wade conclusion LESSONS FROM EAST ASIA

CONCLUSIONS (2): LESSONS FROM EAST ASIA THE DEBATE about the role of the state in economic development demonstrates the power of infinite repetition as a weapon of modem scholarship. The issue is nonnally posed in tenns of the "amount" of state intervention or the "size" of government. The neoclassical side says that more successful cases show relatively little intervention in the market, while less successful cases show a lot (Brazil and Mexico compared to East Asia; or sub-Saharan Africa at the bottom). It uses this evidence to urge governments to shrink the size of the state and remove many of its interventions in the market. The political economy side says that the neoclassicals have their facts wrong: the most successful cases show "heavy" or "active" intervention. It concludes from this evidence that governments can, in some circumstances, guide the market to produce better industrial perfonnance than a free market, even in the absence of neoclassical-type market failure. But neither side has been noticeably enthusiastic to specify just what evidence would be consistent with its position and what would not. Both have exercised a selective inattention to data that would upset their way of looking at things. So the debate about the role of the state is less a debate than a case of paradigms ("parrot-times") talking past each other. I have shown for Taiwan-and suggested for Korea and Japan-that ample evidence is available in support of both the free market/simulated free market (FMlsM) and the governed market (GM) theories. This poses an identification problem. How important are those facts which are consistent with the FMisM interpretation, and how important are those which are consistent with the GM approach? My argument is simply that the GM facts are too important to ignore in an explanation of Taiwan's (and Korea's and Japan's) superior performance. This challenges economics to deploy--or invent-theories which will make the non-neoclassical facts of East Asia analytically tractable. But does it also support the prescription that other middle-income countries should try to govern the market in a broadly similar way (with appropriate adjustment for national circumstances)? That depends on the answers to three questions. First, are the conditions of the international economy as favorable to a rapid, forced, and export-dependent industrialization today as they were for Taiwan and Korea? Second, is there a general economic rationale for GM policies? Third, can governments significantly improve their administrative and political capacity to govern the market? 346 CHAPTER II CONDITIONS OF THE WORLD EcONOMY Both the FMlSM and the GM interpretations of East Asian success emphasize the importance of domestic factors, in particular "right" policies-though they differ on what constitutes' 'right." Implicitly, they assume that the trajectories of states are parallel and theoretically independent, each separately subject to the same economic tendencies. Development is a kind of marathon race, in which each runner's position is a function of his internal resources and in which all runners could in principle cross the finish line at the same time. Yet it is clear from what has been said that a good part of the reason for East Asian success has to do with international factors. These created opportunities for relatively low-cost industrial production sites to be integrated into the world economy. In the 1960s several conditions came together to produce at one and the same time relatively favorable access to industrial country markets, dramatically increased access to international finance, and increasing relocation of production by multinational corporations to low-wage sites. These conditions created opportunities, but did not detennine which countries would seize them. Which countries seized them can be explained partly in tenns of domestic factors-including the existence of an industrial base resulting from prior import-substitution and the existence of a hard state pursuing GM policies. Location and geopolitical importance are also relevant. The United States "invited" Taiwan, Korea, and Japan to become economically strong because of their location on the West's defense perimeter (which made them more strategic than, say, the Philippines, Indonesia, or Brazil). Japan, the most dynamic economy of the postwar era, had special ties with Taiwan and Korea derived from proximity and colonial history. Hence, part of the success of GM policies in East Asia is due to the favorable historical and international conditions in which they were implemented (Bienefeld 1982; Brett 1985; Cumings 1984). To the extent that these factors are different at other times and places, this throws doubt on the possibilities for other countries at other times to emulate East Asian success. A central difference between the world economy of today and that of the 1960s, when Taiwan, Korea, and Japan made big inroads into Western markets, is that it is no longer in an expansionary phase. There has also been a dramatic fall in the demand for unskilled labor and raw materials per unit of industrial production. Consequently, developing countries in the 1980s face an external environment more hostile than in any previous decade since the Second World War (IMP 1988; Stewart 1988). They are doubly squeezed on trade and on capital. Growth in world output slowed from 4.1 percent in 1970- 79 to 2.6 percent in 1980-87. Tenns of trade for nonfuel exports from developing countries deteriorated from a 1.1 percent per year decline in 1970-79 to ".a 1.7 percent per year decline in 1980-87. Protection in developed country markets has increased since the early 1970s, accelerating in the early 19805 to LESSONS FROM EAST ASIA 347 the point where by 198621 percent of manufactured goods imported into the United States and Europe were restricted by quantitative barriers (UNCTAD 1987: table IVA). This protection is being applied with special discrimination against developing countries. Eighteen percent of manufactured imports from developed countries were covered by quantitative restrictions, and 31 percent of manufactured imports from developing countries. Yet manufactured imports from developing countries account for a mere 1.5 percent of manufactured consumption in developed countries. Meanwhile the microelectronics revolution has reduced the advantage of cheap labor sites, slowing the inflow of foreign direct investment to developing countries at large. The debt service burden for indebted developing countries increased sharply in the 1980s, while voluntary private lending to developing countries almost stopped (US$3.5 billion in 1987 compared to $73.4 billion in 1980). On top of these trends has come a sharp increase in the volatility of the international economy, and therefore much more uncertainty facing developing country governments and producers (UNCTAD 1988a). With the internationalization and deregulation of financial markets, financial capital is ricocheting around the world in amounts thirty to forty times bigger than trade flows. The relationships between exchange rates and trade, interest rates and investment, and fiscal and monetary policies have become unhinged (Drucker 1986). Governments' ability to control as fundamental a parameter of economic activity as money supply is diminished, and long-term investment is depressed. Dealing with currency fluctuations "is like changing the handicap in golf on every hole," protested the president of Sony recently. "Wouldn't you lose interest in playing golf eventually? If money scale expands or shrinks every day in different currencies, how can we make up our minds to invest?" (Toronto Globe and Mail, 1 June 1987). If Sony finds long-range investment difficult in current conditions, think of the predicament of would-be exporters and investors in developing countries with free trade and capital movements. They are forced to adjust and readjust to signals from the international economy which are essentially short-term. These adj,ustments to price signals that turn out to be misleading guides to economic fundamentals may cause high costs. They are "distortionary" in a sense different from but as important as the conventional meaning in economics, of price misalignments which arise when an economy has not adjusted sufficiently to international price signals (Bienefeld 1988). The implications are ominous for those developing countries that would seek to "follow the NICS." If many countries are to succeed in increasing their exports of light manufactures, world trade would have to expand fast; but all the signs are it will not. The new protectionism is directed especially at the light manufactured goods which the next tier countries are urged to make their leading export sectors. Moreover, the East Asian four are "stretching" their industrial structures as they expand into more advanced sectors, using tech- 348 CHAPTER 11 nology to remain competitive in light manufactures and thereby only slowly vacating these sectors for others to enter. It will therefore be more difficult for the others (such as Mexico, Brazil, Thailand, Poland, and Hungary) to use industrial and trade policies "successfully." Underlying these ominous trends are shifts in technology which imply potentially far-reaching effects on the competitive position of nations. It is difficult to forecast these effects, for the implications appear to be mixed. The increase in the capabilities of machines to perfonn the tasks of unskilled labor may facilitate a shift in the location of production back to the developed countries; but it may also enable some developing countries to become more competitive through better quality control and cheaper engineers. The reduced importance of unprocessed inputs worsens the export prospects of raw material producers; but the new technologies enable them to process the raw materials in-country, and free some regions from some of nature's constraints (biotechnologies can make deserts bloom). Faced with these new dangers and opportunities, what broad lines of economic policy should developing countries follow? We can be fairly sure that policies to impart an East Asian kind of directional thrust will have a smaller effect than they did in East Asia, if for no other reason than the less favorable conditions of the international economy. On the other hand, this does not mean that FM/sM policies are the better alternative. We have seen that the confidence with which the neoclassical school prescribes liberalization and privatization cannot be grounded in theory, for the theory which shows how liberalization and privatization generate growth is scarcely developed. We have also seen that it cannot be grounded in the experience of the East Asian NICS, for with the partial exception of Hong Kong they have pursued a mix of policies-many of which are inconsistent with neoclassical prescriptions. Hong Kong, though historically much wealthier than Taiwan and Korea, has not been doing as well in tenns of income growth and industrial transfonnation. But what about the other 98 percent of countries? Does the experience of a broad cross-section of countries provide solid grounds for the neoclassical confidence? My review of the cross-sectional evidence in chapter 1 suggests not. Given this, we might usefully deploy a more inductive approach to policy and policy-making. We can ask what policies the most successful countries actually adopted, and then construct a rationale for why those policies may have helped their growth. Due recognition has to be given to the Darwinian or Malinowskian fallacy in this exercise-the assumption that because something exists it must be vital to the survival of the organism or society in which it exists. Translated into East Asian terms, this leads one to argue that because these most successful countries used high protection, tightly controlled finan- ~"cial systems, and the like, such measures must have been vital to their success. But what could be called the Ptolemaic fallacy is more prevalent and more LESSONS FROM EAST ASIA 349 inhibiting of learning: the assumption that only those features of economic policy consistent with neoclassical prescriptions could have contributed to superior economic performance, so that everything else can be safely ignored. It seems useful to err for a change on the side of the former. A distinction must be made between what is consistent with neoclassical theory and what is consistent with neoclassical prescription. There is room within the confines of neoclassical theory for practically any mix of markets and intervention. Most neoclassical economists argue the costs of government regulation and industrial targeting and the virtues of very wide choice by individual market agents on (often implicit) empirical grounds. The empirical significance of market failure has been exaggerated, they say, and government efforts to repair such failure are likely to make matters worse because government failure is empirically more acute than market failure. 1 In what follows 1 use ideas that are familiar in neoclassical analysis, as well as some that are unfamiliar, to reach different conclusions about the possible economic benefits of governed market policies. These conclusions have the merit of being consistent with what the governments of very successful economies-the East Asian ones-actually did. I state the argument in the form of six prescriptions for micro- and mesoeconomic interventions. 2 The argument is most relevant to the circumstances of newly industrializing or newly industrialized countries with per capita incomes in the middle-income range (World Bank 1987a). For the most part I duck the question of its relevance to the industrialized market economies. I am unclear about how a world economy would work in which the leading economies, especially the United States, adopt the kinds of actions endorsed here. Past experience suggests it could be benign: the laissez-faire world of the 1920s had disastrous economic consequences, while the postwar era of stable but negotiable exchange rates and national controls over capital movements generated steady expansion. Keynes used his understanding of the prewar boom and bust to argue passionately in favor of import controls and central bank control of international capital movements as instruments <;>f postwar economic management (Crotty 1983; Polanyi 1957; Bienefeld 1988). But in any case, the lateI On the frontiers of economics the theoretical analysis of market failure has become a minor growth industry in the 1980s. But note what I call the "Helleiner effect." Gerry Helleiner ob· serves: "On the frontiers of the discipline. vigorous ex.perimentation can be found in the juggling of assumptions, the empirical testing of hypotheses. and the adaptation and improvement of both theoretical and quantitative economic models. Once in the difficult world of policy fonnation, however. students of economics are prone to forget all of the qualifications and assumptions, and frequently ajlply instead the simplest and crudest versions of the models they were taught, using, as they would put it, only 'the basic principles' .• (1981a: 541). 2 Here. as in earlier discussion, I say little about agriculture and agriculture· industry links. But in many countries the starting point for an internal demand· led strategy must be in farming. See Adelman 1984; Lipton 1977. I also say little about macropolicies. though such policies affect competitiveness both directly and indirectly. 350 CHAPTER II comer industrializers, which constitute only a tiny part of world income and world markets, can use different principles from those appropriate to the older industrialized market economies. They can, as have many latecomers before them, free ride on the (more or less) liberal norms embraced by the latter out of self-interest. After all, the per capita income of the fifty-two middle-income countries is still, after four decades of self-conscious development, only 10 percent of the per capita income of the "industrialized market economies" (2 percent for the low-income countries; World Bank 1988: table I). My argument is also relevant mainly to noncrisis conditions, when a longerterm view can be taken. Sadly, that currently excludes many countries of subSaharan Africa, the Caribbean, and Latin America. Fully two-thirds of middle-income countries had negative growth of gross domestic investment in 198{}-86, compared with only 2 percent in 1965-80; one-third had negative growth of GDP (World Bank 1988:table 4). Many have not been investing enough to maintain essential infrastructure. Many are unable to obtain even basic economic statistics, because so much activity has moved into grey and black markets. I also assume benign political leaders, whose concerns go beyond using state power to support the affluence of a small group. Some rulers, it is true, are predatory, in the sense that their efforts to maximize the resource flow under their control erode the ability of the resource base to deliver future flows. In these cases enhancing the power and autonomy of the state could be disastrous. But states vary in terms of the benignness or maliciousness of their leaders,3 and the more any particular case is toward the benign end of the spectrum the better the argument applies. Prescription 1 : Use national policies to promote industrial investment within the national boundaries, and to channel more of this investment into industries whose growth is important for the economy's future growth. We must note at the outset that the objective of such policies is not efficient allocation of resources in a Pareto-optimum sense, but growth and innovation. This means that the theorems of welfare economics about the conditions of market failure are largely irrelevant for development purposes. They judge failure in terms of the allocation of resources to the most efficient uses, rather than in terms of the generation of new resources (Kaldor 1972). Theories about market failure in a growth context are not well developed. Here I shall do no more than sketch out a plausible rationale for national industrial policies. The first step is to consider why the government should take steps to intenl A vampire state at one extreme, a ruminant state at the other? The vampire extracts so much as to debilitate; the ruminant grazes the resource base while fertilizing it at the other end. "Ruminant" is John Waterbury's happy term, while "vampire" is the word used by a senior Kenyan fInance official to describe his own state in the late 19805. LESSONS FROM EAST ASIA 351 sify the level of investment and reinvestment within the national boundaries. The point of interest is not the causal effect of investment on growth, for that is well established theoretically and empirically (Romer 1987). So also is the connection between high investment and growth, on the one hand, and high labor demand; and between high labor demand and high wages. The point of interest, rather, is the need for political power to focus the investment process on the national tenitory, which means channeling the options of both domestic capitalists and foreign capitalists by means of import restrictions, domestic content requirements, foreign exchange controls, conditions on the admission of foreign investment, export incentives, technology incentives, and the like. The reason is that as capital becomes internationally mobile, its owners and managers have less interest in making long-term investments in any specific national economy, and hence less interest in the overall development of any specific economy-including their home base. As wages rise, they may be inclined to relocate their assets abroad, or divert them into short-term speculative uses at home, or use their influence over state power to keep labor costs lower than otherwise. From the perspective of a national interest, however, they should be encouraged or cajoled to reinvest at home, and specifically to invest in technological improvements as a way of remaining internationally competitive despite higher wages. For the domestic workforce is not internationally mobile, and its rising real wages are a primary indicator of developmental success. This argument has to be qualified in several ways-by the desirability of some outward foreign investment in terms of a national interest test, and by the problems of overcoming purchasing power constraints on the investment cycle if large export markets are not available. But the qualifications do not change the basic point. Empirically, the work of Alexander Gerschenkron (1962) and Dieter Senghaas (1985), among others, supports the proposition that the more successful European latecomers used a political mechanism to channel the competitive process in the direction of higher-wage and higher-technology activities. (The United States is an exception, where the same result occurred because of the scarcity pf labor in relation to land and capital. Hong Kong is also a partial exception. Wages have risen and the benefits of growth have been widely diffused even with capital operating wholly in terms of an international perspective partly because of its very small size and partly because of its role as a regional service center. And to repeat, in the past decade or so it has been less successful in transfonning its industrial base than Taiwan and Korea.) The next step is to consider why government efforts to concentrate investment in selected industries may help overall growth and productivity. One reason involves economies of scale and learning. Whereas neoclassical analysis normally assumes rising cost curves, in many manufacturing processes a doubling of production volume per unit of time gives rise to a substantial fall in unit costs, commonly on the order of 20 percent. But the size of plant or 352 CHAPTER II firm required to achieve these economies of scale is typically large in relation to the existing assets of firms in developing countries. 4 The risks confronting potential investors are therefore high, and the investment process will be slowed if the risks are not partially lifted. If domestic producers are given assistance to enable them to compete against foreign suppliers in the domestic market despite higher costs, they may be able to expand their production volume to the point where, thanks to economies of scale and the transactions cost of imports, they can compete without further assistance. The Japanese-targeted industry strategy takes this logic further. With the government providing protection and socializing some of the risks of largescale investment, firms are able to price exports at below current average costs in order to· gain market share against foreign rivals. The government in effect carries the firms' negative cash flow through these various forms of subsidy. As the firms capture market share and increase production, costs eventually fall to match this "forward" price. Assistance is then removed and applied to the next set of higher-technology industries to be nurtured to international competitiveness (Magaziner and Hout 1980). Gains in productivity come not only from static economies of scale but also from technological effort based on experience. Even without increases in volume, repetition of production can lead to productivity improvements. Decreases in real unit costs of 100 percent in less than a decade are not unusual in infant industries that have become internationally competitive (Pack and Westphal 1986:106). This gain cannot be obtained simply by buying into the technology; it requires prolonged experience of production, and generally also investment in a deliberate effort to adapt the technology-so the gain does not accrue simply by dint of repetition. (There is a difference between, say, steel and electronics. Learning in a steel plant, where once bought the basic technology will not be changed for a decade, consists of continuous incremental improvements in operating procedures and equipment. Learning in electronics consists not only in how to make existing products more efficiently but also in how to design and produce a new product every nine months-production and innovation learning, perhaps, or static and dynamic learning.) In short, the forces that lie behind the orthodox assumption of rising cost curves are in many manufacturing processes overwhelmed by economies of scale and learning. Producers who expand production can have falling unit costs. They can therefore race down these falling cost curves and capture market share from existing producers. However, when the international productivity frontier is itself advancing rapidly (as in electronics), the time needed for an infant industry to catch up may be long and the amount of assistance large. This strengthens the case for selective rather than across-the-board assistance. 4 The new flexible manufacturing techniques decrease economies of scale at product level; but tend to go with increases in scale at firm level, because of rising indirect costs (R&D, marketing). LESSONS FROM EAST ASIA 353 Of course, if producers have perfect foresight and if capital and insurance markets are perfect, these potential cost gains need not warrant state assistance. When markets covering all possible future conditions are available to all and when lenders base lending decisions on expected future costs and earnings rather than on a firm's existing assets, neoclassical theory could not justify state assistance to finns facing potential economies of scale and learning. The justification comes from the unreality of these assumptions in developing country conditions. Borrowing is typically constrained by a finn's existing assets and lack of reputation. Equity and insurance markets are weak or absent. Building them up is a slow and difficult business. In the meantime government comes in as the "second-best" risk insurer and capital provider, so as to bring a private cost/benefit calculus more into line with a social cost! benefit calculus. Max Corden, impeccable neoclassical credentials notwithstanding, concludes that, "in spite of many qualifications, a valid, practically relevant infant industry argument for subsidization of new manufacturing industries resting on capital market imperfections can be made for many lessdeveloped countries" (1974:255). Governments, in other words, can substitute for missing and difficult-to-develop capital markets (Stiglitz 1989). A third justification, in addition to scale and learning economies and capital market imperfections, comes from that most elastic of concepts, "externalities" or "spillovers." In the general sense, external costs or benefits are those which are created by a firm or other economic agent but which do not bear on or accrue to it. Simultaneous externalities occur where a finn's potential gains from an investment are contingent upon complementary decisions by other finns. Even if all the parties know they would gain by coordinating their investments to capture the externalities, they may face inherent contradictions of interest, as in a Prisoner's Dilemma game.s Hence market prices may not adequately signal the interdependence that exists among these investment decisions, and uncoordinated finns may invest at sUboptimal levels from a national perspective. A big push, involving simultaneous expansion of several industries, can insure the profitability of each investment, even though each on its own would be unprofitable. One important reason is that such simultaneous expansion helps to overcome the constraint of a small domestic market, when entry and participation in world trade entails significant costs. There is also a second kind of externality, sequential rather than simultaneous. Sequential externalities occur where a large upstream plant would, if built, induce the entry of downstream finns to make use of new profit opportunities created by the upstream finn but not appropriable by it. The upstream plant brings greater social benefit, in the fonn of induced downstream growth, than is reflected in its private profit (Biggs and Levy 1988). S For an introduction to the literature on Prisoner's Dilemma, see Rapoport and Chammah 1965; Lipton 1985; Wade 1988d. 354 CHAPTER II To the extent that simultaneous and sequential externalities are spatially concentrated the investment process may become cumulative as more spillovers are generated and more finns enter to secure access to them. Market size increases, helping to reduce the risk of large-scale investments. In practice, diseconomies of agglomeration (such as congestion costs) may arise and some factors will be immobile, so that competitive advantage can shift from established centers to new ones (Brett 1983). We can expect the advantages of coordinating investment to capture these externalities to be particularly large where the domestic market is small, where the input-output structure is "holey," where entrepreneurs with the resources and experience to undertake large-scale investment are few and far between,6 and where access to world markets is limited by high transactions costs and trade restrictions. The role of government is not only to push the process by coordinating upstream industries, but also to lower entry barriers at the downstream end to facilitate the induced response. Conventional notions of "efficiency" are poor guides to what the government should do, because when externalities are considered one may find that a set of microeconomically less efficient industries, considered individually, produce a macroeconomically efficient result through the linkages between them; while a set of microeconomically efficient industries may produce a macroeconomically inefficient result. 7 Another justification for governing the market has to do with the adverse effects of market instabilities on long-tenn investment. Any moderately complex economic system encounters a source of instability arising from the uncertainty inherent in the attempt to match present supply decisions with future demand decisions. One would expect that if prices and quantities are left wholly to the instabilities of the market, investment in industries or technologies which require a large commitment of time or capital may not be made, and a higher than desirable proportion of the economy's investment will go into quick return projects. Individual finns on their own may be more inclined to stick within a narrow range of familiar product lines than branch out into 6 In situations characterized by lack of efficient risk-sharing institutions such as stock markets, • 'the size of the entrepreneurial class is smaller than would be first-best optimal" (Grossman 1984:613). 7 A longer discussion would distinguish between pecuniary and technological externalities. Pecuniary externalities occur through market transactions. They make for market failure, in the neoclassical sense, only in the presence of economies of scale. Gove1'l1ment coordination can help to reduce this source of market failure provided the structure of domestic prices is not closely fixed by the structure of international prices (perhaps because of the investment costs of getting into exporting). Technological externalities are not transmitted through market transactions. Examples include labor training, when a firm invests in training workers who then join another firm, or machinery adaptation, when the firm invests in modifications to existing machines which are then easily copied by others. Technological externalities may be more important causes of neoclassical market failure than pecuniary ones, and may occur independently of capital market imperfections. See Pack and Westphal 1986 and references therein. LESSONS FROM EAST ASIA 355 new industries and products. It' may well be that, within limits, price instability has a more adverse effect on growth than price distortions as conventionally defined. A context of deliberately created stability, achieved by riskspreading mechanisms such as protection or subsidies, can facilitate industrial deepening, export expansion, and political compromises to share adjustment costs. Again, a role for industrial targeting can be warranted by the fact of differences between industries in prospects for long-term growth in output, profits, and wages. Unassisted entrepreneurs may not have either the foresight or the access to capital to follow long-term potential. Their decisions may lock the country into specialization in industries with inferior prospects (an issue beyond the scope of comparative advantage theory). Given a world of technical change, falling cost curves, and differential rates of growth across industries, it can be rational for a government to select from within the plausible industries those which have high growth potential and to use the powers of government to supplement those of the market in marshalling resources for entry and successful participation (Scott 1985:95). This means diverting resources from currently profitable activities into ones that might be fast-growing and/or profitable in the future-which is risky. But any successful large company follows a strategy of diversifying from currently profitable activities into new ones, on the assumption that the future will probably be different from the present. Governments at the national level can aim.to carry through a parallel strategy of diversification. The scarcer the supply of capital and the higher the entry barriers of the new industries, the stronger the case for selective assistance. But can comparative advantage really be modified, made, or achieved in this way? Traditional theory takes comparative advantage as exogenous, largely determined by "factor endowments." In a gross way these considerations are still relevant: Burundi should not go in for computer production just yet. But as Bela Gold says, Virtually all empirical findings of comparative advantage represent no more than ex post facto rationalizations of past trade patterns, often reflecting market interventions rather than substantial differentials in efficiency and costs. Moreover, even the demonstrable comparative advantages prevailing in a given period have frequently been undermined and even reversed thereafter through determined efforts to advance technologies, shift input requirements, alter transport costs, and develop new markets. The very identification of current comparative disadvantages often represents the first step in developing means of overcoming them. (1979:311-12, emphasis added) William Cline reaches a similar conclusion: "Increasingly, trade in manufactures appears to reflect an exchange of goods in which one nation could be just as likely as another to develop comparative advantage, and the actual outcome is in a meaningful sense arbitrary" (l982a:39-40). In place of "ar- 356 CHAPTER 11 bitrary" I would say "subject to strategies of firms and governments." Talk of "revealed comparative advantage" (measured by the relative preponderance in a country's exports of product x compared to its preponderance in the trade of the world as a whole) is hence misleading, for the export pattern may reveal government assistance as much as factor endowments. And factor endowments, it should be remembered, can themselves be arranged on a spectrum from unalterable to alterable, with sunshine at one end and knowledgeable brains at the other. The classic case of Portuguese wine and British sheep reflects unalterable natural endowments; the modem case of British whiskey and Japanese electronics reflects human capital build-up, long-term horizons, and other acquired advantages. Government assistance can create new advantages of the acquired kind, some of which are industry-specific. The popular belief that governments cannot "make winners" rests on remarkably little empirical research into the record of different governments in selective industrial promotion. Many governments, especially in small countries, routinely target industrial assistance at specific industries and even at specific firms, particularly where economies of scale call for a minimum level of subsidy per firm. Yet we do not have systematic data on the performance record of different governments which would allow us to distinguish those with one failure out of four from those with seven failures out of eight. (No failure is itself failure, because it means that the targeters are not taking enough risk.) Research on this question has to balance the record of government failure against the record of failure by private business; and examine, too, what happens to economies where few transformation projects are attempted because the government declines to take an initiative and private business declines to take the risk. In short, several considerations~conomies of scale and learning, capital market imperfections, externalities, market instabilities, and differential growth potential-give grounds for state assistance to industry and to some industries more than others. 8 8 The question of the degree of selectivity of protection and promotion has hardly begun to be studied empirically. Pack and Westphal say of Korea: "The set of promoted infants has changed over time, but it has generally been small at anyone point in time" (1986:94). They prescribe that "selecti ve intervention must indeed be selective," focused on .. a select few extensive changes" (1986:118), those that result in distinctly new capabilities and occur through investments with large indivisibilities. On the other hand, Shinohara contrasts the general prescription for infant industries with Japanese practice. "In general, the nurturing of infant industries is limited to a certain period of time and to a certain number of industries. In Japan, however, these measures were across the board and applied to almost all industries. Because of the vastly extended promotion of infant industries and across-the-board encouragement of exports, MITI'S approach ran counter to the basic principles of modern international economics" (1982:49). One of the theoretical points at issue is the industry-specificity of the dynamic factors; the more they .¥e industry-specific the smaller should be the set of promoted industries, while the more they are -a function of the size of the entire industrial sector the more widely should assistance be spread. LESSONS FROM EAST ASIA 357 These arguments for assistance to particular industries are especially relevant to capital goods industries and microelectronics. In medium- and largesized countries, developing a domestic capital goods industry is a necessary condition for enhancing a country's capacity to develop embodied industrial innovations. The social costs of not doing so are not fully captured in freemarket prices used to choose between domestic and imported capital goods. Such prices do not include either the costs of the chronic trade deficits of medium- and large-sized countries that do not produce capital goods or the costs of the lack of capacity to innovate or adapt technologies associated with the capital goods industries. The relative absence of a machinery industry means that an increase in investment becomes an increase in imports. Since much of the multiplier effect on aggregate demand leaks abroad while the machinery imports increase productive capacity, the profitability of investments is reduced. Moreover, the relative absence of a capacity to adapt imported technologies means that when domestic demand arises for a new product which has just appeared in developed country markets (the lag is typically short because those with discretionary purchasing power in developing countries tend to imitate the consumption patterns of developed countries), it is met by imports. These imports, being in the initial fast-growth-of-demand stage of the product cycle, tend to grow faster than manufactured exports, which tend to be in a later and slower-growing stage; and tend to grow all the faster than agricultural exports, most of which face very slow growth of demand. These elements cause a tendency toward chronic trade imbalances. The state therefore has a role in directing the sectoral allocation of investment so as to reduce imports and thereby increase the multiplier effect, check the tendency toward trade imbalances, and enhance domestic capacity to adapt imported technologies. Furthermore, over the 1970s and 1980s the development of capital goods and microelectronics have become intertwined, many innovations taking place in the course of applying microelectronics to capital goods. National policy has to playa still bigger role to push domestic producers into the new microelectronics technology than was true of the older electromechanical technology; for entry barriers are typically much bigger (as in the case of numerically controlled machine tools compared to conventional machine tools, for exampie). But microelectronics constitutes more than just a radical change of technology. It constitutes a change in technoeconomic paradigm, a set of changes which not only leads to new industries and products but pervades almost every branch of the economy (Freeman and Perez 1988). The last such change of paradigm was in the 1930s and 1940s, associated with the utilization of cheap energy (especially oil). The change underway since the 1970s is associated See Rodrik (1987) for a theoretical and empirical study of the implications of imperfect competition and economies of scale for trade policy. 358 CHAPTER II with microelectronics. In the early stages of diffusion of this paradigm "windows of opportunity" open for countries which have already built up a certain level of infrastructure, industry, and technical knowledge but which have not made heavy physical and organizational investments in the now superseded technology system. Newcomers which meet these difficult threshold conditions can catch up with the previous leaders by early entry into the new information technologies (Perez and Soete 1988). A necessary but not sufficient condition is substantial amounts of government assistance in concentrating investment and socializing risk. Much of the assistance may have to be in a leadership rather than a followership mode, because private entrepreneurs are unwilling to take enough risk, quickly enough, to assure that the new opportunities are taken. Of course, whether national welfare is actually enhanced by the assistance depends on the wisdom of the choice of targeted industries and on the effectiveness of implementation. And industrial policies are only one set of factors on which the competitive success of an industry, or of the country's whole industrial sector, depends. In the total pattern of causation they are much less important than the capabilities of the country's private companies. But those capabilities are themselves able to be augmented by assistance from the public sector. We now consider several guidelines about how governments can assist industries to be more competitive. No attempt is made at completeness. The subjects chosen for discussion meet two criteria: they loom large in East Asian policies, and they involve some skewing of market processes by changes in prices or opportunities for exchange. Prescription 2: Use protection to help create an internationally competitive set of industries. Two of the things which economists disagree least about are that protection, whether for restraining the demand for imports or for promoting domestic industries, is always second-best, and that quantitative restrictions (QRS) are always inferior to tariffs. When unrestrained demand for imports leads to balance-of-payment difficulties, the solution is devaluation plus restrictive expenditure (fiscal and monetary) policy. If for some reason it is deemed necessary to promote specific industries, credit subsidies should be used (Corden 1974). These prescriptions are backed by an impressive body of theoretical reasoning. But once one moves beyond a concern for what is logically consistent with the theoretical system of neoclassical economics, they are not compelling guides to decision-making in the real world. As regards devaluation, the first problem is that experts often disagree by large margins as to what the "desirable" exchange rate should be, not only in developing countries but in industrialized countries as well (notably the United States). Second, even where ~,experts agree that the exchange rate is substantially overvalued, "markets" often seem to be poor at correcting the imbalances. Third, the policy instru- LESSONS FROM EAST ASIA 359 ment is the nominal exchange rate, but there may be no close connection between changes in the nominal rate and changes in the real rate except in the very short run; and it is the real exchange rate which counts for resource allocation. Fourth, the neoclassical argument recognizes no limits on how far the exchange rate can be made to fall. But a fall in the real exchange rate means a fall in the price of non internationally tradable goods and services relative to the price of tradables. The most important nontradable is labor, so a fall in the real exchange rate tends in practice to cause a fall in the real wage. Workers may revolt. More generally, inflexibilities of import-dependent production processes and consumption patterns may mean that the needed fall in the exchange rate is not possible without disruption of production, inflation, social unrest, and political conflict, fear of which may induce a well-meaning government to find other methods of maintaining external balance. The argument to replace protection with credit subsidies as a means of assisting particular industries is also not compelling. First, there can be no presumption that the subsidies needed for infant industries to compete equally against foreign suppliers would match the finance available. Unless a close connection is assumed between the revenue-raising capacity of government and the amount of subsidies needed, the subsidies may exceed the capacity. Second, the advantages of subsidies cannot be presumed to outweigh the "distortionary" effects of raising revenue through the existing tax system. Third, protection through tariffs raises revenue in an administratively simple way, compared to the difficulties of raising revenue through direct taxes; and is probably no more difficult to administer effectively than a subsidy program (Luedde-Neurath 1986). Fourth, subsidies are generally a more visible means of transferring resources and may therefore generate more political conflict than protection, which transfers resources more invisibly. (Whether this is desirable depends on whether the pattern of protection makes national sense.) Finally, insofar as changes in subsidies are more contested politically than changes in protection, subsidies are unlikely to be changed enough to buffer short-term external fiuctations. There are indeed many cases where protection has not had any noticeable innovation- or investment-enhancing effect (e. g., India). This reflects the failure to integrate protection with a wider industrial policy, or link it to export performance, or make the quid pro quo conditions credible, or to maintain macroeconomic stability. If protected producers know that in the foreseeable future protection will be much reduced or that government will pressure them to enter export markets, then protection may give them breathing space in which to undertake the necessary investment and innovation. They can use higher than normal profits in the domestic market to subsidize their entry into export markets, practicing discriminatory pricing. The same effect may be induced by awarding import licenses for targeted products only on evidence that the product could not be obtained from domestic producers within some rea- 360 CHAPTER 11 sonable margin of the import price. Such an "approval" mechanism or "law of similars" at least forces would-be importers to obtain full information about domestic supply capability. It also helps to stabilize demand for domestic producers of import substitutes, thereby lowering their risk and encouraging them to invest enough for economies of scale. But at the same time, the price criterion means that international competitive pressures are brought to bear on domestic producers, though in a modulated way. There is, of course, a tension between stimulating demand for nationally made products by protection (or domestic content requirements, or government procurement) and stimulating the international competitiveness of users of those nationally made products. Supply-side measures of assistance to the domestic· producers can help to reduce the conflict. But in any case, it is important that exporters be exempt from most import restrictions, the exemption being greater the bigger the price and quality differential between imports and ! domestic substitutes. The government can, however, use its import-restricting ability to encourage users of imported inputs to negotiate with local suppliers for upgraded production or lower prices in return for guaranteed sales. Repeated across many products, this mechanism can nudge the production structure of the country upwards. Notice that the mechanism uses QRS rather than tariffs. QRS (and domestic content requirements) have merit when the acquisition of technological capacity and subsequent adaptive innovation depend on extensive interaction between users and suppliers (Lundvall 1988; Pack and Westphal 1986). However, QRS have the costly consequence of amplifying the volatility of price signals, because with changes in domestic prices the tariff-equivalent of any given QR also changes. But where macroeconomic stability prevails, as in East Asia, this familiar cost of QRS is much less significant. The East Asian experience supports the argument that QRS have lower costs in stable than in unstable macroeconomic conditions. The desirable degree of import liberalization is much affected by country size. For most small countries-most of the time and in most industries-a relatively liberal trade regime is a necessity because of the lack of domestic economies of scale. Bigger countries have a wider latitude of choice. In general, the wider the latitude of choice, the more the overall degree of trade freedom should emerge as the result of calculations of the appropriateness of lowering protection to particular industries, bearing in mind that domestic competition can substitute for foreign competition, as in Japan, and that domestic competition, even between oligopolists, can be stimulated by government policies. Taiwan and Korea show how liberal trade policies in some industries can be combined with import substitution policies for other industries, resulting in different incentives to different industries. They also show how a rapidly industrializing country can soften pressures from its trading partners LESSONS FROM EAST ASIA 361 to open its markets or face retaliation, by a judicious combination of camouflage, statements of intent, and real liberalization. Some developing countries, particularly in sub-Saharan Africa, are unable to earn enough foreign exchange to cover import demands at any politically viable exchange rate, because of the limited supply of internationally saleable products. Here it makes no sense to talk of protection only as a temporary measure to assist the emergence of infants able within five to ten years to compete against international competition with no protection. Protection has to be seen as a part of longer-term measures to gain experience of industry and large-scale organization. In its absence resources may remain largely unemployed or confined to very small-scale production. The trick is to use such longer-term protection in a way which does not eliminate all competitive pressures. In short, import protection is, as neoclassical theory says, a powerful tool. Like any powerful tool it can be badly used, producing a trade regime full of inconsistencies. But that is not the end of the story. The East Asian evidencewhose challenge to mainstream trade theory has produced little more than an elliptical pirouette by way of response-suggests that protection can also be used in combination with other measures to foster the creation of internationally competitive industries. 9 Where such industries are not on the horizon, protection can at least help to begin the process of acquiring the capital needed to make new capital, the knowledge needed to absorb new knowledge, the skills needed to acquire new skills, and the level of ~evelopment needed to create the infrastructure and agglomeration economies that make further development possible. All this suggests an important analytical point, that the international trade literature is wrong in identifying some policy instruments as unambiguously better or worse than others without regard to the way those instruments are administered. QRS administered in a conditional way are not the same as unconditional QRS. Protection may be administered more easily than subsidies, and so more reliably achieve the intended effects. Prescription 3: If the wider strategy calls for heavy reliance on trade, give high priority to export promotion policies. East Asia's fast growth and equitable distribution was undoubtedly helped by the rapid growth of exports. Ex9 It is surprising that Jagdish Bhagwati, one of the most theoretically creative proponents of nearly free trade regimes, does not attempt to grapple with the effects of Korea's, Taiwan's, and pre-1970 Japan's protection system on their industrial growth. In writing a book called Protectionism (1988), or in addressing the question of "Is free trade passe after all?" (1989), he might have been expected seriously to address the empirical association between substantial and selective protection in East Asia and superior industrial perfoJIDance, this being critical evidence and not just another set of cases. But his oversight does keep him consistent with his own law of economic miracles: "Economic miracles [Taiwan et al.l are a public good; each economist sees in them a vindication of his pet theories" (1988:99). For a descriptive account of the East Asian trade regimes, see Wade 1988b. 362 CHAPTER 11 ports faced less of a demand constraint than output in general; they provided a channel for technical assistance from buyers; and they gave more scope for labor utilization than the manufacturing sector as a whole or the existing import-substituting industries in particular. However, export growth is not the only important reason for fast and equitable growth, and other countries with different natural resource endowments and larger economic size may be able to achieve "good" growth and distribution with less reliance on exports (Adelman 1984; Sen 1981). Indeed, they may not have much choice in the matter, because Western countries will probably intensify protection to avoid big (especially China-scale) increases in competing imports from developing countries. Where, nevertheless, heavy reliance is to be placed on trade, the government must recognize that successful exporting of manufactured goods to richer countries is not just a matter of getting the exchange rate right and keeping labor cheap, even in the absence of protection. 10 This is because many kinds of manufactured exports to richer countries are only saleable as complete packages meeting all buyer specifications, including packaging, labeling, colors, raw materials, finishes, and technical specifications. Costs rule out the option of importing an incomplete or defective package and correcting the defects in a subsequent stage of manufacturing. Thus, marketing, transmission of information, and quality control tum out to be key activities for export success. Buyers can supply some of these services; but especially because of the externalities the government also has an important role. The government can arrange for information about foreign markets and about domestic suppliers to be easily and freely available; it can directly help the promotion of some products (e.g., through trade fairs); and it can help to curb the tendency of firms without brandnames to compete by producing shoddy goods, spoiling the country's reputation for other producers. Very importantly, the government can also inspire producers to seek out export markets as a normal part of their operations (Keesing 1988). All this holds even in the absence of protection. If the economy is protected, cheap labor and a proexport exchange rate are still less likely to be sufficient. Without quick and automatic access to imported inputs at world market prices, free of customs duty, quantitative restrictions, and indirect taxes, would-be exporters will be handicapped in world markets by being forced to pay more than competitors for the same inputs or by being forced to use inferior domestic substitutes. Since manufactured exports from developing countries are normally sold in intensely competitive markets, producers in a country without a scheme for duty drawback and relaxation of quantitative restrictions are unlikely to obtain big export orders. Buyers for industrialized countries will sim10 There is disagreement between economists on whether conventional neoclassical trade theory does or does not support export subsidies. See Bhagwati 1988:95, n.ll. LESSONS FROM EAST ASIA 363 ply pass them by. However, even once export sales have near-free trade conditions producers of manufactured goods may still face net incentives to sel! on the protected domestic market, and exports may still be uncompetitive because the costs of nontradable inputs (especially labor) are raised by demand for those same inputs from the protected and hence larger-than-otherwise domestically oriented industries. An export sUbsidy scheme may be needed to make export sales as attractive as domestic market-related sales. Combining this discussion of export promotion with the preceding discussion of import protection, we see how misleading it is to present import substitution and export promotion as mutually exclusive strategies, as in Anne Krueger's claim that "export promotion outperforms import substitution" (1981:5). They are mutually exclusive only if defined to refer to the overall balance of incentives between domestic and foreign sale. But at the individual industry level, import-substituting incentives and export-promoting incentives can be complementary. On the one hand, development of the supply side through import substitution may be a prerequisite for the demand-side growth of exports. On the other hand, export growth can be helpful for the further development of industries that are nearing the limits of import substitution. Likewise, export promotion in one industry can complement import substitution in another by providing foreign exchange, for example. At anyone time export promotion and import substitution should coexist, reflecting the different development stages of different industries. We also see how misleading is the common assumption that policy-induced neutrality (as when export incentives "counteract" the effects of import controls) is equivalent to free trade. It is not clear how the many kinds of incentives for export- and import-substituting industries can be commensurated (effective protection rates are hardly adequate). It is fairly clear that the structure ofrelative prices at the time when "neutrality" is achieved reflects the prior rounds of intervention, and differs from that of an economy with untrammeled prices and exchanges throughout. Therefore we cannot presume that relative prices and resource allocations would be unchaflged if the entire array of incentives and protection were eliminated at a stroke. Prescription 4: Welcome multinational companies. but direct them toward exports. Multinationals are the primary source of knowledge about technology and production and an important source of knowledge about marketing. No country is going to get far in knowledge-intensive manufacturing and services without their help. Hence the government of a newly industrializing country should establish attractive policies for foreign capital, whether as subsidiaries, joint ventures, or licensors. However, foreign firms should be under pressure to direct their sales toward exports and their input purchases toward local suppliers. For if their products dominate the domestic market the developmental consequences of the protection system may well be worse than if domestically owned firms dominate the domestic market. 364 CHAPTER 11 First, with multinationals restricted in terms of their dominance of the domestic market, government efforts to promote the growth and restructuring of domestic production do not have to go through the multinationals, whose objectives will not wholly coincide with the development of national production capability. The government is able to use investible funds according to priorities designed to further integrate the domestic market, through having more influence over the firms that produce for the domestic market than if those firms were predominantly multinationals. (The multinationals should also be under foreign exchange controls, for if they are free to move funds in and out of the country they may start to function as de facto bankers for their domestic customers and suppliers, eroding the government's own credit policy.) Second, multinationals operating in the domestic market tend to follow marketing strategies that have little to do with average incomes or traditional consumer behavior, thereby accentuating income inequalities. By limiting their access to the domestic market the diversification of goods made available to consumers can be a gradual process geared to the population's purchasing power. In particular, the most modern of consumer goods, produced chiefly by multinationals, should be restricted to exports until basic needs in food and clothing have been met-as the Korean government did under Park (Ikonicoff 1985). Third, export requirements on multinationals not only generate foreign exchange, but also, less obviously, insure that the companies adopt an internationally competitive technology, rather than one which is viable only on the protected domestic market. However, export requirements may be eased in return for higher domestic content. The government should attempt to tie the magnitude of direct foreign investment incentives to either export performance or local content performance. Application of these principles has to recognize that, given the increasing •• footlooseness" of much industrial production, the balance between the costs a country imposes on enterprises and the facilities it provides has to compare with other countries' If requirements on foreign firms have to be relaxed in order to attract and keep them, it is important to compensate with discriminatory state support for competing domestically owned firms to prevent them from being back washed out of existence. Prescription 5: Promote a bank-basedfinancial system under close government control. A closely regulated bank-based financial system has several advantages in industrializing country conditions. 11 First, it permits higher in- " For a discussion of "bank-oriented" financial systems (such as Japan and Germany) and "market bank" systems (United States and Great Britain), see Mayer 1987; Zysman 1983. One of Mayer's themes is that "the separation between investment and finance, which is the basis of most existing models, is untenable" (p. iii). My discussion tries to bridge this separation by showing how a certain type of financial system may affect the "real" economy in developing ~.ountry conditions. LESSONS FROM EAST ASIA 365 vestment than would be possible if investment depended on the growth of firms' own profits or on the inevitably slow development of securities markets. In a capital market-based system, the decentralized preferences of the public largely determine the allocation of potential savings into productive investment, financial speculation, or consumption. In a bank-based system, in which enterprises depend heavily on banks for finance and less on a broad public of shareholders, the long-term growth preferences of government officials and/or bank executives have more weight. Investment decisions are hence more insulated from the preferences of the public. Credit can be more cheaply provided for productive investment, in the context of a long-term approach to the economy's investment activity. In a capital market-based system, on the other hand, government attempts to stimulate investment by tax cuts and deregulation may have only a modest effect on investment, as in the economic reforms of President Reagan and Prime Minister Thatcher. Second, a bank-based system encourages more rapid sectoral mobility and permits the government to guide that mobility insofar as it can influence the banks. Even small changes in the discount rate or in concessional credit supply between sectors can have a significant effect on resource allocation (provided the use of credit is controlled enough to prevent unlimited fungibility), because the effect of such changes on firms' cash flow position is greater than where firms have smaller debt/equity ratios. Where the government is trying to foster key sectors, a bank-based financial system gives it a powerful mechanism for inducing firms to enter sectors they otherwise would not. Where, on the other hand, capital is allocated mainly in decentralized markets, the government's ability to extend a visible and vigorous hand in the functioning of the industrial economy is limited, because firms are less susceptible to state influence (Zysman 1983). Third, a bank-based system can help to avoid the bias toward short-term company decision-making inherent in a stock market system. The creditor needs the borrowing company to do well: it is concerned about the company's market share and ability to repay loans over the long term, which depend on how well the company is developing new products, controlling costs and quality, and so on. So these become the criteria which managers are concerned with, rather than stock market quotations (Johnson 1986; Dore 1985). The fourth advantage is more directly political. Industrial strategy requires a political base. Control over the financial system, and hence over highly leveraged firms, can be used to build up the coalitions needed to support the government's objectives-thus helping to implement the industrial strategy. Firms are dissuaded from opposing the government by the knowledge that opponents may find credit difficult to obtain. Of course, such a practice is easily abused. If it becomes common to allocate credit for "loyalty" rather than for economic performance or potential the legitimacy of the administra- 366 CHAPTER 11 rive discretion will be impugned. Sparing but well-publicized use may reap the political gains without the legitimacy costs. These are four potential advantages of a bank-based, administered-price financial system. However, such a system contains certain imperatives for government action which have far-reaching implications for the government's role in the economy. The first is that the government must help to ease the downside risk of debtfinancing. Higher deposit interest rates can increase the flow of financial savings; but at the new rates the private sector may not be prepared to borrow the savings unless the government intervenes to socialize some of the prospective private losses. Even if in the short run the savings are translated into loans, the higher savings and investment made possible by the higher rates may not be sustainable in the longer run without measures to spread risk. This is because highly indebted (or leveraged) firms are vulnerable to decline in current earnings to below the levels required by debt repayment, repayments on debt being fixed (whereas payments on equity are a share of profits). With firms vulnerable in this way, so are the banks which carry the "nonperforming" loans. So where debt/equity ratios are high, there is an ever-present danger of financial instability in the economy:12 bankruptcies, withdrawal of savings, a fall in real investment, and slower growth. To ease such dangers, firms are likely to borrow less and banks to lend less than if the government were to underwrite some of the risks to which lenders and high debt/equity producers are exposed. If the government does bear some of the risk of private losses, the supply and demand for loanable funds will be greater, so investment, technical change, and hence growth can be higher. The need to socialize risk applies especially in the case of highly correlated risks, to which most firms in major sectors are exposed. So it applies especially to interest rate changes, or major recession, or changes in major export markets, or political risks. Therefore the impetus for government to shoulder some of the risks of investment and saving in an economy with high debt/ equity ratios is especially strong in countries which are trade-dependent and! or under external threat (like Taiwan, Korea, and Japan). The impetus is reinforced in industries where both entry and exit take a long time. This impetus then leads the government to provide a battery of ways to reduce the risks of financial instability-not only in the form of deposit insurance and lender-of-Iast-resort facilities, but also in the form of subsidies to banks imperiled by loan losses, product and credit subsidies to firms in financial difficulties, banks' share-holding in companies, government share-holding in banks and in lumpy projects, and even government ownership of banks. 11 The implications of high debt/equity ratios also depend on profitability at the firm level. In an economy where profitability is higher and more secure the danger of overall financial instability is less. The same applies to the implications of high debt/equity ratios for the relationship between banks and firms. I am indebted to discussions with Frank Veneroso on these matters. LESSONS FROM EAST ASIA 367 Government can also, of course, control interest rates and exchange rates to dampen firms' exposure to market fluctuations in these two important sources of correlated risk. The second imperative is for the supplier of credit to become involved with company management. The supplier of credit may for this purpose be the government (Korea), or the banks (Germany), or some of both (Japan). In any case, the reason for involvement with management is that the creditor cannot simply withdraw when a company runs into difficulties by selling the securities in the secondary capital market; the secondary capital markets are too thin. Given that the "exit" option of the capital market is not available, the alternative is the "voice" option, to try to restructure company management so as to make it more competitive and to take the long-term view (Hirschman 1970). Nevertheless the government and/or the banks must, third, develop an institutional capaCity to discriminate between responsible and irresponsible borrowing, and to penalize the latter. Firms which borrow without due commercial caution and run into trouble must not expect the government or the banks to continue to bail them out (the so-called moral hazard problem). The government must also develop mechanisms of bank supervision to curb the tendency for banks faced with big loan losses to conceal them in the "performing" part of the balance sheet while making even riskier loans in the hope of getting back enough to offset the losses. This is the path that turns good bankers into bad ones, solvent banks into insolvent ones. Once market signals are blunted by administered pricing and socialized risk, the government must, fourth, create a central guidance agency capable of supplementing market signals by its own signals as to which sectors will be most profitable--but in a way which allows plenty of scope for private pursuit of opportunities not seen by the guidance agency. Finally, the government must maintain a cleavage between the domestic economy and the international economy with respect to financial flows. Without control of these flows, with firms free to borrow as they wish on international markets and with foreign banks free to make domestic loans according to their own criteria, the government's own control over the money supply and cost of capital to domestic borrowers is weakened, as is its ability to guide sectoral allocation. Speculative inflows seeking exchange rate gains can precipitate accelerating movements in exchange rates, with damaging consequences for the real economy. Uncontrolled outflows can leave the economy vulnerable to an investment collapse and make it difficult for government to arrange a sharing of the burden of adjustment to external shocks between the owners of capital and others; "the others" are likely to be made to take the burden, with political unrest, repression, and interrupted growth as the likely result. More generally, foreign exchange controls are needed to intensify the cycle of investment and reinvestment within the national territory, with outflows only where they can be shown to meet national economic priorities. 368 CHAPTER 11 Otherwise domestic interest rates come to be determined in large part by V. S. interest rates, and therefore make the economy sUbject to the kind of macroeconomic mismanagement of the V.S. economy seen during the 1980s. Although presented here as just one in a list of several requirements, this cleavage between the domestic financial system and the international financial system is a prior condition for all the others. If free markets for foreign exchange and other financial assets were clearly efficient one might hesitate to recommend such a controlled system. But "belief in the efficiency of the foreign exchange market is a matter of pure faith; there is not a shred of positive evidence that the market is efficient," concludes Paul Krugman. Similarly for bonds and stocks: "there is no positive evidence in favor of efficient markets" (1989:65). On the other hand, the disadvantages of controis can be partly offset by allowing an unregulated curb market to operate in the interstices (Biggs 1988)-and perhaps by tying a sizable portion of subsidized bank credit to export performance. Joseph Stiglitz's new work on developing country financial systems provides a cogent rationale for- policies broadly in line with those recommended here. He concludes, "the LDCS should not set their sights on imitating the capital markets of the most developed countries, but rather should adapt themselves to the reality that capital markets will most likely, if not necessarily, work poorly within their country. Adopting this view suggests a major redirection of several policies which have been widely adopted within the third world" (1989:56). He emphasizes more than I do, however, the difficulties to improving the incentives on government and large public credit institutions to do a good job in selecting and monitoring loans. Prescription 6: Carry out trade and financiaL liberalization graduaLLy, in line with a certain sequence of steps. Many neoclassical analysts urge largescale and quick liberalization, to get a whole package of reforms in place before opposition builds up. And many urge that comprehensive import liberalization should be carried out before export earnings increase, so as to flush away the inefficiencies generated by protective barriers and enable a subsequently better response to export demand (Krueger 1978; Snape 1988; La! 1983; Michaely 1988). By contrast, the East Asian experience is consistent with a prescription for more gradual change and a different sequence. It suggests the following: (1) macroeconomic stabilization should come before trade liberalization; (2) substantial external financial assistance greatly eases the transition from stabilization to liberalization; (3) liberalization of imports of export inputs should come before deprotective competition-providing import liberalization; (4) import liberalization of the latter type is not a prior condition for successful exporting; it should follow the growth of exports; (5) successful exporting requires a large promotional role for public agencies; (6) gradual trade liberalizations can be sustained; and (7) financial liberalization should come late in the queue, after a substantial measure of import liberalization (Sachs 1987; Helleiner 1988; Edwards 1985; Wade 1988b). LESSONS FROM EAST ASIA 369 With reference to financial liberalization, our knowledge of its effects in segmented and imperfect capital markets is thin. Modeling the connections between the regulated financial markets and the curb market in Korea has generated "unconventional" results from orthodox monetary and interest rate policies. Higher (regulated) interest rates and monetary restraint led to a serious slowdown in investment and growth, the effects of which exceeded any positive effects for household savings (van Wijnbergen 1983). Painful experience with rapid and far-reaching financial liberalization in the Southern Cone countries of Latin America has bred a new respect for government supervision and control of the domestic financial system, and caution about lifting external capital markets controls. These results support the gradual approach to domestic and external financial liberalization adopted in East Asia (Helleiner 1988; Mayer 1987; Krugman 1989). More generally, the gradualness of economic liberalizations in all three countries further undermines the view that if only the government of a developing country shrinks from influencing prices or exchanges, it too can expect much improved economic performance. The East Asian liberalizations were gauged to the competitiveness of domestic industry, which was itself promoted by preceding and simultaneous industrial policies. Without them, letting prices work would have been like pushing on a piece of string. There is also a political case for gradualism, which should be weighed against the political argument for quick and deep liberalization to preempt opposition. Liberalization typically involves changes-removal of food subsidies, for example-whose costs affect the general public widely and directly but whose benefits are more concentrated and slower to appear. If several such policies are introduced at once and entail major rather than incremental change, they can be expected to generate opposition from many interest groups, including some whose support is important for the regime's survival. They therefore carry high stakes for the incumbent political leaders, all the more so when the government is insecurely legitimated. Gradual introduction of the reforms, with some sequencing and camouflaging of who gets the costs and benefits, can provide less fuel for opposition organization and thereby make sustained implementation more likely. In the general case, this seems as plausible a scenario as that of the neoclassicals. 13 In considering issues of liberalization we need to make a distinction between shrinkage of the public sector and reduced state capacity to manage the J3 Indira Ghandi' s post· 1980 government carried out significant economic liberalization without drawing sharp political reaction partly because she made the changes look like marginal, "technical" ones. Her Son and successor, by boldly proclaiming a decisive shift of development strategy to a liberal model, aroused such widespread opposition from groups on whom the Congress depended for votes as to force a drastic slowdown in the liberalization. However, in the early period Rajiv Ghandi's govemment did successfully introduce a number of liberalizing policies-those which "were brought about quietly, without much fanfare, as seemingly technical changes in a piecemeal fashion" (Kohli 1989:314). 370 CHAPTER II market. The size of the public sector, in terms of employment, share of GOP, and other such measures, is not closely associated with state capacity to manage the market. 14 Indeed, shrinkage of the fonner-which would be called "liberalization" in the gross way the term is generally used-may help to expand the latter; and the latter is what counts. The argument for economic liberalization-whether in trade, finance, or other spheres-also needs to address the question of what kinds of private sector groups will gain from the change. It cannot be assumed that they will wish to be entrepreneurial investors rather than lUXUry consumers. Nor can it be assumed that they will wish to place limits on the arbitrary actions of the state and discipline the state to provide effective services. Liberalization may lead to the capture of economic power by less accountable cliques around the power-holders, Marcos-style. The analytical dichotomy between "state" and "economy" can lead us to overlook the point that the same people or groups may have feet planted finnly on both sides of the divide, in which case a shrinkage of the state and expansion of the private sector may further remove economic power fonnerly in the hands of the state from some degree of accountability. It may further erode a "center" -a cohesive organizational structure-where collective interests can be articulated and followed. These are six broad economic prescriptions supported by the experience of Taiwan, Korea, and Japan. But we must note another lesson to do with differences rather then similarities. While the three East Asian states all governed the market, they used somewhat different methods for doing so. Taiwan used large upstream public enterprises and selected foreign finns to provide "unbalanced" pushes in certain sectors, arms-length incentives to steer the response of myriad small downstream firms, and stable prices and real effective exchange rate. Korea used huge private business groups as the spearheads, steering them with massive credit subsidies and more direct cajoling (recently switching to more of a negotiation mode). It obtained more of its technology under license than through direct foreign investment, and sacrificed some macroeconomic stability for faster industrial transformation. Japan, which already had huge private business groups in place in the 1930s, pioneered the route that Korea was later to follow, except that consultative decision-making procedures linking government and business were in place from much earlier on. So there is more than one way to govern the market effectively. IMPROVING STATE EFFECTIVENESS My argument is that a necessary but not sufficient condition for more rapid industrialization is state deployment of a range of industrial promotion policies, including ones to intensify the growth of selected industries within the national territory. This is not to say that effectiveness increases with the sheer 14 See chapter I, n.l2. LESSONS FROM EAST ASIA 371 amount of intervention, nor that it increases the more the state imposes its will on society, ignoring other groups. State effectiveness is a function ofthe range of options, given by the number and force of policy instruments, and the flexibility with which those policy instruments are used. Flexibility means that the capacity to intervene, as given by the number and force of policy instruments, is used to varying degrees, more in some industries than in others at anyone time, and more in one industry at some times than at others, always with an eye on the costs of interventions in political as well as economic terms. In particular, high effectiveness requires the flexibility to withdraw assistance from industries as they become internationally competitive, and the ability not to intervene in some industries at all in the interests of concentrating assistance and limiting costs. Behind these proximate determinants of effectiveness are others of a more organizational and political kind. First is the competence and coherence of the central economic bureaucracy. Second is the degree to which political authority is institutionalized. Third is the connection between the central bureaucracy and other major economic interests, especially the owners and managers of capital. We now consider four more prescriptions to do with these organizational and political determinants of state effectiveness. Again, the prescriptions are rooted in what East Asian governments actually did. Prescription 7: Establish a "pilot agency" or "economic general staff" within the central bureaucracy whose policy heartland is the industrial and trade profile of the economy and its future growth path. For an industrial policy to be effective one or two agencies should steer the formulation and application of the policy instruments. Taiwan, Korea, and Japan have all used the "few agencies" model, in contrast to the "many agencies" model of the United States and Great Britain. The pilot agency should have a fairly small staff (Japan's MlTI had only about two thousand in the 1960s). It should be in a position to recruit from among the best and the brightest. Once a competitively selected economic bureaucracy acquires a reputation for attracting the best and brightest, the system develops a momentum of its own. It continu€1s to attract such people (even at much lower salaries than the private sector) because selection is the stamp of outstanding talent. Its personnel need to be motivated by the belief that what they are doing promotes the national welfare. A sense of national mission, combined with a meritocratically based esprit de corps, can motivate the central bureaucracy to use its powers in line with national goals, prOViding a substitute for the motivational force of profits in private firms. A vigorous national press, free to criticize the economic bureaucracy (even if not the political leaders) can help to keep its actions in line with the public interest. The conflict between life-time employment and up-to-date technical and managerial knowledge can be moderated by using parabureaucratic task forces to complement in-house capabilities. The more the government intends to intervene in a leadership rather than a followership mode, the more important are the staffing, motivation, authority, and responsibilities of the pilot agency. 372 CHAPTER II The pilot agency should be concerned with fonnulating operational goals, such as diversification of industries, diversification of markets, reduced dependence on raw material imports, and greater employment in certain industries; and with analyzing how various policies affect these goals. It should think of itself as a strategic oligopolist, scrutinizing the actions of rival governments and taking account of those actions and reactions in framing its policies for investment, trade, and technology. It should have some power of implementation rather than devolve all concern for implementation to the ministries. It should focus on certain key industries at anyone time, more or less ignoring the rest; but should situate policies for these industries within an analysis of the whole economy, bringing multiple policy instruments to bear on them. This can be done without detailed quantitative targets for investment and output for particular industries, which are only likely to distract from the more substantive business of fonnulating the broad vision of the appropriate directions for growth and choosing the specific industries to be promoted. Multibusiness, multinational corporations undertake a broadly similar type of strategic (rather than comprehensive) planning as a matter of course, and their techniques can, with modifications, be extended to multi-industry, multi market countries. In addition to the organizational factors considered above, the effectiveness of such an agency is related to the decision criteria it uses. One of the great merits of using export perfonnance is its simplicity and clear connection to competitiveness. It can be used to make a first judgment on assistance to firms or industries; those that are doing well in export markets will be treated more favorably than those which are not, other things being equal. When a country is pursuing a domestic market-based strategy, particularly one which is led by agriculture, simple and sensible decision criteria are more difficult to find (Pack and Westphal 1986). Above all, one must deemphasize criteria that, for ease of measurement, focus on inputs, not on outputs. The activities of such an agency are likely to be uncongenial to economists trained to believe that targeting by officials will generally fail. That is one good reason for curbing the number and influence of economists in the industrial policy-making process, as was done in Japan, Korea, and Taiwan. The other reason is that neoclassical economics has little to say about the issues raised by the present shift of technoeconomic paradigm, of how to exploit the opportunities opened by the new information technologies (Dosi, et al. 1988). Of course, in Cuba, North Korea, the Soviet Union, China, and other centrally planned economies, the more neoclassically trained economists the better. Prescription 8: Develop effective institutions of political authority before the system is democratized. The class structure of many developing countries implies a cruel choice between faster economic development and well-defended civil and political rights. Power and wealth are often concentrated in groups engaged in socially unproductive activities (including renting out of LESSONS FROM EAST ASIA 373 land, money-lending, exploitation of bureaucratic or military office}. Groups based on industry must grow on the margins of power, perhaps as part of a coalition of sections of the state bureaucracy and sections of the military. Often the rise of this" growth coalition" is attended by conflict as it tries to displace groups with real coercive power, capable of taking the law into their own hands. It may try to harness a popular political movement using nationalism or revolt against exploitative class relations as a rallying cry. Once it can influence state power, it has to use that influence to shape a social structure which is conducive to wealth accumulation through productive investment. Most likely this will require some curtailment of the political and civil rights of those who oppose the changes, and of the powers of democratically elected legislatures (Huntington 1968). The argument in favor of such a state is uncomfortable to those (including myself) who cherish the civil and political freedoms of North American and Western European countries-especially when "democracy," more than any other term in political discourse, now generates such a universal hurrah. "It has been difficult for me to comprehend how free people can choose tyranny for others," declares A. M. Rosenthal, former executive editor of the New York Times. "There is for me only one question that really matters about any government: does it allow political freedom? Does it permit its citizens to breathe and think and talk and write as free people?" (1986:23). He goes on to say that "the apologists for tyrannies argue that economic progress has to come first, and that democracy is really too expensive a way of achieving it. This is said in all seriousness, as if there were evidence that despotism somehow is more efficient than freedom. The exact opposite seems so obviously true " (p. 24). (Rosenthal includes the Park and Chun regimes of Korea as tyrannies, and would presumably so call the Nationalist government of Taiwan.) Rosenthal's Manichean assertions notwithstanding, the balance of argument and evidence seems to me to point the other way. People who live in societies where, for a whole century or more, they have been able to see "freedom of opinion" as about whether editors get sent to jailor not, find it hard to comprehend the priorities of people in societies where freedom is also restrained by fear of the assassin's bullet and fear of being thrown off one's land by trumped-Up suits and corrupt judges. In such societies the priority is to institutionalize a system of order before it is democratized-to move from a system where the press is controlled by people with wealth and private armies to one where it is controlled by the state, before reaching one where it is controlled by people with wealth but without private armies. The executive branch needs to be stronger than the legislative branch, to "rule" while the legislature "reigns." An elected legislature is likely to be directed less by a view of the common good than by, in Adam Smith's phrase, "the clamorous importunity of partial interests" (1775:438). A state in which 374 CHAPTER 11 the legislature is strong relative to the executive will find it difficult to hold the line against unbalanced increases in consumption at the expense of investment, and difficult to direct government assistance to uses which can meet a national interest test rather than a clamorous importunity test. Conversely, where the executive is relatively strong, there is a better chance that policies will not careen from side to side because of frequent turnover in power. This makes for a more stable business environment, facilitating longer-term corporate investment. And it helps the bureaucracy to oversee the operation of the economy. Historically, individual property rights, constitutional restraints .on the state, 'and the rise of the bourgeoisie occurred before the advent of mass democracy (Huntington 1984). Both Taiwan and Korea now provide support for the proposition that the stability of a new democracy depends upon the development of broad-gauged political institutions prior to the expansion of political participation. As we noted, they were in the middle of a rights ranking of middle-income countries in the early 1970s; Taiwan was in the same position in the early 1980s while Korea had fallen a little lower. As of the late 1980s, they are making a transition to stable and partially democratic systems, with the prospect of more democratization to come. Japan, of course, has had much better defended rights since World War H. But Japan's primary industrial revolution took place during the late nineteenth and early twentieth centuries, at which time its political and civil rights were quite restricted, its government distinctly authoritarian and oligarchic. For the past thirty years Japanese voters have gone to the polls with slimmer expectations that the result could be a change of government than in any other industrialized democracy; and the representatives whom they elect have had less influence on the major decisions affecting the national welfare than in any other industrialized democracy. Large cross-sectional studies of developing countries show that democratic regimes tend to grow more slowly than authoritarian ones. Robert Marsh, using a sample of ninety-eight developing countries and several development indicators for the period from 1960 to 1970, concludes that "political competition/democracy does have a significant effect on later rates of economic development; its influence is to retard the development rate rather than facilitate it" (1979:244). Erich Weede, using a different method 'and a different but also large sample of countries, concludes in part that' 'political democracy looks like a major barrier to economic growth in those countries where the state strongly interferes in the economy" (1983:312).15 But there is also probably J~ See also Adelman and Morris 1967: HUntington and Dominguez 1975. Dick (1974) finds the reverse relationship. For a useful discussion, see Kohli 1986. Interventionist governments in Weede's analysis had revenue equal to or greater than 20 percent of GDP in 1965 at all income levels. He finds thirty-four such cases (thirty-five for one regression). He uses Bollen'S (1980) classification of countries as democratic or authoritarian in 1965 (excluding the centrally planned economies). The methodological difficulties of establishing a causal connection between regime LESSONS FROM EAST ASIA 375 more variation in the performance of authoritarian regimes compared to democratic ones at similar per capita incomes. some having mediocre to calamitous results. Many authoritarian governments do not give high priority to economic development. being preoccupied with the task of excluding most of the population from power. 16 As per capita income approaches $4,000, the economic advantages of continued restrictions of civil and political rights---of continued consolidation of authority in the hands of the state-are probably offset by the costs of growing conflict, weak legitimacy, and overburdened state decision-making. This, at least, is a possible interpretation of the fact that over the 1980s there have been virtually no noncommunist countries at or above this income level without fairly competitive political systems. Singapore is an exception, and Taiwan would have been had it not initiated democratic reforms in 1986. Prescription 9: Develop corporatist institutions as or before the system is democratized. State effectiveness depends on the coherence of state policies, which is difficult to maintain when important parts of the state are beholden to sectoral, ethnic, or regional interests. Effectiveness is therefore a function of the degree of insulation (or "autonomy") from the surrounding social structure. Insulation is a function of, among other things (1) officials' dependence on the state for their income, not on interest groups; (2) officials' expertise, which gives them grounds for asserting their own preferences for state action against those of interest groups; and (3) the extent to which the nation faces a threat to "national interests" from other states, in response to which nons tate groups are likely to confer substantial autonomy on state officials. Even in a highly pluralist regime some bureaucratic insulation is conferred by these conditions. In an authoritarian regime the insulation can be much greater because the coalitional basis of the state is narrower. But insulation of central officials from pressures in the wider society also carries costs. It may erode feedback on economic conditions at the point of production and sale, and may remove a potentially strong basis for the formation of a consensus on the dominant factors influencing the course of the economy and the order of type and economic growth are fonnidable; see Hicks and Patterson (1989) and the papers to which they are responding. The gross categories of authoritarian and democratic regimes are too crude to capture some important aspects of civil and political rights. In particular, by focusing on the freedom or lack of freedom to choose rulers they miss the greater importance in the lives of ordinary people of having power to shape the rules which govern the immediate or local aspects of their lives. Perhaps some authoritarian regimes with strong national-level executives neverthe- -less grant or tolerate considerable latitude in the choice of local rules. 16 Dick (1974: table 1) provides some evidence of greater dispersion in the perfonnance of authoritarian regimes. One might relate the dispersion not only to goals (exclusion, economic development, etc.), but also to the type of military they are associated with-the swashbuckling, tribute-raising kind or the strategic-planning, military-engineering kind. In the decade following the mid-1970s most Latin American countries have swung from authoritarian to partially democratic regimes (GastiI1986). 376 CHAPTER II socioeconomic priorities. Even in the most centralized government, the course of the economy is influenced by the decisions of many separate agenciesdecisions about the exchange rate, tariffs, interest rates, availability of capital, wages, public expenditure, public borrowing, taxes, and more. Any attempt at comprehensive control from a single point would overstrain the government's information capacity and power. Without a roughly common view about decision premises, national economic policy decisions are unlikely to achieve the degree of coherence needed for their overall success. Moreover, when the thinking classes are wholly familiar with a certain range of ideas and assumptions about the longer-term future there is a base of legitimacy for government-sponsored measures with long-term payoffs. This base of legitimacy is especially important when some will be losers. But such a consensus is always fragile, vulnerable to the whims of important interest groups because of changes in circumstances, priorities, or evaluations. A corporatist structure·? to represent a limited set of major economic interest groups-and thereby also to channel the demands placed upon the state--<:an help to insure acceptance of negotiated outcomes. It can help to mediate the uneasy tension between unrestricted market forces and social peace, buffering the costs of economic adjustments to external market changes while allowing adjustments to occur (Ruggie 1982). It does so by facilitating reciprocity between big firms, government, and perhaps unions, \8 in which government help is made conditional on stipulated performance by the other parties; and by facilitating the use of state authority to steer groups away from hostile strategies that yield the worst outcomes in Prisoner's Dilemma situations. However, such arrangements are difficult to sustain where the idea of the public or national interest is used primarily as a cover for advancing private, class, or ethnic interests, where a core meaning shared by all the parties is lacking. The only evidence I know of on the economic performance of corporatist 17 In a corporatist structure. the state chaners a limited number of major economic interest groups, granting them a near monopoly of representation. In authoritarian regimes of the common garden variety, the leaders control and appeal to the people directly, without this intermediating structure of representation. And pluralist regimes differ from both in having a geographical rather than functional basis of representation, and in having' 'free trade" in interest group access to the state. I. Inclusion of labor is obviously desirable in principle. But note that if labor exclusion is part of a set of arrangements which generate high-speed growth, workers are protected to some extent by high labor demand. Labor exclusion also gives a government more room to maneuver when austerity comes, and that latitude can be used to restore fast growth more quickly (compare Mexico and Taiwan). But the more mature an economy becomes the less likely is economic growth to be sufficiently fast to meet the norm3.J aspirations for economic security. It is then of the first importance that the institutional arrangements made to provide security do not erode the pressure on people to work hard or convert security guarantees into obstacles to adjustment. The key is to build up labor commitment, so that effort is not based simply on an exchange of effort for pay. LESSONS FROM EAST ASIA 377 regimes comes from already wealthy and democratic countries, whose corporatism is of the "social corporatist" type. Here peak associatiol)s have more of a policy-initiating role than under' 'state corporatism." The evidence suggests that countries with social-corporatist arrangements (such as Austria, Norway, Switzerland, and Sweden) have enjoyed above-average incomes and economic growth, with lower inflation and unemployment, than pluralist countries. In particular, those countries with arrangements for centralized bargaining, politically dominant social democratic parties, and centralized unions, tended to weather the international economic crises of the 1970s better than others. Pluralist countries like the United States and Great Britain, which tolerate raw adversarial outcomes between economic interest groups, have been less successful in adjusting to these crises. 19 But this conclusion cannot readily be extended to cover corporatism in developing country conditions, about which there is little evidence. I expect that state corporatist regimes vary more in their economic perfonnance than pluralist ones. Salazar's Portugal and Per6n's Argentina, both state corporatist, had dismal records, while Taiwan and Korea have had exceptionally good ones (also Japan in its extreme state corporatism of 1940-41). The variation may be closely connected to rulers' objectives. In East Asia a shared sense of the external vulnerability of the nation has helped to concentrate the rulers' minds on perfonnance-enhancing measures as a means of their own survival. The Taiwan case shows an interesting variant of state corporatist arrangements. Fonnal industrial associations are weak, as we saw; but there is plenty of communication between government officials and finns of an infonnal, often dyadic kind. Industrial Development Bureau officials often spend several days a month visiting industry associations and finns in "their" industry, for example. 20 These and other means of keeping in touch with finn-level facts are vital if the government is to intervene to promote specific industries, whether as leader or follower. Prescription 10: Make piecemeal reforms even in soft states so as to create an institutional configuration better able to support a modest industrial policy. Suppose one agrees that governed market policies can improve on the economic results of free or simulated free market policies. Whether such gains are realized depends on the existence of certain organizational arrangements, as just indicated. When such arrangements are not already in existence, administrative and political reforms can create them. But these refonns cannot be simply willed into existence, and their effects take a long time to come 19 Wilensky 1981; Schmitter 1981; Schmidt 1982; Wilson 1985:110; cf. Zeigler 1988:chapters 3--4. I have not examined the Latin American evidence. U.S. and British economic performance has improved during the 1980s in terms of many aggregate indicators, though market share in many of their high-technology industries continues to erode. 20 For a South Korean example of little formal communication combined with abundant informal communication, see Wade 19&2a:54-56, 90. 378 CHAPTER II through. Given this, one might argue that-even accepting the potential gains of governed market policies-most states should move toward free market policies as fast as possible, on grounds that softer states can sustain these policies more easily than governed market ones. Paul Krugman, for one, argues that even though recent work in trade theory provides a rationale for an activist trade strategy in certain conditions (which are, however, unlikely to apply to developing countries), the gains may not be large and the strategy is difficult to implement because of its vulnerability to hijacking by special interest groups. Hence, he says, free trade rules are best for" a world whose politics are as imperfect as its markets" (1987: 143). Instead of political factors being blamed for the inability of governments to follow economically rational free trade policies, political factors are now brought in to say why free trade remains politically best even after it is shown to be not always economically best (Helleiner 1988). This is but a special case of the "practical optimality" of free markets. Even if free markets can be shown to be suboptimal according to some ideal allocation system, they are the method which produces the least inefficient resource allocation in practice, the argument runs. The alleged alternative, administrative allocation, will produce worse results because it is not subject to anybody's bottom-line constraint. There is something to be said for this argument. The image of the highsalaried official who knocks off to play golf at 4:30 leaving a pile of files on his desk and then, when he gets to your application three weeks later, casually says no, is more outrageous than that of the wheeler-dealer entrepreneur who is continually creating monopolies, rigging prices, making false advertising, and so on. But proponents of this argument fail to explain why, if vested interests are strong enough to defeat sensible selective interventions, they will not also be strong enough to distort markets and defeat free trade. Free trade is not self-enforcing. Vested interests seek to maintain the structures in which their interests are vested, rent-seekers seek to preserve the conditions that create rents. Karl Polanyi reminds us that Adam Smith's "natural propensity to truck and barter" had not sufficed to produce free markets in England. "The road to the free market was opened and kept open by an enormous increase in continuous, centrally organized and controlled interventionism" (1957:140). A passionate exponent of free trade agrees that "a courageous, ruthless and perhaps undemocratic government is required to ride roughshod over special interest groups" that stand in the way (Lal 1983:33). In practice, then, free trade may be no easier to sustain than sensibly managed trade. (One wonders, incidentally, what fraction of GDP the United States spends on competition policies, including not only the cost of the relevant government agencies but also the "chilling" effect of antitrust legislation in inhibiting agreements which might have been beneficial, as well as the transaction cost of the legions ofdawyers involved in every move.) LESSONS FROM EAST ASIA 379 Even in the context of a relatively soft state it should be possible to institute higher levels of effectiveness in limited parts of the state. It should be possible for the state's industrial policies effectively to do more than put a seal of approval on what the private sector would have done anyway. By bringing to the negotiations its own sense of the appropriate direction of the economy, and by paying for more rather than less of the costs of a new project which fits that direction, it can still impart a directional thrust. The organizational arrangement might take the form of a pilot agency which, though lacking the implementation power OfMITI or Taiwan's Industrial Development Bureau, still has some statutory power which makes it necessary to the making of policy, so that it can force its long-term perspective into pluralistic bargaining. One would need to insure that appointment to the senior positions of such an agency is by merit criteria; that the incumbents cannot be rotated in and out according to the sway of factions or money; that the standard operating procedures allow the agency to stick to a purpose, so as to deal with surprises and opposition in line with enduring goals; and that it has secure funding. Elsewhere in the economic bureaucracy effectiveness and responsiveness can be improved by several kinds of measures. There are the standard institution-building ones, including more financial resources, higher status, and bigger supervisory infrastructure. But more interesting are ones which attempt to introduce more market-like features into bureaucratic incentive systems, while remaining incremental in nature and not requiring sweeping changes in order to have some effect (Lamb 1982; Murray 1989). These include ways of changing internal incentives and options, such as decentralizing managerial accountability within bureaus or agencies, inducing competition between sections within a bureau or between agencies and parabureaucratic task forces, establishing more performance-based salary and promotion rules, creating a super cadre with better pay and more exacting performance standards than in the normal civil service, establishing public management accounting systems which record output and performance as well as costs, and developing a random-check performance auditing capability. by senior managers. These changes can be complemented by ones which seek to put more market-like pressure on the bureaucracy from outside, such as publicizing performance targets and rates of achievement, organizing would-be beneficiaries into industrial associations and turning over much of the high-cost "retailing" end of industrial service delivery to industry associations or private firms. Shifting from "control by ownership" to "control by contract" can both improve work incentives and reduce the administrative burden on the state. The state retains control over a key segment of a service operation while subcontracting out the other parts. Taiwan and Korea show how some of these ideas can be operationalized (for Korea, see Wade 1982a). These organizational requirements are no doubt difficult to meet in many states, even if focused on only a small part of the civil service. But they are 380 CHAPTER 11 not of the kind that ask elephants to fly; they should not cause us to say that outside the configuration of a hard state political reality dictates a close approximation to free market policies as the best practical way forward. We tend to assume, wrongly, that the ills of public bureaucracy are intrinsic to this generic type of organization, and therefore are inclined to embrace "the market" as the preferred alternative. In fact, many of those ills are features of one particular-though predominant-model of organization found in the corporate as well as the public sector. This "mechanical" model, characterized by elaborate specialization of tasks and standardized procedures, by extended chains of command and written communications, is now being radically reformed in the corporate sector of the West, so as to create organizations better adapted to the exigencies of the new information and production technologies (Hoffman 1989). These new forms of corporate organization can provide experience for new forms of public administration. The current economic difficulties, and budgetary pressures in particular, are making many governments more prepared to tackle difficult institutional issues than would be the case in more normal times; they are running out of alternatives. At the same time, the current conditions in the world economy increase the potential advantages of pursuing GM policies-to modulate the volatility of the world economy on the domestic economy, to help domestic firms compete internationally in the face of increasingly fierce international competition, and to force an early entry into information technologies. Firms whose governments decline to provide assistance may relocate, or they may resort to squeezing labor costs and intensifying work practices in order to avoid losing market share, enlisting the power of the state to help them do so. This higher premium on GM policies raises the advantages of undertaking the organizational improvements needed to support them. In any case, whether government seeks to promote particular industries or seeks only to make all markets freer, it is likely to have to make organizational changes along these lines. Even a government committed to free trade must be purposeful, must have a system of policy management that recognizes the effects of interactions among its own activities, and must be able to insure that desired responses are forthcoming from the commitment of public resources used as side-payments to those who would otherwise block market processes (Heclo 1986). Free trade policies are no means of escape from the need to improve the capabilities of governments. The United States is a telling example. Lacking these capacities, the U.s. government uses leaky protectionist instruments as much as most other industrialized economies. But its departures from free trade are largely a case-bycase response to domestic political pressures rather than being part of a strategy for gestating or nurturing future competitive industries. With no one being required to explain or defend what is being done, its industrial policies remain ad hoc and implicit. Indeed, the philosophical repugnance against government LESSONS FROM EAST ASIA 381 involvement in industrial promotion is such that the government lacks both detailed knowledge of industries and analytical capacity to select appropriate actions. If an industry is in trouble, the government is meant to become involved only when trade is said to be "unfair" or when national security is involved; otherwise, whether the industry becomes more competitive, moves offshore, or goes out of existence is a matter of government indifferencebecause the market outcome is assumed to be best. This repugnance is translated into and then confirmed by organizational incompetence at carrying out concerted industrial policies. The personnel policies of the federal government compound the problem, for they are designed to attract (in the approving words of a former associate director of the Office of Personnel Management) "competent people, not the best and most talented people," who should be encouraged to migrate to the private sector, "the true vehicle for prosperity" (cited in Allison 1989). In these several ways, the United States is a model of what developing countries should avoid. Yet the thinking of most of the international aid community, including USAlD and the World Bank, is profoundly shaped by American conceptions of the role and competence of government. And the political meta-assumptions of neoclassical prescriptions are calibrated to those same ideas. All the more reason for developing countries to study the East Asian experience to see how government and capitalism are arranged where economic development has been a top national priority for decades. And all the more reason for economists to accept the challenge of constructing a theoretical rationale for the non-neoclassical East Asian facts. When the next Wealth of Nations comes to be written, it will look more favorably upon governing the market. The first Adam Smith would surely approve. It was he who warned from his study of the history of astronomy, "The learned give up the evidence of their senses to preserve the coherence of the ideas of their imagination" (in Lindgren 1967:77).