The Quarterly National Income And Product Accounts Of The United States 1942–621

The purpose of this paper is to describe and evaluate the quarterly national income and product accounts of the United States (to be referred to henceforth as N.I.P.). The historical development of these accounts is reviewed first. Next, a summary of the statistical methodology underlying them is provided. An analysis of the errors to which they have been subject follows. An attempt is then made to define the area of their usefulness. Finally, suggestions for their improvement are formulated. The most general conclusions are as follows: 1 The quarterly N.I.P. estimates were developed in close response to an urgent need for them in economic analysis and policy formulation. Their development was a gradual process covering about two decades in which experience gained at one stage suggested improvements and extensions for the next. 2 The quarterly figures are based on abundant data sources. To be sure, the information is not as comprehensive as that available for the preparation of the annual estimates, but we are dealing only with differences in degree ‐ there is no sharp contrast. 3 The series are subject to a considerable margin of error. There is bias in some of them ‐ the initial estimates tend to be too low; and quite apart from bias they differ from the revised figures. 4 The margin of error attaching to the statistics disqualifies them from serving as precision instruments, but they are indispensable as a systematic framework for the order‐of‐magnitude analysis of the major forces determining the short‐run movements of the economy. At present, the outstanding handicap in such analysis is not statistical error, but the inadequacy of economic theories and of our ability to test them. 5 The opportunities for mechanizing estimating techniques are assessed; the work on seasonals is discussed in this context; and a thorough investigation is proposed of the problems involved in the synchronization of the various components of the accounts. In addition, proposals for improving the data available for the estimation of specific income and product flows are formulated. Quarterly estimates of the physical volume of national output by industry, and quarterly accounts showing the financial transactions that mediate between saving and tangible investment are considered the two high‐priority major extensions of the present accounts. Other, less basic, extensions of these accounts are also proposed.