T HE publication of Keynes' General Theory in I936 not only ushered in a new era of theoretical thinking in economics, but also proved a stimulus to statistical work of both explanatory and prognostic nature. When it was noticed that aggregate savings retrospectively always must be equal to aggregate investment (despite investment decisions and saving decisions being mutually independent), the idea arose to use investment series for forecasting aggregate income and employment, provided that aggregate saving as a function of income should prove sufficiently stable over time. The first preparatory step in this direction was an analysis of how the current saving propensities were allowed to materialize into actual savings by offsetting investment activities, which thus assisted in maintaining an aggregate income just large enough to generate the actual amount of saving in question. When such analysis was undertaken for the United States, it was at once realized that, besides private investmenkt, other forces were operating to the ssaving propensity of the community. The offset series, prepared in the Division of Research and Statistics of the Board of Governors of the Federal Reserve System, about which the public was informed by Dr. L. Currie's testimony before the T. N. E. C.1 contained, in addition to three private investment series (i) plant and equipment, (2) private housing expenditure, (3) net change inventories three other offsets, which we may briefly call (4) government net contribution, (5) net foreign balance (on current account), (6) net change in consumers' credit. The introduction of number (5) is necessary because the basic identity

Introduction: types of rationing, 378. — Over-all rationing: pure over-all dollar rationing, 379; pure over-all point rationing, 381; over-all point dollar rationing, 382. — Over-all rationing vs. specific rationing, 386. — Group rationing the bases for grouping, 389; value vs. point rationing, 391; point prices, 392. — Conclusions, 394. — Appendices, 394.

IN a growing economy, current replacement falls short of depreciation. The implications of this fact were discussed at length in a paper by E. D. Domar.' The nature of the subject did not allow Domar to indicate the relationship between the two magnitudes in explicit form; in their place he had to present numerical illustrations. In the case of the individual firm, we can go one step further. Obviously a firm that buys its equipment first-hand will accumulate depreciation funds ahead of the replacement necessity. If the firm decides to reinvest the depreciation allowance, its gross or operating capital, in terms of performance, will increase for some time, although of course the value of the net capital stock by definition would remain constant. This is so because the performance of adequately maintained equipment declines less in proportion to depreciation. At the same time the average lifetime of the equipment items making up the gross capital stock will change. Since the initially installed equipment has to be replaced at some time, the rise in the gross stock from reinvestment of depreciation allowances will be interrupted discontinuously: the gross stock will suffer an abrupt decline, after which it will rise again. Let us give a simple illustration, the basic premises of which will be specified later on. Suppose a railroad invests at the beginning of I957 $IO million worth of rolling stock of a lifetime of exactly ten years, and applies straightline depreciation, amounting at the beginning to $i million per annum. It reinvests after the end of each year the depreciation allowance which was set aside in the preceding year, in rolling stock of the same kind. Assuming that prices do not change, the stock would grow as follows: Beginning of I957 $IO million I958 i i million I959 I2.I million, etc.