«Industrial Policies, the Creation of a Learning Society, and Economic Development1 B. Greenwald and J. E. Stiglitz Industrial policies—meaning ...»
A key issue of industrial strategy is not only the direction (should Korea have attempted to reinforce its comparative advantage in rice, or to create a comparative advantage in some other area?), but also the size of the step. Should it try a nearby technology (product), nudging along a gradual, evolutionary process that might eventually have occurred anyway? Or should it take a big leap? The latter is riskier: perhaps greater returns if successful, but a higher probability of failure.
We have not formally modeled this critical decision, so the following remarks are only meant to be suggestive: The ability to learn (costs of learning) increase significantly the bigger the leap;
but so may the benefits. There are natural non-convexities in the value of
information/knowledge (Radner-Stiglitz, 1984), implying that it pays to take a moderate step:
small incrementalism is not optimal.
By the same token, using another analogy, to corporate strategic policy, it pays to move to a part of the product space where there are rents which can be sustained (e.g. as a result of entry barriers, arising, for instance, out of returns to scale and/or specific knowledge.) This almost surely entails not doing what others are or have been doing.
The inevitability of industrial policy We have argued that government cannot escape thinking about its industrial structure. It is necessary as it makes decisions about public investments (in education, technology, and infrastructure). But the legal framework of a society too inevitably shapes industrial structure.
If, as in the United States, derivatives are given seniority in bankruptcy, while student debts cannot be discharged, and large banks are effectively allowed to undertake high risks, with governments bearing the downside, and speculators are taxed at lower rates than those in manufacturing, the financial sector is encouraged at the expense of other sectors. This is an industrial policy.
Developing countries have to think carefully about every aspect of their economic policy, to make sure that they shape their economy in a way which maximizes learning. But their learning challenge is markedly different from that of the advanced industrial countries, where one of the main objectives is moving out the knowledge frontier. The focus of developing countries should be to close the knowledge gap between them and the more advanced countries (though for some of the more advanced among the emerging markets, one of the challenges it to be at the forefront, at least in some particular areas, something at which both China and Brazil have succeeded.) But this in turn has one important implication: legal frameworks and institutional arrangements (such as for intellectual property) that are appropriate for developed countries are not likely to be appropriate for developing countries and emerging markets.
Industrial Policies and Government Failures We began the discussion of this paper arguing that industrial policies are, in part, a response to market failures. The sectoral allocations resulting from unfettered markets are not in general optimal. But some of the inefficiencies in markets arise, as well, from government policies. A natural response is to remove the government distortions, rather than to create a new, offsetting distortion. But such an approach ignores the complexity of political economy and the difficulty of fine-tuning public policies. Earlier, we referred to the impact of intellectual property. But a country’s intellectual property regime is greatly affected by TRIPS, the WTO agreement, in ways which may not accord with its own best interests. It may, accordingly, attempt to undo or “correct” the distortions arising from that intellectual property regime.
The objectives of industrial policy
Industrial policy is usually conceived of as promoting growth, but it should be seen more broadly, as any policy redirecting an economy’s sectoral allocation where market incentives (as shaped by rules and regulations) are misaligned with public objectives. Governments are concerned about employment, distribution, and the environment in ways in which the market is often not. Thus, in those countries with persistent high levels of employment, it is clear that something is wrong with market processes: labor markets are not clearing. Whether the explanation has to do with inherent limitations in markets (e.g. imperfect information giving rise to efficiency wages), unions, or government (e.g. minimum wages), the persistence implies that “correcting” the underlying failures may not be easy. The social costs of unemployment can be very high, and it is appropriate for government to attempt to induce the economy to move towards more labor intensive sectors or to use more labor intensive processes.
In each of these instances, shadow prices differ from market prices. This is evidently the case in many areas of the environment, where firms typically do not pay for the full consequences of their action. The consequences for investment—including investments in R & D—are obvious.
Firms in many countries are searching for labor saving innovations, even in countries with high unemployment, when from a social perspective, there are high returns to innovations that protect the environment.
A persistent criticism of industrial policies is that, even if market allocations are inefficient, even if market prices differ from shadow prices, government attempts to correct these failures will simply make matters worse. There is neither theory nor evidence in support of this conclusion.
To be sure, there are instances of government failure, but none on the scale of the losses resulting from the failures of America’s financial market failure before and during the Great Recession. Virtually every successful economy has employed, successfully, at one time or another, industrial policies. And this is most notable in the case of East Asia. (Stiglitz, Wade, Amsden, Chang).
In the sequel to this paper (Greenwald and Stiglitz, 2012) we explain that limitations in government capacity (“political economy problems”) should play an important role in shaping the design of industrial policies—what kinds of instruments should be employed.
In short, the debate today should not be about whether governments should pursue policies that shape the industrial structure of the economy. Inevitably, they will and do. The debate today should center around the directions in which it should attempt to shape the economy and the best way of doing so, given a country's current institutions and how they will evolve-recognizing that the evolution of the institutions themselves will be affected by the industrial policies chosen.
Appendix: A simple model of investment in R & D
In the text of this paper, we focused on how learning spill-overs affected the optimal production structure—leading to an industrial structure that might be markedly different from that which might emerge in an unfettered market economy. Here, we extend this work by looking at how knowledge spill-overs affect the optimal pattern of R & D.
Assume there are two products, produced by a linear technology
Total employment in sector i is the sum of production and research workers. If Aij 0 (i ≠ j) implies there are spill-over benefits for product i from research on product j. For simplicity, we assume Ri = Lir, the amount of labor devoted to research in sector i.
Social welfare maximization entails Max U(Q1, Q2) – L After some manipulation, the first order conditions can be written α 11 (L1 /L1r) + α12(L2/L1r) = 1 α 12 (L1 /L2r) + α12(L2/L2r) = 1, where ∂ln Ai /∂ ln Ljr = αji.
Role of spill-overs With no spill overs αij = αji = 0, so (Lir /Li) = α ii.
The ratio of employment in research in sector i to production labor is directly related to the own elasticity of productivity. If the elasticity is high—research increases productivity a lot— then a large fraction of labor should be devoted to research.
It is easy to see that if there are externalities (i.e. αij 0), research is increased. Consider the symmetric case, where L1 = L2 in equilibrium. Then L1r/L1 = α 11 + α12.
With perfect spillovers, α 11 = α12, so the effect is to double the ratio of research workers to production workers.
Comparison with a market economy In a perfectly competitive economy with a large number of firms and perfect within-industry spillovers, there would be no research, as each would try to free ride on others: Lir = 0—clearly an underinvestment in research.
At the other extreme, assume that there were no spill-overs. Then each firm would engage in some research. It would maximize output for any given input, i.e.
an equation that identical to that derived earlier for the optimal allocation, in the case of no spillovers—highlighting the crucial role of spillovers in industrial policies. (The overall level of employment may, however, differ in the two situations.) But there is another critical issue: whether there are spill overs or not is, in part, a matter of industrial policy, e.g. concerning compulsory licensing, cooperative research efforts, and disclosure policies.
Thus, assume there are n firms in the industry, and that Ai = Aii + βΣAij. Government policy can increase β (the spillovers from research j to sector i) and thus the optimal amount of research.
Moreover, if sector i has learning as well as research potential, and the other sector does not, then Li will be much greater with β 0, and hence so will Lir.
More typically, sectors in which research is important are imperfectly competitive. Assume that again there is no knowledge spill over, and that each sector faces an elasticity of demand of ε. Then, as before, we can show that Lir/Li = αii.
Production (output) is lower, i.e. for any given level of productivity (Ai ), Li is smaller; and hence Lir is correspondingly smaller. The exploitation of market power results in under production, and thus underinvestment in research, since the value of research is related to the cost savings—i.e. the level of production.
In the case was identical learning functions but differences in demand elasticities, interestingly, the percentage reduction in output is the same, and hence relative increases in productivity stay the same. The monopoly engages in less than optimal research29—but more than the competitive market ( with full spill-overs, where there is no research.) We note, however, that we have assumed implicitly the ability to impose lump sum taxation. With distortionary taxation, the optimal amount of research will obviously be less than with lump sum taxation. See Stiglitz  The long term growth and structure of the economy depends critically on the nature of competition (which itself is endogenous) and spillovers. A Cournot duopoly with full spill-overs may, for instance, result in more R & D than a monopoly with a similar R & D functionOver time, the effects can be cumulative, i.e. the more monopolized sector has lower productivity growth.
Its scale is, as a result, diminished, with resulting diminution in incentives to engage in research.
It is (in this case) not because monopoly has induced laziness, but simply that it does not pay to do as much research.
References Amsden, Alice, 1989, Asia's Next Giant: South Korea and Late Industrialization, Oxford, UK. Oxford University Press.
Arrow, K.J. (1951). ‘An Extension of the Basic Theorems of Classical Welfare Economics.’ In J. Neyman (ed.), Proceedings of the Second Berkeley Symposium on Mathematical Statistics and Probability.
. Berkeley: University of California Press, 507–532.
___. (1962a). ‘The Economic Implications of Learning by Doing’. Review of Economic Studies. 29/3: 155– 173.
___. (1962b). ‘Economic Welfare and the Allocation of Resources for Invention’. In R. Nelson (ed.) The Rate and Direction of Inventive Activity: Economic and Social Factors, National Bureau of Economic Research (NBER). Princeton, NJ: Princeton University Press.
Atkinson, A.B., and J.E. Stiglitz. (1969). ‘A New View of Technological Change’. Economic Journal 79/315: 573–578.
Chang, Ha-Joon, 2002, Kicking Away the Ladder: Development Strategy in Historical Perspective, London,
Coase, R. (1937). “The Nature of the Firm,” Economica, 4(16), pp. 386-405.
Council of Economic Advisers, 1995, “Supporting Research and Development to Promote Economic Growth: The Federal Government’s Role,” Washington, DC: 1995.Dasgupta, P.S., and J.E. Stiglitz.
1980 “Industrial Structure and the Nature of Innovative Activity,” with P. Dasgupta, Economic Journal, 90(358), June, pp. 266-293 _______(1988). ‘Learning by Doing, Market Structure, and Industrial and Trade Policies’. Oxford Economic Papers 40/2: 246–268.
Dasgupta, P. and P.A. David. (1994). "Toward a new economics of science," Research Policy, Elsevier, 23(5): 487-521.
Dickey, D.A., and W. Fuller. (1981). “Likelihood Ratio Statistics for Autorregresive Time Series with a Unit Root.” Econometrica, 49:1057-1072.
Emran, M.S. and J.E. Stiglitz. (2009). “Financial Liberalization, Financial Restraint and Entrepreneurial Development,” working paper. Available at http://www.gwu.edu/~iiep/assets/docs/papers/Emran_IIEPWP20.pdf Furman, Jason, and J. E. Stiglitz, “Economic Crises: Evidence and Insights from East Asia,”, Brookings Papers on Economic Activity, 1998(2), pp. 1-114 Greenwald, B. and J. Kahn. (2005). Competition Demystified : a Radically Simplified Approach to Business
Greenwald, B., and J.E. Stiglitz. (1986). ‘Externalities in Economies with Imperfect Information and Incomplete Markets’. Quarterly Journal of Economics 1/2: 229–264.
—— and ——. (1993). “Financial Market Imperfections and Business Cycles,” Quarterly Journal of Economics, 108(1), pp. 77–114.
___. (2003). Towards a New Paradigm in Monetary Economics. Cambridge, UK: Cambridge University
___. (2006). ‘Helping Infant Economies Grow: Foundations of Trade Policies for Developing Countries’.
American Economic Review: AEA Papers and Proceedings 96(2): 141–146.