The far reaching effects of this failure to adjust for differences in price bases may be illustrated by the perennial bundle of problems arising in measuring the share of government. Unlike business enterprises and individuals, governments sell most of their services on a compulsory basis—setting the volume, composition, and price by legislative fiat. On the buying side too, governments are not subject to the same rules as private firms or individuals, incurring deficits without the economic penalties attached to them in the world of private enterprise. Government is not unique in this respect—some public utilities also exercise compulsion in their selling and are likewise, for brief periods, immune to the fateful consequences of deficits. But the far higher degree to which government possesses these peculiarities than any other sector of the economy puts several stumbling blocks in the way of national income measurement.
First, to differentiate between intermediate and final products of government activity is exceedingly difficult. Were enterprises free to buy and pay for government services as they do coal or services of legal firms, their payments could be considered the value of intermediate services rendered by government. But the government exacts a compulsory price in the form of taxes; and in the organization and accounting of its activities does not separate services to enterprises from those to ultimate consumers. Likewise, were individuals free to buy and pay for services of governments as they do those of physicians or of domestic servants, their payments could be considered the value of the final product of government activity. But the element of compulsion and the difficulty of a functional classification of government activity exists here also.
Second, are the prices paid by government for productive factors engaged under its auspices similar to those paid by business firms to their employees, capital, etc.? Third, should the government be treated like a corporation, and its undistributed profits and losses considered compensation to the factor of enterprise engaged in the government 'industry'? The mode of settling these various problems will affect the size of national income, its composition, and changes in it over time. As a greater or smaller allowance is made for intermediate services of government, there is a greater or smaller allowance for duplication, with corresponding effects on national income as a net total. If we adjust for the differences between the prices government and business firms pay for the things they buy, we will get one national income total; if we do not, another. If we recognize the validity of the concept of net government savings, national income will be larger when the savings are positive and smaller when the savings are negative. And there will be corresponding effects on the share of government in national income, as well as on temporal changes in both.
The arbitrary solutions given these questions in the national income estimates prepared at the National Bureau in recent years were, in our judgment, the least unsatisfactory for what is essentially an insoluble problem. It was thought that because of the joint character of most government activities (e.g., national defense, provision of justice), services by government to business enterprises cannot be differentiated from direct services to individuals. The simplest way seemed to be to take paymeans by business firms to government as measuring intermediate services by the latter, thereby including in the calculation of national income (at the payments level) net business profits after all taxes. Likewise, direct
payments by individuals to government were taken to measure final services, i.e., services by government to individuals as ultimate consumers. This meant including in national income (again at the payments level) all payments to individuals, gross of direct taxes paid by them.
Differences between the prices paid by government for the goods it purchased and those paid by other sectors of the economy were not adjusted for, except in the case of war products in National Product in Wartime. This is in line with the failure to make similar adjustments for price differentials among other sectors of the economy (e.g., between pricing on farmers' and on urban industries' markets).
Finally, it was deemed consistent to treat government as if it were a business corporation, and allow for its net savings or losses. This is tantamount to measuring the total value of government services by the taxes paid for them, rather than at cost—a basis exactly parallel to that applied in valuing the total output of other industries in the economy.
These answers leave plenty of room for doubt and contention, and as the character of government activity changes and its functional characteristics are better analyzed, a less arbitrary approach may become feasible. But the essential difficulty will remain, viz., governments (and related semipublic sectors) and the private business sectors (both firms and individuals) do not and cannot operate under the same rules, any more than do or can the business and what may roughly be called the family sectors. The difficulties in handling the latter are reduced by excluding it almost completely from national but national income includes both the private business and the public sectors. The fundamental difference in the principles on which these sectors operate means that some arbitrary decisions will always be called for in order to put the two together—by applying the private market or public economy base to both, or by devising sonic common denominator.
Though the bearing of differential pricing upon international comparisons of national income is obvious, we cannot end without stressing it. Differences in price bases among the several sectors of a nation's economy reflect differences in institutional characteristics and principles of operation. Even an industrially advanced nation with a democratic social system cannot be described as virtually one free business market. If the family and its economic life is omitted, there is still the farm sector, the public utility, the government—each with its peculiarities. When we consider more than one national economy, differences in the relative importance of various sectors within each are perhaps most prominent. Failure to adjust for price differentials among the several sectors in each country means that international comparisons cannot be made properly unless it is assumed that the relative distribution of national income among these various sectors as well as the extent of the intra-national price differentials are similar. Obviously, such an assumption is not valid. Hence international comparisons of levels, composition, and even changes in national income cannot be sound until some advance has been made toward adjusting for price differentials within each country. The customary basis for international comparisons of national income, exchange rates or market prices for a few commodities, is obviously so crude that only the biggest differences can be deemed significant.