The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
FEDERAL RESERVE BANK OF S T . L O U IS O C T O B ER 1970 Stabilization Policy and Inflation I n FLATION, caused by an excessive rise in total spending, has been a serious national problem since the mid-Sixties. The monetary and fiscal actions of the past two years have been successful in moderating the surge in total spending, and initial adjustments thereto were manifested by reduced real output. More recent adjustments apparently have included a de cline in the rate of price increase. An immediate consequence of the batde against inflation has been insufficient total spending to absorb the growing labor force and plant capacity. But suc cess in correcting established inflation has always been costly. If the growth of total demand for goods and services is restrained to a reasonable rate, output and employment growth will gradually accelerate as in flationary forces recede. Two strategic questions for stabilization policy in the immediate future are: what is an appropriate rate of spending growth? and what rate of growth in money will foster this spending growth? This Bank has estimated that a 3 per cent rate of increase in money will provide about a 4.5 per cent rate of spending growth a year from now, while a 5 per cent rate of increase in money will provide about a 6.5 per cent rate of spending growth.1 Monetary Actions Monetary actions have been a major factor in mod erating the growth of total spending in 1969 and 1970, just as in 1966 and 1967. The money stock did not grow from April 1966 to January 1967 and then ex panded at about a 7 per cent annual rate in 1967. Similarly, eight months of little expansion in the money stock from June 1969 to February 1970 have been followed by about a 6 per cent growth rate in the money stock since February. In early 1967, total spending growth slowed markedly to a 3.6 per cent annual rate in the first half of the year. Similarly, after autumn 1969, total spending growth slowed to a 4.1 per cent rate. In both instances, monetary au thorities reacted to the slowdown by shifting to a relatively less restrictive monetary policy, with a view iSee “A Monetarist Model for Economic Stabilization,” this Review (April 1970), pp. 7-25; and “Economic Slowdown and Stabilization Policy,” this Review (September 1970), pp. 6 and 7. Page 2 M o n e y Stock Ratio Scale Billions of D o lla rs 23 0 Ratio Scale Billions of D olla rs 230 M o n t h ly A v e r a g e s o f D a i l y F ig u r e s S e a s o n a lly A d ju ste d 225 225 220 220 215 215 210 210 205 205 200 200 195 195 190 190 185 185 180 180 175 175 170 170 165 160 Jan. 67 Jan.'69 tl I 1 I l 1 I I I l 1967 June'69 Feb .7 0 ^ Sept 70 — 1968 1969 :— 165 160 1970 P e r c e n t a g e s a r e a n n u a l ra t e s o f c h a n g e fo r p e r i o d s in d ic a t e d . L a te st d a t a p lo tte d : S e p te m b e r to avoiding inordinate restraint on total spending and output. The more rapid growth of the money stock since February 1970 has been fostered by increases in Fed eral Reserve credit and member bank reserves and has been accompanied by rapid growth in time de posits. Total Federal Reserve credit grew at a 9 per cent annual rate from February to September, com pared with a 2.7 per cent rate from January 1969 to February 1970. Total member bank reserves grew at a 9.5 per cent annual rate from February to Septem ber 1970. Time deposits at commercial banks have increased rapidly in 1970, because banks have been better able to compete for funds than in 1969. Certain interest rate ceilings were relaxed in January and completely suspended on 30- to 89-day large CD’s in June, and short-term market interest rates have declined. Time deposits grew at a 12 per cent annual rate from January to June and at a 35 per cent rate from June to September. About half of the increase in time de posits since February of this year has been in the form of large negotiable certificates of deposit, repre senting largely a reintermediation of funds previously flowing through nonbank channels. FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER. 1970 The inflow of time deposits to commercial banks has not been at the expense of savings flows into nonbank financial intermediaries. Net savings flows to nonbank financial intermediaries have increased rapidly in 1970 compared with 1969. Savings and loan shares grew at an 8 per cent annual rate in the first eight months of 1970 compared with 2.6 per cent in 1969. Mutual savings bank deposits increased at a 6.3 per cent annual rate in 1970 compared with 3.9 per cent in 1969. Short-term market interest rates have declined in 1970, because of more rapid monetary expansion and slower growth of total credit demand. Three-month Treasury bill rates declined from 7.S7 per cent in January to 5.84 per cent in early October. The de cline in short-term market interest rates relative to the discount rate, together with the greater ability of banks to attract time deposit funds, has induced mem ber banks to reduce their borrowings at the Federal Reserve from about $900 million last spring to about $500 million in late September and early October. In contrast to short-term interest rates, yields on long-term corporate bonds have changed little on balance in 1970, because of continued heavy demands for funds relative to the supply. Interest rates on highest grade seasoned corporate bonds, after reach ing a peak of 8.60 per cent in mid-1970, declined to about 8 per cent in early October, about the same as at the beginning of the year. Total Spending and Output Monetary and fiscal restraint in 1969 and 1970 have been successful in moderating the growth of total spending. Total spending expanded at a rapid 9.4 per cent rate from mid-1967 to mid-1968. After imposition of the surtax and some Government spending cuts in T a b le 1 C h a n g e s in S p e n d i n g a n d Output (a n n u a l rates of c h a n g e in current d o lla rs) 11/67 to 11/68 Total S p e n d in g (G N P ) 11/68 to 111/69 111/69 to 111/70 9 .4 % 7 .8 % Real O u tp u t 5 .0 2.9 -0 .5 C on su m p tio n 4 .5 % 8.1 7 .9 6 .9 G ro s s Private Dom estic Investm ent 1 5 .0 10.2 - 4.9 G o ve rn m e n t Purchases of G o o d s a n d Services 1 1 .2 6 .3 3 .4 Federal G o ve rn m en t Purchases of G o o d s a n d Services 9.8 2.9 — 3 .4 1 2 .7 9 .7 9 .7 State a n d Local G o ve rn m e n t Purchases of G o o d s a n d Services mid-1968, spending grew at a slower, but still rapid, annual rate of 7.8 per cent until the third quarter of 1969. Following reduced monetary expansion begin ning in early 1969, spending slowed to a more moder ate 4.5 per cent rate of increase from the third quarter of 1969 to the third quarter of 1970, a rate slightly higher than the commonly assumed trend growth of productive capacity (Table 1). Growth in real output slowed simultaneously with the reduction of total spending. The unsustainable 5 per cent increase of real output from mid-1967 to mid-1968 slowed to about a 3 per cent annual rate from mid-1968 to the third quarter of 1969, and then to a 0.5 per cent decline from the third quarter of 1969 to the third quarter of 1970. Within the private sector, the reduction in spend ing growth in the last two years was greater in investment expenditures than in consumption. Con sumer spending growth slowed gradually to a 6.9 per cent rate in the last year. The trend of gross private domestic investment changed sharply, slowing from a 15 per cent rate of expansion from the second quarter of 1967 to the second quarter of 1968, to a 10 per cent annual rate from the second quarter of 1968 to the third quarter of 1969, and then to a 4.9 per cent rate of decline from the third quarter of 1969 to the third quarter of 1970 (Table 1). The initial absolute de cline in investment expenditures occurred about two quarters after monetary restraint was initiated and simultaneously with early adjustments in total spend ing. The decline was comparable to investment be havior at the end of 1966 and the beginning of 1967. Page 3 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER. 1970 Employment Employment responded only slowly to reduced growth in spending and output. Payroll employment growth slowed slightly in the last half of 1969 but nonetheless continued to grow strongly through March of 1970, before declining at a 2.4 per cent rate from March to September. In 1969 and early 1970 firms adjusted mainly overtime and hours of work rather than reduce their trained labor force. Reduced spending and production since last spring have forced employers to reduce their labor force. Despite recent declines, however, employment relative to population of working force age (16-64) was higher this past summer than at any time in the Fifties or the Sixties before 1967. Corporate profits after taxes, a major source of in vestment funds, responded sharply to reduced total spending and the continued rise in costs. They de clined 12 per cent from the second quarter of 1969 to the second quarter of 1970 after showing littie net change in the previous three years. Profits declined from 6.8 per cent of GNP in 1965 to 5.2 per cent in 1969 and to 4.6 per cent in the first half of 1970. Purchases of the public sector have also grown more slowly in the last two years, and priorities have been shifted. Total government purchases (Federal, state, and local) have grown at about a 5 per cent annual rate since mid-1968, compared with about a 14 per cent rate from 1965 to 1968. The trend of Federal Government purchases has shifted most markedly, remaining about unchanged since mid1968 after growing at a 14.8 per cent annual rate from mid-1965 to mid-1968. State and local government purchases have grown at about a 9.7 per cent annual rate since mid-1968, compared with a 13 per cent rate in the preceding three years. Defense purchases have been cut back severely, while nondefense purchases of the Federal Govern ment have continued to grow. Recentiy, defense spending has been about $2.4 billion lower than at its peak at the end of 1968, and 14 per cent fewer people are currently employed in defense and defense-related industries than in mid-1968. Defense spending rose from 8.5 per cent of total spending (GNP) in the 1962-64 period to about 9 per cent at the peak in 1967-68, and has since declined to 7.6 per cent in the third quarter of 1970. 4 Page The unemployment rate averaged 5.2 per cent in the third quarter of this year compared with 3.6 per cent a year earlier. The recent 5.2 per cent rate com pares with an average of 5.8 per cent in 1961-64 when prices were not rising significantly. The unemployment rate for married men in the third quarter of this year was 2.8 per cent compared with 1.6 per cent a year earlier and an average of 3.6 per cent in 1961-64. Prices Price trends over a considerable period of time generally reflect the trend of spending relative to the trend of productive capacity. Many prices, how ever, are slow to change, and hence overall prices do not move up quickly when spending growth acceler ates, nor stop rising soon after the growth in spending moderates. Restrictive monetary actions of 1969 and early 1970 have been sufficient to halt the acceleration of inflation, and there are scattered indications that the rate of inflation may be slowing. Consumer price advances may have moderated since last April. From April to August, consumer prices rose at a 4.5 per cent annual rate, compared with a 6 per cent rate in the preceding year. The decline in recent months has been the result of a less rapid rise in the prices of services and food. Although these data sug gest that the rate of consumer price advance may be slowing, it is difficult to say at this time whether this constitutes the beginning of a definite change in trend or is merely an irregularity in the series. Such an irregularity occurred in farm products and food prices in the summer of 1969. Wholesale prices of industrial commodities rose at a 2.1 per cent annual rate from May to September, compared with a 3.9 per cent rate in the preceding year. While these developments are encouraging, sim- OCTOBER, 1970 FEDERAL RESERVE BANK OF ST. LOUIS Prices Ratio S ca le 1 9 5 7 -5 9 = 1 00 Ratio S c a le 1 9 5 7 -5 9 = 1 00 S o u r c e : U .S . D e p a r t m e n t o f L a b o r P e r c e n t a g e s a r e a n n u a l ra t e s o f c h a n g e fo r p e r i o d s in d ic a t e d . L ate st d a t a p lo tte d : C o n s u m e r - A u g u s t ; W h o l e s a l e - S e p t e m b e r ilar consecutive months of improvement in industrial prices in 1969 did not prove to be lasting. Probably the most significant improvement in wholesale prices in 1970 has been the lessening of upward price pres sures in the markets for metals. An additional group ing of wholesale prices, the daily spot market index of thirteen raw industrial prices which are believed to be highly sensitive to demand conditions, has declined sharply since February of this year. The index reflects world as well as domestic market conditions. One reason for some optimism regarding price prospects in the near future, despite only uncertain improvements so far, is that growth in total spending has remained moderate for over a year. Past experi ence indicates that changes in spending growth have usually affected output initially, and about three quarters after the reduction in spending and output, a reduction in the rate of inflation has generally be come evident. Since monetary restraint was initiated in early 1969, and did not materially reduce spending until the last half of 1969, the lags in price adjust ments have not been unexpectedly long, especially in view of the strong upward price momentum which had developed. An increasingly larger impact on prices can be expected as the period since the onset of restraint lengthens, provided stimulative stabiliza tion policies are not pursued too actively. Inflationary Expectations and Monetary Policy The course of total spending is influenced primarily by monetary and fiscal actions. Inflationary expecta tions, which are formed and dissipated only slowly, have their main effect on the division of total spend ing between prices and real output. Some of the im pact of inflationary expectations since mid-1969 has been manifested in a slowing of real output growth. Inflationary expectations continue to have a bearing on the formulation of monetary policy. Monetary au thorities are currently faced with five years of cumu lated inflationary expectations. Monetary actions to combat inflation must now be sufficiently strong to break the pattern of inflationary expectations if the ground work is to be laid for resumed real growth with relative price stability. A vital current question is whether the slowing of real output growth from mid-1968 through mid-1969 and then three or four quarters of no expansion in late 1969 and in 1970 has been sufficient to reduce infla tionary expectations adequately. The more expansion ary monetary actions thus far in 1970 may contribute to real output growth and greater employment, and probably do not run a risk of fostering a growth of total spending which reinforces inflationary expecta tions. But significantly more rapid monetary expan sion than has prevailed so far in 1970, if continued for any length of time, is likely to prolong the inflation problem unduly. Total spending growth must be appropriately limited if we are to complete within a reasonable period the transition to an economy of relative price stability. Page 5 Some Lessons to be Learned from the Present Inflation A Speech by DARRYL R. FRANCIS, President, Federal Reserve Bank of St. Louis, to the University of Miami Commercial Bank Forum, Doral Country Club, Miami, Florida, September 24, 1970 I t IS GOOD to have this opportunity to discuss with you some of my views regarding the inflation which has plagued our economy for the past five or six years. It seems important to me that all of us understand current economic stabilization problems and the efforts of our public officials to handle them. If our economy is ever to contain inflation, we must have leaders who are aware of the causes of inflation, of its costs to our society, and of the difficulties inherent in reducing inflation once it is allowed to run rampant. Since late 1964 prices have been rising with increas ing rapidity, with effective attempts to control this situation only in the last two years. Now that we have undergone for several years our worst inflation since World War II, it is timely to reflect on this experience and to draw some conclusions. If the obvious lessons of the last six years are remembered in the future, the likelihood of repeating unnecessary mistakes should be reduced considerably. Only by avoiding such mistakes can our economy experience economic growth at a high level of employment with a reasonable stable price level. These, of course, are widely accepted national economic goals. The Current Inflation Before going to the main part of my remarks, let me review the course of the present inflation and steps we have taken to curb it. Our economy had a period of substantial price stability from 1958 to 1964. During this period the wholesale price index was virtually unchanged, the consumer price index rose about one per cent a year, and the GNP price de flator rose at only a slightly faster rate than con sumer prices. Those six years of relatively stable prices marked the end of the inflation generated during World War II and the Korean War. By 1964 we had Page 6 achieved a high level of resource utilization, and prospects were good for continuing price stability. Our economy was moving into a period when it could be said that the goals of economic stabilization had been essentially achieved. But, then, from 1964 to just recently, an era of ever more rapidly increasing prices developed. This inflation was caused by growth of total spending for goods and services at an eight per cent annual rate from 1964 to 1968, or about twice as rapid as our economy’s ability' to increase the production of goods and services. This excessive expansion in total spend ing was fostered by very stimulative monetary actions of the Federal Reserve System, supported by the Ad ministration, the Congress, and public opinion. The nation’s money stock, except for a brief interlude in 1966, rose at rates which approximated those prevail ing during the World War II and the Korean War inflations. One of the main reasons for such extremely high rates of monetary growth appears to have been a de cision to expand welfare programs and the Vietnam War simultaneously and to finance these increases, in large part, by inflating the monetary system rather than exclusively by taxes or borrowing from the planned saving of the public. These latter two sources of Government finance are basically noninflationary, because most of an increase in Government expendi tures is then made at the expense of private expendi tures.® On the other hand, when increased Govern ment expenditures are accompanied by excessive monetary expansion, there is little, if any, direct reduc tion in private spending. In fact, there are strong secondary repercussions from such a method of Gov ernment finance which greatly enlarge spending by business firms and consumers. ‘ For further elaboration of this point, see “The ‘Crowding Out’ of Private Expenditures by Fiscal Policy Actions,” on pages 12-24 of this Review. FEDERAL RESERVE BANK OF ST. LOUIS Significant actions to curb inflation, either fiscal or monetary, were delayed well into 1968, despite acknowledgment of the existence of a serious inflation and often-expressed desires to do something about it. Then, in mid-1968 a program of reducing the rate of increase in Government spending and increasing taxes was adopted with a view to bringing the excessive rate of growth in total spending more into line with growth in potential output of goods and services. However, it was not until the rate of growth in the money stock was reduced markedly in 1969 that the stage was set to bring total spending more in line with growth in our economy’s productive potential. Curbing such a long inflation has proven, once again, to be both slow and costly. Only in recent months has there been any evidence of a slowing in the rate of price increase, and the period is still too short to conclude definitely that there has been a marked waning of inflation. I am confident, though, that if monetary growth is limited to a moderate rate for the next several months, there will be significant, but slow, abatement of inflation. On the cost side of setting the stage for reducing inflation, there has been a slowdown in output of goods and services accom panied by a rise in unemployment. It should be pointed out, however, that the present slowdown has been much less than during any of the other such slowdowns during the past twenty-five years. Some Lessons I turn now to the main theme of this discussion — some lessons to be learned from the present inflation. In developing these lessons, I will point up some of the failures of commonly accepted economic ideas regarding economic stabilization which have been in strumental in permitting our present inflation to de velop. By the commonly accepted economic ideas, I mean the form of analysis taught in the majority of undergraduate economics courses for the past twentyfive years. Although most economists have now ad vanced beyond this rather limited analysis, it still per meates the thinking of the general public, many busi ness and financial leaders, news writers, politicians, and public policymakers. Inflation a Monetary Phenomenon One lesson, and I believe the most important, is that inflation is primarily a monetary phenomenon, where as the conventional view has placed great stress on Government deficits, union power, and business monopolies as causes of inflation. There is now con OCTOBER, 1970 siderable evidence from studies at our Bank and by others that the excessive total spending which led to a high and accelerating rate of price advance was generated, for the most part, during 1964 to 1968 by the exceedingly high rate of monetary expansion of that period. As I pointed out earlier, rapid growth in Govern ment spending and deficits is not a major source of inflation unless accommodated by growth in the money stock. Likewise, upward pressures on prices from union or business monopoly actions are not likely to initiate a period of inflation unless accompanied by rapidly rising total spending. Such a rise in total spending requires expansive actions on the part of monetary authorities. Thus, the price level effect of often mentioned fiscal and monopoly causes of infla tion can be contained if they are not validated by monetary actions which generate a rapid growth in total spending. Popular Economic Analysis Inadequate Another lesson to be learned from our recent in flation is that the popular economic analysis of the past quarter century has been ill-equipped to correct inflation. A major aspect of this conventional analysis, as I mentioned earlier, is that the general price level is believed to be only remotely influenced by mone tary actions. Instead, in addition to fiscal actions, con siderable emphasis is given to controlling undesired price level movements by measures to reduce monop oly power or by exertion of Government pressure, such as guidelines, to induce those who set prices to act in a manner consistent with national objectives. This view, which was developed in large part from the experience of the Great Depression of the 1930’s, is still prevalent in the economic theory which underlies much of popular thought regarding economic stabili zation. By being developed within such a depression orientation, this body of theory is not particularly useful, in my opinion, in developing programs to cope with an inflationary situation such as we have ex perienced since 1964. Also, reliance on such devices as the wage-price guidelines during the 1960’s became a substitute for sound stabilization policy, and thereby, contributed to the emergence of inflation. Roles of Monetary and Fiscal Actions A further lesson from our experience of recent years is that monetary actions rather than fiscal actions should be given the major role for stabilizing Page 7 FEDERAL RESERVE BANK OF ST. LOUIS the economy. Until recently, fiscal actions in the form of Government spending and taxing programs have been given the main emphasis in economic stabiliza tion efforts to the virtual exclusion of monetary ac tions. Such a development was an outgrowth of con ventional economics, which for the past thirty-five years has taught that Federal Reserve actions exer cise little influence on total demand for goods and services. According to this conventional thought, changes in the money stock bring about changes in market interest rates, while total demand is little influenced by interest rate movements. Consequently, monetary actions have been thought to be of little use in any program of economic stabilization. On the other hand, increased Government expenditures are viewed as adding directly to total demand, and tax reductions are thought to add to disposable income which subsequently is used to purchase goods and services. Consequently, this view has argued that fiscal actions have an immediate and powerful influence on total spending. This conventional analysis, possibly because of its simplicity which helps in the teaching of under graduate economics, has received wide acceptance as evidenced in discussions of economic stabilization by the general public, in the press, in the Congress, and even in some of the Reports of the Council of Economic Advisers during the mid-1960’s. It should be pointed out that this view of the influence of fiscal actions does not take into consideration the importance of choice among the three alternative means of financing Government expenditures — taxes, borrowing from the public, and monetary expansion. At the St. Louis Federal Reserve Bank we have reported several studies regarding the relative im portance of monetary and fiscal actions for economic stabilization. Our empirical studies for the United States economy from 1919 to 1969 and for several foreign countries in the post-World War II period support the view that monetary actions, measured by changes in the money stock, should receive the main emphasis in economic stabilization, not fiscal actions. The accelerating inflation of the last half of the 1960’s can be attributed, in large part to the great emphasis given to fiscal actions and the downgrading of monetary influence. Monetary authorities did not reduce the rapid rate of monetary expansion during a large part of that period because there was a desire to let fiscal actions curb inflation and a belief by some that only fiscal actions would be effective. Then, when restrictive fiscal actions were taken in mid-1968 — the surtax and slower increases in Government Page 8 OCTOBER, 1970 spending — many economists, on the basis of preva lent theories, predicted “fiscal over-kill” by early 1969. In response to such predictions, monetary authorities engaged in even more expansionary actions in the last half of 1969. Continuation of accelerating inflation after fiscal actions had been expected to provide a quick cooling of the inflationary fires should bum firmly into our memories the lesson that monetary actions are more effective than fiscal actions in promoting economic stability. But I do not want there to be any misunderstanding regarding our view concerning fiscal actions. Some have interpreted us as saying that Government spend ing and taxing have no influence on the course of the economy, but this is not the case. Our research indi cates that accelerations and decelerations in the rate of increase in Government spending, even if there is no accommodating change in money, cause cor responding short-run changes in total spending. Also, financing of large Government deficits has in the past caused the Federal Reseive to expand the money stock at excessive rates. This was one reason for rapid monetary growth in 1967 and 1968. Finally, Govern ment spending and taxing programs, insofar as they affect the amount of resources allocated to private investment and to Government outlays of a similar nature, may have a significant influence on long-run economic growth. ltnplementation of Monetary Policy An additional lesson we have learned from our pres ent inflation is that the usual method of carrying out monetary policy in the 1950’s and 1960’s was faulty. Although stated monetary policy was to control infla tion, the method used for implementing this policy actually contributed to the inflation rather than to its control. Discretionary monetary policy was reinstated in 1951, after its suspension during World War II and up through the early part of the Korean War. The pur pose of the 1951 change was to permit monetary authorities to fight the inflation of the Korean War. In conducting its monetary policy responsibilities since then, the Federal Open Market Committee until very recently has relied largely on measures of money market conditions as guides to its operations. I am sure that most of you are familiar with the view that falling interest rates or rising free reserves indicate easy monetary actions, while tight actions are indi cated by rising interest rates or falling free reserves. FEDERAL RESERVE BANK OF ST. LOUIS Such a view was in general agreement with the widely held belief that monetary actions work pri marily through changes in market interest rates. It also was in agreement with the view that the Federal Reserve has great ability to “set” market interest rates. Recent research and experience, however, have tended to reject these propositions. For example, rapid monetary expansion, such as in 1967 and 1968, stimulates total spending and thereby generates rap idly growing demand for credit and rising interest rates. By using market interest rates to indicate the thrust of monetary actions, many public policymakers con cluded that despite very rapid monetary growth, rising interest rates were evidence of monetary re straint during 1967 and 1968. In fact, there was a belief by many that the extent of the increase in market interest rates was too great because of the dislocations which occurred in the savings and hous ing industries. There was a desire to hold back the extent of interest rate increases, but attempts to do so required injections of bank reserves which contributed to a rapid growth in the money stock. This, in turn, fostered excessive total demand and fed further the fires of inflation. In retrospect, it is now apparent that the traditional reliance on such measures of money market conditions as market interest rates contributed to our present inflation. Sound economic stabilization requires guides to the thrust of monetary actions other than money market conditions. Recent experience demonstrates that use of a monetary aggregate, such as the money stock, would have produced far better results than we got during the last half of the 1960’s. Excessive total spending followed the very rapid rates of monetary expansion from early 1965 to early 1966 and during 1967 and 1968. But when money ceased to grow in the last eight months of 1966 and grew only slowly in 1969, total spending slowed markedly after a short lag. Conclusions I have advanced from this casual analysis have been supported by a growing body of empirical research. Importance of Price Anticipations A further lesson concerns the importance of price anticipations in the inflationary process and in the curbing of inflation. As I mentioned earlier, much of economic theory upon which recommendations for stabilization actions during the 1960’s rested did not give adequate consideration to the basic forces in fluencing the price level. Little consideration was OCTOBER, 1970 given to the well-known fact that consumers, business men and labor unions do take into consideration anticipated price level changes when making decisions to purchase goods and services in the present and when negotiating contracts for the future. Once growth of total demand exceeds growth of potential output and inflation has been underway for a period of time, these decision-makers tend to extrapolate the past trend of prices into the future in an attempt to protect their positions from the ravages of inflation. This process provides a momentum to inflation which causes prices to continue to rise. This infla tionary momentum may carry on well after public policy steps have been taken to bring total demand growth into line with potential output growth. Such a development has been seen in recent experience in which, after a year or so of reduced rate of growth of total spending, the price level has continued to rise rapidly. Another manifestation of inflationary expectations during the past several years has appeared in financial markets. There is a considerable body of economic theory which holds that market interest rates are greatly influenced by expected price level movements. This proposition was not incorporated into the con ventional theory underlying stabilization efforts of the 1960’s. We who maintain that market interest rate movements reflect inflationary expectations argue that when prices are expected to rise, borrowers are willing to pay higher interest rates because they will pay back with depreciated dollars. In addition, any delay in making purchases using borrowed funds will result in high costs in the future. We also argue that lenders will ask for higher interest rates in order to protect the purchasing power of their funds. Thus, both de mand and supply forces during a period of inflation lead to higher and higher interest rates. Many who followed conventional views were at a loss to explain the marked rise in interest rates from 1965 to 1968 at a time when the money stock was rising rapidly. At the St. Louis Federal Reserve Bank we have reported empirical evidence that inflation caused almost all of this increase in market interest rates. The recent experience demonstrates that rapid monetary expansion produces high, not low, interest rates. The truth is the reverse of conventional wisdom regarding interest rate movements. This lesson leads to the conclusion that the theo retical foundation of economic stabilization must give adequate recognition to the pervasive influence of Page 9 FEDERAL RESERVE BANK OF ST. LOUIS price level expectations. Not to do so, would be to repeat the mistakes of the past. I believe that this lesson has already shown up in the expressed views of many policymakers, but there has been little evi dence that it has been learned by the general public, by the Congress, or by economic commentators in the news media and market news letters. OCTOBER, 1970 Regulation Q and Control of Inflation reduced purchasing power of fixed income groups and of holders of wealth in the form of fixed money claims. It is true that if inflation is anticipated correctly, a large number of individuals can adjust their contracts and wealth holdings so as to avoid most of the effects of rising prices. And, in recent years Congress has kept Social Security benefits more or less abreast of the price increase, thereby helping to maintain the purchasing power of a large number of retired persons. Another lesson from the recent experience is that ceilings on interest rates paid by commercial banks on time deposits, set by the Federal Reserve under Regulation Q, are not an effective device for slowing growth in total spending, as many maintained in the late 1960’s. Instead, such ceilings merely create in efficiencies in our financial markets. Commercial bank ers are well aware of the rechanneling of loan funds away from banks and into such markets as the one for commercial paper when free market interest rates exceed Regulation Q ceiling rates. However, when the rate of inflation is changing rapidly, and holders of wealth attempt to adjust their holdings, there are losses in addition to that from reduced purchasing power. For example, the great drop in the bond market during our present inflation and the recent bear stock market are partly a con sequence of attempts of investors to adjust to ac celerating inflation. This recent experience demon strates that even the stock market may not be a very good hedge against inflation when the rate of price advance is accelerating. Some have argued that since banks can make fewer loans under such circumstances, total spending will be restrained. The fact of the matter is that while spend ing by potential bank borrowers may be reduced, spending by those who have access to the money markets will rise by about the same amount. As a result, total spending is little affected by manipulation of Regulation Q interest rate ceilings. There are also considerable losses to the whole economy resulting from the adjustment process which accompanies accelerations in the rate of price advance. Inefficiencies develop in product, resource, and finan cial markets in the process of adjusting prices and contracts to rising prices. Normal business transac tions become more difficult. For instance, we have reports that some business firms in recent years quoted list prices only on a day-to-day basis. Their salesmen were required to contact the home office before any price could be quoted. Labor contract negotiations become more difficult to settle. In financial markets investors have to pay greater attention to ascertaining the impact of inflation on their portfolios and on alternative outlets for their funds. These ceiling rates on time deposits, however, have led to inefficiencies in the flow of funds and in utiliza tion of real resources in recent years. As a result of Q, customary movements of loan funds from one corporation to another through commercial bank channels flow instead through the more direct com mercial paper market or through the less direct and less efficient Euro-dollar market. Such rechanneling of loan funds reduces the size of the commercial bank ing system relative to the total market for funds, a process which is not essential for stabilization policy. Other interest rate ceilings, including state usury laws and ceilings set on funds raised by savings and loan associations and mutual savings banks, also lead to less efficient channeling of funds and use of resources. Costs of Adjusting to Inflation Another lesson from recent experience is that there are great costs of adjusting to accelerating inflation. Everyone is familiar with such losses from inflation as Page 10 Costs of Curbing Inflation A final lesson is that curbing a rapid, prolonged and accelerating inflation is a slow and difficult proc ess, and is not without considerable costs. As I mentioned earlier, anticipations of price increases provide a powerful momentum to inflation. Such an ticipations respond slowly to actual price movements and are not reduced until the rate of inflation has actually subsided for some time. As a result of this slow process of reducing anticipated rates of price increase, the general price level continues to rise rapidly for some time after restraint is applied to growth in total spending. FEDERAL RESERVE BANK OF ST LOUIS Many have been surprised and disappointed that restraint of the past two years has not produced greater results in terms of the price level. Some have even expressed despair at ever seeing relative price stability again. It should be pointed out, however, that inflation was permitted to develop for almost five years before effective restraint was applied. By then, inflation was moving along under its own mo mentum, and only moderate restraint was applied. It should, therefore, not be surprising that five years of inflation cannot be eliminated in a short period of time. Moreover, it should be remembered that the inflations of World War II and the Korean War were not curbed until the late 1950’s and that much greater restraint was applied in that effort. There is also considerable cost in eliminating infla tion. With restricted growth in total spending and with prices continuing to rise for some time, output of goods and services stagnates or is reduced. As a consequence, there is a loss of jobs and income to many individuals and a loss of goods and services to the whole of our society. Labor strife is accentuated, as we now see, when unions attempt to catch up with inflation and to anticipate further inflation at a time of declining corporate profits. OCTOBER, 1970 Conclusions Let me now draw a few general conclusions from this discussion of our recent experience. Inflation, be cause of the many problems and costs it creates, should never be permitted to start. This may seem obvious and trivial, but many have argued that these costs are small compared to alleged large gains flow ing from a high level of employment. Our research indicates, however, that inflation is not required for our economy to have a high level of employment. Another conclusion is that the main body of eco nomic thought of the 1950’s and the 1960’s has not proven very useful in handling economic stabilization problems. In fact, there is considerable evidence that reliance on this body of thought contributed greatly to the present inflation —both as a cause of rapid price level advances and as a hindrance to their control. Finally, monetary policy has a major responsibility for promoting price level stability. If such policy is to be applied in an effective manner, the public, the Congress, the Administration, and the Federal Re serve should reflect on the lessons to be learned from the experience of the past six years. Page 11 The “Crowding Out” of Private Expenditures by Fiscal Policy Actions by ROGER W. SPENCER and W ILLIAM P. YOHE Fiscal policy — F ed eral G overnm ent spending and taxing program s — was given the dom inant role in econom ic stabilization efforts during th e d ec a d e o f the I9 6 0 ’s. T h e in com e tax cut o f 1964 was design ed to accelerate the m ovem en t tow ard full em ploym en t after abou t th ree years o f what was considered by som e a rather slow rate o f econom ic expansion follow ing the recession o f 1960-61. T he incom e tax surcharge an d a reduction in th e rate o f increase in G overnm ent spending w ere ad o p ted in 1968 to curb th e inflation o f the last h alf o f th e I9 6 0 ’s. T he theoretical rationale frequ ently given by stabilization officials for such reliance on fiscal actions w as the sim ple K eynesian m ultiplier analysis fou n d in a large num ber o f econom ic textbooks. T h e sim ple form o f the m ultiplier process holds that an increase in G overnm ent expenditures or a d ecrease in the rate o f taxation induces rep ea ted rounds o f spending by consum ers and business firms, resulting in a m ultiple expansion o f total spending. A m ultiple reduction o f total spending is said to result from fiscal changes opposite those just m entioned. This analysis gives little recognition to the influ en ce on total spending o f financing a deficit, or disposing o f a surplus, by altering th e am ount o f G ov ernm ent borrow ing from the gen eral p u blic or th e rate o f m onetary expansion. T he extent to w hich this analysis g u ided th e recom m endations o f stabilization authorities dur ing the 1960’s is in dicated by an exam ination o f th e A n n u a l R e p o r t s o f the President’s Council o f E conom ic Advisers (CEA). T he m ultiplier process just m en tion ed p la y ed an im portant role in sh a p ing the CEA ’s view o f econom ic stabilization and was sp elled out several tim es in th e A n n u a l R epo rts. T he CEA ’s A n n u a l R e p o r t for 1964, A ppendix A, outlined the m ultiplier process by w hich an $11 billion tax cut w ould produ ce a $30 billion increase in total dem and. A ccording to this outline, a tax cut o f this size w ou ld first increase consum ption by $9 billion, w hich in turn “w ou ld g en erate still fu rther increases in incom es and spending in an endless, but rapidly diminishing, chain.” This w ould result in adding $18 billion to GNP from in creased consum ption alon e “ — not just on ce, but year-in and year-out . . But the m ultiplier process is not com p lete at this point. In response to rising incom e, expenditures fo r plant and equipm en t, inventories, residential construction, and state and local governm ent program s w ould expand, and by the multiplier process w ou ld a d d another “$10 to $14 billion to GNP.” In this analysis, how ever, no m ention w as m ad e regarding the G overnm ent financ ing requirem ents o f the prop osed tax reduction. In discussing th e prop osed incom e surtax as a m eans o f reducing inflation, th e A n n u a l R e p o r t for 1968 argued that the sam e m ultiplier process w as relevant. T he CEA asserted, . . th e e c o nom ic effects o f a tax increase are the mirror-image o f the expansionary effects accom p lish ed by tax reduction.” T he view that changes in G overnm ent expenditures an d tax rates exercise a pow erful influence on total spending, w ithout regard to changes in th e volum e o f G overnm ent d eb t outstanding or in the rate o f monetary expansion, has also b een prevalent am ong many others. A bout four hundred professional econom ists signed a statem ent supporting th e R evenue Act o f 1964 as a m eans of stimulating total dem and. Politicians, public figures, an d econom ic com m entators have com e g en erally to support fiscal actions explicitly, but m ore frequ en tly implicitly, on th e basis o f th e sim ple m ultiplier analysis. This view has b een challen ged by a num ber o f econom ists on th e grounds that it d oes not give ad equ ate recognition to the financing o f G overnm ent expenditures. T hey argue that G overn m ent spending financed by taxes or borrow ing from saving o f th e gen eral pu blic m ay red u ce other spending to such an extent that there will b e little, if any, net increase in total spending. This is frequ ently referred to as the “crow ding out” o f private expenditures by fiscal actions. A ccording to these econom ists, stabilization recom m endations b ased on th e prevailing m ultiplier analysis o f th e 1960’s are erroneous. T he folloioing article by theory from A dam Smith to ity. T hey find that “crow ding M oreover, the em phasis on lization program s during the post-Keynesian econom ics. R oger W . Spencer and W illiam P. Y ohe provides a survey o f econom ic the present regarding th e influence o f fiscal actions on econom ic activ out” has been the dom inant view during th e past two hundred years. the sim ple Keynesian m ultiplier analysis in d evelopin g econom ic stabi 1960’s is not in gen eral agreem en t w ith K eynes’ own view s or w ith Roger W . S pencer is an econom ist o f th e F ed eral Reserve B ank o f St. Louis. W illiam P. Y ohe is Professor o f E conom ics at D uke University and was a visiting scholar at th e F ed eral R eserve Bank o f St. Louis for a period o f tim e in 1969 and 1970. Page 12 FEDERAL RESERVE BANK OF ST. LOUIS Fiscal policy provides additional spending in a w orld o f sparse spending opportunities. But it does not provide a new source of finance in a w orld w here spending is con strained by sources o f finance. T he govern m ent expenditures are financed in d eb t m arkets in com petition with private expen di tures. T he case least favorable to fiscal policy is that in w hich the additional governm ent borrow ing sim ply CROW DS OUT o f the m arket an equ al (or conceivably even greater) volum e o f borrow ing that w ould h ave financed private expenditures.1 (Italics and capitaliza tion added) Changes in Government expenditures and taxes, the policy arm of Keynesian economics, have come under attack recently for their apparent failure to secure desired stabilization goals. Several studies, in this Review and elsewhere, have concluded that in creases in Government expenditures, which are not accompanied by money creation, induce temporary increases in nominal GNP with no net effect over a longer period of time.2 Monetary actions, in contrast, have been found to exert an important influence on economic activity quite apart from fiscal develop ments. Much of the rebuttal of these articles has been focused on debating the strengths of the mone tary variables rather than the weaknesses of the fiscal variables.3 “The authors wish to acknowledge the helpful comments on earlier drafts of: Leonall C. Andersen, Karl Brunner, Keith M. Carlson, E. G. Corrigan, John M. Culbertson, R. Alton Gilbert, Thomas M. Havrilesky, Jerry L. Jordan, Michael W. Keran, Laurence H. Meyer, and. Clark A. Warburton. The authors are solely responsible for the analysis and conclu sions presented in the article. ijohn M. Culbertson, Macroeconomic Theory and Stabiliza tion Policy (New York: McGraw-Hill Book Company, 1968), p. 463. 2Leonall C. Andersen and Jerry L. Jordan, “Monetary and Fiscal Actions: A Test of Their Relative Importance in Economic Stabilization,” this Review (November 1968); Leonall C. Andersen and Keith M. Carlson “A Monetarist Model for Economic Stabilization,” this Review (April 1970); Michael W. Keran, “Monetary and Fiscal Influences on Eco nomic Activity —The Historical Evidence,” this Review ( No vember 1969), and “Monetary and Fiscal Influences on Economic Activity — The Foreign Experience,” this Review (February 1970); John Deaver, “Monetary Model Building,” Business Economics (September 1969). Also, in this vein, see Milton Friedman and David Meiselman, “The Relative Sta bility of Monetary Velocity and the Investment Multiplier in the United States, 1897-1958,” in Stabilization Policies, Com mission on Money and Credit (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1963). 3The monetary variables most often employed in these arti cles are changes in money and the monetary base. The primary fiscal spending variable is changes in high-employment Government expenditures. For a defense of the fiscal position, see Murray Weidenbaum, “Is Fiscal Policy Dead?,” Financial Analysts Journal ( March-April, 1970) and E. G. Corrigan, “The Measurement and Importance of Fiscal Policy Changes,” Federal Reserve Bank of New York Monthly Re view (June 1970). OCTOBER, 1970 This article surveys a large body of economic litera ture, in search of some reasonable rationale which could explain the poor showing of the Government spending multipliers in the previously mentioned reduced-form statistical studies. Two possible reasons why these multipliers fail to achieve the high magni tudes promised by a mathematical derivation of the basic “Keynesian” multiplier are that many Keynesian models do not capture (1) the displacement of private spending by Government spending (the “crowding-out” effect), or (2) leakages associated with consumer or Governmental propensities to spend. The leakages are usually accounted for in the more sophisticated econometric models, resulting in lower multipliers than those derived from the elementary models, but failure to allow for the possibility of crowding out makes even these multipliers higher than warranted. Worldwide acceptance of the Keynesian theory, and increasing Government intervention in economic affairs, represent the trend of post-depression thinking on stabilization policy.4 There seems to be a growing belief, however, that fiscal policies developed and ap plied in a society operating at a low level of resource utilization are not necessarily applicable under highemployment, inflationary conditions. Further, the evi dence presented by one analyst casts doubt on the efficacy of fiscal policies even during the deflationary 1930’s.5 Pure fiscal actions, which entail changes only in Government expenditure and tax programs, are rare; most stabilization actions involve a mixture of (1) 4See, for example, Robert Lekachman, The Age of Keynes (New York: Random House, 1966). The difficulty of gen eralizing “Keynes’ view” or “the Keynesians’ view” has been pointed out by Axel Leijonhufvud ( “Keynes and the Key nesians: A Suggested Interpretation,” American Economic Review, May 1967, Papers and Proceedings, p. 401). One must be careful in applying the epithet ‘Keynesian’ nowadays. I propose to use it in the broadest possible sense and let ‘Keynesian economics’ be synonymous with the ‘majority school’ macroeconomics which has evolved out of the debates triggered by Keynes’s General Theory . . . . Within the majority school, at least two major factions live in recently peaceful but nonetheless uneasy coexistence. With more brevity than accuracy, they may be labeled the ‘Revolutionary Orthodoxy’ and the ‘Neoclassical Resurgence’. . . . . The orthodoxy tends to slight monetary in favor of fiscal stabilization policies. The neoclassical faction may be sufficiently characterized by negating these statements. As described, the orthodoxy is hardly a very reputable position at the present time. Its influence in the currently most fashionable fields has been steadily diminishing, but it seems to have found a refuge in business cycle theory — and, of course, in the teaching of undergraduate macro economics. 5Keran, “Monetary and Fiscal Influences on Economic Activ ity — The Historical Evidence,” p. 23. Page 13 FEDERAL RESERVE BANK OF ST. LOUIS fiscal (tax-financed Government spending), (2 ) mone tary (Federal Reserve actions) and (3) debt man agement (changes in the maturity or value of the debt) policies. For example, the offering of Treasury bonds to finance a deficit constitutes debt manage ment policy (or fiscal policy, by more conventional definitions) at first, but the subsequent purchase of such bonds (or perhaps the failure to purchase such bonds) by the Federal Reserve is generally construed to be monetary policy. Thus die existence of Govern ment bonds clouds the distinction between the two basic economic stabilization policies — monetary and fiscal. The method of financing Government expenditures influences economic stabilization efforts. There are two conflicting views regarding the extent of this in fluence. Funds for Government spending are obtained by taxing the public, borrowing from the public, and/or money creation. Monetarists (those who favor monetary over fiscal stabilization policies) generally assert that Government spending financed by either taxing or borrowing from the public is mainly a re source transfer from the private sector to the Govern ment, with little net effect on total spending. Monetary expansion, even if unaccompanied by an increase in Government spending, has a strong, stimulative in fluence on the economy. Contemporary Keynesian theory holds that all of the three techniques of Government finance involve something more than a simple resource transfer from the private to the Government sector. An expansion ary effect on the economy may be achieved by a rise in Government spending matched by an increase in tax receipts, or by a rise in Government spending financed by bond issuance either to the public or the monetary authorities. The monetarists’ view that Government spending financed by taxes or borrowing from the public merely displaces, or “crowds out,” private spending is not a new one. It was, in fact, the dominant view before the Keynesian revolution of the 1930’s. Classical econ omists including Adam Smith and David Ricardo, and neo-classicists including F. A. Hayek and R. G. Hawtrey, found little use for fiscal stabilization efforts. Keynes, at first a fairly orthodox neo-classical econo mist, altered his view's on many issues prior to the publication of The General Theory o f Employment, Interest and Money, including a downgrading of the fiscal crowding-out concept.6 Since it was Keynes’ 6John Maynard Keynes, The General Theory of Employment, Interest and Money (New York: Harcourt, Brace and Com pany, 1936). His book will be referred to as The General Theory throughout the remainder of this article. Page 14 OCTOBER, 1970 theory of the proper use of fiscal policy which even tually became the dominant view, we will develop at some length Keynes’ thinking on the subject. Because Keynes was strongly aware of the traditional neo classical views on fiscal theory, he hedged his argu ments more carefully than many of his successors. Keynes’ General Theory analysis, in contrast to the focus of the classical school on long-run supply factors, was oriented toward short-run demand considerations. Both views were couched predominantly in real (ra ther than nominal) terms. Consequently, crowdingout analysis derived from either of these two ap proaches deals primarily with the displacement of real private spending by Government spending.7 A summary of classical, neo-classical and Keynes’ early views on fiscal crowding out will be presented first. Next, we will trace Keynes’ later fiscal views, underscoring the assumptions on which his final posi tion was built. The subsequent section examines bondand tax-financed crowding out, based principally on a Keynesian analytical framework outlined by Rich ard Musgrave. Finally, we discuss alternative frame works and some empirical evidence bearing on the crowding-out issue. Summary of Classical, Neo-Classical and Keynes’ Early Views on Fiscal Crowding Out The mainstream of economic thought prior to the publication of Keynes’ General Theory in 1936 did not favor Government spending for stabilization pur poses.8 There was some opposition to an enlarged "The crowding-out phenomenon may be described in conven tional mathematical symbols in the following manner: a) Real crowding out dY° ~ 0 dG* dM” = 0 b) Nominal crowding out dM = 0 These relations imply that: c) d(Y»-G») _ dG* dM = 0 or d) d(Y-G) dG I I , _ 1 dM = 0 Y = total spending, G = Government spending, and M = money supply. Relation (b ) most closely represents the empirical crowdingout results reported in studies cited in footnote 2. 8A more detailed discussion of classical, neo-classical and Keynes’ early views on crowding out will be presented in a forthcoming Working Paper of the Federal Reserve Bank of St. Louis. FEDERAL RESERVE BANK OF ST. LOUIS role for Government spending purely from a philoso phical view, but much of the criticism of increased Government intervention was based on crowding-out theory. Adam Smith, for example, opposed extensive Gov ernment involvement for both philosophical and crowding-out reasons. For the most part, Smith (writ ing in 1776) considered Government labor “unproduc tive,” and condemned the transfer of resources from the private sector, whether through taxation or bor rowing. Borrowing funds from the public to finance Government spending was asserted to involve the “destruction of some capital which had before existed in the country; by the perversion of some portion of the annual produce which had before been destined for the maintenance of productive labour, towards that of unproductive labour.”0 Smith believed that “saving is spending,” because one man’s saving becomes another man’s investment. Later classical economists, such as John Stuart Mill and J. B. Say, writing primarily in the first half of the nineteenth century, saw in Adam Smith’s maxim a guarantee of full employment. That is, Government spending was considered unnecessary as a stabilization tool, because private investment was sufficient to utilize the funds provided by private saving. The most elementary case for crowding out may be examined in a “Say’s Law” framework. Say’s Law is widely known as “supply creates its own demand.” More specifically, if output (supply of goods and services) is determined by the behavior of profit maximizing producers, competitive labor markets, the existing stock of capital goods, and the state of tech nology, then relative prices will tend automatically to adjust so as to eliminate a deficiency or excess of demand. In an economy in which Say’s Law is op erative, attempts by the Government to increase total spending, by raising Government expenditures and financing the increasing budget by either borrowing from the public or taxation, merely induce changes in relative prices so as to reallocate the same level of real output. The two cases —bond- and tax-financed Govern ment expenditures — involve changes in the structure of prices (principally interest rates) to restore equilibrium at full employment, so they represent the tendency of a market economy ultimately to neu tralize disturbances ( in this instance budgetary °Adam Smith, The Wealth of Nations (New York: Random House, Inc., 1937), p. 878. OCTOBER. 1970 shocks). Without the Government’s presence, private propensities to spend the full-employment level of real income for either consumption or investment sum to one as a consequence of automatic price adjust ments; adding a Government propensity to spend correspondingly reduces private propensities, in order to maintain total propensities to spend equaling one.10 Neo-classical business cycle theories became an im portant part of economic stabilization literature in the early twentieth century. One of these theories, over-investment financed by “forced” saving, was equally applicable for any sector favored over others for loan extension by the banking system. If, for ex ample, the Government were to borrow from banks to finance its investment spending, the increased pur chasing power of the Government would allow it to bid resources away from other sectors and, under fullemployment conditions, drive up the price level. The higher price level would serve as a deterrent to “real” consumer or private investment spending which would otherwise have taken place. A good example of this view is found in the testi mony of the English economist, R. G. Hawtrey, be fore the Macmillan Committee in 1930. Hawtrey denied the usefulness of Government spending, re gardless of financing even under depression conditions: On the matter of government spending [to bring England out of her stagnation], Hawtrey stated that whether the spending came out of taxes or loans from savings, the in creased governm ental ex p en di tures w ould m erely rep lace private expenditures. He even considered the “radical” idea of government spending out of new bank credit, but predicted that the result of such a policy would be inflationary and a threat to the gold standard, thus forcing up the bank rate of interest and causing credit contraction. Such a plan, for him, would only defeat itself, since government expenditures out of bank credit would mean the end of cheap money for free enterprise.11 (Italics added) Keynes advocated Government spending as a stim ulative economic measure twelve years before the publication of The G eneral Theory, but he attached the “rider” that such spending should be financed by monetary expansion.12 He emphasized (in 1929) that 10See, for example, Culbertson, p. 333. n Lawrence R. Klein, The Keynesian Revolution (London: The Macmillan Company, 1968), pp. 45-46. Hawtrey’s business cycle theory was based on easy money creating additional working capital and inducing inventory investment. 12Roy Forbes Harrod, The Life of John Maynard Keynes (New York: Harcourt, Brace and Company, 1951), p. 441. Keynes was only one of a number of economists of the FEDERAL RESERVE BANK OF ST. LOUIS the central bank had the power to defeat expan sionary fiscal actions and thus . . ensure that the expenditure financed by the Treasury was at the ex pense of other business enterprise.”13 The British Treasury, however, like Hawtrey, took the position that Government spending, regardless of financing, simply displaced private spending. Keynes continued to press his fiscal policy argu ments in the years preceding The General Theory, but eventually downgraded the necessity of mone tary expansion accompanying increased Government spending in order for such spending to have an ex pansionary effect. Development of the liquidity pref erence theory of interest and the fiscal multiplier theory, particularly the latter, were the keys to Keynes’ ability to shift money into the background of his analysis. The liquidity preference concept led to the idea that more efficient utilization of the existing money stock (increased velocity) was encouraged by increased Government expenditures, while the mathe matical formulation of the multiplier (by R. F. Kahn, a student of Keynes) indicated that the increased taxes and saving generated by the rise in Govern ment-induced income would be just enough to cover the financing of the deficit. Keynes’ General Theory View of Fiscal Crowding Out A detailed explanation of the workings of the mul tiplier was one of the chief contributions of The General Theory, as was the discussion of the qualifica tions of multiplier analysis. The investment multiplier (k ), wrote Keynes, “tells us that, when there is an increment of aggregate investment, income will in crease by an amount which is k times the incre ment . . .”14 The principle of the multiplier provides OCTOBER, 1970 “an explanation of how fluctuations in the amount of investment, which are a comparatively small propor tion of the national income, are capable of generating fluctuations in aggregate employment and income so much greater in amplitude than themselves.”15 The multiplier concept of an increment of new in vestment spending being transmitted from pocket to pocket and thereby increasing total spending by some multiple is easier to grasp intuitively than mathemati cally. Keynes, himself searching for the proper mathe matical rationale for his deficit spending ideas, was quick to seize and popularize Kahn’s version. He and Kahn realized some of the limitations of their formal multiplier theory, but the scores of articles published during the past three decades, which clarify or repu diate “the Keynesian multiplier,” suggest the inade quacy of its treatment in The G eneral Theory. The significance of the multiplier may be examined from the crowding-out point of view, to the extent that the multiplier represents causal relations rather than the ex post tautology that the change in income must, by definition, equal some multiplier times the increment of investment.16 Keynes, in The General Theory, provided one of the most cogent and clear crowdingout arguments to be found. Although he wished to move money to a supporting rather than leading role in his General Theory analysis, he recognized that monetary influences could overcome his newly de veloped multiplier and liquidity preference constructs. Contrary to expressions found in the present “con ventional wisdom,” Keynes hedged his arguments considerably. He pointed out that, If, for example, a Government employs 100,000 additional men on public works, and if the multi plier . . . is 4, it is not safe to assume that aggregate employment will increase by 4 0 0 ,0 0 0 . F o r the new policy may have adverse reactions on investm ent in other directions. . . . The method of financing the policy and the increased working cash, required by the increased employment and the associated rise of prices, may have the effect of increasing the rate of interest and so retarding investm ent in oth er d irec tions, unless the monetary authority takes steps to the contrary; whilst, at the same time, the increased cost of capital goods will reduce their marginal efficiency Twenties and Thirties who advocated governmental deficit spending to spur economic activity. Twelve University of Chicago economists, for example, recommended deficit spending in a 1932 memorandum to a member of Congress: “Convinced that pump-priming deficits would induce at best only temporary revival, the Chicago economists recom mended that deficits should continue until recovery was in disputably established. These would be financed by selling new issues of bonds to the reserve banks or by exchanging them for bank deposits and Federal Reserve notes,” See J. Ronnie Davis, “Henry Simons, the Radical: Some Docu mentary Evidence,” History of Political Economy, Fall, 1969, pp. 389-90. 13See John Maynard Keynes, Essays in Persuasion (New York: Harcourt, Brace and Company, 1932), p. 126. i5Jbid., p. 122. 14Keynes, The General Theory, p. 115. The multiplier con cept may be formalized as AY = kAI where AY is the change in income, AI is the change in investment and k is the multiplier. Investment refers to outlays on newly cre ated plant and equipment and housing, and additions to inventories. A similar multiplier is developed for changes in Government expenditures (A G ). 16R. F. Kahn ( Economic Journal, 1931, p. 188) was quite aware of the causal vs. tautological nature of the equality between the increase in the rate of investment expenditure and the subsequent streams of generated saving. The equal ity, he wrote, “far from being the logical consequence of summing an infinite geometrical progression, is in reality self-evident in nature . . .” Page 16 FEDERAL RESERVE BANK OF ST. LOUIS OCTOBER. 1970 F ig u r e 1 C rowding Out in an IS-LM Framework (A) (B) IN C O M E IN C O M E to the private investor, and this will require an actual fall in the rate of interest to offset it.17 The multiplier limitation outiined above by Keynes can be summarized within the conventional IS-LM curve framework. In the IS-LM system, a set of equations is postulated regarding propensities to spend, to save, and to hold money balances.18 The system may be reduced to two equations. One con tains various combinations of income (Y ) and interest rates (r) at which investment equals saving; this is referred to as the IS curve and represents the real sector of the economy. The other equation contains various combinations of Y and r at which money de manded equals the money stock in existence; this is labeled the LM curve and represents the financial sector. The two equations may then be solved simul taneously to determine equilibrium values of the in terest rate and income. If the LM curve is steeply sloped as shown in part (A) of the accompanying figure, the increase in Gov ernment spending reflected in the rightward shift of the IS curve affects a sharp rise in interest rates (from 17Keynes, The General Theory, pp. 119-20. (Italics added except for “fall” ) 18See, for example, Gardner Ackley, Macroeconomic Theory (New York: The Macmillan Company, 1961), pp. 359-98. r! to r2) and little or no change in income (Yx to Y2). A sharply rising LM curve implies that the retarding effects of such actions on private investment are sub stantial. Only by shifting the LM curve to the right, through monetary expansion for example, will a sig nificant rise in income occur (from Yx to Y3). Keynes, by mentioning this multiplier limitation, supplied a strong theoretical basis for the crowding-out thesis, and thereby shifted the discussion to a quantitative plane involving determination of the slopes of the IS-LM curves and the substitutability of Government and private spending. Keynes recognized a second way, based on business psychology, in which Government spending could crowd out private spending. “With the confused psychology which often prevails, the Government pro gramme may, through its effect on ‘confidence’, in crease liquidity-preference or diminish the marginal efficiency of capital, which, again, may retard other investment unless measures are taken to offset it.”19 I!lKeynes, The General Theory, p. 120. Daniel Throop Smith, “Is Deficit Spending Practical?”, Harvard Business Review (Vol. 18, 1939), pp. 38-9, was particularly concerned about the business confidence aspect of the crowding-out thesis. A continued experience with deficits which do not produce sustained recovery, as in this country, or a recent infla tion and collapse, as in continental European countries, is likely to make a deficit matter for concern and anxiety, And, if there is disbelief in the benefits of a deficit, then Page 17 FEDERAL RESERVE BANK OF ST. LOUIS If the increase in Government spending (shown by the shift of IS to IS' in part B of Figure 1) has an adverse effect on liquidity preference, the LM curve shifts to the left (LM ') and income increases only slightly from Y, to Y2. If the marginal efficiency of capital is adversely influenced, this may be shown by a leftward shift of the IS curve from IS' to IS", also resulting in a small increase in income from Yt to Y2. The magnitude of these shifts of the curves depends, of course, on the assumed response of “business con fidence” to increased Government spending. Keynes also noted a potential multiplier leakage in the hypothesized tendency of the marginal propensity to consume to diminish as employment increases. This introduces the state of economic activity as an influ ence on the efficacy of the multiplier. Thus Keynes could, without reservation, recommend any sort of public works (pyramid building, wars) during periods of unemployment, “if only from the diminished cost of relief expenditure, provided that we can assume that a smaller proportion of income is saved when unemployment is greater; but they may become a more doubtful proposition as a state of full employ ment is approached. Furthermore, if our assumption is correct that the marginal propensity to consume falls off steadily as we approach full employment, it follows that it will become more and more trouble- the new money spent by the government may well be more than offset by additional withdrawals of private money which would otherwise be spent. Likewise, if con sumer incomes do increase immediately as a result of the deficit, business may anticipate that the increase is tem porary and refrain from long-term commitments. . . . Consumer income arises primarily out of the pro cesses of the production and distribution of goods and services. Therefore, if private spending decreases by as many dollars as government spending increases (and it takes only a very small proportionate decrease in private spending to offset a large proportionate increase in gov ernment spending), then on balance there is no net in crease in consumer income. Furthermore, even if there is an immediate net increase in consumer income, there may be enough flexibility in our economic system in the form of idle capacity, especially during a depression, to obviate any necessity of spending on capital goods to provide consumer goods. Also, if there is an actual increase in consumer income and spending but it is believed that the demand is an artificial one dependent upon a con tinuation of government deficits which cannot last in definitely, business will be loath to make long-term capital commitments. In such circumstances business at most will make up deferred maintenance and build up inventories creating an artificial “boomlet” which is not likely to develop into a sustained recovery. This seems to have been particularly true in 1936 and 1937. (Italics added) This analysis implies that, dependent upon the time pe riod specified, Government spending may not only crowd out an equal magnitude of private spending but an even greater amount. The multiplier would then be negative rather than zero or positive. Page 18 OCTOBER, 1970 some to secure a further given increase of employ ment by further increasing investment.”20 A related leakage was that of an increase in em ployment which tends, “owing to the effect of dimin ishing returns in the short period, to increase the proportion of aggregate income which accrues to the enterpreneurs, whose individual marginal propensity to consume is probably less than the average for the community as a whole.”21 Moreover, the impact of the multiplier might be reduced by those unemployed consumers formerly existing on negative savings, pri vate or public, whose employment would diminish negative savings, and thereby, the marginal propen sity to consume. An additional caveat to those who would mechan istically apply the multiplier was that, “In any case, the multiplier is likely to be greater for a small net increment of investment than for a large increment; so that, where substantial changes are in view, we must be guided by the average value of the multi plier based on the average marginal propensity to consume over the range in question.”22 These multiplier modifications and many more can be found in the post-General Theory literature,23 but enough has been introduced here to indicate that the value of the Government spending multiplier may be positive, zero or negative. It should be pointed out that although Keynes was aware of some of the lim itations of the multiplier (as demonstrated above), he did not emphasize them, partly because of his desire to put a simple rationale for Government spend 20Keynes, The General Theory, p. 127. One implication of this analysis is that increased Government spending, for the rea sons given, may not only be less effective at stimulating the economy near full employment than at substantially less than full employment, but also relatively ineffective at slowing it. zlIbid., p . 121. 22 Ibid. 23See Hugo Hegeland The Multiplier Theory (New York: Augustus M. Kelley Publishers, 1966) p. 73: “The basic weakness of the multiplier theory lies in its assumptions, which in fact eliminate the real problems involved and make the theory almost a truism.” One example of post-Keynesian criticism of the mechani cal application of the multiplier can be found in Gardner Ackley, “The Multiplier Time Period: Money, Inventories, and Flexibility,” American Economic Review (June, 1951) pp. 350-368. Ackley notes that the existence of inventories serves as a buffer providing a potentially long and variable lag between a change in consumer expenditures and a change in production rates. An argument made by Milton Friedman is that Government expenditure increases may be viewed by consumers as adding only to transitory income (i.e., temporary income.) He argues that the marginal propensity to consume out of transitory income is zero. See Milton Friedman, A Theory of the Consumption Function (Princeton: Princeton University Press, 1957), p. 26. OCTOBER, 1970 FEDERAL. RESERVE BANK OF ST. LOUIS ing during the depression in the hands of the policy makers as expeditiously as possible.24 A Basic Keynesian Framework for Crowding-Out Analysis A number of attempts were made after Keynes to spell out more formally the relationships governing the conditions under which Government spending would add to total spending or crowd out a significant volume of private spending. Richard Musgrave de veloped one of the most efficient frameworks for analyzing the crowding-out phenomenon of those con structed in the approximately three decades follow ing The General Theory.25 There are three basic sectors of the economy and three basic “effects” in the Musgrave framework.26 24The crowding-out issue emerged in a different context in at least one of Keynes’ post-General Theory essays. In Paul Davidson, “Keynes’s Finance Motive,” Oxford Economic Papers (March 1965), pp. 48-49, Keynes is quoted from an April 18, 1939 letter to the London Times, in which he explains his approach to financing Government expenditures for rearmament: “If an attempt is made to borrow them [the savings which will result from the increased produc tion of non-consumption (war) goods] before they exist, as the Treasury have done once or twice lately, a stringency in the money market must result, since, pending the ex penditure, the liquid resources acquired by the Treasury must be at the expense of the normal liquid resources of the banks and of the public.” 25See Richard A. Musgrave, The Theory of Public Finance (New York: McGraw-Hill, Inc., 1959), pp. 526-55. Others outlining frameworks under which crowding out might be analyzed include Alvin H. Hansen, Monetary Theory and Fiscal Policy (New York: McGraw-Hill Inc., 1949), pp. 167-73, and Abba P. Lerner, “The Burden of the National Debt,” in Income, Employment and Public Policy, Essays in Honor of Alvin H. Hansen (New York: W. W. Norton and Company, Inc., 1948), pp. 255-75. Lemer, for example (p. 26 9 ), states that “The sale of gov ernment bonds diminishes liquidity, tends to raise interest rates, and discourages investment. The government should therefore not borrow unless it wishes to bring about these deflationary effects.” 20The following is the linear form of Musgrave’s basic equations: (1 ) Y = C + I + G (2 ) C = a + c ( Y - T ) + wM (3 ) I = d — ei (4 ) i = U/Ma • ( 1 - P ) / P ( 5) Ma = M - M t ( 6) Mt = 1/ V • Y ( 7) G = T Y = income, C = consumption, I = investment, G = Government spending, T = tax yield, M = total money supply, Ma = asset money, Mt = transactions money, V = income velocity, i = “the” market interest rate, c = marginal propensity to consume, w = a wealth para meter, e = an investment parameter, U = bond coupon bill, P = fraction of claims ( asset money and Government debt) people wish to hold in form of Government debt, a = intercept term, d = intercept term. The Government budget constraint, G = T, could be modified to permit deficit finance through changes in money or Government bonds. A rough approximation of the Gov ernment budget constraint for the above system of equa tions might take the form G = T + AM + AU. Government, consumer, and investment spending com prise the three sectors, and income, wealth, and sub stitution are the three effects. Income effects are those related to changes in disposable income. Wealth ef fects occur with changes in the level of wealth, the ratio of wealth to income or changes in the level or structure of debt claims (money and/or Government bonds). Substitution effects result from saving or spending incentives provided by a particular Govern mental tax or expenditure policy. Income and wealth effects are interdependent; the initial effect may be one or the other, but both effects become operative as the adjustment proceeds. For example, an increase in investment may occur as a result of a change in income because of a change in the propensity to in vest. The income effect would be followed by a wealth effect as changes in the interest rate and the ratio of wealth to income occur over the period of adjustment. Musgrave assumes that consumption is altered through income effects and/or wealth effects. Invest ment is influenced by changes in income and/or (via interest rates) changes in the level or structure of claims, while the substitution effect, which does not appear explicitly in the model, may be operative through changes in profitability due to taxation. Wealth and income effects are seen most clearly by comparing the private sector’s response to a variety of policy actions. The accompanying table illustrates a basic cataloguing of budget policies within the Mus grave framework. It gives various combinations of stabilization actions by arraying changes in the supply of Government debt against changes in the supply of money. T a b le 1 Cataloguing Fiscal Finance C h a n g e in S u p p ly of M o n e y C h a n g e in * S u p p ly o f D ebt + 0 - + 4 Deficit finan ce d b y mix of d eb t a n d m on ey 3 Deficit financed b y new m on ey 8 D ebt retirem ent finan ce d b y new m on e y 0 - 2 9 Deficit financed M o n e y retirem ent b y debt financed b y b o rro w in g 1 6 Eq u a l c h a n g e in M o n e y retirem ent tax y ie ld a n d finan ce d b y e x p e n d itu re s current su rp lu s 5 D eb t retirem ent finan ce d b y current surp lu s 7 C urrent su rp lu s used to retire m ix of m on e y a n d debt ♦Changes in the supply of debt refer to changes in coupon bill or m aturity value rather than to changes in the m arket value of the total debt outstanding. Page 19 FEDERAL. RESERVE BANK OF ST. LOUIS We assume that the ratio of consumption to income will increase (that is, the ratio of saving to income will fall) with increases in the supply of money, or (with some qualification) the supply of public debt and the level of investment will rise with an increase in the supply of money and will fall with an increase in the supply o f debt. It is then possible to determine whether each policy mix is expansive or restrictive. Policies 3, 4, and 8 are expansionary, policies 6, 7, and 9 are restrictive, and policies 1, 2, and 5 contain ambiguous effects. Policy 2, a deficit financed by debt issue, most closely approximates the bond-financed’ crowding-out case. Policy 1 approximates the tax-financed crowding-out case. Note that neither policy 1 or 2 involves changes in the supply of money. These two cases will be examined separately, to ascertain which factors are involved in determining whether purely budgetary actions result in an expansionary effect on the economy, or simply crowd out private spending. Bond-Financed Changes in Government Expenditures What actually happens in the case of bond-financed Government expenditures depends to a large extent on the movement of the interest rate. There will be an expansionary income effect (Government spend ing adds directly to income) which may be aug mented by a stimulative wealth effect on consump tion (more bonds held by the public), if the interest rate remains relatively unresponsive.27 If the increase in the supply of Government bonds is accompanied by a sharp rise in the interest rate, the expansionary income effect could be countered by: (1) a smaller net wealth effect (since the market value of the Gov ernment debt varies inversely with the interest rate); (2) displacement of private investment, which re sponds negatively to a rise in the bond interest rate; and (3) a diversion of funds from consumption into 27The wealth effect, as described by Musgrave, does not consider the issue of future tax liabilities associated with Government debt. Even if the interest rate responds negli gibly to an increase in Government bonds, the rise in the value of the debt (positive wealth effect) may be partially offset by the discounted future tax liability required to pay the interest on the debt. The impact of the future tax liability on aggregate demand in the current period depends upon (a) the degree to which the bond-holding public considers future tax liabilities, (b ) the maturity of the debt, and (c ) the anticipations of the public as to the nature of the future taxes to be levied (on the income from labor or capital) to pay the interest on the debt. See Boris P. Pesek and Thomas R. Saving, Money, Wealth and Economic Theory (New York: The Macmillan Company, 1967), Chapter 10, who argue that the tax liability offset to the wealth effect is only partial and not total as some analysts have suggested. Page 20 OCTOBER, 1970 Government bond purchase (voluntary saving), hav ing an initially restrictive effect. The response of total spending to bond-financed Government expenditures is ambiguous. If savers are indifferent between the holding of bonds and money, the interest rate does not rise, the net market value of the debt increases by this means of financing, and a strong wealth effect may be realized. At the other extreme, savers insisting on holding a fixed ratio of money and debt would result in a rise in the interest rate, no change in the value of the debt (gains from the acquisition of new bonds being offset by capital losses suffered from the holding of old bonds), and no wealth effect. Actual conditions, according to Musgrave, fall some where in the middle. During a depression, the interest rate is hkely to be relatively unresponsive. The im plication is that the interest rate is likely to be most responsive during periods of rapid expansion. A more detailed model might consider the effects of price expectations and varying default risks on interest rates (and, consequently, private spending), and present a wider spectrum of private and public assets. The introduction of fractional reserve banking (which is excluded from the above discussion) into the system complicates the argument somewhat, but the essential principles remain unchanged. In prac tice, as Buchanan points out, commercial bank pur chases of Government bonds out of excess reserves permit the full expansionary effect to take place ( money creation), but bonds purchased when reserves are not in excess have the same effect as borrowing from the public. In the latter case, “The expansionary effects of the public spending side of the deficit are, to a considerable extent, offset by the reduction in private investment caused by the tightening up of funds available for private securities.”28 Tax-Financed Changes in Government Expenditures The most obvious crowding out of private expendi tures by Government spending, involving the involun tary transfer of funds from the private to the Govern ment sector, is the case in which expenditures are financed by taxation. A balanced-budget multiplier equal to one (total spending increases by an amount equal to an increase in Government spending financed by taxes) is the usual upper limit assigned to taxfinanced Government expenditures. This result is 28James M. Buchanan, The Public Finances 111.: Richard D. Irwin, 1965), p. 99. (Homewood, FEDERAL RESERVE BANK OF ST. LOUIS generally deduced from the assumption that the Government’s propensity to spend for goods and services out of tax revenue is 100 per cent, while private propensities to spend after-tax income are less than 100 per cent.29 Most simple models, including Musgrave’s, do not allow adequately for potential leakages from the Gov ernment balanced-budget multiplier, but by injecting debt considerations into the discussion of an increase in Government expenditures met by a rise in taxes, Musgrave uncovers potentially adverse influences on private spending. If an initial rise in income is gen erated by a balanced-budget change, a drain on asset money would develop (because of the increased transactions demand for money) which tends to have a detrimental effect on investment. The rise in income relative to a constant money supply may have a de pressing effect on consumption, especially so the more the drop in net wealth (that is, a fall in the value of the debt) produced by a rising interest rate. Most observers, however, ignore debt considera tions in their analysis of the limitations of the unitary balanced-budget multiplier. The large number of as sumptions necessary to the determination of a uni tary balanced-budget multiplier, disregarding debt problems, reflects to some extent its implausibility.30 Baumol and Peston discuss the issue in terms of leak 29One observer has relied strongly on a positive balancedbudget multiplier to support his argument for the expan siveness of bond-financed Government expenditures. See Robert Eisner, “Fiscal and Monetary Policy Reconsidered,” American Economic Review (December 1969), pp. 897905. Eisner finds Government expenditures, regardless of the financing, to be the controlling exogenous variable in slowing or stimulating the economy ( “monetary measures are likely ultimately to be as limited in their impact in combating inflation as they have long been recognized to be limited in combating a deep depression. . . . “Public works or, more generally, government investment and con sumption, reemerge in their early role as prime weapons in the arsenal against depression and take on analogous im portance in any struggle against inflation,” p. 904). His model shows that whether Government spending financed by bond issue is expansionary (increase in prices) or offset depends on the interest and wealth elasticities of real money and commodity demand. He presents no empirical justification for his conclusion that bond-financed Govern ment expenditures are expansionary, but concludes that to argue against an expansion “would be to argue that a deficit-financed increase in Government expenditures can be deflationary while the same increase in Government expenditures would be inflationary if fully supported by taxes.” (p. 903) His balanced-budget multiplier analysis ignores its numerous limitations as discussed above. 30James Buchanan provides a list of seven assumptions under lying the unitary balanced-budget multiplier in The Public Finances, pp. 78-80. These are: 1) The entire amount of the Government spending change must take the form of purchases of real goods and services currently produced; 2 ) the balanced-budget change must be financed through taxes having about the same effects as the personal income OCTOBER, 1970 ages from the multiplier, and conclude that, in es sence, a rise in Government spending financed by taxation may crowd out an equivalent or greater magnitude of private spending.31 The basis for their reasoning is that, although the Government’s marginal propensity to spend exceeds that of the private sec tor, Government spending is subject to significant leakages before its influence on private spending is realized. The existence of these “leakages” in a governmental tax-expenditure program — the nonredistributional effects on private consumption, the purchase of items on capital account and of goods from abroad — is, of course, well known, though no attempt seems to have been made to take account of them in the literature. Possibly some of the writers had them in mind but considered them unimportant, since pre sumably none of them considered the unit multiplier figure as more than an approximation to the empiri cal magnitude in view of the recognized qualifica tions. The magnitude of the leakages may indeed be rather small, and it is very tempting to conclude that if 10 per cent of the government’s balanced budget expenditure is “leaked,” the multiplier will be reduced from unity to say about 0.9. However, we argue now that fairly small leakages can even pro duce a negative balanced budget multiplier. . . . Thus the multiplier will be positive, zero, or negative as the marginal propensity to save and im port of the private sector is greater than, equal to, or less than k, the government’s ‘marginal propensity to leak.’32 Alternative Frameworks for Analysis of the Crowding-Out Effect If, as has been maintained, the method by which Government spending is financed matters to both private and overall spending, it would seem that cur rent analysis should explicitly account for it in dis cussions of the Government expenditure multiplier. Quite often, however, current analysis ignores the potential crowding out of private spending when Gov ernment expenditures are financed by bond sales or taxation. tax; 3) the public spending must not exert substitution ef fects on the pattern of private spending; 4) the marginal propensity to save for taxpayers must be equal to the marginal propensity to save for the suppliers of Govern ment goods and services; 5 ) investment spending must not be altered significantly by the budgetary change; 6 ) the monetary-banking framework must permit attempted changes in spending to be carried out; 7) individual behavior in earning income must not be directly affected by the bud getary change. 31W. J. Baumol and Maurice H. Peston, “More on the Multi plier Effects of a Balanced Budget,” American Economic Review (March 1955). 32Ibid., pp. 144 and 145. Page 21 FEDERAL RESERVE BANK OF ST. LOUIS One possible reason for this oversight is that the numerous attempts to clarify the principles enunciated in Keynes’ General Theory have resulted in serious oversimplification of the complete Keynesian doctrine. This oversimplification has taken place at both basic and advanced levels of economic analysis. At the basic level, the “45-degree diagrams” popularized in principles text books leave no room for budget financ ing considerations.33 Similarly, the fundamental twoequation Keynesian model popularized in elementary text books, and its usual extensions, encourage the mechanical “cranking out” of fiscal multipliers which omit adequate treatment of the financing issue.34 The Hicksian IS-LM analysis, formulated to sum marize Keynesian macroeconomic theory, represents oversimplification at a fairly advanced level (see figure 1). An increase in Government spending is demonstrated by a rightward shift of the IS curve, which, as explained above, may result in a rise in both the interest rate and income. If the analysis stops here (as it frequently does), it is difficult to discern how the increased expenditure is financed. In reality, the shift from an equilibrium to a disequilibrium po sition may bring about simultaneous changes in in come, wealth, interest rate, money demand and money supply variables, and tax functions, depend ing on the source of finance. An approximation of the effects of movements in the key variables, as a new equilibrium position is established, may be achieved by altering the slopes and positions of the IS and LM curves, but the framework of analysis, without an explicit Government budget constraint, is not condu cive to such shifts.35 William L. Silber sets forth conditions under which Government spending can be contractionary in a properly specified IS-LM framework.30 Utilizing a framework of analysis similar to that of Musgrave, he 33See, for example, Paul Samuelson, Economics, 8th edition (New York: McGraw-Hill Book Company, Inc., 1970), p. 317. M( l ) C = a + bY ( 2) Y = C + I + G Substituting (1 ) into (2 ) and solving for Y gives: Y = a /1 —b + 1 /1 —b + G / l —b where 5Y/5G = 1/1 — b = Government spending multiplier 35“The Hicks-Hansen diagram has elegant simplicity which appeals to many. It has the disadvantage, however, that most of the ‘works’ are out of sight. This means that we need to use another diagram (or an extra mental calcula tion) to determine the effect of a displacement of the equilibrium on the other variables of our system.” See Ackley, p. 372. 38William L. Silber, “Fiscal Policy in IS-LM Analysis: A Cor rection”, forthcoming in the Journal of Money Credit and Banking. Page 22 OCTOBER, 1970 finds that a rise in the supply of Government bonds which increases wealth also increases the demand for money (represented in the IS-LM framework by a leftward shift of the LM curve). Silber concludes: It has been demonstrated that traditional IS-LM analysis has not treated the bond-finance and new money-finance cases of government deficits symmetri cally. W hen proper treatment is given to the former case, we found that in the simple world of IS-LM analysis government expenditures financed by selling bonds to the public can be contractionary. Even when GNP does go up due to AG, the increase that occurs is overstated in the traditional (but incomplete) IS-LM model of income determination. The failure to incorporate the monetary effects of d e b t finance into the LM function is the major source of confu sion. While other studies have treated this ques tion, it has never been formally incorporated into IS-LM models. This had led to incorrect conclusions regarding the multiplier affects of government spending. The many income-expenditure econometric models which have emerged in the 1960’s appear to be wellsuited to handle the problems of simultaneity and the Government budget constraint. Unfortunately, some explicit provision for the fact that Government expenditures must be financed by taxing, borrowing from the public, or money creation has often been omitted from monetary and fiscal policy simulations of these large models. A model which includes, for example, Government purchases of goods and services, transfer payments, taxation, the change in the sales of Government bonds to the public, and the change in the stock of high-powered money, should be closed by the specification of the relationship between the items representing Government expenditures and the items representing Government “income.” Closure of the above system implies that in simu lating the economy’s response to alternative monetary and fiscal actions, it is not possible to change only one of the five variables at a time when evaluating mul tipliers. Carl Christ provides the following example relating Government expenditures to Government receipts: . . . if government purchases are increased, then either transfers must be cut, or tax payments must rise, or government debt must be issued to the pri vate sector, or high powered money must be issued, or some combination of these. The effect of the in crease in government purchases will depend upon what combination of them is chosen.37 37Carl Christ, “Monetary and Fiscal Policy in Macroeconomic Models” (Paper presented at the Sixteenth Annual Con ference on the Economic Outlook, Ann Arbor, Michigan, November 14-15, 1968), p. 102. FEDERAL RESERVE BANK OF ST. LOUIS Thus, those models which purport to be able to say something like “A $1 billion increase of Government purchase of goods and services results in a one-quarter increase in income of $2.5 billion” should also make statements regarding another fiscal/monetary variable which changes in order for the statement to have meaning. With no way to determine how the $2.5 billion increase in income is generated, it may be falsely attributed to a fiscal action when, in reality, a monetary action is responsible. That is, if changes in money are not held constant in the analysis, the esti mated increase in income arising from a change in the independent fiscal policy variable may not be ac curately captured. Consequently, the model would not measure accurately fiscal crowding out. One analyst gives several examples of how, by failure to incorporate correctly the Government bud get constraint, a number of income-expenditure mod els overestimate the impact of Government actions on income. His own reduced form empirical evidence leads him to conclude that “These results are consist ent with a simple Keynesian multiplier from a deficit financed by bond purchases, with respect to the in come of the private sector equal to zero.”38 Neglect of the Government budget constraint is not the only reason why fiscal crowding out has not ap peared in some structural econometric models. If such monetary variables as money demand and supply are omitted completely from the model, the “real” sector of the economy —which normally includes Gov ernment spending actions —will, by default, dominate income changes. In terms of IS-LM analysis, omission of the LM curve (which reflects money demand and supply functions) would severely restrict the possible emergence of the crowding-out effect. Evidence presented in this Review is consistent with the thesis that failure to identify monetary vari ables adequately leads to suppression of the crowd ing-out effect. Michael Keran found that the use of interest rates rather than monetary aggregates as monetary variables in his reduced-form equations re sulted in the elimination of the fiscal crowding-out phenomenon.39 His study suggests that proper spe cification of monetary actions, in terms of monetary aggregates rather than interest rates, assists empirical estimators in uncovering fiscal crowding-out influences. 38Robert Auerbach, The Income Effects of the Government Deficit (Ph.D. Dissertation, University of Chicago, 1969), Appendix A and p. 49. 39Michael W. Keran, “Selecting a Monetary Indicator —Evi dence from the United States and Other Developed Coun tries” (this Review), September 1970. OCTOBER, 1970 Karl Brunner and Allan Meltzer have developed (in a preliminary paper) a framework of equations couched in terms of elasticities of interest rates, prices, wealth, and anticipations. Their approach to macroeconomic analysis, in contrast to the standard Keyne sian framework, emphasizes the whole spectrum of relative price adjustments (which includes changes in the prices of goods and services as well as interest rate responses) to monetary and fiscal actions.40 “The role of the relative price process is particularly exam ined together with the responses resulting from the interaction of the asset markets. The orthodox Keyne sian view of a reliable and direct’ effect of fiscal policy on income disolves rather thoroughly. The an alysis establishes that monetary and fiscal policy are equally ‘indirect’ and dependent on stimuli conveyed by relative price changes and adjustments in wealth positions.”41 Brunner and Meltzer believe their analysis “pro vides a foundation for the proposition that changes in budgetary variables (Government expenditures or taxes) exert by themselves relatively litde effect on economic activity or price-levels. The crucial effect depends on the financing.” They find that the Gov ernment expenditures elasticity of aggregate demand “varies between less than 1/5 up to unity . . with the higher values achieved through the injection of base money. “The pure fiscal effect of Government expenditures thus amounts to at most 1/5 measured in terms of elasticities.”42 Even the more conventional econometric models may uncover crowding-out influences, if interest rate effects are permitted to develop over a substantial period. Private spending is curtailed over time by a rise in interest rates generated by expanded Govern ment spending, but the result is not immediate. This 40Crowding out in a relative price context may be illustrated by the following simple example: If a newly-issued Gov ernment construction contract enables a Government enter prise to bid against private firms for construction and other workers in a near-fully employed community, the private firms must raise wages to retain their workers. Those pri vate firms losing workers will probably have to curtail out put, while those retaining workers with higher wages may attempt to pass along the higher costs in the form of price increases. The higher prices would be most noticeable in the construction sector, but, with the general rise in pur chasing power, could be transmitted throughout the com munity. Also, the higher prices of residential construction could cause households to shift expenditures away from housing into less expensive alternatives. 41 Karl Brunner and Allan Meltzer, “Fiscal and Monetary Policy in a Non-Keynesian World,” (paper prepared for private circulation), p. 6. *2Ibid., pp. 57 and 57a. Page 23 FEDERAL RESERVE BANK OF ST. LOUIS is particularly true of corporations whose decisions to invest are based on the economic climate which exists well before actual outlays are made. The specification of the lags involving interest-sensitive private spend ing plans may play a strong role in determining the time lag necessary for crowding out to occur. The FRB-M IT model, which is characterized by rather long lag structures, demonstrated complete crowding out of private spending by Government actions, when model simulations were tracked over a long period. “In the long run, and by this we mean a long run of ten or fifteen years, the [FRB-M IT] model is classical in that the only permanent effect of fiscal policy is to raise prices and the transactions demand for money and, in the presence of a fixed supply of money, interest rates sufficiently to crowd out enough real private expenditures that the ultimate real income effect of Government spending is zero.”43 Fiscal crowding out emerges in the reduced form equations published in this Review only after a pe riod of time, even though it is a much shorter period of time than that of the FRB-M IT model. Govern ment spending, as measured by high-employment ex penditures, exercises a strong influence on GNP (assuming a constant money supply) in the current quarter and the next quarter, but the Governmentspending effects wash out over approximately a oneyear period. These results should not be interpreted to suggest that “Government spending doesn’t matter.” It matters very much over a certain period. Moreover, if Government spending were to accelerate rapidly rather than be held to a once-and-for-all increase, the impact on GNP would be considerable over the period of acceleration and somewhat beyond. Conclusion A number of plausible theories have been devel oped over the years which substantiate the view that Government expenditures, depending on the source of finance, may crowd out a roughly equivalent magnitude of private expenditures. This view, in fact, was the dominant classical and neo-classical view, persisting at least until the publication of Keynes’ General Theory. Keynes himself discussed crowding out in detail in T he General Theory. He indicated that the fiscal OCTOBER, 1970 multiplier might not generate the increases in em ployment given by a mechanical manipulation of the equations centering on his consumption function, be cause of the restrictive assumptions upon which his equations were based. Keynes noted several factors tending to offset the influence of increased Govern ment expenditures: possible adverse reactions on pri vate investment, “confused” business psychology, and a tendency of the marginal propensity to consume to decline with rises in employment. A significant number of additional limitations to the Keynesian multiplier have been pointed out in the post-General Theory literature. One of the most seri ous deficiencies of the fiscal multiplier appears to be its asymmetry; that is, the crowding out of private spending is theoretically more likely at full-employment than at considerably less than full-employment conditions. Since the unemployment rate has rarely exceeded seven per cent in the past three decades, compared with an average unemployment rate of 18 per cent in the 1930’s, the crowding-out effect has probably been much stronger in recent years than during the period in which Keynesian multiplier the ory was developed. Wealth, income, and substitution effects, important factors in the determination of the degree of crowd ing out, are often incorporated in econometric models, but failure to impose the Government budget con straint or treat adequately the monetary variables in the system of equations probably has led to under statement of the crowding-out effect. More research on the time interval of crowding-out influences should be conducted to improve stabiliza tion policy recommendations. Articles published in this Review suggest that Government expenditures financed by taxes or borrowing from the public are important over a very short period, but their tendency to crowd out private expenditures obviates any significant, lasting influence. This conclusion is sup ported by other research, which indicates that crowd ing out does indeed occur, but over a much longer time period.44 These results suggest that the main dispute regarding the crowding out effect centers on the length of time involved. The rationale and some empirical verification of the existence of crowd ing out have been established —precise relationships and time periods remain a subject of further research. i4Ibid. 43Edward M. Gramlich, “Recent Experience with the FRBMIT Model,” (paper presented to the Committee on Bank ing and Credit Policy, New York, November 6, 1969), p. 6. Page 24 This article is available as Reprint No. 60.