Examination of witnesses (Questions 120-139)|
TUESDAY 6 FEBRUARY 2001
120. This Committee spoke to the Congressional
Budget Office in Washington and the Federal Reserve in Washington.
They are quite clearly making forecasts ten years out. You may
argue that they are not terribly reliable but nevertheless they
are making those forecasts of trillion dollar surpluses. The driver
to all of that is quite clearly the fairly heroic assumptions
in the US about productivity. I am trying to get a sense from
you of the extent of that mode of thinking as adopted by the researchers
to the MPC and the kind of work they are doing.
(Professor Buiter) I hope not very far because these
numbers are clearly crazy.
121. They are clearly crazy. That is very helpful.
We will send a note to Alan Greenspan on that.
(Mr Flemming) I would like to pick up one hypothesis
that Professor Buiter overlooked and was alluded to, but only
tangentially, by Sir Alan. You may recall that he stressed "as
recorded" in looking at the comparisons. The fact is that
the statistical conventions differ somewhat on the two sides of
the Atlantic. That is relevant, for instance, to the proposition
that Professor Buiter made that there has not been a great boom
in the quantity of investment of the relevant kind, particularly
in the UK. That is partly a statistical matter. For a given change
in expenditure, particularly in IT, the US would allocate a larger
proportion to an increase in quantity and a small proportion to
an increase in price. That may be one reason, and also because
the US and the Fed are projecting figures that are going to be
generated by their conventions and not our conventions. That is
one important caveat. Then there is a question about comparison,
not only with the United States but across the Channel too, where
you find a pattern much more similar to ours and where, incidentally,
the statistical conventions are more similar. The Germans in particular
have done quite a lot of work on the impact of those conventions.
The other question is: given that there is a puzzle here, whose
job is it to address it? It is certainly true that the Monetary
Policy Committee should be thinking about it and should be monitoring
research and perhaps commissioning some. As you mentioned in your
question, the particular peg you have chosen to hang your question
on was an expression of puzzlement by the Treasury and that would
suggest that the Treasury might do it. It is also, of course,
very important to the Treasury's own policies and judgments on
fiscal policies. Equally, I do not think that the question ought
to be: should the Bank of England and the Monetary Policy Committee
be doing it or should the Treasury be doing it? What we want is
an open discussion and an open debate to encourage people everywhere
to address this problem and to try to work, to some extent competitively,
and to some extent collaboratively on narrowing the range of uncertainty.
As Professor Buiter said, this is not something where there is
going to be a clear answer. Even the economic historians still
debate what was happening 100 years ago, and that will be true
in 100 years' time.
(Mr Barrell) May I give a slightly more simplistic
view of the policy making process? Of course we should be doing
research on the new economy. It is very difficult to put a handle
on it but, as practical forecasters, we can look at what happened
in the US over the last six years and what happened in the UK.
In the UK on average recently inflation has been about or very
slightly below target and about or very slightly below forecast.
That is interesting information because if, for reasons that we
cannot get at, productivity was growing more rapidly, output would
grow more rapidly, and therefore inflation would probably be well
below target. In the US over the same period, since about 1995/96,
in each year from 1995 through to 2000 our forecast of US inflation
was 1 per cent too high year after year and so was that of almost
everybody else. It has taken us some time to acknowledge that
there has been a genuine change in productivity growth in the
US over those five years. One of the ways you can read that in
is by saying: "We were wrong on our growth and inflation
forecasts and perhaps there has been something going on in the
US." Conversely we can say: "We were round about right
on our inflation and growth forecasts for the UK, and therefore
not much seems to be happening that we were not anticipating."
The US has seen very different things with very much lower inflation
than anticipated, which is a signal something is happening temporarily
in productivity. The amount organisations can produce is suddenly
jacked up. Probably it has stopped moving up now but they have
suddenly gone through this rearrangement of the way they do things.
As a result, their inflation is lower and their growth higher.
The Monetary Policy Committee can read that into existing inflation
and forecast serious mistakes, like we were all making for the
Mr Ruffley: That is very helpful.
122. Do you think the transparency of the current
monetary framework could be improved and, if so, how? We can start
with Sir Alan, as you already have some views on record.
(Sir Alan Budd) These are my views and they are only
added to what I believe is an excellent system, possibly the best
system in the world for setting monetary policy. I had expected
there to be rather more in the way of individual accountability
by the nine members of the Monetary Policy Committee. Their votes
are recorded. The reasons why they vote in the way they do are
reported in the minutes. They are reported in fairly general terms
and they are reported anonymously. I have no particular problem
with these views being anonymous in the minutes but I think it
is perfectly reasonable that subsequently individual members should
be asked to explain why they voted in the way they did at a particular
meeting, and even more importantly, at a sequence of meetings,
so that we can test the extent to which they are acting consistently
and coherently. There was a time when I think the lack of transparency
became quite serious but that has got better since then. This
was in May of last year when there was an inflation report which
published a forecast in the usual way. The fan charts and what
is known as the best collective judgment gave a path for inflation
which had inflation at the end of the two-year period at the target
level of 2.5 per cent. But there was also a table 6B published
in the inflation report which showed how individual members, with
their own views, differed from that central view because they
differed about the path of the exchange rate, or they differed
about the path of productivity, or they differed about what was
going to happen to retail margins. These are perfectly reasonable
disagreements but if you summed these disagreements and could
believe that some people were at one extreme and some at another,
there appeared to be at the May meeting some people who thought
that inflation in two years' time would be 3 per cent and some
who thought that inflation in two years' time would be 2 per cent.
In policy-making terms, even given the uncertainty, that is a
very wide range of views of where inflation is going to be, yet
there was a unanimous vote not to change interest rates. That
puzzled me greatly. I had no way of explaining this myself and
I think it would have been helpful if individual members had been
asked which of those views they themselves held and why, despite
holding those views, they had decided that the appropriate policy
response was no change. In fact we know rather more about this
than we did then because the most recent inflation report has
a rather helpful box in it which explains how the forecasts relate
to policy decisions. It may well have been that the individuals
would have given the explanations at the time but at the time
I was simply confused about why the Monetary Policy Committee
was behaving in the way that it was.
(Professor Buiter) It is certainly possible for somebody
to be in the box twice. With inflation there could be one deviation
from the norm which would cause inflation to be higher and another
to be lower, more optimistic about productivity and more pessimistic
about the exchange rate. I do not think there is a logical problem
at all necessarily.
(Sir Alan Budd) No, but again I do not think there
is any reason why we should not have been told that.
(Professor Buiter) I agree but there are two aspects
to that. First, I think we could have at the end of the minutes
where the votes are recorded a short paragraph written on the
day the vote is taken because you do not want people to re-rationalise
their positions with the benefit of several weeks' hindsight.
There could be a short paragraph explaining why each individual
member voted that way. The Japanese do that to a certain extent.
They ask the dissenting members to write an account of the reasons
for their dissent at the end of the minutes. The second option,
of course, is to be very open. When I joined the MPC I always
expected the first question to be: Explain why you voted the way
you did. I was never asked that question. That really staggered
(Sir Alan Budd) That was really the point I wanted
123. Did you think it would be good for us to
do these things?
(Professor Buiter) Yes.
124. To be honest, I am not sure that is correct.
We have asked individual members why they vote a particular way.
(Professor Buiter) You did not ask me.
125. Perhaps we just thought we knew about you!
(Mr Barrell) Can I make the main distinction that
Sir Alan made between transparently and clarity? Something that
is transparent is not necessarily clear. For instance, a Swatch
watch is transparent but, I think, extremely difficult to read.
Clarity is much more important than transparency. Why the final
decision was made and what the reaction would be to situations,
like for instance the collapse in the US, are much more important
to know than individual votes and comments during the process
of decision-making. It is rather like the Cabinet taking collective
responsibility for their decisions. It may be that no individual
member of the Cabinet actually supported the final decision but
there is a reason for the decision the Committee makes. Committees
do different things from individuals. I think, rather than at
the bottom of the minutes having a paragraph, at the top of the
minutes have a page explaining the collective decision, why a
decision was made collectively and what would have happened if
different information had been available. Then, if anybody wishes
to go on and read what individuals said and what they voted on,
that may be of interest but it is the collective view that matters.
You have to remember that the UK framework is very distinct. It
is largely a panel of experts. These are people who do not have
separate interests. For instance, the Bundesbank operated quite
successfully for some years but one could never say that its operations
were transparent. You hardly ever found out what people thoughtthere
was a committee making decisionpartly because they had
regional interests. However, the Bundesbank's response function
was largely clear. People knew what it might do and that is very
important in a monetary policy framework, knowing what you might
do rather than knowing who voted for it. One might say the same
with the Fed. It is clearer what the Fed might do because of its
experience. Although we have needed perhaps some transparency
in the early stages of the Monetary Policy Committee, perhaps
it and other people should concentrate on the clarity of its responses
and the reasons for its responses rather than how each individual
contributed to a joint decision.
(Sir Alan Budd) Can I just say that that may be what
has happened but it is not what was intended. These are not committee
decisions in the normal sense of committee decisions, and they
are certainly not Cabinet decisions in which there may be dissent
in the course of reaching the decision, but there is not supposed
to be dissent thereafter. These are quite clearly individual votes.
They are added up to give a majority result and this is very distinct
for me from the way in which one might take committee decisions,
and they were intended to be.
(Professor Buiter) The reason the collective decision
comes out that way is because the Governor puts the proposition
that demands a majorityend of storyso there are
at least five in favour.
126. Mr Barrell has made various comments about
how it compares with other countries. Can you make more comments
about the transparency and clarity compared with those other countries
and whether it makes our system more or less effective than the
framework in other regimes around the world?
(Mr Barrell) One has to think of the institutional
setting in which something is constructed like the European Central
Bank, which I think would find it difficult to have an independent
panel of experts. I think also that it would be politically difficult
to publish the votes because the individual members of councils
have regional or country interests. They represent in some sense
their country. Publishing their vote might influence their behaviour
in ways that would be unwise. We are not in that situation. We
do not have representatives of Yorkshire and of Scotland and so
on, and therefore there is no regional pressure. In the US it
is the same. To an extent, governors have a regional interest.
One has to be a little more cautious in the US system. In the
UK one can be distinctive. We do not have to copy the ECB or the
Federal Reserve. We can build our own institutions and we can
have dissent because it is an academic panel. That is also quite
distinctive. If one has an academic panel, one can have all sorts
of decisions. It is harder when it is a political panel. We have
built a different institution. It may not be the best institution
we could build but it is definitely not the worst. It does seem
to be doing a great deal better than the institutions we have
had for the previous 25 years.
(Professor Buiter) It is not of course true that the
national governors on the ECB board represent regional interests.
They are mandated by the Treaty only to look at the EU-wide interests.
If they do represent national interests, they are violating the
constitutional mandate. Secondly, they do not vote. They could
vote but they do not. They have a consensus, a collective way
of taking decisions. I think the procedure is bad and unnecessary.
The argument that it would be awkward because of national political
pressures I think could be put the other way round. If there were
to be votesand of course there are notthe governments
of the capitals would know in five minutes of the vote. The pressure
would be there. The effect of accountability would not be there
because you could hide behind the fig leaf of formal responsibility.
I think the system, from the point of view of accountability and
openness, is worse than it need be.
127. That is in the European Cental Bank?
(Professor Buiter) Yes. The American system is a one-man
show. That is its weakness.
128. Does that make it transparent?
(Professor Buiter) Certainly it is pretty transparent,
though of course the chair is extremely opaque in his statements.
Yes, it is pretty predictable.
Chairman: We have discovered that for
129. To return to forecasting, John Flemming,
if the Bank no longer assumes the inflation targets to be unchanged
at the end of the two-year horizon, is it right to continue with
its policy of assuming that interest rates are also unchanged
from the forecast?
(Mr Flemming) As you probably know, this is an issue
that has concerned me considerably, It has not been easy. I do
not think it is one on which this group will necessarily agree.
I am not sure the problem is so much with a prospective change,
although I did suggest that there might be a change in the inflation
target. I do think that there is a problem, occasionally at least,
with the internal coherence of the formal assumption that the
forecast is generated by assuming unchanged interest rates. I
do not have the date in my head as well as Alan did with his problems.
There was an occasion when the famous chart of inflation expectations
was on target for two years out but it was essentially a crash
landing. Given what one knows about the sort of inertia in the
system, it was fairly clear that the next month or the next quarter's
forecast was going to show the inflation rate being considerably
above target. Therefore, it would be natural for someone to draw
the conclusion that interest rates were going to have to change
within the next quarter. It should have been natural for the members
of the MPC to make that assumption and to build in that in some
way. I do believe that there might be some advantage and that
it could be done. The traditional central banking view is that
we never make any statements implicit or explicit about prospective
movements in interest rates. I think that, in the context of a
structured arrangement with explicit forecasts and assumptions
for them, it might well be possible to do that. Indeed, I actually
argued this when the MPC was launched. I think their success in
holding inflation very close to the centre of the target range
reinforces it, that actually publishing inflation forecasts is
not very helpful. The natural assumption for someone to make now
is that inflation will be at the target rate. The thing that people
want to know about, and might want to know what the experts on
the Monetary Policy Committee think, is what the trajectory of
interest rates will have to be to deliver that. One could actually
turn the whole process upside down and abandon the inflation forecasts
and say: "We are going to assume that we are going to meet
our target"it would not be a bad assumption, actually"and
we will publish, with wide dispersement probably, a trajectory
of short-term interest rates which we think is compatible with
that delivery." That is an extreme position that I have taken
up from time to time. You will probably hear from my neighbours
why it is not a sensible way to go.
130. If I could press you a little on that,
in your article you point out that publishing an interest rate
forecast is not a statement necessarily of intent. How would you
deal with the problem of market perception that it might be seen
almost inevitably as a statement of bias?
(Mr Flemming) Of course, as you indicate, the Americans
have from time to time made a statement of bias. That did not
prove fatal. So if it were so interpreted, it would not matter
very much. Again, this is going back some time, but I did suggest
that with the dispersement that you would have in there, you would
actually be able to make some comparisons between the uncertainty
reflected in the MPC's forecast done on this basis and the uncertainty
about future interest rates reflected in certain derivative products
and options over future Treasury Bill contracts or whatever. Therefore
it would be impossible to see whether or not there was alignment
between the MPC's estimates of bias, as it were, or expression
of bias, and the expectations of the market. You may recall, again
in the early days of the MPC, that Deputy Governor Mervyn King
did take some satisfaction from the fact that most of the MPC's
policy changes, or rather interest rate changes, had been correctly
anticipated by the markets. If one did this, one would also be
able to see whether that was true not only in the primary market
but also in the markets for these derivative option instruments.
Coming back to what he said and I said then, the MPC could prove
to be boring to the second degree.
131. Sir Alan Budd, do you see any difficulty
(Sir Alan Budd) The difficulty that has always been
given when this sort of discussion is held about the recognised
problems of having a fixed interest rate assumption throughout
the forecast period is that of getting the Monetary Policy Committee
to agree on the path. I think what John Flemming has suggested
is that this, too, should be a fan chart. It would completely
change the way in which the Monetary Policy Committee conducts
its discussions compared with what it was when I was there, which
was about the path for inflation and the unchanged interest rate,
which would achieve a satisfactory inflation outcome. They would
have to somewhat re-think it. I am not saying this is impractical
but it would be a very different process I think from the one
that I shared. Perhaps it was changing at the time Professor Buiter
(Professor Buiter) First of all, I think it is important,
and it seems to be impossible to get this across, that the two-year
horizon is neither here nor there. The target is not aiming to
hit the inflation target two years from now. Two years is just
part of monetary folklore, that monetary policy will have its
maximal effect. I do not think that is true for the British economy,
if it ever was. So two years is just where the paper runs out.
It has no significance apart from that. I do agree that the constant
interest rate assumption is not a forecast, not a prediction,
not a commitment. It is at best simply a conditional projection.
It is one that makes very little sense because if we really stuck
to a constant interest rate, we know that the model would blow
up. All these models have this problem. At some point you have
to start responding to inflation in order to stabilise the economy.
Any constant or any exogenous path of nominal interest rates is
likely to lead to explosive behaviour. I would favour making explicit
what is implicit beyond the horizon the fact that we do respond
to deviation from the inflation target or of output from where
we think its capacity level or of unemployment from the natural
rate and produce inflation forecasts using a rule. It is hard
to get the Committee to agree on the rule but the current agreement
I think is largely illusory. We do not agree on the constant interest
assumption; it is just there for obvious reasons. Central bankers
do get wet feet about appearing to make a commitment. That is
something one has to get around because clearly you do not want
to commit yourself to any path or, even to any particular rule.
A best collective judgment type fudge can be very helpful. It
is just a question of educating the markets, and indeed also our
political masters, about what these forecasts mean, so you do
not get shot with the petard of your own forecast.
(Mr Barrell) Of course, I cannot know how the Monetary
Policy Committee or the Bank of England produces a forecast but,
as a practitioner, if I was asked to produce a forecast for constant
interest rates for two years, I could do it. In order to hit the
target, as Professor Buiter says, after two years and one quarter,
I have interest rates adjusting to get to the target but I do
not have to publish that. Everybody in my forecast is expecting
it and therefore it works. It works because it is something beyond
the forecast horizon and the Policy Committee, if it is using
models, may have done something like that, to have a sudden jump
in interest rates after the forecast. It might be better to take
the leap and smooth that jump in interest rates over the forecast
horizon we publish, with some degree of uncertainty and all the
normal caveats about our not knowing what is going to happen.
The problem for a central bank is that if the central bank says
the interest rate will go up in six months' time, everybody else
will say, "Why don't you do it now?" That is a very
difficult question to answer. I do accept they have constraints
that are different from the constraints that I as a forecaster
face, but the uncertainty bounds around the interest rate would
help them disguise their embarrassment about not moving now if
they thought it was necessary.
(Mr Flemming) Although I said it could turn everything
upside down, I am not sure that it need do so. One of the points
that emerged from my colleagues' comments is the importance of
emphasising the dispersement. In fact the question that I would
like to put to each member of the MPC, and then have an algorithm
that produced the fan chart, would be to give, say, at six-monthly
intervals over the two-year forecasting horizon their best estimates
of the quartile points, to get away from the central line, which
might indicate the bias. So you would ask each person: what is
the interest rate that is so high that there is only a 25 per
cent chance of its being higher than that and what is the one
that is so low that there is only a 25 per cent chance of its
being lower than that? Then you can combine the nine sets of figures.
I do not think it would be very difficult using an algorithm.
If it were explained, I think that would minimise any kind of
commitment or damage that would be done by what I believe would
be a practical procedure.
132. Could I ask each of you if you think it
is time for the Chancellor to reconsider the current definition
of price stability?
(Mr Flemming) I suggested that you might earlier.
It seems to me a slightly curious use of language. I believe that
the 2.5 per cent target when it was set was a very sensible level
at which to set it. A number of the developments which might make
it appropriate to consider lowering it now were then not reliably
to be anticipated. I think that it is slightly curious to define
2.5 per cent as price stability, but that is purely a linguistic
matter. The other issue that you may be alluding to is the question
of whether RPIX is the ideal indicator.
133. No, I was thinking about the target.
(Mr Flemming) In that case, I will stop there.
(Mr Barrell) In the remit that the Monetary Policy
Committee had, it was made clear that the target was set contingently
on existing inflation but for this Parliament; in other words,
it took account of an existing situation when inflation had been
high. I think 2.5 per cent cannot really be described as price
stability in the medium term and there would be a very strong
case for reducing the target to 1-ish for the next parliament,
without there being a significant tightening of monetary policy
during that process. There are numbers of reasons for changing
the target. Partly if what one wants is to encourage long-term
contracting with the knowledge that prices will be stable over,
say, a five or ten year contract in any real sense, then, 1 is
much better than 2.5 because 1 per cent inflation is round about
maybe the upper band of the quality drift one sees in products.
So prices are on average stable if we measure inflation to be
running at 1 per cent. I think we should be suggesting that there
should be some revision of the target at the beginning of the
next parliament and that revision should be downwards.
(Professor Buiter) I would not be in favour of that
for the following reason. It is true that 2.5 per cent RPIX is
probably a high rate of inflation if you want consistency with
price stability, even if you allow for the kind of bias that inevitably
creeps into that sort of index. First of all, changing the target
should only be done very infrequently; otherwise the Chancellor
by repeated changes in the target can regain discretionary power
over monetary policy. You do not have the instrument but you just
move the target around. I would do it infrequently. If he were
going to stick to this regime for the foreseeable future, at some
point a change to something lower than 2.5 per cent might be desirable,
say to 2 per cent, and not much less in my view because the problem
is that there is an asymmetry in the economy as normal interest
rates cannot go below zero. I think that is a constraint that
we have to respect. The reason I would not do that now is that
there will, I hope in the not too distant future, be an opportunity
to revise the target when Britain joins EMU. I would save my credibility,
my capital, for that day rather then spend it in intermediate
target changes will be good only for a few years and try to stick
to 2.5 for the next three years or so, however long it takes.
(Sir Alan Budd) I think I am perhaps slightly closer
to Professor Buiter than the other two. We know from experience
that 2.5 per cent on the retail price index excluding mortgage
interest payments seems to be equivalent to about 1.5 per cent
on the harmonised index of consumer prices used around Europe.
So we are not far from the 1 per cent that the European Central
Bank has in mind. One might at some stage close that gap. I very
much take Willem Buiter's view about the desirability of stability
in this inflation targeting activity.
134. As you know, the Treasury Committee has
a central role in this accountability process of the MPC. How
do you think we can improve our role? Sir Alan Budd I know has
views on this, so perhaps you could start.
(Sir Alan Budd) It is linked to the previous discussion
and indeed linked to the comment that Willem Buiter made, that
the minutes are the minutes and that is fine. That can be judged
as a good account of the way in which the Monetary Policy Committee
is making decisions. Individuals have voted individually and it
seems to me perfectly reasonable to ask them why they have done
so and which of the views expressed in the minutes they happen
to hold. The most obvious body to do this is your Committee. I
think that it could do this rather more rigorously and indeed
regularly than it has done so far. As I have pointed out, you
do not have all the members all the time; you take a subset of
135. It would be impossible to have all the
members, given that they are quite a long-winded lot!
(Sir Alan Budd) I understand. I think that all the
members of the Monetary Policy Committee must understand that
they are at risk, if you like, of having to explain to you why
they did what they did in one month and why they did something
else in another month. The obvious place for this to happen is
in this Committee.
136. That is very useful and clear. We will
ponder on that one. Does anybody else have views on this?
(Professor Buiter) First on procedures that could
be improved to make it easier for you to call us to account effectively:
there are the individual explanations in writing at the end of
the minutes by MPC members explaining their votes. Then the possibilityand
it would require a change in the lawof having real confirmation
hearings. This would only work of course if Parliament somehow
could commit itself, this Committee, not to turn it into a three-ring
circus. That requires restraint that I do not know is available.
It is a real issue because American confirmation hearings sometimes
have degenerated into a rather unseemly exercise. I think in principle
it would be valuable. For the rest, I think there is a real problem
of expertise. Parliament in general, the TSC being no exception,
is under-resourced and under-staffed and inevitably you do not
have a committee that is staffed by people who are monetary policy
137. If I may interrupt you there, in fact one
of our advisers is now on the Monetary Policy Committee.
(Professor Buiter) I know but it is still a very small
staff compared to what I think is necessary. On the substance
of monetary policy decisions, the MPC has been given far too easy
a ride by the Committee. We were given a far harder time to explain
ourselves by the other place.
138. Why do you think that is?
(Professor Buiter) They have more expertise.
139. Than is available to the Committee?
(Professor Buiter) Yes, and more time to concentrate
on it. You people have real jobs.
(Mr Flemming) May I comment on one of the latter points
that Willem Buiter made about real confirmation hearings? I am
not sure about that. I think that there is an important question.
Chairman: We are just coming on to confirmation.