PROBABLY more than half of capital expenditures involve in one sense or another the replacement of existing stock. Among competing hypotheses as to the timing and determinants of replacement expenditures are: (1) they are a fairly constant proportion of capital; (2) they substitute for expansion expenditures, thus stabilizing the annual rate of investment, falling when expansion rises and rising when expansion decreases; (3) they are tied closely or are essentially equal to depreciation charges; (4) they vary with the current rate of profit or flow of funds; (5) they are positively related to the age of capital stock. McGraw-Hill capital expenditure survey data and collateral statistics offer a unique opportunity to test these and related hypotheses. In a recent article, Feldstein and Foot (1971) have utilized McGraw-Hill aggregative reports, along with series from the Department of Commerce on planned capital expenditures. and from the Federal Trade Commission and Securities Exchange Commission on flow of funds, in an analysis of replacement expenditures. This paper offers a partly parallel analysis of both replacement and expansion expenditures on the basis of individual firm data. Key to the analysis is a question which has been included in the McGraw-Hill spring surveys in the years 1952 through 1955 and 1957 to date: Of the total amount you now plan to invest in new plants and equipment in [the current year] how much is for: expansion %; replacement and modernization % ? By applying the indicated proportions to anticipated and actual capital expenditures, estimates have been obtained of expenditures for replacement and modernization and expenditures for expansion. For actual expenditures these estimates related to from 112 to 254 firms in each of the fourteen years from 1954 to 1968, excluding 1956; 1 estimates of anticipated expenditures were available for approximately the same firms. The basic data, price-deflated and otherwise processed as previously reported,2 are as follows:

Politicians almost all talk about the deficit, and almost all decry it. Very few, literally, know what they are talking about. To my dismay I have felt over some years that too many economists fall in the same category. I shall insist, contrary to Ricardian views, that deficits do matter and can matter very much. They can be too small as well as too large, and you cannot even begin to tell what they are until you measure them right. At this time, the real is too small. One can pick from a huge variety of deficits. The federal for the 1991 fiscal year reported by the Office of Management and Budget (OMB), including off-budget and on-budget items, was $269 billion. This compares with an economically more meaningful federal on national income accounts of $190.3 billion, which was just 3.3 percent of gross domestic product. If you were to follow Congressional legislation and arbitrarily exclude social security (and the postal service) from the unified or total OMB budget you can work the up to $321 billion. More sensibly, one can exclude $67 billion for that is, the savingsand-loan bailout, which is at this point merely a financial transaction substituting explicit federal debt for the debt implicit in deposit guarantees. This would get the down to $202 billion. If one looks at a (measured at 5.5 percent unemployment, which I would consider too high), eliminating the effects of the recession along with deposit insurance, the would be $124 billion. Looking at what is called the primary, standardized-employment budget, excluding interest payments along with deposit insurance, one actually finds a substantial surplus, of $71 billion.' There are other, more meaningful measures of the that might well be advanced. These would entail: 1) adjustment for the inflation tax on the holders of existing debt; 2) including the offset of state and local government surpluses, particularly since federal grants contributing to those now meager surpluses comprise a major element in the federal deficit; and 3) excluding net capital expenditures, as would be consistent with private business accounting. These most appropriate adjustments, as shown in Table 1A, bring the deficit down from its 1991 figure of $269 billion to a paltry $17 billion. Still another way of looking at the budget is to note that an appropriate concept of balance for the government in a growing economy, like that for any business, is that the debt grow no faster than income or output, so that the debt:income ratio does not rise, as shown in Table 1B. The 7-percent growth that the economy has experienced in previous, nonrecession years would then imply an increase in debt-or deficit, aside from the effects of the recession-of $188 billion. This in a meaningful sense would be balance; but that is again 3.3 percent of GDP, almost precisely the actual federal on the national income account. Furthermore, that includes a substantial component due to the recession. By standards of constant debt: GDP ratio, a high-employment, cyclically adjusted budget would be in substantial surplus. The one sophisticated objection frequently offered to budget deficits without, I must say, paying much attention to how

A couple of years ago a New York Stock Exchange study (1974) pointed to a of some $650 billion by 1985. Treasury Secretary William E. Simon, comparing his estimates of capital requirements current dollars over the next decade with capital expenditures current dollars over the last decade, came out with a gap of over 2-1/2 trillion dollars without noting the noncomparability of prices (p. 3871). We have indeed a host of estimates from a number of econometric models, government bodies and private institutions, from Barry Bosworth, James Duesenberry and Andrew Carron and many others. A major Bureau of Economic Analysis study under the direction of Vaccara projected a total of $986.6 billion, 1972 prices, for business fixed investment from 1975 to 1980, or 12.0 percent of cumulative gross national product, in order to insure a 1980 capital stock sufficient to meet the needs of a full employment economy, and the requirements for pollution abatement and for decreasing dependence on foreign sources of petroleum (p. 7). Scarcities are sometinmes seen terms of sources of financing. Benjamin Friedman wrote 1975, To an unusually great extent, financial considerations may act during this period [1977-811 as effective constraints on the amount of fixed investment which the economy aggregate is able to do (1975, p. 52). In May 1976, however, Allen Sinai declared, There are no financial shortages of any consequence (p. 2). But with the plethora of articles, studies, claims and warnings, what meaning can we attach to the notion of a capital shortage'? In what sense can there be a shortage a free economy where markets are cleared by the impetus of price movements? In an uncontrolled, competitive system, the rate of investment is not imposed as a prior constraint. Business investment, particular, is the resultant of the utility-maximizing saving propensities of households and the profit or wealth-maximizing production decisions of business. These are subject to the constraints of the general economic atmosphere determined by the monetary and fiscal authorities of government, particular tax and monetary influences, and general currents of the world. Any argument that there is a capital shortage must either imply a literal failure of market clearing or some standard external to the economic system. A failure of markets to clear an equilibrium sense implies fixed or sticky prices. If government were to control prices and set those for capital goods too low, the quantity of capital goods demanded could exceed the quantity of capital goods supplied. Perhaps more to the point, government regulatory agencies might hold prices of certain products, such as electric power, so low that, while the quantity of electric power demanded might be very high, firms anticipating continued low prices would not find it profitable to invest the capacity to meet future needs. Similarly, there may be price fixing financial markets. If the monetary authority and/or inflation force up interest rates while regulatory *Williarn R. Kenan Professor of Economics, Northwestern University, and Senior Research Associate, National Bureau ot Economic Research. I amil indebted to Martin Feldstein, Benjamin Friedman, Marc Nerlove and Beatrice Vaccara for helpful commiients.