1
University of Groningen
& Newcastle University Business School
Master Thesis
The impact of institutional distance on
stock market success of acquisitions
by Jesley Hagedoorn
Groningen student number: S2974959
Newcastle student number: B9018545
Supervisors: Joshua Haist & Rudi de Vries
Date: December 2, 2019
2
The impact of institutional distance on stock market success of
acquisitions
Abstract: Foreign acquisition deals have grown in popularity due to the widespread
trends of globalisation. While there is ample research available regarding success
factors of these deals, there is little attention for its effect on the stock market. This
paper researches the effect of institutional distance between the acquirer and the
target on stock market response after acquisition announcements. An event study
was conducted including 500 cross-border acquisition deals from 95 different
nationalities in the period of 2008-2019. The study uses stratified sampling to
maximise the variety of institutional distance across the sample. It was found that
unlike research in the past, institutional distance may no longer a significant barrier
for acquisition success. While positive stock market response was widespread, this
positive response was not found to be affected by institutional distance. A minor
negative relationship between institutional distance and stock market response is
suspected, but no significant evidence is found. While there may be several
alternative explanations for this, it does imply that institutional distance between
parties might no longer be primary factor that defines stock price reactions following
an announced acquisition.
3
Table of Content
1. Institutional Distance and stock market success of acquisitions ... 4
2. Literature Review ... 7
Institutional Distance & Alliances Theory ... 7
Mergers, Acquisitions & Alliance Success Theory ...12
3. Hypotheses & Conceptual Model ...14
4. Methodology ...19 Research Strategy ...19 Ethical Considerations ...20 Variables ...20 Sampling ...27 Analysis ...30 5. Results ...33
6. Discussion & Conclusion ...39
Conclusive remarks ...41
Limitations & Future Research ...43
7. Bibliography ...46
4
1. Institutional Distance and stock market success of acquisitions
Cross-border mergers and acquisitions have grown in popularity since the turn of the decade, as discussed by Hitt et al (2009). However, mergers and acquisitions tend to
succeed only half of the time (Bauer & Matzler, 2014), and the success rate is even lower for international M&As (Rottig & Reus, 2018). This shows that international cooperation can be a complex and difficult concept to understand and implement. Therefore, it is more and more important for firms to understand the success factors that lead to maximisation of gain from such an international alliance, merger or acquisition. This increase of cross-border activity is illustrated by the growing amount of research conducted on the topic.
One concept that has been studied extensively in business is the concept of distance, as pointed out by Lavie and Miller (2008). As such, it has been suggested that multinational firms suffer from liability of foreignness. The reason for this is that firms are unfamiliar with the local environment and the various differences between home and host country (Eden & Miller, 2004). This leads to a disadvantage for the foreign firm, as they will incur additional costs compared to local firms. Such costs can be divided in four groups; transportation costs; costs based on the unfamiliarity of the firm in the new country; costs resulting from home country environment pressures; and costs resulting from pressures related to the host country environment (Zaheer, 1995).
Zhang et al (2010) found that whether an alliance is international can have a positive impact on how well new knowledge is able to result in innovative performance. The research did not actually find evidence on the amount of knowledge creation or knowledge acquisition itself. This suggests that it is possible that international alliances have an indirect positive effect on innovation rather than a direct one.
Research has been conducted on the actual success of alliances and specifically mergers and acquisitions. Bauer & Matzler (2014) found that strategic and cultural fit and
5 This suggests that a (national) institutional framework of a country may have a similar impact on alliances, as its concept is similar. Therefore, further research on the effect of the country-level institutional environment that the firm resides in could yield noteworthy results.
This paper focuses on how institutional differences or institutional distance affect the success of an acquisition in the short term. Institutional distance in this case relates to the distance between the institutional environment of the firm of the acquirer and the institutional
environment of the target. This research is conducted by focussing on stock price movement after the announcement of such an acquisition, based on the effect of institutional distance.
This potential relationship between stock price shifts and institutional distance is important to examine. One reason for this is that the stock market was proven to have at least some extent of predictive power for firm success (Guo, 2002). Moreover, it may show either trust or distrust that investors have in the firm following a certain acquisition. After all, shifts in stock value of the firm show the outside perception of the business prospects of the firm. It is therefore a good indication of the expectations that people outside the firm have from the firm. As pointed out by Rosen (2006), investors will have a certain grasp of whether they expect the merger or acquisition to result in positive benefits such as synergies. This should especially be the case in an efficient market.
Lastly, Kale et al (2002) found a correlation between alliance success based on stock
performance and alliance success measured through managerial assessment. This validates the fluctuation of stock value as a predictor of alliance success. Factors that impact such a stock price fluctuation have already received significant attention (Campbell et al, 2016). Many factors were examined previously, such as method of payment (Loughran & Vijh, 1997) and acquisition premium (Schijven & Hitt, 2012). However, these were found to be insignificant predictors. It is important to find factors that do actually affect these investor reactions. Institutional distance is potentially such a factor.
6 This paper offers some unique contributions to acquisition literature. The research uses multiple measurements of institutional distance and multiple measurements of stock price shift. In that sense, this research offers a unique method to assess a potential relationship between institutional distance and stock price shift after an acquisition announcement. As prior research struggled to find explanations for stock price fluctuations after announcements (Campbell et al, 2016), there is clearly a research gap that this paper attempts to fill.
The research question associated with this paper is formulated as follows:
How does institutional distance between both parties of an acquisition affect stock price reaction to the announcement?
This introduction is followed by an overview of the literature in the field of institutional
7
2. Literature Review
Institutional Distance & Alliances Theory
First of all, it is useful to provide a definition of institutional distance. Kostova (1996) defined institutional distance as ‘’the extent of similarity or difference between a host country and a home country in its institutional context.’’ (Xu & Shenkar, 2002: 608). Kostova (1999) described that institutional distance can be divided into three different pillars, namely regulatory distance, cognitive distance and normative distance. Each of these pillars has their own impact on firm behaviour.
Scott (1995) defined that the regulative pillar of institutions refers to rules and laws that ensure the order in a society or economy, while the normative pillar of institutions refers to the norms and values that exist in a society to govern people’s behaviour. Lastly, the cognitive pillar refers to the cognitive processes that relate to the nature of reality and the meaning that is given to it. (Xu et al, 2004). These three pillars together form the institutional framework in which a firm has to survive. Differences between these pillars in different environments together form the concept of institutional distance. In order to form
expectations regarding the research question, it is important to assess the prior literature regarding the impact of this concept of institutional distance on alliances.
Moreover, it is important to be able to measure institutional distance. Kostova (1997) was the first notable research to make an attempt of a measurement of institutional quality. The research resulted in an 18-item questionnaire using 7-point Likert scales. As it included only 10 countries, it needed substantial future extension. Yet, it was the first step towards an overview of institutional quality on a national level. A more recent method of measuring institutional distance using institutional quality was created by Kuncic (2014). Kuncic divided the world in five groups based on institutional strength, with the first group signalling the weakest institutional environment, while the fifth group consists of countries with the strongest institutional frameworks. Kuncic came up with such a measure of institutional quality by collecting a large set of institutional proxies and carefully analysing these per country. Using this method, the author divided the world in five groups based on institutional quality.
8 The researchers involved in this project collected data on 6 different variables that together make up institutional quality. This has been used by researchers before in order to measure institutional distance. For instance, Liou et al (2016) were able to create a measurement of institutional quality and distance combining the data of the six variables. These authors used it to assess the effect of institutional distance on ownership and were able to find a
relationship.
The concept of institutional distance has been proven to have an effect on other aspects of business, as well such as entry mode (Xu & Shenkar, 2002) and the likelihood to enter a foreign country at all (Holburn & Zelner, 2010). A short overview will now be provided outlining the established research on this subject.
Research in international business increasingly focusses on variations regarding the institutional environment (Meyer et al, 2011). Meyer et al (2009) found that differences in particularly the quality of institutional frameworks may lead to differences in firm behaviour. Institutional arrangements can be considered ‘strong’ if they in a way support the natural market mechanism in a country, and logically ‘weak’ if they fail to ensure such a market mechanism or even undermine it, in some way or another. One example of the effect of institutional quality on firm behaviour is the selection of entry mode: A stronger institutional environment will encourage acquisition and greenfield entry of a foreign country, rather than entering through a joint venture. (Meyer et al, 2009).
Brouthers (2013) has conducted further research on entry mode of firms, based on the institutional environment among other factors. Entry mode selection was found to be
influenced directly by the environment and context (both institutional and cultural) of the host country. When there are fewer legal restrictions, there is logically less legal control in a country. It was found that in an environment with less control, firms prefer a wholly owned mode of entry. On the other hand, if there are a higher number of legal restrictions, a joint venture was found to be common. Countries with lower institutional quality will most likely be entered through a wholly owned subsidiary instead of a joint venture mode. While both papers may have slightly contrasting conclusions, the papers by Meyer et al (2009) and Brouthers (2013) show that a joint venture mode is unlikely when institutional differences increase. This signals that firms would prefer more control and less cooperation when such differences in institutional frameworks are significant.
9 equity alliances in environments that are less restrictive. As pointed out by the authors, firms from emerging markets are knowledgeable about doing business in less restrictive countries, meaning they may actually have an advantage in this foreign market over firms from
countries with high institutional quality.
Following this logic, a firm from Indonesia may have an advantage over a firm from the United Kingdom when investing in a Chinese subsidiary. According to this research by Ang & Michailova, the reason relates to prior experience of the Indonesian firm with low institutional quality or institutional voids in the home country. Institutional voids refer to a lack of the rule of law in a certain environment. (Carney et al, 2009). These are more present in countries with lower institutional quality such as China and Indonesia, meaning an Indonesian firm would have knowledge regarding dealing with them. However, the firm from the UK would be less experienced with business in such an environment, as these institutional voids are not as present in the home country. Therefore, a UK firm would have a disadvantage in such an environment, making a successful acquisition less likely.
Yet, this finding is not entirely in line with findings by Wu & Chen (2013). These authors researched the behaviour of firms from emerging markets and how the institutional
environment in the home country affects the prospects in developed economies. It was found that more developed institutions in the home country lead to lower government intervention and therefore a more natural flowing market. In that case, firms will be more prepared for doing business in advanced economies. Although if there is significant institutional instability in the home country, this will limit the development of the emerging market firm. This is even the case if it relates to quick development of the institutional framework in the country. The prospects of these firms in developed countries will in that case be limited.
It should also be noted that unlike firms, institutional systems are not mobile (Mudambi & Nevarra, 2002). MNEs will have to adapt their organisations and governance to suit the new institutional environment (Meyer et al, 2011). As a result, every location is associated with costs and benefits in the new institutional environment. Following this notion is the finding that institutional frameworks affect the location choice of a firm, as found by Bevan et al (2004), This results in international or even local competition, as countries attempt to make business as attractive as possible, for instance by providing measures against corruption (Meyer & Nguyen, 2005).
self-10 enforcing such as credit and long-distance trade. According to the authors, a reason for this is that strong and stable institutions reduce opportunism in high-risk transactions. Moreover, these institutions provide a template for multilateral creation of reputation supported by frameworks of coordination, enforcement and credible commitment. Together, this shows how important the institutional environment may be for businesses, particularly for
multinationals.
In line with this is research conducted by Xu & Shenkar (2002), who also emphasise how institutions within and across countries play a massive role in the success of an organisation. Institutional theory says that organisations must conform to the rules and belief systems prevailing in the environment, in order to gain legitimacy and thereby success in the market. This relates to isomorphism (Dacin, 1997), which is the idea that adhering to the same institutional norms will create structural similarities. Therefore, firms may be more
interconnected within a nation than across different nations. This implies that conducting business across borders may be significantly more difficult than within a country, given that the different institutional environment will result in differences in firm behaviour.
As noted, institutional distance roughly describes the differences between institutional environments of two different countries (Kostova, 1999). It was proposed that this concept of institutional distance is a significant barrier for successful transfer of organisational practice and thereby will hinder business success. Institutional distance in this case is defined as the difference between the institutional profiles of the two countries involved in the deal. Kostova (1999) found several explanations for the idea that such international distance may
negatively affect business success. These reasons relate to the three different pillars that were established at the start of this section. The regulatory dimension is that, when the regulations affecting a certain practice are in severe conflict with ones in the home country, there will be increased difficulties of implementing these practices. The cognitive distance in turn will result in an increased difficulty of interpretation of events related to institutions. Moreover, it makes it harder to get accustomed to practices. Thirdly, practices need to be consistent with different assumptions and value systems of the national cultures of the host country. This is the normative perspective of why institutional distance could affect the extent of business success. In the context of this research, it might also affect the success of an alliance or acquisition.
11 foreign parents. The foreign parents need to maximise control in this case. The authors found that there are two factors that can impact the success and survival of these
subsidiaries. The first factor is host country experience and the second one is institutional distance. Interestingly, these authors found a reversed U-form relationship between institutional distance and subsidiary survival. In the optimal situation, there needs to be at least a small amount of institutional distance. After all, these researchers found that a certain amount of institutional distance may be positive. However, high institutional distance would be counterproductive, therefore leading to a reversed U-formed relationship. It is important to consider that these authors specifically focussed on Japan and on subsidiaries of
multinationals in Japan. This may have an impact on the results.
Research suggests that similarly to institutions, national culture also plays a role in the success of mergers & acquisitions as both firms will need to culturally fit together (Weber, 1996), in line with how differences in institutional environments may affect the success. This would be something that investors will also be aware of, meaning that the other forms of distance could also affect stock market performance after an announcement.
In summary, institutional distance can be seen as a hindrance for exploitation, but also as an advantage for exploration (Krammer, 2018). On one hand, distant firms will be less attractive, as a large number of resources will need to be devoted to understand and deal with the severe difficulties related to the institutional distance (Chan & Makino, 2007). On the other hand, these cognitive and normative differences between two partners may also be desirable if the firm is looking for knowledge that is different from their current knowledge sets. The reasoning is that more cognitive and normative differences result in more heterogeneous resources, thereby increasing the absorptive capacity of the firm (Krammer, 2018). It was found that while such deals involving severe institutional distance between partners is likely to result in difficulties, such differences may be beneficial as well.
12
Mergers, Acquisitions & Alliance Success Theory
Literature on stock response to mergers and acquisitions remains relatively limited, yet there is ample research regarding success factors of particularly alliances.
Among the research that is available regarding stock data, Das et al (1999) researched the impact of alliance announcements on firm valuation, a research that resembles the one of this paper. The authors note, that as claimed by Kogut (1988), firms enter alliances mostly for long-term strategic considerations, rather than short-term quick return. However, as Das et al (1999) pointed out, long-term advantage should be recognised by an efficient stock market, therefore resulting in a shift in stock market value shortly after any announcement. The authors did find abnormal returns for regular alliance announcements, but these returns were not statistically significant. Yet, they did note differences based on different types of alliances, with announcements of technological alliances resulting in the largest returns while marketing alliances had lower returns. The explanation that was offered here is that
technological alliances may be perceived as more attractive in the long-term for investors.
Rosen performed a similar research in 2006, which was specifically focused on mergers and the momentum in the market. It was found that the moment of announcement is another factor for positive response by the stock market. When the stock market is so-called ‘hot’, meaning there is more of a positive reaction to mergers than usually, the market responds better to merger announcements in general. Although this research specifically focused on mergers, expectations would be similar for acquisitions, meaning that the general stock market trends may play a role.
A more recent research related to firm value shifts and alliances was conducted by Cabral and Almeida (2019). These authors found that alliances increase the dispersion of the distribution of firm value. Aspects such as resource complementarity and partner commitment and compatibility were found to positively impact firm value resulting from alliances. This suggests that institutional distance may play a part here as well, given that more similarity of the firms is likely to impact the compatibility of partners.
13 Hitt et al (2009) emphasized the importance of synergies as well. These authors argued that in some cases, firms will be prepared to pay a premium price for an acquisition, given that there are potential synergies that may be created through the merger of two firms. However, it can be difficult to realise such synergies because of the challenges of achieving integration. As a result, those deals involving a large premium often are less successful than expected (Datta et al, 1992).
Whipple & Frankel (2002) also researched several alliance success factors, but from a supply chain perspective. They found that the most common reason for alliance failure is organisational problems, often related to the structure of the firm. Technical and financial problems are less often the reason for failure. These organisation problems relate to firms sticking to their own habits rather than adopting a new way of business more complimentary to that of their business partner. Other factors that are crucial for a successful partnership are the presence of trust, senior management support, clear goals and the ability to actually fulfil these goals.
Furthermore, Anand & Khana (2000) found that firms can ‘learn’ how to manage an alliance optimally. It was found that prior experience with all forms of alliances can enhance the success of future alliances, particularly if the partnership is concerned with R&D or
technology alliances. Also, these authors found that so-called ‘alliance capabilities’ may be a factor. For instance, some firms are better at creating value from alliances than others. This finding has been backed up by other researchers, as for instance Heimeriks & Duysters (2007) argue that experience helps firms to deal with issues related to alliances. Another finding is that experience allows firms to select more appropriate partners (Simonin, 1997) and moreover it increases their ability to ease conflict situations (Mohr & Spekman, 1994).
A similar relationship was found between prior acquisition experience and acquisition performance (Haleblian & Finkelstein, 1999). However, the condition is that the experience must be somewhat similar to the new acquisition, as dissimilar acquisition experience was found to have an adverse effect, due to inappropriate application of learnings from these dissimilar experiences.
14
3. Hypotheses & Conceptual Model
Considering the preceding research on institutional distance, it can be concluded that institutions in general can have a massive effect on firm behaviour, such as entry mode as researched by Brouthers (2013). Moreover, another conclusion based on the literature, is that institutional distance between partners may also significantly affect business success. As pointed out by Meyer et al (2011), in a new environment, MNEs will have to adapt their entire organisation to suit the new institutional environment in the host country. This means that larger institutional differences require more significant adjustments and will result in more difficulties, which in turn decreases investor confidence. The papers by Brouthers (2013) and Meyer et al (2009) are in line with this finding. They found that when institutional distance is high, firms prefer to maximise control and attempt to avoid significant
cooperation. This may have a significant effect on alliance success as well.
Additionally, according to Rosen (2006), the investor response to a merger being announced depends on whether the investors believe it will result in new synergies that can enhance business success. As pointed out by Hitt (2009), firms typically engage in such mergers and acquisitions to capture new synergies, yet often with disappointing results. The reason for this is that to capture new synergies, integration is crucial. However, if firms differ
significantly from each other, successful integration may be too difficult or even impossible. Mudambi & Nevarra (2002) state that in such a situation, firms will have to adapt their entire system to a new institutional environment. Therefore, larger differences in firms will be found when institutional differences between the countries the firms reside in increase. This leads to larger integration problems and less possibilities to capitalise on potential synergies.
Lastly, the direction of the stock response can be predicted based on preceding research. Dyer et al (2001) found that on average, at every alliance announcement that is made by firms, the company’s stock price jumps by 1%, corresponding to 54 million US dollars per alliance that is announced. While alliances may not have the same effects as acquisitions do, the idea of being able to extract more value that may result in business success is similar. Therefore, it can logically be hypothesised that the stock market generally responds
15 Hypothesis 1: Smaller institutional differences between acquirers and targets will be
perceived as beneficial and will therefore lead to a larger increase in stock price in the period after the announcement.
As pointed out by Ang & Michaliova (2008), firms from less developed countries have more than just an advantage when doing business in their home country. Given that institutional arrangements and severe regulation is lacking in their home country, many emerging market firms have the distinct advantage of being able to thrive when institutions are lacking. As a result, these firms from countries of low institutional quality will have an advantage when working in another country without sophisticated institutions, compared to firms that are used to higher institutional quality.
It makes sense to assume the existence of an opposite effect too. Wu & Chen (2013) mentioned that emerging market firms may have a disadvantage in advanced economies, where institutional quality is higher. The reason is that poor institutions in the home country may limit the natural flow of the market, meaning firms from these countries may not be as well-prepared for competing in advanced economies compared to firms from advanced economies themselves. This leads to an advantage for firms from countries with high institutional quality when merging with other firms from a country with sophisticated institutions, while firms from emerging economies may have a distinct advantage when conducting business in a country of low institutional quality.
However, Gubbi et al (2009) specifically researched how Indian firms acquiring European firms impact the stock value of these Indian firms, in a quite similar research. Contrary to the hypothesis based on the work by Wu & Chen (2013), the researchers found positive
abnormal results after the announcement of such acquisitions. The reasoning provided here is that the perceptions of these acquisitions depend on the new resources and possibilities such a firm acquires in a new country.
This relates to the idea of springboard MNEs as illustrated by Luo and Tung (2017), who state that many firms from emerging economies benefit from expanding to advanced economies, as they are able to capture strategic advantages related to the different
institutional contexts and resources. Furthermore, such an idea would be in line with ideas by Demirbag & Yaprak (2015), who state that upmarket acquisitions are crucial for emerging multinationals. The reason for this is not only the quick access to new resources and
16 2013). Demirbag & Yaprak (2015) argued that some aspects of institutional distance, such as political, economy and knowledge, would give emerging market firms the opportunity to escape from home-country institutional constraints. The stock market can be expected to recognise this.
It is important to keep in mind that the study by Gubbi et al (2009) was specifically focused on Indian springboard firms. Therefore, it may not be representative for the entire cluster of emerging markets. However, the evidence so far suggests that for emerging market firms acquiring a firm from a more developed country, the relationship will be reversed. Therefore, acquisitions involving institutional distance may actually be considered promising rather than simply difficult.
This results in two contrasting hypotheses. H2a is specifically focused on situations where the acquirers are rooted in countries with highly developed institutional frameworks. On the other hand, H2b is specifically focused on acquirers rooted in countries where there is lower institutional quality.
Hypothesis 2a: Positive stock price reaction to an acquisition announcement will be larger when an acquiring firm from a high institutional quality country targets a firm from a different high institutional quality country, compared to when the acquiring firm targets a firm from a low institutional quality country.
Hypothesis 2b: Positive stock price reactions to an acquisition announcement will be smaller when an acquiring firm from a low institutional quality country targets a firm from a different low institutional quality country, compared to when the acquiring firm targets a firm from a high institutional quality country.
Together, these hypotheses provide an overview about the effect of institutional distance on stock price after an acquisition announcement. Therefore, the results of the tests of these hypotheses will allow me to answer the general research question.
17 Figure 3.1: The conceptual model of this research
In order to better understand this model, it is important to realise that stock price in general tends to increase after an announcement, regardless of institutional distance. This
expectation relates to the findings by Dyer et al (2001). This effect specifically deals with the coefficient of the stock price, in other words the percentage of change. Due to the prior literature, this paper shares the assumption that there is a general increase of stock price in this sample as well.
However, the hypothesis is that when there is higher institutional distance, this leads to a lower expected increase of stock price. Stock price is still expected to rise even when institutional distance is higher, but the coefficient is predicted to be smaller than with lower institutional distance. It is therefore proposed that a larger institutional distance results in a lower increase of stock price, while a small institutional distance may lead to the largest increase of stock price following an acquisition announcement.
18 The relationship suggested by H1 is moderated by hypotheses 2a or 2b.
Hypotheses 2a applies in the situation of advanced countries, while hypothesis 2b applies in developing and emerging countries. These assumptions relate to the institutional quality in the home country. It is important to remember that these hypotheses each affect the negative relationship between institutional distance and the stock price increase. Therefore,
considering 2a is positive, this implies that high institutional quality in the home country strengthens the negative relationship between institutional distance and stock price increase. In other words, if the institutional quality in specifically the home country is higher, this effect is stronger and stock price increase may be lower or even non-existent.
However, if the institutional quality in the home country is lower, institutional distance may be perceived as a positive thing, meaning the negative effect of institutional distance on stock price may be lower. As a result, the increase of stock price is logically larger in general.
In short, the moderator that relates to H2a and H2b can be described as the institutional quality in the home country of the acquiring firm. As contrary effects are expected, these hypotheses are listed separately in the model. Moreover, separating the two sub-hypotheses maximises the clarity of the conceptual model.
19
4. Methodology
Research Strategy
In order to be able to answer the research question, stock market data is combined with data related to institutional quality of the involved countries and data related to variables such as deal size. This is a quantitative study based on development of stock pricing. The reason for the focus on stock market analysis is that it is considered an objective predictor of firm success (Guo, 2002).
An event study was conducted, observing these changes in stock pricing ten days after the announcement of an acquisition. This part of the study resembles the methodology used by Kale et al (2002), who measured stock response to alliance announcements in order to find out to what extent stock value fluctuation impacts long-term acquisition success. As
established earlier in this paper, the stock market is a reasonable predictor of future success of a deal. Therefore, measuring stock response in situations of varying institutional
differences is an appropriate methodology to assess the effect of institutional distance in acquisition deals.
In order to collect the data, an extensive sampling procedure was used based on several groups. These groups are made up of certain institutional quality levels in home and host country. As a large variation of institutional quality is desirable for the research, the data needs to include a large number of countries as well. The procedure is explained in more detail in the sampling paragraph of the methodology section. Before that, ethical
considerations of this research are considered and more is explained about the variables selected for this research.
20
Ethical Considerations
Research conducted within Newcastle University needs to adhere to General Data Protection Regulation (GDPR). This research uses secondary data that was obtained from a publicly available database and does not contain any sensible or confidential data. For this reason, GDPR regulations might not always apply.
Also, this research does not involve primary data such as questionnaires and interviews. It does not include human participants or sensitive data either, given all the data was willingly published by firms over the 2008-2019 period. It does include the names of the firms (both the acquirer and target) as they appear in the Zephyr database. However, names are not included in this research to ensure the privacy of the firms is not harmed. Therefore,
participants included in this database are not believed to be subject to any harm and dignity is respected.
For all these reasons, consent related to the data can be assumed. More information can be found in the ethics approval form that was submitted for this research. This form can be found in appendix B.
Variables
Independent variables
Institutional quality
21 Another benefit that Kuncic’s work offers is that this research was published in 2014. This means that the research is still relatively recent. Older research may be outdated given that institutional frameworks change gradually over time (Mahoney & Thelen, 2010). Kuncic’s work is especially appropriate to use given the fact that the deals selected for this research took place between 2008 and 2019, with the median (250th and 251th) deal of the sample taking place in 2014. For this reason, using Kuncic’s work avoids certain issues that are related to outdated data. A more detailed description of the clusters used for institutional distance can be found in Appendix A. The table in Appendix A lists the 5 groups of institutional quality as described by Kuncic (2014).
Furthermore, institutional distance is checked using an alternative measurement. It is possible that clusters are too narrow as a measurement of institutional quality. Considering the needs of this research and its diverse sample, the data needs to be both recent and broad. These requirements are fulfilled by the Worldwide Governance Indicator project (World Bank, 2018), which matches the countries included in Kuncic (2014) optimally. Therefore, the data from this research are used as an alternative measurement of
institutional quality. Specifically, the 2014 data of the World Bank is used. Data from this year is complete and provides the best possible fit with the time period of the research. This averts the problem of potential institutional change (Mahoney & Thelen, 2010) between the moment of data collection and the time of acquisition announcement.
Using two different measurements enhances the accuracy of this research, as it increases the chance of more accurate institutional distance measurements. Moreover, an added benefit is that it is able to showcase how clustering institutional distance may lead to different results than accounting for institutional distance individually for every country.
Acquisition data
22
Dependent Variables
Stock Price Shift
Stock pricing appears to be one of the best available metrics to assess the success of deals, and it is one that has been repeatedly used. Measuring acquisition performance using different methods than stock pricing is challenging and complex (Gubbi et al, 2009). Moreover, stock pricing has successfully been used in similar M&A performance studies before (Moeller & Schlingemann, 2005).
In order to collect comprehensive and accurate historic stock data, the usage of stock price databases is required. This research uses the comprehensive online stock databases of investing.com and Yahoo Finance (Investing.com, 2019) (Yahoo Finance, 2019). These two online databases turned out to have the most complete and regular stock data of firms during the entire period of 2008-2019, which is a requirement for this research.
Two different methods for measuring stock price shifts were used, namely basic stock price shift and abnormal stock price shift.
1. Basic Stock Price Shift
The basic stock price shift is measuring the shift without controlling for market trends. It is measured by comparing the stock price before the acquisition announcement to the stock prices after the announcement. In order to measure this, the stock price 1 day before the announcement and 10 days after the announcement were selected. Such a 10-day event window has been used before by Boateng et al (2019). In the large majority of cases, the opening price of the stock market was used on both of the days. However, in some cases, opening price data was missing. In this case, the average trading price on the day was measured instead. This would be done for both measured dates, in order to remain
consistent in measuring the stock price shift. In order not to distort the data, the stock price 1 day before announcement is used, rather than the opening price at the date of
announcement.
23 2. Abnormal Stock Price Shift
As the preceding stock data studies of Kale et al (2002) and Rosen (2006) have
demonstrated, it is appropriate to control for the preceding stock price movement before the announcement date as well. Therefore, stock pricing is monitored 3 days before the
announcement until 10 days after the announcement. This should be enough capture the change in stock price trends from before and after the announcement, relative to the
possibilities of this research. This two-week period was selected as it fits this paper’s limited scope and resources. Measuring these trends results in a more appropriate measurement, given that it allows controlling for abnormal stock price movement. This is a method to control for the regular ‘expected’ movement within this period, which is the trend that was already present without any acquisition announcements.
To be more precise, the overall price trend of the stocks before announcement is observed and taken into account as well, to be able to focus at stock responses that actually deviate from the trend of the stock market. Therefore, what happens is an observation of the difference between the actual value post-announcement and the expected value. This expected value is obtained when extending the existing daily stock price movements 3 days before the announcement towards the trading date ten days post-announcement.
To obtain the expected value at t+10, the following formula is followed:
Then, to find the group average of abnormal stock price shift, this is the calculation that was used:
24
Control Variables
Deal Size
The first control variable that needs to be included is a size-related variable. Based on research by Mueller & Schlingemann (2004), size was found to have an impact on abnormal stock reactions after acquisition announcements. Multiple reasons are proposed for this, for instance it was found that size impacts the acquisition premium. Alternatively, smaller firms may have strictly positive synergies related to such alliances, while large firms may actually experience negative synergies related to their acquisitions announcements.
Moreover, Kusewitt (1985) found comparative size to matter for alliance performance as well, due to deal size mismatches having severe negative impact on alliance success. This means that that deals that are too small or large for certain firms, will likewise result in different abnormal returns on the stock market.
Controlling for size in this research is important, as size was already found to be a factor in similar research. Prior research has used deal size as a control variable (Chen et al, 2014). As this seems suitable, deal size is also included in this paper. If the size of the deal is larger, meaning the target firm involved in the deal is bigger in revenue and value, this will logically also have a larger impact on the stock price effect following the announcement. Deal size is reported in million euros. This is the most convenient option given that the Zephyr database reports deals in millions of euros as well.
Prior alliance and acquisition experience
Heimeriks & Duysters (2007) found that alliance performance can be enhanced through the development of alliance capabilities. This suggests that prior alliance experience is a factor, as these capabilities typically are developed through learning from prior partnerships. It can be proposed that similar learning effects will apply for acquisition deals. Generally it was found that acquisition experience should also help firms, on the conditions that these acquisition experiences are similar to the current deal (Haleblian & Finkelstein, 1999).
25 stake or its international outlook to maximise the effects of learning. Therefore, this research specifically includes all majority-stake cross-country deals that firms in this sample were involved in as acquirer, in the 21st century, preceding the year of acquisition in the sample. Zephyr-listed deals in this particular timeframe are used to measure this variable.
Considering the completeness of the Zephyr database, it is accurate estimation of recent acquisition experience that the firms involved in the deals have acquired.
On average, the firms in this sample had 12 cross-border acquisition experiences until the year preceding the deals in this sample. This includes 145 firms that did not have any prior experience at all. The sample includes both a lot of firms for which the deal is their first cross-border experience, but it also includes very experienced firms. Therefore, prior experience seems to be well-balanced across the sample.
Country of Origin
Given the fact that institutions are only one aspect of the entire environment of the country, it would be worthwhile to control for the firm’s country of origin. As suggested by Rossi & Volpin (2004), there are different country-level factors that may play a role. For instance, investor protection in a country was found to affect the mergers and acquisitions activities. Also, as mentioned by Xia (2011), there is often a certain dependence on other firms in the home market of the firm. This could lead to different effects of acquisitions based on the business environment in the home country and the function that the acquired firm has for the parent firm. Together, elements like these portray several effects that the country of origin could have on the relationship examined in this paper. Therefore it is valuable to control for the home country of the firm. In order to do so, each country involved in the sample was obtained a unique number. As a result, it could be used in the regression analysis, as a nominal variable.
26 Table 4.1: Variables data
Variable Type Source Obs Std. Dev Mean Min Max
Inst. Distance (clu) Ordinal Kuncic (2014) 500 1,20 1,6 0 4 Inst. Distance (sc)
Ratio World Bank (2018) 500 17,12 32,62 5,41 80,12 Stock Price Shifts Ratio Yahoo Finance (2019) 500 0,095592 1,0159 0,66233 1,78674 Abnormal Stock Price Shifts Ratio Yahoo Finance (2019) & Investing.com (2019) 500 0,229939 1,0296 0,24174 3,6
Deal size (in million euros)
Ratio Bureau van Dijk (2012)
500 2372,36 478,37 10 34164
Prior cross-border experience
Ratio Bureau van Dijk (2012)
500 43,2156 12,094 0 651
Country of Origin
Nominal Bureau van Dijk (2012)
500 / / / /
Table 4.2: Correlation Matrix
Inst. Dis (cluster) 1 Inst. Dis (score) .637** 1 Stock price shift .019 .014 1 Abn stock price shift .019 -.051 .288** 1 Deal size -.007 -.006 .014 -.079 1 Prior XP .010 -.012 .017 .022 .016 1 Country of origin -.096* -.181** .069 .049 .029 .035 1
I.D. (cl.) I. D. (sco.) SPS ASPS DV AXP COI
* = significant at 5%, ** = significant at 1%
27 As expected, there is a high positive correlation, as these are two different methods to
measure the same variable.
The second pairing where this applies is stock price shift and abnormal stock price shift. Similar to the institutional distance measurements, these two variables are expected to correlate, and therefore it matches the expectation that they do indeed somewhat correlate. These two variables are separated in the tests, meaning there are no issues related to multicollinearity in this sample.
No other major correlations are found. The only other significant relationship that was found is those between the institutional distance metrics and country of origin. This was largely expected, given that institutional distance is calculated on a country level and is therefore part of the unique characteristics of a country. Not a lot can be said about the direction of this correlation, given that country of origin is a nominal variable.
Sampling
Countries from all different levels of institutional quality are in the sample, in order to maximise the diversity of institutional quality across the sample. This is required to capture the effect of institutional distance across the entire spectrum of institutional quality.
After that, as many combinations as possible were be made, pairing up acquirers and targets from the different groups of institutional quality. This leads to 25 different groups. Mergers and acquisitions from the time period 2008-2019 are taken into account, which leads to a large availability of data. This time period allows sufficient data in order to complete this research. Moreover, the deals in this sample are recent, and as data primarily relates to the current decade, data is as complete and accurate as possible for this research.
The complete sample size of this research is 500 deals, which are divided over 25 groups of 20 deals. It is exactly 500 deals as symmetry is desirable for clustered analysis, and
28 different nationalities of target firms. In total, the sample is made up of 94 different firm
nationalities.
Table 4.3 shows the sample divided in the 25 groups that each include 20 deals. The group of target and group of acquirer values in the table again refer to Kuncic (2014), as group 1 refers to the group with the weakest institutional quality while group 5 refers to the group of strongest institutional quality. The table includes the average stock response and corrected abnormal stock response as they were found in the data. The average stock price shift in the sample of 500 was found to be +1,59%. When adjusted for the stock price trends, this figure was found to be an increase of 2,96%. The formulas used to calculate these shifts per groups are those explained in the variables section.
Table 4.3: Sample Overview
Group Group of target Group of
acquirer
Sample size Avg. stock response
29 This research uses stratified sampling. Based on the research strategy, this may be the most reasonable way to create a varied sample (Trost, 1986). Symmetry in this research is
required, yet the fixed groups have high variation in commonality, as advanced market firms in this population turned out to have higher internationalisation rates than emerging and developing market firms. If stratified sampling is used, every individual cluster has the same impact on institutional distance. This is desirable for this paper as it attempts to evenly take all of the different pairings of institutional quality in account.
High quality randomisation is conducted using randomisation software supplied by Random.org (Randomness and Integrity Services, 2019). Random.org is one of the most popular true random numbers generators available (Antu Annam & Varghese, 2018). True random number generators generate numbers based on statistically random ‘’noise’’ signals, rather than through a (semi-random) algorithm. Instead of a predictable series of numbers, true random number software generates a unique digital bit stream (Pain, 2019). Considering the scope of this research, random.org is the most accessible option to obtain true random numbers. The randomisation results in a sample that is the most representative of the population of worldwide acquisition deals. The downside of this sampling method is that it may result in an uneven balance of countries involved in deals.
As a result, a country like China may seem somewhat over-represented in the sample, as it is being represented by 88 acquirers. The reason for this is that China is of lower institutional quality, while it has high international activity. This overrepresentation relates to the clustered approach of Kuncic (2014). Within the weakest group of institutional distance, China is by far the country with the largest international presence in the business world. Subsequently, acquirers from the weakest groups are primarily Chinese firms. While the US is arguably as important in the business world, there are more countries in the group of the United States that are also internationally present, meaning the US is not as overrepresented. However, each level of institutional distance is equally present in the sample and a control variable for country factors is in place, meaning it is not considered problematic for this research. In order to ensure sufficient effect of alliances on the course of business, deals need to fit certain requirements. First of all, this research makes use of a minimum deal size. Strictly larger deals are taken into account, and additionally deals with unknown deal values will be left out. The cut-off value for deal size was decided to be 10 million euros. This ensures that the sample size of this research will be large enough to appropriately represent the
30 acquisitions in this research should be expected to impact the stock market in general, potentially based on the institutional distance involved in the deal.
Requirement 1: Deal size > 10 million euros.
Another crucial aspect of this research relates to the fact that the research strictly takes majority acquisitions into account. Minority acquisitions may not attract the same attention as a majority deal and therefore are not expected to have a comparable impact on stock price. Controlling for acquisition stake is therefore an important method to ensure that true effects of acquisitions are taken into account.
Requirement 2: Acquisition stake > 50% of the firm.
Lastly, given that changes in the stock market are measured in this research, it is important that the acquiring firm is listed on the stock market. As this is tracked by the Zephyr database (Bureau van Dijk, 2012), this is relatively easy to check for. As this research focusses on the effect on acquiring firms and not the target firm, it is acceptable for the target not to be listed on a stock market during the deal.
Requirement 3: The acquiring firm needs to be listed on a stock market before, during and after the announcement of the deal.
The final sample based on these requirements is listed in Table 4.3 earlier in this section. As the sampling and the nature of the data is now cleared up, the next section will zoom in on the analysis that will be conducted based on this data.
Analysis
After the data collection process, different types of analysis are used to analyse the
relationship between institutional distance and stock price effect based on the collected data. The analysis contains different types of regression analyses and ANOVA tests. These are explained in detail later in this section.
31 filters out potential country factors related to the shift in stock prices. Using the method of abnormal returns, general stock market movements are also controlled for using the dependent variable. This is be done by comparing the expected value based on the pre-announcement stock price movement of the firm with the actual stock price 10 days after the announcement. Using this method, the general trend of the stock market regardless of an acquisition announcement is also controlled for in this research.
Hypothesis 1 is firstly tested using six different models of linear regression. Stock market response and abnormal stock market response are tested as dependent variable in a linear regression. The first three models are related to regular response, while models IV to VI relate to abnormal response. Model I and IV test the relationship between the control variables and the dependent variable. Models II and V solely include institutional distance into account as the independent variable related to stock response. Lastly, models III and VI include both the control variables and institutional distance as dependent variables. This means that for these models, deal size, prior acquisition experience of firms and country or origin are taken into account as well.
Additionally, the means of the different groups based on Kuncic (2014) are compared using ANOVA testing. This is done in order to determine if there are statistically significant
differences between the predetermined groups. The Tukey HSD is used for this reason, given the fact that the aim of the test is to find out if the means of 4 or more groups (in this case 25) statistically deviate from each other.
32 Lastly, hypotheses 2a and 2b are tested using a moderator analysis. This is done using the Process macro for SPSS, which was written by Andrew F. Hayes (2019). It is widely used in business science to assess several types of interaction in models involving moderators and mediators. The tool integrates many functions of established tools for mediation and
moderation (Hayes, 2012). This SPSS add-on is therefore used for the moderator analysis. Similar to hypothesis I, institutional quality metrics by Kuncic (2014) and The World Bank (2018) are taken into account. The home country cluster is used for Kuncic, while the home country institutional quality index is used for the test involving The World Bank (2018) data.
Slightly different parts of the data are used compared to H1, as the institutional quality measurements in the home country are taken into account as a moderator, rather than calculations of the institutional distance. The control variables that were used in H1 are likewise used for these tests, as covariates in the different models. The institutional quality in the home country is used as a moderator for the relationship between institutional distances and (abnormal) stock price shifts. These tests are conducted for both of the aforementioned measurements of institutional distance.
The hypothesis testing includes eight different models. Models I and II measure the effect of the institutional quality cluster variable on regular stock shift, while models III and IV measure the effect of this variable on abnormal stock shift. This pattern is repeated for institutional quality score. Model V and VI measure this variable’s effect on regular stock shift, while models VII and VII measure the effect on abnormal stock shift. Two models are used for these four different groups to compare the effects both without the control variables and including the control variables. Together, these eight models capture a potential moderating effect of institutional quality in the home country on the stock price shift after the
33
5. Results
In order to test the hypotheses, the procedure as explained in the methodology section was taken out.
First of all, it was already proven that stock prices typically increase after an acquisition announcement. The data in this research is in line with that idea as well. A regular increase of 1,59% was found, and when controlling for the stock trends a 2,96% abnormal shift was found. These numbers are obtained using the formulas of basic stock price shift and abnormal stock price shift, as explained in the variable section of this paper. The positive shifts that were found in the sample are in line with the expectations based on prior literature. The next step is to assess whether this shift in stock price is actually affected by institutional distance, along with the control variables selected for this research.
Hypothesis 1
The first step towards testing hypothesis 1 is six models of linear regression. Model I checks the relationship of the control variables, model II checks the dependent variable while model III checks all the relationships. The results can be found in table 5.1.
Table 5.1: Regression analysis examining the relationship between regular stock market response and institutional distance (cluster).
Dependent variable: Stock Market response
Model I Model II Model III
34 These three models did not show any relationships that are statistically significant. Rather than that, it suggests that the control variables somewhat explain the stock market response, while institutional distance struggles to do so. Model II 0,019, p = 0,679) and model III (-0,012, p = 0,788) suggest that there may be a slight negative relationship between the regular stock market response and the institutional distance based on the cluster. However, the relationship is clearly not statistically significant.
Together, these three models do not seem to adequately explain the shift in stock prices after an acquisition announcement, given the low R square scores of 0,000 and 0,005. This suggests that either there are other variables that may capture the stock shift, or that regular stock price shift is simply difficult to capture in general.
As mentioned previously, the following step is to re-run these models, based on the other dependant variable. This variable is abnormal stock market response, which is controlled for stock price trends prior to announcement.
Table 5.2: Regression analysis examining the relationship between abnormal stock market response and institutional distance.
Dependent variable: Abnormal stock market response
Model IV Model V Model VI
Institutional Distance 0,019 0,024 Deal size -0,081 -0,081 Prior XP 0,022 0,021 Country of Origin 0,051 0,053 R Square 0,009 0,000 0,010 N 500 500 500 * = significant at 5%, ** = significant at 1%
Using abnormal stock market response did not drastically change the results. However, the models appear to have more explanatory power than the models in Table 5.1.
Model V resulted in a beta of 0,019 and a p-value of 0,665. Model VI (0,024, p-value = 0,597) did not result in a significant relationship between abnormal stock market response and institutional distance.
35 None of the variables were found to be significant at a 5% confidence interval, but along with a larger R square, the coefficients are found to be higher. In this sample, deal size has the most notable impact on the stock response, at a p-value of 7,2%. While this effect is not significant enough in order to consider a confirmed relationship at a 5% confidence interval, it does lead to the suspicion that there might be a relationship here.
As explained, an ANOVA test is now performed to assess group differences. Findings are reported in table 5.3.
Table 5.3: ANOVA test examining group differences regarding abnormal stock market response among the 25 groups based on Kuncic (2014).
Stock Price shift Abnormal Stock Price shift
Lowest Group 0,9885 0,9715
Highest Group 1,0617 1,1610
P-value 0,764 0,616
These two tests result in p-values of 0,764 and 0,616. This shows that the result of this test is not significant. This signals insufficient differences between the 25 groups. Moreover, two more ANOVA tests were conducted specifically based on the 5 groups of institutional quality of the home countries.
Table 5.4: ANOVA test examining group differences regarding abnormal stock market response among the 5 home country groups based on Kuncic.
Stock Price shift Abnormal Stock Price shift Lowest Acquirer Group 1,0114 1,0056 Highest Acquirer Group 1,0281 1,0518 P-value 0,735 0,616
36 Table 5.5: Regression analysis examining the relationship between regular stock market response and institutional distance (score).
Dependent variable: Stock Market response
Model VI Model VII
Institutional Distance -0,014 -0,002 Deal size 0,012 Prior XP 0,014 Country of Origin 0,068 R Square 0,000 0,005 N 500 500 * = significant at 5%, ** = significant at 1%
Results here seem to be in line with results from Table 5.1. Effects of institutional distance appear to be very minor, at a beta of -0,014 and -0,002, and p-values of 0,746 and 0,966. It is notable to mention that these models once again found negative beta scores.
Table 5.6: Regression analysis examining the relationship between abnormal stock market response and institutional distance (score).
Dependent variable: Stock Market response
Model IX Model X Institutional Distance -0,051 -0,044 Deal size 0,081 Prior XP 0,021 Country of Origin 0,043 R Square 0,003 0,011 N 500 500 * = significant at 5%, ** = significant at 1%
37 somewhat relate to the abnormal stock market response as well. Country of origin positively relates at a level of 0,349, while deal size positively relates at a probability of 0,072.
In sum, there is insufficient evidence to assume that institutional distance has a significant impact on the stock market response after an acquisition announcement.
However, in the vast majority of the models in place, a negative relationship between institutional distance and the stock market response is suspected. This would be in line with hypothesis 1, which yet cannot be confirmed.
Hypothesis 2a and 2b
Hypotheses 2a and 2b are tested using moderator analysis. As explained, the moderator analysis is conducted using the external Process SPSS add-on (Hayes, 2019). Using the home country cluster by Kuncic, four models were tested, of which results are displayed in table 5.7.
Table 5.7: Moderator analysis examining the moderator effect of acquirer institutional quality (cluster) on the relationship between institutional distance and stock price effect
Model I (regular stock shift) Model II (regular stock shift) Model III (abnormal stock price shift) Model IV (abnormal stock price shift) Inst. Distance (cluster) 0,002 0,004 -0,007 -0,005 AcquirerIQ 0,003 0,001 -0,005 -0,009 Inst.Dis x AcquirerIQ -0,001 -0,001 0,004 0,003 Acquirer XP 0,000 0,000 Deal value 0,000 0,000 Country of origin 0,003 0,001 * = significant at 5%, ** = significant at 1%
38 Table 5.8: Moderator analysis examining the moderator effect of acquirer institutional quality (cluster) on the relationship between institutional distance and stock price effect
Model V (regular stock shift) Model VI (regular stock shift) Model VII (abnormal stock price shift) Model VIII (abnormal stock price shift) Inst. Distance (score) -0,000 -0,000 -0,002 -0,002 IQhomecntry 0,000 0,000 -0,001 -0,001 Inst.Dis x IQhomecntry 0,000 0,000 0,000 0,000 Acquirer XP 0,000 0,000 Deal value 0,000 0,000 Country of origin 0,000 0,000 * = significant at 5%, ** = significant at 1%
The outcomes of the models in table 5.8 are also insignificant. The p-values of these models varied between 0,2950 and 0,8011.
Therefore, insufficient evidence was found for hypotheses 2a and 2b. This means that hypotheses 2a and 2b can be assumed to be false. This could be expected given that no significant evidence was found for hypothesis 1 either, meaning that significant moderation of that relationship is also improbable. A notable finding is that for both models I until IV and models V-VIII, the abnormal stock price shift seems to pick up effects better than the regular stock shift can do. It is possible that abnormal stock price is a more accurate reflection of these relationships.
In sum, neither hypothesis 1 nor hypotheses 2a and 2b were found to be significant
39
6. Discussion & Conclusion
These results allow for answering the research question of this paper, which is:
How does institutional distance between both parties of an acquisition affect stock price reaction to the announcement?
Evidence from hypothesis 1 is inconclusive and insufficient regarding the impact of
institutional distance on stock price reaction. In 6 out of 8 models that involved institutional distance, the beta of institutional distance was negative. While the results are insignificant, it does suggest the direction of institutional distance’s impact on the stock market. Considering the relatively consistent negative betas, this suggests that an increase of institutional
distance will potentially have a negative impact on the stock price shift, rather than a positive impact on the stock shift. This theory is still unproven by this research, but the direction of the findings is in line with hypothesis 1. It is in line with literature such as Meyer et al (2011), who also suggested the relationship would be negative.
Furthermore, it is notable that in this model, prior acquisition experience, country factors and firm size were also not found to be significant predictors of the shift in the stock market. A potential reason for that is that the stock market is unpredictable. Due to intrinsic volatility, the task of predicting stock market shifts is challenging. (Basak et al, 2019). However, some control variables, particularly deal size (p= 0,072), have a small p-value in the results of this paper. It is therefore a possibility that these variables have an effect on the stock market. This may be in different or maybe very specific circumstances other than those in this research.
Another interesting finding in this research relates to the conceptualisation of institutional distance. Kuncic (2014) and the World Bank (2018) approached differences in institutional quality in different ways, and this could be seen in the test results. While these two methods correlated heavily (0,637), the institutional distance proxy of the World Bank found stronger relationships. The p-value and beta values of institutional distance in the models using the institutional distance score were generally more convincing than those using the cluster approach. This suggests that the different methodology in place for both methods could be significantly different, and may potentially underline the limitations of a clustering approach.
