This chapter, in the tradition of comparative studies of trade and development (Little, Scott and Scitovsky, Bhagwati and Krueger, and Balassa, among others) is a welcome one for anyone interested in the trade–development–institution nexus. It highlights the difficulty of generalising from the inherent uniqueness brought out in case studies. It also points to the limitations of the lessons to be drawn from cross-country studies. As documented by Ravaillon (Reference Ravaillon2005) in the related trade-and-poverty debate, the case study approach illustrates the necessity to ‘look beyond averages’.
Bangladesh is the success story of a single sector, the RMG sector. It is also an old-fashioned story of success: endowments (an ample pool of low-skill, mostly female, labour) matter. This success story is now facing growing challenges. On the domestic front, there is a strengthening real exchange rate appreciation and a capture of the policy process by the RMG company owners. On the foreign front, there is increased competition from low-wage suppliers (Cambodia, Ethiopia, etc.) and stricter compliance requirements in destination markets following multiple incidents that have hampered the use of subcontracting to balance production capacity and order size. To these, one must add the challenge of shifting production towards higher value-added sectors (domestic value addition in RMG is about 40–50%, while it is estimated in the 80% range for leather, a candidate sector for diversification). This chapter also discusses the difficult obstacles to successful upgrading into the fashion industry.
This chapter describes the genesis of this success story, via the acquisition and mastery of technology in the garment industry by the stay of trainees in South Korea in 1978. Policies, including the long list of export-incentive schemes (see Box 4.1) are well documented. The importance of these incentives squares well with the view that an active government was at the origin of the Asian ‘miracle’, rather than the simple provision of neutral incentives as suggested by the neoclassical recipe.Footnote 51 This chapter offers a narrative of the role of institutions in which informality played, and continues to play, a key role in this success story. It is one in which informal institutions were preponderant in a young country (independence occurred in 1971).
This informality is plausible, perhaps even unavoidable, in a young country with low educational levels, where people were – to use Raihan’s terminology – operating in a ‘deals’ rather than in a ‘rules’ space. In the narrative, deals were initially open (access to all) and ordered (deals are respected), but they then became progressively closed. So, if there is a surprise, it is that the deals space remained open for so long while, as expected, the traditional determinants of comparative advantage (endowments and productivity), accompanied by generous incentives, were the proximate cause of the early surge of the RMG sector.
My remarks seek to place Bangladesh within the landscape of some of the current ‘received wisdom’ on the role of trade and institutions in successful industrialisation strategies. Where does the RMG success story fit in relation to this received wisdom? I focus on four aspects identified as important ingredients during the take-off stages of a country’s industrialisation:
initial conditions;
trade and macro policies;
institutions; and
supply chain trade.
Initial Conditions
The extensive cross-section of literature on growth, trade, and institutions has been unable to provide convincing evidence to confirm two often-cited anecdotal conjectures: that countries with ‘better’ institutions and countries that trade more grow faster; and that countries with better institutions also tend to grow faster. Initial conditions shape the paths of countries’ development. Here, Bangladesh’s high-growth experience has not been accompanied or preceded by the development of formal institutions. To illustrate the importance of initial conditions, contrast the paths of North and South America and those of Bangladesh and Korea.
Taking a long-run view, Engerman and Solokoff (Reference Engerman, Sokoloff and Haber1997) document the diverging paths of the development of institutions in North and South America since the arrival of Europeans. The spread of education and political participation was much slower in the South than in the North, where wealth inequality and rents were lesser, and their subsequent growth paths diverged. Bangladesh and Korea, when they started to industrialise, were both resource-poor and in a post-conflict environment with relatively low inequality. In both, industrialisation started from a low base (in Korea, in 1945 manufacturing production was less than a fifth of its level in 1940).Footnote 52 But Korea, which had started industrialisation in the colonial period (1910–1945) under the Japanese, had a much higher level of education than Bangladesh, where enrolment rates in schools were still in the lowest quintile of countries in 2007.Footnote 53 Ironically, the lower level of education in Bangladesh helped the ‘management’ of labour in the RMG sector, and hence was conducive to deeper specialisation in RMG.
Conditions for the development of formal institutions were far more favourable in Korea than in Bangladesh. Initial differences in educational levels matter for the growth of formal institutions. Together with the growth of formal institutions, this much higher level of human capital must have contributed to the spectacular transition out of RMG and footwear into heavy manufacturing in Korea, and likewise to the persistence of specialisation in the RMG sector in Bangladesh.
Trade and Macro Policies
The stellar performance of RMG in Bangladesh is an indication of an outward-looking development outcome, yet in the outward-looking vs. inward-looking industrialisation debate, Bangladesh only shares some of the policies of outward-looking successes.
On the micro-trade incentives side, Bangladesh still has relatively high tariff and non-tariff protection. Average applied tariffs went down from 58% in 1992 to 22% in 2000, but still stood at 13.9% in 2015. Non-trade barriers (NTBs) are still pervasive, though Bangladesh not being in the UNCTAD database precludes the estimation of ad-valorem equivalent of NTBs. As for the anti-export bias, in this chapter, it is captured by indicators of effective exchange rates for exports and imports, rather than for sales in the domestic and export markets at the individual product level. This better indicator of the whole array of incentives received by the RMG sector would probably show that producing and exporting apparel was more profitable than producing and exporting other manufactures, whereas in Korea during the early stages of exporting, export incentives were quite uniform across all exporting activities (Westphal, Reference Westphal1990, Table 1).
As in outward-looking East Asian countries, export incentives were widespread. The RMG sector ‘enjoyed quite extraordinary privileges’ (p. 26). Progressively, the sector captured export incentives that were de jure available for all. Typically, in outward-looking countries, one sector does not capture all export incentives. For example, in Korea’s export-led industrialisation, a virtual free-trade regime was established for ALL export activities. Neutral policies were available for all established export activities, whereas for infant industries, ‘non-neutral’ special incentives were provided (access to preferential credit, exemptions from taxes) (Westphal, Reference Westphal1990). However, Korean policymakers removed these special incentives when they realised that the so-called HCI (High Chemical Industry) drive had given excessive power to firms in the domestic market.
By contrast, In Bangladesh, the RMG sector progressively captured incentives, in principle available to all. This has been, and continues to be, a hurdle to diversification towards higher value-added activities as incumbents in the RMG sector have, in effect, raised barriers to entry into other potential export activities.Footnote 54 Raihan describes the process of policy capture as one of moving from an open, ordered deals environment to a closed-deal–ordered environment. This ‘raising the cost of rivals’, observed elsewhere, is not particular to a ‘deals’ environment. For example, Djankov et al. (Reference Djankov, La Porta, Lopez-de-Silanes and Shleifer2002) display a positive scatter between an index of corruption and the extent of paperwork. This suggests that politically well-represented groups take advantage of their position to protect their rents by raising barriers to entry to newcomers.
Since in their early phases of development, Asian countries’ exports of textiles and apparel generated rents (perhaps in the 20–40% range under the MFA), Bangladesh stood out, with higher quota utilisation rates under the MFA than other developing countries outside of East Asia (83% fill rate in the US market in 2001/2002 and no quotas in the European Community). Informal institutions operated smoothly in Bangladesh – perhaps aided by the relative homogeneity of the population – to capture these rents. In Korea, productivity-led rises in wages, and the acquisition and mastery of imported technology in manufacturing, led to a rapid shift out of RMG. By contrast, as an LDC, exporting apparel has continued to be profitable for Bangladesh. In the EU and US markets, LDCs still have a 10–15% preferential margin over competitors exporting under Most Favoured Nation (MFN) tariffs (see Brunelin et al. (Reference Brunelin, de Melo and Portugal-Perez2019), Tables A2 and A3). Also, the double transformation rule for textiles and apparel was removed for LDCs in their exports to the EU in 2011. All this contributed to maintaining high profits from RMG exports for Bangladesh, acting as a brake on diversification.
On the macro front, Bangladesh had a competitive real exchange rate, at least until recently, a sign of sound macro policies, a key ingredient observed across successful outward-looking strategies.Footnote 55 Bangladesh also had an ‘export surge’ over the period 1980–2006. The surge took place relatively early on, in 1986, supporting the hypothesis that a relatively long spell of sustained under-valuation of the exchange rate is necessary for exports to take off, to compensate for the information disadvantage that tradeables face relative to non-tradeables.Footnote 56
In conclusion, looking at the incentive side, exchange rate policies were similar to those of the other successful Asian exporters. Export incentives were also widespread. Policy capture by lobbies was typical of those found in weak institutional environments. However, the distinguishing characteristic here is that rents were obtained from export sales rather than from sales in protected domestic markets. This put a lid on the inefficiency associated with rent-seeking activities on domestic markets. The persistence of these rents over a long period also served as a brake on diversification.
An Institution Puzzle?
Raihan argues that Bangladesh’s performance in the RMG sector stands out against its poor ranking on various institutional indicators (Doing Business, Transparency International, and WGI). Doubts are widespread about what is captured by these multi-dimensional institutional indicators, which typically mix policy stances and outcomes.Footnote 57 Early on, in their comments on Kaufmann and Kraay, Devarajan and Johnson (Reference Devarajan and Johnson2008) use the example of Bangladesh to discuss the downside of indicators like the WGI that allow for intercountry comparisons at the cost of not capturing the multifaceted way in which governance affects development in a particular country – the focus of the country case studies in the EDI project. Devarajan and Johnson note that Bangladesh has a ‘vibrant and active civil society that not only delivers services, but provides some accountability to the government’ (p. 32). They also note that the Bangladeshi ‘…people have worked around the country’s governance problem to spur development’ (p. 32). In their remarks, Devarajan and Johnson (Reference Devarajan and Johnson2008) note that ‘…some countries have managed to sustain rapid growth in spite of weak initial governance because of a strong focus on exports, particularly of manufacturing goods’ (p. 35). They conclude that greater efforts should be put into measuring governance more at the local level, both for sectors and for cities.
More recently, in their sample of 10 countries, including Bangladesh, Pritchett et al. show that, at the country level, indicators of institutional quality are only correlated with GDP levels. When it comes to GDP growth, they detect only a weak bivariate correlation with GDP growth at the country level. And they fail completely to detect any correlation between improvements in indicator values of institutions and GDP growth (see their Figure 1.8). For Bangladesh, digging in on the measurement of indicators of institutional quality, one only observes weak improvements in rankings of institutional indicators during the growth acceleration episode of 1995–2010, when per capital GDP grew at an average rate of 3.5% per annum. Bangladesh’s long sustained period of growth in spite of low and slowly evolving rankings on institutional indicator values evidences the limitations of these indicators as predictors of performance.Footnote 58
However, indicators of the quality of formal institutions could contribute to the explanation of the patterns of comparative advantage in manufacturing. In Bangladesh, concentration of the export basket has been on the rise while the values of indices of complexity have been low and falling (Hassan and Raihan, Reference Hassan, Raihan, Pritchett, Sen and Werker2018, Figure 4.4). Nunn and Trefler (Reference Nunn, Trefler, Gopinath, Helpman and Rogoff2014) report cross-country evidence supporting the overall importance of contracting institutions. They show that their indicators of the contract intensity of sectoral production across markets (product market, labour market, and financial markets), when entered interactively with indicators of governance, as captured by the rule of law of Kaufmann and Kraay (Reference Kaufmann and Kraay2008), are all significant determinants of the patterns of revealed comparative advantage. Although these results call for careful interpretation, the governance indicators are as important quantitatively as the traditional indicators of comparative advantage (Heckscher–Ohlin and technology indicators).Footnote 59
Supply Chain Trade Under the ‘RMG-Centric’ Model
In the introduction to this chapter, Raihan points out the importance for success of a diversified export basket, following the adage that ‘you get rich by producing the goods the rich consume’. The argument is that participation in supply chain trade (or global and regional value chains (GVCs and RVCs)) provides new opportunities for developing countries to participate in global trade and to diversify their export baskets (Bangladesh has the most concentrated exports in textiles and apparel among comparators). Without participation in supply chain trade, a country has to be able to produce a complete product before entering a new line of business. The fragmentation of production allows countries to enter a product chain without having to carry out all the stages of production. Supply chain trade can then be a lift for a country to shift rapidly from labour-intensive to capital-intensive, skill-intensive, and information-intensive activities. The World Bank World Development Report of 2020 (World Bank, 2020a) documents the higher growth of countries that have transitioned out of commodities using imported inputs through GVC participation.Footnote 60
When a country’s participation in supply chains is high, foreign imports have a high share in a country’s gross exports (backward participation or ‘upstreamness’) and a high share of its gross exports enter into other countries’ exports (forward participation or ‘downstreamness’). So, to succeed, Bangladesh should strive for a high participation rate in supply chain trade. This has not been the case over the past 25 years, at least relative to comparators. For its level of development, Bangladesh’s GVC participation is low, and Figure 4.5 also shows Bangladesh’s low GVC participation relative to other RMG exporters. In contrast with Ethiopia and Vietnam, Bangladesh (like Cambodia) has not caught up with the sample average of GVC participation during the 25-year period. By contrast, Ethiopia and Vietnam have caught up with the trend, and now have above-average GVC participation for their income level. Both display a positive correlation between supply chain participation and GDP growth, suggesting that participation in GVC trade could have been growth-enhancing for these countries.Footnote 61
It is also mentioned by Melo and Twun (Reference Melo and Twun2020) that compared with Ethiopia and Vietnam, Bangladesh displays vertical industrialisation: a low share of imported intermediates in gross exports and a low share of gross exports entering further processing in destination countries. Also, a low share of Bangladesh exports goes through further processing in the destination countries of Bangladesh exports: a reflection of the RMG-centric model that has delivered final goods (t-shirts, trousers). The domestic vertical specialisation of RMG in Bangladesh has been at the expense of growth in backward integration experienced by competitors, especially Ethiopia.
Connectivity is important for both the physical supply chain of goods, but also effective communication among participants in GVCs, and certainly in any quest to move up the fashion industry. This low GVC participation in the aggregate also holds for the RMG sector since, in textiles and apparel, Bangladesh’s GVC participation is the lowest in the group and has not evolved over the period. GVC participation is also low for wood and paper, a candidate sector for easy diversification (Melo and Twun, Reference Melo and Twun2020).
Compared with these other successful exporters of RMG, Bangladesh has a low participation in supply chain trade. In view of the success of RMG, this low participation seems surprising as most of the literature evidences the productivity gains from a reduction in the tariffs of imported varieties of intermediate inputs (Amiti and Konigs, 2007 for Indonesia and Goldberg et al., Reference Goldberg, Khandewal, Pavcnik and Topalova2010 for India), and most recently from participation in supply chain trade. The incentive system has guaranteed duty-free access to imported inputs for the RMG sector that has also largely bypassed participation in the slicing up of the value chain. Kee (Reference Kee2015) shows that the performance of domestic firms in the RMG sector benefitted from non-pecuniary externalities through contacts with FDI firms. Drawing on a stratified random sample of 10% of domestic firms and 100% of FDI firms in the apparel sector, she shows that firms that shared local suppliers with foreign firms benefitted from significant spillovers. She estimates that from 1999 to 2003, the spillover effects helped explain a quarter of the expanded scope and a third of the productivity gains of domestic firms in the apparel sector. Automation in the RMG sector is increasing. Some factories employ several thousand, mostly unskilled, workers. In the low-end of the garment sector, production runs are long, helping reap economies of scale and economies of scope.
A homogenous population, a stable real exchange rate, incentives that allowed exporters to earn rents on sales abroad rather than on a sheltered domestic market, all contributed to the RMG success. Can Bangladesh replicate this success in other sectors? With a growing number of activities built around networks across several countries, acquiring the necessary mastery of technology will require the competences associated with a functioning education system. As in Korea’s industrialisation, a helpful hand from the government will also be required.
Open access