Growth Slowdowns and Stagnation in the Middle: The Middle-Income Trap Revisited.

• Author: Ekanayake, Raveen
• Position: RESEARCH NOTES

1. Introduction

   Over the last decade or so, the notion of the middle-income trap (MIT) has reinvigorated the development policy debate, particularly with respect to the sequencing and timing of policy reforms needed to successfully navigate the middle-income transition. Despite a large body of work, there is little consensus on the existence of a so-called middle-income "trap". Rather and perhaps more aptly, a few studies (notably Eichengreen, Park, and Shin 2012, 2014; Aiyar et al. 2013) have postulated the phenomenon as the incidence of growth slowdowns at the middle-income stage of development. This study goes a step further and argues that the middle-income trap paradigm could be better reframed as the incidence of growth slowdowns and "stagnation" in the middle-income stage of development. In this context, stagnation means that some middle-income economies have not been making any significant headway in catching up with their peers, owing to their inability to reinvigorate and sustain robust growth dynamism at the middle-income stage of development.

   The main objective of this paper is to investigate the determinants of growth slowdowns and middle-income stagnation with a view of gaining a better understanding of the drivers of economic growth at the middle-income stage of development. At the outset, the paper draws on the existing literature to reframe the middle-income trap narrative as the incidence of growth slowdowns and stagnation in the middle. Thereafter, through a cross-country panel data analysis, which employs a theoretically informed empirical model, the determinants of growth slowdown episode and middle-income stagnation cases are evaluated.

   The findings are broadly consistent with the theoretical interpretation of growth dynamics at the middle-income stage of development. While capital accumulation does contribute to reducing the likelihood of middle-income stagnation, its effect is marginal. Moreover, the analysis suggests that quality of human capital stock and international cost competitiveness, measured by the real effective exchange rate, contributes far more significantly to reducing the likelihood of both middle-income slowdown episodes and stagnation. The upshot is that growth dynamism at the middle-income stage of development cannot be sustained through factor accumulation alone. Countries need to shift their growth strategies from labour-intensive to skill- and knowledge-intensive activities to maintain economic dynamism at the middle-income phase of development. The policy emphasis should therefore be on developing human capital capabilities while maintaining a conducive environment for international trade and investment to facilitate industrial upgrading towards high value-added economic activities.

   The rest of the paper is structured as follows. The next section reframes the middle-income trap paradigm based on the existing body of work on economic growth and structural change. The third section discusses the methodology and data employed to assess the determinants of growth slowdowns and stagnation in the middle. The subsequent section reports and discusses the empirical results. And the final section concludes.

2. Refraining the Middle-Income Trap Narrative: Growth Slowdowns and Stagnation in the Middle

   As originally conceived by Gill and Kharas (2007), the "middle-income trap (MIT)" refers to a situation where fast-growing developing countries find it increasingly difficult to sustain their growth dynamism upon reaching the middle-income stage of development. Some countries face difficulty in competing in labour-intensive activities with low-income countries and in skill-, capital- and technology-intensive activities with high-income countries; hence the notion of being "trapped". Following this line of reasoning, the authors argue that several fast-growing Southeast Asian economies such as Malaysia and Thailand are already caught-up in the middle-income trap.

   Although Gill and Kharas (2007) have spawned a plethora of empirical studies probing the existence of a "middle-income trap",' however, there is actually no real consensus on the existence of such a trap. Some authors have argued that "middle-income growth slowdown" is the more apt label for the phenomenon (Riedel 2016).

   Middle-income growth slowdowns are, by no means, novel. As Pritchett (2000) argues, the process of transitioning from low- to middle- and high-income status is not straightforward, and could be better characterized as a process of searching for "hills" amongst "plateaus", "mountains" and "plains", suggesting that countries typically feature abrupt and sustained changes in growth regimes rather than a consistent convergence process. Building on these insights, in a major study covering a large number of countries during the post-World War period, Pritchett and Summers (2014) uncover a pattern of "regression to the mean" in growth trajectories, which implies that growth slowdown at middle-income levels are a part and parcel of the natural long-run growth convergence process. Their findings support the view that growth slowdown in the middle-income stage of development could be better thought of as growth rates reverting to the average level for all countries.

   Theoretically, the factors behind middle-income growth slowdowns can be explained using the Lewis-Solow-Kuznets framework. At low-income levels, growth is sustained through the process of structural transformation, whereby surplus labour in traditional (primary) sectors are reallocated to modern ones. However, as Lewis (1954) highlighted, there are natural limits to growth driven by structural transformation. As surplus labour in the traditional sectors is exhausted, owing to the reallocation effect (i.e., after reaching the "Lewis turning point"), wages across the board begin to rise, deeming labour intensive activities uncompetitive. As such, growth beyond this point must be sustained through the process of technological change supplemented by capital accumulation (Solow 1956, 1957; Kuznets 1971). In the case of many present-day developing countries, reaching the middle-income stage of development often coincides with the economy reaching the Lewis turning point (Gill and Kharas 2015).

   In light of the above, it could be inferred that growth slowdowns at the middle-income stage of development are triggered by natural forces, namely, the passing over of the easy stage of labour-intensive growth. Countries, however, have the ability to circumvent these natural constraints through appropriate policy choices. The notion of middle-income stagnation, where some developing countries take a much longer time compared to their peers to make the graduation to high-income status, is often reflective of the difficulty faced by developing countries to make the timely and successful transition from a factor accumulation driven growth model to one driven by productivity and endogenous technological changes.

   There are two competing explanations for the inability of countries to make this transition--policy failure and the lack of opportunity. Proponents of the policy failure argument claim that some countries have failed to align growth strategies with prevailing structural characteristics of their economies. These nations are accused of attempting to sustain labour-intensive growth, despite eroding labour cost advantages. On the other hand, some countries appear to be prematurely pushing for advanced skill-, capital- and knowledge-intensive industries without fully developing the necessary complementary human capital, and physical and institutional infrastructure, resulting in suboptimal growth outcomes (Gill and Kharas 2015). Scholars who advocate the lack of opportunity argument claim that the global economic landscape and the circumstances under which some economies have integrated into global production networks provide little scope for the countries to move into higher value-added activities (Doner and Schneider 2016).

   In order to better engage in this debate, it is important to gain an understanding of the drivers of economic growth at the middle-income stage of development. The next section discusses in detail the methodology and data employed to evaluate the determinants of growth slowdowns and middle-income stagnation.

3. Methodology and Data

   Eichengreen, Park, and Shin (2012, 2014) are amongst the first to empirically examine the existence of middle-income growth slowdowns and analyse their determinants. Their analysis of economic performance of 189 countries between 1956 and 2010 identifies eighty-four slowdown episodes concerning forty-one countries, which generally tend to occur at around two income ranges--one at around the US$10,000-US$11,000 range and another at around US$15,000-US$16,000.2 The incidence of multiple slowdown episodes in the case of some countries suggests that middle-income countries tend to experience slowdowns in several steps. The authors go on to examine a number of determinants, including human capital, investment, consumption, trade openness, macroeconomic, demographics, and political variables, and find that high levels of tertiary education and a greater proportion of high-technology exports tend to significantly reduce the likelihood of slowdown episodes (Eichengreen, Park, and Shin 2014).

   Using an alternative approach, Aiyar et al. (2018) identify the incidence of growth slowdown episodes at all income levels. They show that middle-income countries are disproportionately more likely to experience growth slowdowns than countries at other income levels. Their analysis sheds light on the incidence of some 120 middle-income slowdown episodes between 1960 and 2005. Determinants that they examine include quality of institutions, demography, trade openness, and macroeconomic variables.

   While empirical approaches to identify growth slowdowns of these studies are methodologically sound, econometric models employed to...