along with the percentage differences. In 1935 prices, these incomes would be approximately 20 billion dollars smaller, the same as, and 20 billion dollars greater than, the consumer income level of 1935-36.

The most striking feature of table III is the apparent lack of any indication of a limit to any of the wants reflected in the items. At the highest level of expenditure, for every one of the separate items there would be a great increase in expenditure over the lowest. The expenditure for food, the item showing the smallest rate of increase, would be 44 percent greater with the two-thirds greater level of total expenditure. This 44 percent increase in expenditure on food is particularly significant since it is so often stated that the demand for food is limited. Undoubtedly part of the increased food expenditure would go into an improvement in quality-more milk, meat, and fresh vegetables -and only part into an increase in quantity, but in either case it would call for greater farm production for domestic consumption. For each major category of consumption, the structure of wants appears to be such that a big lift in consumer expenditures would create a greatly increased domestic market for every broad class of products.

From a study of this table it is also possible to add a new dimension to the structure of wants. While the demand for all broad classes of foods and services expands with larger national income, it is not a uniform expansion. Greater buying power, with price and other market conditions not significantly changed, is likely to produce a smaller than proportionate expansion in food and housing expenditures and transportation other than automobile. The largest proportionate increase indicated is in expenditures for automobiles and for education. Furnishings, clothing, medical care, recreation, and other miscellaneous items also show a much more than proportionate increase. Increases in the remainder of the items listed would be more nearly proportionate to the increase in total expenditure.

As in the case of individual differences in buying power, the greater the total buying power the less the proportionate emphasis on expenditure for the basic necessities and the greater the emphasis on expenditure for better clothing, automobiles, recreation, and private education.

An actual change in the level of national expenditure would be unlikely to be reflected in just the proportions shown above, because of the adjustments in expenditure habits involved in the transition from one level to another. There is little specific data on the transitional

With the actual variation in consumer income experienced in recent years this expansibility in the demand for food is not apparent, as farm production has been maintained at a relatively constant level, and food prices have dropped markedly, in relation to other prices, with declines in consumer income. This has allowed consumption to be maintained at a relatively constant level.

effect of a change in income on the direction of current expenditure. Probably its most important effect would be on the purchase of durable and semidurable goods, whose purchase can so often be postponed. When the income of a family that has become adjusted to a $2,000 level of spending is suddenly reduced, the family is likely to carry over reserves of durable and semidurable goods which can contribute for a time to the family living without involving current expense. Clothing will be worn longer and the family automobile will be tinkered with and repaired instead of being replaced by a later model. Gradually as this extra reserve of durable and semidurable goods is used up, the family will have to adapt its living more nearly to a balanced pattern of expenditures. But in the transition period, a smaller proportion of its expenditure is likely to go into durable and semidurable goods than either before the change in income level or after adjustment has been made to the lower level.

Similarly, an increase in income, that represents a return to levels to which a family was accustomed at some earlier date, is likely to involve a greater than balanced expenditure on durable and semidurable goods. Even a sudden expansion of income to unaccustomed levels appears likely to lead to the sudden acquisition of durable goods to correspond with the higher level of living made possible by the increased income. Data are not available to show the nature and magnitude of these transition changes in the pattern of expenditure, but they are likely to be of sufficient importance to the structure of wants to warrant further research. The evidence of the actual changes in consumption of durable and nondurable goods from 1929 to 1932 point in the direction indicated above.

Two other major dimensions of the structure of wants require exploration-the effect of price relationships on the structure of wants and the trend of change as wants shift through time. On neither of these points are data available for the major categories of consumer expenditure. Pioneer work has been done on specific price relationships in relation to consumption, especially for agricultural products, and on the trend in consumption for a number of specific items. With respect to both these aspects the data are insufficient, and the delineation of these dimensions of the structure of wants must wait upon future research.

The main aspects of the structure of wants can now be summarized in terms of the major items of expenditure as follows: Food, clothing, and shelter are wants finding expression in the largest body of consumer expenditure, accounting for nearly two-thirds of the total.

See Bureau of Agricultural Economics, Agricultural Economics Bibliography No. 58, October 1935; The Theory and Measurement of Demand, by Henry Schultz, Chicago, 1938.

At lower income levels the want for food requires a larger proportion of expenditure while at higher income levels food expenditure is relatively smaller, and expenditure on clothing, travel (including automobile), recreation, and private education play a larger role. Changes in the distribution of income appear unlikely to produce important changes in the direction of expenditure among the major items, provided the total expenditure remains the same. On the other hand, changes in the level of the total consumer expenditure appear likely to produce considerable changes in the direction of expenditure, reflecting the shift in expenditure as the buying power of each income group is shifted upward or downward. A higher national expenditure would involve a somewhat smaller proportion spent on food and a larger proportion spent on clothing, travel, recreation, and private education. Few data are available on the effect on expenditure of the transition from one level of consumer income to another, but it seems likely that a transition to a lower level of expenditure would decrease the proportion of expenditure going to more durable goods and a transition to a higher

level would increase the proportion going to such goods. The influences of likely changes in price relationships on the direction of expenditure and trends of change still remain to be worked out. When these are determined the main aspects of the structure of wants will have been covered for the major items of consumer expenditure.

Major Items of Social Expenditure

In addition to the wants of consumers expressed through private expenditure, there are many wants, such as those for education, sanitation, highways, church services, and hospital care, which find expression largely through social expenditure. A rough estimate of the principal items of social expenditure indicates that the total social expenditure amounted to at least 5 billion dollars in 1935-36. If this figure is added to the 51 billions of direct consumer expenditure, it gives a total expenditure on consumption of 56 billion dollars, of which at least 9 percent was social expenditure. In addition, consumption by institutional residents amounted to approximately 200 million dollars, and by quasi-institutional groups, such as army, navy, and civilian conservation corps, to 500 million dollars.

The principal items obtained through social expenditure are shown in table IV. Some of these items are supplied only on a social basis, such as roads. Others are partly furnished socially, and partly out of individual expenditures. The total of individual expenditures for the roughly corresponding items is placed beside the social expenditure.

TABLE IV.-Major items of social expenditures 1935-36
(millions of dollars)

100

Operation and maintenance of public service enterprises. Capital expenditures..

Millions

833

1,081

940

999

1,845

Some items included within these categories might properly be included as items of social consumption, e. g., that part of expenditures for police and fire departments which goes to the protection of homes. Some parts of the items included in the table should not be included, e. g., the part of expenditure for highways which should be allocated to business rather than to consumption. The whole field of social expenditure is one upon which much work needs to be done before even a roughly adequate analysis can be made.

2 Consumer Expenditures in the United States, National Resources Committee. Nearly half of this amount represents interest on Federal debt incurred through loans made to Allies during the World War. This should not be considered as an expenditure for social consumption.

The amount and character of social expenditure is, in large measure, a reflection of the wants of consumers, although it does not involve direct purchase by them. Contributions in the form of taxes and gifts to governments and institutions for purposes of social expenditure are part of the expenditure patterns of individuals. The wants of consumers, and the ability to make those wants effective, are reflected in large expenditures for public education in a wealthy community, or in one sufficiently eager for education to sacrifice direct private expenditures in favor of indirect expenditures via taxes, and large expenditures for religion in a wealthy or especially pious parish. The variation in social expenditure for public education in relation to differences in per capita retail sales, shown in chart VII, emphasizes this close relation between individual and social expenditure.

Durability and Consumer Wants

So far both consumer and social expenditure have been considered primarily in terms of the types of wants which they are aimed to satisfy. An equally important grouping of consumer wants would be one that took account of the durability of the goods which are purchased to satisfy them. As will become apparent as the structure of the American economy is further examined, important structural characteristics revolve around the factor of durability.

For purposes of analysis, all consumer goods can be grouped into four degrees of durability: durable commodities, semidurable commodities, nondurable commodities, and services. Automobiles and furniture would fall into the durable group since they usually render a series of services over a considerable period of years. Clothing and automobile tires fall into the semidurable group, yielding a series of services but not usually lasting more than a year or so. Food and gasoline fall into the nondurable category since they are usually consumed in what is, for practical purposes, a single operation instead of rendering a series of services during a period of time.10 Services include the items which do not take an intermediate physical form but are rendered directly to consumers, such as education, music, medical service, and personal service.

The data on consumer expenditure are not yet avail. able in a sufficiently detailed form to make possible an accurate estimate of the proportion of consumer expenditures falling into each class of durability. In chart VIII and in table V, the major items of expenditure are grouped on the basis of the durability categories into

10 Durability in the sense of rendering a series of services should not be confused with non-perishability. Commodities like salt or wine may be capable of being stored for long periods yet are usually classed as nondurable if they are customarily consumed in a single use.

Source: Based on data from the report, Consumer Expenditures in the United States, National Resources Committee.

which they most nearly fall. These groupings reflect degrees of durability only roughly, since the available expenditure categories are, in many cases, combinations of goods of varying degrees of durability. Only half of the total automobile expenditures in 1935-36 went for the purchase of automobiles, and half went for semi- and nondurable goods, such as tires, oil and gas, or for garage and similar services. Expenditures for household furnishings are classed as dominantly durable because they consist to such a large extent of furniture, but they contain such semidurable goods as dish towels and sheets. Recreation, classed as dominantly nondurable, includes the purchase of radios. Housing is separately listed since expenditure for rent constitutes payment for a service provided by a durable good and does not fall clearly into the four categories of durability.

In spite of the mixed character of the data, it is apparent from chart VIII that the bulk of consumer expenditure in 1935-36 was directed to the purchase of goods which were not of a durable or semidurable character. Over 60 percent of expenditure went into categories dominated by nondurable commodities and services and nearly 20 percent into housing. Only approximately one-tenth each went to categories dominated by semidurable and durable consumers goods."

The importance of the distribution between durable and nondurable commodities for the structure of wants lies in the greater sensitivity of expenditure on consumer durable goods to variations in consumer income. As in the case of the major items of consumer expenditure, a shift in the distribution of consumer income would produce little alteration in the proportion of a given expenditure going to durable as against nondurable goods.12 However, a variation in the level of total expenditure could be expected to make a significant difference, the

11 It is probable that the grouping of items here used throws more nondurable items into the durable category than vice versa, and tends to minimize rather than to exaggerate the predominance of expenditure for nondurable items.

12 If the whole of the 50 billion dollars expended by consumers in 1935-36 had been spent according to the expenditure pattern of consumers with incomes between $1,250 and $1,500, 0.8 percent less would have been spent on the items grouped into the durable goods category, 0.6 percent less on semidurable, 0.7 percent less on housing, and 2.1 percent more on goods and services in the nondurable category.

sponding to an income of 40 billion dollars is compared with that of 63,198 million dollars corresponding to an income of 80 billion dollars, and both are compared to the 1935-36 expenditure corresponding to approximately 60 billion dollars. Each expenditure is calculated on the assumption that prices and proportionate income distribution were the same as in 1935-36 and that consumers followed the patterns of consumption at the different income levels already indicated in chart V. Such calculations take account of the difference in level of expenditure but, of course, take no account of the transitional changes in expenditure.

In practice, rapid changes in the total of consumer expenditure are usually accompanied by significant transitional and price changes so that the expenditure on consumer durable goods is more sensitive to variations in consumer expenditure than these figures would indicate. In chart IX below the consumer expenditure on durable, semidurable, and nondurable goods and for services is indicated for recent years. For each series the estimates of dollar expenditure have been deflated by a price index for the particular category so as to give a quasi-physical measure of quantity of production stated in 1929 dollars. The greater sensitivity of durable goods expenditure to variation in total expenditure is at once apparent in the data for the depression period. The quantity of durable goods purchased dropped 52 percent from 1929 to 1932, whereas the semidurable goods purchased dropped only 20 percent and the nondurable dropped 8 percent. The purchase of services, represented by a less reliable figure, appears to have dropped approximately 33 percent. If the drop in total expenditure had been simply one of level without significant transitional or price changes and had followed the 1935-36 pattern for the separate income groups, the percentage drops would not have corresponded to those given above.

Until it is possible to disentangle the influence of prices from that of the transition from one level to another it is not possible to delineate that dimension of the structure of wants which has to do with the influence of prices on the relative expenditures on durable and nondurable goods. Likewise the trends