Trade Protection and Productivity Differentials between Multinationals and Local Firms in Vietnamese Manufacturing.

• Author: NgocThuyen, Truong Thi

1. Introduction

   Multinational enterprises (MNEs) are key players in the process of global economic integration. Production of MNEs has tended to grow faster than production of local firms in Vietnam and many other economies in Asia and worldwide (Ramstetter 2012; Ramstetter and Phan 2013). Theory suggests that to become an MNE, a firm must first own assets such as advanced production technology, management skills, distribution networks and marketing know-how (Dunning 1993; Markusen 1991). In addition, MNEs also tend to have relatively high R&D propensities (R&D-sales ratios). These firm-specific assets are crucial for MNEs to enter and expand in a host country (Wang and Blomstrom 1992) If MNEs do indeed possess these firm-specific assets in relatively large amounts, it follows that they will tend to be more productive than non-MNEs.

   Although economy theory suggests that foreign affiliates will have higher productivity than non-MNEs, existing empirical evidence is mixed. For example, analyses of large samples of manufacturing plants in Mexico (Blomstrom 1986) and Indonesia (Takii 2004), or manufacturing firms in Vietnam (Ramstetter and Phan 2013), respectively, found that MNEs tended to have relatively high productivity. However, industry-level analyses for Indonesia and Vietnam suggested that MNE-local productivity differentials were seldom significant statistically. Moreover, evidence for manufacturing plants in Malaysia (Haji Ahmad 2010; Menon 1998; Oguchi et al. 2002) suggested small and generally insignificant differentials in productivity levels or growth. MNE-local differences were also found to be generally insignificant for manufacturing plants in Thailand (Ramstetter 2004).

   This paper contributes to the growing literature on the firm-level productivity effects of MNEs and import protection in Vietnam and other developing economies. It extends previous analyses of Vietnam's manufacturing firms to 2005-10, comparing results from both large samples of all manufacturing firms and smaller samples of industry subgroups of industries and sub-periods. (1) Vietnam is an interesting case to study because it experienced substantial market-oriented reforms after the economic renovation policy (Doi Moi) was introduced in 1986. Trade liberalization and corporate reform then accelerated substantially in the early to mid-2000s as Vietnam prepared to join the World Trade Organization in January 2007. (2) Attracting MNEs and spurring productivity improvements have also been key policy focuses in the country's economic transition.

   One of this study's more important contributibutions is the analysis of how trade policy affects productivity differentials between MNEs and non-MNEs. Literature on the role of MNEs in economic development often suggests that liberalization of foreign investment should go hand in hand with liberalization of trade policy in order to encourage MNEs to employ technologies more consistent with the host country's comparative advantage (Athukorala and Chand 2000; Balasubramanyam, Sapsford, and Salisu 1996; Bhagwati 1968, 1973). This also creates a higher probability of MNEs generating positive spillovers for local non-affiliated firms; there is some evidence supporting this conjecture in Vietnam (Truong, Jongwanich, and Ramstetter 2015). Thus, MNE-local productivity differentials are expected to be smaller under the trade liberalization regime. There is no clear evidence regarding this issue in Vietnam.

   The rest of this paper is organized as follows. The next section reviews the previous literature in more detail. The third section presents data on ownership and productivity of Vietnamese manufacturing. The subsequent section discusses the empirical model and results. The final concludes and highlights policy implications.

2. Literature Review

   As described above, foreign MNEs in developing economies are expected to be more productive than local firms or plants in developing economies such as Vietnam, largely because MNEs possess relatively large amounts of firm-specific assets than non-MNEs, and most local firms are not MNEs themselves. The empirical evidence on this point, however, is mixed. Evidence from manufacturing plants in Indonesia and Vietnam is somewhat more consistent with the hypothesis that MNEs have relatively high productivity. But, that is not the case when all production function coefficients are allowed to vary among industry groups (by estimating separate equations for each group) compared to the scenario when inter-industry differences are estimated with intercept dummies alone in large samples of all firms.

   The lack of significant productivity differentials between foreign MNEs and non-MNEs could be due to two reasons. First, MNE parent companies may be reluctant to transfer their proprietary assets to minority-owned affiliates (Caves 2007; Moran 2001). If this is true, productivity differentials may become significant only when parent firms hold large ownership shares and control access of local firms to access such assets. However, empirical evidence on this issue is also unclear. For example, in large samples of Indonesian manufacturing plants, Takii (2004) found that positive and significant productivity differentials were most common for majority MNEs, whereas Takii and Ramstetter (2005) found that productivity was not always larger in plants with relatively large foreign ownership shares. Similar evidence also exists for Thai plants (Ramstetter 2004) and Vietnamese firms (Ramstetter and Phan 2013).

   It is important to note that the Vietnamese studies usually distinguish between two distinct groups of MNEs: wholly-owned MNEs (WOs) and joint venture MNEs (JVs). In WOs, the parent company has control of all business decisions and the firm's managerial system usually resembles that of the parent. Moran (2001) argues that WOs tend to be more efficient than JVs because the former are more tightly integrated into the parent company and have greater access to the MNE's intangible assets. Particularly in the case of a minority-foreign JV, the fear of technology leakages

   and/or the lack of managerial control may make the parent company reluctant to share its most advanced technology with the JV On the other hand, JVs can clearly be beneficial for MNEs, especially when the parent has limited knowledge of a host economy. For example, if the JV partner has intangible assets that complement the MNE's assets, and/or when the MNE and the JV partner(s) can cooperate closely and share information when making decisions.

   However, empirical studies on the relationship of productivity differentials and foreign ownership have been inconclusive. Blomstrom and Sjoholm (1999), for instance, found that ownership did not affect manufacturing labour productivity differentials in Indonesia in 1991. Similarly, Takii and Ramstetter (2005) showed that labour productivity during the 1986-2001 period was usually highest in majority-foreign (50-89 per cent foreign) MNEs and lower in minority-foreign MNEs or in heavily foreign (more than 90 per cent foreign) MNEs. In contrast, Takii's (2004) more general evidence from Indonesia in 1995 indicated that WOs tended to have higher productivity than JVs in large samples of all manufacturing plants, but this difference was rarely significant at the industry level. Other studies of Europe (for example, De Backer and Sleuwaegen (2003) for Belgium in 1990-95; and Javorcik and Spatareanu (2008) for Romania in 1998-2003) found that productivity in JVs improved faster than in WOs. For Vietnam, Nguyen et al. (2006) found no significant labour productivity differentials between WOs and JVs, while Ramstetter and Phan (2013) indicated that WOs were generally more productive than JVs, but the results differed depending on the industry and subperiod examined.

   Documented evidence on productivity spillovers from MNEs to local firms is the second reason productivity differentials may be relatively small or insignificant. This is thought to be a major factor behind the lack of productivity differentials in Malaysia and Thailand (Ahmed 2010; Ramstetter 2006). Spillovers occur when MNE presence affects the productivity of local firms. This can take place through three channels. First, spillovers may result from forward or backward linkages between MNEs and local firms, although backward linkages are usually thought to be more important (Dunning 2008). Second, labour mobility can be an important source of spillovers, especially when relatively skilled workers move from MNEs to local firms or start-up new local firms (Chen 1983; Gorg and Strobl 2005; Katz 1987; Kohpaiboon 2006a). However, MNEs often seek to minimize the turnover of high-quality employees by paying higher wages and offering better benefits than local firms (Fosfuri, Motta, and Ronde 2001; Javorcik 2004; Moran 2002). Third, MNE presence often increases competition and encourages domestic firms to improve efficiency--often by imitating MNEs (Kokko 1994; Wang and Blomstrom 1992). When spillovers occur, the productivity gap between MNEs and local firms reduces over time.

   The theoretical literature on immiserization (Bhagwati 1973, 1985; Brecher and Findlay 1983; Brecher and Alejandro 1977) also suggests that the extent of spillovers from MNEs is likely to be affected by the host's trade policies. Import protection distorts resource allocation and reduces motives for productivity improvement in all firms, including MNEs. Correspondingly, some studies suggest that negligible or negative spillovers are more common when protection is high (Moran 2002; Kohpaiboon 2006b). This is because MNEs often produce for relatively limited local markets, which they often dominate. In addition, when protection is high, MNEs and local firms are often more motivated to lobby for protection than to pursue productivity gains. In contrast, export promotion usually entails lower levels of protection, greater emphasis on international...