Article Shared by

Allocation of resources is one of the central problems faced by all economies. Establishing criteria for investment of a nation’s resources are crucially important for poor countries.

One of the most traditional of the investment criteria uses the capital- output ratio as a key economic indicator. In this view, investment will be forthcoming or — what is the same thing — the least investment per unit of output. Economic decision makers, private or public, would seek activities in which the capital-output ratio is lowest and concentrate their resources there.

In the 1930s and 1940s, when the Harrod-Domar model was becoming increasingly popular, it was thought that one could use the capital-output ratio as the basis for choice of investment programme. If G = s/v where g is the growth rate and s the savings ratio and v the incremental capital-output ratio (ICOR), the lower the v the higher the g. So, the first approximation to a criterion for investment choice was a project’s ICOR.

These common-sense notions, however superficial they may prove to be on more detailed examination, have a secure place in the world of economic theory. The Harrod-Domar model focuses on the capital-output ratio as one of the two central parameters in determining the rate of economic growth.

The capital-output ratio is often used as an investment criterion and plays a key role in the Harrod-Domar model. For most purposes, we use the marginal or incremental capital-output ratio (ICOR) rather than the aver­age capital-output ratio (ACOR).

We want to know how much additional output is associated with a given addition to the capital stock, i.e., with a given amount of investment, for an additional amount of investment. Let us suppose that there is no difficulty in defining national production or capital formation. Is there then any problem in calculating the ICOR? The answer, as one might expect, is yes.

First, suppose we have three observations on capital and output, as in Fig. 13 (a). Is it possible, by drawing a line between them and extending it to the origin, to conclude that the ICOR is the same as the ACOR, and steady? It is not. To arrive at this conclusion is possible only if other things are equal.

If technology has been changing, as in Fig. 13 (b), the ACOR and the ICOR may each be steady, but widely different. Or, technology may be un­changed, as in Fig. 13(c) but with very different ICORs in each period and the same ACOR. The model assumes other things equal; in the real world, other things do change.

Second, should we take ICOR gross of depreciation or net? It evidently makes a considerable difference since, unless D or depreciation is very small.

Assume there are two projects, each costing Rs.100 but of very different lives: one lasting twenty years, the other four. Suppose that, there is no ambiguity about what depreciation should be charged and that straight-line cost is the appropriate basis, physically as well as financially. Then as the Table 2 shows, the gross ICORs favour project B, the net ICORs project A.

The answer to the question can be provided by adding a qualifying clause, “It all depends”. Here, it depends on whether the capital structure is likely to fairly stable or whether transformations in the structure of the economy are frequent.

With stability, it is sufficient to deal net output, since depreciation allowances are not needed to shift capital to other sectors. If there is a large possibility that capital will be shifted into other industries, however, as transformation occurs, then gross production is the most rele­vant concept.

Another aspect of this distinction is raised by the question whether one ought to consider the capital or the current cost of the capital input in choosing between two competing investments. The longer the life of the capital, the smaller the depreciation charge in any one year and the lower the capital cost, calculated gross. But, this type of reasoning may lead into danger, since it favours long-lived capital.

A number of other significant questions are raised by depreciation. For one, there is no fixed scientific basis for calculating depreciation for a given capital asset. In most instances, depreciation is charged fairly uniformly through the life of the assets, either in a straight line or with constant percentage formula.

Physically, however, some capital tends to wear out more rapidly at the end of its life than in the beginning. It is possible, therefore, in a growing economy, to reinvest depreciation allowances from recent investments in new capital formation and rely on the greater productivity of the economy later when development has proceeded some distance to make good the physical exhaustion of capital.

If an economy is consistently growing, moreover, straight-line depreciation will continuously provide more depreciation allowances than is needed to make good physical wear­ing out. But this is a simple property of geometric growth.

Third, there is the drastic oversimplification involved in comparing this year’s investment to this year’s increases in output. This is a helpful device but analytically unsatisfactory. Some additional reality can be introduced by a lagged model, in which inputs in period t lead to output in period t+1 and inputs in period t+1 yield their outputs in period t+2.

Says Kindleberger, “When an economy undertakes all three types of investment — point-input, point-output; continuous-input, continuous- output; and point-input, continuous-output — the capital-output ratio that relates this year’s output to this year’s investment is evidently wide of the mark.”

In fact, in a system of instantaneous production, or with a fixed lag, it is appropriate to take account of the capital-output ratio. When, however, output is received in a different time sequence, however, the investment problem becomes one of comparing the cost of a given input with the present value of its future output.

There may also rise a question whether to use a simple or a compound interest formula in comparing the present value of two investments with different time profiles of income.

Fourth, in a disaggregated model — one that examines individual pro­jects in the various sectors of the economy — the task of associating specific outputs with appropriate inputs becomes still more difficult as a result of complementarities and external economies.

Manufacturing and distribu­tion may have measured ICORs as low as 2, while electricity and railways may be as high as 16. But manufacturing has the low ratio only under marginalist conditions or if it is assumed that markets have already been linked by transport, that materials can be cheaply assembled, and that energy is available as needed at constant cost.

To the extent that each sector uses intermediate products and services from other sectors and that big (supra-marginal) projects are the most crucial ones for the development process, a focus on sectoral capital-output ratios as guides to investment decision has limited value.

One obvious objection which can be made to the method of calculating the aggregate capital requirements of the LDCs on the basis of a stable overall capital-output ratio is that this implies the assumption of constant returns to scale for the expansion of the economy as a whole.

This assump­tion is justified by the mature phase of the advanced countries to which the Harrod-Domar growth model is intended to apply. In LDCs, agriculture predominates in national production and is subject to diminishing returns. In Myint’s language, “We cannot apply the assumption of a stable overall capital-output ratio to the LDCs unless we can show at the same time how to counteract the general tendency towards diminishing returns.”

In the language of the growth models, the assumption of a stable overall capital-output ratio for the LDCs requires not only that a continual stream of innovations is taking place, but that they are of a land-saving character, enabling a progressive substitution of capital and labour for natural re­sources.

In spite of this objection, the concept of an overall capital-output ratio has enjoyed considerable popularity. It is often defended on the ground that it offers a useful basis for testing (1) the consistency of the desired target rate of growth in national income, and (2) the available resources of a developing country. But, in reality, we cannot get very far in testing the economic development plans of a country unless we are prepared to go behind the overall ratio into the structural factors which determine it.

The national income or output is not a homogeneous thing but is made up of different goods and services, each having widely varying capital-out­put ratios. The sectoral capital-output ratios are very high for some items, notably transport and communications and public utilities. Next, in order of high capital-output ratios, come housing and capital-goods industries.

Manufactured consumers goods industries together with other distributive and service industries generally have lower capital-output ratios. The capi­tal-output ratio in the agricultural sector of the LDCs is generally likely to be low, although some of the big irrigation and river-valley projects require vast sums of capital.

Characteristically, the expansion of agricultural output in these countries depends not only on capital inputs such as fertilizers and improved equipment, but also on improvements in technical knowledge, marketing credit, or land tenure, for example, which are not directly re­flected in the capital-output ratio.

Now, the overall capital-output ratio is nothing but the average of these different sectoral capital-output ratios weighted according to the quantities of the different goods and services which are to be produced. Thus, before we can calculate the overall ratio we must specify the proportions of the different constituent items which are to make up a proposed rate of increase in the national output.

But, this barely scratches the surface of the problem of testing the consis­tency of an integrated economic development plan. For one thing, the target figures of increase in outputs of various items are not given independently of each other. Many of them are required not only for final consumption but also as intermediate goods or inputs in the production of other items.

In testing, therefore, the consistency of the target figures of items and the resources available for them, we must take into account not only the direct requirements but also the indirect requirements of capital. Furthermore, these complex input-output relationship should be tested not only for a given year, but continuously over the whole period of the plan.

This means that for each of the intervening years, say, during a five-year plan, the rates of expansion of the different sectors must be phased so that they dovetail into each other, without any sector lagging behind their concerted time-table and holding up the others.

For, if this happens, shortages and excess capacities will develop and this will alter the effective capital-output ratios in the sectors which have gone out of alignment with the general plan. By the time we have gone through the consistency of an integrated development plan in this way, it does not help us much further to ‘sum up the whole thing’ in the form of an overall capital-output ratio.

Yet, a great deal of importance has been attached to this ratio. Perhaps, the proximate reason seems to be that it offers a convenient shorthand basis for making out the case for increasing economic aid to the LDCs.

Despite these objections — assuming other things equal—when they are not, inability to decide between the gross and net measurements and the difficulty of imputing outputs to inputs, either in time or by sectors, the ICOR is widely calculated and used as device for projecting the overall investment requirements associated with plans for the expansion of out­puts from estimates of investment.

Says Kindleberger “Its predictability as suggested is weak. In the short run, it is markedly variable. But over longer periods of time, averaging the annual marginal rate appears to produce meaningful overall results. This is in large part, no doubt, the result of the law of large numbers in which opposing movements cancel out. The ICOR has, in fact, been remarkably similar in a large number of countries, averaging somewhere near 3.3:1.”