Industrial Upgrading and Innovation Capability for ...

Industrial Upgrading and Innovation Capability for Inclusive Growth: Experience in East Asia and its Lessons

May 10 2014

Keun Lee (Seoul National University; Center for Economic Catch-up)

Revision of this paper has benefitted from the feedback from several scholars, including Barbara Stallings and Eric Hershberg, around the workshop held in Santiago, Chile, November 2013.

1

1. Introduction

Generating and sustaining economic growth can be a pre-condition for ‘better and more’ distribution and thus inclusive growth. If you sustain growth not just for a decade but more than two decades, such growth is sure to lead to better equity because it will push wage rates to higher levels.

Examples are South Korea and Taiwan in the 1970s and 1980s, where

sustained growth push wage rates and reduces poverty, and eventually lead to lower inequality (Lee 2010; Bai 1982). China also seems to show similar signs as sustained growth has been causing labor shortage and pushing wage rates in the coastal regions (Jin and Lee 2013). Bhagwati and Panagariya (2013: 8) also observed that it is growth since the 1990s that tended to create more jobs and tax revenues to be paid for social welfare expenditure. But, why some growth is not inclusive; one answer is that many of growth episodes in the South tend to be short-lived or not sustained; Jones and Olken (2005) and Hausman et al. (2005) pointed out that many developing countries were able to show growth spurts for a certain period of time (usually less than a decade) but were unable to sustain it over a longer period. The other answer is that many growths have few backward and forward linkages or spillovers. Examples are mining or natural resource-based growth. And example countries are those with somewhat high growth but with high inequalities such as Nigeria or Namibia. In comparison, manufacturing, especially with locally-owned ones, tend to have stronger domestic linkages and spillovers. Low educational attainment of population can be another source of weak spillover and high inequality. Cornia (2014: 70) points out the falling inequality in Latin America since the 2000s and attributed it to lowering educational inequality and rise in the demand for unskilled labor following the adoption of competitive exchange rates for more trades. In sum, having growth more inclusive should start from making growth sustained over a longer period of time and with stronger linkages and spillovers; for instance, more manufacturing than resources; more and wider education; more locally owned enterprises than FDI although FDI can be a channel for initial learning and tech transfer. Now, innovation based on better educated workforce can be called for to make growth more sustained over longer period of time. Without innovation to realize upgrading of the products 2

and sectors, growth based on lower costs (wages) in labor-intensive products tend not to be sustained because its own success will push up wages rates and other competing countries will offer lower wages than yours (Lee and Mathews 2012). This is a case of the middle income trap defined as a situation in which middle-income countries face a slowdown of growth as they get caught between low-wage manufacturers and high-wage innovators because their wage rates are too high to compete with low-wage exporters and the level of their technological capability is too low to enable them to compete with advanced countries (Lin 2012a; Paus 2012; Yusuf and Nabeshima 2009; World Bank 2010 and 2012). The situation is also similar to the so-called ‘adding-up’ problem (Spence 2011) among developing countries which jammed international market by similar products, thus competing each other with ever lower prices. Korea and Taiwan has been an exception to this problem as they have been able to realize two kinds of upgrading, intra-sector upgrading (moving into higher value segments in the same sector), and inter-sector upgrading (entries into new industries with higher value-added), as noted by Lee and Mathews (2012, Lee 2013). China is also showing some signs of similar upgrading (Lee et al 2011). These two kinds of upgrading can also be considered as intrasector and inter-sector diversification (Lee 2013). In this view, pursuing initially resourcebased growth is acceptable but it should also realize these kinds of upgrading to make it sustained with stronger backward and forward linkages. This is where innovation and STI can help. However, switch from low-end, labor intensive production to high-end, knowledgeintensive production is a big jump or break-up, which is not easy. In general, there might be a situation that can be called a ‘capability failure’ (Lee 2013; Lee 2014), which are different from neo-classical market failure (Cimoli, Dosi, and Stiglitz 2009) and from NeoSchumpeterian system failure (Metcalfe 2005). Typical market failure justification of R&D subsidy arises from the perceived positive externality of R&D and its resulting undersupply (Greenwald and Stiglitz 2014). In this view, firms are assumed to be ‘already capable’ of conducting R&D. The problem is considered to be simply about their inability to produce the optimal amount. The reasons for this are sought outside the firm, such as in the capital market or risk market, where government’s corrective action is recommended. However, in most 3

developing countries, private firms are not able to conduct in-house R&D. They consider it an uncertain endeavor with uncertain returns. Thus, the problem is not less or more R&D but ‘zero’ R&D. In fact, R&D-to-GDP ratio becomes flat among the middle-income countries, which means that they are not doing enough R&D (Lee 2013). This is serious because middle-income countries are the ones that should start paying more attention to innovation. This suggests that the failure to innovation is the root of the middle-income trap, as verified by Lee and Kim (2009). In contrast to the typical argument for government activism based on market failure or system failure, “capability failure” is a stronger justification for government activism. In middle income countries where firms have a low R&D capability, a safer way of doing business is to buy or borrow external technologies or production facilities and specialize in less technical methods or assembly manufacturing. To move beyond this stage, effective forms of government activism is needed, not simple by providing R&D funds but using various ways to cultivate R&D capability itself. One form is public and private joint R&D consortium which served as an important vehicle to enter new industries as in the cases of Korea, Taiwan and recently China (Mathews 2001; Lee et al 2005; Mu and Lee 2005). So, this paper will discuss this public and public R&D consortium and other channels of learning for technological upgrading. In what follows, section 2 discusses the linkages between economic growth and equity with the examples from east Asia in comparison with the situation in Latin America. Then, section 3 explains why simple low-end good production is not enough to realize sustainable growth, and why a middle income country should upgrade into higher-ends segments or sectors. Section 4 compares three rationales for government activism and points out capability failure as a more unique justification in the context of developing countries. Then section 5 discusses several ways to overcome the capability failure situation by providing opportunity for technological learning for the latecomer firms. Section 6 concludes the paper.

2. Growth Sustained is a Basis for Inclusive Growth: Asian Experience

4

Literature on the relationship between economic growth and income inequality are huge and diverse, and a strand of the literature that follows the pioneering work of Kuznets (1955) has concentrated mainly on the causal effect of economic growth on income distribution. In this line of study, the conventional wisdom (the so-called Kuznets curve) is that income inequality and per-capita GDP have a relation in the form of an inverted-U: income inequality increases over time while a country develops. Then, when a certain income level is attained, income inequality begins to decrease. Following Kuznets’ insight, a number of studies that use data from developed and developing countries have found evidence in support of Kuznets’ theory (Adelman and Morris, 1973; Chenery and Syrquin, 1975; Lindert and Williamson, 1985).

Although this inverted-U hypothesis has become ambiguous in more recent

literature, especially with regard to developing countries (Oshima, 1991; Ravallian, 1995; Deininger and Squire, 1998; Riskin 2007), there does exist a couple of countries that have followed the prediction, such as Korea and Taiwan realizing the so-called equalizing growth. Given that both countries used to be an agricultural economy, rural surplus labor was an important condition that determines the inequality. The turning point concept proposed by Lewis (1954) predicts that the income gap between the urban and rural sectors will continue until the modern urban sector absorbs surplus labor in the traditional sector and that this turning point will narrow the gap. This conjecture has been confirmed in Korea (Lee 2010; Bai 1982). Korea also had rural surplus labor in the 1950s and 1960s before its take-off since the mid 1960s. Then, Korea had an initial period of worsening inequality caused by the gap between those who stayed in rural area as a part of surplus labor and those who emmigrated to urban areas for higher paid jobs. However, once the fast growth in urban areas had fully absorbed the migrant workers from rural areas by the mid-1970s, Korea soon reached the Lewis turning point with emergence of labor shortages, and inequality started to decline with rising wage rates. The Gini coefficient of Korea increased from 0.32 in 1961 to 0.40 in 1972 but steadily decreased to 0.32 in 1997, the year of the Asian financial crisis. 1 In China, the three decades of sustained growth relieved several hundreds of millions of people out of poverty, thus reducing ratio of population under poverty from a quarter to less

1

Source: WIDER DB (www.wider.unu.edu) 5

than 5% of the population.

However these growth decades have also seen the rise of

inequality, especially urban-rural inequality. In this light, China faces a similar dualistic problem as the past Korea. The urban–rural income gap in China is primarily formed by the considerable amount of agricultural surplus labor, as confirmed by econometric analysis by Jin and Lee (2013) using province-panel data. The analysis reveals a robust and positive relationship between the initial size of surplus labor and the degree of urban–rural inequality. Thus, slow and controlled urbanization and the resulting slow reduction in rural surplus labor have widened the urban–rural income gap and thus increased overall income inequality. Whether China has passed the Lewis turning point and has moved from a period of unlimited supply to an era of labor shortage has been a subject of heated debate (Zhang et al. 2011; Yao and Zhang 2010; Cai 2010; Zhao 2010; Das and N’Diaye 2013). Jin (2013) estimates the size of surplus labor in each province and in the three major regions of China and finds that the eastern region already shows a shortage of labor or a lack of surplus labor, whereas surplus labor remains in the central and western regions. In other words, the fast growing coastal areas have a lower inequality than the slow-growing inland areas, due to the different speed of absorbing surplus labor and sustaining the demand for workers. Actually, Jin and Lee (2013) find that the urban–rural income inequality is aggravated by economic growth during the earlier reform period, but the impact of economic growth on the urban–rural income inequality has become insignificant or even negative during the later period when openness, higher education, or technology variables are controlled. This result is consistent with the original insights of the Kuznets hypothesis. With wage rates increasing every year more than 25% in coastal regions of China, firms are turning to innovation to move into higher end segment of value-chains and sectors (Lee et al 2011), with overall R&D expenditure ratio to GDP approaching 2%. Jin et al (2008) also find that the engine of growth in China has moved from FDI, denationalization, export to innovation and exports. The experience of Asian countries suggests a following pattern of a successful transformation for sustained growth: Initial take-off based on cost advantages -> wage rates increase and control of further inequality -> increased expenditure on education and innovation -> entries into higher value segments and sectors

-> growth sustained over

longer period of time, accompanied by improved inequality.

In this light, the sources of 6

problems in countries fallen into the so-called middle income trap is failure to pay enough for education and innovation and thus to move into higher value activities. In this regards, contrast between east Asia and Latin America is clear, as discussed below. Korea, like the other developing countries, faced external imbalances with persistent trade deficits during the first two decades of industrialization in the 1960s and 1970s. Since the mid-1970s however, the government placed emphasis on technological development by conducting publicly funded and conducted R&D. The results were made available to private firms. This initiated private R&D. These private R&Ds were further encouraged through tax incentives and through initiating public-private joint R&D for bigger and risky projects in the 1980s. In the mid- 1980s, Korea’s R&D/GDP ratio surpassed 1%. The shares of the private sector in the total R&D expenditure surpassed 50%, and even reached around 80% in the 1990s. Through government actions, the number of college students suddenly doubled in the mid-1980s. 2 In five years, the enrollment ratio in tertiary education more than doubled; from 14.7% in 1980 to 34.1% in 1985. This ratio is higher the average level of the developed countries (Lee, 2006). This intensification of R&D expenditure and higher education laid the basis for knowledge-driven growth. The rise in US patents filed by Koreans is proof. In the early 1980s, the number of US patent applications by Koreans were within the range of other developing countries (around 50 or so), or Latin American countries (Table 1 of Lee and Kim 2009). Throughout the 1980s and 1990s however, the number increased sharply to more than 10 times the average of the other middle income countries in Latin America, which for their part, showed little or no increase during the same period. Finally, the same source indicates that in 2000, Korea and Taiwan filed more than 5,000 US patent applications; the other middle income countries filed less than 500 applications. This often unnoticed policy initiative finally succeeded in bringing in for the first time, a trade surplus to modern Korean history; briefly, in the late 1980s, and in a more stable manner in the mid-1990s onwards, as it

2

The exact background for this change in college enrollement is not clear, whether or not the government had a idea that this is a basis for a knowledge-based economy and industrial upgrading. At least it is in sharp contrast to the policy initiatives in the 1960s and 70s when the government promoted many vocational high schools to bring up human capital for low or middle-end manufacturing. 7

laid a basis for upgrading of Korea’s export items from low-end to higher end goods so that it may be able to afford the rising wage rates. 3 Through this, Korea was able to overcome the persistent trap of external imbalances (the stop-go cycle of imbalances) and market reforms like devaluation. In contrast, most of the Latin American countries failed to increase their R&D expenditure. In 2000, none of them had passed the 1% mark in the R&D to GDP ratio (Table 1 of Lee and Kim 2009). In fact, Korea’s R&D/GDP ratio exceeded 2% in the early 1990s, with private R&D accounting for more than 80% of the total. Recently, the same thing happened in another East Asian miracle country - China. China has been strongly pushing for more R&D expenditure. It finally reached the 1% ratio in 2000, earlier than most of the Latin American middle-income countries, and then around 2% recently. In terms of tertiary school enrollment ratio, China had a record 3.4% in 1990, much lower than that of the nine middle income countries’ average of 13.5% in 1980 (Lee, 2006). However, China is catching up very rapidly with the average of middle income countries. This remarkable progress has to do with the higher education revolution in the late 1990s and the increase in the number of students entering college (20% every year) since 1998 (Lee et al 2011).

3. Limitations of Low Cost based Growth Models and Need for ’Double’ Upgrading

Dilemmas and the Crisis of the ‘Standard development’ (OEM) Model In the early days of their take-off, firms from Asia faced at least two important competence disadvantages; notably their dislocation from the major international sources of innovation, and their distance from advanced markets and the user-producer links essential to innovation. However, the OEM (original equipment manufacturing) has been one of the chief institutional mechanisms used to overcome these barriers to entry and to enable technological learning to occur (Hobday 2000); OEM is a specific form of subcontracting where finished products are made to the precise specification of a particular buyer who then markets the products under their own brand name, through their own distribution channels. In Taiwan and 3

Additional exogenous factor was the rapid appreciation of Japanese Yen, which helped to increase price competitiveness of Korean goods that were in competition with Japanese goods, such as memory chips. 8

Korea, OEM accounted for a significant share of electronics exports during the 1970s, 1980s and even the 1990s (Hobday 2000: 133). While latecomer firms readily achieve an early stage of development through producing products designed by others (the so-called OEM model), they face uncertain long term prospects as potential technology suppliers refuse to sell designs or licenses or switch production orders to other lower-wage sites or countries (Lee 2005). The fundamental reason for the unfolding of an “OEM crisis” has to with the rising wage rates upon its success and the difficulty of upgrading into higher value added segment to pay for higher wages. But, upgrading requires acquisition of design capabilities. In the Korean case, products existed to imitate but no designs were forthcoming from incumbent producers since they were reluctant to transfer design technology to potential rivals. In Taiwan, the crisis was such that foreign vendors switched their OEM orders to firms in other lower wage economies like Malaysia. The Taiwanese firms then realized that they had to upgrade their design capability if they were to keep their customers. Specifically, they had to design an “imitative” product by themselves and to start to sell this under their own brand name. Design capability is not easily acquired simply be continuing as a subcontractor or through networking with local producers. The case of Acer in Taiwan shows how difficult it is to move out of the OEM phase, and to move into OBM or Own Brand Manufacturing (Khan 2002; Hobday 2000; Mathews 2002). Acer was the most representative high-tech start-up to emerge in Taiwan in the late 1970s and the early 1980s. Initially, ACER relied on product innovation and OEM with experience gained by individuals who had worked overseas in the USA. The firm started with just 11 engineers in 1976 and achieved total sales of 1.4 billion US dollars in 1993. ACER led the Taiwanese computer industry in the 1980s with a 60% market share in Taiwan. Since then, it has begun to distribute directly to customers abroad in order to challenge other brand leaders and move beyond OEM. But, it has had to retreat from OBM back to OEM or ODM (Own Design Manufacturing) after heavy losses between 1990 and 1993. 4 After that, ACER has been stuck in ODM with a large percentage its sales made to IBM and others. Only very recently, has it been reported to be trying to return to an OBM-

4

The original design manufacturing (ODM) firms carry out a majority of the product design, while allowing their customers’ firms to carry out the marketing functions. 9

based strategy.

Why are successive upgrading and entry needed? The Korean and Taiwanese cases reveal that upgrading in the same industry and successive entry into new promising new industries occurred over the course of industrial development. Our proposition is that unless these two kinds of upgrading are pursued, the chances for successful and sustained catch-up are slim. There are two issues involved here: one from the perspective of the latecomer and the other from that of the front-runner, or incumbent firm. First, from the latecomer perspective, we note that while the current success with the OEM strategy tends to make wage rates rise accordingly, new cheaper labor sites in “next-tier down” countries can emerge to replace a country’s position in the global value-chains. This condition forces firms to move up to higher value-added activities in the same industries. Second, we note that the innovators in the “front-runner” countries tend to generate new higher value-added industries. As innovations arise established industries mature and may degrade into lower-value-added activities, forcing firms to enter newly emerging industries and higher valued added activities. The need for these two kinds of upgrading stems partly from the typical international industrial life cycles such that new industries have tended to be created by the developed world, and partly from the fact that it always used to be the case that latecomer countries and firms tended to inherit these industries after they had become mature and their products standardized (Suarez and Utterback 1995). Given this life cycle, an important feature of successful catch-up has been to be able to enter at an earlier (higher-value-added) stage of cycles as time goes on, which is possible only with enhanced absorptive capabilities. Otherwise, a company may be doomed to remain stuck in lower-wage activities or lowerwage industries, with few chances for long-term success. First, examples of the upgrading in the same industries are numerous in East Asia. For example, semiconductor firms in Korea and Taiwan started from IC-packaging or testing (low-value-added activities), then moved to IC-fabrication and eventually to IC-design (highest valued-added) (Mathews 2004, 2006).

10

Next, there are many cases of successive entry into higher value-added activities found in Taiwan and Korea. For instance, the Tatung company in Taiwan has made successive entries since the 1960s into new activities, starting with black & white TVs in 1964, color TVs in 1969, VCRs and PCs in the mid 1980s, hard disk drives in the mid 1980s, TV Chips/ASICs in the late 1980s, and workstation clones in 1989 (Khan 2002). The Samsung group in Korea is well-known to have made successive entries into new industries over the 60 years of its history. Samsung started with involvement in light manufacturing industries, such as textiles, but then entered consumer electronics, then semiconductors, then telecommunications equipment, and then flat panel displays. The next issue is how to make these double upgrading to happen. Upgrading and structural transformation do not occur automatically even if a country is open to trade and FDI. Rather, it always involves deliberate learning and risk-taking by companies and other public actors, combined with the exogenously open windows of opportunity. The market mechanism serves not as a triggering factor but as a facilitating factor that stimulates risktaking and rewards the successful actors. For example, Taiwan’s successful entries into higher value-added industry segments would have taken a longer time had there been no public-private R&D cooperation, the first successful example of which is the consortium to develop laptop computers (Mathews 2002). However, we should note that there were several attempts and failures prior to this achievement. Such private-private joint effort does not guarantee immediate success, but is the only way out of the old specialization in low end goods sectors, and hence, out of the middle-income trap. In Korean history, the first case of a successful public-private R&D consortium was the development of digital telephone switches. This marked the beginning of the country's emergence as a leader in telecommunication and IT devices, because that success was the source of learning and confidence that, in turn, led to further public-private cooperation in the production of memory chips, mobile phones, and digital TVs. With this series of public-private R&D collaborations to enter new industries,

11

4. Market Failure, System Failure, and Capability Failure 5 The new structural economics of Lin (2012) as well as the initiatives put forward by Cimoli, Dosi, and Stiglitz (2009) argue for the pro-active role of the government in the presence of market failure. 6 Governments are advised to promote infant industries as well as facilitate industrial upgrade and diversification, which are justified by identifying issues of information and coordination failure, as well as external conditions that can be regarded as instances of broadly defined market failure. In Greenwald and Stiglitz (2014), the source of market failure is the fact that knowledge is a public good, and thus industrial policy is necessary due to possible underinvestment in learning when there are flaws in the capital and risk markets, as well as market failure associated with imperfectly competitive industries and a spillover in learning. From this perspective, the actual amount of R&D is often less than the optimal amount that would prevail without market failure. Therefore, government subsidies to support R&D are suggested given the externality involved in the production of knowledge. Another view on the role of

government is the system failure view based on neo-

Schumpeterian economics, specifically the concept of the national innovation system of Nelson (1993), and Lundvall (1992). One of its earliest proponents is Metcalfe (2005), who argues that the process of innovation depends on the emergence and success of innovation systems connecting the various actors (components) engaged in the process., wherein innovation systems consist of firms, universities, public research laboratories, and government and financial institutions. Then, they argue that effective interaction among the actors in the innovation systems does not exist naturally but has to be constructed, instituted for a purpose. In particular, some scholars (i.e., Bergek et al., 2008; Dodgson et al., 2011) observe that system failure often exists where missing or weak connections (and synergies) among actors tend to lead the system to lower performance. 7 In this situation, the main function of the government is not to promote individual innovation events but to set the 5

This section relies on Lee (2014) This nature of market failure is similar to the concept of appropriation failure discussed in Sabel et al 2012 (Ch. 1). 7 A neo-classical counter-part concept to this concept of system failure would be coordination failure, discussed in Sabel et al 2012 (Ch. 1). 6

12

framework conditions in which innovation systems can better self-organize across a range of economic activities. There is a need to re-assess the aforementioned views whether they can be considered as an effective forms of government activism in developing countries, and applicable to their context. For instance, their common and hidden presumption is that the firms and other economic actors are already capable of production and innovation, and that the government must simply try to modify the extent of their activities or promote interaction among them. Especially, in the market failure view, the reasons for failures are sought outside the firm, such as in the capital market or risk market, and these are the areas where the government’s corrective action is suggested. However, the stark reality in developing countries is that the actors, especially the firms, have extremely weak levels of capability. In a number of developing countries, private firms are unable to pursue and conduct in-house R&D, which they consider as an uncertain endeavor with uncertain returns. Thus, the problem is not less or more R&D but ‘zero’ R&D. This is a serious condition because middle-income countries are the ones that should start paying more attention to innovation efforts. This information clearly suggests that this is the root of the middle-income trap. In contrast to the concept of market failure or system failure, this paper emphasizes “capability failure” and suggests specific ways to raise the capabilities of firms in developing countries, including some role by the public sector. In developing countries where firms have a low R&D capability, a safer way of doing business is to buy or borrow external technologies or production facilities, as well as to specialize in less technical methods or assembly manufacturing. To move beyond such states, effective forms of government activism had better include not the simple provision of R&D funds but various ways to cultivate R&D capability itself. More effective and alternative forms of intervention may include the transfer of R&D outcomes performed by public research institutes and a publicprivate R&D consortium, which gained success in Korea and Taiwan. 8 Such direct intervention is important because learning failure happens not only due to the fact that knowledge is a public good but also the fact that there has been no opportunity for 8

For details, see Mathews (2002), Lee and Lim (2001), Lee et al. (2005), and OECD (1996). 13

effective learning due to historically-inherited conditions or policy failure. Seen from this angle, industrial policy is not about choosing winners but about choosing good students and matching them with good teachers or bringing them to good schools. Good schools may be in the form of licensing-based learning (of tacit knowledge) or public-private joint R&D projects, in which direct and cooperative learning take place. By contrast, banks that merely supply R&D money might not serve as good schools. Continuing with this analogy, the market failure view can be expressed as, “I will pay for your school so that you may take more classes,” whereas the system failure may be expressed as, “Go to school and make more friends.” However, both views do not pay enough attention to such factors as the initial aptitude of students, what is taught to them in schools, who the teachers are, and how they teach their students. In the capability view, these aspects are crucial to a successful industrial policy. Thus, the capability failure view essentially believes in the importance of raising the level of capabilities of the firms (students) and the various learning methods to be provided over the dynamic course of learning, not only in the elementary schools but also in the secondary and tertiary institutions. In sum, we need both tuition fees (R&D money) and good friends (linkages to other components in the system) in schools, but the critical factors are the student himself, a good curriculum, a knowledgeable teacher, and an effective teaching method. Table 1 summarizes the aforementioned arguments.

Table 1. The Three Types of Failure Market failure

System failure

Capability failure

Focus

Market institutions

Interaction among actors

Actors (firms)

Source

Knowledge as

Cognition failure

historically given;

public good

from tacitness of knowledge

No learning opportunity

Sub-optimal R&D

Lower R&D effects

No R&D

Solutions

R&D subsidies

Reducing cognitive distance

School Analogy

Tuition support

Making more friends

Relevance

Developing and

Developing and

More unique to

advanced countries

advanced countries

developing countries

Example problem

Access to knowledge and help in learning Targeting student learning

Source: Lee (2014) 14

In next section, we discuss various ways to cultivate innovation capability of the latecomer firms.

5. Ways to Cultivate technological capability: East Asian Experience

An essential requirement to upgrading and needed learning is to establish and initiate its own in-house R&D center. Independent R&D efforts are required because foreign firms would become increasingly reluctant to grant technology licenses to the rising latecomer firms, especially when the latter attempt to enter the skill-intensive markets dominated by the advanced countries. Thus, investment in R&D is required not only for the further absorption of advanced technology, but also for the development of the latecomers’ own technological capabilities. Developing in-house R&D capabilities is critical also because the initial success leads to an increase in local wage rates, resulting in losing competitiveness compared to other economies offering cheaper costs or wages (Lee and Mathews, 2012). With the establishment of in-house R&D labs, the firms may explore more diverse channels of learning and access to foreign knowledge arise from licensing. Arranging access to foreign knowledge and trying new modes of learning is critical, because isolated in-house R&D efforts are often insufficient to build indigenous R&D capabilities. Alternative mode of learning are diverse, including co-development contracts with foreign R&D specialist firms and/or with public R&D institutes, gaining mastery of the existing literature, setting up overseas R&D outposts, and initiating international M&As. For example, it was also from the early 1990s that a small number of Korean firms began to establish overseas R&D posts, mainly to obtain easy and faster access to foreign technology that used to be difficult to acquire through licensing. These overseas posts also served as a window on recent trends in technological development (OECD, 1996). In what follows, we will be introducing several alternatives.

1) Forming and Participation in a public-private R&D consortium.

15

Forming and participation in a public-private R&D consortium can also be an effective school for private firms when their capability is low. Given their low R&D capabilities, the private firms cannot take the lead in the consortium, in which public research agencies play the key R&D roles and teach and transfer the outcomes to participating private firms. We can find many examples of this process from Korea, Taiwan, and other catching-up countries. A noteworthy example would be the government-led R&D consortia in the telecommunication equipment industry, specifically the accompanying local development of telephone switches. This led to the successful localization of telephone switches in the 1980s and 1990s in several latecomer countries, including China, Korea, India, and Brazil (Lee, Mani, & Mu, 2012). Most of the developing countries used to have serious telephone service bottlenecks in the 1970s and 1980s; they had neither their own telecommunication manufacturing equipment industry nor their own R&D program. As a result, they used to import expensive equipment and related technologies, and local technicians merely installed foreign switching systems into the country’s domestic telephone networks. With industrial and commercial bases developing rapidly – along with population growth – a number of countries decided to build their own manufacturing capabilities. Starting with Brazil in the 1970s, followed by Korea and India in the mid-1980s, and finally by China toward the late 1980s, all of these countries crafted a state-led system of innovation in the telecommunication equipment industry, with a government research institute at the core. The research institute developed more or less “indigenous” digital telephone switches that were then licensed to public and private domestic enterprises. In these four countries, a common pattern in the indigenous development of digital switches was the tripartite R&D consortium among the government research institutes (GRIs) in charge of R&D functions, state-owned enterprises (SOEs) or the ministry in charge of financing and coordination, and private companies in charge of manufacturing at the initial or later stages. However, the subsequent waves of industry privatization and market liberalization in Brazil and India versus the consistent infant industry protection in Korea and China differentiated the trajectory of the industries in these four countries (Lee, Mani, & Mu, 2012). At one extreme, the indigenous manufacturers of China and Korea took over from the importers and MNCs. Their enhanced capabilities in wired telecommunication, which were accumulated 16

over the preceding decades, led to the growth of indigenous capabilities in wireless telecommunication as well. At the other extreme, Brazil and India have increasingly become net importers of telecom equipment, and their industries are now dominated by affiliates of the MNCs. As noted by Lee and Mathews (2012), examples from Taiwan include the cases of calculator and laptop PC production. The calculator case is an example of the acquisition of more fundamental design capability or the basic design platform, which is made possible with the help of a government entity such as the Industrial Technology Research Institute (ITRI). Another example is the public-private R&D consortium to develop laptop PCs from 1990 to 1991(Mathews, 2002). This consortium developed a common mechanical architecture for a prototype that could easily translate into a series of mass-produced standardized components. The consortium represented an industry watershed, and even after several failed attempts, it succeeded in establishing new "fast follower" industries in Taiwan.

2) Co-development contracts with foreign/external R&D specialist agencies or firms. A good example of this mode (co-development) is the case of Hyundai Motors of Korea. The main business of the Hyundai group used to be construction, a long-cycle technologybased sector. Hyundai entered the business of automobiles in the early 1970s as an assembly maker for Ford, the US car manufacturer. Such a story is common in developing countries. However, Hyundai Motors and Korea's current status as stronghold of the automobile business would not have been possible without the company's brave decision to cut its ties with Ford and to sell its own brand of automobiles equipped with its own engines. Hyundai then became a joint venture with the Japanese car maker, Mitsubishi, wherein the Japanese company provided engines and other key components, while Hyundai merely assembled them. In that partnership, Hyundai was a licensed producer but not an OEM producer, as it used its own brand in the local and export markets.

However, when Hyundai wanted to

develop its own engines, Mitsubishi (which held 20 percent of the equity) refused to teach the former how to design and produce these engines on its own. Most developing country businessmen would have given up at that point, but Hyundai's founding chairman, Chung Juyung, did not. He decided to spend an enormous amount of money on R&D, with efforts 17

focused on engine development. 9 Fortunately, Hyundai was then able to gain access to the external knowledge of specialized R&D firms, such as Ricardo in England. The process was not easy; Ricardo did not just provide an engine design. It was basically a co-development of a completely new design by the two companies. In fact, the partners had to try more than 1,000 prototypes until they finally succeeded seven years after the project was launched in 1984 (Lee and Lim 2001).

3) Promoting Indigenous Firms by Learning from FDI Firms: the Chinese way Realizing the attractiveness of China's market size and the bargaining power associated with it, the Chinese government has actively approached multinational suppliers to engage in technology transfer and JV negotiations, adopting a purposeful strategy of ‘trading the (domestic) market for (foreign) technology’ (Mu and Lee 2005). Although this strategy has not been entirely successful, there are cases in which it worked and contributed to technological catch-up, with the telecommunications equipment industry as an excellent example. China took advantage of its large market size to pressure the foreign partner to transfer core technology to the local partner. Shanghai Bell and other JV establishments fostered the diffusion of technological know-how on digital telephone switches across the country. Thus, indigenous manufacturers emerged and began to compete directly with JVs in the mid-1990s, initially in rural markets and subsequently in urban markets. This process is called ‘parallel learning’ (Eun et al. 2006). Although a similar diffusion of knowledge also occurred in Southeast Asian countries, China was more successful in turning diffusion into the promotion of indigenous companies. In this sector, China achieved a stage-skipping catch-up. As China had limited experience in developing and producing electromechanical switches, it skipped the development and production of analogue electronic switches and jumped directly to digital automatic switch production. Similar phenomena are taking place in other sectors. Chinese authorities regard a JV as a channel through which learning about technology can take place. Thus, even after its entry to the World Trade Organization, the Chinese government has made no commitment to lift the restriction on the maximum amount of foreign shares (which are usually 50% as in autos) in JVs in key industries, including the 9

For details on the history of Hyundai Motors, see Lee and Lim (2001). 18

automobile, telecommunications, and banking sectors. This continuing restriction on foreign shares is in sharp contrast to the market opening exemplified by the present lowering of tariffs at about 10% or less on average, which is lower than the average in most developing countries.

4) Promotion of the Academy-run Enterprises in Forward Engineering China has successfully reared a number of national champion firms in high technology sectors by exploiting their own scientific knowledge base as exemplified by Lenovo, Founder, Tsinghua Tongfang, and Dongruan. These firms have all been established by and affiliated with academic institutions. These academy-run enterprises are widespread in China. Although their share in the national economy is still minimal, their importance in key high-tech regions, such as Beijing and Shanghai, is substantial. The direct involvement of academic institutions in industrial business is called ‘forward engineering’ (Eun et al. 2006; Lu 2000). In the ‘reverse engineering’ strategy, latecomer firms acquire technological principles by conducting autopsies on final (typically imported) products. Reverse engineering is a bottom-up mode, whereas forward engineering is a top-down mode of technological development, in which the creators (academic institutions) who already possess scientific knowledge further process nascent knowledge until it can be applied to commercial uses. Forward engineering is an inherently Chinese characteristic that differentiates China from other east Asian countries. Taiwan and Korea have rarely exploited their academic institutions for technological development, with the academia mainly supplying engineers to local firms. By contrast, Chinese universities and research institutes, such as those under the banner of the Chinese Academy of Sciences, have played an active role in commercializing new technologies using the results of their research projects.

5) Acquisition of Foreign Technologies and Brand by M&As Until the 1990s, the Chinese outward direct foreign investments were highly regulated compared with the major source countries for FDI. However, a significant shift in policy was made at the Chinese Communist Party’s 16th Congress in 2002 when the Premier announced a new strategy for encouraging Chinese companies to ‘Go Global’ by investing overseas. The 19

policy change seemed to reflect a desire on the part of the Chinese government to acquire foreign technologies and brands, as can be seen from many M&As targeting foreign companies in the manufacturing sector. This strategy serves the objective of saving time for catch-up considering the fact that it will take a long time and great effort to build brands and technologies of its own. A well-known case is Lenovo’s purchase of the PC division of IBM in 2004 and TCL’s acquisition of an European company (Schneider) for electricity technology. The move by BOE, a Chinese cathode ray tube (CRT) maker, to acquire the Korean company Hynix’s TFT-LCD division (HYDIS) had more to do with the technology than the brand. Similar cases of targeting foreign technologies include Geerly’s acquisition of Volvo, D’rong’s acquisition of a German passenger airplane maker (Fairchild-Dornier), and Shanghai Automobile’s acquisition of a Korean automaker SsangYong.

6. Criteria for Sectoral Specialization and Leapfrogging beyond Resource-based development Strategy

Besides building innovation capabilities, a thorny issue is how to choose right sectors or businesses. That is because capability building does not take place in a vacuum but in specific areas of business and sectors. This nature and criterion of sectoral specialization has been a classical issue in economics, in particular unbalanced growth theories. There is an established answer for the low-income groups: specialization based on initial endowments, such as labor and natural resources or comparative advantages associated with resource endowments (Lin, 2012). These industries usually produce low-value-added or low-end goods in the global division of labor, which essentially resembles trade-based specialization. Now, an intriguing issue is to identify a criterion of specialization that can be applied to the group of middle-income countries that strive to upgrade their industrial structure from low to higher value-added. Value-added per worker or labor productivity might be a criterion, but it is too general and there are too many sectors with similar levels of labor productivity. Lin’s structural economics (2012) points out the need for dynamic comparative advantage, suggesting that latecomers should target the industries with latent comparative advantage or mature industries from the countries slightly ahead of them. Nevertheless, though this is a 20

good practical guideline, we still need a more theoretically grounded criterion, or a more specific, differentiating criterion for middle- income countries attempting to mobilize new tools of technology policy. For example, suppose that a country is ready to form a private-public R&D consortium to develop certain technologies or products. In this case, one thorny issue is identifying which technologies or products to target. This question of specialization has also been raised in Greenwald and Stiglitz (2014) who suggest that an economy should choose a sector with greatest learning possibility and capacity but they did not go for further specification. Some scholars, such as Hausman et al. (2007), argue that developing countries should diversify their export products into more sophisticated products, which should be a way for sustained growth. Hausman et al. (2007) developed a measure of the sophistication of tradable products using income level as the weighting factor. However, this made the measure somewhat tautological. That is, it suggests that if a country wants to be rich, it has to move into goods currently being produced by richer countries. In other words, the gradual nature of diversification begs the question of which sectors diversification should encompass first. Another criterion has been suggested by other scholars, and that is technological opportunity as measured by the growth rate of patents per field. This opportunity variable has been used in the literature as an indicator of ‘good’ technological specialization, but a robust relationship has yet to be found; Meliciani (2002) failed to confirm any significant relationship between such specialization and economic growth. Lee (2013: ch 4) also confirms the absence of a significant relationship between the variable of technological opportunity and technological catch-up in the cases of Korea and Taiwan. From a latecomer’s point of view, specialization in high-opportunity sectors is desirable, but involves a greater risk because such sectors are more crowded with already established companies. Having discussed several alternatives, now let me to suggest the cycle time of technology as a criterion for the specialization of middle-income countries. Conceptually, the length of cycle time of technologies refers to the speed by which technologies change or become obsolete over time, causing new technologies to emerge more often. A long cycle time indicates greater importance of old knowledge, hence the greater need for latecomers to study such knowledge, as is the case of medicine, pharmaceuticals and machineries.. When 21

knowledge in the field changes quickly (i.e., essentially meaning of short cycle time), as in typical IT products, the disadvantages for the latecomer might not be susbstantial. Thus, it is advantageous for qualified latecomers to target and specialize in these sectors. Technologies based on short cycle time possess two key properties, namely, the sector has less reliance on existing technologies, and it have a greater opportunity for the continued emergence of new technologies. New opportunities indicate more growth prospects, and less reliance on existing technologies may lead to the faster localization of a knowledge creation mechanism. Additionally, this criterion satisfies the condition of viable profitability and competitiveness. This is because it indicates lower entry barriers and the possibility of higher profitability brought about by fewer collisions with advanced countries’ technologies, less royalty payments, and even first- or fast-mover advantages or product differentiation. The validity of this argument and the criterion for specialization has been verified by extensive econometric analysis conducted by Lee (2013) at the firm, sector, and country levels. Lee’s book show that Korea and Taiwan were similar to other typical developing countries specializing low-end long cycle sectors (such as apparels) but increasingly turn to shorter and shorter cycle technology based sectors (such as Its) from the mid 1980s to seek their own niche different from the incumbent high income countries; table 3.7 of Lee (2013) show that there is no overlaps at all between the top 10 technological fields where largest number of US patents are filed by G5 countries and two latecomers (Korea and Taiwan), indicating a completely different specialization of these two groups.

Only recently in the

2000s these two catching-up economies are trying to get similar to advanced countries by entering into more hard science-based sectors, such as bio and medical products. This long journey, or a detour as called by Lee (2013), is described in figure 1.

[ Figure 1 ]

Typical research, such as diversification thesis, in this area tends to focus on how quickly a developing country can become similar to a rich country in its areas of economic activity. For instance, they point out that the industrial structure of advanced economies are highly diversified and thus recommend that developing countries should also try to diversify. 22

However, these studies do not tell us how latecomers can maintain competitiveness and succeed against the incumbents in the same industries. Furthermore, they do not tell us how to diversify and what direction to first take. In comparison, we suggest that latecomers should avoid both the immediate emulation of rich countries and direct market competition with them. Instead, they should look for their own niches in short-cycle sectors, where they can enjoy a certain level of profitability. In sum, rather than telling developing country to become similar to a rich country, we state the opposite: a developing country’s transition strategy should involve the entering of sectors with short-cycle technologies instead of entering sectors already being dominated by rich countries, namely those with long-cycle technologies. Given that the specialization in short cycle sectors and diversification proceed together, it is not clear which one is ‘causing’ which. We see the cycle times of technologies as a better policy-guide variable because it indicates to us in which direction an economy should diversify. In other words, developing countries should diversify by moving into short-cycle technology-based sectors. Last, the rationality of all middle-income countries specializing in the same short-cycle technologies can be questioned. Such an inquiry is analogous to the adding-up problem, which refers to the risks involved in the labor-intensive specialization practiced by all lowincome countries. In other words, developing countries that compete against each other in the same area of specialization risk eliminating their initiatives and disrupting the very industries in which they have advantages. Thus, a specialization based on factor endowments is relatively fixed, with few opportunities for change. However, specializing in short-cycle technologies does not entail a fixed list of technologies. Instead, the implication in sectors with short-cycle technologies is that new technologies always emerge to replace existing ones. In other words, the criterion for technological specialization is less about the cycle length itself, but more about the technological sectors that rely less on existing technologies and offer greater opportunities with newly emerging technologies. Continuous technological emergence suggests the availability to new entrants of fresh windows of opportunity that are not confined to the old, dominant technologies. This concept is the exact opposite of the product life cycle concept of Vernon (1966) in which latecomers merely inherit old or mature industries (or segments thereof) from the incumbent economies (Lee 2013). 23

So far the discussion has often used Korean and Taiwanese firms and industries as examples of successful catch-up, leaving us with an intriguing question: Did the policy makers in these countries have the criterion of short cycle time in mind as they planned their economic development? While the answer to this question is no, they were in fact always asking themselves, ‘What’s next?’. They looked keenly at which industries and businesses were likely to emerge in the immediate future and thought carefully about how to enter the emerging ones. New or emerging industries or businesses are often the ones with short-cycle technologies because they rely less on existing technologies. Therefore, without specifically planning to do so, in effect the policy makers were always pursing the short-cycle industries. According to Lin’s framework, if policy makers choose to target an industry that is new to a latecomer country but mature in the forerunning countries, the industry must offer a latent comparative advantage (Lin 2012b). After a certain amount of technological capability is built up in the latecomer economy, it can then target another industry that is new to both the latecomer and forerunning economies. This is an effort at leapfrogging, and China is already doing this in various industries. Its efforts to this end are exemplified in particular by its solar power and wind power initiatives. Korea also succeeded in leapfrogging in the 1990s into mobile phones and digital TV for which they used to be a latecomer without solid experiences and knowledge (Ch. 7 of Lee 2013). In general, leapfrogging strategy has a higher chance of success when it is conducted during the time for shifts of technological paradigm or generations (Perez and Soete 1988), and such shifts are happening with higher frequency in short cycle technology sectors. So, one final policy idea for Latin America is to utilize resource-based development as a seed for leapfrog into emerging technologies that combine renewable energies, nanotechnology, bioelectronics, and new materials. Experts on innovation issues in Latin America, such as Perez (2008), have also described resource-based development as offering only a short window of opportunity. However, such development can also serve as the basis for a self-funded developmental leap, possibly into new emerging industries. Perez (2008) makes the following analogy between east Asia and LA: east Asian tigers acquired their initial capability in the fabrication industries of the 1960s and 1970s, and they used this capability to place themselves at an advantageous position within the emerging ICT 24

paradigm-based industries of the 1990s. LA economies can similarly use their current resource exports as a platform and a source of funding with which to begin to enhance their capabilities in preparation for entering the next technological revolution. Other than manufacturing-based leapfrogging, IT service may also be a promising sector to consider as it is also a short-cycle technology based businesses and thus leapfrogging into service bypassing manufacturing.

The advantages of It services, such as low entry barriers,

has already been taken advantage of by Indians in their promotion of IT services, which was really a leapfrogging bypassing IT manufacturing (Lee 2013; Ch. 8 on India). There is already a success case from LA, such as software company, ARTech Consultores in Uruguay, based on public-private collaboration (Sabel et al 2012; Ch. 10).

7. Concluding Remarks

This paper has argued that sustaining growth over longer period of time is one of the key necessary conditions for inclusive growth. Sustained growth can lead to more jobs and wage rate increases, and also higher equality especially when it entails higher backward and forward linkages and is more locally-owned.

A desirable process of successful

transformation for inclusive growth would be as follows: Initial take-off based on cost advantages -> wage rates increase and control of further inequality -> increased expenditure on education and innovation -> entries into higher value segments and sectors

-> growth

sustained over longer period of time, accompanied by improved inequality. However, to sustain growth to move beyond the middle income trap situation, or to be free from the adding-up problem, a country has to accomplish double upgrading, moving into higher-end segment in the same industry and new higher end industries, which require enhanced effort for learning and innovation capabilities. To overcome such ‘capability failure,’ private sectors of a country had better be provided an opportunity for learning and ‘good teachers’ or access of foreign knowledge. The paper has discussed several ways to cultivate technological capabilities of the private sectors of a developing countries or Latin America. They include public-private joint R&D, co-development with foreign R&D entities, promotion of indigenous firms’ learning from FDI firms, promotion of academia-run 25

enterprises, international M&As, and setting up overseas R&D outposts. These channels are more diverse than just public labs, and a key aspect is that these learning initiatives are to be locally-owned and grounded, mobilizing whatever knowledge resources available within a territory, but with arranging access to foreign knowledge bases without which they are doomed to fail. For instance, Hyundai’s engine development did not involve public labs but it worked together with an UK-based R&D company (Ricardo Co.); Fixed line telephone switch development in the 1980s in Korea is an outcome of public-private joint R&D but with licensing arrangement for some foreign technology. Korean consortium developed a world-first CDMA-based mobile phones but a source technology provided by an US-based Qualcomm Co. (see Ch. 7 of Lee 2013 for details of these cases); This paper also suggests that for a developing country to go beyond the middle-income stage, they should implement a technological specialization in a shorter cycle-technology based sector and even leapfrogging into emerging technologies and sectors. This strategy makes sense due to the fact that in sectors with shorter technology cycle times, new technologies emerge frequently and existing ones become obsolete quickly. Thus, the latecomers do not have to master existing technologies dominated by the incumbent. As a caveat, we should also mention the double-edged nature of short-cycle or frequently changing technologies, that is, they can serve either as windows of opportunity or as additional barriers to entry. Although Korea and Taiwan achieved successful catching-up in short-cycle sectors, other lower-tier countries, such as those in Latin America, did not experience success in those sectors (Lee, 2013; Ch. 4). This condition has to do with the notion of truncated learning (Lall, 2000), in which frequent technological changes interfere with the effectiveness of learning, and acquired knowledge becomes obsolete or useless with the advent of new technologies. That is why typical developing countries feel safe or comfortable with staying with low-end long cycle technology-based sectors (such as apparel or footwear) where they do not have to worry about disruptive technology change which might made their past learning or skill obsolete. This explains why the latecomers should build certain level of innovation capabilities through diverse learning modes discussed above before they try leapfrogging. So, for low-income countries, it makes sense to specialize in such low-end, 26

stable long cycle technology based sectors, but if you are a middle-income country aiming ‘high,’ you should start taking some risk. Keeping entering and leapfrogging into emerging technologies requires accumulation of some capabilities in the preceding stages. Several LA countries seem to have already accumulated certain level of capabilities in some sectors required for upgrading or leapfrogging. Although it involves lots of risk, it is the only available way toward high-income country status.

27

References Adelman, I. and C. T. Morris , 1973, Economic Growth and Social Equity in Developing Countries. Stanford University Press, Stanford, CA. Bai, M. K., 1982, "The Turning Point in the Korean Economy," The Developing Economies, 117-140. Bergek, Anna., Staffan Jacobsson, Bo Carlsson, Sven Lindmark, and Annika Rickne, 2008, “Analyzing the functional dynamics of technological innovation systems: A scheme of analysis”, Research Policy, 37(3), 407-429. Bhagwati, J., A Panagariya, 2013, Why Growth Matters: How Economic Growth in India Reduced Poverty and the Lessons for Other Developing Countries. New York: Public Affairs: Cai, F., 2010, “Demographic Transition, Demographic Dividend, and Lewis Turning Point in China”, China Economic Journal 3: 107-119. Chenery, H. B. and M. Syrquin, 1975, Patterns of Development, 1950-1970. Oxford University Press, Oxford. Cimoli, Mario, Giovanni Dosi, and Joseph E. Stiglitz, (eds.), 2009, Industrial Policy and Development. New York: Oxford Univ. Press. Cornia, GA, 2014, eds. Falling Inequality in Latin America: Policy Changes and Lessons: Oxford Univ. Press. Das, M. and P. N’Diaye , 2013, “Chronicle of a Decline Foretold: Has China Reached the Lewis Turning Point?” IMF Working Paper, January 2013. Deininger, K. and L. Squire, 1998, “New Ways of Looking at Old Issues: Inequality and Growth,” Journal of Development Economics 57: 259-287. Dodgson, Mark, Alan Hughes, John Foster, and J. S. Metcalfe, 2011, “Systems thinking, market failure, and the development of innovation policy: The case of Australia”, Research Policy 40(9), 1145-1156. Eun, Jong-Hak , K. Lee, and G. Wu 2006. ‘Explaining the “University-run Enterprises” in China: A Theoretical Framework for University-Industry Relationship in Developing Countries and its Application to China’, Research Policy 35 (9): 1329-46. Greenwald, Bruce, and Joseph Stiglitz. (2014) “Industrial Policy, Creation of a Learning Society and Economic Development,” in J. Stiglitz and J. Lin, (eds), Industrial Policy Revolution I (New York: Palgrav MacMillan). Hausmann, R., Hwang, J., & Rodrik, D.,2007. “What you export matters”. Journal of economic growth, 12(1), 1-25. Hausmann, R., L. Pritchett, and D. Rodrik, 2005, “Growth accelerations”, Journal of Economic Growth 10, 303-329. Hobday, Michael 2000. “East versus Southeast Asian Innovation Systems: Comparing OEM- and TNC-led Growth in Electronics,” in L. Kim and R. Nelson (eds.), Technology, Learning and Innovation: Experiences of Newly Industrializing Economies, pp. 129–69. Cambridge University Press. Jin, Furong, (2013), “Four Alternative Estimates of Surplus Labor and Their Influence on Urban-Rural

28

Inequality in China,” Seoul Journal of Economics 26 (4): 453 - 482 Jin ,Furong, Keun Lee (2013). Growth – Inequality Nexus in China: Lewis and Kuznets Hypotheses. A paper presented at the International conference on the Inequality and the Middle-income trap in China, hosted by the CCER of the Peking University. Jin Furong, Keun Lee, Y Kim (2008). Changing Engines of Growth in China: From Exports, FDI and Marketization to Innovation and Exports. China and World Economy, 16(2): 31-49 Jones, Benjamin F., and Benjamin A. Olken, 2005, “The Anatomy of Start-Stop Growth”, NBER Working Papers 11528, National Bureau of Economic Research, Inc. Khan, Haider 2002. ‘Innovation and Growth: A Schumpeterian Model of Innovation applied to Taiwan’, Oxford Development Studies 30 (3): 289-306. Kuznets, S. (1955), “Economic Growth and Income Inequality,” American Economic Review 45: 1-28. Lee, Keun 2014. “Capability Failure and Industrial Policy to Move beyond the Middle-Income Trap: From Trade-based to Technologybased Specialization,” inJ. Stiglitz and J. Lin, (eds), Industrial Policy Revolution I (New York: Palgrav MacMillan). Lee, Keun 2013. Schumpeterian Analysis of Economic Catch-up: Knowledge, Path-creation and middle income trap. Cambridge: Cambridge University Press. Lee, Keun 2010. ‘Thirty Years of Catch-up in China, compared with Korea’, in Ho-Mao Wu and Yang Yao (eds.), pp.224-42, Reform and Development in New Thinking in Industrial Policy China. New York: Routledge. Lee, Keun 2006. ‘The Washington Consensus and East Asian Sequencing: Understanding Reform in East and South Asia’, in J. Fanelli and G. McMahon (eds.), pp.99-140, Understanding Market Reforms Volume 2: Motivation, Implementation and Sustainability. Gordonsville, US: Palgrave MacMillan. Lee, Keun 2005. “Making a Technological Catch-up: Barriers and Opportunities,” Asian Journal of Technology Innovation 13(2): 97–131. Lee Keun, M Jee, J H Eun (2011). Assessing China's Economic Catch-Up at the Firm Level and Beyond: Washington Consensus, East Asian Consensus and the Beijing Model. Industry and Innovation, 18 (5): 487-507 Lee, Keun, and B. Y. Kim 2009. “Both Institutions and Policies Matter but Differently at Different Income Groups of Countries: Determinants of Long Run Economic Growth Revisited,” World Development 37(3): 533–49. Lee, Keun, and C. Lim 2001. “Technological Regimes, Catching-up and Leapfrogging: Findings from the Korean Industries,” Research Policy 30(3): 459–83. Lee, Keun, and J. Mathews 2010. “From the Washington Consensus to the BeST Consensus for World Development,” Asian-Pacific Economic Literature 24(1): 86–103. Lee, Keun, and J. Mathews 2012. “Ch 6. Firms in Korea and Taiwan” in John Cantwell and Ed Amann (eds.), The Innovative firms in the Emerging Market Economies, pp. 223–48. New York: Oxford University Press. Lee, Keun, and J. Mathews 2013. “STI for Sustainable Development,” UN: committee for development policy, Background paper No. 16.

29

Lee, Keun, C. Lim, and W. Song 2005. “Emerging Digital Technology as a Window of Opportunity and Technological Leapfrogging: Catch-up in Digital TV by the Korean Firms,” International Journal of Technology Management 29(1–2): 40–63. Lee, Keun, S. Mani, and Q. Mu 2012. “Divergent Stories of Catchup in Telecom: China, India, Brazil, and Korea,” in F. Malerba and R. Nelson (eds.), Economic Development as a Learning Process, pp. 21–71. Cheltenham, UK: Edward Elgar. Lewis, W. A. (1954), “Economic Development with Unlimited Supplies of Labour,” The Manchester School 22: 139-191. Lin, Justin Yifu, 2012, New Structural Economics: A framework for Rethinking Development and Policy, Washington DC: World Bank. Lindert, P. and J. Williamson, 1985, “Growth, Equality, and History,” Explorations in Economic History 22: 341-377. Lu, Qiwen 2000. China’s Leap into the Information Age: Innovation and Organization in the Computer Industry. Oxford: Oxford University Press. Lundvall, B.-A., 1992, National System of Innovation-Toward a Theory of Innovation and Interative Learning, London: Pinter Publishers. Mathews, J. A., 2006. Dragon multinationals: New players in 21st century globalization. Asia Pacific journal of management, 23(1), 5-27. Mathews, J. A., 2004, June. Industrial dynamics and corporate strategy: the case of the flat panel display industry. In DRUID summer conference, Elsinore, Denmark , 14-16. Mathews, John A., 2002, “The Origins and Dynamics of Taiwan’s R&D Consortia”, Research Policy 31, 633-651. Mathews, J. A., 2001. National systems of economic learning: The case of technology diffusion management in East Asia. International Journal of Technology Management, 22(5), 455-479. Meliciani, V. 2002. ‘The impact of technological specialisation on national performance in a balanceof-payments-constrained growth model’, Structural Change and Economic Dynamics 13 (1): 101– 18. Metcalfe, JS., 2005, “Systems failure and the case for innovation policy”, in Patrick Llerena, Mireille Matt, Arman Avadikyan (eds.), Innovation Policy in a Knowledge-based Economy: Theory And Practice, Springer. Mu, Qing, and Keun Lee 2005. ‘Knowledge diffusion, market segmentation and technological catchup: The case of the telecommunication industry in China’, Research Policy 34 (6): 759-83. Nelson, R. 1993. National Innovation Systems: A Comparative Analysis. New York: Oxford University Press OECD, 1996, Reviews of National Science and Technology Policy: Republic of Korea, Paris: OECD. Oshima, H., 1991, “Kuznets Curve and Asian Income Distribution” in T. Mizoguchi, N. Takayama, M. Kuboniwa, and T. Tsuru (eds.) Making Economies More Efficient and More Equitable: Factors Determining Income Distribution. Kinokuniya and Oxford University Press. Paus, Eva, 2012, “Confronting the Middle Income Trap: Insights from Small Latecomers,” Studies in Comparative International Development 47 (2): 115-138 Perez, Carlota, 2008. ‘A Vision for Latin America: A Resource-based Strategy for Technological Dynamism and Social Inclusion’, Globelics Working Paper No.WPG0804. 30

Ravallion, M., 1995, “Growth and Poverty: Evidence for Developing Countries in the 1980s,” Economic Letters 48: 411-417. Riskin, Carl, (2007), “Has China reached the Top of the Kuznets Curve,” in V. Shue and C. Wong, eds. Paying for Progress in China. Oxon: Routledge. Sabel, Charles, E. Fernandez-Arias, R. Hausmann, A. Rodriguez-Clare, and E. Stein, eds. 2012, Export Pioneers in Latin America. Washington DC: IDB. Spence, Michael, 2011, The Next Convergence: the future of economic growth in a multispeed world. New York: FSG Books. Suarez, F. F., & Utterback, J. M., 1995. Dominant designs and the survival of firms. Strategic management journal, 16(6), 415-430. Vernon, Raymond, 1966, “International Investment and International Trade in the Product Cycle”. The Quarterly Journal of Economics 80, 190-207. World Bank 2012. China 2030: Building a Modern, Harmonious, and Creative High-Income Society. Washington, DC: The World Bank. World Bank, 2010, “Exploring the Middle-Income-Trap”, World Bank East Asia Pacific Economic Update: Robust Recovery, Rising Risks, vol. 2, Washington, DC: World Bank. Yao, Yang, and Ke Zhang, 2010, “Has China passed the Lewis Turning Point: A Strutural Estimation based on Provincial Data,” China Economic Journal 3, No 2: 155-62. Yusuf, S. and Nabeshima, K., 2009, “Can Malaysia Escape Middle Income Trap? A Strategy for Penang”, Policy Research Working paper 4971, Washington, DC: World Bank. Zhang, Xiaobo, Jin Yang, and Shengjin Wang, (2011), “China has reached the Lewis Turing Point,” China Economic Review 22, No 4: 542-554. Zhao, X., 2010, “Some Theoretical Issues on Lewis Turning Point,” Economist 5: 75-80.

31

Figure 1: Korea’s Successive Entry into Shorter Cycle Technology Sectors and the Resulting Diversification

Notes: The numbers in vertical axis is the cycle time of technologies measured by the mean backward citation lags of the Korean held US patents. For instance, 9 means the cycle time is average 9 years after which patents in the field are no longer cited or useful. The names of sectors are positioned along the average cycle time of Korea near the year when Korea entered that sectors, ranging from low end long cycle sectors (apparel), medium-cycle sectors (steel and autos), short cycle sectors ( many IT goods), and finally high-end long cycle sectors (bio and medicine). The first turning point indicates the point from when (mid 1980s) Korea started to turn decisively into shorter cycle technologies, and the second turning point indicate that since the 2000s Korean industries are getting mature or close to advanced countries and are trying to enter into more hard science based (long cycle) sectors. For detailed explanation, see Lee (2013: Ch 1, Ch 6, Ch 9).

32