Since estimates of real national income may have several possible objects, one should not expect to find perfectly general solutions to many of the problems. The correct, or most appropriate, solution for one purpose may often differ from the correct, or most appropriate, solution for some other purpose. There would be general agreement that the particular purpose of the estimates may affect the choice of weights, as well as the deflation of what Is Included within the boundaries of economic activity.
But it may also influence the best method of making SE of the available data; and one should even bear in mind the possibility that, in certain cases, it may affect the choice of indicators. The purposes of estimates of the real national income or product seem to fall into two main categories. Such estimates may be needed in connection with (a) problems concerned with the potential or actual welfare or satisfaction derived by consumers and purchasers of final products; or (b) problems of productivity, and the effective utilization of resources, the emphasis In this case being on the producer, not the consumer.
It must be admitted that the usefulness of estimates of the 145 46 INCH MEAN D WEALTH real national income, in either connection, is very limited. They do not, certainly not without modification, and perhaps cannot provide general quantitative measures either of economic welfare, or of total output in the productivity sense. The estimates usually correspond with, or derive from, some generally accepted definition of the national income, which in turn implies a definition of economic activity.
Not all economic activities need contribute, while many activities that are not regarded as economic can contribute, to total production or economic welfare. According to the peck& or general purposes In mind, the concepts of productivity and welfare can also be defined In different ways. But the real national product does at least provide a starting point frown which a measure, or dudes, of one kind or the other may be earlier; Ana can Itself, odometer, De given a mover – to put It no enlarger – letter AT productivity or of welfare.
So much would, perhaps, he fairly widely admitted. But the practical implications of this dichotomy need to be considered rather carefully. And we must be prepared for some seemingly paradoxical results. It is worth, first, briefly considering the ratio?sale national of income aggregation. In order to obtain an aggregate of all the goods and services which together comprise the total national product, each item must he valued at a certain price, which begging a few questions for the moment – is the price at which it changes hands.
Some common unit of measurement is obviously necessary. But what is the Justification of valuing each item at this particular price? We would naturally use these prices if we merely wanted to describe the actual exchanges taking place at a particular time, or to study the interrelationships of different parts of the economic system. But there must be more to it than this. What is it that the exchange price is intended to measure?
From the point of view of the consumer the answer must, I think, be that the distribution of his expenditure over different commodities is assumed to be such that the marginal utility of each purchase is proportional to its price; that, in other words, the last shill?spent on any one item yields, broadly speaking, the same satisfaction as the last shilling spent on any other item. Frown the producer’s point of view the answer must, I imagine, be that the distribution of factors of production is assumed to be such that the marginal product of each is proportional to its price. J. L.
If this is the rationale of the usual process of aggregation it may be worth recalling the conditions to which it applies. It may be noticed at once that, if the aggregate is to be given a welfare connotation, different goods and services must be valued at their market prices, including the effects of indirect taxes and subsidies; but that, if the aggregate is to be interpreted in a productivity sense, the valuation of final goods and services must exclude indirect taxes and subsidies levied after the h a 1 stage of reduction – otherwise the proportionality conditions will not hold good.
It may be noted, in passing, that the marginal products and the costs of different factors will be proportional only if the costs are those actually incurred in purchasing one factor rather than another. An entrepreneur who is deciding on how much of different factors and materials (which embody factors used at earlier stages) to purchase will have regard to their market prices, including any indirect taxes, and will compare these costs with the value of the product, excluding any taxes that are levied on the product.
In the aggregate, therefore, all indirect taxes (less subsidies) that are necessarily incurred in the process of production, I. E. All except those imposed after AU the processes of production have been completed, must be included in the valuation of the national product. Marginal products are assumed, in other words, to be proportional to factor costs, not to factor rewards. L It is clear, therefore, that even in a closed system the two aggregates themselves will generally differ – by the net amount of any indirect taxes that are imposed after the final stage of prod?action. ? consider ten welfare aspect TLS. I nee assumption Tanat tense Is perfect competition In the markets for all consumer goods is perhaps not too wide of the mark. But the assumption about proportionality applies, of course, only at the margin. Although for each consumer the marginal utility of any item is supposed to be proportional to its price, its total utility certainly cannot be assumed to be proportional to his total expenditure on it. The first unit purchased may not give the same satisfaction as the tenth, or hundredth, or last unit purchased.
It follows that two ‘The term factor costs is commonly, and misleadingly, interpreted to mean factor rewards. A fuller discussion of this problem will be found in my article June on ‘National Income at Factor Cost or Market Pence’, Echo?Lombardi?oral, 1955. The only taxes of this kind, in the United Kingdom, are those which represent LLC price of a of a good or service (e. G. Marriage, dog and gun licenses). Aggregates are not strictly comparable if they contain widely different quantities of several commodities.
Furthermore, although the proportionality condition may hold good for a single consumer, it obviously cannot be assumed to hold good as between different consumers. For these two reasons, it is difficult to attach any meaning to an aggregate which covers all consumers, and all units consumed of every commodity. The most that it can do is to serve (I) as an index, for the Penrose of arranging different aggregates of commodities in order; and (it) as a starting point from which relatively small changes in the neighborhood of that starting point can, on certain assumptions, be compared.
For a single consumer whose tastes remain the same, small changes of apparently the same amount, measured from points which are not far apart, will have approximately the same significance, provided that substantial hanged do not occur in the relative quantities of the different commodities comprising the aggregate. A change from 98 to 100 may be assumed to meal pretty much the same as a change from 100 to 102, and will perhaps have something in common with a change from 120 to 122, but cannot be directly compared with a change from 200 to 202.
On the same assumptions, marginal changes in an aggregate covering all consumers may be compared in a similar way, and the aggregate can be used for purposes of ordering, provided further that there is no change in the distribution of incomes. The last qualification is important: the limited significance which can be attached to changes in the aggregate is likely to be very sensitive to changes in the distribution of incomes; indeed it is difficult to define what is meant by ‘no change in the distribution of incomes’ if the situation changes at all.
It is important to recognize that the significance of a national income aggregate, whether for purposes of ordering or for comparison of marginal changes, is bound up, firstly, with a given composition of this aggregate, although limited changes in composition are permissible, and secondly with a particular distribution of incomes. Such an aggregate is bound to reflect the distribution of incomes, not only because it is without body or substance unless a particular distribution of Incomes Is postulated, out also Decease ten Illustration AT Incomes affects ten prices of different commodities and hence the weights used in calculating the index.
These limitations apply equally to the 149 welfare and to the productivity aspects of national income. If substantial changes take place in the distribution of (real) incomes, or if movements in the output of different commodities -especially commodities which are known to have largely efferent markets – are widely different, a general measure of changes in real income can have little meaning. In these circumstances we can hope to obtain valid indices only for particular classes of the community which have not experienced such changes.
A general measure embracing everyone has validity only if everyone has experienced the same change in real income. It is obvious, indeed, that a general measure, covering all members of the community, involves some assessment of the relative importance of each member. Why should the rich (poor) be given greater weight simply because they are rich (poor)? And yet who is to say what should be the relative weights for different people?
Should everybody – men and women, healthy and infirm, young babies and octogenarians with one foot in the grave, spinsters and profligates, scholars and fools – should they all be given the same weight (they are not at present) simply because no good reason can be found for giving them different weights? Clearly the assumptions made – and assumptions of this kind, whether stated or not, are often made-can very materially affect the results. The interpretation of national income aggregates in terms of productivity rests on assumptions which re closely analogous to those already considered.
All industries are assumed to enjoy perfect competition, in the sense that there are supposed to be no restrictions on the entry of new producers. I t is also assumed that there is perfect competition between purchasers, as well as between sellers, of all factors of production. Proportionality is assumed between the price and the marginal product of every factor; but of course the marginal product of a factor, or combination of factors, may vary with the level of production: constant returns to scale are unlikely to prevail at all levels of output.
It follows, again, that only marginal changes in the aggregate, measured from points in the same neighborhood, are directly comparable; and that changes in the aggregate lose their significance when the changes in the different commodities comprising the aggregate are markedly different. As a measure of changes in productivity the index is therefore bound to be influenced by the extent to which different industries L 150 INCOME A N D WEALTH or factors enjoy monopolistic or monopolistic positions.
If the degree of monopoly were ten same In all Institutes, or IT changes In ten output o T Deterrent commemorate ere approximately the same, there would be need to worry. But when changes in the output of different industries show substantial variations, the inclusion of monopoly earnings in the weights for each industry is likely to distort the general picture. This type of distortion may not often be serious, but is likely to become important in periods of radical change. All this sounds rather discouraging. But the position is not, perhaps, as bad as it appears from a recital of the underlying assumptions.
It would be a very tops-truly world in which the assumptions were so wide of the mark as to destroy the value of all national income estimates. It is as well, nevertheless, to recognize their full limitations from the outset. The qualifications that have been mentioned above apply, it insist be remembered, not only to comparisons of the national product at different points of time within a given country, but also, and indeed with much greater force, to comparisons between the national products of different countries.
To assume that competition is the general rule in a single country is one thing. But to assume a similar degree of competition between the products of industries, or between the factors of production, situated in entries that are separated by oceans, that have different climates, traditions, customs and, maybe, different industrial systems; and to assume even broad similarity in the tastes of consumers, or in the distribution of incomes, in different countries is to place a very severe strain on the imagination.
It follows from what has already been said that a measure of the change in the aggregate should be such that, if marginal quantities of some commodities, or factors of production, arc substituted for marginal quantities of others, the value of the aggregate, whatever the point of view adopted, should be unaffected. This, again, shows that both the aggregate itself and a measure of the change may be different, according to whether we are thinking of productivity or of welfare.
If, however, there are large-scale substitutions the whole method breaks down and it becomes impossible to measure the change at all accurately. It is useful to remember these principles when a sudden 1. L. NICHOLSON 151 change occurs in the composition of the aggregate and it becomes necessary, unless all attempt at comparison is abandoned, to resort to some arbitrary method of bridging the gap.
When, for instance, there is a fundamental change in the character f a product it is possible that a measure of the ‘change’ in output Welch seems appropriate whew considering welfare may not be appropriate when considering productivity. Take the familiar example of beer. Suppose that water is added to beer, so that the number of bulk barrels increases relatively to the Illumine of standard barrels. There has been a fundamental change in i.e. quality of the product and it is necessary to find some measure of equivalence between the two kinds of beer.
From the welfare point of view we need to decide how much of the new, weaker beer gives he same satisfaction as a given quantity of the old, stronger beer (or vice versa). From the point of view of productivity, we need to decide how much of the new, weaker Deer cool a nave Eden producer Walt n ten resources wanly went Into producing the old, stronger beer (or vice versa). The answer in the one case may approximate to the number of standard barrels, and in the other to the number of bulk barrels. At all events the two measures may be different.
It is also worth considering, against this background, the suggestion that the units in which the various goods and services are measured should always be determined by reference o the market contract. At 31st glance this seems a sound enough general principle; but it can lead to rather curious results. Consider its application to the case of doctors. The productivity of doctors would then depend on the form of the institutional organization which arranges medical attention for patients whether privately, or through a national health service.
In an epidemic, when doctors in the health service are called on to do more work, the index would not show any rise in output. Even from the welfare point of view it seems unduly Riemannian to argue hat because the consumer, under a national health service, pays the same as before, therefore he obtains the same quantity of services as before. It seems more reasonable to assume, even from the point of view of welfare, that when there is an epidemic the public obtains a greater quantity of the doctors’ services; and that the price in effect is reduced.
Doctors will be prevented, during an epidemic, from doing the work they would otherwise be able to do in their spare time, which we may 152 INCOME AND WEALTH suppose they spend in writing their autobiographies. The output of autobiographies ill fall and, on the market contract principle, there will be no compensating increase in the output of medical services supplied under the national health scheme. Doctors may be more or less efficient at writing their autobiographies than at doctoring.
But the proposition that their autobiographies are far more valuable than the extra medical services which they provide during an epidemic would not, I think, command general assent. Suppose that, in a very severe winter, consumers spend more than they usually do on clothing. No one suggests that the extra output, thus induced, would not be regarded as part of the national product. Or take the case of life assurance. What the consumer purchases -life cover – is the same in each year that his policy is in force.
But the work of the insurance company is much greater in the first year, when the policy is drawn up, and in the last year, when the claim is met, than in the intermediate years. This is, indeed, recognized in the indicators that are generally used to represent the output of the insurance industry, which take into account the number of new policies and the number of claims met, as well as the number of policies in force. The allocation of output between deterrent years may also depend, therefore, on which of the two problems, to which we attach the labels productivity and welfare, we have in mind. There is a further widespread illusion which ought to be dispelled. The main test of a price or quantity index should be whether it stands up to the demands made on it in actual, practical problems; whether, that is to say, it will behave, in given conditions, in a way that satisfies the requirements wanly are Inherent In ten nature AT ten problem. It Is sometimes TA for granted, without proof being offered, that a quantity index and a price index of he same group of commodities should be such that their product equals the (unique) ratio of total values.
Yet the possession of this property is not, in practice, felt to be a necessary attribute of either price or quantity indices; nor is there any reason in logic why price and quantity indices should *The distinction between a measure which relates to changes in productivity elaborated by ?emulsion Economic Papers, 1950). Valuation of Real ?commemoration'(Oxford J . L. NICHOLSON 153 have this property. L The change in the value of a group of commodities can in fact he analyses into (I) the effect of changes in prices, the effect of changes in quantities, and (iii) the interaction between changes in prices and changes in quantities.
There is no reason to assume that the price and quantity indices should be such as to make the interaction term equal to zero. 1 . THE FUNDAMENTAL NATIONAL INCOME AGGREGATES 1 Something should be said, before proceeding further, about the main national income aggregates, and in particular about the effects of foreign trade. We are concerned with the following three categories: (A) gross domestic product=constitution+ home investment *exports–imports; (B) gross national income?gross domestic product+net (property) income from abroad; (C) gross domestic resources?consumption+home investment.
Each of these categories is important and has its own uses. The difference between the gross national income and the gross domestic product is simply net income from abroad. This is all property income since it is part of the gross domestic product of some other country. The difference between gross domestic product and gross domestic resources is exports-imports?net investment abroad-net income from abroad+net unilateral transfers abroad. The difference between gross national income and gross domestic sources is net foreign investment (I. E. Net purchase of foreign assets)+net unilateral transfers abroad.
The importance of the first two aggregates is well recognized. The third – gross domestic resources -provides a measure of the total resources actually used during the year, either for investment or for consumption, and includes any resources ‘This was one of the wholly arbitrary tests which Irving Fisher introduced in that lamentable example of misplaced energy, ‘The Making of Index Numbers’. The book was properly castigated by Yule Rainproof talc RoyalStatirticnlSociely, 1923, p. 424) and Bowled (Economic Journal, 1923, p. 0) on its TLS appearance, out ten criticisms nave since apparently Eden Torment.
It was plotted out, for instance, that if two sets of rectangular areas are compared, the average difference in length multiplied by the average difference in width does not equal the average difference in area, except under peculiar conditions. 154 NCO M E AND WEALTH obtained as income from abroad or from foreign disinvestment. Since it includes all imports and excludes al 1 exports, it automatically reflects any changes in the terms of trade. We can conceive of changes in both (A) and (C) being measured in real terms. But the measurement, in real terms, of changes in (B) presents a real conceptual difficulty.
Income as such does not, strictly, have any real (physical) counterpart. Until income is spent, it cannot be identified with real goods and services. Of the three possible aspects of national income product, income and expenditure – only two – product and expenditure – can be identified in real terms. Any attempt, deforest, to measure national income as such in real terms is bound to be very arbitrary. It involves, in particular, the deflation of net income from abroad, the conventional method being to use the index of the prices of imports.
But there seems no good reason to use the prices of imports, in preference to the prices of exports or of domestic expenditure; income from abroad can be used either to finance additional imports or to reduce exports, or consequently to male use of more home-produced goods, or even to increase foreign investment. Since it is impossible to know how this income is used- it is not earmarked for any particular purpose – deflation by any of these indices is scarcely better than multiplying by the number of snakes in Ireland. Any adjustment for changes in the terms of trade is also bound to be arbitrary and conventional.
The adjustment usually consists in replacing exports by imports of the same total (current) value, although some exports may represent property income payable abroad and some Nina be balanced by foreign investment. The gross domestic product and gross domestic resources can each be valued either at constant market prices, or at constant factor costs. It may appear, at first sight, that the gross domestic product has more meaning if valued at constant factor costs; and that total domestic resources has more meaning if valued at constant market prices.
The method generally followed for the valuation of imports at constant market prices a necessary stage in estimating the gross domestic product at constant market prices – is indeed somewhat arbitrary when some imports are subject to heavy customs duties. Conventionally, customs duties are not included in the valuation of imports, on the grounds -presumably – that all taxation is a matter which is purely internal to the country concerned. And yet the actual 1 55 cost of those I?inputs, measured against the possible alternative use of home- produced goods, clearly includes the customs duties.
But, apart from this type of difficulty, there appears to be no reason, in principle, why the gross domestic product would not be valued at constant market prices, or gross domestic resources al constant Doctor costs I nee Locomotion Detente product Ana tied to the distinction between productivity and welfare. J. L. NICHOLSON 111. ESTIMATION BY TWO ROUTES resources n In practice there are only two routes by which the gross domestic product (or gross domestic resources) can be estimated: one proceeding from the production side, the other from the expenditure side.
Although the two routes should end up at the same aggregate, they proceed from quite different directions, at almost no point do they nonacid, and detailed comparisons between the individual components of the two estimates are therefore, in the nature of the case, impossible. The production method consists in summing the net contributions of all industries to the gross domestic product (so-called net output, including depreciation), valued at base year prices.
All industries are included, irrespective of whether they produce raw materials or intermediate products for use by other industries, or goods for final use in investment or consumption, or – as often happens -some combination of these. The expenditure method, on the other hand, consists in estimating all the monuments of final expenditure on goods and services consumption, government expenditure, investment in fixed capital and stocks, and exports – at constant prices.
No attempt is made to determine the contributions of different industries to any particular item of final expenditure. If we had complete, accurate and detailed information about all transactions, the two methods should, in theory, lead to the same answer. But without this information – without in fact a full Limonite matrix for two or more years – any detailed reconciliation of the two sets of results is impossible. There are, in fact, various reasons why the two methods must be expected to give different results.
In the production method, we have to measure the change in the net contribution of each industry to the gross domestic product. In practice, the change in 156 net output is generally represented by a measure which relates to gross output, including the contributions of earlier stages of production. Any change, therefore, in the ratio of net to gross output would result in the measure being, to that extent, inaccurate. For example, if a firm decides to install its own electricity generating plant, he resulting increase in its net output would not be reflected in the measure of gross output.
Changes in the ratio of net to gross output can also be caused by changes in the amount or quality of workmanship incorporated in the product; by economy or extravagance in the use of materials; by changes in the technical processes of production; by cabbages in the degree of integration of industry not already reflected in the available statistics of gross output; or by changes in selling costs or payments Tort toner services per unlit AT output. N ten expenditure sloe, also, it is often necessary to make assumptions and approximations. Arbitrary assumptions, for instance, usually have to he made about the proportions of certain items of consumption that are chargeable as business expenses, and which are not, therefore, part of personal constitution. Changes in quantities are sometimes measured directly and sometimes indirectly from money expenditure and estimated cabbages in prices.
In either case it is seldom possible to be ‘sure that full account has been taken of changes in the quality of the product; and there may also be changes in the amount, or quality, of services purchased with the product, the accurate agreement of which is virtually impossible. More important, there is usually 110 direct information about physical changes in stocks, which therefore have to be estimated indirectly, for the United Coinciding at least, from estimates of the value of stocks and rather precarious estimates of ‘stock appreciation’.
Further work has since been undertaken, a new set of estimates has been compiled by the expenditure method and a virtually independent set of estimates has been compiled by the production method. Both sets of estimates relate to the gross domestic product of the United Kingdom. Two estimates were at first made by each method, one at factor cost and the other at market prices; but the allocation of indirect taxes and subsidies between different industries, in the production method, was necessarily somewhat arbitrary and has not since been reattempted.
All the estimates have been expressed in terms of 1948 prices, as this was the most recent year for which we had the full results of a Census of Production. The estimates have been published in a summarized form in recent issues of the Blue Book. = Importance was attached to keeping the estimates derived by the two methods as independent as possible. Co-ordination has resulted in the elimination of discrepancies due to differences in edition (e. G. What is included in the field of economic activity); the adoption as far as possible of the same principles for measuring changes in the output of final goods and services; the derivation of all base year figures and weights for both calculations from the same source, namely the estimates for 1948 given in the Blue Book; and the avoidance of arbitrary differences in the use or treatment of the available data. Where, for example, we were measuring changes in the same thing and there was a choice between two or more possible indicators (e. G. Physical quantities and deflated values), the same method was used in both sets of estimates.