«* Archanun Kohpaiboon Faculty of Economics, Thammasat University, Bangkok Thailand archanun ABSTRACT: This paper examines Foreign ...»
Vertical and Horizontal FDI Technology Spillovers:
Evidence from Thai Manufacturing
Faculty of Economics, Thammasat University,
ABSTRACT: This paper examines Foreign Direct Investment (FDI) spillover, using an
unbalanced panel data set of the manufacturing survey of Thailand during the period
2001-03. In this paper, not only are both horizontal and vertical FDI technology
spillovers examined, but the former is also assumed to vary across industries. The key hypothesis is that horizontal FDI spillovers depend on the trade policy regime as well as the absorptive capability of locally owned plants. Our panel data econometric analysis highlights the important role of the trade policy regime as a conditional gain of horizontal FDI spillovers. In particular, positive horizontal FDI spillovers are found only in an industry operating in a relatively liberal environment. Interestingly, imposing an assumption of identical horizontal FDI spillovers across industry could result in biased estimates of vertical FDI spillovers. The key policy inference highlights the relative importance of the trade policy regime in harnessing the gain from foreign presence. Liberalizing the foreign investment regime thus has to go hand in hand with liberalizing the trade policy to gain FDI technology spillovers. Our finding here gives a warning not to overemphasize the role of linkages. It is the quality rather than magnitude of linkages that should be used a proxy of the magnitude of vertical FDI spillovers.
JEL Classification: F14, O24, D24 * I would like to thank Professor F. Kimura, Mr. S. Umezaki and all participants at the Workshop ‘Deepening East Asia Economic Integration: Part II Firm-level Analysis, organized by Economic Research Institute for ASEAN and East Asia (ERIA), December 23, 2008 and February 9, 2009. The opinions expressed in this paper are the sole responsibility of the author and do not reflect the views of the ERIA.
1. Issues Enticing multinational enterprises (MNEs) to set up affiliations is placed high on the policy agenda in many countries, especially developing ones, as their entry would bring in much-needed capital, new production technologies, marketing techniques and management knowhow. While all of these potential benefits of Foreign Direct Investment (FDI) are viewed as important, particular emphasis is placed on technological gains in the productivity and competitiveness of the domestic industry, known as FDI technology spillovers (henceforth referred to as FDI spillovers). As a result, the expectation of gaining from technology spillover persuades many developing countries to offer variousincentives in order to attract FDI. Nonetheless, only in some investment-receiving (host) countries are FDI spillovers empirically found.
While tangible efforts have recently been made to gain a better understanding of the factors that determine the presence of FDI spillovers, they have not thus far borne fruit (Crespo& Fontoura, 2007). The existing literature divides into two broad themes.
First, horizontal FDI spillovers are assumed not to be automatic but are hypothesized as being a function of the economic environment and domestic policies in host countries.
In this literature, two determinants have been generally recognized as conditioning gains from FDI. These are the trade policy regime and the absorptive capability of locally owned enterprises.1 While both of these factors are acknowledged, most researchers have examined only the role of absorptive capability. This may be because of the difficulty of finding a reliable proxy for protection across industries. So far only a few studies (e.g. Kokko et al., 2001; Kohpaiboon, 2006) have examined empirically the role of thetrade policy regime. Additionally, there is a dearth of studies that bring absorptive capacity and the trade policy regime together in examining FDI spillovers. A major caveat of literature in this field is that it concentrates only on spillovers taking place within a given industry, (i.e. horizontal FDI spillovers).
See the comprehensive survey in Görg & Greenaway (2004) Crespo & Fontoura (2007) Hayakawa et al. (2008).
In fact, a number of recent studies2 argue that it is more likely that FDI spillovers would take place through backward and forward linkages (i.e. vertical FDI spillovers) as opposed to horizontal ones. That is, where foreign investors involve themselves with indigenous enterprises in upstream and/or downstream industries, it is very likely that the latter will gain technological benefit from the former. MNEs would have an incentive to prevent information leakage to their competitors, including local enterprises, thereby reducing the possibility of horizontal spillover taking place. By contrast, there would be incentive for them to transfer knowledge to their local suppliers because such knowledge transfer would benefit the MNEs in terms of getting better input quality and/or cheaper costs, and receiving inputs on time. It is also plausible that spillovers from MNEs in upstream industries exist to provide inputs that either were previously unavailable in the country or to make them technologically more advanced or less expensive, or to ensure that they are accompanied by the provision of complementary services (Javorcik, 2004).
Empirical studies examining the presence of vertical FDI technology spillovers are sparse (Blomström et al. 2000; Lin & Saggi, 2005). The notable exception is Javorcik (2004) and Blalock & Gertler (2008) which examined cases in Lithuanian and Indonesian manufacturing sectors, respectively. Their key finding supports the relative importance of vertical against horizontal FDI spillovers. In particular, it was found that vertical FDI spillovers were statistically significant. Nevertheless, a major caveat in these two studies is that their empirical model contains the implicit assumption that horizontal FDI spillovers are identical for all industries. As argued above such an assumption is rather restrictive. In addition, the correlation between protection and the extent of industries generating backward linkages tends to be positive, and omitting the trade policy regime in examining FDI spillovers could create bias in the results.
Against this backdrop, this paper examines the presence of FDI technology spillover in Thai manufacturing. Panel data econometric analysis is conducted, using They are Rodŕigueze-Clare (1996), Markusen & Venables (1999), Javorcik (2004), Lin & Saggi, (2005), Blalock & Gertler (2008) the Industrial Survey conducted by the Office of Industrial Economics, Ministry of Industry, during the period 2001-2003. This is the most up to date and reliable plant survey available so far. In the empirical model, we follow the general practice in this research area, in which the productivity equation of locally owned plants in the manufacturing sector is estimated and the statistical relationship between plants’ productivity and the extent of foreign presence is examined. This paper contributes to the existing literature in two ways. First, in our econometric analysis both horizontal and vertical FDI spillovers are examined. So far there have been few studies (e.g.
Javorcik (2004) and Blalock & Gertler (2008) examining both spillovers simultaneously. Additionally, our measure of backward and forward linkages takes into consideration both direct and indirect (inter-sectoral) repercussions. This is different from Javorcik (2004) and Blalock & Gertler (2008) in which only the direct linkage is included. Secondly, we allow horizontal FDI spillovers to vary across industries. Trade policy regime and absorptive capability are included in the empirical model as the key factors determining the extent of horizontal FDI spillovers.
Thai manufacturing is a good laboratory for the issue in hand for two reasons.
First, Thailand has been a large FDI recipient throughout the past three decades.
However, few studies have examined technology spillover in Thai manufacturing. So far there have been two studies, Kohpaiboon (2006) and Kohpaiboon and Jongwanich (forthcoming), both of which are based on the Industrial Census of 1996. Hence, this paper not only provides up-to-date evidence but also re-examine the relative importance of spillover channels, and horizontal versus vertical spillovers. Secondly, Thai manufacturing is broad-based as opposed to neighbouring countries, covering a wide range of industries from traditional labour- intensive industries like garment and footwear to several key industries in the machinery and transport equipment sector such as automotive, electronics, and electrical appliances. Hence, evidence drawn from Thai manufacturing would provide an insightful lesson for other countries.
The paper is organized as follows: Section 2 provides an analytical framework illustrating possible channels where FDI spillover could take place as well as the role of key determinants conditioning FDI spillovers. In Section 3, patterns of labour productivity across industries are discussed and related to the extent of the foreign presence and the effective rate of protection. The following section explains the empirical model used in this paper (Section 4). Section 5 presents data and variable construction and regression results are in Section 6. Conclusion and policy inferences are in the final section.
2. Analytical Framework While MNEs have the potential to generate considerable impact on host countries’ economies, it is often argued that spillovers are the most desirable benefit of all. In general, there are at least three channels through which FDI spillovers can occur. The first channel is the demonstration effect. The presence of foreign firms can have a demonstration effect that allows local firms to become familiar with superior technologies, marketing and managerial practices used in foreign affiliates. Thus, spillover can take place in the form of imitating the foreign subsidiaries’ technology.
Over and above this, the presence of foreign affiliates can exert pressure on local firms exhibiting technical or allocation inefficiencies to adopt more efficient methods. This allows local firms to survive successfully or even compete with foreign firms. Since both demonstration and competition effects are likely to occur simultaneously, these two effects are regarded in the literature as a single channel of spillover.
Linkage is the second channel of FDI spillovers. Where foreign investors are linked to upstream and downstream industries in host countries, the linked indigenous firm has the possibility of gaining technological benefits. The former is referred to as backward linkage and the latter as forward linkage. By backward linkage, foreign investors establish an inter-firm relationship with local suppliers and create demand for inputs from local suppliers in upstream industries. When these local firms are engaged to supply certain raw materials, the high quality, reliability and speed of delivery that MNE affiliates demand force them to enhance productivity. Moreover, in some cases, local suppliers in upstream industries receive technical and managerial training in the production of the required inputs. This is likely to generate additional economic activity and income, and to transfer technological and management skills to the host country.
Similarly, forward linkage effects are created when one industry uses another industry’s output as its inputs. Every activity that does not by its nature cater exclusively to final demand induces attempts to utilize its outputs as inputs in other industries. Benefits for domestic suppliers resulting from the presence of MNEs may be extended to other domestic firms that produce end-user consumer goods. The most evident link is observed in the MNEs’ supply of higher quality inputs and/or at a lower price to domestic producers of end-user consumer goods. The sum of the backward and forward linkages gives a total linkage effect, which can be seen as the growth in other new industries induced by establishing an MNE affiliates.
The last channel is labour mobility. Foreign affiliates generally play a more active role than local firms in educating and training local labour. Through this training and subsequent work experience, workers become familiar with the foreign affiliates’ technologies and production methods. FDI spillovers through this channel occur when employees of foreign affiliates move on to local employers or set up their own business, using knowledge gained during their previous employment.
Empirically, most econometric studies have only examined the presence of FDI spillovers through the demonstration and linkage channels simply because of data availability. Analysis of labour mobility is very limited as researchers must have access to information about top managers’ backgrounds. Unfortunately, such information is not usually available.3 Secondly, in theory, FDI spillovers through the demonstration effect can take place either within the same industry or across industries. In practice, it is very difficult to measure the demonstration effect across industries so that spillovers through demonstration effects are usually referred to as horizontal FDI spillovers. On the other hand, FDI spillovers through linkage occur when MNEs are located in a given industry, To the best of our knowledge so far, the only econometric analysis of spillovers through labour mobility is undertaken by Görg and Strobl (2002), using firm level data in Ghana.
and benefit upstream and downstream industries. These are regarded as FDI vertical spillovers.
The recent studies such as Rodŕigueze-Clare (1996); Markusen & Venables (1999); Lin & Saggi (2005); Javorcik (2004); and Blalock & Gertler (2008) highlight the relative importance of vertical FDI spillovers as opposed to horizontal ones. In particular, they argue that vertical FDI spillovers are likely. For example Blalock & Gertler (2008) argue that it is hard to believe that horizontal FDI spillovers are likely.
Firstly, the technology gap between foreign and domestic firms may often be wide.