Beware of Gaps in Our Knowledge
This paper by Jean-Claude Trichet, President of the European Central Bank (ECB), delivered at this year's symposium at Jackson Hole contains lots of interesting commentary about differences in monetary policy between the Fed and the ECB, and also on monetary policy more generally. So much so that it is not possible to cover it all in a single post and do the material justice. So I am going to focus on particular aspects of the paper in individual posts. Still, this post is a bit long. But there's an interesting graph as a reward...
There has been lots of talk about the pitfalls of relying on short-run data (e.g. here and here) to make decisions, particularly relying upon any single measure of activity, prices, or expectations. The remarks below highlight problems with this approach. The remarks by ECB president Trichet begin from the premise that credibility is the most important commodity a central bank has in its possession and it should not give up that credibility easily. To follow the often false signals provided by short-run data risks having to reverse the course of policy as new data arrives. Such reversals undercut central bank credibility and reduce the ability of markets to predict central bank behavior. Transparent and consistent behavior is essential in establishing credibility, and the predictability that results is essential in anchoring expectations. It may not be the particular policy rule the central banks follow that is important, but rather the certainty that transparency and commitment to a rule give, a point made elsewhere in the remarks.
When we use the term transparency, at least as I understand it, we mean more than simply explaining what the central bank is doing, more than a portal to watch what is happening. Transparency also means giving markets the ability to duplicate the central bank's decision-making process and accurately predict how the central bank will respond in a particular circumstance. If the central bank is tempted into following false short-run signals, then such predictability is undermined to the detriment of output stability. Even policy that seems appropriate based upon the data at hand could look foolish in hindsight as better data become available. And such data revisions are not trivial nor confined to a short time period after the data are released. As the remarks by ECB president Trichet note, substantial revisions of 1999 data are still occurring in 2004. A very informative example of the problems with relying upon short-run measures of activity, in the example here the output gap is the focus, is provided from the experience of the ECB. This section is part of a larger discussion:
Monetary Policy and 'Credible Alertness,' by Jean-Claude Trichet, President of the ECB at the Panel Discussion “Monetary Policy Strategies: A Central Bank Panel”, at the Symposium sponsored by the Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, 27 August 2005: A corollary: beware of false signals In my view, the statement that a reputation for consistent behaviour is an asset carries an important corollary with it. We should guard against over-reacting to indicators that might give a distorted picture of economic reality and thus promote policies that might be regretted – and hastily reversed – in retrospect. To be more specific, … we would ideally need accurate, quantitative, contemporaneous readings of the current pertinent economic, monetary and financial data. For instance, knowledge of the level of production that would be consistent with stable prices, the so-called “natural” rate of output, would be an important tool for assessing whether the stance of policy is broadly appropriate. However, estimates of the time-varying natural rate of output, and by implication measures of economic slack, are notoriously very imprecise. Even in hindsight, different estimation methods yield quite dispersed figures. Allowing for data revisions … only makes the uncertainty surrounding those measures more pervasive.
Now, acting too strongly on such indicators, only to be forced to reverse gear soon thereafter as new estimates become available, risks the danger of an erratic policy process. Importantly, it confounds the markets, which … try to make sense of our past behaviour using the wisdom that is only available in hindsight. If this gives indications that differ markedly from those which could be reasonably inferred in real time, ... Even prudent policy conduct could appear ex post – and be deplored – as evidence of incorrect decisions. This could drain the central bank’s stock of credibility.
All this is not new and revives memories of the 1970s. Were the spectacular policy mistakes of that unfortunate decade due to poor economic theory, poor policy or a biased representation of the economy? It was probably a blend of all of these factors, with the last one carrying substantial weight. Advances in theory since the 1970s and more sophisticated econometrics have not immunised the policy process from these potential pitfalls. And here is a concrete example of what I mean. The first estimates of the output gap are typically available during the reference year, and are then revised over subsequent years. It is important to note that the final estimate is not available for a very long period of time: 1999 estimates, for example, were still revised by non-negligible amounts in 2004. Revisions are also large: they can often affect the sign, as well as the magnitude of the estimates. In the case at hand, for example, the revised output-gap measure available to us today is positive in almost all estimates for 1999, 2000 and 2001, but it was estimated to be negative in real time [as shown in Chart 1].
(Click on graph for larger version)
Uncertainty is even higher for output gap projections, which are the only measures of output gap existing at the beginning of each year. For example, the real-time estimates of the output gap for 2005, published in spring by the European Commission, the IMF and the OECD, differ on average more than 25% from the output-gap projections that these institutions made at the end of 2004. And, if we add to the real-time measures of the output gap for 2005 the average size of revisions observed in the past, we can conclude that it may ex post turn out to be anything between -3.6% and +0.8%.
Clearly, output-gap measurement issues are not a fact of the past. Central banks also need to cope with this problem at present. I take this as a warning that the central bank should not rely on any simple indicator of economic slack in taking its policy decisions. To conclude this point I would say that there are, undoubtedly, great differences in terms of monetary policy strategy between us at the ECB and, to use Alan Blinder terms, the Greenspan’s Fed. ... It is true that, in that sense, we belong to two different schools of thoughts on each side of the Atlantic.
But there is an important point on which on the contrary we share exactly the same views. We both want to be as comprehensive as possible ... We both do not want to neglect any information ... We both do not want to rely exclusively on a single particular model of the economy, as sophisticated as it may be. I have myself said several times that the Governing Council of the ECB has no intention of being the “prisoner” of a single system of equations. We both highly praise “robustness”. There is no substitute for a comprehensive analysis of the risks to price stability that pays due attention to all relevant information. It stands better chances of anchoring policy and – through that channel – expectations in a durable fashion.
Posted by Mark Thoma on Sunday, August 28, 2005 at 05:04 PM in Economics, Monetary Policy |
Permalink
TrackBack (4)
Comments (1)
You can follow this conversation by subscribing to the comment feed for this post.