Multinationality and systematic risk: a literature review and meta-analysis

Höge-Junge, Christin; Eckert, Stefan Article — Published Version Multinationality and systematic risk: a literature review and meta-analysis Management Review Quarterly Provided in Cooperation with: Springer Nature Suggested Citation: Höge-Junge, Christin; Eckert, Stefan (2022) : Multinationality and systematic risk: a literature review and meta-analysis, Management Review Quarterly, ISSN 2198-1639, Springer International Publishing, Cham, Vol. 74, Iss. 1, pp. 377-414, https://doi.org/10.1007/s11301-022-00304-6 This Version is available at: https://hdl.handle.net/10419/309262 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) Management Review Quarterly (2024) 74:377–414 https://doi.org/10.1007/s11301-022-00304-6 1 3 Multinationality andsystematic risk: aliterature review andmeta‑analysis ChristinHöge‑Junge1· StefanEckert1 Received: 18 January 2022 / Accepted: 24 October 2022 / Published online: 18 November 2022 © The Author(s) 2022 Abstract In the literature, the impact of multinationality on the valuation of multinational companies is heavily debated. To understand this impact on valuation, we need to clarify whether and how multinationality affects systematic risk. For this purpose, we analyze the state of research concerning the impact of corporate multinationality on systematic risk, conducting a systematic literature review of 35 studies and a univariate meta-analysis based on 20 studies. We test the predictions of the upstream– downstream hypothesis and the increasing capital market integration hypothesis on the basis of a meta-regression analysis of 17 studies. Our results provide no empirical support for the upstream–downstream hypothesis. However, they corroborate the capital market integration hypothesis in a more radical manner than expected: whereas multinationality seemed to have a risk-reducing effect until the beginning of the 1990s, since then its impact appears to have shifted. We find a risk-increasing effect for multinationality from 1990 on. Our results have important implications for academic research and managerial practice. Keywords Multinationality· Systematic risk· Capital market integration· Upstream–downstream-hypothesis· Univariate meta-analysis· Meta-regression analysis JEL Classification F23· F30· G11· G31· G32· L25· M16 * Christin Höge-Junge [email protected] Stefan Eckert stefan.ecker[email protected] 1 International Institute Zittau, Chair ofInternational Management, TU Dresden, Markt 23, 02763Zittau, Germany 378 C.Höge-Junge, S.Eckert 1 3 1 Introduction The impact of corporate multinationality on systematic risk is an essential issue for academic research as well as for business practice (Reeb etal. 1998; Song etal. 2017). Systematic risk, also referred to as market risk or beta, is the component of a company’s total risk that is inherent in a market and affects all companies in that market to a specific degree (Shapiro and Balbirer 2000). A geographical expansion of a company beyond the borders of the home country may lead to a change in the level of the company’s systematic risk. Surprisingly, the impact of multinationality on systematic risk has been analyzed relatively rarely in the academic literature up to now (Kwok and Reeb 2000). However, it is a topic of utmost importance. We have to clarify whether and how multinationality affects systematic risk in order to really be able to understand the valuation impact for multinational companies, a heatedly debated issue in the literature (Contractor 2007; Eckert etal. 2016; Hennart 2007). Since finance scholars generally use the stock market value of the firm as a measure of firm performance, the issue also has implications for answering the question of the relationship between multinationality and performance (Li 2007). Furthermore, the question of the relationship between multinationality and systematic risk bears profound practical implications: modern methods to estimate the market value of a firm build on valuation models, which employ the systematic risk of a company in order to generate adequate discount rates for its estimated future cash flows. If a company is multinational, we need to know how multinationality affects systematic risk to calculate the correct weighted average cost of capital and formulate a profound investment decision. The objective of this paper addresses this research gap. We contribute to the debate on the impact of multinationality on systematic risk. We review extant research by conducting a systematic literature review as well as a univariate meta-analysis and a meta-regression analysis. We focus on two theoretical arguments to explain differences in the effect of multinationality on systematic risk. The first argument is the so-called upstream–downstream hypothesis: Kwok and Reeb (2000) argue that firms from stable economies investing in unstable countries might experience increases in systematic risk, while firms from unstable home countries investing in more stable target countries might experience decreases in systematic risk. We refer to the second argument as capital market integration hypothesis: Following this line of reasoning, the effects of multinationality on systematic risk may have diminished over time as a consequence of an increase in global integration of national capital markets. Our findings indicate that international diversification arising from multinationality contributed to a reduction of systematic risk in the past; however, it has resulted in an increase in systematic risk since the 1990s. Increasing market integration could be an explanation for the shift regarding the impact of multinationality on systematic risk. One important implication of our findings is that future research has to take consideration of an international market portfolio as the reference market for estimating systematic risk. Furthermore, corporate executives should carefully rethink their internationalization steps if they want to increase the firm’s capital market value. Considering the recommendations of Fisch and Block (2018), Kuckertz and Block (2021), Hansen et al. (2022), this paper is divided into six sections. Theoretical arguments why and how multinationality might affect a firm’s systematic risk are presented 379 1 3 Multinationality andsystematic risk: aliterature review… in the second section. In the third section, we explain the selection of academic papers included in our analysis and our research design. In the fourth section, the results of our systematic literature review are provided. Building on the findings of our literature review, we develop hypotheses and report the results of our univariate meta-analysis and metaregression analysis. In the sixth section, we discuss our findings and show implications. 2 Theoretical background According to modern portfolio theory, the risk of a security can be divided into two mutually exclusive components: systematic risk and unsystematic risk. Following the Capital Asset Pricing Model (CAPM) only systematic risk has to be compensated whereas unsystematic risk can be eliminated through diversification. Empirical research has shown that international diversification contributes to this elimination of unsystematic risk (Amit and Livnat 1988; Michel and Shaked 1986). International business (IB) scholars have used portfolio theory to examine the effect of a firm’s internationalization and thus firm’s multinationality on its risk (Song et al. 2017). Multinational companies can diversify the portfolio of their activities by means of international geographical dispersion and reduce their systematic risk due to the reduced volatility of earnings (Bution etal. 2015; Olibe etal. 2008; Song etal. 2017). However, the portfolio concept was originally developed in finance theory under the assumption of perfect markets (Song etal. 2017). IB scholars like Rugman (1976), Reeb etal. (1998) argue that expanding internationally, whether by establishing new subsidiaries or acquiring foreign firms, can be compared to the construction of a stock portfolio. Although this argumentation has raised severe criticism and, in the meantime, there has been a proliferation of theoretical explorations to explain the performance impact of multinationality, portfolio theory and the CAPM still provide the theoretical foundation to consider the impact of multinationality on firm risk. From an investor’s point of view systematic risk is defined as: with: ρjm: correlation coefficient between security j and the market m; σj: standard deviation of security j; σm: standard deviation of the market m. In order to understand the impact of multinationality on systematic risk, we must assess how internationalization and thus international diversification affects the individual components of beta. Assuming that the effect of internationalization on σm is marginal, we can concentrate on the numerator. Conventional argumentation in the literature is that ρjm decreases through international diversification due to the low correlation of foreign markets with the home market of multinational corporations (MNCs) (Aggarwal 1977). However, this effect presupposes that the reference market m is the home market of the MNC. As our literature review will show, the overwhelming majority of extant research contributions follows this presumption. 𝛽 i= 𝜌 jm ×𝜎j 𝜎 m 380 C.Höge-Junge, S.Eckert 1 3 In addition to the effect on ρjm, multinationality has an impact on σj, the standard deviation of firm j. This component can on the one hand be expected to increase through internationalization due to exchange rate risk (Bartov etal. 1996; Reeb etal. 1998) and political risk (Choi and Severn 1992; Reeb etal. 1998)).1 In this regard the typical argument in the literature is that multinationality leads to an increase in a firm’s foreign exchange exposure which raises the variations of foreign returns in domestic currency (Reeb etal. 1998). Furthermore, it is argued that international firms experience an increase in exposure to political risk due to the dangers of the appropriation of foreign assets by host country governments, unanticipated regulatory changes induced by host countries, corruption, and so on (Reeb etal. 1998). However, this argumentation implicitly considers the case of a multinational corporation based in a developed home country entering host markets in developing countries. In contrast, emerging market MNCs may not suffer from significant increases in exposure to political risk when entering developed country markets (Kwok and Reeb 2000). On the other hand, if country markets are not perfectly correlated, multinationality may also lead to a compensation of country-specific variations in cash flows and therefore contribute to a reduction in cash flow variance. 3 Methodology Our first step in identifying the overall effect of multinationality on beta is to systematically review existing empirical research on the impact of multinationality on the market-based risk measurement of beta available as of August 2021. Papers relevant to our interest are detected by researching “Business Source Premier” of EBSCO, the databases of Elsevier and Springer, and the web search engine Google Scholar. We browse for English articles including specific terms in title or abstract,2 which are relevant to our research question. These terms are depicted in Fig.1. We understand “multinationality” as a firm’s geographic diversification of operations. Papers about the diversification of other firm related aspects like the international diversification of ownership remain unconsidered.3 Academic contributions focusing on firms from the financial services industry are also neglected due to industryspecific aspects. In addition to the database query, leading journals are searched by scanning the titles and abstracts of published papers. Since the topic is related to the field of international business as well as financial research, we consider journals from both fields. The journals are chosen using the SCImago Journal Ranking 2020 for business, management, and accounting journals as well as economics, econometrics, 1 Reeb, Kwok, and Baek (1998) furthermore argue for an increase in risk due to a weaker ability to monitor and discipline firm management. 2 Search options in the database of Springer and Google Scholar are limited to searches of publication titles. 3 In this context, emphasis should be placed on Chira and Marciniak (2014), which neglect a firm’s geographic diversification but discuss possible differences in systematic risk between domestic and crosslisted firms. In light of an effect on systematic risk, cross-listing and its potential benefits should be borne in mind for future studies. 381 1 3 Multinationality andsystematic risk: aliterature review… and finance journals. We confine our retrieval to the journals of the first 35 and 25 ranks, respectively. As the journals of Elsevier and Springer are already included in the database query, we concentrate on the following journals (journals preventing a search limited on title or abstract are excluded): • Academy of Management Journal • Strategic Management Journal • Journal of International Business Studies • Journal of Management Studies • Academy of Management Perspectives • Strategic Organization • Journal of Finance • Review of Financial Studies • Journal of Management • Review of Finance • Review of Corporate Finance Studies To complement our research, we scan the citations and the list of references of the papers already identified. Finally, we send a request to the AIB list servers and scan titles and abstracts of working papers available on the Social Science Research Fig. 1 Search terms and their combination. Term 1 and Term 2 are combined in an And-Boolean search 382 C.Höge-Junge, S.Eckert 1 3 Network (SSRN) server to uncover unpublished research in order to address the problem of publication bias.4 The database queries result in 2282 papers and 1911 papers after removing duplicates. These studies are subsequently screened by title and abstract and 1888 studies are excluded due to deviating research questions and inappropriate measures of risk (see Fig.2). A close examination of the reference list of the remaining papers identifies another 12 studies.5 Finally, 35 studies are included in the analysis (see Table1) notwithstanding a broad range of search terms. These papers are subsequently classified according to different aspects relevant for our analysis, e.g., the statistical methods applied, measurements of multinationality, and sample characteristics, and qualitatively reviewed in respect of their empirical results. The literature review is quantitatively summarized. We use vote counting in order to identify possible structural similarities between studies disclosing similar findings in order to give indications for subgroups and moderating variables. We do not intend to make conclusive statements about the true effect of multinationality on this basis. In addition to the qualitative review, we summarize the empirical findings with the help of a univariate meta-analysis and meta-regression analysis using the software package Comprehensive Meta-Analysis (CMA).6 We consider the current discussion on the possible impact of multinationality and base our summary on hypotheses related to firms’ home country and time. Univariate meta-analysis is a suitable method to summarize, integrate, and interpret the results of various studies, but it only applies to empirical research studies with quantitative findings that can be meaningfully compared in the form of an effect size (Lipsey and Wilson 2000). In view of this, the statistical analysis is restricted to certain papers so that the data basis of meta-analysis decreases further compared to the qualitative review. Due to the differences regarding the statistical methods applied and the statistical data reported, we divide the remaining papers into groups as the type of effect size must be the same to allow a meaningful analysis (Lipsey and Wilson 2000). We translate the specific research method employed as a study descriptor into codes and conduct two separate effect size statistics for analyzing sample comparison findings and regression results.7 In the light of meta-regression analysis we further code the home countries of the firms in the respective samples, distinguishing between nonUS and US data as well as developing and emerging country data. In addition, we classify the samples according to the time period of analysis between 1960–1989 and 1990–present. 7 The results of further statistical methods, e.g., (M)ANCOVA and semiparametric regression, are not examined. 4 See Fisch and Block (2018) as well as Kuckertz and Block (2021) on the need for a detailed description of the review process with the goal of transparency and repeatability of the literature review. 5 Haegele (1974) and Thomsen (2012) may also deal with our research question. Their results, however, cannot be obtained. 6 Specifically, CMA v.3 is used. 383 1 3 Multinationality andsystematic risk: aliterature review… Fig. 2 PRISMA flowchart of the systematic literature review 384 C.Höge-Junge, S.Eckert 1 3 Table 1 Overview of publications Publication Origin of companies Time Horizon Statistical method Effect size ProxybRelation-shipcNumber of observations 1 Aggarwal (1979)USA 1974 Simple linear regression COR FIR −** 171 FSR −* 149 FAR −*** 165 2 Agmon and Lessard (1977)USA 1959–1972 Simple linear regression COR FSR −** 217 3 Bany-Ariffin etal. (2016) Malaysia 2002–2009 Semiparametric regression COR Other −**5581 4 Barone (1983)USA 1974 Simple linear regression COR FIR − 162 1975 FIR − 168 1976 FIR + 154 1977 FIR + 166 1978 FIR + 166 1979 FIR + 164 1974 FSR − 141 1975 FSR − 164 1976 FSR + 165 1977 FSR + * 179 1978 FSR + ** 174 1979 FSR − 170 1974 FAR −*** 156 1975 FAR −** 174 1976 FAR − 171 1977 FAR − 181 1978 FAR + 175 1979 FAR − 170 391 1 3 Multinationality andsystematic risk: aliterature review… 1981; Broaden and Samii 2001; Fatemi 1984; Madura and Rose 1989; Michel and Shaked 1986) and the S&P index (Collins 1990; Kohers 1976; Siegel etal. 1995; Theerathorn etal. 1992). Krapl (2015), Shaked (1985) build a value-weighted U.S. market portfolio including all firms contained in their samples. In addition to the data source for the market index, the exclusive use of the domestic market as the reference portfolio has to be considered as the standard method throughout the papers of our study. Only Hughes etal. (1975), Jacquillat and Solnik (1978), Kwok and Reeb (2000) use an international market index as benchmark. Agmon and Lessard (1977) Joliet and Hübner (2008), Shyu and Ou (2009), Thompson (1985) as well as Chambliss etal. (1994) combine an analysis of systematic risk based on an international market index to the one based on the domestic market index through the application of a two-index model and multi-factor model, respectively. Bühner (1987) also considers a national and international market index within a single index model, however, concentrates on the domestic market index as the results of the regression are no more statistically significant when the international market index is used.9 With regard to the calculation of the systematic risk, the Carhart four-factor model (Jung 2015; Jung etal. 2018a, b; Song etal. (2017) and two-index model (Amon and Lessard 1977; Joliet and Hübner 2008; Thompson 1985; Shyu and Ou 2009) are used alongside the single-index model. In contrast, Aggarwal (1979), Barone (1983), Broaden and Samii (2001), Madura and Rose (1989) refer to published betas. Moreover, various approximations are used to determine multinationality. Nevertheless, the ratio of foreign to total sales is still the most common measure Fig. 5 Scope of countries surveyed (grouped) 9 The choice of the market index could not be identified in case of Jung (2015), Jung, Dalbor and Lee (2018a, b), Jung, Kim, Kang and Kim (2018), Severn (1974) and Song, Park and Lee (2017) in case of US-data studies as well as Bany-Ariffin, Matemilola, Wahid and Abdullah (2016), Shafigullin (2016) and Marciano and Herlambang (2016) in case of non-US studies. 392 C.Höge-Junge, S.Eckert 1 3 followed by the ratio of foreign to total assets. Referring to Goldberg and Heflin (1995), the ratio of foreign sales to total sales would generally best reflect the portion and significance of business transactions conducted in foreign countries versus total world transactions. It is the basic metric measuring the degree of foreign involvement (Nguyen 2017). This ratio is a consistent, relatively widespread, and accepted measure of foreign activities. It is publicly available in many cases and relatively unbiased by national accounting regulations (Fatemi 1984; Goldberg and Heflin 1995; Kwok and Reeb 2000). Employing the most widely used proxy also guarantees consistency with previous research (Olibe etal. 2008). Notwithstanding the numerous advantages of the foreign sales ratio, the majority of authors reject its sole usage and rely additionally on foreign assets, as foreign sales include export sales as well as foreign subsidiary sales (Kwok and Reeb 2000; Nguyen 2017; Olibe etal. 2008; Reeb etal. 1998). Further proxies often complement the measurement of international involvement. Using different measures of international involvement is preferable as each measure captures a different aspect of the complex structure of multinationality. Therefore, observing differences in effects across the measures provides valuable insight into the relationship between international involvement and risk (Krapl 2015). However, the multifaceted nature of multinationality also complicates the assessment of an overall impact of international diversification on systematic risk. In terms of the time horizon examined, the US data studies look back a time horizon of 55years covering data from 1959 to 2013. While Agmon and Lessard (1977) and Severn (1974) rely on the earliest US data covering a time horizon from 1959 to 1972 and 1966, respectively, Jung (2015), Jung etal. (2018a, b), Song etal. (2017) are the most recent studies covering a common period with the years 2000 to 2013. Within these 55years, several years were analyzed by different papers simultaneously, leading to a concentration on the years 1963 to 1987. Compared with US studies, the time horizon of non-US studies from 1966 to 2015 has a similar length, though it is considerably less analyzed. Only Bühner (1987), Thompson (1985) take up the issue for non-US corporations in the early years. A greater interest has arisen in recent years. Kwok and Reeb (2000) have initiated the research focus on nonUS firms with their upstream–downstream hypothesis analyzing non-US firms from 1994 to 1996. Referring to the specific length of the period of analysis, the papers can also be assigned to different categories. Severn (1974), Song etal. (2017) select the length of their time horizon to encompass all possible economic conditions. A similar long period of seven years is also used by Bany-Ariffin etal. (2016) to generate a stable result.Agmon and Lessard (1977), Krapl (2015), in contrast, gather data for very long time periods. Since long data series raise the question of beta stability, Michel and Shaked (1986), Reeb etal. (1998), Siegel etal. (1995) take temporal stability into account by portioning the time horizon into two or more subperiods in order to test the stability of the results over time. Broaden and Samii (2001), Hughes etal. (1975), Olibe etal. (2008), Shaked (1985) as well as Bution etal. (2015), Kwok and Reeb (2000) limit the time horizon for their survey to three to five years. Harjito etal. (2018), Joliet and Hübner (2008), Marciano and Herlambang (2016), Shyu and 393 1 3 Multinationality andsystematic risk: aliterature review… Ou (2009), Thompson (1985) use a similar short time horizon. Aggarwal (1979) and Barone (1983) even restrict their sample to a one-year period. Another problem regarding measurement of multinationality concerns those studies which rely on sample comparisons. Some studies use the same measures of multinationality, but different thresholds for classification. Siegel etal. (1995), Theerathorn etal. (1992), for example, rely on the foreign tax ratio. However, since Siegel etal. (1995) apply lower thresholds for the three classification groups, results are hard to compare. Similar problems also apply to all other sample comparison papers except for the comparison between Michel and Shaked (1986) and Shaked (1985). Fatemi’s (1984) classification scheme additionally leads to a discontinuity in the data spectrum, as firms with more than 25% foreign sales are referred to as multinationals and firms with no foreign operations are referred to as uninationals. Firms with foreign sales between 0 and 25% are not included at all (Siegel etal. 1995). In contrast, for the studies using multiple regression analysis, the choice of control variables leads to differences. As pointed out by Olibe etal. (2008), several factors, e.g., size, financial leverage, international diversification, growth, and currency fluctuations, have been shown to affect a firm’s systematic risk from an accounting and financial perspective. Therefore, multiple control variables should also be considered to clarify the relationship between multinationality and systematic risk. This position is also reflected in the increasing awareness of the importance to control further variables influencing the systematic risk within the papers over time. While the early regression analysis studies neglect any form of control variables, more and more control variables have been included in the studies since the mid-1980s. Thompson (1985) is the first paper that examines the effect of control variables. Subsequent studies also include additional explanatory variables in their regression analysis. However, the control variables examined differ. While Bühner (1987) examines the influence of firm ownership and growth, Madura and Rose (1989) and Theerathorn etal. (1992) consider the influence of R&D and advertising expenses. The frequent consideration of certain control variables, in detail firm size and leverage, has only been apparent since Goldberg and Heflin (1995) and later studies. With firm growth as well as profitability and liquidity, further explanatory variables have been regularly included in the analyses since 1998 and 2015, respectively. In addition to the commonly used control variables, several recent papers include control variables previously unconsidered. For example, Bany-Ariffin etal. (2016), Harjito et al. (2018) examine the influence of firm age. Krapl (2015) adds financial distress as a control variable as distressed firms appear to have higher standard deviations and market betas than less distressed firms. Operating efficiency, dividend policy and capital intensity are just some of the other variables that are also included in the multiple regression in several studies. However, it is difficult to make a general statement about the direction and intensity of the impact of these control variables, because the results of the individual studies differ due to the inclusion of further control variables in varying combinations. Moreover, despite seemingly identical control variables, the papers may nevertheless be not comparable due to the different measurement basis of the variables. While most papers use the logarithm of total assets as the basis for firm size, Theerathorn etal. (1992) use total sales and Shafigullin (2016), Jung (2015), Jung 394 C.Höge-Junge, S.Eckert 1 3 etal. (2018a, b), and Jung etal. (2018a) use the logarithm of total sales. Other studies, e.g., Bany-Ariffin etal. (2016), Harjito etal. (2018), Jung etal. (2018a, b) and Theerathorn etal. (1992), deviate from the usual calculation of the leverage in form of debt ratio by using the logarithm of total debt and the debt-to-equity ratio. With regard to liquidity, two different ratios are also used equally: the quick ratio and the current ratio. This also applies to less common control variables. For example, growth prospects and dividend payout are measured by MTB (market to book) and EBIT growth, and dividend payout and dividend yield, respectively. 4.2 Results In view of the methodological differences described above, it is not surprising that existing research provides inconsistent findings and contradictory conclusions. To allow an initial indication of the effect of multinationality on systematic risk, the papers are grouped systematically by the home country of the firms analyzed. According to the upstream–downstream hypothesis of Kwok and Reeb (2000) and considering the premise that the US market is the most stable (Kwok and Reeb 2000) we expect different effects for US and non-US studies, as the latter mainly refer to developing countries. Within US studies we furthermore distinguish concerning the statistical method of analysis (sample comparison vs. regression analysis). 4.2.1 Results ofnon‑US studies BANY-ARIFFIN etal. (2016) expect a negative impact of multinationality on beta for MNCs from Malaysia as outward foreign direct investments from there are often associated with lower risk. The results, however, show a positive effect of international diversification. The authors argue that a concentration of investments in the ASEAN region may provide an explanation for their findings (Bany-Ariffin et al. 2016). This presumption of a home region concentration is confirmed by Harjito etal. (2018). While firms investing only in Asia experience an increase in risk, firms investing outside are able to reduce their systematic risk thanks to asynchronous business cycles (Harjito etal. 2018). The positive effect of international diversification on systematic risk indicated by Bany-Ariffin etal. (2016) is also confirmed by Marciano and Herlambang (2016). Marciano and Herlambang (2016) provide substantial evidence for an S-shaped relationship between multinationality and beta. In order to estimate the risk exposure to domestic and international factors, Shyu and Ou (2009) use a hybrid capital asset pricing model. The results show that foreign shareholding and psychic distance lead to less domestic systematic risk. Conversely, there is no significant effect of the foreign sales proportion on beta. Bution etal. (2015) and Shafigullin (2016) examine the multinationality effect for firms from developing countries beyond the ASEAN region, specifically Brazil and Russia. Both papers hypothesize that systematic risk will decrease with increasing international exposure. However, their findings do not support these assumptions. Notwithstanding a significant positive effect, the small sample size of 17 firms 395 1 3 Multinationality andsystematic risk: aliterature review… weakens the robustness of the results in the case of Bution etal. (2015). The results of Shafigullin (2016) are insignificant. Thompson (1985) and Bühner (1987) instead make British and West German firms, respectively, and thus developed countries a subject of their research. The results of Thompson (1985) support the hypothesis that systematic risk declines with an increase in multinationality. The potential for risk reduction, however, may be quite small (Thompson 1985). Bühner (1987), in contrast, argues that the risk effect of multinationality appears to be contingent on growth in sales. In order to confirm the supposed influence of multinationality in line with their upstream–downstream hypothesis, Kwok and Reeb (2000) extend their study to a worldwide sample. International involvement seems to increase systematic risk in the case of firms from more stable economies like the US but decreases the systematic risk of firms from more volatile economies. The authors also include firms based in emerging markets and discuss the influence on these markets separately. The results confirm a significant negative relationship for firms from emerging markets. Due to a substantially larger and wider dataset of firms from developed and developing countries as well as due to the different periods of analysis, the findings of Kwok and Reeb (2000) are difficult to compare with the results of Thompson (1985) and Bühner (1987). In addition, the effect on systematic risk varies according to country and industry characteristics as shown by Joliet and Hübner (2008). According to their results, inter-industry differences in systematic risk are even more pronounced than inter-country differences. 4.2.2 Results ofUS studies A closer look at US studies that conduct sample comparisons reveals a significant risk-reduction effect among the papers of Fatemi (1984), Hughes et al. (1975), Michel and Shaked (1986), Shaked (1985). Brewer (1981), Kohers (1976), in contrast, find no significant evidence. Kohers (1976) even finds no evidence within an industry-specific analysis. In contrast to the studies mentioned above, more recent papers make finer distinctions. Collins (1990) differentiates between domestic US firms and US firms with investments in developed and developing countries. He finds no significant effect for firms focusing on developed countries, but a negative effect for firms investing in developing countries. This result directly contradicts the assumption of Kwok and Reeb (2000) and points out the need for a closer look at the target countries. Theerathorn etal. (1992) as well as Siegel etal. (1995) differentiate between UNCs, MNCs, and “intermediates,” firms with a degree of multinationality higher than 10% (0%) but less than 30% (21.5%).10 Their results are contradictory as MNCs have significantly higher systematic risk according to the findings of Theerathorn etal. (1992) whereas UNCs always have the highest risk according to the findings of Siegel etal. (1995). Similar conflicting results arise within studies based on US firms using regression analysis. Concentrating on firms from the restaurant industry, Song etal. (2017) 10 The thresholds of Siegel etal. (1995) are reported in parentheses. 396 C.Höge-Junge, S.Eckert 1 3 support the theory of risk reduction, as the results indicate a decrease in systematic risk through multinationality. Furthermore, a curvilinear relationship between international diversification and risk is confirmed since the risk-reduction effect diminishes as the level of diversification increases (Song etal. 2017). Jung (2015) also finds a significant negative effect on systematic risk. The effect, however, turns out to be linear. Specifically, international diversification appears to decrease restaurants’ systematic risk in a lagged-linear manner. Furthermore, the risk-reduction effect is greater for limited-service restaurants than for full-service restaurants as shown by Jung etal. (2018a, b).11 According to Jung etal. (2018a, b) it is also important to distinguish between positive corporate social responsibilities policies (CSR), e.g., organizing corporate philanthropy, and negative CSR policies, e.g. violations of desirable practice. While a significant effect of internationalization and positive CSR policies on systematic risk cannot be supported, negative CSR policies have a positive but insignificant moderating effect on the relationship between international diversification and systematic risk. The risk-reduction effect of multinationality is also supported for studies that do not concentrate on a specific industry. The regression results of Fatemi (1984) confirm the risk-reduction effects already discovered, employing sample comparison methodology. Similar results are provided by Aggarwal (1979) and Goldberg and Heflin (1995). Referring to Goldberg and Heflin (1995), a 10% increase in DOI induces a beta decrease of 0.025. Although the research question does not directly target the impact of multinationality on systematic risk, the results of Agmon and Lessard (1977) and Chambliss etal. (1994) point towards some potential delineations of the effect of an international involvement on the home market beta. Agmon and Lessard (1977) analyze, if investors respond to the assumed risk reduction associated by companies’ foreign investments. Based on the results of a hybrid capital asset pricing model, they find that the regression coefficient of the US index is much higher for portfolios with little international involvement. Chambliss etal. (1994) apply cross-sectional regressions in order to identify changes in US firms’ systematic risk due to increased integration across European markets. An increase in foreign sales leads to a decrease in sensitivity to the US market. Parameter estimates of foreign assets and leverage are not significant. In contrast to these findings, Olibe etal. (2008), Reeb etal. (1998) provide empirical evidence that international involvement increases systematic risk. Furthermore, Theerathorn etal. (1992) are also able to validate the risk-increasing effect of international diversification discovered on the basis of a sample comparison by regression analysis. Moreover, the authors suggest an increasing integration of the world economy as firms’ systematic risk decreases between earlier and more recent investigation periods. Weak differences in systematic risk for UNCs, intermediates, and MNCs during the last period of analysis (1983–1987) may indicate an increase in global market integration. In addition to the research contributions mentioned above, there are some contributions, based on US data, which provide inconsistent results concerning the impact of 11 Pearsons’ product-moment correlation coefficient of Jung (2015) and Jung etal. (2018a, b) strengthen the theory of a risk-increasing effect contrary to the above-mentioned statement. 397 1 3 Multinationality andsystematic risk: aliterature review… international diversification on systematic risk. (Krapl 2015; Madura and Rose 1989; Barone 1983; Broaden and Samii 2001; Jacquillat and Solnik 1978; Severn 1974). In summary, it can be noted that there is no clear tendency concerning empirical results about the impact of international diversification on systematic risk. Table2 depicts the outcome of the vote-counting method. While empirical evidence for a significantly positive effect of multinationality on systematic risk is found in 17 studies, a significantly negative effect is shown based on the results of 15 studies. In the case of 15 other studies no significant relationship between the two variables could be found. A similar heterogeneous result is obtained for the applied proxy for multinationality as well as the applied regression method. An indication of a positive overall effect is only given for studies of firms from developing countries, where four of seven results are significantly positive. Overall, the inconsistent findings cannot be prima facie explained by certain characteristics of the studies alone. The relationship appears to be more complex and potential moderating effects have to be considered. The following univariate metaanalysis and meta-regression analysis will be used to obtain additional information about this relationship. 5 Univariate meta‑analysis 5.1 Development ofworking hypotheses Univariate meta-analysis evaluates12 the results of different research papers on a particular topic through objective statistical tests and serves as a quantitative literature review. Research dimensions such as measure of research quality and model adequacy, which cannot be used in the original research studies due to the absence of variation, are used in univariate meta-analysis and meta-regression analysis in order to explain the observed variation in results (Stanley etal. 2008). Despite the observed countervailing effects, common argumentation in IB literature based on portfolio theory posits that multinationality leads to a higher degree of cash flow stability and therefore that the risk-decreasing effects on beta are stronger than the risk-increasing effects (Doukas and Kan 2006; Gande etal. 2009). Considering this argumentation in the univariate meta-analysis, we hypothesize: H1 Multinationality has a negative impact on systematic risk. Nevertheless, we do not expect extant empirical results to be completely homogenous. Instead, we would expect that a certain degree of variation can be observed due to differences in the institutional environment of the home country as well as the host countries of MNCs. Kwok and Reeb (2000) argue, that it is important to differentiate between economically upstream foreign investments (i.e., lower degree of risk abroad) and economically downstream foreign 12 See for the coining of the term and further meta-analytical methods Hansen etal. (2022). 398 C.Höge-Junge, S.Eckert 1 3 investments (higher degree of risk abroad). They provide evidence that firms from more stable economies investing in more unstable countries might experience an increase in systematic risk whereas firms from unstable home countries investing in more stable target countries might realize a corresponding decrease in systematic risk. Their findings are further supported by Gande et al. (2009) who show that the valuation benefits from international diversification are higher if a firm invests in a country where corporate governance standards are stronger. Based on these considerations we hypothesize: H2 Firms from countries with a low degree of risk will tend to experience an increase in systematic risk through multinationality. In contrast, firms from countries with a high degree of risk will tend to experience a decrease in systematic risk through multinationality (upstream–downstream hypothesis). Furthermore, we would expect a certain degree of heterogeneity in empirical findings due to changes in market integration overtime. The world’s financial and as well as economic markets have become increasingly integrated in recent decades (Kearney and Lucey 2004) and these changes cannot be neglected. During the early 1990s in particular, the world experienced a radical transformation towards more integration. Market reforms, liberalization, and regional and global integration like that of the European Union, especially the common market of 1992, as well as the opening up of Eastern European economies in 1990 or the formation of NAFTA in 1994 led to a removal of many barriers to the free flow of capital, restrictions on foreign ownership, and economic integration (Errunza and Miller 2000). Innovations in information and communication technology during the past few decades, starting with the invention of the Internet, have contributed to an increase in the degree of integration of financial and economic markets worldwide. The consequences of these developments are empirically well confirmed. Bordo etal. (1998) claim that in the period after 1983 financial markets worldwide "are more stably integrated than in any other" (p. 12) time period between 1880 and 1998. Albuquerque etal. (2005) find that for a sample of 94 countries’ foreign direct investment inflows the contribution of global factors in explaining these inflows has substantially increased at the beginning of the 1990s. The authors explain this development with an increasing integration of financial markets worldwide. Carrieri etal. (2007) find an increasing extent of capital market integration for emerging market countries after 1990. Due to this increasing integration in capital markets which is driven by liberalizations in the global economic environment and revolutionary innovations in information and communications technology, we would expect an increasing correlation between the world’s markets over time, especially since 1990. This expectation is confirmed through empirical studies. Goetzmann etal. (2005) document that international equity correlations have reached a peak in the late twentieth century and conclude that the international diversification potential should be rather low today. Bekaert etal. (2005) provide empirical evidence of increased correlation between national equity markets. Therefore, the risk-reducing effect of corporate multinationality 399 1 3 Multinationality andsystematic risk: aliterature review… can be expected to weaken over this period due to a broader investment horizon available to investors but also an increased co-movement of economic markets. As a consequence, we would expect a weakening effect of MNCs’ international diversification on systematic risk and propose: H3 The negative impact of multinationality on systematic risk has decreased over time (capital market integration hypothesis). 5.2 Findings ofsample comparison The results of the papers using sample comparisons are analyzed by comparing the means and evaluating the significance on differences between the mean values of the samples of UNCs and MNCs. Since the identified papers investigate differences between the mean value on the systematic risk of two or more samples, we use the standardized mean difference as the effect size statistic. Referring to Borenstein etal. (2009), the standardized mean difference is estimated by with d = X1−X 2 SD within Table 2 Vote counting—results Joliet and Hübner (2008), Barone (1983), Olibe etal. (2008), Madura and Rose (1989), Kwok and Reeb (2000), Krapl (2015), Jung etal. (2018a, b) and Shyu and Ou (2009) are included in the vote counting several times due to contradictory results within sub-studies or proxies Significantly positive Significantly negative No significant relationship Number of studies—total 17 15 15 Analyzed markets Number of studies—US 12 11 10 Number of studies—Non-US 5 4 4 Number of studies—emerging market 4 1 2 Analyzed time period Number of studies 1960s–1980s 8 8 7 Number of studies 1990s–2000s 9 7 8 Applied proxy for multinationality Number of studies—foreign sales ratio 5 5 6 Number of studies—foreign asset ratio 2 3 1 Number of Studies—other proxy 5 5 5 Applied regression method Number of studies—simple regression 3 5 4 Number of studies—multiple regression 8 8 8 Number of studies—other 1 0 0 400 C.Höge-Junge, S.Eckert 1 3 Computation of the effect size statistics requires the means X of the two groups, the standard deviation SD , and the sample size n upon which the groups are based (Lipsey and Wilson 2000). In the case of missing values, the effect size can be estimated from other reported statistics.13 Shaked (1985), for example, gives information on t-values for differences in means. Brewer (1981) and Theerathorn etal. (1992), however, do not provide full information in order to calculate the effect size statistics and, thus, are excluded from the analysis. Collins (1990) is removed due to the lack of comparability within the comparison groups, as the group of MNCs are differentiated between MNCs with international operation in developed countries and MNCs with international operation in developing countries. Michel and Shaked (1986) are excluded because of sample identity with the data used by Shaked (1985). The analysis of the remaining papers suggests the use of a random-effects model as the studies included are unlikely to be functionally identical. Due to different data sources (selected companies, time horizon and proxy for multinationality) and variations in the procedures presented in Sect.4.1, we assume that the studies do not share a common effect size, but the true effect size varies and the studies represent a random sample.14 Thus, our goal is not to estimate one true effect, but to estimate the mean of a distribution of effects. This is confirmed by the common statistical test of homogeneity in effect sizes.15 The Q-value of 20.053 is greater than the 0.05 critical value of 11.07 for a chi-square distribution with five degrees of freedom. Thus, the variability among the effect sizes results not merely from a subject-level sampling error, but also from notable random differences between studies (Lipsey and Wilson 2000). Furthermore, I2, the percentage of total variation due to heterogeneity (Higgins etal. 2003), has a value of 75.066%. The application of a fixed effect model is not recommended. A random-effects model is more suitable, nevertheless, it has its imperfections in the case of a small number of studies.16 Irrespective of this, we concentrate on the random-effects model in order to calculate the overall effect size. The required weighted mean of the effect size is computed as (Borenstein etal. 2009): SD within = √( n1−1 ) SD2 1+ ( n2−1 ) SD2 2 n1+n2−2 M∗= ∑k i=1W∗ iYi ∑ k i=1 W∗ i 15 The choice of model should not be based on the test for heterogeneity, which often suffers from low power. Furthermore, the random-effects analysis is reduced to a fixed-effect model if the between-studies variance turns out to be zero (Borenstein etal. 2009). See for further information and further options for handling a statistically significant test of homogeneity Lipsey and Wilson (2000). 16 If the number of studies is very small, the estimate of the between-studies variance will have poor precision. The information needed to apply the random-effects model correctly is missing. However, the random-effects model is still the appropriate model (Borenstein etal. 2009). 13 See for further information Lipsey and Wilson (2000), Appendix B. 14 See for further information Borenstein etal. (2009). 407 1 3 Multinationality andsystematic risk: aliterature review… from developed countries, represent rather crude measures for a test of Kwok and Reeb’s hypothesis (Kwok and Reeb 2000). Furthermore, we assume that the risk-reducing effect of multinationality has weakened over the period of analysis, especially since the beginning of the 1990s. Our empirical results suggest that the development has been even more radical. Whereas we are able to confirm a risk-reducing effect of multinationality for the time period up to 1990, we find a risk increasing effect of multinationality for the period from 1990 onwards. We assume that these results can be attributed on the one hand to the increasing international integration of national capital markets and henceforth the increasing correlations between them. As a consequence, portfolio diversification benefits of firm internationalization as well as country-specific effect differences decrease since the countries are increasingly exposed to the same risks. And, on the other hand these results can be attributed to the overwhelmingly dominant methodological practice to rely on a national market index as the reference for estimating the systematic risk of a firm. The view of limited investment opportunities within national borders, however, cannot be further recommended in light of the increasing capital Table 5 Results of Meta-Regression—stepwise Regression Analysis *, **, ***Indicate that the result is significant at 10, 5, and 1% levels, respectively Model 1 Model 2 Model 3 Intercept −0.0096 −0.1751** −0.2086*** US/ Non-US [Non-US] Developed/ Emerging [Emerging] 0.1812 0.0613 0.2244** Time [90/00er] 0.2853*** 0.2153** RegTyp: [MR] 0.1170 RegTyp: [other] −0.3082* Τ20.0322 0.0214 0.0106 R2 between-study variance 0.00 0.23 0.62 Number of studies 17 17 17 Model 1 Model 2 Model 3 Model 4 Intercept −0.0748 −0.2081* −0.2688*** −0.3242*** US/ Non-US [Non-US] Developed/ Emerging [Emerging] 0.2470* 0.1278 0.2599** 0.1083 Time [90/00er] 0.2527 0.2154* 0.3929*** RegTyp: [MR] 0.1238 0.2608** RegTyp: [other] −0.2836 0.0364 INT measure [FSR] −0.2907* Τ20.0448 0.0500 0.0195 0.0144 R2 between-study variance 0.07 0.00 0.59 0.70 Number of studies 12 12 12 12 408 C.Höge-Junge, S.Eckert 1 3 market integration. Nevertheless, the choice of market index cannot fully explain the change in the effect. Investors seems to assign a higher risk to firms multinationality. Notwithstanding the early stage of this research object, our results can be used to provide a first hint to the true effect of international diversification on firms’ systematic risk and, thus, capital market value. Our results, nevertheless, require further analyses as the small number of previous studies and the underlying methodological variety indicate limitations. This is also confirmed by the meta-regression analysis results, where time is found to be a significant moderator variable, but other study characteristics, such as regression type, also seem to have an impact on the effect. In order to enable a reliable statement on the risk-reducing effect over time, academic research also has to provide new empirical results based on systematic risk estimates using international reference markets. Furthermore, a more sophisticated view of the relationship between the geographical structure of multinational firms and systematic risk seems to be necessary in order to provide more insight regarding the upstream–downstream theory. Verbeke and Brugman (2009) argue that there is an important difference between the degree of internationalization and the degree of international diversification within multinationality. While the degree of internationalization is related to the firm’s international expansion, the degree of international diversification is related to the firm’s geographic dispersion. Despite of the relationship between internationalization and international diversification, the impact of international expansion on systematic risk can be expected to be different depending on the level of geographic dispersion. The majority of the studies considered, nevertheless, refer to proxies measuring international expansion. Similar to Li’s (2007) point about the research on the relationship between multinationality and performance, a parallel consideration of several operationalizations of multinationality in regression analyses is also a promising research perspective. Using the common measures, e.g., foreign sales ratio or foreign asset ratio, in isolation can only capture part of the multinational phenomenon. For comparative purposes, future analyses should still be based on similar procedures and comparative data. However, it is recommended to further enlarge analyses and to go into detail referring to the recent findings. The question of a curvilinear relationship, for example, needs further investigation since Song etal. (2017) finds evidence while Fatemi (1984), Jung (2015) cannot verify it. Furthermore, research should expand the idea of Jung (2015), Song etal. (2017) and investigate whether the results would differ across different industries, different types of MNCs and further possible moderating variables. Referring to Collins (1990), the difference between past studies may be due to sample specific differences in the level of involvement of the sample MNCs in developing countries. In view of this, separate analyses of investments in developing and developed countries should be deepened, too. Besides extending the common procedure, procedures of similar research questions may be helpful in assessing the impact of multinationality on systematic risk. Morck and Yeung (1992), for example, examines stock price reaction to announcements of foreign acquisitions through an event study test. This methodology is an alternative, however, raises several difficulties (Goldberg and Heflin 1995; Binder 1998; McWilliams etal. 1999). 409 1 3 Multinationality andsystematic risk: aliterature review… Moreover, our findings also have practical implications. It is common practice in the course of company valuations to use betas of peer companies and adjust them according to the company specifics. Our findings show that adjustments that were appropriate in the past based on portfolio theory do not seem to be appropriate any more. In the case of multinationality, betas need to be adjusted upward rather than downward. Contrary to current expectations, the additional risks associated with multinationality outweigh the potential reduction in variations of earnings. Executives should therefore carefully consider their internationalization steps, if they intent to increase firm’s capital market value through risk diversification. Appendix Robustness test While univariate meta-analysis supports a change concerning the effect of multinationality on systematic risk: from risk-reducing to risk-increasing, a final inference must await further evaluation of the reliability of these results. In this context, the problem of publication bias has to be considered due to a conceivably too small scope of studies. If the identified studies are a biased sample of all relevant studies, the mean effect computed will reflect this bias (Borenstein etal. 2009). In order to accept or reject the existence of the publication bias, we refer to the visual identification by funnel plot asymmetry with the help of funnel plots with standard error at the vertical axis as measure of study precision.22 In addition to the graphical analysis, we perform an Egger test in order to avoid a subjective conclusion on asymmetry. However, the Egger test is limited to studies of regression analysis using data from the 1990s to the present, as the number of sample comparison studies and the number of linear regression studies from 1960 and 1980s are too small.23 A closer look at Fig.8 demonstrates that the effect sizes of sample comparison studies are within or near the 95% confidence limits around the mean effect centered around the middle of the funnel plot. In case of a symmetrical funnel plot effect sizes will spread uniformly within the 95% confidence limits as effect sizes from smaller studies will scatter more widely at the bottom of the 95% confidence limits with a spread narrowing with increasing precision among larger studies (Sterne and Egger 2001). Larger and smaller study sizes seem underrepresented. Funnel plot asymmetry can be assumed. Small sample size, however, impedes further analysis. This is different for the case of studies employing linear regression methodology. As can be seen in Fig.9, these studies are centered around the top of the plot and 22 As recommended by Sterne and Egger (2001, p. 1053). 23 Stanley etal. (2008) illustrate how a meta-regression model can capture the problem of publication bias by making standard error an important independent variable in the model. However, due to the need to average the effect sizes of several studies, there is a further reduction in sample size, so this approach was not considered further. 410 C.Höge-Junge, S.Eckert 1 3 Fig. 8 Sample comparison findings—funnel plot Fig. 9 Regression analysis findings—funnel plot 411 1 3 Multinationality andsystematic risk: aliterature review… several studies are outside the 95% confidence limits around the mean effect.24 The graphical impression of the funnel plot asymmetry is now confirmed by an Egger test since intercept 𝛽0 having a value of −2.16764 deviates significantly from zero (two-sided p-value: 0.06459). The asymmetry in the funnel plot, however, can be caused by further sources than a publication bias, for instance a true heterogeneity in results.25 An explanation of true heterogeneity is supported by the studies’ horizontal distribution and an improvement of funnel plot asymmetry through the establishment of subgroups. Concentrating on studies using data from the 1990s to the present, we estimate an insignificant Egger test intercept 𝛽0 of −0.49728 (two-sided p-value: 0.67558). In light of the above, we assume that true heterogeneity made an important contribution to the funnel plots’ asymmetry. Based on these results, we conclude that the mean effect of the subgroup of linear regression studies from the 1990s to the present is not affected by publication bias and, thus, represents a reliable reflection of the true impact of multinationality on systematic risk. Although funnel plot asymmetry cannot be denied for sample comparison studies as well as for the subgroup of linear regression studies from 1960 to the 1980s, we assume robustness of the direction of the mean effects because of their statistical significance and an almost unambiguous direction of the single effect sizes. Author contributions Both authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by CH-J. The first draft of the manuscript was written by CH-J and SE. Both authors commented on previous versions of the manuscript. Both authors read and approved the final manuscript. Funding Open Access funding enabled and organized by Projekt DEAL. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for profit sectors. Data availability The dataset generated and analyzed during this study is available from the corresponding author upon request. Declarations Conflict of interest The authors have no relevant financial or non-financial interests to disclose. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. 24 See for further information on design and interpretation of a funnel plot Sterne and Harbord (2004). 25 See for further information Egger etal. (1997). 412 C.Höge-Junge, S.Eckert 1 3 References Albuquerque R, Loayza N, Servén L (2005) World market integration through the lens of foreign direct investors. J Int Econ 66(2):267–295. https:// doi. org/ 10. 1596/ 181394503060 Aggarwal R (1977) Theories of foreign direct investment: a summary of recent research and a proposed unifying paradigm. Econ Aff 22(1):31–45 Aggarwal R (1979) Multinationality and stock market valuation: an empirical study of U.S. companies and markets. Manag Int Rev 19(2):5–21. https:// doi. org/ 10. 1111/j. 15406288. 1977. tb019 69.x Agmon T, Lessard DR (1977) Investor recognition of corporate international diversification. J Finance 32(4):1049–1055. https:// doi. org/ 10. 1111/j. 15406261. 1977. tb033 08.x Amit R, Livnat J (1988) Diversification and the risk-return trade-off. Acad Manag J 31(1):154–166. https:// doi. org/ 10. 5465/ 256502 Bany-Ariffin AN, Matemilola BT, Wahid L, Abdullah S (2016) International diversification and firm’s value: evidence from developing nations. Rev Int Bus Strategy 26(2):166–183. https:// doi. org/ 10. 1108/ RIBS0120140016 Barone RN (1983) Risk and international diversification: another look. Financ Rev 18(2):184–195. https:// doi. org/ 10. 1111/j. 15406288. 1983. tb001 45.x Bartov E, Bodnar GM, Kaul A (1996) Exchange rate variability and the riskiness of U.S. multinational firms: evidence from the breakdown of the Bretton woods system. J Financ Econ 42(1):105–132. https:// doi. org/ 10. 1016/ 0304405X(95) 00873-D Bekaert G, Harvey CR, Ng A (2005) Market integration and contagion. J Bus 78(1):39–69 Binder JJ (1998) The event study methodology since 1969. Rev Quant Financ Acc 11(2):111–137. https:// doi. org/ 10. 1023/A: 10082 95500 105 Bordo M D, Eichengreen B, Kim J (1998) Was there really an earlier period of international financial integration comparable to today?, NBER Working Paper Series, 6738, September 1998 Borenstein M, Hedges LV, Higgins JPT, Rothstein HR (2009) Introduction to meta-analysis. Wiley, Hoboken Brewer HL (1981) Investor benefits from corporate international diversification. J Financ Quant Anal 16(1):113–126. https:// doi. org/ 10. 2307/ 23306 69 Broaden CB, Samii MV (2001) Measuring the impact of globalization: an analysis of the risk and return of multinational firms. Working paper, Southern New Hampshire University Bution J, Masiero G, Oliva F (2015) The systematic risk of the Brazilian textile industry: consequences of their increasing international exposure. Int Proc Econ Dev Res 85:138–145 Bühner R (1987) Assessing international diversification of west german corporations. Strateg Manag J 8(1):25–37. https:// doi. org/ 10. 1002/ smj. 42500 80104 Carrier F, Errunza V, Hogan K (2007) Characterizing world market integration through time. J Financ Quant Anal 42(4):915–940. https:// doi. org/ 10. 1017/ S0022 10900 00034 46 Chambliss K, Madura J, Wright FW (1994) The changing risk profile of U.S.-based multinational corporations exposed to European community markets. J Financ Res 17(1):133–146. https:// doi. org/ 10. 1111/j. 14756803. 1994. tb001 79.x Chira I, Marciniak M (2014) Risk change during crises: How do purely local companies differ from cross-listed firms? Evidence from the European Crisis of 2010–2012. J Econ Bus 74:11–23. https:// doi. org/ 10. 1016/j. jecon bus. 2014. 03. 001 Choi JJ, Severn AK (1992) On the effects of international risk, segmentation and diversification on the cost of equity capital. J Multinatl Financ Manag 3(1):1–19 Collins JM (1990) A market performance comparison of US firms active in domestic, developed and developing countries. J Int Bus Stud 2(2):271–287. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84903 35 Contractor FJ (2007) Is international business good for companies? The evolutionary or multi-stage theory of internationalization vs. the transaction cost perspective. Manag Int Rev 47(3):453–475. https:// doi. org/ 10. 1007/ s115750070024-2 Doukas JA, Kan OB (2006) Does Global Diversification Destroy Firm Value? J Int Bus Stud 37(3):352–371. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84002 03 Eckert S, Ral-Trebacz A, Sefrin Y, Trautnitz G (2016) Two tales about one world? Reconstructing financial analysts’ view about the valuation impact of corporate multinationality through the lens of international business research. Int J Bus Res 16(3):31–60. https:// doi. org/ 10. 18374/ IJBR16-3.3 413 1 3 Multinationality andsystematic risk: aliterature review… Egger M, Smith GD, Schneider M, Minder Ch (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629–634. https:// doi. org/ 10. 1136/ bmj. 315. 7109. 629 Errunza VR, Miller DP (2000) Market segmentation and the cost of capital in international equity markets. J Financ Quant Anal 35:577–600. https:// doi. org/ 10. 2307/ 26762 56 Fatemi AM (1984) Shareholder benefits from corporate international diversification. J Financ 39(5):1325–1344. https:// doi. org/ 10. 1111/j. 15406261. 1984. tb049 10.x Fisch C, Block J (2018) Six tips for your (systematic) literature review in business and management research. Manag Rev Q 68:103–106. https:// doi. org/ 10. 1007/ s113010180142-x Gande A, Schenzler C, Senbet LW (2009) Valuation effects of global diversification. J Int Bus Stud 40(9):1515–1532. https:// doi. org/ 10. 1057/ jibs. 2009. 59 Glass GV, McGaw B, Smith ML (1981) Meta-analysis in social research. Sage Publications, Thousand Oaks Goetzmann WN, Li L, Rouwenhorst KG (2005) Long-term global market correlations. J Bus 78(1):1–38 Goldberg SR, Heflin FL (1995) The association between the level of international diversification and risk. J Int Financ Manag Acc 6(1):1–25. https:// doi. org/ 10. 1111/j. 1467646X. 1995. tb000 47.x Haegele MJ (1974) Exchange Rate Expectations and Security Returns. PhD Dissertation, University of Pennsylvania Hansen C, Steinmetz H, Block J (2022) How to conduct a meta-analysis in eight steps: a practical guide. Manag Rev Q 72:1–19. https:// doi. org/ 10. 1007/ s1130102100247-4 Harjito A, Bany-Ariffin AN, Matemilola BT, Ashari M (2018) International diversification and multinational firms’ risk: Do locations of investments matter? J Econ Manag 12(1):357–368 Hennart J-F (2007) The theoretical rationale for a multinationality-performance relationship. Manag Int Rev 47(3):423–452. https:// doi. org/ 10. 1007/ s115750070023-3 Higgins JPT, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 237:557–560. https:// doi. org/ 10. 1136/ bmj. 327. 7414. 557 Hughes JS, Logue DE, Sweeney RJ (1975) Corporate international diversification and market assigned measures of risk and diversification. J Financ Quant Anal 10(4):627–637. https:// doi. org/ 10. 2307/ 23306 11 Jacquillat B, Solnik B (1978) Multinationals are poor tools for diversification. J Portf Manag 4(2):8– 12. https:// doi. org/ 10. 3905/ jpm. 1978. 408629 Joliet R, Hübner G (2008) Corporate international diversification and the cost of equity: european evidence. J Int Money Financ 27(1):102–123. https:// doi. org/ 10. 1016/j. jimon fin. 2007. 04. 008 Jung S (2015) An Examination of U.S. Restaurant Firms’ Internationalization in a Risk Context, PhD Dissertation, University of Nevada Jung S, Dalbor M, Lee S (2018a) Internationalization as a determinant of systematic risk: the role of restaurant type. Int J Contemp Hosp Manag 30(8):2791–2809. https:// doi. org/ 10. 1108/ IJCHM0620170321 Jung S, Kim JH, Kang KH, Kim B (2018b) Internationalization and corporate social responsibility in the restaurant industry: risk perspective. J Sustain Tour 26(7):1105–1123. https:// doi. org/ 10. 1080/ 09669 582. 2017. 14212 01 Kearney C, Lucey BM (2004) International equity market integration: theory, evidence and implications. Int Rev Financ Anal 13(5):571–583. https:// doi. org/ 10. 1016/j. irfa. 2004. 02. 013 Kohers T (1976) A risk-return comparison: U.S. multinational and U.S. domestic corporations. Univ Michigan Bus Rev: 25–27 Krapl AA (2015) Corporate international diversification and risk. Int Rev Financ Anal 37:1–13. https:// doi. org/ 10. 1016/j. irfa. 2014. 11. 005 Kuckertz A, Block J (2021) Reviewing systematic literature reviews: ten key questions and criteria for reviewers. Manag Rev Q 71:519–524. https:// doi. org/ 10. 1007/ s1130102100228-7 Kwok ChCY, Reeb DM (2000) Internationalization and firm risk: an upstream-downstream hypothesis. J Int Bus Stud 31(4):611–629. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84909 25 Li L (2007) Multinationality and performance: a synthetic review and research agenda. Int J Manag Rev 9(2):117–139. https:// doi. org/ 10. 1111/j. 14682370. 2007. 00205.x Lipsey MW, Wilson DB (2000) Practical meta-analysis. Sage Publications, Thousand Oaks Madura J, Rose LC (1989) Impact of international sales degree and diversity on corporate risk. Int Trade J 3(3):261–276. https:// doi. org/ 10. 1080/ 08853 90890 85236 69 Marciano D, Herlambang A (2016) The Effect of Internationalization to Risk in Industrial Manufacturing Company the Listed on the Indonesia Stock Exchange 2010–2014 Period. 13th Ubaya 414 C.Höge-Junge, S.Eckert 1 3 International Annual Symposium on Management, March 18–20, 2016, Ho Chi Minh City, Vietnam: 344–358 McWilliams A, Siegel D, Teoh SH (1999) Issues in the use of the event study methodology: a critical analysis of corporate social responsibility studies. Organ Res Methods 2(4):340–365. https:// doi. org/ 10. 1177/ 10944 28199 24002 Michel A, Shaked I (1986) Multinational corporations vs domestic corporations: financial performance and characteristics. J Int Bus Stud 17(3):89–100. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84904 35 Morck R, Yeung B (1992) Internationalization: an event study test. J Int Econ 33(1–2):41–56. https:// doi. org/ 10. 1016/ 00221996(92) 90049-P Nguyen Q (2017) Multinationality and performance literature: a critical review and future research agenda. Manag Int Rev 57(3):311–347. https:// doi. org/ 10. 1007/ s115750160290-y Olibe KO, Michello FA, Thorne J (2008) Systematic risk and international diversification: an empirical perspective. Int Rev Financ Anal 17(4):681–698. https:// doi. org/ 10. 1016/j. irfa. 2007. 09. 004 Reeb DM, Kwok ChCY, Baek HY (1998) Systematic risk in the multinational corporation. J Int Bus Stud 29(2):263–279. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84900 36 Rugman AM (1976) Risk reduction by international diversification. J Int Bus Stud 7(2):75–80. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84907 02 Severn AK (1974) Investor evaluation of foreign and domestic risk. J Finance 29(2):545–550. https:// doi. org/ 10. 1111/j. 15406261. 1974. tb030 68.x Shafigullin I (2016) Impact of corporate diversification on company performance and risk: evidence from Russia. Master’s Thesis, St. Petersburg State University Shaked I (1985) Are multinational corporations safer? J Int Bus Stud 17(1):83–106. https:// doi. org/ 10. 1057/ palgr ave. jibs. 84904 18 Shyu Y-W, Ou N (2009) Corporate internationalization and systematic risk. Working Paper, Chang Gung University, Tao-Yuan, Taiwan Siegel PH, Omer K, Rigsby JT, Theerathorn P (1995) International diversification: a review and analysis of the evidence. Manag Financ 21(9):50–77. https:// doi. org/ 10. 1108/ eb018 533 Stanley TD, Doucouliagos C, Jarrell SB (2008) Meta-regression as the socio-economics of economics research. J Socio-Econ 37(1):276–292. https:// doi. org/ 10. 1016/j. socec. 2006. 12. 030 Song S, Park S, Lee S (2017) Impacts of geographic diversification on restaurant firms’ risk: domestic vs. international diversification. Int J Hosp Manag 61:107–118. https:// doi. org/ 10. 1016/j. ijhm. 2016. 11. 011 Theerathorn P, Alcerreca-Joaquin C, Siegel PH (1992) Sources of risk in U.S.-based multinational companies: a comparative cross-sectional time-series analysis. Adv Int Account 5:193–209 Thomsen AH (2012) Risk and Internationalization: An Empirical Study of Large Firms in the Euro Countries 2006–2011. Unpublished Thesis, Aarhus University Thompson RS (1985) Risk reduction and international diversification: an analysis of large UK multinational companies. Appl Econ 17(3):529–541. https:// doi. org/ 10. 1080/ 00036 84850 00000 55 Shapiro AC, Balbirer SD (2000) Modern corporate finance: a multidisciplinary approach to value creation. Prentice Hall, Upper Saddle River Sterne JAC, Egger M (2001) Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. J Clin Epidemiol 54:1046–1055. https:// doi. org/ 10. 1016/ S08954356(01) 00377-8 Sterne JAC, Harbord RM (2004) Funnel plots in meta-analysis. Stand Genomic Sci 4(2):127–141. https:// doi. org/ 10. 1177/ 15368 67X04 00400 204 Valentine JC, Pigott TD, Rothstein HR (2010) How many studies do you need? A primer on statistical power for meta-analysis. J Educ Behav Stat 35(2):215–247. https:// doi. org/ 10. 3102/ 10769 98609 346961 Verbeke A, Brugman P (2009) Triple-testing the quality of multinationality-performance research: an internalization theory perspective. Int Bus Rev 18(3):265–275. https:// doi. org/ 10. 1016/j. ibusr ev. 2009. 01. 005 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.