Employment and Consumption
HE purpose of this paper is to explore T the effects of variations in employment on family and aggregate consumption. The results illustrate a general thesis that the effects of income variation on consumption expenditures depend on the sources of such variation. This is true in cross-sections as well as in time series, even though the major factors related to changes of income over time are not equally important as determinants of income differences at a point of time, and conversely. Growth of productivity is, of course, the essence of long-run changes in real per capita income, and fluctuations in amounts of factor inputs, particularly labor, dominate the short-run changes in income. In a cross-section, a long list of factors responsible for differences in income can be named, and, once again, differences in the degree of employment among individuals and families play an important role. For purposes of emphasis and brevity, we shall abstract from other factors in tracing the effects of the employment variable. Note that the degree of employment of members of a consumer unit observed in a given short period (say, a year) is a very unreliable indicator of the unit's longer-run income position, compared with other income-determining characteristics, such as education, occupation, property ownership, or even age (experience). This observation points to an obvious way of introducing the employment factor into consumption analysis. This is achieved by a special interpretation of the theory according to which a family's aggregate consumption is determined by its income.' As a first approximation, we may define expected or income as the income which a family receives per unit of time during which its labor input is normal.2 This definition is likely to be quite satisfactory for analytical purposes, if we restrict ourselves to the wage-earning group, particularly the unskilled. If we include the whole range of skills up to the highly trained professions in our population, we must take account of another factor which makes for a difference between current and expected income, namely changes in income with age3 (experience), quite apart from the effects of variations in employment. These age-changes are more pronounced the higher the skill level of an occupation, so that in the top occupation groups (professional and managerial) they are much more important than employment changes in distinguishing between current and expected income. Thus, in each individual case the previously defined measure of expected income should be corrected upward whenever the individual is located on the upward phase of his age-income curve, the correction being larger the steeper the curve, and conversely. In the case of income from self-employment or from property, expected income is best identified with normal returns in a given industry, and the differences between current and expected income are cyclical for groups as well as both random and age-associated for individuals. Let us now specify a model of consumption behavior along the lines of expected income theory, using this particular approximation of the concept of expected income. Because of its commitment to a different interpretation of expected income, the Modigliani-Brumberg model is not useful in the present context. While Friedman's framework is more appropriate, some of its assumptions which the deliberate non-specificity of the concept of permanent income made possible will be changed to suit the purposes at hand. It is of interest to note that the modifications do not involve complica* This paper was presented at the Boston meetings of the Econometric Society, August I958. Research embodied in it was carried out as part of the Rockefeller Foundation Consumption-Income Distribution Research Project at the University of Chicago. The author is indebted to Dorothy S. Brady and Margaret G. Reid for valuable comments. 'As expounded by Friedman in A Theory of the Consumption (Princeton, I957); and by ModiglianiBrumberg in Utility Analysis and the Consumption Function in Post-Keynesian Economics, ed. K. Kurihara (New Brunswick, I954). 2 We abstract from property income throughout the analysis. 'Reference here is made to age and occupation of the family head.
- DOI
- 10.2307/1926091
- Volume
- 42 (1)
- Pages
- 20
- Export
- BibTeX
- Sources
- openalex crossref