The “Middle Income Trap” in Economic Growth in Different Economies
Abstract: Nowadays, more than 50% of the world population live in middle-income economies. Economies in the middle-income development stage are confronted with a number of challenges, such as economic restructuring, industrial upgrade and income growth. Therefore, academia around the world have paid much attention to theoretical and empirical researches of the “middle income trap” and long-term trends of global economic growth. Based on analyses about the long-term growth path of the world economy and the characteristics of different stages of economic development, this paper defines the“middle income trap” and its characteristics and examines, through a large number of cases from different economies, the reasons why “growth slowdown” arises during the middle-income development stage. These theoretic and empirical researches have reference value and heuristic meaning for China to actively respond to the challenges of the “middle income trap”.
Keywords: “middle income trap”, middle-income development stage, growth slowdown JEL Classification: F11
In the 1950s, rapid development of the global economy brought many countries and regions into the middle-income rank. However, few have since advanced into high- income economies. Instead, most fell into the “middle income trap”. According to the World Bank (2012), among 101 economies that were in the middle-income rank in 1960, only 13 (Equatorial Guinea, Israel, Mauritius, Greece, Ireland, Portugal, Spain, Puerto Rico, Japan, South Korea, Singapore, Hong Kong and Taiwan) successfully attained high- income status in 2008. In recent years, academia around the world paid much attention to theoretical and empirical researches of the “middle income trap” and global economic growth trends (e.g. Yusuf and Nabeshima, 2009; Tho, 2013; Egawa,2013; Berg, Ostry and Zettelmeyer,2012 ).
China currently is the largest middle-income economy and most populous country in the world. How to tackle the challenge of the “middle income trap” is now and will in foreseeable future be a formidable economic development problem for itself and has significant impact on the global economy as well (World Bank, 2012; Wu, 2013; Kuijs, 2010). “The rapid economic growth of so-called emerging markets is one of the leading storylines of our age and arguably the most important economic development affecting the world’s population in the first decade of the 21st century. It has lifted millions of households out of poverty. It has accounted for the vast majority
of global growth in a period when the advanced countries have been economically challenged and financially troubled. For some time now the question on everyone’s mind has been how long this rapid growth can continue, in emerging markets in general and the group’s largest and most economically dynamic member, China, in particular” (Eichengreen et al., 2013). After growing rapidly for more than 30 years, the Chinese economy now faces transition from high-speed growth to a “new normal” of midhigh- speed growth. During 2001- 2012, the Chinese economy grew at an average annual rate of 10.1%, and from 2013 to 2015, China’s GDP grew by 7.4%, 7.3% and 6.9%, respectively. The research group of the Frontier Theories of China Economic Growth forecast that China will grow at an annualized rate between 6.4% and 7.8% in the next five years, while the Development Research Centre of the State Council’s “Midand Long- Term Growth Research Group” predicted that China’s average annual economic growth rate would be around 6.5% in the coming decade. Since 2011, China’s double- digit economic growth has given way to mid- high speed growth, which is thought to be a symptom of the “new normal” of the Chinese economy.
For all of this, this paper will put the “middle income trap” under the context of longterm economic growth trend to look into its characteristics, its root course, and the ways to cope with it, through which we expect to offer a wider horizon and reference frame for researches on the transition of China’s economic growth model at the present stage.
2. “Middle Income Trap”: Definition and Characteristics 2.1 What Is the “Middle Income Trap”?
The idea of a “middle income trap” was first put forward in the World Bank’s East Asian Visions 2006. In its subsequent reports, the World Bank described the “middle income trap” as the inability of developing economies to overcome the heavy obstacles in their course of economic development. In particular, after entering the middle- income stage, their economic development runs out of steam so that they cannot successfully change their economic development model. This makes it impossible for them to cross the growth threshold and transition into the high- income stage. Instead, they are trapped in secular stagnations. For example, in the 1970s, some Latin American economies including Brazil, Argentina, Chile and Mexico, entered themiddle- income stage one after another. However, by 2007, their GDP per capita was still hovering around 3,0005,000 USD. In Asia, Malaysia and Thailand were the poster children of how an economy slowed down and fell into the “middle income trap”. Despite the Asian financial crisis in 1997, the productivity rates of these two countries had almost caught up with developed countries by the end of the 20th century. Nonetheless, their production and export activities, mainly based on labor-intensive models, have never changed in past twenty-odd years. Meanwhile, they had to face fierce competitions from other lowcost producers, such as China, India and now Vietnam and Cambodia. As a result, they fell into secular stagnations.
Some researchers have inquired into the “poverty trap” ( Azaridis, 2004; Bowles, Durlaufand Hoff, 2006) and gained insights as to the reasons why some economies fell into permanent and deep poverty. Others have analyzed the growth dynamics and industrial conversions in the long-term economic growth path ( Galor and Moav, 2002; Hansen and Prescott, 2002). These researches inspired researchers to describe the “middle income trap” from the perspective of industrial structure upgrade by summarizing the development stages in different economies. Ohno (2009), borrowing from the “catching-up industrialization” theory, divided the development process of an economy into four stages. Initially there is the primary stage in which society is mono- cultural and agriculture is only subsistence agriculture. Once FDI flows in, it would enter the first stage, in which the state, under external technological guides, has the competence to engage in some simple manufacturing production. Then, after industries begin to take advantage of clustering,
the second stage is reached. By then, pillar industries have emerged within the economy, though external technological guides are still needed. To enter the third stage, the state has to absorb technologies, proficiently harness management skills and technologies, and develop the ability to produce high- quality goods. At last, after acquiring innovating capabilities, the state hasentered the fourth stage, in which it becomes fully competent to design new products and innovate and has the potential to become the global leader of product innovation in the industry.
According to Ohno ( 2009), there existed a “glass- ceiling”, an invisible threshold to cross the “middle income trap”, for ASEAN economies when they were crossing the seco nd stage into the third stage. Furthermore, a large number of economies in the primary stage fell into stagnation due to the lack of FDI flowing into the manufacturing sector. Even if they successfully crossed the first stage, it would be more difficult for them to cross subsequent stages. For example, most economies were trapped in the second stage because they failed to raise their human capital level. Few ASEAN economies, including Thailand and Malaysia, managed to break the invisible “glass- ceiling” of the manufacturing sector between the second and the third stages. Similar situations also occurred in most Latin American economies. Although they had achieved rather high-income levels in the 20th century, they are still in the middle-income stage today.
In conventional growth theories, “traps” are typically thought to be a hyper- stable equilibrium. A number of academics argue that, according to endogenous growth models and cross- national comparative analyses, rich economies tend t o drive growthby technological advances, which in turn make them richer, while poor economies grow more quickly due to their comparative advantages in the manufacturing sector. Economies between these two groups will have a dilemma (Acemoglu,1997, 2005; Barro,1997). Eechhout and Javanovic ( 2007) further emphasized that the labor force in rich economies tended to be more skilled and therefore more competent to undertake management and R& D jobs, while poor economies could increase the supply of unskilled labor more quickly. Economies neither rich nor poor have no advantages in supplying either kind of labor. The Asian Development Report ( ADB, 2010) described economies caught in the “middle income trap” as follows: they cannot compete with low-income and lowwage economies in manufacturing export, nor can they compete with developed economies in high-tech innovations. These economies failed to transition in time from the low-cost labor and low-capital resource-driven growth model to the model driven by productivity and innovation. Spence ( 2011) argued that economies whose national income per capita have reached 5,00010,000 USD range would confront growth transitions. According to Spence, “in this stage, continually rising wages led industrial development that previously fuelled economic growth to lose its strengths. The economy should move these labor-intensive industries to countries with lower wages and turn to develop new capital-, human capital-, or knowledgeintensive industries that create values. ”
2.2 Characteristics of the “Middle Income Trap”
The so- called “middle income trap” correlates with an economy’s national income per capita and its growth rate. In 2012, the World Bank set the standard classifying different economies into different income stages. According to the Atlas method, economies with a GNI per capita of no more than 1,005 USD are classified as low-income, a GNI per capita between 1,005 and 12,276 USD as middleincome, a GNI per capita of more than 12,276 USD as high-income. Middle-income economies can be further divided into lower-middle income (with a GNI per capita between 1,005 and 3,975 USD) and upper- middle- income ( a GNI per capita between 3,975 and 12,276 USD). Based on this standard, in 2010 there were 29 lowincome economies, 31 lower- middle- income economies, 30 upper-middle-income economies, and 34 high-income economies.
Based on the World Bank’s classifying standard and considering the availability of data, Felipe et al. (2012) divided economies in the world into four groups according to each economy’s GDP (measured by 1990 Purchasing Power Parity ( PPP)): those with a GDP per capita under 2,000 USD are categorized into the low-income group; a GDP per capita between 2,000 and 7,250 USD into the lower- middleincome group; a GDP per capita between 7,250 and 11,750 USD into the upper-middle-income group, and a GDP per capita above 11,750 USD into thehigh- income group. Thus, among 124 economies covered by them in 2012, 40 were low-income, 52 were middle-income (38 lowermiddle and 14 upper- middle), and 32 were high-income. Compared with the World Bank’s classification, the number of high- income economies was nearly the same, while the number of middle-income economies, especially upper-middle-income economies, was less than half.
Determining an economy’s current development stage based on its income level requires, for the convenience of comparison, converting its income using the PPP method. For example, data in Penn World Table (PWT) (2013) and Maddison (2003) were converted by PPP to a comparable basis. WDI (World Development Indicators), the database of the World Bank, offers two sets of data calculated by different methods, with the Atlas method exchange rate method as its distinguishing criterion. Due to the World Bank’s influence, this distinguishing criterion has been adopted by many academics. However, to compare data from different sources, researchers often faced the problem of cumbersome conversions and measurement benchmarks. Therefore, Ohno (2009) and Woo ( 2012) introduced another classifying method. They used an economy’s income level relative to the most developed country ( often the U. S.) as an indicator to determine its current development stage. Economies whose relative income levels fell in the middle range should be considered middle-income economies. This relative income makes up the “catch-up index” (CUI). A decreasing CUI for a given economy during its developing process means it might have fallen into the “middle income trap”. This classifying method, using the relative income level of an economy as the criteria to define its development stage, was somewhat arbitrary.
A number of researchers examined the “middle income trap” problem from the perspective of economic convergence. They concluded that middle- income economies could cross the “middle income trap” through converging toward developed economies. The time the convergence would take can be calculated through the formula below:
(1) Where GI stands for the gap with the income level of developed economies; gM is the average growth rate of overall GDP in the middle- income economy; TC is the time the convergence would take. We can see from formula ( 1) that the lower GI is, the shorter the convergence would take. Meanwhile, if a middle- income economy achieved relatively more rapid growth than its developed
counterparts, that is, the larger is, the shorter TC would be, and thus the easier it would be for it to cross the “middle income trap”.
Other researchers examined how much time it would take for different economies to cross a certain development stage. Felipe et al. ( 2012) measured how long an economy that had successfully attained upper- middle or high- income status had stayed at the previous development stage. If it spent more time than average to cross the lower- or upper- middleincomestage, then it was thought to have fallen into the “lower- middle income trap” or the “upper- middle income trap”. Studies in this area showed that the average time the lowermiddle-income economies spent to attain uppermiddle status was 28 years, while economies that successfully entered the high- income ranks stayed 14 years on average at the upper middle- income stage. That is, lower- middleincome economies had to grow at least 4.7% a year to avoid falling into the “lower- middle
income trap”; meanwhile, upper-middle-income economies had to grow at least 3.5% a year to avoid falling into the “upper- middle income trap”. Thus, the problem of whether an economy would fall into the “middle income trap” could be converted to another problem – whether or not a lower- middle- income economy could cross the threshold of upper- middle- income in no more than 28 years, and whether or not an upper- middle- income economy could cross the threshold of high- income in no more than 14 years. That said, it is never an easy task to catch up with developed economies. Im and Rosenblatt ( 2013) argued that, based on the current gap between middle- income countries and developed economies, namely the U. S. and other OECD members, and assuming that developed economies on average grow at an annual rate of 1.8% while middle- income countries keep their average annual growth rate in the past 30 years, it will take at least 50 years for them to catch up with their developed counterparts.
Still other academics examined the problem of the “middle income trap” from the perspective of growth path divergence. These researches could be tracked back to Quah ( 1993). He creatively employed the income transformation matrix to estimate the probabilities of various economies retreating to, remaining at or jumping into different development stages, and found that in the long- term, rich economies would become richer and poor economies poorer, while economies in the middle range would diverge into these two extremes, so that in the end there would be no middle- income economies at all. Advocates of this theory noted that there existed a rather big difference between the effects of growth in middle-income economies and that in developed ones. Therefore, the “middle income trap” could be interpreted as the divergence of long- term growth among different middle-
income economies, some of these economies downgraded to low- income economies while others leapt upward into high-income economies. Researches about income distributions among various economies in 1962—2008 using Quah’s method showed that the probability of a middleincome economy falling back to the low-income group was far higher than the probability of it leaping upward into the high-income group.
2.3 The “Middle Income Trap” and Growth Slowdowns in Various Economies
The “middle income trap” correlates closely with the loss of growth momentum and growth slowdown. Aiyaret al. （ 2013） compared a number of East Asian economies that successfully crossed the “middle income trap” with Latin American economies that suffered growth stagnation. Figures 1 and 2 show the results. Figure 1 depicts how GDP per capita in these economies evolved relative to the U.S. after they had achieved the 3,000 USD threshold. Compared with other economies, Latin American countries, including Mexico, Peru and Brazil, experienced prolonged growth paths after their national income had reached 3,000 USD. That is, their income per capita achieved 3,000 USD very early, but had swung around that number since. Among the so-called “Asian Tigers”, South Korea and Taiwan took off late but grew rapidly, and their national income as a percentage of the US’s leapt from 10%-20% to 60%-70%. On the contrary, Latin American economies suffered stagnation (Brazil and Mexico) and even recessions (Peru).
Other Asian middle- income economies performed midway between rapidly growing East Asian Tigers and sluggish Latin American countries. According to the World Bank’s classification of growth stage for different economies published in 2012, China crossed the low-income threshold (a GNI per capita of 1,005 USD) and became a middle- income economy in 2002, and then in 2010 crossed the lowermiddle-income threshold (a GNI per capita of 3,975 USD) thus attained upper-middle-income status. By far for China, while it has been in the middle-income stage for no more than 20 years,
its economic growth trajectory looks far more remarkable than that of other Eastern Asian economies that successfully crossed into the high-income ranks in their early take-off times. Meanwhile, Malaysia’s growth performance, whether in absolute or relative levels, is far more impressive than that of Latin American countries. Thailand’s growth path looked like that of Brazil and Mexico in their early days. Indonesia experienced mediocre economic growth even compared to Latin American countries.
Figure 2 further compares logarithmically the national incomes of these economies, with the slope of each curve representing the corresponding economy’s growth rate. We can see in Figure 2 that the twenty- year- or- more period of higher economic growth (though not as high as East Asia) enabled most Latin American countries to rapidly enter the middle- income stage. However, after that they suffered obvious growth slowdowns, which made their growth paths diverge from those of other East Asian economies. As a result, growth stagnation, or even falling into growth traps, correlates closely with growth slowdowns.
3. An Inquiry into Long-Term Economic Growth Trends
3.1 Different Stages in Economic Growth
(1) Growth acceleration
Research by Hausmann et al. ( 2005) was among the earliest analyses of stages in economic growth. It described the process of rapid growth in an economy by identifying the conditions that have to be met for growth to accelerate. First, the growth rate of period t, gt, is defined as the growth rate of the minimum variance of income per capita (y) from period t to period t+1, which is presented by gt, t+ n. That is,
So the change of growth rate in period t can be presented as the difference of average growth rates of different n years:
Then, the period of growth acceleration can be further defined as the period of rapid growth that satisfies the three conditions below:
The first condition, inequality (4), requires that the average growth rate of the n years after t has to be more than 3.5%; the second condition, inequality ( 5), requires that the growth rate has to accelerate by at least 2%; and the third condition, inequality ( 6), requires that the n year average income per capita after the growth acceleration has to be no less than the maximum n year average income per capita before t.
(2) Growth slowdown Eichengreen ( 2012), based on the work of Hausmann ( 2005), demonstrated that growth slowdowns have to meet three similar conditions. First, the average GDP growth rate before slowdown should be no less than 3.5%. Second, when growth slowdown happens, the average GDP should reduce by 2 or more percentage points. Lastly, when growth slowdown happens, national income per capita should be higher than 10,000 USD.
In this equation, yt is GDP per capita measured by constant 2005 USD, gt is GDP growth rate, gt- n, t and gt, t+ n represent the n year average GDP growth rates from t-n to t and from t to t+n, respectively.
Based on this criterion, Eichengreen screened the growth slowdown periods of economies all around the world from 1957— 2007, and found that most economies (42) had experienced at least one growth slowdown
period; some had experienced as many as a dozen of times, such as Greece (10 times), Japan ( 12 times), Puerto Rico ( 10 times); and most economies had experienced prolonged growth slowdowns, such as Israel (from 1970—1975), Chile (1994—1998) and Taiwan (1994—1999). However, Eichengreen only focused on growth slowdowns suffered by economies with national income per capita higher than 10,000 USD. (3) Stagnation
Reddy and Minoiu ( 2006) studied the problem of stagnation using time- series data of GDP per capita of economies all around the world from 1960 to 2001. They defined the stagnation period as the period between the onset of stagnation and the time when the first turning point happens. According to their definition, when the real national income per capita of an economy fell below that of any time in past two years but was still above that of any time in the subsequent four years, the economy could be said to have entered a stagnation spell. When the real national income per capita of an economy was at least one percentage point higher than that of the previous year, but still at least one percentage point lower than that of the subsequent year, the economy could be said to have reached a turning point. Meanwhile, the authors also discussed the length and depth of stagnation. “The length of stagnation” measured how much time passed after the onset of stagnation. “The depth of stagnation” was represented by the difference of national income per capita during stagnation as a percentage of the corresponding income at the end of the stagnation, in which the difference of national income per capita means the difference between the corresponding income per capita at the onset of stagnation and the lowest income during stagnation.
Based on the definition of stagnation described above, Reddy and Minoiu ( 2006) screened the available growth spells of a GDP per capita in constant LCUs of 119 economies from 1960 to 2001. The situations of stagnation in various economies are shownin Table 1. Their study showed that among the 119 economies they covered, 72, or 60.5%, experienced significant stagnation. OECD members in general experienced less stagnation: only 4 economies in 24 of them, or 16.67%, experienced stagnation. All other economies experienced somewhat serious stagnation, 62.51% of landlocked economies and 80% or more of other economies experienced stagnation. The most serious stagnation happened among Latin American economies, 22 of 24, or 91.67%, of which experienced stagnation. Meanwhile, we found that growth stagnation was widespread among economies whose growth relied heavily on exporting primary goods. For example, 8 of the ten members of OPEC, or 80%, experienced stagnation . While the reexisted some differences between the populations of primary
goods exporters screened out by different measurements (the number of the primary goods exporters screened out by the first measurement, primary goods exporters I, was 32, and that screened by the second measurement, primary goods exporters II, was 12), most of them (87.5% and 83.33%, respectively) suffered stagnation.
Table 1 specifically describes the characteristics of stagnation, including the depth and length, in different types of economies. We can see that there existed a lot of differences among the characteristics of stagnation in different types of economies. In terms of the average depth of stagnation, OECD economies were 0.03, while Latin American economies and OPEC members were 0.24 and 0.97, respectively. The average length of stagnation varied from 7 years ( OECD members) to 18 years ( primary goods exporters II). Some economies experienced as few as 1.3 times of stagnation (Latin American countries and OECD members) while others experienced as many as 1.8 times of stagnation (OPEC economies).
Generally speaking, among all economies, those that relied on primary goods development suffered stagnation with typical characteristics. For example, OPEC suffered stagnation with the most extreme depth (0.97) as well as the most frequent stagnation spells (1.8 times). Primary goods exporters II screened out by the second measurement suffered the second deepest (0.89 in average) and the longest (18 years on average, almost covering half of the period covered in Table 1) stagnation. Meanwhile, SubSaharan African economies experienced deeper and longer stagnation than their Latin American counterparts, which means the former contracted more and suffered longer in stagnation than the latter, thus their stagnation was more serious.
3.2 The Root Course of Slowdowns in the Middle-Income Stage
Researches about stagnation show that after an economy has reached the middle or uppermiddle-income stage, factors and advances that once fueled its high speed growth in the primary growth stage may disappear. In other words, stagnation links closely to a time spot in the growth process. After this very time point has passed, the driving factors that lead to stagnation may arise. The essential reason why some economies may suffer slowdowns, stagnation or even recessions is that they are short of new growth impetus. Therefore, researchers in various countries made growth slowdowns the breakthrough point for their studies about growth impetus. Important work in this area includes those done by Ben- David and Papell (1998), Pritchett (2000), Hausmanm et al. (2008) and Eichengreen et al. ( 2000, 2013), among others.
Some academics tried to find the turning points in the growth series of sample economies through statistical instruments and the rule of thumb. Ben- David and Papel l ( 1998) searched through a sample of 74 developed and developing economies to look for significant breakthrough points in the growth series. They found that most of the discontinuous points defined by them linked to stagnation. For developed economies, most of the structural discontinuous points concentrated in the 1970s, while those of developing economies concentrated in the 1980s. Later, Alesina and Rodrik (1994) found that frequent conflicts and weak conflict management mechanisms made economies vulnerable to sharp downturns. Reddy and Miniou ( 2006) investigated the stagnation spells of real income and found that 60% of the economies had suffered stagnation, especially those that were in poverty, with frequent conflicts, or relied on exporting primary goods. Meanwhile, they argued that economies that had suffered stagnation at one time would frequently suffer subsequent stagnation.
Recent researches were also based on statistical instruments, though their subjects were about the characteristics of economic growth paths at various stages. Berg et al. (2012) screened out the growth spells in economic growth paths to investigate their attributes. Abiad et al. (2012) tried to determine whether the economies were in the expanding, recessing or recovering stage by imposing certain conditions that the lengths of growing periods or stages had to meet. Pritchett ( 2000) and
Felipe et al. (2012) quantified the conditions for crossing the “middle income trap”. They found that to avoid falling into the “lower- middle income trap”, lower-middle-income economies should achieve average annual growth rates of no less than 4.7%, while upper-middle-income economies should grow by at least 3.5% per year on average to avoid falling into the “uppermiddle income trap”. These studies showed inherently that after an economy has attained middle- income status, it will fall into the “middle income trap” if it fails to realize growth acceleration and thus is unable to transition into the high-income stage.
Other researches focused on the root of growth slowdowns. For example, Hausmann et al. ( 2008) discussed the course of stagnation. Eichengreen et al. (2012) examined the growth spells of rapidly growing economies since 1956 that satisfied the conditions of growth slowdown, and found that economies with a GDP per capita of 16,700 USD ( or between 15,000 and 16,000 USD) were vulnerable to growth slowdowns. That means their GDP was about 58% of that of developed economies, and manufacturers contributed 23% of employment. The authors then predicted when stagnation would happen in China based on the conclusion. If, as they assumed, China keeps growing at the average annual growth rate of GDP per capitaat 9.3% in the last decade, it will cross the threshold of growth slowdown in 2015. Meanwhile, if the U. S. grows on average at 1.9% per year in the same period, the GDP per capita of China would have been 58% as much as that of the U.S. by 2023. Eichengreen et al. (2013) further noted that a number of middleincome economies would experience two times of growth slowdown, with their growth rates of GDP per capita reduced from 5.6% to 2.1% and an average reduction of 3.5%. That is, before rapidly growing middle- income economies enter the “middle income trap”, it is very likely that they would decelerate in steps rather than have their national income per capita suddenly collapse at a certain time point. Meanwhile, they argued that when slowdowns happen, labor movement from agriculture to industries would not lead to additional productivity improvement, the total factor productivity would fall significantly, and the benefits from importing technologies from abroad would also decline.
Therefore, generally speaking, as Eichengreen et al. ( 2012) showed, stagnation was typically accompanied by previous high speed growths, undesirable demographic dynamics ( especially low or even negative growth of the labor force and high dependency ratios), high investment ratios, and undervalued exchange rates. By contrast, economies with a large percentage of the populationbeing midand highly- educated and whose exportswere mainly high- tech products seldom suffered stagnation. As a result, continuously improving the education level of the labor force and promoting technological progress is essential for avoiding the “middle income trap”.
To be sure, growth impetus is not the only factor driving growth. Fostering and improving growth involves a number of factors, such as industrial structures, factor structures, productivities, economic conditions, and institutional arrangements. The World Bank argued in its 2010 report titled “Robust Recovery, Increasing Risks” that “the global economic crisis brought into sharper focus the need for East Asia’s middle- income countries to accelerate structural reforms needed to transition through the crowded middle of industrial development and emerge as highincome economies. This will not be easy. For decades, many economies in Latin America and the Middle East have been struck in this middleincome trap, where countries are struggling to remain competitive as high- volume, lowcost producers in the face of rising wage costs, but are yet unable to move up the value chain and break into fast- growing markets for knowledge- and innovation-based products and services.” Ohno (2009) interpreted development regressions as follow: economies failed to gain enough potential impetus to upgrade their industries during their industrial developments; as a result, they were not able to transform their growth models smoothly. Therefore, to avoid the “middle income trap”, economies have to
adapt to the structural factors in time, transition their industries to produce goods and services with high added value, promote independent innovations and industrial upgrade, and look for new ways to improve their total factor productivity.
 Abiad, A., J. Bluedorn, J. Guajardo, and P. Topalova. 2012. World Economic Outlook. Washington: International Monetary Fund, October.
 Acemoglu, D. 1997. “Training and Innovation in an Imperfect Labor Market. Review of Economic Studies, 64 (3): 445-464.
 ——. 2005. “Institutions as a Fundamental Cause of LongRun Growth. In Handbook of Economic Growth edited by Philippe Aghion and Steven Durlauf. Amsterdam: Elsevier.  Aiyar, S. et al. 2013. “Growth Slowdowns and the MiddleIncome Trap. IMF Working Paper, WP/13/71.  Alesina, A., and D. Rodrik. 1994. “Distributional Politics and Economic Growth.” Quarterly Journal of Economics,
 Asian Development Bank. 2010. “The Future of Growth in Asia.” Asian Economic Outlook (October): 37-89.  Azariadis, Costas. 2004. “The Theory of Poverty Traps: What Have We Learned?” In Poverty Traps, edited by Bowles, S., Durlauf, S. & Hoff, K. New Jersey: Princeton University Press.
 Barro, R. J. 1997. Determinants of Economic Growth: A Cross-Country Empirical Study. Cambridge: MIT Press.  Ben- David, and David Papell. 1998.“Slowdowns and Meltdowns: Postwar Growth Evidence from 74 Countries.”
Review of Economics and Statistics, 80: 561-571.
 Berg, A., Ostry, J., and J. Zettelmeyer. 2012. “What Makes Growth Sustained? Journal of Development Economics, 98 (2): 149-166.
 Bowles, S., Steven Durlauf, and Karla Hoff. 2006. “Poverty Traps.” Princeton and New York: Princeton University Press and Russell Sage Foundation.
 Eeckhout, J., Jovanovic, B. 2007. “Occupational Choice and Development.” Journal of Economic Theory, 147 (2): 657683.
 Egawa, A. 2013. “Will Income Inequality Cause a Middle
Income Trap in Asia? Bruegel Working Paper, No.2013/06.  Eichengreen, Barry, Donghyun Park, and Kwanho Shin. 2012. “When Fast Economies Slow Down: International Evidence and Implicationa for China.” Asian Economic Papers, 11: 42-87.
 Eichengreen, Barry, Donghyun Park, and Kwanho Shin. 2013. “Growth Slowdowns Redux: New Evidence on the Middle-Income Trap.” NBER Working Paper, No. 18673.  Felipe, J., Abdon, A., and U. Kumar. 2012. “Tracking the Middle-Income Trap: What Is It, Who Is in It, and Why?”
Levy Economics Institute of Bard College Working Paper,
 Galor, Oded, and Omer Moav. 2002. “From Physical to Human Capital Accumulation: Inequality in the Process of Development.” Brown University Working Paper, No.99-27.  Hansen, G. D., and E. Prescott. 2002. “Malthus to Solow.”
American Economic Review, 92: 1205-1217.
 Hausmann, Ricardo, Francisco Rodriguez, and Rodrigo Wagner. 2008. “Growth Collapses.” In Money, Crises and Transition, edited by Carmen Reinhart Carlos Vegh and Andres Velasco. Cambridge: MIT Press.
 Hausmann, Ricardo, L. Pritchett, and Dani Rodrik. 2005. “Growth Accelerations.” Journal of Economic Growth, 10: 303-329.
 Im, F. G., and D. Rosenblatt. 2013. “Middle-Income Traps: A Conceptual and Empirical Survey.” Policy Research Paper, the World Bank, No.6594.
 Kuijs, Louis. 2010. “China through 2020-A Macroeconomic Scenario.” World Bank China Research Working Paper, No.9.
 Maddison, A. 2003. The World Economy: Historical
 Ohno, K. 2009. “Avoiding the Middle- Income Trap: Renovating Industrial Policy Formulation in Vietnam.”
ASEAN Economic Bulletin, 26 (1): 25-43.
Pritchett, Lant. 2000. “Understanding Patterns of Economic Growth: Searching for Hills among Plateaus,Mountains and Plains.” World Bank Economic Review, 14: 221-250.
 PWT. 2013. The Center for International Comparisons, University of Pennsylvania, https: //pwt.sas.upenn.edu/ php_ site/ pwt_index.php.
 Quah, D. 1993. “Galton’s Fallacy and Tests of the Convergence Hypothesis.” Scandinavian Journal of Economics, 95 (4): 427-443.
 Reddy, Sanjay, and Camelia Miniou. 2006. “Real Income Stagnation of Countries, 1960- 2001.” Unpublished Manuscript,Columbia University.
 Spence, M. 2011. “The Next Convergence: The Future of Economic Growth in a Multispeed World.” Farrar, Straus and Giroux: New York.
 Tho, T. V. 2013. “The Middle- Income Trap: Issues for Members of the Association of Southeast Asian Nations.”
ADBI Working Paper Series, No.421.
 World Bank . 2012 . “China 203 0 : Building a Modern,Harmonious, and Creative High-Income Society.”
World Bank (Washington DC).
 Wu, Y. 2013. “Productivity, Economic Growth and Middle Income Traps : Implications for China . ”
WorkingPaper, Business School of University of Western Australia.
 Yusuf, S., and K. Nabeshima. 2009. “Can Malaysia Escape the Middle Income Trap? A Strategy for Penang.” Policy Research Working Paper Series, No.421.