This is an open access work published under the terms of the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0), which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use (https://creativecommons.org/licenses/by-sa/4.0/)

1. Introduction

Since the implementation of Renovations or “Doi Moi” towards a market economy in 1986, Vietnam has gained significant achievements in economic development and international integration. Vietnam has become a reliable destination of multinational corporations (MNCs) as a result of its political and macroeconomic stability and improved ease of business. In response, several domestic enterprises have started investing abroad since 1989. This is considered as initial steps in Vietnamese Investment Development Path (IDP), where Outward Foreign Direct Investment (OFDI) flows have increased  at a level much lower than inward foreign direct investment (IFDI) (Dunning, 1981, 1986).

Due to its large economic openness and reliance on international cooperation (Quang, 2023), Vietnam’s OFDI has become an important factor, contributing to its process of extensive economic integration and value chain enhancement. From strategically investing into neighbours such as Laos and Cambodia in the 1990s, Vietnam has expanded its OFDI locations to approximately seventy countries and territories by 2021 (VFIA, 2022). While there have been many studies investigating the motives of foreign direct investment (FDI) in Vietnam (Quang et al., 2022; Nguyen, 2008a) or its effects on Vietnamese economy (Nguyen, 2021; Jenkins, 2006; Le & Pomfret, 2011; Nguyen, 2008b; Nguyen et al., 2020),  few have endeavoured to analyse Vietnam’s OFDI motivations.

OFDI from emerging and developing countries are implemented in distinctive strategies, as prevailing theoretical approach to OFDI largely focuses on MNCs in developed countries (Thu Ha et al., 2021; Abdulrazak et al., 2023; Eatessam et al., 2023). Empirical research strands on OFDI in developing economies are nearly on prominent powers’ strategies such as India and China. These giant economies themselves hold geopolitical senses and consequently tend to design distinctive OFDI plans, which might not be applied to medium-scale economies like Vietnam. For example, China showed significant interests on African markets (Mourao, 2018), while Indian MNCs largely invested in developed countries (Hansen, 2009). Therefore, the analysis of OFDI motives in Vietnam, a medium-scale economy, contributes to the current literature on OFDI from developing countries, and serves as a measure of comparison with other nations with similar economic sizes at different stages of IDP. Using regression models with panel data of 159 economies during the period 2007-2021, this paper aims to investigate the determinants of Vietnamese OFDI’s location choice and the amount of new-registered OFDI capital in host economies.

2. Literature Review

The IDP model is a classic theoretical framework analysing worldwide FDI patterns, in which the values of FDI flows inward and outward depend on their levels of economic development (Dunning, 1981; 1986). Developing countries are at IDP’s initial stages IDP. Their OFDI scale begins flourishing; however, it would be much lower the IFDI one. Dunning et al. (2010) provide empirical evidence on Trade Development Path (TDP) in parallel with IDP; accordingly, levels of trade increase with expansion of cross-border investment flows as economies constantly grow.

However, Svetličič (2003) argues that some countries under the socialist regime have made efforts to leapfrog steps in the IDP model via forceful government interventions in international cooperation. Furthermore, deviating from the conventional electric paradigm theory and specifically the IDP model, Hansen (2009) introduces a theoretical proposition by utilizing India as an anomaly. The hypothesis posits that during the initial stages of development, the imposition of stringent policies on Indian Inward Foreign Direct Investment (IFDI) alongside the establishment of non-equity linkages with multinational corporations impedes the capacity of domestic enterprises to initiate and proactively engage in overseas investment through Outward Foreign Direct Investment (OFDI). Additionally, it suggests that the promotion of OFDI can, in turn, catalyze the augmentation of domestic industries through the enhancement of IFDI. In empirical aspects, most studies show the universality of the IDP model. However, in the  application of the IDP model, selected specific characteristics of economic structures, monetary policies and especially governance institutions should be taken into account to provide a comprehensive analysis on OFDI’s features in each country (Bellak & Jel-code, 2001; Liu et al., 2005; Stoian, 2013). In essence, the empirical literature links economic development with OFDI in the IDP model with large considerations on good governance. When a developing economy grows at significant rates, several aspects of good governance, such as its abilities on control of corruption tend to be improved  (Paldam, 2003). In addition, the state would be under increasing civilian pressure  (Acemoglu & Robinson, 2000), or the economy can face vital conditions such as better education (Fukuyama, 2001) and a rising middle class (Pittaluga et al., 2015) to undertake institutional reforms towards more transparency and democracy. Therefore, economic development becomes a requisite for good governance in terms of improving democratic and transparent institutions (Boix & Stokes, 2003). These good governance conditions establish crucial fundamentals in appealing IFDI as well as nurturing domestic enterprises to become MNCs for OFDI. For the former, it is confirmed that MNCs are interested in foreign countries with good government infrastructure during their location choice (Globerman & Shapiro, 2002). For the latter, Das (2013), while studying developing countries’ behaviours on OFDI, shows that good governance in home economies plays an indispensable role in encouraging these flows. However, weak governance in source countries might be a thrust for MNCs to escape bad institutions and invest abroad (Stoian & Mohr, 2016).

In a firm level, foreign investors can be divided into four categories based on their motives: natural resource seekers; market seekers, efficiency seekers, or strategic asset seekers (Dunning, 1977; Dunning & Lundan, 2008). This is also a key topic in various empirical studies assessing host countries’ features and determining OFDI flows from developing and emerging economies. This is because MNC’s motives and strategies are diverse across countries, depending on host countries’ barriers such as corruption and ease of doing business, and MNCs’ own capacity (Kravtsova, 2010).

For the first type, MNCs tend to select locations with ample natural resources and cheap labour to match their technical and administrative efficacy. Studies on China’s OFDI flows show that this country is a typical case of resource-seeking motives (Hu, 2013; Kolstad & Wiig, 2012; Zhang & Daly, 2011). Weak governance and large values of natural resource exports in host economies are key drivers of its OFDI flows. Wang et al. (2012) affirm that Chinese government has implemented state-ownership interventions in outward investment activities to orient its OFDI capital toward developing countries rather than developed ones. In doing so, China has succeeded in promoting investments in African countries, where there are natural resource advantages and limited governance (Mourao, 2018) .

Market seekers invest in various countries with their desire to expand market shares abroad by providing products to host economies’ domestic markets. This market entry usually comes after an increase in export to the host countries. Empirical evidence of Eem et al. (2019) and Goh & Wong (2011) reveals that Malaysian MNCs employ market-seeking motives, as host countries’ market sizes and income levels positively affect Malaysian’ OFDI flows. In addition, Ibrahim et al. (2019) vividly indicate that selected ASEAN countries, such as Malaysia, Singapore, Indonesia, and Thailand boost their OFDI flows via the market-seeking motives.

The efficiency seeking OFDI aims to rationalize firms’ operation by technology transferred to host countries and taking advantages of their cheap labour conditions. These enterprises employ economy-to-scale features and a natural monopoly to increase their competitiveness. Among developing countries, Ma et al. (2020) argue that the motives of China's OFDI in the period 2005-2016 in ASEAN countries shifted from resource-seeking to efficiency-seeking.

Beside a traditional exploitative view based on the three aforementioned motives, Kim et al. (2015) investigate a novel theoretical concept, in which businesses can access an explorative view to conduce to their growth via seeking new knowledge and favourable conditions in host countries. Regarding the theory of  Dunning & Lundan (2008), the approach of Kim et al. corresponds to the forth motive-strategic asset seeking. In host countries, strategic asset seekers use ownership methods that are ineffective or infeasible in their source country, generate gains from adjusted ownership, and improve their positions in international markets by weakening competitors. An analysis of Indian OFDI reveals that its MNCs had a tendency to implement market entries through strategic asset seeking, and largely invested in non-tradable high-tech services such as software and telecommunications in developed countries, which resulted in a significant increase of its OFDI flows from 2001 (Hansen, 2009). This strategy was reinforced by Indian high-tech workforce migrating to developed countries, promoting strategic OFDI from India into those industries and in turn boosting the Indian economy (Khadria, 2002).

Theoretical and empirical literature review reveals that countries design different investment strategies: Although the IDP model is basically universal across countries, specific investigations should be carried out when governments implement intervention policies in developing countries as well as countries under the socialist regime in transition. Vietnam’s outward FDI, to a certain extent, follows the IDP model. However, the OFDI’s motives have not been thoroughly investigated. To the knowledge of this research’s author, only two articles by Phung (2016) and Thanh Thu et al. (2018) empirically analyse the determinants of Vietnamese OFDI in its two neighbours, Laos and Cambodia, in which Vietnam's OFDI activities are positively influenced by the State interventions as well as Vietnamese relative economic development levels compared to Laos and Cambodia. Such studies cannot fully reflect Vietnam's OFDI motives and strategies in its development path.

For investigating the motivations of Vietnam’s OFDI, we propose three main hypotheses, divided into two strategy groups: location selection and the amount of new registered OFDI capital.

For Vietnamese location selection for OFDI flows, we proposed one hypothesis:

H1: Political stability in one country has positive impacts on Vietnam's selection of investment locations.

For new registered OFDI amount, once this country is selected for OFDI flows, we proposed two hypotheses:

H2: Natural resource advantages have positive impacts on the amount of registered OFDI.

H3: Good governance in host middle income country groups has negative impacts on the new registered OFDI amount.

We follow the gravity model of Zhang & Daly (2011) to test the aforementioned hypotheses. As this paper is the first quantitative research, investigating the determinants of Vietnam's OFDI strategies, the coefficients of the controls in the gravity model like population, geographical distance, GDP per capita are also analysed in terms of statistical significance levels to investigate associations of these factors with Vietnamese OFDI.

3. Methodology

The Vietnamese OFDI strategies will be investigated in this research via the construction of a panel data regarding all countries and territories during the 2007-2021 period. The main reason for selecting the year 2007 as the beginning of our time span is the fact that this year marks Vietnam’s joining of the WTO. It can be said that the year 2007 is a unique step in Vietnam’s globalization and economic openness. We follow the empirical gravity models of Kolstad and Wiig (2012) as well as Zhang and Daly (2011), and add other independent variables on good governance to test our hypotheses:

where i is an index by country or territory, ε is an independent and identically distributed random variable.

The calculation and data sources of variables: OFDI, Governance, Natural resources and Controls are described in Table 1. We calculate the average value of these indicators in the period 2007-2021 to produce long-term variables (Zhang & Daly, 2011). In our regression models, we use natural logarithmic numbers, as suggested by Goh & Wong (2011) to be able to compare their impacts and elasticities in relative terms, which cannot be investigated via the original model of Zhang & Daly (2011).

There are four clarifications on our data and methodology for calculating the variables. Firstly, we use three data sources to calculate the Controls: Centre d'Études Prospectives et d'Informations Internationales (CEPII) for the distance variable, World Integrated Trade Solution (WITS) of the WB for the variables on imports and exports, and World Development Indicators (WDI) for the other Controls. Secondly, we calculate our dependent variable, OFDI, as the annual average of new-registered OFDI capital from Vietnam to each country and territory in the world. Data on the new-registered OFDI capital is extracted from FIA data set, which records new OFDI registration licenses during the period of 2007-2021. The measure of new registered capital is employed to reflect the Vietnamese MNCs’ expectations at the starting points of planning their investments abroad. In the period of 2007-2021, Vietnam has registered to invest in 59 countries and five territories (Macao, Hong Kong, Taiwan, British Virgin Islands (BVI), and Camay Islands). Thirdly, for governance variables, we use the WB’s Database on Good Governance to measure 6 sub-indicators: Control of Corruption (COC), Governance Effectiveness (GEF), Political Stability (POS), Rule of Law (ROL), Regulatory Quality (RQU), and Voice and Accountability (VOA) (Kaufmann et al., 2010). Finally, we calculate natural logarithms of (1+ GDPG) and (1+ |Inflation|) as proxies for GDP growth and inflation to run our regressions. A one percentage point increase in (1+ GDPG) corresponds to an increase of one percent in GDP growth (GDPG). In addition, the absolute value of CPI is employed to test a hypothesis that the more stable CPI one country maintains (an inflation rate is near zero), the more motives Vietnamese firms have to invest via OFDI.

We use two regression techniques to test two hypothesis groups: probit for the determinants of location choice and Fully Modified Ordinary Least Squares (FMOLS) for the determinants of the OFDI amount. In both regression techniques, we use robust standard errors to avoid heteroskedasticity, which leads regression techniques to become inefficient (White, 1980).

For the determinants of location choice, the panel data is constructed in all countries and territories with available data. WITS records countries and territories with whom Vietnam trades in significant amounts of exports and imports; therefore, we exclude countries which are not included in the WITS data set due to uncertainties in Vietnamese export and import values (see Table 2 for the variables’ description)[2]. In addition, we use observations with missing no figures in all of the variables in Table 1 to run probit regressions. The OFDI variable in probit regressions has binary values of 0 if there are no OFDI flows from Vietnam to an economy in the period of 2007-2021, and 1, reversely.

To examine the factors that impact the OFDI amount, we employ the FMOLS estimate to regress Ln (OFDI) with independent variables. In this model, only observations in Vietnamese host countries-or OFDI =1 in probit regressions-are taken into account. Correlation matrix of the variables for this model is presented in Table 3. It can be seen that good governance variables are strongly correlated with each other.

4. Results and Discussion

Regarding the determinants of location choice, the probit technique is employed with regressing the binary OFDI variable with Controls, Natural resources, and Governance. Results of six regressions are presented in Table 4.

Among Controls, coefficients of three variables, inflation rate, distance, and EX, are statistically significant at different levels of 1%, 5%, and 10% in all 6 columns. At a 1% significance level, all 6 regressions show that geographical distance to Vietnam negatively impacts the probability of Vietnam’s choice of the country to become its investment location. Meanwhile, export value from Vietnam to one country has positive association with the probability of Vietnamese investment location, with the coefficients ranging from 0.41 to 0.45 at significance levels of 5% in the two columns (4.1) and (4.3) and 1% in the four columns (4.2), (4.4), (4.5) and (4.6). These results support the correlations between IDP and TDP models in Vietnam. In addition, countries with inflation rates farther than zero percent are more likely to be used for investment by Vietnamese enterprises. In other words, Vietnam tends to invest in countries with relatively large inflation rates. Finally, a coefficient of the population variable, POP, in column (4.3) is statistically significant at a 5% significance level with a magnitude of 0.0658. Population also has a positive effect on Vietnamese OFDI’s location choice. However, this effect cannot be confirmed by other results in the remaining five columns of Table 4[3].

Regarding Governance and Natural resources, only a coefficient of POS in column (4.3) is statistically significant at a 1% level with its magnitude of 0.176. Political stability positively influences the probability of Vietnamese location choice. In addition, it cannot be confirmed that natural resource advantages are a factor that induces Vietnamese firms to choose their investment locations. Similarly, the five other sub-indicators of Governance, control of corruption (COC), governance effectiveness (GEF), regulatory quality (RQU), rule of law (ROL), and voice and accountability (VOA) are not significantly associated with Vietnamese OFDI’s location choice.

To compare impacts’ magnitudes among variables’ coefficients with statistical significance at a 10% level or lower, the two variables POS and DIST have the largest effects on the location choice. In other words, political stability and distance are the most important factors in selecting investment location in the period of 2007-2021. The possibility of being selected as Vietnamese OFDI’s host countries would be large if their political systems remain stable with their close geographical distance to Vietnam. Interestingly, the coefficients of the inflation rate variable are positive and statistically significant, but lower than ones of POS. This means that Vietnamese’s enterprises give priorities to stable political systems over stable economic ones, at least in terms of inflation, when choosing a location to invest abroad.

Regarding the determinants of new registered OFDI amount, Table 5 shows results of regressing Vietnam’s new registered OFDI capital to host economies with dependent variables Controls, Natural resources, and Governance. While geographical distance and political stability play important roles in the location selection model, we cannot confirm that these factors have impacts on the new registered OFDI amount during the period 2007-2021.

Among Controls, population and export value have the strongest associations with the amount of new registered OFDI capital. Coefficients of POP are around -1.1 and statistically significant at a 5% level or lower in all six columns of table 5. If a host country’s population increases by 1%, the amount of the new registered capital decreases by 1.1 percentage points. In addition, coefficients of GDP per capita, GDPP, are negative and statistically significant at a 5% level in two regressions (5.3) and (5.6). Therefore, Vietnam's OFDI strategies are different from Malaysian ones, since Vietnam does not seem to invest via market-seeking motives. Instead, Vietnam tends to increase the amount of the OFDI capital in host countries with lower degrees of both population sizes and GDP per capita. This investigation partially corresponds to our observations that Vietnam invests a significant amount of capital in low middle-income economies.

In addition, export value is positively associated with the amount of the OFDI capital, with coefficients proximately 1.0 and statistically significant at a 5% level or lower in all six regressions. If the export value to a host country increases by 1%, the amount of Vietnamese new registered OFDI capital in that country increases by approximately 0.93-1.13 percentage points.

The OFDI capital amount is largely affected by a host country’s natural resource potential. Coefficients of Natural resources, NR, are ranging between 0.78 and 0.88 at 1% and 5% significance levels in all six columns, as illustrated in Table 5. If a host country increases the ratio of ores and metals exports to merchandise exports by 1%, the amount of Vietnamese OFDI capital increases by 0.78-0.88 percentage point. Therefore, Vietnam's OFDI flows hold resource-seeking motives.

While coefficients of selected Governance sub-indicators, COC, POS, and VOA, are not statistically significant, coefficients of the other Governance ones, GEF, RQU, and ROL, are all negative and statistically significant at a 5% level. If a host country’s good governance sub-indicators GEF, RQU, and ROL decrease by 1%, there are increases of 1.93, 2.11, and 1.80 percentage points, respectively in the amount of Vietnamese new registered OFDI capital in that country. Compared to each other in a group of Governance, the regulatory quality, RQU, has the largest negative impact on the OFDI amount. Vietnamese strategies on the OFDI amount are designed to increase the capital into host countries with weak governance, especially low regulatory quality; however, these countries must ensure high political stability, as shown in the location choice model.

In comparison with impacts’ magnitudes among all of the independent factors, three governance variables, GEF, RQU, and ROL also have the strongest impacts on Vietnamese OFDI amount with negative coefficients, following by export value with positive coefficients, and population size with negative ones.

5. Conclusion

Since the implementation of Renovations or “Doi Moi” in 1986, the Vietnamese economy has grown considerably and started to participate in OFDI process, which has been promoted since the 2000s. This process follows the IDP model with its FDI flows larger than its OFDI ones. This study investigated the factors affecting Vietnam’s OFDI strategies on location selection and the amount of new registered OFDI capital in host countries. The data were based on panel data of 159 countries and territories in the period 2007-2021 with a probit technique for investigating the determinants of the location selection and the data of 55 host economies with FMOLS technique for investigating ones of the new registered OFDI amount.

Trade relations, measured by export values, are positively associated with both probabilities that Vietnam chooses investment locations and the new registered OFDI amount in host economies. These findings support a hypothesis on TDP and IDP, since trade and investment values have positive correlation with each other and correspond to the level of economic development.

In terms of location selection, proximity to Vietnam and high levels of political stability are the two most important determinants of probability for being chosen as a host economy. Priorities are given to political stability over macroeconomic stability, at least in terms of inflation. In addition, the potential on natural resources cannot be confirmed to have significant effects on the choice of investment locations.

Regarding the choice of the new registered OFDI amount, high levels of political stability in host countries no longer affect this variable. Instead, other factors of good governance, including government effectiveness, regulatory quality, and rule of law have negative impacts on this variable. If degrees of variables GEF, RQU, and ROL decrease by 1%, the OFDI value in host countries increases by 1.93, 2.11, and 1.80 percentage points, respectively. Furthermore, natural resource advantages are a crucial factor that positively impacts the value of new registered OFDI flows. If the ratio of ores and metals exports to merchandise exports in host countries increases by 1%, there is an increase of 0.78-0.88 percentage point in the Vietnamese new registered OFDI value. Comparing impact magnitudes of all dependent variables, Controls, Natural resources, and sub-indicators in Governance, low levels of regulatory quality in host countries have the strongest effects on the OFDI amount.

Our empirical analysis contributes to the understanding of OFDI motives in developing countries in IDP initial stages. Before reaching a higher level of economic development like other countries in ASEAN, such as Singapore, Malaysia, Thailand and Indonesia, OFDI from Vietnam aimed at resource seeking rather than market seeking. In these first stages, Vietnamese strategies are similar to Chinese strategies in the early 2000s. However, with its rapid economic development, China has shifted its strategies on the OFDI motives to market-seeking in selected regions (Ma et al., 2020). This also implies research topics and hypotheses for future studies on Vietnam's OFDI motives, as Vietnam climbs to higher stages of economic development.

[1]. logarithms of these figures to compare the results in regression models in relative terms.

[2]. The remaining five economies in which Vietnam invested included Camay Islands, BVI, Hong Kong, Taiwan, and Macao. See more at Appendix 1.

[3]. OFDI = 1 if a country is chosen for Vietnam’s OFDI during 2007-2021; otherwise, OFDI = 0. Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. See List of 159 observations in Appendix 1.

[4]. Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1.