Content area
While scholars extensively study the middle-income trap, gaps persist, particularly regarding foreign direct investment and its role in transitioning from middle- to high-income status. This mixed-methods research examines the relationship between foreign direct investment inflows and escaping the middle-income trap. Utilizing a comparative case study of the automotive industries in Mexico and South Korea during the 1960s to 2000s, alongside logistic regressions, we consistently observe a negative correlation; higher foreign direct investment inflows are associated with a decreased likelihood of escaping the middle-income trap. This confirms the World-Systems Theory’s view of the relationship between foreign direct investment inflows and the middle-income trap. The implications of this study should induce caution on the part of middle-income countries when accepting foreign direct investment in strategic, high-value-adding industries. This study recommends opting for purchasing foreign technology or for joint ventures that ensure technology transfers. This research addresses a critical gap, offering insights into the nuanced dynamics of foreign direct investment’s effects on the transition from middle to high-income status.
Introduction
The World Bank classifies countries into four categories based on their gross domestic product (GDP) per capita level. For the fiscal year 2023, low-income (LI) countries are those that have a GDP per capita under 1086 USD, lower middle-income (MI) countries are those with a GDP per capita between 1086 USD and 4255 USD, upper MI countries’ GDP per capita range is between 4256 USD and 13,205 USD, and any country with a GDP per capita over 13,205 USD is considered high-income (HI) (The World Bank, 2023).
101 out of the 114 countries classified as MI in 1960 stayed there until 2008 (The World Bank and Development Research Center of the State Council, the People’s Republic of China, 2013). Meaning during that almost 50-year period, merely 13 countries, so-called breakout cases, have been able to graduate to the ranks of HI status.1 This phenomenon has been dubbed the middle-income trap (MIT), a term first coined by the World Bank economists, Indermit Gill, Kharas (2007). According to the term’s originators, scholars have since latched onto the idea of MIT. As of 2015, papers on Google Scholar with the term in their title were at around 300, while papers with the term in its contents exceeded 3000 (Gill & Kharas, 2015). One can imagine the number has grown significantly since then. To define the MIT, a country is deemed trapped by remaining at lower MI status for more than 28 years, upper MI status for 14 years, or in MI status in general for longer than 42 years (Felipe et al., 2012).
In general terms, MI countries fail to transition to HI status and get trapped because they lack the ability to compete with either LI or HI countries. As MI countries develop and their wages grow, they lose their competitive edge in low-value-adding industries, yet unless they have successfully upgraded their industries to compete in higher-value-adding ones, they lack the ability to compete with HI countries in said industries. Accordingly, the intuitive logic behind understanding MIT as a result of the failure of industrial upgrading is evident (Anastasi, 2024b: p. 48).
Despite the popularity of MIT among scholars, some aspects and avenues of inquiry have yet to be fully investigated. One such aspect is that of foreign direct investment’s (FDI) effect on MI countries’ prospects for transitioning to HI status. As various models of state-led growth started to decline around the 1980s, the rise of global supply chains integrated LI countries and MI countries into the production processes of multinational corporations (MNCs), which chiefly originate from HI countries (Oks & Williams, 2022). This process has happened to such a great extent that to produce a single iPhone, various parts are needed from 43 countries on six different continents (Petrova, 2018).
FDI is used to set up fully foreign-owned and controlled firms in the host country or, in some cases, create joint ventures in partnership with local firms. As of 2021, the flows of FDI sat at 2.2 trillion USD, or 2.3 percent of global GDP, according to the World Bank (2023). In the same year, MI countries welcomed 754 billion USD of FDI inflows or 2.1 percent of MI countries’ GDP. MNCs’ and local firms’ interests often diverge, and with the rise of international agreements such as TRIMS (Trade-related Investment Measures), the host country’s state capacity to regulate and influence foreign firms has shrunk considerably. With the substantial amount of inflows, not to mention the stock, of FDI into the MI world and the limited power MI countries’ states have on MNCs’ operations within their borders, a rigorous investigation into its effect on MI countries’ relationship with the MIT is warranted.
While some scholars have theorized the effect FDI has on an MI country’s prospects for transitioning to HI status (Anastasi, 2023; Henley, 2018; Doner & Schneider, 2016; Zeng & Fang, 2014), to the best of the author’s knowledge, the number of empirical studies on the relationship is rather small (CM et al., 2024; Nguyen-Huu & Pham, 2021). The empirical studies, while they do look at the effects of FDI on the transition from MI to HI status, the relationship between MIT and FDI is not their main focus, leaving gaps in the scholarship. This paper, deploying a mixed-methods approach, seeks to fill this gap, first by a comparative case study and then by several logistic regressions.
The literature primarily focusing on FDI and MIT is rather limited. The scholarship on FDI and industrial upgrading as a whole is much more expensive, yet falls outside of the scope of this research, as we are primarily focused on MI countries and the industrial upgrading necessary to make the transition to HI status. Due to this, a brief comparison of the same industry during the same time period, in two different countries with two distinctly different relationships with both FDI and MIT should shine some light on this issue. We will look at how South Korea, a break-out case, was able to build a world-class automotive industry, while Mexico, a victim of MIT, was unable to, with special attention paid to FDI.
The rationale for choosing the automotive industry2 is that during the time period of this comparative case study (1960s–2000s), it was considered a late industry, due to its demanding technological prerequisites, representing modernity (Chenery & Taylor, 1968). Yet this also implies that the firms and states that possess this technology will guard it fiercely. However, breaking into this industry is not impossible for countries and firms, although it is arduous. This study focuses on the case of South Korea and Mexico within the automotive sector, both having successfully established domestic industries, albeit with marked differences.
The purpose of this comparative case study is to analyze the impact of FDI on industrial upgrading. Acknowledging the imperfect nature of using industrial upgrading as a proxy for avoiding MIT, the study aims to tease out the distinctions in FDI acceptance models and their effects on industrial upgrading.
When constructing this comparative case study two additional variables (besides time) were controlled for: level of development and population size. South Korea and Mexico were around the same level of development in the early 1960s in terms of GDP per capita, which is also around the same time both attempted to use industrial policy to develop an automotive industry (The World Bank, 2024). Their population sizes were comparable during this period, however, Mexico’s population size grew much larger in subsequent years (The World Bank, 2024). This might be in part due to the fact that developing countries’ population growth is higher than developed countries.
The quantitative section seeks to confirm or deny the results of this study’s comparative case study. An explanation of the data, estimation methods, and model construction methods is given in that section of the paper.
The paper will proceed as follows: the “Literature review” section will provide a literature review of the relationship between FDI and MIT. The following section will qualitatively compare two different models of welcoming FDI by one break-out case, South Korea, and one country stuck in the MIT, Mexico. The next two sections will explore and analyze the data used and the logistic regression models built to test this relationship and discuss their results respectively. The succeeding section covers implications, and the last section serves as a conclusion.
Literature review
Positive relationship between FDI inflows and the MIT
As failure to advance towards HI status is a failure to upgrade domestic industries (Anastasi, 2024b: p. 48), the question of the relationship between MIT and FDI thus turns to does FDI aids or poses a hindrance to the type of industrial upgrading necessary to make the transition to HI status. According to Zeng and Fang’s (2014) analysis, FDI presents an obstacle. Using World-Systems Theory as a theoretical lens, the pair examines how China’s relationship with MIT might shape up to be. Zeng and Fang argue that world-systems theory does a great deal of explaining MIT since, “during the last two decades of the 20th century there appears to have been less upward mobility than occurred in the 1960s and 1970s, which suggests that ‘globalization’, rather than spreading wealth and facilitating development, may create barriers for nations on the lower rungs of the global economy to move upwards.”
Proponents of world-systems theory argue that the international division of labor serves as the fundamental organizing principle of the global economy. Consequently, peripheral and semi-peripheral countries specialize in the production of ‘periphery-like products,’ while core countries specialize in ‘core-like products.’ These core-like products necessitate “(high levels of) industrialization, advanced technology, and well-developed financial systems” for their production, thereby limiting their creation to only a handful of countries and companies with the requisite capabilities. As a result, core-like products hold a “quasi-monopolized” status, allowing for significant extraction of profits due to the lack of competition. In contrast, periphery-like products have lower production requirements, fostering genuinely competitive markets for them. When exchanges occur between the periphery and semi-periphery and the core, core-like products possess a significantly stronger position relative to periphery-like products, resulting in a flow of surplus value from the developing world to the developed world.
MNCs will utilize FDI in the developing world to capitalize on cheaper inputs, such as labor and land, resulting in narrow profit margins. The true profit-generating activities, such as product development and marketing, are often reserved for their home countries, primarily in the developed world. Zeng and Fang specifically examine China and characterize its development strategy (as of their paper in 2014) as “dependent development” due to China’s reliance on FDI and foreign technology. They argue that China’s substantial inflow of FDI could pose challenges for its transition to HI status. According to them, these challenges can manifest in three ways.
First, the type of economic activities that MNCs use China for are low-profit generating in nature, such as low-end commodity manufacturing. For example, as of 2012, Foxconn, a manufacturer of Apple products operating in China, has a profit margin of 1.5 percent, while Apple’s operating margin is more than 30 percent (Culpan, 2012). As of 2020, Apple’s profit margins have risen to over 40 percent, and while Foxconn’s profit margins have increased since 2012, they remain in the low single digits (Ma, 2020). In fact, another study looking a Brazil found that despite large inflows of FDI, FDI’s contributions to the Brazilian economy’s growth have been marginal at best, and its contributions to industrial upgrading are virtually unobservable (Arbix & Laplane, 2003). This is because FDI went towards purchasing already existing assets for discounted prices, rather than investing in new production capabilities.
Zeng and Fang’s (2014) second reason is that FDI increases China’s ‘economic vulnerability,’ as MNCs can move production outside of China as their manufacturing wages rise too high. This puts downward pressure on local workers’ wages, suppressing the creation of a vibrant domestic market. Lastly, high levels of FDI in certain sectors, namely high-technology ones, could suppress the emergence of local firms in those sectors.
Anastasi (2023) contributes to the ongoing theoretical discussion on the relationship between MIT, FDI, and the hierarchical international system. Drawing upon world-systems theory, I have argued that HI countries and their MNCs have incentives to impede the industrial upgrading of MI countries, thereby creating barriers to their progression towards HI status. When MNCs enter developing nations, they often strive to safeguard their technology and prevent knowledge spillovers to local firms. This strategy serves two purposes: firstly, to maintain a quasi-monopolistic hold on these markets by limiting competition, and secondly, to prevent their subsidiary’s input costs from escalating in the developing world.
Similar to the findings of Zeng and Fang (2014), Anastasi (2023) underscores the potential negative consequences of FDI in impeding the emergence of domestic national champions. This viewpoint aligns with the arguments put forth by Doner and Schneider (2016), who contend that the current stock of FDI in MI countries exceeds the levels observed during the corresponding stage of development in previously industrialized nations. While large firms play a crucial role in achieving economies of scale and averting the MIT for MI countries, reliance on MNCs to fulfill this role presents its own set of challenges.
MNCs dominate the manufacturing sectors in MI countries and significantly contribute to their exports of manufactured goods. In some MI countries, such as those in Latin America, MNCs constitute 33 percent to 50 percent of the top 100 firms (Schneider, 2013). This situation creates obstacles to the growth of domestic manufacturing firms while generating distinct, opposing incentives for MNCs and local companies.
Local firms have an interest in improving their country’s educational system and elevating the skill level of potential employees. In contrast, MNCs lack the same incentive since they can attract highly talented workers from the local labor pool by offering higher wages. Moreover, the stance on research and development (R&D) spending varies between MNCs and local firms. Local companies generally have an incentive to allocate some of their profits to R&D, as it can lead to long-term profitability, despite the associated risks. Conversely, MNCs tend to avoid R&D investments in host countries, reserving such endeavors primarily for their home nations. In fact, there is evidence that increases in FDI lower R&D spending in MI countries, creating additional barriers to industrial upgrading and transitioning to HI status (Anastasi, 2024a; Azman-Saini et al., (2010)).
Henley (2018) explores the historical origins and consequences of dependent industrial development in Southeast Asia, focusing on the impact of ethnic capitalism and FDI. Drawing contrasts with Northeast Asian countries like Korea, the paper contends that Southeast Asian nations, marked by ethnic business specialization, face challenges in pursuing non-dependent industrial development. It delves into the political economy of countries like Malaysia and Indonesia, highlighting the complexities arising from the dominance of ethnic Chinese minorities in the private sector. Henley (2018) argues that this phenomenon has hindered cohesive national industrialization efforts and led to a reliance on FDI, which has suppressed indigenous technology creation, particularly in the electronics sector.
Following the studies of Anastasi (2023), Henley (2018), Doner & Schneider (2016), and Zeng & Fang (2014), this paper proposes the following hypothesis:
Hypothesis 1: There is a positive relationship between the level of FDI inflows (relative to GDP) and the likelihood of being trapped in MI status.
Negative relationship between FDI inflows and the MIT
CM et al. (2024) examine the relationship between innovation (defined as labor productivity, the share of the industrial workforce, the percentage share of the population using the internet, the number of patents, and the number of scientific and technical articles) and economic globalization (defined as trade openness, FDI, high-technology exports, and taxes on international trade) on the transition from MI to HI status. Utilizing Bayesian model averaging (BMA) and the generalized method of moments (GMM), the authors found that FDI was positively correlated with transitioning from MI to HI status in both their BMA and GMM models. They also used Cox regressions in a novel fashion to test different variables’ effects on the speed of transition from MI to HI status. The results showed that the time-dependent covariate FDI increased the transition speed, thus reducing the transition duration.
Tampakoudis et al. (2017) aim to contribute to the scholarship on FDI and the MIT. However, the authors acknowledge that, at the time of their writing, “there is no empirical evidence regarding the relationship between FDI inflows and the middle-income trap.”
This study does not attempt to fill the gap that the authors identified; instead they “investigate the effects of certain determinants on FDI inflows to middle-income countries, with respect to avoiding the middle-income trap.” The study itself assumes accepting FDI is a precondition for avoiding MIT. They find that openness and GDP growth are positively correlated with FDI inflows, while population growth is negatively correlated.
Nguyen-Huu and Pham (2021) present interesting results. They test the effect of FDI inflows (among other independent variables) on GDP per capita in MI and LI countries. They find that, in the long run, FDI inflows positively correlate with GDP per capita growth for their full sample and upper MI countries. However, in the short run, for both upper and lower MI countries, FDI inflows negatively correlate with GDP per capita growth. Despite this short-term negative correlation, the long-term positive correlation with GDP per capita growth aligns Nguyen-Huu and Pham (2021) with other studies that found a negative correlation between FDI inflows and being trapped in the MIT.
Following the studies of CM et al. (2024), Tampakoudis et al. (2017), and Nguyen-Huu & Pham (2021), this paper proposes a second Hypothesis:
Hypothesis 2: There is a negative relationship between the level of FDI inflows (relative to GDP) and the likelihood of being trapped in MI status.
Comparative case study: variations in FDI reception in MI countries
The comparative analysis begins by examining Mexico’s foray into the automotive sector, characterized by the dominance of MNCs. Notably, Mexico’s potential national champion, Automex, faced challenges within this competitive landscape. In contrast, the study delves into South Korea’s experience, where the creation of national champions, epitomized by the success story of Hyundai, played a pivotal role. This exploration seeks to understand how FDI contributed to fostering a competitive domestic automotive industry, thereby influencing industrial upgrading and overall development. The chronological scope spans from the early 1960s to the 2000s, but starting a bit earlier for Mexico, providing a comprehensive backdrop for assessing the impact of various FDI models on development and industrial progress within the same industry.
The Mexican automotive industry
The Mexican automotive industry got its start with the creation of a Ford assembly plant in 1925, yet the plant did not become fully operational until 1932. During this time, most automobiles were still imported and industrial policy for the industry was fractured. Yet foreign and domestic capitalists piled into Mexico’s small automotive market due to the number of highways being built, protective tariffs, and the potential size of the market. Before the Mexican state’s first well-formulated attempt at industrial policy in 1962, 12 assemblers and 29 importers made over 100 different models. Consequently, no firm was able to hit economies of scale, which is thought to be at 100,000 units per year (Doner, 1992).
In the year 1960, the Mexican state produced a report, entitled the NAFIN report, on the global automotive industry and steps the state can take to ensure the development of a vibrant domestic industry. The report had a few suggestions including, (i) In order to achieve economies of scale, the state implemented a restriction on the number of assemblers, capping it at four to five. Furthermore, the range of models was also limited, and the lifespan of each model was extended to five years. (ii) Assemblers were required to integrate with local parts and components firms, with the exception of engines. The responsibility for manufacturing parts was exclusively assigned to these local firms. (iii) The standardization of auto parts was introduced, enabling parts and components firms to benefit from economies of scale and supply the same parts to different assemblers. (iv) The concept of “Mexicanization” was promoted, allowing joint ventures between local firms and MNCs, but with local Mexicans in charge of management. This blueprint provided a solid foundation for Mexico’s inaugural attempt at an industrial policy for the automotive industry, offering valuable guidance to the president and his office.
While the industrial policy for the Mexican automotive industry was well crafted, its implementation was less than desired. Despite the report on the automotive industry suggesting that only four firms be allowed to participate in auto assembly, the four largest domestic firms,3 the government allowed many more. Two American MNCs, Ford and GM, and one Japanese MNC, Nissan were allowed to operate. By allowing these MNCs entry, the state had a hard time saying no to more domestic firms attempting to enter. This failure to limit firms in a relatively small market ensured that no firms were able to hit economies of scale.
The inclusion of more than the ideal four firms can be blamed on successful lobbying attempts by Ford. Before the report, the three largest automotive firms in Mexico were two MNCs Ford and GM, and one domestic firm, Automex. Simultaneously with the compilation of the NAFIN report, Ford crafted its own report of the Mexican state concerning the automotive industry’s development. Predictably, Ford’s report advocated a significant role for the company (Back, 1990: p. 278). Adding to this influence, Edgar Molina, Ford’s general manager for its Mexican operations, maintained a close personal friendship with Raúl Saliana Lozano, the head of the Ministry of Industry (SIC). Exploiting this connection, Molina ensured Ford’s approval (Bennett & Sharpe, 1985: pp. 109, 114). Thomas Mann, the U.S. ambassador to Mexico, supported this effort by conveying to the SIC that excluding U.S. firms from the industry would be viewed as an unfriendly act (Bennett & Sharpe, 1985: p. 110). The involvement of MNCs complicated the rejection of other Mexican assemblers. Representaciones Delta, with ties to the president, gained approval after the President explicitly communicated to the SIC that they should be allowed to assemble automobiles regardless of the associated costs (Bennett & Sharpe, 1985: 123). Additionally, Impulsore Mexicana Automotriz and Reo received approval. However, Reo, which assembled Toyotas, went out of business two years later in 1964. Subsequently, the Japanese government pressured the Mexican state to approve Nissan, threatening to halt the import of Mexican cotton if they resisted. Nissan officially entered the Mexican market in 1964, two years beyond the legal deadline (Bennett & Sharpe, 1985: p. 111).
The failure to limit firms ensured that no firms operating in the Mexican market could hit economies of scale, which naturally benefited MNCs, as they had profitable businesses elsewhere in the world.4 This led to the failure of many local assemblers, most notably, Automex, Mexico’s potential national champion. Automex built its business by purchasing complete-knock-down (CKD) kits from Chrysler, an American MNC.5 At the time, Automex was Chrysler’s only exposure to the Mexican market, and refused to leave the market to Ford and GM. As Automex was looking as if it would go under, Chrysler worked with Automoex to form a plan to keep Automex afloat. The proposed plan entailed a merger among Automex, VAM, DINA, and FANASA (all three were stated owned at the time), with the distribution of equity and management rights as follows: Automex would retain one-third of both equity and management rights, Chrysler would acquire a third of the equity, and the Mexican state would also be allocated a third of the equity. The two firms approached the Mexican state with this plan and the state set up a committee to study the effects this plan would have. The committee’s report’s findings indicated that the proposed merger plan held the potential to address inefficiencies prevalent in the industry, specifically stemming from numerous firms producing an excess of models at a scale that was too low. Furthermore, the merger plan was anticipated to bolster the utilization of locally sourced parts, offering a potential resolution to Mexico’s balance-of-payment issue.
During this period, Ford grew apprehensive about the potential implications of the proposed merger plan on its position in the Mexican auto market. Specifically, Ford was concerned about disruptions to its supply chain from the United States. The worry centered around the likelihood of increased local content requirements if the merger plan proceeded. This would compel Ford to acquire more locally sourced parts, potentially forcing price hikes or extending model life due to the perceived lower quality and limited scale of Mexican domestic parts and components firms.
In response, Ford, mindful of Campos Salas’s, the head of Mexico’s SIC, efforts to encourage more exports, approached him with an alternative proposal. Capitalizing on concerns about the balance of payments, Ford devised a plan where exports would offset imports. The proposition stipulated that starting in 1972, firms must export an amount equivalent to their imports to maintain basic production quotas. Those exceeding this requirement could see an increase in their production quotas. This approach allowed Ford to prioritize higher exports over increasing locally sourced parts for its Mexican operations.
In the end, the state chose Ford’s plan over Automex’s merger plan. This led to Automex failing and Chrysler acquiring it. Automex’s failure was the last nail in the coffin for Mexico’s attempt at creating a national champion in its automotive industry. Instead, MNCs and FDI grew to dominate the sector. Though the state did not give up on industrial policy for this sector, the goals of the state shifted from trying to create a Mexicanized automotive industry, to one that could alleviate balance-of-payments concerns of the country, as these were the main targets of 1972, 1977, 1983, and 1989 presidential decrees on the automotive industry. Within these goals, the state had some successes. In regards to industrial upgrading, not so much. The industry could assemble automobiles efficiently and even manufacture engines, yet the true high-value-adding parts of the production process never moved into Mexico, as the MNCs saved that for their home countries.
The 1980s brought about a debt crisis in Mexico and most of Latin America entitled the Latin American Debt Crisis. The Mexican state pursued a strategy of liberalization after this (Morton, 2003). After 1994, the Mexican automotive sector was heavily deregulated, and decision-making power for the industry moved even further away from Mexico City, and even closer to Detroit (Moreno Brid, 1996). According to Chiquiar and Tobal’s (2019) analysis of Mexico’s engagement in Global Value Chains (GVC), the country’s automotive industry held a 6.3 percent share of the international industry’s GVC in 1994, a figure that significantly rose to 27.2 percent by 2017. Notably, there was a strategic shift from manufacturing capital-intensive parts in 1994 to labor-intensive assembly of automobiles in 2017. The transformation was also reflected in employment dynamics. In 1990, Mexico accounted for seven percent of North American automotive jobs, a percentage that surged to 42 percent by 2017 (King’s College London, 2021). However, this influx of jobs was accompanied by a wage squeeze. In 2007, the average wage in the Mexican automotive industry stood at 3.95 USD per hour, which decreased to 3.60 USD in 2013 and further dropped to just 2.30 USD in 2019 (King’s College London, 2021). This decline in wages corresponded with a reduction in Mexico’s domestic value added (Blyde, 2014), both on the aggregate and for the automotive industry in particular, with domestic value added falling by 50 percent between 2003 and 2016 for the automotive industry (King’s College London, 2021).
The involvement of FDI and MNCs in the Mexican automotive industry had a detrimental impact on industrial upgrading. Ford, in particular, vehemently opposed Automex’s success, depriving Mexico of a national champion and diminishing its prospects of advancing toward the higher-value-added segments of automotive production. The neoliberal shift in Mexico resulted in increased production volume within the industry, but a simultaneous decline in value-added components. This shift indicated that, once the state withdrew, MNCs were inclined to exploit Mexico for its relatively cheap labor, reserving the genuinely high-value-adding aspects for their home countries. Consequently, Mexico found itself confined or even downgraded within the division of labor in the automotive industry.
The South Korean automotive industry
South Korea’s automotive industry had its start in 1961 when the newly established interim government led by Park Chung Hee following his military coup in the same year, constructed an automotive assembly plant and passed the Law for the Protection of the Automobile Industry and the Five-Year Plan for the Promotion of the Automobile Industry. Instead of having a foreign firm or a joint venture, a domestic firm, Saenara, was given monopoly rights to assemble automobiles in Korea, and it started importing semi-knock-down (SKD) kits from Nissan to assemble its Bluebird model. For reasons mostly outside of Saenara’s control, the firm went under after assembling just 2,773 automobiles (Yülek et al., 2020). South Korea faced a currency crisis and depleted foreign reserves rendered it impossible for Saenara to continue importing SKD kits from Nissan. Consequently, Saenara faced closure and was ultimately acquired by South Korea’s Ministry of Commerce and Industry (MCI) and Hanil Bank, a state-owned bank.
After Parks’s 1963 presidential victory and the creation of the 3rd Republic of Korea, South Korea had its second attempt at industrial policy for the automotive industry. The following year saw the launch of the Promotion Plan for the Automobile Industry and the search for a new monopolist assembly firm. After intense disagreement within the South Korean state, Sinjin emerged victorious after securing the support of President Park himself (Kim & Lee, 2013). In 1965, Sinjin took over Saenara and started importing SKD kits and assembling Toyota’s Corona model the following year under a licensing agreement with 21 percent of parts being locally produced (Chang, 1985). The local automotive market experienced a surge in demand, but Sinjin maintained low production levels, and reaped significant profits, attracting new players such as Asia Motors and Hyundai Motors (Kim & Lee, 2013).6 These entrants, arguing that Sinjin’s dependence on Toyota hindered the growth of the domestic parts industry, persuaded the government to allow their entry. Consequently, Asia Motors and Hyundai Motors, importing SKD kits from Fiat and Ford, respectively, entered the market.
Increasing the number of assembly firms by 200 percent made the task of building a vibrant local auto parts and components industry much more difficult. A market as small as South Korea split between three assemblers, would not have enough production of a single model of automobile needed to allow the local auto parts firms to hit economies of scale. Understanding this, Park’s government launched the 1969 Basic Promotion Plan for the Automobile Industry. The plan focused on increasing the number of local parts bought by the three assembly firms and building out local engine manufacturing capability. Yet as part of the 1969 plan, local firms would need to partner with MNCs in order to attain foreign engine manufacturing know-how, breaking with South Korea’s previous strategy of simply purchasing foreign technology.
Sinjin, the current market leader, partnered up with General Motors (GM), an American MNC. President Park figured that this joint venture would ensure US defense commitment to South Korea. GM knowing this took advantage of Sinjin and scored a great deal. Both companies would invest 24 million USD and split the ownership 50/50, yet GM also controlled all of the financing, received 750,000 USD a year in management fees, and received three percent of total sales. This resulted in the creation of GM Korea (GMK). Hyundai, as a new entrant in the automotive industry, saw this plan as an opportunity to score a larger market share and leapfrog Sinjin. Hyundai entered negotiations with its partner Ford, yet the two never reached an agreement, leaving Hyundai completely domestically owned.
The 1969 plan was a disaster, as almost none of the goals were met. This failure of industrial policy was also in the context of an economic and political crisis, which led to a regime change (though not a change of government). After declaring a state of emergency, Park pushed through the Yushin (유신) constitution, creating the Fourth Republic of Korea (4RoK). This essentially got rid of any checks and balances from the judiciary or the legislators, and allowed President Park to rule for life (Im, 2013). The idea of ‘politicians ruling and bureaucrats governing,’ still held as the fourth republic was just as technocratic as the third, yet the capacity of the South Korean state was increased dramatically (Kim & Lee, 2013). This change in state structure also brought about a change in economic priorities, namely a heavy industry and chemicals (HCI) drive.
The 4RoK’s HCI drive introduced a new plan for the automotive industry, and in 1973 the Long-Term Plan for the Promotion of the Automobile Industry was launched. The underpinning of this plan was the creation of a ‘citizen’s car,’ a fully domestically designed and made model. Along with requirements for the price, fuel efficiency, restrictions on the number of models, and length of time for which each model could be made, it had the requirement of being made with 95 percent locally sourced parts. The plan was met with mixed reviews from the assemblers. In 1973, GMK was clearly the leader out of the three domestic assemblers, and had no real interest in export promotion, as that could damage GM’s (the MNC) position in the global automobile market. Limiting the number of and extending the production cycle for models, could allow Hyundai and Asia Motors to cut into GMK’s market share, which would run counter to GMK’s interest. Additionally, GMK had little faith in local parts and components firms, as they just purchased these parts from GM’s global network from much more efficient firms, and the local content requirements would put their car at the same level of quality as the other two assemblers (Kim & Lee, 2013). Additionally, GMK complained that the investment requirements were too high, and domestic demand and domestic technology were too low (Yülek et al., 2020).
Hyundai’s reaction was much different. In fact, due to its loose ties with MNCs and its drive to become a world-class brand, the MCI worked with Hyundai’s management team as they crafted the Long-Term Plan. Asia Motors and GMK did in fact increase the amount of locally sourced parts in their assembled automobiles, yet the models they produced were just previous models from their licensors with minor changes. Hyundai had their sights set on creating their own car, and they did.
In 1974 Hyundai introduced a prototype of its first original model, the Pony. By 1976, the Hyundai Pony was ready for the domestic market, and it was an instant hit. The pony was first introduced with about 85–90 percent of locally sourced parts (Yülek et al., 2020). In one year, Hyundai’s market share went from 19 percent to 39 percent. In 1979, it had a majority market share of 51 percent (Hyundai Jadongcha Isipnyeonsa, 1987). Hyundai’s success and rising incomes within Korea led to a large increase in auto production.
The development of the Pony was a large task, as the technology levels of Hyundai and the South Korean economy were quite low at the time. Hyundai was able to acquire foreign technology without becoming a pawn of MNCs. They did this through licensing agreements and diversifying where technology came from, Hyundai’s pony was designed by an Italian car maker, a design for an engine from the U.K., and other important parts designed by the Japanese. Also by hiring foreign talent, in 1974 Hyundai hired George Turnbull, an automobile maker executive from the U.K. and he brought with him a team of engineers from the U.K. Lastly, they struck a mutually beneficial deal with Japan’s Mitsubishi. Mitsubishi, lagging behind its Japanese competitors, saw a partnership with Hyundai as an opportunity to leapfrog them (Kim & Lee, 2013).
Soon after Hyundai found success in its home market, it turned its sights to overseas markets, namely in the developing world. However, South Korea was dragged into both a political and economic crisis following the 1979 assassination of Park Chung Hee. After a small window of political openness, Chun Doo-hwan took control of power via a coup and immediately set up an interim government. The interim government was keen on consolidation within the automotive industry. In August and September of 1980, GM (the foreign owner of GMK) and Hyundai held six meetings, yet hit a dead-end over the issues of exports and equity ratio. Hyundai wanted GM to control only 20 percent of equity and have no management rights, while GM wanted 50 percent of equity and veto power over management decisions. GM also wanted to incorporate Hyundai into its global network of suppliers, while Hyundai wanted to become an international brand. GM wanted to turn Hyundai into something it fought hard and took risks to make sure it would not become, a mere pond of an MNC.
After Chun Doo-hwan established the Fifth Republic of Korea, the merger plan for the automotive industry was abandoned. Hyundai entered the United States market in 1984 and was able to compete with MNCs in their own backyard. This success was due to Hyundai’s competitiveness and a bit of luck. The United States and Japanese trading relations took a bit of a dive, creating an opening for cheap passenger automobiles, which Hyundai filled. Many developed economies’ automakers started to focus on larger, more expensive vehicles, which allowed Hyundai to excel in the small passenger vehicle market. Additionally, Hyundai strengthened its relationship with Mitsubishi. Hyundai got to access Mitsubishi’s global marketing network and some of Mitsubishi’s superior technology, while Mitsubishi saw this as an opportunity to leap-frog its domestic competitors within the US market by taking a 10 percent stake in Hyundai.
At the same time, Hyundai began investing heavily in R&D in order to start developing its own indigenous technology. Starting from 23.8 billion won in 1982 and by 1987, Hyundai’s R&D spending hit 467 billion won, which accounted for 72 percent of the South Korean automotive industry’s total R&D spending in 1982 and 1987, 50 percent (Yülek et al., 2020). Hyundai continues to excel and continues to use more and more of its indigenous technology rather than licensed technology for its automobiles.
The Korean automotive industry took another hit during the 1997 Asian Financial Crisis, yet was able to return to growth. Currently, South Korea’s automotive industry is one of the world’s biggest and most developed. The country is the fifth-largest auto producer in the world and the sixth-largest exporter of automobiles (Yoon, 2023).
A comparison of the two
The South Korean and Mexican automotive industries developed on different trajectories, and this is, in part, due to their different models of accepting FDI. South Korea’s industry has developed into one of the world’s largest and most advanced. The Mexican automotive industry did in fact develop into something laudable. Currently, Mexico’s automotive industry is the sixth-largest producer of passenger automobiles in the world, just below South Korea (International Trade Administration, 2023). In 2019, Mexico produced 3.8 million units, most of which were exported (Carlier, 202). On average, 89 percent of automobiles produced in Mexico are exported to the rest of the world, yet a majority of them are exported to the United States, as 80 percent of all automobiles produced in Mexico are exported there (International Trade Administration, 2023). Mexico does not have any national champions, so most of the automotive industry is linked to global supply chains, namely for North America, with ‘The American Big Three’ (Ford, GM, and Chrysler), playing a large role. This means that FDI plays a huge role in the Mexican automotive industry, making up 5 billion USD a year on average (Carlier, 2023).
Despite this notable success, it has been unable to move up the division of labor in the industry and is stuck to performing relatively low-value-adding economic activities. This is clearly due to the level of FDI and the number of MNCs Mexico allowed to operate within its automotive industry. One of the largest challenges when first developing this industry is that of economies of scale, consequently, large firms, or national champions, are needed to overcome this obstacle. We saw that Hyundai was able to grow into a large firm by purchasing foreign technology and keeping an arm’s length away from MNCs. In Mexico, Ford was able to ensure that Automex, Mexico’s potential national champion, went under. This is due to its superior market power, which came from profits it earned elsewhere in the world, and its ability to co-op industrial policy, again due to its profits it earned elsewhere.
It would be irresponsible to assume that if Mexico never allowed such large amounts of FDI, its automotive industry would have developed much like South Korea’s. However, without Ford, the driving force that brought about Automex’s destruction would not have been present, giving Mexico a greater chance of creating a national champion within its automotive industry. Taking a step back, if this trend of FDI negatively affecting industrial upgrading in MI countries continues, then we should see similar results in the quantitive models.
Data and research design
Data
The data collected for the four models built in this study come from the World Bank’s Databank website. Four breakout cases: South Korea, Uruguay, Greece, and Japan, and four countries that are trapped: Mexico, Turkey, Malaysia, and Brazil were chosen.7 The years for this data span from 1970 to 2000. The starting year was chosen due to the availability of data concerns, and the last year was chosen to minimize the years when the breakout cases were at HI status.
Estimation methods
The question of whether a country has or has not advanced toward HI status results in a simple, binary answer; consequentially a logistic regression is best suited for this conundrum. This study employs a cross-sectional approach. Although the data covers multiple years, the MIT outcomes are constant over time for each country, with breakout cases coded as 1 and trapped cases coded as 0 for the entire period. This approach allows us to capture variations between countries and different levels of FDI inflows relative to GDP, enhancing the robustness of our findings regarding the impact of FDI inflows on overcoming the MIT. Descriptions of the variables used can be found in Table 1, and descriptive statistics and the variable’s Pearson’s correlations can be found in Table 2.
Table 1. Description of variables.
Variables | Symbols | Description | Reference | observations |
|---|---|---|---|---|
Middle-income trap outcomes | MITθ | Binary choice variable where avoiding the middle-income trap is coded as “1”, and failing to advance to high-income status is coded as “0”. | The World Bank; World Development Indicators - Coded by Author (2023) | 248 |
Foreign direct investment, net inflows (% of GDP) | FDI%G | Foreign direct investment are the net inflows of investment to acquire a lasting management interest (10 percent or more of voting stock) in an enterprise operating in an economy other than that of the investor. It is the sum of equity capital, reinvestment of earnings, other long-term capital, and short-term capital as shown in the balance of payments. This series shows net inflows (new investment inflows less disinvestment) in the reporting economy from foreign investors and is divided by GDP. | The World Bank; World Development Indicators | 248 |
Gross capital formation (% of GDP) | GCF | Gross capital formation consists of outlays on additions to the fixed assets of the economy plus net changes in the level of inventories. Fixed assets include land improvements (fences, ditches, drains, and so on); plant, machinery, and equipment purchases; and the construction of roads, railways, and the like, including schools, offices, hospitals, private residential dwellings, and commercial and industrial buildings. Inventories are stocks of goods held by firms to meet temporary or unexpected fluctuations in production or sales, and "work in progress." According to the 1993 SNA, net acquisitions of valuables are also considered capital formation. | The World Bank; World Development Indicators | 248 |
Age dependency ratio (% of working-age population) | ADR | Age dependency ratio is the ratio of dependents--people younger than 15 or older than 64--to the working-age population--those ages 15–64. Data are shown as the proportion of dependents per 100 working-age population. | The World Bank; World Development Indicators | 248 |
Exports of goods and services (% of GDP) | E | Exports of goods and services represent the value of all goods and other market services provided to the rest of the world. They include the value of merchandise, freight, insurance, transport, travel, royalties, license fees, and other services, such as communication, construction, financial, information, business, personal, and government services. They exclude compensation of employees and investment income and transfer payments. | The World Bank; World Development Indicators | 248 |
Manufacturing, value added (% of GDP) | MVA | Manufacturing refers to industries belonging to ISIC divisions 15–37. Value added is the net output of a sector after adding up all outputs and subtracting intermediate inputs. It is calculated without making deductions for depreciation of fabricated assets or depletion and degradation of natural resources. The origin of value added is determined by the International Standard Industrial Classification (ISIC), revision 3. Note: For VAB countries, gross value added at factor cost is used as the denominator. | The World Bank; World Development Indicators | 186 |
Source: The World Bank DataBank; World Development Indicators.
Table 2. Descriptive statistics and Pearson’s correlations.
Variables | Mean | SD | Max | Min | 1 | 2 | 3 | 4 |
|---|---|---|---|---|---|---|---|---|
1) FDI%G | 1.116 | 1.501 | 8.760 | −0.167 | ||||
2) GCF | 25.972 | 7.889 | 48.284 | 3.396 | 0.131 | |||
3) ADR | 64.283 | 15.047 | 102.018 | 38.937 | 0.173 | −0.452 | ||
4) E | 21.832 | 20.082 | 121.311 | 3.218 | 0.715 | 0.263 | −0.021 | |
5) MVA | 21.654 | 4.945 | 34.271 | 8.989 | 0.041 | 0.246 | −0.151 | 0.245 |
None of the correlation coefficients takes a considerably high value (<0.9), suggesting that multicollinearity, generally, is not a problem.
Note: Numbers in bold are Pearson’s r.
Model construction
Figure 1 simply shows the levels of FDI%G of the countries that advanced to HI status (1) and those trapped at MI status (0). Figures 2 and 3 are the logistic regression curve and the marginal effects plot, respectively, of a simple logistic regression between MITθ (dependent variable) and FDI%G (independent variable). All figures show a clear negative relationship between acceptance of FDI%G and MITθ, being the advancement to HI status. Model 1 is written as:
Fig. 1 [Images not available. See PDF.]
FDI’s effect on transitioning from MI to HI status.
Notes: The y-axis is MIT outcomes (0:trapped, 1:not trapped) and the x-axis is foreign direct investment inflows (% of GDP).
Fig. 2 [Images not available. See PDF.]
Logistic regression curve.
Notes: The y-axis is the probability of MIT outcomes and the x-axis is foreign direct investment inflows (% of GDP).
Fig. 3 [Images not available. See PDF.]
Marginal effects plot.
Notes: The y-axis is MIT outcomes and the x-axis is foreign direct investment inflows (% of GDP).
To further test this relationship, three additional models were built. Model 2 adds two additional variables that have been argued to be correlated with development, the ADR and GCF.8 Model 2 is written as:
The third model adds E and MVA to simulate the characteristics of the developmental state, which has been credited with enabling the quick, robust development found in East Asia (Haggard, 2018). Model 3 is written as:
The fourth model combines all the control variables, which is written as:
The results of the models can be found in Table 3, and the 95 percent confidence intervals of the odds ratios for each variable in the models can be found in Table 4.
Table 3. FDI inflows’ effects on middle-income trap outcomes of select countries (1970–2000).
Variables | MITθ (Model 1) | MITθ (Model 2) | MITθ (Model 3) | MITθ (Model 4) |
|---|---|---|---|---|
FDI%G | −1.026*** (−5.280) [0.359] | −0.866*** (−4.192) [0.420] | −2.119*** (−5.521) [0.120] | −1.521*** (−3.532) [0.219] |
GCF | 0.054 (1.691) [1.055] | 0.014 (0.359) [1.014] | ||
ADR | −0.208*** (−6.639) [0.812] | −0.158*** (−5.672) [0.854] | ||
E | 0.071*** (4.462) [1.074] | 0.034 (1.768) [1.035] | ||
MVA | −0.024 (−0.615) [0.976] | −0.079 (−1.397) [0.924] | ||
Observations | 248 | 248 | 186 | 186 |
McFadden R-squared | 0.159 | 0.607 | 0.253 | 0.548 |
Nagelkerke R-squared | 0.264 | 0.759 | 0.383 | 0.698 |
AIC | 293.090 | 142.951 | 184.858 | 118.945 |
BIC | 300.117 | 157.005 | 197.761 | 138.299 |
Notes: Numbers in brackets are odds ratios, numbers in parentheses are z-values, and numbers not in parentheses or brackets are coefficient estimates. The asterisks *** indicate a significance level (p-value) at <0.001.
Source: The World Bank DataBank; World Development Indicators.
Table 4. 95 Percent confidence intervals of the odds ratios for models 1–4.
Variables | MITθ (Model 1) | MITθ (Model 2) | MITθ (Model 3) | MITθ (Model 4) |
|---|---|---|---|---|
FDI%G | 0.24 0.51 | 0.27 0.60 | 0.05 0.24 | 0.08 0.45 |
GCF | 0.99 1.13 | 0.94 1.10 | ||
ADR | 0.76 0.86 | 0.80 0.90 | ||
E | 1.04 1.11 | 0.99 1.08 | ||
MVA | 0.90 1.05 | 0.82 1.03 | ||
Observations | 248 | 248 | 186 | 186 |
Notes: The top numbers are the lower end of the 95 percent confidence interval, and the bottom numbers are the higher end.
Source: The World Bank DataBank; World Development Indicators.
Robustness test
To test the robustness of these four models, this study deployed four different subsample analysis tests, dividing the data by region (Table 5). The initial test mirrored the first model of this study, using only one independent variable (FDI%G), but focused solely on countries in East Asia (South Korea, Japan, and Malaysia). FDI%G continued to show significance in this context.
Table 5. Robustness test: subsample analysis.
Variables | East Asia MITθ (Test 1) | East Asia MITθ (Test 2) | Latin American MITθ (Test 3) | Latin American MITθ (Test 4) |
|---|---|---|---|---|
FDI%G | −2.274*** (−4.543) [0.103] | −1.818*** (−3.595) [0.162] | −1.118** (−2.754) [0.327] | −1.465** (−2.964) [0.494] |
GCF | 0.057 (0.362) [1.058] | −0.011 (−0.069) [0.989] | ||
ADR | −0.079 (−1.568) [0.924] | −0.794** (−2.709) [0.452] | ||
Observations | 93 | 93 | 93 | 93 |
McFadden R-squared | 0.747 | 0.804 | 0.096 | 0.736 |
Nagelkerke R-squared | 0.852 | 0.890 | 0.160 | 0.844 |
AIC | 33.948 | 31.207 | 111.045 | 39.305 |
BIC | 39.013 | 41.338 | 116.110 | 49.435 |
Notes: Numbers in brackets are odds ratios, numbers in parentheses are z-values, and numbers not in parentheses or brackets are coefficient estimates. The asterisks ***, ** indicate the significance level (p-value) at <0.001 and 0.01, respectively.
Source: The World Bank DataBank; World Development Indicators.
The second subsample test also used East Asian countries but added ADR and GCF, similar to the second model of this study. FDI%G remained significant; interestingly, ADR, which was significant in Models 2 and 4, appeared nonsignificant in this test. Due to data availability concerns, the third and fourth models of this study were not tested.
The third test replicated Model 1 (and Test 1) but examined only Latin American countries (Mexico, Brazil, and Uruguay). FDI%G was significant, although slightly less so, with a p-value of 0.006. The fourth and final test mirrored Model 2 (and Test 2) but focused on observations from Latin American countries. Similar to Test 3, FDI%G was significant, though less so than in the first two tests (and the four models), with a p-value of 0.003.
Empirical results
Descriptive statistics and Pearson’s correlations
The mean values and standard deviations of key variables are as follows: FDI%G (1.116, 1.501), GCF (25.972, 7.889), ADR (64.283, 15.047), E (21.832, 20.082), and MVA (21.654, 4.945). These indicate moderate variability in FDI%G, significant differences in GCF, a broad range in ADR, considerable disparity in E, and moderate variability in MVA. Pearson’s correlation coefficients show: FDI%G has slight positive correlations with GCF (0.131) and ADR (0.173), moderate positive correlation with E (0.715), and weak correlation with MVA (0.041). GCF has a negative correlation with ADR (−0.452) and positive correlations with E (0.263) and MVA (0.246). ADR has weak negative correlations with E (−0.021) and MVA (−0.151). E has a moderate positive correlation with MVA (0.245). None of the correlation coefficients are considerably high (<0.9), indicating no significant multicollinearity. This suggests that the independent variables are not highly correlated, ensuring more reliable regression analysis results.
Model 1
The initial model, which considers only FDI%G as a predictor, revealed a significant negative association. The odds ratio (OR) for FDI%G was 0.36 (95% CI: 0.24–0.51), indicating that a one-unit increase in FDI%G is associated with a 64% decrease in the odds of positive MITθ. An OR of 0.36 means that for each one-unit increase in FDI%G, the odds of transitioning to HI status decreased by 64% (since 1–0.36 = 0.64). This is a substantial effect, highlighting that FDI inflows might be counterproductive to achieving economic advancement in the context of MI countries.
The 95% confidence interval for the OR (0.24–0.51) indicates that the true effect of FDI%G on the odds of escaping the MIT is highly unlikely to be greater than 0.51 or less than 0.24. This range suggests robustness in the negative association between FDI%G and positive MITθ, providing strong evidence that higher FDI levels are consistently associated with lower odds of transitioning to HI status. This model’s goodness of fit is reflected in the McFadden R² (0.16) and Nagelkerke R² (0.20).
Model 2
Expanding the model to include GCF and the ADR introduced new insights. FDI%G remained negatively associated with MITθ, with an odds ratio (OR) of 0.42 (95% CI: 0.27–0.60). This indicates that for each one-unit increase in FDI%G, the odds of transitioning to HI status decrease by 58% (since 1 − 0.42 = 0.58). The consistency of the negative association across models reinforces the robustness of the finding that higher FDI inflows are detrimental to achieving economic advancement.
The ADR also exhibited a significant negative association with MITθ, with an OR of 0.81 (95% CI: 0.76–0.86). This means that for each one-unit increase in the ADR, the odds of transitioning to HI status decreased by 19% (since 1 − 0.81 = 0.19). This suggests that higher ADRs, indicative of a larger proportion of dependents (both young and old) relative to the working-age population, hinder economic progression to HI status. The inclusion of GCF in the model did not yield a significant association with MITθ, implying that variations in gross capital formation do not substantially impact the likelihood of escaping MIT in this context.
The goodness of fit for Model 2, as indicated by the Akaike Information Criterion (AIC) of 142.95 and the Bayesian Information Criterion (BIC) of 157.01, suggests a good balance between model fit and complexity. The lower AIC and BIC values relative to Model 1 indicate improved model performance while maintaining parsimony. The pseudo-R² values for Model 2 (McFadden: 0.60, Nagelkerke: 0.76) show a substantial increase compared to Model 1, reflecting a better explanatory power of the model with the additional predictors. These values suggest that the model explains a significant portion of the variance in the outcome variable, providing a more comprehensive understanding of the factors influencing the transition to HI status.
Model 3
Incorporating E and MVA alongside FDI%G demonstrated that FDI%G retained a negative association (OR = 0.12, 95% CI: 0.05–0.24). This indicates that for each one-unit increase in FDI%G, the odds of transitioning to HI status decrease by 88% (since 1 − 0.12 = 0.88).
E exhibited a positive association with the odds of transitioning to HI status, with an OR of 1.07 (95% CI: 1.04–1.11). This means that for each one-unit increase in E, the odds of transitioning to HI status increase by 7%. This positive association suggests that higher export levels contribute to economic progression, potentially by fostering greater economic activity and integration into global markets. MVA did not show significance.
The model exhibited a reasonable fit (AIC: 184.86, BIC: 197.76). These criteria suggest that the model balances fit and complexity well, providing a robust framework for understanding the factors influencing the transition to HI status.
Model 4
The final model included all predictors, and FDI%G continued to exhibit a negative association (OR = 0.22, 95% CI: 0.08–0.45). This suggests that a one-unit increase in FDI%G is associated with a 78% decrease in the odds of moving to HI status (since 1 − 0.22 = 0.78). Notably, the ADR retained its significant associations. The model’s fit was well-balanced, as indicated by the AIC of 118.94 and the BIC of 138.30. These values suggest that the model effectively captures the relationship between the predictors and the outcome without excessive complexity. The relatively high pseudo-R² values (McFadden: 0.55, Nagelkerke: 0.70) further indicate that the model provides a good fit to the data, capturing a substantial proportion of the variability in the likelihood of transitioning to HI status.
Discussion
The consistent negative correlation between FDI and the likelihood of transitioning to HI status across all models supports the first hypothesis, allowing us to reject the first null hypothesis. Higher levels of FDI appear to hinder a country’s ability to escape the MIT. However, the results seem to counter the second hypothesis, indicating that we cannot reject the second null hypothesis.
The AIC and BIC values offer valuable insights into model fit and complexity. Lower AIC and BIC values across subsequent models suggest improved goodness of fit without an unwarranted increase in complexity. Thus, the selected models strike a balance between explanatory power and parsimony, enhancing the robustness of the findings. These results collectively contribute to a nuanced understanding of the factors influencing MITθ and the role FDI plays in the transition from MI to HI status.
The findings from the qualitative comparative case study, specifically that FDI played a negative role in the formation of Mexico’s automotive industry and hindered its transition from MI to HI status, were confirmed by the results of the multiple logistic regression models. This finding is particularly significant in the context of the global economic landscape, where FDI is often considered a key driver of economic growth (Bénétrix et al., 2023).
The results of both the qualitative and quantitative studies have several theoretical implications. First, they run counter to CM et al. (2024) and Tampakoudis et al. (2017). Second, they confirm the theoretical studies of Anastasi (2023), Henley (2018), Doner & Schneider (2016), and Zeng & Fang (2014), validating the claim by Anastasi (2023), Zeng & Fang (2014) and World-Systems theorist that the free movement of capital seems to harm developing countries’ prospects of development rather than aid it. Finally, these findings open up new questions, particularly regarding the mechanism by which FDI harms the host countries’ transition from MI to HI status.
As for the exact mechanism by which higher levels of FDI lead to a lower chance of advancing to HI status, future research is needed. However, based on this study’s comparative case study, it seems that FDI prevents the materialization of large domestic firms capable of performing high-value-adding economic activities within the host country. This was evident in the case of the Mexican automotive industry, where MNCs, through both economic (increased competition) and political (successful lobbying to influence industrial policy) fronts, prevented Automex from succeeding. This stands in contrast to South Korea’s model of FDI acceptance, where Hyundai grew into a large firm capable of performing the highest value-adding economic activities within the automotive industry.
Beyond the primary independent variable of this study, FDI%G, there were some interesting findings with the other control variables. GCF was not found to be significant in any model it was included in (Models 2 and 4), nor was it significant in the subsample analysis tests. ADR was significant in Models 2 and 4 but was insignificant in the subsample analysis test focused solely on East Asia. E was significant in Model 3, but once other control variables were added in Model 4, it was not significant. Lastly, MVA was not found to be significant in any model.
Implications
In 2021, according to the World Bank (2023), MI countries received $754 billion of FDI inflows, which represented 2.1% of their GDP. While this is a substantial amount, the results of this mixed-method study suggest that MI countries should exercise caution in accepting FDI.
It would be unwise for MI countries to completely shut down FDI inflows; instead, they should adopt a more balanced approach to advance towards HI status.
If the primary mechanism by which FDI traps MI countries at MI status is indeed the suppression of large domestic firms capable of performing high-value-adding economic activities, then MI countries should be selective about the FDI they allow or accept FDI in a way that promotes industrial upgrading. China’s automotive industry provides a useful example. Starting in 1994, China crafted an industrial policy for the automotive sector that included MNCs and sought to attract FDI through joint ventures (JVs), where MNCs partnered with local firms.9 With state support in negotiating technology transfers, China’s young domestic firms entered the market and moved towards higher-value-adding economic activities (Chin, 2018). Today, Chinese automakers are world-class in terms of electric vehicles, showcasing how properly managed FDI can enable industrial upgrading.
However, not every MI country has a market the size of China, which might make MNCs reluctant to transfer technology merely to enter a smaller market. In such cases, MI country governments should strategically decide which sectors to allow FDI. For higher-value-adding industries, whose creation is essential for avoiding the MIT, MI countries should initially aid local firms in purchasing foreign technology and subsequently support the development of indigenous technology, much like the strategy employed by South Korea (Anastasi, 2024b: p. 116).
From a global economic governance perspective, the global trading system and the system of global flows of capital are broken. This harms both developed countries, such as the United States (Klein & Pettis, 2021; p. 190) and developing countries alike. The implications of the study imply that capital flows into the MI world, in the form of FDI are harmful to industrial upgrading, economic advancement, and thus making the transition to HI status. As stated above there are some domestic policy implications (such as the use of JVs, being selective as to which sectors of the economy can welcome FDI inflows, and obtaining foreign technology through technology purchases and licensing agreements), although there are some global economic governance implications. MI should attempt to bring their domestic laws in line surrounding FDI inflows in a manner that increases the likelihood of technology transfers (namely in regards to JVs and intellectual property rights) and force a global norm change regarding FDI. This could stop a potential “race-to-the-bottom” in regards to laws governing FDI inflows within the MI world. We have seen something like this in the case of the “Global Minimum Corporate Tax” (GMCT) (Partington, 2021), which major countries agreed a few years back. This could serve as a model for a new international structure to govern FDI inflows into the MI world.
Conclusion
While the concept of MIT has received well-deserved attention from scholars during its almost two decades in existence, some large questions about it remain. One of these is the effect FDI has on the transition from MI to HI status. Through a mixed-method approach, this research attempted to fill this gap.
In examining the trajectories of the automotive industries in Mexico and South Korea, this comparative case study reveals nuanced insights into the role of FDI in shaping industrial development. The Mexican automotive industry, heavily reliant on FDI, found itself entwined with MNCs, facing challenges in achieving economies of scale and failing to create national champions. In contrast, South Korea strategically navigated its industrial policy, fostering the emergence of successful domestic players like Hyundai.
The Mexican experience highlights the complexities of managing FDI and underscores the impact of MNCs’ influence on policy decisions. The failure to limit the number of firms and the dominance of MNCs hindered industrial upgrading, leaving Mexico performing relatively low-value-adding economic activities. The shift in industrial policy focus from creating a Mexicanized industry to addressing balance-of-payments concerns marked a significant turning point. Conversely, South Korea’s journey showcases the potential for strategic policy interventions to foster a vibrant domestic industry. The emphasis on a “citizen’s car” with high local content, successful collaborations with MNCs, and Hyundai’s innovative approaches allowed South Korea to become a major global player. The ability to balance domestic priorities with international partnerships contributed to South Korea’s position as one of the world’s largest and most advanced automotive industries.
The comparative analysis underscores the critical importance of policy choices and the nature of FDI acceptance models in determining industrial outcomes. While Mexico achieved success in becoming a major automotive producer, it struggled to move up the division of labor and remained dependent on global supply chains. In contrast, South Korea’s strategic policies facilitated the creation of a robust industry with substantial technological advancements.
The logistic regression models, analyzing the relationship between FDI and a country’s likelihood of advancing to HI status, consistently demonstrated a negative association. Higher levels of FDI appear to hinder a country’s ability to escape the MIT. The inclusion of control variables provided a nuanced understanding of the factors influencing outcomes, emphasizing the delicate interplay between FDI and other economic indicators. While the results of this study run counter to some of the earlier ones (CM et al., 2024; Tampakoudis et al., 2017), it validated the claim by World-Systems theorists (Anastasi, 2023; Zeng & Fang, 2014), that the free movement of capital seems to harm developing countries’ prospects of development rather than aid it.
The limitations of this study are that in cases such as the Chinese automotive industry, we have indeed seen FDI inflows lead to industrial upgrading (Chin, 2018). While on the aggregate, FDI inflows seem to pose a threat to industrial upgrading thus avoiding the MIT in terms of MI countries, there seem to be cases that do no comfort to the results of this study. Future research should investigate why this is so, to better understand under which conditions can FDI aid development and industrial upgrading in MI countries.
In conclusion, the experiences of Mexico and South Korea in the automotive sector and the logistic regression models point to the fact that states should exercise caution while accepting FDI. States should devise domestic laws governing FDI and MNCs to facilitate technology transfers to local firms or opt to push domestic firms to purchase foreign technology through licensing agreements, avoiding FDI altogether. Careful policy considerations and strategic planning that empower the creation of national champions and indigenous technology development are essential for ensuring long-term success. The findings of this study contribute to the broader discourse on the impact of FDI on industrialization and provide valuable insights for policymakers and scholars alike.
Author contributions
All the work of this study was performed by the sole author.
Data availability
The data used in the quantitative part of this study can be found on the World Bank Databank’s website or alternatively can be made available by requesting it from the author. The qualitative data can be found through in-text citations.
Competing interests
The author declares no competing interests.
Ethical approval
Ethical approval was not required as the study did not involve human participants. This article does not contain any studies with human participants performed by the author.
Informed consent
This article does not contain any studies with human participants performed by the author.
These break-cases are; Israel, Japan, Ireland, Spain, Hong Kong, Singapore, Portugal, Taiwan, Mauritius, Equatorial Guinea, South Korea, Greece, and Puerto Rico. The number of HI countries has since changed a bit from the time of the writing of the World Bank’s report. New entries mostly include post-Soviet states, island states, and Chile and Uruguay in South America. Equatorial Guinea and Mauritius, two of the 13 break-out cases, have since declined in terms of GDP per capita and have been downgraded to MI status.
2For a comprehensive look at automotive industry building in the developing world, please see Doner (1992).
3The ideal four firms would be: Automex, DINA, Promexa, and VAM.
4In fact, one study shows that between the years 1966 to 1972, MNCs’ profits were about 16.5% higher than local firms within the Mexican automotive industry (Jenkins, 1997).
5Complete-knock-down kits are kits that contain most, if not all, of the necessary parts needed to assemble automobiles. They differ from semi-knock-down kits, as complete-knock-down kits are completely unassembled, while semi-knock-down kits come with some of the parts already assembled to some degree.
6Asia Motors was bought out by Kia in 1976, yet will be referred to as Asia Motors for the sake of consistency.
7Trapped countries are according to Felipe et al.’s (2012) definition of being trapped in the MIT.
8See Lawton (2019) for the age dependency ratio’s link to economic growth in developing countries and Topcu et al. (2020) for the link between gross capital formation and economic growth.
9This was the 1994 Automotive Industry Policy (AIP), or the 1994版《汽车工业产业政策》. The 1994 AIP details were; (i) local firms could only partner with world-leading automotive MNCs, (ii) the MNCs could not have more than a 49 percent stake in the joint venture, (iii) the MNCs could not be involved in the distribution and sales of the automobiles, as that would be the sole responsibility of the local firm, (iv) at least 40 percent of automotive’s part should be bought locally, and the higher the local content level, the lower the import duties would be for the rest of the imported parts, (v) local firms are allowed to set up joint ventures with multiple MNCs (yet this was there to increase the local firm’s power within their relationship with the MNC), and (vi) the MNC should perform R&D in China.
Supplementary information
The online version contains supplementary material available at https://doi.org/10.1057/s41599-024-03662-6.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Anastasi, AWD. Foreign direct investment’s effect on Research & Development Spending in middle-income countries. J Polit Law; 2024; 17,
Anastasi AWD (2024b). How to avoid the middle-income trap: a state-business relations and state capacity explanation—a case study of South Korea and Mexico [Thesis]. Researchgate. Retrieved May 2024, from https://www.researchgate.net/publication/380940772_How_To_Avoid_The_Middle-income_Trap_A_State-Business_Relations_and_State_Capacity_Explanation_-_A_Case_Study_of_South_Korea_and_Mexico
Anastasi, AWD. Trapped in the semi-periphery: understanding the middle-income trap from a world-systems theory perspective. J World-Syst Res; 2023; 29,
Arbix, G; Laplane, M. Stagnation, liberalization and foreign investment in Latin America. Compet Change; 2003; 7,
Azman-Saini, WNW; Law, SH; Ahmad, AH. FDI and economic growth: New evidence on the role of financial markets. Econ Lett; 2010; 107,
Back JG (1990) Politics of late industrialization: the origins and processes of automobile industry policies in Mexico and South Korea. Ph.D. thesis, University of California
Bénétrix A, Pallan H, Panizza U (2023) The elusive link between FDI and economic growth. Policy Research Working Papers. https://doi.org/10.1596/1813-9450-10422
Bennett, DC; Sharpe, KE. Transnational corporations versus the state: The political economy of the Mexican auto industry; 1985; Princeton, NJ, Princeton University Press: [DOI: https://dx.doi.org/10.1515/9781400857807]
Blyde, JS. The participation of Mexico in global supply chains: the challenge of adding Mexican value. Appl Econ Lett; 2014; 21,
Chang D-J (1985) Economic control and political authoritarianism: the role of Japanese corporations in Korean politics, 1965–1979. Sogang University, Seoul. pp. 118–132 https://books.google.com/books/about/Economic_Control_and_Political_Authorita.html?id=0H-3AAAAIAAJ
Carlier M (2023) Topic: automotive industry in Mexico. Statista. Available at: https://www.statista.com/topics/7249/automotive-industry-in-mexico/#dossierKeyfigures. Accessed: 04 Dec 2023
Chenery, HB; Taylor, L. Development patterns: among countries and over time. Rev Econ Stat; 1968; 50,
Chiquiar D, Tobal M (2019) Global value chains in Mexico: a historical perspective. Banco de México. Available at: https://www.banxico.org.mx/publications-and-press/banco-de-mexico-working-papers/%7B0E644F45-DB5A-077B-3C4C-81E98A43AD6F%7D.pdf
Chin G (2018) The evolution of government–MNC relations in China: the case of the automotive sector. In: Zhu T & Zhang X (eds.). Business, government and economic institutions in China. Springer International Publishing. pp. 81–104
CM, J; Hoang, NT; Yarram, SR. The transition from middle-income trap: role of innovation and economic globalisation. Appl Econ; 2024; 56,
Culpan T (2012) Apple profit margins rise at Foxconn’s expense: Chart of the day. Bloomberg.com. Available at: https://www.bloomberg.com/news/articles/2012-01-04/apple-profit-margins-rise-at-foxconn-s-expense. Accessed: 08 Jun 2023
Doner, RF. Limits of state strength: toward an institutionalist view of economic development. World Polit; 1992; 44,
Doner, RF; Schneider, BR. The middle-income trap. World Polit; 2016; 68,
Felipe J, Abdon A, Kumar U (2012) Tracking the middle-income trap: what is it, who is in it, and why? SSRN Electronic J. https://doi.org/10.2139/ssrn.2049330
Gill I, & Kharas H (2007). 10 Years after the Crisis. East Asia and Pacific Economic Update, The World Bank. http://documents.worldbank.org/curated/en/679431468038153232/10-years-after-the-crisis
Gill I, Kharas H (2015). The middle-income trap turns ten. Policy Research Working Papers. https://doi.org/10.1596/1813-9450-7403
Haggard, S. Developmental states; 2018; Cambridge, United Kingdom, Cambridge University Press: [DOI: https://dx.doi.org/10.1017/9781108552738] ISBN: 9781108449496
Henley, D. Foreign investment and the middle-income trap in Southeast Asia. Lembaran Sej; 2018; 14,
Hyundai Jadongcha Isipnyeonsa (1987) HMC. Appendix, 60, Seoul
Im HB (2013) The origins of the Yushin regime: Machiavelli unveiled. In: The Park Chung Hee Era: the transformation of South Korea. Harvard University Press, Cambridge, MA. pp. 233–261. http://www.tinyurl.com/y4nsjoy6
International Trade Administration (2023) Mexico—automotive industry. International Trade Administration | Trade.gov. Available at: https://www.trade.gov/country-commercial-guides/mexico-automotive-industry. Accessed 04 Dec 2023)
Jenkins RO (1977) Dependent industrialization in Latin America: The automotive industry in Argentina, Chile and Mexico. Praeger. https://books.google.com/books/about/Dependent_Industrialization_in_Latin_Ame.html?id=Yh5GAAAAYAAJ
Klein MC, Pettis M (2021) Trade wars are class wars: how rising inequality distorts the global economy and threatens international peace. Yale University Press
Kim P, Lee NY (2013) The automobile industry. In: The Park Chung Hee Era: The Transformation of South Korea. Harvard University Press. pp. 295–327 http://www.tinyurl.com/y4nsjoy6
King’s College London (2021) Climbing global value chains: Industrial upgrading in the Mexican automotive industry. Master’s thesis. Available at: https://www.kcl.ac.uk/eis/assets/climbing-global-value-chains-industrial-upgrading-in-the-mexican-automotive-industry.pdf
Lawton P (2019) Age dependency ratios and GDP growth in developing countries. SSRN Electron J [Preprint]. https://doi.org/10.2139/ssrn.3370526
Ma W (2020). Inside Apple’s eroding partnership with Foxconn. The Information. Available at: https://www.theinformation.com/articles/inside-apples-eroding-partnership-with-foxconn?utm_content=article-4988&utm_campaign=article_email&utm_source=sg&utm_medium=email. Accessed: 08 Jun 2023
Moreno Brid JC (1996) Mexico’s auto industry after NAFTA: a successful experience in restructuring? Working Paper Number [Preprint]. https://kellogg.nd.edu/documents/1527
Morton, AD. Structural change and neoliberalism in Mexico: ‘Passive revolution’ in the global political economy. Third World Q; 2003; 24,
Nguyen-Huu TT, Pham N-S (2021). Escaping the middle income trap and getting economic growth: how does FDI can help the host country? https://shs.hal.science/halshs-03143087
Oks, D; Williams, H. The long, slow death of global development. Am Aff; 2022; 6,
Partington R (2021) OECD deal imposes Global Minimum Corporate tax of 15%, The Guardian. Available at: https://www.theguardian.com/business/2021/oct/08/oecd-deal-imposes-global-minimum-corporate-tax-of-15. Accessed: 24 Jul 2024
Petrova M (2018) We traced what it takes to make an iPhone, from its initial design to the components and raw materials needed to make it a reality. CNBC. Available at: https://www.cnbc.com/2018/12/13/inside-apple-iphone-where-parts-and-materials-come-from.html. Accessed: 06 Jun 2023
Schneider BR (2013) Hierarchical capitalism in Latin America. Cambridge University Press
Tampakoudis, IA; Tampakoudis, I; Dafermos, V. Determinants of foreign direct investment in middle-income countries: new middle-income trap evidence. Mediterranean J Soc Sci; 2017; 8,
Topcu, E; Altinoz, B; Aslan, A. Global evidence from the link between economic growth, natural resources, energy consumption, and Gross Capital Formation. Resour Policy; 2020; 66, 101622. [DOI: https://dx.doi.org/10.1016/j.resourpol.2020.101622]
The World Bank & Development Research Center of the State Council, the People’s Republic of China (2013) China 2030: Building a modern, harmonious, and creative society. China 2030. https://doi.org/10.1596/9780821395455_overview
The World Bank (2023) World Bank country and lending groups. World Bank Country and Lending Groups—World Bank Data Help Desk. Available at: https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups. Accessed 07 Jun 2023
World Development Indicators (2023) DataBank. Available at: https://databank.worldbank.org/source/world-development-indicators. Accessed: 06 Jun 2023
The World Bank (2024) GDP per capita (current US$)—Mexico, Korea, rep. World Bank Open Data. Available at: https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=MX-KR
The World Bank (2024) Population, total—Mexico, Korea, rep. World Bank Open Data. Available at: https://data.worldbank.org/indicator/SP.POP.TOTL?locations=MX-KR
Zeng, J; Fang, Y. Between poverty and prosperity: China’s dependent development and the “middle-income trap. Third World Q; 2014; 35,
Yoon JS (2023) Topic: Automotive industry in South Korea. Statista. Available at: https://www.statista.com/topics/5249/automotive-industry-in-south-korea/#topicHeader__wrapper. Accessed 04 Dec 2023
Yülek, MA; Yülek, E; Ugur, A. State capacity and the role of industrial policy in automobile industry: a comparative analysis of Turkey and South Korea. J Ind Compet Trade; 2020; 20,
© The Author(s) 2024. This work is published under http://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.