Monetary policy in uncharted territory
Speech by Professor Otmar Issing, Stone Lecture, 3 November 2003, London
Richard Stone contributed since the early 40s to the development of national accounts systems. Sixty years later it is clear that national accounts systems have greatly contributed to empirical and policy-relevant economic research. Timely and accurate statistics are, in Europe and elsewhere, as important as at the time of Stone's original contributions. Even more remarkably, these contributions, which were part of the dissemination of Keynesian ideas, have been accepted and used by all schools of economic thought. They are, in this respect, "robust" analytical tools, of the sorts which are mostly appreciated and needed in all fields of economics, but especially in monetary policy analysis. It is a pleasure and a privilege for us, Vitor Gaspar and me, to give the Stone lecture for 2003.
My plan is to elaborate on the difficulties of designing monetary policy under uncertainty, with special emphasis on those created by the occurrence of unprecedented historical events. As a former professor of monetary economics, my first instinct in these instances is to rush back to the comfortable world of academic research, where one has the freedom to draw a line and only discuss issues that fall on one's favourite side of that line. I suppose I still have a soft spot for the beauty of impeccable demonstrations descending from one's postulated assumptions by force of logical consistency. As a policy-maker for 13 years now, however, I have also learnt that real world economies are rarely seduced by that notion of beauty. When formal elegance becomes an end - rather than a means to an end - for theoretical research, theory risks being of little help as a guide for practical decision making. Thus, almost 300 years ago, Adam Smith warned: "The man of system (...) is often so enamoured with the supposed beauty of his own ideal plan of government, that he cannot suffer the smallest deviation from any part of it. (...) He seems to imagine that he can arrange the elements of a great society with as much ease as the hand arranges different pieces upon a chessboard. He does not consider that the pieces upon the chess board have no other principle of motion besides that which the hand impresses upon them; but that, in the great chess board of human society, every single piece has a principle of motion of its own, altogether different from that which the legislature may choose to impress upon it". 
In this lecture, I will try to illustrate how "science" based on academic theories and evidence and the "art of central banking" can usefully interact to produce superior outcomes in the field of monetary policy. I will argue that some fundamental results of monetary theory, which are robust across empirical analyses and theoretical models, must be placed at the core of any monetary policy strategy. I will also argue, however, that the filter of experience and judgement is necessary to bridge the gap from the naivety of available theoretical models to the complexities of real world economies. This mix of theoretical foundations and pragmatic application is especially useful at times of extraordinarily high uncertainty, as in the case of historical transformations. I will illustrate these ideas with reference to two such events, which I had to face in my central banking life: German unification and the launch of the European Monetary Union.
2. The forms of uncertainty
How does monetary theory suggest central banks should deal with uncertainty? The answer is dependent on the particular form of uncertainty depicted in the analysis.
First, there is uncertainty on prevailing economic conditions. This is mostly related to the need of filtering from observable data the level of latent variables - like the output gap, equilibrium real interest rates, or various measures of excess liquidity - and to the process of identification and interpretation of the nature of the shocks driving observed economic developments. By and large, the recent literature seems to consider this source of uncertainty as easy to deal with, by virtue of certainty equivalence. The world in which certainty equivalence is known to apply, however, is one where the model of the economy is linear, shocks are additive and the policy- maker preferences are quadratic.  More importantly, certainty equivalence relies on perfect knowledge of the true model of the economy.
When this assumption does not hold, monetary theorists speak of model uncertainty. This, in turn, can take the form of parameter uncertainty, namely imperfect knowledge of the parameters which characterise elasticities and functional dependencies within any particular model. Brainard's so-called conservatism principle is probably the most widely quoted result on how to deal with parameter uncertainty.  A large body of recent evidence, however, has pointed out that conservatism is not a robust outcome.  An aggressive, rather than cautious, response to inflationary shocks, for example, is advisable when there is uncertainty on the exogenous degree of inflation persistence. 
More fundamentally, model uncertainty implies ambiguity of the model which should provide a suitable description of the structural relationships in the economy. On this issue, I will simply quote a distinguished researcher like Ben McCallum,  who states: "it is not just that the economics profession does not have a well-tested quantitative model of the quarter-to-quarter dynamics, the situation is much worse than that: we do not even have any basic agreement about the qualitative nature of the mechanism." Once all sources of state and model uncertainty are compounded, as is the case in the real world, monetary analysis can offer hardly any robust guideline on how policy should react to shocks of unknown nature. The lack of consensus on the most appropriate model of economic dynamics leads to conceptual ambiguities and empirical difficulties when trying, for example, to determine the best course of policy, to identify the relevant economic shocks and their persistence, to estimate unobservable variables.
But this is not the end of uncertainty. A third class of uncertainty is that due to the endogeneity of expectations. The central bank sometimes wonders about the reaction of economic agents and financial markets to its own policy decisions and announcements. Conversely, economic agents may be unsure about the precise motivations and actions of central banks and other economic agents. In the realm of monetary policy, the level of uncertainty stemming from endogenous expectations is mainly, and inversely, related to central bank credibility. Luckily, in this field we do have some robust policy results. In order to ensure credibility and anchor inflation expectations, a central bank should be independent and the goal of pursuing price stability should be attributed priority in its statute or mandate. A firm response to inflationary shocks is also necessary, especially if agents take into account the possibility that the anti-inflationary determination of the central bank may wane over time. Finally, the announcement of a monetary policy strategy also contributes to dispel uncertainty on the actions of the central bank.
All the aforementioned sources of uncertainty constantly play a role in day-to-day monetary policy making in all countries, but they reach exceptional heights at the time of historical events such as German unification and EMU. How can central banks deal with these situations? How did the Bundesbank and the ECB do it?
3. Pragmatic monetarism: the Bundesbank
In October 1990 I left university where I had spent all my professional life so far and was appointed member of the Executive Board of the Deutsche Bundesbank.
From the first day I was entrusted with the portfolio which had been in the hands of Helmut Schlesinger - at that time Vice President - for almost twenty years. This implied not only responsibility for the business areas Economics and Statistics but also the presentation of economic analysis and monetary policy conclusions to the Board and the bimonthly meetings of the Council ("Zentralbankrat"), i.e. the highest decision making body. Responsibility for the Monthly Report, the very influential publication of the Bundesbank was also part of the portfolio.
At that time the Bundesbank was already one of the most respected central banks in the world with a unique track record as an inflation fighter. The Bundesbank had been able to counter inflationary shocks, most of which were international, more effectively than other central banks. German year-on-year inflation never in fact reached double-digit figures in the seventies. Germany experienced a prolonged inflationary shock, but the so-called Great inflation that plagued the United States and many other countries never really took hold of Germany.
3.1 The framework
The institutional arrangement of the Bundesbank undoubtedly played a fundamental role in achieving these outcomes. The law on the Bundesbank provided independence to the central bank and mandated it with the goal of "safeguarding the stability of the currency". The Bundesbank had been very successful in communicating that this was to be understood as safeguarding the "internal stability", i.e. synonymous to maintaining price stability.
In December 1974, as the first central bank in the world, the Bundesbank pre-announced a target for the growth rate of money for the following year - more precisely, 8% for the central bank money stock in the course of the year. It repeated this practice every year until the end of being responsible for its currency the D-Mark, i.e. until entering European Monetary Union with the introduction of the euro.
At the core of the monetary targeting strategy was Milton Friedman's dictum that in the long run inflation is always a monetary phenomenon and that therefore developments of prices over longer horizons are determined by the expansion of money. The Bundesbank had never considered the idea of controlling any form of base money or of pursuing monetary targets over the short-term. In fact, over 24 years of monetary targeting, it adapted some details of its strategy several times, e.g. moving to announce target corridors rather than single-figure targets or, from 1988 onwards, choosing the broad aggregate M3 instead of the central bank money stock. Moreover, the Bundesbank did never react in a mechanical way to M3 developments. The term "pragmatic monetarism" was sometimes used to describe this strategy.
The strategy was based on an application of the quantity equation. It required that, for given price norm, the growth of the money stock adjusted for the long-term change in velocity should be in line with potential output growth. Thus, the derivation of the monetary target took trend variables and normative elements into account. The Bundesbank always considered its strategy of monetary targeting as a medium-term approach. Incidentally, an important role for money based on the quantity equation can be seen as consistent with the ideas of Sir Richard Stone. In his Nobel lecture,  Sir Stone emphasised that accounting can be useful "in describing and understanding society." He added that "by organising our data in the form of accounts we can obtain a coherent picture of the stocks and flows, incomings and outgoings of whatever variables we are interested in [….] and thence proceed to analyse the system of which they form part." Along these lines, the quantity equation is the accounting identity which provides a coherent picture of the "incomings and outgoings" of inflation. For the Bundesbank, it formed the starting point for a broader analysis of future inflationary prospects, based on a set of monetary, financial and real economic indicators.
3.2 Monetary targeting in the aftermath of German unification
Up to 1990 the strategy as such had proven to work well and seemed superior to any available alternative. So, what could be easier than simply continuing with such a successful strategy? The first major challenge for the newcomer in the Executive Board was to prepare the document for the proposal for the monetary target for 1991 and present it to the Council. Should the case of German unification make any difference?
No doubt, the situation was substantially dissimilar from previous years. The introduction of the D-Mark on 1 June 1990, which preceded political unity on 3 October 1990, had influenced the environment for monetary policy in the direction of increased uncertainty. Whereas the introduction of the D-Mark in the still GDR was a logistical masterpiece, the conversion of GDR-Mark holdings into D-Mark initially led to an expansion in the money stock which had to be considered as overly generous. The money supply in East Germany increased by an amount roughly equivalent to 15% of the West German money stock. However, based on the estimated national product of the former GDR at market prices, - which later turned out to be far too optimistic - an increase of the money supply of only some 10% would have been appropriate. But, this was the result of applying standards from West Germany in the past. Was this appropriate? In deriving its monetary target for 1991 — a 4 to 6 % band as in the previous year - the Bundesbank ignored this monetary overhang. Over 1991, as it observed a limited reversal in east German demand for money when households in the new Länder geared their portfolios to the longer term, the Council - for the first and only time - made use of the regular proviso when announcing the target and revised its target corridor to 3-5 % in the context of the mid-year review of the monetary target.
The greater uncertainty was related to the possibility, and if realised the extent, of a change in the trend of the demand for money in reunited Germany with respect to that in the old Federal Republic. Would reconstruction require increased monetary support, particularly in view of the East German enterprises' lack, or at least very inadequate provision, with own funds? The diverging trends in the individual sectors of the east, especially in the first years after unification, tended to support this conjecture, as they suggested a need for increased intermediation by the banking system between lenders and borrowers - a phenomenon known as the "straddle effect" which had been a subject for discussion since the sixties. One had also to expect that the demand for money might be affected by the dramatic swing in the current account, from a surplus of over 100 billion DM in 1989 to a deficit of over 30 billion DM in 1991. And finally, but pointing rather in the opposite direction, there was a marked difference between income and wealth in the east and the west. How could these factors be taken into account when estimating money demand for the broad aggregate M3?
The ultimate question was: could the Bundesbank still assume that the long-term stability of money demand would be unchallenged by the impact of reunification and allow for continuation of monetary targeting? An answer had to be given without the help of formal econometric evidence. It was only ex post that empirical studies by Scharnagl  showed that these disruptions in monetary developments were only of a temporary nature and did not fundamentally call into question the monetary targeting strategy.
A final aspect contributed to specific difficulties related to monetary targeting in reunified Germany. Unification led to a massive expansionary shock and to high deficits in public budgets. Indeed, the public sector moved from a balanced position in 1989 to a huge deficit, while public sector indebtedness, including the previously off budget funds, increased from 41% to 60% of GDP between 1989 and 1996.
As a consequence of these expansionary policies and strong wage increases, inflation in Germany quickly rose. In West Germany, it exceeded 4% in the second half of 1991 and the first half of 1992 and remained at levels above 2% until early 1994. When setting the monetary targets during this period it was obvious that inflation may temporarily exceed the normative rate set so far at 2 %. The Bundesbank's medium- term oriented policy would not aim to reduce the inflation rate in a short period of time by a dramatic restrictive monetary policy move, but it would try in the first place to prevent inflation from running adrift, not allowing financing of second round effects of price hikes driven by increases in direct taxation and bringing down inflation towards the norm over the medium term.
How to preserve credibility for monetary policy under these circumstances? The first option was, as had already happened at the beginning of the practice of monetary targeting, to derive the monetary target based on a rate for "unavoidable inflation" which had to be above 2 %. The risk implicit in this approach was that a rate declared "unavoidable" by the Bundesbank, and whose exact value was difficult to estimate ex ante, would set a kind of floor to inflation expectations and enter as such all kinds of nominal contracts. In this context, it had to be taken into account that the inflation rise following German reunification had been due to demand pressures while in the middle of the seventies and the early eighties oil price shocks had been driving inflation upwards. The advantage of the approach would be that of acknowledging ex ante the possibility of temporary increases in the rate of growth of money, because of the inflationary effects of reunification.
The other option was to stick to the normative rate of 2 % for inflation in the derivation of the monetary target. Given the orientation of bringing inflation down gradually, this implied that the Bundesbank would be ready, for a limited period of time, to accept a higher rate of monetary expansion than announced in its target. This kind of tolerable, near-future overshooting was obviously inconsistent with a dogmatic interpretation of the monetary targeting strategy. At the same time, however, it would convey the central role of the inflation norm in the strategy and the fact that this was not subject to short-run revisions simply due to the arrival of shocks - not even exceptional ones. This option could therefore be seen as more effective in anchoring expectations at a delicate historical juncture.
The Bundesbank chose the second approach giving a clear signal that, even under the specific circumstances of German unification, 2 % remained the norm for inflation. In its publications of all kinds the Bundesbank tried to explain the implications of this approach. To my - certainly partisan, but not necessarily biased - judgement, this decision convinced the public and financial markets of the Bundesbank's unchanged orientation towards price stability. It also contributed to the result that in the end inflation was overcome faster and with a smaller output loss than anticipated by most observers.
3.3 Towards EMU
Great concerns on the European level arose when the Bundesbank tightened monetary policy substantially to counter strong inflationary pressures after unification. Within the Exchange Rate Mechanism of the EMS, the impact of the Bundesbank's monetary policy was immediately transmitted to the other members. Exchange rates came under pressure, as these countries were not able to react accordingly by adjustment of other macroeconomic variables. Objections against, and criticism of, German monetary policy were soon being voiced. The argument was that the Bundesbank, as the de facto European Central Bank, should orient its policy not only to the situation in Germany but also with a view to the partner countries. This combined with arguments from inside Germany, by quite a number of politicians and also business leaders, to put aside the focus on price stability under the unique historical circumstances, until the major challenges by German unification had been mastered.
How did the Bundesbank react under strong pressure from inside and outside the country? The Bundesbank explained that there was a law with a clear mandate for the stability of the D-Mark, a domestic goal. Moreover, it was convinced that it would be wrong to see a contradiction between this orientation and the interests of the European partner countries.
May I just quote from a speech I gave on 5 June 1992 at the Paolo Baffi Centre for Monetary and Financial Economics of Bocconi University in Milan: "A central bank's reputation substantially results from its past policy record and its corresponding success in fighting inflation. This reputation provides the basis for the credibility of its current and future work. The European Central Bank is starting out into monetary union from scratch as it were; it has yet to earn a reputation by conducting a sound monetary policy. The European Central Bank can only win confidence in this difficult initial phase by "inheriting", as it were, the tradition of the Bundesbank and other major central banks with a high reputation. It will no doubt succeed in doing this only if the Community enters monetary union at a point in time characterised not only by virtual stability but also in which partner countries can look back on as long a period as possible of stable conditions.…… The Bundesbank therefore believes that, in pursuing its counter-inflationary policy, it is acting responsibly both nationally and internationally. The ambitious goal of replacing national currencies by a single common European currency by the end of this decade and of the century can only be met satisfactorily if such a step is backed by nations which are deeply convinced that in the final analysis inflation is not a means of really solving problems but an evil which at the end of the day merely compounds existing problems and creates new ones. It would be a dangerous illusion to hope that this lesson could be learnt overnight, as it were, namely with the beginning of the European Monetary Union."
One important aspect - supporting this view - has to be added. It was obvious that it would be very difficult to convince the German public of the need to give up the D- Mark for the sake of a single currency in Europe. Germans with the bad experience of their history were suspicious of any kind of "monetary reform" and, with all European enthusiasm which was prevalent, they were extremely reluctant to embrace the project of EMU. Any notion that a "European orientation" might imply less price stability would have been disastrous for the project of monetary union. In the end this might have undermined the acceptance of the single currency also in other countries as people were eager to have a common, but in the first place a stable, new currency. The German central bank had created a stability culture, which set a benchmark. Eventually, all European central banks embraced this culture of stability. Only against this successful experience it was feasible that the statute of the ECB was designed in the same spirit as the Bundesbank law, with the major elements of independence and a clear mandate to maintain price stability.
4. A stability-oriented strategy: the ECB
Hence, what about the experience of the Bundesbank's monetary policy as a recommendation, an advice for the future ECB's policy?
As early as 1993 I was invited to address the issue of "Monetary policy strategy in the EMU" at a conference organised by De Nederlandse Bank. Here I will not go into the details, but I tried to explain the special circumstances the new central bank would be confronted with, stressed the need for a clear concept - a strategy -, discussed the prerequisites for monetary targeting - stability of money demand etc. -, and came to the conclusion that the ECB should indeed adopt a strategy of monetary targeting. One major argument in favour of this approach was to transfer by that as far as possible credibility and reputation to the new institution which had to start without a track record of its own.
In the end, when the time came however, I did not propose monetary targeting as a strategy for the ECB. While the empirical evidence available in late 1998 suggested that the demand for M3 for a sample of currencies joining EMU was stable and that M3 growth displayed leading indicator properties for inflation, taking such an approach appeared to be very risky - in the end much too risky. Unavoidably, all these econometric studies were based on a period when monetary policy had not been in the hands of a unique institution. Hence, behavioural changes due to a regime shift after the start of Stage Three had to be given high probability.
Moreover, the pragmatic approach the Bundesbank could take towards monetary targeting without risking a loss of credibility was less suited for the ECB, a new institution which still had to built up its reputation. In the same vein, any significant change of the monetary policy strategy, such as a change of the key monetary aggregate soon after the start of monetary union, could have seriously hampered the ECB's credibility. Against this background, there were a number of good reasons to choose a broader-based and more robust monetary policy strategy.
4.1 The framework
The ECB's monetary policy strategy is nevertheless another illustration of a commitment to a procedural framework, which may overcome some of the limitations and risks associated with more narrowly defined monetary policy rules.
First and foremost, the strategy includes a clear commitment to the goal variable, i.e. the primary objective of price stability. In October 1998, the Governing Council declared that price stability would be defined "as year-on-year increase in the Harmonised Index of Consumer Prices (HICP) for the euro area of below 2%". In May 2003, after an extensive review of the strategy, the Governing Council clarified that, within the definition of price stability, monetary policy would aim to maintain the inflation rate below, but close to, 2%. This clarification aims to shed light on the ECB's commitment to maintain a sufficient "safety margin to guard against the risks of deflation", while also addressing "the issue of the possible presence of a measurement bias in the HICP and the implications of inflation differential within the euro area".
A fundamental aspect of the definition is that "price stability is to be maintained over the medium term". The medium term orientation shows two important general features of the strategy. The first is the choice to adopt a gradual approach in the return to price stability after a disturbance, hence avoiding any unnecessary volatility in output and interest rates. The second feature is an explicit acknowledgement of the existence of lags in the monetary policy transmission mechanism: shorter-term oriented policies would only have the consequence of introducing further noise in the economy.
The strategy also puts forward a robust approach, especially important given the need to cope with the particularly high degree of uncertainty and imperfect knowledge prevailing at the beginning of Stage Three of EMU. This is reflected in the choice to base the assessment of the risks to price stability on two analytical perspectives, referred to as "pillars": "economic analysis" and "monetary analysis". Accordingly, the stability-oriented strategy recognises the need for scrutinising all available information, based on a wide set of economic indicators, in order to improve policy decisions. At the same time, the strategy acknowledges the connection between money and price developments in the medium to long run as one of the most robust known economic relationships.
Naturally, the link between money and prices tends to be stronger over the medium to long term, whereas in the short run several factors may lead to unexpected shifts in the velocity of circulation. A variety of indicators of domestic and international cost and price factors often provide a reasonably accurate account of near-term price developments. Therefore, it seems natural to attribute a relatively more important role to the latter indicators in forming a judgement about short term price developments, while shifting progressively the emphasis on "money" as the horizon of the assessment lengthens and uncertainty widens.
There are obvious similarities between the ECB strategy and the Bundesbank's monetary targeting. In particular, the ECB also relies on money to keep a steady sense of direction for the medium term. This was a deliberate element of continuity with the past, which has arguably enhanced the initial credibility of the ECB.
At the same time, the ECB strategy is clearly different from monetary targeting in its attributing greater emphasis to the need to take all relevant information and models into account when formulating policy decisions. From the very beginning, the ECB strategy also made the medium term relevance of money more explicit, as conveyed by the announcement of a reference value for the rate of growth of M3 over the medium term, rather than a monetary target for the year ahead. Since the medium term trend assumptions underlying the derivation of the reference value for M3 are not expected to change frequently, in May 2003 the ECB announced its decision to discontinue its regular annual review of the reference value.
4.2 The experience so far
In the first few years of EMU, economic and monetary analyses provided useful, but difficult to extract, information in equivalent respects.
For example, the output gap, which plays a central role in neo-keynesian recommendations for monetary policy, has proven difficult to estimate precisely for the euro area - the same is true elsewhere. Two causes of this difficulty are well known. The first is that the output gap is a latent variable, which is never observed exactly over time. The second is that the gap is, from a theoretical viewpoint, an elusive concept. Its definition is clearly dependent on the model used to define the "equilibrium output" benchmark and various possible benchmarks have been proposed over time. Existing evidence shows that different definitions - for example the theoretically-consistent ones derived within general equilibrium models and the more empirically oriented ones based on less restrictive economic assumptions - yield substantially different estimates for the gap.
All in all, it seems prudent to look at the results produced by a wide diversity of alternative techniques, but it is difficult to see how the output gap could play in practice the fundamental role that it is sometimes attributed in theory.
Has monetary analysis provided easy-to-extract information? Not really. As expected, monetary analysis has proven to be also difficult, although these difficulties are of a different nature.
M3 growth, in particular, has had a bumpy ride over the first years of EMU, buffeted by a succession of temporary shocks due to special factors. At the beginning of 1999, these were related to the increase in uncertainty occurring at the beginning of Stage Three, involving inter alia a change to the new minimum reserve requirements system, which was partly reflected in portfolio shifts into liquid instruments. An additional factor that might have played a role in the high growth of M3 in the beginning of 1999 was the existence of statistical distortions in the data. Other instances of portfolio rebalancing emerged in the autumn 2001 period of financial market instability. The persistent weakness in global stock markets drove investors away from equity and induced a reallocation of portfolios of euro area non-MFI investors towards low-risk assets. This was further enhanced by the rise of uncertainty in financial markets in the wake of the 11 September terrorist attacks.
These portfolio shifts have so far been interpreted as having a temporary character. If shifts are temporary, they will likely be reversed in the future and M3 growth, ceteris paribus, should return to more moderate levels. This interpretation is partly based on the evidence that long-run money demand in the euro area remains stable, as confirmed by extensive studies based on euro-area-wide data, including recent ones conducted in the context of the review of the strategy.
As already mentioned above, after a 4-year period which allowed us to accumulate some experience, the ECB decided to conduct an evaluation of its monetary policy strategy whose outcome was announced on 8 May. An important part of the review was to come to an assessment of the experience with the implementation of the strategy in the first years of EMU. The evidence accumulated led the Governing Council to conclude that the experience with the two-pillar strategy has been successful.
More specifically, the Council noted that the strategy, in particular the quantitative definition of price stability, had been very effective in anchoring market expectations. The direct and indirect measures of expected inflation in the euro area have shown a remarkable stability in the last few years, at a level just below 2%. This is a very important result, perhaps the most important, for a central bank whose mandate is to maintain price stability over the medium term. It is especially remarkable because we know that reputation - "living up to one's words" - is the ultimate way to achieve credibility in the opinion of both academics and central bankers.  In this respect, the ECB's lack of a track record should have represented a crucial drawback in the quest for credibility. Thanks to its successful strategy, we can say today that the ECB was born credible.
Expectations remained anchored in spite of a considerable number and intensity of adverse price shocks that hit the euro area economy during the first years of EMU. Since 1999, we experienced economic or financial turbulence world-wide, a sequence of substantial oil price and food price shocks, a very persistent foreign exchange depreciation and, more recently, a marked correction in stock prices and a pronounced foreign exchange appreciation - not to forget two wars and the terrorist attack of September 11. Amidst such shocks, the ECB used its measured, medium-term oriented approach, which did allow for some deviation of the inflation rate above the upper bound of the definition of price stability. Looking back, a correction of current short-term inflationary trends remains necessary to bring our track record fully in line with our definition of price stability. Nonetheless, there is evidence that economic agents understand the medium term orientation of the monetary policy of the ECB and they deem it appropriate. The recent developments in the headline HICP and in short-term inflation expectations left long term expectations largely uninfluenced. The ECB preserved a degree of monetary and economic stability that would have hardly been conceivable before the advent of Monetary Union.
5. The lessons
My short account can only provide a very rough summary of the challenges for monetary policy I faced in my central banking career. For an academic joining a policy making institution, German developments in the nineties provided almost every day an opportunity to learn and gain experience. As it turned out later, my years at the Bundesbank were not the end of a professional career but - totally unplanned - also a preparation for an even greater challenge.
For an academic becoming a policy maker in a field where he had made research himself, immediate questions arise: To what extent are theoretical insights applicable in policy making? What are the traps of too naïve theorising on the one hand and of atheoretical pragmatism on the other hand? This list is easy to expand.
Niehans once formulated the warning towards the academics at the end of an excellent treatise on the theory of money:  "Economics should be under no illusion that central banking will ever become a science. Academic critics love to chide central bankers for their lack of a fully articulated doctrine of monetary policy, based on testable - and perhaps even tested - hypotheses. These critics mistake central bankers for what they are themselves, namely teachers and intellectuals. In fact, a good central banker is a doer and a politician, for whom even ambiguity and inconsistency may sometimes serve his purposes. His field of action is the ever-changing stream of economic history, where every day may pose new problems requiring new solutions….. This treatise may thus end on a role of humility: however far monetary theory may progress, central banking is likely to remain an art."
On the one hand this is an honest statement supported by powerful arguments by a leading academic. On the other hand, if misunderstood and misused, it opens a pandora box. Centuries ago Kant has already marked the other pole: "No man has the right to pretend that he is practically expert in a science and yet show contempt for theory without revealing that he is an ignoramus in his field." 
I think that from these two positions the message is quite straightforward: try to bridge the gap between theoretical insight and need for action in a thoughtful, responsible way. This is especially demanding for somebody responsible for organising research and assessment in the central bank and presenting the "collective wisdom" of economists to the decision making bodies. In this assessment, it is crucial to separate noise from underlying trends, while being cautious towards potential structural breaks. In considering policy actions, Friedman's dictum is still valid:  "in first place try to avoid major mistakes".
Theory gives very few robust, and convincing, qualitative results and very little quantitative guidance on how to deal with compounded uncertainty. If asked to take decisions - i.e. precise questions such as e.g. whether to raise the policy rate by 25 basis points, to leave it unchanged,- strictly grounded on model-based results, the policy maker would probably have to give up and conclude with a typical two-handed non-decision. In order to make progress, policy makers have to base their decisions on their best judgement. This will partly be guided by general and robust theoretical results, and partly be informed by their belief on the determinants and effects of a particular economic development.
Such pragmatic approach is characterised by a firm reliance on those few robust and general results which can be distilled from theory - the importance of credibility, the ultimate link between inflation and money growth, the existence of long and variable transmission lags. At the same time, pragmatism relies on the act of judgement to build the bridge from general theoretical results to practical actions, which have to be taken in highly uncertain and unknown circumstances.
The adjective pragmatic should therefore not be interpreted in a pejorative fashion, i.e. as monetary policy purely based on animal spirits and ignorant of any theory.
The Bundesbank's "pragmatic monetarism" was firmly grounded in the monetarist message that inflation is always and everywhere a monetary phenomenon. This implied a "rediscovery" of money and nominal variables, after decades in which prices had often disappeared from economic models. At the same time, the nature of the approach kept the Bundesbank away from a dogmatic and extreme applications of monetarism. As I have argued elsewhere,  pragmatic monetarism "constitutes a successful synthesis of a theoretical basis and practical implementation."
Similarly, the ECB's stability-oriented strategy is aware of, and exploits, the advances in monetary policy theory. Partly motivated by the theoretical emphasis on their importance, the ECB does publish its internal projections for future GDP growth and inflation. At the same time, however, the ECB remains sceptical towards the idea that all problems of practical monetary policy making have been solved by the current generation of theoretical models - and towards the conclusion that good monetary policy must invariably coincide with inflation targeting.
If I had to summarise the core messages from research to central bankers and from policy-makers to academics, I would claim that these are broadly consistent with each other. The can be stated as follows:
don't try tricks, don't try to be too clever;
keep steady, keep committed to your mandate even in exceptional circumstances;
say as much as you can of what you are going to do: announce a strategy;
don't be dogmatic, but follow a policy which is always in line with your strategy.
Theoretical insights and central banking experience are both fundamental components of good monetary policy making, but especially so at the time of historical events.
In order to deal effectively with these events, sound institutional arrangements play a central role. Historical evidence and theory agree in pointing out that central bank independence and a clear mandate for price stability are the basic elements of a sound institutional set-up. They provide the premises to establish credibility, to anchor inflationary expectations and ultimately deliver price stability and foster a stable macroeconomic environment.
We must never forget this message, nor ever take credibility for granted, even at times when price stability is established and there seem to be minor challenges ahead. Credibility is hard to gain, but it is easily lost. To be maintained, it requires continuous vigilance. If lost, it can be regained only at high costs to society.
 A. Smith (1759), Theory of Moral Sentiments.
 See, for example, L. Svensson and M. Woodford (2003), "Indicator variables for optimal policy," Journal of Monetary Economics 50, 1177-88.
 W. Brainard (1967), "Uncertainty and the effectiveness of policy," American Economic Review 57, 411-25.
 For a recent survey, see C. Walsh (2003), "Implications of a Changing Economic Structure for the Strategy of Monetary Policy", Proceedings of the Jackson Hole Symposium sponsored by the Federal Reserve Bank of Kansas City on Monetary Policy and Uncertainty: Adapting to a Changing Economy.
 For an example, see I. Angeloni, G. Coenen and F. Smets (2003), "Persistence, the transmission mechanism and robust monetary policy", Scottish Journal ofPolitical Economy, forthcoming.
 B. T. McCallum (1996), "Inflation targeting in Canada, New Zealand, Sweden, the United Kingdom and in general", NBER Working Paper No. 5579, p. 17.
 Richard Stone (1986), "Nobel Memorial Lecture 1984: The Accounts of Society," Journal of Applied Econometrics 1, 5-28.
 M. Scharnagl, 1998, "The stability of German monetary demand: Not just a myth," Empirical
Economics 23, 355-370.
 See A. Blinder (1999), "Central bank credibility: Why do we care? How do we build it?," NBER Working Paper No. 7161.
 J. Niehans (1978), The Theory of Money (Baltimore), p. 294.
 E. Kant (1793), Selections from Concerning the Common Saying: This May be True in Theory But Does Not Apply in Practice, in Allen Wood (ed.), 2001, Basic Writings ofKant, translated by Carl J. Friedirich, New York: Modern Library, pp. 417-8.
 M. Friedman (1968), "The role of monetary policy", American Economic Review 58, p. 12.
 O. Issing (1997), "Monetary targeting in Germany: The stability of monetary policy and of the monetary system," Journal of Monetary Economics 39, 67-79.