After the Crisis: Industrial Policy and the Developmental State in Low-Income Countries
The current global economic crisis has been disastrous for many millions of people. But it has also had the desirable effect of prompting a little more skepticism towards the economic beliefs that have constituted the mainstream view about public economic strategy for the past three decades, both in the major western states and in international lending organizations like the World Bank and the IMF. They have at their core the proposition that ‘government failure is generally worse than market failure’, which supports the default policy setting of ‘more free market’ in most countries most of the time. The new crisis-induced skepticism is good news because the previous confidence rested more on what J. S. Mill called ‘the deep slumber of a settled opinion’ than on a solid empirical base. The present article begins by summarizing some powerful pieces of evidence that challenge core mainstream propositions in the context of developing countries, which have received less attention than they deserve. Having shown why the mainstream prescription for the role of government in development is questionable, the article describes some key points about the nature of industrial policy in East Asia and about the general rationale for a certain kind of industrial policy even where state capacity is relatively weak. The rationale is all the stronger in the world economy after the crisis, when a major surge of innovation around energy, water, nanotechnology and genetics is likely, rendering many existing specializations unviable. The article then presents an argument about the institutional arrangements of a ‘developmental state’ through which national strategies can be formed and implemented. It ends by describing small signs of new flexibility in World Bank and IMF thinking.
- •Liberalizing markets, attracting FDI and promoting good governance are not necessary conditions of long-term economic growth and development in low-income countries.
- •In the wake of the global financial crisis and the impending surge of new technologies, the role of industrial policy – promoting some sectors or products ahead of others – should be expanded.
- •Import replacement as well as export orientation are crucial components of a successful industrial policy.
- •Four organizational features are important in a developmental state: (1) an even balance between the state and business groups; (2) a public service mindset among state officials; (3) delivery of patronage resources separately from the economic bureaucracy; and (4) an industrial extension service, with tight limits on its officials’ discretionary control of resources.
- •Developing country governments and firms should be prepared to push back against the shrinking latitude for industrial policy instruments allowed in international trade and investment agreements.
A New York Times journalist covering the World Economic Forum meeting in 2002 wrote that among business executives and government leaders attending the forum, the prevailing view was that ‘A nation that opens its economy and keeps government’s role to a minimum invariably experiences more rapid economic growth and rising incomes’ (New York Times, 9 February 2002, p. 1).
In 2006 the American economist Gregory Mankiw assured readers of the Wall Street Journal that ‘Adam Smith was right when he said that “Little else is required to carry a state to the highest degree of opulence from the lowest barbarism but peace, easy taxes and a tolerable administration of justice” ’ (Mankiw, 2006). Mankiw is no crank right winger. He is Professor of Economics at Harvard, former chair of the US Council of Economic Advisors and author of a popular textbook.
Martin Wolf, perhaps the most influential economic commentator in the English-speaking world through his columns in the Financial Times, declared: ‘Bad government is the single most important cause of failure in developing countries’ (Wolf, 2005).1
Finally, the finance ministers of the Group of Eight industrial countries decried what they called the ‘new investment protectionism’ in their draft communiqué for the G8 meeting in June 2007. They prescribed that ‘restrictions to market access for foreign investment should only apply to exceptional cases where national security is at stake’ (Beattie, 2007).
These four quotations express in stripped-down form the mainstream view among western (particularly American and British) economists about the open-ended nature of the world economy and about the need for limited government in all economies, including low-income countries seeking to rise up the world income scale. In this view, all countries could conceivably be rich at the same time, just as all runners could conceivably cross the finishing line of a marathon at the same time; and the relative position of a country in the development race is determined by factors substantially under the control of the government and private agents within the country, just as each runner’s position in the marathon race is in his or her own hands, or legs. The key is to get the government to do well the narrow range of tasks that Adam Smith prescribed, and not much more.
The strength of attachment to these beliefs is attested to by the fact that the current global economic crisis has done little to shake them in the academy (though there are signs of a new flexibility in International Monetary Fund (IMF) thinking; see below) (Wade, 2010). A survey of American economists in 2009 concluded:
Free market theory, mathematical models and hostility to government regulation still reign in most economics departments at colleges and universities … The belief that people make rational decisions and the market automatically responds to them still prevails … Graduate students who stray too far from the dominant theory and methods seriously reduce their chances of getting an academic job (Cohen, 2009).
Gregory Mankiw declared in the pages of the New York Times: ‘Despite the enormity of recent events, the principles of economics are largely unchanged … Students still need to learn about … the efficiency properties of market outcomes’ (Mankiw, 2009). One wonders which markets he had in mind when he talked of their ‘efficiency properties’. Subprime mortgages? Over-the-counter derivatives?
What we are seeing in the academy since the American and then global crisis erupted in 2007–08 is continuing efforts to tweak established economic theory in an effort to save it. We remain in the equivalent of the Earth-centric planetary model named after Ptolemy. In this article I challenge some of the core ideas of the mainstream view about development, and offer some more positive prescriptions which move in what might modestly be described as a Copernican direction.
1. What is wrong with the mainstream view of development strategy?
To illustrate the shaky foundations of the mainstream view I consider four core propositions.
- •The world economy is an open system, in the sense of giving plenty of room for upward income mobility over periods of two or more decades.
- •Opening the economy and liberalizing markets are important conditions for accelerating the rate of economic growth, in virtually all developing countries.
- •Foreign direct investment (FDI) is an important engine of growth for developing countries and should be granted unrestricted access, except where national security is involved.
- •Clean, effective government is an important cause of economic growth.
For each of them I briefly consider some powerful contrary evidence, which has tended to be either ignored or treated with the annoyance one might direct towards a fly.
Is the world economy an open-ended system, with plenty of opportunity for states to move up the hierarchy of national income?
Table 1 shows two state mobility matrices, for 1960–78 and 1978–2000. They divide countries into four income categories (in $PPP (purchasing power parity) per person). The ‘rich’ category includes the western European, North American and Japanese core. The ‘contenders’ are countries whose incomes are two-thirds or more of that of the poorest rich country (Portugal or Greece). The ‘Third World’ comprises countries with incomes between one-third and two-thirds of that of the comparator. The rest are ‘Fourth World’.
State mobility matrices for the periods 1960–78 and 1978–2000
The results are clear. The majority of states, unsurprisingly, remained in the same income category over two decades. But the stability was concentrated at the ends: at the top, few states fell into a lower income category; at the bottom, few states rose into a higher category. The middle two categories showed plenty of movement, but most of it downwards. Eighty per cent of the states in the ‘contender’ category in 1978 fell into a lower income category by 2000, and two-thirds of the ‘Third World’ states.
Much the same exercise has been done with ten income categories and incomes converted at market exchange rates, and the results similarly show little downward mobility at the top, little upward mobility at the bottom and strong downward mobility in between (Ikeda, 2004).
The ‘middle-technology’ trap: the case of Malaysia
The proposition about the world economy being an ‘open-ended system’ can be further assessed by asking about the chances of today’s middle-income countries moving up global value chains and vacating lower value-added economic activities. The mainstream view tends to presume that the middle-income countries’ continued advance is assured, the assurance being expressed, for example, in alternatively rosy or worried talk about the BRICs (Brazil, Russia, India, China) or the BIMCs (Brazil, India, Mexico, China) ascending to developed country status. This presumption carries the important implication that the low-income countries will face improved prospects for faster growth as the middle-income countries catch up to the high-income ones.
The state income mobility matrix presented earlier questions this implication. It shows that of the ‘contender’ countries in 1978 – those with average income in the range of more than two-thirds of that of the bottom country of the ‘old’ Organization for Economic Cooperation and Development (OECD) – only 13 per cent ascended to ‘rich’ country levels of income by 2000, while 80 per cent descended to a lower income category. By 2000 there were fewer countries in the ‘contender’ income category than there were in 1978. This is not a misprint.
The statistics suggest the difficulties of upward mobility in the world income hierarchy, and the experience of specific middle-income countries confirms the difficulties. Many middle-income countries are today caught in what could be called a ‘middle-technology trap’, their firms stuck in the relatively low value-added segments of global production chains, unable to break into innovation-intensive activities or into the market for branded products, where the high profits are to be made.
Take Malaysia, the most advanced of the second cohort of ‘Tiger’ economies, which also comprises Thailand, Indonesia and the Philippines. Malaysia’s transformation can be seen in its export mix. In 1970 more than 95 per cent of its exports were resource based. By 1990, only about half of its exports were natural resource-based products. By 2007 the share of such exports was less than 30 per cent. Thailand and the Philippines have undergone a similar transformation in their exports from natural resource based to industrial, but not Indonesia (Yusuf and Nabeshima, 2009, table 1.3). Malaysia’s strategy relied heavily on foreign direct investment, as western and Japanese multinational corporations established subsidiaries or joint ventures.
Looked at superficially, Malaysia has been very successful. Its GDP grew at 7.7 per cent a year over the 1970s, 5.9 per cent in the 1980s, 7.3 per cent in the 1990s and 5.2 per cent in 2000–06. But during the past decade, according to a new World Bank book, Malaysia’s ‘technological capabilities are relatively static (and may even be declining) and … industrial competitiveness is marking time’ (Yusuf and Nabeshima, 2009, p. 26). More specifically,
- •Malaysia’s export value added has been stagnating, and its export diversification by product and destination has slowed;
- •the FDI factories remain largely ‘enclaves’, and have not increased domestic value added through forwards or backwards linkages – nor has FDI helped to boost national research capacity;
- •domestic innovation capacity has increased little if at all over the past decade despite higher spending on R&D.
These findings about Malaysia have important implications for low-income countries aiming to achieve emerging market status.
- •They underline the idea that the world economy is not an open-ended system in which all can raise their long-term growth rate and reduce the gap with the rich countries (if they get their prices and governance right).
- •They suggest that middle-income countries may get caught in the middle-technology trap, specialized in relatively low value-added industrial activities that lower-income countries might have moved up to, amplifying the difficulties of the latter in moving up.
- •They highlight the difficulties of relying on spillovers from FDI firms to power national development. Malaysia would seem to have been well placed to get these spillovers, but the spillovers are not much in evidence. The reason may be that the Malaysian state has not done many of the things needed in order to reap them; or it may be that FDI firms find it in their private interests to curtail spillovers and have learned how to do so despite the government’s best endeavors.
Are liberalization and economic opening important conditions for accelerating economic growth?
The mainstream view says, to repeat, that ‘A nation that opens its economy and keeps government’s role to a minimum invariably experiences more rapid economic growth and rising incomes’ (New York Times, 9 February 2002, p. 1, emphasis added). To assess this proposition would take the whole article and more. Here is just one example of a recent empirical study that runs strongly against it. Dani Rodrik, Ricardo Hausmann and Lant Pritchett identified more than 80 episodes since 1950 in which a country’s growth rate increased by at least two percentage points and sustained the increase for at least seven years. Then they looked for causes. They found that almost all these spells of accelerated (national) growth occurred spontaneously, without any preceding ‘reforms’, or only marginal ones, whether to open the trade regime or to deregulate or privatize or any of the other prescriptions of the Washington Consensus (Hausmann et al., 2005). They did not discover a ‘magic bullet’ for growth accelerations. But their negative finding is still important, in justifying countries experimenting with economic policy and not accepting one-size-fits-all conditions from international organizations.
Is foreign direct investment such an important engine of growth that it should be granted more or less unrestricted access?
The Washington Consensus says that foreign direct investment is good for development and more FDI is almost always better. Theodore Moran, in his Harnessing Foreign Direct Investment for Development takes a more nuanced view (Moran, 2006). He distinguishes between FDI that is oriented to selling on a protected domestic market and not integrated into the parent firm’s global production chain, on the one hand, and FDI that is oriented to exporting and integrated into the parent firm’s global production chain, on the other. The former generates net costs for the host economy, the latter net benefits, says Moran.
Moran’s argument improves on ‘the more FDI the better’, but it is misleading in several ways. First, in suggesting that FDI can play a major role in the development of developing countries in general, it downplays the fact that FDI to developing countries is highly concentrated in a small number of developing countries. Roughly 80 per cent goes to only ten countries.2 The vast majority receive very little. In 1980 the concentration was more or less the same as today – indicating that the evolutionary hope that FDI would spread out across more and more developing countries has not been realized.
Second, Moran’s argument ignores evidence on the harmful effects of FDI in Latin America, where foreign firms have dominated the most dynamic manufacturing sectors since their inception; indeed, their share of sales relative to national firms increased with trade liberalization over the 1990s. Without a strong, East Asian-type push from government to build up domestic suppliers, transnationals’ global strategies have led them to source many of their specialized inputs from abroad, and even to disintegrate existing vertical supplier links within the national economy. The effect has been to shrink intermediate and supplier industries. Yet much evidence suggests that the growth of intermediate inputs and producer services within the national economy is an essential part of industrialization, for these are typically rich sites of innovation (Ciccone and Matsuyama, 1996).
In the case of Argentina, for example, Bernardo Kosacoff says:
the data show that the manufacturing sector has itself utilized trade openness and economic deregulation to increase its imports not only of parts and components but of finished production too. This is indicative of a trend towards the vertical de-integration of activities that affects both manufacturing activities … and commercialization activities (Kosacoff, 2000, p. 188).
The upshot is that reliance on transnational corporations (TNCs) can produce an import-intensive or deficit-prone industrialization. We see this deficit-prone industrialization throughout Latin America, where exports of natural resource processing industries, foodstuffs and primary commodities have grown fast, while imports of capital goods and intermediate goods have grown even faster. Also, this pattern of TNC-led growth has caused a rapid increase in economic concentration, as small and medium enterprises which earlier had supplied big national firms were marginalized by imports. Mexico’s income elasticity of import demand has doubled over the past 15–20 years. All these negative effects are seen even where the growth of output and exports from TNC activities is high, because the multiplier effects and technological spillovers from them to the rest of the economy turn out to be typically low (as noted for Malaysia).
In short, FDI can bring large benefits to a developing country host economy, but it can also – individually and in aggregate – bring large costs. The government has to build up capacity to be able to manage FDI strategically so as to raise the benefits and reduce the costs. But in the majority of developing countries FDI will continue to remain marginal, which puts more onus on building up the capacity of national firms to respond to international competition.
Is clean, effective government a cause of growth?
Finally, under the heading of ‘What is wrong with the mainstream view of development strategy?’, consider another of the core prescriptions of the international development organizations for accelerating economic growth: improve the performance of the public administration. We can agree that a clean and effective public administration is desirable in itself; but the international organizations emphasize improved governance on grounds that improvements contribute to higher subsequent economic growth. Paulo Mauro puts it the other way around in a recent IMF Staff Paper: ‘a consensus seems to have emerged that corruption and other aspects of poor governance and weak institutions have substantial, adverse effects on economic growth’ (Mauro, 2004).
On the basis of this consensus the World Bank and IMF have built governance reforms into their lending conditionalities. As Mitchell Seligson says, ‘So widespread is confidence in these findings that international lending agencies have embarked upon major efforts to reduce corruption, conditioning many of their loans on formal, widespread efforts to clean it up’ (Seligson, 2002, p. 410).
There are indeed a large number of academic studies which claim that improvements in public administration cause higher subsequent growth. But these studies suffer from two basic weaknesses (Kurtz and Schrank, 2007).3 First, they are based on dubious measures of administrative performance. Second, even using these dubious measures, the evidence is that the causality is more from growth to improvements in governance than from improvements in governance to growth.
On the dubious measures of governance, the problem is that many measures are based on perceptions of foreign businesspeople. The problem is obvious: to base the assessment of state capacity on the opinions of foreign businesspeople implies that the interests of foreign investors, buyers and sellers are the same as the national interest. Yet foreign businesspeople may judge any state regulation of their activities to be ‘growth inhibiting’, and they may score state capacity as ‘low’ when the state is effective in imposing taxes and regulations that they do not like, and state capacity as ‘high’ when they can easily evade taxes and regulations.
One of the main data sets about governance is provided by the World Bank. This data set uses ‘absence of red tape’ as one of its main criteria of ‘the quality of bureaucracy’, and it measures absence of red tape by: (1) how quickly decisions are made in regulatory agencies; and (2) how easily ‘foreign investors can go about their business’ (Kaufman et al., 2005a, p. 93, 2005b).
However, if one applies this reasoning to South Korea and Taiwan of the 1950s to the 1980s one would have to score them quite low on administrative capacity. Foreign businesspeople complained loudly about delays. But as I document in my book, Governing the Market (Wade, 2004), delays were often ‘tactical’ in pursuit of industrial policy objectives. For example, an FDI subsidiary might experience delays in its previously automatically approved permission to import certain components when the industrial policy authorities considered that a number of domestic firms could – if given long-term supply contracts by the FDI subsidiary – upgrade to achieve the same price and quality as the imports. Without the ‘interference’ of the state the FDI subsidiary would have been less likely to switch suppliers from foreign to domestic.
The larger point is that the World Bank’s data set cannot discriminate between situations in which ‘delay’ or ‘red tape’ is an expression of bad governance and in which it is an expression of an effective state seeking to channel the behavior of foreign investors in line with the national interest rather than short-term profit.
So there are several reasons why the measures of administrative performance are unreliable for drawing any conclusion about its connection with economic growth.
The second basic problem relates to the direction of causality. We can get an intuition of the weakness of the mainstream view from the history of the USA. From 1860 (the end of the Civil War) to 1910 the US judicial and police system was riddled with corruption; yet the US economy underwent the fastest expansion in the world during this time. The improvement in US governance came after, not before, in the form of the professionalization of the civil service, federal oversight of banking and interstate commerce and the like.
It turns out that people’s opinions about administrative performance are correlated with recent past economic performance. A government that presides over a period of good economic performance tends to be judged competent regardless of its actual practice; and one that presides over a period of bad economic performance tends to be judged incompetent regardless of its actual practice.
Indeed, Marcus Kurtz and Andrew Schrank find that measures of governmental effectiveness are not at all correlated with subsequent rates of growth. Nor do they find the converse: bad government – by the standard measures – is not correlated with higher subsequent economic growth. The conclusion is simply that the empirical foundation of the argument that governance reforms should be a high priority in efforts to accelerate economic growth is shaky – notwithstanding vigorous mainstream claims to the contrary.
2. The case for industrial policy
The combination of the new ambiguity prompted by the global crisis and evidence of the kind just described opens the way to reconsidering the case for industrial policy, meaning targeted efforts to promote some sectors or products ahead of others.
Part of that case rests on the successful trajectories of older developmental states. The East Asian capitalist economies in the 1950s to the 1980s – the phase of rapid industrialization – experienced intensive government ‘intervention’ in markets, including high rates of effective protection (though not in the city states of Hong Kong or Singapore) and active technology-upgrading policies (though not in Hong Kong). Further, these interventions were not focused mainly on ‘making the market work better’ but on the production capacity of firms or industries, aiming to accelerate diversification and upgrading (Lee et al., 2007). They relied heavily on price-distorting incentive policies, including managed trade and managed FDI, and an array of sectorally specific incentives for exports. Taiwan also had a large public enterprise sector spanning the commanding heights of the economy.
The fact that East Asian capitalist governments practiced intensive ‘intervention’ does not necessarily mean that the intervention was important to subsequent growth, of course. Disentangling the impact of industrial policies from other things – including heavy investment in education, as well as cold war-facilitated aid and entry to the American market for manufactured exports – is difficult. But we now have detailed studies of how these industrial policies worked, and these detailed studies make it plausible that the policies had their intended effects (Chang, 2009; Harrison and Rodriguez-Clare, 2009; Wade, forthcoming).
For example, Alice Amsden argues in The Rise of the Rest (Amsden, 2001) that East Asia’s success in manufactured exports – and in leveraging export success into product diversification and upgrading – was related to the governments’ use of ‘reciprocal control mechanisms’. Firms had to meet performance targets in exchange for special favors – such as targets for exporting, or local content requirements, or product specifications.
My own work on East Asia provides lots of evidence about the nitty-gritty of industrial policies and the various kinds of links between policy support and outcomes. I argue that East Asian industrial policy comprised two kinds –‘leading the market’ and ‘following the market’– where ‘leading’ refers to the government making an investment decision that private actors would not make, and ‘following’ refers to the government supporting some of the bets of private firms or supporting a marginal extension of the production frontier in a given product.
The classic example of industrial policy of the ‘leadership’ kind is Posco, the Korean integrated steel firm, which no private firm wanted to undertake and which the World Bank in the 1960s advised the Korean government not to undertake on the grounds that Korea had no comparative advantage in steel (and should stick to its comparative advantage in … radios and the like). By 1987 the World Bank itself swallowed hard and described Posco as ‘arguably the world’s most efficient producer of steel’, without revisiting its earlier advice (Wade, 2004, p. 319).
However, a lot of East Asian industrial policy was not the leadership kind. The ‘followership’ kind was aimed at accelerating movement in some of the directions that private entrepreneurs wanted to move in. For example, the Taiwan government used a fiscal incentive scheme which gave fiscal incentives to firms for that part of their production comprising specified frontier products. Firms producing products that met the specifications were eligible for tax holidays or accelerated depreciation or both. In 1982 the list of eligible products included ‘high-efficiency fluorescent tubes, limited to those which had an intensity of 80 lumen or above’. Later the threshold of fluorescent tube eligibility for fiscal incentives was raised as the volume of production of 80 lumen tubes increased (Wade, 1990, 2004, appendix A).
East Asian industrial policy officials also intervened by using administrative discretion to ‘nudge’ subsidiaries of foreign firms to link up with domestic suppliers. In one Taiwan case, Industrial Development Bureau (IDB) officials in charge of the input–output chain that included specialized glass considered that at least two Taiwan glass makers could meet the price and quality of the glass imported by the Philips TV factory if given a long-term supply contract and some technical help. They suggested this to Philips, which refused to consider the idea, saying it was happy with its existing arrangement to import from a Philips factory elsewhere. But then Philips began to experience delays in authorization to import the glass, which previously had been granted without delay. Philips protested. The delays lengthened. The IDB officials reminded Philips of the advantages of switching to domestic suppliers. To cut a long story short, eventually Philips got the message, entered into a long-term supply contract with a domestic producer and in so doing built up goodwill in the IDB. Meanwhile Taiwan’s glass-making capacity had made an incremental move up the value chain, which was the IDB’s objective.
While East Asian governments did undertake a lot of industrial policy of a ‘leading’ the market kind, they also undertook a lot of this incremental pulling and pushing kind, using a range of industrial policy instruments to steer the incentive environment of firms; and they implemented it constantly, decade after decade, nudging firms to upgrade and diversify. Followership industrial policy is far from ‘bureaucrats picking winners’, and therefore not much open to the familiar jeer, ‘bureaucrats can’t pick winners’. On the other hand, it is also very difficult to measure in terms of its impacts. Certainly none of the existing measures of industrial policy pick up its effects.
The development strategy combined import replacement with export orientation – notwithstanding that these are understood as mutually exclusive in the mainstream view. Firms were given a range of incentives to replace imports but were not thereby insulated from international competitive pressures (as tended to be the case in India and Latin America). Industrial policy officials monitored the price and quality of domestic replacements against international comparators, and judged the assistance in relation to the trajectory of the gap. Firms understood that assistance was conditional on them either reducing the price–quality gap or showing good export performance.
My argument is that the governments of low-income countries today should undertake industrial policy of the same general kind as the first and second cohort of East Asian capitalist economies. It could be called ‘open economy’ industrial policy – provided that ‘open economy’ is interpreted to include ‘import replacement’. (To repeat, East Asian industrial policy stressed import replacement every bit as much as export promotion; the two things were seen as complementary, like the two wings of a bird.) And a large part of it should be of the incremental, nudging, followership kind rather than picking the winners kind.
Going beyond the East Asian experience, the more general arguments for industrial policy today include the following. First, countries with large agricultural or mining sectors show a strong positive correlation between the growth of manufacturing output and the growth of productivity in both manufacturing and nonmanufacturing. This suggests that measures to accelerate manufacturing output can contribute to productivity increases in agriculture and some services, as well as in manufacturing itself.
Second, China’s ability to outcompete manufacturers across a wide range of products suitable for low-income economies – thanks partly to an undervalued exchange rate – means that manufacturing sectors throughout the developing world are under intense competitive pressure and many are shrinking. Saying that a developing country government should adopt policies of maximum openness (to trade, investment) – also known as a ‘level playing field’– is a recipe for the economy losing much of its manufacturing.
Third, the industrial policy should be of the ‘open economy’ kind, not the inward-looking kind used in many developing countries through the 1950s and the 1970s, and it should use ‘reciprocal control mechanisms’ where state assistance is given against performance. Just granting trade protection without performance conditions yields the preliberalization Indian automobile industry, which for decades produced replicas of 1955 British Morris cars.
Fourth, it is likely that a surge of innovations relating to environmentally sustainable living will come on stream and be deployed around the world over the next two decades. The surge will open new supply opportunities and raise the need to find new sources of demand, including previously excluded people in developing countries. A strategic industrial policy that combines the information, perspective and objectives of the public sector with those of the private sector can constitute the foundation of a new national development project.
3. Establishing a developmental state
It is one thing to make a case for a more proactive role of the state than neoliberalism and the Washington Consensus allow; it is another to identify how to do it. As we saw earlier, the Washington Consensus came to be extended over the 1990s to include a ‘good governance’ agenda for strengthening the capacity of the state to provide certain public goods (though not for strengthening its capacity to act as the coordination point for a national development project). The content of the good governance agenda was based on the argument that developing countries should undertake administrative reforms intended to move them towards the degree of formalization, or depersonalization, of economic rules found in developed countries. That meant formalizing and enforcing property rights, reducing corruption, improving the effectiveness of the bureaucracy, the judiciary and the police, and strengthening transparency and political accountability.
The argument postulated that as rules become more formalized, more impersonal, actors will become more confident in undertaking economic transactions; which will encourage them to invest more; which will raise growth rates – out of which the institutions that improve the security of transactions can themselves be further strengthened, in a virtuous circle.
Indeed, when developing and transitional countries as a group are compared to developed countries, they show markedly less formalization of rules. But – to repeat the earlier point about the direction of causality – this does not mean that the Washington Consensus good governance agenda is correct in urging developing and transitional countries to give high priority to establishing a system of formalized and impersonal rules as a means of promoting growth. If we use a weaker criterion of catch up than in the earlier discussion – average growth between 1990 and 2004 greater than or less than that of the developed country average – we find that roughly a third of a sample of 70 developing and transitional countries experienced catch-up growth (Khan, 2006). We also find that the key institutional features that differentiate the catch-up countries from the falling-behind countries are not those of the good governance agenda. Many catch-up countries have low scores in terms of the formalization of their rules, yet have grown relatively fast over a sustained period. Countries with equivalent levels of governance (or formalization of rules) show very different economic performance.
The key features that discriminate between the institutions of the catch-up countries and those of the fall-behind countries seem to be related to the state’s capacity to coordinate agents, stabilize their confidence in the state’s behavior and establish national development as an urgent overarching project (Meisel, 2008).4 These seem to substitute for across-the-board formalization of rules (the priority of the mainstream approach). On the face of it, raising the state’s capacity to coordinate a selected set of economic agents is a more feasible task than across-the-board formalization and enforcement of rules – a task which requires high fixed costs and many decades, and which often provokes fierce resistance, especially from those already in the elite.
Developing countries tend to be regulated by what Douglass North calls ‘limited access social orders’ (as distinct from ‘open social orders’ more characteristic of developed countries), or what could also be called ‘insider systems’ (North et al., 2006). It is important to stress that the East Asian catch-up cases, such as South Korea and Taiwan, were also regulated by limited access social orders. All through their high growth decades they operated with strikingly informal, personalized rules of the game and with well-developed insider systems, oiled by plenty of corruption – which nevertheless generated high and sustained economic growth – and carried out a subsequent transformation towards a formal and depersonalized mode of regulation.
To oversimplify, many of the catch-up countries created a system of coordination for the insider groups, which constituted what has been called the ‘developmental state’. The groups whose specific interests counted most in shaping the content of the ‘national good’ were incorporated in the coordination system, and their sustained interaction and negotiation in one or more ‘focal points’ encouraged them to mute their oligopolistic struggles for access to rents, to define convergent interests and thereby to forge a sense of common interest. Their interaction in the focal point created information and incentives such that it served wider interests than their own specific ones.
They interacted inside the focal point in the same way as in the wider society, using informal, personalized rules, but now disciplined by the logic of repeated interaction in the focal point and the emerging sense of a common interest. Their interactions would not score well by the criteria of the ‘good governance’ agenda.
The involved public officials acted so as to steer the private calculations of private coalitions in the direction of a larger and longer-term national interest. ‘Steer’ is different from top-down instruction (as in the conventional meaning of dirigisme), as it is also different from bottom-up capture; and it spans the conventional western opposition between private and public elites. It is a notion wholly unfamiliar in the US and the UK civil service, except in the defense industries where industrial policy flourishes though it dare not speak its name.
The basic bargain of the state–business alliance was that business got assistance from the state in finance, technology and marketing, in return for delivering ‘performance’ as measured by indicators like exporting, or import replacing, or reducing the gap between international and domestic prices, or increasing the proportion of local content.
From its role in helping business to succeed the state gained legitimacy for further interventions, and the circle became virtuous rather than vicious. On the other hand, where the state provided protection and subsidies without performance conditions, failure was more likely and the same circle became vicious. India’s automobile industry for several decades after the Second World War is Exhibit A of the vicious circle, as mentioned earlier.
One of the tasks of the public officials was to steer the insiders to support measures of ‘inclusionary’ growth, or sufficiently inclusionary to offset discontent among outsiders which could be exploited by one or other coalition of insiders in a way that would destabilize the insider system. In East Asia heavy public spending in rural areas after land reforms helped to buy off discontent (Wade, 1982, 2004), as did social housing in Singapore and the social security system in France.
Concrete manifestations of this kind of institutionalized coordination include Japan’s MITI, Taiwan’s Economic Planning Council and its Industrial Development Bureau, South Korea’s Economic Planning Board, Singapore’s Economic Development Board; and also numerous industry associations. In France the Commissariat General au Plan, established in 1946, was another important example. In some countries sectoral focal points have been more important than national ones: in Brazil – automobiles and aircraft, Columbia – coffee, Chile – several agroprocessing sectors, and Italy – IRI (the giant public sector holding company) and sector associations of a decentralized kind mainly in the center and north of the country.
In South Korea, the Economic Planning Board acted as the peak state organization, headed by the deputy prime minister, with authority over the other ministries. It was complemented by a dense array of subordinate forums of consultation between state agencies and business groups. Through these forums the big business groups (chaebol) sought help from the state in order to assist them to compete internationally and replace imports, and fairly standard kick-back percentages helped to oil the wheels of officialdom.
In the Philippines, by contrast, the big landed families – lacking such institutionalized focal points – used personalized rules of the game to obtain political protection in order to oppose industrial transformation, and used corruption to protect their existing sources of rents and to assassinate interlopers.
In India the Planning Commission was intended to have a similar peak coordinating and strategizing function, but the offensive against it by business in the 1950s opened the way for ministries to revolt against its authority, and it was never able to function in this way (Chibber, 2003).
Four conditions are important for the coordination mechanism to have pro-developmental effects. First, the state and the business groups should be relatively evenly balanced. The state must insist that business respect the quid pro quo; which means that in return for political patronage and protection business does not challenge the government politically or try directly to penetrate the political sphere (e.g. by placing family members in state agencies to act on businesses’ behalf). Failure to insist on this separation leads to the Philippines.
Where a small set of government and business elites, organized in peak organizations and organizationally separate from each other, are evenly balanced, with neither side having the upper hand, political protection and corruption (‘rent seeking’) can have relatively benign effects, as in Korea before the transition to democracy in 1988. Where one side has the upper hand, the effects may be less benign. Democratization in Korea had the effect of fragmenting the state and strengthening big business, breaking the previous balance. Thereafter the state lost effectiveness as a development agency and big business escaped national discipline (e.g. via borrowing abroad), both of which contributed to the 1997–99 crash. In the Philippines, on the other hand, democratization has somewhat restrained a state which under Marcos plundered from the top down, and has moved the nation a little way in the direction of a joint state–business project of a developmental state.
The second condition relates to the mindset of public officials. The public officials engaged in governing the market should operate with an activist, public service oriented mindset of the kind suggested by the quote displayed in the entrance to the Industrial Development Bureau in Taipei: ‘The most important thing in life is to have a goal, and the determination to achieve it’ (Weiss and Thurbon, 2004). In its assumption of activism this mindset is the opposite to that of senior British civil servants working on economic issues, as articulated by one who declared in 1930: ‘If I leave the office on Saturday feeling confident that in the past week I have done no harm, then I am well content’ (Allen, 1975).
The descendants of this breed of ‘do no harm’ civil servants came to shape the world view of the World Bank, and were key to the restoration of ‘right thinking’ in the Bank after the McNamara–Chenery interregnum. They allied themselves with the increasingly powerful neoliberal economists riding to power on the back of the Reagan and Thatcher revolution in economic policy. One of the early events in the rise of neoliberalism in the global development agenda was the World Bank’s Accelerated Development in sub-Saharan Africa: An Agenda for Action (1981), which portrayed postcolonial policy and performance in Africa as disastrous and ushered in what became the conventional wisdom on the state in Africa: an undifferentiated source of pathologies (Mkandawire, 2001).
Third, the state should create bifurcated political and economic administrative structures, such that political patronage can be given via political channels without sacrificing economic efficiency. In South Korea the government created in the early 1970s the New Community (Saemaul) movement, a giant hierarchy of offices, budgets, plans and activities centered in the president’s office and extending down to private firms, urban boroughs, government departments and rural counties (Wade, 1982). President Park declared it a ‘national spiritual revolution for a better way of life’, ‘a driving force for nation building’ via increased productivity, community participation in local infrastructure projects and education in the policies of the government and its great leader (himself). Saemaul groups from across the country were brought together periodically at various levels of the hierarchy, culminating in the Saemaul Institute near Seoul, for education and motivation. Through this channel much patronage and protection flowed, rewarding supporters and not rewarding dissidents, separate from the economic departments of government (Chen, 2008; Wade, 2004).
Finally, the state officials who are doing the nudging of private firms must not have significant public resources under their discretionary control. They may recommend that a certain factory buy a new type of machine tool, but they must not be able to decide on the use of a subsidy for the purchase. The decision about a grant or cheap credit must rest with some other body. Taiwan’s Industrial Development Bureau officials had few resources under their direct control.
The mainstream view, as expressed in the quotations at the beginning, continues to dominate development policy making. It is an Old Testament kind of religion, underpinned by dichotomies like ‘import substitution vs export orientation’ and ‘public vs private’, with the virtues loaded on to one side and the vices on to the other. It presumes that the typical developing country state is ‘neopatrimonial’ and therefore even worse than the state-in-general, and so should be constrained as much as possible by private markets.
The preceding discussion illustrates how we can get beyond such crudities by examining actual states and their links with business interests. Such examination suggests rules of thumb for the institutional configurations out of which appropriate development strategies are likely to emerge with some chance of being effectively implemented.
The preceding discussion also illustrates why one should not equate industrial policy with ‘picking winners’, and should not assume that it is like a light switch, either ‘on’ or ‘off’, ‘present’ or ‘absent’. Industrial policy can be big or small, discontinuous or incremental. It can ‘lead’ the market or – less riskily –‘follow’ the market. Much East Asian industrial policy was of the ‘followership’ kind, and involved nudging private producers to extend their production capabilities by degrees (producing more sophisticated fluorescent lights, for example) and nudging foreign affiliates to switch supplies of intermediate goods from imports to domestic firms. East Asian industrial policy provides concrete examples of how to do small-scale, market-following industrial policy of a nudging kind, which may be relevant in countries with embryonic manufacturing sectors and few highly trained engineers. But before this can be done, politicians and civil servants have to stop being ‘the slaves of some defunct economist’, in Keynes’ phrase, or more accurately, the slaves of the ideas of many very living economists located in or trained in the centers of the discipline in the United States and Britain – not coincidentally the two hegemonic powers of the past 150 years.
The global economic crisis that began in 2008 is forcing even liberal market states like the US and the UK to overcome their long aversion to taking responsibility for structural outcomes and lead the effort to save their plunging economies. As captured in a New Yorker cartoon, a messenger runs into the castle just as the executioner is about to behead the king and shouts: ‘Stop! Wait! Government is not the problem, it is the solution’. As the governments of the US and UK move beyond ad hoc emergency responses towards more strategic ones, their moves may open the way for a larger reconsideration of the role of the state in development in places like the World Bank and the IMF. As the imperative of global behavior change towards environmentally sustainable living takes hold, the role of the state in steering private responses is likely to become even more important.
No alternative policy paradigm to neoliberalism has yet attracted wide support. Instead, what we see is a new ambiguity, with states beginning to experiment more than in the past few decades. We see small signs of this new ambiguity even within the World Bank and the IMF. The World Bank’s chief economist and senior vice-president, Justin Lin, a Chinese national, recently published a paper titled ‘New Structural Economics: A Framework for Rethinking Development’ (Lin, 2010), which is mainly about industrial policy, an almost taboo subject in the World Bank for the past three decades. The paper emphasizes the role of government in helping firms exploit the country’s existing comparative advantage – but says little about the role of government in creating comparative advantage. It is written in the spirit of ‘crossing the river one stone at a time’. Remarkably, though published in a World Bank outlet, the paper identifies Lin by his Peking University affiliation, not by his influential World Bank affiliation.
As for the IMF, its Stand By Agreements (SBAs) with countries in the current crisis show more flexibility than those it deployed in the East Asian crisis, in several ways (Grabel, 2010). First, the SBAs give more flexibility to governments in terms of fiscal policy, even as they continue to insist – as in the East Asian crisis – on stringent monetary policy, including high interest rates. Contrary to the deficit phobia that prevails in many western governments, the IMF’s new study, Exiting from Crisis Intervention Policies, argues:
Unwinding public intervention too early could jeopardise progress in securing a sustained economic recovery, while maintaining intervention for too long could distort private incentives and pose risks to price, financial and fiscal stability. One of the key lessons from experiences of similar crises is that a premature withdrawal of policy stimulus can be very costly, particularly if the financial system is weak. Thus, in the current context, the potential risks associated with an early withdrawal of policy stimulus seem to outweigh the risks of maintaining it for longer than possibly needed (http://www.imf.org).
A second important difference from the Fund’s interventions in earlier crises is that today its SBAs give more emphasis to social protection for the poor and vulnerable. The third difference is capital controls. The Fund has opposed restrictions on capital mobility for decades; it even tried to secure a change of its Articles of Agreement in 1997 to establish a presumption that all members of the Fund would move to and maintain open capital accounts as a condition of membership of the Fund. The effort was aborted by the East Asian crisis, then beginning, which made clear to many governments the dangers of open capital accounts.
In the current crisis, none of the SBAs endorse capital controls, with the exception of Iceland’s; so in this respect the Fund mostly continues to push its old agenda. Furthermore, in all the many reports about the current crisis written by the IMF and the World Bank, there is almost no discussion of capital controls. Both organizations have gone strangely silent on the issue. But if – as is likely – Iceland’s stringent capital controls are seen to have helped more than they hindered, the experience may help to change the Fund’s mind more generally.
The Fund’s reaction to Brazil’s imposition of capital controls in October 2009 may be a harbinger. When the Brazilian government in October 2009 introduced a modest tax on money inflows for equities and bonds in order to curb the appreciation of the exchange rate and the blowing up of asset bubbles, and then extended the controls to cover other kinds of capital inflows, the Fund’s reaction was quite mild, far from the outrage it expressed when Malaysia introduced capital controls in 1998. The managing director, Dominique Strauss-Kahn, commented cryptically: ‘There is no reason to believe that no kind of control is always the best kind of situation’ (quoted in Guha, 2009).
He went on to express sympathy with policy makers facing a sudden upsurge in capital inflows. He acknowledged that the standard policy prescription to limit bubbles – to allow the currency to appreciate – might be difficult if competitor nations such as China did not also allow their currencies to appreciate at a similarly rapid pace (Guha, 2009). So by IMF standards, its reaction to Brazil was not even a slap on the wrist. The fact that Brazil has recently become an important financier of the IMF no doubt helped to mute the IMF’s disapproval.
Indeed, in February 2010 the IMF published a staff report espousing a significantly more positive view of capital controls than it has taken for the past three decades. In the words of the IMF announcement of the report, ‘controls on capital inflows to emerging market economies can usefully form part of the policy toolkit to address the economic or financial concerns surrounding sudden surges in capital’ (IMF, 2010). If by 2020 it becomes an international norm that governments in financial difficulties can legitimately impose capital controls, this one change will mark a big departure from the neoliberal world of today.
And a fourth indicator of recent change at the IMF in a more flexible direction is its much greater effort during this crisis to engage with civil society organizations beyond governments and find out what they want, including with NGOs. This is quite new, and when combined with the new emphasis on protecting the poor and the opening of a chink of tolerance for capital controls, it may indicate a willingness on the part of the IMF to give more policy space to national governments. Having acquiesced in the economic policies and financial engineering that led to the Great Recession, the IMF acknowledged recently that ‘In general, fiscal and monetary stimulus may need to be maintained well into 2010 for a majority of the world’s economies, including several of the largest’ (Keegan, 2010).
My main point is that, on the one hand, there is no sign of a coherent policy alternative to the mainstream view of the past three decades. But on the other hand, the erosion of confidence in this mainstream view – at least in policy-making circles, if not yet within the academy – may widen the room for countries to experiment. How this translates into reforms of the international financial system depends heavily on whether the new economically powerful states assert their preferences or buy into the agenda of the developed countries. In the optimistic view, we will have established by 2020 an international financial system that sanctions more national and regional diversity of institutions than was the direction of travel for the three decades after 1980, which serves rather than feasts on the productive economy and which provides room for a new generation of developmental states.
Wolf echoes many economists and political scientists who have claimed that ‘administrative competence’ is the ‘single most important explanatory variable’ of differences between countries in development performance.
Note that ‘80 per cent in ten countries’ is from the mid-1990s. See Roberts et al., 2003.
This section is drawn largely from Kurtz and Schrank (2007).
I acknowledge the tension between this proposition and the earlier-reported finding about growth accelerations.
Robert Wade is Professor of Political Economy at the London School of Economics and Political Science, and winner of the Leontief Prize in Economics 2008. His book Governing the Market won the American Political Science Association’s prize for Best Book in Political Economy. He is a member of the Financial Times Economists Forum.