The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
The Review is published 10 times per year by the Research Department o f the Federal Reserve Bank o f St. Louis. Single-copy subscriptions are available to the public free o f charge. Mail requests for subscriptions, back issues, or address changes to: Research Department, Federal Reserve Bank o f St. Louis, P.O. Box 442, St. Louis, Missouri 63166. Articles herein may be reprinted provided the source is credited. Please provide the Bank’s Research Department with a copy o f reprinted material. FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 A Monetary Analysis of the Administration’s Budget and Economic Projections KEITH M. CARLSON Teconomic JL H E adm inistration s budget proposals and report, presented to Congress and the This article analyzes the role of monetary actions in the current administration’s economic framework. nation in early February, have generated consider The discussion evaluates the consistency of the able controversy.1 The prospect of historically large adm inistration’s economic projections, given the deficits through 1987 has especially unsettled many structure of the economy and past experience with observers. Many question the plausibility of the lags in the effect of economic policy. The basis for adm inistration’s econom ic forecast, w hich they this evaluation is a monetary model of the U.S. econ consider too optimistic. omy developed at the Federal Reserve Bank of St. Louis.3 The implications of the analysis also are Economic forecasts have always been a critical applied to the federal budget outlook. part of the budget process. One can see, however, how their importance is magnified in an inflationswollen economy. A re-estimate of GNP growth by MONETARY ANALYSIS AND only 1 percent, for example, results in a change of THE ECONOMIC REPORT $13 billion in federal budget receipts within two The Economic Report of the President and The years.2 In addition, federal expenditures in recent years have become more sensitive to the pace of Annual Report of the Council of Economic Advisers inflation and output, as the num ber of inflation- (CEA Report) together provide a concise summary of indexed programs and income-security programs, the economic philosophy behind the adm inistra which automatically change in response to economic tion’s decision-making. President Reagan’s report devotes relatively little space to the subject of conditions, has increased. monetary policy, although the president states sup Aside from the budget issue, the administration’s port for “ . . . a policy of gradual and less volatile projections are of general interest because they reduction in the growth of the money supply.”4 This reflect the philosophy that guides the administra support contrasts with President Carter’s statem ent a tion’s economic policies. This year’s budget and year earlier “ . . . that public opinion not hold the economic report provide the first detailed statement Federal Reserve to such a rigid form of monetary of the administration’s economic philosophy. One targeting as to deprive it of the flexibility it needs to key difference from the previous administration’s conduct a responsible monetary policy.”5 philosophy is in the interpretation and role of mone tary actions in the determination of economic events. The most explicit discussion of the role of mone tary actions in the ad m in istratio n ’s econom ic framework is in the CEA Report. For example, in the opening chapter, which summarizes current eco 1Budget o f the United States Government for Fiscal Year 1983 nomic conditions, the CEA singles out the varying (hereafter referred to as Fiscal 1983 Budget) and the 1982 Economic Report o f the President, which also includes the 1982 Annual Report of the Council of Economic Advisers (hereafter referred to as CEA Report). 2See Fiscal 1983 Budget, p. 2:9. 3For details of this model, see the appendix. 41982 Economic Report o f the President, p. 8. s1981 Economic Report o f the President, p. 13. 3 MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS and generally restrictive rate of monetary expansion as the chief culprit responsible for the economy’s unsatisfactory performance in the 1979-81 period. The CEA goes on to say that “continued monetary restraint and a reduction of the within-year vari ability of money growth . . . are necessary both to reduce inflation and provide the basis for sustained economic growth.”6 The CEA Report’s overall theme is that the federal government’s role in economic affairs should be reduced. Consistent with that them e is a program to control inflation, which, as the CEA states forcefully, is essentially a monetary phenomenon. Thus, . . a decrease in money growth is the necessary strategy to end inflation.”7 In light of the important role that expectations play in the inflationary process, the CEA is very specific: “For the Federal Reserve, this means setting money growth targets consistent with a sustained decrease in the rate of inflation and then adhering to those targets.”8 After establishing these guidelines for an antiinflationary monetary policy, the CEA details the economic prospects for 1982, 1983 and beyond. Assumptions about money growth, however, do not play an explicit role in its economic forecasts. In stead, the CEA’s forecasts follow the traditional “ adding-up” approach typical of previous CEA Reports; that is, the activity of individual sectors are forecast and summed to obtain an aggregate forecast. Oddly enough, the CEA, after em phasizing the con nection betw een money growth and nominal magni tudes like GNP and the price level, and recognizing the relationship betw een deviations of m oney growth from trend and the movements ot real GNP, slights the role of money growth in their projections, particularly for 1982 and 1983.9 61982 CEA Report, pp. 24-25. ’ Ibid., p. 55. aIbi(L, pp. 59-60. 9The CEA attempts to correct for this oversight. It notes that: Concerns have been expressed that the Federal Reserve’s targets for money growth are not compatible with the vigorous upturn in economic activity envisioned late in 1982. . . We believe that such fears, while understandable on the basis of recent history and policies, are unjustified in light of current policies and the Admin istration’s determination to carrv them through. (1982 CEA Report, p. 25.) This statement contrasts sharply with a statement found else where in its report: Indeed, changes in the trend of the growth rate of nominal GNP over the period 1960 to 1981 are almost entirely attributable to changes in the trend of the growth rate of the money stock (M l), as opposed to changes in the trend of the growth rate of velocity (Chart 3-3). (1982 CEA Report, p. 63.) 4 ADMINISTRATION ECONOMIC PROJECTIONS E ver since enactm ent of the C ongressional Rudget and Im poundm ent Control Act of 1974 (hereafter referred to as the Control Act), the in cum bent administration has been required each year to present five-year projections ofthe federal budget. Thus, the current budget and econom ic reports cover the period through 1987. The administration also must set five-year num er ical goals for several key economic indicators under the provisions of the Full Em ployment and Ralanced Growth Act of 1978 (Humphrey-Hawkins). This act originally specified the following goals: an unemploym ent rate of 4 percent and a rate of in crease in consumer prices of 3 percent by 1983, and an interim goal for federal outlays to equal 21 percent of GNP by 1981. However, the act allowed a change in this tim etable if deem ed necessary, and, in January 1980, President Carter extended the timetables for unemploym ent to 1985 and for infla tion to 1988. A Review o f Previous Long-Term Projections Incum bent administrations have been presenting long-term economic projections since the passage of the Control Act in 1974. Table 1 summarizes these projections.10 They represent the efforts of three different administrations: President Ford’s.in 197577, President C arter’s in 1978-81 and President Reagan’s in early 1982. The table indicates that, for each administration, the one-year forecasts have been quite accurate for all of the indicators.11 In fact, the record for GNP is good as far as four years ahead. For all the other major indicators, the forecasts tend to deteriorate beyond the two-year horizon. This may reflect the practice 10The table is limited to the official reports published in January or February of each year and thereby excludes revised esti mates when a new administration comes into power and those contained in the mid-session review of the budget. 11The root-mean-squared errors for table 1 are as follows: Real GNP Unemployment GNP GNP deflator rate 1 year ahead 0.92 1.00 0.97 0.22 2 years ahead 1.01 1.32 1.72 0.45 3 years ahead 1.14 2.77 1.16 2.6.3 4 years ahead 0.98 3.72 .3.59 1.75 5 years ahead 2.46 4.45 4.88 1.97 6 years ahead 2.16 5.16 5.10 2.22 FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 Table 1 Administration Economic Projections (percent) Date of forecast 1975 1976 1977 1978 1979 1980 10.8 1981 1982 1983 1984 1985 1986 1987 G NP Early 1975 7.2 1976 12.6 12.4 12.0 10.8 12.4 12.2 12.4 11.9 10.9 9.1 11.0 11.3 11.6 10.5 7.9 6.4 11.0 11.2 10.8 10.5 9.6 11.3 9.5 10.1 9.4 7.9 6.3 8.3 10.7 12.8 12.9 12.0 11.0 11.4 13.1 12.3 11.8 11.0 10.2 8.1 11.5 10.2 9.7 9.2 1977 1978 1979 1980 1981 1982 A ctua l1 8.0 12.4 12.0 8.8 8.5 9.0 11.4 10.9 11.6 4.8 5.6 6.5 6.5 6.5 6.2 5.7 5.9 6.5 6.5 4.9 5.2 5.1 5.9 5.5 3.9 3.5 4.7 4.8 4.8 5.0 4.7 4.2 3.3 2.5 4.2 4.7 4.4 3.4 -0 .6 1.7 4.3 5.0 4.9 0.9 3.5 3.5 3.7 3.7 3.7 0.2 5.2 5.0 4.7 4.4 Real G NP Early 1975 -3 .3 1976 1977 1978 1979 1980 1981 1982 A ctua l1 -1 .1 5.4 5.5 4.8 3.2 -0 .2 4.0 4.7 4.3 2.0 Price d eflato r Early 1975 10.8 1976 7.5 6.5 5.1 4.1 5.9 6.2 6.1 5.0 4.2 4.0 5.6 5.9 5.4 4.7 3.8 2.8 6.1 6.2 5.7 5.2 4.7 4.2 7.7 6.8 5.7 4.5 3.4 2.8 8.9 8.8 8.2 7.4 6.8 6.1 10.5 9.3 8.5 7.8 7.0 1977 1978 1979 1980 1981 1982 A ctua l1 7.9 9.3 5.2 6.0 5.0 4.7 6.3 4.6 4.5 5.8 7.3 8.5 9.0 9.2 7.9 7.5 6.9 6.2 5.5 7.7 6.9 6.4 5.8 5.2 4.9 7.3 6.6 5.7 4.9 4.8 4.7 6.3 5.9 5.4 5.0 4.5 4.1 6.0 6.2 5.7 4.9 4.2 4.0 7.0 7.4 6.8 5.9 5.1 4.3 7.8 7.5 7.1 6.7 6.3 6.0 7.1 6.4 5.8 Unem ploym ent rate Early 1975 8.1 1976 1977 1978 1979 1980 1981 1982 A ctua l1 8.9 8.5 7.7 7.1 6.1 5.8 7.1 7.9 5.3 7.6 1As of February 1982 5 FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 economic policies. According to the most recent CEA report, “The events of the past 15 years are a good illustration of the danger of pursuing economic policies based on short-run analysis and focused on immediate problems. Sound policy requires em phasis on a time horizon during which the some times lengthy, and usually unpredictable, lags in economic processes can work.”12 C h a rt 1 Inflation an d U n em p lo y m e n t In f la t io n ro te LL P trC S lt lAnm m l Current Projections 0 (( 3 4 5 6 7 8 9 U n e m p lo ym e n t rate 12 Percent S o u r c e s : U .S. D e p a r t m e n t o f C o m m e r c e a n d U .S . D e p a r tm e n t o f L a b o r L i P e r c e n t a g e c h a n g e in th e G N P i m p l i c i t p r i c e d e f l a t o r . [2^ P e r c e n t o f c i v i l i a n l a b o r fo r c e . whereby assumptions for the current and next year are called “forecasts,” but beyond the next year are labeled “projections consistent with moving grad ually toward relatively stable prices and maximum feasible employment.” For the longer term, these projections seemingly ignore or seriously misjudge some fundamental economic constraints. The failure of the U.S. economy to achieve relative price stability and “full employment” is obvious when one compares the projection record for these two indicators with actual performance. (For addi tional historical perspective, see chart 1.) Since the start of publishing long-term projections, each ad ministration has projected a general decline of both inflation and unemployment. The actual perform ance of the economy, of course, has been far different. Though the rate of inflation declined from 1975 to 1976, it has accelerated on an annual average basis each year since then. The unem ploym ent rate did fall from 1975 through 1979, but since then has risen sharply. Such persistent forecast errors are probably a reflection of the fact that each administration gives insufficient weight to the long-term effects of its Table 2 summarizes the Reagan adm inistration’s economic projections. The nominal GNP goal for fourth quarter 1987 is $5,248 billion, which would mean a 9.8 percent average annual rate of increase from 1981 to 1987. This rate would be distributed as a 4.4 percent rate of expansion in real GNP and a 5.2 percent rate of increase in the GNP deflator. In 1987, according to these projections, real GNP would be growing at a 4.3 percent rate, the GNP deflator would be rising at a 4.4 percent rate and the unemployment rate would decline to 5.2 percent by the fourth quarter. As a part of its program, the administration has proposed a budget plan aim ed at a year-by-year reduction in the size of the federal deficit. Federal outlays are projected to decline to 19.7 percent of GNP in fiscal 1987 compared with an estimated 23.5 percent in fiscal 1982. More importantly, however, the administration announced its support of a m one tary policy that will produce continued gradual reductions in the rate of monetary growth. F ro m th e fo u rth q u a rte r o f 1979 to th e fo u rth q u a rte r of 1980, M l (cu rren cy p lu s c h eck a b le d e posits) g rew at a 7.3 p e rc e n t an n u al rate. T h e A d m in istration assum es a g radu al b u t stead y red u c tio n in th e grow th o f m on ey to o n e -h alf th a t rate b y 1986.13 The CEA notes that inflationary expectations must adjust speedily to the anti-inflationary monetary regime in order to attain these economic goals.14 A MONETARY ANALYSIS OF ADMINISTRATION PROJECTIONS In sharp contrast to previous administrations, the present administration has explicitly spelled out a target path for monetary growth. It is therefore of 121982 CEA Report, pp. 49-50. l3Ibicl„ p. 206. 14Ibid., p. 26. FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 Table 2 Administration’s Economic Projections: 1982-87 (from fiscal 1983 budget)1 GNP (billions of dollars) IV/1981 Actual Real GNP (billions of 1972 dollars) Prices (1972=100) Unem ploym ent rate M1 (billions of dollars) $2995 (9.7) $1498 (0.8) 200.0 (8.8) 8.4% IV/1982 3307 (10.4) 1543 (3.0) 214.4 (7.2) 8.4 457.4 (4.7) IV/1983 3671 (11.0) 1623 (5.2) 226.2 (5.5) 7.6 477.9 (4.5) IV/1984 4038 (10.0) 1702 (4.9) 237.2 (4.9) 6.8 498.1 (4.2) IV/1985 4417 (9.4) 1781 (4.6) 248.1 (4.6) 6.2 517.8 (4.0) IV/1986 4819 (9.1) 1857 (4.3) 259.5 (4.6) 5.6 537.0 (3.7) IV/1987 5248 (8.9) 1937 (4.3) 270.9 (4.4) 5.2 555.5 (3.4) 1981-87 (9.8) (4.4) (5.2) 6.6 (4.1) $436.7 (5.0) NOTE: All GNP data adjusted to February 1982 revision of NIA accounts; M1 reflects revision of February 1982. M1 figures correspond to m onetary policy assum ption stated in the 1982 CEA Report. 'Rates of change in parentheses. interest to see how the administration’s projections compare with those derived from an explicitly mone tarist model. The framework used for this compar ison is a revised and updated version of the “ St. Louis model.”15 According to the St. Louis model, nominal GNP is determ ined directly by a reduced-form equation relating the percent change in GNP to current and past changes in money (M l) and high-employment federal expenditures (national income accounts basis). Estimates of this equation indicate that the growth of federal spending has little net effect on GNP over a period of a year or m ore.16 The primary factors affecting GNP growth are the rate of change of money and trend velocity, as em bodied in the coefficients of the equation. The change in GNP is distrib u ted betw een changes in the price level and output via a price equation. The price equation specifies the percent change in the GNP deflator as a function of energy prices, demand pressure and the recent history of price change.17 Over the long run, the estimated change in the price level is dominated by the trend of money growth. Given the change in GNP and the change in the price level, the change in output is found via the GNP identity; that is, GNP equals price level times output. The unem ployment rate also is solved for as a part of the St. Louis model. Estim ated changes in output along with assumptions about the growth of poten tial output provide the basis for calculating the unemploym ent rate via Okun’s Law.18 15For a discussion of the original model, see Leonall C. Andersen and Keith M. Carlson, “A M onetarist Model for Economic Stabilization,” this Review (April 1970), pp. 7-25. For a detailed summary of the model in revised and updated form, see the appendix. 18For a recent study of the impact of fiscal actions on GNP, see R. W. Hafer, “The Role of Fiscal Policy in the St. Louis Equation,” this Review (January 1982), pp. 17-22. 17For a further discussion of the role of energy prices in the determination of the price level, see John A. Tatom, “Energy Prices and Short-Run Economic Performance,” this Review (January 1981), pp. 3-17. 18Arthur M. Okun, “Potential GNP: Its M easurem ent and Sig nificance,” 1962 Proceedings o f the Business and Economic Statistics Section o f the American Statistical Association, pp. 98-104. 7 FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 Table 3 St. Louis Model Projections for 1976-81: An Ex Post Comparison Adm inistration Projections as of Mid-1977 GNP Real GNP Prices Unem ploym ent rate M1 1976 Actual 1977 11.6% 6.0% 5.3% 7.7% 11.3 5.1 5.9 7.0 e xp licit 1978 11.9 5.3 6.3 6.3 assum ption 1979 11.3 5.0 6.1 5.7 1980 10.6 5.2 5.1 5.2 1981 9.8 4.9 4.3 4.8 1982 8.6 4.3 4.2 4.5 11.0 5.1 5.5 6.1 1976-81 No 1977 St. Louis M odel Projections with Adm inistration G N P Path GNP Real GNP Prices Unem ploym ent rate M1 1976 Actual 11.6% 6.0% 5.3% 7.7% 1977 11.2 5.2 5.7 7.1 6.8 1978 12.1 5.7 6.1 6.1 7.7 1979 11.1 4.5 6.5 5.7 7.8 1980 10.7 2.9 7.6 5.6 6.8 1981 9.7 0.5 9.1 6.5 6.0 1982 8.7 -0 .8 9.5 8.2 5.1 11.0 3.8 7.0 6.5 7.0 1976-81 5.1% A ctual Perfo rm ance Using D ata as of F ebruary 1982 GNP Real GNP Prices Unem ploym ent rate M1 1976 Actual 1977 10.9% 5.4% 5.2% 7.7% 11.6 5.5 5.8 7.1 7.7 1978 12.4 4.8 7.3 6.1 8.2 1979 12.0 3.2 8.5 5.8 7.8 1980 8.8 -0 .2 9.0 7.1 6.3 1981 11.4 2.0 9.2 7.6 6.9 1976-81 11.2 3.0 7.9 6.9 7.4 5.7% NOTE: Adm inistration and St. Louis Model projections taken from November 1977 Review. To illustrate the projection performance of the St. Louis model, table 3 presents an ex post summary of projections made in this Review in the fall of 1977.19 The relevant projection period at that time was 197781. The administration’s GNP projections at that time implied a path of declining growth in money, a 19Keith M. Carlson, “ Economic Goals for 1981: A Monetary Analysis,” this Review (November 1977), pp. 2-7. The major differences in the model used at that time and the version described in the appendix are in the treatm ent of energy prices and the adjustm ent for serial correlation. 8 path that was used in simulating the St. Louis model. Since the actual path of monetary expansion was similar to that assumed in simulating the model and that implicit in the administration’s projections, the growth of GNP was forecast with considerable accuracy by both the administration and the model. There w ere differences, however, betw een the administration’s and the St. Louis m odel’s forecasts for real GNP, the price level and the unem ploym ent rate, particularly after 1978. In contrast to the ad ministration’s forecast, the model projected a slow FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 Table 4 St. Louis Model Simulations: 1982-87 (assuming administration’s GNP path)1 GNP (billions of dollars) IV/1981 Actual Real GNP (billions of 1972 dollars) Prices (1972=100) Unem ploym ent rate M1 (billions of dollars) $2995 (9.7) $1498 (0.8) 200.0 (8.8) 8.4% $436.7 (5.0) IV/1982 3306 (10.4) 1538 (2.7) 215.1 (7.5) 8.8 471.2 (7.9) IV/1983 3670 (11.0) 1603 (4.3) 229.1 (6.5) 8.1 507.0 (7.6) IV/1984 4037 (10.0) 1662 (3.7) 243.2 (6.2) 7.7 540.0 (6.5) IV/1985 4416 (9.4) 1720 (3.5) 257.1 (5.7) 7.5 572.9 (6.1) IV/1986 4819 (9.1) 1787 (3.9) 270.2 (5.1) 7.2 606.7 (5.9) IV/1987 5249 (8.9) 1861 (4.1) 282.8 (4.6) 6.8 641.3 (5.7) 1981-87 (9.8) (3.7) (5.9) 7.7 (6.6) ’ Rates of change in parentheses. ing in output and an acceleration of the price level in the latter part of the period, both of which occurred. The results of this simulation, shown in table 4, should be compared with those in table 2. It should be noted first that the path of money growth required to attain the administration’s projected GNP path is substantially higher than what they explicitly state Simulation Using Administration as desired. Assuming that this GNP path is attained, GNP Growth Path however, the St. Louis model indicates that the The first issue addressed here is the feasibility of administration’s projections are indeed optimistic. the output and inflation scenarios. The analysis does The model indicates an unem ploym ent rate of 6.8 not, at this point, examine the question whether percent in late 1987 in contrast to the adm inis GNP can be attained with the administration mone tration’s projected 5.2 percent rate, with annual real tary assumptions; it focuses exclusively on its pro growth averaging 0.7 percent lower for the model jections of inflation and output growth, given its path simulation. The model is also more pessimistic on for the growth of GNP. The assumptions used for the inflation, indicating an annual average inflation rate other exogenous variables in the St. Louis model are of 5.9 percent instead of the administration’s esti as follows: potential GNP is assumed to grow 3.3 mated 5.2 percent. percent per year from late 1981; growth in highemployment federal expenditures is projected at 6.3 Alternative Simulations percent per year; and the change in the relative price Since the adm inistration explicitly supports a of energy is assumed to be zero.20 monetary policy of gradual reduction in the rate of monetary growth, the results of this scenario, in 20These assumptions are designed to be consistent with the ad which M l growth is reduced gradually and steadily ministration’s, even though they do not provide specific esti mates of these variables in either the CEA Report or the Fiscal to a 3.7 percent rate in 1986, are summarized in table 1982 Budget. For a discussion of prospects forreal GNP growth, 5. All other assumptions are the same as in the previous simulation. see 1982 CEA Report, pp. 115-17. MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS As might be expected, the model shows a growth rate of nominal GNP much less than the adminis tration has projected (compare with table 2). The CEA is aware of this discrepancy, but does not ex plain why the assumed growth of velocity should far exceed its historical rates of growth (see chart 2).21 For this scenario of a gradual reduction of money growth, the model indicates that the administration’s inflation goal is easily achieved; in fact, the simulated inflation rate falls well below the administration’s projected rate after 1983.22 The simulated path for real GNP, however, is considerably different than the administration has projected. In the early years, 1982-84, the model simulates much slower output growth, followed by faster growth in the later years. As a result, the simulated unem ploym ent rate is still at a high 6.9 percent in late 1987 compared with an administration estimate of 5.2 percent. Finally, a third simulation was run, based on a constant 5 percent annual growth in money through 1987. The results are shown in table 6. This steady money growth path comes closer to attaining both of the adm inistration’s inflation and unem ploym ent goals than either of the simulations summarized in tables 4 and 5. With steady 5 percent money growth, inflation averages 3.9 percent per year for the pro jectio n period, and the unem ploym ent rate is brought to near 6 percent by late 1987. C h a rt 2 R ate o f C h a n g e of M l V e lo c ity 11 LL D a ta a r e t w o - y e a r r a te s o f c h a n g e , u s in g fo u r th q u a r t e r d a t a . D a s h e d l i n e is i m p l i c i t in a d m in is t r a t io n p r o je c tio n s . V e lo c ity is G N P d i v i d e d b y M l . The more fundamental question yet to be an swered is how the administration expects GNP to Money Growth and the Administration's grow rapidly if money growth gradually declines. With the administration making explicit statements Projections: The Basic Conjiict about interest rates falling in future years, appar The administration has emphasized that it is im ently the result of declining inflation, velocity portant to establish credibility in economic policy in growth might be expected to slow rather than accel order to “ break the back” of inflation expecta erate. Furthermore, velocity growth historically has tions. Behind this strategy is the presumption that, if been remarkably stable over time, an observation inflation can be reduced more rapidly than past that the CEA itself has em phasized.23 Thus, while relationships would indicate (e.g., faster than is the output-inflation breakdown of GNP in the St. em bodied in the estimates from St. Louis model), Louis model may be open to question, there seems to greater output growth would result. This prospect be little reason to question its GNP projections. would produce a brighter outlook for the interim years than shown in the simulations em ploying THE FEDERAL BUDGET OUTLOOK gradual money reduction (table 5). There is little AND ECONOMIC PROJECTIONS likelihood, however, that the unem ploym ent rate would be reduced to as low as the administration’s The administration’s economic projections are of estimate of 5 percent. interest because they indicate how the nation’s economic welfare can be expected to change in coming years. They are also of interest because of 21See footnote 9. their impact on estimates of the budget deficit. The 22Over the long run in the St. Louis model, the inflation rate approxim ates the rate of m onetary grow th. Prior to the achievem ent of this equilibrium, however, the St. Louis model oscillates. Digitized for 10 FRASER 23See footnote 9. MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS Table 5 St. Louis Model Simulations: 1982-87 (assuming declining growth rate of money from 5.0 percent rate in 1981-82)1 GNP (billion s of dollars) Real GNP (billions of 1972 dollars) Prices (1972=100) $2995 (9.7) $1498 (0.8) 200.0 (8.8) 8.4% IV/1982 3227 (7.7) 1501 (0.2) 215.0 (7.5) 9.7 457.4 (4.7) IV/1983 3472 (7.6) 1528 (1.8) 227.3 (5.7) 9.9 477.9 (4.5) IV/1984 3727 (7.3) 1581 (3.5) 235.8 (3.7) 9.7 498.1 (4.2) IV/1985 3989 (7.0) 1659 (4.9) 240.7 (2.1) 9.0 517.8 (4.0) IV/1986 4259 (6.8) 1754 (5.8) 242.9 (0.9) 8.0 537.0 (3.7) IV/1987 4534 (6.5) 1860 (6.0) 244.0 (0.4) 6.9 555.4 (3.4) 1981-87 (7.2) (3.7) (3.4) 8.9 (4.1) IV/1981 Actual U nem ploym ent rate M1 (billion s of dollars) $436.7 (5.0) 1Rates of change in parentheses. Table 6 St. Louis Model Simulations: 1982-87 (assuming steady growth rate of money of 5.0 percent)1 GNP (billion s of dollars) Real GNP (billions of 1972 dollars) Prices (1972=100) Unemploym ent rate M1 (billion s of dollars) $2995 (9.7) $1498 (0.8) 200.0 (8.8) 8.4% IV/1982 3233 (8.0) 1504 (0.4) 215.0 (7.5) 9.6 458.6 (5.0) IV/1983 3495 (8.1) 1537 (2.2) 227.5 (5.8) 9.7 481.5 (5.0) IV/1984 3779 (8.1) 1580 (4.0) 236.6 (4.0) 9.3 505.6 (5.0) IV/1985 4085 (8.1) 1683 (5.3) 243.0 (2.7) 8.4 530.8 (5.0) IV/1986 4416 (8.1) 1784 (6.0) 247.7 (2.0) 7.3 557.4 (5.0) IV/1987 4774 (8.1) 1895 (6.2) 252.3 (1.8) 6.1 585.3 (5.0) 1981-87 (8.1) (4.0) (3.9) 8.4 (5.0) IV/1981 Actual $436.7 (5.0) ’ Rates of change in parentheses. 11 MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS Table 7 Alternative Budget Estimates: Fiscal 1987 (billions of dollars) Receipts Adm inistration estimates from fiscal 1983 budget Outlays Surplus/D eficit $926 $979 $ -5 3 St. Louis model sim ulation using adm inistration's GNP path 926 1028 -1 0 2 St. Louis model sim ulation assuming declining grow th rate of money 781 925 -1 4 4 St. Louis model sim ulation assuming steady 5 percent grow th of money 829 940 -11 1 At the same time, federal outlays have become increasingly sensitive to variations in economic activity. The usual effect via unem ploym ent insur ance continues to operate, but, like the revenue side, a given unem ploym ent rate now involves a greater amount of dollar expenditures than before. In addi tion, automatic changes in outlays for a num ber of welfare programs occur when the economy slows down or speeds up. In fact, approximately 30 percent Economic Activity and the Budget of federal outlays now are indexed to inflation. Although the effect of the budget on economic Finally, interest payments on the national debt, an growth is still an open issue, there is no question that important endogenous component of the budget, the budget is sensitive to the pace of economic ac reflect both the size of the deficit and the level of tivity. This relationship received added emphasis in interest rates. this year’s budget docum ent as budget figures appear to have become more and more sensitive to Budget Implications of economic conditions. Alternative Simulations In prior years, analyses of the connection betw een To examine the sensitivity of budget estimates to the budget and the economy focused on government revenues. Given our tax laws, different revenue alternative economic assumptions, budget equa estimates depend on the assumptions made about tions were added to the St. Louis model. The growth GNP and such related indicators as wages and of receipts was specified as a function of the growth salaries, and corporate profits. The relationship still of nominal GNP, using the elasticity im plied in the holds, of course, but the size of today’s economy is so administration’s budget document.25 The growth of large that a given growth rate of GNP translates into a outlays was expressed as a function of the growth of much different dollar amount of federal revenues output and the rise in prices, again using the relevant than it did just a few years ago. This relationship elasticities from the budget document. betw een GNP and government revenues is impor Table 7 summarizes the budget results for fiscal tant because public attention seems to focus on the 1987 for all three simulations. Only results for fiscal dollar size of the federal deficit. size of prospective deficits has become an issue among economic analysts, presumably because they consider it an indicator of the government’s impact on credit markets and, thus, on long-term economic growth.24 However, as is shown below, the process of estimating the deficit is an imprecise exercise. 24Such an effect is not in the St. Louis model; incorporation of this presum ed relationship betw een the size of the deficit and the rate of economic growth would require specifying potential output as a function of either the size of the deficit or the size of government. The only role for federal deficits in the St. Louis model is their possible relationship to the rate of money growth. Digitized for 12 FRASER 25Fiscal 1983 Budget, pp. 2:6-13. The im plied elasticities are found by com paring the budget effects of three econom ic scenarios. These scenarios are higher inflation/same growth, higher growth/lower inflation, and lower growth/higher infla tion, w ith all alternatives defined w ith reference to the administration’s basic economic projections (summarized in table 2). FEDERAL RESERVE BANK OF ST. LOUIS 1987 are given to ease the comparison of alternative policy scenarios. Moreover, focusing on 1987 illus trates the imprecision that encompasses any budget estimates, because a small change in growth rates can translate into a difference of many billions of dollars. All sim ulations assume that the basic proposals contained in the fiscal 1983 budget are enacted.26 The differences in results reflect only the impact of differing economic assumptions. The first simulation, using the administration’s GNP path as shown in table 4, yields a deficit of $102 billion; the administration estimates $53 billion. The estimate for receipts is the same as the adm in istration’s because the growth of nominal GNP is the same. Outlays are higher for this simulation because of higher inflation estimates, which push up outlays for indexed programs, and lower real growth esti m ates, w hich boost outlays for unem ploym ent com pensation and other unem ploym ent-related welfare programs. The second simulation, based on a gradual re duction of money growth (see table 5), yields a much larger deficit in 1987 than the administration pro jects. Outlays are less than projected by the admin istration because inflation is slower, but receipts fall 26This also assumes the Economic Recovery Tax Act of 1981 is left intact. The basic proposals them selves have been revised since February, but details await the outcome of negotiations be tw een Congress and the administration. The purpose of the estimates presented here is to illustrate the budget impact of alternative economic assumptions without actually attempting to forecast the size of the deficit. MAY 1982 even more sharply because the growth of nominal GNP is much less rapid. As a result, the deficit is estimated at $144 billion for 1987 — despite the incorporation of the administration’s proposals to reduce government programs in the 1983 budget. The third simulation, based on steady 5 percent money growth (see table 6), yields a slightly larger deficit than the simulation using the administration’s GNP path. However, both outlays and receipts are lower than in that case. SUMMARY AND CONCLUSIONS The administration has presented a controversial set of economic assumptions and budget projections for the years through 1987. Some simulations of a monetarist model, however, demonstrate that the adm inistration’s projections contain fundamental inconsistencies. Based on U.S. economic experience since 1960, (1) the administration’s estimates for GNP growth are inconsistent with its stated monetary targets; and (2) given its GNP growth path, its estimates of real growth, unemployment and, to a lesser extent, infla tion appear too optimistic. These conclusions also indicate that the admin istration’s estimates of the size of the federal deficit are imprecise. Given the administration’s budget plan, the pattern of declining growth in money that it supports will result in a deficit of about $144 billion in 1987, $93 billion more than is projected in the fiscal 1983 budget. 13 MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS Appendix Revised Form of St. Louis Model 1 The version of the St. Louis model used for the slack variable is entered in real rather than nominal simulations in this article is summarized in table 1, terms; and (3) where relevant, the m odel’s equations with the coefficients given in table 2. Equations 1, 2 have been corrected for serial correlation problems. and 4 are estimated with Almon constraints on the coefficients. Equation 1 is estim ated with ordinary least squares. Three characteristics differentiate this 'F o r further discussion, see Keith M. Carlson and Scott E. Hein, “An Analysis of a Modified St. Louis M odel,” a paper prepared model from the original version published in 1970: for the Spring Conference on Comparing the Predictive Per (1) most variables are entered in rate-of-change form formance of Macroeconomic Models at Washington University rather than first-difference form; (2) the dem and in St. Louis (April 20, 1982). Table 1 Table 2 The Model In-Sample Estimation: 1/1960-IV/1980 (absolute value of t-statistic in parentheses) (1) Y, = C 1 + 2 CMj (M,.j) + i= 0 4 (2) P, = C2 + 2 CPE, (PE,.j) + 1=1 2 CE; (E,_i) + el, i= 0 5 2 CDi (X,. - XF,V) i= 0 (1) Y, = 2.44 + 0.40 M, + 0.39 M,., + 0.22 M,.2 + 0.06 fClt.3 (2.15) (3.38) (5.06) (2.18) (0.82) - 0.01 M,_4 + 0.06 f£, + 0.02 Em - 0.02 E,_2 (0.11) (1.46) (0.63) (0.57) + CPA (PA,) + CDUM1 (DUM1) + CDUM2 (DUM2) + e2, - 0.02 E,_3 + 0.01 E,_4 (0.52) (0.34) 21 (3) PA, = 2 CPRL,(PH) i= 1 (4) RL, = C3 + 20 2 CPRL, (P,.;) + e3, i= 0 (5) U, - UF, = CG (GAP,) + CG1 (GAP,.,) + e4, R 2 = 0.39 SE = 3.50 DW = 2.02 (2) P, = 0.96 + 0.01 PE,., + 0.04 PE,.2 - 0.01 PE, .3 (2.53) (0.75) (1.96) (0.73) + 0.02 PEt.4 - 0.00 (X, - XF,*) + 0.01 (X,., - X F *,) (1.38) (0.18) (1.43) (6) Y, = (P,/100) (X,) (7) Y, = ( (Y^Y,.,)4 - 1) 100 (8) X, = ( (X,/X,.,)4 - 1) 100 (9) P, = ( (P,/P,.1)4 - 1) 100 (10) GAP, = ( (XF, - X,)/XF,) 100 (11) XF* = ( (XF,/X,.,)4 - 1) 100 + 0.02 (X,.2 - XFt*.2) + 0.02 (Xt_3 - XFf. 3 ) (4.63) (3.00) + 0.02 (X,_4 - XF,*4) + 0.01 (X, 5 - XFt*5) + 1.03 (PA,) (2.42) (2.16) (10.49) - 0.61 (DUM1t) + 1.65 (DUM2t) (1.02) (2.71) R2 = 0.80 Y M E P PE X XF RL U UF DUM1 DUM2 = = = = = = = = = = = = nom inal GNP money stock (M1) high em ploym ent expenditures GNP deflator (1972 = 100) relative price of energy ou tput in 1972 dollars potential ou tput (Rasche/Tatom) corporate bond rate unem ploym ent rate unem ploym ent rate at fu ll employment con tro l dumm y (111/1971-1/1973 = 1; 0 elsewhere) post control dum m y (1/1973-1/1975 = 1; 0 elsewhere) Digitized for 14 FRASER SE = 1.28 (4) RLt = 2.97 + 0.96 DW = 1.97 p = 0.12 DW = 1.76 p = 0.94 20 2 P,., i= 0 (3.12) (5.22) R2 = 0.32 SE = 0.33 (6) U, - UFt = 0.28 (GAP,) + 0.14 (GAP,.,) (11.89) (6.31) R 2 = 0.63 SE = 0.17 DW = 1.95 p i = 1.43 pz = 0.52 Short-Run Money Growth Fluctuations and Real Economic Activity: Some Implications for Monetary Targeting DALLAS S. BATTEN AND R. W. HAFER h er e is ample evidence that the rate of infla tion is directly related to the long-term growth of the money supply. Indeed, this relationship has been demonstrated for various countries.1 The implica tion of this finding is that the control of money growth over the long term is vital to the control of inflation, a realization that undoubtedly helps to explain the fairly recent announcements of monetary growth targets in m ost of the major industrial countries.2 Although the money growth/inflation connection is fairly well-documented, the relationship between short-run movements in money growth and eco nomic activity is less well-known. Even though this connection has been demonstrated for the United States, its general applicability has not been tested.3 The purpose of this article, therefore, is to in vestigate the relatio n sh ip b etw een short-run movements in the growth of the money stock and T 'D allas S. Batten,“Money Growth Stability and Inflation: An International Comparison,” this Revieiv (October 1981), pp. 712. See also Richard T. Selden, “Inflation and Monetary Growth: Experience in Fourteen Countries of Europe and North America Since 1958,” Federal Reserve Bank of Richmond Economic Review (November/December 1981), pp. 19-35. zO f the Group of Ten countries plus Switzerland, only two, Belgium and Sweden, do not formally announce monetary growth targets of some kind. See Organization for Economic Co-operation and Development, Monetary Targets and Inflation Control (Paris:OECD, 1979). 3Milton Friedman and Anna J. Schwartz, “ Money and Business Cycles,” Revieiv o f Economic* and Statistics (February 1963), pp. 32-78; W illiam Poole, “The R elationship of M onetary Decelerations to Business Cycle Peaks: Another Look at the Evidence,” Journal of Finance (June 1975), pp. 697-712; and Leonall C. Andersen and Keith M. Carlson,“A M onetarist Model for Economic Stabilization,” this Review (April 1970), pp. 7-25. fluctuations in real economic activity.4 Although the evidence presented in this article is not de rived from a rigorous empirical analysis, it indicates quite convincingly that virtually every downturn in economic activity in recent years in each of the countries examined was preceded by a significant reduction in the growth of its narrowly defined money supply. MONEY AND ECONOMIC ACTIVITY: THE THEORY T here is little disagreem ent that significant changes in the growth of the money supply influence economic activity. Changes in the long-term growth of money, measured by some moving average of money growth over a num ber of years, affect the rate of inflation. Indeed, several empirical studies of the United States indicate that it may take as long as five years for the rate of inflation to reflect completely the impact of a change in money growth.5 More recent 4The evidence presented also sheds light on the debate about the impact of M l growth during periods of financial innovation and institutional change. By examining the connection between short-run fluctuations in M l growth and real economic activity across countries with different financial institutions and regu lations, some understanding of the relationship’s robustness in a changing financial environm ent may be gained. For a good example of the uncertainty that pervades current thinking on the future efficacy of targeting on M l, see Anthony M. Solomon, “ Financial Innovations and Monetary Policy,” Federal Reserve Bank of New York, Annual Report, 1981 (1982), pp. 3-17; and Edward Yardeni, E. F. Hutton Economics Alert (January 29, 1982). 5See Denis S. Karnosky, “The Link Between Money and Prices — 1971-76,” this Review (June 1976), pp. 17-23; Keith M. Carlson, “The Lag From Money to Prices,” this Review (October 1980), pp. 3-10; and John A. Tatom, “Energy Prices and Short-Run Economic Performance,” this Revieiv (January 1981), pp. .3-17. 15 FEDERAL RESERVE BANK OF ST. LOUIS studies also have dem onstrated that a lengthy lag betw een money growth and inflation is common in several industrial countries.6 This evidence indi cates that changes in current money growth have a relatively small impact on prices in the short run. For short-run changes in money growth to affect economic activity, they must initially influence the real economy more significantly than they influence prices.7 Indeed, studies have shown that, at least for the United States, sizable reductions in money growth below its established trend rate for only a few quarters have preceded declines in real economic activity.8 The economic theory that “predicts” the results just described is as intuitively appealing as it is empirically verifiable. A marked and sustained de cline in the growth of the money supply creates a “monetary disequilibrium ” : the quantity of money that individuals desire to hold exceeds the quantity' that they are actually holding. By reducing their spending, they can increase their money holdings to a desired level. Eventually, this reduced spending will cause, the rate of inflation to fall. In the short run, however, producers who cannot tell immediately whether this decline in aggregate dem and (spending) is perm anent or just a temporary aberration initially react to the reduction in money growth (and spending) by reducing output. There fore, the decline in money growth results in a slow down in economic activity; if it is pronounced enough and sustained long enough, it can produce a recession. Only w hen the decline in spending (motivated by the monetary disequilibrium asso ciated with the reduction in money growth) has been identified as permanent will producers reduce their prices and increase production back to “normal” levels. Thus, the impact of the monetary contraction on output eventually vanishes, and, in the long run, only the rate of inflation is affected by a sustained reduction in money growth.9 The potential usefulness of monetary targeting for economic policy purposes is evident from this dis 6Batten, “ M oney Growth Stability and Inflation;” and also Selden, “ Inflation and Monetary Growth.” 7This article discusses only the impact of changes in money growth on the real output of the economy. It does not investigate the impact of money growth changes on financial markets. 8Poole, “The Relationship of Monetary Decelerations to Business Cycle Peaks.” See also Economic Report o f the President (Gov ernm ent Printing Office, 1982), pp. 192-96, for another use of the theory presented here. 9The empirical problem here, of course, is dating the “long run.” Digitized for 16 FRASER MAY 1982 cussion. First, in the long run, perm anent changes in the rate of money growth are reflected by equivalent changes in the rate of inflation, other things equal. Second, if short-run money growth is volatile, the growth of real output and em ploym ent w ill be similarly volatile. In other words, sufficiently un stable money growth in the short run, that is, a re duction in money growth relative to its trend rate, may cause recessions. Consequently, minimizing the variability of short-run money growth appears to be essential in establishing a stable, non-inflationary environm ent for economic growth. SHORT-RUN MONEY GROWTH AND ECONOMIC ACTIVITY: THE EVIDENCE We now investigate the validity of the conceptual analysis presented in the preceding section. To examine the relationship betw een short-run fluctu ations in money growth and real economic activity, a sample of four industrialized countries was selected: the U nited States, the U nited Kingdom, W est Germany and Italy. Moreover, to make the results of the analysis directly comparable, the narrow defini tion of money for each country is used.10 To illustrate the relationship betw een short-run money growth and real output growth, charts for each country are presented for the period 1973 to the present.11 These charts depict the deviations of short-run money growth from its trend, m easured by subtracting the 20-quarter moving average growth rate of money from its two-quarter moving average growth rate. In addition, the quarter-to-quarter, compounded annual rate of growth of real GNP is 10The M l definition is used throughout. It should be noted that even though the narrow definition is used, it is not the variable used by all the central banks in their policy deliberations. The countries and their respective monetary target(s) are: United States (M l, M2), U nited Kingdom (Sterling M3), Germany (Central Bank Money Stock) and Italy (Total Domestic Credit). "T h e period since 1973 is used for two reasons. First, it is char acterized as a flexible exchange rate period, a condition giving each country more control over its own domestic money supply and, hence, economic activity than in a fixed exchange rate period. W hile the analysis also applies to a fixed exchange rate period, economic activity of open economies during such a period may merely reflect economic activity in the United States. Consequently, we chose the post-1973 period because we are concerned with examining the impact of changes in short-run money growth that are m otivated by changes in factors indigenous to the dom estic economy. Second, this period covers the tim e in w h ich each co u n try ’s cen tral bank announced a monetary aggregate policy target. Prior to 1973, announced money supply growth targets were not universal. MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS plotted. Periods in which real output growth was negative for two consecutive quarters or more are denoted by the shaded areas; these designate periods of recession in these countries.12 The individual charts reveal that there is a com mon relationship betw een sharp reductions in the short-run growth of money (the two-quarter moving average) relative to its trend (the 20-quarter moving average) and real economic activity.13 Despite the wide differences among these countries in terms of their financial structures, regulations and monetary policy objectives, the relationship betw een shortrun deviations in their money growth from trend and declines in their real economic activity is quite similar. To see this more clearly, we briefly examine the historical record of each country in our sample. The United States The chart for the United States reveals three re cessions since 1973. As predicted by the theoretical discussion, each recession was preceded by a sharp slowing in short-run money growth. Prior to the 1974 recession, for example, short-run money growth fell from slightly over 2 percentage points above trend to about 2 percentage points below trend, a change that is mirrored in the reduction in real GNP growth in 1973. While one may argue that the recession of 1974 was supply-oriented — a reaction to the unexpected OPEC oil shock — the chart indicates that the depth and breadth of the downturn was exacerbated by short-run money growth well below trend in late 1974.14 12The recessions in the United States are those defined by the National Bureau of Economic Research. Since recessions are not formally defined in the other countries in the sample, the generally accepted rule of thumb is that a recession is indicated by at least two consecutive quarters of declining real GNP. 13The purpose of this article is not to employ statistical methods to investigate rigorously the money/real output relationship in those countries. Instead, we are simply applying the general implications of the research that has been conducted for the United States to an analysis of these countries, as a first attempt to see if empirical relationships similar to those in the United States can be found. Obviously, the timing of the money growth/ real output relationship may be different across countries and, in fact, the 20-quarter and two-quarter distinctions may not be completely applicable to all. These results, however, appear to be quite robust and, consequently, we shift to the unconvinced reader the obligation of an alternative interpretation of the data. 14The oil price shocks of 1973-74 and 1979-80 resulted in dis similar monetary growth rates in the United States. For a dis cussion of this, see R. W. Hafer, “The Impact of Energy Prices and Money Growth on Five Industrial Countries,” this Review (March 1981), pp. 19-26. The most recent downturns in economic activity also are associated with declines in short-run money growth. For example, prior to the onset of the II/ 1980-III/1980 recession, money growth fell from about 3 percentage points above trend to over 4 percentage points below trend. Although money growth’s sharp rebound during late 1980 helped produce the turnaround in real GNP growth in early 1981, the equally dram atic dow nturn in m oney growth relative to trend during 1981 has precipitated yet another reduction in real economic activity. Indeed, since 1/1980, short-run money growth has fallen short of trend almost 90 percent of the time, and real GNP growth has been negative almost 40 percent of the time. Clearly, the dramatic slowing in short-run money growth relative to its long-run trend and the increase in its volatility during the past two years have been associated with substantial reduc tions in real economic activity over this period. The United Kingdom The accompanying chart indicates that the United Kingdom has experienced a num ber of “recessions” during the brief period studied. Of the six recessions shown, all but one were preceded by sharp reduc tions in short-run money growth. For instance, prior to the IV/1973-I/1974 downturn, money growth fell from about 5 percentage points above trend to more than 10 percentage points below trend, a reversal of about 15 percentage points in less than one year. Likewise, the 1/1977-11/1977 recession came on the heels of a drop in money growth to more than 5 percentage points below its trend. The period since late 1978 is interesting because it reveals the effect on the economy of a sustained reduction in short-run m oney growth below its trend. Although money growth did not dip far below trend prior to the IV/1978-I/1979 recession, shortrun money growth fell from over 15 percentage points above trend in IV/1977 to its trend level in only three quarters, a change that is associated with the drop in real GNP growth from IV/1977 to 1/1979. Also, the impact of the nature of the money growth decline during the period from IV/1977 to 1/1981 is reflected by relatively stagnant output growth during this period. Finally, the IV/1974-III/1975 recession represents an anomaly to the theory. The recession was not preceded by a downturn in short-run money growth relative to its trend; instead, money growth in creased faster than its trend rate prior to this reces17 FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 M o n e y an d O utput G ro w th in Selected Countries United States Percent Percent 1973 Li C o m p o u n d e d 1974 annual [2 T w o - q u a r t e r m o v i n g Shaded 18 areas 1975 rates 1976 1977 of change. average money g row th 1978 1979 Source: rate m inus the represent p erio d s o f e c o n o m ic dow nturn. 2 0 -q u a r te r m oving 1980 1981 I n t e r n a t i o n a l F i n a n c i a l S ta t is t ic s average m oney grow th rate. FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 M oney and Output Grow th in Selected Countries West Germany Percent 1973 |J_ C o m p o u n d e d 1974 [2 T w o - q u a r t e r m o v i n g Shaded areas 1975 1976 1977 a n n u a l rates o f c h a n g e . average m oney g row th 1978 1979 Source: rate m inus th e 2 0 - q u a r t e r m oving 1980 1981 I n t e r n a t i o n a l F i n a n c i a l S t atis tic s average m oney grow th rate. represent p e rio d s o f e c o n o m ic d o w n tu rn . 19 sion. This may have been an attempt to use monetary policy to offset, at least partially, the dislocations created by the OPEC oil shock that lowered the growth of real GNP. Interestingly, the U.K. response to the 1978-79 OPEC oil shock was to decrease the short-mn growth of the money stock, as shown in the chart.15 growth relative to its trend rate. This pattern is es pecially evident for the II/1974-II/1975 and 11/1980III/1980 recessions. CONCLUSIONS The evidence presented here suggests that sizable and sustained reductions in short-run money growth West Germany below its trend rate portend declines in the growth of real GNP. Of the 14 recessions in the four countries The chart for West Germany again supports the examined, only one — the IV/1974-III/1975 reces theoretical discussion. Each of the two recessions is sion in the United Kingdom — was not preceded by a preceded by periods of money growth below trend. substantial decline in short-run m oney growth. Although the timing is different for each episode, the Moreover, in only one instance — the III/1975-IV/ reaction of the real economy to declines in short- 1976 period for W est Germany — did short-run run money growth is clear and consistent. money growth fall substantially below trend without W est Germ any also presents a case in which a recession following. In that instance, however, money growth fell below trend and no technical West German real GNP growth fell from about 10 recession occurred. From III/1975 to IV/1976, percent to zero, a result consistent with the theo money growth fell from about 7 percentage points retical discussion. above trend to about 5 percentage points below Thus, the evidence indicates that policymakers trend. Although no recession followed, the level of should be concerned with short-run fluctuations in real GNP growth fell sharply as the theory predicts: the growth of the m oney supply relative to its the growth rate of real GNP fell from about 10 per trend.16 If this evidence is at all useful, it dem on cent in IV/1975 to zero in 11/1977. Thus, while strates how robust the relationship betw een money technically no recession followed the decline in growth and real economic activity is over the short m oney growth, real GNP growth was curtailed run. Coupled with previously reported research sharply, an example of a “growth recession.” indicating a direct, positive link betw een longerterm money growth and inflation, the em pirical evidence favors a steady growth of the money stock Italy in both the short and long run as the most effective The relationship between real GNP growth and means of achieving economic stability. money growth relative to trend in Italy, once again, is consistent with theoretical expectations. Of the three recessions since 1973, each was preceded by a 16This evidence contradicts the recent claim that “the [money growth] volatility issue itself is a hoax. No one as yet has been period of sharp reductions in short-run m oney 15Hafer, “Impact of Energy Prices and Money Growth.” 20 able to demonstrate that the reported volatility in money has any impact on either the pace of economic activity or inflation.” Aubrey G. Lanston & Co., Inc., Newsletter (March 22, 1982). Money, Credit and Velocity MACK OTT S h a k e s p e a re : “ N e ith e r b o rro w e r, n o r a le n d e r b e ” (H am let, I, iii, 75, P olo nius to L aertes) G o eth e: “ L e t us live in as sm all a circle as w e w ill, w e are e ith e r deb to rs or cred itors b efo re w e h av e h ad tim e to look aro u n d .” (.E lective A ffinities, Bk. II, C h. 4) I^ .E C E N T L Y , many critics of monetary policy, and some monetary policymakers as well, have as serted that the links betw een monetary aggregates and national economic policy variables—that is, GNP, inflation and real economic growth—have been severed by a host of financial and credit market innovations. If these critics are correct, then a monetary policy based on targeting the growth of a m onetary aggregate would becom e increasingly ineffective and inappropriate, as credit arrange ments are substituted for monetary payments.1 The puipose of this article is to provide a theo retical framework in which to assess these claims The author, an associate professor of economics at The Pennsyl vania State University, is a visiting scholar at the Federal Reserve Bank of St. Louis. •For examples, see Neil G. Berkman, “Abandoning Monetary Aggregates,” in Controlling Monetary Aggregates III, pro ceedings of a conference sponsored by the Federal Reserve Bank of Boston (October 1980), pp. 76-100; Benjamin M. Friedman, “The Relative Stability of Money and Credit ‘Velocities’ in the United States: Evidence and Some Speculations,” NBER Work ing Paper No. 645 (March 1981); Anthony M. Solomon, “Finan cial Innovation and Monetary Policy” (remarks before the Joint Luncheon of the American Economic and American Finance Associations, Decem ber 28, 1981); James M. Tobin, “ Inflation,” in E ncyclopedia oj Economies, D ouglas G reenw ald, ed. (McGraw-Hill, 1982), pp. 510-23. For the contrary position—i.e., that monetary policy should be undertaken through effective control ofa monetary aggregate—see Milton Friedm an, “Mone tary Policy: Theory and Practice,"Journal o f Money, Credit and Banking (February 1982), pp. 98-118; and Allan H. Meltzer, Robert H. Rasche, Stephen H. Axilrod, and Peter Sternlight, “Money, Credit, and Banking Debate: Is the Federal Reserve’s M onetary Control Policy M isdirected?” journal o f Money, Credit and Banking (February 1982), pp. 119-47. and to examine empirical evidence bearing on their purported policy consequences. The analysis pre sented in this article does not support the critics’ assertions. This conclusion rests on two arguments. F irst, the relation betw een m oney and credit requires that the amount of credit granted match the anticipated amount of money that will be available to settle the debt when it comes due. Thus, regulating the rate of monetary growth, which in turn regulates the anticipated future quantity of money, deter mines the amount of credit and the conditions under which it is granted. This constraining influence of monetary growth on credit would be undone only if the relation betw een money and income growth departed from its historical pattern. That it has not is the second argument: the em pirical evidence on velocity, the relation between money growth and income growth, reveals no sig nificant change during the last two years from its previous history. C onsequently, despite recent claims to the contrary, the growth of the monetary aggregates is still reliably linked to the economic variables of interest to policymakers. MONEY, CREDIT AND EXCHANGE In contemporary societies, the exchange of goods is indirect. The purchase or sale of goods, w hether in organized markets or through informal arrange ments, is almost always in exchange for money or m oney-denom inated prom ises. D irect bartering of one good for another is either nonexistent or unimportant. The reason for this is at once obvious, yet theo retically challenging to elucidate. In the intro duction to his book, The Theory o f Money, Jurg Niehans observes: Economists (and laymen) have always felt that the use of a medium of exchange increases the efficiency 21 FEDERAL RESERVE BANK OF ST. LOUIS o f an econom y. T h e gain w as u su ally co n sid ered to b e large. It has bo th q u a litativ e an d q u a n titiv e aspects. T h e q u a litativ e aspects ap p e a r w h en m o n etary ex ch ange is co m p ared w ith b arter. C lassical an d n e o classical econo m ists w ere g raph ic in d e scrib in g th e “ d o u b le co in cid e n ce o f w an ts” of th e h u n g ry tailo r an d th e sh iv erin g b a k er w h ich w o u ld be n ecessary for an ex ch an ge in a b a rte r econom y an d th e narrow lim itation s it im po ses on th e d iv isio n o f labor. T he u se o f m o n ey w o u ld in c re ase w e lfa re by fre ein g ex chan ge from th e shack les of th e d o u b le co in ci d e n ce o f w an ts.2 Robert Clower succinctly summarized the results of these advantages as imposing a constraint on*the exchange process: “Money buys goods and goods buy money; but goods do not buy goods.”3 In other words, it is the nature of a system of monetary ex change to replace the cumbersome barter exchange of goods with two non-synchronized m onetized exchanges: a sale of goods for money and a later purchase of goods by money. This exchange attrib ute in turn has implications for both the appropriate definition of money and for the monetary arrange ments used in exchange.4 First, the period betw een the sale of one good for money and the subsequent purchase of another good may be long enough or predictable enough to allow the interim holding of funds in a non-transaction account. This implies that the appropriate monetary aggregate may not be narrowly defined money (i.e., M l), but a broader aggregate (e.g., M2) which con2Jurg Niehans, The Theory o f Money (Johns Hopkins University Press, 1978), p. 2. 3Robert W. Clower, “A Reconsideration of the Microfoundations of Monetary Theory,” Western Economic Journal (March 1967), p. 6. Also, see Karl Brunner and Allan H. Meltzer, “The Uses of Money: Money in the Theory of An Exchange Economy,” Amer ican Economic Review (Decem ber 1971), pp. 784-805. 4Milton Friedman and Anna Schwartz described this attribute as “the separation of the act of purchase from the act of sale,” but criticized the medium of exchange approach as being too narrow to capture the essential nature of money: In order for the act of purchase to be separated from the act of sale, there m ust indeed be something that will be generally accepted in payment—this is the feature emphasized in the “medium of ex change” approach. But also there must be something that can serve as a temporary abode of purchasing power, in which the seller holds the proceeds in the interim betw een sale and subsequent purchase or from which the buyer can extract the general purchasing power with which he pays for what he buys. . . . Both features are necessary to perm it the act of purchase to be separated from the act of sale, but the ‘something’ that is generally accepted in paym ent need not coincide with the ‘something’ that serves as a temporary abode of purchasing power; the latter may include the former and more besides. Milton Friedman and Anna Schwartz, Monetary Statistics o f the United States: Estimates, Sources, Methods (NBER, 1970), pp. 106-07. Digitized for 22FRASER MAY 1982 tains w hat M ilton F riedm an characterizes as “temporary abodes of purchasing power” that are readily convertible at low cost into an exchange medium.5 Second, if the purchase of the good to be financed by the proceeds from the sale of another good pre cedes the sale of that other good, then the anticipated future sale proceeds may be used to mediate the earlier purchase. Of course, an exchange arrange m entlike this is a familiar part of modern economies; such purchases are said to be made “on credit.” Credit is granted by sellers or other third party lenders to buyers precisely on the basis of the buyer’s anticipated future receipts (with the lender concurring) and, of course, is measured in monetary units. As a consequence, credit is as much of a medium of exchange as is money.6 While both credit and money are used to mediate exchange, they are obviously different entities. The quantity of money circulating in an economy is a stock; its units are used repeatedly in a sequence of exchanges. Credit, on the other hand, is a flow and is transaction-specific; it can only mediate the trans action for which it was created.7 5Two goods that are perfect substitutes are economically the same good. If two durable goods are costlessly transformable, one into the other, then they are perfect substitutes in an inventory. On this criterion, if the cost of transferring funds from a savings account to a demand account or to currency were zero, then, clearly, savings accounts would be economically indistinguish able from demand accounts or currency and would be exchange media. Conversely, if the costs of transfer were prohibitively large, savings accounts would not be a close substitute for demand deposits. Hence, as Friedm an and Schwartz argue, the question of what money is cannot be settled on an a priori basis, but is an empirical question which, in part, depends on how costly inter-deposit transfers are. 6This observation has led Clower and others to argue that some measure of credit availability or line of credit be included in the policy relevant concept of money: . . for most practical pur poses, ‘money’ should be considered to include trade credit as well as currency and demand deposits.” Robert W. Clower, “Theoretical Foundations of Monetary Policy,” in Monetary Theory and Monetary Policy in the 1970s, George Clayton, John C. Gilbert and Robert Sedgwick, eds. (Oxford University Press, 1971), p. 18. See also Arthur B. Laffer,“Trade Credit and the Money Market” (March 1970), pp. 239-67; and J. Stephen Ferris, “A Transactions Theory of Trade Credit Use,” Quarterly Journal o f Economics (May 1981), pp. 243-70. 7It has been argued that credit is not an exchange medium, but m erely an arrangem ent that raises the velocity of money. Ironically, the same argum ent was once used against including demand deposits in money. As Friedm an and Schwartz point out, much of the 19th century debate betw een the banking and cur rency schools centered on whether bank notes and deposits were money or merely “means of raising the velocity of bank vault cash but not as adding to the quantity of money.” Friedm an and Schwartz, Monetary Statistics o f the United. States, p. 95. FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 BOTH CREDIT AND MONEY ARE NECESSARY FOR MONETIZED EXCHANGE receive money or a promise to deliver money at a specified future time. If only commodities were exchanged, it would be possible always to use money alone and never incur The epigraphs from Shakespeare and Goethe rep a debt. Services, however, by their very nature, resent conflicting views on the desirability and cannot be exchanged without one party, either seller inevitability of credit; to wit, while money and credit or buyer, extending credit to the other. Hence, a law are alternative exchange media, would either be attem pting to enforce Shakespeare’s adm onition sufficient to m ediate all exchanges w ithout the would not prohibit the sale of apples, automobiles or other? Could any of us, as Polonius suggests, avoid clothing; it would, however, prohibit the renting of a credit transactions completely? Conversely, could house, the purchase of a ski-lift ticket or the hiring of credit function as we know it without a monetary labor. In each of these latter examples, the trans framework? Not surprisingly, the answer to both action entails the exchange of money before or after questions is no. Hence, the advice of Polonius is as the completion of the activity with, necessarily, a fatuous as the character offering it. Both credit and concomitant issuance of credit.9 money are necessary in the exchange process, each fulfilling functions that the other could not. Thus, Goethe was right; each of us inevitably engages credit transactions every day. For ex In order to establish this com plem entarity of ample, weinextend to our employer and receive money and credit, consider the exchange process as it from our electriccredit utility. If services of any form are a contractual arrangem ent betw een buyer and to be exchanged, credit must be offered either by the seller.8 Under this characterization, the exchange seller—as in the typical employm arrangem ent and the settlem ent of the contract need not coincide where wages are received after theentservices have in time so that either credit or money can mediate an been delivered—or by the buyer—as in entertain exchange. In the case of a credit transaction, at the activities where the purchase of a ticket pre time of the exchange the buyer incurs a contractual ment cedes the concert, game or movie.10 liability tor ^.subsequent settlement toclearhis debt. Using this contractual approach, we can now dem credit is inextricably bound up with sell onstrate why Goethe’s claim of the inevitability of ingClearly, services in a monetized economy in order to credit in any society is correct. avoid the problem of making an indefinitely large num ber of infinitesimal cash payments. Yet money and credit are simply alternative means of lowering Credit and the Exchange o f Services the cost of exchanging goods relative to a primitive barter system. Thus, even some commodities might Two types of goods are voluntarily offered for be too costly to exchange in customary ways if credit exchange in markets: commodities and services. By were ruled out (e.g., hom e-delivered newspapers or definition, a commodity is a tangible physical entity raw materials purchased by firms).11 not intrinsically dependent on time (e.g., an apple, a phonograph record or an automobile), while a ser 9Note that this would also rule out the existence of any firm other vice is an activity or process that is intangible and than owner-operated producers of commodities. intrinsically sensible only with the passage of time 10Barter exchange of services is conceivable as suggested in the maxim, “You scratch my back and I’ll scratch yours.” Yet, even (e.g., a gardener’s chores, a concert or a taxi ride). In a here, credit sneaks in unless the exchange is simultaneous. m onetized economy, sellers of either type of good 8Under Anglo-American law, an enforceable contract must have three elements: (1) There must be an offer; (2) There must be an acceptance precisely matching the offer— else it is a counter-offer; (3) There m ust be consideration—i.e., the offeror or acceptor must make some performance that would be a detrim ent to him if the agreem ent were not fulfilled. See “Contract” and other referenced citations thereunder in Henry Campbell Black, Black’s Law Dictionary, 5th ed. (West Publishing Co., 1979), pp. 291-94, 277. “ Credit extended by sellers of raw materials is an especially important example. If credit were not extended to producers, either deliveries would have to be made more frequently (in smaller lots) to match producers’ cash flow from sale s of output, or the material-using firms would have to tie up more of their capital in raw material inventories and, hence, less in the capital to process these materials. Alternatively, firms would find it more advantageous to be vertically integrated—i.e., to own their suppliers—than to acquire these materials from other firms. See “Credit Allocation: An Exercise in the Futility of Controls” (Citibank Economics Dept., 1979), p. 40. In any case—more frequent delivery, larger inventories in capital, or more vertical integration—resources would be less productively allocated than when credit is extended. 23 FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 The Relationship Between Money and Credit exchange system functions.12 From this vantage, they have docum ented that, in moving from barter to indirect exchange, the m ost useful function of prim itive monies is the com m only-agreed-upon Money and credit are both substitutes and com valuation unit.13 plements in the exchange process. On the individual level, money and credit are potential substitutes for Finally, credit mediation of exchange is facilitated mediating any exchange of commodities. On the by the universal acceptability of money as a means of societal level, money and credit are complements in settlem ent—the standard of deferred payment func the exchange process; each provides a function tion. All credit contracts can be settled (directly or necessary to some exchanges that the other cannot through civil courts) by means of a money payment; fulfill. In fact, credit is a more general medium of that is, money is legal tender in our economy. This exchange than money in that it facilitates exchange general agreem ent on the m eans of settlem ent involving time—both in permitting the sale of ser makes credit less costly to extend, thereby increas vices and in perm itting differing delivery dates in ing its availability for exchange mediation. A decen exchanges of commodities; money without credit tralized use of credit requires that individuals and can act as the exchange medium only for a com firms be able to clear their debts individually (i.e., modity. Yet, money is likewise crucial to the func pairwise) with some mutually agreeable means of tioning of credit through its role as the primary settlement; without such agreem ent on the means of means of settlement. settlement, credit clearing would require a costly centralized M onetary theorists generally have agreed that “barter club.”system of record-keeping much like a money in modern economies is anything that fulfills all of the following functions: 1. 2. 3. 4. Medium of exchange, Store of value, Unit of account, Standard of deferred payment. THE RELATION OF CREDIT EXPANSION TO MONETARY POLICY Credit is not money, but the promise of future money to the lender in return for the temporary use Most economists have argued that the crucial char of current purchasing power—goods or money— acteristic in this list is its functioning as a medium of extended to the borrower. Two errors that violate exchange. Typically, they have argued that any this logic occur every day in the financial press: durable good can fulfill the remaining three func tions, but only money can fulfill the first. 12See Philip Grierson, “The Origins of Money,” Research in However, we have seen that credit also fulfills the Economic Anthropology, Vol. 1 (JAI Press, Inc., 1978), espe cially pp. 9-12 for evidence on the importance of standard of medium of exchange function. Credit in our dis value in explaining early monetary systems. See also George cussion has taken a special form—namely, credit Dalton, “Primitive Money,” American Anthropologist (1965:1), measured in units of money and, implicitly, with the pp. 44-65; and Denise Schmandt-Besserat, “The Earliest Pre cursors of W riting,” Scientific American (June 1978), pp.50-59. deferred payment to be made in units of money. In exchange systems with money and credit acting as 13In this context, it is ironic and revealing that contemporary “barter clubs” use dollars as the unit of account but not as an exchange m edia, the other three functions in exchange medium. Consider these descriptions from “As Barter money’s repertoire take on an importance not ap Boom Keeps Growing,” U.S. News and World Report (Sep tem ber 21, 1981), p. 58: parent in the conceptual monetary exchange models A participant lists items for sale, and they are advertised to the other without credit. m embers. If a listed item is sold, the former owner is issued trade credits—sometimes called trade dollars. These credits can later be used to purchase goods and services from other m em bers. . . “We W ithout agreement on the unit of account, credit don't make outright trades; we perform a banking function. . .” transactions would have all the disadvantages of This is also the method by which every “barter exchange” in the article appears to be organized: barter except simultaneity. Anthropologists, in con profiled Besides credits, most barter exchanges issue barter cards that can trast to economists, have placed more emphasis on be used for purchases at participating merchants. Through the Trade Bank International exchange, a Memphis dentist began receiving the unit of account function because their focus is on customers who used their barter cards for dental work. W ithin a how a monetized exchange system evolves from a year’s time, the dentist accum ulated enough trade dollars to buy carpeting for his office, install new signs and pay for flying lessons. barter system rather than how an extant monetized Digitized for 24FRASER FEDERAL RESERVE BANK OF ST. LOUIS 1. R eferrin g to th e in te re st rate as th e p rice of m oney; 2. Id en tify in g av ailab le cre d it as m o n ey .14 The first error is so commonplace that its repeti tion makes it seem valid; nonetheless, the interest rate is not the price but the rental rate for a dollar or, properly expressed, any other good. The price of a dollar is a dollar’s worth of something—certainly more than a mere percentage of a dollar. No one would refer to the rental rate at Hertz as the price of a new Ford, or to the rent on a house as its purchase price, but the confusion of interest on credit with the price of money has become so common that the error no longer jangles our sensibilities. Yet the distinc tion is not only obvious but as important for money and credit as for owned and rented automobiles. Similarly, the second error, referring to available credit as money, also escapes rebuke through fre quent use. The annual total of credit extensions is many times larger than the year-to-year increases in either M 1 or M2, and, in recent years, has been larger than the stock of M l. Considering the consumer sector (which accounts for over 60 percent of na tional income), a large share of credit extensions, almost two-thirds, are by institutions other than commercial banks and, therefore, do not entail monetary expansion. Considering only installment consumer credit, about 40 percent of such credit is extended by non-depository institutions with about 20 percent being extended by retailers and gasoline companies. In these retail extensions, money affects the transaction only through the anticipated mone tary settlem ent.15 These errors are substantive for they focus the public’s evaluation of monetary policy on regulating the flow of credit instead of controlling the growth of the stock of money. Controlling the rate of growth of the money stock in a predictable fashion enhances the predictability of the future availability of the means of settlement. This regularity of monetary expansion makes for better-informed, intertemporal decision-making and, therefore, contributes to the stabilization of credit markets. When non-monetary shocks occur, the predictable availability of quan tities of m oney in the system allow s m arket- MAY 1982 determ ined signals—that is, interest rate changes— to allocate credit efficiently to adjust to the shocks. Conversely, attempting to control interest rates requires the monetary authority, in effect, to allocate credit at the cost of making the growth rate of mone tary expansion less predictable; since this makes the real future value of the means of settlem ent more variable, credit transactions become riskier, and credit markets less stable. W hen non-m onetary shocks occur, the less predictable quantities of means of settlem ent with relatively fixed interest rates impede market signals from efficiently allo cating credit. Since both money and credit are exchange media, the key to effectively controlling either or both of them must be first to isolate their interconnections and mutual dependencies. This article has argued that credit is unavoidable and that a money means of settlem ent is necessary for a decentralized credit system. What it now addresses is how monetary and credit expansion relate to each other and how both of these relate to national income. Credit and Money Creation In contemporary market economies, the money supply grows through two types of credit transac tions: the central bank creating deposits (money) and bank reserves by buying government securities, and depository institutions creating deposits (money) from increased reserves by granting loans.16 Of course, not all credit extensions entail mone tary expansion. There are three distinct sources of credit extension: (1) bank and non-bank depository institutions (commercial banks, savings and loans, credit unions, m utual savings banks); (2) non depository financial interm ediaries (finance com- 16In other words, modern monetary systems have a fiat base— literally money by decree—with depository institutions, acting as fiduciaries, creating obligations against them selves with the fiat base acting in part as reserves. The decree appears on the currency notes: “This note is legal tender for all debts, public and private.” W hile no individual could refuse to accept such money for debt repayment, exchange contracts could easily be 14Recent examples are (1) "The price of money—the interest to thwart its use in everyday commerce. However, a rate—reflects, therefore the interaction of millions of partici composed forceful explanation as to why money is accepted is that the pants in the credit market. . H em y Kaufman, Washington federal government requires it as payment for tax liabilities. Post, Septem ber 23, 1981; (2) “As long as the Federal Reserve Anticipation of the need to clear this debt creates a demand for Board maintains its current course, credit—or money available the pure fiat dollar, guaranteeing its exchange value. See Abba to lend—will remain tight." Harry B. Guis, Christian Science P. Lerner, “Money as a Creature of the State,” American Eco Monitor, Septem ber 21, 1981. nomic Review (May 1947), pp. 312-17; and Ross M. Starr, “The 15Souree: Federal Reserve Bulletin (January 1982), Tables 1.21, Price of Money in a Pure Exchange Monetary Economy with Taxation,” Econometrica (January 1974), pp. 45-54. 1.56, 1.57, 1.58, 2.16. 25 FEDERAL RESERVE BANK OF ST. LOUIS panies, investm ent banks, brokerages, insurance companies); and (3) sellers of goods (retail and trade credit). In the first case, a depository institution lends money to a borrower who in turn uses these funds to purchase goods or repay debts; the credit extension entails monetary expansion of purchasing pow er because it consists of checkable deposit expansion. During the last three decades, loans by such depository institutions have accounted for between 35 and 50 percent of the annual total of credit market funds extended to the non-financial sector.17 Alternatively put, more than half of the credit extended annually in U.S. financial markets does not entail deposit expansion. In the second case, a non-depository institution (e.g., a consumer finance company) issues the credit or buys the accounts receivable of a credit-issuing seller. The latter method of credit extension is called factoring, and non-depository institutions fund this activity by either selling debentures directly or by acting as an agent for a depository institution. Under either method, the extension of credit does not entail an expansion of deposits but a reallocation of exist ing deposit holdings.18 Finally, in case three, credit may be extended directly by the seller of goods and held as accounts receivable. Often this credit is financed by the sale of commercial paper issued by the seller/credit-issuer (e.g., firms with their own financial subsidiaries such as Sears or General Motors). In these instances, whether the firm holds its own accounts receivable, factors its accounts receivable or sells commercial paper, the extended credit represents an increase in purchasing power not created by checkable deposit expansion. 17Source: Board of Governors, Federal Beserve System. Of course, this credit expansion is limited by bank reserves under a given set of reserve requirements and is consequently directly controlled by the monetary authority. For this form of credit, additional credit control authority would be superfluous. This case also covers bank credit card usage since credit issued by a sellerto abuyer against a bank card becomes a demand deposit increm ent as soon as the seller/credit-issuer submits the credit invoice to the agent bank. In both types of credit extension, direct or credit card, a depository institution creates money matching the extended credit. 18If a depository institution issues a loan to a creditor using the accounts or debt as collateral, then the credit extension has the same one-for-one expansion of deposits as if the loan were directly placed. From 1977 through 1980, the percentage of installm ent loans by non-depository institutions was .39, .37, .40, .45 respectively; source: Federal Reserve Bulletin (Sep tem ber 1981), table 1.57. A breakdown for non-installm ent credit has not been present in the Bulletin since 1975, but from 1965 tol975, commercial banks extended only about one-third of single-payment non-installment loans. Digitized for 26 FRASER MAY 1982 In the second and third cases, credit extensions substitute for monetary mediation, while, in the first case, a dollar of money is created by each dollar of credit extended. Thus, for the case of loans by de posit creation, credit expansion has no apparent impact on the relation betw een the narrowly defined money supply and income since M l and credit move together; however, in the latter two cases, credit substitutes for money w hich apparently w ould change the ratio of income to money supply. Yet, to the extent that credit arrangements in creasingly provide as ready a source of purchasing power as narrowly defined money (M l), the ap pearances of these cases are somewhat misleading. There should be an incentive to reduce M l holdings and to increase the non-M l portion of M2 holdings. For example, given the rising acceptability of bank credit cards—about 30 percent of U.S. retail and service establishments accepted them in 1972, ap proximately 50 percent in 1981—the utility of hold ing a reserve of currency or dem and deposit balances in order to m ediate unforeseen or spur-of-themoment purchases has been significantly reduced for consumers.19 Still, to clear the short-term credit card debt at m onth’s end, a ready source of funds to shift to demand or other checkable deposits remains necessary. Consequently, even if the proportions of cash and credit purchases were constant, given the increasing acceptability of credit as an exchange medium, it would not be surprising to see consumer holdings of dem and deposits decline relative to purchases (i.e., to have had a rising velocity). IMPLICATIONS OF RISING CREDIT FOR MONETARY GROWTH AND ECONOMIC ACTIVITY If all credit extensions represented monetary ex pansion, then controlling monetary growth would control credit. The same constraint that limiting reserves imposes on deposit expansion also limits 19The total num ber of merchant (i.e., retail and service) estab lishments in the United States rose less than 2 percent per year during the 1960s and 1970s, while the num ber of merchant outlets accepting MasterCard and VISA rose at over 8 percent and 9 percent per year, respectively. (Sources: Statistical Abstract o f the United States, 1980 (U.S. Dept, of Commerce, Bureau of the Census), 101st ed., and data supplied by VISA and MasterCard). To estimate the percentage of merchants accept ing bank cards, we estimated total merchants for 1981 by ex trapolating the 2 percent annual growth rate from 1977 forward. This was then divided into the num ber of merchant outlets that accept MasterCard. MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS credit extensions, and inflation policy can properly focus on controlling m oney growth, leaving the market to allocate credit. As we have seen, however, depository institutions account for less than half of the credit annually extended in the United States. C onsequently, m ight not the purchasing power created by non-deposit credit extensions render monetary policies undertaken through control of monetary growth rates ineffective? The answer is no: money in its role as the means of settlem ent constrains non-depository as well as depository credit. If an increase in the use of credit alters the moneyincome relationship, the income velocity of money will rise. That is, if a larger share of transactions by households or firms can be m ediated by credit, those households and firms, relative to their incomes, will plan to hold less M l and more of other assets, includ ing non-M l deposits. As this substitution occurs, the ratio of nominal income to M l (velocity) will rise. W hether such a change will occur for all monetary aggregates, narrow and broad, depends on the extent to which substitutions of non-M l assets for M l com prise deposits included in other m onetary aggregates.20 Velocity, v, which is the ratio of nominal gross national product, Y, to money, M, (.3) v = P + y - M, from equation 1, w here' indicates the annualized growth rate of each variable. From equation 3, we obtain (4) P = v — y + M, which shows the significance of velocity for m one tary policy with the inflation rate, P, as its target. As is obvious from equation 4, if velocity is con stant (v = o), then the inflation rate will be equal to the difference betw een the growth rates of real output, y, and money, M; if v is relatively constant but non-zero, then inflation would be the difference betw een the growth rates of money and real output plus that of velocity. If v does not depend on M or y, then equation 4 im plies that if v is simply pre dictable, even if not constant, then controlling the money supply is tantamount to controlling inflation.22 This interpretation abstracts from variations in real output, but, to the extent that fluctuations in the growth rate of money exacerbate such variations, setting a constant growth rate of money reduces that source of disturbance. Non-monetary disturbances to real output growth (e.g., the OPEC oil embargo), of course, may cause inflation to deviate from its anticipated path, but over longer periods of time, a steady growth rate of money will smooth real income growth as well as facilitate inflation predictability. is the rationale for a policy of targeting on the measures the turnover rate of the average dollar in This growth of money and why its effectiveness M, that is, how many times a dollar was used in a dependsrate upon the predictability of velocity.23 transaction involving Y during the year.21 Express Assessing the predictability of a variable involves ing nominal income as the product of the price level, two separate evaluations: point forecasts and vari P, and real output, y, ability. The shorter the time period considered, the (2) Y = Py, relatively more important is the latter characteristic; we obtain an equation for the growth rate of velocity, that is, while a short-run forecast of a variable may rarely be precise, if that variable does not fluctuate wildly in a fashion out of keeping with its history, “ Essentially, this is again Friedm an’s argument that the defini then describing it as predictable is sensible. tion of money is not an a priori hut an empirical issue. “The selection [of money’s definition] is to be regarded as an em pirical hypothesis asserting that a particular definition will be most convenient for a particular purpose because the magnitude based on that definition bears a more consistent and regular relation to other variables relevant for the purpose than do alternative magnitudes of the same general class. . . . It may well be that the specific meaning it is most convenient to attach to the term money differs for different periods, under different institutional arrangements, or for different purposes.” Fried man and Schwartz, Monetary Statistics o f the United States, p. 91. 21The reciprocal of velocity measures the average holding period of a dollar, how long betw een final income transactions. This period is germane to the Friedm an notion of temporary abode of purchasing power. 22Note that for policy purposes we need not know precisely why the growth rate of velocity is predictable; for the purpose of formulating an inflation policy through control of a monetary aggregate, it is sufficient that it is predictable. 23For a more detailed statement, see Milton Friedman, “A Theo retical Framework for Monetary Analysis, ’’Journal o f Political Economy (March/April 1970), pp. 193-238. Friedm an also argues that monetary policy is not useful in counter-cyclical policy because of lags in its impacts and that, consequently, it is more useful if steady or predictable; see his American Eco nomic Association Presidential Address, “The Role of M onetaiy Policy,” American Economic Review (March 1968), pp. 1-17, and his “ Monetaiy Policy” lecture cited in footnote 1. 27 FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 C h art 1 Income, M o n e y an d Credit Billions of D ollars 3 0 0 0 1------ ------ r HAS RISING CREDIT SIGNIFICANTLY AFFECTED THE RELATIONSHIP RETWEEN MONEY AND INCOME? markets by firms, consumers and the government, plus trade credit extended betw een firms.24 On a sem i-log chart, constant grow th rates graph as straight lines, and equal growth rates appear as There are several ways to assess the impact of parallel lines. In this format, it is plain that from 1959 rising credit on the money-income link. Three dif to 1981 credit’s growth was the fastest of the aggre ferent procedures are used here: (1) a consideration gates, that GNP and M2 have grown at roughly equal of the levels of GNP, money and credit; (2) an ex rates, and that all three grew somewhat faster than am ination of consum er deposit holdings, credit M l. The credit magnitude grew at an average rate of extensions and purchases; (3) observations of the 9.2 percent per year, while M2 grew at about the same rate as GNP during the last two decades—8.3 growth rates of M l and M2 velocities. First, we can see w hether the relationship be that it is the flow of credit—i.e., extensions—not the stock tween money and income growth appears to have 24Note of debt that is relevant here. Credit, as discussed earlier, is changed in recent years by simply looking at the data transaction-specific and can mediate only that transaction for which it is extended. Even if the promissory note from a pre on income, money and credit presented in chart 1. vious credit transaction were subsequently used as collateral for Chart 1, using a semi-log scale, depicts annual GNP, another credit transaction, there would be another credit ex M l and M2 holdings, and credit flows, with the last tension for that transaction. Unlike past money expansion, the defined as the quantity of funds raised in credit stock of past extensions is, in itself, irrelevant. Digitized for 28 FRASER FEDERAL RESERVE BANK OF ST. LOUIS MAY 1982 C hart 2 V elo cities of M l a n d M 2 The ratio of credit to income, while persistently percent and 8.2 percent per year, respectively. In contrast, M l grew at a 5.2 percent rate. rising, probably understates the importance of credit explaining the rise of M l velocity. The credit total In chart 2, the velocities of M l and M2 are dis in is misleadingly low since it represents quarterly played. The approximate constancy of the M2 veloc balance sheet changes in debt. If credit is extended ity is clearly evident here, as well as the persistent and repaid within the period of observation (one rise of M l velocity. Not so evident, however, is the quarter for the data in chart 1), there is no change in relatively constant rate of M l velocity growth. Over the credit balance and, thus, no evidence that this the 1959-81 period, M l velocity grew at around credit extension took place; nonetheless, such ex 3.2 percent. Indeed, except for a noticeable slowing tensions of credit would have m ediated exchanges in the late ’60s, the velocity growth rate of both old and contributed to spending and economic activity. M l and new M l has been betw een 3 percent and 4 percent since 1950.25 A second way to assess the impact of credit use is to focus on the behavior of individuals and families— 25Recently, Robert E. W eintraub, senior economist for the Joint in particular, to examine their holdings of demand Economic Committee of the U.S. Congress, made a similar and other checkable deposits as compared to credit point in a letter to the Wall Street Journal, October 14,1981: “As a matter of logic, offshore and other new financial developments in mediating consumer purchases. Table 1 presents can contribute to inflation only if they contribute to the rate of data on consumer deposit holdings, credit exten rise of money’s velocity. However, they have not. Since the early 1950’s, the rate or rise of M IB’s velocity has been quite sions and purchases in the U.S. economy during the 1970s. By focusing on the consumer sector, three steady, 3.2% yearly.” 29 FEDERAL CO o RESERVE C onsum er Expenditures and M ediations _ V e lo c ities (6) (7) (8) (9) (10) (11) (12) Consumer M2 deposits (5) Total consum er credit extensions Personal consum ption expenditures Total cash purchases Percent cash purchases 6 + 1 6 -3 7 -3 6 -4 $ 54.0 $ 458.5 $187.1 $ 634.1 $ 447.0 59.1 532.3 215.8 692.6 476.8 (1) (2) (3) (4) Year Demand deposits O ther checkable deposits Total checkable deposits 1970 $ 53.6 $ 0.4 1971 58.6 0.5 70.5% 11.83 11.74 8.27 1.38 68.8 11.82 11.72 8.07 1.30 1972 65.4 0.6 66.0 609.8 240.8 767.0 526.2 68.6 11.73 11.62 7.97 1.26 1973 70.1 0.8 70.9 654.8 269.0 834.3 565.3 67.8 11.90 11.77 7.97 1.27 1974 73.3 0.9 74.2 694.4 269.4 914.1 644.7 70.5 12.47 12.32 8.69 1.32 1975 78.0 1.6 79.6 796.2 280.7 1016.9 736.2 72.4 13.04 12.78 9.25 1.28 1976 82.6 3.2 85.8 921.2 318.2 1127.9 809.7 71.8 13.65 13.15 9.44 1.22 1977 91.0 4.8 95.8 1034.8 373.5 1254.5 881.0 70.2 13.79 13.09 9.20 1.21 1978 97.4 7.8 105.2 1117.5 424.2 1416.6 992.4 70.1 14.54 13.47 9.43 1.27 1979 99.2 17.7 116.9 1200.1 465.8 1582.3 1116.5 70.6 15.95 13.54 9.55 1.32 1301.7 74.3 17.10 13.49 10.03 1.36 1432.3 74.1 22.05 11.86 8.90 1.36 1980 1981 102.4 86.6 27.4 74.4 129.8 161.0 1286.2 1400.8 449.3 477.2 Notes: (1) Gross IPC Consum er demand deposits, year-end figures. Source: Federal Reserve B ulletin. Figure fo r 1981 is preliminary. (2) NOW and ATS accounts, credit union share drafts and demand deposits at m utual savings banks. Source: Federal Reserve Board. (3) IPC consum er demand deposits plus other checkable deposits. (4) M2 m inus o ve rn ig h t E urodollars m inus overnight RPs m inus money m arket m utual funds minus currency minus demand de posits plus IPC consum er demand deposits plus other checkable deposits. Source: Federal Reserve Bulletin. 1909.5 (5) Consum er in sta llm e nt cre d it extensions plus no n-installm en t consum er credit outstanding. The installm ent fig u re is 12 times the December total fo r that year, w hile the non-installm ent figure is tw o times the December total (under the assum ption of a sixm onth, term -to-m aturity structure of non-installm ent credit, on average). Source: Federal Reserve Board. (6) Expressed at annual rates. Source: Departm ent of Commerce. (7) Personal consum ption expenditures less total consum er credit [Col (6) - Col (5)]. (8) The ratio of total cash purchases to personal consum ption ex penditures [Col (7) - Col (6)[. MAY 1982 1751.0 OF ST. LOUIS Co nsu m er Deposits and Credit BANK Table 1 Consumer Deposits, Credit, Expenditures and Deposit Velocities (amounts in billions of dollars) FEDERAL RESERVE BANK OF ST. LOUIS technical national income accounting and com parability problems are avoided. First, all personal consumption expenditures are final goods trans actions and appear in GNP; in fact, they are over 60 percent of this measure. Hence, all the credit ex tensions to consum ers are used for final goods purchases. In contrast, commercial credit and trade credit may be financing intermediate goods. Second, a direct comparison of credit use and demand de posit holdings for an identifiable set of buyers is made possible; hence, characterizations about the relative use of credit and dem and deposits in rela tion to income are facilitated. Third, data on credit extensions are available so that a truer picture of credit utilization can be obtained than when using balance sheet changes in debt. The data in table 1 characterize the m anner in which households have made their purchases and held their deposits during the last 12 years; these data are based on fourth quarter and D ecem ber observations in each year. Clearly evident is the recent substitution of non-bank checkable deposits for dem and deposits (columns 1 and 2), as well as the steady decline in holdings of demand deposits rela tive to total purchases (column 6) m easured by their velocity (column 9). Conversely, the ratio of pur chases to total consumer checkable deposits, the velocity of total checkable deposits (column 10), rose much more gradually and fell abruptly in 1981 to about its level in 1970. As the data indicate, the proportions of consumer transactions initially mediated by money and credit (column 6) varied only slightly during the 1970s; the share of purchases that were mediated by currency and demand deposits rem ained around 70 percent (assuming a six-month term to m aturity in non installment credit) over the decade. Thus, over this period of rough constancy in the distribution oftypes of mediation, the ratio of consumer expenditures to demand deposit holdings by consumers (column 9) increased by almost 45 percent. Conversely, the ratio of purchases to total checkable deposits rose only 15 percent through 1980 (column 10). More over, in 1981, demand deposits fell abruptly (column 1) and other checkable deposits rose even more sharply (column 2) after the institution of NOW accounts nationwide. As a result, the velocity of total cheekables fell in 1981 to approximately its 1970 value. If we assume a narrow or transactions medium definition of money, M l, the observations over 197080 would be evidence of a decline in the quantity of MAY 1982 m oney dem anded by households. On the other hand, if we consider total cheekables in 1981 or assume a broader temporary-abode-of-purchasingpower definition, M2, then the ratios of consumer expenditures to the consumer deposit holdings pro vide contrary evidence. As shown in column 12 of the table, the ratio of consumer expenditures to the sum of household dem and deposits, saving and small time deposits, and money market mutual funds varied comparatively little relative to the demand deposit and total cheekables ratios. Thus, under the broader definition, the quantity of money dem anded —at least the consumer portion—does not appear to have declined during the 1970s. In particular, 1980 and 1981 do not appear to be qualitatively different than the earlier years. The third m anner of assessing credit’s impact is to determ ine w hether the trends in the income veloc ities of the m onetary aggregates have changed significantly in recent years. As we saw in the slopes of M l and M2 velocities in chart 2, monetary aggre gate velocities had strong trends in their growth over the two decades 1959-81. While on a quarter-toquarter basis velocity growth rates exhibit signifi cant variability, chart 2 suggests that over longer periods velocity growth is fairly regular. This trend regularity is substantiated in chart 3, which plots the growth rates of M l and M2 velocities. In this chart, quarter-to-quarter (QQ), four-quarter moving aver age (4QMA) and 20-quarter moving average (Trend) growth rates appear. While QQ is highly variable for both M 1 and M2, the 4QMA for each has a markedly smaller amplitude; considering ± 4 percent bands, only one observation for M l’s velocity growth and three observations for M2’s velocity growth lie beyond them . Also, the trend for each strongly underscores the apparent tendencies in chart 2; in each case, M l and M2 velocities have stable trends, especially when m easured over periods longer than a year. In particular, the charts do not reveal recent velocity growth to have been qualitatively different than in earlier years. This lack of change in M l and M2 velocity growth is even more apparent in table 2, which displays velocity growth rates, their standard deviations, and their ranges for 1961-81, for five-year subperiods, and for the year 1981; growth rates are computed for two observation frequencies: quarter-to-quarter (QQ) and four-quarter moving average (4QMA). Consider the behavior of M l velocity computed on a quarterly basis. Over the entire 1961-81 period, it has had an average growth rate of 3.16 percent per 31 FEDERAL RESERVE BANK OF ST. LOUIS C hart 3 Ve lo city G r o w t h Rates Note: QQ - qu arter-to -q uarter 4 Q M A = fo u r-q u a rte r m o v in g a v e r a g e 32 MAY 1982 MAY 1982 FEDERAL RESERVE BANK OF ST. LOUIS Table 2 Annual Growth Rates of M1 and M2 Velocities During 1961-81 Aggregate i Observation frequency M1 QQ Mean SD Range 3.25 3.62 -4 .4 7 , 13.90 3.71 2.62 - 1.18, 9.44 1.96 3.02 -4 .0 6 , 7.08 4QMA Mean SD Range 3.12 1.58 -1 .0 1 ,6 .9 8 . 3.17 1.79 - 1.01, 5.86 QQ Mean SD Range .17 4.05 -8 .2 3 ,1 1 .7 5 4QMA Mean SD Range .04 2.36 -5 .3 2 , 6.31 M2 Velocity G rowth at Annual Percentage Rates during: 1961-81 1961-65 1966-70 1971-75 1976-80 1981 3.64 3.64 -3 .3 6 , 9.00 3.39 3.49 -4 .1 6 , 10.02 4.74 9.13 -4 .4 8 , 13.90 2.36 1.52 -.0 9 , 5.15 2.94 1.16 .70, 5.47 3.77 1.51 1.86, 6.98 4.35 1.39 2.61, 6.01 -.5 9 2.62 - -4.32, 4.36 .68 3.54 -7 .8 1 , 5.75 -.2 8 4.13 -8 .2 3 , 6.26 .81 4.87 -6 .0 9 , 11.75 .47 8.03 -7 .0 0 , 10.83 -1.25 1.87 5.32, 2.29 1.06 1.72 -2 .7 6 , 3.01 -.6 5 2.41 -4 .4 6 , 3.96 .60 2.68 -3 .8 4 , 6.31 year. As was apparent in chart 3, quarter-to-quarter fluctuations can be significant; yet, over the two decades, the standard deviation of its growth rate has remained about 3.00. W hile extrapolating the longrun velocity growth rate of M l to 1981 underesti mates the observed growth rate, the 4.74 percent rate is well within one standard deviation of either the 1976-80 mean or that of the full 1961-81 period, and represents a fluctuation that is comparatively small in terms of the range of observed growth rates during either the subperiod or the full period as shown in chart 3. For M l, QQ and 4QMA have roughly the same average growth rates; for M2, the 4QMA growth rate is relatively more volatile than the QQ growth rate. Yet, in absolute terms the difference between QQ and 4QMA is about equal for M l and M2 for the entire 1961-81 period ( —.13) and for each subperiod except 1976-80 and 1981. For both M l and M2, the variability (SD) of 4QMA is naturally significantly less than that of QQ. The standard deviations of velocity growth computed on a four-quarter moving average are about one-half of the quarterly version for M l and the base and betw een one-half and twothirds for M2. Moreover, the standard deviation for 1981 is smaller than for the preceding subperiod. The implication is, as usual, that quarterly monetary statistics are a less useful guide to the longer-run behavior of m oney than averages over longer periods. 2.06 1.41 .23, 3.66 In summary, w hether we look at M l or M2, the information displayed in chart 3 and compiled in table 2 conveys the same message: namely, the behavior of monetary aggregate velocities in 1981 is not qualitatively different than over the preceding 20-year period or any of the subperiods. This is clearest when considering the four-quarter moving average growth rates, though the more volatile quarter-to-quarter rates tell essentially the same story. While velocity growth rates were higher in 1981 than in preceding subperiods during 1961-81, there is no evidence that credit use and financial innovations have severed the link betw een mone tary aggregates and the inflation rate. CONCLUSION Much of the current debate over U.S. economic policy has focused on the wisdom of targeting a monetary aggregate to control inflation. Some critics of such policies have alleged that financial innova tions have both made money uncontrollable and severed its predictable link with national income and prices. Others have argued that non-monetary assets or liabilities may have a closer link than money to income over the long run. This article has focused on the predictable linkage issue by exam ining the principal function of money and credit, the mediation of exchange. Since credit’s mediation 33 FEDERAL RESERVE BANK OF ST. LOUIS function depends crucially on the predictable source of monetary settlement, there is no theoretical sup port for assertions that the increasing use of credit has severed money’s link to income. In terms of the empirical evidence for the year 1981, both M l and M2 velocities grew reasonably close to their trend rates. This is grossly inconsistent with assertions that monetary policy is ineffective. While the controllability issue has not been ad dressed in this article, an analysis of the changes in monetary aggregates in relation to Federal Open Market Committee (FOMC) directives during 1981 suggests that both M l and M2 movements were 34 MAY 1982 strikingly in accord with the intentions of the FOMC.26 Consequently, there appears to be no reasonable foundation—theoretical or empirical—for abandon ing the use of a monetary aggregate as the vehicle for monetary policy. Unless or until velocity becomes more unpredictable or fluctuates over ranges not previously observed, the usefulness of monetary aggregates in controlling inflation and maintaining economic stability will be undiminished. 26See Daniel L. Thornton, "The FOMC in 1981: Monetary Con trol in a Changing Financial Environm ent,” this Review (April 1982), pp. 3-22.