Heikensten: Monetary policy in Sweden and the United States
Thank you for your invitation to Lund University. I am looking forward to discussing monetary policy and its framework with you. It is not that often I get the chance to do that in academic circles.
In our efforts to develop monetary policy in Sweden over the last ten years or so we have often had reason to look beyond our own borders, to make comparisons and to try to learn from the approaches of other countries. On those occasions the monetary policy in the US has naturally been one of the key comparison norms. In the Swedish debate it has also been common to refer to US policy along the lines that all would have been better if we had acted like them. Against that background I think it might be interesting to discuss our policy in comparison to theirs. Personally, the issue has also taken on particular relevance since I returned from the Jackson Hole Symposium in the US, which this year paid tribute to the soon-to-retire Federal Reserve chairman, Alan Greenspan.
Broad consensus on fundamental principles
To begin with, it is important to underline that the differences that exist between Swedish and US monetary policy – or for that matter the policy of any other comparable country – do not apply to the more general principles for how monetary policy should be conducted. Today, there is wide international agreement on these principles. The crucial issue is that monetary policy be conducted in such a way that confidence that inflation will remain low is not shaken – that, as it is usually put, the economy does not lose its nominal anchor.
However, this is not a mechanical process, of course. Since the world is complex and unexpected events happen all the time, each decision-making occasion is unique. For that reason monetary policy has to contain elements of discretion, as we usually say; the decisions are taken on a case-to-case basis and are dependent on the circumstances that prevail at the time in question. It is a question of taking into account, in every situation, the consequences for inflation, but of also considering what the wider effects on the economy will be – to quite simply adopt a wise approach to stabilising inflation.
Sometimes extraordinary events occur too, events whose final consequences are difficult to take stock of and incorporate into traditional analyses and forecasts. The terrorist attacks in the US on 11 September 2001 are an example of such an event. In the wake of these attacks, many central banks around the world were quick to cut interest rates. The powerful response can be interpreted as a kind of “insurance” against a really weak outcome. On a number of occasions Alan Greenspan has said that monetary policy is very much about quantifying and managing various risks. The following quote refers to this and gives a good idea of how Greenspan views the task of monetary policy, a view that is consistent with my own.
“This […] framework emphasizes understanding as much as possible the many sources of risk and uncertainty that policymakers face, quantifying those risks when possible, and assessing the costs associated with each of the risks.” (1)
Differences and similarities in the frameworks
I am fairly convinced that most central banks would describe the general principles for their policy in roughly the same way I have done here. But that does not mean that there are not differences between countries when it comes to the details of their monetary policy frameworks. I would like to discuss some aspects of the specific frameworks for Swedish and US monetary policy where we differ from one another but also some aspects in which we are similar.
The inflation target
Perhaps the most obvious difference is the objective of policy. Unlike the US, Sweden has a numerical target for the inflation rate. The Federal Reserve, like the Riksbank, does have price stability as its objective, but it has chosen not to specify a target level for inflation.
In the US there was a shift in policy around 1980, when the Federal Reserve under Paul Volcker managed to bring down inflation from high, occasionally double-digit, levels. At that point inflation targeting did not exist anywhere, not even in the academic literature. The only really concrete experience was our own experiment with price stability in Sweden in the 1930s, based on Knut Wicksell’s price norm, but this of course was practically unheard of outside Sweden. Volcker referred in his argumentation to monetarist ideas, but in practice he conducted policy in a pragmatic, albeit firm, manner. Through this policy and that which has been conducted since, the Federal Reserve has managed to convince economic agents that inflation will be held at a low level, without having to formulate an explicit target for the rate of price increases.
One could see the policy shift in the US as a different way of giving the economy a nominal anchor than through the numerical target that we have used, even though the US case can be said to be an implicit rather than an explicit undertaking.
I think it goes without saying that both the strategy of an explicit inflation target and the US approach have been successful and that both currently enjoy considerable confidence. Against that background it is interesting – and perhaps also less well-known here at home – that in the United States a debate is nonetheless ongoing as regards whether it may be worth taking a further step and introducing an explicit target for inflation. (2)
One argument in the US debate has been that the adoption of an explicit inflation target would help to increase the transparency of monetary policy. That could include making policy more predictable in advance for economic agents. But it could also mean making the decisions easier to evaluate afterwards.
Another argument has been that an inflation target would contribute to anchoring inflation expectations even better(3) . The more firmly anchored these are, the lower the risk that, for example, a sudden rise in costs would feed through to other prices and the more expansionary monetary policy could be allowed to be in such a situation. So by further strengthening the confidence that inflation will be kept low, the introduction of an inflation target – and this might seem paradoxical – can increase rather than limit the room for manoeuvre of monetary policy.
Yet another argument has been that an inflation target would help to focus the political debate on what monetary policy in the long term actually can achieve – contain inflation – rather than on what it cannot accomplish – a long-term rise in economic growth and employment.
The arguments against have primarily revolved around the fact that the current regime has worked well. Other arguments are mainly political and not all that easy for a non-American to evaluate. They have to do with the risks of appearing preoccupied with inflation, as an “inflation nutter”, to borrow an expression from my British colleague Mervyn King. Some also are concerned that a discussion about the target would raise other issues in Congress that are not wanted on the table. But the main argument seems nonetheless to be “if it ain’t broke, don’t fix it”. And you can probably infer from this that if some form of inflation targeting had been invented and used in the US at the end of the 1970s, it is likely that the policy would have been enjoying the same high confidence as the present one.
Formulation of the policy objective
Another difference is how the Riksbank’s and Federal Reserve’s mandates have been formulated in the underlying legislation. The Federal Reserve has a dual mandate, which means that the aim of its monetary policy is to achieve both “maximum employment” and “stable prices”. In Sweden and many other countries with inflation targets the legislation has primarily focused on inflation. Goals for employment and output are either not specified or subordinate to the objective of price stability – one usually speaks of a so-called lexicographic order of the goals, where any goals for the real economy are to be fulfilled given that the price stability objective is not threatened.
The differences in how the central banks' mandates have been formulated have largely historical explanations. Many countries during the 1990s adopted new central bank legislation. This legislation has been influenced by the practical experiences and theoretical insights that evolved during mainly the 1970s and 1980s. During that time it became clearer that it was not possible through monetary policy to bring about a long-term rise in output and employment in a country. What central banks could do, on the other hand, was ensure price stability. In that way it was also possible to lay a foundation for an otherwise good economic performance. In addition to this came increased independence in central bank decision-making, which also was justified by practical experiences chiefly in Germany, and theoretical insights that the design of the institutional framework was of central importance for the capability to keep inflation at a low level.(4)
The emphasis on inflation in the legislation sent out a clear signal that policy would be focused in a considerably more decisive way than before on keeping inflation low and stable. My impression is that there also were concerns, not least in Sweden, that generally formulated targets for the real economy would be interpreted to mean that monetary policy also was responsible for the more long-term performance of, e.g. employment, which it naturally cannot be.
In this context it is also worth noting that in the preparatory works to the new Swedish legislation for monetary policy that entered into force in 1999, the assessment was that it was implied that the Riksbank, in its capacity as a public authority accountable to the Riksdag (the Swedish parliament), had the task of supporting the general economic policy that the Riksdag had adopted – to the extent that it did not conflict with the price stability objective. Consequently this obligation, it was said, did not have to be specified in the legislation.(5)
The developments in the US have followed a somewhat different course. The Federal Reserve’s dual mandate came about through an amendment to the Federal Reserve Act in 1977. The policy shift around 1980 that marked the starting point for the credible low-inflation policy that has been conducted since then was able to occur within the framework of the existing monetary policy legislation with dual objectives. That this was possible was most likely due to a combination of factors, which I already have discussed in some measure. Firstly, that inflation was increasingly being seen as the most acute problem. Another contributory factor was probably that the Federal Reserve Board at the time was decisive enough to tackle the inflation problem, despite the fact that the slowdown in inflation was highly likely to entail a sharp rise in unemployment for a period. Perhaps it could also be said that it reflects the fact that the US target formulation leaves fairly wide scope for the Federal Reserve to make its own interpretations of what needs to be prioritised.
Monetary policy in practice
So what are the practical implications of the differences in how the policy objectives have been formulated? One reason to ask this question is that sometimes in the Swedish debate the US solution with a dual mandate, where the central bank takes account of both employment and inflation, has been highlighted as a model approach. The discussion was particularly intense in connection with the drawing up of the new regulatory framework for the Riksbank, which came into force in 1999, although similar arguments have appeared in the debate in recent years as well.
The fact that the Riksbank legislation does not say anything about real economic objectives does not mean, however, that no account is taken of real economic developments. The Riksbank conducts what has been called a flexible rather than a strict inflation-targeting policy. So the focus is not only on stabilising inflation; rather, the policy is flexible in the sense that other factors are also taken into consideration. The approach that we take at the Riksbank in these respects has been set out in a clarification that we published at the start of 1999.(6)
This clarification says that we conduct monetary policy in such a way that certain deviations from the inflation target are accepted at the same time as we normally aim to bring inflation back to target within a period of two years. That can be explained both by the fact that we are not able to control inflation exactly in the short term and by the need for flexibility – the ability to take account of factors other than inflation, e.g. stability in the real economy. At the same time it is important that this flexibility does not lower the credibility of the inflation target in the longer run but that it can serve its purpose as a “nominal anchor”.
The two-year horizon can be seen as a restriction on how much consideration normally can be taken of other objectives than price stability, a restriction that the Riksbank itself – like the explicit inflation target – has set with a view to generating credibility for its policy of achieving price stability.
We use the word “normally” because in some cases and under some circumstances situations may arise when we have to have more flexibility if the overall economic outcome is to be acceptable. In extreme cases the deviations from the inflation target may be so large that it actually is not possible to return inflation to target within the usual two-year perspective. In general, the pace with which it is possible and desirable to return inflation to target after a deviation is dependent on the size of the shocks hitting the economy and what kind of shocks they are. During certain periods, e.g. when major shocks occur on the supply side of the economy, it therefore is possible that inflation will deviate from the two-per-cent target for a longer time.
The view that the differences are small in practice is also supported indirectly by the ongoing debate in the US, which I touched upon earlier. There, it has been claimed that the Federal Reserve gradually has begun to conduct a policy that essentially could be described as implicit inflation targeting.(7) The advocates of this view say that one benefit of going a step further and adopting inflation targeting officially – over and above what I mentioned earlier – is that it would be one way to ensure that the same measured policy that has been conducted under Alan Greenspan will continue to be conducted in the future. In other words it would essentially only entail a codification and institutionalisation of the current policy, which would make it less dependent on single individuals.
So, in my opinion, the differences between the formulations of the Swedish and US objectives are far greater on paper than in practice. In my opinion, it basically is a case of two different ways to describe the same kind of policy rather than a case of there being crucial differences in the policies themselves.(8) Mervyn King expressed this in a slightly different way when he said that flexible inflation targeting is “a generic form encompassing different monetary policy regimes as special limiting cases”.(9)
Openness and communication
In terms of the degree of openness and communication, however, there are distinct differences between monetary policy in Sweden and the US. Internationally, the Riksbank has been of the leading advocates among central banks when it comes to openness, especially during the latter half of the 1990s. As regards the Federal Reserve the opposite has rather been the case, at least in comparison with other industrialised countries. To give an example, it was not until 1994 that the Federal Reserve began to publish its interest rate decisions immediately after they were taken. But the Federal Reserve has also taken a great number of steps to increase openness and clarity. After 1994 there has been a gradual transition to more substantial press releases, which explain the interest rate decisions. Since 2002 the Federal Reserve also publishes how the individual members have voted as well as any reservations against the decision. From 2005 the minutes of the monetary policy meetings are published with a shorter lag than before – they are now published three weeks after the meetings. In scientific studies that try to determine the degree of openness, however, the Riksbank still ranks among the best and regularly a bit higher than the Federal Reserve. (10)
What then can we say about the differences in how policy is described in Sweden and the US? I would like to make the following rough interpretation. In somewhat simplified terms one could say that in the US monetary policy is described and communicated in a way that resembles the textbook model's description of a monetary policy with a dual mandate; the policy rate should be set in such a way that the expected development of inflation and the real economy overall “looks good”. This fairly general way to describe policy and its consequences gives the Federal Reserve extensive "room for manoeuvre". The other side of the coin is that policy sometimes can be perceived as unclear – it can be difficult to see exactly what it is trying to achieve. This problem would presumably be even worse if policy in recent years had not been so closely connected to one individual whose communication has been the market’s “law”. To put it another way: when you have a group of policymakers that also communicate their own positions on the policy rate, the value from a communication point of view of a clear analytical framework is probably even higher than in a system dominated by a single policymaker.
In Sweden the description and communication of monetary policy has the advantage that it can take the inflation target as its point of departure. Decisions regarding the policy rate can be explained and justified in terms of their ultimate aim to bring inflation towards the target. At the same time, there is, as I have mentioned, scope to take account of real economic developments in that inflation temporarily can be allowed to deviate from target. A central issue in this context is that the framework with an inflation target forces us in such situations to present good, clear arguments for our actions – we have to explain why we have chosen to deviate from target and in which time frame we intend to try to bring inflation back in line. Apart from what this means for our communication, it has been a useful process that ensures a better-planned position than would have been the result from a less coherent procedure; that is my conviction at any rate.
All in all, I think that this is a system that has worked well. The clarity and stability that the inflation target has entailed has presumably also been an important explanation for why we in Sweden have managed to avoid the fixation on one person that has characterised monetary policy in the US. Neither Urban Bäckström's decision to step down a few years ago nor my own recent decision has given rise to any serious speculation that the fundamental monetary policy strategy would be changed in the future.
At the same time, I should say that not everyone agrees that inflation targeting has made it easier to understand and interpret how monetary policy is conducted. For example, there is more general criticism at international level which says that an explicit inflation target makes central banks’ communication more obscure and difficult to understand as regards considerations other than those that have to do with inflation developments.(11)
In my opinion, the task facing the Riksbank is to further elucidate the principles set out in 1999. The discussion in the recent two years shows that monetary policy, despite the clarification in 1999, often is evaluated in a mechanical way, solely on the basis of the outcome for the CPI and UND1X.(12) This may be a reasonable point of departure for a discussion, but the discussion should be carried further and also deal with how the policy outcomes relate to those considerations that have been taken of various temporary price fluctuations such as the electricity price increases in 2002-2003. Real and financial developments should also be taken account of in those situations where they have influenced policy. Exactly how this should be done is something the Riksbank and those who evaluate us have reason to ponder. And work along these lines is currently ongoing at the Riksbank. At the same time I cannot imagine that this work will lead to any fundamental changes in the monetary policy strategy. Rather, it is a question of increasing transparency regarding how we work so as to further facilitate the understanding and predictability of policy and also to enable a more appropriate and effective evaluation.
Summary and concluding remarks
Allow me now to summarise what I have said and round off with a couple of final remarks.
There is broad consensus today regarding the fundamental principles that should characterise a well-designed monetary policy. Crucial in this context is that policy be conducted in such a way that confidence that inflation will remain low and stable is not shaken. However, the fact that there is wide agreement on the more general principles does not mean that there are no differences between countries in terms of the details of how policy is formulated. Differences between the US and Sweden exist as regards the formulation of the policy objective, both in that the US does not have an inflation target and that the trade-off between low, stable inflation and real economic stability is described differently. When it comes to openness the Federal Reserve, despite a number of recent steps towards increased openness, has not come as far as the Riksbank. But in spite of that, there are, in my opinion, considerable similarities in how the two central banks implement policy; in both cases the aim is long-term price stability and to balance that ambition against primarily real economic stability, but also against other risks, e.g. threats to financial stability.
The difficult task facing us at the Riksbank, and one we probably share with the Federal Reserve, is to be clearer about how we handle these considerations. The misunderstandings that have occurred sometimes have generally stemmed from this. The task is not simple, mainly because the world seldom repeats itself. Every situation tends to be a bit different. Nor do I know of any other central bank that has tackled the problem of trying to specify a clear strategy for how it weighs up these considerations and how it wants to be evaluated against that background. Personally, I regard this work to be the next big challenge for the Riksbank in the field of monetary policy.
Thank you.
Notes:
1) Greenspan, A. (2004), “Risk and Uncertainty in Monetary Policy”, American Economic Review 94 (May), 33-40.
2) See, for example, Bernanke, B.S. (2003), “A Perspective on Inflation Targeting”, Remarks at the Annual Washington Policy Conference of the National Association of Business Economists, Washington D.C. March 25 and Mishkin, F.S. (2004), “Why the Federal Reserve Should Adopt Inflation Targeting”, International Finance 7, 117-127.
3) See, for example, Gurkaynak, R., A.T. Levin and E.T. Swanson (2005), ”Inflation Targeting and the Anchoring of Long-Run Inflation Expectations: International Evidence from Daily Bond Yield Data”, Paper prepared for the Sveriges Riksbank conference on ”Inflation targeting: implementation, communication and effectiveness”, Stockholm, June 2005.
4) The theoretical insights are based on works by Nobel Laureates Finn Kydland and Edward Prescott; see, for example, Kydland, F. and E. Prescott (1977), ”Rules Rather Than Discretion: The Inconsistency of Optimal Plans”, Journal of Political Economy 85, 473-490.
5) See Ds 1997:50 Riksbankens ställning (the Riksbank’s position), pp. 73.
6) See Heikensten, L. (1999), “The Riksbank’s inflation target – clarification and evaluation”, Sveriges Riksbank Quarterly Review 1.
7) See Goodfriend, M. (2005), “Inflation Targeting in the United States?”, in Bernanke, B.S. and M. Woodford (eds.), The Inflation-Targeting Debate, The University of Chicago Press.
8) See also Svensson, L.E.O. (2002), “Monetary Policy and Real Stabilization”, in Rethinking Stabilization Policy, A Symposium Sponsored by the Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, August 29-31, 261-312.
9) See King, M. (1997), “The Inflation Target Five Years on”, Speech at the London School of Economics to mark the tenth anniversary of the LSE Financial Markets Group.
10) See, for example, Eijffinger, S.C.W. and P.M. Geraats (2003), ”How Transparent Are Central Banks?”, CEPR Discussion Paper Series No. 3188 or van der Cruijsen, C. and M. Demertzis (2005), “The Impact of Central Bank Transparency on Inflation Expectations”, Paper presented at the National Bank of Poland conference “Central Bank Transparency and Communication: Implications for Monetary Policy”, Warsaw, June 2-3, 2005.
11) See, for example, Fischer, S. (1996), “Why Are Central Banks Pursuing Long-Run Price Stability”, in Achieving Price Stability, symposium sponsored by the Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming. A recent reference that has brought the problem to the fore once again is Faust, J. and D.W. Henderson (2004), “Is Inflation Targeting Best-Practice Monetary Policy?”, Federal Reserve Bank of St. Louis Review 86(4), 117-143.
12) UND1X is an alternative measure of inflation that is defined as the CPI excluding household mortgage interest expenditure and the direct effects of changes in indirect taxes and subsidies.
