F 349-469. General note.

The national balance sheet is derived by summing similar balance sheets for sectors in the economy-nonfarm households, agriculture, unincorporated business, corporations, etc. The balance sheet of each group is in turn derived by summing the balance sheets of the constituent units, based as far as possible on a comparable valuation of assets and liabilities. In deriving the balance sheet, no creditordebtor or owner-issuer relationships among units are eliminated; for example, the debts of households to corporations appear on one side as assets of corporations and on the other as liabilities of households. When all relationships among constituent units are canceled, whether these units be in the same or different groups, the balance sheet reduces to a national wealth statement. (In the series shown, the estimate for total tangible assets in the national balance sheet differs very slightly from that in the statements of national wealth and national tangible assets because of a minor disparity in the treatment of monetary metals.) Thus, the national balance sheet adds to the national wealth statement a comprehensive summary of the various types of financial obligations outstanding at a particular date, and provides perspective on the magnitude of financing activities in the Nation's economy.

The national balance sheet falls somewhat short of the goal of a comprehensive summary of the assets, liabilities, and net worth of all transactors in the economy, since, for lack of data, obligations among households are not included, and in the case of corporations with subsidiaries, the balance sheet of the parent company is used, thus eliminating relationships among the subsidiary units. In addition, intangibles such as goodwill and patent rights are excluded from the balance sheet. Finally, and this limits the comprehensiveness of the national wealth statement as well, inventories of nondurable goods in the hands of consumers, expenditures on soil improvement, subsoil assets, and military and naval equipment held by the government are omitted.

The value for "equity" in the national balance sheet exceeds total national wealth, that is, consolidated net national worth. This is primarily because, in the balance sheet, the net worth of the various constituent units are added together. For example, the net worth of a corporation is added to the net worth of the stockholders. In the national wealth statement, however, they are consolidated. That is, the outstanding stock of the corporation is canceled against the holdings of the owners, leaving only the net worth of the stockholders and the undistributed earnings of the corporations. Stated differently, the "equity" entry in the balance sheet includes the equity of intermediaries as well as of ultimate owners.

F 349-364. National tangible assets, in current prices, 1952-1968. Source: U.S. Congress, Institutional Investor Study Report of the Securities and Exchange Commission, Supplementary Volume I, House Document 92-64, Part 6, March 10, 1971.

Estimates of reproducible assets shown in series F 349-364 and F 365-376 were made using the perpetual inventory method. This method involves the computation of a weighted sum of a time series of gross investments in the asset; the weights are determined by the particular life and depreciation assumptions employed in the calculation. The difference between the gross investment of a given year and the change in stock during that year is, by definition, the depreciation which has occurred. To derive the replacement cost estimates used in series F 349-364, the calculation was first made in terms of constant dollars (series F 365-376), and then the stock and depreciation estimates were reflated to current year prices.

The gross investment series used for the estimates of the private stock of depreciable assets are in all cases those used in the gross investment component of the income and product accounts produced by the U.S. Bureau of Economic Analysis (BEA). For public sector estimates, the construction data and equipment series were taken from the income and product accounts wherever possible. Such data are published regularly in the Survey of Current Business, although the two government sectors are not credited with capital formation in the BEA accounts.

For a detailed description of the method used to obtain estimates for the various components of depreciable assets, see the source publication cited above, pp. 252–259.

Land estimates shown in series F 349-364 and F 365-376, with few exceptions, are those given in Appendix II, "Estimates of the Value of Land in the United States Held by Various Sectors of the Economy, Annually, 1952 to 1968," of the source publication. The land of financial corporations was estimated by multiplying the Internal Revenue Service estimates of the book value of land of all financial institutions by the market-to-book ratio developed in Appendix II for "finance, insurance, and real estate." No adjustment was made for unincorporated financial institutions, which tend to be brokerage houses, as the land holdings of the finance, insurance, and real estate aggregate for partnerships and proprietorships are accounted for primarily by the holdings of real estate firms.

The estimated value of farmland shown in these series was derived by subtracting the value of buildings from the U.S. Department of Agriculture's estimate of the value of farm real estate.

Transactions were measured by first differences in the holdings.

F 365-376. National reproducible tangible assets in constant (1958) prices, 1952-1968.

Source: See source for series F 349-364.

See also general note for series F 349-469 and text for series F 349-364.

F 377-421. National balance sheet, in current prices, 1900-1968. Source: Raymond W. Goldsmith, et al., Studies in the National Balance Sheet of the United States, vol. II, Princeton University Press, tables I and Ia (copyright 1963 by National Bureau of Economic Research, New York); and unpublished data.

The national balance sheet is derived by summing similar balance sheets for various transactor groups in the economy-nonfarm households, agriculture, unincorporated business, etc. (see general note for series F 349-469). For most of these groups, however, balance sheets of the constituent units are nonexistent, so that in practice the group balance sheet is compiled from separate estimates of the various categories of assets and liabilities, net worth being derived as a residual. Only in the case of corporations and the Federal Government does a substantial proportion of the items come from their own financial statements. Military assets, i.e., military structures and equipment and the assets of the Atomic Energy Commission, are excluded from these balance sheets.

The estimates presented are in current prices rather than original cost. Essentially, this means that reproducible tangible assets are valued at reproduction cost, and nonreproducible tangible assets and intangibles at market value, though some intangibles, particularly short-term claims, are valued at par or face value.

In deriving the estimates, a problem sometimes arose because of a difference between two groups in the value at which the same item is carried on the balance sheet, a difference not attributable to bad debt

reserves alone. Where this was the case, no attempt was made to force consistency. Both valuations were carried over into the national balance sheet on the appropriate sides. This, together with the treatment of net holdings of foreign assets and liabilities, principally accounts for differences between the asset and liability totals for certain intangible items-differences which are generally small compared with the balance sheet totals.

The source provides considerable additional detail-in particular, balance sheets for separate transactor groups, such as nonfarm households, agriculture, etc., and makes it possible to trace the patterns of claims and counterclaims among the various groups.

Figures for 1958-1968 have been derived principally from the Federal Reserve Board's flow-of-funds data and differ from earlier data mainly because of statistical revisions in the basic data. However, differences in the following items are the result of conceptual differences.

F 381, monetary metals. Data for 1900-1958 include all gold and silver coin; data for 1958-1968 include gold and official foreign exchange reserves.

F 382, other currency and demand deposits. The earlier estimates include cash items in process of collection and other interbank claims within the private financial sector which are not included in the flow-of-funds data. These items amounted to $35.1 billion in 1958.

F 397, U.S. Government securities, long-term. Data for 19001958 include special issues held by U.S. Government pension and trust funds; data for 1958-1968 do not. Data for 1958-1968 include issues of U.S. Government credit agencies, while 1900-1958 estimates include these in "other bonds and notes" below.

F 401, other bonds and notes. See above.

F 405, equity in other business. Data for 1900-1958 include equity in unincorporated broker-dealers; 1958-1968 data cover non-farm, nonfinancial business only.

F 410, private life insurance reserves. The 1900-1958 data include the pension reserves of life insurance companies and the policy reserves of fraternal insurance organizations which are not included in the flow-of-funds figures. However, the policy reserve estimates are available in Appendix I of the Institutional Investor Study (see source for series F 349-364).

F 411, private pension and retirement funds. Data for 1958-1968 include the pension reserves of life insurance companies which are included in F 410 for 1900-1958.

F 412, Government pension and insurance funds. Data for 1900-1958 include the reserves of Old Age Survivors Insurance, about $21.9 billion in 1958; data for 1958-1968 omit these, although data are available in the Monthly Treasury Statement.

F 422-445. National wealth, by type of asset, in current prices, 18501958.

Source: Raymond W. Goldsmith, 1850-1900, "The Growth of Reproducible Wealth of the United States of America From 1805 to 1950," International Association for Research in Income and Wealth, Income and Wealth of the United States: Trends and Structure, Income and Wealth Series II, Bowes and Bowes, Cambridge, England, 1952, p. 306 (estimates for 1805 presented in this publication have not been reproduced here because of questionable reliability); 1900-1958, The National Wealth of the United States in the Postwar Period, Princeton University Press, App. A and B (copyright 1962 by National Bureau of Economic Research, New York).

The estimates for 1900 to 1958 were constructed by Goldsmith by means of the "perpetual inventory method." In this method, the stock of an asset in existence at a given point in time is estimated from annual output totals extending back over a period equal to the average life of the asset, the output total for every year being depreciated to the end of the period, and the results summed. (See also text for series F 349-364.) Military assets are excluded.

The underlying estimates for 1850 appear in the Census Office, Preliminary Report of the Eighth Census, 1862, p. 195; and those for 1880, 1890, and 1900 in Simon Kuznets, National Product Since 1869,

National Bureau of Economic Research, New York, 1946, pp. 202215. In every case, the original estimates were adjusted by Goldsmith (for 1880 substantially) to improve comparability with the estimates for 1900-1958. The basic sources for these earlier estimates were returns on stocks of various assets in the industrial censuses and censuses of wealth. Hence, there is a sharp break in the method of derivation between the earlier and later estimates. However, the figures for the overlap year, 1900, agree reasonably well. The figures for 1850 exclude the value of slaves.

The estimates for 1900-1958 are in "current prices," that is, each asset is valued at its replacement cost in the given year. This is preferable to valuation at original cost, whether depreciated or undepreciated. Assets appearing in the wealth statement for any given date were produced in different years, and since prices change from year to year, summation of original cost values would often result in an arithmetic aggregate without economic meaning.

For the estimates for 1850 to 1900, which are primarily from the Federal censuses, the basis of valuation is not always certain, and is not uniform among types of assets and among industries. It is possible that the figures may approximate either current market values or original cost, depreciated or undepreciated, or some combination of the two. Some assurance as to the comparability of the earlier and later sets of figures on this score is provided, however, by the overlapping values for 1900, though this comparison applies only to a single year.

As to the reliability of the estimates for 1850 to 1900, the source (Income and Wealth of the United States: Trends and Structure) states that the margin of error amounts to hardly less than 10 to 20 percent at any date, that this relative margin increases going back in time, and that it is not certain that comparability is impaired by as much as the size of the margin may imply because the error probably tends in the same direction for most if not all benchmarks, although the understatement is probably more pronounced in the early part of the period than in the latter. Concerning the estimates for 1900 to 1958, derived by the perpetual inventory method, the most important source of error is considered to reside in the estimates of construction expenditures. For some of the components of total wealth, reliability is strengthened because of the availability of checks against alternative estimates, as is the case for residential real estate, farm structures, inventories, and international assets. Checks are less satisfactory for nonfarm business structures and equipment but the information in corporate balance sheets submitted to the Internal Revenue Service gives assurance that the perpetual inventory estimates are not too far off for recent years. The only sectors of reproducible tangible wealth in which the perpetual inventory estimates are not subject to checks, or only to very unsatisfactory checks are consumers' durables and government fixed assets.

The source also presents considerably greater detail than given here (for example, annual estimates for 1896-1949). Estimates of national wealth by contemporaries are also available for various dates during the 19th century. See, for example, Samuel Blodget, Jr., Economica; A Statistical Manual for the United States, 1806 edition, and Annual Report of the Director of the Mint, 1881.

F 446-469. National wealth, by type of asset, in 1929 and 1947-49 prices, 1850-1958.

Source: See source for series F 422-445.

These estimates were derived by adjusting the current dollar figures for a given class of assets in series F 422-445 for the change in price or cost of construction of that type of asset between each year and the base year. Thus, conceptually, changes over time in the constant price value of a category of assets reflect changes in the physical stock of that asset and not in its value. For 1945-1958, a different base year was necessary because estimates in 1929 prices for the most recent years were not available. This shift in base years introduces some element of incomparability, since the relative weights of individual assets in the price index differ between the two years.

For 1900-1958, an attempt was made to adjust for price changes by fairly narrow classes of assets, using construction cost or price indexes referring specifically to the assets in each class. For 1880, 1890, and 1900, a more summary adjustment was used. Only three separate deflators were employed for construction (residential, other private, and farm), and a single deflator was used for all types of equipment. For 1850, the same price index (Snyder's index of the general price level) was applied to all types of structures and equipment, although for the adjustment of inventories the wholesale price index was used.

Goldsmith states that the conceptual significance of a constant price estimate for land is open to question. If land is carried for all dates at its absolute value in the base year, the relation to the constant price value of reproducible assets tends to become unrealistic, particularly at dates fairly far removed from the base year. In the present estimate, an alternative procedure is followed, a constant price value of land being derived, generally speaking, as a fixed proportion of the constant price value of structures. This permits derivation of a constant price series for aggregate national wealth, but it should be recognized that the deflated estimates of land values included in the totals cannot be conceived as reflecting changes in physical units alone.

The adjustment for price changes introduces errors in the estimates in addition to those discussed in connection with series F 422445.

On balance, any error is likely to lead towards an overstatement of the price rise over the period and hence an understatement of growth rates because the techniques used in adjusting for price change fail to make adequate allowance for improvement in the quality of the assets, and there is no evidence that the error is larger for one part of the period than for another, although the possibilities of error are certainly greater in the 19th century than the 20th. In addition, it is likely that the failure to allow for quality improvement has a differential effect on the different components of wealth. In particular, it leads to a more serious understatement in the growth of components such as producer and consumer durables than for structures and inventories.

F 470-534. General note.

Although estimates of capital stocks are less well developed than those of economic flows, in recent years a number of capital stock estimates have been prepared and published by the U.S. Bureau of Economic Analysis (formerly the U.S. Office of Business Economics), as part of a project to measure the entire tangible wealth of the Nation. BEA estimates have been published for (1) residential capital, (2) fixed nonresidential business capital, (3) provisional estimates of consumer durable goods, and (4) stocks of business inventories. References for these studies are as follows: (1) John C. Musgrave, "New Estimates of Residential Capital in the United States, 1925-73," Survey of Current Business, October 1974; (2) Bureau of Economic Analysis, Fixed Nonresidential Business Capital in the United States, 1925-73, 1974; (3) Henry Shavell, "The Stocks of Durable Goods in the Hands of Consumers, 1946-1969," 1970 Proceedings of the Business and Economics Section of the American Statistical Association, 1971; (4) Shirley F. Loftus, "Stocks of Business Inventories in the United States, 1928-71," Survey of Current Business, December 1972, with updating in August 1974 Survey of Current Business. Also, estimates of inventories owned by nonfinancial corporations, as of midyear for the years 1948-71, in constant (1958) prices and the current prices of each year, appeared in "Nonfinancial Corporations: New Measures of Output and Input," by John A. Gorman, Survey of Current Business, March 1972.

Series F 470-479, F 480-515, F 516-527, and F 528-534 provide selected series from these BEA capital stock studies. Series F 470-479 is a summary table providing series on gross and net stocks of nonresidential structures and equipment, residential structures and equipment, and inventories in both current and constant prices. Greater detail on nonresidential structures and equipment and residential structures is provided in series F 480-515, F 516-527, and F 528-534.

Fixed nonresidential structures and equipment estimates are computed by the perpetual inventory method, which derives capital stock estimates for a given year by cumulating past investment and deducting the cumulated value of the investment that is used up. The data used to implement this method are taken from the national income and product accounts since 1929 and from various private studies prior to that time. Included are all privately owned nonresidential structures and producers' durable equipment. Estimates shown are on the secondhand price method of valuing business purchases of government surplus assets, variant 1 deflators for structures, straight-line depreciation, and 85 percent of service lives given in Bulletin F, Internal Revenue Service. The series published here is just one of a number of variants of capital stock estimates reflecting different valuations, service lives, and depreciation techniques. (See source for additional estimates.)

The residential capital estimates are also computed by the perpetual inventory method. The data used to implement this method are taken from the national income and product accounts since 1929 and from various private studies prior to that time. Included are all residential structures, both privately and publicly owned. Depreciation is estimated by a declining balance formula.

The stocks of business inventories shown in series F 470-479 were calculated by cumulating the annual inventory changes, in book values and in constant (1958) prices, respectively, that are estimated in the national income and product account. An estimate of the level of each book value and constant price stock series was made for some single point in time for which appropriate data were available; that stock was then moved forward through time by adding the estimated annual changes and backward through time by substracting the annual changes.

Series F 516-527 and F 528-534 give information on the age structure of the capital stock. Such information is essential for gauging the extent to which capital is up-to-date in terms of both physical condition and technological characteristics. Two measures of age structure are presented in the publications cited above: an average age series of the capital stock and the ratios of the net stock of the capital to the gross.

These two measures of age can be used interchangeably for many purposes, but each provides specific information. The net-gross ratios show the extent to which the services initially embodied in capital goods remains intact, on the assumption that the purchase price is a measure of the value of the services bought initially and that depreciation reflects the value of the services that have been used up. The average age measures, which are given in series F 516-527 and F 534, provide information on absolute age.

F 470-479. Private capital stocks, current and constant (1958) cost valuation, 1925-1970.

Source: Series F 470, sum of series F 471-474; series F 475, sum of series F 476-479. Series F 471, 472, 476, and 477, U.S. Bureau of Economic Analysis, Fixed Nonresidential Business Capital in the United States, 1925-1973, 1974, pp. 1-12 and unpublished data. Series F 473 and 478, "New Estimates of Residential Capital in the United States, 1925-73," Survey of Current Business, October 1974. Series F 474 and 479, "Stocks of Business Inventories in the United States, 1928-71," Survey of Current Business, December 1972, pp. 29–32, and August 1974.

For a description of the conceptual framework and estimating techniques used to derive these data, see the general note for series F 470-534.

F 480-515. Fixed nonresidential business capital-current and constant (1958) cost valuation, 1925-1970.

Source: U.S. Bureau of Economic Analysis, Fixed Nonresidential Business Capital in the United States, 1925-1973, 1974, pp. 1-12 and 48-51, and unpublished data.

For a description of the conceptual framework and estimating

techniques used to derive these data, see the general note for series F 470-534.

F 516-527. Fixed nonresidential business capital-average age of
gross stocks, constant (1958) cost valuation, 1925-1970.
Source: U.S. Bureau of Economic Analysis, Fixed Nonresidential
Business Capital in the United States, 1925-73, 1974, pp. 1-12.

For a description of the conceptual framework and estimating techniques used to derive these data, see the general note for series F 470-534.

F 528-534. Residential capital, current and constant (1958) cost valuation, 1925-1970.

Source: U.S. Bureau of Economic Analysis, "New Estimates of Residential Capital in the United States, 1925-73," Survey of Current Business, October 1974.

For a description of the conceptual framework and estimating techniques used to derive these data, see the general note for series F 470-534.

F 535–539. Value of stock of structures and equipment in specified sectors, in 1929 prices, 1880-1948.

Source: Simon Kuznets, Capital in the American Economy: Its Formation and Financing, National Bureau of Economic Research, New York, 1961 (copyright).

These estimates fall somewhat short of the value of all reproducible wealth in each sector, since the value of inventories is omitted, and considerably short of total wealth, since land is excluded. Also, data are not available for other business sectors; for example, trade and the service industries are omitted. However, it is estimated that the four sectors included here accounted for about 80 percent of the stock of structures and equipment in 1880.

The underlying sources of the estimates are three monographs prepared in connection with the National Bureau of Economic Research Study of Capital Formation and Financing: Alvin S. Tostlebe, Capital in Agriculture: Its Formation and Financing Since 1870, Princeton University Press, 1957; Daniel Creamer, Israel Borenstein, and Sergei P. Dobrovolsky, Capital Formation and Financing in Manufacturing and Mining, 1960; and Melville J. Ulmer, Capital in Transportation, Communication, and Public Utilities: Its Formation and Financing, 1960. With the exception of the last monograph, the approach followed in deriving the estimates of capital stock differed rather noticeably from that chiefly employed in obtaining the figures presented in series F 422-469, since the basic data, such as census returns on capital or balance sheet items in Statistics of Income, related to stocks rather than output flows. Further detail on capital investment by type and/or minor industrial sector is given in these monographs.

Statistics of saving provide the link between the statements of national income or product, on the one hand, and the national wealth statement and balance sheet, on the other. Generally speaking, for the Nation as a whole, aggregate saving, which equals national income less national consumption, is identical with net national investment, and the latter, in turn, is equal to the change in real national wealth. For the individual economic unit, however, saving is equal not to the change in holdings of real assets, but to the difference between the change in total assets (both tangible and intangible) and total liabilities. The national balance sheet registers the effect of saving on the stock of intangibles as well as tangibles.

The link provided by the saving statistics is imperfect for both conceptual and statistical reasons. To note only some of the principal conceptual differences, there are, first, variations in the treatment of government. In the Department of Commerce estimates of income and saving, government investment and government saving are excluded, while in the Kuznets income estimates, and the Gold

smith saving and wealth estimates, government saving and investment are included, though the Goldsmith estimates exclude military assets. Another important difference is in the treatment of consumer durables, which in both the Department of Commerce and Kuznets income estimates is not considered investment, but in the Goldsmith estimates of saving and wealth is so considered. Finally, there are important differences in the scope and valuation of capital consumption allowances. Beyond the conceptual differences, there are variations in the sources and techniques employed by the different estimators. The broad outlines of the relationships among the different social accounts can, nevertheless, be distinguished. In addition, the saving statistics throw important light on the nature of the different groups of savers in the economy and the forms that saving takes.

Source: Raymond W. Goldsmith, A Study of Saving in the United States, vol. I, 1955, p. 345 (saving, excluding consumer durables, computed by subtraction of estimates of saving in consumer durables for nonagricultural individuals, p. 359, and for agriculture, p. 756). Reprinted by permission of Princeton University Press.

These series provide an estimate of saving by government (thus permitting the derivation of aggregate national saving), and estimates of personal saving subdivided among three major groups-nonagricultural individuals (including private nonprofit institutions and personal trust funds), agriculture, and unincorporated business.

The saving concept underlying these estimates differs somewhat from the concept represented by series F 552-565. While these estimates include all forms of saving covered in series F 552-565, they also cover saving in the form of consumer durables, and of brokers' and dealers' commissions and profits on change of hands of existing assets. In addition, in deriving these estimates of net saving, capital consumption allowances have been valued at replacement cost. Neither set of figures, however, includes saving in the form of soil improvement or additions to military assets. An important difference also exists between the two sets of estimates in the technique of derivation. The estimates in series F 552-565 were derived by the income approach; these figures, with the exception of those for corporate saving, were obtained by the balance sheet method. In this respect, they are similar to the estimates of personal saving presented in series F 638-667, though differences in techniques and in concept cause the actual estimates to differ between the two tables, e.g. because of inclusion in series F 659 (but not in series F 623) of stock issues of small corporations not distributed by security dealers. The following statement from the source (pp. 40-41) provides an indication of the reliability of the estimates:

Evaluation of the possible errors in the individual series from which the estimates of group and national saving have been constructed indicates that the margin of error is hardly under 10 percent for any given year or for the average annual figure in any series, that it is probably in the order of magnitude of 20 to 30 percent in many of them, that it may run even higher in not a few cases, but that the relative margin of error in most cases is reduced for sequences of several years and generally the smaller the longer the period...

Most of the components utilized in building the estimates of saving of any of the major saver groups are statistically independent; and the estimates for the major saver groups are very largely independent of each other except for those on nonfarm households and unincorporated business enterprises. Accordingly since the number of components of saving is large for each of the groups, running to several dozen even if only those of substantial quantitative importance are taken into account, there is reason to assume that errors in one direction, i.e. overstatements or underestimates of saving, made in any one year in some of the comAs ponent series will be offset by errors in the opposite direction in other series. a result, the relative error in the estimates of saving by the major groups, and still more the estimates of broad aggregates such as national or personal saving, may be expected to be considerably lower than the average of the relative errors in the component series. Indeed, it is quite possible that, if we take account of the number of independent component series and their relative size, and even take a pessimistic view of errors in constituent series, the relative error of national or personal saving in any one year does not on the average exceed something like 10 percent. The quality of most of the individual series used in the measurement of saving has undoubtedly improved. It would seem to be substantially poorer for the period before the thirties than for the last two decades, and within the earlier period, in turn, to be particularly poor for the years before approximately 1905. Nevertheless, there is no statistical evidence, such as might be provided by the difference between estimates of saving and investment, that the estimates of aggregate saving have larger relative errors in the earlier part of the period than in the later part. Indeed, from that point of view, the relative error in the estimates would have to be regarded as substantially the same through the thirties, and as considerably lower only for the last decade. There is, however, evidence... that the error is ... in the direction of an overstatement of saving in the first three decades and an understatement during the thirties.

F 552-565. Sources and uses of gross saving, 1929-1970.

Source: U.S. Bureau of Economic Analysis, 1929-1963, The National Income and Product Accounts of the United States, 1929-1965; 1964-1967, U.S. National Income and Product Accounts, 1964-67; and 1968-1970, Survey of Current Business, July issues, table 5.1.

The following are definitions used by the source:

Gross private saving is the total of household and business saving. Saving through government, including government insurance funds, is excluded. Household expenditures for consumer durables, except on residential construction, are not treated as saving. The series is "gross" in that it includes business capital consumption allowances and depreciation on residences.

Personal saving represents the excess of personal income over the sum of personal consumption expenditures and personal tax and nontax payments. It includes the current saving of individuals (including owners of unincorporated business), nonprofit institutions, and private health, welfare, and trust funds. Personal saving may be in such forms as changes in cash and deposits, security holdings, indebtedness, and reserves of life insurance companies and mutual savings institutions, the net investment of unincorporated enterprises, and the acquisition of real property net of depreciation. Inventory profits and other capital gains are excluded.

Gross business saving includes undistributed corporate profits, corporate inventory valuation adjustment and capital consumption allowances, and, for 1943–1953, the excess of wage accruals over disbursements.

Undistributed corporate profits represent the difference between corporate profits after taxes and dividends. Corporate profits after taxes are the earnings of corporations organized for profit which accrue to the residents of the Nation, measured after Federal and State profit taxes, without deduction of depletion charges and exclusive of capital gains and losses. Dividends measure cash dividend disbursements by corporations organized for profit to stockholders who are residents of the United States.

Corporate inventory valuation adjustment is the excess of the value of the change in the volume of nonfarm corporate business inventories, valued at average prices during the period, over the change in the book value of nonfarm corporate inventories.

Capital consumption allowances represent the sum of business depreciation charges and accidental damage to fixed business capital. Business depreciation charges are charges made by private business against receipts for the current consumption of durable capital goods and comparable allowances for nonprofit institutions. They include depreciation charges against owner-occupied houses. Depreciation reported by business is not adjusted for changes in the replacement value of capital goods, except for farm enterprises. Accidental damage to fixed business capital represents the value of the physical losses by fire, natural events, and other accidents to fixed capital of private business not covered by depreciation charges.

Government surplus or deficit is the excess of government receipts over government expenditures as defined in the national income and product accounts. As such, it equals the acquisition of financial assets less borrowing by general government and government enterprises. It also includes new government purchases of land. Net acquisitions of reproducible assets are excluded here because they are included in government purchases of goods and services.

Capital grants received by the United States in 1970 are the Special Drawing Rights allocated to the United States by the International Monetary Fund. These allocations represent additions to the foreign assets of the United States that are not matched by corresponding liabilities. They are considered part of the U.S. net foreign investment and are shown as a source by means of this special entry.

Gross private domestic investment consists of the net acquisitions of fixed capital goods by private business and nonprofit institutions; including commissions arising in the sale and purchase of new and existing fixed assets, principally real estate; and the value of the change in the volume of inventories held by business. It covers all

private dwellings including those acquired by persons for their own

occupancy.

Net foreign investment is numerically equal to the balance on goods, services, and unilateral transfers as measured in the balance of payments statistics. As such, it is equal to the acquisition of foreign assets by U.S. residents less the acquisition of U.S. assets by foreign residents. It also includes the "errors and omissions" item in the balance of payments statistics.

With respect to reliability of these estimates, the Department of Commerce notes that the margin of error in the estimates of gross private saving and its components tends generally to be high. Because personal saving is derived as the difference between two much larger totals, it is subject to large percentage error in both level and movement. Undistributed corporate profits is more accurate, but the corporate inventory valuation adjustment is liable to considerable error, so that the reliability of the two items combined is not high. Furthermore, while approximately half of the estimate for capital consumption allowances is based on fairly solid data, the remainder is estimated on the basis of a variety of sources and methods, and some of these are subject to a wide margin of error.

Series F 566-594. Individuals' saving, by components, in current prices, 1946-1970.

Source: Board of Governors of the Federal Reserve System, Flow of Funds Accounts: Annual Flows, 1946-1971, August 1972, pp. 69-71. Series F 566-594 presents an estimate of personal saving that is conceptually equivalent to the amounts derived in national income accounts (NIA), series F 553, but statistically it is almost entirely independent of NIA data. The NIA series for personal saving is calculated as a residual in current transactions: Personal income less personal taxes less consumption and other current outlays. That residual measures the net flow of funds from current activities of persons and is used for acquisition of capital assets, both tangible and financial, and for repayment of personal debt. Series F 566–594 is a direct measure of those capital acquisitions and debt flows and is thus an estimate of the same net flow of personal saving in terms of the capital uses to which it is put. The basic identity relationship reflected in series F 566-594 is that personal saving equals net acquisition of capital assets less net increase in debt, or alternatively that total sources of funds from saving and net borrowing equals total uses of funds for capital asset purchases.

Series F 566-594 divides these capital account transactions into three sections: Net increase in financial assets, net investment in tangible assets, and increase in debt owed by individuals. Financial

assets consist of claims on others (including other individuals), mainly in the form of money, deposits, securities, corporate equities, and equities in insurance and pension reserves. The amounts shown are net transaction flows, the excess of acquisitions over liquidations; and changes in values of holdings through market price movements. Unrealized capital gains are not included in the figures.

Investment in tangible assets (series F 583) appears net of capital consumption allowances, which are mainly book depreciation charges. These charges are reflected in personal consumption in NIA and personal saving is thus smaller because of them. Increases in personal debt, in the third section, are offsets to asset acquisitions. Funds acquired from borrowing are used either for the asset purchases shown in the table or for consumption or tax payments, both of which decrease saving.

The table on p. 251 compares this capital-account calculation of net investment-which is equal to saving-with the NIA estimate. The capital account version differs in a few conceptual aspects from the NIA definition, and adjustments are made for these differences. The adjustments allow for equities in government life insurance and retirement fund reserves, which are included in assets in this table but not in NIA personal saving; investment in consumer durables, which are treated as capital goods in this table but as current consumption in NIA; and investment company capital gains dividends,