Multinational Enterprises and Local Firms’ Export Market Entry: A Panel Data Analysis of Vietnam’s Food Processing Industry

Dao Thi Hong Nguyen

Faculty of Economics, Nha Trang University,

2 Nguyen Dinh Chieu Street, Nha Trang, Vietnam

Email: daonth@ntu.edu.vn

Ari Kokko

Copenhagen Business School, Department of International Economics, Government and Business, Porcelænshaven 24A, 2000 Frederiksberg, Denmark

Email: ako.egb@cbs.dk

Thong Tien Nguyen

Copenhagen Business School, Department of International Economics, Government and Business, Porcelænshaven 24A, 2000 Frederiksberg, Denmark

Email: ttin.egb@cbs.dk

Abstract: This paper develops new insights into the spillover effects of multinational enterprises (MNEs) on local firms’ export market entry, using the case of Vietnam – a notable global manufacturing hub located in Southeast Asia. The study utilizes a disaggregated and more recent firm-level panel dataset of the food processing industry. This represents an essential part of Vietnam’s thriving manufacturing sector, with enormous potential for exports. The random-effects Probit estimation results reveal that the presence of MNEs significantly boosts the chance for local counterparts, notably the privately owned group, to become exporters and thereby start integrating into global value chains. Further regressions suggest that in evaluating MNEs-linked spillovers, measurement of foreign presence and local firms’ export experience matter. The findings imply the considerable potential for MNEs to make a difference in local firms’ export prospects and validate the policy efforts to attract FDI inflows to the food processing industry.

Keywords: Multinational enterprises; Spillovers; Export market entry; Food processing industry; Vietnam

JEL codes: F14, F23, L66

Acknowledgements: This study is part of the project “Innovating Vietnam’s TVET system for sustainable growth (Vietskill)” funded by Danida Fellowship Centre, Denmark. We are grateful for this financial assistance. All remaining errors are the authors’ sole responsibility.

 

1. Introduction

In small open economies, it is vital that many firms can find their way to the export market and progressively integrate into global value chains (Broocks & Van Biesebroeck, 2017). There is a growing body of literature exploring the determinants of export behaviour at the firm level, which highlights the importance of firms’ innate characteristics (e.g., size, age, ownership, capital intensity, and labour quality) (Bernard & Jensen, 2004; Haddoud et al., 2021; Kneller & Pisu, 2006). Firm-level variables tend to capture the proven capability and inherent competitiveness of individual firms, enabling them to overcome entry barriers (which are often in the form of sunk costs) to start serving the foreign market. We know much about firm attributes that can explain their export decisions, but surprisingly little about the impact of external forces, including the rising presence of multinational enterprises (MNEs) in the host economies. As MNEs have played an increasingly important role in global value chains (Giroud & Mirza, 2015; Scoppola, 2021), they can serve as catalysts and help establish contacts with the global economy to facilitate the export entry of local firms.

MNEs are widely perceived to possess superior advantages and perform better than their local counterparts, notably in transition and developing countries. Their presence with advantageous assets (e.g., technological know-how, managerial and human resource expertise, marketing techniques and export experience) represents a powerful source of knowledge that can spill over and affect local firms’ various performance aspects (Anwar & Sun, 2014; Javorcik, 2004; Nguyen, 2021). Such spillovers can occur over the course of co-location and interactions among firms via major channels of industrial linkages, interpersonal contacts and labour mobility, competition effects, information, demonstration and imitation effects (for a detailed discussion, see Balsvik (2011); Blomström and Kokko (1998); Villar et al. (2020)). While MNEs’ impact on local firms’ productivity has been exhaustively examined, scant evidence has been marshalled so far on their effect on local peers’ exporting or so-called export spillovers, particularly in the case of Vietnam.  

This study contributes new insights into the role of MNEs as a determinant of local firms’ export market entry, using data from the food processing industry in Vietnam. Since the launch of an economic reform known as Doi moi in 1986, Vietnam’s economic performance has been remarkable, and it has been identified as one of the world’s fastest-growing economies over the past three decades (Nguyen, 2020; WB, 2020). Accelerating the integration into the regional and global economy has had a prominent position in the country’s policy agenda, and has resulted in a series of agreements on trade and investment liberalization (ADB, 2019). With strong economic fundamentals and a favourable investment environment, Vietnam has become a notable global manufacturing hub and a preferred destination for MNEs. Foreign direct investment (FDI) inflows have gradually increased from a few hundred million USD per year in the early 1990s to USD 31 billion in 2020 (GSO, 2018). Despite the adverse impacts of the Covid-19 pandemic, foreign investment in 2021 maintained strong growth momentum, recording a 9.2 percent increase in registered capital compared to the previous year.   

This paper utilizes a rich firm-level panel dataset focusing on a significant part of Vietnam’s thriving manufacturing sector, namely food processing. Vietnam has strong comparative advantages in the food processing industry with a favourable climate and a diverse range of products (e.g., rice, vegetables, coffee, pepper, cashew, tropical fruits and seafood). Although Vietnam’s agricultural, forestry, and aquatic product exports have increased over time, the proportion of deeply processed products with higher value added remains modest. Our data shows that only 13% of domestic food processing firms were exporting in 2011-2016, while the corresponding share among the foreign MNEs in the industry was over 75%. Thus, exploring whether spillovers from more export-oriented foreign MNEs could raise the export participation among local firms is important for capitalizing on the country’s competitive advantages and upgrading its position in global value chains.  

Compared to most previous studies relying on large data sets covering the entire manufacturing sector or economy, our use of more disaggregated data for an emerging sub-sector presents several advantages. It may provide deeper and more relevant insights into MNE-induced spillover effects on local firms at the same time as it reduces potential problems related to aggregation bias. Furthermore, unlike most earlier contributions to the export spillover literature, it allows us to shed light on the measurement of MNE presence in evaluating their impacts on local firms’ export market entry. We use a set of four proxies to capture the importance of foreign presence: the output share, employment share, export share and asset share of MNEs across the industry-region-year dimensions. In addition, we explore the role of domestic firms’ export experience in moderating the impact from foreign MNE presence, which is rarely examined in extant literature.

The remainder of the paper proceeds as follows. Section 2 reviews the related literature on MNEs and exports by domestic firms. Section 3 describes the data and empirical modelling. Section 4 presents the estimation results and Section 5 summarizes and provides some concluding remarks.

2. MNEs and exports by local firms: A brief review of the literature

The existing literature on MNEs and local firms’ exports is relatively scarce compared to the literature on productivity spillovers. Additionally, previous findings are rather mixed, which to some extent could be attributed to heterogeneity in country settings, data used, and measurement of foreign presence. Aitken et al. (1997) provided one of the first quantitative analyses of the role of MNEs as catalysts for domestic firms’ exports. They examined the Mexican manufacturing industry, using panel data of 2,104 firms during the period 1986–1990, and found that the probability that domestic firms export was positively correlated with proximity to foreign MNEs. At the same time, they found no correlation found between the firms’ export decisions and the concentration of exporters in the industry. Therefore, they suggested that the government should encourage prospective exporters to locate near foreign firms (by creating export processing zones), which was expected to reduce the costs of foreign market entry and raise the export propensity of domestic firms.

Kokko et al. (2001) investigated export spillovers from MNEs to domestic Uruguayan manufacturing firms, using a cross-sectional dataset of 1,243 firms in 1998. They found evidence of positive export spillovers from foreign-owned to domestic firms in Probit estimations. Their study also explored the spillover effects by splitting foreign firms based on the period in which they were established in Uruguay. The results suggested that only foreign firms established after 1973 (i.e., the outward-oriented period of Uruguay’s policy regime) had a positive effect on the export participation of domestic firms. They also explored whether the geographical destination of exports mattered. They found stronger evidence of export spillovers for destinations outside of Uruguay’s neighbouring markets (Brazil and Argentina). The proposed reason was that foreign MNEs did not have any particular information advantages with respect to Uruguay’s neighbours, where low transaction costs and preferential institutional arrangements instead contributed to relatively strong export performance among local firms.

Greenaway et al. (2004) focused on UK manufacturing firms during the period 1992–1996, and used three measures of foreign presence to capture the impact of foreign MNEs on local firms: the MNEs’ expenditure on R&D, their employment share, and their export share. The empirical results suggested that export spillovers were present, with increased competition from MNEs as the most important channel. Furthermore, information externalities were found to affect firms’ export participation decisions but not their export intensity. Kneller and Pisu (2007) further explored export spillovers for UK firms by incorporating both horizontal and vertical spillovers. Their findings suggested that domestic firms’ export participation decisions were not affected by contacts with MNEs, but that their export intensity appeared to be influenced by the MNEs’ presence in upstream and downstream industries.

Several studies have focused on Asia. Ma (2006) explored the role of MNEs in determining Chinese manufacturing exports. Using comprehensive firm-level data for the 1993-2000 period, the analysis suggested that foreign firms from OECD countries had a positive influence on local firms’ decisions to export, but the activities of overseas-based Chinese firms did not have any corresponding effects. Sun (2009) also examined the case of China and extended the analysis by including interaction terms between foreign presence and firm-specific characteristics to search for determinants of export spillovers. The findings indicated positive and significant FDI export spillover effects on the export intensity of local firms. In addition, the scale of spillovers was positively correlated with a geographical location in central China, but negatively correlated with firm-level variables such as the ratio of costs to revenue, the ownership structure, and a location in western China.

While a majority of studies seem to find positive export spillovers from foreign presence, there are also analyses where no or even negative export spillovers are found. Barrios et al. (2001) investigated the export behaviour of Spanish manufacturing firms during the period 1990–1998. Foreign presence was proxied by the export activity and the R&D expenditures of MNEs at the industry level. Results from Probit and Tobit estimations indicated that a firm’s own R&D activity was a significant determinant of exports. However, there was no evidence that indigenous firms would have benefited from export spillovers from MNEs.

Similarly, Phillips and Ahmadi-Esfahani (2010) examined export spillovers from MNEs to domestic food manufacturing firms in Australia, using a cross-sectional dataset for the year 2005. Their Probit estimation results did not find any support for the hypothesis that MNE export concentration would raise the probability that domestic firms start exporting. They also concluded that the export decisions of local firms were unaffected by the overall presence of foreign firms at the national level. A possible explanation for the lack of export and competition effects from MNEs was the possibility that foreign firms may have chosen to operate in isolation from domestic firms, thus protecting their firm-specific assets.

Ruane and Sutherland (2005) studied export spillovers from foreign MNEs to domestic manufacturers in Ireland, which could be characterized as a third-country export platform. Their analysis suggested that domestic firms’ decisions to enter the export market were positively correlated with the foreign presence in the relevant subsector. However, both export participation decisions and the export intensity of domestic firms were negatively associated with the export intensity of foreign firms. These negative spillovers were believed to be linked to the dominance and extremely high export intensity of US-owned firms in the traditional export sectors.

Regarding the case of Vietnam, there is scant and inconclusive evidence on MNEs-linked spillovers on exports by local firms. Anwar and Nguyen (2011) and Nguyen and Sun (2012) both examined the role of MNEs as a determinant of local manufacturing exports, using data for 2000 and 2003-2004, respectively. While both studies found some positive effects from foreign to domestic firms, the export spillovers were heterogeneous and conditional on firm- and industry-specific characteristics such as firm age, firm size, ownership structure, labour quality, industry technology and competition levels. A more recent contribution by Ha et al. (2020), using a large panel dataset covering both manufacturing and services during the period 2010-2015, found mixed results for different spillover channels. Specifically, the spillovers through backward linkages were found to be positive, but effects through forward linkages were strongly negative, while horizontal export spillovers were insignificant.

3. Data and empirical modelling
   • Data

This study utilizes a panel data set of firms in the food processing industry at the three-digit level of Vietnam’s Standard Industrial Classification (VSIC 2007). The data were obtained from the General Statistics Office (GSO), which commissions nationwide enterprise surveys covering all business entities in the state, private and foreign sectors. This data set is the most comprehensive and reliable firm-level database for empirical studies on the Vietnamese economy. After screening the data for systematic missing values and outliers, the constructed final sample is a six-year unbalanced panel (2011-2016), comprising 25,032 firm-year observations. All monetary variables are converted to constant prices using the consumer price index. Statistical software package Stata 17 is employed for data management and analysis.

Table 1 displays the distribution of firms by ownership across the eight three-digit food processing industries covered by the analysis. Looking first at the domestic Vietnamese actors, the private sector was dominant in terms of the number of firms across all industries, accounting for nearly three-quarters of all firms. Industries C107 (Manufacture of other food products), C106 (Manufacture of grain mill products, starches and starch products) and C102 (Processing and preserving of fish, crustaceans and molluscs) accounted for most of both state-owned and private. Together, these three major industries comprised roughly 50% of all domestic firms in the Vietnamese food processing sector. In addition to these three sub-sectors, industry C103 (processing and preserving of fruit and vegetables) and C108 (manufacture of prepared animal, fish, poultry feeds) attracted many foreign enterprises. In terms of relative importance, foreign MNEs held particularly large shares in C108 (prepared animal, fish, and poultry feeds) and C104 (manufacture of vegetable and animal oils and fats), representing 12.3% and 11.2% of all the firms in these two sub-sectors. As far as the output share of MNEs is concerned, their relative contributions were much more sizable across all industries, reaching up to over 70% in C108. On average, MNEs made up 31.4% of the total output in the examined industry.    

Table 1. Firm distribution across ownership types and three-digit industries

VSCI Code	Three-digit food processing industry	Local firms		Number of MNEs	Share of MNEs	Output share of MNEs
		State	Private
C101	Processing and preserving of meat	315	880	68	0.054	0.150
C102	Processing and preserving of fish, crustaceans and molluscs	1,265	3,250	175	0.037	0.060
C103	Processing and preserving of fruit and vegetables	716	2,307	179	0.056	0.178
C104	Manufacture of vegetable and animal oils and fats	105	259	46	0.112	0.495
C105	Manufacture of dairy products	212	496	41	0.055	0.411
C106	Manufacture of grain mill products, starches and starch products	1,590	3,343	76	0.015	0.060
C107	Manufacture of other food products*	1,976	5,814	648	0.077	0.454
C108	Manufacture of prepared animal, fish, poultry feeds	682	1,822	350	0.123	0.701
	Total	6,861	18,171	1,583	0.059	0.314

* Including bakery products; sugar; cocoa, chocolate and sugar confectionery; macaroni, noodles, couscous and similar farinaceous products; prepared meals and dishes.

 

Table 2 shows the distribution of firms by ownership and region. Based on geographical and socio-economic conditions, Vietnam is divided into six regions. Of these, Southeast (including Ho Chi Minh City) is the country’s economic hub and Mekong River Delta is highly productive in agriculture and aquaculture. Hence, it is no surprise that these two regions represented the largest number of food processing firms across all ownership types. Together, they made up around 64% and 60% of domestic state-owned and private firms in the industry, respectively. Likewise, MNEs were largely concentrated in the Southeast, accounting for over 55% of all foreign firms. The Red River Delta and Mekong River Delta also attracted a significant number of MNEs, comprising around 13-15% of foreign firms in all regions. The relative share of MNEs in terms of firm number is rather modest, showing the highest proportion in Central Highlands (11.9%). Meanwhile, their output share was more substantial with a regional average of around 32%.

Table 2. Firm distribution across ownership types and regions

Region	Local firms		Number of MNEs	Share of MNEs	Output share of MNEs
	State	Private
Red River Delta	1,286	3,723	210	0.040	0.396
Northern Midland and Mountain	307	987	54	0.040	0.419
Central Coast	757	2,030	102	0.035	0.221
Central Highlands	154	560	96	0.119	0.381
Southeast	2,094	5,884	875	0.099	0.324
Mekong River Delta	2,263	4,987	246	0.033	0.165
Regional average	6,861	18,171	1,583	0.059	0.318

Table 3 presents descriptive statistics on the share of exporting firms in the eight three-digit food processing industries. It can be seen that there are local exporters in all industries, although their share of the total number of firms in the industry varies. Domestic exporters are mainly concentrated in the top three sub-sectors, namely C102, C107 and C103, which jointly account for nearly three-quarters of all domestic exporters in the food processing industry. The share of exporters is much larger among foreign firms: on average, three out of four foreign firms are engaged in exporting.  

Table 3. Export participation of local firms and MNEs by three-digit industries

VSCI Code	Three-digit food processing industry	Local firms		MNEs
		Number of exporters	Share of exporters	Number of exporters	Share of exporters
C101	Processing and preserving of meat	73	0.061	41	0.612
C102	Processing and preserving of fish, crustaceans and molluscs	1,024	0.227	137	0.787
C103	Processing and preserving of fruit and vegetables	555	0.184	148	0.827
C104	Manufacture of vegetable and animal oils and fats	54	0.148	39	0.848
C105	Manufacture of dairy products	65	0.092	34	0.850
C106	Manufacture of grain mill products, starches and starch products	348	0.071	54	0.761
C107	Manufacture of other food products*	853	0.110	460	0.714
C108	Manufacture of prepared animal, fish, poultry feeds	320	0.128	270	0.778
	Total	3,291	0.131	1,183	0.754

* Including bakery products; sugar; cocoa, chocolate and sugar confectionery; macaroni, noodles, couscous and similar farinaceous products; prepared meals and dishes.

• Empirical modelling

Following previous studies, a baseline model is set up in Equation (1) below to examine the determinants of local firms’ export market entry, with a focus on the role of foreign MNEs (Aitken et al., 1997; David Greenaway et al., 2004; Kokko et al., 2001; Phillips & Ahmadi‐Esfahani, 2010). This general specification takes into account the impact of the presence of foreign firms ( ) and a set of other variables that can potentially induce local firms to start exporting. It should be noted that the dependent variable is constrained to local firms only. This separation can help capture the influence of MNEs as well as eliminate possible bias due to higher export participation among MNEs in the industry. (1)

The variable proxying local firms’ export market entry (  is a dichotomous variable taking the value of 1 if local firm i in three-digit industry j exports at time t, and 0 otherwise. The explanatory variables include foreign presence (MNE_kjt), domestic firms’ innate characteristics ( ), sub-industry-specific features ( ), and other control variables captured by regional dummies (RD_j) and year dummies (YD_t), as well as an error term ( ). The subscripts i, k, j, and t denote variations across firm, three-digit industry, region and time, respectively.

The presence of foreign firms (MNE_kjt) is the focal variable in the model, intended to capture the spillover effects of MNEs on local firms’ exports in the examined industry. Based on the GSO’s enterprise surveys, MNEs in this data set are defined as foreign affiliates with full foreign ownership or joint venture participation. The importance of MNEs is proxied by the log of their output share in the specific industry-region-year context, as summarized in equation (2) below.

(2)

where s is the total sales of firm i in three-digit food processing industry k in region j at time t, and Z and Y denote the subsets of MNEs and local firms, respectively. This is a widely adopted measure of MNE presence, intended to capture intra-industry effects in the existing literature (Aitken et al., 1997; Ha et al., 2020; Nguyen & Sun, 2012). The construction of this key variable as in equation (2) allows variations in three dimensions (k, j, t) while most previous papers only control for differences across industry (k) and/or time (t). The additional layer of measurement enables us to better capture the heterogeneous impacts of MNEs on local firms. After the first round of estimations, we will introduce three alternative proxies of MNEs, namely their employment share, export share and asset share, to explore the robustness of MNEs’ role in explaining local firms’ decision to enter the foreign market.

As underscored by earlier studies, local firms’ inherent characteristics are likely to play a crucial part in determining an individual firm’s capability and competitiveness to become an exporter, which is generally captured by  in Equation (1) and further detailed in Equation (3) below (Aitken et al., 1997; David Greenaway et al., 2004; Kokko et al., 2001; Nguyen & Sun, 2012; Phillips & Ahmadi‐Esfahani, 2010). These firm-level characteristics can be important indicators of domestic firms’ absorptive capacity, which allows them to adapt to, learn from, and compete effectively with the foreign MNEs in the local industry and gain beneficial export spillovers from the latter’s presence.(3)

Export experience ( ) is a dummy variable, taking the value of 1 if local food processing firm i has already engaged in exporting in the past and 0 otherwise. Firm age ( ) denotes the number of years in operation in log form. Firm size ( ) is calculated by total sales in log form. Capital intensity ( ) is the log of the ratio of fixed assets to total employment. Average wage ( ) is measured by the cost of labour (including wages, salaries and bonuses) per employee. Industrial zone ( ) refers to the firm’s location, taking the value of 1 if it is in an industrial/economic zone and 0 otherwise. Ownership structure ( ) takes the value of 1 for privately owned firms and 0 for state ownership.

Existing empirical evidence largely supports the hypothesis that these firm-level underlying attributes will have a significant role in determining the firms’ odds to successfully enter the export market (Aitken et al., 1997; Ha et al., 2020; Kneller & Pisu, 2007; Nguyen & Sun, 2012; Phillips & Ahmadi-Esfahani, 2010; Sun, 2009). Notably, firms with prior export experience have already incurred the fixed costs of market entry, and are therefore more likely to continue exporting in the following year. Compared to young, small, labour-intensive and low-paying firms, older, larger, more capital-intensive and higher-paying firms are perceived to be more financially competent and profitable, with a stronger position in the local market. Therefore, the latter group of firms tends to exhibit proven capability, enabling them to compete in the foreign market. Similarly, firms located in an industrial/economic zone generally benefit from export promoting policies instituted by the local government, encouraging them to reach out to serve overseas markets. Finally, compared to privately owned firms, state-owned firms may have better prospects for becoming exporters due to easier access to funding, export promotion incentives, and the network of diplomatic missions abroad.

Equation (1) also controls for sub-industry-specific features ( ) such as industry-level exports and industry concentration. Of these, industry exporting ( ) is measured as the export share of three-digit industry j in the total food processing industry’s exports. It proxies the importance of each sub-industry in the export structure of the entire food processing industry and accounts for the possibility that a) firms operating in more export-oriented industries are more likely to export because of the presence of stronger comparative advantages, and b) that MNEs may tend to choose to invest in industries with stronger export competitiveness. If these characteristics are not controlled there is a risk that results will be biased because of endogeneity problems (Kneller & Pisu, 2007). Industry concentration ( ) captures the competitive pressures within the local industry and is proxied by the Herfindahl index. This is illustrated in Equation (4), where s_ikjt is the sales of domestic firm i and S_kt is the total sales of three-digit food processing industry k at time t. Fierce competition in the local market might induce firms to expand their presence overseas through exporting. It is also worth noting that a highly concentrated industry structure would give export incentives for both the incumbents and the new industry entrants. The incumbents have motives to export because of their plausible market power (i.e., more domestic sales can only be achieved if the price is reduced) and the new entrants might consider the entry barriers to be high in the local industry.(4)
          To execute the benchmark specification, this study employs a random-effects Probit model as shown in Equation (5). Compared to the alternative pooled Probit, Logit, Poisson, Linear Probability and GMM models, this is a preferred approach for dealing with the fact that the dependent variable, firm export market entry ( ), is constrained and binary in nature and the examined data is an unbalanced panel. Notably, this method allows us to control for unobserved time-invariant firm-specific factors that are correlated with independent variables (Forte & Salomé Moreira, 2018; Greenaway et al., 2007; Kim & Xin, 2021; Wooldridge, 2019). The model is estimated by the maximum-likelihood technique with robust standard errors computed to address potential arbitrary heteroscedasticity. As Equation (5) consists of a large number of explanatory variables, multicollinearity might arise, causing misleading estimates. We address this issue by using correlation coefficients and collinearity measures of main regressors. (5)

4. Estimation results and discussions
   • Main estimation results

Table 4 presents the random-effects Probit estimation results on the impacts of MNEs and other factors on Vietnamese food processing firms’ export market entry. Given the nonlinearity of the dependent variable (EME_ikjt), Table 4 also reports the marginal effects (M.E) calculated at the sample means of the covariates to facilitate the interpretation and comparison of MNEs’ impacts. Model [1] covers all local firms and models [2] and [3] provide further insights with separate estimations for privately owned and state-owned enterprises. The robust standard errors were clustered by firm identity to allow for unspecified serial correlation within firms while assuming independence among them (D. Greenaway et al., 2004; Sun, 2009). Three-digit industry, region and year dummies are included to control for industry, region and time effects. The Wald-test for overall model significance indicates that the coefficients of the regressors are jointly significant at the one percent level across the three regressions. Regarding potential multicollinearity, the reported low values of correlation coefficients (<0.35) and variance inflation factors (<2.0) suggest that severe multicollinearity is not an issue of concern in this study (Alin, 2010; Dormann et al., 2013; Fox, 1991; Morrow-Howell, 1994).

Table 4. Baseline results from the random-effects Probit model

	Regressor	Dependent variable [Local firms' export market entry]
		[1] All firms		[2] Privately owned firms		[3] State-owned firms
		Coef.	M.E	Coef.	M.E	Coef.	M.E
	Export experience  ( )	0.9882*** [0.0438]	0.0738*** [0.0045]	1.3379*** [0.0578]	0.1016*** [0.0066]	0.3191*** [0.0873]	0.0186*** [0.0053]
	Firm age ( )	0.1690*** [0.0309]	0.0126*** [0.0023]	0.1634*** [0.0359]	0.0124*** [0.0027]	0.2682*** [0.0594]	0.0156*** [0.0034]
	Firm size ( )	0.7297*** [0.0319]	0.0545*** [0.0020]	0.7798*** [0.0421]	0.0592*** [0.0024]	0.6188*** [0.0469]	0.0360*** [0.0025]
	Capital intensity ( )	0.1950*** [0.0230]	0.0146*** [0.0016]	0.2299*** [0.0302]	0.0175*** [0.0022]	0.1549*** [0.0350]	0.0090*** [0.0019]
	Average wage ( )	0.2666*** [0.0400]	0.0199*** [0.0030]	0.2324*** [0.0519]	0.0176*** [0.0039]	0.3123*** [0.0666]	0.0182*** [0.0038]
	Industrial zone ( )	0.5871*** [0.0893]	0.0438*** [0.0066]	0.5863*** [0.1070]	0.0445*** [0.0081]	0.6231*** [0.1222]	0.0363*** [0.0071]
	Ownership structure  ( )	0.3185** [0.1530]	0.0238** [0.0114]	______	______	______	______
	Foreign presence  ( )	0.2915*** [0.0932]	0.0218*** [0.0070]	0.3789*** [0.1123]	0.0288*** [0.0085]	0.0229 [0.1249]	0.0013 [0.0073]
	Industry exports ( )	0.0006*** [0.0001]	4E-05*** [6.56E-06]	0.0002** [0.0001]	2E-05** [8.48E-06]	0.0012*** [0.0002]	7E-05*** [0.0000]
	Industry concentration  ( )	-0.3137** [0.1423]	-0.0234** [0.0106]	-0.3607** [0.1668]	-0.0274** [0.0126]	0.0225 [0.1911]	0.0013 [0.0111]
	Industry dummies ( )	Yes		Yes		Yes
	Regional dummies ( )	Yes		Yes		Yes
	Year dummies ( )	Yes		Yes		Yes
	Wald chi2	1617.2100***		1295.3500***		289.1000***
	Log pseudolikelihood	-4604.9540		-3306.5371		-1172.8941
Notes: Marginal effects (M.E) are evaluated at the sample means of the covariates; ***, **and  * denote significance at 1%, 5% and 10% levels, respectively; estimates are efficient for arbitrary heteroscedasticity with computed robust standard errors (R.S.E.) in square brackets.

 

For the full sample in model [1], the Probit estimates show strong evidence that the presence of MNEs in the same industry and geographic location significantly raises the likelihood that local firms become exporters and thereby start integrating into global value chains. The estimated coefficient of the key variable ( ) is positive and statistically significant at the one percent level. The estimated marginal effect suggests that a one percentage point increase in MNE presence leads to a 2.18 percentage point increase in the probability that local firms engage in exporting. This result is mostly consistent with earlier studies in other countries (Aitken et al., 1997; David Greenaway et al., 2004; Kokko et al., 2001). It implies that MNEs diffuse valuable export information or generate competitive pressure that motivates local firms to expand to foreign markets.

This finding is potentially important for Vietnam’s food processing industry given the modest share of exporters among local firms. Earlier studies for Vietnam’s aggregate manufacturing sector have largely failed to identify any significant intra-industry export spillovers from MNEs to domestic firms’ export market entry. It is possible that the differences between earlier findings and our results are explained by the decision to focus on a specific industry, food processing, where the gap between foreign and local technology is smaller than in many other industries, and by the measurement of foreign presence accounting for regional variations. Earlier contributions have shown that the spillover effects are to some extent dependent on the technology gap between foreign and local firms (Imbriani et al., 2014; Kohpaiboon, 2006; Smeets, 2008), and suggested that many types of spillover effects decrease with geographic distance (Aitken & Harrison, 1999; Jaffe et al., 1993; Sun et al., 2011).

Furthermore, the regressions for the sub-samples suggest that the beneficial export spillovers from MNEs only hold for privately owned firms but not the sample of state-owned enterprises. The estimated coefficient of  is positive and significant in model [2] while being insignificant in model [3]. The estimated marginal effect indicates that a one percentage point increase in MNE presence is associated with a 2.88 percentage point increase in the likelihood that local privately owned firms start exporting. The finding that foreign presence does not seem to have any impact on state-owned firms might be attributable to the various privileges related to state ownership, which may, to some extent, insulate them from fierce competition from foreign MNEs, and which could make them less dependent on knowledge flows from MNEs as a source of information about foreign markets. As noted earlier, it is possible that state-owned enterprises have privileged access to funding and support from various public export promotion programs (Anwar & Nguyen, 2011; Nguyen & Sun, 2012; Sun, 2009).

Table 4 also confirms the important role of firm-specific attributes in predicting export decisions. The estimated coefficients and marginal effects of all firm-level variables are positive and significant at either 1% or 5% levels. Notably, compared to new exporters, firms with prior export experience have a higher likelihood of 7.38 percentage points to continue exporting in the following year. This is the largest marginal effect estimated in the model. As expected, well-established and larger firms have a better chance to become exporters. A one percentage point increase in firm age and firm size leads to an increase in their probability to engage in exporting by 1.26 and 5.45 percentage points, respectively. Similarly, capital-intensive and high-paying firms are more capable to serve the foreign market. A one percentage point increase in firms’ capital intensity and average wage raises the likelihood that they will start exporting by 1.46 and 1.99 percentage points, respectively. Being located in an industrial zone is also beneficial as it raises firms’ propensity to enter the export market by 4.38 percentage points. Unexpectedly, compared to state-owned firms, privately owned peers are 2.38 percentage points more likely to participate in global value chains. This highlights the emerging role of private food processing firms in exporting, and suggests that industry dynamics have changed during the past decades: previous studies for Vietnam’s manufacturing sector strongly supported the opposite evidence (Anwar & Nguyen, 2011; Ha et al., 2020; Nguyen & Sun, 2012).   

Regarding three-digit industry characteristics, Table 4 confirms their considerable influence on firms’ export market entry. Unsurprisingly, firms operating in more export-oriented industries are more likely to start exporting as the estimated coefficient of the variable industry exports ( ) is positive and significant at 1% level. Meanwhile, the impact of industry concentration ( ) is more sizable and negative. Accordingly, a one percentage point increase in competitive pressure (i.e., decrease in concentration level) results in a 2.34 percentage point increase in domestic firms’ propensity to start exporting.  This finding for the case of Vietnam’s food processing industry is consistent with some earlier studies suggesting that intense competition in the domestic market might force firms to expand their presence overseas through exporting (Anwar & Nguyen, 2011; Ha et al., 2020; Sun, 2009).

• The role of MNE measurement

The variable of interest in this study is presence of foreign firms (MNE_kjt), which is expected to capture potential spillovers of MNEs on local counterparts’ decisions to become exporters. How this key variable is measured can play a significant role in estimation results. Previous studies on intra-industry spillovers largely rely on one single proxy of MNEs, which is typically their output share in the local industry (Aitken et al., 1997; Ha et al., 2020; Nguyen & Sun, 2012). However, the presence and importance of MNEs can be measured in different ways, which may highlight different channels of spillover transmission to local firms (David Greenaway et al. (2004). It can be hypothesized that the competition effect is primarily captured by the overall presence of MNEs measured by the output or asset share, while the information or knowledge spillover effect may be better reflected by the export share of foreign firms. Similarly, the employment share of MNEs may be a better proxy for spillover effects related to interpersonal contacts and labour mobility. Therefore, to test the robustness of the findings, we use three additional measures of MNE presence (apart from the sales share of MNEs used in Table 4), namely their export share, employment share and asset share in each three-digit food processing industry k in region j at time t.

Table 5 reports the estimation results with the three new proxies for MNE presence. There is high consistency in the signs of the estimated coefficients for all variables, including foreign presence ( ) across three regressions. This affirms the robustness of the main finding of positive export spillovers from MNEs to local food processing firms, regardless of the measurement approach. In terms of the size of the estimated coefficients and marginal effects, the estimations report notably different findings, implying the high measurement sensitivity of MNE spillovers. Of these, model [3] using the asset share yields the largest estimate with a one percentage point increase in foreign presence boosting local firms’ chances to start exporting by 4.62 percentage points. Meanwhile, the spillover magnitude of MNEs using the other two proxies is less sizable. Specifically, a one percentage point increase in MNE presence measured by export share and employment share leads to a corresponding increase of 1.09 and 1.41 percentage points in domestic firms’ propensity to start exporting. Given the largest estimated marginal effects from MNEs’ output and asset shares, this analysis suggests a more profound effect of competitive pressure from MNEs, inducing domestic firms to engage in more efficient production techniques and thereby facilitating entry into the foreign market. While the concentration of MNE export activities and employment also generates beneficial information and interpersonal externalities on export products and markets, these spillover channels seem less pronounced. It is reasonable that export information and interpersonal contacts by MNEs might not be readily available and accessible to all local food processing firms.  

Table 5. Estimation results – The role of MNE measurement

	Regressor	Dependent variable [Local firms' export market entry]
		[1]		[2]		[3]
		Coef.	M.E	Coef.	M.E	Coef.	M.E
	Export experience ( )	1.0012*** [0.0452]	0.0745*** [0.0046]	1.0028*** [0.0438]	0.0730*** [0.0044]	0.9882*** [0.0438]	0.0738*** [0.0045]
	Firm age ( )	0.1976*** [0.0327]	0.0147*** [0.0024]	0.1775*** [0.0315]	0.0129*** [0.0023]	0.1690*** [0.0309]	0.0126*** [0.0023]
	Firm size ( )	0.7207*** [0.0326]	0.0536*** [0.0021]	0.7305*** [0.0319]	0.0532*** [0.0019]	0.7297*** [0.0319]	0.0545*** [0.0020]
	Capital intensity ( )	0.1898*** [0.0237]	0.0141*** [0.0017]	0.1930*** [0.0229]	0.0141*** [0.0016]	0.1950*** [0.0230]	0.0146*** [0.0016]
	Average wage ( )	0.2716*** [0.0418]	0.0202*** [0.0031]	0.2636*** [0.0404]	0.0192*** [0.0029]	0.2666*** [0.0400]	0.0199*** [0.0030]
	Industrial zone ( )	0.5963*** [0.0933]	0.0443*** [0.0069]	0.5935*** [0.0901]	0.0432*** [0.0065]	0.5871*** [0.0893]	0.0438*** [0.0066]
	Ownership structure ( )	0.2283 [0.1516]	0.0170 [0.0113]	0.3295** [0.1477]	0.0240** [0.0108]	0.3185** [0.1530]	0.0238** [0.0114]
	Foreign presence ( )	0.1466*** [0.0242]	0.0109*** [0.0018]	0.1942*** [0.0326]	0.0141*** [0.0024]	0.6194*** [0.1981]	0.0462*** [0.0149]
	Industry exporting ( )	0.0006*** [0.0001]	4E-05*** [6.88E-06]	0.0005*** [0.0001]	4E-05*** [6.48E-06]	0.0006*** [0.0001]	4E-05*** [6.56E-06]
	Industry concentration ( )	-0.2711** [0.1170]	-0.0202** [0.0087]	-0.4182*** [0.1168]	-0.0305*** [0.0085]	-0.4401** [0.1782]	-0.0329** [0.0133]
	Industry dummies ( )	Yes		Yes		Yes
	Regional dummies ( )	Yes		Yes		Yes
	Year dummies ( )	Yes		Yes		Yes
	Wald chi2	1514.7400***		1593.1400***		1617.2100***
	Log pseudolikelihood	-4325.3566		-4507.3745		-4604.9540
Notes: Models [1], [2] and [3] adopt the export, employment and asset share of MNEs in each three-digit food processing industry k in region j at time t, respectively; marginal effects (M.E) are evaluated at the sample means of the covariates; ***,  **,  * denote significance at 1%, 5% and 10% levels, respectively; estimates are efficient for arbitrary heteroscedasticity with computed robust standard errors (R.S.E.) in square brackets.

 

• The role of prior export experience

The estimation results in Table 4 indicate the crucial role of export experience in determining local firms’ export market entry. This is highly consistent with the existing literature, suggesting that firms with prior export experience have already incurred the fixed costs of market entry, and hence are more likely to continue exporting in the following year. The sunk entry costs can include the costs of identifying buyers, establishing distribution and logistics networks, product compliance/modifications and regulations, market research on consumer preferences and competitors, marketing campaigns, and so on. The greater the importance of prior export experience, the more important fixed entry costs are likely to be. While past experience has been extensively examined in predicting firm export behaviour, there is a notable scarcity of evidence on its role in moderating the export spillovers from MNEs. In Table 6, we add to the existing literature by splitting the sample and carrying out separate estimations for sub-samples of firms with and without prior experience of exports.

Table 6. Estimation results – The role of prior export experience

Regressor	Dependent variable [Local firms' export market entry]
	Local firms with prior export experience		Local firms without prior export experience
	Coef.	M.E	Coef.	M.E
Firm age ( )	0.1165*** [0.0395]	0.0396*** [0.0134]	0.1201*** [0.0307]	0.0050*** [0.0013]
Firm size ( )	0.5319*** [0.0325]	0.1810*** [0.0089]	0.6326*** [0.0327]	0.0261*** [0.0014]
Capital intensity  ( )	0.0964*** [0.0305]	0.0328*** [0.0103]	0.1872*** [0.0233]	0.0077*** [0.0009]
Average wage ( )	0.2646*** [0.0579]	0.0900*** [0.0196]	0.2373*** [0.0399]	0.0098*** [0.0017]
Industrial zone  ( )	0.4760*** [0.1054]	0.1619*** [0.0354]	0.4456*** [0.0848]	0.0184*** [0.0036]
Ownership structure ( )	0.4642*** [0.1646]	0.1579*** [0.0561]	0.4197* [0.2265]	0.0173* [0.0094]
Foreign presence ( )	0.0875 [0.1063]	0.0298 [0.0361]	0.3228*** [0.0841]	0.0133*** [0.0035]
Industry exporting ( )	0.0003** [0.0001]	1E-04** [0.0000]	0.0005*** [0.0001]	2E-05*** [3.90E-06]
Industry concentration ( )	-0.1309 [0.1726]	-0.0445 [0.0587]	-0.3150** [0.1284]	-0.0130** [0.0053]
Industry dummies ( )	Yes		Yes
Regional dummies ( )	Yes		Yes
Year dummies ( )	Yes		Yes
Wald chi2	690.7100***		628.7500***
Log pseudolikelihood	-1382.2215		-2945.3585

Notes: Marginal effects (M.E) are evaluated at the sample means of the covariates; ^***, ^**and  ^* denote significance at 1%, 5% and 10% levels, respectively; estimates are efficient for arbitrary heteroscedasticity with computed robust standard errors (R.S.E.) in square brackets.   

The estimation results reported in Table 6 suggest that the effects of MNE presence on export market entry differ between the two groups of local firms. The estimated coefficient of foreign presence ( ) is positive but insignificant in the estimation for local firms with prior export experience. Hence, once firms already entered a foreign market, it seems they have overcome the entry barriers related to fixed costs of exporting and gained sufficient experience and knowledge to keep exporting the next year. MNE presence does not seem to exert any significant additional effect on the export decisions of these firms. Meanwhile, local firms without prior export experience are more strongly affected by foreign presence. The estimated marginal effect indicates that a one percentage point increase in the presence of MNEs in the industry raises the probability that firms will start exporting by 1.33 percentage points. This implies that spillovers from MNEs are likely to lower the market entry barriers for local food processing firms that are about to start exporting for the first time. As nearly 87% of local firms in the industry have no prior export experience, the evidence from this sub-sector analysis is potentially important to the prospect of Vietnam’s food exports and the opportunity for domestic firms to accelerate the integration into global value chains.

5. Concluding remarks

This paper develops new insights into the spillover effects of MNEs on local firms’ export market entry, using a newer and more disaggregated dataset on Vietnam’s food processing industry than what earlier studies have done. Food processing represents an essential part of Vietnam’s thriving manufacturing sector, with enormous potential for exports. Following earlier literature, we specify an econometric model which includes a measure of MNE presence plus a set of control variables capturing standard determinants of firms’ export decisions. The study adopts a random-effects Probit estimator to deal with the fact that the dependent variable is constrained and binary in nature and the examined data set is an unbalanced panel. To deepen the understanding of MNEs-linked effects, extended analyses are conducted to further examine the role of the measurement of MNE presence and local firms’ prior export experience in moderating the magnitude of any spillovers from MNEs.

The estimation results show strong evidence that the presence of MNEs in Vietnam’s food processing industry significantly raises the chances for local firms to become exporters and start integrating into global value chains. A one percentage point increase in foreign presence leads to a 2.18 percentage point increase in the probability that local firms engage in exporting. Further analysis indicates the existence of positive export spillovers from MNEs to the privately owned group of firms but not the state-owned ones. Moreover, the empirical results confirm findings from earlier studies regarding firm-level export determinants: firms with prior export experience, older, larger, more capital-intensive and higher-paying firms tend to expand to the foreign market at a higher rate. Being located in an industrial zone also raises firms’ propensity to start exporting. In addition, firms operating in more export-oriented and highly competitive three-digit industries are more likely to reach out to the foreign market.  

The key result from this paper highlights the beneficial export information externalities and valuable competitive pressure generated by MNEs as catalysts for local food processing firms’ export entry decisions. The findings from this sub-sector analysis are noteworthy as earlier studies for Vietnam’s manufacturing sector largely failed to present such evidence. These positive export spillovers from MNEs can be expected to be relatively important in the examined industry, given the low proportion of existing exporters among local firms. Contrary to previous work, this study shows a prominent role for privately owned firms in exports as well as in absorbing positive MNE spillovers. Therefore, to amplify potential spillovers, we suggest that policymakers should focus on promoting stronger linkages between MNEs and this group of local food processing firms. Possible mechanisms are via co-location (e.g., industrial and export zones) and collaboration in organizing regular professional platforms to share knowledge on export products, technology and markets (e.g., agro/food export fairs and exhibitions).  

Further analysis suggests that in evaluating MNEs-linked export spillovers, measurement of foreign presence and local firms’ past experience matter. While the evidence of a positive impact from MNEs is robust across our four proxies of foreign presence, the magnitude varies markedly. The estimated marginal effects imply the relatively more profound importance of the competitive pressure compared to the export information and interpersonal externalities. Additional regressions reveal that the presence of MNEs appears to play an insignificant role for domestic firms with previous export experience. However, MNEs exert a significant and valuable impact on local firms without prior export knowledge, enabling them to overcome market entry barriers to start exporting. These findings imply the considerable potential for MNEs to make a difference in local firms’ export prospects and validate the policy efforts to attract FDI inflows to the food processing industry.      

References

Aitken, B., Hanson, G. H., & Harrison, A. E. (1997). Spillovers, foreign investment, and export behavior. Journal of International Economics, 43(1), 103-132.

Aitken, B., & Harrison, A. E. (1999). Do Domestic Firms Benefit from Direct Foreign Investment? Evidence from Venezuela. American economic review, 89(3), 605-618.

Alin, A. (2010). Multicollinearity. Wiley Interdisciplinary Reviews: Computational Statistics, 2(3), 370-374.

Anwar, S., & Nguyen, L. P. (2011). Foreign Direct Investment and Export Spillovers: Evidence from Vietnam. International Business Review, 20(2), 177-193.

Anwar, S., & Sun, S. (2014). Heterogeneity and curvilinearity of FDI-related productivity spillovers in China's manufacturing sector. Economic Modelling, 41, 23-32.

Asian Development Bank (ADB). (2019). Asia Integration Center - Tracking Asian Integration - Vietnam Profile. Retrieved from https://aric.adb.org/vietnam/data

Balsvik, R. (2011). Is labor mobility a channel for spillovers from multinationals? Evidence from Norwegian manufacturing. The Review of Economics and Statistics, 93(1), 285-297.

Barrios, S., Görg, H., & Strobl, E. (2001). Explaining firms export behaviour: the role of RD and spillovers. Retrieved from

Bernard, A. B., & Jensen, J. B. (2004). Why some firms export. Review of Economics and Statistics, 86(2), 561-569.

Blomström, M., & Kokko, A. (1998). Multinational corporations and spillovers. Journal of economic surveys, 12(3), 247-277.

Broocks, A., & Van Biesebroeck, J. (2017). The impact of export promotion on export market entry. Journal of International Economics, 107, 19-33.

Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., . . . Leitão, P. J. (2013). Collinearity: A Review of Methods to Deal with it and a Simulation Study Evaluating their Performance. Ecography, 36(1), 27-46.

Forte, R., & Salomé Moreira, A. (2018). Financial constraints and small and medium-sized firms’ export propensity: evidence from Portuguese manufacturing firms. International Journal of the Economics of Business, 25(2), 223-241.

Fox, J. (1991). Regression Diagnostics: An Introduction (Vol. 79): Sage.

Giroud, A., & Mirza, H. (2015). Refining of FDI motivations by integrating global value chains’ considerations. The Multinational Business Review.

Greenaway, D., Guariglia, A., & Kneller, R. (2007). Financial factors and exporting decisions. Journal of International Economics, 73(2), 377-395.

Greenaway, D., Sousa, N., & Wakelin, K. (2004). Do domestic firms learn to export from multinationals? European Journal of Political Economy, 20(4), 1027-1043.

Greenaway, D., Sousa, N., & Wakelin, K. (2004). Do Domestic Firms Learn to Export from Multinationals? European Journal of Political Economy, 20(4), 1027–1043.

GSO. (2018). Foreign Direct Investment Projects Licensed by Kinds of Economic Activity. Retrieved from http://gso.gov.vn/default_en.aspx?tabid=776.  

Ha, V., Holmes, M. J., & Hassan, G. (2020). Does foreign investment benefit the exporting activities of Vietnamese firms? The World Economy, 43(6), 1619-1646. doi:https://doi.org/10.1111/twec.12912

Haddoud, M. Y., Onjewu, A.-K. E., Nowiński, W., & Jones, P. (2021). The determinants of SMEs’ export entry: A systematic review of the literature. Journal of Business Research, 125, 262-278.

Imbriani, C., Pittiglio, R., Reganati, F., & Sica, E. (2014). How much do technological gap, firm size, and regional characteristics matter for the absorptive capacity of Italian enterprises? International Advances in Economic Research, 20(1), 57-72.

Jaffe, A. B., Trajtenberg, M., & Henderson, R. (1993). Geographic localization of knowledge spillovers as evidenced by patent citations. The Quarterly Journal of Economics, 108(3), 577-598.

Javorcik, B. S. (2004). Does Foreign Direct Investment Increase the Productivity of Domestic Firms? In Search of Spillovers through Backward Linkages. The American Economic Review, 94(3), 605-627.

Kim, M., & Xin, D. (2021). Export spillover from foreign direct investment in China during pre-and post-WTO accession. Journal of Asian Economics, 75, 101337.

Kneller, R., & Pisu, M. (2006). The role of experience in export market entry: Evidence for UK Firms. University of Nottingham Research Paper(2006/48).

Kneller, R., & Pisu, M. (2007). Industrial Linkages and Export Spillovers from FDI. The World Economy, 30, 105–134.

Kohpaiboon, A. (2006). Foreign direct investment and technology spillover: A cross-industry analysis of Thai manufacturing. World Development, 34(3), 541-556.

Kokko, A., Zejan, M., & Tansini, R. (2001). Trade Regimes and Spillover Effects of FDI: Evidence from Uruguay. Review of World Economics, 137, 124–149.

Ma, A. C. (2006). Export spillovers to Chinese firms: Evidence from provincial data. Journal of Chinese Economic and Business Studies, 4(2), 127-149.

Morrow-Howell, N. (1994). The M word: Multicollinearity in multiple regression. Social Work Research, 18(4), 247-251.

Nguyen, A. L. T. (2020). FDI inflows and intellectual property rights for MNEs in emerging markets: an alternative approach through the lens of trademarks in Vietnam (1986-2016). Multinational Business Review.

Nguyen, D. T. H. (2021). Firm performance in the financial industry: the role of foreign entrants and their origin. Multinational Business Review. doi:10.1108/MBR-04-2020-0078

Nguyen, D. T. H., & Sun, S. (2012). FDI and Domestic Firms’ Export Behaviour: Evidence from Vietnam. Economic Papers: A journal of applied economics and policy, 31(3), 380-390.

Phillips, S., & Ahmadi-Esfahani, F. Z. (2010). Export Market Participation, Spillovers, and Foreign Direct Investment in Australian Food Manufacturing. Agribusiness, 26(3), 329– 347.

Phillips, S., & Ahmadi‐Esfahani, F. Z. (2010). Export market participation, spillovers, and foreign direct investment in Australian food manufacturing. Agribusiness, 26(3), 329-347.

Ruane, F., & Sutherland, J. (2005). Foreign Direct Investment and Export Spillovers: How do Export Platforms Fare? Paper presented at the IIIS Discussion Paper No. 58 Dublin.

Scoppola, M. (2021). Globalisation in agriculture and food: the role of multinational enterprises. European Review of Agricultural Economics, 48(4), 741-784.

Smeets, R. (2008). Collecting the Pieces of the FDI Knowledge Spillovers Puzzle. The World Bank Research Observer, 23(2), 107-138. doi:10.1093/wbro/lkn003

Sun, S. (2009). How does FDI Affect Domestic Frms’ exports? Industrial Evidence. The World Economy, 32, 1203–1222.

Sun, S., Song, L., & Drysdale, P. (2011). The role of geographical proximity in FDI productivity spillovers in China. In: Edwar Elgar Publishing.

Villar, C., Mesa, R. J., & Barber, J. P. (2020). A meta-analysis of export spillovers from FDI: advanced vs emerging markets. International Journal of Emerging Markets.

1. (2020). World Development Indicators. Washington DC: World Bank. .

Wooldridge, J. M. (2019). Correlated random effects models with unbalanced panels. Journal of Econometrics, 211(1), 137-150.