Another simple way to deal with the problem would be to accept market prices at their face value. One good reason for confining national income to products of market-bound activities is the availability of market prices, which surely approximate the relative economic significance of various goods more closely than any physical property. If we could assume further that market prices reflect the relative economic significance of different categories of goods consistently as between two years or within the same year among distinctly different types of market, consistent, that is, with the input of real resources such as labor and the satisfaction consumers derive from the product, the problem would vanish. All we would have to do would be to find market prices for any given year, weight the various items that enter national income, and add.
The problem exists precisely because market prices are not consistent from period to period, and at a given point of time in different types of market. Prices of a physically identical commodity change from year to year. Even if the change is due to shifts in real costs or tastes, not to a general price inflation or deflation, we cannot use changing price totals and derive a comparable national income series. We still have to use prices at some point of time, i.e., the weights expressing the relative economic significance in a fixed base period. But most price changes over time are due to the general imperfections of our money mechanisms which create widespread movements in the price level. And society itself acknowledges this defect in trying to improve the mechanism and to curb the more extreme variations or trends.
Nor are market prices a consistent valuation base at a point of time. For example, a surgeon charges his private patient $1,000 for an operation he performs free in the hospital clinic. Such differentials in prices of physically identical commodities and services abound. While they may correspond to differences in the value of the monetary unit to would-be purchasers, a fixed base is as essential for different categories of goods as for changes over time.
These two aspects of the valuation problem—changes over time and differences in pricing bases among several sectors of the economy at a given time—are distinct. Treated differently in the historical development of quantitative economics and of national income measurement, they are kept separate even in the brief discussion below.
The problem of adjusting for changes in prices over time has long been recognized; and the availability of series for many groups of commodities has facilitated statistical solutions. Though difficult questions
remain, their character is well known. All we need stress here is that to adjust national income for temporal changes in prices is far more difficult than to adjust any one component. Because of the very comprehensiveness of national income totals, they cannot be estimated directly in terms of quantities, then weighted by some constant prices; and the 'deflation' of the current price totals for changes in prices over time requires price indexes of a comprehensiveness virtually impossible. It is difficult enough to get comparable prices for groups of commodities fairly standardized in quality, such as steel or cotton of specific grades. But when one visualizes the variety of goods included in national income, and the importance in it of services or commodities subject to rapid qualitative changes, the difficulty of 'deflating' the current dollar totals precisely seems insurmountable. National income totals in constant prices are consequently rough approximations, though always measuring net product more accurately than unadjusted totals in current prices.
Although the second aspect of the problem—the differentials in price bases among the various sectors of the economy—has also long been recognized, few statistical attempts have been made to solve it, chiefly for lack of proper data—prices of identical goods in different markets; e.g., prices of meat of the same grade to consumers in lower and higher income neighborhoods; or prices for an identical service paid by governments and by corporations. Rich though our price data are, comparison is possible for merely a few items. Precisely because of the different characteristics of markets, the share of common, identical goods is likely to be small in each, and that of different (though comparable) goods hence, for just the categories for which price differentials may be large, prices cannot be compared except through analytical experiments designed to render two qualitatively different goods truly comparable.
Whatever the reason, national income totals, even in constant prices, are not adjusted for differences in price bases of the component sectors. Consequently, the weights assigned components at any given period may be affected by differences in price bases; i.e., the components would be weighted differently were an attempt made to put them on a comparable price basis. And whatever the relative weights of components characterized by different price bases change materially, the movements of national income in constant prices are also affected.