An acceleration of money growth was intended to stimulate demand and provide themonetary scope necessary for thedesired real growth of the economy.On the otherhand, the target was also intended to show that no precipitate action would betaken to ease monetary conditions, in order not to jeopardize further progresstowards containing the inflationary tendencies. (Deutsche Bundesbank 1976a,p.5)It is worth noting,however, that this explanation and the statement cited in the previousparagraph were made after the targets were announced, not contemporaneously with theannouncementTHEOPERATIONALFRAMEWORKOur historical and institutional analysis in this section and the following one (whichdiscusses German monetary policy in the 1990s) independently confirms the impression ofGerman monetary policymaking raised in Bernanke and Mishkin (1992)and argued by latereconometric observers. That is, the Bundesbank does not behave according to a reduced-form-reaction function as though price stability were its sole short-to-medium-term policygoal, or as though the monetary growth-goal correlation were strong enough to justifystrictly following the targets, ignoring wider information. In fact, in the followingdiscussion we bring out the operational reality and implications: that the monetary targetsprovide a framework for the central bank to convey its long-term commitment to pricestability.From 1975 until 1987, the Bundesbank announced targets for the growth of centralbank money (CBM). CBM is defined as currency in circulation plus sight deposits, timedeposits with maturity under four years, and savings deposits and savings bonds withmaturity of less than four years (the latter three components are weighted by theirrespective required reserve ratios as of January 1974).CBM is different from the monetarybase in that banks' excess balances are excluded and the weights of deposits subject toreserve requirements are historical, not current, ratios.Since 1988, the Bundesbank has used growth in M3 as its intermediate target. M3is defined as the sum of currency in circulation, sight deposits, time deposits with maturityunder four years, and savings deposits at three months' notice. Apart from not includingsavings deposits with longer maturities and savings bonds, the major differencebetween M3 and CBM is that the latter is a weighted-sum aggregate, while the former isa simple sum. By definition, therefore, CBM moves very closely with M3.Because theweights on the three types of deposits are fairly smal, the only source for large divergencesbetween the growth of the two aggregates is significant fluctuation in che holdings ofcurrency as compared with deposits.This potential divergence became critical in 1988in the face of shifting financial incentives, and again in 1990-91, after German monetaryunification.The Bundesbank has always set its monetary targets at the end of a calendar year forthe next year. It derives the monetary targets from a quantity equation, which states thatthe amount of nominal transactions in an economy within a given period of time is15
identically equal to the amount of the means of payment times the velocity at which themeans of payment changes hands. In rate-of-change form, the quantity equation states thatthe sum of real output growth and the inflation rate is equal to the sum of money growthand the change in (the appropriately defined) velocity. The Bundesbank derives the targetgrowth rate of the chosen monetary aggregate (CBM or M3) by estimating the growth ofthe long-run production potential over the coming year, adding the rate of price change itconsiders unavoidable (described below), and subtracting the estimated change in trendvelocity overthe year.Two elements of this procedure deserve emphasis.First, the Bundesbank does notemploy forecasts of real output growth over the coming year in its target derivation, butinstead estimates thegrowth in production potential.7This"potential-oriented approach"is based on the Bundesbank's conviction that it should not engage in policies aimed atshort-term stimulation. This approach allows the Bundesbank not only to claim that it is notmaking any choice about the business cycle when it sets policy, but also to de-emphasize anypublic discussion of its forecasting efforts for the real economy, further distancing monetarypolicy from the course of unemployment. The transparency of the quantity approach,therefore, gets certain items off the monetary policy agenda (or at least moves in thatdirection) by specifying the central bank's responsibilities.The second noteworthy element of the Bundesbank's procedure for deriving thetarget growth rate of its chosen monetary aggregate relates to the concept of"unavoidableprice increases,"where prices are measured by the all-items consumer price index (CPI),These goals for inflation are sec prior to the monetary target each year and specify theintended path for inflation, which in turn motivates monetary policy.In view of the unfavorable underlying situation, the Bundesbank felt obliged until1984 to include an "unavoidable" rate of price rises in its calculation. By so doing, ittook due account of the fact that price increases which have already entered into thedecisions of economic agents cannot be eliminated immediately, but only step bystep. On the other hand, this tolerated rise in prices was invariably below the currentinflation rate, or the rate forecast for the year ahead. The Bundesbank thereby made itplain that, by adopting an unduly "gradualist" approach to fighting inflation, it didnot wish to contribute to strengthening inflation expectations. Once price stabilitywas virtually achieved at the end of 1984, the Bundesbank abandoned the conceptof "unavoidable" price increases. Instead, it has since then included ...a medium-term price assumption of 2%. (Deutsche Bundesbank 1995c, Pp.80-1)The setting of the annual unavoidable price increase thus embodies four normativejudgments by the Bundesbank.First,a medium-term goal for inflation motivates policydecisions. Second, convergence of the medium-term goal to the long-term goal should begradual since the costs of moving to the long-run goal cannot be ignored. Third, themedium-term inflation goal has always been defined as a number greater than zero. Fourth,if inflation expectations remain contained, there is no need to reverse prior price-level rises.16
The target for1975was a point target forCBMgrowth from December1974 toDecember 1975. Since this targer definition was susceptible to short-term fluctuations inmoney growth around year-end, the targets from 1976 to 1978 were formulated as pointtargets for the average growth of CBM over the previous year.In1979,two changes tothetargetformulationweremade.First,with theexceptionof1989, all targets have been formulated in terms of a target range of plus or minus 1 or 1.5percent around the monetary target derived from the quantity equation.In view of the oil price hikes in 1974 and 1979-80, the erratic movements in "real"exchange rates and the weakening of traditional cyclical patterns, it appearedadvisable to grant monetary policy from the outset limited room for discretionarymaneuver in the form of such target ranges. To ensure that economic agents areadequately informed . .. the central bank must be prepared to define from the start asdefinitely as possible the overall economic conditions under which it will aim at thetop or bottom end of the range. (Schlesinger 1983, p.10)In moving to a target range rather than a point target, the Bundesbank believed that, bygiving itself room for response to changing developments, it could hit the target range; infact, the tone of its explanation suggests that it was conferring some discretion upon itselfrather than buying room for error in a difficult control problem.The second change made in 1979 was to reformulate the targets as growth rates ofthe average money stock in the fourth quarter over the average money stock in the previousyear in order to indicate "the direction in which monetary policy is aiming more accuratelythan an average target does" (Deutsche Bundesbank 1979b, January, P. 8). Chart 3 (p.98)depicts quarterly growth rates of CBM (through1987)and M3 (thereafter)over thefourth-quarter level of the previous year and the targets since 1979 (the earlier targets are omittedbecause they were not formulated in terms of year-on-year rates).The Bundesbank has repeatedly stressed that situations may arise where it wouldconsciously allow deviations from the announced target path to occur in order to supportother economic objectives.These allowances are beyond and in addition to those implicit inthe setting of a targer range and of a gradual path for movements in unavoidable inflationA case in point is the year 1977, when signs of weakness in economic activity, combinedwith a strong appreciation of the deutsche mark, prompted the Bundesbank to tolerate theovershooting of the target. As said at the time:However, the fact chat the Bundesbank deliberately accepted the risk of a majordivergence from its quantitative monetary target does not imply that it abandonedthe more medium-term orientation which has marked its policies since 1975. . ..There may be periods in which the pursuit of an "intermediate target variable," asreflected intheannouncedgrowth rateof thecentral bankmoneystock,cannotbe given priority. (Deutsche Bundesbank 1978a, p.22)The main reason why CBM was initially chosen as the target aggregate was theBundesbank's perception of CBM's advantages in terms of transparency and communicationto the public.The Bundesbank explained its choice of CBM in the following words:17
[CBM) brings out thecentral bank's responsibilityformonetary expansion especiallyclearly.The money creation of the banking system as a whole and the money creationof the central bank are closely linked through currency in circulation and the banksobligation to maintain a certain portion of their deposits with the central bank.Central bank money,which comprises these two components, can therefore readilyserve as an indicator of both. A rise by a certain rate in central bank money shows notonly the size of the money creation of thebanking system but alsothe extent towhich the central bank has provided funds for the banks'money creation. (DeutscheBundesbank1976a,p.12)Although at any point in time CBM is a given quantity from the Bundesbank'spoint ofview because of the minimumreserve requirements,the choice of CBMnevertheless also reflects the monetary policy stance in the recent past. It is worth notingthat this use of CBM to publicly track the monetary stance is consistent with theBundesbank's focus on having minimum reserve requirements (as seen in the Bank'sadvocacy of such requirements for the unified European currency). The information beingconveyed by CBM in this context, however, is not so much to prevent either the public orthecentral bank from makinga largemistakeabout the unclear stance of monetary policy(a major concern in the framework design of inflation targeters such as Canada), but to giverapid feedback about the state of monetary conditions in general. The mindset is thatmonetary control provides useful information about policy and lowers policy uncertainty.The Bundesbank's confidence that it can explain target deviations and redefinitionsto the public is teflected in the design of its reporting mechanisms. There is no legalrequirement in the Bundesbank Act or in later legislation for the Bundesbank to give aformal account of its policy to any public body. The independence of the central bank inGermany limits government oversight to a commitment that "the Deutsche Bundesbankshall advise the Federal Cabinet on monetary policy issues of major importance, and shallfurnish it with information upon request" (Act Section 13). The only publications that theBundesbank is required to produce are announcements in the Federal Gazette of the settingof interest rates,discount rates,and the like (Act Section 33).According toAct Section18,the Bundesbank may at its discretion publish the monetary and banking statistics thatitcollects.The Bundesbank chooses to make heavy use of this opportunicy. On the inside frontcover, the Montbly Report is described as a responseto Section 18 of the Bundesbank Act, butit does much more than report statistics. Every month, after a"Short Commentary" onmonetary developments, securities markets, public finance, economic conditions, and thebalance of payments,there appear two to fourarticles on a combination ofonetimetopics(for example,"The State of External Adjustment after German Reunification")andrecurring reports (for example,"The Profitability of German Credit Institutions"[annual]and"The Economic Scene in Germany" [quarterly]). Each year in January, the monetarytarget and its justification are printed (between 1989 and 1992, the target and justificationwere available in December).The Annual Report gives an extremely detailed retrospective of18
economic, not just monetary, developments in Germany for the year, lists all monetarypolicy moves, and offers commentary on the fiscal policy of the federal government and theLander." Between these two publications, and regularly updated"special publications" suchas The Monetary Policy of the Bundesbank (an explanatory booklet), no Bundesbank policydecision is left unexplained with respect to both its immediate impact and its short- andlong-termeffects.The Bundesbank's commitment to transparency does not come without self-imposed limits on its accountability.Two limitations in particular provide a strongcontrast to the inflation report documents prepared by central banks in Canada, theUnited Kingdom, and other countries in recent years. First, no articles in the Montbly Reportare signed eicher individually or collectively by authors, and the Annual Report has only abrief foreword signed by the Bundesbank President (although all Council members arelisted on the pages preceding ic). Speeches by the President or other Council members arenever reprinted in either document. This depersonalization of policy is to some extent madeup for by the enormously active speaking and publishing schedule that all Councilmembers (not just the President and Chief Economist)and some senior staffers engage in,but the fact of depersonalized reports still weakens the link between the main policystatements and the responsible individuals.The second limitation on accountability is that the Montbly Report and the AnnualReport always deal with the current situation or assess past performance?no forecasts ofany economic variable are made public by the Bundesbank, and private sector forecasts oreven expectations are not discussed. The Bundesbank makes itself accountable on the basisof its explanations for past performance, but it does not leave itself open to be evaluated as aforecaster.In fact, its ex post explanations, combined with its potential GDP and normativeinflation basis for the monetary targets, enable the Bundesbank to shift responsibility forshort-term economic performance to other factors at any time.Nevertheless, those samemonetary targets are seen by the Bundesbank as the main source of accountability andtransparency because they commit the Bundesbank to explaining policy with respect to abenchmark on aregularbasis.GERMAN MONETARY POLICY UNDER MONETARY TARGETINGThe history of the German experience with inflation and monetary targeting up until 1990has been discussed elsewhere (for example,see Bernanke and Mishkin [1992] and Neumannand von Hagen [1993]). Rather than review the entire history of German monetarytargeting, we start by highlighting events through the 1970s and 1980s that areillustrative of certain themes discussed above-particularly the treatment of themonetary targets not as rigid rules but as a means of structured transparency for monetarypolicy.Then, the bulk of our discussion focuses on the challenging episode of Germanmonetary unification. In that instance, the Bundesbank successfully handled a (bydefinition) onetime inflationary shock of great magnitude and politically sensitive19