separately, to study if similar acquisition patterns between the two groups can be observed. Weitzel et al. (2014) couple the global-crisis and euro-crisis as one crisis-period and only include acquisitions with both an EU-target and EU-acquirer. My study differs on both of these aspects, with acquisitions from outside the EU included and the two crises studied separately. I argue that these differences are of importance.
10
banking sector in the EMU and the flow of bank credit to the private sector became impaired. Other regions, most importantly the other Triad regions, where not facing this crisis (Allen & Moessner, 2013; Gorea & Radev, 2014). Because of the differences between these two crises, acquisitions patterns in the EMU are likely to differ between these crises.
The insider/outsider narrative in the fire-sale FDI literature leads me to argue that the exclusion of non-EU acquirer acquisitions in the EMU by Weitzel et al. (2013) might not capture the full story. Because the global-crisis affected most countries world-wide (Mariana, 2011) you could argue that nearly all countries were insiders during this crisis. On the other hand, Payne (2012) finds that the BRIC-countries, i.e. Brazial, Russia, India and China, did not suffer so severely from the global-crisis and improved their position on several economic fronts relative to the Triad regions, e.g. their share of cross-border acquisitions. This could indicate that the BRIC-countries could be considered outsiders. During the euro-crisis all regions other than the EMU did not face this crisis and could be considered outsiders, e.g. the BRIC and Triad. This means that Weitzel et al. (2013) affectively exclude the most important outsiders and predominantly focus on the acquisitions from and to insiders, i.e. the EMU, and EU outsiders which arguably were also partly impaired. This is why in this study I will include all acquisitions in the EMU and differentiate between the two crises.
2.2 Fire-sale acquisition patterns
In this study I am interested in possible fire-sale FDI effects during the recent global-crisis and euro-crisis. Munnichs (2014) finds evidence for the fire-sale FDI theory in the PIIGS, during the euro-crisis but not for the global-crisis, but excludes other EMU countries as targets in his study. Weitzel et al. (2014) find little evidence of fire-sale acquisitions, in a study were both crises are coupled and most acquisitions by outsiders are ignored. As discussed in section 2.1.3, I argue that excluding cross-border acquisitions from countries located outside the EMU, does not provide the total picture.
11
of acquisitions of each group, to all acquisitions in country A. At T+2 the crisis economies starts to recover and the share of domestic acquisitions and the share of cross-border acquisition by crisis insiders starts to increase again.
These predicted patterns show that at the start of the crisis, the share of domestic acquisitions is expected to decline and the share of the cross-border acquirers to increase. Of the cross-border acquirers we can expect that the share of insiders will also fall while the share of outsiders will rise.
Figure 1: Predicated fire-sale acquisition pattern in a crisis region
This figure shows the share of three different groups in the total acquisitions of country A. The horizontal axes shows the progression of time, where T denotes the outbreak of a crisis, T-1 one year before the crisis and t+1 one year after the start of the crisis, etc. At T+2 the economies in the region and of country A start to recover from the crisis.
In terms of EMU countries during the euro crisis we can predict, from the theory of fire-sale FDI, that the share of domestic acquisitions in individual EMU countries diminishes, similar to country A in figure I. At the same time other countries during the euro-crisis increased their share of cross-border acquisitions into EMU countries. During the global-crisis this expectation is less obvious2.
The effects of the euro-crisis spread among the EMU, making the whole EMU an insider in this crisis. Figure I visualizes the prediction that the share of acquisition made by insiders also decreases, next to the share of domestic acquisitions, and that only outsiders will increase their share of cross-border
2
12
acquisition in crisis countries. With respect to figure 1, country A can be any country in the EMU, e.g. Greece or Germany, the insiders are all other countries in the EMU and the outsiders are all countries outside the EMU.
What is important to emphasize is, that the countries which should be identified as insiders or as outsiders differ during both crises. An important difference between the euro-crisis and the global-crisis is that the euro-global-crisis specifically manifested itself in the EMU. This means that other countries should not be affected so severely by this crisis. Outsiders during the euro-crisis are those countries which did not face the same liquidity problems, and can thus be identified as all non-EMU countries.
During the global-crisis the most important regions for international acquisitions, the Triad regions, were all severely affected by the crisis, leading to economy-wide liquidity problems in all these countries, and thus they can all be considered insiders. Some countries and regions were less affected than others during this crisis, e.g. in the degree of bank lending problems arising from the financial downturn. These countries less affected by the crisis did not face those severe liquidity problems as faced by the insiders (Payne, 2012, p. 146 & 148). Such countries could be considered outsiders during the global-crisis.
Examples of such countries which can be considered outsiders during the global-crisis are the BRIC-countries, i.e. Brazil, Russia, India and China (Payne; 2012). Wu (2014) analyses the shares of cross-border M&A activity undertaken by firms from emerging countries and finds a steady increase from the beginning of this century, which proceeds throughout the crises. Although these countries did suffer from an economic downturn during the crisis in various degrees, through trade and investment linkages with the rest of the world, existing financial restrictions imposed by their governments before the outbreak of the crisis limited the liquidity problems which their firms faced during the global-crisis (Payne, 2012, p. 148). This resulted in a situation where firms from emerging economies might have had the opportunity to increase their share of Acquisitions into the insiders of the global-crisis as money supply was not a major issue in these countries during the crisis periods (Payne, 2012, p. 144).
13
The first group consist of the EMU countries. In this study I am interested in the developments of acquisition patterns in this group during the two crisis periods. The theory of fire sale FDI predicts that during a regional crisis we should find evidence of a decreased share of domestic acquisitions and insider border acquisitions. These decreases should be offset by increases in the share of cross-border acquisitions targeting these countries from other country groups, i.e. outsiders. During both crises this group can be considered an insider.
The second group of which I will study the cross-border acquisitions targeting the EMU are the non-EMU EU-countries3, which are identified by Other-EU. I make this distinction because of the strong economic linkages within the EU. As the euro-crisis unfolded a distinguishing factor between these EU countries and the EMU countries is that these EU countries still had their own currencies. Having an own currency gives a country the possibility to set financial policies tailored to their own economy, such as extending credit. This is of importance, especially for the asset and liability prices of small and medium sized enterprises which are usually denominated in the residing country domestic currency. Through a common currency the economic linkages between EMU countries are stronger and as a result the crisis spread more contagious among them (Missio & Watzka, 2011). I consider this group an insider during the global-crisis and an outsider during the euro-crisis. I also investigate Norway and Switzerland, identified by NS, combined as a separate third group, because of similar economic arguments, their enclosure within the EU region and their relative large share of cross-border acquisitions in the EMU. The EU countries which belong to the EMU and EU are listed in appendix A
The fourth group which I distinguish are the Triad-regions, minus the European countries. This region is particularly important for cross-border acquisition amongst the countries belonging to this group. During the global-crisis the Triad-regions can be considered insiders whereas during the euro-crisis they can be considered outsiders.
The fifth and sixth groups consist of the BRIC-countries and other countries Wu (2014) identifies as the Emerging Economies EE, listed in appendix B. I distinguish between these groups because the BRIC-countries are considerably larger economies than the other EE and because Payne (2012)
3
14
describes the role of the BRIC-countries during the crises as one which challenges the authority of the traditional economic powers. During both crises these groups can be considered outsiders.
The final group consists of countries which can be identified and grouped as the Rest Of the World (ROW). ROW-countries consist of all countries which do not belong to the above mentioned groups. During the euro-crisis the ROW-countries can predominantly be considered outsiders; however this classification during the global crisis is unclear because of the wide variety of included countries.
Table 1: Country group insider/outsider classification Country group Global-crisis Euro-crisis
EMU Insider Insider
Other-EU Insider Outsider
NS Insider Outsider
Triad Insider Outsider
BRIC Outsider Outsider
EE Outsider Outsider ROW Unclear Outsider This Table shows the insider/outsider classification of each separate country group during the two crises.
The predicted acquisitions patterns of insiders and outsiders during a crisis, which follow from the theory of fire-sale FDI, leads to the following expectation:
Expectation I. The acquisition shares of domestic acquisitions and insider cross-border
acquisitions decreased during the euro-crisis in the EMU while the share of outsider cross-border acquisitions increased.
2.3 Acquisitions targets during crises and the legal and financial environment
The second objective of this thesis is to explore if differences in cross-border acquisition shares during crisis periods can be explained by the legal and financial environments of the target countries. Section 2.3.1 introduces the legal and financial environment. Next, I introduce my hypotheses regarding the legal environment (section 2.3.2) and the financial environment (2.3.3).
2.3.1 Acquisitions and the legal and financial environment
15
macro illiquidity. This in turn affects the possible desirability of firms as acquisition targets. Studies investigating how acquisition patterns depend on these environments have done so during relatively stable times and not during crises. In this paper I will do both.
The paper by Hyun & Hyuk (2010) incorporates both the quality of legal institutions in a country and the development of the financial market as determinants of cross-border M&A volume. I follow an approach using similar variables. I will investigate if the legal and financial environments help determine if firms from different EMU countries are more likely targets for acquirers outside the EMU. I will investigate if these relations hold during the euro-crisis.
2.3.2 The legal environment
Differences in the legal environment matter in the acquisition target decision making process of acquirers. In general, acquirers favour legal environments which provide market-friendly policies and institutions to foreign investors. These environments enhance the confidence of foreign investors in the private sector and their willingness to invest (Wu, 2014, p 110-111).
In my analysis of the legal environment I will broadly follow La Porta et al. (1997)4. This article is referenced commonly throughout literature on firm capital structures and literature on determinants of cross-border acquisitions (see e.g., Demirgüç-Kunt & Maksimovic, 1999; Rossi & Volpin, 2004; Hyun & Hyuk, 2010; Rose et al., 2012; Wu, 2014). La Porta et al. (1997) distinguish between three important determinants in the legal environment, namely: origin of the legal system, i.e. common law or civil law5; quality of legal investor protection; and quality of law enforcement. I will only use the last two determinants because Ireland is the only common law country within the EMU6.
In an extension to the La porta et al. (1997) study, Rossi & Volpin (2004) find that with higher investor protection, the probability that a given acquisition in a country is cross-border, rather than domestic, decreases. More specifically, they find that the share of cross-border acquisitions targeting
4
The article is related to Schleifer & Vishny (1997) which are also the co-authors of the La Porta et al. (1997) paper.
5 Common law is made by judges and subsequently incorporated into legislature. In contrast, laws in civil law countries are
created through a scholar and legislator tradition (La Porta et al., 1997).
6 With respect to the origin of the legal system, this EMU study does not lend itself to test this empirically. La Porta et al.
16
domestic firms is higher in countries with a poorer legal environment. The same can be argued for the quality of law enforcement, with a higher quality leading enhancing the confidence of foreign investors in the private sector and their willingness to invest.
Based on these findings, I would expect that EMU countries with higher quality legal environments have a lower share of inward cross-border acquisitions during non-crisis periods. More specifically, in this EMU orientated study I will investigate if the share of cross-border acquisitions in EMU countries by non-EMU acquirers is larger for those countries which have a relatively lower quality legal environment. This leads to my first hypothesis:
Hypothesis I The probability that a given deal in the EMU has a non-EMU acquirer
dddddd decreases with the quality of the legal environment of the EMU target country.
I will explore this relationship during both the pre-crisis period and the euro-crisis period holds.
2.3.3 The financial environment and likely targets
“Presumably, the willingness of an entrepreneur to sell his equity, or to assume debt, depends to a large extent on the terms at which he can obtain external finance” (La Porta et al., 1997, p. 2) Firms can become financially constrained when the terms for raising external funds become unacceptable. Di Giovanni (2005) argues that good terms, low barriers to raise external finance in the country of the acquirer, can provide a more favourable environment for buyers to gain access to capital for cross-border acquisitions. The good terms result from larger and mature stock markets and/or easier access to bank credit. I will explore if unfavourable terms in a country, lead to better opportunities for border acquirers to target firms in that country. This should be represented by a higher share of cross-border acquisitions in that country.
In order to estimate the financial market developments I follow Di Giovanni (2005) and Hyun & Hyuk (2010) in their selection of variables for financial market developments. Di Giovanni (2005) examines the influence of financial deepening in determining cross-border M&A flows. He considers the stock market capitalization, relative to GDP, and the amount of credit provided by banks and other financial institutions to the private sector, relative to GDP, as financial environment development indicators which determine financial deepening, i.e. the availability of financing through debt and equity to GDP.
non-17
EMU acquirers is larger for those countries which have relatively more shallow financial markets. This leads to my first hypothesis:
Hypothesis II The probability that a given deal in the EMU has a non-EMU acquirer
ddddddddddd decreases with deeper financial markets of the EMU target country.
Again, I will explore this relationship during both the pre-crisis period and the euro-crisis period holds.
3 Data and methodology
In this chapter the variables (section 3.1) and methodologies (section 3.2) for my study are described. In order to increase the comparability with recent fire-sale FDI studies, variables and methods are used which are coherent with previous studies, most notably with Alquist et al. (2013), Munichs (2014) and Weitzel et al. (2014). Section 3.3 provides a summary of the dataset.
3.1 Data
Data on FDI flows and M&A deals both have their advantages and disadvantages. This study, following most fire-sale FDI literature, will rely on M&A data. A benefit of focussing on M&A’s, which is a part of FDI, rather than data on FDI flows, is that M&A data excludes greenfield investments. The fire-sale FDI theory is interested in the acquisitions of existing firms and thus greenfield investment should be ignored. M&A’s include the majority of FDI flows accounting for up to 80% during merger waves (Stiebale and Reize, 2011). For data on the number of M&A’s in EMU countries I use the Thompson One database. The Thompson One database has data available on: the date of the acquisition, the amount of the acquired shares, the targets’ and acquirers’ firm nationality and the targets’ and acquirers’ SIC-industry classification. Although I am interested in acquisitions and not in mergers the Thompson One database does not discriminate between the two. In practice this should not be a problem since mergers are much rarer in practice. I will refer to the M&A data from the Thompson One database as acquisitions.
18
Because purchases in the banking sector have different determinants for acquisitions than non-banking firms, I chose to exclude all acquisitions were banks are either the acquirer or the target (e.g. Beccalli & Frantz, 2013; Hernando et al., 2009).
To identify the fire-sales effects during the crises I depend on the comparisons of transactions undertaken during the crises and non-crisis periods in the EMU countries. Therefore, to identify the fire-sale effects through the variation in the cross-sectional data during the crisis periods it is needed to exclude other effects which possibly influence domestic and foreign acquisition behaviour. With several countries joining the EMU just before the beginning of the crisis, during the crisis, or pending to introduction of the currency, possible cross-border acquisitions effects arising from joining the EMU are a concern. To exclude the effects of joining the EMU, I will mainly test the fire-sale FDI effects, in what I call, the Old-EMU countries. However to see if my results hold for the whole EMU, I will also perform my analysis in a combined Old-EMU, New-EMU sample. The Old-Emu countries include the original members of the Eurozone. These Old-EMU and New-EMU countries are listed in Appendix A.
A problem which arises when studying the fire-sale effects for the EMU is that the euro-crisis immediately followed the global-crisis. Since both crises may have had different results on acquisition patterns it is of importance to distinguish between the two. I follow Munnichs (2014) his demarcation of the euro-crisis. He argues that euro-crisis started between December 2009 and February 2010 and therefore uses the 1st of January 2010. I also use the 31st of December 2013 as the end date of the euro-crisis, which is the last date in the dataset. In the current state of the ongoing debt crisis one could argue that the euro-crisis by now has mainly become a Greece crisis (Trayner et al., 2015). In my analysis I will also use a robustness check in which I analyse my results ending the euro-crisis at the end of June 2013.
I demarcate the beginning of the global-crisis on the 1st of January 20097 and the starting date of the euro-crisis as its end. In my data sample, I choose to expel all observations during the year 2008 for three reasons; First, the effects of the global-crisis did not start simultaneously in all countries, for example its effects were apparent earlier in the U.S. than in the EU; Second, a large number of
7
19
countries tried to counter the negative effects of the crisis by employing Keynesian policies which at first softened the effects of the crisis; Third, the effects of the global-crisis on GDP and liquidity is not immediate but results at a delay.
This third point is of particular important for the results of fire-sale acquisitions, where illiquidity is an important concept. At the outbreak of the crisis most firms8 still have orders and liquid current assets so illiquidity is not a big problem. As the crisis progresses, orders fall, internal financing dries up and external financing becomes less accessible. Appendices B and C illustrate that the effects of the global-crisis indeed appear to be delayed. Appendix C shows that the number of corporate bankruptcies, which depend on illiquidity, in the Netherlands sharply rises in 2009 with respect to 2008, which was a year with a relative low amount of bankruptcies. Appendix D shows the GDP of the EMU, with Germany and Greece also shown individually, and illustrates that GDP still rises in 2008 and only starts to fall in 2009.
With respect to the data on the legal and financial environment I follow Hyun & Hyuk (2010), they copy the variables used by Di Giovanni (2005) and use approximates for the determinants of La Porta et al. (1997). The variables for the legal environment are the ‘Strength of legal rights index’, which rates how the rights of borrowers and lenders are protected and ‘Rule of Law’ which indicates how society has confidence in and abide by the rules of the society, and in particular the quality of contract enforcement, property rights, the police, and the courts. The variables used for the financial environment are the ‘market capitalization of listed companies’ as a percentage of GDP and the ‘domestic credit provided by the financial sector’ as a percentage of GDP. The data is retrieved via the freely accessible WorldBank database.
I add several control variables to control for macroeconomic conditions and their effects on acquisition patterns. In the fire-sale literature it is common to include the natural logarithm of the targets’ country real GDP per capita and real GDP-growth lagged one year and the development of the exchange rate9, of which I take the natural logarithm (see e.g., Aguiar & Gopinath, 2005; Alquist et al., 2013;
8
This might be different for banks which were affected more immediate trough balance sheet contagion effects. This is not a problem since bank-acquisitions are excluded in my sample.
9
20
Munnichs, 2014; and Weitzel et al., 2014). The natural logarithm is used to evaluate the effect of a 1 precent change of the regressor on the regressand. To control for the euro-crisis specific conditions, which is a sovereign debt crisis, I also control for public debt as a percentage of GDP. The data is retrieved via the freely accessible WorldBank database except for data on the exchange rate which is retrieved from the freely accessible IMF database. Appendix E provides an overview of the variables used and their definitions.
3.2 Methodology
This section is divided up into 2 subsections. The first subsection (3.2.1) explains the methods used to answer my expectation on insider and outsider acquisitions-shares during the euro-crisis. The second subsection (3.2.2) explains the methods used check if these results hold while controlling for macroeconomic variables and to test my hypotheses.
3.2.1 The Two-Sample t-Test Assuming Unequal Variances
In order to test if the acquisition shares of insiders and outsiders changed during the crises and pre-crisis periods I will use a variant of the student’s t-Test to examine if the shares of domestic acquisitions and cross-border acquisitions of different groups are significantly different during the three periods of time. To make the distinction between the pre-crisis period, the global-crisis and the euro-crisis I will use dummy variables, assigning a 1 during the specific periods and a 0 outside this period. The t-Tests compare means which are calculated by summing up all observations for which the dummy variable equals 1 and dividing this by all observations during that period per group. The groups which I have identified, in section 2.2.2, are also designated using dummy variables, assigning a value of 1 if a firm from an acquirer country belongs to a certain group and 0 when it does not.
The variant of t-test which will be employed is the Two-Sample t-Test Assuming Unequal Variances. The convenience of t-procedures in testing binominal data10 is that it can be used for clearly non-
10
21
normal data11 when the means of two samples are compared (Moore et al., 2003, p. 469). If the samples are large enough, the Central Limit Theorem holds and the t-Test can be used for non-normal data. This condition is easily met in my entire set of data test which usually include hundreds to thousands of observations. I prefer the Two-Sample t-Test Assuming Unequal Variances over a test which assumes equal variances because the sample sizes in my test are large and have unequal sizes (Coombs et al., 1996). In the Two-Sample t-Test Assuming Unequal Variances the t-statistic is calculated as follows 𝑡′ = 𝑥̅̅̅̅ −𝑥1 ̅̅̅̅ −∆2 0 √𝑠12 𝑛1 ⁄ +𝑠22 𝑛2 ⁄ 12 (1)
and significance is found when the t-statistic lies outside the value associated with the critical t-value at the chosen level of significance, for which I use 𝛼 = 0.05 throughout this study.
In my analyses I test the differences between two samples of binominal data. This means that the sample mean is equal to the share of the observation being true13 for which the dummy takes a value 1. Thus, the t-Test compares means but in the tests presented in thus study also compares the shares of different groups, which are the same in this study.
3.2.2 The Linear Probability Model
For answering my expectations and hypotheses, I follow a methodology used by Alquist (et. al., 2013) and Munnichs (2014) who use a Linear Probability Model (LPM). I use different adaptations of equation 2 which I will specify below.
Equation 2 tests if the results of the different t-Tests, comparing the two crisis periods with the pre-crisis period, hold when I control for macroeconomic conditions during these different periods. To test if the share of cross-border acquisitions of insiders and outsiders changed during the crises with respect to the pre-crisis period I use the LPM. In the LPM the dependent variable is a dummy variable
11
Using the Sign Test, which is sometimes proposed for non-normal distributions, does not provide an outcome as this test compares medians. In cases with binominal data the median will always be 1 if more than 50 percent of the data is assigned a 1, and 0 if less than 50 percent is assigned a 1, and 0.5 otherwise. This makes tests comparing medians unusable in the comparison of binominal data.
12
t’ Is the resulting test statistic which is calculated by 𝑥̅̅̅ and 𝑥1 ̅̅̅ which are the sample means of the two sets which will be 2
compared, ∆0= 𝜇1− 𝜇2 which are the hypothesized differences in population means, 𝑠12 & 𝑠22 are the variances of sample 1
and sample 2 and 𝑛1 & 𝑛2 are the number of observation in set 1 and 2. 13
22 (𝐷𝑘𝑐𝑡) , and my starting equation is estimated by:
` 𝑃 (𝐷𝑘𝑐𝑡= 1| ·) = 𝑎 + 𝛽1𝐷𝑔𝑡+ 𝛽1𝐷𝑒𝑡+ 𝛽𝑚𝑐𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑐𝑡+∈ (2)
Here k, c and t stand for transaction, target country, and time respectively and a is the intersect. The dependent dummy variable takes the value 1 if the acquisition is a cross border acquisition of the tested group and a value of 0 otherwise. To clarify, when comparing the likelihood of an acquisitions being completed by a foreign acquiring firm from a specific group during a certain crisis to the same likelihood in a non-crisis period, the dependant variable is a “Cross-border Acquisition Dummy”,(𝐷𝑘𝑐𝑡), that takes the value 1 when the acquiring firm in the transaction k, in insider/outsider
country c, at date t, and belongs to the category “cross-border”, and otherwise takes the value 0.
The vector of explanatory variables includes two crisis dummies 𝐷𝑔𝑡 and 𝐷𝑒𝑡 which respectively
indicates if the acquisition is done during the global-crisis or euro-crisis. In addition, a vector of country-level macroeconomic controls, 𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑐𝑡, which consists if per capita GDP, GDP growth, exchange rate, and debt to GDP are included in the regressions as proxies for aggregate conditions in the country of the target firm. ∈ is the error term. The coefficients of interests are 𝛽1 and 𝛽2.
Next, I employ an adaptation of equation 2, equation 3. This allows me to test if the results of the different t-Tests on the shares of insiders and outsiders during the euro-crisis with respect to the global-crisis period hold when I control for macroeconomic conditions during the different periods. This adaptation is estimated by:
𝑃(𝐷𝑙𝑘𝑐𝑡= 1| ·) = 𝑎 + 𝛽2𝐷𝑒𝑡+ 𝛽𝑚𝑐𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑐𝑡+∈ (3)
In the setting of this model only crisis observations are included for the different insiders and outsiders. Through this method differences in the acquisition shares are tested during the euro-crisis,
𝐷𝑒𝑡, compared with the global-crisis. Here, the coefficient of interests is 𝛽2. The rest of the equation
follows a similar explanation to equation 2. Comparing the outcomes of equations 2 and 3 clarifies if the developments in the share of cross-border acquisitions, controlling for macroeconomic conditions, follows a similar pattern compared to the results of the t-Test.
23
𝑃(𝐷𝑙𝑘𝑐𝑡= 1| ·) = 𝑎 + 𝛽3𝑅𝑖𝑔ℎ𝑡𝑠 + 𝛽4𝑅𝑢𝑙𝑒 + 𝛽5𝐶𝑎𝑝 + 𝛽6𝐶𝑟𝑒𝑑𝑖𝑡 + 𝛽𝑚𝑐𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑐𝑡+∈ (4) The dependant variable in this regression is a dummy variable which takes the value 1 when the EMU target is acquired by an acquirer outside the EMU. The continuous variables Rights, Rule, Cap and Credit indicate the score on the ‘legal rights index’, the score of ‘Rule of Law’, ‘Market Capitalization’ and ‘Credit supplied’ respectively. To study if there are differences between the pre-crisis period and the different crises periods on the sign and significance of the score on the ‘legal rights index’, the score of ‘Rule of Law’, ‘Market Capitalization’ and ‘Credit supplied’ in explaining the regressand, I will estimate equation 4. The analysis relies on four models including cross-border acquisitions only which study the effects on the regressand by the regressors during four different time periods. These periods include: the whole time period ranging from 1999 to 2013; the pre-crisis period exclusively; the global-crisis exclusively; and euro-crisis exclusively. The coefficients of interests are 𝛽3 to 𝛽6. The rest of the equation follows a similar explanation as the explanation of equation 2.
While the LPM works similar as an OLS-regression, which makes it simple to estimate and intuitive to interpret, the outcomes of the model might lead to wrong inferences. The LPM allows coefficient outcomes outside the range (0,1), such as negative values and values greater than 1. This cannot be true as the outcome should strictly lie between 0 and 1, because the possibility range lies between 0 and 1. Here, 0 indicates that we are 100 percent sure of the possibility that an outcome, where the dummy is true, will never happen, and 1 indicates that we are 100 percent sure of the possibility that an outcome, dummy is true, will always happen14.
To control for this shortcoming of the LPM, I use a logit model15 for the same regression to check the robustness of the outcomes. Appendix F provides the estimated logit equations, which are logit adjustments to the equations presented above and explains the econometric method in short.
14
Setting the values outside the range (0,1) equal to the closest border 0 or 1, a process which is called truncation in econometrics might lead to faulted inferences from the model, i.e. can we really be certain that the probability of an outcome is exactly 0 or 1? The answer is of course no.
15
24
3.3 Data set summaries
My final dataset includes 97.234 observations of acquisitions in EMU countries, of which 83.797 are in the Old-EMU. Over the whole time period 32 percent of the acquisitions in the Old-EMU are cross-border acquisition which numbers to a total of 27.002 cross-cross-border acquisitions.
Figure 2 shows a graph depicting the number of acquisitions per year in the old-EMU of different insiders and outsiders. We can observe that the observation of Mariana (2011), that global-crisis caused an overall fall in acquisitions holds. From the graph we can see that this drop is larger for cross-border acquisitions than for domestic acquisitions in the EMU.
Figure 2: Number of acquisitions in the Old-EMU
Graph showing the number of acquisitions of different groups. The dotted lines correspond to the left-hand vertical axis and the figure-lines to the right-hand vertical axis. Domestic refers to the total number of domestic acquisitions in the EMU; Cross-border refers to the total number of cross-border acquisition in the EMU. The figure-lines are the number of different (combined) groups their cross-border acquisitions in the EMU. The groups Other-EU and NS are combined, as well as BRIC and EE, since the movements of this groups correspond narrowly and to prevent an overflow of information.
Table 2 shows the share of cross-border acquisitions to for each EMU country and different country groups within the EMU region. When we compare the four different country groups in the bottom part of table 2 we can see a clear difference between the averages of the New-EMU, the Old-EMU and the whole EMU. Throughout the pre-crisis and crisis periods the new-EMU countries clearly have a higher share in cross-border acquisitions. This indicates evidence of other acquisition effects at play in these countries, e.g. effects of joining the EMU, justifying their exclusion as this might distort the results. To check the robustness of the results a combined data-set including both Old-EMU and
25
EMU is used on which similar test are performed. To promote the comparison with the results of the study by Weitzel et al. (2014)16 and Munnichs (2014) I have also included the averages of the PIIGS-countries. From here onwards, I will refer to the Old-EMU as EMU.
Table 2: Table 1: Percentage of cross-border acquisitions to all acquisition
Country pre-crisis global-crisis euro-crisis Country pre-crisis global-crisis euro-crisis
Austria 43.69% 42.07% 40.19% Latvia* 60.49% 47.06% 49.24% Belgium 48.62% 44.81% 51.81% Lithuania* 57.89% 41.18% 42.93% Cyprus* 39.52% 43.94% 66.79% Luxembourg* 78.13% 74.47% 76.67% Estonia* 55.44% 28.57% 47.45% Malta* 77.14% 41.67% 74.07% Finland 28.74% 28.18% 31.37% Netherlands 40.82% 32.50% 37.17% France 29.76% 19.88% 19.41% Portugal 33.86% 25.24% 33.80%
Germany 34.72% 29.50% 33.11% Slovak Rep* 72.63% 87.50% 73.13%
Greece 17.66% 27.18% 27.00% Slovenia* 27.86% 62.50% 55.88%
Ireland 49.63% 45.68% 57.88% Spain 27.02% 23.87% 25.26%
Italy 31.22% 23.92% 33.38%
Group pre-crisis global-crisis euro-crisis Group pre-crisis global-crisis euro-crisis EMU average 34.46% 28.46% 31.33% PIGGS average 30.24% 26.02% 30.97% Old-EMU average 33.67% 27.90% 30.43% New-EMU average 53.92% 45.21% 56.02%
The top part of this table show the individual shares of cross-border acquisitions to all acquisitions in the EMU countries, the New-EMU countries are indicated by *. The bottom part of the table shows the average shares of different groups. Average is the total average of all countries, Old-EMU is the average of the Old-EMU countries, PIIGS is the average of the PIIGS countries and New-EMU is the average of the New-EMU countries during these periods.
Table 3, summarizes the results of the shares of acquisitions. The top part of the table shows the development of the shares of domestic and cross-border acquisitions in the EMU. These two acquisitions types combined are all acquisitions and together sum to a 100 percent. The bottom part of the table shows the shares of cross-border acquisitions in the EMU of the identified groups. This means that the observed 41 percent share of EMU acquisitions in other EMU countries during the pre-crisis period is 41 percent of all cross-border acquisitions in that period, which was 34 percent of all acquisitions. Each period sums to a 100 percent.
16
26
Table 3: Geographical origin of acquirers targeting the EMU during the three periods
Acquisition type share Pre-crisis Global-crisis Euro-crisis
Domestic 66.33% 72.10% 69.57%
Cross-border 33.67% 27.90% 30.43%
Cross-border group share Pre-crisis Global-crisis Euro-crisis
EMU 40.61% 41.23% 34.96% Other-EU 23.54% 17.68% 18.31% NS 5.83% 7.76% 7.11% Triad 22.83% 20.68% 25.92% BRIC 1.64% 3.85% 4.97% EE 1.38% 3.78% 3.77% ROW 4.17% 5.02% 4.96%
This table shows the share of acquisitions in the EMU during the three identified time periods. The top part shows the share of domestic and cross-border acquisitions, to total acquisitions. The bottom part shows the shares of cross-border acquisitions in the EMU of the different identified groups to the total of all cross-border acquisitions. As introduced before EMU refers to the Old-EMU.
At first glance, we see that in each period the majority of the acquisitions are domestic acquisitions. We can see that at the start of the global crisis this share further increases. From figure 2 displays that this increase in the share of domestic acquisitions is accompanied by a sharp decline in cross-border acquisitions.
During the euro-crisis a decline of domestic and EMU cross-border acquisitions is depicted, which might be evidence of fire-sale FD. These patterns will be analysed in the next chapter. Appendix G summarizes further information on the shares and number of acquisitions for individual acquirer groups. Similar to table 3, appendix H summarized the results of the shares of acquisitions in EMU countries minus the PIIGS, identified by EMU*, and the PIIGS separately and shows similar acquisitions patterns for both groups which correspond to the patterns in table 3.
With respect to the data which will be used in the regression analyses, Appendix I shows the descriptive statistics in Table I.1, for the dummy variables, and Table I.2, for the continuous explanatory and control variables. Table I.3 tabulates the correlations between all the regression variables. The correlations between all the variables are neatly below the .75 level which makes them appropriate to use with respect to possible multicollinearity.
4 Empirical Results
27
EMU and the results of cross-border acquisitions shares of different groups in the EMU countries. Next, Section 4.2 describes the test results for the legal and financial environment and section 4.3 finally provides a brief summary of the results for the PIIGS-countries and the rest of the EMU to enhance comparability with the studies by Munnichs (2014) and Weitzel (2014).
4.1 Cross-border acquisitions in the EMU
In this section I first analyse the difference between the shares of cross-border acquisitions to the total acquisition in the three time periods (section 4.1.1). Next, I will analyse the difference in the shares of cross-border acquisitions into the EMU by the different groups (section 4.1.2). In my analyses a range of t-Tests is employed, to test if acquisition shares are significantly different from each other during these periods. The hypothesis which is tested is that the acquisitions shares of the two sets compared in the t-Test are equal to each other. Several robustness and sensitivity checks are also included to test the strength of the results.
4.1.1 The share of cross-border acquisitions in the EMU
To visualize if the acquisition pattern follows the expected pattern of fire-sale acquisition, figure 3 displays the actual acquisition patterns for the EMU in a similar fashion as to how figure 1 is presented. Again, we can observe is that the majority of the acquisitions are domestic acquisitions. The group EMU is showed individually since this group is considered the only insider during the euro- crises. The black vertical line represents the start of the global-crisis and the white line the start of the euro-crisis. With respect to figure 1, the lower shares of domestic and cross-border insider acquisitions during the euro-crisis, with respect to the pre-crisis period, is quite the opposite of what we expect to happen in a situation of fire-sale acquisitions, i.e. we observe the opposite of what is expected. However, this fall in the share of domestic acquisitions comes forth from the global-crisis and is explained by the drop in the total number of cross-border acquisitions in the EMU, shown in figure 2.
28 Figure 3: Acquisitions in the EMU
This graph shows the development of the shares of domestic EMU and the development of cross-border acquisition by EMU members and non-EMU members separately. The black vertical line shows the beginning of the global crisis and the white vertical line the beginning of the euro-crisis. The vertical axis starts at 60%.
The top part of table 4 provides the results of the t- tests were the share of cross-border and domestic acquisitions during the pre-crisis period, the global-crisis and euro-crisis are compared. These acquisition shares for each period are tested against the acquisitions shares of the other periods. The signs >> and designates significance for both the one-tail test as for the two-tail test, > indicates significance for the one-tailed test and Χ indicates no significance is found at the 0.05 level. The results of all the individual t-Tests are listed in appendix J.
The results indicate that the shares of cross-border and domestic acquisitions differ significantly during all three periods. The share of cross-border acquisitions during the global-crisis and euro-crisis are significantly smaller than this share during the pre-crisis period. The share of cross-border acquisitions during the euro-crisis is significantly larger than the mean during the global-crisis. Appendices J and K include robustness checks for allowing a different time period of the euro-crisis and inclusion of the whole of the EMU, i.e. the Old-emu and New-EMU countries combined.
29 Table 4: Results of t-Test comparing acquisition shares
The top part of the table shows results of the t-Tests comparing pre-crisis, global-crisis and euro-crisis periods domestic and cross-border acquisition shares with respect to total acquisitions. The bottom part of the table shows the results of the t-Tests comparing the cross-border shares of the identified groups with respect to total cross-border acquisitions. The < and > signs show if A is significantly smaller or larger than B, i.e. A is smaller than B is denoted by A<B. A single < or > sign indicates significance for the one-tale tests and a double << or >> sign significance for the one-tale and the two-tale test at the 5% level. X indicates no significance is found.
Table 5: LPM, partial regression showing the crisis dummy variables results Dependent Variable: Cross-border
LPM 1. All observations Obs. 83797
Crisis dummy Coefficient Std. Error Prob.
Global-crisis -0.09 0.01 0.00* Euro-crisis -0.04 0.01 0.00* LPM 2. Excluding pre-crisis
observations Obs. 29510
Crisis dummy Coefficient Std. Error Prob.
Euro-crisis 0.03 0.01 0.00*
This table presents a summary of the test results of the two Linear Probability Models, LPM, corresponding to equations 2 and 3 for the coefficients of interest. The results show the increase in cross-border acquisitions estimated by crisis dummies. A positive (negative) signs means that the dummy variable leads to an increase (decrease) in cross-border acquisitions The first regression, LPM 1, consists of all acquisitions observations in the EMU during the whole time range of the main data-set. The second regression, LPM 2, consists of all acquisitions observations in the EMU during the two crisis periods. The coefficients off the control variables are not included in this table to prevent an overload on information but can be found in appendix K, table K.1. Coefficients marked * are significant at the 5% level.
These results of the two LPM analysis control for macroeconomic conditions and confirm the findings of the t-Tests. Compared with the pre-crisis period, the likelihood that an acquisition was cross-border is lower during both crises. During the euro-crisis the likelihood that an acquisition was cross-border is higher than during the global-crisis. Appendix K also shows that the results are robust when using a logit model opposed to a LPM or t-Test.
Pre-crisis Global-crisis Pre-crisis Euro-crisis Global-crisis Euro-crisis
Domestic 66.33% << 72.10% 66.33% << 69.57% 72.10% >> 69.57%
Cross-border 33.67% >> 27.90% 33.67% << 30.43% 27.90% << 30.43%
Cross-border
group share Pre-crisis Global-crisis Pre-crisis Euro-crisis Global-crisis Euro-crisis
EMU 40.90% X 41.40% 40.90% >> 35.64% 41.40% >> 35.64% Other-EU 23.54% >> 17.68% 23.54% >> 18.31% 17.68% X 18.31% NS 6.02% << 8.33% 6.02% << 6.88% 8.33% X 6.88% Triad 22.73% >> 20.67% 22.73% << 25.92% 20.67% << 25.92% BRIC 1.63% << 3.81% 1.63% << 4.92% 3.81% << 4.92% EE 1.38% << 3.78% 1.38% << 3.77% 3.78% X 3.77% ROW 4.17% < 5.02% 4.17% << 4.96% 5.02% X 4.96%
Pre-crisis & Global-crisis Pre-crisis & Euro-crisis Global-crisis & Euro-crisis
30
4.1.2 The cross-border acquisitions shares of different groups in the EMU
In this section the results from the t-Tests are presented which compares the shares of cross-border acquisitions from the identified groups. The analysis works similar to the t-Test discussed in the previous subsection. The bottom part of table 4 provides the results of these t-Tests where the cross-border shares of different groups during the pre-crisis period, the global-crisis and euro-crisis are compared. Here, the cross-border acquisitions shares are a part of all cross-border acquisitions for each period. The shares of each group during each period are tested against the shares of the other periods from the same group. Again, the results of all the individual t-Tests are listed in appendix J.
The results indicate that the share of EMU cross-border acquisitions in the EMU did not significantly differ during the pre-crisis-period and global-crisis. The cross-border acquisition share of the euro-crisis compared to the pre-euro-crisis period and global euro-crisis is significantly lower.
For the group Other-EU we find that during the global-crisis and euro-crisis the share of cross-border acquisitions is significantly lower than this share during the pre-crisis period. No such significance is found comparing the means of the global-crisis and euro-crisis. For the group NS, we find the same results as for the group Other EU.
For the BRIC group the test shows that the cross-border share in the global-crisis period is significantly higher than this share of the pre-crisis period. The cross-border share of the global-crisis period is significantly lower than this share during the euro-crisis.
For the EE group the test shows that the cross-border means of the global-crisis period and euro-crisis period are significantly higher than the mean of the pre-crisis period. The means of the global-crisis and euro-crisis are not significantly different
31
Appendix K shows corresponding results follow from the LPMs for the identified groups, with respect to the results of the t-Test17. These checks indicate that the results are robust across different econometric methods. Appendix L presents sensitivity checks which includes 2008 and follows Munnichs (2013) by letting the start date of the global-crisis be the 1st of October 2008, and the rest of 2008 as the pre-crisis period. Appendices M and N shortly present the findings of the results of more sensitivity checks In appendix M, the end of June 2013 is used as the end date of the euro-crisis. Appendix N lists the results using a dataset where the whole-EMU, i.e. Old- and New-EMU are combined. All the sensitivity and robustness checks yield similar results. These checks show that the results hold across different data demarcation. The robustness and sensitivity checks indicate that the results can be considered strong.
4.2 PIIGS, EMU* and Germany
Appendix O presents the results for t-Tests comparing the acquisition shares of the PIIGS and the other EMU countries, i.e. excluding the PIIGS. The EMU without the PIIGS is denoted as EMU*. The acquisition patterns in the PIIGS and EMU* both follow a similar pattern as the EMU, i.e. a decrease in cross-border acquisitions during the global-crisis and an increase during the euro-crisis. The only difference is that the cross-border acquisition share in the PIIGS during the euro-crisis is larger than this share in the pre-crisis period, but not significantly so.
Appendix P includes the results of the t-Tests comparing the acquisitions shares in the EMU* and the PIIGS. Table 6 presents the results. A clear difference, which can be observed together with table 4, is that during the pre-crisis and global-crisis the cross-border acquisition shares in the EMU* are higher than in the PIIGS. During the euro-crisis these shares are not significantly different, signalling a large rise in the share of cross-border acquisitions in the PIIGS than in the EMU*.
Furthermore, we find that during the pre-crisis periods the EMU* was relatively targeted less by the EMU firms than the PIIGS, which could indicate that the problems in these countries were so severe that firms from the EMU region still saw changes to exploit the possible fire-sale in the PIIGS. The opposite is true for the NS and the Triad. The other identified groups targeted both the PIIGS and the
17
