The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Review ____ Vol. 69, No. 2 __ February 1987 5 T h e D ollar’s Effective E xch an ge Rate: A ssessing the Im p act o f Alternative Weighting S ch e m e s 15 The FOMC in 1986: Flexible Policy for U n certain T im es T he Review is p u b lis h e d 10 tim es p e r y e a r by the R esea rch an d P ublic In form ation D epartm en t o f the F e d e ra l R eserv e B ank o f St. Louis. S in gle-copy su bscrip tion s a r e available to th e p u b lic f r e e o f ch arg e. M ail req u es ts f o r su bscrip tion s, b a c k issues, o r a d d ress ch a n g es to: R esea rch an d P ublic In form ation D epartm ent, F e d e r a l R eserv e B ank o f St. Louis, P.O. Box 442, St. Louis, M issouri 63166. T he views e x p r e s s e d a r e th o se o f the individual au th ors an d d o not n ecessarily r e fle c t official p osition s o f th e F e d e r a l R eserv e B ank o f St. Louis o r the F ed era l R eserv e System . A rticles h erein m ay b e rep rin ted p r o v id ed th e s o u r c e is cred ited . P lease p rov id e the B an k’s R esea rch an d P ublic In form ation D epartm ent with a co p y o f rep rin ted m aterial. F ed eral R eserve Bank of St. Louis R eview February 1987 In This Issue . . . Effective exchange rate indexes are widely used in econom ic analysis, policy evaluation and financial planning and forecasting. Because these indexes are weighted averages of a num ber of exchange rates, their use avoids mistaken generalizations about the dollar's value that might arise by looking at fluctuations in a single exchange rate. Their value is generally conceded to depend on both the currencies included and the structure of the weighting schem e. In the first article of this Review, “The Dollar’s Effective Exchange Rate: Assess ing the Im pact of Alternative Weighting Schem es," Mack Ott com pares the effect of varying the weighting schem e in an effective exchange rate of the dollar against the 10 leading industrial econom y currencies. Ott finds that several alternatively weighted effective exchange rates using the same 1 0 currencies yield closely sim ilar results w hen used in an export equation. This somewhat surprising outcom e implies that the relative im portance of the weighting schem es used to derive effective exchange rate indexes needs to be more thoroughly investigated. In the second article, Philip A. Nuetzel reviews the deliberations and m onetaiy policy decisions of the Federal Open Market Committee (FOMC) last year. In "The FOMC in 1986: Flexible Policy for Uncertain Tim es,” Nuetzel discusses the factors that led the FOMC to take an accommodative policy stance. To illustrate the FOMC’s approach to policy in 1986, Nuetzel first reviews its long-run policy decisions for the year. This is followed by a more detailed exam ination of the FOMC's reasons for reducing its em phasis on M l in guiding policy, and a discussion of other variables used by the FOMC as policy guides. Finally, the FOMC’s short-run policy decisions are reviewed chronologically, each in the context of the prevailing econom ic clim ate in w hich it was made. 3 FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 The Dollar’s Effective Exchange Rate: Assessing the Im pact of Alternative Weighting Schem es M ack Ott M,r \NY analysts of international econom ics m ain tain that a multilateral weighted exchange rate is more useful than any single bilateral exchange rate in as sessing the value or changes in the value of the dollar .1 A multilateral or effective exchange rate (EER), which com prises many exchange rates, avoids mistaken gen eralizations that can result from changes peculiar to a single currency. Moreover, the EER reflects thirdcountry im pacts on the dollar’s exchange value, which are excluded in a bilateral exchange rate. The construction of an EER entails two analytic problems. First, w hich currencies should be in cluded? Second, how should the included currencies be weighted? These issues appear to be inextricably related, so that the correct choice for one issue would seem always to be conditional on the correct choice for the other. Yet, some insights about the relative im portance of the choice of weights can be obtained by examining the effects of changing the weights for a given set of exchange rates. This article exam ines the weighting issue using the Federal Reserve Board’s Trade-W eighted Exchange Rate (TVVEX). In particular, EERs constructed with trade weights, capital-flow weights and equal ("naive”) weights are com pared in terms of their explanatory Mack Ott is a senior economist at the Federal Reserve Bank of St. Louis. James C. Poletti provided research assistance. 'See Black (1976), Hooper and Morton (1978), Maciejewski (1983), Dutton and Grennes (1985), Belongia (1986), Cox (1986) and Rosensweig (1986). power and out-of-sample forecasts in a trade equa tion. THE USE O F EFFECTIVE EXCHANGE RATES The usefulness of an EER can be illustrated by asking w hether the dollar has strengthened or weak ened during some interval .2As chart 1 shows, the value of the dollar has appreciated against som e currencies and depreciated against others since 1973. For exam ple, the dollar has appreciated against the Canadian dollar and sterling, is about the sam e in 1986 as it was in 1973 against the DM, and has depreciated vis-a-vis the yen and Swiss franc. W ithin this 13-year span, most currencies have exhibited similar relative pat terns against the dollar, peaking in 1980 and bottom ing out in 1985. In contrast, the yen, Canadian dollar, Swiss franc and sterling each have had substantial departures from the com m on patterns. The Swiss currency has been notable for its consistently strong dollar value — the dollar buying roughly half the num ber Swiss francs in 1986 that it could in 1973. Moreover, as chart 2 shows, adjusting these bilateral exchange rates for different rates of inflation between the United States and the respective countries yield 2For expository purposes, therefore, we will use levels of the constit uent exchange rates in illustrating and explaining EERs. For many analytical applications, levels of the EER are less useful than their changes; consequently, the remainder of the article will focus on changes in the variously defined EERs. 5 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS C hart 1 N o m in a l Dollar Exchange Rates for G -7 Countries an d S w itzerlan d Perc e n t sim ilar patterns. The dollar’s real exchange rates against these currencies (adjusted by consum er price indexes) also dem onstrate disparate assessm ents of the change in the dollar’s value during this period. Still, m ost analysts believe that the dollar appreci ated during 1973-86. Such an assessm ent m ust be based on som e type of weighting schem e — that is, an average of the curren cies’ exchange rates is implicitly evaluated. The use of EERs is simply an explicit forma lization of this principle. CONSTRUCTING AN E E R : SOME GENERAL ISSUES In order to construct an EER, several questions m ust be answered: W hich currencies should be in cluded? What m easure of international com m erce 6 P ercent should be used to weight these currencies ? 3 Should the weights be based on bilateral or multilateral ex change? Should the weights be arithm etic or geom et ric? What time period should be used for the weights? It has been com m only argued that the answers to each of these questions depends upon the purpose of the analysis — that is, the use to w hich the EER will be applied .4 Which C urrencies? This choice generally has been governed by a com promise between com pleteness of the set of trading 3 With the exception of the IMF’s Multilateral Effective Exchange Rate (MERM), which has weights generated from the solution of a trade model, all major EERs are trade-weighted. "See Hooper and Morton (1978), Belongia (1986) and Rosensweig (1986). FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS C hart 2 Real Dollar Exchange Rates for G - 7 Countries and Switzerland Percent Percent 1973=100 175 1973=100 175 A n n u a l D ata Germa ny // 150 \ / /X . A v Cana da / /> jnce if 100 ^ \\ \\ iL \ \\ \ \ \ 125 Italy 150 \ \ \ t t \\ \\ / 125 100 \ Japan / \ V \ 7 Switze rland \ 75 75 \ U.K. 50 50 1973 74 75 76 77 78 79 partners and data availability. Most indexes use the principal industrial econom ies' currencies. The Inter national Monetary Fund’s (IMF’s) Multilateral Effective Exchange Rate (MERM) covers 21 countries, Morgan Guaranty Trust of New York uses 15 industrial cou n tries’ currencies and the Federal Reserve Board's TWEX, the best known example of such an index, is based on the Group of 1 0 countries plus Sw itzerland .5 The currencies in TWEX are used both because of the availability of data and because these countries ac count for m ost international trading activity. More over, the 10 U.S. trading partners in the G-10 countries plus Switzerland also account for m ost U.S. foreign trade. For example, in 1973, these countries accounted for 60.1 percent of U.S. exports plus im ports; including the United States, these 11 countries accounted for 5 See Belongia (1986) for a fuller discussion of these indexes and their characteristics. In contrast, Cox (1986) has recently formulated an index covering all 131 of the U.S. trading partners. 80 81 82 83 84 85 1986 67.2 percent of world exports plus imports. In 1983, these proportions fell to 52.5 and 53.8 percent, respec tively; then rose to 58.5 and 62.4 percent in 1985. What M easure o f C om m erce? Except for the IMF’s MERM, all existing EERs are weighted by some m easure of traded goods and ser vices, the sum of exports plus im ports .6 Yet, either capital flows or trade flows — that is, either side of the balance of payments statistics — would seem to be reasonable bases for weighting exchange rates. As Hooper and Morton observe, The total supply of and dem and for dollars on foreign exchange markets derive from U.S. dem ands for for eign goods and fo reign currency-denom inated finan cial assets and foreign dem ands for U.S. goods and dollar denom inated financial assets. . . . An excess sup- 6See Dutton and Grennes (1985). 7 FEDERAL RESERVE BANK OF ST. LOUIS ply of dollars resulting from a decline in dem and for U.S. goods or dollar denom inated assets would tend to cause a decline in the foreign currency price of the dollar.7 Thus, using capital flows, m easured as the sum of dom estic investment flows abroad and foreign invest m ent flows in the hom e country, provides an alterna tive approach for weighting each currency’s im por tance. Consider, briefly, the arguments in favor of each. T ra d e flow w eights Trade flow weights for the EER m easure the direct im pact on incom e (through net exports) of the foreign sector. Thus, a country w hose trade share is large is one w hose econom y’s im pact on U.S. markets is large, while a country with a sm aller trade share has less impact. The larger this share, the greater is the competitive im portance of that country’s producers for U.S. producers. Hence, the EER should also reflect these relative rankings of U.S. com petitors’ currencies. Capital flow w eigh ts Capital flow weights for the EER scale the currencies by the m agnitude of the financial flows betw een the respective countries. The currencies of countries with larger investment and portfolio flows are m ore im portant com petitors for the dollar in international transactions than are curren cies of countries with sm aller investm ent and portfo lio activity. Unlike trade weights, w hich em phasize an incom e approach to exchange rate determination, capital flow weights em phasize a financial approach to the dollar’s valuation .8 An EER using capital flow weights will reflect these financial market consider- 7Hooper and Morton (1978), p. 784; italics added. FEBRUARY 1987 ations and the relative im portance of the non-U.S. currencies in international finance. Multilateral o r Bilateral Weights? Under multilateral weighting, each country receives a weight equal to its proportion of total trade or capital flows. Under bilateral weighting, each country re ceives a weight equal to its proportion of the flows to and from the United States. Bilateral flows seem closer to the notion of measuring the im portance of individ ual U.S. trading partners to U.S. econom ic activity; however, they omit third-party effects. For example, if the DM-price of autos rises, other things the same, the German share of U.S. auto imports would fall, and the Japanese, Italian and Swedish share of U.S. imports would increase. Analogously, considering financial as sets, a multilateral weighting schem e is preferable because it includes these m ulticountry financial m ar ket im plications .9 B ase P eriod? This choice may depend on the period of the analy sis. If the relative size of trade or capital flows of the included countries are changing, it would seem that the base period should be chosen so that the weights characterize the structure of com m erce or investment throughout the period of analysis. If the structure shifted, the weights from an earlier period conceivably would no longer reflect the current trade or capital relations .10 Arithmetic o r G eom etric? The form of the index carries im plications for the comparative im portance of absolute vs. percentage changes. Most indexes, in particular the TVVEX, are weighted geometrically, so that proportional changes are em phasized." 8While capital flow weights emphasize the financial side of the balance of payments flows, they are not completely consistent with the modern asset market view of exchange rate determination which emphasizes stocks rather than flows; see Dornbusch (1976), Frenkel (1976, 1981), and Mussa (1979, 1982, 1984). As summa rized by Mussa (1979, p. 38): The asset market approach views the exchange rate as being deter mined by essentially the same forces that determine the prices of other assets that are traded in organized asset markets, such as the stock markets and the commodity exchanges. In such markets, prices are determined not by the balancing of How demands and flow supplies, but rather by the prices at which the market as a whole is prepared to hold the total outstanding stocks of the assets in question. Since the assets in question are durable, the currently determined price of an asset is tightly linked to the market’s expectation of the future price of that asset, (italics added) The measures of capital flows used in constructing the capital weights are the annual net increment in national asset portfolios by financial asset class. To the extent that the relative national asset holdings (stocks) of these financial assets do not change, the relative net flows would be proportional to the unobserved stocks. 8 9See Black (1976) and Hooper and Morton (1978). Hooper and Morton also note that the bilateral construction assigns Canada a 20 percent weight in the EER, which is probably distorted by the crossborder trade in partially completed automobile assemblies. Re cently, much attention has been focused on the dispute between adherents of bilateral vs. multilateral trade flows; see Belongia (1986), Cox (1986) and Rosensweig (1986). 1 Based on this possibility, Cox (1986) uses a moving-average 0 weighting scheme. This makes evaluation of the dollar problematic since changes in its value may result from changes in weights, not from changes in exchange values. 1 See Board of Governors of the Federal Reserve System (1978) and 1 Belongia (1986). Among widely used EERs, only the IMF's SDR is arithmetically weighted. FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 Table 1 Alternative Weights for Dollar EER Based on Trade Flows and Capital Flows Jase years BelgiumLuxembourg Canada France 1972-762 1979-83 .074 .071 .090 .078 .129 .130 1972-76 1979-83 .152 .093 .072 .046 .243 .149 Germany TWEX: Italy Netherlands United Sweden Switzerland Kingdo Trade Weights1 .206 .203 CWEX: Japan .090 .095 .084 .080 .040 .032 .034 .033 .118 .119 .058 .032 .136 .156 .014 .016 .067 .075 .125 .362 Capital Weights3 .103 .061 .095 .063 .070 .103 NOTES: 'Computed as the ratio of the five-year exports plus imports for the country divided by the sum of exports plus imports for all 10 countries, based on IMF data from the International Financial Statistics tape, July 1986. 2 These weights differ slightly from the Board of Governors' trade weights (see Board of Governors of the Federal Reserve System [1978]) due to data revisions. 3 Computed as the sum of the negative of non-official capital outflows ( - ) plus non-official capital inflows ( + ) for the five years for each country divided by the sum of such capital flows for all 10 countries based on data from the IMF Balance of Payments Statistics, July 1986. COMPUTATION OF TRADE AND CAPITAL WEIGHTS FO R E E R INDEXES As noted above, the weighting schem es generally applied in EERs are derived from data on trade flows, not capital flows. Yet, for the reasons offered above, capital flows offer a potentially useful alternative for weighting the exchange rates in an EER. The construction of the Capital Weighted Exchange Rate (CWEX) essentially parallels that of the TWEX. Since TWEX is familiar to m ost readers, we briefly review its construction, then exam ine that of CWEX. Following this, we show how each index is put into real term s; this deflation results in the priced-adjusted indexes, RTWEX and RCWEX. TWEX This index is constructed by com puting the trade flows (imports plus exports) of each of the non-U.S. G-10 countries as a percent of the total for all of these countries. These weights are com puted as the average for a five-year base period; two periods were used, 1972-76 and 1979-83. TWEX is then com puted as the product of these weights multiplied by the natural log (In) of the respective exchange rates, indexed to March 1973. Thus, 111 TWEX, = 100 10 w, T R, t i= 1 10 = 100 exp 2 w, In Hit, i= l where the weight for country i is 10 w, = (Imports, + Exports,)/ 2 (Imports, + Exports,), j =l and R,, = price in U.S. cents of currency i in M arch 1973 divided by its price at time t. The alternative forms of the exchange rate index, equation 1, are show n to em phasize that TWEX is a geometric rather than an arithm etic average of the constituent exchange rates. Also, note that TWEX is specified in average foreign currency units per dollar and is indexed to its value at the beginning of the floating-rate period, March 1973. Thus, a rise in TWEX m eans the dollar’s value is increasing, and values over 1 0 0 mean that its weighted foreign currency value is greater than it was in M arch 1973. The weights for the two base periods, 1972-76 and 1979-83, are displayed in table 1. CWEX This index is constructed by com puting the nonofficial net capital flows (imports plus exports) of each of the non-U.S. G-10 countries as a percent of the total for all of these countries. These capital flows include direct investment, portfolio investment, other long- 9 FEDERAL RESERVE BANK OF ST. LOUIS and short-term capital flows of deposit m oney banks and nonbank sectors as reported in the International Monetary Fund’s B a la n ce o f P aym en ts S tatistics; a detailed breakdown of the included item s appears in the appendix.1 2 Only non-official capital flows were used. This re striction is based on the assum ption that private agents will buy and sell assets based on rationally formed forecasts of relative asset values and antici pated changes in those values in order to maximize their wealth. Official flows, in contrast, may be driven by attem pts to change values or offset market anticipa tions. To the extent that these interventionist policies are successful, they will be reflected in non-official flows; otherwise, they are merely noise.1 3 Thus, the index is defined parallel to TWEX as 10 (2) CWEX, = 100 exp 2 x, log, R„ i= l w here the weight for cou ntiy i is x, = (Capital Outflows, + Capital Inflows,!/ 10 2 (Capital Outflows, -4 Capital Inflows,). - j=l The weights for CWEX for the two base periods are also displayed in table l . 1 4 Real EERs For many analytic purposes, price adjusted EERs, here RTWEX and RCWEX, are m ore useful than nom i nal EERs. These real indexes, in principle, are co n structed by weighting the real (price-deflated) ex1 One reservation about the capital weighting scheme is that it adds 2 net capital outflows plus net capital inflows while trade weights are, in principle, based on gross exports plus gross imports. Yet, the capital flows used in constructing the CWEX EERs are the sum of narrowly specified asset categories; hence, while inflows and out flows within any category (e.g., foreign holdings of corporate equi ties) are netted out, there is no cancellation across asset categories (e.g., foreign holdings of corporate equities and foreign holdings of public sector bonds). ,3See Batten and Ott (1984) for a general discussion of both the motivation for and the limitations on the efficacy of central bank foreign exchange intervention. ,4An alternative version of the capital weights was computed because Switzerland reported no data on direct investment — overseas investment by the Swiss and foreign investment in Switzerland. Since direct investment constitutes a substantial portion of the capital flows for the other countries, this would be likely to bias downward the weight for the Swiss franc. To compensate for this omission, a capital-weighted exchange rate index with net errors and omissions (CWEXO) was computed in the same manner as TWEX and CWEX; see appendix. The results of its comparative performance in the tests below, however, were indistinguishable from those reported and are omitted. 10 FEBRUARY 1987 change rates; however, this is equivalent to dividing the nom inal index by the ratio of a weighted index of foreign CPIs to the U.S. CPI. Thus, the real TWEX (RTWEX) is obtained as 10 (3) RTWEX, = 100 exp 2 w, [lnR„ - lnCPI,, + lnCPIust) i= l 10 10 = 100 exp [ 2 w, lnR„ - 2 w, (lnCPI,, - lnCPIus,)] i= l i= l 10 = TW EX/[100 exp 2 w, In (CPVCPI^,)]. i= 1 = TWEX/TWCPI,. The real CWEX (RCWEX) is obtained analogously as 10 (4) RCWEX, = CWEX,/[100 exp 2 x, In (CPI,/CPIust) i= 1 = CWEX/CWCPI,. COMPARISONS OF THE ALTERNATIVE E E R s In order to determ ine w hether different weighting schem es yield different results, em pirical assessm ents were made of their comparative usefulness. These em pirical analyses focused on changes, rather than levels, in the alternative EERs. First, correlation coefficients were com puted for the change in the natural logarithm (delta In) of the EERs, both nom inal and real. Second, the real EERs were each included as an explanatory variable in a trade equation with changes in U.S. agricultural exports as the dependent variable.1 These estim ates and their 5 out-of-sample forecasts provide m easures of the rela tive explanatory pow er of the different weighting schem es. In each of these empirical exercises, a “ naive” EER, in w hich each currency received equal weight, was also included as a benchm ark (or null hypothesis) to see w hether the theoretically based weights yielded superior results. Correlation am ong the EERs The correlations among these five exchange rate series, both nom inal and real (CPI-deflated), are re ported in table 2, and the results are striking. The delta In of these alternative EERs’ tim e series of changes are nearly perfectly correlated: the correlation coefficients 1 The choice of model was made to facilitate further comparisons with 5 the related work by Belongia (1986) on alternative exchange rate measures. FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 Table 2 Correlation Coefficients for ALn Samples of Alternatively Weighted G-10 EERs — Trade, Capital and Naive______________ TWEX72 TWEX79 CWEX72 CWEX79 NAIVE NOMINAL (January 1972 - August 1986 ; n = 175) TWEX72 TWEX79 CWEX72 CWEX79 NAIVE 1.000 1.000 1.000 .996 .994 1.000 .973 .973 .972 1.000 .998 .997 .995 .971 1.000 1.000 1.000 1.000 .995 .994 1.000 .970 .970 .970 1.000 Since this article focuses on the usefulness of the percentage change on delta In EER series, a delta In version of Belongia’s (1986) model was used to com pare the results of the alternative EERs. The purpose of this test was not to determ ine the best trade equa tion or test the validity of the specific equation esti mated. Rather, the purpose was simply to see how differently each EER series performed using a typical trade equation from the trade literature. The estim ated equation is REAL (January 1972 - December 1985; n = 167)’ TWEX72 TWEX79 CWEX72 CWEX79 NAIVE such as that used by Belongia, provides one direct way to determ ine determining w hether REERs that vary only in their weighting schem es also produce dispar ate regression and out-of-sample forecast results. .997 .996 .994 ,968 1.000 l (5) Ain X, = a„ + 2 auAin FGNP,_, i= 0 2 + NOTE: All correlations are significant at the .0001 level. ’Deflated by ratio of weighted foreign CPI to U.S. CPI. 2 Ain (USAGP/USCPI),_, j=l 3 2 + a3t Ain REER,.k + e,, k= 1 between the five EERs vary from .968 to 1.000 (rounded to 3 significant digits); this relationship holds for both nominal and real changes specifications. While the extremely high correlations both among the EERs and among the REERs may seem to imply that they will be virtually identical in any em pirical application, this generally is not correct. For example, Belongia found that, although different REERs were highly correlated, they generated different regression coefficients and highly divergent out-of-sam ple forecasts. C onse quently, the regression and forecast com parisons are included here in order to determ ine w hether or not these REERs perform identically. R egression an d Forecast R esults f o r the EERs In Belongia (1986), an equation explaining U.S. agri cultural exports was estim ated utilizing, in turn, five different REERs: the Federal Reserve Board’s TWEX, the IMF's MERM and SDR, Morgan Guaranty's EER, and the U.S. Department of Agriculture’s AG-Export weighted EER. These estim ates afford a com parison of the explanatoiy power and out-of-sam ple forecasts of the five REERs. Belongia found that these real ex change rates had substantially different regression and out-of-sample properties, even though the corre lation coefficients among them ranged from .853 to .983. Consequently, estim ating some trade equation, where, X, FGNP USAGP USCPI REER = = = = = real exports of U.S. farm comm odities, foreign real GNP“, index of U.S. farm prices, U.S. CPI, real TVVEX72, TWEX79, CWEX72, CWEX79, NAIVE and e = random error term. The results of estim ating this equation on quarterly data over I/1973-IV/1981 are reported in table 3; statis tics for out-of-sample forecasts over 1/1982— 1/1985 are reported in table 4. The sum m ed coefficients are displayed for FGNP and USAGP/USCPI and the individual coefficients for the Ain REERs. These coefficients and their signifi cance levels, as reported in table 3, are very similar across the five specifications for the non-REER vari ables, as are the R2 and Durbin-Watson statistics. The latter imply that the residuals do not have significant first-order correlation. The magnitude, signs and tratios for the REERs are also similar, although the sums of the REER coefficients differ slightly — the 1 This series, obtainable from the Federal Reserve Board, is a hybrid 6 of weighted foreign GNPs, containing Mexican and other oilexporting countries’ GNPs as well as the (non U.S.) G-10 plus Switzerland industrial countries’ GNPs. 11 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS Table 3 Estimates of U.S. Agricultural Exports Equation Using Alternatively Weighted G-10 REERs (1/1973— IV/1981) Weighting scheme Intercept lAlnFGNP XAIn(USAGP/USCPI) AlnREER R2 t t-1 DW t —2 TWEX72 0.004 (0.202) 1.597 (0.782) -0 .4 9 5 (2.156)* 0.326 (0.448) 0.751 (1.035) -1.642 (2.027) + 0.335 2.16 TWEX79 0.004 (0.202) 1.603 (0.782) -0 .4 8 9 (2.119)* 0.321 (0.442) 0.789 (1.090) -1 .5 8 9 (1.971) + 0.332 2.16 CWEX72 0.005 (0.242) 1.494 (0.730) -0 .4 9 8 (2.194)* 0.253 (0.368) 0.664 (0.959) -1 .5 4 9 (2.018) + 0.324 2.16 CWEX79 0.003 (0.153) 1.712 (0.624) -0 .4 7 2 (2.138)* 0.461 (0.635) 0.791 (1.084) -1 .6 4 4 (2.026) + 0.344 2.16 NAIVE 0.004 (0.212) 1.585 (0.775) —0.486 (2.113)* 0.363 (0.493) 0.771 (1.051) -1 .5 7 4 (1.940) + 0.329 2.16 NOTE: Absolute t-ratios in parentheses; * indicates significance at 5% level and + indicates significance at 10% level. CWEX72 being sm aller and CWEX79 larger than the other three, although not significantly so .17 The out-of-sam ple forecast properties of the five estim ates of equation 5, differing only in their REERs, are shown in table 4. Error statistics and Theil statis tics from the forecast series are displayed. The error statistics — the m ean error, the m ean absolute error and the root-m ean-square error (RMSE) — are nearly identical for the five equations. Thus, the accuracy of the forecasts does not vary with the weighting schem e used for the REER. The Theil statistics decom pose the forecast errors into three com ponents .18 As shown in these error decom positions, there is no substantive difference in the pattern of the forecast errors. The close conformity of the regression and forecast results for the variously weighted versions of the G-10 "Some differences in these sums may reflect scale differences, as the CWEX72 and CWEX79 have weights which differ most from TWEX72;see table 1. ,8Bias measures the proportion of the mean square error (RMSE squared) due to a tendency to estimate too high or too low the level of the forecast. Variance measures the proportion of the MSE due to the variance of predictions differing from the variance of actual levels. The covariance is, essentially, the residual error proportion. 12 REERs contrasts starkly with the divergent results reported for different REERs in Belongia (1986). There are two key differences betw een Belongia’s and those shown here. First, the REERs in this study contain the same currencies; differences betw een them are lim ited solely to alternative weighting schem es. In con trast, Belongia used REERs that differed both in their currencies and in their weighting. Second, the analy sis here focuses on changes in the In REER; Belongia focused on levels of these data. CONCLUSION Trade-weighted effective exchange rates are widely used to assess both the value of the dollar as an end in itself and to provide a broad measure for use in analyz ing and explaining trade and capital flows. Surpris ingly, while questions often arise about w hich curren cies to include or how to weight them, alternatives to asymmetrically trade-weighted EERs have seldom been examined. Several alternative EERs have been exam ined in this article. An important finding is that the equally weighted naive EER is highly correlated with both the traditional trade-w eighted EERs and alternative capital-flow-weighted EERs over the range of consid FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 Table 4 Out-of-Sample Statistics for Projected Farm Exports Using Alternatively Weighted G-10 REERs (1/1982-1/1985) Error Statistics Theil Statistics Fraction due to: Weighting scheme Mean error Mean absolute error RMSE Bias Variance Covariance TWEX72 TWEX79 CWEX72 CWEX79 NAIVE 0.031 0.031 0.030 0.031 0.032 0.058 0.059 0.058 0.059 0.058 0.080 0.081 0.080 0.082 0.081 0.147 0.151 0.147 0.139 0.155 0.419 0.424 0.434 0.392 0.436 0.434 0.425 0.420 0.469 0.409 ered weights. Moreover, the explanatory and predic tive power of the alternative EERs, including the naive EER, were found to be statistically equivalent in an agriculture export equation. Since these results are for one set of currencies and for one historical period, generalizations m ust be advanced with care; however, these results suggest that further research — both empirical and theoretical — on the comparative im portance of the choice of weighting schem es vs. the choice of currencies to be included in the EER is w arranted .19 REFERENCES Artus, Jacques R., and Rudolf R. Rhoneberg. “A Multilateral Ex change Rate Model,” International Monetary Fund Staff Papers (November 1973), pp. 591-611. Batten, Dallas S., and Mack Ott. “What Can Central Banks Do About the Value of the Dollar?” this Review (May 1984), pp. 16-26. Belongia, Michael T. “Estimating Exchange Effects on Exports: A Cautionary Note,” this Review (January 1986), pp. 5-16. Black, Stanley W. “Multilateral and Bilateral Measures of Effective Exchange Rates in a World Model of Traded Goods," Journal of Political Economy (June 1976), pp. 615-21. Board of Governors of the Federal Reserve System. “Index of the Weighted-Average Exchange Value of the U.S. Dollar: Revision,” Federal Reserve Bulletin (August 1978), p. 700. Cox, Michael. “A New Alternative Trade-Weighted Dollar Exchange Rate Index,” Federal Reserve Bank of Dallas Economic Review (September 1986), pp. 20-28. Dornbusch, Rudiger. “Expectations and Exchange Rate Dy namics,” Journal of Political Economy (December 1976), pp. 1161— 76. Dutton, John, and Thomas Grennes. “The Measurement of Effec tive Exchange Rates Appropriate for Agricultural Trade,” mimeo graph, Department of Economics and Business, North Carolina State University (August 1985). Fisher, Irving. The Making of Index Numbers (Houghton-Mifflin Company, 1922). Frenkel, Jacob A. “Flexible Exchange Rates, Prices and the Role of News’: Lessons from the 1970s,” Journal of Political Economy (August 1981), pp. 665-705. Hooper, Peter, and John Morton. “Summary Measures of the Dol lar’s Foreign Exchange Value,” Federal Reserve Bulletin (October 1978), pp. 783-89. Maciejewski, Edouard B. “ ‘Real’ Effective Exchange Rate Indices,” International Monetary Fund Staff Papers (September 1983), pp. 491-541. 1 lronically, Irving Fisher (1922, p. 365) first advanced the importance 9 of this question 65 years ago after arriving at an analogous empirical finding for weighting schemes for price indexes: Among the consequences of the surprising agreement between the various legitimate methods of calculating index numbers are two which need emphasis here. The first is that all discussion of “ different formulae appropriate for different purposes” falls to the ground. The second is that, the supposed differences among formulae once banished, the real problem of accuracy is shifted to the other features of an index number, — the assortment of the commodities included, their number, and data. . .. Thus the figures for weights in particular may usually be tenfold or one tenth of the true figures without appreciably disturbing the accuracy of the resulting index number. Henceforth, the effort to improve the accuracy of index numbers must center chiefly on the assortment of the items to be included. Mussa, Michael. “Empirical Regularities in the Behavior of Ex change Rates and Theories of the Foreign Exchange Market,” Policies for Employment, Prices and Exchange Rates, in Karl Brun ner and Allan Meltzer, eds., Carnegie-Rochester Conference Se ries on Public Policy, Volume 11 (1979), pp. 9-58. ________ _ “A Model of Exchange Rate Dynamics,” Journal of Political Economy (February 1982), pp. 72-104. _________ "The Theory of Exchange Rate Determinations," John F. O. Bilson and Richard C. Monston, eds., Exchange Rate Theory and Practice, University of Chicago Press, 1984. Rosensweig, Jeffrey A. “A New Dollar Index: Capturing A More Global Perspective,” Federal Reserve Bank of Atlanta Economic Review (June/July 1986), pp. 12-22. 13 FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 Appendix Sources of Data and Specification of Weights TWEX: Data are from the July 1986 edition of the In tern ation al F in an cial Statistics tape of the IMF. The data utilized are the im ports of goods and services plus the exports of goods and services in billions of U.S. dollars, annual during 1972-76 and 1979-83. CWEX: Data are from the July 1986 edition of the B alan ce o f P aym ents S tatistics tape of the IMF. The data utilized are: Line D ata D escrip tio n Sign/Code D irect Investm en t Direct Investment Abroad Foreign Direct Investm ent at Home -3L.X4 3Y.X4 P ortfolio Investm en t 53 54 55 56 57 58 59 60 Public Sector Bonds Assets Official Liabilities Other Liabilities Other Bonds Assets Official Liabilities Other Liabilities Corporate Equities Assets Official Liabilities Other Liabilities -6A1X4 6T1X4 6Q1X4 -6B1X4 6U1X4 6R1X4 -6D1X4 6V1X4 6S1X4 O th er L on g -T erm Capital of D eposit M oney Banks 69 70 Drawings on Loans Extended Repayments on Loans 14 -5C1Y4 -5C1W4 71 72 73 74 75 76 Other Assets Liabilities Official (National Currency) Liabilities Official (Foreign Currency) Drawings on Other Loans Repayments on Other Loans Other Liabilities -5K1X4 5U1X4 5V1X4 5P1W4 5P1Y4 5S1X4 O th er L on g-T erm Capital of O th er S e cto rs 77 78 79 80 81 82 83 Drawings on Loans Repayment on Loans Other Assets Liabilities (Foreign Official) Drawings on Loans Repayments on Loans Other Liabilities -8C1Y4 -8C1W4 -8K1X4 8W1X4 8P1W4 8P1Y4 8X1X4 O ther S h o rt-T erm Capital of D eposit M oney Banks 89 90 91 92 Assets Liabilities (National Currency) Liabilities (Foreign Currency) Other Liabilities -5L2X4 5U2X4 5V2X4 5X2X4 O th er S h o rt-T erm Capital of O th er S e cto rs 93 94 95 96 97 Loans Extended Other Assets Liabilities (Foreign Reserves) Other Loans Received Other Liabilities -8C2X4 -8K2X4 8W2X4 8P2X4 8S2X4 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS The FOMC in 1986: Flexible Policy for Uncertain Times Philip A. N uetzel T J L HE Federal Reserve’s m onetary policy actions during 1986 were influenced by indications of weak econom ic growth and m oderate inflation. The incom e velocity of money, defined as the ratio of nom inal GNP to the narrowly defined m oney supply M l, declined even more rapidly in 1986 than it had in the previous year .1 Interest rates declined on balance over 1986, and the Federal Open Market Com m ittee (hereafter “Com m ittee” or “FOMC") viewed their decline and the asso ciated rapid growth of M l as a desirable development in light o f the sluggish econom y. As the year pro gressed, the Committee deem phasized M l as a guide to policy while focusing on the broader monetary aggregates, M2 and M3, and several indicators of eco nom ic and financial conditions. In the uncertain eco nom ic environm ent that prevailed in 1986, the Com mittee was flexible in its approach to m onetary policy. This article reviews the FOMC’s m onetary policy decisions during 1986. The C om m ittee’s annual growth objectives for the m onetary aggregates are discussed in the next section, and the target ranges for 1986 are com pared with actual money growth during Philip A. Nuetzel is an economist at the Federal Reserve Bank of St. Louis. Laura A. Prives provided research assistance. NOTE: Citations referred to as “Record” are to the "Record of Policy Actions of the Federal Open Market Committee" found in various issues of the Federal Reserve Bulletin. Citations referred to as “Re port” are to the “Monetary Policy Report to Congress,” also found in various issues of the Federal Reserve Bulletin. ’M1 growth from IV/1985 to IV/1986 was 15.3 percent, up from 12.1 percent over the preceding four quarters. Over the same two peri ods, nominal GNP grew at rates of 4.2 percent and 6.3 percent, respectively. the year. Then, the Com m ittee’s views concerning the rapid growth of M l are considered in more detail, and other variables that had a significant influence on policy are discussed. Finally, the short-run directives issued by the FOMC during 1986 are reviewed chrono logically. ANNUAL TARGETS FO R 1986 Each February, the Board of Governors appears b e fore Congress to report on the annual growth targets that the FOMC has established for the monetary and credit aggregates for the com ing year. In July, the Board reports on the progress made toward meeting these goals and announces the FOMC’s provisional growth targets for the following calendar year .2 The Committee states its annual targets in term s of growth ranges from the fourth quarter of the previous year to the fourth quarter of the current year .3 The dates of the three meetings at w hich the annual target ranges for 1986 were considered are listed in table 1 along with the ranges established for M l, M2 and M3, and the actual growth rates of these aggregates during 1986. 2 These reports are required under the Full Employment and Bal anced Growth Act of 1978, also known as the Humphrey-Hawkins Act. 3 The Committee's use of the fourth quarter of the previous year as the base period for establishing the current year’s growth targets leads to an upward drift in money growth if the growth of an aggregate during the previous year exceeded the target range for that year. The "base drift” problem is discussed by Broaddus and Goodfriend (1984). For a viewpoint that favors base drift, see Walsh (1986). 15 FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 July 1985 M eeting The tentative ranges for 1986 established at the July 1985 FOMC m eeting reflected the Com m ittee’s feeling, at that time, that continuation of the rapid money growth in the first half of 1985 might be inconsistent with sustainable econom ic expansion and reasonable price stability. The tentative 1986 range for M l of 4 to 7 percent was 2 percentage points narrower than the 3 to 8 percent rebased range adopted in July 1985 for the 11/1985— IV/1985 period ;4 the upper limit of the tentative M3 range of 6 to 9 percent was one-half percentage point lower than the upper limit of its 1985 range. The growth of the three m onetary aggregates slowed somewhat during the second half of 1985. M l growth exceeded the upper limit of its rebased range, however, and its velocity declined even m ore sharply than it had during the first half. February 1986 M eeting The annual growth ranges for 1986 were reconsid ered at the Com m ittee’s m eeting on February 1 1 - 1 2 , 1986. For M l growth, the Committee chose a target range of 3 to 8 percent, w hich was based on expecta tions that M l growth would slow while nom inal GNP growth would accelerate. This range was 2 percentage points wider than the FOMC had tentatively planned, reflecting continuing uncertainty about the future b e havior of M l and its velocity. The Board’s report to Congress stated that: The width of the M l range reflects continuing u n cer tainty about the behavior of M l under varying eco nomic and financial circu m sta n ce s.. . . While the range for M1 is wide enough to allow for some variation in behavior of the aggregate's incom e velocity in re sponse to changing conditions, the range was set on the assumption that there would not be a large drop in velocity, such as occu rred in 1985. In that connection, the Committee will evaluate behavior of M l in light of its consistency with other monetary aggregates, e co nomic and financial developments, and the potential for inflationary pressures.3 The 6 to 9 percent ranges for M2 and M3 established tentatively in July were affirmed by the Committee at “Because of the apparent decline in M1 velocity during the first half of 1985, the FOMC decided in July 1985 that the annual range origi nally established for M1 growth was undesirable. The Committee voted to move the base period of that range from IV/1984 to 11/1985. See Hafer (1986) for a discussion of the decision to rebase the M1 range and the influence that declining M1 velocity had on the FOMC’s decisions in general during 1985. 5Report (April 1986), p. 214. 16 Table 1 FOMC Annual Monetary Growth Ranges and Actual Money Growth: IV/1985-IV/1986 Dates of meeting M1 M2 M3 July 9-10,1985' February 11-12, 1986 July 8-9, 1986 4-7% 3 -8 [3— 8]2 6-9% 6-9 6-9 6-9% 6-9 6-9 Actual Growth Rate 15.3% 8.9% 8.7% 'The ranges established for 1986 at this meeting were tentative. Mr. Martin dissented from this action because he preferred a somewhat higher growth range for M1 to guard against the possibility that velocity would continue to decline or grow slug gishly in 1986. Ms. Seger dissented because she felt that higher growth ranges were more consistent with sustaining economic expansion and reducing financial strains in the economy. 2 The Committee did not establish any new range for M1 growth at this meeting and agreed that growth in excess of the existing range would be acceptable, taking account of the growth of the broader aggregates and other economic and financial develop ments. the February m eeting. The growth of these broader aggregates had been within their target ranges in 1985, and it was observed that "on balance over the past few years, the behavior of M2 and M3 seem ed to have been less affected by in stitu tion al and in terest rate changes .”6 July 1986 M eeting Over the first half of 1986, the growth rates of M2 and M3 were roughly in the middle of their target ranges. M l growth slowed in the first quarter from its pace in 1985, but accelerated sharply in the second quarter, leaving its annualized growth rate from Decem ber through Ju n e at 13.3 percent, m ore than 5 percentage points above the upper bound of the Com m ittee’s target range for the year. Other evidence reviewed by the Com mittee at the m eeting of July 8-9, 1986, sug gested that the rate of econom ic growth had slowed considerably in the second quarter from the 2.9 per cent growth of real GNP registered in I/1986.7 Further 6Record (June 1986), p. 410. The Committee also adopted a moni toring range for the growth of total domestic nonfinancial debt of 8 to 11 percent. 'Revisions showed that real GNP grew at a rate of 3.8 percent in 1/1986 and 0.6 percent in 11/1986. All of the data used in the text in discussing the Committee’s deliberations are those that were avail able to the Committee at the time. FEDERAL RESERVE BANK OF ST. LOUIS more, wage and price increases continued to m oder ate over the first half of the year, even w hen the direct effects of declines in food and energy prices were eliminated. The Com m ittee reaffirmed the long-run growth ranges of 6 to 9 percent for M2 and M3. In their discussion of the range for M l growth, however, the m em bers again em phasized uncertainties affecting the outlook for M l velocity and the changes in M l’s com position resulting from the rapid growth of its deregulated, interest-bearing com ponent. Interest rates had declined by 1 to 2 percentage points during the first half and were thought to be associated with the rapid growth of M l. Some m em bers thought that the M l range should be eliminated, "at least pending the reestablishm ent of a more predictable relation ship with overall m easures of econom ic activity .”8 A majority, however, preferred to retain a range for M l “even though they believed its operational signifi cance could only be judged in the perspective of concurrent econom ic and financial developments, in cluding the behavior of M2 and M 3 .”9 Rather than raising or rebasing the existing M l range, the Commit tee acknowledged its desire to accom m odate u ncer tain changes in M l dem and by agreeing that, after accounting for the behavior of the broader aggregates and other developments, including trends in interest rates, growth of M l in excess of 8 percent would be acceptable for 1986. Actual M oney Growth in 1986 The actual growth rates of the m onetary aggregates for 1986 are reported at the bottom of table 1. M l growth of 15.3 percent was 7.3 percentage points above the upper bound of its range set early in the year. This growth rate represents a significant acceler ation from the already rapid 12.1 percent growth of M l from IV/1984 to IV/1985. In com parison, M l’s average annual growth rate over the 1960-84 period was only 5.8 percent. One reason why the Com m ittee accepted the rapid M l growth during 1986 was that the growth rates of M2 and M3 were quite close to the 9 percent upper limit of their target ranges for the year. M2 growth of 8.9 percent and M3 growth of 8.7 percent were slightly more rapid than their respective growth rates of 8.7 FEBRUARY 1987 percent and 7.7 percent over the IV/1984-IV/1985 p e riod. The Committee felt, however, that the behavior of these aggregates was generally consistent with its overall policy objectives. The Board’s mid-year report to Congress stated that “during a period of greater overall price stability and adequate capacity relative to the dem ands placed upon it,” m onetary policy had been able "to accom m odate dem ands for money and credit, helping facilitate further declines in interest rates___”10 THE FOMC’s GUIDES FOR POLICY IMPLEMENTATION The Com m ittee’s use of a variable as a policy guide is based on assum ptions about the effect of that vari able on real econom ic growth and the rate of inflation, w hich are of ultimate concern, as well as the ability of policy actions to influence that variable. As noted previously, the deemphasis of M l as a policy guide was a reaction to the instability of M l velocity. In this section, we review the Com m ittee’s explanation for the unusually rapid growth of M l relative to nominal GNP and its reasons for believing that the relationship between M l and econom ic activity would be subject to a high degree of uncertainty for some time. We then discuss the Com m ittee’s reasons for continuing to use other variables as policy guides, such as M2 and M3, the level of interest rates, the foreign exchange value of the dollar, real GNP and indicators of output and prices. Institutional Changes and Ml In accounting for the rapid growth of M l in 1986, the Committee em phasized that the com position of M l had changed in part because of the deregulation of deposit interest rates and minimum balance require m ents that had taken place under the Monetary Con trol Act of 1980 and the Garn-St. Germain Act of 1982." From 1980 to 1985, for example, the proportion of interest-bearing checkable deposits (other checkable deposits, or OCDs) in M l rose from 5.5 percent to 27.6 percent, a trend that continued in 1986. Table 2 shows the annualized growth rates of M l and its three m ajor com ponents for each quarter of ,0Report (September 1986), p. 603. aRecord (October 1986), p. 708. 9lbid. "For a discussion of federal policy on deposit interest rate ceilings, or Regulation Q, and the phaseout of deposit regulation, see Gilbert (1986). 17 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS Table 2 Growth of M1 and Its Major Components in 1986 (seasonally adjusted, compounded annual rates) Period Currency IV/1984— IV/1985 7.5% IV/1985— 1/1986 1/1986-11/1986 11/1986-111/1986 111/1986— IV/1986 7.5 6.6 7.5 8.3 IV/1985— IV/1986 Demand deposits 7.5 1986 and for the year as a whole. The OCD com ponent of M l grew at 28.6 percent, 6.4 percentage points faster than its 1985 growth rate. Growth in dem and deposits was also quite rapid, accelerating from 8.9 percent in 1985 to 11.6 percent in 1986. Growth of the public’s currency holdings was unchanged at 7.5 percent. The Com m ittee attributed m uch of the rapid growth of OCDs and dem and deposits, and hence declining M l velocity in 1986, to declining interest rates and subsiding inflationary expectations. In the m idyear report to Congress, the Board stressed that “the advent and expansion of interest-bearing check ing accounts over the years have attracted more savings-type balances and have increased the respon siveness of M l to interest rate changes .” 12 Com m ent ing on the rapid growth of dem and deposits, Chair m an Paul Volcker said that th ere were “som e indications of a greater willingness of businesses to hold dem and deposits at a time of lower interest rates, partly because, with interest rates down, a larger bal ance is necessary to com pensate banks for a given am ount of services .” 13 Chart 1 shows the paths of selected short-term Other checkable deposits 22.2% 12.1% 4.7 15.4 13.3 13.4 17.9 28.1 34.2 34.9 9.1 16.4 17.6 18.2 11.6 28.6 15.3 8.9% interest rates and the growth of M l over 1985 and 1986. There were m ajor accelerations in M l growth accom panied by declining short-term rates during the spring of 1985, the late w inter and spring of 1986, and the sum m er of 1986. On the other hand, there were also accelerations of M l growth during periods of relatively stable interest rates in the sum m er and the late fall of 1985, and the fall of 1986. The consensus at the Com m ittee’s m eeting in July 1986 was that rapid M l growth during the first half of the year reflected lagged adjustm ents to declining inflationary expecta tions and interest rates; it did not seem to hold the usual potential for reigniting inflationary pressures w hen judged in the context of other developments, including m ore restrained growth of the broader aggregates .14 While the Committee thought that M l still had some informational value, it did not wish to restrain the process of adjustm ent that it felt was responsible for rapid M l growth and declining velocity. Chairman Volcker stated in July 1986 that “. . . a firm conclusion concerning the nature and stability of future velocity characteristics may take years of experience in the new institutional and econom ic setting .” 15 The Chair m an w ent on to summarize the Com m ittee’s attitude toward the use of M l as a policy tool: ,2Report (September 1986), p. 614. Support for the view that the interest elasticity of money demand has increased with deposit deregulation can be found in Keeley and Zimmerman (1986), Roth (1985), Mehra (1985) and Wenninger (1986). 1 See the Record (October 1986), p. 708. 4 1 Volcker (1986), p. 640. 3 t5Volcker (1986), p. 641. 18 M1 FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 Chart 1 Short-Term Interest Rates and M l Growth J F M A M J J A S O N D 1985 J F M A M J J A S O N D 1986 1J_ M l g r o w t h r a t e s a r e c o m p o u n d e d a n n u a l r a t e s o f c h a n g e o f f o u r - w e e k m o v i n g a v e r a g e s o f M l f r o m f o u r w e e k s p r e v i o u s . Experience over the first half of 1986 underscored the difficulty — I would say impossibility — of conducting m onetary policy in current circum stances according to one or two simple, preset criteria . . . the weight of the evidence strongly suggests that M l alone during this period of econom ic and institutional transition is not today a reliable m easure of future price pressures (or indeed a good short-term "leading indicator” of business activity). The m ore restrained perform ance of the broader aggregates, as well as the perform ance of the econom y and prices themselves, point in a differ ent direction.'6 M2 and M3 In recent years, the velocity behavior of M2 and M3 16lbid., p. 642. have not changed as radically as M l’s. Apparently, the shifts in asset holdings resulting from the com bina tion of falling interest rates and deposit deregulation have occurred largely within the broader aggregates and have not led to as large a surge in their growth as in M l’s. The FOMC has responded by assigning more weight to M2 and M3 as policy guides .17 In 1986, the Commit - 1 Monetary policy actions have a weaker influence on the broader 7 aggregates than on M1, however, because most of the non-M1 deposit liabilities included in M2 and M3 are not subject to reserve requirements. Moreover, information on these aggregates is avail able at a longer lag. See Lawler (1981). Also, Hafer (1981) presents empirical evidence suggesting that M2 is less controllable through policy actions than M1. 19 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS Organization of the Committee in 1986 The Federal Open Market Committee (FOMC) consists of 1 2 m em bers: the seven m em bers of the Federal Reserve Board of Governors and five of the 12 Federal Reserve Bank presidents. The chairm an of the Board of Governors is, by tradition, also chairm an of the Committee. The president of the New York Federal Reserve Bank is, also by tradition, its vice chairm an. All Federal Reserve Bank presi dents attend Com m ittee m eetings and present their views, but only those who are m em bers of the Committee may vote. Four m em berships rotate among Bank presidents and are held for one-year term s beginning on March 1 of each year. The president of the New York Federal Reserve Bank is a perm anent voting m em ber of the Committee. Members of the Board of Governors at the begin ning of 1986 included Chairman Paul A. Volcker, Vice Chairman Preston Martin, Hemy C. Wallich, J. Charles Partee, Em m ett J. Rice and Martha R. Seger .1 The term of J. Charles Partee expired on February 1. The two open seats on the Board were filled by Wayne D. Angell and Manuel H. Johnson on Febru ary 7. Preston Martin resigned from the Board effec tive April 30.2 On August 19, H. Robert Heller joined the Board, and, on August 22, Manuel H. Johnson took the office of vice chairm an. Henry C. Wallich resigned from the Board effective D ecem ber 15, and Em mett J. Rice resigned effective D ecem ber 31.3 The following Bank presidents voted at the m eet ing on February 11 and 12, 1986: Robert P. Black (Richmond), E. Gerald Corrigan (New York), Robert P. Forrestal (Atlanta), Silas Keehn (Chicago) and Robert T. Parry (San Francisco). The Committee m em bership changed in March and the presidents’ voting positions were filled by E. Gerald Corrigan (New York), Roger Guffey (Kansas City), Karen N. Horn (Cleveland), Thom as C. Melzer (St. Louis) and Frank E. Morris (Boston ).4 The Committee m et eight times at regularly scheduled m eetings during 1986 to discuss eco nom ic trends and to decide upon the future course of open-market operations .5 As in previous years, telephone or telegram consultations were held o c casionally betw een scheduled meetings. During each regularly scheduled meeting, a directive was issued to the Federal Reserve Bank of New York. Each directive contained a short review of eco nom ic developments, the general econom ic goals sought by the Committee, the Com m ittee’s longrun monetary growth objectives and instructions to the Manager of the System Open Market Account at the New York Bank for the conduct of open-market operations. T h ese in stru ction s typically were stated in terms of the degree of pressure on reserve positions. The latter were associated with expected short-term growth rates for M l, M2 and M3 that were in turn considered to be consistent with de sired longer-run growth rates of the m onetary ag gregates .6 The Committee also specified interm eet ing ranges for the federal funds rate. These ranges provide a m echanism for initiating consultations between meetings whenever it appears that the constraint of the federal funds rate is proving in consistent with the objectives for the behavior of the m onetary aggregates. The account manager has the m ajor responsibil ity for formulating plans regarding the timing, types and am ount of daily buying and selling of securities in fulfilling the Com m ittee’s directive. Each m orn ing the m anager and his staff plan the open-market operations for that day. This plan is developed on the basis of the Com m ittee’s directive and the latest developments affecting m oney and credit market conditions, the growth of the monetary aggregates and bank reserve conditions. The m anager also consults with the Board of Governors’ staff. Present market conditions and open-m arket operations that the m anager proposes to execute are dis- 'Ms. Seger did not attend the May FOMC meeting. 2Mr. Martin did not attend the April meeting. 3 Mr. Wallich did not attend the May and November meetings. Mr. Rice did not attend the December meeting. “Robert H. Boykin (Dallas) voted as an alternate for Mr. Melzer at the May meeting. 20 5No formal meetings were held in January, March, June and October. 6ln July, the FOMC agreed that M1 growth in excess of its annual target range would be acceptable and discontinued statements regarding the Committee's expectations for short-run growth in M1. FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS cussed each m orning in a telephone conference call involving the staff at the New York Bank, the Board and one voting president and his staff. Other m em bers of the Com m ittee may participate and are informed of the daily plan by internal m em o or wire. The directives issued by the Committee and a summary of the reasons for Committee actions are published in the "Record of Policy Actions of the Federal Open Market Com m ittee.” The “Record” for each m eeting is released a few days after the Com mittee meeting. Soon after its release, it appears in the F e d e ra l R eserv e Bulletin. In addition, "Records” for the entire year are published in the annual report of the Board of Governors. The "Record” for each meeting during 1986 included: 1) a staff summary of recent econom ic develop m ents — such as changes in prices, employment, industrial production and com ponents of the na tional incom e accounts — and projections of gen eral price, output and em ploym ent developments for the year ahead; tee often evaluated potential changes in policy on the basis of the recent growth of these aggregates, among other factors. For example, in April, the Committee viewed the growth of M2 and M3 at rates within their target ranges as evidence that the rapid growth of M l did not represent an excessive buildup of liquidity .18 At later meetings, when M2 and M3 were near the upper limits of their ranges, m em bers expressed m ore con cern about the im plications of accom m odative policy for inflation .19 Interest Rates The Com m ittee also evaluated potential changes in its policy stance during 1986 in the context of recent movements in interest rates. By im plem enting an ac commodative policy in 1986, the Committee sought to provide the reserves necessary to support increases in money dem and .20 At two meetings, though it felt that 2 ) a summary of recent international financial developments and the U.S. foreign trade balance; 3) a summary of open-market operations, growth of the m onetary aggregates and bank reserves and m oney m arket conditions sin ce the previous meeting; 4) a summary of the Com mittee's discussion of the current and prospective econom ic and finan cial conditions and the current policy consider ations, including money market conditions and the movement of m onetary aggregates; 5) decisions of the Committee; 6 ) a policy directive issued by the Committee to the Federal Reserve Bank of New York; 7) a list of the m em bers’ votes and any dissenting com m ents; and 8 ) a description of any actions regarding the Com m ittee’s other authorizations and directives and reports on any actions or consultations that may have occurred between the regularly sch ed uled meetings. less restrictive reserve conditions were desirable, the Committee contem plated that the easing would be achieved through discount rate reductions by the Board. This was expected to facilitate a market ten dency toward lower interest rates .21 Moreover, the appropriateness of potential interm eeting adjust m ents in the provision of reserves were viewed as conditional on near-term changes in rates, including potential discount rate reductions. While it sought to prom ote easier credit market conditions, the Committee also took account of move m ents in long-term interest rates as indicators of changes in inflationary expectations. For example, an increase in interest rates before the Septem ber m eet ing, together with some other developments, was in terpreted as a sign of an increase in expected inflation. This was one factor that led to a vote for “maintaining unchanged conditions of reserve availability, ” and one 1 See the Record (September 1986), pp. 649-50. 8 ,9See the following Records: November 1986, pp. 784-85; January 1987, pp. 34-35; February 1987, pp. 120-21. “ See Report (September 1986), pp. 612-13. 2 See the following Records: October 1986, p. 710; November 1986, 1 p. 784. The same point is discussed in Record (June 1986), pp. 411-12. 21 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS m em ber believed that less ease in policy im plem enta tion was desirable at that tim e .22 International D evelopm ents Concerns about the external sector also influenced the FOMC's policy actions in 1986. The foreign ex change value of the dollar had declined substantially on a trade-weighted basis since early 1985. As a result, the trade deficit was expected to decline during 1986, but it widened instead and was a continuing elem ent of uncertainty in the econom ic outlook. Throughout the year, the Committee expressed doubts about the timing and magnitude of any improvement in the trade balance because it believed that foreign pro ducers would attem pt to m aintain their U.S. market shares by reducing their profit margins ,a The Commit tee also felt that increases in U.S. net exports would be hindered by sluggish growth in the econom ies of some m ajor U.S. trading partners in the absence of more stimulative policies abroad .24 While the falling dollar improved the prospects for a sm aller trade deficit, Committee m em bers noted that it also had a potential inflationary im pact .25 Further more, the trade deficit implied large capital inflows from abroad that were financing dom estic econom ic activity. Loss of confidence in the dollar might require sharp increases in dom estic interest rates to maintain the inflow of foreign capital; m em bers of the Commit tee often m entioned this risk to econom ic expansion .26 In general, the Com m ittee felt that greater caution should be exercised in providing reserves in the event of sharp dollar depreciation during interm eeting periods .27 “ See the Record (January 1987), p. 34. ^Mann (1986) presents evidence in support of this view. See Report (April 1986), p. 216, Report (September 1986), p. 610, and the following Records: June 1986, p. 409; July 1986, p. 480; January 1987, p. 33. 2 See Report (September 1986), pp. 604-05 and p. 610, and the 4 following Records: July 1986, pp. 480-81; October 1986, p. 707; January 1987, p. 33. Also see the Record in Press Release, Federal Reserve Board of Governors, February 13,1987, p. 7. There were, in fact, limited moves toward expansionary policies in some major industrial nations during the year, including coordinated rounds of discount rate reductions with the Federal Reserve. 2 See the following Records: June 1986, p. 409; November 1986, pp. 5 783-84; January 1987, pp. 33-34; February 1987, p. 120. “ See the following Records: June 1986, p. 412; October 1986, p. 710; November 1986, p. 784. Also, see Report (September 1986), p. 641, for a brief discussion of the risks of sharp depreciation of the dollar on foreign exchange markets, and how those risks influenced the Board’s decisions on discount rate reductions. 2 See the following Records: June 1986, p. 412; October 1986, p. 710; 7 November 1986, pp. 784-85; January 1987, p. 35; Press Release, February 13, 1987, p. 11. 22 Real Activity and Prices Finally, the FOMC’s decisions in 1986 were guided by available data on production, employment, wages and prices, and projections of econom ic activity based on recent developments. Consum er prices rose by only 1.1 percent in 1986, the lowest rate of inflation by this m easure since 1965. Steep declines in oil prices early in the year contributed to the low rate of in flation, but wage and price pressures were otherwise quite m oderate. Monthly data on production and em ployment and the growth of real GNP indicated, over m uch of the year, that the expansion was proceeding at a slower pace than the Com mittee had anticipated .28 Throughout the year, the m em bers generally antici pated more rapid real growth and inflation in later quarters. Nevertheless, current indications of sluggish growth, m oderate price pressures and downside risks in the econom ic outlook weighed heavily in the Com m ittee’s decisions to m aintain the accommodative stance of policy .29 The Com m ittee's views on the desir ability of interm eeting adjustm ents in policy im ple m entation also depended upon indications of the pace of econom ic growth and inflation, among other developments ,30 SHORT-RUN POLICY O BJECTIVES The FOMC m eets eight tim es each year to review econom ic developments and discuss the status of policy and its im plem entation. At each meeting, the Com mittee's decisions are sum m arized in a directive issued to the Federal Reserve Bank of New York. The directive is then used by the Manager for Domestic Operations, System Open Market Account, to guide the day-to-day im plem entation of m onetary policy through open-m arket operations during the intermeeting period. Becent policy directives specify the Com m ittee’s decision about the appropriate "degree of pressure on reserve positions” of depository institutions for the “ Real GNP grew 2.1 percent over the IV/1985— IV/1986 period. At the February 1986 meeting, the forecasts of real GNP growth of the members of the Committee and Federal Reserve Bank presidents had a central tendency of 3 to 31 percent. See the Record (June /2 1986), p. 408. At the July 1986 meeting, the forecasts for the year had a central tendency of 21 to 3 percent. See the Record (October fe 1986), p. 706. ^For example, see the following Records: June 1986, p. 411; Sep tember 1986, p. 650; October 1986, p. 710. “ For example, see the following Records: June 1986, p. 412; July 1986, p. 482; September 1986, p. 650; February 1987, pp. 120-21; Press Release, February 13,1987, p. 11. FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS Table 3 FOMC Short-Run Operating Specifications Meeting date Expected growth rates Target period M1 M2 M3 Intermeeting federal funds range Degree of reserve pressure December 16 -1 7 ,1 9851 November 1985March 1986 7-9% about 6-8% about 6-8% 6-10% decrease somewhat February 1 1 -1 2 ,19862 November 1985March 1986 about 7 about 6 about 7 6-10 unchanged April 1,1986 March 1986June 1986 about 7-8 about 7 about 7 6-10 unchanged May 2 0 ,19863 March 1986June 1986 about 12-14 about 8-10 about 8-10 5-9 unchanged July 8 - 9 , 19864 June 1986September 1986 not specified5 about 7-9 about 7-9 4-8 decrease somewhat August 1 9 ,19866 June 1986September 1986 not specified about 7-9 about 7-9 4-8 decrease slightly September 2 3 ,19867 August 1986December 1986 not specified 7-9 7 -9 4-8 unchanged November 5, 1986 September 1986December 1986 not specified 7 -9 7-9 4 -8 unchanged December 15-16,1986 November 1986March 1987 not specified about 7 about 7 4-8 unchanged 'Mr. Black dissented because he felt that a decrease in the degree of reserve pressure was undesirable, given the rapid growth of M1. 2Mr. Martin and Ms. Seger dissented because they believed that the risks to the economic expansion would be lessened by reductions in short-term interest rates, including an eventual reduction in the discount rate. They favored some easing of reserve conditions in order to facilitate these reductions. 3Mr. Wallich dissented because he was concerned about the inflationary implications of rapid monetary expansion and felt that open-market operations should be directed toward somewhat greater restraint. “Mr. Melzer dissented because he felt that easing under current circumstances could have an adverse impact on inflationary expectations and lead to an undesirably sharp depreciation in the value of the dollar on foreign exchange markets. He also noted that the outlook for the quarters ahead appeared to be consistent with the economy's long-run growth potential, and that further ease, in his view, would generate inflationary pressures without encouraging much faster growth in real output. He therefore favored maintaining the existing degree of reserve pressure. 5 The directive stated that "While growth in M1 is expected to moderate. . . that growth will continue to be judged in the light of the behavior of M2 and M3 and other factors.” See Record (October 1986), p. 711, and subsequent Records referenced in the footnotes to the text. 6Mr. Melzer dissented because he was concerned that further ease might heighten inflationary expectations and put excessive downward pressure on the foreign exchange value of the dollar. He also felt that the prospects for economic growth had improved during the intermeeting period. Mr. Wallich dissented because he felt that policy should be directed toward slowing the growth of the monetary aggregates and reducing the potential for inflation. Mr. Wallich and Mr. Melzer preferred to maintain the existing degree of pressure on reserve positions. 7 Mr. Wallich dissented because he believed that a slight tightening of reserve conditions was desirable in light of the persistence of rapid monetary growth and the threat that it presented to continued price stability. period before the next m eeting .31 In addition, each directive gives the Com m ittee’s expectation of the growth of various m onetaiy aggregates for som e shortrun target period, contingent upon the stated degree 3 See the directives contained in any of the Records of Policy Actions 1 referenced below. of reserve pressure, and specifies an interm eeting range for the federal funds rate. If reserve conditions proved to be inconsistent with a federal funds rate in that range, the chairm an could call for a Committee consultation before the date of the next meeting. Table 3 shows the Com m ittee’s expectations for m oney growth at each m eeting in 1986. The table also shows the interm eeting federal funds range established at 23 FEDERAL RESERVE BANK OF ST. LOUIS each m eeting and the degree of pressure to be applied to reserve positions. February M eeting The econom ic data reviewed at the February m eet ing indicated a m oderate and, perhaps, improving rate of econom ic growth. Real GNP grew at an annual rate of 2.4 percent in IV/1985, w hich was slower than its 3 percent growth in the previous quarter. Several indica tors of real activity, however, such as industrial pro duction, housing starts and new orders for nonde fense capital goods had shown strength in D ecem ber after performing sluggishly in prior m onths. Moreover, substantial gains in em ploym ent were reported for January. The Com m ittee discussed a num ber of uncertain ties that clouded the outlook at the February meeting. The sharp decline in oil prices in early 1986 was expected to have broadly favorable effects on the econ omy. These effects were difficult to assess, however, and energy-producing regions of the country and some oil-producing, developing countries with large debt burdens were likely to suffer. While prospective fiscal restraint associated with deficit reduction under the Gramm-Rudman-Hollings legislation was thought to have had a beneficial im pact on financial markets, it was expected to have adverse effects on aggregate demand. Some m em bers expressed concern over the strength of business investment in light of the u ncer tainties surrounding tax reform legislation, w hich was likely to tilt the com position of tax liabilities toward businesses and away from households. Despite these considerations, m em bers generally agreed that the econom y was likely to grow at a faster rate during 1986 than it had in 1985. Some of the positive factors cited were the effects of rapid M l growth, lower interest rates, higher stock prices and further declines in the foreign exchange value of the dollar. In fact, some m em bers felt that inflationary pressures might crop up several quarters ahead. The growth of M l and M2 had slowed in the weeks prior to the February meeting, and m oney growth was close to the rate expected by the Com m ittee in Decem ber for the November-to-March period. While that fact was encouraging, some m em bers were concerned about the failure of short-term interest rates to decline further in recent m onths in response to the relatively accom m odative stance of m onetary policy that had prevailed for some time. There was general accord on the desirability of im plem enting policy “in a m anner that would not in itself signal or encourage higher 24 FEBRUARY 1987 interest rates or impede the tendency for som e market rates to d ecline .”33 There was, however, a perceived risk of "a cum ulating decline in the exchange rate that might discourage willingness to hold dollars at declin ing interest rates. In these circum stances, nearly all participants agreed that little or no change in reserve availability was w arranted .”33 The Com mittee viewed its policy stance as accom modative, but som e m em bers felt that further easing might be necessary in light of the risks of a weakening econom y. In fact, “the point was made that the dis count rate might need to be reduced to permit or accom m odate a market tendency toward lower inter est rates and that such a move would be a desirable com plem ent to open market operations. .. .’’MOn the other hand, m em bers felt that the desirability of a reduction in the discount rate would depend on evolving circum stances and the prospects for similar action by m ajor foreign central banks .35 April M eeting Interest rates of all m aturities declined during the period betw een the February and April meetings, with long-term rates falling more sharply than short-term rates. The foreign exchange value of the dollar also declined on balance. At the tim e of the April meeting, however, there were conflicting signals about the pace of econom ic activity. Spending and real output were thought to have grown more rapidly during the first quarter than in the sluggish fourth quarter .36 On the other hand, growth was clearly weak in som e key sectors, and production and em ploym ent data for February were disappointing. On the bright side, de clines in oil prices dom inated the inflation outlook, and were viewed as instrum ental in lowering in flationary expectations. The Committee decided to m aintain "about the existing degree of pressure on reserve conditions .”37 This was felt to be consistent with the long-run ob jec “ Record (June 1986), pp. 411-12. “ Ibid., p. 411. Mlbid. 3 The Federal Reserve announced a reduction of the discount rate 5 from 71 to 7 percent on March 7. Subsequently, this action was /2 matched by several foreign central banks. “The Commerce Department had reported a downward revision in real GNP growth in IV/1985 to 0.7 percent. 3 Record (July 1986), p. 482. 7 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS tives for growth in the m onetary aggregates. The growth of M l had accelerated in Februaiy and March, and was near the upper end of the Com m ittee’s an nual range, but the growth of the broader aggregates had been m oderate. The Com m ittee discussed the possibility that dem ands for M l balances could grow substantially if interest rates continued to decline. Furthermore, the velocity of M l rem ained weak, and some m em bers suggested that a m ore accommodative posture with respect to m oney and reserve growth might well becom e desirable. For the weeks im m edi ately ahead, however, "m ost of the m em bers felt that there should be no presum ptions about the likely direction of any interm eeting adjustm ents .”38 ing at some point in the future .”41 The growth of the broader aggregates had been well w ithin their respec tive target ranges for 1986, however, w hich “raised questions as to w hether the growth of M l really repre sented a potentially excessive buildup in liquidity or was more of a shift in the com position of liquid hold ings in response to relative movements in interest rates .”42 The m em bers generally agreed that some slowing in M l growth was likely in the weeks ahead, but, because of uncertainties about the timing and extent of the slowdown, “som e proposed omitting num erical references in the directive to the Commit tee’s expectations for m onetaiy growth in the second quarter ." 43 This proposal was rejected by the majority. The Board announced another reduction in the discount rate of one-half point, to 6 V2 percent, effec tive April 21. The Com m ittee held a telephone confer ence on that date and agreed to make no changes in the current directive. Becent data indicated that growth in the m onetary aggregates had accelerated, however, and the m em bers felt that, in im plem enting open m arket operations, “a degree of caution should be exercised to avoid an im pression that a further change in the discount rate was sought over the p e riod immediately ahead .”39 While som e evidence of slower real growth had emerged by the time of the May meeting, a num ber of factors pointed toward more rapid growth later in the year. These factors included rapid money growth, higher prices of financial assets, lower energy prices and further depreciation of the dollar against the currencies of m ajor trading partners. In response to these elem ents in the outlook, “m ost of the mem bers indicated that they were in favor of continuing to direct open market operations at least initially toward m aintaining the existing degree of reserve availabil ity .”44 In their discussion of possible interm eeting ad justm ents, m ost m em bers em phasized a potential need for restraint in response to signals of a strength ening econom y if growth in the m onetary aggregates did not slow as anticipated. The directive stated that, under those circum stances, “somewhat greater re serve restraint would be acceptable,” but in the event of slower m oney growth and sluggish econom ic activ ity, “somewhat lesser reserve restraint might be ac ceptable .”45 May M eeting The acceleration in the growth of M l continued, and the appropriateness of using that aggregate as a guide to policy was prom inent in the discussion at the Com m ittee’s m eeting on May 20. As of early May, M l was well above the 8 percent upper limit of its target range for 1986. Some m em bers noted that the rela tively rapid growth of M l balances needed to be ac com m odated in light of the continuing adjustm ents to earlier declines in inflationaiy expectations and inter est rates and som e indications of weakness in the econom y .40 Other m em bers suggested that the rapid m oney growth might represent excessive growth in liquidity that eventually would have inflationary consequences. In this view, rapidly growing cash balances "were available to support a considerable pickup in spend- Ju ly M eeting Some optimism was expressed at the July meeting about the prospects for econom ic growth over the second half, and the outlook for inflation rem ained favorable. There was concern, however, about the ’'Record (September 1986), p. 648. “ Ibid. 42lbid., p. 649. wlbid., p. 483. “ Ibid., p. 650. “ The Commerce Department’s preliminary estimate of real GNP growth in the first quarter was 3.7 percent, but industrial production declined on balance over the three months ending in April. More over, weakness among oil producers and uncertain but potentially adverse changes in the tax code were retarding business fixed investment. ■ Ibid., pp. 651-52. In response to rapid growth in required reserves “ and currency in circulation, the limit on changes in System Account holdings of U.S. government and federal agency securities between Committee meetings was temporarily raised by the Committee from $3 billion to $9 billion on June 18. "Ibid., p. 650. 25 FEDERAL RESERVE BANK OF ST. LOUIS sluggish pace of business investment, the lack of im provement in the trade balance, and the “sharp con trasts in the econom ic perform ance of different sec tors and regions of the country and . . . strains on financial institutions that serviced the depressed industries. At the previous meeting, the m em bers had antici pated that a move toward restraint might be n eces sary. By July, the favorable inflation outlook and con cern over slow econom ic growth led m ost of the m em bers to believe ‘‘that som e easing was desirable,” which they preferred to im plem ent “at least initially, through a lower discount rate rather than through open market operations."4 The m em bers accepted a 7 directive “that called for some decrease in the existing degree of reserve pressure, recognizing that relaxation could be accom plished in the first instance by a re duction in the discount rate.”4 8 The Committee continued to anticipate a slowdown in the growth of M l, w hich had decelerated somewhat in Ju n e but was still quite rapid (see chart 1). Consider able doubt rem ained about the extent and timing of such a slowdown, however. With the growth of the broader aggregates around the midpoints of their ranges for the year, and in the context of an unex pectedly sluggish economy, the Committee mem bers agreed that rigid adherence to the original M l target was inconsistent with their objectives. Because of their uncertainty about the usefulness of M l as a guide to policy under prevailing conditions, “ m ajority of a the m em bers expressed a preference for not indica ting a specific rate of expected growth for M l in the short-run operational paragraph of the Com m ittee’s directive. ”8 4 August M eeting The rapid growth of M l continued through July and into early August. The broader aggregates also grew “ Record (October 1986), p. 707. 47lbid., p. 710. Of course, a cut in the discount rate should increase the demand for borrowed reserves by depository institutions. Under the borrowed reserve operating procedure currently used by the openmarket desk, this would result in open-market purchases of securi ties unless the borrowed reserves target is increased. See Gilbert (1985) for a discussion of the current operating procedure and two others that have been used by the open-market desk since 1970. The borrowed reserves targets used by the desk during each year are published during the following year in the Federal Reserve Bank of New York Quarterly Review. “ Record (October 1986), p. 710. On July 10, the Federal Reserve announced a 1 2-point reduction in the discount rate to 6 percent. / 49lbid. 26 FEBRUARY 1987 quite rapidly, leaving them near the upper lim its of their target ranges. At the Com m ittee’s August m eet ing, there was some concern about w hether the rapid growth in all three aggregates had inflationary im pli cations. Moreover, there had been further deprecia tion in the foreign exchange value of the dollar during the interm eeting period. The cheaper dollar was ex pected to put some upward pressure on prices, even though there were not yet any signs of the longawaited reduction in the trade deficit. There was some evidence that econom ic growth was accelerating from the weak pace of the second quarter, including strong consum ption dem and and housing activity.5 Never 0 theless, the data reviewed at this meeting continued to indicate a lack of balance in term s of growth among different sectors of the econom y and only m oderate wage and price pressures. In view of the fact that interest rates had resum ed their decline since early June, the m em bers agreed that m oney growth had not been excessive. The m em bers considered a num ber of uncertainties that continued to cloud the econom ic outlook in Au gust. These included downside risks related to the effects of tax reform legislation, rising consum er debt burdens and sluggish growth of the econom ies of several m ajor U.S. trading partners. There was also uncertainty about the course of the federal budget deficit and its im pact on the econom y. The Committee agreed that “some slight easing in the degree of reserve pressure” was appropriate, and once again stated that this “might be accom plished through a reduction in the discount rate."5 The m em 1 bers felt that an interm eeting adjustm ent in either the direction of ease or restraint might be warranted, depending on ensuing developments. It was noted, however, that in the event of a further cut in the discount rate, a significant depreciation of the dollar on foreign exchange markets would call for “a little greater caution in the provision of reserves through open market operations . . . .”5 The Board reduced the 2 discount rate by Vz percentage point to 5Vz percent effective August 21. Septem ber M eeting Short-term interest rates fell somewhat after the reduction in the discount rate, but long-term rates “ The preliminary estimate of real GNP growth in 11/1986 was 1.1 percent. 5 Record (November 1986), p. 785. 1 “ Ibid. FEDERAL RESERVE BANK OF ST. LOUIS FEBRUARY 1987 rose sharply, and by the time of the Committee s m eeting on Septem ber 23, the value of the dollar on foreign exchange m arkets had changed little. The growth rates of M2 and M3 decelerated in August but were still fairly rapid, and M l growth accelerated from its already rapid pace before slowing sharply in early September. in the trade balance rem ained elusive and, in large part, dependent upon stronger econom ic growth overseas to spur demands for U.S. exports. One mem ber referred to increasing protectionist sentim ent as a threat to real growth and price stability. In addition, tax reform legislation appeared to be deterring busi ness investment, particularly in structures. Most Committee m em bers believed that, despite an apparently stronger econom y in the third quarter, an improvement in the trade balance was critical to sus tained growth. At the same time, the m em bers felt that there were com pelling reasons for expecting some upward price pressures in the quarters ahead, includ ing the likelihood of price increases for imports and im port-com peting goods stem m ing from the decline of the dollar, and a continuing reversal of the earlier decline in world oil prices. In addition, there had been indications of a resurgence of inflationary expecta tions in financial markets and in markets for precious m etals. Because "m onetary policy had moved toward an increasingly accommodative posture over the course of recent m onths,” several m em bers believed “that it was now time to pause and observe develop ,5 m en ts___1 3 The Committee voted for “no change in the current degree of pressure on reserve positions.”5 4 In fact, while not ruling out the possibility of a move toward ease, most of the m embers believed that any potential intermeeting adjustm ent would more likely involve some restraint, depending on the behavior of a num ber of guides reflecting econom ic and financial conditions. The Committee expected inflation to accelerate somewhat over the quarters ahead because of the lagged im pacts of the dollar’s depreciation and energy price developments. On the other hand, relatively low rates of capacity utilization in m ost industries, m oder ate wage growth and continuing efforts by businesses to reduce costs and improve productivity were factors that would help to hold inflation in check. Moreover, the value of the dollar on foreign exchange markets had stabilized during the interm eeting period. If con tinued, that stability would limit a potential source of upward price pressure. N ovem ber M eeting The Committee's expectation that money growth would fall somewhat from its exceptionally rapid pace during the sum m er m onths was fulfilled in September and October: M2 and M3 advanced at annual rates of 9.3 and 8 percent over the two m onths, and M l growth slowed to a rate of 12.5 percent. Meanwhile, econom ic activity appeared to be growing at a moderate rate.5 5 At the Com m ittee’s November meeting, the m em bers saw a continuation in the moderate pace of econom ic expansion as a likely outcome, but certain aspects of the outlook were disturbing. Improvement “ Record (January 1987), p. 34. “ Ibid., p. 35. 5 Real GNP grew at an annual rate of 2.4 percent in 111/1986, accord 5 ing to the preliminary estimate, after growth of only 0.6 percent in 11/1986. “ Record (February 1987), p. 120. Given the prospects for sustained, moderate growth in econom ic activity and recent moderation in the growth of the monetary aggregates, the Committee voted for “maintaining unchanged conditions of re serve availability.”5 With regard to possible interm eet 6 ing adjustm ents, some m embers felt that an easing might be desirable in the context of indications of weakness in the economy, while others felt that money growth below the Com m ittee’s expectations should be tolerated in the absence of rising interest rates or a weak economy. The directive did not incor porate any presum ption, however, about the likely direction of any interm eeting adjustm ent in policy.5 7 D ecem ber M eeting The data reviewed at the Committee’s Decem ber meeting showed that employment growth, industrial production and consum er spending had strength ened in recent m onths. Sluggishness in business spending and the housing sector were elem ents of concern, however, and the balance of trade showed no convincing signs of improvement. To a considerable extent, the discussion focused on downside risks to econom ic growth, particularly for the early part of 1987. The earlier decline of the dollar had enhanced the international competitiveness of many U.S. firms, 5 The Committee approved a temporary increase from $6 billion to $7 7 billion in the limit on changes in System Account holdings of govern ment securities during the next intermeeting period, effective De cember 3. Outright purchases through December 1 had left insuffi cient leeway for additional purchases that would be necessary to provide for seasonal increases in required reserves and currency in circulation. 27 FEBRUARY 1987 FEDERAL RESERVE BANK OF ST. LOUIS but a num ber of m em bers expected only m inor im provement in foreign trade over the quarters ahead. The growth of consum er debt was viewed as a factor that might inhibit dom estic dem and in 1987. More over, consum ers and businesses were thought to have shifted some purchases originally planned for 1987 into 1986 to take advantage of certain provisions of the tax code that were scheduled for rescission under the new tax legislation.5 While the reduction in personal 8 tax rates for 1987 was good news for consum ers, the new tax code along with high vacancy rates had nega tive im plications for spending on multifamily housing and nonresidential construction. In their discussion of policy im plem entation, the m em bers noted that the broader m onetaiy aggregates, w hose growth had slowed in November, were within their target ranges for the year. On the other hand, the growth of M l had accelerated in November. Some m em bers felt that a continuation of the rapid growth of that aggregate, and the reserves needed to support it, carried an inflationary risk. There was a strong likelihood, however, that M l velocity would continue to decline even with some slowing in M l growth. Once again, the Com m ittee agreed that the growth of M l would be appraised in the context of the growth of the broader aggregates and other developments. Given the econom ic outlook and the fact that M2 and M3 were w ithin their long-run ranges, the Com mittee directed the desk “to m aintain the existing degree of pressure on reserve positions.”5 In light of 9 the downside risks to the economy, severed m em bers em phasized that in subsequent weeks, developments might call for some easing of reserve conditions. These m em bers noted that flexibility in the direction of ease was afforded by the recent firming of the dollar’s value. M embers recognized, however, that circum stances might call for a small adjustm ent in either direction. The Com m ittee’s deem phasis of M l as an interm e diate target and guide for policy was underscored at the Decem ber meeting. A tentative range of 3 to 8 percent for M l growth in 1987 had been reported to Congress in July as more tentative than usual. In December, a m ajority of the Com m ittee indicated that they opposed “establishing a formal target range for M l growth in 1987.” Many of those m em bers believed, however, that M l growth “should continue to be m on itored or evaluated in light of information about the economy, prices, and the broad m onetaiy aggregates and other financial variables.”6 0 CONCLUSION The FOMC deem phasized M l and placed relatively more weight on the broader m onetaiy aggregates and various econom ic and financial indicators in estab lishing its overall approach to policy and in guiding short-run policy im plem entation during 1986. A state m ent typical of the 1986 directives issued by the Com m ittee was that changes in the direction of policy im plem entation would depend “on the behavior of the aggregates, taking into account the strength of the business expansion, developments in foreign ex change markets, progress against inflation, and condi tions in dom estic and international credit m arkets.”6 1 The lengthy list of factors guiding policy underscored the Com m ittee’s desire to take a flexible approach in providing reserves in what it viewed as a highly u n cer tain econom ic environment. The recent changes in the relative weights attached to various policy guides have reflected the Commit tee’s evaluation of the im portance and reliability of these variables in influencing real growth and in flation. The accom modative thrust of policy was m oti vated by sluggish econom ic growth and a num ber of risks to sustained expansion. While the Committee was wary of inflationary risks in the outlook, price p ressu res over the co u rse of 1986 w ere w ellcontained. W hether m onetary policy can continue to provide sufficient liquidity to sustain econ om ic growth without an acceleration in the rate of inflation is a m ajor issue confronting the Federal Reserve in 1987. REFERENCES Broaddus, Alfred, and Marvin Goodfriend. “Base Drift and the Longer-Run Growth of M1: Experience from a Decade of Monetary Targeting,” Federal Reserve Bank of Richmond Economic Review (November/December 1984), pp. 3-14. Gilbert, R. Alton. “Operating Procedures for Conducting Monetary Policy,” this Review (February 1985), pp. 13-21. ________ _ “Requiem for Regulation Q: What It Did and Why It Passed Away,” this Review (February 1986), pp. 22-37. Hafer, R. W. “Much Ado About M2,” this Review (October 1981), pp. 13-18. _________ “The FOMC in 1985: Reacting to Declining M1 Veloc ity," this Review (February 1986), pp. 5-21. Keeley, Michael C., and Gary C. Zimmerman. “Deposit Rate Dereg ulation and the Demand for Transactions Media," Federal Reserve Bank of San Francisco Economic Review (Summer 1986), pp. 4 7 62. Lawler, Patrick J. “The Large Monetary Aggregates as Intermediate “ See Tax Reform Act of 1986. “ Ibid., p. 302. “ Record (April 1987), p. 304. 6 Record (February 1987), p. 122. 1 28 FEDERAL RESERVE BANK OF ST. LOUIS Policy Targets,” Voice of the Federal Reserve Bank of Dallas (November 1981), pp. 1-13. Mann, Catherine L. “Prices, Profit Margins, and Exchange Rates,” Federal Reserve Bulletin (June 1986), pp. 366-79. FEBRUARY 1987 Tax Reform Act of 1986. 1986). H. Rept. 99-841, 99 Cong. 2 Sess. (GPO, Volcker, Paul A. Statement before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate, July 23, 1986, Federal Reserve Bulletin (September 1986), pp. 635-42. Mehra, Yash. “The Recent Financial Deregulation and the Interest Elasticity of the Simple M1 Demand Function: An Empirical Note,” Federal Reserve Bank of Richmond, Working Paper No. 85-3 (1985). Walsh, Carl E. “In Defense of Base Drift," American Economic Review (September 1986), pp. 692-700. Roth, Howard. “Effects of Financial Deregulation on Monetary Pol icy,” Federal Reserve Bank of Kansas City Economic Review (March 1985), pp. 17-29. Wenninger, John. “Responsiveness of Interest Rate Spreads and Deposit Flows to Changes in Market Rates," Federal Reserve Bank of New York Quarterly Review (Autumn 1986), pp. 1-10. 29