The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Federal Reserve Bank of Atlanta Septem ber/October 1977 eral Reserve Bank of Atlanta eral Reserve Station inta, Georgia 30303 Bulk Rate U .S. Postage PAID Atlanta, G a. Perm it 292 Iress Correction Requested Barbara Turnbull Federal Reserve Bank of Philadelphia P.O. Box 66 Philadelphia, Pa. 19105 FEATURES: Banking Note: Bank Earnings Recover Slightly in 1976 . . Member bank earnings improved last year despite lower returns on earning assets. Energy Dependence and Southeastern Economic Growth: An Input-Output A n a ly s is ......................... Component Ratio Estimation of the Money Multiplier . . 103 The industrial composition of the south eastern economy suggests that rising energy costs may have a stronger impact here than in the nation as a whole. 120 This abstract of a technical working paper summarizes the results of an em pirical investigation into (1) the relation ship between the monetary base and monetary aggregates and (2) the money supply's sensitivity to interest rates. D irector of Research: Harry Brandt Expansion of the Miami Edge Act Corporations.................... Editing: Patricia Faulkinberry 112 Rapid growth of Edge Act corporations has added to Miami's international flavor in recent years, furthering the city's de velopment toward a full-service, spe cialized Latin American banking center. Productions and G raphics: Susan F. Pope Eddie W . Lee, Jr. E c o n o m ic R eview , V o l. LX II, No. 7. Free su b scription and ad d itio n a l co p ies a v a ila b le upon req uest to the R e se a rch D ep artm en t, Fed eral R ese rv e Ban k of A tlan ta, A tla n ta , G e o rg ia 30303. M a te ria l herein m ay be reprinted or ab stracted , p rovided this R e vie w , the Bank, and the au th o r are cred ited . P lea se p rovide this Bank's R ese arch D ep a rtm e n t w ith a co p y of a n y p u b licatio n in w h ic h su ch m aterial is reprinted. NEW FEDERAL RESERVE BANK OF ATLANTA WORKING PAPERS: Component Ratio Estimation of the Money Multiplier by Stuart G. Hoffman Convenience and Needs: Holding Company Claims and Actions by Joseph E. Rossman, Jr., and B. Frank King Interested parties may place their names on a subscription list for future studies in the series. For further information, copies of or subscriptions to any of the publications men tioned above, write to the Research Department, Federal Reserve Bank of Atlanta, Atlanta, Georgia 30303. Please include a complete address with ZIP Code to ensure delivery. ENERGY DEPENDENCE AND SOUTHEASTERN ECONOMIC GROWTH: AN INPUT-OUTPUT ANALYSIS by James T. Fergus This past winter's brush with energy shortages heightened our awareness of the central role played by energy supplies in econom ic growth. Recent proposals to curtail energy consumption are likely to be passed by Congress in the relatively near future (see Box 1). Suggested energy-saving measures have generally included raising the price of fuels in shortest supply (petroleum and natural gas) and providing incentives to use alternative fuels, par ticu la rly coal. How seriously would in creased energy costs a ffe ct the Southeast?1 W ould prices of some goods and services produced in the Southeast rise relative to those provided by other regions? W ould substitutes displace some of its products, dampening the region's rate of economic progress? One cannot claim to present definitive answers to these questions because of the limited inform ation presently available and the uncertainties surrounding key aspects of future energy developments. This analysis takes an initial step toward un derstanding the regional impact of energy price increases. First, the major U. S. in dustries which are heavy energy users are identified. Next, an exam ination of the relative im portance of these industries in the Southeast suggests that the region may be somewhat more vulnerable to energy cost increases than the country as a whole because our region's industries are more dependent on petroleum and natural gas and less dependent on coal. A final section sketches im plications for southeastern econom ic growth and outlines various uncertainties. ’ The Southeast is defined as those states contained either wholly or partially within the Sixth Federal Reserve District —Alabama, Florida, Georgia, Louisiana, Mississippi, and Tennessee. ENERGY DEPENDENCE IN MAJOR ECO N O M IC SECTORS Table 1 and Chart 1 show the pattern of energy usage by m ajor econom ic sectors. The lower panel in Chart 1 pictures, for seven m ajor sectors (industries) of the U. S. economy, the total cost of energy inputs expressed in cents per dollar of industry output. Energy costs relative to production value are greatest in the mining, excluding fuels, sector, follow ed by nondurable goods m anufacturing and agriculture. Durable goods m anufacturing and the combined transportation and trade category consume relatively less energy. The construction and services sectors have the lowest energy cost components. Industries' use of each of the major categories of fuels exhibits the same general pattern. Agriculture, nondurables m anufacturing, and nonfuel mining are the most intensive users of petroleum and natural gas. Transportation and trade, durables m anufacturing, and construction are moderate consumers of these fuels; the services sector again has the lowest cost component. The pattern differs somewhat for coal use, with durable goods manu facturing leading in consumption. Agriculture fa lls into the low-use category for coal inputs, accom panied by con struction, transportation and trade, and services. Nondurable goods m anufacturing and mining, excluding fuels, rely moder ately on coal. Ele ctricity is consumed most vigorously by the mining sector, follow ed by the two m anufacturing sectors and transportation and trade businesses. Services, agriculture, and construction require a relatively low volum e of e lectricity. In the follow ing section, these energy input cost measures w ill be used to assess CHART 1 CHART 2 A D JU S T E D D IR E C T AND IN D IR E C T EM P LO Y M EN T S H A R E S O F M A JO R EC O N O M IC R E Q U IR E M E N T S FO R E N E R G Y IN PU TS, S E C T O R S : U .S. AND S O U T H E A S T U N ITED S T A T E S , 1967 Input Cost in Cents, Per Dollar of Output Coal Mining - 6.0 - 4.0 - — — - U.S. Southeast 2.0 0 6.0 Petroleum and Natural Gas — 4.0 - — Electric Utilities - 2.0 6.0 2.0 6.0 - 4.0 - 2.0 Mining, excluding fuels E. Transportation and Trade Nondurable Goods Manufacturing P- Construction C. Agriculture G. Services D. Durable Goods Manufacturing A. B. SO U R C E: Computed from data contained in “The input-Output Study of the U.S. Economy, 1967: Energy Model.” See Technical Note for explanation of adjustment pro cedure. energy dependence in southeastern states. Under ideal conditions, one would use energy input measures specific to the region under study. Regional energy measures would reflect differences from the national average in the processes and e fficie n cy of industries in the region. How ever, lacking su fficien tly precise, closely com parable state data, the national in form ation has been used (see Box 2 for a description of the nature, source, and lim itations of the data). NONDURABLE GOOOS MFG. MINING. EXCLUDING FUELS CONSTRUCTION TRANSPORTATION AND TRADE SERVICES DURABLE GOODS MFG. OTHER 4.0 0 - AGRICULTURE 8 SO U R C E: Computed from data contained in Employment and Earnings, U.S. Department of Labor. REGIO N AL ENERGY DEPENDENCE Assessment of the regional im pact of energy price changes requires a device which relates national energy use data to the econom ic structure of the region. One significant measure of an industry's im portance is the share of regional em ployment it provides.2 Table 2 and Chart 2 present the shares of total employment represented by the m ajor econom ic sectors in the U. S. and the southeastern states. Combining inform ation previously presented about sectoral energy requirements with the measures of their significance in each geographical area provided by these em ploym ent shares, we can form some tentative impressions of relative energy dependence. Petroleum and Natural Gas Energy. The emerging outlines of a national energy policy suggest that users of oil and natural gas w ill face more rapidly rising costs. Employment is one of several standards which could be used to assess the relative importance of economic sectors Alternative, possibly preferable, measures include value added, wage and salary payments, and physical output. Employment has been used in this study because it is a basic determinant of regional economic activity and because recent, reasonably comparable data are available by industry and geographic division Thus, a key facto r in evaluating the potential im pact of energy costs is the degree to which m ajor industries depend on these fuels. Chart 1 showed relatively high reliance on petroleum and natural gas inputs in the agricultural, nonfuel mining, and non durable goods m anufacturing sectors. There is little difference among geographic areas in mining's share of employment. But for both agriculture and nondurable goods m anufacturing, the employment share in the Southeast is appreciably greater. This provides an initial hint that this region may be subject to a greater degree to energy-induced cost increases. The remaining m ajor difference in employment shares reinforces this interpretation, since durable goods m anufacturers, moderate users of petroleum and natural gas, are a much less important source of em ploym ent in the Southeast than nationally. To the extent that em ploym ent con centration in petroleum- and natural gas intensive industries is greater in the Southeast and its share of moderate-user industries is lower, the region is more vulnerable to energy-induced cost increases. W ithin the D istrict states, Florida's position appears most advantageous at first glance (Chart 3). Its employment shares in both agriculture and nondurable goods m anufacturing are the lowest of the six states, w hile its share is greatest in the services sector, the industry least hungry for petroleum and natural gas inputs. However, a sizable part of Florida's service business is tourism -related. Although the cost structures of these businesses may not be greatly affected by energy cost in creases, the volum e of their tourist business could easily be eroded by higher costs of auto and air transportation. The transportation and trade industries, which account for a relatively large proportion of Florida em ploym ent because of the im portance of to u rism ,3 w ill be affected even more directly. Louisiana's economy, like Florida's, has important services, transportation, and 3ln economic terminology, transportation is a "complementary good to trade and services because they are consumed together like shoes and shoelaces, left and right gloves, etc. CHART 3 EM P LO Y M EN T S H A R E S O F M A JO R EC O N O M IC S E C T O R S : S IX S O U T H E A S T E R N S T A T E S Alabama Florida Georgia Louisiana M ississippi AGRICULTURE NONDURABLE GOODS MFG. MINING. EXCLUDING FUELS CONSTRUCTION TRANSPORTATION AND TRADE SERVICES DURABLE GOOOS MFG. OTHER SO U R C E: Computed from data contained in employment releases from individual state labor departments. trade industries, portions of which are also dependent on tourism . This state would probably experience sim ilar effects from changing energy costs. It also has the highest concentration of em ploym ent in contract construction, another light consumer of petroleum and natural gas. O f course, Louisiana enjoys another great advantage — its position as a major energyproducing and processing state. TA BLE 1 A D JU S T E D D IR E C T AND IN D IR E C T R E Q U IR E M E N T S F O R E N E R G Y IN P U TS, U N ITED S T A T E S , 1967 (input co st exp ressed in cen ts per dollar of output) PRODUCING IN D U STR IES: CONSUMING IN D U STRIES. Mining, Nondurable Excluding Goods M anufacturing1 Fuels Agriculture Durable Goods Manufacturing Transportation and Trade Construction Services Total Petroleum and Natural G as: Crude Petroleum and Natural G as _____ Petroleum R e fin in g __ G a s Utilities 1.1 0.8 1.2 1.4 1.2 0.8 1.6 1.6 0.3 0.7 0.5 0.8 1.0 1.0 0.4 0.9 0.8 0.4 0.6 0.5 0.4 Coal Mining Electric Utilities 3.1 0.3 2.5 3.4 0.3 1.2 3.5 0.1 0.8 2.0 0.5 1.3 2.4 0.1 1.1 2.1 0.2 0.7 1.5 0.1 0.9 5.9 4.9 4.4 3.8 3.6 3.0 2.5 TO T A L E N ER G Y R EQ U IREM EN TS Foods, tobacco, textiles, apparel, paper and allied products, printing and publishing, chem icals, petroleum, rubber, and leather. 2 Ordnance, lumber and wood products, furniture and fixtures, primary m etals, fabricated metal products, machinery, transportation equipment, instrum ents, and stone, clay, and glass. Source: Computed from data contained in “The Input-Output Study of the U. S. Econom y, 1967: Energy Model," U. S. Dept, of Com m erce, Bureau of Econom ic Analysis, mimeo, 4 pp. See Technical Note for explanation of adjustment procedure. TA BLE 2 EM P L O Y M E N T S H A R E S O F M A JO R E C O N O M IC S E C T O R S : U. S ., S O U T H E A S T , AND IN D IVID U A L S T A T E S (May 1977) (percentage of total em ploym ent, by geographic division) A griculture Mining, Exclu ding Fu e ls' Transportation and Trade2 Durable G o o ds M anufacturing N ondurable G o o ds M anufacturing C o n struction S e rv ice s3 O ther U .S . Southeast Ala. Fla. Ga. La. M iss. Tenn. 4.8 5.6 6.9 3.3 4.5 4.4 11.3 7.6 0.3 0.3 0.3 0.3 0.4 0.3 0.1 0.5 24.2 24.5 21.5 28.6 25.3 26.2 19.4 19.9 13.4 8.7 11.2 6.2 7.5 6.3 14.5 11.6 9.5 4.5 19.2 24.1 11.1 4.8 19.1 25.9 13.9 5.1 15.9 25.2 5.6 4.7 24.0 27.3 16.0 4.2 16.9 25.2 7.5 6.6 18.9 29.8 10.8 4.6 14.6 24.7 15.7 3.7 18.2 22.8 Becau se of unavailability of data, the percentages for som e employment categories do not correspond exactly to the title of the category. Exceptions in content occu r in the following c ase s: ’ Mining, Excluding Fuels: Alabama - Mining, excluding only bituminous coal Florida - Nonmetallic minerals, except fuels Georgia and Tenn essee - Includes all mining Lo uisiana - Nonmetallic m inerals M ississippi - Mining, excluding oil and gas extraction 'Transportation and Trade: Tenn essee - Trade only; transportation is included with services ’ Services, Including Com m unications, Water, and Sanitary Services: Alabam a and Louisiana - Services, comm unications, and public utilities Florida, Georgia, and M ississippi - Services, com m unications, and electric, gas, and sanitary services Tenn essee • Includes transportation in addition to services, com m unications, and public utilities Source: Computed from data contained in Employment and Earnings (Bureau of Labor Statistics, U. S. Department of Labor) and in employment releases from individual state labor departments. The other four states have notably larger shares of industries w hich are heavily dependent on petroleum and natural gas inputs. Tennessee and Alabam a have large shares of nondurable goods m anufacturing em ploym ent and above-average shares of jobs in agriculture. The agricultural character of M ississippi's economy makes the state vulnerable to cost increases. A high concentration of nondurables m anufacturing and a large share of touristrelated transportation and trade jobs place Georgia among the more dependent states. Furthermore, the services sector is below average in im portance in each of these states. C o a l C o m fo rt? Another likely thrust of future energy policy is encouragement of conversion from petroleum-related sources of power to alternative fuels such as coal. BOX 1 CURRENT STATUS OF ENERGY LEGISLATION r The National Energy Act, with most of President Carter's proposed energy pro gram intact, was passed by the House of Representatives in early August. The most important provisions affecting in dustrial energy costs are outlined below. Since Senate com m ittee hearings have just begun, these programs remain sub ject to substantial alterations. 1. Crude O il Prices The controlled price of dom estically produced oil sold to refiners would be increased by a tax to be applied in three steps. By 1980, the price would reach a level equal to the uncontrolled price of crude oil sold in the international market. The tax would term inate on September 30, 1981, along with the President's power to control oil prices. Incom e tax credits and other payments would o ffset the purchasing power loss w hich consumers would suffer as a con sequence of higher energy prices. 2. Natural Gas Prices Natural gas price regulation would continue. The ceiling price for newly discovered gas would rise from the cur rent level of $1.46 per thousand cubic feet to $1.75 im m ediately. In the future, the price of natural gas would corre spond to the price of the amount of dom estically produced crude oil which would yield the same amount of energy. To forestall regional gas shortages, the ceiling price would apply to gas pro duced and sold w ithin a state as w ell as to gas sold for delivery to another state. The im pact of rising natural gas prices would be fe lt prim arily by industrial users in itially. 3. Coal Conversion Penalties and Incentives New u tility and industrial plants would be prohibited from burning oil or natural gas, with some exceptions based on environm ental or econom ic con siderations. Existing utilities would be re quired to cease burning natural gas by 1990. Plants w hich are now capable of burning coal could be required to use coal rather than oil or natural gas. A system of penalty taxes would be ap plied to industrial users of oil and natural gas beginning in 1979 and to u tility com panies beginning in 1983. These taxes would increase year by year to m otivate conversions to coal power. Plants using sm all quantities of oil and gas and firm s whose m anufacturing process or product q u ality would be seriously impaired by use of other fuels could be exempt from the oil- and gasusers' tax. A com pany could credit expenditures for conversion to alternate fuel sources against its user taxes (dis qualifying the investm ent for the 10percent general investment tax credit) or it could take an additional 10-percent tax credit for investments in energy equipment. To hasten conversions, the latter option would apply only through 1982. In addition, any oil- or gas-burning boiler purchased after June 20, 1977, would no longer q u alify for the regular 10-percent investment tax credit or for depreciation at an accelerated rate. For summaries of the provisions of the National Energy Act, see The Wall Street Journal, August 8, 1977, p 4, and Congressional Quarterly Weekly Report, Vol 35, No.2 (August 6.1977), pp 1624-1625 Since the costs of coal-burning processes seem likely to fall relative to oil and natural gas, an effort to assess potential price effects must consider the reliance on coal energy by major sectors.4 4This analysis does not include the influence of fuels used to generate elec tricity, which is derived from a variety of primary energy sources Relative to the U S., the Southeast derives a much higher share of its electric power from coal and a much lo w e r proportion from natural gas. The shares of electricity supplied by oil and nuclear power are about the same regionally as nationally, while the national percentage of hydroelectric power is significantly greater On balance, smaller cost increases in electric power are likely to be ex perienced in the Southeast than in the nation as a whole See Table VI, pp 18-19, in 1977 Annual Electric Power Survey, published by the Edison Electric Institute. A lower price for coal relative to petroleum-related fuels would favor areas where coal-using industries are con centrated. We can be fa irly sure that coalburning fa cilitie s are concentrated in the older m anufacturing centers of the North east and M idwest. The "n e w " centers in the South, Southwest, and W est grew in the postwar period of inexpensive, readily availab le, e fficie n tly burning natural gas. Therefore, the change in relative energy cost in favor of coal w ill probably penalize m anufacturers in these regions and BOX 2 KEY TO THE ENERGY-DEPENDENCE RIDDLE Input-Output Data. How does one spot a heavily energy-dependent industry? "Inputoutput" studies provide a key to this prob lem in the form of a detailed "shopping list" for each major sector of the economy. These data indicate the "ingre dients" required for each sector's produc tion, including the value of key raw m aterials obtained from other industries, labor compensation, profits, and taxes. One variety of input-output data shows the value in cents of each input directly con sumed to produce one dollar's value of output for each industry.1 A second variety of "shopping list" shows the value in cents of the direct input requirements plus the indirect requirements generated by a onedollar increase in output for each industry. The in d u stry energ y re q u ire m e n ts presented in Table 1 and Chart 1 of the ac com panying article are based on such input-output data. Recency and Other Reservations About the Data. Although the input-output in form ation seems well suited for an in vestigation of energy dependence, it has been used with some reservations. First, the data are not very current. The most recently published data are for the 1971 calendar y e a r.2 Furthermore, even older data have been used for this study. A ’ For a useful description of input-output data, see "The Input-Output Structure of the U .S . Economy, 1967, Survey of Current Business, Vol 54, No 2 (February 1974), pp 24-56. zPaula C. Young and Philip M. Ritz, ' Input-Output Table of the U. S. Economy, 1971" (Bureau of Economic Analysis Staff Paper No. 28), U. S. Department of Commerce, March 1977 special tabulation of the 1967 input-outpi data is available, w hich offers two majc advantages: First, it provides greater deta for energy-producing industries; second, reduces the degree of detail for nonenerg industries by aggregating them into majc econom ic sectors.3 Use of ten-year-old data creates som risks, of course. Changes in technology prices, demand patterns, and product mi may have significantly altered the patter of industrial input use since 1967. A second reservation concerns th representativeness of the data. The inpul output numbers discussed below are broai averages w hich apply to the entire Unite< States. However, conditions w ithin pai ticu lar regions, states, industries, and firm may cause input-output patterns to diffe sharply from these average values. On would expect fa irly wide variations be tween areas in industry composition and i the e fficie n cy of particular firm s. But i the absence of su fficien tly precise stat and regional data, the national inform atio supplies a useful indication of sectore energy dependence. In discussing thes numbers, we do not wish to imply tha they are typical of all areas.4 3' The Input-Output Study of the U. S Economy, 1967: Energy Model,” U. S. Department of Commerce, Bureau of Economic Analysis, mimeo, 4 pp. ‘ Although input-output studies have been prepared for some states, most do not provide a sufficiently detailed breakdown of energy usage to permit the type o analysis pursued in this study. Also, studies for particular states are usually not comparable. See W illiam A. Schaffer, Eugene A. Laurent, and Ernest M. Sutter, Using the Georgia Economic Model, Atlanta, Ceorgia, Georgia Institute of Technology, 1972. enhance the com petitive positions of the coal-burning fa cilitie s of the Northeast and Midwest. However, use of national data on industry coal requirements tends to ob scure these regional differences. As Chart 1 indicated, coal is used most intensively in the durable goods m anufacturing sector. But the share of em ploym ent provided by durable goods producers is significantly lower in the Southeast. Industries w hich use coal to a moderate degree include mining and nondurable goods m anufacturing. As noted D IRECT AND IN D IREC T ENERGY CONSUMPTION BY S T E E L M ANUFACTURERS Adjustment to Eliminate Understatement r Overstatement. The input-output data la t appear in Table 1 and Chart 1 have een adjusted to avoid understating or verstating energy requirements. The steel idustry provides a convenient example of le need for such adjustm ents. Steel production processes use large uantities of both coal and e lectricity, 'mitting other energy sources for the molent, how would one properly represent the eel industry's energy dependence? One Duld sum the values of the direct rejirem en ts for coal and ele ctricity (1 and 3 i the diagram). Although the value of coal >ed to generate e le ctricity (2) is included in le value of the e le ctricity used in steelmakg (3), this procedure would result in a previously, mining's em ploym ent shares vary little, either between the Southeast and the nation or among the southeastern states. Nondurables is a sector of con centrated em ploym ent in the Southeast; but this industry's moderate use of coal only serves to soften somewhat the disadvantage of its heavy reliance on petroleum-related energy. R elatively heavy concentrations of coalusing industries slightly improve the posi tion of some southeastern states. Tennessee and Alabam a, heavily dependent serious underestim ate of energy consump tion, since coal is also used to produce numerous other "ingredients" for steelmaking. These inputs are represented in the diagram by shipping, w hich is assumed to be co al-po w ered. M easurin g the energy dependence of steel m anufacturing by sum ming only its direct energy inputs would om it the energy content of these nonenergy inputs (4). Then, why not add together the direct and indirect requirements for energy inputs? This a p p ro a c h w o u ld overstate energ y dependence because of double counting. In the preceding diagram, steel m anufacturers' direct and indirect consum ption of coal (1, 2, and 4) would be included. But the value of coal used to generate e le ctricity (2) would be counted a second time as part of the value of the e le ctricity input directly con sumed in steel production (3). The direct and indirect input coefficients (cents per dollar of output produced) for each m ajor energy source have been modified to elim inate this source of overstatem ent. In sim ple term s, the m odification employed here takes the sum of input flows (1) and (4) as the measure of coal input dependency and classifies seg ment (3) as the measure of dependence on e le c tric ity .5 That is, the value of energy used to produce energy is measured in its final form as a direct input. sFor a more detailed explanation'of the adjustment method used, please consult the technical note at the conclusion of this article on petroleum-related inputs, also have greater-than-average employment shares in durable goods and nondurable goods m anufacturing. M ississippi, the D istrict state with the largest durable goods job concentration, should benefit to some extent from any shift of energy prices in favor of coal. Despite the potential benefits to coal users, the present role of coal inputs is minor compared to that of petroleum and natural gas energy sources. In the m anufacture of durable goods, the value of the coal used to generate one dollar of output is one quarter of the cost of petroleum and natural gas required. For the other industries, it amounts to only one-tenth or less. Unless the mix of fuel consumption is altered m arkedly, coal price incentives w ill cushion the im pact of increasing energy costs only slightly. IM PLICATIO N S FOR SOUTHEASTERN GROW TH Although producers throughout the nation w ill face increasing energy costs, incipient changes in energy prices and use patterns w ill probably have a stronger im pact in the Southeast than in the nation as a whole. H eavy use of petroleum and natural gas w ill be discouraged, but southeastern industries are more depen dent on these fuels than are their national counterparts. Fuel consumption w ill be shifted to nonpetroleum sources, especially coal; and with the exception of .-T E C H N IC A L NOTE-------------------As shown in the example in Box 2, the problem of double counting arises when energy sector A uses inputs from energy sector B, resulting in an overlap of direct consumption of sector A's product with indirect consumption of sector B's product. The key to identifying these areas of duplication is Table T-1, w hich gives the direct and indirect energy input requirements of the five energy-producing sectors. The procedure used in this study to adjust for double counting began by summing the coefficients for the inputs provided by the other four energy sectors to the particular sector under con sideration (calculating a total for each colum n of Table T-1). Thus, adding the requirements of the coal mining industry from the crude petroleum and natural gas, petroleum refining, electrical u tility, and gas u tility industries provided a sum for coal mining. This sum indicates the relative importance of other energy inputs, directly and indirectly required, in the production of coal. Note that this fraction, about 6 percent in the case of coal mining, measures the extent to which another sector's direct consumption of coal would overlap the indirect consumption (via direct use of coal) of the other four forms of energy. That is, for each $1 of coal supplied to coal-consuming industries, approxim ately 6 cents would be counted again as energy inputs obtained indirectly by those industries from other energyproducing sectors. The adjustm ent required to correct this overlap is to reduce the value of coal inputs by about 6 percent. The exact facto r is obtained by subtracting the sum of the energy input co efficients to the co; industry from 1.0 (see Table T-1). In this case, about 94 percent of the value of co. inputs supplied to other industries is not counted w ithin other energy sector input coefficients. In each of the remaining colum ns of the table, an identical procedure is follow ed to obtain the ad justm ent factors shown on the bottom Iin< of the table. This explains how the adjustm ent facto were derived; but how were they applied? For each of the five energy sectors, the original requirement co efficie n t for each input (shown in Table T-2) was m ultiplied by the adjustm ent facto r for that input (a given in Table T-1). The resulting adjusted direct and indirect input requirements are presented in Table 1 of the article. Thus, for coal mining inputs, the agricultural sector's co efficien t was changed from .00125 to .00117 (= .9373 x .00125) and th durable goods m anufacturing sector's co efficie n t was altered from .00549 to .00515 (= .9373 x .00549), etc. enter new markets, offer new or improved products, and increase production ef ficie n cy via investm ent is hampered, at least tem porarily. For some sm all-scale producers, coal conversion may be so costly as to be unprofitable and d ifficu lt to finance. W ithout the conversion option, higher prices of fuels presently used could force them to curtail operations. Thus, a number of uncertainties cloud a definitive conclusion concerning the ef fects of energy costs on the outlook for southeastern econom ic growth. The most certain aspect of the outlook is that considerable turbulence is in store before adjustments to the new realities of energy supplies and prices are completed. ■ electrical power generation, coal-burning fa cilitie s are relatively scarce in the Southeast. As their production costs begin to in crease, relatively heavy users of petroleum and natural gas w ill face an unappealing choice. They can attem pt to absorb rising energy costs by controlling other costs, increasing prices, or sacrificing net income. O r they can undertake m ajor capital in vestment programs to convert to alter native fuels, prim arily coal. But tax in centives for coal conversion would only partly offset the additional financing costs incurred. Such investments would absorb capital that could be invested in new or expanded fa cilitie s. Producers faced with conversion may find that their ab ility to TA BLE T-1 E N E R G Y S E C T O R S : D IR E C T AND IN D IR E C T IN PU T R E Q U IR E M E N T S , 1967 CONSUMING IN D USTRIES: Coal Mining Crude Petroleum and Natural G as Petroleum Refining Electrical Utilities G as Utilities PRODUCING IN D USTRIES: Coal Mining - .00124 .00162 .06684 .00110 .00960 .01800 .03062 .00448 - .49695 .00654 .01264 .00719 - .02111 .02104 .01476 .02234 .28325 .00435 .00818 .06191 - .06270 .02761 .53567 .17090 .29688 .93730 .97239 .46433 .82910 .70312 Crude Petroleum and Natural Gas Petroleum Refining Electrical Utilities G as Utilities - Sum of Energy Input Coefficients 1.0 -(Su m of Energy Input Coefficients) Source: “The Input-Output Study of the U. S. Economy, 1967: Energy Model" and computations. TA BLE T-2 D IR E C T AND IN D IR E C T R E Q U IR E M E N T S P E R D O L LA R O F D E L IV E R Y TO F IN A L D EM AN D CONSUM ING IN DUSTRIES: Mining, Excluding Agriculture Fuels Construction Durable Goods Manufacturing Nondurable Goods Manufacturing Transportation and Trade Services PRODUCING IN DUSTRIES: Coal Mining Crude Petroleum and Natural Gas Petroleum Refining Electric Utilities G as Utilities .00125 .00365 .00215 .00549 .00305 .00114 .00127 .01678 .03440 .00940 .00448 .01127 .01722 .03022 .01772 .00945 .01809 .00845 .00557 .00687 .01019 .01509 .01136 .01423 .02578 .01506 .01080 .01069 .02048 .01335 .00573 .00627 .01010 .01077 .00536 Source: “The Input-Output Study of the U. S . Economy, 1967: Energy Model.” EXPANSION OF MIAMI EDGE ACT CORPORATIONS by Donald E. Baer In 1919, Senator W alter Edge successfully sponsored a Federal Reserve A ct amend ment w hich permitted banks, singly or jo in tly, to establish international banking subsidiaries outside their home states. The resulting "Edge A ct corporations" were confined to international a ctivity and required to show a minimum cap italizatio n of $2 m illio n .1 For 40 years, the provision was v irtu a lly unused. Since 1960, however, the number of Edge A ct corporations has expanded rapidly as U. S. banks have become increasingly involved in in ternational finance. By June 1977, 113 Edge A ct corporations were operating in U. S. cities. The Edge A ct expansion came in two distinct stages. In the first stage, banks and bank holding companies, prohibited from investing directly in foreign banks and corporations, established "investm ent Edges" to accom plish the same result indirectly. These Edges were generally located in the same city as the head quarters of the parent bank. Subsequent amendments in 1966 and 1970 to the Federal Reserve and Bank Holding Com pany A cts, however, elim inated some of the major advantages of these investment-oriented Edge A ct corporations by allow ing bank holding companies to invest in foreign companies under guidelines sim ilar to those governing Edges and further permitting national banks' investment in foreign banks. The second stage of Edge A ct cor poration development has been con centrated on "ban kin g" Edges, where a parent bank (or banks) establishes in ternational banking facilitie s in regional fin an cial centers outside the parent's home 1A previous study on Edge Act corporations in the Southeast was published in the Federal Reserve Bank of Atlanta's September 1974 Monthly Review. state .2 By June 1977, there were 62 of these Edges, 10 of which are in M iam i. The Miam i Edges have all been established since 1969 by parent banks in C alifo rn ia, Georgia, Illino is, M assachusetts, and New York (see Table 1), placing M iam i second only to New York C ity in the number of away-from-headquarters Edge A ct cor porations (see Appendix). f TA BLE 1 MIAMI E D G E A C T C O R P O R A T IO N S (as of A ugust 1977) A Commenced Bu sin ess • • • • • • • • C itize n s and Southern International Bank Bank of A m erica International of Flo rid a Citib ank Interam erica Irving Interam erican Bank W ells Fargo Interam erican Bank Bank of Boston International of Miami C h a se B ank International (Miami) Bankers Trust International (Miami) Corporation • Northern Trust Interam erican Bank • Morgan G uaranty International Bank of Miami 3-24-69 3-8-71 5-3-71 8-2-71 8-16-71 5-15-72 10-19-72 8-19-74 9-16-74 2-15-77 V__________________ J Latin America and the Miami Edge Corporations. Latin Am erican linkages have drawn the Edge corporations to M iam i. Its bilingual population, relative proxim ity to Latin Am erica, and heavy South and Central Am erican visitor tra ffic make it an attractive and convenient site for the Latin Am erican operations of large U. S. banks. Indeed, nearly all of the foreign loans and deposits of M iam i Edges involve Latin Am erica or the Caribbean. H isto rically, much of Latin Am erican international fin an cial transactions has been handled directly by the largest U. S. banks, p articu larly those in New York. Now M iam i banks provide not only traditional servicing of deposits but also trade and medium-term financing. In terbank com petition and increased fin an cial sophistication of international clients are stim ulating M iam i's develop ment into a full-service, specialized Latin Am erican banking center. ^ Banking" Edges are defined as subsidiaries whose aggregate demand deposits and acceptance liabilities exceed capital and surplus A ll M iam i Edge corporations operate as subsidiaries of their parent bank, with activities in M iam i dependent on their parents' network of international offices. W here the parent bank m aintains a substantial number of branches in Latin Am erica, M iam i Edge activities are necessarily coordinated with those branches and with the Latin Am erican headquarters of the parent bank. For some banks, the M iam i Edge serves as the Latin Am erican headquarters itself. O ther M iami Edges deal only with clients in specific countries, sharing Latin Am erican responsibilities with other Edge cities with Latin Am erican links (e.g., Houston, Los Angeles, and New Orleans). Edge Act Corporation Regulations3 and Effect on Activity. By law, Edge A ct corporation activities are confined to servicing the international financing requirements of U. S.-based customers and accounts of foreigners. The Edges may o ffer demand and time deposits to foreigners but not passbook savings accounts. Edge deposits are subject to reserve requirements no less than those prevailing for Federal Reserve System member banks. The reserve requirement on aggregate Edge deposits, however, can never be less than 10 percent. This reserve requirement is not a severe constraint in the case of demand deposits. Due to their size, their parent banks are subject to higher than 10-percent marginal requirement rates on demand deposits (see Table 2). Some of the Edge demand deposits may a ctu ally represent accounts previously held at the parent bank and now subject to lower marginal reserve requirements at the Edge. Tim e deposits, however, are also subject to the minimum 10-percent reserve requirement. Since Federal Reserve member bank reserve requirements on time deposits vary from 1 to 6 percent (see Table 2), the Edge A ct corporations' cost of m aintaining time deposits surpasses that of com m ercial banks. This is im portant to M iami because time deposits represented nearly half of M iam i's 1976 Edge total liabilities and capital. 3Edge Act corporations are regulated by Section 25(a) of the Federal Reserve Act and by the Federal Reserve System's Regulation K T O T A L L IA B IL IT IE S , R E S E R V E S , AND C A P IT A L A C C O U N T S , MIAMI E D G E A C T C O R P O R A T IO N S December 1976 (Mil.$) A. Demand Deposits Due to Other than Directly Related Institutions B. Tim e and Savings Deposits Due or Issued to Other than Directly Related Institutions C. Other Liabilities, Reserves, and Capital Accounts V ______________________________________________ S As with national banks, banking Edges cannot lend more than a tenth of their capital and surplus to one borrower. Since most M iam i Edges were established with close to the $2-million minimum ca p ita l ization, their initial loan lim it was approx im ately $200,000. This lim itation has played an important part in shaping the character of the Edges. Some large loans are arranged by a M iam i Edge, with participations by the head o ffice or other subsidiaries. M any Edges, however, direct larger credit requests im m ediately to head o ffices, particu larly large-scale public sector credits. Likew ise, few Edges pur chase participations from nonrelated institutions; however, they do join sub sidiaries or branches of their parent institutions in some participations. The capital and reserves of the M iam i Edges are expanding as they receive new capital infusions from parent banks and plow back earnings into reserves. This growing fin an cial base gives the Edges greater independence and enables them to make larger individual loans. Growth of Miami Edge Corporation Liabilities. Both the number of M iam i Edge A ct corporations and the level of M iam i Edge fin ancial a ctivity have expanded rapidly in recent years. Growth of liab ilities was greatest in 1974, when liab ilities, reserves, and capital grew more than 70 percent. Even the more modest 25percent 1976 growth was striking (see Table 3). Edge A ct corporation growth has in creased com petition for foreign accounts w ithin M iam i's fin ancial com m unity; local banks have also benefited from the rising level of international activity. One index of M iam i's com m ercial bank international expansion is the growth of demand and time deposits maintained by foreign banks, foreign governments, o fficia l institutions, and central banks. Such deposits increased some 85 percent between Decem ber 1973 and Decem ber 1976. M iam i's com m ercial banks in mid-1976 maintained nearly tw ice the volum e of foreign deposits of the Edge corporations and approxim ately equal amounts of foreign loans and foreign trade financing activity. Uses of Miami Edge Act Funds. The M iam i Edge A ct corporations have con sistently accepted more deposits than could be placed lo cally. Deposits and loans to parent or other related institutions have constituted from a quarter to a third of M iam i Edge asset portfolios since 1973. In June 1977, about 17 percent of M iami Edge A ct assets were deposits and loans to directly related institutions in the U. S., with another 12 percent allocated to directly related institutions abroad (see Table 4). Still, an increasing proportion of the M iam i Edge assets has been committed to loans abroad; the share had reached 40 percent by June 1977. The bulk of these foreign loans is made to firm s and in dividuals, although loans to nonrelated foreign banks and foreign public in stitutions have also expanded in relative significance. Loans to U. S.-based entities have declined in importance; in June 1977, such loans represented less than 10 per cent of Edge assets. Miami Edges have found trade financing opportunities in Flori da and the Southeast but generally have placed greater emphasis on foreign loans. Conclusion. Edge A ct corporations are increasingly performing as international TA BLE 2 FE D E R A L R E S E R V E SYSTEM M EM B ER B A N K R E S E R V E R E Q U IR E M E N T S 1 (percent of deposits) R equirem ents in Effect Ju n e 30,1977 Type of Deposit and Deposit Interval (million $) Net Demand:2 0-2............................................ 2 - 1 0 ....................................... 1 0 - 1 0 0 .................................. 1 0 0 - 4 0 0 ................................ Over 4 0 0 ................................ Time:23 S a v in g s .................................. Other Time: 0-5, maturing in 30-179 d a y s .................... 180 days to 4 years. . . . 4 years or m o re............... Over 5, maturing in 30-179 d a y s .................... 180 days to 4 years. . . . 4 years or m o re............... Percent Effective Date 7 9 ’/2 1% 12% 16% 12/30/76 12/30/76 12/30/76 12/30/76 12/30/76 3 3/16/67 3 42 V t 41 3/16/67 1/8/76 10/30/75 6 42 ’/2 41 12/12/74 1/8/76 10/30/75 'Fo r changes in reserve requirements beginning 1 96 3, see Board’s Annual Statistical Digest, 1971-1975, and for prior changes, see Board's Annual Report for 1976, Table 13. 2{a) Requirement schedules are graduated, and each deposit interval applies to that part of the deposits of each bank. Demand deposits subject to reserve requirements are gross demand deposits minus cash items in process of collection and demand balances due from domestic banks. (b) The Federal Reserve Act specifies different ranges of requirements for reserve city banks and for other banks. Reserve cities are designated under a criterion adopted effective November 9, 197 2, by which a bank having net demand deposits of more than $ 4 0 0 million is considered to have the character of business of a reserve city bank. The presence of the head office of such a bank constitutes designation of that place as a reserve city. Cities in which there are Federal Reserve Banks or Branches are also reserve cities. Any banks having net demand deposits of $ 4 0 0 million or less are considered to have the character of business of banks outside of reserve cities and are permitted to maintain reserves at ratios set for banks not in reserve cities. For details, see the Board's Regulation D. (c) Member banks are required under the Board's Regulation M to maintain reserves against foreign branch deposits computed on the basis of net balances due from domestic offices to their foreign branches and against foreign branch loans to U. S . residents. Loans aggregating $ 1 0 0 ,0 0 0 or less to any U. S . resident are excluded from computations, as are total loans of a bank to U. S. residents if not exceeding $1 million. Regulation D imposes a similar reserve requirement on borrowings from foreign banks by domestic offices of a member bank. A reserve of 4 percent is required for each of these classifications. Negotiable Orders of Withdrawal (NOW) accounts and time deposits such as Christmas and vacation club accounts are subject to the same requirements as savings deposits. 4The average of reserves on savings and other time deposits must be at least 3 percent, the minimum specified by law. Note: Required reserves must be held in the form of deposits with Federal Reserve Banks or vault cash. Source: Federal Reserve Bulletin, July 1977. banking rather than international in vestment entities; each M iam i Edge has a banking Edge perspective. The character of the Edge A ct corporations is shaped by Edge A ct regulations as well as by the local environment. Regulations necessarily lim it a ctivity to international finance. Reserve requirements and loan lim itations tied to capital accounts also affe ct the character of Edge operations. The M iami Latin Am erican linkages--both locational and demographic--have been prime in ducements for Edge A ct corporations' operations in the city. The rapid growth of international finance in M iam i, shown by Edge A ct corporations and com m ercial banks alike, is increasingly giving the city the character of a specialized Latin Am erican banking center * TA BLE 3 MIAMI E D G E A C T C O R P O R A T IO N L IA B IL IT IE S AND C A P IT A L S E LE C T E D ACCO UN TS (million $) December 1973 Foreign Individuals, Partnership s, and Corporation Demand D ep o sits. December 1974 December 1975 December 1976 June 1977 24.5 35.8 55.2 76.8 95.8 71.9 165.7 212.3 285.6 288.1 O ther Lia b ilitie s. 76.8 99.7 152.8 166.2 159.0 C ap ital A cco u n ts (Stock, S urplus, and Undivided P ro fits)____________ 28.7 47.8 52.2 59.5 88.3 . 201.9 349.0 472.5 588.1 631.2 72.9 35.4 24.5 Foreign Individuals, Partnership s, and Corporation Tim e D e p o sits____ . Total L iab ilitie s and C a p ita l_____________ A nnual Growth R a te , Total Liab ilities and Capital A cco u n ts (Percent)_____ 7.3* ‘ Sem iannual TA BLE 4 MIAMI E D G E A C T A S S E T S (percent distribution) Foreign Lo an s O ther than to Directly R elated Institutions a. To Foreign B a n k s. b. To Foreign G overnm ents, Central Banks, and International Monetary In stitu tio n s— c. O ther Lo a n s O ther than to D irectly Related In stitu tio n s. d. Sum of Foreign Lo a n s O ther than to D irectly Related In stitu tio n s__________ U. S . Lo a n s O ther than to D irectly Related In stitu tio n s______________ C u sto m e rs’ L ia b ilitie s on A cce p ta n ce s O utstand in g. December 1973 December 1974 December 1975 December 1976 June 1977 4.1 4.4 4.8 5.8 6.2 1.6 2.0 1.8 2.1 2.3 26.2 30.1 30.1 29.1 31.8 31.9 36.5 36.7 37.0 40.3 13.5 12.0 11.1 7.5 9.2 1.7 12.8 5.6 7.9 8.4 12.7 13.1 12.4 14.9 15.0 18.8 13.5 20.2 11.8 17.1 25.8 27.3 33.8 33.7 28.9 27.1 11.4 12.8 13.9 13.2 100.0 100.0 100.0 100.0 100.0 D ep o sits Due from and Lo an s to D irectly Related Institutions a. In Foreign C o un tries b. In U .S . c. Sum of D ep o sits from and Loan s to D irectly Related In stitution s O ther A s s e ts (Inclu des R eserve Rfiquirpmfintfi) V Total A ss e ts Appendix E D G E S B Y C IT Y C O R P O R A T IO N S L O C A T E D O U T S ID E H E A D Q U A R T E R S C IT Y (as of Jun e 1977) CH ICA G O HOUSTON LO S A N G E LE S MIAMI NEW YORK SAN FR A N CISC O O TH ER Allied Bank International Bank of Am erica International of Chicago Bank of America International of Texas Bank of America International of Florida Bank of Am erica Bank of California International Bankers Trust International (Midwest) Bankers Trust International (Southwest) Bankers Trust International (Pacific) Bankers Trust International (Miami) ★ Central Cleveland International Bank Chase Bank International Chicago Ch ase Bank International Houston Chase Bank International Los Angeles Chase Bank International Miami Chem ical Bank International of Chicago Citibank International Chicago * * Citibank International Houston Citibank International Los Angeles Citibank Interamerica Chem ical Bank International of San Francisco Citibank International San Francisco Citizens and Southern International Bank Citibank Overseas Investment Corporation (Wilmington) Citizens and Southern International Bank of New Orleans Connecticut Bank International * Continental Bank International (Texas) Continental Bank International (Pacific) Continental Bank International Crocker MidAmerica International Bank Crocker International Bank ★ Fidelity International Bank Bank of Boston International of Los Angeles * First Chicago International Los Angeles Bank of Boston International of Miami Bank of Boston International First Chicago International Banking Corporation First Chicago International San Francisco First W isconsin International Bank New York Girard International Bank Harris Bank V * 116 International Corporation ★ H eadquarters C ity _ J CH ICA G O HOUSTON LO S A N G E LE S Irving Trust Company International Pacific MIAMI NEW Y O RK SAN FR A N C ISC O O TH ER Irving Interamerican Bank Manufacturers Hanover Bank (Los Angeles) * Marine Midland International Corporation Mellon Bank International I Morgan Guaranty Morgan Guaranty International Morgan Guaranty International Bank of Houston International Bank of Bank of Miami San Francisco North Carolina National Bank (NCNB) International Banking Corporation Northern Trust International Bank Northern Trust International Banking Corporation Philadelphia International Bank Rainier International Bank New York Rainier International Bank Los Angeles Security Pacific International Bank * State Street Bank of Boston International United California Bank International United Virginia Bank International (Norfolk) Wachovia International Banking Corporation (New York) 6 6 *H e a d q u arters C ity 9 W ells Fargo Interamerican W ells Fargo Bank Bank International 10 (SU M TO TA L BY C ITY) 24 4 3 SIXTH DISTRICT BANKING NOTES Bank Earnings Recover Slightly in 1976 Sixth D istrict member banks took a sm all step tow ard a recovery in earnings last year. A ccording to their operating ratios, returns on equity cap ital advanced to 6.8 percent from 6.3 percent in 1975. Despite the im provem ent, the earnings rate rem ained sig n ifican tly below the 10 percent experienced in the early part of the Seventies. The earnings advance resulted from expenses rising more slow ly than revenues. The ratio of total operating incom e to total assets a ctu a lly declined from 7.59 percent to 7.46 percent. But expenses charged against that income dropped even further, from 7.05 percent to 6.84 percent of total assets. The relative decline in operating incom e reflected lower rates of return on earning assets. M em ber banks sharply increased holdings of U. S. Treasury securities, in vestm ents that earn much less than loans, to 14 percent of total assets from 9.3 percent in the previous year. A t the same tim e, the interest return on these in vestm ents averaged 6.73 percent, down from 7.1 percent in 1975. Because of the shift toward governm ent securities, such interest incom e com prised nearly 13 percent of total operating incom e, com pared to a 9-percent share in the year before. Lagging loan demand pulled loan in come down from 64.2 percent of operating incom e to 62.2 percent last year. A reduction in the proportion of earning assets accounted for by lending out weighed a slightly higher rate of return on loans. The im portance of total loans dim inished despite sustained increases in real estate and consum er loans; com m ercial and industrial loans continued w eak until late in the year. O perating expenses consumed a sm aller proportion of total operating incom e last year. Reduced interest costs fo r borrowed m oney and lower interest paym ents on de posits contributed to the savings. Average interest paid on tim e and savings deposits dropped as banks experienced inflow s of lower cost funds. This helped to counter the increased expense of additional interest-bearing deposits. Provisions for loan losses, w h ile still high, w ere sm aller last year. W age and salary expenses, the second largest bank expense, remained unchanged as a percent of total operating incom e. There continues to be considerable variatio n in earnings among banks in the d ifferen t Sixth D istrict states. W hile member banks in the D istrict part of Louisiana still lead the Sixth D istrict in earnings perform ance, their earnings declined slightly last year. Banks in A lab am a, Georgia, and the Sixth D istrict portions of M ississippi and Tennessee posted a m oderate earnings gain. However, Florida's member banks had below-average earnings of 3.5 percent, a bit less than in the previous year. Poor perform ance by Flo rid a's member banks reflects in part the large number of very sm all member banks in that state. Sm aller banks have tended to earn lower rates of return than medium- and large sized banks. W hile the sm aller D istrict banks had slightly higher operating income/asset ratios, they also spent sig n ifica n tly more for em ployees' salaries, o ccu p an cy of their fa c ilitie s, and "a ll o ther" operating expenses. M any of these types of expenses are re la tive ly fixed and in divisible, and the larger banks can spread them over a larger asset base. N early 50 percent of the Sixth D istrict mem ber banks w ith assets of less than $10 m illion failed to generate su fficie n t incom e to meet all of their expenses last year. In contrast, less than 15 percent of banks w ith total assets of $10 m illion to $50 m illion earned less than they spent. M em ber banks apparently have turned the corner on earnings and, according to prelim inary reports fo r the first half of 1977, are on the w ay back toward the higher returns of previous years. Sharply im proved earnings, however, w ill depend on banks' a b ility to expand their most pro fitable a ctiv ity , lending, w h ile reducing provisions fo r loan losses. Note. D ata based on "1976 O perating Ratios, Sixth D istrict M em ber B an ks" now a v a ila b le upon request. John M, Godfrey COMPONENT RATIO ESTIMATION OF THE MONEY MULTIPLIER by Stuart G . H offm an This article summarizes a staff analysis that may interest those in the economics and banking professions as well as others. It is more technical than the typical Economic Review article. The analysis and conclusions are those of the author. Studies of this kind do not necessarily reflect the views of the Federal Reserve Bank. The complete study is available as part of a series of Federal Reserve Bank of Atlanta Working Papers. Single copies of this and other studies are available upon request to the Research Department, Federal Reserve Bank of Atlanta, Atlanta, Georgia 30303. The notion that money growth significantly influences econom ic activity is the heart of much recent economic doctrine. A c cordingly, the Federal Reserve System selects monetary aggregate growth ranges consistent with the goals of price stability, low unem ployment, and sustained real econom ic growth. The Federal Reserve's emphasis on money supply growth necessitates research on procedures for controlling the rate of money growth. The “ m ultiplier-base" fram ework is one approach to the analysis of money stock behavio r.1 This fram ework is so called because the money stock is viewed as a m ultiple of the "m onetary base." Sp e cifi cally, the money stock (M) is related to the monetary base (B) through the following identity: M = mB The monetary base is the sum of member bank deposits at Federal Reserve Banks, member and nonmember bank vault cash, cash in the hands of the nonbank public, and a reserve adjustm ent.2 An important assumption in this fram ework is that the Federal Reserve is capable of controlling the magnitude of the monetary base. The ’ See, for instance, Karl Brunner and Allan Meltzer, "Some Further Investi gations of Demand and Supply Functions of Money, journal of Finance, May 1964, pp 240-283, and Albert E Burger, The Money Supply Process, Belmont, California, Wadsworth PublishingCompany, 1971. It is important to point out that the Federal Reserve does not use the "multiplier-base” approach to controlling money at this time. 2The reserve adjustment accounts for reserve requirement ratio changes and shifts in deposits between classes and sizes of banks over time This paper was completed prior to the recent change in the method by which the reserve adjustment magnitude is computed. See Albert E. Burger and Robert H Rasche, "Revision of the Monetary Base, Review, Federal Reserve Bank of St. Louis, July 1977. "m oney m u ltip lier" (m) is the link con necting the base to the stock of money. This study is prim arily concerned with the predictability of the money m ultiplier. The assumption of a controllable monetary base, combined with a predictable money m ultiplier, suggests that the Federal Reserve should be able to control the money stock. Predicting the Money Multiplier. To help explain the determ ination of the money m ultiplier, the D efinitional-Behavioral technique is em ployed.3 Starting with the above "m ultiplier-base" identity, the m ultiplier is defined in terms of ten component ratios w hich specify the in fluences of the behavior of the U. S. Treasury, com m ercial banks, and the nonbank public on the money stock (see Appendix). The actual values of these component ratios can be computed at any moment in time. However, the essence of the "m ultiplier-base" fram ework is that the public, banks, and U. S. Treasury have a desired value for each ratio under their control. Each desired value depends on the values of other econom ic, institutional, and policy variables. When actual ratio values d iffer from desired values, the sectors respond by taking actions to elim inate the discrepancy, bringing the ratios back toward their desired levels. After the money m ultiplier form ulas 3Albert Burger, Lionel Kalish, 111, and Christopher Babb, "Money Stock Control and Its Implications for Monetary Policy, Reprint No 72 from the Review, Federal Reserve Bank of St Louis, October 1971, p 8 have been developed (see Appendix), the D efinitional-Behavioral technique requires specifying the structural relationship between each component ratio and its causal determ inants. Each behavioral equation is then estimated with regression analysis using monthly, seasonally unadjusted data from January 1969 to Decem ber 1975.4 (This sample period encompasses most of the period since the September 1968 Amendment to Regulation D of the Federal Reserve Act, which in stituted lagged reserve accounting.) A post sample "fo re c a st" of values of each component ratio for the first nine months of 1976 was constructed using the estimated reduced form regression coefficients, the actual post-sample values of the explanatory variables, and the predicted values of the lagged ratio. The final step in the DefinitionalBehavioral method is to substitute the predicted values for each component ratio from its estimated reduced form equation into the money m ultiplier form ulas (equations (1) and (2) in the Appendix) to calcu late the predicted values for the m ultipliers associated with the narrowly and broadly defined money sto cks.5 Summary results of this final calculation for the sample and post-sample periods are presented in Table 1. Comparing predicted with actual m ultipliers shows that the model's predictive accu racy in the sample period is com parable for the narrow and broad money m ultipliers (m, and m2, respectively). The in-sample average monthly prediction error equals 0.37 percent for both m ultipliers, which is nearly equal to their respective average quarterly prediction errors. The model consistently overpredicts each m ultiplier before mid-1971 and consistently under estimates them thereafter. This result implies that monthly misses do not cum ulate nor do they offset one another. Still, the annualized prediction error for any month would be greater than the 4The specification and empirical estimation of the behavioral equation for each component ratio are not discussed in this summary article A thorough discussion of this important step in the Definitional-Behavioral method can be found in Sections II and III of the Working Paper 5The narrow money stock, M ,, equals currency and demand deposits held by the nonbank public. The broad money stock, M 2. equals M, plus time and savings deposits other than large negotiable certificates of deposit (CDs) held by the public TA BLE 1 SU M M A R Y S T A T IS T IC S F O R T H E M U L T IP L IE R P R E D IC T IO N E R R O R S 1 Period (m? - m,)/m? (mi1- m2)/ m! (percent) (percent) 0.37 0.71 0.78 0.37 0.74 0.79 0.38 0.69 0.69 0.37 0.70 0.72 0.58 0.58 0.39 1.69 1.69 0.41 S A M P L E (Monthly) ME M AE R M SE S A M P L E (Quarterly) ME MAE R M SE PO ST -S A M P LE (Monthly) ME M AE R M SE M E = Mean (Average) Error M AE = Mean (Average) A bso lute Error R M S E = Root Mean Sq uare Error 1m^ and denote the actual values of the narrow and broad money stocks, respectively, divided by the nonborrowed monetary base, m, and m2 denote the predicted values of the multipliers calculated from Appendix equations (1) and (2), respectively, using the predicted values for each component ratio. annualized error for any q u arter.6 In the post-sample period, the average monthly prediction errors for the narrow and broad money m ultipliers are 0.58 and 1.69 percent, respectively (see Table 1). For the narrow m ultiplier, the error is only slightly higher than the in-sample prediction error. However, the broad m ultiplier estimation error is nearly five times as great as the com parable in-sample average error. The model performs much less satisfacto rily for the broad than the narrow money m ultiplier in tracking post sample movements because the broad m ultiplier is more sensitive to the public's holdings of time and savings deposits relative to demand deposits (t1)--a ratio that was relatively d iffic u lt to predict in the post-sample period. The Federal Reserve would be interested in estimating the money m ultiplier because it is the connecting link between the money stock and the presumed con trollable nonborrowed monetary base. The 6For example, the narrow multiplier s monthly mean absolute error of 0 71 percent equals an annualized error of 8 5 percent while the quarterly mean absolute error of 0 69 percent equals an annualized error of only 2 8 percent '"■ ........ "\ F IG U R E 1 A CTU AL AND P R ED IC T ED NARROW M ONEY STO C K VALU ES F IG U R E 2 A C T U A L AND P R E D IC T E D V A L U E S F O R T H E BR O A D M O N EY S T O C K FOR T H E (January 1969-September 1976) (January 1969-September 1976) Seas.Adj. B i 1.$ S'~ 300 — Mi — M* - 250 - 200 r i 1969 i i 1971 i 1 1973 1 1 1975 — J ab ility of the Federal Reserve to control movements in M, and M 2 is related to the accu racy of the money m ultiplier predictions. M ultiplying the predicted m ultipliers by the actual nonborrowed monetary base produces predicted values for the narrow and broad money stocks (Mi and M 2, respectively). These predicted values can be compared to the actual values of each money stock, with the difference between the two measuring the dollar prediction error. These results are graphed in Figures 1 and 2 and sum marized in Table 2 .7 The predicted values of the seasonally adjusted narrow and broad money stocks deviate from their actual values by $1.1 billion and $2.4 billion, respectively, on both a monthly and quarterly average basis for the 1969-75 period. In the post-sample period, the average monthly Mi and M 2 prediction errors rise to $1.8 billion and $11.7 billion, resp ectively.8 The big increase in the average M 2 error reflects the large post sample (under)prediction error for the broad money m ultiplier. 7Note that the table and figures compare the actual and predicted values for the seasonally a dju sted money stocks The predicted seasonally ad|usted values were computed in the following manner The predicted seasonally unadiusted money stock (the predicted seasonally unadjusted multiplier times the actual seasonally unad|usted nonborrowed monetary base) was multiplied by the implicit seasonal factor for that month. This seasonal factor was computed by dividing actual seasonally adjusted money stock by its actual seasonally unadjusted value. 8The average percentage error between the actual and predicted money stock is. of course, equal to the average percentage error between the actual and predicted money multiplier in both periods. Multiplier Interest Rate Elasticity. The main issue in this study is the feasib ility of money stock control by the Federal Reserve, given its ab ility to determine the magnitude of the nonborrowed monetary base through open market operations. In conducting an open market purchase, the Federal Reserve induces the com m ercial banks and the nonbank public to sell U. S. Governm ent securities in exchange for reserves or demand deposits, respectively, by bidding up the price of the securities (or forcing down the yield). In an open market sale, the Federal Reserve prompts just the opposite exchange by forcing down the price of the securities (or forcing up the yield). In this analysis, the m arket yield on three-month U. S. Treasury bills (TB R ) is used as a proxy for the many different yields on government securities of varying m aturities. Estim ation of the behavioral equations for the m ultiplier component ratios revealed that certain ratios were significantly related to the Treasury bill rate.9 Thus, open market operations, undertaken to control the magnitude of the nonborrowed monetary base, nec essarily involve changes in the bill rate. Those changes, in turn, alter the values of the related ratios and, thus, the values of the money m ultipliers. Are these interest S p e c ific a lly, the h, k, t1, t2, e, and b ratios were found to be significantly related to the three-month Treasury bill rate TA BLE 3 TA BLE 2 SU M M A R Y S T A T IS T IC S FO R T H E M O N EY S T O C K P R ED IC T IO N E R R O R S (billion $) Period M, - M, \ IM PA CT AND LON G-RUN IN T E R E S T R A T E E L A S T IC IT IE S O F T H E M O N EY M U L T IP L IE R S 1 Period Im pact* Long-Run' 0.016 0.005 0.055 -0.018 0.014 0.003 0.034 -0.051 M2-M 2 Jan . 1969- April 1973 S A M P L E (Monthly) ME M AE R M SE S A M P L E (Quarterly) ME M AE R M SE P O ST -S A M P LE (Monthly) ME M AE R M SE 1.1 1.8 2.0 2.4 3.9 4.0 1.1 1.8 1.8 2.4 3.7 3.7 1.8 1.8 1.2 11.7 11.7 3.0 M E = Mean (Average) Error M A E = Mean (Average) A bso lute Error R M S E = Root Mean Sq uare Error E(m ,,T B R ) E(m2, TBR) May 1 973-D ec. 1975 E(m ,, TBR) E(m2, T B R ) ’ The elasticity coefficient of the narrow money multiplier (m,) with respect to the Treasury Bill Rate (TBR), denoted by E(m,, TBR), is defined as the p e rc e n t change in m, divided by th e p e rc e n t change in TB R . The larger the elasticity coefficient the greater the response of the multiplier to a given change in the bill rate. * Valued at sample means TA BLE 4 rate-induced changes in the money m ultipliers large enough to offset the effects of Federal Reserve policy on the monetary aggregates? To answer this question, the interest rate e la stic ity 10 of each m ultiplier was ca lcu lated over the sample period using the money m ultiplier form ulas and the behavioral equation specified for each component ratio. Likew ise, the elasticity of each money m ultiplier with respect to the Federal Reserve Discount Rate (D ISC) was calculated. These results are summarized in Tables 3 and 4 .11 The impact elasticity measures the immediate or initial response of each m ultiplier to a change in the Treasury bill or discount rate. The long-run e lasticity measures the fu ll or complete response of each m ultiplier after all subsequent adjustm ents have taken place. The results confirm the hypothesis that the narrow money m ultiplier is positively related to the bill rate and inversely related to the discount rate, although the m ultiplier's response to movements in either rate is sm a ll.12 In both the January ' “Elasticity measures the degree to which one variable (multiplier) responds to a change in another variable (bill rate). "F o r a detailed description of the calculation of the money multipliers' interest and discount rate elasticities, see Section V and Appendix 11 of the Working Paper. ’ zThe low interest and discount rate elasticities for the narrow money multiplier found in this study are consistent with the results of previous empirical studies of the money supply process. For a summary of the results of these other studies, see Table 7 of the Working Paper IM PA CT AND LON G-RUN D ISC O U N T R A T E E L A S T IC IT IE S O F T H E M O N EY M U L T IP L IE R S 1 Period Im pact* Long-Run' -0.016 -0.017 -0.059 -0.061 Jan . 1969- Dec. 1975 E(m ,, D ISC) E(ma, D ISC) ’ The elasticity coefficient of the narrow money multiplier (m,) with respect to the Federal Reserve Discount Rate (DISC), denoted by E(m,, DISC), is defined as the p e rc e n t change in m,divided by the p e rc e n t change in DISC. The larger the elasticity coefficient the greater the response of the multiplier to a given change in the discount rate. 'V alued at sample means 1969-April 1973 and M ay 1973-December 1975 subperiods,13 the narrow m ultiplier's im pact interest e lasticity is very low and even its long-run response to changes in the bill rate is very inelastic. Likewise, for the fu ll sample period, the narrow m ultiplier's initial and long-run responses to changes in the discount rate are slight. In contrast to the narrow m ultiplier, the broad m ultiplier's long-run interest rate e lasticity is negative, although also very sm all. A rise in the bill rate ultim ately 13The May 1973 suspension of the Regulation Q ceiling rate on 90-day, large negotiable CDs caused a shift in each multiplier's responsiveness to bill rate changes Therefore, each multiplier's interest rate elasticity was calulated separately for the subperiod prior to ceiling suspension (January 1969-April 1973) and the subperiod after the suspension (May 1973-December 1975). causes a significant reduction in the public's ratio of time and savings to demand deposits (t1), which leads to a sm all decline in the broad money m ultiplier. This inverse relationship tends to reinforce the im pact of changes in the monetary base on M 2. However, the im pact interest e lasticity of the broad m ultiplier in both subperiods is slightly positive. For the total sample period, the broad m ultiplier's impact and long-run discount rate elasticities are nearly equal to the com parable discount rate elasticities of the narrow m ultiplier. That r is, discount rate movements have sim ilar minor impacts on each m ultiplier. These results confirm very low interest and discount rate elasticities for the money m ultipliers, implying that feedback effects on m, and m 2 via changes in the bill rate induced by open market oper ations w ill be quite sm all. Therefore, the use of open m arket operations by the Federal Reserve to determine the magni tude of the nonborrowed monetary base in an attem pt to control Mi and M 2 w ill not induce large offsetting changes in the money m ultipliers. ■ APPENDIXThe narrow money multiplier (mi) linking the nonborrowed monetary base with the narrow money stock, Mi (currency and demand deposits held by the nonbank public), is approximated by the following formula: 1 + k (1)mi = 7 Z -------- IT----------------------------r (h + dg) + r n (t1 + t2 ) + (e — b)[h + dg + n (t1 + t2 ) ] + ck where h, n, and g are the fractions of total private demand deposits, time and savings deposits, and U.S. Government demand deposits, respectively, held in member banks; k, t1, t2, and d are the ratios of public currency holdings, consumer time and savings deposits, large negotiable CDs, and U.S. Government demand deposits, respectively, to the demand deposit component of the money stock; e and b are the ratios of member bank excess reserves and Federal Reserve Bank borrowings, respectively, to total member bank liabilities subject to reserve re quirements; and c is the ratio of currency outside member banks to public currency holdings. rd and r* are the average reserve requirement ratios against demand and time and savings deposits, respectively. They are held constant at their August 1954 levels, since the reserve ad justment magnitude included in the nonborrowed base presumably captures changes in reserve requirement ratios. The general form of equation (1) is m t = f (rd, r l, h, n, g, d, k, t1, t2, e, b, c), where d ' = component ratios determined by the behavior of the U. S. Treasury, g. e b component ratios determined by the behavior of commercial banks, and c _ h' n k component ratios determined by the behavior of the nonbank public. t1 t2 The broad money multiplier (m 2 ) linking the nonborrowed monetary base with the broad money stock, M2 (Mi plus time and savings deposits less large negotiable CDs held by the non bank public), is approximated by the following formula: 1 (2 ) m a = ~7 r d (h + dg) + 7 + + t1 ----------- r ln(t1 + t2 ) + ( e - b ) [h + dg + n (t1 + t2 ) ] + ck where all component ratios are defined above. k