Dynamic versus Static
Matthew McCartney (SOAS, University of London)
© Copyright 2004 Matthew McCartney
This paper begins by outlining the neoclassical theory of efficiency, using international trade in Bangladesh as a case-study. This notion of efficiency is extremely narrow, and concerned only with the allocation of a given quantity of resources. Competition is better modelled as a dynamic process. This difference is considered in the context of Bangladesh. The phase-out of the WTO’s Multi-Fiber Arrangement (MFA) quota regime will lead to intensified international competition for textile exports. Dynamic efficiency can be defined as a virtuous circle of increasing productivity, output and wages (the high-road). Likewise a vicious circle of reduced wages, longer hours and intensified working conditions is possible (the low-road). Neoclassical economics has no means to distinguish between these two processes, if all returns are equalised at the margin it is perfectly possible for both to be considered efficient. Dynamic efficiency is argued here to be an alternative paradigm to neoclassical economics. The implications for economic analysis and policy making are briefly considered. The principal conclusion of this paper is that the narrow view of efficiency has restricted the relevance of neoclassical economics. A more realistic interpretation of how economies function as dynamic not static entities is important in properly evaluating the conflicting and complementary roles of government intervention and the free-market. The most important implication of dynamic efficiency is in setting a theoretical basis of for the economic analysis of the developmental state.
The explicit theoretical rationale of liberalisation according to neoclassical economics is to achieve an efficient (static) allocation of resources. The link to economic growth is implicit, rational individuals will save according to criteria such as the life-cycle hypothesis, profit maximising firms will utilise these available resources to invest efficiently. In a free market there is no such thing as growth that is too slow, growth reflects the time preferences of individual agents. Price signals link the short and long-run and there is no need to consider the two separately.
In international trade neoclassical economics offers a strong theoretical prediction. The theory of comparative advantage states that a country will export goods intensive in its abundant factor, and import those intensive in its scarce factor. For South Asia with a relatively low area of land per person1 and abundant labour, exports should principally comprise labour-intensive manufactured goods rather than primary sector products. Structural adjustment should see a shift in the composition of production from capital-intensive import substituting industries2 to export-orientated labour-intensive industries. A well-documented and lauded example of such growth3 is the phenomenal expansion of the ready-made garment sector (RMG) in Bangladesh. Exports were negligible in 1979/80, by the late 1990s Bangladesh had become the twelfth largest apparel exporter in the world, the RMG sector accounted for about 76% of total export earnings. By the late 1990s the industry employed 1.5m people, 90% of them women. The change is efficient from a neoclassical perspective, the abundant resource (unskilled/ female labour) has been re-allocated (rural-urban migration) in a rational response to price incentives.
Efficiency in neoclassical Economics
The neoclassical concept of efficiency is extremely narrow, this is revealed in striking clarity by an examination of four well-used microeconomics text-books4. In general ‘efficiency’ gets only a passing mention and is entirely subsumed by the concept of Pareto efficiency.
In Gravelle and Rees (1992) and Kreps (1990) efficiency is solely a static concept concerned with the efficient level of output of public and private goods, efficient risk sharing, solution to bargaining, the Edgeworth Box, and Walrasian equilibrium5. In Mas-Colell (1995) efficiency gets six entries in the index, the Pareto concept appears 76 times. Kreps (1990) doesn’t bother to separate them, “Efficiency, see Pareto efficiency” (p824) notes the index, Pareto efficiency in its various forms appears 26 times. Also in Varian (1992) efficiency appears only as Pareto efficiency (p225).
The necessary requirements for Pareto efficiency (Gravelle and Rees, 1992, p479-485) are ‘efficient consumption’, ‘efficient input supply’ and ‘efficient input use’ (production efficiency) and ‘efficient output mix’. Theses are the “three types of efficiency embodied in a Pareto optimal exercise” (Mas-Colell, 1995, p564). The first is consumption efficiency, where consumers have allocated their budgets to maximise their own well-being (utility maximisation). The marginal rate of substitution between any two goods equals their price ratio. The second is production efficiency, where producers cannot alter the ratio of inputs to raise output or reduce the cost of a given volume of production. The marginal rate of technical substitution between any two inputs equals their price ratio. The final measure is aggregate output efficiency, where resources are allocated simultaneously to achieve both production and consumption efficiency. Where for example in a society of bipeds an equal number of right and left shoes are produced. Utility and profit maximisation will ensure consumption efficiency and the efficient use of inputs and composition of outputs.
Glancing again at the index in Gravelle and Rees (1992) at ‘dynamic’ reveals only a set of references that give more mathematical rigour to the concept of static equilibrium. By dynamic efficiency neoclassical economics means the existence, stability and uniqueness of equilibrium. Dynamic analysis is shorn of any substance and asks simply whether an economy in equilibrium (existence) subject to an exogenous shock will return (stability) to its original position (unique). There are a few cases such as the cobweb model which has a unique equilibrium but any deviation from which can produce an explosive divergence of price and output, such are at most given passing attention.
Imperfect Information and Market Failure: A Radical Departure?
Theorising on imperfect information and markets failures appears to be a radical departure from the neoclassical paradigm. However this analysis implicitly accepts efficiency as being a static concept, Pareto Efficiency as the benchmark and government policy as a means to make the world look more like the neoclassical theory.
If there exists a wedge between social and private costs (an externality), a taxation, subsidy or regulation can push the economy towards the overall social optimum. An optimal Pigouvian tax can replicate efficient allocation, see for example Mas-Colell et al (1995, p355). Similarly government policy may help solve the preference revelation problem for public goods, see Varian (1992, p425). There may be some problems government policy is unable to overcome such as moral hazard and asymmetric information in the market for bank loans. The market is then constrained to allocate resources in a second best world, see Stiglitz and Weiss (1981). The very notion of ‘second-best’ illustrates the striking normative preference for Pareto Efficiency
It is not with the analysis of market and information imperfections that we are forced to confront the implications of an alternative paradigm. The crucial assumption is an economy that is static, where efficiency is measured at a moment in time. In an alternative world, when we consider the dynamics of competition, investment and growth, what we mean by efficiency takes on a radically new meaning. An implication of this proposition is introduced in the context of future prospects for the Bangladeshi textile industry.
Competition is a Dynamic Process
As of December 31st 2004 textile and clothing products will be subject to WTO rules, with the final abolition of the MFA6. When the MFA was being implemented in the 1970s Bangladesh was not considered to be a viable exporter, consequently it was never subject to its strictures. Other potentially competitive exporters such as Sri Lank, India, Pakistan and China have been subject to binding MFA quotas on apparel and textile exports. Bangladesh has been able to export into an open niche in world markets since the late 1970s, after 2005 Bangladesh will face intensified competition on world markets. There are broadly two potential outcomes, the low and high-road of competition. The latter is ‘dynamically efficient’, leading to rising wages and productivity over time. The concern of neoclassical economics with efficient allocation has no theoretical means to distinguish these two processes, as long as marginal equalities are retained according to neoclassical criteria even the low-road of competition could be judged efficient.
Dynamic Efficiency and the Low and High-Roads of Competition
Bangladesh is currently most competitive in price sensitive, low-value, low-priced items7. Bangladesh has two options to compete after 2005, raising productivity or reducing costs.
Bangladesh could react to intensified competition by trying to enhance it price competitiveness within its existing niche by extending hours, reducing overheads (subcontracting) and intensifying work conditions (a low-road of competition). There is some evidence this path has already been pursued in the Indian textile industry.
The fragmentation, ruralisation and casualisation consistent with a low-road of competition has already had a profound impact in India. As early as the 1960s textile mills in Ahmedabad and Bombay began putting out weaving work to decentralised power-loom units. Pharmaceutical firms in Bombay passed on work to smaller units located away from the high-wage industrial belt. From the 1970s there was a general increase in the use of contract, temporary and casual workers. The share of casual workers in large factory employment rose from 4.6% in 1980/81 to more than 12% in 1993/948. Subcontracting was not a significant activity prior to 1970, by 1978 it was a prominent activity in large factories with a share of 21% of total employment9. In India especially sharp has been decline of large urban cotton mills and ruralisation of the industry10. This ruralisation of labour is reflected in the fall in the average size of industrial units from 3.2 to 2.5 workers between 1961 and 1991. The fall in average employees per factory is true for most industries and has persisted throughout the 1970s and 80s11.
b) The High-Road of Competition
A high-road of competition could consist of remaining in an existing production niche and raising productivity, or upgrading to a less (price) competitive market niche to capture rents. In the RMG sector Bangladesh may compete by capacity building to enhance skills in fashion, design, cutting and technology upgrading, developing backward linkages to suppliers to shorten lead times, and improving the skills and training of management and workers.
Good policy can be defined as that which helps achieve a high-road response to competition. Dynamic efficiency is a situation characterised by a virtuous circle of higher productivity, output growth and higher wages rather than having a rigorous mathematical definition.
Dynamic Efficiency, Rents and Learning
Dynamic efficiency is an alternative paradigm to neoclassical efficiency. In fact there is likely for various reasons to be trade-offs between static and dynamic efficiency. When we are considering dynamic efficiency good policy cannot be mechanically judged in terms of whether it liberalises the economy, encourages competition or expands the freedom of decision making. Policy is a far more nuanced process that has to be carefully evaluated in terms of its effect on the dynamics of investment, growth and competition.
One neoclassical assumption immediately disposed of when we consider dynamic efficiency is that no allocation or industrial structure is preferable to any other. In fact, while many allocations may be efficient, some are more (dynamically) efficient than others.
Neoclassical theory argues that export structures are simply a product of comparative advantage and factor prices. The composition of exports does not matter; no set of activities are more desirable than any other. There are no externalities, so returns are equalised at the margin (efficient allocation). Lall (1999, p1775) notes that spill-over benefits for the whole economy are positively related and ease of market entry of competitors negatively related to the technological complexity of a product. The consideration of dynamic efficiency is, much more than market or information failures, what creates the potential for industrial policy by the government. We can broadly define industrial policy as a deliberate action by the state to shift the structure of the economy away from its static comparative advantage to a structure offering more dynamic potential. We generate the first strong implication of our alternative paradigm: there may exist a trade-off between static and dynamic efficiency.
b) Profits and Efficiency
Profits in the neoclassical model are a temporary aberration of the market. Profits may exist temporarily before resources and factors flow into a sector and compete them away. In a dynamic world profits (or more correctly rents) are useful to induce and reward learning in order to raise productivity or upgrade to higher value-added and less price sensitive sectors. Learning is much like patents that reward innovation in a developed country.
Neoclassical economics assumes innovations take place in advanced countries and learning in less developed countries (LDC) is no more difficult than selecting the most appropriate. Innovation (shifting the production frontier) is distinct from mastering/ adapting technology13. In truth, though much technology is tacit, experimentation and learning are necessary to understand the tacit elements and adapt them to local conditions. In practise there is less difference between innovation in developed countries and industrialisation based on learning already commercialised technology.
Investment in learning by one entrepreneur in discovering a commercial niche that can be profitably exploited is likely to lead to rapid imitation14. If such learning requires investment, the returns to which cannot be fully appropriated, entrepreneurs in LDC’s face similar problems to innovators in developed countries. While neoclassical economics subscribes to the need for patent protection to generate an incentive for innovation, it advocates complete freedom of market entry in all other scenarios. LDC investors should not get patent protection no matter how high the (external) social return. Entrepreneurial learning is likely to be under-supplied. Profits / rents that reward and motivate learning may lead to a more dynamically efficient economy even if they are a sign of resource misallocation according to considerations of static allocative efficiency.
c) How do we Evaluate Policy?
The analysis of policy intervention in the static neoclassical model is easy, anything that increases the scope of the free market and free decision making is a good thing. When we consider our alternative paradigm, that of dynamic efficiency, the analysis of policy is much less clear. Policy needs to increase the expected payoff to learning, hence it is important to distinguish firms that are engaged in costly learning and those who simply imitate the results of others’ learning. The parallels with innovation and patent protection are evident.
Temporary trade protection may increase profits from learning but only for firms producing for the domestic market14. Trade protection does not discriminate between innovators and imitators. This will promote early entry and lower the expected return to learning. Export subsidies avoid the anti-export bias of trade protection but likewise do not discriminate between learners and imitators. Export subsidies can be relatively good at discriminating between successful and unsuccessful performers ex-post. Providing subsidies or government credit contingent on exporting can allow policy makers to discriminate between firms.
d) Dynamic Efficiency and Liberalisation
Neoclassical analysis of efficiency pre-supposes that good economic policy consists of removing constraints on the operation of the free market. Individuals are rational, so any constraints on voluntary options and mutual exchange can only reduce welfare and efficiency. One exception is that of game theory, or more precisely the ‘Prisoners Dilemma’. This illustrates a situation in which individuals acting in their own self-interest generate a socially sub-optimal outcome. Some sort of constraint is necessary to prevent individuals rationally defecting to maximise the social return. The literature has typically analysed this in terms of extra-economic factors such as trust, culture or coercion. For example:
“an economy can perform well only to the extent that it is embedded in a well integrated society, and that a society exists only to the extent that it is capable of imposing normative constraints, or social obligations, on the pursuit of individual interest.”15
This question is of immediate relevance to Bangladesh. Currently unions are forbidden from operating and organising in Export-Promotion-Zones where many of the RMG factories are located. Japanese and Korean foreign investors are threatening to withhold FDI should the law be amended. Concerned institutions in the US motivated by ‘fair trade’ rhetoric are pressing for this to happen under threat of countervailing import duties. Placing a floor under the process of cost-cutting, longer hours and intensified working conditions may force producers to pursue a high-road to international competition.
Successful policy cannot simply be judged in terms of the degree to which markets are liberalised. Once we consider economies as dynamic rather than static entities and evaluate policy in terms of achieving dynamic not static efficiency, what we conceive of as good policy becomes far more nuanced. The impact of policy on learning and imitation is relatively clear, but the relative merits of trade protection, export and government subsidies are more subtle and complex. Certainly we can say liberalisation has to be carefully compared and evaluated against other possible policies; certainly liberalisation can only ever be a policy means to achieve a given end; it certainly cannot be judged an end in itself. Beginning with a benchmark of ‘dynamic efficiency’ we have arrived at the theory of the developmental state, this is the archetype of a dynamically efficient economy. The developmental state is an alternative to neoclassical efficiency, not an occasional aberration and second best-solution to allocative inefficiency16.
1. A.Wood and M.Calandrino, ‘When the Other Giant Awakes: Trade and Human Resources in India’ (2000), University of Sussex, IDS Mimeo.
2. Which were prominent parts of the domestic industrial structure in India especially before liberalisation.
4. These are H.Gravelle and R.Rees (1992), Microeconomics (2nd Edition) (London, Longman, 1992), H.R.Varian, Microeconomic Analysis, (Third Edition) (London, W.W.Norton and Co, 1992), D.M.Kreps, A Course in Microeconomic Theory, (London, Harvester Wheatsheaf, 1990), A.Mas-Colell, M.D.Whitston and J.R.Green, Microeconomic Theory, (Oxford, Oxford University Press, 1995).
8. K.V.Ramaswamy, ‘The Search for Flexibility in Indian Manufacturing: New Evidence on Outsourcing Activities’, Economic and Political Weekly, (1999), 34:6, pp. 363-8.
9. B.Harriss-White, India Working: Essays on Society and Economy, (Cambridge, Cambridge University Press, 2003).
13. Y.W.Rhee (1990) notes that the number of export-orientated RMG factories in Bangladesh exploded after the single firm Desh proved it was a profitable proposition at the end of the 1970s, by 1985 there were 700 such firms.
15. W.Streeck in J.R.Hollingsworth and R.Boyer, Contemporary Capitalism: The Embeddedness of Institutions, (Cambridge, Cambridge University Press, 1999) p199.