I think I need hardly remind you that there is a crucial respect in which Soviet industrial development since 1929 has diverged from the traditional road of development followed by capitalist economies in the past. The latter has been summed up as the road of ‘textiles first’,¹ whereas the former embarked, on principle, upon the building up of basic power and metals and engineering first. Indeed, what has come to be known as the ‘Soviet way of industrialization’ was essentially characterized by this priority to development of heavy industry or the industrial sector producing capital goods.

Lenin was quoted as having said that a crucial feature of capitalist development was that the ‘production of means of production’ proceeded faster than the ‘production of means of consumption’ i.e. ‘production for production’. Ever since the [19]20’s (save for a brief period of questioning in the middle [19]50’s) it has been taken as an unquestioned principle of Soviet economic orthodoxy that if this is a feature of capitalist development, it is a feature also of socialist development, only much more so.

Added to this unorthodoxy (from a western viewpoint) was a further one: that from 1929 onwards Soviet planning showed a preference for adopting the latest types of western technique. They built the most modern type of available steel plants and rolling mills, built textile mills in Uzbekistan equipped with automatic looms, and mechanized agriculture with tractors and combine harvesters. In the period of the First Five-Year Plan it looks as though handicraft production, instead of expanding, actually contracted and raw materials etc. were directed to factory industry.

Now the reasons why in the past economists in the west have regarded these characteristic traits of Soviet development as uneconomic (as examples of rejecting economic principles in favour of ideological prejudices, of national aggrandizement, or, as many have thought of the pre-war decade, for military reasons) can be summed up in terms of what one may call the theory of factor proportions. This is virtually a doctrine of ‘comparative costs’ in terms of marginal productivity. Those factors of production that are relatively plentiful have a low marginal productivity and hence a low price and conversely with factors that are relatively scarce. Consequently those lines of production and those technical forms of production that use relatively more of the plentiful factors and economize on the scarce ones would have the lowest costs. Hence in a country like Russia with plentiful labour and scarce capital (there was of course both urban unemployment and a substantial surplus of labour in the countryside in the Russia of the late [19]20’s), relatively labour-using techniques are most economical (rather than capital-expensive ones). For the same reason handicrafts and light industries rather than heavy industries, requiring large expenditure of fixed capital (plant and equipment), are to be preferred.

Since then this doctrine has been generalized into a general principle of development in underdeveloped countries in successive U.N. publications and in most of the American text books on development which have appeared in recent years.

The strength of this factor proportion theory is, of course, its strong appeal to common sense: it seems obvious common sense to adapt your development plans and methods so as to make the most use of those economic factors that are most plentiful – in this case labour. To which can be added the humanitarian argument that to do this will create the maximum employment in conditions where there is a large reserve of unemployed (or underemployed). The immediate objection to it that springs to mind is that the factor proportion theory, in common with any comparative cost doctrine derived from it, is a static theory, which refers to a particular factor endowment at some given date. Isn’t it inappropriate to derive therefrom a criterion of development, since one is here dealing essentially with dynamic situations, where the factor endowment is subject to change (the essence of development, e.g. is a growing accumulation of capital and hence a change in the capital-labour ratio)? If one tries to do so, won’t you find a catch in it? Won’t you at any rate, get possibly a different answer according as you relate your corollaries to the factor endowment of today, or of a decade or two decades from now.

Now I think this objection doesn’t really unseat the traditional corollaries provided that the rate of change of the factor situation – speaking specifically, the increase in invested capital – is substantially unaffected by the choice you make initially (a choice on the basis of existing comparative costs if you follow the theory’s advice). If this is not the case (i.e. if the choice you make does affect the rate at which the economy can grow) then the objection does have force – it has force to the extent that your choice cannot be based simply on the prevailing cost situation; but must take into account the difference made to that situation at various dates in the future according to the difference in your choice.

How you make your choice would depend on what you conceive to be the essential determinant of the amount, that an economy can invest at any time – of its ‘investment potential’ (and hence to a large extent of its growth potential). I don’t want to go into the technicalities of the so-called theory of choice of technique, but one can put the matter a little more precisely in the following way. Choice in most of the relevant cases will be between techniques that cost more initially (the more capital intensive and labour saving ones) but yield a higher level of productivity, when in use, and initially cheaper techniques of lower productivity. Now the fact that the former gives a promise of a higher level of productivity affords a prima facie reason for supposing that it will to this extent make more available for investment. Whether it will in fact do so depends, of course, on what happens to consumption when productivity rises (as it would depend also on the exact relationship between the higher productivity and the higher initial cost of the more capital intensive machine – after a point the effect of the latter may quite swamp the former). If any rise in productivity involves an equivalent rise in consumption, then the investment potential of the economy will not derive any benefit therefrom.

Similarly there may be no advantage in having the more labour intensive technique, if total consumption remains unaffected by the number employed: i.e. if a man’s wages in employment are no higher than what he would consume if unemployed.

However, if one defined a certain level of necessary consumption (allowing for the fact that this is probably higher when a man is at work then when he is idle), then a rise in production relative to this level of consumption is bound in some sense to enhance the investment potential, in the sense of the resources that could be devoted to growth and investment without infringing on the level of necessary consumption. I will only add that the notion being used here is essentially the same as that of ‘net revenue’ as employed by the classical economists (e.g. in Ricardo’s charge against Adam Smith that he constantly magnifies the advantages which a country derives from a large gross rather than a large net income) and the later notion of ‘taxable capacity’ in the theory of public finance.

As some of you are doubtless aware, there has been a discussion on this matter in choice of techniques in recent years in the economic literature, both here and in America, in which I and my colleague Dr. A.K. Sen have argued that (up to a point) the choice of higher capital intensity will increase the growth rate, even if it fails to maximize national income or output in the present (because of its labour saving influence on the level of employment). I don’t know how far this can be said to be a still open or settled question. But I think one can at any rate say that it is recognized that what is optimum from the growth standpoint may not be identical from what is optimum from the employment angle as is implied by the theory of factor proportions.

I should add, almost parenthetically, that some seem to have supposed that the conclusion of the static factor proportion theory (i.e. when labour is the plentiful factor) can be derived from a theory cast in explicitly dynamic terms: namely from the famous Harrod-Domar type of growth equations.* This represents the growth rate as the product of the savings ratio and the capital output ratio. From this it would seem to follow that with a given savings ratio, the growth rate would be higher the lower you make the capital output ratio. Hence, there would seem to be only one choice consistent with maximizing growth – the least capital-intensive form of production. This, however, only follows if one treats the savings ratio and the capital output ratio as independent of one another. For the reasons we have stated they are not so independent or at least not unless a rise in production and a rise in consumption per head are equi-proportional (or the latter increase is proportionately greater than the former).

During the First Five-Year Plan, as a matter of fact employment expanded very rapidly none the less – by about a double in industry and building alone; and though more slowly than this, it increased over the decade as a whole by three times. To start with, this was because the stepping up of the rate of investment increased employment, while the labour saving effect of new techniques was yet to show its impact. Even so it might look as though the choice was not labour saving enough and might have been more so, since the expansion in employment gave rise to pronounced inflationary tendencies, especially during the period of the first plan. However these were largely due to miscalculations of the extent to which labour productivity would rise during the period (with an unforeseen swelling of employment in consequence), and the tendency of enterprises (which retained a lot of financial autonomy in this period) to overbid one another for labour and thereby bring about unbudgeted increase in wages. For the rest, it was due to the drastic upward revision of output targets half way through the quinquennium which were largely at the expense of supplies for the output of consumer goods.

The question of the distribution of investment between the industrial sectors, in particular between the capital goods sector and the consumer goods sector, is really quite distinct from that of the choice of industrial technique, although in the past the distinction has been clouded by illicit analogies. Nevertheless there is a certain analogy between them; and the general argument we outlined above can be applied in an even simpler and more direct form in this case. Again it depends on what you assume to be the effective limitations on the investment potential of the country. It is arguable that the crucial factor is the size of its capital goods sector – let us say for short its productive capacity in steel, so that the rate at which it can grow depends upon the amount of steel it can turn out annually. You may question whether this is so for a number of reasons – e.g. you may think that labour (or perhaps raw material) is likely to be a more crucial bottleneck than production equipment, or you may deny that there can be any such limit so long as capital goods can be imported from abroad. But in a country with labour reserves, in the shape of rural overpopulation and one with limited possibility of foreign borrowing and narrow possibilities of enlarging its imports of machinery without adverse effects on its terms of trade (all of which characterized the USSR in the early Five-Year Plan period), it may well be true that its steel capacity and engineering capacity constitute the main bottleneck in development.

If this be so, then of course it will follow that the larger slice of any given investment potential you allocate to the capital goods sector – the more of the annual steel output you plough back to enlarge steel producing capacity – the larger the rate of investment that can be maintained in future years and hence the higher the growth-rate that the economy in the future can sustain. Investment in this sector will have a growth inducing effect, which investment elsewhere will not have. In the interests of attaining a high rate of growth it will be quite right, therefore, to give investment priority to the capital goods sector, as the USSR in the pre-war decade so markedly did, and as it has continued to do, if in a less pronounced degree, in the 1950s. And I would throw it out as a provocative suggestion, without intending to pursue it here, that this was a path or mode of development that a planned economy can follow, whereas for a number of reasons, I think it would be extremely difficult and an improbable course for an unplanned private enterprise economy in normal circumstances to follow.

Just one additional remark in the context as to what such a policy implies for consumption. It might seem that, since it can only be pursued at the expense of consumption and the standard of life, it is dependent on a particular welfare-judgement (or value-judgement to use that overworked fashionable phrase) equivalent to saying that the raising of per capita consumption has a low priority or to discounting the future at a very low rate, or some such. (A consumption of Rupees 100 today gives more satisfaction than consumption of Rs. 100 one year hence. Therefore it is normal practise in economics to discount future consumption. A high rate of discount implies you give more importance to consumption in the present and vice versa. J.M.) To put this matter in proper perspective, two things are to be made clear. Firstly, although under this method consumption will increase more slowly than employment and hence involve ceteris paribus a declining tendency in real wages, it does not involve any fall in total consumption (and not necessarily a fall in per capita consumption, if there has previously been unemployment). It is quite possible for consumption both in total and per capita terms to grow simultaneously with the devotion of the major portion of investment to the development of the capital goods sector. Only in the boundary case, where the whole of the investment in industry is allocated to the latter, will consumption fail to rise at all – and even then it will not fall absolutely.

Secondly, the date in the future at which consumption under this method (the high growth rate method) rises above what it would have been by this date under the alternative method may well be a much earlier one than is commonly supposed. If so, the conflict between growth and a rising standard of life only applies within a fairly narrow time-horizon. This is a question, of course, partly of the actual magnitude assumed. But if I may quote a simple estimate of Domar on the assumption about relevant magnitudes that does not seem at all far fetched, the time horizon in question may not extend further into the future than 12 to 15 years.** Domar (in one of his recent ‘Essays on the Theory of Economic Growth’) takes a simplified two-sector model, supposes a capital output ratio of 3, uniform for the two sectors, and a rate of net investment of 10% (1/10 of the National Income is devoted to net investment). He assumes that investment was exclusively devoted to the two industrial sectors of his two-sector model, and denotes by the symbol ‘g’ the proportion of this investment devoted to the capital goods sector. He takes first as a term of comparison the case where this coefficient is 0; all investment going directly into the consumer goods sector. Then the consumption will rise by constant absolute amount each year (but the relative growth rate will fall); over the first decade the consumption would rise by about a third, and by the end of the second decade by some 75%; but not until year 27 will the initial consumption level be doubled. With values of g between 0.2 and 0.5 the consumption curve will rise above the g=0 line by the end of the first decade; and by the end of the second decade it will reach a level between double and treble the initial level. With a very high value of 0.9 for g, consumption will of course scarcely rise at all during the first quinquennium and will remain substantially below the g=0 for the whole of the first decade; but after year 12 it will rise above it and by year 20 will have risen to some 4 times the initial level of more.

**E. Domar in his book, ‘Essay in the Theory of Economic Growth’ includes an essay on ‘A Soviet Model of Growth’. It is based on an article by the Soviet economist G.A. Fel’dman in the journal. ‘The Planned Economy’ in 1928. Fel’dman’s model divides the economy into capital and consumer goods industries based on the Marxian categories of Department I and Department II. His key variable is (g) the fraction of total investment allocated to the capital goods industry. The National Income (g) and consumption (c) as derived in the model have a constant term and an exponential in t (time period). As time goes on, the exponential will dominate the scene and the rates of growth of g and c will gradually approach g/v1 where v1 is the capital output ratio in the Department I capital goods industry. Hence, after a certain time period the higher the value of g the higher will be the growth rate of consumption.

What this type of consideration indicates is that for the growth of total output and consumption over the next two decades or so, how you distribute your investment fund – where you put it – can be more important than the initial rate of investment with which you start, upon which theories of development have traditionally focused attention. Implicit in this approach is the proposition that any given date there is not much one can do about the rate of investment or the average saving ratio. This is already decided for you within fairly narrow limits by the past history and past decisions. What one should focus attention upon is the so-called marginal investment ratio – to what purpose you devote the increment.

As regards the essential facts of Soviet development there is not now very much dispute. In the attached table I have tried to set out in quantitative terms the distribution of investment between the main sectors by comparison of what is known about this distribution in the USA, this country and (as an example of an underdeveloped country) India under her two Five-Year Plans of the past decade. These figures must not be taken too strictly – both the Soviet and the American cases are approximate estimates only (not mine, but those of Norman Kaplan and for the USA by Kaplan out of Kuznets) and the Indian ones are only those embraced in the Plans. These include targets for the private sector as well as the public sector. But I think the approximate order of magnitudes stands out clearly enough, independently of the possible margins of error in estimating these or the effects of price disparities etc. The salient facts are of course the higher proportion of total investment devoted to industry in U.S.S.R., and of this the much higher proportion devoted to the metal and metal production industries. (The latter proportion of course somewhat understates the proportion devoted to the capital goods sector as a whole, if one includes in the latter heavy chemicals and fuel and power; though on the other hand some metal goes into consumer goods and transport). In the USA this latter proportion is only attained in the war years (despite the importance of the motor industry as a consumer of steel). Outside wartime only India so far as I am aware even contemplated so high a proportion – namely the Mahalanobis draft plan for India, which was, however altered a good deal before it reached the final stage. The final Second Plan showed a smaller ratio for industry but a higher proportion of this in metals.

Kaplan’s conclusions are, as regards the USSR, that higher growth-rates are not explained primarily in terms of the higher rate of investment². The higher rate of increase of the industrial output in the USSR has been due, basically, not due to differences in the USSR-US ratios of investment, but rather to the differences in the direction of investment (Norman Kaplan, ‘Capital Formation and Allocation’, in ed. Abram Bergson, ‘Soviet Economic Growth’, White Plains, New York, 1953, p. 80).

1. Agricultural processing and light industry plus a little transport development preceding the growth of metallurgical and engineering industries and extensive power development.

2. Gross investment as a proportion, he estimated, was not very different from the USA. The net may have been a higher proportion of gross in the Soviet Union than in the USA.

Click here to return to the April 2002 index.