experience. Such things as household production of cleaning services or participation in educational activities have market analogues and are (relative to, say, leisure) more closely aligned with what is currently viewed as output in the accounts, meaning that the hurdles to clear in devising sensible measurements are lower. Because improving output (and corresponding input) measures is a prerequisite to any vision for an expanded set of accounts, this is where the panel focused its energies. Even if the accounting objective is limited primarily to measurement of output, there is still a formidable amount of work to be done.
In defining the boundary of what should be counted when developing measures of nonmarket output, some have advocated the application of Margaret Reid’s (1934) third-party criterion: Is the output something that a person could have hired someone else to produce? This criterion seems appropriate for certain areas—for example, a household production account could be designed to include such things as meals, clothing services, shelter services, and a basic component of child care, but to exclude fertility, studying, and exercise. For other areas of nonmarket activity, such as education and health, the third-party criterion is clearly inappropriate: there is no replacement for self involvement in the activities required to enhance a person’s cognitive skills or improve his or her health—activities that nonetheless produce valuable but nonmarketable capital outputs.
In considering how broadly nonmarket accounting work should be cast, one must ask what standards of accuracy and reliability should be applied to measures of inputs and outputs. We do not have a full answer to this question. Traditionally, the statistical agencies responsible for economic accounting—the Bureau of Economic Analysis of the Department of Commerce and the Bureau of Labor Statistics of the Department of Labor—set high standards of accuracy. At least initially, nonmarket satellite accounts presumably will have to be constructed under a more forgiving standard, but there undoubtedly will be debate about how imprecise a number can be and still remain useful. In market accounting, outside researchers often produce accounts from ingredients supplied by the statistical agencies—for example, a consulting firm publishes monthly GDP from data supplied by the BEA, though the BEA does not consider the data to be sufficiently reliable in all sectors to produce an official monthly measure. We anticipate similar private accounting efforts with nonmarket data.
As detailed in this volume, potentially valuable areas of nonmarket accounting are at different levels of development with respect to measurement concepts and available data. For that reason, the panel favors a staged approach. Work should begin in areas where potential gains are high and costs (mainly in terms of new data collection needs) are low. For example, we believe that accounting for certain types of home production should begin soon, based on data from the American Time Use Survey. In contrast, neither the conceptual understanding