Creating your own path to move beyond the middle-income trap: lessons from Korea

Abstract The development experience observed in Korea has been a symbol of successful catch-up for several decades. This process allowed its upward transition from middle income to high-income status and has drawn the attention of many streams of scholars. More recently, emergent research has improved our understanding of this experience and its policy implications for developing countries (Lee, 2013; 2016; 2019). This paper proposes a review of what this literature has to say about the mechanisms behind the successful path followed by Korea and a discussion of lessons to overcome the middle-income trap. It is argued that latecomers do not limit themselves to follow the path of technological development of the advanced countries and that alternative paths are possible. The main policy implication for latecomers is that a successful catch-up is possible yet difficult to achieve because it requires taking detours and leapfroging into new technologies.


Introduction
The development experience undertaken in South Korea (hereafter Korea) is recognized as one of the most successful strategies of narrowing the gap with forerunning economies in terms of per capita income and avoiding a fall into the middle-income trap 1 . It is argued that no nation has tried harder than Korea and come so far so quickly (Vogel, 1991) and this process led to the transition from a poor and technologically backward economy to a prosperous and modern country.
There are different interpretations of the rapid development achieved by Korea, yet the approaches that place learning and the mastery of new technologies at the centre of their analysis had a growing prominence and infl uenced the academic and policy debate (Amsden, 1989;Hobday, 1995;Kim, 1997;Pack;Westphal, 1986). More recently, emerging studies have further expanded these ideas to reveal that in the catching-up process the latecomer does not simply follow the path of technological development of the advanced countries . Thus, catch-up is more than a matter of relative speed in a race along a fi xed track, and the technology is not designed into a cumulative unidirectional process.
Accordingly, the performance achieved by Korea is also explained by its ability to explore alternatives in the process of catch-up by skipping some stages and creating their own individual path, which differs from that taken by the forerunners. These fi ndings reveal specifi c conditions for a successful transition to high-income status and have important implications for support policy design for countries that are stuck in the middleincome trap.
Although these new studies have expanded what we know about the successful Korean experience, the systematic attempts to take advantage of these advances in terms of policy lessons for latecomers are remarkably thin. This article fi lls this gap by offering a review of what this emerging view say about the successful path followed by Korea and critically discussing the policy recommendations that are derived to overcome the middle-income trap by less developed countries.
The review of this literature reveals that the income transition achieved by Korea was made possible by building capabilities and promoting business groups and that these efforts have enabled the creation of a development path that differs from the one taken by advanced countries. The main lesson for latecomers is that a successful catch-up is possible yet diffi cult to achieve because it requires taking detours and leapfrogging into new technologies.
The rest of the paper is structured as follows. Section 2 provides a brief historical overview of the upward transition process experienced in Korea. Section 3 describes the path that Korea has followed that has enabled it to achieve its results. A synthesis of lessons for latecomers taking detours and leapfrogging into new technologies is examined in Section 4. Section 5 summarises the conclusions and discusses their implications.

A historical perspective
The concept of "catch-up" refers to the seminal work of Gerschenkron (1962) who describes the economic growth of continental Europe until it reached the United Kingdom at the end of the 19 th century. Later on, Abramovitz (1986) popularized this concept by testing the hypothesis of convergence in productivity between industrialized countries and, consequently, infl uencing its adoption by several development scholars. Thus, catch-up is defi ned as the process of reducing the gap between the forerunning and latecomer economies. The narrowing gap vis-à-vis a leading country is measured in terms of productivity and income (Fagerberg;Godinho, 2005).
This defi nition is coherent with that proposed by Odagiri et al. (2010) when describing catch-up as the process by which a less developing country narrows its gap in income ("economic catch-up") and technological capability ("technological catch-up") in relation to the forerunner. Accordingly, the catch-up could be assessed by several indicators such as income, productivity, and technological capability and the measurement method depends on the objective and scope of the research. Figure 1 compares the gross domestic product (GDP) per capita of select countries as percentages of US income levels between 1950 and 2008. In 1950, the per capita income of Korea was about 9% which was slightly higher than China and India, respectively 5% and 6%, but was substantially lower than that of most other less developed countries such as Brazil (17%) and South Africa (27%). In the early 1980s, Korea reached the level of middle-income countries and by 1983 it had already surpassed the percentages of countries such as Brazil and South Africa 2 . By the year 2000, the per capita incomes as percentages of US in countries such as Brazil and South Africa was only one-half or one-third that of Korea. An examination of this measure leads us to conclude that positive growth does not guarantee catch-up since Korea has also relied on faster growth in order to reduce its gap with its forerunners. Even in the 1980s, Korea had escaped the middle-income trap, unlike many comparable countries that were trapped at an income level of about 20-30 percent of the US level. This achievement is remarkable since between 1960 and 2008 as many as 2 These results are also observed by other measures.  presents a comparative view based on measures such as GDP per capita in constant dollars and purchasing power parity (PPP). 1950(PPP). 1952(PPP). 1954(PPP). 1956(PPP). 1958(PPP). 1960(PPP). 1962(PPP). 1964(PPP). 1966(PPP). 1968(PPP). 1970(PPP). 1972(PPP). 1974(PPP). 1976(PPP). 1978(PPP). 1980(PPP). 1982(PPP). 1984(PPP). 1986(PPP). 1988(PPP). 1990(PPP). 1992(PPP). 1994(PPP). 1996(PPP). 1998(PPP). 2000(PPP). 2002(PPP). 2004(PPP). 2006(PPP). 2008 Brazil thirty upper middle-income countries fell into this trap and only twelve out of 101 economies joined the high-income club (World Bank, 2012). Once these results are realized, an inevitable question is: how did Korea manage to escape the trap, and how did they continue to catch-up despite the odds? How were middle-income countries, based on this experience, able to make a breakthrough in this dismal situation, and build their own path in the upward transition to high-income status? This article attempts to answer these questions based on recent advances in research on how Korean catch-up was enabled by the country´s creating its own individual path.

General and specifi c conditions
The review of the literature that attempts to explain the upward transitions from low to high-income status shows that some general conditions were common to the newly industrialized economies in Southeast Asia (Hobday, 2000). Based on the general conditions investigated, this article discusses three main ones. First, there was a relatively stable macroeconomic environment of which companies, workers and governments were able to take advantage (i.e. when compared to Latin American countries). This scenario was characterized by low interest rates, low infl ation patterns and high savings rates. This relative stability provided an environment for long-term planning and investment (World Bank, 1993).
The second condition was the commitment of governments to eliminate illiteracy, ensure good quality basic education and provide vocational training for industry 3 . There is an apparent consensus on the rationale that the skilled workers (also in the public sector), are a critical element for creation and diffusion of the knowledge that fosters the catch-up with their forerunners. This human-centered perspective, providing universal education to the entire population, ensured an adequate supply of workers, technicians and engineers required by industrial development (Amsden, 1989;Wade, 1990).
The third condition was the industrial policies of each country that were outward-oriented and benefi ted fi rms through the promotion of ex-ports. These interventions encouraged local fi rms to overcome their initial gap when compared to international markets and functioned as an instrument for learning and technological investments. In the case of Korea, this industrial policy is an example of a successful intervention since the exports went from a mere $ 40 million in 1960 to $ 125 billion in 1995 Nelson, 2000). The country migrated from exports in labour-intensive industries such as garments, toys, wigs, footwear in the 1960s to computers, semiconductors, automobiles, and other technology-intensive industries in the mid-1980s.
The successful export promotion policies are also explained by the role of the equipment manufacturing (OEM) system. These systems are a specifi c form of subcontracting where the product is made to exact specifi cations and the latecomer fi rms do not add much value. The Korean OEM system in 1990 represented between 70% and 80% of the total exports of the electronic sector and the three largest chaebols 4 depended on this system in 60% of their exports (Hobday, 2000).
The general conditions were also combined with specifi cities that are fundamental to understanding Korean evolution. According to Kim (1997), four main peculiarities are highlighted. First, the Korean War (1950)(1951)(1952)(1953) generated consequences that transformed a rigid social class system into a more fl exible and less classist society. Moreover, socio-cultural factors were shaped as a result of the War, such as perseverance in situations of confl ict and deprivation, the organizational discipline inherited from compulsory military service, the learning opportunities arising from the management of complex logistical support systems and the reinforcement of a nationalist feeling.
Second, the existence of a strong government that led the industrialization process, setting ambitious targets and guiding the private sector with incentives and sanctions 5 . Firms received incentives and the export performance was the practical yardstick for assessing progress toward the objective of international competitiveness (Pack;Westphal, 1986). The decision-making process was centralized by the president but was relatively 4 Korean chaebols are large conglomerate business groups centered around diversifi ed trading companies. Since the mid-1970s companies have been Korea's engine of exports and economic growth. On the other hand, the emergence of this organizational model also raises inequality indicators, since growth is driven by a few large companies. 5 In contrast to the industrial policy of many countries, for example in Latin America, that have typically used too much of the carrot and too little of the stick (Rodrik, 2004). effective and effi cient since it was supported by a competent bureaucratic apparatus who formulated and executed the development programs 6 .
The third specifi city is the role of the government in using crises as a mechanism of learning and technological transformation. This strategy consists of the deliberate imposition of several crises on fi rms as a way to boost the adoption of ambitious goals. These crises generated intense pressure to create the urgency for changes by causing administrative coalitions to reach consensus on certain organizational goals and leading other members to accept them. Government agencies did not hesitate to impose these strategies by authoritarian means and also by adopting coercion, which made Korean fi rms understand that it would be better to agree to be able to continue to exist 7 . In addition, the creation of crises has generated an increase in the intensity of efforts in the search, both at the individual and organizational level, of alternative actions to transform them into creative ones. Consequently, the government has become an important facilitator of fi rm learning (Kim, 1997).
Finally, a fourth feature is the existence of skilled and dedicated workers who were responsible for running the engines of the Korean industry. These workers were hired for long workdays to make Korea's success a reality. For instance, workers in the manufacturing sector had an average workload of 53.8 hours per week in 1985, while in other OECD countries the average was between 33.1 and 42.8 hours. This determination of Korean workers can be explained by at least fi ve circumstantial factors: obstinacy, which is a national feature; the han psyche 8 ; conditioning during school life; the physical environment; the collective desire to "overcome Japan"; the experience of deprivation (Kim, 1997).

The Korean transition pathway
The existing literature addresses different reasons for explaining the path followed by Korea that enabled the rapid growth of its backward econo-6 See  for a discussion of the role Confucian state in Korea. 7 In other words, the government was much more authoritarian in Korea than in Japan and Taiwan, especially in the 1960s and 1970s. 8 The Korean word han means "resentment" or "grudges" and defi nes a common feature of the mentality of Koreans that has multiple causes and was triggered for different reasons. For a detailed discussion see Kim (1997). my 9 . According to authors such as Young (1995), Kim and Lau (1994) and Krugman (1994), the explanation lies in the high investment rates that allowed for rapid growth through movements throughout the production function 10 . Nelson and Pack (1999) classifi ed these arguments as "accumulation theories" where the relevant portion of the growing production can be explained by the increase in physical and human capital, which almost automatically brings technology as a by-product. Furthermore, what this approach often omits is that such accumulation would have occurred in severely decreasing returns, had it not been for improvements in technological knowledge, ensuring that the available factors were used effi ciently (Pack, 2000).
In contrast, another stream of research argues that physical and human capital accumulation is insuffi cient to account for this unprecedented growth. Accordingly, the learning of new technologies and how to master them is crucial in their analysis (Amsden, 1989;Hobday, 1995;Kim, 1997;Pack;Westphal, 1986). Nelson and Pack (1999) classify these arguments as "assimilation theories" since they stress, in addition to the investment rate, risk-taking efforts, effective learning and innovation itself.
More recently, following these assimilationist views, renewed efforts have been made to improve our understanding of this process and its policy implications in developing countries . These efforts are committed to fi nding "binding constraints" of countries according to their income levels and structural differences (Rodrik, 2006), in contrast to attempts of the mainstream to fi nd a universal factor for economic growth. Based on the Asian experience, these studies have argued that catch-up is not a matter of relative speed in a race along a fi xed track, and technology is not developed into a cumulative unidirectional process, as was assumed by previous studies 11 (Perez, 1988). Therefore, latecomers do not limit themselves to follow the path of technological development of the advanced countries and could skip some stages or even create their 9 As expected, this work is not able to fully cover the reasons faced by the literature and may not adequately deal with approaches such as those centered on the relative endowment of natural resources or geopolitics. An example of this lack is the debate concerning the role of the favourable external context faced by Korea in critical periods of its development. 10 Moreover, some claim this growth was not so remarkable and "mainly a matter of perspiration rather than inspiration-of working harder, not smarter" (Krugman, 1997). 11 An exception is Dahlman et al. (1987) who adequately recognize the specifi cs of the process of combining foreign and local technological elements. own individual path, which differs from that taken by the forerunners Lim, 2001).
The following sections explore in a historical perspective how overcoming obstacles such as the intrinsic diffi culty of building innovation capabilities and the lack of world-class businesses have enabled Korea to cross the narrow transition path to high-income stage. From this experience, some lessons are also presented for the design of policies in developing countries to free themselves from the middle-income trap.

Building innovation capabilities
The crucial step in the Korean upward transition from middle income to high-income status was the building innovation capabilities at the fi rm, sector, and national level. The country was successful in dealing with an obstacle named "capability failure", that is a consequence of the lack of opportunity for effective learning and capability building . This failure is a distinctive feature of latecomer context, where fi rms are endowed with extremely weak levels of capacity, which limits their ability to search and lead in-house R&D. Since incentives are absent, undertake R&D becomes an unsafe investment and with high uncertainty about its return (Lee, 2019). Thus, the Korean public intervention for promoting the enhancement of capabilities was able to offer different methods over their dynamic course of learning 12 .
Another obstacle that deserved special action was the "system failures" that occur when missing or weak connections (and synergies) among actors produce a poor performance of an entire national innovation system (NIS) . In less developed NIS there is a lack of interaction between scientifi c and technological knowledge and institutions that catalyse this interaction are required to overcome specifi c thresholds (Bernardes;Albuquerque, 2003). Therefore, the system failure arises when the country builds up a certain level of capacity and the virtuous circle related to the functioning of dynamic complementarities is not able to work (Lee, 2019).
The main explanation for this failure is the increasing mismatches or misalignments in the accumulation of tacit knowledge among the NIS 12 An example of the results achieved by this policy is the number of business R&D units that expanded from 12 to 1,690 in less than two decades, between 1975 and 1993 (Kim, 1997). agents, which result in an increase in their cognitive distance and a vicious circle of low interaction and learning (Lee, 2019;Nooteboom et al., 2007). The possible evidence that Korea knew how to deal with system failure is that the main indicators of innovation activities such as spending on R&D and the number of researchers observed signifi cant growth in the private and government sectors, albeit in different proportions .
There is a rich literature that demonstrates the Korean process of building capacities at sector level (Hobday, 2000;Westphal et al., 1985) and also the national level (Kim, 1993). More recently, studies have focused on the role of Korean fi rms as a device for monitoring and assimilation of the knowledge produced abroad (Lee, 2019). Since the early 1990s, these fi rms began to establish overseas R&D posts as well as inhouse R&D laboratories so as not to their access to knowledge through simple licensing. This active learning process involves the establishment of Co-Development Contracts and was crucial for the Korean Hyundai Motors in the joint venture with the Japanese Mitsubishi, where the former simply assembled the cars and the latter offered engines and other key components (Lee, 2019).
The fi rm-level accumulation of technological knowledge was also made possible by an adequate intellectual property regime (IPR) which protects both conventional and minor innovations. The 1961 revision of the Korean IPR enabled the intense exploration of utility models (Petit Patents) instead of regular innovation patents 13 . Since the utility models may function as useful alternative outlets for emerging innovation, this reform changed the incentives for innovation 14 . In a context where fi rms are lagging behind and heavily dependent on imported technology and reverse engineering, the reformulation of IPR was critical for learning from foreign technologies (Lee, 2019).
Other forms of intellectual property such as the trademark are also candidates to reveal the variety of capabilities used for Korean technological development. While patents are an indicator for inventive activities, trademarks are an effort to enhance quality and reputation in product markets. Since the 1960s, the change in the application patterns of intel-13 For a more details about the Korean IPR evolution, see Lee and Kim (2016). 14 The adoption of such strategies is much more diffi cult today, especially after Trade-Related Aspects of Intellectual Property Rights (TRIPS), a multilateral agreement towards the global harmonization of IPRs that increased the level of IPR protection in developing countries. lectual property (regular, utility model and trademark), when considering the sectoral differences, reveals the 1990s as the transition period for the Korean catch-up model from imitative to innovation-based technologies (Kang et al., 2019).
Korean ability to build innovative capabilities is also explained by the time cycle of technologies that it has specialized in during different stages of the catch-up process. Since the knowledge becomes obsolete over time, the speed of this obsolescence tends to affect the chances of learning. Over the last three decades, these capabilities were built from specialization in short-cycle, less tacit, and high modularity technologies and made it possible to enter industries such as memory chips, cell phones, and digital televisions. This specialization enabled this country to increase the local creation of knowledge and thus local value-added, given the capacity of short-cycle technology-based sectors to broaden the chances of an economy to rapidly increase its degree of knowledge localization .
However, this is a game of "chasing a moving target" and once a certain level of technological diversifi cation is achieved, new frontiers for building innovation capabilities are open for exploration. Since the 2000s, Korean fi rms like Samsung have been pursuing long-cycle technologies such as pharmaceuticals and medical equipment (Lee, 2019).
The way in which Korea has seized learning opportunities and build capacities from global value chains (GVC) is a process that deserves attention. The evidence reveals that the joining GVCs was successful because it was able to avoid getting stuck in performing low value-added activities without value upgrading. Moreover, this integration was based on "in-out-in-again" strategy, where at the initial stage, the participation in GVCs was undertaken to allow learning from outside, then a reduction in participation creates spaces for an increase of domestic value-added and fi nally a reintegration was promoted after the construction of its own chains (Lee, 2019;Lee et al., 2018).

Promotion of business groups
The upward transition from middle to high income depends on business models of all sizes, however big businesses have a signifi cant causal effect on economic growth compared with SMEs. According to the Schumpete-Feitosa rian framework, large and quasi-monopolistic fi rms are the organizational model capable to exploit economies of scale and scope in the catch-up process 15 . Korea is an example since the number of Fortune 500 companies has increased from eight in the early 1990s to fi fteen in the mid-2010s. The econometric exercises proposed by  shows that this increase had a signifi cant positive effect on the economic growth of that country 16 .
The lack of world-class businesses in developing countries is conceptually defi ned as "size failure". Within these contexts, the market space that would be occupied by large companies (as in developed countries) is fi lled by small and medium-sized fi rms, which are perceived as an insuffi cient organizational form to free latecomers from the middle-income trap. Following this rationale, the catch-up process requires a certain number of big business and many emerging economies have a smaller number of big business than that predicted by their sizes, while many developed countries tend to have a larger number of BBs than that predicted by their sizes, as shown by Table 1. The growth and consolidation of these large businesses are also critical factors that enable the building of innovative capabilities described in the previous section. As big businesses take the form of business groups (BGs) they become a mechanism by which countries achieve high levels of knowledge production and capacity to undertake R&D and marketing activities with higher value-added (Lee, 2019;. 15 Despite this, the benefi ts of big business are a controversial topic in literature. For instance, Fogel et al. (2008) fi nd faster economic growth in countries where big business is less stable over time. 16 The elasticity of Fortune's BBs with respect to GDP per capita growth rate is about 0.01. Put differently, if the number of companies increases by 10%, then the GDP per capita growth rate increases by 0.1% points .
The business group can be defi ned as collections of fi rms bound together in some formal and/or informal ways, characterized by an "intermediate" level of binding (Granovetter, 1995). In this business model, affi liated fi rms are encouraged to enter new markets and for this purpose they receive subsidies at different stages of business evolution, creating incentives for growth and mitigating the risks involved. Moreover, bind fi rms can enjoy the advantages of resource sharing and knowledge spillovers among themselves which increase their innovation capabilities (Cheong, Choo;Lee, 2010). One example is the case of Samsung when it entered the memory chip ventures in the 1980s. The group suffered seven years of losses that were only dampened by the profi ts of other affi liates (Lee, 2019).
Korean business groups, also known as chaebols, originated mostly from small businesses and expanded dynamically through rapid organizational learning 17 . Despite criticism for being one of the causes of the 1997 Asian fi nancial crisis, are still seeing as symbols of economic growth. Many scholars investigate the reasons why many chaebols continued to prosper after surviving the crisis and their re-emergence in the 2000s. Since chaebols have a persistent family-controlled structure, a common argument for this post-crisis turnaround was the focus on corporate governance issues. However, a capability-based view tends to recognize the role of resources and seeks to explain this prosperity due to enhancing the technological capabilities of the chaebols (Choo et al., 2009).
Big business in Korea was also boosted by the creation of state-owned companies (SOEs), although with different potentials to build innovation capacities, according to the technological intensity of the sectors in which they operate. These companies are nurtured by the state apparatus and privatized as they become more competitive by international standards. Leading Korean fi rms that were originally SOEs are POSCO (a global steel fi rm), Korean Air (a global air-carrier), Doosan Heavy Industry (a turbine producer) and SK-Telecom (a top telephone service fi rm).

Skipping stages and path-creating
The fi nal stage of the Korean catch-up was reached after the 1980s when 17 For example, Samsung and Daewoo started as a small import and export business, Hyundai as a small construction company and LG as a face cream maker (Kim, 1997). it virtuously entered into a newly emerging techno-economic paradigm, bypassing investment in an old one and leading the new paradigm era. This leapfrogging was possible due to its ability to promote radical reversals of market shares and occupy central spaces in industries dominated by high-income economies. This "long jump" into higher-end goods was the reason why this country was able to rise beyond middle-income status to become a high-income economy.
The leapfrogging was preceded by a long process of capacity building from the 1960s to the 1980s, in which the country adopted a safe mode of "path-following" catch-up. Since the fi nal stage of the catch-up is surrounded/permeated? by high uncertainties about its results, fi rms and industries were only able to move in that direction when they had built a certain level of capabilities.
At this fi nal stage, the country also registered turning points in innovation activity indicators. In the mid-1980s the share of the private sector in total R&D expenditure surpassed 50% and the R&D/GDP ratio surpassed 1% (Lee, 2016). Table 2 shows that in the early 1980s other potential candidates for the leapfrogging were investing more than Korea in R&D as a proportion of GDP, but this scenario was reversed in the 2000s when the country became a leader in these efforts.  The turning point is also seen from an outcome measure such as patent applications. Table 3 reveals that the selected countries had a small difference in the number of patent applications in the early 1980s and an expressive gap emerges with the Korean leadership in the 2000s. Put differently, these differences represent the priorities that countries allocate to the enhancement of their long-term growth potential, particularly innovation capability. The capacity building and existence of business groups are necessary but not suffi cient to achieve the leapfrogging, which also relies on changes in industry leadership. This dynamic is explained from the so-called catchup cycle framework and deals with diverse factors at the fi rm, industry, and even national institution levels and the interactions among them Malerba, 2017). According to the notion of "windows of opportunity" (Perez;Soete, 1988), the chances for late entrants can appear thanks to three factors: the rise of a new techno-economic paradigm that tends to threaten the advantage of existing fi rst movers or incumbents; the business cycle and/or rapid change in market demand, including the increase in new consumers; and the government generating an asymmetric environment for incumbents and entrants through a range of regulations. Following this framework, the Korean leapfrogging is also explained by taking advantage of the windows of opportunities associated with the rise of digital technologies . Korean fi rms have benefi ted from the growing demand for consumer electronics. An example was the case of Samsung, that was able to take over the Japanese incumbent Sony 18 . The replacement of analogue technologies by digital ones in products such as digital televisions and cell phones provided the real chance to Korea to leapfrog ahead of Japan He, 2016;Jung, 2016).
These results have been achieved in several "paths" along with the evolution of industries and countries. Latecomers fi rms have some options for a possible entry or catch-up, such as path-following, stage-skipping, and path-creating, in which "path" means the trajectory of technologies and "stage" pertains to phases in trajectories Lim, 2001). The path-following catching-up means that the latecomer fi rms follow the same path as that taken by the forerunners but in a shorter period. In the case of stageskipping catching-up, the latecomer fi rms follow the path to an extent but skip some stages, and consequently save time. Finally, the path-creating catching-up option enables the latecomer fi rms to explore their own path of technological development. Kim and Lim (2001) use this model to explain the different technological evolution of the selected industries in Korea and identify the path-creating catching-up in CDMA mobile phones, the path-skipping catching-up in D-RAM and automobiles, and the path-following catching-up in consumer electronics, personal computers and machine tools. These authors also recognize the fi rst two case of catching up as "leapfrogging" 19 .
Another example is the steel industry in the late 1990s, when the stateowned enterprise POSCO was successful in surpassing the leadership of the Japanese Nippon Steel. Posco achieved this position in two phases, initially adopting a path-following strategy of importing mature technologies from Japan and then moving to the stage-skipping strategy by utilizing the price reduction caused by the economic downturn as an opportunity for increasing its facilities and updating its technologies Ki, 2017).

Lessons for latecomers
The early stage of Korean development is one of the main experiences of catch-up based on learning and mastering ways of doing things that are in use in the leading countries. In advanced stages, the success came from its ability to go beyond merely following the path of technological development in advanced countries Lim, 2001). This is due to the fact that although practices in forerunners usually provide a model, the needs of latecomer countries inevitably differ in various and important ways from the existing templates (Mazzoleni;Nelson, 2007). This is also illustrated by the paradox that states "to be similar, you've got to be different" which means that while catch-up means trying to be similar, the long-term success requires a different path from that adopted by developed countries (Lee, 2019).
The Korean upward transition from middle to high-income status was a narrow pathway since these are possible events but at the same time rare to occur. Despite this, lessons on possible paths for latecomers are increasingly needed in a world characterized by cumulative polarization between rich and poor economies in terms of innovative intensity (Castellacci, 2011) and persistent heterogeneity in R&D intensities among fi rms in the same sector (Coad, 2019).
One of the main messages of the model proposed by  is that the successful transition to high-income status in Korea was due to "detours" and "leapfrogging", the former due to its ability to do different things from those done by the forerunners and the latter by doing something new ahead of the forerunners. Based on a comprehensive theory of catch-up economics, this approach consists of "late entry -detours -leapfrogging" and this elaboration is critical for latecomers since it deals with the paradox that is "one can never catch up if they keep catching up, where the former "catch up" means closing the gap or overtaking and the latter "catching up" means imitation" (Lee, 2019, p. 182).
The fi rst detour is the establishment of an intellectual property regime that facilitates imitative innovation instead of following the rigid patent regimes in advanced countries. This goal would be achieved by promoting petit patents instead of regular invention patents, aiming to increase their innovation capabilities in the intermediate stage and generate patentable innovations of higher level in the later stage. This proposal is based on the notion that different forms of intellectual property regime have varying effects on different stages of economic development.
The promotion of trademarks rather than regular patents for invention is another promising strategy for latecomers. This registration effort is a brand strategy and can encourage fi rms to make high-quality products and adhere to a steady level of quality. Depending on the sector, a product made using tacit knowledge can be protected and distinguished from competitors in the market, which guarantees the establishment of market power that can be critical for the growth of fi rms.
The second detour consists in the specialization in short-cycle technology-based sectors and products (i.e., information technology) in the fi rst stages of development and, only at a later stage, in long-cycle sectors and segments (i.e., pharmaceuticals). In sectors such as information technology, the specifi c knowledge and technologies tend to be outdated quickly and frequently and therefore the extensive experience of fi rms in the frontier countries is no longer considered a great advantage. This directing of efforts towards short-cycle technologies is promising because the entry barriers are low and the possibilities for growth are greater due to the high innovation frequency that often disrupts the dominance of the incumbent.
Finally, the third detour refers to joining the global value chains (GVC) by the creation of domestic value-added and reducing the dependence on foreign value-added. There is a certain enthusiasm about the possibility of overcoming the middle-income trap from joining value chains (Baldwin, 2016). In contrast, scholars state that merely joining does not guarantee entry into higher value-added segments and involves the risk of being stuck in low value-added activities (Lee, 2019;Lee et al., 2018). The review of experiences shows that Korea (also Taiwan and China) have been able to create more domestic value-added after learning through their participation in these value chains. In contrast, the integration model with a high degree of GVC participation pursued by Mexico and Thailand has resulted in performing low value-added activities without achieving success in escaping the middle-income trap.
Accordingly, policies for joining value chains in latecomers should pay attention to the Korean lesson on the importance of local creation of knowledge and thus local value-added. This detour is related to the former since the specialization in short-cycle technologies can augment the possibility for latecomers to increase their degree of knowledge localization. Latecomers can take advantage of the link between knowledge localization and short-cycle technologies since the adoption of such technologies suggests a reduced reliance on existing or old knowledge stock.

Concluding remarks
This article reviews emerging studies on the path followed by Korea in the upward transition to high-income status and critically discusses the resulting policy recommendations for less developed countries to overcome the middle-income trap. It is argued that this transformation required some general conditions -such as macroeconomic stability, education system and industrial policies -be combined with circumstantial factors and particularities in Korea. However, these conditions were necessary but not suffi cient to reduce the gap between the forerunning and latecomer economies since failures can arise during the process.
The review of emerging studies reveals that the income transition achieved by Korea since the 1980s was possible by building capabilities at the fi rm, sector, and national level and promoting big businesses in the form of business groups . These efforts have enabled the country to create a different path of development and not merely follow the path of technological development of the advanced countries.
For latecomers, the main lesson from Korea is that a successful catchup is possible but rarely achieved because the transition path is very narrow and requires taking detours and leapfrogging into new technologies. Based on a comprehensive theory of economic catch-up comprising "late entry -detours -leapfrogging" (Lee, 2019), three distinct but complementary detours are proposed, namely: promote minor innovations instead of a high level of innovation via regular patents; specialize in short-cycle technologies rather than long-cycle technologies; and increase the share of domestic value-added rather than depend on GVCs with a low level of domestic value-added. Furthermore, it is also crucial for latecomers to pursue leapfrogging, which involves accomplishing something in advance of the forerunners. YOUNG, A. The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience. The Quarterly Journal of Economics, v. 110, n. 3, p. 641-680, 1995.
The author would like to thank Eduardo Albuquerque and two anonymous referees who provided useful comments on an early version of this manuscript. Usual disclaimers apply.

About the article
Submission received on June 09, 2020. Approved for publication on October 10, 2020.