Multidimensional approach to factors affecting Japanese firms’ expansion in the MENA region

This paper investigates whether factors specific to firms, countries, mode of entry, and the timing of entry contribute to value-creation from the investment of a sample of Japanese firms in the MENA economies over the period 2006–2014. Results show that while this expansion was on average associated with positive wealth effects, such effects were relatively stronger in recent years. Also, results from firms’ choice of entry mode show that the least risky modes of entry were rewarded by the highest returns. Specifically, expansion through acquisitions and new plants was on average associated with negative wealth effects compared to expansion through non-FDI modes of entry. Besides, there is an heterogeneity in wealth effects across countries. In addition, due to changing economic and institutional environments, differences in wealth effects from investments in the same country are observed overtime. Subjects: Middle East Studies; International Relations Theory; Economics; Finance


Introduction
The Middle East and North Africa (MENA) region stands on the threshold of a decisive economic development stage. The World Bank estimated that the region needs to create about 300 million jobs by 2050 to accommodate with rapid population growth. In this regard, the role of foreign investment can be critical to the future of the region's economic growth (Neaime & Marktanner, 2009). Foreign investment by multinational enterprises (MNEs) in the MENA region has been low in comparison to other regions. In 2000, the region received only Anis Khayati ABOUT THE AUTHOR Anis Khayati received his Ph.D. in Economics from Nagoya City University, Japan. He has been a faculty member at different universities in the Middle East, where he taught many courses in economics and finance. He is currently an Assistant Professor in Economics at the College of Business, University of Bahrain. His main research interests are in the areas of FDI, multinational firms, economic growth, technology transfer, investment decisions, educational economics and monetary economics. He published many articles in regional and international journals.

PUBLIC INTEREST STATEMENT
Because of its various advantages (employment, technological transfer, access to international markets, productivity gains and competitiveness, etc.), foreign investment is coveted by developed as well as developing countries. Japanese investment outflows in MENA countries are limited comparing to other regions. However, Japan remains a relevant investor in the region notably in some specific sectors. This paper attempts to investigate the factors that influence the shareholder wealth from Japanese investments in MENA economies. Those factors are found to be closely related to the specificities of countries and firms, as well as the type and timing of investment.
0.4% of global investments (Dichtl & Drobetz, 2011). This rate increased to 2.7% in 2017 (World Bank Group, 2018). Despite this upsurge, MENA countries are still facing significant structural challenges to secure their right share in global investments.
The FDI inflow distribution among MENA countries has been uneven. The main FDI destinations into the region have been, and remain, Israel and the United Arab Emirates (World Bank, 2018). The main sectors related to FDI inflows into MENA countries are oil manufacturing and non-tradable sectors, such as construction and real estate. Those sectors received more than 50% of FDI inflows during the period 2000-2012 (Anyanwu, Yameogo, & Ali, 2016). Yet, they are not creating enough potential jobs to absorb the high unemployment rate in the region.
For most MENA countries, the United States and European countries, notably France and UK, are the main investors. Japan is also considered a relevant investor mainly in the oil manufacturing industry in Saudi Arabia and the UAE. Nevertheless, Japanese investments in MENA region represent only 0.59% of total Japanese investment outflows, in comparison to 27.55% for Asia and 6.63% for Latin America (JETRO, 2018). The sectorial distribution of Japanese FDI in MENA is different from other regions, where most investments are destined to communication and financial sectors, in addition to the sector of machinery and equipment. Also, the number of Japanese firms undertaking FDI modes of investment in MENA countries has been always limited if compared to Asian and South-American countries. Indeed, Japanese firms are known to be typical followers rather than first-movers. This includes even their investments in South-East Asia and more recently in geographically and culturally neighbouring China (Belderbos, Fukao, & Iwasa, 2006). A question that arises in face of this evidence is whether expansion in MENA economies has been associated with value creation.
The objective of the paper is to investigate the consequences of Japanese firms' decisions to expand into MENA countries by separately evaluating the wealth effects associated with when, where and how firms expand. We use a sample of 115 announcements of expansion into the region over the period 2006-2014.
The framework for the empirical analysis is a two-stage estimation procedure. In the first stage, event study methodology is used to estimate abnormal stock returns to shareholders of Japanese publicly traded firms. In the second stage, the estimated abnormal returns are used as the dependent variable to examine what factors contribute to value creation or value reduction from expansion in the MENA region.
The outline of this study is as follows: In section 2, a conceptual framework of the study is constructed. Section 3 states the data sources and estimation methodology. Section 4 presents the empirical results and discusses the findings and section 5 concludes.

Literature review and theoretical framework
While the literature suggests that the characteristics of firms, host countries, mode and timing of entry are important factors that affect expansion in MENA countries (Kinda, Plane, & Véganzonès-Varoudakis, 2011;Al-Khouri, 2015;Bouyahiaoui & Hammache, 2014;Pigato, 2009;Mohamed & Sidiropoulos, 2010;Jabri, Guesmi, & Abid, 2013), comprehensive studies related to these factors have been lacking. This paper contributes to the previous literature through the examination of expansion decisions on shareholder wealth.
According to the general theory of internalization, the existence of imperfections in the goods and factors markets serve to undermine the theoretical reasons for free trade (Buckley & Casson, 2009). The development of MNEs can be seen as a response to such imperfections. For example, setting up a subsidiary in the foreign country allows the MNE to create an internal market and thus overcome imperfections in the goods market such as tariffs and other types of government-imposed distortions. Given this theoretical framework, multinational expansion in the presence of intangible assets should be viewed by the market as increasing firm value (Pugel, Kragas, & Kimuri, 1996).
The eclectic or OLI paradigm embraces the internalization theory and traditional trade theories (Dunning, 2002), and systematizes the benefits for firms that operate internationally (Faeth, 2009). Specifically, the ownership advantage ensues when grander technology and management skills permit the firm to participate in an overseas market despite the transaction costs. Dunning also noted the business had domination over its own explicit advantages, and applying them overseas results in increased marginal cost-effectiveness in relation to its opponents. However, several academics consider that ownership advantages tend to obsolete as markets become global, and postulate that other variables can also perform decisive and significant parts in firms' decision to invest abroad (Pedersen, 2003).
This analysis suggests that gains to shareholders can be partially explained by the presence of firm-specific assets. To test whether this effect is present for the firms in our sample, two proxies are used: R&D and advertising intensity. The null hypothesis is formulated as follows: Hypothesis 1. The wealth effects associated with announcements of expansion in MENA economies are positively related to proxies for firm-specific assets.
Concerning the mode of entry factors, theoretical and empirical analyses of the operation of multinational enterprises give rise to some predictions about what mode of entry firms will choose when expanding into foreign countries. For example, firms that wish to achieve speedy entry are more likely to enter via a joint venture or an acquisition. Joint ventures are also more likely when firms expand into unfamiliar host countries (Caves, 1996). In addition, the presence of proprietary technologies makes entry more likely in the form of a new plant or a subsidiary (Eicher & Kang, 2005). Finally, if a firm wishes to secure a stream of short-term profits, then licensing or sales and service contracts are preferable (Conconi, Sapir, & Zanardi, 2016).
The hypotheses formulated below aim to capture two ways in which mode of entry factors may contribute to differences in wealth effects. First, some modes of entry may be more appropriate than others in a given time period because of differences in the expected risk and return they entail. Specifically, expansion through FDI in the form of joint ventures, acquisitions, and new plants, is associated with higher expected return than non-FDI expansion such as licensing, since it affords the firm increased flexibility in the face of changing market conditions, and it presents advantages such as ownership and control of assets (Reddy, 2014). However, FDI modes are also associated with higher investment risk environment such as greater susceptibility to fluctuating exchange rates and macroeconomic instability. In fact, non-FDI modes represent a choice that can minimize investment risk, give access to difficult markets, induce low commitment of recourses and generate higher returns on investment (Anderson & Gatignon, 1986;Yang & Maskus, 2009). Therefore, hypothesis 2 is formulated as: Hypothesis 2. There is no difference in the wealth effects associated with expansion through different modes of entry in a given time period.
Second, as the number of firms entering the region increases overtime, availability of information on difficulties and successes with FDI might improve. Therefore, the choice of entry mode in subsequent investments may be affected by experience with a particular entry mode. This notion, often referred to as path dependency, suggests that once a firm undertakes a successful acquisition, greenfield or licensing entry, it may prefer to continue with the same mode of entry. Studies have found that prior entries and know-how transfer through a partner increased the probability of subsequent transfer through similar mechanisms (Westney, 1988). However, other authors such as Chang and Rosenzweig (2001) consider that this approach has many limitations and they underscore the importance of experience in foreign investment, as companies learn from early entries and adapt the modes of subsequent ones.
To test whether the perception of FDI through a given entry mode changes overtime, the null hypothesis is formulated as follows: Hypothesis 3. The wealth effects associated with expansion through FDI and non-FDI modes are the same over time.
Also, a potential key component of a firm's decision to expand into foreign countries is the timing of entry (Gallego, Hidalgo, Acedo, Casillas, & Moreno, 2009).
Empirically, various studies have provided supporting evidence for the positive effect of entry timing on performance indicators, such as sales growth, profitability, market share, and establishment of a competitive position in host markets (e.g., Cui & Lui, 2005;Delios & Makino, 2003;Luo, 2001;Luo & Peng, 1998;Pan, Li, & David, 1999). For example, Luo (2001) finds that early entrants to a foreign market outperform late entrants in local market expansion, as measured by sales growth. Pan et al. (1999) find that early entrants achieve higher market shares and profitability. Delios and Makino (2003) conclude that the earlier the market entry, the greater is a firm's subsidiary size relative to that of its competitors. However, early investors might face risks in the start-up phase and can be disadvantaged in terms of overall profitability (Zachary, Gianiodis, Payne, & Markman, 2015). Therefore, we hypothesize the following: Hypothesis 4. The wealth effects associated with expansion in MENA economies are the same across time.
Besides, wealth effects for shareholders of Japanese firms expanding in MENA economies may vary with the choice of location. If the host country offers a stable economic and political environment and more developed instutions, this reduces the investment risk for foreign firms, and therefore expansion into that country may lead to positive wealth effects (Benassy-Quere et al., 2007;Cleeve, 2008;Mohamed & Sidiropoulos, 2010). But other studies show inconclusive results for some of the mentioned variables (Biswas, 2002;Mhlanga, Blalock, & Christy, 2010;Vijayakumar, Sridharan, & Rao, 2010). In this study, to test whether openness to FDI is associated with positive wealth effects, we use as a proxy the ratio of FDI to GDP, and the null hypothesis is formulated as follows: Hypothesis 5. The wealth effects associated with announcements of expansion in MENA economies are positively related to FDI openness in the host country.
In addition, overtime, international experience in a specific country might enhance a firm's understanding, competence, and confidence, as well helping it to develop a more accurate perception of foreign risks and returns. This means that an experienced multinational firm is confident and willing to take risks to enter foreign markets. Therefore, a firm's cumulative international experience is positively related to the degree of control it exerts on the foreign business entity. However, changes in the economic, political, and investment environment over time may entail different wealth effects for shareholders of early versus recent entrants. Therefore, to test for this effect, the null hypothesis is formulated as: Hypothesis 6. The wealth effects associated with investments in a given country are the same overtime.

Data collection, analytical context and methodology
The sample of Japanese companies is drawn from the editions of Kaigai Shinshutsu Kigyo soran (Japanese Overseas Investment Data). The information about firms is matched with the availability of firm-specific variables used in the study and sourced from the Japan Company Handbook, documented annually by Toyo Keizai. The stock exchange data was obtained from the Nikkei Economic Electronic Databank System (NEEDS), and host country data were obtained from the editions of the World Development Indicators compiled by the World Bank.
The framework for the empirical analysis is a two-stage estimation procedure. In the first stage, event study methodology is used to obtain abnormal announcement returns for each firm. These returns are corrected for heteroscedasticity in the second stage and regressed on a set of explanatory variables that aim to evaluate what factors contribute to value creation from foreign expansion.
In the first-stage estimation, a basic assumption about security markets is that they are informationally efficient, and therefore the effect of an unanticipated economic event on the value of a firm will be instantaneously reflected in its stock price. The economic event of interest here is an announcement of expansion by a Japanese firm into a MENA economy. An investment that is expected to result in higher profitability for the firm should be associated with increases in the value of equity. In contrast, an investment that is expected to lower the firm's profitability should be associated with a decrease in the value of equity. Such increases or decreases in the firm's value constitute abnormal returns reflecting the impact of the investment announcement on the firm's value (MacKinlay, 1997).
The date associated with the event of interest is referred as the event date (henceforth labeled as t = 0). Since information leaks can occur prior to the event date, it is conventional to analyze the cumulative abnormal returns (CARs) over a number of days around t = 0, known as an event window.
For firm i and event date t the abnormal return is: where AR it , R it and E (R it /R mt ) are the abnormal, actual and normal returns respectively for date t, and R mt is the market return. By conditioning the firm's expected return on the market return, the market model assumes a stable linear relation between the market return and the security return. For any security i, the market model is: In order to estimate the parameters of the market model, it is necessary to define an estimation window. In this paper, an estimation window of 30 trading days before the event day is used. The abnormal return for firm i observed on date t, AR it , is simply the residual of the estimated model: For each firm, the cumulative abnormal return, CAR t , is measured as the sum of the abnormal returns over the days in the event window. Then, as mentioned above, it is important to perform sensitivity analyses by estimating the CARs with event windows of various lengths. In this paper, the CARs for each firm are estimated for two different event windows. The first of these is simply the event date and so: CAR t = AR it . The second event window measures the abnormal returns three days including the day before, the day of and the day after the announcement; and is calculated as follows: In the second stage, the estimated abnormal returns are used as the dependent variable in a regression equation that seeks to evaluate what factors contribute to value creation or value reduction from expansion in MENA economies. Unlike a regular dependent variable, however, the abnormal returns in the first stage are measured with error which introduces a heteroscedasticity problem in the second stage regression. This may render estimates of the coefficients of the independent variables inefficient when applying the ordinary least squares estimation technique.
To correct for heteroscedasticity, the second stage uses weighted least squares, where observation of the CAR vector and the independent variables is weighted by the inverse of the standard error from the regression used to estimate the parameters of the market model in the first stage.
The second stage regression is written as follows: Where i refers to firms, j to countries, t to years and m to entry mode.
Time-invariant country effects are included in the regression through dummy variables for each country, which take on the value of 1 if an investment was made in that country and 0 otherwise. How firms choose to enter is included in the regression through three entry modes corresponding to non-FDI expansion, acquisitions and new plants. Equation (5) uses non-FDI modes as the benchmark for comparison. Also, a dummy for each year in the sample period 2006-2014 is included to capture time effects and the year 2010 is taken as the benchmark for evolution across time. The regression also includes two sets of explanatory variables specific to firms and host countries in the sample. Firm-specific explanatory variables are motivated by empirical tests of the internalization theory. Specifically, intensity of expenditures on R&D and advertising are included in the regression as proxy variables for intangible assets. In order to control for firm size, construction of these variables involves scaling expenditures on R&D and advertising for each firm by the firm's total assets. The country-specific variable, FDI intensity, measures FDI in the host country as a percentage of GDP. Inclusion of this variable in Equation (5) is done to capture the relative openness of the investment environment in the host country where firm i made an investment for the year in which the investment was announced. Table 1 reports summary statistics for abnormal returns for the day of the announcement and three-day windows over the periods 2006-2009 and 2010-2014. The mean CARs from announcements of FDI expansion in Table 1, show that in the 2010--2014 period, abnormal returns over event windows are consistently higher than those over the 2006-2009 period. The standard deviation in the 2010-2014 sub-sample is also higher, ranging from 4.6% on the announcement day to 5.5% over the three-day window. Overall, mean CARs over the period 2006-2014 are on average positive, but still the relative split between positive and negative returns suggests that differences in wealth effects may exist with respect to when, where and how firms choose to expand in MENA economies. The empirical results from the second-stage estimation are presented below.

Differences in wealth effects associated with firm-specific factors
Japanese firms investing in MENA economies display heterogeneity in terms of firm-specific characteristics such as the intensity of advertising and R&D expenditures. Inclusion of these variables in the estimation of Equation (5) was motivated by the predictions of the internalization theory.
On average, the estimated coefficients of firm-specific variables in Table 2 do not provide strong support for predictions of the internalization theory in this sample. Abnormal returns from expansion are positively related to R&D intensity and significant at the 10% level only on the day of the announcement in the 2010-2014 sub-sample. Advertising intensity in this sub-sample is negative and significant over the three-day window at the 10% level. But, as with R&D intensity, this effect disappears in the full sample. In face of this weak evidence, hypothesis 1, that wealth effects associated with announcements of expansion in MENA economies are positively related to proxies for firm-specific assets is not confirmed.

Differences in wealth effects associated with mode-of-entry factors
To address mode-of-entry effects, dummies corresponding to announcements of both FDI (acquisitions and new plants) and non-FDI expansion were used in the estimation of Equation (5). The estimated coefficients on mode-of-entry dummies suggest that acquisitions and new plant investments in the full sample are associated with statistically significant abnormal returns that are, on average, lower compared to expansion through non-FDI modes. Therefore, hypothesis 2 that wealth effects are similar with respect to how firms expand into the region is rejected for acquisitions and new plants in the full sample. Overall, the estimated coefficients on mode-of entry variables indicate that expansion by Japanese firms through entry modes associated with large long-term equity commitment is on average less value enhancing compared to operations that aim to secure a steady stream of short-term profits.

Differences in wealth effects associated with country-specific factors
The discussion of country effects here focuses on the behavior of the variable measuring the intensity of FDI inflows in the host country. The estimated coefficient on FDI intensity is consistently positive and significant at the 10% level in the full sample. As may be expected, expansion into countries with more openness to FDI is on average associated with higher abnormal returns. This points to the benefits of operating in an environment of better developed institutions and higher surrounding economy. In view of these results, hypothesis 5, that wealth effects are positively related to FDI openness in the host country is confirmed.

Differences in wealth effects associated with the timing of expansion
Along with firm characteristics and mode-of-entry dummies, Equation (5)  In general, the significance and magnitude of positive wealth effects in 2013 and 2014 suggest that relative improvement in the economic and political environment in MENA countries is a key factor influencing abnormal returns. Overall, hypothesis 4, that the wealth effects associated with expansion are the same across time is rejected.

Country and mode-of-entry differences over time
As discussed above, it is also of interest to test whether expansion through a given mode of entry or in a given country entails differences in abnormal returns over time. Differences in estimated coefficients on mode-of-entry and country dummies over the periods 2006-2009 and 2010-2014 are computed and reported in Table 3. The results suggest that valuation of investments through new plants and acquisitions when compared with non-FDI clearly differ over time. Therefore, hypothesis 3, that wealth effects associated with expansion in MENA economies are the same across time is rejected for new plants and acquisitions. Differences in the perception of entry modes over time suggest that FDI in MENA economies may be accompanied by a learning process: improved availability of information about consequences from entry through a given mode may, over time, entail reevaluation of the reliability of investments through acquisitions and new plants.
Finally, to establish whether abnormal returns from expansion in a given country differ over time (hypothesis 6), differences in the coefficients of the country dummies over the 2006-2009 and 2010-2014 are also reported in Table 3. The results from Wald tests for differences in coefficients lead to rejection of hypothesis 6 over both the one and three-day windows for Saudi Arabia, Israel, Turkey, Egypt and the U.A.E, and for Morocco over the three-day window. Overall, these results are consistent with evidence of changing economic environments in MENA economies.

Conclusion
The focus of the empirical analysis in this paper is to investigate whether factors specific to firms, countries, mode of entry and the timing of investments contribute to value-creation or valuereduction for shareholders of Japanese firms expanding in MENA economies.
Several conclusions emerge from the empirical analysis. First, the timing of entry is an important factor that influences shareholder wealth effects for firms in the sample: Value creation from recent expansion highlights a relative improvement in the investment climate in some MENA countries. Second, differences in shareholder wealth are evident with respect to the choice of entry mode. Expansion through FDI in the form of new plants and acquisitions is generally perceived as less value enhancing than expansion through non FDI modes such as licensing. Also, the perception of entry through acquisitions and new plants relative to non-FDI expansion changes overtime. This result suggests that expansion into MENA economies may be accompanied by a learning process: improvements in availability of information about the business environment induce changes in investment strategies over time. Finally, while expansion into countries with more openness to FDI seems to be a factor contributing to value creation; overtime, expansion into most countries in the sample is associated with differences in wealth effects. This result is consistent with the gradually changing economic and institutional environment in many MENA economies over the past decade.
The present study could be considered as a promising starting point for further analysis of the factors that may affect the returns from Japanese investment in the MENA region. Also, this analysis can be extended to other set of countries and regions, in a dynamic way that includes both inward and outward perspectives. However, the study is limited by the exclusive use of market openness as a proxy for the country-specific factors related to foreign investment. The research reveals the importance of separate studies for dissimilar levels of economic growth, i.e. various country-specific variables. In fact, by grouping diverse economies, there is a likehood some significant impacts might not appear. Indeed, MENA countries generally differ in many aspects such as infrastructure, market size, endowments with human capital and natural resources, institutional reforms, economic and political stability, etc. The availability of those factors can reduce the investment risk for foreign firms and therefore the expansion into the country may lead to positive wealth effects, all else being equal (Sin & Leung, 2001). On the contrary, other studies showed that the effects of some of the mentioned variables could be negative or insignificant (Assuncao et al., 2011). Thus, further empirical analyses that include more country-specific factors would be useful to strengthen the results obtained and to delve into the different aspects that have emerged from this study. This also would allow recognizing the main types of country-specific factors that would induce higher returns from Japanese investment, and therefore generate higher potentialities of FDI inflows into the region.