The Country-of-Origin effect

(1)

The Country-of-Origin effect

An investigation of the Country-of-Origin effect on post-acquisition performance of

cross-border acquisitions

Brian Blömer (S2381419)

b.r.blomer@student.rug.nl

Supervisor: P.J. (Paulo) Marques Morgado

Co-assessor: Dr. G. (Gjalt) de Jong

MSc. International Business and Management

(2)

Abstract

Technological development and globalisation have vastly contributed to the popularity of cross-border acquisitions. However, theory prevailing in existing literature suggests that unidentified variables may explain significant variance in post-acquisition performance. Therefore, this research investigates the concept of the country-of-origin effect and whether it holds explanatory value for post-acquisition performance. More specifically, we argue that firms from developing countries experience a negative country-of-origin effect, compared to their developed country counterparts. In light of theoretical perspectives, we propose that a negative country-of-origin effect results in lower post-acquisition performance due to associated negative perceptions of shareholders. Therefore, a framework to test the influence of country-of-origin on shareholder value is developed. Quantitative research in combination with event study is proposed to analyse the research question.

Key words:_{Cross-border acquisitions, post-acquisition performance, country-of-origin, liability of}

(3)

1. Introduction

Once, Lenovo was an obscure Chinese PC manufacturer, barely making any money. Today, it is the world’s largest PC maker with close to $40 billion in sales, leaping ahead of its closest competitors, HP and Dell (Panos MourdouKoutas, 2015). How did Lenovo overcome their negative brand image and gain legitimacy in the international arena? In 2002 the Chinese government announced its ‘go global’ policy by encouraging Chinese companies with the capabilities and know-how to expand abroad (Liu, 2007). Lenovo, as a leader in the Chinese IT sector, responded with a string of aggressive acquisitions, which helped the company assembling advantages in scale and scope (Panos MourdouKoutas, 2015). Most famously, in 2005, Lenovo acquired the PC hardware business unit of IBM. Global expansion has helped Lenovo’s products to gain recognition in foreign markets, and overcome the Country-of-Origin effect (Panos MourdouKoutas, 2015). Based on these and other examples, we observe the contemporary gain in voice of so called emerging-market multinational enterprises (EM MNEs), international firms which originate from emerging markets. These firms engage in outward foreign direct investment (FDI) in one or more foreign countries (Luo & Tung, 2007).

Technological development and globalisation have vastly contributed to the popularity of mergers and acquisitions (M&A) and cross-border M&As (Shimizu, Hitt, Vaidyanath, & Pisano, 2004). A typical and controversial issue regarding acquisitions in general is whether acquiring firms create value for their shareholders. Empirical pieces of evidence are widely available, however, no theoretical framework can completely explain the relationship between acquisition antecedents and subsequent performance (Hitt, Harrison, Ireland, & Best, 1998). Therefore, King et al. have suggested that unidentified variables may explain significant variance in post-acquisition performance, suggesting the need for additional theory development and changes to M&A research methods (King et al., 2004).

According to institutional theory, the institutional environment of a country influences MNEs’ practices and behaviour in foreign markets (Dimaggio & Powell, 1983; Kostova & Roth, 2002; North, 1990). As long as companies do not gain social acceptance, credibility, and support from host country nationals, they are in a disadvantageous position compared to domestic companies. This disadvantageous position is consequently reflected in higher liability of foreignness (LoF) (Brouthers, O'Donnell, & Hadjimarcou, 2005; Suchman, 1995; Zaheer, 1995).

(6)

MNEs) (Axinn & Matthyssens, 2002). Other scholars have called for research into strategies for firms from emerging markets (Peng, 2005). Most importantly, decades of research scrutiny have led to ambiguous conclusions regarding the country-of-origin concept. We agree that it is important to address these gaps in the literature for various reasons.

First, a surge in cross-border acquisitions was the principal factor behind a rebound in global FDI flows in 2015. Secondly, we contribute to the understanding of the behaviour of EM MNEs (Axinn & Matthyssens, 2002; Gubbi et al., 2010) and answer the call for more research into EM MNEs (Peng, 2005). Finally, we go beyond the perceptual measurement of the country-of-origin concept and

investigate whether the concept holds explanatory value with regard to post-acquisition performance of cross-border acquisitions. Hereby we answer the call for additional theory development and research into M&A research methods (King et al., 2004).

Based on the aforementioned reasons, we come to the following research question:

“Does the Country-of-Origin influence the cumulative abnormal returns after cross-border acquisitions?”

We expect that cross-border acquisitions result in lower cumulative abnormal returns when shareholders perceive the acquiring firm to originate from a developing country. While, cross-border acquisitions should result in higher cumulative abnormal returns when shareholders associate the acquirer with a developed country.

From a methodological point of view, our research makes a contribution by identifying that the country-of-origin concept could explain a portion of the unidentified variance in post-acquisition performance. Furthermore, we identified the World Governance Indicators as a proxy for the

country-of-origin concept, because the WGI data sources comprise both perceptual and non-perceptual measures. From a managerial point of view, we noted that not all firms will derive the same benefits from internationalisation. MNEs face costs due to liability of foreignness (Brouthers, O'Donnell, &

(7)

2. Literature Review

In order to understand whether current literature is applicable to the country-of-origin effect, we first have to review theory that prevails in existing literature. This chapter starts with introducing FDI.

Subsequently, we draw on international expansion theory and review entry mode strategies. After which, we will examine the relationship between cross-border acquisitions and performance. More specifically, we build on institutional distance and the concept of LoF in order to expand our knowledge of the country-of-origin effect. Finally, we provide the reader with an illustrative synopsis of theoretical perspectives and we introduce our corresponding propositions.

2.1 Introducing the acquisition strategy

Foreign direct investment is an investment that takes place when an entity in one country establishes a business operation, in the form of controlling ownership, in another country. The World Investment Report, provided by the United Nations Conference on Trade and Development (UNCTAD), presents financial data for FDI all over the world. Global FDI flows rose by 38 per cent to $1.76 trillion in 2015, their highest level since the global economic and financial crisis of 2008-2009 (UNCTAD, 2016). A surge in cross-border M&As to $721 billion, from $432 billion in 2014, was the principal factor behind the global rebound in 2015 (UNCTAD, 2016).

The geographic pattern of FDI flows tilted in favour of developed economies in 2015, although developing Asia remained the largest recipient of FDI flows. Flows to developed economies nearly doubled rising from $522 billion in 2014 to $962 billion (UNCTAD, 2016). FDI flows to North America and Europe registered particularly large increases. The large-scale increase in FDI flows to Asia

contrasted with a more modest performance in other developing regions (see Figure 1) (UNCTAD, 2016). A primary catalyst of decreasing inflows in developing and transition economies was the continued decline in commodity prices, especially for crude oil, metals and minerals. In 2015, MNEs from

developed economies invested abroad $1.1 trillion, with MNEs from Europe and Japan contributing to the growth (UNCTAD, 2016). By contrast, almost all developing and transition regions saw their FDI

(8)

Figure 1: FDI inflows, by region, 2013-2015 (billions of dollars)

Source: (UNCTAD, 2016)

Drawing from international expansion theory, various scholars have analysed entry mode choice as a strategy for internationalisation. According to Chang and Rosenzweig, MNEs can choose between a variety of entry modes when entering a host country via FDI (Chang and Rosenzweig, 2001):

● Greenfield investments, which entails setting up a new plant from scratch. ● Acquisitions, which denotes the purchase of a controlling interest in a local firm.

● Joint ventures, which can be defined as the pooling of assets of two or more firms in a separate entity.

2.2 Cross-border Acquisitions and Performance

Technological development and globalisation have vastly contributed to the popularity of M&As and cross-border M&As (Shimizu et al., 2004). As aforementioned, a surge in cross-border M&As was the principal factor behind the global rebound in 2015. Several factors are responsible for fuelling the growth of cross-border M&As. Among these factors are the worldwide phenomenon of industry consolidation and privatisation, and the liberalisation of economies (Shimizu et al., 2004). In 2015, the surge in

acquisitions was partly driven by corporate reconfigurations (i.e. changes in legal or ownership structures of MNEs, including tax inversions) (UNCTAD, 2016).

(9)

acquisitions are an effective strategy to acquire competences, expertise, technology, products, and achieve economies of scale and scope (Wirtz, 2003). What is more, according to Luo and Tung acquisitions are primarily used to secure brands (Luo & Tung, 2007). Thus, acquisitions add differentiation and brand advantages to the existing cost advantage. This strategy of buying-in establishes international reputations and accelerates market entry and legitimacy. But contrary to the expected advantages, around 80 percent of M&As fail and do not result in improved shareholder value (Cartwright & Cooper, 1993).

2.2.1 Cross-border Acquisitions and Value Creation

In this section we analyse a controversial issue regarding acquisitions in general. The primary question is do acquiring firms create value for their shareholders. The theoretical foundation for positive returns from cross-border acquisitions are based on the assumption that firms enter foreign markets to exploit the firm’s specific resources to take advantage of imperfections in the market (Buckley & Casson, 2016; Morck & Yeung, 1992; Wilson, 1980). Several researchers state that cross-border acquisitions provide integrating benefits of internalisation, synergy, and risk diversification and thereby create wealth for both acquirer and target-firm shareholder (Kang, 1993; Markides & Ittner, 1994; Morck & Yeung, 1991). In general acquisition performance is defined in the accounting and financial return context of the

shareholder value, which can be derived from the cumulative abnormal return (Zollo & Meier, 2008).

Empirical pieces of evidence of cross-border acquisition activities are widely available, however, they vary widely in conclusion with regard to either negative or positive effects on shareholder value. Morck and Yeung examined 332 foreign acquisitions by U.S. firms between 1978 and 1988 (Morck & Yeung, 1992). Using a transaction cost perspective, they found that the acquirer’s R&D intensity, advertising intensity and management quality were positively associated with the acquirer’s abnormal returns. Furthermore, Kang examined 119 Japanese firms that bid on 102 U.S. firms between 1975 and 1988. He found that Japanese acquisitions of U.S. firms created wealth for both acquirer and target firm shareholders (Kang, 1993). On the contrary, following an integrative theoretical approach, Datta and Puia reported opposite results from those reported above (Datta & Puia, 1995). They concluded that

cross-border acquisitions on average do not create value for the acquiring firm shareholders, aligning the results with those of domestic M&A activity.

(10)

1991-2004 (Aybar & Ficici, 2009). The results show that, on average, cross-border expansions of EM MNEs through acquisitions do not create value, but point to value destruction for more than half of the transactions analysed. Furthermore, they also found that target size, ownership structure of the target (private vs public), and structure of the bidder (diversified vs non-diversified) positively affect bidder value. While, high-tech nature of the bidder and the pursuit of targets in related industries negatively affect bidder value.

An example of positive effects is the study of Gubbi et al.. This study for cross-border M&As of Indian companies showed that companies from emerging-markets can acquire advantages in resources, technologies, and skills of the target company (Gubbi, Aulakh, Ray, Sarkar, & Chittoor, 2010). What is more, the acquiring company can get access to a new customer segment and gain market knowledge (Morosini, Shane, & Singh, 1998). Moreover, Morosini et al. also found that firms from emerging

economies are highly motivated to have access to managerial skills, advanced business knowledge, and to gain access to natural resources (Morosini, Shane, & Singh, 1998). Furthermore, Nicholson and Salabar analysed global M&A activities from India and China, and found positive abnormal returns for the shareholder of the acquiring company during a short-term, four days event window (Nicholson & Salabar, 2013). Finally, the study of Gu and Reed found that Chinese cross-border M&A activity resulted in improved shareholder value, based on cumulative abnormal returns (Gu & Reed, 2013).

On the other hand, Luedi established that Chinese cross-border M&As tend to perform worse in comparison with Western cross-border M&A activity (Luedi, 2008). They also concluded that the cumulative abnormal return tended to be negative. This is confirmed in the study by Alexandridis et al., who conclude that M&A announcement returns in general tend to be insignificant or negative

(Alexandridis, Petmezas, & Travlos, 2010). Other studies (Deng, 2010) indicate that high rates of overseas failure are generated by weak competencies and cultural differences. Other reasons mentioned were institutional differences, language, experience, and a lack of transparency.

2.2.2 Post-acquisition Performance

(11)

variables may explain significant variance in post-acquisition performance, suggesting the need for additional theory development and changes to M&A research methods (King et al., 2004).

Based on the abovementioned evidence we can conclude that no theoretical framework can completely explain the relationship between acquisition antecedents and subsequent performance (Hitt, Harrison, Ireland, & Best, 1998). However, existing literature is in agreement that cross-border

acquisitions do have an impact on post-acquisition performance.

The dynamics of cross-border acquisitions are largely similar to those of domestic M&A activity (Shimizu et al., 2004). However, due to their international nature, they also involve unique challenges, as countries have different economic, institutional, and cultural structures (House, Javidan, Hanges, & Dorfman, 2002; Lonner, Berry, & Hofstede, 1980). Therefore, the pursuit of cross-border acquisitions is not without risk. Firms engaging in cross-border acquisitions are faced with unique risks which are defined as “the costs of doing business abroad (CDBA)” (Hymer, 1976) and the “liability of foreignness” (Zaheer, 1995). In the following sections we will elaborate on these concepts and their accompanying theoretical perspectives in an effort to build our case for the country-of-origin effect and its potential impact on cross-border acquisitions.

2.3 Challenges to Cross-border Acquisitions

As aforementioned, due to the international nature of cross-border acquisitions, these cross-border acquisitions involve unique challenges as countries have different economic, institutional, and cultural structures (House et al., 2002). One of the first theories prevailing in the existing literature concerning complications to international expansion can be found in the work of Hymer (1976). Namely, Hymer argued that MNEs, because they were foreign, faced barriers to entry in a host country. Moreover, he reasoned that in order for FDI to take place, MNEs have to possess certain firm-specific advantages which differentiated them from local competitors (Hymer, 1976).

(12)

Figure 2: The Costs of Doing Business Abroad

Source: (Eden & Miller, 2004)

According to institutional theory, the institutional environment of a country influences MNEs’ practices and behaviour in foreign markets (Dimaggio & Powell, 1983; Kostova & Roth, 2002; North, 1990). As long as companies do not gain social acceptance, credibility, and support from host country nationals, they are in a disadvantageous position compared to domestic companies. This disadvantageous position is consequently reflected in a higher Liability of Foreignness (Brouthers, O'Donnell, &

Hadjimarcou, 2005; Suchman, 1995; Zaheer, 1995).

(13)

2.4 Institutional Distance and Liability of Foreignness (LoF)

Institutions are the “rules of the game” in the society in which MNEs operate (North, 1990).

Understanding differences in institutional environment between home and host countries is critical for firms expanding their business across national boundaries (Bae & Salomon, 2010). Therefore, it is assumed that institutional distance has a key impact on entry mode choice as well as cross-border acquisitions (Bae & Salomon, 2010; Kogut & Singh, 1988).

Furthermore, it has been observed that LoF faced by MNEs often stems from a lack of knowledge about local culture and regulations. Consequently, the LoF would increase the transaction costs in the operation of an MNE (Bae & Salomon, 2010). In order to overcome the LoF firms need to seek

legitimacy in an effort to survive the social environment through corresponding to regulations and norms, which means efforts should be made to manage both formal and informal institutional distance.

2.4.1 Sources of LoF

According to Eden and Miller LoF can be seen as a combination of three hazards which altogether comprise the costs that affect foreign firms disproportionately, namely: unfamiliarity, relational, and discrimination hazards (Eden & Miller, 2004). First, unfamiliarity hazards reflect inadequate knowledge about the host country’s practices and culture (Eden & Miller, 2004). Placing foreign firms at a

disadvantage compared to local firms. This LoF is not related to the age of an MNE, but rather to the longevity of its experience in the host country (Eden & Miller, 2004). Short tenure in the host country causes additional costs that the MNE must incur to achieve the same level of host-market knowledge as a local firm.

Secondly, relational hazards are the result from increased internal (intra-firm) and external (inter-firm) transaction costs. From the perspective of intra-firm relations, costs are created due to difficulties of managing employees at a distance and from different cultures (Anderson & Gatignon, 1986). From an inter-firm perspective, costs of trust-building are incurred as well as negotiation and monitoring costs (Luiken, 2016). These uncertainties create relational hazards in the form of higher administrative costs in managing the relationship between parties involved (Buckley & Casson, 1998; Henisz & Williamson, 1999).

(14)

discrimination can result in terms of specific restrictions of the host government to protect domestic firms or specific industries. Indirect discrimination occurs in preferential treatment of local companies over foreign companies (Held & Berg, 2014). Discriminatory costs therefore focus on challenges of obtaining external legitimacy.

2.4.2 Dual Liability of Foreignness

Previously we have argued that existing literature does not provide the same insights into the behaviour of EM MNEs as DM MNEs (Axinn & Matthyssens, 2002). When EM MNEs enter developed-market countries, they commonly do so from a dual liability of foreignness position. More specifically, EM MNEs would have to compete from a weaker resource base position, are accustomed to a weaker home-country institutional environment, and suffer from a negative country-of-origin effect (Luiken, 2016). Altogether these effects can be defined as a “dual liability of foreignness position”.

We would like to briefly discuss these sources of a dual liability of foreignness position. First, apart from coping with LoF, it is argued that EM MNEs commonly compete from a relatively weaker resource base position when compared to DM MNEs (Barnard, 2010). More specifically, EM MNEs typically have a relatively weaker asset base to draw from, which increases the difficulty to operate freely.

Secondly, extant literature has suggested that emerging-market countries often lack the institutions needed for efficient market-based transactions resulting in institutional voids (Khanna & Palepu, 2006). As a consequence, EM MNEs commonly develop skills and adopt their organisational structures to match their environments, characterised by institutional voids (Luiken, 2016). These skills and structures may not be transferrable to developed-markets with well developed institutions (Gaur & Kumar, 2010). Therefore, EM MNEs need to make greater investments to become familiar and comply with well developed regulations of developed-market countries, compared to their emerging-market counterparts (Gaur & Kumar, 2010).

Finally, apart from a weaker resource base and being accustomed to a weaker institutional environment, several authors have argued that EM MNEs are more likely to possess a negative image and less legitimacy, relatively to DM MNEs (Kalasin, Dussauge, & Rivera_{-Santos, 2014). Subsequently, we} assume that EM MNES suffer from a negative country-of-origin effect in the sense that developed-market country nationals associate EM MNEs with negative home country characteristics such as poorer

(15)

Wilkinson, 2009). In contrast to that, numerous firms have used positive associations with the country-of-origin as an advantage in the marketing of goods (N. Papadopoulos & Heslop, 2014). We propose a more thorough investigation of the concept which goes beyond the perceptual judgement of consumers and investigate the effect on post-acquisition performance.

2.5 The Country-of-Origin effect

The country-of-origin effect, also known as the made-in image or the nationality bias, originally is a psychological effect describing how consumers’ attitude, perceptions, and purchasing decisions are influenced by products’ country-of-origin labelling (Cai, Cude, & Swagler, 2004).

2.5.1 The importance of analysing the Country-of-Origin effect on cross-border acquisitions

Several scholars have suggested that EM MNEs should expand internationally to countries with similar institutional frameworks (Cuervo-Cazurra & Genc, 2008; Khanna & Palepu, 2006). Furthermore, we have observed that existing literature argues that EM MNEs are more likely to possess a negative image and less legitimacy, relative to DM MNEs (Kalasin et al., 2014). However, we have observed that multiple EM MNEs, such as Lenovo, have successfully overcome their negative country-of-origin image when expanding internationally (Khanna & Palepu, 2006). Moreover, positive country-of-origin images have enabled numerous firms to gain an advantage in the marketing of their goods.

Many researchers have analysed the internationalisation process of DM MNEs. However, now scholars have called for research into strategies for firms from emerging economies (Peng, 2005). Furthermore, decades of research scrutiny have led to ambiguous conclusions regarding the

country-of-origin concept. Therefore, we propose that it is important to address this gap in the existing literature and investigate the influence of the country-of-origin concept on post-acquisition performance of cross-border acquisitions for various reasons.

(16)

Second, by including an analysis of FDI from emerging-market firms to developed-market countries, we go against the nature of conventional wisdom, which describes the direction of FDI within the global economy (Ramamurti & Singh, 2009). Furthermore, we contribute to the understanding of the behaviour of EM MNEs (Axinn & Matthyssens, 2002; Gubbi et al., 2010) and answer the call for more research into EM MNEs (Peng, 2005).

Third and most importantly, with this research we go beyond the perceptual measurement of the country-of-origin concept and apply it outside the marketing and behavioural field. Finally, we investigate whether the concept of country-of-origin holds explanatory value with regard to post-acquisition

performance of cross-border acquisitions and thereby answer the call for additional theory development and research into M&A research methods (King et al., 2004).

2.5.2 Definition and evidence

When MNEs enter into foreign countries through acquisitions, they do so from a liability of foreignness position. However as we argued previously, existing literature has led to ambiguous conclusions

regarding consequences of the country-of-origin effect. What is more, the country-of-origin effect could hold explanatory value for post-acquisition performance, since several scholars (King et al., 2004) indicate that unidentified variables may explain significant variance in post-acquisition performance.

Since 1965, the concept of the country-of-origin effect has been extensively studied by scholars (Dinnie, 2004). Initially Schooler concluded that the country-of-origin of a product can have an impact on consumers’ opinion of a product (Schooler, 1965). A significant conceptual advance was made by

(17)

One weakness of the studies mentioned above is the lack of focus on what would later become conceptualised as product-country image. For example, the ‘made in Scotland’ informational cue operates more effectively for traditional product categories such as whiskey rather than for new technology

product categories (Dinnie, 2004). In an effort to overcome this weakness, Papadopoulos contributed to the country-of-origin literature by criticising the concept of country-of-origin as being narrow and misleading, since it assumes a single place of origin for a product when a product may well be

manufactured in one country but designed, assembled and branded in another country (N. Papadopoulos & Heslop, 1993). In order to account for this multidimensional character Papadopoulos proposed the term “product-country image”.

Furthermore, Thakor and Kohli argue that consumer perception may differ from the actual location where the products carrying the brand name are manufactured (Thakor, 1996). Thus consumer perceptions may not coincide with reality for various reasons: ignorance, lack of salience of origin information for a particular brand, or deliberate obfuscation by companies concerned about consumer reactions to unfavourable origin (Dinnie, 2004).

The influence of culture on the country-of-origin effect was studied by Zhang. Zhang concluded that rather than cultural similarity acting as a positive factor influencing the choice of a product, it is the status of the country-of-origin as a developed economy that makes it attractive to buyers (Zhang, 1996).

In the context mentioned above, the country-of-origin effect refers to the extent to which the place of manufacturing influences perceptions of an organisation (Kinra, 2006). Moreover, Papadopoulos et al. have suggested that the perception of a company’s country-of-origin are based on three components associated with the standard attitude model, namely people’s “cognitions” which include knowledge about specific products and brands, “consumer affect” or favourable/unfavourable attitude towards the country-of-origin, and the “conative” behaviour which is related to the actual purchase of foreign brands (N. Papadopoulos & Heslop, 2014).

2.5.3 The Country-of-Origin effect on Acquisitions

(18)

poorer economic development, infrastructure bottlenecks, and inadequate legal rights (Chang et al., 2009). While reputations for products vary from country to country, people tend to generalise their attitudes and opinions across products from a given country. Which is based on their familiarity and background with that country. They do this based on their own personal experiences of product attributes such as

“technological superiority”, “product quality”, “design”, “value for money”, “status and esteem”, and “credibility of country-of-origin brand” (Kinra, 2006).

In line with international expansion theory, firms from emerging-markets should engage in acquisitions, as this mode of entry will provide them with access to market knowledge. Consequently, this reduces unfamiliarity and increases legitimacy of emerging-market firms. Thus acquisitions as a mode of entry should be used when an emerging-market firm experiences a negative country-of-origin effect (Eden & Miller, 2004; Lorenzen & Mahnke, 2002; Sea-Jin Chang & Philip M. Rosenzweig, 2001). Furthermore, according to Luo and Tung acquisitions are used primarily to secure brands (Luo & Tung, 2007). This strategy of buying-in establishes international reputations and accelerates market entry and legitimacy.

However, a persistent challenge for developing country firms as they globalise is the perception that their brands are inferior. Many factors contribute to this image, such as: a country’s economy (most countries with a positive country-of-origin image are developed), technological advancement (higher technological capability implies a positive country-of-origin effect resulting from the form of

government), the success of capitalism (usually a negative connotation follows countries that do not follow capitalism), and the reputation of the government and corporate governance (associated with bureaucracy and efficiency). Therefore, countries that are less developed tend to have a negative country-of-origin effect (Lumb & Lall, 2006).

Considering the literature, it is logical to assume that when a firm’s country-of-origin image is negative, there will be a negative perception among shareholders if an acquisition occurs, resulting in a negative effect on post-acquisition performance. Thus, the hypothesis is formulated as:

Hypothesis 1a: Cross-border acquisitions result in lower cumulative abnormal returns if the acquiring firm is perceived to be from a less developed country.

(19)

post-acquisition performance. Thus, the hypothesis is formulated as:

Hypothesis 1b: Cross-border acquisitions result in higher cumulative abnormal returns if the acquiring firm is perceived to be from a more developed country.

2.6 Integrated overview

All in all, we conclude that cross-border acquisitions have an effect on post-acquisition performance. Furthermore, we reasoned that the country-of-origin effect stems forth from a liability of foreignness position. We found evidence that the more developed the country-of-origin is, the more likely it is that a cross-border acquisition will result in higher cumulative abnormal returns. Subsequently, we provide the reader with a synopsis of our propositions in Figure 3:

(20)

3. Methodology

In this section, we describe the methods proposed to be used to answer our research question. We start with introducing the proposed research design, subsequently we present our data collection, and proposed method to analyse the data respectively.

3.1 Research Design

The research is designed based on a positivism assumption on the phenomena. Positivism holds that what can be perceived as true is based on the reality (Ryan, Scapens, & Theobald, 2002). Due to the positive assumption, which permits the formulation of hypotheses and the statistical testing of expected results to an acceptable level of probability, a deductive approach will be suitable to apply.

3.1.1 Research Strategy and Time Horizon

In order to have a better understanding on the influence of country-of-origin on post-acquisition performance, by testing hypotheses, a quantitative research will be employed. The application of

quantitative research implies a necessity to collect a quantity of data to establish a statistically significant result. Thus, our research will apply an archival research strategy, which is conducted based on existing material. Since we apply quantitative research, our research can be characterised as mono-method research.

(21)

Eastern Asia region.

3.2 Data Collection

3.2.1 Data Sources

For the analysis, several databases are proposed to be used to collect data, namely the Zephyr and Orbis Databases as well as Thomson Reuter’s DataStream. The first two databases aim to provide detailed information about cross-border acquisitions, such as deal announcement and complete dates, value, countries of acquiring and acquired companies, industry and other relevant information. Thomson Reuter’s DataStream provides detailed stock market information of selected companies. This database allows us to identify and examine trends, generate and test hypotheses, and develop viewpoints on the market. Finally, the data to measure the country-of-origin estimator of the acquiring country is derived from the World Governance Indicators and is provided by the World Bank.

3.2.2 Sample Selection

Based on information presented above we have proposed to use secondary data. The reasoning to choose for secondary data is time efficiency in the collection and analysing, compared to primary data. Collecting primary data from companies would be very time consuming.

For the purpose of eliminating alternative influences, we employ a number of criteria to determine an appropriate sample. First of all, the time span between 2009 till 2015 was proposed for a number of reasons, which are observed above. The nature of the research question makes it necessary to collect data of cross-border acquisitions, completed within the selected time-span. The acquiring company in turn should be listed on the stock market. Finally, an acquisition must be publicly disclosed in order to attract attention of capital markets and should exceed the threshold of 50 per cent ownership after the acquisition, while owning less than 50 per cent before the acquisition in order to solely keep cases where a majority position was acquired.

(22)

Furthermore, another reason to propose an investigation of cross-border acquisitions with a Central and Eastern Asian acquirer, is that the region accounts for a majority of FDI inflows and outflows, and remains relatively stable over time compared to other respective world regions (UNCTAD, 2016).

3.2.3 Sample Statistics

Our initial sample size comprised 698 cross-border acquisitions. However, in an attempt to ensure homogeneity of our sample, a number of observations needed to be eliminated. Firstly, in order to reduce the possibility for outliers, deals that targeted companies in ‘offshore financial centres’ - the Bermuda Islands and the Cayman islands - were removed. Secondly, the sample size was reduced further by excluding deals that could be characterised as ‘bankruptcy deals’ (based on deal value) - deals where the deal value was extremely low (often between €1,- and €10,-) in an effort to merely obtain the property rights of nearly bankrupt target companies - from the total sample. Finally, deals without sufficient information on variables (such as pre-deal total assets) after conducting a cross-check were omitted.

What is more, a visual (see Figure A1 in Appendix A) and manual check of the data in SPSS for possible outliers and extreme cases resulted in a further reduction of the dataset. More specifically, we identified several observations that could be characterised as outliers. However, according to Anscombe and Tukey we should not simply reject outliers (Anscombe & Tukey, 1963). Furthermore, Hoaglin and Iglewicz demonstrate that the interquartile range function used in SPSS might be inaccurate in up to 50 per cent of cases (Hoaglin & Iglewicz, 1987). An alternative measure was suggested to us by professor McCarthy of the University of Groningen, namely winsorization. From the literature it becomes apparent that only up to 5 per cent of data can be winsorized (Dixon, 1980), otherwise data will be distorted. Therefore, in line with (Field, 2009) we have opted to remove outliers identified by the larger interquartile range function (3), while winsorizing ±2 standard deviations. All of the above, resulted in a significantly reduced sample, containing the total sum of 182 observed cross-border acquisitions.

(23)

Figure 4: Distribution of Cross-border Acquisitions by Acquirer

(24)

As can be observed in Figure 4, cross-border acquisitions by acquirers in our sample are most frequently performed by Japanese, Korean, and Chinese companies subsequently. By far a company located in the United Kingdom is targeted by Central and Eastern Asian firms. The United Kingdom is followed by Germany, China and Malaysia correspondingly. However, the remaining half of the

acquisitions are spread over Western Europe and Central Eastern Asia, displaying an increasing scope of internationalisation.

Furthermore, Table 1 below presents an integrated overview of several specific transactional characteristics. More specifically, the table below provides information on the aspects of industry relatedness and the stake after the acquisition.

Table 1: Transactional Characteristics

Transaction characteristic Number of observations

Industry Relatedness

- Related - Unrelated

55 127

Stake after the acquisition

- 50 to 59 per cent - 60 to 69 per cent - 70 to 79 per cent - 80 to 89 per cent - 90 to 99 per cent - 100 per cent 13 13 17 11 14 114 3.3 Measurement of Variables 3.3.1 Dependent Variable

Based on our conceptual model, we can identify post-acquisition performance as the dependent variable. Acquisition performance is broadly prevailing in existing literature and is measured in one of three categories, namely accounting-based measures, socio-cultural integration outcomes, and stock-based measures (Verzhykovskiy, 2016).

(25)

and reporting standards differ across countries, we propose to exclude this method (Weetman & Gray, 1991). Next, socio-cultural integration outcomes (such as employee satisfaction and voluntary turnover) capture the level of conflict and therefore can be considered a measure for M&A success (Stahl & Voigt, 2004). However, this method requires extensive access to internal company information. Considering the limited time-span and resources of my research, we propose not to pursue this method.

Stock-based measures observe the stock market’s reactions (abnormal returns) to the acquirer’s share price after a takeover announcement (Stahl & Voigt, 2004). A weakness of this measure is that it is under assumption of efficient capital markets (share prices reflect all publicly available information immediately and help to conclude about expected outcome of takeovers) (Verzhykovskiy, 2016). Despite this weakness, we adopt this measure for the following reasons. First, this measure has been widely used in acquisition research (Moeller & Schlingemann, 2005). Secondly, stock-based measures are ex ante measures that correlate with ex post performance (Haleblian, Kim, & Rajagopalan, 2006). Finally, this method allows us to obtain widely available and verifiable data.

More precisely, we decided to use cumulative abnormal returns as a proxy for post-acquisition performance. Taking into account shareholder and market efficiency theories (Brown & Warner, 1980), we assume that the markets are efficient and any change in value of a firm is directly reflected in stock prices and subsequently in rates of return (Verzhykovskiy, 2016).

3.3.2 Independent Variable

Based on our conceptual model, we can identify cross-border acquisitions as the independent variable. We wish to measure cross-border acquisitions by the stake owned after the acquisition, in other words the level of control by the acquiring firm. This variable is obtained through the Zephyr Database respectively. Typically, an acquiring firm does not hold target shares prior to the offer, but held a majority of

outstanding target shares upon successful execution of the offer (Bradley, Desai, & Kim, 1988).

(26)

Moreover, it is recognised by Eden and Miller (Eden & Miller, 2004), amongst others, that the ownership strategy is a key decision when facing a liability of foreignness position. Finally, we are most confident in our selection of the independent variable because multiple scholars argue that a higher level of control (gained through higher stake in the target) could generate significantly more value for the acquirer (Aybar & Ficici, 2009; Chari, Ouimet, & Tesar, 2004).

3.3.3 Moderating Variable

Based on our conceptual model, we can identify the country-of-origin and its effect as the moderating variable. Within the literature review we have observed that there is no developed measurement, besides perceptual measurements. However, we concluded that while reputations for products vary from country to country, people tend to generalise their attitudes and opinions across products from a given country, based on their familiarity and background with that country. A persistent challenge for developing country firms as they globalise is the perception that their brands are inferior. Many factors contribute to this image, such as a country’s economy, but also the reputation of the government and corporate governance (associated with bureaucracy and efficiency).

Due to the abovementioned reasons, we have decided to focus on the World Governance

Indicators (WGI) as a proxy for the level of development for the country under investigation with regard to the country-of-origin effect. Our decision is justified because existing literature concluded that rather than cultural similarity acting as a positive factor influencing the choice of a shareholder, it is the status of the country-of-origin as a developed economy that makes it attractive to buyers (Zhang, 1996). Finally, the WGI data sources comprise both perceptual and non-perceptual measures (Worldbank Group, 2017), which will complement the original perceptual nature of the country-of-origin concept, while adding a non-perceptual measurement as well.

The WGI were created to report aggregate and individual governance indicators for over 200 countries, divided over six dimensions of governance (Worldbank Group, 2017). The WGI are especially valuable for cross-country comparison because it scales the six aggregate WGI indicators in standard normal units, ranging from -2.5 to 2.5 (Kaufmann, Kraay, & Mastruzzi, 2009). The six aggregate indicators are the following (see Appendix C for the complete description):

● Voice and Accountability.

(27)

● Government Effectiveness. ● Regulatory Quality. ● Rule of Law.

● Control of Corruption.

We have observed that each of the aggregate indicators would be able to measure a specific aspect of the governance of a country (which we will discuss later on). However, since our research question infers an analysis of the country-of-origin concept as a whole we wish to compute a sum variable of the six aggregate indicators, in an effort to measure a country’s overall development using the WGI. In order to assess the degree to which the six aggregate indicators of the overall concept are consistent with each other, a reliability analysis is performed (see Figure A2 of Appendix A). After which we computed the sum variable ‘WGI’ to measure the overall concept of country-of-origin.

3.3.4 Control Variables

Post-acquisition performance may be influenced by a number of variables other than the country-of-origin effect. Thus we conducted extensive research and propose to include the following control variables in our research.

Relative Size, a variety of papers argue that the size of an acquisition company can affect

post-acquisition performance (King et al., 2004). Therefore, we propose to measure relative size by calculating the acquiring and acquired companies’ total assets in the year before acquisition, and dividing the acquired company’s total assets by acquiring company’s total assets.

Industry Relatedness, this variable aims to control whether companies which are operating in the

same industry impact post-acquisition performance, as several scholars argue (King et al., 2004). We include a dummy variable, value “1” when companies are in the same or related industries, while a value of “0” appears when companies operate in opposite industries. In line with Datta, we compare SIC standard codes to define industry relatedness (Datta, 1991).

3.4 Estimation and Model of Proposed Analysis

(28)

is the difference between the expected return and actual return of a stock. The cumulative abnormal return then is the sum of those abnormal returns over a specified event window. The aim of cumulative

abnormal return is to determine the effect that events such as the announcement of an acquisition have on stock prices.

As abovementioned, cumulative abnormal returns are calculated over a specified event window. Therefore, the calculation of cumulative abnormal returns occurs through an event-study. Since such an event-study follows market efficiency theory, where we expect that particular events, such as the announcement of an acquisition in our case, can influence share prices (Brown & Warner, 1980).

An event study methodology employs time-series, which are divided into several phases as can be seen in Figure 6. More specifically, the estimation window, event window, and post event window. The several phases serve a specific purpose in event study methodology, on which we will elaborate further in the following sections. However, we exclude the post event window, since it is employed to measure long-term performance.

Figure 6: Event study time-series

(MacKinlay, 1997)

The estimation window provides the information necessary to identify the normal returns (Verzhykovskiy, 2016). Pursuing Doukas and Travlos, we compute the daily returns according to the following formula:

Figure 7: Daily returns model

α R

R_xt = x+ βx mt+ Ext

(29)

Where,

R_xt Expected rate of return for stock _xon event day _t ,

αx βx Parameter values for stock x

R_mt Return of portfolio on event day _t

E_xt Standard error for stock _x on event day _t

In accordance with several authors (Doukas & Travlos, 1988; Verzhykovskiy, 2016), we have opted for an estimation window of 256 trading days. More specifically, the calculation is performed over the period between 256 (T0) and 1 (T1) trading day(s) before the announcement [-256;-1]. For all observations we employed the Thomson Reuter’s Datastream to calculate target stock prices and market portfolio. Furthermore, the target stock prices and market portfolio were subject to a check for reliability of the data, by means of visual inspection of major stock indices on which our observations were listed. After the model definition, we computed abnormal returns using the following formula:

Figure 8: Abnormal returns model

α R )

AR_xt = R_xt− ( x+ βx mt + Ext

According to (Doukas & Travlos, 1988)

Where, R

A _xt Daily abnormal returns for stock _xon event day _t

R_xt Expected rate of return for stock _xon event day _t ,

αx βx Parameter values for stock x

R_mt Return of portfolio on event day _t

(30)

As specified previously, the cumulative abnormal returns are the sum of the abnormal returns over a specified event window. A crucial step in the research design is the length of an event window used in an event study (McWilliams & Siegel, 1997). More specifically, Mcwilliams and Siegel argue that long event windows severely reduce the power of the test statistic, which in turn can lead to false

inferences about the significance of an event (McWilliams & Siegel, 1997). Therefore, we have opted for a short event window of 5 days (T2) after the event, in addition to the event date itself (T1) [0;+5]. Finally, we computed the cumulative abnormal returns according to the following formula:

Figure 9: Cumulative abnormal returns model

CARx = ∑

N x = 1

ARx

According to (Doukas & Travlos, 1988)

Where,

CARx Cumulative abnormal return for stock x over the event window

Finally, this section explains the statistical tests used to test the hypotheses. For our study we propose to employ a special type of regression model, namely moderation analysis. A moderation analysis is a multiple regression model used to determine a moderator effect and whether it is statistically significant (Aguinis, 2004). You can observe the empirical model for our research in the following figure:

Figure 10: Empirical Model

) (SaA ) (IR ) (RS )

CAR_j = β₀+ β1(SaA )_j + β₂(W GI_j + β₃ _j _{* W GI}_j + β₄ _j + β₅ _j + ε

Where,

CAR_j Cumulative abnormal returns in acquisition _j

(31)

W GI_j World Governance Indicator in acquisition _j

Ind_j_{* M}_j Interaction effect between independent and moderator variable in acquisition _j

IR_j Industry relatedness control variable in acquisition _j

RS_j Relative size control variable in acquisition _j

We have opted to propose a moderation analysis for a variety of reasons. First of all, it is within the nature of our research question to discover a relationship among our variables, as illustrated in the integrated synopsis (Aguinis, 2004). Secondly, our independent, dependent, and moderator variables are continuous in nature, implying that the variables have an infinite number of possible values (Aguinis, 2004).

(32)

4. Analysis

In this section we cover the empirical results of our research paper. We initiate this chapter with the descriptive statistics, after which we will examine the correlation analysis. Before we proceed to the outcome of the regression analysis, we investigate the underlying assumptions of the moderation analysis.

4.1 Descriptive Statistics

As can be observed in Table 2, our final sample consisted out of 182 observations of cross-border acquisitions, containing the necessary values for all required variables. The total of 182 cross-border acquisitions are distributed over 11 ‘acquirer’ country’s in the Central and Eastern Asian region. Since 2009, the most common ‘acquirer’ countries are Japan (N=54), the Republic of Korea (N=34), and China (N=28). The United Kingdom (N=34) is by far most frequently targeted among all targets in Western Europe and Central and Eastern Asia.

Post-acquisition performance (N=182), as measured by the cumulative abnormal returns, ranged from 22.20 per cent to -17.23 per cent (M=0.35, SD=7.46). The largest cumulative abnormal returns are achieved by acquirers in Singapore, China, and Indonesia, while the lowest cumulative abnormal returns are also reported by an acquirer in Singapore.

(33)

developed country, in our sample has been observed by a firm in Taiwan in the year of 2015.

Finally, the stake after the acquisition held by the acquiring company ranged from 50 per cent to 100 per cent (M=89.99, SD=15.80). The most common observation in our sample is 100 per cent (N=114). This would entail that the majority of cross-border acquisitions by Central and Eastern Asian firms were a full acquisition.

4.2 Correlation

In an attempt to obtain more information about the preliminary indicators of data patterns, we have performed a Pearson correlation test. Table 3 reports the main correlations. The significant correlations are marked with one (P<0.05) asterisk. As can be observed from the table, there are two significant correlations. In particular, there is a significant positive relationship between our moderator (

) and cumulative abnormal returns (r=0.169, P<0.05). This preliminary indicator suggests

SaA_j _{* W GI}_j

that an increase in the stake after the acquisition leads to an increase in cumulative abnormal returns, and this relationship will be stronger when the level of development of the acquiring country is higher as opposed to low. This indicates that the moderating effect as hypothesised in H1a and H1b may be supported in subsequent analysis. In addition, there is a significant negative relationship between our control variable relative size and cumulative abnormal returns (r=-0.167, P<0.05). This indicates that an increase in the relative size of the target firm respective to the acquiring firm may lead to a decrease in cumulative abnormal returns. In other words, we would expect a comparatively larger acquirer to have higher post-acquisition performance.

On the other hand, the stake after an acquisition does not have a statistically significant

relationship with any of the variables. What is more, we observe no correlation for the World Governance Indicators when treated as an independent variable. Moreover, the level of control, measured by the stake after the acquisition, also does not appear to correlate to any other variable. Finally, the industry

(34)

4.3 Testing Assumptions

When conducting statistical analysis, different statistical models assume some certain characteristics about the data, also known as assumptions. If these models are going to reflect reality accurately then these assumptions need to be true, violation of these assumptions changes the conclusion of research and interpretation of results (Field, 2009). In this section we will briefly deal with the assumptions of

moderation analysis, as specified by Aguinis (Aguinis, 2004). More specifically, we will deal with multicollinearity, linearity, homoscedasticity, outliers, and normality.

Firstly, while examining multicollinearity we wish to find no evidence of a high correlation between independent and dependent variable. While examining the Pearson’s correlation coefficients in Table 3 we did not observe a value of 0.8 or higher. This provides a strong indication that the assumption is met. However, in order to check for multicollinearity, a VIF test is conducted (see Table A1 in

Appendix B). VIF scores above 10 would indicate strong evidence of multicollinearity. As evidenced in Table A1, there are no scores greater than 10 and therefore we have sufficient evidence that

multicollinearity is not an issue.

(35)

acquisitions would have occurred.

Thirdly, we must assess the assumption of homoscedasticity. To this end, we inspected a plot of the studentised residuals against the dependent variable. We can observe that the residual values are normally distributed in Figure A4 (in Appendix B) and therefore conclude we meet the assumption.

Fourthly, we must examine our data for outliers. However, as mentioned previously we incorporated this step at an earlier stage in an effort to alleviate the possibility of alternative influences. More specifically, we conducted a visual and manual check and investigated several observations that could be characterised as outliers. In accordance with several scholars (Anscombe & Tukey, 1963; Dixon, 1980; Field, 2009; Hoaglin & Iglewicz, 1987) we opted to remove outliers identified by the larger

interquartile range function (3), while winsorizing ±2 standard deviations. Still, a new visual inspection of the data in SPSS demonstrated several outliers following the smaller interquartile range function (1.5) (see Figure A5 in Appendix B). However, as established previously the nature of cumulative abnormal returns implies we should not simply reject outliers (Anscombe & Tukey, 1963), and it was demonstrated that the interquartile range function in SPSS might be inaccurate (Hoaglin & Iglewicz, 1987). Therefore, we conclude that we have a good model without outliers, following our previous definitions.

Finally, we will examine whether we meet the assumption of normality. Normality entails that the distribution of the test is normally distributed. In our pursuit of normality we have performed a

Kolmogorov-Smirnov test (see Table A2 in Appendix B) and will interpret our results using a Q-Q plot (see Figure A6 in Appendix B). From Table A2 we can conclude that we fail the normality assumption. Further inspection of Figure A5 reveals that positive outliers might be the cause of a non-normal distribution. However, once again the nature of cumulative abnormal returns implies that we should not simply reject outliers (Anscombe & Tukey, 1963).

(36)

4.4 Regression Results

Initially in an effort to observe the impact of the country-of-origin concept, we ran a moderation analysis of each of the six aggregate World Governance Indicators. We argue that each of the aggregate indicators would be able to measure a specific aspect of the governance of a country and further investigation is warranted. Our regression analyses are presented in Table A3 to A8 in Appendix D. Model 1 merely includes control variables, whilst Model 2 includes the predictor variables. Finally, Model 3 includes the interaction effect between the independent and predictor variables. More specifically, the interaction effect for Voice and Accountability, Political Stability, Government Effectiveness, Regulatory Quality, Rule of Law and, Control of Corruption are presented in Table A3 to Table A8 respectively.

From our regression analyses we observe that the control variable Relative Size has a significant influence on post-acquisition performance in Tables A3 to A8, throughout all models. On the other hand, our control variable Industry Relatedness and every single predictor variable does not show any

statistically significant influences. However, overall the stake after the acquisition is negatively associated with post-acquisition performance, while all aggregate indicators except Voice and Accountability are positively associated with cumulative abnormal returns when the aggregate indicators are treated as predictor variables.

Finally, in an attempt to check whether the country-of-origin effect indeed moderates the relationship between our independent and dependent variable, we examine Model 3 for every aggregate indicator. First, the moderator for Voice and Accountability has a positive significant influence on cumulative abnormal returns (B=1.076, P<0.1). Secondly, the moderator Political Stability is not statistically significant. Third, the Government Effectiveness moderator is statistically significant (B=1.130, P<0.1). Fourth, the moderator for Regulatory Quality is statistically significant (B=1.132, P<0.05). Fifth, the Rule of Law moderator has a statistically significant influence on cumulative abnormal returns (B=1.136, P<0.1). Lastly, the moderator for Control of Corruption significantly influences

post-acquisition performance (B=1.066, P<0.1). In the end we computed a sum variable of the six aggregate indicators, in order to measure a country’s overall development using the WGI.

(37)

interaction effect between the independent variable and the moderator variable. Therefore, Model 3 specifically tests for the hypothesised effect of the level of development of the acquiring country on the cumulative abnormal returns, while Model 2 tests the direct effects that the predictor variable might have on the dependent variable. The various fit parameters show that our full models fit the data better, the R2 improves from 4 per cent in Model 1 to 6.4 per cent in Model 3. Furthermore, the improved adjusted R2, from 2.9 per cent to 3.8 per cent, allows us to conclude that the explanatory power of our model is not simply the result of more predictor variables but overall fit of the model.

(38)

As abovementioned, the second model introduces the independent and moderating variables as predictor variables. We observe a negative relationship between the Stake after the Acquisition and cumulative abnormal returns. This indicates that, in line with several scholars (Alexandridis et al., 2010; Datta & Puia, 1995; Deng, 2010; King et al., 2004), cross-border acquisitions would result in negative performance. More specifically, a higher level of control would result in lower cumulative abnormal returns. Furthermore, when the moderator is treated as an independent predictor variable, we find a positive relationship between the acquiring country level of development and cumulative abnormal returns. However, we must conclude that the effect of the both predictor variables independently has an insignificant effect on the dependent variable.

In an effort to check whether the acquiring country level of development indeed moderates the relationship between our independent and dependent variable, we wish to examine Model 3. A closer examination of Model 3 in Table 4 reveals there is a statistically significant result of the acquiring country level of development on cumulative abnormal returns (B=1.152, P<0.05). This means that a higher level of control will lead to an increase in cumulative abnormal returns, and this relation will be even more pronounced when the acquiring country level of development is higher. Since we are dealing with a non-normal distribution, an inspection of the R2 indicates 6.4 per cent of effects can be explained by Model 3. These findings preliminary confirm Hypothesis 1a which states that cross-border acquisitions result in lower cumulative abnormal returns if the acquiring firm is perceived to be from a less developed country. While simultaneously providing support for Hypothesis 1b which states that cross-border acquisitions result in higher cumulative abnormal returns if the acquiring firm is perceived to be from a more developed country.

Although the analysis provided above can be utilised to draw conclusions about our sample (Field, 2009), we are not yet able to generalise these findings to the general population due to the non-normal distribution of data in our sample. Therefore, we have decided to delete nine cases

(39)

A careful examination of the new regression in Table 5 reveals that a negative relationship between our control variable Industry Relatedness and our dependent variable. From this we can infer that if the acquired and acquiring company are from a similar industry, we would expect a decrease in

performance. Furthermore, our control variable Relative Size appears to have a positive relationship with cumulative abnormal returns. This would imply that an increase in the relative size of the acquired company when compared to the acquiring company, entails an increase in cumulative abnormal returns. However, neither of the control variables shows any statistically significant influences.

In Model 2 we observe a positive relationship between the Stake after the Acquisition and cumulative abnormal returns. What is more, Model 2 also reveals that there is a positive relationship between the acquiring country level of development and the dependent variable. Once more, both variables do not have any statistical significant influences.

What is more, in order to examine whether the acquiring country level of development moderates the relationship between our independent and dependent variable, we continue to investigate Model 3 in Table 5. Once more, a close examination reveals a significant positive relationship between our

(40)

control will lead to an increase in cumulative abnormal returns, and this relation will be even stronger when the acquiring country level of development is higher. We would also observe that the full models fit the data better, as R2 improves from 0.6 per cent to 2.9 per cent. However, the final value for the adjusted R2 is 0 per cent, which leads us to believe that the relation merely holds explanatory power because we added more predictor variables and not because of the real explanatory power of the variables itself. Which means we could accept Hypothesis 1a and Hypothesis 1b for the general population, but our model doesn’t hold high explanatory value of post-acquisition performance variance.

Finally, the fit parameters for a general population show that our full models fit the data better, the adjusted R2 improves from -0.006 per cent in Model 1 to 0 per cent in Model 3. However, the negative result would imply that Model 1 and Model 2 do not hold any power to explain the effects. Finally, an inspection of Model 3 reveals that the explanatory power is no longer negative, but accounts for precisely 0 per cent of all effects. Therefore, the result in Table 5 imply that our model might no longer holds explanatory value for a general population, and therefore, we propose to initiate an investigation of of the removed cases in the discussion section.

(41)

5. Conclusions

The following section presents a discussion on the outcomes of the empirical analysis. Subsequently we will address the theoretical and managerial implications of our research. Finally we will address the limitations of our study and highlight possibilities for further research.

5.1 Discussion

We have observed that technological development and globalisation have vastly contributed to the

popularity of cross-border acquisitions (Shimizu, Hitt, Vaidyanath, & Pisano, 2004). A controversial issue regarding acquisitions is whether acquiring firms create value for their shareholders. Although empirical pieces of evidence are widely available, theory prevailing in existing literature might not be able to adequately explain post-acquisition performance. In line with King et al. we have suggested that unidentified variables may explain significant variance in post-acquisition performance. What is more, several scholars argue that theories on international expansion do not provide the same insights into the behaviour of EM MNEs as DM MNEs (Axinn & Matthyssens, 2002).

Therefore, we proposed a more thorough investigation of the country-of-origin concept and whether the concept can significantly explain a portion of the variance in post-acquisition performance. More specifically, our research aimed to address the research question whether the country-of-origin influences the cumulative abnormal returns of cross-border acquisitions.

Through an empirical analysis using a sample of 182 cross-border acquisitions (187 collected) of Central and Eastern Asian firms into Western Europe and Central and Eastern Asia between 2009 and 2015, this research sought a linkage between cross-border acquisitions and post-acquisition performance. In line with institutional theory we investigated whether the country-of-origin concept, measured by the acquiring country’s level of development, moderates the relationship between the final stake after an acquisition and cumulative abnormal returns.

The Country-of-Origin effect