developments in the real economy through flexibility and communication. Closeexamination of this episode also illustrates how the Bundesbank has operated its monetarytargeting regime in the 1990s and provides a baseline for the three inflation targeters weexamine next. Charts 1-4 (pp. 97-8) track the path of inflation, interest rates, monetarygrowth, GDP growth, and unemployment before and after monetary union.It is fair to generalize that in the 1970s and 1980s the Bundesbank frequently over-and undershot its annual monetary targets; it reversed overshootings in most but not allcases. In addition, the Bundesbank responded to movements in other variables besidesinflation. From the beginning of CBM targeting in 1975, the Bundesbank was aware of therisk that"central bank money is prone to distortions caused by special movements incurrency in circulation" (Deutsche Bundesbank 1976a, p. 11). In 1977, the Bundesbankallowed CBM growth to exceed the target in the face of an appreciating deutsche mark andweak economic activity.10 At that early time, only two years after the adoption of thetargets, the Bundesbank relied on the power of its explanation that"there may be periods inwhich the pursuit of an 'intermediate target variable' .. . cannot be given priority,"acknowledging the importance of intervening real and foreign exchange developments inits decision making (Deutsche Bundesbank 1978a, p.2).In 1981 and early1982, CBM grewmuch more slowly than M3 because ofweakness in the deutsche mark, leading to large-scale repatriation of deutsche mark notesand an inverted yield curve that caused portfolio shifts out of currency into high-yieldingshort-term assets. Accordingly, the monetary target for 1981 of 4 to 7 percent wasundershot (Chart 3, P. 98); since during this period the Bundesbank was pursuing adisinflationarycourse,and pro-gress wasbeing madeon the inflationfront,thecentralbank didnot act to bring money growth up into target range.In 1986 and 1987, the reverse situation—a strong deutsche mark combined withhistorically low short-term interest rates-led to CBM growth of 7.7 percent and8 percent, respectively, while M3 grew ac 7 percent and 6 percent during those twoyears, so that all measures exceeded the target monetary growth range. The Bundesbank'sallowance of this overshooting could be seen as part of the results of the Plaza Accord on theGroup of Seven exchange rates as well. The latter development prompted the Bundesbankto announce a switch in 1988 to monetary targets for the aggregate M3:The expansion of currency in circulation is in itself of course a significantdevelopment which che central bank plainly has to heed. This is, after all, the mostliquid form of money...and not least thekind of money which the central bankissues itself and which highlights its responsibility for the value of money. On theother hand, especially at times when the growth rates of currency in circulation anddeposit money are diverging strongly,there is no reason to stress the weight ofcurrency in circulation unduly.(Deutsche Bundesbank1988b,March,"MethodologicalNotes on the Monetary Target Variable'M3,"pp.18-21)The fact that the Bundesbank changed the target variable when CBM grew too fast,but did nor do so when it grew too slowly, can be interpreted as an indication of the20
importance chat the Bundesbank attaches to the communicative function of its monetarytargets. Allowing the target variable to repeatedly overshoot the target because of specialfactors to which the Bundesbank did not want to react might have led to the misperceptionon the part of the public chat the Bundesbank's attitude toward monetary control andinflationhadchanged.llAn econometric argument has been made by Clarida and Gertler (1997)that theBundesbank has displayed an asymmetry in reacting to target misses; that is, it usuallyraises interest rates in response to an overshooting of the target, but it does not lowerinterest rates in response to an undershooting. In any event, the switch in targetedmonetary aggregates was not accompanied by any other alterations in the monetaryframework,and cheperceivedneed for the switch did notseemto occasion much concernIn short, as long as the underlying inflation goal was met over the medium term, theexistence of the monetary targets rather than their precise functionality was sufficient.As noted in the previous section's discussion of unavoidable price increases (latertermed normative levels of price increase) underlying the Bundesbank's monetary targets,the Bundesbank has tended to pursue disinflation gradually when inflationary shocksoccur. The Bundesbank's response to the 1979 oil-induced supply shock was very gradualand publicly stated co be sothe Bundesbank set its level ofunavoidable price inflation for1980 at 8 percent, clearly below the then-prevailing rate, but also clearly above the level ofprice inflation that was acceptable over the longer term. The target inflation level wasbrought down in stages, eventually returning to the long-run goal of 2 percent only in1984. Even though the underlying intent was clear, each year's target unavoidable inflationlevel (as well as the monetary target and interest rate policies determined by that level) wasactually set only a year ahead, allowing the Bundesbank still further flexibility to respond toevents and to rethink the pace of disinflation. Although what turned out to be four years ofmarked inflation reduction is hardly an instance of the Bundesbank going easy on inflation,it is an illustration of flexibility and concern for the real-side economic effects of GermanmonetarypolicyThe economic situation in the Federal Republic of Germany during the two yearspriorto economic and monetary union with theGerman Democratic Republic (GDR)onJuly 1, 1990, ("monetary union") was characterized by GDP growth of around 4 percentand the first significant fall in unemployment since the late 1970s (Chart 4, P.98). After aprolonged period of falling inflation and historically low interest rates during the mid-1980s, inflation had increased from-1 percent at the end of 1986 co slightly more than3 percent by the end of 1989. The Bundesbank had begun tightening monetary policy inmid-1988, raising the repo rate in steps from 3.25 percent in June 1988 to 7.75 percentin early1990.After thefirst M3 target of 3to6percent had been overshot in 1988 by1 percent, the target for M3 growth of around 5 percent in 1989 was almost exactlyachieved, with M3growing at 4.7percent.M3growth was certainly not high in view oftheprevailing rateof economicgrowth.21
InresponsetotheuncertaintiesresultingfromtheprospectofGermanreunification, long-term interest rates had increased sharply from late 1989 until March1990, with ten-year bond yields rising from around 7 percent to around 9 percent in lessthan halfa year.Combined with a strong deutsche mark, this rise in long-term interest ratesallowed the Bundesbank to keep official interest rates unchanged during the monthsimmediately preceding monetary union.In the immediate aftermath of monetary union itkept official interest rates unchanged as well, despite the fact that the effects of themassively expansionaryfiscal policyaccompanying reunification werebeginning to propelGDPgrowthtorecord levels.To some extent,the Bundesbank's decision to keep official interest rates unchangedfor thefirst few months following monetary union was due to the fact that the inflationarypotential resulting from the conditions under which the GDR mark had been convertedinto deutsche marks was very difficult to assess.The Bundesbank had been opposed to theconversion rate agreed to in the treaty on monetary union (on average about 1 to 1.8) andhad been publicly overruled on this point by the federal government.12 The moneystock M3 had increased almost 15 percent because of monetary union. The rate ofconversion chosen turned out to be almost exactly right. While GDP in the former GDRwas estimated to be only around 7 percent of the Federal Republic's once reunification tookplace, wich the vast government transfers to the east all of the money was absorbed (seeKonig and Willeke [1996).During the first few months following monetary union, theBundesbank was preoccupied as well with assessing the portfolio shifts in east Germany inresponse to che introduction nor only of a new currency, but also of a new financial systemand a broad range of assets that had not previously existed there.As the east German banks were adjusting to their new institutional structure,andvelocity was destabilized by portfolio shifts in east Germany, monetary data that includedeast Germany were hard to interpret. The Bundesbank therefore continued during thesecond half of 1990 to calculate monetary aggregates separately for east and westGermany, based on the retutns of the banks domiciled in the respective parts. AlthoughM3growth in west Germany accelerated in late1990 as a result of the moderate growthrates during the first half of the year, growth of M3 during 1990 of 5.6 percent was wellwithin the target range of4to6percent.During the fall of 1990, the repo rate had approached the lombard rate, whichmeant that banks were increasingly using the lombard facility for their regular liquidityneeds and not as the emergency facility for which the Bundesbank intended lombard loanstobe used.OnNovember2, 1990, theBundesbankraised thelombard ratefrom8 to8.5 percent as well as the discount rate from 6 to 6.5 percent. Within the next few weeks,however, banks bid up the interest rate (Mengemtender), and the repo rate rose above thelombard rate, prompting the Bundesbank to raise the lombard rate to 9 percent as ofFebruary 1, 1991. With these measures, the Bundesbank was reacting to both the volatileGDP growth rates and the faster M3 growth in the last part of 1990. Inflation had untilthen remained fairly steady, but it seems likely that the Bundesbank at that point was22
probably expecting inflationary pressures to develop in the near future given the fiscalexpansion, the overstretched capacities in west Germany, and the terms of monetary union.At the end of 1990, the Bundesbank announced a target range for M3 growth of4 to 6 percent for the year 1991, applying a monetary target for the first time to the wholecurrency area. The target was based on the average all-German M3 stock during the lastquarter of 1990. As this stock was stll likely to be affected by ongoing porfolio shifts in eastGermany, the target was subject to unusually high uncertainty. It is worth noting thatneither the basic inputs into the quantity equation that generates the Bundesbank's moneygrowth targets' normative inflation nor the potential growth rate of the German economywas changed.13Following German unification, the monetary targets set by the Bundesbank weredecidedly ambitious as they left normative inflation, on which these targets are based, unchanged at 2% during this period, even though it was obvious from theoutset that this rate could not be achieved in the target periods concerned.(Issing1995a)This statement was one of policy-the reunification shock did not fundamentallyalter the basic structures of the German economy. Moreover, this statement communicatedto the public at large that any price shifts coming from this shock should be treated as aonetime event and not be passed on to inflationary expectations.This stance required faith in the public's comprehension of, and the Bundesbank'sability to credibly explain, the special nature of the period. It is important to contrast thisadherence to the 2 percent medium-term inflation goal with the Bundesbank's responseto the 1979 oil shock, when, as already noted, unavoidable inflation was ratcheted up to8 percent and brought down only slowly. There are two explanations for the difference inpolicy response in the 1990-93 period, neither of which excludes the other:first, themonetary unification shock was a demand rather than a supply shock, and so theBundesbank was correct not to accommodate it; and second, after several years of monetarytargeting, the Bundesbank's transparent explanations of monetary policy had trained thepublic to discern the differences between onetime price-level increases and persistentinflationary pressures. In any event, the Bundesbank was clearly allowing its short-termmonetary policy to miss the targets in pursuit of the longer term goal.Following the Bundesbank's target announcement stressing its continued adherenceto monetary targeting after reunification and the lombard rate increase on February 1,long-term interest rates started falling for the first time since 1988.In hindsight, it isapparenc that this was the beginning of a downward trend that continued until the bondmarket slump in early 1994. Although the highest inflation rates were still to come, at thispoint financial markets were apparently convinced that the Bundesbank would succeed incontaining, if not reducing, inflation in the long run. By making it clear that it would notaccommodate further price increases in the medium term, the Bundesbank bought itselfflexibility for short-term easing without inviting misinterpretation. This link betweentransparency and enhanced flexibility, of course, depends upon the central bank's23
commitmenc ro price stability being credible, but it emphasizes how even a credible centralbank may gain through institutional design to increase transparency.Until mid-August 1991, the Bundesbank left the discount and lombard ratesunchanged, while the repo rate steadily edged up toward che lombard rate of 9 percent. CPIinflation in west Germanyhad still remained around 3percent during the first halfof1991,while GDP growth remained vigorous. M3 growth, by contrast, was falling compared withits upward trend during late 1990,in part because offasterthan expected portfolio shiftsinto longerterm assets in east Germany.These portfolio shifts, as well as the sharper than expected fall in the GDR'sproduction potentiai, led the Bundesbank for the first cime ever to change its monetarytarget on the occasion of its midyear review. The target for 1991 was lowered by 1 percent,to 3 to 5 percent. The fact that monetary targets are rarely reset is critical to any changebeing accepted without being perceived as a dodge by the central bank.In this instance, the Bundesbank was able to invoke the implicit escape clause buileinto the semiannual target review. That formalized process, which required a clearexplanation for any shift in targets,gave a framework for the Bundesbank to justify itsadjustment.The discipline of the monetarytargeting framework displayed theframework'sdisadvantages as well: that is,the difficultyofmeeting short-run targets stemming from theinstability of money demand and the inability toforecast changes in the monetaryaggregate's relationship togoal variables.As the repo rate approached the lombard rate again, the Bundesbank, on August 16,1991, raised the lombard rate from 9 to 9.25 percent and the discount rate from 6.5 to7.5 percent.The discount rate was raised to reduce the subsidy character of banksrediscount facilities, which the Bundesbank had tolerated as long as the east German banksrelied mostly on rediscount credit for the provision of their liquidity.Despite the fact that GDP growth started to slacken during the second half of 1991,M3growth accelerated.To some extent, the faster gtowth of M3 was a result of the by-theninverted yield curve,which led to strong growth of time deposits and prompted banks tocounter the outflow from savings deposits by offering special savings schemes with attracciveterms, This period was the first time that the yield curve had become inverted since theearly 1980s and since the Bundesbank had been targeting M3. In this situation, the conflictarose for the Bundesbank that increases in interest rates were likely to foster M3 growth.This problem was all the more acute since banks' lending to the private sector was growingunabated despite the high interest rates, probably, to a large extent, because loan programswere subsidized by the federal government in connection with the restructuring of the eastGerman economy and housing sector.This conundrum, of the Bundesbank's instrument tending to work in the "wrong"direction, brought the underlying conflict of monetary targeting to the forethe targetmust be critically evaluated constantly in relationship to the ultimate goal variable(s)However, if the target is cast aside regularly with reference to changes in that relationship24