Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

The timing, in theory, of advancing science and technology

Presuming that it is desirable for developing countries to pursue science and technology, the next issue that arises is when they should do so. Here, again theory has something to say. The economic theory that we shall draw upon is the so-called optimal growth theory, 'growth' in that it attempts to describe, in the abstract, the growth of an economy over the long run; and 'optimal' in that it attempts, for the economy so described, to determine how scarce resources should be allocated through time so as best to achieve the economy's overall objective. The objective is often specified as maximizing the discounted stream of consumption per capita, over some long, perhaps infinite, horizon; in less technical language as obtaining as high a standard of living as possible throughout the future.

Since the optimal programme is somewhat surprising and, at first at least, extremely austere, it might be useful to describe the theory of optimal economic growth, under advancing science and technology, in some detail. The first piece of theory that we shall draw upon is that formulated by Phelps (1966).

Phelps' objectives in formulating his theory were twofold. First, he wanted to represent, in the abstract, the mechanism by which advances in science and technology led to increases in the productivity which scarce resources achieved in production. He called the activity which generated increases in productivity 'research': research is one of the two sectors comprising his economy. In the mathematical model describing the economy, laid out in detail in the article, research is undertaken by only one of the two factors of production: labour, allocated to the research sector, yields technical progress, or, as we would call them, advances in science and technology.

These advances have profound effects on the performance of the economy. In the research sector already identified, advances in science and technology make subsequent research easier to undertake. Advances achieved today enable labour engaged in research to be more productive tomorrow; in other words, undertaking research now creates a momentum which leads to more effective research in the future. The other effect of research is to make the remaining portion of the labour force, engaged in the (second) sector of the economy which produces its physical output, more productive also. The advances in science and technology are assumed to disseminate throughout the productive sector immediately and costlessly, so that all the labour there employed becomes more efficient as a result.

Having represented the pursuit of science and technology in this manner, Phelps then asks how should an economy allocate its labour supply between the research sector and the productive sector so as to maintain, throughout all time, maximum consumption per capita, for the growing population. Phelps is not concerned with how that economy reaches this optimum state, merely with the properties of this optimum. This optimum he calls the 'golden age', since it can, without changes in the system or its parameters, exist in this happy state forever. It is the dynamic equivalent of static equilibrium: and can be thought of as a momentary equilibrium for all moments of time. Mathematicians call this a 'steady state'.

What is the allocation of labour among the two sectors in the steady state? In that steady state characterizing the 'golden age' (there may be other steady states which fail to maximize consumption per capita) it happens that the portion of the labour force devoted to research is exactly half the total. This coincidental result comes about from the mathematics, in particular the functional forms chosen to express the outputs of the research and productive sectors, but may not be as unrealistic as seems. If, instead of science and technology (our collective term for technical progress; or 'research', the term used by Phelps) one thinks of 'information', the term used by information scientists, one can recognize their claim that approximately half of all labour in the developed countries is devoted to generating, processing and applying information. They say that the Japanese, US or Western European economies are 'information economies'; Phelps would say that they are 'research economies', and we that they are scientific and technological economies. And Phelps' prescription for an underdeveloped economy is that it be, somehow, an 'information economy'.

Phelps also illustrates the benefit that accrues to the research economy in its 'golden age': with the labour force growing at a constant rate (as elements do in a steady state) and without any research being conducted, i.e. with all the labour force allocated to production, the economy grows at a slower, although constant, rate. In other words, the economy grows at the same rate as the labour force, or in this abstract economy, as the population. Consumption, per capita, is static. The economy is growing, to be sure, but no faster than its population. However, with research undertaken so extensively as to occupy half the labour force, the rate of growth of the economy, and of consumption too, is twice the rate of growth of the population. In other words, the economy is growing twice as fast as the population; and consumption at the same rate. How superior an outcome this is!

There are two qualifications that we may wish to impose upon Phelps' model: that it has a narrow view of what resources are utilized in undertaking research and that it does not address the issue of how the economy attains the 'golden age'. The first of these qualifications does not really hold, neither in theory nor, perhaps unfortunately, in practice in the part of the world that we are investigating. In his development of the theory, Phelps does consider how his results would differ if capital, a second factor of production already necessary for production in his model, is also necessary for 'research'. He thus considers an augmented model in which both inputs, labour and capital, are needed in both sectors. When capital, as well as labour, is a necessary ingredient for the conduct of 'research', a 'golden age' is still attainable in principle. The results indicate that it is even more golden - 24 carat as against say 16 carat in the case of 'research' requiring labour alone - in that the steady state rate of growth of the economy can exceed twice the rate of growth of the labour force. The reason would seem to be that the scarcest input in this economy is labour; and if something can be substituted for labour, then the labour constraint on growth can be relaxed. The something that can substitute for labour is capital.

The qualification that Phelps took too narrow a view of the inputs necessary for research may not hold in Sub-Saharan Africa, because labour seems to be almost the sole input to research. We are anticipating the results of our inquiries into the conduct of R&D there, but one of our findings is that almost all the expenditures of the institutes are devoted to wages and salaries. What Phelps would designate as capital - such things as laboratory equipment, supplies, scientific books and journals, etc; the wherewithal of research - are severely lacking in practice. Advances in science and technology are secured almost solely by the unassisted labour of scientists and technologists.

The second qualification to Phelps' analysis is that it neglects what in a practical sense would be the development of the economy. How does the economy reach perfection? What happens before the 'golden age'? For an answer to this question we must move to a different sort of growth model, one which considers the path to be followed from any initial situation, to that of the 'golden age'. In mathematical terms, this means considering the transition from a non-steady state to that particular steady state characterized as the 'golden age'. Posing the task as a question: what is the optimal path to be followed by an economy, given that it commences from a non-optimal position? And what allocations, through time, will ensure that the economy adheres to this path?

To answer these related questions we shall move on to the work of Uzawa and Enos. Uzawa's was much the more original and grand contribution (Uzawa, 1965). He, like Phelps, assumes that, of the two inputs the economy utilizes, only labour is needed to advance science and technology. The sector in which science and technology are advanced is called, by Uzawa, 'education'. When allocated to the educational sector, labour achieves an output which raises the productivity of all the (remaining) labour assigned to the productive sector. As in Phelps' case, the productivity of labour in the productive sector is raised, through 'education', universally and immediately. The cost thereof is in the output of the productive sector forgone through the alternative allocation of otherwise productive labour to 'education'.

In Uzawa's model there are two sequences of decisions which the economy must make at every instant of time, if it is to achieve its objective. The objective is rather more complicated than in Phelps' model; in Uzawa's it is to maximize the discounted sum of consumption per capita over an infinite horizon. In the 'golden age', the issue of discounting does not arise, since every instant is the same, and the future identical to the present. In the interval, perhaps infinite in time, before the 'golden age' is attained, however, a little more consumption today, and its necessary concomitant, a little less in the future, must be weighed against a lime less today and a little more in the future. Discounting future units of goods consumed against present consumption makes some sense as a weighting process. The two decisions the economy must make are in what portions to allocate its scarce labour supply among the two sectors, 'education' and production, and in what portions to allocate the current output of goods among the two competing uses, as consumption and as investment. Although education in Uzawa's model needs only labour, production needs both labour and capital; and the capital stock available for production is enlarged only by investment.

What are the two optimal allocations through time, for an economy that is initially short of 'education'? Remember that being short of 'education' means having a low productivity of labour in production, and consequently producing little output with the labour and capital initially available. In our terms, it means having initially a low level of science and technology, less, probably far less, than would be accessible in the 'golden age'.

Taking first the output of the productive sector, which is either consumed or invested, it is optimal initially, and for some extended length of time, to allocate the entire amount to investment. In other words, the capacity of the productive sector to produce goods should be built up as quickly as possible. Current consumption should be halted; only the ability to produce for future consumption should be considered. Only after an appropriate stock of capital has been accumulated should some of current output be made available for consumption.

In Uzawa's artificial world, consumers (= labour) are compensated for not consuming today by consuming more. in the future; but in reality such abstinence is not practicable. Economists immediately recognized the extreme nature of the optimal programme, the absurdly high value of the optimal savings rate. A more realistic savings/investment rate, one which might be politically and socially feasible, is to maintain consumption per capita constant and allocate all the remaining output to investment; i.e. to invest the economy's surplus. But, whatever allowance is made for maintaining current consumption, the optimal policy continues unrepentant start by investing everything that one can, for as long as is necessary, or, in more emotive terms, sacrifice today's pleasures for tomorrow's good.

Of more concern to us is the second sequence of decisions, derived as the optimal policy for the allocation of labour between production and 'education'. The choice is not easy, nor simple. If labour is allocated to the productive sector, investment can flourish; if it is allocated to the educational sector, the productivity of labour in the productive sector can flourish. Unlike the decision to consume or to save/invest, this seems not so much to involve a trade-off between present and future: the trade-off is between utilizing one means to the end - applying more labour to production - and the other means to the same end - applying more labour to 'education', thereby raising the productivity of the fewer remaining workers in the productive sector.

What is the optimal programme for the assignment of labour? In Uzawa's model it is not so clear as in Phelps': given Uzawa's assumptions of labour being perfectly mobile and supplied in a perfectly competitive market, its allocation will depend upon its relative contribution, at the margin, in the two sectors. Its contributions will be measured by its shadow prices, which are established according to the discounted future value of the consumption their output permits, be it via the investment goods it produces in the goods sector, or via the increase in labour productivity through the advance in science and technology in the educational sector. There is no magic result, as in Phelps' case; but just as the optimal policy for the allocation of goods involves maximizing, initially, their assignment to investment, so the optimal policy for labour involves assigning initially more than the steady state amount to the educational sector.

The optimal allocation of labour in a model reflecting the advance of science and technology can be seen somewhat more clearly in another growth model, this one attributable to Enos (Enos, 1991). In this model, there are the customary two sectors: one, in which goods are produced; and the other, in which 'training' takes place. The difference between Enos' model, on the one hand, and Phelps' and Uzawa's on the other, lies in what are identified as the scarce inputs: in the latter's models the scarce inputs are labour and capital; in the former unskilled labour and skilled labour. Both types of labour are assumed to be necessary for the production of goods, and both for the training of unskilled labour so that it acquires skills. Progress occurs through the increasing prevalence in the economy of the more productive skilled labour.

What is the optimal policy in Enos' model, given that the objective of the economy is, as in the other models, to maximize the discounted sum of future consumption per head? Not surprisingly, given the previous results, for the economy initially short of skilled labour the optimal policy is to initially allocate all labour to the 'training' sector, provided there is not too great a difference in the degrees to which the two types of labour can be substituted one for the other in the two sectors. If the degree of substitutability in the 'training' sector is low, and that in the production sector high, almost all the skilled labour will be assigned to training, until the proportion of skilled to unskilled labour has become sufficiently high. Again, this is an austere outcome, for with the assignment of most of the skilled labour to training unskilled labour, little physical output is produced, and little consumption takes place. Moreover, another issue arises, one that also arises in the case of Uzawa's and Phelps' models but is left implicit there: the issue of absorption. Along the optimal path, incomes are maintained as both skilled and unskilled labour earn wages in both sectors, but there are few goods to purchase with these earnings. Total income of the economy is undiminished, but the total volume of goods available falls: under the optimal programme it falls to a very low level. The immediate effect of reallocating large amounts of labour to 'training', therefore, would be to put the economy under great stress, as existing incomes chase fewer goods.

The implications of Enos' model for our study are twofold. The first is that the best policy for the developing country would be to assign as much labour as feasible to raising the skill level of the population, or in our words, to advancing science and technology. The second is that the degree of substitutability of the relatively more scarce input for the relatively less scarce is of importance. This phenomenon - substitutability - will appear again in the next model we draw upon, the model that comes closest to including Structural Adjustment.

Before we go on to consider theory devoted to explaining Structural Adjustment we may find it useful to recapitulate the theory, and the few statistical results already described. The various theories have three elements in common: first they are all rigorously developed, using mathematical language to secure consistency and exactness. Secondly, they are all sufficiently powerful to yield interesting, and usually testable, hypotheses. These two elements are the characteristics of all theory in economics; it is the third that divides our theories from the very much larger universe of all economic theories, namely that our sample focuses on the long run. Short term phenomena are missing, and consequently there are also missing those fluctuations in economic activity that occur from month to month or from year to year, and those policies that governments and other economic agents adopt in response. Our theories are concerned with trends, with what can happen over the longer horizon when economic fluctuations are irrelevant.

Given that the theories take the long view, they yield a set of mutually consistent ideas about economic development. Most germane to our study is the idea that the advance of science and technology, to use our term, contributes greatly to the growth of an economy. The demonstrations were initially made by Solow and Denison. The next ideas, that we have attributed to Phelps, are that advances in science and technology can be modelled as a separate activity, utilizing the resources that are scarce in the economy, and therefore useful elsewhere (for example in production), but organized in its own fashion and following its own rules and procedures. When organized separately, the evolution of the sector that generates advances in science and technology can be related to that of the rest of the economy, and their joint progress observed. In the steady state, described by Phelps, the growth of the economy and the advance of science and technology move in step, at a rate greater than that at which physical resources, specifically the labour force, multiply. Equally consequently, a substantial portion of the labour force is always allocated to further and further advancing the scientific and technological level of the society.

To determine how the economy best reaches the steady state, that state where the economy is organized so as to achieve throughout all time the highest level of material satisfaction, we drew upon two optimal growth theories, those of Uzawa and Enos. Like Phelps, they formulated models of long-run economic growth in which the sector generating advances in science and technology was identified separately, but unlike Phelps they addressed the issue of how the economy attained the 'golden aye' of steady, maximal growth. Their simple but self-denying conclusion was that the best course of action for the economy to follow was to allocate all the resources that it could spare foremost to building up its scientific and technological potential. Using their language, the optimal policy was first to allocate all the available resources to investing in capital and 'education' (Uzawa) or 'skills' (Enos). Until capacity to advance science and technology had been created in sufficient amount to permit its rapid advance, little attention should be given to increasing physical output. Only afterwards should producing goods lay claim upon the economy's scarce resources.

So, in summary, advancing science and technology is vital; its advance can be considered separately from the growth of the productive economy; the economy in its highest state has a flourishing ambience of science and technology, in which are usefully employed a substantial fraction of the working population; and finally, to reach this highest state from one less abundantly endowed, with the least sacrifice and in the quickest time, the entire surplus of the economy should, immediately and for an extended interval, be channelled to accelerating scientific and technological progress.

These are the implications of the theory that we have chosen to describe, and these are the theses that underlie our analysis. But we have not yet exhausted our search among economic theories, for we need additional guidance in conducting our inquiry among developing countries with very low levels of science and technology and under great stress. In the language of the theorists, we need guidance for an inquiry into economies formulating their optimal policies with distressing parameter values and under dreadful initial conditions.

Contents - Previous - Next