A learning portal model of emerging markets multinationals

Research Summary: The purpose of the article is to explain the mechanisms underlying the internationalization springboard strategies of China's emerging multinationals in the automobile industry. Using a unique combination of location bound (country-specific) and non-location bound (firm-specific) advantages, companies in this industry have overcome their latecomer disadvantages by exploring knowledge from mature markets through backward and forward integration of the value chain, and exploiting this knowledge to enhance their competitive position, first at home, then in emerging markets and later in developed countries. We test and refine springboarding theory and propose the learning portal model. The learning portal model can be used as a new theory of emerging markets multinationals beyond China and the automobile industry. Management Summary: The globalization of Chinese enterprises is upon us. Chinese enterprises have not only successfully internationalized to other emerging markets, but also to advanced economies. To do so, these firms used springboard strategies to leapfrog cer-tain technological development stages and accelerate their catch-up trajectory with incumbent firms from the advanced economies. This article investigates the catch-up strategies of China's leading automotive companies. We find that Chinese enterprises establish learning portals in mature markets to acquire knowledge and assets, which they first exploit in their home market, before seeking to expand their market share in other emerging markets and finally advanced economies.

. The next stage is radical outward FDI, the transfer and upgrading of 1. How do EMNEs overcome their firm-specific disadvantages in international markets? 2. What international modes of entry do EMNEs use abroad and how do they use these modes as learning portals? 3. How are learning portals used to augment EMNEs' firm-specific, non-location bound advantages first at home and later in international markets, both developing and developed?
More specifically, this article shows how Chinese MNEs in the automobile industry establish learning portals in mature markets to acquire knowledge, assets, and channels of distribution to expand sales. Knowledge gained from these two types of markets is then used to further their industrial capabilities and global competitiveness.

| THE SPRINGBOARD THEORY OF EMERGING MARKET MULTINATIONALS
Internalization theory and its successor, the eclectic theory, assume that multinational companies possess FSA and ownership advantages through technology, know-how, or intellectual property that they can exploit through investment in countries with institutional voids to deal with market imperfections. However, this assumption is, in fact, often wrong when it comes to EMNEs (Peng, 2012). For example, Geely did not buy Volvo from Ford out of a position of strength, but rather out of a position of weakness because it was searching for better organizational capabilities in production and marketing (Alon, Fetscherin, & Sardy, 2008). EMNEs have stirred calls for new theories or the extension of existing theories to better explain the behavior of these new entrants in the global economy (Child & Rodriguez, 2005;Cuervo-Cazurra, 2012). One of the most prominent theories providing a new view is Luo andTung's (2007, 2018) springboard perspective. This theory is particularly suitable to the study of EMNEs because it is dynamic and shows the evolution of the firm's internationalization over time and space. It suggests that EMNEs invest abroad in search of strategic assets such as technologies or brands. Companies can capture more added value when moving toward the market (downstream), or toward basic technology and know-how (upstream; Hertenstein, 2019;Wu & Ang, 2020). EMNEs that have low costs for manufacturing and labor, energy and land, but also limited added value, lack the opportunities to develop FSA at home. To catch up, they have to link to advanced country multinationals, technologies, and markets, leverage access to their own markets' resources, and learn from best practices (Bartlett & Ghoshal, 1989;Hertenstein et al., 2017;Liu & Meyer, 2020;Mathews, 2006). The springboard perspective provides a plausible explanation of how EMNEs use asset-seeking investments, especially cross-border M&As, as a tool to overcome their competitive weaknesses and latecomer disadvantages.
Springboarding builds on the idea of leapfrogging, through which firms can skip stages of their lengthy development process and jump to the newest technologies to close the gap with incumbent industry-leading firms (Anderson & Engers, 1994;Dore, 1973;Khanna & Palepu, 2006;Luo, 1998). The concept of leapfrogging and rapid development originates from studies on late industrialization at the level of the national economy, and has been adapted to describe the catching up of EMNEs (Abramovitz, 1986;Bell & Pavitt, 1993;Kumaraswamy, Mudambi, Saranga, & Tripathy, 2012;Lall, 1990). Ramamurti and Hillemann (2018) suggested that EMNEs turn lateness into a competitive advantage through labor cost arbitrage, skipping to the latest technological frontier in smokestack industries and the newest platforms in sunrise industries. What starts as an original equipment manufacturer turns into an original design manufacturer or original brand manufacturer. Luo and Tung (2007) maintained that springboarding involves: (a) internationalization, (b) proactive international M&A and greenfield investments to seek strategic assets (radical internationalization), (c) outward FDI to bypass national protectionism, (d) overcoming domestic institutional voids, (e) obtaining national and governmental support, (f) reverse investment back into the home country, (g) exploiting local competitive advantages, such as low cost manufacturing, (h) obtaining capabilities from international partners through joint ventures, inwards investment, and cooperative agreements, and (i) achieving national champion status. Luo andTung (2007, 2018) maintained that firms internationalize to satisfy overseas demand (often termed "market-seeker"), and that outward FDI is used to establish an international operating firm. In their model, EMNEs buy strategic assets such as technology outside their own country in order to quickly catch up with the global competition on an international scale: "At the core of this framework is our argument that [EMNEs] use outward investments as a springboard to acquire strategic assets needed to compete more effectively against global rivals" (Luo & Tung, 2007, p. 482;emphasis added).
A recent in-depth study has questioned the implicit, underlying assumption that EMNEs invest abroad to become globally competitive (Williamson & Raman, 2013). It appears that at least some EMNEs use the assets they acquire through cross-border M&As almost exclusively to strengthen their competitive positions in their domestic markets. Moreover, the springboard model focuses on using FDI, particularly cross-border M&As, to acquire knowledge and build capabilities abroad to the exclusion of other learning modes such as exports, which, according to the Uppsala model of internationalization, is a key mechanism for learning (Johanson & Vahlne, 1990, 2011.
Therefore, we posit that the springboard perspective is underspecified in three aspects. First, the question remains how and where EMNEs use the assets they acquire abroad. The literature on springboarding maintained that EMNEs seek to become competitive in the global economy, competing directly with global rivals, because they are seeking markets outside their home country. They argued that EMNEs' early expansion into mature markets improves their "international competitiveness" (Buckley et al., 2007;Deng, 2009;Rui & Yip, 2008;Taylor, 2002;Zhu, Tse, & Li, 2019), "competing in international markets" (Buckley, Cross, Tan, Xan, & Voss, 2008), or "becoming global players in international markets" (Child & Rodriguez, 2005;Deng, 2009;Rui & Yip, 2008). While secondary or intermediate motives such as resource seeking or strategic asset seeking are acknowledged, the studies state that the primary goal is to secure the competitive advantage of the firm in their global market. The strategic asset-seeking motive is thereby combined with a market-seeking motive for the outward FDI of EMNEs. The springboard perspective inherited this view from Dunning's (1993Dunning's ( , 2000 eclectic paradigm, in which FDI and non-FDI entry modes are seen as two alternative routes to satisfy overseas demand. We question this view of the linked asset and market seeking behavior and posit that both objections can be independently pursued. Second, the springboard perspective is underspecified as to how EMNEs combine location bound (country-specific advantages, CFA) and non-location bound (FSA) to create new FSAs. As Luo and Tung (2018) predicted, EMNEs engage in unique activities. They start with joint ventures with international competitors or engage in "radical" greenfield and brownfield investments abroad, which allows them to combine location bound advantages with non-location bound advantages to dynamically and indirectly help the firm develop new FSA that help with global competitiveness (Rugman & Verbeke, 1992). In practice, disentangling firm-and country-specific advantages and disadvantages is difficult, especially for state-owned enterprises and industries relying on state funding, purchases and/or subsidies (Sutherland et al., 2020). Rugman and Verbeke (1992) suggested that the impact of CFA is "not restricted to direct andor static benefits accruing to the foreign subsidiary location in that country but may actually contribute to the development of new FSA" (p. 766). Ramamurti and Hillemann (2018) maintained that Chinese multinationals are more likely to be state-owned or controlled, depend on home-country specific advantages, but lack non-location bound FSA, which they therefore seek abroad. But Chinese enterprises are not without unique FSA including ownership-specific and proprietary know-how, as well as the unique assets or transactional advantages that they coordinate and control (Rugman & Verbeke, 1992). While Chinese automobile companies lack brand name recognition, global channels of distribution, and manufacturing superiority, they have the advantage of home country access when operating in the world's fastest growing automobile market (China), government support and funding, cheap labor costs at home, and a strong desire to compete on a global scale.
Third, the springboard theory assumes entry is via FDI rather than other means of accessing and accumulating knowledge and other assets in mature markets. Further studies, mostly using aggregated FDI data, seemed to support the theory. However, an expanding body of literature using in-depth studies struggles with the model, as they report different internationalization strategies for EMNEs (Anderson & Sutherland, 2015;Hertenstein et al., 2017;Huang & Wang, 2011;Williamson & Raman, 2013). Apart from the home market perspective of outward FDI, other entry modes seem to play an equally important role in the learning and development of Chinese MNEs. This finding is not surprising, given that previous research that examined the catch-up of South Korean and Japanese firms established how exports can help create knowledge for the development of FSA (Amsden, 1989;Cusumano, 1985). Alon et al. (2020) noted that the Chinese establishment mode varies with the home country's institutional environment and the learning that comes from the scope and scale of internationalization.

| RESEARCH METHODS
An inductive approach seemed most appropriate to explore the complementary questions regarding the springboard perspective (Gibbert, Ruigrok, & Wicki, 2008). For inductive research questions, and to answer "how" questions, a qualitative approach is best. Moreover, to explore the internationalization trajectories and adjustments in the strategic objectives of the companies, we selected a longitudinal study design. For these reasons, we chose a case study approach. Utilizing this method also augments existing research on China's outward FDI that so far has been of a descriptive nature, using official aggregated data and econometric analysis to understand the trend (Buckley et al., 2007;Taylor, 2002). Reviews on this topic revealed that there is a dearth of qualitative in-depth studies on the internationalization of Chinese MNEs (Alon et al., 2018;Deng, 2012), which would benefit from more "longitudinal and qualitative work" (Deng, 2011, p. 423).
We chose a grounded approach for theory development (Eisenhardt & Graebner, 2007;Glesne, 1999). We did so by exploring the internationalization strategies of multiple case studies of China's automotive industry, as multiple cases "yield more robust, generalizable, and testable theory than single-case research" (Eisenhardt & Graebner, 2007, p. 27). Using multiple case studies also improves the study's external validity and provides a good basis for analytically generalizable findings (Eisenhardt, 1989;Gibbert et al., 2008).

| Sample selection and data collection
We limited our sample to a single industry to obtain comparable results for a better understanding of the different internationalization strategies. By focusing on a single industry from a single home country, we implicitly control for industry variations in internationalization and home country development, allowing us to identify the effects of location-bound and non-location bound FSA. We selected 12 assembly firms from China's automotive industry, which we identified from the Chinese Automotive Yearbook (Hang et al., 2011). To avoid selection bias based on internationalization strategies, we selected the companies according to size. Originally, we approached 12 companies that included passenger car and commercial vehicles manufacturers, as well as state-owned and privately owned companies. Of the companies we approached, eight agreed to participate in the research, resulting in a good mix between passenger car and commercial vehicle manufacturers and private as well as state-owned enterprises. The companies were at different stages of their internationalization at the time of the research and used various approaches to enter new markets.
Each case study is based on primary data from semi-structured interviews, and secondary data from annual reports, company web pages, magazine and newspaper articles, published interviews with company officials, or previously published works on the Chinese automotive industry (Eisenhardt & Graebner, 2007;Gibbert et al., 2008;Yin, 2003). Between July and December 2012, we conducted 18 semi-structured interviews with the companies and industry experts in China. To avoid retrospective sense-making, interview partners were chosen from different functional areas, hierarchical levels and from different units of the companies (Eisenhardt & Graebner, 2007). Moreover, to improve internal validity, interviews were also conducted with industry experts from outside the actual companies. Table 1 provides a brief background on each of the cases we studied. Table 2 summarizes the knowledge and strategic assets accessed through internationalization, the internationalization modes, and the markets in which each firm sought to leverage its expanded portfolio of advantages.

| Exploring knowledge through M&A
Multiple companies used joint ventures and FDI to learn about and transfer technology. One example of this approach is SAIC's cooperation with MG Rover. The SAIC Motor Corporation Limited is the largest Chinese auto producer with approximately 4 million units a year and was among the first Chinese state-owned enterprises to actively acquire foreign firms for assetseeking reasons. Before its acquisition of MG Rover, SAIC planned to create a joint venture with the British automotive manufacturer. According to the interviewee, the purpose of such a joint venture was to gain access to MG Rover's world-class technology (Interview, August 6, 2012). At the time of the planned joint venture, SAIC already had two joint ventures with Western OEMs: one with Volkswagen, and one with General Motors. Why, then was SAIC planning a third joint venture to gain access to such knowledge? As the manager explained, the previous joint ventures were structured so that their Western partners did all of the research and development for the vehicles and their platforms, including the powertrain, and the external and internal designs, with no involvement of SAIC. While SAIC was able to learn about different production processes and technologies, learning about platform and vehicle development was restricted, and did not meet the company's demand for learning. The new SAIC-MG joint venture would allow SAIC to be involved in the platform and vehicle development, filling the T A B L E 1 Profiles of the companies

Passenger Car OEMs
Brilliance, founded in 1992, is a state-owned automotive manufacturer that had an annual output of 434,182 vehicles in 2010, of which approx. half were passenger cars and commercial vehicles. The group was the first Chinese passenger car producer to expand into mature markets. In 2007, the group started exporting passenger cars to Germany but had to withdraw by 2009. By gradually adapting the group's strategy from gaining market share to selfimprovement in technology and experience, Brilliance set up a learning portal through an R&D center in Germany in 2011.
Geely, established in 1997, is arguably the most successful privately owned OEM in China. The international breakthrough came with the acquisition of Volvo in 2010, from which the company gained international attention and prominence overnight. Geely first exploited the FSA gained from the Volvo acquisition to expand its production footprint in China, before eventually expanding its sales in the European market.
SAIC is one of the oldest state-owned OEMs in China, and one of the first companies to get involved in international M&A activities. In 2004, SAIC quired a 51% stake in the Korean company Ssangyong, but was not able to turn the company around, and sold its shares again in 2010. In 2006, SAIC acquired key technologies from MG Rover, which was split between SAIC and NAC. However, when NAC had financial problems, SAIC acquired NAC and with it, all properties of the former British MG. Further internationalization for knowledge acquisition followed, with the founding of a U.S. based R&D center in 2015, and an Israel based R&D center in 2017.

Truck OEMs
Dongfeng is a state-owned company and one of China's largest OEMs, producing both passenger cars and commercial vehicles. The latter division began internationalizing. In 2009, Dongfeng wanted to expand to the EU, but failed. At the same time, the company started exporting commercial vehicles to multiple emerging markets such as Pakistan, Indonesia, Malaysia, Myanmar, Paraguay, Philippines, and Vietnam. Moreover, the company changed its strategy to build learning portals in advanced economies, and in 2012 established an international R&D center in Sweden.
FOTON is a subsidiary of the state-owned BAIC, and is the largest commercial vehicle enterprise in China, with an annual output of 615,725 commercial vehicles. As one of the earliest exporters, FOTON has been able to capture considerable market share outside of China, most notably in Africa, Southeast Asia, South Asia and Central & South America. Moreover, the company uses small-scale exports to North America, Europe, and Japan to establish a learning portal and gain knowledge from the markets.
FAW is China's oldest vehicle manufacturer, with a focus on commercial vehicles. It has established representatives in Germany, Japan and the United States, to which it has started small-scale exports of small trucks in order to learn from its competitors in these markets.

Bus OEMs
King-Long is a semi-privatized, state-owned company and the largest bus producer in China. It is also the largest exporter of buses, with significant exports to the Middle East and East Asia. The company's products were approved by the EU in 2005 with a British Commercial Vehicle certificate UCA. In 2011, King-Long entered a partnership with Arriva and has supplied buses to Malta on a small scale as a way to establish a learning portal.
learning gap in the previous joint ventures. However, in 2005 during the negotiations for the joint venture, MG Rover had to file for bankruptcy. To retain its access to MG's technology and development capabilities, SAIC quickly moved from negotiations about a joint venture to negotiations for the acquisition of MG. With Nanjing Automobile (NAC) also bidding for MG, SAIC initially acquired some of MG's technology, the R&D center and the Rover brand name, while NAC acquired the majority of MG's technology together with MG's equipment and machinery. However, after running into financial trouble of its own, NAC was also merged into SAIC in a shotgun merger driven by the government. After SAIC acquired NAC, together with the remaining assets from MG Rover, SAIC focused on transferring the knowledge and technology to China, while maintaining a small MG assembly plant in the United Kingdom. The purpose of the U.K. production plant is to manufacture a few hundred to thousand vehicles locally for sales in the EU. The goal of these sales is not to seek markets outside of China, but to maintain the quality and technological standards and requirements for the EU market, so as to prevent the MG brand from losing value. As a former SAIC senior manager put it: "People will think our cars are good enough for the European market. This way we hope to be able to compete with the global firms in China." Other than that, SAIC has "no plans to further expand or serve the European market" (Interview, August 6, 2012).

| Exploring knowledge through R&D centers in developed countries
Another objection behind FDI focused on establishing R&D centers in advanced markets. For instance, Brilliance set up a European center in 2011 to conduct market research, build knowledge, and learn how to compete in mature markets. The company took this step after an attempt at large-scale exports to Germany that failed in 2009. The knowledge created at the center focuses on understanding current trends and customer demands in the mature market and feeding the knowledge to the R&D center in China to improve their quality and safety standards (Interview, July 26, 2012c).
Other companies also focused on establishing strong learning loops in advanced economies through localized R&D. For instance, while SAIC moved the majority of MG's production operations to China, MG's R&D center that develops the MG branded vehicles for the United Kingdom and China remained in the United Kingdom to be able to harvest the knowledge spillovers from close geographic proximity to "innovation drivers." Examples include cutting-edge firms such as other auto assembly groups or top-tier supply firms, leading universities, or independent research institutions like the Fraunhofer Institute in Germany (Hertenstein & Williamson, 2018;Zhang, 2009). Being located close to these organizations enables the R&D center to exchange know-how with other players in the market, and to develop cutting-edge technologies and products (Kogut, 1991;Penner-Hahn & Shaver, 2005;Szulanski, 1996). Additionally, SAIC opened R&D centers in Silicon Valley in 2015 and in Israel in 2017 to tap into local knowledge and talent pools for future technologies such as autonomous vehicles.

| Exploring knowledge through small-scale exporting to developed markets
All of the firms in our case studies used exports to enter advanced economies. However, they exported only a limited number of vehicles. In doing so, they used the exports for strategic purposes other than market seeking. One firm (Brilliance) initially exported vehicles for market seeking purposes but had to retreat within 2 years. In contrast, all other firms used small-scale sales of for learning, like safety standards, government regulations regarding emissions, and the purchasing criteria for customers in the top segment, who focus on quality, safety, interior design, and performance of the cars.
King-Long, a state-owned company founded in 1988, and China's largest manufacturer of commercial coaches, is a good case in point. In 2005, the company received a British commercial vehicle certificate for the European market and partnered with Arriva in 2011 to export its first buses to Malta. Although technically not a developed country as defined by the OECD, Malta is a member state of the European Union and therefore has the same technological requirements and similar customer requirements as a mature market. The total number of vehicles exported to Malta at that time was around 80-a mere 0.1% of the annual output of King-Long. Moreover, King-Long had to establish a technical service team in Malta to provide after-market services for the vehicles sold. For such a small number of exports, the expenses exceeded the profits. Still, the manager stressed the importance of their small number of exports to the mature market, and claimed that the objectives had been metwhich were not market seeking in overseas advanced economies: Most of our knowledge we gain from our customers. Our engineers need to be involved with our customers early on, so that we can understand their needs … In the future, the emerging markets, even the Chinese market, will gradually improve their requirements. So we will have the same requirements here as we have with the European customers now. So going to the European market was not because we wanted to do some business in European market, but we wanted to give us a push, or pull, to improve our current technical standards as well as manufacturing technicians, so we can be prepared for the future. We think that the European market will force us to improve now, but we are only doing a little business in Europe. Because if we compete with the top brands, we will learn from them. We cannot just watch from outside; we need to be in the market, so that we can improve every day. (Interview, July 24, 2012).
The direct competition with the global leading assemblers helped the Chinese firm develop its own strategic assets and build sustainable, FSA by combining the knowledge acquired in the mature market about issues such as safety and emissions with the unique advantages in their home market such as low cost production (Hertenstein & Williamson, 2018).
Exporting to accelerate learning instead of seeking market expansion may seem counterintuitive at first. However, the firms in our sample were able to learn and increase their capabilities on multiple fronts. They acquired a deeper understanding of the technological requirements to meet safety expectations of customers, learned about the standardized European vehicle tests, and the role of specific components such as side airbags, head supports, and the structural stability of the chassis for good test results. They learned about shifts in customer purchasing criteria as the industry matures, such as an increasing focus on total cost of ownership for commercial vehicles rather than on the purchasing price. Responding to these insights, the Chinese companies redesigned some of their vehicles to reduce their weight, increasing fuel efficiency and reducing wear on the tires, two key aspects that increased the total cost of ownership. Furthermore, they learned how Western clients approach joint projects, laying out detailed specifications to customize the vehicles. In one case, the interviewee explained that the company needed to redesign the driver's door on its coaches. The customer rejected the first two redesigns because the design was off by 2 g, well within what would be tolerated by Chinese customers. "It took us six months of development to finally meet all requirements. We learned a lot during the project, how the companies operate and collaborate, but mostly how to push the developers to meet stricter requirements" (Interview, July 26, 2012a). The lessons learned through the close collaboration abroad helped the company to prepare new processes at home that helped to move up the value chain by developing better performing vehicles for increasing demand in China.
The knowledge the companies learned from their exports differs from that gained from other entry modes. Through the exports, the companies came to understand the demand requirements from advanced markets, and by extension, about the expected demand for changes in their home country, as the market matures. By acquiring foreign targets, the companies gained access to leading edge technologies. Through greenfield investments, often to establish R&D centers, the companies acquired product development capabilities, adding to the insights into customer demand gleaned from their experience with their exports. While the knowledge gained from exports differs from that of M&As or R&D centers, we conclude that through small-scale exports to mature markets, EMNEs can establish a learning portal in the advanced economy. Through that portal, EMNEs can accelerate their learning, improve their capabilities, and enhance perceptions about their brand among potential customers at home to become more competitive in their home market.

| Exploiting gained advantage at home
The overarching objective of the firms in our sample was to capture knowledge and technology in advanced economies (location-bound FSA), and then recombine them with the non-location bound FSA to increase their market share in the home market. This was the case for both FDI and non-FDI internationalization approaches. The firms used the assets they acquired primarily to drive sales in their home market and to improve the firms' competitiveness against rivals from other local and global players in China. As one interviewee said: I think it's a feasible way for Chinese companies to acquire a champion company, not necessarily a big company, but it has to be a hidden champion in a niche market. If they have technology, we take over the company, we get IP, get the knowhow, get the product, and then we develop the China market (Interview, July 7, 2012, emphasis added).
Even in one of the most discussed cases in the literature, the Volvo takeover by Geely, we identified a dual purpose. According to our understanding, the primary objective was to bring Volvo to China and thereby strengthen the position of the Geely group in China. The second objective was to access Volvo's technology to use for other Geely brands such as PolarStar. The primary goal becomes apparent when looking at the adjustments made in the production footprint since the takeover in 2010. While two production sites were closed in Thailand and Sweden, three new plants were opened in China: Chengdu, Daqing, and Luqiao, with a fourth plant in Chengdu under construction. Only after the rapid expansion of Volvo to exploit the Chinese market did the Geely group consider a further expansion to other markets by establishing the first U.S. based Volvo plant in South Carolina in 2018. While the reverse market-oriented expansion of the acquired Volvo branded vehicles in China was executed swiftly after the acquisition, steps to cross-sell Chinese vehicles in Europe (the home of Volvo) began only in 2020-10 years after the acquisition. Similarly, a senior manager of Yutong, a Chinese coach and bus manufacturer, said: From the European market we can learn a lot for the company. We just try to learn something from the product side. For sales, every year we sell only 30-40 buses. Even if we could sell a few hundred, this would not be good enough from a sales perspective. … At the moment, we are not profitable in Europe (Interview, July 26, 2012a).
He went on to explain how he learned more about the automotive industry during the 2 years since they started exporting to Europe than in the entire 20 years of his career before. He also described how the company used this knowledge to gain a better position in the home market and in their most important overseas market: the Middle East.

| Exploiting gained advantage in emerging markets
Our firms also entered markets outside of China with the aim of growing foreign sales, but only after establishing a strong position in their home market. Successful market-seeking expansion has focused entirely on other developing markets in South East Asia, the Middle East, Latin America, and Africa (Table 2) instead of advanced economies. Two of our case study firms attempted market-seeking expansion into advanced economies, both of which failed. Unlike its experience in Europe, Brilliance was able to successfully export cars to other emerging markets. Of its total annual revenue of 6 billion RMB in 2013, 6.2% came from sales in the Middle East, 2.1% from Africa, 1.9% from South America, and 0.8% from other Asian countriesall regions with predominantly emerging markets. Only a mere 0.01% of the revenue came from "other countries," including sales in advanced economies. After its failed entrance and subsequent exit from the German market, the focus of Brilliance's international expansion has been on other emerging markets.
Similarly, King-Long expanded to other emerging markets. As a senior manager from King-Long explained, the company has successfully sought markets outside of China with a focus on emerging markets, noting that, "approximately 20% of our revenue is from overseas" (Interview, July 24, 2012). To date, King-Long is active in other East Asian markets and the Middle East. As one interviewee said: "Emerging markets are the largest overseas markets for all Chinese OEMs at the moment, because this is the best match between our products and the market requirements" (Interview, July 24, 2012). In emerging economies, Chinese MNEs are able to exploit the unique competitive advantages gained from combining the location-bound advantages from home (such as low-cost manufacturing in China) and abroad (such as access to knowledge and technology) with its firm specific advantages (such as R&D capabilities). Accessing and exploring location-bound advantages in advanced economies requires the creation of local exposure through knowledge portals that the firm can thereafter exploit these transferrable advantages elsewhere (Gupta, Smith, & Shalley, 2006;March, 1991). Through absorptive capacity at home, the Chinese MNEs can transfer the knowledge gained to their capabilities at home, to re-combine them with existing advantages (Cohen & Levinthal, 1990). By re-contextualizing the learning from abroad, the companies are able to avoid overshooting the development for vehicles for the demand in other emerging markets, to develop vehicles that are "just right" for the local demand and compete directly with international competitors (Hertenstein & Williamson, 2018;Thun, 2018). Only after succeeding in countries at similar stages of development with limited psychic distance did they expand further to more distant emerging economies, and eventually into mature markets. It is notable that none of the firms in our sample followed the mixed strategy suggested by the springboard perspective, where a firm expands into mature markets to seek assets and local market share simultaneously.

| TOWARD A LEARNING PORTAL MODEL
Based on our case study findings, we propose a model to explain the early expansion of Chinese MNEs into mature markets via outward FDI and exports. We posit that Chinese MNEs expand into mature markets early on to create a learning portal that will help them access locationbound advantages from advanced markets and combine them with the FSA in China to create new capabilities Learning portals are established in geographic locations where EMNEs can access location-bound advantages, such as the technology or market knowledge needed for the development of their own FSA. Due to the "stickiness" of such knowledge, it is usually found in close proximity to cutting-edge excellence clusters, most of which are in mature markets (Kogut, 1991;Penner-Hahn & Shaver, 2005;Szulanski, 1996). Acquiring assets from overseas is important when the home country cannot create such assets (Rui & Yip, 2008;Young, Huang, & McDermott, 1996). Alliances with other leading firms, and joint research programs with top universities or independent research institutions can also help create new knowledge (Hertenstein & Williamson, 2018;Zhang, 2009). The MNEs exploit their new advantages in different stages, first in their home market, second in other emerging markets, and third in advanced economies (Figure 1). This model is situated between the traditional theory of internationalization that assumes a linear trend of internationalization from psychically similar countries to less familiar ones and the springboard perspective that assumes a sudden surge in internationalization leapfrogging into developed countries for market seeking purposes.
Only two of our cases support the argument that Chinese companies use their acquisitions as a springboard to compete immediately in the global market. Indeed, they both failed to capitalize on their investments and had to retreat from the overseas market. In these cases, leapfrogging or springboarding to global markets by utilizing strategic assets acquired via cross-border M&As, as the theory suggests, did not occur (Luo, 1998;Luo & Tung, 2007). Instead, as Table 2 indicates, the vast majority of our firms focused on market exploitation in other emerging markets. (China)

Home market
Emerging market Advanced market Step 1: Establish learning portal via FDI & export for advantage exploration Step 2: Transfer knowledge & technology for market exploitation at home Step 3: Exploit combined advantage in other emerging markets Step 4: Exploit combined advantage in advanced markets F I G U R E 1 EMNE catch-up stages through learning portals For instance, after its failure in Germany, Brilliance successfully exported to emerging markets such as Bangladesh, Jordan, and Turkmenistan. In 2019, Geely exported almost 63.000 vehicles to emerging markets such as Russia, Ukraine, the Middle East, and South America. Truck exports from Dongfeng focused on East Asia and South America, and FOTON internationalized its production operations in emerging markets such as Brazil, Paraguay, Vietnam, Malaysia, and Pakistan, with no significant expansion to mature markets yet. The bus producer Yutong successfully exported large numbers of coaches to other emerging markets for over 10 years, before announcing a partnership with Flixbus in 2018 to finally increase exports to advanced economies. Like Bartlett and Ghoshal (2000), we found that springboarding EMNEs do not immediately compete against global leading firms in their home turf. Instead, they rapidly establish learning portals in advanced markets, to accelerate their learning and catch-up in a timely process move up the value chain before eventually expanding into mature markets In our view, springboarding needs to be understood as one step in an evolutionary process through which EMNEs can accelerate their gradual learning to catch up with the leading global companies. We further suggest that the springboarding process needs to be combined with a strengthening of the position at home, to combine the acquired location-bound advantages with the FSA developed at home, which results in unique new FSA that allows the EMNE to catch up with AMNEs. As Figure 2 illustrates, early internationalization in advanced economies serves the purpose of knowledge building through learning portals, followed by market seeking internationalization as suggested by contemporary theory.

| The learning portal model and contemporary internationalization theory
The sequential utilization of the knowledge created through the learning portal and the subsequent market-seeking internationalization of the firms in our sample aligns well with the common transaction cost-based and behavioral internationalization theory (Buckley & Casson, 1976;Dunning, 1993;Johanson & Vahlne, 1977). As this theory predicts, the firms in our sample started expanding to close markets in South East Asia, then gradually expanded into more distant markets such as the Middle East, South America, Africa, or Eastern Europe. Other Chinese OEMs such as JAC and Chery have also recently entered the Brazilian automotive industry, which supports our analysis. Thus, Chinese MNEs do follow the internationalization patterns predicted by well-established theories and augmented by our learning portal model for early knowledge exploration activities in advanced economies that help enhance and extend existing FSA. We make two contributions to the understanding of springboarding: First, the abovementioned distinction between knowledge seeking and market seeking, and second by arguing that non-FDI entry modes play a similar role for accelerated learning. What is unique about our model and is not fully specified in the springboarding perspective is the use of exports for learning. Previously, exports were thought to be a function of market development, or an extension of the home market. However, in our model, we show that Chinese firms cared less about the actual sales and profits that could be generated in developed countries and more about exploring potential opportunities to learn about such things as risk management, product development, quality standards, management skills, branding, supplier management, distribution models, and customer relations. The location specific advantage of meeting European customer demands, as well as standards regarding emissions or crash tests sells well with home-country consumers. Exports are used for the same purpose as FDI in mature markets, which is to acquire knowledge and accelerate the EMNE's learning and strengthening its position in the home market. This non-FDI addition to the model also differentiates it from the double handspring model developed by Raman (2011, 2013), who found similar objectives of home market oriented FDI as we did but limited their model to crossborder M&As from China.

| The learning portal model and the Uppsala process model
Why do EMNEs today appear to internationalize faster than earlier studies on the catch-up process suggested? Slower development paths followed during the earlier internationalization of AMNEs or during the catch-up phase of companies from the so-called "Asian tiger states" may no longer be a feasible option for EMNEs in today's globalized environment. Today, EMNEs already face competition from AMNEs in their own markets -something most AMNEs never experienced in their formative stages of internationalization (Amsden, 1989;Hertenstein et al., 2017). Additionally, pressures from the globalized economy encourage fast-paced internationalization, which is also evident in new entrants into the global market coming from developed countries (Cuervo-Cazurra, 2012;Hennart, 2014;Loane & Bell, 2006). As Vahlne (2009) argued, andThaijongrak (2013) and Hertenstein et al. (2017) demonstrated, this situation underscores the need to recalibrate the internationalization process model to allow for an accelerated internationalization process. Our learning portal model extends these advances by providing an explanation for accelerated learning, the building of FSA and catching-up, as well as adding an explicit distinction between knowledge-seeking and market-seeking objectives.

| Marrying the internationalization and springboarding perspectives
In order to marry the merits of both internationalization and springboarding perspective one should consider that firms can cycle through periods of exploitation and exploration sequentially and dynamically and through temporal sequencing, as the ambidexterity view contends (Raisch, Birkinshaw, Probst, & Tushman, 2009). According to this view, firms need to continuously reconfigure their activities to meet changing demands in their internal and external environments through a dynamic process of combinative capabilities, synthesis, access to new knowledge, and absorptive capacity. Akin to this view is the learning portal model, suggesting that EMNEs explore new capabilities in developed countries through exports and FDI, and exploit market opportunities through their existing technologies. Before large-scale internationalization, there is a process of incubation where the EMNE absorbs and integrates capabilities through learning portals and uses them at home, and tests them in other emerging markets. Only after developing differentiated and valuable capabilities on par with the incumbents will the EMNE be ready to face the global AMNEs on their own turf. The process begins with internationalization to obtain a domestic competitive advantage. Large emerging markets, such as BRICS, are by themselves sizable for scale. Looking at internationalization from the perspective of initially seeking a domestic advantage also allows us to reinterpret the likely nature and impact of FDI by EMNEs. The findings from the cases suggest that, for EMNEs, most of the benefits from some of the early outward FDI that is usually associated with bringing new assets and technologies from the host country actually spill over into the investor's home country. Outward FDI by Chinese firms creates benefits such as managerial and technological know-how fueling growth and job creation back in China (Luo & Tung, 2007;Young et al., 1996). The Chinese government has encouraged such FDI through favorable financing and policies such as the Belt and Road Initiative (Sutherland et al., 2020).
Knowledge obtained is combined with CFA to create new FSA with every acquisition and international encounter, including failure. The ability to transfer knowledge from the host country investment is conditioned on its embeddedness. According to the literature in international business (Driffield, Love, & Menghinello, 2009;Liu & Meyer, 2020;Nair, Demirbag, & Mellahi, 2015;Simonin, 2004) and knowledge transfer, advantages are location-bound when their transfer is costly (Criscuolo, 2009;Iwasa & Odagiri, 2004;Kogut & Zander, 1993). Knowledge transfer barriers are created when advantages are tacit in nature, or embedded organizationally at the local level, or in the local environment.
We argue that a distinction is needed between location boundedness and exploration and exploitation activities. According to March (1991), exploration is associated with "search, variation, risk taking, experimentation, play, flexibility, discovery, innovation" (p. 71). In other words, exploration includes research and development and knowledge creation, which are by definition adaptive processes and belong to a higher order of dynamic capabilities (Eisenhardt & Martin, 2000;Teece, 2014;Teece, Pisano, & Shuen, 1997). Exploitation, by contrast, is about the efficient execution or implementation of the new advantage. While the processes involved with the exploration can be location-bound and require a learning portal in an advanced economy, their exploitation can be non-location bound and can be used elsewhere. An example of such location-bound exploration is the establishing of R&D centers in knowledge clusters to tap into that knowledge and develop new features or products. While the knowledge cluster is location-bound, the gained advantage of new features and products and be exploited elsewhere, and is non-location bound. Similarly, the learnings gained from doing business in advanced economies can be transferred to and applied in other contexts. Their exploitation, thereby, becomes non-location bound. However, not all location-bound advantages are transferrable. For instance, a local brand reputation may not have the same value in a different country. For an advantage to be exploitable elsewhere, it needs to carry the same value to customers from different regions. In Figure 3 we distinguish between the location boundedness of exploration and the exploitation of advantages.
In the sample of our case studies, the Chinese MNEs were able to access advantages for which the exploration was location-bound in advanced economies but could be exploited internationally. The bundling of the advantages is where new firm-specific advantage is created. It is created by exploiting country-specific factors with varying degrees of comparative advantage in different functional areas, for example, technology from Europe, labor and manufacturing capabilities from China with building the brand in the target market.

| The learning portal and international modes of entry
Learning portals can be created via FDI and non-FDI entry modes, such as direct exporting. In our case studies, they were created via cross-border M&As, greenfield investments, and smallscale exports. While the learning opportunities in the different entry modes are not mutually exclusive, they provide different strengths in creating learning opportunities for different types of knowledge. Cross-border M&As provide access to cutting-edge technologies and R&D capabilities, sometimes from small "hidden champions" that occupy a leading position in their niche technology, but also through acquisitions of large firms to acquire entire platforms and a broad portfolio of technologies. Greenfield investments allow the creation of R&D centers in mature markets, with which firms can access location-bound capabilities in the form of engineers and can absorb knowledge spillovers from the local excellence clusters. The primary objective of exports is to gain knowledge about customer demands in the mature market that can be fed back to the R&D process in China to improve the product and its technology (Hertenstein & Williamson, 2018). Learning about customer demands and safety and emissions regulations in the mature market helps the firms close the technological gap with Western OEMs and thereby move up the value chain. This knowledge also helps the company position Chinese vehicles differently in the Chinese market to compete with Western OEMs in China directly. Table 3 highlights the different learning opportunities created via the three entry modes.

| Generalizability of the learning portal model and future research
The learning portal model is a model of EMNEs. While we focused on a single country (China) and a single industry (automotive), there is some anecdotal evidence of its applicability to other sectors and other emerging markets. For instance, by 2005, Huawei, which started in 1988, had more than half of its sales overseas (Sun, 2009). Huawei set up R&D centers in developed countries and sold its products rapidly in international markets, both developed and developing. In that sense, Huawei established learning portals for upstream and downstream learning through R&D and sales in a similar fashion as our automotive companies (Zhang, Alon, & Lattemann, 2020 was instrumental in developing the telecommunication industry in China. Later, Huawei established R&D centers with various international companies and had 10 joint research labs by the end of 2005 in addition to six domestic R&D centers across China (Sun, 2009). Huawei also joined 83 international standard organizations with 100 different roles (Sun, 2009). These learning portals helped Huawei launch itself as a formidable global competitor. Other emerging markets have also followed a similar internationalization process. For example, the Indian automotive producer TATA established R&D centers in the United Kingdom and Italy prior to a jump in international exploitation that saw its revenues from international markets increase from just 9% in 2016 to almost 80% in 2018 (Sushil & Garg, 2019). Similarly, Hindalco, Essar, Tata Steel, Mahindra, Ranbaxy, and other Indian multinationals matured by serving domestic markets first (Sun, 2009).
Even firms from mature markets have recently established international learning portals, not as a way to catch up and internationalize, but to keep up with young, leading tech enterprises. For instance, Volkswagen, which centralized its international R&D in the early 2000s (Hertenstein, 2019;Hertenstein & Williamson, 2018), recently established international learning portals in Israel and Silicon Valley. Similarly, younger enterprises such as automotive startups focusing on electric vehicles have followed similar trajectories. For instance, XPeng, founded in 2014, established an office in Mountain View, California early on and made headlines when it hired former Tesla machine learning leader Junil Gu to build an autonomous driving R&D center in the United States in 2018. In 2020, it entered the Norwegian market, which is considered the most developed EV market worldwide. This export experience will help XPeng learn about the most discerning consumers of EVs (non-location-bound advantage), while benefiting from the EV infrastructure (location bound).
One caveat in generalizing our results is that China is truly institutionally different from other countries. As a communist party ruled nation, the second largest economy in the world, and the most influential emerging market, its companies are uniquely positioned. Its Chinaspecific location-bound advantages provide its automobile companies with the largest and fastest growing market and a base for large-scale, high-quality, low-cost manufacturing. In addition, the Chinese government's policy of providing financial support for technological upgrading is an additional benefit. Large state-owned companies could bail out private ones while absorbing their capabilities, as SAIC did with NAC.
Therefore, future research is needed to test the learning portal model in varying contexts. There is some research with similar findings in the chemical industry (Williamson & Raman, 2013) and the wind-turbine industry (Awate, Larsen, & Mudambi, 2015). There are also quantitative studies on EMNEs in fledging industries that used this approach (Anderson & Sutherland, 2015;Huang & Wang, 2011). However, scholars have not specifically pointed out the learning mechanisms of such investments. Moreover, further research is needed to better understand the reverse knowledge transfer and feedback loops to domestic R&D, and the absorptive capacity of EMNEs that is needed to successfully create advantages by establishing such learning portals. Finally, future research could explore the varying behavior of EMNEs and AMNEs in establishing learning portals, as some AMNEs also seem to create R&D centers to access location-bound knowledge, particularly in environments involving radical or disruptive innovation.

| CONCLUSION
Chinese automobile companies are gearing up for the global competitive race through industrial upgrading, innovation, and internationalization. For example, Guanzhou-based Xpeng has built an electric vehicle together with Haima Motors and sells it in the Norwegian EV market. The company has an R&D center in Silicon Valley. The marketing of the car has the look and feel of Tesla, and the company is publicly traded as an ADR under the XPEV symbol on the U.S. stock market. Geely, the Chinese company that bought the Volvo brand in 2010, has developed its R&D by hiring foreign talent from its international competitors and buying distressed assets in developed countries during volatile times (Ramamurti & Williamson, 2019).
In this article, we sought to explain how Chinese MNEs catch up by establishing learning portals in mature markets to acquire knowledge and assets. In doing so, we sought to refine and test Luo andTung's (2007, 2018) springboarding theory. We propose that (a) Chinese MNEs seek to acquire strategic assets such as knowledge, technologies, brands, and R&D capabilities via FDI (cross-border M&As and greenfield investments) and small-scale exports to advanced economies to establish a learning portal inside these economies. Such internationalization steps are for knowledge-seeking reasons. Chinese MNEs do not expand into advanced economies with the aim of building market share in the host country. (b) The knowledge created through the learning portal in the advanced economies strengthens the competitiveness and enhances the firms' position in the domestic market. Chinese MNEs then use their improved position at home as a platform to expand sales into other psychically close emerging economies, before eventually expanding into mature markets.