level of taxes or permit fees for pollution emissions or waste disposal. Similarly, decisions about public investments in education would benefit from information about the externalities associated with a more educated population. It is equally clear, however, that accurate measurement of these externalities is apt to be a challenge.

Measuring Quantities

Dollar values are relatively easy to obtain for the market inputs to nonmarket production. If one wants quantity indexes for these market inputs, they can be constructed by using appropriate price indexes as deflators for the nominal expenditure data. In contrast, for both nonmarket inputs and nonmarket outputs, quantity measurement often will be a necessary first step in the development of monetary valuations.

Complications can arise even in the case of market inputs. Purchases of capital equipment by households, for example, are categorized as final consumption in the NIPAs. But measuring the inputs to household production requires an estimate of the stock of consumer durables. To create such a stock estimate, one must combine information on spending over time for dishwashers, refrigerators, vacuum cleaners, washing machines, and other capital equipment used in home production with information on these items’ useful lives. Although there are practical difficulties that complicate estimation of the stock of capital equipment used in home production, the basic approach is well developed.5

An especially important nonmarket input on which, until very recently, quantity data have been lacking is the time devoted to nonmarket production. Fortunately, the American Time Use Survey (ATUS), launched at the start of 2003 by the Bureau of Labor Statistics, should go a long way toward filling this gap. The ATUS, described more fully in Chapter 2, can be expected to provide good data on the inputs of adult time to various sorts of nonmarket production in households of various types.

These data would be even more useful if the Census Bureau were to produce frequently updated information on the distribution of demographic characteristics in the population, designed to complement the new information on time and support accounting efforts generally. A complete demographic database might include information on the age, gender, school enrollment, years of education and degrees completed, occupation, household structure, immigrant status, employment status, and possibly other dimensions of the population. Knowing about the distribution of demographic characteristics and changes in that distribution over time would be of value, as an example, for determining whether observed changes


This is a case for which the BEA already maintains a suitable data series, albeit not as a part of the core accounts. See Katz (1983) for a discussion of measuring the stock of consumer durables.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement