The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
ECONOMIC REVIEW . i s m m s m Ms i m m k JANUARY FEBRUARY 1973 The E ffect o f Federal Reserve Membership on Earnings o f F ourth D is tric t Banks, 1 9 6 3 - 1970 A Measure O f M onetary Policy CONTENTS The E ffect o f Federal Reserve Membership on Earnings o f Fourth D istrict Banks, 196 3- 1 9 7 0 ................................... 3 A d d itio n a l copies o f the ECONOMIC REVIEW may be obtained from the Research Departm ent, Federal A Measure o f Monetary Policy . . . .19 6387, granted Reserve Bank o f Cleveland, P. O. Box Cleveland, to publication Ohio reproduce providing 44101. any credit Permission material is given. in is this Please provide this Bank's Research Departm ent w ith a Annual Index to Economic Review 1972 ............................................ 31 copy o f any published material in w hich an article is reprinted. J A N U A R Y - F E B R U A R Y 1973 THE EFFECT OF FEDERAL RESERVE M E M B E R S H IP ON EARNINGS OF FOURTH DISTRICT BANKS, 1963-1970 Marvin Phaup Considerable a tte n tio n has been focused in recent months on the issue o f membership in the Federal Reserve System. A frequently voiced view is th a t the reserve requirements associated w ith membership have a negative im pact on bank earnings. This article reports on a study made o f member and nonm em ber bank earnings in the F o u rth D is tric t fo r the 1963-1970 period. The study fo u n d no significant difference between the earnings o f member and nonmem ber banks a fte r the effects o f differences in lo cation, size, grow th rate, and asset and lia b ility com position were taken in to account. This fin d in g is co n tra ry to the view th a t membership is costly to a ll com m ercial banks, b u t i t is consistent w ith the rather stable membership experience o f the F o u rth D is tric t during the 1960's. Inasmuch as this study was concerned w ith earnings o f members and nonmembers as groups, however, i t is possible th a t fo r some in d ivid u a l banks, earnings m ay be affected either positive ly o r negatively b y membership. 3 ECONOMIC REVIEW 11 is frequently asserted th a t Federal Reserve membership adversely affects the earnings o f basis o f the experience o f F ourth D istrict com m er cial banks fo r the years 1963-1970. individual commercial banks. The cost o f member ship is though t to result prin cip a lly from the GENERAL APPROACH OF STUDY stringency o f Federal Reserve regulations regarding The objective is to measure the e ffe ct o f the level and com position o f required reserves. Federal Reserve membership on net income; i.e., That is, some state required reserve levels are lower 1 than Federal requirements. Moreover, while actual net o f all operating expenses, provisions fo r or loan losses, securities transactions, and member bank reserves must be held in vault cash taxes. Final earnings are considered the appro or as deposits w ith the Federal Reserve, many priate measure o f p ro fita b ility on the assumption state chartered th a t this is the p ro fit figure pertinent to a bank's nonmembers are perm itted to count interest-earning assets as reserves. Professor Lucille S. Mayne has offered evidence membership decision. T hat is, it is irrelevant w hether or n o t member and nonmem ber bank th a t membership does impose a burden on mem earnings happen to be d iffe re n t before securities ber examining the earnin.gs o f gains and losses, safe deposit earnings or taxes if, member and nonmember banks in Illinois during after all charges th a t are custom arily involved in 1961-1963, she concluded: banking, th e ir earnings are n o t d iffe re n t. A d d i banks.2 A fte r ...the injurious effe ct o f Federal Reserve membership on the u ltim ate measure o f p ro fita b ility , after-tax return on equity, is tio n a lly , Professor Mayne found differences in the earnings of members and nonmembers to be greatest when earnings net o f securities trans n o t restricted merely to the smallest banks b u t extends through all size classes and is statistically significant, in specific instances, among banks w ith $2 m illion to under $25 O m illio n in deposits. tors include the nature o f demand facing a bank, The present w ork subjects the "m em bership is the c o s tly " hypothesis to additional testing on the actions and taxes were compared.4 A m u ltip lic ity o f factors affects the net income o f commercial banks. Fundam entally, these fac prices the bank must pay fo r factors o f production, and the technology employed. These basic determinants o f bank p ro fita b ility are related to the size o f the bank, its location, the size and i The U. S. banking system is a "d u a l” one. Banks may be chartered by either Federal or state banking authorities. The U. S. C om ptroller of the Currency charters national banks, which are required to join the Federal Reserve System. State chartered banks may choose to be non members and, hence, to be subject to state reserve requirements. 2 Lucille Stringer Mayne, The E ffect o f Federal Reserve System Membership on the P ro fita b ility o f Illin o is Banks, 1961-1963, (U niversity Park, Pennsylvania: Pennsylvania State University, 1967). 3 Ibid, p. 132. 4 location o f its com petitors, its rate o f grow th, and the regulatory structure to w hich it is subject. In order to isolate the effect o f Federal Reserve membership on bank earnings, it is necessary to take in to account the im pact o f all other c o n tri b u to ry factors. To illustrate the im portance o f taking other influences in to account, a simple comparison o f the average earnings o f all member 4 Ibid, p. 130. J A N U A R Y - F E B R U A R Y 1973 and nonmember banks reveals th a t member earn measuring the relationship between variables. The ings are much higher. However, if one examines process enables one to obtain estimates o f the the relative size o f member and nonmembers, it is impact o f each o f the designated independent apparent th a t member banks on average are much variables on the dependent variable—in this case, larger—and larger banks have higher absolute earn bank net income. ing levels. Regression analysis also permits one to make Meaningful comparisons, therefore, require consideration be given to all causes o f statements variations in earnings. Yet, at the same tim e, one attached to an estimate o f the relationship be about the confidence th a t may be must avoid separately accounting fo r those factors tween an independent variable and the dependent through w hich membership has an effect. For variable. In particular, it is desirable to know w ith example, if is thought to reduce some specified degree o f p ro b a b ility if the im pact earnings via the im pact on cash holdings, the effect o f the independent variable differs from zero. In o f cash holdings on p ro fits should not be separated the present study, if an estimated im pact differs from the effect o f membership itself. from zero w ith 95 percent confidence, the variable membership Ideally, one w ould like to have a complete list is considered to be a statistically significant deter o f independent, d ire ctly measurable variables th a t m inant o f net income. Special attention is given to jo in tly determine each bank's p ro fit performance. the significance o f the estimated im pact o f Federal Practically, it is impossible to take all variables Reserve membership on bank earnings. th a t influence the earnings outcome in to account; some variables cannot be measured, e.g., com INDEPENDENT VARIABLES EMPLOYED m u nity goodwill towards a bank; many o f the determinants lack independence in the sense o f unilaterally causing an outcome, e.g., high salaries may be a cause or an effect o f p ro fits; and some variables are not independent o f one another but jo in tly represent some other cause, e.g., the ratios o f cash to deposits and securities to deposits may both reflect bank regulation. Necessary com promise between the ideal and practical requires Membership. Federal Reserve membership is thought to penalize earnings o f member banks in tw o ways: first, through higher reserve ratios than those required by states and second, by the provision th a t member reserves be held in non earning assets, i.e., vault cash and deposits w ith the Federal Reserve. In fact, many state adminis tered reserve requirements are higher than Federal Reserve requirements. the analyst to select a manageable number o f For example, Ohio reserve requirements at the measurable variables th a t are considered to repre end o f 1970 were 12 percent on demand deposits sent the most im p ortant determinants o f bank and 6 percent on tim e deposits. Federal Reserve p ro fit. In this study, a number o f such variables requirements fo r co u n try banks at th a t tim e were have 12 1/2 percent on the firs t $5 m illio n o f demand been chosen fo r inclusion in regression analysis. Regression is a statistical technique5 useful in 5 An in tro d u ctio n to regression that employs neither calculus nor m atrix algebra may be found in Econometrics, by Thomas H. W onnacott and Ronald J. W onnacott, (New Y o rk: John W iley and Sons, 1970) Chs. 1-3. 0 Andrew F. Brim mer, Member of the Board of Governors o f the Federal Reserve System, "The Rationalization of Commercial Bank Reserve Requirem ents," Paper pre sented to the 67th Annual Convention of the National Association of Supervisors of State Banks, A p ril 4, 1968. ECONOMIC REVIEW deposits (13 percent thereafter) b u t 3 percent on laws,” govern the establishment o f branches,7 and savings deposits and the firs t $5 m illio n o f time prescribe the organizational structure o f banks. deposits (5 percent thereafter). However, non These laws—as in the case o f state reserve require member banks many states, including the ments—lack u n ifo rm ity . During the credit squeeze D istrict states o f K entucky, Ohio, and o f 1969, fo r example, the interest rate ceiling on Pennsylvania are perm itted to hold some portion home mortgages in Ohio and West V irginia was 8 Fourth in o f reserves in the form o f earning assets. Again percent, the K entucky m axim um contract rate was using Ohio as an example, 60 percent o f reserves 8 1/2 percent, while the Pennsylvania lim it was 6 fo r tim e deposits may consist o f securities issued percent. or guaranteed by the U. S. Government. (F or a tiguous co u n ty branching, Ohio has county-w ide more detailed comparison o f the various Fourth branching, K entucky has D istrict branching law, West branching. S im ilarly, reserve requirements in effect during 1963-1970, see A ppendix.) panies Moreover, and Pennsylvania a m odified V irginia m ulti-bank are n o t allowed in permits con c o u n ty prohibits holding com Pennsylvania, West Virginia or K entucky,8 b u t have existed in Ohio Location. A bank's p ro fita b ility in the short- since 1929. run w ill be influenced by its location. D ifferent localities have d iffe re n t demands fo r bank services, both w ith respect to level and com position. The degree o f com petition among banks may also vary by m arket areas. In addition, the cost o f doing business is higher in urban regions than in rural areas. However, there is a force at w o rk that, over tim e, tends to moderate earnings stemming economy grows, changes, new fro m and banks differences in such its bank causes. As the spatial com position are established and new branches o f existing banks are opened. While the opening o f a new bank or branch requires approval of the regulatory authorities, requests fo r new charters and permission to branch tend to be concentrated in high earnings markets, where the com petitive effe ct w ill be to bank is an im p o rta n t determ inant o f the level o f bank earnings. As holdings o f assets increase, other things equal, the gross and net income derived from these assets also increases. However, it is n o t clear th a t earnings increase in d ire ct p roportion w ith bank size—especially over the wide range o f bank sizes observed in the Fourth D istrict. Econ omies or diseconomies o f scale may be encoun tered. In order to avoid the problems o f estimation th a t are encountered where such economies or diseconomies exist and to generally increase the homogeneity o f the samples, this analysis is carried o u t on fa irly narrow size groupings o f banks: Group I, total assets under $5 m illio n ; Group II, total assets o f $5 to $10 m illio n ; Group III, to ta l reduce earnings toward the norm. Differences in state regulations may also cause earnings differentials. For example, state laws th a t apply w ith equal force to member and nonmember banks specify legal interest rate ceilings—"usury 6 Bank Size. The q u a n tity o f assets held by a 7 For a study th a t examines the effects of branching on bank cost, see Frederick W. Bell and Neil B. M urphy, Economies o f Scale in Commercial Banking, (Boston: Federal Reserve Bank of Boston, 1967). g A Louisville-based bank holding company w ith tw o banks was form ed in 1925 and is exempt from the present p ro h ib itio n under a grandfather clause. assets o f $10 to $20 m illio n ; Group IV , to ta l assets Analysis (FCA) program c o n flic t w ith the assump o f $20 to $30 m illio n ; Group V, to ta l assets o f tio n th a t loans are always more profitable than $30 to $50 m illio n ; and Group V I, total assets o f investments.11 A ccording to FCA Average Banks reports, the net rate o f return on investments $50 m illio n to $100 m illio n .9 exceeded the return on loans fo r all bank size Bank Growth. The rate at which a bank's size is increasing may affect its earnings, although the categories in 1969 and 1970 and fo r tw o o u t o f three bank size categories in 1968. direction o f im pact is somewhat ambiguous. For Liability Composition. The ratio o f tim e and example, if bank grow th occurs because o f an increase in demand fo r banking services, the contemporaneous effect o f growth on earnings seems likely to be positive. The effect o f " b o o t strap” grow th at the expense o f one's com petitors savings to to ta l deposits is also tho u g h t to have a m ajor im pact on bank earnings, since interestbearing deposits are much more costly to a bank than demand deposits. 12 because o f price and service concessions, however, is lik e ly to be quite diffe re n t. Perhaps, the statistical d iffic u lty o f distinguishing these kinds In comparing the earnings o f member and o f growth account fo r the sparsity o f evidence nonmember banks, consideration w ill have to be regarding the im pact o f grow th on bank p ro fits .10 given to the effects o f location (by state), size, the Asset Composition. It is often asserted th a t the the ratios o f loans to investments and tim e and pace at w hich individual banks are growing, and make-up p o w e rfu l of a bank's asset holdings exerts a influence on its earnings. High- savings deposits to ences on total deposits. These in flu bank earnings may be accounted for performance banks are associated w ith a high proportion o f instalm ent loans (if the bank serves a retail clientele) o r commercial and industrial loans (if the bank is large enough to enter such a market). Generally, loans are thought to be more profitable than securities. Yet, accounting estimates o f net rates o f return earned on investments and loans by banks p a rtici pating in the Federal Reserve Functional Cost 11 Functional Cost Analysis, Average Banks, 1966-1970. Functional Cost Section, Federal Reserve Bank of Cleve land. The FCA estimates do not include any adjustment fo r the effect of compensating balances on net rates of return, b u t Professor Mayne, op. cit., Ch. 8, also found evidence consistent w ith hypothesis that the net rate of return on investments exceeded that on loans fo r Illinois banks in 1961 -1963. 9 Homogeneous size groupings are also useful in dealing w ith the econom etric problem of heteroscedasticity. 10 Mayne, op. c it . , Ch. 8 and Lyle E. Gramley, "G ro w th and Earnings at Individual Commercial Banks," Essays on Commercial Banking, (Kansas C ity: Federal Reserve Bank of Kansas C ity , 1962), pp. 9-19. 12 See fo r example, "T h e Cost of Demand Deposits, 1969-1970" Econom ic C om m entary, Federal Reserve Bank o f Cleveland, August 23, 1971 and "T h e P ro fit a b ility o f Time Deposits—A Functional Cost Approach," Economic Commentary, Federal Reserve Bank of Cleveland, October 25, 1971. ECONOMIC REVIEW and the im pact o f membership measured by estimating the fo llo w in g equation: located in the Fourth Federal Reserve D istrict. The fo llo w in g insured banks were excluded: ♦ those opening fo r business since 1960 ♦ those w ith incom plete data Nl. = a + a . M. + a „ O. + a~ P. + a„ K. + a r TA. i o i l 2 i o i 4 i b i ♦ those w ith to ta l assets o f $100 m illio n or more. + ac G- + a-, L /l- + aQ TS/TD. + e. b I / I o I I New banks were excluded from the analysis on where the terms are defined as follow s: Nl. = M. = O. = net income after loan losses, securities gains and losses and taxes, the i1*1 bank in thousands of dollars. an index variable fo r membership taking on the value 1 fo r member banks and 0 fo r non-member banks. a dum m y variable fo r Ohio where if a bank is located in O hio, 0 = 1 ; otherwise O- = 0. i i P. and K. sim ilarly are dummies fo r banks located in i i Pennsylvania and K entucky. West Virginia banks constitute the standard against which banks in other states are compared. TA. = the average of total assets fo r the ith bank in a given year. TA . fo r the t**1 year is calculated by adding total assets reported at the end of Decem ber t-1 to total assets at the end of December t to tw ice total assets reported in June t and dividing by four. Gj = the annual percent change in total assets fo r the it *1 bank. Gj fo r the t**1 year is calculated by subtracting total assets reported in December t-1 from total assets reported in December t and dividing the results by December t-1 total assets. the grounds th a t the earnings performance o f a new bank is substantially d iffe re n t from th a t o f an established bank. Involvem ent in a merger and a change o f main member-nonmember status were the reasons fo r incom plete data. The largest banks in the D istrict were excluded because o f the desire to restrict the regressions to rather narrow size ranges o f banks in order to maintain the homogeneity o f the sub-samples and because the small num ber o f banks w ith assets over $100 m illio n . The number o f banks in each bank group fo r each year, after these deletions, is reported w ith the regression results. REGRESSION RESULTS The c o e fficie n t estimates and summary sta tistics by year and size group are presented in Tables I-V I. The most im p o rta n t aspect o f these results fo r this study is the failure o f the analysis L / l. = the ratio o f loans to investments fo r the i**1 bank as reported on the December call report. to detect any systematic influence o f membership TS/TD . = the ratio o f tim e and savings deposits to total I +K deposits fo r the i bankas reported on the Decem ber call report. Federal e. = an error term reflecting the effect o f errors in measurement and om itted independent variables. on net earnings o f commercial banks in the Fourth Reserve D is tric t during the years 1963-1970. The estimate o f the c o e fficie n t fo r the membership variable is significantly d iffe re n t from zero at the 95 percent confidence level in o n ly 5 o f 48 cases. In tw o o f those instances, the sign o f the co e fficie n t is positive (G roup IV banks in 1964 and 1967), meaning th a t member banks had higher DATA EMPLOYED earnings than non-members. The three negative The equation was estimated fo r each year, and significant coefficients are observed fo r the 1963-1970 inclusive, fo r groups o f banks th a t tw o smallest bank-size groups in 1966 and 1967 together (Group 8 constitute nearly all insured banks II) and 1968 (G roup I). O f the 48 J A N U A R Y - F E B R U A R Y 1973 T A B LE I Net Income Regression Results fo r Fourth D istrict Banks W ith Assets Less Than $5 M illio n (Group I) 1963-1970 Membership Ohio Pennsylvania Kentucky Total assets Growth Loans/1 nvestments Time and savings deposits/Total deposits Intercept Standard error o f estimate R2 Mean o f dependent variable n 1963 1964 1965 1.596 (1.260) -4 .6 0 5 (3.147) -1 .3 6 4 (3.617) -3 .9 5 9 (3.319) .0 0 7 5 3 ** (.00049) - 7 .4 0 4 (6.851) -1 .2 3 4 (.917) .103 (1.459) -2 .8 3 5 (3.772) - .0 0 9 (4.310) -4 .8 5 7 (3.900) .0 0 80 6 ** (.00058) 22.316* (9.850) -2 .3 2 2 * (.951) -.8 9 1 (1.568) -6 .3 7 9 (4.075) 2.409 (4.664) -8 .8 2 1 * (4.123) .0 0 83 3 ** (.00064) 3.567 (11.539) -2 .1 1 4 * (.835) 1966 1967 .170 -3 .5 5 5 (1.507) (1.873) - 1 1 .8 9 7 * * 9.573* (3.875) (4.781) -7 .9 3 2 10.901 (4.524) (5.657) 6.054 - 1 0 .8 4 7 * * (3.891) (4.870) .0 0 81 3 ** .0 1 00 3 ** (.00074) (.00060) 7.133 34.051* (8.484) (13.121) - 2 .4 8 7 * * -1 .6 3 7 * (.779) (.715) 1968 1969 1970 -4 .2 3 4 * (1.740) -2 .6 4 7 (4.147) 3.931 (4.990) -6 .1 3 9 (4.343) .0 0 91 2 ** (.00067) 1.020 (10.981) - .5 1 6 (.994) - .1 5 7 (2.348) 1.708 (5.551) 10.604 (6.577) - 1 .4 0 3 (5.875) .0 1 1 9 1 ** (.00097) 6.041 (14.698) 1.077 (.888) .481 (3.899) -2 .5 4 7 (9.624) 5.614 (11.336) -1 3 .1 7 9 (10.000) .0 1 30 1 ** (.00163) 3.151 (13.991) - 5 .9 8 9 * * (1.814) -1 2 .4 4 1 ** - 1 3 .9 2 2 * * - 2 2 .0 8 7 * * - 1 5 .5 1 3 * * - 2 1 .3 4 3 * * - 2 3 .5 0 5 * * - 4 2 .0 5 9 * * —4 9 .86 8 1 (3.630) (6.964) (11.174) (4.205) (4.823) (4.361) (5.435) (4.971) 10.290 10.970 30.774 17.981 21.466 - .8 6 8 16.432 13.672 10.840 .45 11.910 .42 12.055 .43 10.887 .48 12.684 .47 11.055 .50 13.928 .52 20.560 .44 20.973 367 23.482 334 22.362 290 23.970 263 26.411 236 27.906 213 33.363 179 32.053 151 NOTE: Standard errors o f the coefficients are in parentheses. * Significant at 95% confidence level. * * Significant at 99% confidence level. Source: Federal Reserve Bank o f Cleveland estimates o f the effect o f membership, only 20 are be w ith o u t substantial empirical support. There negative. Eleven o f the 20 negative im pact cases was some irregular tendency fo r West Virginia and occur in consistent Groups I w ith the and II. is Pennsylvania banks to achieve higher earnings than Federal those located in Ohio and K entucky. The pre This evidence hypothesis th a t Reserve membership has no effect on bank earn ponderance o f negative signs o f the state dummies ings. Yet, because o f the concentration o f negative (89 o u t o f 120 cases) reflects the higher earnings signs in Groups I and II and of positive signs in o f West Virginia banks, and the size and sign o f the Groups 111-VI, the likelihood th a t membership is Pennsylvania dum m y relative to the Ohio and costly decreases w ith bank size. K entucky variables indicates the somewhat better earnings record o f banks located in Pennsylvania. THE EFFECT OF VARIABLES OTHER THAN MEMBERSHIP However, o n ly 14 o f the state variable coe fficie n t estimates are significantly d iffe re n t from zero. The Location by State.The conjecture regarding the state variables were o m itted from the Group VI effect o f diversity in state regulations was found to regressions because some Fourth D istrict portions 9 ECONOMIC REVIEW T A B LE II Net Income Regression Results fo r Fourth D istrict Banks W ith Assets From $5 M illio n to Less Than $10 M illio n (Group II) 1963-1970 Membership Ohio Pennsylvania Kentucky Total assets G rowth Loans/In vestments Time and savings deposits/Total deposits Intercept Standard error of estimate R2 Mean o f dependent variable n 1963 1964 1965 1966 1967 1968 1969 1970 -3 .6 5 6 (3.059) 6.146 (9.570) 14.015 (9.966) - .2 0 4 (9.749) .0 0 50 8 ** (.00102) 6.569 (24.758) - .6 4 0 (1.928) 1.173 (3.052) 2.594 (10.412) 12.308 (10.855) 2.618 (10.711) .0 0 56 4 ** (.00103) -4 0 .7 2 0 (24.966) - .4 9 8 (1.742) -6 .1 0 8 (3.725) 5.684 (11.369) 12.104 (12.098) 4.061 (11.763) .0 0 4 7 2 ** (.00131) -6 .5 0 5 (26.953) -3 .8 9 3 (2.222) -8 .2 6 5 * (3.933) -1 0 .6 0 9 (11.210) -1 4 .6 1 5 (12.146) -1 3 .7 2 3 (11.629) .0 0 46 0 ** (.00138) 54.794 (31.642) -2 .8 3 2 (2.865) - 8 .3 9 3 * (3.914) 12.263 (11.296) 19.773 (12.223) 3.565 (11.652) .0 0 36 9 ** (.00128) 33.865 (30.190) -5 .0 0 8 (2.943) -4 .3 8 3 (3.883) -1 5 .7 2 3 (10.910) -1 5 .3 8 7 (11.885) -1 2 .1 8 2 (10.967) .0 0 6 8 3 ** (.00132) 15.021 (31.178) -1 0 .5 8 4 * * (2.823) -3 .3 6 6 (4.378) 3.217 (12.239) 12.206 (13.745) 4.548 (12.396) .0 0 88 3 ** (.00144) -5 0 .8 3 0 (32.624) - 5 .3 2 3 (2.736) - .4 8 6 (4.238) -2 .6 9 9 (10.283) -5 .5 5 9 (12.025) -4 .0 0 2 (10.862) .0 0 9 2 9 ** (.00142) 6.057 (25.398) - 5 .5 7 9 * * (2.094) - 3 5 .0 2 5 * * - 2 8 .5 4 9 * (10.920) (11.539) 28.660 26.027 -3 1 .0 5 4 * (13.249) 36.360 -1 3 .8 3 7 (14.610) 42.941 - 3 1 .4 9 0 * (15.285) 45.678 -2 6 .7 2 7 (15.790) 56.812 - 4 8 .0 2 1 * * -5 4 .3 3 7 * (16.481) (17.146) 51.201 44.523 20.585 .18 20.237 .22 24.701 .12 26.438 .11 26.161 .11 24.969 .26 28.315 .26 27.574 .29 48.025 199 53.800 195 48.515 200 51.955 202 56.846 202 60.375 184 70.184 185 75.086 185 NOTE: Standard errors o f the coefficients are in parentheses. * Significant at 95% confidence level. * * Significant at 99% confidence level. Source: Federal Reserve Bank o f Cleveland o f states did n o t have any banks in this size range There was some tendency fo r the to ta l assets fo r some o f the years. co e fficie n t to increase w ith bank size. Total Assets. Bank size as measured by total Growth. This study finds no evidence o f a assets proved to be a highly significant deter constant relationship between bank grow th and m inant o f the level o f bank earnings w ith in size bank earnings. The grow th c o e fficie n t appears to groupings. The coefficients range fro m a low o f be highly unstable both in size and sign. The 0.3 percent (recorded by Group III banks in 1968) co e fficie n t is negative in 22 o u t o f 48 cases. to a high o f 1.3 percent (Group I banks in 1970). Moreover, the value is sig n ifica n tly d iffe re n t from Digitized for 10 FRASER J A N U A R Y - F E B R U A R Y 1973 T A B LE III Net Income Regression Results fo r Fourth D istrict Banks W ith Assets From $10 M illio n to Less Than $20 M illio n (Group III) 1963-1970 Membership Ohio Pennsylvania Kentucky Total assets G rowth Loans/Investments Time and savings deposits/Total deposits Intercept Standard error of estimate R2 Mean o f dependent variable n 1963 1964 1965 1966 1967 1968 1969 1970 3.702 (7.579) -4 4 .4 9 4 (30.362) -3 3 .5 0 5 (31.081) -4 3 .3 1 1 (31.400) .0 0 65 9 ** (.00133) -9 2 .6 3 9 (62.087) 1.005 (4.707) 11.192 (7.274) -4 8 .5 3 0 (30.756) -2 9 .7 2 0 (31.301) -3 4 .9 3 3 (31.628) .0 0 69 7 ** (.00125) -1 5 .6 6 4 (64.261) 1.265 (4.751) 10.545 (6.456) -2 8 .3 8 0 (22.575) -2 8 .6 7 0 (23.168) -3 4 .2 6 2 (23.752) .0 0 75 1 ** (.00107) .0003 (.002) -3 .0 3 3 (3.906) 9.487 (6.987) - .0 8 6 (29.358) 8.217 (30.004) -1 0 .9 5 4 (30.373) .0 0 45 7 ** (.00112) 85.755 (51.311) - 8 .8 9 2 * (4.463) 7.367 (7.202) 30.879 (28.719) 45.923 (29.498) 14.450 (29.391) .0 0 7 1 7 ** (.00119) 22.145 (46.247) - 1 1 .1 3 9 * (5.079) -1 1 .8 2 8 (19.672) -5 0 .8 8 1 (70.535) -3 4 .9 7 5 (73.231) -3 9 .4 8 5 (72.673) .00313 (.00350) 6 2 5 .0 4 1 ** (158.927) - 8 7 .3 7 6 * * (15.663) - 1 .7 3 0 (7.292) -3 .3 6 5 (24.000) 30.353 (25.209) -6 .3 8 0 (24.814) .0 0 83 0 ** (.00126) - 3 6 .1 2 5 (68.181 -1 1 .1 0 2 * (5.730) -5 .0 0 8 (6.368) 8.122 (24.509) 29.520 (25.347) -4 .9 4 9 (24.991) .0 1 1 1 7 ** (.00112) 69.267 (43.811) - 2 0 .6 7 9 * (4.915) - 7 6 .0 7 6 * * - 9 3 .9 1 1 * * - 9 3 .6 9 8 * (27.134) (27.066) (27.713) 89.998 71.123 84.238 - 6 2 .8 0 0 * (29.852) 72.910 -4 7 .5 5 2 (30.051) 13.885 67.255 (88.433) 124.161 - 1 1 5 .1 0 0 * * -5 4 .8 5 4 (33.965) (29.348) 94.866 32.567 41.635 .24 41.856 .29 38.042 .32 40.995 .20 39.908 .31 119.580 .23 46.157 .32 41.394 .48 91.201 154 102.414 162 98.686 172 101.175 171 109.547 159 96.813 166 122.153 183 138.492 187 NOTE: Standard errors o f the coefficients are in parentheses. * Significant at 95% confidence level. * * Significant at 99% confidence level. Source: Federal Reserve Bank o f Cleveland zero at the 95 percent confidence level in o n ly 6 cases (4 w ith a positive sign and 2 negative). Asset Composition. The ratio o f loans to instance (15) in which the c o e fficie n t is signifi cantly d iffe re n t fro m zero, it is also negative.This evidence is consistent w ith the hypothesis th a t the investments in bank portfo lio s appears as an often net rate o f return on securities is at times higher significant determ inant o f the level o f bank earn than th a t on loans, at least fo r banks w ith less than ings, b ut w ith a co e fficie n t sign th a t defies a $100 m illio n in to ta l assets. This may result from w idely held view th a t loans are more profitable the higher costs o f making, servicing, collecting than investments. O ut o f 48 cases, there are 33 and absorbing losses on loans. The tax-exem pt negative signs. But more im p o rta n tly, in every status o f some security holdings is another factor 11 ECONOMIC REVIEW T A B LE IV Met Income Regression Results fo r Fourth D istrict Banks W ith Assets From $20 M illio n to Less Than $30 M illio n (Group IV) 1963-1970 1963 Membership Ohio Pennsylvania Kentucky Total assets G rowth Loans/Investments Time and savings deposits/Total deposits Intercept Standard error o f estimate R2 Mean o f dependent variable n 1964 1965 1967 1966 1968 1970 1969 35.903 (25.139) -4 4 .1 8 6 (64.263) - .2 1 5 (64.844) -1 2 5 .5 9 7 (76.583) .00421 (.00425) -3 6 .4 4 3 (131.944) 12.431 (17.753) 33.604 53.590* 11.988 (17.864) (19.932) (20.361) -9 0 .6 1 9 -2 4 .4 6 0 -7 2 .6 2 9 (56.104) (52.881) (51.118) -4 4 .9 9 5 7.650 -9 .4 5 9 (56.651) (52.411) (54.576) -5 3 .4 9 5 -1 3 0 .0 3 5 * -1 0 2 .1 8 6 (63.293) (59.832) (60.665) .0 0 8 8 0 ** .00579* .00585* (.00280) (.00252) (.00327) -1 3 9 .8 4 9 -4 0 2 .7 4 8 * 290.405 (132.840) (162.309) (177.656) -2 .5 6 4 26.157 13.856 (12.298) (11.503) (18.183) -2 0 3 .1 3 7 (154.775) 162.974 -1 0 7 .3 1 8 (92.731) 104.069 -3 1 .5 6 3 (77.475) 97.315 57.198 .30 47.875 .33 45.686 .41 65.231 .38 55.406 .41 68.622 .24 75.339 .24 69.064 .36 163.195 36 170.201 40 169.978 44 180.868 53 184.177 68 181.442 86 223.138 80 240.449 87 4 2 .3 7 9 ** -2 7 .6 1 0 (14.866) (16.910) - 9 2 .7 6 5 * * -5 3 .1 4 8 (50.780) (34.926) - 8 1 .4 5 0 * -3 6 .6 7 9 (38.059) (53.412) -5 1 .0 4 4 -4 2 .8 3 7 (40.542) (54.616) .0 1 06 3 ** .00503 (.00284) (.00289) -1 0 0 .7 6 1 - 5 .0 1 9 (130.936) (111.625) 2.911 -1 5 .1 1 9 (9.878) (12.086) -3 7 .6 5 3 (19.069) -4 4 .3 2 1 (45.602) -1 2 .3 5 0 (49.137) 1.827 (51.810) .00897* (.00361) 289.184 (176.669) -1 9 .7 3 6 (13.511) 8.338 (16.693) - 8 4 .0 2 5 * (37.900) -6 2 .9 2 6 (41.746) - 1 1 1 .8 8 1 * * (42.049) .0 1 0 9 5 ** (.00278) -6 .1 0 7 (111.747) -1 8 .5 8 9 * (9.114) -1 9 8 .1 7 5 * -4 3 .0 0 3 - 2 4 5 .4 6 5 * * -9 9 .4 8 3 (85.370) (94.502) (70.612) (78.165) 22.732 300.904 139.398 79.568 - 2 6 3 .7 7 9 * * (80.153) 234.037 NOTE: Standard errors of the coefficients are in parentheses. * Significant at 95% confidence level. * * Significant at 99% confidence level. Source: Federal Reserve Bank of Cleveland producing the negative sign.13 sample correlation coefficients calculated fo r the tw o variables and shown below: 13 James W. Leonard ("Federal Reserve Membership and D iscrim ination—A Com m ent,” Journal o f Finance, X X , September 1965, pp. 483-484) and others have argued that if member banks are required to hold greater than desired ratios of cash to to ta l assets, they w ill adjust the liq u id ity of their p o rtfo lio s to the desired level by reducing liq u id security holdings and increasing their holdings o f illiq u id loans. Thus, the ratio o f loans to investments w ill be positively associated w ith member ship. If such a relationship exists one possible conse quence would be fo r the loan/investm ent variable to "p ic k u p " some o f the effect o f membership on earnings and, reduce the significance of the membership variable. However, fo r Fourth D istrict banks during 1963-1970, the relationship between membership and the loan/investm ent ratio was not strong as indicated by the 12 1963 1964 1965 1966 1967 1968 1969 1970 Group I Group 11 G roup IV -.0 3 .00 .02 - .0 2 - . 0 3 - .0 2 - . 6 3 - .0 2 .16 .03 .03 .20 .06 .02 .00 .11 .00 .13 .02 .10 .10 - . 0 4 .19 .14 Consequently, when the relationship between net income and its postulated determinants (excluding loan/investm ent) was re-estimated fo r Group I banks, almost identical results were obtained. Specifically, there were no changes in either the signs or the significance of the membership dum m y. Related to this, in some trial regressions o f operating revenue on the postulated determ inants o f net income fo r Group I banks, the membership variable was negative in each case and significant (at 5%) in 1963, 1966, and 1968. j A !MU A RY —F E B R U A R Y 1973 TA B LE V Net Income Regression Results fo r Fourth D istrict Banks With Assets From S30 M illio n to Less Than $50 M illio n (Group V) 1963-1970 Membership Ohio Pennsylvania Kentucky Total assets Growth Loans/Investments 1963 1964 1965 1966 55.332 (30.970) -4 0 .2 2 3 (63.362) -3 5 .6 1 4 (64.681) -1 5 .7 4 4 (75.890) .00571 * (.00221) -1 4 4 .4 7 9 (265.729) 22.728 (19.591) 39.623 (31.119) -1 0 8 .5 1 4 (79.058) -9 .5 6 3 (80.040) -1 5 4 .9 7 4 (91.603) .0 0 87 7 ** (.00259) -1 1 .4 9 6 (315.198) - .4 2 4 (19.771) -7 .2 4 0 (26.470) -1 2 4 .6 7 9 (65.396) -1 8 .9 4 6 (67.231) -9 2 .0 7 2 (70.192) .0 0 7 2 2 ** (.00179) -5 0 .0 8 1 (200.780) 5.176 (17.814) 13.892 (38.079) -2 4 8 .5 4 6 * (98.278) -2 1 2 .3 1 4 * (102.113) -2 7 7 .6 3 5 * (107.030) .0 0 84 5 ** (.00292) -1 1 3 .1 1 0 (382.888) 11.917 (27.869) 1967 1968 1969 9.983 51.410 -8 .7 6 8 (32.658) (45.749) (32.083) -1 3 1 .7 6 8 -1 2 2 .1 3 0 -1 2 4 .5 3 5 (75.162) (91.700) (63.346) -1 0 5 .5 7 4 -8 6 .1 0 7 -2 0 .3 1 6 (72.004) (93.939) (78.133) -1 8 0 .9 8 8 -1 3 3 .1 1 1 -1 7 5 .5 4 3 * (82.424) (103.778) (72.360) .00572* .00773* .0 1 09 6 ** (.00284) (.00224) (.00217) -2 4 .2 8 2 -3 7 8 .1 7 9 * 151.748 (166.327) (239.776) (194.069) -3 3 .7 4 0 19.926 --5 4 .0 2 4 ** (23.534) (18.540) (25.366) Time and savings deposits/Total —3 8 6 .6 1 5 ** -3 3 2 .9 1 3 * * -1 3 6 .0 8 5 -1 8 6 .5 3 1 -1 1 4 .9 7 8 -2 9 6 .7 0 0 -1 2 1 .9 7 8 deposits (117.156) (118.847) (119.003) (227.133) (143.393) (161.570) (129.361) Intercept 182.748 159.419 160.427 269.940 341.912 205.093 217.972 Standard error of estimate 57.525 71.807 60.912 90.465 84.261 90.195 82.285 R2 .54 .66 .57 .41 .42 .32 .57 Mean o f dependent variable 231.698 267.487 262.530 268.343 275.1 12 306.477 335.397 n 33 34 37 38 45 42 53 1970 43.372 (43.171) -1 1 5 .5 8 7 (87.284) -1 1 6 .8 9 1 (94.076) -3 0 3 .6 6 1 * (97.766) .00651* (.00280) 377.715* (173.984) - 6 1 .5 8 9 * (25.580) -4 7 7 .6 6 9 (193.914) 584.581 113.069 .49 377.365 55 NOTE: Standard errors o f the coefficients are in parentheses. * Significant at 95% confidence level. * * Significant at 99% confidence level. Source: Federal Reserve Bank o f Cleveland Liability Composition. The ratio o f tim e and w ith a choice o f conclusions: either this study is savings deposits to total deposits is confirm ed as a seriously flawed or the general belief th a t Federal significant factor affecting bank earnings. In 47 Reserve membership is costly out o f 48 cases,the estimate has a negative sign;and banks is w ro n g .14 fo r all member in 29 o f those instances, the coefficient is sig n ifi cantly d iffe re n t from zero at the 95 percent confidence level. C O R R O B O R A T I V E E VI DE NC E Given the above failure to id e n tify a systematic cost o f Federal Reserve membership, one is le ft 14 A study that finds net benefits in membership for Alabama banks is John G. Fulmer, Jr., “ An Investigation into the Effects of Federal Reserve System Membership o f Individual Commercial Banks" (unpublished Ph.D. dissertation. Graduate School o f Business, University of Alabama, 1970). 13 ECONOMIC REVIEW T A B LE VI Net Income Regression Results fo r Fourth D istrict Banks W ith Assets fro m $50 M illion to Less Than $100 M illio n (Group V I) 1963-1970 Membership Total assets G rowth Loans/Investments Time and savings deposits/Total deposits Intercept Standard error of estimate R2 Mean o f dependent variable n 1963 1964 1965 277.067 (141.690) .0 0 80 1 ** (.00147) 344.570 (699.323) 47.902 (75.992) 102.969 (71.020) .0 0 6 7 2 ** (.00159) -1 8 4 .2 2 6 (482.199) 55.754 (62.500) -6 9 .2 1 0 (77.360) .0 0 7 9 5 ** (.00250) -9 2 3 .2 6 9 (959.695) 61.824 (70.748) -7 4 .2 2 4 38.760 (117.749) (77.018) .00785* .00666* (.00362) (.00266) 1827.415 690.490 (1202.917) (632.184) -9 1 .4 9 5 -1 0 4 .7 7 5 (91.396) (60.998) -3 1 5 .0 6 6 (253.544) -3 2 1 .7 8 4 -3 8 5 .0 9 8 (276.377) 7.126 -6 5 1 .7 2 3 (314.810) 315.690 -6 3 2 .4 6 0 --1033.013* *-1 1 0 6 .1 9 4 *-6 9 4 .1 9 1 * -8 4 4 .5 6 4 (482.814) (266.153) (421.893) (326.423) (253.024) 378.134 739.257 646.456 384.356 383.722 96.381 .82 116.045 .60 151.574 .50 260.200 .29 162.700 .60 271.825 .31 218.575 .34 162.705 .46 452.294 17 443.300 20 462.682 22 445.097 31 509.448 29 498.150 40 628.805 41 638.237 38 1967 1966 1968 1969 171.267 (101.826) .00661 (.00362) 510.587 (763.226) -6 1 .5 8 0 (79.754) 1970 78.241 58.022 (81.408) (62.927) .0 0 97 3 ** .0 1 0 0 9 ** (.00306) (.00246) 570.833 -2 2 6 .3 9 0 (735.433) (477.546) -5 0 .5 2 8 10.165 (54.182) (44.988) NOTE: Standard errors o f the coefficients are in parentheses. * Significant at 95% confidence level. * * Significant at 99% confidence level. Source: Federal Reserve Bank o f Cleveland One possible defect in the present study is that the conclusion may depend on compensating good candidate fo r a state in w hich Reserve membership is costly. Federal K entucky also errors introduced by the inclusion o f banks from seems to be a lik e ly choice since so few o f the several states in the sample. That is, given the banks there (less than 30 percent) are members. diversity o f state regulations, membership may be Consequently, small to medium size banks in costly in some states and profitable in others. If the Fourth D is tric t portions o f these states are so, the estimated im pact o f membership w ill be analyzed separately. Estimates o f the membership understated. co e fficie n t (and standard error) are presented fo r Pennsylvania—because o f the relatively liberal these tw o groups o f banks in Table V II. None o f provision th a t 40 percent of required reserves may these estimates o f the cost o f membership are be held in obligations o f the U. S. Government, significantly d iffe re n t fro m th e dence is found th a t Federal Reserve membership C o m m o n w e a lth of Pennsylvania or Pennsylvania political sub-divisions—seems to be a 14 in Pennsylvania or zero. Thus, no evi K entucky is costly fo r all the regression results are consistent w ith stable Fourth D istrict membership during this period. A sharp drop in Federal Reserve membership w ould 1963-1970 require some explanation other than p ro fit motive, 1963 1964 1965 1966 1967 1968 1969 1970 Pennsylvania K entucky -1 .0 7 6 (8.278) 5.337 (12.786) -2 5 .1 9 2 (13.153) 7.344 (26.664) -5 .6 3 1 (16.499) -1 .4 3 7 (24.336) -2 6 .7 8 7 (19.745) -6 .5 2 0 (15.351) - .7 5 5 (5.072) 8.692 (5.740) -7 .4 1 6 (5.250) -1 0 .9 4 7 (7.776) 10.406 (9.957) -7 .4 2 6 (12.003) 15.561 (11.384) 2.794 (11.181) NOTE: The estimated impact is in thousands o f dollars. None o f the estimates are significantly d iffe re n t from zero at the 95% confidence level. * Banks w ith $5-$30 m illio n in total assets and located in the Fourth D istrict portions of Pennsylvania and Kentucky. Source: Federal Reserve Bank o f Cleveland if the regression results are correct. A t firs t glance, it is startling to discover th a t the number o f member banks in the Fourth D istrict declined by 57, from 527 to 470, between the beginning o f 1963 and the end o f 1970. This figure is put in to perspective, however, by noting th a t the number o f nonmember banks in the Fourth D is tric t declined by 32, from 357 to 325 during the same period. Thus, member banks constituted 59.6 percent o f all Fourth D istrict banks in 1963 and 59.1 percent in 1970. Bank mergers accounted fo r v irtu a lly all o f the reduc tio n in the number o f banks, and there was no noticeable tendency fo r member banks to dis appear through merger faster than nonmembers. It is also interesting that, w hile seven member banks converted to nonmem ber status during the period under review, six nonmembers converted to mem bership. Nor was this sym m etry o f numbers disturbed by asymmetries o f size. A t the beginning o f 1963, member banks held 88.4 percent o f to ta l bank assets in the D istric t and 89.2 percent o f (i.p.c.) demand deposits. A t the end o f 1970, these banks, even when banks in these states are treated separately. One way o f evaluating the present study is by considering the im plications o f the postulate th a t ratios were o nly slig h tly lower, 87.1 percent and 1R 87.6 percent, respectively. In summary, there was no substantial net loss o f members in the Fourth D istrict between 1962 and 1971. membership is n ot costly fo r banks in the Fourth D istrict and attem pting to confirm one or more o f those im plications. For example, the absence o f significant negative coefficients fo r the member ship variable implies th a t Fourth D istrict member banks had little p ro fit incentive to leave the Federal Reserve System during 1963-1970. Thus, 15 The m inor declines that did occur may be a ttrib u te d to the com position of new insured commercial banks form ed in the D istrict since 1962. Nine national (member) banks and 18 state nonmember new banks have opened. No new state-member bank opened during this period. ECONOMIC REVIEW SUMMARY This study cannot claim to have proved th a t either in to or o u t o f membership. S im ilarly, there Federal Reserve membership is never costly to any may be some individual banks th a t w ould be bank. Federal Reserve harmed and state regulations are if they had to change th e ir member constantly changing. Consequently, the findings o f ship status. A ll th a t can be concluded here is th a t this study and others like it need n o t hold fo r the earnings o f Fourth D is tric t member banks as a banks in all regions at all times. A d d itio n a lly , this group were n o t significantly d iffe re n t from those paper deals w ith the earnings o f member banks as o f nonmembers during 1963-70 where allowance is a group compared w ith the earnings o f nonmem made fo r differences in location by state, size, ber banks. It is possible th a t there are individual grow th, loan to investment com position, and the banks—both members and nonmembers—whose p ro fita b ility could be improved by a conversion Digitized for16 FRASER p roportion o f total deposits made up o f tim e and savings accounts. Appendix Federal Reserve and State Reserve Rei Percent o f Deposits and Composition 1963-1970 (Shaded areas denote changes.) A. Federal Reserve Time Deposits Net Demand Deposits" Reserve C ity Banks Country Banks Effective Date:): In effect Jan. 1, 1963 1966—July 14, 21 Sep. 8, 15 1 9 6 7 -M a r. 2 Mar. 16 1968—Jan. 1 1 ,1 8 1969—Apr. 17 1970—Oct. 1 Other Time Deposits! Under $5 Over $5 Under $5 Over $5 Under $5 Over $5 M illion M illion M illion M illion Savings M illion M illio 1 61/ 2% 16/4% 12% 12 % 161/2 16 1/z 16 V2 I 6 V2 161/2 17 17% 17% 12 12 12 12 12 12 12 12 12 1 61/2 161/2 161/2 1 6 V2 17 17 1 2 V2 1 2 1/2 12'A 13 13 4% 4% 4 4 3 1/2 3 3 3 3 4 4 3 'A 3 3 3 3 4% Com position o f Reserves Deposits w ith the Federal Reserve and vault cash. 5 6 6 6 6 6 5 * Demand deposits subject to reserve requirements are gross demand deposits minus cash items in process o f collection and demand balances due from domestic banks, t Effective January 5, 1967, tim e deposits such as Christmas and vacation club accounts became subject to the same requirements as savings deposits. X When tw o dates are shown, the first applies to the change at reserve c ity banks and the second to the change at c o u n try banks. Source: Federal Reserve Bulletin B. K entucky Demand Deposits 1 Reserve C ity Banks Country Banks Time Deposits In effect Jan. 1, 1963 10% 7% 3% Demand deposits w ith commercial banks and vault cash. June 18, 1970 10 7 3 Up to 25% o f the reserve against demand deposits may be in U. S. Treasury or Agency securities w ith a m a tu rity o f one year or less or certificates o f deposit issued by banks located in K entucky. The reserve against time deposits may be invested in securities issued by the U. S. Treasury, the Common wealth o f Kentucky, or certificates o f deposit issued by banks located in Kentucky. Effective Date f Deposits payable w ith in th irty days. Composition o f Reserves Source: O ffice o f the Kentucky Commissioner o f Banking and Securities ECONOMIC REVIEW C. Ohio Effective Date Demand Deposits1 In effect Jan. 1, 1963 15% Jan. 18, 1968 Aug. 18, 1970 12 12 Ti me De posits t Com position o f Reserves 10% Demand deposits w ith commercial banks and vault cash. Up to 60% o f reserves against tim e deposits may be in obligations issued or guaranteed by the U. S. 8 6 * Deposits payable w ith in 30 days exclusive o f U. S. deposits, t Exclusive o f U. S. postal savings deposits. Source: O ffice o f the Ohio Superintendent o f Banking D. Pennsylvania Tim e Deposits Other Tim e Deposits Effective Date Demand Deposits* Savings Under $5 M illion Over $5 M illion 14% 6% 6% 6% In effect Jan. 1, 1963 Jan. 10, 1963 July 21, 1966 March 7, 1967 12 12 12 Com position o f Reserves Demand deposits w ith commercial banks and vault cash. Up to 40% o f reserves may be in obligations o f the U. S. Treasury, the Commonwealth of Pennsylvania or any Pennsylvania p o litical subdivision. * Deposits payable w ith in 30 days. Source: O ffice o f the Pennsylvania Secretary o f Banking E. West V irginia Effective Date Demand Deposits1 Time Deposits In effect Jan. 1, 1963 10% 5% July 1, 1969 7 3 Com position o f Reserves Demand deposits w ith commercial banks and vault cash. * Deposits payable w ith in 30 days. Source: O ffice o f the West Virginia Commissioner o f Banking Digitized for18 FRASER J A N U A R Y - F E B R U A R Y 1973 A MEASURE OF MONETARY POLICY Lorraine E. Duro The p o lic y actions o f m onetary authorities have an im p o rta n t influence on economic grow th, em ploym ent, prices, and the in te rn a tio n a l balance o f payments. In a d d ition , they have an e ffe ct on the nation's financial markets. F or many analysts, the in te rp re ta tio n o f p o lic y actions in terms o f tim ing and im pact on the economy is d iffic u lt from existing financial series. Essentially, there are three approaches to measuring the direction o f m onetary policy. One—an eclectic approach—consists o f weighing the evidence fro m a variety o f financial variables before reaching a conclusion. This approach is somewhat lim ite d , however, because d iffe re n t variables can provide c o n flic tin g signals over the short term. A second approach involves the selection o f one variable th a t is believed to be least affected b y non p o lic y forces and therefore least lik e ly to be misleading. A lthough a num ber o f analysts favor this approach, they do n o t agree on the variable to be used. A th ird approach involves the m odifica tio n o f an existing financial variable so th a t i t reflects p o lic y actions. This stud y employs the la tte r approach. The im pact o f economic a c tiv ity on the observed money stock is estimated and then removed to derive a measure o f p o lic y action. M onetary p o lic y o f the past 2 0 years is then reviewed in terms o f this "d e rive d ” indicator. The results should be o f interest to those who are studying methods o f measuring m onetary p o licy and who participate in the debate on the v a lid ity o f measures o f m onetary policy. 19 ECONOMIC REVIEW I he Federal Reserve System has among its goals Some economists believe th a t an improved the p rom otion of economic grow th, high em ploy measure of m onetary action can be derived by ment, in the removing the effects o f economic a c tiv ity from the nation's balance o f payments. The System is also available financial variables. While there is not concerned like ly to be a perfect measure o f monetary actions, price s ta b ility , w ith other and eq u ilib riu m goals, such as orderly performance o f financial markets. Because o f these it is nevertheless useful to explore the im plications m ultip le objectives, it may be d iffic u lt to infer of this alternative by constructing a m odified policy in te n t from p o licy action, at least u n til the indicator—w hich is the approach th a t underlies records o f the Federal Open Market Com m ittes are this study. made available to the public three months after The article firs t discusses the lim ita tio n s o f the each meeting. Therefore, many observers try to indicators in current use. A measure th a t was assess the effect o f policy actions—w ith o u t regard derived in an e ffo rt to correct fo r the bias o f these to in te n t—because o f policy's influence on a broad indicators is then considered. Finally, monetary spectrum o f economic and financial measures. Frequently, this focuses on the questions o f policy o f the past 20 years, as it w ould be classified by this new measure, is reviewed. whether, and to w hat degree, policy is stim ulating or restraining economic activity. The im pact o f monetary policy can be studied w ith econom etric models o f the economy. Such models allow observers to THE CHOICE OF AN INDICATOR For a variable to serve as a good indicator of make estimates of monetary p olicy, it should be d ire ctly affected by impact o f a given policy action—as measured by actions o f m onetary authorities and have some changes in the tra d itio n a l central bank policy predictable tools—on o u tp u t, prices, and em ploym ent. But fo r should not be simultaneously influenced by other those w ho do not have access to, or do not prefer economic forces. relationship to policy goals, but it If such an indicator existed, to rely on models, an alternative approach is to policy ease or restraint w ould be signaled by a select an individual or group of financial variables direction change o f the indicator; policy strength fo r use as a measure o f policy actions. w ould be measured by the magnitude and duration For many decades, simple indicators—such as o f the changes; and adequacy o f p o licy w ould be the level or change in the q u a n tity o f money and judged by contrasting the posture and strength to certain interest rates—have been used to assess an estimate o f the economy's needs. policy action. None o f these indicators has been an ideal index o f policy actions or im pact, because Two broad classes o f variables are used to appraise monetary policy actions: one relating to they are biased by th e ir jo in t dependence on both money market variables and another relating to policy and nonpolicy influences. In ad d ition , the money stock variables. Generally speaking, the impact o f a given policy action is not constant firs t are m ainly prices (such as the Federal funds over tim e. Therefore, the "in fo rm a tio n '' about the rate and Treasury b ill rate) and the second, mainly effect quantities (such as reserves and money stock). of policy actions on the behavior of financial measures could be misleading. Digitized for20 FRASER U n fo rtu n a te ly, neither o f these classes nor any one B R U A R Y 1973 variable from one o f these classes solely reflects in an industry trigger an increase in demand fo r policy influence. Because both market and money bank loans and if the banks' demand fo r excess stock related variables show the combined impact reserves has some interest elasticity, interest rates of changes in policy and economic a ctivity, they w ould rise and the money stock w ould increase. cannot indicate the effect o f one independently o f These the opposite policy stance to some observers; i.e., other. To com pound the problem, the movements, in isolation, could signal influence o f economic activity on both classes of restrictive action on the basis o f rising rates or variables is such that they w ould lead to opposite expansive action A ccordingly, money stock. Ideally, policy should be defined in m arket variables are biased toward a terms o f the response o f the monetary authorities conclusions "m oney as p olicy indicators. Reserve policy to these changes. But the resulting observed levels related variables are of interest rates and money w ill o n ly reflect the favorable assessment o f Federal actions on the basis o f an increased and money-stock combined impact o f a u th o rity response and the biased tow ard an unfavorable assessment."1 Prices and quantities o f many financial variables tend to move in the same direction as economic behavior o f the public and the banks. Since economists are well aware of this activity. Interest rates typ ic a lly rise in periods o f problem, it is not surprising th a t many prefer an economic expansion and fall in recessions. For eclectic approach, subjectively weighing a broad those who prefer to classify m onetary policy based range o f indicators to assess the stance of policy on money m arket indicators, changes in interest actions. However, this approach can lead to diverse rates w ould suggest counter-cyclical actions by the estimates monetary of tim in g and, since it is not stock quantifiable, provides little indication o f magni related variables also tend to rise and fall w ith tude or sufficiency o f policy actions in relation cyclical ship to policy targets or goals. a u th o rity. However, money a ctivity. Those who prefer to classify policy according to money stock related variables w ould therefore conclude th a t policy was A new approach was taken by economist Patric Hendershott who chose to construct a "n e u tra l is ized money sto c k " to serve as an indicator of necessarily correct. Cyclical behavior w ould be m onetary p o lic y —rather than to jo in the debate as observed in both classes o f variables even if the to which economy a u th o rity. misleading. This approach has its counterpart in Therefore, policy should be defined in terms o f fiscal policy in the high-em ploym ent surplus or the alteration o f this cyclical pattern by policy d e ficit concept. In his construct, deviations from pro-cyclical. had Neither judgm ent no central of m onetary policy actions. For example, assume banks have some level o f excess reserves. Should some technological change indicator or set o f indicators is least the trend values o f components o f the money stock were removed. The and conventional Patric H. Hendershott, The N eutralized Money S tock: An Unbiased Measure o f Federal Reserve Policy Actions (Homewood, Illin o is: Richard D. Irw in, Inc., 1968) p. 1. money stock” represented the in itia l e ffo rt to m o d ify a variable 1 "neutralized ever, the correct m onetary acceptance the bias problem of policy indicators. How and usefulness of this m odified variable has been lim ited, partially due to the substantial data lag o f the variables used in its 21 ECONOMIC REVIEW form u la tio n. Nevertheless, the basic thrust of The form er, or endogenous com ponent, reflects Hendershott's w o rk market changes in overall a c tiv ity th a t are transm itted to variables and money stock related variables have the money stock by increased bank lending, shifts predictable biases th a t due both to money cyclical economic in currency and deposits, and the use o f reserves.4 The p o licy, or exogenous com ponent, reflects activity remains a substantive challenge. actions o f the monetary authorities. Thus, the observed money stock is the sum o f components determined by endogenous and exogenous factors (M = M n + M x ). A MODIFIED MEASURE, Mx Numerous In this article, a new m odified variable th a t can be sim ply calculated on a m o n th ly basis studies indicate th a t income is related to money w ith a lag. The d is trib u tio n and is va ria b ility o f the lag are open questions, b u t the considered as an indicator o f monetary policy. existence o f a lag is n o t.5 Therefore, current dollar Monetary policy o f the past 20 years is discussed GNP (Y t ) is a fu n ctio n o f the money stock of in terms o f this indicator. Periods o f ease and previous periods (Mt n ), and the net effect o f restraint are delineated by increases and decreases other variables (z), including all other current and of the new series; i.e., by directional movement. lagged values o f endogenous and exogenous forces Strength other than values o f the money stock (Y t = f(M t p , and persistance o f p o licy periods are inferred from the percentile rank o f the m o n th ly Zt l) . values. Statements of the appropriateness o f policy w ith respect to the stabilization needs o f the In contrast, current economic a c tiv ity has some immediate im pact on the money stock. It is not economy also can be made, b u t o nly after some im p o rta n t to specific target or goal assumption is e x p lic itly endogenous stated. It should be stressed at this p o in t th a t the reflected in the current m o n th ly period, b u t rather choice o f an indicator, conventional or m odified, th a t some is separate and distinct from a theory o f income present. For example, firm s may meet their need determ ination. this fo rm u la tio n force of impact fro m th a t all o f the economic a c tiv ity income to be money is 3 The observed money stock can be thought o f as the sum resulting fro m tw o major forces: current fo r increased growing loans, w orking economic and balances in a c tiv ity consumers by may periods of obtaining finance bank durable economic a ctiv ity and current monetary policy. 4 9 Ronald L. Teigen, "R eview o f The Neutralized Money Stock, by Patric H. H endershott." Journal o f Finance, Vol. 25, No. 4, September 1969, pp. 744-46. 3 Richard Zecher, "Im p lica tio n s o f Four Econom etric Models fo r the Indicators Issue," Am erican Economic Review, May 1970, pp. 47-54. Digitized for22 FRASER The values of endogenous variables are determined by the sim ultaneous interaction o f the relations in a statistical model; exogenous variables are determ ined outside the model. 5 John H. Kareken and Robert M. Solow, "P art 1. Lags in Fiscal and M onetary P o licy," S tabilization Policies, Commission on Money and Credit, Research Studies o f the Commission, (Englewood Cliffs, New Jersey: Prentice Hall, Inc., 1963.) J A N U A R Y - F E B R U A R Y 1973 purchases by instalm ent debt. Therefore, the endogenous com ponent o f the money stock is a fu n ction of current economic a c tiv ity certainly seem to be a good p ro xy fo r GNP. Using this index, the steps taken to infer the policy p o rtio n o f the money stock (M ^ ) can be demonstrated. (M N t = g(Y t )). The Thus, the observed money stock is the sum of objective is to remove the effects o f the exogenous policy p o rtio n o f the same time economic a ctivity from the money stock so as to period and the endogenous portion th a t results derive a measure o f policy actions. As from th a t an observed value, the money stock is regressed on induced by previous policy (Mt = M ^ t + g (Y t )). the composite index and a trend fa cto r as follow s: current cyclical a ctiv ity , including Because o f the lagged relationship, it is assumed is not M = a + b-| Cl + b 2 T + u there is no interaction between the components o f where, the sum in the same tim e period. This suggests that the current exogenous p o licy p o rtio n can a, b^, and b 2 are regression estimates and be thought o f as the residual of the observed money stock not related to cyclical and secular trends and past Mt policy - actions (M x t = Mt — M = m o n th ly seasonally adjusted money stock = in billions o f dollars Cl = m o n th ly composite index o f coincident g(Y t )). indicators T = dum m y variable fo r tim e To use this procedure, it was necessary firs t to u = calculated residual term select a measure o f economic a c tiv ity th a t is current (i.e., available w ith subject to m inim al little tim e lag and revisions), and then to The term " u " may be viewed as being composed of and v, a random error. In calculating u, the demonstrate th a t a significant relationship exists endogenous between However, since Although this measure the economic and the money stock. most comprehensive measure of Mt _ influence is reflected in " u . " (How well is (GNP), calculations o f a m onetary indicator using the expected value o f v should be close to zero, and its variance should be small. Also, be National from this considerable tim e Gross removed reflected depends at least on tw o conditions. First, w ould is is is not in the regression, its Product series a ctivity influence lim ited because of a lag. A series th a t is readily available and subject to minimal revisions had to be found to serve as a reasonable GNP proxy. Also, since the fo rm u la tio n specifies independence between current monetary policy and economic activity, the shortest tim e span com patible w ith data availability is a prerequisite. The composite index o f coincident indicators published by the D epartm ent o f Commerce has the advantage o f being available on a m o n th ly basis and w ould should be independent o f the cyclical variable. If these 0 See BCD series No. 820, published m on th ly in Business Conditions Digest, U. S. Departm ent of Commerce, Washington, D. C. However, it is not clear th a t this index w ill adequately reflect price changes since o n ly one of the five series used in its construction reflects current price changes. Consequently, the regressions discussed in the te x t were duplicated using an index form ed by the product o f the Industrial Production Index (BCD No. 47) times the Consumer Price Index (BCD No. 781) divided by 100. This economic a ctivity index w ill be noted as EAI and the regression results using this index w ill be footno te d . 23 ECONOMIC REVIEW conditions are not seriously violated, movements relationship must be stable if it is to be used to in u w ill reflect movements in M ^ .) derive a m onetary policy measure as the residual. As changes in money and economic a ctiv ity are If the im pact o f income on m oney is highly more m eaningfully expressed in percentage terms, variable—w hich rather than levels, the equations are expressed in m oney on incom e—the exogenous policy portion log form . For example, a $10 b illio n increase in o f the money stock could not be represented by money stock th a t totals $100 b illio n expansive in policy increase in terms than a stock is the relationship expected o f is more the residual. Therefore, another set o f estimates a $10 b illio n was com puted. The m o n th ly data fo r the period th a t totals $200 b illio n . 1952-1971 were disaggregated in to periods of Therefore, the log form is used in the regressions, economic expansions and economic contractions, w hich perm its the coefficient fo r the cyclical based on the reference cycles as defined by the indicator to be interpreted as an elasticity value National Bureau o f Economic Research (NBER). and the tim e co efficie n t to be translated into an For the 197 months o f economic expansion, the annual rate o f change. For the period 1952-1971, results are: the results are:7 (2) Log M = 2.89152 + .51577 Log C l- .0 0 0 3 3 T (10.76) (1) Log M = 3.03692 + .47677 Log Cl -.0 0 0 0 7 T (15.91) (-.4 4 ) R2= .96 (-1 .1 4 ) R2 = .95 and fo r the 43 contraction months: (3) Log M = 3.0184 + .48472 Log Cl - .00019 T (23.89) Equation 1 suggests th a t a 5 percent change in the (-.2 8 ) R2 = .98 level o f economic a c tiv ity w ould be associated w ith a money The association between the cyclical in d ica tor and Because the period chosen frequently money was stable. The tim e trends are consistently affects regression results, a fu rth e r test was made insignificant, b u t the negative signs do suggest the to check the sta b ility o f the association between increasing velocity economic a ctivity and the money stock.9 This period.10 Equation 1 can now be solved fo r the predicted stock.8 2.4 percent change in the th a t prevailed in this tim e values o f the money stock, or the endogenous 7t-statistics appear in parentheses. O Using EAI as explained in foo tno te 6, the results are: (1a) Log M = 2.82093 + .54375 Log EAI - .00064 R2 = .97 (18.77) (-4 .0 6 ) com ponent (Mpg). The difference between M ^ and the observed value o f money (M) represents the policy influence on the stock (M ^ ). 9 One technique fo r a sta b ility test is that proposed by Gregory C. Chow. It consists o f estimating long-run coefficients fo r the independent variables over the fu ll time span under study. These are then tested fo r equivalence against coefficients estimated fo r subsets of the data. Gregory C. Chow, "Test o f Equality Between Sets o f Coefficients in Two Linear Regressions,” Econometrica, Vol. 28, No. 3, July 1960, pp. 591-605. 24 10 Using EA I, the results fo r expansions and contractions respectively are: (2a) Log M = 2.7960 + .55032 Log EAI - .00070T R2 = .97 (17.53) (-4 .1 8 ) (3a) Log M = 2.88691 + .52693 Log E A I-.0 0 0 4 1 T R2 = .99 (9.88) (-1 .3 6 ) J A N U A R Y - F E B R U A R Y 1973 BEHAVIOR OF THE MEASURE in tro d u ctio n , direction and change o f direction of The crucial question is whether the exogenous money stock monetary (M ^ ) is an accurate indicator of posture and select policy periods. When is dilemma growing as a pro p o rtion o f the money stock so cannot be measured against the ideal that an upward trend is indicated in Chart 1, this p o lic y —and returns. the indicator series w ill be used to define policy the original has the pattern o f increase w ill be classified as ease. A desirable characteristic o f exogeneity by construc downward trend w ill be treated as a period of tio n . It is also assumed in this article th a t actions restraint. In periods when indicator since none exists. However, of m onetary stock, authorities do affect the money and th a t changes in the influence aggregate o u tp u t, money stock regardless of the transmission mechanism. Therefore, because o f its narrow moves w ith in a range, in te rru p ting but not reversing a basic trend, the interpretation may appear to be ambiguous because only relative terms are appropriate. For example, in 1963 and 1964, exogeneity and its lin k to economic a ctivity, remained a relatively constant share o f the money has some o f the basic characteristics sought fo r an stock. Since a downward trend had been evident indicator. p rior to this tim e, this period could be considered The inform ational content indicated by the pattern o f as ease based on past values, less restrictive if the values over the past 20 years should basic trend is emphasized, or neutral since policy also be plausible and reasonable. The indicator is direction was not reversed. In terms o f periods expected to d iffe r from other single indicators selected by directional change, this indicator is in initial premise o f this study is that close accord w ith a number of comprehensive conventional indicators are "biased." However, studies th a t descriptively classify monetary policy since the ought to reflect policy periods o f the past as categorized by eclectic appraisal and hindsight. In periods by weighing 11 retrospect. general, this policy indicator meets these expecta However, it should be noted that when policy is described and periods are designated, definitional problems imm ediately arise. The descriptive terms used in policy discussions are notoriously im pre cise; e.g., "leaning against the w in d ,” "n e u tra l," "easy,” "stim u la tive .” personal Subjective interpretation, biases, or d iffe re n t goals can lead to d iffe re n t conclusions and descriptive categories by several observers o f a single chart or set o f data. Therefore, the scope o f the discussion in the next three sections should the discussed indicator policy are those (M ^ ). be carefully interpreted. periods and categorizations suggested by the derived Second, as stated in the of the variables in Estimates o f policy strength and duration can be made by ranking tions. First, all classes. However, to values in to percentile make broader judgmental statements about the appropriateness o f policy from this indicator, or any other, requires the comparison of the indicator values to some desired goal. For example, if the direction o f the indicator is upward but still negative, the policy posture m ight be categorized as one o f ease—but questions remain: How easy? The th ird section or Is there sufficie n t ease? th a t fo llo w s shows how an assumed goal may alter policy classifications made only by direction, change, or percentile rank. 11 One example o f an excellent survey is contained in Lester V. Chandler, The Economics o f Money and Banking, (New Y o rk: Harper & Row, Publishers, 1969), Chapters 24 and 25. 25 ECONOMIC REVIEW pattern o f the exogenous money stock (M ^ ) from Comparison of Leads and Lags of Policy Indicator (M ^) With Respect to Economic Activity 1952 NBER Reference Dates Policy Ease or Restraint: Direction of M ^ . The in to 1972, stated as a percent o f the observed money stock, is illustrated in Chart 1. When the pattern o f is compared to the NBER 8/53 7/53 11/69 the m onetary p o lic y 6/60 The major turning points of the policy measure (M ^ ) and the NBER reference dates are listed in 4/61 2/61 1 1/69* 9/71 11/70 makers can be appraised by contrasting directional change in M X to the economic peaks and troughs. 6/58 4/58 timeliness by 9/57 7/57 5/60 action 9/54 8/54 pro- or counter-cyclical can be made. Also, the of Restraint Ease Trough Peak designation o f economic expansions and contrac tions, an estimate o f the basic posture o f policy as Policy Indicator (M x) * Trend towards ease resumed. See text. Sources:National Bureau o f Economic Research and Federal Reserve Bank of Cleveland the Table. As measured by the exogenous money stock, the m onetary a u th o rity responded to the beginning o f the 1953, 1957, and 1960 recessions by in stitu tin g a movement towards ease in the firs t or second m onth fo llo w in g the business peaks. In except fo r b rie f in te rru p tion s in the firs t half o f shows the resum ption o f a 1956 and the th ird quarter o f 1959. These tw o policy o f ease, after little change in the previous periods are coincident w ith policy actions taken to November 1969, 14 months. (The trend towards ease was instituted counter both the anticipation in late 1966.) Policy changes in the direction o f major steel strikes. The policy measure also shows restraint occurred about tw o months after the movement w ith in 1954, 1958, and 1961 troughs and 10 months range from mid-1961 through 1965. This was a after the 1970 trough. Therefore, generally,policy period when public policy-makers recognized the makers responded q u ickly to major periods o f need fo r stim ulus to encourage domestic expan change in economic activity. In fact, in the period sion, b u t th e ir p o licy from 1952 to 1965, m onetary policy is basically account the large deficits in the nation's balance o f classified as counter-cyclical by the indicator. This payments. O f this period, one study states: "Even judgm ent is consistent w ith conclusions frequently the United States could have its freedom o f action a relatively narrow, negative posture also to o k reached by money market observers, even though lim ited of-payments and international reserve p o lic y ." variable. This lim ited between the movement ended w ith balance- the move economic a ctivity, the policy measure also reflects toward ease occurred in 1967 and 1968, follow ed specific by a interlude o f relative n e u tra lity in 1969 and a authorities. It longer-term trends action is evident in by the points to toward additional restraint in 1966. A movement of tu rn in g related in periods major considerations into the indicator is a m odified money stock related In by and actuality o f m onetary Chart 1 th a t the in the m id-1950's and late 1950's in this policy indicator were downward, 26 n ib id ., p. 511. CHART 1 EXOGENOUS MONEY STOCK (MX ) AS A PERCENT OF OBSERVED MONEY STOCK Percent .12 .10 .06 .02 .00 -.02 -.04 -.08 1952 '54 '56 '58 '60 '62 '64 '66 '68 '70 '72 NOTE: Shaded areas indicate periods o f business contraction as defined by the National Bureau o f Econom ic Research, Inc. Last en try: December 1972 Source: Federal Reserve Bank o f Cleveland continuation o f ease since 1970. Consequently, leads to the classification o f monetary policy as the exogenous money stock indicator w ould im ply pro- or counter-cyclical. th a t at least fo u r o f the seven years fro m 1965 Policy Strength: Percentile Values of the M « . through 1972 were pro-cyclical m onetary policy A nother advantage o f a quantifiable measure is years 1971, and 1972). In that it permits a comparison o f policy move. The (i.e., 1967, 1968, summary, the direction o f the measure indicates question can be asked: How do changes in various policy periods o f ease or restraint, w hile a contrast policy p e rio d sd iffe r in magnitude from other policy o f these periods to the economy's cyclical stage periods? If the values o f the in d ica tor are analyzed ECONOMIC REVIEW in terms o f their d istrib u tio n , each m o n th ly value drawn by using a percentile rank procedure w ould o f the series can be ranked relative to all other have to be m odified or even rejected. values occurring during the 20-year span. Any m onthly be appraised Chart 2 illustrates the money stock series and in tw o the exogenous money stock series in terms of dimensions, the percentile rank and the persistence annual rate o f change adjusted to a three-month o f p olicy moving average. Using values o f values. value can then through an appraisal o f surrounding For example, in money stock reached 1966, the exogenous its lowest value o f the and percentile rank o f Chart 1, monetary policy during the 1954, 1957, and 1960 recessions moved tow ard ease. 20-year period, and 11 o f the m o n th ly values fall However, the grow th o f the money stock in each in the lowest decade o f these periods was sharply slowing, as shown in Yet, percentile rank may give an Chart 2, and an actual money decline follow ed. incom plete story if policy posture is not measured This demonstrates th a t, although the annual rate against policy goal. of change in d istrib u tio n . decile of the tw o (also shown was positive and relatively high in Chart 2), p olicy was not as stim ulative as it m ight have been if the goal was to encourage fu lle r u tiliza tio n of resources by Appropriateness o f Policy: Value o f M ^ vs. avoiding a decline in the money stock. In other Target. While classification in terms o f percentile words, the im pact o f exogenous policy was not rank leaves unanswered the su fficie n t to offset the decline in the endogenous question o f appropriateness o f policy actions in p o rtio n , much less to add consistently to a positive relationship rate of grow th. is inform ative, it to the desired goals of resource u tiliza tio n in the economy. The firs t differences o f the derived policy indicator can The same type o f analysis could be done fo r be used to any predetermined desired rate o f change in the appraise the co n trib u tio n o f the exogenous money money stock. If a target rate o f grow th is specified stock to the observed rates of change in the money fo r a period, based on em ploym ent and price level stock. This w ould provide some indication o f the goals o f th a t period, an estimate can be made o f adequacy o f policy action—if the desired rate of the c o n trib u tio n o f exogenous policy to th a t goal. growth o f the money stock were specified. H ow This is n o t meant to suggest an adoption o f the ever, even the desirability of specifying a growth " ru le " position fo r monetary policy. It is instead rate fo r the money stock is an unresolved issue in an e x p lic it recognition o f the System's stabiliza economics today and the topic o f vigorous debate. tio n Some observations can be made by assuming a ultim ate goals o f high em ploym ent, price sta b ility, desired goal. For example, assume th a t the stock and grow th. of money should not decline in periods o f substan tia lly less than fu ll em ploym ent. Given some price rigidities in the economy, it is u n lik e ly that velocity changes could compensate fo r any consis tent decline in the money stock. Therefore, using on ly this assumption, some o f the conclusions 28 role w ith changeable targets to achieve J A N U A R Y - F E B R U A R Y 1973 CHART 2 OBSERVED MONEY STOCK AND EXOGENOUS MONEY STOCK (MX ) Percent .15 O B S E R V E D M O N E Y S TO C K .10 .05 .00 -.05 E X O G E N O U S S TO C K -.10 -.15 Seasonally Adjusted Annual Rate 3-M onth Moving Average i m -.20 i i i I I I I I I I I I I Ia ll I N O T E : Shaded areas indicate periods of business contraction as defined by the National Bureau o f Econom ic Research, Inc. Last entry: December 1972 Sources: Board of Governors of the Federal Reserve System and Federal Reserve Bank of Cleveland CONCLUSION developed as to w hich p o licy variable is most One approach in appraising the impact of m o n e ta ry p o lic y suitable as an indicator o f m onetary actions. As actions on the economy is to one economist stated, "T h e indicators issue cannot variable th a t reflects policy. be expected to die so long as the tw o sets o f prime However, all o f the financial measures th a t are indicator candidates—interest rates and monetary readily available are jo in tly dependent upon policy stocks—frequently and nonpolicy influences.Therefore, a debate has about the stance o f monetary p o lic y .''13 A choice select a financial yield c o n flic tin g in form ation Zecher, op. cit., p.47. 29 JOMIC o f any one interest rate or any one money stock counter-cyclical in this tim e span, as contrasted to related variable is even more like ly to receive a the rate o f change in the money stock (Chart 2). skeptical reception. However, it is also evident th a t policy was not as This article considers an approach to obtain a stim ulative as it m ight have been in some recession variable th a t reflects policy periods if a goal assumption o f positive growth in actions only. The derived indicator is subject to the money stock is imposed. In the la tte r half of measurement error because o f its residual tech the 1960's, nique, but this error is believed to be considerably counter-cyclical action. It is also interesting that, m odified m o n th ly shows a mixed pattern o f pro- and endogenous fo r the 1966-1972 period, the exogenous money economic activity to conventional indicators. (This stock substantially m irrored the rate o f change o f does n ot preclude the fa ct th a t policy actions w ill the observed be unexpected during a period o f high em ploym ent smaller than taken a c tiv ity —in th a t w ith this introduced respect sense, to by desired policy and economic economic activity are never independent.) money stock. This result is not (1966-1969); b u t since 1970, the pattern implies a departure from previous history. F inally, the The derived exogenous money stock indicator indicator derived fo r this study suggests th a t there suggests stabilization policy in the 1950's and the are grounds fo r examining alternative m odified firs t half o f the 1960's was generally counter financial variables as a means fo r interpreting cyclical, as shown in Chart 1. When viewed in monetary policy actions. terms o f rate o f change, was still basically J A N U A R Y -F E B R U A R Y 1973 ANNUAL INDEX TO ECONOMIC R E V I E W - 1 9 7 2 MONTH ARTICLE TITLE JA NU AR Y Excess Reserves and Bank Size Capital Market Developments, 1952-1970 FEBRUARY Federal Agency Issues: Newcomers in the Capital Market MARCH Banking Structure and Performance: Some Evidence From Ohio The Structure o f State Revenue A P R IL — M AY JUNEJULY AUGUSTSEPTEMBER OCTOBERNOVEMBER A Newcomer's View o f the U. S. Banking Industry The Nature and Use o f Forward Exchange What Happens When the U nem ploym ent Rate Changes? Defining the Product Market in Commercial Banking Policy Influence on the Money Stock in 1971 The Market fo r State and Local Government Bonds The Im pact o f In fla tio n on the Elderly A natom y o f Profitable Medium-Size Banks in the Fourth D istrict, 1966-1970 Economic G rowth and Change in K entucky, 1960-1970